Food and Beverage Stability and Shelf Life (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food and beverage stability and shelf life ß Woodhead Publishing Limited, 2011 Related titles: Chemical deterioratio...

Author: Persis Subramaniam | David Kilcast

887 downloads 3696 Views 8MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Food and beverage stability and shelf life

ß Woodhead Publishing Limited, 2011

Related titles: Chemical deterioration and physical instability of food and beverages (ISBN 978-1-84569-495-1) For a food product to be a success in the marketplace, it must be stable throughout its shelf life. Changes due to food chemical deterioration and physical instability are not always recognised by food producers, who are more familiar with microbial spoilage, yet can be just as problematic. This book provides an authoritative review of key topics in this area. Chapters in Parts I and II focus on the chemical reactions and physical changes which negatively affect food quality. The remaining chapters outline the likely effects on different food products. Food spoilage microorganisms (ISBN 978-1-85573-966-6) Action by microorganisms is a common means of food spoilage and ensuring that a product has a suitable shelf life is a critical factor in food quality. With current trends towards less severe processing techniques, reduced use of preservatives and higher consumption of perishable foods such as fresh fruit and vegetables, the deterioration of foods by microbial spoilage is an increasing problem for the food industry. Methods to detect, analyse and manage food spoilage are reviewed in the opening parts of this collection. The following chapters focus on important yeasts, moulds and bacteria, their classification, growth characteristics and detection and the implications of these factors for their control in food products. Understanding and measuring the shelf life of food (ISBN 978-1-85573-732-7) The shelf life of a product is critical in determining both its quality and profitability. This important collection reviews the key factors in determining shelf life and how it can be measured. Part I examines the factors affecting shelf life and spoilage, including individual chapters on the major types of food spoilage, the role of moisture and temperature, spoilage yeasts, the Maillard reaction and the factors underlying lipid oxidation. Part II addresses the best ways of measuring the shelf life of foods, with chapters on modelling food spoilage, measuring and modelling glass transition, detecting spoilage yeasts, measuring lipid oxidation, the design and validation of shelflife tests and the use of accelerated shelf-life tests. Details of these books and a complete list of Woodhead titles can be obtained by: · visiting our web site at www.woodheadpublishing.com · contacting Customer Services (e-mail: [email protected]; fax: +44 (0) 1223 832819; tel.: +44 (0) 1223 499140 ext. 130; address: Woodhead Publishing Limited, 80 High Street, Sawston, Cambridge CB22 3HJ, UK) If you would like to receive information on forthcoming titles, please send your address details to: Francis Dodds (address, tel. and fax as above; e-mail: [email protected]). Please confirm which subject areas you are interested in.


Woodhead Publishing Series in Food Science,Technology and Nutrition: Number 210


Edited by David Kilcast and Persis Subramaniam


Published by Woodhead Publishing Limited, 80 High Street, Sawston, Cambridge CB22 3HJ, UK www.woodheadpublishing.com Woodhead Publishing, 1518 Walnut Street, Suite 1100, Philadelphia, PA 19102-3406, USA Woodhead Publishing India Private Limited, G-2, Vardaan House, 7/28 Ansari Road, Daryaganj, New Delhi ± 110002, India www.woodheadpublishingindia.com First published 2011, Woodhead Publishing Limited ß Woodhead Publishing Limited, 2011 The authors have asserted their moral rights. This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but the authors and the publisher cannot assume responsibility for the validity of all materials. Neither the authors nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from Woodhead Publishing Limited. The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited for such copying. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 978-1-84569-701-3 (print) ISBN 978-0-85709-254-0 (online) ISSN 2042-8049 Woodhead Publishing Series in Food Science, Technology and Nutrition (print) ISSN 2042-8057 Woodhead Publishing Series in Food Science, Technology and Nutrition (online) The publisher's policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp which is processed using acidfree and elemental chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. Typeset by Godiva Publishing Services Limited, Coventry, West Midlands, UK Printed by TJI Digital, Padstow, Cornwall, UK


Contents

Contributor contact details

............................................

xv

Woodhead Publishing Series in Food Science, Technology and Nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xxi

Preface

................................................................

xxix

Part I Deteriorative processes and factors influencing shelf life 1

Microbiological spoilage of foods and beverages . . . . . . . . . . . . . . . . G-J. E. Nychas and E. Panagou, Agricultural University of Athens, Greece 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Spoilage of foods and beverages; a microbiological approach: microbes vs indigenous enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Factors affecting the rate of microbiological spoilage of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Evaluating, monitoring and measuring microbiological spoilage of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5 Predicting microbiological spoilage of foods and beverages 1.6 Preventing microbiological spoilage of foods and beverages 1.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . 1.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


3 4 4 5 12 14 15 17 18 19

vi

Contents

2

Chemical deterioration and physical instability of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Kong and R. P. Singh, University of California, Davis, USA 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Chemical deterioration and physical instability of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Factors affecting the rate of quality loss due to chemical deterioration and physical instability . . . . . . . . . . . . . . . . . . . . . . . 2.4 Measuring chemical deterioration and physical instability of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Predicting and monitoring chemical deterioration and physical instability of foods and beverages . . . . . . . . . . . . . . . . . 2.6 Preventing chemical deterioration and physical instability of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . 2.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

4

5

Moisture loss, gain and migration in foods . . . . . . . . . . . . . . . . . . . . . . G. Roudaut, University of Burgundy, France and F. Debeaufort, University of Burgundy, France and IUT-Dijon, France 3.1 Introduction: moisture loss, gain and migration in foods and quality deterioration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Mechanism of the moisture transfers in food products . . . . . 3.3 Measuring, monitoring and predicting moisture loss, gain and migrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Moisture loss, gain and migration related to the shelf life . . 3.5 Conditions for moisture migration and foods affected by moisture transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Insect and mite penetration and contamination of packaged foods C. H. Bell, Food and Environment Research Agency, UK 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Insects and mites contaminating stored food products . . . . . . 4.3 Combating critical points in the food chain . . . . . . . . . . . . . . . . 4.4 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Sources of further information and advice . . . . . . . . . . . . . . . . . . 4.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

29 29 30 39 44 49 53 55 55 56 63

63 64 70 93 97 100 106 106 107 121 126 127 128

The influence of ingredients on product stability and shelf life . 132 N. W. G. Young, Danisco A/S, Multiple Food Applications, Denmark and University of Chester, UK and G. R. O'Sullivan, Danisco A/S, Multiple Food Applications, Denmark 5.1 Introduction to shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 ß Woodhead Publishing Limited, 2011

Contents 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 6

7

vii

Methods of shelf life extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Movement of moisture in food systems . . . . . . . . . . . . . . . . . . . . Food spoilage due to water activity . . . . . . . . . . . . . . . . . . . . . . . . Edible moisture barriers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preservation of foods by freezing . . . . . . . . . . . . . . . . . . . . . . . . . . Sweetener ingredients as humectants or cryoprotectants . . . . Ingredients for shelf life extension . . . . . . . . . . . . . . . . . . . . . . . . . Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of further information and advice . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

133 135 144 147 148 150 154 161 176 177 179

Processing and food and beverage shelf life . . . . . . . . . . . . . . . . . . . . . M. Brown, MHB Consulting, UK 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Main quality change factors and their interaction with processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Shelf life and stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Product and process design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6 Unit operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7 Production of low and intermediate moisture foods . . . . . . . . 6.8 Thermal processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.9 Filling and packaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.10 Novel processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.11 Hygiene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.12 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.13 References and further reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

184

Packaging and food and beverage shelf life . . . . . . . . . . . . . . . . . . . . . G. L. Robertson, University of Queensland and FoodPackagingEnvironment, Australia 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Role of packaging in extending food and beverage shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 Major packaging materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Key package properties related to shelf life . . . . . . . . . . . . . . . . 7.5 Predicting shelf life of packaged foods and beverages . . . . . . 7.6 Packaging migrants and food and beverage shelf life . . . . . . 7.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . 7.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


184 187 189 198 200 204 209 211 228 230 234 235 236 244 244 245 249 253 260 266 269 269 270

viii 8

9

Contents Effects of food and beverage storage, distribution, display and consumer handling on shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. Evans, London South Bank University, UK 8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Overview of the cold chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Storage life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Sectors of the cold chain and their influence on food quality and safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . 8.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smart packaging for monitoring and managing food and beverage shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P. S. Taoukis, National Technical University of Athens, Greece 9.1 Introduction: smart packaging ± time-temperature integrators (TTIs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Principles of the application of time-temperature integrators (TTIs) for shelf life monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Requirements and selection of time-temperature integrators (TTIs) for food and beverage products . . . . . . . . . . . . . . . . . . . . . 9.4 Use of time-temperature integrators (TTIs) for shelf life management and optimization in the cold chain ± case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Part II

273 273 274 276 283 298 299 300 303 303 304 309 311 320 321 321

Methods for shelf life and stability evaluation

10 Food storage trials: an introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. M. D. Man, London South Bank University, UK 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Food deterioration and spoilage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Storage trials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Sensory evaluation methods for food shelf life assessment . . . . . . D. Kilcast, Consultant in Food and Beverage Sensory Quality, UK 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Principles of sensory evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Basic requirements for sensory analysis . . . . . . . . . . . . . . . . . . . . 11.4 Discrimination tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.5 Quantitative descriptive tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ß Woodhead Publishing Limited, 2011

325 325 328 334 346 347 350 350 352 354 358 361

Contents 11.6 11.7 11.8 11.9 11.10 11.11 11.12 11.13 11.14 11.15

Consumer acceptability testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Operation of sensory shelf life tests . . . . . . . . . . . . . . . . . . . . . . . . Design of sensory shelf life tests . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpretation of sensory shelf life data . . . . . . . . . . . . . . . . . . . . . Instrumental methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standardisation in sensory shelf life testing . . . . . . . . . . . . . . . . Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of further information and advice . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12 Advances in instrumental methods to determine food quality deterioration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Kong and R. P. Singh, University of California, Davis, USA 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Assessing food appearance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3 Measurement of relative humidity (RH), moisture, and water activity (a w) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.4 Texture evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.5 Evaluation of rheological properties of liquid and semi-solid foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.6 Assessing lipid oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.7 Electronic nose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.8 Electronic tongue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.9 Infrared (IR) spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.10 Microbiological testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.11 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.12 Sources of further information and advice . . . . . . . . . . . . . . . . . . 12.13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Modelling microbiological shelf life of foods and beverages . . . . A. AmeÂzquita, D. Kan-King-Yu and Y. Le Marc, Unilever R&D Colworth, UK 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Classification of predictive models by microbial response . . 13.3 Development of predictive models for microbiological safety and stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4 Modelling approaches, applications and opportunities for shelf life prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5 Usage considerations and access to predictive microbiology electronic resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.6 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.7 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


ix 366 366 368 370 372 374 375 376 376 378 381 381 382 384 385 387 388 392 394 395 398 399 399 399 405 405 407 412 426 437 446 451 452

x

Contents

14 Modelling chemical and physical deterioration of foods and beverages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. J. Sousa Gallagher, P. V. Mahajan and Z. Yan, University College Cork, Ireland 14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.2 Factors influencing shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.3 Development of mathematical models . . . . . . . . . . . . . . . . . . . . . . 14.4 Predictive mathematical models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Accelerated shelf life testing of foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. Mizrahi, Technion-Israel Institute of Technology, Israel 15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.2 Basic principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.3 Initial rate approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.4 Kinetic model approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.5 Single accelerating factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.6 Glass transition models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.7 Multiple accelerating factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.8 Dynamic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.9 The `no model' approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.10 Combination of approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.11 Problems in accelerated shelf life tests . . . . . . . . . . . . . . . . . . . . . 15.12 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Microbiological challenge testing of foods . . . . . . . . . . . . . . . . . . . . . . . E. Komitopoulou, Leatherhead Food Research, UK 16.1 Introduction: role of challenge testing in shelf life evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.2 Basic principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.3 Challenge testing limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4 Challenge testing and the use of mathematical models . . . . . 16.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . 16.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Part III

459 459 460 462 466 476 477 482 482 483 483 485 487 493 493 496 498 500 501 502 503 507 507 508 519 519 521 521 522

The stability and shelf life of particular products

17 Beer shelf life and stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G. G. Stewart and F. G. Priest, Heriot-Watt University, UK 17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2 Biological instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


527 527 528

Contents 17.3 17.4 17.5 17.6 17.7 17.8 17.9

xi

Physical instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flavour stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Foam stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gushing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Light stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

529 531 533 536 536 537 537

18 Shelf life of wine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R. S. Jackson, Brock University, Canada 18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2 Factors affecting wine stability and shelf life . . . . . . . . . . . . . . . 18.3 Changes during the shelf life of wine . . . . . . . . . . . . . . . . . . . . . . 18.4 Evaluating wine shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.5 Preventing wine quality deterioration at or post-bottling . . . 18.6 Sensory significance of shelf life changes . . . . . . . . . . . . . . . . . . 18.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . 18.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

540

19 The stability and shelf life of fruit juices and soft drinks . . . . . . . P. Ashurst, Ashurst and Associates, UK 19.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.2 Factors influencing the stability of fruit juices and soft drinks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.3 Ensuring product stability and extending shelf life . . . . . . . . . 19.4 Shelf life determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . 20 Practical uses of sensory evaluation for the assessment of soft drink shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L. L. Rogers, Consultant, UK 20.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.2 Using a risk-based approach to shelf life for soft drinks . . . 20.3 Estimating shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.4 Determining shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.5 Monitoring shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.6 Considerations before developing the shelf life plan . . . . . . . 20.7 Developing the sensory plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.8 Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.9 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


540 542 551 558 559 563 564 565 565 571 571 574 584 588 591 593 594 594 596 597 597 598 599 603 603 613 614

xii

Contents

21 The stability and shelf life of coffee products . . . . . . . . . . . . . . . . . . . L. Manzocco, S. Calligaris and M. C. Nicoli, University of Udine, Italy 21.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Main critical events affecting the stability and shelf life of coffee products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Ensuring stability and extending the shelf life of coffee . . . . 21.4 Evaluating the shelf life of coffee . . . . . . . . . . . . . . . . . . . . . . . . . . 21.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . 21.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 The stability and shelf life of fruit and vegetables . . . . . . . . . . . . . . M. J. Sousa Gallagher and P. V. Mahajan, University College Cork, Ireland 22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Stability and shelf life of fruit and vegetables . . . . . . . . . . . . . . 22.3 Extending the shelf life of fruit and vegetables . . . . . . . . . . . . . 22.4 Controlled and modified atmosphere packaging for longer shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 The stability and shelf life of bread and other bakery products S. P. Cauvain and L. S. Young, BakeTran, UK 23.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2 A brief overview of the manufacture of bakery products . . . 23.3 The key `fresh' characteristics of bakery products . . . . . . . . . . 23.4 Factors affecting the stability of bread and other bakery products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.5 Evaluating the shelf life of bread and other bakery products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.6 Ensuring stability and extending the shelf life of bread and other bakery products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 The stability and shelf life of fats and oils . . . . . . . . . . . . . . . . . . . . . . . G. Talbot, The Fat Consultant, UK 24.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2 Mechanisms of oxidation and hydrolysis in fats and oils . . . 24.3 Factors affecting the stability and shelf life of fats and oils 24.4 Evaluating the shelf life of fats and oils . . . . . . . . . . . . . . . . . . . . 24.5 Ensuring stability and extending the shelf life of fats and oils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.6 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ß Woodhead Publishing Limited, 2011

615 615 617 620 624 636 636 637 641 641 642 648 650 652 653 657 657 658 662 663 669 673 678 680 683 683 684 693 703 707 712

Contents 24.7 24.8

xiii

Sources of further information and advice . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

714 714

25 The stability and shelf life of confectionery products . . . . . . . . . . . P. Subramaniam, Leatherhead Food Research, UK 25.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.2 Factors affecting shelf life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3 Chocolate and chocolate products . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4 Sugar glass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5 Toffee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6 Gums and jellies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7 Aerated confectionery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . 25.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

716

26 The stability and shelf life of vitamin-fortified foods . . . . . . . . . . . . R. Burch, Leatherhead Food Research, UK 26.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.2 Factors affecting the stability and shelf life of vitaminfortified foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.3 Ensuring stability and extending the shelf life of vitaminfortified foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.4 Evaluating the shelf life of vitamin-fortified foods . . . . . . . . . 26.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . 26.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 The stability and shelf life of milk and milk products . . . . . . . . . . D. D. Muir, Consultant, UK 27.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27.2 Chemical composition and principal reactions of milk . . . . . 27.3 Bacteria in milk and related enzyme activity . . . . . . . . . . . . . . . 27.4 Raw milk enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27.5 Control of the quality of short shelf life products . . . . . . . . . . 27.6 Factors influencing the stability of long shelf life products . 27.7 Control of the stability of long life milk products . . . . . . . . . . 27.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27.9 Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27.10 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 The stability and shelf life of seafood . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. ToldraÂ, Institute of Agro-chemical Technology and Food (CSIC), Spain and M. Reig, Universidad PoliteÂcnica de Valencia, Spain 28.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28.2 Factors affecting the stability and shelf life of seafood . . . . . 28.3 Microorganisms involved in seafood spoilage . . . . . . . . . . . . . . ß Woodhead Publishing Limited, 2011

716 717 720 727 729 733 737 740 740 743 743 745 749 750 752 752 753 755 756 756 761 763 764 768 770 777 777 777 779 779 780 783

xiv

Contents 28.4 28.5 28.6 28.7

Evaluation of the shelf life of seafood . . . . . . . . . . . . . . . . . . . . . . Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of further information and advice . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

784 787 788 788

29 The stability and shelf life of meat and poultry . . . . . . . . . . . . . . . . . M. G. O'Sullivan, University College Cork, Ireland 29.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29.2 Factors affecting the stability and shelf life of meat and poultry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29.3 Evaluating the shelf life of meat and poultry . . . . . . . . . . . . . . . 29.4 Ensuring stability and extending the shelf life of meat and poultry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29.5 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29.6 Sources of further information and advice . . . . . . . . . . . . . . . . . . 29.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

793

804 808 809 809

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

817


793 794 800


Chapter 1

(* = main contact)

Editors Dr David Kilcast Consultant in Food and Beverage Sensory Quality E-mail: [email protected] Persis Subramaniam Leatherhead Food Research Randalls Road Leatherhead KT22 7RY UK E-mail: psubramaniam@ leatherheadfood.com

George-John E. Nychas* and Efstathios Panagou Department of Food Science Technology & Human Nutrition Laboratory of Microbiology and Biotechnology of Foods Agricultural University of Athens Iera odos 75 Athens 11855 Greece E-mail: [email protected]

Chapters 2 and 12 Fanbin Kong and R. Paul Singh* Department of Biological and Agricultural Engineering University of California, Davis Davis, CA 95616 USA E-mail: [email protected]


xvi


Chapter 3

Chapter 5

GaeÈlle Roudaut* Department of Water, Active Molecules, Macromolecules and Activities EMMA EA 581 AgroSup-Dijon UniversiteÂ de Bourgogne 1 esplanade Erasme 21 000 Dijon France E-mail: [email protected]

Professor Niall W. G. Young* Danisco A/S, Multiple Food Applications Edwin Rahrs Vej 38 8220 Brabrand Denmark E-mail: [email protected]

FreÂdeÂric Debeaufort Department of Water, Active Molecules, Macromolecules and Activities EMMA EA 581 AgroSup-Dijon UniversiteÂ de Bourgogne 1 esplanade Erasme 21 000 Dijon France and IUT-Dijon 7 Blvd Docteur Petitjean BP 17867, 21078 Dijon Cedex France

Chapter 4 Dr Christopher H. Bell Food and Environment Research Agency Sand Hutton York YO41 1LZ UK E-mail: [email protected]

and University of Chester Environmental Quality and Food Safety Research Unit Department of Biological Sciences Parkgate Road Chester CH1 4BJ UK Geoffrey R. O'Sullivan Danisco A/S, Multiple Food Applications Edwin Rahrs Vej 38 8220 Brabrand Denmark E-mail: geoffrey.royston.osullivan@ danisco.com

Chapter 6 Martyn Brown MHB Consulting 41 College Road The Historic Dockyard Chatham ME4 4JS UK E-mail: [email protected]


Contributor contact details xvii

Chapter 7

Chapter 11

Professor Gordon L. Robertson University of Queensland and FoodPackagingEnvironment 6066 Lugano Drive Hope Island QLD 4212 Australia E-mail: [email protected]

Dr David Kilcast Consultant in Food and Beverage Sensory Quality & Leatherhead Food International E-mail: [email protected]

Chapter 8 Dr Judith Evans London South Bank University Churchill Building Langford Bristol BS40 5DU UK E-mail: [email protected]

Chapter 13 Alejandro AmeÂzquita*, Denis KanKing-Yu and Yvan Le Marc Safety & Environmental Assurance Centre Unilever R&D Colworth Sharnbrook MK44 1LQ UK E-mail: Alejandro.Amezquita@ unilever.com

Chapter 14

Chapter 9 Dr Petros S. Taoukis National Technical University of Athens School of Chemical Engineering Division IV ± Product and Process Development Laboratory of Food Chemistry and Technology Iroon Polytechniou 5 15780 Athens Greece E-mail: [email protected]

Dr Maria J. Sousa Gallagher*, Dr Pramod V. Mahajan and Dr Zhengyong Yan Department of Process & Chemical Engineering College of Science, Engineering and Food Science University College Cork, UCC College Road Cork Ireland E-mail: [email protected]

Chapter 15

Chapter 10 Dr Dominic Man London South Bank University London E-mail: [email protected]

Professor Shimon Mizrahi Technion-Israel Institute of Technology Israel E-mail: [email protected]


xviii


Chapter 16

Chapter 20

Dr Evangelia Komitopoulou Leatherhead Food Research Randalls Road Leatherhead KT22 7RY UK E-mail: ekomitopoulou@ leatherheadfood.com

Lauren L. Rogers Consultant 17 Alma Street Leek ST13 8EH UK E-mail: [email protected]

Chapter 21

Chapter 17 Professor Graham G. Stewart* and Professor Fergus G. Priest International Centre for Brewing and Distilling Heriot-Watt University Riccarton Edinburgh EH14 4AS UK E-mail: [email protected]

Chapter 18 Dr Ronald S. Jackson Cool Climate Oenology and Viticulture Institute (CCOVI) Brock University 500 Glenridge Avenue St Catharines Ontario Canada L2S 3A1 E-mail: [email protected]

Lara Manzocco*, Sonia Calligaris and Maria Cristina Nicoli Department of Food Science University of Udine via Sondrio 2a 33100 Udine Italy E-mail: [email protected]

Chapter 22 Dr Maria J. Sousa Gallagher* and Dr Pramod V. Mahajan Department of Process & Chemical Engineering College of Science, Engineering and Food Science University College Cork, UCC College Road Cork Ireland E-mail: [email protected]

Chapter 23

Chapter 19 Dr Philip Ashurst Ashurst and Associates Reachfar Middleton-on-the-Hill Ludlow SY8 4BD UK E-mail: [email protected]

Dr Stanley P. Cauvain* and Dr Linda S. Young BakeTran 97 Guinions Road High Wycombe HP13 7NU UK E-mail: [email protected] [email protected]



xix

Chapter 24

Chapter 28

Geoff Talbot The Fat Consultant Suite 250 St Loyes House 20 St Loyes Street Bedford MK40 1ZL UK E-mail: [email protected]

Fidel ToldraÂ* Instituto de AgroquõÂmica y TecnologõÂa de Alimentos (CSIC) Avenue AgustõÂn Escardino 7 46980 Paterna (Valencia) Spain E-mail: [email protected]

Chapter 25 Persis Subramaniam Leatherhead Food Research Randalls Road Leatherhead KT22 7RY E-mail: [email protected]

Milagro Reig Institute of Food Engineering for Development Universidad PoliteÂcnica de Valencia Camino de Vera s/n 46022, Valencia Spain E-mail: [email protected]

Chapter 29 Chapter 26 Dr Rachel Burch Leatherhead Food Research Randalls Road Leatherhead KT22 7RY UK E-mail: [email protected]

Dr Maurice G. O'Sullivan School of Food and Nutritional Sciences University College Cork Ireland E-mail: [email protected]

Chapter 27 Professor David Donald Muir Consultant 26 Pennyvenie Way Girdle Toll Irvine KA11 1QQ UK E-mail: [email protected]


Woodhead Publishing Series in Food Science, Technology and Nutrition

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Chilled foods: a comprehensive guide Edited by C. Dennis and M. Stringer Yoghurt: science and technology A. Y. Tamime and R. K. Robinson Food processing technology: principles and practice P. J. Fellows Bender's dictionary of nutrition and food technology Sixth edition D. A. Bender Determination of veterinary residues in food Edited by N. T. Crosby Food contaminants: sources and surveillance Edited by C. Creaser and R. Purchase Nitrates and nitrites in food and water Edited by M. J. Hill Pesticide chemistry and bioscience: the food±environment challenge Edited by G. T. Brooks and T. Roberts Pesticides: developments, impacts and controls Edited by G. A. Best and A. D. Ruthven Dietary fibre: chemical and biological aspects Edited by D. A. T. Southgate, K. W. Waldron, I. T. Johnson and G. R. Fenwick Vitamins and minerals in health and nutrition M. Tolonen Technology of biscuits, crackers and cookies Second edition D. Manley Instrumentation and sensors for the food industry Edited by E. Kress-Rogers Food and cancer prevention: chemical and biological aspects Edited by K. W. Waldron, I. T. Johnson and G. R. Fenwick Food colloids: proteins, lipids and polysaccharides Edited by E. Dickinson and B. Bergenstahl Food emulsions and foams Edited by E. Dickinson Maillard reactions in chemistry, food and health Edited by T. P. Labuza, V. Monnier, J. Baynes and J. O'Brien The Maillard reaction in foods and medicine Edited by J. O'Brien, H. E. Nursten, M. J. Crabbe and J. M. Ames Encapsulation and controlled release Edited by D. R. Karsa and R. A. Stephenson Flavours and fragrances Edited by A. D. Swift Feta and related cheeses Edited by A. Y. Tamime and R. K. Robinson Biochemistry of milk products Edited by A. T. Andrews and J. R. Varley Physical properties of foods and food processing systems M. J. Lewis


xxii


24 Food irradiation: a reference guide V. M. Wilkinson and G. Gould 25 Kent's technology of cereals: an introduction for students of food science and agriculture Fourth edition N. L. Kent and A. D. Evers 26 Biosensors for food analysis Edited by A. O. Scott 27 Separation processes in the food and biotechnology industries: principles and applications Edited by A. S. Grandison and M. J. Lewis 28 Handbook of indices of food quality and authenticity R. S. Singhal, P. K. Kulkarni and D. V. Rege 29 Principles and practices for the safe processing of foods D. A. Shapton and N. F. Shapton 30 Biscuit, cookie and cracker manufacturing manuals Volume 1: ingredients D. Manley 31 Biscuit, cookie and cracker manufacturing manuals Volume 2: biscuit doughs D. Manley 32 Biscuit, cookie and cracker manufacturing manuals Volume 3: biscuit dough piece forming D. Manley 33 Biscuit, cookie and cracker manufacturing manuals Volume 4: baking and cooling of biscuits D. Manley 34 Biscuit, cookie and cracker manufacturing manuals Volume 5: secondary processing in biscuit manufacturing D. Manley 35 Biscuit, cookie and cracker manufacturing manuals Volume 6: biscuit packaging and storage D. Manley 36 Practical dehydration Second edition M. Greensmith 37 Lawrie's meat science Sixth edition R. A. Lawrie 38 Yoghurt: science and technology Second edition A. Y. Tamime and R. K. Robinson 39 New ingredients in food processing: biochemistry and agriculture G. Linden and D. Lorient 40 Benders' dictionary of nutrition and food technology Seventh edition D. A. Bender and A. E. Bender 41 Technology of biscuits, crackers and cookies Third edition D. Manley 42 Food processing technology: principles and practice Second edition P. J. Fellows 43 Managing frozen foods Edited by C. J. Kennedy 44 Handbook of hydrocolloids Edited by G. O. Phillips and P. A. Williams 45 Food labelling Edited by J. R. Blanchfield 46 Cereal biotechnology Edited by P. C. Morris and J. H. Bryce 47 Food intolerance and the food industry Edited by T. Dean 48 The stability and shelf life of food Edited by D. Kilcast and P. Subramaniam 49 Functional foods: concept to product Edited by G. R. Gibson and C. M. Williams 50 Chilled foods: a comprehensive guide Second edition Edited by M. Stringer and C. Dennis 51 HACCP in the meat industry Edited by M. Brown 52 Biscuit, cracker and cookie recipes for the food industry D. Manley 53 Cereals processing technology Edited by G. Owens 54 Baking problems solved S. P. Cauvain and L. S. Young 55 Thermal technologies in food processing Edited by P. Richardson 56 Frying: improving quality Edited by J. B. Rossell 57 Food chemical safety Volume 1: contaminants Edited by D. Watson 58 Making the most of HACCP: learning from others' experience Edited by T. Mayes and S. Mortimore 59 Food process modelling Edited by L. M. M. Tijskens, M. L. A. T. M. Hertog and B. M. NicolaõÈ



xxiii

60 EU food law: a practical guide Edited by K. Goodburn 61 Extrusion cooking: technologies and applications Edited by R. Guy 62 Auditing in the food industry: from safety and quality to environmental and other audits Edited by M. Dillon and C. Griffith 63 Handbook of herbs and spices Volume 1 Edited by K. V. Peter 64 Food product development: maximising success M. Earle, R. Earle and A. Anderson 65 Instrumentation and sensors for the food industry Second edition Edited by E. Kress-Rogers and C. J. B. Brimelow 66 Food chemical safety Volume 2: additives Edited by D. Watson 67 Fruit and vegetable biotechnology Edited by V. Valpuesta 68 Foodborne pathogens: hazards, risk analysis and control Edited by C. de W. Blackburn and P. J. McClure 69 Meat refrigeration S. J. James and C. James 70 Lockhart and Wiseman's crop husbandry Eighth edition H. J. S. Finch, A. M. Samuel and G. P. F. Lane 71 Safety and quality issues in fish processing Edited by H. A. Bremner 72 Minimal processing technologies in the food industries Edited by T. Ohlsson and N. Bengtsson 73 Fruit and vegetable processing: improving quality Edited by W. Jongen 74 The nutrition handbook for food processors Edited by C. J. K. Henry and C. Chapman 75 Colour in food: improving quality Edited by D. MacDougall 76 Meat processing: improving quality Edited by J. P. Kerry, J. F. Kerry and D. A. Ledward 77 Microbiological risk assessment in food processing Edited by M. Brown and M. Stringer 78 Performance functional foods Edited by D. Watson 79 Functional dairy products Volume 1 Edited by T. Mattila-Sandholm and M. Saarela 80 Taints and off-flavours in foods Edited by B. Baigrie 81 Yeasts in food Edited by T. Boekhout and V. Robert 82 Phytochemical functional foods Edited by I. T. Johnson and G. Williamson 83 Novel food packaging techniques Edited by R. Ahvenainen 84 Detecting pathogens in food Edited by T. A. McMeekin 85 Natural antimicrobials for the minimal processing of foods Edited by S. Roller 86 Texture in food Volume 1: semi-solid foods Edited by B. M. McKenna 87 Dairy processing: improving quality Edited by G. Smit 88 Hygiene in food processing: principles and practice Edited by H. L. M. Lelieveld, M. A. Mostert, B. White and J. Holah 89 Rapid and on-line instrumentation for food quality assurance Edited by I. Tothill 90 Sausage manufacture: principles and practice E. Essien 91 Environmentally friendly food processing Edited by B. Mattsson and U. Sonesson 92 Bread making: improving quality Edited by S. P. Cauvain 93 Food preservation techniques Edited by P. Zeuthen and L. Bùgh-Sùrensen 94 Food authenticity and traceability Edited by M. Lees 95 Analytical methods for food additives R. Wood, L. Foster, A. Damant and P. Key 96 Handbook of herbs and spices Volume 2 Edited by K. V. Peter 97 Texture in food Volume 2: solid foods Edited by D. Kilcast 98 Proteins in food processing Edited by R. Yada 99 Detecting foreign bodies in food Edited by M. Edwards


xxiv


100 Understanding and measuring the shelf-life of food Edited by R. Steele 101 Poultry meat processing and quality Edited by G. Mead 102 Functional foods, ageing and degenerative disease Edited by C. Remacle and B. Reusens 103 Mycotoxins in food: detection and control Edited by N. Magan and M. Olsen 104 Improving the thermal processing of foods Edited by P. Richardson 105 Pesticide, veterinary and other residues in food Edited by D. Watson 106 Starch in food: structure, functions and applications Edited by A.-C. Eliasson 107 Functional foods, cardiovascular disease and diabetes Edited by A. Arnoldi 108 Brewing: science and practice D. E. Briggs, P. A. Brookes, R. Stevens and C. A. Boulton 109 Using cereal science and technology for the benefit of consumers: proceedings of the 12th International ICC Cereal and Bread Congress, 24±26th May, 2004, Harrogate, UK Edited by S. P. Cauvain, L. S. Young and S. Salmon 110 Improving the safety of fresh meat Edited by J. Sofos 111 Understanding pathogen behaviour in food: virulence, stress response and resistance Edited by M. Griffiths 112 The microwave processing of foods Edited by H. Schubert and M. Regier 113 Food safety control in the poultry industry Edited by G. Mead 114 Improving the safety of fresh fruit and vegetables Edited by W. Jongen 115 Food, diet and obesity Edited by D. Mela 116 Handbook of hygiene control in the food industry Edited by H. L. M. Lelieveld, M. A. Mostert and J. Holah 117 Detecting allergens in food Edited by S. Koppelman and S. Hefle 118 Improving the fat content of foods Edited by C. Williams and J. Buttriss 119 Improving traceability in food processing and distribution Edited by I. Smith and A. Furness 120 Flavour in food Edited by A. Voilley and P. Etievant 121 The Chorleywood bread process S. P. Cauvain and L. S. Young 122 Food spoilage microorganisms Edited by C. de W. Blackburn 123 Emerging foodborne pathogens Edited by Y. Motarjemi and M. Adams 124 Benders' dictionary of nutrition and food technology Eighth edition D. A. Bender 125 Optimising sweet taste in foods Edited by W. J. Spillane 126 Brewing: new technologies Edited by C. Bamforth 127 Handbook of herbs and spices Volume 3 Edited by K. V. Peter 128 Lawrie's meat science Seventh edition R. A. Lawrie in collaboration with D. A. Ledward 129 Modifying lipids for use in food Edited by F. Gunstone 130 Meat products handbook: practical science and technology G. Feiner 131 Food consumption and disease risk: consumer±pathogen interactions Edited by M. Potter 132 Acrylamide and other hazardous compounds in heat-treated foods Edited by K. Skog and J. Alexander 133 Managing allergens in food Edited by C. Mills, H. Wichers and K. HoffmanSommergruber 134 Microbiological analysis of red meat, poultry and eggs Edited by G. Mead 135 Maximising the value of marine by-products Edited by F. Shahidi 136 Chemical migration and food contact materials Edited by K. Barnes, R. Sinclair and D. Watson 137 Understanding consumers of food products Edited by L. Frewer and H. van Trijp 138 Reducing salt in foods: practical strategies Edited by D. Kilcast and F. Angus 139 Modelling microorganisms in food Edited by S. Brul, S. Van Gerwen and M. Zwietering



xxv

140 Tamime and Robinson's Yoghurt: science and technology Third edition A. Y. Tamime and R. K. Robinson 141 Handbook of waste management and co-product recovery in food processing Volume 1 Edited by K. W. Waldron 142 Improving the flavour of cheese Edited by B. Weimer 143 Novel food ingredients for weight control Edited by C. J. K. Henry 144 Consumer-led food product development Edited by H. MacFie 145 Functional dairy products Volume 2 Edited by M. Saarela 146 Modifying flavour in food Edited by A. J. Taylor and J. Hort 147 Cheese problems solved Edited by P. L. H. McSweeney 148 Handbook of organic food safety and quality Edited by J. Cooper, C. Leifert and U. Niggli 149 Understanding and controlling the microstructure of complex foods Edited by D. J. McClements 150 Novel enzyme technology for food applications Edited by R. Rastall 151 Food preservation by pulsed electric fields: from research to application Edited by H. L. M. Lelieveld and S. W. H. de Haan 152 Technology of functional cereal products Edited by B. R. Hamaker 153 Case studies in food product development Edited by M. Earle and R. Earle 154 Delivery and controlled release of bioactives in foods and nutraceuticals Edited by N. Garti 155 Fruit and vegetable flavour: recent advances and future prospects Edited by B. BruÈckner and S. G. Wyllie 156 Food fortification and supplementation: technological, safety and regulatory aspects Edited by P. Berry Ottaway 157 Improving the health-promoting properties of fruit and vegetable products Edited by F. A. TomaÂs-BarberaÂn and M. I. Gil 158 Improving seafood products for the consumer Edited by T. Bùrresen 159 In-pack processed foods: improving quality Edited by P. Richardson 160 Handbook of water and energy management in food processing Edited by J. KlemesÏ, R. Smith and J.-K. Kim 161 Environmentally compatible food packaging Edited by E. Chiellini 162 Improving farmed fish quality and safety Edited by é. Lie 163 Carbohydrate-active enzymes Edited by K.-H. Park 164 Chilled foods: a comprehensive guide Third edition Edited by M. Brown 165 Food for the ageing population Edited by M. M. Raats, C. P. G. M. de Groot and W. A. Van Staveren 166 Improving the sensory and nutritional quality of fresh meat Edited by J. P. Kerry and D. A. Ledward 167 Shellfish safety and quality Edited by S. E. Shumway and G. E. Rodrick 168 Functional and speciality beverage technology Edited by P. Paquin 169 Functional foods: principles and technology M. Guo 170 Endocrine-disrupting chemicals in food Edited by I. Shaw 171 Meals in science and practice: interdisciplinary research and business applications Edited by H. L. Meiselman 172 Food constituents and oral health: current status and future prospects Edited by M. Wilson 173 Handbook of hydrocolloids Second edition Edited by G. O. Phillips and P. A. Williams 174 Food processing technology: principles and practice Third edition P. J. Fellows 175 Science and technology of enrobed and filled chocolate, confectionery and bakery products Edited by G. Talbot


xxvi


176 Foodborne pathogens: hazards, risk analysis and control Second edition Edited by C. de W. Blackburn and P. J. McClure 177 Designing functional foods: measuring and controlling food structure breakdown and absorption Edited by D. J. McClements and E. A. Decker 178 New technologies in aquaculture: improving production efficiency, quality and environmental management Edited by G. Burnell and G. Allan 179 More baking problems solved S. P. Cauvain and L. S. Young 180 Soft drink and fruit juice problems solved P. Ashurst and R. Hargitt 181 Biofilms in the food and beverage industries Edited by P. M. Fratamico, B. A. Annous and N. W. Gunther 182 Dairy-derived ingredients: food and neutraceutical uses Edited by M. Corredig 183 Handbook of waste management and co-product recovery in food processing Volume 2 Edited by K. W. Waldron 184 Innovations in food labelling Edited by J. Albert 185 Delivering performance in food supply chains Edited by C. Mena and G. Stevens 186 Chemical deterioration and physical instability of food and beverages Edited by L. Skibsted, J. Risbo and M. Andersen 187 Managing wine quality Volume 1: viticulture and wine quality Edited by A. G. Reynolds 188 Improving the safety and quality of milk Volume 1: milk production and processing Edited by M. Griffiths 189 Improving the safety and quality of milk Volume 2: improving quality in milk products Edited by M. Griffiths 190 Cereal grains: assessing and managing quality Edited by C. Wrigley and I. Batey 191 Sensory analysis for food and beverage control: a practical guide Edited by D. Kilcast 192 Managing wine quality Volume 2: oenology and wine quality Edited by A. G. Reynolds 193 Winemaking problems solved Edited by C. E. Butzke 194 Environmental assessment and management in the food industry Edited by U. Sonesson, J. Berlin and F. Ziegler 195 Consumer-driven innovation in food and personal care products Edited by S. R. Jaeger and H. MacFie 196 Tracing pathogens in the food chain Edited by S. Brul, P. M. Fratamico and T. A. McMeekin 197 Case studies in novel food processing technologies: innovations in processing, packing and predictive modelling Edited by C. J. Doona, K. Kustin and F. E. Feeherry 198 Freeze-drying of pharmaceutical and food products T.-C. Hua, B.-L. Liu and H. Zhang 199 Oxidation in foods and beverages and antioxidant applications Volume 1: understanding mechanisms of oxidation and antioxidant activity Edited by E. A. Decker, R. J. Elias and D. J. McClements 200 Oxidation in foods and beverages and antioxidant applications Volume 2: management in different industry sectors Edited by E. A. Decker, R. J. Elias and D. J. McClements 201 Protective cultures, antimicrobial metabolites and bacteriophages of food and beverage biopreservation Edited by C. Lacroix 202 Separation, extraction and concentration processes in the food, beverage and nutraceutical industries Edited by S. S. H. Rizvi 203 Determining mycotoxins and mycotoxigenic fungi in food and feed Edited by S. De Saeger 204 Developing children's food products Edited by D. Kilcast and F. Angus



xxvii

205 Functional foods: concept to product Second edition Edited by M. Saarela 206 Postharvest biology and technology of tropical and subtropical fruits Volume 1 Edited by E. M. Yahia 207 Postharvest biology and technology of tropical and subtropical fruits Volume 2 Edited by E. M. Yahia 208 Postharvest biology and technology of tropical and subtropical fruits Volume 3 Edited by E. M. Yahia 209 Postharvest biology and technology of tropical and subtropical fruits Volume 4 Edited by E. M. Yahia 210 Food and beverage stability and shelf life Edited by D. Kilcast and P. Subramaniam 211 Processed meats: improving safety, nutrition and quality Edited by J. P. Kerry and J. F. Kerry 212 Food chain integrity: a holistic approach to food traceability, safety, quality and authenticity Edited by J. Hoorfar, K. Jordan, F. Butler and R. Prugger 213 Improving the safety and quality of eggs and egg products Volume 1 Edited by Y. Nys, M. Bain and F. Van Immerseel 214 Improving the safety and quality of eggs and egg products Volume 2 Edited by Y. Nys, M. Bain and F. Van Immerseel 215 Feed and fodder contamination: effects on livestock and food safety Edited by J. Fink-Gremmels 216 Hygiene in the design, construction and renovation of food processing factories Edited by H. L. M. Lelieveld and J. Holah 217 Technology of biscuits, crackers and cookies Fourth edition Edited by D. Manley 218 Nanotechnology in the food, beverage and nutraceutical industries Edited by Q. Huang 219 Rice quality K. R. Bhattacharya 220 Meat, poultry and seafood packaging Edited by J. P. Kerry 221 Reducing saturated fats in foods Edited by G. Talbot 222 Handbook of food proteins Edited by G. O. Phillips and P. A. Williams 223 Lifetime nutritional influences on cognition, behaviour and psychiatric illness Edited by D. Benton


Preface

The stability of a food product and its consequent shelf life depends on many factors including the quality of ingredients, product composition and structure, processing conditions used during manufacture, packaging characteristics and finally the storage, handling and distribution conditions. All these factors need to be firstly understood and then controlled to achieve the optimal or target quality and shelf life. The food industry has a great responsibility firstly to ensure that the products it manufactures are safe at every occasion over its entire shelf life and, additionally, that the products are of a sensory quality acceptable to and expected by the consumer. Manufacturers need to address both these issues in setting the shelf life for their products. Short shelf life products that spoil due to microbial activity, such as chilled products, are marked with a ùse by' date, whereas the more stable products, that degrade but do not pose a health risk, are given a `best before' date. These date marks are set by the manufacturers after much testing to determine the product shelf life. A general problem faced by manufacturers is the time constraint of the product development stage. Often when dealing with relatively stable products, real time tests cannot be performed that can run to the end of life. In these cases, the manufacturer has to rely on a combination of previous experience and results of accelerated stability tests to set the shelf life. If the shelf life is set too conservatively, there is an impact on unnecessary food waste, but if set too generously there may be an impact on loss of quality and consumer acceptability. Shelf life dating of food is currently a hot topic for both government authorities and the food industry. Industry is facing pressures from the authorities and consumer groups to reduce the amount of packaging and food waste. Consumers are better informed and more demanding with regard to products they buy and consume. The constant demand by consumers for healthier products with


xxx

Preface

reduced fat, sugar and salt, and the removal of artificial additives, including preservatives, puts further pressure on manufacturers in terms of ensuring food safety and a target shelf life. Achieving the targets requires a good knowledge of the impact of ingredients and processing for the different categories of products. This book has been produced to provide useful information to those addressing issues relating to the stability of products. Theoretical and practical aspects of product stability are brought together in making decisions about the storage stability and setting of shelf life. The experience and help of experts are invaluable in this process as it is not possible for any one individual to hold all the knowledge and answers to all questions. It was clear at the time of publication of our first book, The Stability and Shelf-life of Food (Woodhead Publishing, 2000) with its contributions from experts in the field that there was a great need for information that specifically related to shelf life. The continued interest in this book in the ten years since publication has suggested to us that there is an ongoing need for an expanded and updated reference book dedicated to discussing food stability issues and shelf life measurement, which can be used as a resource by all those needing help or requiring reassuring advice. With this aim in mind, we have brought together contributions from individual experts in their own fields both from academia and industry to produce in a single volume a comprehensive book covering deteriorative processes, shelf life measurement techniques and specific issues related to a wide range of products. The book is structured into three parts. Part I describes the various types of deteriorative processes that can limit the shelf life of products, including physiochemical aspects, insect contamination, processing, packaging and storage and distribution. Part II describes methods for shelf life evaluation, including sensory, instrumental and microbiological tests, accelerated testing and shelf life modelling procedures. Part III covers a combination of productrelated shelf life issues and case studies related to a wide range of product categories, including beer, wine, fruit juices, soft drinks, coffee, fruit and vegetables, bread and baked products, oils and fats, confectionery products, milk and milk products, seafood and vitamin-fortified products. Users of The Stability and Shelf-life of Food will consequently find not only updated versions of essential chapters from this publication, but a much wider overview of stability and shelf life issues, and covering a much wider range of product categories. We would like to thank all the contributors to the book, each one an expert in their own field. It is our hope that you will find the book to be a useful and lasting resource on shelf life evaluation. D Kilcast, Consultant P Subramaniam, Leatherhead Food Research


1 Microbiological spoilage of foods and beverages G-J. E. Nychas and E. Panagou, Agricultural University of Athens, Greece

Abstract: Food spoilage may be defined as a process or change which renders a product undesirable or unacceptable for consumption. This complex ecological phenomenon is the outcome of the biochemical activity of microbial chemical processes which will eventually dominate according to the prevailing ecological determinants. To ensure the safety and quality of foods and beverages, the effective monitoring of the chill chain through production, transportation, distribution and storage in retail cabinets and home refrigerators is essential. Currently, a variety of different methodologies are used for assessing food spoilage, in which microbiological methods play a decisive role. Recently, the relationship between microbial growth and the chemical changes occurring during food storage has been recognised as a potential indicator which may be useful for monitoring freshness and safety. For this purpose, interesting analytical approaches have been developed for rapid and quantitative assessment of food spoilage. These are based on biosensors, sensor arrays and spectroscopy techniques in tandem with chemometrics. Various processes have been utilised to prevent the microbiological spoilage of foods and beverages, amongst which low temperature storage and heat treatment seem to be the most effective. The application of a rich carbon dioxide atmosphere as part of a modified atmosphere packaging system is also effective in suppressing spoilage microorganisms. Key words: chill chain, ephemeral spoilage organisms, metabolomics, microbial inhibition, modified atmosphere packaging, shelf life.


4

1.1


Introduction

Despite the technological progress made in recent decades, changes in consumer lifestyles have made it necessary for the food industry to fulfil seemingly contradictory market demands. Consumers now expect food products of superior sensory quality and increased functional and nutritional properties, combined with a traditional, wholesome image and guaranteed safety. However, there is also a demand for less heavily preserved or processed foods, for fewer additives and technological interventions, as well as for increasingly competitive prices. At the same time, consumers expect an extended product shelf life (i.e. inhibition or control of spoilage which is mainly microbiological) and a high level of convenience in preparation and use. In a recent consumer survey, `fresh/not spoiled' and `quality' were the second and third most important criteria with 37% and 33%, respectively, while `price' was the most important purchase criterion for food (mentioned by 66% of respondents) (RoÈhr et al., 2005). This is a straightforward message as to the importance of the successful management of spoilage. To date, there are no food spoilage management systems per se, food spoilage control being linked with many other safety and hygiene systems, processes and practices, the majority of which are commonplace within the food industry.

1.2 Spoilage of foods and beverages; a microbiological approach: microbes vs indigenous enzymes It must be emphasised that the contribution of indigenous food enzymes to spoilage is negligible when compared to the activity of microbial flora. This is mainly the case in food of animal origin (e.g., meat, fish and dairy) (Nychas and Tassou, 1997; Tsigarida and Nychas, 2001). For example, in meat and fish, the post-mortem glycolysis, caused by indigenous enzymes, ceases after the death of the animal when the final pH reaches a value of 5.4±5.5. On the other hand, the indigenous proteolytic and lipolytic enzymes are not sufficient to affect food spoilage. However, these enzymes or other chemical or mechanical means are utilised in the artificial tenderising of meat (Nychas et al., 2007). As far as spoilage due to proteolysis is concerned, the soluble sarcoplasmic proteins probably form the initial substrate for proteolytic attack (Hasegawa et al., 1970a,b; Jay and Shelef, 1976). The proteolytic activity of bacterial action on meat and its impact on spoilage has been clearly demonstrated (Schmitt and Schmidt-Lorenz, 1992a,b; Nychas and Tassou, 1997). Proteolytic bacteria may gain an ecological advantage through penetration which gives them access to newly available resources (e.g., nutrients) which would not be accessible or available to non- or less proteolytic bacteria (Nychas et al., 2007). There is no doubt that microbiological activity is by far the most important factor influencing the changes which cause spoilage in a food system (Nychas et al., 1998). However, it is microbial activity (growth) per se, rather than the


Microbiological spoilage of foods and beverages

5

activity of microbial enzymes and the accumulation of metabolic by-products that identifies food spoilage (Braun and Sutherland, 2004). In the context of meat spoilage, therefore, it is important to include interactions between microbial growth and its respective enzymatic activity.

1.3 Factors affecting the rate of microbiological spoilage of foods and beverages Generally, food spoilage may be considered to be an ecological phenomenon which encompasses changes in the available nutrients (e.g., low molecular compounds) during proliferation of the bacteria which constitute microbial processes in the product regardless of its origin (e.g., animal or plant). The dominance of a particular microbial process on these products depends upon factors which persist during processing, transportation and storage. It is a wellestablished fact that any food ecosystem includes five categories of ecological determinants: intrinsic, processing, extrinsic, implicit and the emergent effect. These influence the establishment of particular microbial processes and determine the rate at which a maximum population is attained. This is known as èphemeral/specific spoilage micro-organisms' (E(S)SO), i.e., those which are able to adopt various ecological strategies (Koutsoumanis and Nychas, 2000; Nychas et al., 2007). These ecological strategies, developed by the ESO, are the consequence of environmental determinants (e.g., stress, the limitation or availability of nutrients and oxygen) and allow them to proliferate in all available niches. In fact, all the determinants mentioned above constitute a virtual ecological niche (n-dimensional) in which an organism is influenced in (micro) space and time (Boddy and Wimpenny, 1992). This ecological approach is pertinent to the understanding of the changes that occur in products throughout the food chain, from farm to fork. In practice, scientists and technologists involved in food industries should attempt to control or modify some or all of the parameters (e.g., temperature) noted above in order to extend the shelf life of these products. In this chapter, emphasis will be placed on implicit (intrinsic biotic parameters) as well as extrinsic factors. 1.3.1 Implicit (intrinsic biotic parameters) factors Ephemeral spoilage organisms (ESO) Among the different types of spoilage, the microbial contributes greatly to the huge amount of food which is wasted and to the associated financial losses (Kantor et al., 1997). As mentioned above, a vast number of studies in food microbiology have established that spoilage can be attributed to a relatively small group of micro-organisms existing in the microbial processes within foods (for a review, see Nychas et al., 1998). This concept has contributed significantly to our understanding of food spoilage. Microbiological spoilage


6


of foods and beverages is caused by a great variety of bacteria, moulds and yeasts. The latter group of micro-organisms, yeasts and moulds can affect a wide range of products which have low pH or water activity. Spoilage from moulds and yeasts is often manifested by their growth on the surface of products such as cheese and meat, as well as by fermentation of sugars in liquid and semi-liquid products. Products with a high sugar or salt content or with low pH, such as soft drinks, syrups, dips, salad dressings and olives, are frequently spoiled by species of Zygosaccharomyces and Torulaspora, Brettanomyces, Saccharomyces, Debaryomyces, Yarrowia and Rhodotorula (Table 1.1). Further to the information provided in the following chapters, Fleet (1992) has provided comprehensive lists of yeast species which have been isolated from various foods and beverages. In view of this diversity of taxa, the correct identification of species is often a challenge. Yeasts also contribute to spoilage in foods of animal origin (e.g., cheese, meat and fish), but this is mainly due to bacterial activity (Table 1.2). The most important factors determining the microbiological quality of foods are: · · · · ·

the the the the the

physiological state of an animal at slaughter, condition of fruits and vegetables at harvesting, spread of contamination during slaughter, processing of both animal and plant origin raw materials, and temperature and other conditions of storage and distribution.

In most raw or fresh foods, a consortium of bacteria, commonly dominated by Pseudomonas spp., is in most cases responsible for spoilage during aerobic storage of these products at different temperatures (ÿ1 to 25 ëC). It is now well established that under aerobic storage three species of Pseudomonas spp., namely P. fragi, P. fluorescens and P. lundensis, are the most important in producing slime and odours as the main signs of spoilage (Stanbridge and Davies, 1998). Cold-tolerant Enterobacteriaceae (e.g., Hafnia alvei, Serratia liquefaciens, Enterobacter agglomerans) also occur in chilled food storage. These bacteria have been found to contribute to spoilage in fresh vegetables, dairy products and foods of animal origin. Lactic acid bacteria have been detected in the aerobic spoilage flora of chilled meat, fish, dairy and freshly cut vegetable products (Holzapfel, 1998). Both lactic acid bacteria and B. thermosphacta are the main, if not the most important, cause of spoilage, which can be recognised as souring rather than putrefaction (Table 1.2). This type of spoilage is one of the two distinct situations that occur in meat and is commonly associated with vacuum or modified atmosphere packaging (MAP) and is the result of competition between facultatively anaerobic Gram-positive flora. The second situation is that of competition between different Gram-negative flora. The physiological attributes of the organisms in the latter case, under imposed ecological determinants, are reported in Nychas et al. (1998).


Microbiological spoilage of foods and beverages Table 1.1

7

Yeasts and moulds in various commodities

Products

Micro-organisms

Fruit juices, fruit concentrates, drinks

Alicyclobacillus acidoterrestris, Saccharomyces cerevisiae, S. bayanus, S. pastorianus, S. kluyveri, S. unisporus, S. exiguous, Z. mellis, Z. rouxii, Lachancea cidri, L. fermentati, L. thermotolerans, Torulaspora delbrueckii, T. microellipsoides, Zygosaccharomyces bailii, Z. lentus Candida albicans, Penicillium expansum, Penicillium funiculosum, Saccharomyces cerevisiae, Mucor plumbeus

Apples, apple juice, apple cider Orange juice Soft drinks Alcoholic beverages (beer, wine and cider), wine, spirits, sweet and sparkling wines High sugar products, honey Products with low sugar and high salt Dried fruit Fruits and vegetables Strawberries, pears, citrus, potatoes, carrots, sweet potatoes, cassava, guavas, yams, kola nuts Vegetable salads, salad dressing, salad vegetable with mayonnaise, condiments, ranch dressing Black olives Bakery products, bread, British bread, sourdough bread Dairy products Butter, European cheeses, cheese, yogurts Meat products

Saccharomyces cerevisiae, Hanseniaspora uvarum Z. bailii, Lachancea fermentati, Torulaspora microellipsoides, Zygosaccharomyces bisporus, Z. kombuchaensis, Z. florentinus Saccharomyces cerevisiae, S. bayanus, P. brevicompactum, A. flavus, Torulaspora delbrueckii, Zygosaccharomyces bailii, Z. lentus, Z. rouxii Zygosaccharomyces bailii, Z. mellis, Z. rouxii, S. cerevisiae, Lachancea thermotolerans, Torulaspora delbrueckii Torulaspora delbrueckii, Zygosaccharomyces bisporus, Z. rouxii Lachancea thermotolerans, Torulaspora delbrueckii, Zygosaccharomyces bailii, Z. rouxii Moulds (Penicillium and Aspergillus) Rhizopus sexualis, Mucor pirifomis, Mucor racemosus, Mucor hiemalis, Mucor circinelloides, Cunninghamella elegans Z. bailii, S. exiguust, S. dairenensis, Z. lentus, Z. bisporus, Torulaspora delbrueckii, S. bayanus, S. unisporus Candida famata Penicillium roqueforti, Hansenula anomala, Pichia anomala (and Candida guilliermondii, C. parapsilosis, Saccharomyces cerevisiae), S. cerevisiae, S. exiguus, S. unisporus, S. bayanus, S. pastorianus S. cerevisiae, S. dairenensis, S. exiguus, S. kluyver, Rhodotorula, Cryptococcus, Candida, Penicillium commune, Mucor racemosus, Mucor circinelloides, Penicillium solitum, Mucor plumbeus, Torulopsis candida, Kluyveromyces fragilis, Mucor hiemalis S. cerevisiae, S. exiguus

Sources: Arias et al. (2002); Basaran et al. (2004); Dennis and Buhagiar (1980); Elez-MartõÂnez et al. (2005); Filtenborg et al. (1996); Fleet (2006); Fleet and Mian (1987); Gouws et al. (2005); Hocking and Faedo (1992); ICMSF (1998); King and Mabbitt (1982); Kurtzman (1990); Kurtzman and Fell (1998), Kurtzman et al. (1971, 2001); Legan and Voysey (1991); Lund et al. (1995); Magan and Aldred (2006); Panagou (2006); Pitt and Hocking (1997); Rankine and Pilone (1973); Rodrigo et al. (2001); Sampedro et al. (2007); Sand and van Grinsven (1976); Spicher (1980); Steels et al. (1999); Suriyarachchi and Fleet (1981); Thomas (1993); Tran and Farid (2004); Van der Horst (2001); Waite et al. (2009) and Wiley (1994).


8


Table 1.2

Microbial association in foods

Products

Conditions

Micro-organisms

Fish

Aerobic storage, 0±4 ëC

Shewanella putrefaciens, Pseudomonas spp. Brochothrix thermosphacta B. thermosphacta, S. putrefaciens Photobacterium phosphoreum, lactic acid bacteria B. thermosphacta, lactic acid bacteria Pseudomonas spp. Leuconostoc gasicomitatum Lactobacillus sakei, Brochothrix thermosphacta, Photobacterium phosphoreum, Aeromonas spp., Serratia spp. Lactobacillus alimentarius

MAP, chill storage >50% CO2 and O2, 0±4 ëC 50% CO2, 0±4 ëC 6.0%, addition of sorbate and/or benzoate, pH < 5.0 Cold-smoked fish, vacuum packed under low NaCl and light acidification Iced fish, high pH Shrimp, brined Milk

Raw Raw, refrigerated Pasteurised Bulk tank sampling From mastitis infected animal

Cream

Pasteurised

Lactobacillus spp., Carnobacterium spp., Photobacterium phosphoreum, psychrotrophic Enterobacteriaceae Shewanella putrefaciens-like organisms Lactic acid bacteria Streptococcus spp., P. fluorescens, P. putida, P. fragi, P. aeruginosa, Staphylococcus spp., Micrococcus spp. Bacillus spp., Paenibacillus spp. B. cereus, B. circulans, B. mycoides, B. licheniformis Streptococcus uberis Streptococcus agalactiae, S. uberis, S. aphaureus Alcaligenes spp., Acinetobacter spp., Aeromonas spp., Enterobacteriaceae

Butter and reduced fat dairy spreads

P. fragi, P. putrefaciens

Dairy products

P. fragi, P. fluorescens, P. hundensis

Cheese

Brine salted, hard and semihard Hard cheese Cheese rind

Clostridium tyrobutyricum, Clostridium spp. P. aeruginosa

Lettuce and ready-to-eat vegetables

Minimally processed

Pseudomonas fluorescens, Pantoea agglomerans, Rahnella aquatilis

Meat and poultry

Aerobic, chill storage

Pseudomonas spp., P. fragi, P. fluorescens, Lactobacillus sakei


Microbiological spoilage of foods and beverages Table 1.2 Products

Continued Conditions

Micro-organisms

Vacuum/MAP packed

Lactic acid bacteria, Enterobacteriaceae, Hafnia alvei, Lactobacillus sakei, L. curvatus Pseudomonas spp., Brochothrix thermosphacta, but also Lactobacillus sakei, L. curvatus, Leuconostoc mesenteroides, Hafnia alvei, Enterobacter amnogenus Clostridium esterteticum, C. algidicarnis Clostridium gasigenes, Cl. Algidixylanolyticum Pseudomonas fragi, P. lundensis, P. fluorescens biovars A, B, C, P. lundensis-like and P. fluorescenslike bacteria Serratia liquefaciens, Hafnia alvei Shewanella putrefaciens, Bro. thermosphacta Clostridium algidicarnis Shewanella putrefaciens Acinetobacter johnsonii, A. lowfii

Beef, aerobic storage, 5 ëC

Beef and pork, vacuum packed Lamb, raw Poultry carcasses

DFD meat, vacuum/high O2/ MAP packed Pork, vacuum packed Fresh meat, high pH Fresh meat and poultry Meat products

Cooked, vacuum packed Modified atmosphere packaging Sliced ham and turkey breast fillets, vacuum packed Blood sausage (Morcilla de Burgos), vacuum/MAP packed

Lactobacillus sakei, Leuconostoc citreum Leuconostoc gasicomitatum, Lactobacillus oligofermentans Leuconostoc mesenteroides subsp. mesentaroides Lactic acid bacteria especially Leuconostoc mesenteroides

Fruit juice

Pasteurised

Alicyclobacillus acideoterrestris

Sous vide products

Mild heat treatment

Spore forming bacteria (Clostridium spp., Bacillus spp.)

Table olives Anchory stuffed

Lactobacillus brevis

Vegetable sausage

Leuconostoc gasicomitatum

Soda bread

9

Partially baked

Bacillus subtilis, B. pumilus, B. licheniformis

Sources: Ben Embarek (1994); BjoÈrkroth et al. (2000); Borch et al. (1996); Bramley et al. (1984); Brocklehurst et al. (1987); Broda et al. (1999, 2000a,b); Chai et al. (1968); Champagne et al. (1994); Chenoll et al. (2007); Cogan and Beresford (2002); Cousin (1982); Dainty and Mackey (1992); Dalgaard (2000); Dalgaard et al. (1993); Deeth et al. (2002); Drosinos and Nychas (1996); Gardner (1981); Gill and Newton (1979); Gram and Huss (1996, 2000); Harmon et al. (1987); Jùrgensen et al. (2000); Kalchayanand et al. (1993); Korkeala et al. (1988); Lafarge et al. (2004); Lahellec and Colin (1979); Legan (1993); Leroi et al. (1998); Lyhs et al. (2004); McMeekin (1977); Meer et al. (1991); Muir (1996); Nguyen-the and Carlin (1994); Ogunnariwo and Hamilton-Miller (1975); Samelis et al. (1998); Santos et al. (2005); Shaw and Latty (1988); Shay and Egan (1981); Stohr et al. (2001); Sundheim et al. (1998); Sutherland et al. (1975); Tekinsen and Rothwell (1974); Torriani et al. (1996); Truelstrup Hansen et al. (1995); Tsigarida and Nychas (2001); Walker and Stringer (1990) and Walls and Chuyate (2000).


10


In general, the metabolic activity of the ephemeral microbial processes which prevail in a food ecosystem under certain aerobic conditions, or are generally introduced during processing, leads to changes or spoilage in the food. These changes or spoilage are related to (i) the type, composition and population of the microbial process, and (ii) the type and availability of the energy substrates in meat. The type and extent of spoilage are governed by the availability of low molecular weight compounds (e.g. glucose, lactate) which exist in meat (Nychas et al., 1998). By the end of this phase, changes and subsequent overt spoilage are due to the catabolism of nitrogenous compounds and amino acids as well as to secondary metabolic reactions. 1.3.2 Extrinsic factors Effect of temperature Temperature seems to be a major factor in influencing spoilage as well as the safety of foods (Nychas et al., 2008). Modern lifestyles, and the development of consumer requirements during the past decade, have led to a significant increase in demand for fresh, high quality food products. The mass consumption of fresh food products of both animal and plant origin, as well as new consumer trends such as reduced cooking times for minimal quality loss, and microwave cooking, have accentuated the need for the constant and systematic control of temperature handling of these products throughout the chill chain from production (slaughterhouse, field) to consumption. Several studies have recently been carried out to assess the importance of handling highly perishable food products at low temperatures. Additionally, emphasis has been placed on the effect which temperature fluctuations or temperature abuses during handling will have on product quality (Koutsoumanis and Taoukis, 2005; Koutsoumanis et al., 2006; McMeekin et al., 2006). Thus an important aspect of the distribution and consumption of food, whether fresh or raw, is the effective monitoring of time/temperature conditions which affect both safety and overall quality (spoilage). It is generally recognised by the European industry, retailers, food authorities and consumers, that there are stages in the chill chain, such as transfer points or storage rooms, which are likely to be the weakest link in the management of chilled or perishable food. All food products, unless appropriately packaged, transported and stored, will spoil in a relatively short time. Food chill chain The chill chain, especially of animal (e.g., fish and meat) and plant origin foods includes two main steps: the primary and secondary chilling. Both steps are important for microbiological stability, eating quality and production yield (Koutsoumanis and Taoukis, 2005). Primary chilling is the process of cooling food, e.g. meat carcasses after slaughter from body to refrigeration temperatures (ca. 3 ëC). During primary chilling, the rapid growth of both pathogenic and spoilage micro-organisms may



11

occur. The rapid reduction of temperature on the food surface can prevent microbial growth and extend the shelf life of a product. It is clear that rapid chilling offers a number of other advantages in product quality and production economics. After primary chilling, any subsequent procedures, such as handling, cutting and mincing, etc., will increase food temperature, therefore secondary chilling is necessary to reduce temperature below 7 ëC. Secondary chilling is also of great importance in pre-cooked food products (e.g., meat, fish, vegetables). Different technologies used to chill food products before transportation are: (i) air chilling, (ii) immersion chilling, (iii) spray chilling, and (iv) vacuum cooling. The effectiveness of air chilling applications depends on a number of factors including air temperature and velocity, relative humidity, the weight and fat cover of the products, as well as product loading. Immersion chilling is probably the least expensive method and provides very rapid cooling with no risk of freezing. Spray chilling is an alternative method to immersion chilling which has been increasingly used, especially in the USA, for meat and meat products (Allen et al., 1987; Johnson et al., 1988). It is based on a combination of sprays and air during the initial stage of the chilling cycle, and the use of air alone in the remainder of the chilling period. Finally, vacuum cooling is a rapid batch process in which moist products (e.g., meat and bakery products, fruits and vegetables) containing free water are cooled by the evaporation of moisture in a vacuum (Mellor, 1980). Rapid cooling in a vacuum has the advantage of significantly reducing the count of phychrophile and mesophile bacteria, even after several days storage (McDonald et al., 2000). Transportation During the meat marketing (transportation) route to the final user for preparation and consumption, food products are stored in tracks, retail cabinets and home refrigerators. These points are of great concern regarding quality and safety. Industrial and track chambers differ in characteristics and performance (Koutsoumanis and Taoukis, 2005). The size of cabinets, initial temperature of the incoming food (which depends on the type of food), targeted storage temperature, temperature of the surroundings, mechanical characteristics (location of refrigeration machinery, compressors, ventilation and insulation) and energy/cost factors are all issues of the greatest importance when considering cold storage requirements. The management approach dominating most food markets is the principle of `first in±first out'. This approach is strictly adhered to in all stages of the chill chain, primarily (but not always) through properly designed handling procedures in the chill storage rooms. The different points of transportation, from cold storage to retail outlets, and then to the consumer's refrigerator, are critical for the product's overall quality and safety. The transporting vehicle must be supplied with a good refrigeration system, and another weak point in distribution is the transportation period from the product purchase to the consumer's refrigerator.


12


1.4 Evaluating, monitoring and measuring microbiological spoilage of foods and beverages So far, a great number of different methodologies, e.g. microbiological, physical and biochemical, have been applied to evaluate food spoilage (Jay, 2000). Among these, microbiological methods have been used almost exclusively in the actual evaluation of spoilage (European Commission, 2005). Initially for the enumeration of microbial populations, the actual number of colonies grown on a Petri dish has played a critical role in the evaluation of food spoilage. Recently this evaluation has been based on existing knowledge of the microbial process which contributes to spoilage in genus (specific spoilage) or species (ephemeral spoilage) level (Nychas et al., 2008; Tassou and Boziaris, 2002; Ercolini et al., 2008). The idea of seeking correlation(s) between microbial growth and the (bio)chemical changes which occur during spoilage has been recognised throughout as a means of revealing specific substrates and/or end products which may be useful for assessing food quality (Jay, 1986; Dainty, 1996; Nychas et al., 1998; Ellis et al., 2002). The ideal indicator (microbial metabolite) should meet, among others, the following criteria (Jay, 1986): the compound (i) should be absent or at least occur at low levels in meat, (ii) should increase with storage time, and (iii) should be produced by the dominant flora and have a good correlation with organoleptic assessment. During the last two decades, numerous attempts have been made to associate certain metabolites with the microbial spoilage of various muscle and vegetable origin foods, and yet there is not a single one available to quantify spoilage/ quality of specific products. There are many reasons for this: for example, (i) the proposed methods are so slow that they give retrospective information and hence cannot be used for on- or at-line monitoring, and (ii) changes in the technology of preservation (e.g., vacuum, modified atmospheres, etc.), are quite likely to affect the application of the chosen methodology. Thus it is evident that the identification of the ideal metabolite for spoilage assessment is a difficult task for the following reasons: (a) most metabolites are specific to certain organisms (e.g. gluconate to pseudomonads) and when these organisms are not present or are inhibited by the food ecology, whether natural or man-made, incorrect spoilage information results; (b) metabolites are the result of the consumption of a specific substrate, but the absence of the given substrate or its presence in low quantities does not preclude spoilage; (c) the rate of microbial metabolite production and the metabolic pathways of these bacteria are affected by the imposed environmental conditions (e.g., pH, oxygen tension, temperature, etc.); (d) the accurate detection and measurement requires sophisticated procedures, highly educated personnel, time and equipment; and (e) many of them provide retrospective information which is not satisfactory.



13

However, it could be stipulated that regardless of the methodology employed for the quantitative evaluation of spoilage, an understanding of where specific metabolites (metabolomics) originate (i.e., responsible organism, substrate), how they are regulated in cell levels (genomics ± proteomics), what is the effect on food characteristics as well as the microbial association on the rate and type of metabolite formation, it is essential to know when and how to exploit them for the benefit of the industry, authorities and consumer. In general, food industries need rapid analytical methods or tools for quantifying these indicators to determine the processing which will be suitable for their raw material and for predicting the remaining shelf life of their products. Inspection authorities need reliable methods for control purposes. Retailers and wholesalers need valid methods to ensure the freshness and safety of their products as well as for the settling of disputes between buyers and sellers. Reliable indicators of the safety and quality status of food from retail to consumption are desirable. It is therefore crucial to have valid methods to monitor freshness and safety so as to ensure what quality is, regardless of whose perspective you take, i.e. that of the consumer, the industry, the inspection authority, or the scientist. Recently, some interesting analytical approaches have been put forward for the rapid and quantitative monitoring of meat spoilage. These include: biosensors (enzymatic reactor systems), electronic noses (array of sensors), Fourier transform infra-red spectroscopy (FT-IR), integration of the FT-MIR Attenuated Total Reflectance bio-sensors or other bio-sensors in tandem with an information platform and development of an èxpert system' to automatically classify the sensory input into a `diagnosis' based on extracted pre-processing features. However, the enormous amount of information provided by the last mentioned technology makes the data produced unmanageable. The application of advanced statistical methods (discriminant function analysis, clustering algorithms, chemo-metrics) and intelligent methodologies (neural networks, fuzzy logic, evolutionary algorithms and genetic programming) may be used as qualitative rather than quantitative indices since their primary aim is to distinguish objects, groups or populations (Goodacre et al., 2004). This is an unsupervised learning method (Ellis and Goodacre, 2001). Nowadays, modern machine learning procedures are based on supervised learning algorithms (Beavis et al., 2000; Goodacre, 2000; Shaw et al., 1999). The last mentioned approach, together with the development of artificial neural networks (ANN), could soon be used for evaluation of food spoilage. However, the quest of the food industry and of food scientists is for techniques and/or instruments that will either rapidly predict or detect microbial spoilage and will therefore eliminate traditional time-consuming and retrospective microbiological methods. Recently, the ability to use mathematical models which describe spoilage has been advanced by developing and validating models which quantitatively estimate the growth of these ephemeral organisms and are consequently able to predict the shelf life of various foods (Devlieghere et al., 2001; Koutsoumanis et al., 2006). However, the food environment can be


14


very complex and it may be difficult to quantify or even categorise some of its features and their potential effects on microbial population dynamics or the ability to recover a target organism from a particular food. For example, food structure (e.g., gels, mayonnaise-based products, drinks, fruit juices), may affect the environmental limits for growth (Koutsoumanis et al., 2004).

1.5 Predicting microbiological spoilage of foods and beverages The perception of spoilage in foods is rather subjective partly because of the lack of general agreement on the early signs of incipient spoilage for every product and partly because of the changes in the technology of food preservation (e.g., vacuum, modified atmospheres, etc.) which makes objective evaluation a difficult task. Shelf life, an indirect measurement of spoilage, may be quantitatively predicted using mathematical models. In the last decade, a significant number of mathematical models for the growth of various spoilage bacteria such as Photobacterium phosphoreum (Dalgaard, 1995; Dalgaard et al., 1997), pseudomonads (Ratkowsky et al., 1982; 1983; Neumeyer et al., 1997; Pin and Baranyi, 1998; Koutsoumanis et al., 2000; Koutsoumanis, 2001), Shewanella putrefaciens (Dalgaard, 1995; Koutsoumanis et al., 2000) and Brochothrix thermosphacta (McClure et al., 1993; Koutsoumanis et al., 2000) have been published. However, despite this progress, predictive spoilage models remain a research tool rather than an effective industrial application (McDonald and Sun, 1999). The reasons for this include: · The lack of information required for the application of models which predict the shelf life of foods (e.g. SSO, spoilage domain, spoilage level); · The development of most models is based on observations in a well-controlled laboratory environment, using microbiological media. Predictions based on such models are not necessarily valid in complex food environments, such as meat, as significant factors for microbial growth in food structure (Robins and Wilson, 1994; Pin et al., 1999; Wilson et al., 2002), and interactions between micro-organisms (Gram and Melchiorsen, 1996; Pin et al., 1999) are not taken into account. As a result, validation of the models in food products often shows a low level of accuracy which limits their application. · The majority of developed models have focused on the effect of environmental factors on the maximum specific growth rate of a micro-organism without taking into account the lag phase. It has been shown, however, that the lag phase duration of the SSO can be a significant part of the total shelf life of foods (Koutsoumanis and Nychas, 2000; Koutsoumanis, 2001). Ignoring the lag phase may lead to an underestimation of shelf life, with significant economic losses for the food industry. · Most models are developed and validated under static temperature conditions. In practice, however, temperature fluctuations are often encountered



15

during the storage and distribution of foods. Thus, validation under changing temperatures is of great importance for evaluating the performance of the model in predicting shelf life under real chill chain conditions. Predictive, or alternatively quantitative, food microbiology (McMeekin et al., 1997) involves knowledge of microbial growth responses to environmental factors expressed in quantitative terms through mathematical equations (models). Data and models can be stored in databases and used to interpret the effect of processing, distribution and storage conditions on microbial growth (McMeekin et al., 1997). This approach offers precision in estimating the shelf life of foods. In addition, the combination of data on the environmental history of the product with the mathematical models may lead to ìntelligent' product management systems for the optimisation of food quality and safety at the time of consumption (Koutsoumanis et al., 2002; 2003; Giannakourou et al., 2001). Finally, to facilitate the implementation of Hazard Analysis Critical Control Point (HACCP) systems in the meat industry, recent attempts have been made to evaluate the risk in the consumption of meat and meat products contaminated with pathogens, especially E. coli O157:H7 (http://www.fsis.usda.gov/OPHS/ ecolrisk/home.htm) and L. monocytogenes. According to existing literature data on meat (specifically ground beef), risk assessments have been conducted for E. coli O157:H7 on hamburgers. These risk assessments aimed either to identify data gaps in evaluating the risk of illness by consumption of contaminated and improperly cooked hamburgers (Marks et al., 1998), or to model the exposure of consumers to this pathogen from farm to fork (Cassin et al., 1998).

1.6 Preventing microbiological spoilage of foods and beverages Various processes and methods have been applied to prevent the microbiological spoilage of foods and beverages. In fact, food preservation may be defined as the process of treating and handling food in such a way as to stop, control or greatly slow down spoilage, and of course, to minimise the possibility of food-borne illness whilst maintaining the optimum nutritional value, texture and flavour. To be effective, preservation must be equal to, or greater than, the microbial `challenge' which the food product presents. 1.6.1 Application of temperature Temperature is the main method used in the food industry for the prevention of spoilage. Low temperature Food products are generally divided into three main categories depending on storage conditions:


16


· Frozen foods. These are often stored at ÿ18 ëC and have a shelf life of six months to two years. At these low temperatures (frozen foods), the growth of micro-organisms is not supported (although they may survive) and therefore the shelf life will not be limited by microbial activity. Deterioration of foods may be due to enzymatic reactions and these are also slowed down at low temperatures. The shelf life is likely to be limited by textural changes such as the formation of ice crystals, by moisture loss or by biochemical changes such as rancidity. · Chilled foods. These are generally stored at temperatures below 8 ëC and typically have a shelf life of one to six weeks, depending on the characteristics of the food product. Storage at chill temperatures will reduce the growth rates of micro-organisms but many spoilage organisms and/or pathogenic bacteria are able to grow at refrigeration temperatures. Additional factors (low pH, water activity, etc.) may be applied to control the activity of these organisms. · Ambient stable foods. These will be subjected to a suitable heat treatment, which is intended to target and destroy micro-organisms in the product. Few products (e.g., fermented, acid foods, etc.) will be sufficiently inhibitory in terms of pH, water activity or preservative level to prevent the growth of micro-organisms likely to survive the heating process. These products will typically have a shelf life of six months to two years. Heat treatment With the exception of raw products, most foods will be subjected to some heat process during manufacture. There are three main categories of heat treatment used to stabilise foods. · Pasteurisation to inactivate vegetative micro-organisms. Typically a process of 70 ëC/2 min or equivalent (z value of 7.5 ëC). The decimal reduction time (D) value is the time required at a given temperature for the surviving population to be reduced by one log cycle or 90%. The z value is the number of degrees Celsius which will result in a 10-fold change in the D value given to chilled food products. With this process, a six-log reduction (6D) in Listeria monocytogenes and other vegetative pathogens can be achieved. It is also sufficient to inactivate most spoilage bacteria such as Enterobacteriaceae, Pseudomonas spp., lactic acid bacteria and yeasts. · Pasteurisation to inactivate psychrotrophic or acid tolerant spore-formers. A process of 90 ëC/10 min or equivalent (z value of 9 ëC) is also applied to chilled food products that are vacuum packed or modified atmosphere packed and which have a shelf life of more than 10 days. With this process, the inactivation of vegetative spoilage bacteria can be achieved in parallel with a six log reduction of psychrotrophic strains of Clostridium botulinum. A process of 95 ëC/5 min or 95 ëC/10 min or equivalent (z value of 8.3 ëC) is applied to acidic ambient stable products. It is designed to inactivate acid-tolerant spore-formers that could grow and spoil the product if they remain present after heat treatment, i.e., Clostridium butyricum, Bacillus polymyxa, etc.



17

· Sterilisation to achieve commercial sterility in canned goods. A process equivalent to 3 min at 121.1 ëC (Fo = 3) based on a z value of 10 ëC is used to achieve a 12-log reduction of mesophilic C. botulinum. It will also inactivate all vegetative micro-organisms and the majority of spore-forming organisms.

1.6.2 Application of carbon dioxide The addition of carbon dioxide has been used to control or inhibit the growth of spoilage micro-organisms in many raw, fresh and treated food products (e.g., pasteurised milk, olives, sausages, cottage, yoghurt and cottage cheese; Chen and Hotchkiss, 1991; Loss and Hotchkiss, 2002; Daniles et al., 1985; Roberts and Torrey, 1988; Nychas and Skandamis, 2005). The packaging system in which combinations of carbon dioxide, nitrogen, oxygen and other harmless gases are introduced within a high barrier film, is called `modified atmosphere packaging' (MAP), and can extend the shelf life and sustain the visual appearance of refrigerated food products. Trace gases such as carbon monoxide, nitrous oxide and sulphur dioxide have also been used. Carbon dioxide may be used in combination with refrigeration, pasteurisation and high barrier packaging to further extend the shelf life of processed milk products without a negative affect on quality (Loss and Hotchkiss, 2002). Much work has focused on the prevention of spoilage in animal origin muscle foods when compared with fruit and vegetables. However, in both types of products, the target organisms were pseudomonads (Nguyen-the and Carlin, 1994; Garcia-Lopez et al., 1998; Holzapfel, 1998; Francis et al., 2011). For example, in refrigerated raw milk, the addition of carbon dioxide may be a valuable technique for controlling the growth of psychrotrophic microflora (e.g. coliforms and pseudomonads) and for reducing the occurrence of heat-resistant microbial proteinases and lipases which diminish the sensory quality of the processed product (King and Mabbitt, 1982; Ruas-Madiedo et al., 1998; Espie and Madden, 1997). However, concerns have been expressed by regulatory authorities (Gill, 1988), food industry groups (Anon., 1988) and others, that this practice may represent an undue safety hazard. Indeed, despite the increasing commercial interest in the use of MAP to extend the shelf life of many perishable products such as meat and poultry, concern about the potential growth of pathogenic bacteria able to survive and grow even at refrigeration temperatures (Silliker and Wolfe, 1980; Palumbo, 1987) remains the limiting factor to further expansion of the method. This is also the case with dairy products (Hotchkiss et al., 1999; Werner and Hotchkiss, 2002).

1.7

Future trends

The introduction of converging technologies in the food industry is among the priorities of the 7th Framework Programme, which is expected to predominate


18


in the future, resulting in substantial changes in the manner in which research is designed. This can be achieved through the integration of modern analytical and high throughput platforms with computational and chemo-metric techniques. Multivariate statistical analyses (e.g., partial least squares regression, discriminant function analysis, cluster analysis) and intelligent methodologies (e.g., artificial neural networks), contribute to the development of a decision support system for prompt determination of the safety and quality of meat products. They may also prevent unnecessary economic losses. Furthermore, the development of computational research platforms and online experimental databases such as ComBase (Baranyi and Tamplin, 2004) and Sym'Previus (Leporq et al., 2005), provide research scientists with a fast and efficient means of storing and exchanging knowledge, whatever their geographic location. The partial least squares analysis (PLS) and artificial neural networks (ANNs) are widely employed modelling approaches due to their ability to relate the input and output variables without pre-knowledge of the system under study, provided that an accurate and adequate amount of data on the system variables is available (Singh et al., 2009). When compared to other areas, the application of ANNs in the field of food science is still in the early development stage (Huang et al., 2007). Nevertheless, interest in using ANNs in food microbiology is increasing as promising results have been produced in several applications, such as growth parameter estimation of micro-organisms (Geeraerd et al., 1998; HervaÂs et al., 2001), bacterial heat resistance (Lou and Nakai, 2001; Esnoz et al., 2006), production of metabolites and simulation of survival curves (Palanichamy et al., 2008; Panagou, 2008). With regard to food safety, the management of food spoilage needs to be applied throughout the supply chain (from `farm to fork' and from `plough to plate'). It must also be implemented during the transition from food product development to manufacture (from `concept to consumer'). A quality assurancebased approach to identifying and controlling relevant spoilage hazards can be integrated with that for safety hazards. The use of a `stable by design' approach and implementation by means of HACCP principles, together with all the associated prerequisite programmes (PRPs), can also be harnessed to help in the management of food spoilage. Food spoilage is part of a continuum involving the identification of potential microbial, chemical and physical hazards, followed by their control to prevent a spectrum of consequences ranging from product spoilage to consumer illness, injury or even death. Consequently, food spoilage should be managed using an integrated approach, but in the context of this book, the focus of this chapter has been on microbial spoilage hazards and their control.

1.8

Sources of further information and advice

Several predictive models for spoilage micro-organisms and food-borne pathogens are available through the internet. Some useful sites include the following.



19

· The Seafood Spoilage Predictor (SSP) program (provided in many languages), available at http://sssp.dtuaqua.dk/ contains predictive models for spoilage of fresh fish (Dalgaard, 1995; Gram and Dalgaard, 2002). · The Food Spoilage Predictor (FSP), which includes the Pseudomonas model of Neumeyer et al. (1997), can be found at http://www.hdl.com.au/ html.body_fsp.htm. · The Pathogen Modelling Program (PMP), for growth predictions for foodborne pathogens is available through a link at http://www.ars.usda.gov/ Services/docs.htm?docid=11550. · The basic free web-based database of food microbiology data is ComBase: www.combase.cc. The ComBase Initiative is a collaboration between the Food Standards Agency and the Institute of Food Research in the United Kingdom; the USDA Agricultural Research Service and its Eastern Regional Research Center in the United States; and the Food Safety Centre in Australia. Moreover, the growth predictor (GP), provided from the UK is available at www.ifr.ac.uk/safety/growthPredictor and the DMfit Program, which matches growth data by linear and non-linear regression with the curve of Baranyi et al. (1993) is accessible through the website of the UK Institute of Food Research at www.ifr.bbsrc.ac.uk. · Many EU-funded projects deal with predictive modelling: ± FAIR CT98-4083 project (accessible in the future through the website of London Metropolitan University (www.londonmet.ac.uk)) ± SMAS (http://smas.chemeng.ntua.gr/start.php?module=overview) ± SYMBIOSIS-EU (www.symbiosis-eu.net).

1.9

References

and GOLLS S J (1987), Èffects of spray chilling and carcass spacing on beef carcass cooler shrink and grade factors', J Anim Sci, 64, 165±170. ANON (1988), `Safety considerations for new generation of refrigerated foods', Nat. Food Processors Assoc. Dairy Food Sanitation, 8, 5±7. ARIAS C R, BURNS J K, FRIEDRICH L M, GOODRICH R M and PARISH M E (2002), `Yeast species associated with orange juice: evaluation of different identification methods', Appl Envir Microb, 68, 1955±1961. BARANYI J and TAMPLIN M L (2004), `ComBase: A common database on microbial responses to food environments', J Food Protect, 67, 1967±1971. BARANYI L, ROBERTS T A and MCCLURE P (1993), À non-autonomous differential-equation to model bacterial-growth', Food Microb, 10, 43±59. BASARAN N, QUINTERO-RAMOS A, MOAKE M M, CHUREY J J and WOROBO RW (2004), Ìnfluence of apple cultivars on inactivation of different strains of Escherichia coli O157:H7 in apple cider by UV irradiation', Appl Envir Microb, 70, 6061±6065. BEAVIS R C, COLBY S M, GOODACRE R, HARRINGTON P B, REILLY J P, SOKOLOW S and WILKERSON C W (2000), Àrtificial intelligence and expert systems in mass spectrometry', in Meyers R A, Encyclopedia of Analytical Chemistry, Chichester, ALLEN D, HUNT M C, LUCHIARI FILHO A, DANLER R J


20

Food and beverage stability and shelf life John Wiley & Sons, 11558±11597. (1994), `Microbiological safety and spoilage of sous vide fish products', PhD Thesis, Royal Veterinary and Agricultural University, Frederiksberg, Denmark.

BEN EMBAREK K

È RKROTH K J, GEISEN R, SCHILLINGER U, WEISS N, DE VOS P, HOLZAPFEL W H, KORKEALA H J BJO

and VANDAMME P (2000), `Characterization of Leuconostoc gasicomitatum sp. nov., associated with spoiled raw tomato-marinated broiler meat strips packaged under modified atmosphere conditions', Appl Envir Microb, 66, 3764±3772. BODDY L and WIMPENNY J W T (1992), Ècological concepts in food microbiology', J Appl Bacteriol, 73, 23S±38S. BORCH E, KANT-MUERMANS M-L and BLIXT Y (1996), `Bacterial spoilage of meat and cured meat products', Int J Food Microbiol, 33, 103±120. BRAMLEY A J, MCKINNON C H, STAKER R T and SIMPKIN D L (1984), `The effect of udder infection on the bacterial flora of the bulk milk of ten dairy herds', J Appl Bacteriol, 57, 317±323. BRAUN P and SUTHERLAND J P (2004), `Predictive modelling of growth and measurement of enzymatic synthesis and activity by a cocktail of Brochothrix thermosphacta', Int J Food Microb, 95, 169±175. BROCKLEHURST T F, ZAMAN-WONG C M and LUND B M (1987), À note on the microbiology of retail packs of prepared salad vegetables', J Appl Bacteriol, 63, 409±415. BRODA D M, LAWSON P A, BELL R G and MUSGRAVE D R (1999), `Clostridium frigidicarnis sp nov., a psychro-tolerant bacterium associated with `blown pack' spoilage of vacuum-packed meats', Int J Syst Evol Bacteriol, 49, 1539±1550. BRODA D M, SAUL D J, LAWSON P A, BELL R G and MUSGRAVE D R (2000a), `Clostridium gasigenes sp nov., a psychrophile causing spoilage of vacuum-packed meat', Int J Syst Evol Microbiol, 50, 107±118. BRODA D M, SAUL D J, BELL R G and MUSGRAVE D R (2000b), `Clostridium algidixylanolyticum sp nov., a psychro-tolerant, xylan-degrading, spore-forming bacterium', Int J Syst Evol Microbiol, 50, 623±631. CASSIN M H, LAMMERDING A M, TODD E C D, ROSS W and MCCOLL R S (1998), `Quantitative risk assessment for Escherichia coli O157:H7 in ground beef hamburgers', Int J Food Microb, 41, 21±44. CHAI T, CHEN C, ROSEN A and LEVIN R E (1968), `Detection and incidence of specific spoilage bacteria on fish. II. Relative incidence of Pseudomonas putrefaciens and fluorescent pseudomonads on haddock fillets', Appl Microbiol, 16, 1738±1741. CHAMPAGNE C P, LAING R R, ROY D, MAFU A A and GRIFFITHS M W (1994), `Psychrotrophs in dairy products: their effects and their control', Crit Rev Food Sci, 34, 1±30. CHEN J H and HOTCHKISS J H (1991), Èffect of dissolved carbon-dioxide on the growth of psychrotrophic organisms in cottage cheese', J Dairy Sci, 74, 2941±2945. Â N M C, ELIZAQUIÂVEL P and AZNAR R (2007), `Lactic acid bacteria CHENOLL E, MACIA associated with vacuum-packed cooked meat product spoilage: population analysis by rDNA-based methods', J Appl Microbiol, 102, 498±508. COGAN T M and BERESFORD T P (2002), `Microbiology of hard cheese', in Robinson R K, Dairy Microbiology Handbook, 3rd edn, Chichester, Wiley-Interscience, 515±560. COUSIN M A (1982), `Presence and activity of psychrotrophic micro-organisms in milk and dairy products: a review', J Food Protect, 45, 172±207. DAINTY R H (1996), `Chemical/biochemical detection of spoilage', Int J Food Microb, 33, 19±34. DAINTY R H and MACKEY B M (1992), `The relationship between the phenotypic properties



21

of bacteria from chill-stored meat and spoilage processes', J Appl Bacteriol, 73, 103±114. DALGAARD P (1995), `Qualitative and quantitative characterization of spoilage bacteria from packed fish', Int J Food Microb, 26, 319±333. DALGAARD P (2000), `Fresh and lightly preserved seafood' in Man C M D and Jones A A, Shelf-Life Evaluation of Foods, 2nd edn, London, Aspen Publishers, 110±139. DALGAARD P, GRAM L and HUSS H H (1993), `Spoilage and shelf life of cod fillets packed in vacuum or modified atmosphere', Int J Food Microbiol, 19, 283±294. DALGAARD P, MEJLHOLM O and HUSS H H (1997), Àpplication of an iterative approach for development of a microbial model predicting the shelf-life of packed fish', Int J Food Microb, 38, 169±179. DANILES J A, KRISHHNAMURTHI R and RIZVI S S H (1985), À review of the effects of carbon dioxide on microbial growth and food quality', J Food Protect, 6, 532±537. DEETH H C, KHUSNIATI T, DATTA N and WALLACE R B (2002), `Spoilage patterns of skim and whole milks', J Dairy Res, 69, 227±241. DENNIS C and BUHAGIAR R W M (1980), `Yeast spoilage of fresh and processed fruits and vegetables', in Skinner F A, Passmore S M and Davenport R R, Biology and Activities of Yeasts, London, Academic Press, 123±133. DEVLIEGHERE F, GEERAERD A H, VERSYCK K J, VANDEWAETERE B, VAN IMPE J and DEBEVERE J (2001), `Growth of Listeria monocytogenes in modified atmosphere packed cooked meat products: a predictive model', Food Microb, 18, 53±66. DROSINOS E H and NYCHAS G-J E (1996), `Brochothrix thermosphacta, a dominant microorganism in Mediterranean fresh fish (Sparus aurata) stored under modified atmosphere', Ital J Food Sci, 4, 323±330. Á -HERNAÁNDEZ J, SOLIVA-FORTUNY R C and MARTIÂN-BELLOSO O ELEZ-MARTIÂNEZ P, ESCOLA (2005), Ìnactivation of Lactobacillus brevis in orange juice by high intensity pulsed electric fields', Food Microbiol, 22, 311±319. ELLIS D I and GOODACRE R (2001), `Rapid and quantitative detection of the microbial spoilage of muscle foods: current status and future trends', Trends Food Sci Tech, 12, 414±424. ELLIS D I, BROADHURST D, KELL D B, ROWLAND J J and GOODACRE R (2002), `Rapid and quantitative detection of the microbial spoilage of meat by Fourier Transform Infrared Spectroscopy and machine learning', Appl Envir Microb, 68, 2822±2828. ERCOLINI D, FRISSO G, MAURIELLO G, SALVATORE F and COPPOLA S (2008), `Microbial diversity in natural whey cultures used for the production of caciocavallo silano PDO cheese', Int J Food Microb, 124, 164±170. ESNOZ A, PERIAGO P M, CONESA R and PALOP A (2006), Àpplication of artificial neural networks to describe the combined effect of pH and NaCl on the heat resistance of Bacillus stearothermophilus', Int J Food Microb, 106, 153±158. ESPIE W E and MADDEN R H (1997), `The carbonation of chilled bulk milk', Milchwissenschaft, 52, 249±253. EUROPEAN COMMISSION (2005), `Commission Regulation (EC) No. 2073/2005 of November 2005 on microbiological criteria for foodstuffs', OJ L 338, 1±26. FILTENBORG O, FRISVAD J C and THRANE U (1996), `Moulds in food spoilage', Int J Food Microbiol, 33, 85±102. FLEET G H (1992), `Spoilage yeasts', Crit Rev Biotechnol, 12, 1±44. FLEET G H (2006), `Saccharomyces and related genera', in Blackburn C, Food and Spoilage Microorganisms, Cambridge, Woodhead Publishing, 311. FLEET G H and MIAN M A (1987), `The occurrence and growth of yeasts in dairy products',


22

Food and beverage stability and shelf life Int J Food Microbiol, 4, 145±155.

and SOFOS J N, (2011), `Factors affecting quality and safety of fresh-cut produce', Crit Rev Food Sci Nutr, in press. GARCIA-LOPEZ M L, PRIETO M and OTERO A (1998), `The physiological attributes of Gramnegative bacteria associated with spoilage of meat and meat products', in Board R G and Davies A R, The Microbiology of Meat and Poultry, London, Blackie Academic and Professional, 1±34. GARDNER G A (1981), `Brochothrix thermospacta in the spoilage of meats: a review', in Roberts T A et al., Psychotrophic Microorganisms in Spoilage and Pathogenicity, London, Academic Press, 139±173. GEERAERD A H, HERREMANS C H, CENENS C and VAN IMPE J F (1998), Àpplication of artificial neural networks as a non-linear modular modelling technique to describe bacterial growth in chilled food products', Int J Food Microb, 44, 49±68. GIANNAKOUROU M, KOUTSOUMANIS K, NYCHAS G-J E and TAOUKIS P S (2001), `Development and assessment of an intelligent shelf life decision system (SLDS) for quality optimization of the food chill chain', J Food Protect, 64, 1051±1057. GILL C O (1988), `The solubility of carbon-dioxide in meat', Meat Sci, 22, 65±71. GILL C O and NEWTON K G (1979), `Spoilage of vacuum-packaged dark, firm, dry meat at chill temperatures', Appl Envir Microb, 37, 362±364. GOODACRE R (2000), Àpplications of artificial neural networks to the analysis of multivariate data', in Cartwright H M, Intelligent Data Analysis in Science: A Handbook, Oxford, Oxford University Press, 123±152. GOODACRE R, VAIDYNATHAN S, DUNN W B, HARRIGAN G G and KELL D B (2004), `Metabolomics by numbers: acquiring and understanding global metabolite data', Trends Biotechnol, 22, 245±252. GOUWS P A, GIE L, PRETORIUS A and DHANSAY N (2005), Ìsolation and identification of Alicyclobacillus acidocaldarius by 16S rDNA from mango juice and concentrate', Int J Food Sci Technol, 40, 789±792. GRAM L and DALGAARD P (2002), `Fish spoilage bacteria ± problems and solutions', Curr Opin Biotech, 13, 262±266. GRAM L and HUSS H H (1996), `Microbial spoilage of fish and fish products', Int J Food Microbiol, 33, 121±137. GRAM L and HUSS H H (2000), `Fresh and processed fish and shellfish', in Lund B M, Baird-Parker A C and Gould G W, The Microbiological Safety and Quality of Foods, 1st edn, London, Chapman and Hall, 472±506. GRAM L and MELCHIORSEN J (1996), Ìnteraction between fish spoilage bacteria Pseudomonas sp. and Shewanella putrefaciens in fish extracts and on fish tissue', J Appl Microb, 80, 589±595. HARMON S M, KAUTTER D A and MCKEE C (1987), `Spoilage of anchovy-stuffed olives by heterofermentative pathway', J Food Safety, 8, 205±210. HASEGAWA T, PEARSON A M, PRICE J F and LECHOWICH R V (1970a), Àction of bacterial growth on the sarcoplasmic and urea soluble proteins from muscle. I. Effects of Clostridium perfringens, Salmonella enteritidis, Achromobacter liquefaciens, Streptococcus feacalis and Kurthia zophii', Appl Microb, 20, 117±122. HASEGAWA T, PEARSON A M, RAMPTON J H and LECHOWICH R V (1970b), Èffect of microbial growth upon sarcoplasmic and urea-soluble proteins from muscle', J Food Sci, 35, 720±724. HERVAÂS C, ZURERA G, GARCIA R M and MARTINEZ J A (2001), Òptimization of computaFRANCIS G A, GALLONE A, NYCHAS G-J, COLELLI G, AMODIO M L, SPANO G



23

tional neural network for its application in the prediction of microbial growth in foods', Food Sci Technol Int, 7, 159±163. HOCKING A D and FAEDO M (1992), `Fungi causing thread mould spoilage of vacuum packaged Cheddar cheese during maturation', Int J Food Microbiol, 16, 123±130. HOLZAPFEL W H (1998), `The Gram-positive bacteria associated with meat and meat products', in Board R G and Davies A R, The Microbiology of Meat and Poultry, London, Blackie Academic and Professional, 35±84. HOTCHKISS J H, CHEN J H and LAWLESS H T (1999), `Combined effects of carbon dioxide addition and barrier films on microbial and sensory changes in pasteurized milk', J Dairy Sci, 82, 690±695. HUANG Y, KANGAS L J and RASKO BA (2007), Àpplications of artificial neural networks (ANNs) in food science', Crit Rev Food Sci, 47, 113±126. INTERNATIONAL COMMISSION ON MICROBIOLOGICAL SPECIFICATION FOR FOODS (ICMSF)

(1998), `Micro-organisms in foods', Volume 6, Microbial Ecology of Food Commodities. New York: Blackie Academic and Professional. JAY J M (1986), `Microbial spoilage indicators and metabolites', in Pierson D and Sterm N J, Foodborne Microorganisms and Their Toxins: Developing Methodology, New York, Marcel Dekker, 219±240. JAY J M (2000), Modern Food Microbiology, 6th edn, Gaithersburg, MD, Aspen Publishers. JAY J M and SHELEF L A (1976), Èffect of micro-organisms on meat proteins at low temperatures', J Agr Food Chem, 24, 1113±1116. JOHNSON J L, DOYLE M P and CASSENS R G (1988), `Survival of Listeria monocytogenes in ground beef', Int J Food Microb, 6, 243±247. JéRGENSEN L V, HUSS H H and DALGAARD P (2000), `The effect of biogenic amine production by single bacterial cultures and metabiosis on cold-smoked salmon', J Appl Microbiol, 89, 920±934. KALCHAYANAND N, RAY B and FIELD R A (1993), `Characteristics of psychrotrophic Clostridium laramie causing spoilage of vacuum-packaged refrigerated fresh and roasted beef', J Food Protect, 56, 13±17. KANTOR L S, LIPTON K, MANCHESTER A and OLIVEIRA V (1997), Èstimating and addressing America's food losses', Food Review, January±April 1997, http://www.ers.usda.gov/ publications/foodreview/jan1997/jan97a.pdf (accessed 8 August 2005). KING J and MABBITT L A (1982), `Preservation of raw milk by the addition of carbon dioxide', J Dairy Res, 49, 439±447. È KELA È P (1988), `Ropy slime formation in vacuum-packed KORKEALA H, SUORTTI T and MA cooked meat products caused by homofermentative lactobacilli and a Leuconostoc species', Int J Food Microbiol, 7, 339±347. KOUTSOUMANIS K P (2001), `Predictive modelling of the shelf life of fish under non isothermal conditions', Appl Envir Microb, 76, 1821±1825. KOUTSOUMANIS K and NYCHAS G J E (2000), Àpplication of a systematic experimental procedure to develop a microbial model for rapid fish shelf life predictions', Int J Food Microb, 60, 171±184. KOUTSOUMANIS K P and TAOUKIS P (2005), `Meat safety, refrigerated storage and transport: modelling and management', in Sofos J N, Improving the Safety of Fresh Meat, Cambridge, Woodhead Publishing, 503±561. KOUTSOUMANIS K P, TAOUKIS P S, DROSINOS E H and NYCHAS G-J E (2000), Àpplicability of an Arrhenius model for the combined effect of temperature and CO2 packaging on the spoilage microflora of fish', Appl Envir Microb, 66, 3528±3534.


24


and NYCHAS G-J E (2002), Àpplication of SLDs (shelf life decision system) to marine cultured fish quality', Int J Food Microb, 73, 375±382. KOUTSOUMANIS K, TAOUKIS P S and NYCHAS G-J E (2003), `Development of a safety monitoring and assurance system (SMAS) for chilled food products', in Van Impe J F M, Geeraerd A H, Leguirinel I and Mafart P, Proceedings of the Fourth International Conference on Predictive Modelling of Foods, France, Quimper, 244±246. KOUTSOUMANIS K P, KENDALL P A and SOFOS J N (2004), À comparative study on growth limits of Listeria monocytogenes as affected by temperature, pH and aw when grown in suspension or on a solid surface', Food Microb, 21, 415±422. KOUTSOUMANIS K P, GEORNARAS I and SOFOS J N (2006), `Microbiology of land muscle food', in Hui Y H, Handbook of Food Science, Technology and Engineering, New York, Marcel Dekker. KURTZMAN C P (1990), `DNA relatedness among species of the genus Zygosaccharomyces', Yeast, 6, 213±219. KURTZMAN C P and FELL J W (eds) (1998), The Yeasts, A Taxonomic Study, 4th edn, Amsterdam, Elsevier Science. KURTZMAN C P, ROGERS R and HESSELTINE C W (1971), `Microbiological spoilage of mayonnaise and salad dressings', Appl Microbiol, 21, 870±874. KURTZMAN C P, ROBNETT C J and BASEHOAR-POWERS E (2001), `Zygosaccharomyces kombuchaensis, a new ascosporogenous yeast from ``Kombucha tea'' ', FEMS Yeast Research, 1 (2), 133±138. LAFARGE V, OGIER J-C, GIRARD V, MALADE N V, LEVEAU J-Y, GRUSS A and DELACROIX-BUCHET A (2004), `Raw cow milk bacterial population shifts attributable to refrigeration', Appl Envir Microb, 70, 5644±5650. LAHELLEC C and COLIN P (1979), `Bacterial flora of poultry. Changes due to variations in ecological conditions during processing and storage', Arch Lebensmittelhyg, 30, 95±98. LEGAN J D (1993), `Mould spoilage of bread: the problem and some solutions', Int Biodet Biodegr, 32, 35±53. LEGAN J D and VOYSEY P A (1991), `Yeast spoilage of bakery products and ingredients', J Appl Bacteriol, 70, 361±371. LEPORQ B, MEMBRE J M, DERVIN C, BUCHE P and GUYONNET J P (2005), `The ``Sym'Previus'' software, a tool to support decisions to the foodstuff safety', Int J Food Microb, 100, 231±237. LEROI F, JOFFRAUD J J, CHEVALIER F and CARDINAL M (1998), `Study of the microbial ecology of cold-smoked salmon during storage at 8 ëC', Int J Food Microbiol, 39, 111±121. LOSS C R and HOTCHKISS J H (2002), Èffect of dissolved carbon dioxide on thermal inactivation of microorganisms in milk', J Food Protect, 65, 1924±1929. LOU W G and NAKAI S (2001), Àpplication of artificial neural networks for predicting the thermal inactivation of bacteria: a combined effect of temperature, pH and water activity', Food Res Int, 34, 573±579. LUND F, FILTENBORG O and FRISVAD J C (1995), Àssociated mycoflora of cheese', Food Microbiol, 12, 173±180. È RKROTH K J (2004), `Leuconostoc gelidum and LYHS U, KOORT J M, LUNDSTROM H S and BJO Leuconostoc gasicomitatum strains dominated the lactic acid bacterium population associated with strong slime formation in an acetic-acid herring preserve', Int J KOUTSOUMANIS K P, GIANNAKOUROU M, TAOUKIS P S



25

Food Microbiol, 90, 207±218. and ALDRED N (2006), `Managing microbial spoilage in cereal and baking products', in Blackburn C, Food Spoilage Microorganisms, Cambridge, Woodhead Publishing, 195. MARKS H M, COLEMAN M E, LIN C T J and ROBERTS T (1998), `Topics in microbial risk assessment: dynamic flow tree process', Risk Anal, 18, 309±328. MCCLURE P J, BARANYI J, BOOGARD E, KELLY T M and ROBERTS T A (1993), À predictive model for the combined effect of pH, sodium chloride and storage temperature on the growth of Brochothrix thermosphacta', Int J Food Microb, 19, 161±178. MCDONALD K and SUN D W (1999), `Predictive food microbiology for the meat industry: a review', Int J Food Microb, 52, 1±27. MCDONALD K, SUN D W and KENNY T (2000), `Comparison of the quality of cooked beef products cooled by vacuum cooling and by conventional cooling', Food Sci Technol, 33, 21±29. MCMEEKIN T A (1977), `Spoilage association of chicken leg muscle', Appl Envir Microb, 33, 1244±1246. MAGAN N

MCMEEKIN T A, BROWN J L, KRIST K, MILES D, NEUMEYER K, NICHOLS D S, OLLEY J, PRESSER K,

and SOONTRANON S (1997), `Quantitative microbiology: a basis for food safety', Emerg Infect Dis, 3, 541±550. MCMEEKIN T A, BARANYI J, BOWMAN J, DALGAARD P, KIRK M, ROSS T, SCHMID S and ZWIETERING M H (2006), Ìnformation systems in food safety management', Int J Food Microb, 112, 181±194. MEER R R, BAKER J, BODYFEL T F W and GRIFFITHS M W (1991), `Psychrotrophic Bacillus spp. in fluid milk products: a review', J Food Protect, 54, 969±979. MELLOR J D (1980), `Vacuum techniques in the food industry', Food Technol Aust, 32, 397±398. MUIR D D (1996), `The shelf life of dairy products: 1. Factors influencing raw milk and fresh products', J Soc Dairy Technol, 49, 24±32. NEUMEYER K, ROSS T and MCMEEKIN T A (1997), `Development of a predictive model to describe the effects of temperature and water activity on the growth of spoilage pseudomonads', Int J Food Microb, 38, 45±54. NGUYEN-THE C and CARLIN F (1994), `The microbiology of minimally processed fresh fruits and vegetables', Crit Rev Food Sci Nutr, 34, 371±401. NYCHAS G-J E and SKANDAMIS P (2005), `Fresh meat spoilage and modified atmosphere packaging (MAP)', in Sofos J N, Improving the Safety of Fresh Meat, Cambridge, Woodhead Publishing, 461±502. NYCHAS G-J E and TASSOU C C (1997), `Spoilage process and proteolysis in chicken as noted by HPLC method', J Sci Food Agr, 74, 199±208. NYCHAS G-J E, DROSINOS E H and BOARD R G (1998), `Chemical changes in stored meat', in Board R G and Davies A R, The Microbiology of Meat and Poultry, London, Blackie Academic and Professional, 288±326. NYCHAS G-J E, MARSHALL D and SOFOS J (2007), `Meat poultry and seafood', in Doyle M P and Beuchat L R, Food Microbiology Fundamentals and Frontiers, 3rd edn, Washington, DC, ASM Press. NYCHAS G-J E, SKANDAMIS P N, TASSOU C C and KOUTSOUMANIS K P (2008), `Meat spoilage during distribution', Meat Sci, 78, 77±89. OGUNNARIWO J and HAMILTON-MILLER J M (1975), `Brown- and red-pigmented Pseudomonas aeruginosa. Differentiation between melanin and pyrubrin', J Med Microbiol, 8, 199±203. RATKOWSKY D A, BOSS T, SALTER M


26


and HOLLEY R A (2008), `Predicting survival of Escherichia coli O157:H7 in dry fermented sausage using artificial neural networks', J Food Protect, 71, 6±12. PALUMBO S A (1987), Ìs refrigeration enough to restrain foodborne pathogens?', J Food Protect, 49, 1003±1005. PANAGOU E Z (2006), `Greek dry-salted olives: monitoring the dry-salting process and subsequent physicochemical and microbiological profile during storage under different packing conditions at 4 and 20 ëC', Lebensm. Wiss. Technol, 39, 322±329. PANAGOU E Z (2008), À radial basis function neural network approach to determine the survival of Listeria monocytogenes in Katiki, a traditional Greek soft cheese', J Food Protect, 71, 750±759. PIN C and BARANYI J (1998), `Predictive models as means to quantify the interactions of spoilage organisms', Int J Food Microb, 41, 59±72. PIN C, SUTHERLAND J P and BARANYI J (1999), `Validating predictive models of food spoilage organisms', J Appl Microb, 87, 491±499. PITT J I and HOCKING A D (1997), Fungi and Food Spoilage, 2nd edn, London, Blackie Academic and Professional. RANKINE B C and PILONE D A (1973), `Saccharomyces bailii, a resistant yeast causing serious spoilage of bottled table wine', Am J Enol Viticult, 24, 55±58. RATKOWSKY D A, OLLEY J, MCMEEKIN T A and BALL A (1982), `Relationship between temperature and growth-rate of bacterial cultures', J Bacteriol, 149, 1±5. RATKOWSKY D A, LOWRY R K, MCMEEKIN T A, STOKES A N and CHANDLER R E (1983), `Model for bacterial culture-growth rate throughout the entire biokinetic temperaturerange', J Bacteriol, 154, 1222±1226. ROBERTS R F and TORREY G S (1988), Ìnhibition of psychrotrophic bacterial growth in refrigerated milk by addition of carbon dioxide', J Dairy Sci, 71, 52±60. ROBINS M M and WILSON P D G (1994), `Food structure and microbial-growth', Trends Food Sci Tech, 5, 289±293. RODRIGO D, MARTIÂNEZ A, HARTE F, BARBOSA-CAÂNOVAS G V and RODRIGO M (2001), `Study of inactivation of Lactobacillus plantarum in orange±carrot juice by means of pulsed electric fields: comparison of inactivation kinetic models', J Food Protect, 62, 259±263. È HR A, LUDDECKE K, DRUSCH S, MULLER M J and ALVENSLEBEN R V (2005), `Food quality RO and safety ± consumer perception and public health concern', Food Control, 16, 649±655. RUAS-MADIEDO P, BASCARAN V, BRANA A F, BADA-GANCEDO J C and DE LOS REYES-GAVILAN C G (1998), Ìnfluence of carbon dioxide addition to raw milk on microbial levels and some fat-soluble vitamin contents of raw and pasteurized milk', J Agr Food Chem, 46, 1552±1555. SAMELIS J, KAKOURI A, GEORGIADOU K G and METAXOPOULOS J (1998), Èvaluation of the extent and type of bacterial contamination at different stages of processing of cooked ham', J Appl Microbiol, 84, 649±660. SAMPEDRO F, RIVAS A, RODRIGO D, MARTIÂNEZ A and RODRIGO M (2007), `Pulsed electric fields inactivation of Lactobacillus plantarum in an orange juice±milk based beverage: effects of process parameters', J Food Eng, 80, 931±938. SAND F E M J and VAN GRINSVEN A M (1976), Ìnvestigation of yeast strains isolated from Scandinavian soft drinks', Brauwissenschaft, 29, 353±355. È RKROTH J (2005), `CharacSANTOS E M, JAIME I, ROVIRA J, LYHS U, KORKEALA H and BJO terization and identification of lactic acid bacteria in morcilla de Burgos', Int J PALANICHAMY A, JAYAS D S



27

Food Microbiol, 97, 285±296. and SCHMIDT-LORENZ W (1992a), `Degradation of amino acids and protein changes during microbial spoilage of chilled unpacked and packed chicken carcasses', Lebensm Wiss Technol, 25, 11±20. SCHMITT R E and SCHMIDT-LORENZ W (1992b), `Formation of ammonia and amines during microbial spoilage of refrigerated broilers', Lebensm Wiss Technol, 25, 6±10. SHAW A D, WINSON M K, WOODWARD A M, MCGOVERN A C, DAVEY H M and KADERBHAI N (1999), `Rapid analysis of high-dimensional bio-processes using multivariate spectroscopies and advanced chemometrics', in Scheper T, Advances in Biochemical Engineering/Biotechnology, Berlin, Springer-Verlag, 83±114. SHAW B G and LATTY J B (1988), À numerical taxonomic study of non-motile, nonfermentative Gram-negative bacteria from foods', J Appl Bacteriol, 65, 7±21. SHAY B G and EGAN A F (1981), `Hydrogen sulfide production and spoilage of vacuumpackaged beef by a Lactobacillus sp.', in Roberts T A, Hobbs G, Christian J H B and Skovgaard N, Psychrotrophic Microorganisms in Spoilage and Pathogenicity, London, Academic Press, 241±251. SILLIKER J H and WOLFE S K (1980), `Microbiological safety considerations in controlledatmosphere storage of meats', Food Technol, 34, 59±63. SINGH K P, OJHA P, MALIK A and JAIN G (2009), `Partial least squares and artificial neural networks modeling for predicting chlorophenol removal from aqueous solution', Chemometr Intell Lab, 99, 150±160. SPICHER G (1980), `Zur AufklaÈrung der Quellen und Wege der Schimmelkontamination des Brotes im Grossbackbetreib', Zentralblatt fuÈr Bakteriologie Parasitenkunde, Infektionskrankheiten und Hygiene, 1 Abt. Original Reiheb Hygiene Betriebshygiene Preventive Medizin, 170, 508±528. STANBRIDGE L H and DAVIES A R (1998), `The microbiology of chill-stored meat', in Board R G and Davies A R, The Microbiology of Meat and Poultry, London, Blackie Academic and Professional, 174±219. STEELS H, BOND C J, COLLINS M D, ROBERTS I N, STRATFORD M and JAMES S A (1999), `Zygosaccharomyces lentus sp. nov., a new member of the yeast genus Zygosaccharomyces barker', Int J Syst Bacteriol, 49, 319±327. STOHR V, JOFFRAUD J J, CARDINAL M and LEROI F (2001), `Spoilage potential and sensory profile associated with bacteria isolated from cold-smoked salmon', Food Res Int, 34, 797±806. SUNDHEIM G, SLETTEN A and DAINTY R H (1998), Ìdentification of pseudomonads from fresh and chill-stored chicken carcasses', Int J Food Microbiol, 39, 185±194. SURIYARACHCHI V R and FLEET G H (1981), Òccurrence and growth of yeasts in yogurts', Appl Envir Microb, 42, 574±579. SUTHERLAND J P, PATTERSON J T and MURRAY J G (1975), `Changes in the microbiology of vacuum-packaged beef', J Appl Bacteriol, 20, 286±298. TASSOU C C and BOZIARIS J S (2002), `Survival of Salmonella enteritidis and changes in pH and organic acids in grated carrots inoculated or not with Lactobacillus sp. and stored under different atmospheres at 4 degrees C', J Sci Food Agr, 82, 1122±1127. TEKINSEN O C and ROTHWELL J (1974), À study of the effect of storage at 5 ëC on the microbial flora of heat-treated market cream', J Soc Dairy Technol, 27, 57±62. THOMAS D S (1993), `Yeasts as spoilage organisms in beverages', in Rose A and Harrison J, The Yeasts, London, Academic Press, 517±561. TORRIANI S, VAN REENEN C A, KLEIN G, REUTER G, DELLAGLIO F and DICKS L M T (1996), `Lactobacillus curvatus subsp. curvatus subsp. nov. and Lactobacillus curvatus SCHMITT R E


28


subsp. Melibiosus subsp. nov. and Lactobacillus sake subsp. sake subsp. nov. and Lactobacillus sake subsp. carnosus subsp. nov., new subspecies of Lactobacillus curvatus Abo-Elnaga and Kandler 1965 and Lactobacillus sake Katagiri, Kitahara, and Fukami 1934 (Klein et al. 1996, emended descriptions), respectively', Int J Syst Bacteriol, 46, 1158±1163. TRAN M T T and FARID M (2004), Ùltraviolet treatment of orange juice', Innov Food Sci Emerg Technol, 5, 495±502. TRUELSTRUP HANSEN L, GILL T and HUSS H H (1995), Èffects of salt and storage temperature on chemical, microbiological and sensory changes in cold-smoked salmon', Food Res Int, 28, 123±130. TSIGARIDA E and NYCHAS G-J E (2001), Ècophysiological attributes of a Lactobacillus sp. and a Pseudomonas sp. on sterile beef fillets in relation to storage temperature and film permeability', J Appl Microb, 90, 696±705. VAN DER HORST H C (2001), `Membrane processing', in Tamime A Y and Law B A, Mechanisation and Automation in Dairy Technology, Boca Raton, FL, CRC Press, 296±317. WAITE J G, JONES J M and YOUSEF A E (2009), Ìsolation and identification of spoilage micro-organisms using food-based media combined with rDNA sequencing: ranch dressing as a model food', Food Microbiol, 26, 235±239. WALKER S J and STRINGER M F (1990), `Microbiology of chilled foods', in Comley T R, Chilled Foods: The State of the Art, Barking, Elsevier, 269±304. WALLS I and CHUYATE R (2000), `Spoilage of fruit juices by Alicyclobacillus acideoterrestris', Food Aust, 52, 286±288. WERNER B G and HOTCHKISS J H (2002), Èffect of carbon dioxide on the growth of Bacillus cereus spores in milk during storage', J Dairy Sci, 85, 15±18. WILEY R C (1994), Ìntroduction to minimally processed refrigerated fruits and vegetables', in Wiley R C, Minimally Processed Refrigerated Fruits and Vegetables, New York, Chapman & Hall, 1±27. WILSON P D G, BROCKLEHURST T F, ARINO S, THUAULT D, JAKOBSEN M, LANGE M, et al. (2002), `Modelling microbial growth in structured foods: towards a unique approach', Int J Food Microbiol, 73, 275±289.


2 Chemical deterioration and physical instability of foods and beverages F. Kong and R. P. Singh, University of California, Davis, USA

Abstract: Food deterioration and spoilage during storage and distribution are a result of a variety of chemical, biochemical, and/or physical changes. These changes are often the result of product composition, environmental factors, and processing conditions. A variety of measures are available to a food technologist to obviate the deteriorative effects of extended storage on food quality. This chapter gives an overview of these chemical and physical reactions. Intrinsic and extrinsic factors affecting these reactions are summarized. Methods to measure and model these changes are discussed. Key words: chemical deterioration, physical instability, shelf life, quality evaluation.

2.1

Introduction

The quality of a food changes over time, which impacts its shelf life. Many deterioration and spoilage problems of foods are related to chemical, biochemical, and/or physical changes, such as lipid oxidation, enzymatic and non-enzymatic browning, and moisture absorption/loss. These reactions change the overall food appearance, texture, and flavor/aroma, and cause loss of nutrients such as vitamins. Some of the commonly observed deteriorative changes include offodors and rancidity developed in fatty foods, browning and darkening of meat and fruit juices, bread staling, and color fading and texture softening in fruits and vegetables (Singh and Anderson, 2004; Yang, 1998; Roos, 2001). These reactions are strongly affected by environmental conditions such as oxygen availability,


30


temperature, relative humidity, as well as food composition including water content, pH, and ingredients. The quality changes can be evaluated by sensory panels, and more frequently, by using instruments, which are more cost effective and have better reproducibility. Good correlation between sensory and instrument methods can be achieved with careful selection of the instrument methods. Approaches to minimize deterioration of foods and extend shelf life include modification of formulation, processing, packaging, and storage conditions. However, foods are unstable in the thermodynamic sense, and tend to change from a low entropy, high enthalpy state to a high entropy, low enthalpy state (van Boekel, 2008). Therefore, the deterioration can slow down but will never stop. The ability to predict the deterioration rates is critical for food processors to estimate the shelf life of food under various storage conditions. Kinetic modeling is an important tool for predicting the changes in food quality. It involves the use of chemical kinetics to study the rates and mechanisms, and the Arrhenius relationship to describe the influence of temperature on the reaction rate constants (van Boekel, 2008; Kong and Chang, 2009). This chapter gives an overview of the chemical and physical reactions that cause quality loss in foods during storage thus limiting their shelf life. Intrinsic and extrinsic factors affecting these reactions are summarized. Methods to measure and model these changes are presented. More details on these topics are reviewed extensively in other articles (Singh and Anderson, 2004; Singh and Cadwallader, 2004). They are also described in other chapters in this book.

2.2 Chemical deterioration and physical instability of foods and beverages A series of chemical and physical changes can occur in foods during storage. Major chemical deterioration of foods include lipid oxidation and hydrolysis that cause rancidity and off-flavor, enzymatic degradation leading to color and texture changes, non-enzymatic browning, light-induced reactions that catalyze lipid oxidation, and protein hydrolysis and oxidation. Some of the common physical changes include mechanical damage of fruits and vegetables during harvesting and post-harvest handling, moisture migration that changes food texture and other physical properties, crystal growth due to temperature fluctuation in frozen stored food products, and viscosity changes and phase separation in emulsion systems such as mayonnaise. These changes can occur simultaneously in food systems, affecting color, flavor, aroma, and/or texture of the food product, leading to reduced shelf life of foods. Table 2.1 presents an overview of some major reactions and their influence on food quality. 2.2.1 Chemical deterioration Lipid oxidation Lipids are the least stable macro-constituents in foods. Oxidative rancidity is one of the major issues affecting the shelf life of fatty foods, especially those


Chemical deterioration and physical instability of foods and beverages

31

Table 2.1 Overview of reactions in foods affecting quality (adapted from van Boekel, 2008) Example

Type of reaction

Consequences

Non-enzymatic browning

Chemical reaction (Maillard reaction)

Color, taste and aroma, nutritive value, formation of toxicologically suspect compounds (acrylamide)

Fat oxidation

Chemical reaction

Loss of essential fatty acids, rancid flavor, formation of toxicologically suspect compounds

Fat oxidation

Biochemical reaction (lipoxygenase)

Off-flavors, mainly due to formation of aldehydes and ketones

Hydrolysis

Chemical reaction

Changes in flavor, vitamin content

Lipolysis

Biochemical reaction (lipase)

Formation of free fatty acids, rancid taste

Proteolysis

Biochemical reaction (proteases)

Formation of amino acids and peptides, bitter taste, flavor compounds, changes in texture

Enzymatic browning

Biochemical reaction of polyphenols

Browning

Separation

Physical reaction

Sedimentation, creaming

Gelation

Combination of chemical and physical reaction

Gel formation, texture changes

containing fatty acids with high levels of unsaturation. The susceptible food products include meats, seafood, fried foods, nuts, mayonnaise, and margarine. Other products include biscuits, cookies, ice cream powder, dried whole milk, dried fruits, milk powder, and coffee (Yang, 1998). Meat, poultry, and seafood are susceptible to oxidative reactions due to relatively high concentrations of unsaturated lipids, heme pigments, metal catalyst, and various other oxidizing agents present in the tissue. Lipid oxidation causes the development of rancidity and `warmed-over' flavors in meat. In addition to the development of off-flavor, the oxidation process also leads to loss of vitamins, alteration in color, degradation of proteins, and even the production of toxic substances (Singh and Cadwallader, 2004; Yang, 1998; Singh and Anderson, 2004). Lipid oxidation occurs when the double bonds of a fatty acid are attacked by oxygen, hydrogen, and enzymes. The general mechanisms of the oxidative processes involve three steps of autoxidation: (i) initiation, (ii) propagation, and (iii) termination, catalyzed by transition metals, enzymes, and photosensitizers. The primary products of lipid oxidation are hydroperoxides, which are unstable compounds that tend to break down into secondary oxidation products, including aldehydes, ketones, and alcohols that are the volatile products causing off-flavor in products (Rahman et al., 2009). Temperature and oxygen are the two critical factors influencing the rate of oxidation. The rate of oxidation increases exponentially with an increase in


32


temperature. Storage at low temperature retards lipid oxidation. Proper package design is important for the package to act as a barrier to oxygen transmission. Fatty acid composition, especially the number and location of double bonds on the fatty acids or triglycerides, greatly affects the rate of oxidation. Light, and trace metals such as copper and iron can greatly catalyze the oxidation reaction. Enzymes such as lipoxygenase can catalyze the reaction between polyunsaturated fatty acids and oxygen to produce hydroperoxides (Gordon, 2004). Heat inactivation, e.g. pasteurization and sterilization of milk, and blanching of vegetables, is a commonly used strategy to inhibit enzyme activity. Presence of water significantly affects lipid oxidation, and the oxidation occurs at high rates at very low water activities. Antioxidants, such as -tocopherol, citric acid, and vitamin C, are able to slow or prevent oxidation by reacting with radical oxygen in a product. Application of tocopherol and oil of rosemary can extend shelf life of salted potato chips from 10 weeks to 12±14 weeks (Yang, 1998). Enzymatic degradation Another mechanism of lipid degradation involves lipolytic/hydrolytic rancidity. For example, lipolytic enzyme lipases catalyze lipolysis resulting in off-odors and off-flavors in foods such as meats and meat products. In this reaction, lipases cleave off free fatty acids (FFA) from triglyceride molecules in the presence of water. The FFAs have shorter chain lengths, lower flavor thresholds, and sometimes off or rancid flavors and odors. For example, lauric acid (C12:0) can be generated in rancid coconut and coconut oil with a strong soapy flavor. Butyric acid, often produced in butter, has a very strong and undesirable odor and flavor at lower levels. Lipid hydrolysis can occur at frozen temperatures (Rahman et al., 2009; RodrõÂguez et al., 2007). A free fatty acid test based on the titration method is commonly used to assess lipid hydrolysis of foods during storage. The lipid hydrolysis reactions can be reduced by minimizing moisture and using heat. Most lipolytic enzymes can be inactivated by heating above 60 ëC. Certain enzyme catalyzing reactions can occur in fruits and vegetables causing degradation in color and texture. Peeled ripe bananas or sliced apples, pears or some vegetables can develop unappealing brown color when exposed to the air. These reactions are catalyzed by phenoloxidase enzymes, which react with phenol compounds and oxygen to form undesirable brown pigments. These reactions take place readily when cells are fractured by bruising, cutting, or peeling (Singh and Anderson, 2004). Minimally processed vegetables are expected to maintain firm and crunchy texture attributes that are associated with freshness and wholesomeness. However, enzyme degradation often leads to a soft or limp product that gives rise to consumer rejection prior to consumption. The reactions involve enzymatic degradation of pectins, in which pectin is first partially demethylated by pectin methylesterase, and later depolymerized by polygalacturonase to polygalacturonic acid causing a loss of firmness (Rico et al., 2007).



