Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems (Wiley Series Sponsored by IUPAC in Biophysico-Chemical Processes in Environmental Systems)

BIOPHYSICO-CHEMICAL PROCESSES OF ANTHROPOGENIC ORGANIC COMPOUNDS IN ENVIRONMENTAL SYSTEMS Wiley-IUPAC Series in Biophy...

Author: Baoshan Xing | Nicola Senesi | Pan Ming Huang

5 downloads 819 Views 200MB 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

BIOPHYSICO-CHEMICAL PROCESSES OF ANTHROPOGENIC ORGANIC COMPOUNDS IN ENVIRONMENTAL SYSTEMS

Wiley-IUPAC Series in Biophysico-Chemical Processes In Environmental Systems

The Division of Chemistry and the Environment of the International Union of Pure and Applied Chemistry (IUPAC) is sponsoring this series which addresses the fundamentals of physical-chemical-biological interfacial interactions in the environment and the impacts on: (1) the transformation, transport, and fate of nutrients and pollutants, (2) food chain contamination and food quality and safety, and (3) ecosystem health, including human health. In contrast to classical books that focus largely on separate physical, chemical, and biological processes, this book series is unique in integrating the frontiers of knowledge of both fundamentals and impacts of interfacial interactions of these processes in the global environment. Books in the series: Biophysico-Chemical Processes of Heavy Metals and Metalloids in Soil Environments, edited by Antonio Violante, Pan Ming Huang, and Geoffrey M. Gadd Biophysico-Chemical Processes Involving Natural Nonliving Organic Matter in Environmental Systems, edited by Nicola Senesi, Baoshan Xing, and Pan Ming Huang Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems, edited by Baoshan Xing, Nicola Senesi, and Pan Ming Huang

BIOPHYSICO-CHEMICAL PROCESSES OF ANTHROPOGENIC ORGANIC COMPOUNDS IN ENVIRONMENTAL SYSTEMS

Edited by

BAOSHAN XING NICOLA SENESI PAN MING HUANG

Copyright Ó 2011 by John Wiley & Sons, Inc. All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Biophysico-chemical processes of anthropogenic organic compounds in environmental systems / edited by Baoshan Xing, Nicola Senesi, Pan Ming Huang. p. cm. Includes index. ISBN 978-0-470-53963-7 (cloth) 1. Environmental chemistry. 2. Bioorganic chemistry. 3. Anthropogenic soils. I. Xing, Baoshan. II. Senesi, N. (Nicola) III. Huang, P. M. TD193.B5475 2011 628.5—dc22 2010033320 Printed in Singapore oBook ISBN: 978-0-470-94447-9 ePDF ISBN: 978-0-470-94446-2 ePub ISBN: 978-1-118-00211-7 10 9 8 7 6 5 4 3 2 1

CONTENTS

SERIES PREFACE

vii

PREFACE

ix

CONTRIBUTORS

xi

PART I FUNDAMENTAL BIOPHYSICO-CHEMICAL PROCESSES OF ANTHROPOGENIC ORGANIC COMPOUNDS IN THE ENVIRONMENT

1

1

Interactions of Anthropogenic Organic Chemicals with Natural Organic Matter and Black Carbon in Environmental Particles

3

Joseph J. Pignatello

2

Comprehensive Study of Organic Contaminant Adsorption by Clays: Methodologies, Mechanisms, and Environmental Implications

51

Stephen A. Boyd, Cliff T. Johnston, David A. Laird, Brian J. Teppen, and Hui Li

3

The Role of Organic Matter–Mineral Interactions in the Sorption of Organic Contaminants

73

Myrna J. Simpson and Andre J. Simpson

4

Photocatalytic Degradation of Organic Contaminants on Mineral Surfaces

91

Chuncheng Chen, Zhaohui Wang, Wanhong Ma, Hongwei Ji, and Jincai Zhao

PART II ANTHROPOGENIC ORGANIC COMPOUNDS IN AIR, WATER, AND SOIL, AND THEIR GLOBAL CYCLING 5

Sorption of Anthropogenic Organic Compounds to Airborne Particles

113 115

Hans Peter H. Arp and Kai-Uwe Goss

6

Measurement and Modeling of Semivolatile Organic Compounds in Local Atmospheres

149

Songyan Du and Lisa A. Rodenburg

7

Pharmaceuticals and Personal Care Products in Soils and Sediments

185

Bo Pan and Baoshan Xing

v

vi

8

CONTENTS

Fate and Transport of Organic Compounds in (to) the Subsurface Environment

215

Peter Grathwohl

9

Pharmaceuticals and Endocrine-Disrupting Compounds in Drinking Water

233

Daniel W. Gerrity, Mark J. Benotti, David A. Reckhow, and Shane A. Snyder

10

Intermedia Transfers and Global Cycling of Persistent Organic Pollutants

251

Claudia Moeckel and Kevin C. Jones

11

Emission of Polycyclic Aromatic Hydrocarbons in China

267

Shu Tao, Bengang Li, Yanxu X. Zhang, and Huishi Yuan

PART III ANALYTICAL TECHNIQUES 12

Principles and Guidelines of Sampling, Extraction, and Instrumental Analysis Techniques for Measurements of Organic Pollutants in Environmental Matrices

283

285

Eddy Yongping Zeng, Zhaohui Wang, and O. Samuel Sojinu

13

NMR Application in Environmental Research on Anthropogenic Organic Compounds

315

Robert L. Cook

14

Synchrotron-Based X-Ray and FTIR Absorption Spectromicroscopies of Organic Contaminants in the Environment

341

John R. Lawrence and Adam P. Hitchcock

15

Application of Solid-Phase Microextraction in Determination of Organic Compounds from Complex Environmental Matrices

369

Sanja Risticevic, Dajana Vuckovic, and Janusz Pawliszyn

16

Application of Biosensors for Environmental Analysis

413

Marinella Farre´, Sandra Pe´rez, Lina Kantiani, and Dami a Barcelo´

17

Analyses of Drugs and Pharmaceuticals in the Environment

439

Imran Ali, Hassan Y. Aboul-Enein, and Klaus Kummerer

PART IV RESTORATION OF NATURAL ENVIRONMENTS CONTAMINATED BY ORGANIC POLLUTANTS 18

Biochemistry of Environmental Contaminant Transformation: Nonylphenolic Compounds and Hexachlorocyclohexanes–Two Case Studies

463

465

Hans-Peter E. Kohler

19

Biodegradation of Anthropogenic Organic Compounds in Natural Environments

483

Jose-Luis Niqui-Arroyo, Marisa Bueno-Montes, and Jose-Julio Ortega-Calvo

20

Phytoremediation of Soils Contaminated with Organic Pollutants

503

Jason C. White and Lee A. Newman

21

Bioavailability of Hydrophobic Organic Contaminants in Soils and Sediments

517

Wesley H. Hunter, Jay Gan, and Rai S. Kookana

22

Abiotic and Biotic Factors Affecting the Fate of Organic Pollutants in Soils and Sediments

535

Richard E. Meggo and Jerald L. Schnoor

INDEX

559

SERIES PREFACE

Scientific progress is based ultimately on unification rather than fragmentation of knowledge. Environmental science is the fusion of physical and life sciences. Physical, chemical, and biological processes in the environment are not independent but rather interactive processes. Therefore, it is essential to address physical, chemical, and biological interfacial interactions in order to understand the composition, complexity, and dynamics of ecosystems. Keeping separate these domains, no matter how fruitful, one cannot hope to deliver on the full promise of modern environmental science. The time is upon us to recognize that the new frontier in environmental science is the interface, wherever it remains unexplored. The Division of Chemistry and the Environment of the International Union of Pure and Applied Chemistry (IUPAC) has approved the creation of an IUPAC-sponsored book series entitled Biophysico-Chemical Processes in Environmental Systems published by John Wiley & Sons, Hoboken, New Jersey. This series addresses the fundamentals of physical--chemical--biological interfacial interactions in the environment and the impacts on (1) the transformation, transport, and fate of nutrients and pollutants; (2) food chain contamination and food quality and safety; and (3) ecosystem health, including human health. In contrast to classical books that focus largely on separate physical, chemical, and biological processes, this book series is unique in integrating the frontiers of knowledge of both fundamentals and impacts of interfacial interactions of these processes in the global environment. With the rapid developments in environmental physics, chemistry, and biology, it is becoming much harder, if not impossible, for scientists to follow new progress outside their immediate area of research by reading the primary research literature. This book series will capture pertinent research

topics of significant current interest and will present to the environmental science community a distilled and integrated version of new developments in biological physical, and chemical processes in environmental systems. This book, Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems, is volume 3 of the Wiley-IUPAC series. This book comprises 22 chapters by renowned experts in the topic and is unique in integrating both fundamentals and impacts of interfacial interactions of physical, chemical, and biological processes pertaining to adsorption, transformation, bioavailability, toxicity, and transport processes of anthropogenic organic compounds in the air--water--soil environment and their global cycling. Further, the most modern techniques used for sampling, extraction, and instrumental analyses and various means for the restoration of natural environments contaminated by organic pollutants are treated. This book can be used as an advanced reference source on biological, physical, and chemical processes and performance, analytical techniques, and restoration of anthropogenic organic compounds in the global environment for senior undergraduate and graduate students in environmental sciences and engineering. It is an essential reference for chemists and biologists studying environmental systems, as well as for geochemists, environmental engineers; and soil, water, and atmosphere scientists. It will serve as a useful resource book for professors, instructors, research scientists, professional consultants, and other individuals working on environmental and ecological systems. P. MING HUANG NICOLA SENESI Series Editors vii

PREFACE

Anthropogenic organic compounds (AOCs) are synthetically made organic chemicals. They range from gasoline components (e.g., benzene, toluene, xylene) to emerging contaminants such as endocrine-disrupting chemicals. Because of their wide use and disposal, AOCs are commonly found in our environment such as the water we drink, the air we breathe, and the soil from which we obtain our food. These compounds are often toxic and can severely deteriorate an ecosystem. They can also bioaccumulate through food chains and cause various diseases (and even death) to organisms, including humans. AOCs behave differently in various environmental media which differ in their different physical, chemical, and biological components and processes. Therefore, an in-depth and more complete understanding of the biological, physical, and chemical processes of AOCs in environmental systems is essential for the development of innovative management strategies for sustaining the environment and ecosystem integrity. Physical, chemical, and biological interfacial interactions and processes govern the fate, transport, availability, exposure, and risk of AOCs. However, the fundamentals of many physicochemical and biological interfacial reactions of AOCs and their impacts on ecosystems remain largely unknown. As a result, predictive models for their fate, transport, and risk in different media are often off target. To advance the frontiers of knowledge on the subject matter would require a concerted and comprehensive effort of scientists in relevant physical and life sciences such as chemistry; mineralogy; geochemistry; microbiology; ecology; environmental engineering; and soil, atmospheric, and aquatic sciences. In addition, physical, chemical, and biological reactions and processes of AOCs in the environment are not independent but rather interactive and closely interrelated. Therefore, it is essential to systematically address

these interactive processes and interactions through an interdisciplinary approach. Scientific progress in advancing the understanding of environmental fate and behavior of AOCs is based ultimately on integration rather than separation of knowledge across scientific disciplines. To achieve the goal of knowledge integration, this book entitled, Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems, brings together world-renowned scientists on the subject matter across scientific disciplines to integrate the current stateof-the-art knowledge, especially the latest discoveries, development, and future prospects on the research of AOCs in the environment. By virtue of the complex nature of the interactions of AOCs with different environmental components and matrixes, no single available technique, instrument, or model is satisfactory yet for determining their fate, transport, availability, and risk in the environment. In order to fully understand the biological, physical, and chemical interactions and processes of AOCs in the environment, it is critical to know the chemical, physical, and biological properties of AOCs and their analytical techniques. This book is unique because of its multidisciplinary approach, which provides a comprehensive and integrated coverage of biological, physical, and chemical reactions and processes of AOCs in various environments, associated sampling and analytical techniques, and restoration of natural environments contaminated by AOCs. There are 22 chapters in this book, and these chapters are divided into four parts. Part I contains four chapters focusing on the fundamental biological, physical, and chemical processes of AOCs in the environment; Part II, with seven chapters, presents the occurrence and distribution of AOCs in air, water, and soil, and their global cycling; Part III, containing six chapters, discusses the state-of-the-art ix

x

PREFACE

sampling methods and current analytical, biological, spectroscopic, and microscopic techniques for monitoring and studying AOCs; and Part IV consists of five chapters emphasizing the restoration of natural environments contaminated by organic pollutants. This book is an informative and important reference book for scientists, engineers, and professionals who are interested in the biological, physical, and chemical processes and interactions of AOCs in environmental systems. This book is also a critical addition to the existing literature on the subject matter. Further, this book can be used by undergraduate and graduate students, instructors and professors in the disciplines of environmental science and engineering; aquatic, soil, marine, and atmospheric sciences; geosciences; and

ecological, biological, and chemical sciences. Again, the book chapter authors are leading authorities in their respective fields of research. Each chapter was rigorously reviewed externally as for refereed journal articles. We sincerely thank all chapter authors and reviewers who graciously volunteered their time and effort, and contributed their knowledge and wisdom to improve the quality and clarity of this book. We are also highly grateful to the staff of IUPAC and John Wiley & Sons for their strong support and great cooperation in the publication of the book. BAOSHAN XING NICOLA SENESI PAN MING HUANG

CONTRIBUTORS

Dr. Hassan Y. Aboul-Enein, Pharmaceutical and Medicinal Chemistry Department, Pharmaceutical and Drug Industries Research Division, National Research Center, Cairo, Egypt Dr. Imran Ali, Department of Chemistry, Jamia Millia Islamia (Central University), New Delhi, India Dr. Hans Peter H. Arp, Norwegian Geotechnical Institute, Department of Environmental Engineering, Oslo, Norway Dr. Dami a Barcelo´, Department of Environmental Chemistry, IDAEA-CSIC c/Jordi Girona, Barcelona, Spain Dr. Mark J. Benotti, Applied Research and Development Center, Southern Nevada Water Authority, River Mountain Water Treatment Facility, Las Vegas, Nevada Dr. Stephen A. Boyd, Department of Crop and Soil Sciences, Michigan State University, East Lansing, Michigan Dr. Marisa Bueno-Montes, Institute of Natural Resources and Agrobiology of Seville—CSIC, Seville, Spain Dr. Chuncheng Chen, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Photochemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China Dr. Robert L. Cook, Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana Dr. Songyan Du, Department of Environmental Sciences, Rutgers, the State University, New Brunswick, New Jersey Dr. Marinella Farre, Department of Environmental Chemistry, IDAEA-CSIC c/Jordi Girona, Barcelona, Spain

Dr. Jay Gan, Department of Environmental Sciences, University of California, Riverside, California Dr. Daniel W. Gerrity, Applied Research and Development Center, Southern Nevada Water Authority, River Mountain Water Treatment Facility, Las Vegas, Nevada Dr. Kai-Uwe Goss, Department Analytical Environmental Chemistry, Helmholtz-Center for Environmental Research—UFZ, Leipzig, Germany Dr. Peter Grathwohl, Center for Applied Geosciences, University of T€ubingen, T€ubingen, Germany Dr. Adam P. Hitchcock, Brockhouse Institute for Materials Research,McMasterUniversity,Hamilton,Ontario,Canada Dr. Pan Ming Huang (deceased), Department of Soil Science, University of Saskatchewan, Saskatoon, Saskatchewan, Canada Dr. Wesley H. Hunter, Department of Environmental Sciences, University of California, Riverside, California Dr. Hongwei Ji, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Photochemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China Dr. Cliff T. Johnston, Department of Agronomy, Purdue University, West Lafayette, Indiana Dr. Kevin C. Jones, Lancaster Environment Centre, Lancaster University, Lancaster, United Kingdom Lina Kantiani, Department of Environmental Chemistry, IDAEA-CSIC c/Jordi Girona, Barcelona, Spain Dr. Hans-Peter E. Kohler, Department of Environmental Microbiology, Eawag, Swiss Federal Institute of Aquatic Science and Technology, D€ubendorf, ZH, Switzerland xi

xii

CONTRIBUTORS

Dr. Rai S. Kookana, CSIRO Land and Water, Glen Osmond, Australia Dr. Klaus Kummerer, Department of Health Sciences, University Medical Center, Freiburg, Germany Dr. David A. Laird, National Laboratory for Agriculture and the Environment, USDA-ARS, Ames, Iowa Dr. John R. Lawrence, Environment Canada, National Hydrology Research Centre, Saskatoon, Saskatchewan, Canada Dr. Bengang Li, College of Urban and Environmental Sciences, Peking University, Beijing, China Dr. Hui Li, Department of Crop and Soil Sciences, Michigan State University, East Lansing, Michigan Dr. Wanhong Ma, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Photochemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China

Dr. Jerald L. Schnoor, Department of Civil and Environmental Engineering, University of Iowa, Iowa City, Iowa Dr. Nicola Senesi, Department of Biology and Environmental and Agroforestal Chemistry, University of Bari, Bari, Italy O. Samuel Sojinu, Department of Chemical Sciences, Redeemer’s University, Mowe, Ogun State, Nigeria Dr. Myrna J. Simpson, Department of Chemistry, University of Toronto, Toronto, Ontario, Canada Dr. Andre J. Simpson, Department of Chemistry, University of Toronto, Toronto, Ontario, Canada Dr. Shane A. Snyder, Applied Research and Development Center, Southern Nevada Water Authority, River Mountain Water Treatment Facility, Las Vegas, Nevada Dr. Shu Tao, College of Urban and Environmental Sciences, Peking University, Beijing, China

Richard E. Meggo, Department of Civil and Environmental Engineering, University of Iowa, Iowa City, Iowa

Dr. Brian J. Teppen, Department of Crop and Soil Sciences, Michigan State University, East Lansing, Michigan

Dr. Claudia Moeckel, Lancaster Environment Centre, Lancaster University, Lancaster, United Kingdom

Dr. Dajana Vuckovic, Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada

Dr. Lee A. Newman, Biology Department, Brookhaven National Laboratory, Upton, New York

Dr. Ji-Zhong Wang, School of Earth and Space Sciences, University of Science and Technology of China, Hefei, Anhui Province, China

Dr. Jose-Luis Niqui-Arroyo, Institute of Natural Resources and Agrobiology of Seville—CSIC, Seville, Spain Dr. Jose-Julio Ortega-Calvo, Institute of Natural Resources and Agrobiology of Seville—CSIC, Seville, Spain Dr. Bo Pan, College of Environmental Science and Engineering, Kunming University of Science and Technology, Kunming, China Dr. Janusz Pawliszyn, Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Dr. Sandra Perez, Department of Environmental Chemistry, IDAEA-CSIC c/Jordi Girona, Barcelona, Spain Dr. Joseph. J. Pignatello, Department of Soil and Water, Connecticut Agricultural Experiment Station, New Haven, Connecticut Dr. David A. Reckhow, Department of Civil and Environmental Engineering, University of Massachusetts, Amherst, Massachusetts Sanja Risticevic, Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada Dr. Lisa A. Rodenburg, Department of Environmental Sciences, Rutgers, the State University of New Jersey, New Brunswick, New Jersey

Dr. Zhaohui Wang, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Photochemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China Dr. Jason C. White, Department of Analytical Chemistry, Connecticut Agricultural Experiment Station, New Haven, Connecticut Dr. Baoshan Xing, Department of Plant, Soil, and Insect Sciences, University of Massachusetts, Amherst, Massachusetts Huishi Yuan, College of Urban and Environmental Sciences, Peking University, Beijing, China Dr. Eddy Yongping Zeng, State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy Science, Guangzhou, China Yanxu X. Zhang, College of Urban and Environmental Sciences, Peking University, Beijing, China Dr. Jincai Zhao, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Photochemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China

Figure 1.14. Equilibrium configurations for the adsorption of methane and water in a carbon slit pore of width 1 nm with varying active site densities at 300 K. See page 27 for text discussion of this figure.

Figure 2.14. Illustration of sorbed dinitro-o-cresol (DNOC) molecule laying flat on the siloxane surface showing the siloxane surface and water molecules surrounding the exchangeable cations. See page 64 for text discussion of this figure.

Figure 2.15. Molecular dynamics snapshot of dinitro-o-cresol (DNOC) in K-smectite (K-SWy-2) clay interlayer. See page 65 for text discussion of this figure.

Figure 5.1. A SEM scan of a particle sample indicating a large porous conglomerate of organic particles with some salts (yellow circle), a sharp-edged, smooth-surfaced salt crystal (blue circle), a pollen grain (red circle) and a smooth-edged, smooth-surfaced mineral particle (green circles). See page 120 for text discussion of this figure.

Figure 5.5. Illustration of nonspecific interactions. See page 126 for text discussion of this figure.

Figure 5.2. Simultaneous particle and air-phase chemical collectors used in sample-and-extract methods. See page 120 for text discussion of this figure.

Figure 5.6. Illustration of specific interactions. See page 127 for text discussion of this figure. Figure 5.7. Examples of adsorptive interactions at surfaces. See page 128 for text discussion of this figure.

Figure 5.8. The influence of relative humidity (RH) on adsorption to minerals and metal oxides, showing the increase of the water layer thickness with increasing RH. See page 131 for text discussion of this figure. Montmorillonite

Kaolinite

Mica

Water and exchangeable cations O

OH

Si, Al

K

Al, Mg, Fe

Figure 7.3. Structures for typical clay minerals. See page 196 for text discussion of this figure.

Oven dried montmorillonite Silica tetrahedra Alumina octahedra Silica tetrahedra Silica tetrahedra ~1.0 nm

Alumina octahedra Silica tetrahedra

Silicon Oxygen Hydrogen Aluminium

Wet montmorillonite Silica tetrahedra

Cations


Water molecules Hydrated exchangeable cations

3 nm or more

Silica tetrahedra

Oxytetracycline


Figure 7.4. Interlayer spacing of montmorillonite at different conditions. See page 197 for text discussion of this figure. 1628 1709

1580

Free CIP at pH 5 1385 CIP on HFO at pH 5 1360

1633

Absorbance

1619

CIP on HAO at pH 5

1354

1640

1627 1358

1800

1600

1400

1200

Wavenumber (cm-1)

Figure 7.5. ATR-FTIR evidence for ciprofloxacin sorption on hydrous oxides of Al (HAO) and Fe (HFO). See page 198 for text discussion of this figure.

Aqueous phase

+ + -

+

Electrostatic attraction + +

-

Adsorbent Cation exchange

+ - +

-

+

+

+ - + +

PPCP molecule Cation

Electrostatic attraction

+

-

+

+

-

+

Exchangeable cation -

-+

Ligand exchange

Negatively charged site

Cation bridging

Figure 7.7. Schematic illustration of adsorption of PPCPs on solid particles in the presence of cations. See page 200 for text discussion of this figure.

Figure 10.1. Global migration processes of POPs. See page 252 for text discussion of this figure.

Figure 10.2. Schematic representation of the key processes affecting concentrations of POPs in the air, vegetation and soil systems. See page 256 for text discussion of this figure.

Figure 10.3. Schematics of (a) different deposition types supplying airborne POPs to vegetation and soil surfaces and (b) important partitioning processes in the terrestrial environment. See page 257 for text discussion of this figure.

Figure 10.4. The Maximum Reservoir Capacity for HCB, January and July. See page 258 for text discussion of this figure.

Figure 11.8. Six zones for analysis of seasonal variations on different regions. See page 278 for text discussion of this figure.

Figure 11.10. PAH emission density maps for the major emission sources in different seasons. See page 279 for text discussion of this figure.

Figure 14.8. Color coded composite maps (in each case with rescaling within each color) assembled from the component maps presented in Figure 14.7. See page 352 for text discussion of this figure.

Figure 14.9. STXM analysis of a colloid sample from 100 m below Lake Brienz. See page 354 for text discussion of this figure.

Figure 14.13. Contour diagrams from infrared mapping obtained at the end of the experiment showing the spatial distribution of the infrared absorption peaks corresponding to (top) Mycobacterium sp. JLS bacteria, (middle) ESHA, and (bottom) pyrene completely degraded. See page 362 for text discussion of this figure.

Figure 15.15. The comparison of different SPME derivatization approaches for the determination of pesticides in rainwater samples with PDMS/DVB fiber coating and PFBBr derivatization reagent. See page 387 for text discussion of this figure.

Figure 22.4. Conceptual depiction of increasing sequestration of organic pollutants in soil pores with the passage of time. See page 542 for text discussion of this figure.

PART I FUNDAMENTAL BIOPHYSICO-CHEMICAL PROCESSES OF ANTHROPOGENIC ORGANIC COMPOUNDS IN THE ENVIRONMENT

1 INTERACTIONS OF ANTHROPOGENIC ORGANIC CHEMICALS WITH NATURAL ORGANIC MATTER AND BLACK CARBON IN ENVIRONMENTAL PARTICLES JOSEPH. J. PIGNATELLO 1.1. Introduction 1.1.1. Nature of the Sorbents 1.1.2. The Sorption Process in General 1.1.3. Overview of Weak Intermolecular Forces in Physisorption 1.2. Sorption from the Perspective of the Sorbate 1.2.1. Thermodynamic Driving Forces in Sorption 1.2.2. Quantification of Driving Force Contributions 1.3. Sorption from the Perspective of the Sorbent 1.3.1. Shape of the Isotherm 1.3.2. Sorbent Properties of Natural Organic Matter 1.3.3. Sorbent Properties of Black Carbon 1.3.4. Competitive Sorption 1.3.5. Hysteresis 1.3.6. Apportionment of Sorption between NOM and BC in Environmental Samples 1.3.7. Mass Transfer Rates 1.4. Summary and Concluding Remarks

matter (NOM) and black carbon (BC) materials. These organic substances usually dominate the sorption of nonionic compounds except when present in very low levels or when water is scarce. The chapter does not cover sorption to anthropogenic organic waste liquids or semiliquids, such as weathered oil residues (Jonker and Barendregt 2006) and coal tar (Bayard et al. 2000; Khalil et al. 2006), which can play a significant role as sorbent when present in soil or sediment at high levels. The chapter emphasizes sorption from the aqueous phase, since sorption at the relative humidities existing in most environments approaches that under water-saturated conditions. The chapter is not meant to be an exhaustive review, but a personal perspective of the author that emphasizes the author’s own work and the more recent literature. It considers sorption separately from the perspective of the sorbate and the sorbent. Although this distinction is artificial (obviously, they form a complex), it is useful here as an organizational aid.

1.1.1. Nature of the Sorbents 1.1. INTRODUCTION Sorption refers to the process of molecular exchange between molecules in a gas or liquid phase and a solid phase (the sorbent). Sorption to natural solids typically plays a major, fundamental role in a compound’s transport, reactivity, and bioavailability in the environment. The topic of sorption to natural solids is huge and extends over many decades. This chapter emphasizes sorption to organic substances in soils and sediments, including natural organic

Natural organic matter (NOM) is defined as the organic substances remaining after advanced decomposition of biomass below temperatures where pyrolysis becomes important. It includes humic substances (humic and fulvic acids, humin) and geologically older organic matter present in kerogen and coal. Biomass sources of NOM derive mainly from photosynthetic processes of plants and algae and secondary processes of fungi and heterotrophic bacteria. Natural organic matter can exist in a variety of states: dissolved molecules or molecular aggregates, colloidal particles,

Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems. Edited by Baoshan Xing, Nicola Senesi, and Pan Ming Huang. 2011 John Wiley & Sons, Inc. Published 2011 by John Wiley & Sons, Inc.