33

Non-enzymatic browning Non-enzymatic browning (Maillard reaction) occurs due to the interaction between reducing sugars and amino acids. The reaction scheme involves reactions to form an unstable Schiff's base, then transformation through the Amadori rearrangement. The reactions continue further through the Strecker degradation and polymerization reactions to form volatiles and dark pigments (Singh and Anderson, 2004). This reaction leads to the development of a brown color and the accompanying flavor in foods, which is highly desirable in the baking of bread, brewing of beer and roasting of coffee. However, the browning of foods in storage is an undesirable effect. It mostly occurs in dehydrated and semi-moist foods, such as dried fruits and vegetables, powdered eggs and milk, fruit juice concentrates, certain beverages, jams, jellies, certain canned vegetables, and meat products (Yang, 1998). In addition to the darkening of color, the Maillard reaction also leads to loss of protein solubility, bitter offflavor, textural alterations, and even the production of toxic substances (Yang, 1998; Singh and Cadwallader, 2004). Among the factors that affect browning reactions during storage are the structure of amino acids and sugars, temperature, moisture content and water activity, and pH value (Gordon and Davis, 1998; Arnoldi, 2004). The browning rate increases as water activity (aw) increases, and reaches a maximum at water activity between 0.6 and 0.8. A further increase in water activity, however, results in a decrease in rate due to reactant dilution. The Maillard browning reaction is strongly affected by pH, and tends to occur at high pH. The reaction is also catalyzed with metal ions such as copper and iron (Singh and Anderson, 2004). Light-induced chemical changes Milk, chocolate, butter, and other foods, when exposed to light, such as sunlight or fluorescent light, may develop a characteristic off-flavor caused by photooxidation. Photooxidation may occur due to photolytic free radical autoxidation and/or photosensitized lipid oxidation. Both reaction pathways may lead to formation of free lipid radicals (Mortensen et al., 2004), thus initiating autocatalytic oxidative processes. In particular, dairy products, such as milk and cheese, are very sensitive to light oxidation because of the presence of riboflavin (vitamin B2), which functions as a strong photosensitizer. A photosensitizer is able to absorb visible and UV light and transfer this energy into highly reactive forms of oxygen such as singlet oxygen, subsequently inducing a cascade of oxidation reactions, leading finally to lipid and protein oxidation, significant losses of vitamins and amino acids, discoloration, and formation of strong offflavors, and even toxic products (e.g., cholesterol oxides) (Yang, 1998; Borle et al., 2001). Light-induced oxidative processes are a major cause of deterioration in alcoholic beverages such as wines, degrading the flavor and color (Refsgaard et al., 1995). The potential of photooxidation is related to spectral distribution and intensity of a light source, its wavelength, presence of sensitizers, temperature, exposure time and the amount of available oxygen (IFST, 1993; Mortensen


34


et al., 2004). Appropriate packaging material that properly protects foods from both light and oxygen is important to minimize the photooxidative deterioration of these products. Protein degradation During storage, protein degrades in various ways. Enzyme activity such as proteolysis by protease is an important contributing factor. A protease can cause hydrolysis of the peptide bonds that link amino acids together in the polypeptide chain, thus digesting long protein chains into short fragments. In milk, plasmin is a protease that can cause degradation of dairy proteins, leading to coagulation and gelation (Singh and Anderson, 2004). Oxidation of proteins is another way of protein degradation. In meats, myoglobin and oxymyoglobin can be oxidized into metmyoglobin, resulting in the meat color turning from bright red to brown. This color change can be unappealing to consumers. Protein oxidation is caused by reactive oxygen species that are generated via lipid oxidation, metal- or enzyme-catalyzed oxidative reactions, and other chemical and biological processes. Physical and chemical changes in oxidized proteins include amino acid destruction, decrease in protein solubility due to protein polymerization, loss of enzyme activity, and increases in protein digestibility (Xiong, 2000). Protein oxidation can cause formation of amino acid derivatives, such as carbonyls. Measurement of carbonyl concentration is commonly used to indicate the extent of protein oxidation. Protein oxidation is linked to lipid oxidation, and a significant correlation between carbonyl content and the thiobarbituric acid reactive substances (TBARS) values in meat has been reported (Mercier et al., 1998). 2.2.2 Physical deterioration Mechanical damage Mechanical damage (physical injury), such as bruising of fresh fruits and vegetables, and crushing or breaking of dried snack foods, represents a serious hazard that significantly reduces the value of a product. Physical injury is possibly the most important cause of loss in fruits and vegetables. It can occur during harvesting, packing house operations, handling and transportation, in which the produce is subjected to one or more types of loading compression, impact, and vibration. The damage is caused by various forces such as pressure between fruit and machinery, surface abrasion, and package handling. For most fruit, bruising is a common type of postharvest mechanical injury (PeÂrezVicente et al., 2002). Bruising takes place when the shear stress exceeds the mechanical strength of cells (i.e., yield stress) leading to cell wall ruptures, cell bursting and/or cell deflation as a result of loss of cell fluid (Szczesniak, 1998). Bruising causes water loss, and may induce color changes due to enzymatic browning and microbial growth that occur at the injury area. Damage susceptibility of fruit is determined by their mechanical properties. Studman



35

(1997) noted that apple bruising can result in a typical economic loss in the 10± 25% range, even as high as 50%. Mechanical damage can also occur in other foods. Cracks can develop on the thin coat of the soybean seed during handling, leading to deterioration (Parde et al., 2002). Dry, brittle products, such as crackers, potato chips, and ready-to-eat cereals, are susceptible to breakage during transportation and distribution that can make many products unacceptable. Bruising and breakage can be minimized by the use of well-designed packaging systems which protect the products from vibrations and mechanical damage during distribution and handling. For example, fruit injury can be reduced by using cushioning in a package that will absorb much of the mechanical energy (Singh and Anderson, 2004; Szczesniak, 1998). Moisture change and glass transition Moisture content and water activity are critical factors influencing food stability and shelf life. Water activity (aw) is defined as the vapor pressure of water above a sample (p) divided by that of pure water at the same temperature (p0 ), i.e. aw p=p0 . It describes the degree to which water is free or bound to other components. Water activity depends on the composition, temperature and physical state of the compounds. For a fixed water content, the weaker the water interactions, the greater the water activity, and the product becomes more unstable (Fabra et al., 2009). The relative humidity (RH) of the immediate environment directly affects the moisture content of food. The difference between the RH of the surrounding environment and water activity (aw) of the food determines whether a food gains or loses moisture during storage. The higher the difference between aw and RH during storage, the more potential for moisture migration to or from the environment. The water migration occurs continuously until equilibrium is reached. Moisture migration and the change in the water activity directly impact food shelf life and quality when foods are consumed. Moisture loss causes fresh produce to wilt and shrivel, and experience increased senescence. Freezer burn is also a consequence of moisture migration from the surface of frozen foods. Moisture migration changes food texture. Dry products such as breakfast cereals and potato chips lose their crispness after gaining moisture above the 0.35±0.5 aw range, while dried fruits and bakery goods become unacceptably hard upon losing moisture to below 0.5±0.7 aw (Taoukis et al., 1997). For foods containing multiple domains with widely different sorption isotherms, moisture transfer can occur between components with different water activities until the same aw is reached. Staling of bread results from moisture migration from the crumb (high aw) to the crust (low aw), leading to a dryer, firmer crumb and a tougher and less crisp crust. Deli salads can also deteriorate from migration of water from the vegetable component into the dressing. Moisture transfer can cause changes in glass transition temperature (Tg), a property of food reflecting molecular mobility, significantly affecting stability


36


and shelf life. Glass transition infers material changing from the glassy state to the rubbery state or vice versa, accompanied by changes in thermodynamic properties, molecular mobility, dielectric constant, and mechanical properties. As an illustration, when a flexible rubber band is put into liquid nitrogen, it will change to a solid, brittle and shatterable state; but after warming to room temperature, the rubber band will again become flexible and rubbery (rubbery state). The glass transition concept was originated from polymer science, initially appearing in the literature in the 1960s, but it has now been extensively used with foods since the 1980s (Levine and Slade, 1992). The glass transition concepts and water activity together provide a strong scientific basis for food stability during drying and freezing. Based on the glass transition theory, many low-moisture foods are in the amorphous metastable state, where the material lacks long-range molecular order. These foods include sugar-based products (hard candy, toffee, boiled sweets), dried products (milk and whey powder, fruit juice powder), starchbased products (baked products such as bread, crackers, pasta), and frozen foods. These foods can be in an amorphous glassy state or amorphous rubbery state, depending on temperature and moisture content. Foods in an amorphous glassy state have a high internal viscosity and low internal mobility, whereas in the rubbery state the foods have a viscous, more fluid-like state. Increase in temperature can cause transition of a food from glassy to rubbery state. The temperature (or range of temperatures) where the transition between glassy state and a more fluid-like rubbery state occurs is the glass transition temperature (Tg), which can be determined using differential scanning calorimetry (DSC). A major assumption in food shelf life and quality is that stability of food is maintained in the glassy state. Above Tg, various changes could occur in a food during storage, such as crystallization, collapse and increased stickiness, due to an increase in molecular mobility and decrease in viscosity. The rates of these changes are determined by the temperature difference, T ÿ Tg , i.e. how far the storage temperature T is above Tg (Kalichevsky-Dong, 2000; Meste et al., 2002). Water directly affects Tg by acting as a plasticizer. Water content determines glass transition temperatures of sugars and other carbohydrates in foods, which are common glass formers (Roos, 2010). Tg increases when the water content decreases. For low moisture foods, small amounts of water can decrease Tg tremendously. Absorption of moisture will cause crisp crackers to undergo glass transition and become tough and soggy. On the contrary, food materials can be rendered glassy by removing water. When soft bakery products lose moisture (raised Tg) to the point where they undergo glass transition, they will become glassy, hard, and brittle. Caking of food powder is closely related to its glass transition: when dry powder gains moisture, it undergoes glass transition and becomes amorphous, causing powder to stick together and cake. A recent review published on this subject is given by Roos (2010). The glass transition of food can be identified using a state diagram, a stability map showing different states and phases of a food such as freezing



37

point and glass transition as a function of solids content and temperature. The state diagram defines the moisture content and temperature region at which a food domain is glassy, rubbery, crystalline, or frozen. It helps in understanding the complex changes when the water content and temperature of foods are changed, and also assists in identifying the stability of food during storage as well as selecting suitable conditions of temperature and moisture content for processing. State diagrams have been reported for different fruits, such as grape, strawberry, apple, pineapple, persimmon, and grapefruit (Rahman, 2006; Fabra et al., 2009). Starch gelatinization and retrogradation Starch retrogradation refers to the reassociation or the recrystallization of the polysaccharides in gelatinized starch, i.e. amylase and amylopectin. It occurs when starch-based foods are exposed to freeze/thaw cycles, or when moisture migration occurs in starchy foods, impacting textural and nutritional attributes of foods. Starch retrogradation is one of the main mechanisms for staling of bakery products, increasing firmness of crumb, changing flavor and aroma, and causing loss of crispiness (Morris, 1990). The retrogradation of starch also occurs during the tempering of half-products of many snack products and breakfast cereals, producing textural changes such as increased hardness and reduced stickiness (Farhat, 2000). In addition to the changes in texture and flavor, retrogradation of starch also decreases the starch digestibility. The retrogradation of starch is affected by storage temperature, compositions such as water content, sugars, lipids, salts, and anti-staling enzymes. Bread staling occurs most rapidly at 0±4 ëC. Retrogradation occurs more readily with amylose than with amylopectin since amylose is a smaller unbranched molecule; therefore the use of waxy starches (low amylose content) in foods can decrease the level of retrogradation (Singh and Anderson, 2004). Light microscope and X-ray diffraction analysis are commonly used to study the retrogradation process. Chill injury When fruits and vegetables are stored at low, but non-freezing, temperatures (generally at temperatures of 5±15 ëC), the tissues are unable to carry on with normal metabolic processes. As a result, various physiological and biochemical alterations can occur leading to the development of a variety of chilling injury symptoms, including surface pitting, discoloration, internal breakdown, failure to ripen, growth inhibition, wilting, loss of flavor, and decay (Wang, 1989). Chilling injury is affected by growing conditions, maturity, storage temperature, and storage period. The symptoms develop mainly during fruit ripening after cold storage. For fruits such as peaches and nectarines, major symptoms include mealy or woolly texture, internal browning and reddening, and flesh tissue separation and cavity formation (Lurie and Crisosto, 2005). Chilling injury can be alleviated by various methods such as temperature preconditioning, intermittent warming, chemical treatments, hormonal regula-


38


tion, controlled atmosphere storage, and genetic manipulation (Wang, 1989). Lowering O2 and raising CO2 in the storage atmosphere is effective in delaying or preventing chilling injury (Crisosto et al., 1995). Fruits are injured more rapidly when stored at temperatures between 2.2 and 7.6 ëC, and can be minimized when stored at near or below 0 ëC (Lurie and Crisosto, 2005). Crystal growth Ice crystals formed during freezing affect the quality parameters of stored frozen foods. For frozen meat or fish, growth of ice crystals may lead to high drip loss during thawing and cooking. The rate of crystal growth is more severe when slow freezing processes or multiple freeze/thaw cycles are applied, due to liquid migration from inside of the cell fluids to the extracellular space that contributes to the formation of large damaging extracellular ice crystals, rupturing membranes and disrupting the ultrastructure of cells and tissues. A fast freezing rate minimizes the migration of water into the extracellular spacing and, consequently, promotes formation of smaller intracellular ice. The formation of large ice crystals can also be minimized with the addition of emulsifiers and other water-binding agents. Moreover, when the temperature of the product is kept below its glass transition temperature during frozen storage, water in the food has much less mobility and will tend not to form ice crystals (Singh and Anderson, 2004). Recent studies have indicated that pressure shift freezing (PSF) allows a large and uniform supercooling over the entire volume of the sample and subsequently a much preserved microstructure in food can be achieved. The PSF process includes cooling the sample under pressure to reach a temperature just above its freezing temperature at the applied pressure. Pressure is then released rapidly resulting in supercooling, which enhances instantaneous and homogeneous crystallization throughout the cooled material. PSF is reported to significantly reduce drip loss in seafood (Chevalier et al., 2000). Other problems related to crystal growth include sugar bloom and fat bloom. Sugar bloom occurs when chocolates experience glass transition, i.e., the glassy sugar changes to a rubbery state by uptake of moisture or by increase in temperature. The sugar crystallizes on the surface that give a gray or white appearance. Fat bloom, on the other hand, indicates migration and recrystallization of fat (e.g., cocoa butter) in chocolate which appears as a whitish, greasy haze. This defect is related to improper tempering during the manufacturing process which leads to less stable forms of fat crystals. Recent studies on fat bloom in chocolate have been published by Hodge and Rousseau (2002), and Lonchampt and Hartel (2004). Emulsion breakdown An emulsion is traditionally defined as a dispersion of droplets of one liquid in another, the two being immiscible (Dickinson and Stainsby, 1982). Foods with an emulsion system include mayonnaise, margarine, salad dressings, cake batter, and ice cream. These emulsions are either oil-in-water (O/W) emulsions (cream,



39

dressings) or water-in-oil (W/O) emulsions (butter, margarine). Emulsion stabilization is usually achieved by adding emulsifying agents and thickeners to the emulsion, such as small surfactant molecules (e.g., polysorbates, phospholipids), proteins (e.g., milk proteins), and thickening agents (e.g., gums, gelatin) (Rousseau, 2000). Emulsifiers lower the interfacial tension between the oil and water phases, and/or form a mechanically cohesive interfacial film around the droplets that prevents coalescence (Dickinson, 1992). For example, egg yolk is frequently used as an emulsifier since it contains ends that are both hydrophilic and hydrophobic acting at the surface of the droplets to lower surface tension. Food emulsion is a thermodynamically unstable system. After enough time, an emulsion will collapse, leading to phase separation. The rate at which an emulsion breaks down is strongly influenced by composition, processing condition, and environmental condition (e.g., temperature and pH). Violent vibration, partial freezing or extremely high temperatures all contribute to destabilization of the emulsion. The stability of emulsion can be improved by increasing the viscosity of the continuous phase and reducing the average droplet size in emulsion via homogenization (e.g., milk). Emulsions are usually more stable at lower temperatures due to increased phase viscosity. Some recent literature in this area includes reviews given by Rousseau (2000) and McClements (2007).

2.3 Factors affecting the rate of quality loss due to chemical deterioration and physical instability The rate and extent of physical and chemical reactions depend on many factors, which can be categorized into intrinsic and extrinsic factors. Extrinsic factors are characteristics of the environment as food moves through the food chain, such as temperature, relative humidity, light exposure, and composition of gaseous atmosphere within packaging. Intrinsic factors are characteristics of the food itself, including moisture and water activity, pH and total acidity, availability of oxygen, and additives and preservatives. 2.3.1 Extrinsic factors Temperature Temperature is one of the most important factors affecting the shelf life of foods. Freeze damage can occur in fresh fruits and vegetables. Ice crystals may grow in meats and seafood causing structural disruption and drip loss. Solid fat will become liquid at higher ambient temperature and act as a solvent facilitating a faster deterioration. Increased temperature can also destabilize emulsion systems and change the crystallization characteristics of foods containing sugar syrups. Moreover, increasing temperature generally increases the rate of chemical reactions, resulting in faster deterioration. The dependence of reaction rate on temperature is described by the Arrhenius equation, which will be discussed further later.


40


During storage, foods are often subjected to fluctuating ambient temperature. Storage losses under fluctuating temperature condition can be significantly greater than those at mean temperature. For example, pasta stored under a square wave fluctuating temperature condition had a faster rate for thiamine loss. This is because rates of reactions and physical processes increase exponentially with temperature, thus the average reaction rate under fluctuating temperatures is slightly higher than the reaction rate at the equivalent average temperature (Baeza et al., 2007; Ergun et al., 2010). A change in temperature in relation to Tg will change the physical state of foods and subsequently affect many different reaction rates. As discussed in the previous section, when an amorphous glassy food material is warmed above its Tg, the molecular mobility of the food increases and the food becomes rubbery. Below Tg the molecular mobility is much less and reaction rates are generally much lower, and foods can be regarded as stable (Kalichevsky-Dong, 2000). Relative humidity Relative humidity of air is defined as the ratio of the vapor pressure of air to its saturation vapor pressure. The equilibrium relative humidity (ERH) of a food product is defined as relative humidity of the air surrounding the food that is in equilibrium with its environment. When the equilibrium is obtained, the ERH (in percent) is equal to the water activity multiplied by 100, i.e. ERH (%) aw 100. When a food is exposed to a constant humidity, the product will gain or lose moisture until the ERH is reached. The moisture migration significantly affects the physical and chemical properties of the food, as previously described. Light Food products may be exposed to daylight (or artificial light) at various points in the supply chain. As described in the previous section, light accelerates the oxidation process and therefore the rate at which rancidity develops, especially for fatty foods, causing off-flavor, color fading, or degradation of vitamins. Although transparent packaging materials are generally favored by consumers due to the convenience in observing the contained product, there is an increased risk of light-induced oxidation of foods in such packaging (Mortensen et al., 2004). Packaging The use of appropriate packaging is most important in maintaining the quality of foods and achieving the required shelf life. The principal function of packaging is to protect food from light, oxygen, temperature, moisture, and microorganisms. Packaging also shields food from mechanical damage and protects foods from shock and vibration encountered during distribution. Additionally, food packaging serves to communicate to the consumer information about the food as a marketing tool, and provides consumers with ease of use and convenience (Yam et al., 2005).



41

Controlled and modified atmosphere packaging (MAP), including vacuum packaging, nitrogen flushing or a gas mixture consisting of N2, O2, and CO2, are commonly used to slow down undesirable reactions by limiting the availability of oxygen. The mixture of gases in the package depends on the type of product, packaging materials, and storage temperature. For some packaged foods (such as potato chips), the gas contains 99.9% nitrogen. Fruits and vegetables are respiring products, and the gas atmosphere in a MAP package often consists of N2, O2, and CO2, with lowered level of O2 and an increased level of CO2. This atmosphere can potentially reduce respiration rate, ethylene sensitivity and production, and minimize physiological changes such as oxidation, thus extending the shelf life. The presence of other gases, especially CO2, strongly affects biological and microbial reactions in fresh meats, fruits and vegetables, partly due to increased surface acidification (Singh and Cadwallader, 2004). Comprehensive reviews of controlled and modified atmosphere packaging (CAP/MAP) technology are given by McMillin (2008), Brody et al. (2008) and Fonseca et al. (2002). Materials that have traditionally been used in food packaging include glass, metals (aluminum, foils and laminates, tinplate, and tin-free steel), paper and paperboards, and plastics. Modern food packages often combine several materials to exploit each material's functional or aesthetic properties (Marsh and Bugusu, 2007). The main considerations in selection of packaging materials are gas permeability, water vapor transmission rate, mechanical properties, transparency, and type of package and sealing reliability. Oxygen and moisture permeability of materials are crucial factors when selecting packaging films. For non-respiring products such as meat, fish, and cheese, high barrier films are used. But for fruits and vegetables, permeable films are used to allow gases to transmit from and to the package; proper permeability (for O2 and CO2) of the packaging film that is adapted to the product's respiration rate is critical to establishing a desirable equilibrium modified atmosphere in the package and increasing the shelf life of the product. Some of the new packaging technologies include active and intelligent food packaging. Active packaging allows packages to interact with food and the environment thus playing a dynamic role in food preservation; examples include carbon dioxide absorbers/emitters, odor absorbers, ethylene removers, and aroma emitters (Brody et al., 2008). Intelligent or smart packaging is capable of monitoring food properties or package environment, and informing the processor, retailer and/or consumer of the food status and the environment. Examples include time±temperature indicators (TTIs), ripeness indicators, biosensors, and use of radio frequency identification detectors (McMillin, 2008). A summary of recent innovations in food packaging materials is given by Brody et al. (2008).


42


2.3.2 Intrinsic factors Moisture and water activity Moisture content and water activity (aw) are the most important factors in addition to temperature that affect microbial growth as well as the rate of chemical and physical deteriorative reactions. Water activity and moisture content are correlated through sorption isotherms. Most fresh foods can be considered as high-moisture foods, with more than 50% w/w (water and water) activity of 0.90 to 0.999. These foods include beverages, fresh meat and seafood, dairy products, and fruits and vegetables. Intermediate moisture foods (IMF) have a water content of 10±50%, and aw of 0.60±0.90. These foods include grains, nuts, dehydrated fruits, and a number of processed foods. Table 2.2 summarizes the water activity of some common foods. As described in previous section, the difference between water activity (aw) of the food and the relative humidity (RH) of the immediate environment determines whether a food gains or loses moisture during storage. Similarly, in foods with multiple domains of different aw, migration occurs between the domains until an equilibrium is reached. As mentioned earlier, moisture migration of foods can cause deterioration in texture, promote chemical deterioration reactions, and change molecular stability, thus limiting shelf life of foods. Use of improved packaging materials minimizes moisture migration to the environment. Moisture migration within multi-domain foods can be retarded through the use of edible films and/or reformulation to balance aw of the different domains (Ergun et al., 2010). In foods, water functions as a solvent, reaction medium, and reactant. Increasing aw generally enhances deteriorative reactions. Food deterioration due to microbial growth is not likely to occur at aw < 0.6. However, chemical reactions and enzymatic changes may occur at considerably lower aw values. For example, lipid oxidation occurs at aw < 0.30, and Maillard browning reaction accelerates as the aw increases above 0.25±0.3.

Table 2.2

Summary of the water activity of some common foods

aw

Typical foods

1±0.95 0.95±0.9 0.95±0.85 0.85±0.8 0.8±0.75 0.75±0.65 0.65±0.6 0.5 0.3 0.2

Fresh and canned fruits, meat, milk, breads, fish, cooked sausages Cheese, cured meat Margarine, fermented sausage, sponge cakes Salted meats, syrup, flour Jam, glace fruits Nuts, jelly, molasses Honey, caramel, toffee Pasta Cookies, crackers Dried vegetable



43

The water activity concept proposed that a food product is most stable at its monolayer moisture content. Most reactions have minimal rates at the monolayer value. For example, the rate of lipid oxidation has a minimum rate at aw 0.35, corresponding to a moisture level of 8±10%, which is the water content to form a monolayer. This monolayer acts as a barrier to protect foods from oxygen attack on the unsaturated lipids, which varies with the food composition, structure, and temperature (Esse and Saari, 2004; Rahman, 2009). A food is most stable at and below its glass transition point. As described in the previous section, Tg decreases with increasing aw. The relationship between Tg and aw is often linear for foods with intermediate moisture contents (Rahman, 2009; Kalichevsky-Dong, 2000). pH and total acidity The pH of a food strongly affects protein solubility and functionality. Protein solubility has a direct effect on the reaction behavior, and the solubility of protein is usually at a minimum near the isoelectric point. Changes in food pH also change the shape or charge properties of proteins, thus significantly affecting food storage stability. pH strongly affects enzymatic activity, and each enzyme has a region of pH for optimal stability. Enzymes can completely lose activity in extremely high or low pH environments. The pH of food also impacts the stability of constituents present in the food. For example, anthocyanin, a flavonoid pigment with antioxidant capacity, is most stable under acidic condition. At high pH (e.g., pH 7.01), anthocyanins are unstable and will lose color completely after 20 days' storage at ambient temperature (Pang et al., 2001). Oxygen The presence of oxygen in a package not only facilitates the growth of aerobic microbes and molds, but also triggers or accelerates oxidative reactions that result in food deterioration, developing off-odors, off-flavors, undesirable color changes, and reduced nutritional quality. Oxygen affects both the rate and apparent order of oxidative reactions (Labuza, 1971). Most food packaging is concerned with keeping oxygen out of the pack, by nitrogen flushing, vacuum packaging or modified atmospheric packaging. Oxygen scavengers can be used inside a package to reduce headspace oxygen levels to marginal levels. They are mostly agents that can react with oxygen to reduce its concentration. Ferrous oxide is the most commonly used scavenger. Other oxygen scavengers include ascorbic acid, sulfites, and enzymes such as glucose oxidase (Brody et al., 2008). Sometimes oxygen is needed to maintain desirable food quality characteristics. For beef, oxygen is needed to develop oxymyglobin that maintains the bright red color associated with freshness. The red color is due to the presence of oxymyglobin, which develops whenever the meat is exposed to air (Emblem, 2000).


44


Product formulation and composition The shelf stability of a food is governed by its composition. For example, the presence of fats, especially polyunsaturated fats, will make the product prone to chemical and physical changes. Lipid oxidation of oils can be retarded by blending with other types of oils which are more resistant to reactions. RodrõÂguez et al. (2007) found that blending of Moringa oleifera oil (MOO) with sunflower oil and soybean oil significantly increased the oxidative stability of both the substrate oils. This improvement was attributed to the high contents of oleic acid (C18:1) present in MOO. Oleic acid (C18:1) is more resistant to oxidation compared with polyunsaturated fatty acids (RodrõÂguez et al., 2007). The stability of some food products is related directly to the stability of particular ingredients. Incorporating lactose into confectionery can cause premature crystallization and graining of products such as toffee (Subramaniam, 2000). The shelf life of thiamin-containing beverages can be improved by using an appropriate type of buffer based on the pH of the beverage. For example, at pH 4 and 5, thiamin stability was greater in phosphate buffer than citrate buffer. While in high pH beverages, citrate buffer is better for thiamin stability (Pachapurkar and Bell, 2005). Additives and preservatives are commonly used to maintain food quality and flavor and keep food from spoilage by bacteria and yeasts. More than 3000 food additives and preservatives are available in the market, and the most commonly used ones are salt and sugar. These additives are classified as antimicrobial agents, antioxidants, artificial colors, artificial flavors and flavor enhancers, chelating agents, and thickening and stabilizing agents. Antioxidants including vitamin C, E, butylated hydroxytoluene (BHT), and butylated hydroxyanisole (BHA) are mainly used in foods containing high fats, which are compounds that are able to inhibit oxidation reactions by interrupting the radical chain reaction. Chelating agents such as malic acid, citric acid, and tartaric acid are used to prevent flavor changes, discoloration and rancidity of the foods. Other additives are used as humectants to retain moisture, and emulsifiers to reduce separation of water and oil from products (Subramaniam, 2000). Some of the additives are manufactured from the natural sources such as corn, beet and soybean, others are artificial and man-made. Despite their wide application, the benefits and safety of many artificial food additives (including preservatives) are vigorously debated (Wuttke et al., 2007).

2.4 Measuring chemical deterioration and physical instability of foods and beverages 2.4.1 Sensory panels Sensory evaluation infers measurement, analysis and interpretation of characteristics of food materials as they are perceived by the senses of sight, smell, taste, touch and hearing. It is the most comprehensive way of assessing



45

the quality of food (O'Mahony, 1979). Traditional sensory methods of texture evaluation involved assessment and grading by èxpert' tasters, in which one or two trained èxperts' assign quality scores on the appearance, flavor and texture of the products based on the presence or absence of predetermined defects. The shortcomings of this method include inability of predicting consumer acceptance and the lack of objectivity in quality assessment. Two products with different relative intensities of sensory characteristics may receive similar quality scores based on defects detected by the èxperts' (Claassen and Lawless, 1992). Sensory evaluation by a trained panel usually gives a good estimate of the overall quality of a food. Descriptive analysis is commonly used that deals with the total profile of a food product. It refers to a collection of techniques that seek to discriminate between a range of products based on their sensory characteristics, and to determine a quantitative description of the sensory differences that can be identified. Descriptive analysis requires at least three evaluative processes: discrimination of the trait; description of the trait; and quantitation of the trait. External standards are usually used to define attributes and standardize the scale for each assessor. Developing and refining a vocabulary is an essential part of sensory profile work. Panel training is then performed to increase panelist sensitivity and memory and helps panelists to make valid, reliable judgments independent of personal preferences. Sample testing is usually carried out in replicated (commonly three) sessions, using experimental designs that minimize biases. Descriptive analysis results are subjected to univariate statistics (e.g., multi-way analysis of variance) or multivariate statistics (e.g., principal component analysis) (Hugi and Voirol, 2010; Borgognone et al., 2001). For details of the sensory technique, the reader is referred to the book by Stone and Sidel (2004). Sensory methods have been, and will be for the foreseeable future, the primary means of measuring the range of sensory characteristics of food that are important to consumer acceptance. However, the limitations of sensory testing exist that include high cost, excessive time consumption, high variability, ethical restrictions, and health risk to the panel when exposed to spoiled or potentially hazardous samples. In addition, sensory data are subjective in nature, and the testing results often lack consistency (Singh and Anderson, 2004; Singh and Cadwallader, 2004). These limitations make instrumental methods important in evaluating food quality changes during storage. 2.4.2 Instrumental methods Compared with sensory analysis, instrumental methods usually have improved accuracy and reproducibility (Gordon, 2004). Coupling sensory analysis with chemical analysis data can provide more insights than using either technique alone. A reliable instrumental technique is expected to be well correlated with relevant sensory attributes.


46


Physical measurements Color is the first sensory attribute of most foods that customers can appraise. It often degrades during storage as a result of enzymatic and non-enzymtic reactions, oxidation, and physical reactions. Color is commonly measured using a tristimulus colorimeter or a reflectance spectrophotometer. The color data can be obtained in terms of tristimulus values, chromaticity coordinates, hue, and chroma (Clydesdale, 1998). Good correlation between color and food quality has been reported. For fresh produce, color measurement is one of the few instrumental tests that give results correlating well with consumer assessment of quality (Aked, 2002). Kong and Chang (2009) reported that soybean color can be used to predict soybean quality as well as tofu-making properties. Recently, machine vision systems using a conventional CCD camera have been used in color assessment as well as to categorize products with respect to size and other appearance factors (Chen et al., 2002; Kong et al., 2007b). Machine vision systems are also used to locate bruising in fruits. In particular, spectroscopy and hyperspectral imaging have emerged as powerful techniques in that they greatly enhance our ability to identify materials. These methods can detect subtle and/or minor features of an object that are only sensitive to specific wavelengths (Chen et al., 2002; Van Zeebroeck et al., 2007). Methods of measuring moisture content fall into two categories: direct measurements and indirect measurements. Direct measurement, including the oven-drying method, mostly involves weighing the sample before and after removing the water. Indirect methods measure a property of the food that is itself related to moisture content, for example, the electrical resistance and the dielectric constant of a sample. Water activity values are often obtained by either a capacitance or a dew point hygrometer (Mathlouthi, 2001). Texture is one of the most commonly used physical indicators of food quality. Texture degradation occurs due to moisture migration, enzymatic hydrolysis, and other physical or chemical deteriorations. The texture of a food is often defined based on the stress/strain or force/deformation relationship obtained when food is subjected to an instrumental determination. Most of the instrumental texture measurements involve mechanical tests quantifying the resistance of the food to applied forces, from which quality attributes such as hardness, crispness, and cohesiveness are derived. A large number of instruments are available for testing food texture, and the most popular ones include Instron universal testing machine (Yuan and Chang, 2007), and Texture Technologies' TA.XT2 Universal Texture Analyzer (Kong et al., 2007b). More sophisticated methods are also available, such as the acoustic method, involving measuring the perception of air-conducted sounds to establish its contribution to the sensation of crispness, and recording the sounds while performing a mechanical test (Juodeikiene and Basinskiene, 2004). The acoustic technique is a non-destructive test suitable for on-line texture measurement. Good correlation between sensory and instrumental results for texture can be established when the



47

measurement method is carefully chosen (Rahman et al., 2007). An extensive review of the principles and appliations of texture measuring methods was published by Bourne (2002). The rheological properties of liquid and semi-solid foods are characterized in terms of viscosity, flow behavior index, and consistency index, which may experience significant change during storage. For example, the flow behavior index of concentrated milk significantly changes with storage time (VeÂlez-Ruiz and Barbosa-CaÂnovas, 1998). Rheometers and viscometers are commonly used to quantify the flow properties of the food by measuring the change in stress at either changing or constant shear rate. Glass transition plays a crucial role in modifying the physical property of a food. Glass transition temperature Tg is mostly measured by using differential scanning calorimetry (DSC). DSC defines the glass transition as a change in the heat capacity as the food matrix goes from the glass to the rubbery state. Dynamic mechanical analysis (DMA) is also commonly used as a sensitive and versatile thermal analysis technique which measures the modulus (stiffness) and damping (energy dissipation) properties of materials as the materials are deformed under periodic stress. The DMA storage and loss moduli provide valuable information on food properties, such as the softness of bread, as well as the cooking characteristics of pasta. Other methods, such as X-ray diffraction, microscopy, and dilatometry, are also used to study crystalline structure and glass transition which enable full three-dimensional characterization (Farhat, 2004). Nuclear magnetic resonance (NMR) spectroscopy is increasingly used to monitor the molecular mobility of the components of a food over a range of temperatures encompassing Tg. The principle involves proton relaxometry, which is related to the glass transition of food-related systems (Kou et al., 2000). Chemical measurements Chemical analysis is used to measure the end points of chemical reactions occurring in food during storage, or to confirm the results obtained by the sensory panels. The level of rancidity in lipids is often measured using the peroxide value (PV) and the free fatty acid content (FFA) (Singh and Anderson, 2004). The FFA method measures the liberation of fatty acids as a result of hydrolytic rancidity development. The PV method determines oxidative deterioration of oils by measuring the hydroperoxide, the primary oxidation products. As hydroperoxides quickly decompose to secondary products, PV is often combined with other measurements to reveal the whole picture of oxidation, such as thiobarbituric acid value (TBA). TBA measures malonaldehydes, one of the secondary products of lipid oxidation, representative of aldehydes. Other methods monitoring oxidative deterioration of an oil include using p-anisidine to quantify aldehydes, analyzing conjugable oxidation products (Visioli et al., 1995), and determining octanoate value (Peers and Swoboda, 1979). For all of these tests, standard


48


methods of analysis have been established by several organizations such as the Association of Official Agricultural Chemists (AOAC), the American Oil Chemists' Society (AOCS), the International Union of Pure and Applied Chemistry (IUPAC) and the members of the International Standardization Organization (ISO). Analysis of volatiles in headspace of closed food containers with gas chromatography (GC) methods is a common method to monitor oxidative deterioration and determine fatty acid composition, and to correlate with offflavor. Headspace sampling is widely employed, including static, dynamic headspace or solid-phase microextraction sampling followed by GC separation of volatiles generated during lipid oxidation. Solid-phase microextraction sampling is especially preferred due to relatively simple sample preparation, sensitivity, and rapidity (Singh and Cadwallader, 2004). GC-mass spectrometry (GC-MS), and GC-flame ionization detector (GC-FID), and GC-olfactometry are widely used to analyze the composition of volatiles and to estimate the sensory contribution of a single aroma compound to food flavors (Limpawattana et al., 2008; van Ruth, 2001). Good correlations between sensory assessment and chemical measurements are reported. For example, rancid odor and flavor detected by sensory testing have been correlated with aldehydes, and particularly hexanal, which is therefore called a marker molecule (Fritsch and Gale, 1977; Morales et al., 1997). However, most often sensory descriptions are not related directly to individual compounds (Limpawattana et al., 2008). Consequently, more information on compounds that lead to the prediction of the sensory properties of foods needs to be elucidated. A recent development in detecting odors and aromas is the electronic nose, based on the GC volatile methods. It is able to determine the odor intensity of mixtures of a variety of volatile oil degradation compounds, due to its special detection system consisting of an array of gas sensors (mainly semiconductors). It may function as a rapid and non-destructive tool for on-line flavor characterization, especially for rancidity analysis for foods during storage. The electronic nose has been commercialized, and its use is widely reported for detecting lipid oxidation of foods and change in aroma, such as wine (GarcõÂa et al., 2006), meat (Vestergaard et al., 2007), and peach (Infante et al., 2008). A recent review on this subject was written by Peris and Escuder-Gilabert (2009). An èlectronic tongue' has also been developed that may be used to detect taste and olfaction in foods. An electronic tongue consists of an array of crosssensitive (or partially selective) sensors. Good correlation was found between the instrument output and sensory descriptors pertaining to the global quality of a product (body, overall quality, and astringency) (Rudnitskaya et al., 2009). Recent published reviews on this subject include those by Li et al. (2006) and Ampuero and Bosset (2003).



49

2.5 Predicting and monitoring chemical deterioration and physical instability of foods and beverages The shelf life of many foods is limited by the chemical and physical changes that lead to deterioration in appearance, texture, and odor/flavors. The rate of deteriorative reactions depends on product composition as well as environmental factors, i.e., temperature, water activity, and ambient atmosphere. Quantitative prediction of chemical deterioration and physical instability is critical for estimating the shelf life of food products and designing new processes and packaging. This can be done by the use of mathematical modeling. Mathematical modeling of quality deterioration is commonly conducted to describe the fate of quality indicators as a function of intrinsic and extrinsic factors in the food chain. There are a number of modeling methods, of which kinetic modeling is the most commonly used. Kinetic modeling implies that characteristic kinetic parameters are contained in the mathematical models, such as rate constants and activation energies. Kinetic modeling has been used to characterize microbiological growth, changes in texture and color, as well as chemical/ biochemical reactions in foods during processing and storage. The derived models are either empirical or semi-empirical (van Boekel, 2008). To develop these models, experiments are needed to collect data relating the change in food quality with given storage conditions. The model can be developed by analyzing the experimental data statistically to determine kinetic parameters and seek mathematical relationships. Validation of the model is needed to determine how well it describes the original data. It is important to note that these empirical models may not be valid when used outside the region of influencing parameter on which the models are based. Other types of mathematical modeling are available, such as multivariate statistical tools. They are not covered in this chapter and readers are referred to books or papers in literature such as van Boekel (2000) and Martins and van Boekel (2005).

2.5.1 Kinetic modeling of food quality attributes Modeling chemical reactions The rate of chemical reactions is an important determinant of food quality changes and shelf life. Chemical kinetics involves the study of the rates and mechanisms by which a chemical species converts to another. It is characterized by the rate constant and the order of the reaction. The rate of a chemical reaction (or deterioration of a quality indicator) is defined as the change of concentration of a reactant (or quality factor) (C) at a given time (t): ÿ

dC kC n dt

2:1

where k is the rate constant in appropriate units, and n is the order of the


50


chemical reaction of the quality factor. Solutions of Eq. 2.1 for zero-, first-, or second-order reactions are, respectively: C C0 ÿ kt

2:2

C C0 eÿkt

2:3

1 1 kt C C0

2:4

where C0 is the initial concentration. In zero-order reactions, the rate of loss of the quality factor is constant. An example of a zero-order reaction is the formation of brown color in foods as a result of the Maillard reaction (van Boekel, 2008; Kong and Chang, 2009). First-order reactions are frequently reported for reactions in foods, including lipid oxidation and development of rancidity, microbial growth, vitamin losses in dried foods, and loss of protein quality (Sewald and DeVries, 2003). Second-order reactions are relatively less common. Examples include changes of amino acids involved in the Maillard reaction (van Boekel, 2008) and decay of thiamin during heating (Kong et al., 2007a). Modeling temperature dependence of chemical reactions Increase in storage temperature will accelerate many quality deteriorative reactions in stored food. The relationship between reaction rate constant k and temperature can be described by the Arrhenius equation: k AeÿEa =RT

2:5

where A is a so-called `pre-exponential factor', Ea the activation energy, and R and T the gas constant and absolute temperature, respectively. The Arrhenius equation is derived from thermodynamic laws and statistical mechanics principles, and it is the most prevalent and widely used model describing the temperature dependence of chemical reactions that occur in foods during processing and storage. High activation energy implies that the reaction is strongly temperature dependent, i.e., accelerates greatly with increase in temperature. It should be noted that there are situations where the temperature effect on food quality loss does not follow Arrhenius behavior (Labuza and Riboh, 1982). These situations often involve phase change such as the melting of fats, crystallization of amorphous carbohydrates, and denaturation of proteins as well as increased water activity. These changes may increase or decrease the mobilization of reactants, thus complicating the effect of temperature. For example, temperature decrease may cause crystallization of carbohydrates that will reduce the amount of available sugars for reaction but create more free water for other reactions. It is therefore important to test the validity of the Arrhenius equation whenever it is used for modeling the temperature effect. This is particularly important when the result of accelerated testing is used to estimate the deterioration characteristics under ambient storage conditions (Mizrahi, 2000).



51

An alternative way of expressing temperature dependence of a reaction is to use the concept of `Q10 '. `Q10 ' is defined as the ratio of the reaction rate constants at temperatures differing by 10 ëC (18 ëF). It indicates how faster a reaction will occur if the temperature is raised by 10 ëC, and thus can be used to predict expected product shelf life. For example, if a food attribute is stable for 10 weeks at 30 ëC, and has a Q10 of 2, then its stability at 20 ëC will be 2 10 weeks = 20 weeks. The Q10 and Arrhenius equation together are the principles behind accelerated shelf life testing, a method commonly used for rapid estimation of the shelf life. Determination of kinetic parameters Kinetic parameters of food quality loss are determined through experimental and statistical means, including reaction constant k, reaction order n, activation energy Ea , initial quality C0 , and Q10 . When conducting shelf life experiments, stress variables are defined depending upon the factors that affect the concerned reactions. For example, to study lipid oxidation, influencing factors that should be considered include temperature, water activity, antioxidants, oxygen, pH, and light or even the presence of catalysts. Accelerated testing is often needed when the product shelf life is relatively long (e.g., canned foods). Packaging materials and geometrical shapes significantly affect the rate of quality loss reactions by allowing different levels of heat/mass transfer and light exposure. These factors need to be carefully selected simulating industrial practice. During storage experiment, the quality attributes of food samples are measured at different temperatures and time periods. For each testing temperature, at least five to six data points should be taken over time to make the study results statistically reliable. The higher the storage temperature, the more frequent should be the testing. Determination of the model parameters is usually carried out by statistical regression calculations based on the principles of temperaturedependent chemical reaction kinetics. For details, readers are referred to numerous papers and books in this regard (Sewald and DeVries, 2003; Kong et al., 2007a). Other kinetic models In addition to reaction order and the Arrhenius equation, other models are also available for kinetic modeling of reactions. One of them is Michaelis±Menten equation, which is mostly applied to model enzymatic reactions. As mentioned previously, enzymatic catalyzed degradations cause hydrolytic rancidity, and discoloration and texture degradation in fruits and vegetables. These reactions can be described by Michaelis±Menten kinetics (van Boekel, 2008): v0

vmax S S KM

2:6

where v0 is the initial rate of the reaction, vmax the maximum rate under the conditions studied, [S] is the substrate concentration, and KM the Michaelis constant. vmax and KM are the parameters of the equation. The Michaelis±


52


Menten equation relates the initial reaction rate v0 to the substrate concentration [S]. The rate of the enzymatic reaction can be predicted by nonlinear regression estimation. Glass transition significantly affects food stability through changes between glassy and rubbery state. For foods that undergo glass transitions, the Williams± Landel±Ferry (WLF) model is mostly used to describe the temperature dependence of mechanical properties. Specifically, the WLF equation describes the relationship between viscosity and temperature T and the glass transition temperature Tg0 : C1 T ÿ Tg 2:7 ln C2 T ÿ Tg g where is the viscosity, g, is the viscosity at Tg, and the parameters C1 and C2 are empirical constants. By calculating the viscosity in the glass transition range, the WLF equation relates the molecular mobility of food to the temperature range where glass transition occurs. Molecular mobility is closely related to rate of reactions causing food deterioration. Therefore, the WLF model can be used to obtain valuable information about physical processes such as recrystallization, loss of flavor and desired textural attributes caused by such structural changes (Mizrahi, 2000; van Boekel, 2008).

2.5.2 Time±temperature history The food distribution chain includes several stages involving storage, transport, and handling, where food is often exposed to varying temperatures. Since temperature is one of the most important environmental factors that influence a number of quality attributes in foods, it is critical to know the temperature exposure of a food consignment during storage and distribution. The time± temperature indicator (TTI) is a device that can be attached to foods to indicate the time±temperature history of the food. It is a reliable tool for continuous temperature monitoring and shelf life prediction, and has been selectively used as a food quality monitor for various perishable and semi-perishable foods, particularly chilled and frozen foods which are sensitive to temperature, such as fresh milk, frozen fish, meat, and seafood (Wells and Singh, 1998; Taoukis and Giannakourou, 2004). The principle of TTI operation involves irreversible biological, chemical, or physical reactions that are accelerated at elevated temperature, resembling the temperature dependence of most quality loss reactions of foods (Yan et al., 2008). TTI reflects the cumulative time±temperature history of foods by different means, including development of a specific color or movement of a dye (of known color) along a scale. TTI can indicate full history, partial history, or critical temperature. It is important to know that TTI can reflect the quality status of the food only if the activation energy of quality loss reaction is close to that of the TTI response; therefore successful simulation of the food quality loss kinetics determines the effectiveness of TTI in monitoring



53

quality deterioration (Taoukis and Labuza, 1989; Taoukis and Giannakourou, 2004). More details on TTI devices can be obtained from many books and papers (Taoukis and Labuza, 1989; Yan et al., 2008). With TTI, the time±temperature history of the product can be monitored continuously. This information, plus modeling, primarily kinetic modeling of different deteriorative reactions that occur in food systems, allows us to assess the extent of quality loss of a product, and estimate the remaining shelf life at any point along the distribution chain of products (Taoukis and Giannakourou, 2004). An approach presented by Wells and Singh (1992) involves using the response of a TTI at a constant reference temperature. This information is used along with the activation energy of the indicator to calculate a constant temperature equivalent of the change in the indicator response during the inspection interval. The amount of food quality attribute remaining at the end of the interval is then predicted using the calculated temperature equivalent. Wells and Singh (1992, 1998) conducted detailed research in this area and explored the mathematical derivations that describe this approach for zero-order and firstorder reactions.

2.6 Preventing chemical deterioration and physical instability of foods and beverages As described earlier, chemical and physical reactions occur that lead to food spoilage. Whereas the key factors controlling food stability are temperature, time, and water content, other extrinsic and intrinsic factors, such as pH, light, ingredients, product formulation, oxygen, and packaging, also significantly impact quality changes of food during processing and storage. Deteriorative reactions can be retarded by controlling these factors through the food chain from product manufacturing, processing, to packaging and storage. Strategies that are often employed include control of temperature and water activity; addition of chemicals such as salt, sugar, carbon dioxide, or antioxidants; removal of oxygen; modification of initial headspace gas composition and its retention during distribution and storage; or a combination of these with effective packaging (Brody et al., 2008; Singh and Cadwallader, 2004). The design of product formulation is fundamental to the safety and quality of the food. Specifically, ingredients should be selected or tailored to meet clearly defined quality characteristics. Antioxidants can be used to control oxidation reactions and minimize rancidity. Water activity in foods is critical for food stability, especially for low-moisture foods. Additives (salts, sugars, and glycerol) can be added to lower aw, thus increasing the stability of foods. A variety of processing methods are available for food preservation. Traditional approaches include thermal processing (pasteurization, blanching, cooking, and sterilization), drying, refrigeration (chilling and freezing), extrusion, and separation (filtration, centrifugation, and membranes). New processing technologies are being researched including high hydrostatic pressure, pulsed electric


54


fields, microwave heating, and ohmic heating (Kong and Singh, 2009; Sun, 2005). These technologies provide various options for food preservation and extended shelf life. For example, heat treatments such as blanching effectively inhibit enzymatic activity and preserve color and texture of fruits and vegetables; high hydrostatic pressure can significantly improve the shelf life of minced pressurized albacore muscle (Ramirez-Suarez and Morrissey, 2006); plums treated with putrescine are likely to be more resistant to mechanical damage during handling and packaging (PeÂrez-Vicente et al., 2002). Packaging is an essential element of the food preservation chain. The technique of modified atmospheric packaging (MAP) extends vacuum packaging to more sophisticated gas flush packaging. Roasted coffee can be stored for as long as 18 months after being flushed with CO2 and N2 then vacuum packaged, up from 3 days without packaging (Winger, 2000). As shown earlier in this chapter, light can trigger and accelerate oxidation of unsaturated fatty acids. Therefore, suitable packaging for fatty foods should be designed to reduce the intensity of the incident light, thus retarding or eliminating lightinduced reactions. Beers packed in amber glass bottles have a longer shelf life than those packed in clear glass. When designing modified atmosphere packaging, a number of variables need to be taken into account: the characteristics of the product, the optimum atmosphere composition, the permeability of the packaging materials to gases, sensitivity to temperature, and the respiration rate of the product as affected by different gas composition and temperature (Fonseca et al., 2002). It should be recognized that MAP is a dynamic process, and the gas composition will alter to a certain level after storage is initiated. It is also important to impose tight quality control through package testing to avoid any potential problems that may cause product spoilage. Routine package testing should be conducted to analyze headspace gas composition, oxygen levels, and examine package seals that assure package quality (Fonseca et al., 2002). Active packaging with special gas and moisture absorbers to maintain proper gas composition provide more options (Emblem, 2000). Many of the physical, chemical, or biochemical changes that occur in foods are difficult to control using only one control measure. Hurdle technology, although defined for microbial aspects of the shelf life extension, is also applicable for preservation of food quality. Hurdle technology employs a number of individual hurdles in such a way as to minimize the deterioration reactions. Low temperatures and modified atmosphere packaging together are applied to decrease the rate of quality degradation in minimally processed vegetables during storage. It is also found that the effectiveness of the modified atmosphere on fruits and vegetables may be enhanced by the appropriate use of anti-browning agents (Ragaert et al., 2007). The use of PPP (product, process, and packaging) concepts has been suggested to be the most important overall consideration for frozen product quality, in which precise integration of the product formulation, processing, package, and distribution is required to alleviate food quality loss (Jul, 1984). For example, in manufacturing potato



55

chips, different approaches are combined to minimize deterioration during storage, including use of high solid potato varieties, proper selection of frying oil and the use of antioxidants (Winger, 2000).

2.7

Future trends

One of the aims of current food research is to develop new technologies for high quality shelf-stable food products. An example is microwave sterilization technology. The US Food and Drug Administration (FDA) has recently approved the use of microwave energy for producing pre-packaged, low-acid foods (Harrington, 2010). Other technologies that are being vigorously researched include high-pressure processing, pulsed electric fields, ohmic heating, and ultrasound. These technologies aim to inactivate microorganisms in foods with improved quality attributes, by either reducing heating time (e.g. microwave processing), or non-thermal processing (e.g., high-pressure processing). The reduced heat load will benefit the preservation of nutrients, and reduce the rates of deteriorative reactions such as lipid oxidation, thus extending the shelf life. These technologies also offer the potential for improving existing processes. Packaging is essential for food stability. Synthetic polymers are the most common packaging materials due to their flexibility, light weight and transparency. However, they are non-biodegradable and impose serious ecological problems (Siracusa et al., 2008). With heightened social and environmental consciousness, and strict regulations on pollutants and disposal of municipal solid waste, active research is being conducted to develop innovative packaging approaches. Natural polymers, such as films made of polysaccharides and proteins, are used in packaging to replace petroleum-based polymers. So far, the application of biodegradable films for food packaging is limited because of their poor barrier and weak mechanical properties (Brody et al., 2008; Marsh and Bugusu, 2007). The use of nanomaterials in packaging is being actively studied. Nano-particles such as titanium dioxide and silver when combined into natural packaging materials have shown to be effective in both antibacterial activity and preservation of quality for fresh fruits, thus extending the shelf life of packaged products (Yang et al., 2010). Ongoing research in this area is expected to provide new opportunities to food manufacturers to extend the shelf life of foods with minimal quality loss.

2.8


Several recently published reference books and review articles that are focused on the area of shelf life of foods (in addition to the chapter's references) are as follows:


56


and LASEKAN, O. (2009). The relationship between water activity and fish spoilage during cold storage: a review. Journal of Food, Agriculture and Environment, 7(3±4): 86±90. AMPUERO, S. and BOSSET, J.O. (2003). The electronic nose applied to dairy products: a review. Sensors and Actuators B, 94: 1±12. ESKIN, M. and ROBINSON, D. (2001). Shelf-life Stability: Chemical, Biochemical and Microbiological Changes. Boca Raton, FL: CRC Press. È TKE ENTRUP, M. (2005). Advanced Planning in Fresh Food Industries. Heidelberg: LU Physica-Verlag. MARTINS, R., V. LOPES, V., VICENTE, A. and TEIXEIRA, J.A. (2008). Computational shelf-life dating: complex systems approaches to food quality and safety. Food and Bioprocess Technology, 1(3): 207±222. MESTDAGH F., DE MEULENAER, B., DE CLIPPELEER, J., DEVLIEGHERE, F. and HUYGHEBAERT, A. (2005). Protective influence of several packaging materials on light oxidation of milk. J. Dairy Sci., 88: 499±510. OHLSSON, T. and BENGTSSON, N. (2003). Minimal Processing Technologies in the Food Industry. Weimar, TX: C.H.I.P.S. ROBERTSON, G.L. (2006). Food Packaging: Principles and Practice, 2nd edn. New York: Marcel Dekker. ROBERTSON, G.L. (2009). Food Packaging and Shelf Life: A Practical Guide. Boca Raton, FL: CRC Press. SKIBSTED, L., RISBO, J. and ANDERSEN, M. (2010). Chemical Deterioration and Physical Instability of Food and Beverages. Cambridge: Woodhead Publishing. ABBAS, K.A., SALEH, A.M., MOHAMED, A.

2.9

References

(2002). Maintaining the post-harvest quality of fruits and vegetables. In Jongen, W. (ed.), Fruit and Vegetable Processing: Improving Quality. Cambridge: Woodhead Publishing. AMPUERO, S. and BOSSET, J.O. (2003). The electronic nose applied to dairy products: a review. Sensors and Actuators B, 94: 1±12. ARNOLDI, A. (2004). Factors affecting the Maillard reaction. In Steele, R. (ed.), Understanding and Measuring the Shelf Life of Food. Boca Raton, FL: CRC Press. BAEZA, R., MIELNICKI, D., ZAMORA, M.C. and CHIRIFE, J. (2007). Effect of environmental daily temperature fluctuations over one year storage on the prediction of non-enzymatic browning in reduced-moisture foods stored at àmbient' temperature. Food Control, 18(12): 1532±1537. BORLE, F., SIEBER, R. and BOSSET, J.O. (2001). Photo-oxidation and photoprotection of foods, with particular reference to dairy products: an update of a review article (1993±2000). Sci. Aliments, 21: 571±590. BORGOGNONE, M.G., BUSSI, J. and HOUGH, G. (2001). Principal component analysis in sensory analysis: covariance or correlation matrix? Food Quality and Preference, 12(5±7): 323±326. BOURNE, M. (2002). Food Texture and Viscosity, Food Echnology, International Series, 2nd edn. San Diego, CA: Academic Press. BRODY, A.L., BUGUSU, B., HAN, J.H., SAND, C.K. and MCHUGH, T.H. (2008). Scientific status summary. Journal of Food Science, 73(8): R107±R116. AKED, J.



57

and KIM, M.S. (2002). Machine vision technology for agricultural applications. Computers and Electronics in Agriculture, 36(2±3): 173±191. CHEVALIER, D., SEQUEIRA-MUNOZ, A., LE BAIL, A., SIMPSON, B.K. and GHOUL, M. (2000). Effect of freezing conditions and storage on ice crystal and drip volume in turbot (Scophthalmus maximus): evaluation of pressure shift freezing vs. air-blast freezing. Innovative Food Science & Emerging Technologies, 1(3): 193±201. CLAASSEN, M.I. and LAWLESS, H.A.T. (1992). Comparison of descriptive terminology systems for sensory evaluation of fluid milk. Journal of Food Science, 57(3): 596±600. CLYDESDALE, M.F. (1998). Color: origin, stability, measurement and quality. In Irwin, A.T. and Singh, R.P. (eds), Food Storage Stability. Boca Raton, FL: CRC Press. CRISOSTO, C.H., MITCHELL, F.G. and JOHNSON, R.S. (1995). Factors in fresh market stone fruit quality. Postharvest News Inform, 5: 17N±21N. DICKINSON, E. (1992). An Introduction to Food Colloids. Oxford: Oxford University Press. DICKINSON, E. and STAINSBY, G. (1982). Colloids in Food. London: Applied Science. EMBLEM, A. (2000). Predicting packaging characteristics to improve shelf-life. In Kilcast, D. and Subramaniam, P. (eds), The Stability and Shelf-life of Food. Cambridge: Woodhead Publishing. ERGUN, R., LIETHA, R. and HARTEL, R.W. (2010). Moisture and shelf life in sugar confections. Critical Reviews in Food Science and Nutrition, 50(2): 162±192. ESSE, R. and SAARI, A. (2004). Shelf life and moisture management. In Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Cambridge: Woodhead Publishing. FABRA, M.J., TALENS, P., MORAGA, G. and MARTIÂNEZ-NAVARRETE, N. (2009). Sorption isotherm and state diagram of grapefruit as a tool to improve product processing and stability. Journal of Food Engineering, 93(1): 52±58. FARHAT, I.A. (2000). Advanced instrumental methods: the use of H relaxation NMR to monitor starch retrogradation. In Kilcast, D. and Subramaniam, P. (eds), The Stability and Shelf-life of Food. Cambridge: Woodhead Publishing. FARHAT, I.A. (2004). Measuring and modelling the glass transition temperature. In Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Cambridge: Woodhead Publishing. FONSECA, S.C., OLIVEIRA, F.A.R. and BRECHT, J.K. (2002). Modelling respiration rate of fresh fruits and vegetables for modified atmosphere packages: a review. Journal of Food Engineering, 52(2): 99±119. FRITSCH, C. and GALE, J. (1977). Hexanal as a measure of rancidity in low fat foods. Journal of the American Oil Chemists' Society, 54(6): 225±228. GARCIÂA, M., ALEIXANDRE, M., GUTIEÂRREZ, J. and HORRILLO, M.C. (2006). Electronic nose for wine discrimination. Sensors and Actuators B: Chemical, 113(2): 911±916. GORDON, J. and DAVIS, E.A. (1998). Biochemical processes: carbohydrate instability. In Taub, I.A. and Singh, R.P. (eds), Food Storage Stability. Boca Raton, FL: CRC Press, pp. 191±244. GORDON, M.H. (2004). Factors affecting lipid oxidation. In Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Cambrdige: Woodhead Publishing. HARRINGTON, R. (2010). Microwave sterilisation system may revolutionise food processing. http://www.foodqualitynews.com/Innovation/Microwave-sterilisationsystem-may-revolutionise-food-processing (accessed 3/11/10). HODGE, S.M. and ROUSSEAU, D. (2002). Fat bloom formation and characterization in milk chocolate observed by atomic force microscopy. J Am Oil Chem Soc, 79: 1115± 1121. CHEN, Y.-R., CHAO, K.


58


and VOIROL, E. (2010). Instrumental measurements and sensory parameters. In Kress-Rogers, E. and Brimelow, C.J.B. (eds), Instrumentation and Sensors for the Food Industry, 2nd edn. Boca Raton, FL: CRC Press. IFST (1993). Shelf Life of Foods ± Guidelines for its Determination and Prediction. London: Institute of Food Science and Technology. INFANTE, R., FARCUH, M. and MENESES, C. (2008). Monitoring the sensorial quality and aroma through an electronic nose in peaches during cold storage. Journal of the Science of Food and Agriculture, 88: 2073±2078. JUL, M. (1984). The Quality of Frozen Foods, Orlando, FL: Academic Press, pp. 33± 153. JUODEIKIENE, G. and BASINSKIENE, L. (2004). Non-destructive texture analysis of cereal products. Food Research International, 37(6): 603±610. KALICHEVSKY-DONG, M.T. (2000). The glass transition and microbial stability. In Kilcast, D. and Subramaniam, P. (eds), The Stability and Shelf-life of Food. Cambridge: Woodhead Publishing. KONG, F. and CHANG, S.K.C. (2009). Statistical and kinetic studies of the changes in soybean quality during storage as related to soymilk and tofu making. Journal of Food Science, 74(2): S81±S89. KONG, F. and SINGH, R.P. (2009). Emerging food technologies. In Stadler, R. and Lineback, D. (eds), Process-Induced Food Toxicants: Occurrence, Formation, Mitigation, and Health Risks. New York: Wiley. KONG, F., TANG, J., RASCO, B. and CRAPO, C. (2007a). Kinetics of salmon quality changes during thermal processing. Journal of Food Engineering, 83(4): 510±520. KONG, F., TANG, J., RASCO, B., CRAPO, C. and SMILEY, S. (2007b). Quality changes of salmon (Oncorhynchus gorbuscha) muscle during thermal processing. Journal of Food Science, 72(2): S103±S111. KOU, Y., DICKINSON, L.C. and CHINACHOTI, P. (2000). Mobility characterization of waxy corn starch using wide-line 1H nuclear magnetic resonance. Journal of Agricultural and Food Chemistry, 48(11): 5489±5495. LABUZA, T.P. (1971). Kinetics of lipid oxidation in foods. CRC Review of Food Technology, 2: 335±405. LABUZA T.P. and RIBOH D. (1982). Theory and application of Arrhenius kinetics to the rediction of nutrient losses in foods. Food Technology, 36: 66±74. LEVINE, H. and SLADE, L. (1992). Glass transiton in foods. In Schwartzberg, H.G. and Hartel, R.W. (eds), Physical Chemistry of Foods. New York: Marcel Dekker. LI, Z., WANG, N. and VIGNEAULT, C. (2006). Electronic nose and electronic tongue in food production and processing. Stewart Postharvest Review, 4(7): 7.1±7.5. LIMPAWATTANA, M., YANG, D.S., KAYS, S.J. and SHEWFELT, R.L. (2008). Relating sensory descriptors to volatile components in flavor of specialty rice types. Journal of Food Science, 73(9): S456±S461. LONCHAMPT, P. and HARTEL, R.W. (2004). Fat bloom in chocolate and compound coatings. European Journal of Lipid Science and Technology, 106(4): 241±274. LURIE, S. and CRISOSTO, C.H. (2005). Chilling injury in peach and nectarine. Postharvest Biology and Technology, 37(3): 195±208. MARSH, K. and BUGUSU, B. (2007). Food packaging ± roles, materials, and environmental issues. Journal of Food Science, 72(3): R39±R55. MARTINS, S.I.F.S. and VAN BOEKEL, M.A.J.S. (2005). Kinetics of the glucose±glycine Maillard reaction pathways: influences of pH and reactant initial concentrations. Food Chem, 92: 437±448. HUGI, A.



59

(2001). Water content, water activity, water structure and the stability of foodstuffs. Food Control, 12(7): 409±417. MCCLEMENTS, D.J. (2007). Critical review of techniques and methodologies for characterization of emulsion stability. Critical Reviews in Food Science and Nutrition, 47(7): 611±649. MCMILLIN, K.W. (2008). Where is MAP going? A review and future potential of modified atmosphere packaging for meat. Meat Science, 80(1) : 43±65. MERCIER, Y., GATELLIER, P., VIAU, M., REMIGNON, H. and RENERRE, M. (1998). Effect of dietary fat and vitamin E on colour stability and on lipid and protein oxidation in Turkey meat during storage. Meat Science, 48(3±4): 301±318. MESTE, M.L., CHAMPION, D., ROUDAUT, G., BLOND, G. and SIMATOS, D. (2002). Glass transition and food technology: a critical appraisal. Journal of Food Science, 67(7): 2444± 2458. MIZRAHI, S. (2000). Accelerated shelf-life tests. In Kilcast, D. and Subramaniam, P. (eds), The Stability and Shelf-life of Food, Cambridge: Woodhead Publishing. MORALES, M.T., RIOS, J.J. and APARICIO, R. (1997). Changes in the volatile composition of virgin olive oil during oxidation: flavors and off-flavors. Journal of Agricultural and Food Chemistry, 45(7): 2666±2673. MORRIS, V.J. (1990). Starch gelation and retrogradation. Trends in Food Science & Technology, 1: 2±6. MORTENSEN, G., BERTELSEN, G., MORTENSEN, B.K. and STAPELFELDT, H. (2004). Lightinduced changes in packaged cheeses ± a review. International Dairy Journal, 14(2): 85±102. O'MAHONY, M. (1979). Psychophysical aspects of sensory analysis of dairy products: a critique. Journal of Dairy Science, 62: 1954±1962. PACHAPURKAR, D. and BELL, L.N. (2005). Kinetics of thiamin degradation in solutions under ambient storage conditions. Journal of Food Science, 70(7): c423±c426. PANG, X.Q., ZHANG, Z.Q., DUAN, X.W. and JI, Z.L. (2001). Influence of pH and active oxygen on the stability of anthocyanins from litchi pericarp. Acta Hort. (ISHS), 558: 339± 342 PARDE, S.R., KAUSAL, R.T., JAYAS, D.S. and WHITE, N.D.G. (2002). Mechanical damage to soybean seed during processing. J. Stored Prod. Res., 38: 385±394. PEERS, K.E. and SWOBODA, P.A.T. (1979). Octanoate: an assay for oxidative deterioration in oils and fats. Journal of the Science of Food and Agriculture, 30(9): 876±880. MATHLOUTHI, M.

Â ., SERRANO, M., RIQUELME, F., PEÂREZ-VICENTE, A., MARTIÂNEZ-ROMERO, D., CARBONELL, A Â GUILLEN, F. and VALERO, D. (2002). Role of polyamines in extending shelf life and

the reduction of mechanical damage during plum (Prunus salicina Lindl.) storage. Postharvest Biology and Technology, 25(1): 25±32. PERIS, M. and ESCUDER-GILABERT, L. (2009). A 21st century technique for food control: electronic noses. Analytica Chimica Acta, 638(1): 1±15. RAGAERT, P., DEVLIEGHERE, F. and DEBEVERE, J. (2007). Role of microbiological and physiological spoilage mechanisms during storage of minimally processed vegetables. Postharvest Biology and Technology, 44(3): 185±194. RAHMAN, M.S. (2006). State diagram of foods: its potential use in food processing and product stability. Trends in Food Science & Technology, 17(3): 129±141. RAHMAN, M.S. (2009). Food stability beyond water activity and glass transtion: macromicro region concept in the state diagram. International Journal of Food Properties, 12(4): 726±740. RAHMAN, M.S., AL-WAILI, H., GUIZANI, N. and KASAPIS, S. (2007). Instrumental-sensory


60


evaluation of texture for fish sausage and its storage stability. Fisheries Science, 73: 1166±1176. RAHMAN, M.S., AL-BELUSHI, R.M., GUIZANI, N., AL-SAIDI, G.S. and SOUSSI, B. (2009). Fat oxidation in freeze-dried grouper during storage at different temperatures and moisture contents. Food Chemistry, 114(4): 1257±1264. RAMIREZ-SUAREZ, J.C. and MORRISSEY, M.T. (2006). Effect of high pressure processing (HPP) on shelf life of albacore tuna (Thunnus alalunga) minced muscle. Innovative Food Science & Emerging Technologies, 7(1±2): 19±27. REFSGAARD, H.H.F., BROCKHOFF, P.M., POLL, L., OLSEN, C.E., RASMUSSEN, M. and SKIBSTED, L.H. (1995). Light-induced sensory and chemical changes in aquavit. LebensmittelWissenschaft und -Technologie, 28(4): 425±435. RICO, D., MARTIÂN-DIANA, A.B., BARAT, J.M. and BARRY-RYAN, C. (2007). Extending and measuring the quality of fresh-cut fruit and vegetables: a review. Trends in Food Science & Technology, 18(7): 373±386. RODRIÂGUEZ, A., LOSADA, V., LARRAIÂN, M., QUITRAL, V., VINAGRE, J. and AUBOURG, S. (2007). Development of lipid changes related to quality loss during the frozen storage of farmed coho salmon (Oncorhynchus kisutch). Journal of the American Oil Chemists' Society, 84(8): 727±734. ROOS, Y.H. (2001). Water activity and plasticization. In Eskin, N.A.M. and Robinson, D.S. (eds), Food Shelf Life Stability. New York: CRC Press, pp. 3±36. ROOS, Y.H. (2010). Glass transition temperature and its relevance in food processing. Ann. Rev. Food Sci. Technol., 1: 469±496. ROUSSEAU, D. (2000). Fat crystals and emulsion stability ± a review. Food Research International, 33(1): 3±14. RUDNITSKAYA, A., POLSHIN, E., KIRSANOV, D., LAMMERTYN, J., NICOLAI, B., SAISON, D.,

DELVAUX, F.R., DELVAUX, F. and LEGIN, A. (2009). Instrumental measurement of beer taste attributes using an electronic tongue. Analytica Chimica Acta, 646(1±2): 111± 118. SEWALD, M. and DEVRIES, J. (2003). Shelf life testing. In Medallion Laboratories Analytical Progress. Available at http://www.medallionlabs.com/Downloads/ shelf_life_testing_web.pdf (accessed 4/10/10). SINGH, R.P. and ANDERSON, B.A. (2004). The major types of food spoilage: an overview. In Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Cambridge: Woodhead Publishing. SINGH, T.K. and CADWALLADER, K.R. (2004). Ways of measuring shelf-life and spoilage. In Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Cambridge: Woodhead Publishing. SIRACUSA, V., ROCCULI, P., ROMANI, S. and ROSA, M.D. (2008). Biodegradable polymers for food packaging: a review. Trends in Food Science & Technology, 19(12): 634±643. STONE, H. and SIDEL, J. (2004). Sensory Evaluation Practices, 3rd edn. Orlando, FL: Academic Press. STUDMAN, C.J. (1997). Factors affecting the bruise susceptibility of fruit. In Jeronimidis, O. and Vincent, J.F.V. (eds), Proceedings of Conference on Plant, University of Reading, Reading, pp. 273±281. SUBRAMANIAM, P. (2000). Confectionery products. In Kilcast, D. and Subramaniam, P. (eds), The Stability and Shelf-life of Food. Cambridge: Woodhead Publishing. SUN, D.-W. (ED.) (2005). Emerging Technologies for Food Processing. London: Elsevier. SZCZESNIAK, A.S. (1998). Effect of storage on texture. In Taub, I.A. and Singh, R.P. (eds), Food Storage Stability. Boca Raton, FL: CRC Press, pp. 191±244.



61

and GIANNAKOUROU, M.C. (2004). Temperature and food stability: analysis and contol. In Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Boca Raton, FL: CRC Press. TAOUKIS, P.S. and LABUZA, T.P. (1989). Applicability of time±temperature indicators as shelf life monitors of food products. Journal of Food Science, 54: 783±788. TAOUKIS, P., LABUZA, T.P. and SAGUY, I. (1997). Kinetics of food deterioration and shelf-life prediction. In Valentas, K.J., Rotstein, E. and Singh, R.P. (eds), The Handbook of Food Engineering Practice, Boca Raton, FL: CRC Press, pp. 361±403. VAN BOEKEL, M.A.J.S. (2000). Kinetic modelling in food science: a case study on chlorophyll degradation in olives. J. Sci. Food Agric., 80: 3±9. VAN BOEKEL, M.A.J.S. (2008). Kinetic modeling of food quality: a critical review. Comprehensive Reviews in Food Science and Food Safety, 7(1): 144±158. VAN RUTH, S.M. (2001). Methods for gas chromatography-olfactometry: a review. Biomolecular Engineering, 17(4±5): 121±128. VAN ZEEBROECK, M., VAN LINDEN, V., RAMON, H., DE BAERDEMAEKER, J., NICOLAIÈ, B. M. and TIJSKENS, E. (2007). Impact damage of apples during transport and handling. Postharvest Biology and Technology, 45(2): 157±167. Â NOVAS, G.V. (1998). Rheological properties of VEÂLEZ-RUIZ, J.F. and BARBOSA-CA concentrated milk as a function of concentration, temperature and storage time. Journal of Food Engineering, 35(2): 177±190. VESTERGAARD, J.S., MARTENS, M. and TURKKI, P. (2007). Analysis of sensory quality changes during storage of a modified atmosphere packaged meat product (pizza topping) by an electronic nose system. LWT ± Food Science and Technology, 40(6): 1083±1094. VISIOLI, F., BELLOMO, G., MONTEDORO, G. and GALLI, C. (1995). Low density lipoprotein oxidation is inhibited in vitro by olive oil constituents. Atherosclerosis, 117(1): 25± 32. WANG, C.Y. (1989). Chilling injury of fruits and vegetables. Food Reviews International, 5(2): 209±236. WELLS, J.H. and SINGH, R.P. (1992). The application of time±temperature indicator technology to food quality monitoring and perishable inventory management. In Thorne, S. (ed.), Mathematical Modelling of Food Processing Operations, Amsterdam: Elsevier Applied Science. WELLS, J.H. and SINGH, R.P. (1998). Quality management during storage and distribution. In Taub, I.A. and Singh, R.P. (eds), Food Storage Stability, Boca Raton, FL: CRC Press, pp. 369±386. WINGER, R.J. (2000). Preservation technology and shelf life. In Man, D. and Jones, A. (eds), Shelf Life Evaluation of Foods, 2nd edn. Gaithersburg, MD: Aspen Publishers, pp. 73±86. WUTTKE, W., JARRY, H. and SEIDLOVAÂ-WUTTKE, D. (2007). Isoflavones ± safe food additives or dangerous drugs? Ageing Research Reviews, 6(2): 150±188. XIONG, Y. (2000). Protein oxidation and implications for muscle food quality. In Decker, E., Faustman, C. and Lopez-Bote, C.J. (eds), Antioxidants in Muscle Foods: Nutritional Strategies to Improve Quality. New York: John Wiley & Sons. YAM, K.L., TAKHISTOV, P.T. and MILTZ, J. (2005). Intelligent packaging: concepts and applications. Journal of Food Science, 70(1): R1±10. YAN, S., HUAWEI, C., LIMIN, Z., FAZHENG, R., LUDA, Z. and HENGTAO, Z. (2008). Development and characterization of a new amylase type time±temperature indicator. Food Control, 19(3): 315±319. TAOUKIS, P.S.


62


and HU, Q.H. (2010). Effect of nano-packing on preservation quality of fresh strawberry (Fragaria ananassa Duch. cv Fengxiang) during storage at 4 ëC. Journal of Food Science, 75(3): C236±C240. YANG, T.C.S. (1998). Ambient storage. In Taub, I.A. and Singh, R.P. (eds), Food Storage Stability. Boca Raton, FL: CRC Press, pp. 435±458. YUAN, S. and CHANG, S.K.C. (2007). Texture profile of tofu as affected by instron parameters and sample preparation, and correlations of instron hardness and springiness with sensory scores. Journal of Food Science, 72(2): S136±S145. YANG, F.M., LI, H.M., LI, F., XIN, Z.H., ZHAO, L.Y., ZHENG, Y.H.


3 Moisture loss, gain and migration in foods G. Roudaut, University of Burgundy, France and F. Debeaufort, University of Burgundy, France and IUT-Dijon, France

Abstract: The loss, gain and transfer of moisture often affect food materials. Whether arising from interaction with the atmosphere or with another component of the food, such changes always cause deterioration of the overall quality of the food through softening, toughening, breakdown, swelling or shrinkage due to phase transitions or dissolution. In most cases, water migration leads to organoleptic or microbiological changes in the food. With a view to better understanding the physical deterioration of food and to providing a tool for better control of food quality (and therefore of longer shelf life), this chapter reviews the water relationships in foods with particular attention to, and illustration of, glass transition-related phenomena. It also considers examples of foods affected by moisture exchanges with the atmosphere or within the product itself. The mechanisms controlling these migrations are presented together with some experimental approaches (measurements of moisture content, water activity and migration and modelling). Key words: moisture migration, texture, stability, model.

3.1 Introduction: moisture loss, gain and migration in foods and quality deterioration Who has never experienced a shrivelled piece of fruit or a soggy pastry base while eating a cream cake? Whether dry or moist, all foods may be affected by moisture loss or gain during handling or storage. Moisture transfer, which increases or decreases the water content of a food material may affect the food


64


through softening, toughening, emulsion breakdown and swelling or shrinkage due to phase transitions such as glass transition, crystallisation or dissolution (Petersen et al., 1999). In most cases, water migration leads to a deterioration of the overall organoleptic or hygienic quality of the food. Moisture transfers compromise the quality, stability and safety of the product and limit its shelf life. Chemical deterioration (e.g. oxidative reactions, hydrolyses) is caused mainly by moisture gain which leads to greater reaction rates due to increased water activity. These aspects are reviewed elsewhere in this book. With a view to better understanding the physical deterioration of food and for the provision of a tool for better control of food quality (and longer shelf life), this chapter will review the water relationships in foods with particular attention to, and illustration of, glass transition-related phenomena. It also considers examples of foods affected by moisture exchange with the atmosphere or within the product itself. The mechanisms controlling these migrations will be presented together with some experimental approaches (measurements of moisture content, water activity and migration and modelling).

3.2

Mechanism of the moisture transfers in food products

3.2.1 General considerations on moisture transfers in foods In all heterogeneous products, water transfer occurs from `wet' to `dry' areas. The water may be in the form of either liquid or vapour, or may form a `reservoir' when in the solid state, as in frozen products. For example, in an ice cream cone, moisture migrates in the liquid and vapour states. To control such transfers, a coating, usually chocolate or fat similar to cocoa, is applied. This controls both the migration of liquid water from the cryo-concentrated solution of the ice cream in contact with the wafer, and the exchange of water vapour inside and outside the cone. It also ensures an airtight seal with the packaging paper or aluminium-paper-based complex. Whatever the physical state of the migrating water, the temperature or the nature of the products, the humidity of the two areas tends to reach an equilibrium. However, it should be noted that the balance is established between the water chemical potentials, often expressed by their water activity, and not between the water contents. Indeed, as shown by Karathanos et al. (1995) when considering the interface between raisins and a brioche dough, the two areas do not balance at the same moisture content and the initially dry area reaches a water content higher than that of the initially wet area (Fig. 3.1). However, when at equilibrium, both sides of a heterogeneous product stabilise at the same water activity of 0.82 when the two areas were initially at 0.95 and 0.4 respectively for dough cereal and raisins. This can be described and predicted from the sorption isotherms and diffusion coefficients of the two areas. By using the sorption isotherms of the two products and knowing their relative proportions in the mixture and the initial water activity, it may be foreseen that the raisins will rehydrate when the dough dehydrates. The adverse effects on the product are numerous: development of mould spores which are naturally present


Moisture loss, gain and migration in foods

65

Fig. 3.1 Moisture transfer from dough to dry raisins after 3 days contact. Initial water contents and water activities were 0.96 and 55% for the dough, and 0.40 and 20% for the raisins, respectively (adapted from Karathanos et al., 1995).

on raisins, the swelling and bursting of the raisins, drying of the dough and problems with rising during the baking process. This example highlights the fact that the driving force for water transfer is not the differential in concentration or volume, but the differential of the chemical potential of water, generally expressed as the water activity, the mole fraction, or the partial pressure for the gaseous phases. This mechanism of water transfer is described by laws which differ according to the structure of the environment. When considering water migration in a porous open system, i.e. when channels or pores connect both sides of the product, the flow of water is described by Knudsen diffusion, Hage-Poiseuille's law and by Darcy's law as the hydration level increases as shown in Fig. 3.2. Darcy's law equation directly links the water flow rate (F) to the pressure on both sides of the porous film (p), while taking into account the viscosity (), the density () and the specific environment resistance (R) which depends on the diameter of pores and their tortuosity: p F R However, Darcy's law only applies to fluid flowing through the pores of the film, i.e. for very high relative humidity. Indeed, several mechanisms are involved as the moisture level of the medium increases. Thus, as shown in Fig. 3.2, with increasing relative humidity we successively observe: first, a gaseous diffusion of water molecules associated with monomolecular adsorption on the surface of the pores; secondly, moisture disseminated to both the gas in the pores and the surface of the water film adsorbed on the surface of the pores and, finally, a flow as a condensed phase occurs (QueÂnard and SalleÂe, 1991). Although many food products are considered to be porous (cereal products, extruded material or foams such as ice cream or mousse and some vegetables like zucchini or leeks), it appears that this very simple model does not apply


66


Fig. 3.2 Moisture transfer mechanism in porous media as a function of the relative humidity or water activity level (adapted from QueÂnard and SalleÂe, 1991).

because the pores never cross the structure of the product throughout its thickness. Most food products have an alveolar or cellular structure instead of truly porous systems. The model which better fits these products is the `simple' diffusion model which usually applies for dense, homogeneous and isotropic products described by Fick's laws (Crank, 1975) and shown in Fig. 3.3. There are two basic types of diffusion. The first is random and uncoordinated molecular motion, often referred to as Brownian motion or self-diffusion. The self-diffusion coefficient of water at the molecular level is commonly measured using pulsed field gradient nuclear magnetic resonance (NMR) spectroscopy and

Fig. 3.3 Diffusion through homogeneous medium: concentration profiles in transient and steady flow rates from Fick's laws, with x distance and C concentration.



67

magnetic resonance imaging (MRI). The second type of diffusion, which is the focus of this chapter, is directional and reveals a concerted molecular motion which is attributed to the presence of a driving force (e.g., differences in chemical potential), often referred to as flow or bulk diffusion. Determination of the bulk moisture diffusion coefficient (D) of water in food materials is commonly obtained from Fick's laws applied to kinetic data from the three processes of drying, sorption and permeation. Fick's first law applies to a permanent flow rate or steady state (no accumulation, i.e., no change of concentration with time) with unidirectional diffusion and the absence of any chemical reaction. @C @C @C J ÿD grad C ÿD @x @y @z Fick's first law is often simplified as unidirectional to describe the moisture flow (J) or transfer rate (TR) in food or through packaging films as follow: dm dC ÿD k dC J TR A dt dx where J is the moisture diffusion flow or current (kg sÿ1 mÿ2); m is the water mass (kg); t is the time (s); A is the surface (m2); D is the water diffusion coefficient (m2 sÿ1); C is the moisture concentration (kg mÿ3); x is the distance (m) and k is the moisture transfer coefficient (m sÿ1) with k D=dx. For transient flow rates (variation of mass flux with time), a generalisation of Fick's second law, which is a derivative of the first law, applies: 2 @J @C @ C @2C @2C D ÿ @x; y; z @t @x2 @y2 @z2 where D is the moisture diffusion coefficient or diffusivity (m2 sÿ1); C is the moisture concentration (kg mÿ3) as a function of x, y, z and t; x, y and z are the spatial coordinates (m). Solutions to this differential equation can be found for various boundary conditions (Crank, 1975). In all techniques for determining diffusion and mass transfer, water content, water activity and sorption, isotherm measurements are required.

3.2.2 Sorption isotherm characterisation of foods A moisture sorption isotherm maps the complex, product-specific relationship between water content and water activity. Considered as the `fingerprint' of a food product, this isotherm curve (usually S-shaped), shows how water content changes as water activity increases or decreases. The sorption isotherm is the key to understanding and controlling product formulation, product stability, moisture sensitivity, temperature effects and drying characteristics (Simatos, 2002). The relationship between total moisture content and the water activity of a food, considered over a range of values and at a constant temperature, yields a


68


Fig. 3.4

Typical food adsorption and desorption isotherms.

moisture sorption isotherm when expressed graphically. This isotherm curve may be obtained in one of two ways (Fig. 3.4): 1. An adsorption isotherm is obtained by placing a completely dry material into differing atmospheres of increasing relative humidity and measuring the weight gain due to water uptake. 2. A desorption isotherm is found by placing initially wet material under the same relative humidity and measuring the weight loss. The adsorption and desorption processes are not fully reversible; therefore, a distinction can be made between the adsorption and desorption isotherms by determining whether the moisture levels within the product are increasing, thus indicating wetting, or whether the moisture is gradually lowering to reach equilibrium with its surroundings, implying that the product is being dried (Simatos, 2002). Sorption isotherms are usually divided into three zones corresponding to different ways in which moisture fixation takes place on the solid substrate as shown in Fig. 3.4: · Zone 1: monomolecular monolayer of adsorbed water on the product surface. This corresponds to the van der Waals interactions between the hydrophilic parts of the product and the water molecules. Water adsorption occurs progressively until it creates a continuous monolayer of water molecules on both the external surface of the product and on the surface of its infractuosities. Water is considered as strongly fixed and in a `semi-rigid' state because of the interactions between water molecules and the surface. The second zone starts when the whole surface is saturated (Van den Berg, 1991). · Zone 2: pluri-adsorption on the initial monolayer. This represents water molecules which are less firmly bound, initially as multi-layers above the



69

monolayer. In this zone, water is held in the solid matrix by capillary condensation. This water may act as a solvent for low molecular weight solutes and for some biochemical reactions. The quantity of water that does not freeze at normal freezing temperatures is often associated with this zone. · Zone 3: water in liquid state. Excess water is present in macro-capillaries or as part of the liquid phase in high moisture materials. This exhibits nearly all the properties of bulk water, and thus is able to act as a solvent (Al Muhtaseb et al., 2002). On the basis of the van der Waals adsorption of gases and vapours on various solid substrates, Brunauer et al. (1940), the International Union of Pure and Applied Chemistry (www.iupac.org) classified adsorption isotherms into five general types (Fig. 3.5). The classification of isotherms corresponds to different types of interactions between adsorbed substances and the adsorbent or its porosity. Type I is the Langmuir, and Type II the sigmoid shaped adsorption isotherm. However, no specific names have been attached to the other three types. Types II and III are closely related to types IV and V, with the exception of the maximum adsorption occurring at a pressure lower than the vapour pressure of the gas. However, if the solid is porous and has a significant internal surface, then the thickness of the adsorbed layer on the walls of the pores is limited by the width (diameter and length) of the pores. The form of the isotherm is modified correspondingly and instead of type II and III, is classified as types IV and V. The moisture sorption isotherms of most foods are non-linear, generally sigmoidal in shape, and are classified as type II isotherms. Some mathematical models have been developed for sorption isotherm description as given in Table 3.1. These are often used in models developed for the prediction of mass transfer or diffusion equations.

Fig. 3.5

The IUPAC classification of isotherms.


70


Table 3.1

Mathematical expression for sorption isotherm description

Model

Mathematical expression

Langmuir

X X0

BET Halsey

C aw 1 C aw

X0 C aw 1 ÿ aw 1 ÿ C ln1 ÿ aw 1=B ÿA X T lnaw X

X A B log10 1 ÿ aw ÿln1 ÿ aw 1=B Henderson X A aw B Oswin X A 1 ÿ aw 1=

Ferro-Fontan X ln=aw Smith

GAB

X

Reference Brunauer et al. (1940) Aguerre et al. (1989) Halsey (1948) Smith (1947) Henderson (1952) Oswin (1946) Ferro-Fontan et al. (1982)

X0 C K aw Guggenheim (1966) 1 ÿ K aw 1 ÿ K aw C K aw

with C c0 eHe =RT and C k0 eHk =RT Peleg

X K1 anw1 K2 anw2

Peleg (1993)

A is a constant (dimensionless), aw, the water activity, B, a constant (dimensionless), C, the GAB model parameter (dimensionless), c0, the constant adjusted to the temperature effect (dimensionless), Hc, the difference in enthalpy between mono-layer and multi-layer sorption (J molÿ1), Hk, difference between the heat of condensation of water and the heat of sorption of the multi-layer (J molÿ1), K, GAB model parameter (dimensionless), k0, the constant adjusted to the temperature effect) (dimensionless), n1 and n2, the equation parameters (dimensionless), R, the universal gas constant (8.314 J molÿ1 Kÿ1), r, the equation parameter (dimensionless), T, the temperature (K), X, the equilibrium moisture content (kg kgÿ1 dry solid), X0, the mono-layer moisture content (kg kgÿ1 dry solid), , the equation parameter (dimensionless), , the equation parameter (dimensionless).

3.3 Measuring, monitoring and predicting moisture loss, gain and migrations 3.3.1 Water content measurement The water content of foodstuffs may be determined using either direct or indirect methods. Direct determinations may be based on some physical separation techniques such as distillation, drying or chemical reactions (Mathlouthi, 2001). Indirect determinations rely on the spectroscopic properties of water molecules. This is the case for NMR, infrared and Raman spectroscopy, which are nondestructive techniques, as well as microwave spectroscopy or microwave resonator methods.



71

Physical techniques for water content measurement Oven drying over a fixed period (from 1 to 18 h) at a standard temperature (from 102 to 105 ëC) is very often the legal-standard method for the determination of water content. The period of drying is specified for each type of product. The main sources of error or inaccuracy are usually the incomplete removal of water, the loss of volatiles other than water during the drying period, the formation of a crust at the surface of the product which slows down the evaporation of water, the decomposition of the product and the Maillard reaction, which also produces water. Vacuum-oven drying takes place at lower temperatures (70 ëC) for longer periods. This technique is particularly useful in preventing the destruction of heat-sensitive samples. However, the duration of the drying period may not be sufficient to allow the food to reach a steady state as this is strongly dependent on the size of particles. Solvent extraction may be used to extract water from food using an organic solvent having a strong affinity for water prior to its analysis by chemical titration. Chemical analysis of water content Karl Fischer titration is a classical titration method in analytical chemistry which uses coulometric or volumetric titration to determine trace amounts of water in a sample. It was invented in 1935 by the German chemist Karl Fischer. The main compartment of the titration cell contains the anode solution plus the analyte. The anode solution consists of an alcohol (ROH), a base (B), SO2 and I2. Typical alcohols for this purpose are methanol or diethylene glycol monomethyl ether, and a common base is imidazole. The titration cell also consists of a smaller compartment with an anode immersed in the anode solution of the main compartment. The two compartments are separated by an ionpermeable membrane. The Pt anode generates I2 when current is provided through the electric circuit. The net reaction as shown below, is the oxidation of SO2 by I2. One mole of I2 is consumed for each mole of H2O. In other words, two moles of electrons are consumed per mole of water as described in the following equations: BI2 + BSO2 +B + H2O ÿ! 2BH+Iÿ + B+SOÿ3 B+SOÿ3 + ROH ÿ! BH+ROSOÿ3 The end point is detected most commonly by a bipotentiometric method. A second pair of Pt electrodes is immersed in the anode solution. The detector circuit maintains a constant current between the two detector electrodes during titration. Before the equivalence point, the solution contains Iÿ but little I2. At the equivalence point, excess I2 appears and an abrupt voltage drop marks the end point. The amount of current needed to generate I2 in order to reach the end point can then be used to calculate the amount of water in the original sample.


72


Volumetric titration is based on the same principles as coulometric titration, except for the anode solution being used as the titrant solution. The titrant consists of an alcohol (R-OH), base (B), SO2 and a known concentration of I2. One mole of I2 is consumed for each mole of H2O. The titration reaction proceeds as above. The end point may be detected by a bipotentiometric method as described above. Quantitative chemical reactions which produce gas may be used to quantify water content, provided that the released gas is accurately analysed. Among these reactions, we find: H2O + CaH2 ÿ! CaO + 2H2 H2O + CaC2 ÿ! CaO + C2H2 The volume of C2H2, which is directly linked to water content, may change as a function of temperature in the analysed medium and is easily determined (Mathlouthi, 2001). Thermal analysis Thermal analysis using either differential thermal analysis (DTA) or differential scanning calorimetry (DSC) may be used in the heating of a frozen sample to determine its freezable water content, which is approximately that fraction of water considered as being mobile or `free'. These techniques also provide information on the physical state of water in foodstuffs, which may be helpful in interpreting the behaviour of the product, for instance, during drying (Mathlouthi, 2001). Gas chromatography Gas chromatography has been applied to the determination of water content in freeze-dried or dried products. However, water has to be extracted by organic solvents before analysis and the sample must be homogeneous. The extraction solvent should have a high affinity for water and be protected from the surrounding atmospheric humidity. Using methanol or dimethylformamide together has proved to be an efficient method of extraction. Generally, Porapaq-Q or Carbowax is used as a stationary phase in the column and a thermal conductivity detector (TCD) or a mass spectrometer is required. This method is relatively quick, but limited to a sensitivity of approximately 10 ppm and inaccuracies are usually due to poor peak separation or water traces in the solvent when the TCD is used. GC-MS now permits a far greater sensitivity, but it remains a technically complex and expensive method. Spectroscopic techniques For all spectroscopic techniques, calibration to standards or products of known water content is required. NMR spectroscopy is informative on hydrogen atoms which are more easily detected in a liquid environment. This technique is better adapted to distin-



73

guishing water of different mobilities than to the accurate determination of water content. Moreover, it is necessary to obtain a precise calibration which is specific to the analysed product and based on a good reference method. Timedomain nuclear magnetic resonance (TD-NMR) has proved a quick, reproducible, accurate and non-invasive technique which is particularly well suited to measuring moisture content. Nuclear spin±spin relaxation times (T2) are an excellent probe of molecular mobility, which in turn can be directly correlated to moisture content (Hickey et al., 2006). Near infrared (NIR) absorption of water occurs at different wavelengths (1950, 1450 and 977 nm). The ratios of the intensities of the bands at 1950 and 1450 nm are used to measure water content (Vornhof and Thomas, 1970). Automatic or on-line NIR spectrometers are used in different food industries for the determination of water content as well as for other food constituents such as proteins, fats, minerals, caffeine and sugars. This method requires a specific calibration for the analysed food. Several parameters affect the outcome of the results (colour, particle size, thickness and texture). The reflectance technique allows for the detection of surface water and may not be representative of the whole if the product is not homogeneous. Microwave spectroscopy uses the dipolar character of water molecules. Water content is measured by the shift in wavelength and the attenuation of the amplitude of the waves when a sample is placed between the microwave emitter and receiver. Parameters such as water concentration, density and thickness of the analysed sample, may have an effect on the result. Only mobile water can be measured. The method may be used for on-line measurements when the thickness and permittivity of the sample are known (Isengard, 1995). 3.3.2 Water activity determination Water activity or aw is a measurement of the energy status of water in a system. It is defined as the vapour pressure of water above a sample divided by that of pure water at the same temperature. Therefore, pure distilled water has a water activity of exactly one. Water activity is defined as: p RH aw p0 T 100 T where p is the vapour pressure of water in the food product, and p0 is the vapour pressure of pure water at the same temperature, and RH is the relative humidity above the food sample when equilibrium between product and atmosphere is reached. If there is no difference in the interaction between water and water, and between water and solute, the determination of water activity is easy and its expression becomes: nw Xw aw ntotal


74


which is directly obtained from the mole fraction Xw of water molecules (nw) to total molecules in the solutions (ntotal). For real solutions, aw Xw , where is an activity coefficient. The higher the modification of water binding by the solute, the greater the difference between the coefficient and 1. There are several factors which control water activity in a system: · The colligative effects of dissolved substances (e.g. salt or sugar), which interact with water through dipole-dipole, ionic, and hydrogen bonds. · The capillary effect, where the vapour pressure of water above a curved liquid meniscus is less than that of pure water because of changes in the hydrogen bonding between water molecules. · Surface interactions, in which water interacts directly with chemical groups in undissolved ingredients (e.g. starches and proteins) through dipole-dipole forces, ionic bonds (H3O+ or OHÿ), van der Waals forces (hydrophobic bonds), and hydrogen bonds. It is the combination of these three factors in a food product which reduces the energy of the water and thus reduces the relative humidity when compared to pure water. These factors may be grouped under two broad categories with osmotic and matrix effects. Due to varying degrees of osmotic and matrix interactions, water activity is the term which describes the continuum of energy states of the water in a system. The water appears `bound' by forces in varying degrees. This is a continuum of energy states rather than a static `boundness'. Water activity in a system is sometimes defined as `free', `bound', or àvailable water'. Although these terms are easier to conceptualise, they fail to adequately define all aspects of the concept of water activity (Van den Berg and Bruin, 1981). Water activity is temperature dependent. Temperature changes the partial water vapour pressure, and then the water activity because of modifications in water binding, the dissociation of water, the solubility of solutes in water, or the state of the matrix. Although the solubility of solutes can be a controlling factor, the state of the matrix is usually more significant. As the state of the matrix (glassy vs. rubbery state) depends on temperature, the temperature affects the water activity of the food. The effect of temperature on the water activity of a food is product specific. While most high moisture foods exhibit negligible aw change with temperature, for some products, water activity increases with increasing temperature and in others, it will decrease with increasing temperature. Therefore the direction of the change in water activity with temperature cannot be predicted, since it depends upon the way in which temperature affects the factors controlling water activity in the food under consideration. As a potential energy measurement, water activity is a driving force for water movement from regions of high to low water activity. For example, if honey (aw 0:6) is exposed to humid air (aw 0:7) it will absorb water from the air. Other examples of this dynamic property of water activity are, as previously described, moisture migration in multi-domain foods (e.g. cracker-cheese sandwiches), the movement of water from soil to the leaves of plants, and cell



75

turgor pressure. Since microbial cells have a high concentration of solute surrounded by semi-permeable membranes, the osmotic effect on the free energy of the water is important in determining microbial water relations and therefore growth rates (Fennema, 1985). Most of the methods used for determining the water activity or equilibrium relative humidity (ERH) of foods were originally devised by meteorologists for the measurement of relative humidity in atmospheric air. When water activity (aw) is measured, it is generally necessary to know whether or not the product has reached the critical zone where spoilage reactions may occur. For this reason, accuracy within 0.01 aw is sufficient for most food-related applications (Mathlouthi, 2001). Water activity values are obtained by a capacitance hygrometer, a dew point hygrometer or a mano-vacuometer. Capacitance or resistive hygrometer Capacitance hygrometers consist of two charged plates separated by a dielectric polymer membrane. As the membrane adsorbs water, its ability to hold a charge increases and the capacitance is measured. This value is roughly proportional to the water activity as determined by a sensor-specific calibration. Most volatile chemicals do not affect capacitance hygrometers which can be much smaller than other sensors. They do not require cleaning, but are less accurate ( 0.015 aw) than dew point hygrometers (< 0.01 aw). They require regular calibration and can be affected by residual water in the polymer membrane. Their accuracy is often limited to the range of 0.15±0.97. Novasina has specialised in the accurate measuring of air and material humidity. Their LabMaster AW is very accurate (Fig. 3.6). Due to the newly developed electrolytic measurement cell with saved calibration data as well as a temperature controlled measuring chamber, the instrument determines high precision and repeatable aw values. The integrated pre-conditioning chamber improves the measurement speed and increases the efficiency of the measurement process (www.novasina.com). Dew point hygrometer The temperature at which dew forms on a clean surface is directly related to the vapour pressure of the air. Dew point hygrometers work by placing a mirror over

Fig. 3.6

Principle of capacitive sensor for relative humidity measurement used in the LabMaster AW meter from Novasina (www.novasina.com).


76


Fig. 3.7 Chilled mirror dew point principle used in the Aqualab (Decagon, www.decagon.com) and in the FA-st lab (GBX, www.GBXinstru.com) aw meters.

a closed sample chamber. The mirror is cooled until the dew point temperature is measured by means of an optical sensor. This temperature is then used to find the relative humidity of the chamber, using psychrometric charts. This method is the most accurate ( 0.003 aw) and often the fastest. The sensor will require cleaning if debris accumulates on the mirror. The AquaLab Series from Decagon is a lab-grade water activity meter (www.decagon.com). The chilled-mirror dewpoint sensor offers unparalleled accuracy ( 0.003 aw) and an extremely rapid reading time (less than five minutes, including equilibration time). Equivalent apparatus (Fast-Lab) were also developed by GBX, offering similar performance (www.gbxinstru.com) as displayed in Fig. 3.7. Manometric hygrometer As water vapour pressure is given in tables for different temperatures, a direct measurement of water vapour pressure in the food should give the best direct tool for the determination of aw. To achieve this measurement, it is necessary to establish a vacuum and to work at very low temperatures. Working at zero pressure on one side with the moisture freeze trap, and leaving the sample on the other side to release its vapour, permits an accurate measurement of aw. This method requires accurate temperature measurement and the device is extremely fragile (Troller and Christian, 1978).



77

Spectroscopy determination of relative humidity The determination of relative humidity in equilibrium above a food product directly gives its water activity. Absorption spectroscopy is a relatively simple method of passing light through a gas sample and measuring the amount of light absorbed at the specific wavelength. Traditional spectroscopic techniques have not been successful at doing this in natural gas because methane absorbs light in the same wavelength regions as water. But by using a very high resolution spectrometer, it is possible to find water peaks that are not overlapped by gas peaks. The Tunable Diode Laser Absorption Spectroscopy (TDLAS) analyser is the only instrument that can meet all the following: the necessity for an analyser that will not suffer from interference or damage from corrosive gases, liquids or solids, that will react very quickly to drastic moisture changes and will remain calibrated for very long periods of time. SpectraSensors, Inc., is a manufacturer of optical-based gas sensors for the industrial processing, environmental monitoring and clean technology markets. The company's sensors measure the absorption of laser light at specific wavelengths to detect carbon dioxide and water vapour (RH or aw) in industrial process control and environmental monitoring applications (www.spectrasensors.com). 3.3.3 Isotherm determination Sorption isotherm determination requires either the water content or the water activity of the product to be measured. Discontinuous methods are most often used in factories and research laboratories. However, automatic systems have been developed and commercialised since the end of the 1990s. Conventional methods Micro-climates gravimetric method The gravimetric method involves the measurement of weight changes. Weight changes can be determined both continuously and discontinuously in dynamic or static systems (i.e., air may be circulated or stagnant). In the discontinuous systems, salt or sulphuric acid solutions are placed in vacuum or atmospheric systems with the food material, to give a measure of the equilibrium relative humidity. A static gravimetric technique was developed and standardised in the Water Activity Group of the European COST 90 project, using saturated salt solutions to control the relative humidity of the atmospheres in contact with the food samples (Jowitt and Wagstaffe, 1989). Continuous methods use electrobalances or quartz spring balances and are the basic principles behind commercial automatic devices. Manometric method The manometric method measures the vapour pressure of water in the vapour space surrounding the food. To improve accuracy, the fluid used in the Umanometer is often oil instead of mercury. The whole system is maintained at a constant temperature and the food sample will lose water to equilibrate with the


78


vapour space. This will be indicated by the difference in height on the manometer. Hygrometric method The hygrometric method measures the equilibrium (balanced) relative humidity of air in contact with a food material at a given moisture content, in contrast to the previous methods which measure the water content of sampled equilibrated to fixed RH atmospheres. Dew-point hygrometers or capacitance (or resistive) hygrometers are often used. Diaphragm pressure drop method set up for very high relative humidity range isotherms A fast and accurate method has been set up to measure sorption isotherms of solid foods in the water activity range 0.9±1 by Beaucour and Daudin (2000). This method avoids the main problem generated by the micro-climate technique which is that it is unsuited to high humidity ranges owing to (i) the lack of ventilation which results in excessive equilibration times, and (ii) its inability to produce and control high relative humidities. In a purpose-built device, thin slices of solid material are subjected to a high velocity air flow (more than 10 m sÿ1), at temperatures ranging between 4 and 40 ëC. The relative humidity of the air is controlled by a reduction in total pressure of the saturated air by use of diaphragms. The humid air is obtained from saturated air which is expanded at a constant temperature. The relative humidity of the air may thus be controlled by precisely measuring the total pressure. For constant temperature and water molar fraction, RH1 Ptot1 , the relative humidity and total pressure vary in the same way: RH2 Ptot2 where Ptot is the total pressure and RH the relative humidity. Air relative humidity is very sensitive to pressure change, whereas the water activity of a food product does not vary significantly and this can be assessed as follows: aw2 Mw Ptot2 ÿ Ptot1 ln aw1 w R T with Mw is the molecular weight of water (kg kmolÿ1), w is the density of water (kg mÿ3), T is the temperature (K) and R is the gas constant (8.134 J Kÿ1 molÿ1). The apparatus has two parts (Fig. 3.8): (i) a saturation column, which produces saturated air, and (ii) a sequence of 10 cells where the samples are placed. By making successive pressure drops from cell to cell, a specific range of air relative humidity can be covered from 0.9 to 1. Automatic commercialised apparatus The Autosorp from Biosystemes is an automatic device for the determination of water sorption isotherms of chemicals, food or biological products based on the gravimetric and micro-climate technique as standardised by the COST-90, which has been on the market since 1985 (www.biosystemes.com). The relative



79

Fig. 3.8 General diagram of the experimental device and design of the 10 cells, sample holder and a cross-view of a calibrated diaphragm (from Beaucour and Daudin, 2000).

humidity in the cabinet where the samples are stored is fixed by mixing a wet saturated stream and a dry stream of air, in a range from 2 to 98%. The RH is controlled using a capacitance hygrometer. The temperature ranged from 10 to 45 ëC. Up to 34 samples can be defined simultaneously, with weights ranging


80


from 1 to 220 g with a sensitivity of 0.1 mg. Samples are automatically and periodically weighed up to a constant mass which corresponds to the sorption equilibrium. The software can determine the equilibrium time, then automatically changes the RH for the next measurement, either in sorption or desorption determination for a period ranging from a few hours up to ten weeks. An example of results is given in Fig. 3.9. The introduction of the Dynamic Vapour Sorption (DVS) apparatus by Surface Measurement Systems (www.thesorptionsolution.com or www.smsuk.co.uk) in 1994, revolutionised the field of gravimetric moisture sorption measurement, replacing outdated time and labour intensive desiccators with cutting-edge instrumentation and overnight vapour sorption isotherms (Fig. 3.10). It is the main commercialised apparatus, cited and compared to the standardised methods in the scientific literature. Indeed, Yu et al. (2008) compared the performance of the DVS with the saturated salt solution method and concluded that the DVS provides accurate results and allows good estimations of diffusivity (Table 3.2).

Fig. 3.9 Example of sorption and desorption kinetics measured simultaneously on a model food sample (polysaccharide based) and the resulting sorption isotherm.



Fig. 3.10

81

Schematic design of the DVS apparatus (from www.thesorptionsolution.com).

Table 3.2 Comparison between the saturated salt solution method and the DVS instrument for collecting data for determining the bulk moisture diffusion coefficient (from Yu et al., 2008) Experimental feature Saturated salt solution method DVS instrument Average sample size 1±2 g Experiment time Weeks to months Experiment design Less flexible; discrete relative humidity values dependent on salts selected and experimental temperature Air flow Data collection

Labor Cost

Static to slow movement with use of a fan or stir bar inside the chamber Wt measurements disturb the environmental relative humidity; fewer, discrete data points

5±100 mg Days More flexible relative humidity and temperature control; can obtain absorption and desorption data on the same sample in a short time Dynamic; air continuously flows past the sample

Wt measurements do not disturb environmental relative humidity; numerous, nearly continuous data points (e.g., 30±60 s intervals) More labor intensive (periodic Less labor intensive (instrument weighing of samples for weights sample) weeks to months) Low High


82


DVS combines high quality micro-balance, gas flow and vapour measurement technologies to deliver excellent performance in terms of experimental design as well as instrument accuracy and repeatability. The Advantage set up uses a dry carrier gas, usually nitrogen, and the user can select one of any two vapour sources. Precise control of the ratio of saturated and dry carrier gas flow is enabled by mass flow control, combined with the use of unique real-time vapour concentration monitoring for both water and organics. A known concentration of the selected vapour then flows over a sample suspended from a recording ultramicro-balance, which measures the weight change caused by sorption or desorption of the vapour molecule. It is these dynamic flow conditions that enable the sorption/desorption process to be so rapidly studied. DVS allows sample weights from 1 to 150 mg, the weight variation sensitivity is 0.1 g, temperatures range from 5 to 50 ëC, RH can vary from 0 up to 98% with an accuracy of 0.5% and for the organic vapour from 0 to 96% of the saturation with an accuracy of 0.7%. The IGA series from Hidden Isochema is a fully automated benchtop gravimetric analysis system designed to quickly and accurately measure the magnitude and kinetics of moisture sorption by materials (www.hidenisochema.com). The IGA200 Multicomponent Gas/Vapour Analyser is a fully integrated gravimetric sorption analyser designed for dynamic multi-component gas and vapour sorption analysis. In addition to the multiple inlet mass flow controller system for multicomponent gas sorption experiments, the IGA-200 also features a unique multistream vapour generation module, allowing the integration of both gas and vapour inlet streams in pressure control and flow control modes. Anti-condensation protection up to 50 ëC is incorporated in the IGA-200, which allows operation with water and an extensive range of hydrocarbon vapours. The IGA-200 comes complete with a degas heater reactor with an integral refrigerated recirculating water bath for fully automated sample temperature control from ÿ15 to 500 ëC. The combination of the multi-component gas and vapour sorption capability with closecoupled quadruple mass spectrometry makes the IGA-200 the most complete and sensitive sorption analysis system. Using the Dynamic Dewpoint Isotherm method (DDI), AquaSorp from Decagon (www.decagon.com) produces highly accurate adsorption and desorption isotherm curves in 24 hours or less and is affordable for researchers and formulators. In contrast, an isotherm generator will cost around $100,000 and take between two and five weeks to create a single isotherm by hand. The relative humidity ranges from 10% to 90% with an accuracy of 0.5%, and the temperature can vary from 15 to 40 ëC. Very little information is available on this new apparatus which was presented in late 2007. 3.3.4 Measurement of water migration Concentration profile The concentration profile method uses a tube in a standardised design where two cylinders of solid or semi-solid products are brought into contact. These may



83

have a different initial water content or activity and may, or may not, be separated by a barrier layer. The compartments can then be thinly sliced at timed intervals and either or both the water content and water activity of these slices be measured (Voilley and Bettenfeld, 1985) by the methods previously described. The concentration profiles are measured over time and distance from the interface in the multi-compartment system and the water transfer is defined through a concentration±distance curve. Nuclear magnetic resonance (NMR) imaging In contrast to the gravimetric methods previously described, NMR allows nondestructive measurements of water migration to be performed. The principle of NMR imaging is to offer spatial information on spins (Ruan and Chen, 1998), thus providing a spatial measurement of NMR parameters (such as T2, for example). When these parameters are sensitive to the moisture content of the sample, the spatial distribution of water in a sample can be defined. The recording of NMR imaging signals at timed intervals enables measurement of changes in moisture distribution within a sample (Ruan and Chen, 1998, Lodi et al., 2007), especially in intermediate and high moisture systems. Stray field nuclear magnetic resonance imaging (STRAFI-NMRI) The application of NMR imaging is generally limited to intermediate and high moisture systems by technical constraints (the necessity of rapidly switching high field gradients) in low moisture systems. However, the technique has been modified to expand its field of application to systems having a high solute concentration (and thus very short T2) (Hopkinson et al., 2001). As explained in detail in Hopkinson et al. (2001), the water profile of the sample can then be obtained either by moving the sample in the field gradient to excite successive slices, or by changing the frequency of the excitation pulse. The water content of the sample is measured as a function of position, by probing the sample with a pulse sequence of varying frequency. This displays the spatial sensitivity of the signal and thus the spatial distribution of the water. 3.3.5 Predicting moisture transport phenomena in food products From foodstuff to surrounding atmosphere and within foodstuff (from a moist area towards an area of lower hydration) Studies have recently been published describing the modelling of moisture transfers by sorption (wetting) and desorption (drying) behaviours, inside compartmented hetereogeneous food products (Guillard et al., 2003; Roca et al., 2008; Bourlieu et al., 2008). These models were established from either sorption experiments and/or the concentration profile methods, assuming the food products to be layers or sheets. Depending on the assumptions made for the description of product deformation and external mass transfer phenomena, four alternative models were developed in order to predict moisture migration inside products, and between


84


products and environment, thus determining the effective moisture diffusivity (Roca et al., 2008; Karbowiak et al., 2006a, 2008). Considering the description of the influence of water uptake on the state of the product deformation, the product volume was assumed to remain constant during water sorption (no deformation hypothesis). Secondly, it was assumed that the solid structure of the product would swell according to the local value of moisture content: in this case, the volume occupied by both liquid and solid phases was assumed to be equal to the sum of partial volume occupied by each phase. Equations describing mass transport phenomena were therefore written in a referential (; t) moving at the velocity of the solid phase. The Lagrangian coordinate is related to the dry matter in the main direction of mass transport (Boudhrioua et al., 2003). Dealing with the description of external moisture transfer (between the product surface and the surrounding environment), the external resistance to mass transfer was assumed to be negligible. The product surface moisture content was then assumed to be constant and equal to the equilibrium moisture content for the relative humidity of the air (no external resistance hypothesis). The external resistance to mass transfer was taken into account using an external mass transfer coefficient (kmass, in m/s), expressing the resistance to diffusion through the mass transfer boundary layer for the vapour emitted by the product surface (external resistance hypothesis). Combining these different assumptions, four models (Table 3.3) were developed for describing mass transport during moisture sorption or desorption experiments in a sample assumed to consist of sheets or thick layers: · · · ·

Model Model Model Model

1: 2: 3: 4:

no deformation and no external resistance to mass transfer. deformation but no external resistance to mass transfer. external resistance to mass transfer but no deformation. deformation and external resistance to mass transfer.

For all models, the system is assumed to be composed of a continuous aqueous phase moving through a solid phase (so-called dry matter). A Fickian theory with a constant moisture apparent diffusivity is used. Whatever the hypotheses chosen to describe product deformation and external mass transfer, the mass conservation equation in the system may be written as the generalised Fick equation. Whatever the time t, @X @ @X Deff for 0 x xmax @t @x @x and Deff

@X 0 @x

for x 0

at the interface between the food and its surrounding atmosphere (or between two phases of the composite food), where x and Deff are respectively the modified space coordinate (m), and the modified apparent diffusivity (m2/s),



85

Table 3.3 Assumptions and boundary conditions of models for effective moisture diffusivity prediction (from Roca et al. 2008) Model 1

Model 2

Model 3

Model 4

Assumptions No deformation No external resistance Eulerian coordinate x Deff Deff

Simple deformation No external resistance Lagrangian coordinate

No deformation External resistance

Simple deformation External resistance

Eulerian coordinate X

Lagrangian coordinate

Deff Deff Deff Deff 0dm 1 0 X x

Deff Deff 2 0 1 dm X 0x

Boundary conditions at x xmax whatever the time t

@X @X @ ÿdm Deff 2 0 @x dm 1 0 X x hm M hm M ... ... aw pvsat aw pvsat RT RT

X X1

X X1

ÿ0dm Deff

whose expressions depend on the hypothesis chosen to describe the product deformation. This hypothesis influences the nature of the boundary condition at x xmax . For the four mass transport models developed, the expressions for and for the boundary condition at the surface of the product may be found in Table 3.2 where Deff stands for the effective diffusivity (m2/s) and dm, 0dm and 0x stand, respectively, for dry matter concentration in the binary mixture (kg mÿ3), pure dry matter and pure water intrinsic densities (kg mÿ3). When the hypothesis of simplified deformation is retained, the Eulerian coordinates xi corresponding to Lagrangian coordinate i are recalculated as follows: Z i 0dm 1 0 X d xi x 0 Concerning model 1, with the assumption of no deformation and no external resistance to mass transfer, the generalised Fick equation may be solved analytically in the case of an infinite slab; the evolution of moisture content with time is given by Crank (1975): 1 2 X ÿ X1 8X 1 2 Deff 2 exp ÿ2n 1 t X0 ÿ X1 n0 2n 12 L2 where X is moisture content (g/g), X1 equilibrium moisture content (g/g), X0 initial moisture content (g/g), Deff effective moisture diffusivity (m2/s), t time


86


Fig. 3.11 Moisture uptake of dry biscuit when put in direct contact with 0.75 aw agar (), corresponding model fitting (± - ± -); moisture uptake of dry component when put in indirect contact with 0.75 aw agar with hydrophobic barrier at the interface (n), corresponding model fitting (ÐÐ) (from Bourlieu et al., 2008).

(s), and L thickness of the slab (m). When moisture transport occurs only from one surface of the slab, the thickness must be substituted by 2L. Concerning models 2, 3 and 4, with the assumption of a deformation and/or an external resistance, the generalised Fick equation can be solved numerically. The domain 0 x xmax is divided into N sub-regions of equal thickness x xmax =N . The first- and second-order spatial derivatives of the previous system of equations are then discretised. For example, the second-order derivation in the generalised Fick equation once discretised may be expressed as: Deff ;i Deff ;i1 dXi Deff ;iÿ1 Deff ;i X ÿ X Xi1 ÿ Xi iÿ1 i dt 2x2 2x2 These models were solved and usually applied to experimental data through MatlabÕ algorithms. Figure 3.11 shows how the moisture migration model is effective in describing the moisture uptake of dry biscuits in direct contact with a wet agar gel phase. Through barrier packaging and/or edible coatings In dense materials, small molecules will permeate thin layers (such as packaging films or edible films and coatings), implying a molecular diffusion due to the chemical potential differential between the two sides of the layer (Chao and Rizvi, 1988). The moisture transfer mechanism follows a three step-process: 1. sorption, whether coupled or not to condensation, 2. diffusion of the solute in a liquid state, and 3. desorption whether coupled or not to evaporation (Fig. 3.12). Sorption and desorption are considered as instantaneous, whereas condensation and/or evaporation require energy for the diffusion process. The whole phenomenon of mass transfer through barrier layers is usually described by simplified first Fick's laws:



87

Fig. 3.12 The three-steps mechanism of moisture transfer through barrier layers.

TR ÿD

@C m @x A t

where TR is the transfer rate (g mÿ2 sÿ1), D the diffusivity (m sÿ1), C the concentration (g mÿ3), x the distance (m), m the weight variation (g), A the surface exposed to the transfer (m), and t the time (s). In the case of gases and vapours, Henry's law gives the relationship between the concentration in the solid or liquid phase and the vapour pressure. It represents the partition coefficient between the condensed and vapour phases. The Henry coefficient S is also called the solubility coefficient. The value of concentration can be related to the vapour pressure, giving a transfer rate equation which can be expressed as: TR ÿD

@C @p ÿD S @x @x

and the permeability coefficient is defined as P D S, where S is the solubility coefficient (g mÿ3 Paÿ1), and p the partial vapour pressure (Pa). It therefore appears that permeability depends on both a kinetic parameter (diffusivity which represents the migration speed of the migrant molecule within the barrier layer), and a thermodynamic parameter (the solubility (sorption) coefficient which represents the affinity of the migrant molecule for the barrier layer). The water vapour permeability (WVP) is often expressed as a function of the water vapour transfer rate (WVTR) related to the thickness (L) and the vapour pressure differential at both faces of the barrier layer: L WVP WVTR p Elsewhere, the water activity is defined as the ratio of vapour pressure: p RH aw p0 T 100 The water transfer rate can be expressed as: WaterTR ÿD S

@p @aw aw P p0 k A aw ÿD S p0 @x L @x

where k is the overall mass transfer coefficient. However, knowledge of the sorption isotherm (solubility coefficient) and of the moisture diffusivity in packaging films is necessary to predict permeability.


88


Sorption may be measured using the previously described methods. The diffusivity coefficients of moisture in films may be measured according to several techniques. Moisture diffusivity in packaging films may be measured from the permeation kinetics according to Felder's solutions of Crank's equations (Felder, 1978). In the case of integral permeation which plots the cumulative amount of water vapour transferred through the film as a function of time, the diffusivity of moisture is equal to D L2 =6 where is the time lag before the constant transfer rate (stationary flow) is reached. When the transfer rate is directly measured through continuous techniques (differential permeation), the moisture diffusivity is determined from D L2 =7:199 t1=2 , where t1/2 is half the time taken to reach the stationary phase of the permeation process. When no apparatus for permeability measurement is available, the diffusivity can be determined from the sorption or desorption kinetics from the following solution of the second Fick's law solved by Crank (1975): " # 1 M1 ÿ Mt 8X 1 2n ÿ 12 2 DA t exp ÿ M1 ÿ M0 2 n1 2n ÿ 12 4L20 These methods for the diffusivity of moisture in packaging only apply for water vapour transfer. However, most foods contain liquid water and are in direct contact with the barrier films. The permeability of films by water sometimes depends on the physical state of water in contact with the packaging as shown by Morillon et al. (1998, 2000). Indeed, the liquid transfer rate through most packaging films is often much greater than that of water vapour permeability, and permeability is always measured for water vapour (Fig. 3.13). This discrepancy between liquid and vapour permeability for the same activity differential is called Schroeder's paradox. Karbowiak et al. (2008) proposed techniques ranging from the macroscopic to the molecular scale for the measurement of moisture diffusivity in films. The concentration±distance curves (or concentration profiles) method consists of bringing two cylinders of solid or semi-solid products into contact, either of which may contain a different initial water content or activity, and which are separated by the barrier layer. The concentration profiles are measured from their distance to the interface, as a function of time, along a onedimensional axis. From experimental kinetics obtained along the x-axis, and assuming D to be constant and independent of the concentration, the simplest analytical solution to Fick's second law can be applied (Crank, 1975) in order to estimate D app for the different times from a simple measurement of concentration at different points of the system: C ÿ C1 x erf p C0 ÿ C1 2 Dapp t where erf x is the integral error function (Karbowiak et al., 2008). The possibility of studying mass transfers in edible films is considered at a mesoscopic scale using goniometry (Karbowiak et al., 2006b), which entails



89

Fig. 3.13 Schroeder paradox: discrepancies of moisture transfers through cellulosebased and polyethylene films as a function of the state of water in contact with the films while the water activity gradients are the same (from Morillon et al., 2000).

investigation on a square-millimeter scale. The wetting phenomenon of a solid surface by a liquid was initially described by Young 200 years ago as the thermodynamic equilibrium measurement of the contact angle (Young, 1805). Although the value of the contact angle gives information on the surface hydrophobicity, it is not directly informative on transfers. However, the use of this technique in a dynamical way enables one to define mass transfers occurring on the surface of the studied material (Muszynski et al., 2003; Karbowiak et al., 2006b). In this case, it is important to take into account the evaporation flux involved at the liquid/air interface in order to correctly determine the absorption flux occurring at the solid/liquid interface. The investigation achieved by goniometry permits a rapid determination of the absorption flux of liquids within the tested materials and identification of the role of various constituents on the surface properties of the films. This approach is more relevant to food products in which more water is in a liquid state, as compared to vapour permeability. Fourier transform infrared (FTIR) spectroscopy is a very useful analytical technique, particularly for the identification of chemical groups within a molecule. Using the attenuated total reflectance (ATR) mode enables one to detect the chemical composition of a film surface. The infrared radiation penetrates the surface of the film in contact with the crystal. By putting liquid water on the other side of the film, the detection of a characteristic absorption band of water molecules enables the monitoring of water absorption over time. It then becomes possible to determine the apparent diffusion coefficient of water in the film, assuming a Fickian mechanism. This original approach has been developed by Fieldson and Barbari (1993), studying water diffusion in polyacrylonitrile. Diffusivity values obtained by this method are in accordance with those obtained


90


from gravimetric measurements and have been recently applied by Karbowiak et al. (2009) for moisture diffusivity measurement through thin biopolymer films. The increasing absorption of the hydroxyl band is thus characteristic of the water transfer occurring through the film. The observed phenomenon can be described by the classical Fickian diffusion equation through a membrane of L thickness, with a uniform initial distribution, and different surface concentrations. It may be solved using an analytic solution to Fick's law as applied to the transient state (Crank, 1975): " # 1 X Mt 8 ÿD2n 12 2 t 1ÿ exp 2 2 M1 4L2 n0 2n 1 In this case, the absorbance ratio At =A1 , obtained from the integration of the corresponding infrared absorption spectra, can be used in place of the mass ratio. However, such a classical Fickian model does not fit the data, because of structural changes or penetration depth of the detection IR beam. In this case, other factors need to be integrated to the model. 1. The penetration depth of the infrared radiation, dP: s dp 2 n2 2 n1 sin2 ÿ n1 with n1 refractive index of the crystal (ZnSe), n2 refractive index of the packaging film, angle of the incident radiation, and wavelength of the radiation. 2. The field evanescence, E: x Ex E0 exp ÿ dp E decreases exponentially as a function of the distance to the surface of the crystal. The integration of these parameters into Fick's diffusion equation leads to the following model: At 8 1ÿ L A1 dp 1 ÿ exp ÿ2 dp ! 3 ÿD2n 12 2 t 2n 1 L n 2 ÿ1 exp exp ÿ2 6 1 4L2 2L dp dp 7 X 6 7 6 7 ! exp6 2 7 4 5 4 2n 1 n0 2n 1 2 dp 2L 2



91

Fig. 3.14 Diffusion coefficient calculated from moisture absorption in edible carrageenan-based films determined from ATR-FTIR: experimental values, ÐÐ calculated values (from Karbowiak et al., 2009).

When these factors are taken into account, the model makes a good fit with the experimental values and allows an accurate prediction of moisture diffusivity in thin barrier layers as shown in Fig. 3.14. Knowledge of the solubility and diffusivity or permeability of moisture in film makes it possible to predict the shelf life of moisture-sensitive food products. Initially developed for the prediction of the shelf life of products packaged in plastic flexible or semi-rigid packaging, there are several simple models which also fit well with coated foods. These consider edible films and coatings as conventional packaging and take into account the characteristics of the food product, the surrounding medium, and the barrier layer (Labuza, 1982; Cardoso and Labuza, 1983; Hong et al., 1991). For instance, for water it is necessary to know: 1. the film permeability (WVP), or the water vapour transfer rate (WVTR) of the coatings, as well as the coating thickness; 2. the dry matter of the product to be coated, its initial and critical water contents, its sorption isotherm, and possibly its density and moisture diffusivity; 3. the water activity of the wet area and its possible diffusivity. These methods consist of evaluating the amount of transferred water necessary to induce product degradation, such as, e.g., the loss of crispness in a dry biscuit, and determining the time necessary for that amount of water to go through the coating (Debeaufort et al., 2002). The simplest model uses the water transfer rate and the critical water content of the dry part of the product according to the following equation: time

Mc ÿ Mi m Mc ÿ Mi m L WVTR A WVP A p


92


Fig. 3.15 Schematic representation of shelf life prediction for a model of a wrapped food (dry biscuit) in a plastic film or of a composite food (stuffed biscuit) whose dry and wet compartments are separated by a barrier coating.

In this equation, WVTR is the water transfer rate (g mÿ2 sÿ1), A is the surface exposed to the transfer; L is the coating thickness (m); m is the dry matter of the product (g); Mi and Mc are respectively the initial and critical water contents; and p is the water vapour partial pressure differential across the barrier. However, this model supposes that the water activity differential between the two faces of the coating remains constant over storage, which in reality never occurs. The progressive change in the water activity in the dry area (e.g., a cereal-based biscuit) has to be taken into account from the sorption isotherm modelled as linear between the initial and critical water contents if the predictions are to accord with reality (Labuza and Contreras-Medellin, 1981). Parameters used in this model are illustrated in Fig. 3.15, where Mi, Mc and Me are the initial, critical and equilibrium water contents, respectively; awi, awc, awe are the water activities; and bs is the slope of the linear-considered sorption isotherm between the initial and critical moisture content points. In an attempt to consider the water activity variation in the dry area, Labuza and Contreras-Medellin (1981) and Labuza and Altunakar (2007) suggested the following equation: Mi ÿ Me ln Mi ÿ Me P A p0 Mc ÿ Me time then time Lmb ln P A p0 Mc ÿ Me Lmb In this equation, P is the coating permeability (g mÿ1 sÿ1 Paÿ1), A is the exposed surface (m2), p0 is the saturated vapour pressure at the experiment temperature, L is the coating thickness (m), m is the dry matter of the product (g), and b is the average slope of the sorption isotherm of the dry area (gwater/gproduct). This model was used by Biquet and Labuza (1988) for chocolate coatings applied between agar-gel and microcrystalline cellulose. Morillon et al. (1998) also compared the estimated shelf life calculated from this latter model, with



93

Fig. 3.16 Prediction of moisture transfers in composite foods: application of cocoabased coatings to sugar wafer in contact with ice-cream at ÿ10 ëC (bars are estimated values from Contreras-Medellin and Labuza model and triangles are experimental values obtained from sensory analysis).

experimental values determined from sensory experiments for a wafer coated with a cocoa-based coating, as shown in Fig. 3.16. Even though many models have been developed for predicting either moisture diffusivity, sorption, permeability or shelf life of food in relation to moisture transfers, most of the works dealing with the control of moisture transfer between foods and the surrounding atmospheres, or between different areas within a composite food (sandwiches, stuffed biscuits, meat pies, etc.) remain empirical.

3.4

Moisture loss, gain and migration related to the shelf life

Water controls many properties of food materials at both the molecular and macroscopic levels. Such a role in both thermodynamic and dynamic properties is due to the interaction of water with the other food components. Food materials usually exist either in a liquid state (in solution or above melting temperature) or in a solid form. In the latter state, products may be either in the thermodynamically stable crystalline state or in the amorphous state (which is not a true equilibrium). Water uptake affects the structure of solid food through its plasticising effect which consists of the replacement of solute±solute interactions by solute±water interaction. This results in a softening, or even the dissolution of the material of increased hydration. Water relationships are particularly important in amorphous food systems. Such products are numerous, as many processes (baking, dissolution/drying) lead to a loss of crystallinity, which causes many foods to become amorphous below or above their Tg under storage conditions. Due to the plasticising effect of water, the glass transition temperature decreases with increasing water activity or content. Thus a food area which was initially glassy (hard and brittle)


94


will become rubbery (malleable) following a moisture-induced decrease of Tg below the ambient temperature. The physical properties of the foods will then be particularly affected by water migration. 3.4.1 Plasticisation A decrease of Tg below the ambient or working temperature results in a marked decrease in the rigidity of the hydrated product which is then plasticised. However, in certain cases, water may exhibit an anti-plasticising effect. Instrumental and sensory measurements have shown that for many products, rigidity increases with increasing water activity in a limited humidity range, generally below 11% of water (Kapsalis et al., 1970; Harris and Peleg, 1996; Roudaut et al., 1998; Suwonsichon and Peleg, 1998; Valles-Pamies et al., 2000; Waichungo et al., 2000). Indeed, water keeps its ability to decrease the glass transition temperature. This mechanism, described as anti-plasticising for its opposite effect to plasticisation, is of practical importance but its physical origin is as yet not fully understood. 3.4.2 Crispness Among the textural changes occurring due to water migration or water activity changes, crispness is a well-known critical property. By definition, this attribute of texture may be associated with a combination of the high-pitched sound and the crumbling of the product as it is crushed through. Such behaviour is generally encountered with puffed and brittle food products at low hydration (25%), in the rubbery plateau region, they are soft at room temperature and not very sensitive to changes in moisture content (Roudaut, 2007). They become progressively


96


more leathery as moisture content decreases. Finally, below 20% moisture (this critical value depends, of course, on their composition), as their Tg becomes greater than ambient temperature, their Young's modulus increases sharply and they become brittle. 3.4.4 Stickiness, caking and collapse Generally used to describe powder behaviour at a macroscopic scale, caking (formation of hard masses of greater size) might be observed as being a result of stickiness, when individual solids of a free flowing powder stick to one another and ultimately form a mass of solids. The two main conditions under which caking is known to occur are drying and storage. However, the key parameter of the phenomenon is the same: a critical hydration level. When powder hydration increases, the surface viscosity of particles falls below 108 Pa s (Downton et al., 1982) which tends to cause a merging with neighbouring particles through interparticle liquid bridging. The latter will take place if sufficient flow occurs during contact and the bridge resists subsequent deformation. Caking can be evaluated by a caking index which corresponds to the amount of a sample (expressed as a percentage) retained by a given mesh. Figure 3.18 shows the effects of humidity on the caking kinetics of fish hydrolysates: the higher the humidity (or temperature) the higher the caking index (Aguilera et al., 1993). These structural changes will also be accompanied by changes in physical properties, such as an alteration of the visual aspect, flowability and water dispersibility. Caking and structure collapse processes generally occur in products with high levels of soluble sugars, minerals or protein hydrolysates (such as milk powders, instant coffee, and dehydrated fruit juices). Structural collapse is driven by the same mechanisms as caking and should be seen as an advanced stage of caking. As a result of the lower viscosity, the walls

Fig. 3.18 Caking index versus time for fish protein hydrolysates stored at various aw at 30 ëC (adapted from Aguilera et al., 1993).



97

of the porous material become closer and the inter-pore space which controls the porosity of the material is progressively lost, leading to a more compact, collapsed material. 3.4.5 Crystallisation Given the threshold molecular mobility (both rotational and translational) required for crystallisation to take place, it is generally accepted that mobility below Tg is not sufficient to allow crystallisation. When water activity increases, the solute molecules become increasingly mobile and collide in the correct orientation for the formation of a nucleus. This is the starting point for crystallisation which eventually spreads to the entire matrix. Anhydrous crystal formation would expel water from the matrix, and depending on the close environment of the sample, the water molecules will diffuse towards the neighbouring phase or to the atmosphere when exposed to the open air. Among the products other than spray-dried milk powders (which are the subject of most publications on crystallisation in foods) for which crystallisation is particularly critical, ice cream, hard candies, soft cookies and baked products can be mentioned. The crystallisation rate compared with the temperature is known to exhibit a maximum between Tg and the melting temperature Tm: at low temperatures, the high viscosity hinders the diffusion required for crystal growth, whereas nucleation is limited at a temperature approaching Tm. A similar bell-shape behaviour has been described for isothermal crystallisation rate as a function of water content (Roos, 1995).

3.5 Conditions for moisture migration and foods affected by moisture transfer As soon as a relative humidity (RH) gradient (or water activity gradient) exists between a food and its environment, or between two food areas, there will be water migration: water will diffuse from the high to the low RH phase until equilibrium is reached between the two phases. Two situations exist for the occurrence of moisture migration: moisture transfer with the atmosphere and moisture transfer within the product. 3.5.1 Moisture transfer with atmosphere Water transfer may take place between the food and the surrounding atmosphere. Packaging has been developed to prevent various exchanges with the atmosphere and the permeability of the packaging material will control the dynamics of these transfers. The best candidates with regard to moisture exchange with the atmosphere are the food products having an RH range very different from the ambient RH. Owing to their high optimal storage RHs (85±95% and 90±98% respectively for fruits and vegetables), these foods are prone to moisture loss when stored under


98


normal conditions. Moisture loss or uptake causes wilting, shrinkage and loss of the firmness, crispness and succulence of the product. However, too high a moisture barrier will cause a high RH in the packaging, leading to microbial growth (Petersen et al., 1999). Owing to the changes in water chemical potential values with temperature, refrigerated or frozen products are likely to be subject to moisture migration. Dehydration of unwrapped butter (aw: 0.75 at 25 ëC, Shukla et al., 1994) is quite usual when stored in the fridge. As a result, the surface of the butter dries out and becomes more yellow, which is often seen by the consumer to be an indication of lower quality. In cheeses such as Brie or Camembert, water exchanges with the environment must be maintained if the surface flora is to be kept alive and avoid an alteration in the appearance and organoleptic properties (Mousavi et al., 1998). The optimisation of packaging for such products is therefore of primary importance. Similarly when frozen products are submitted to temperature variations during storage, they may suffer from moisture migration to or from the atmosphere within the pack. This results either in freezer burns (when the surrounding temperature is greater than that of the food) or in frost deposition at the surface of the products (when the temperature of the latter is lower than that of the atmosphere). Moisture loss through sublimation from the surface of the product leads to freezer burns and this is avoided by using packaging material which is highly impermeable to water vapour and sticks tightly to the surface of the frozen food. In general, moisture exchange will impair the visual appearance of the product and lead to unacceptable changes in the weight of the product. 3.5.2 Moisture transfer within the product Owing to the RH gradient existing between different areas, composite food materials are excellent candidates for moisture migration between the different parts of the food itself. Multi-domain systems may be encountered at different levels: molecular or macroscopic (Labuza and Hyman, 1998). Different examples of macroscopic or composite materials which are candidates for moisture transfer are: genuinely multiple layer foods, or multiple food components packed together in a single container (breakfast cereals, salad or soup mixes). The difference between the two categories may be defined by the extent of the contact between the components. Multi-component foods are generally designed to have attractive contrasts of texture. These are controlled primarily by the association of multiple components which differ in texture, such as in-filled products (breakfast cereals, biscuits, sandwiches, ice cream wafers), dairy products with fruits and/or cereal balls and soups with croutons. Textural contrasts are generally achieved by differences in structure and/or composition with water being a key ingredient. One of the simplest examples of such contrast is bread. The contrast between the soft crumb and the crisp crust results from the baking process, during which



99

Fig. 3.19 Moisture transfer in bread from crust to centre as a function of time (adapted from Piazza and Masi, 1995).

the surface of the bread is exposed to a higher temperature than the crumb. This causes both a RH and water content gradient between the bread surface and its core which is responsible for a change in the moisture distribution in the bread over time (Fig. 3.19) (Piazza and Masi, 1995). This is the cause of the familiar result of a softened crust and a stiffened crumb. It should be mentioned that starch retrogradation (recrystallisation) also contributes significantly to textural changes in the crumb as it is likely that the lower water content will have an effect on the starch reorganisation (Baik and Chinachoti, 2000). The aforementioned molecular level at which moisture migration may occur has also been described as playing a part in the deterioration of bread during storage. It has been claimed that moisture migration from gluten to starch occurs in the early stage of crumb evolution (Chen et al., 1997). When considered separately, the different components of composite foods (for example, a biscuit and a strawberry jam filling) differ in their water content and thus often by their water activity (respectively 0.4 and 0.84). When in contact with a multilayer food, water will migrate from the jam to the cake, inducing a softening of the cake and a thickening of the jam as a result of increasing concentration and the consequent crystallisation of sugar. Moreover, the latter may exacerbate the loss of quality, since sugar crystallisation may cause syneresis from the filling. Similarly, composite foods such as dairy products or ice creams (both of which have high RH) with confectionery additions (lower RH) will also be affected by moisture migration between the components. As a result, the added sugary components will lose their edible quality, becoming soft or even disappearing through dissolution into the surrounding aqueous environment. In some cases, the migration may have a deleterious impact on the other component which originally had the greater RH. Indeed, its decreased hydration may promote solute crystallisation in the vicinity of the addition. Increased granularity in ice cream may result from lactose crystallisation caused by the partial `dehydration' of the ice cream, to the benefit of the added components.


100


Avoiding or limiting the above mentioned physical and chemical changes resulting from water uptake or loss may be achieved by controlling moisture migration into foods, or between the different components of the food. As the water activity gradient is one of the driving forces for moisture migration, one of the main solutions for stability in composite foods consists of limiting the gradient which exists between the components, often by lowering the water activity of the moist phase. This is generally achieved by the use of solutes, such as sugars, salts or polyols, which interact strongly with water, thus depressing the water activity of the highest RH phase. The use of humectants is generally limited by their solubility (no effect above saturation concentration), their reactivity (e.g., reducing sugars will result in the Maillard reaction), their effect on texture and taste and by regulatory requirements. Further information on water activity is given in the course of this chapter. There are many textbooks on water activity itself, and on water activity in foods which may be referred to for further information (Barbosa-CaÂnova et al., 2007, and references therein) on ways of manipulating water activity gradients. In addition to the thermodynamic control of moisture migration, the dynamic properties of food may also provide a means of influencing the extent of water transfer (Labuza and Hyman, 1998). The kinetics of water changes may be affected by the matrix composition and structure (through an increased viscosity, and decreased diffusivity of water) or by the application of a water barrier between the phases of differing water activity.

3.6

References

and LILLFORD PJ (1986), `The significance of water in the baking process'. In: Blanshard JMV, Frazier PJ and Galliard T (eds), Chemistry and Physics of Baking: Materials, Processes and Products, Royal Society of Chemistry, London, 30±41. AGUERRE RJ, SUAREZ C and VIOLLAZ PE (1989), `New BET type multiplayer sorption isotherms. Part II: Modelling water sorption in foods', Lebensmittel-Wissenchaft und Technology 22: 192±195. AGUILERA JM, LEVI G and KAREL M (1993), Èffect of water content on the glass transition and caking of fish protein hydrolyzates', Biotechnology Progress 9: 651±654. AL MUHTASEB AH, MCMINN WAM and MAGEE TRA (2002), `Moisture sorption isotherm characteristics of food products: a review', Transaction of the Institution of Chemical Engineers 80: 118±128. ATTENBURROW GE, DAVIES AP, GOODBAND RM and INGMAN SJ (1992), `The fracture behaviour of starch and gluten in the glassy state', Journal of Cereal Science 16: 1±12. BAIK MY and CHINACHOTI P (2000), `Moisture redistribution and phase transitions during bread staling', Cereal Chemistry 77(4): 484±488. Â NOVAS GV, FONTANA AJ, SCHMIDT SHELLY J and LABUZA TP (2007), Water BARBOSA-CA Activity in Foods ± Fundamentals and Applications. John Wiley & Sons, New York. ABLETT S, ATTENBURROW GE



101

and DAUDIN JD (2000), `Development of a new method for fast measurement of water sorption isotherms in the high humidity range validation on gelatin gel', Journal of Food Engineering 44(2): 97±107. BIQUET B and LABUZA TP (1988), Èvaluation of the moisture permeability characteristics of chocolate films as an edible moisture barrier', Journal of Food Science 53: 989± 998. BOUDHRIOUA N, BONAZZI C and DAUDIN JD (2003), Èstimation of moisture diffusivity in gelatin-starch gels using time-dependent concentration±distance curves at constant temperature', Food Chemistry 82: 139±149. BOURLIEU C, GUILLARD V, POWELL H, VALLES-PAMIES B, GUILBERT S and GONTARD N (2008), `Modelling and control of moisture transfers in high intermediate and low composite food', Food Chemistry 106(4): 1350±1358. BRENNAN JG, JOWITT R and WILLIAMS A (1974), `Sensory and instrumental measurement of ``brittleness'' and ``crispness'' in biscuits', Proceedings 4th International Congress Food Science Technology, Vol. 11: 130±143. BRUNAUER S, DEMING LS, DEMING WE and TROLLER E (1940), Òn the theory of Van der Waals adsorption of gases', Journal of American Chemical Society 62: 1723± 1732. CARDOSO G and LABUZA TP (1983), `Prediction of moisture gain and loss for packaged pasta subjected to a sine wave temperature/humidity environment', Food Technology 18: 587±606. CHAO RR and RIZVI HH (1988), Òxygen and water vapour transport through polymeric films: a review of modelling approaches'. In Hotchkiss JH (ed.), Food Packaging Interactions, Am. Chem. Soc., Washington, DC, 216±242. CHEN PL, LONG Z, RUAN R and LABUZA TP (1997), `Nuclear magnetic resonance studies of water mobility in bread during storage', Lebensmittel-Wissenschaft und Technologie 30(2): 178±183. CRANK J (1975), The Mathematics of Diffusion, 2nd edn, Oxford University Press, Oxford. DEBEAUFORT F, VOILLEY A and GUILBERT S (2002), `Les proceÂdeÂs de stabilisation des produits par des films ``barrieÁre'''. In Lorient D, Simatos D, Le Meste M (eds), ProprieÂteÂs de l'eau dans les aliments, Tec and Doc Lavoisier, Paris, 549±600. DOWNTON GE, FLORES-LUNA JL and KING CJ (1982), `Mechanism of stickiness in hygroscopic, amorphous powders', Industrial Engineering and Chemistry Fundamentals 21: 447±451. FELDER RM (1978), Èstimation of the gas transport coefficients from differential permeation, integral permeation, and sorption rate data', Journal of Membrane Science 3: 15±27. FENNEMA OR (1985), Food Chemistry, 2nd edn, Marcel Dekker, New York, 46±50. FERRO-FONTAN C, CHIRIFE J, SANCHO E and IGLESIAS HA (1982), Ànalysis of a model for water sorption phenomena in foods', Journal of Food Science 47: 1590±1594. FIELDSON GT and BARBARI TA (1993), `The use of FTIR-ATR spectroscopy to characterize penetrant diffusion in polymers', Polymer 34(6): 1146±1153. GUGGENHEIM EA (1966), Application of Statistical Mechanics, Clarendon Press, Oxford. GUILLARD V, BROYART B, BONAZZI C, GUILBERT S and GONTARD N (2003), `Modelling of moisture transfer in a composite food: dynamic water properties in an intermediate aw porous product in contact with high aw filling', Chemical Engineering Research and Design 81(9): 1090±1098. HALSEY G (1948), `Physical adsorption on non-uniform surfaces', Journal of Chemistry Physics 16: 931±937. BEAUCOUR P


102


and PELEG M (1996), `Patterns of textural changes in brittle cellular foods caused by moisture sorption', Journal of Texture Studies 73: 225±231. HENDERSON SM (1952), À basic concept of equilibrium moisture', Agricultural Engineering 33: 29±32. HICKEY H, MACMILLAN B, NEWLING B, RAMESH M, VAN EIJCK P and BALCOM B (2006), `Magnetic resonance relaxation measurements to determine oil and water content in fried foods', Food Research International 39: 612±618. HONG YC, KOELSCH CM and LABUZA TP (1991), Ùsing the L number to predict the efficacy of moisture barrier properties of edible coating materials', Journal of Food Processing and Preservation 15: 45±62. HOPKINSON I, JONES RAL, MCDONALD PJ, NEWLING BA, LECAT A and LIVINGS S (2001), `Water ingress into starch and sucrose:starch systems', Polymer 42: 4947±4956. HUTCHINSON RJ, MANTLE SA and SMITH AC (1989), `The effect of moisture content on the mechanical properties of extruded food foams', Journal of Material Science 24: 3249±3253. ISENGARD HD (1995), `Rapid water determination in foodstuffs', Trends in Food Science and Technology 6: 155±162 JOWITT P and WAGSTAFFE PJ (1989), The Certification of the Water Content of Microcrystalline Cellulose at 10 Water Activities. Commission of the European Communities, Community Bureau of Reference, CRM 302, EUR 12429 EN. KALETUNC G and BRESLAUER KJ (1993), `Glass transition of extrudates: relationship with processing-induced fragmentation and end-product attributes', Cereal Chemistry 70: 548±552. KAPSALIS JG, WALKER JE and WOLF JC (1970), À physico-chemical study of the mechanical properties of low and intermediate moisture foods', Journal of Texture Studies 1: 464±483. KARATHANOS VT, KOSTAROPOULOS AE and SARAVACOS GD (1995), `Diffusion and equilibrium of water in dough/raisin mixtures', Journal of Food Engineering 25: 113± 121. KARBOWIAK T, DEBEAUFORT F, CHAMPION D and VOILLEY A (2006a), `Wetting properties at the surface of iota-carrageenan-based edible films', Journal of Colloid and Interface Science 294(2): 400±410. KARBOWIAK T, HERVET H, LEGER L, CHAMPION D, DEBEAUFORT F and VOILLEY A (2006b), Èffect of plasticizers (water and glycerol) on the diffusion of a small molecule in iota-carrageenan biopolymer films for edible coating application', Biomacromolecules 7(6): 2011±2019. KARBOWIAK T, GOUGEON R, RIGOLET S, DELMOTTE L, DEBEAUFORT F and VOILLEY A (2008), `Diffusion of small molecules in edible films: effect of water and interactions between diffusant and biopolymer', Food Chemistry 106(4): 1340±1349. KARBOWIAK T, DEBEAUFORT F, VOILLEY A and TRYSTRAM G (2009), `From macroscopic to molecular scale investigations of mass transfer of small molecules through edible packaging applied at interfaces of multiphase food products', Innovative Food Science and Emerging Technologies 10: 116±127. KATZ EE and LABUZA TP (1981), Èffect of water activity on the sensory crispness and mechanical deformation of snack food products', Journal of Food Science 46: 403±409. KIRBY AR, PARKER R and SMITH AC (1993), Èffects of plasticizers on the fracture behaviour of wheat starch', Special publication of the Royal Society of Chemistry 133: 250. HARRIS M


Moisture loss, gain and migration in foods LABUZA TP

97.

103

(1982), `Moisture gain and loss in packaged foods', Food Technology 36: 92±

and ALTUNAKAR B (2007), `Diffusion and sorption kinetics of water in foods'. In Barbosa-Canovas GV, Fontana AJ, Schmidt SJ and Labuza TP (eds), Water Activity in Foods: Fundamentals and Applications, Blackwell Publishing, Oxford, 215±237. LABUZA TP and CONTRERAS-MEDELLIN R (1981), `Prediction of moisture protection requirements for foods', Cereal Food World 26: 335±343. LABUZA T and HYMAN CR (1998), `Moisture migration and control in multi-domain foods', Trends in Food Science and Technology 9: 47±55. LE MESTE M, HUANG VT, PANAMA J, ANDERSON G and LENTZ R (1992), `Glass transition of bread', Cereal Food World 37: 264±267. LE MESTE M, ROUDAUT G and ROLEÂE A (1996), `Physical state and food quality ± an example: the texture of cereal-based-foods'. In Fito P, Ortega-Rodriguez E and Barbosa-CaÂnovas GV (eds), Food Engineering 2000, Chapman and Hall, New York, 97±113. LEVINE H and SLADE L (1990), Ìnfluence of the glassy and rubbery states on the thermal, mechanical, and structural properties of doughs and baked products'. In Faridi H and Faubion JM (eds), Dough Rheology and Baked Product Texture, Van Nostrand Reinhold, New York, 157±330. LI Y, KLOEPPEL KM and HSIEH F (1998), `Texture of glassy corn cakes as function of moisture content', Journal of Food Science 63: 869±872. LODI A, ABDULJALIL A and VODOVOTZ Y (2007), `Characterization of water distribution in bread using magnetic resonance imaging', Magnetic Resonance Imaging 25: 1449±1458. MATHLOUTHI M (2001), `Water content, water activity, water structure and the stability of foodstuffs', Food Control 12: 409±417. MORILLON V, DEBEAUFORT F, CAPELLE M, BLOND G and VOILLEY A (1998), Èffect of the water properties on the moisture barrier efficiency of lipid-sugar based edible coatings'. ISOPOW VII ± Water Management in the Design and Distribution of Quality Foods. 30 May±4 June, Helsinki, Finland. MORILLON V, DEBEAUFORT F, CAPELLE M, BLOND G and VOILLEY A (2000), Ìnfluence of the physical state of water on the barrier properties of hydrophilic and hydrophobic films', Journal of Agricultural and Food Chemistry 48(1): 11±16. MOUSAVI SM, DESOBRY S and HARDY J (1998), `Mathematical modelling of migration of volatile compounds into packaged food via package free space. Part I: Cylindrical shaped food', Journal of Food Engineering 36(4): 453±472. MUSZYNSKI L, WALINDER MEP, PIRVU C, GARNER DJ and SHALER SM (2003), Àpplication of droplet dynamics analysis for assessment of water penetration resistance of coatings'. In Mittal KL (ed.), Contact Angle, Wettability and Adhesion. VSP, Utrecht, 463±478. NICHOLLS RJ, APPELQVIST IAM, DAVIES AP, INGMAN SJ and LILLFORD PJ (1995), `Glass transitions and fracture behaviour of gluten and starches within the glassy state', Journal of Cereal Science 21: 25±36. NIKOLAIDIS A and LABUZA TP (1996), `Glass transition state diagram of a baked cracker and its relationships to gluten', Journal of Food Science 61(4): 803±806. OSWIN CR (1946), `The kinetics of package life III. The isotherm', Journal of Chemical Industry 65: 419±421. PELEG M (1993), Àssessment of a semi-empirical four parameter general model for LABUZA TP


104


sigmoid moisture sorption isotherms', Journal of Food Process Engineering 16: 21±37. PELEG M (1994), À mathematical model of crunchiness/crispness loss in breakfast cereals', Journal of Texture Studies 25: 403±410. PELEG M (1998), Ìnstrumental and sensory detection of simultaneous brittleness loss and moisture toughening in three puffed cereals', Journal of Texture Studies 29: 255±274. PETERSEN K, VáGGEMOSE NIELSEN P, BERTELSEN G, LAWTHER M, OLSEN MB, NILSSON NH and MORTENSEN G (1999), `Potential of biobased materials for food packaging', Trends in Food Science and Technology 10(2): 52±68. PIAZZA L and MASI P (1995), `Moisture redistribution throughout the bread loaf during staling and its effect on mechanical properties', Cereal Chemistry 72(3): 320±325. QUEÂNARD D and SALLEÂE H (1991), `Le transfert isotherme de la vapeur d'eau condensable dans les mateÂriaux microporeux du baÃtiment', Cahier CSTB 323: 1±53. ROCA E, BROYART B, GUILBERT S and GONTARD N (2008), Èffective moisture diffusivity modelling versus food structure and hygroscopicity', Food Chemistry 106: 1428± 1437. ROOS Y (1995), Phase Transitions in Foods, Academic Press, New York. ROOS Y, ROININEN K, JOUPPILA K and TUORILA H (1998), `Glass transition and water plasticization effects on crispness of a snack food extrudate', International Journal of Food Properties 1: 163±180. ROUDAUT G (2007), `Water activity and physical stability'. In Barbosa-CaÂnova GV, Fontana AJ, Schmidt SJ and Labuza TP (eds), Water Activity in Foods: Fundamentals and Applications, Blackwell, Oxford, 199±213. ROUDAUT G, DACREMONT C and LE MESTE M (1998), Ìnfluence of water on the crispness of cereal based foods acoustic, mechanical, and sensory studies', Journal of Texture Studies 29: 199±213. RUAN RR and CHEN PL (1998), Water in Foods and Biological Materials: A Nuclear Magnetic Resonance Approach. Technomic Publishing, Lancaster, PA. SAUVAGEOT F and BLOND G (1991), Èffect of water activity on crispness of breakfast cereals', Journal of Texture Studies 22: 423±442. SHUKLA A, BHASKAR A, RIZVI S and MULVANEY S (1994), `Physicochemical and rheological properties of butter made from supercritically fractionated milk fat', Journal of Dairy Science 77(1): 47±54. SIMATOS D (2002), `ProprieÂteÂs de l'eau dans les produits alimentaires: activiteÂ de l'eau, diagrammes de phases et d'eÂtats', In Le Meste M, Simatos D and Lorient D (eds), L'eau dans les aliments, Tech and Doc Lavoisier, Paris, 49±79. SLADE L and LEVINE H (1993), `The glassy state phenomenon in food molecules'. In Blanshard JMV and Lillford PJ (eds), The Glassy State in Foods, Nottingham University Press, Nottingham, 35±102. SMITH SE (1947), `The sorption of water vapour by high polymers', Journal of the American Chemical Society 69: 646±651. SUWONSICHON T and PELEG M (1998), Ìnstrumental and sensory detection of simultaneous brittleness loss and moisture toughening in three puffed cereals', Journal of Texture Studies 29: 255±274. TESCH R, NORMAND MD and PELEG M (1996), `Comparison of the acoustic and mechanical signatures of two cellular crunchy cereal foods at various water activities levels', Journal of the Science of Food and Agriculture 70: 347±354. TROLLER JA and CHRISTIAN JHB (1978), Water Activity and Food, Academic Press, New York, 13±47.



105

and MITCHELL JR (2000), Ùnderstanding the texture of low moisture cereal products: mechanical and sensory measurements of crispness', Journal of the Science of Food and Agriculture 80: 1679±1685. VAN DEN BERG C (1991), `Food±water relations: progress and integration, comments and thoughts'. In Levine H and Slade L (eds), Water Relations in Foods, Plenum Press, New York, 21±28. VAN DEN BERG C and BRUIN S (1981), `Water activity and its estimation in food systems'. In Rockland LB and Stewart GF (eds), Water Activity: Influences on Food Quality, Academic Press, New York, 147±177. VOILLEY A and BETTENFELD ML (1985), `Diffusivities of volatiles in concentrated solutions', Journal of Food Engineering 4(4): 313±323. VORNHOF DW and THOMAS JH (1970), `Determination of moisture in starch hydrolysates by near-infrared and infrared spectrophotometry', Analytical Chemistry 42: 1230. WAICHUNGO WW, HEYMANN H and HELDMAN DR (2000), Ùsing descriptive analysis to characterize the effects of moisture sorption on the texture of low moisture foods', Journal of Texture Studies 15: 35±46. WU S (1992), `Secondary relaxation, brittle-ductile transition temperature, and chain structure', Journal of Applied Polymer Science 46: 619±624. YOUNG T (1805), Àn essay on the cohesion of fluids', Philosophical Transactions of the Royal Society 95: 65±87. YU X, SCHMIDT AR, BELLO-PEREZ LA and SCHMIDT SJ (2008), `Determination of the bulk moisture diffusion coefficient for corn starch using an automated water sorption instrument', Journal of Agricultural and Food Chemistry 56: 50±58. VALLES-PAMIES B, ROUDAUT G, DACREMONT C, LE MESTE M


4 Insect and mite penetration and contamination of packaged foods C. H. Bell, Food and Environment Research Agency, UK

Abstract: The importance of packaging as a protective measure for stored food products and the currently available packaging materials are described. The pest status and life histories of the principal insect and mite species attacking packaged foods at various points in the supply train are outlined. Some of the most frequently occurring circumstances affecting the vulnerability of packaged foods are discussed together with measures to combat pest problems. Complete sealing of the package is needed and means for the early detection of pest populations whenever products are held in storage along the path between manufacturer and consumer need to be in place. Materials acting as attractants or repellents have a useful part to play in the protection of the product. Key words: insects, mites, packaged foods, MAP, storage, penetration of plastics.

4.1

Introduction

Any consideration of how best to protect and preserve foodstuffs by packaging needs to take into account measures designed to avoid infestation by insects or mites. These pests can locate and enter any flaw in packaging and are able to reproduce at an alarming rate. A widely dispersed residual population in the fabric of the storage or holding premise is the source of attack. Though they may be widely separated, individuals of many insect pest species release chemicals known as pheromones which attract the opposite sex and, once breeding starts, a population increase rate of twenty to thirty-fold per month is not uncommon, and even higher rates of increase may be expected from mite species under


Insect and mite penetration and contamination of packaged foods

107

favourable conditions. With the increased stringency for hygiene in food trading standards, the discovery of a single insect in an imported consignment can lead to rejection by the port health authority of the entire batch, with severe economic and legal consequences for all involved in the trade chain. The risk for packaged products is highest in warm climates encountering periods of high humidity and where the residence time for products is prolonged. Today most non-perishable foods are packaged prior to distribution to the consumer and with the exception of canned products, most are susceptible to attack by stored product insects (Highland, 1984). One of the first synthetic packaging materials was cellophane, produced as an extrusion from a solvent solution of the plant polymer cellulose, (C6H10O5)n. Today there are many other materials in use, all loosely termed as plastics and often utilising cellulose ethers in their manufacture. Many are laminated, some layers providing tensile strength and others impermeability to moisture or gases (Greengrass, 1993). The materials in most frequent use are listed in Table 4.1. The desire to achieve gas-tight as well as moisture-proof barriers led to the development of modified atmosphere packaging (MAP). First practised on fruit in the 1920s, followed by meat and fish in the 1930s, the technique remained as a specialist option until its combination with vacuum packaging in the 1970s (Parry, 1993). The sudden popularity of shrink wrapping under vacuum of products such as poultry prior to freezing paved the way for an exponential expansion of not only modified atmosphere packaging but a much wider consideration of packaging applications, until by the end of the 1990s nearly every food product and even newspapers and magazines found themselves in plastic wraps. The sealing of the packaged food must remain intact to be effective, but insects possess some formidable equipment to gain entry. Mites and the younger stages of insects are very small and can exploit the tiniest openings. Characterised as arthropods by having an exoskeleton, some insects are heavily sclerotised as adults. Many beetles possess powerful mandibles, both as adults and larvae, which can cut through edges and folds in the package. Moth larvae also have strong biting and chewing mouthparts. In cockroaches and some beetles, a secondary set of biting or chewing apparatus is provided by the galeae and laciniae, distal lobes of the maxillae head appendages, which have developed sclerotised hooks and cutting surfaces (Fig. 4.1). In beetles which attack whole grains the combination of mouthparts allows holes to be `drilled' through the testa of seeds, and, of concern here, through various packaging materials.

4.2

Insects and mites contaminating stored food products

4.2.1 Coleoptera Beetles comprise the largest group of stored product pest species with nearly 200 species from 14 families having been implicated in storage problems. Only some of these qualify as pests of major importance (Bell, 2003) and not all of this


Table 4.1

Some properties of materials in common use for the packaging of food products (after White and Roberts, 1992; Greengrass, 1993)


Material

Polyethylene LD Polyethylene HD Polypropylene Ethylene vinyl acetate (EVA) Polyvinyl chloride (PVC) (unplasticised) Polyvinylidene chloride (PVdC) Nylon 6 (polyamide) Polyethylene tetraphthalate (PET) Polystyrene-butadiene Ethylene vinyl alcohol (EVOH)

Tensile strength

Thermoformability

Permeability at 25 ëC (litres/m2.day.atm) for 25 micron film Oxygen

Nitrogen

Carbon dioxide

Moisture transfer at 38 ëC, 90% r.h. (g/m2.day)

Good Good Fair Poor Good

Good Poor Good Good Good

7.8 2.6 2.0±3.7 12.5 0.15±0.35

2.8 0.65 0.4±0.7 4.9 0.06±0.15

42.0 7.6 8.0±10.0 50.0 0.45±1.0

16±24 6±10 6±12 40±60 22±40

Fair Very good Very good Good Fair

Good Fair Poor Good Poor

0.002±0.01 0.04±0.08 50±130 5.0 0.001±0.005

0.001±0.002 0.014 0.015±0.018 0.8 na

0.01±0.03 0.15±0.19 0.18±0.39 18.0 na

0.8±3.2 80±300 20±50 100±125 16±80


109

Fig. 4.1 Mouthparts of Blatta orientalis. 1. Mandibles: pr, prostheca; ab.m, abductor muscle; ad.m, adductor muscle. 2. Left maxilla: mx.p, maxillary palp; g, galea; l, lacinia; s, stipes; sg, subgalea; c, cardo. 3. Labium: gl, glossa; pg, paraglossa; l.p., labial palp; pm, prementum; pgr, palpiger; m, mentum; sm, submentum. After Imms (1951).

smaller group are a threat to packaged products. The principal pests are described below by family. Anobiidae Lasioderma serricorne (F.) The cigarette or tobacco beetle has long been associated with dried vegetable products, and is one of the most widespread of all stored product pests. Of tropical origin and a strong if slow flyer, it favours warm, moderately humid conditions, and in temperate zones readily colonises any heated building where food material is processed or stored. Damage is caused by the larvae which have powerful biting mouthparts. Adults, whilst able to perforate dry tobacco leaves, do not feed in store, but may feed on nectar in the open. The biology of L. serricorne was reviewed by Ashworth (1993). The short-lived adults are about 3 mm in length and reddish brown in colour with smooth elytra. Though notorious as a pest of tobacco, L. serricorne is also a pest of cocoa, soybeans, various cereals, spices, textiles and many other products. It breeds rapidly, multiplying 20-fold in a 4-week period at 32±35 ëC with up to six generations per year in the tropics. A minimum temperature of 22 ëC and, at optimal temperatures, a minimum relative humidity (r.h.) of 30% are needed for a population increase (Howe, 1965).


110


Stegobium paniceum (L.) The drug store, biscuit beetle or bread beetle is one of the most long-standing pests of homes and stores in temperate zones with numerous records going back to the birth of applied entomology. Excavations of archaeological sites have found dead specimens in leather artefacts from Roman times and in the remains of food left in tombs in ancient Egypt. Similar in appearance to L. serricorne except for its striated elytra, the 2±3 mm beetle is found on a very wide range of stored products, favouring finely divided materials for oviposition. The larvae cause damage to leather, textiles, tobacco, papers, wood, rubber and cork as well as food materials. S. paniceum was a particular pest of biscuits stored on sailing ships in the days of long voyages, but today it is found in bakeries, commercial retail stores and domestic larders. Here, for development to lead to a population increase a minimum temperature of 17 ëC is needed and development proceeds fastest (7.5 times increase over 28 days) at 25±28 ëC (Howe, 1965). High humidity is favoured. Bostrichidae Rhyzopertha dominica (F.) The lesser grain borer (Fig. 4.2a) is a serious pest of cereal grains, adults and larvae both boring into grain in which development is completed. Damaged grains are preferentially selected for attack and, if unchecked, populations can cause total loss of product. R. dominica can also complete development on various flours, meals and macaroni, and is able to penetrate most packaging materials. Development is rapid (c. 25 days at 34 ëC) with a high temperature optimum of 32±35 ëC and temperature and r.h. minima of 19 ëC or 30% r.h. (Howe, 1965; Arbogast, 1991). A 20-fold population increase can occur every 4 weeks and, in spite of the high temperature developmental range, the adults are cold hardy. Eggs are laid in batches on the kernels of grain or singly in frass or meal. The larvae are unusual in that they start life as active, campodeiform types while later instars are scarabeiform and immobile. The shiny dark brown 2±3 mm adults fly readily in warm conditions and a single female can lay over 400 eggs over a 5-month period. Bruchidae There are a very large number of beetles from this family that infest leguminous crops in the field but only a few can continue breeding in store. Problems have occurred when infested beans have been packaged and the non-feeding adults may then appear within the pack or larvae may actually bore out (de Luca, 1977). Clearly this is not a problem of packaging per se, and to avoid such instances corrective measures should have been applied to the commodity prior to packaging. Three species commonly implicated in this way are Acanthoscelides obtectus (Say) a serious pest of bean seeds (Phaseolus vulgaris) in store, Callosobruchus maculatus (F.) the cowpea weevil, and C. chinensis (L.) the adzuki bean weevil. All are about 3±4 mm in size and can increase at a rate of 25 times a month under optimal conditions. The developmental range extends



111

Fig. 4.2 Some beetle pests of stored food pictured on grains of wheat: (a) Lesser grain borer, Rhyzopertha dominica; (b) Saw-toothed grain beetle, Oryzaephilus surinamensis; (c) Rust-red flour beetle, Tribolium castaneum.

from about 15 to 33 ëC for A. obtectus, 20±37 ëC for C. maculatus and 17±37 ëC for C. chinensis (Arbogast, 1991). Cleridae Necrobia spp. Two species of this genus, Necrobia ruficollis (F.), the red shouldered ham beetle, and N. rufipes (Degeer) the copra beetle or red-legged ham beetle, occur as pests of animal products such as dried meat and fish products, milk powder, cheese, hides and bone meal. N. rufipes is also frequently encountered on poorly-stored vegetable products such as copra, dried fruit, cocoa, rice bran, palm nuts and cassava. For optimal development the beetles require high humidity and warm conditions and the best method to avoid their establishment is to keep vulnerable products below 20 ëC and 50% r.h. Temperatures between 30 and 34 ëC give the maximal rate of population increase of 25 times in a 4week period (Howe, 1965), but Necrobia spp. are highly predatory and cannibalism can slow population growth. Mature larvae move out of the food medium and produce a characteristic strong white protective sheath for pupation. It is often at this point that their presence is noted. Cucujidae Cryptolestes ferrugineus (Stephens) The rust-red grain beetle is a common pest of stored grain, but is seen less often than may be expected because, quite apart from its small size (2.5 mm including long antennae), it will move away from areas of disturbance and, as its generic Latin name suggests, hide. This refuge-seeking behaviour of adults is promoted by low temperature, food shortage, and high population density, and is strongest


112


in females and younger individuals (Cox et al., 1990). A highly adaptable species, in addition to grain C. ferrugineus has been recorded on flour, meals, oilseeds, dried fruit, cassava root and other dried vegetable materials. It is tolerant of low humidity and is cold hardy in the long-lived adult stage. Both adults and larvae cause damage. Above 25 ëC flight is initiated, enabling the rapid dispersal of populations (Cox and Dolder, 1995). The eggs of Cryptolestes spp. are elongate (over three times as long as wide, length c. 0.6 mm) and up to 400 may be laid by one female. For optimal development, temperatures of 32± 35 ëC at moderate r.h. can permit population growths of up to 60-fold in a 4week period (Howe, 1965). Other Cryptolestes species responsible for disrupting the trading of food commodities are the flat grain beetles C. pusillus (Schoenherr) and C. capensis (Waltl), and C. turcicus (Grouvelle), the Turkish grain beetle. The latter two species are often found in flour mills, C. turcicus throughout temperate Eurasia and America and C. capensis, like C. ferrugineus tolerant of low humidity, from Europe and North Africa (Arbogast, 1991). Cathartus quadricollis (Guenin-Meneville) The square-necked grain beetle is another cosmopolitan cucujid species, common in the USA. It attacks maize in the field and in store, but also wheat, rice and dried fruits. The developmental period varies according to the food medium and high humidity is preferred, but a generation can be completed in about six weeks at 30 ëC, 70% r.h. (Yoshida, 1976). Curculionidae Sitophilus spp. (grain weevils) Three species, the granary weevil, Sitophilus granarius (L.), the rice weevil S. oryzae (L.) and the maize weevil, S. zeamais Motschulsky, rank among the most serious pests of cereal grains in the world. Weevils develop inside the grain, females digging a tunnel into the grain with the chewing mouthparts at the end of the elongated snout before laying an egg and cementing over the opening with a gelatinous plug that rapidly hardens. On completing development inside the grain, the mature beetle chews through the grain shell to mate and start the next generation, this often being the first sign of infestation. Fortunately the group are confined to the harvested crop before processing, and do not attack packaged products, although production line hygiene failures may occasionally result in their discovery in packaged products such as popcorn. Dermestidae Dermestes lardarius L., D. maculatus Degeer These moderate sized (8±10 mm) beetles infest dried meat, bacon, sausages, fish and other animal products such as sheepskins, hides, furs, feathers, bone meal or cheese. Both larvae and adults feed on the product. Dried plant materials may also be attacked but females require a high protein diet for optimal and continued egg production. Eggs of D. lardarius, up to 80 per female, are laid



113

singly on food at the rate of one every one or two days, and larvae pass through a variable number of instars to reach maturity, within 7 weeks after oviposition at 25±30 ëC (Coombs, 1978). Development can be completed between 15 and 32.5 ëC and adults may live for six months. Females of D. maculatus tend to lay eggs in small clusters and can produce over 300 eggs during their lifetime when both food and water are available (Arbogast, 1991). The developmental range of D. maculatus is slightly higher than that of D. lardarius, egg production and development proceeding up to 35 ëC. If a suitable material is present, Dermestes larvae often bore tunnels in which to pupate, a habit which can cause a great deal of damage to materials on which they do not feed (Hinton, 1945). Four other species are widespread, D. ater De Geer, D. frischii Kugelann, D. haemorrhoidalis Kuster, and D. peruvianus Laporte de Castelnau. Trogoderma granarium (Everts) The khapra beetle is one of the most destructive pests of whole grain and cereal products in warmer areas of the world, particularly North Africa and the Indian subcontinent. Many ships harbour endemic populations, larvae remaining concealed behind paint or rust scale, and because of its notoriety the species is on the quarantine pest list of many nations. The adult stage, 2±3 mm in length, is relatively short-lived but larvae can survive periods of adversity for some years (Burges, 1962). Poor food, high population density or lowered temperature can trigger a developmental arrest during which larvae may feed intermittently and even undergo moults to reduce as well as increase size while remaining for the most part in cracks or crevices (Beck, 1971; Karnavar, 1984). The arrest, variously described as a diapause or quiescence, may be terminated by a substantial temperature rise or the renewal of food (Nair and Desai, 1973). Under high temperature conditions development can be very rapid on foodstuffs, less than 30 days from egg to adult at 35 ëC, and 50±70% r.h. Very low r.h. can be tolerated but for populations to increase a minimum temperature of 22.5 ëC is required at 70% r.h. (Burges, 2008). Many other species of Trogoderma have been recorded on stored products. The warehouse beetle T. variabile Ballion is an important pest in the USA and Australia while T. inclusum Le Conte, T. angustum (Solier), T. anthrenoides (Sharp) and T. glabrum Herbst have caused problems in Europe and elsewhere. Ptinidae Ptinids are collectively known as spider beetles because the narrow waist between the abdomen and thorax gives a spider-like appearance to the 3±4 mm adults. Spider beetles are scavengers associated with long-term infestations of granaries, flour mills, biscuit manufacturing plants and other related food processing premises. The golden spider beetle Niptus hololeucus (Faldermann) may often occur in undisturbed areas in domestic stores and larders. Other species commonly encountered are the white marked spider beetle Ptinus fur (L.), the Australian spider beetle (though actually more common in Europe) P.


114


tectus Boieldieu, the black spider beetle Mezium affine Boieldieu, and the shiny spider beetle Gibbium aequinoctiale Boieldieu. Ptinids are characterised by a long life cycle and an ability to survive in cool as well as warm conditions. Male ptinids are strong flyers but sexual dimorphism occurs, some females being sedentary (Howe, 1991). Silvanidae Oryzaephilus mercator (Fauvel) and Oryzaephilus surinamensis (L.) These are small, slim beetles, about 3 mm in length, strongly striated and medium brown in colour. The merchant grain beetle O. mercator is more commonly found on products such as oilseeds, dried fruit, nuts and cocoa beans than on grain or cereal products. It can complete development between 17 and 38 ëC and is tolerant of low r.h., though at 20 ëC no larvae develop at 30% r.h. or below (Lale et al., 1996). Fecundity is increased at moderate to high r.h. (not above 90%), larvae utilising stored product fungi in their diet. Females lay eggs singly or in small clusters and each may produce up to 300 (Arbogast, 1991). The saw-toothed grain beetle O. surinamensis (Fig. 4.2b) is a serious pest of stored grain in Europe, infesting wheat, oats and barley. It is also one of the commonest pests of cereal products, dried fruits, nuts, and oilseeds such as sunflower. It can develop rapidly, achieving a population increase of 50-fold over a 4-week period at 31±34 ëC (Howe, 1965). The minimum temperature for a population increase is about 20 ëC at 70±80% r.h., lower humidities being tolerated at higher temperatures (Arbogast, 1991). The cold tolerant adults are able to overwinter in cracks and crevices in the fabric of buildings, leaving their refuges in milder conditions to search for food residues at the onset of darkness (Bell, 1991). Tenebrionidae Tribolium confusum J. du Val and Tribolium castaneum (Herbst) A common pest of flour mills, the confused flour beetle was thought to be conspecific with T. castaneum the rust-red flour beetle until 1868, being of similar size (c. 4 mm) and colour. In addition to cereals and cereal products, T. confusum is also known to infest copra, groundnuts, sesame and oilseeds. Fungi and other insect remains can be utilised in the diet. Adults are extremely longlived (1±3 years), and tolerant of cold and very low humidity. The developmental range is 20±38 ëC with an optimum of 30±32 ëC, at which a 60-fold increase in population is achievable on an optimal food in a 28-day period at 70% r.h. (Howe, 1965; Arbogast, 1991). Worldwide the rust-red or red flour beetle T. castaneum (Fig. 4.2c) is perhaps the most frequently intercepted pest of stored products. It is the primary pest of flour mills, maltings and food processing premises, adults and larvae feeding on all cereal products, groundnuts, cacao, spices, dried figs and dates, copra, dried yam, palm kernels, various nuts, oilseeds and cotton seed. Its rapid development and readiness to breed in the laboratory have made it a popular tool in physiological and genetic studies. Like many other tenebrionids, the free ranging



115

larvae and adults are predatory on other species. The life cycle can be completed between 22 and 40 ëC with an optimum of 32±35 ëC at which, on an optimal food at 70% r.h., a population increase of up to 70 times can be achieved over 28 days (Howe, 1965; Arbogast, 1991), the highest rate of increase achieved by any stored product insect. Adults live for 1±2 years, are capable of flight in warmer conditions and are cold hardy. Adults of Tribolium spp. have long been known to produce quinones, which at high population densities tend to trigger dispersion. Trogossitidae Tenebroides mauritanicus (L.) The cadelle beetle is a major pest of grain and cereal products in the USA. It also infests nuts, seeds, and dried fruit and vegetables of various kinds. Both adult and larval stages are voracious feeders, the former being predatory on any other insect stage they encounter, including larvae of their own kind. The adult is a shiny black medium sized (8±10 mm) beetle with an obvious waist dividing the body in two. The life cycle is relatively long, the time for completion under optimal conditions being about 10 weeks, while individuals hatching in late summer generally do not complete development until the following summer. Larvae are notorious for burrowing into structural materials, plastics and fabrics and have caused extensive damage to equipment in flour mills (Mueller, 1998). Adults are long-lived and females may produce as many as 1000 eggs in their lifetime (Arbogast, 1991). 4.2.2 Lepidoptera Many moth species are associated with stored or finished vegetable and animal products, but those in the families Gelecheidae and Tinaeidae attack only cereal grains, decaying food residues, textiles, furs, fabrics or household furnishings and do not occur on packaging. Two families remain for consideration here. Oecophoridae Hofmannophila pseudospretella (Stainton) The brown house moth (length 9±15 mm) occurs widely in buildings in northern Europe and is a pest of premises such as flour mills where food is stored or processed. In domestic houses it infests food residues, textiles, wool, dried decaying organic matter and various seed products at high humidity. The larvae can chew through many fabrics and packaging materials, being equipped with powerful sclerotised mandibles. The life cycle is complex and involves a larval diapause and stages of quiescence rendering it univoltine in nature (Woodroffe, 1951). Development can proceed between 10 and 29 ëC, r.h. at 80% being preferred (Howe, 1965). The related white-shouldered house moth Endrosis sarcitrella (L.) which develops more rapidly also occurs widely in European mill basements and domestic premises.


116


Pyralidae This family contains most of the moth pests of stored products. Damage is caused by the larval stage which features a heavily sclerotised head capsule with biting and chewing mouthparts while, as in other stored-product Lepidoptera, the rest of the elongated body is unsclerotised and of a whitish colour. Corcyra cephalonica (Stainton) The rice moth is a serious pest of mills in hot damp climates but occurs widely on imports of cereals, cereal products, dried fruit, seeds, cocoa and groundnuts and can become established in heated premises anywhere in the world. It is capable of a population increase between 18 and 35 ëC (Howe, 1965; Cox et al., 1981). Under optimum conditions (30 ëC, 80% r.h.) the life cycle can be completed in 4 weeks and a population increase of 50-fold in 28 days can be achieved. Oviposition occurs from dusk onwards, but is inhibited by daylight (Bell, 1981). Eggs are not cold tolerant and will not hatch at 15 ëC or below. On completing their development, larvae spin a tough double-layered cocoon in preparation for pupation. Ephestia cautella (Walker) The tropical warehouse or almond moth is the most frequently intercepted moth pest on imports into the developed world and occurs on a very wide range of products including dried fruit, nuts, cereals and cereal products, dried vegetables, cocoa beans, spices, copra, pulses and carobs. The life cycle can be rapid (50 ëC) or vacuum drying, sometimes assisted by microwaves. When heating is used to assist drying, it may cause irreversible changes to vitamins and loss of volatile flavours and aromas. Many different types of continuous and batch equipment are used, spray dryers and heated drums or belts. Dryers are normally used at high or low temperatures that prevent the growth of, or to kill, microorganisms, and conditions are usually a compromise between preventing quality damage to the product from high temperature processing (e.g., burning or surface desiccation) and minimizing the time to reach the target aW (e.g., time available for chemical and microbiological reactions to occur). An exception to this is the production of potato and vegetable crisps by deep frying (see http://www.hyfoma.com/en/content/ food-branches-processing-manufacturing/fruit-vegetable-potato/potato/crisps/). The addition of glycerol, sorbitol, sugar and salt, is used to produce semimoist or intermediate moisture foods (IMF) which have a soft texture and 15± 30% moisture content with aw levels from 0.65 to 0.90. They are preserved because the humectants or water-binding substances, such as sugar and salt, restrict the amount of water available for microbial growth and during processing some water may also be removed by drying. Typical foods are dried beef, peaches and apricots. There are many IMF pet foods. To formulate IMF products, direct measurements of aw are needed because the effectiveness of combinations of humectants cannot be calculated reliably. The effectiveness of IMF preservation can be increased by the addition of organic acids. (See also http://www.fao.org/docrep/005/Y4358E/y4358e07.htm for information on the preparation of stable intermediate or high moisture fruits.) They receive a pasteurization treatment and have an aw of 0.94±0.98 and low pH 3.0±4.1. They


Processing and food and beverage shelf life

211

contain preservatives (e.g., citric acid and potassium sorbate) and an antibrowning agent (sodium bisulphite) to provide colour stability. The stability of IMF foods relies on packaging providing effective oxygen and water barriers. Clear or opaque fruit syrups (e.g., cassis and peach) are low aw products made from concentrated fruit juice. They are often ambient stable because of a high concentration of sugar and solutes (50±70%) and a low pH (90 ëC) to about 3±5 ëC in 90 minutes or so, but cooling rates depend on interaction between the equipment and the product units or product flow being cooled. Minimum cooling rates should be at least fast enough to prevent the outgrowth of any spores present after heat processing. Blast chillers should be designed to cool without re-contaminating material; often they will receive `naked' product and so their design and hygiene measures should ensure that food contact surfaces (e.g., belts and racks) remain free of contaminants. Freezers reduce or maintain temperatures below the freezing point of food. Freezing rates are determined by the quantity and temperature of the heat transfer medium, its contact with the product and the thermal path within the product. Heat transfer characteristics and the rate of freezing of packaged products will be affected by the insulating effects of packaging. In addition blast freezers may produce low quality product due to slow freezing if they are overloaded or undersized and some product shapes may allow excessive air to by-pass the product surfaces, leading to slow freezing. Freezing reduces the level of free (unfrozen) water present (low aw) and so will slow (or stop) microbial metabolism, but freezing does not kill microorganisms. This reduction in the level of unfrozen water accelerates the rate of some quality change reactions by concentrating soluble components. The best quality is usually achieved by using the fastest possible freezing rates and steady storage temperatures (Redmond et al., 2004). At fast freezing rates freeze concentration and cell damage are minimized and solutes are trapped within the ice crystals. Freezing rate determines whether ice crystals form inside or outside cells, and this will determine the extent of cell damage to cell walls and the proportion of intact cells. Temperature fluctuations mechanically destroy the structure of



227

cellular materials by causing cyclic growth and thawing of ice crystals. Cellular damage causes a loss of texture and the release of intracellular materials, including enzymes and fats, which promote quality loss. Where high air velocities are used for freezing naked product, such as fish fillets, there is a risk of surface dehydration and this may lead to colour loss (freezer burn) and other oxidative changes which, once initiated, continue during storage. Slow freezing and temperature fluctuations during storage will drive sublimation and concentrate reactants and catalysts in the remaining unfrozen water, which dehydrates intact cells and increases reaction rates (Rahman, 1999). It may also change the configuration of proteins and break emulsions. Primary packaging with a minimum headspace can be used to reduce these effects. Some products may be physically damaged by handling in belt freezers, leading to poor appearance. Heat may be removed from products either by convection (air and cryogen cooling) or by conduction (belt freezers or mixer freezers). Equipment for freezing product can use a variety of agents for either direct or indirect freezing. Blast freezers use very cold air as the freezing medium (mechanical refrigeration, air typically ÿ20 to ÿ40 ëC). This air is in direct contact with the product or pack, which is transported by belts in a spiral or straight configuration or as a fluidized bed for small particles (such as broccoli florets). IQF freezers (for freezing individual items like pepper slices or florets) may use liquid nitrogen (ÿ190 ëC) or CO2 as the freezing agent (cryogenic freezing); these systems can produce very rapid freezing and high quality, reducing losses by evaporation during the freezing process. Products may also be contact frozen (indirectly) by plates, belts or other heat exchange surfaces (e.g., scraped surface heat exchangers). Often the choice of freezer will be determined by production volume demands and the type of material to be frozen. Vegetables are often frozen using fluidized bed tunnel freezers; poultry, meat and prepared, packaged foods are frozen on trays or directly on belts using spiral freezers. Solid stainless steel belt and drum freezers can be used for moist, soft moulded or extruded products, such as fish fillets. 6.8.9 Thawing It is usually necessary to thaw frozen ingredients and products, with the exception of ice-cream and frozen confectionary before further processing or consumption. Frozen raw materials are usually thawed before heat processing to ensure consistent heating times and also ensure that unfrozen ingredients are not over-processed. To retain quality, thawing should be carried out as fast as possible, if material being thawed is packaged, for example in cartons, then thawing rates will be reduced. Thawing is most rapid when there is direct contact of the heat transfer medium (air or water) with the product surface, but when naked product is handled special attention should be paid to hygiene and the temperature of the heat transfer medium. Thawing temperatures should not exceed 20 ëC, higher temperature thawing in hot air ovens or water baths


228


(40±60 ëC), may reduce the quality of some materials (e.g., bell peppers) and can lead to texture, loss and possible growth of spoilage microorganisms and pathogens. If materials are to be cut, minced or flaked then their temperature during processing exerts a major effect on the type of particle and the distribution of particle sizes produced; generally the lower the temperature, especially below freezing, the smaller the particle size produced. Thawing can be speeded up by minimizing the thermal path within the product (40 mm or so maximum). More recently commercial scale equipment using microwaves (915 or 2450 MHz) and radio frequency radiation in cabinets or tunnels has become available and this allows deep frozen foods to be tempered (to about ÿ10 ëC from ÿ20 ëC) or defrosted (ÿ2 to ÿ3 ëC) uniformly in a few minutes. If overhigh power levels are used then run-away heating can occur with some parts of the materials being cooked or burnt.

6.9

Filling and packaging

Many types of filler are available and they may often include pack sealing, capping or closing mechanisms. Choice will be determined by the line speed required and the type of fill (e.g., liquid, viscous and with or without particulates) and the type, shape and weight of the pack and its type of closure. Different products may present different problems to filling and closing equipment (e.g., weight or volume variation, the degree of splashing or dripping during filling, product viscosity and particles) and lead to differences in performance. Fillers may be single or multistage and use either vacuum, auger, piston or overflow filling heads in a rotary, multi-head or in-line (intermittent motion) configuration. Control of the quantity filled can be either volumetric or by weight. Some fillers may be linked to closing or sealing machines that close or complete the pack and may add a modified headspace atmosphere containing a mixture of gases such as CO2 or nitrogen to preserve the product. Metal tops or bases may be seamed onto cans or plastic packs and pouches may be induction or heat sealed. Where hot filling is the preferred option, control of minimum temperatures after filling and temperature drop during any stoppages are critical to safety and quality. If product is supplied to filling heads by pipe work, it is necessary to ensure that there are neither unacceptable temperatures, leading to temperatures within the growth range of microorganisms or hold-up of material. Recirculation loops returning product to heated tanks or vessels may be used to prevent this, but their hygiene must be carefully monitored. When packs are filled for in-pack pasteurization or sterilization, consistent dosing plays a major role in ensuring that heating characteristics, weight and headspace are uniform. If pack dimensions or fill quantity are variable, then quality loss may be caused by over- or under-processing.



229

6.9.1 Primary packaging Most food products are sold packaged. Primary packaging is often in direct contact with the product and must ensure that it is protected from contamination with microorganisms. This depends on good sealing, mechanical integrity of the pack (e.g., the absence of breaks or holes in the packaging material and seal imperfections), and on resistance of packs to physical damage, which can lead to contamination with microorganisms. Packaging materials coming into direct contact with food must not contaminate it chemically and there is legislation governing the materials that may be used for food contact packaging (see http:// www.kunststoffverpackungen.de/08_pressemitteilung/news/040616_ReferatBruder.pdf). Key pack properties include: · · · · · · · · ·

shape stability in use consistent dimensions stability after manufacture or forming to allow clean filling chemical and UV resistance and exclusion barrier properties (e.g., oxygen, CO2 and water vapour) reliable interaction with filling and closing equipment (machinability) freedom from pin holes and other defects toughness and rigidity to protect the product food product resistance to prevent staining.

Many products rely on their particular packaging to achieve their expected shelf life and this may be a major factor in the selection of a packaging material, for example if moisture and oxygen barrier properties are required or if aseptic packaging is used with UHT processing (see www.tetrapak.com, http:// www.sig-group.com/site/en/kartonpackung/Kartonpackung.jsp; http:// www.oystar.gasti.de/dogaseptic.html). If foods are intended for a long storage life, then the properties of the packaging material, especially barrier properties, must be retained over the shelf life (e.g., to prevent oxidative changes or fat staining in long shelf life ambient packs). In some cases the type (and cost) of barrier materials can be adjusted depending on the shelf life required (e.g., 6 weeks or 6 months). Whatever packaging material is used, the expected shelf life will be dependent on the integrity of the package seal to maintain the atmosphere and moisture level within the package (beyond any expected gas transmission across the packaging film) for the designed shelf life. Packaging material may also determine the response of a pack to processing; if packs are made of an unsuitable material, have an incorrect headspace volume or weakened seals are processed in retorts or ovens, burst or weakened packs may result. Some products designed for extended chilled storage (e.g., chilled meats) rely on the modified atmosphere surrounding the product forming part of the preservation system (e.g., vacuum, low PO2, high PN2 and CO2; see http:// www.eufic.org/page/en/faqid/what-is-modified-atmosphere-packaging-map/). Packs for microwave or oven heating carry very different temperature and migration risks to packs used for the distribution of chilled products, as they will


230


be subjected to different temperature conditions during use and must be tested for suitability. 6.9.2 Secondary packaging It is essential that the secondary packaging protects the primary pack from damage during palletization, transport and storage. Secondary packaging may also contribute to limiting quality change during storage, for example cardboard sleeves may limit light contact with products in transparent pouches or trays lidded with clear film. Rigid overwrapping may prevent texture change due to particle breakage.

6.10

Novel processes

6.10.1 Ohmic, impedence or inductive heating During ohmic heating, an electric current is passed either continuously, or intermittently, through the foodstuff. In principle, this gives rapid and uniform heating, because unlike conventional in-pack heating (where heat is transferred from the surface into the pack by conduction and/or convection), there is no thermal gradient across the pack (see http://ohioline.osu.edu/fse-fact/0004.html). Usually liquid foods (e.g., liquids including soups, stews, fish, vegetable and fruit particles, up to 2.5 cm across, in liquid and sauces) are pumped through a pipe system past electrodes and volumetric heating occurs because the food has an electrical resistance (Samaranayake et al., 2005). Ohmic heating causes a temperature rise in the product, cooking occurs and microorganisms are killed predictably by heat. In practice, there are difficulties in reliably identifying the coldest or slowest heating spot in the pack and hence the process conditions needed to give a minimum heat treatment can be severe and in multi-component foods, some components may heat faster than others. Because of this uncertainty, extended holding times are used to allow temperature equilibration within the product and therefore the quality benefits of milder processing may be lost (see Fryer and De Alwis, 1989). Additionally quality assurance procedures have been designed to ensure that electrical conductivity of all components is controlled reliably, specific control procedures may be needed (e.g., presoaking) and has to be monitored to measure the ionic content of ingredients. 6.10.2 Pressure-assisted thermal sterilization (PATS) of food This technology combines mild heat with high pressure (138±483 MPa for 5±20 min at 25±70 ëC) to produce commercially sterile low-acid food products. Product is preheated to a specified temperature (50+ ëC) and then is pressurized to raise its temperature to above 121 ëC which is maintained throughout the pressure stage. When pressure is released, the product returns to the preheat temperature. Because of the short cycle times, loss of quality attributable to



231

heating and cooling times is reduced. Research from the US Army Natick Soldier Systems Centre (NSSC) (http://www.natick.army.mil/soldier/media/fact/ food/HPP.htm) has found that the quality of these foods was better than that of traditional retorted foods. The effectiveness of the technology has been validated by NCFST researchers (see http://www.ncfst.iit.edu/pdfdocs/PressRelease PATSLACF.pdf for details of a recent FDA LACF process filing). Very little other work has been done on quality benefits, although the milder process conditions used offer the potential for improved quality. The killing mechanism is thought to involve either pressure-induced germination of spores, which results in 2- to 5-log reductions of resistant bacilli, alternatively the thermal expansion of liquids may cause structural damage to spores (see http:// researchspace.auckland.ac.nz/handle/2292/5294). This mechanism may also damage cellular food materials, but some authors consider it may preserve quality (see http://fshn.illinois.edu/food_processing_forum/presentations/ c1_Bala_astract.pdf). 6.10.3 Pulsed electric fields Pulsed electric field (PEF) processing is another non-thermal method for decontaminating liquid (grape juice: MarselleÂs-Fontanet, 2009 and http:// www.foodscience.csiro.au/pef-technology.htm) and semi-liquid food products (see Altunakar et al., 2007). It uses short pulses of electricity (20±80 kV for >1 microsecond) to inactivate vegetative microorganisms (e.g., > 5-log reduction of Saccharomyces, Lactobacilli, etc.) by creating or enlarging pores in their cell membranes (see Barbosa-CaÂnovas et al., 1999). Because heat is not involved, it is said to cause minimal changes in food quality (see http://ohioline.osu.edu/fsefact/0002.html). It may be used in combination or sequence with mild heating and its effects are changed by the properties of the food (Kristina and RoÈnner, 2001). At present it is predominantly used for acidic liquids and fruit juices. Application is limited to liquids free of gas bubbles, because bubbles allow electric arcing between the electrodes, burning the material being processed and potentially generating unwanted materials. Sanchez-Vega et al. (2009) compared the effect of UHT and PEF treatments on enzyme inactivation and quality of apple juice. They found that UHT treatment was more effective at inactivating enzymes than PEF treatment, but that the sensory quality of PEF processed juice was higher after processing. In a simulated skim milk, very severe PEF treatments had to be used to inactivate a microbial lipase, and the maximum reduction in activity was 62% (Soliva-Fortuny et al., 2006). Results indicate that enzymes are more resistant to PEF than microorganisms and it is not clear whether levels of enzyme inactivation are high enough to ensure high quality during storage. 6.10.4 Microwave and radio frequency processing Microwave and radio frequency heating use electromagnetic radiation (900±2450 MHz; see http://www.worldscibooks.com/etextbook/4763/4763_chap01.pdf) to


232


heat food and water (Ramaswamy and Tang, 2008). This type of heating can be used for thawing, cooking, re-heating and pasteurization. Because heating is volumetric, these processes can take less time to heat solid and semi-solid foods than conventional methods (Fu, 2004). The interaction of the equipment, magnetron, product composition and pack shape and size can dramatically influence the location and temperature of the coldest point and hence the process time needed for a heat treatment. This uncertainty coupled with the complexity of energy absorption and heat transfer makes it difficult to design and specify the process conditions needed for a product and the means of handling process deviations. Many techniques have been tried to improve the uniformity of heating. Microwave sterilization has to be carried out in water-filled retorts, where heating is indirect because most microwave energy is absorbed by the immersion water, but this indirect method of heating is claimed to have the advantage that the energy distribution in the vessel can be tailored to the product and pack distribution to achieve more uniform quality. Although industrial and pilot microwave pasteurization and sterilization systems have been available for many years, with claims for improved quality, commercial systems are not known to be in use. 6.10.5 Irradiation ± pasteurization and sterilization Much information is available about food irradiation by gamma rays, X-rays and high energy electron beams (http://www.iaea.org/programmes/nafa/d5/public/ foodirradiation.pdf), but there is considerable consumer opposition to its use (see http://www.irradiation.info/). UV and electron beams have been used to reduce surface contamination on meat products and ionizing radiation has been used to pasteurize or sterilize pork rolls and chops (Shults et al., 1998). During exposure to ionizing radiation, the food absorbs energy (absorbed dose) and this causes the formation of free (or hydroxyl) radicals, which kill microorganisms, but can also interact with other food molecules and accelerate quality change in fatty materials by initiating rancidity. Hence it is not suitable for use with products containing unsaturated fats, where quality change is also caused by free radical reactions. However, some foods (e.g., poultry, red meat, spices, and fruits and vegetables) have been pasteurized by gamma rays, X-rays and electron beams to increase their storage life (Lund et al., 2005). 6.10.6 Ultra high pressure (UHP) treatment High pressure (HP) or ultra high pressure (UHP) processing involves applying 5000±11 000 bar pressure to a food to decontaminate (pasteurize) it and can be used much like heat. It destroys only vegetative microorganisms including pathogens and spoilage bacteria (He et al., 2002) but it is not effective at killing microbial spores (Patterson, 2005; Margosch et al., 2006). UHP denatures proteins in a manner similarly to heat (Balny and Masson, 1993) and can damage cell membranes (Luscher et al., 2005). It can slow biochemical and



233

enzymatic reactions, by inactivating enzymes (Hendrickx et al., 1998). It can also unfold protein chains, by disrupting hydrophobic and ion-pair bonds, whereas the off-flavours formed by denaturation of proteins by heat result from the formation or destruction of covalent bonds. Foods, such as juices, salsas and other moist foods, can be decontaminated without affecting heat labile vitamins and flavours. Commercial equipment operates at pressures up to about 600 bar and causes about 5-log reduction of vegetative cells: see Avure Technologies ± http://www.avure.com/food/default.asp. Foods are treated by placing them in a liquid medium (usually water) within a thick-walled pressure vessel and compressing the medium. Treatment may be performed either by · heating at constant volume (adiabatic) or · compression at constant temperature (isothermal) conditions. Batch equipment can process products in bulk bags or consumer packs (see http://www.avure.com/archive/documents/Food-products/qfp_215l-eu-may2007.pdf). Semi-continuous equipment can be used for pumpable products followed by aseptic packaging. If food is air-free and contains water, the hydrostatic pressure does not crush the texture, because the water in the food protects it from physical damage. In 2009 applications of UHP cover a range of foods ± ready-to-eat meats, oysters (He et al., 2002), juices (orange, apple, tomato; see Deliza et al., 2005), fruit (melon; Wolbang et al., 2008); and guava (Yen and Lin, 1996) and prepared salads and dips (see http://www.fst.ohio-state.edu/1108feat_preserving foods.pdf). Commercial processes exist for guacamole (see http:// www.avure.com/pdf/Guacamole.pdf). Pressure treatment of non-pasteurized citrus juice results in no loss of vitamin C and extends the shelf life. See http:// www.hpp.vt.edu/references%5CHPPReferencesA.pdf for a list of references on the use of UHP. UHP may also alter the texture of proteins, and milk proteins may be coagulated without acidification or the use of rennet and gels of varying strengths were formed depending on hold time and the rate of ramping up and down of pressure (Pereda et al., 2007). UHP treatments have been used to process fruit and vegetables (Butz et al., 2003), but they accelerate browning in some foods. UHP treatment gave microbiologically stable single-strength tomato juice with improved viscosity and colour properties in comparison with conventional heat-processed juice (Poretta et al., 1995), but there was less enzyme inactivation than with conventional hot-break treatment; and UHP processing gave higher n-hexanal and cis-3-hexenal levels (from FFA oxidation) than heat processing. Pulsed light Pulsed light is a surface irradiation technology which uses very intense and short flashes of light in the UV to NIR range from xenon discharge lamps. This light affects only the surface of the material being treated and so it is only useful for killing surface contaminants or for clear liquids processed as a very thin film.


234


Experimental work has shown it can reduce numbers of vegetative microorganisms (9-log) and bacterial spores (7-log) on smooth, non-porous surfaces, such as packaging materials, but if the surfaces are rougher, as in the case of foods, effectiveness is reduced to 2- to 3-logs (see http://www.fda.gov/Food/ ScienceResearch/ResearchAreas/SafePracticesforFoodProcesses/ucm103058.htm). Pulsed light treatment can be effective in extending the shelf life of a variety of foods, but despite its approval for surface decontamination by the US FDA, pulsed light is not yet commercially used (see http://www.embeddedstar.com/ press/content/2002/9/embedded5062.html and http://www.steribeam.com/), mostly due to the lack of knowledge regarding the critical factors that influence effectiveness and lack of knowledge of the mechanism of inactivation.

6.11

Hygiene

To ensure food quality and stability during storage, the cleaning and hygienic design of equipment need to prevent re-contamination of the main product flow with food debris remaining from previous processing sessions, because these will initiate quality changes or cause quality loss, for example if there are strong flavourings or allergens in a previous batch (Hayes, 1995). Hygienic design should keep material in the main product flow and minimize the range of residence times. Food debris may deteriorate if it remains in equipment (e.g., spoil, oxidize, desiccate or burn) or is retained adjacent to areas of high temperature (e.g., motors, etc.) which will accelerate rates of deterioration. To prevent this, equipment should not have dead ends or low spots to trap food and should be self draining. Construction materials (see http://www.ehedg.org/ uploads/DOC_08_E_2004.pdf) and maintenance should ensure that the food materials are not retained in cracks or corroded areas or absorbed into plastics. Under their anticipated operating conditions, product contact materials must be inert and not absorb either product or any detergents or disinfectants used (see Council Directive 89/109/EEC of 21 December 1988 relating to materials and articles intended to come into contact with foodstuffs). Hence processing and hygiene form an interconnected system that must be managed to ensure quality and limit quality change. If equipment cleaning does not remove residual material there may also be an adverse effect on equipment performance (e.g., a reduction in heat transfer by heat exchangers caused by fouling or ineffective sealing by sealing heads: see http://www.food-info.net/uk/eng/docs/doc8.htm). Cleaning between batches and product sequencing should prevent intermixing of different products and allergen or flavour contamination which could result from a product change-over. Optimum sequences of production should ensure that down time and the chances of contamination are minimized and product wastage is minimized. If successive products are very similar, full cleaning may not be necessary, and rinsing or scraping down between batches may be effective. The effectiveness of cleaning will be reduced if operatives need to prevent the wetting of sensitive parts of machines (e.g., electronic



235

controls), which may be rendered inoperative by water penetration, especially if high-pressure cleaning is used. Hence an integral part of the hygienic design of machines is good access for cleaning and inspection and effective waterproofing of controls and sensors.

6.12

Future trends

Future trends in processing will focus on preserving quality, reducing conversion costs and meeting consumer expectations for fresher, more natural foods. This could lead to the wider use of integrated systems for ingredient sourcing, processing, including non-thermal technologies, and packaging to ensure product quality and stability and meet consumer wishes that food should be fresh and not over-packaged. Extended shelf life will be an issue in the developed world, as there is still distrust by consumers of foods that remain stable for longer periods than expected. However, pressure to reduce travel for shop visits and ensure that logistic chains are fully utilized will counter this pressure, and technology and formulations providing increased stability will gradually become accepted. In the undeveloped world àppropriate' technology will be developed to reduce food wastage through inadequate storage, processing, preservation and packaging. New and existing process technologies will focus on ensuring better nutrient retention, stability with lower levels of preservatives (such as synthetic antioxidants) and the capability to process novel ingredients (see USDA; Center for Food Safety and Applied Nutrition, 2000). Currently, the most widely used technology, thermal processing, provides high levels of microbiological stability, but tends to reduce the quality of foods. Often this occurs because there is poor process control and insufficient understanding of the underlying mechanisms of heating by developers. Freezing and frozen food distribution will continue to be used to retain nutrient quality and overcome seasonality in supply. Wider use will come from preventing unwanted reactions and especially texture changes during storage. The freezing process itself consumes high amounts of energy and more energy efficient techniques will be sought. Chilled foods offer quality and convenience but require highly complex hygienic areas for manufacture and energy intensive logistics systems to cope with their short shelf life, but continue to find favour with consumers. Many of the non-thermal technologies will continue to have a limited scope of application because of their limitations (e.g., their ability to penetrate food materials) and poor knowledge of their mechanisms and possible interactions which may reduce their effectiveness. There is only limited availability of equipment for commercial-scale production and current equipment cannot yet produce commercial volumes and products that command a premium price that would justify investment. There is always a risk of adverse consumer risk± benefit perception of non-conventional processing (e.g., ionizing radiation) which increases commercial risk. UHP treatment is attractive from a product


236


quality point of view, but its application is currently limited to pasteurization processes, limiting it to low pH and chilled products. Similarly PEF can be used for non-thermal inactivation, but in common with microwave and ohmic heating, it requires complex control and fluid handling systems with extensive routine ingredient characterization. More realistically, advances in quality are likely to come from the development of predictive models for the interactions between microorganisms, quality, materials and process conditions during conventional processing. This will lead to the better use of the synergies, including natural ingredients, in hurdle technology. The development of functional foods is a challenge to processing, because current processes and process management systems cannot ensure that any beneficial compounds found in raw materials remain active after processing, or at consumption, because the impact of commercial processing on these compounds is largely unknown. Consumer acceptance of any new process is likely to be increased if consumer benefits are accepted, but this is likely to be a difficult process as there is an information gap between food technologists, marketers and consumers and activists.

6.13

References and further reading

and RICHARDS P. (1999) Numerical simulation of natural convection heating of canned food by computational fluid dynamics. Journal of Food Engineering 41, 1, 55±64. ABDUL GHANI, G., FARID, M.M., CHEN, X.D. and RICHARDS P. (2001) A computational fluid dynamics study on the effect of sterilization temperatures on bacteria deactivation and vitamin destruction. Journal Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering 215, 1, 9±17. ABDUL GHANI, G., FARID, M.M. and CHEN, X.D. (2003) A computational and experimental study of heating and cooling cycles during thermal sterilization of liquid foods in pouches using CFD. Journal Proceedings of the Institution of Mechanical Engineers, Part E: Journal of Process Mechanical Engineering 217, 1, 1±9. ALLEN, C.E. and FOEGEDING, E.A. (1981) Some lipid characteristics and interactions in muscle foods ± a review. Food Technology 35, 253±257. ALTUNAKAR, B., GURRAM, S.R. and BARBOSA-CaÂNOVAS, G.V. (2007) Applications of pulsed electric fields processing of food. In Food Preservation by Pulsed Electric Fields: From Research to Application. Eds Lelieveld, H.L.M. and Notermans, S. Woodhead Publishing, Cambridge. APAIAH, R.A. and BARRINGER, S.A. (2007) Quality loss during tomato paste production versus sauce storage. Journal of Food Processing and Preservation 25, 4, 237± 250. AWUAH, G.B., RAMASWAMY, H.S. and ECONOMIDES, A. (2007) Thermal processing and quality: principles and overview. Chemical Engineering and Processing 46, 6, 584±602. BALNY, C. and MASSON, P. (1993) Effects of high pressure on proteins. Food Rev. Int. 9, 4, 611±628. Â NOVAS, G.V. et al. (1999) PEF inactivation of vegetative cells, spores and BARBOSA-CA ABDUL GHANI, G., FARID, M.M., CHEN, X.D.



237

enzymes in foods. In Preservation of Foods with Pulsed Electric Fields. Eds. Barbosa-CaÂnovas, G.V., GoÂngora-Nieto, M.M., Pothakamury, U.R. and Swanson, B.G. Elsevier, Amsterdam. BERRY, M.R. JR. and BUSH, R.C. (1987) Establishing thermal processes for products with broken-heating curves from data taken at other retort and initial temperatures. Journal of Food Science 52, 4, 958±961. BOROWSKA, J., KOWALSKA, M., CZAPLICKI, S. and ZADERNOWSKI, R. (2003) Effect of hydrothermal processing on carrot carotenoid changes and interactions with dietary fiber. Nahrung 47, 1, 46±48. BOSKOU, D. and ELMADFA, I. (1999) Frying of Food: Oxidation, Nutrient and Non-nutrient Antioxidants, Biologically Active Compounds and High Temperatures. CRC Press, Boca Raton, FL. BOURNE, M.C. (1987) Effect of blanch temperature on kinetics of thermal softening of carrots and green beans. Journal of Food Science 52, 3, 667±668. BRISKEY, E.J. and WISMER-PEDERSEN, J. (1961) Biochemistry of pork muscle structure. I. Rate of anaerobic glycolysis and temperature change versus the apparent structure of muscle tissue. Journal of Food Science 26, 297±305. Â NDEZ GARCIÂA, A., LINDAUER, R., DIETERICH, S., BOGNA Â R, A. and TAUSCHER, B.J. BUTZ, P., FERNA (2003) Influence of ultra high pressure processing on fruit and vegetable products. Journal of Food Engineering 56, 2±3, 233±236. BYRNE, D.V. (1999) Development of a sensory vocabulary for warmed-over flavour: Part ii. In chicken meat. Journal of Sensory Studies 14, 1, 67±78. CACACE, D., PALMIERI, L., PIRONE, G., MASI, G.D.P. and CAVELLA, S. (1994) Biological validation of mathematical modelling of the thermal processing of particulate foods: the influence of heat transfer coefficient determination. Journal of Food Engineering 23, 1, 51±68. CERF, O., DAVEY, K.R. and SADOUDI, A.K. (1996) Thermal inactivation of bacteria ± a new predictive model for the combined effect of three environmental factors: temperature, pH and water activity. Food Research International 29, 3±4, 219± 226. CFA (2006) Best Practice Guidelines for the Production of Chilled Foods, 4th edn. CHANDARANA, D.I., GAVIN A. (III) and WHEATON, F.W. (1990) Particle/fluid interface heat transfer under UHT conditions at low particle/fluid relative velocities. Journal of Food Process Engineering 13, 3, 191±206. CHEAH, P.B. and LEDWARD, D.A. (1997) Catalytic mechanism of lipid oxidation following high pressure treatment in pork fat and meat. Journal of Food Science 62, 6, 1135± 1139. CHONG H.H., SIMSEK, S. and REUHS, B.L. (2009) Analysis of cell-wall pectin from hot and cold break tomato preparations. Food Research International 42, 7, 770±772. CLARK, P.J. (2009) Dry mixing. In Case Studies in Food Engineering. Springer, New York. COPPOLA, S., MAURIELLO, G., APONTE, M., MOSCHETTI, G. and VILLANI, F. (2000) Microbial succession during ripening of Naples-type salami, a southern Italian fermented sausage. Meat Science 56, 4, 321±329. COROLLER, L., LEGUERINEL, I., METTLER, E., SAVY, N. and MAFART, P. (2006) General model based on two mixed Weibull distributions of bacterial resistance, for describing various shapes of inactivation aurves. Applied Environmental Microbiology 72, 10, 6493±6502. COX, P.W. and FRYER, P.J. (2002) Heat transfer to foods: modelling and validation. Journal of Thermal Science 11, 4, 320±330.


238


and BEAN, P.G. (1977) Method for the immobilization of bacterial spores in alginate gel. Lab Pract. 26, 773±775. DATTA, A.K. and TEIXEIRA, A.A. (1988) Numerically predicted transient temperature and viscosity profiles during natural convection heating of canned liquid food. Journal of Food Science 53, 1, 191±195. DE KOCK, S., MINNAAR, A., BERRY, D. and TAYLOR, J.R.N. (1995) The effect of freezing rate on the quality of cellular and non-cellular par-cooked starchy convenience foods. Food Science and Technology 28, 1, 87±95. DELIZA, R., ROSENTHAL, R.A., ABADIO, F.B.D., SILVA, H.O.S. and CASTILLO, C. (2005) Application of high pressure technology in the fruit juice processing: benefits perceived by consumers. Journal of Food Engineering 67, 1±2, 241±246. DEPARTMENT OF HEALTH (1994) Guidelines for the Safe Production of Heat Preserved Foods. HMSO, London. DEVLIEGHER, F., GEERAERD, A.H, VERSYCK, K.J., BERNAERT, H, VAN IMPE, J.F. and DEBEVERE, J. (2000) Shelf life of modified atmosphere packed cooked meat products: addition of Na-lactate as a fourth shelf life determinative factor in a model and product validation. International Journal of Food Microbiology 58, 1±2, 93±106. EVANS, G.C. and RANKEN, M.D. (2007) Fat cooking losses from non-emulsified meat products. International Journal of Food Science and Technology 10, 1, 63±71. FAO (1998) Food Quality and Safety Systems ± A Training Manual on Food Hygiene and the Hazard Analysis and Critical Control Point (HACCP) System. FAO, Rome. FARAG, K.W., DUGGAN, E., MORGAN, D.J., CRONIN, D.A. and LYNG, J.G. (2009) A comparison of conventional and radio frequency defrosting of lean beef meats: effects on water binding characteristics. Meat Science 83, 2, 278±284. FESSAS, D. and SCHIRALDI, A. (2000) Starch gelatinization kinetics in bread dough. DSC investigations on `simulated' baking processes. Journal of Thermal Analysis and Calorimetry 61, 2, 411±423. FRYER, P. and DE ALWIS, A. (1989) Validation of the APV ohmic heating process (APV Baker). Chemistry and Industry 2 October. FU, Y-C. (2004) Fundamentals and industrial applications of microwave and radio frequency. In Food Processing: Principles and Applications. Eds Smith, J.S. and Hui, Y.H. Blackwell Publishing, Oxford, pp. 79±100. GEORGIADIS, M.C., PAPAGEORGIOU, L.G. and MACCHIETTO, S. (2000) Optimal cleaning policies in heat exchanger networks under rapid fouling. Industrial & Engineering Chemistry Research 39, 2, 441±454. GRAY, J.A. and BEMILLER, J.N. (2003) Bread staling: molecular basis and control. Comprehensive Reviews in Food Science and Food Safety 2, 1, 1±21. GUPTA, M.K., GRANT, R. and STEER, R.F. (2004) Critical factors in the selection of an industrial fryer. In Frying Technology and Practices. Eds Gupta, M.K., Warner, K. and White. P.J. AOCS, Urbana, IL. HANSEN, E., LAURIDSEN, L., SKIBSTED, L.H., MOAWAD, R.K. and ANDERSEN, M. L. (2004a) Oxidative stability of frozen pork patties: effect of fluctuating temperature on lipid oxidation. Meat Science 68, 2, 185±191. HANSEN, E., JUNCHER, D., HENCKEL, P., KARLSSON, A., BERTELSEN, G. and SKIBSTED, L.H. (2004b) Oxidative stability of chilled pork chops following long term freeze storage. Meat Science 68, 3, 479±484. HAYES, R. (1995) Design of food processing equipment. In Food Microbiology and Hygiene. Springer, Berlin. HE, H., ADAMS, R.M., FARKAS, D.F. and MORRISSEY, M.T. (2002) Use of high-pressure DALLYN, H., FALLOON, W.C.



239

processing for oyster shucking and shelf-life extension. Journal of Food Science 67, 2, 640±645. HENDRICKX, M.E.G. and KNORR, D. (2002) Ultra High Pressure Treatments of Foods. Kluwer Academic/Plenum Publishers, New York. HENDRICKX, M., LUDIKHUYZE, L., VAN DEN BROECK, I. and WEEMAES, C. (1998) Effects of high pressure on enzymes related to food quality. Trends in Food Science & Technology 9, 5, 197±203. HIDDINK, J., SCHENK, J. and BRUIN, S. (1976) Natural convection heating of liquids in closed containers. Applied Scientific Research 32, 3, 217±237. HUGHES, J.P., JONES, T.E.R. and JAMES, P.W (2003) Numerical simulation and experimental visualization of the isothermal flow of liquids containing a headspace bubble inside a closed cylinder during off-axis rotation. Trans. IChemE 81, Part C, 119±128. INTERNATIONAL COMMISSION ON MICROBIOLOGICAL SPECIFICATIONS OF FOODS (ICMSF)

(1996) Microorganisms in Foods 5, Microbiological Specifications of Food Pathogens. Eds Roberts, T.A., Baird-Parker, A.C. and Tompkin, R.B. Blackie Academic and Professional, New York. KACEM, B., CORNELL, J.A., MARSHALL, M.R., SHIREMAN, R.B. and MATTHEWS, R.F. (1987) Nonenzymatic browning in aseptically packaged orange drinks: effect of ascorbic acid, amino acids and oxygen. Journal of Food Science 52, 6, 1668±1672. KOUTSOUMANIS, K. (2001) Predictive modelling of the shelf life of fish under nonisothermal conditions. Applied and Environmental Microbiology 67, 1821±1829. KOUTSOUMANIS, K. and NYCHAS, G.J.E. (2000) Application of a systematic experimental procedure to develop a microbial model for rapid fish shelf life predictions. International Journal of Food Microbiology 60, 171±184. KRALL, S.M. and MCFEETERS, R.F. (1998) Pectin hydrolysis: effect of temperature, degree of methylation, pH, and calcium on hydrolysis rates. Journal of Agricultural and Food Chemistry 46, 4, 1311±1315. È NNER, U. (2001) Influence of pH, water activity and temperature on the KRISTINA, A. and RO inactivation of Escherichia coli and Saccharomyces cerevisiae by pulsed electric fields. Innovative Food Science & Emerging Technologies 2, 2, 105±112 and erratum in Innovative Food Science & Emerging Technologies 3, 1, 2002, 101±102. KUMAR, B.R., KALAIKANNAN, A. and RADHAKRISHNAN, K.T. (2007) Studies on processing and shelf life of pork nuggets with liquid whey as a replacer for added water. American Journal of Food Technology 2, 1, 38±43. KUPPAN, T. (2000) Heat Exchanger Design Handbook. Marcel Dekker, New York. LABUZA, T.P., HEIDELBAUGH, N.D., SILVER, M. and KAREL, M. (1971) Oxidation at intermediate moisture contents. Journal of the American Oil Chemists' Society 48, 2, 86±90. LALAS, S. (2008) Quality of frying oil. In Advances in Deep-Fat Frying of Foods. Eds Sahin, S. and Servet, G.S. CRC Press, Boca Raton, FL. LANOISELLE, J.-L., CANDAU, Y. and DEBRAY, E. (2006) Predicting internal temperature of canned foods during thermal processing using a linear recursive model. Journal of Food Science 60, 4, 833±835. LEA, A.G.H. (1994) Apple juice. In Production and Packaging of Non-Carbonated Fruit Juices and Fruit Beverages, 2nd edn. Ed. Ashurst, P.R. Springer, Berlin. LEISTNER, L. (2000) Basic aspects of food preservation by hurdle technology. International Journal of Food Microbiology 55, 10, 181±186. LEISTNER, L. and GOULD, G.W. (2003) Hurdle Technologies. Combination Treatments for Food Stability, Safety and Quality. Kluwer Academic/Plenum Publisher, New York.


240


and HEPPELL, N. (2001) Continuous Thermal Processing of Foods: Pasteurization and UHT Processing. Aspen Publications, Gaithersburg, MD. LUKASSE, L.J.S. and POLDERDIJK, J.J. (2003) Predictive modelling of post-harvest quality evolution in perishables, applied to mushrooms. Journal of Food Engineering 59, 2±3, 191±198. LUND, B.M., BAIRD-PARKER, A.C. and GOULD, G.W. (2005) Microbiological Safety and Quality of Food, Volumes 1±2, Springer-Verlag, Berlin. LUSCHER, C., SCHLuÈTER, O. and KNORR, D. (2005) High pressure±low temperature processing of foods: impact on cell membranes, texture, colour and visual appearance of potato tissue. Innovative Food Science & Emerging Technologies 6, 1, 59±71. MAFULEKA, M.M., OTT, D.B., HOSFIELD, G.L. and UBERSAX, M.A. (1991) Dry bean (Phaseolus vulgaris) hardening and the consequences of pectin methylesterase activity in storage. Journal of Food Processing and Preservation 15, 1, 1±18. MANKAD, S., NIXON, K.M. and FRYER, P.J. (1997) Measurements of particle-liquid heat transfer in systems of varied solids fraction. Journal of Food Engineering 31, 1, 9± 33. È NZLE, M.G. (2006) MARGOSCH, D., EHRMANN, M.A., BUCKOW, R., HEINZ, V., VOGEL, R.F. and GA High-pressure-mediated survival of Clostridium botulinum and Bacillus amyloliquefaciens endospores at high temperature. Applied and Environmental Microbiology 72, 5, 3476±3481. Á .R., PUIG, A., PAOLA OLMOS, P., MIÂNGUEZ-SANZ, S. and MARTIÂNMARSELLEÂS-FONTANET, A BELLOSO, O. (2009) Optimising the inactivation of grape juice spoilage organisms by pulse electric fields. International Journal of Food Microbiology 130, 3, 159± 165. MARTINEZ, A., RODRIGO, D., FERNANDEZ, P.S. and OCIO, M.J. (2006) Time±temperature integrators for thermal process evaluation. In Thermal Food Processing: New Technologies and Quality Issues. Ed. Sun, D-W. CRC Press, Boca Raton, FL. MATARAGAS, M., DROSINOS, E.H., VAIDANIS, A. and METAXOPOULOS, I. (2006) Development of a predictive model for spoilage of cooked cured meat products and its validation under constant and dynamic temperature storage conditions. Journal of Food Science 71, 6, 157±167. MENG, Y. and RAMASWAMY, H.S. (2007) Visualization of particle/liquid movements in high viscous fluids during end-over-end rotation. Journal of Food Engineering 80, 2, 545±552. È SER, C. and ADLER-NISSEN, J. (2001) Efficiency of enzymatic and other MEYER, A.S., KO alternative clarification and fining treatments on turbidity and haze in cherry juice. Journal of Agricultural and Food Chemistry 49, 8, 3644±3650. MIELCHE, M.M. (1993) Effects of heat treatment on warmed-over flavour in ground beef during aerobic chill storage. Zeitschrift fuÈr Lebensmittel-Untersuchung und -Forschung 197, 1, 8±13. LEWIS, M.

MIERS, J.C., WAGNER, J.R., NUTTING, M-D., SCHULTZ, W.G., BECKER, R., NEUMANN, H.J., DIETRICH, W.C. and SANSHUCK, D.W. (2006) Field processing of tomatoes: 2. Product quality and composition. Journal of Food Science 36, 3, 400±404. MOLIN, G. (2000) The effect of carbon dioxide on microorganisms. In The Microbiological Safety and Quality of Food. Eds Lund, B.M., Baird-Parker, A.C. and Gould, G.W. Springer, Berlin. È FTU È GIL, N. (2006) The peroxidase enzyme activity of some vegetables and its resistance MU to heat. Journal of the Science of Food and Agriculture 36, 9, 877±888.



241

and ARCHBOLD, D.R. (2006) Biosynthesis of trans-2hexenal in response to wounding in strawberry fruit. Journal of Agricultural and Food Chemistry 54, 4, 1442±1448. PACHECO, C DE P. and DE MASSAGUER, P.R. (2005) Biological validation of tomato pulp continuous heat process. Journal of Food Process Engineering 27, 6, 449±463. PATTERSON, M.F. (2005) Microbiology of pressure-treated foods ± a review. Journal of Applied Microbiology 98, 6, 1400±1409. PEARSON, A.M., LOVE, J.D. and SHORELAND, F.B. (1977) Warmed over flavour in meat poultry and fish. In Advances in Food and Nutrition Research. Ed. Stewart, G.F. Academic Press, New York. PELEG, M. and COLE, M.B. (1998) Reinterpretation of microbial survival curves. Critical Reviews in Food Science and Nutrition 38, 353±380. PEREDA, J., FERRAGUT, V., QUEVEDO, J.M., GUAMIS, B. and TRUJILLO, A.J. (2007) Effects of ultra-high pressure homogenization on microbial and physicochemical shelf life of milk. Journal of Dairy Science 90, 1081±1093. PETERS, R.E. (1998) The broader application of HACCP concepts to food quality in Australia. Food Control 9, 2±3, 83±89. PETROTOS, K.B., QUANTICK, P.C. and PETROPAKIS, H. (1999) Direct osmotic concentration of tomato juice in tubular membrane ± module configuration. II. The effect of using clarified tomato juice on the process performance. Journal of Membrane Science 160, 2, 171±177. PORRETTA, S., BIRZI, A., GHIZZONI, C. and VICINI, E. (1995) Effects of ultra-high hydrostatic pressure treatments on the quality of tomato juice. Food Chemistry 52, 1, 35±41. PRADHAN, A.K., LI, Y., MARCY, J.A., JOHNSON, M.G. and TAMPLIN, M.L. (2007) Pathogen kinetics and heat and mass transfer-based predictive model for Listeria innocua in irregular-shaped poultry products during thermal processing. Journal of Food Protection 70, 3, 607±615. RAHMAN, M.S. (1999) Food preservation by freezing. In Handbook of Food Preservation. Ed. Rahman, M.S. CRC Press, Boca Raton, FL. RAMASWAMY, H. and TANG, J. (2008) Microwave and radio frequency heating. Food Science and Technology International 14, 423±427. REDMOND, G.A., GORMLEY, T.R. and BUTLER, F. (2004) The effect of short- and long-term freeze-chilling on the quality of cooked green beans and carrots. Innovative Food Science & Emerging Technologies 5, 1, 65±72. REID, D.S. (1997) Overview of chemical/physical effects of freezing. In Quality in Frozen Food. Eds Erickson, M.C. and Hung, Y-C. Springer, Berlin. RICHARDSON, P.S. (2004) Modelling and optimizing retort control. In Improving the Thermal Processing of Foods. Woodhead Publishing, Cambridge. RIVA, M., FRANZETTI, L. and GALLI, A. (2001) Effect of storage temperature on microbiological quality and shelf-life of ready to use salads. Annals of Microbiology 51, 1, 39±52. RODRIÂGUEZ-LUNA, G., SEGURAJAUREGUI, J.S., TORRES, J. and BRITO, E. (2006) Heat transfer to non-Newtonian fluid foods under laminar flow conditions in horizontal tubes. Journal of Food Science 52, 4, 975±979. ROSS, E.W. (1998) Mathematical modelling of quality loss. In Food Storage Stability. Eds Taub, I.A. and Singh, R.P. CRC Press, Boca Raton, FL. SALTVEIT, M (1997) Physical and physiological changes in minimally processed fruit and vegetables. In Phytochemistry of Fruit and Vegetables. Eds TomaÂs-BarberaÂn, F.A and Robins, R.J. Oxford University Press, Oxford. MYUNG, K., HAMILTON-KEMP, T.R.


242


and ZHANG, H.Q. (2005) Pulsed ohmic heating: a novel technique for minimization of electrochemical reaction during processing. Journal of Food Science 70, 8, E460±E465. SANCHEZ-VEGA, R., MUJICA-PAZ, H., MARQUEZ-MELENDEZ, R., NGADI, M.O. and ORTEGA-RIVAS, E. (2009) Enzyme inactivation of apple juice treated by ultrapasteurization and pulsed electric fields technology. Journal of Food Processing & Preservation 33, 4, 486±499. SARANG, S., SASTRY, S.K., GAINES, J., YANG, T.C.S. and DUNNE, P. (2007) Product formulation for ohmic heating: blanching as a pre-treatment method to improve uniformity in heating of solid-liquid food mixtures. Journal of Food Science 72, 5, E227±E234. SASTRY, S.K. (1986) Mathematical evaluation of process schedules for asceptic processing of low-acid foods containing discrete particulates. Journal of Food Science 51, 5, 1323±1328. SASTRY, S.K. (2007) A model for heating of liquid-particle mixtures in a continuous flow ohmic heater. Journal of Food Process Engineering 15, 4, 263±278. SASTRY, S.K. and CORNELIUS, W.D. (2002) Aseptic Processing of Foods Containing Solid Particulates. Wiley Interscience, New York. SATYANARAYAN, V.T. and HONIKEL, K-O. (1992) Effect of different cooking methods on warmed-over flavour development in pork. Zeitschrift fuÈr Lebensmitteluntersuchung und -Forschung A 194, 5, 422±425. SHULTS, G.W., COHEN, J.S, HOWKER, J.J. and WIERBICKI, E. (1998) Development of Radappertized Pork Items. US Army Natick Research Development and Engineering Center, MA. Available at: http://handle.dtic.mil/100.2/ADA356509 SIEBERT, K.J. (2006) Haze formation in beverages. LWT ± Food Science and Technology 39, 9, 987±994. SINGH, R.P. and HELDMAN, D.R. (2008) Fluid flow in food processing. In Introduction to Food Engineering, 4th edn. Eds Singh, R.P. and Heldman, D.R. Academic Press, New York. SOARES, N.F.F. and HOTCHKISS, J.H. (1999) Comparative effects of de-aeration and package permeability on ascorbic acid loss in refrigerated orange juice. Packaging Technology and Science 12, 3, 111±118. SOLIVA-FORTUNY, R.S. BENDICHO-PORTA, S. and MARTIN-BELLOSO, O. (2006) Modelling high-intensity pulsed electric field inactivation of a lipase from Pseudomonas fluorescens. Journal of Dairy Science 89, 4096±4104. TATTIYAKUL, J., RAO, M.A. and DATTA, A.K. (2001) Simulation of heat transfer to a canned corn starch dispersion subjected to axial rotation. Chemical Engineering and Processing 40, 391±399. TATTIYAKUL, J., RAO, M.A. and DATTA, A.K. (2002) Heat transfer to a canned corn starch dispersion under intermittent agitation. Journal of Food Engineering 54, 4, 321± 329. TUCKER, G.S. and HOLDSWORTH, S.D. (1991a) Optimization of quality factors for foods thermally processed in rectangular containers. Technical Memorandum No. 627. Campden & Chorleywood Food Research Association, Chipping Norton, Glos., UK. TUCKER, G.S. and HOLDSWORTH, S.D. (1991b). Mathematical modelling of sterilization and cooking process for heat preserved foods ± application of a new heat transfer model. Food & Bioproducts Processing, Trans. IChemE 69, Cl, 5±12. SAMARANAYAKE, C.P., SASTRY, S.K.

US FOOD AND DRUG ADMINISTRATION. CENTER FOR FOOD SAFETY AND APPLIED NUTRITION

(2000) Kinetics of Microbial Inactivation for Alternative Food Processing Technologies. US FDA, Washington, DC.



243

and PIECKOVAÂ, E. (2001) Growth modelling of heat-resistant fungi: the effect of water activity. International Journal of Food Microbiology 63, 1±2, 11±17. VAN BOEKEL, M.A.J.S. (2008) Modelling the food matrix. In Kinetic Modelling of Reactions in Foods. CRC Press, Boca Raton, FL. VAN LINDEN, V., SILA, D.N., DUVETTER, T., DE BAERDEMAEKER, J. and HENDRICKX, M. (2008) Effect of mechanical impact-bruising on polygalacturonase and pectin methylesterase activity and pectic cell wall components in tomato fruit. Postharvest Biology and Technology 47, 1, 98±106. WICHERS, H.J. and BOERIU, C. (2004) Controlling the texture of fruit and vegetables: the role of oxidizing enzymes. In Texture in Food: Solid Foods. Ed. Kilcast, D. Woodhead Publishing, Cambridge. WOLBANG, C.M., FITOSA, J.L. and TREEBY, M.T. (2008) The effect of high pressure processing on nutritional value and quality attributes of Cucumis melo. Innovative Food Science & Emerging Technologies 9, 2, 196±200. YANG, W.H. and RAO, M.A. (1998) Numerical study of parameters affecting broken heating curve. Journal of Food Engineering 37, 1, 43±61. YEN, G.C. and LIN, H-T. (1996) Comparison of high pressure treatment and thermal pasteurization effects on the quality and shelf life of guava pureÂe. International Journal of Food Science and Technology 31, 2, 205±213. ZEUTHEN, P. and BéGH-SéRENSEN, L. (2003) Food preservation techniques. p. 117 Woodhead Publishing, Cambridge. ZHANG, L. and FRYER, P.J. (1995) A model for conduction heat transfer to particles in a hold tube using a moving mesh finite element method. Journal of Food Engineering 26, 2, 193±208. VALIÂK, L.


7 Packaging and food and beverage shelf life G. L. Robertson, University of Queensland and FoodPackagingEnvironment, Australia

Abstract: The role of packaging in extending the shelf life of foods and beverages is outlined and the major food packaging materials (metals, glass, paper, plastics) described. The key package properties related to shelf life are discussed including barrier, surface area:volume ratio and closure integrity. Three examples illustrating how the shelf life of packaged foods and beverages can be predicted are given for situations where the end of shelf life is determined by moisture gain, oxygen gain and microbial growth. Finally, the way in which packaging migrants can lead to end of shelf life is illustrated using as examples epoxidised soy bean oil, antimony, tin and photoinitiators. Key words: food packaging, shelf life, metals, glass, paper, plastics, barrier, surface area:volume ratio, closure integrity, food contact materials.

7.1

Introduction

Packaging is a socio-scientific discipline which ensures delivery of goods to the ultimate consumer of those goods in the best condition appropriate for their use. In today's society, packaging is both pervasive and essential as it protects the foods we buy from the moment they are processed and manufactured through storage and retailing to the final consumer. The importance of packaging hardly needs stressing because in developed countries it is almost impossible to find more than a handful of foods that are sold in an unpackaged state. A primary package is one which is in direct contact with the contained product. It provides the initial and usually the major protective barrier.


Packaging and food and beverage shelf life

245

Examples of primary packages include metal cans, paperboard cartons, glass bottles and plastic pouches. Frequently it is only the primary package which the consumer purchases at retail outlets. A secondary package contains a number of primary packages, e.g. a corrugated case or box. It is the physical distribution carrier and is increasingly being designed so that it can be placed directly onto retail shelves for the display of primary packages (so-called shelf-ready packaging). A tertiary package is made up of a number of secondary packages, the most common example being a stretch-wrapped pallet of corrugated cases. This chapter will confine itself to a consideration of the primary package. Packaging has a major impact on food and beverage shelf life and a recent book was devoted solely to this topic (Robertson, 2010a). This chapter will briefly review the key aspects of packaging and its influence on food and beverage shelf life.

7.2 Role of packaging in extending food and beverage shelf life The package must protect its contents from outside environmental effects, be they water, water vapour, gases, odours, microorganisms, dust, shocks, vibrations, compressive forces, etc., and protect the environment from the product. For many food products, the protection afforded by the package is an essential part of the preservation process. In general, once the integrity of the package is breached, the product is no longer preserved. Knowledge of the kinds of deteriorative reactions that influence food quality is the first step in developing food packaging that will minimise undesirable changes in quality and maximise the development and maintenance of desirable properties. Once the nature of the reactions is understood, knowledge of the factors that control the rates of these reactions is necessary in order to minimise the changes occurring in foods during storage, that is, while packaged (Robertson, 2010b). Deteriorative reactions can be enzymic, chemical, physical (typically as a result of moisture gain or loss), and biological (both microbiological and macrobiological, that is, due to insect pests and rodents). Biochemical, chemical, physical and biological changes occur in foods during processing and storage, and these combine to affect food quality. The most important quality-related changes are as follows (van Boekel, 2008): · Chemical reactions: due mainly to either oxidation or non-enzymatic browning reactions. · Microbial reactions: microorganisms can grow in foods which is desirable in the manufacture of fermented foods such as cheese or beer; otherwise, microbial growth will lead to spoilage and, in the case of pathogens, to unsafe food. · Biochemical reactions: many foods contain endogenous enzymes that can potentially catalyse reactions leading to quality loss (enzymatic browning,


246


lipolysis, proteolysis, etc.). In the case of fermentation, enzymes can be exploited to improve quality. · Physical reactions: many foods are heterogeneous and contain particles. These particles are unstable, and phenomena such as coalescence, aggregation and sedimentation usually lead to quality loss. Also, changes in texture can be considered physical reactions, although the underlying mechanism may be of a chemical nature. The deterioration of packaged foods depends largely on transfers that can occur between the external environment, which is exposed to the hazards of storage and distribution, and the internal environment of the package. For example, there may be transfer of moisture vapour from a humid atmosphere into a dried product, or transfer of an undesirable odour from the external atmosphere into a high-fat product, or development of oxidative rancidity if the package is not an effective oxygen (O2) barrier. Also, flavour compounds can be absorbed by some types of plastic packaging materials (a phenomenon referred to as scalping), and chemical contaminants can migrate from the packaging material into the food (e.g., plasticisers from plastic film). In addition to the ability of packaging materials to protect and preserve foods by minimising or preventing these transfers, packaging materials must also protect the product from mechanical damage and prevent or minimise misuse by consumers (including tampering). Although certain types of deterioration will occur even if there is no transfer of mass (or heat, as some packaging materials can act as efficient insulators against fluctuations in ambient temperatures) between the package and its environment, it is possible in many instances to prolong the shelf life of the food through the use of packaging. Preservation is a means of protecting a product, usually against microbiological deterioration. It is important to understand the differences between biotic deterioration which refers to changes in a food brought about by biological agents such as enzymes (e.g., ripening of fruit, respiration of vegetables) or microorganisms (e.g., moulds, bacteria, and yeasts), and abiotic deterioration which is brought about by physical or chemical agents (e.g., atmospheric O2, moisture, light, odours and temperature). Common insect pests are attracted by food odours and some insect species have the ability to bore through flexible packaging materials (Riudavets et al., 2007). Both biotic and abiotic deterioration can lead to food spoilage, albeit by different methods. Packaging can often (but not always) provide a barrier to, or inhibit the action of, those agents that lead to deterioration. Deteriorative reactions in foods are influenced by two factors: the nature of the food and its surroundings. These factors are referred to as intrinsic and extrinsic parameters. Intrinsic parameters are an inherent part of the food and include water activity (aw), pH, oxidation-reduction potential (Eh), O2 content and product formulation, including the presence of any preservatives or antioxidants. The parameter aw is defined as the ratio of the water vapour



247

Fig. 7.1 Schematic of a typical moisture sorption isotherm showing effect of temperature on water activity and moisture content. ß 2006. From Robertson G. L. Food Packaging Principles and Practice, 2nd edn. Reproduced by permission of Routledge/ Taylor & Francis Group, LLC.

pressure of a food to the vapour pressure of pure water at the same temperature and is an intrinsic property of the food. A plot of the moisture content (expressed as mass of water per unit mass of dry matter) against the corresponding relative humidity (RH) or aw at constant temperature is known as a moisture sorption isotherm. Such plots are very useful in assessing the stability of foods and selecting effective packaging. As aw is temperature dependent, it follows that moisture sorption isotherms must also exhibit temperature dependence (see Fig. 7.1). Thus, at constant moisture content (which is the situation existing in a food packaged in an impermeable package), aw increases with increasing temperature. As rates of deteriorative reactions depend on both aw and temperature, the increase in rate in such situations will typically be greater than that due solely to an increase in temperature. This has important implications for shelf life. Extrinsic factors that control the rates of deteriorative reactions include temperature, RH, gas atmosphere and light; packaging can, to varying degrees, influence the impact of these factors on the rates of deteriorative reactions, depending on the specific packaging material. Temperature is a key factor in determining the rates of deteriorative reactions, and in certain situations the packaging material can affect the temperature of the food. The RH of the


248


ambient environment is important and can influence the aw of the food unless the package provides an impermeable barrier to water vapour. Many flexible plastic packaging materials provide good moisture barriers, but none is completely impermeable, thus limiting the shelf life of low aw foods. The presence and concentration of gases in the environment surrounding the food have a considerable influence on the growth of microorganisms, and the atmosphere inside the package is often modified. Atmospheric O2 generally has a detrimental effect on the nutritive quality of foods, and it is therefore desirable to maintain many types of foods at a low O2 tension, or at least prevent a continuous supply of O2 into the package. With the exception of respiring fruits and vegetables and some flesh foods such as meat, changes in the gas atmosphere of packaged foods depend largely on the nature of the package. Adequately sealed metal and glass containers effectively prevent the interchange of gases between the food and the atmosphere. With flexible packaging, however, the diffusion of gases depends not only on the effectiveness of the closure but also on the permeability of the packaging material, which depends primarily on the physicochemical structure of the barrier. Many deteriorative changes in the nutritional quality of foods are initiated or accelerated by light. The intensity of light and the length of exposure are significant factors in the production of discoloration and flavour defects in packaged foods (Manzocco et al., 2008). Modification of plastic materials can be achieved by incorporation of dyes or application of coatings that absorb light at specific wavelengths. Glass is frequently modified by inclusion of colourproducing agents or by application of coatings. In this way a wide range of light transmission characteristics can be achieved in packages made of the same basic material. There have been many studies demonstrating the effect of packaging materials with different light-screening properties on the rates of deteriorative reactions in foods. Many of the chemical reactions that occur in foods can lead to deterioration in food quality (both nutritional and sensory) or the impairment of food safety. The rates of these chemical reactions are dependent on a variety of factors amenable to control by packaging, including light, O2 concentration, temperature and aw. Therefore, the package can, in certain circumstances, play a major role in controlling these factors, and thus indirectly the rate of the deteriorative chemical reactions. In designing suitable packaging for foods, it is important to first define the indices of failure (IoFs) of the food, that is, the quality attributes that will indicate that the food is no longer acceptable to the consumer (Robertson, 2010c). An IoF could be development of rancid flavours in cereals due to oxidation, loss of red colour (bloom) in chilled beef due to depletion of O2, reduction of carbonation in bottled soft drinks due to permeation of CO2 through the bottle wall, caking of instant coffee due to moisture ingress, development of microbial taint in chilled poultry, or moisture loss in green vegetables resulting in wilting. Once the IoFs for a particular food have been defined, the next step is to



249

attempt to quantify the magnitude of the particular degradation, for example, how much moisture or O2 can react with the food before it becomes unacceptable. The final step is to ascertain which (if any) of the IoFs might be influenced by the packaging material, as packaging cannot prevent all undesirable changes in foods. If, for example, the IoF of a snack food was loss of crispness, then the packaging material could influence this by the extent to which it permitted the ingress of moisture. Different plastic films, for example, have different water vapour transmission rates (WVTRs), and thus the shelf life obtained varies depending on the particular polymer selected. Similar considerations apply to foods for which the IoF is oxidation, as different packaging materials have different O2 transmission rates (OTRs). However, if the IoF of a snack food was non-enzymatic browning, then it is unlikely that different packaging materials would influence the extent of this reaction.

7.3

Major packaging materials

The protection offered by a package is determined by the nature of the packaging material and the format or type of package construction. A wide variety of materials is used in packaging and primary packaging materials consist of one or more of the following materials: metals; glass; paper; and plastic polymers. These are briefly described below; more detailed information is available elsewhere (Robertson, 2006; Yam, 2009). 7.3.1 Metals Four metals are commonly used for the packaging of foods: steel, aluminium, tin and chromium. Tin and steel, and chromium and steel, are used as composite materials in the form of tinplate and electrolytically chromium-coated steel (ECCS), the latter sometimes being referred to as tin-free steel (TFS). Aluminium is used in the form of purified alloys containing small and carefully controlled amounts of various metals. The term tinplate refers to low carbon mild steel sheet varying in thickness from around 0.15±0.5 mm with a coating of tin between 2.8 and 17 gsm (g mÿ2) (0.4±2.5 m thick) applied electrolytically on each surface of the material. After plating, the coating is passivated by electrolytic treatment in sodium dichromate to render the surface more stable and resistant, and then lightly oiled. The combination of tin and steel produces a material which has good strength combined with excellent fabrication qualities as well as a corrosion-resistant surface of bright appearance due to the unique properties of tin. ECCS consists of a duplex coating of metallic chromium and chromium sesquioxide to give a total coating weight of approximately 0.15 gsm. Although the surface of ECCS is more acceptable for protective lacquer coatings, printing inks and varnishes than tinplate, it is less resistant to corrosion and therefore must be lacquered.


250


Aluminium is used to manufacture both metal cans and thin foil, the latter ranging in thickness from 4 to 150 m. Foils thinner than 25 m contain minute pinholes that are permeable to gases and water vapour. In both applications alloying agents including silicon, iron, copper, manganese, magnesium, chromium, zinc and titanium are added to impart strength and improve formability and corrosion resistance. 7.3.2 Glass Glass is an amorphous, inorganic product of fusion that has been cooled to a rigid condition without crystallising. Although rigid, glass is a highly viscous liquid that exists in a vitreous or glassy state. A typical formula for soda-lime glass is silica, SiO2 68±73%; calcia, CaO 10±13%; soda, Na2O 12±15%; alumina, Al2O3 1.5±2%; and iron oxides, FeO 0.05±0.25%. The two main types of glass containers used in food packaging are bottles (which have narrow necks) and jars (which have wide openings); about 75% of all glass food containers are bottles. Today's glass containers are lighter but stronger than their predecessors, and through such developments the glass container has remained competitive and continues to play a significant role in the packaging of food products. The container finish is the glass surrounding the opening in the container that holds the cap or closure and can be broadly classified by size (i.e. diameter) and sealing method (e.g., twist cap, cork, etc.). The type of closure can have a significant impact on the shelf life of foods and beverages packaged in glass. 7.3.3 Paper Paper is the general term for a wide range of matted or felted webs of vegetable fibres (mostly wood) used for the production of paper, paperboard, corrugated board and similar products. When its grammage exceeds 224 gsm, paper is referred to as board. Since it is obtained from plant fibre it is therefore a renewable resource. The properties of an individual paper or paperboard are extremely dependent on the properties of the pulps used (e.g. whether from hardwood or softwood species). These pulps may be used unbleached or bleached to varying degrees by various techniques. Almost all paper is converted by undergoing further treatment after manufacture such as embossing, coating, laminating and forming into special shapes and sizes such as bags and boxes. While paper that has been laminated or coated with plastic polymers can provide a good barrier to gases and water vapour, other paper packaging provides little more than protection from light and minor mechanical damage. Multi-ply boards are produced by the consolidation of one or more web plies into a single sheet of paperboard which is then subsequently converted into rigid boxes, folding cartons, beverage cartons and similar products.



251

7.3.4 Plastics Plastics are organic polymers with the unique characteristic that each molecule is either a long chain or a network of repeating units. The properties of plastics are determined by the chemical and physical nature of the polymers used in their manufacture; the properties of polymers are determined by their molecular structure, molecular weight, degree of crystallinity and chemical composition. These factors in turn affect the density of the polymers and the temperatures at which they undergo physical transitions. Polymer chains can and do align themselves in ordered structures, and the thermodynamics of this ordered state determine such properties as melting point, glass transition temperature, and mechanical and electrical properties. However, it is the chemical nature of the polymer which determines its stability to temperature, light, water and solvents, and hence the degree of protection it will provide to food when used as a packaging material. A wide range of polymers is used in food packaging and the major categories are briefly reviewed below. Polyolefins These form an important class of thermoplastics and include low, linear and high density polyethylenes (LDPE, LLDPE and HDPE) and polypropylene (PP). The polyethylenes have the nominal formula ±(CH2±CH2)n± and are produced with a variable amount of branching, each branch containing a terminal (±CH3) group that prevents close packing of the main polymer chains. LDPE is a tough, flexible, slightly translucent material that provides a good barrier to water vapour but a poor barrier to gases. It is widely used to package foods and is easily heat sealed to itself. LLDPE contains numerous short side chains and has improved chemical and puncture resistance and higher strength than LDPE. HDPE has a much more linear structure than LDPE, is stiffer and harder and provides superior oil and grease resistance. It is used in both film form where it has a white, translucent appearance, and as rigid packs such as bottles. PP is a linear polymer with lower density, higher softening point and better barrier properties than the polyethylenes. In film form it is commonly used in the biaxially-oriented state (BOPP) where it has sparkling clarity; it can also be blow and injection moulded to produce closures and thin-walled containers. Substituted olefins Monomers in which each ethylene group has a single substituent are called vinyl compounds; those with two substituents on the same carbon are called vinylidene compounds. The properties of the resultant polymers depend on the nature of the substituent, molecular weight, crystallinity and degree of orientation. The simplest is polyvinyl chloride (PVC) with a repeating unit of (±(CH2± CHCl)n±). A range of PVC films with widely varying properties can be obtained from the basic polymer. The two main variables are changes in formulation (principally plasticiser content) and orientation. Thin, plasticised PVC film is


252


widely used for the stretch wrapping of trays containing fresh red meat and produce. The relatively high WVTR of PVC prevents condensation on the inside of the film. Oriented films are used for shrink wrapping of produce and fresh meat, but in recent years LLDPE films have increasingly replaced them in many applications. Unplasticised PVC rigid sheet is thermoformed into a wide range of inserts from chocolate boxes to biscuit trays but recently they have been substituted by PET or starch-based biopolymers. Unplasticised PVC bottles have better clarity, oil resistance and barrier properties than those made from HDPE, but in recent years they too have been increasingly replaced by PET for a wide range of foods including fruit juices and edible oils. Polyvinylidene chloride (PVdC) has a repeating unit of (±(CH2±CCl2)n±) and the homopolymer yields a rather stiff film which is unsuitable for packaging purposes. When PVdC is copolymerised with 5±50% (but typically 20%) of vinyl chloride (VC), a soft, tough and relatively impermeable film results. Although the films are copolymers of VdC and VC, they are usually referred to simply as PVdC copolymer and their specific properties vary according to the degree of polymerisation and the relative proportions of the copolymers present. Properties include a unique combination of low permeability to water vapour, gases and odours, as well as greases and alcohols. They also have the ability to withstand hot filling and retorting and so find use as a component in multilayer barrier containers. Ethylene vinyl alcohol (EVOH) copolymers are produced by a controlled hydrolysis of ethylene vinyl acetate (EVA) copolymer, the hydrolytic process transforming the VA group into VOH; there is no VOH involved in the copolymerisation. EVOH copolymers offer not only excellent processability, but also superior barriers to gases, odours, fragrances, solvents, etc., when dry. It is these characteristics that have allowed plastic containers incorporating EVOH barrier layers to replace many glass and metal containers for packaging food. Polystyrene (PS) has the general formula (±(CH2±CHC6H5)n±). Crystal grade PS can be made into film but it is brittle unless the film is biaxially oriented. While a reasonably good barrier to gases, it is a poor barrier to water vapour. The oriented film can be thermoformed into a variety of shapes. To overcome the brittleness of PS, synthetic rubbers (typically 1,3-butadiene isomer CH2=CH± CH=CH2) can be added during polymerisation at levels generally not exceeding 25% w/w for rigid plastics. The chemical properties of this toughened or high impact polystyrene (HIPS) are much the same as those for unmodified or general purpose polystyrene (GPPS); in addition, HIPS is an excellent material for thermoforming into tubs which find wide use in food packaging. Polyesters Poly(ethylene terephthalate) (PET) is a condensation product of typically ethylene glycol (EG) and terephthalic acid and has the general formula (±OOC± C6H5±COOCH2±CH2±)n. The outstanding properties of PET film as a food packaging material are its great tensile strength, excellent chemical resistance, light weight, elasticity and stability over a wide range of temperatures (ÿ60 to



253

220 ëC). PET films are most widely used in the biaxially-oriented, heatstabilised form. To improve the barrier properties of PET, coatings of LDPE and PVdC copolymer have been used but today the use of MXD6 (see next section) and nanoclays are increasingly common for bottles. PET film extrusion-coated with LDPE is very easy to seal and very tough. PET is also used to make òvenable' trays for frozen foods and prepared meals. PET bottles are stretch blow moulded, the stretching or biaxial orientation being necessary to get maximum tensile strength and gas barrier, which in turn enables bottle weights to be low enough to be economical. Polyamides Polyamides (PA) are condensation, generally linear thermoplastics made from monomers with amine and carboxylic acid functional groups resulting in amide (±CONH±) linkages in the main polymer chain that provide mechanical strength and barrier properties; they are commonly referred to as nylons. Nylon 6 films have higher temperature, grease and oil resistance than nylon 11 films. A relatively new polyamide is MXD6 made from m-xylylene diamine and adipic acid; it has better gas barrier properties than nylon 6 at all humidities, and is better than EVOH at 100% RH, due to the existence of the benzene ring in the MXD6 polymer chain. Biaxially-oriented film produced from MXD6 is used in several packaging applications as it has significantly higher gas and water vapour barrier properties, and greater strength and stiffness, than other PAs. MXD6 film is also suitable as a base substrate for laminated film structures for use in lidding and pouches, especially when the film is exposed to retort conditions. Recently MXD6 has found use as a barrier layer in PET bottles. Regenerated cellulose Regenerated cellulose film (RCF) is made from cellulose and is therefore a natural and renewable polymer. It is not a plastic because it does not soften when heated but undergoes thermal decomposition. However, since it competes with synthetic polymers in food packaging applications it is discussed here. It is commonly referred to by the generic term cellophane which is still a registered trade name in some countries. RCF can be regarded as transparent paper and for food packaging applications it is plasticised (typically with ethylene glycol) and coated on one or both sides, the type of coating largely determining the protective properties of the film. The most common coatings are LDPE, PVC and PVdC copolymer.

7.4

Key package properties related to shelf life

7.4.1 Barrier In the selection of suitable packaging materials for a particular food or beverage, the focus is typically on the barrier properties of the packaging material. In


254


contrast to packaging materials made from glass or metal, packages made from thermoplastic polymers are permeable to varying degrees to small molecules such as gases, water vapour, organic vapours and other low molecular weight compounds. A plastic polymer that is a good barrier has a low permeability. The following expression can be derived from Fick's first law (Robertson, 2006): Q

DSp1 ÿ p2 At X

7:1

Here Q is the quantity of gas or vapour permeating through a polymer of thickness X and surface area A in time t under a pressure gradient of p1 on one side and p2 on the other where p1 > p2 . D is the diffusion coefficient and S the solubility coefficient of the permeant; the product DS is referred to as the permeability coefficient and is represented by the symbol P. Thus: QX 7:2 P Atp1 ÿ p2 or: Q P Ap t X

7:3

The term P=X is referred to as the permeance. P is a property of the polymer while P=X is a property of the packaging material. Typical values for the permeability coefficient of commercial food packaging polymers are presented in Table 7.1. The permeability coefficient defined above is independent of thickness, since the thickness is already accounted for in the calculation of P. However, the total amount of protection afforded by unit area of a barrier material approaches zero only asymptotically. Consequently, as polymer thickness X is increased beyond a certain value, it becomes uneconomical to increase it further to obtain lower permeability. For example, to equal the O2 barrier of a 25 m film of a high barrier material such as PVdC copolymer would require 62 500 m of PP or 1250 m of PET or 1250 m of PVC or 250 m of nylon 6. In recent years rigid and flexible polymers have been coated with a variety of compounds to improve their barrier properties including aluminium oxides (Hirvikorpi et al., 2010), oxides of silicon (SiOx) (Deilmann et al., 2008) and amorphous carbon (Boutroy et al., 2006). Nanoclays have also been added to polymers to produce polymer nanocomposites which have improved barrier and mechanical properties (Thellen et al., 2009). Literature data for gas transport coefficients (permeability, diffusion and solubility coefficients) vary generally with parameters that are intrinsic to the polymer such as degree of crystallinity, nature of the polymer, and the thermal and mechanical histories of samples such as orientation. Sorption and diffusion phenomena take place exclusively in the amorphous phase of a semicrystalline polymer and not in its crystalline zones.



255

Table 7.1 Typical permeability coefficients of various food packaging polymers and permeants at 25 ëC P 1011 [mL(STP) cm cmÿ2 sÿ1 (cm Hg)ÿ1] Polymer

O2

Low density polyethylene 15±30 Linear low density 31±36 polyethylene High density polyethylene 6±12 Ethylene vinyl acetate 27±54 (12% VA) Polypropylene 9±16 Poly(vinyl chloride) 0.3±1.2 Polystyrene (high impact) 15±27 Nylon 6 (0% RH) 0.09±0.11 Nylon MXD6 0.01 Poly(ethylene terephthlate) 0.3±0.75 Polycarbonate 10±15 PVdC/PVC copolymer 0.005±0.07 EVOH copolymer (0% RH) 27 mol% ethylene 0.0018 44 mol% ethylene 0.0033

CO2

N2

SO2

H2O (90% RH)

60±160 54

4±12 0.6

200

800

45 170

3.3

57

180

92 1.2±3.0 60±150 0.2±0.3

4.4 0.0093 2.4±7.8 0.015±0.05

0.7±1.2 0.02±0.06 47±66 1.7 0.23±0.48 0.006±0.012 0.024 0.012

7 680 1.2 93 220 12±18,000 22* 7,000 1,300 14

0.0005

* Nylon 11

However, in the published literature it is rare to find many details about a particular plastic packaging material apart from its name, sometimes the resin supplier and maybe if it has been oriented. This makes it virtually impossible to replicate the experimental conditions described in the literature since the range of polymers available is vast. For example, the website www.ides.com contains data sheets on over 80,000 commercial polymers from 694 resin manufacturers. Of course, not all of these polymers are approved or suitable for use in food packaging. The temperature dependence of the permeability coefficient can be represented by an Arrhenius-type relationship: P P0 exp ÿEp =RT

7:4

where Ep is the apparent activation energy for permeation, R is the gas constant and T is the absolute temperature. The permeability coefficient of a specific polymer-permeant system may increase or decrease with increases in temperature depending on the relative effect of temperature on the solubility and diffusion coefficients. Generally, the solubility coefficient increases with increasing temperature for gases and decreases for vapours, and the diffusion coefficient increases with temperature for both gases and vapours. For these reasons, permeability coefficients of different polymers determined at one temperature may not be in the same relative order at other temperatures.


256


The above treatment of steady state diffusion assumes that both D and S are independent of concentration but in practice deviations do occur when there is interaction such as occurs between hydrophilic materials (e.g., EVOH and some of the PAs) and water vapour, or for heterogeneous materials such as coated or laminated films. The property is then defined as the transmission rate (TR) of the material, where: Q 7:5 TR At where Q is the amount of permeant passing through the polymer, A is the area and t is the time. In the case of water and oxygen, the terms WVTR (water vapour transmission rate) and GTR (gas transmission rate) or more specifically OTR or O2TR (oxygen transmission rate) are in common usage. It is critical that the thickness of the film or laminate, the temperature and the partial pressure difference of the gas or water vapour are specified for a particular TR. To convert a measured WVTR or OTR to P, it is necessary to multiply by the thickness of the film and divide by the partial pressure difference used when making the measurements. Example: Calculate the permeability coefficient of an LDPE film to O2 at 25 ëC given that the OTR through a 2:54 10ÿ3 cm thick film with air on one side and inert gas on the other is 1:5 10ÿ6 mL cmÿ2 sÿ1 at 50% RH. O2 partial pressure difference across the film is 0.21 atm 16 cm Hg OTR thickness p 1:5 10ÿ6 mL cmÿ2 sÿ1 2:54 10ÿ3 cm 16 (cm Hg) 2:4 10ÿ10 [mL(STP) cm cmÿ2 sÿ1 cm Hg)ÿ1 ]

P

24 10ÿ11 [mL(STP) cm cmÿ2 sÿ1 cm Hg)ÿ1 ] Therefore: P 1011 24 [mL(STP) cm cmÿ2 sÿ1 cm Hg)ÿ1 ] which is within the range given in Table 7.1.

The OTRs of packaging materials used for modified atmosphere packaging (MAP) of chilled products vary extensively with temperature, RH and material thickness after the thermoforming of packages. Jakobsen et al. (2005) studied two different polymer combinations: APET/LDPE (tray) and PA/LDPE (lid). A temperature reduction of 8 ëC (in the interval 7±23 ëC) caused an OTR reduction of 26±48% depending on material type, degree of thermoforming and RH. An



257

increased OTR was observed as a result of material thinning; however, the increase was not always directly proportional to the degree of material thinning. The changes in OTR observed emphasise the necessity of evaluating the performance of packaging materials under realistic storage conditions, in order to estimate the real O2 content of a chosen package solution. 7.4.2 Surface area:volume ratio The dimensions of the package for a given weight of food can have a large influence on shelf life. While a spherical shape will minimise the surface area of the package (and thus the quantity of moisture or O2 that will permeate through the package wall) it is not a practical shape for commercial use and in practice most packages tend to be rectangular or cylindrical. Table 7.2 gives the surface areas for a range of different shapes which all have the same volume (approximately 450 mL). Compared to the surface area of a sphere, the surface area of a cylinder is 16% greater; a cube 24% greater; a tetrahedron 49% greater; a rectangular shape 58% greater and a thin rectangular shape 246% greater. Extremely thin packages have a much greater surface area:volume ratio and thus require a plastic with better barrier properties to get the same shelf life than if the same quantity of product were packaged in a thicker format. For different quantities of the same product packaged in different sized packages using the same plastic material, the smallest package will have the shortest shelf life as it inevitably has a greater surface area per unit volume. Many food companies still seem unaware of this fact as they continue to launch smaller-sized packages without changing the packaging material and then wonder why the shelf life is shorter for the smaller-sized package.

Table 7.2 Surface areas of different package shapes all having a volume of 450 mL. ß 2010. From Food Packaging & Shelf Life edited by G.L. Robertson. Reproduced by permission of Routledge/Taylor & Francis Group, LLC Shape Sphere Cylinder

Dimensions (cm)

Diameter 9.52 Diameter 7.3 Height 10.8 Cube Sides 7.67 Tetrahedron Sides 15.65 Rectangular pack (1) Height 3 Length 15 Width 10 Thin rectangular Height 1 pack (2) Length 20 Width 22.5

Surface area (m2) (cm2)

Increase (%)

Surface area: volume ratio

285 331

0.0285 0.0331

0 16

0.63 0.73

353 424 450

0.0353 0.0424 0.0450

24 49 58

0.78 0.94 1.0

985

0.0985

246

2.18


258


Example: A food powder having a density of 1 is to be packaged in a plastic film which has a WVTR of 3.1 g mÿ2 dayÿ1 at 25 ëC and 75% RH. The initial moisture content of the powder is 2% and the critical moisture content is 8%. Assuming that each pack will contain 450 g of powder and will be exposed to an external environment at 25 ëC and 75% RH, calculate the shelf life if the shapes of the packs are the same as those listed in Table 7.2. For simplicity, assume that the driving force for WVT remains constant and that there are no moisture gradients in the powder. Weight of dry solids 98% of 450 441.0 g Initial weight of water in powder 2% of 450 9.0 g Final weight of water in powder 441.0/0.92 ÿ 450 479.3 ÿ 450 29.35 g Therefore weight of water permeating into powder is: 29.35 ÿ 9.0 20.35 g For a spherical-shaped package: Quantity of water permeating into package per day is: 0:0285 3:1 0:08835 g dayÿ1 Therefore shelf life s

20:35 0:08835

230 days For the other package shapes: Cylinder: Cube: Tetrahedron: Rectangle 1: Rectangle 2:

s s s s s

198 days 186 days 155 days 146 days 67 days

Thus the shelf life for the same quantity of product packaged in the same film varies by a factor of 3.4 from 67 to 230 days depending on the shape of the package.

7.4.3 Package closures and integrity While the choice of suitable packaging material is critically important to achieve the desired product shelf life, adequate closure or sealing of the package after filling is crucial since the quality of the resultant seal is of paramount importance to the ultimate integrity of the package.



259

For glass containers, a wide range of closures made from either metal or plastics is available. Metal closures are stamped out of sheets of tinplate, ECCS or aluminium and can take four forms: screw caps, crowns, lug caps and spin-on or roll-on closures. Plastic closures are generally compression or injection moulded, the former being based on urea-formaldehyde or phenolicformaldehyde resins, and the latter on a variety of thermoplastic polymers including PS, LDPE, HDPE, PP and PVC. The closure used to retain internal pressures of 200±800 kPa as found in carbonated drinks and beer has traditionally been the crown cork, a crimp-on/ pry-off friction-fitting closure made from tinplate with a fluted skirt and a cork or plastisol liner. A roll-on tamper-evident (ROTE) aluminium or plastic closure is used where critical sealing requirements such as carbonation retention, vacuum retention and hermetic sealing are to be met and is especially popular for soft drinks in large containers where reopening is common. The same closures are applied to glass and plastic bottles. The most common closure designed to contain and protect the contents with no internal pressure (e.g. wine in a bottle) has been the traditional bark cork obtained from the holm oak tree Quercus suber, but it can present problems such as cork dust, leakage and cork taint. In recent years increasing quantities of wine in glass bottles have been sealed using an aluminium roll-on pilfer-proof (ROPP) closure (Brajkovich et al., 2005). The main routes of O2 ingress through different closures into wine bottles is now well established (Lopes et al., 2007). Three types of closures made from metal (either tinplate or ECCS) are used to maintain a vacuum inside a glass container which typically contains heat processed food: a lug-type or twist cap; a press-on twist-off cap held on mainly by vacuum with some assistance from the thread impressions in the gasket wall; and a pry-off (side seal) cap widely used on retorted products and consisting of a cut rubber gasket held in place by being crimped under the curl. Vacuum closures often have a safety button or flip panel consisting of a raised, circular area in the centre of the panel that provides a visual indicator to the consumer that the package is properly sealed. For metal containers, the end is mechanically joined to the cylindrical can body by a double seaming operation. The final quality of the double seam is defined by its length, thickness and the extent of the overlap of the end hook with the body hook. Heat sealable films are considered to be those films which can be bonded together by the normal application of heat. Non-heat-sealable films obviously cannot be sealed this way, but they can often be made heat sealable by applying a heat-sealable coating. In this way the two facing coated surfaces become bonded to each other by application of heat and pressure for the required dwell time. Methods to heat seal plastic films include conduction, impulse, induction, ultrasonic, dielectric and hot-wire (Robertson, 2006). Paper packages are typically sealed by the use of adhesives which can be made from either natural (e.g., starch, protein or rubber latex) or synthetic materials (e.g., PVA). The latter category can be either water- or solvent-borne;


260


hot-melt and cold-seal type adhesives are also widely available. To confer gas and/or water vapour barrier properties, paper is coated with a continuous film of typically LDPE which also makes it possible to heat seal the coated layers.

7.5

Predicting shelf life of packaged foods and beverages

As discussed earlier, the shelf life of a food is controlled by product characteristics including formulation and processing parameters (intrinsic factors); the environment to which the product is exposed during distribution and storage (extrinsic factors); and the properties of the package. Examples of extrinsic factors include temperature, RH, light, total pressure and partial pressure of different gases, and mechanical stresses including transportation and consumer handling. Many of these factors can affect the rates of deteriorative reactions which occur during the shelf life of a product. The properties of the package can have a significant effect on many of the extrinsic factors and thus indirectly on the rates of the deteriorative reactions. Thus the shelf life of a food can be altered by changing its composition and formulation, processing parameters, packaging system, or the environment to which it is exposed. Foods can be classified according to the degree of protection required which focuses attention on the key requirements of the package such as maximum moisture gain or O2 uptake. This enables calculations to be made to determine whether or not a particular packaging material would provide the necessary barrier required to give the desired product shelf life. Examples of such shelf life calculations for moisture and oxygen exchange and microbial growth are given in the following sections. The use of mathematical modelling to design modified atmosphere packaging (MAP) has recently been reviewed (Torrieri et al., 2009); the use of such an approach enables a systematic approach to the design of packaging systems which is still all too rare. 7.5.1 Moisture exchange and shelf life When a food is placed in an environment at a constant temperature and RH, it will eventually come to equilibrium with that environment. The corresponding moisture content at steady state is referred to as the equilibrium moisture content. A plot of the moisture content (expressed as mass of water per unit mass of dry matter) against the corresponding aw at constant temperature gives a moisture sorption isotherm which is very useful in assessing the stability of foods and selecting effective packaging. The expression for the steady state permeation of a gas or vapour through a thermoplastic material presented above (see Eq. 7.3) can be rewritten as: w P 7:6 A p1 ÿ p2 t X where w=t is the rate of gas or vapour transport across the film, the latter term corresponding to Q=t in the integrated form of the expression (Eq. 7.3).



261

The prediction of moisture transfer either to or from a packaged food requires analysis of the above equation given certain boundary conditions. If it is assumed that P=X is constant, that the external environment is at constant temperature and humidity, and that p2, the vapour pressure of the water in the food, follows some simple function of the moisture content, then a simple analysis can be made. However, because external conditions will not remain constant during storage, distribution and retailing of a packaged food, P=X will not be constant. If the food is being sold in markets in temperate climates, then WVTRs determined at 25 ëC/75% RH or 23 ëC/50% RH can be used. In tropical countries analysis can be made using WVTRs determined at 38 ëC/90% RH. A further assumption is that the moisture gradient inside the package is negligible, i.e. the package should be the major resistance to water vapour transport. This is the case whenever P=X is less than about 10 g mÿ2 dayÿ1 (cm Hg)ÿ1, which is the case for most films under high humidity conditions. The internal vapour pressure is not constant but varies with the moisture content of the food at any time. Consequently the rate of gain or loss of moisture is not constant but falls as p gets smaller. Thus to be able to make accurate predictions, some function of p2, the internal vapour pressure, as a function of the moisture content, must be inserted into the equation. Assuming a constant rate results in the product being overprotected. In low and intermediate moisture foods, the internal vapour pressure is determined solely by the moisture sorption isotherm of the food. In the simplest case the isotherm can be treated as a linear function as shown in Fig. 7.2: m b aw c

7:7

where m is the moisture content in g H2O per g solids; aw is the water activity; b is the slope of curve; and c is a constant. The moisture content can be substituted for water gain and, after some mathematical manipulation, the following expression is obtained: m e ÿ m i P A p0 7:8 t ln me ÿ m X Ws b where me is the equilibrium moisture content of the food if exposed to the external package RH; mi is the initial moisture content of the food; m is the moisture content of the food at time t; and p0 is the vapour pressure of pure water at the storage temperature (NOT the actual vapour pressure outside the package). The end of product shelf life is reached when m mc , the critical moisture content, at which time t s , the shelf life. Although this equation has been extensively tested for foods and found to give reasonable predictions of actual weight gain (Labuza and Altunakar, 2007), it is clear from Fig. 7.2 that linearising the isotherm results in the use of a pseudo-equilibrium moisture content m0e that is less than what would be experienced in practice. Therefore the calculated shelf life will be longer than what would be achieved in practice.


262


Fig. 7.2 Schematic of a typical moisture sorption isotherm for breakfast cereal with a superimposed straight line of slope b. Initial (mi), critical (mc) and equilibrium (me) moisture contents are indicated together with the pseudo-quilibrium (m0e ) moisture content used for package shelf life calculations.

Equations such as Eq. 7.8 can be used to calculate the effect on shelf life of various packaging films, different external conditions such as temperature and humidity, changes in the surface area:volume ratio of the package, and variations in the initial moisture content of the product. The following example will illustrate this. Example: A breakfast cereal has an initial moisture content mi of 2.5% and a critical moisture content mc of 8% due to loss of crispness. The equilibrium moisture content me at 25 ëC is 14.8% and the pseudo-equilibrium moisture content m0e obtained by extension of the linear portion of the isotherm is 11%; the slope of the line (b) is 0.147 g H2O/g solids/unit aw (see Fig. 7.2). Calculate the shelf life of the cereal if it is packaged in a 50 m (micron) LDPE film or a 50 m OPP film. The weight of dry cereal in the package is 400 g and the dimensions of the package are 20 cm 30 cm. The packed product is to be stored at 25 ëC and 75% RH. Surface area of the packs are 20 30 600 cm2 0.06 m2 Vapour pressure of pure water at 25 ëC 2.3756 cm Hg Data from a plastic film supplier indicated that WVTRs determined at 25 ëC/75% RH are: 50 m LDPE 8.0 g mÿ2 dayÿ1 50 m OPP 1.35 g mÿ2 dayÿ1



263

These WVTRs must be converted into water vapour permeances P=X by dividing by the driving force for water vapour transfer: Driving force at 25 ëC/75% RH 2:3756 0:75 1:782 cm Hg For LDPE film: P 8:0 g 1 X m2 day 1:782 (cm Hg) 4.489 g H2O mÿ2 dayÿ1 (cm Hg)ÿ1 For OPP film: P 1:35 g 1 2 X m day 1:782 (cm Hg) 0.758 g H2O mÿ2 dayÿ1 (cm Hg)ÿ1 Substituting into Eq. 7.8 for cereal packed in LDPE film: ln

11 ÿ 2:5 0:06 2:3756 4:489 s 11 ÿ 8 400 0:147

7.9

Solving for shelf life s : s ln [2.833]/1.088 10ÿ2 1:0413=1:088 10ÿ2 96 days If the cereal were packed in OPP film instead: s ln [2.833]/1.837 10ÿ2 567 days The shelf life is inversely related to the water vapour permeances of the film; since P=X for LDPE is 5.9 times that for OPP, the shelf life in the latter film is 5.9 times that in the former. If the required shelf life were, say, 300 days then Eq. 7.8 could be recalculated using s 300 and solved for P=X . From this the corresponding WVTR could be calculated and the film supplier requested to supply a film that met this specification at 25 ëC and 75% RH. As noted earlier, the shelf lives calculated above will be longer than what would be achieved in practice because the pseudo-equilibrium moisture content used in the calculations is less than the actual equilibrium moisture content which is the real driving force for water vapour transport. Because of the simplifying assumptions made in the above calculations, the calculated shelf lives should be verified by actual shelf life testing.


264


7.5.2 Oxygen exchange and shelf life It is also possible to calculate the shelf life of a food where the major mode of deterioration is oxidation as demonstrated in the following example.

Example: The six-layer PP/EVOH squeezable GammaÕ bottle became the first barrier food bottle to replace glass when it debuted for Heinz ketchup in 1983. About seven years later, the ketchup bottle moved to coinjection blow moulding with PET instead of PP to enhance both clarity and recyclability. Burgess et al. (1990) determined the O2 permeability coefficient for the GammaÕ bottle, studied the change in the colour of ketchup after long-term exposure to an oxygen-rich environment and established a minimum acceptable redness level for ketchup. The combined results were then used to determine the reaction rate between O2 and ketchup and to predict the shelf life for colour stability of ketchup packaged in the GammaÕ bottle. The OTRs of 12 bottles ranged from 0.31 to 0.98 with an average of 0.56 mL dayÿ1 at 22.7 ëC. The corresponding permeability coefficient P ranged from 4:1 10ÿ11 to 13:9 10ÿ11 with an average of 8:3 10ÿ11 mL cm cmÿ2 secÿ1 (cm Hg)ÿ1. The rate of colour change from red to brown fitted a second-order equation with a rate constant k 2:1 10ÿ6 day cmÿ2. The colour was judged to be unacceptable once 7:1 10ÿ6 mol of O2 per cm2 of bottle surface was absorbed by the ketchup. The surface area of exposed ketchup inside the bottle was 671 cm2. The maximum number of moles of O2 that could be consumed through permeation in order to reach an unacceptable colour is: (671 cm2) (7:1 10ÿ6 mol cmÿ2) 0.00476 mol The ideal gas law was used to convert 0.00476 mol into mL at standard conditions: v nRT=p 0:00476 82:06 298=1 116 mL where R 82:06 mL atm molÿ1 Kÿ1. A bottle of ketchup can therefore consume 116 mL of O2 before changing to an unacceptable colour. As O2 permeates into the bottle, it can both build up inside the bottle and be consumed in a reaction with the ketchup. In the worst case, the ketchup will instantaneously absorb all of the permeating O2, thereby maintaining an O2 partial pressure of zero inside the bottle. The maximum value for the OTR found in the permeation experiment was 0.98 mL dayÿ1. Therefore, the predicted shelf life is at least 116/0.98 118 days at 22.7 ëC.



265

7.5.3 Microbial shelf life The influence of packaging on the microbial shelf life of food has recently been reviewed (Lee, 2010). For packages in which the atmosphere has been modified to exclude O2 either by applying a vacuum or gas flushing to suppress the growth of aerobic microorganisms and minimise oxidative quality changes, packaging materials with a poor gas barrier act to promote microbial growth of aerobes and facultative anaerobes. Even microaerophiles such as Lactobacillus spp. which dominate in vacuum and CO2 packaging of meat products may have enhanced growth rates with higher OTR film or packaging (Tsigarida and Nychas, 2006). The effect of gas permeability on microbial spoilage is shown clearly in Fig. 7.3 in which sous vide packages with a high OTR favoured the growth of aerobic and anaerobic bacteria. The high microbial load consisted of thermoduric Bacillus spp. facultative anaerobes which survived the pasteurisation process and were presumed to have been responsible for the microbial spoilage (Kim et al., 2003). When the microbial lag time was used to estimate shelf life in Fig. 7.3, a package with an OTR three times less extended the shelf life to twice that of the more permeable one. Uncertainty in estimating the microbial shelf life of chilled foods exposed to changing temperature is due to the experimental variability of the model parameters (Almonacid and Torres, 2010).

Fig. 7.3 Effect of gas permeability on evolution of aerobic and anaerobic bacterial counts of sous vide packaged seasoned spinach soup (600 g pouch pack) at 10 ëC containing thermoduric organisms. s: aerobic bacteria with high OTR film package (6.3 mL mÿ2 hÿ1 at OTR partial pressure differential of 1 atm); D: anaerobic bacteria under high O2 permeability film package; l: aerobic bacteria under low OTR film package (OTR 2.3 mL mÿ2 hÿ1); : anaerobic bacteria under low OTR film package. Adapted from Kim et al. (2003).


266


7.6

Packaging migrants and food and beverage shelf life

7.6.1 Introduction There has been a long history of so-called food contact substances migrating from packaging materials into foods (Grob et al., 2006). Not surprisingly, food packaging materials are highly regulated in many countries to ensure consumer safety. Risk assessment of food contact materials (FCMs) in the EU and USA has recently been discussed (Barlow, 2009). Migration is the transfer of molecules originally contained in the packaging material (e.g. plasticiser, residual monomer, antioxidant, catalyst) into the food and possibly to the external environment. Overall migration (OM) is the sum of all (usually unknown) mobile packaging components released per unit area of packaging material under defined test conditions, whereas specific migration (SM) relates to an individual and identifiable compound only. OM therefore is a measure of all compounds transferred into the food whether they are of toxicological interest or not, and will include substances that are physiologically harmless. One of the complications from a legislative viewpoint is that many of the substances that migrate (especially components migrating from can coatings) are neither the starting materials, nor obvious derivatives therefrom, and are, therefore, not covered by existing systems based on positive lists of substances which can be used in food contact materials. The migration of molecules from packaging material into food is a complex phenomenon, and most mathematical treatments of transport processes are derived initially from a consideration of gaseous diffusion as discussed in Section 7.4.1. It is worth noting that diffusion in liquids is approximately one million times slower than in gases, and in solids about one million times slower than in liquids. In the initial stages when up to 60% of the migrant is lost from a polymer to a food, the amount of substance migrating into the food is typically proportional to the square root of time. The extent of migration is strongly controlled by the diffusion and partition coefficients which are influenced by the identity of the packaging material and its chemical structure, molecular weight, polarity and concentration of the migrant in the packaging, the kind of food, any interaction between the food and the packaging, the volume of the packed food and the time and temperature of storage (Ossberger, 2009). Therefore it is possible in both theory and practice that migrants in packaged foods will increase during storage and when they exceed the legal limit, the food will have reached the end of its shelf life and can no longer be legally sold. Space does not permit a detailed discussion of all the possible situations where migration may lead to the premature end of shelf life of packaged foods. Therefore several examples will be presented to demonstrate the diversity of migration of FCMs. 7.6.2 Epoxidised soy bean oil (ESBO) Many types of foods are sold in glass jars with metal lids. To ensure tight closure and fairly easy opening, the lids contain a gasket of PVC with 40±45%



267

plasticiser, usually epoxidised soy bean oil (ESBO). Migration from these lids has repeatedly been an issue of concern. A widely publicised incident in 1998 concerned the migration of ESBO and ESBO derivatives into baby foods packed in glass jars with metal closures, the amounts sometimes exceeding the tolerable daily intake (TDI). Fankhauser-Noti et al. (2005) reported that the migration of ESBO into food products with some free oil far exceeded the SM and OM limit. When the gasket is tightened against the rim of the jar, 60±250 mg (average 165 mg) is in contact with food and on average 70 mg ESBO was in food contact. After exposure to olive oil for four weeks at ambient temperature, all the ESBO was transferred; 70 mg ESBO in a 250 g jar resulted in a concentration of 280 mg kgÿ1; in a 100 g jar it was 700 mg kgÿ1. In oily foods such as garlic, chilli or olives in oil, these predicted concentrations are approached. The estimated exposure of infants aged 6±12 months to ESBO migrating into baby foods can sometimes exceed the TDI by up to 4±5-fold. A SM limit of 30 mg kgÿ1 for ESBO in baby foods has been in effect in the EU since November 2006; for other foods a SM limit of 60 mg kgÿ1 applies. ESBO migration into food containing free oil in contact with the gasket has been reported with a mean of 166 mg kgÿ1 in 86 samples and a maximum of 580 mg kgÿ1 (Fankhauser-Noti et al., 2005). Recently Graubardt et al. (2009) reported further insights into the mechanism of migration from the PVC gaskets of metal closures into oily foods in glass jars. 7.6.3 Antimony (Sb) Antimony trioxide (Sb2O3) is used as a catalyst in 90% of PET manufactured worldwide. As a result most commercial PET material typically contains 190± 300 mg Sb kgÿ1. Antimony trioxide is a suspected carcinogen and is listed as a priority pollutant by the US EPA and the EU. A background level of antimony in pristine ground water in Canada is around 2 ng Lÿ1, but once filled into PET bottles, these levels rise to around 50 ng Lÿ1 in 37 days and 566 ng Lÿ1 after 6 months storage at room temperature (Shotyk et al., 2006). Up to 626 ng Lÿ1 have been found in German brands of water in PET bottles. Antimony residues in ready-to-eat meals heated in PET trays of up to 38 g Lÿ1 have been reported which is the SM limit in the EU. However, the migrated amounts of Sb relative to the accepted TDI give no cause for toxicological concern. Shotyk and Krachler (2007) determined antimony concentrations in 132 brands of bottled water from 28 countries; two of the brands were at or above the maximum allowable Sb concentration for drinking water in Japan (2 g Lÿ1). All of the bottled waters were found to contain Sb in concentrations well below the guidelines recommended for drinking water by the WHO (20 g Lÿ1), US EPA (6 g Lÿ1), as well as the German Federal Ministry of Environment (5 g Lÿ1). Although the extent of contamination of bottled waters by leaching of Sb from PET increased with duration of storage, the reactivities of the bottles were variable for reasons which are not apparent.