3

4

INTERACTIONS OF ANTHROPOGENIC ORGANIC CHEMICALS

surface patches or coatings on minerals, intimate complexes with clay-size minerals, and discrete particles. At the primary level NOM is a heterogeneous mixture of functional units within charged, polydisperse molecules that include nonpolar alkyl, carbohydrate-like, protein-like, lignin-like, heterocyclic, and polyaromatic moieties (Hayes and Clapp 2001; Schulten and Schnitzer 1997). This author hesitates to illustrate a hypothetical NOM structure for the reader, since consensus does not exist for one. Most of our knowledge about the structure and composition of NOM has been gained from studies on “dissolved” NOM (DNOM) from natural waters or soil extracts. The distinction between dissolved and colloidal is arbitrary but operationally it is defined by many researchers as the filtration membrane cutoff of 0.45 mm. Most DNOM molecules do not exist individually, but rather are self-associated by hydrogen-bonding and other weak forces in aggregates. Molecules of NOM complex strongly with trivalent metal cations (e.g., Fe and Al) and less strongly with divalent cations and other inorganic ions. The size of individual NOM molecules depends on the source, and may range from a few hundred to 105 daltons (Da) or larger. Most researchers believe that solid NOM can be characterized as a three-dimensional “phase,” consisting primarily of macromolecules with a fraction of molecules below 103 Da. The reader is referred to various treatises and reviews on humic substances (Aiken et al. 1985; Greenland and Hayes 1981; Hayes et al. 1989; Leenheer 2009; Schnitzer and Khan 1978; Stevenson 1994; Sutton and Sposito 2005). Valuable background on geologically aged NOM is given by Allen-King et al. (2002). Solid NOM may be characterized as a randomnetwork macromolecular organic solid. The closest analogy is lignin, the random–network “polymer” in plants derived from phenylpropane units substituted randomly with methoxyl, phenolic, carbonyl, and quinoid groups and giving plants their woody character (Glasser and Kelley 1987). Black carbon is the carbonaceous byproduct of the incomplete combustion of biomass or fuels (Goldberg 1985). It includes char, in which the original material is carbonized in the solid state, and soot, which is formed by the condensation of precursors from the gas phase. Black carbon is a small but important component of total organic carbon in undisturbed soils and sediments, but can be more prevalent in deep-sea and marine shelf sediments, in areas of frequent or recent fires, and in areas of high industrial activity (Masiello and Druffel 1998; Skjemstad et al. 1999). Soot is a component of atmospheric aerosols (Schmidt and Noack 2000). Black Carbon plays important roles in various geo- and bio-geochemical processes, the global carbon cycle, carbon sequestration, radiative heat balance of the planet, and pulmonary toxicity of aerosols (Lighty et al. 2000; Ramanathan and Carmichael 2008). The potential importance of BC as an adsorbent of environmental pollutants was first noted in the pioneering work of Gustafsson and Gschwend (Gustafsson and Gschwend 1997; Gustafsson et al. 1997).

Black carbon is not a single material, but a continuum of materials whose properties depend on source stock and formation conditions. The BC body is composed of single and short stacks of polyaromatic (graphene) platelets rimmed with ketone, ether, hydroxyl, quinoid, carboxyl, and other functional groups. The BC platelet—averaging 7–17 fused rings for biomass chars depending on formation conditions (Brewer et al. 2009; Kaneko et al. 1991)—is considered well represented by the terraced (0001) basal surface of graphite (Donnet et al. 1993). Thus, the BC surface may be considered micrographitic in character. The stacks are arranged in a highly disordered fashion creating a pore network (Boehm 1964; Goldberg 1985; Palotas et al. 1996). The platelets may contain five- and seven-membered rings that introduce curvature, contributing to the disorder (Harris and Tsang 1997; Shibuya et al. 1999). Raw BC is typically highly porous and thus highly surface-active. The pores lie mainly in the micropore ( xg

Step 2

0

S>S 0

xg Step 1

xw Step 2

water-wet sorbent

xg Step 3

x

Figure 1.1. Sorption as a series of steps in a thermodynamic cycle: (1) transfer from water to gas phase; (2) preparation of the (waterwet) sorbent; (3) transfer from the gas phase to the prepared site.

DGgw DGreorg 0

Step 3 xg þ S > x S DGinteract Sum

5

xw þ S > x S0 DGinteract ¼ DGgw þ DGreorg þ DGinteract

ð1:1Þ

Step 1 is the transfer of x from water to the gas phase, encompassing desolvation of x and collapse of the preexisting solvation shell to the bulk water structure. Step 2 is the reorganization of the sorbent matrix needed to accommodate x, such as the formation of a cavity to fit x, or displacement of existing water molecules or other organic molecules or ions at the site. Step 3 establishes interactions of x with the nowprepared sorbent. Sorption from the vapor phase is the same except without step 1. The Gibbs free energy of sorption DGsorp is the net change in free energy summed over all the steps. The DGsorp is related to the dimensionless thermodynamic sorption coefficient Ksorp by

DGsorp ¼ RTlnKsorp ;

Ksorp ¼

as =as afl =afl

ð1:2Þ

where R is the universal gas constant, T is temperature [in Kelvins], a is the activity in the solid or fluid state, and a is the corresponding reference state activity. The value of DGsorp will depend on the choice of the reference state for the sorbed phase. The relationship between Ksorp and a conventional sorption coefficient defined by one of the

isotherm models (see Section 1.3.1) will likewise depend on the choice of the sorbed phase reference state. Note that if the sites are heterogeneous in energy, DGinteract and possibly DGreorg will be dependent on concentration; hence, Ksorp will be concentration-dependent. However, sorption always tends toward linearity (i.e., a constant ratio of sorbed to fluid-phase concentrations) as the concentration in the fluid phase tends toward zero. In the environment many organic contaminants can be present at concentrations in the fluid phase ranging from vanishingly small to the limit of solubility or vapor pressure. 1.1.3. Overview of Weak Intermolecular Forces in Physisorption Noncovalent intermolecular attractive forces are listed in Table 1.1. For physisorbing compounds these so-called weak forces act simultaneously and additively in combinations appropriate to the structures of the interacting species. 1.1.3.1. London–van der Waals and Coulombic Forces. The forces between a nonionic molecule and an uncharged site may include: dipole–dipole, the interaction between permanent dipoles (Type 1 in Table 1.1); dipole–induced dipole, the attraction of a permanent dipole with the dipole that it induces in its neighbor; (Type 2 in Table 1.1); and induced dipole-induced dipole (Type 3 in Table 1.1), the

6

Term (synonym)

Dipole–dipole (dipolar, “Keesom”)

Dipole–induced dipole, (induction, “Debye”)

Induced dipole–induced dipole (dispersion, “London”)

Charge–dipole

Charge–induced dipole (a form of induction)

Charge–charge (coulombic)

H bond

Type

1

2

3

4

5

6

7

TABLE 1.1. Weak Intermolecular Attractive Forces

AH :B

r

Depiction

2ð4pe0 Þ2 e2 r6

Q2 a

6ð4pe0 Þ2 e2 kTr6

Q2 u2

3 a1 a2 I1 I2 2 ð4pe0 Þ2 e2 r6 ðI1 þ I2 Þ

ð4pe0 Þ2 e2 r6

u2 a

3ð4pe0 Þ2 e2 kTr6

u21 u22

Proportional to 1/r2; directional; 70 kJ/mol

Q1 Q2 ð4pe0 Þer

Energy Averaged Over All Orientationsa

Charge–quadrupole

9 (+)

− + −

sandwich

− + −

parallel-displaced

quadraupole vector

+ − +

+ − +

T-shaped

− + −

attraction between like polarized rings

− + −

attraction between oppositely polarized rings

− + −

− + −

− + −

− + − − + −

3M n2 1 2D 4prNA nD þ 2

where M is molecular weight, r liquid density, and NA ¼ Avagadro’s number, 6.02 1023 mol1; the index of refraction is the ratio of the velocity of light in a vacuum to that in the substance]; I—first ionization potential of molecule; (i.e., energy needed to remove the least tightly held electron to infinity: i ! i þ þ e ). Source: After Israelachvili (1992).

a ¼ ð4pe0 Þ

Notation: e0—permittivity (ability to transmit an electric field) of a vacuum, 8.85419 1012C (coulomb) m1 V1; e—dielectric constant of the surroundings (ratio of permittivity in the medium to permittivity of a vacuum); r—intermolecular distance; k—Boltzmann constant, 1.380658 1023 J K1; T—temperature; Q—Coulombic charge ze, where z is formal charge ( þ 1, þ 2, etc.) and e is charge of an electron; u—dipole moment; a—distortion polarizability [response of electron cloud to an applied electric field, related to index of refraction, nD:

Quadrupole

8 − + −

7

8


mutual attraction of momentary dipoles produced by the synchronization of electronic motion in interacting neighbors. These are commonly called dipolar, induction, and dispersion forces, respectively. They are often referred to collectively as London–van der Waals, or simply van der Waals (vdW) forces. The strength of individual vdW forces depends on the separation distance to the inverse sixth power, how the molecules are oriented, and applicable molecular properties of the interacting species such as dipole moment, ionization potential, and polarizability (Table 1.1). Except for very small molecules the most important of the vdW interactions is thought to be dispersion (Hunter 2004). Placing a charge on either the sorbate or the site leads to charge–dipole (Type 4 in Table 1.1) and charge–induced dipole (Type 5 in Table 1.1) interactions, both proportional to the charge and the sixth power of the separation distance. When the sorbate and site are fully and oppositely charged, a Coulombic force (Type 6 in Table 1.1) exists between the charges that is proportional to the first power of the separation distance. While the Coulombic force usually dominates the energy of interaction between ions, it is important to realize that since ions are not ideal point charges, vdW forces contribute significantly. This is true even for simple inorganic ions, as witnessed in chromatographic separations (Fritz 2005). Moreover, the uncharged regions of the organic ion will undergo the other weak interactions permitted by structure and orientation. 1.1.3.2. Hydrogen (H-) Bonding. Hydrogenbonding(type7 in Table 1.1) is the interaction between an acidic proton of a donor and the lone pair electrons of an acceptor (AH :B).It is included among the noncovalent forces; however, whereas weak H bonds are due mainly to dipolar interactions, the covalent nature of the H bond increases with its strength, such that very strong H bonds are essentially three-center, fourelectron covalent bonds (Gilli and Gilli 2000). The most important type of H bond in environmental systems occurs when the atoms A and B terminate in O, N, or S. These “ONS” H bonds are highly oriented (A--H--B angle, 180 15 , except in intramolecular H bonds), and in most relevant environmental systems range in enthalpy from 6 to 70 kJ/mol. Gilli et al. (2009) categorize ONS H bonds into ordinary (Ohb), charge-assisted ( þ /CAhb, þ CAhb, and CAhb), and resonance-assisted (RAhb) leitmotifs: AH : B ½Að1=2Þ H : Bð1=2Þ þ

ð1:3Þ

Ohb þ =CAhb

ð1:4Þ

½A H : B þ

þ CAhb

ð1:5Þ

½A H : B

CAhb

ð1:6Þ

RAH : B¼R () R¼A HBR

RAhb

ð1:7Þ

The Ohb and þ /CAhb leitmotifs differ only in the degree of proton transfer between A and B. Formation of an RAhb takes place when H bonding is accompanied by p–conjugated bond delocalization, such as in the carboxylic acid dimer: H O

R2

H O

R1

O O O H

R1

O R2

O O H

The leitmotifs þ CAhb, CAhb, or RAhb are the strongest H bonds. The strength of the H-bond increases as the difference in pKa values of AH and BH þ (DpKa ¼ pKa,AH – pKa,BH þ ) approaches zero (Gilli et al. 2009). Thus, water (DpKa ¼ 14) forms only weak H-bonds with itself [8 kJ/mol (Silverstein et al. 2000)] and with most ONS functional groups. Water forms stronger H bonds with carboxylic acids, phenols having strongly electron-withdrawing groups, and protonated nitrogen groups. Table 1.2 groups H bonds in terms of strength between organic functional groups relevant to the structures of NOM and many common contaminants. Particularly strong H bonds exist in carboxylic acid dimers, in amide dimers, in conjugate pairs (RCO2H/RCO2, ArOH/ ArO), between RCO2H or RCO2 and certain nitrogen or protonated nitrogen compounds, and between phenols and certain nitrogen compounds (Table 1.2). Hydrogen bonds are also possible between ONS donors and the p bond in alkenes or aromatic rings (the so-called p--H bond); between ONS acceptors and acidic C--H groups, TABLE 1.2. Estimated Strengths of Intermolecular H-Bonded Pairs for Some Relevant Functional Group Pairs Weak H Bonds (4–17 kJ/mol) Most pairwise combinations of aliph-OH, R3N (1 , 2 , and 3 ), R--O--R, R2C¼O, RNO2, RC(¼O)NR2, RCN H Bonds of Moderate Strength (17–29 kJ/mol) Pairs of RC(¼O)NHR or RC(¼O)NH2 ArNR2 with Ar--OH or RCO2H Ar--OH with RCO2 aliph-OH with ¼N-Strong H Bonds (29-67 kJ/mol) Dimers of RCO2H RCO2H with RCO2 ArOH with ArO RCO2H with ¼N-RCO2 with ¼NH þ-- or R3NH þ ArOH with ¼N-ArO with ¼NH þ-- or R3NH þ Notation: aliph—is aliphatic; Ar—is aromatic; R—is either aliphatic or aromatic group bonded through C; ¼N-- is N that is doubly bonded to C, such as in azine- and azole-type (i.e., heterocyclic aromatic N not in conjugation with the ring), or oxime (R2C¼N--OH or RCH¼N--OH) functional groups. Source: Gilli et al. (2009).

INTRODUCTION

and between acidic ONS protons and aliphatic halogen atoms. As these H bonds are, with some exceptions, much weaker than ONS-type H bonds, they may not be able to compete with water when water is abundant. 1.1.3.3. Interactions between p-Conjugated Systems. It has been realized for a long time that special weak interactions may occur between arene units. Arene is defined as aromatic and related cyclic p conjugated systems. Arene–arene interactions are believed to play a role in many chemical and biological phenomena (Hunter et al. 2001; Hunter and Sanders 1990; Janiak 2000; Meyer et al. 2003; SchmidtMende et al. 2001). It now appears likely that they play a role in environmental partitioning as well. The nature of arene–arene interactions is not fully understood. The strongest interactions occur when one unit is p-electron rich (donor, D) and the other is p-electron poor (acceptor, A). These interactions are known as p-stacking or p–p electron donor–acceptor (p–p EDA) interactions. The units undergoing p–p EDA interaction typically associate in a sandwich, but more often in a parallel displaced fashion (Table 1.1). According to current understanding (Cockroft et al. 2007; Gung and Amicangelo 2006; Hunter et al. 2001; Hunter and Sanders 1990; Sinnokrot and Sherrill 2006), the p–p EDA complex is a hybrid structure that can be written n o D A ¼ ðD AÞquad () ðD AÞvdW () D þ A CT

The 1 : 1 C6H6: C6F6 complex forms a solid that crystallizes in a parallel displaced arrangement and melts 20 K higher than either monomer (Williams 1993). While neither monomer has a permanent dipole moment, the 1 : 1 gas–phase complex has a dipole moment of 0.44 Debye, about one-eight of a charge (Steed et al. 1979). All VdW forces between rings and their substituents are believed to be minor because they largely cancel out with those previously existing between momomers and displaced solvent molecules (Cockroft et al. 2007; Hunter 2004). (However, they obviously would play an important role in gas-solid adsorption.) The CT structure is the dative bond formed by one electron transfer from the highest occupied molecular orbital of D to the lowest unoccupied molecular orbital of A. It contributes primarily to the excited state, often yielding an absorbance band in the near UV–visible region, but may contribute appreciably to the ground state energy in strong D A complexes (Gung and Amicangelo 2006). The p–p EDA bond enthalpy approaches that of a strong H bond (Foster 1969; Perry et al. 2007). It scales with the respective donor and acceptor ability of the opposing units determined by substituents that donate or withdraw electron density, especially through the p system. The flow of electron density is illustrated below: a

ð1:8Þ representing quadrupolar, vdW, and charge transfer interactions between ring systems. In solution, ðD AÞquad is believed to dominate the energy. In general, p systems have a quadrupole vector perpendicular to the plane of the nuclei, where the p cloud on either side of the plane is polarized oppositely to the s-bond framework (type 8 in Table 1.1). Complexation may result if interactions between quadrapoles is net attractive. A classic example is the 1 : 1 complex between benzene and hexafluorobenzene (Williams 1993). The monomers have quadrapole moments of opposite sign and magnitude (33.3 and 31.7 C/m2, respectively (Vrbancich and Ritchie 1980). Benzene is electron rich in the p cloud and electron poor in

+

N

9

d

Multiple alkyl and/or --NR2 groups confer strong donor character. Donor strength also trends with p cloud polarizability; thus, polyaromatic hydrocarbons (PAHs) are p donors whose strength appears to increase with fused ring number, at least up to 4. Especially strong electron withdrawing groups include --NO2, --NR3 þ , --CF3, --C(¼O)R, --SO2R, and --CN. Halogen has opposing s-withdrawing and p-donating effects; yet a ring may become net p-accepting as the number of halogens exceeds a critical number, as occurs with F. Especially favorable complexation occurs when the acceptor is a charged heteroaromatic ring. In this case stability is aided by a charge–quadrupole (“pcation”) interaction (Type 9, Table 1.1) (Qu et al. 2008). An example is the complex in

N

H

N H

the s bond system, while hexafluorobenzene is polarized in the opposite sense. Thus, attractive forces occur between the p clouds and to a lesser extent between the s systems.

H

N H

water between phenanthrene and o-phenanthroline mono- or dication (Wijnja et al. 2004), leading to a dramatic increase in the apparent solubility of phenanthrene.

10


The p–p EDA complexes form readily in water and in both polar and apolar organic solvents (Breault et al. 1998; Ferguson and Diederich 1986; Foster 1969), although the relation with solvent polarity is complex (Breault et al. 1998). Like-polarized arenes (D D; A A) associate much more weakly than do oppositely polarized arenes, and their complexes tend to orient T-shaped in order to maximize p–s and minimize p–p interactions. Thus, benzene and small PAHs orient roughly T–shaped in their crystals (Newcomb 1994), and benzene orients T-shaped in its most stable gas–phase complex (Iimori et al. 2002; Morimoto et al. 2007).

that raises an entropy and/or enthalpy penalty for dissolution in water. It should be noted that hydrophobicity may have a different connotation when speaking of the distribution of an organic molecule among qualitatively different microdomains within a solid (i.e., NOM phases, narrow pores, and near surfaces) because water exists in a different organizational state in those environments than it does in bulk water. The following are the salient features of the hydrophobic effect according to contemporary understanding. .

1.2. SORPTION FROM THE PERSPECTIVE OF THE SORBATE 1.2.1. Thermodynamic Driving Forces in Sorption The contribution of each force listed in Table 1.1 to the net driving force for sorption from aqueous solution depends on the difference in free energy of the interaction with water compared to that with the sorbent. Teasing apart the contributions of individual noncovalent forces in sorption has been a challenge. The following discusses the state of our knowledge about the contributions of individual noncovalent forces to the sorption free energy. Such a discussion must consider both positive and negative driving forces; examples of the latter are solvation and steric hindrance to sorption. 1.2.1.1. Desolvation and the “Hydrophobic Effect”. A major driving force for sorption from the aqueous phase of nonionic compounds, whether to minerals or organic matter, is the exclusion of the organic solute from water known as the hydrophobic effect. This is supported by a vast and relentless stream of data. The term hydrophobic effect refers to the forces that limit the solubility of apolar molecules—or parts thereof—in water. It is also terminology that refers to the clustering of hydrophobic units such as surfactants (surfaceactive agents) into micelles, the folding of biological molecules, and the behavior of solutes and water near apolar surfaces. Because of the central role of the hydrophobic effect in chemistry and biology, many studies have been devoted to a resolution of its underlying nature. For reviews, see Chandler (2005), Lazaridis (2001); and Southall et al. (2002). While concordance has not yet been reached on all its aspects, researchers are in general agreement that the hydrophobic effect arises from disruption of the cohesive energy of water, not from any special attraction between hydrophobic molecules nor of any special repulsion between apolar entities and water molecules. The disruption originates in the greater ordering, and fewer—if not stronger— water–water H bonds within the hydration shell surrounding the hydrophobic entity than in bulk water itself, a situation

.

.

.

.

There is no evidence for long-range forces between hydrophobic entities. The solvation free energy for alkanes in water is pairwise-additive (Wu and Prausnitz 2008); that is, the Henry law constant, which is proportional to the free energy of solvation, is linearly related to the number of C--C bonds. This means that the hydrophobic effect is primarily a local phenomenon limited to only a few angstroms (Chandler 2005; Wu and Prausnitz 2008). Aggregates of nonpolar molecules less than 1 nm in radius are unstable in water (Chandler 2005; Southall et al. 2002). This means that nonpolar entities such as alkanes (Wu and Prausnitz 2008) and PAHs (Wijnja et al. 2004) have no tendency to self-associate in water much below their solubility limit. This and the previous point are consistent with a solvent-centered, rather than solute-centered origin, of the hydrophobic effect. Since there appears to be no special driving force for association of hydrophobic entities, terms such as “hydrophobic bonding” and “hydrophobic interactions” are misleading. The hydrophobic effect is sometimes described as being entropy-driven. In fact, both entropy and enthalpy play a role, depending on the solute and temperature (Southall et al. 2002). The dissolution of hydrocarbons is unfavorable in cold water because of entropy, and in hot water because of enthalpy. Smaller solutes tend to force ordering of neighboring water molecules, resulting in an entropy penalty for dissolution, whereas larger solutes tend to be more effective at breaking water–water H bonds, resulting in an enthalpy penalty. For water at the interface of an organic liquid or an extended hydrophobic surface, each water molecule participates in about one fewer H bonds than in bulk water (Chandler 2005). However, the H bonds near a hydrophobic surface appear to be stronger; the enthalpy of the H bond of water in the first hydration shell of argon in water is 2.1 kJ/mol greater than in bulk water (Silverstein et al. 2000). Some OH groups of water at thewater–liquid interface of hydrocarbons or chlorinated hydrocarbons orient toward the organic phase in the fashion, --O--H organic. That this interaction is attractive is indicated by the shift to

SORPTION FROM THE PERSPECTIVE OF THE SORBATE

.

.

lower energy in the stretching frequency of such OH groups relative to that of the O--H protruding into air at the water–water vapor interface (Moore and Richmond 2008). This is consistent with a generally attractive interaction between water and “hydrophobic” surfaces. Due to vdW interactions of water with the hydrophobic entity, the interfaces between water and organic liquids or between water and hydrophobic surfaces are sharp, with no evidence for a vacuum-like gap (Chandler 2005; Moore and Richmond 2008). While the hydration shell of apolar entities may be more ordered than bulk water, both the “iceberg” structure originally proposed by (Frank and Evans 1945) and the “crystalline water cage” structure that forms around the clathrate hydrates of noble gases and some small hydrocarbons at suitable temperature and pressure (Bontha and Kaplan 1999; Jeffrey 1984) are probably exaggerations of its true nature.

Any given functional unit of a molecule introduces opposing forces for sorption. On one hand, its size and polarizability tend to disrupt the structure of bulk water and drive the molecule to the solid; on the other hand, its permanent polarity tends to drive the molecule to the aqueous phase. This is best exemplified by the halogens. Halogen (except F) is large and polarizable, contributing to molecular hydrophobicity. By contrast, halogen polarizes the C--X bond because of its electronegativity. We know that bond polarization plays a role because hydrophobicity (as octanol–water partition coefficient, Kow) increases with halogen substitution more so for sp2- than sp3-substituted carbon (alkane < alkene < aromatic ring). The same is true for most functional groups bonded to carbon through O, N, or S. This is so for two main reasons. (1) , sp2 carbon is more electronegative and competes better for the s-bond electrons than does sp3 carbon; and (2) , in sp2 systems, delocalization of a halogen electron pair into the p system counteracts the bond dipole moment. δX δ+ sp3 halogen

δX δ+

+ X sp2 halogen

The question arises about the importance of dispersion as a driving force for sorption. Many will point to the general relationship between molecular size of a series of apolar compounds (or apolar functional groups) and sorption intensity as support for the importance of dispersion interactions with the sorbent. However, it is believed that dispersion forces between a hydrocarbon entity and its water solvation shell are similar in magnitude to those between the entity and its hydrocarbonaceous neighbors when dissolved in a

11

hydrocarbon solvent (Israelachvili 1992). It follows that, in the process of exchange between the aqueous phase and the sorbent, dispersion interactions roughly cancel out, and dispersion per se is not a major driving force in sorption, but rather the effect of molecular size is manifested in the hydrophobic effect. By contrast with solution–solid sorption, gas–solid sorption is highly driven by dispersion, since intermolecular forces in the gas phase are negligible. 1.2.1.2. Hydrogen Bonding and Dipolar Interactions. Hydrogen bonding favors sorption only to the extent that the net free energy of H bonding with the sorbent is greater than that with water. Hydrogen bonding opportunities are greater in water because of the sheer abundance of H-bond donor and acceptor groups—111 and 55.5 mol/L, respectively. On the other hand, NOM and BC both are rich in carboxylate and phenolate groups, that tend to form stronger H bonds than water, especially with solutes that have phenolic, aromatic amine, amide, and carboxylate groups (Table 1.2).