268


Keresztes et al. (2009) determined Sb in 10 brands of Hungarian still and sparkling mineral water stored in PET bottles under various conditions. Generally, the Sb concentration of still mineral water was lower than that of sparkling; under certain extreme light and temperature storage conditions, the Sb concentration of some samples exceeded 2 g Lÿ1. The extent of Sb leaching from the PET of different brands of mineral water differed by one order of magnitude in experiments conducted under the same conditions. However, it is not only Sb in bottled water that has received attention. Recently Krachler and Shotyk (2009) reported levels of 23 elements in 132 brands of bottled water from 28 countries; trace metal levels of most bottled waters were below guideline levels currently considered harmful for human health. 7.6.4 Tin The chemical structure which gives metals their valuable practical properties is also responsible for their main weakness: susceptibility to corrosion, the chemical reaction between a metal and its environment. All metals are affected to a greater or lesser extent. Foods and beverages are extremely complex chemical systems covering a wide range of pH and buffering properties, as well as a variable content of corrosion inhibitors or accelerators. The most important corrosion accelerators in foods include O2, anthocyanins, nitrates, sulfur compounds and trimethylamines. While high concentrations of tin in food may cause stomach upsets in some individuals, this is unlikely to be the case where tin concentrations remain below the legal limit of 200 mg kgÿ1 (100 mg kgÿ1 in canned beverages and 50 mg kgÿ1 in canned baby foods). Grassino et al. (2009) reported maximum values of tin in cans of tomato pureÂe up to 301 mg kgÿ1 after 180 days at so-called elevated storage temperatures (36 ëC) which in countries near the equator is the ambient temperature. Based on the legal limit for tin, the shelf life of these canned foods would be less than five months. 7.6.5 Photoinitators Printing inks are incredibly complex materials and their detailed composition a closely-guarded trade secret. Over the past 20 years there has been a move away from solvent-based inks towards those that are cured by UV radiation or (less commonly) electron beams (EB). Photoinitiators are components widely used in UV-cured inks for printing food packaging and a number of food contamination incidents resulting from migration of photoinitiators into food have occurred. Johns et al. (2000) studied the migration of ink components from cartonboard to food during frozen storage and observed that under low temperature conditions (ÿ20 ëC) the migration of benzophenone (a widely-used photoinitiator) occurred even when there was no direct contact between the packaging and the food. Isopropylthioxanthone (ITX) is another photoinitiator used in UV-cured



269

offset printing inks; it is not prohibited for use in food packaging by the EU; it is also not listed on the WHO's prohibited list. In 2005 NestleÂ undertook a recall of over 30 million cartons in four European countries of UHT baby milk packaged in multilayer laminate Tetra Brik Aseptic cartons following the discovery by Italian food safety authorities of the presence of ITX at levels ranging from 120 to 305 mg Lÿ1 for baby milk and from 74± 445 mg Lÿ1 in milk for babies aged 12 months and over; ITX was found at 600 mg Lÿ1 in a single sample of flavoured milk tested. The European Food Safety Authority (EFSA) later said that the presence of the chemical in packaged foods does not pose a health risk. Rothenbacher et al. (2007) detected ITX in 36 of 137 packages (26%) not limited to multilayer laminate cartons (e.g. it was found in sausage skins and plastic cups), and significant migration occurred in 75% of the packaging materials that tested positive. The levels of ITX ranged up to 357 mg Lÿ1 in orange juice. The authors concluded that industry should utilise other, lessmigrating photoinitiators, and that the implementation of legislative standards for GMP with a positive list for printing inks and maximum migration limits, especially for substances with incomplete toxicological assessment, is essential.

7.7

Future trends

Numerous factors including political and legislative changes as well as global demand for foods and the likely move towards a low carbon economy will influence the development and success of new packaging materials. However, there is no doubt that the use of existing food packaging materials will increase, but as part of the drive towards more sustainable packaging, food manufacturers will reduce the amount of packaging per unit of food. This will have obvious implications for shelf life. Major supermarket chains are already leading the way by encouraging their suppliers to use less packaging material and this trend is likely to accelerate. The use of bio-based materials which generally have poorer barrier properties will create challenges for food manufacturers who need to meet target shelf lives for their products to ensure orderly distribution and marketing. Finally, there will be greater legislative knowledge and oversight about potential migrants from food contact materials that will lead inevitably to some chemicals being banned or restricted in the manufacture of packaging materials.

7.8


(2006). Food Packaging Principles & Practice, 2nd edn. Boca Raton, FL: CRC Press. ROBERTSON G.L. (ed.) (2010). Food Packaging and Shelf Life. Boca Raton, FL: CRC Press. YAM K.L. (ed). (2009). The Wiley Encyclopedia of Packaging Technology, 3rd edn. New York: John Wiley & Sons Inc. ROBERTSON G.L.


270


7.9

References

(2010). Uncertainty of microbial shelf-life estimations for refrigerated foods due to the experimental variability of the model parameters. Journal of Food Process Engineering 33: 66±84. BARLOW S.M. (2009). Risk assessment of food-contact materials: past experience and future challenges. Food Additives and Contaminants Part A 26: 1526±1533. ALMONACID S.F., TORRES J.A.

BOUTROY N., PERNEL Y., RIUS J.M., AUGER F., VON BARDELEBEN H.J., CANTIN J.L., ABEL F.,

ZEINERT A., CASIRAGHI C., FERRARI A.C., ROBERTSON J. (2006). Hydrogenated amorphous carbon film coating of PET bottles for gas diffusion barriers. Diamond & Related Materials 15: 921±927.

BRAJKOVICH M., TIBBITS N., PERON G., LUND C.M., DYKES S.I., KILMARTIN P.A., NICOLAU L.

(2005). Effect of screwcap and cork closures on SO2 levels and aromas in a Sauvignon Blanc wine. Journal of Agricultural and Food Chemistry 53: 10006± 10011. BURGESS C., BURGESS G., OFOLI R. (1990). Oxygen diffusion rate through the Gamma bottle and associated kinetic effects on tomato ketchup. Packaging Technology & Science 3: 233±239. DEILMANN M., THEIû S., AWAKOWICZ P. (2008). Pulsed microwave plasma polymerization of silicon oxide films: application of efficient permeation barriers on polyethylene terephthalate. Surface & Coatings Technology 202: 1911±1917. FANKHAUSER-NOTI A., FISELIER K., BIEDERMANN S., BIEDERMANN M., GROB K., ARMELLINI F.

(2005). Epoxidized soy bean oil (ESBO) migrating from the gaskets of lids into food packed in glass jars. European Food Research & Technology 221: 416±422. GRASSINO A.N., GRABARIC Z., PEZZANI A., SQUITIERI G., FASANARO G., IMPEMBO M. (2009). Corrosion behaviour of tinplate cans in contact with tomato pureÂe and protective (inhibiting) substances. Food Additives and Contaminants 26: 1488±1494. GRAUBARDT N., BIEDERMANN M., FISELIER K., BOLZONI L., CAVALIERI C., GROB K. (2009). Further insights into the mechanism of migration from the PVC gaskets of metal closures into oily foods in glass jars. Food Additives and Contaminants Part A 26: 1217±1225. GROB K., BIEDERMANN M., SCHERBAUM E., ROTH M., RIEGER K. (2006). Food contamination with organic materials in perspective: packaging materials as the largest and least controlled source? A view focusing on the European situation. Critical Reviews in Food Science & Nutrition 46: 529±535. È HA È -NISSI M., MUSTONEN T., IISKOLA E., KARPPINEN M. (2010). Atomic layer HIRVIKORPI T., VA deposited aluminium oxide barrier coatings for packaging materials. Thin Solid Films 518: 2654±2658. JAKOBSEN M., JESPERSEN L., JUNCHER D., MIQUEL BECKER E., RISBO J. (2005). Oxygen and light barrier properties of packaging materials used for modified atmosphere packaging: evaluation of performance under realistic storage conditions. Packaging Technology & Science 18: 265±272. JOHNS S.M., JICKELLS S.M., READ W.A., CASTLE L. (2000). Studies on functional barriers to migration. 3. Migration of benzophenone and model ink components from cartonboard to food during frozen storage and microwave heating. Packaging Technology & Science 13: 99±104. Â R E., MIHUCZ V.G., VIRAÂG I., MAJDIK C., ZA Â RAY G. (2009). Leaching of KERESZTES S., TATA antimony from polyethylene terephthalate (PET) bottles into mineral water. Science of the Total Environment 407: 4731±4735.



271

(2003). Effect of different oxygen permeability packaging films on the quality of sous-vide processed seasoned spinach soup. Food Science and Biotechnology 12: 312±315. KRACHLER M., SHOTYK W. (2009). Trace and ultratrace metals in bottled waters: survey of sources worldwide and comparison with refillable metal bottles. Science of the Total Environment 407: 1089±1096. LABUZA T.P., ALTUNAKAR B. (2007). Diffusion and sorption kinetics of water in foods. In: Water Activity in Foods: Fundamentals and Applications (Ed. by G.V. BarbosaCaÂnovas, A.J. Fontana, S.J. Schmidt and T.P. Labuza). Oxford: Blackwell Publishing, pp. 215±237. LEE D.S. (2010). Packaging and the microbial shelf life of food. In: Food Packaging and Shelf Life (Ed. by G.L. Robertson). Boca Raton, FL: CRC Press, pp. 55±79. LOPES P., SAUCHIER C.D., TEISSEDRE P.-L., GLORIES Y. (2007). Main routes of oxygen ingress through different closures into wine bottles. Journal of Agricultural and Food Chemistry 55: 5167±5170. MANZOCCO L., KRAVINA G., CALLIGARIS S., NICOLI M.C. (2008). Shelf life modelling of photosensitive food: the case of colored beverages. Journal of Agricultural and Food Chemistry 56: 5158±5164. OSSBERGER M. (2009). Migration from food contact materials. In: The Wiley Encyclopedia of Packaging Technology, 3rd edn (Ed. by K.L. Yam). New York: John Wiley & Sons, pp. 765±772. RIUDAVETS J., SALAS I., PONS M.J. (2007). Damage characteristics produced by insect pests in packaging film. Journal of Stored Products Research 43: 564±570. ROBERTSON G.L. (2006). Food Packaging Principles & Practice, 2nd edn. Boca Raton, FL: CRC Press. ROBERTSON G.L. (ed.) (2010a). Food Packaging and Shelf Life. Boca Raton, FL: CRC Press. ROBERTSON G.L. (2010b). Food packaging and shelf life. In: Food Packaging and Shelf Life, (Ed. by G.L. Robertson). Boca Raton, FL: CRC Press, pp. 1±16. ROBERTSON G.L. (2010c). Food quality and indices of failure. In: Food Packaging and Shelf Life, (Ed. by G.L. Robertson). Boca Raton, FL: CRC Press, pp. 17±30. ROTHENBACHER T., BAUMANN M., FUGEL D. (2007). 2-Isopropylthioxanthone (2-ITX) in food and food packaging materials on the German market. Food Additives and Contaminants Part A 24: 438±444. SHOTYK W., KRACHLER M. (2007). Contamination of bottled waters with antimony leaching from polyethylene terephthalate (PET) increases upon storage. Environmental Science & Technology 41: 1560±1563. SHOTYK W., KRACHLER M., CHEN B. (2006). Contamination of Canadian and European bottled waters with antimony from PET containers. Journal of Environmental Monitoring 8: 288±292. THELLEN C., SCHIRMER S., RATTO J.A., FINNIGAN B., SCHMIDT D. (2009). Co-extrusion of multilayer poly(m-xylylene adipimide) nanocomposite films for high oxygen barrier packaging applications. Journal of Membrane Science 340: 45±51. TORRIERI E., MAHAJAN P.V., CAVELLA S., GALLAGHER M.D.S., OLIVIERA F.A.R., MASI P. (2009). Mathematical modelling of modified atmosphere packaging: an engineering approach to design packaging systems for fresh-cut produce. In: Advances in Modeling Agricultural Systems (Ed. by P.J. Papajorgii & P.M. Pardalos). New York: Springer Science + Business Media, pp. 455±482. TSIGARIDA E., NYCHAS G.-J.E. (2006). Effect of high-barrier packaging films with different KIM G.T., PAIK H.D., LEE D.S.


272


oxygen transmission rates on the growth of Lactobacillus sp. on meat fillets. Journal of Food Protection 69: 943±947. VAN BOEKEL M.A.J.S. (2008). Kinetic modeling of food quality: a critical review. Comprehensive Reviews in Food Science and Food Safety 7: 144±158. YAM K.L. (ed). (2009). The Wiley Encyclopedia of Packaging Technology, 3rd edn. New York: John Wiley & Sons.


9 Smart packaging for monitoring and managing food and beverage shelf life P. S. Taoukis, National Technical University of Athens, Greece

Abstract: Cost effective monitoring of food and beverage products in the cold chain can be realized by time-temperature integrators (TTIs), inexpensive, active `smart labels' based on physicochemical, chemical or biological principles of operation and exhibiting an easily measurable response that integrates the temperature history of the product. Prerequisite for application of TTI is a correlation system to translate TTI response to the quality status of the food at any point of the chain. Basic structural elements of this system are validated kinetic models of TTI response and kinetics of the degradation indices of the food or beverage, such as predictive models of microbial growth. TTIs can serve as temperature monitors and tools for the optimization of stock rotation policies and cold chain management in general. As an example case study, a kinetic model for growth of spoilage bacteria in modified atmosphere packed (MAP) minced beef was used to select appropriate time-temperature integrators (TTIs) in order to monitor the meat product quality during refrigerated distribution. Key words: smart labels, time-temperature integrators, TTI, shelf life, kinetics, Arrhenius, cold chain, FIFO, SMAS.

9.1 Introduction: smart packaging ± time-temperature integrators (TTIs) In addition to the protection required for ensuring the safety and integrity of foods and beverages, current packaging technology aims to provide additional functionality. Smart packaging contributes to shelf life extension and provides valuable information about the quality and safety status of food products for better management of the food chain, reduction of food waste, and increased


304


protection of the consumer. The `smartness' of packaging refers to its ability to communicate information about the requirements known to ensure product quality, such as package integrity (leak indicators) and time-temperature history of the product (time-temperature integrators, `TTIs'). Smart packaging can also give information on product quality directly. For example, freshness indicators provide a direct indication of the quality (Smolander, 2003) by providing a signal that is the result of a reaction between the indicator and specific chemical compounds or metabolites produced by the deteriorative mechanism (chemical or microbial) of the food or beverage. Such direct or indirect indicators of quality or safety of the products are based on the recognition and thorough study of the deteriorative phenomena that define spoilage processes of foods and beverages throughout their intended shelf life. For perishable food products and beverages, temperature is the main parameter that determines post-processing food quality. Shelf life can be shortened considerably if products are not stored and distributed appropriately at controlled temperatures throughout their entire life cycle, from production to consumption. Monitoring temperature is therefore an essential prerequisite for effective shelf life management. A cost-efficient way to monitor and continuously communicate the temperature conditions of individual food and beverage products throughout distribution is time-temperature integrators (TTIs). Based on having available reliable models of the product shelf life and information on the kinetics of a TTI's response, temperature can be monitored, recorded, and translated into its effect on quality, all the way from production to the consumer's table. Implementing a TTIbased system could lead to realistic control of the chill chain, optimization of stock rotation and reduction of waste, and efficient shelf life management. TTIs are inexpensive, active `smart labels' that can easily show measurable, time- and temperature-dependent changes that reflect the full or partial timetemperature history of a food product to which it is attached (Taoukis and Labuza, 1989). TTIs are based on mechanical, chemical, enzymatic or microbiological changes that are irreversible and expressed usually as a response in a visually quantifiable identifier in the form of mechanical deformation, color development, or color movement. The rate of change in the system underlying the TTI is temperature dependent, increasing with higher temperatures, in a manner similar to the deteriorative reactions responsible for food spoilage. Overall, the visible response of the TTIs reflect the cumulative time-temperature history of the food products they accompany. TTIs are an integral part of an interactive intelligent package and can serve as part of an active shelf life signal in conjunction with the ùse-by date' on the label.

9.2 Principles of the application of time-temperature integrators (TTIs) for shelf life monitoring Since the potential for significantly improving quality and shelf life by monitoring and controlling temperature in the food cold chain was realized,


Smart packaging for monitoring and managing food/beverage shelf life

305

reliable, cost-effective temperature history integrating systems have been sought. The first application of a `device' to indicate handling abuse dates from World War II, when the US Army Quartermaster Corps used an ice cube placed inside each case of frozen food. The deformation or disappearance of the cube indicated mishandling (Schoen and Byrne, 1972). The ideal TTI would have the following properties (Taoukis and Labuza, 2003): · It responds with a continuous time-temperature dependent change. · Its response is easily measurable and irreversible. · Its response mimics or can be correlated with the extent of quality deterioration and residual shelf life of the food or beverage. · It is reliable, giving consistent responses when exposed to the same temperature conditions. · It has low cost. · It is flexible, so that different configurations can be adopted for various temperature ranges (e.g., frozen, refrigerated, room temperature) with useful response periods of a few days as well as up to more than a year. · It is small, easily integrated as part of the food or beverage package and compatible with a high speed packaging process. · It has a long shelf life before activation and can be easily activated. · It is unaffected by ambient conditions other than temperature, such as light, RH and air pollutants. · It is resistant to normal mechanical abuses and its response cannot be altered. · It is nontoxic, posing no safety threat in the unlikely situation of product contact. · It is able to convey in a simple and clear way the intended message to its target, be that distribution handlers or inspectors, retail store personnel or consumers. · Its response is both visually understandable and adaptable to measurement by electronic equipment for easier and faster information, storage and subsequent use. For more than three decades, numerous TTI systems have been proposed, of which only a few reached the industrial prototype and even fewer the commercial application stage. The history of TTI development is outlined by Taoukis (2010). Systems that are currently available commercially are the following: · The CheckPointÕ TTI (VITSAB AB, MalmoÈ, Sweden) is an enzymatic system. This TTI is based on a color change caused by a pH decrease which is the result of a controlled enzymatic hydrolysis of a lipid substrate. Different combinations of enzyme-substrate and concentrations can be used to give a variety of response lives and temperature dependencies. Upon activation, the enzyme and substrate are mixed by mechanically breaking a separating barrier inside the TTI. Hydrolysis of the substrate (e.g., tricaproin) causes


306

·

·

·

·

·


acid release (e.g., caproic acid), and the corresponding pH drop induces a color change in a pH indicator from deep green to bright yellow to orange-red (Fig. 9.1). A visual scale of the color change facilitates visual recognition and evaluation of the magnitude and significance of the color change. The continuous color change can also be measured with instrumentation and the results can be used in a shelf life management scheme. The Fresh-CheckÕ TTI (Temptime Corp, NJ, USA) (successor to FreshCheck of Lifelines) is based on a solid state polymerization reaction. The TTI function is based on the property of disubstituted diacetylene crystals (R± C=C±C=C±R) to polymerize through a lattice-controlled solid-state reaction, resulting in a highly colored polymer. The response of the TTI is the color change as measured in terms of a decrease in reflectance. The color of the àctive' centre of the TTI is compared with the reference color of a surrounding ring (Fig. 9.2). Before using the indicators, which are active from the time of their production, the TTIs have to be stored at deep frozen temperatures, where change is very slow. The OnVuTM TTI (Ciba Specialty Chemicals & Freshpoint, SW) is a newly introduced solid state reaction-based TTI. It is based on the inherent reproducibility of reactions in crystal phase. Photosensitive compounds such as benzylpyridines are excited and colored by exposure to low wavelength light. This colored state reverses to its initial colorless state at a temperaturedependant rate (Fig. 9.3). By controlling the type of photochromic compound and the time of light exposure during activation, the length and the temperature sensitivity of the TTI can be set. This TTI can take the form of a photosensitive ink and be very flexible in its application. The (eO)Õ TTI (CRYOLOG, Gentilly, France) is based on a time-temperature dependent pH change that is expressed as color changes using suitable pH indicators. The pH change is caused by controlled microbial growth occurring in the gel of the TTI (Louvet et al., 2005; Ellouze et al., 2008). The parameters of the TTI are adjusted for select microorganisms by appropriate variations in the composition of the gel. The response of the TTI is claimed to mimic microbiological spoilage of the monitored food products, as its response is based on the growth characteristics of similar microorganisms, such as select patented strains of the micoorganisms Carbonbacterium piscicola, Lactobacillus fuchuensis, and Leuconostoc mesenteroides. The pH drop occurs with a color change of the pH indicator from green to red (Fig. 9.4). A visual scale of the color change can facilitate visual recognition and evaluation of the significance of the color change. The continuous color change can also be measured instrumentally and be used in a shelf life management scheme. The TT SensorTM TTI (Avery Dennison Corp., USA) is based on a diffusionreaction concept. A polar compound diffuses between two polymer layers and the change in its concentration causes the color change of a fluorescent indicator from yellow to bright pink (Fig. 9.5). The 3M Monitor MarkÕ (3M Co., St. Paul, MN, USA) is based on diffusion of proprietary polymer materials. A viscoelastic material migrates into a



307

Fig. 9.1 Response scale of enzymatic CheckPointÕ TTI from green at time of application (left) to orange-red (right) indicating end of shelf life.

Fig. 9.2

Fig. 9.3

Polymer-based Fresh-CheckÕ TTI.

Solid state photochromic OnVuTM TTI.

Fig. 9.4 Response scale of Microbial TTI (eO)Õ.

Fig. 9.5

TT SensorTM TTI.


308


Fig. 9.6

Diffusion-based 3M Monitor MarkÕ TTI.

light-reflective porous matrix at a temperature dependent rate. This causes a progressive change in the light transmissivity of the porous matrix and provides a visual response (Fig. 9.6). The TTI is activated by adhesion of the two materials that, before use, can be stored separately for a long period at ambient temperature. TTIs can be used to monitor the temperature exposure of food and beverage products during distribution, from production up to the time they are displayed at the retail level. Attached to product cases or bulk units, they give a measure of the preceding temperature conditions at selected control points. Information from TTIs can be used for continuous, overall monitoring of the distribution system, leading to identification and correction of weak links in the chain. Additionally, it serves as a confirmation of compliance with contractual requirements by the producer and distributor. It can guarantee that a properly handled product was delivered to the retailer, thus disallowing unsubstantiated rejection claims by the latter. The presence of the TTI itself would probably improve handling, serving as an incentive and reminder to the distribution personnel throughout the distribution chain of the importance of proper temperature storage. The same TTIs can be used as shelf life end-point indicators readable by the consumer and attached to individual products. Tests using continuous instrumental readings to define the end-point under constant and variable temperatures showed that such end-points could be reliably and accurately recognized by panellists (Sherlock et al., 1991). However, for a successful application of this kind there is a much stricter requirement that the TTI response matches the behavior of the food. In this way the TTI attached to individually packaged products can serve as active shelf life labeling in conjunction with, open date labeling. The TTI assures the consumer that the product was properly handled and indicates the remaining shelf life. Consumer surveys have shown that consumers can be very receptive to the idea of using TTIs on dairy products along with the date code (Sherlock and Labuza, 1992). Use of TTI can thus also be an effective marketing tool.



309

9.3 Requirements and selection of time-temperature integrators (TTIs) for food and beverage products A number of experimental studies have sought to establish correlations between the response of specific TTIs and quality characteristics of specific products. They tested foods at different temperatures, plotting the response of the TTI vs time and the values of selected quality parameters of the foods before testing the statistical significance of the TTI response correlation to the quality parameters (Taoukis and Labuza, 2003). Such studies offer useful information but do not involve any modeling of the TTI response as a function of time and temperature. They are thus applicable only for the specific foods and the conditions that were used. Extrapolation to other similar foods or quality loss reactions, or even use of the correlation equations for the same foods at other temperatures or for fluctuating conditions, is not accurate. A kinetic modeling approach allows the potential user to develop an application scheme specific to a product and to select the most appropriate TTI without the need for extensive testing of the product and the indicator. This approach emphasizes the importance of reliable shelf life modeling of the food to be monitored. Shelf life models must be obtained with an appropriate selection and measurement of effective quality indices and based on efficient experimental design at isothermal conditions covering the range of interest. The applicability of these models should be further validated at fluctuating, nonisothermal conditions representative of the real conditions in the distribution chain. Similar kinetic models must be developed and validated for the response of the suitable TTI. Such a TTI should have a response rate with a temperature dependence, i.e. activation energy EA1, in the range of the EA of the quality deterioration rate of the food. The total response time of the TTI should be at least as long as the shelf life of the food at a chosen reference temperature. TTI response kinetics should be provided and guaranteed by the TTI manufacturer as specifications of each TTI model they supply. The basic principles of TTI modeling and application for quality monitoring are detailed by Taoukis and Labuza (1989) and Taoukis (2001). The shelf life of a food or beverage product evaluated by the measurement of a characteristic quality index, A, can be expressed as: ÿEA 1 1 t 9:1 ÿ f A kt kref exp R T Tref where f A is the quality function of the food or beverage and k the reaction rate constant; t is an exponential function of inverse absolute temperature, T, given by the Arrhenius expression shown, where kref is the reaction rate constant at a reference temperature Tref, EA is the activation energy of the reaction that controls quality loss and R the universal gas constant. The activation energy of food-related chemical reactions and spoilage or microbial growth usually falls within 30± 120 kJ/mol. The reference temperature used is characteristic of the storage range of the food or beverage, e.g. for chilled products Tref 273 K can be used.


310


Similarly to food or beverage quality, a response function FX can be defined for TTI such that FX kI t, with kI an Arrhenius function of T. The value of the functions, f At at time t, after exposure at a known variable temperature exposure, Tt, can be found by integrating Eq. 9.1. Introducing the term of the effective temperature Teff, which is defined as the constant temperature that results in the same quality value f At , as the variable temperature distribution over the same time period, Eq. 9.1 gives: Z t EA 1 1 t 9:2 kTdt kref exp ÿ ÿ f A R Teff Tref 0 For an indicator exposed to the same temperature distribution, Tt, as the food/beverage product, and corresponding to an effective temperature Teff, the response function FX can be expressed as: Z t EA 1 1 dt exp ÿ I ÿ FX t kIref R T Tref 0 EAI 1 1 t 9:3 ÿ kIref exp ÿ R Teff Tref where X is the measured response of the TTI and kIref and EAI are the TTI Arrhenius parameters. A generalized scheme, illustrated in Fig. 9.7, was used (Taoukis and Labuza, 1989; Giannakourou and Taoukis, 2003) translating TTI response to food shelf life status. Based on the developed algorithm, from the measured response X of the TTI at time t, the value of the response function is calculated, from which by solving Eq. 9.3, the Teff of the exposure is derived. The underlying requirement for the reliable prediction of the effective temperature of the food is that the

Fig. 9.7 Schematic representation of the systematic approach for applying TTI as quality monitors.



311

activation energy of the food and the TTI, EA(food) and EA(TTI) should be similar (EA(food) ÿ EA(TTI) < 25 kJ/mol) (Taoukis, 2001).

9.4 Use of time-temperature integrators (TTIs) for shelf life management and optimization in the cold chain ± case study The information provided by the TTI smart labels, translated to remaining shelf life at any point of the cold chain can be used to manage shelf life by improving distribution control and a stock rotation practices. The approach currently used is the first in first out (FIFO) system according to which products received first and/or with the closest expiration date on the label are shipped, displayed and sold first. This approach aims to establish a `steady state' with all products being sold at the same quality level. The assumption is that all products have gone through uniform handling, thus quality is basically a function of time. The use of the indicators can help establish a system that does not depend on this unrealistic assumption. The objective is again a `steady state' situation with the least remaining shelf life products being sold first. This approach was coded LSFO (least shelf life first out). The LSFO reduces the number of rejected products and largely eliminates consumer dissatisfaction since the fraction of product with unacceptable quality at the time of use or consumption is minimized. LSFO aims to reduce the rejected products at the consumer end, by promoting, at selected decision making points of the product life cycle, those product units with the shorter shelf life, according to the response of the attached TTI (Taoukis et al., 1998; Giannakourou and Taoukis, 2003). LSFO allows the calculation of the actual remaining shelf life of individual product units at strategic control points of the chill chain. Based on the distribution of the remaining shelf life, decisions can be made for improved handling, shipping destination and stock rotation. A further improvement of the LSFO approach is a chill chain management system coded SLDS (shelf life decision system) (Giannakourou et al., 2001). Compared to LSFO, SLDS policy takes additionally into account the realistic variability of the initial quality state Ao of the product. The state of the TTI technology and of the scientific approach with regard to the quantitative safety risk assessment in foods allows the undertaking of the next important step: the study and development of a TTI-based management system that will assure both safety and quality in the food chill chain (Koutsoumanis et al., 2005; Tsironi et al., 2008). The development and application of such a system coded with the acronym SMAS, was the target of the multinational European research project `Development and Modeling of a TTI based Safety Monitoring and Assurance System (SMAS) for Chilled Meat Products' (project QLK1CT2002-02545, 2003-2006; http://smas.chemeng.ntua.gr). SMAS uses the information from the TTI response at designated points in the chill chain, ensuring that the temperature-burdened products reach consumption at acceptable quality level. Although SMAS is developed for meat products, the


312


same principles can be effectively applied to the management of the chill chain of all chilled food or non-food perishable products. The effectiveness of the TTI-based SMAS system was evaluated by running a large number of chill chain scenarios using a Monte Carlo simulation approach. Field test experiments to demonstrate and quantify the improvement at the time of consumption in comparison to the conventional FIFO rule, were also conducted. The SMAS decision-making routine at a specified control point of the chill chain is based on the microbial growth that has potentially occurred within the period between production and arrival of the product at the control point. The growth is estimated based on the product's characteristics and the timetemperature history of the product using the appropriate predictive model. The above elements form the core program of integrated software that allows the calculation of growth in individual product units (e.g. small pallets, 5±10 kg boxes or single packs) at strategic control points in the chill chain. Based on the relative growth, it is possible to make decisions for optimal handling, shipping destination and stock rotation, aiming to obtain a narrow distribution of quality at the point of consumption. At a certain point in the chill chain, e.g. at a distribution center, product from the same initial shipment is split in half and is forwarded to two different retail markets, a close and a distant one that requires long transportation. The split could be random according to conventional, currently used FIFO practice or it can be based on the actual microbial growth of the product units and the developed decision system. For all units, the time-temperature history of the product monitored by TTI is input. This information fed directly into a portable unit with the SMAS software, is translated to microbial status, Nt, based on the growth models of the pathogen of concern. Having calculated Nt for all the n product units, a microbial load distribution for the products at the decision point is constructed. Based on the load of each product unit relative to this distribution, decisions about its further handling are made (Fig. 9.8). In order to simulate the results of the application of the developed SMAS system and quantitatively prove its effectiveness, the Monte Carlo method can be applied (Koutsoumanis et al., 2005). By taking into account the status of the product after production and various temperature distributions at different steps and alternative storage conditions, the distribution of the quality of the studied set of products at the final stage of consumption can be estimated. To confirm the effectiveness of SMAS appropriate experiments were also designed and conducted. One such experiment is described below to demonstrate the SMAS approach and the kind of information required for its application. A kinetic model for growth of spoilage bacteria in modified atmosphere packed (MAP) minced beef was developed and appropriate enzymic and photochromic TTI were studied and selected in order to monitor the meat product quality during distribution. The applicability of the SMAS system for chill chain management and optimization of the studied products was demonstrated (Taoukis et al., 2010).



313

Fig. 9.8 Classification of products based on their microbial load and rationale of SMAS-based split at the decision point (where A and B are the two possible destinations, the distant and the local market).

In order to present the applicability of the selected TTIs as shelf life monitors for MAP minced beef, a realistic distribution scenario in the chill chain was simulated. The simulated chill chain conditions consisted of five different timetemperature scenarios with effective temperatures, Teff, between 2 and 15 ëC, conducted in programmable temperature cabinets. Products were split at a designated point in the simulated chill chain, 24 hours from packing (corresponding to the distribution center) and followed a simulated path to a `local' and a `distant' market. Products with TTIs were split based on TTI response translated into temperature history based on TTI kinetics. According to the TTI-based system, the more temperature burdened products were diverted to the `local' market, shortening their shelf life cycle in order to be consumed first (Giannakourou et al., 2001). Sixty packages without TTI were split randomly, according to conventional, currently used FIFO practice. Half of the samples were subsequently stored at 4 ëC and half at 8 ëC for different times simulating the different final consumption times of products sold at the local and the distant market (Fig. 9.9). Lactic acid bacteria (LAB) level at these times, assumed as the end of storage period and time of consumption, was calculated using the kinetic models from the study on MAP minced beef. LAB level for packs simulating transport to a local market was calculated at 12, 24 and 36 hours and for packs simulating transport to the distant market was calculated at 48, 72 and 96 hours after split. The LAB loads of the overall 60 SMAS managed packs with TTI were compared to the LAB loads of the 60 FIFO managed packs without TTI. The response of the enzymic CheckPointÕ TTI, type M was described by the normalized a b value of the CIELAB scale (Eq. 9.4): X1

a b ÿ a bmin a bmax ÿ a bmin


9:4

ß Woodhead Publishing Limited, 2011 Fig. 9.9

Field test simulated chill chain conditions.


315

Equation 9.4 represents the M-type TTI response, X1 , which, when plotted as a function of time, has a sigmoidal shape, of specific pattern, that was described by the following logistic type equation: 1 9:5 X1 k1I ÿ t 1 exp k2I where k1I and k2I are functions of storage temperature and enzyme concentration. Note that 1/k2I is the exponential rate constant, i.e. the slope of the phase in which the response of the TTI changes exponentially with time (Fig. 9.10). Values of k1I and k2I at each temperature were determined, by non-linear regression analysis (Sigma Plot 8.0). The response rate constants k1I and k2I showed the same temperature dependence expressed by the same value of

Fig. 9.10 (a) Response of Check Points TTIs at different enzyme concentrations and isothermal storage at 5 ëC (experimental points and potential fit) · M-50U, M-100U and t M-200U, (b) total response time of Check Point TTIs as a function of temperature at different enzyme concentrations (ÐÐ 50U, ± ± ± 100U, 200U) calculated by the composite model.


316


activation energy of the Arhenius equation (Ea 96±107 kJ/mol). The value of k1 was expressed as a function of enzyme concentration and a composite model was developed that can determine the TTI response at a selected timetemperature scenario and known enzyme concentration: 1 X FXC 9:6 0 1 Ea 1 1 ÿB1 ÿt k C exp ÿ B 1;ref C R T Tref C 1 expB @ A E 1 1 a k2;ref C ÿB2 exp ÿ R T Tref where T is the absolute temperature (K), Ea is the activation energy (kJ/mol), R is the universal gas constant, Tref is a reference temperature (4 ëC), C is the enzyme concentration (50±200 U) and B1,2 are constants. The value of Ea for all different enzyme concentrations of the TTI was defined at 105:5 15:8 kJ/mol. The visual color change of the photochromic OnVu TTIs was adequately described by the E value (Eq. 9.7). E was modelled by an exponential decay function with time (Eq. 9.8) as representatively shown in Fig. 9.3. q 9:7 E L ÿ Lmax 2 a ÿ a2min b ÿ bmax 2 E Eo expÿk t

9:8

where k is a function of initial charging time, tc, and storage temperature T (K). The Lmax, amin and bmax values are the values of the `white' colored, uncharged TTI (Lmax 80, amin ÿ3.5 and bmax 1.2). The end-point of the OnVu TTI was determined at E 11:9 (Lf 69, af ÿ5 and bf ÿ6), corresponding to the visual end-point. The response rate constants were plotted as a function of temperature in Arrhenius plots. The total response time (time from activation to end-point) of the TTIs is shown in Fig. 9.11. Based on the results of the testing of the TTI, a composite model that allows the calculation of the response rate, k, at any selected charging time was developed. The model was based on the observation that the response rate k at any temperature is a power function of the charging time, tc. It was also observed that the effect of charging time on the Ea values of the TTIs is within the statistical variation. Thus the Ea value can be assumed not to change with charging time. The form of the composite model is expressed as: ÿEa 1 1 9:9 ÿ k kref ;Tref ;1s tcÿA exp R T Tref where T is the absolute temperature (K), Ea is the activation energy (kJ/mol), R is the universal gas constant, Tref is a reference temperature (4 ëC), kref,Tref,1s is the TTI response rate constant at Tref (with charging time tc 1 s). The composite model allows the calculation of the charging time needed in order to achieve the required color change rate and response time for the TTI in order to



317

Fig. 9.11 Response of OnVu TTI at different charging times, tc (n 1 s, s 2 s, l 3 s) and isothermal storage at 5 ëC (experimental points and potential fit), (b) total response time of OnVu TTI as a function of temperature at different charging times, tc (ÐÐ 0.5 s, 1 s, ± ± 2 s, ± ± 3 s and === 4 s) calculated by the composite model.

better match the respective kinetics of the product. The activation energy value, Ea, for all different charging levels of the TTI was defined at 122 16:6 kJ/mol. Lactic acid bacteria were used as the spoilage index for MAP minced beef. Starting from an initial level of 103 CFU/g in raw fresh meat, LAB reached high populations of 107±108 CFU/g at the end of the storage period. The dominance of LAB in the CO2-rich atmosphere has been repeatedly reported in the literature for meat and meat products (Borch et al., 1996; Devlieghere et al., 1998; Mataragas et al., 2006; Nychas et al., 2008; Vaikousi et al., 2009). The experimental data for LAB of the MAP minced beef are shown in Fig. 9.12(a) with the fitted growth curves. The end of shelf life, i.e. the limit of sensory acceptability, was correlated to a 7 logCFU/g level, in agreement with published data (Borch et al., 1996; Devlieghere et al., 1998; Mataragas et al., 2006; Nychas et al., 2008; Vaikousi et al., 2009). The specific growth rates (k) of LAB


318


Fig. 9.12 (a) Growth of lactic acid bacteria on MAP minced beef packed at ± ± 0 ëC, ÐÐ 2.5 ëC, 5 ëC and ± ± 10 ëC, (b) shelf life (h) is MAP minced beef at different storage temperatures, calculated using the Arrhenius model, together with the matching TTI response curves (dotted lines: OnVu TTI, tc 2 s and dashed lines: ± ± ± Check Point TTI M-50U).

were modeled as a function of temperature using the Arrhenius equation. The activation energy value was defined at 122:5 14:0 kJ/mol and the growth rate of LAB at the reference temperature (4 ëC), kref, was 0:0171 0:0014 hÿ1. The applicability of the developed kinetic equations was validated at variable time-temperature profiles. The growth rate of the LAB predicted by the Arrhenius model, for the Teff of the profile, was compared with the rate obtained by the non-isothermal experiment. The relative error value was below the limit of 20% that is used in the literature as the criterion of applicability (Dalgaard et al., 1997; Gougouli et al., 2008), indicating that the Arrhenius model can describe satisfactorily the growth of LAB in MAP minced meat during refrigerated storage in the range of 0±10 ëC. To select the appropriate charging time, based on the kinetic studies of MAP minced beef and the response profiles of the TTIs, the composite models (Eq. 9.9) can be solved. To obtain an exact match between the product shelf life and response time of the TTI at a reference temperature in the chilled range (e.g. 4 ëC), it was calculated that charging time of 2 s for the OnVu TTI and 50 U for



319

the M-type CheckPoint TTI could lead to suitable TTIs for monitoring the quality of MAP minced beef during refrigerated storage. In Fig. 9.12(b) the shelf life of MAP minced beef is shown together with the matching TTI total response time curves. If the products are stored at very low temperatures of 0±2 ëC the end of shelf life will be determined and limited by the expiration date on the food package. On the other hand, if abuse temperatures of 6±10 ëC prevail, then the TTI will conservatively signal poor quality products slightly before the end of shelf life. The SMAS system and FIFO approach were applied to MAP minced beef. The effectiveness of the SMAS system was evaluated based on the level of lactic acid bacteria at the end of storage. Samples with LAB levels above 7 logCFU/g were characterized as spoiled. The initial level of LAB was 3.1 logCFU/g. Using the FIFO approach, 12% of 60 samples ± `local' and `distant' promoted ± reached the spoilage level at the end of the storage period. When SMAS-based sorting is applied, only 3% of the 60 samples reached the spoilage level, significantly reducing the number of rejected products before the `time of consumption'. In Fig. 9.13 the observed LAB log counts of MAP minced beef samples at the `time of consumption' are depicted. In total (for both `local' and `distant' markets), the SMAS system resulted in a reduction in the number of spoiled products. SMAS uses the information from the TTI, at appropriate points in the chill chain (e.g., at a central distribution center), to make decisions for the further management of products based on their temperature history and hence microbial status and remaining shelf life of the products (Fig. 9.5b). The conducted simulated field test demonstrates the applicability of the TTI-based SMAS approach to improve the meat chill chain. The overall field test result showed that SMAS-based sorting at a decision point resulted in more uniform acceptable quality at the time of consumption in comparison to the conventional FIFO approach. In a previous study, growth of LAB and Listeria monocytogenes in fresh ground lamb (MAP) was modeled (Taoukis et al., 2006). Growth was measured on naturally contaminated products inoculated with the pathogen, at isothermal and dynamic conditions from 0 to 15 ëC. Enzymatic TTI with suitable response was also modeled. Some 120 products of MA packed ground lamb (20% CO2) were then tested, on half of which TTIs were attached at packing. All products were stored, in programmed cabinets simulating the conditions of the real chill chain to the consumption point. At the decision point, at 64 h from packing, the products were split in half and were stored for three different short times and three longer times (local and distant market scenarios). The products without TTI were split randomly. Split and further handling of products with TTI was based on their integrated temperature history and the SMAS scheme. According to the final microbiological measurements of all products at `consumption time', the spoilage profile of the products with TTI was significantly improved. A total of 22% of randomly handled samples were spoiled at consumption time (LAB > log8) compared to less than 6% handled with the SMAS approach. Respectively, 28% exceeded a set limit for Listeria compared to less than 3% handled with


320


Fig. 9.13 (a) Distribution of the lactic acid bacteria log counts and (b) distribution of the remaining enzyme activity based on FIFO and PMAS approach.

SMAS. These experimental results demonstrate the applicability of the TTIbased system to optimize quality and reduce risk.

9.5

Future trends

TTI-based syatems could replace the current FIFO practice and lead to risk minimization and quality optimization by improving distribution logistics and management of the food chill chain. It improves stock rotation in selected points of the chill chain. It ensures that the temperature abused products are consumed before they reach unacceptable risk. When recommended chill chain conditions are maintained, TTI-based practices do not differ from the FIFO practice. However, in case of incidental temperature abuse, TTI-based systems can manage the chain by diverting abused products so that final rejection and risk are minimized. Cold chain optimization and effective management will be a central issue in research, industrial practices, and regulatory efforts, as industry continuously



321

strives to deliver high quality foods and other perishable items to consumers. Integrated systems, like the proposed SMAS based on the availability of quality data and temperature history of individual product units, will be applied and validated in practice, and TTIs can be combined with RFID technology to supplement the current traceability requirements mandated by regulation or developed by industry initiatives.

9.6

Acknowledgements

This chapter includes results partly carried out with the financial support of the Commission of the European Communities, specific RTD program `Quality of Life and Management of Living Resources', Key Action 1 ± Health Food and Environment, Project No. QLK1-CT2002-02545 (http://smas.chemeng.ntua.gr), FP6 Collective Research Project COLL-CT-2005-012371 (http:// www.freshlabel.net) and FP7 Capacities Project IQ-Freshlabel (G.A. no: 243423).

9.7

References

(1996), Bacterial spoilage of meat and cured meat products. International Journal of Food Microbiology 33: 103±120. DALGAARD P, MEJLHOLM O, HUSS HH (1997), Application of an iterative approach for development of a microbial model predicting the shelf-life of packed fish. International Journal of Food Microbiology 38: 169±179. DEVLIEGHERE F, DEBEVERE J, VANIMPE J (1998), Effect of dissolved carbon dioxide and temperature on the growth of Lactobacillus sake in modified atmospheres. International Journal of Food Microbiology 41: 231±238. BORCH E, KANT-MUERMANS ML, BLIXT Y

ELLOUZE M, PICHAUD M, BONAITI C, COROLLER L, COUVERT O, THUAULT D, VAILLANT R

(2008), Modelling pH evolution and lactic acid production in the growth medium of a lactic acid bacterium: application to set a biological TTI. International Journal of Food Microbiology 128: 101±107. GIANNAKOUROU MC, TAOUKIS PS (2003), Application of a TTI-based distribution management system for quality optimisation of frozen vegetables at the consumer end. Journal of Food Science 68(1): 201±209. GIANNAKOUROU MC, KOUTSOUMANIS K, NYCHAS GJE, TAOUKIS PS (2001), Development and assessment of an intelligent shelf life decision system for quality optimization of the food chill chain. Journal of Food Protection 64(7): 1051±1057. GOUGOULI M, ANGELIDIS AS, KOUTSOUMANIS K (2008), A study on the kinetic behavior of Listeria monocytogenes in ice cream stored under static and dynamic chilling and freezing conditions. Journal of Dairy Science 97: 523±530. KOUTSOUMANIS K, TAOUKIS PS, NYCHAS GJE (2005), Development of a Safety Monitoring and Assurance System (SMAS) for chilled food products. International Journal of Food Microbiology 100: 253±260. LOUVET O, THUAULT D, VAILLANT R (2005), Method and device for determining whether or not a product is in condition to be used or consumed. Patent. International Publication Number WO 2005/026383 A1. MATARAGAS M, DROSINOS EH, VAIDANIS A, METAXOPOULOS I (2006), Development of a


322


predictive model for spoilage of cooked cured meat products and its validation under constant and dynamic temperature storage conditions. Journal of Food Science 71: M157±M167. NYCHAS GE, SKANDAMIS PN, TASSOU CC, KOUTSOUMANIS KP (2008), Meat spoilage during distribution. Meat Science 78: 77±89. SCHOEN HM, BYRNE CH (1972), Defrost indicators: many designs have been patented yet there is no ideal indicator. Food Technology 26(10): 46±50. SHERLOCK M, LABUZA TP (1992), Consumer perceptions of consumer time-temperature indicators for use on refrigerated dairy foods. Journal of Dairy Science 75: 3167± 3176. SHERLOCK M, FU B, TAOUKIS PS, LABUZA TP (1991), Systematic evaluation of time temperature indicators for use as consumer tags. Journal of Food Protection 54 (11): 885±889. SMOLANDER M (2003), The use of freshness indicators in packaging. In Ahvenainen R (ed.), Novel Food Packaging Techniques. Woodhead Publishing Ltd, Cambridge, 127±143. TAOUKIS PS (2001), Modelling the use of time temperature indicators in distribution and stock rotation. In Tijskens L, Hertog M, Nicolai B (eds), Food Process Modelling, Woodhead Publishing Ltd, Cambridge. TAOUKIS PS (2010), Commercialization of active food packaging (Time-Temperature Integrators, `TTIs'). In Doona C, Kustin K, Feeherry F (eds), Case Studies in Novel Food Processing Techniques. Woodhead Publishing Ltd, Cambridge. TAOUKIS PS, LABUZA TP (1989), Applicability of time temperature indicators as shelf life monitors of food products. Journal of Food Science 54: 783±788. TAOUKIS PS, LABUZA TP (2003), Time-temperature indicators (TTI). In Ahvenainen R (ed.), Novel Food Packaging Techniques. Woodhead Publishing Ltd, Cambridge, 103±126. TAOUKIS PS, BILI M, GIANNAKOUROU M (1998), Application of shelf life modeling of chilled salad products to a TTI based distribution and stock rotation system. In Tijskens LMM, Hertog MLATM (eds), Proceedings of the International Symposium on Applications of modeling as an innovative technology in the Agri-food chain. Acta Horticulturae, 476: 131±140. TAOUKIS PS, KATSAROS G, GOGOU E, TSIRONI T, TSEVDOU M (2006), Application and Experimental Validation of the TTI Based Chill Chain Management System SMAS for MAP Lamb Products. Food Micro 2006: The 20th International ICFMH Symposium Food safety and food biotechnology: diversity and global impact International Committee on Food Microbiology and Hygiene (ICFMH), 29 August±2 September 2006, Bologna, Italy, Book of Abstracts, p. 529. TAOUKIS PS, TSIRONI TH, GIANNOGLOU M, METAXA I, GOGOU E (2010), Historical review and state of the art in time temperature integrator (TTI) technology for the management of the cold chain of refrigerated and frozen foods. In Proceedings of Cold Chain Management, 4th International Workshop, Bonn, 27±28 September. TSIRONI T, GOGOU E, VELLIOU E, TAOUKIS PS (2008), Application and validation of the TTI based chill chain management system SMAS (Safety Monitoring and Assurance System) on shelf life optimization of vacuum packed chilled tuna. International Journal of Food Microbiology 128: 108±115. VAIKOUSI H, BILIADERIS CG, KOUTSOUMANIS KP (2009), Applicability of a microbial time temperature indicator (TTI) for monitoring spoilage of modified atmosphere packed minced meat. International Journal of Food Microbiology 133: 272±278.


10 Food storage trials: an introduction C. M. D. Man, London South Bank University, UK

Abstract: This chapter reviews definitions of shelf life and key concepts such as `best before' and ùse by' dates. It reviews key factors affecting food spoilage and deterioration. It then goes on to discuss the principles and practices of shelf life testing and storage trials to establish accurate shelf life dates which manufacturers can use in food labelling. Key words: shelf life, stability, food spoilage, best before dates, use by dates, storage trials.

10.1

Introduction

Because of their perishable nature, all foods (taken to mean food and beverages) deteriorate throughout storage, albeit at different rates. For this and other related reasons, the behaviour of foods during storage is of immense interest to the consuming public as well as to all those who make, prepare, handle, buy, sell and distribute foods. The period during which a food remains safe and enjoyable to eat has been called its shelf life, which is now a legal term within the EU. Regulation (EC) No. 852/2004 of the European Parliament and of the Council on the hygiene of foodstuffs, implemented in England along with other associated regulations as the Food Hygiene (England) Regulations 2006, requires food business operators to adopt as appropriate a number of specific hygiene measures (Article 4(3(a))), which include `compliance with microbiological criteria for foodstuffs' as laid down in Commission Regulation (EC) No. 2073/ 2005 on microbiological criteria for foodstuffs. The latter regulation defines `shelf life' as èither the period corresponding to the period preceding the ùse


326


by' or the minimum durability date, as defined respectively in Articles 9 and 10 of Directive 2000/13/EC, the most recent European labelling Directive. A much more useful and informative definition of shelf life of food has been available for some time (IFST, 1993): it is the period of time during which the food will · remain safe; · be certain to retain its desired sensory, chemical, physical, microbiological and functional characteristics; · where appropriate, comply with any label declaration of nutrition data, when stored under the recommended conditions. Clearly, safety and quality are the two main aspects of shelf life of food, and food safety must always take priority over quality as it is both a fundamental and a legal requirement. According to Article 14 of the General Food Law Regulation (EC) 178/2002, food must not be placed on the market if it is unsafe. Food is deemed to be ùnsafe' if it is considered: · injurious to health; · unfit for human consumption. Food can be injurious to health by virtue of the presence of a hazard, which may be microbiological, chemical or physical in nature. Article 14(4) of the Regulation goes on to say: in determining whether any food is injurious to health, regard shall be had: (a) not only to the probable immediate and/or short-term and/or longterm effects of that food on the health of a person consuming it, but also on subsequent generations; (b) to the probable cumulative toxic effects; (c) to the particular health sensitivities of a specific category of consumers where the food is intended for that category of consumers. In terms of ùnfitness' for human consumption the central concept is unacceptability. Food can become unfit for human consumption because of contamination, whether by foreign objects (e.g., glass, plastics), by chemicals (e.g., cleaning chemicals, agrichemical residues) or by microbes causing putrefaction, decomposition or decay. Consequently, there should be little doubt in the mind of a food business operator as to what safe food means. Quality, on the other hand, is not usually regulated by law unless it has to do with compositional/marketing standards. British Standard BS EN ISO 9000:2005 (Quality management systems ± Fundamentals and vocabulary) defines quality as `degree to which a set of inherent characteristics fulfils requirements'. In practice, therefore, it is the job of every food business operator to establish as fully as possible the requirements of its target consumer and to ensure that the characteristics of its food product in question reflect and fulfil those require-


Food storage trials: an introduction

327

ments consistently. The need to provide an acceptable, reliable and consistent product shelf life represents an obligation as well as a challenge to every food business operator. In the UK the legal responsibility to assign an acceptable shelf life is contained in the Food Labelling Regulations 1996 as amended, such that pre-packed foods that are required to carry a durability indication (i.e., most foods) must indicate either: · `Best before' followed by the date up to and including which the food can reasonably be expected to retain its specific properties if properly stored, or · Ùse by', for foods which are, from the microbiological point of view, highly perishable and in consequence are likely after a short period to constitute an immediate danger to health, followed by the date up to and including which the food, if properly stored, is recommended for use. All such declarations must by followed by an indication of any storage conditions that need to be observed if the unopened food is to retain its specific properties up to the date indicated. This is understandable as all foods are perishable, they will naturally deteriorate in an unexpected manner, or faster or both, if they are stored under harsher conditions (usually warmer and more humid) for which they are not intended. The decision as to whether a food requires a ùse by' date is one for those who manufacture, pack and therefore mark it in the first place. Useful guidance, however, is available and it has been suggested that the following food groups, essentially all chilled foods, are likely to require a ùse by' date (Crawford, 1998): · · · ·

dairy products, e.g. fresh cream-filled desserts cooked products, e.g. ready-to-heat meat dishes smoked or cured ready-to-eat meat or fish, e.g. hams, smoked salmon fillets prepared ready-to-eat foods, e.g. sandwiches, vegetable salads such as coleslaw · uncooked or partly cooked savoury pastry and dough products, e.g. pizzas, sausage rolls · raw ready-to-cook products, e.g. uncooked products comprising or containing either meat, poultry or fish, with or without raw prepared vegetables · vacuum or modified atmosphere packs, e.g. raw ready-to-cook duck breast packed in modified atmosphere In order to arrive at an acceptable, reliable and reproducible shelf life, a food business operator will need to have answers for the following questions: 1. Is my product safe to eat throughout its intended shelf life? (Essentially, an unsafe food product has no useful/meaningful shelf life.) 2. How long will my product last for before it becomes unacceptable to the target consumer? In order to answer these separate and yet related questions satisfactorily, a food business operator needs to have sufficient knowledge about its product, in


328


particular, a thorough understanding of the shelf life limiting mechanism of deterioration and factors that influence it. Typically, shelf life is determined by conducting shelf life studies commonly but not exclusively by experimentation using storage trials under defined conditions. The final shelf life that is ultimately assigned may be decided dependent upon commercial considerations such as product category and associated image perceptions, and the margin of safety required. This shelf life is then expressed legally either as a ùse by' or `best before' date. Such is the importance of the legal requirement to set appropriate and accurate date marks that the UK Food Standards Agency launched a consultation on 25 March 2010 on the latest revision of its existing guidance on the application of date marks to food (FSA, 2010).

10.2

Food deterioration and spoilage

Changes in the characteristics of food inevitably occur during its storage. With very few exceptions such as cheeses and wines, these changes result in deterioration and spoilage of the food to the point when it is no longer acceptable to the target consumer and are usually classified as: · microbiological · non-microbiological ± biochemical ± chemical ± physical ± temperature-related. When they happen these changes effectively constitute the underlying mechanism(s) of deterioration and spoilage of the food in question, which if allowed to continue, will either singly or in combination cause the food to be rejected by the consumer. Figure 10.1 provides a picture of the progression of these changes, which cause the food to deteriorate and spoil during storage. In practice, a number of such changes can take place simultaneously or in sequence; in many cases, though, a particular type of change is likely to be the predominant one, which turns out to be the shelf life limiting change. The primary aim of a shelf life experiment, therefore, is to learn about these changes as they impact on the behaviour of the food during storage. As time goes on, a point is eventually reached when the food becomes unacceptable to the consumer, which marks the end of its shelf life and which has to be determined. A brief review of the different types of changes that can occur in food is given in the following sections. 10.2.1 Microbiological changes Besides the initial load or level of contamination, microbial growth depends on a number of well-known factors, which have been summarised by Mossel (1971):



329

Fig. 10.1 A picture of changes in food during storage.

· intrinsic properties of the food (e.g., nutrients, pH, total acidity, water activity, structure, presence of preservatives and/or natural antimicrobials, redox potential) · extrinsic factors (e.g., environmental temperature, relative humidity, gaseous atmosphere) · processing factors (e.g., heat destruction, freezing, packaging) · implicit factors (e.g., physiological attributes such as specific growth rate of the micro-organisms and microbial interactions). Micro-organisms, be they pathogenic or spoilage, share the same factors for growth. However, the growth of pathogenic organisms in food such as Salmonella species and Listeria monocytogenes is not necessarily accompanied by changes in appearance, smell, and even taste or texture that human senses can detect, posing serious health concerns. On the other hand, growth of spoilage organisms in food is often associated with signs that can readily be recognised as changes in sensory properties, for example, visual mould growth and production of objectionable odours and flavours. Examples of some common food spoilage organisms and changes they cause in food are given in Table 10.1. 10.2.2 Biochemical and chemical changes Raw materials from which practically all food products are manufactured are biological in origin, and it may be unappreciated by the average consumer that food is composed of chemicals. Some biochemical and/or chemical changes in food are therefore inevitable. These changes can occur arising from reactions within the food or from reactions between food components and external species or factors such as oxygen or light respectively. In packaged food, interactions, many of which are chemical in nature, can occur between packaging and the food. With a few exceptions such as maturing of wines and cheeses, and post-harvest ripening of fruits, most biochemical and chemical changes in food are undesirable, deteriorative and effectively shelf life limiting.


330


Table 10.1 Examples of food spoilage organisms and changes they cause in food (adapted from Huis in't Veld, 1996) Food spoilage organisms

Changes in food

Gram-negative rod-shaped bacteria, e.g. Pseudomonas spp.

Production of off-flavours, visible slime and pigmented growth in red meat, fish, poultry, milk and dairy products

Gram-positive spore-forming bacteria e.g. Bacillus and Clostridium spp.

`Sweet curdling' and `bitty cream' in milk (Bacillus cereus) Gas production ± `late blowing' of hard cheeses (Clostridium spp.)

Other Gram-positive bacteria, e.g. Brochothrix thermosphacta

Off-flavour development in MAP and VP meat products

Lactic acid bacteria, e.g. Lactobacillus, Streptococcus, Leuconostoc and Pediococcus spp.

Slime formation, generation of CO2, production of lactic acid, causing a drop in pH and off-flavour development in some dairy products

Yeasts and moulds

Production of soft rot in fruit, pigmented growth in baked goods, production of acid, gas or alcohol in some soft drinks and jams, development of off-odours in beer

Examples of some biochemical and chemical changes in food are given in Table 10.2. 10.2.3 Physical changes Significant transfer of moisture (or water vapour) and/or other substances in or out of food can often cause deteriorative changes in food. These changes are very common and can affect short-, medium- as well as long-life products. Most of these changes are important from a product quality point of view while a few can have food safety implications such as in the case of migration of chemical components from the packaging material into food, particularly when the latter has a long shelf life. In the EU, both overall and specific migrations of chemical components from packaging materials into food are controlled by Regulation (EC) No. 1935/2004 on Materials and Articles intended to come into contact with Food. Examples of some physical changes in food are given in Table 10.3. 10.2.4 Temperature-related changes Temperature, arguably the most important environmental factor, affects all of the above changes and not always in the same way. Micro-organisms, pathogens and spoilage organisms, exhibit a range of minimum growth temperatures below which they cannot grow. For instance, temperature selects for the types of organisms that can survive and grow at refrigerated temperatures. Table 10.4



331

Table 10.2 Some biochemical and chemical changes in food (adapted from Man, 2002) Biochemical/chemical reactions

Changes in food

Oxidative rancidity, e.g. oxidation of fats and oils involving a catalyst such as copper ions; oxidation of fats and oils initiated by light in the presence of a photo-sensitizer such as myoglobin; oxidation of fats and oils catalysed by the enzyme lipoxygenase

Rancidity (off-flavour development) in fatty food and food products

Oxidation-reduction reactions with atmospheric oxygen

Degradation and loss of vitamins C, B1, A and E

Hydrolysis of aspartame (sweetener)

Reduction in sweetness of calorie-free/ low-calorie soft drinks

Non-enzymic browning (Maillard reaction)

Browning (discoloration) in dehydrated fruits and vegetables, instant potato powder, dried egg white and dried milk products

Enzymic browning

Browning in pre-cut vegetables and fresh fruit salads

Chemical breakdown caused by light

Colour fading

Electrochemical reactions between foods and tinplate cans

Gas production, discoloration of food, etc., depending on the food and type of metal can

Table 10.3

Examples of physical changes in food (adapted from Man, 2002)

Product

Quality change

Underlying mechanism

Fresh vegetables

Wilting

Moisture loss

Biscuits

Softening, loss of crunchiness

Moisture gain

Carbonated soft drinks

Loss of fizziness

Loss of carbonation (CO2) to the environment

Orange juice

Reduction in citrus flavour intensity

Sorption of limonene and other aroma compounds by the packaging material

Dressed salads, e.g. coleslaw

Changes in texture of vegetables, changes in consistency of dressing

Moisture migration from vegetables to dressing

Chilled composite desserts, Gradual loss of distinctive e.g. trifle layers

Bleeding of colours, migration of moisture/syrup


332


Table 10.4 Some pathogenic micro-organisms known to be associated with chilled foods (Betts et al., 2004; Voysey, 2007) Micro-organism

Minimum growth temperature (ëC)

Salmonella Staphylococcus aureus Bacillus cereus (spores/heat resistant) Clostridium botulinum (non-proteolytic B, E, F) Listeria monocytogenes Escherichia coli Escherichia coli (O157:H7) Vibrio parahaemolyticus Yersinia enterocolitica Aeromonas hydrophila

4 5.2 (10 for toxin) 4 3 ÿ0.4 7±8 6.5 5 ÿ1.3 ÿ0.1

gives a list of pathogenic micro-organisms and their minimum growth temperatures that are known to be associated with chilled foods. In consequence, compliance with the relevant temperature control requirements, i.e., a maximum storage temperature of 8 ëC, of the current food hygiene regulations (TSO, 2006) is essential in assuring the microbiological safety and stability of chilled foods. The effect of elevating temperature on many chemical reactions, and hence potential adverse chemical changes in food during storage, is well known; increasing the temperature generally increases the rate of chemical reactions by a factor of 10. This empirical relation between the rate of reaction (k) and temperature was first proposed by Svante Arrhenius in 1889: Ea k A exp ÿ RT where A the frequency factor (or pre-exponential factor), Ea the activation energy, R the universal gas constant (0.001987 kcal molÿ1 Kÿ1 or 8.31 J molÿ1 Kÿ1), and T the absolute temperature in K (kelvin). Converting this relationship to logarithmic form, the following is obtained: Ea log10 k log10 A ÿ 2:303RT or ln k ln A ÿ

Ea RT

In theory, a plot of lnk versus the reciprocal of absolute temperature should give a straight line, the slope of which is the activation energy divided by the gas constant (Ea/R). A graph of ln k against 1/T is called an Arrhenius plot; many chemical reactions have been found to show Arrhenius behaviour, i.e. their Arrhenius plots show a straight line. Thus, by studying a reaction and measuring k at two or three different temperatures, one could extrapolate with a straight line to a lower temperature and predict the rate at this temperature. This is the



333

basis of accelerated storage trials for shelf life at an elevated temperature. Often though, reactions in real food systems are far more complex than can be easily modelled by the Arrhenius equation. For certain non-microbiological changes in food, lower temperatures do not automatically mean lower rates of change or insignificant changes. For instance, bread stales fastest at refrigerated temperatures; increased temperatures can slow the development of bread staling, which is thought to be due to re-distribution of moisture and retrogradation of starch molecules. Fluctuating temperatures can cause ice crystal formation in frozen foods such as ice cream, and changes in fat crystallinity are promoted by fluctuating storage temperature, which encourage bloom to develop in chocolate. 10.2.5 A summary The changes that bring about deterioration and spoilage in food as outlined in the previous sections can be summarised as in Fig. 10.2. Microbiological and non-microbiological changes can take place in parallel or in sequence. More than one type of change can take place at the same time, and changes in many foods can be complex. Nevertheless, a number of well-known mechanisms broadly classified as microbiological and non-microbiological changes can be used to explain deterioration, spoilage and subsequent loss of shelf life in many food products (Man, 2004): · microbiological changes · non-microbiological changes ± biochemical and chemical changes including light-induced changes ± moisture and/or water vapour transfer leading to gain or loss ± physical transfer of substances such as oxygen, odours or flavours other than moisture and/or water vapour ± other mechanisms or changes such as loss of pack integrity. The question as to which of the above changes and indeed what predominant change will take place in a food, will depend on many shelf life influencing

Fig. 10.2

A basic model for food deterioration and spoilage (Ellis and Man, 2000).


334


factors, which can be categorised into product and external factors (IFST, 1993). Product factors are related to the composition, make-up and properties of the final product. They include the following: · · · · · · · ·

raw materials (their microbiology and biochemistry) product composition and formulation (e.g., use of preservatives) food structure (i.e., homogeneous versus heterogeneous) product assembly (i.e., composite, multi-component product) pH value, and total acidity including type of acid water activity (aw) redox potential (Eh) oxygen availability.

External factors are those that the final product is subject to or comes into contact with as it moves through the food chain up to the point of consumption. They include the following: · hygienic conditions during preparing, processing, storage and distribution · type and extent of processing (e.g., time-temperature combination of heat treatment) · conditions within packaging (i.e., composition and pressure of atmosphere) · packaging materials and system · exposure to light (UV and IR) during processing, storage and distribution · temperature control throughout the food chain · relative humidity during processing, storage and distribution · consumer handling, preparation and use.

10.3

Storage trials

The most common and direct way of determining shelf life is to carry out experimentally storage trials of the product in question under conditions that simulate those it is likely to encounter during storage, distribution, retail display and consumer use. The aims of all storage trials of food are the same, which are, as indicated earlier in this chapter: · to establish the safety of the food throughout its intended shelf life, whatever its length, and · to arrive at a period of time during which the food will be certain to retain its sensory, chemical, physical, microbiological and functional characteristics that meet the target consumer requirements, and where appropriate, comply with any label declaration of nutrition data, when stored under the recommended conditions. 10.3.1 Safe shelf life In order to establish food safety, the most effective way, which is also a legal requirement within the EU/UK, is to use the internationally recognised system



335

based on the Hazard Analysis and Critical Control Points (HACCP) principles as detailed in Article 5 of the EU Regulation (EC) No. 852/2004 on the hygiene of foodstuffs. The principles consist of the following (European Commission, 2004): (i) identifying any hazards that must be prevented, eliminated or reduced to acceptable levels; (ii) identifying the critical control points (CCPs) at the step or steps at which control is essential to prevent or eliminate a hazard or to reduce it to acceptable levels; (iii) establishing critical limits at CCPs which separate acceptability from unacceptability for the prevention, elimination or reduction of identified hazards; (iv) establishing and implementing effective monitoring procedures at CCPs; (v) establishing corrective actions when monitoring indicates that a critical control point is not under control; (vi) establishing procedures, which shall be carried out regularly, to verify that the measures outlined in (i) to (v) above are working effectively; and (vii) establishing documents and records commensurate with the nature and size of the food business to demonstrate the effective application of the measures in (i) to (vi) above. Earlier, Article 4 of the same Regulation requires food business operators to adopt as appropriate a number of specific hygiene measures, which include among others compliance with microbiological criteria for foodstuffs as set out in Commission Regulation (EC) No. 2073/2005 on microbiological criteria for foodstuffs. This Regulation establishes two types of microbiological criteria: food safety criteria, and process hygiene criteria (FSA, 2005). A food safety criterion is one that defines the acceptability of a product or a batch of foodstuff, applicable to products placed on the market. Applicable food safety criteria in Regulation No. 2073/2005 should therefore be used to establish safe microbiological shelf life during product development and to assess the microbiological safety of a food product or batch of products within the framework of an effective HACCP-based food safety management system. Examples of the microbiological (food safety) criteria set out in Annex I of Regulation No. 2073/ 2005 are given in Table 10.5. In an effort to assist food businesses of all levels of expertise to assign appropriate and correct date marks, the Chilled Food Association in the UK recently published a good practice guide on `Shelf life of ready to eat food in relation to L. monocytogenes', which was produced by a stakeholder drafting group chaired by the British Retail Consortium and which has been endorsed by the FSA (CFA, 2010). Benefiting from a wide knowledge and experience base, and taking advantage of collective wisdom, this guide effectively develops and expands the requirement in Regulation No. 2073/2005 for food business


Table 10.5 Examples of food safety criteria applicable to products placed on the market during their shelf life (taken from FSA, 2005 ± Annex 1, Chapter 1) Criterion

1.2

Micro-organism and food category

Examples of foods

Sampling plan

Limits m

Analytical reference method


n

c

M

5

0

100 cfu/g

EN/ISO 11290-2

5

0

*Absence in 25 g

EN/ISO 11290-1

5

0

100 cfu/g

EN/ISO 11290-2

Listeria monocytogenes Ready-to-eat foods able to support the growth of L. monocytogenes, other than those intended for infants and for special medical purposes

Chilled ready-to-eat products with more than 5 days' life Pre-packed delicatessen products Pre-packed sliced cooked meat Smoked salmon PaÃteÂ Soft cheese

1.3

Listeria monocytogenes Ready-to-eat foods unable to support the growth of L. monocytogenes, other than those intended for infants and for special medical purposes

Yoghurt Hard cheese Products with a pH less than 4.4, e.g. coleslaw Products with shelf life less than 5 days, e.g. sandwiches

1.4

Salmonella Minced meat and meat preparations intended to be eaten raw

Steak tartare

5

0

Absence in 25 g

EN/ISO 6579

1.8

Salmonella Meat products intended to be eaten raw, excluding products where the manufacturing process or the composition of the product will eliminate the salmonella risk

Salami Parma ham Cold smoked duck

5

0

Absence in 25 g

EN/ISO 6579


1.20

Salmonella Unpasteurised fruit and vegetable juices (ready-to-eat)

Freshly squeezed unpasteurised fruit juices, mixed fruit juices; smoothies; vegetable juices

5

0

Absence in 25 g

1.21

Staphylococcal enterotoxins Cheeses, milk powder and whey powder, as referred to in the coagulase-positive staphylococci criteria in Chapter 2.2 of Annex 1 of Regulation No. 2073/2005

Cheeses, excluding processed cheese and non-fermented cheese

5

0

Not detected in 25 g

1.23

Enterobacter sakazakii Infant milk and dairy products, as referred to in the Enterobacteriaceae criterion in Chapter 2.2 of Annex 1 of Regulation No. 2073/2005

Dried infant formulae and dried dietary foods for special medical purposes intended for infants below six months of age

30

0

Absence in 10 g

ISO/DTS 22964

1.24

E. coli Live bivalve molluscs and live echinoderms, tunicates and gastropods

Oysters, clams, sea urchins, winkles and welks

1

0

230 MPN/100 g of flesh and intra-valvular liquid

ISO TS 16649-3

1.25

Histamine Fishery products from fish species associated with a high amount of histidine

Tuna, mackerel, sardines, mahi

9

2

100 mg/kg

100 mg/kg

EN/ISO 6579

European screening method of the CRL for milk

HPLC

* This criterion applies to products before they have left the immediate control of the producing food business operator, when he is not able to demonstrate, to the satisfaction of the competent authority, that the product will not exceed the limit of 100 cfu/g throughout the shelf life.

338


operators to `conduct studies' (Article 3(2)) as necessary to ensure `that the food safety criteria applicable throughout the shelf life of the products can be met under reasonably foreseeable conditions of distribution, storage and use' (Article 3 1(b)). Annex II of the same Regulation says The studies referred to in Article 3(2) shall include: · specifications for physico-chemical characteristics of the product, such as pH, aw, salt content, concentration of preservatives and the type of packaging system, taking into account the storage and processing conditions, the possibilities for contamination and the foreseen shelf life, and · consultation of available scientific literature and research data regarding the growth and survival characteristics of the micro-organisms of concern. When necessary on the basis of the above-mentioned studies, the food business operator shall conduct additional studies, which may include: · Predictive mathematical modelling established for the food in question, using critical growth or survival factors for the microorganisms of concern in the product, · Tests to investigate the ability of the appropriately inoculated microorganism of concern to grow or survive in the product under different reasonably foreseeable storage conditions. · Studies to evaluate the growth or survival of the micro-organisms of concern that may be present in the product during the shelf life under reasonably foreseeable conditions of distribution, storage and use. The above-mentioned studies shall take into account the inherent variability linked to the product, the micro-organisms in question and the processing and storage conditions. In practice, therefore, in order to establish safe shelf life of ready-to-eat food in relation to L. monocytogenes and indeed other pathogens, a food business operator should use all or any suitable combination of the following: · product characteristics and relevant scientific literature and research data · historical data pertinent to the control of the pathogen of concern (i.e., in this case L. monocytogenes) · predictive microbiology, i.e. internet-based predictive microbiological models e.g. ComBase (http://www.combase.cc) · specific laboratory shelf life studies ± challenge testing ± storage trials under controlled conditions · Collaboration between food businesses in conducting shelf life studies. In view of the primary importance to assure microbiological safety of food, further guides in relation to other pathogens are likely to be produced in future, in particular, for ready-to-eat foods.



339

10.3.2 Challenge testing In general, a challenge test is a laboratory investigation of the behaviour of a product when subjected to a set of controlled experimental conditions. Challenge testing, in the context of shelf life studies, almost always refers to microbiological challenge testing, the aim of which is to simulate what can happen to a food product during processing, distribution and subsequent handling, following inoculation with one or more micro-organisms of concern. As such, it is a very useful tool for determining the ability of a food to support the growth of pathogens or spoilage organisms. The main areas of application of microbiological challenge testing include: · determining microbiological safety and assessing the risk of food poisoning after HACCP has identified the organisms likely to be the microbial hazards for the product at some stage during production and distribution; for example, this is useful in determining the safe shelf life of chilled foods (Uyttendaele et al., 2004) · establishing quality shelf life by inoculating the product with food spoilage organisms known or likely to contaminate it; for example, this is useful in evaluating the microbiological stability of emulsified and non-emulsified condiment sauces intended for ambient distribution (Jones, 2000) · studying the effects of different formulations of the food on a target organism, i.e. either a pathogen or a spoilage organism, during product development with a view to achieving an acceptable shelf life · validating processes such as aseptic processing and packaging that are intended to deliver some degree of lethality against a target organism or group of target organisms. In all situations, relevant expertise and skills together with the necessary laboratory facility must be available to produce meaningful results from challenge testing. When conducting a microbiological challenge test, a number of factors need to be carefully considered: · · · · · · ·

the selection of appropriate pathogens/spoilage organisms the level of challenge inoculum the inoculum preparation and method of inoculation duration of the study formulation factors and storage conditions sample examination data analysis and interpretation, and pass/fail criteria.

Useful and detailed guidelines for the design and planning of microbiological challenge testing are available (Anon., 2003; 2010; Notermans et al., 1993). 10.3.3 Quality shelf life and storage trials Ideally, storage trials aimed at establishing the quality shelf life of a food product can begin once its safety has been established. In practice, and more


340


often than not, storage trials will be run in parallel to food safety evaluation based on HACCP principles as required by law. While in principle shelf life storage trials should employ conditions that mimic those the product in question is expected to encounter during storage, distribution, retail display and consumer use, in practice, and in many small and medium-sized companies, a fully comprehensive storage trial is rarely possible as conditions during distribution and retail display, for instance, are difficult and expensive to reproduce. Consumer storage, handling and use, too, are often highly variable and unpredictable, and over which the manufacturer has little control. What the manufacturer must be certain about is the objective of the storage trial, which, after all, is a controlled experiment, and the manufacturer must be clear about what variables he can control and what he cannot. Storage conditions Storage conditions can be fixed or cyclical, or a combination of both. For a given set of storage conditions, the following variations should ideally be available (Man, 2002): · Optimum conditions: These are the most desirable conditions of temperature, humidity, light and so on under which the most optimistic shelf life data should be obtained. · Typical or average conditions: These are the conditions that are expected to be most commonly experienced by the product and under which shelf life data that apply to the bulk of future production should be generated. · Worst case conditions: These are the most extreme but not abuse conditions that the product is expected to encounter and under which the most conservative shelf life data should be obtained. The latter, if used to assign a shelf life, should give it a margin of safety ensuring that product failures due to insufficient shelf life are highly unlikely in practice. For cost reasons, storage trials tend to employ fixed conditions, which, in the absence of universal standards, commonly include: · Frozen: ÿ18 ëC or lower (relative humidity is usually near 100%). · Chilled: 0 to 5 ëC, with a maximum of 8 ëC (relative humidity is usually very high: ~90%+). · Temperate: 25 ëC, 75% relative humidity. · Tropical: 38 ëC, 90% relative humidity. · Control: control conditions (for storage of control samples) are usually the optimum conditions, be they ambient, chilled or frozen. Samples for storage trials As outlined in Section 10.2.5, there are product (e.g., raw materials, product composition) as well as external (e.g., packaging, processing) factors that can influence shelf life. As such, they need to be known, controlled and standardised across replicate storage trials or trials conducted during product development,



341

otherwise results from these trials could be misleading or even meaningless. A corollary of this is that every time a significant change is made in any of these factors, for instance, in the microbiological quality of a major raw material or in the time-temperature combination of a thermal treatment, fresh storage trials will have to be conducted. The number and size of samples to be laid down for trials need to be carefully chosen. The type of product, its end-use application, the anticipated or required shelf life and the tests planned for assessing changes during storage are some of the factors that need to be taken into account. Should there be great uncertainty about the shelf life, it is better to be generous with the number of samples retained than to run out of samples before the storage trial ends. Frozen storage at ÿ18 ëC or lower is often used as a means of keeping control samples. However, if freezing and thawing are known to affect the product adversely, facilities must be available for the preparation of fresh reference/control samples that are identical to the test samples, at any time during a storage trial. Experimental design and sampling schedule At present, there are few universally accepted protocols for storage trials for shelf life testing, be it legal or industrial. A number of designs have been put forward (Kilcast and Subramaniam, 2000), including some based on a statistical approach (Gacula, 1975). All have advantages and disadvantages, as well as varying implications on resources that include number of samples, storage facilities, development and maintenance of a trained taste panel and the amount of testing required. When conventional profiling is used to study sensory changes during storage, difficulty can arise due to the taste panel generating inconsistent responses over time, particularly if the storage time is long and test frequency low. Difficulties such as this further underline the importance of assuring the quality of stored samples, both test and control, if storage trials were not to produce at best inconclusive and at worst incorrect shelf lives. Nevertheless, the following are some possible protocols (Man, 2002): · Short shelf life products: For chilled foods with shelf life of up to one week (e.g., ready meals), samples can be taken off daily for testing. · Medium shelf life products: For products with a shelf life of up to three weeks (e.g., some ambient cakes and pastry), samples can be taken off on days 0, 7, 14, 19, 21 and 25. · Long shelf life products: For products with a shelf life of up to one year (e.g., some breakfast cereals and heat-processed shelf-stable foods), samples can be taken off at monthly intervals or at months 0, 1, 2, 3, 6, 12 and (perhaps) 18. The exact frequency will depend on the product and on how much is already known about its storage behaviour. Accelerated storage trials Sometimes, accelerated storage trials, mostly based on the Arrhenius equation (see Section 10.2.4), may be used to shorten the time required to estimate a shelf


342


life, which otherwise can take an unrealistically long time to determine. In principle, accelerated storage trials are applicable to any deterioration process, biochemical, chemical, microbiological or physical, that has a valid kinetic model. In practice, because of their obvious advantages over direct storage trials, validated accelerated storage trials may be viewed as commercially sensitive such that only a few are available in the literature. The latter include the following: · shelf life and safety of minimally processed CAP/MAP chilled foods over a limited temperature range (Labuza et al., 1992) · aspartame stability in commercially sterilised flavoured dairy beverages (Bell and Labuza, 1994) · accelerated storage of commercial orange juice in 1 litre TetraBrikTM (Petersen et al., 1998) · accelerated shelf life testing of whey-protein-coated peanuts (Lee et al., 2003). The limitations of accelerated storage tests are well known; they tend to be product-specific and their results have to be interpreted with care based on detailed product knowledge and sound scientific principles. Fuller accounts of the limitations are available (IFST, 1993; Mizrahi, 2000). Accelerated tests must not be mistaken for àbuse tests'. An accelerated test is only of value if the shelf life limiting mechanism of deterioration under accelerated conditions is the same as that under normal/ambient conditions, and the relationship between changes under accelerated conditions and those under normal storage needs to be confirmed and validated using food products of known quality. An accelerated storage model that has enjoyed notable commercial success and is widely used in the baking industry is called ERH CalcTM (Fig. 10.3). The model is part of a computer-based `Cake Expert System' for the baking industry originally developed by the UK Flour Milling and Baking Research Association (now part of Campden BRI). ERH Calc allows users to run simulations on flour confectionery formulations and rapidly calculate their theoretical equilibrium relative humidities (ERHs) and estimate their mould-free shelf lives. The latter, though, do not necessarily mean that the products themselves are organoleptically acceptable. Shelf life tests As pointed out earlier, besides food safety, an acceptable shelf life is expected to retain desired sensory, chemical, physical, functional or microbiological characteristics of the product and which, where appropriate, should comply with any label declaration of nutritional information throughout its shelf life. Thus, tests employed to measure shelf life tend to be product-specific, reflecting the quality characteristics of the product being studied. In a sense, the tests to be used are informed by the knowledge and understanding of the ways the food product deteriorates and spoils, including the mechanism of deterioration that is shelf life limiting. A systematic and structured approach based on the HACCP principles



343

Fig. 10.3 Predicting mould-free shelf life of baked goods using ERH Calc (reproduced with kind permission of Campden BRI).

has been used to implement a control system designed to prevent rancidity in confectionery and biscuits (Frampton, 1994). Essentially, this approach follows the same principles of HACCP; here a hazard is taken to mean a microbiological, chemical or physical agent in, or condition of, the food with a potential to cause it to deteriorate and spoil, terminating its shelf life. Applying the principles systematically leads to the determination of the critical control points at which control can be exercised and which are necessary to eliminate or delay the shelf life limiting hazard, preventing it from ending the required shelf life prematurely. Given the nature of the potential and possible hazards, the following types of tests can be used individually or in combination to measure the progress of shelf life: · · · ·

microbiological examination, including challenge testing chemical analysis physical/instrumental testing, measurement and analysis sensory evaluation.

Many shelf life studies together with the tests employed have been published in both the primary and secondary literature. Table 10.6 gives some examples that illustrate the specific tests used to measure shelf lives in light of the underlying mechanisms of deterioration.


Table 10.6 Examples of published shelf life studies and their tests Product

Storage conditions

Shelf life tests

Approximate shelf life

Reference


Orange juice (11.2 ëBrix) reconstituted from concentrate, then exposed to thermosonication and pulse electric fields

25 ëC

Total bacterial counts Conductivity Soluble solids pH Colour attributes (tristimulus colorimeter) on day 0, 14, 28, 84 and 168

168 days

Walkling-Ribeiro et al. (2009)

Fresh Pacific salmon slices treated with salts of organic acids

1 ëC

pH ATP breakdown products Total volatile base nitrogen (TVB-N) and trimethylamine (TMA) Sensory analysis of cooked slices on day 0, 3, 6, 9, 12 and 15

15 days

Sallam (2007)

2 1 ëC

Composition of gas mixtures pH of meat Colour instrumental measurement and metmyoglobin percentage Lipid oxidation analysis Counts of aerobic psychrotrophic flora Sensory evaluation on day 0, 4, 8, 12, 16 and 20

20 days

Martinez et al. (2006)

Fresh pork sausages packaged in various modified atmospheres


345

Assigning shelf life The main aim of a storage trial for shelf life is to find out as accurately as possible, under specified storage conditions, the point in time at which the product has become either unsafe or unacceptable to the target consumers, and if the product meets its shelf life objectives. In terms of microbiological safety and stability, the following should be useful in helping to fix an end-point for the shelf life of the food being studied: · Relevant food legislation, e.g. Commission Regulation (EC) 2073/2005 on microbiological criteria for foodstuffs. · Guidelines for the microbiological quality of some ready-to-eat foods (Gilbert et al., 2000) given by enforcement authorities or agencies in support of their work, e.g. those given by the UK Health Protection Agency (previously the UK Public Health Laboratory Service). · Guides on microbiological criteria for foods produced by independent food research associations such as Campden BRI (Voysey, 2007). · Current industrial best practice as published in the primary literature, which suggests probiotic functional foods and drinks should contain at least 107 live and active bacteria per g or ml for their functional claims to be maintained over the shelf life period (Birollo et al., 2000). · Predictive models, e.g. ComBase. Non-microbiological criteria that are used to set shelf life tend to be relatively more product-specific. In an ideal situation, these criteria are either contained in the original marketing brief or can be developed from it. Crucially, the criteria, be they physical, chemical or sensory, need to be correlated to the quality attributes that are critical to product acceptability/consumer requirements, and hence quality (as opposed to safe) shelf life and, where appropriate, they should be agreed between the manufacturer and its customer. Once product safety has been established, sensory evaluation is the most popular means by which the end of shelf life is determined. A detailed treatment of sensory evaluation to study shelf life, either using a trained panel, or a sample of consumers, is beyond the scope of this chapter. 10.3.4 A summary Success in determining the shelf life of a food product depends on the following factors: · confidence in assuring food safety · ability to define the critical quality characteristics that determine product acceptability and meet customer requirements · knowledge and understanding of the pertinent mechanisms of deterioration and spoilage including the shelf life limiting mechanism · adequate capability, either in-house or external, in terms of both technical know-how and appropriate resources (skilled staff, testing facility etc.), to measure shelf life either directly through storage trials or indirectly through prediction and estimation, or both.


346


Storage trials for shelf life determination are controlled experiments, to which the basic principles of experimental design that include use of control, randomisation and replication apply. An estimate of shelf life without an indication of its variability is of little value. A safe food product of acceptable quality that consistently pleases its consumers has its origin in good product design that must include carefully planned and professionally executed shelf life testing. Replication of the storage trial experiment on sufficient food samples of agreed and consistent quality is essential for the setting of reliable and reproducible shelf life.

10.4

Future trends

In the past two decades, as a result of major research efforts in a number of countries, notably the US, UK and Australia, coupled with ever-increasing power of personal computers, the use of Internet-based predictive microbiological models as an aid to HACCP and microbiological risk assessment has had a significant and positive impact on the management of microbiological safety of foods. Food safety, which includes chemical and microbiological safety, is of fundamental importance and will always remain so. Recent research has focused on sensory shelf life in an effort to maximise consumer acceptance and minimise food waste due to inaccurate shelf life or shelf life that is too conservative. Apart from catastrophic circumstances, food products do not usually fail all at once such that for a given product there is a distribution of shelf lives over time, and concomitantly, an increasing proportion of the consumers are expected to reject the product over the same period. Realisation of this has led researchers to use survival analysis statistics to study sensory shelf life of foods (Hough et al., 2003). Since then, Bayesian methods and the Arrhenius equation have been used separately to study sensory shelf life of foods and to analyse data based on consumers' acceptance or rejection of samples stored at different times and different temperatures, respectively (Luz Calle et al., 2006; Hough et al., 2006). The number of consumers necessary for shelf life estimations based on survival analysis statistics has also been determined in a simulation study that assumes a Weibull distribution for the data model (Hough et al., 2007). Advantages of applying survival analysis statistics to sensory shelf life estimations include relatively simple sensory work with say 50±100 consumers and that the estimations are based directly on consumer data. The disadvantages are that the underlying mechanism of deterioration that limits shelf life will not be provided by the consumer data if it is unknown, and specialised statistical software and expertise are required for the calculations and interpretation of the results (Hough, 2006). Even more recent research has begun to look at the possibility of integrating the modelling of safety and quality of foods, taking a complex systems approach to estimating shelf life (Martins et al., 2008). In the meantime, storage trials for



347

estimating shelf life remains a cornerstone of the shelf life determination of foods with which all responsible food businesses should be conversant.

10.5

References

(2003) Microbiological Challenge Testing. Comprehensive Reviews in Food Science and Food Safety, vol. 2 (Supplement), 46±49, IFT, Chicago, IL. ANON. (2010) Challenge testing protocols for assessing the safety and quality of food and drink. Guideline No. 63, Campden BRI, Chipping Campden, UK. BELL L N and LABUZA T P (1994) Aspartame stability in commercially sterilised flavoured dairy beverages. Journal of Dairy Science, 77, 34±38. BETTS G D, BROWN H M and EVERIS L K (EDS) (2004) Evaluation of Product Shelf-life for Chilled Foods. Guideline No. 46, Campden and Chorleywood Food Research Association, Chipping Campden, UK. BIROLLO G A, REINHEIMER J A and VINDEROLA C G (2000) Viability of lactic acid microflora in different types of yoghurt. Food Research International, 33, 799±805. CFA (2010) Shelf life of ready to eat food in relation to L. monocytogenes ± Guidance for food business operators, 1st edn. Chilled Food Association, Kettering, UK. CRAWFORD C (1998) The New QUID Regulations. Chandos Publishing, Oxford. ELLIS M J and MAN C M D (2000) The methodology of shelf-life determination. In: Shelf-life Evaluation of Foods, 2nd edn. Man D and Jones A (eds). Aspen Publishers, Gaithersburg, MD, pp. 23±49. EUROPEAN COMMISSION (2004) Regulation (EC) No. 852/2004 of the European Parliament and of the Council on the hygiene of foodstuffs. Official Journal of the European Union, 25 June 2004, L 226/3± L226/21. FRAMPTON A (1994) Prevention of rancidity in confectionery and biscuits ± a Hazard Analysis Critical Control Point (HACCP) approach. In: Rancidity in Foods, 3rd edn. Allen J C and Hamilton R J (eds). Blackie Academic & Professional, London, pp. 161±178. FSA (2005) General Guidance for Food Business Operators. EC Regulation No. 2073/ 2005 on Microbiological Criteria for Foodstuffs. Food Standards Agency, UK (www.food.gov.uk/). FSA (2010) Food Standards Agency guidance on the application of date marks in food. [Online]. Available from: http://www.food.gov.uk/consultations/consulteng/2010/ fsaguidanceappdatemarksfoodeng (accessed 1 April 2010). GACULA M C (1975) The design of experiments for shelf-life study. Journal of Food Science, 40, 399±403. ANON.

GILBERT R J, DE LOUVOIS J, DONOVAN T, LITTLE C, NYE K, BIBEIRO C D, RICHARDS J, ROBERTS D

and BOLTON F J (2000) Guidelines for the microbiological quality of some ready-toeat foods sampled at the point of sale. Communicable Disease and Public Health, 3(3), 163±167. HOUGH G (2006) How does survival analysis help us in estimating the probability of a consumer rejecting a stored product? In: Workshop summary: sensory shelf-life testing. Food Quality and Preference, 17, 644±645. Â MEZ G and CURIA A (2003) Survival analysis applied to sensory HOUGH G, LANGOHR K, GO shelf life of foods. Journal of Food Science, 68, 359±362. Â MEZ G (2006) Sensory shelf-life predictions by survival HOUGH G, GARITTA L and GO


348


analysis accelerated storage models. Food Quality and Preference, 17, 468±473. and CURIA A (2007) Number of consumers necessary for shelf life estimations based on survival analysis statistics. Food Quality and Preference, 18, 771±775. HUIS IN'T VELD J H J (1996) Microbial and biochemical spoilage of foods: an overview. International Journal of Food Microbiology, 33, 1±18. IFST (1993) Shelf life of Foods ± Guidelines for its Determination and Prediction. Institute of Food Science & Technology, London. JONES A A (2000) Ambient-stable sauces and pickles. In: Shelf-life Evaluation of Foods, 2nd edn. Man D and Jones A (eds). Aspen Publishers, Gaithersburg, MD, pp. 211± 226. KILCAST D and SUBRAMANIAM P (2000) Introduction. In: The Stability and Shelf-life of Food, Kilcast D and Subramaniam P (eds). Woodhead Publishing, Cambridge, pp. 1±19. LABUZA T P, FU B and TAOUKIS P S (1992) Prediction for shelf-life and safety of minimally processed CAP/MAP chilled foods. Journal of Food Protection, 55, 741±750. LEE S-Y, GUINARD J-X and KROCHTA J M (2003) Relating sensory and instrumental data to conduct an accelerated shelf-life testing of whey-protein-coated peanuts. In: Freshness and Shelf-life of Foods. Cadwallader K and Weenen H (eds). American Chemical Society, Washington, DC, pp. 175±187. Â MEZ G (2006) Bayesian survival analysis LUZ CALLE M, HOUGH G, CURIA A and GO modelling applied to sensory shelf life of foods. Food Quality and Preference, 17, 307±312. MAN C M D (2002) Shelf Life. Food Industry Briefing Series, Blackwell Science, Oxford. MAN C M D (2004) Shelf-life testing. In: Understanding and Measuring the Shelf-life of Food. Steele, R (ed.). Woodhead Publishing, Cambridge, pp. 340±356. MARTINS R C, LOPES V V, VICENTE A A and TEIXEIRA J A (2008) Computational shelf-life dating: complex systems approaches to food quality and safety. Food Bioprocess Technol., 1, 207±222. Â N J A and RONCALEÂS P (2006) Effect of varying MARTINEZ L, DJENANE D, CILLA I, BELTRA oxygen concentrations on the shelf-life of fresh pork sausages packaged in modified atmosphere. Food Chemistry, 94, 219±225. MIZRAHI S (2000) Accelerated shelf-life tests. In: The Stability and Shelf-life of Food, Kilcast D and Subramaniam P (eds). Woodhead Publishing, Cambridge, pp. 107± 128. MOSSEL D A A (1971) Physiological and metabolic attributes of microbial groups associated with foods. Journal of Applied Bacteriology, 34, 95±118. NOTERMANS S, IN'T VELD P, WIJTZES T and MEAD G C (1993) A user's guide to microbial challenge testing for ensuring the safety and stability of food products. Food Microbiology, 10, 145±157. PETERSEN M A, TéNDER D and POLL L (1998) Comparison of normal and accelerated storage of commercial orange juice ± changes in flavour and content of volatile compounds. Food Quality and Preference, 9 (1/2), 43±51. SALLAM K I (2007) Chemical, sensory and shelf life evaluation of sliced salmon treated with salts of organic acids. Food Chemistry, 101, 592±600. TSO (2006) The Food Hygiene (England) Regulations (SI 2006/14), The Stationary Office, London. UYTTENDAELE M, RAJKOVIC A, BENOS G, FRANCËOIS K, DEVLIEGHERE F and DEBEVERE J (2004) Evaluation of a challenge testing protocol to assess the stability of ready-to-eat HOUGH G, LUZ CALLE M, SERRAT C



349

cooked meat products against growth of Listeria monocytogenes. International Journal of Food Microbiology, 90, 219±236. VOYSEY P A (2007) Establishment and Use of Microbiological Criteria (Standards, Specifications and Guidelines) for Foods. Guideline No. 52, Campden and Chorleywood Food Research Association, Chipping Campden, UK. WALKLING-RIBEIRO M, NOCI F, CRONIN D A, LYNG J G and MORGAN D J (2009) Shelf life and sensory evaluation of orange juice after exposure to thermosonication and pulsed electric fields. Food and Bioproducts Processing, 87, 102±107.


11 Sensory evaluation methods for food shelf life assessment D. Kilcast, Consultant in Food and Beverage Sensory Quality, UK

Abstract: The shelf life of foods that are microbiologically stable is limited by changes in perceived sensory characteristics (appearance, texture and flavour), which in turn result from physicochemical changes in the product in question. The main sensory methods used in shelf life testing are described, together with the design of shelf life testing programmes and the ways in which shelf life data can be analysed and interpreted. Published standards relevant to shelf life testing are outlined, together with associated instrumental test methods. Key words: sensory test methods, sensory standards, ethical procedures, end point, analysis and interpretation. Note: This chapter is a revised and updated version of Chapter 4 `Sensory evaluation methods for shelf-life assessment' by D. Kilcast in The Stability and Shelf-life of Food, ed. D. Kilcast and P. Subramaniam, Woodhead Publishing Limited, 2000, ISBN: 978-1-85573-500-2.

11.1

Introduction

The various available definitions of shelf life present some difficulties to the food industry when investigating the shelf life of microbiologically stable foods, in which the shelf life-limiting factors are usually changes in sensory characteristics. The definition of shelf life from the Institute of Food Science and Technology in the UK (IFST, 1993) specifies a single sensory criterion, that `be certain to retain desired sensory characteristics'. This requires identification and measurement of `desired' sensory characteristics, but does not expand on the meaning of `desired'. This definition also implies that sensory characteristics


Sensory evaluation methods for food shelf life assessment

351

should not change over the shelf life of the product, but in practice most foods undergo deterioration following production, and this must be recognised by defining the ranges of required characteristics. Further, some notable foods, such as cheese and wine, undergo changes that generate desired product characteristics during storage. ASTM E 2454-05, which deals specifically with sensory shelf life, defines sensory shelf life as the `time period during which the products' sensory characteristics and performance are as intended by the manufacturer'. This definition itself adds little to the earlier IFST definition, but then the standard introduces more practical terminology, in which shelf life is described as the `time period that a product may be stored before reaching its end point', and then further defines the end point as the `point at which a product no longer meets predetermined criteria as defined by test data (for example, discrimination, descriptive, or affective, or a combination thereof)'. Three types of sensory shelf life end points are then described: `(1) the product's overall sensory profile has changed; (2) a product attribute(s) that is known or suspected to be key to the consumers' perception of the product has changed; and (3) the acceptability of the product is too low'. These concepts offer sensory analysts an improved springboard from which to design and operate sensory shelf life tests. When we eat food, we perceive a whole range of different characteristics relating to the appearance, flavour and texture of the food. Physiological differences between individuals result in a range of responses to these stimuli, and we must expect these different responses to be encountered within a given consumer population. Further, differences in ethnic and cultural backgrounds and in experiences of foods will broaden further the response of consumers to foods. In selecting and using sensory methods, we must be prepared not only to encounter and work within this wide response, but also to interpret data generated by sensory measurements in the context of the target consumer population. Changes in all the different sensory modalities can occur throughout the shelf life of foods. Appearance changes are commonly seen on storage of, for example, red meat (browning), fruit juices (darkening), dairy gels (syneresis) and emulsions (separation). Odour loss is a particular problem in products such as bread and coffee, whereas the development of off-odours is particularly important as an index of deterioration in many products. Odour changes are frequently accompanied by flavour changes, but flavour is a complex characteristic that is perceived in different ways and consequently flavour changes can occur independently of odour changes. Textural changes can be seen as positive (e.g., maturation and softening of fruit), but are more frequently deteriorative (e.g., staling of bread and loss of crispness in snack foods). A more complete set of examples of deteriorative changes in different product types is shown in Tables 11.1±11.5 in the Appendix. There is often a temptation to interpret measured sensory changes in terms of perceived quality, but this must be given careful consideration. In general, we dislike extremes, preferring intermediate levels of a sensory characteristic,


352


leading to an inverted-U relationship between liking and attribute intensity; simple linear relationships are not often seen within a typical consumer population, although different relationships can be seen in segmented populations. Assessment of sensory shelf life can therefore be approached in one of two ways: from measurement of sensory characteristics, or from measurement of consumer liking. In this chapter, the principles underlying the measurement of sensory characteristics will be described, together with practical measurement systems and the interpretation of the measured data in terms of sensory shelf life.

11.2

Principles of sensory evaluation

Human beings employ a range of senses in perceiving food quality. A schematic diagram of the main senses, and how they can interact, is shown in Fig. 11.1. The discussion below summarises these senses briefly. Fuller descriptions can be found in the references in Section 11.13. 11.2.1 The human senses The visual senses are particularly important in generating an initial impression of food quality that often precedes the input from the remaining senses. If the appearance of the food creates a negative impact, then the food might be rejected without the other senses coming into play at all. The visual sense is often equated only with colour, but provides input on many more appearance attributes that can influence food choice, for example size, shape, surface gloss,

Fig. 11.1 The main human senses, and how they interact.



353

and clarity. In particular, the visual senses can provide an early, and strong, expectation of the flavour and textural properties of foods. Taste (gustation) is strictly defined as the response by the tongue to soluble, involatile materials. These have classically been defined as four primary basic taste sensations: salt, sweet, sour and bitter, but these have now been joined by umami, the savoury sensation associated with monosodium glutamate. In some countries this list is extended to include sensations such as metallic and astringency. The taste receptors are organised groups of cells, known as taste buds, located within specialised structures called papillae. These are located mainly on the tip, sides and rear upper surface of the tongue. Sweetness is detected primarily on the tip of the tongue, salt and sour on the sides of the tongue and bitter on the rear of the tongue. Other oral surfaces have lower sensitivities. Taste stimuli are characterised by the relatively narrow range between the weakest and the strongest stimulants (ca 104), and are strongly influenced by factors such as temperature and pH. The odour response is much more complex, and odours are detected as volatiles entering the nasal cavity, either directly via the nose or indirectly through the retronasal path via the mouth. The odorants are sensed by the olfactory epithelium, which is located in the roof of the nasal cavity. Some 150± 200 odour qualities have been recognised, and there is a very wide range (ca 1012) between the weakest and the strongest stimulants. The odour receptors are easily saturated, and specific anosmia (blindness to specific odours) is common. It is thought that the wide range of possible odour responses contributes to variety in flavour perception. Both taste and odour stimuli can only be detected if they are released effectively from the food matrix during the course of mastication. The chemical sense corresponds to a pain response through stimulation of the trigeminal nerve. This is produced by chemical irritants such as ginger and capsaicin (from chilli), both of which give a heat response, and chemicals such as menthol and sorbitol, which give a cooling response. With the exception of capsaicin, these stimulants are characterised by high sensory thresholds. The combined effect of the taste, odour and chemical responses gives rise to the sensation generally perceived as flavour, although these terms are often used loosely. Texture is perceived by the sense of touch, and comprises two components: somesthesis, a tactile, surface response from skin, and kinesthesis (or proprioception), a deep response from muscles and tendons. For many foods, visual stimuli will generate an expectation of textural properties. The touch stimuli themselves can arise from tactile manipulation of the food with the hands and fingers, either directly or through the intermediary of utensils such as a knife or spoon. Oral contact with food can occur through the lips, tongue, palate and teeth, all of which provide textural information. The descriptions given above, whilst appropriate for the individual sensing modalities, fail to take into account their interactive nature. These interactions have been extensively reviewed by Cardello (1996). Colour, which is obviously an important appearance characteristic, can be shown to have an influence on


354


flavour perception. For example, Dubose et al. (1980) found significant increases in perceived flavour intensity in beverages with increasing colour intensity. Textural properties of foods have substantial effects on the perception of flavour, and sound emission from crisp and crunchy foods has been shown to be of great importance in the perception of their texture (e.g., Vickers, 1991). The importance of the interaction between the texture of foods and their perceived flavour can be clearly seen if the time course of events during food consumption is considered. As already indicated, strong expectations of the flavour and texture characteristics can be generated before the food is introduced into the mouth. As food enters the mouth, and is either bitten or manipulated between tongue and palate, catastrophic changes occur to the structure of the food that strongly influence the way in which tastants and odorants are released from the food. These processes can result in important effects on perceived flavour, and can produce substantial changes in flavour and texture quality if changes to food structure occur on storage. 11.2.2 Factors influencing the quality of sensory data The complex nature of food quality perception creates many difficulties for the sensory analyst, whose primary task is to use human subjects as an instrument to measure the sensory quality of foods. The factors that should be considered in assessing the performance of human subjects in this way are accuracy, precision and validity (Piggott, 1995). Sensory measurements are a direct measure of human response, and have an inherently higher validity than instrumental measures, which are nonetheless of value as a complement to sensory data in shelf life assessment. In measuring human responses, low precision must be expected, but variation can be reduced by careful selection of a range of human subjects who can produce a response with lower variability, and by extensive training. Improving accuracy (giving the correct answer without systematic error or bias) can be achieved by recognising the various sources of physiological and psychological biases that can influence human subjects. The effect of physiological differences between individuals can be reduced, but not completely eliminated, by careful selection procedures. Psychological factors can introduce systematic biases that might not be recognised. These include those arising from unwanted interaction between panellists, and those from more subtle sources. These can be greatly reduced by choice of sensory test procedure and by careful experimental design and operation of sensory test procedure. Such factors play a major role in generating sensory data that can be interpreted reliably in terms of shelf life.

11.3

Basic requirements for sensory analysis

In developing and implementing a high-quality sensory evaluation system a number of inter-related requirements can be defined; these are discussed below,



Fig. 11.2

355

The components of a sensory evaluation system.

and more detailed discussions can be found in standard texts, examples of which are referenced in Section 11.13. The requirements are shown schematically in Fig. 11.2. 11.3.1 Clear definition of the objectives of the sensory evaluation system Clear objectives are central to the establishment of any system that will be sufficiently accurate to measure the required sensory characteristics with the required precision, but with the important proviso that it should be practical and cost-effective. This is particularly important in shelf life assessments, in which repeated measurements over a period of time demand substantial resources and commitment. Large amounts of sensory data can be generated over the test period, and careful planning must be given to how these data are produced and handled if a meaningful interpretation is to be achieved. Problems commonly seen in industry include: underestimation of panellist requirements, including enforced changes in personnel over the test period; ambiguity in the type of sensory information to be generated; and absence of guidelines on the interpretation of storage changes in terms of shelf life. 11.3.2 Provision of a dedicated sensory testing environment A suitable environment is essential for generating high-quality sensory data with minimal bias. The environment is important not only in providing standardised working conditions for the assessors, but also in providing a work area for


356


sample preparation and for data analysis. The three main components of an ideal sensory evaluation environment are: · a preparation area of adequate size and appropriately equipped · a testing environment, adjacent to, but separated from, the preparation area · individual booths to eliminate panellist interaction. 11.3.3 Selection of suitable test procedures Many sensory test methodologies are available, but fall into two main classes, shown schematically in Fig. 11.3. · Analytical tests. These tests are used to measure sensory characteristics of products by providing answers to the questions: ± Is there a difference? ± What is the nature of the difference(s)? ± How big is(are) the difference(s)? · Hedonic/affective tests. These tests are used to measure consumer response to sensory characteristics of the products by providing answers to the questions: ± Which product is preferred? ± How much is it liked? The two classes comprise tests that satisfy completely different objectives, and which are subject to different operating principles. Analytical tests use human subjects as a form of instrument to measure properties of the food. Hedonic tests measure the response of consumer populations to the food in terms of likes or

Fig. 11.3 Main classes of sensory testing procedures.



357

dislikes. Different psychological processes are used for each type of test, and in general there is no simple linear relationship between the two types of data, an inverted U-curve being typical. Of great practical importance, the type and numbers of subjects used for the analytical and hedonic tests are quite different. Use of each test type for shelf life determination is described in more detail in subsequent sections. 11.3.4 Selection and training of suitable test subjects The subjects to be used are defined by the objective of the test and by the consequential choice of test. The number of subjects to be used depends on the level of expertise and training of the panellists. General guidance on numbers is given in ISO 6658 (2005), which also discriminates between assessors, selected assessors and experts, but for guidance on specific tests, relevant individual standards should be referred to. (Note: the term assessor is used throughout the ISO standards, but for the remainder of this chapter, the more common term panellist will be used.) Analytical tests Both discriminative and descriptive tests use small panels of panellists chosen for their abilities to carry out the tests. Guidelines for establishing such assessors are given in ISO 8586-1 (1993). A general scheme for establishing a panel requires the following steps: · Recruitment. Panellists can be recruited from within the company, or dedicated part-time panellists can be recruited from the local population (company employees should not be compelled to participate). · Screening. These preliminary tests are used to establish that sensory impairment is absent, to establish sensitivity to appropriate stimuli and to evaluate the ability to verbalise and communicate responses. These tests will depend mainly on the defined objectives of the sensory testing, but will typically consist of the following: ± the ability to detect and describe the five basic tastes: sweet, sour, salt, bitter and umami; these may be extended to cover metallic and astringent ± the ability to detect and recognise common odorants, especially those characteristic of the product range of interest ± the ability to order correctly increasing intensities of a specific stimulus, for example increasing sweetness or increasing firmness ± the ability to describe textural terms characteristic of relevant food types ± absence of colour vision deficiencies. Approximately 8% men, but only 0.4% women, suffer colour vision deficiencies. Tests can be carried out using Ishihara charts (available from opticians or booksellers). Selection of suitable panellists is usually made on the basis of a good performance across the entire range of tests, rather than excellence in some and poor response to others. If the panel is to be used for a specific purpose, then the tests relevant to that purpose can be weighted appropriately.


358


· Training. In the initial stages, training is limited to the basic principles and operations, following which further selection can be made. More closely targeted training can then be carried out using the products of interest and aimed towards the specific tests to be used in practice. · Monitoring. Close monitoring of panel performance is essential, and any drift that is identified must be corrected by retraining procedures. Hedonic tests Subjects (respondents) for hedonic tests are chosen to represent the target consumer population, and to reflect any inhomogeneity in that population. Consequently, they need to be used in sufficient numbers to give statistical confidence that they are representative, and they must be given the opportunity to behave as they would in a real consumption environment. In particular, they must not be selected on the basis of sensory ability and must not be given any training. Numbers in excess of 100 respondents are normally used. For the early stages of concept development, qualitative studies using focus groups with small numbers of respondents can be used, but the data generated should be treated carefully and conclusions must not be generalised. The same subjects must not be used for both types of test and, in particular, in-house staff must not be used to generate hedonic data that may be viewed as consumer-related. 11.3.5 Data handling, analysis and presentation Sensory experiments can generate large amounts of data, and reliable conclusions require validation using statistical techniques. Different types of sensory test procedures generally utilise specific analysis procedures, but, in the case of the more sophisticated profiling techniques, a wide range of options is available, ranging from basic univariate analysis to sophisticated multivariate analysis. Many statistical software packages are now available. The most sophisticated require a sound understanding of statistical principles, but more user-friendly packages are available that satisfy most requirements. In practice it is usual to find that no single package can cover the entire range of basic requirements. Clear and effective presentation of sensory data, including the results of statistical tests, is essential. Most standard spreadsheets are now able to offer a wide range of presentation possibilities for both univariate and multivariate data.

11.4

Discrimination tests

Discrimination tests are perceived as one of the easiest classes of sensory testing to apply in an industrial environment, and are consequently heavily used. The tests can be used in two ways: to determine whether there is an overall difference between two samples, or to determine whether one sample has more or less of a specific attribute than another. However, there are inherent limitations of such tests, and they are often overused in circumstances in which



359

alternative methods such as profiling would be superior. In addition to testing for a difference, new versions of ISO standards also permit testing for similarity (contrary to common belief, absence of a difference does not imply similarity). In general, however, uses of similarity testing are more limited in practice as a consequence of the need for higher numbers of assessors and a lack of agreement in the choice of statistical test criteria. In this section, the main types of test with practical value for shelf life assessment will be described. 11.4.1 Paired comparison test In the most common form of the test (less commonly referred to as the 2-AFC, alternative forced choice, test), two coded samples are presented either sequentially or simultaneously in a balanced presentation order, AB and BA (ISO 5495, 2007). There are two variations on the test. In the directional difference variant, the panellists are asked to choose the sample with the greater or lesser amount of a specified characteristic. The panellists are usually instructed to make a choice (forced-choice procedure), even if they have to make a guess. In the directional form, it is important that the panellists clearly comprehend the nature of the attribute of interest. 11.4.2 Duo-trio test In the most common variant of the duo-trio test, the panellists are presented with a sample that is identified as a reference followed by two coded samples, one of which is the same as the reference and the other different (ISO 10399, 2004). These coded samples are presented in a balanced presentation order, i.e. A (reference) AB A (reference) BA The panellists are asked to identify which sample is the same as the reference. The duo-trio test is particularly useful when testing foods that are difficult to prepare in identical portions. Testing such heterogeneous foods using the triangle test, which relies on identical portions, can give rise to difficulties, but in the duo-trio test it is possible to ask the question: `Which sample is most similar to the reference?' 11.4.3 Triangle test The panellists are presented with three coded samples, two of which are identical, using all possible sample permutations (ISO 4120, 2007): ABB AAB BAB ABA BBA BAA The panellists are asked to select the odd sample, preferably using a fixed-choice procedure. The increased number of samples compared with a paired com-


360


parison test can result in problems with flavour carry-over when using strongly flavoured samples, making identification of the odd sample more difficult. Difficulties can also be encountered in ensuring presentation of identical samples of some foods. 11.4.4 Difference from control test The panellists are presented with an identified control and a range of test samples. They are asked to rate the samples on suitable scales anchored by the points `not different from control' to `very different from control'. The test can be used to identify overall differences, or differences in specified attributes. The test results are usually analysed as scaled data. 11.4.5 Analysis of discrimination tests The basic principle underlying the analysis of difference is to test the actual response obtained against the response that would have occurred purely by chance: for the paired comparison and duo-trio tests this is 1 in 2, for the triangle test this is 1 in 3. One consequence of the different probabilities is that the statistical power of the tests differs, together with the numbers of responses that are needed in order to give a meaningful and reliable result. These numbers are related to the levels of risk that are deemed acceptable, and these are the Type 1 risk (incorrectly concluding that there is a difference that does not exist) and the Type 2 risk (not identifying a difference that is present). It is sometimes possible to generate the required number of judgements by replicated tests with a smaller number of panellists. Such a procedure should be used with care (e.g., generating 15 responses by using 3 panellists in 5 replicates is not recommended), and each replicate should be set up as an independent test. Individual ISO standards should be referred to for further details of minimum responses. The test results are usually analysed using tables of the binomial expansion, although other distributions have been used. The 5% level of significance is frequently used in sensory tests, but an increasingly common procedure is to calculate exact probability levels. If a strict statistical interpretation is required, a forced-choice response must be used. Similarly, if relatively inexperienced panellists, or consumers, are being used, then a forced-choice test must be used to prevent `fence-sitting'. However, if highly experienced panellists are used, a no difference response can be highly informative in specific circumstances. Extended variants of discrimination tests are often used, although some concerns have been raised in moving away from the simple test procedure. Descriptions of the nature of any difference can provide useful guidance for further testing. A simple scaled assessment of the degree of confidence in the decision (absolutely sure/fairly sure/not very sure/only guessed) is very useful, especially when using forced-choice procedures. Assessment of the degree of difference is only likely to be of value if panellists have been trained in scaling procedures. More controversially, panellists can be asked which samples they



361

prefer, but this type of procedure is of value only for crude guidance; preference tests should be set up separately as consumer tests. There is potential value in acquiring this information in shelf life assessments, but the hedonic information should be used with great care. All such information must only be used in support of the main difference test data, and can only be used from panellists who have given the correct response.

11.5

Quantitative descriptive tests

The major advantages of discrimination tests are their relative simplicity to set up and operate, and their high sensitivity. However, they have two important limitations. Firstly, only two sample treatments are compared together. Secondly, the information content of discrimination tests is limited, even when operated in an extended format, incorporating a range of questions. More informative tests can produce more quantitative data that can be subjected to a wider range of statistical treatments. 11.5.1 Scaling procedures Quantification of sensory data is needed in many applications, and the recording of perceived intensity of attributes or liking requires some form of scaling procedure. These procedures should be distinguished from quality grading systems, which are used to sort products into classes defined by a combination of sensory characteristics. Such systems are not open to quantitative numerical analysis. Scaling procedures are mainly used to generate numeric data that can be manipulated and analysed statistically. Before this can be carried out, however, thought must be given to how the scales used are seen and interpreted by the assessors, and how this may influence the type of analysis that can be safely applied. The different types of scale used are described below. · Category scales use a defined number of boxes or categories (often 5, 7 or 9, although other numbers are often used). The scale ends are defined by verbal anchors, and intermediate scale points are often given verbal descriptions. · Graphic scales (line scales) consist of a horizontal or vertical line with a minimum number of verbal anchors, usually at the ends. Other anchors can be used, for example to define a central point, or to denote the position of a reference sample. · Unipolar scales have a zero at one end, and are most commonly used in profiling, especially for flavour attributes. · Bipolar scales have opposite attributes at either end. Definition of the central point can often give rise to logical difficulties, as can ensuring that the extreme anchors are true opposites. This can be a particular problem for textural attributes, for example when using soft . . . . . . hard type scales. Bipolar scales are frequently used for consumer acceptability testing, especially using the like extremely . . . . . . dislike extremely format.


362


· Hedonic scales are use to measure consumer liking or acceptability. Category scales are usually used. · Relative to ideal scales are a type of hedonic scale which measures deviation from a personal ideal point. The type of scale used, and its construction, depends on a number of factors: · Purpose of test. Both category and graphic scales are commonly used with trained panels. In consumer testing, category scaling methods are usually used. · Expertise of assessors. Untrained assessors are generally poor discriminators, and can discriminate only over a small number of scale points. Trained panels can start with 5- or 7-point category scales, but, as their discrimination ability increases, they can use effectively more scale points or graphic scales. When using inexperienced assessors, scales incorporating a `neutral point', such as the central point in an odd-numbered category scale, are sometimes avoided in order to minimise the risk of `fence-sitting'. · Number of assessors. Using small assessor numbers with a low number of category scale points will limit statistical analysis options. · Data-handling facilities. Category scaling responses can be entered relatively quickly onto a spreadsheet, whereas data from line scales must be measured, and this can be a time-consuming procedure. Computerised data acquisition, either directly from a terminal or indirectly from optical readers, can avoid this problem. In practice, establishing a trained sensory panel can often proceed from a category scale with a small number of scale points (e.g. 5), through a category scale with more points (e.g. 9) to a line scale. Sensory analysts should be aware of difficulties that panellists have in using scales, and careful training is needed to ensure that scales are unambiguous and can measure the intended response. 11.5.2 Simple descriptive procedures Scaling may often be needed in order to quantify a single, well-defined attribute. However, it should be established that there is no ambiguity in the attribute of interest. This is particularly relevant during product development or modification, when the assumption that a process or ingredient modification will change only a single attribute is frequently violated. Such changes are especially common when textural changes are a consequence of process or ingredient modifications. If it is suspected that several attributes might be of interest, then the profiling procedures described in the subsequent sections should be considered. 11.5.3 Quantitative Descriptive Analysis (QDA) Variants of the original QDA procedures are probably used more than any other profiling procedure. The QDA technique uses small numbers of highly trained



363

panellists. Typically, 6±12 people are screened for sensory acuity and trained to perform the descriptive task to evaluate the product. Three major steps are required: development of a standardised vocabulary, quantification of selected sensory characteristics and statistical analysis of the results. Vocabulary development Development of the vocabulary is a group process for creating a complete list of descriptors for the products under study. Panellists freely describe the flavour, appearance, odour, mouthfeel, texture and aftertaste characteristics of different samples. No hedonic (good or balanced), general (full or typical) or intensitybased (strong or weak) terms are permitted. Terminology should be consistent from product to product and tied to reference materials. The references decrease panellist variability, reduce the amount of time necessary to train sensory panellists, and allow calibration of the panel in the use of intensity scales. References should be simple, reproducible and clear to the assessors, and illustrate only a single sensory descriptor. They can be single chemical substances or finished products, and are made available during both the training and the testing phase, at various concentrations or intensity. One requirement for the use of QDA in shelf life testing is the use of training samples that illustrate quality changes that occur on storage. This is often difficult to achieve in practice, especially for long shelf life foods. The attributes are collected and compiled into a master list. This individual preliminary evaluation of the samples may be revised during an open discussion to eliminate any redundant or synonymous descriptors. New terms might be added and physical references proposed. The panel leader condenses and formats the information into a proposal for standardised vocabulary. This vocabulary is then modified and improved in several interactive sessions. Multivariate statistical methods (e.g., factor analysis) are sometimes used to reduce the number of descriptors. Finally, definitions for the attributes are agreed. Intensity measurement Once agreement is reached on the vocabulary, further training is performed. The number of training sessions is dependent on the subject's performance, product and attribute difficulties and the time allowed for QDA testing. Panel training increases panellist sensitivity and memory and helps panellists to make valid, reliable judgements independent of personal preferences. Once the training sessions have established satisfactory panel performance, and removal of ambiguities and misunderstandings, the test samples can be evaluated. This is usually carried out in replicated (commonly three) sessions, using experimental designs that minimise biases. 11.5.4 The SpectrumTM method The SpectrumTM method resembles QDA in many respects; for example, the panel must be trained to fully define all product sensory attributes, to rate the


364


intensity of each and to include other relevant characterising aspects such as change over time, difference in the order of appearance of attributes, and integrated total aroma and/or flavour impact. However, the perceived intensities are recorded in relation to absolute or universal scales, which allow the comparison of relative intensities among attributes within a product and among products tested. Panellists develop their lists of descriptors by first evaluating a broad array of products that define the product category. The process includes using references to determine the best choice of term and to best define that term so that it is understood in the same way by all panellists. Words such as vanilla, chocolate or orange must describe an authentic vanilla, chocolate or orange character for which clear references are supplied. All terms from all panellists are then compiled into a list that is comprehensive yet not overlapping. The SpectrumTM method is based on an extensive use of reference points. The choice of scaling technique may depend on the available facilities for computer manipulation of data and on the need for sophisticated data analysis. Whatever the scale chosen, it must have at least two, or preferably three or five reference points distributed across the range. 11.5.5 Free choice profiling (FCP) Free choice profiling is a very different concept, which removes the need to generate a compromise consensus vocabulary (Williams and Langron, 1983), and which can also be used in consumer research (Jack and Piggott, 1992). Assessors are allowed to develop their own individual vocabularies to describe sensory perceptions of sample sets and to assign intensity scores to these attributes. As a consequence of removing the need to agree vocabularies, free choice profiling requires little training ± only instruction in the use of the chosen scale. Assessors merely have to be objective, capable of using line scales, and able to use their developed vocabulary consistently. Thus, assessors can be still regarded as representing naõÈve consumers. Characteristics being judged can be restricted by the panel leader, but the number of descriptors produced is limited only by the perceptual and descriptive skills of the assessors. A range of sensory characteristics such as appearance, flavour, aroma or texture can be examined. One particular advantage of the technique for shelf life assessment is that new attributes that develop on storage can readily be incorporated into the profile. Disadvantages include the need to use a complex statistical analysis technique (generalised Procrustes analysis) and the absence of any agreed terminology. 11.5.6 Time-related methods Time-intensity methods are used to measure intensity of a specific attribute as a function of time in the mouth, and have been used extensively to investigate the temporal behaviour of tastants, such as sweet and bitter molecules, and the release of volatile flavour materials from foods. Such studies are particularly



365

important in the reformulation of foods that results in structural modifications, and in changes that can occur on storage. These structural modifications are often accompanied by textural changes, and these often result in complex perceptual phenomena that are direct consequences of the changes in texture with time producing different flavour release phenomena. Although the use of time-intensity for flavour measurement is relatively well established, textural changes can also be monitored using the method. A major limitation of the time-intensity method is that only a single attribute can be tracked with time, or, with some software packages, two attributes. If several important attributes are thought to be time-dependent, separate sessions are needed for each attribute. Difficulties encountered in time-intensity profiling prompted the development of a hybrid technique, progressive profiling (Jack et al., 1994). In this technique, assessors carried out a profile on a set of texture descriptors at each chew stroke over the mastication period. Such a method has a number of potential advantages: several attributes can be assessed in one session; scaling is reduced to a unidimensional process; and the most important aspects of the shape of a time-intensity curve are retained. 11.5.7 Statistical analysis of scaled data Univariate procedures are the starting point for the analysis of any sensory data. The procedures can be used at different stages of a sensory programme, but are particularly useful in assessing the performance of panellists undergoing training for profile panels, and for exploratory investigation of scaled data. An important consideration in selecting appropriate analysis techniques is the nature and distribution of the data. Prior to the use of any statistical procedure, the form of the data should be examined by visualisation techniques, such as the use of scatter plots. Data that are not normally distributed are analysed by nonparametric methods. It is frequently assumed that sensory data are normally distributed, and that parametric tests can be used. The distribution of all sensory data should be examined, however, especially when relatively small numbers of responses are being used. If in doubt, non-parametric tests can be employed. The most commonly used procedures used to examine sensory data are t-tests, analysis of variance (ANOVA) and multiple comparison tests. The t-test procedure can be used to compare the mean scores from two samples, usually used in the paired format if the same panellists have assessed both samples. If more than two samples are to be compared, two-way ANOVA is used with the panellists and samples as factors. Panellist sample interactions are also usually examined. If significant differences for a given attribute are identified by ANOVA, multiple comparison tests can be used to identify which samples differ. In most applications of any form of sensory testing, the intensities of many attributes are being measured, leading to highly complex data sets. Multivariate analysis methods such as principal component analysis (PCA) are increasingly being seen as essential in interpreting such data sets, and several different uses are evident:


366


· Assistance in panel training, including assessment of panellist performance and reduction of attribute lists in forming profile vocabularies. · Simplifying the complexity of data presentation. Visualisation of the relationships between two attributes is easily accomplished, and visualisation of three attributes presents few difficulties, but greater numbers of attributes present substantial problems. · Identification of redundancy in the use of descriptive attributes. · Investigation of the underlying structure between products and between the attributes characterising them. · Construction of `maps' visualising the similarities and dissimilarities between products. As sensory shelf life tests are usually carried out in conjunction with instrumental measures of product change, statistical methods can be used to explore data relationships. Methods include correlation analysis, regression analysis and multivariate analysis. Further details are given in Kilcast (2010).

11.6

Consumer acceptability testing

Consumer tests give a direct measure of liking that can be used more directly to estimate shelf life. The most common procedure is to ask consumers representative of the target population to scale acceptability on a 9-point category scale, anchored from like extremely to dislike extremely. A minimum of 50 consumers should be used, and preferably well over 100, although lower numbers (32±40) have been reported. Suitable experimental designs should be used, in conjunction with appropriate statistical analysis. Other information on individual modalities (appearance, odour, flavour and texture) can also be obtained, together with attribute intensity information, but it is preferable to keep such tests simple and to focus on overall acceptability. The most common procedure for operating the tests is to recruit consumers from a convenient high street or mall location and to carry out the tests in a convenient hall. Alternatively, a mobile test laboratory can be used to increase the degree of control.

11.7

Operation of sensory shelf life tests

11.7.1 Selection of tests for shelf life assessment The choice of tests for shelf life assessments depends on the purpose of the assessment, and on the way in which the sensory storage changes are to be interpreted in terms of shelf life. Quality grading schemes are available for some foods, for example fish (MartinsdoÂttir, 2010), but cannot be regarded as suitable systems for the shelf life assessments of most foods. Difference tests can be used if the shelf life criterion is defined in terms of the first detectable change, but in general difference tests will detect changes that are small and of little relevance to shelf life. Consequently, most sensory tests employ quantitative measures of



367

change that are more open to interpretation in consumer terms. It is also possible to use hedonic tests to generate consumer acceptability directly. Such tests can be expensive, especially if used at repeated time points, and an alternative is to use quantitative sensory tests to measure change, and at critical change points to carry out consumer tests to evaluate the impact of the changes on consumer acceptability. 11.7.2 References for sensory shelf life assessment The variability of sensory data can be reduced substantially if a reference standard can be made available at each assessment session. Unless a very high level of panel training is feasible, memory of sensory quality is unreliable for most shelf life testing, especially over medium- to long-term storage periods, and reference samples should be provided for all tests. Ideally, a reference standard should be used from the same batch of product under test that can be stored under conditions in which changes do not occur. This is rarely achieved in practice, and more frequently it must be assumed that a stored reference undergoes quality change. Care must be taken to choose conditions that minimise the change. An alternative procedure is to manufacture a new reference for each test point. This is a valid procedure only in circumstances in which batch-to-batch variation is minimal; substantial variations will prejudice data interpretation. An increasingly common alternative to physical reference standard is a written specification (Beeren, 2010), generated by sensory techniques such as QDA. Whilst considerably superior to reliance on memory, successful use of such a standard requires extensive panel training and maintenance of a stable panel performance over the storage period. The problems described above are inevitably more serious in the case of shelf life tests carried out over long storage periods. 11.7.3 Ethical considerations Any sensory testing of foods must be carried out under a defined ethical policy for the use of human subjects, and general guidelines have been issued by the Institute of Food Science and Technology in the UK (IFST, 2005). This is particularly important in the case of storage testing, especially when the test protocol takes the products close to, or even past, the shelf life of the products. In particular, it is essential to assess any microbiological hazards that might be associated with testing, especially near the end of shelf life and under accelerated (especially elevated temperature) storage conditions. If necessary, microbiological testing should be carried out prior to sensory testing, preferably on the same samples to be used for sensory testing. Under no circumstances should samples of questionable microbiological quality be submitted for sensory testing. If there are any residual questions regarding microbiological quality, sensory testing should be limited to assessment of appearance and odour only.


368

11.8


Design of sensory shelf life tests

The experimental determination of shelf life can require a considerable amount of experimentation, with consequent costs and demands on time. Efficient design of such experiments is important if such tests are to be cost-effective. A statistical approach has been outlined by Gacula (1975), which describes a number of options for controlling the number of necessary measurements. In the most commonly operated type of test, a single batch of product (or replicate batches) is put on test at time zero, and samples are taken off for testing at intervals determined by the expectation of the likely shelf life (partially staggered design, Fig. 11.4a). If there is no prior knowledge of the shelf life, it may be necessary to take sufficient samples at each time point, therefore requiring extensive experimentation. In a variant of this procedure, the number of samples tested is increased up to the acceleration point, at which failure is expected, and after which a constant number of samples is tested. A further variant uses an expansion in sample numbers determined by the number of failed units. This basic type of design has the clear advantage that data related to shelf life are generated at intervals and build up to give a moving picture of deteriorative change. Whilst this carries few problems in circumstances in which instrumental measurements are the primary information source, problems are frequently encountered when sensory analysis techniques are being used to assess shelf life. This is related to the difficulties in generating consistent panel responses over time, these difficulties increasing over long storage times and if the test periods are infrequent. Several factors can contribute, mainly inconsistent use of scoring scales, changing panel composition and learning effects. If sensory profiling is the method of choice, the appearance of new attributes not present at the initial panel training stage (e.g., off-flavours) can give rise to difficulties. The ideal design for sensory testing would involve having all samples from all storage treatments and from all timepoints tested together in a balanced design. In principle, this can be achieved in three different ways: 1. Samples can be drawn from successive production batches and put into storage for an appropriate time. At the end of shelf life, all samples can be tested in an appropriate design. This design is of course susceptible to fluctuations in production quality, and can only be used in situations in which production consistency can be assured. 2. A more practical variant of this design is shown in Fig. 11.4b (drawn sample design), in which a single large batch is held under conditions under which quality changes are effectively zero, for example frozen storage. Samples are removed at appropriate intervals and stored under the desired conditions. 3. Another variant is shown in Fig. 11.4c (stored sample design). A large batch is put into storage, and samples are drawn at appropriate intervals and held under non-changing conditions (e.g. frozen) until the required storage time has been reached. The major difficulty in the last two designs is identifying appropriate non-changing storage conditions, as few foods can be stored in such a way without changes in some important quality attribute.



369

Fig. 11.4 Shelf life test designs: (a) partially staggered design, (b) drawn sample design, (c) stored sample design.


370


Whilst these last two designs all have the important advantage of delivering an internally consistent picture of changes in sensory characteristics, they suffer from a number of disadvantages. First, no information is generated on stability and shelf life until the very end of the storage trials, a situation which is unlikely to be tolerated in most industrial environments. Second, the setting up of the trials requires a prior knowledge of expected shelf life. Third, a large global test can impose severe logistical problems for both sensory and instrumental laboratory measurements.

11.9

Interpretation of sensory shelf life data

The various sensory test procedures generate information on whether changes are occurring, the nature of the changes that are occurring and the magnitude of the changes. Such information cannot be transformed into shelf life information unless two criteria are satisfied. First, the pattern of the changes must be understood, in terms of both the form and the direction of the change. Second, there must be a company policy on sensory quality that forms a framework within which the data can be interpreted. This is essential when interpreting analytical sensory data in terms of consumer response. These two issues are closely related, and are discussed subsequently. Important quality changes on storage are often assumed to be linear, but this is rarely the case in practice. It is also often erroneously assumed that any change represents quality deterioration, but this is clearly not the case with foods such as cheese, and beverages such as wines. Changes in product attributes with forms such as those shown in Fig. 11.5 are not uncommon, especially in the period immediately following manufacture. Clearly, the form of such changes must be known before any reliable interpretation can be made. The criteria that can be used for interpreting sensory shelf life data have been reviewed by Dethmers (1979), and fall into three categories: first detectable

Fig. 11.5

Illustrative changes in sensory attributes following manufacture.



371

Fig. 11.6 Non-linear attribute change: similar differences are seen at three timepoints A, B, C.

change, measured attribute change and change in consumer acceptability. The first detectable change (or just noticeable difference) in product quality can be measured using difference tests, assuming that a suitable reference sample is available. However, difference tests can be over-sensitive to changes that have little relevance to sensory quality as perceived by consumers, and give limited information on the nature of the change. An additional problem can be encountered when non-linear changes occur, as shown in Fig. 11.6. In this case, spot difference tests carried out at timepoints A, B and C would all identify the same level of difference. This illustrates an underlying problem with the use of difference tests, which is that a quantitative picture of change is rarely attainable. If quantitative measures of relevant sensory attributes are made, a fixed level of change can be used as a criterion. This is illustrated in Fig. 11.7(a) for two products showing a decreasing attribute intensity. The decrease of this attribute is faster for product 1, reaching a critical limit at a shorter storage time. The critical limit needs to be agreed as representing the end of shelf life. Figure 11.7(b) shows an analogous situation in which an attribute that is absent at the start of storage increases in intensity. This typifies the situation in which an offflavour develops on storage. Growth of a non-characteristic attribute is often more easily detected than decrease of a characteristic attribute, and is likely to be of great importance to consumer acceptability. An alternative approach to shelf life assessment is to measure consumer acceptability directly. Figure 11.8(a) shows how direct measurement of consumer acceptability can be used to compare the shelf life of two products. Greater difficulty in interpretation is encountered, however, when the changes in acceptability of two products of different initial quality are measured. This is illustrated in Fig. 11.8(b), in which product 1 represents an economy product, and product 2 a premium product. The use of a single critical acceptability level fails to recognise the different quality levels, and in these circumstances it may be preferable to define critical levels for each product that reflect its market.


372


Fig. 11.7 Level of change of a given attribute: (a) the attribute decrease is faster for product 1 than for product 2, and a critical intensity is reached more quickly, (b) the attribute increase is faster for product 1 than for product 2, and a critical intensity is reached more quickly. The curve form for product 2 is typical of off-flavour growth, for example rancidity.