OH

H-bonding with water

H-bonding with sorbent

Given the abundance of H-bonding groups in NOM, H bonding of contaminants with H-bonding substituents almost certainly takes place in this phase. Nevertheless, direct evidence for this has been elusive. Dixon et al. (1999) observed a progressive downfield shift in the 19 F NMR signal of 4-fluoroacetophenone with increasing fulvic acid concentration in methanol–water, but whether this shift was due to H bonding is questionable since it was not reversed by addition of a six fold excess of acetophenone, which arguably should have outcompeted the fluoro derivative for H-bond acceptor sites interacting with the ketone O. Welhouse and Bleam (1993a, 1993b) showed by 1 H NMR that atrazine dimerizes in CCl4 and also forms H bonds (as donor or acceptor) of moderate strength with mono functional molecules. With bifunctional molecules such as carboxylic acids and amides, atrazine forms strongly H-bonded cyclic complexes by simultaneously accepting and donating a H bond, as in the structure below: Cl N (CH3)2CHNH

N NCH 2CH3

N

H H O

O R

12


Working with the trifluoroethyl derivative of atrazine in 10% humic acid at pH 11.8, Chien and Bleam (1997) inferred H bonding of atrazine to DNOM by the apparent absence of 19 F NMR resonances characteristic of atrazine dimers. The contribution of H bonding to atrazine sorption, however, remains unclear. In order to gain insight into the driving forces for sorption to NOM, Borisover and Graber (2003) transformed regular soil–water isotherms to soil–hexadecane isotherms (Fig. 1.2) by converting aqueous concentration to the thermodynamically equivalent n-hexadecane concentration via Henry’s constant ratios or solubility ratios. This amounts to switching the reference state from the pure liquid/subcooled liquid to the dilute solution in an “inert” solvent (n-hexadecane) capable only of dispersion. The resulting “rebuilt” isotherms (Fig. 1.2) therefore represent the difference in free energy of interactions in the water-wet solid and the inert solvent, and thus highlight interactions with NOM of a more polar nature. Compounds showing exceptionally strong sorption to a peat soil (45% OC) were those capable of forming strong H bonds with carboxyl and phenolic groups on NOM: 3-nitrophenol, phenol, pyridine, benzoic acid, benzyl alcohol, atrazine, and 2,4-dichlorophenol. Compounds of intermediate sorption strength included trichloromethane and those capable of serving as H-bond acceptors to form weaker H bonds: acetophenone, anisole, 2-chloronitrobenzene, and

NPh Py Atr AP CNB

106 105

nitrobenzene. Compounds sorbing more weakly were apolar aliphatic and aromatic compounds. Interestingly, chlorinated aliphatic compounds sorbed more strongly than did chlorinated aromatic compounds. This suggests that dipolar interactions between the bond dipoles of chlorinated aliphatics, which are greater than those of chlorinated aromatics, engage in dipolar interactions with NOM to favor their transfer from the hydrocarbon phase. 1.2.1.3. p–p EDA Interactions. When permitted by structure, p–p EDA interactions can only add to the driving force of sorption, since water is incapable of p–p EDA interactions. Many pesticides, explosives, antibiotics and other classes of environmental contaminants contain arene groups with strong p–donor or p-acceptor character. In addition, certain natural compounds in soil can act as strong acceptors—there are many examples of quinones among plant signaling chemicals and allelochemicals, such as juglone (5-hydroxy-1,4naphthoquinone). Heteroaromatic amines have been detected in surface and subsurface waters contaminated by shale oil or coal liquification wastes (Sims and O’Loughlin 1989). Humic substances are rich in p-acceptor units, including quinones, charged heterocyclic amines, and aromatic rings having multiple electron-withdrawing carbonyl groups. It has been postulated that the CT bands from internal quinone–hydroquinone complexation partially account for absorbance of

Phe BA DCP NB Ari

Solution concentration, mg/kg

104 103 102 101 100 DBP TCM CH DCCH 1,2-DCB 1,3-DCB 1,4-DCB 1,2,4-TCB 1,2,3-TCB 1,3,5-TCB Bcnz Phcn Naphi Tol

10-1 10-2 10-3 10-2

10-1

100

101

102

103

104

105

106

Solution concentration in n-hexadecane, mg/L

3-nitrophenol NPh Phe phenol py pyridine BA benzyl alcohol Atr atrazine DCP 2,4-dichlorophenol AP acetophenone NB nitrobenzene CNB 2-chloronitrobenzene Ani anisole TCM trichloromethane DBP 1,3-dibromopropane CH cyclohexane trans-1,2-dichlorocyclohexane DCCH Phen phenanthrene Napht naphthalene Tol toluene Benz benzene 1,2-DCB 1,2-dichlorobenzene 1,3-DCB 1,3-dichlorobenzene 1,4-DCB 1,4-dichlorobenzene 1,2,4-TCB 1,2,4-trichlorobenzene 1,3,5-TCB 1,3,5-trichlorobenzene 1,2,3-TCB 1,2,3-trichlorobenzene

Figure 1.2. Pahokee peat-n-hexadecane sorption isotherms (Borisover and Graber 2003). The concentration in hexadecane is calculated by the concentration in water times the ratio of the solubility in hexadecane to that in water.


humic substances in the visible region (Del Vecchio and Blough 2004). Humics may also contain strong p-donor units such as alkyl-substituted rings, polyaromatic rings, and pyrrole-type heterocyclic rings. The graphene platelets of black carbon, with their polyaromatic surfaces, may contain both electron-rich and electron-poor regions, depending on their size and the distribution of functional groups along their rims, which may attract p-acceptor and p-donor molecules, respectively (Zhu and Pignatello 2005a) (see Fig. 1.3). The term “p-p interactions” has been used rather loosely in the more recent environmental literature. Speculation is the norm, and the nature of the force is frequently misunderstood. Nevertheless, evidence is emerging for a contribution of p–p EDA interaction to sorption of some compounds by NOM, as well as by BC. Mixing of substituted pyridines and triazine herbicides with DNOM is reported to generate CT absorbance at 460 nm, suggestive of p–p EDA (M€ uller-Wegener 1987). Sorption of the p-donor compounds, pentamethylbenzene, naphthalene, and phenanthrene, to several soils increased with decreasing pH from 7 to 2.5 (Zhu et al. 2004), whereas no similar pH effect occurred for non-p-donor hydrophobic compounds. Other possible systematic affects varying with pH were ruled out. These p-donor solutes may interact with p-acceptor sites in NOM, such as aromatic rings with multiple carboxyl groups or charged aromatic and heteroaromatic amines, whose acceptor ability increases with degree of protonation (i.e., --CO2H and --NH þ ¼ are more electronegative than --CO2 and --N¼, respectively). Pairwise complexation in methanol–water between donors, pentamethylbenzene, naphthalene or phenanthrene, and each of the following model NOM acceptors was identified spectroscopically: 1,3,5-benzenetricarboxylic acid, 1,4,5,8-naphthalenetetracarboxylic acid, and pyridine. Donor complexes with these model acceptors was pH-dependent and gave a CT band in the UV/visible spectrum and upfield 1 H NMR chemical shifts indicative of face-to-face association. On the basis of the free-energy relationship to be discussed in Section 1.2.2, p-p EDA interactions made a small contribution (5%–8%) to the free energy of phenanthrene sorption in a soil (Zhu and Pignatello 2005b). Other, more indirect evidence for p–p EDA interactions of PAHs with NOM or DNOM exists. Polycyclic aromatic hydrocarbons (PAHs) consistently sorb more strongly to soils than do PCBs of comparable hydrophobicity (Kow) (AllenKing et al. 2002; Chiou et al. 1998; Cornelissen et al. 2004; van Noort 2003); PAHs are strong p-donors, while PCBs are not strong donors or acceptors. Isotherms of PAHs to DNOM are frequently nonlinear (Laor and Rebhun 2002; Polubesova et al. 2007), suggesting that specific interactions take place. In terms of the n-hexadecane reference state, aromatic hydrocarbons sorption to Pahokee peat (Borisover and Graber 2003) follows the order, phenanthrene naphthalene > benzene, the same order as their p-donor capablility.

13

A number of papers have postulated “p–p charge transfer” processes between pesticides (triazine, urea, and bipyridylium herbicides) and humic substances based on elevated free radical concentrations measured by ESR spectroscopy (e.g., Senesi et al. 1995; Sposito et al. 1996). Whatever the source of free radicals, their presence cannot be related to the reversible p–p EDA complex referred to here, as no electron transfer to produce free radicals takes place. Moreover, free radicals would likely react with O2 or couple with NOM, resulting in loss of solute identity. Evidence also exists for p–p EDA interactions with elemental carbonaceous materials. Zhu and Pignatello (2005a) found that adsorption of nitroaromatics on nonporous microcrystalline graphite and on BC (wood charcoal) is far greater than predicted by the hydrophobic effect, based on a calibration set of compounds, and in accord with their p-acceptor strength (mono- < di- < trinitrotoluene). Hydrogen bonding of the nitro groups was ruled out. Adsorption of PAHs on the same solids was likewise greater than predicted by the hydrophobic effect and followed the p-donor strength (naphthalene < phenanthrene). Complexation between the PAH donors and the nitroaromatic acceptors was observed in chloroform. These complexes displayed upfield shifts of NMR spectral frequencies due to ring current effects, which is indicative of face-to-face association. They also gave CT bands in the visible region often seen with p–p EDA complexes. The association constant followed the order in expected strength of D A interaction; namely, mono- < di< tri-nitrotoluene with a given PAH, and naphthalene < phenanthrene < pyrene with a given nitroaromatic. Figure 1.3 shows the strong relationship between the free energy of molecular complexation in chloroform solution and the excess free energy of nitroaromatic adsorption on graphite based on the hexadecane reference state. Chen et al. (2007b) confirmed p–p EDA interactions between polynitroaromatic compounds and the graphene-like surface of carbon nanotubes using a similar approach as Zhu and Pignatello (2005a). Taken together, the results indicate that the graphene surface may be amphoteric with respect to p-interactive adsorbates; referring to Figure 1.3, electron rich regions of the surface attract strong p acceptors while electron-poor regions attract strong p donors. Polarization of the graphite surface near defects and edges is visible by scanning tunneling microscopy (McDermott and McCreery 1994). Figure 1.3 further presents a hypothetical structure of BC showing a potential donor region overlying the polarizable polyaromatic center of the sheet, and acceptor regions in the vicinity of p-acceptor moieties along the rims. 1.2.1.4. Steric Effects. Steric hindrance is possible both in partitioning and adsorption, but very little attention has been paid to it until relatively recently. Steric effects in partitioning to NOM have not been systematically investigated. Partitioning into a solid phase involves the opening of a cavity for


-Ge of adsorption on graphite relative to hexadecane reference state (J/mol)

14

26000 24000 TNT

22000 20000

DNT

18000 16000 14000 with D = NAPH with D = PHEN with D = PYR

MNT

12000 10000

-4000 -2000 0 2000 -ΔG of molecular complexation with model donor in chloroform solution (J/mol) (a)

step defect

⇑

⇓

δ−

⇑

δ+

δ−

edge defect

(edge-on view of graphite layers)

π-donor surface

CO2

O O

NH

O OH O

CO2

π -acceptor surface

Figure 1.3. (a) Free-energy relationship between excess adsorption on nonporous graphite and complexation with p donors, naphthalene, phenanthrene, or pyrene in chloroform measured by 1 H NMR for mononitrotoluene, 2,4-dinitrotoluene, and trinitrotoluene. The excess adsorption free energy is postulated to originate from p–p EDA interactions on the graphene surface. (bottom) Strong p acceptors (e.g., trinitrotoluene) interact with electron-rich regions of the surface, while strong p donors (e.g., phenanthrene) interact with electron-poor regions. Surface charge separation may be due to step/ edge defects (graphite) or rim functionality [black carbon (BC)]. A hypothetical BC platelet with potential donor and acceptor regions is indicated.

the incoming molecule if one does not already exist. The free energy of partitioning from water must include a term for the difference in free energy of cavitation in the solid and in the aqueous phase. The excess free energy required for opening a spherical cavity of radius r in a liquid is the sum of a volumetric term (4/3 pr3E), where E is the volumetric energy density, and a surface term (4pr2c), where c is surface tension (Hummer et al. 1998; Southall et al. 2002). The volumetric term

dominates for small molecules, whereas the surface term dominates for very large molecules or extended surfaces. Cavitation of a solid phase strongly depends on its viscoelastic properties and free volume distribution. For a flexible solid in its equilibrium state, the free-energy cost for forming a cavity (“cavity penalty”) increases systematically with penetrant size for a homologous series of penetrants; atoms or very small penetrants may fit into existing free volume, thus requiring little or no cavitation, whereas complete


Cl

vs A

Cl

B

Cl

Cl Cl

Cl vs Cl

A

pore network

15

B

Cl Cl

Cl

A can achieve a flatter conformation on the surface and better fit into pore throats than B

Figure 1.4. Illustration of the contact area and size exclusion hypotheses for steric effects in adsorption.

exclusion from the internal phase would be reached at some very large size. Suppressed sorption on the basis of molecular size for large molecules has been reported to occur in phospholipid bilayers (Dulfer and Govers 1995; Gobas et al. 1988; Kwon et al. 2006; Yamamoto and Liljestrand 2004) ordinarily regarded as being a partition medium. Steric effects are known to play a role in the adsorption of organic compounds to matrix-impenetrable solids (K€arger and Ruthven 1992), both porous and nonporous, and have been claimed for natural solids. There are at least two sources of steric effects, illustrated in Figure 1.4: deviation from molecular planarity size exclusion. (1) Since interactions depend on close molecular approach, deviation from molecular planarity should reduce contact area with a flat surface, other molecular properties being equal (2) Steric effects in microporous solids can be manifested by size exclusion at pore throats. The contact area hypothesis for adsorption to a smooth surface has been tested for alkanes. As they are able to readily adopt a planar conformation, normal alkanes are expected to achieve closer contact with a surface than cycloalkanes, which are restricted by ring puckering. Closer contact favors vdW forces, whose magnitudes are inversely related to separation distance to the sixth power (Table 1.1). The difference in sorption coefficient between n- and c-alkanes, however, is small: the ratio Kn/Kc for gas-phase transfer to the surfaces of liquids and nonmicroporous inorganic solids is close to unity (0.83–1.62) and tends to decrease with the number of carbons (Endo et al. 2008b). However, the difference in interaction enthalpy may be greater than it would appear by the ratio Kn/Kc, since the entropy penalty for transfer to a surface or phase is greater for the n-alkane because it has more degrees of freedom in the gas phase than does the cycloalkane. The contact area hypothesis has also been invoked to explain sorption trends for BC materials and sediments with high levels of BC. For example, in sediment systems polychlorinated biphenyl (PCB) congeners in which coplanarity of the rings is impaired due to ortho chlorine substitution have a lower sorption coefficient (Barring et al. 2002; Bucheli

and Gustafsson 2001; Jonker and Smedes 2000; Jonker et al. 2004; van Noort et al. 2002), greater “fast-desorbing” fraction (van Noort et al. 2002), and higher bioavailability (Jonker et al. 2004) than do coplanar congeners with the same number of chlorines. However, noncoplanar congeners have an inherently weaker tendency to interact with themselves, as witnessed by their higher subcooled vapor pressures (Schwarzenbach et al. 2002). They also partition less favorably into octanol from water (Schwarzenbach et al. 2002) for reasons that are not clear, since close approach applies to both solvents. The coplanarity effect for PCB adsorption to soot and soot-like materials from water held after normalizing the sorption coefficient by Kow (Jonker and Koelmans 2002b). Since BC materials are highly porous, an alternative explanation for the coplanarity effect is size exclusion (molecular sieving) in pores; a noncoplanar congener has a larger critical diameter and therefore would be more restricted in the passages that it could enter. It is well known that diffusion in zeolites becomes severely hindered as the minimum critical molecular diameter approaches the pore diameter (K€arger and Ruthven 1992). Anthracene and phenanthrene are both planar and have nearly identical KOW values (Schwarzenbach et al. 2002), yet anthracene consistently sorbs more strongly than does phenanthrene to BC materials (Jonker and Koelmans 2002b) and BC present in sediments (Cornelissen et al. 2004). Size exclusion may also explain the decline in Langmuir adsorption capacity of PAHs on activated carbon with increasing size (Walters and Luthy 1984), although no mechanism was offered in that study. Adsorption of n-hexane in some activated carbons is much greater than is cyclohexane, said to reflect size exclusion (Endo et al. 2008b). Systematic evidence has now been presented for size exclusion in charcoal BC for a series of planar aromatic compounds, both polar and apolar (Pignatello et al. 2006a; Zhu and Pignatello 2005a). For this sample of BC about 80% of the porosity exists in pores up to 2 nm wide. To normalize for hydrophobic effects, nonporous graphite was used as the reference state; BC–graphite isotherms were constructed

16


Mean Kch-gr, m2/m2

10 TOL XYL 124 TMB 1235 TeMB 1245 TeMB 12 DCB 124 TCB BNTL MNT DNT TNT BEN

1

0.1 0

1 2 3 Number of substituents

4

10 Mean Kch-gr, m2/m2

NAPH BEN PHEN

1

0.1

1

2 3 Number of fused rings

Figure 1.5. Char-graphite distribution coefficients normalized by surface area of aromatic compounds as a function of molecular size. Each point represents the mean and standard deviation of 11-18 data measured over a range of concentrations. [Data from (Zhu and Pignatello, 2005).] TOL ¼ toluene; XYL ¼ 1,4-dimethylbenzene; 124 TMB ¼ 1,2,4-trimethylbenzene; 124 TCB ¼ 1,2,4-trichlorobenzene; 1235 TeMB ¼ 1,2,3,5-tetramethylbenzene; 1245 TeMB ¼ 1,2,4,5-tetramethylbenzene; 12 DCB ¼ 1,2-dichlorobenzene; BNTL ¼ benzonitrile; MNT ¼ 4-nitrotoluene; DNT ¼ 2,4-dinitrotoluene; TNT ¼ 2,4,6-trinitrotoluene; BEN, benzene; NAPH ¼ naphthalene; PHEN ¼ phenanthrene.

from the experimental BC–water and graphite–water isotherms: xw > xBC xgr > xw xgr > xBC

KBC w 1=Kgr w KBCw qBC KBCgr ¼ ¼ Kgrw qgr

ð1:9Þ

In this reaction, K is the distribution concentration ratio between the sorbed and the solution phases, and q is the sorbed concentration normalized by surface area. The BC–graphite isotherms are shown in Figure 1.5. Benzene and monosubstituted benzenes sorb somewhat more strongly on BC than on graphite, which could be true or merely an artifact

of the technique for measuring surface area (Braida et al. 2003). Nevertheless, in Figure 1.5a it can be seen, that the char-graphite distribution ratio decreases—that is sorption becomes weaker relative to graphite—increasing number of ring substituents, regardless of the compound’s polarity. The effect is considerable; for example, adsorption of tetramethylbenzene to BC is about an order of magnitude weaker than benzene at constant sorbed concentration on graphite. Likewise, the char-graphite distribution ratio decreases with increasing fused ring size (Fig. 1.5b), revealing a one-order-of-magnitude difference in benzene and phenanthrene affinities for BC at constant concentration on graphite. Since these compounds are all planar (although some of the nitro groups may orient in a nonplanar relationship with the benzene ring), these results conclusively show that a size exclusion effect is operative that restricts the internal pore network surface area available for adsorption as molecular size increases. Support for steric effects is also evident in the work of Nguyen et al. (2007), who found that in two chars the maximum sorption capacity increased in the order of decreasing molecular diameter among planar compounds: phenanthrene < naphthalene < 1,2-dichlorobenzene/1,2,4trichlorobenzene < 1,4-DCB (Nguyen et al., 2007). 1.2.1.5. Sorption of Ionic and Ionizable Compounds. Because of the deprotonation of carboxyl and phenoxyl groups NOM contains an abundance of charged sites where ion exchange of organocations with native cations (M þ ) may occur. It is often difficult to quantify the contribution of NOM to cation exhange in whole soils because cation exchange sites are abundant on mineral surfaces as well. Organic anions may also show a tendency to sorb, particularly if the nonionic parts of the molecule are large; the classic examples are tetra- and pentachlorophenoxide (Schellenberg et al. 1984), whose sorption may occur as an ion pair with an inorganic cation and is weaker than sorption of the neutral molecule. Perfluoroalkanoic acids and perfluoroakane sulfonic acids (Ahrens et al. 2009; Higgins and Luthy 2006) also undoubtedly sorb as the ion pair, since their pKa values are well below 1. For further information on ion sorption, refer to Schwarzenbach et al. (2002). Sorption, of ionic or ionizable molecules containing multiple functional groups presents a much more complex situation. This will be illustrated with two antibiotic compounds whose sorption to NOM has been studied in some detail—sulfamethazine and tetracycline and their analogs. Sulfamethazine (Fig. 1.6) may exist in water as a cation, neutral molecule, or anion. Therefore the possible modes of interaction depend strongly on pH. (The zwitterion is a small fraction of the net-zero charged molecule.) NOM and, to a lesser extent clay minerals, play a role in its sorption, whereas sesquioxides have not been studied. The mean log KOC of

17


1.0 pKa3 = 9.7

8

O

N

9

7

6 6a

5 5a

4 OH 3

10a 10

11a 11

12a 12

1

OH

O

OH

O

OH NH2

2

S NH pKa1 = 2.46

O

Fraction

N

HO H pKa2 = 7.45

TCH22

0.8 0.6

TCH-

TCH2+

TC2-

0.4 0.2

N

O 0.0

H2N Sulfamethazine

pKa1 = 3.3

pKa2 = 7.7

Tetracycline

2

8

4

5

6

7

8

9

10 11

pH

Figure 1.6. Structures of the antibiotics and speciation diagram for tetracycline in solution in the absence of complexing metal ions (Gu et al. 2007). Metal ion coordination sites on tetracycline are as follows. Depending on pH and stoichiometry, Ca2 þ coordinates through the N4/O12a or O12/O1 atomic pairs, and Mg2 þ through the N4/O3 or O10/O12 atomic pairs. As the pH is raised from 2, Cu2 þ successively coordinates to O3, then to O10/O12, and finally to N4/OH-12a.

sulfamethazine in five soils is 2.11 (Carda-Broch and Berthod 2004). The KOC of a structural analog, sulfathiazole, in compost, manure, and solid humic acid follows the order KOCcation > KOCneutral > KOCanion (Kahle and Stamm 2007) and is hardly affected by K þ or Ca2 þ , suggesting that the interaction of sulfathiazole with NOM occurs by ordinary weak forces and the hydrophobic effect. This is consistent with the predominance of the neutral molecule over the zwitterion in solution. Sorption of the cation is enhanced relative to the neutral form due to electrostatic interactions with negatively charged sites on NOM, whereas sorption of the anion is suppressed for the same reason. Sorption of sulfamethazine to a charcoal BC followed the order KOCneutral > KOCcation > KOCanion (Teixido´ Planes and Pignatello unpublished). The KOCneutral was almost six orders of magnitude greater than the Kow of 0.27 for neutral sulfamethazine (Carda-Broch and Berthod 2004), a result that underscores the high affinity that polar compounds can have for BC. Charcoals have some cation exchange capacity owing to carboxyl groups (Chan and Xu 2009) that increases with weathering. However, sorption of the sulfamethazine cation was unaffected by NH4 þ , indicating that cation exchange was unimportant. Sorption of tetracyclines is even more complex than the sulfonamides because of the greater number of exchangeable protons and the tendency to coordinate with bound and free metal ions (Fig. 1.6). Sorption of the cationic form of tetracyclines may dominate sorption in most soils over the pH range 4–8, according to a more recent model (Sassman and Lee 2005). Sorption of tetracycline to dissolved humic acid is maximal for the zwitterion (Gu et al. 2007), which predominates between pH 4 and 7. The strong competition by NaCl indicates that the zwitterion sorbs primarily by ion exchange at carboxylate sites, with little contribution from hydrophobic effects. Sorption of the cation is weaker than sorption of the zwitterion, due to competition with H þ at lower pH values. Sorption of the anion is weaker still as a

result of charge repulsion and unfolding of humic macromolecules at higher pH values. In the presence of metal ions Al and Fe, metal ion bridging between the antibiotic and NOM (Tet–Mn þ –NOM) becomes important (MacKay and Canterbury 2005), and may be responsible for the strong binding of oxytetracycline to manure via the 5-OH (Loke et al. 2002). 1.2.2. Quantification of Driving Force Contributions A critical tool for separating the contributions of individual forces in sorption to natural solids is the free-energy relationship (FER). One obvious FER for testing hydrophobic effects is water solubility log KOC ¼ a log Sw þ b

ð1:10Þ

where a and b are fitting parameters. One problem with the solubility FER is that the reference state implied by the FER—the pure liquid or subcooled liquid—is different for each compound; hence, the relationship only works well for, (1) apolar compounds whose activity coefficient in each other’s liquid is typically not much different than in itself (unity) or (2) a series of compounds with a systematic change in the size of the nonpolar portion of the molecule (Allen-King et al. 2002). Other problems include the limited number of accurate solubility values and the necessity of converting the solubility of solids to that of the sub-cooled state, a conversion that rests on an estimation of the free energy of fusion (Schwarzenbach et al. 2002). A more popular predictive model for contaminant sorption is the FER with n-octanol–water partitioning, which takes the form log KOC ¼ c log Kow þ d

ð1:11Þ

where Kow is the octanol–water partition coefficient (L/L) and c and d are fitting parameters. Because of its importance in biology, the Kow is known for many thousands of

18


compounds, is relatively easy to measure, or can be calculated with available algorithms to a fairly high degree of accuracy (SRC). The KOC–Kow correlation has mechanistic significance in the sense that the free energy of octanol–water partitioning reflects the hydrophobic effect and to some degree polar forces (H bonding, dipolar), owing to the --OH group and the presence of a large amount of water in the octanol (21 mol% at equilibrium) in the test. Nevertheless, octanol partitioning cannot effectively represent all the driving forces for sorption of all types of compounds; indeed, while KOC–Kow correlation is reasonably good within a series of structurally related compounds, it deteriorates when compounds of different polarities are included in the dataset. [For tabulations of KOC–Kow FERs, see Allen-King et al. (2002) and Schwarzenbach et al. (2002).] In recent years researchers have tested polyparameter FERs (ppFERs) for various environmental partitioning phenomena, including sorption to NOM (Nguyen et al. 2005; Poole and Poole 1999; Endo et al. 2009b; Niederer et al. 2006b) and to activated carbon (Shih and Gschwend 2009) (see also Chapter 5 in this book for a fuller review). The ppFER takes the following form log KOC ¼ eE þ sS þ aA þ bB þ vV þ c

ð1:12Þ

where the uppercase letters are molecular descriptors and the lowercase letters are regression coefficients whose values reflect the net difference of interactions between the sorbed and dissolved states. The E descriptor combines dispersion and induction forces, S is the dipolarity/polarizability (D/P) descriptor encompassing dipolar and induction forces, A represents combined H-bond acidity (i.e; as A--H) and electron acceptor ability, B represents combined H-bond basicity and electron donor ability, and V represents the energy required to form a cavity to accommodate the molecule in the respective phase. The descriptors were calculated from physical constants or derived from extensive chromatographic datasets in the literature. While its main purpose has been predictive—the fiveparameter Equation (1.12) generally gives a better fit to KOC data than does the single-parameter OC–octanol FER—it is possible to draw some limited conclusions about the contributions of intermolecular forces in sorption to NOM through the application of Equation (1.12). The greatest influence on log KOC among apolar and weakly polar compounds appears to be the cavitation energy term V, followed by E. The contribution of the S term is variously reported to be positive (Endo et al. 2009b) or negative (Nguyen et al. 2005; Niederer et al. 2006b). Among the polar compounds, the H-bond donor/electron acceptor ability (A) term contributes little or nothing to log KOC, while the H-bond acceptor/electron donor ability term (B) contributes substantially, but negatively—that is, favoring water. For example (Nguyen et al. 2005), B contributed negatively by up to 28%

of log KOC, and especially for the ureas, the benzamides, and benzyl alcohol (Nguyen et al. 2005). At least the first two are strong H-bonders and this property favors water not NOM. Equation (1.12) applied to adsorption of a set of 14 diverse compounds from water to granular activated carbon (Shih and Gschwend 2009) showed adsorption depended positively on the sorbate’s V term, negatively on its B term, and weakly on its S term (the sign depended on concentration). A major obstacle to interpreting the ppFER in Equation (1.12) is the appreciable overlap of fundamental forces among the descriptors. Dispersion appears in both E and V descriptors. Dipolar and induction appear in S, A, B, and—to the extent that H-bonding and dipolar interactions play a role in water organization in the hydration shell—V. Another obstacle is the unclear meaning of A and B, which seems to have morphed from a pure H-bond acidity–basicity meaning to the dual H-bond and electron donor/acceptor meaning. Moreover, the contribution of those terms is sometimes incongruous with structural theory; for example, some PCBs and chlorinated alkanes and alkenes contribute several percent to log KOC via the B term (Nguyen et al. 2005), yet have negligible ability to accept a H bond, and are clearly not noted for their electron donor ability. Zhu and Pignatello (2005b) applied a multiparameter FER to sorption of apolar and polar compounds to polyethylene and three high organic soils in which NOM was presumed to be the dominant sorbent. Following the lead of others (Borisover and Graber 2003; Kleineidam et al. 1999a), they found it advantageous to switch the thermodynamic reference state from the pure liquid state to the dilute solution in n-hexadecane, as discussed above. The free energy of sorption to OC was partitioned into terms representing the hydrophobic effect (hyd), dipolarity/polarizability (D/P), H-bonding, and p–p EDA interactions, all in excess of the reference state (notated by the superscript e). The FER is written as: RTln KOC ¼ DGOC ðhydÞ þ GeOC ðD=PÞ þ GeOC ðHbondÞ þ GeOC ðppEDAÞ þ c

ð1:13Þ

where c is a constant. The DGOC(hyd) was taken to be a linear function of the free energy of hexadecane–water partitioning, so that, ln KOC ðhydÞ ¼ aln KHD þ b. The GeOC ðD=PÞ was taken to be proportional to the D/P descriptor S used in Equation (1.9). The isotherms, which were nonlinear, were constructed over a wide concentration range, which established the concentration- dependence of individual freeenergy contributions. Not surprisingly, sorption of all compounds to polyethylene was linear and could be modeled by considering hydrophobic effects only. For sorption of apolar compounds to the three natural sorbents, hydrophobic effects still predominated, but D/P effects contributed from 15%–40% of

SORPTION FROM THE PERSPECTIVE OF THE SORBENT

Amherst soil

Beulah Zap lignite

Pahokee

1.0

1.0

}

0.8

DCCH Fractional contribution

Fractional contribution

BEN HYD

0.6

0.4

}

0.2

0.0 10 -4

10 -3

10 -2

10 -1

10 0

D/P

101

}

0.8

0.4

}

0.2

0.0 10 -5

10 2

10 -4

10 -3

10 -2

10 -1

D/P

10 0

10 1

Solute Concentration, mmol/L 1.0

1.0

MCB vs DCB vs TCB in BZL

}

0.8

Fractional contribution

BEN, NAPH, PHEN in AS and BZL Fractional contribution

HYD

0.6

Solute Concentration, mmol/L

HYD

0.6

0.4

0.2

19

}

D/P + EDA for PHEN in AS

D/P

0.8

}

12 DCB MCB

0.6

HYD

14 DCB 124 TCB

0.4

135 TCB

}

0.2

D/P

EDA for PHEN in AS

0.0

10 -5

0.0

10 -4

10 -3

10 -2

10 -1

10 0

10 1


10 -5

10 -4

10 -3

10 -2

10 -1

10 0

10 1


Figure 1.7. Contributions of hydrophobic effects (HYD), dipolar/polarizability (D/P), and p–p EDA to the sorption free energy for the natural sorbents in selected systems. Compounds include benzene, naphthalene, phenanthrene, trans-1,2-dichlorocyclohexane, chlorobenzene, 1,2- and 1,4-dichlorobenzenes, and 1,2,4- and 1,3,5-trichlorobenzenes.

log KOC. Some trends are summarized below, with selected examples given in Figure 1.7: .

.

.

. .

In most cases, with increasing sorbed concentration the percent contribution of the hydrophobic term (% hyd) increases, while % D/P decreases (Fig. 1.7). This indicates a transition to loading of sites that are poorer in D/ P influence—possibly less polar and less aromatic. These results are consistent with those of Endo et al. (2008a). The % D/P increases with dipole moment (CCl4 < CHCl3 < CH2Cl2) and polarizability (cyclohexane < benzene < naphthalene < phenanthrene) (Fig. 1.7). Both % hyd and % D/P for mono through trichlorinated benzenes were insensitive to the number and position of Cl atoms (Fig. 1.7). The % D/P was greater for 2,4-dichlorophenol than for the less polar, less polarizable n-nonanol. The absolute contribution of H-bonding was roughly equal for n-nonanol and 2,4-dichlorophenol, even through the latter was a better H-bond donor toward

carboxylate and phenolate groups on NOM (Table 1.1.) However, it is not possible to tell whether H bonding of these compounds with NOM is more favorable than with water, since the free energy of H bonding with water is embedded in the hydrophobic term.

1.3. SORPTION FROM THE PERSPECTIVE OF THE SORBENT 1.3.1. Shape of the Isotherm When constructed in sufficient detail and over sufficiently wide range in solute concentration, sorption isotherms on natural solids, as well as NOM and BC reference materials, are often found to be nonlinear in solute concentration. Since the activity coefficient in water cw [see Eq (1.2)] of solutes is typically independent of concentration under dilute conditions, nonlinearity implies a change in the activity coefficient in the sorbed state cs. That, in turn, can mean that either sorption sites are heterogeneous in energy

20


or the nature of the solid changes with the degree of loading. Both are possible. Heterogeneity can arise from a distribution in pore widths where condensation occurs. Heterogeneity may even be a property of pure nonporous materials due to surface irregularities, defects, and edge effects. Partitioning into a flexible solid can result in swelling that feeds back on the solid’s sorbent properties. In addition, solid phases may not be internally homogeneous with respect to sorption potential. How sorption varies with solute (or vapor-phase) concentration and the concentration of other solutes has important implications for modeling or predicting the behavior and bioavailability of chemicals, as well as for understanding the sorption process mechanistically. The variation in sorption distribution coefficient ranging in aqueous concentration from infinite dilution to water solubility can be as much as three orders of magnitude depending on the degree of nonlinearity in the isotherm. Since, in most cases, bioavailability depends on the concentration in the fluid phase, not the sorbed phase, nonlinearity can have a profound influence on bioavailabilty (Pignatello 2009). Nearly all isotherms models are based on the Langmuir, Freundlich, or Polanyi–Manes equations or variations thereof: 1. The Langmuir equation assumes a limited number of sites of a single energy (Adamson and Gast 1997) and is given by QL KL C q¼ ð1:14Þ 1 þ KL C where q is the equilibrium sorbed concentration, C the equilibrium solute concentration, and QL and KL are, respectively, the Langmuir maximum capacity and affinity parameters. 2. The Freundlich isotherm may be derived from the Langmuir equation by assuming a distribution of site energies (Adamson and Gast 1997). It is given by q ¼ KF C n

ð1:15Þ

where KF is the affinity coefficient and n the exponent reflecting the degree of linearity. The linear isotherm is a special case of the Freundlich isotherm where n ¼ 1. 3. The Polanyi–Manes isotherm is based on the concept of condensation in small pores and is given by q ¼ QP 10½aðe=N Þ b

ð1:16Þ

where e ¼ RT lnðCmax =C Þ and a and b are fitting parameters; QP is maximum capacity; N is a normalizing factor, often taken to be the adsorbate molar volume; e is adsorption potential corrected for adsorption potential of an equal volume of water displaced; Cmax is maximum water solubility of the pure liquid/ subcooled liquid; R is the thermodynamic gas constant;

and T is temperature [see Xia and Ball (2000) and references cited therein]. For vapor–solid isotherms, Cmax/C is replaced by p0/p, where p is partial pressure and p0 is the pure liquid or sub-cooled liquid vapor pressure. Developed originally for AC adsorbents, the Polanyi– Manes model assumes chemical condensation in a liquidlike state in the pore at a pressure governed by capillary forces, beginning in pores of the smallest diameter; water, if present, is displaced (Allen-King et al. 2002; Manes 1998; Xia and Ball 1999). Sorption is nonlinear because of the wide range of pore widths in the solid. In some cases sorption of compounds whose thermodynamic state is solid at the experimental temperature has been observed to be weaker than predicted on the basis of the Polanyi characteristic curve for liquids of similar structure (e.g., benzene vs. PAHs). To explain this, the solid solute is postulated to condense in a crystalline state, filling the pore less efficiently than would a liquid (Xia and Ball 2000). At this time there is no independent evidence for crystallization, and there are examples where the phenomenon of weaker sorption of the solid solute is not observed [e.g., for example, in the chlorinated benzene series (Nguyen et al. 2007; Xia and Pignatello 2001)]. While crystallization seems plausible in mesopores, which are filled at higher concentrations, the concept of crystallization in a micropore is not a compelling one. As we have seen, size exclusion in pores may limit the available adsorption space of the larger (solid) molecules and rationalize their observed weaker sorption. Cation exchange of simple inorganic ions on soils follows a Langmuir-like isotherm, leveling off at the maximum cation exchange capacity of the sample (CEC, measured in equivalents per unit mass). However, for organocations, one often finds continued sorption beyond the CEC related to molecular size. Since charge cannot accumulate on the solid, this additional sorption must be due to sorption of the ion pair. The following three equations (1.17 to 1.19) illustrate the simple monovalent counterion case: x þ þ M þS > x þ S þ M þ x þ þ y > ðx þ y ÞS qtot ¼ qxS þ qðxyÞS ¼

Kex KIP

ð1:17Þ ð1:18Þ

CEC Kex ½x þ þ KIP ½y ½x þ ½M þ þ Kex ½x þ ð1:19Þ

In these equations, y is a counterion forming an ion pair with x þ , q refers to solid-phase concentrations, brackets refer to aqueous-phase concentrations, Kex is the ion exchange equilibrium constant; and KIP is the ion pair sorption equilibrium


constant. The value of KIP itself may be concentration-dependent, depending on the tendency of the organocation to form ordered layers (hemimicelles, admicelles) on the surface, as is the case for organocation surfactants, such as tetraalkylammonium salts. In theory as well as experimentally, sorption always tends toward linearity as solute concentration approaches infinite dilution. Note that the Freundlich and the Polanyi models are unrealistic at very low concentrations because they fail to linearize. Assigning mechanism predominantly on the basis of data fit to a particular model is not advised; for one thing, the above three models do not have the same number of fitting parameters. 1.3.2. Sorbent Properties of Natural Organic Matter Solid-phase dissolution (“partitioning”) is the historic paradigm for sorption to NOM. The basis of this paradigm is the conceptualization of NOM as a hydrated, “loosely-knit” gel phase capable of “dissolving” organic solutes and permitting them move about freely in the milieu, as if it were a liquid (Chiou 1989, 2002; Chiou et al. 1998). Since liquid–liquid partitioning is generally linear under dilute conditions (i.e., the activity coefficient in each liquid is concentration-independent), the frequent finding of sorption “linearity” has reinforced the solid-phase dissolution concept. However, in many of these studies, isotherms were constructed over a limited range in solute concentration, where curvature, if any, would have been difficult to discern, or where number and/or quality of data were inadequate to statistically distinguish linearity from nonlinearity. In many studies linearity had been merely assumed for the sake of simplicity in modeling, especially when sorption was not the main focus. More detailed studies show that sorption to soils when NOM is likely to be the predominant sorbent material (i.e., high OC content, nonpolar solute, abundant water) is usually nonlinear. The nonlinearity of sorption in Pahokee peat, a high-organic reference soil from the International Humic Substances Society, is such that the distribution coefficient can range over two orders of magnitude from very low to concentrations approaching water solubility (Endo et al. 2008a; Xia and Pignatello 2001). In addition, sorption is nonlinear and competitive even for nonpolar solutes in BC-free humic acid isolates (e.g., Huang et al. 1997; Pignatello et al. 2006b; Xing and Pignatello 1997; Zhao et al. 2001). Finally, isotherms constructed using a variety of experimental techniques of PAHs—mere hydrocarbons—in DNOM, arguably the simplest and most homogeneous form of NOM, are convincingly nonlinear and competitive (Laor and Rebhun 2002; Pan et al. 2007; Polubesova et al. 2007). Recognizing that the solid-phase dissolution paradigm is useful, but likely a simplification, a new paradigm for sorption to NOM has emerged—one in which SOM exists as a heterogeneous mixture of physical states that provides a

21

hierarchy of sites, each limited in capacity. Several postulates have been offered to explain this heterogeneity. They are based on (1) domain-based preferential sorption, (2) functional-group-based preferential sorption, and (3) the physical state in which NOM exists. A clear choice among these is hampered by a lack of molecule-scale evidence. 1.3.2.1. Domain-Based Preferential Sorption. This hypothesis holds that strands of NOM associate together on the basis of functional group identity to form domains large enough in scale to act independently as micropartition phases (a “carbohydrate-like domain,” an “aromatic domain,” etc.). Domain selectivity can clearly depend on domain chemistry—witness the wide range in solvent–water partition coefficient for a given compound in different solvents (Schwarzenbach et al. 2002). Partition domains in NOM would have to be limited in maximum capacity, however; otherwise, sorption would merely be a linear combination of linear terms—still linear overall. Evidence for functional unit homogeneity on a scale large enough to serve as separate partition domains is mixed. Several research groups have identified crystalline and amorphous polymethylene domains in humic acids (Gunasekara et al. 2003; Hu et al. 2000; Mao et al. 2002; Mao and SchmidtRohr 2006; Schaumann et al. 2005) and whole soils (Lattao et al. 2008) by nuclear magnetic resonance (NMR) spectroscopy. These domains originate from the remnants of plant cuticular waxes, cutan, cutin, and subarin, which are long preserved in NOM and often make up percent levels of NOM (Chen and Xing 2005). It is reasonable to postulate sorption in the amorphous polymethylene domains. Sorption in the crystalline domain, by definition, is not possible. On the other hand, carbohydrate-like and lignin-like moieties in humic acid (Mao and Schmidt-Rohr 2006) and whole soils (Lattao et al. 2008) appear to be intimately mixed and therefore not likely to form independent domains. Most studies testing domain-based preferential sorption rely on correlations between sorption and bulk functional group composition of NOM determined by solid-state NMR techniques. Overall, however, the results are conflicting or inconclusive in many respects: on whether sorbates prefer aromatic or aliphatic domains (Chefetz et al. 2000; Kile et al. 1999; Mao et al. 2002; Ran et al. 2007), or have no preference (Wen et al. 2007); on whether desorption rate correlates or not with aromaticity (Cornelissen et al. 2000; Lucht and Peppas 1987); and on whether sorption intensity correlates with an index of NOM polarity (Chen et al. 2005; Kile et al. 1999). Cook, in Chapter 13 of this book, discusses many of the technical factors that limit the success of the NMR approach. There are also a number of conceptual limitations: (1) a disregard for the role of local “physical organization” of NOM, (2) the importance of excess free volume (porosity) (see below), (3) the often-arbitrary defini-

22


tion of functional units (e.g., “aromatic” may be hydrophobic or hydrophilic depending on substituents), (4) a neglect of functional unit mixing, and (5) the possible influence of hydration on partitioning in a given domain. In a review of the literature (Chefetz and Xing 2009), no significant correlations existed between either aromaticity or aliphaticity of NOM and sorption affinity for hydrophobic compounds. However, the authors emphasize that hydrophobic compounds do partition readily into alkane phases and therefore the polymethylene domains are probably important sorbents. 1.3.2.2. Functional-Group-Based Preferential Sorption. According to this hypothesis, molecules may interact with individual functional groups capable of specific or directional interactions. Nonideal behavior would result if the number of sites of each discrete type were limited. The case that first comes to mind is organocation exchange at negatively charged sites, which are inherenetly limited in number. So, too, compounds capable of specific weak interactions such as H-bonding or p–p EDA interactions with NOM may encounter a matrix with a limited number of complementary sites, and therefore could potentially show nonlinearity and competition. The relatively few available data available lead tentatively to the conclusion that nonlinearity seems not to require the existence of functional groups in the molecule that can undergo specific interactions; rather, nonlinearity seems to be due to a change in the nature of sites in NOM with loading. In their study of many polar and nonpolar compounds on Pahokee peat (45% OC), Borisover and Graber (2003) found no basis for predicting nonlinearity by the presence of strongly-interacting functional groups. Using Pahokee peat, a second high-organic soil, and a lignite, Zhu and Pignatello (2005b) observed concentration-dependent contribution of driving forces to sorption. However, the concentration dependence was relatively insensitive to solute structure. They found a general decrease in the contribution of the “dipolarity/polarizability” (D/P) term to log KOC with loading, and attributed it to preferential affinity for sites that are rich in functional groups with D/P influence, such as polar or aromatic groups. Endo et al. (2008a) came to similar conclusions in their study of polar and apolar compounds on a lignite and Pahokee peat. In both sorbents, H bonding played no role in nonlinearity. For lignite, the cause of nonlinear sorption was fairly compound-independent. For Pahokee peat, n was inversely linearly dependent on the S descriptor of Equation (1.12); they attributed this to a transition with increasing sorbate loading to sites of less aromatic and/or less polar character. 1.3.2.3. Heterogeneity Due to the Physical State of the Matrix—the Glassy Polymer Concept. In principle, macromolecular organic solids can exist in four interconvertible physical states—melted, rubbery, glassy, and crystalline (Eisenberg 1993). Matrix structure and dynamics play an

important role in the sorbent properties of synthetic polymers, and researchers, mainly of the author’s group and the group of Weber and coworkers, have postulated that this is true for NOM, as well. The polymer analogy for NOM has been made by many investigators. The most relevant states of solid NOM are the amorphous rubbery and glassy states. A critical property of amorphous polymers is the glass-torubber transition temperature (Tg), which correlates with the rigidity of the matrix. The rubbery state (T > Tg) has relatively high segmental flexibility of the macromolecules, allowing equilibrium to be achieved quickly following changes in an applied external condition such as temperature. The glassy state is the opposite. As a rubbery solid cools through the Tg to a temperature below the Tg, macromolecular rearrangements necessary for reaching the thermodynamic state require more time than is available by the imposed cooling rate, ultimately becoming prohibitive. As a result, excess free volume is frozen into the matrix in the form of poorly interconnected micropores or “holes.” The glassy state is, thus, a perpetual nonequilibrium state. Hong et al. (1996) confirm the existence of holes in glassy bisphenol A polycarbonate using positron annihilation lifetime spectroscopy (PALS) during the sorption of CO2. In the PALS technique positrons (b þ ) emitted from 22 Na interact with matter to form positronium atoms (Ps, electron þ positron), whose triplet state lifetime correlates with hole size, and whose intensity is a measure of free volume. Growing evidence indicates that NOM solids are—at least in part—glassy, a property that significantly affects their behavior as sorbents. Thermal transitions in heat capacity or coefficient of expansion that are characteristic of glassy synthetic and biopolymers have been observed for humic and fulvic acids, whole soils, shales, and coals, as well as for lignin, a major precursor of humic substances (DeLapp and Leboeuf 2004; LeBoeuf and Weber 1997; Schaumann and Antelmann 2000; Schaumann and LeBoeuf 2005; Zhang et al. 2007;) Lucht et al. 1987; Yun and Suuberg 1993). Based on solid-state wide-line 1 H NMR spectroscopy of dry humic and fulvic acids, a soil NOM, and a whole soil, Mao and Schmidt-Rohr (2006) found that segments undergoing fast, large-amplitude motions are a minor component of the sample, consistent with the postulated glassy character of NOM (Xing and Pignatello 1997). Glassy polymer theory has long been used to describe sorption of chemicals by coal (Milewska-Duda 1993). One important property of polymers that bears on the behavior of NOM is that the glass-to-rubber conversion takes place, not only with increasing temperature but also with increasing concentration of a penetrant. Technically known as plasticization, this is due to the softening effect resulting from progressive replacement of polymer–polymer interactions with penetrant–polymer interactions. Plasticization and its influence on sorption properties of polymers is well documented by the close correspondence between sorption and dilation isotherms (Fleming and Koros 1990). Evidence


Sorbed concentration

rubbery state

glassy state

“plasticization” to the rubbery state at the Sg

Pressure or dissolved concentration

Figure 1.8. Hypothetical sorption isotherms of a penetrant in identical glassy (solid-line curve) and rubbery (dashed-line curve) polymers. The Sg indicates the glass-to-rubber transition concentration. The dashed line below the Sg would correspond also to the thermodynamic state of the glassy polymer.

for swelling of humic substances during sorption is based on sedimentation volumetric changes in cosolvent–water mixtures (Lyon 1995) and “conditioning effects” (see Section 1.3.5). Pyridine—a known swelling agent of coals—promotes rapid macromolecular motions and expansion of void volume as shown by 1 H NMR studies of Argonne premium coals (Xiong and Maciel 2002). Sorption of PAHs by plant cuticular waxes coated on montmorillonite causes transitions from a less flexible to a more flexible amorphous state (Chen and Xing 2005). Figure 1.8 schematically presents the sorption isotherms for a penetrant in hypothetically identical polymers, except that one is rubbery (T > Tg; curved line) and the other is glassy (T < Tg; solid line) at the experimental temperature T. Sorption in the rubbery state obeys the Flory-Huggins equation (Chiou 2002) h i C w¼ exp ðwpoly þ xw2poly Þ ð1:20Þ Csat where w is the volume fraction of penetrant in the solid, wpoly is the volume fraction of polymer, and x is the–polymer-penetrant interaction parameter. Strictly speaking, linearity is not characteristic of the rubbery state; while appearing linear over most of the range, the Flory–Huggins curve becomes nonlinear, concave-up at sufficiently high concentrations dependent on the value of x. The concave-up trend occurs because (1) plasticization reduces the cavity penalty for partitioning and (2) loading makes the solid more “penetrant-like” (i.e., drives the penetrant activity coefficient in the solid toward that of the pure liquid penetrant). An isotherm of chlorobenzene in rubbery polyethylene showing the concave-up bending at high concentration has been published (Sander et al. 2006). For the glassy polymer, the isotherm initially takes on a nonlinear, concave-down shape at low concentration (Fig. 1.9). Fit to the Freundlich equation in this range would

23

give an exponent n less than 1. At sufficiently high concentration, the solid is converted to the rubbery state through plasticization, and the isotherm inflects at the point where it joins the rubbery isotherm—the glass transition concentration Sg in Figure 1.8—and from there on, the isotherm would follow the Flory–Huggins equation. Figure 1.9 shows the isotherms of compounds in polymers and in OC-rich soils that show the inflection point and the reverse-S shape. As a result of this characteristic shape, the Freundlich n varies depending on the experimental concentration range over which it is calculated—typically, starting out near unity at the lowest concentrations, reaching a minimum at an intermediate concentration, and then approaching unity again at high concentration (Xia and Pignatello 2001). The isotherm of trans-1,2-dichlorocyclohexane in Figure 1.10 shows the difficulty of observing the characteristic shape unless the isotherm is constructed with many data points over a wide range in concentration (Xia and Pignatello 2001):, when the number of data is reduced from 64 to 32 or to 16, the resulting isotherm could easily be accepted as being “linear.” The proposed cause of the concave-down curvature at low concentration is the presence of holes due to incomplete relaxation relative to the thermodynamic state. These holes act as adsorption sites toward solute molecules. (Note that adsorption is not exactly accurate here, as within such a pore there can be no real distinction between containing the wall and not containing the wall.) “Holes” are defined as semipermanent cavities of subnanometer size resulting from the incomplete relaxation of the solid to its thermodynamic (rubbery) state that may lie within folds of individual macromolecules, between macromolecules, or between the organic phase and a mineral surface. This is illustrated in Figure 1.11 and is applicable to macromolecular organic matter solids— ordinary humic substances and ancient organic matter, kerogen, and soft coals. The model often used to describe sorption to glassy polymers is the dual-mode model (DMM), which combines the linear Freundlich and Langmuir isotherms: q ¼ KD C þ

QH KH C 1 þ KH C

ð1:21Þ

Here, the subscripts D and H refer to dissolution domain and hole domain, respectively. When the solid is a composite of rubbery and glassy materials, KD represents the sum of the solid–water partition coefficient for the rubbery-phase and that for the dissolution domain of the glassy phase. A conceptually similar model has been introduced by Weber and co-workers called the dual-reactive-domain model (DRDM) (Huang et al. 1997; Johnson et al. 2001; Weber et al. 1992; Young and Weber 1995). The DRDM combines linear partitioning into a “soft” state with nonlinear partitioning into a “condensed” state of NOM. Sorption occurs preferentially in the holes because unlike the dissolution domain it requires little or no cavity penalty .

24


(a)

1,2-DCB sorbed to PVC (mg/kg)

(b) 2e+5 10000

q (µg/g)

2e+5

5000

0

1e+5

0

5

10

5e+4

0 0

200

400

600

800

6x10 4 5000

0 0

5

10

2x10 4

1000 1200 1400

0

C (µg/mL)

10

20

30

1,2-DCB in solution (mg/L)

(c)

(d)

15000

DCCH sorbed (mg/kg)

2500

4x10 4

dual-mode fit below nmin

dual-mode fit to data below nmin

linear fit above nmin, thru origin

linear fit to data above nmin, thru 0

150000

10000

100000

5000

50000

0

0 0

100

200

300

DCCH in solution (mg/L)

0

1000

2000

DCP in solution (mg/L)

Figure 1.9. Isotherms of organic compounds in polymers and NOM (Pahokee peat soil) showing reverse-S shape. (a) Nitrobenzene in Tenax [Tg (glass transition temperature) ¼ 227 C]; (b)1,2dichlorobenzene in polyvinylchloride (Tg ¼ 85 C); (c) trans-1,2-dichlorocyclohexane in NOM; (d) 2,4-dichlorophenol in NOM. The nmin is close to the inflection point. [Data for nitrobenzene from Zhao and Pignatello (2004); all others from (Xia and Pignatello 2001).

The free energy of cavitation for a molecule the size of 1,2,4-trichlorobenzene or naphthalene in the dissolution domain of soil organic matter has been estimated to be 15–20 kJ/mol (Lu and Pignatello 2004a). Since holes are finite in number, the preference for holes predicts concave-down nonlinearity and a competitive effect when multiple solutes are present, behavior that is quite commonly observed. Polymers in the glassy state that trap a large amount of interconnected free volume are said to behave like microporous materials in many respects (Budd et al. 2005). In whole soils the unrelaxed free volume should be considered the sum of unrelaxed free volume within the NOM solid phase and that between the mineral surface and NOM molecular strands (Fig. 1.11). Equation 1.21 represents a simplification of reality: it assumes two discrete types of domains, a single hole energy, and does not reflect plasticization. The extended dual-mode model (EDMM) of Kamiya (Kamiya et al. 1992, 1986, 1998;

Wang et al. 1998) based on sorption and dilation isotherms of gases and hydrocarbons in polymers includes terms for progressive plasticization. As penetrant is loaded, the extra hole free volume is gradually eliminated through plasticization and the solid changes toward an equilibrium (rubbery) configuration. The cavity penalty is inversely dependent on the sorbed concentration, consistent with a plasticization effect (Zhu and Pignatello 2005b). The Tg of macromolecular solids increases with molecular weight, chain branching, degree of unsaturation, ring content, and interstrand crosslinking—properties that introduce chain stiffness and/or strand interconnectedness to the solid, and therefore reduce its flexibility. Dissolved NOM (DNOM) studied in its natural state or isolated from waters (Leenheer 2009) tends to consist of smaller, more readily hydrated molecules than those that are immobilized on geosolids. As those properties favor flexibility, DNOM aggregates or colloids are expected to be on the rubbery end of


(a)

64 data

DCCH sorbed (mg/kg)

15000

dual-mode fit below nmin linear fit above nmin, thru origin

dissolution domain

10000

5000

25

unrelaxed free volume (“hole” domain)

NOM

0 0

100

200

300

mineral

DCCH in solution (mg/L) (b)

15000 DCCH sorbed (mg/kg)

Figure 1.11. Artist’s conception of holes in natural organic matter associated with a mineral surface.

32 data (every second one)

10000 Kp = 57.7 2

r = 0.9797

5000

320

0 0

100

200

300

DCCH in solution (mg/L)

DCCH sorbed (mg/kg)

(c)

16 data (every fourth one)

15000

10000 Kp = 60.6 r2 = 0.9847

5000

218 0 0

100 200 300 DCCH in solution (mg/L)

Figure 1.10. Effect of isotherm detail on appearance of linearity. (a) Isotherm of trans-1,2-dichlorocyclohexane, the same as in Figure 1.7, consisting of 64 data; (b,c) successive removal of data and fit to linear isotherm (number at bottom of graph is the intercept).

the scale. Likewise, humic and fulvic acids extracted from soil tend to be the smaller, more soluble molecules of SOM favored by the commonly used extraction techniques. Nevertheless, sorption isotherms of apolar compounds to solid soil humic acid particles (Lu and Pignatello 2004a,b; Pan et al. 2007) are nonlinear, show competition, and are hysteretic (Lu and Pignatello 2004b; Pignatello et al. 2006b; Sander et al. 2006)—all of which are consistent

with solid humic acid existing in the glassy state. Crosslinking of humic acid with trivalent metal ions increases the Tg slightly (Pignatello et al. 2006b) and increases sorption nonlinearity and the contribution of hole filling to total sorption (Lu and Pignatello 2004b). Figure 1.12 illustrates these effects for 1,2,4-trichlorobenzene sorption in H þ exchanged soil humic acid (H-HA), Al-exchanged HA (Al-HA), and a brown coal [Beulab–Zap lignite (BZL) (Lu and Pignatello 2004a). Figure 1.12 shows the isotherms [plotted as the sorption distribution ratio Kd(L/kg) vs.the reduced concentration] fit to the DMM [Eq (1.21)]; the ratio QHKH/KD, which corresponds to the ratio of concentrations in the hole domain to the dissolution domain in the limit of infinite dilution; and a plot of the fraction sorbed in the hole domain versus reduced concentration. Notice that fit to the DMM captures the inflection of the isotherm. The collective results of thermal and thermomechanical analysis of various samples (DeLapp and Leboeuf 2004; Schaumann and Antelmann 2000; Schaumann and LeBoeuf 2005; Zhang et al. 2007; Lucht et al. 1987; Yun and Suuberg 1993) indicates that Tg increases as a function of diagenetic or geothermal alteration (Zhang et al. 2007) in the following order (rough range in Tg): Fulvic acid particles < humic acid particles (30 C–70 C) 150 C) Soil organic matter (i.e., recently deposited material in nearsurface horizon soil samples) seems to fall between humic acid and types I and II kerogens (DeLapp and Leboeuf 2004; Schaumann and Antelmann 2000; Schaumann and LeBoeuf 2005). Sorption of apolar compounds in soft coal (BZL) is greater overall and has a larger hole-filling component than does sorption in soil humic acid (Fig. 1.12). It is known that moisture greatly suppresses adsorption of organic vapors to ordinary soils, since water competes strongly with organic compounds for mineral surfaces, which

26

INTERACTIONS OF ANTHROPOGENIC ORGANIC CHEMICALS 10 5

6

both donor and acceptor properties) give the opposite behavior, while monopolar compounds show nonuniform behavior (Niederer et al. 2006a). Using a concentrationspecific FER, Zhu and Pignatello (2005b) found that, with increasing loading from aqueous solution, organic molecules fill sites in NOM of progressively greater hydrophilic character. These observations may be explained as follows. Water initially has an anti-plasticizing effect on NOM (increase in Tg) due to crosslinking of polar functional groups, and then a plasticizing effect above 12% moisture (Schaumann and LeBoeuf 2005). Water competes for adsorption space of organic compounds within glassy matrices. Hence, the decrease in sorption with increasing moisture content may be due to the progressive transition from a more to a less glassy material, coupled with the competitive effect of water. In the case of bipolar compounds, the extra H bonding afforded by the presence of water molecules may overcome the competitive and plasticizing effects of water.