11.10

Instrumental methods

Sensory measures of quality changes on storage are essential as the only valid reflection of perceived quality, but are expensive and time-consuming to operate. In addition, they suffer from high variability when carried out over long time periods, and need regular panel calibration, especially if the panel composition changes. If valid instrumental measurement methods are available, these can be of great value in augmenting the sensory data, although they are only rarely sufficiently reliable in replacing sensory data (e.g., Kilcast, 2010; Kress-Rogers, 1993). Their value can most clearly be found in long-term measurement of shelf life, which poses substantial challenges to the sensory analyst. 11.10.1 Appearance Overall appearance changes on storage can readily be tracked using either conventional or digital still photographs. This is a particularly powerful means of monitoring change in form of a product, and can be used to monitor visual



373

Fig. 11.8 Changes in consumer acceptability: (a) for products starting with the same acceptability, product 1 has a shorter shelf life than product 2, (b) for products starting from different acceptabilities, different critical levels should be used.

colour changes. However, accurate rendition of colour changes requires careful standardisation of lighting conditions and photographic technique, and ideally should be carried out by a professional photographer. Successful imaging of appearance has the benefit of providing accurate visual standards that are of great value in shelf life measurement. For colour assessment alone, many instruments are available that can give relevant measurements of product colour characteristics. In addition, extensive use is made of standard colour atlases, although there are problems in applying these to wide ranges of foods. Consequently, many sectors of the food and drinks industry have devised colour matching charts specifically for their own products. Colour measurement and the use of colour atlases are discussed in detail in MacDougall (2002) and Hutchings (1999). 11.10.2 Aroma and flavour The complexity of the flavour response presents enormous difficulties for those needing a rapid and simple assessment. Measurement of the wide range of volatiles that contribute to food flavour is technically feasible, but even the most sophisticated techniques, such as gas chromatography-mass spectrometry, carry


374


the risk of not identifying trace volatiles that have low detection thresholds. In principle, analysis of involatile tastants should pose a lesser problem, but, even though there are few basic tastes, the taste response can be stimulated by a wide range of food components. As a consequence, generalised analysis plays a limited role in shelf life assessment studies. If the deterioration mechanism is known, however, analysis for specific deterioration indicators, such as chemical compounds produced from rancidity development, can be highly effective. Contrary to popular expectations, the use of the so-called electronic noses and electronic tongues has been minimal in the food and beverage industries (Kilcast, 2010; RoÈck et al., 2008). 11.10.3 Texture Changes in physical properties that are perceived as textural changes can be measured using a range of techniques. Properties of fluid foods can be measured by a range of rheological techniques; properties of solid foods can be measured using mechanical techniques that typically measure force-deformation behaviour (Bourne, 2002; McKenna, 2003; Kilcast, 2004). Many of the techniques are capable of measuring change, but not necessarily change that is relevant to perceived texture. If a valid relationship can be established, such measurements can be a valuable adjunct to sensory testing.

11.11

Standardisation in sensory shelf life testing

The basic requirements for setting up and operating sensory testing procedures for foods and beverages are described in both standard texts (see Section 11.13) and in a series of Standards published by the International Standards Organisation (a full list of available standards can be found at http://www.iso.org/iso/ iso_catalogue/catalogue_tc/catalogue_tc_browse.htm?commid=47942, and some specific standards have been referred to earlier in this chapter). These provide general guidance on core sensory operations, but offer little practical advice on specific applications to activities such as shelf life testing. The American Standard (ASTM E 2454, 2005) is unique in its approach to sensory shelf life testing, in that in addition to describing appropriate sensory testing approaches, it also gives advice on possible decision criteria for establishing sensory shelf life of consumer products. The focus on the end point concept, as described in Section 11.1, forces the sensory analyst to give careful consideration to both the sensory changes that are likely to occur on storage, and to their importance to consumer perception of quality. Such focus is needed to avoid one of the most common problems in shelf life assessment, which is that studies are set up without any clear idea of the criteria that define end of shelf life, and of which measurements will be needed to generate data that relate to shelf life. Failure to define criteria and appropriate measurements very frequently leads to the generation over time of large numbers of measured parameters (both sensory



375

and instrumental) that, even with the availability of multivariate statistical techniques, leads to intractable analysis problems. Use of the end point concept should circumvent many of these difficulties. This standard also gives advice on the types of test design shown in Fig. 11.4, although the notation of designs is changed to `multiple-point' and `single-point' designs.

11.12

Future trends

Progress in sensory shelf life measurement in recent years has focused on the increasing use of sensory quality specifications and on the use of consumer acceptability data. Considerable effort continues into developing simpler, costeffective test methods, and into researching accelerated test methods with the intention of developing predictive sensory shelf life models. However, the high resource levels required to design reliable methods introduce the temptation to cut too many corners, with the consequence that unstable and unreliable models are generated. In addition, warnings have been given on the validity of predictive models using sensory data (Guerra et al., 2008). Other shelf life modelling procedures have also been investigated, for example those based on artificial neural network (ANN) algorithms (Siripatrawan and Jantawat, 2009). 11.12.1 Sensory specifications In the years following the publication of the first edition of this text, the use of sensory specifications has become increasingly widespread (Beeren, 2010). Numerous different approaches have been used, although at present there has been no attempt to produce a uniform, standardised procedure, mainly as a result (at least in the UK) of conflicting requirements from the retail sector. However, a well-designed sensory specification provides the basis for not only sensory quality systems that are stable and continuous over long time periods, but also offers similar stability for shelf life testing over long time periods, with obvious benefits when dealing with long shelf life foods. 11.12.2 Survival analysis As consumer quality requirements form the basis of the most generally accepted definitions of sensory shelf life, it is of little surprise that there have been major efforts in employing consumer data in the measurement of shelf life. Particular progress has been made through the use of statistical procedures based on survival analysis (e.g., Hough et al., 2003, Hough, 2010), in which samples with different storage times are presented to consumers. Consumers are asked a question such as `Would you normally consume this product? Yes or no?' and a survival function is defined as the probability of consumers accepting a product beyond a certain storage time. The technique has been applied to a wide range of long shelf life foods, and has given superior results to other techniques in studies on bread (GimeÂnez et al., 2007).


376


11.13


Literature on setting up and operating sensory evaluation systems is extensive. The following texts offer extensive information, but other literature is also available. Sensory evaluation

and HEYMANN H (1998). Sensory Evaluation of Food. Principles and Practices. Chapman & Hall. MEILGAARD M, CIVILLE C V and CARR B T (2006). Sensory Evaluation Techniques, 4th edn. CRC Press. Ä OZ A M, CIVILLE G V and CARR B T (1992). Sensory Evaluation in Quality Control. Van MUN Nostrand Reinhold. PIGGOTT J R (1988). Sensory Analysis of Foods, 2nd edn. Elsevier Applied Science. LAWLESS H T

Sensory perception BOURNE M C

Press.

(2002). Food Texture and Viscosity: Concept and Measurement. Academic

(1999). Food Colour and Appearance, 2nd edn. Springer. (1999). Food Texture: Measurement and Perception. Aspen. MACDOUGALL D B (2002). Colour in Food. Improving quality. Woodhead Publishing. MCKENNA B M (ed.) (2003). Texture in Food. Volume 1: Semi-solid foods. Woodhead Publishing. KILCAST D (ed.) (2004). Texture in Food. Volume 2: Solid foods. Woodhead Publishing. TAYLOR A J and ROBERTS D D (2004). Flavor Perception. Blackwell. HUTCHINGS J B

ROSENTHAL A J

Many contract laboratories offer sensory testing services that can be used for sensory shelf life assessment, but it is advisable to use those laboratories that also offer associated services relevant to shelf life measurement, in particular microbiological and physicochemical testing. In the UK, the main contract laboratories that can offer a full service package are, in alphabetical order: Campden BRI, Leatherhead Food Research, and Reading Scientific Services.

11.14

References

(2005). Sensory Evaluation Methods to Determine the Sensory Shelf Life of Consumer Products. ASTM E 2454-05. BEEREN C J M (2010). Establishing product sensory specifications. In Sensory analysis for food and beverage quality control, ed. D Kilcast, Woodhead Publishing, 75±96. BOURNE M C (2002). Food Texture and Viscosity: Concept and Measurement. Academic Press. CARDELLO A V (1996). The role of the human senses in food acceptance. In Food Choice, Acceptance and Consumption, ed. H L Meiselman and H J H MacFie, Blackie A&P. DETHMERS A E (1979). Utilizing sensory evaluation to determine product shelf life. Food Technology, September, 40±42. DUBOSE C N, CARDELLO A V and MALLER O (1980). Effects of colorants and flavorants on ASTM E 2454



377

identification, perceived flavor intensity and hedonic quality of fruit flavored beverages and cake. Journal of Food Science, 45, 1393±1415. GACULA M C (1975). The design of experiments for shelf life study. Journal of Food Science, 40, 399±403. GIMEÂNEZ A, VARELA P, SALVADOR A, GASTOÂN A, FISZMAN S and GARITTA L (2007). Shelf life estimation of brown pan bread: a consumer approach. Food Quality and Preference, 18, 196±204. GUERRA S, LAGAZIO C, MANZOCCO L, BARNABAÁ M and CAPPUCCIO R (2008). Risks and pitfalls of sensory data analysis for shelf life prediction: Data simulation applied to the case of coffee. LWT - Food Science and Technology, 41(10), 2070±2078. HOUGH G (2010). Use of survival analysis statistics in analyzing the quality of foods from a consumer's perspective. In Processing Effects on Safety and Quality of Foods, ed. E Ortega-Rivas, CRC Press. Â MEZ G and CURIA A (2003). Survival analysis applied to sensory HOUGH G, LANGOHR K, GO shelf life of foods. Journal of Food Science, 68, 359±362. HUTCHINGS J B (1999). Food Colour and Appearance, 2nd edn. Springer. IFST (1993). Shelf Life of Foods ± Guidelines for its Determination and Prediction. IFST, London. IFST (2005). Ethical and Professional Practices for the Sensory Analysis of Foods (http:// www.ifst.org/documents/policy_statements/practicesforsensoryanalysis_ policystat.pdf). IFST, London. ISO 8586-1 (1993). Assessors for sensory analysis. Part 1. Guide to the selection, training and monitoring of selected assessors. ISO 10399 (2004). Sensory analysis. Methodology. Duo-trio test. ISO 6658 (2005). Sensory analysis. Methodology. General guidance. ISO 4120 (2007). Sensory analysis. Methodology. Triangle test. ISO 5495 (2007). Sensory analysis. Methodology. Paired comparison test. JACK F R and PIGGOTT J R (1992). Free choice profiling in consumer research. Food Quality and Preference, 3, 129±134. JACK F R, PIGGOTT J R and PATERSON A (1994). Analysis of textural changes in hard cheese during mastication by progressive profiling. Journal of Food Science, 59(3), 539± 543. KILCAST D (ed.) (2004). Texture in Food. Volume 2: Solid Foods. Woodhead Publishing. KILCAST D (2010). Combining instrumental and sensory methods in food quality control. In Sensory Analysis for Food and Beverage Quality Control, ed. D Kilcast, Woodhead Publishing, 97±117. KRESS-ROGERS E (1993). Instrumentation and Sensors for the Food Industry. Woodhead Publishing. MACDOUGALL D B (2002). Colour in food. Improving quality. Woodhead Publishing. MCKENNA B M (ed.) (2003). Texture in Food. Volume 1: Semi-solid Foods. Woodhead Publishing. Â TTIR E (2010). Sensory quality management of fish. In Sensory Analysis for MARTINSDO Food and Beverage Quality Control, ed. D Kilcast, Woodhead Publishing, 293315. PIGGOTT J R (1995). Design questions in sensory and consumer science. Food Quality and Preference, 6(4), 217±220. È CK F, BARSAN N and WEIMAR U (2008). Electronic nose: current status and future trends. RO Chem. Rev., 108, 705±725. SIRIPATRAWAN U and JANTAWAT P (2009). Artificial neural network approach to simul-


378


taneously predict shelf life of two varieties of packaged rice snacks. International Journal of Food Science and Technology, 44, 42±49. VICKERS Z M (1991). Sound perceptions and food quality. Journal of Food Quality, 14(1), 87-96. WILLIAMS A A and LANGRON S P (1983). A new approach to sensory profile analysis. In Flavour of Distilled Beverages: Origin & Development, ed. J R Piggott. Ellis Horwood Ltd.

11.15

Appendix

Tables 11.1±11.5 show typical physicochemical and sensory factors that can change on storage and consequently limit the shelf life of different product types. Table 11.1 Fruit and vegetable products Product

Deterioration mechanisms

Limiting changes

Soft fruit

Enzymic breakdown Mould growth Moisture loss

Textural softening Visible mould Dry appearance

Hard fruit

Enzymic action Moisture loss

Textural softening, bruising Dry texture

Potatoes

Enzymic action Sprouting

Softening, poor cooking Sprouting, toxin production

Cucumber

Enzymic action

Loss of crispness, gross structure breakdown

Coleslaw

Moisture loss from vegetables Fat oxidation

Loss of viscosity in dressing, appearance changes, microbial growth Rancidity

Prepared salads

Moisture loss Oxidation

Loss of crispness, drying Browning

Fruit preserves

Syneresis Oxidation

Serum separation, mould growth Flavour loss

Dried fruit

Enzymic action Chemical reactions

Browning Flavour changes


Sensory evaluation methods for food shelf life assessment Table 11.2

Meat and meat products

Product


Limiting changes

Fresh red meat

Oxidation Microbial growth

Loss of red colour, rancidity Off-odours and off-flavours

Frozen meat

Oxidation Ice sublimation

Rancidity Freezer burn

Fresh fish

Microbial growth Chemical reactions

Microbial Off-odours Appearance changes

Fresh poultry

Microbial growth

Microbial Off-odours

Fresh sausages

Microbial growth Oxidation

Microbial Rancidity

Fresh bacon

Microbial growth Oxidation

Microbial Rancidity, colour change

Canned ham

Chemical reactions Can deterioration

Flavour loss Gas generation

Table 11.3

379

Beverages

Product


Limiting changes

Carbonated beverages

Gas evolution Hydrolysis/oxidation

Carbonation loss Flavour loss, off-flavours, rancidity

Beer

Oxidation Microbial growth

Off-flavours Turbidity

Coffee

Volatile loss Oxidation

Flavour change Rancidity

Fruit juices

Oxidation Enzymic reactions

Flavour and nutrient loss Cloud instability

Tea

Volatile loss Volatile absorption

Flavour loss Off-flavours

Wine

Oxidation

Off-flavours Colour change

Low-calorie soft drinks

Hydrolysis

Sweetness loss


380


Table 11.4 Cereal and other dry products Product


Limiting changes

Bread

Starch retrogradation Moisture migration Moisture uptake Oxidation Moisture loss Starch changes Microbial growth Starch changes Protein changes Moisture migration Starch retrogradation Oxidation Moisture uptake

Stale texture and flavour Dry texture, mould growth Loss of crispness Rancidity Drying and hardening Stale flavour and texture Mould formation Texture changes, breakage Staling Softening (cereal), hardening (fruit) Stale flavour and texture Rancidity Caking Non-enzymic browning Mould, microbial growth Flavour changes Colour loss Fat crystallisation (bloom) Texture changes Staling, rancidity Texture changes Rancidity

Snack foods Cakes Dried pasta Breakfast cereals Dry mixes Spices Chocolate confectionery Sugar confectionery

Microbial growth Volatile loss Chemical reactions Fat migration Oxidation Moisture uptake Oxidation

Table 11.5 Dairy products Product Ice cream


Moisture migration Oxidation Fluid milk Oxidation, hydrolytic reactions Microbial growth Dried milk powder Moisture uptake Oxidation Butter Oxidation Cheese Oxidation Lactose crystallisation Microbial growth Low-fat spreads Microbial growth Oxidation Yoghurt Syneresis Oxidation Fruit yoghurt Syneresis Oxidation Microbial

Limiting changes Ice crystal formation Rancidity Rancidity and other off-flavours Caking Flavour change, rancidity Rancidity Rancidity Gritty texture Mould Mould Rancidity Serum separation Rancidity Serum separation Rancidity Mould


12 Advances in instrumental methods to determine food quality deterioration F. Kong and R. P. Singh, University of California, Davis, USA

Abstract: Instrumental techniques are widely used to assess the changes in quality attributes and evaluate remaining shelf life of the foods. This chapter presents some of the major instrumental techniques for evaluating shelf life of food, including determination of color, appearance, texture, rheological properties, lipid oxidation, and microbiological analysis. Recent developments in instruments and their applications such as electronic nose, electronic tongue, and Infrared (IR) Spectroscopy are discussed along with the limitations of selected instrumental techniques. Key words: instrumental determination, shelf life, electronic nose, electronic tongue, infrared (IR) spectroscopy.

12.1

Introduction

During food storage, chemical, biochemical and physical deteriorative reactions can occur that cause changes in food color, appearance, texture, and flavor, significantly impacting the overall quality attributes and consumer acceptability of foods. On the other hand, microbiological forms of deterioration can also occur that cause food spoilage and safety issues. Although sensory evaluation is routinely used in the industry to evaluate the quality of foods, it can be expensive and time consuming. Alternatively, instrumental techniques are widely used to assess quality attributes and determine changes in quality to evaluate the remaining shelf life of a food.


382


Temperature and moisture are critical factors affecting the quality deteriorative reactions in foods. Food appearance, especially color, determines the first impression of consumers about food products. Texture and rheological properties such as viscosity significantly influence the taste of the food and handling characteristics of the product. All these factors may change during food storage and distribution. Lipid oxidation is one of the major spoilage reactions limiting the shelf life of foods. The unsaturated lipids in food can easily oxidize, resulting in alterations in smell, taste, texture, color and nutritional value. A number of chemical testing methods have been developed for determining the level of lipid oxidation by measuring primary and secondary lipid oxidation products. Volatile compounds are produced during lipid oxidation that are responsible for rancidity flavor. Gas chromatography (GC) is commonly used to differentiate and quantify volatile compounds. Rapid analytical techniques such as electronic nose and tongue, and infrared (IR) spectroscopy are becoming popular and attracting more attention of food processors. The advantages of these methods are their ability to provide rapid analysis and simultaneous evaluation of several parameters, and their potential for on-line or at-line use. Developments in computer science and chemometrics have enhanced the ability to analyse food quality. Instrumental methods usually have higher accuracy and reproducibility than sensory analysis. However, when instrumental techniques are used to measure sensory quality factors, they can be regarded as reliable only if the measured parameters can be correlated with the relevant sensory attributes (Kilcast, 2001). Although many instrumental methods are available to measure quality loss, some of the methods are more routinely applied in the food industry due to their simplicity, rapidity and convenience, such as color and texture determination. Other methods such as GC are used more often for research, because they are time-consuming and require expensive laboratory equipment and trained personnel. Different chemical reactions occur simultaneously during storage, but only the key reactions influencing major product quality attributes need to be measured during shelf life testing. These include crispness in biscuit, tenderness and drip loss in meat, and appearance (color, texture) in fruits and vegetables. Microbiological analysis is a primary indicator of safety in shelf life studies. The objective of this chapter is to provide a brief summary of the current status of a number of selected techniques that are often used to evaluate food quality with regard to shelf life. Key features of instrumental methods and their advantages and limitations are discussed.

12.2

Assessing food appearance

Color and other visual aspects of the appearance give consumers their first impression of the food significantly influencing their decision. The human eye perceives color as the reflected radiation in the visible region of the


Advances in instrumental methods

383

electromagnetic spectrum (400±700 nm). During storage, food color often degrades as a result of enzymatic and non-enzymatic reactions, oxidation, and other physical and chemical reactions. It is regularly used as a quality control measure of food during processing and storage. For fresh produce, color measurement is one of the few instrumental tests that gives results that correlate well with consumer assessment of quality (Aked, 2002). Kong and Chang (2009) and Kong et al. (2008) reported that soybean color can be used to predict the change of soybean quality during storage as well as tofu-making properties. Color can be assessed by using either a tristimulus colorimeter or a reflectance spectrophotometer to take color measurements of a single spot, and the data can be obtained in terms of tristimulus values, chromaticity coordinates, hue, and chroma (Clydesdale, 1998). A tristimulus colorimeter, such as the commonly used Minolta Chromameter, has three glass filters that correspond to the three primary colors in the spectrum (red, green and blue). The measured color data are expressed in a three-dimensional color space, L*, a* and b*, where the L* axis (luminance) indicates the brightness ranging from black to white, the a* axis ranging from green to red and the b* axis ranging from yellow to blue. Reflectance spectrophotometry determines the ratio of reflected light at specific wavelengths from a sample to that from a known reference standard. The spectrophotometer uses an integration sphere to collect light reflected from the sample, and normalizes the light to the source light of the reflectance. The light is calibrated with a pure white standard (100% reflection) and a black box (zero reflection) over the entire wavelength spectrum in the visible region. Results are expressed as a percentage, displayed as a graph showing reflectance versus wavelength. Different wavelength intervals may be used, typically 10 nm or 20 nm. Results from the spectral data can be converted to colorimetric values in the L*a*b* system. Spectrophotometry is more expensive than a simpler colorimetry, but it measures full reflectance curves, and is able to make colormatching and define tolerance volumes (Brimelow and Joshi, 2001). Machine vision system (MVS), also called computer vision system, is being increasingly used to assess food color and appearance. A MVS incorporates a camera to capture an image, and a computer to process and analyze the image, facilitating objective and nondestructive assessment of visual quality characteristics in food products (Brosnan and Sun, 2004). Using a camera, MVS can measure the color of the whole product surface, and is able to characterize food size, shape, roughness of surface, and defects (Chen et al., 2002; Kong et al., 2007). It has been successfully adopted to analyze the quality of various food products, including meat, fish, pizza, cheese, bread, and grain (Chen et al., 2002; Davies, 2009; Moreda et al., 2009). MVS-based online sorting systems have been developed to inspect, grade, and classify fruits, vegetables and fish (Brosnan and Sun, 2004). It is also used to locate bruising in fruits. In particular, spectroscopy and hyperspectral imaging have emerged as powerful techniques in that they can detect subtle and/or minor features and constituents in the products that are only sensitive at specific wavelengths (Van Zeebroeck et al.,


384


2007; Chen et al., 2002). Moreover, when applying radiation of various wavelengths to food material, it will penetrate into the food, thus exposing the internal structure and fractures (Davies, 2009). Near infrared (NIR) and time resolved spectroscopy are used to detect internal defects such as cracks and hollow regions in melon, and brown heart in pears. Magnetic resonance imaging (MRI) and X-ray imaging techniques can also be used to evaluate global internal quality with better resolution, but are so far limited by the high equipment costs (Hertog et al., 2008).

12.3 Measurement of relative humidity (RH), moisture, and water activity (aw) The moisture content and water activity of stored food products are among the most important factors affecting food quality stability. The differences between food water activity (aw) and the relative humidity (RH) of the ambient environment will cause food to absorb or lose moisture to the ambient environment. The moisture change in the food will alter its texture. An increase in moisture content will increase molecular mobility in food and induce microbial growth. The methods for determining RH, moisture, and aw are closely related. A variety of methods have been explored over many years to obtain meaningful humidity measurements. These methods are employed either by probing the fundamental properties of water vapor or using various transduction methods which are capable of giving humidity-related measurements. Some of the available instruments include mechanical hygrometer, based on the use of materials which expand or contract in proportion to the humidity change, and chilled mirror hygrometer based on an optical technique for the determination of the dew point temperature. The chilled mirror hygrometer is known to provide the most accurate and reliable measurements, and is often used for measurements setting a calibration standard (Yeo et al., 2008). Wet and dry bulb psychrometry is also used as a simple and relativity low cost method. It consists of two thermometers, one of which measures the temperature of the sampled air (dry bulb temperature) and the other, covered with a damp wick, determines the wet bulb temperature. Absorption types of optical hygrometers, including infrared and ultraviolet hygrometers, are available based on water vapor absorption of radiation in certain optical wavelengths (Wiederhold, 1997). Electronic sensors are commonly used for RH and aw determination. Some of the most popular sensors for humidity measurements are capacitive- and resistive-based humidity sensors. Capacitive-based humidity sensors incorporate a polymer capacitor in a measuring chamber, and its capacitance can change as a function of humidity. Resistive-based sensors rely on hygroscopic materials such as conductive polymers whose conductivity changes with absorption of moisture. A conductivity hygrometer is also available, which measures electric impedance of a liquid hygroscopic substance (e.g., salt solutions) affected by the relative humidity or water activity (Chen and Lu, 2005). Recently, with the



385

advent of optical fibre technology, a considerable level of research has focused on fibre-optic (FO)-based techniques for humidity sensing. FO humidity sensors employ materials whose optical property changes with humidity change (Yeo et al., 2008). As water activity of food is equal to equilibrium relative humidity divided by 100, the methods for RH measurement are also used for water activity determination, mostly involving the use of capacitance or dew point hygrometers (Mathlouthi, 2001). Moisture determination in food can use direct methods such as hot air oven drying, which is based on the weight loss of the sample when water is removed. Karl Fischer titration determines the water content in a sample based on an iodine/iodide redox reaction. It is not affected by volatile compounds, thus is advantageous compared with determination based on weight loss. Indirect measurement of moisture content includes measurement of the electrical resistance of a sample. In particular, infrared (IR) spectroscopy is increasingly used to quantify moisture of foods as a rapid and non-destructive approach. The principle of IR spectroscopy will be covered in Section 12.9.

12.4

Texture evaluation

Texture is one of the most commonly used physical indicators of food quality. Textural change may occur in stored food due to moisture migration, enzymatic hydrolysis, and other physical or chemical deteriorations that make food unacceptable for consumers. For example, fish muscle may become tough as a result of frozen storage, or soft and mushy as a result of autolytic degradation. Bread staling may occur due to moisture migration which causes a firming crumb and softer crust (Singh and Anderson, 2004). Szczesniak and her co-workers were the first to establish the relationship between the mechanical properties of a food and its texture profile (Friedman et al., 1963). They developed the method of texture profile analysis (TPA), in which a Texturometer was used to conduct a double-compression test to obtain a force±displacement curve (Fig. 12.1). A number of texture features can be obtained from the TPA profile, including hardness, cohesiveness, viscosity, elasticity, adhesiveness, brittleness, chewiness, and gumminess. TPA analysis is still frequently referred to in the literature as a standard method for texture characterization. Since Szczesniak's work, a variety of instrumental methods have been developed for texture measurements based on the mechanical parameters of food as determined by using the stress±strain or force±deformation relationship. A large number of texture measurement devices are available for different products for quality control purposes. The most popular ones include Texture Technologies' TA.XT2 Universal Texture Analyzer (Kong et al., 2007) and Instron universal testing machine (Yuan and Chang, 2007). Most measurements are based on empirical methods, in which the measured variables and procedures are from practical experience that are related to some aspect of textural quality for a


386


Fig. 12.1

A typical force±displacement curve obtained from a TPA test.

certain range or specific products (Kilcast, 2001). Examples include using Magness±Taylor testers to assess fruits, Warner±Bratzler shear fixture to measure meat tenderness, and FMC Pea Tenderometer to grade pea quality. These devices measure the resistance of food during penetration, shearing, or compression by the machine. Extrusion devices are also used in which the food is forced through one or more orifices. Kong et al. (2007) developed a multipleblade texture probe to measure shear force of fish muscle as a representative indicator of tenderness. The results of empirical methods are ill-defined physical properties, i.e., the measured properties are dependent on the method of measurement. On the other hand, fundamental methods have been developed that measure well-defined physical properties of food. In these cases, the measured properties are independent of measurement method, thus they can be measured with general-purpose testing machines. The commonly measured fundamental parameters include Young's modulus, shear modulus, and bulk modulus (Kilcast, 2001). Physical measurement of textural characteristics can be of practical value only if it is shown to relate to some relevant sensory textural property. Statistical techniques are used to analyze and describe relationships between instrumental data and sensory texture profiling scores. Significant correlation has been reported between sensory and instrumental testing of texture for some foods and selected texture parameters (Adhikari et al., 2003; Rahman et al., 2007). However, the correlation is often dependent on the methods used in the measurements, such as type of probe, cross-head speed, sample position and alignment, and distance of penetration or strain. Therefore these methods must be chosen carefully to obtain the best possible correlation with sensory measurements. An extensive review of the principles and applications of texture measuring methods was published by Bourne (2002). Advances in food research and practice have resulted in new approaches for texture determination. Acoustic property of food during fracture corresponds closely to human perception of food crispness and crunchiness; therefore sound



387

signal of foods during crushing can be used to evaluate crispness (Chen, 2009; Juodeikiene and Basinskiene, 2004; Liu and Tan, 1999). Sonic resonance testing is used as a non-destructive measurement to evaluate firmness, rigidity and the ripening for fruits and vegetables (GoÂmez et al., 2005). Spectroscopic and related methods are also developed for texture assessment. For example, nearinfrared (NIR) spectroscopy has been used to assess firmness of apples and other fresh fruits. The mechanism involves the backscattered radiation spectrum which is affected by both the scattering and absorption properties of the tissue, thus providing information about the physical structure (Hertog et al., 2008). Multivariate statistical techniques such as partial least squares regression are needed to analyze the relationship between the backscattered radiation spectrum and the texture. Mid-infrared (MIR) spectroscopy coupled with chemometrics has been successfully used to predict instrumental texture and meltability attributes of processed cheese samples (Fagan et al., 2007; Qing et al., 2007). A recent development in textural instruments is the application of the electromyography (EMG) technique. An acoustic-EMG system quantifies food texture by monitoring the mastication process through imaging techniques such as real-time magnetic resonance imaging (MRI), and acoustic sensors to record the auditory signals produced during mastication. This technique makes it possible to correlate food physics with the physiology of oral processing and food sensory perception, with the potential to become an alternative to texture profile analysis (TPA) and sensory texture measurements (Chen, 2009; Jessop et al., 2006).

12.5 Evaluation of rheological properties of liquid and semisolid foods Liquid and semi-solid food materials are generally non-Newtonian in nature, that is, the determined viscosity value is dependent on the shear rate. The majority of food materials display shear-thinning behavior. The Power Law model is commonly used to describe the relationship between viscosity and shear rate, from which the flow behavior index and consistency index can be derived. The rheological properties for liquid and semisolid foods are characterized in terms of viscosity, flow behavior index, consistency index and yield stress, which may experience significant change during storage. For example, the flow behavior index of concentrated milk changes significantly with storage time (VeÂlez-Ruiz and Barbosa-CaÂnovas, 1998). The complex viscosity of yogurts increases during storage due to the increase in lactic acid and production of exopolysaccharides (Saint-Eve et al., 2008). Fundamental methods determining food rheological properties are conducted by applying a well-defined stress on a food sample and measuring the shear rate or alternatively by measuring the developed shear stress on the food in a range of well-defined shear rates. The capillary viscometer is the simplest form of viscometer. The principle involves measuring the time taken for a fixed volume


388


of the test fluid to pass through a length of capillary tubing. Rotary viscometry is the most commonly used instrument for testing food rheological properties, in which a sample is loaded between two surfaces, and one of which undergoes an applied rotary motion. The geometry of these surfaces can be a pair of parallel plates, cone and plate, or concentric cylinders. Based on how the rotating surface is controlled, these instruments may be rate- or stress-controlled (McKenna and Lyng, 2003; Steffe and Daubert, 2006). For more details, the reader is referred to the books by Bourne (2002) and Rao (2007). When taking measurements, it is important to select a range of shear rates covering the whole scope of interest, as power law fluids may not exhibit the same behavior over the entire range of shear rates. Empirical methods of measurement are also widely used for determination of rheological properties and quality control. Foods with non-homogeneous complex structures unsuitable for fundamental methods are commonly measured with empirical methods. They are used to obtain an index of product rheology, correlate with results from sensory analysis and sometimes even considered as official identification standards. For example, dough testing equipment is commonly used to determine flour specifications, such as the strength and mixing properties of dough, by using specialized testing including farinography, mixography, extensiography, and alveography. Other examples include rapid visco analysers, viscoamylographs, falling ball viscometers, Bostwick consistometers, Adams consistometers, Zhan viscometers, and Hoeppler viscometers. Description of these devices can be found in the books by Steffe (1996) and Rao (2007).

12.6

Assessing lipid oxidation

Lipid oxidation is one of the major forms of spoilage in foods resulting in development of objectionable flavors and odors known as òxidative rancidity', thus leading to degraded quality and reduced shelf life of the product. Food lipids are composed mainly of triacylglycerols, which are esters of three fatty acids and a glycerol molecule. The fatty acids vary in chain length, degree of unsaturation and position on the glycerol molecule. When oxygen is present, lipid oxidation may occur consisting of a series of complicated autocatalytic processes, as shown by: Reactants (unsaturated lipids and O2) # Primary products (hydroperoxides and conjugated dienes) # Secondary products (ketones, aldehydes, alcohols, hydrocarbons) Peroxides (R±OOH) are primary reaction products formed during the initial stages of oxidation. They easily decompose into the secondary products, most of which, especially aldehydes, have strong off-flavors. In addition to rancid flavor, lipid oxidation also causes loss of vitamins, formation of potentially toxic



389

compounds, and eventually unacceptability of the food. The oxidative stability of an oil or fat depends on the degree and nature of the unsaturation of its triglycerides, its antioxidants content, the presence of pro-oxidants such as trace metals (copper and iron), and storage conditions (such as temperature and light). A number of methods have been developed to characterize the extent of lipid oxidation in foods. The methods of peroxide value (PV) and thiobarbituric acid (TBA) are mostly used for general quality estimation. Other methods measure the content of conjugated dienes or use p-anisidine. Gas chromatography (GC) is commonly used for headspace analysis of the volatile oxidation products giving results that correlate well with sensory evaluation, but the method requires access to gas chromatographic equipment, and therefore it is more often used in research. 12.6.1 Peroxide value The peroxide value (PV) determines the concentration of hydroperoxide, the primary oxidation products. The principle involves peroxides liberating iodine from potassium iodide, i.e. ROOH KI ÿ! ROH KOH I2 The amount of ROOH is then determined by measuring the amount of iodine formed, which is done by titration with sodium thiosulfate and using a starch indicator: I2 starch 2Na2S2O3 (blue) ÿ! 2NaI starch Na2S4O6 (colorless) The amount of peroxides is calculated back by the amount of sodium thiosulfate (Na2S4O6) consumed. It is expressed as peroxide value (PV) in units of milliequivalents (meq) peroxide per 1 kg of fat extracted from the food. A general rule is that PV should not be above 10±20 meq/kg fat to avoid rancidity flavor (Connell, 1975). 12.6.2 Thiobarbituric acid (TBA) The peroxide value only indicates the amount of primary products. Because hydroperoxides are easily broken down into secondary products, the PV value may not reflect the whole extent of lipid oxidation. TBA is used to quantify the secondary oxidation products. Specifically, TBA measures the concentration of malonaldehyde, a secondary reaction product and a reactive aldehyde. During the TBA test, the thiobarbituric acid reacts with malonaldehyde resulting in the formation of a pink colored complex. The intensity of this pink color is determined by measuring its absorbance at 540 nm with UV-visible spectrophotometer, which is directly related to the concentration of malonaldehyde in the original sample. For this reason, this test is also referred to as the thiobarbituric acid reactive substances (TBARS) method. The results are expressed as micromoles malonaldehyde present in 1 g of fat. Foods with TBA above 1±2 mol MDA-equivalent per g fat will probably have rancid flavour (Connell, 1975).


390


12.6.3 Conjugated dienes This method measures primary oxidation products; therefore it is only useful for monitoring the early stages of lipid oxidation. Along with the formation of peroxides, conjugated dienes containing conjugated double bonds (C=C±C=C) are formed based on the non-conjugated double bonds (C=C±C±C=C) that are present in the natural unsaturated state. The conjugated dienes strongly absorb ultraviolet radiation at 233 nm. Thus oxidation can be simply measured by dissolving the lipid in a suitable organic solvent and measuring the change in its absorbance with time using a UV-visible spectrophotometer. Conjugated dienes are broken down into secondary products in the later stages of lipid oxidation, leading to a decrease in absorbance. As some secondary products also have this conjugated structure and will contribute to the absorbance, the method is less specific than PV measurement. 12.6.4 p-anisidine Aldehydes deriving from the secondary oxidation of fats can react with panisidine to give products that absorb at 350 nm. By using a UV-visible spectrophotometer to measure the absorption, one can estimate the amount of secondary oxidation products of fat. The result is expressed as p-anisidine value or AnV. Measurements of p-anisidine value are commonly combined with peroxide value measurements to describe the total extent of lipid oxidation, as expressed by `Totox value'. Totox value is an empirical parameter equivalent to the sum of the p-anisidine value plus twice the peroxide value. 12.6.5 Analysis of volatiles with GC Gas chromatography (GC) is the most powerful method to identify and quantify individual aroma components, and monitor volatile lipid oxidation products of food. It is commonly used to quantify the secondary oxidation products including aldehydes, ketones, alcohols, short carboxylic acids and hydrocarbons. Some of these volatile compounds are highly specific to the oxidative degradation of a particular polyunsaturated fatty acid family. For example, hexanal, a main aldehyde formed during the oxidation of linoleic, gammalinolenic and arachidonic acids, is often measured as a good marker of oxidative rancidity (Pastorelli et al., 2006, 2007). The GC approach consists of three steps: recovery of volatile components, separation using GC column, and detection using mass spectrometry (MS), flame ionization detection (FID), or olfactory method. Volatile oxidization products can be recovered by means of extraction, for example, simultaneous steam distillation extraction (SDE) is one of the commonly used extraction methods. During SDE, a sample is distilled and the volatiles are collected in the extraction solvent. The solvent is then dried using a drying agent and the volatile components are concentrated by slow evaporation. The concentrated volatile analytes are then injected into the GC column. Compared to SDE, headspace



391

analysis is more rapid and labor saving. Headspace analysis can be performed by static headspace (SH), dynamic purge-and-trap headspace (DH) or solid phase microextraction (SPME) techniques. SH involves collecting volatile compounds evaporated from a sample placed in an airtight vial. The headspace volatiles are harvested after equilibrium is reached, and are then injected into the GC column. This method can only quantify a fraction of the target compounds and the sensitivity is limited (Plutowska and Wardencki, 2008). DH involves using inert gas continually purging sample to extract volatile compounds, and the volatile analytes are then collected by passing the gas effluent through a porous polymer trap. This method can yield a high concentration of volatile compounds thus leading to high sensitivity (Plutowska and Wardencki, 2008). SPME utilizes adsorptive polymeric film to absorb volatile analytes, and the analytes are then released into the GC column. It is a simple, effective tool for detecting low levels of flavor compounds in foods and beverages, requiring less complex equipment than SH and DH (Laguerre et al., 2007a). In a gas chromatography column, the volatiles are separated, and the separated analytes are labeled and quantified by a detector. Flame ionization detector (FID) or mass spectrometry (MS) are the most often used detectors. GC-MS allows one to quantify aldehydes in g kg±1 concentrations (Varlet et al., 2007). GC-olfactometry is also used to determine the active odor compounds, in which sensory evaluation of the elute from the chromatographic column is conducted by a trained panel. This method involves installation of an olfactometric port that allows the sample to be split into two parts, for detector and for sensory evaluation, respectively (Fig. 12.2). GC-olfactometry allows qualitative and quantitative evaluation of the odor for each analyte leaving the chromatographic column, which helps to determine the intensity of the odor and understand the sensory properties of a certain compound at a given concentration (Plutowska and Wardencki, 2008). As MS is capable of interference-free detection and quantitation of each individual compound in a complex sample, a recent advance in the analysis of volatile fractions of foods is to directly couple MS with static headspace (SH),

Fig. 12.2

Scheme of the gas chromatograph equipped with the olfactometric detector.


392


dynamic headspace (DH), or solid-phase microextraction (SPME). In this way, the GC separation step is omitted, leading to a fast characterization of food volatile fractions, and a rapid classification or prediction of quality factors. The mass spectrum obtained without prior chromatographic separation forms a `spectral fingerprint' from which qualitative or quantitative information can be extracted to predict food quality, using multivariate analysis, such as principal component analysis, linear discriminant analysis, partial least squares, and soft independent modeling of class analogy (Laguerre et al., 2007b). This method has been used to characterize cheeses (Peres et al., 2001) and study off-flavors in milk (Marsili, 1999).

12.7

Electronic nose

Electronic nose (EN) is an artificial olfactory system based on the GC volatile methods. It can detect and recognize a wide spectrum of odor patterns, and determine the odor intensity of mixtures of a variety of volatile oil degradation compounds. An EN can function as a rapid and non-destructive tool for on-line flavor characterization, especially, for rancidity analysis of foods during storage. The application of EN in the food industry has been increasing due to its rapidity, cost-effectiveness, objectivity and simplicity. An EN is composed of three elements: a sample handling system, a detection system, and a data processing system. The sample handling system introduces the volatile compounds present in the headspace of the sample into the detection system, by using the method of static headspace (SH) technique, dynamic headspace (DH) techniques, or solid-phase microextraction (SPME). The detection system consists of an array of gas sensors, which are electronic chemical sensors based on conducting polymers, metal oxides, surface acoustic wave devices, quartz crystal microbalances, or combinations of these devices. Of the various types of sensor, those based on metal oxides appear to be most suitable for the discrimination of different stages of lipid oxidation, and hence for shelf life prediction (MuÈller and Steinhart, 2007; Vinaixa et al., 2005). In the data processing system, responses generated by each sensor from the detection system are subject to analysis by pattern recognition (PR) techniques, in which principal component analysis or artificial neural network are commonly employed (Peris and Escuder-Gilabert, 2009). The partial specificity of gas sensors toward volatile components and the appropriate PR system make EN capable of recognizing simple or complex odors, and characterizing and discriminating products by their volatile components. Electronic noses have been commercialized. Some of the major manufacturers include Win Muster Airsense (WMA) Analytics Inc. (Schwerin, Germany), Alpha M.O.S. (Toulouse, France) and Cyrano Sciences Inc. (Danbury, CT, USA) (Tamaki et al., 2008). Figure 12.3 shows a portable electronic nose (PEN2) produced by WMA Analytics Inc. PEN2 consists of a sampling apparatus, a



Fig. 12.3

393

Schematic diagram of electronic nose measurement.

detector unit containing the array of sensors, and pattern recognition software (Win Muster v.3.0) for data recording. The sensor array system is composed of 10 metal oxide semiconductors (MOS) with different specificity for volatile compounds. EN systems are gaining applications in the food industry in process monitoring, shelf life investigation, freshness evaluation, authenticity assessment and other quality control studies. It is used to detect lipid oxidation of foods and change in aroma, in, for example, wine (GarcõÂa et al., 2006), meat (Tikk et al., 2008; Vestergaard et al., 2007) and fruits (Infante et al., 2008; Saevels et al., 2004). It is also used to monitor quality change and evaluate shelf life in nut (Pastorelli et al., 2007), olive oil (Mildner-Szkudlarz and Jelen, 2008), and cheese (Limbo et al., 2009). Other uses include characterization and classification of wines (Buratti et al., 2004; GoÂmez et al., 2006), and determination of fruit ripeness such as tomato (GoÂmez et al., 2008), mandarin (GoÂmez et al., 2007) and apple (Peris and Escuder-Gilabert, 2009). In addition to differentiating volatiles, EN can be used to assess other quality properties such as texture. This is done through multivariate statistical analysis to correlate electronic nose signals and quality indicators measured by other instruments. For example, good correlation was obtained between electronic nose signal and apple firmness (Brezmes et al., 2001; GoÂmez et al., 2008). Recently, a new type of EN system based on mass spectrometry has been investigated. As mentioned earlier, SPME-MS provides rapid classification or prediction of volatile components. This methodology is used in EN as an alternative design (Marsili, 1999). Similarly, multivariate statistic analyses (MVS) including principal component analysis are required to analyze signals from MS to differentiate between samples (Mildner-Szkudlarz and Jelen, 2008; Pastorelli et al., 2007). This new mode of EN composed of SPME-MS-MVS has been used to discriminate types of off-flavor problems in milk (Marsili, 1999) and apple (Saevels et al., 2004). The main problem with EN technology is èlectronic drift' (Hertog et al., 2008), referring to the instability of signal over time. It is caused by physical


394


changes in the sensors and the effects of the environment as well as the composition of the gaseous sample. One possible solution to this problem is to view sensor arrays as time-varying dynamic systems, and track the variation using adaptive estimation algorithms (Dutta et al., 2003).

12.8

Electronic tongue

Corresponding to electric nose, èlectric tongue' (ET), also called artificial tongue, has been developed to detect taste and olfaction in foods. By mimicking the human tongue to differentiate tastes of sourness, saltiness, bitterness, sweetness and umami, ET is capable of both qualitative recognition and quantitative determination of taste. ET is based on an array of sensors displaying high cross-sensitivity to various substances in aqueous media (Li et al., 2006), which allows fast recognition and classification of multiple components, as well as quantitative determination of concentrations of these components. Compared with a sensory panel, the advantage of ET lies in its cost-effectiveness, rapidity, small sample volume requirement, objectivity, and ease of use. The basic structure of ET is shown in Fig. 12.4. The core part of the ET is the array of non-specific chemical sensors with a high cross-sensitivity. Crosssensitivity means that the sensor responds not to a single analyte but to several substances simultaneously present in the analyzed media. Two commonly used sensor types are potentiometry and voltametry. The potentiometric sensors are more selective than a voltametric sensor that makes data interpretation easier to operate; while the voltametric technique has higher sensitivity, versatility, and robustness. The digitized signals are recorded in the computer, and processed by statistical software to interpret the sensor data into taste patterns. ETs have been used in various areas in the food industry, and have proven to be successful in discrimination and classification, quality evaluation and control, process monitoring and quantitative analysis of foodstuffs and beverages. It has been applied to classify types of wines and discriminate varieties of apples (Buratti et al., 2004). Beullens et al. used an ET originally developed at Saint Petersburg University to classify tomato cultivars (Beullens et al., 2008; Vlasov et al., 1994). The system contained a sensor array of 18 potentiometric chemical

Fig. 12.4 Schematic diagram of electronic tongue.



395

sensors. These sensors exhibit sensitivity to organic acids and minerals. A conventional Ag/AgCl electrode was used as a reference. When the sensors were exposed to tomato juice, equilibrium was obtained within 3 minutes and the chemical composition of the sample was determined. The 18 potential values were recorded in computer data files. Multivariate statistical data analysis techniques including principal components analysis, canonical discriminant analysis and partial least squares regression were used to quantitatively relate the taste compounds to the sensory panel scores and to classify tomato cultivars based on similarity in taste profile. The result showed that the electronic tongue was very well suited to classify tomato cultivars. Furthermore, the ET was suitable to quantify individual sugars, acids and minerals in a complex mixture (Beullens et al., 2008). Other researchers also found good correlation between instrument output and sensory descriptors pertaining to the global quality of a food sample (body, overall quality, and astringency) (Rudnitskaya et al., 2009). Recent reviews in this subject include Li et al. (2006) and Ampuero and Bosset (2003). EN and ET are sometimes combined as fusion sensors for food flavor detection, simulating the coexistence of the two sensory functions in humans (Li et al., 2006). Buratti et al. (2004) reported that combination of EN and ET in classification of Barbera wines gave 100% correct assignation. However, Cosio et al. (2007) reported that the ET did not seem to improve classification performance when electronic nose and electronic tongue are both used to evaluate olive oil samples stored under different conditions and periods (Cosio et al., 2007). Therefore, the success of combining electronic nose and tongue may be more dependent on the specific product concerned and storage conditions involved.

12.9

Infrared (IR) spectroscopy

Infra-red light is part of the broad spectrum of energy known as electromagnetic radiation. Within the infrared wavelengths of light, the waveband between 4 and 400 cm±1 is categorized as far infrared, between 400 and 4000 cm±1 is categorized as mid-infrared (MIR) and between 4000 and 14 000 cm±1 is categorized as near infrared (NIR). Identification of compounds in food by IR spectroscopy is based on the property of molecules to absorb the infrared light and experience a wide variety of vibrational motion characteristic of their composition. When coupled with chemometric data analysis techniques, NIR and MIR spectroscopy are rapid techniques that possess potential selectivity for screening products for qualitative attributes. IR spectroscopy in the mid- and near-infrared regions has become a powerful, fast, and non-destructive tool, and is widely used for quantitative analysis and quality evaluation of foods. The basis of spectroscopic techniques to study chemical composition of the food relies on wavelength-dependent interaction of light with the food material. In IR spectroscopy, a beam of infrared light passes through the sample.


396


Radiation interacting with a sample may be absorbed, transmitted or reflected. The reflected or transmitted radiation can be measured, and the spectra contain information of molecular groups and chemical composition. Water, sugar, acids and a range of other organic substances absorb near infrared (NIR) light in proportion to their concentration. These absorption features make NIR spectroscopy able to discriminate the constituent parts of a foodstuff, and assess the concentration of each constituent. Absorption bands in the MIR region are generally due to intra-molecular phenomena and specific molecular composition and structure. Analysis of a food sample using the MIR spectrum (4000± 400 cmÿ1) reveals information about the molecular bonds present and can therefore give details of the types of molecules present in the food. Compared with MIR radiation, NIR can typically penetrate much further into a sample, which makes it suitable for probing bulk material with little or no sample preparation. The NIR spectrometer is generally composed of a light source, a monochromator, a sample holder or a sample presentation interface, and a detector, allowing for transmittance or reflectance measurements (Fig. 12.5). The light source is usually a tungsten halogen lamp or laser emission diode (Huang et al., 2008). The monochromator instrument may be a grating or a prism used to separate the individual frequencies of the radiation either entering or leaving the sample. The wavelength separator rotates so that the radiation of the individual wavelengths subsequently reaches the detector. Detector types include silicon, lead sulfide and indium gallium arsenide, with different size, speed, sensitivity and signal-to-noise properties (Reich, 2005). As radiation may be absorbed, transmitted or reflected, there are different measurement modes fitting different applications, and the two most frequently used are diffuse transmittance and diffuse reflectance (Huang et al., 2008). With the improvement of computer technology, Fourier transform infrared (FTIR) spectrometry has been developed and increasingly used in food

Fig. 12.5 Basic NIR spectrometer configurations.



397

applications. Instead of a monochromator used for a NIR spectrometer, in FTIR, the IR light is guided through an interferometer to generate modulated light. By collecting interferograms of a sample, the FTIR spectrometer measures all infrared wavelengths simultaneously, rather than individual wavelengths of the radiation each time for NIR. The interferogram is then converted into a conventional spectrum using the Fourier transform algorithm. FTIR spectrometers are cheaper than conventional spectrometers but more rapid in response since information at all frequencies is collected simultaneously. A critical part in IR spectroscopy is to build a reliable and stable calibration model through using chemometrics (multivariate statistical techniques) to analyze a great deal of spectral data, i.e., to extract the information of quality attributes from the NIR spectrum (NicolaõÈ et al., 2007). To obtain accurate and reproducible calibration models, partial least squares regression and principal component regression are frequently used to estimate the component concentration and chemical and physical properties from the infrared spectra. Classification methods such as soft independent modeling by class analogy, K-nearest neighbors and artificial neural networks are powerful tools for the characterization, differentiation and classification of complex spectral data (Shiroma and Rodriguez-Saona, 2009). IR spectroscopy is a well-established technology and its use started 50 years ago. Now it is commonly used for rapid analysis of moisture, protein and fat content of a wide variety of agricultural and food products. It is a major method for determining authenticity of food products such as fats and oils, soluble coffee, green coffee and fruits. Rudnitskaya et al. (2006) reported that FTIR with attenuated total reflection (ATR-FTIR) can discriminate apples of varieties Jonagold and Golden Delicious despite the quite similar composition of these two apple varieties (Rudnitskaya et al., 2006). IR spectroscopy allows the measurement of many important parameters within a short period of time. For example, with NIR and MIR, moisture and fat content in potato chips can be determined within 5 min, compared to the 10±16 h required for conventional methods (Shiroma and Rodriguez-Saona, 2009). This feature, plus its noninvasive and non-destructive nature, makes the IR spectroscopic technique very suitable for online monitoring processes. The NIR technique is widely accepted as one of the most promising on/in-line process control techniques detecting fat, moisture, and protein content in meats, fruit and vegetables, grain and grain products, milk and dairy products, and beverages and other products (Huang et al., 2008). IR spectroscopy, coupled with the use of chemometric techniques, provides a reliable, accurate method for predicting the shelf life of foods under different storage conditions. The chemical and physical deteriorative reactions in foods during storage cause changes in a number of quality attributes that are reflected in IR spectra, thus IR spectrometry is suitable for evaluating food shelf life, including evaluation of loss of freshness, and onset of spoilage of various foods (Farkas and Dalmadi, 2009). IR spectroscopy was used to study shelf life of green asparagus (SaÂnchez et al., 2009), Crescenza cheese (Cattaneo et al., 2005),


398


ricotta cheese (Sinelli et al., 2005), and fresh-cut pineapple (Di Egidio et al., 2009). New applications are being developed with IR spectroscopy. As the propagation of NIR radiation in food is affected by their microstructure, NIR spectroscopy is used to measure microstructure-related attributes, such as stiffness and internal damage, and even sensory attributes (NicolaõÈ et al., 2007). Other developments in NIR spectroscopy include multi- and hyperspectral imaging techniques to attain both spatial and spectral information from food material, thus obtaining a global measurement of quality attributes (Gowen et al., 2007). In addition, variations on spectroscopic techniques, such as time resolved spectroscopy and space resolved spectroscopy, have been developed to interpret the backscattered radiation spectrum differently, which may lead to novel and better calibration models for various fruit quality attributes including texturerelated attributes (NicolaõÈ et al., 2007).

12.10

Microbiological testing

Routine microbiological testing of food has traditionally involved enumeration of total numbers of organisms by direct microscopic methods or viable counts. These conventional methods are time consuming and labor intensive, and provide limited information about the behavior of microorganisms in food that may be important for assessing food quality and predicting shelf life. In recent years, much interest has been focused on development of rapid and automated methods in microbiology. These include automation and mechanization of traditional methods to facilitate easy handling of large numbers of samples and processing data using computational statistical techniques (White, 1993). A variety of rapid and sensitive methods have been developed. One of them is to use a special culture medium, which incorporates bio-chemical reaction substrate, antibody, fluorescence, reaction substrate, or enzyme substrate to the culture medium, enabling a better separation and identification of target microorganisms (Atlas, 1993). Another fast developing area is immunological techniques, which discriminate the bacteria by the differential combination reactions of antigen and anti-body involving immune magnification through immunofluorescence and irradiation immunity (Yali et al., 2009). Traditionally, the microorganisms in food have been studied by culture-based methods. The limitation of this method is its inability to detect non-culturable cells and failure to characterize minor populations of microorganisms. Cultureindependent techniques are thus developed, such as the polymerase chain reaction (PCR) technique based on the amplification of polymerase chain reaction and detection of nucleic acids. Compared with traditional culture-based methods, these new methods are generally faster, more specific, more sensitive and more accurate, and are now increasingly applied in food microbiology (JusteÂ et al., 2008).



12.11

399

Future trends

An important trend in shelf life evaluation is to develop rapid, simple, low-cost analytical tools that are suitable for on-line monitoring or quality control (Farkas and Dalmadi, 2009). Examples include machine vision system, electronic nose and tongue, and FTIR spectroscopic methods discussed in this chapter. Similar approaches include the use of chemical `markers', which employ biosensors and other types of sensors to track and analyze shelf life deterioration. Readers can obtain information on these topics from numerous literature resources, including Kress-Rogers (2001). These methods are rapid, non-destructive, non-invasive and cost-effective, representing the direction of instrumental development for assessing quality changes and enhancing control during processing and storage of foods. They are also suitable for on-line or rapid at-line measurement of quality attributes relevant to shelf life of foods which are of increasing importance to the food industry.

12.12


Additional references in the area of instrumentation for shelf-life evaluation are listed below: (2006). Rapid methods of assessing as a tool for quality improvement and standardization of food products. Electronic Journal of Polish Agricultural Universities 9(3), #07. MAN, C.M.D., JONES, A.A. (1999). Shelf Life Evaluation of Foods. Gaithersburg, MD: Aspen Publisher Inc. Â ZARO, D., LOMBARD, B. ET AL. (2007). Trends in analytical methodology in RODRIÂGUEZ-LA food safety and quality: monitoring microorganisms and genetically modified organisms. Trends in Food Science & Technology 18(6): 306±319. STEELE, M. (2004). Understanding and Measuring Shelf-life of Food. Boca Raton, FL: CRC Press. KONIECZNY, P., BILSKA, A., UCHMAN, W.

12.13

References

(2003). Textural characteristics of low fat, full fat and smoked cheeses: sensory and instrumental approaches. Food Quality and Preference 14(3): 211±218. AKED, J. (2002). Maintaining the post-harvest quality of fruits and vegetables. In: Jongen, W. (ed.), Fruit and Vegetable Processing: Improving Quality. Cambridge: Woodhead Publishing. AMPUERO, S., BOSSET, J.O. (2003). The electronic nose applied to dairy products: a review. Sensors and Actuators B: Chemical 94(1): 1±12. ATLAS, R.M. (1993). Handbook of Microbiological Media. Boca Raton, FL: CRC Press. ADHIKARI, K., HEYMANN, H., HUFF, H.E.

Â ROS, P., VERMEIR, S., KIRSANOV, D., LEGIN, A., BUYSENS, S., CAP, N., BEULLENS, K., MEÂSZA NICOLAIÈ, B.M., LAMMERTYN, J. (2008). Analysis of tomato taste using two types of

electronic tongues. Sensors and Actuators B: Chemical 131(1): 10±17.


400


(2002). Food Texture and Viscosity, Food echnology, International Series, 2nd edn. San Diego, CA: Academic Press. BREZMES, J., LLOBET, E., VILANOVA, X., ORTS, J., SAIZ, G., CORREIG, X. (2001). Correlation between electronic nose signals and fruit quality indicators on shelf-life measurements with pinklady apples. Sensors and Actuators B: Chemical 80(1): 41±50. BRIMELOW, C., JOSHI, P. (2010). Colour measurement of foods by colour reflectance instrumentation. In Kress-Rogers, E. and Brimelow, C.J.B. (eds), Instrumentation and Sensors for the Food Industry, 2nd edn. Boca Raton, FL: CRC Press. BROSNAN, T., SUN, D.-W. (2004). Improving quality inspection of food products by computer vision ± a review. Journal of Food Engineering 61(1): 3±16. BURATTI, S., BENEDETTI, S., SCAMPICCHIO, M., PANGEROD, E.C. (2004). Characterization and classification of Italian Barbera wines by using an electronic nose and an amperometric electronic tongue. Analytica Chimica Acta 525(1): 133±139. CATTANEO, T.M.P., GIARDINA, C., SINELLI, N., RIVA, M., GIANGIACOMO, R. (2005). Application of FT-NIR and FT-IR spectroscopy to study the shelf-life of Crescenza cheese. International Dairy Journal 15: 693±700. CHEN, J. (2009). Food oral processing ± a review. Food Hydrocolloids 23(1): 1±25. CHEN, Y.-R., CHAO, K., KIM, M.S. (2002). Machine vision technology for agricultural applications. Computers and Electronics in Agriculture 36(2±3): 173±191. CHEN, Z., LU, C. (2005). Humidity sensors: a review of materials and mechanisms. Sens. Lett. 3: 274±295. CLYDESDALE, M.F. (1998). Color: origin, stability, measurement and quality. In: Irwin, A.T. and Singh, R.P. (eds), Food Storage Stability. Boca Raton, FL: CRC Press. CONNELL, J.J. (1975). Control of Fish Quality. Farnham, Surrey: Fishing News (Books) Ltd. COSIO, M.S., BALLABIO, D., BENEDETTI, S., GIGLIOTTI, C. (2007). Evaluation of different storage conditions of extra virgin olive oils with an innovative recognition tool built by means of electronic nose and electronic tongue. Food Chemistry 101(2): 485±491. DAVIES, E.R. (2009). The application of machine vision to food and agriculture: a review. Imaging Science Journal 57: 197±217. DI EGIDIO, V., SINELLI, N., LIMBO, S., TORRI, L., FRANZETTI, L., CASIRAGHI, E. (2009). Evaluation of shelf-life of fresh-cut pineapple using FT-NIR and FT-IR spectroscopy. Postharvest Biology and Technology 54(2): 87±92. DUTTA, R., KASHWAN, K.R., BHUYAN, M., HINES, E.L., GARDNER, J.W. (2003). Electronic nose based tea quality standardization. Neural Networks 16(5±6): 847±853. BOURNE M.

FAGAN, C.C., EVERARD, C., O'DONNELL, C.P., DOWNEY, G., SHEEHAN, E.M., DELAHUNTY, C.M.,

(2007). Prediction of processed cheese instrumental texture and meltability by mid-infrared spectroscopy coupled with chemometric tools. Journal of Food Engineering 80(4): 1068±1077. FARKAS, J., DALMADI, I. (2009). Near infrared and fluorescence spectroscopic methods and electronic nose technology for monitoring foods. Progress in Agricultural Engineering Sciences 5(1): 1±29. FRIEDMAN, H.H., WHITNEY, J.E., SZCZESNIAK, A.S. (1963). The Texturometer ± a new instrument for objective texture measurement. Journal of Food Science 28(4): 390±396. GARCIÂA, M., ALEIXANDRE, M., GUTIEÂRREZ, J., HORRILLO, M.C. (2006). Electronic nose for wine discrimination. Sensors and Actuators B: Chemical 113(2): 911±916. O'CALLAGHAN, D.J., HOWARD, V.



401

(2005). Impulse response of pear fruit and its relation to Magness±Taylor firmness during storage. Postharvest Biology and Technology 35(2): 209±215. Â MEZ, A.H., WANG, J., HU, G., PEREIRA, A.G. (2006). Electronic nose technique potential GO monitoring mandarin maturity. Sensors and Actuators B: Chemical 113(1): 347± 353. Â MEZ, A.H., WANG, J., HU, G., PEREIRA, A. (2007). Discrimination of storage shelf-life for GO mandarin by electronic nose technique. LWT ± Food Science and Technology 40(4): 681±689. Â MEZ, A.H., WANG, J., HU, G., PEREIRA, A. G. (2008). Monitoring storage shelf life of tomato GO using electronic nose technique. Journal of Food Engineering 85(4): 625±631. GOWEN, A.A., O'DONNELL, C.P., CULLEN, P.J., DOWNEY, G., FRIAS, J.M. (2007). Hyperspectral imaging ± an emerging process analytical tool for food quality and safety control. Trends in Food Science and Technology 18(12): 590±598. Â MEZ, A. H., WANG, J., PEREIRA, A. G. GO

HERTOG, M.L.A.T.M., NICOLAIÈ, B.M., DE KETELAERE, B., LAMMERTYN, J., DE BAERDEMAEKER, J.

(2008). Non-destructive techniques and quality models for the supply chain: a review. Acta Hort. (ISHS) 768: 375±384. HUANG, H., YU, H., XU, H., YING, Y. (2008). Near infrared spectroscopy for on/in-line monitoring of quality in foods and beverages: a review. Journal of Food Engineering 87(3): 303±313. INFANTE, R., FARCUH, M., MENESES, C. (2008). Monitoring the sensorial quality and aroma through an electronic nose in peaches during cold storage. Journal of the Science of Food and Agriculture 88: 2073±2078. JESSOP, B., SIDER, K., LEE, T., MITTAL, G. S. (2006). Feasibility of the acoustic/EMG system for the analysis of instrumental food texture. International Journal of Food Properties 9(2): 273±285. JUODEIKIENE, G., BASINSKIENE, L. (2004). Non-destructive texture analysis of cereal products. Food Research International 37(6): 603±610. JUSTEÂ, A., THOMMA, B.P.H.J., LIEVENS, B. (2008). Recent advances in molecular techniques to study microbial communities in food-associated matrices and processes. Food Microbiology 25(6): 745±761. KILCAST, D. (2001). Modern methods of texture measurement. In: Kress-Rogers, E. and Brimelow, C.J.B. (eds), Instrumentation and Sensors for the Food Industry, 2nd edn. Boca Raton, FL: CRC Press. KONG, F., CHANG, S.K.C. (2009). Statistical and kinetic studies of the changes in soybean quality during storage as related to soymilk and tofu making. Journal of Food Science 74(2): S81±S89. KONG, F., TANG, J., RASCO, B., CRAPO, C., SMILEY, S. (2007). Quality changes of salmon (Oncorhynchus gorbuscha) muscle during thermal processing. Journal of Food Science 72(2): S103±S111. KONG, F., CHANG, S.K.C., LIU, Z., WILSON, L.A. (2008). Changes of soybean quality during storage as related to soymilk and tofu making. Journal of Food Science 73(3): S134±S144. KRESS-ROGERS, E. (2001). Chemosensors, biosensors, immunosensors and DNA probes. In: Kress-Rogers, E. and Brimelow, C.J.B. (eds), Instrumentation and Sensors for the Food Industry, 2nd edn. Boca Raton, FL: CRC Press. LAGUERRE, M., LECOMTE, J., VILLENEUVE, P. (2007a). Evaluation of the ability of antioxidants to counteract lipid oxidation: existing methods, new trends and challenges. Progress in Lipid Research 46(5) : 244±282.


402


(2007b). Rapid discrimination of scented rice by solid-phase microextraction, mass spectrometry, and multivariate analysis used as a mass sensor. Journal of Agricultural and Food Chemistry 55(4): 1077±1083. LI, Z., WANG, N., VIGNEAULT, C. (2006). Electronic nose and electronic tongue in food production and processing. Stewart Postharvest Review 2: 1±5. LIMBO, S., SINELLI, N., TORRI, L., RIVA, M. (2009). Freshness decay and shelf life predictive modelling of European sea bass (Dicentrarchus labrax) applying chemical methods and electronic nose. LWT ± Food Science and Technology 42(5): 977± 984. LIU, X., TAN, J. (1999). Acoustic wave analysis for food crispness evaluation. Journal of Texture Studies 30(4): 397±408. MARSILI, R.T. (1999). SPME±MS±MVA as an electronic nose for the study of off-flavors in milk. Journal of Agricultural and Food Chemistry 47(2): 648±654. MATHLOUTHI, M. (2001). Water content, water activity, water structure and the stability of foodstuffs. Food Control 12(7): 409±417. MCKENNA, B.M., LYNG, J.G. (2003). Introduction to food rheology and its measurement. In: McKenna, B.M. (ed.), Texture in Foods, Vol. 1. Semi-solid Foods. Cambridge: Woodhead Publishing Ltd, pp. 130±160. MILDNER-SZKUDLARZ, S., JELEN, H.H. (2008). The potential of different techniques for volatile compounds analysis coupled with PCA for the detection of the adulteration of olive oil with hazelnut oil. Food Chemistry 110(3): 751±761. MOREDA, G.P., ORTIZ-CANÄAVATE, J., GARCIÂA-RAMOS, F.J., RUIZ-ALTISENT, M. (2009). Nondestructive technologies for fruit and vegetable size determination ± a review. Journal of Food Engineering 92(2): 119±136. È LLER, A., STEINHART, H. (2007). Recent developments in instrumental analysis for food MU quality. Food Chemistry 102(2): 436±444. LAGUERRE, M., MESTRES, C., DAVRIEUX, F., RINGUET, J., BOULANGER, R.

NICOLAIÈ, B.M., BEULLENS, K., BOBELYN, E., PEIRS, A., SAEYS, W., THERON, K.I., LAMMERTYN, J.

(2007). Nondestructive measurement of fruit and vegetable quality by means of NIR spectroscopy: a review. Postharvest Biology and Technology 46(2): 99±118. PASTORELLI, S., VALZACCHI, S., RODRIGUEZ, A., SIMONEAU, C. (2006). Solid-phase microextraction method for the determination of hexanal in hazelnuts as an indicator of the interaction of active packaging materials with food aroma compounds. Food Additives and Contaminants 23(11): 1236±1241. PASTORELLI, S., TORRI, L., RODRIGUEZ, A., VALZACCHI, S., LIMBO, S., SIMONEAU, C. (2007). Solid-phase micro-extraction (SPME-GC) and sensors as rapid methods for monitoring lipid oxidation in nuts. Food Additives and Contaminants 24(11): 1219±1225. PERES, C., VIALLON, C., BERDAGUE, J.-L. (2001). Solid-phase microextraction-mass spectrometry: a new approach to the rapid characterization of cheeses. Analytical Chemistry 73(5): 1030±1036. PERIS, M., ESCUDER-GILABERT, L. (2009). A 21st century technique for food control: electronic noses. Analytica Chimica Acta 638(1): 1±15. PLUTOWSKA, B., WARDENCKI, W. (2008). Application of gas chromatography-olfactometry (GC-O) in analysis and quality assessment of alcoholic beverages ± a review. Food Chemistry 107(1): 449±463. QING, Z., JI, B., ZUDE, M. (2007). Predicting soluble solid content and firmness in apple fruit by means of laser light backscattering image analysis. Journal of Food Engineering 82(1): 58±67.



403

(2007). Instrumental-sensory evaluation of texture for fish sausage and its storage stability. Fisheries Science 73(5): 1166± 1176. RAO, M.A. (2007). Rheology of Fluid and Semi Solid Foods. Principles and Applications, 2nd edn. Gaithersburg, MD: Aspen Publishers. REICH, G. (2005). Near-infrared spectroscopy and imaging: basic principles and pharmaceutical applications. Advanced Drug Delivery Reviews 57(8): 1109±1143. RAHMAN, M.S., AL-WAILI, H., GUIZANI, N., KASAPIS, S.

RUDNITSKAYA, A., KIRSANOV, D., LEGIN, A., BEULLENS, K., LAMMERTYN, J., NICOLAIÈ, B.M., IRUDAYARAJ, J. (2006). Analysis of apples varieties ± comparison of electronic tongue with different analytical techniques. Sensors and Actuators B: Chemical 116(1±2): 23±28.

RUDNITSKAYA, A., POLSHIN, E., KIRSANOV, D., LAMMERTYN, J., NICOLAIÈ, B., SAISON, D.,

(2009). Instrumental measurement of beer taste attributes using an electronic tongue. Analytica Chimica Acta 646(1±2): 111± 118.

DELVAUX, F.R., DELVAUX, F., LEGIN, A.

SAEVELS, S., LAMMERTYN, J., BERNA, A.Z., VERAVERBEKE, E.A., DI NATALE, C., NICOLAIÈ, B.M.

(2004). An electronic nose and a mass spectrometry-based electronic nose for assessing apple quality during shelf life. Postharvest Biology and Technology 31(1): 9±19. SAINT-EVE, A., LEÂVY, C., LE MOIGNE, M., DUCRUET, V., SOUCHON, I. (2008). Quality changes in yogurt during storage in different packaging materials. Food Chemistry 110(2): 285±293. Â NCHEZ, M.-T., PEÂREZ-MARIÂN, D., FLORES-ROJAS, K., GUERRERO, J.-E., GARRIDO-VARO, A. SA

(2009). Use of near-infrared reflectance spectroscopy for shelf-life discrimination of green asparagus stored in a cool room under controlled atmosphere. Talanta 78(2): 530±536. SHIROMA, C., RODRIGUEZ-SAONA, L. (2009). Application of NIR and MIR spectroscopy in quality control of potato chips. Journal of Food Composition and Analysis 22(6): 596±605. SINELLI, N., BARZAGHI, S., GIARDINA, C. CATTANEO, T.M.P. (2005). A preliminary study using Fourier transform near infrared spectroscopy to monitor the shelf-life of packed industrial ricotta cheese. Journal of Near Infrared Spectroscopy 13: 293±300. SINGH, R.P., ANDERSON, B.A. (2004). The major types of food spoilage: an overview. In: Steele, R. (ed.), Understanding and Measuring the Shelf-life of Food. Cambridge: Woodhead Publishing. STEFFE, J.F. (1996). Rheological Methods in Food Process Engineering, 2nd edn. East Lansing, MI: Freeman Press. STEFFE, J.F., DAUBERT, C.R. (2006). Bioprocessing Pipelines: Rheology and Analysis. East Lansing, MI: Freeman Press. TAMAKI, K., SONOKI, S., TAMAKI, T., EHARA, K. (2008). Measurement of odour after in vitro or in vivo ingestion of raw or heated garlic, using electronic nose, gas chromatography and sensory analysis. International Journal of Food Science and Technology 43(1): 130±139. TIKK, K., HAUGEN, J.-E., ANDERSEN, H.J., AASLYNG, M. D. (2008). Monitoring of warmed-over flavour in pork using the electronic nose ± correlation to sensory attributes and secondary lipid oxidation products. Meat Science 80(4): 1254±1263. VAN ZEEBROECK, M., VAN LINDEN, V., RAMON, H., DE BAERDEMAEKER, J., NICOLAIÈ, B.M.,

(2007). Impact damage of apples during transport and handling. Postharvest Biology and Technology 45(2): 157±167.

TIJSKENS, E.


404


(2007). Volatile aldehydes in smoked fish: analysis methods, occurence and mechanisms of formation. Food Chemistry 105(4): 1536± 1556. VEÂLEZ-RUIZ, J.F., BARBOSA-CAÂNOVAS, G.V. (1998). Rheological properties of concentrated milk as a function of concentration, temperature and storage time. Journal of Food Engineering 35(2): 177±190. VESTERGAARD, J.S., MARTENS, M., TURKKI, P. (2007). Analysis of sensory quality changes during storage of a modified atmosphere packaged meat product (pizza topping) by an electronic nose system. LWT ± Food Science and Technology 40(6): 1083± 1094. VINAIXA, M., LLOBET, E., BREZMES, J., VILANOVA, X., CORREIG, X. (2005). A fuzzy ARTMAPand PLS-based MS e-nose for the qualitative and quantitative assessment of rancidity in crisps. Sensors and Actuators B: Chemical 106(2): 677±686. VLASOV, Y.G., BYCHKOV, E.A., LEGIN, A.V. (1994). Chalcogenide glass chemical sensors: Research and analytical applications. Talanta 41(6): 1059±1063. WIEDERHOLD P.R. (1997). Water Vapor Measurement. New York: Marcel Dekker. WHITE, C.H. (1993). Rapid methods for estimation and prediction of shelf-life of milk and dairy products. Journal of Dairy Science 76: 3126±3132. YALI, T., LIXIN, L., WEI, Z., JUN, W. (2009). Rapid detection techniques for biological and chemical contamination in food: a review. International Journal of Food Engineering 5(5): 1±13. YEO, T.L., SUN, T., GRATTAN, K.T.V. (2008). Fibre-optic sensor technologies for humidity and moisture measurement. Sensors and Actuators A: Physical 144(2): 280±295. YUAN, S., CHANG, S.K.C. (2007). Texture profile of tofu as affected by instron parameters and sample preparation, and correlations of instron hardness and springiness with sensory scores. Journal of Food Science 72(2): S136±S145. VARLET, V., PROST, C., SEROT, T.


13 Modelling microbiological shelf life of foods and beverages A. AmeÂzquita, D. Kan-King-Yu and Y. Le Marc, Unilever R&D Colworth, UK

Abstract: Predicting the fate of spoilage and pathogenic microorganisms can be very beneficial for food manufacturers if they are to market their products without harm to consumers and damage to the brand, through loss of quality. This chapter discusses the different biological ènd-points' that are relevant to the application of predictive models in food manufacture, presents an overview of the mathematical modelling approaches available for microbial shelf life prediction of foods and beverages, and describes the key considerations for development of predictive microbiological models, as well as the main limitations and practical considerations for the sound application and usage of models. Key words: predictive microbiology, mathematical modelling, microbial growth, microbial spoilage, hurdle technology.

13.1

Introduction

Food is inherently perishable and, depending on its physical and chemical properties and the storage conditions, there will come a point when either its quality will be unacceptable or it will become harmful to the consumer. At this point it has reached the end of its shelf life and the ability to predict this is of great value to the food industry when defining storage and distribution conditions and limits, formulating products, assessing manufacturing processes and doing quantitative risk assessment. Furthermore, having the ability to predict accurately the shelf life of a product would reduce the risk of unnecessary disposal of wholesome food due to a conservative estimate of its shelf life. In


406


designing a food product, it is important to identify which factors determine its shelf life: these may be microbiological, chemical or physical depending on the product, the process, the packaging and the storage conditions. This chapter focuses only on aspects related to the prediction of microbiological shelf life (safety and stability aspects), and it does not address chemical or physical factors affecting the shelf life of a food product. Depending on the product, process and storage conditions, the microbiological shelf life may be determined by either the growth of spoilage or pathogenic microorganisms. Traditional methods for determination of shelf life due to spoilage microorganisms include storage of the product at different temperatures and determining spoilage by sensory evaluation or microbial count. In the case of shelf life set by growth of pathogenic microorganisms, such methods may involve challenge testing the product with the organism prior to storage and microbiological analysis at intervals. These methods are labourintensive, time-consuming and expensive. Therefore, being able to predict the behaviour or fate of these various groups of microorganisms is very beneficial if manufacturers are to market their products without harm to consumers and damage to the brand, through loss of quality. Although challenge tests or product storage trials are often still used, the use of mathematical models to predict microbial behaviour or fate can help to reduce the need for them. Mathematical modelling in food microbiology is nowadays a well-established discipline, and has extended beyond academic research interests to real added-value industrial applications. There are several dedicated reference books (Brul et al., 2007; McKellar and Lu, 2004; McMeekin et al., 1993; Peleg, 2006), a number of electronic resources (which will be discussed later), and a plethora of peerreviewed publications, with some relevant recent review papers presented by Marks (2008), McMeekin (2007), McMeekin et al. (2002), and a complete special issue of the International Journal of Food Microbiology presenting selected papers from the Fifth International Conference on Predictive Modelling of Foods (PMF5), with the preface written by Koutsoumanis et al. (2008). In developing predictive models for spoilage microorganisms, the concept of specific spoilage organisms (SSOs) has proven to be very valuable in prediction of shelf life of certain types of foods such as seafood and meat products (Dalgaard, 1995; Koutsoumanis and Nychas, 2000; Kreyenschmidt et al., 2010; Nychas et al., 2008). An SSO can be defined as the fraction of the total microflora in a specific food product that is able to establish itself as the dominating population and is responsible for spoilage. The application of this concept assumes that the SSO produces the metabolites responsible for spoilage, that the rate of metabolite production is proportional to its growth rate, and that spoilage is noticeable when the SSO reaches a minimal spoilage level (MSL). As such, the end of shelf life can be defined on the basis of a spoilage criterion which could be either given by the time the SSO requires for multiplication from an initial population level to the MSL or by the time required for the production of a certain metabolite by the SSO to a level which results in sensory rejection (Dalgaard, 1995; Koutsoumanis and Nychas, 2000). Therefore, kinetic growth


Modelling microbiological shelf life of foods and beverages

407

models as a function of the spoilage domain (i.e. the space of environmental conditions within which the SSO is responsible for spoilage) are relevant and often used. For products where the shelf life may be set by the growth of pathogenic microorganisms, the range of relevant organisms is wide and includes infectious pathogens (e.g., Listeria monocytogenes), toxin-producing organisms (e.g., Staphylococcus aureus) and toxico-infectious agents (e.g., Bacillus cereus). Depending on the pathogen of concern, growth/no-growth models are often more appropriate, since manufacturers need to design foods that do not support even small amounts of growth during shelf life (e.g., where L. monocytogenes is tolerated in foods, such as in Europe, this is generally up to a level of 100 cfu/g of food). For organisms that can be tolerated at higher levels, such as B. cereus, kinetic models are more relevant, to predict time to a given log increase in the population. Predictive models are often used at early stages in the design of a food, to identify means through which a product developer can control relevant target microorganisms and set a shelf life that delivers the required safety and quality characteristics. Validation of the predicted behaviour is a common subsequent step that is carried out, to provide additional evidence and confidence that the design is appropriate. When used as part of the product and process design, predictive models can greatly facilitate ìn-silico' assessments of the effect that product reformulations and process modifications would have on the product shelf life, before planning laboratory or pilot-scale experiments. As such, predictive models offer a systematic, cost-effective approach for the development of microbiologically safe and stable food and beverage products. The main purpose of this chapter is to present an overview of the mathematical modelling approaches available for microbiological shelf life prediction of foods and beverages, and to describe the key considerations for development of predictive microbiological models, as well as the main limitations and practical considerations for the sound application and usage of models.

13.2 Classification of predictive models by microbial response Predictive models can be defined as a set of mathematical equations that describe the number of microorganisms present in laboratory media or in a food product, as a function of some environmental variables. Such variables (or controlling factors) as will be described later (on pp. 414±15) can be intrinsic to the food product or medium (e.g., pH or water activity of the food product formulation) or extrinsic (e.g., storage temperature of the food product). The vast majority of predictive models available in the literature quantify microbial populations or probabilities of presence of microorganisms. Little development has been made to model the behaviour of single organisms, partly due to the limited knowledge available at the cellular level.


408


Predictive models are commonly classified as primary, secondary or tertiary models (Whiting, 1995). Primary models characterise the number of microorganisms in a population as a function of time under specific conditions. Secondary models establish a mathematical relationship between the primary model parameters and the environmental variables of interest (e.g., pH, product storage temperature, etc.). By successively applying primary and secondary modelling, it is possible to describe the complete evolution of a microbial population as a function of the environmental conditions. The implementation of both primary and secondary models in a user-friendly software application is referred to as `tertiary models'. Several notable tertiary models exist in the public domain (see Section 13.5). For the food industry, the availability of ready-to-use tools is particularly useful for assessing the potential risk associated with some identified microbiological hazards in food products and processes. For food developers, tertiary models are particularly attractive for optimising new products and processes. Predictive models can be developed in many ways, depending on the purpose serving the model development and the information available at the time of development. With regard to the latter point, another level of model classification is commonly described in the literature (Peleg, 2006). Depending on the current scientific knowledge about the system to be modelled and the quality and quantity of data available for that particular system, predictive models (primary, secondary and tertiary) are usually described as empirical or mechanistic. Empirical models describe the data in a purely statistical way, with no representation of the current scientific knowledge about the underlying biological phenomena occurring in the system. These models are intended to provide the most accurate fit to the data. They are therefore very good for predicting microbiological behaviour within the range of the observed data (i.e., interpolation region). Microbial mechanistic models, on the other hand, mathematically characterise some known microbiological phenomena. They theoretically provide an exact mathematical translation of a microbiological behaviour, as understood and described by the broader scientific community. In theory, mechanistic models have the major advantage of providing reliable predictions of microbiological behaviours under any environmental conditions within and outside the range of the observed data (i.e. extrapolation region). Currently, insufficient knowledge is available to implement purely mechanistic models for microbial growth or inactivation. Most models used are either purely empirical or semi-empirical, meaning that some microbiological knowledge is incorporated, to a certain extent, into the model. Another way of classifying predictive microbiological models is based on the nature of the different biological ènd-points' (or microbial responses). Subjected to certain intrinsic and/or extrinsic conditions, observed microbial responses over time can be predicted and quantified by the use of growth models, inactivation or death models, survival models, growth/no-growth models or combined models. Growth models describe an increase in a population over time whereas inactivation models describe a decrease over time. Survival models also



409

describe death over time, but in this case death occurs relatively slowly when no lethal treatment is applied deliberately to the product (e.g., ambient-stable acidpreserved products such as mayonnaise or salad dressings). Growth/no-growth interface models (also called growth boundary models) describe the probability of survival (or growth) of microorganisms in the situation when the environmental conditions can either way result in microbial growth or inactivation (Dang et al., 2010; McKellar and Delaquis, 2002; McMeekin et al., 2000; Ratkowsky and Ross, 1995). Combined models describe the changes of behaviour in a microbial population subjected to conditions that can vary from growth to inactivation (Ross et al., 2005; Whiting and Cygnarowic-Provost, 1992). These models require particular attention in order to avoid some discontinuities at the interface between growth, growth/no-growth and inactivation. If such models are attractive for describing the complete range of microbial responses, they represent, however, a major ìnvestment' in terms of data requirements and modelling effort. Such investment may not be justifiable in practice for most food process models, which would usually focus on either microbial inactivation (for instance to ensure that pathogenic microorganisms are reduced by a certain amount (e.g. 6 logs) after a specific period of time) or growth (for instance to ensure that spoilage microorganisms are not able to grow to the extent of spoiling the food product before the end of shelf life). In the context of modelling the growth of spoilage microorganisms, if the population reaches the death phase (see Fig. 13.1), one might have no interest in quantifying the subsequent inactivation rate of the microbial population as the product would have already been spoiled.

Fig. 13.1 Schematic representation of a typical microbial growth curve.


410


13.2.1 Growth models Growth models have been a major area of development in predictive microbiology over the past 25 years. Traditionally, these rely on generation of kinetic data that enable a description of the whole growth curve that includes lag time, exponential growth phase, stationary phase and death phase (see Fig. 13.1). In practice, most growth models are based on the assumption of a sigmoidal growth function; in other words, the death phase is ignored in practice. The simplest form of sigmoidal response has been described by Buchanan et al. (1997) with a three-phase linear model, representing the lag phase, the exponential growth phase and the stationary phase. Zwietering et al. (1990) reparameterised different forms of sigmoidal curves in order to relate the model parameters to biologically meaningful terms, namely the lag time, the maximum specific growth rate and the maximum cell density. In the context of predicting microbiological shelf life, kinetic growth models are the most relevant ones, and these will be described in more detail in Section 13.4.1. Some of the most commonly used primary growth models are the Baranyi model (Baranyi and Roberts, 1994), the logistic model (Dalgaard, 1995), the modified Gompertz model (Gibson et al., 1987) and the three-phase linear model (Buchanan et al., 1997). 13.2.2 Growth/no-growth or growth boundary models The use of kinetic growth models might be appropriate for spoilage microorganisms or pathogens for which some growth may be tolerated up to a certain level. The situation becomes different when dealing with foodborne pathogens with a very low infective dose (e.g., E. coli O157:H7), or with toxigenic organisms where even a small amount of growth is considered potentially hazardous (e.g., Clostridium botulinum). In that respect, the ability to predict whether a foodborne pathogen might grow or not becomes more relevant with regard to consumer safety. Similarly, the ability to determine the growth limits of spoilage microorganisms with a high spoilage potential under certain environmental conditions might be of relevance (e.g., Zygosaccharomyces bailii in fruit concentrates and juices). The latter is particularly useful when designing ambient-stable products with long shelf lives. Being able to predict the location of the boundary between growth and no growth means being able to determine, in a multi-factorial space, the combination of environmental conditions (such as temperature, water activity, pH, etc.) that will influence the probability of microbial growth. This underpins the concept of `hurdle technology' (Leistner, 2000) to design combinations of environmental factors that will inhibit growth, ensuring the stability and the safety of the foods without compromising their nutritional and sensorial qualities (see Section 13.4.2). Models that define the combinations of environmental factors that prevent growth are known as `growth/no-growth' or `growth boundary' models. To some extent, when designing food products, growth/no-growth models are



411

preferred to kinetic growth models because any growth implies the possibility to cause food spoilage or illness. There is therefore an economical interest for food manufacturers to be able to design product recipes that would exclude the possibility of microbial growth without the need for challenge testing. Various authors have described and highlighted the importance of growth/no-growth models for designing safe and stable foods through `hurdle technology' (McMeekin et al., 2000; Ratkowsky and Ross, 1995). Ross and Dalgaard (2004) classify growth/no-growth models under three broad groups: deterministic approaches, logistic regression techniques and artificial neural network. Deterministic approaches to growth/no-growth interface modelling involve determining the set of environmental conditions that will predict a probability of growth of 50% as the boundary (Augustin and Carlier, 2000b; Le Marc et al., 2002; MembreÂ et al., 2001). Another technique involves using combined models (describing growth and death) where both the growth rate and death rate are estimated at the same time. The set of conditions predicting a growth rate equal to the death rate constitute the growth/no-growth interface (Battey et al., 2001). Logistic regression is a common statistical technique (falling under the family of generalised linear models) that is used to model binary data. Data available under the form `growth' or `no-growth' are adequate for the use of logistic regression. By using the logistic link function, it is possible to estimate the probability of growth under specific environmental conditions and the growth/no-growth boundary can be specified at any level of confidence. For example, Dang et al. (2010) proposed the use of linear logistic regression to help develop guidelines for designing new shelf-stable foods without the need for chemical preservatives. The model used in the study describes the growth/nogrowth boundary of the food spoilage yeast Z. bailii at 30 ëC, given a range of pH, water activity and (total) acetic acid (Ac), based on the following model: P b0 b1 aw b2 pH b3 Ac b4 a2w b5 pH2 logitP ln 1ÿP b6 Ac2 b7 aw pH b8 aw Ac b9 pH Ac 13:1 where P is the probability of growth, and b1 ; . . . ; b9 represent the parameters to be fitted to the experimental data. As can be observed in Eq. 13.1, the logistic regression approach relates logit(P) to a polynomial expression of the explanatory variables. Artificial neural networks (ANNs) have been proposed as an alternative to logistic regression (see for instance FernaÂndez-Navarro et al., 2010). They are mainly algorithms that can be used to describe complex nonlinear relationships between a large set of covariates and responses variables. Despite the efforts of some authors to demonstrate the fitting performance of ANNs, they are still not as widely used as logistic regression models. Advantages and disadvantages of ANN are described in Ross and Dalgaard (2004).


412


13.2.3 Inactivation and survival models Modelling inactivation is an important aspect of microbiological safety and stability by design, and it has received considerable attention in the field of food research. However, within the scope of this chapter, inactivation models are of little relevance with a view to modelling the microbiological shelf life of foods and beverages. A top level description is therefore given in this section, but no further details are covered throughout the chapter. Further information can be found in the literature (Peleg, 2006; Ross and Dalgaard, 2004). Kinetic inactivation models describe the decrease in microbial population over time, as is often observed after the application of a lethal process such as a heat treatment. When no lethal treatment is applied deliberately, but there is a death response under conditions where growth is prevented, the models describing this response are often called `survival' models. These are particularly useful when modelling the fate of microorganisms in products where the preservation system is based on non-thermal intervention processes, such as fermented meats, fermented dairy products (e.g., yoghurts and cheeses), dressings or mayonnaises and beverages (e.g., fruit juices). Likewise, survival models are useful when designing multiple-use food products where the shelf life can extend for many weeks after opening and it is critical that the same consideration for design is applied to òpen' as well as `closed' shelf life. Unlike growth, where the shape of the microbial response is generally the same i.e. sigmoid, the kinetics of survival are not easily predicted. Different shapes of microbial inactivation/survival curves can be observed, and generally they can be described as having (a) a tailing pattern, (b) an increase in population followed by die-off (commonly observed when spore activation occurs), (c) a shoulder or lag prior to inactivation, or (d) a sigmoidal pattern with both a lag and a tail (see Fig. 13.2). Despite these observed shapes of inactivation/survival curves, most inactivation data available in the literature are assumed to be log-linear and are modelled as such. This assumption is likely to be dominated by mathematical considerations so that a simple log-linear form can be used for modelling purposes. At the moment, there is no clear consensus about a suitable choice of model for describing inactivation data. This may change with the proposal of new mechanistic models that would relate biological responses to inactivation patterns and incorporate the natural variability of the behaviour of microorganisms under lethal stress.

13.3 Development of predictive models for microbiological safety and stability Although there are different approaches to develop predictive models in food microbiology, this section discusses key aspects that should be considered when developing a model. As mentioned in the previous section, mechanistic models aim at describing the theoretical basis of the microbial response, but owing to the complexity of microbial physiology and our current level of



413

Fig. 13.2 Common shapes of survivor curves: (a) initial rapid decrease with a subsequent tail, (b) initial increase in numbers followed by a phase of rapid decrease (observed typically during heat inactivation studies of spore-forming bacteria), (c) initial shoulder (or lag) followed by a phase of rapid decrease, (d) sigmoidal curve.

understanding, these types of models are rare. Consequently, most predictive microbiological models are, generally speaking, empirical. By èmpirical' we mean that the way in which the microbial behaviour or fate is predicted is based on observing the effects of various factors on microorganisms in systems (either laboratory media or foods), usually under well-controlled conditions, and then fitting these data with mathematical functions (models). In most cases, these models have been formulated in such a way that the fitted parameters have biological relevance (e.g., specific growth rate or lag phase duration). In that context, although they are observational models, they are not, strictly speaking, `black box' models as is the case with purely statistical models (e.g., polynomial models). In some instances, though, statistical models offer the most suitable alternative to describe the dataset used for their generation, and therefore, are still often used; however, their parameters have no biological significance. One of the fundamental premises on which predictive modelling in food microbiology is based is that the microbial responses to environmental conditions are reproducible. Therefore, if those environmental conditions can be properly characterised in terms of the factors that control the main physiological events (i.e., growth, survival and inactivation), then it is possible to predict the microbial responses on the basis of past observations when similar environmental conditions can be properly monitored. To that end, having high-quality data describing the full range of responses to the


414


environmental conditions of interest is a fundamental step when building a predictive model. In order to gather sufficient information to build a good database for development of a model, the main options are either to generate new data via experimentation or to use data reported in the literature or in public databases. The latter option has become in recent years more accessible to non-academic users with the creation of web-based databases such as ComBase (http:// www.combase.cc) (Baranyi and Tamplin, 2004) or Sym'Previus (http:// www.symprevius.net) (Leporq et al., 2005), which will be discussed later in the chapter (see Section 13.5). When using data from literature to develop a predictive model, the main disadvantage is that, in most cases, the experiments reported are not designed primarily to produce a predictive model or some details of the experiments are not always described adequately. Nonetheless, good examples where only data from literature and/or public databases were used for model development have been reported (Augustin et al., 2005; Le Marc et al., 2005; Ratkowsky and Ross, 1995), though due to the availability of data in the public domain, these are typically limited to pathogenic organisms. In spite of some of the possible limitations of such an approach, oftentimes a thorough investigation of the data available in the literature or in public databases can provide useful information about certain patterns of microbial behaviour under the environmental conditions of interest. That information may in turn indicate areas where gaps exist, thereby facilitating the design of experiments to generate the data required for model development. 13.3.1 Experimental considerations Experimental design Due to the empirical nature of predictive models in food microbiology, they cannot be used reliably to make predictions beyond the area defined by the conditions used to generate the model (see page 422 for a more in-depth discussion on this topic). Therefore, the intended range of model usage is a key consideration for the design of experiments to generate the data required for model development. When a model is developed without much prior thought to the scope of the subsequent applications, this may result in inappropriate choice of key controlling factors and limitation of its use. A better strategy is to decide on the food or range of foods to be targeted and ensure that the controlling factors are selected to reflect this. The many factors that can affect the fate of microorganisms in food can be grouped into three categories: · Intrinsic factors: characteristics of the food itself (e.g., pH, water activity (aw), concentration of preservatives). · Extrinsic factors: characteristics of the environment in which the food is stored (e.g., temperature, gaseous atmosphere, humidity). · Implicit factors: the characteristics of the microorganism itself and how it behaves in the presence of combinations of the intrinsic and extrinsic factors (e.g., specific growth rate of the microorganisms).



415

In many cases, the fate of microorganisms in foods is determined by a small number of factors, such as pH, aw and temperature, and it has therefore been relatively straightforward to characterise responses in laboratory media and to develop models that predict their fate in foods, provided that the food environment can be adequately simulated. An experimental system is required in which these factors can be altered easily. In most cases microbiological media are used because they are of consistent composition and can be easily and reproducibly modified to the required conditions. In some cases there may be different methods of applying factors, e.g. the choice of acidulant and humectant for adjusting pH and aw, respectively. If the model is to be applied across a wide range of foods, then the use of less inhibitory chemicals, e.g. hydrochloric acid, are more likely to avoid fail-hazardous predictions (where slower microbial growth is predicted than actually happens). However, if the model is intended for specific foods then the choice of factors may need to be more focused, e.g. specific organic acid in order to include the effects related to the undissociated molecule. This approach allows inclusion of additional inhibitory factors that may be the difference between a safe or stable formulation and a potentially hazardous or unstable formulation. When considering aw as a controlling factor, it must be noted that different microbial responses, at a given aw, can be observed when different humectants are used to control the aw value (Mattick et al., 2001; Stewart et al., 2002). This specific solute effect not only can affect the minimum tolerated aw (i.e., the minimum aw level allowing growth), but also the growth rate. This may be particularly difficult to deal with when it comes to designing an experiment with a view to developing a predictive model, because for some organisms, a particular humectant may be more inhibitory than others on an aw basis but the opposite effect is observed for another organism. For example, glycerol is more inhibitory than NaCl for some Gram-positive cocci such as Staphylococcus aureus, with minimum aw for growth reported as 0.86 with NaCl as humectant and 0.89 with glycerol (Marshall et al., 1971), whereas for Listeria monocytogenes the opposite effect is observed, with minimum aw for growth given as 0.92 with NaCl as humectant and 0.90 with glycerol (Tapia de Daza et al., 1991). Simplifying the experiments by choosing only the most relevant three or four factors driving the microbial response in laboratory media is common practice, and this is typically done for cost reasons. However, growth models developed using those simplified systems tend to give conservative predictions (i.e., failsafe predictions), and this results inevitably in restrictive estimates of product shelf life. To that end, models that are developed in a matrix that closely simulates the food product of interest provide more accurate predictions, but these usually require more complex experimental conditions and have a restricted domain of validity for their application. The choice of strain(s), size of inoculum and culturing conditions of the microorganism used will all affect the responses measured and subsequent predictions. Different strains have different phenotypic responses and so the inclusion of mixtures of strains (i.e., cocktails) or some form of strain selection or screening needs to be carried out. Cocktails are more likely to contain a range


416


of physiological responses that may not be fully characterised, but may be more representative the diversity found in naturally occurring microbial populations. The strains used in the cocktail should be, as much as reasonably possible, representative of the target group of organisms likely to be found in the product for which the model is intended to be used. For shelf life studies, a commonly used approach is to select the fastest growing strain in the spoilage domain, which will simulate a worst-case scenario. Likewise, strains that have been isolated from spoiled products representative of those under investigation are often used. For example, Neumeyer et al. (1997) screened nine strains of psychrotrophic pseudomonads of dairy origin, and selected the fastest growing strain (Pseudomonas putida 1442) as the main isolate for model development; however, the authors also modelled the slowest of the strains (P. fluorescens 1412) to provide an indication of strain-to-strain variability of growth rates. The size of the inoculum has to ensure that the expected microbial response can be measured rather than necessarily actually reflecting the numbers commonly present. For instance, in generating growth curves for developing a growth model, an initial inoculum of 2±3 log10 cfu/ml or g is appropriate because it allows enumeration during the lag phase whilst reducing the chance of increasing unrealistically the probability of growth which could occur at high initial levels. In contrast, if a survival model is of interest (i.e., a non-thermal inactivation model under mild conditions), for example in the case of a pathogen of concern in a product designed to be ambient-stable, then an appropriate inoculum can be around 4±6 log10 cfu/ml or g to allow either growth or death responses to occur (as the experimental conditions of interest may fall very near the growth/no-growth boundary). The pre-history (growth or storage conditions including temperature and growth medium) can affect the microorganism's response to the controlling factors and it should be carefully selected to reflect as far as possible the likely conditions of naturally contaminating microorganisms. The statistical design to be used is crucial for the subsequent stages in the process of developing a predictive microbiological model. In recent years, various sources of detailed information about statistical design of experiments for microbiological modelling have become available (Rasch, 2004; van Boekel and Zwietering, 2007). We limit our discussion to basic principles that may be important to be considered in the context of growth and growth boundary models, which are of special interest for microbiological shelf life prediction. In growth models, full factorial designs are often not necessary because the region of interest may be in the area where the response that is being measured is more variable (e.g., near a boundary of growth/no-growth), and some treatments in the design will involve combinations of factors that do not support growth. Therefore, fractional factorial designs are often used, particularly Box±Behnken designs as these not only allow a reduction in the number of experimental units (compared to a complete factorial design) but also the design points fall on the edges of the cuboidal region rather than on the corners (i.e., where combinations of conditions are likely to result in a no-growth response) (see Fig. 13.3). Battey and Schaffner (2001) reported the development of a growth model for the



417

Fig. 13.3 Example of a Box±Behnken design for the case of three factors (e.g., temperature, pH and water activity). Several replications of the design centre are typically included.

spoilage bacteria Gluconobacter oxydans and Acinetobacter calcoaceticus in cold-filled ready-to-drink beverages, using a Box±Behnken design for the generation of growth curves. Their design consisted of 42 experimental conditions (in duplicate) to cover five factors and three levels (pH 2.8, 3.3, 3.8; titratable acidity 0.2%, 0.4%, 0.6%; sugar content 8, 12, 16 ëBrix; sodium benzoate 100, 225, 350 ppm; and, potassium sorbate 100, 225, 350 ppm), resulting in a more manageable study than a full factorial design (i.e. 53 experimental conditions), yet providing sufficient information for model development. In growth/no-growth models, the main interest when designing an experiment is the detection of growth, without the need to quantify it. For this type of model, it is crucial that the experimental design matrix includes many replicates with about half of the conditions resulting in observable growth on one side of the interface, and the other half resulting in non-detectable growth on the other side of the interface. Therefore, automated methods are often used primarily in screening experiments to elucidate the choice of experimental conditions that are `marginal' (i.e., very near the growth/no-growth interface), and ideally, these could have complete factorial designs which can be handled relatively easily in the automated system. For generating the dataset that will ultimately be used for model development, central composite designs may be more manageable, but ideally they should be centred on the `marginal' growth conditions where the greatest variability is expected. Data generation The most labour-intensive stage is the generation of data (e.g., growth or survival) of the organism in the model system. Quantification of microorganisms


418


at selected time points is usually by standard colony count methods, but optical density and conductance measurement have also been used. When the target microorganism is in pure culture, methodology for enumeration is usually straightforward, but for survival models provision to enumerate sub-lethally injured cells may need to be made. Although total viable count (TVC) methods are expensive and labour intensive, they have the advantage of best describing growth kinetics of a bacterial population. The accuracy of the growth parameter estimates depends on the number of counts made per growth curve and also the quality (uncertainty on the experimental data points) of the measurement (Poschet et al., 2004). For instance, Walker and Jones (1993) suggested a minimum of 10 points per growth curve. This is not always achievable in practice due to time and/or budget constraints. It is important to note that having a large amount of observations can be of little use if these are wrongly positioned in time. For example, the accuracy of the lag time estimate will be optimal if one can sample sufficient data points during the transition zone between the lag phase and exponential phase. Automated optical density (OD) methods have the advantage of being rapid, inexpensive and, generally speaking, allow for quicker data generation with less need for human resources. However, they present the main disadvantage of having a limited range of validity. This is because the experimental detection limit usually corresponds to a bacterial concentration in the range of 106±107 bacteria/ml (Begot et al., 1996). The growth rate obtained using OD will typically represent the late-exponential growth phase which will be less than the maximum specific growth rate, leading to underestimation of the `true' value. For this reason, OD measurements are often used in combination with time-to-detection (TTD) measurements (Cuppers and Smelt, 1993). Using serial dilutions (i.e. changing the inoculum size), TTD measurements can be used to obtain values for both the growth rate and the lag time (Baranyi and Pin, 1999; Cuppers and Smelt, 1993; Wu et al., 2000). This technique has been applied by several researchers in recent years, and comparisons between TVC and automated OD methods have shown that both types of techniques can be used reliably depending on the objectives of the study (Augustin et al., 2005; Biesta-Peters et al., 2010; Metris et al., 2006). For instance, automated OD measurements can be used to determine growth parameters for certain factors such as pH or preservative concentration, where a large number of combinations can be assessed quickly. However, for some specific growth conditions, TVC methods will remain necessary, such in the case of modified atmospheres which cannot easily be achieved in OD measurements, or for certain temperatures where evaporation or condensation of liquid media may affect the reliability of the method. 13.3.2 Data analysis and modelling The next stage involves mathematical analysis of the data to produce a model and determine the quality of the data and the goodness of fit of the data to the model. There are a number of different modelling techniques for growth,



419

probability of growth, survival and inactivation, and these are now well established and thoroughly described in dedicated books and review papers (Brul et al., 2007; McKellar and Lu, 2004; McMeekin et al., 1993; Whiting, 1995). These techniques essentially comprise the development of primary and secondary models (see Section 13.2) from a database (either built from experiments specifically designed to build the model or from literature or independent sources), and turning those data into knowledge by developing user-friendly software (or tertiary models) through which application is made possible (i.e. they enable users to ìnterrogate' primary and secondary level models in order to obtain predictions). ComBase (http://www.combase.cc) and Sym'Previus (http:// www.symprevius.net) are examples of current systems that combine that database-modelling-software functionality, and these will be described later (see Section 13.5). Initial inspection of the raw data using plots or tabulations is often used to check for quality and consistency. The fitting process typically involves linear or nonlinear regression techniques where the use of statistical software is required. For kinetic models (primary level), some tools are available free of charge in the public domain for both growth and inactivation models. For example, GInaFiT (Geeraerd et al., 2006) is a freeware tool to fit inactivation data to different mathematical models that describe all the known survivor curve shapes shown for vegetative bacterial cells. DMFit is another example of a freeware tool developed by the Institute of Food Research in the UK, and available as part of the ComBase Modelling Toolbox (http://www.combase.cc/toolbox.html). DMFit allows the user to fit bacterial (growth or survival) curves (i.e., logarithmic cell counts vs time) where a linear phase is preceded and followed by a stationary phase to different primary models. In growth models, traditionally the fitting process is done in two steps, i.e. fitting kinetic data to a primary model first, and then fitting the parameters of interest (e.g., growth rate) in terms of the independent variables (e.g., temperature, pH, etc.) via a secondary model. However, one-step global regression procedures are now used more frequently, where the response can be fitted directly as function of time and the independent variables in one step. This approach offers the flexibility of selecting parameters that can be fitted in common for all the data (according to the assumptions or constraints of each model), as well as selecting parameters that can be fitted individually for each curve. Although this approach is more demanding in terms of both software and expertise, it offers the advantage of preventing accumulation of fitting errors (Valdramidis et al., 2005), and it is particularly useful when the dataset contains incomplete growth or survival curves. Currently, there are freeware user-friendly tools that facilitate this onestep regression process, and one example of such a tool is OptiPa (Hertog et al., 2007, available at http://perswww.kuleuven.be/~u0040603/optipa). An important consideration when starting the development of a predictive model relates to the stochastic assumption used when fitting a mathematical equation to experimental data, i.e., to obtain the best fit of the model to the data, the error in the estimate of the selected response must be independent of the


420


value of the measured response. In microbiology, due to the exponential nature of growth, the growth rates over the range of conditions tested may span several orders of magnitude (from the slowest to the fastest growth rate), and the measurement error tends to be a constant proportion of the magnitude of the growth rate (Ross, 1999). As such, data for longer generation times will tend to have a greater influence on the fitting process than data for shorter generation times, because the differences between measured and expected values are usually larger. Therefore, a suitable transformation should be used to normalise this variance. This is particularly relevant when comparing the performance of different models. Two transformations are commonly used: logarithmic and square root of rate. Whichever modelling technique is used, the model should describe the data as accurately as possible without being overly complicated. Models should use the minimum number of parameters which describe the response adequately (i.e., be parsimonious), where an adequate fit is defined by certain criteria. 13.3.3 Model validation Mathematical testing and examination for biological sense Mathematical testing is the process of quantifying how well the model describes the data and one approach has been described by McClure et al. (1994). There are a number of sources of variability that may be the inherent variability of the microorganism, systematic errors due to analytical laboratory methods or bias due to inappropriate modelling techniques not adequately describing the data. There is a degree of acceptance or rejection at this stage and any requirement for additional or repeated microbiological data, or the use of a more appropriate modelling technique, can be highlighted. Visual comparison of predicted values against observations under the same conditions is always a good starting point in assessing the performance of the model. However, more systematic measures of goodness of fit are necessary. A commonly used measure for both linear and nonlinear regression models is the root mean square error (RMSE, Eq. 13.2), which measures the àverage' discrepancy between observed data (transformed if necessary) and their predicted values (Ratkowsky, 2004). The magnitude of the RMSE is useful in assessing whether the model fits the data well; the smaller its value, the better the fit of the model to the experimental data. s P predicted ÿ observed2 13:2 RMSE df where df is degrees of freedom (i.e., the number of observations minus the number of parameters estimated). In order to assess the performance of a model in a more systematic way, Ross (1996) introduced two dimensionless indices: the bias and the accuracy factors (Eqs 13.3 and 13.4, respectively), later modified and generalised by Baranyi et al. (1999) to quantify the confidence in the model predictions:


Modelling microbiological shelf life of foods and beverages 421 P bias factor 10 logGTpredicted =GTobserved =n 13:3 P accuracy factor 10 jlogGTpredicted =GTobserved j=n

13:4

where GTpredicted is the predicted generation time and GTobserved is the observed generation time, and n is the number of observations. The bias factor indicates systematic over- or under-prediction. For example, a bias factor of 1.15 indicates that the predicted generation times, on average, were 15% longer than the observed values, indicating that the model is `faildangerous'. It also indicates that the predictions exceed the observations by 15% on average. Equation 13.3 can equally be used for the maximum specific growth rate, and in that case a bias factor greater than 1 indicates a `fail-safe' model. The accuracy factor provides a measure of the average difference between the observed and predicted values. The larger the value of the accuracy factor, the less accurate the prediction. An accuracy factor of 1.1 would indicate that the observed and predicted values differ by 10% on average, without indicating if the discrepancy was an over- or under-prediction. Ross (1999) proposed general criteria defining the acceptability of a validated model on the basis of the bias factor, as follows: · a bias factor of 0.9±1.05 can be considered good · a bias factor of 0.7±0.9 or 1.06±1.15 can be considered acceptable · a bias factor < 0.7 or > 1.15 can be considered unacceptable. The criteria above are consistent with recommendations given by Dalgaard (2000) for spoilage models in seafood, which considered that bias factor values between 0.75 and 1.25 can be used as a criterion for successful model validation. The criteria proposed by Ross (1999) have been used by various authors in validating their models (see, for instance, Mejlholm et al., 2010). It is also important that the model predictions make biological sense. This is often done by plotting predictions in three dimensions via surface plots or bar graphs of the response variable of interest (e.g., growth rate) as function of two of the independent variables included in the model. Several such plots can be created for different conditions, and it is simply recommended to examine if the predictions from the model behave as expected on the basis of microbiological experience. This may readily help in understanding unexpected model predictions in particular circumstances where the presence of a preservative at a certain level may hinder the expected effect of another one (i.e., opposite to the `hurdle' effect). One such effect, for example, has been reported for E. coli and Salmonella in acid-preserved foods containing a combination of NaCl and acetic acid (a commonly used hurdle strategy) (Chapman and Ross, 2009). In that study, the authors reported that NaCl at intermediate concentrations in an aqueous solution containing acetic acid at lethal acidity levels has a protective effect on both pathogens, delaying the `lag' time prior to inactivation.


422


Domain of validity The domain of validity of a predictive model is defined by the minimum convex polyhedron (MCP) of the experimental design (Baranyi et al., 1996), which can be understood as the minimum space enclosed by the actual combinations of conditions used to generate the model. In other words, the MCP defines the interpolation region of the combinations tested. Predictions outside the MCP (i.e., extrapolation region) may not be reliable (see Fig. 13.4). When models are used outside the interpolation region, this should be made explicit and considered in the interpretation of the outcome. In the case of growth models, it is often the case that conditions resulting in no growth are omitted from the model fitting, but these may still be of interest to the food industry as they may be very close to the growth/no-growth interface. In that case, comparisons made with literature data outside the model interpolation region are useful to highlight areas where model data could be added to the original dataset in order to make the model more applicable to a wider range of products. Le Marc et al. (2005) proposed the utilisation of the MCP concept to define a `growth' region, which could contain data for which growth has been reported in the literature or recorded in a public database such as ComBase. This `growth' region may include conditions that could be used for fitting the predictive growth model, plus additional growth data (from independent sources) not used in the model development. The advantage of this approach is that this new `growth' MCP region may highlight new combinations of factors where new data could be generated for model refinement (i.e., a growth region where previously no data was available from the experimental design). The region outside the `growth' MCP can be considered a `no-growth' region. However,

Fig. 13.4 Schematic representation of the model interpolation region (minimum convex polyhedron ± MCP) for the simple case with two variables (temperature and pH).



423

Fig. 13.5 Schematic representation in two dimensions of the space of the environmental variables as divided into the `Model' (M), `Growth' (G), Ùncertainty' (U) and `No Growth' (NG) regions (after Le Marc et al., 2005).

this region can be further divided into an ùncertainty' region and a `no-growth' region, where there is certainty that the conditions will result in no growth based on microbiological knowledge and experience (see Fig. 13.5). Product validation Product validation involves the comparison of predictions from a model with growth, survival or death data of the relevant organism in food. The most rapid and inexpensive way of acquiring these data is the use of scientific publications, although the amount of data can be limited and is often incomplete with no measurement of some of the necessary physicochemical factors such as pH, sodium chloride concentration or aw. However, for some pathogenic microorganisms such as L. monocytogenes, there is nowadays a considerable amount of data available in the public domain (from the literature or public databases) in different food matrices. This has recently allowed a joint international effort to validate various proposed growth models for this organism (Mejlholm et al., 2010). Figure 13.6 shows an example of model validation using literature data. The figure compares the predicted growth/no-growth interface from a model developed in the authors' laboratory for L. monocytogenes as a function of pH, aw and preservative concentration against data reported by Tienungoon et al. (2000). The problems of incomplete or lack of data from the literature to validate a predictive model can be overcome by the use of challenge tests specifically designed for the purpose of product validation. In this way, the data are often more accurate, reliable and complete. Specific challenge tests can be time consuming and are relatively expensive, so they are often used to supplement


424


Fig. 13.6 Effects of pH, aW and lactic acid (50 mM) on the growth boundary of Listeria monocytogenes at 25 ëC. The solid line represents predicted values of a model developed in our laboratory; dark grey and light grey points are reported growth and no-growth values, respectively (data from Tienungoon et al., 2000).

published data. Obtaining quantitative microbial growth or survival data can be problematic if the target organism or group of organisms is outnumbered by the natural food microflora. This may require the use of selective agars, which in themselves may not completely prevent overgrowth by competitor organisms, but may lead to an underestimate of any injured cells that are present. One way of eliminating the problem of the natural flora in the food is simply by purchasing sterile or commercially sterile foods (Walls and Scott, 1997), or using a heat, filtration or irradiation process. While this approach enables nonselective agars to be used, it can be criticised for not reflecting the ecology of most foods. The use of antibiotic-resistant strains of the target organism and the incorporation of the antibiotics in non-selective or minimally selective agar has enabled specific enumeration in the presence of outnumbering background flora (Blackburn and Davies, 1994). Ideally, the validation should include the foods in which the organism is considered a hazard or the cause of spoilage and the physicochemical properties of the foods and storage/heating temperatures should, as far as possible, cover the range of the controlling factors of the model. Physicochemical analysis of the food and monitoring of the storage conditions are required and as a minimum, these must include the controlling factors of the model (e.g., temperature, pH, aqueous sodium chloride, aw). Measurement of other factors that are not in the model and that may affect growth/survival (e.g., preservatives) can be useful to help explain any deviations between predictions and challenge test data. There are a number of reasons why significant deviation between predictions and observed data may be seen. Published data are usually not designed for validation purposes and are, therefore, often incomplete. There can be considerable variation between species and strains. There may be growth-inhibitory factors in the food that are not accounted for in the model, e.g. the presence of an organic acid or different humectant. This tends to lead to fail-safe predictions for



425

growth models. The history of the inoculum can affect the subsequent lag phase or survivability of the population. The natural food microflora can affect the physicochemical properties of the food when they reach spoilage levels. The inappropriate use of physicochemical data (e.g., the use of an aqueous salt measurement for a food in which aw is affected by other humectants) or the use of the model outside its domain of validity can account for some of the differences between predictions and experimental data. An understanding of all these factors greatly enhances the interpretation and application of predictions. In the past, there has been considerable scepticism towards the application of predictions from models developed in laboratory media to foods. Product validation has gone some way to redress the balance and demonstrate the value of predictive models. More specifically it determines the applicability of a model for use with different foods and can highlight foods or conditions where care is needed in applying predictions. In this way the data can be used as a means of accepting, rejecting or modifying the model. Ironically, it is often when conducting product validation that any limitations of a model, in terms of the choice of controlling factors and their ranges, are realised. In fact, an initial, limited product validation study is useful as an integral part of experimental design. 13.3.4 Limitations of using models Models can obviate the need to generate results in all conditions of interest, but should be interpreted with care and by those with a certain level of microbiological knowledge. Models are most often not identical to reality; they are a simplified representation of reality so that their outcomes have to be interpreted with a good understanding of the window of conditions they were generated for. Model predictions should only be used as a guide to the response of an organism under a particular set of conditions. One reason for this is that the strain of a contaminant and the conditions in the food are unlikely to be the same as those used to generate the model. Some models have been built using single strains, whereas others use cocktails containing several strains of an organism. When a cocktail has been used, the response will be determined by the fastest growing or the longest surviving strain under the particular conditions tested. The experimental methods and conditions used to generate the data will also affect the model (e.g., pre-treatment of strains, type of recovery media, plus numerous other experimental variables). Growth models, which have been generated from data in liquid microbiological media, oftentimes give predictions on the side of safety, as the media are normally designed to give optimum growth of the microorganisms, whilst growth may be slower in some foods. Therefore, the user may want to validate the model predictions in their specific food of interest (i.e., use the models to define parameters for a challenge test), as discussed on pages 420±5. Even if the models have been generated from experiments in a food, the user should still validate these, as it is unlikely that the food used in the model will be identical to


426


the food of interest. Completeness errors are a common source of disagreement between model predictions and observed values in actual food matrices, i.e., there may be factors in some foods which could have a big effect on the fate of the microorganism of interest but which may not have been considered in the generation of the model. Other limitations are related to the design and structure of the model itself. For example, over-parameterised models, where the principle of parsimony has not been observed, provide good descriptions of the data in the interpolation region (i.e., acceptable fit), but can result in highly erratic predictions near the limits of the interpolation region or for conditions not originally tested. Finally, a model is only as good as the data used to build it, so the fit of the data to the model is important, and it may be better in one region of the model than another (or there may have been more data generated in one region, e.g. at a specific temperature). When using growth models, care should be taken when predicting at the boundaries of the model, as predictions are not always so good at the extremes of parameters. However, in these regions, the growth model predictions are generally fail-safe. Where there is more than one model for an organism, the user should choose the model and the parameters which most closely reflect the food of interest, but should be aware of the limitations described above. Despite the limitations described, several advantages can be gained from using model predictions. For example, to start a challenge test from scratch often involves setting up numerous conditions, which can be very lengthy and costly. Use of models can narrow down the range of conditions which need to be investigated in a challenge study.

13.4 Modelling approaches, applications and opportunities for shelf life prediction This section focuses on those predictive models and approaches that are of particular interest and applicability in the prediction of microbiological shelf life. It discusses growth models and how they can be used to predict shelf life of foods and beverages, growth boundary models for quantification of hurdle effects, and finally, relative rate of spoilage (RRS) models. 13.4.1 Modelling microbial growth The food industry is increasingly asked to provide fresher or more natural foods and beverages to the consumer. The general demand for authentic foods with less or no preservatives implies milder treatments without compromising the microbiological stability and safety of food products during their shelf life. The design of such foods requires a thorough characterisation of the growth kinetics of relevant microorganisms in order to establish a shelf life that is suitable for both the food company and the consumer. For example, toxigenic microorganisms such as S. aureus may be present as long as they are not able to grow



427

to a level where they produce toxin. Spoilage microorganisms such as yeasts and lactic acid bacteria can be tolerated as long as no visible signs of spoilage are observed during the entire shelf life. Being able to model accurately the growth rate and the lag time of microorganisms present in the food is therefore fundamental in determining a shelf life. Primary growth models In 1825, a new mathematical function was proposed by Gompertz to describe the human mortality rate as a function of age. Known as the Gompertz function, this equation was used by insurance companies to calculate the cost of life. A modified version of this equation was first used in predictive microbiology by Gibson et al. (1987), and is given by: log10 N t A C expfÿexpÿBt ÿ Mg

13:5

where N t is the number of bacteria at time t, A is the asymptotic log-count as t decreases to zero, C is the asymptotic amount of growth that occurs as t increases indefinitely, and B is the relative growth rate at M, where M is the time at which the absolute growth rate is a maximum. From the modified Gompertz model presented in Eq. 13.5, other quantities of interest are derived as follows: BC 13:6 growth rate (log10 count/h) exp1 lag time (h) M ÿ

1 B

13:7

In the early years of predictive microbiology, some authors considered this model to be one of the best sigmoidal models for growth curves (McMeekin et al., 1993; Zwietering et al., 1991). However, it presents the main drawback of overestimating the specific growth rate and the lag time due to an inflection curve inherent to the Gompertz curve (Baranyi et al., 1993; McKellar and Knight, 2000; MembreÂ et al., 1999; Whiting and Cygnarowic-Provost, 1992). Some attempts have been made to propose more biologically-based growth models. The Baranyi model (Baranyi and Roberts, 1994) is one of the most popular population growth models that was developed to describe the process of adjustment of microbial cells by hypothetical adjustment functions (Eqs 13.8 and 13.9). dN Qt N t N t 13:8 max 1 ÿ dt 1 Qt Nmax dQ max Qt dt

13:9

where N t is the cell density at time t (cfu/ml), max the maximum specific growth rate (hÿ1), Nmax the maximum cell density (cfu/ml) and Qt is a dimensionless quantity related to the physiological state of the cells at time t (Q(0) is the initial physiological state of the cells at inoculation).


428


Qt ) is a 1 Qt monotonic increasing function taking values from 0 to 1. This function enables the transition from lag phase to exponential phase by gradually diminishing the effect of the previous environment embedded in Qt. Similarly, the adjustment N t enables the transition from the exponential to the function bt 1 ÿ Nmax stationary phase, inhibiting growth when N t approaches Nmax. According to the Baranyi model, the lag time can be calculated as follows: 1 1 ÿln0 ln 1 13:10 max Q0 max The first term on right-hand side of Eq. 13.8 (i.e., at

The definition of lag time presented in Eq. 13.10 is related to the concept of relative lag time (RLT), which assumes that the product of lag time () and the maximum specific growth rate (max) is constant. This product is a measure of the amount of work that a bacterial cell has to do before it can initiate growth. As such, the dependence of the lag time on the environmental conditions can be derived from the dependence of the maximum specific growth rate on the environmental conditions as long as the pre-culturing conditions remain unchanged. The parameter 0 has the role of an initial value, quantifying the history of the cells. Its value ranges from 0 to 1. When 0 ! 0, there is no growth and the lag time is infinite. Conversely, when 0 ! 1, there is no lag phase and growth will start immediately (see Fig 13.7). The assumption of the product of max constant, although practical, may not always be valid under all conditions and should therefore be used with care (Delignette-Muller, 1998). The popularity of the Baranyi model resides in its capacity to provide good fits. It is also applicable under dynamic environmental conditions and from a biological point of view, most of the model parameters are interpretable (Lopez et al., 2004; Pin et al., 2002; Van Impe et al., 2005).

Fig. 13.7 Effect of the ìnitial physiological state' parameter (0 ) on duration of lag phase according to the Baranyi model.



Fig. 13.8

429

Graphical representation of the Buchanan three-phase linear model.

A simpler three-phase linear model was proposed by Buchanan et al. (1997), which can be expressed by the following system of equations: 8 t > < log10 N0 13:11 log10 N log10 N0 t ÿ < t < tmax > : log10 Nmax t tmax where N0 is the initial population density (cfu/ml), Nmax is the maximum population density supported by the environment (cfu/ml), t the elapsed time, the time when the lag phase ends (h), tmax the time when the maximum population density is reached (h), the specific growth rate (log10 cfu mlÿ1 hÿ1) (see Fig. 13.8). When fitting the three-phase linear model to experimental data, the parameters to estimate are , tmax, N0 and Nmax. The (maximum) specific growth rate is then calculated as: Nmax ÿ N0 13:12 tmax ÿ This model embeds stochastic aspects with respect to the lag time, i.e. it assumes that each individual cell has a lag time i that depends on its adaptation time i and its generation time tmi , so that i i tmi . If the variances of i and tmi are large, the transition from lag phase to exponential phase for the whole population will be smooth, while for very small variances, the transition will be abrupt (Buchanan et al., 1997). Another type of primary growth model commonly used is the four-parameter Logistic model, which is given by Eq. 13.13: Nmax ÿ Nmin 13:13 log10 N t log10 Nmin 1 expÿmax t ÿ ti


430


where Nmin is the minimum asymptotic cell concentration (cfu/ml), ti is the time (h) when half the maximum cell concentration is reached, and the other parameters have been already explained. From the four-parameter Logistic model, the population lag time has been derived as follows (Dalgaard, 1995): 1 Nmax Nmax expmax ti ÿ1 13:14 ln ti ÿ max Nmax Nmin expmax ti A three-parameter version of the Logistic model, which does not consider the lag phase (i.e., the parameter Nmin is not included), has been used by Dalgaard (1995) to calculate the end of shelf life in seafood products, as the time the specific spoilage organisms (SSO) requires for multiplication from N(0) to a minimal spoilage level (MSL), as given by Eq. 13.15: shelf life (days)

log10 (MSL) ÿ log10 N 0 ln10 max 24

13:15

Obviously, application of Eq. 13.15 requires the determination of the MSL. This can be achieved by controlled storage trials, where the concentration of the SSO is monitored over time, and then correlated to the time where products are spoiled as determined by sensory rejection. For example, Dalgaard et al. (1997) determined that cod fillets stored in modified atmosphere conditions at 0 ëC, on average, spoiled four generation times (tg) after the SSO, Photobacterium phosphoreum, reached the inflection point (ti) of the three-parameter Logistic growth curve. A spoilage criterion was then defined as the storage time equal to ti 4tg as shown in Eq. 13.16: Nmax ÿ1 4 ln2 ln N 0 spoilage criterion 13:16 max The correlation between a cell concentration and the point of sensory rejection is not always straightforward as variable levels of cells may be measured when spoilage occurs. This makes it difficult to predict the end of shelf life accurately. For particular product types, it makes more sense to predict the end of shelf life as the time required for the SSO to produce a certain level of a metabolite that will be responsible for sensory rejection. For example, Dalgaard (1995) reported that in packed cod, a concentration of approximately 30 mg of N-trymethylamine (N-TMA) per 100 g is typically found at the time of sensory rejection. Therefore, the end of shelf life was predicted as the time required for P. phosphoreum and Shewanella putrefaciens (the SSOs identified for spoilage) to produce 30 mg of N-TMA, as given by Eq. 13.17: shelf life (days) 30 mg N ÿ TMA/100 g N 0 ÿ log10 N 0 ln10 log10 YTMA=cfu 100 13:17 max 24



431

Equation 13.17 combines the exponential growth model (see Eq. 13.15) for the SSO with the yield factor for TMA production (YTMA/cfu) which needs to be measured experimentally with appropriate analytical techniques. Individual cell lag time While the growth rate has been extensively studied in the literature, the study of the lag time is rather recent in comparison. This observation seems quite surprising as the characterisation of the lag phase (in combination with the growth rate) will determine whether growth of spoilage or pathogenic microorganisms can occur before the end of a target shelf life. It is therefore necessary to better understand the lag phase and the factors most likely to affect it. The development of such knowledge is only recent, which is the reason why the study of the lag time has received relatively less attention than the study of the growth rate. There is a greater difficulty in estimating the lag time as this not only depends on the environmental conditions but also on the previous history of the cell, the composition of the medium in which the inoculum was cultivated or the stress conditions before inoculation into the new medium. These various factors make the lag time inevitably more prone to variability. In the previous subsection, we presented several population-based lag models, which are derived from deterministic primary growth models. However, it has been shown that the lag time depends on the inoculum level as the inoculum decreases to less than 100 cells per ml (Augustin et al., 2000; Robinson et al., 2001). This may be caused partly by the fact that not all the cells are able to divide, so what we observe is not a `true lag' phase but rather an àpparent' one (Pirt, 1975). The behaviour of an individual cell can be represented by using stochastic models. Such models have the major advantage of incorporating differences between individual cells so that a change in individual cell behaviours can be used to predict a change in population behaviour. Two stochastic growth models are mentioned in the following: the Baranyi lag phase model and the McKellar continuous discrete continuous (CDC) model. The Baranyi lag phase model (Baranyi, 2002) was developed to describe the transition between lag phase and exponential phase of the cell population by applying an integral formula for lag time distribution of the individual cells in a bacterial population. Assuming an exponential distribution for individual lag times, the model successfully predicted the growth curves of Brochothrix thermosphacta. However, Baranyi indicated that the assumption about exponential distribution of individual lag times had not been experimentally validated as it was not in practice feasible. Developed for individual cells, the McKellar continuous discrete continuous (CDC) model (McKellar et al., 2002a, 2002b; McKellar and Knight, 2000) combines a continuous adaptation phase with a discrete step making the transition to a continuous exponential growth phase. In this model, each individual cell follows a positively truncated normal distribution. At the end of their lag time, the discrete step shifts each cell into an exponential growth phase. The population growth curve is then derived from the pool of individual cell growth and the population lag time can be estimated. The


432


advantage of the McKellar CDC model is that the model is dynamic in the population lag, and the specific growth rate max can change with the environmental conditions during the lag phase. This approach facilitates the study of single cell behaviour subjected to certain environmental factors such as pH and temperature (McKellar et al., 2002a). Secondary growth models Secondary models describe the effect of environmental factors on one or more parameters of the primary models (e.g., lag time, growth rate, death rate). Most of the studies have focused on the description of the bacterial growth rate as it is considered an implicit characteristic of an organism in a specific environment. To homogenise the variance of the parameters, square root and logarithmic transformations were usually applied on the growth rate before fitting. A number of models have been used for this purpose, including response surface models, square-roots and gamma-type models. Each type of model has its supporters, merits and disadvantages being often discussed in the literature. Equation 13.18 shows an example of a response surface model, proposed by Gibson et al. (1988) for the effects of temperature, pH and NaCl on the growth rate of Salmonella: ln y a b1 s b2 t b3 p b4 s2 b5 t2 b6 p2 b7 st b8 sp b9 tp e 13:18 where ln y denotes the natural logarithm of the modelled growth response variable (i.e., the modified Gompertz model parameters B or M); s, t, and p represent NaCl (% w/v), temperature (ëC) and pH, respectively; a, b1, b2, . . . b9 are the coefficients to be estimated; and e represents a random error, assumed to have a zero mean and constant variance. Figure 13.9 depicts generation time predictions for Salmonella at pH 6.1 as a function of temperature and NaCl concentration derived from the secondary model presented in Eq. 13.18 and including the observed values reported by Gibson et al. (1988). One advantage of the polynomial models, such as the one illustrated in Eq. 13.18, is that they allow in almost every case the development of a model for any environmental factor to be taken into account. However, the disadvantages of this approach lie in the relatively large number of parameters and their lack of biological significance. An example of a square root type model is that proposed by Ratkowsky et al. (1983), which relates the specific bacterial growth rate to temperature (in the entire temperature range allowing the growth of the considered microorganism), and is given by: p 13:19 max b T ÿ Tmin f1 ÿ expc T ÿ Tmax g where T is the temperature, b and c are model parameters without biological meaning, Tmin and Tmax are the theoretical minimum and maximum temperatures for growth, respectively. The inclusion of other factors such as pH



433

Fig. 13.9 Salmonellae generation time model at pH 6.1 with temperature and NaCl concentration as controlling factors. Solid squares are observed values (secondary model and data from Gibson et al., 1988).

and water activity in square root and gamma models is discussed in Section 13.4.2. Secondary lag models Secondary lag time modelling is complicated, because the lag time is influenced not only by actual environmental factors, but also by the history or preincubation condition of the cells (i.e., their physiological state) and the inoculum size. Thus, during the secondary modelling, the lag phase can be taken into account in different ways: (i) models that assume that the product max is constant (as shown, for instance, in Eq. 13.10), and (ii) models where the lag time and growth rate are modelled independently. Assuming that the product of the lag time and the maximum specific growth rate max is constant, the dependence of the lag time on the environmental conditions can be derived from the dependence of the maximum specific growth rate on the environmental conditions as long as the pre-culturing conditions remain unchanged. As such, the concept of relative lag time (RLT, Eq. 13.20) is useful when developing secondary lag models, but, as mentioned before, it should be used with caution, because distributions of RLT rather than


434


single constant values have been reported for various organisms under a wide range of experimental conditions (Augustin and Carlier, 2000a; DelignetteMuller, 1998; Ross, 1999). In spite of that fact, the RLT concept simplifies the growth modelling process, because it enables the effects of and max to be predicted by a single growth model. max 13:20 RLT ln 2 The cardinal parameter models (CPMs) (Le Marc et al., 2002; Rosso et al., 1995) can be used in such a way and have become an important group of empirical secondary models. The general principle of CPMs is to use model parameters that have a biological interpretation. CPMs rely on the assumption that the inhibitory effects of the environmental factors have a multiplicative effect on the maximum specific growth rate max (see Section 13.4.2 for further details). The inhibitory effect of each environmental factor is represented by a numerical function taking values between 0 (at inhibitory condition for that factor) and 1 (at optimal condition for that factor). Square-root type models have also been used to derive lag time from existing growth rate models based on the same assumption that the product max is constant (Devlieghere et al., 2000; Zwietering et al., 1991). Davey (1991) also applied this concept to change the linear Arrhenius model into a lag time model. Other authors developed secondary models independently for the generation time and the lag time. For example, Gibson et al. (1988), McClure et al. (1993), Zaika et al. (1998) followed such an approach using polynomial models, and Geeraerd et al. (1998) and GarcõÂa-Gimeno et al. (2002) followed this approach using artificial neural networks. It can be easily argued that pre-growth conditions are not taken into account in the estimation of the lag phase from these models. This will subsequently make the lag time estimate subject to greater variability. 13.4.2 Quantification of hurdle effects The concept of `hurdle technology' was developed by Professor Leistner and colleagues (Leistner, 1995; Leistner and Gorris, 1995). The basis of this approach is that the microbiological safety of food products can be achieved by applying a combination of different preservative factors (called hurdles). This hurdle concept is of particular interest for mildly preserved foods with minimal inactivation/intervention during processing (e.g., mild heat or fermentation process; Leistner, 2000). Predictive microbiology models should aim at quantifying the effects of the combined hurdles on the bacterial growth rate and the ability of the microorganisms to grow under specific environmental conditions. In this section, we will focus on gamma type models and growth/ no-growth (G/N-G) models which are commensurate with the assumptions underlying the `hurdle concept' (Bidlas and Lambert, 2008; McMeekin et al., 2000).



435

McMeekin et al. (1987) observed that temperature and water activity had independent effects on the growth rate of Staphylococcus xylosus and that their overall effect could be obtained by multiplying separate temperature and aw terms: p max b T ÿ Tmin aw ÿ aw;min 13:21 where Tmin and aw,min are the theoretical minimum temperature and minimum water activity for growth, respectively and b is a model parameter without biological meaning. The independence of the effects of temperature and aw on the bacterial growth rate was also demonstrated for various microorganisms by Davey (1989). Adams et al. (1991) drew the same conclusions for the effects of temperature and pH on the growth rate of Yersinia enterocolitica. Zwietering et al. (1996) formalised this independent/multiplicative effect through the `Gamma concept'. The approach consists in multiplying the separate effects (normalised between 0 and 1) of the environmental factors on the bacterial growth rate. For temperature, pH and water activity, the model is written as follows: max opt T pH aw

13:22

where opt is the growth rate at optimum conditions and (T), (pH) and (aw) are the relative effects of temperature, pH and water activity, respectively. The model is consistent with the hurdle concept: each environmental factor produces its own hurdle and the overall effect is obtained by multiplying the individual hurdles (Witjzes et al., 2001). If the assumption of multiplicative effects usually provides reasonable results for the growth rate (Zwietering, 2002), experimental observations suggest a synergistic rather than an independent effect at the growth/no-growth boundary (McMeekin et al., 2000). For example, it has been observed that the minimum pH and the water activity at which growth of L. monocytogenes and E. coli could be observed were increasing at low temperatures (Salter et al., 2000; Tienungoon et al., 2000). For this situation, quantification of the hurdles requires the modelling of the boundary between growth and no growth. As mentioned previously (see Section 13.2.2), logistic regression procedures have been commonly used to build probability of growth models for both pathogens and spoilage microorganisms. Available growth boundary models include equations for Shigella flexneri (Ratkowsky and Ross, 1995), E. coli (Presser et al., 1998; Salter et al., 2000), L. monocytogenes (Tienungoon et al., 2000), Salmonella (Koutsoumanis et al., 2004), and S. aureus (Valero et al., 2009). Deterministic approaches for which a binary response (either growth or nogrowth) rather than a probability were also proposed by Augustin and Carlier (2000b) and Le Marc et al. (2002). In these models, the equation for the G/N-G interface is defined on the cardinal parameters or the functions used in the growth rate modelling. Mejlholm and Dalgaard (2009) and Mejlholm et al. (2010) extended the approach developed by Le Marc et al. (2002) to model the effects of 12 environmental factors on the growth rate and the growth/no-growth boundary of L. monocytogenes in lightly preserved seafood and ready-to-eat


436


meats. Most of the authors reported an abrupt transition between the growthpermitting conditions and the no-growth conditions. Therefore at the growth limits, small variations in the factors have a huge impact in terms of microbiological stability. G/N-G models can be used to identify which levels of factors will guarantee the microbiological stability and quantify the effects of variations in the formulations and storage of food products. 13.4.3 Relative rate of spoilage models The concept of relative rate of spoilage (RRS) is defined as the shelf life at a reference temperature divided by the shelf life at a temperature of interest T. As such, mathematical RRS models can be developed on the basis of shelf life data obtained during product storage trials at different temperatures using sensory evaluation as criterion for evaluating shelf life. Therefore, they are not really predictive microbiological models, as they do not consider the kinetics of growth of microorganisms, nor the types of reactions that cause spoilage. However, they are practical, easy to use, and have proven to be useful in predicting shelf life of fresh and preserved seafood at different storage temperatures (Dalgaard, 2002). When developing RRS models, the only information required is the product shelf life determined at one single known storage temperature (as determined by sensory evaluation), whose reciprocal value is the rate of spoilage (RS, expressed in daysÿ1). RS values are then fitted to different types of models (Eqs 13.23 to 13.25) to estimate temperature characteristic parameters, which will allow for the prediction of shelf life at different storage temperatures. The main types of RRS models used are: Square-root model shelf life at Tref RRS shelf life at T

T ÿ Tmin Tref ÿ Tmin

2

Exponential model shelf life at Tref expa T ÿ Tref RRS shelf life at T

13:23

13:24

Arrhenius model

shelf life at Tref ÿEA 1 1 exp RRS ÿ shelf life at T R TK Tref ;K

13:25

where Tmin (ëC), a (Cÿ1), and EA (kJ molÿ1) are the temperature characteristics for the square-root, the exponential and the Arrhenius models, respectively; Tref and Tref,K are the reference temperature (ëC and Kelvin, respectively), T and TK are the temperature (ëC and Kelvin, respectively); and, R is the universal gas constant (8.31 J molÿ1 Kÿ1).