4

1.3.3. Sorbent Properties of Black Carbon

TCB

Kd, L/kg

10 4

10 3

10 2 10 -5

10 -4

10 -3

10 -2

10 -1

C/Sw 8

(QHKH)/KD

non-conditioned conditioned

2

0 H-HA

Al-HA

BZL

1.0 TCB

Fraction in Hole-domain

BZL 0.8 Al-HA 0.6 H-HA 0.4

0.2

0.0

-5

10

-4

10

-3

10

C/Sw

-2

10

-1

10

0

10

Figure 1.12. Sorption of 1,2,4-trichlorobenzene in H þ -exchanged soil humic acid (H-HA), the Al-exchanged HA (Al-HA), and a brown coal [Beulah–Zap lignite (BZL)] before and after “conditioning.” Fit is to the dual-mode model [see Eq. (1.21)]. Conditioning effect is discussed in Section 1.3.5.

are usually abundant in soils (Chiou 2002). Water, however, has a more modest effect on sorption to NOM, which can absorb as much as one-fourth its weight in water. Apolar compounds typically experience up to three fold suppression of sorption from the dry state to the fully humid (98% relative humidity) or water-saturated state (Niederer et al. 2006a; Rutherford and Chiou 1992). Bipolar compounds (those with

As mentioned, sorption to raw BC is highly nonlinear and competitive, more so than NOM. Nonlinearity can be so severe that the distribution coefficient can range over many orders of magnitude. For example, the Freundlich n of the neutral form of sulfamethazine is 0.27 and so the KBC varies by 3 orders of magnitude from 1 mg L1 (near the limit of quantification) to 30 mg L1 (about 6.7% of water solubility) (Teixido´ Planes and Pignatello unpublished). Nonlinearity is due to a combination of surface site heterogeneity and pore size heterogeneity. The distribution of adsorption potentials on a graphene surface can be pronounced even on what would appear to be a smooth, homogeneous surface; for example, the Freundlich n values for sorption of benzene and a number of other apolar aromatic compounds from water onto nonporous graphite (99.95% C; BET specific surface area, 4.5 m2/g) are below 0.60 (Zhu and Pignatello 2005a). Graphite has at least three different kinds of sites: basal plane, defect sites along the basal plane (pits), and edge sites along step elevations. At higher concentrations, adsorbate–adsorbate interactions are possible as is confirmed by atomic force microscopy for substituted benzoic acids (Martin 2003). A detailed isotherm of benzene sorption from water to a wood charcoal over the range in concentration from 1.8 107 times to 0.6 times water solubility is shown in Figure 1.13 (Braida et al. 2003). The isotherm is highly nonlinear, and, owing to its reverse-S shape, does not fit any of the conventional isotherm models in section 1.3.1. The log BC–water distribution ratio ranges from a maximum of 4 to a minimum of 2.7; these values are, respectively, about 200 and about 10 times greater than KOC for benzene calculated on the basis of an OC–octanol FER. Note that the Freundlich n for benzene approaches 1 at low concentrations but declines as the range over which it is calculated increases.


1e+6

q(μg/gsoot)

log K BC = 2.7

Experimental Freundlich Fit Langmuir Fit

1e+5

27

1e+4

1e+3

log KBC = 4

Freundlich n

1e+2

0.706 0.825

1e+1

0.759

0.903

1e+0 0.00010.001

0.01

0.1

1

10

100

1000

Benzene dissolve conc. (g/mL)

Figure 1.13. Sorption isotherm of benzene on charcoal powder suspended in water. The highest concentration is about 60% of water solubility. Horizontal arrows indicate the range over which the Freundlich n is calculated. KBC (L/kg) is the charcoal-water distribution ratio. [After Braida et al. (2003).]

Adsorption to BC is strongly affected by the polarity of the surface. Offgassing the polar C by catalytic reduction in a stream of H2 hardly affected the surface area, but enhanced adsorption on a surface area basis, regardless of the polar or nonpolar nature of the adsorbate (Zhu et al. 2005). An inverse linear relationship exists between the oxygen content of multiwalled carbon nanotubes and their maximum adsorption capacity for naphthalene (Cho et al. 2008). Black carbon undergoes oxidation as it weathers in soil. As proposed for AC (M€ uller and Gubbins 1998; M€ uller et al. 2000), polar groups on micrographitic sorbents apparently exert a “crowding out” effect resulting from the clustering of water molecules around polar sites that limits the available adsorption surface for the organic compound. Grand canonical Monte Carlo simulations of water competition with methane are consistent with this hypothesis (Fig. 1.14) and also show that competition diminishes as pore size increases (M€ uller et al. 2000). The surface activity of raw BC can be greatly attenuated by the presence of other substances, including, (1) incompletely carbonized biomass polymer fragments, (2) unburned liquid fuel or fuel byproducts, and (3) natural substances in the environment. For a given biomass source material, the specific surface area of the char varies with pyrolysis temperature, ramping steeply above 300 C (Fig. 1.15). Sorption trends with specific surface area, although this correlation is not very useful because it is complicated by pore size distribution and surface chemistry (Jonker and Koelmans 2002b), and there are conceptual issues with the customary use of N2 as at 77 K as the probe gas for surface

Figure 1.14. Equilibrium configurations for the adsorption of methane and water in a carbon slit pore of width 1 nm with varying active-site densities at 300 K (M€ uller et al. 2000). Methane, red balls; water, blue balls. Gray asterisks represent active sites on surface in foreground or background, depending on their size. (See insert for color representation of this figure.)

area (Kwon and Pignatello 2005; Pignatello et al. 2006a). The uncarbonized material is a weaker sorbent than the carbonized material. Semivolatile oil-like and tar-like materials can condense in the pores of soot particles during synthesis and/or cooling (Akhter et al. 1985b). Semivolatiles can also condense on charcoal if the offgases do not vent during charring. Aerosol BC particles can accumulate secondary photolysis products of semivolatile organic compounds in the atmosphere that subsequently affect sorption (Dachs and Eisenreich 2000; Kamens et al. 1995; Strommen and Kamens 1997). On weathering in the soil/sediment environment, the ability of BC particles to sorb organic compounds is attenuated by humic substances, metal ions, and possibly metal sesquioxides that may adsorb or coat their surfaces. Dissolved humic substances are well-known to foul activatedcarbon adsorbents used in the water treatment industry (Kilduff and Wigton 1999; Li et al. 2003; Newcombe et al. 1997). Field evidence supporting a weathering effect on BC include the finding of weaker sorption than expected on the basis of raw BC (up to nine fold) for PAHs and PCBs

28


N2B.E.T. Surface Area

(a)

100

400 qEC, mmol/kgEC

Specific surface area, m2/g

500

300 200 100

10 qEC,char = 103.72 Cw0.38 1 qEC,cofloc = 10 4.12 Cw0.66 0.1 char alone Al-HA-char co-floc

0

0.01 10-9 200

300

400 T, oC

500

600

700

Figure 1.15. Effect of pyrolysis temperature on the specific surface area of charred maple wood shavings. The SSA is based on the N2 probe gas adsorption isotherm at 77 K and calculated by the Brunauer–Emmett–Teller equation (Brunauer et al. 1938).

in sediment believed to be rich in BC (Cornelissen and Gustafsson 2004; Jonker et al. 2004). Direct evidence for attentuation of BC surface activity by natural substances has been obtained (Koelmans et al. 2009; Kwon and Pignatello 2005; Pignatello et al. 2006b; Wen et al. 2009). [Negative evidence is also reported (Cornelissen and Gustafsson 2006).] For example, wood char particles placed in a soil–water suspension experienced after a few weeks a 13-fold decline in the N2–BET-specific surface area and about a three-fold decline in benzene sorption coefficient. Loading of humic acid, fulvic acid or triglycerides (as surrogates for the humic lipid fraction) from the aqueous phase resulted in up to two-orders-of-magnitude reduction in char-specific surface area and a modest suppression of naphthalene and phenanthrene sorption (after correcting for partitioning to DNOM colloids) (Pignatello et al. 2006b). Sorption suppression increased with solute molecular size, consistent with competition between humic substances and the solute on external surfaces. The greatest suppression—an order of magnitude or more for phenanthrene—is found for samples of char that had been coflocculated with humic acid by Al3 þ (Fig. 1.16a) or merely coated with humic acid (Figure 1.16b). Tannic acid, a model for humic substances, decreased surface area and sorption of hydrophobic compounds (Qiu et al. 2009). It is also known that metal ions compete with nonionic compounds on the surfaces of both BC (Chen et al. 2007a) and carbon nanotubes (Chen et al. 2009); in these cases, it is believed that the adsorbed metal ion and its hydration sphere “crowd out” the organic. In summary, these results collectively suggest that NOM, mineral–NOM complexes, and metal ions play a role in sorption suppression by competition for surface area and/or by pore blockage.

10-7

10-6

10-5

10-4

C w, mmol/L

(b) 10000 Sorbed concentration (mg/kg)

100

10-8

100 1E-3

0.01

0.1

Equilibrium concentration (mg/L)

Figure 1.16. (a) Strong suppression of phenanthrene sorption by coating black carbon with humic substances: elemental carbon– normalized isotherms in wood charcoal and in the complex obtained by adding Al ion to a suspension of black carbon in a soil humic acid solution (Al-HA-char co-floc). The EC-normalized isotherm of the latter is obtained after subtracting out the OC contribution on the the basis of Freundlich parameters of the AlHA co-floc.[From Pignatello et al. (2006a)]. (b) Freundlich fit of phenanthrene isotherm by a char (&), HA (.), lipid (~), and chars coated with HA at low (&), medium (*), or high (~) levels, and char coated with lipid at low (!), medium (x), or high ( ) levels. [From Wen et al. (2009).]

1.3.4. Competitive Sorption Contaminants in the environment more often exist in mixtures than singly. Competition is treated here as a sorbentcentered phenomenon because it impacts the sorbent reorganization free-energy term (step 2, equation 1.1), creating a negative driving force for sorption of the principal solute. Put simply, competition between the principal solute and a cosolute will be experienced when sites are distributed in energy and sites of a given energy are limited in number. Nonlinearity, after all, is simply a manifestation of selfcompetition due to site energy distribution. Competitive

29


QL1 KL1 C1 P q1 ¼ 1 þ KL1 C1 þ n2 KLn Cn

ð1:22Þ

Sander and Pignatello (2005a) showed that benzene and toluene compete for the same sites on a charcoal, and that, when normalized for the hydrophobic effect via the hexadecane–water partition coefficient, the single-solute benzene and toluene data, together with the bisolute data (qben þ qtol, molar basis), fall onto a single isotherm (Fig. 1.17b). Hence, benzene and toluene are “ideal” competitors and replace each other on the charcoal surface in a 1 : 1 molar ratio. Nitrobenzene was also found to be an ideal competitor for both toluene and benzene on the same char (Sander and Pignatello 2005a). All three competition models mentioned above predict the greater the competitive effect, the more linear the isotherm of the principal solute will be [apparent by inspecting Eqs. (1.22) and (1.23)]. This is generally confirmed experi-

(a)

10 1

10 0 q [mol kg-1]

sorption in soils and isolated humic substances has been noted by many investigators. Competitive effects among PAHs for BC is argued to account for the lower-than-expected sorption of individual PAHs in sediments (Cornelissen and Gustafsson 2006). Competition has also been observed between contaminants and small natural molecules such as aromatic acid plant exudates in soils (Xing and Pignatello 1998), and between contaminants and dissolved humic acid on BC (Koelmans et al. 2009; Pignatello et al. 2006a). Competition has also been shown to increase the bioavailability of the principal solute in soils (White et al. 1999a) and is expected to play an even greater role in the bioavailability of chemicals sorbed to BC because nonlinearity is typically more pronounced than in NOM (Pignatello 2009). Competition may also occur between contaminants and unburned liquids accompanying soot particles; for example, phenanthrene sorption was weaker and more linear in a reference diesel soot having the greater amount of native extractable oils (20% vs. 2% by weight) (Nguyen and Ball 2006). Competition for sorption of a solute 1 is manifested in the Langmuir model by the sum of terms in the denominator corresponding to solutes 2 through n:

10 -1

-1

CW_0(TOL)= 0.00 mol L -1 CW_0(TOL)= 5.24E-6 mol L -1 CW_0(TOL)= 1.11E-5 mol L CW_0(TOL)= 2.30E-5 mol L-1 -1 CW_0(TOL)= 4.88E-5 mol L CW_0(TOL)= 1.07E-4 mol L-1

10 -2

Assuming that each site of a discrete energy is limited in number, a competitive Freundlich isotherm can be derived (Fritz and Schlunder 1981; Sheindorf et al. 1981), !n1 1 n X q1 ¼ KF1 C1 C1 þ a1n Cn ð1:23Þ

10 -3

benzene competitor: toluene

10

CW_0(TOL)= 4.23E-3 mol L

-4

10 -8

10 -7

10 -6

10 -5

10 -4

10 -3

-1

10 -2

Cw[mol L-1]

2

(b) 10 1 Benzene Toluene

q(BEN) + q(TOL) [mol kg-1]

where a1n represents the competition coefficient between solute 1 and solute n. In a bisolute system, the competition coefficients can be obtained by regressing a linearized form of equation 23 (not shown). In theory, a12 ¼ 1/a21 in a bisolute system. The model has been tested successfully on carbon adsorbents (Fritz and Schlunder 1981; Sheindorf et al. 1981). Another important competitive model is ideal adsorbed solution theory (IAST), which is derived from the Gibbs equation. This theory has been applied successfully to multisolute sorption on activated carbon (Crittenden et al. 1985; Radke and Prausnitz 1972) and soil (e.g., McGinley et al. 1989; Xing et al. 1996). The main assumption of IAST is that in dilute conditions the adsorbed phase forms a twodimensional ideal solution, such that the fugacity of each adsorbed component in a mixture is proportional to its adsorbed-phase mole fraction times its fugacity in the single-solute system at the same spreading pressure. Spreading pressure is defined as the difference in interfacial tension between pure solvent–solid and solution–solid systems. The advantage of IAST is that, in principle, it is adaptable to any isotherm model, although the mathematics get quite complicated for all except the simplest models.

CW_0(TOL)= 2.47E-4 mol L -1 -1 CW_0(TOL)= 6.18E-4 mol L -1 CW_0(TOL)= 1.57E-3 mol L

10 0

CW_0 (TOL)=8.24E-6 mol L

-1


-1


-1


10 -1 -1


-1


10 -2

-1


-1


-1

10 -3 10 -5


10 -4

10 -3

10 -2

10 -1

10 0

10 1

-1 CH(Ben)+CH'(TOL) [mol LH ]

Figure 1.17. (a) Sorption isotherm of benzene in the benzene–toluene bisolute system. CW_0(TOL) is the time zero molar concentration of toluene. (b) Sorption data plotted as the sum of sorbed concentrations of the two solutes versus the sum of the corrected solution concentrations. Corrected solution concentration is the equivalent concentration in n-hexadecane, which was used as a reference solvent to normalize for hydrophobic effects (isotherms also overlapped when are used benzene as the reference solvent).

30


mentally (e.g., McGinley et al. 1989; Sander and Pignatello 2005a, 2007) (see Fig. 1.17a). Competition-forced linearity has three major origins: 1. At a given co-solute concentration, the co-solute becomes less and less effective as a competitor with increasing principal solute concentration. This has the effect of increasing the slope of the principal solute’s isotherm on a log scale. 2. When both a dissolution phase and an adsorbing phase are present, suppression of adsorption occurs to the benefit of dissolution-type sorption. 3. When the sorbing phase is glassy, the cosolute competitively fills holes and, at sufficiently high concentration, plasticizes the solid to the rubbery state, which exhibits more linear behavior. An assumption common to all three competition models is that the respective sorption domains fully overlap; that is, all sites are available to all molecules, albeit with unequal affinity. If this condition is not met, the respective competition isotherms cannot be constructed from a simple combination of the single-solute isotherms. This hypothesis has not been fully tested, but some results are supportive. Using IAST, Xing et al. (1996) modeled competition between atrazine and either a triazine analog, prometon, or the structurally unrelated compound, trichloroethene, on a high-organic soil, humic acid particles, the commercial glassy polymer poly(2,6-diphenyl-p-phenylene oxide) known as Tenax, and amorphous silica gel as shown below. They found that, whereas IAST simulated competition by prometon satisfactorily, it greatly overpredicted competition by trichloroethene. Whether for steric or electronic reasons, the shared sorption domain of atrazine and trichloroethene on these sorbents is smaller than that of atrazine and promoton. The importance of steric effects in sorption to NOM is suggested by the inverse relationship between the competitive effect and the difference in molecular free surface area between competing halogenated aliphatic compounds (Pignatello 1991). Moreover, whereas benzene shares the same sorption domain with nitrobenzene and toluene on a charcoal, the size exclusion effect on this adsorbent (Section 1.2.2.6 and Fig. 1.5) clearly prevents

benzene and naphthalene or benzene and phenanthrene from fully sharing the same sorption domain. Schaefer et al. (2000) suggested that solute polarity and/or sorbent geometry may have contributed to the failure of IAST to simulate competition between trichloroethene and tetrachloroethene on natural solids. Adsorption of hydrophobic compounds by BC is competitively suppressed by water (Endo et al. 2009a). Water molecules compete for surface area by interacting with the graphene surface through dispersion forces and especially by clustering around polar functional groups along the platelet rims and defect sites. Finally, competition can lead to apparent hyteresis if the cosolute concentration is altered in the desorption step (see Section 1.3.5). 1.3.5. Hysteresis Hysteresis means the lagging of a physical effect on a body behind its cause (Merriam-Webster 2000). The phenomenon of sorption hysteresis is widely encountered. It results when chemical released by dilution of the solute in the fluid phase is less than predicted by the isotherm constructed in the “forward” (uptake) direction. Hysteresis may be true or artificial. True, and in some cases artificial hysteresis has important implications for environmental behavior. Among its consequences are lower bioavailability, greater “tailing” of a contaminant plume traveling through porous medium, and stronger resistance toward remediation than predicted assuming reversibility. 1.3.5.1. Artificial and True Hysteresis. Artifacts fall into three categories: degradation, insufficient time allowed for mass transfer to reach a static condition, and perturbation of a competitive situation. If in an experiment sorbed concentration is calculated by mass balance rather than measured directly, uncorrected chemical or biological transformation of the parent compound will lead to overestimation of the actual sorbed concentration. Mass transfer limitations can result in underestimation of the equilibrium sorbed concentration in the uptake step, overestimation of the equilibrium sorbed concentration in the release step, or both. Such artifact has been documented in experiments where the contact time was deliberately shortened (Altfelder et al. 2000), and underscores the importance of taking mass

OCH3

Cl N CH3CH2HN

N N

Atrazine

N NHCH(CH 3)2

(CH3)2CHNH

N N

Prometon

Cl

Cl

H

Cl

NHCH(CH3)2 Trichloroethene


transfer rates into account in models that include sorption terms. The plasticizing effect of moisture on NOM takes a number of days to complete (Schaumann and LeBoeuf 2005), so artificial hysteresis can occur if sufficient hydration time is not allowed. A third type of artificial hysteresis is known as the competitor dilution effect (Sander and Pignatello 2007). The competing substance may be a solute or a nonseparable microsorbent in the liquid phase. Examples of a competing microsorbent include colloidal NOM (Curl and Keolelan 1984) and surfactant micelles. In any case, because competitive pressure is partially relieved when the competitor concentration is reduced in the dilution step, the principal solute sorbs more strongly than predicted by the uptake branch of the isotherm. Sander and Pignatello (2007) showed that pairwise sorption to a charcoal of benzene, nitrobenzene, and toluene resulted in much greater hysteresis in the biosolute case than in the corresponding single-solute case owing to the competitor dilution effect. The competitor dilution effect—while artificial in the sense that it is a secondary effect and competitor concentration is a controlled variable—has real implications because competing solutes and nonseparable microsorbents are more often than not a part of natural and disturbed environments. The underlying cause of true hysteresis is the formation of a metastable state or states during the sorption process. True hysteresis may be distinguished from artificial hysteresis caused by mass transfer limitation because the metastable state persists indefinitely in the absence of a perturbing force, like the book balanced on its edge. Sorption in such cases is deemed “irreversible” in the thermodynamic context of the word in that sorption does not follow the same pathway in the forward and reverse directions. (“Irreversible” should not be confused with its more common meaning, that the chemical cannot be retrieved without drastic means.) The existence of different pathways in the forward and reverse directions implies that sorption is dependent on the history of the system. (Interestingly, the words hysteresis and history are not related etymologically.) At least two types of true hysteresis have been identified for organic compound sorption. One is capillary condensation hysteresis (Fig. 1.18a) (Rouquerol et al. 1999). This occurs when vapors condense in a fixed mesopore to form a metastable film that collapses at a threshold pressure to form the thermodynamic meniscus plug state, which has a lower vapor pressure. It gives rise to the classic hysteresis loop observed at high relative pressures. The desorption branch represents the true equilibrium path. The other (Fig. 1.18b) is referred to as pore deformation hysteresis and occurs with sorbents whose matrix is sufficiently flexible to undergo physical changes by the incoming solute. At ordinary temperatures deformable sorbents typically include organic materials such as polymers, NOM, and possibly carbons.

31

1.3.5.2. True Hysteresis in Natural Organic Matter. First applied to glassy polymers, the concept of pore deformation hysteresis holds that permeant molecules exert pressure on internal pores or proto-pores smaller than the adsorbate molecular volume, causing pore expansion or creation (Fig. 1.18b). At the same time, the local matrix is softened by interaction of permeant with segments, increasing their mobility. On desorption from the local matrix the process attempts to reverse, but is unable to do so fully because the matrix stiffens before relaxation of the expanded or created pores is completed. Thus, swelling and shrinking is inelastic (irreversible) and leaves the local matrix with greater excess free volume after the cycle than before that persists indefinitely. The greater excess free volume leads to higher affininity for the solute in a subsequent desorption or repeat sorption step, manifesting as sorption hysteresis. For uptake of CO2 by polycarbonate, positron annihilation lifetime spectroscopy showed that the polymer is left with a greater hole volume and hole size after a sorption–desorption cycle than in the original sample (Hong et al. 1996).) Swelling of flexible matrices like gels and rubbers, on the other hand, is fully elastic (reversible) and shows no sorption hysteresis. In thermodynamic terms, during pore deformation hystersis in a glassy material. part of the free energy of sorption in a glassy material goes into matrix expansion that is not fully recovered during the cycle. The thermodynamic index of irreversibility (TII) via pore deformation mechanism may be defined in terms of the hysteresis loss in solute chemical potential over the cycle (Sander et al. 2005): TII ¼

lnCc lnCD lnC S lnC D

ð1:24Þ

where, respectively, CS, CD, and Cc represent solute concentration at the observed sorption point, observed desorption point, and the hypothetical reversible desorption point corresponding to the same sorbed concentration as at the observed desorption point (i.e., where qc ¼ qD). Considered another way, TII is the ratio of the observed to the upper-limit loss of free energy due to irreversibility. The TII is 0 for completely reversible systems and approaches 1 as mass desorbed approaches zero. Pore deformation hysteresis does not occur in melts, gels, or rubbery solids, nor in rigid, fixed-pore solids whose pores are not strained beyond their elastic limit (Bailey et al. 1971). These endpoints suggest a trend, illustrated in Figure 1.18c, in which the hysteresis potential of a solid matrix capable hypothetically of varying continuously between the rubbery and fixed-pore states, while holding all else constant, would reach a peak somewhere in between. This trend predicts that solids lying on the rubber side of the apex would show decreasing hysteresis with increasing solute concentration, while solids lying on the fixed-pore side of the apex would

32


(a) q

p/p

0

sorption

desorption

(b)

(reversible if “rubbery”) +

swelling

sorbate carbonaceous solid

(irreversible if “glassy”) shrinking

Part of Gsorp goes into matrix expansion

+

hysteretic effect

(c)

increasing matrix stiffness rubbery solid

glassy solid

fixed-pore solid

Figure 1.18. True hysteresis of organic compounds. (a) Capillary condensation hysteresis in a mesopore and the resulting hysteresis loop. The condensate film is metastable and collapses at a threshold pressure to the meniscus plug. The desorption branch represents the thermodynamic isotherm. (b) Schematic illustrating pore deformation hysteresis during a sorption–desorption cycle in a glassy solid. Sorption causes swelling that is reversible in very flexible (rubbery) solids and incompletely reversible (irreversible) in stiffer solids. (c) Schematic showing that the magnitude of the hysteretic effect in a solid whose matrix is able to vary continuously between the rubbery and very rigid (fixed-pore) states reaches a peak.

show the opposite. Sander found trends consistent with this prediction (Sander and Pignatello 2009). True hysteresis has been confirmed for NOM-rich solids by isotope tracer exchange experiments performed on two sorbents and two compounds, 1,4-dichlorobenzene, and naphthalene (Sander and Pignatello 2005b, 2009). Following equilibration at each step of a normal–sorption-desorption cycle, a slight amount of the 14 C-labeled chemical (tracer) is added and equilibrated under the same conditions, keeping the bulk chemical concentration constant. Such experiments showed that, while sorption of the bulk chemical is hysteretic

(depending on its concentration), tracer exchange was complete and nonhysteretic under all conditions. Thus, hysteresis is due to changes in the sorbent induced by uptake by the bulk chemical. This is further supported by differences in rates between bulk chemical and tracer in both the uptake and release directions (Sander and Pignatello 2005b, 2009). Convincing evidence for the pore deformation mechanism in NOM has been obtained in the so-called conditioning effect experiments (Lu and Pignatello 2002, 2004a,b; Sander et al. 2006; Xia and Pignatello 2001), where a sample of humic acid, whole SOM, or coal is “conditioned” by


subjecting it to a swelling–shrinking cycle performed by uptake–release of an organic solute. The response isotherm of a test compound—either the same or a similar compound—is constructed before and after conditioning. Such experiments show enhanced second-time sorption, consistent with irreversible free volume expansion by the conditioning agent. Figure 1.12 shows overall sorption enhancement and an increase in the contribution of hole filling relative to dissolution sorption as a result of conditioning in three organic matter materials, including a humic acid. Figure 1.19 shows sorption enhancement in a peat soil. As mentioned, the incompletely relaxed system represents a persistent metastable condition. Sorption enhancement through conditioning persists for months if the solid is stored at room temperature (Lu and Pignatello 2002). If the solid is heated (annealed), the conditioning effect diminishes in relation to the annealing temperature and time (Sander et al. 2006) (see Fig. 1.19). As in polymer systems, relaxation at a given temperature follows a double exponential rate law with a nonzero constant term descriptive of the final state that itself varies inversely with temperature. This means that at temperatures below the glass transition temperature, annealing does not fully relax the solid even after long annealing times. 1.3.5.3. Hysteresis in Black Carbon. True hysteresis may also be characteristic of carbons, although more study is needed to confirm it. Bailey et al. (1971) observed hysteresis of nonpolar organic vapor sorption by activated carbon; terming it “low-pressure hysteresis” to distinguish it from capillary condensation hysteresis in mesopores at higher pressures, they attributed it to “intercalation of molecules of adsorbate in narrow pore spaces leading to irreversible

changes in the pore structure.” Environmental BC clearly can be swelled by organic solvents (Akhter et al. 1985b; Braida et al. 2003; Razouk et al. 1968), confirming that BC platelets are held together in part by noncovalent forces. Swelling of wood charcoal by uptake of benzene dissolved in water affected the shape of the isotherm, and was likely responsible for benzene adsorption hysteresis (Braida et al. 2003; Sander and Pignatello 2007). Swelling may open up new sectors previously not connected with the external fluid. On desorption, some molecules become trapped as the polyaromatic scaffold collapses during desorption. The ability of some solvents to extract PAHs and other compounds from BC has been attributed to swelling (Akhter et al. 1985b; Jonker and Koelmans 2002a). Although NMR studies show an increase in molecular motion of sorbed 13 C-benzene with its concentration in charcoal (Smernik et al. 2006), other causes are possible. Finally, LeBoeuf and co-workers (Zhang et al. 2007) report thermal transitions indicative of a glass transition for several chars at 140 C–160 C, but interestingly not for diesel and hexane soot up to 200 C. 1.3.5.4. Hysteresis in Related Model Systems. Studying plant cuticular waxes deposited on montmorillonite Chen and Xing (2005) found sorption hysteresis of naphthalene and phenanthrene above but not below the approximate point where the sorbate induces a phase transition from a more to a less flexible amorphous state. This behavior is the opposite that observed for NOM. These authors proposed that hysteresis is due to “trapping” of molecules caused by collapse of the flexible amorphous state to the less flexible state on desorption.