437

A few RRS models have been implemented in the Seafood Spoilage and Safety Predictor (Dalgaard et al., 2002), software made available free of charge by the National Institute of Aquatic Resources in Denmark (http:// sssp.dtuaqua.dk/), which will be discussed later (see Section 13.5.2).

13.5 Usage considerations and access to predictive microbiology electronic resources Several predictive microbiology electronic resources and software systems are nowadays available in the public domain around the globe. They fall mainly into three general categories: (i) searchable databases, (ii) fitting tools, and (iii) userfriendly applications of mathematical models (i.e., tertiary models) to predict the behaviour of microorganisms under different conditions in food and food environments. The various systems have been built in different ways and for different purposes. They also have quite different scopes and levels of quality control. Some of the systems have been developed with the support of government agencies in different countries, whereas others have been supported by academic institutions, independent research organisations or private companies. The various software systems cover a range of microorganisms and conditions, which sometimes overlap. However, as the systems have not been built in exactly the same way and the mathematical models for seemingly identical microorganisms or species are different, the predictions can be different. This highlights that predictions should be interpreted with care, recognising the limitations and assumptions behind each model and the conditions under which each model is valid, thereby avoiding misinterpretation of model predictions. Available predictive microbiology application software packages are described below. 13.5.1 ComBase and ComBase modelling toolbox ComBase (Baranyi and Tamplin, 2004) is a relational database of predictive microbiology information (http://www.combase.cc). ComBase contains a large volume of data on bacterial growth, survival and death under a range of conditions of temperature, pH, water activity and atmosphere as well as in a variety of different foods, including meat and fish, dairy products, fruit and vegetables. Using an Internet interface, users identify criteria that they are interested in for a food microbiology scenario(s). This includes identifying a type or species of organism, a type or class of food, pH, temperature, water activity (or NaCl concentration), and specific food conditions. Alternatively, ComBase customers may be interested in retrieving data donated by a specific source (publication, organisation or researcher). The data in ComBase can be used not only for model development but also to validate newly created models, to compare the predictions of established models developed in a specific medium (e.g., broth) against data recorded in other matrices, or to identify boundary growth values of specific organisms in a variety of foods.


438


This initiative is a collaboration between the Food Standards Agency in the UK, the Institute of Food Research (IFR) in the UK, the USDA Agricultural Research Service, and the Australian Food Safety Centre of Excellence. The datasets have been donated by microbiology laboratories in academia, government and industry and have been derived from the published literature. The database is continuously updated. Through the ComBase website, users can find the ComBase Modelling Toolbox (http://www.combase.cc/toolbox.html) which includes three main components: (i) ComBase Predictor, (ii) Perfringens Predictor, and (iii) DMFit. ComBase Predictor ComBase Predictor is, in fact, a modified and expanded version of another stand-alone application previously developed by IFR, known as Growth Predictor, which in itself was the successor of the former Food MicroModel (McClure et al., 1994). It has a set of twenty growth models, seven thermal death models and two non-thermal survival models (see Table 13.1). ComBase Predictor allows for predictions under static and dynamic temperature conditions. In this case, the user simply enters a time±temperature profile (e.g., from a temperature data-logger) into the main interface of ComBase Predictor, and the predictions are presented both in graphic and tabulated form. The predictions can be easily copied into any spreadsheet software for further analyses. Another improvement over the Growth Predictor is the fact that ComBase Predictor can simultaneously produce predictions for up to four microorganisms, thereby facilitating comparisons amongst several scenarios. Perfringens Predictor Perfringens Predictor is specifically designed to predict the response of Clostridium perfringens during the cooling of cooked meats. Predictions from Perfringens Predictor are based on the heat treatment applied to the meat product being in the range of 70 ëC (for up to 6 hours) to 95 ëC (for up to 1.5 hours). The inputs to the model are temperature (dynamic profile to < ln N0 ln N t Nmax > ÿ 1 expÿmax t ÿ lag t > lag : ln Nmax ÿ ln 1 N0 13:28 The parameters of Eq. 13.28 have already been explained in the context of other primary models described earlier in the chapter. The effect of the three main environmental factors (i.e., temperature, pH and aw) on growth of the microorganism is modelled using the Cardinal Model of Rosso et al. (1995). The input values for these three environmental factors can be entered as static (i.e., constant) or dynamic (e.g., a time±temperature profile) values. The output of this growth simulation module is presented as a list of growth parameters, a graphical display of the growth curve (including 80%, 90% and 95% confidence intervals), and a histogram of the growth rate distribution. Both the growth curve values (with confidence intervals) and the growth rate distribution can be easily exported to a MicrosoftÕ Excel file. Growth/no-growth interface simulation module This module calculates the combinations of environmental factors (temperature, pH, and aw) that represent the limits of growth/no-growth for a bacterial species. There are models available for the following microorganisms: Bacillus cereus, Clostridium botulinum (both proteolytic and non-proteolytic), Clostridium perfringens, Escherichia coli, Listeria monocytogenes, Salmonella, and Staphylococcus aureus. Additionally, the user can create a personalised model by manually entering the cardinal values for temperature (minimum, optimum, maximum), pH and aw (minimum, optimum). The personalised input screen also gives the option of defining the inhibitory effect of an organic acid, which the user can define by entering the following parameters: minimum inhibitory concentration (MIC, in mol/L), pKa and alpha (a shape parameter). The user enters a range of values for the two environmental factors for which a growth/no-growth interface prediction is desired (e.g., temperature and pH), and specifies a fixed value for the third factor (e.g., aw). The simulation result is provided as a graph showing three iso-probability growth curves (10%, 50%, and 90%). On this plot, the user can also add his/her own observations for comparison with the predicted probabilities of growth. A matrix containing the percentage probability of growth can be downloaded to MicrosoftÕ Excel. This module can be useful to evaluate whether a given pathogenic microorganism is capable of growing in a product (as defined by temperature, pH and/or aw), to evaluate the effect of changing a formulation, and to optimise the formulation (pH and/or aw) or the storage temperature.



445

Other relevant features of Sym'Previus In addition to the three modules described above, Sym'Previus also includes the following facilities. Growth curve fitting tool This module allows the user to fit experimental growth data to the primary Logistic model shown in Eq. 13.28. The input (bacterial counts vs time) can be imported from a spreadsheet, or simply typed in directly on the input screen. The outputs are fitted parameters and standard deviations for growth rate, lag time, initial population and maximum population. The software displays the fitted curve and experimental inputs graphically. The fitted growth curve can be exported to MicrosoftÕ Excel. The second fitting facility in this module allows for secondary modelling of the growth rate as a function of environmental factors (temperature, pH or aw). Input values (growth rate vs one selected environmental factor) are fitted to the Cardinal Model of Rosso et al. (1995). The outputs are the fitted parameters for optimum growth rate (opt ), and the minimum, optimum and maximum values (temperature, pH or aw) for growth with corresponding standard deviation for each fitted parameter. A graphical depiction of the fitted model is also displayed on-screen. Probabilistic module This module simulates the evolution of a microbiological contaminant throughout the food product's shelf life and indicates the probability of exceeding a critical threshold at different stages of the shelf life. The inputs to this module are specific to the bacterial species, the characteristics of the food product and the environment it is in, as well as the processing conditions, so that the results can be tailored for each product and/or process relevant for a food company. The module is organised as a series of four sequential steps. Each step consists of a screen where specific questions need to be answered or for which information/data need to be entered. Outputs include the probability of growth in the food (per portion size), simulation of growth (with 90% confidence bands), evolution of contamination density (distribution) vs time, prevalence distribution of contaminated products at the end of shelf life, simulation of lag time and growth rate, factors affecting the growth simulation (i.e., evolution of gamma values and environmental factors), comments and remarks, and estimated contamination level at the end of shelf life. Thermal destruction simulation module This module simulates the destruction kinetics of a bacterial species during heat treatment, and calculates the reduction rate at the end of the treatment and the probability of a microorganism to survive. 13.5.4 Pathogen Modeling Program The Pathogen Modeling Program (available at: http://www.ars.usda.gov/naa/ errc/mfsru/pmp) is a stand-alone software package of microbial models and a


446


research product of the Microbial Food Safety Research Unit (MFS) of the Agricultural Research Service (ARS) of the US Department of Agriculture. The model is available free of charge. The latest version (PMP 7.0) currently contains 38 models for 11 bacterial pathogens, including one model for spoilage flora in chicken (inactivation by irradiation). Table 13.2 presents a summary of the microorganisms and models included in PMP 7.0. Most models included in the PMP are isothermal; however, it contains four models that are able to predict the growth of C. botulinum and C. perfringens under time-varying temperature conditions (cooling models). Once downloaded, user-friendly features allow users to easily input food-relevant criteria and then to receive predictions about how pathogenic bacteria react to specific food environments. To further assist food processors in meeting regulatory requirements, references are provided for each model via direct Internet access to PDF files. A drawback of the PMP is the lack of information from validation studies showing the performance of models in specific foods as well as more general facilities to predict the effect of time-varying temperature conditions on growth and inactivation. Recently, an online version of PMP has been made available (http://pmp.arserrc.gov/PMPOnline.aspx), which does not include all the original models implemented in the stand-alone version; however, it includes new models that are not covered by the stand-alone version (e.g., a model for the transfer of Listeria monocytogenes during slicing of salmon).

13.6

Future trends

In spite of the scepticism that surrounded the field of predictive microbiology during the early years of research, predictive models are nowadays well established and widely accepted by the food industry, researchers in academia and government organisations. The international efforts to create databases such as ComBase have been fundamental to facilitate the development and validation of new models, but more importantly, to support the operation of quantitative microbiological risk assessments. However, there are still many models that could potentially be very useful to support product and process design in the food industry, but that have not been made available in a user-friendly software application. This is particularly true for models related to spoilage organisms. Other than a handful of applications (such as the Seafood Spoilage and Safety Predictor described in Section 13.5.2), most available systems and databases include models and data for pathogenic organisms only. Although many predictive spoilage models are available in the public domain, the majority of those are still only available via scientific publications. This poses a considerable barrier for wider application of models that are potentially very useful for the food industry. Although in some cases the implementation of published models in a spreadsheet format by potential users is possible, this can sometimes be limited by the lack of data (used for model development and/or validation) or other


Table 13.2 Summary of microorganisms and models included in the stand-alone Pathogen Modeling Program (PMP 7.0) Microorganism

Physiological event

Response variablesa

Independent variables and ranges


Temperature

pH

NaCl (% w/v)

Other

5±42 ëC

5.3±7.3

0.5±4.5

NaNO2 (0±150 ppm)

5±30 ëC

5.3±7.3

0.5±3.5

NaNO2 (0±150 ppm)

5±42 ëC

4.5±7.5

0.5±5.0

NaNO2 (0±150 ppm)

10±42 ëC

5.0±9.0

0.5±5.0

NaNO2 (0±150 ppm)

Time±temperature history (up to 50 data pairs) 15±34 ëC

±

±

±

5.0±7.2

0.0±4.0

±

Time to n-log reduction

70±90 ëC

5.0±7.0

0.0±3.0

Na-pyrophosphate (0.0±0.3%w/v)

Probability of growth (time to turbidity)

Pmax, ,

5±28 ëC

5.0±7.0

0.0±4.0

±

Length of lag phase

Lag (shelf life of fresh fish in modified atmospheres)

4±30 ëC

±

±

±

Growth (anaerobic)

GT, lag, time to n-log increase Net growth (in beef broth)

19±37 ëC

6.0±6.5

1.0±3.0

Time±temperature history (up to 50 data pairs)

±

±

Na-pyrophosphate (0.1±0.3%w/v) ±

Aeromonas hydrophila

Growth (aerobic)

Aeromonas hydrophila

Growth (anaerobic)

Bacillus cereus (vegetative cells) Bacillus cereus (vegetative cells) Proteolytic Clostridium botulinum (spores)

Growth (aerobic)

Proteolytic Clostridium botulinum Non-proteolytic Clostridium botulinum (spores) Non-proteolytic Clostridium botulinum (spores) Non-proteolytic Clostridium botulinum (spores ± types E & F) and aerobic competitive flora Clostridium perfringens (vegetative cells) Clostridium perfringens (spores)

Probability of growth (time to turbidity) Heat inactivation

Growth (anaerobic) Growth (during cooling)

Growth (during cooling)

GT, lag, time to n-log increase GT, lag, time to n-log increase GT, lag, time to n-log increase GT, lag, time to n-log increase Net growth Pmax, ,

Table 13.2 Continued Microorganism

Physiological event

Response variablesa

Independent variables and ranges Temperature


Clostridium perfringens (spores)


Clostridium perfringens (spores)


Escherichia coli O157:H7 Escherichia coli O157:H7 Escherichia coli O157:H7 Escherichia coli O157:H7 Escherichia coli O157:H7 Listeria monocytogenes

Growth (aerobic) Growth (anaerobic) Heat inactivation Survival Inactivation by Gamma-irradiation Growth (aerobic)

Listeria monocytogenes

Growth (anaerobic)


Heat inactivation


Survival

Salmonella spp.

Growth (aerobic)

Salmonella spp.

Survival

Net growth (in cured beef)

Time±temperature history (up to 50 data pairs) Net growth Time±temperature (in cured chicken) history (up to 50 data pairs) GT, lag, time to 5±42 ëC n-log increase GT, lag, time to 5±42 ëC n-log increase Time to n-log 55±62.5 ëC reduction Time to n-log 4±37 ëC reduction Net reduction ÿ20±10 ëC (in beef tartar) GT, lag, time to 4±37 ëC n-log increase GT, lag, time to 4±37 ëC n-log increase Time to n-log 55±65 ëC reduction Time to n-log 4±42 ëC reduction GT, lag, time to 10±30 ëC n-log increase Time to n-log 5±42 ëC reduction

pH

NaCl (% w/v)

Other

±

±

±

±

±

±

4.5±8.5

0.5±5.0

NaNO2 (0±150 ppm)

4.5±8.5

0.5±5.0

NaNO2 (0±150 ppm)

4.0±8.0

0.0±6.0

3.5±7.0

0.5±15.0

±

±

Na-pyrophosphate (0.0±0.3%w/v) NaNO2 (0±75 ppm); lactic acid (0.0±2.0%w/w) Irradiation dose (0±2 kGy)

4.5±8.0

0.5±5.0

NaNO2 (0±150 ppm)

4.5±8.0

0.5±5.0

NaNO2 (0±150 ppm)

4.0±8.0

0.0±6.0

3.2±7.3

0.5±19.0

5.6±6.8

0.5±4.5

Na-pyrophosphate (0.0±0.3%w/v) NaNO2 (0±150 ppm); lactic acid (0.0±2.0%w/w) ±

3.5±7.2

0.5±16.0

NaNO2 (0±200 ppm)

Salmonella Typhimurium

Growth ± previous NaCl (aerobic)


Growth ± previous temperature (aerobic)


Growth ± previous pH (aerobic) Inactivation by Gamma-irradiation

Salmonella Typhimurium ß Woodhead Publishing Limited, 2011


Inactivation by Gamma-irradiation



Shigella flexneri

Growth (aerobic)

Shigella flexneri

Growth (anaerobic)

Staphylococcus aureus

Growth (aerobic)


Growth (anaerobic)


Survival

Yersinia enterocolitica

Growth (aerobic)

Spoilage (normal flora)


a

GR, lag, time to 10±40 ëC n-log increase (in sterile chicken breast) GR, lag, time to 16±34 ëC n-log increase (in sterile chicken breast) GR, lag, time to 15±40 ëC n-log increase Net reduction (in ÿ20±10 ëC sterile chicken MDM) Net reduction (in non-sterile chicken MDM) Net reduction (in non-sterile chicken legs) GT, lag, time to n-log increase GT, lag, time to n-log increase GT, lag, time to n-log increase GT, lag, time to n-log increase Time to n-log reduction GT, lag, time to n-log increase Net reduction (in non-sterile chicken MDM)

±

±

Previous growth NaCl (0.5±4.5%w/v)

±

±

Previous growth temperature (16±34 ëC)

5.2±7.4

±

±

±

ÿ20±10 ëC

±

±

Previous growth pH (5.7±8.6) Irradiation dose (0±3.6 kGy); two models (air and vacuum atmospheres) Irradiation dose (0±3.6 kGy)

ÿ20±10 ëC

±

±

Irradiation dose (0±3.6 kGy)

10±37 ëC

5.0±7.5

0.5±5.0

NaNO2 (0±150 ppm)

12±37 ëC

5.5±7.5

0.5±4.0

NaNO2 (0±150 ppm)

10±42 ëC

4.5±9.0

0.5±12.5

NaNO2 (0±150 ppm)

12±42 ëC

5.3±9.0

0.5±16.5

NaNO2 (0±150 ppm)

4±37 ëC

3.0±7.0

0.5±20.0

5±42 ëC

4.5±8.5

0.5±5.0

NaNO2 (0±200 ppm); lactic acid (0.0±1.0%w/w) NaNO2 (0±150 ppm)

ÿ20±10 ëC

±

±

Irradiation dose (0±3.6 kGy)

GT generation time; GR growth rate; Pmax maximum probability of growth, growth rate of individual samples showing growth; time to ÝPmax (inflection point), MDM mechanically deboned meat; where there is no indication of a specific food matrix or medium in brackets it means the model was developed in laboratory media (broth).

450


necessary modelling details, which are typically not included in the publication due to space limitations. Moreover, small processors may not have sufficient and/or qualified resources to use published models in an effective way. Also, in some cases, the model itself may indeed require software tools more sophisticated than a spreadsheet format in order to be implemented effectively. The submission of raw data to databases such as ComBase for newly developed and published models should be taken as `good modelling practice'. Moreover, details about conditions of the experiment, strains, growth or inactivation media, etc., used in the development of a model should also be reported clearly in order to prevent improper use of the data. Likewise, the presentation of experimental results used to develop a predictive model should, as much as possible, include information about sources of variability and the uncertainty of the results obtained. Of particular relevance to the prediction of shelf life, the effective integration of temperature data under real supply chain conditions with predictive models represents another important gap. Temperature conditions vary amongst different geographical regions and along different stages of the supply chain. However, it is still common practice to use single temperature values, simulating the worst-case temperature conditions along the supply chain, as inputs to deterministic predictive models for estimation of shelf life. This limits product design as the predictions obtained in this way are overly conservative. Instead, temperature distribution data collected during the different stages of the supply chain (e.g., storage, transport, distribution and retail) could be directly used as probabilistic inputs to predictive models to allow for a more realistic estimation of shelf life. To that extent, more studies dealing with stochastic models and associated implementation tools are necessary in order to account for the variation of supply chain data. Moreover, the ability to collect temperature data in real time for process control in situ remains an important challenge in the estimation of shelf life. In order to overcome this limitation, the use of radio frequency identification (RFID) tools (tags and readers) has gained considerable interest in recent years, not only for traceability of products through the supply chain, but also as a tool to collect product information (e.g., temperature) in real time. Other tools such as time±temperature integrators (TTIs) for monitoring temperature history and shelf life through the supply chain have been used in the food industry for several years, particularly in chilled foods. However, in recent years, excellent developments in the application of enzymatic TTIs (e.g., VITSAB Check PointÕ) and microbiological TTIs (e.g., Cryolog TRACEOÕ and (eO)Õ) to validate predictive spoilage models have been reported (see, for instance, Ellouze and Augustin, 2010; Nuin et al., 2008). This will allow food processors to bridge the gap between product traceability and management of food safety and stability. Development and validation of models in real food matrices continue to be another important gap that needs to be filled. Many of the models available in the public domain have been developed in liquid laboratory media (broth), which, oftentimes, supports the growth of microorganisms better than actual



451

food matrices. Hence, many of these models are perceived by the food industry as very conservative, to the point that they have limited practical application. Although many efforts have been made in recent years to generate microbiological data in food matrices for model development and validation, there is still a need for more models developed in food materials for a wider range of applications. Likewise, understanding of microbial ecology in real food matrices, particularly as it relates to the interactions between populations of microorganisms under real manufacturing and shelf life conditions, remains a gap which needs to be filled. This knowledge would contribute to increasing the reliability of predictive models to support the design of microbiologically safe and stable food products. Undoubtedly, the development and application of predictive models in food microbiology have come a long way over the last 30 years. However, for real added value to food innovations at the stage of product/process design, predictive microbiological models need to be integrated with sensory response models, physical property models and process models with a view to providing the necessary tools for ìn-silico' design of high-quality products. We are long way from achieving that. This is a need that is gaining increasing attention in the food industry as new preservation strategies need to be evaluated in a short amount of time in order to meet consumer demands for convenient, healthy, mildly preserved, and chemical-free labelled products. Computer-aided product/ process design, via an integrated user-friendly software application, could not only reduce considerably the time a new product takes to reach the market (from concept to delivery to consumer), but it could also allow for more realistic estimates of shelf life, reducing the potential for product wastage and for the occurrence of incidents in the market place. On a more futuristic note, it is worth highlighting that the evolution of science in the area of food microbiology has taken us to a point where there is a plethora of data and information on microbial physiology, genomics, microbial ecology and quantitative microbiology available in the scientific literature or other public sources. These areas will need to be linked in a systematic way for the development of models that can describe in a mechanistic way the behaviour of microorganisms in foods. This is by no means a small challenge, and it will require collaboration between food scientists, predictive modellers, network analysts and systems biologists to interpret and process the vast amount of bioinformatics information available in the context of realistic conditions found in food manufacturing.

13.7

Acknowledgements

The authors gratefully acknowledge Dr Clive de W. Blackburn, author of the first edition of this chapter and esteemed colleague at Unilever.


452

13.8


References

and EASTER M C (1991), `Modelling the effects of pH, acidulant and temperature on the growth rate of Yersinia enterocolitica', Journal of Applied Bacteriology, 71, 65±71. AUGUSTIN J-C and CARLIER V (2000a), `Mathematical modelling of the growth rate and lag time for Listeria monocytogenes', International Journal of Food Microbiology, 56, 29±51. AUGUSTIN J-C and CARLIER V (2000b), `Modelling the growth rate of Listeria monocytogenes with a multiplicative type model including interactions between environmental factors', International Journal of Food Microbiology, 56, 53±70. AUGUSTIN J-C, BROUILLAUD-DELATTRE A, ROSSO L and CARLIER V (2000), `Significance of inoculum size on the lag time of Listeria monocytogenes', Applied and Environmental Microbiology, 66, 1706±1710. AUGUSTIN J-C, ZULIANI V, CORNU M and GUILLIER L (2005), `Growth rate and growth probability of Listeria monocytogenes in dairy, meat and seafood products in suboptimal conditions', Journal of Applied Microbiology, 99, 1019±1042. BARANYI J (2002), `Stochastic modelling of bacterial lag phase', International Journal of Food Microbiology, 73, 203±206. BARANYI J and PIN C (1999), Èstimating bacterial growth parameters by means of detection times', 65, 732±736. BARANYI J and ROBERTS T A (1994), À dynamic approach to predicting bacterial growth in food', International Journal of Food Microbiology, 23, 277±294. BARANYI J and TAMPLIN M L (2004), `ComBase: a common database on microbial responses to food environments', Journal of Food Protection, 67, 1967±1971. BARANYI J, ROBERTS T A and MCCLURE P (1993), À non-autonomous differential equation to model bacterial growth', Food Microbiology, 10, 43±59. BARANYI J, ROSS T, MCMEEKIN T A and ROBERTS T A (1996), Èffects of parameterization on the performance of empirical models used in ``predictive microbiology''', Food Microbiology, 13, 83±91. BARANYI J, PIN C and ROSS T (1999), `Validating and comparing predictive models', International Journal of Food Microbiology, 48, 159±166. BATTEY A S and SCHAFFNER D W (2001), `Modelling bacterial spoilage in cold-filled ready to drink beverages by Acinetobacter calcoaceticus and Gluconobacter oxydans', Journal of Applied Microbiology, 91, 237±247. BATTEY A S, DUFFY S and SCHAFFNER D W (2001), `Modelling mould spoilage in cold-filled ready-to-drink beverages by Aspergillus niger and Penicillium spinulosum', Food Microbiology, 18, 521±529. BEGOT C, DESNIER I, DAUDIN J D, LABADIE J C and LEBERT A (1996), `Recommendations for calculating growth parameters by optical density measurements', Journal of Microbiological Methods, 25, 225±232. BIDLAS E and LAMBERT J W (2008), `Quantification of hurdles: predicting the combination effects ± interaction vs. non interaction', International Journal of Food Microbiology, 128, 78±88. BIESTA-PETERS E G, REIJ M W, JOOSTEN H, GORRIS L G M and ZWIETERING M H (2010), `Comparison of two optical-density-based methods and a plate count method for estimation of growth parameters of Bacillus cereus', Applied and Environmental Microbiology, 76, 1399±1405. BLACKBURN C D W and DAVIES A R (1994), `Development of antibiotic-resistant strains for ADAMS M R, LITTLE C L



453

the enumeration of foodborne pathogenic bacteria in stored foods', International Journal of Food Microbiology, 24, 125±136. BRUL S, VAN GERWEN S and ZWIETERING M (2007), Modelling Microorganisms in Food, Cambridge, Woodhead Publishing Limited. BUCHANAN R L, WHITING R C and DAMERT W C (1997), `When is simple good enough: a comparison of the Gompertz, Baranyi, and three-phase linear models for fitting bacterial growth curves', Food Microbiology, 14, 313±326. CHAPMAN B and ROSS T (2009), Èscherichia coli and Salmonella enterica are protected against acetic acid, but not hydrochloric acid, by hypertonicity', Applied and Environmental Microbiology, 75, 3605±3610. CUPPERS H G A and SMELT J P P M (1993), `Time to turbidity measurement as a tool for modeling spoilage by Lactobacillus', Journal of Industrial Microbiology, 12, 168± 171. DALGAARD P (1995), `Modelling of microbial activity and prediction of shelf-life for packed fresh fish', International Journal of Food Microbiology, 26, 305±317. DALGAARD P (2000), `Fresh and lightly preserved seafood', in Man C M D and Jones A A, Shelf-Life Evaluation of Foods, London, Aspen Publishers, Inc., 110±139. DALGAARD P (2002), `Modelling and predicting the shelf-life of seafood', in Bremmer H A, Safety and Quality Issues in Fish Processing, Cambridge, Woodhead Publishing Ltd, 191±219. DALGAARD P, MEJLHOLM O and HUSS H H (1997), Àpplication of an iterative approach for development of a microbial model predicting the shelf-life of packed fish', International Journal of Food Microbiology, 38, 169±179. DALGAARD P, BUCH P and SILBERG S (2002), `Seafood Spoilage Predictor ± development and distribution of a product specific application software', International Journal of Food Microbiology, 73, 343±349. DANG T D T, MERTENS L, VERMEULEN A, GEERAERD A H, VAN IMPE J F, DEBEVERE J and DEVLIEGHERE F (2010), `Modelling the growth/no growth boundary of Zygosaccharomyces bailii in acidic conditions: a contribution to the alternative method to preserve foods without using chemical preservatives', International Journal of Food Microbiology, 137, 1±12. DAVEY K R (1989), À predictive model for combined temperature and water activity on microbial growth during the growth phase', Journal of Applied Bacteriology, 67, 483±488. DAVEY K R (1991), Àpplicability of the Davey (linear Arrhenius) predictive model to the lag phase of microbial growth', Journal of Applied Bacteriology, 70, 253±257. DELIGNETTE-MULLER M L (1998), `Relation between the generation time and the lag time of bacterial growth kinetics', International Journal of Food Microbiology, 43, 97± 104. DEVLIEGHERE F, GEERAERD A H, VERSYCK K J, BERNAERT H, VAN IMPE J F and DEBEVERE J (2000), `Shelf life of modified atmosphere packed cooked meat products: addition of Na-lactate as a fourth shelf life determinative factor in a model and product validation', International Journal of Food Microbiology, 58, 93±106. ELLOUZE M and AUGUSTIN J-C (2010), Àpplicability of biological time temperature integrators as quality and safety indicators for meat products', International Journal of Food Microbiology, 138, 119±129. EMBORG J and DALGAARD P (2008a), `Growth, inactivation and histamine formation of Morganella psychrotolerans and Morganella morganii ± development and evaluation of predictive models', International Journal of Food Microbiology,


454

Food and beverage stability and shelf life 128, 234±243. and DALGAARD P (2008b), `Modelling the effect of temperature, carbon dioxide, water activity, and pH on growth and histamine formation by Morganella psychrotolerans', International Journal of Food Microbiology, 128, 226±233.

EMBORG J

Â NDEZ-NAVARRO F, VALERO A, HERVAS-MARTINEZ C, GUTIERREZ P A, GARCIA-GIMENO FERNA

and ZURERA-COSANO G (2010), `Development of a multi-classification neural network to determine the microbial growth/no growth interface', International Journal of Food Microbiology, 141, 203±212. GARCIA-GIMENO R M, HERVAS-MARTINEZ C and DE SILONIZ M I (2002), Ìmproving artificial neural networks with a pruning methodology and genetic algorithms for their application in microbial growth prediction in food', International Journal of Food Microbiology, 72, 19±30. GEERAERD A H, HERREMANS C H, CENENS C and VAN IMPE J F (1998), Àpplication of artificial neural networks as a non-linear modular modeling technique to describe bacterial growth in chilled food products', International Journal of Food Microbiology, 44, 49±68. GEERAERD A H, VALDRAMIDIS V P and VAN IMPE J F (2006), `GInaFiT, a freeware tool to assess the non-log-linear microbial survivor curves', International Journal of Food Microbiology, 102, 95±105. GIBSON A M, BRATCHELL N and ROBERTS T A (1987), `The effect of sodium chloride and temperature on the rate and extent of growth of Clostridium botulinum type A in pasteurized pork slurry', Journal of Applied Bacteriology, 62, 479±490. GIBSON A M, BRATCHELL N and ROBERTS T A (1988), `Predicting microbial growth: growth responses of salmonellae in a laboratory medium as affected by pH, sodium chloride and storage temperature', International Journal of Food Microbiology, 6, 155±178. HERTOG M L A T M, VERLINDEN B E, LAMMERTYN J and NICOLAIÈ B M (2007), ÒptiPa, an essential primer to develop models in the postharvest area', Computational Electronics Agriculture, 57, 99±106. KOUTSOUMANIS K and NYCHAS G J E (2000), Àpplication of a systematic experimental procedure to develop a microbial model for rapid fish shelf life prediction', International Journal of Food Microbiology, 60, 171±184. KOUTSOUMANIS K P, KENDALL P A and SOFOS J N (2004), `Modelling the boundaries of growth of Salmonella typhimurium in broth as a function of temperature, water activity and pH', Journal of Food Protection, 67, 53±59. KOUTSOUMANIS K, MCMEEKIN T and DALGAARD P (2008), Ìntroduction to predictive modelling special issue', International Journal of Food Microbiology, 128, 1. È BNER A, BEIERLE E, CHONSCH L, SCHERER A and PATERSEN B (2010), KREYENSCHMIDT J, HU `Determination of the shelf life of sliced cooked ham based on the growth of lactic acid bacteria in different steps of the chain', Journal of Applied Microbiology, 108, 510±520. LEISTNER L (1995), `Principles and applications of hurdle technology', in Gould G W, New Methods of Food Preservation, London, Blackie Academic, 1±21. LEISTNER L (2000), `Basic aspects of food preservation by hurdle technology', International Journal of Food Microbiology, 55, 181±186. LEISTNER L and GORRIS L G M (1995), `Food preservation by hurdle technology', Trends in Food Science and Technology, 6, 41±46. LE MARC Y, HUCHET V, BOUGEOIS C M, GUYONNET J P, MAFART P and THUAULT D (2002), `Modelling the growth kinetics of Listeria as a function of temperature, pH and RM



455

organic acid concentration', International Journal of Food Microbiology, 73, 219± 237. LE MARC Y, PIN C and BARANYI J (2005), `Methods to determine the growth domain in a multidimensional environmental space', International Journal of Food Microbiology, 100, 3±12. LEPORQ B, MEMBREÂ J-M, DERVIN C, BUCHE P and GUYONNET J P (2005), `The ``Sym'Previus'' software, a tool to support decisions to the foodstuff safety', International Journal of Food Microbiology, 100, 231±237. LOPEZ S, PRIETO M, DIJKSTRA J, DHANOA M S and FRANCE J (2004), `Statistical evaluation of mathematical models for microbial growth', International Journal of Food Microbiology, 96, 289±300. MARKS B P (2008), `Status of microbial modeling in food process models', Comprehensive Reviews in Food Science and Food Safety, 7, 137±143. MARSHALL B J, OHYE D F and CHRISTIAN J H B (1971), `Tolerance of bacteria to high concentrations of NaCl and glycerol in the growth medium', Applied Microbiology, 21, 363±364. MATTICK K L, JORGENSEN F, WANG P, POUND J, VANDEVEN M H, WARD L R, LEGAN J D, LAPPIN-

SCOTT H M and HUMPHREY T J (2001), Èffect of challenge temperature and solute type on heat tolerance of Salmonella serovars at low water activity', Applied and Environmental Microbiology, 67, 4128±4136. MCCLURE P J, BARANYI J, BOOGARD E, KELLY T M and ROBERTS T A (1993), À predictive model for the combined effect of pH, sodium chloride and storage temperature on the growth of Brochothrix thermosphacta', International Journal of Food Microbiology, 19, 161±178. MCCLURE P J, BLACKBURN C D W, COLE M B, CURTIS P S, JONES J E, LEGAN J D, OGDEN I D, PECK

and WALKER S J (1994), `Modelling the growth, survival and death of microorganisms in foods: the UK Food Micromodel approach', International Journal of Food Microbiology, 23, 265±275. MCKELLAR R C and DELAQUIS P J (2002), À probability model describing the interface between survival and death of Escherichia coli O157:H7 in a mayonnaise model system', Food Microbiology, 19, 235±247. MCKELLAR R C and KNIGHT K (2000), À combined discrete-continuous model describing the lag phase of Listeria monocytogenes', International Journal of Food Microbiology, 54, 171±180. MCKELLAR R C and LU X (2004), Modeling Microbial Responses in Food, Boca Raton, FL, CRC Press. MCKELLAR R C, LU X and KNIGHT K P (2002a), `Growth pH does not affect the initial physiological state parameter (p(0)) of Listeria monocytogenes', International Journal of Food Microbiology, 73, 137±144. MCKELLAR R C, LU X and KNIGHT K P (2002b), `Proposal of a novel parameter to describe the influence of pH on the lag phase of Listeria monocytogenes', International Journal of Food Microbiology, 73, 127±135. MCMEEKIN T A (2007), `Predictive microbiology: quantitative science delivering quantifiable benefits to the meat industry and other food industries', Meat Science, 77, 17±27. MCMEEKIN T A, CHANDLER R E, DOE P E, GARLAND C D, OLLEY J, PUTRO S and RATKOWSKY D A (1987), `Model for the combined effect of temperature and water activity on the growth rate of Staphylococcus xylosus', Journal of Applied Bacteriology, 62, 543± 550. M W, ROBERTS T A, SUTHERLAND J P


456

Food and beverage stability and shelf life and RATKOWSKY D Theory and Application, New York, Wiley.

MCMEEKIN T, OLLEY J N, ROSS T

A

(1993), Predictive Microbiology:

and TIENUNGOON S (2000), `Quantifying the hurdle concept by modelling the bacterial growth/no growth interface', International Journal of Food Microbiology, 55, 93±98. MCMEEKIN T A, OLLEY J, RATKOWSKI D A and ROSS T (2002), `Predictive microbiology: towards the interface and beyond', International Journal of Food Microbiology, 73, 395±407. MEJLHOLM O and DALGAARD P (2007), `Modeling and predicting the growth of lactic acid bacteria in lightly preserved seafood and their inhibiting effect on Listeria monocytogenes', Journal of Food Protection, 70, 2485±2497. MEJLHOLM O and DALGAARD P (2009), `Development and validation of an extensive growth and growth boundary model for Listeria monocytogenes in lightly preserved and ready-to-eat shrimp', Journal of Food Protection, 72, 2132±2143. MCMEEKIN T A, PRESSER K, RATKOWSKY D, ROSS T, SALTER M

MEJLHOLM O, GUNVIG A, BORGGAARD C, BLOM-HANSSEN J, MELLEFONT L, ROSS T, LEROI F,

and DALGAARD P (2010), `Predicting growth rates and growth boundary of Listeria monocytogenes ± an international validation study with focus on processed and ready-to-eat meat and seafood', International Journal of Food Microbiology, 141, 137±150. MEMBREÂ J-M, ROSS T and MCMEEKIN T A (1999), `Behaviour of Listeria monocytogenes under combined chilling processes', Letters in Applied Microbiology, 28, 216±220. MEMBREÂ J-M, KUBACZKA M and CHENE C (2001), `Growth rate and growth-no-growth interface of Penicillium brevicompactum as functions of pH and preservative acids', Food Microbiology, 18, 531±538. METRIS A, GEORGE S M and BARANYI J (2006), Ùse of optical density detection times to assess the effect of acetic acid on single-cell kinetics', Applied and Environmental Microbiology, 72, 6674±6679. NEUMEYER K, ROSS T and MCMEEKIN T A (1997), `Development of a predictive model to describe the effects of temperature and water activity on the growth of spoilage pseudomonads', International Journal of Food Microbiology, 38, 45±54. NUIN M, ALFARO B, CRUZ Z, ARGARATE N, GEORGE S, LE MARC Y, OLLEY J and PIN C (2008), `Modelling spoilage of fresh turbot and evaluation of a time-temperature integrator (TTI) label under fluctuating temperature', International Journal of Food Microbiology, 127, 193±199. NYCHAS G J E, SKANDAMIS P N, TASSOU C C and KOUTSOUMANIS K P (2008), `Meat spoilage during distribution', Meat Science, 78, 77±89. PELEG M (2006), Advanced Quantitative Microbiology for Foods and Biosystems: Models for Predicting Growth and Inactivation, Boca Raton, FL, CRC/Taylor & Francis Group. PIN C, DE FERNANDO G D G, ORDONEZ J A and BARANYI J (2002), Ànalysing the lag-growth rate relationship of Yersinia enterocolitica', International Journal of Food Microbiology, 73, 197±201. PIRT S J (1975), Principles of Microbe and Cell Cultivation, Oxford, Blackwell. POSCHET F, BERNAERTS K, GEERAERD A H, SCHEERLINCK N, NICOLAIÈ B M and VAN IMPE J F (2004), `Sensitivity analysis of microbial growth parameter distributions with respect to data quality and quantity by using Monte Carlo analysis', Mathematics and Computers in Simulation, 65, 231±243. PRESSER K A, ROSS T and RATKOWSKY D A (1998), `Modelling the growth limits (growth/no growth interface) of Escherichia coli as a function of temperature, pH, lactic acid ELSE T, VISSER D



457

and water activity', Applied and Environmental Microbiology, 64, 1773±1779. (2004), Èxperimental design and data collection', in McKellar R C and Lu X, Modeling Microbial Responses in Food, Boca Raton, FL, CRC Press, 1±20. RATKOWSKY D A (2004), `Model fitting and uncertainty', in McKellar R C and Lu X, Modeling Microbial Responses in Food, Boca Raton, FL, CRC Press, 151±196. RATKOWSKY D A and ROSS T (1995), `Modelling the bacterial growth/no growth interface', Letters in Applied Microbiology, 20, 29±33. RATKOWSKY D A, LOWRY R K, MCMEEKIN T A, STOKES A N and CHANDLER R E (1983), `Model for bacterial culture growth rate throughout the entire biokinetic temperature range', Journal of Bacteriology, 154, 1222±1226. ROBINSON T P, ABOABA T P, KALOTI A, OCIO M J, BARANYI J and MACKEY B M (2001), `The effect of inoculum size on the lag phase of Listeria monocytogenes', International Journal of Food Microbiology, 70, 163±173. ROSS E W, TAUB I A, DOONA C J, FEEHERRY F E and KUSTIN K (2005), `The mathematical properties of the quasi-chemical model for microorganism growth-death kinetics in foods', International Journal of Food Microbiology, 99, 157±171. ROSS T (1996), Ìndices for performance evaluation of predictive models in food microbiology', Journal of Applied Bacteriology, 81, 501±508. ROSS T (1999), Predictive Food Microbiology Models in the Meat Industry, Sydney, Meat and Livestock Australia. ROSS T and DALGAARD P (2004), `Secondary models', in McKellar R C and Lu X, Modeling Microbial Responses in Food, Boca Raton, FL, CRC Press, 63±150. ROSSO L, LOBRY J R, BAJARD S and FLANDROIS J P (1995), `Convenient model to describe the combined effects of temperature and pH on microbial growth', Applied and Environmental Microbiology, 61, 610±616. SALTER M A, RATKOWSKY D A, ROSS T and MCMEEKIN T A (2000), `Modelling the combined temperature and salt (NaCl) limits for growth of a pathogenic Escherichia coli strain using nonlinear logistic regression', International Journal of Food Microbiology, 61, 159±167. STEWART C M, COLE M B, LEGAN J D, SLADE L, VANDEVEN M H and SCHAFFNER D W (2002), `Staphylococcus aureus growth boundaries: moving towards mechanistic predictive models based on solute-specific effects', Applied and Environmental Microbiology, 68, 1864±1871. TAPIA DE DAZA M S, VILLEGAS Y and MARTINEZ A (1991), `Minimal water activity for growth of Listeria monocytogenes as affected by solute and temperature', International Journal of Food Microbiology, 13, 333±337. TIENUNGOON S, RATKOWSKY D A, MCMEEKIN T A and ROSS T (2000), `Development and evaluation of a model describing the growth limits of Listeria monocytogenes as a function of temperature, pH, NaCl and lactic acid', Applied and Environmental Microbiology, 66, 4979±4987. RASCH M

VALDRAMIDIS V P, BELAUBRE N, ZUNIGA R, FOSTER A M, HAVET M, GEERAERD A H, SWAIN M J,

and KONDJOYAN A (2005), `Development of predictive modelling approaches for surface temperature and associated microbiological inactivation during hot dry air decontamination', International Journal of Food Microbiology, 100, 261±274.

BERNAERTS K, VAN IMPE J F

VALERO A, PEREZ-RODRIGUEZ F, CARRASCO E, FUENTES-ALVENTOSA J M, GARCIA-GIMENO R M

and ZURERA G (2009), `Modelling the growth boundaries of Staphylococcus aureus: effect of temperature, pH and water activity', International Journal of Food Microbiology, 133, 186±194.


458


and ZWIETERING M H (2007), Èxperimental design, data processing and model fitting in predictive microbiology', in Brul S, van Gerwen S and Zwietering M, Modelling Microorganisms in Food, Cambridge, Woodhead Publishing Limited, 22±43. VAN IMPE J F, POSCHET F, GEERAERD A H and VEREECKEN K M (2005), `Towards a novel class of predictive microbial growth models', International Journal of Food Microbiology, 100, 97±105. WALKER S J and JONES J E (1993), `Protocols for data generation for predictive modeling', Journal of Industrial Microbiology, 12, 273±276. WALLS I and SCOTT V N (1997), `Validation of predictive mathematical models describing the growth of Listeria monocytogenes', Journal of Food Protection, 60, 1142± 1145. WHITING R C (1995), `Microbial modeling in foods', Critical Reviews in Food Science and Nutrition, 35, 467±494. WHITING R C and CYGNAROWIC-PROVOST M A (1992), À quantitative model for bacterial growth and decline', Food Microbiology, 9, 269±277. WITJZES T, ROMBOUTS F M, KANT-MUERMANS M L T, VAN'T REIT K and ZWIETERING M H (2001), `Development and validation of a combined temperature, water activity, pH model for bacterial growth rate of Lactobacillus curvatus', International Journal of Food Microbiology, 63, 57±64. WU Y, GRIFFITHS M W and MCKELLAR R C (2000), À comparison of the Bioscreen method and microscopy for the determination of lag times of individual cells of Listeria monocytogenes', Letters in Applied Microbiology, 30, 468±472. ZAIKA L L, PHILLIPS J G, FANELLI J S and SCULLEN O J (1998), `Revised model for aerobic growth of Shigella flexneri to extend the validity of predictions at temperatures between 10 and 19 degrees C', International Journal of Food Microbiology, 41, 9± 19. ZWIETERING M H (2002), `Quantification of microbial quality and safety in minimally processed foods', International Dairy Journal, 12, 263±271. ZWIETERING M H, JONGENBURGER I, ROMBOUTS F M and VAN'T REIT K (1990), `Modeling of the bacterial growth curve', Applied and Environmental Microbiology, 56, 1875± 1881. ZWIETERING M H, DE KOOS J T, HASENACK B E, DE WIT J C and VAN'T REIT K (1991), `Modeling bacterial growth as function of temperature', Applied and Environmental Microbiology, 57, 1094±1101. ZWIETERING M H, DE WIT J C and NOTERMANS S (1996), Àpplication of predictive microbiology to estimate the number of Bacillus cereus in pasteurised milk at the point of consumption', International Journal of Food Microbiology, 30, 55±70. VAN BOEKEL M A J S


14 Modelling chemical and physical deterioration of foods and beverages M. J. Sousa Gallagher, P. V. Mahajan and Z. Yan, University College Cork, Ireland

Abstract: Foods are perishable and there are many factors that can deteriorate the quality and safety of food products during storage and distribution. These can be categorized into chemical and physical factors. To minimize the degradation of food during processing or storage, kinetic models which describe degradation rates and the dependence of intrinsic factors (i.e., critical factors) on extrinsic factors such as temperature and moisture content must be determined. The essential purpose of kinetic models is, first to describe sufficiently a set of experimental data obtained, and second to use the models for prediction, process control, optimization and simulation of food processing, packaging and storage operations. Key words: food degradation, shelf life, mathematical models, quality, temperature.

14.1

Introduction

The shelf life of food is the period during which the food retains an acceptable quality from a safety and organoleptic point of view, and depends on four main factors, namely formulation, processing, packaging and storage conditions. Foods are perishable and there are many factors that can deteriorate the quality and safety of food products during storage and distribution. These can be categorized into chemical and physical factors. To minimize the degradation of food during processing or storage, kinetic models which describe degradation rates and the dependence of intrinsic factors on extrinsic factors, such as temperature and moisture content, must be determined. The essential purpose of


460


kinetic models is to describe sufficiently a set of experimentally obtained data. These kinetic models can then be used for prediction, process control, optimization and simulation of various food processing operations. It allows what-if scenarios and insights into the food systems to be explored and simplifies the process design for maximizing shelf life. Shelf life models are mathematical equations which describe the relationship between the food, the package and the environment. These models are based on different degradation factors (i.e., critical factors) and are essential in predicting the shelf life of food, in designing the packages, and providing useful insights about the foods under abnormal circumstances along the supply chain. Most of the efforts in terms of mathematical modelling have focused on microbiological safety and spoilage (Blackburn, 2000) and this chapter will be devoted to mathematical modelling of chemical and physical deterioration of foods.

14.2 Factors influencing shelf life 14.2.1 Factors responsible for degradation There are different types of product changes that can limit the shelf life of food and these can be classified into four main categories: formulation, processing, packaging and storage conditions. All four categories are critical, but their relative importance depends on the perishability of the food. Generally, a perishable food has less than 14 days of shelf life, but with aseptic technology and controlled atmosphere/modified atmosphere packaging (CA/MAP) such foods may last up to 90 days. A packaging system prevents the deterioration of packaged products to extend shelf life, to maintain quality and to increase the safety of the packaged foods. Some characteristics are considered important in assessing the quality of perishable food products, and knowledge of optimal environmental conditions and adequate packaging materials can be selected to guarantee high-quality product throughout the shelf life. The extent to which packaging can be successfully used to maintain quality and reduce spoilage and extend shelf life depends on: · an assessment of the product properties and identification of the critical quality parameters, i.e., how sensitive it is to changes in temperature, moisture, oxygen (air) and light; · an evaluation of the conditions to which the packed product is likely to be exposed in the supply chain, in order to design and develop packaging with an appropriate barrier and other relevant characteristics. Many factors can influence shelf life and these can be categorized into intrinsic and extrinsic (IFST, 1993). Intrinsic factors are the properties of the final product (e.g., water activity (aw), pH value and total acidity, available oxygen, redox potential (Eh), nutrients, natural microflora and surviving microbiological counts, natural biochemistry of the product formulation (enzymes,


Modelling chemical and physical deterioration of foods and beverages

461

chemical reactants), use of preservatives in product formulation (e.g., salt)) (Kilcast and Subramaniam, 2000). Intrinsic factors are influenced by such variables as raw material type and quality, and product formulation and structure. Extrinsic factors are those factors that the final product encounters as it moves through the food chain (e.g., time±temperature profile during processing; relative humidity (RH), exposure to light and/or environmental microbial counts during processing/storage/distribution, temperature control during storage/distribution, composition of atmosphere within packaging, pressure in the headspace, subsequent heat treatment and consumer handling). The interaction of such intrinsic and extrinsic factors as these either inhibits or stimulates a number of processes which limit shelf life. These can be conveniently classified as: chemical, physical and microbiological. In modern processing, these factors are addressed in the HAACP (Hazard Analysis And Control Point) concept, a comprehensive quality control assurance methodology that aims to ensure both safety and high quality (Frampton, 1989; Stewart et al., 2003; Koutsoumanis et al., 2005). Therefore, the quality of products depends not only on its original state but also on the extent of changes during processing and storage. Environmental conditions like air humidity, air temperature, light, and oxygen affect the quality changes of foods during storage. Therefore, knowledge of optimal environmental conditions and selection of optimum packaging materials can be considered to prevent food degradation and ensure a high-quality product throughout the shelf life. 14.2.2 Identifying critical quality parameters Food is inherently perishable and, depending on its properties (e.g., physical and chemical) and the storage conditions, there will come a point when either its quality will become unacceptable or it will become harmful to the consumer. At this point it has reached the end of its shelf life and the ability to predict this is of great value to the food industry when defining storage and distribution conditions and limits, formulating products, assessing manufacturing processes and carrying out quantitative risk assessment. Therefore, it is important to identify which are the critical factors in order to determine the shelf life of the product. Depending on the nature of the product, various properties or quality indices must be experimentally followed as a function of time in order to evaluate the degradation of the product quality in terms of the physical-chemical properties. In order to fully account for all the degradation criteria, a well-planned experimental design must be adopted. Taking this into consideration, a standard and comprehensive protocol using the accelerated shelf life testing (ASLT) method has been outlined (Labuza and Schimdl, 1985; Corradini and Peleg, 2007). In general, the analysis is performed by following the variation of each quality index in time during storage and then comparing the measured value to a threshold, which is officially set by standards bodies or commonly accepted by a


462


consumer panel as the limit of acceptance. As such the quality degradation of the product is judged by looking independently at the variations of the individual properties of the product during storage, with particular attention to the rapidly varying properties. The use of chemical kinetics, the study of the rates and mechanisms by which one chemical species converts to another and the Arrhenius relationship, which describes the influence of temperature on the reaction rate constants, have been used to model changes in food quality (Singh, 1994).

14.3 Development of mathematical models The essential purpose of mathematical models is to describe sufficiently a set of experimental data. Empirical models are useful in simulating food systems due to their complexity of reactions and non-homogeneous structure. Also from a pragmatic perspective, for practical purposes simple mathematical expressions can be easily used to control the process. The use of mathematical models can help to: · simplify design systems, e.g., product reformulations and process modifications, which are labour intensive, time consuming and expensive; · reduce number of experimental trials required to achieve optimal systems, e.g., packaging; · explore what-if scenarios and insights into the systems, e.g., possible packaging options worthy of testing; · develop a prototype, e.g., package and explore several choices of film types, package size, and product quantity, creating combinations that will result in beneficial atmospheres. 14.3.1 Basic reaction kinetics The loss of food quality for most foods can be represented by a mathematical equation: dC kC n 14:1 dt where C is the quality factor measured, t is time, k is a constant which depends on temperature and water activity, n is a power factor called the order of the reaction, and dC=dt is the rate of change of C with time. A negative sign is used if the deterioration is a loss of C and a positive sign if it is for production of an undesirable end-product. In zero-order reactions the rate is independent of the quality factor. For a degradation reaction: dC k 14:2 ÿ dt



463

which by integration results in: C C0 ÿ kt

14:3

where C0 is quality factor at time zero. In many cases C is not a very quantifiable or measurable value and is based solely on human panel evaluation. In this case Ci (initial value of C) is assumed to be 100% quality and Ce (value of C at end of shelf life) is just unacceptable quality. Thus, the rate of deterioration or the rate constant is given by: k

100% constant % per day time

14:4

Some deterioration is applicable directly to zero-order kinetics, which include enzymatic degradation (fresh fruits and vegetables, some frozen foods, some refrigerated dough), non-enzymatic browning (e.g., dry cereals, dry dairy products, dry pet foods, loss of protein) and lipid oxidation (rancidity development in snacks, dry foods, pet foods, frozen foods). The mathematical expression for a first-order degradation reaction can be described by: ÿ

dC kC dt

14:5

which by integration becomes: C Ci eÿkt

14:6

The types of deterioration that follow n 1 include rancidity (as in salad oil or dry vegetables), microbial growth and death, microbial production of offflavours and slime (meat, fish and poultry), vitamin loss (canned and dried foods) and loss of protein quality (dry foods). Very few data exist describing food degradation by orders other than zero or first. Lee et al. (1977) and Singh et al. (1976) have described the degradation of vitamin C in liquid foods such as tomato juice or canned infant formulas by a second-order reaction. Waloddi Weibull introduced the family of Weibull distributions in 1939. Determination of shelf life by hazard analysis has been extensively used in the mechanical and electrical industries. This approach has been adapted and developed for shelf life determination of foods. The procedure entails fitting failure time to a Weibull distribution and determining the time at which the product is expected to fail at different probability levels. This methodology has been utilized in the determination of the shelf life of foods (Thiemig et al., 1998; Hough et al., 1999; Cardelli and Labuza, 2001; Duyvesteyn et al., 2001). The Weibull model has been widely used in food research, e.g. describing water uptake and soluble solids losses during rehydration of dried apple pieces (Ilincanu et al., 1995), shrinkage of potato during frying (Costa et al., 2001), and enzyme inactivation under high pressure (Lemos et al., 1999).


464


The probability product failure Ft is related to storage time t according to:

Ft 1 ÿ eÿt=

14:7

where is a scale constant and is the shape constant, or behaviour index. The Weibull model corresponds to the first-order degradation kinetics for specific case of 1 (Nelson, 1969). The is related to the kinetic mechanisms and may be expected to be temperature independent, at least within a limited range of temperatures, as found by Ilincanu et al. (1995) and Machado et al. (1999). This parameter gives the model a wide flexibility, making it a potentially good model to describe different reaction kinetics. Among empiric models, the Weibull model has been used to describe the behaviour of systems or events that have some degree of variability, such as quality parameters. The quality parameters can be described as: C ÿ Ce t 14:8 exp ÿ C0 ÿ Ce where C0, C and Ce are the initial, at time t, and at equilibrium quality parameter, respectively, is the Weibull scale parameter, is the shape parameter (dimensionless), and t is the sampling time (Corzo et al., 2008). 14.3.2 Dependence of rate constant on temperature The above basic mathematical analysis of quality loss assumed a constant temperature. The rate constant k is dependent on temperature. Theoretically, k obeys the Arrhenius relationship, which states: k ko eÿEa =RT

14:9

where ko is a pre-exponent constant, Ea is the activation energy in cal/mol, R is the gas constant in cal/mol K and equals 1.986, and T is the temperature in K. Labuza (1982) enumerated potential causes for non-linear Arrhenius plots when the food product is held at high abuse temperature. Furthermore, when reaction or quality loss mechanisms change with temperature, the activation energy may be variable (Karel, 1983). The Arrhenius model does not use a finite reference temperature and the value of the rate constant (ko) at an infinite temperature is often very high. Van Boekel (1996) explained that to reduce correlation between ko and Ea, the temperature can be centred about the mean value of temperatures studied (Tref, reference temperature), and the Arrhenius equation can be reformulated as: ÿ ÿERa T1 ÿT 1 ref 14:10 k kref e where kref is the reparameterized pre-exponential factor: kref ko e

ÿERa T 1

ref


14:11


465

14.3.3 Q10 and accelerated shelf life tests Mathematical models have been used in the food and beverages industry to describe how much faster the product quality will proceed if this product is held at some high abuse temperature. If the temperature-accelerating factor is known, then extrapolation to lower temperatures could be used to predict true product shelf life in those conditions. This is the most often used methodology and yet it is commonly abused in design and interpretation of results, where the objectives are to store the finished product with certain packaging under some test abuse condition, examine the product periodically until the end of shelf life occurs, and then predict the product shelf life under true storage and distribution conditions. In the study of food shelf life, this accelerating factor is sometimes called the Q10 factor and is defined as: Q10

s at T

14:12

s at T10

or Q10

kat T10 kat T

14:13

where T is the temperature (ëC), s is the shelf life at the indicated temperatures (Labuza, 1982), and k is rate constant of reaction. For any temperature difference , different from 10 ëC, this becomes: QsT1 =10 14:14 Q10 QsT2 If it is impossible to establish an Arrhenius plot, then, as shown by Labuza (1982), a simple plot of the log of the time to end of shelf life s (as established by some criteria) vs the temperature in ëC can be used, as long as the extrapolation does not go beyond the 30 ëC range. This plot also helps to establish the Q10. The value of either plot (log k vs Kÿ1) or (log s vs ëC) is that data can be collected at high temperature and used to extrapolate to shelf life at some lower temperature. However, when the product/package system is tested, the package also controls shelf life so that the true shelf life of the food itself is unknown; thus, if a new package with different permeability to oxygen, water, or carbon dioxide is chosen, Eq. 14.14 may not be applicable. If the accelerated shelf life test (ASLT) conditions are chosen properly, however, and the appropriate algorithms for extrapolation are used, then shelf life under any ùnknown' distribution should be predictable. 14.3.4 Evaluation of the goodness of fit of the model to experimental data The criteria commonly followed to evaluate the goodness of fit of the model to the experimental data is the mean relative percentage deviation modulus (E), expressed as:


466

Food and beverage stability and shelf life E

N jmi ÿ mpi j 100 X mi N i1

14:15

where mi is the experimental value, mpi is the predicted value, and N is the number of experimental data. The mean relative percentage deviation modulus (E) is widely adopted throughout the literature, with a modulus value below 10% indicative of a good fit for practical purposes. Boquet et al. (1978) suggested the use of the root mean square of deviations (Srm), expressed by Eq. 14.16 to compare the fitting abilities of the different equations when applied to the same experimental data: v u N u1 X mi ÿ mpi 2 Srm t 14:16 N i1 The relative percentage root mean square (r) has also been used (Bizot, 1983), expressed as: v u N u 1 X mi ÿ mpi 2 100 14:17 rt mi N i1 which combines both concepts described above.

14.4 Predictive mathematical models Kinetic studies dealing with the modelling of degradation of food components and nutrient losses during storage have received increasing attention in recent years. Labuza (1973), Saguy and Karel (1980) and Taoukis et al. (1997) have reviewed modelling of quality deterioration of foods during storage. 14.4.1 Colour Colour is one of the most important appearance attributes of food materials, since it influences the degree of consumer acceptability or it can even be harmful to health. Colour development is the result of various reactions such as non-enzymatic browning reactions and pigment destruction (Cornwell and Wrolstad, 1981; Wong and Stanton, 1992). In non-enzymatic browning reactions, the browning rates are closely related to temperature and relative humidity (Labuza and Saltmarch 1981; Rapusas and Driscoll, 1995; Soponronnarit et al., 1998). Prachayawarakorn et al. (2004) pointed out that the relative humidity at low levels played an important role in retarding browning reactions. Prachayawarakorn et al. (2004) and Kumar and Mishra (2004) found that the lightness and redness of dried garlic slices, and colour change of mango soy fortified yoghurt powder during storage followed a first-order and zero-order kinetic model, respectively. Dependence of the reaction rates on temperature



467

could be described by the Arrhenius equation (Kaymak-Ertekin and Gedik, 2005). Krokida et al. (2001) found that the changes in redness (a) and yellowness (b) were found to follow a first-order kinetic model during drying of apple, banana, potato and carrot. Chen and Ramaswamy (2002) concluded that the changes of Hunter L and E values of ripening banana followed a logistic model, while a and b values were well described by a simple zero-order and fraction conversion models. Yan et al. (2008b) showed that the colour (L and E) degradation of intermediate moisture content (IMC) banana under different relative humidities at 20, 30 and 40 ëC showed a good fit to the zero-order reaction model and a secondary model (Eq. 14.18) was also developed to describe the changes of L values during storage, which could be used to predict the colour of IMC banana under a range of different relative humidities and temperatures during storage: h i L Lo a1 ln RH a2 t e

c1 c2 RHc3 RH2 R

1 1 Tref ÿT

14:18

14.4.2 Moisture content and water activity The moisture and/or water activity of dried or IMC product has a critical influence on its storage stability and moisture migration from the environment (Yan et al., 2008a). Textural quality, chemical and biochemical reactions, and microbial growth rates are greatly affected by moisture content of products. IMC banana is hygroscopic which absorbs or desorbs moisture under different environmental conditions. Therefore, control of moisture adsorption in the distribution chain is critical to control the quality of IMC banana. The effects of temperature (10, 15, 20, 30 and 40 ëC) and relative humidity (76% RH) on the kinetics of the changes of moisture content and water activity of IMC banana during storage were investigated (Yan et al., 2008b). It was found that moisture was absorbed faster when stored at high temperature and at high relative humidity due to higher driving force (awe ÿ aw ) at the beginning stage of storage (Fig. 14.1). However, when moisture content of IMC banana nearly reached equilibrium, the change in moisture content at different relative humidities all levelled off. To understand the physical-chemical relationship between water and the various components in foods, water activity is more revealing than moisture content. The equilibrium relationship between the water activity and moisture content of foods at constant temperature and pressures is shown by the sorption isotherms of foods. Thus, with the knowledge of the moisture content sorption isotherm, it is possible to predict the maximum moisture that the food can be allowed to gain during storage. Therefore, the moisture content of IMC banana was converted into water activity by sorption isotherms (Yan et al., 2008b). Various mathematical models have been proposed in the literature to describe sorption isotherms. Some were developed with a theoretical basis to describe adsorption mechanisms (e.g., GAB), whereas others are just empirical or a simplification of more elaborate models. In some ranges of water activity,


468


Fig. 14.1 Influence of temperature on the moisture uptake by IMC banana under 75% RH (source: Yan et al., 2008b).

sorption isotherms can be approximated to linear equations. Sorption isotherms are usually classified according to their shape in five different types: I, II, III, IV and V (Brunauer et al., 1940). Dried food products usually show isotherms of Type II or III. A summary of the equations that have been reported in the literature to describe sorption isotherms of dried food products is shown in Table 14.1. It should be noted that in some cases these equations predict non-zero moisture content for zero water activity and some of the equations take into account the effect of temperature. The modified Chung±Pfost, modified Henderson, modified Halsey, modified Oswin and GAB equations have been adopted as standard equations by the American Society of Agricultural Engineers for describing sorption isotherms (ASAE, 1995). The equations of BET and GAB provide the monolayer moisture content, which can be considered to be the most useful for determining the optimum moisture conditions for good storage stability, especially for dehydrated foods (Arslan and TogÏrul, 2006). Yan et al. (2008b) found that water activity change of samples stored under different relative humidities and temperatures could be fitted by the lumped capacity model (i.e., first-order kinetics due to the predominance of the external resistance to mass transfer) (Eq. 14.19) in agreement with Costa and Oliveira (1999) and as shown in Fig. 14.2: daw Kawe ÿ aw 14:19 dt where awe and aw are the water activity of sample at equilibrium and at time t, respectively; the rate constant, K (dayÿ1), changed with temperature according to an Arrhenius type equation:


Table 14.1 Mathematical models used to describe sorption isotherms Model

Equation

Constants

Type

Halsey

A 1=B MC ÿ ln aw

A RT=a B

II

Halsey (1948)

A B

II

Chung and Pfost (1967)

Mo Caw 1 ÿ aw aw C ÿ 11 ÿ aw

C ± related to heat of sorption Mo ± monolayer moisture content

II

Brunauer et al. (1938)

A 1 ÿ Baw

A Mo BC B

III

Weisser (1986)

C C 0 eH1 ÿHm =RT K K 0 eH1 ÿHq =RT Mo ± monolayer moisture content

II or III

Van den Berg and Bruin (1981)

Mo ± monolayer moisture content Mo C=TKaw K 1 ÿ Kaw 1 ÿ Kaw C=TKaw C

II or III

Jayas and Mazza (1993)

Chung & Pfost* ß Woodhead Publishing Limited, 2011

BET

MC ÿ MC

BET (modified)

GAB

GAB (modified)

Oswin Henderson

ln ÿA ln aw B

MC

MC

MC

CKMo aw 1 ÿ Kaw 1 ÿ Kaw CKaw

B aw MC A 1 ÿ aw ln 1 ÿ aw 1=B MC ÿ A

Reference

A B

B < 1 II B > 1 III

Oswin (1946)

A aT B

B > 1 II B < 1 III

Henderson (1952)

Table 14.1 Continued Model

Equation

Constants

Type

Smith*

MC A ÿ B ln 1 ÿ aw

A B

II

Smith (1947)

A B

II

Iglesias and Chirife (1978)

A B

II


A B C

II

Pfost et al. (1976)

ÿexpA Bt 1=C ln aw

A B C

II


C aw MC A Bt 1 ÿ aw

A B C

C < 1 II C > 1 III

Oswin (1946)

A B C

C > 1 II C < 1 III

Thompson et al. (1968)

A B

II

Iglesias and Chirife 1*


Iglesias and Chirife 2*

Chung±Pfost* (modified) Halsey (modified)

Oswin (modified)

Henderson (modified)

Freundlich

MC

e2ABaw ÿ M0:5 2eABaw

MC A B

MC ÿln

MC

MC

aw 1 ÿ aw

ÿB t ln aw =C A

ÿln 1 ÿ aw AB t

MC Aaw 1=B

1=C

Reference

Freundlich (1926)

MC ± moisture content (g water/g dry solids); aw ± water activity; T ± temperature (K); t ± temperature (ëC); H1 ± heat of condensation of pure water vapour; Hm ± total heat of sorption of the first layer; Hq ± total heat of sorption of the multilayers; M0.5 ± moisture content at aw 0.5. * The moisture content for zero water activity is different from zero.


471

Fig. 14.2 Variation of water activity with time predicted by sorption isotherm and by the approximation of the lumped capacity model at 10 ëC (a), 15 ëC (b), 20 ëC (c), 30 ëC (d) and 40 ëC (e) (source: Yan et al., 2008b).

aw awe awo ÿ awe expKref

Ea 1 1 t ÿ exp ÿ R T Tref

14:20

where T is temperature (K); t is time (day); Ea is activation energy in cal/mol, Tref and Kref are reference temperature (K) and K at reference temperature, respectively. The effect of temperature on the changes in water activity was also evaluated by using the secondary model (Eq. 14.20), which correlated the lumped capacity model (Eq. 14.19) with the Arrhenius equation. The comparison between water activity predicted by sorption isotherm and the secondary model, the plot of frequency distribution of residuals and normal probability plot of residuals are presented in Fig. 14.3. The secondary model was shown to fit very well (R2 0:982 and E 1:642) all the data under different relative humidities at


472


Fig. 14.3 Diagnosis plot between experimental data and the secondary model predictions of aw; small plots show the frequency distributions of the residuals (top left corner) and distribution of the residuals with the predicted values of water activity (bottom right corner) (source: Yan et al., 2008b).

different temperatures. Moreover, p-levels of all the constants estimated were less than 0.05 and the correlation matrix of the constant estimates (Kref and Ea) of the secondary model was low (ÿ0.275). Ea and Kref were determined as 26.71 kJ molÿ1 and 0.234 dayÿ1, respectively. Johnson and Brennan (2000) found that the changes in moisture content of plantain flour under certain relative humidities and temperatures could be represented by a quadratic polynomial model. Muthukumarappan and Gunasekaran (1996) and Sapru and Labuza (1996) used the finite difference method analysis to solve for moisture migration in corn kernel, cereals and raisins, respectively. 14.4.3 Respiration The processing of fresh-cut produce is different from intact produce, due to which interference occurs in tissue and cell integrity, with the associated increase in enzymatic, respiratory and microbiological activity, and therefore reduced shelf life. The respiration rates (RO2, RCO2) are measured by the difference in O2 and CO2 concentrations at different time intervals using the following equations:


Modelling chemical and physical deterioration of foods and beverages yO2 yiO2 ÿ

RO2 W t ÿ ti Vf

yCO2 yiCO2

RCO2 W t ÿ ti Vf

473 14:21 14:22

where yiO2 , yiCO2 , yO2 and yCO2 are, respectively, the O2 and CO2 concentrations (volumetric fraction) in the gas mix at the initial time ti (h) (or time zero) and at time t (h). RO2 and RCO2 are the respiration rates and W is the weight of the product (kg) and Vf is the free volume inside the package (ml). The effect of temperature is described by an Arrhenius-type equation as shown by: ÿ ÿREac T1 ÿT 1 ref 14:23 RR Rref e where RR is respiration rate (ml/(kg.hr)), Rref is reference respiration rate (ml/ (kg.hr)), Ea is activation energy (kJ/mol), Rc is universal gas constant (0.008314 kJ/(mol.K)), T is temperature (K) and Tref is reference temperature (average temperature 283.15 K). By adjusting Eq. 14.23 in Eqs 14.21 and 14.22, the global mathematical model, as shown in Eqs 14.24 and 14.25, is then used to predict the respiration rate, for example, of mushrooms as a function of temperature and to estimate the Arrhenius equation parameter directly from the raw experimental data, thus minimizing errors in parameter estimates. EO ÿ W ÿ 2 1ÿ 1 14:24 yO2 yiO2 ÿ RO2;ref e Rc T Tref t ÿ ti Vf ECO ÿ W ÿ 2 1ÿ 1 14:25 yCO2 yiCO2 RCO2;ref e Rc T Tref t ÿ ti Vf The dependence of respiration rate on O2 concentration has been widely expressed by a Michaelis±Menten type equation, which is the simplest enzymatic kinetic mechanism. The role of CO2 in respiration was suggested to be mediated by an inhibition mechanism of the Michaelis±Menten equation, as shown in Table 14.2. The maximum rate () of O2 consumption or CO2 production and the dissociation constant () of the enzyme±substrate complex corresponding to the concentration of half the maximal respiration rate may be estimated by linearization of the equation. Competitive inhibition occurs when both the inhibitor (CO2) and the substrate (O2) compete for the same active site of the enzyme. Thus, the maximum respiration rate is lower in high CO2 concentrations. Uncompetitive inhibition occurs when the inhibitor reacts with the substrate±enzyme complex. In this case, the maximum respiration rate is not much influenced at high CO2 concentrations. Non-competitive inhibition occurs when the inhibitor reacts both with the enzyme and with the enzyme±substrate complex (Fonseca et al., 2002). 14.4.4 Water loss or transpiration Water loss or transpiration is an important physiological process that affects the main quality characteristics of fresh produce, such as saleable weight,


474


Table 14.2 Enzyme kinetic equations for different types of inhibitory mechanisms to model the influence of O2 and CO2 on respiration rate (Fonseca et al., 2002), where is the maximum rate of O2 consumption or CO2 production and is the dissociation constant of the enzyme-substrate complex or the concentration corresponding to the half maximal respiration rate Equation R R

R

R

R

Type of inhibition

yO2 yO2

No inhibition

yO2 yCO2 yO2 1 yc yO2 yCO2 yO2 1 yu yO2 yCO2 yO2 1 yn yO 2 yCO2 yCO2 yO2 1 1 yc yu

Competitive

Uncompetitive

Non-competitive

Competitive + Uncompetitive

appearance and texture. Transpiration rate is influenced by factors such as temperature, humidity, surface area, respiration rate and air movement. The relative humidity of the ambient atmosphere has a considerable effect on water loss of fresh mushrooms during storage and the measurement and model development of mass loss rate (transpiration rate) was reported by Mahajan et al. (2008). Low RH was found to increase moisture loss in mushrooms causing shrinkage and quality deterioration, whereas high RH causes moisture to persist on caps supporting microbial growth and causing browning or yellowing of surface. The results stress the importance of maintaining proper in-pack humidity levels as well as storage temperature in order to extend the shelf life of mushrooms. Transpiration rate is expressed in terms of change in mass of mushrooms per unit time per unit surface area of mushroom as shown in Eq. 14.26: 1 dM 14:26 TR ÿ As dt where TR is the transpiration rate in mg/(cm2.h). Integrating Eq. 14.26 with the limits of initial mass Mi to mass M at time t and re-arranging, we have the mass at time t as: 1=1ÿb 14:27 M Mi1ÿb d b ÿ 1 TR t



475

According to Ben-Yehoshua (1987) the flow of water vapour through a fruit surface was proportional to the difference between humidity of fruit internal atmosphere and humidity of the storage atmosphere and behaves according to Fick's law of gas diffusion. In this model the relative humidity of fruit internal atmosphere was considered as a first approximation to be 1.0 (or 100% RH). This parameter depends on solute content of the fruit and is slightly less than 1.0. dV 14:28 ÿKi As awi ÿ aw dt where V is the volume of water given off by the mushrooms; aw is water activity of the container, RH/100; awi is the water activity of mushrooms and Ki is mass transfer coefficient. The moisture content of fresh mushrooms was measured and found to be 92.4%; therefore the system developed for fruits could be applied to mushrooms. It showed that the aw of mushrooms remained constant throughout the storage period despite storing mushrooms at different combinations of temperature and humidity. This gives a constant humidity gradient across the mushroom yielding uniform mass loss during the storage period. Equation 14.27 was expressed in terms of change in mass of mushroom with respect to time as: dM ÿ Ki As awi ÿ aw 14:29 dt where is the density of water. Equation 14.29 was rearranged and combined with Eq. 14.26 yielding Eq. 14.30 where TR is the transpiration rate: dM TR Ki awi ÿ aw 14:30 dt As Equation 14.30 was then integrated with the limits Mi to M for mushroom mass and 0 to t for storage time yielding: 1=1ÿb 14:31 M Mi1ÿb b ÿ 1 Ki d awi ÿ aw t The mass transfer coefficient Ki of this model was estimated by fitting Eq. 14.31 to the experimental data by non-linear regression using Statistica software (Statsoft, Tulsa, USA). The values of mass of mushrooms predicted by Eq. 14.31 were in close agreement with those obtained experimentally (R2 > 99:8). A good agreement was found between observed and predicted transpiration rates as shown in Fig. 14.4. The distribution of residuals was normal (Fig. 14.4), indicating that the trend is not biased. The coefficient Ki, as determined for each set of experimental conditions, was found to increase with temperature. Hence, Eq. 14.31 was modified in order to incorporate the overall effect of temperature on Ki yielding: 1=1ÿb M Mi1ÿb b ÿ 1 Ki d awi ÿ aw 1 ÿ eÿaT t 14:32


476


Fig. 14.4 Relationship between experimental and predicted mass of mushrooms for all the experimental data. The upper left graph shows residual versus predicted values of mass of mushrooms and bottom right graph shows the distribution of residuals obtained from the secondary model (source: Mahajan et al., 2008).

14.5 Future trends It is widely accepted that consumer acceptance of foods is determined mainly by their sensory perception. Lots of researchers have been trying to explain the relationship between sensory characteristics and instrumental measurements. Auerswald et al. (1999), Bozkurt and Bayram (2006), Sousa et al. (2007) and Yan et al. (2008a) found a significant correlation between instrumental and sensory evaluation for blackberries, tomato and Sucuk, and banana, respectively. Repeated measurements are data where individuals have multiple measurements over time or space. Analysing these data requires recognizing and estimating variability both between and within individuals. Further, it is not uncommon for the relationship between an explanatory variable (e.g., time) and a response variable (e.g., colour or texture change) to be non-linear in the parameters. Non-linear mixed effects models provide a tool for analysing repeated measurement data by taking into consideration these two types of variability as well as the non-linear relationship between the explanatory variable and the response variable. Non-linear mixed-effects (NLME) models are hierarchical, involving both fixed effects associated to the population



477

variability and random effects accounting for unexplained inter- and intraindividual variability, respectively. In a mixed-effect model, fixed effects are the parameters associated with an entire population or with certain repeatable experimental factors, and random effects are associated with individual experimental units (Lang, 2007; Mohapatra et al., 2008). Computational fluid dynamics (CFD) is a numerical solution of heat, mass and momentum (fluid flow) transfer equations simultaneously with the given boundary conditions. Use of CFD in food processing/preservation is currently limited to food processing operations. There is a need for application of CFD to food storage/packaging and shelf life studies (Erdogdu, 2009).

14.6

References

and TOGÏRUL H (2006). The fitting of various models to water sorption isotherms of tea stored in a chamber under controlled temperature and humidity. Journal of Stored Products Research, 42, 112±135. ASAE (1995). Moisture relationship of plant-based agricultural products. ASAE Standard D245.5. St. Joseph, MI. È CKNER B, KRUMBEIN A and KUCHENBUCH R (1999). Sensory AUERSWALD H, PETERS P, BRU analysis and instrumental measurements of short-term stored tomatoes (Lycopersicon esculentum Mill.). Postharvest Biology and Technology, 15, 323± 334. BEN-YEHOSHUA S (1987). Transpiration, water stress, and gas exchange. In: Postharvest Physiology of Vegetables, J Weichmann (ed.). Marcel Dekker, New York, pp. 113± 170. BIZOT H (1983). Using the GAB model to construct sorption isotherms. In: Physical Properties of Foods, R Jowitt, F E Escher, B Hallstrom, H F T Meffert, W E L Spiess and G Vos (eds). Applied Science Publishers, London. BLACKBURN W C (2000). Modelling shelf-life, In: The Stability and Shelf-Life of Food, D. Kilcast and P. Subramaniam (eds). Woodhead Publishing Limited, Cambridge, pp. 249±278. BOQUET R, CHIRIFE J and IGLESIAS H A (1978). Equations for fitting water sorption isotherms of foods. II. Evaluations of various two-parameter models. Journal of Food Technology, 14, 319±327. BOZKURT H and BAYRAM M (2006). Colour and textural attributes of Sucuk during ripening. Meat Science, 73, 344±350. BRUNAUER S, EMMETT P H and TELLER E (1938). Adsorption of gases in multimolecular layers. Journal of the American Chemical Society, 60, 309±319. BRUNAUER S, DEMING L S, DEMING W E and TELLER E (1940). On a theory of the van der Waals adsorption of gases. Journal of the American Chemical Society, 62, 1723± 1732. CARDELLI C and LABUZA T P (2001). Application of Weibull hazard analysis to the determination of shelf life of roasted and ground coffee. Lebensmittel-Wissenschaft und -Technologie, 34, 273±278. CHEN, C R and RAMASWAMY H S (2002). Colour and texture change kinetics in ripening bananas. Lebensmittel-Wissenschaft und -Technologie, 35, 415±419. ARSLAN N


478


and PFOST H B (1967). Adsorption and desorption of water vapour by cereal grains and their products. Part II: development of the general isotherm equation. Transactions of the American Society of Agricultural Engineers, 10, 552±555. CORNWELL C J and WROLSTAD R E (1981). Causes of browning in pear juice concentrate during storage. Journal of Food Science, 46, 515±518. CORRADINI M G and PELEG M (2007). Shelf-life estimation from accelerated storage data. Trends Food Sci. Technol., 18, 37±47. Â SQUEZ A (2008). Weibull distribution for modeling CORZO O, BRACHO N, PEREIRA A and VA air drying of coroba slices. Food Science and Technology, 41, 2023±2028. COSTA R M and OLIVEIRA F A R (1999). Modelling the kinetics of water loss during potato frying with a compartmental dynamic model. Journal of Food Engineering, 41, 177±185. COSTA R M, OLIVEIRA F A R and BOUTHCHEVA G (2001). Structural changes and shrinkage of potato during frying. International Journal of Food Science and Technology, 36, 11±23. DUYVESTEYN W S, SHIMONI E and LABUZA T P (2001). Determination of the end of shelf life for milk using Weibull hazard method. Lebensmittel-Wissenschaft undTechnologie, 34, 143±148. ERDOGDU F (2009). Computational fluid dynamics for optimization in food processing. In: Optimisation in Food Engineering, F Erdogdu (ed.). CRC Press, Boca Raton, FL, pp. 219±227. FONSECA S C, OLIVEIRA F A and BRECHT J K (2002). Modelling respiration rate of fresh fruits and vegetables for modified atmosphere packages: a review. Journal of Food Engineering, 52, 99±119. FRAMPTON A (1989). Prevention of rancidity in confectionery and biscuits. The Manufacturing Confectioner, 129±136. FREUNDLICH H (1926). Colloid and Capillary Chemistry, Methuen, London. HALSEY G (1948). Physical adsorption on non-uniform surfaces. Journal of Chemical Physics, 16, 931±937. HENDERSON S M (1952). A basic concept of equilibrium moisture. Agricultural Engineering, 33, 9±32. HOUGH L, PUGLIESO M L, SANCHEZ R and DA SILVA O M (1999). Sensory and microbiological shelf life of commercial ricotta cheese. Journal of Dairy Science, 82, 454±459. IFST (1993). Shelf Life of Foods ± Guidelines for its Determination and Prediction. IFST, London. IGLESIAS H A and CHIRIFE J (1976). Prediction of the effect of temperature on water sorption isotherm of food materials. Journal of Food Technology, 11, 109±116. IGLESIAS H A and CHIRIFE J (1978). Delayed crystallization of amorphous sucrose in humidified freeze-dried model systems. Journal of Food Technology, 13, 137±144. IGLESIAS H A and CHIRIFE J (1981). An equation for fitting uncommon water sorption isotherms in foods. Lebensmittel-Wissenschaft und -Technologie, 14, 111±117. ILINCANU L A, OLIVEIRA F A R, DRUMOND M, MACHADO M F and GEKAS V (1995). Modelling moisture uptake and soluble solids losses during rehydration of dried apple pieces. In: Proceedings of the First Main Meeting of the Copernicus Programme Project `Process Optimisation and Minimal Processing of Foods. Vol. 3: Drying', J C Oliveira (ed.). Escola Superior de Biotechnologia-Universidade Catolica Portuguesa, Porto, Portugal, pp. 64±69. JAYAS D S and MAZZA G (1993). Comparison of five, three-parameter equations for description of dsorption data of oats, Transactions of the ASAE, 36, 119±125. CHUNG D S



479

and BRENNAN J G (2000). Kinetics of moisture absorption by plantain flour. Journal of Food Engineering, 45, 33±36. KAREL M (1983). Quantitative analysis and simulation of food quality losses during processing and storage, In: Computer-aided Techniques in Food Technology, I Saguy (ed.). Marcel Dekker, New York, pp. 117±135. KAYMAK-ERTEKIN F and GEDIK A (2005). Kinetic modelling of quality deterioration in onions during drying and storage. Journal of Food Engineering, 68, 443±453. KILCAST D and SUBRAMANIAM P (2000). Introduction. In: The Stability and Shelf-Life of Food, D Kilcast and P Subramaniam (eds). Woodhead Publishing Limited, Cambridge, pp. 249±278. KOUTSOUMANIS K, TAOUKIS P S and NYCHAS G J E (2005). Development of a safety monitoring and assurance system for chilled food products, International Journal of Food Microbiology, 100, 253±260. KROKIDA M K, MAROULIS Z B and SARAVACOS G D (2001). The effect of the method of drying on the colour of dehydrated products. International Journal of Food Science and Technology, 36, 53±59. KUMAR P and MISHRA H N (2004). Storage stability of mango soy fortified yogurt powder in two different packaging materials, HDPP and ALP. Journal of Food Engineering, 65, 569±576. LABUZA T P (1973). Effects of dehydration and storage. Food Technology, 27, 20±26, 51. LABUZA T P (1982). Open Shelf Life Dating of Foods. Food and Nutrition Press Inc., Westport, CT. LABUZA T P and SALTMARCH M (1981). The nonenzymatic browning reaction as affected by water in foods. In: Water Activity: Influences on Food Quality, L B Rockland and G F Stewart (eds). Academic Press, New York, pp. 605±650. LABUZA T P and SCHMIDL M K (1985). Accelerated shelf life testing of foods. Food Technol. 39, 57±64. LANG W (2007). A computationally efficient method for nonlinear mixed-effects models with nonignorable missing data in time-varying covariates. Computational Statistics & Data Analysis, 51 (5), 2410±2419. LEE Y C, KIRK J R, BEDFORD C L and HELDMAN D R (1977). Kinetics and computer simulation of ascorbic acid stability of tomato juice as functions of temperature, pH and metal catalyst. Journal of Food Science, 42, 640±644, 648. LEMOS M A, OLIVEIRA J C, VAN LOEY A and HENDRICKX M E (1999). Influence of pH and high pressure on thermal inactivation kinetics of horseradish peroxidase. Food Biotechnology, 13, 13±32. MACHADO M F, OLIVEIRA F A R, GEKAS V and SINGH R P (1999). Kinetics of moisture uptake and soluble-solids loss by puffed breakfast cereals immersed in water. International Journal of Food Science and Technology, 33, 225±237. MAHAJAN P V, OLIVEIRA F A R and MACEDO I (2008). Effect of temperature and humidity on the transpiration rate of the whole mushrooms. Journal of Food Engineering, 84 (2), 281±288. MOHAPATRA D, FRIAS J M, OLIVEIRA F A R, BIRA Z M and KERRY J (2008). Development and validation of a model to predict enzymatic activity during storage of cultivated mushrooms (Agaricus bisporus spp.). Journal of Food Engineering, 86 (1), 39±48. MUTHUKUMARAPPAN K and GUNASEKARAN S (1996). Finite element simulation of corn moisture adsorption. Transactions of the ASAE, 39, 2217±2222. NELSON W (1969). Hazard plotting for incomplete failure data. Journal of Quality Technology, 1, 27±30. JOHNSON P N T


480


(1946). The kinetics of package life. III. Isotherm. Journal of the Society of Chemical Industry, 65, 419±421. PFOST H B, MAURER S G, CHUNG D S and MILLIKEN G A (1976). Summarizing and reporting equilibrium moisture data for grains. Transactions of the American Society of Agricultural Engineers, 76, 3520±3532. OSWIN C R

PRACHAYAWARAKORN S, SAWANGDUANPEN S, SAYNAMPHEUNG S, POOLPATARACHEWIN T,

SOPONRONNARIT S and NATHAKARAKULE A (2004). Kinetics of colour change during storage of dried garlic slices as affected by relative humidity and temperature. Journal of Food Engineering, 62, 1±7. RAPUSAS R S and DRISCOLL R H (1995). Kinetics of non-enzymatic browning in onion slices during isothermal heating. Journal of Food Engineering, 24, 417±429. SAGUY I and KAREL M (1980). Modelling of quality deterioration during food processing and storage. Food Technology, 37, 78±85. SAPRU V and LABUZA T P (1996). Moisture transfer simulation in packaged cereal-fruit systems. Journal of Food Engineering, 27, 45±61. SINGH R P (1994). Scientific principles of shelf life evaluation, In: Shelf Life Evaluation of Foods, C M D Man and J A Jones (eds). Blackie Academic and Professional, London, pp. 3±36. SINGH R P, KIRK J and HELDMAN D R (1976). Kinetics of quality degradation: ascorbic acid oxidation in infant formula. Journal of Food Science, 41, 304±308. SMITH S E (1947). The sorption of water vapour by high polymers. Journal of the American Chemical Society, 69, 646±651. SOPONRONNARIT S, SRISUBATI N and YOOVIDHYA T (1998). Effect of temperature and relative humidity on yellowing rate of paddy. Journal of Stored Products Research, 34, 323±330. Â NDEZ C (2007). Effect of processing on the SOUSA M B, CANET W, ALVAREZ M D and FERNA texture and sensory attributes of raspberry (cv. Heritage) and blackberry (cv. Thornfree). Journal of Food Engineering, 78, 9±21. STEWART C M, COLE M B and SCHAFFNER D W (2003). Managing the risk of staphylococcal food poisoning from cream-filled baked goods to meet a food safety objective. Journal of Food Protection, 66, 1310±1325. TAOUKIS P S, LABUZA T P and SAGUY I S (1997). Kinetics of food deterioration and shelf life prediction. In: Handbook of Food Engineering Practice, K J Valentas, E Rotstein and R P Singh (eds). CRC Press, Boca Raton, FL, pp. 361±403. THIEMIG F, BUHR H and WOLF G (1998). Characterization of shelf life and spoilage of fresh foods. First results with the Weibull hazard analysis. Fleischwirtschaft, 78, 152± 155. THOMPSON T L, PEART R M and FOSTER G H (1968). Mathematical simulation of corn drying: a new model. Transactions of the American Society of Agricultural Engineers, 11, 582±586. VAN BOEKEL M A J S (1996). Statistical aspects of kinetic modelling for food science problems. Journal of Food Science, 61, 477±485. VAN DEN BERG C and BRUIN S (1981). Water activity and its estimation in food systems: theoretical aspects. In: Water Activity: Influences on Food Quality, L B Rockland and G F Stewart (eds). Academic Press, New York, pp. 1±61. WEISSER H (1986). Influence of temperature on sorption isotherms. In: Food Engineering and Process Applications, vol. 1, M Le Maguer and P Jelen (eds). Elsevier, New York, pp. 189±200. WONG M and STANTON D W (1992). Effect of removal of amino acids and phenolic



481

compounds on non-enzymatic browning in stored kiwi fruit juice concentrates. Lebensmittel-Wissenschaft und -Technologie, 26, 138±144. YAN Z, SOUSA-GALLAGHER M J and OLIVEIRA F A R (2008a). Identification of critical quality parameters and optimal environment conditions of dried banana during storage. Journal of Food Engineering, 85, 168±172. YAN Z, SOUSA-GALLAGHER M J and OLIVEIRA F A R (2008b). Mathematical modelling of the kinetic of quality deterioration of intermediate moisture content banana during storage Journal of Food Engineering, 84, 359±367.


15 Accelerated shelf life testing of foods S. Mizrahi, Technion-Israel Institute of Technology, Israel

Abstract: The food industry has a great need to obtain, in a relatively short time, the necessary information for determining the shelf life of its products. The chapter reviews the approaches available for running accelerated shelf life tests. The scientific basis behind each of these approaches is discussed as well as the problems and challenges that are involved. The chapter reviews methods that are based on the traditional linear kinetic models as well as on the recently emerging ones that are derived from non-linear ones. Key words: accelerated shelf life tests, non-linear kinetic models, kinetic model approach, dynamic shelf life testing. Note: This chapter is a revised and updated version of Chapter 5 Àccelerated shelf life tests' by S. Mizrahi in The Stability and Shelf-life of Food, ed. D. Kilcast and P. Subramaniam, Woodhead Publishing Limited, 2000, ISBN: 978-1-85573-500-2.

15.1 Introduction The food industry has a great need to obtain, in a relatively short time, the necessary information for determining the shelf life of its products. It has a very important impact on handling of the products' storage, distribution and shelf life dating.1 Moreover, it provides an essential tool to probe the possibilities of extending shelf life through proper product formulation and processing techniques. For practical reasons, especially when the actual storage time is long, the industry resorts to accelerated test techniques that considerably shorten the process of obtaining the necessary experimental data. In the context of this chapter, therefore, accelerated shelf life testing (ASLT) will refer to any method that is capable of evaluating product stability, based on data that are


Accelerated shelf life testing of foods

483

obtained in a significantly shorter period than the actual shelf life of the product. This chapter will discuss first the scientific basis of accelerated shelf life testing. It will indicate what tools are available for carrying out the tests and explain the problems encountered when using them. At the end, an attempt is made to suggest where this important area of accelerated shelf life testing is heading and what expectations one should have with regard to developing novel practical and reliable tools that the industry will find convenient to use.

15.2

Basic principles

ASLT is applicable to any deterioration process that can be quantitatively expressed by a valid model. This model may follow the changes in a shelf life expressing the value of a marker of deterioration or the extent of product failure under a given storage and handling history. The deterioration processes may be chemical, physical, biochemical or microbial. The principles of the ASLT will be the same in all cases. However, a larger range of different ASLT approaches are available for chemical deterioration of foods and therefore the examples in this chapter will be based on them. The following sections will discuss these approaches, all of which have a common goal of getting reliable deterioration data in a short period and selecting the proper model to use for predicting the shelf life of the product. In addition, cases of ASLT that have no need for a model will also be discussed.

15.3

Initial rate approach

Conceptually, one of the simplest techniques for obtaining useful data in a relatively short time for predicting shelf life is the ìnitial rate approach'.2 It may be applicable to cases where there is a shelf life determining deterioration marker that can be accurately monitored at levels well below failure. Such an approach may require an extremely accurate and sensitive analytical method. This method should be capable of measuring minute changes in the extent of deterioration after a relatively short storage time at actual conditions. The data of the initial rate of the deterioration process can serve to evaluate the parameters of a valid quantitative model. If a conventional kinetic model is used, the most important parameters to evaluate are the order of reaction (n) and the kinetic constant (K). In the case of monitoring the changes in the concentration (C) of a deterioration marker, the kinetic equation may be expressed as: dC 15:1 KC n dt where t is time. For sake of simplicity, let us define an index of deterioration (D) that has the form:


484

Food and beverage stability and shelf life dD

dC K dt Cn

15:2

By doing that, the index of deterioration will be always linear with time and will have the following form: D ÿ D0 Kt

15:3

where D0 is the initial level of the index of deterioration. Equation 15.3, if valid, is the only kinetic model that is required to employ this approach to ASLT and the extrapolation process, after evaluating the value of K from the initial rate, is obviously very simple. The product shelf life (ts ) is therefore: D ÿ D0 15:4 ts K Fortunately, information about the order of reactions in many food systems is available in the literature. Many of the chemical deterioration reactions in foods follow either a zero- or a first-order kinetics. The value of the index of deterioration will be in these cases: · Zero order (n 0) D C · First order (n 1) D ln C

15.5 15.6

On a time scale it is translated to a linear or semi-logarithmic relationship, respectively (Fig. 15.1). When the order of reaction is unknown, a simple accelerated test procedure may be used to evaluate it empirically. In that case the simplest version of the kinetic model approach, which is discussed in the following sections, may be used. Such a method uses any convenient kinetically active factor to accelerate the deterioration process. The initial rate method, when applicable, can provide practically an ideal accelerated shelf life testing technique. It has the advantage of obtaining, in a relatively short time, the kinetic data at the actual storage conditions.

Fig. 15.1 Extent of deterioration as a function of time for zero- and first-order kinetics.



485

An example of using a relatively sensitive analytical method was attempted by Teixeira Neto et al. to determine the rate of oxygen uptake during oxidation of dehydrated foods.3 The commonly used manometric techniques are notorious for being insensitive to minute changes in the relatively large mass of oxygen in the headspace.4 To overcome this problem, Teixera Neto et al.3 determined the rate of oxygen uptake by analyzing the changes in the mass of the oxygen, which was adsorbed or entrapped in the product.5 Since that mass is relatively much smaller than that of the headspace, the data of the rate of oxygen uptake by the product were obtained in only a few days. The discussion of the initial rate approach may serve also as an appropriate reminder as to why there is a need to have other accelerated shelf life testing methods for food systems having a shelf life determining marker. In the absence of a very sensitive and accurate analytical technique, the deterioration process should be allowed to progress for a longer time to enable the available method to detect the changes in a statistically significant way. The minimal time required to obtain significant data is therefore dependent on the accuracy and sensitivity of the analytical method; the worse they are the longer the time needed to obtain the data. In a way, accelerated shelf life testing is required to overcome the shortcomings of the analytical methods that are used by the industry. Therefore, the selection of the proper analytical techniques for monitoring the deterioration process is of great importance to shorten the period of the ASLT.

15.4

Kinetic model approach

The kinetic model approach is the most common method for accelerated shelf life testing. The basic process involves the following steps: · Selection of the desired kinetically active factors for acceleration of the deterioration process. · Running a kinetic study of the deterioration process at such levels of the accelerating factors that the rate of deterioration is fast enough. · By evaluating the parameters of the kinetic model, extrapolating the data to normal storage conditions (Fig. 15.2). · Use the extrapolated data or the kinetic model to predict shelf life at actual storage conditions. The absolute requirement for using this procedure is to have a valid kinetic model for the deterioration process. The general and most comprehensive kinetic model for chemical reactions in foods includes all the factors that may affect their rate. These factors may be divided into two main groups, namely compositional (CFi) and environmental factors (EFj).6 The model may be generally expressed as follows: dD 15:7 KCFi ; EFj dt


486


Fig. 15.2

Schematic diagram of data extrapolation in accelerated shelf life testing.

This equation indicates that the kinetic constant K is a function of these factors. In practice, however, one does not need a comprehensive kinetic model. For prediction of shelf life at actual storage conditions, the model should include only those factors that change during storage (SFi). Therefore, the required model should only be as follows: dD 15:8 KSFi dt The list of SFi should include factors such as temperature, moisture content, light intensity, composition and others, but only if they change during storage. Obviously, when one is interested in predicting the shelf life of a product at a constant temperature, it is of no interest to have a kinetic model that includes this factor. Yet, temperature can be used very effectively to accelerate the rate of the deterioration process. Therefore, the demands from a kinetic model for ASLT may be different from one that is used only to predict shelf life. The model for accelerated shelf life testing should contain two groups of factors. The first comprises those that are changing during storage (SFi), as is in Eq. 15.8, and the second those that are used to accelerate the rate of reaction (AFj). The kinetic model for ASLT therefore has the form: dD KSFi ; AFj 15:9 dt The kinetic model for accelerated shelf life testing may therefore be different



487

from the one usually used to predict product stability at normal storage conditions. Obviously, any of the factors that are changing during storage may be used to accelerate the rate of reaction. Equation 15.9 expresses a concept of great practical importance for ASLT. It indicates that it is possible to use any desired factor to accelerate the process of deterioration regardless of whether it is active during normal storage conditions. Weissman et al.7 have suggested that one might even use compositional factors to accelerate the rate of deterioration. Corradini and Peleg8 have suggested reducing the salt concentration in order to accelerate bacterial growth. This implies that the composition of a product may be altered just for the benefit of accelerating the deterioration rate. Clearly, the information obtained is useful only if a valid kinetic model is available for these compositional factors. Such a concept can open a large number of creative avenues for conducting accelerated shelf life testing.

15.5

Single accelerating factor

The simplest and most commonly used method of ASLT is based on employing only a single factor, mostly temperature, to expedite the deterioration process. The simplicity of such a method is related to both the experimental procedure and the availability of valid models. It should be emphasized that in ASLT, the validity of the kinetic model is crucial to obtaining accurate prediction of shelf life. Unfortunately, the validity of the model cannot be fully verified by the ASLT procedure, because the levels used for the accelerating factor do not include those of actual storage conditions. This is in contrast to the situation where the quantitative model is established and verified for actual storage conditions. Therefore, it is of great advantage if the selection of a model for ASLT is based on prior knowledge of its validity. The following sections will discuss the pros and cons of using temperature as a single accelerating factor and of the available models. 15.5.1 Arrhenius model The Arrhenius model that relates the rate of a chemical reaction to the changes in temperature is the best example of what is believed to be a generally accepted model with experimentally proven validity. It is a linear model expressing the effect of temperature on the rate constant (K) of different reactions in many food systems, represented by: Ea 15:10 K K0 exp ÿ RT where K0 is a constant, Ea the energy of activation, R the gas constant and T the absolute temperature. The Arrhenius model requires the evaluation of two parameters only, K0 and Ea, that are supposed to be temperature independent.


488


Therefore, it is very convenient that these parameters can be accurately evaluated by accelerated tests at high temperatures. Moreover, since this model has been applied to many cases, a large database is available, mainly of the energy of activation of different reactions. One may conveniently use this information to get a reasonable estimate of the extent a change in temperature may affect the rate of reaction. To simplify the process further, one may avoid the need to evaluate K0 by using a ratio between the rates of reaction when the temperature is changed by any arbitrary value. The most commonly used value is 10 ëC and therefore the ratio between the rate of reactions is known as Q10. The value of Q10 may be calculated by using Eq. 15.8 to express the rate of reaction first for a temperature of (T 10) and then for T and divide the two, namely: Ea dD2 exp 10Ea RT 10 dt 15:11 exp Q10 dD1 Ea RTT 10 exp dt RT The simplicity of using Q10 has made it a very popular method for estimating shelf life. If prior knowledge or estimates of the value of the energy of activation are relied on, the accelerated tests may be run only at one elevated temperature. When choosing the maximal possible temperature, for which the Arrhenius model is still valid, the data are obtained in the shortest possible time by minimal experimental efforts. To improve the accuracy of this version of tests further, the energy of activation may also be evaluated. In that case, the rate of reaction must be obtained at a number of different temperatures below the maximal one in order to be within the range where the model is valid. Obviously, such a procedure takes a much longer time to run. The rule in accelerated stability tests is that to get more accurate data requires a longer experimental time. The use of the Arrhenius model, as will be discussed later, is questionable when changes in the mechanism of reaction take place due to phase transition and competitive reactions. However, even if it is valid, its use, or rather any approach that is based on a single accelerating factor, may be problematic with regard to the accuracy of the extrapolated data. To demonstrate that problem, let us consider first a simple case where the kinetic constant of the reaction is linearly related to the accelerating factor (Fig. 15.3). In this figure, the solid line represents the true relationship between the kinetic constant (y) and the accelerating factor (x). The point at the top end of the line represents the true kinetic constant (Ye) at the level (Xe), which may be estimated from the experimental data. To extrapolate the data, the slope (a) of the line must be evaluated by curve fitting of the accelerated test's kinetic data. That value of the slope is used to extrapolate the line to actual storage conditions (Xs) where the true rate of reaction is supposed to be (Ys). However, the error in the slope (a) may cause the extrapolated line to produce a predicted kinetic constant (Yp



489

Fig. 15.3 Analysis of extrapolation error in linear plot.

(high) or Yp (low)) which deviates from that true value (Ys) by Y (Fig. 15.3). For the line that has a slope of a ÿ a, which is symmetrical to the one with a slope of a ÿ a, the following expression should hold: Ye ÿ Yp high Ye ÿ Ys Y a ÿ a Xe ÿ X s X e ÿ Xs For the true line: Ye ÿ Ys a Xe ÿ X s

15:12

15:13

Subtracting Eq. 15.13 from Eq. 15.12 one obtains: Y a X e ÿ Xs

15:14

To find how the error in evaluating slope (a) affects the accuracy of the extrapolated value, one should divide Eq. 15.14 by Eq. 15.13, resulting in the following expression: a Y Y a Ye ÿ Ys Ys YYe ÿ 1 s Therefore, the error in the extrapolated value is: Y a Ye ÿ1 Ys a Ys

15:15

15:16

Let us define the acceleration ratio (AR) as the rate of the accelerated reaction in reference to that at normal storage conditions. In the case of the linear


490


relationship between the kinetic constant and the accelerating factor, the value of that acceleration ratio is expressed as: Ye 15:17 AR Ys Therefore, the relative error of the predicted value of the kinetic constant is: Y a AR ÿ 1 Ys a

15:18

The extrapolation process multiplies the experimental error of evaluating the slope of the line by the acceleration ratio minus one. The error of the predicted kinetic constant may be extremely high, especially when a very high acceleration ratio is used and if special care is not taken to reduce the experimental error to a very low value (Fig. 15.4). The magnitude of the error changes when the relationship between the kinetic constant and the accelerating factor is no longer linear. For example, when that relationship is exponential (Arrhenius model) or a power law, the extrapolation error may be different and it can be estimated by turning these models into their linear form and then using the above equations. The only necessary step is to assign the y-axis the value of lnK. In such a case, Eq. 15.18 will read: ln K ln Kp ÿ ln Ks ln Kp =Ks a ln Ke ÿ1 15:19 ln Ks ln Ks ln Ks a ln Ks Therefore: ln

Kp a a Ke a ln Ke ÿ ln Ks ln ln AR Ks Ks a a a

Fig. 15.4 Error in linear extrapolation.


15:20


491

Fig. 15.5 Error in exponential or power law extrapolation.

That results in: Kp ARa=a Ks

15:21

The error in the extrapolated data is: Kp ÿ Ks Ks ARa=a ÿ 1 ARa=a ÿ 1 Ks Ks

15:22

It appears, therefore, that using a model like the Arrhenius equation involves a lower error in extrapolating data (Fig. 15.5) than in the case of a simple linear model (Fig. 15.4). 15.5.2 Non-linear kinetic models The popularity of using the Arrhenius model, especially when the order of reaction is known, has made it synonymous with ASLT and therefore temperature has been the major accelerating factor. The practical aspects and data interpretation of such tests have been recently reveiwed by Saguy and Peleg.9 Most of the reported ASLTs are based on this model.10±15 However, as indicated above, this model may not be valid especially when changes in the mechanism of reaction take place due to phase transition, competitive reactions, glass transition, chemical changes in the food, etc.16 Therefore, large experimental efforts and lengthy procedures are required mainly to validate the linear Arrhenius model. In such cases a recently proposed approach of using non-linear kinetic versatile models may become attractive. This approach, developed mainly at the University of Massachusetts by Peleg and colleagues, is based on


492


empirical models that are relatively simple to handle and are remarkably capable of quantitatively expressing the progress of a wide range of deterioration processes including very complex ones.8,17,18 Moreover, their usage does not require any assumption or knowledge about the order of reaction or the nature of the mechanisms that control food spoilage. There are a number of empirical models of that kind that fit the data of the deterioration processes and thus may be considered as a possible basis for ASLT. The one that got most of the attention is the Weibullian distribution that in our context has the following form in its decay mode: C expÿZt 15:23 D C0 where Z and are constants known as the `rate' and the shape parameters, respectively. The kinetics of first-order decay is a special case of Eq. 15.23 where the shape parameter 1. In this case the rate parameter Z K. For exponential growth, the sign of the exponent is reversed. Equation 15.23 has two temperature dependent coefficients, which must be determined experimentally instead of one as in the Arrhenius model. Fortunately, in many cases, the shape parameter () is much less sensitive to temperature changes and therefore can be considered constant. The determination of the temperature dependence of Z, and wherever needed of as well, may not require any additional ASLT experiments but only different mathematical treatment. This issue was discussed by Corradini and Peleg,8 who suggested a number of simple versatile secondary models for this purpose. One model, for example, which they used to fit the rate parameter (Z) vs temperature data is the log logistic equation: Z loge f1 expBT ÿ Tc g

15:24

where B and Tc are constants. Sometimes, evaluating these two constants by regression is all that is required when the shape parameter () can be considered constant. However, if the shape parameter is temperature sensitive, its temperature dependence too ought to be expressed by an empirical secondary model, a power law type, for example. In view of the available information for ASLT tests based on temperature, the non-linear kinetic approach has the potential of becoming generally accepted for many deterioration processes. Its use may become more common, especially when the deviations from constant order kinetics are large and the Arrhenius model's validity might come into question. In such a case, validation of the model also requires carrying out experiments at temperatures close to actual storage conditions, thus requiring a lengthy procedure. So far, the single accelerating factor that was discussed was temperature. As already indicated above, any factor that affects the deterioration process may be used in ASLT but only if it has a valid model for data extrapolation. The nonlinear kinetic models apply also for other accelerating factors but need more work to establish their validity.



15.6

493

Glass transition models

One of the most interesting approaches to kinetic studies and their use for ASLT is based on glass transition models, which were borrowed from polymer science. Clearly, this approach may be applicable only to products that are in the physical state for which such models are valid. These models, such as the Williams, Landel and Ferry (WLF) model, relate changes in the system properties, which are connected with the polymer molecular mobility, to the temperature within the range of the transition of the product from its glassy to rubbery state.19 Based on the assumption that the rate of the deterioration reactions should relate to molecular mobility in much the same way, this approach yielded valuable information about processes of recrystallization, and losses of flavor and desired textural attributes caused by such structural changes.20 When applicable, glass transition models offer a number of very attractive features with regard to kinetic studies and ASLT. The first one is the fact that it combines both the effects of the temperature and the moisture content into one relatively simple equation.21 The second one, which is even more interesting, is that the rate of the deterioration is related only to the physical state of the system, which can be independently determined in a very short time by readily available physical techniques. That considerably simplifies the experimental work since one needs only the kinetic data, at one high level of temperature or moisture content, and the physical characterization of the system. Unfortunately, that kind of interesting approach to ASLT has, so far, found very limited use. In general, the glass transition model was found to correspond closely to a stability limit with respect to physical processes, such as the ones mentioned above.22 On the other hand, the glass transition model proved inadequate to account for different deterioration kinetics.20,23±27 In general, the glass transition model failed to account for diffusion of some small molecules, especially water. However, it has been proposed that the glass transition model may be applicable to predict changes in the rate of chemical reactions in food deterioration but only if proven to be diffusion limited.

15.7

Multiple accelerating factors

The use of multiple accelerating factors presents an effective approach to obtain a high acceleration ratio of the deterioration reaction at a minimal cost of prediction error. To demonstrate this fact, let us consider a simple theoretical case of a kinetic model that has the following form: K c1 F1 c2 F2 c1 c2 F1 F2

15:25

where c1 and c2 are the estimated parameters of the accelerating factors F1 and F2 , respectively. In order to evaluate the error in the kinetic constant due to that of the estimated parameters, Eq. 15.25 is differentiated with regard to these parameters, resulting in:


494

Food and beverage stability and shelf life dK c2 F1 F2 dc1 c1 F1 F2 dc2

15:26

When dividing Eq. 15.26 by Eq. 15.25 and combining it with Eq. 15.18, the estimated error is found from the following expression: K c1 c2 AR1 ÿ 1RE1 AR2 ÿ 1RE2 c1 c2 K

15:27

where RE1 and RE2 are the experimental relative errors for the factors F1 and F2 , respectively. By using multiple factors, a 100-fold acceleration of the deterioration reaction, e.g. a single one may be replaced by two factors each having an acceleration ratio of only 10. This one order of magnitude reduction in the acceleration ratio decreases considerably the extrapolation error. If, for example, the error in estimating the model parameter for each of these factors is only 1%, the extrapolated data might deviate from the real value by 99% (Eq. 15.18) for a single as compared to 18% (Eq. 15.27) for two accelerating factors. While the total acceleration effect of using two or more factors is a multiplication of their effect, the error is only the summation. Moreover, the required relatively low acceleration ratio is achieved by a much smaller change in the level of the kinetic factors and thus the system stays much closer to the actual storage conditions. Furthermore, when a narrower range of the accelerating factor is used, not only is the validity of the kinetic model better maintained but also the kinetic model may have a simpler form. The advantages of the multiple factors approach are obtained at a cost of running a more complicated experimental procedure. That is the result of the need to evaluate not only the effect of each factor on the reaction kinetics but also a possible interaction between them. The procedure, therefore, lacks the simplicity that makes such a technique more practical for the food industry. A multiple factor acceleration of the deterioration reaction was carried out by Mizrahi et al. by combining the effect of temperature and moisture content (m).10 It enabled a shelf life that lasts for over one year to be predicted based on an experimental study that required only 10 days. The basic kinetic equation had the following general form: Km; T f mTr exp Ea =R1=Tr ÿ 1=T

15:28

where Tr is a reference temperature. Since moisture content in a food product is related to the water activity (aw) by the sorption isotherm, the kinetic function at the constant reference temperature ( f mTr ) could be expressed also in terms of that water activity. One form of such a function for non-enzymatic browning of cabbage is:10 K K0 aw s

15:29

The kinetic model shown in Eq. 15.28 indicates that the evaluation of the kinetic effect of moisture content is performed for a constant reference temperature (Tr). Theoretically, therefore, the evaluation of the kinetic model may be as simple as first running an experiment at an elevated constant temperature and changing only



495

the moisture content and then keeping the latter constant at any desired level and varying the temperature. In many cases, especially when the range of temperature and moisture content changes are kept within a relatively narrow range, that procedure may be adequate. However, when that range is relatively large, a possible interaction between the two factors might play an important role in determining the accuracy of the shelf life prediction. Such was the case in the study of the non-enzymatic browning of cabbage where the energy of activation happened to be affected by the moisture content.10 The empirical expression that was used to describe the effect of the moisture content on the energy of activation was: Ea c1 exp ÿc2 m

15:30

where c1 and c2 are constants. That interaction between the factors greatly complicates the experimental procedure since the effect of the moisture content on the energy of activation should be tested by changing both factors at the same time. That requires a much longer time and more experimental work, which may make this method very unattractive for practical use. However, as stated before, when a narrower range of accelerating factors is used, that elaborate and cumbersome procedure may not be necessary. Another example of using two factors for accelerating the deterioration process could have been based on the data of Dattatreya et al.28 that quantitatively evaluated how pH and temperature affect the rate of browning of sweet whey powder (SWP). Their data enabled an ASLT procedure to be devised where very high acceleration ratio could be obtained by acidifying this product and exposing it to elevetaed temperatures. In this case too, the Arrhenius model may be used for expressing the effect of temperature on the rate of reaction. Other empirical equations are used for the effect of pH. In a similar way to the former case where temperature and moisture content were combined to further accelerate the deterioration process, in this case too the Arrhenius model must be modified in order to account for the effect of the pH on the value of the energy of activation. But again, this multiple-factors ASLT method facilitates changing them only within a relatively narrow range in which the Arrhenius model is practically valid. The above discussion indicates that when using multiple accelerating factors, their interaction is complicating the deterioration models. In fact, one has to use mostly empirical models that are derived by curve fitting of data that are relevant to a specific case. Moreover, in order to establish their validity, the experimental conditions should include actual storage conditions. This requires a long experimental time that makes it impractical as an ASLT method. The same problem exists when using non-linear kinetic models.8 In this approach, too, the parameters of the empirical models should be expressed as a function of the accelerating factors. So far there are no such models that proved to be universal. Until this situation changes, the non-linear approach, therefore, does not look attractive when considering ASLT by multiple accelerating factors. So far, the practical use of multiple accelarating factors seems to be limited only to cases having well-established valid models.


496

15.8


Dynamic methods

In many ways the simplest method of accelerating the rate of reaction is by placing the product at elevated constant temperatures. Data fitting establishes the parameters of the kinetic model which in turn helps in estimating the shelf life at storage conditions. More explicitly, the latter means storage at constant temperature. Based on the evaluated model, prediction of shelf life in a dynamic situation where the temperature is changing during handling and storage, may be inaccurate in cases where history effects exist. Quast and Karel29 found such an effect in oxidation of potato chips. Similarly, Labuza and Ragnarsson30 reported a significant history effect in lipid oxidation when transfering samples from one temperature to another. This, in turn, casts doubts about using Arrhenius or any other model for dynamically changing temperatures. Therefore, one must consider the possibility that the mode of changing the storage conditions might have an extra effect on the extent of deterioration. Regular kinetic models do not account for such extra effects. Therefore, for predicting shelf life under dynamic storage conditions one has not only to establish a kinetic model but also to determine if there might be a history effect and how it might be accounted for. In cases where the kinetic experiments are carried out at constant levels of the accelerating factor, for example temperature, one must include a step where samples are transferred from one temperature to another and monitoring whether their kinetics are the same as those of the samples that are kept all the time at that constant condition. Another way of establishing the existence of history effects may be by dynamic testing as will be discussed later in this section. For sake of convenience, the following discussion will use temperature as the main example of the kinetic factor in dynamic testing. The common term for such procedures is non-isothermal testing.8,31±40 Before getting into the discussion about dynamic testing, it may be helpful to consider first some of the pros and cons of such methods compared to the ones that are using constant conditions. When using constant temperature, for example, the effect of the come up and cooling time is assumed to be negligible due to slow kinetics. However, when the kinetics is very fast, one faces an experimental problem for which dynamic testing presents a good way to overcome such a problem. On the other hand, constant temperatures are much easier to maintain. One way, therefore, to improve the accuracy of the dynamic non-isothermal testing is to change the temperature in steps.39 Another advantage of dynamic testing is when it is possible to continously monitor the extent of deterioration. In such a case, one needs a much smaller number of samples as compared to isothermal tests where samples are withdrawn periodically for analysis. When dynamic testing is used as an ASLT procedure, shortening of the experimental time simply means that samples are not allowed to stay at the lower range of temperature for long enough to validate the model. Most of the significant kinetic data are obtained from the higher range of temperatures. In dynamic testing, the product is subjected to conditions where the kinetically active factor is programmed to change with time in any desired way.



497

Fig. 15.6 Schematic diagram of dynamic testing of deterioration processes.

That creates a situation where both the extent of deterioration and the value of the kinetic factor are changing with time (Fig. 15.6). At any given time, namely at a given level of the kinetic factor, the rate of reaction can be obtained by a numerical or graphical derivative of the deterioration curve. When running a non-isothermal ASLT procedure and following the extent of deterioration, the rate of reaction is the slope of the curve of the extent of deterioration at a momentary value of temperature and time. The obtained data may be fitted with models that have been established or believed to be valid for the deterioration processes of the tested product. In order to test for history effect, one has to run another dynamic experiment having a different pattern of how the kinetically active factor is programmed and checking for a discrepancy between the calculated results and the actual ones. Another important way to establish the deterioration model and evaluate its parameters is to express the temperature history and the kinetics in a single rate equation. In cases where the Arrhenius model is valid, its coefficients K0 and Ea can be evaluated by one non-isothermal procedure. This is because the rate of reaction is dependent only on temperature. In case of the non-linear kinetic model, the rate of reaction is dependent also on time. That requires a different data analysis approach. In this case too the temperature history and the kinetics are expressed in a single rate equation. Solving this equation will result in an expression relating the extent of deterioration as a function of the model parameters that need evaluation. The technique was described by Corradini et al.41,42 In their case, the combined equation contained three adjustable parameters. Therefore, they had to have the results of four tests having four different dynamic temperature histories. They used the results of three experiments to create three equations with the three deterioration parameters (B, Tc and ) as the three unknown. By solving these three equations, the model parameters were determined. The model validity was thereafter tested by comparing the curve


498


generated with these parameters inserted into the model equation with the actual data from the fourth experiment. Non-isothermal procedures, as well as dynamic moisture content tests have been reported in the food and in the pharmaceutical related literatures.1,8,31±41 A combination of non-isothermal tests and compressed oxygen has been reported for products that are susceptible to oxidation.39 A combination of dynamic tests involving temperature and moisture content have been succesfully used for accelerating the procedure for kinetic model evaluation.43±45

15.9

The `no model' approach

The `no model' approach is a term used for the accelerated shelf life testing method that assumes that a valid kinetic model exists but does not require experiments to evaluate it. This approach, which is a sort of dynamic test, may apply only to cases where the kinetically active factor (F) is changing during storage in a monotonically, continuous way. The ASLT technique is based on monitoring the extent of deterioration in the same product in which that factor is programmed to change in such a way that it goes through the `storage' cycle in a shorter period. The obtained data are then converted into real storage conditions by a calculation that is based only on knowing how the kinetically active factor (F) is changing with time (t), namely on having the following function (g): F gt

15:31

The inverse of that equation yields the function (f) of how time relates to the changing factor: t f F

15:32

It should be noted that this equation might have an analytical expression, but may as well represent a numerical or graphical datum. Assuming that a valid kinetic model exists for the deterioration reaction, it will have the following form: dD KFdt

15:33

The value of dt may be replaced in this equation by using the derivative of Eq. 15.32, namely: dt f 0 FdF

15:34

Thus Eq. 15.33 changes into: dD KF f 0 KdF

15:35

When we have two samples of the same product, one at actual storage conditions and the other at accelerated test conditions (denoted by subscript s and a, respectively), the ratio between their rate of deterioration is:


Accelerated shelf life testing of foods dDs KF f 0 FdFs dDa KF f 0 FdFa

499 15:36

Thus, the rate of deterioration at actual storage conditions is related to that at accelerated ones by: dDs

KF f 0 FdFs dDa KF f 0 FdFa

15:37

Let us consider first a situation where the kinetic factor (F) is changing linearly with time both in storage and accelerated test conditions, thus having the following respective expressions: F F 0 bs t

15:38

F F0 ba t

15:39

where b is a constant. Using the inverse form of these equations, the ratio of their derivative is: fs0 F ba fa0 F bs

15:40

Therefore, the ratio between the extent of deterioration in this case is: R F KFdF F0 ba b s D ÿ D0 a a D ÿ D0 a D ÿ D0 s R F bs bs KFdF F0

15:41

a

Since both integrals in this equation are only functions of the factor F, they have the same value and therefore cancel out. The extent of deterioration at storage conditions is therefore obtained by accelerating the change in the kinetically active factor with time and multiplying the obtained data by the ratio of the rates of change. So far, this method is applicable only to cases where the kinetic factor is changing linearly with time. The application of this approach may be extended also to the general situation, which is expressed by Eq. 15.37. In that case, it is possible to divide the whole range of these equations into n sections, each of which may be approximated by a straight line with a slope, which can be calculated from the derivative of this equation. The basic equation in this case will be: Dj s

fs0 Fj dF bj a Dj a Dj a 0 fa Fj dF bj s

The extent of deterioration is therefore: n n X X fs0 Fj Dj a Dj s D ÿ D0 s f 0 Fj j1 j1 a

15:42

15:43

This `no model' approach was developed and successfully tested for a moisture-sensitive dry product.46 The product was packaged in a water vapor


500


permeable plastic film. Since the water activity in common storage conditions of such a product is higher than that of the packaged foods, the product will continuously absorb moisture through the film. The accelerated shelf life testing in this case was carried out by packing the same product in a film that has significantly higher water vapor permeability than the original one. In both the actual storage and the accelerated test conditions, the change in moisture content with time is not linear. In fact, the derivative of the relationship between time and moisture content for the samples that were kept at external constant water activity (ae) can be expressed:46 f 0 m kPae ÿ hmÿ1

15:44

where h denotes a function of moisture content (m), k is a constant and P is the packaging film permeability to water vapor. If different films are used for storage and for accelerated tests having a permeability of Ps and Pa, respectively, then: fs0 m Pa fa0 m Ps

15:45

In that case the extent of deterioration is given by: Pa D ÿ D0 s D ÿ D0 a Ps

15:46

This is the same solution as the linear case owing to the fact that the external water activity is the same for storage and accelerated tests. Such an accelerated shelf life testing method is simple to perform, especially since it does not require the evaluation of the kinetic model. However, there is one important problem that should be considered. It has to do with the fact that the higher the rate that one programs the change of the kinetic factor, namely the moisture content in this example, the lower the extent of deterioration. That is simply the result of the fact that the deterioration reaction is given less time to develop. This approach is therefore more effective, the better the accuracy and sensitivity of the analytical method used to monitor the deterioration process. In any case, the acceleration ratio in this approach is very dependent on how small a fraction of the total acceptable extent of deterioration may be significantly determined.

15.10

Combination of approaches

The application of a combination of methods to accelerated shelf life testing has the same advantages as using multiple accelerating factors. Such a combination may provide an effective approach in obtaining a high acceleration ratio of the deterioration reaction at a minimal cost of prediction error by staying closer to actual storage conditions. Moreover, this approach provides potentially the largest number of avenues to ASLT. One may use a combination of multiple factors together with initial rate and `no model' approaches. Mizrahi and Karel



501

have used a combination of the `no model' approach together with elevated temperature for accelerated stability tests of moisture-sensitive products.47 This combination presents an interesting case of how to link the effect of two methods where one requires evaluation of the kinetic model and the other one does not. The assumption was that the Arrhenius equation is a valid kinetic model for the rate of deterioration at different temperatures when the moisture content is kept constant. The procedure is based on packing the product in films of different permeability and placing them in an environment of the same, or different, water activity and elevated temperatures. The temperature changes not only the rate of reaction but also the moisture gain. Therefore, in order to evaluate the parameters of the Arrhenius equation one has to separate the two processes. The technique is based on the following steps:47 · Arbitrarily select a reference moisture gain curve. It may be, for example, the moisture gain of the product at actual storage conditions. For some cases, one may conveniently select a straight line. · At each temperature, transform the extent of deterioration to the reference moisture gain line by using the procedure outlined in the `no model' approach, namely by using Eq. 15.37 or 15.46 for the simple case where the ratio of the moisture gain is constant. · Use the transformed data, which are now normalized to the same reference line, to obtain the parameters of the Arrhenius equation. · Use the combination of the reference data and Arrhenius equation to extrapolate the data to actual storage conditions.

15.11

Problems in accelerated shelf life tests

The problems that are related to ASLT may be classified into three main groups. The first has to do with those cases where no valid kinetic model is believed to exist for any accelerating kinetic factor. No accelerated test procedure is available for such a case. The second kind of problem is encountered when a model does exist but it is very complicated and requires the evaluation of too large a number of parameters. The experimental procedure in such a case may prove very cumbersome to a point where the ASLT procedure may not be practical. The third group of problems relates to the application of valid ASLT methods. These problems are discussed in the following section. 15.11.1 Absence of a deterioration index Food products may be judged on the basis of sensory evaluation that is influenced by the combined effect of a multitude of different reactions. In many cases, a measurable deterioration index, which correlates well with the sensory evaluation, is unavailable. The product may therefore be judged only on the basis of being acceptable or unacceptable and not by a continuous measurable scale, thus eliminating the possibility of using the ìnitial rate', the `no model',


502


or the `dynamic' approaches to accelerated stability tests. However, the kinetic model approach may be used in such cases simply by assigning the kinetic constant (K), at constant conditions, a value of: 15:47

K 1=tc

where tc is the critical time that marks the end of the shelf life of the product. This approach arbitrarily assigns the point of product failure a value of one. As in any other kinetic study, this kinetic constant is evaluated by an experimental procedure that is carried out at different constant storage conditions. The obtained data of the values of the kinetic constant as a function of these conditions provide the basis for evaluating the kinetic model and its parameters. That model can be used for predicting shelf life by integrating the kinetic equation and finding the time it takes to reach a degree of deterioration of one. This approach is exactly the same as the time±temperature tolerance (TTT) that has been extensively used to predict shelf life mainly in frozen products.48,49 15.11.2 Statistical problems Statistics is an essential part of designing the experimental procedures and analyzing the data both in common kinetic studies as well as in ASLT. Statistical methods have been critically evaluated, for example, for their use in data fitting of the Arhhenius model.50,51 In any case, it is essential that the proper statistical methods be used in ASLT. One particular subject in that respect, which relates to the validation of kinetic models, should be especially noted. The validity of the model is best established when kinetic data are available for both the actual storage and the accelerated test conditions. Obviously, the ASLT technique by itself lacks the capability of verifying the validity of the model, especially an empirical one, for actual storage. Moreover, when any model is used its parameters are evaluated only by using the data of the very high rate of reaction. That may produce a large deviation of the extrapolated data from normal conditions. One should therefore use statistical methods that test the sensitivity of the model by a cross-validation method. In principle, these methods use part of the data to verify the validity of the model. This requires a wider range of accelerated storage conditions. The closer they are to the actual storage conditions the better. Such an approach costs more both in time and in experimental effort.

15.12

Future trends

The subject of prime importance in ASLT and in the prediction of shelf life will continue to be the pursuit of simple-to-handle versatile kinetic models of product deterioration and establishing their validity. Such models are ideally expected to be a priori valid for any deterioration process, regardless of how complex it is, under any handling and storage conditions. In fact, this is what the non-linear



503

kinetics approach claims to offer, as discussed earlier in this chapter. As already indicated, the non-linear kinetic models' approach is based on empirical equations that potentially cover a broad range of deterioration processes. The main difficulty with these models is that the resulting rate equation requires a numerical solution. This need not be a problem, though, because almost all modern mathematical software can handle such differential equations. Furthermore, there is a need now for secondary versatile models to correlate between the parameters of the nonlinear models and the accelerating factors. Therefore, it is reasonable to believe that the major thrust in the future will be in two areas. The first is to further establish the versatility of the non-linear kinetic models. The second area is to establish the validity and versatility of secondary equations preferably having a minimal number of constants. As already indicated, such work has been done with regard to temperature as the accelerating factor. In a similar way, it is expected that future research work will provide more information about the behavior of other accelerating factors that may be considered in ASLT. In future efforts to establish versatile secondary equations, one might expect also an attempt to reduce the number of parameters in the kinetic model. The best example so far is the shape parameter in the Weibullian model that shows little sensitivity to temperature changes. This greatly simplifies the deterioration rate model, which with a constant , has only two temperature-dependent terms that need to be expressed algebraically. A practical usage of non-linear models for ASLT and shelf life prediction must obviously be based first and foremost on the validity of the models. For cases where these models are proven valid, one should expect that future efforts will be dedicated also to making the non-linear kinetics much more popular. This may not be a simple task, especially when the curve fitting model combines both the primary and the secondary equations. It may be even more complex when using such models in dynamic ASLT procedures. The group of Peleg and his colleagues at the University of Massachusetts has been handling these problems using available mathematical software. It should be expected that their work and of other people will help in formulating experimental and data analysis ASLT procedures as well as in developing software that will make their use relatively simple.

15.13 1. 2.

References

Shelf-life Dating of Foods, Westport, CT, Food & Nutrition Press, 1982. Novel approaches to accelerated storage tests. In Engineering and Food, Vol 2: Preservation Processes and Related Techniques, Spiess W E L and Schubert H, eds, pp. 794±803, New York, Elsevier Applied Science, 1990. 3. TEIXEIRA NETO R O, KAREL M, SAGUY I and MIZRAHI S, Òxygen uptake and 3-carotene decoloration in a dehydrated food model', J of Food Science, 1981 46 665±76. 4. QUAST D G and KAREL M, Èffects of environmental factors on the oxidation of potato chips', J of Food Science, 1972 37 584. LABUZA T P, MIZRAHI S,


504 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21. 22. 23.

Food and beverage stability and shelf life SAGUY I, GOLDMAN M, HOREV B and KAREL M, Àn improved method for the determination of adsorbed/entrapped gases in dehydrated foodstuffs', Z Lebensm Untess Forsch, 1983 177 359±63. SAGUY I and KAREL M, `Modeling of quality deterioration during food processing and storage', Food Technology, 1980 34 (2) 78±85. WEISSMAN I, RAMON O, KOPELMAN I J and MIZRAHI S, À kinetic model for accelerated tests of Maillard browning in a liquid model system', J of Food Processing and Preservation, 1993 17 455±70. CORRADINI M G and PELEG M, `Shelf-life estimation from accelerated storage data', Trends in Food Science and Technology, 2007 18 37±47. SAGUY S and PELEG M, Accelerated and parallel storage in shelf life studies. In An Integrated Approach to New Food Product Development, Moskowitz H R, Saguy S and Straus T, eds, pp. 429±55, Boca Raton, FL, CRC Press/Taylor & Francis, 2009. MIZRAHI S, LABUZA T P and KAREL M, `Feasibility of accelerated tests for browning in dehydrated cabbage', J of Food Science, 1970 35 804±7. MIZRAHI S and KAREL M, Àccelerated stability tests of moisture sensitive products in permeable packages at high rates of moisture gain and elevated temperatures', J of Food Science, 1977 42 1575±9. LABUZA T P and RIBOH D, `Theory and application of Arrhenius kinetics to the prediction of nutrient losses in food', Food Technology, 1982 36 66±74. LABUZA T P and KAMMAN J, Reaction kinetics and accelerated tests simulation as a function of temperature. In Application of Computers in Food Research, Saguy I, ed., New York, Marcel Dekker, 1983. LABUZA T P and SCHMIDL M K, Àccelerated shelf-life testing of foods', Food Technology, 1988 39 (9) 57±62, 64. TSOUMBELI M N and LABUZA T P, Àccelerated kinetic study of aspartame degradation in the neutral pH range', J of Food Science, 1991 56 1671±5. NELSON K and LABUZA T P, `Water activity and food polymer science: implications of state on Arrhenius and WLF models in predicting shelf-life'. J of Food Engineering, 1994 22 271±90. PELEG M, ENGEL R, GONZALES-MARTINEZ C and CORRADINI M G, `Non-Arrhenius and non-WLF kinetics in food', J of the Science of Food and Agriculture, 2002 82 1346± 55. CORRADINI M G, NORMAND M D and PELEG M, Non linear growth and decay kinetics ± principles and potential food applications. In Food Engineering: Integrated Approaches, GutieÂrrez Lopez G F, Barbosa-CaÂnovas G V, Welti-Chanes J. and Parada Arias E, eds, pp. 47±71, New York, Springer, 2008. WILLIAMS M L, LANDEL R F and FERRY J D, `The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids', J of Chemical Engineering, 1955 77 3701±7. KAREL M, ANGLEA S, BUERA P, KARMAS R and LEVI G, `Stability related transitions of amorphous foods', Thermochimica ACTA, 1994 246 (2) 249±69. BUERA P and KAREL M, Àpplication of the WLF equation to describe the combined effect of moisture, temperature and physical changes on non-enzymatic browning rates in food systems', J of Food Processing and Preservation, 1993 17 31±47. KAREL M, Food research tasks at the beginning of the new millennium ± a personal vision. In Water Management in the Design and Distribution of Quality Foods, Roos Y H, Leslie R B and Lillford P J, eds, pp. 535±59, Lancaster, PA, Technomic, 1999. CHIRIFE J and BUERA P, `Water activity, glass transition and microbial stability in


Accelerated shelf life testing of foods 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

505

concentrated semimoist food systems', J of Food Science, 1994 59 921±7. BUERA M D P, CHIRIFE J and KAREL M, À study of acid-catalyzed sucrose hydrolysis in an amorphous polymeric matrix at reduced moisture contents', Food Research International, 1995 28 (4) 359±65. KARMAS R and KAREL M, `Modeling Maillard browning in dehydrated food systems as a function of temperature, moisture content and glass transition temperature', ACS Symp Ser, 1995 610 64±73. CARDONA S, SCHEBOR C, BUERA M P, KAREL M and CHIRIFE J, `Thermal stability of invertase in reduced-moisture amorphous matrices in relation to glassy state and trehalose crystallization', J of Food Science, 1997 62 (1) 105±12. SCHEBOR C, BUERA M P, KAREL M and CHIRIFE J, `Color formation due to nonenzymatic browning in amorphous glassy anhydrous model systems', Food Chemistry, 1999 65 (4) 427±32. DATTATREYA A, ETZEL M R and RANKIN S A, `Kinetics of browning during accelerated storage of sweet whey powder and prediction of its shelf life', International Dairy Journal, 2007 17 177±82. QUAST D G and KAREL M, `Computer simulation of storage life of foods undergoing spoilage by two interacting mechanisms', J of Food Science, 1972 37 679±83. LABUZA T P and RAGNARSSON J O, `Kinetic history effect on lipid oxidation of methyl linoleate in model system', J of Food Science, 1985 50 (1) 145±7. LABUZA T P, À theoretical comparison of losses in foods under fluctuating temperature sequences', J of Food Science, 1979 44 1162±8. LABUZA T P, BOHNSACK K and KIM M N, `Kinetic of protein quality loss stored under constant and square wave temperature distributions', Cereal Chemistry, 1982 59 142±8. RIBOH D K and LABUZA T P, `Kinetics of thiamine loss in pasta stored in a sine wave temperature condition', J of Food Processing and Preservation, 1982 6 (4) 253±64. TUCKER I G and OWEN W R, `High information kinetic studies: non-isothermal programmed acid concentration kinetics', International Journal of Pharmaceutics, 1982 10 323±37. ZHAN X, YIN G, WANG L and MA B, Èxponential heating in drug stability experiment and statistical evaluation of nonisothermal and isothermal prediction', J of Pharmaceutical Sciences, 1997 86 709±15. CORRADINI M G and PELEG M, À model of non-isothermal degradation of nutrients, pigments and enzymes', J of the Science of Food and Agriculture, 2004 84 217±26. OLIVA A, LLABRES M and FARINA J B, `Data analysis of kinetic modeling used in drug stability studies: isothermal versus nonisothermal assays', Pharmaceutical Research, 2006 23 2595±602. ARANGUIZ M Y F, TORRE S and BERRAONDO M R, À simulation study with statistical evaluation for the determination of non-isothermal kinetics conditions in drug stability', European Journal of Pharmaceutical Science, 2007 31 277±87. LIN B, ZHAN X C, LI L L, LI C R, QI H J and TAO J L, `Step nonisothermal method in kinetic studies of captopril oxidation under compressed oxygen', Yakugaku Zasshi, 2008 128 617±24. PELEG M, CORRADINI M G and NORMAND M D, Ìsothermal and non-isothermal kinetic models of chemical processes in foods governed by compeing mechanisms', J Agricultural and Food Chemistry, 2009 57 7377±86. CORRADINI M G, NORMAND M D and PELEG M, `Prediction of an organism's inactivation pattern from three single survival ratios determined at the end of three non-isothemal


506 42.

43. 44. 45. 46. 47. 48. 49. 50. 51.

Food and beverage stability and shelf life heat treatments', International Journal of Food Microbiology, 2008 126 98±111. CORRADINI M G, NORMAND M D, NEWCOMER C, SCHAFFNER D W and PELEG M, Èxtracting survival parameters from isothermal, isobaric, and `ìso-concentration'' inactivation experiments by the ``3 end point method''', J of Food Science, 2009 74 R1±R11. SAGUY I, MIZRAHI S, VILLOTA R and KAREL M, Àccelerated method for determining the kinetic model of ascorbic acid loss during dehydration', J of Food Science, 1978 43 1861±4. HARALAMPU S G, SAGUY I and KAREL M, Ìdentification of moisture sensitivity models of packaged materials under simulated storage conditions', Mathematical Modelling, 1986 7 1±13. HARALAMPU S G, SAGUY I and KAREL M, `The performance of a dynamic stability test for moisture sensitivity', Mathematical Modelling, 1986 7 15±25. MIZRAHI S and KAREL M, Àccelerated stability tests of moisture-sensitive products in permeable packages by programming rate of moisture content increase', J of Food Science, 1977 42 958±63. MIZRAHI S and KAREL M, Àccelerated stability tests of moisture sensitive products in permeable packages at high rates of moisture gain and elevated temperatures', J of Food Science, 1977 42 1575±9. VAN ARSDEL W B, COPLEY M J and OLSON R L, Quality and Stability of Frozen Foods Time-Temperature Tolerance. New York, Wiley-Interscience, 1969. JUL M, The Quality of Frozen Foods, London, Academic Press, 1984. COHEN E and SAGUY I, `Statistical evaluation of Arrhenius model and its applicability in prediction of food quality losses', J of Food Processing and Preservation, 1985 9 273±90. HARALAMPU S G, SAGUY I and KAREL M, Èstimation of Arrhenius model parameters using three least squares method', J of Food Processing and Preservation, 1985 9 129±43.


16 Microbiological challenge testing of foods E. Komitopoulou, Leatherhead Food Research, UK

Abstract: The chapter discusses the basic principles of microbiological challenge testing and provides some useful tips for the effective design and application of challenge testing protocols for the assessment of food safety, quality and stability. Differences between shelf life and challenge testing are highlighted. Specific advantages and limitations of challenge testing are discussed with emphasis on the use of the results in the production of microbiological predictive models. Key words: shelf life, challenge testing, predictive models, food safety, food quality.

16.1 Introduction: role of challenge testing in shelf life evaluation According to Notermans et al. (1993), microbiological challenge testing is an important tool used to: · determine the ability of the food matrix to support, or not, microbial growth or survival, i.e., to determine its safety and stability during storage until consumption, · establish the product's shelf life, · aid in product formulation in terms of intrinsic control factors (e.g. pH and water activity), and · establish critical points in a processing line. There is often confusion around the applicability of shelf life analysis versus that of microbiological challenge testing. In shelf life analysis, the product is


508


stored under the normal conditions and analyzed over time to ensure that it is safe and stable. This approach assumes good manufacturing conditions under a HACCP plan which would limit the chances of micro-organisms, other than the normal background flora of the product (not pathogens), contaminating the product. Therefore, in shelf life trials, one assumes that analysis will target the naturally present spoilage micro-flora growing during storage under stipulated conditions. On the other hand, challenge testing is designed to answer the question of whether the product could be safe and stable if accidentally contaminated with pathogenic or spoilage micro-organisms, i.e., whether a specific product formulation would favor or inhibit their growth. The aim of challenge testing is to simulate what could happen to a product during production, processing, distribution or subsequent handling by consumers following inoculation with relevant micro-organism(s) and storage under the representative conditions, from production to consumption (Notermans and In't Veld, 1994).

16.2

Basic principles

The usefulness or appropriateness of challenge testing depends on factors such as the probability of the product supporting microbial growth (spoilage and pathogenic micro-organisms), and knowledge of its previous history, e.g., raw materials, processing, etc. This chapter discusses the major factors that need to be considered in the design of a microbiological challenge test, starting from the selection of the challenge micro-organisms, product inoculation methodology, sample analysis and results interpretation (Fig. 16.1). 16.2.1 Factors affecting microbial growth and survival The ability of a food matrix to support, or not, microbial growth and survival is a complex process that involves a combination of intrinsic and extrinsic factors. Water activity, pH, nutrient availability, oxidation-reduction (redox) potential, the presence of naturally occurring antimicrobial compounds and background micro-flora (competitive micro-flora) are all factors characteristic of the food itself (intrinsic). Changes in those factors as a result of the food's environment, e.g., packaging, processing, storage time and temperature (extrinsic factors), are those that finally determine its shelf life. Knowledge of the minimum, maximum and optimum microbial growth conditions is important in effective risk assessments, when decisions need to be made as to the possibility of specific micro-organisms growing in the product during its shelf life. Indicative water activity and ranges of pH values of specific foods are shown in Tables 16.1 and 16.2, respectively. Approximate water activity and pH values for the growth of major pathogenic and spoilage microorganisms are shown in Table 16.3 (IFT/FDA, 2003a). Dominant micro-organisms in food are those able to utilize available nutrients with carbohydrates and amino acids being utilized first, followed by more complex molecules. Gram-negative bacteria are generally able to derive


Microbiological challenge testing of foods

509

Fig. 16.1 Steps in microbiological challenge testing.

their nutritional requirements from available nutrients, while on the other hand, Gram-positive bacteria are generally considered more fastidious in their requirements. Despite its importance in determining microbial growth rates, due to the complexity of its effects on micro-organisms, nutrient availability cannot be considered to be a strong predictive tool per se of microbial behavior in food. Practical difficulties in measuring redox potential in foods and in the interpretation of the redox potential values obtained have led to this factor being Table 16.1 Typical water activity (aw) values of selected food categories Food category

aw

Bread crust Bread white Cake icing Dried fruit Cereal Jam Fresh meat, poultry, fish Cured meat Fresh fruit and vegetables

0.30 0.94±0.97 0.76±0.84 0.55±0.80 0.10±0.20 0.75±0.80 0.99±1.00 0.87±0.95 0.97±1.00

Source: Adapted from IFT/FDA (2003a)


510

Food and beverage stability and shelf life Table 16.2 Typical pH values of selected food categories Food category

pH

Milk Mayonnaise Fish (most species) Tuna fish Ham Beef Chicken Apples Bananas Parsley Tomatoes

6.3±7.0 3.0±4.1 6.6±6.8 5.2±6.1 5.9±6.1 5.1±6.2 6.2±6.4 2.9±3.3 4.5±4.7 5.7±6.0 4.2±4.3


Table 16.3 Approximate pH and water activity (aw) values permitting growth of selected micro-organisms Micro-organism Campylobacter spp. Salmonella spp. Vibrio parahaemolyticus Bacillus cereus Clostridium perfringens Listeria monocytogenes Staphylococcus aureus (growth) Staphylococcus aureus (toxin)

pH

aw

4.9±9.0 3.8±9.5 4.8±11.0 4.3±9.3 5.5±9.0 4.39±9.4 4.0±10.0 4.5±9.3

0.98±0.99 0.94±0.99 0.94±0.99 0.91 (min.) 0.943±0.97 0.92 (min.) 0.83±>0.99 0.87±>0.99

Source: Adapted from IFT/FDA (2003a) and Micro-Facts (2007)

the least well-defined and most under-utilized parameter in microbial growth studies. Typical values of redox potential of some foods are shown in Table 16.4. However, these are highly dependent on changes in the food's pH, packaging, presence of background micro-flora, atmosphere and storage temperature and, therefore, presented values should only be taken as indicative. The ability of a dominant background micro-flora to inhibit the growth of the minority organisms in a food matrix was first described in the 1960s, when Jameson described the growth inhibition of salmonellae in the presence of a majority Gram-negative background population (Jameson, 1962). Later on, this inhibition was correlated to a rapid decrease in the redox potential caused as the dominant flora entered the stationary-phase in a mixed culture (Komitopoulou et al., 2004a, 2004b). In a food matrix, antimicrobial compounds can be part of the product's formulation (e.g., herbs and spices), can be added in the form of chemical preservatives, such as sorbate and benzoate, acetic acid, nitrite/nitrates, sulfur



511

Table 16.4 Typical redox potential (Eh; mV) values of selected food categories Food category Milk Cheddar Raw minced meat Canned meat Whole grain cereals Grape juice Lemon juice

Eh (mV) 300 to 340 300 to ÿ100 225 ÿ20 to ÿ150 ÿ320 to ÿ360 409 383


dioxide, etc., or can involve naturally occurring compounds, as a result of fermentation, produced by certain micro-organisms. Bacteriocins mainly produced by lactic acid bacteria (e.g., nisin) are the best-described naturally occurring compounds in foods, with proven antimicrobial activity against Gram-positive micro-organisms. Numerous studies on the antimicrobial activity of bacteriocins against Listeria monocytogenes have been published (e.g., ArqueÂs et al., 2005; Tahiri et al., 2009; Maks et al., 2010). However, their limited range of activity, restrictions in regulations covering their use in foods and their limited compatibility with most food matrices have restricted their application in food preservation. The effective use of antimicrobial compounds in food preservation and extension of a product's shelf life can only be a result of a combination of synergies occurring between the added compound(s) and a combination of other factors covering the food's intrinsic and extrinsic characteristics. The microbiological stability of certain shelf-stable processed cheese formulations, for example, can be taken as the best proof of Leistner's hurdle concept. This concept states that several inhibitory factors (hurdles), whilst individually unable to inhibit micro-organisms, will nevertheless be able to do so in combination (Leistner, 1995). The prolonged safe storage of processed cheese products in ambient conditions is a result of a combination of the appropriate heat process, water activity, salt and pH conditions. The application of the hurdle concept, with or without the use of preservatives, has been the basis for the development of predictive models as important tools in shelf life predictions. The effect of the external environment on the food's shelf life is determined mainly by microbiological challenge testing whereby products are intentionally contaminated with micro-organisms possessing implicit properties most closely related to the intrinsic and extrinsic properties of the food. The contaminated, spiked, product is then stored at the recommended storage temperature, and within the recommended packaging conditions (appropriate gas atmosphere), for a specific time period which would correspond to the current, or intended or desired shelf life of the product. Important parameters in the determination of the safe shelf life of foods using challenge testing involve the choice of


512


challenge micro-organisms, the choice of inoculation methodology, including the inoculum levels and preparation conditions. 16.2.2 Selection of challenge organisms The choice of micro-organisms to be used in challenge testing should be a result of an assessment of the risk for food contamination and of the intrinsic and extrinsic characteristics of the food to support microbial growth. Knowledge of previous history of the food and any previous indications of pathogen growth, e.g. implication in episodes of food poisoning, is a very important determinant of the choice of microbial inocula to be used in any challenge testing. A list of indicative strains previously used in challenge testing of different foods is shown in Table 16.5. It is generally accepted that the ideal challenge micro-organisms are natural isolates having previously been isolated from similar type products. Use of strains from recognized culture collections (e.g., ATCC, NCTC) can be an alternative solution when natural isolates are not available. However, it is important that strain source and availability are clearly indicated within the challenge trial design. The latter may then need to be adjusted to accommodate a couple of extra steps in the preparation of the microbial inoculum to allow for a Table 16.5 foods

Typical pathogenic micro-organisms used in challenge testing of various

Food category

Challenge micro-organism(s)

Salad dressings

Salmonellae, S. aureus

Dairy products

Salmonellae, S. aureus, C. botulinum, enterohemorrhagic E. coli, L. monocytogenes

Confectionery products

Salmonellae

Sauces and salsas stored at ambient temperature

Salmonellae, S. aureus

Cooked or dried meat and poultry

C. botulinum, C. perfringens, L. monocytogenes, Salmonellae, S. aureus, enterohemorrhagic E. coli

Fish and seafood

B. cereus, C. botulinum, L. monocytogenes, Salmonellae, Shigella spp., S. aureus, Vibrio spp.

Fruits and vegetables

B. cereus, C. botulinum, enterohemorrhagic E. coli, L. monocytogenes, Salmonellae, Shigella spp., Y. enterocolitica

Cereal grains and related products (e.g., fresh pasta, cooked rice)

B. cereus, C. botulinum, Salmonellae, S. aureus

Source: Adapted from IFT/FDA (2003b) and NACMCF (2009)



513

certain degree of adaptation of the strains to the product's intrinsic conditions (e.g., low pH, high sugar, etc.) prior to inoculation. Pre-adaptation, or microbial `conditioning' can be a lengthy, nevertheless essential, process in producing the worst-case inoculum scenario using otherwise `matrix-unfamiliar' strains. Use of conditioned strains can have its own benefits and drawbacks. In the absence of natural isolates, use of conditioned strains would represent the most valid alternative. However, such strains are generally characterized by faster growth rates and shorter lag-phases than their natural counterparts. This can seriously affect the results of challenge testing and although in reality this would represent a fail-safe result, it would not be particularly useful in trials designed to establish the time, within the storage period, required to achieve a specific log-increase in microbial numbers. When challenge trials are designed to assess the effect of another stress application (e.g., heating) on microbial loads, the use of artificially conditioned strains can result in false results and the indication of thermal death times significantly smaller than normal. In these types of challenge trials, it is advisable to allow conditioned isolates some adjustment time to survive, or even grow by one or two logs, in the food matrix prior to carrying out any thermal death trials. Use of combinations of challenge micro-organisms in microbial cocktails has generally been recommended, and has most often been included in the inoculation methodology part of challenge trial designs. Provided there is a lack of any mutual antagonism, which could affect the fate of the individual strains during storage, use of such a mixed inoculum represents the most realistic scenario that can also have its own limitations. In a mixed culture, individual strains may grow faster as a result of utilizing nutrients produced by the other strains in the cocktail. They may, however, grow more slowly as a result of a certain degree of competition, especially in cases where a cocktail of Gram-negative strains is used. To minimize the effect of strain competition, it is recommended that where possible, mixtures of Gram-negative strains are avoided. Use of a microbial cocktail can affect the recovery and enumeration of the individual strains during sampling, significantly compromising the results of the challenge testing. This can be the case when the work requires distinguishing between two or more different species used in a single cocktail, in which cases selective or indicator media need to be used. Media selectivity represents an additional hurdle to already stressed micro-organisms (e.g., those originally spiked in and recovered from a low pH, preservative-containing food matrix), significantly underestimating their ability to grow in the product. The choice of a mixed or single strain inoculum has generally been determined by available funds (mixed strain studies being the most cost-effective), rather than by any specific benefits overriding the limitations. In some cases, use of selective media to distinguish between strains in a cocktail can be avoided by using genetically modified strains and utilizing specific genetic markers, e.g., luminescence/fluorescence or antibiotic resistance. When this is the case, care should be taken to ensure that the genetically modified organisms used share the same growth/survival behavior as their parent strains under the same conditions.


514


The use of surrogate strains, e.g. use of Clostridium sporogenes as a surrogate for Clostridium botulinum, or Listeria innocua as a surrogate for Listeria monocytogenes, is common when pathogenic strains cannot be used in processing plants. Care should be taken to choose the right surrogate strain(s) in the challenge trials so that meaningful results are obtained. The choice of surrogate strains should be based on the following key attributes of the successful surrogate (IFT/FDA, 2003b): · lack of pathogenicity · stable and consistent growth/survival characteristics, bearing strong similarity to those of the target pathogen when exposed to the same formulation or processing factors (e.g., pH and temperature sensitivity, oxygen tolerance, etc.) · inactivation characteristics and kinetics similar to those of the target pathogen · easy to produce, time-stable, high-density population that is easy to enumerate and distinguish in a mixed culture environment · genetically stable to ensure reproducibility of results independent of the laboratory or time of experiment · stress and injury susceptibility similar to that of the target pathogen. 16.2.3 Inoculum level and preparation The preparation and level of inoculum are both important parameters in the design of the challenge testing protocols. Inoculum levels per unit weight or volume of a product need to be realistic and in direct correlation to the purpose of the challenge testing. If the aim of the trials is to evaluate the safety and stability of a product during a specified period, then initial levels of inoculum should be between 100 and 1000 cells per gram or per ml of the product. Levels lower than 100 cells per gram or ml may be below the limits of detection in many sampling methodologies employed, thus making the incorrect assumption that the product is safe and stable, when it is not. On the other hand, levels any higher than 1000 cells per gram or ml may overcome the intrinsic preservation properties of the food matrix leading to the false assumption that the product is not safe and stable, when in reality it is. Use of inoculum levels higher than 1000 cells per gram or ml of a product applies to those trials aiming to determine microbial log-reductions following the application of a specific stress, such as heating, irradiation, etc. In these cases inoculum levels can reach 106±108 cells per gram or ml of a product, depending on the scale of the log-reduction the system tested needs to demonstrate. For example, if the aim of the trials is to confirm that a specific heating process can result in a five-log reduction in numbers of Listeria monocytogenes in cheese, then the product needs to be inoculated with a minimum of 106 cells per gram of the product and then be subjected to the specific heating process. Culture and inoculum preparation prior to inoculation is the second most important part of the challenge testing protocol following the choice of the challenge strains. There are no specific guidelines that govern inoculum



515

preparation as generally this is affected by the type of inoculum and the nature of the food matrix to be challenged. A basic rule in all inoculation trials is standardization of the protocols used in culture preparation to involve clear instructions on culture maintenance (e.g., refrigerated cultures, slants, cultures frozen in glycerol, freeze-dried cultures etc.), isolate sub-culturing and recovery. Specific guidelines for culture maintenance exist (e.g., Kirsop and Doyle, 1991) and the choice of the best protocol to follow depends mainly on available facilities. The most important aspect in culture maintenance involves strains previously isolated from extreme environments (e.g., low pH, high sugar matrices). Storage of these strains should ensure that specific characteristics developed and expressed as a result of previous exposure to extreme environments are maintained and are adequately reproduced by the appropriate media throughout storage. For example, a yeast strain previously isolated from a low pH beverage (pH 3.5) needs to be stored and maintained in a medium of similar pH. Prolonged exposure of this strain to pH conditions higher than those previously encountered might render it unable to subsequently grow under low pH conditions. This is more so for isolates (not natural) previously conditioned to grow in extreme environments as one cannot assume that adaptation can be maintained in the absence of the appropriate trigger. Cell cultivation needs to take place under the optimum growth conditions for each strain. In most cases, resuscitation involves growth for 18 to 24 h under optimum temperature and atmospheric conditions; however, procedures that involve using 48 or 72 h cultures (e.g., certain yeasts) are often used. Culture enumeration at this stage is important to determine the scale of required dilutions to achieve the target inoculum in the challenge product. Spores washed and stored in distilled water to prevent germination may need to be heat shocked immediately prior to inoculation, if prompt germination and growth are to be achieved and represent what happens in food processing. Depending on the strain, spores may also need to be thoroughly washed to minimize the transfer of free toxin in the product during inoculation, as in the case of botulinal toxin produced by Cl. botulinum. In most protocols, cultures are centrifuged and the cell pellets are washed thoroughly to avoid the transfer of any compounds (e.g., nutrients) from the laboratory growth media to the food matrix during inoculation. For liquid products, washed pellets are then resuspended in a volume of the same food product, and any subsequent dilutions of the resuspended culture are also carried out using the appropriate volumes of the same product. Product inoculation can then be carried out without alteration of the intrinsic characteristics of the matrix. For solid products, cell pellets are resuspended in appropriate diluents, also used to carry out any necessary dilutions. Product inoculation is then carried out using the minimum possible volume of inoculum to ensure that its main characteristics, e.g., pH, water activity, remain unaffected. An alternative method of inoculum preparation involves microbial cell recovery from lawn plates prepared from 24±48 h cultures. Following this protocol, lawn plates are prepared using grown cultures and are incubated under


516


optimum growth conditions. Cell recovery is then carried out using an appropriate carrier depending on the type of the challenge product, and the resuspended cells or spores are appropriately diluted and used in the challenge trials. Studies have shown that lawn-collected cells are characterized by an enhanced survival potential compared to broth-collected cells and can therefore represent a worst-case scenario in challenge experimentation (Komitopoulou and Penaloza, 2009). 16.2.4 Choosing the right method of inoculation The choice of the inoculation method is another crucial parameter in the design of challenge trials. A challenge test can be considered successful if inoculation does not affect any of the product's intrinsic or extrinsic properties. The methodology employed also needs to be reproducible and properly validated, one that also includes pre- and post-inoculation analysis of the critical characteristics of the product's formulation (e.g., moisture content, water activity and pH), to ensure that these have remained unaffected by the inoculation process. The choice and volume of the liquid inoculum carrier used in the inoculation methodology are crucial parameters affecting the success of the trials. Independent of the type of carrier used, the aim should always be to use the minimum possible volume of carrier to also ensure the least possible change in the product characteristics. In liquid products the inoculum can be suspended in a sample of the product matrix itself, creating a stock which is then used to inoculate different samples of the same product. In cases where maintaining the moisture level is important, the inoculum carrier can be the same diluent used to adjust the moisture content of the product formulation in the first case. The successful inoculation methodology needs to ensure even distribution of the inoculum within the product matrix to minimize errors in the subsequent sampling and enumeration of the challenge organism, whilst at the same time maximizing microbial exposure to the product's environment. This can prove particularly problematic when inoculation is carried out using a syringe through the packaging wall containing a rubber septum. When adequate mixing cannot be achieved, samples are first inoculated and then re-packed making sure that packaging after inoculation matches the normal packaging conditions of the product. Mixing of inoculum within the product matrix can easily be done in products with water activities higher than aw 0.96 (e.g., sauces) using the minimum volume of the inoculum carrier possible. Spraying has also been suggested and employed for inoculation of product surfaces, although care should be taken to ensure that this is carried out using the appropriate protective equipment to make sure that exposure to pathogenic aerosols is minimized. 16.2.5 Duration of trials and storage conditions The duration of the challenge trials will depend on the current, desired or intended shelf life of the product, needs to cover all of this period and also



517

include a margin beyond this time to account for the possibility of consumers storing the product beyond its recommended life. Storage of the inoculated product during this time needs to replicate the conditions the product is likely to be exposed to under normal storage conditions. However, slight abuse storage conditions, as imposed by consumers, during transport and/or at retail, can often occur and, therefore, the challenge trials need to account for the possibility of the worst, within reason, scenario of slightly abused storage conditions. Storage conditions, and storage temperature in particular, can seriously affect the duration of the challenge trials; the duration of the challenge trials of chilled products should not be expected to be the same when the products are stored at normal (e.g. 4 ëC) and slightly abusive chilled conditions (e.g., 8 ëC). The extended lag-phase sometimes observed by micro-organisms, e.g. when they are exposed to extreme conditions within the food matrix when they are introduced, can also delay the duration of the challenge trial. This delay is deemed necessary to allow the micro-organisms time to overcome the preliminary shock, reflected by the lag-phase, recover and then grow depending on the ability of the food matrix to support or inhibit growth. Other storage conditions that need to be considered in a challenge trial involve the type of packaging and nature of the gas atmosphere within the packs. A number of challenge trials are carried out to determine the safety of a film type used in packaging or to determine the role of a gas mixture on the safety and stability of a product during storage under normal temperature conditions. When the inoculation methodology used in the challenge test involves interfering with the packaging characteristics of the product, these will then need to be replicated to mimic the exact conditions normally applied to the specific product, for the challenge trial to produce any meaningful results. 16.2.6 Sample analysis and data interpretation Duration of the challenge trial will affect the frequency of sampling during storage. In all cases, sampling needs to include time zero analysis, i.e., analysis immediately after product inoculation, to verify inoculum levels and allow calculation of log differences in the numbers of the challenge micro-organisms during storage. Depending on whether the duration of the challenge trial is measured in days, weeks or months, a sampling regime needs to cover representative time points during storage. When products have a short shelf life measured in days, sampling is required daily, while a more prolonged trial duration may require sampling once or twice a week. A minimum of five to seven points are required in a challenge trial to obtain an accurate indication of the behavior of the inoculum during storage. The results from sample analysis during storage can also be used as an indicator of duration of the challenge trial; there is no point to continue sampling a product when levels of inoculum have already reached 108 cells per gram or ml of the product. When levels below the limits of detection are obtained following the standard enumeration techniques, detection methodologies specific to the target micro-


518


organism(s) need to be used instead. This is to confirm the absence of the organism and ensure that no subsequent recovery will occur that can compromise the shelf life of the product. When looking for toxin production during storage, this can be tested for at some but not necessarily all sampling points. Sample replication is an important factor contributing to the validity of the challenge trial. Duplicate or triplicate samples should ideally be analyzed at each sampling point, while it is recommended that the challenge trial is repeated twice or three times in independent trials (i.e., performed at different times, using fresh inocula and ideally carried out by different staff) in cases where a high degree of certainty is required. The selection of the media and reagents for sampling and enumeration of the challenge micro-organisms will depend mainly on the type of micro-organisms and also the nature of the challenge matrix itself. The use of selective media needs to be done with care to avoid imposing additional stresses on already stressed challenge micro-organisms in the food matrix. Absence of background flora in a product can allow the use of non-selective media. Monitoring of the background flora is carried out by setting up and analyzing un-inoculated, negative control, samples of the same product alongside the spiked samples. Positive controls, in which a portion of the same cell inoculum is used for the inoculation of laboratory media that can support maximum microbial growth, are also recommended. Both negative and positive controls can help interpret results. The presence or absence of background flora can be important in allowing or inhibiting/delaying the growth of major pathogens, respectively. Background flora can also have an indirect effect on the target micro-organisms by changing the product's critical parameters (e.g., lowering the pH). Monitoring of those parameters (mainly aw and pH, sugar and salt levels, preservatives, gas atmosphere) throughout the challenge trials can help validate the results of the challenge trials. If growth of a particular challenge micro-organism in the product is not observed, growth of the same cell inoculum in the positive control trials will indicate that absence of growth was related to the inhibitory/ preservative nature of the product and was not a result of a deficient inoculum. Interpretation of the results of a challenge trial should involve the use of all sets of data obtained during the trials; microbial growth, change in physicochemical and other intrinsic/extrinsic properties of the product, positive and negative controls, so that a comprehensive evaluation of the microbiological safety and stability of the product is achieved. Upon completion of the challenge trial, presentation of the results obtained can be in different forms depending on the individual writing up the report and any specific requirements, e.g. when the report is intended to reach a customer. Trend analysis and graphical plotting of mean log counts against time or mean survivor curves, depending on the experimental design, are undoubtedly the preferred way of presenting results. Prior to the start of the challenge trial, it is important to define the pass and fail criteria for the product and use those as the basis of any decision made around the duration of the trials and the need to reformulate. Pass/fail criteria



519

and the significance of a population increase during storage depending on the hazard significance of that population. The outcome of the challenge trials can be very important in determining the shelf life of a product and indicating any potential changes in the product formulation that could affect product safety and stability. The use of challenge trials in the development of predictive models is a confirmation of the use and application of the challenge trials in the real-time determination of a current product formulation and prediction of the microbiological safety and stability of future formulations.

16.3

Challenge testing limitations

Challenge testing is a time-consuming process both in terms of required effort but also in terms of duration and elapsed time before any results are obtained. Use of accelerated storage conditions should not be considered an option in challenge testing. Designing a challenge testing protocol requires certain skills and knowledge, starting from the initial risk assessment, choice of challenge micro-organisms, protocols of product inoculation and result interpretation. Since most of the challenge trials involve use of pathogenic micro-organisms, it is important that these trials are conducted by experts, certainly qualified microbiologists, familiar with the necessary precautions that need to be taken for safe handling of pathogenic isolates and their potential toxins. Automatically, this makes challenge testing an expensive test to perform. The most important limitation of challenge testing is that its results are only valid for the specific product formulation challenged. Any changes in the product formulation, and/or handling of the product (e.g., processing, packaging), no matter how minor these may be, can render the results of the trial invalid. Inevitably this has made challenge testing an expensive test to perform in the development of a safe product formulation and has reinforced the need for the development of predictive models.

16.4

Challenge testing and the use of mathematical models

In the 1980s, predictive microbiology was made a priority as a result of major foodborne outbreaks caused by the so-called `traditional' pathogens (e.g., Salmonella in eggs) and other emerging pathogens with unusual characteristics (e.g., ability of Listeria monocytogenes to grow at chilled conditions). Since then, development of predictive microbiological models has found applications in a wide industrial context, making predictive microbiology an established scientific discipline. In product innovation, predictive mathematical models are used in the assessment of microbial growth characteristics or death kinetics (inactivation


520


rates) associated with a particular food formulation or process condition, respectively. Their aim is to develop new products/processes, reformulate existing products and determine their shelf life and stability. In operational support, predictive models are used in support of food safety decisions around the implementation of specific food manufacturing operations. The latter may involve setting up critical control points (CCPs) in HACCP, designing heat processing protocols and assessing the impact of any deviation from the standard procedures on the microbiological safety and quality of the products. The impact of issues associated with products already in the market on consumer safety is often evaluated using predictive models, making incident support another distinct application of predictive modelling (MembreÂ and Lambert, 2008). Available predictive models (e.g., Pathogen Modelling Programme (PMP), Growth Predictor, SymPrevius and ComBase) are user-friendly and easy-to-use software packages. The choice of the most appropriate software to use will need to be based on some comparison of preliminary predictions versus real-time experimentation (involving challenge testing), using the same input of information, to confirm findings and validate applicability of the system for the particular product. The majority of the existing predictive models describe population dynamics (continuous models) and serve as comprehensive analytical tools, requiring only basic expert mathematical skills and involving the application of mathematical equations. Classical mathematical models used in predictive microbiology usually describe the population through a `top-down approach'. They deal with equations that apply to the whole population and reflect essential microbial characteristics taking into account external variables such as water activity, pH and cell density (Li et al., 2007). Mathematical models that fit within the population top-down approach involve the mechanistic, empirical and probabilistic models. Mechanistic models are those that focus on studying microbial population dynamics and relate microbial lag-phase and growth rate with the state of the inoculum used and temperature (Baranyi and Roberts, 1994; Daughtry et al., 1997). Empirical models are those that predict the lag-phase and microbial growth rate incorporating data available from external databases (e.g., ComBase and PMP) and therefore focus on describing, reproducing and also predicting microbial behaviors (Buchanan, 1991; Baranyi and Tamplin, 2004). Finally, probabilistic models focus on the study of microbial communities in the growth/no-growth interface for risk assessment and spoilage (McKellar et al., 2002). It has been stated that a predictive model can provide results at least 1000 times faster and considerably cheaper than a traditional challenge test (Zwietering et al., 1996); however, use of predictive models cannot be considered as an alternative to challenge testing. Predictions can only provide indications of the likelihood for microbial growth and survival and should only be considered as such, unless these are tailor-made models produced using results obtained from real-time challenge tests of specific products. The source of information used in the construction of the predictive models is the most



521

significant limitation to their application. The use of microbial growth and survival data from the literature and their incorporation into publicly available models have significantly restricted their industrial application. Nevertheless, they can be used to provide some preliminary indication of microbial behavior. Similarly, there are models that have been produced solely using data obtained under laboratory conditions, in broth systems, using combinations of parameters of no practical use and application to industry, using a limited number of microorganisms, mainly pathogenic, whilst no attempt has been made to validate those models in real-time food matrices. Production of tailor-made models is an expensive and lengthy process requiring the setting up of numerous challenge trials. However, it is thought to be the only precise way of producing models relevant to a specific product matrix. If a model is already in existence for a particular product design then it can be a very valuable tool for any subsequent reformulations of the same product.

16.5

Future trends

The demand for natural ingredients has been fuelled by a growing consumer preference for healthy foods. The loyal users of natural products are increasing and this trend is predicted to continue followed by an increased demand for new products to be of natural ingredients and old products to be reformulated. Replacement of chemical preservatives with natural alternatives is not the only need for product reformulation that industry is currently facing. Salt reduction in processed foods, including meats and baked products, has been on the agenda of the Food Standards Agency. Sugar (sucrose) is well known to increase the heat resistance of vegetative cells of microbes, and decrease microbial growth rates, by reducing water activity. Replacement of sucrose by an intense sweetener, or reduction by use of a sugar of higher sweetness intensity (e.g., fructose), will allow the growth of many microbes, including pathogens. As the demand for `clean labels' linked to significant product reformulation is on the rise, the requirements for verification of the safety and stability of the new formulations will increase dramatically.

16.6


Particularly useful sources of information on microbial characteristics (growth, survival, resistance) and their association with different foods are: · Wareing, P. and Fernandes, R. (eds) (2007). Micro-Facts. The Working Companion for Food Microbiologists. Leatherhead Food International, UK. RSC Publishing. · Leatherhead Food Research ± Microbiology Handbooks on: Meat Products (2009). Fernandes, R. (ed.), RSC Publishing.


522


Fish and Seafood (2009). Fernandes, R. (ed.), RSC Publishing. Dairy Products (2009). Fernandes, R. (ed.), RSC Publishing.

16.7

References

and NUNÄEZ, M. (2005). Effect of combinations of high-pressure treatment and bacteriocin-producing lactic acid bacteria on the survival of Listeria monocytogenes in raw milk cheese. International Dairy Journal, 15, 893±900. BARANYI, J. and ROBERTS, T.A. (1994). A dynamic approach to predicting bacterial growth in food. International Journal of Food Microbiology, 23, 277±294. BARANYI, J. and TAMPLIN, M.L. (2004). ComBase: a combined database on microbial responses to food environments. Journal of Food Protection, 67, 1967±1971. BUCHANAN, R.L. (1991). Using spreadsheet software for predictive microbiology applications. Journal of Food Safety, 11, 123±134. DAUGHTRY, B.J., DAVEY, K.R. and KING, K.D. (1997). Temperature dependence of growth kinetics of food bacteria. Food Microbiology, 14 (1), 21±30. IFT/FDA (2003a). Chapter III: Factors that influence microbial growth. Comprehensive Reviews in Food Science and Food Safety, 2, 21±32. IFT/FDA (2003b). Chapter VI: Microbiological challenge testing. Comprehensive Reviews in Food Science and Food Safety, 2, 46±50. JAMESON, J.E. (1962). A discussion of the dynamics of Salmonella enrichment. Journal of Hygiene, 60, 193±207. KIRSOP, B.E. and DOYLE, A. (1991). Maintenance of Microorganisms and Cultured Cells ± A Manual of Laboratory Methods, London: Academic Press. KOMITOPOULOU, E. and PENALOZA, W. (2009). Fate of Salmonella in dry confectionery raw materials. Journal of Applied Microbiology, 106 (6), 1892±1900. KOMITOPOULOU, E., BAINTON, N. and ADAMS, M.R. (2004a). Oxidation-reduction potential regulates RpoS levels in Salmonella Typhimurium. Journal of Applied Microbiology, 96(2), 271±278. KOMITOPOULOU, E., BAINTON, N. and ADAMS, M.R. (2004b). Premature Salmonella Typhimurium growth inhibition in competition with other Gram-negative organisms is redox potential regulated via RpoS induction. Journal of Applied Microbiology, 97 (5), 964±972. LEISTNER, L. (1995). Principles and applications of hurdle technology. In: Gould, G.W. (ed.), New Methods of Food Preservation. London: Blackie Academic & Professional, 1±21. LI, H., XIE, G. and EDMONDSON, A. (2007). Evolution and limitations of primary mathematical models in predictive microbiology. British Food Journal, 109 (8), 608± 626. MAKS, N., ZHU, L., JUNEJA, V.K. and RAVISHANKAR, S. (2010). Sodium lactate, sodium diacetate and pediocin: effects and interactions on the thermal inactivation of Listeria monocytogenes on bologna. Food Microbiology, 27(1), 64±69. MCKELLAR, R.C., LU, X. and DELAQUIS, P.J. (2002). A probability model describing the interface between survival and death of Escherichia coli O157:H7 in a mayonnaise model system. Food Microbiology, 19, 235±247. MEMBREÂ, J-M. and LAMBERT, R.J.W. (2008). Application of predictive modeling techniques ARQUEÂS, J.L., RODRIÂGUEZ, E., GAYA, P., MEDINA, M.



523

in industry: from food design up to risk assessment. International Journal of Food Microbiology, 128(1), 10±15. NACMCF (2009). Parameters for Determining Inoculated Pack/Challenge Study Protocols, National Advisory Committee on Microbiological Criteria for Foods, Washington, DC. NOTERMANS, S. and IN'T VELD, P. (1994). Microbiological challenge testing for ensuring safety of food products. International Journal of Food Microbiology, 24, 33±39. NOTERMANS, S., IN'T VELD, P., WIJTZES, T. and MEAD, G.C. (1993). A user's guide to microbial challenge testing for ensuring the safety and stability of food products. Food Microbiology, 10, 145±157. TAHIRI, I., DESBIENS, M., KHEADR, E., LACROIX, C. and FLISS, I. (2009). Comparison of different application strategies of divergicin M35 for inactivation of Listeria monocytogenes in cold-smoked wild salmon. Food Microbiology, 26(8), 783±793. ZWIETERING, M.H., DEWIT, J.C. and NOTERMANS, S. (1996). Application of predictive microbiology to estimate the number of Bacillus cereus in pasteurised milk at the point of consumption. International Journal of Food Microbiology, 30, 55±70.


17 Beer shelf life and stability G. G. Stewart and F. G. Priest, Heriot-Watt University, UK

Abstract: Brewing was one of the earliest processes to be undertaken on a commercial scale and consequently it became the first biological process to develop from a craft into a technology. Although the production of beer is a relatively simple process, the finished product is unstable when in its final package ± bottle, can or keg. As well as being susceptible to microbial infection, non-biological instability involves a number of complex reactions with proteins, carbohydrates, polyphenols, metal ions, thiols and carbonyls. Although our understanding of these reactions has progressed over the past 25 years, we are still far from a complete comprehension of beer instability/ stability reaction systems. Key words: beer, clarity, flavour, instability, infection.

17.1

Introduction

Brewing was one of the earliest processes to be undertaken on a commercial scale and, of necessity, it became one of the first processes to develop from a craft into a technology. Beer production is divided into five distinct processes: · malting is the germination of barley or other cereal and drying (or kilning) of the germinated cereal; · mashing is the extraction of the ground malted barley with water and separation from the insoluble material to produce wort; · wort boiling with the inclusion of hops or hop extracts; · fermentation, maturation and filtration; · packaging (used generally to mean kegging, bottling and canning). The production of beer is a relatively simple process. Yeast cells are added to the nutrient medium (the wort) and the cells take up the nutrients and utilise


528


them to increase the yeast population. The cells excrete ethanol and carbon dioxide into the medium along with a host of minor metabolites, many of which contribute to beer flavour. The fermented medium, generally after the yeast has been removed for re-use, is often called `green' beer because it usually has the aroma of green apples (due to acetaldehyde and other carbonyl compounds). The beer is then aged (conditioned, matured or lagered), maybe diluted, clarified (filtered), carbonated and packaged.

17.2

Biological instability

Biological instability involves contamination by bacteria, yeast or mycelial fungi. There is always a risk during brewing that beer can become contaminated by microorganisms. However, beer is an inhospitable environment for microbial growth: it has a low pH (less than 4.4), ethanol is present in a range of concentrations (2.5±6.5 (v/v)), there are limited nutrients due to yeast growth, hop acids are present that are bacteriostatic, the environment is anaerobic, and the liquid is carbonated. The advent of low/no alcohol beers (LAB/NABs) and products with pHs above those of traditional beers has exposed them to a greater susceptibility of infection. Most potential contaminants originate from the raw materials or unclean brewing equipment. Barley can contain Fusarium fungi that can release mycotoxins or cause gushing (see Section 17.6). It can also carry bacteria that contribute nitrosamines (potentially carcinogenic agents) and cause filtration problems. Contaminants can cause flavour deterioration, turbidity and health problems. Of the microflora found in a brewery, the Gram-positive lactic acid bacteria are the most feared. In addition to being potential beer spoilers, the lactic acid bacteria have a reputation for being `difficult' in terms of detection, recovery from spoilt beer and identification. The concerns reflect the nutritional fastidiousness of these bacteria and their variable response to the anti-microbial effects of hop iso--acids. The major bittering (and antimicrobial) substances in beer include isohumulone, isocohumulone and isoadhumulone and their cis and trans isomers. Generally, Gram-positive bacteria are sensitive to these isomerised hop acids and accordingly cannot grow in hopped beers. However, strains of Lactobacillus and Pediococus able to spoil beer are significantly more resistant to these acids. Studies by Simpson (1993) showed great variation in the sensitivity of a selection of Gram-positive bacteria to one of the major hop acids, trans-isohumulose. Although many questions remain to be answered, typically lactic acid bacteria isolated from beer will not grow when the colony is transferred to beer. This is unsatisfactory, as the spoilage status of the isolate remains unclear. It is noteworthy that hop-sensitive and hop-resistant lactic acid bacteria are indistinguishable from each other in terms of morphology, physiology and metabolism. The molecular mechanisms of hop toxicity are becoming clear (Behr and Vogel, 2009) and the genetics of hop resistance in Gram-positive bacteria are being unravelled (Sakamoto et al., 2001; Iijima et al.,


Beer shelf life and stability

529

2006). Such studies will enable the rapid molecular detection of hop-resistant strains. The major contaminating Gram-negative bacteria are acetic acid bacteria (Acetobacter and Gluconobacter) and various genera in the Enterobacteriaceae (Obesumbacterium, Citrobacter, Klebsiella) as well as Zymomonas, Pectinatus and Megasphaera (Priest and Campbell, 2003). In comparison to the Grampositive bacteria, the threat of the various Gram-negative bacteria is under reasonable control. Day-to-day management of this microbiological threat is achieved through regular acid washing (Simpson and Hammond, 1989) of pitching yeast and scrupulous attention to process hygiene. A wild yeast is defined in the brewing industry as àny yeast not deliberately used and under full control' (Gilliland, 1971). This definition of wild yeast is divided into Saccharomyces and non-Saccharomyces groupings. Irrespective of classification, wild yeast contamination of process and product can be a major cause for concern. Generally, the Saccharomyces wild yeasts are regarded as more hazardous than the heterogeneous grouping of the non-Saccharomyces wild yeasts. It is important to exclude these contaminants from the brewing process. Modern plant and good hygiene will help. Many breweries pasteurise and others membrane filter their beer to ensure biological stability. With good hygiene, the use of expensive and potentially beer damaging processes can be reduced. However, inefficient operation of either pasteurisation or membrane filtration can negatively affect a beer's non-biological stability.

17.3

Physical instability

With a few notable exceptions, consumers prefer their beer to be bright and free of particles. When beer is stored it has the potential to produce haze and the brightness is compromised. Beer's physical stability, also called colloidal stability or simply haze formation, cannot be ensured by treating beer with one `super-product' that will solve everything. Stability will be affected by the whole brewing process; consequently, care must be taken at every stage. However, raw materials are typically the source of haze precursors. There are a number of types of beer haze including: -glucan, starch, pentosan, oxalate, microorganisms and can lid lubricants. However, the primary reaction is the polymerisation of polyphenols and their interaction with specific (sensitive) proteins. When beer is cooled below 0 ëC, chill haze will form that consists of a reversible association of small polymerised polyphenols and proteins. When the beer is restored to room temperature, this haze re-dissolves and the beer becomes bright again. If beer is chilled and warmed a number of times, or if beer is stored at room temperature for an extended time period (six months or longer), permanent haze will form. This haze does not re-dissolve even when the beer is warmed to 30 ëC or higher. The balance between flavenoid polyphenols (tannoids) (Fig. 17.1) and sensitive proteins largely dictates physical or colloidal stability. Beers can differ


530


Fig. 17.1 A typical beer polyphenol.

widely in the content of these species, the relative levels of which depend upon raw materials and the process conditions employed. Haze formation is increased by a number of factors (Bamforth, 1985) but storage temperature has the greatest influence on haze formation because an increase in temperature raises the rate of the reaction. For example, pasteurisation accelerates colloidal haze formation. Oxidation (the presence of oxygen) has a great effect on beer haze formation. Extensive oxidation can increase the rate of haze appearance manyfold. Heavy metal ions (particularly iron) can promote the formation of colloidal haze. Movement of beer accelerates haze formation because of rapid interaction of colloids. Light encourages oxidation and consequently haze formation as well as other reactions that are dealt with later. Beer chill haze consists of a loose bonding of high molecular weight proteins with highly condensed polyphenols (predominantly anthocyanogens). In these loosely bound aggregates, small quantities of carbohydrates and inorganic materials are included. This loose binding is broken on warming. Haze formation occurs as a result of dissolved colloidal particles colliding and hydrogen bonds forming between them. In the course of time, increasingly large aggregates come together until they are visible as haze. Haze formation correlates with the presence of sensitive proteins (substances that precipitate with tannic acid) and tannoids (polyphenols adsorbed by polyvinyl-polypyrolidone or PVPP). The driving force for haze formation is the interaction of hydrophilic groups on these sensitive proteins with polyphenols. There are also hydrophobic proteins in beer. These surface-active species are important in foam formation (see Section 17.5). There are a number of procedures that can be employed to retard or prevent haze formation: · prevent the formation of large quantities of complex protein degradation products during beer production; · enzymatically hydrolyse the complex `sensitive' protein degradation products; · remove some of the polyphenols or sensitive proteins during brewing; · store packaged beer in cold temperatures to retard haze formation. Employing stabilisers can produce beers with a longer shelf life. The main stabilising agents, which can be used singly or together are: silica gel preparations, PVPP and proteolytic enzymes. Silica gel preparations are important



531

stabilising agents that bind hydrophilic polypeptides. They are employed in quantities of 50 to 150 g/hL and are usually dosed into the beer before filtration. There are two types of silica gel preparations used in brewing: hydrogels that have a moisture content of more than 30% and xerogels (dry gels) with a 5% water content. PVPP selectively removes phenol-containing substances. PVPP binds to polyphenols as it has a similar structure to the amino acid proline (Siebert and Lynn, 1997). Both have five-membered saturated, nitrogencontaining rings with amide bonds and no other functional groups. It is not certain whether PVPP binds to the same part of the polyphenol molecule to which polypeptides bind. This selection depends on the pH-sensitive formation of hydrogen bonds that are broken again in alkaline solution with the release of the adsorbed phenol compounds. Regeneration of PVPP with hot caustic is very effective. PVPP and silica gel preparations have been used together with good results because both polyphenol and sensitive protein components are removed (McMurrough and O'Rourke, 1997). Proteolytic enzymes are also employed as stabilising agents, but because of the advent of silica gel preparations, their use today is not as common as it was a number of years ago. The enzymes employed include papain (from papaya), bromelain (from pineapple) and ficin (from figs). These enzyme preparations are not very specific and, as well as hydrolysing haze-specific proteins, they often hydrolyse the hydrophobic foamspecific polypeptides (see Section 17.5). Consequently, the use of these enzymes often requires the addition of a foam-enhancing agent such as propylene glycol alginate. The use of propylene glycol alginate has to be treated with care because hazy beer can result if pH control is not effective (Jackson et al., 1980).

17.4

Flavour stability

The flavour stability of a beer depends primarily on the oxygen content of packaged beer. However, it is now clear that flavour stability is influenced by all stages of the brewing process (Narziss et al., 1993): · preservation of reducing substances by minimising oxygen pickup during mashing, lautering or mash filtering (the separation of unboiled wort from solid grain material (spent grains)) and wort boiling; · elimination of substances that are prone to react with flavo-active compounds like carbonyls by good mashing and wort separation procedures; · prevention of ion accumulation, such as iron and copper (Irwin et al., 1991); · controlled exposure of the wort to heat to limit the formation of Maillard reaction products (produced as a result of heating sugars with amino acids) and related substances. The role of the above reaction products in beer flavour staling reactions is ambiguous and there are reports (Bright, 2001) of their positive and negative effects.


532


In many foods, such as milk, butter, vegetables, vegetable oils, and beverages, staling is caused by the appearance of various unwanted unsaturated carbonyl compounds. It is now becoming increasingly clear that the same is true of beer staling. As already discussed, packaged beer has a limited shelf life. The phenomenon of beer aging or staling has been intensively investigated by the brewing industry with the objective to understand and control it (Bamforth, 2004) but the mechanism(s) of staling are still not fully understood. The actual compounds responsible for stale flavour vary during prolonged storage as evidenced by changes in the flavour profile of beer (Fig. 17.2) (Dalgleish, 1977). The compounds causing the sweetish, leathery character of very old beers have not been identified. However, there is evidence that the papery cardboard character of 2±4-month-old beer is due to unsaturated aldehydes. The most flavour-active aldehyde that has been conclusively proven to rise beyond flavour-threshold levels is trans-2-nonenal (Dalgleish, 1977). Other aldehydes such as nonadienal, decadienal, and undecadienal may also exceed threshold levels. Although there are many factors that will influence the flavour stability of beer, the oxygen level in the final package is of paramount importance. It is critical that this level in beer, immediately prior to packaging, is as low as possible (less than 100 mg/L) and that oxygen accumulation during filling is minimal. The adverse effects of oxidation on the flavour of finished beer have been known for a long time and some brewers add bisulfites or other antioxidants, such as ascorbic acid, to beer prior to packaging to provide protection against oxygen. This can improve flavour stability. The effectiveness of bisulfite, besides its antioxidant properties, is also its ability to bind carbonyl compounds into flavour-neutral compounds (Barker et al., 1983). The reaction is reversible and excess bisulfite will increase yields of the adduct (Fig. 17.3).

Fig. 17.2 Sensory changes in beer flavour during aging.



533

Fig. 17.3 Binding of bisulfite to carbonyls.

Bisulfite addition to fresh beer minimises the increase of free aldehyde concentration during aging. In addition, when added to stale beer, bisulfite lowers the concentration of free aldehydes and affects the removal of the cardboard flavour. However, over time, the bisulfite will be oxidised to sulphate, thus increasing the concentration of free aldehydes again (Barker et al., 1983).

17.5

Foam stability

When beer is sold, the stability of the foam in a glass of beer is considered by many consumers to reflect the quality of the product. The increasing use of adjuncts (unmalted sources of carbohydrate) and the associated decrease in malt being used today, together with the employment of high gravity brewing techniques, have had a negative effect on foam values in many beers. There are many foam-promoting compounds in beer, such as iso--acids from hops, protein/polypeptides, metal ions, and polysaccharides, and all have an important role to play in foam formation and stability. However, the backbone of foam is protein. Many methods have been tried and extolled for their virtues in the isolation and characterisation of foam positive beer polypeptides, for example, separation of foaming proteins by hydrophobic interaction chromatography has long been a standard technique (Bamforth, 1999). Once separated, the proteins are investigated to discover whether they are related to beer foam potential or stability. On the basis of such experiments, it has been proposed that certain sizes of proteins are important in the formation and stabilisation of foam; for example, 40, 10 and 8 kD proteins have been postulated as major foam stabilising molecules (Lusk et al., 1995). However, it is now widely accepted that the polypeptides of greatest hydrophobic character produce the most stable foam and it is the hydrophobic property that is more important than size (Bamforth, 1985). The use of high gravity brewing techniques is essential for the present and future economic viability of the brewing industry internationally (Murray and Stewart, 1991). High gravity brewing is a procedure that employs wort at higher than normal concentration and therefore requires dilution with water later in the process. By reducing the water employed in the brewhouse, this process increases production capacity without adding to the existing brewing, fermenting and maturation plants. Therefore, most major brewing companies worldwide have revised their production processes to accommodate high gravity


534


Fig. 17.4

Change in the levels of hydrophobic peptides during the brewing process (final high gravity beer diluted to 4.5% alcohol by volume).

brewing procedures as a means to reduce capital expenditure and increase process sustainability. Although this process has many advantages (Stewart, 1999), one of the problems that still exists is that beers brewed at higher gravities exhibit poor foam stability (d'Amore et al., 1991). The effect of high gravity brewing on head retention with respect to hydrophobic polypeptide levels has been examined using phenyl sepharose liquid chromatography (Bamforth, 1985) throughout the brewing, fermentation and finishing of high and low gravity worts (Fig. 17.4) (Cooper et al., 1998a). Three notable features of the data are highlighted. At the kettle full stage (when `run-off' from the lauter tun or mash filter is complete and wort boiling begins), the level of hydrophobic polypeptides was similar in the high gravity and low gravity worts despite the use of twice the quantity of malt grist to produce the high gravity wort. (In this case, the high gravity wort was measured at 20ë Plato and the low gravity wort employed was measured at 10ë Plato. 1ë Plato is equivalent to 1 g of glucose dissolved in 100 mL of distilled water at 20 ëC.) This implies there was a major failure to extract hydrophobic polypeptides during the high gravity mash. There was much greater loss of hydrophobic polypeptides during fermentation of the high gravity wort so that by the end of fermentation, the hydrophobic polypeptide content of the high gravity fermented wort was just over 50 mg/L, markedly lower than that of the low gravity fermented wort (90 mg/L). When the high gravity beer was diluted to 4.5% alcohol by volume, equivalent to the low gravity beer, it contained a level of hydrophobic polypeptide less than 50% of the low gravity brewed beer (Cooper et al., 1998b). The head retention of the diluted high gravity brewed beer was less than that of the low gravity brewed beer. This contrasts with the low gravity brewed beer where the hydrophobic polypeptide in this foam accounted for over 40% of the total polypeptide. Therefore, not only is the polypeptide of the high gravity brewed beer reduced, but so is the hydrophobic content of its foam that would adversely influence its stability (Bamforth, 1995; Cooper et al., 1998b). The



535

amino acid profiles of hydrophobic polypeptides recovered from beer foam, unlike polypeptides involved in haze formation (where glutamic acid and proline account for 40±50% of the total amino acid composition) have no amino acids present in a distinctive quality (Leiper, 2002). It has already been discussed that the fermentation stage is a key stage where hydrophobic polypeptides are lost during the brewing process (Fig. 17.4). Two factors could account for the loss of hydrophobic polypeptides during fermentation. First, fermentation is known to be responsible for the loss of a large quantity of foam-active substances and this problem is exacerbated during the fermentation of high gravity worts. This loss is principally due to adsorption of foam onto the side of the fermenter. Second, yeast secretes proteolytic enzymes into the fermenting wort and these enzymes have a negative effect on foam stability of finished beer through protein degradation (hydrolysis) that occurs during fermentation and storage (Stewart, 2004). Analysis of proteinase A activity (using a fluorimetric method described by Kondo et al., 1998) in wort and beer during the brewing process (Fig. 17.5) showed, as would be expected, that freshly boiled wort did not contain enzyme activity. However, during fermentation proteinase A was secreted into wort by yeast cells. Proteinase A increased throughout fermentation with the highest enzyme activity occurring at the end. Considerably larger quantities of proteinase A were released during 20ë Plato fermentations compared to the 10ë Plato wort fermentations. During high gravity brewing, increased stress on the yeast, in the form of both elevated osmotic pressure and ethanol concentrations, stimulated the secretion of proteinase A into the wort during fermentation. In vitro studies in our laboratory have shown that both ethanol and increased osmotic pressure (simulated using sorbitol that is not metabolised by brewer's yeast) stimulated the secretion of proteinase by brewer's yeast strains (Brey et al., 2002).

Fig. 17.5

The effect of wort gravity on proteinase A release during fermentation of low (12ë Plato) and high (20ë Plato) gravity worts.


536

17.6


Gushing

Excess foam in a beer is regarded as deleterious and is known as gushing or `wild beer'. Gushing is the violent, uncontrolled ejection of beer from the package at the time it is opened and involves the loss of a significant portion of the contents. There may be two classes of gushing namely, sporadic and epidemic. Sporadic gushing may occur as a result of minor production deviations that are generally difficult to pinpoint. Epidemic or long-term serious gushing may be caused by several factors. Perhaps the most widely discussed cause is the use of weathered (damp) barley. If barley is harvested when wet, Fusarium or other fungus infection can develop during the malting process, resulting in beer susceptible to serious gushing. The formation of mycotoxins such as deoxynivalent (DON) has been paralleled with the development of gushing potential. The screening of barley and malt for these metabolites may offer a means of reducing beer gushing problems (Heikara, 1980). Other factors, such as increased levels of carbonation or a carbonating system operating without proper controls, can produce beers that have the potential to gush. Calcium oxalate microcrystals in beer are another cause of gushing. These crystals are thought to form nuclei for carbon dioxide gas emissions, but excess treatment and filtration will overcome this cause of gushing. Excessive levels of iron and other nuclei forming particles such as sediments will contribute to gushing problems.

17.7

Light stability

Beer is sensitive to light, especially in the 350±500 nm range. Light at these wavelengths can penetrate clear and green glass containers and cause a nauseous off-flavour in beers bottled in such glass containers and drinking glasses. The beer is said to be `sunstruck' and the aroma and taste referred to as `skunky'. Light instability in beer results from hop components. As already discussed, hops in brewing have a number of roles: they impart bitterness to beer; provide characteristic hop aromas; suppress growth of certain microorganisms, particularly Gram-positive bacteria; assist in beer foam stability and contribute polyphenols to the protein-polyphenol complex during wort boiling. When beer is exposed to light, one of the side chains on the iso--acid (a component of hops) is cleaved and the highly reactive radical that is liberated combines with sulphur-containing compounds (Fig. 17.6) to produce 3-methyl2-butene-1-thiol (MBT). MBT has a skunky-like aroma. MBT has a flavour threshold in the order of parts per trillion, making it one of the most flavouractive substances in beer (Wilson et al., 2001). Specialised hop extracts (produced using liquid CO2 or ethanol as a solvent) have been developed to combat this sensitivity to light (Wilson et al., 2001). In essence, pairs of hydrogen atoms are catalytically added to the isomerised iso-acid, tetrahydro-iso--acid. There are three principal types of such extracts



Fig. 17.6

537

Mercaptan formation in light-sensitive iso--acid.

(called reduced extracts) currently available on the market: RHO iso--acid, tetrahydro-iso--acid, and hexahydro-iso--acid. All of these materials are bitter to varying degrees, some improve beer-foam cling and stability and protect beer against light-struck-sun-induced skunky flavours. Normally all of these materials are used as a post-fermentation addition to achieve maximum benefit and optimum utilisation. To achieve complete light-strike protection, no iso--acid can be used in any other part of the process. Even repitched yeast (Lusk et al., 1995) with iso--acid absorbed onto its surface will provide sufficient material for photolytic cleavage to occur and the resultant production of MBT and skunky flavours.

17.8

Conclusions

Beer instability involves a number of complex reactions involving proteins, carbohydrates, polyhphenols, metal ions, thiols and carbonyls. There are many diverse types of beer instability involving a number of different microorganisms, chemical species and reactions. Our understanding of these reactions has progressed over the past 25 years, but we are far from a complete comprehension of beer instability reaction systems.

17.9

References

(1985), The foaming properties of beer, J. Inst Brew., 93, 216±219. (1995), Foam: method, myth or magic?, The Brewer, 81, 396±389. BAMFORTH, C.W. (1999), Beer haze, J. Amer. Soc. Brew. Chem., 57, 81±90. BAMFORTH, C.W. (2004), A critical control point analysis of flavour stability of beer, Tech. Quart. Master Brew. Assoc. Amer., 41, 97±103. BARKER, R.L., GRACEY, D.E.F., IRWIN, A.J., PIPASTS, P. and LEISKA, E. (1983), Liberation of staling aldehydes during storage of beer, J. Inst. Brew., 89, 411±415. BEHR, J. and VOGEL R.F. (2009), Mechanisms of hop inhibition: hop ionophores, J. Agric Food Chem., 57, 74±81. BAMFORTH, C.W. BAMFORTH, C.W.


538


and STEWART, G.G. (2002), The loss of hydrophobic polypeptides during fermentation and conditioning of high gravity and low gravity brewed beer, J. Inst. Brew., 108, 424±433. BRIGHT, D. (2001), A study of the antioxidant potential of speciality malt. PhD Thesis, Heriot-Watt University, Edinburgh, Scotland. COOPER, D.J., STEWART, G.G. and BRYCE, J.H. (1998a), Hydrophobic polypeptide extraction during high gravity mashing ± experimental approaches for its improvement, J. Inst. Brew., 104, 283±287. COOPER, D.J., STEWART, G.G. and BRYCE, J.H. (1998b), Some reasons why high gravity brewing has a negative effect on head retention, J. Inst. Brew., 104, 221±228. D'AMORE, T., RUSSELL, I. and STEWART, G.G. (1991), Advances in the fermentation of high gravity wort. Proceedings of the European Brewing Convention Congress, Lisbon, pp. 337±344. DALGLEISH, C. (1997), Flavour stability. Proceedings of the European Convention Congress, Amsterdam, pp. 623±659. GILLILAND, B. (1971), Yeast classification, J. Inst. Brew, 77, 276±284. HEIKARA, A. (1980), Gushing induced by fungi. European Brewing Convention Monograph VI, pp. 251±258. IIJIMA, K., SUZUKI, K., OZAKI, K. and YAMASHITA, H. (2006), HorC confers beer-spoilage ability on hop-sensitive Lactobacillus brevis ABBC45cc, J. Appl. Microbiol., 100, 1282±1288. IRWIN, A.J., BARKER, R.L. and PIPASTS, P. (1991), The role of copper, oxygen and polyphenols in beer flavour instability, J. Amer. Soc. Brew. Chem., 49, 140± 149. JACKSON, G., ROBERTS, R.T. and WAINWRIGHT, T. (1980), Mechanism of beer foam stabilization by propylene glycol alginate, J. Inst. Brew., 86, 34±37. KONDO, H., YOMO, H., FURUKUBO, S., FUKUI, N., KAWASAKI, Y. and NAKATANI, K. (1998), Advanced method for measuring proteinase A in beer. Proceedings of the 25th Convention, The Inst. of Brewing, Asia Pacific Section, pp. 119±124. LEIPER, K. (2002), Beer polypeptides and their selective removal with silica gels. PhD Thesis, Heriot-Watt University, Edinburgh, Scotland. LUSK, L.T., GOLDSTEIN, H. and RYDER, D. (1995), Independent role of beer proteins, melanoidins and polysaccharides in foam formation, J. Amer. Soc. Brew. Chem., 53, 93±103. MCMURROUGH, I. and O'ROURKE, T. (1997), New insight into the mechanism of achieving colloidal stability, Tech, Quart. Master. Brew. Assoc. Amer., 34, 271±277. MURRAY, C.R. and STEWART, G.G. (1991), Experience with high gravity lager brewing, Birra et Malto, 44, 52±64. NARZISS, L., MIEDANER, H., GRAF, H., EICHFORN, P. and LUSTIG, G. (1993), Technological approach to improve flavour stability, Tech. Quart. Master Brew. Assoc. Amer., 30, 48±53. PRIEST, F.G. and CAMPBELL, I. (2003), Brewing Microbiology. Academic Press, New York. SAKAMOTO, K., MARGOLLES, A., VAN VEEN, H.W. and KONINGS, W.N. (2001), Hop resistance in the beer spoilage bacterium Lactobacillus brevis is mediated by the ATP-binding cassette multidrug transporter HorA, J. Bact., 183, 5371±5373. SIEBERT, K.J. and LYNN, P.Y. (1997), Mechanisms of beer colloidal stabilization, J. Amer. Soc. Brew. Chem., 55, 73±78. SIMPSON, W.J. (1993), Cambridge prize lecture. Studies on the sensitivity of lactic acid bacteria to hop bitter acids, J. Inst. Brew., 99, 405±411. BREY, S.E., BRYCE, J.H.



539

and HAMMOND, J.R.M. (1989), The response of brewing yeast to acid washing, J. Inst. Brew., 95, 347±354. STEWART, G.G. (1999), High gravity brewing, Brew. Guard., 128, 31±37. STEWART, G.G. (2004), The chemistry of beer instability, J. Chem. Edu., 81, 963±968. WILSON, R.J.H., ROBERTS, I., SMITH, R.J. and BIENG, M. (2001), Improving hop utilization and flavour control through the use of pre-isomerized products in the brewery. Tech. Quart. Master Brew. Assoc. Amer., 38, 11±21. SIMPSON, W.J.


18 Shelf life of wine R. S. Jackson, Brock University, Canada

Abstract: What constitutes shelf life in wine is difficult to define. It is often based on consumer or critic expectation, not objective criteria. Initially, changes are generally beneficial, resulting in a reduction in bitterness, astringency, and a loss of yeasty odors. Subsequent modification usually results in a diminution of fruity aromas. Wines with a longer shelf life typically possess a distinctive varietal aroma and develop an aged bouquet. However, in the absence of these features, or with the development of oxidized and other unpleasant odors, the shelf life of wine may be measured in terms of several months to a few years. The latter may develop as a consequence of failures in the bottle closure, exposure to sunlight, the presence of high temperatures, environmental contaminants, or microbial spoilage. Key words: cork closure, wine aging, wine shelf life, wine off-odors, wine oxidation.

18.1

Introduction

Wine differs from most foods and beverages in not possessing a `best before' date. This reflects ambiguity as to what distinguishes wine deterioration from the maintenance or improvement of its sensory attributes. Opinion varies markedly between wine authorities and consumers, usually based as much on experience as on preference. This situation indicates inconsistency in how wine quality is perceived and rated. It also mirrors the relative importance placed on wine retaining its initial, youthful attributes, versus attributes that develop upon extended aging. The relative weight given these factors can also vary depending on the type of wine. For example, most white wines and roseÂs are preferred with the presence of a fruity fermentation bouquet, combined with a mild varietal


Shelf life of wine 541 fragrance. In contrast, most red and some white wines are often preferred when the varietal fragrance is more marked and supplemented with oak flavors. The eventual disappearance of varietal and fermentation fragrances in older wines can be accepted (even desired) if replaced by a refined aged bouquet. It is this complex, subtle flavor that provides much of the intrigue and exclusivity to mature, premium vintage wine. Although aging potential is highly regarded by afficionados, there are few objective criteria by which its development can be predicted. It is far easier to explain shelf life loss than its retention. What is considered an acceptable shelf life, and when the wine reaches its optimum, also depends on the production process. For example, beaujolais nouveau (made via carbonic maceration) is considered to be optimal just after production. It maintains its character for several months, but often deteriorates quickly thereafter. In contrast, premium barolos or bordeaux often require many years before developing their desired and traditional characteristics. However, this also depends on the expectations and preferences of the taster. British experts usually prefer well-aged, mellow versions, whereas French commentators often express a favor of the young, austere, astringent version. Similar differences often apply to other wines. Older champagnes are (or were) much appreciated in England, but depreciated in France. Fino sherries have short shelf lives upon bottling (if one prefers its initial attributes), whereas oloroso sherries can retain their character for decades, and for months after being opened. Thus, what is an appropriate or acceptable shelf life for one wine may be unacceptable for another. It is more typical to refer to the shelf life of wines designed to be consumed young than for those produced with long aging in mind. In the former, the desired properties are the retention of the sensory attributes expressed at bottling. These typically refer to a fresh fruity to floral fragrance, with a nonaggressive flavor. In contrast, premium wines often possess attributes unpleasant at bottling, at least to the novice drinker. Prolonged storage is expected to result in an enhancement of sensory quality, and the development of a complex, rich flavor and bouquet. In addition, flavors considered a fault in the majority of wines may be accepted, or even highly regarded in a premium wine. Thus, no precise set of chemical or sensory attributes consistently define wine shelf life. Wine is often considered a natural and artisanal creation of sun and soil, not a standardized commercial product. Another feature distinguishing wine from other beverages is the participation of consumers in its maturation. The shelf life of most products relates to their acceptability at the time of purchase or shortly thereafter. This applies equally to most wines, but to a select number of purchasers, shelf life reflects the wine's aging potential ± how long the wine can be stored before it passes its `peak' (in reality, an extended but changing plateau), before losing its desirable sensory attributes.


542

18.2


Factors affecting wine stability and shelf life

18.2.1 Prior to bottling Viticultural conditions One of the current credos of modern wine making is that quality originates in the vineyard. It is the rationale given for emphasis on the wine's provenance (terroir). While partially true, the marketing advantage of the terroir concept is even more significant. That aside, the quality of grapes arriving at the winery door does set outer limits on the attributes the wine may possess, including its potential shelf life. However, this topic is too large for detailed discussion in this chapter. For details, the reader is directed to sources such as Jackson (2008) and White (2003). Some generalities are worth noting, however. Warm climatic regions have usually been considered to produce wines unsuitable for long shelf life, especially white versions. Warm climates tend to result in grapes being harvested with low acid levels. Low acidity not only favors the oxidation of wine phenolics (favoring the presence of the more readily oxidated phenolate state), but also promotes the breakdown of fruit esters. The latter donate most of the fruity fragrance of young white wines. pH values above 3.5 also favor the growth of spoilage lactic acid bacteria. Their action not only further reduces wine acidity (producing of a `flat' taste), but also results in the generation of off-flavors. Warm conditions may also promote the loss of desired and distinctive varietal fragrance compounds in grapes, and the development of high sugar contents. High sugar contents increase the likelihood of stuck fermentation (their premature termination) and the presence of high residual sugar contents. The latter give dry table wines an undesired sweet aspect. It also makes the wine much more liable to microbial spoilage. Wine produced in cool climate conditions seldom experience these problems, and historically have had longer shelf lives. Modern viticultural practices can help reduce these limitations, before and during harvest, and advanced wine-making procedures can further subdue the historic disadvantages of warm climate conditions. To offset the development of low acid/high pH juice, grapes can be picked earlier in the season, before grape metabolism consumes its malic acid content. Earlier harvesting also helps limit the development of undesirably elevated sugar contents, as well as the degradation of flavorants in the skins. Harvesting early in the morning can result in grapes being cooler when reaching the winery. Alternatively, the grapes may be cooled on reaching the winery. Fermenting the juice cool favors the production and retention of higher amounts of fruit (acetate) esters. In addition to macroclimatic factors, seasonal variations and microclimatic conditions can significantly modify the potential characteristics and shelf life of a wine. These are well recognized, and often used to advantage, when favorable, in marketing wine from a particular region. Some factors such as drainage conditions, susceptibility to frost, and slope of the land cannot be easily modified. However, some negative features can be modified without excessive difficulty. Disease control is a major example. Healthy grapes are far more


Shelf life of wine 543 likely to produce wines with longer shelf lives than diseased grapes. The latter may be contaminated with laccase, a polyphenol oxidase with a much wider range of phenolic substrates than grape polyphenol oxidase. In addition, infected grapes are much more susceptible to the formation of light-diffracting, brownish tannin colloids. Infected grapes also tend to possess higher concentrations of soluble proteins (thaumatin-like and chitinases). These can generate a protein haze in bottled wines. More intractable are the disturbed sugar, acid and flavor constituents of diseased grapes. The major exception to the undesirability of infection involves `noble rotted' white grapes, given very slow pressing to avoid the release of fungal mucopolysaccharides. Such Botrytis-infected grapes, under special late-season conditions, can generate the most precious, and expensive, white wines, for example sauternes and trockenbeerenausleses. Training and pruning choices also affect optimal grape ripening, as do fertilization and irrigation decisions. An example of how shelf life can be adversely affected by growing conditions involves the development of an ùntypical aged' (UTA) off-odor, about a year after bottling. It is variously considered to be reminiscent of naphthalene, furniture polish, or wet wool. It appears to be correlated with grapes having suffered stress during the growing season (Sponholz and HuÈhn, 1996). Minimizing undue (usually drought) stress is the principal means by which the development of this fault can be avoided. In addition, there is evidence that the incidence of UTA can be reduced by the addition of thiamine and diammonium phosphate at fermentation, as well as the addition of ascorbic acid (150 mg) (Rauhut et al., 2001). Modern technology, in combination with GPS, has led to a new term in viticulture: `precision viticulture'. It is beginning to permit the grape grower to understand and regulate one of the major limiting factors in grape quality ± vineyard variability, and thus wine shelf life. Variable conditions throughout the vineyard are reflected in non-uniform grape maturity ± inclusion of unripe and overripe grapes reduce overall wine quality. Understanding the origins of vineyard variability offers the opportunity for selective adjustment of viticultural practice for particular regions of a vineyard, or selective timing of harvest throughout the vineyard to provide fruit of more uniform quality. Vinification conditions As with viticultural conditions, fermentation procedures can significantly affect shelf life. For example, it has become standard practice for white wines to minimize oxygen exposure after crushing, during fermentation, and maturation. Red wines may be permitted limited air exposure during fermentation, usually associated with pumping over (mixing of the grape seeds and skins with the fermenting juice). Pumping over helps equilibrate temperature differences throughout the fermenting must and extract anthocyanins. The oxygen absorbed can assist fermentation. Pumping over is unnecessary for white wines due to the absence of seeds and skins during fermentation. The absence of oxygen tends to favor the production of acetate (fruity smelling) esters. These are especially important to the flavor and shelf life of


544


white wines with limited or no distinctive varietal aroma. Cool fermentation temperatures also favor the formation of acetate esters and their retention. Correspondingly, most white wines are fermented at cool temperatures. In addition, oxygen favors the synthesis of C6 alcohols and aldehydes, which can donate herbaceous off-odors, as well as oxidize varietal aromatics. The production and retention of acetate esters, and reduced generation of C6 alcohols and aldehydes, can be further accentuated by the addition of antioxidants such as sulfur dioxide and ascorbic acid (Moio et al., 2004), the judicious selection of yeast strain (high ester producers) and maturation at cool temperatures. Nonetheless, with some cultivars, limited oxygen access during fermentation has been associated with the development of more complex flavors that develop in the bottle. Thus, the style and design of the wine can influence the procedures used during vinification. Wine intended for immediate enjoyment (but limited shelf life) accentuate development of a fruity flavor, whereas alternative procedures are used when a more mature bouquet is desired (providing and benefiting from prolonged storage). Where extended shelf life is desired for a white wine, the juice are often left in contact with the crushed grapes, rather than rapid separation from the seeds and skins upon crushing. This skin contact (maceration) period permits the increased extraction of distinctive (varietal) aromatics from the skins (their predominant location). Coincident with skin contact is also the increased uptake of phenolic compounds from the skins (those in the seeds are rarely extracted due to the maceration seldom lasting more than 24 hours). This practice must be used judiciously, however, to avoid the excessive uptake of flavonoid phenolics. These can result both in the wine taking on a bitterish aspect (undesired in most white wines), and enhancing the likelihood of early browning (Simpson, 1982). To minimize the latter possibility, juice exposed to skin contact is often given slight aeration. Early oxidation of flavonoids favors their precipitation during fermentation. Gentle pressing of the crushed grapes (as with pneumatic presses) further limits the uptake of flavonoid phenolics. A technique increasingly used in the production of premium white wines is sur lies maturation. This involves leaving wine in contact with the lees (dead and dying yeast cells) for several months at the end of fermentation. Typically sur lies maturation occurs in small cooperage (~250 l), to avoid the generation of hydrogen sulfide and reduced sulfur odors (notably mercaptans). These can easily form in the thick layer of yeast cells that develop in large-volume cooperage. The treatment has several benefits in terms of shelf life. Primarily, it reduces the wine's browning potential. This appears to involve the interaction between membrane sterols, released by autolysing yeast cells, and precursors of brown phenolics (notably catechins and their dimers) (MaÂrquez et al., 2009). The procedure can also add aromatic yeast metabolites, such as ethyl octanoate and ethyl decanoate. These ethyl esters could extend shelf life by augmenting the fruity aspect of the wine. Lees may also reduce the sensory defects generated by 4-ethylphenols (Chassagne et al., 2005), and the sensory impact of carbonyl compounds such as diacetyl. Exposure to yeast lees also diminishes the


Shelf life of wine 545 production of sotolon (possessing a curry-like odor) during bottle aging (Lavigne et al., 2008). The benefit/detriment value of the latter depends on the type of wine and the expectations of the consumer. Most wines designed for early consumption (short shelf life) are matured in inert containers (often stainless steel) prior to bottling. Maturation in oak cooperage is normally reserved for wines designed for medium- to long-bottle aging. This is frequently viewed as contributing to the aromatic complexity of the wine. Although true, it can also extend shelf life by supplying small quantities of oxygen (stabilizing a red wine's color) and increasing the concentration of ethyl esters (Salinas et al., 1996). Red wines are considerably less susceptible to oxidation than white wines. This is partially accounted for by their higher phenolic concentration. oDiphenols are particularly active in consuming oxygen and protecting other wine aromatics from oxidation. In addition, the generation of acetaldehyde (a consequence of phenol oxidation) helps stabilize the wine's color, by forming oxidation resistant anthocyanin-tannin polymers. With most red wines, fermentation and maturation procedures are aimed at producing wine that possesses a shelf life of up to five years. This usually means that the wine will show a fresh berry to jammy fragrance, combined with any distinctive aroma the grape cultivar may possess. For wines designed for extended aging (shelf life greater than 5±10 years), production procedures are adjusted to give the wine an enhanced phenolic content. Although desirable in promoting long aging potential, it also demands patience on the part of the consumer. It takes many years for the wine to develop its optimal attributes. In contrast, when the intent is to generate a red wine that is drinkable almost immediately, carbonic maceration may be employed. One of the principal drawbacks of carbonic maceration is that it also results in a wine with a comparatively short shelf life (rarely more than a few years). To give most red wines their deep red color, the juice is fermented in contact with the seeds and skins. The anthocyanins that provide the red color are almost exclusively found in the grape skins. Coincident with the extraction of anthocyanins are other phenolic compounds. The latter donate red wine's typical bitter, astringent character. They are also crucial to giving the wine most of its antioxidant potential (and extended shelf life). These phenolics (mostly flavonoids) are also directly involved in stabilizing the wine's color, and in the gradual shift from the purplish cast of a young wine to the brickish shade of a mature red wine (Fig. 18.1). When these changes are gradual, they are an accepted and expected aspect of aging. If they occur early, and are associated with an oxidized flavor, they are a fault and severely shorten the wine's shelf life. The shelf life of all wines is enhanced by a desirable acid content (and low pH). Where juice acidity is undesirably low, it can be adjusted upwards by the addition of tartaric acid. Tartaric acid is preferred because it is a natural grape constituent, but primarily due to its being poorly metabolized by most bacteria (thereby reducing the likelihood of microbial spoilage).


546


Fig. 18.1 The shift in color of red wines during aging as measured by a change in the absorption spectra: I, 1-year-old; II, 10-years-old; III, 50-years-old. (From RibeÂreauGayon `Shelf-life of wine', pp. 745±772, in Handbook of Food and Beverage Stability: Chemical, Biochemical, Microbiological and Nutritional Aspects (G. Charalambous, ed.). Copyright Elsevier (1986), reproduced by permission).

18.2.2 Storage conditions of bottled wine Oxygen exposure It is essential to minimize oxygen uptake to maximize shelf life. Despite all precautions, sufficient oxygen eventually enters to overpower the wine's antioxidant potential, inducing irreversible sensory degradation. Currently, it is impossible to predict how much, or when, oxidation will reach a level sufficient to negatively impact shelf life. It depends on the amounts of natural or added antioxidants, the rate and extent of oxygen ingress, the presence of spoilage microbes, the sensitivity of impact aromatics to oxidation, and the sensory threshold of oxidative deterioration. What is feasible is limiting the rate and degree of oxygen ingress, and adjusting wine conditions to either restrict the activation of oxygen or favor the consumption of oxygen in sensorially neutral or beneficial ways (for example, mollifying a red wine's astringency). Molecular oxygen is itself relatively stable under wine conditions. To be involved in oxidation, it usually must be converted to a more active form. These may include hydrogen peroxide, singlet oxygen, superoxide, or the hydroxyl radical. Their formation is facilitated primarily by trace amounts of iron and copper ions, and to a lesser extent, by light exposure. The latter is thought to involve the interaction of compounds such as riboflavin. The best understood oxidative reactions in wine involve phenolics, notably odiphenols. The oxidation of flavonoid phenolics, notably catechins (and their polymers) to quinones, changes their chromic properties, generating pigments with a yellowish to brownish color. These may also react with other wine constituents, such as acetaldehyde, that favor polymerization between flavonols, or between flavonols and anthocyanins. The latter tend both to stabilize the color of red wines and generate the brickish shift associated with bottle aging. The color shift may also involve the polymerization of oxidized tartaric acid


Shelf life of wine 547 (glyoxylic acid) with flavonoid phenolics, to generate orangish to yellow byproducts, notably xanthylium derivatives (Oszmianski et al., 1996). In white wines, the oxidative generation of brownish pigments is largely thought to involve nonflavonoids, such as caftaric and fertaric acids, flavonoids such as quercetin and kaempherol, and small amounts of catechins. Some additional browning products may arise from the oxidation of galacturonic acid, and the formation of Maillard products from residual sugars. The presence of limited amounts of oxygen can also favor the degradation of the fragrance generated by fruit esters and various terpenes. In addition, less desirable fragrances, derived from aldehydes, diethylesters, heterocycles and thiols, may accumulate. Temperature Storage temperature is possibly second only to oxygen as a factor unfavorably affecting shelf life. Traditionally, wine has been stored under relatively constant, cool conditions. Temperatures above 25 ëC are generally viewed as not only speeding maturation, but also generating undesirable flavor changes (Singleton, 1962). Because aging is primarily physicochemical, heat can both activate and speed the reactions involved. However, different reactions possess distinct activation energies. Thus, a temperature change does not affect all reactions equivalently (Table 18.1). Cool storage tends to retain the fresh, fruity character of most young wines. For example, the concentrations of fragrant acetate esters, such as isoamyl and hexyl acetates are stable at 0 ëC, whereas they rapidly hydrolyze at 30 ëC (Fig. 18.2). In contrast, the formation of less aromatic ethyl esters is rapid at 30 ëC, but negligible at 0 ëC. Temperature also has marked influences on the liberation of norisoprenoid aromatics from their glycosidic precursors (Leino et al., 1993). This may account for some of the increased concentration of trimethyl-1,5dihydronaphthalene (TDN) in `Riesling' wine aged at 30 versus 15 ëC (Marais et al., 1992), as well as the content and types of monoterpene alcohols found in some wines (Rapp and GuÈntert, 1986). High temperatures also favor the degradation of sugars to furfurals and pyrroles. Whether similar activation affects the conversion of norisoprenoid precursors to spiroesters, such as vitispirane and theaspirane, or to hydrocarbons such as TDN and ionene, is unknown. For most wines, exposure to temperatures 40 ëC and above prompts rapid quality deterioration. Carbohydrates in the wine undergo Maillard and thermal degradation reactions, turning brown and producing a baked (maderized) flavor. The wine also tends to develop a sediment. Even temperatures about 30 ëC produce evident losses in fragrance within a few months. From the limited data available, it appears that traditional cellar temperatures (about 10 ëC) permit the prolonged retention of most fruit esters, while not excessively inhibiting other desired aging reactions. Nevertheless, temperatures up to 20 ëC do not appear inimical to the sensory modifications associated with aging, at least for red wines. Some of the changes that accrue under different storage temperatures are illustrated in Fig. 18.2.


548


Table 18.1 Differences in aroma composition of several Riesling wines as affected by aging and storage temperature 1982

1978

1973

Acetate esters i-amyl acetate i-butyl acetate 2-phenylethyl acetate

107a 16.7 38.7

58.4 4.2 25.1

Ethyl esters butanoic acid ester hexanoic acid ester succinic mono-acid ester

15 47 128

Diethyl esters succinic acid ester malic acid ester

41 96

Carbohydrate degradation products 2-furfural 4.1 furan-2-carbonic acid ethyl 0.4 ester 2-formylpyrrole ± 5-hydroxymethylfurfural (HMF) ± Monoterpenes linalool -terpineol 3,7-dimethyl-1,5-octadien3,7-diol

16.8 8.4 33.3

Year

1964

1976 (frozen)

1976 (cellar storage)

5.9 2.8 1.9

10.9 3.1 5.7

243 32.1 27.2

27.1 6.0 3.2

19 72 338

16 47 438

30 77 415

40 48 152

41 63 339

384 640

656 1375

738 969

117 262

407 729

13.9 0.6

39.1 2.4

44.6 2.8

2.2 0.7

27.1 2.0

2.4 ±

7.5 1.0

5.2 2.2

0.4 ±

1.9 0.5

1.0 3.2 12.6

± 7.0 9.2

± 8.3 15.1

19.4 10.8 28.6

2.8 16.6 28.3

Data from Rapp and GuÈntert (1986). a Relative peak height on gas chromatogram (MM).

While cool temperatures are normally viewed as promoting the slow, favorable maturation of wine, storage under cold conditions can promote the precipitation of tartrate and other salt crystals. Occasionally, consumers may mistake potassium tartrate crystals as glass slivers. Because of this potential misinterpretation, wines are normally cold stabilized before bottling to minimize, though not entirely eliminate, their development. In addition to affecting the rate and direction of wine development, rapid and marked temperature fluctuations can sufficiently affect wine volume to loosen the cork seal. It can lead to wine seepage, oxygen ingress, and wine spoilage by permitting the reactivation of dormant microbes in the wine. If the wine freezes, the volume increase can be sufficient to force the cork out of the bottle.


ß Woodhead Publishing Limited, 2011 Fig. 18.2 Effect of storage temperature and duration on the concentrations of (a) hexyl acetate, (b) i-amyl acetate, (c) diethyl succinate, and (d) dimethyl sulfide in a Colombard wine. (From Rapp and Marais `The shelf life of wine: changes in aroma substances during storage and ageing of white wines', pp. 891±921, in Shelf Life Studies of Foods and Beverages (G. Charalambous, ed.). Copyright Elsevier (1993), reproduced by permission).

550


Light exposure Light exposure may have two, independent, detrimental effects on wine. Near ultra-violet and blue radiation can activate oxidative reactions, whereas direct exposure to sunlight may provoke heat-induced damage. Most light-activated reactions are thought to involve the production of potent oxidants, notably singlet oxygen. Its production appears to involve the activation of an electron on a pigment (presumably riboflavin) and its association with molecular oxygen. Examples of light-associated problems include the formation of a copper-induced haze and the shrimp to skunky thiol off-odor of `lightstruck' champagne (Carpentier and Maujean, 1981). Additional off-odors, unrelated to thiols, associated with champagne exposed to light have been noted by D'Auria et al., (2003). Light exposure, notably full sunlight, can cause rapid changes in wine temperature. The increased temperature can both promote undesirable reactions, generating off-odors, and produce pressure on the bottle closure. Repeated and rapid shifts in temperature (and wine volume) weaken the adherence of the closure to the bottle neck, leading to wine seepage and oxygen ingress. The best protection against light-induced spoilage is keeping the wine in lowlit conditions or in the dark. Alternately, bottling in UV and blue absorbing glass can limit photo-induced shortening of shelf life. pH and acidity Relatively low pH values are preferred for several reasons. They provide wines with a fresh taste, promote microbial stability, reduce browning potential, and diminish the need for SO2 addition. Regrettably, it also favors the breakdown of acetate (fruit) esters to their constituent moieties (an alcohol and acetic acid) (Marais, 1978). The monoterpene content may also be affected. For example, the concentration of geraniol, citronellol, and nerol may rise, whereas those of linalool, -terpineol, and hotrienol decline at low pH values. In red wines, color intensity and hue are enriched at lower pH values. Low pH also minimizes the concentration of the more readily oxidizable, phenolate form of phenolics. For example, there are nine times more phenolate ions at a pH of 4.0 than a pH of 3.0. This also helps to explain why white wines are less susceptible to browning at low pH values. Vibration It is commonly thought that vibration is detrimental to wine shelf life. Other than the resuspension of sediment in older wines, induced by agitation during serving, there is no evidence that vibration itself is detrimental. The one research paper on the issue notes some minor physicochemical changes that may occur with marked and prolonged vibration (Chung et al., 2008). However, even those noted seem unlikely to be of sensory significance.


Shelf life of wine 551 Environmental contaminates Contaminants affecting wine shelf life can arise from a host of sources, but most can be grouped as originating from winery equipment or additives, bottle closures, or the storage environment. Probably the principal reason given for wine rejection results from the presence of corky odors. These may be generated by a range of compounds, but the most common appears to be 2,4,6-trichloroanisole (TCA) (Sefton and Simpson, 2005). It typically diffuses into wine from cork exposed to the pesticide pentachlorophenol at some stage during bark growth, stopper production, or storage. Several microbes metabolize the pesticide into TCA (a nontoxic but moldy-smelling compound at trace amounts). It may also be absorbed from winery cooperage or other wooden structures treated with the insecticide. Another source of corky odors, seemingly arising from winery cooperage, is 2,4,6tribromoanisole (Chatonnet et al., 2004). In this case it is the by-product of the microbial transformation of a fire-retardant and wood preservative, 2,4,6-tribromophenol. Another moldy odor noted as a frequent contaminant is 2-methoxy3,5-dimethylpyrazine (Simpson et al., 2004). Other moldy odors, potentially provoking consumer rejection, are 1-octen-3-one, 2-methylisoborneol, guaiacol, 1-octen-3-ol, and geosmin. These are all microbial by-products generated on cork or winery equipment. The shift to artificial corks and screw caps has markedly reduced, but not eliminated, their incidence in wine. Synthetic corks and screw caps stored in an environment contaminated by volatile microbial by-products may absorb these compounds and subsequently release them into wine. Winery equipment is also the predominant source of metallic contamination. Alternative origins include agents used in fining, notably bentonite. As noted later in the chapter, metal ions participate in the induction of several forms of casse (haze) in bottled wine.

18.3

Changes during the shelf life of wine

The chemical changes that commonly and negatively affect shelf life are those associated with oxidation and the hydrolysis of esters. Those involved with reduction, polymerization, structural rearrangement, and volatilization are more likely to be initially beneficial. Their relative importance depends on the type of wine, its production, its varietal origin, storage conditions, and the expectations of the consumer. Initially, most modifications enhance wine sensory characteristics. Subsequently, changes in the varietal aroma, fermentation bouquet, and color progressively modify the wine's original attributes. These alterations may or may not be viewed as desirable. Eventually, varietal aroma, fermentation bouquet and color begin to fade. If these sensory attributes are not replaced by nuances, termed the aged bouquet, the wine's shelf life will be comparatively short (1±5 years). However, if the aged bouquet develops and is appreciated, it supplies the wine with long-aging potential. Most premium wines possess this feature, supplying a shelf life that may extend beyond 20 years.


552


18.3.1 Appearance All wines change color with age, becoming more brownish (white wine becoming darker, whereas red wines become lighter). The relative acceptability of these changes depends on how rapidly they develop and if they are normally expected. The most studied series of reactions involve select flavonoids. Their oxidation results in the formation of o-diquinones and hydrogen peroxide. The o-diquinones can polymerize with other phenolics, some of which can rearrange to form oligomeric o-diphenols. These, in turn, can oxidize and polymerize in a similar series of reactions forming larger polymers. Because quinones absorb light non-uniformly, tending to be brown colored, successive oxidations explain the expression `brown begets brown' (Singleton, 1987). The hydrogen peroxide generated during o-diphenol oxidation can activate the oxidation of other constituents, notably ethanol to acetaldehyde (Timberlake and Bridle, 1976), and presumably glycerol to glyceraldehyde and dihydroxyacetone (Laurie and Waterhouse, 2006). In addition, hydrogen peroxide may be consumed in Fentonlike reactions with phenolics (Walling and Johnson, 1975), involving the catalytic activity of iron and possibly copper, or reactions leading to phenolic degradation, such as the conversion of gallic acid to muconic acid derivatives (Singleton, 1987). It is in these secondary oxidation reactions that oxygen is incorporated into organic wine constituents. Color changes in white wines The most common and undesirable color change in white wines is termed premature browning. It is normally measured as a noticeable increase in absorption at 420 nm. The exact chemical nature of the, presumably multiple, chromophores involved still remains unclear. Most are thought to be derived from oxidation of the wine's limited phenolic content. The most prevalent of these is caftaric acid, a common constituent of grape juice. After hydrolysis, caffeic acid can significantly increase oxidative browning by polymerizing with flavonoids. Because the concentration and precise chemical nature of a wine's phenolic content vary from cultivar to cultivar, from year to year, vinification procedures, as well as storage conditions, predicting the occurrence of browning has proven intractable. Because molecular oxygen is thought to be essential for premature browning, its erratic appearance among samples of the same wine has often been ascribed to closure problems. Although white wines have a lower content of oxidizable phenolic compounds, they are more susceptible to oxidative browning than red wines. This apparent anomaly results from the ability of most flavonoid phenolics (more common in red wines) to consume large amounts of oxygen without noticeable browning, retarding the undesirable sensory consequences of oxidation. In addition, the paler color and milder flavor of most white wines make the consequences of oxidation evident sooner and of greater sensory significance. Besides oxidized phenolics, additional sources of colored material include phenolics that have a yellowish color in their natural (non-oxidized) state, notably kaempferol and quercetin. These are not considered a fault unless they


Shelf life of wine 553 occur at abnormally high concentrations. For example, quercetin may occasionally induce a yellowish flavonol haze due to the precipitation of its fine crystals (Somers and Ziemelis, 1985). Another source of yellowish to golden coloration in white wines comes from the formation of Maillard products. In wine, this usually results from a reaction between residual sugars and organic bases, such as amines, amino acids, or free amino acid groups on proteins. The initial step involves the dehydration of sugars to deoxysones. These are more electrophilic than their parent sugars and react with nitrogen bases to generate various chromophores. In addition to generating a diverse range of pigmented compounds, some Maillard by-products are aromatic. Furaldehydes, with caramel-like flavors, are a classic example. At cellar and room temperatures, Maillard reactions occur slowly, but can eventually add to the admired golden coloration of aged, sweet, white wines. In baked fortified wines, notably madeiras, Maillard reaction products contribute significantly to the basic color and flavor of these wines. Another potential source of color change comes from metal-induced structural modifications of galacturonic acid (derived from pectin breakdown). These age-induced color changes are considered normal and often highly regarded by wine connoisseurs. The development of a slightly pink cast in white wines is an infrequent problem with wines produced from cultivars such as Sauvignon blanc, Pinot gris, and AlbaÄrino. It can usually be corrected, but at considerable expense, by opening the effected bottles, treating with polyvinylpolypyrrolidone (PVPP), and rebottling. It appears to be associated with exposure of the wine to oxygen. Its exact chemical nature is unknown but is suspected to be derived from flavan3,4-diols (leucoanthocyanins). These slowly dehydrate to flavenes under reducing conditions. Upon exposure to oxygen they can readily oxidize to their corresponding colored flavylium state. Color changes in red wines The color changes in red wines are much more chemically complex than those found in white wines (Fulcrand et al., 2006; Jackson, 2008). Nonetheless, the ultimate visual effect is relatively similar, the eventual shift toward the brown. Initially the changes are considered beneficial, with the depth of color generally increasing and possessing a purplish cast. Subsequently, color depth decreases as the cast becomes increasingly brickish. As long as the change is not associated with the development of an oxidized flavor, and does not occur atypically rapidly, the color change is not considered a fault. It is expected and appreciated as a sign of proper maturation. Rapid color shifts are, however, viewed as indicating a defect and can markedly influence consumer acceptance. The color of red wines comes primarily from anthocyanins. Five types occur in wine, of which the most prevalent and stable is malvin. In grapes, anthocyanins typically are bound with glucose, and are arranged in loose, stacked complexes. During and after fermentation, these complexes tend to break down, and the sugar moiety may be lost. This leaves the pigment more susceptible to oxidation and color degradation. This tends to be avoided by polymerization


554


with the other principal group of flavonoids in wine, catechins and their polymers (procyanidins and condensed tannins). Polymerization may involve direct bonding between catechins with anthocyanins, or indirectly via the interaction of a by-product of phenol oxidation (the oxidation of ethanol to acetaldehyde). A host of other reactions products have recently been discovered that generate yellowish to brickish-colored pigments (Fulcrand et al., 2006; Jackson, 2008). The relative significance of these to the color shift of red wines is still being assessed. As noted, a color shift is considered normal and expected. They are not considered to contribute to a shortened shelf life. It is only when there may be insufficient flavonoids to stabilize the color (limit oxidative color loss), or there is excessive oxygen penetration, that an unacceptable premature browning and an oxidized flavor develop. Hazes and deposits Shelf life is negatively influenced by haze formation. To avoid its development, winemakers often go to great lengths. It is such a regular part of enologic practice that haze formation in bottled wine is now rare. Even the sediment (precipitated tannin/potassium tartrate complexes) that used to characterize most red wines is now relatively uncommon. If it occurs, it is not considered a fault. Deposition of sediment may even be considered a sign of quality. The only evidence that this view might possibly have some validity is the increased presence of several fruit esters in unfiltered wine (Moreno and Azpilicueta, 2006). Whether these esters would remain at levels sensorially significant by the time sediment might form is unestablished. Turbidity can be provoked by the presence of excessive amounts of iron or copper ions. Examples of colloidal hazes induced by the presence of metal ions are those produced by insoluble ferric phosphate in white wines (from the oxidization of ferrous phosphate); complexes formed between oxidized (ferric) ions and anthocyanins or tannins in red wines; and the development of a reddishbrown haze/deposit by the interaction of sulfides, copper ions and suspended proteins in white and occasionally roseÂ wines. The development of any of these is likely to result in consumer rejection. Microbial growth may also generate a haze. Although extensive yeast growth is usually necessary to produce evident turbidity (~105 cells/ml), Brettanomyces has been reported to form a distinct haze at 102 cells/ml (EdeleÂnyi, 1966; Amerine et al., 1980). Bacterial growth may also provoke haze generation. A classic example is the problem called ropiness. It is produced by the profuse synthesis of mucilaginous polysaccharides by some strains of Oenococcus oeni and Pediococcus. When the bottle is undisturbed, the polysaccharides hold the bacteria together in long silky chains that appear like floating threads. When dispersed, the polysaccharides give the wine an oily look and possibly a viscous texture. Occasionally wine may form crystals of potassium tartrate on the underside of corks, or form slender crystals that may be resuspended if the bottle is


Shelf life of wine 555 agitated. Tartrate crystal formation is more likely to form if the wine is stored at abnormally low temperatures. Additional sources of innocuous crystallization include a flavonol haze in white wines (fine yellow crystals of quercetin), a phenolic deposit generated by a fine precipitate of off-white to fawn-colored ellagic acid crystals (derived from extended exposure to oak chips), and crystals of the calcium salts of saccharic and mucic acids (in botrytized wines). Another precipitation problem occasionally observed in bottles of red wine is a lacquer-like finish on the inner surfaces of the bottle. The film apparently consists of a thin layer of tannins, anthocyanins and protein (Waters et al., 1996). Occasionally a film-like deposit, termed `masque', may also develop on the inner surfaces of champagne bottles. It appears to originate from the interaction of albumin (added as a fining agent) and fatty acids from yeasts (Maujean et al., 1978). Both may limit sales as they can give the false impression that the wine is turbid. 18.3.2 Gustatory changes One of the typical benefits derived from aging red wine is a reduction in bitterness and astringency. This benefit has usually been ascribed to the polymerization and precipitation of flavonoid phenolics. Conversely, if small phenolics remain in solution, or tannin polymers hydrolyze, perceived bitterness may increase. New evidence, however, suggests that bitterness arises more from nonflavonoids (ethyl esters of hydroxybenzoic and hydroxycinnamic acids) than flavonoids (Hufnagel and Hofmann, 2008). Other than the gustatory significance of phenolic polymerization, few other marked changes in taste occur during the shelf life of wine. Occasionally, a loss of sweetness in dry white wines may be noted. This probably arises from the hydrolysis of acetate esters. Their fruity fragrance in wine often induces the perception of sweetness, due to a learned association between fruity odors and sweetness (Prescott, 2004). Perceived sourness may also decrease slightly due to the slow esterification of tartaric acid (Edwards et al., 1985). Otherwise, noticeable gustatory changes usually accrue from microbial spoilage. 18.3.3 Olfactory changes The most significant modifications affecting shelf life involve aromatic deterioration. These have been most studied in white wines. The fresh, fruity fragrance of most young white wines, ascribed to the presence of acetate esters, is primarily the by-products of yeast metabolism. Because they often accumulate to concentrations significantly above their sensory threshold, and equilibrium with their acetic acid and alcohol constituents, they slowly hydrolyze back to their moieties after fermentation. This results in an eventual loss in their sensory impact. The rate at which these events affect the sensory attributes of wine depends on their importance to the fragrance (especially in white wines with


556


little distinctive varietal aroma), the concentration of the individual esters, their specific sensory thresholds, the storage temperature, and the wine's pH. In Riesling wines, most esters fall to equilibrium levels within about six years at 14 ëC (Rapp et al., 1985). Similar findings have been found with other cultivars (Marais and Pool, 1980; Ferreira et al., 1997). Oxidation accentuates their decline (Ferreira et al., 1997). Other esters may hydrolyze even earlier. For example, the sulfur-containing (3-methylthio-propyl)-acetate was detected only during the first month after fermentation. The degree to which a wine experiences these changes may partially depend on its phenolic content. For example, caffeic acid can markedly retard the degradation of acetate esters, and especially ethyl esters of fatty acids (Roussis et al., 2005, 2007). The presence of caffeic acid also retards the oxidation of terpenes. The detection and olfactory significance of compounds depends not only on their sensory thresholds, and how other wine constituents affect their perception (synergistically or antagonistically), but also on the individual sensitivity of the taster and their interest in assessing such details. In contrast to the decline in acetate ester concentration, the content of diethylesters of some dicarboxylic acids tends to rise (Table 18.1). Some components of an aged bouquet appear to depend on the degradation of grape-derived carotenoids. A major example is 1,1,6-trimethy1-1,2-dihydronaphthalene, a sensorially important vitispirane. Other sources of aging-derived fragrances include carbohydrate dehydration products such as 2-furfural and 2hydroxymethy1-5-furfural. The latter are best known in relation to heat exposure (madeira maturation and oak barrel toasting), but can also form slowly at room and cellar temperatures. Although some aromatic changes are unaffected by the wine's redox potential, others definitely are. Generally low redox (reductive) potentials are considered beneficial, whereas high redox (oxidative) potentials are detrimental. For example, wines depending primarily on monoterpene alcohols for their varietal character are damaged by oxygen ingress. Frequently, oxygen converts monoterpene alcohols into their corresponding oxides. Not only are these less volatile, but they possess qualitatively different aromas. For example, linalool oxides have flavor thresholds in the 3000±5000 g/l range, versus 100 g/l for linalool (Rapp, 1988). -Terpineol, an oxide of linalool, possesses a musty, pine-like odor, in contrast to the floral aspect of linalool. The result is a loss in the desired varietal (floral) distinctiveness of wines produced from cultivars such as Riesling, GewuÈrztraminer, Viognier, and Muscat. Another example of fragrance loss involves the oxidation of thiol flavorants (conversion to disulfides or other constituents). In Sauternes, most polyfunctional thiols, typical of the young wine, are no longer detectable two years after bottling (Bailly et al., 2009). The principal exception is 3-sulfanyhexan-1-ol. It remains at concentrations above its threshold value for several years. Other varietally significant aromatics, such as -terpineol, sotolon, several fermentation esters, ethyl esters, and oak maturation volatiles tend to remain for at least five years. Occasionally, thiol compounds accumulate during aging, at least in wines aged


Shelf life of wine 557 on lees. Examples are phenylmethanethiol and ethyl 3-sulfanylpropionate (Tominaga et al., 2003). Like most white wines, roseÂs have comparatively short shelf lives. RoseÂs often lose much of their fruity character within one to two years. This has been correlated with the degradation of 3-mercaptohexan-1-ol and 3-mercaptohenyl acetate (Murat et al., 2003; Murat, 2005). The limited presence of anthocyanins and tannins appears to be involved, their concentration being insufficient to provide adequate oxidative protection. This also probably explains another feature of the usual short shelf life of roseÂ wines ± their tendency to rapidly take on an orangish coloration due to oxidation. That the fruity character of roseÂ wines partially depends on acetate esters (notably phenethyl acetate) further clarifies the poor aging characteristics of roseÂs. Maintaining a low redox potential is generally considered essential for retaining the desirable attributes of a wine. Nonetheless, the adoption of screw cap closures may induce the development of a reduced sulfur odor, shortening shelf life. Its occurrence may partially depend on the oxygen impermeability of the foam liner used in the aluminum cap (Lopes et al., 2009). Alternatively, others consider the problem results from conditions associated with grape cultivation or fermentation. Although mercaptans are universally considered a fault, hydrogen sulfide can be considered acceptable, or even desirable, at just detectable values. In addition, the slow buildup of dimethyl sulfide (DMS) is often considered desirable, if not too obvious. In `late harvest' white wines, du Plessis and Loubser (1974) found DMS contributed to the full-bodied, bottle-aged characters of the wines. Segurel et al. (2004) found it contributed to the fruity, as well as the aged truffle and black olive attributes of Syrah and Grenache wines. Storage at cellar temperatures retards the formation of DMS. In addition to the loss or modification of grape- and yeast-derived fragrances, a wide range of new compounds develop during in-bottle aging. Some are beneficial, others undesirable, notably those resulting from oxidation. Although the origin of acetaldehyde is the best understood, its relevance to the òxidized' odor of table wines remains doubtful (Escudero et al., 2002). Its distinctive odor is typically not detectable in table wines considered oxidized, nor does the free (volatile) concentration of acetaldehyde rise upon short-term, intentional oxidation (Singleton et al., 1979; Escudero et al., 2002). Although acetaldehyde is produced as an indirect by-product of phenol oxidation, free acetaldehyde is rapidly removed by its reaction with other wine constituents. It is only in some fortified wines, notably sherries, that an aldehydic odor is considered appropriate and expected. When wines are exposed to oxygen for several weeks, the treated wines develop characteristics described as `cooked vegetables,' and `pungent.' These attributes have been correlated with the presence of furfural and hexanal, respectively. Their pronounced odors may mask the wine's natural fragrance. Other aromatic constituents potentially involved in the development of an oxidized odor appear to be 2-nonenal, 2-octenal, and benzaldehyde (Ferreira et al.,


558


1997). Escudero et al. (2000) also associated methional with a cooked cabbage odor in oxidized wine. Additional studies have correlated oxidative odors with the production of methional, phenylacetaldehyde, 4,5-dimethyl-3-hydroxy-2(5H)furanone (sotolon), and trimethyl-1,5-dihydronaphthalene (TDN), particularly at high (45 ëC) temperatures and/or low pH values (Silva Ferreira et al., 2003). Quinones generated during oxidation may react with amino acids. These reactions could potentially enhance flavor by generating aldehydes and ketones in Strecker-type degradations. It is a moot point whether some of the attributes associated with oxidation, such as honey (phenylacetaldehyde), hay and wood (eugenol), should be viewed as oxidation off-odors, or the benefits of aging. This situation is typical of issues relating to wine shelf life ± there is no clear definition of what constitutes an oxidized odor. In addition to the development of undesirable oxidized odors, and the loss of varietal and fermentation fragrances, associated with a shortened shelf life, there may be the formation of an attribute, termed the aged bouquet, that enhances shelf life. Examples of desirable compounds, formed during bottle-aging in red wines, matured in oak, include 4-ethylphenol, 2-methoxy-4-ethylphenol, 2furaldehyde, 5-butyl-4-methyldihydrofuran-2(3H)-one (whisky lactone) (PeÂrezPrieto et al., 2003; Fernandez de Simon et al., 2006). In Sauternes, desired, agerelated constituents include homofuraneol, theaspirane, -decalactone and abhexon (Bailly et al., 2009). The conditions that favor an aged bouquet development are far less understood than those that limit their formation. Examples of the latter are high pH, limited skin contact before or during fermentation, warm storage temperatures, exposure to sunlight, ingress of oxygen, and low alcohol content. Examples of terms used to describe an aged bouquet may include `leather,' `cigar box,' `truffle' for red wines, and `sun-dried linen' for white wines. These terms are used almost universally, regardless of the wine's varietal origin. Microbial spoilage is thankfully comparatively rare, but can rapidly shorten shelf life. Not only can it generate unsightly turbidity, noted above, but also can generate a variety of off-odors. The most common spoilage organisms are strains of acetic acid bacteria, some lactic acid bacteria, and yeasts such as Brettanomyces. At near threshold values, certain of their metabolic by-products, notably ethyl acetate, acetic acid, 4-ethyl phenol and related compounds, may be viewed as donating a desirable complexity to the flavor. Their presence has occasionally been mistaken as a component of the wine's geographic (terroir) attributes, presumably by those having high threshold values for these compounds. At recognition values they are highly undesirable.

18.4

Evaluating wine shelf life

Estimating the aging potential (shelf life) of a wine is one of the favorite pastimes of wine critics and connoisseurs. Regrettably, these predictions are


Shelf life of wine 559 based on past recollections, not on objective or verifiable criteria. Thus, they are only a general guide, assuming ideal storage conditions. Examples are views that fino sherry and beaujolais nouveau wines possess short shelf lives (several months); that most white wine begins to show oxidative browning within two to three years, and experience a perceptible loss in fruity odors within a year; most regular red wines lose much of the fruity and varietal aromas within five years; sparkling wines begin to lose their characteristic effervescence with a few years; and most ports show little or no aromatic improvement subsequent to bottling ± with the exception of vintage ports (reaching `maturity' within 15 to 20 years). Despite being a useful guide, nothing at present can accurately predict how any particular wine will develop. Nonetheless, a few researchers are investigating nondestructive, in situ methods of assessing some aspects of shelf life. These include nuclear magnetic resonance (NMR) and spectroscopy. NMR has the potential to quantify certain wine constituents, indicating for example the presence of spoilage amounts of acetic acid (Weekley et al., 2003) and acetaldehyde (Sobieski et al., 2006). Because of its expense, its application may be appropriate only to verify the drinkability of rare vintage wine sold at auction. More affordable and readily available is spectroscopy. Regrettably, its analytic range is currently limited to assessing the degree of oxidative browning in white wines ± absorption at 420 nm (Skouroumounis et al., 2003), and total and free sulfur dioxide contents (Cozzolino et al., 2007). Spectroscopy is most readily applicable to wine in clear flint glass bottles. For amber and antique green glass, which absorbs intensely in the blue range, absorption at 540 nm and 540 nm or 600 nm, respectively, can be substituted (Skouroumounis et al., 2003). Nonetheless, with empty versions as a control, it can be applied to wine in colored glass bottles. Absorption at 420 nm is typically used as a rough indicator of oxidation, but Silva Ferreira et al. (2003) recommend forced oxidation as an indicator of the potential of a wine to undergo premature oxidation. It can be used to assess the need for additional antioxidants prior to bottling.

18.5

Preventing wine quality deterioration at or post-bottling

18.5.1 Fining, filtration and disinfection Wine is typically fined to remove soluble proteins, excess tannins, metal ions, or other constituents that could lead to haze formation or off-odor production. In addition, the wine is usually given a final polishing filtration prior to bottling, to remove particulate material that might subsequently precipitate in the bottle. Filtration is usually not designed to remove microorganisms. Such disinfection is typically unnecessary due to their relatively low numbers and the unfavorable conditions for growth in bottled wine. Correspondingly, other than regular cleaning after manufacture, bottles are typically not sterilized prior to filling. Only in sweet wines of low alcohol content may sterile (> 0.5 ) filtration into sterilized bottles be advisable to avoid microbial spoilage.


560


Occasionally, some low priced, sweet wine may be pasteurized prior to bottling for the same reason. 18.5.2 Sulfur dioxide addition Once bottled, shelf life largely depends on the storage conditions and protective nature of the enclosure. Retarding oxygen ingress is one of the primary methods, as it favors optimal maturation and limits the development of undesirable flavors. Nonetheless, oxygen can gain access to bottled wine from air present in the headspace of the bottle, present in cellular voids in natural cork, seep in via faults or creases in the cork, or slowly diffuse through the cork. Although oxygen ingress may be minimized by a range of procedures, sulfur dioxide is typically added to limit the damage caused by what does enter. The amount added depends largely on the type of wine. White wines, having lower phenolic contents, need more protection than red wines. Also, red wines typically receive less sulfur dioxide to avoid partial bleaching of the wine's anthocyanin content. Sulfur dioxide owes much of its antioxidant activity to its reaction with peroxide, generated by the oxidation of phenols. This limits the oxidation of important fermentation odors, notably acetate esters, and varietally important aromatics, such as monoterpenes, and thiols (Blanchard et al., 2004). Sulfur dioxide also limits the oxidation of ethanol to acetaldehyde, binds acetaldehyde in a nonvolatile complex, and participates in the regeneration of phenols from quinones. Sulfur dioxide is also used in low alcohol, sweet table wines to reduce the incidence of microbial spoilage. It often remains at a level sufficient to be effective for periods of up to one to two years, usually sufficient to cover the normal shelf life of most white wines. Red wines, with long aging potentials, usually have sufficiently low pH values and high tannin contents to provide adequate microbial protection in the absence of oxygen. 18.5.3 Bottle closure and orientation Oxygen uptake is primarily restricted by the closure, but can also be reduced at filling by flushing air out of the bottle, usually with carbon dioxide. Subsequent upright positioning of the bottles for 24 h permits pressurized gas in the headspace, compressed on cork insertion, to escape between the neck and the stopper. Despite the elasticity of the cork, it takes about 24 h for a tight seal to fully develop. Any residual pressurized CO2 in the headspace is absorbed into the wine, equilibrating the pressure within the bottle and the atmosphere. Nitrogen gas is not recommended due to its poor solubility in wine. Pressurized nitrogen in the headspace would continue to exert its force against the cork. Alternatively, bottles may be put under a partial vacuum before filling. De Rosa and Moret (1983) showed that vacuum and flushing, and vacuum application, reduced average SO2 loss after 12 months from 28 to 16 mg/l and 5 mg/l, respectively. Reduction in sulfur dioxide content is often considered an


Shelf life of wine 561 indication of the oxygen ingress. Application of a vacuum also avoids a pressure buildup in the headspace on closure insertion. With the introduction and market acceptance of a variety of bottle closures, there has been an upsurge in research on their effects on shelf life. Previously, the only variables were the quality of natural cork, its degree of compression in the bottle, and stopper length. Currently, the choice of closures has expanded to include agglomerate cork, synthetic corks, glass stoppers, and aluminum screw caps. Because of the importance of limiting oxygen access, most of the studies have concentrated on oxygen penetration and bottle-to-bottle variability. Natural cork appears to be associated with a short period of oxygen uptake, from the cellular voids in cork cells. Uptake subsequently declines to about 2± 6 l O2/day, finally stabilizing within a year at about 0.1±2 l O2/day (Lopes et al., 2007). In contrast, use of screw caps and agglomerate cork seems to be associated with less than 1 l O2 ingress per day, equivalent to that of the highest quality natural corks. Regrettably, the oxygen permeability of natural cork tends to exhibit variation (Caloghiris et al., 1997), occasionally by up to four orders of magnitude (Hart and Kleinig, 2005). In contrast, agglomerate cork and screw caps show high consistency in oxygen permeability. Synthetic corks tend to show undesirably high diffusion rates (Godden et al., 2001), often within the range of 6±12 l O2/day. Such high rates limit shelf life and restrict their use to wines intended for early consumption (less than two years). Nonetheless, where early maturity is desired, the increased oxygen permeability of synthetic corks might have some benefit. Although long, high quality, natural corks can provide good protection from oxygen uptake, corks are still potentially plagued by additional problems. The principal is their periodic association with a corked off-odor. This moldy odor is primarily, but not exclusively, caused by migration of 2,4,6-trichloroanisole (TCA) from the cork (Juanola et al., 2005). It is generally viewed as the most common fault in wine, and can severely shorten its shelf life. Various procedures have been developed to remove this contaminant from corks, as well as coatings developed to retard (if not prevent) its migration into wine. Contamination with other sources of moldy odors and extraneous off-odors is less common. This situation has led to a trend away from natural cork closures in many markets For white wines, there is general consensus as to the detrimental effect of oxygen ingress. There is more controversy concerning the effects of the slow, minimal, oxygen uptake from cork closures on red wine maturation, especially premium versions intended for extended bottle aging. Although opinions differ widely, recent evidence suggests that red wines do not need, at least at detectable levels, oxygen ingress for their maturation (Hart and Kleinig, 2005). Minimal oxygen contact retards the loss of the oxidative protection provided by both sulfur dioxide and tannins. Thus, color density is retained for a longer period and hue changes are slowed. Loss of the fruity fragrance of the wine is also delayed, further extending the wine's youthful characteristics. Bottle orientation is also important in minimizing oxygen ingress. For natural cork closures, horizontal positioning is essential for long-term storage. Contact


562


between the stopper and the wine is required to maintain a stable moisture content. Otherwise, the cork will dry, losing its resilience and elasticity. Both are essential to retaining its protective sealing properties (Bartowsky et al., 2003). Nonetheless, serious consequences from upright storage may not appear for more than two years (Lopes et al., 2006). It may also take several years before upright storage results in distinct browning and the development of an oxidized odor (Skouroumounis et al., 2005). In contrast, hybrid cork stoppers, consisting of a combination of ground cork (often with its lignin content removed) and microspheres of a synthetic polymer, are comparatively moisture independent. Bottles so closed, like those with screw caps, can safely be stored upright. 18.5.4 Storage container The use of glass bottles has a long tradition dating back to the late 1500s. Its prime benefit, relative to shelf life, is its inertness and gas impermeability. Its transparency also facilitates direct observation of visual defects. Nonetheless, light transmittance permits some types of haze development, can induce lightstruck off-odors, and cause heat-induced problems. Surprisingly, white wines, those most susceptible to these disruptions, are typically the least protected, being bottled in clear or light-colored glass. Despite the predominant use of glass bottles, its position is being challenged by other containers. Improvements in the oxygen impermeability of lightweight, PET (polyethyleneterephtalate) bottles has led to their tentative adoption for wine storage (Giovanelli and Brenna, 2007). Currently, bag-in-box containers (Biuatti et al., 1997) are the principal alternative to glass. Here, the principal factor limiting shelf life is the attachment of the spigot that permits oxygen ingress (Armstrong, 1987). Other containers, such as aluminum cans and carton boxes are becoming popular in some markets for the casual wine consumer. Although the principal value of bag-in-box containers is their consumer friendliness, the plastic liner can have a little known, unexpected benefit. It effectively absorbs some wine off-odors. Examples are the absorption of 2,4,6trichloroanisole (TCA), napthalene, and 1,1,6-trimethyl-1,2-dihydronaphthanene (TDN) (Capone et al., 2002). Regrettably, pleasant smelling aromatics may also be `scalped.' This feature has also been detected with cork and artificial closures (Capone et al., 2003). 18.5.5 Temperature control As noted previously, the temperature of wine storage not only significantly affects the speed and characteristics of wine aging, but also can shorten its shelf life. Temperatures above 25 ëC, and especially above 30 ëC, modify the flavor characteristics of table wines in ways unacceptable to most consumers. Typically, wines for long storage are held at a relatively constant temperature of about 10 ëC. This retards most chemical reactions in the wine, providing


Shelf life of wine 563 conditions optimal for what is considered favorable maturation. Colder temperatures may even further extend the shelf life, but at the expense of excessively slow maturation (Table 18.1).

18.6

Sensory significance of shelf life changes

A major concern for wine makers and wine merchants alike is consumer alienation. This not only results in direct financial loss, connected with bottle returns, but also indirect loss, associated with the forfeiture of repeat sales. Nonetheless, in most cases, the source(s) of customer dissatisfaction are unknown. There is no organized system in the wine industry for assessing why consumers have negative reactions to particular wines. When consumer surveys are conducted, it is critical that questions be worded to avoid biasing the response. In addition, the questions should be relevant to real-life situations (KoÈster, 2003). For example, questioning about faults can exaggerate their apparent presence, and preference choices in a laboratory or shopping mall will unlikely generate valid data relating to purchase activity. Of visual faults, oxidative browning is universally unacceptable among wine industry professionals. Despite this, there is little definitive evidence that consumers view this issue as equally important. The significance of wine color deviance appears to depend markedly on the experience or training of the assessor (Williams et al., 1984). Rejection appears most often to be associated with odor and taste faults. Commonly stated reasons are the presence of corked, oxidized, and vinegary attributes. Most corked odors probably arise from the presence of above threshold amounts of 2,4,6-trichloroanisole (though, as noted previously, corky/ moldy odors may be donated by other compounds). Oxidized odors are frequently mentioned, but exactly what consumers mean when they use this term is far from clear. In most instances, the mention of a vinegary odor is probably inappropriate. It probably refers to their perception that the wine is inappropriately or excessively sour. Nonetheless, it could legitimately arise from above threshold amounts of acetic acid (often associated with ethyl acetate). Other potential sources of shortened shelf life include a naphthalene-like odor found in some white wines (termed untypical aging flavor), a prune-like odor in red wines that age prematurely (possibly caused by 3-methyl-2,4-nonanedione) (Pons et al., 2008), baked odors (due to exposure to high temperatures), and the presence of a reduced sulfur (shrimp-like) odor (found in some white wines closed with screw caps). Although there is contention among wine industry professionals as to the latter's origin, it is also uncertain as to its importance to consumers. From my experience, the wine industry is lucky that consumers frequently do not recognize wine faults, and correspondingly do not consider them a shelf life issue. Even wine critics are often ambivalent on wine faults, accepting barnyardy (ethyl phenol) odors as a quality in some famous RhoÃne wines, and nail polish (ethyl acetate) odors in prestigious Sauternes. Thus, wine


564


faults, and their shelf life significance, are often perceptions influenced by experience and exterior biases, not objective decisions.

18.7

Future trends

Because oxidation continues to be one of the principal causes of shortened shelf life, there is considerable interest in new means of addressing this issue. This is especially so, due to the increasing insistence of government and consumer advocacy groups on limiting the use of sulfur dioxide, despite it being the safest and most effective antioxidant currently available. Alternatives under investigation include compounds such as glutathione. Glutathione is a natural grape antioxidant, but is rapidly consumed in oxidative reactions during fermentation. Thus, for it to be of value in extending wine shelf life it must be added at bottling. Its addition, alone or in combination with caffeic acid (the principal hydroxycinnamic acid in wines) has been shown to delay the degradation of important flavorants in wines, notably acetate esters and terpenes (Roussis et al., 2005). Gallic acid is another natural phenolic that can limit oxidation (Lambropoulos and Roussis, 2007). N-acetyl-cysteine also can play an antioxidant role, in association with caffeic acid in wines (Sergianitis and Roussis, 2008). A very different approach, the immobilization of the yeast for fermentation in carragenate beads, has apparently been found successful in limiting browning (Merida et al., 2007). Other approaches are attempts to better predict wine shelf life. Examples are tests to assess a wine's susceptibility to browning (Palma and Barroso, 2002) or temperature damage (Sivertsen et al., 2001). As noted previously, techniques such as NMR and spectroscopy hold promise. They have the potential for assessing the presence of aroma compounds, off-odors, additives, and contaminants in unopened bottles. While still primarily in the research stage, improvements in their use and access may facilitate their transfer from the laboratory to the winery or wholesale merchant. In addition to issues about commercializing NMR and spectroscopic assessment, a precise correlation between wine chemistry and perceived quality is lacking. This requires further investigation to establish the compounds that are essential to a wine's distinctive sensory attributes. This will likely involve techniques such as gas chromatography-olfactometry (GC-O) (Plutowska and Wardencki, 2008), combined with data on odor activity value (OAV). The odor activity value is derived by dividing the concentration of a constituent by its detection threshold. Most significant aromatic constituents tend to possess OAVs above unity. Nonetheless, the sensory importance of a compound can only be confirmed by selectively adding them to a model wine, and assessing its impact on aroma detection (Grosch, 2001). Such a procedure has been successfully used in assessing the chemical nature of the varietal aroma of several wines (Hashizume and Samuta, 1997; Ferreira et al., 1998; Guth 1998; Kotseridis et al., 2000).


Shelf life of wine 565

18.8


Additional information on issues noted in this chapter may be found in books such as Jackson (2008), and reviews by Aldave et al. (1993), Marais (1986), RibeÂreau-Gayon (1986), Rapp and GuÈntert (1986), Rapp and Marais (1993) and Singleton (2000). Another informative review, but dealing with beer, is provided by Vanderhaegen et al. (2006).

18.9

References

and SALVADOR MD (1993), `The shelf-life of young white wine' in G Charakanbous (ed.) Shelf Life Studies of Foods and Beverages. Chemical, Biological, Physical and Nutritional Aspects. Amsterdam, Elsevier, pp. 923±943. AMERINE MA, BERG HW, KUNKEE RE, OUGH CS, SINGLETON VL and WEBB AD (1980), The Technology of Wine Making. AVI, Westport, CT. ARMSTRONG DM (1987), `Leakers, equipment failure and quality assurance procedures for defect supervision' Proc 6th Aust Wine Ind Tech Conf, Adelaide, Australian Industrial Publishers, pp. 275±276. BAILLY S, JERKOVIC V, MEUREÂE A, TIMMERMANS A and COLLIN S (2009), `Fate of key odorants in Sauternes wines through aging' J Agric Food Chem, 57, 8557±8563. BARTOWSKY EJ, XIA D, GIBSON RL, FLEET GH and HENSCHKE PA (2003), `Spoilage of bottled red wine by acetic acid bacteria' Lett Appl Microbiol, 36, 307±314. BIUATTI S, CELOTTI E, FERRARINI R and ZIRONI R (1997), `Wine packaging for market in containers other than glass' J Agric Food Chem, 45, 2081±2084. BLANCHARD L, DARRIET P and DUBOURDIEU D (2004), `Reactivity of 3-mercaptohexanol in red wine: impact of oxygen, phenolic fractions, and sulfur dioxide' Am J Enol Vitic, 55, 115±120. CALOGHIRIS M, WATERS EJ and WILLIAMS PJ (1997), Àn industry trial provides further evidence for the role of corks in oxidative spoilage of bottled wines' Aust J Grape Wine Res, 3, 9±17. CAPONE DL, FRANKE S, MCLEAN HJ and SEFTON MA (2002), `The effect of wine cask bladder packaging on the composition of a white wine' Proc 11th Aust Wine Ind Tech Conf, Winetitles, Adelaide, Australia. pp. 270±271. CAPONE D, SERFON M, PRETORIUS I and HéJ P (2003), `Flavour `scalping' by wine bottle closures ± the `winemaking' continues post vineyard and winery' Aust NZ Wine Indust J, 18, 16, 18±20. CARPENTIER N and MAUJEAN A (1981), Light flavours in champagne wines. In P. Schreier (ed.), Flavour '89. 3rd Weurman Symposium Proceedings of the International Conference. Berlin, de Gruyter, pp. 609±615. CHASSAGNE D, GUILLOUX-BENATIER M, ALEXANDRE H and VOILLEY A (2005), `Sorption of wine volatile phenols by yeast lees' Food Chem, 91, 39±44. CHATONNET P, BONNET S, BOUTOU S and LABADIE M-D (2004), Ìdentification and responsibility of 2,4,6-tribromoanisole in musty, corked odors in wine' J Agric Food Chem 52, 1255±1262. CHUNG H-J, SON J-H, PARK E-Y, KIM E-J and LIM S-T (2008), Èffect of vibration and storage on some physico-chemical properties of a commercial red wine' J Food Comp Anal, 21, 655±659. Â CERES I, GONZA Â LEZ-RAURICH M ALDAVE L, ALMY J, CABEZUDO MD, CA


566


and GISHEN M (2007), À feasibility study on the use of visible and short wavelengths in the near-infrared region for the nondestructive measurement of wine composition' Anal Bioanal Chem, 387, 2289± 2295. D'AURIA M, EMANUELE L, MAURIELLO G and RACIOPPI R (2003), Òn the origin of gou Ã t de lumiere in champagne' J Photochem Photobiol A: Chemistry, 158, 21±26. DE ROSA T and MORET I (1983), Ìnfluenza dell'imbottigliamento in ambiente gas inerte sulla conservazione di uno vino blanco tranquillo' Rev Viticult, Enol, 36, 219-226. DU PLESSIS CS and LOUBSER GJ (1974), `The bouquet of ``late harvest'' wine' Agro Chem Phys, 6, 49±52. EDWARDS TL, SINGLETON VL and BOULTON R (1985), `Formation of ethyl esters of tartaric acid during wine aging: chemical and sensory effects' Am J Enol Vitic, 36, 118± 124. EDELEÂNYI M (1966), `Study on the stabilization of sparkling wines' (in Hungarian) BorgazdasaÂg, 12, 30±32 (reported in Amerine et al., 1980). ESCUDERO A, CACHO J and FERREIRA V (2000), Ìsolation and identification of odorants generated in wine during its oxidation: a gas chromatographic/olfactometric study' Eur Food Res Technol, 211, 104±111. ESCUDERO A, ASENSIO E, CACHO J and FERREIRA V (2002), `Sensory and chemical changes of young white wines stored under oxygen. An assessment of the role played by aldehydes and some other important odorants' Food Chem, 77, 325±331. FERNANDEZ DE SIMON B, CADAHIA E, HERNANDEZ T and ESTRELLA I (2006), Èvolution of oak-related volatile compounds in a Spanish red wine during 2 years bottled, after aging in barrels made of Spanish, French and American oak wood' Anal Chim Acta, 563, 198±203. Â NDEZ P and CACHO JF (1997), `Changes in the profile of FERREIRA V, ESCUDERO A, FERNA volatile compounds of wines stored under oxygen and their relationship to the browning process' Z Lebens-Unters Forsch A, 205, 392±396. Â PEZ R, ESCUDERO A and CACHO JF (1998), `The aroma of Grenache red wine: FERREIRA V, LO hierarchy and nature of its main odorants' J Sci Food Agric, 77, 259±267. Ä AS M, SALAS E and CHEYNIER V (2006), `Phenolic reactions during FULCRAND H, DUEN winemaking and aging' Am J Enol Vitic, 57, 289±297. GIOVANELLI G and BRENNA OV (2007), Òxidative stability of red wines stored in packages with different oxygen permeability' Eur Food Res Technol, 226, 169±179. GODDEN P, FRANCIS L, FIELD J, GISHEN M, COULTER A, VALENTE P, HéJ P and ROBINSON E (2001), `Wine bottle closure: physical characteristics and effect on composition and sensory properties of a Semillon wine. 1. Performance up to 20 months postbottling' Aust J Grape Wine Res, 7, 64±105. GROSCH W (2001), Èvaluation of the key odorants of foods by dilution experiments, aroma models and omission' Chem Senses, 26, 533±545. GUTH H (1998), `Comparison of different white wine varieties by instrumental and analyses and sensory studies' in LA Waterhouse and SE Ebeler (eds), Chemistry of Wine Flavor. ACS Symposium Series #714, Washington, DC, American Chemical Society, pp. 39±52. HART A and KLEINIG A (2005), `The role of oxygen in the aging of bottled wine' Aust NZ Grapegrower Winemaker, 497a, 79±80, 82±84, 86, 88. HASHIZUME K and SAMUTA T (1997), `Green odorants of grape cluster stem and their ability to cause a stemmy flavor' J Agric Food Chem, 45, 1333±1337. HUFNAGEL JC and HOFMANN T (2008), `Quantitative reconstruction of the nonvolatile COZZOLINO D, KWIATKOWSKI MJ, WATERS EJ


Shelf life of wine 567 sensometabolome of a red wine' J Agric Food Chem, 56, 9190±9199. (2008), Wine Science: Principles and Application, 3rd edn. San Diego, CA, Academic Press. Á D, SALVADOÂ V, REGUEIRO JAG and ANTICO Â E (2005), `Migration of 2,4,6JUANOLA R, SUBIRA trichloroanisole from cork stoppers to wine' Eur Food Res Technol 220, 347±352. È STER EP (2003), `The psychology of food choice: some often encountered fallacies' KO Food Qual Pref 14, 359±373. KOTSERIDIS Y, RAZUNGLES A, BERTRAND A and BAUMES R (2000), `Differentiation of the aromas of Merlot and Cabernet Sauvignon wines using sensory and instrumental analysis' J Agric Food Chem, 48, 5383±5388. LAMBROPOULOS I and ROUSSIS IG (2007), Ìnhibition of the decrease of volatile esters and terpenes during storage of a white wine and a model wine medium by caffeic acid and gallic acid' Food Res Intern, 40, 176±181. LAURIE VF and WATERHOUSE AL (2006), Òxidation of glycerol in the presence of hydrogen peroxide and iron in model solutions and wine: potential effects on wine color' J Agric Food Chem, 54, 4668±4673. LAVIGNE V, PONS A, DARRIET P and DUBOURDIEU D (2008), `Changes in the sotolon content of dry white wines during barrel and bottle aging' J Agric Food Chem. 56, 2688± 2693. LEINO M, FRANCIS I, KALLIO H and WILLIAMS PJ (1993), `Gas chromotographic headspace analysis of Chardonnay and SeÂmillon wines after thermal processing' Z Lebensm Forsch, 197, 29±33. LOPES P, SAUCIER C, TEISSEDRE PL and GLORIES Y (2006), Ìmpact of storage position on oxygen ingress through different closures into wine bottles' J Agric Food Chem, 54, 6741±6746. LOPES P, SAUCIER C, TEISSEDRE PL and GLORIES Y (2007), `Main routes of oxygen ingress through different closures into wine bottles' J Agric Food Chem, 55, 5167±5170. JACKSON RS

LOPES P, SILVA MA, PONS A, TOMINAGA T, LAVIGNE V, SAUCIER C, DARRIET P, TEISSEDRE P-L

and DUBOURDIEU D (2009), Ìmpact of oxygen dissolved at bottling and transmitted through closures on the composition and sensory properties of a Sauvignon blanc wine during bottle storage' J Agric Food Chem, 57, 10261±10270. MARAIS J (1978), `The effect of pH on esters and quality of Colombar wine during maturation' Vitis, 17, 396±403. MARAIS J (1986), Èffect of storage time and temperature of the volatile composition and quality of South African Vitis vinifera L. cv. Colombar wines' in G Charalambous (ed.), The Shelf Life of Foods and Beverages, Amsterdam, Elsevier, pp. 169±185. MARAIS J and POOL HJ (1980), Èffect of storage time and temperature on the volatile composition and quality of dry white table wines' Vitis, 19, 151±164. MARAIS J, VAN WYK CJ and RAPP A (1992), Èffect of storage time, temperature and region on the levels of 1,1,6-trimethyl-1,2-dihydronaphthalene and other volatiles, and on quality of Weisser Riesling wines' S Afr J Enol Vitic, 13, 33±44. Â RQUEZ T, MILLA Â N C, SOUQUET J-M and SALMON J-M (2009), Èffect of different yeast MA strains and their culture conditions on the prevention of wine model solution browning by yeast lees' J Agric Food Chem, 57, 3771±3779. Â tude sur un pheÂnomeÁne de masque observeÂ MAUJEAN A, HAYE B and BUREAU G (1978), È en Champagne' Vigneron Champenois 99, 308±313. MERIDA J, LOPEZ-TOLEDANO A and MEDINA M (2007), Ìmmobilized yeasts in -carragenate to prevent browning in white wines' Eur Food Res Technol 225, 279±286. MOIO L, UGLIANO M, GENOVESE A, GAMBUTI A, PESSINA R and PIOMBINO P (2004), Èffect of


568


antioxidant production of must on volatile compounds and aroma shelf life of Falanghina (Vitis vinifera L.) wine' J Agric Food Chem, 52, 891±897. MORENO NJ and AZPILICUETA CA (2006), `The development of esters in filtered and unfiltered wines that have been aged in oak barrels' Int J Food Sci Technol, 41, 155±161. MURAT M-L (2005), `Recent findings on roseÂ wine aromas. Part 1: identifying aromas studying the aromatic potential of grapes and juice' Aust NZ Grapegrower Winemaker, 497a, 64±65, 69, 71, 73±74, 76. MURAT M-L, TOMINAGA T, SAUCIER C, GLORIES Y and DUBOURDIEU D (2003), Èffect of anthocyanins on stability of a key-odorous compound, 3-mercaptohexan-1-ol, in Bordeaux roseÂ wines' Am J Enol Vitic, 54, 135±138. OSZMIANSKI J, CHEYNIER V and MOUTOUNET M (1996), Ìron catalyzed oxidation of (+)catechin in model systems' J Agric Food Chem, 44, 1712±1715. PALMA M and BARROSO CG (2002), Àpplication of a new analytic method to determine the susceptibility of wine to browning' Eur Food Res Technol, 214, 441±443. Â PEZ-ROCA JM, MARTIÂNEZ-CUTILLAS A, PARDO-MIÂNQUEZ F and GO Â MEZPEÂREZ-PRIETO LJ, LO PLAZA E (2003), Èxtraction and formation dynamics of oak-related volatile compounds from different volume barrels to wine and their behavior during bottle storage' J Agric Food Chem, 51, 5444±5449. PLUTOWSKA B and WARDENCKI W (2008), Àpplication of gas chromatography-olfactometry (GC±O) in analysis and quality assessment of alcoholic beverages ± a review' Food Chem, 107, 449±463. PONS A, LAVIGNE V, ERIC R, DARRIET P and DUBOURDIEU D (2008), Ìdentification of volatile compounds responsible for prune aroma in prematurely aged red wines' J Agric Food Chem, 56, 5285±5290. PRESCOTT J (2004), `Psychological processes in flavour perception' in AJ Taylor and D Roberts (eds), Flavour Perception. London, Blackwell Publishing, pp. 256±278. RAPP A (1988), Wine aroma substances from gas chromatographic analysis. In HF Linskens and JF Jackson (eds), Wine Analysis. Berlin, Springer-Verlag, pp. 29±66. È NTERT M (1986), `Changes in aroma substances during the storage of white RAPP A and GU wines in bottles' in G Charalambous (ed.), The Shelf Life of Foods and Beverages. Amsterdam, Elsevier, pp. 141±167. RAPP A and MARAIS J (1993), `The shelf life of wine: changes in aroma substances during storage and ageing of white wines' in G Charakanbous (ed.), Shelf Life Studies of Foods and Beverages. Amsterdam, Elsevier, pp. 891±921. È ber VeraÈnderungen der Aromastoffe È NTERT M and ULLEMEYER H (1985), Ù RAPP A, GU waÈhrend der Flaschenlagerung von Weibweinen der rebsorte Riesling' Z Lebensm Forsch, 180, 109±116. È HNERTZ O È RBEL H, POUR NIKFARDJAM M, LOOS U and LO RAUHUT D, SHEFFORD PG, ROLL C, KU (2001), Èffect of pre- and/or postfermentation addition of antioxidants like ascorbic acid or glutathione on fermentation, formation of volatile sulfur compounds and other substances causing off-flavours in wine' in Proc 26th OIV Conf, Adelaide, pp. 76±82. RIBEÂREAU-GAYON P (1986), `Shelf-life of wine' in G Charalambous (ed.), Handbook of Food and Beverage Stability: Chemical, Biochemical, Microbiological and Nutritional Aspects, Orlando, FL, Academic Press, pp. 745±772. ROUSSIS IG, LAMBROPOULOS I and PAPADOPOULOU D (2005), Ìnhibition of the decline of volatile esters and terpenols during oxidative storage of Muscat-white and Xinomavro-red wine by caffeic acid and N-acetyl-cysteine' Food Chem, 93, 485±492.