Distribution ratio (sorbed divided by solution concentrations, L/kg)

8000

104

6000

20 oC, 21 d o

50 C, 6h o 60 C, 6h 75 oC, 6h 90 oC, 6h

4000

103 2000

10 -3

10 -2

10 -1

33

10 0

10 1

0.01

0.02

-1 Solution concentration, mg L

Figure 1.19. Sorption isotherm of 1,2,4-trichlorobenzene at 20 C on Pahokee peat before (filled circles) and after (open triangles) conditioning with chlorobenzene. The dashed curves are fits to the dual-mode model [Eq. (1.21)]. The inset shows the effects of storage at 20 C for 21 days and heating (annealing) at different temperatures for 6 h. It also shows a tendency for the distribution ratio to “relax” to its original isotherm after annealing. [After Sander et al. (2006).]

34


Sorption of organic compounds by expanding clays may exhibit sorption hysteresis if sorption induces hysteretic swelling or structural reorganization of the quasicrystals (Chatterjee et al. 2008).

1.3.6. Apportionment of Sorption between NOM and BC in Environmental Samples Soil and sediment organic matter may contain a mixture of NOM and BC. While BC may be a minor component (estimates place it at 1%–10% of total organic carbon in undisturbed soils, higher at industrial sites and in marine sediments), BC in its raw state is a much stronger sorbent than NOM by an order of magnitude or more, and therefore may play a role disproportionate to its mass fraction. Researchers have tried to partition sorption in whole soil or sediment by using multiterm equations that account for contributions by NOM and BC (Accardi-Dey and Gschwend 2002, 2003; Allen-King et al. 2002; Xia and Ball 1999). In some cases coal and NOM are distinguished (Cornelissen and Gustafson 2005). The models use a linear “partitioning” term for NOM and either Freundlich or Polanyi–Manes terms for the other carbonaceous sorbents, as, for example, the following two-compartment isotherm qT ¼ fOC KOC C þ fBC KBC CnBC

ð1:25Þ

where qT is the total observed mass-based sorption, OC is NOM carbon, BC is BC carbon, and f is the mass fraction of each type of carbon in the sample. It is important to understand, however, that this model rests a number of assumptions that are questionable: 1. As yet, there is no generally accepted method for quantifying fBC in soils and sediments (Hammes et al. 2007, 2008), especially at low BC levels. A fundamental analytical problem is that “BC” is not a single material but a continuum of materials with different properties, some of which overlap NOM. In fact, raw BC often contains a nontrivial amount of sp3 C and nonaromatic sp2 C from functional groups on graphene rims and from incompletely carbonized precursors in pores (Schmidt and Noack 2000). Thus, a single analytical method may sample only a window of the BC continuum. The most commonly used of the various methods proposed for quantifying BC is chemical–thermal–oxidation (CTO375) method—the sample is first acidified to remove inorganic carbon, then combusted at 375 C in air to volatize NOM, and finally the BC carbon that is assumed to have survived is then quantified by conventional hightemperature combustion (Gustafsson et al. 1997; Nguyen et al. 2004). The CTO-defined fBC must be regarded as operational, however, since studies show that low-temperature combustion also volatilizes most or all of the char and a large

portion of the soot (Nguyen et al. 2004). This means that fBC is likely to be underestimated in this test. Hawthorne et al. (2007) were unable to explain PAH partitioning in 114 historically contaminated and background sediments by a linear KOC–KBC dual model on the basis of the CTO-375 method. Another method that has been attempted is the acid–dichromate method [See Grossman and Ghosh (2009), Knicker et al. (2007), Pignatello et al. (2006b), and references cited therein]. However, acid–dichromate does not oxidize polymethylene groups (Knicker et al. 2007; Pignatello et al. 2006b), which can constitute several percent of NOM; thus, acid–dichromate will likely overestimate fBC. The contribution of BC to total sorption will be overestimated proportionate to the degree fBC is underestimated, and vice versa. Likewise, coal contains substances overlapping in properties with both “ordinary” NOM and BC. So, while there are petrographic methods to identify BC and coal (Karapanagioti et al. 2001, 2000), there are no unambiguous methods yet to quantify the “BC,” “NOM,” and “coal” contents of soil or sediment for purposes of predicting sorption and bioavailability. 2. The validity of using isotherm parameters obtained from existing reference standards for BC in the sample is questionable, as such parameters are not universal, and it is seldom possible to trace the source of BC in a sample. For example, the Freundlich KBC varied by two to three orders of magnitude, and nBC varied by several tenths of a unit (Jonker and Koelmans 2002b) on different raw BCs. Moreover, the value of nBC depends on the concentration range over which it is calculated (see Sections 1.3.2 and 1.3.3). Overestimating KBC would correspondingly overestimate the contribution of BC at any given concentration. Overestimating nBC would overestimate the contribution of BC, increasingly so as solute concentration declines. Some investigators have used the residual carbon remaining after CTO treatment as the reference standard; however, such treatment removes non-BC materials (unburned biomass, NOM, unburned liquids) that can strongly affect sorption (Endo et al. 2009a; Pignatello et al. 2006a). 3. As is usual in applications of Equation (1.25), all nonlinearity is attributed to BC. However, as we have seen, sorption even to BC-free humic substances can be nonlinear. When reasonable nonlinearity is ascribed to NOM, the contribution of BC declines compared to the case where all nonlinearity is assigned to BC. This is illustrated in Figure 1.20. The estimated contribution of BC to sorption in a fabricated dataset in which both the OC and BC terms in Equation (1.25) are nonlinear is considerably less than when all the nonlinearity is ascribed to the BC term (Fig. 1.20a), even as the data remain accurately predicted on forcing the OC term to be linear (Fig. 1.20b). 4. In most applications of Equation (1.25), KOC is estimated from historic OC–octanol FERs. These values are


Fraction sorbed to black carbon

(a) 1.0

(especially in the case of highly hydrophobic compounds). The contribution of BC will be overpredicted proportionate to the degree that KOC is underestimated. 5. As we have seen in Section 1.3.3, adsorption on BC is attenuated by weathering in the environment. Attenuation can be an order of magnitude or more.

0.8

0.6

1.3.7. Mass Transfer Rates

0.4 data, nOC = 0.80; nBC = 0.60 calc assuming linear OC term

0.2

data, nOC = 0.85; nBC = 0.60 calc assuming linear OC term

0.0 -6

-5

-4

-3

-2

-1

0

1

2

3

log solute concentration (b) 6

log sorbed concentration


4


2

0

-2 -8

35

-6

-4 -2 0 log solute concentration

2

4

Figure 1.20. Plots illustrating the effects of forcing sorption by OC to be linear on estimating the fraction sorbed to black carbon in a soil containing 90% OC and 10% BC relative to total organic carbon. (a) Fabricated data are the sum of a Freundlich isotherm to BC, where KBC ¼ 2000 and nBC ¼ 0.6, and a Freundlich isotherm to OC, where KFOC ¼ 300 and where nOC is either 0.80 (circles) or 0.85 (crosshairs). The lines represent calculated values after (1) a linear regression of the OC data forced through zero, which yielded KOC ¼ 96.8 (circles data) or 128 (crosshair data); and (2) regressing log (qT –KOCC) versus log C, which resulted in modified values K0 BC ¼ 3184 and n0 BC ¼ 0.625 (corresponding to circles data), or K0 BC ¼ 3007 and n0 BC ¼ 0.626 (corresponding to crosshairs data) (b) plot shows that the isotherms remain unchanged by linearizing the OC data.

taken to truly represent NOM, yet they were obtained on ordinary soils that contain unknown amounts of BC particles that contribute to sorption to an unknown degree due to the problems mentioned in paragraphs 1–3 and 5. Sorption to OC was generally assumed linear in the historic studies. Consequently, there is a certain degree of circular reasoning behind the acceptance of reference KOC values. Furthermore, too often sorption was underestimated in the historic studies because the equilibration times were insufficiently long

1.3.7.1. General Considerations. Thus far this chapter has dealt with sorption equilibria. However, sorption/desorption rates are often critical factors controlling the transport and bioavailability of contaminants. The rates of sorption and desorption of physisorbing compounds are governed primarily by molecular diffusion and secondarily by matrix flexing properties. Both of these processes are governed by the structure of the sorbent. Diffusion is the tendency of molecules to migrate in response to a gradient in their chemical potential so as to achieve maximum entropy. Diffusivity is a function of molecular structure, the nature and geometry of the diffusing medium, concentration gradients, interfacial boundary conditions, and temperature. Rate laws for soil systems inevitably involve simplifying assumptions because of the system heterogeneity. The simplest situation is the one in which the chemical is stable over the observational timeframe, and soil particles are well dispersed in a fluid such that diffusion and advection in the fluid phase is not limiting. The complexity increases dramatically as one moves to the soil column situation; as water flows or evaporates; and as biological, chemical, or physical processes act to remove the contaminant or change concentration gradients. Diffusive equilibrium in well-mixed water suspensions may require as short as a few hours or as long as several months, depending on the compound and the soil or soil isolate being tested (Pignatello and Xing 1996). Equilibrium after performing a dilution step can take even longer. Many studies have used physical stripping techniques to mimic desorption to an infinitely dilute fluid (sink). Commonly, the stripping technique employs a stream of gas (Werth and Hansen 2002; Werth and Reinhard 1997a,b), or a polymer resin, such as Tenax or XAD added in sufficient quantity to ensure theoretical quantitative mass transfer had equilibrium been achieved (Pignatello 1990a; Zhao and Pignatello 2004). Complete desorption of strongly sorbing contaminants under stripping conditions can require exceedingly long times. This is illustrated in Figure 1.21 for phenanthrene that was desorbed from six different sterilized soils after an uptake step lasting 180 days. Desorption was carried out for up to 606 days in the presence of a large excess of Tenax renewed at each timepoint. Desorption tended eventually toward completion in some soils, but in others was reduced to extremely slow rates and considerable levels of phenanthrene remained.


1

1 Seal Beach 240 µg/gOC 25000 µg/gOC

250 µg/g OC 4100 µg/g OC

Pahokee

0.1

Seal Beach

0.1 1

1



log q / q0

36

0.1

0.1 Mount Pleasant

Modified Seal Beach

0.01 1

1


160 µg/g OC 2000 µg/g OC 13000 µg/g OC

Cheshire

Port Hueneme

0.1

0.1 0

100 200 300 400 500 600

0

100 200 300 400 500 600

Time, Days

Figure 1.21. Desorption of phenanthrene from different sterilized soils after preequilibration for 180 days. Lines are for visualization only. [Data are from Braida et al. (2002).] The kinitial and kfinal were calculated on the basis of the first three and last three data in each case.

q0, mg/gOC Pahokee peat Pahokee peat Mount Pleasant Mount Pleasant Cheshire Cheshire Cheshire Seal Beach Seal Beach Modified Seal Beach Modified Seal Beach Port Hueneme Port Hueneme

250 4,100 280 4,900 160 2,000 13,000 240 25,000 530 25,000 980 21,000

kinitial, d1 0.12 0.20 0.35 0.96 0.29 0.30 0.35 1.0 0.85 0.81 0.97 0.28 0.29

kfinal, day1 0.00073 0.00047 0.0027 0.15 0.00050 0.00050 0.00023 0.00053 0.00079 0.0012 0.00020 0.0012 0.00049


1.3.7.2. Nature and Geometry of the Diffusing Medium. Soils may contain mineral grains with patches and/or coatings of humic substances on their surfaces. These gains may be cemented together in aggregates having a wide range in both pore size and pore connectivity. The aggregates may include microscopic particles of NOM in varying stages of diagenesis, as well as particles of BC. Intraparticle mass transport may involve diffusion through pore fluids (pore diffusion), along pore walls (surface diffusion), or through solid matrices of organic matter (solid-phase, or matrix, diffusion). Pore and surface diffusion are conceptually difficult to distinguish in pores not greatly larger than the width of the diffusant, such as micropores (IUPAC definition, < 2 nm). Diffusion of molecules to or from the furthest reaches of a soil particle requires multiple “jumps” and the crossing of many grain–grain and grain–water interfaces. The length scale over which diffusion is rate-limiting may be much smaller than the macroscopic particle radius and will depend on the micromorphology of the particle (Pignatello 2000). Pore diffusion is retarded by the tortuosity of pore network pathways, sorption on pore walls, and steric hindrance. Steric hindrance begins to appear when the minimum critical molecular diameter is about 10% of the pore diameter and becomes severe as it approaches 100% (K€arger and Ruthven 1992). Steric effects for most molecules of interest will be important in micropores and the smaller mesopores. Matrix diffusion requires cooperative motions between the diffusant and matrix strands as the molecule “jumps” from site to site (Pignatello 2000). Relatively soft NOM represented by humic acid films impedes diffusion by three or four orders of magnitude compared to bulk water (Chang et al. 1997). Glassy solids can impede molecular diffusion by many more orders of magnitude, depending on diffusant diameter, the glass transition temperature Tg (Pignatello 2000), and the magnitude and interconnectivity of the free volume in the solid. The basis for intraparticle diffusion models is Fick’s second law, which is given in radial coordinates for a v-dimensional (v ¼ 1 for a slab, v ¼ 2 for a cylinder, v ¼ 3 for a sphere) particle that is isotropic and homogeneous, as follows @s 1 @ @s ðv1Þ ¼ r Deff ðT; sÞ ð1:26Þ @t rðv1Þ @r @r where r is the thickness (slab) or radius (cylinder, sphere) of the diffusing medium; s (M/L3) is the total local volumetric concentrationin the diffusing medium, including sorbed chemical and chemical dissolved in internal pore water; and Deff(T,s) [L2/T] is the effective diffusion coefficient, or diffusivity, which may be concentration- and/or temperature-dependent.

37

The average sorbed concentration in the particle is obtained by integration; in the example of a sphere it is given by SðtÞ ¼ 3 r3

ðr sðr; tÞr2 dr

ð1:27Þ

0

The boundary condition at the particle–external solution interface is given by, sðtÞr¼r ¼

ds CðtÞ dC

ð1:28Þ

where ds/dC is the distribution coefficient usually presumed equivalent to Kd obtained from the isotherm. Analytical solutions to the diffusion equation when Deff is constant are available for uniform spherical particles applicable to various situations, such as finite external solution, infinite external solution of constant or variable concentration, and desorption to a “vacuum” (Crank 1975; Haws et al. 2006; K€arger and Ruthven 1992; Pignatello 2000) (see also Chapter 8). Numerical solutions have been worked out for less homogeneous media and situations of greater complexity. Typically, the output parameter of the diffusion model is Deff/reff2 (T1) since the characteristic diffusion length scale reff is unknown. If the diffusing medium is an isotropic phase like NOM or a network of micropores, Deff is not broken down further. If the diffusing medium is a network of waterfilled meso- or macropores, then intraparticle diffusion can be regarded to occur only in the pore fluid with retardation due to sorption on or in the walls. The Deff may then be expressed as Deff ¼

wkt1 Dw w þ ð1wÞK

ð1:29Þ

where j is the intraparticle porosity; Dw is the diffusivity in water; K is the sorption distribution coefficient, usually taken to be the bulk Kd; t ( 1) is a tortuosity factor that reflects deviation from straight-line paths and pore interconnectedness (often taken to be proportional to j1); and k (1) is a parameter that reflects steric hindrance by the pore walls. Steric effects become significant when the molecular diameter reaches about 10% of the pore diameter (K€arger and Ruthven 1992). In reality, pore surface diffusion occurs as well, but it is difficult to estimate out. If sorption sites are microdomains of NOM on internal pore surfaces, then Deff should correlate inversely with OC content of the soil, since K would be proportional to fOC [Eq. (1.29)]; so far, such a correlation seems not to exist (Birdwell et al. 2007; Kukkonen et al. 2003; Shor et al. 2003).

38


Soil heterogeneity has an important influence on rates. Theoretically, the uptake or release rate is inversely related to the square of the particle radius. In some studies, the rate is found to increase with decreasing nominal particle radius (Ball and Roberts 1991b; Kleineidam et al. 1999b; Wu and Gschwend 1986), while in others no dependence is observed (e.g., Carroll et al. 1994; Farrell and Reinhard 1994; Steinberg et al. 1987). However, it is usually true that pulverization of the soil increases rates (Ball and Roberts 1991a; Pignatello 1990b; Steinberg et al. 1987). Examining coaly sedimentary sands and gravels separated on the basis of size, color, and porosity, Kleineidam (1999b) found that sorption rate decreased with increasing size, increasing OC content, and decreasing porosity of particles. As described by Cook in Chapter 13 of this book, water wetting of soil is a process that affects the rate of chemical uptake, yet may take days to reach equilibrium. Matrix flexing has an effect on rates. Sander and Pignatello (2005b; 2009) compared uptake and release curves of a chemical and its radiolabeled tracer. The experiments were performed under identical gradient conditions, except that the tracer was added after the bulk chemical had already equilibrated and in an amount sufficiently small so as to not perturb the bulk chemical equilibrium. In all cases (naphthalene and 1,4-dichlorobenzene; two solids; high and low concentrations), tracer rates were faster, especially at the higher bulk chemical concentration. There are three reasons for this result: (1) diffusion is faster for the tracer because the bulk chemical already occupied the higher energy (slower filling/emptying) sorption sites, (2) diffusion of the tracer takes place in an already filled solid that is more open and flexible than it was during diffusion of the bulk chemical because of plasticization (Section 1.3.2.3), and (3) sorption or desorption of the bulk (but much less so for the tracer) is accompanied by matrix swelling or shrinking that prolongs its uptake or release; this is reflected in a distinct biphasic shape of the curves for the bulk (but not tracer) at its higher (but not lower) concentration. 1.3.7.3. Influence of Solute Structure, Solute Concentration, and Competition. Typically, the characteristic rate parameter decreases with increasing molecular size or hydrophobicity. Such behavior is expected in all of the relevant diffusion media—in pores, on surfaces, and in a matrix—and is consistent with the findings of fundamental studies on polymers (Berens 1989; Rogers 1965) and porous inorganic reference materials (K€arger and Ruthven 1992). However, systematic studies of natural solids involving more than three compounds are rare (Carroll et al. 1994; Piatt and Brusseau 1998), as are studies contrasting polar and nonpolar molecules of similar size (Piatt and Brusseau 1998). Compounds with multiple points of interaction are expected to diffuse more slowly because when strong interactions occur

simultaneously at more than one place in the molecule, all such interactions must be broken before the molecule can jump to the next site. Diffusion in glassy materials is far more sensetive to molecular diameter than in rubbery materials. Take, for example, the diffusion of gases and hydrocarbons up to 10 A in diameter in polyvinylchloride (PVC; Tg ¼ 85 C) (Berens 1989): the diffusion coefficient declined at the rate of 3.4 log units per angstrom diffusant diameter in the asprepared (glassy) polymer, while in the plasticized polymer (with added phthalate ester compounds) it declined at the rate of only 0.46 log units per angstrom. The concentration dependence of the diffusion coefficient, D for a chemical species within a given diffusing medium (say, NOM or the interstices of a porous particle) is dictated by the gradient in chemical potential with respect to concentration (K€arger and Ruthven 1992; Pignatello 2000): D ¼ D0

dlnp dlns

ð1:30Þ

Here, p is the pressure in the external fluid, s is the concentration in the diffusing medium, and D0 is the self-diffusivity (also called corrected diffusivity). Thus, D ! D0 when s is a linear function of p, which is always true as s ! 0. Thus, sorption nonlinearity and competition affect sorption and desorption rates (Braida et al. 2001, 2002; Sander and Pignatello 2005b, 2009). Sorption–desorption by a compound exhibiting a linear isotherm with a particular sorbent is symmetric; that is, the normalized rate will be independent of the absolute concentration, and the normalized sorption curve (Mt/M1 vs. t) and the corresponding normalized desorption curve after dilution will coincide. However, the same is not true for a compound exhibiting a nonlinear isotherm. Recalling that a concavedown isotherm indicates weakening affinity with increasing concentration, the normalized sorption or desorption rate increases with absolute concentration because the solid provides progressively lower resistance to diffusion. Thus, for two identical particles placed in separate infinite liquid media, the one in the more dilute medium will take longer to reach equilibrium. Moreover, in an experiment in which desorption follows sorption by diluting the liquid phase, the system will take longer to reach equilibrium in the desorption than in the sorption step, in relation to the deviation from isotherm linearity. This happens because the strongest sites are filled from a relatively high concentration source, but emptied to a relatively low concentration sink. Competition also affects diffusion. A competing cosolute will increase the rate of sorption or desorption of a principal solute, proportionate to cosolute concentration, because the principal solute occupies weaker and weaker sites, and becomes progressively more labile (White et al. 1999c; Zhao et al. 2001).


Diffusion in solids is temperature-dependent. The governing equation is D ¼ AeEa =RT

ð1:31Þ

where Ea is the diffusion activation energy and A is a regression parameter. Diffusion is more temperature-dependent in glassy than rubbery solids and in microporous than mesoporous solids. 1.3.7.4. Strongly Resistant Desorption. Over the years many researchers have discovered that chemicals placed in contact with soil can generate a fraction that becomes highly resistant to desorption and biodegradation. Such fractions may be found in historically contaminated samples as well as in spiked samples of clean soil, and may be exhibited even by small, weakly sorbing molecules like C2 and C3 halogenated hydrocarbons. This topic is the subject of several reviews (Brusseau and Rao 1989; Luthy et al. 1997; Pignatello 1990b, 2000; Pignatello and Xing 1996) and will only be summarized here. Desorption resistance has important implications for environmental transport, natural attenuation, bioavailability, and bioremediation and physicochemical remediation strategies (Alexander 1995, 2000; Loehr and Webster 1997; Pignatello and Xing 1996; Pignatello, 2009). The term resistant—and the converse, labile—are not rigorously defined but depend on the experimental timeframe and methodology. Nevertheless, a fraction showing pronounced resistance often makes up a nontrivial fraction— up to several percent or more of the total concentration present. Historically contaminated samples may be enriched in the resistant fraction, due to the depletion of more labile fractions by dissipation and degradation during the lengthy time that passes before the sample is collected (Pignatello et al. 1993). It has been shown that a highly resistant fraction can be generated after a contact time as short as a few hours (Cornelissen et al. 1997; Kan et al. 1997; 1998; Pignatello 1990a,b; Ten Hulscher et al. 2005). Biodegradation of a chemical, previously allowed to equilibrate with a sterilized soil before inoculation, often tails off to leave a biodegradation-resistant fraction that correlates with the desorption-resistant fraction (Braida et al. 2004; White et al. 1999b) [see also citations in Pignatello and Xing (1996)]. An example of extreme resistance of an otherwise labile compound in a historically contaminated soil is 1,2-dibromoethane. This biodegradable, volatile, and moderately water-soluble compound persisted at microgram per kilogram (mg/kg) levels in topsoil up to at least 19 years after its last known application (Pignatello et al. 1990; Steinberg et al. 1987). Compared to similar concentrations of freshly added 14 C-1,2-dibromoethane, the field residues exhibited far greater soil–water distribution ratios (Kd), far slower rates of desorption, and greatly reduced degradation by native

39

microorganisms. An example of extreme resistance in a spiked system is trichloroethene (TCE); passage of a stream of N2 at 100% relative humidity through preequilibrated soil columns released most of the TCE with 10 min, but a small fraction was extrapolated to take months to years to desorb (Farrell and Reinhard 1994). Various nonexhaustive extraction techniques have been proposed for measuring physical availability or predicting bioavailability of contaminants. A popular approach is the use of polymer adsorbents such as Tenax or XAD resin added in large excess. Using Tenax, Cornelissen et al. (2000) applied an exponential desorption model that assumes “fast,” “slow,” and sometimes “very slow” compartments. Each compartment desorbs in a first-order manner. The twocompartment model is given by q ¼ Ffast ekfast t þ Fslow ekslow t q0

ð1:32Þ

where q0 is the initial sorbed concentration and F represents the mass fraction and k (T1) the rate constant for contaminant in the designated compartment. This approach is equivalent to applying a driving force condition at the surface [i.e., rate is proportional to the difference in chemical activity in the two phases at the interface (near zero in the aqueous phase)] and assuming that the chemical in the particle is always well mixed. In the example of Figure 1.21, one can see that the final desorption rate constant (corresponding to the slope of log q/q0 vs. time) is very much smaller than the initial desorption rate constant—in most cases, three orders of magnitude smaller. Several studies report correspondence between the Tenax-desorbed fraction and the biodegraded fraction (Cornelissen et al. 2000; Braida et al. 2004; Lei et al. 2004; Li et al. 2005; Pignatello 2006). Possible causes of strongly resistant desorption include the following: .

The normal process of retarded diffusion in and out of remote domains . Occlusion of molecules in closed pores during particle genesis . Alteration of the soil matrix during a sorption– desorption cycle that results in highly hindered diffusion or occlusion. Desorption resistance is usually attributed to the limitations of molecular diffusion through highly tortuous and sterically hindered pore networks or through highly viscous organic matter phases. The progress of diffusion through the mesopores of mineral aggregates may be retarded by the presence of microparticles of organic matter within the aggregates (Kleineidam et al. 1999a), which provide “way stations.” NOM phases may contain domains at the microscopic scale that are rigid, and thus present great, although

40


not insurmountable, barriers to diffusion. For highly hydrophobic compounds such as PAHs in stagnant soil columns, sorption equilibrium is not easily reached because the fluid-phase mass (representing the actual diffusant) is a very small fraction of the total mass per unit volume. Sorption intensity increases and diffusivity decreases as soil moisture content decreases. Retarded diffusion can explain the “aging effect,” in which bioavailability (uptake or degradation) decreases with precontact time of the chemical (Alexander 2000). The greater the degree of equilibrium reached in the precontact step, the less will leak out after contact with the organism is initiated. Because of the “random walk” nature of diffusion, incomplete pre-equilibration will always lead to molecules migrating both “inward” and “outward” of the particle after it contacts the organism. Thus, even for very short precontact times, some of the contaminant will end up becoming “bioresistant” to the observer. Diffusion can also explain the “rebound” effect that occurs when a contaminant is seemingly exhaustively removed by a remediation technique, and then redistributes over time, being placed once again in a more available state. Occlusion during particle synthesis is exemplified by PAHs in soot. Soot condenses after a complex series of gas-phase free-radical reactions, in which PAHs are among the intermediates (Akhter et al. 1985b; Lahaye 1990; Smedley et al. 1992). It is possible that some unreacted PAH molecules become trapped in internal pores of the nascent soot condensate—pores that later may have no connection to external fluids. The evidence for occlusion of PAHs during BC synthesis is only circumstantial; it rests mainly on the desorption-resistance of some fraction of PAH content under extraordinary conditions—for example, high temperatures (Harmon et al. 2001) or supercritical carbon dioxide extraction (Jonker et al. 2005)—and is also implied by the ability of organic solvents to facilitate extraction (Akhter et al. 1985a; Jonker and Koelmans 2002a). Native PAHs in sediment thought to be associated primarily with soot particles equilibrated poorly with an isotope-labeled PAH spike (Jonker and Koelmans 2002b). Alteration of the soil matrix leading to occlusion has been little investigated but holds some credence. Farrell et al. (1999) proposed that mineral precipitation leads to blockage of intra-granular micropores in silica, entrapping TCE and PCE. Some have suggested that inelastic swelling of NOM (Sander and Pignatello 2009; Weber et al. 2002) and BC (Braida et al. 2003) may lead to immobilization via an antiplasticization mechanism occurring during the desorption step. In such a case, an abrupt release of sorbed chemical causes the matrix to collapse and stiffen around some molecules before they have a chance to escape. Release in this case may involve cooperative flexing of the matrix of a nature that requires a high activation energy. This hypothesis has not been rigorously tested.

1.4. SUMMARY AND CONCLUDING REMARKS Whether one concludes that we know a little or a lot about sorption to natural organic substances depends on one’s perspective. This author believes we clearly know a lot about individual processes and interactions in well-focused experimental situations (although much remains to be learned about the details); but when it comes to actually predicting in situ sorption–desorption behevior from sorbate and sorbent properties, we are a long way off. This is poignantly illustrated by a recent study (Arp et al. 2009) that found that established models could predict in situ KOC values for apolar compounds in impacted sediments (PAHs, PCBs, PCDD/Fs, and chlorinated benzenes; 55 compounds, 473 data) to within a range of only three orders of magnitude (i.e., a factor of 30) at best! The two best models were based on coal tar as a sorption surrogate for sediment organic matter, not NOM itself; while one was a ppFER, the other was simple singleparameter Raoult’s law model (i.e., KOC inversely proportional to subcooled water solubility). Sorption of uncharged molecules from the aqueous phase to NOM and BC will be driven primarily by the hydrophobic effect, with noncovalent forces directly with the sorbent contributing. Suitable free-energy models to assign contributions of individual driving forces are in an early stage of development. Poly-parameter FERs are better than singleparameter FERs, but are difficult to interpret unambiguously in terms of driving force contributions. Dispersion forces seem to be balanced between water and the sorbent. Hydrogen-bond donating by the sorbate seems to have little effect on sorption intensity, while H-bond accepting apparently disfavors sorption. It is likely that p-p electron donor–acceptor interactions will be found to play an important role in sorption involving moieties with especially strong donor/ acceptor character—those with strong electron-withdrawing substituents (e.g., polynitro, charged aromatic and heteroaromatic amines) and highly polarizable donor ability (e.g., PAHs, graphene surfaces on BC). Steric effects play an important role in sorption to microporous materials such as BC; however, they have not yet been quantified. The importance of steric effects in partitioning to NOM is unknown but could be substantial for very large molecules that require a large cavity. Sorption intensity of ionic and ionizable organic compounds to both NOM and BC is still poorly understood and predictable. Nonlinearity in the xisotherm reflects the heterogeneity of the sorption process. Depending on the degree of nonlinearity, the sorption distribution coefficient can vary considerably— by as much as three orders of magnitude—over the range in concentration from infinite dilution to maximum solubility in water. The implications of nonlinearity for bioavailability and environmental fate are, therefore, quite substantial but have not been fully appreciated. The causes of nonlinear sorption in NOM are still being debated, but appear at this point to be due

REFERENCES

to changes in the chemical and/or physical nature of sorption sites or domains with loading, rather than a saturation limitation of any particular functional group interaction. Evidence for preferential sorption based on domains segragated on the basis of functional unit identity is mixed. The glassy polymer model of NOM provides a rationale for nonlinearity of even hydrocarbons in NOM in regard to its postulate of progressive filling of internal microporosity frozen into the matrix. Competitive sorption is simply the “flip side” of nonlinearity. A competitor will both suppress sorption and increase the isotherm linearity of the chemical of interest. The electronic and steric factors controlling the degree of overlap in the sorption domains of competing sorbates are poorly understood. Competition can give rise to artificial hysteresis if the competitor is diluted. Sorption to BC is typically more highly nonlinear than sorption to NOM. Nonlinearity in BC originates from both nonuniform pore sizes and chemical inhomogeneity of the surface. Adsorption to BC is inversely related to polar atom (mainly O) content, due to competition for adsorption space from water molecules. Environmental BC lies on a continuum between partially charred biomass and graphitized carbon depending on formation conditions, which confounds its quantification in environmental samples. The surface activity of BC can be attenuated by incompletely carbonized biomass, unburned fuel products, and natural substances—organic and inorganic—that adsorb during weathering in the environment. All of these factors make prediction of sorption intensity from the supposed composition of soil organic matter tenuous. The composition of organic matter in soils is best regarded as a continuum of materials ranging from fully flexible macromolecular substances that behave as true partition phases, to fully rigid graphitic substances with extended pore networks that behave analogously to fixed-pore adsorbents. Except for the latter extreme, these materials are dynamic in the sense that they can undergo swelling or shrinking during uptake or release of sorbates. The swelling process becomes inelastic as the ambient temperature approaches the glass transition temperature of the substance. The inelasticity of expansion results in an increase in the excess free volume, which persists, and leads to sorption hysteresis. Demonstration of pore deformation hysteresis in NOM not only represents an example of true hysteresis but also provides convincing evidence for the glassy character of at least some fraction of NOM in soil. Sorption–desorption rates play a critical role in the bioavailability and mobility of contaminants, yet little progress has been made in our ability to predict rates from soil and contaminant properties and existing conditions. Diffusion is sensitive to molecular diameter, possibly polarity, concentration, temperature, competing sorbates, the nature of the organic matter, and how the organic matter is distributed within soil particles and aggregates. At high concentrations,

41

diffusion may be sensitive to matrix changes accompanying the uptake or release of contaminant. Diffusion in soil materials appears not to be symmetric—that is, a single diffusivity is unable to describe the entire uptake or desorption profile. The quantities in the “fast,” “slow,” and “very slow” compartments have been the focus of many papers, but this distinction, while useful for some experimental purposes, is artificial and rarely allows prediction or translation to another system. A small fraction of the total chemical present may be so resistant to desorption that for all practical purposes it is physically and biologically unavailable. The causes of highly resistant desorption are poorly understood, and so predicting this fraction from soil/chemical properties is not yet possible. Sorption hysteresis and resistant desorption are the “elephants in the room” of sorption research— common problems with serious implications that nobody wants to tackle.

REFERENCES Accardi-Dey, A. and Gschwend, P. M. (2002), Assessing the combined roles of natural organic matter and black carbon as sorbents in sediments, Environ. Sci. Technol. 36, 21–29. Accardi-Dey, A. and Gschwend, P. M. (2003), Reinterpreting literature sorption data considering both absorption into organic carbon and adsorption onto black carbon, Environ. Sci. Technol. 37, 99–106. Adamson, A. W. and Gast, A. P. (1997), Physical Chemistry of Surfaces, 6th ed., Wiley, New York. Ahrens, L., Yamashita, N., Yeung, L. W. Y., Taniyasu, S., Horii, Y., Lam, P. K. S., and Ebinghaus, R. (2009), Partitioning behavior of per- and polyfluoroalkyl compounds between pore water and sediment in two sediment cores from Tokyo Bay, Japan, Environ. Sci. Technol. 43, 6969–6975. Aiken, G. R., McKnight, D. M., Wershaw, R. L., and MacCarthy, P., eds. (1985), Humic Substances in Soil, Sediment, and Water, Wiley, New York. Akhter, M., Chughtai, A, and Smith, D. (1985a), The structure of hexane soot I: Spectroscopic studies, Appl. Spectrosc. 39, 143–153. Akhter, M.S.,Chughtai, A. R, and Smith, D. M. (1985b), The structure ofhexanesootII:Extractionstudies,Appl.Spectrosc.39,154–167. Alexander, M. (1995), How toxic are toxic chemicals in soil? Environ. Sci. Technol. 29, 2713–2717. Alexander, M. (2000), Aging, bioavailability, and overestimation of risk from environmental pollutants, Environ. Sci. Technol. 34, 4259–4265. Allen-King, R. M., Grathwohl, P., and Ball, W. P. (2002), New modeling paradigms for the sorption of hydrophobic organic chemicals to heterogeneous carbonaceous matter in soils, sediments, and rocks, Adv. Water Resour. 25, 985–1016. Altfelder, S., Streck, T, and Richter, J. (2000), Nonsingular sorption of organic compounds in soil: The role of slow kinetics, J. Environ. Qual. 29, 917–925.

42


Arp, H. P. H., Breedveld, G. D, and Cornelissen, G. (2009), Estimating the in situ sediment–porewater distribution of PAHs and chlorinated aromatic hydrocarbons in anthropogenic impacted sediments, Environ. Sci. Technol. 43, 5576–5585. Bailey, A., Cadenhead, D. A., Davies,D. H., Everett, D. H., and Miles, A. J. (1971), Low pressure hysteresis in the adsorption of organic vapours by porous carbons, Trans. Faraday Soc. 67, 231–243. Ball, W. P. and Roberts, P. V. (1991a), Long-term sorption of halogenated organic chemicals by aquifer material. 1. Equilibrium, Environ. Sci. Technol. 25, 1223–1235. Ball, W. P. and Roberts, P. V. (1991b), Long-term sorption of halogenated organic chemicals by aquifer material. 2. Intraparticle diffusion, Environ. Sci. Technol. 25, 1237–1249. Barring, H., Bucheli, T. D., Broman, D., and Gustafsson, O. (2002), Soot-water distribution coefficients for polychlorinated dibenzop-dioxins, polychlorinated dibenzofurans and polybrominated diphenylethers determined with the soot cosolvency-column method, Chemosphere 49, 515–523. Bayard, R., Barna, L., Mahjoub, B., and Gourdon, R. (2000), Influence of the presence of PAHs and coal tar on naphthalene sorption in soils, J. Contam. Hydrol. 46, 61–80. Berens, A. (1989), Transport of organic vapors and liquids in poly (vinyl chloride), Makromol.Chem., Macromol.Symp.29, 95–108. Birdwell, J., Cook, R. L., and Thibodeaux, L. J. (2007), Desorption kinetics of hydrophobic organic chemicals from sediment to water: A review of data and models, Environ. Toxicol. Chem. 26, 424–434. Boehm, H. P. (1964), Some aspects of the surface-chemistry of carbon-blacks and other carbons, Carbon. 32, 759–769. Bontha, J. R. and Kaplan, D. I. (1999), Immobilization or recovery of chlorinated hydrocarbons from contaminated groundwater using clathrate hydrates: A proof-of-concept, Environ. Sci. Technol. 33, 1051–1056. Borisover, M. and Graber, E. R. (2003), Classifying NOM-organic sorbate interactions using compound transfer from an inert solvent to the hydrated sorbent, Environ. Sci. Technol. 37, 5657–5664. Braida, W. J., White, J. C., Ferrandino, F. J., and Pignatello, J. J. (2001), Effect of solute concentration on sorption of polyaromatic hydrocarbons in soil: Uptake rates, Environ. Sci. Technol. 35, 2765–2772. Braida, W., White, J. C., Zhao, D., Ferrandino, F. J., and Pignatello, J. J. (2002), Concentration-dependent kinetics of pollutant desorption from soils, Environ. Toxicol. Chem. 21, 2573–2580. Braida, W., Pignatello, J. J., Lu, Y., Ravikovitch, P. I., Neimark, A. V., and Xing, B. (2003), Sorption hysteresis of benzene in charcoal particles, Environ. Sci. Technol. 37, 409–417. Braida, W., White, J. L., and Pignatello, J. J. (2004), Indices for bioavailability and biotransformation potential of contaminants in soils, Environ. Toxicol. Chem. 23, 1585–1591. Breault, G. A., Hunter, C. A., and Mayers, P. C. (1998), Influence of solvent on aromatic interactions in metal tris-bipyridine complexes, J. Am. Chem. Soc. 120, 3402–3410. Brewer, C. E., Schmidt-Rohr, K., Satrio, J. A, and Brown, R. D. (2009), Characterization of biochar from fast pyrolysis and

gasification systems, Environ. Prog. Sustain. Energy 28, 386–396. Brunauer, S., Emmett, P. H., and Teller, E. (1938), Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60, 309–319. Brusseau, M. L. and Rao, P. S. C. (1989), Sorption nonideality during organic contaminant transport in porous media, Crit. Rev. Environ. Control. 19, 33–99. Bucheli, T. D. and Gustafsson, O. (2001), Ubiquitous observations of enhanced solid affinities for aromatic organochlorines in field situations: Are in situ dissolved exposures overestimated by existing partitioning models? Environ. Toxicol. Chem. 20, 1450–1456. Budd, P. M., McKeown, N. B., and Fritsch, D. (2005), Free volume and intrinsic microporosity in polymers, J. Mater. Chem. 15, 1977–1986. Carda-Broch, S. and Berthod, A. (2004), Countercurrent chromatography for the measurement of the hydrophobicity of sulfonamide amphoteric compounds, Chromatographia 59, 79–87. Carroll, K. M., Harkness, M. R., Bracco, A. A., and Balcarcel, R.R. (1994), Application of a permeant/polymer diffusional model to the desorption of polychlorinated biphenyls from hudson river sediments, Environ. Sci. Technol. 28, 253–258. Chan, K. Y. and Xu, Z. (2009), Biochar: Nutrient properties and their enhancement, in Biochar for Environmental Management, Lehmann, J. and Joseph, S., Earthscan, London, UK or Sterling, VA pp. 67–81. Chandler, D. (2005), Interfaces and the driving force of hydrophobic assembly, Nature 437, 640–647. Chang, M., Wu, S., and Chen, C. (1997), Diffusion of volatile organic compounds in pressed humic acid disks, Environ. Sci. Technol. 31, 2307–2312. Chatterjee, R., Laird, D. A., and Thompson, M. L. (2008), Interactions among K þ - Ca2 þ exchange, sorption of m-dinitrobenzene, and smectite quasicrystal dynamics, Environ. Sci. Technol. 42, 9099–9103. Chefetz, B., Deshmukh, A. P., Hatcher, P. G., and Guthrie, E. A. (2000), Pyrene sorption by natural organic matter, Environ. Sci. Technol. 34, 2925–2930. Chefetz, B. and Xing, B. (2009), Relative role of aliphatic and aromatic moieties as sorption domains for organic compounds: A review, Environ. Sci. Technol. 43, 1680–1688. Chen, B., Johnson, E. J., Chefetz, B., Zhu, L. and Xing, B. (2005), Sorption of polar and nonpolar aromatic organic contaminants by plant cuticular materials: Role of polarity and accessibility, Environ. Sci. Technol. 39, 6138–6146. Chen, B. and Xing, B. (2005), Sorption and conformational characteristics of reconstituted plant cuticular waxes on montmorillonite, Environ. Sci. Technol. 39, 8315–8323. Chen, G., Shan, X., Wang, Y., Wen, B., Pei, Z., Xie, Y., Liu, T., and Pignatello, J. J. (2009), Adsorption of 2,4,6-trichlorophenol by multi-walled carbon nanotubes as affected by Cu(II), Water Res. 43, 2409–2418. Chen, J., Zhu, D., and Sun, C. (2007a), Effect of heavy metals on the sorption of hydrophobic organic compounds to wood charcoal, Environ. Sci. Technol. 41, 2536–2541.

REFERENCES

Chen, W., Duan, L. and Zhu, D. Q. (2007b), Adsorption of polar and nonpolar organic chemicals to carbon nanotubes, Environ. Sci. Technol. 41, 8295–8300. Chien, Y. and Bleam, W. F. (1997), Fluorine-19 nuclear magnetic resonance study of atrazine in humic and sodium dodecyl sulfate micelles swollen by polar and nonpolar solvents, Langmuir 13, 5283–5288. Chiou, C. T. (1989), Theoretical considerations of the partition uptake of nonionic organic compounds by soil organic matter, in Reactions and Movement of Organic Chemicals in Soil, Sawhney, B. L. and Brown, K., eds., Soil Science Society of America (special publication, Madison, WI, pp. 1–29. Chiou, C. T. (2002), Partition and Adsorption of Organic Contaminants in Environmental Systems, Wiley, Hoboken, NJ. Chiou, C. T., McGroddy, S. E., and Kile, D. E. (1998), Partition characteristics of polycyclic aromatic hydrocarbons on soils and sediments, Environ. Sci. Technol. 32, 264–269. Cho, H. H., Smith, B. A., Wnuk, J. D., Fairbrother, D. H., and Ball, W. P. (2008), Influence of surface oxides on the adsorption of naphthalene onto multiwalled carbon nanotubes, Environ. Sci. Technol. 42, 2899–2905. Cockroft, S. L., Perkins, J., Zonta, C., Adams, H., Spey, S. E., Low, C. M. R., Vinter, J. G., Lawson, K. R., Urch, C. J, and Hunter, C. A. (2007), Substituent effects on aromatic stacking interactions, Org. Biomol. Chem. 5, 1062–1080. ¨ . (2004), Cornelissen, G., Elmquist, M., Groth, I., and Gustafson, O Effect of sorbate planarity on environmental black carbon sorption, Environ. Sci. Technol. 38, 3574–3580. ¨ . (2005), Importance of unburned Cornelissen, G. and Gustafson, O coal carbon, black carbon, and amorphous organic carbon to phenanthrene sorption in sediments, Environ. Sci. Technol. 39, 764–769. Cornelissen, G. and Gustafsson, O. (2004), Sorption of phenanthrene to environmental black carbon in sediment with and without organic matter and native sorbates, Environ. Sci. Technol. 38, 148–155. Cornelissen, G. and Gustafsson, O. (2006), Effects of added PAHs and precipitated humic acid coatings on phenanthrene sorption to environmental black carbon, Environ. Pollut. 141, 526–531. Cornelissen, G., Hassell, K. A., van Noort, P. C. M., Kraaij, R., van Ekeren, P. J., Dijkema, C., de Jager, P. A., and Govers, H. A. J. (2000), Slow desorption of PCBs and chlorobenzenes from soils and sediments: Relations with sorbent and sorbate characteristics, Environ. Pollut. 108, 69–80. Cornelissen, G., van Noort, P. C. M., and Govers, H. A. J. (1997), Desorption kinetics of chlorobenzenes, polycyclic aromatic hydrocarbons, and polychlorinated biphenyls: Sediment extraction with Tenax and effects of contact time and solute hydrophobicity, Environ. Toxicol. Chem. 16, 1351–1357. Crank, J. (1975), The Mathematics of Diffusion, 2nd ed., Clarendon Press, Oxford, UK. Crittenden, J. C., Luft, P., Hand, D. W., Oravitz, J. L., Loper, S. W., and Arl, M. (1985), Prediction of multicomponent adsorption equilibria using ideal adsorbed solution theory, Environ. Sci. Technol. 19, 1037–1043.

43

Curl, R. L. and Keolelan, G. A. (1984), Implicit-adsorbate model for apparent anomalies with organic adsorption on natural adsorbants, Environ. Sci. Technol. 18, 916–922. Dachs, J. and Eisenreich, S. J. (2000), Adsorption onto aerosol soot carbon dominates gas-particle partitioning of polycyclic aromatic hydrocarbons, Environ. Sci. Technol. 34, 3690–3697. Del Vecchio, R. and Blough, N. L. (2004), On the origin of the optical properties of humic substances, Environ. Sci. Technol. 38, 3885–3891. DeLapp, R. C. and Leboeuf, E. J. (2004), Thermal analysis of whole soils and sediment, J. Environ. Qual. 33, 330–337. Dixon, A. M., Mai, M. A, and Larive, C. K. (1999), NMR investigation of the interactions between 4’-fluoro-1’-acetonaphthone and the Suwannee River fulvic acid, Environ. Sci. Technol. 33, 958–964. Donnet, J.-B., Bansal, R. C., and Wang, M.-J. (1993), Carbon Black, 2nd ed., Marcel Dekker, New York. Dulfer, W. J. and Govers, H. A. J. (1995), Membrane-water partitioning of polychlorinated biphenyls in small unilamellar vesicles of four saturated phosphatidylcholines, Environ. Sci. Technol. 29, 2548–2554. Eisenberg, A. (1993), The glassy state and the glass transition, in Physical Properties of Polymers, Mark, J. E., Eisenberg, A., Graessley, W. W., Mandelkern, L., Samulski, E. T., Koenig, J. L., and Wignall, G. D. eds., American Chemical Society, Washington, DC, pp. 61–95. Endo, S., Grathwohl, P., Haderlein, S. B., and Schmidt, T. C. (2008a), Compound-specific factors influencing sorption nonlinearity in natural organic matter, Environ. Sci. Technol. 42, 5897–5903. Endo, S., Grathwohl, P., and Schmidt, T. C. (2008b), Absorption or adsorption? Insights from molecular probes n-alkanes and cycloalkanes into modes of sorption by environmental solid matrices, Environ. Sci. Technol. 42, 3989–3995. Endo, S., Grathwohl, P., Haderlein, S. B., and Schmidt, T. C. (2009a), Effects of native organic material and water on sorption properties of reference diesel soot, Environ. Sci. Technol. 43, 3187–3193. Endo, S., Grathwohl, P., Haderlein, S. B., and Schmidt, T. C. (2009b), LFERs for soil organic carbon-water distribution coefficients (Koc) at environmentally relevant sorbate concentration, Environ. Sci. Technol. 43, 3094–3100. Farrell, J., Grassian, D., and Jones, M. (1999), Investigation of mechanisms contributing to slow desorption of hydrophobic compounds from mineral solids, Environ. Sci. Technol. 33, 1237–1243. Farrell, J. and Reinhard, M. (1994), Desorption of halogenated organics from model solids, sediments, and soil under unsaturated conditions. 2. Kinetics, Environ. Sci. Technol. 28, 63–72. Ferguson, S. B. and Diederich, F. (1986), Electron donor-acceptor interactions in host-guest complexation in organic solutions, Angew. Chem. Int. Ed. Engl. 25, 1127–1129. Fleming, G. K. and Koros, W. J. (1990), Carbon dioxide conditioning effects on sorption and volume dilation behavior for biphenol a-polycarbonate, Macromolecules 23, 1353–1360.

44


Foster, R. (1969), Organic Charge-Transfer Complexes, Academic Press, London. Frank, H. S. and Evans, M. W. (1945), Free volume and entropy in condensed systems. .3. Entropy in binary liquid mixtures— partial molal entropy in dilute solutions—structure and thermodynamics in aqueous electrolytes, J. Chem. Phys. 13, 507–532. Fritz, J. S. (2005), Factors affecting selectivity in ion chromatography, J. Chromatogr. A (Proc. 17th Int. Ion Chromatography Symp. 1085, 8–17. Fritz, W. and Schlunder, E. U. (1981), Competitive adsorption of two dissolved organics onto activated carbon-I, Chem. Eng. Sci. 36, 721–730. Gilli, G. and Gilli, P. (2000), Towards an unified hydrogen-bond theory, J. Molec. Struct. 552, 1–15. Gilli, P., Pretto, L., Bertolasi, V., and Gilli, G. (2009), Predicting hydrogen-bond strengths from acid-base molecular properties. The pK(a) slide rule: Toward the solution of a long-lasting problem, Acc. Chem. Res. 42, 33–44. Glasser, W. G. and Kelley, S. S. (1987), Lignin, in Encyclopedia of Polymer Science and Technology, Kroschwitz, J., ed., Wiley, New York, pp. 795–852. Gobas, F. A. P. C., Lahittete, J. M., Garofalo, G., Shiu, W. Y., and Mackay, D. (1988), A novel method for measuring membranewater partition coefficients of hydrophobic organic chemicals: Comparison with 1-octanol-water partitioning, J. Pharm. Sci. 77, 265–272. Goldberg, E. D. (1985), Black Carbon in the Environment, Wiley, New York. Greenland, D. J. and Hayes, M. H. B. (1981), Soil processes, in The Chemistry of Soil Processes ed. Greenland, D. J., and Hayes, M. H. B., eds., Wiley, Chichester, UK, pp. 1–35. Grossman, A. and Ghosh, U. (2009), Measurement of activated carbon and other black carbons in sediments, Chemosphere 75, 469–475. Gu, C., Karthikeyan, K. G., Sibley, S. D., and Pedersen, J. A. (2007), Complexation of the antibiotic tetracycline with humic acid, Chemosphere 66, 1494–1501. Gunasekara, A. S., Simpson, M. J., and Xing, B. (2003), Identification and characterization of sorption domains in soil organic matter using structurally modified humic acids, Environ. Sci. Technol. 37, 852–858. Gung, B. W. and Amicangelo, J. C. (2006), Substituent effects in C6F6-C6H5X stacking interactions, J. Org. Chem. 71, 9261–9270. ¨ . and Gschwend, P. M. (1997), Soot as a strong Gustafsson, O partition medium for polycyclic aromatic hydrocarbons in aquatic systems, in Molecular Markers in Environmental Geochemistry, Eganhouse, R. P., ed., American Chemical Society, Washington, DC, pp. 365–381. ¨ ., Haghseta, F., Chan, C., MacFarlane, J., and Gustafsson, O Gschwend, P. M. (1997), Quantification of the dilute sedimentary soot phase: Implications for PAH speciation and bioavailability. Environ. Sci. Technol. 31, 203–209. Hammes, K., Schmidt, M. W. I., Smernik, R. J., Currie, L. A., Ball, W. P., Nguyen, T. H., Louchouarn, P., Houel, S., Gustafsson, O.,

Elmquist, M., Cornelissen, G., Skjemstad, J. O., Masiello, C. A., Song, J., Peng, P., Mitra, S., Dunn, J. C., Hatcher, P. G., Hockaday, W. C., Smith, D. M., Hartkopf-Froeder, C., Boehmer, A., Luer, B., Huebert, B. J., Amelung, W., Brodowski, S., Huang, L., Zhang, W., Gschwend, P. M., Flores-Cervantes, D. X., Largeau, C., Rouzaud, J. N., Rumpel, C., Guggenberger, G., Kaiser, K., Rodionov, A., Gonzalez-Vila, F. J., Gonzalez-Perez, J. A., de la Rosa, J. M., Manning, D. A. C., Lopez-Capel, E., and Ding, L. (2007), Comparison of quantification methods to measure firederived (black/elemental) carbon in soils and sediments using reference materials from soil, water, sediment and the atmosphere, Global Biogeochem. Cycles 21. Hammes, K., Smernik, R. J., Skjemstad, J. O., and Schmidt, M. W. I. (2008), Characterisation and evaluation of reference materials for black carbon analysis using elemental composition, colour, BET surface area and C-13 NMR spectroscopy, J. Appl. Geochem. 23, 2113–2122. Harmon, T. C., Burks, G. A., Aycaguer, A.-C., and Jackson, K. (2001), Thermally enhanced vapor extraction for removing PAHs from lampblack-contaminated soil, J. Environ. Eng. 127, 986–993. Harris, P. J. F. and Tsang, S. C. (1997), High-resolution electron microscopy studies of non-graphitizing carbons, Philos. Mag. A 76, 667–677. Haws, N. W., Ball, W. P., and Bouwer, E. J. (2006), Modeling and interpreting bioavailability of organic contaminant mixtures in subsurface environments, J. Contam. Hydrol. 82, 255–292. Hawthorne, S. B., Grabanski, C. B., and Miller, D. J. (2007), Measured partition coefficients for parent and alkyl polycyclic aromatic hydrocarbons in 114 historically contaminated sediments: Part 2. Testing the k(oc)k(bc) two carbon-type model, Environ. Toxicol. Chem. 26, 2505–2516. Hayes, M. H. B. and Clapp, C. E. (2001), Humic substances: Considerations of compositions, aspects of structure, and environmental influences, Soil Sci. 166, 723–737. Hayes, M. H. B., MacCarthy, P., Malcolm, R. L.and Swift, R. S., eds. (1989), Humic Substances II. In Search of Structure. Wiley, Chichester, UK. Higgins, C. P. and Luthy, R. G. (2006), Sorption of perfluorinated surfactants on sediments., Environ. Sci. Technol. 40, 7251–7256. Hong, L., Jean, Y. C., Yang, H., Jordan, S. S., and Koros, W. J. (1996), Free-volume hole properties of gas-exposed polycarbonate studied by positron annihilation lifetime spectroscopy, Macromolecules 29, 7859–7864. Hu, W.-G., Mao, J., Xing, B., and Schmidt-Rohr (2000), Poly (methylene) crystallites in humic substances detected by nuclear magnetic resonance, Environ. Sci. Technol. 34, 530–534. Huang, W., Young, T., Schlautman, M. A., Yu, H., and Weber, W. J., Jr. (1997), A distributed reactivity model for sorption by soils and sediments. 9. General isotherm nonlinearity and applicability of the dual reactive domain model, Environ. Sci. Technol. 31, 1703–1710. Hummer, G., Garde, S., Garcia, A. E., Paulaitis, M. E., and Pratt, L. R. (1998), Hydrophobic effects on a molecular scale; doi:10.1021/jp982873 þ , J. Phys. Chem. B 102, 10469–10482.