Shelf life of wine 569 and TZIMAS P (2007), `Protection of volatiles in a wine with low sulfur dioxide by caffeic acid or glutathione' Am J Enol Vitic, 58, 274±278. SALINAS MR, ALONSO GL, NAVARRO G, PARDO F, JIMENO J and HUERTA MR (1996), Èvaluation of the aromatic composition of wines undergoing carbonic maceration under different aging conditions' Am J Enol Vitic, 47, 134±144. SEFTON MA and SIMPSON RF (2005), `Compounds causing cork taint and the factors affecting their transfer from natural cork closures to wine ± a review' Aust J Grape Wine Res, 11, 188±200. SEGUREL MA RAZUNGLES AJ RIOU C SALLES M and BAUMES RL (2004), `Contribution of dimethyl sulfide to the aroma of Syrah and Grenache noir wines and estimation of its potential in grapes of these varieties' J Agric Food Chem, 52, 7084±7093. SERGIANITIS S and ROUSSIS IG (2008), `Protection of volatile esters and terpenes during storage of a white wine and a model wine medium by a mixture of N-acetylcysteine and caffeic acid' Eur Food Res Technol, 227, 643±647. SILVA FERREIRA AC, HOGG T and GUEDES DE PINHO P (2003), Ìdentification of key odorants related to the typical aroma of oxidation-spoiled white wines' J Agric Food Chem, 51, 1377±1381. SIMPSON RF (1982), `Factors affecting oxidative browning of white wine' Vitis, 21, 233± 239. SIMPSON RF, CAPONE DL and SEFTON MA (2004), Ìsolation and identification of 2-methoxy3,5-dimethylpyrazine, a potent musty compound from wine corks' J Agric Food Chem, 52, 5425±5430. SINGLETON VL (1962), Àging of wines and other spiritous products, acceleration by physical treatments' Hilgardia 32, 319±373. SINGLETON VL (1987), Òxygen with phenols and related reactions in musts, wines, and model systems: observation and practical implications' Am J Enol Vitic, 38, 69±77. SINGLETON VL (2000), À survey of wine aging reactions, especially with oxygen' in Proc Am Soc Enol Vitic 50th Anniv Annu Meeting, Davis, CA, American Society for Enology and Viticulture, pp. 323±336. SINGLETON VL, TROUSDALE E and ZAYA J (1979), Òxidation of wines. I. Young white wines periodically exposed to air' Am J Enol Vitic, 30, 49±54. SIVERTSEN HK, FIGENSCHOU E, NICOLAYSEN F and RISVIK E (2001), `Sensory and chemical changes in Chilean Cabernet Sauvignon wines during storage in bottles at different temperatures' J Sci Food Agric, 81, 1561±1572. SKOUROUMOUNIS GK, KWIATKOWSKI M, SEFTON MA, GAWEL R and WATERS EJ (2003), Ìn situ measurement of white wine absorbance in clear and in coloured bottles using a modified laboratory spectrophotometer' Aust J Grape Wine Res, 9, 138±148. ROUSSIS IG, LAMBROPOULOS I

SKOUROUMOUNIS GK, KWIATKOWSKI MJ, FRANCIS IL, OAKEY H, CAPONE DL, DUNCAN B,

SEFTON MA and WATERS EJ (2005), `The impact of closure type and storage conditions on the composition, colour and flavour properties of a Riesling and a wooded Chardonnay wine during five years' storage' Aust J Grape Wine Res, 11, 369±384. SOBIESKI DN, MULVIHILL G, BROZ JS and AUGUSTINE MP (2006), `Towards rapid throughput NMR studies of full wine bottles' Solid State Nucl Magnet Reson, 29, 191±198. SOMERS TC and ZIEMELIS G (1985), `Flavonol haze in white wines' Vitis, 24, 43±50. È HN T (1996), Àging of wine: 1,1,6-trimethyl-1,2-dihydroSPONHOLZ WR and HU naphthalene (TDN) and 2-aminoacetophenone' in Proceeding of the Fourth International Symposium on Cool Climate Viticulture and Enology, Geneva, New York, NY State Agricultural Experimental Station, pp. VI-37±57. TIMBERLAKE CF and BRIDLE P (1976), Ìnteractions between anthocyanins, phenolic


570


compounds, and acetaldehyde and their significance in red wines' Am J Enol Vitic, 27, 97±105. TOMINAGA T, GUIMBERTAU G and DUBOURDIEU D (2003), `Role of certain volatile thiols in the bouquet of aged Champagne wines' J Agric Food Chem, 51, 1016±1020. VANDERHAEGEN B, NEVEN H, VERACHTERT H and DERDELINCKX G (2006), `The chemistry of beer aging ± a critical review' Food Chem, 95, 357±381. WALLING C and JOHNSON RA (1975), `Fenton's reagent. V. Hydroxylation and side-chain cleavage of aromatics' J Am Chem Soc, 9, 363±367. WATERS EJ, PENG Z, POCOCK KF and WILLIAMS PJ (1996), `Lacquer-like bottle deposits in red wine' in Proc 9th Aust Wine Ind Tech Conf, Adelaide, Australia, Winetitles, pp. 30±32. WEEKLEY AJ, BRUINS P, SISTO M and AUGUSTINE MP (2003), Ùsing NMR to study full intact wine bottles' J Mag Reson, 161, 91±98. WHITE RE (2003), Soils for Fine Wines. New York, Oxford University Press. WILLIAMS AA, LANGRON SP, TIMBERLAKE CF and BAKKER J (1984), Èffect of color on the assessment of ports' J Food Technol, 19, 659±671.


19 The stability and shelf life of fruit juices and soft drinks P. Ashurst, Ashurst and Associates, UK

Abstract: This chapter considers the microbiological, physical and chemical factors that affect the stability of fruit juices and soft drinks, and the impact those factors have on the shelf life of the products. It looks at the impact that various processing techniques have in extending product shelf lives. Packaging plays a critical role in containing beverages and extending their shelf life and the use of various types of packaging is considered. Shelf life determination and accelerated testing are considered briefly as are possible alternatives to thermal pasteurization. Key words: fruit juices, soft drinks, processing techniques, pasteurization, packaging, shelf life, product stability.

19.1

Introduction

Unlike some food items, fruit juices and soft drinks cannot exist as articles of trade without the packaging that is needed to contain the product, protect it from deterioration, loss and damage and provide a vehicle to advise the consumer of the contents and other essential information. The interaction of product and packaging is thus key to ensuring that these products reach the consumer at the level of quality that the manufacturer intended. In addition to containing the product, the packaging has other important functions which include making the product look attractive to fulfil a marketing objective. In most countries it is now a statutory requirement for the consumer to be advised of the shelf life of the product and this is normally provided by printing the product's shelf life as a `best before' date. As an alternative, a ùse by' date may be employed in products where there is a potential food safety issue. This


572


latter term is only used (in the context of this chapter) for freshly squeezed juices that have a very short shelf life which is maintained by cold temperature distribution. The determination of the shelf life of a product is an essential part of its development phase and there are many considerations that must be taken into account when the shelf life is set. These include: · compliance with any statutory requirement for an indication of product durability · to ensure compliance with any nutritional claims made · to meet customer demands and minimize the risk of product failure and writeoff · to inform the distribution, marketing and retailer demands. Technically, the shelf life is established as the period of time in which the particular combination of product and packaging retains the quality and taste set by the manufacturer. Any processing that the product receives is an integral part of the shelf life determination. Any substitution of raw material, packaging or processing can then be evaluated to ensure maintenance of the desired shelf life. It is very important for manufacturers to understand any changes that occur within the predetermined shelf life under various conditions that should include all anticipated markets. Overall, the shelf life of a beverage is an important indicator of production consistency and the confirmation of quality systems. 19.1.1 Fruit juices At their simplest, fruit juices are obtained by pressing fruit of appropriate ripeness and collecting the expressed juice. Many consumers still carry out this operation in homes and other places where the juice will be consumed within a matter of hours. In such circumstances, the issue of the shelf life and stability of the juice is of little relevance. However, fruit juices are now sold across the globe and represent a multi-billion dollar industry that is dominated by orange juice. In order to transport, package and sell fruit juice to the consumer, its shelf life and stability are of major importance to both ensure that the consumer is provided with a product of acceptable quality and to avoid the loss of product of substantial commercial value. Fruit juices are generally available in various forms. There is a rapidly increasing market sector for fresh juice that is described as `not from concentrate' (NFC). This sector is divided into product that is unpasteurized and must be sold through a cold distribution chain and consumed within a few days of pressing, and product that is subjected to a `light pasteurization' (typically a few seconds at around 90±92 ëC) and packed into clean, but not aseptic, containers that are also distributed through a cold distribution system but with a shelf life of several weeks. The major market sector for fruit juices is based on product that is reconstituted from juice concentrated in the country of processing, shipped to the main market where water is added to reconstitute the typical composition of


The stability and shelf life of fruit juices and soft drinks 573 fresh juice and packaged aseptically into containers that allow a shelf life of up to one year. In all these sectors, the related issues of the stability of the product and its shelf life are of the utmost importance to producer and consumer alike. This chapter will examine in more detail the outline processes that are involved in the production of fruit juices, the treatments needed to achieve the required stability and the factors that will impact on the resulting quality and shelf life of the product. 19.1.2 Soft drinks Traditionally, soft drinks were prepared by dissolving granulated sugar in specially treated water, or alternatively diluting liquid sugar with this water. A variety of ingredients including preservatives, flavouring and colouring agents, carbon dioxide and acidulants (invariably either citric or phosphoric acid) were then added. Other constituents such as fruit juice or comminuted fruit, artificial sweeteners, antioxidants, ingredients to deliver clouding and foam were added, depending on the particular product being made. More recently, `diet' soft drinks in which the sugar has been replaced with an artificial sweetener (typically aspartame or sucralose) have become very popular. Soft drinks are now prepared almost exclusively using the pre-mix system whereby the blended syrup, prepared using ingredients outlined above, after flash pasteurization if necessary (85 ëC for 15±30 seconds would be typical), is mixed in appropriate proportions with treated water prior to delivery to the filler. If the end product is to be carbonated, it has traditionally been cooled to 1±3 ëC before arrival at the filler in order to minimize loss of carbonation and facilitate filling. Fillers and ancillary equipment capable of handling the product at ambient temperatures have recently been introduced. Most food and drink products are supplied to consumers in some form of primary packaging and in many cases secondary packaging as well. Beverages, however, are totally dependent on primary packaging as a means of containing the product. This is true for all types of beverages but for carbonated drinks the packaging plays a greater role by retaining not only the liquid but also the CO2 that gives the product one of its principal characteristics. This chapter provides an insight into the processes used to manufacture and package soft drinks and the factors that affect their stability and shelf life. 19.1.3 Packaging Development of the beverage industry for fruit juices and soft drinks is paralled by the development of suitable packaging that provides physical support for the product in addition to providing a reasonable shelf life for the contents. All early long life fruit juices and carbonated products were packaged in glass which even today provides a performance benchmark for product protection despite its principal disadvantages of weight and brittleness. Today, a significant proportion of all beverages are packed in either some form of plastic container,


574


plastic laminated paperboard or other flexible packing most of which have only made a significant contribution to the markets during the last 30 years. Metal cans still provide an important alternative to the other types of packaging. The packaging of carbonated products is, with few exceptions, limited to glass, metal cans and plastics. It is self-evident that the primary function of any beverage packaging, which must include the closure, is to provide the physical retention of the contents. The consumer expects to retain the amount of product that he or she purchased until the time for its consumption. Container leakage may also result in damage to other property. The primary evaluation of beverage packaging is thus concerned with the retention of liquid content. This performance characteristic is determined not only by the container itself but by the effectiveness of the seal between container and closure. It is rare for containers themselves to be produced with a defect in the body that permits leakage. For packages that are produced on line such as in `form-fill-seal' operations, there is a significantly increased risk of the failure of seals and quality checks need to be strengthened accordingly. Assuming the contents of a beverage container are retained satisfactorily, there are other quality attributes that determine the suitability or otherwise of the beverage/packaging combination. In most countries there is now a statutory requirement to state some form of product durability marking on the label and this is usually in the form of a `best before' date although with some very short shelf life products, such as freshly squeezed juices, a ùse by' date is more appropriate. The performance of any package is thus measured by its ability to keep the contents in a condition that is as close to the taste, appearance and nutritional qualities or other standards as is required by the manufacturer (and expected by the consumer) within the period between the dates of manufacture and expiry. Although this chapter is not primarily concerned with packaging, its importance in ensuring the required product stability and shelf life cannot be ignored.

19.2 Factors influencing the stability of fruit juices and soft drinks 19.2.1 General The factors that affect the stability and shelf life of fruit juices and soft drinks may be conveniently divided into three main areas: microbiological, physical and chemical. Soft drinks and especially fruit juices are particularly susceptible to microbial spoilage and, in certain circumstances, may support the growth of pathogenic organisms. A rapid deterioration may occur if products are not pasteurized. Where microbial deterioration does occur physical and chemical changes are almost always evident. The physical and chemical changes that are discussed later refer to the changes that may be seen when no microbiological activity is present.


The stability and shelf life of fruit juices and soft drinks 575 19.2.2 Fruit juices The processing details of fruit juices are well documented and it is not proposed to go into much detail in this chapter except insofar as process is likely to affect the stability of the end product. Details vary depending on the botanical structure of the fruit being handled and diagrams of typical processes for soft fruit such as apples and pears and citrus appear in Figs 19.1 and 19.2. The key stages in any fruit processing operation are: · Washing and inspecting to reduce microbial load and remove rotten fruit. · Pressing to obtain optimum juice yield that is compatible with the quality required. · Separating juice from pulp and debris. · Pasteurizing juice to deactivate enzymes and secure microbial stability. · Concentration (if required). · Packaging into appropriate containers and storage/distribution.

Fig. 19.1 Outline typical process for soft fruit processing.


576


Fig. 19.2 Outline typical process for citrus fruit processing.

The relevance of these stages to the stability and shelf life of the end product is discussed below. Washing and inspection Most fruit processing operations incorporate some form of washing into the delivery and transport of fruit into the facility. Incoming fruit is often discharged into pits that use a water flume to provide a gentle means of transporting the fruit to the plant. The circulating water gives an initial washing stage that will usually remove gross soil and other debris. When fruit reaches the main facility it will typically be elevated into an inspection area. The elevators usually incorporate a


The stability and shelf life of fruit juices and soft drinks 577 spray of clean water which may also incorporate a mild sterilant. Inspection in some countries is a manual operation where rotten fruit and other debris are removed. Automatic inspection facilities are often used in larger operations. The microbial load on most raw fruit is considerable and an effective washing operation will assist in reducing this load. Since in the pressing and separation stages, juice and pulp/peel are intimately mixed, high levels of initial contamination are likely to be found in the raw juice. Most subsequent pasteurization stages are likely to be based on typical microbial loads and excessive contamination may not be removed leading to later problems. Removal of damaged fruit also assists with the reduction of microbial load and, since damaged fruit may also be partly fermented, its presence, if not removed, may later affect the aroma of juice. Pressing to obtain optimum juice yield that is compatible with the quality required Many different types of fruit presses are used but all involve compressing the fruit to a greater or lesser degree. Soft fruit processes often involve a pretreatment with enzymes to break down the structure of fruit and release as much juice as possible. The development of off-flavours is rarely associated with such treatments but physical stability problems may arise. These include haze or sediments in clear juices or separation of cloudy products. Where clear juices are to be the end product, a later enzyme treatment is normal to assist with clarification. It is possible, by use of an appropriate cocktail of enzymes to procure an almost complete liquefaction of soft fruits such as apples and pears although the legal status of the juices so produced is unclear. Because of the botanical structure of citrus fruit, it is desirable to extract peel oils before juice pressing is undertaken although some extraction methods allow peel oil to be removed later. Peel oils have intense flavour and aroma characteristics which are not generally appropriate in citrus juices although they are likely to be added to citrus comminutes. Apart from the initial aroma and flavour impact of citrus oils, their terpene content renders them vulnerable to rapid oxidation and consequent flavour deterioration. The use of undue pressure in the pressing of oranges is likely to lead to a significant increase in the level of socalled peel extractives. All citrus juices are likely to have a low level of these but higher than normal levels may lead to later problems of chemical or physical stability that will affect shelf life and product authenticity. Citrus comminutes are manufactured by recombining the principal citrus fruit components (juice, peel and oil) in proportions that deliver the characteristics required by the purchaser. They are used as ingredients in citrus drinks, mainly dilute-to-taste products, with other ingredients such as antioxidants and preservatives which limit the rate of deterioration of citrus oils. Separating juice and pulp and debris In most processing operations of all types of fruit, it is necessary to separate juice from the matrix of vegetable matter that forms the structure of the fruit. A


578


number of alternative methods are used to achieve this, including simple screening and the use of centrifuges or rotary separators. These are typically horizontally aligned rotating drums that have perforated walls. Juice is allowed to pass through the perforations whilst pulp and debris are retained. By inclining the drum at a slight angle, the residual material is made to fall from an aperture at the end of the drum onto a moving belt that removes it from the area. Some processors then incorporate the further use of a centrifuge to reduce the pulp content to a predetermined level. Excessive pulp levels in the finished product can have a significant effect on the appearance and mouthfeel of the product and in that way may render it unacceptable to consumers. Pasteurizing juice to deactivate enzymes and secure microbial stability Pasteurization is used for products that have a pH value of 4.5 or less, where the acidic conditions effectively reduce the risk of growth of pathogenic organisms. However, some care needs to be exercised with products that have pH values that are in the region of 3.9±4.5 as it has been demonstrated that some pathogens (e.g., E. coli O157) can survive for a limited period in such conditions. For bulk juice treatment, the normal process used is flash pasteurization, although for product directly packed into consumer units, either flash pasteurization (85±90 ëC for 15±20 seconds is a typical range) coupled with an aseptic packaging unit or in-pack pasteurization of filled products (20 minutes at 70 ëC) may be used. To determine the level of pasteurization needed (in-pack or flash treatment) it is normal to refer to the number of pasteurization units needed. No effective pasteurization occurs below 60 ëC and by holding a product at that temperature for 1 minute it is said to have received 1 pasteurization unit (PU). To calculate the number of pasteurization units for any given time temperature relationship, the following formula applies: Number of pasteurization units 1.389(t-60) time in minutes As an example, a product is held for 20 minutes at 70 ëC (typical in-pack pasteurization conditions). Thus the number of PUs is: 1.389(70ÿ60) 20 (i.e. 1.389(10) 20) 26.73 20 534.6 PUs A simplified approach is to accept that for every rise of 7 ëC over 60 ëC, there is an approximate 10-fold increase in the number of PUs, i.e. 60 ëC 1 PU/ minute; 67 ëC 10 PU/minute; 74 ëC 100 PU/minute, etc. It is, however, difficult to interpolate if the simplified approach is used. Unpasteurized fruit juices have a very limited shelf life of a few days and must be marketed through a cold chain distribution service. They are subject to problems of physical separation and will show rapid microbial spoilage if not


The stability and shelf life of fruit juices and soft drinks 579 stored correctly or are held beyond their use by dates. Despite their relatively low pH, unpasteurized juices are now known to have the potential to support pathogenic microorganisms such as Salmonella, Shigella and Escherichia coli O157. Other than for this relatively small market sector, pasteurization is a critical step in the processing sequence of all fruit juices. Two objectives are secured by this process, an improvement in microbiological stability (and removal of any pathogens) and, provided a temperature of 90 ëC is reached, the destruction by heat of pectolytic enzymes that will otherwise promote physical separation into an upper clear layer and a lower layer of pulpy debris. The stabilization of enzyme activity may not be required if the juice is clarified as it will probably have been through a process that requires the addition of appropriate pectolytic and sometimes other enzymes. Fruit juices that are to be packed without concentration will often be suitably heat treated as part of the overall production and packaging process but juices that are to be concentrated may not necessarily receive the full pasteurization conditions if they are pre-heated as part of the evaporation and concentration stage. Concentrated juice produced in this way will normally be stored frozen, typically at around ÿ18 ëC. Any residual microbiological activity will not normally cause problems at such a temperature and since, when the juice is reconstituted, it will again be pasteurized, no further problems of microbiological activity will be expected. However, if during the concentration process, the juice temperature does not reach that required to deactivate enzymes (at least 90 ëC), they may remain active, albeit at a much reduced rate, during storage and may occasionally become the cause of impaired physical stability when the juice is reconstituted. It should also be noted that any significant delay in pasteurizing juice during the process will allow enzyme activity to proceed and may, even with later apparently adequate heat treatment, result in juice which shows problems of physical stability. Packaging into appropriate containers and storage/distribution The packaging of fruit juices into containers for direct use by the end consumer is the most critical stage in ensuring that the product has the intended shelf life and stability. The use of any containers that are intermediate between the juice producer and final packaging plant also represents a critical stage. Freshly squeezed juices There is a significant market for juices produced by pressing fruit, mostly citrus, placing it in small containers, and selling directly to the consumer without any further heat treatment other than refrigeration. The juice so produced has the freshest possible taste but carries significant risk of microbiological infection and lack of physical stability. The microbiological risk is principally that of spoilage by yeasts and moulds although, as indicated above, there is increasing concern about the ability of certain pathogens to survive in what has been traditionally regarded as a low pH


580


(below 4.5) product. In consequence, this sector of the juice market relies on distribution in a cold chain, sales from refrigerated cabinets and has a very short shelf life of up to about 7 days. Generally, consumers appear willing to accept some physical separation and provided correct distribution and storage temperatures are maintained, microbiological problems are minimized. Direct or `not from concentrate' (NFC) juices NFC juices are products that are obtained by pressing fruit, separating from pulp and debris to the required level then pasteurizing and packaging into containers for consumer use. There may be an intermediate stage where juices are pasteurized and packed into bulk containers under aseptic conditions. Where this occurs the double pasteurization will lead to products that have a slightly less fresh taste. Where juices are packed directly into the consumer packs, they will be typically subjected to a `light' pasteurization, which may involve heating to around 90 ëC for a few seconds, before being packed under clean but not aseptic conditions. Like freshly pressed juice, these products are also stored, distributed and sold under refrigerated conditions and have a typical shelf life of up to 12± 13 weeks. The pasteurizing conditions will inactivate enzymes and thus maintain physical stability and provide a high degree of microbiological protection from spoilage. Pasteurization at the relatively low pH of juice also effectively removes any risk of pathogen survival. Concentrated juices The largest volume of fruit juices are manufactured by processing into concentrates at the fruit processing location, transporting the concentrate to the required packaging location where it will be reconstituted to a standard strength and repackaged as required. The concentration process usually involves heating the juice and evaporating most of its water content under vacuum, although other techniques such as freeze concentration are used on a small scale. In a standard concentration process, the juice may or may not be subjected to the necessary pasteurization temperature prior to the commencement of evaporation. It is therefore possible for concentrated juices to carry a residual active microbial load and, more particularly, to contain active pectolytic enzymes. Citrus juices such as orange and grapefruit are normally concentrated to around 65±66 ëBrix and frozen to around ÿ18 ëC for transportation. Any microbiological activity will be largely suppressed. When reconstituted, most juice is repackaged aseptically which will ensure microbial stability. A smaller proportion of reconstituted juice is marketed through cold chain distribution in similar packs to NFC products and only receives a `light' pasteurization. Pectolytic enzyme activity may continue in the concentrate at a low rate leading to the possibility of reconstituted juice which may not show the required level of physical stability. However, most aseptic packs, such as the 1 litre Tetrapak or Combibloc containers, do not allow the consumer to see the product before it is dispensed and in consequence any loss of physical stability is rarely noticed.


The stability and shelf life of fruit juices and soft drinks 581 The issue of pectolytic enzyme activity and resulting latent physical stability is not normally an issue in pome and berry juices concentrated to around 70 ëBrix. These juices are invariably clarified before concentration and transported in cool but not frozen conditions. The reconstituted juices are also marketed as clear products and packed aseptically. 19.2.3 Physical and chemical changes Whilst microbiological stability is the most critical factor in determining the stability of fruit juices and soft drinks and must be effectively achieved in all products relative to their shelf life, the issues of physical and chemical stability then become significant. Physical stability normally refers to the appearance of a product and includes its propensity to separate into layers or the appearance of sediments, floc or particulates as well as any colour changes. Most of the physical changes that become apparent during the shelf life of a product are related to a chemical or biochemical reaction and it may be more appropriate to refer to this aspect of product stability as `physico-chemical' rather than to being either strictly physical or chemical. Key factors in such changes are the presence of any residual pectolytic or other enzymes, the presence of dissolved oxygen and the temperature at which products are stored. Changes in the appearance of a product relate particularly to its intended nature as being either cloudy or clear. These changes may be ameliorated to some extent by the use of packaging that prevents or limits the consumer from seeing all or part of the product until it is dispensed. Cloudy products Most fruit juices are of a cloudy nature when immediately pressed from the fruit and many consumers prefer cloudy to clear products for certain juice types. For example, there is little if any market for the sale of clear orange or grapefruit juice or related soft drink products whereas lemon and lime juice based products are frequently clarified. Apple and berry juices are also often clarified although `not from concentrate' apple juice is almost always cloudy. Whilst the factors that determine whether a product is to be marketed as either clear or cloudy are mainly subjective, the defects that may occur relating to the intended appearance are not. The main concern relating to the stability of cloudy products is the separation of the product into sediment, which may be pulpy or compacted, and a clear or almost clear supernatant layer. For pure juices and products with a significant juice content such separation is related mainly to the breakdown of natural pectin that is present in variable quantity and quality in juice expelled from fruit. Pectins vary in composition but are essentially methylated polygalacturonic acids. They play an important function in cell wall structures but once cells are ruptured as in the expression of juice, pectolytic enzymes start to break down the chain length as well as initiating the hydrolysis of the methyl esters. The ultimate result is to increase the amounts of galacturonic and polygalacturonic


582


acids in the juice. In pure juices pectins are responsible for the cloud stability of the juice and any breakdown of pectins to galacturonic acids with various levels of demethylation and chain length will adversely affect cloud stability. This breakdown occurs quickly after juice expression as a result of pectolytic enzyme activity which is halted by rapid pasteurization of the juice to around 90 ëC. Thus loss of cloud stability in a pure juice or juice-based product usually relates to delay in pasteurization of the juice component. A clouding preparation may be used in products required to be cloudy but where there is no juice present. Such clouding preparations products are often oil-in-water emulsions produced using citrus oils emulsified with wood rosin ester or permitted synthetic emulsifying agents and an aqueous component incorporating, for example, gum Arabic. These preparations are mostly very stable provided the particle size is kept below about 10 microns, although their stability can be adversely affected by heat and the presence of solvents such as ethanol resulting from the addition of types of flavourings. Clarified products As indicated above, there are some products that consumers expect to purchase clear. In such products the level of clarity must be outstanding as the presence of haze, floc or precipitates will often result in rejection and may be regarded as an index of product failure. Clarification of the juice component is usually achieved by addition of pectolytic and amylolytic enzyme preparations to ensure removal of pectin and starch residues. These residues are frequently the source of haze or precipitates, although polyphenolic substances present may further polymerize on storage and produce sediment. Physical defects in clarified products are thus mainly those which produce haze or other sediments and may be caused by an inadequate treatment. After addition of enzymes, it is often necessary to carry out further treatment such as the addition of gelatin or bentonite to remove any residues of protein that may also be present. The use of a polishing filter is also often employed. Assuming that the product is free from any microbiological contamination that may produce a haze, the later appearance of any haze or other residue may be related to breakdown of other juice components. For example, in juices with a significant red colour, breakdown of anthocyanins may give rise to polyphenolic material that has a sufficiently large molecular weight to come out of solution and produce sediment. Floc can sometimes arise as a result of algal blooms contaminating the water source, although this phenomenon is very unusual and normally only seen during high temperatures and periods of extended sunshine. Other physical and chemical changes Oil rings Products that contain essential oils as part of a flavouring system sometimes suffer separation of some or all of the oily components which then form a socalled neck ring. This is evident as a ring of often coloured oily materials in the


The stability and shelf life of fruit juices and soft drinks 583 neck of a bottle which may sometimes be redispersed by agitating the bottle contents. Essential oils, particularly those from citrus fruits may be added to beverages in the form of a cloudy emulsion where the oils are made to form an oil-in-water emulsion using appropriate permitted emulsifiers and then subjecting the mixture to high pressure emulsification. These cloud emulsions are normally very stable but can sometimes break down if subjected to high pasteurizing temperatures for extended periods (such as may occur when a flash pasteurizer is required to recycle product for 5±10 minutes) or if there is a significant level (over about 0.1% by volume) of alcohol in the product. Colour and flavour changes One of the more commonly observed defects in almost any soft drink or fruit juice is that of product discoloration. Products will be evaluated to establish a shelf life that enables the product to remain of an appearance and taste that the manufacturer decides will be acceptable to consumers. However, all soft drinks and fruit juices will show colour and flavour changes over a period of time which may be slowed or accelerated by the effect of storage conditions. Many manufacturers set the shelf life of their long life products based on the average conditions expected during warehousing and distribution. Typically, it will be assumed that in temperate climates this will be a year round average temperature of between 10 and 20 ëC and conditions of low or no light. Producers in or exporters to tropical or other extreme climates will need to evaluate the range of expected conditions and decide appropriate storage and distribution arrangements to ensure acceptable product quality during the stated shelf life. The most usual colour and flavour defects that are likely to arise in products that have been subjected to inappropriate storage temperatures or excessive pasteurization are the development of browning in the appearance coupled with a pronounced cooked flavour and aroma. In this respect the use of tunnel pasteurization or hot filling carries a greater risk of thermal damage to products than flash pasteurization. Although temperatures are greater in the latter case, it is exposure to longer time at increased temperature that is invariably more damaging. During the period of product development, storage for long periods and at excessive temperatures can be used to establish the likely deterioration of colour and flavour. Any unusual colour or flavour that arises, perhaps as a complaint, can then be readily distinguished from the norm. Light that is allowed to penetrate a product, for example in a clear glass container, is usually very damaging and a product left in direct sunshine will often develop unpleasant off-tastes and suffer colour changes, usually bleaching, in a very short time. This aspect of product deterioration should also be evaluated during the development period by exposing product to both south and north light. Unusual colour development or the presence of an uncharacteristic flavour taint will probably require the services of a specialist laboratory to establish the cause, although much can be done by a close evaluation of the individual ingredients, including water used.


584

19.3


Ensuring product stability and extending shelf life

19.3.1 General principles The most important means of ensuring product stability and delivering a suitable shelf life is by the use of appropriate processing and packaging. Processing It will be evident from earlier sections in this chapter that the processing required for any given product will be determined by its ingredients, intended market, shelf life and packaging to be used. For example, at one end of the continuum will be freshly squeezed fruit juices where, apart from expelling the juice from the fruit and carrying out a simple separation to remove unwanted pulp and other parts of the structure of the fruit, there will be no other processing. Packaging in such instances will usually be a clear plastic (often PET) bottle to allow the consumer to see the product at the point of purchase. The typical shelf life of up to about 7 days will be achieved by storage, distribution and display in cold (2±5 ëC) conditions. At the other end of the continuum will be long life fruit juices which are flash pasteurized, probably at over 90 ëC, followed by immediate packaging into laminated cartons under aseptic conditions to give a shelf life in temperate conditions of at least 12 months. Leaving aside for the moment the issue of packaging, the processing required for any given product can be established by a risk assessment based on the ingredients to be incorporated (including the quality and quantity), the conditions under which it is to be packed, the use or not of permitted preservatives and the shelf life and market conditions for which it is intended. The highest risk ingredients are those which provide nutrients for the growth of microorganisms, although it must be remembered that because of the broad spectrum of conditions under which bacteria and to a lesser extent yeasts and moulds will grow, all soft drinks and fruit juices are at risk to some extent from microbial deterioration. On the assumption that all the products under consideration in this chapter are of a pH lower than 4.0, the greatest microbial risk is that from spoilage by yeasts and moulds. In practical terms, it will be the presence of fruit juices (especially pome and berry juices) at almost any level, but especially over 2% in single strength, that present the highest risk. This risk may be mitigated to a limited degree by the use of concentrated juices, particularly aseptically packed concentrates, but in all instances juice containing beverages require pasteurization. Carbohydrates provide a similar nutrient base for yeasts and moulds but provided that they are of high quality with a low residual count of organisms and there is no separate source of nitrogen, products can be successfully made without pasteurization but with permitted preservatives. The lowest risk products are those that contain synthetic flavours, artificial sweeteners and carbon dioxide as, provided water quality used is at least up to the standard of drinking water in the EU, there is minimal risk of microbial


The stability and shelf life of fruit juices and soft drinks 585 deterioration and no processing is required. If packaging is sterile and equipment hygiene is outstanding, unpreserved products can be successfully made although the prudent manufacturer will always use a permitted preservative in even the lowest risk products. The use of carbon dioxide as an ingredient is worthy of particular mention as it performs a number of significant functions that could be described as `preservative'. Its presence in the headspace of a product container effectively excludes oxygen which substantially rules out mould growth in the product. It also reduces the pH of products in which it is used and as it is a normal byproduct of the metabolism of most yeasts, can be effective in suppressing yeast growth. When processing is deemed necessary for a fruit juice or soft drink, almost the only practically available solution for high volume production at the time of writing is the use of thermal pasteurization. The choice of flash or in-pack treatment or hot filling will largely depend on the packaging required coupled with the facilities available to a given manufacturer. In all instances, the level of pasteurization will be determined by the product risk factors. Typical pasteurizing conditions in use are flash pasteurization at around 85±90 ëC for a period of 3±30 seconds or in-pack treatment at 70 ëC for 20±25 minutes. Other potentially available treatments are discussed in Section 19.5. Packaging For all practical considerations, beverages cannot exist without packaging, except for products for almost immediate consumption; it is this factor more than any other that determines the shelf life of the product. Packaging must selfevidently be capable of physical containment of the product and, if carbonated, the carbon dioxide that gives the product its particular characteristic. It must then primarily protect the contents from microbial contamination and the effects of oxygen which will often produce rapid deterioration. Packaging must also be capable of ensuring adequate physical protection of the contents and is an essential means of communicating both statutory and marketing information to the consumer. Glass For many years glass has been the standard against which alternative packaging materials have been evaluated. Glass containers with an appropriate closure provide almost perfect protection to fruit juices and soft drinks within the agreed shelf life. The most likely cause of failure of glass containers is leakage at the interface between the cap and bottle body. Assuming a satisfactory sealing at the closure interface, there is no loss of contents by transmission of water vapour, carbon dioxide retention is excellent and there is virtually no transfer of oxygen into the product. The advantages of glass containers include its quality image, potential for brand differentiation because of relatively low tooling costs, tamper evidence, recyclability and reuse possibilities. Glass has the added benefit of being practically inert and apart from any risk of contamination of product by


586


glass fragments or misuse of the container for storage of foreign substances, there is minimal risk of tainting. Glass can also be easily coloured or covered with a UV screen to minimize the effect of light on sensitive products. The disadvantages of glass are its weight and its brittleness that render glass containers liable to breakage with the risk of damage to people or property. Few manufacturers now collect, wash and reuse glass containers because of the associated cost and logistical problems, but glass can be easily recycled and most EU countries are now well advanced in schemes for collection of used glass containers for reprocessing as cullet. The effects on shelf life of products packed in glass are thus effectively confined to the physico-chemical changes that occur as a result of time and accelerated by storage at increased temperature, the transmission of light and the presence of dissolved oxygen. Intense light, such as direct sunlight, is extremely damaging to most products; indirect light is less so. The effects of light on most products are to cause breakdown of flavour components resulting in often very unpleasant off-flavours. Metal containers The market for soft drinks packed in cans has declined in recent years but despite that, they remain popular for single serve beverages because of their convenience, robustness and marketing image. Most beverages are now manufactured in aluminium as two piece cans (can body and one end) by the drawn and wall ironed (DWI) process. Such two-piece cans lack the body side seam of their three-piece predecessors and are inherently less prone to leakage of liquid and CO2. Provided the single can end is applied correctly, the risk of leakage of either liquid or gas is minimal. Probably the greatest risk of leakage is likely to arise from can corrosion and eventual pinholing and, if cans are subject to tunnel pasteurization, removal of surface moisture is vital to minimize corrosion. Like glass, cans offer almost perfect protection for soft drinks and, to a lesser extent, fruit juices. When correctly sealed, modern two-piece cans give outstanding retention of contents, prevent ingress of oxygen and are much lighter than glass and are not brittle. The contents are given total protection from the effects of light. Their use today is mostly limited to carbonated drinks because the presence of CO2 provides added physical stability to the can thus enabling a much thinner wall thickness to be used. Canned fruit juices and non-carbonated soft drinks, particularly those requiring tunnel pasteurization, must be injected with a small amount of liquid nitrogen to pressurize the contents and thus prevent can distortion. Like glass, cans may be readily recycled. The principal disadvantages of cans are that they are not easily used for carbonated drinks in volumes in excess of about 500 ml and for soft drinks they rarely exceed 330 ml. Because cans are usually decorated and printed at the time of manufacture and because modern filling lines often run at speeds in excess of 1000 cans per minute (which yields around 300 million units per annum), a substantial investment in stock with its corresponding logistical consequences is required.


The stability and shelf life of fruit juices and soft drinks 587 The effects on shelf life of products packed in cans will, as with any other container, depend on time and temperature, but close attention must be paid to the detail of formulations for canned products and a correspondingly close liaison must be maintained between product formulation teams and the can maker. Any lacquer applied to the internal surface of cans must be compatible with the formulation. It is essential to ensure complete absence of sulphur dioxide from products as, if it is present, chemical reduction is likely to occur at the metal interface with the production of hydrogen sulphide. Perhaps the most significant risk to shelf life in the use of cans is that of metal uptake and the corresponding development of metallic off-tastes. In extreme circumstances, the level of metals could exceed statutory limits. Plastic containers A high proportion of soft drinks and, to a lesser extent, fruit juices are now packed in plastic containers. Whilst there are a number of plastics that have been used for this purpose, the predominant material now in use is polyethylene terephthalate (PET). An alternative, polyethylene naphthalate (PEN) is sometimes used for products requiring in-pack pasteurization as it is much less liable to distort at the temperatures used for pasteurization. Other plastics such as high density polyethylene (HDPE) and polystyrene (usually with co-polymers) find uses in very short shelf life products and form-fill-seal packs such as cup drinks. Since PET is predominant in beverage packaging, the comments in this section relating to shelf life issues will be made in respect of that material. The main advantages of using PET containers are its relative strength, particularly when carbonated, low weight and a clarity that compares favourably with that of glass. Bottles made of this material can also be recycled. The principal disadvantages of PET are its permeability to gases allowing the loss of CO2 and ingress of oxygen. Carbonated products packed in PET are normally given a significantly shorter shelf life than those packed in either glass or metal containers because, apart from the usual factors that determine shelf life, the loss of CO2 will render the product unacceptable to consumers in a shorter time. Manufacturers of carbonated drinks normally specify a maximum loss of 15% over 26 weeks for a 1.5 or 2.0 litre bottle. For smaller sizes, the 15% loss is likely to occur over 10±12 weeks because of the less favourable surface area to volume ratio. For this reason, the limiting shelf life factor for most carbonated drinks packed in PET is the loss of CO2 which is a determining characteristic of these products. Non-carbonated soft drinks packed in PET are susceptible to oxidative deterioration because of the ingress of oxygen through the packaging walls. It will thus be desirable to incorporate an oxygen scavenger such as ascorbic acid into the formulation of products packed in PET. Laminated board containers A significant proportion of fruit juices and non-carbonated soft drinks are now packed aseptically in laminated board packs as exemplified by the use of


588


TetraPak, Combibloc and other similar packs. The material used in these packs is typically a laminate which, starting from the outside, is often made up as follows: Polyethylene Printed design Paper/board Polyethylene Aluminium foil Polyethylene Polyethylene The pre-printed laminate is reel fed through a sterilizing bath into the aseptic part of the machinery where it is formed into the required shape that can be in a limited number of shapes, filled, sealed and discharged ready for any secondary packaging required. These containers provide a very high level of product protection that, in most cases is at least as good as glass (and in some instances better as light is totally excluded) coupled with a large surface for decoration and labelling. Because of their low weight and regular size, they also offer excellent distribution characteristics. The principal disadvantages of these containers are that the maximum size is generally limited to about 1 litre and the enclosed formulation must be free from any substances, such as sulphur dioxide, that might react with the aluminium layer. Close attention must be paid to the box formation and sealing that takes place within the aseptic chamber as any leaking or pinholing will compromise microbiological integrity.

19.4

Shelf life determination

The determination of shelf life relies to a large extent on the use of organoleptic testing which is covered in detail elsewhere in this book. However, in parallel, it is important that other testing is used to confirm the microbiological, chemical and physical characteristics of the product. 19.4.1 Microbiological All fruit juices, which are not permitted to contain preservatives, and soft drinks that are packed without the use of any preservatives, must be free from any microbiological contamination. This will normally be achieved by the use of either aseptic packaging, hot filling or in-pack pasteurization. Samples from each batch of product manufactured by these methods should be taken on an agreed basis and subjected to appropriate testing. Many manufacturers still employ classical plating techniques although some excellent rapid microbiological testing methods are now available. Products without any preservatives


The stability and shelf life of fruit juices and soft drinks 589 are often quarantined during the testing, and to ensure the required shelf life must be free from any microbiological contamination. Some manufacturers also subject products containing preservatives to a similar microbiological testing regime, although quarantining is not usually employed. Low counts, such as less than 10 colony forming units of yeasts and moulds per millilitre of product, may often be discounted particularly when products are carbonated, but the target must always be to obtain a zero count. Experienced technical staff will build up their own assessment of these products and when to release/quarantine or reject because in many cases very low counts will disappear within a short time because of the effects of preservatives. Yeasts and moulds are the organisms of greatest concern as they cause spoilage of the products that may include a build-up of CO2 that can, in extreme circumstances, cause bottle bursting and damage to people or property. Mould spores in a product are difficult to detect particularly by plating techniques and can develop after many weeks of storage. Most bacteria are of little concern provided that product pH values are less than 4 when most pathogens will cease to be viable. Spoilage from Lactobacillus species is not unknown and for non-carbonated ready to drink products Alicyclobacillus can be troublesome in warm conditions. Both these organisms are susceptible to the effects of flash pasteurization. 19.4.2 Physico-chemical testing Physical changes It is essential to ensure that products maintain their physical integrity during the storage period and, since some forms of plastic packaging allow water vapour to pass through the container walls, it will be desirable to ensure the quantity of product remaining in unopened containers at the end of shelf life is as stated on the label. Such losses may be accompanied by distortion of the container such as panelling. Products that are carbonated and packed in PET should also be checked for distortion of the container base particularly during storage in elevated temperatures. Base distortion will lead to the packages becoming unstable on shelves and cause rejection by retailers. Product colour changes during shelf life should also be noted and recorded. Although many coloured products lose `brightness' during storage, the effects of heat either from excessive pasteurization or during storage may cause browning to the extent that consumers may complain. Browning of products is often accompanied by the development of unwanted flavour defects. Colour changes can also arise as a result of interactions between ingredients although these are normally associated with products that contain added vitamins and especially minerals such as iron. Any appearance defects, such as physical separation or oil ring formation that occur during storage, must also be noted and assessed as being acceptable or leading to rejection of the product.


590


Chemical changes The principal concern for chemical changes that occur during storage of soft drinks and fruit juices will relate to products where there are specific claims for ingredients of nutritional value. In most beverages, protein and fat are either absent or present in negligible amounts so the main energy source will be carbohydrates. There may be changes in the carbohydrate spectrum, such as the inversion of sucrose to dextrose and fructose or the partial breakdown of glucose syrup ingredients, but there will be no overall effect on nutritional value. Small changes in sweetness levels may occur but will probably be undetectable to most consumers. Where additions of other nutrients have been made, for example vitamins, and a claim made for their content, it is very important to carry out testing to ensure that the declared level is present at the end of the stated shelf life. In most cases this will mean the addition of an overage at the time of manufacture. Experienced formulators will be likely to know what level of overage addition will be necessary for individual ingredients in particular formulations. Other ingredients that should be checked during initial shelf life assessments include artificial sweeteners and preservatives as any reduction in the amounts of these may have a more significant impact on the product performance. Accelerated shelf life testing Accelerated shelf life testing is often employed to obtain advance indications of the performance of newly formulated products and products destined for tropical markets. Exposure to north light (in the northern hemisphere) or a light box with appropriate wavelength light of the required intensity may be used for coloursensitive products, but the majority of accelerated testing is carried out at increased temperature. If sensitivity testing for light is carried out in a light box the light source may also cause a temperature increase and care must be taken to differentiate between the effects of light and heat in such situations. Robertson (2009) has recently discussed several models that have been developed to demonstrate the effects of heat on the deterioration of products. The most commonly used and generally valid relationship between the effects of temperature and the rates of deterioration of products is that of Arrhenius. The relationship may be shown as: k k0 eÿEa =RT where k rate constant for the deteriorative reaction, k0 constant independent of temperature (the Arrhenius factor), Ea activation energy (J molÿ1), R gas constant, and T absolute temperature. The simplified form expressing the relationship between the rate of deteriorative reaction and temperature is also shown as: k k0 ebTÿT0 where k0 rate at temperature T0 (ëC), k rate at temperature T (ëC), e 2.7183, and b a constant of the reaction.


The stability and shelf life of fruit juices and soft drinks 591 Another term used by Robertson to discuss the response of biological systems to the effects of heat is the temperature quotient. This indicates how much more rapidly a reaction proceeds at temperature T2 than it does at temperature T1. In practical terms, those concerned with the storage and deterioration of soft drinks and fruit juices will often use storage at ambient temperature (in the range of 15± 20 ëC) as typical of normal conditions and a warm room of 35±40 ëC to give an indication of the accelerated effects. At the elevated temperature, an approximate guide is that deterioration proceeds about 3±4 times faster than at ambient.

19.5

Future trends

19.5.1 Fruit juices and ready to drink uncarbonated soft drinks Recent years have seen a very rapid growth in the volume of fruit juices, and to a lesser extent soft drinks, sold through the cold chain distribution system. There is a limited market for so-called `freshly squeezed' juices where (mainly) citrus fruit is pressed, and with further processing limited to screening to remove seeds and some pulp packaged and sold within a few days. Despite their relatively low pH, such products do carry a risk of contamination by pathogenic organisms. The main growth area is for NFC juices and non-carbonated ready to drink soft drinks with a high percentage of fruit juice. The latter are unpreserved but, like the pure juices have been subjected to a `light' pasteurization. Otherwise the juices are almost indistinguishable from their freshly squeezed counterparts. These products will typically be allocated a shelf life of around 8±12 weeks and are typically packaged in laminated cartons. They carry a premium price and the development of their market may well be more limited during times of economic hardship, but otherwise seem set to be a growth feature of otherwise relatively static markets for fruit juices. Such products have a limited shelf life for mainly microbiological reasons, although the packaging employed does not give (and does not need to provide) the protection against oxidative deterioration that products with a longer shelf life need. At the other end of the storage spectrum are the long life fruit juices packed aseptically in laminated cartons. Such products are largely packed in 1 litre boxes, typically carry a shelf life of at least one year and account for a high proportion of all fruit juices sold. Juices packed in this way seem set to continue to dominate the market for many years to come. Despite the dominance of the aseptic pack, high quality adult soft drinks packed in glass continue to occupy a growing niche market where they attract premium prices coupled with long shelf lives. Products fortified with vitamins, minerals and phytochemical extracts (nutraceuticals) also occupy a small niche market and often show growth rates well in excess of mainstream products. Because of complex interactions of reactive ingredients in these products, colour and flavour changes often occur and shelf lives may be reduced accordingly. Non-carbonated RTD soft drinks in small aseptic packs (200±300 ml) are of increasing significance in the market as they offer the opportunity to produce


592


drinks that are free from chemical preservatives. There is increasing resistance to some chemical preservatives, particularly sodium benzoate, in all soft drinks and the use of potassium sorbate is now widespread. In other RTD products that are non-carbonated, the use of dimethyl dicarbonate is now widespread. The use of PET, and other polymers such as PEN that are able to withstand tunnel pasteurization, is likely to increase, and this will facilitate packaging of juices and other microbiologically sensitive products in much cheaper alternatives to the aseptic packs. 19.5.2 Carbonated soft drinks Carbonated soft drinks that are sold in large volumes in PET packages that are mostly up to 2 litres in volume are likely to dominate the soft drinks market for the foreseeable future. However, for smaller PET bottles the development of pasteurizable material will be likely to facilitate the sale of carbonated products that are free from chemical preservatives. The extension of this property to larger PET containers is probably limited by the higher levels of carbonation needed. Additionally, larger containers are frequently left part used and the presence of a chemical preservative in such circumstances is desirable. The limiting shelf life factor for carbonated products packed in PET is likely to remain that of the loss of carbonation rather than other oxidative or deteriorative reactions and there is a significant development potential for PET or similar plastics with improved retention of CO2. As with other premium products, the use of glass is likely to dominate the market for high quality adult oriented products that benefit from the pack image coupled with a long shelf life. For smaller unit volumes, the beverage can is likely to retain an important segment of the market for carbonated drinks as its convenience, recyclability and suitability for pasteurization enable its use for a wide variety of soft drinks. 19.5.3 Dilute-to-taste products Traditionally these products have been packed in glass but the use of PET is now widespread. Glass containers remain the packaging of choice for premium priced products. Most dilute-to-taste products are packed in containers ranging from 75 cl to 2 l and invariably contain chemical preservatives as well as being pasteurized. The use of preservatives is of particular importance as these products are frequently left part used for periods of time. For unopened containers, products packed in PET are likely to suffer some effects of oxidative and other deteriorative reactions in a shorter time than those packed in glass, but once opened and oxygen is allowed to re-enter the headspace, the performance of glass and PET is likely to be similar. For dilute-to-taste products the use of both glass and PET seems set to continue.


The stability and shelf life of fruit juices and soft drinks 593 19.5.4 Processing trends Various alternatives to thermal processing are available although it seems likely that for practical and economic reasons, thermal pasteurization is set to remain the preferred means of stabilizing fruit juice and soft drink products for the foreseeable future. High pressure processing This technique has been used successfully for the production of high value fruit juices where the retention of the characteristics of freshly pressed juice for longer shelf life is considered desirable. Batches of product, which must be packed in flexible containers, are placed in a pressure vessel and subjected to pressures of around 600 MPa. Current processes are both labour and capital intensive and are limited by the necessity of batch operation. Even then, the benefits that may be achieved for a freshly pressed juice only allow for an extension of shelf life by a factor of two or three times that of the unprocessed juice. The consumer benefit is thus limited with a significant increase in processing costs and thus selling price. Overall, this technology is still under development in both academia and the commercial world, but scope for its widespread adoption for relatively low value products such as fruit juices is probably limited. Irradiation Although the use of gamma ray irradiation from a source such as Cobalt 60 or Caesium 137 can be very effective in procuring microbiological stability in a variety of products, the technique is permitted by the US Food and Drug Administration (FDA) for only a limited range of foods which does not include fruit juices or soft drinks. Within the UK/EU, its use is even more limited. The use of high energy electron beams and X-rays is also very effective and may be more acceptable to consumers, although consumer acceptance of any irradiated foods remains a major obstacle. Developments of this technology are progressing widely but are likely, in the short to medium term, to be limited to foods where pathogens present a significant microbial hazard.

19.6


and ASHURST P R (eds) (1996) Fruit Processing, Blackie, Glasgow. (ed.) (2005) Chemistry and Technology of Soft Drinks and Fruit Juices, 2nd edn, Blackwell, Oxford. ASHURST P R and HARGITT R (2009) Soft Drink and Fruit Juice Problems Solved, Woodhead Publishing, Cambridge. GILES G A (ed.) (1999) Handbook of Beverage Packaging, Sheffield Academic Press, Sheffield. PAQUIN P (ed.) (2009) Functional and Speciality Beverage Technology, Woodhead Publishing, Cambridge. ROBERTSON G L (ed.) (2009) Food Packaging and Shelf Life, CRC Press, Boca Raton, FL. STEEN D and ASHURST P R (eds) (2006) Carbonated Soft Drinks, Blackwell, Oxford. ARTHEY D

ASHURST P R


20 Practical uses of sensory evaluation for the assessment of soft drink shelf life L. L. Rogers, Consultant, UK

Abstract: This chapter is concerned with the practical use of sensory techniques for soft drink shelf life setting and confirmation. It introduces the use of a resource efficient risk-based method for shelf life determination and gives an overview of the different approaches to setting shelf life for soft drinks. A list of questions and answers are posed for the sensory scientist to consider prior to setting up their shelf life experiments, and soft drink-based case studies give details of two different experiments to help with the design of future shelf life experiments. Key words: sensory science, shelf life, soft drinks, practical examples, case studies.

20.1

Introduction

The sensory quality of soft drinks can perhaps be regarded as less critical than the microbiological and analytical specifications; however, the sensory aspects of shelf life are incredibly important for product success in the market place. Microbiological safety will always come first for shelf life determination and should always be considered when devising the sensory assessment plan to ensure the safety of the sensory panellists and any consumers taking part in tests. The legal requirements for nutritional labelling, such as vitamin content over shelf life, are also highly important for food companies as these also constitute a legal requirement. The Food Labelling Directive (2000/13/EEC) states that food businesses must guarantee the safety, legality and quality of the product through its shelf life and it is a legal requirement to assign a shelf life, either a ùse by date' or a `best before date', to a food product. The ùse by' requirement is for


Sensory evaluation for the assessment of soft drink shelf life

595

pre-packed foods `which, from the microbiological point of view, are highly perishable and are therefore likely after a short period to constitute an immediate danger to human health'. For soft drinks this could apply to freshly squeezed juices, smoothies and chilled juices. A best before date is the `date up to and including which the foodstuff will retain its optimum condition (e.g., it will not be stale)'. Soft drinks such as cordials, pasteurised ready-to-drink products and heat-treated juices would fall into this category. Sensory quality comes in a very close third place due to the fact that consumers will not repurchase a product that they deem low in quality and they will probably not even take the current shelf life of the product into consideration when making this judgement. The first time a consumer tries a low quality end-of-shelf life product, they will be unhappy with the product quality and may never purchase it again ± even if it was one week before the end of its nine-month shelf life. This is particularly true for ambient soft drinks but can also apply to fresh juices and smoothies. The ASTM state that sensory shelf life is `the time period during which the products' sensory characteristics and performance are as intended by the manufacturer' and they also refer to this as being set to `manage business risk and meet business needs' (ASTM, 2005). There will come a point, depending on the product's physical and chemical properties and its storage conditions, when either the product quality will become unacceptable or it will become harmful to the consumer. These two outcomes may happen in the opposite order or may well happen at the same time. A stronger definition is given by the IFST: `shelf life is defined as the time during which the food product will remain safe; be certain to retain desired sensory, chemical, physical and microbiological characteristics; and comply with any label declaration of nutritional data when stored under recommended conditions' (IFST, 1993). The IFST definition leaves one point for discussion, however: `desired sensory . . . characteristics'. This will depend upon the product, the product type or classification, the branding (premium or own label, for example), and the management decisions around the setting of shelf life in each individual company. The shelf life of products is affected by many different aspects of the product ingredients, processing, packaging and storage conditions (e.g., preservatives, water activity, heat treatments, temperature control, oxygen ingress, light, humidity and pack integrity). The IFST group these effects into intrinsic and extrinsic factors. Intrinsic factors are the properties of the final product (e.g., water activity, pH) whereas extrinsic factors are the external effects on the product such as the time±temperature profile of the product, exposure to light and handling through the supply chain. These factors can all interact ± often unpredictably ± and should be taken into account when planning a sensory shelf life study, particularly in respect of food safety. The microbiological safety and nutritional aspects are heavily regulated, whereas the sensory aspects are not, unless one includes groups such as the Trading Standards in the UK where, for example, calling a child's squash


596


`peach' flavour where at the end of shelf life there is no peach flavour left, could mislead consumers. However, the sensory aspects of the food product: its flavour, texture and appearance, for example, will play a huge role in the consumer purchase behaviour ± both for first and repeat purchase decisions. The only way to measure this consumer àcceptability' is through sensory methods, be they analytical (triangle tests, profiling, difference from control) or consumer (hedonic, just about right scales, survival analysis) sensory methods (Stone and Sidel, 2004). Predictions of sensory shelf life may be possible through purely analytical methods, but this type of predictive method requires validation before use, using real-time storage conditions and sensory analysis to determine the level at which the marker compound results in an unacceptable product. This is generally conducted through the choice of `marker' chemicals, identified through gaschromatograph-mass-spectrometry (GCMS) methods, often linked to GColfactometry (Reineccius and Heath, 2005). This latter method allows the effluent from the GC column to be `sniffed' in an attempt to identify the key compounds responsible for the flavour of the product. Often the marker chosen is representative of the ageing process and gives a cut-off point for the end of shelf life when it reaches a certain level. Within different companies there will be additional considerations for shelf life setting. The supply chain network will probably play a role. For example, if a product has a shelf life of 7 days but it does not reach the supermarket until day 3, then the product quality for the 4 days the product is on sale will be critical. Supermarket requirements will also play a part: many will not accept a product with less than 70±80% of its shelf life remaining. This may result in a requirement to speed up time to market or extend shelf life by some means.

20.2

Using a risk-based approach to shelf life for soft drinks

The development of the shelf life plan for each experiment can be conducted using a risk-based approach. This is a simple way of making the best use of resources and facilities by only conducting comprehensive shelf life testing where required. All other plans will be based upon the amount of risk associated with each individual experiment. For example, for a new product in a new range, the risk assigned may be `high', as there is no existing data on which to base the shelf life determination. But for the change to a new powdered ingredient supplier from an existing supplier, the risk of the shelf life being affected is probably very low and therefore the minimum amount of testing could be conducted. For the confirmation of shelf life for ambient products the risk is generally low to medium, but within this range the sensory scientist may wish to assign further low to high categories, to further utilise resources effectively.



20.3

597

Estimating shelf life

In order to assign shelf life to a new or adapted product, a review of data from previous experiments or new shelf life studies (see below) would be conducted to determine the stability profile of the product. Appropriate storage conditions and the estimated storage period can then be recommended. A review of similar products on the market and their packaging can also give guidance as to the potential shelf life of a brand new product.

20.4

Determining shelf life

It is well known in the sensory field that the use of quantitative descriptive profiling with a trained sensory panel, combined with consumer liking or acceptance methods, provides both research and marketing teams with valuable consumer insight into product behaviour over time. It is recommended that this approach is taken to set and in some cases monitor shelf life. For example, once the key drivers of consumer liking over shelf life are known, a descriptive trained panel can then be used to predict the end of shelf life for both new and any adapted products. See Section 20.5 for more information. Linking the quantitative profiling and the consumer liking data can be extremely helpful in trying to explain the reasons for any drop-off in consumer liking. The end of shelf life is often set at a certain number, for example a score of 6.0 on the nine-point hedonic scale, or a specific drop in consumer acceptability. This can be established dependent upon the product type: for example for a premium smoothie product the limit might be set at `less than a 0.5 drop' on the nine-point hedonic scale, whereas it might be set at `no greater than a 1.0 drop' on the same scale for a product such as UHT orange juice. In some rare cases an action standard of `no detectable change in the sensory characteristics over shelf life' might be used, but this is generally set as an action standard for premium products where the consumer will not accept any deterioration of the product over shelf life. Another action standard for shelf life testing can be to state that: `the product should fail when it no longer represents the product concept'. This can be very useful for long-life products such as cordials and Tetrapak drinks, which, when they change in flavour, and no longer match the concept, are outside a shelf life that is deemed acceptable by consumers. A similar action standard lies in the knowledge of the product's overall sensory profile especially when this is linked to sensory specifications. If the product's overall sensory profile changes then this is the end of shelf life (MunÄoz et al., 1992). For example, if a drink product has a certain intensity of ginger flavour and this drops by, say 10%, and the sweetness also drops by 10%, the point at which this change happens is the end of shelf life. In fact the developer may deem that this is actually beyond the end of shelf life and set the shelf life to be shorter.


598


The overall product profile approach can also be adapted with consumer knowledge, in that only product attributes that are known or suspected to be key to the consumers' perception of the product are taken into account (ASTM, 2005). For example, in a fruit drink which develops a very slight òff-note', but this is not detectable by consumers, the change is not used to set the end of shelf life. But as soon as the appearance starts to change, which consumers will most likely notice, this analytical sensory measurement can be used to set the end of shelf life.

20.5

Monitoring shelf life

When monitoring or confirming shelf life, for example within a quality control environment, there are several methods available to the sensory scientist. These can include, for example, difference from control, rapid profiles, targeted attribute profiles and methods based on quality specifications (Carpenter et al., 2000). A control product of some description is generally required for comparison. When an ingredient change is being explored, for example, and no differences are required, this control would be the original product prior to any changes. Ideally this control should be made at the same time as the trial product and from the same batches of ingredients. This will help eliminate any additional changes to the product other than the change under consideration. The control product will be aged in the same way as the new product but generally a fresh control is included in the tests for production comparisons. For ambient products, storage under chilled conditions is generally suitable for the `fresh' control. Sometimes the control is simply data gathered for the existing product and used in the current experiment for comparison. There are a multitude of reasons for confirming shelf life during production. First, confirmation of shelf life is generally carried out on a regular basis on a certain number of batches a year for the simple reason of ensuring that the consumer is still receiving the high quality product they were expecting. Second, changes to ingredients and ingredient suppliers will result in the need to confirm these changes do not affect the product quality nor the product quality over shelf life. Other production changes such as temperature changes on processing, costsaving experiments, mixing times, equipment changes, will also result in the need to confirm that there has been no effect on shelf life. Packaging changes can also have a major effect on shelf life. Quite often packaging changes may be brought in to extend shelf life as advances in packaging technology become more cost effective. For quality teams heavily involved in new product development and the scale-up to factory production, there may well be involvement in the actual setting of shelf life for the product. This may require the setting up of several tests to determine when the end of shelf life has been reached using certain predetermined action standards linked to the risk level the company has agreed to take for that product. For other quality aspects, such as the change in an



599

ingredient supplier, process changes or packaging changes, the involvement will be in confirming shelf life in comparison to the current standard product (Meilgaard et al., 1999). This can be a much simpler affair as previous work involving the consumer and the changes to key attributes that drive consumer liking, has already been determined and so more simple sensory difference tests might be used to confirm previous findings. However, if the changes under consideration shorten or lengthen the product's shelf life, consumer validation might be required, resulting in a complex experiment involving not only the product under consideration but also the control product or products.

20.6

Considerations before developing the shelf life plan

There are many aspects of the product's shelf life to consider before developing the plan of approach. These considerations are given in Table 20.1. Each of these are considered in turn below. Is the product stable or for how long is the product expected to be stable for? This is an important starting point, as the length of the expected shelf life will determine how many time points are considered, what sensory tests might be used and when decisions will be able to be made. For example, if the product is a powder, stable for one year under usual production conditions, and a change to the packaging is required which is expected to lengthen its shelf life, then conducting tests every month would seem to be rather excessive. If, however, the product has a shelf life of four months and a change to an ingredient is expected to shorten its shelf life, then early, and regular, testing would be recommended.

Table 20.1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Considerations before developing a sensory shelf life plan

Is the product stable or for how long is the product expected to be stable? When does the product start to change? What attributes change and how? What new attributes are introduced (if any)? What further changes happen after the initial change? Can the changing attributes be explained? How long are the products on the shelf? At what point in a product's life is it consumed? How many consumers have access to older products? When do consumers notice the product change? What effect do the changes have on consumer measures? When are the data required? At what project stage does the shelf life need to be set? How will the sampling plan accommodate these requirements? Are there any accelerated methods for storing the product?


600


When does the product start to change? This knowledge is very useful as it can determine the start of the testing procedure, apart from, of course, the initial test on the fresh product to set the baseline data. If a product, with a shelf life of ten months starts to change at eight months generally, and the planned ingredient change is expected to shorten shelf life, the best plan may be to select two or three time points up to say, seven months and then test weekly or fortnightly around the expected change point. Then, if the data indicate the shelf life will be shorter, the time points will be readily available from which to select the new point at which the product has reached its end of shelf life. What attributes change and how? What new attributes are introduced (if any)? What further changes happen after the initial change? The main consideration here in the choice of sensory test, is whether or not quantitative data are required on the planned product change, as these data can be very useful in determining the reasons behind any problems and perhaps working towards eliminating these. If information is only needed about the fact a change has happened, then one of the simple difference tests may be the best choice. If the initial change is then followed quickly by a change that it is known to affect consumers' overall liking of the product, it is important to look for this initial change to help decide the end of shelf life: any later may be too late for the consumer. The answer to these questions will help determine the choice of sensory test. For example, if several attributes change over the product's shelf life, it might be worth considering a rapid profile at each time point, particularly if this method is also used for other quality measurements (Lawless and Heymann, 1999). This profile would be conducted using the attributes drawn up during product development phases or may be an existing `language' for a standard product. The addition of one or two `blank' attributes for each modality can be useful to monitor the introduction and level of new attributes over shelf life. A text box can be inserted into the protocol to enable the panellists to score and describe the new attribute simultaneously. If only one attribute is known to change over shelf life, then several methods are available. Ranking of the changing attribute may be useful, or a simple difference test to determine when the change is happening. Difference from control tests can also be very useful in confirming sensory shelf life, particularly as a direct comparison to the original control product is made within the test. Can the changing attributes be explained? This information can help eliminate changes that severely affect the product's shelf life and may well result in product reformulation in extreme cases. If the changes are due to, for example, flavour loss, tainting, colour changes or packaging effects, then further development work might be required to slow down or eliminate these changes ± particularly if the ingredient or packaging change under consideration itself is the main culprit.



601

How long are the products on the shelf? At what point in a product's life is it consumed? How many consumers have access to older products? These questions are important when determining the time points to test. For example, if the majority of a product with nine months' shelf life is consumed within three months, then there is little requirement to test comprehensively after this point. But if a fair proportion of people may well purchase the same product when it is seven or eight months old, then testing should be extended over the complete shelf life. When do consumers notice the product change? What effect do the changes have on consumer measures? As soon as the product starts to change, as indicated by the sensory tests at each time point, the sensory scientist will need to determine whether this is noticed by the consumer. This might be carried out in any number of ways. First, data may already exist about the sensory attributes that are key drivers of consumer liking and if one of these key attributes is changing in a negative manner then perhaps the end of shelf life has already been reached. If these data exist but due to the experiment, for example a packaging change, additional attributes have been introduced for which no consumer data exist, it might be wise to conduct a small-scale consumer test to determine whether consumers notice the difference. If there are no data on whether the change will be detected by consumers or what effect it will have on their overall liking, again a small-scale consumer test, with around 60 consumers, can give valuable information. In this particular example it will be worthwhile planning for a consumer test at each time point when the sensory attribute data have been gathered and then deciding whether or not to conduct the test depending upon the result of the analytical sensory test. The action standards developed by the project team will be useful here in determining the end of shelf life. When are the data required? At what project stage does the shelf life need to be set? How will the sampling plan accommodate these requirements? It can be very cost and resource effective to conduct sensory shelf life tests by collecting samples throughout shelf life and conducting all the analysis at the hypothesised end of shelf life. This technique is often referred to as a singlepoint shelf life test. This can be conducted through the storage of one batch or through production of several batches. For short shelf life products it is generally fairly easy to set up and complete. The end of shelf life data will be available after a short time period and it is generally not worth considering any of the other approaches which are used for ambient products. For an ambient product, if it is known that the product stores well under chilled conditions with very little change to sensory characteristics, samples can be put on store in the conditions under study (for example, under supermarket type conditions) and then taken off store at the various time points in the sensory


602


plan and put into chilled storage. This has the effect of `pausing' storage. For example, if an ambient product with an estimated shelf life of 26 weeks was stored in supermarket conditions for 13 weeks and chilled (paused) conditions for 13 weeks, although the sample will be 26 weeks old at the end of the shelf life, it is effectively 13 weeks old as it has been ìn stasis' for 13 weeks. In this way all samples, for example 0, 3, 6, 9 weeks, can be saved up for a quantitative profile at the end of the 26-week period. This is obviously very cost effective but not particularly useful if the project manager needs to make decisions at 13 weeks of storage. In this case, especially if the project manager has to actually set the shelf life at, say 13 weeks, for a product that has a hypothesised shelf life of 26 weeks, predictive and accelerated methods might prove useful. If, however, chilled storage of the product is not feasible, another approach is to take production batches at certain time points and collect them for the final profile. In this example, production at week 0 would serve for the 26-week-old product, production at week 7 would serve for the 19-week-old product (26 ÿ 19 7) and production at week 13 would serve for the 13-week-old product, and so on. In this case the production batches must be similar and this method will not work for products with large batch-to-batch variation, unless this can be taken into account during analysis. In cases where the project manager needs information throughout shelf life, the sensory test will need to be conducted at each time point. This method is generally referred to as multi-point shelf life testing, and is generally more resource intensive than the previous sampling methods. Any accelerated samples would also be included at each time point to allow information on the end of shelf life to be available as soon as possible (see Section 20.6.8 for more information). It is also very beneficial to have a sample, generally the fresh product, within the sample set to allow for changes in the panel scoring, particularly if the expected shelf life is over several months. To enable the attribute list to develop over time, the addition of line scales and text boxes to enable the panellists to score and describe the new attribute can be helpful. Are there any accelerated methods for storing the product? Accelerated shelf life testing (ASLT) (Kilcast and Subramaniam, 2000) can be incredibly useful when setting a shelf life for an ambient product, as the developer does not need to wait for the whole of the product's shelf life to determine the end of shelf life. For example, storage at a higher temperature can result in a prediction in, say, half the time of the real shelf life, although some methods claim to deliver the information in substantially less time. In some cases the developers may use the worse-case scenario risk assessment approach to help with the prediction of shelf life, e.g., heat cycling the pack that is most susceptible to change, or the injection of oxygen into the pack combined with heat treatment. However, there are many critics of accelerated shelf life testing. In a recent review (Hough, et al. 2006), Harry Lawless was quoted: àccelerated testing is



603

mostly useless. If it worked, you could make fine aged Bordeaux in an oven. You can't!' The main issue with accelerated conditions is with the production of artefacts: chemical components that would not usually occur under normal storage conditions. Another problem is the amount of resource it takes to validate the accelerated method. The product must be kept at real-time storage conditions for the required length of time to compare to the accelerated test results, thus resulting in a long lead time before the accelerated test can be used with confidence. For confirmatory shelf life tests, this investment will be very worthwhile, but for new product development the resource is less warranted, until the product has been launched and becomes part of standard production. The quandary is that during the new product development phase, the accelerated test is at its most useful. In some cases companies have developed methods that are applicable to a certain group of products, and therefore predictions from accelerated tests are known to give a good enough approximation of the shelf life, until the confirmatory tests are finished.

20.7

Developing the sensory plan

A protocol should be prepared for each shelf life study and it can be very useful to discuss this with the project team prior to its commencement. This could include items from the list below. More details on each listed item are given later in the case studies. · · · · · · · · · · · · · ·

the purpose of the stability study business risk/product risk previous data about the product, similar products or competitor products packaging information and effect on shelf life project stage and when a shelf life decision is required expected shelf life accelerated tests availability storage conditions for each individual sample batches to be selected: number and type which pack type (where there are numerous sizes, for example) sampling plans and time points the testing to be performed at each time point testing specifications and action standards the number of packs required to conduct the specified testing (i.e., how many are required to be put on store and if there is sufficient quantity) · any special requirements (e.g., open shelf life tests, colour assessments).

20.8

Case studies

Using the protocol points above and the risk-based approach to shelf life testing, the sensory scientist can draw up a plan for the shelf life experiment. Two examples are given below and are compared directly in Table 20.2.


Table 20.2 Case study comparison


Point/Question

Case Study 1

Case Study 2

The purpose of the stability study Business risk/product risk

To set shelf life for a new product, a fresh juice product sold in one pack type and size. The business risk is high. The product risk is medium. There are no previous data for the final formulation but there are previous data collected during product development from pilot-plant manufacture but in a different pack size. These data are in the form of sensory profiling data, consumer acceptability tests and also analytical data. The product will be sold in a small PET, fully sleeved pack with a flat yellow-coloured cap. This pack and sleeving are ready and will be used for the trial. As the pack is slightly larger than the pack used in the pilot-plant trial, it is hoped this will help extend the shelf life, albeit only by a fraction. Information required throughout the experiment and a prediction required as early as possible. A final `go/no go' decision will need to be made on day 9 of the full-scale production trial. Initial studies suggest could be up to 10 days. The data from the pilot plant study indicate a shelf life of the experimental product of around 6 to 7 days. It is hoped that full production will give a shelf life of 9 days to help with supply-chain issues. Not necessary for the short shelf life. Multi-point shelf life plan. See Table 20.3 for details.

To confirm shelf life for an ambient juice product sold in three different pack sizes. The business risk is low. The product risk is low. Current product, complete data sets available for all three pack types. These data are in the form of sensory profiling data, consumer acceptability tests, difference from control production records and analytical data.

Previous data about product or similar products

Packaging information

Project stage, shelf life decision requirements Expected shelf life

Accelerated test availability Storage conditions for each individual sample

The current packs are for sale in three sizes: small, medium and large. Generally the smaller pack has more shelf life ìssues' due mainly to the increased contact of the liquid with the pack. This experiment is simply a monitoring exercise as there have been several recent minor changes to ingredients and production and therefore confirmation of shelf life is required. This is fully documented: six months as per current smallest pack, nine for the medium pack and 12 months for largest pack. Available and validated. Single-point reversed shelf life plan. See Table 20.4 for details.

Batches to be selected, number and type Pack type

Three batches will be selected from the large-scale trial for sensory profiling and analytical tests; however, the consumer tests will be conducted on only one of the batches. This product will only be sold in one pack type and this pack will be studied in this experiment.


Sampling plans and time points

As the shelf life is so short, and data are required as the experiment progresses, analysis of the samples will be conducted every day. See Table 20.3 for details.

The testing to be performed at each time point

Sensory profiling will be conducted using the existing attribute list. Small-scale consumer tests will also be carried out at day 1 and planned for day 5 and every day up to the end of shelf life. See Table 20.3 for overview of testing.

Testing specifications and action standards

The action standard for this experiment is: the shelf life for the new juice product will be longer than 7 days. If not the product will be reformulated. The end of shelf life will be set as when the consumer overall liking score drops by 0.5 or more on the 9-point hedonic scale. There will be sufficient quantity of product as the trial is full scale.

The number of packs required to conduct the specified testing (i.e., how many are required to be put on store and if there is sufficient quantity) Any special requirements (e.g., open shelf life tests, colour assessments)

Open shelf life tests will be required in the future. During this experiment the microbiological assessments on open shelf life will be conducted.

Three batches will be selected from the large-scale trial to be conducted. Each batch will be packed into the three different pack sizes. Only the smallest and largest pack sizes will be tested at the end of shelf life. The medium pack size will be kept on store in case the analysis highlights any issues. As the product has such a long shelf life, and data are not required immediately, analysis of the samples will be conducted at the end of shelf life. See Table 20.4 for details. All analysis will be conducted at the end of shelf life as this experiment is for confirmation only and data are not required earlier. A full quantitative profile will be conducted using the current attribute language. A small-scale consumer test is planned in case of any issues. All products should have the same shelf life as per current specifications. The action standard is: shelf life will be the same as per current production.

There will be sufficient quantity of product as the trial is full scale. There will be a requirement to put on store 100 spare units in case the sensory profiling tests indicate any concerns and consumer tests are required. No special requirements for shelf life testing.

606


20.8.1 Case study 1 The purpose of the stability study To set shelf life for a new product, a fresh juice product sold in one pack type and size. Production will be at a new contract packer who currently packs other products for the business. Business risk/product risk The business risk is high as this is a new product with a large advertising campaign. The product risk is medium as the product appears to change very little over the first few days of shelf life, but more data are required for extending the shelf life, hence this full-scale experiment. Previous data about product or similar products There are no previous data for the final formulation but there are previous data collected during product development from an experimental pilot-plant manufacture in a different pack size. These data are in the form of sensory profiling data, consumer acceptability tests and also analytical data. The contract manufacturer is experienced in making similar products and produces for similar ownlabel products. Packaging information The product will be sold in a small PET, fully sleeved bottle with a flat yellowcoloured cap. This pack and sleeving are ready and will be used for the trial. As the pack is slightly larger than the pack used in the pilot-plant trial, it is hoped this will help extend the shelf life, albeit only by a fraction. Project stage, shelf life decision requirements The decision on the shelf life the product might reach is required throughout the experiment and a prediction required as early as possible. A final `go/no go' decision will need to be made on day 8 of the full-scale production trial. Expected shelf life In studies carried out before the pilot-plant shelf life experiment, similar competitor products indicated that the shelf life of this new product could be up to 10 days. The data from the pilot-plant study indicate a shelf life for the experimental product of around 6±7 days. The product developed a dried fruit note and lost the full intensity of its deep red colour. This affected the overall liking in the small-scale consumer tests which dropped from a score of 6.7 on the 9-point hedonic scale at day 0, to 6.0 at day 8. The sensory action standard has been set to indicate the end of shelf life when the consumer overall liking score drops by 0.5 or more on the 9-point hedonic scale. The sensory profiling data will give an early warning of this change when similar levels of dried fruit notes and a drop in red coloration are seen. It is hoped that full production, with its less harsh processing conditions, will give a shelf life of 10 days to help with supply-chain issues although this does



607

not look likely from the previous data. A final `go/no go' decision will need to be made on day 8 of the full-scale production trial as this fits in with the panel working hours and is probably the longest shelf life possible based on the previous data. Accelerated tests availability Not necessary for the short shelf life as results will be ready in around two weeks. Storage conditions for each individual sample The experiment will be set up as a multi-point experiment conducted every day. See Table 20.3 for details. All product is stored chilled as per storage instructions developed during pilot-plant study. Samples from each batch will also be stored frozen for analysis on day 5 and day 9 and also in a later profile to determine whether the product can be frozen for future tests. Batches to be selected, number and type Three batches will be selected from the large-scale trial for sensory profiling and analytical tests; however, the consumer tests will be conducted on only one of the batches. The sensory profiling data will be used to determine the extent of the batch-to-batch variability. The production trial is required to be conducted on a Wednesday (or Thursday) to match with panel attendance days. Which pack type? This product will only be sold in one pack type and this pack will be studied in this experiment. Sampling plans and time points As the shelf life is so short, analysis of the samples will be conducted on day 0, day 1, day 2, day 5 (thus allowing for the weekend), day 6, day 7, day 8, day 9 and possibly day 12. Decision is needed if shelf life can be extended. Decision is also needed regarding panellists' attendance over second weekend if necessary. Further analysis days will be agreed as data are gathered through the joint analytical, micro and sensory meetings. Testing is also planned for day 12 in case the results from full-scale trial look better than the pilot-plant production. The testing to be performed at each time point Sensory profiling will be conducted using the existing attribute list. This attribute list was generated during the previous shelf life trials for the pilot-plant produced juice. Time will be set aside for additional training using the different batches on day 0 of the experiment to confirm the attribute list/vocabulary. Small-scale consumer tests will also be carried out at day 1 and planned for day 5 and every day up to the end of shelf life. See Table 20.3 for an overview of the testing. The tests will be conducted if the sensory profiling data indicate that consumer data will help in the shelf life determination.


Table 20.3 Case study 1 experimental design ß Woodhead Publishing Limited, 2011

Day

Profile notes

Consumer test notes

Day 0 Wednesday

Day 1 Thursday

Full-scale trial: 3 batches of product produced

All product stored chilled Selection stored frozen

Sensory profile conducted after a panel session to confirm vocabulary

Sensory profile: day 1 (3 batches and 1 repeated to confirm panel performance) Small-scale consumer test (n 60) to assess any changes from pilot plant production

Day 2 Friday

Sensory profile: day 2 (3 batches and 1 repeated to confirm panel performance)

Day 3 Saturday

Day 4 Sunday

Weekend

Weekend


Day

Day 5 Monday

Meetings

Joint analytical, micro and sensory meeting to discuss results so far

Profile notes


Notes

Frozen product also assessed in sensory profile

Consumer test notes

Small-scale consumer test (n 60) to assess any effect on consumer liking of shelf life changes (if any)

Day 6 Tuesday



Day 7 Wednesday

Day 8 Thursday

Day 9 Friday

Joint analytical, micro and sensory meeting to discuss results so far

Go/no-go decision ± meeting at 3pm

Final meeting




Decide if panel working weekend

Agree future tests after day 9 (see note below)

Frozen product also assessed in sensory profile




Day 12: Sensory profile: day 12 (3 batches and 1 repeated to confirm panel performance) to assess if a longer shelf life may be possible. In-house consumer test (n 60) to assess any effect on consumers liking of shelf life changes also planned.

610


Testing specifications and action standards The action standard for this experiment is: the shelf life for the new juice product will be longer than 7 days. Any issues in the sensory profiling results will be confirmed by consumer tests. If the shelf life is less than 7 days, the product will need to be reformulated. The later profiling tests may well be performed to help understand the changes happening in the product. These data will be used alongside the analytical testing data to help with the reformulation. The number of packs required to conduct the specified testing (i.e., how many are required to be put on store and is there sufficient quantity) The sensory tests planned will require ten product units at each time point and the consumer tests will require 60 of each at each time point. There will be sufficient quantity of product as the trial is full scale. A temporary refrigerated unit is in place at the contract manufacturer and has been used for other product trials. Any special requirements (e.g., open shelf life tests, colour assessments) Open shelf life tests will be required in the future. During this experiment the microbiological assessments on open shelf life will be conducted. A further open shelf life sensory experiment will be set up once the microbiological data have shown how long the product can be stored open safely. 20.8.2 Case study 2 The purpose of the stability study To confirm the shelf life for an ambient juice product, sold in three different pack sizes, due to recent production changes. The production will be at the current site where the recent production changes have happened. Confirmation of shelf life is required due to the numerous production changes. Business risk/product risk The business risk is low as it is simply a confirmation that the three packs still have the same shelf life. The product risk is also low as the product changes very little over shelf life but does change markedly as the end of its current shelf life is reached. Previous data about product or similar products This experiment concerns a current product and complete data sets are available for all three pack types. These data are in the form of sensory profiling data, consumer acceptability tests, difference from control production records and analytical data. The data span almost three years of production. The sensory vocabulary and consumer questionnaire exist and have been used during other experiments regarding this product. Packaging information The current product is for sale in three sizes: small, medium and large. Generally



611

the smaller pack has more shelf life ìssues' due mainly to the increased contact of the liquid with the pack. Project stage, shelf life decision requirements This experiment is simply a monitoring exercise as there have been several recent minor changes to ingredients and production, and therefore confirmation of shelf life is required as per production protocols. Expected shelf life This is fully documented: six months for the smallest pack, nine for the medium pack and 12 months for largest pack. Accelerated tests availability The accelerated test is available, in current use and has been validated. The acceleration involves storage of the products at an extended temperature only. The accelerated test predicts at double storage time (i.e., 6 months' storage under accelerated conditions is equivalent to 12 months' real-time storage). Generally tests are carried out on product stored chilled, at ambient (under supermarket conditions) and at the accelerated temperature. However, as this experiment does not require data upfront, and as the accelerated test is already validated, no storage at accelerated temperatures will be included in this study for confirmation. However, a simple difference from control test will be conducted with accelerated smaller packs compared to chilled stored controls, three months into the storage time to give an early warning of any issues. It would be a simple case to pick one, or even all, of the pack types to incorporate a validation exercise if needed. Storage conditions for each individual sample The experiment will be set up as a single-point reversed shelf life plan. Table 20.4 gives more details. All samples will be collected and placed on store at the start of shelf life. All samples will be stored chilled and under shelf conditions (i.e., supermarket conditions). They will be taken off store and placed into chilled storage at the required time points therefore `pausing' the storage time. The majority of analyses will be conducted at the end of shelf life. Spare product will be stored under shelf conditions for consumer tests if required. Batches to be selected, number and type Three batches will be selected from one production week. Each batch will be packed into the different pack sizes, labelled and stored accordingly. Which pack type? Only the smallest and largest pack sizes will be tested at the end of shelf life. The medium pack size will be kept on store in case the analysis highlights any issues.


Table 20.4 Case study 2 profiling experimental design


Week

Week 0

Week 1

Week 20

Week 22

Week 24

Storage conditions

Full-scale production. Tests conducted on all three pack sizes. Product stored chilled and under ambient supermarket conditions

Sensory profile conducted after a panel session to confirm vocabulary

Samples (small pack size stored at ambient temperatures) taken off ambient storage and stored chilled (or `paused')



Notes

Single-point reversed shelf life plan starts

All pack sizes included in the same profile

These samples will be equivalent to storage for 20 weeks



Week

Week 26

Week 35

Week 39

Week 48

Week 52

Storage conditions

Samples (all three pack sizes stored at ambient temperatures) taken off ambient storage and stored chilled (or `paused')

Samples (last two pack sizes stored at ambient temperatures) taken off ambient storage and stored chilled (or `paused')

Samples (last two pack sizes stored at ambient temperatures) taken off ambient storage and stored chilled (or `paused')

Samples (large pack size stored at ambient temperatures) taken off ambient storage and stored chilled (or `paused')

Samples (large pack size stored at ambient temperatures) taken off ambient storage and analysed.





Notes

All samples stored chilled/paused taken off store and analysed. Medium pack taken off store and analysed if required. These samples will be equivalent to storage for 52 weeks

Profile notes

Sensory profile conducted on all samples and batches for both smallest and largest pack size (after one panel session to check vocabulary)

Consumer test notes

Small-scale consumer test (n 60): chilled versus 26 (small pack), 39 (medium pack) and 52 week (large pack) ambient to assess any effect on consumers liking


613

Sampling plans and time points The product has a shelf life of 52 weeks in the current pack and 26 weeks in the smallest pack. Shelf-stored samples will be taken at 0, 22, 24 and 26 weeks for the smallest pack and 0, 26, 35, 39, 48 and 52 weeks for the largest pack. Accelerated samples will be taken at 11 and 13 weeks for the smallest pack and 24 and 26 weeks for the largest pack. The testing to be performed at each time point All analysis will be conducted at the end of shelf life as this experiment is for confirmation only and data are not required earlier. A simple difference from control test will be conducted with accelerated smaller packs compared to chilled stored controls, three months into the storage time to give an early warning of any issues. A full quantitative profile will be conducted using the current attribute language; however, the panellists will be presented with all samples to update the language prior to conducting the three replicates on the stored samples. A small-scale consumer test is planned in case of any issues. Testing specifications and action standards All products should have the same shelf life as per current specifications. The action standard is: shelf life will be the same as per current production. Any issues in the sensory profiling results will be confirmed by consumer tests. The number of packs required to conduct the specified testing (i.e., how many are required to be put on store and is there sufficient quantity) The sensory tests planned will require five product units for the large pack and ten for the small pack at each time point. The time points have been kept to a minimum, so there are no store room issues. There will be sufficient quantity of product, as the trial is full scale. There will be a requirement to put on store 100 spare units in case the sensory profiling tests indicate any concerns and consumer tests are required. Any special requirements (e.g., open shelf life tests, colour assessments) No special requirements for shelf life testing.

20.9

Future trends

In the future, consumer expectations and consumer power can only be expected to increase. This may well result in a demand for products of a higher quality over the whole of their shelf life, as opposed to the trend in cheaper `close to the end of shelf life' sales seen recently. A similar trend has been seen in the increase of high quality products even for own label variants, as consumers' high quality demands are met. Therefore the combination of both analytical and consumer sensory methods can only be expected to increase, as companies realise the need for robust end-of-shelf life setting and confirmation.


614


Improvements in the packaging field such as modified-atmosphere and advanced technology in the production of different types of packaging, will also benefit shelf life in the future. These include development of new coatings and barriers, the use of oxygen scavengers, particularly in caps, and the development of new techniques for packaging manufacture. Different production methods, such as microwave or high pressure processing may also help produce foods with a longer shelf life. Changes to the supply chain may also impact on shelf life. New methods of monitoring temperature in the supply chain and intelligent pack labels have the potential to both increase and decrease shelf life where required to help deliver higher quality products to the consumer. Predictive methods will also be more widely available, and using statistical modelling techniques will result in more accurate setting of shelf life, leading to higher quality foods at the end of shelf life. The use of more sensitive analytical techniques will allow shelf life data to be gathered more quickly and easily and will also help develop the understanding of the various mechanisms of the degradation of food and drink products over shelf life. The use of microbiological models and risk-based approaches will be an area of interest for food scientists in the future. New ingredients such as natural antimicrobials, new plant breeding techniques, and genetic engineering may also be areas to watch out for in future shelf life extensions.

20.10

References

(2005), Standard Guide for Sensory Evaluation Methods to Determine the Sensory Shelf Life of Consumer Products, E 2454-05. CARPENTER, R P, LYON D H and HASDELL, T A (2000), Guidelines for Sensory Analysis in Food Product Development and Quality Control, Aspen, Gaithersburg, MD. HOUGH, G, VAN HOUT, D and KILCAST, D (2006), `Workshop summary: sensory shelf-life testing'. Food Quality and Preference, 17, 640±645. IFST (1993), Shelf Life of Foods ± Guidelines for its Determination and Prediction, IFST, London. KILCAST, D and SUBRAMANIAM, P (2000), The Stability and Shelf-life of Food, CRC Press, Boca Raton, FL. LAWLESS, H T and HEYMANN, H (1999), Sensory Evaluation of Food: Principles and Practices, Aspen, Gaithersburg, MD. MEILGAARD, M, CIVILLE, G V and CARR, B T (1999), Sensory Evaluation Techniques, CRC Press, Boca Raton, FL. Ä OZ, M, CIVILLE, G V and CARR, B T (1992), Sensory Evaluation in Quality Control, Van MUN Nostrand Reinhold, New York. REINECCIUS, G and HEATH, H B (2005), Flavor Chemistry and Technology, CRC Press, Boca Raton, FL. STONE, H and SIDEL, J L (2004), Sensory Evaluation Practices, Academic Press, London. ASTM


21 The stability and shelf life of coffee products L. Manzocco, S. Calligaris and M. C. Nicoli, University of Udine, Italy

Abstract: Shelf life assessment of coffee derivatives is a complex task due to the wide number and heterogeneity of products belonging to this food category. For this reason, shelf life assessment strategy must be carefully designed taking into account the peculiarity of the product. Shelf life testing of coffee derivatives will be discussed, focusing initially on the main critical events affecting the stability of coffee products and factors controlling the deterioration rate. Shelf life assessment strategies will then be illustrated by considering the identification of the acceptability limit described by the proper indicator and methodologies for shelf life testing under actual and accelerated storage conditions. Key words: coffee derivatives, acceptability limit, kinetic modelling, accelerated shelf life test.

21.1

Introduction

The shelf life of coffee products may range from a few minutes/hours for an espresso cup or a coffee brew, respectively, to several months for ground and roasted coffee beans, ending up with many years for instant coffee. In the first case, shelf life is too short to need an evaluation, in the last one it is so long that appropriate procedures for shelf life prediction should be developed. Such a variegated situation is the result of the fact that coffee products are wide and heterogeneous as well as their stability and lifespan. Coffee derivatives may exert strongly different shelf lives due to the combination of several variables including intrinsic aspects, such as product composition and characteristics, and


616


Table 21.1 Key unit operations to obtain the main coffee products and relevant effects Unit operation Phenomena

Effect

Product

Roasting

Water removal Non-enzymatic browning Pyrolysis

Bean darkening Volatile and CO2 formation Bean expansion Volatile and CO2 release

Roast coffee

Grinding

Particle size reduction

Surface area increase Volatile and CO2 release

Roast and ground coffee

Brewing

Solid-liquid extraction

Extraction of soluble and Coffee cup emulsifiable substances from the coffee matrix to water

Dehydration

Water removal

Volume decrease Solid concentration

Coffee concentrate, instant coffee

Microbial inactivation

Ready-to-drink coffee drinks

Pasteurization Sanitization Sterilization

extrinsic ones, mainly relevant to packaging and storage conditions. In addition, marketing considerations are also expected to enter into the shelf life decision process since the extent of consumer satisfaction is fundamental for coffee producers. Therefore, shelf life testing of coffee derivatives should be carefully designed considering all these aspects. Green coffee beans from Coffea Arabica, Arabica coffee, and Coffea canephora, Robusta coffee, are the starting material for all coffee derivatives. The latter comprise roasted coffee ± decaffeined or not ± and a wide variety of convenience and semi-manufactured products such as coffee concentrates, instant coffee and ready-to-drink beverages. Table 21.1 shows the key technological steps applied to obtain the most important coffee products. The first operation needed to convert green beans into a beverage is roasting. This is the key technological operation allowing dark beans with the characteristic pleasant flavour and aroma to be obtained. During roasting, coffee beans are put in contact with a hot surface or gases to increase their temperature to 220 ëC. The main reactions occurring are water removal, carbohydrate fragmentation and polymerization, via non-enzymatic browning reactions, and pyrolysis. The chemical composition of the beans is drastically modified with release of large amounts of carbon dioxide (CO2) and the formation of hundreds of substances associated with coffee aroma and taste. Among reaction products, there is a wide number of volatiles and non-volatiles, such as melanoidins and their precursors. The formation of volatiles and CO2 during roasting in combination with the high temperatures of the environment causes the expansion of the beans and the formation of pores and pockets. The beans become very brittle and progressively lose their ability to entrap and retain volatiles. Thus, volatile


The stability and shelf life of coffee products

617

and CO2 are easily released into the vapour phase due to diffusion mechanisms favoured by the pressure gradient between the internal bean pores and atmosphere. For this reason, a degassing step is carried out on roasted coffee before packaging in order to avoid the swelling of the packages during shelf storage. Alternatively, appropriate packaging solutions should be implemented to control these releases and to avoid over-pressure inside the package with possible bursting and loss of package integrity (Nicoli et al., 2010). The whole roasted beans should be ground into smaller fragments before extraction. The purpose of this operation is to increase the specific extraction surface and facilitate the transfer of soluble and emulsifiable substances from the coffee matrix into the brew (Petracco, 2005a). The roast and ground coffee is subjected to a solid-liquid extraction with hot water to obtain a beverage that may be consumed as is or constitute a semimanufactured product. The extraction methods adopted can vary greatly from country to country, strongly affecting the sensorial properties of the coffee (Pictet, 1987; Petracco, 2001, 2005b). Coffee brews prepared at domestic and catering level are generally consumed immediately after preparation. At industrial level, the dehydration of the coffee brews leads to the production of instant coffee or coffee concentrates depending on the degree of water removal. In the case of instant coffee, the dehydration is achieved by freeze drying or by spray-drying, whereas coffee concentrate is obtained through thermal or cryo-concentration (Clarke, 2001). While instant coffee is highly appreciated due to its convenient physical form and long shelf life, coffee concentrates are increasingly used as ingredients for the food industry or as semi-manufactured products for vending machines and catering. Finally, pasteurized or sterilized ready-to-drink coffee beverages have recently become very popular, especially in Asian countries, where the tradition of consuming freshly prepared coffee brews is not widespread. They could be variously formulated containing, besides coffee brews, sugar, dairy products, emulsifiers and other additional ingredients (Petracco, 2001). Although, in term of sold volumes, ready-to-drink coffee drinks are minor coffee products, they represent an interesting and dynamic area for both coffee roasters and soft drink producers.

21.2 Main critical events affecting the stability and shelf life of coffee products Table 21.2 summarizes the main critical events leading to quality depletion of coffee derivatives during their life on the shelf. While the instability of roasted and instant coffee is mainly due to the development of chemical and physical changes including oxidation, volatile loss and physical collapse, the stability of coffee derivatives with higher water content (i.e., coffee concentrates and brews) can also be potentially affected by microbial spoilage as well as by the development of chemical reactions requiring reactants' mobility.


618


Table 21.2 Main critical event leading to quality depletion during shelf life Product

Critical event

Roast and ground coffee

Oxidative reactions Volatile loss Oxidative reactions Physical collapse Microbial spoilage Oxidative reactions Volatile loss Ester hydrolysis Non-enzymatic browning Microbial spoilage Oxidative reactions Volatile loss Ester hydrolysis Non-enzymatic browning

Instant coffee Coffee concentrate

Ready-to-drink coffee beverage

In general terms, all coffee products are known to lose their particular sensory properties due to the occurrence of a defect generally referred to as `staling'. Buffo and Cardelli-Freire (2004) defined coffee staleness as à sweet but unpleasant flavour and aroma of roasted coffee which reflects the oxidation of many of the pleasant volatiles and the loss of others'. Both these phenomena contribute to coffee staling even if the relative weight of one mechanism in comparison with the other is not easy to highlight. The development of oxidative reactions causes not only the loss of pleasant aroma compounds but also the formations of off-flavours (Nicoli and Savonitto, 2005). The sensitivity of roasted coffee towards oxidation reactions is high due to the presence of a large number of strongly active volatile and non-volatile compounds that easily react with oxygen. Among these substances, aromaimpact components, such as aldehydes, ketones and thiols, are particularly prone to oxidation along with the lipid fraction of coffee (Nicoli and Savonitto, 2005; Ortola et al., 1998). The latter ranges from 10 to 14% and contains about 75% of triacylglycerols with a high percentage of unsaponificables, including diterpene alcohols, sterols and tocopherols (Speer and Kolling-Speer, 2001). It should be noted that the oil susceptibility is increased just after roasting due to its migration to the surface of the beans, where the risk of oxidation is maximal. Besides the presence of compounds clearly prone to oxidation, the role of other substances with strong antioxidant capacity should not be underestimated. Several authors (Nicoli et al., 1997; Daglia et al., 2000; Krings and Berger, 2001; Anese and Nicoli, 2003) have attributed the strong antioxidant properties of coffee to the presence of both naturally occurring phenolics, such as chlorogenic acids, caffeic, ferulic and cumaric acid, and heat-induced polyphenol-type structures, which are formed due to non-enzymatic browning reactions during roasting. Coffee was indeed demonstrated to act in vitro as pro-oxidant (Andueza et al., 2004, 2009). These apparent contradictory results can be



619

explained considering that, depending on the extent of non-enzymatic browning, pro-oxidants and/or antioxidants can be formed. In the case of the Maillard reaction, the early stages of the reaction are responsible for the formation of prooxidant compounds while in the advanced stages antioxidant products seem to prevail (Manzocco et al., 2001; Calligaris et al., 2004a). These literature data highlight once again the compositional complexity of roasted coffee and explain why the oxidative reaction pathway in coffee product is not yet completely understood. The occurrence of oxidative reactions could also lead to the quality decay of instant coffee, coffee concentrates and drinks. The susceptibility depends greatly on the characteristics of the coffee derivatives and is due to the presence of sensitive compounds extracted from roasted coffee during brewing as well as of other ingredients added to the formula. Volatile concentration could also change due to their release from bean pores. In this way volatiles could be lost together with CO2 and the intensity of these changes is associated with the technological and packaging solutions adopted after roasting. As regards coffee brews, their aroma change during storage has been related to a supplementary mechanism. In particular, the loss of low-boiling potent aroma compounds, particularly sulfurcontaining key odorants responsible for fresh aroma, has been related to interactions with non-volatile components, such as melanoidins. For instance, odour-active thiols could be rapidly covalently bound by melanoidins just after coffee brew preparation causing a decrease in the overall sulphury-roasty odour (Hofmann et al., 2001; Hofmann and Schieberle, 2002; Mueller and Hofmann, 2007). The quality depletion of instant coffee is also due to its high hygroscopic properties. An increase in moisture content to 7±8% is responsible for caking and collapse of powder or granules which become a pasty or solid mass with reduced aroma compound retention (Clarke, 1987b). Additional detrimental events could take place during storage of coffee beverages. Depending on their water activity and composition, coffee concentrates and brews could present microbial risk. Coffee concentrates having 17% (w/w) water content demonstrated to be very stable from a microbiological point of view up to 1 year at temperatures between 4 and 35 ëC. This was attributed to their low pH and redox potential as well as to the presence of melanoidins with strong antimicrobial activity (Manzano et al., 2000). Coffee beverages also show a very low chemical stability, characterized not only by a change in the flavour profile but also an increase in general perceived sourness. These changes are accompanied by a pH decrease, corresponding to an increase in the titratable acidity (Manzocco and Nicoli, 2007; Perez-Martinez et al., 2008a). Quality depletion of coffee brews as well as of coffee concentrates starts immediately after brewing and proceeds at a significant rate, even at subzero temperatures. At present, very little is known about the mechanisms underlying liquid coffee instability. It has been suggested that the decrease in pH could be the consequence of complex reactions, probably related to nonenzymatic browning pathways involving carbohydrates and amino acids.


620


Additional mechanisms involving lactone hydrolysis could also contribute to pH decrease (Balzer, 2001).

21.3

Ensuring stability and extending the shelf life of coffee

In general terms, the kinetics of deteriorative reaction of foods and thus their shelf life is a function not only of intrinsic factors (Ii) typical of the product but also of extrinsic ones (Ei) linked with environmental and packaging conditions: k f Ii ; Ei

21:1

Thus, in order to ensure or extend the product stability, both classes of factors should be considered. Table 21.3 summarizes the main intrinsic and extrinsic factors controlling the kinetics of the critical events which affect the shelf life of coffee derivatives. 21.3.1 Factors controlling the rate of volatile release in coffee derivatives As previously reported, volatile release could greatly affect the shelf life of coffee products. The pressure gradient between internal bean structure and package atmosphere can be regarded as the main driving force affecting the extent of volatile release. Volatile release is greatly affected by both technological procedures applied after roasting (e.g., processing conditions during Table 21.3 Intrinsic and extrinsic factors controlling the kinetics of the critical events which affect shelf life of coffee products Intrinsic factors

Extrinsic factors

Volatile release

Surface area Glass transition temperature (Tg) Other ingredients affecting Tg

Pressure Temperature Relative humidity

Oxidation

Surface area Redox potential Antioxidant and pro-oxidants Other oxidizable ingredients

Oxygen partial pressure Temperature Relative humidity Light

Physical collapse

Glass transition temperature Anticaking agents Other ingredients affecting Tg

Temperature Relative humidity

Ester hydrolysis and non-enzymatic browning

Water activity pH-regulator agents Formulation

Temperature Light

Microbial spoilage

Water activity Redox potential Antimicrobials

Temperature



621

degassing or grinding) and packaging conditions. For instance, volatiles are easily lost together with CO2 through the one-way safety valve which is fitted on the package to avoid it bursting. Volatile release from roasted coffee can be further decreased by creating a pressure inside the package that is higher than the atmospheric one. Such high pressure can be established upon CO2 and volatile release from roasted coffee packed immediately after air-cooling or by sealing the filled containers in the presence of proper gasses at the desired overpressure (Nicoli and Savonitto, 2005). In addition, for non-hermetically sealed products, storage temperature is also expected to be critical for volatile release. In this case, the temperature dependence of volatile release can be well described by the Arrhenius equation from 4 to 40 ëC (Nicoli and Savonitto, 2005). In particular, the temperature sensitivity, expressed in term of Q10, was about 1.5. This means that for any increase of 10 ëC of temperature, the rate of volatile release increases 1.5-fold. One additional factor affecting volatile release is the physical structure of coffee, which is particularly important in the case of instant coffee. Volatile release is affected by coffee water activity (aw) and, consequently, by its glass transition temperature (Tg). The high volatile retention at low aw values, corresponding to a glassy system, was attributed to the entrapping capacity of the amorphous glass, where diffusion is very low. Plasticization by absorption of water may cause the depression of Tg below the room temperature, and hence the glass± rubber transition of the matrix. In these conditions the structural changes of the matrix may allow initial collapse to occur and volatile to be released (Anese et al., 2005). To control the physical state of coffee matrices, the choice of a proper packaging material with adequate water vapour transmission rate is crucial. For instance, the shelf life of instant coffee packed in flexible films seems to be well correlated to their water vapour transmission rate (Alves and Bordin, 1998). 21.3.2 Factors controlling the rate of oxidation in coffee derivatives The control of oxidative reactions still remains one of the major challenges for food scientists not only for coffee derivatives but also in other foods. The only way to design efficient constraints able to hinder lipid oxidation implies the deeper understanding of the factors affecting oxidative reaction. Unfortunately, the chemical structure of the huge number of compounds that could suffer oxidation in coffee derivatives, as well as the reaction pathways involved, is far from being elucidated. Both intrinsic and extrinsic factors play a critical role in determining the oxidation rate. For this reason, a number of different strategies may be simultaneously applied in order to hinder the development of oxidative reactions in coffee products. In other words, different hurdles should be designed to efficiently control the detrimental effects of oxidation. Since the intrinsic factors are hardly changeable, more possibilities come from the careful definition of the environmental factors. As is well known, oxygen partial pressure, temperature, relative humidity and light are critical in determining the oxidation rate and thus the product shelf life.


622


Table 21.4 Shelf life data of roasted whole and ground coffee packed under different conditions (modified from Nicoli et al., 2010) Packaging technique Air Under vacuum Inert gas Pressurization Active packaging a b

Oxygen percentage

Shelf life (months)

21% 4±6% 1±2% < 1% < 1%

1±3a 4±6a/>12b 6±8a/>12b >18a,b >18b

Shelf life data from Nicoli and Savonitto (2005) Shelf life data claimed on product label

As shown in Table 21.4, decreasing the oxygen percentage inside the packaging, the shelf life of roasted whole and ground coffee could be greatly improved. According to Cardelli and Labuza (2001), the ground coffee shelf life was increased by about 20 times when O2 decreases from 21.3 to 0.5 kPa. For this reason, packing in air is nowadays an obsolete solution due to the very short shelf life of the product. The easiest way of modifying the atmosphere composition is based on the application of vacuum. The level of O2 could be reduced up to 4±6% prolonging the product shelf life. The latter could be further improved by reducing the oxygen level up to 1±2% by replacing the air inside the package with an inert gas, such as N2 or CO2. Active packaging has also been proposed to reach oxygen percentage residue in containers of less than 1%. To achieve this, sachets containing oxygen scavenger systems are included inside the package (Vermeiren et al., 1999). Similar oxygen concentration can also be achieved by high pressure packaging, thus hindering oxidative reactions. The extra pressure inside the package increases the retention of volatiles in the lipid phase leading to their protection. In addition, the quality improvement of coffee is achieved by an additional hurdle: the reduction of oil migration under pressurized conditions (Clarke, 1987a). The oil is thus less prone to oxidation because it tends to remain inside the cells. Besides the oxygen percentage inside the packaging, storage temperature can also affect the oxidation rate, following the well-known Arrhenius equation. The effect of temperature on ground and roasted coffee shelf life was studied by Cardelli and Labuza (2001) determining Q10 values and the energy of activation (Ea) for kinetics of sensory deterioration of roast and ground coffee. The results for Q10 indicated 15±23% acceleration per 10 ëC increase in temperature at an oxygen concentration of 10%. However, if coffee derivatives are packed under modified atmosphere conditions, allowing the decrease of oxygen below 0.5%, the rate of reaction change is almost negligible after up to 12 months of storage, independently of storage temperature in the range from 20 to 45 ëC (Nicoli et al., 2009). This means that oxygen concentration within the package seems to be much more critical than storage temperature.



623

Table 21.5 Shelf life data of roast and ground coffee packed under 3.0 kPa oxygen partial pressure and stored at 22 ëC (elaborated from Cardelli and Labuza, 2001) aw 0.11 0.25 0.41

Shelf life (days) 19 9 5

Relative humidity could also affect the development of oxidative reactions. It is well known that lipid oxidation in dried foods is affected by moisture. In extremely dry and extremely moist environments, lipid oxidation develops rapidly, while at intermediate moisture levels, normally corresponding to the monomolecular water layer, the rate of lipid oxidation goes to a minimum (Labuza et al., 1971). In the case of coffee products, the literature reports that water activity plays an important role in determining the acceptability of ground and roasted coffee. Even if no specific evidence is available on the formation of oxidation products as a function of water activity, Cardelli and Labuza (2001) found a decrease in roasted ground coffee shelf life as the water activity increases from 0.10 to 0.41 (Table 21.5). Similar results were obtained by Anese et al. (2006). 21.3.3 Factors controlling the rate of physical collapse in coffee derivatives The occurrence of physical collapse of instant coffee and its derivatives is strictly related to its glass transition temperature, which is the result of its formulation. Coffee melanoidins are known to be relatively low molecular weight polymers thus showing a glass transition below room temperature (Anese et al., 2005). In order to increase glass transition, instant and soluble coffee is generally added with high molecular weight polysaccharides ensuring the free flowing of the powder. In addition, the presence of ingredients other than coffee, such as milk derivatives could greatly affect the caking rate. Anti-caking agents, such as carbonates, silicates and phosphates are extensively used in powdered drink due to their ability to rapidly absorb water excess or other plasticizers up to 2.5 times their weight yet remaining a free flowing powder (Jaya and Das, 2004). Since coffee collapse occurs when its glass transition temperature is exceeded during storage, temperature and relative humidity are critical. For instance, Anese et al. (2005) reported that soluble coffee stored at room temperature at ERH% lower than 35 is in a glassy state while over this critical value, the glass± rubber transition may allow matrix collapse to initiate. The latter is also favoured by the release of water upon crystallization during storage of sugar ingredients such as lactose and sucrose. 21.3.4 Factors controlling the rate of ester hydrolysis and non-enzymatic browning in coffee derivatives Ester hydrolysis and non-enzymatic browning development during storage tend to be particularly critical for coffee derivatives with high water content, such as


624


coffee concentrates and drinks where they are responsible for the increase in acidity and the change in flavour profile. The rate of pH decrease is strongly affected by storage temperature. The temperature dependence is well described by the Arrhenius equation from ÿ30 to 60 ëC for coffee liquids with water activity within 0.85 and 0.99. The development of these chemical reactions is hardly controllable but can be masked by the addition of pH regulator agents such as sodium and potassium carbonate and bicarbonate (Perez-Martinez et al., 2008b). Similar to other foods presenting microbial risk, even in the case of coffee liquids, microbial spoilage is controlled by applying proper thermal or nonthermal pasteurization or sterilization treatments. The majority of the shelfstable brews present on the market are obtained by thermal sterilization or pasteurization. When pasteurization is not performed (e.g., coffee concentrates), antimicrobial substances and/or chilling are used to achieve adequate shelf life (Matsumiya et al., 2010).

21.4

Evaluating the shelf life of coffee

Despite the worldwide importance of coffee products, only limited and contradictory indications on their shelf life are available in the literature. There is even less information about the methodologies used for their shelf life assessment. Due to the poorness of literature data, the identification of a reliable shelf life for product belonging to the complex world of coffee must be performed by applying proper methodologies, specifically adapted to the product considered. However, a basic systematic approach for a cost-effective shelf life determination can be outlined. 21.4.1 Identification of the acceptability limit Before proceeding to the shelf life testing of coffee products, based on laboratory trials, it is necessary to clarify which is the acceptability limit to adopt in the shelf life study. The acceptability limit can be defined as the quality level discriminating products which are still acceptable for consumption from those no longer acceptable (Manzocco et al., 2010). The acceptability limit is often chosen by company management on the basis of available experience of the product performance on the market or on the emulation of competitors. Despite being simple and inexpensive, such procedure is obviously fraught with the risk of critical overestimations or disadvantageous underestimation of the shelf life. This hazard is much more probable in the case of new foods, for which no previous experience is available. In general terms, the acceptability limit may be the result of the application of different criteria depending on product criticism and on the likelihood that a certain default will cause product unacceptability first (Table 21.6). The acceptability limit may be decided by regulations produced by food/beverage



625

Table 21.6 Main criteria for definition of shelf life acceptability limits for coffee products Product life end

Acceptability limit nature

Legal requirements Compulsory default Label claims default Compulsory Excessive consumer rejection

Volunteer

Subject deciding the acceptability limit

Acceptability limit

Authority

Limit value indicated by current regulation Concentration of molecule voluntarily claimed by the producer Maximum risk of consumer rejection considered tolerable by the producer

Producer Producer

authorities. In this case, the limit has to be compulsory respected by the producer in order to market the product. Due to the absence of specific legal requirements for coffee products, this kind of acceptability limit is rarely relevant to this sector. Indeed, it could find an application in the case of chilled liquid coffee suffering microbial growth during storage. Compulsory shelf life limits can also derive from volunteer label claims. In fact, according to the regulation, producers must guarantee the conformity of the product to any claim reported on the label. For instance, the amount of a bioactive compound, claimed on the label of a coffee beverage to increase its functionality, could be regarded as a shelf life acceptability limit. The latter is thus the result of marketing considerations achieved by merging actual product functionality, product positioning on the market and consumer perception of the claim. Since most coffee products do not present safety risks or special claims, in the majority of cases the producers are free to choose their own acceptability limit according to internal policy and quality targets. This is obviously a question of risk management which undergoes an unavoidable level of subjectivity. To this regard, it has been observed that the hazard should not be focused on the properties of the product, rather on the attitude of the consumers to accept or reject it (Hough et al., 2006). This is particularly true for coffee products because the end of their shelf life is strictly determined by the changes in their overall sensory impact and thus in the relevant level of consumer satisfaction/ dissatisfaction. The latter can be evaluated by studying the evolution of the percentage of consumers rejecting the product upon development of unacceptable quality during storage. For instance, at a given storage time, the product is certainly still acceptable to some consumers, despite being rejected by others. The coffee producer can choose to be exposed to more or less risk of product rejection by selecting, as the acceptability limit, the proper percentage of consumers rejecting the product. In other words, the acceptability limit becomes the maximum percentage of consumers that the company can tolerate to dissatisfy at the end of product shelf life.


626


Table 21.7 Percentage of consumer rejection and relevant proposed risk of product rejection (Manzocco et al., 2010 with permission) Acceptability limit (% consumer rejection)

Risk

0 10 25 50 75

Negligible Very low Low Medium High

It must be pointed out that, since coffee quality perception is strictly related to a number of local traditions and drinking habits, studies dealing with consumer±food interaction, whose purpose is to identify product acceptability limits, should be carried out in the country/market in which the product has to be sold. In fact, depending on the different consumer sensitivities to coffee quality in different geographical areas, the same acceptability limit, expressed as consumer percentage rejection, could correspond to different product quality levels. This is the case for the acceptability limits identified for ready-to-drink coffees in eastern and western countries. Table 21.7 shows the relation between acceptability limit and the proposed rejection risk level. In most shelf life studies a medium risk level (50% consumer rejection) is chosen as a reasonable acceptability limit but it has been suggested that lower percentages of consumer rejection could be much more reliable. According to Guerra et al. (2008), the final shelf life value can be affected from 20 to 100% by selecting different risk levels. 21.4.2 Identification of proper shelf life indicators When the acceptability limits are derived from legal requirement or label claim defaults, the indicators to be monitored during storage in order to assess coffee product shelf life are easily defined. They are instrumental indicators describing the evolution of the property (e.g. microbial count, concentration of bioactive molecule) whose limit is set by the regulation or the label claim. By contrast, when the coffee life end is caused by an excessive quality loss, as indicated by too high a rejection percentage of consumers, the study of consumer±product interactions represents the most suitable indicator accounting for coffee product quality and thus acceptability. However, it should be noted that the evaluation of consumer rejection as a function of storage time is a time-consuming and expensive process. It requires a large sample size, a large number of consumers as well as the application of appropriate statistical techniques. These conditions make such kinds of studies, despite being powerful, hard to conduct by company operators to routinely assess shelf life. In order to meet industrial needs, instrumental or sensory attributes, whose evolution is correlated to the coffee product rejection expressed by consumers,



627

could be identified and routinely assessed to detect the end of coffee product shelf life (Garitta et al., 2004; Calligaris et al., 2007). In other words, the coffee producers, after the identification of the acceptability limit expressed as the maximum tolerated consumer rejection percentage, can define the corresponding internal quality standards, described by instrumental or sensory shelf life indicators. Such instrumental or sensory acceptability limits can be simply used in routine shelf life tests. To do this it is necessary to identify the relationship between consumer rejection and quality indices, either instrumental or sensorial. This approach can be applied to coffee products, by addressing the following issues: · how consumer rejection and analytical indicators evolve during coffee product storage · which analytical indicators best correlate with consumer rejection during storage · what value for these analytical indices causes the maximum tolerable risk of consumer rejection to be reached. Figure 21.1 summarizes a possible methodology to answer these questions. The first critical step implies the identification of analytical quality indices which are easily assessable and potentially correlated to coffee sensory perception and thus to its rejection expressed by consumers. In the case of coffee products, different indicators can be identified depending on coffee products (Table 21.8). Peroxide value, which is a simple and widely used index of oxidation development in several foods, has also been used to follow coffee product

Fig. 21.1 Methodology for the definition of analytical indicators accounting for consumer rejection (modified from Manzocco and Lagazio, 2009).


628


Table 21.8 Main indicators of coffee products quality depletion potentially accounting for consumer rejection during storage (modified from Nicoli et al., 2009) Coffee product

Indicator

Roast whole and ground coffee Instant coffee

Peroxide value (Chafer et al., 1998) Head space volatiles (Amstalden and Leite, 2001; Buffo and Cardelli-Freire, 2004; Holscher and Steinhart, 1992)

Coffee concentrates and drinks

Sensory flavour (Perez-Martinez et al., 2008a; Cappuccio et al., 2001; Manzocco and Lagazio, 2009)

Instant coffee

Water activity and glass transition temperature (Anese et al., 2005) Moisture uptake (Alves and Bordin, 1998; Alves et al., 2000) Particle changes (Saragoni et al., 2007)

Coffee concentrates and drinks

pH (Manzocco and Nicoli, 2007) Titrable acidity (Yamanashi et al., 1992) Sensory sourness (Manzocco and Lagazio, 2009; Perez-Martinez et al., 2008a)

stability. However, peroxides are not sensory perceivable compounds and, due to the bell-shape of their evolution during storage, they are likely to be difficult to relate to consumer rejection. In this regard, it is noteworthy that peroxide value of roast and ground coffee shows a dramatic increase after four months of storage in air (Nicoli et al., 1993) while, according to Table 21.4, its shelf life is expected to be lower than 3 months. By contrast, headspace volatiles could represent typical indicators potentially correlated to coffee product acceptability. Among the volatiles, some specific indicators of coffee aroma freshness have been selected: (a) M/B aroma index as the ratio between methylfuran and 2butanone (Reymond et al., 1962); (b) flavour quality index based on five key odorants (hexanal, vinylpyrazine, pyrrol, furfurylmethylketone and pyridine) which shows an inverse linear relationship with the M/B index (Spadone and Liardon, 1989); and (c) M/M aroma index as the ratio of methanol to 2methylfuran (Vitzthum and Werkhoff, 1978). Steinhart and Holscher (1991) suggested that the loss of coffee aroma freshness is due to the loss of certain aroma volatiles (mainly methyl mercaptan) which can be used as an indicator of freshness. Additional indicators can be derived by sensory evaluation of coffee flavour by a trained panel. Indicators of structure modifications leading to agglomeration and caking of the coffee powder could be water activity (aw), moisture content and the glass transition temperature. The indicators of ester hydrolysis and non-enzymatic browning in coffee liquids may be H3O+ concentration, as assessed by a pH meter or by titration, as well as assessment of sourness by a trained panel. An example demonstrating the possibility to `translate' the acceptability limit identified as a percentage of consumer rejection into an instrumental or sensory



629

Table 21.9 pH limit of coffee brews in correspondence with increasing risk of consumer rejection risk (modified from Nicoli et al., 2009). Test performed in the Italian market Risk Very low Low Medium High Very high

pH 95% confidence interval 5.22 5.19 5.14 5.12 5.08

0.03 0.03 0.02 0.02 0.04

acceptability limit was recently reported by Manzocco and Lagazio (2009). In particular, coffee beverages were assessed during storage for consumer rejection as well as for hydrogen ion concentration and intensity of sensory properties measured by a trained sensory panel. Hydrogen ion concentration and sourness evolution during storage correlated best with the percentage of consumers rejecting the product. Mathematical functions predicting the limit value of hydrogen ion concentration and sourness as a function of the risk of consumer rejection were defined. For instance, Table 21.9 shows the estimated pH values of coffee brews in correspondence with increasing risk of consumer rejection. Similarly, in a study addressed to assess the secondary shelf life of roast and ground coffee at 30 ëC, the acceptability limit accounting for medium consumer rejection risk corresponds to a 60% reduction of initial headspace volatile area (Anese et al., 2006). The advantages of the exploitation of shelf life indicators simply assessable by instrumental or sensory analysis are undoubted. Once the analytical limits have been assessed by correlating with consumer rejection risk, further time-consuming consumer tests can be skipped and the analytical indicator can be routinely applied to evaluate the shelf life of the coffee product in the industry quality control programmes. Unfortunately, to our knowledge, very little information about the relationships between the evolution of simple quality indicators and consumer rejection risk during storage of coffee products is available. 21.4.3 Shelf life testing under actual storage conditions Shelf life testing is performed to estimate the length of time needed to reach the acceptability limit and implies the continuous monitoring of the changes of the shelf life indicator during storage of coffee products under controlled environmental conditions. When there is no necessity to speed up shelf life testing, the latter can be carried out under conditions simulating as much as possible those actually experienced by the coffee product on the shelves. The basic requirement is that storage conditions (e.g. temperature, relative humidity, light) during shelf life testing are kept constant and equal to those of real product storage. Data relevant to the evolution during storage of the shelf life indicator (i.e.,


630


acceptability, instrumental or sensory data) are then modelled to obtain proper parameters describing/predicting the quality depletion kinetics. The latter are necessary to compute the shelf life value once the acceptability limit is known. Different approaches can be followed depending on the nature of the shelf life indicator. When the latter is product rejection expressed by consumers, data are analysed using the statistic methodology of survival analysis (Gacula and Kubala, 1975; Gacula and Singh, 1984; Hough et al., 2003). By contrast, if an instrumental or sensory indicator (derived by regulation, claim based or identified by the producer) is available, its changes over storage time are generally submitted to modelling according to the fundamental kinetic principles. An example of the application of consumer rejection modelling for different coffee products stored at 20 ëC is reported in Fig. 21.2. In particular, consumers were asked to give a response of acceptability/rejection of coffee brew, coffee concentrate and instant coffee stored for increasing times. Failure time, which is the length of time until the occurrence of product rejection, is then estimated considering that data are censored observations (Hough et al., 2003). In fact, the exact failure time cannot be systematically observed for all samples. If a consumer perceives the coffee sample as àcceptable' at a certain time t, that sample would be rejected beyond that time, thus the data are right censored. If the consumer response at time t is `rejection', the consumer started rejecting that coffee before time t and the data are left censored. When the same group of consumers is used to assess samples stored for increasing time, intervalcensoring is very common because the consumer can find the product still acceptable at time t but reject it at a following time. Thus, the data are intervalcensored between the two observation times. The censored nature of

Fig. 21.2 Probability of consumer rejection of coffee brew (1.8% w/w), coffee concentrate (93.7% w/w) and instant coffee (100% w/w) as a function of storage time at 20 ëC.



631

acceptability/rejection data implies that they should not be statistically analysed as observations of exact failure time since they have a share of missing information. For this reason, regression analysis should not be performed and parametric distributions should be fitted to the data to estimate the most likely values of the parameters by appropriate statistical techniques (Hough et al., 2003). Shelf life is then estimated from the parametric survival curve (Fig. 21.2) by identifying the time needed to reach the maximum tolerable risk of consumers rejecting the coffee product. In the example, a low risk level (25% consumer rejection) was chosen, leading to shelf life values of about 3 days, 2.5 months and more than one year for coffee brew, coffee concentrate and instant coffee, respectively. In this regard, a number of different software packages can be used to perform survival analysis and obtain shelf life data. However, it is worth noting that reliable shelf life information with tight confidence intervals requires a large sample size (Hough et al., 2006; Guillet and Rodrigue, 2010). In the case of shelf life assessment using an instrumental or sensory indicator, data describing the changes of the coffee quality under conditions simulating actual storage are submitted to modelling according to the fundamental general rate law integrated to obtain the equations of the pseudo zero, first, second or n order. Z t Z I dI k dt 21:2 n I0 I 0 where k is the rate constant and n the reaction order. By solving the integrated forms of Eq. 21.2 as a function of time, shelf life at the actual storage conditions can be calculated: Z C dI In 0 T cost 21:3 SL C k where I0 is the value of food quality indicator just after production, I is the quality indicator value corresponding to the acceptability limit. Figure 21.3 shows an example of the application of the classic kinetic approach to evaluate the shelf life of a coffee concentrate having 15% soluble solids stored at different temperatures. In this case, consumer rejection risk correlated well with hydrogen ion concentration and hence pH (Manzocco and Lagazio, 2009). Based on this, the changes in this instrumental indicator were assessed during storage for each product. pH data were then transformed into hydrogen ion concentration values, which were plotted as a function of storage time. Data were analysed by linear regression according to the zero-order reaction kinetic equation: [H3 O ]t ÿ [H3 O ]0 kt

21:4

where [H3O+]t is the hydrogen ion concentration at time t, [H3O+]0 is the hydrogen ion concentration of the freshly prepared coffee product, k is the apparent reaction rate and t is the storage time.


632


Fig. 21.3 Hydrogen ion concentration of a coffee brew having 15% (w/w) solid concentration as a function of storage time at 0 (n), 10 (l) and 20 (s) ëC. Figure also shows hydrogen ion concentration in correspondence of the lower confidence interval of 50% consumer rejection and relevant shelf life (elaborated from Manzocco and Nicoli, 2007).

Given an acceptability limit of pH equal to 5.16 (6:95 10ÿ6 M hydrogen ion), corresponding to a medium consumer rejection risk (Table 21.7), the shelf life value at each storage temperature can be calculated as follows: SL

[H3 O ]lim ÿ [H3 O ]0 k

21:5

where [H3O+]lim is the hydrogen ion concentration limit (6:95 10ÿ6 M). A similar approach was also used to evaluate secondary shelf life of ground roasted coffee (Anese et al., 2006). It should be noted that secondary shelf life represents the length of time after opening of the package during which coffee products maintain acceptable quality (Cappuccio et al., 2001). Figure 21.4 shows the evolution of total volatile peak area of coffee stored for increasing time after package opening. Data were analysed by linear regression according to the first-order reaction kinetic equation: ln Vt kt ln V0

21:6

where Vt is total peak area at time t, V0 is total peak area of the just opened coffee, k is the apparent reaction rate and t is the storage time. Considering an acceptability limit of total peak area (Vlim) equal to 866736 mV s, corresponding to a medium consumer rejection risk (Anese et al., 2006), the shelf life value can be calculated as follows: ln Vlim ÿ ln V0 21:7 SL k



633

Fig. 21.4 Total volatile peak area of ground roasted coffee as a function of storage time. Figure also shows total volatile peak area in correspondence of 50% consumer rejection and relevant shelf life (elaborated from Anese et al., 2006).

21.4.4 Shelf life testing under accelerated storage conditions Shelf life testing under actual storage conditions is economically feasible only when coffee product quality decays in a reasonably short time. This generally occurs in the case of coffee derivatives with high water content such as coffee liquids or when assessing secondary shelf life of coffee products. Unfortunately, this methodology does not suit industrial needs when dealing with coffee products having a medium to long shelf life, such as roasted whole and ground coffee packed under modified atmosphere. For this reason, it is convenient to accelerate shelf life experiments by testing coffee products under environmental conditions that speed up food quality depletion and then extrapolating the results to milder conditions usually experienced by the product (Mizrahi, 2000). Accelerated shelf life tests (ASLT) have been proven to be effective when both consumer rejection and instrumental shelf life indicators are used. The basic premises for the application of ASLT are: · the quality decay rate varies only as a function of the accelerating factor, while other environmental and compositional variables are kept constant · an accurate kinetic/descriptive model of the quality decay rate is available · the relationship between the accelerating factor and the quality decay rate is known (Manzocco et al., 2010). In the case of coffee products, temperature seems to be the accelerating factor best meeting these requirements. This is due not only to the fact that temperature


634


is one of the most critical extrinsic factors controlling the kinetics of coffee product quality depletion (Table 21.3), but also to the availability of a theoretical basis for the development of a mathematical description of the temperature sensitivity of chemical reaction rates. The Arrhenius equation can be exploited to estimate, by regression analysis, apparent activation energy and frequency factor of the changes in the quality indicator. These parameters are used to estimate the apparent reaction rate at any temperature in the experimental range tested. Such values can be finally integrated to predict the shelf life of the product under actual storage conditions, given the acceptability limit. Table 21.10 reports some examples from the literature relevant to the application of the Arrhenius equation to describe the temperature dependence of coffee product quality depletion. As can be observed, the Ea values reported in the literature for different combinations of shelf life indicators and coffee products vary greatly. This can be attributed to the differences in compositional and environmental factors taken into account in these studies. The temperature dependence of the chemical stability of instant coffee, coffee paste and coffee concentrate stored in conventional atmosphere was well described by the Arrhenius equation with Ea values higher than 50 kJ/mol (Manzocco and Nicoli, 2007). It is noteworthy that the Arrhenius model correctly predicts the rate of hydrogen ion concentration changes in different temperature ranges depending on the product considered. For example, in the case of instant coffee, the Arrhenius equation result is applicable only above 30 ëC. Between 20 and 0 ëC, a very low value was detected causing a deviation Table 21.10 Examples of the application of the Arrhenius equation to describe the temperature dependence of shelf life indicators of coffee products. Estimated values of Ea and relevant reference are also reported Shelf life indicator

Product

Temperature range (ëC)

Ea Reference (kJ/mol)

Hydrogen ion concentration

Instant coffee

30±60

89.5

Coffee paste (93.7% w/w) Coffee concentrate (78% w/w)

0±60

86.6

ÿ30±60

59.4

Manzocco and Nicoli (2007) Manzocco and Nicoli (2007) Manzocco and Nicoli (2007)

Head space volatiles

Roasted coffee packed in air

4±40

28.1

Nicoli et al. (1993)

Consumer rejection

Roast and ground coffee packed under modified atmosphere Roast and ground coffee packed under modified atmosphere in pods

4±35

13.0

Cardelli and Labuza (2001)

4±40

Food and Beverage Stability and Shelf Life (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Functional and Speciality Beverage Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Maillard Reactions in Chemistry, Food and Health (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Yeasts in Food: Beneficial and Detrimental Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Protective Cultures, Antimicrobial Metabolites and Bacteriophages for Food and Beverage Biopreservation (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Colloids: Proteins, Lipids and Polysaccharides (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Emulsions and Foams (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food, diet and obesity (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Endocrine Disrupting Chemicals in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Determination of Veterinary Residues in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Tracing Pathogens in The Food Chain (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Sensory Analysis for Food and Beverage Quality Control: A Practical Guide (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Processed Meats: Improving Safety, Nutrition and Quality (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Environmentally-Compatible Food Packaging (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Handbook of Food Proteins (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Cereals Processing Technology (Woodhead Publishing in Food Science and Technology)

Developing Children's Food Products (Woodhead Publishing Series in Food Science, Technology and Nutrition - Volume 204)

Functional Foods: Principles and Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Dairy-Derived Ingredients: Food and Nutraceutical Uses (Woodhead Publishing in Food Science, Technology and Nutrition)

The Stability and Shelf-life of Food

The Maillard Reaction in Foods and Medicine (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Meat Products Handbook: Practical Science and Technology (Woodhead Publishing in Food Science, Technology and Nutrition)

Food Intolerance and the Food Industry (Woodhead Publishing in Food Science and Technology)

Determining Mycotoxins and Mycotoxigenic Fungi in Food and Feed (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Encapsulation and Controlled Release (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Phytochemical Functional Foods (Woodhead Publishing in Food Science and Technology)

Food Intolerance and the Food Industry (Woodhead Publishing in Food Science and Technology)

Dietary Fibre: Chemical and Biological Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food and Beverage Stability and Shelf Life (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Functional and Speciality Beverage Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Maillard Reactions in Chemistry, Food and Health (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Yeasts in Food: Beneficial and Detrimental Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Protective Cultures, Antimicrobial Metabolites and Bacteriophages for Food and Beverage Biopreservation (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Colloids: Proteins, Lipids and Polysaccharides (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Emulsions and Foams (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food, diet and obesity (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Endocrine Disrupting Chemicals in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Determination of Veterinary Residues in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Tracing Pathogens in The Food Chain (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Sensory Analysis for Food and Beverage Quality Control: A Practical Guide (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Processed Meats: Improving Safety, Nutrition and Quality (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Environmentally-Compatible Food Packaging (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Handbook of Food Proteins (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Cereals Processing Technology (Woodhead Publishing in Food Science and Technology)

Developing Children's Food Products (Woodhead Publishing Series in Food Science, Technology and Nutrition - Volume 204)

Functional Foods: Principles and Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Dairy-Derived Ingredients: Food and Nutraceutical Uses (Woodhead Publishing in Food Science, Technology and Nutrition)

The Stability and Shelf-life of Food

The Maillard Reaction in Foods and Medicine (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Meat Products Handbook: Practical Science and Technology (Woodhead Publishing in Food Science, Technology and Nutrition)

Food Intolerance and the Food Industry (Woodhead Publishing in Food Science and Technology)

Determining Mycotoxins and Mycotoxigenic Fungi in Food and Feed (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Encapsulation and Controlled Release (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Phytochemical Functional Foods (Woodhead Publishing in Food Science and Technology)

Food Intolerance and the Food Industry (Woodhead Publishing in Food Science and Technology)

Dietary Fibre: Chemical and Biological Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Recommend Documents