REFERENCES

Hunter, C. A. (2004), Quantifying intermolecular interactions: Guidelines for the molecular recognition toolbox, Angew. Chem. Int. Ed. Engl. 43, 5310–5324. Hunter, C. A., Lawson, K. R., Perkins, J., and Urch, C. J. (2001), Aromatic interactions, J. Chem. Soc. Perkin Trans. 2, 651–669. Hunter, C. A. and Sanders, J. K. M. (1990), The nature of pi-pi interactions, J. Am. Chem. Soc. 112, 5525–5534. Iimori, T., Aoki, Y., and Ohshima, Y. (2002). S[sub 1]--S[sub 0] vibronic spectra of benzene clusters revisited. II. The trimer, J. Chem. Phys. 117, 3675–3686. Israelachvili, J. N. (1992), Intermolecular and Surface Forces, 2nd ed., Academic Press, London. Janiak, C. (2000), A critical account on p-p stacking in metal complexes with aromatic nitrogen-containing ligands, J. Chem. Soc. Dalton Trans. 3885–3896. Jeffrey, G. A. (1984), Inclusion Compounds, Academic Press, London. Johnson, M., KeinathII, T., and Weber, J. W. (2001), A distributed reactivity model from sorption by soils and sediments. 14. Characterization and modeling of phenanthrene desorption rates, Environ. Sci. Technol. 35, 1688–1695. Jonker, M. and Smedes, F. (2000), Preferential sorption of planar contaminants in sediments from Lake Ketelmeer, The Netherlands, Environ. Sci. Technol. 34, 1620–1626. Jonker, M. T. and Koelmans, A. A. (2002a), Extraction of polycyclic aromatic hydrocarbons from soot and sediment: Solvent evaluation and implications for sorption mechanism, Environ. Sci. Technol. 36, 4107–4113. Jonker, M. T. O. and Koelmans, A. A. (2002b), Sorption of polycyclic aromatic hydrocarbons and polychlorinated biphenyls to soot and soot-like materials in the aqueous environment: Mechanistic considerations, Environ. Sci. Technol. 36, 3725–3734. Jonker, M. T. O. and Barendregt, A. (2006), Oil is a sedimentary supersorbent for polychlorinated biphenyls, Environ. Sci. Technol. 40, 3829–3835. Jonker, M. T. O., Hawthorne, S. B., and Koelmans, A. A. (2005), Extremely slowly desorbing polycyclic aromatic hydrocarbons from soot and soot-like materials: evidence by supercritical fluid extraction, Environ. Sci. Technol. 39, 7885–7895. Jonker, M. T. O., Hoenderboom, A. M., and Koelmans, A. A. (2004), Effects of sedimentary sootlike materials on bioaccumulation and sorption of polychlorinated biphenyls, Environ. Toxicol. Chem. 23, 2563–2570. Kahle, M. and Stamm, C. (2007), Sorption of the veterinary antimicrobial sulfathiazole to organic materials of different origin. Environ. Sci. Technol. 41, 132–138. Kamens, R. M., Odum, J. R., and Fan, Z.-H. (1995), Some observations on times to equilibrium for semivolatile polycyclic aromatic hydrocarbons, Environ. Sci. Technol. 29, 43–50. Kamiya, Y., Bourbon, D., Mizoguchi, K., and Naito, Y. (1992), Sorption, dilation, and isothermal glass transition of poly(ethyl methacrylate)-organic gas systems, Polymer 24, 443–449. Kamiya, Y., Hirose, T., Mizoguchi, K., and Naito, Y. (1986), Gravimetric study of high-pressure sorption of gases in polymers, J. Polym. Sci., Part B: Polym. Phys. 24, 1525–1539.

45

Kamiya, Y., Mizoguchi, K., Terada, K., Fujiwara, Y., and Wang, J.-S. (1998), CO2 sorption and dilation of poly(methyl methacrylate), Macromolecules 31, 472–478. Kan, A. T., Fu, G., Hunter, M., Chen, W., Ward, C. H., and Tomson, M. B. (1998), Irreversible sorption of neutral hydrocarbons to sediments: Experimental observations and model predictions, Environ. Sci. Technol. 32, 892–902. Kan, A. T., Fu, G., Hunter, M. A., and Tomson, M. B. (1997), Irreversible adsorption of naphthalene and tetrachlorobiphenyl to lula and surrogate sediments, Environ. Sci. Technol. 31, 2176–2186. Kaneko, K., Yamaguchi, K., Ishii, C., Ozeki, S., Hagiwara, S., and Suzuki, T. (1991), Size evaluation of graphitic crystallites in activated carbon fibers from diamagnetic susceptibility measurements, Chem. Phys. Lett. 176, 75–78. Karapanagioti, H., Childs, J., and Sabatini, D. A. (2001), Impacts of heterogeneous organic matter on phenanthrene sorption: Different soil and sediment samples, Environ. Sci. Technol. 35, 4684–4690. Karapanagioti, H. K., Kleineidam, S., Sabatini, D. A., Grathwohl, P., and Ligouis, B. (2000), Impacts of heterogeneous organic matter on phenanthrene sorption: Equilibrium and kinetic studies with aquifer material, Environ. Sci. Technol. 34, 406–414. K€arger, J. and Ruthven, D. M. (1992), Diffusion in Zeolites and Other Microporous Solids, Wiley, New York. Khalil, M. F., Ghosh, U., and Kreitinger, J. P. (2006), Role of weathered coal tar pitch in the partitioning of polycyclic aromatic hydrocarbons in manufactured gas plant site sediments, Environ. Sci. Technol. 40, 5681–5687. Kilduff, J. E. and Wigton, A. (1999), Sorption of TCE by humicpreloaded activated carbon: Incorporating size-exclusion and pore blockage phenomena in a competitive adsorption model, Environ. Sci. Technol. 33, 250–256. Kile, D. E., Wershaw, R. L., and Chiou, C. T. (1999), Correlation of soil and sediment organic matter polarity to aqueous sorption of nonionic compounds, Environ. Sci. Technol. 33, 2053–2056. Kleineidam, S., Rugner, H., and Grathwohl, P. (1999a), Influence of petrographic composition/organic matter distribution of fluvial aquifer sediments on the sorption of hydrophobic contaminants, Sediment. Geol. 129, 311–325. Kleineidam, S., R€ ugner, H, and Grathwohl, P. (1999b), The impact of grain scale heterogeneity on slow sorption kinetics, Environ. Toxicol. Chem. 18, 1673–1678. Knicker, H., Muffler, P., and Hilscher, A. (2007), How useful is chemical oxidation with dichromate for the determination of “black carbon” in fire-affected soils? Geoderma 142, 178–196. Koelmans, A. A., Meulman, B., Meijer, T., and Jonker, M. T. O. (2009), Attenuation of polychlorinated biphenyl sorption to charcoal by humic acids, Environ. Sci. Technol. 43, 736–742. Kukkonen, J. V. K., Landrum, P. F., Mitra, S., Gossiaux, D. C., Gunnarsson, J., and Weston, D. (2003), Sediment characteristics affecting desorption kinetics of select PAH and PCB congeners for seven laboratory spiked sediments, Environ. Sci. Technol. 37, 4656–4663.

46


Kwon, J. H., Liljestrand, H. M., and Katz, L. E. (2006), Partitioning of moderately hydrophobic endocrine disruptors between water and synthetic membrane vesicles, Environ. Toxicol. Chem. 25, 1984–1992. Kwon, S. and Pignatello, J. J. (2005), Effect of natural organic substances on the surface and adsorptive properties of environmental black carbon (char): Pseudo pore blockage by model lipid components and its implications for N2-probed surface properties of natural sorbents, Environ. Sci. Technol. 39, 7932–7939. Lahaye, J. (1990), Mechanisms of soot formation, Polym. Degrad. Stab. 30, 111–121. Laor, Y. and Rebhun, M. (2002), Evidence for nonlinear binding of PAHs to dissolved humic acids, Environ. Sci. Technol. 36, 955–961. Lattao, C., Birdwell, J., Wang, J. J., and Cook, R. L. (2008), Studying organic matter molecular assemblage within a whole organic soil by nuclear magnetic resonance, J. Environ. Qual. 37, 1501–1509. Lazaridis, T. (2001), Solvent size vs cohesive energy as the origin of hydrophobicity, Acc. Chem. Res. 34, 931–937. LeBoeuf, E. J. and Weber, W. J., Jr. (1997), A distributed reactivity model for sorption by soils and sediments. 8. Sorbent organic domains: Discovery of a humic acid glass transition and an argument for a polymer-based model, Environ. Sci. Technol. 31, 1697–1702. Leenheer, J. A. (2009), Systematic approaches to comprehensive analysis of natural organic matter, Ann. Environ. Sci. 3, 1–130. Lei, L., Suidan, M. T., Khodadoust, A. P., and Tabak, H. H. (2004), Assessing the bioavailability of PAHs in field-contaminated sediment using XAD-2 assisted desorption, Environ. Sci. Technol. 38, 1786–1793. Li, J., Pignatello, J. J., Smets, B. F., Grasso, D., and Monserrate, E. (2005), Bench-scale evaluation of in situ bioremediation strategies for soil at a former manufactured gas plant site, Environ. Toxicol. Chem. 24, 741–749. Li, Q., Snoeyink, V. L., Mari~aas, B. J., and Campos, C. (2003), Elucidating competitive adsorption mechanisms of atrazine and NOM using model compounds, Water Resour. Res. 37, 773–784. Lighty, S. J., Veranth, J. M., and Sarofim, A. F. (2000), Combustion aerosols: Factors governing their size and composition and implications to human health, J. Air Waste Manage. Assoc. 50, 1565–1618. Loehr, R. C. and Webster, M. T. (1997), Environmentally Acceptable Endpoints in Soil, American Academy of Environmental Engineers, Annapolis, MD. Loke, M. L., Tjornelund, J., and Halling-Sorensen, B. (2002), Determination of the distribution coefficient (log K-d) of oxytetracycline, tylosin A, olaquindox and metronidazole in manure, Chemosphere 48, 351–361. Lu, Y. and Pignatello, J. J. (2002), Demonstration of the “conditioning effect” in soil organic matter in support of a pore deformation mechanism for sorption hysteresis, Environ. Sci. Technol. 36, 4553–4561. Lu, Y. and Pignatello, J. J. (2004a), History-dependent sorption in humic acids and a lignite in the context of a polymer model

for natural organic matter, Environ. Sci. Technol. 38, 5853– 5862. Lu, Y. and Pignatello, J. J. (2004b), Sorption of apolar aromatic compounds to soil humic acid particles affected by aluminum (III) ion cross-linking, J. Environ. Qual. 33, 1314–1321. Lucht, L. and Peppas, N. (1987), Macromolecular structure of coals. 2. Molecular weight between crosslinks from pyridine swelling experiments, Fuel 66, 803–809. Lucht, L. M., Larson, J. M., and Peppas, N. A. (1987), Macromolecular structure of coals.: IX. Molecular structure and glass transition temperature, Energy Fuels 1, 56–58. Luthy, R. G., Aiken, G. R., Brusseau, M. L., Cunningham, S. D., Gschwend, P. M., Pignatello, J. J., Reinhard, M., Traina, S. J., Weber, W. J., Jr., and Westall, J. C. (1997), Sequestration of hydrophobic organic contaminants by geosorbents, Environ. Sci. Technol. 31, 3341–3347. Lyon, W. G. (1995), Swelling of peats in liquid methyl, tetramethylene and propyl sulfoxides and in liquid propyl sulfone, Environ. Toxicol. Chem. 14, 229–236. MacKay, A. A. and Canterbury, B. (2005). Oxytetracycline sorption to organic matter by metal-bridging, J. Environ. Qual. 34, 1964–1971. Manes, M. (1998), Activated carbon adsorption fundamentals, in Encyclopedia of Environmental Analysis and Remediation, Meyers, R. A., ed., Wiley, New York, pp. 26–68. Mao, J.-D., Hundal, L. S., Thompson, M. L., and Schmidt-Rohr, K. (2002), Correlation of poly(methylene)-rich amorphous aliphatic domains in humic substances with sorption of a nonpolar organic contaminant, phenanthrene, Environ. Sci. Technol. 36, 929–936. Mao, J.-D. and Schmidt-Rohr, K. (2006), Absence of mobile carbohydrate domains in dry humic substances proven by NMR, and implications for organic-contaminant sorption models, Environ. Sci. Technol. 40, 1751–1756. Martin, D. S. (2003), The adsorption of aromatic acids onto the graphite basal surface, Surf. Sci. 536, 15–23. Masiello, C. A. and Druffel, E. R. M. (1998), Black carbon in deepsea sediments, Science 280, 1911–1913. McDermott, M. T. and McCreery, R. L. (1994), Scanning tunneling microscopy of ordered graphite and glassy carbon surfaces: Electronic control of quinone adsorption, Langmuir 10, 4307–4314. McGinley, P. M., Katz, L. E., and Weber, W. J. (1989), Multi-Solute Effects in the Sorption of Hydrophobic Organic Compounds by Aquifer Solutions, Presented before American Chemical Society, pp. 146–149. Merriam-Webster (2000), Unabridged Dictionary, Vol. 2.5. ed. Meyer, E. A., Castellano, R. K., and Diederich, F. (2003), Interactions with aromatic rings in chemical and biological recognition, Angew. Chem. Int. Ed. 42, 1210–1250. Milewska-Duda, J. (1993), The coal-sorbate system in the light of the theory of polymer solutions, Fuel 72, 419–425. Moore, F. G. and Richmond, G. L. (2008), Integration or segregation: How do molecules behave at oil/water interfaces? Acc. Chem. Res. 41, 739–748. Morimoto, T., Uno, H., and Furuta, H. (2007), Benzene ring trimer interactions modulate supramolecular structures, Angew. Chem. Int. Ed. 46, 3672–3675.

REFERENCES

M€ uller, E. A. and Gubbins, K. E. (1998), Molecular simulation study of hydrophilic and hydrophobic behavior of activated carbon surfaces, Carbon 36, 1433–1438. M€uller, E. A., Hung, F. R., and Gubbins, K. E. (2000), Adsorption of water vapor-methane mixtures on activated carbons, Langmuir 16, 5418–5424. M€uller-Wegener, U. (1987), Electron donor acceptor complexes between organic nitrogen heterocycles and humic acid, Sci. Total Environ. 62, 297–304. Newcomb, L. F. (1994), Aromatic stacking interactions in aqueous solution: Evidence that neither classical hydrophobic effects nor dispersion forces are important, J. Am. Chem. Soc. 116, 4993–4994. Newcombe, G., Drikas, M., and Hayes, R. (1997), Influence of characterised natural organic material on activated carbon adsorption: II. Effect on pore volume distribution and adsorption of 2-methylisoborneol, Water. Resour. Res. 31, 1065–1073. Nguyen, T. H. and Ball, W. P. (2006), Absorption and adsorption of hydrophobic organic contaminants to diesel and hexane soot, Environ. Sci. Technol. 40, 2958–2964. Nguyen, T. H., Brown, R. A., and Ball, W. P. (2004), An evaluation of thermal resistance as a measure of black carbon content in diesel soot, wood char, and sediment, Org. Geochem. 35, 217–234. Nguyen, T. H., Cho, H.-H., Poster, D. L., and Ball, W. P. (2007), Evidence for a pore-filling mechanism in the adsorption of aromatic hydrocarbons to a natural wood char, Environ. Sci. Technol. 41, 1212–1217. Nguyen, T. H., Goss, K.-U., and Ball, W. P. (2005), Polyparameter linear free energy relationships for estimating the equilibrium partition of organic compounds between water and the natural organic matter in soils and sediments, Environ. Sci. Technol. 39, 913–924. Niederer, C., Goss, K.-U. and Schwarzenbach, R. P. (2006a), Sorption equilibrium of a wide spectrum of organic vapors in leonardite humic acid: experimental setup and experimental data, Environ. Sci. Technol. 40, 5368–5373. Niederer, C., Goss, K.-U., and Schwarzenbach, R. P. (2006b), Sorption equilibrium of a wide spectrum of organic vapors in leonardite humic acid: Modeling of experimental data, Environ. Sci. Technol. 40, 5374–5379. ´ . B., Rainey, l. C., Feldermann, C. J., Sarofim, A. F., and Palotas, A Vander Sande, J. B. (1996), Soot morphology: An application of image analysis in high-resolution transmission electron microscopy, Microsc. Res. Tech. 33, 266–278. Pan, B., Ghosh, S., and Xing, B. (2007), Nonideal binding between dissolved humic acids and polyaromatic hydrocarbons, Environ. Sci. Technol. 41, 6472–6478. Perry, M., Carra, C., Chretien, M. N., and Scaiano, J. C. (2007), Effect of hexafluorobenzene on the photophysics of pyrene, J. Phys. Chem. A 111, 4884–4889. Piatt, J. J. and Brusseau, M. L. (1998), Rate-limited sorption of hydrophobic organic compounds by soils with well-characterized organic matter, Environ. Sci. Technol. 32, 1604–1608. Pignatello, J. J. (1990a), Slowly reversible sorption of aliphatic halocarbons in soils. I. Formation of residual fractions, Environ. Toxicol. Chem. 9, 1107–1115.

47

Pignatello, J. J. (1990b), Slowly reversible sorption of aliphatic halocarbons in soils. II. Mechanistic aspects, Environ. Toxicol. Chem. 9, 1117–1126. Pignatello, J. J. (1991), Competitive effects in the sorption of nonpolar organic compounds by soils, in Organic Substances and Sediments in Water, Baker, R. A., ed., Vol. 1, Humics and Soils, Lewis Publishers, Chelsea, MI, pp. 291–307. Pignatello, J. J. (2000), The measurement and interpretation of sorption and desorption rates for organic compounds in soil media, in Advances in Agronomy, Sparks, D. L., ed., Vol. 69, Academic Press, San Diego, pp. 1–73. Pignatello, J. J. (2006), Fundamental issues in sorption related to physical and biological remediation of soils, in Soil and Water Pollution Monitoring, Protection and Remediation, Springer, New York, pp. 3–23. Pignatello, J. J. (2009), Bioavailability of contaminants in soil, in Advances in Applied Bioremediation, Soil Biology, Singh, A., ed., Springer-Verlag, Berlin Heidelberg, Vol. 17, Chap. 2, pp. 35–71. Pignatello, J. J., Ferrandino, F. J., and Huang, L. Q. (1993), Elution of aged and freshly added herbicides from a soil, Environ. Sci. Technol. 27, 1663–1671. Pignatello, J. J., Frink, C. R., Marin, P. A., and Droste, E. X. (1990), Field-observed ethylene dibromide in an aquifer after two decades, J. Contam. Hydrol. 5, 195–214. Pignatello, J. J., Kwon, S., and Lu, Y. (2006a), Effect of natural organic substances on the surface and adsorptive properties of environmental black carbon (char): Attenuation of surface activity by humic and fulvic acids, Environ. Sci. Technol. 40, 7757–7763. Pignatello, J. J., Lu, Y., LeBoeuf, E. J., Huang, W., Song, J., and Xing, B. (2006b), Nonlinear and competitive sorption of apolar compounds in black carbon-free natural organic materials, J. Environ. Qual. 35. 1049–1059. Pignatello, J. J. and Xing, B. (1996), Mechanisms of slow sorption of organic chemicals to natural particles, Environ. Sci. Technol. 30, 1–11. Polubesova, T., Sherman-Nakache, M., and Chefetz, B. (2007), Binding of pyrene to hydrophobic fractions of dissolved organic matter: Effect of polyvalent metal complexation, Environ. Sci. Technol. 41, 5389–5394. Poole, S. K. and Poole, C. F. (1999), Chromatographic models for the sorption of neutral organic compounds by soil from water and air, J. Chromatogr. A 845, 381–400. Qiu, Y., Xiao, X., Cheng, H., Zhou, Z., and Sheng, G. D. (2009), Influence of environmental factors on pesticide adsorption by black carbon: pH and model dissolved organic matter, Environ. Sci. Technol. 43, 4973–4978. Qu, X. L., Liu, P., and Zhu, D. Q. (2008), Enhanced sorption of polycyclic aromatic hydrocarbons to tetra-alkyl ammonium modified smectites via cation-pi interactions, Environ. Sci. Technol. 42, 1109–1116. Radke, C. J. and Prausnitz, J. M. (1972), Thermodynamics of multisolute adsorption from dilute liquid solutions, Am. Inst. Chem. Eng. J. 18, 761–768. Ramanathan, V. and Carmichael, G. (2008), Global and regional climate changes due to black carbon, 1, 221–227.

48


Ran, Y., Sun, K., Yang, Y., Xing, B. S, and Zeng, E. (2007), Strong sorption of phenanthrene by condensed organic matter in soils and sediments, Environ. Sci. Technol. 41, 3952–3958. Razouk, R., Saleeb, E., and Said, E. (1968), The heat of wetting and immersional swelling of charcoal, J. Colloid Interface Sci. 28, 487–492. Rogers, C. E. (1965), Solubility and diffusivity, in Physics and Chemistry of the Organic Solid State, Fox, D., Labes, M. M., and Weissberger, A., eds., Interscience Publishers, New York, Vol. II, pp. 509–635. Rouquerol, F., Rouquerol, J., and Sing, K. (1999), Adsorption by Powders and Porous Solids, Academic Press, San Diego. Rutherford, D. W. and Chiou, C. T. (1992), Effect of water saturation in soil organic matter on the partition of organic compounds, Environ. Sci. Technol. 26, 965–970. Sander, M., Lu, Y., and Pignatello, J. J. (2005), A thermodynamically based method to quantify sorption hysteresis, J. Environ. Qual. 34, 1063–1072. Sander, M., Lu, Y, and Pignatello, J. J. (2006), Conditioning annealing studies of natural organic matter solids linking irreversible sorption to irreversible structural expansion, Environ. Sci. Technol. 40, 170–178. Sander, M. and Pignatello, J. J. (2005a), Charcterization of charcoal adsorption sites for aromatic compounds: insights drawn from single-solute and bi-solute competitive experiments, Environ. Sci. Technol. 39, 1606–1615. Sander, M. and Pignatello, J. J. (2005b), An isotope exchange technique to assess mechanisms of sorption hysteresis applied to naphthalene in kerogenous organic matter, Environ. Sci. Technol. 39, 7476–7484. Sander, M. and Pignatello, J. J. (2007), On the reversibility of sorption to black carbon: Distinguishing true hysteresis from artificial hysteresis caused by dilution of a competing adsorbate, Environ. Sci. Technol. 41, 843–849. Sander, M. and Pignatello, J. J. (2009), Sorption irreversibility of 1,4-dichlorobenzene in two natural organic matter-rich geosorbents, Environ. Toxicol. Chem. 28, 447–457. Sassman, S. A. and Lee, L. S. (2005), Sorption of three tetracyclines by several soils: Assessing the role of pH and cation exchange, Environ. Sci. Technol. 39, 7452–7459. Schaefer, C., Schuth, C., Werth, C., and Reinhard, M. (2000), Binary desorption isotherms of TCE and PCE from silicia gel and natural solids, Environ. Sci. Technol. 34, 4341– 4347. Schaumann, G. E. and Antelmann, O. (2000), Thermal characteristics of soil organic matter measured by DSC: A hint on a glass transition, J. Plant Nutr. Soil Sci. 163, 179–181. Schaumann, G. E., Hobley, E., Hurrass, J., and Rotard, W. (2005), H-NMR relaxometry to monitor wetting and swelling kinetics in high-organic matter soils, Plant Soil 275, 1–20. Schaumann, G. E. and LeBoeuf, E. J. (2005), Glass transitions in peat: Their relevance and the impact of water, Environ. Sci. Technol. 39, 800–806. Schellenberg, K., Leuenberger, C., and Schwarzenbach, R. P. (1984), Sorption of chlorinated phenols by natural sediments

and aquifer materials; doi:10.1021/es00127a005, Environ. Sci. Technol. 18, 652–657. Schmidt, M. W. I. and Noack, A. G. (2000), Black carbon in soils and sediments: Analysis, distribution, implications, and current challenges, Global Biogeochem. Cycles 14, 777–793. Schmidt-Mende, L., Fechtenkotter, A., Mullen, K., Moons, E., Friend, R. H., and MacKenzie, J. D. (2001), Self-organized discotic liquid crystals for high-efficiency organic photovoltaics, Science 293, 1119–1122. Schnitzer, M. and Khan, S. U. (1978), Soil Organic Matter, Elsevier, New York. Schulten, H. -R. and Schnitzer, M. (1997), Chemical model structures for soil organic matter and soils, Soil Sci. 162, 115–130. Schwarzenbach, R. P., Gschwend, P. M., and Imboden, D. M. (2002), Environmental Organic Chemistry, 2nd ed., Wiley, New York. Senesi, N., D’Orazio, V., and Miano, T. M. (1995), Adsorption mechanisms of s-triazine and bipyridylium herbicides on humic acids from hop field soils, Geoderma 66, 273–283. Sheindorf, C., Rebhun, M., and Sheintuch, M. (1981), A Freundlichtype multicomponent isotherm, J. Colloid Interface Sci. 79, 136–142. Shibuya, M., Kato, M., Ozawa, M., Fang, P. H., and Osawa, E. (1999), Detection of buckminsterfullerene in usual soots and commercial charcoals, Fullerene Sci. Technol. 7, 181–193. Shih, Y. and Gschwend, P. M. (2009), Evaluating activated carbonwater sorption coefficients of organic compounds using a linear solvation energy relationship approach and sorbate chemical activities, Environ. Sci. Technol. 43, 851–857. Shor, L. M., Rockne, K. J., Taghon, G. L., Young, L. Y., and Kosson, D. S. (2003), Desorption kinetics for field-aged polycyclic aromatic hydrocarbons from sediments, Environ. Sci. Technol. 37, 1535–1544. Silverstein, K. A. T., Haymet, A. D. J., and Dill, K. A. (2000), The strength of hydrogen bonds in liquid water and around nonpolar solutes, J. Am. Chem. Soc. 122, 8037–8041. Sims, G. K. and O’Loughlin, E. J. (1989), Degradation of pyridines in the environment, Crit. Rev. Environ. Control 19, 309–340. Sinnokrot, M. O. and Sherrill, C. D. (2006), High-accuracy quantum mechanical studies of pi-pi interactions in benzene dimers, J. Phys. Chem. A 110, 10656–10668. Skjemstad, J. O., Taylor, J. A., and Smernik, R. J. (1999), Estimation of charcoal (char) in soils, Commun. Soil Sci. Plant Anal. 30, 2283–2298. Smedley, J. M., Williams, A., and Bartle, K. D. (1992), A mechanism for the formation of soot particles and soot deposits, Combust. Flame 91, 71–82. Smernik, R. J., Kookana, R. S., and Skjemstad, J. O. (2006), NMR characterization of 13-C-benzene sorbed to natural and prepared charcoals, Environ. Sci. Technol. 40, 1764–1769. Southall, N. T., Dill, K. A., and Haymet, A. D. J. (2002), A view of the hydrophobic effect, J. Phys. Chem. B 106, 521–533. Sposito, G., Martin-Neto, L., and Yang, A. (1996), Atrazine complexation of soil humic acids, J. Environ. Qual. 25, 1203–1209.

REFERENCES

SRC Syracuse Research Corporation Interactive LogKow (KowWin) Demo; available at http://www.srcinc.com/what-we-do/ databaseforms.aspx?id¼385; accessed 4/14/09. Steed, J. M., Dixon, T. A., and Klemperer, W. (1979), Molecularbeam studies of benzene dimer, hexafluorobenzene dimer, and benzene-hexafluorobenzene, J. Chem. Phys. 70, 4940– 4946. Steinberg, S. M., Pignatello, J. J, and Sawhney, B. L. (1987), Persistence of 1,2-dibromoethane in soils: Entrapment in intraparticle micropores, Environ. Sci. Technol. 21, 1201–1208. Stevenson, F. (1994), Humus Chemistry: Genesis, Composition, Reactions, 2nd ed., Wiley, New York. Strommen, M. R. and Kamens, R. M. (1997), Development and application of a dual-impedance radial diffusion model to simulate the partitioning of semivolatile organic compounds in combustion aerosols, Environ. Sci. Technol. 31, 2983–2990. Sutton, R. and Sposito, G. (2005), Molecular structure in soil humic substances: The new view, Environ. Sci. Technol. 39, 9009–9015. Teixido´ Planes, M. and Pignatello, J. J. (unpublished), Sorption of the antibiotic sulfamethazine to (biochar) black carbon. Ten Hulscher, T. E. M., Vrind, B. A., Van den Heuvel, H., Van Noort, P. C. M., and Govers, H. A. J. (2005), Influence of long contact time on sediment sorption kinetics of spiked chlorinated compounds, Environ. Toxicol. Chem. 24, 2154–2159. van Noort, P. C. M. (2003), A thermodynamics-based estimation model for adsorption of organic compounds by carbonaceous materials in environmental sorbents, Environ. Toxicol. Chem. 22, 1179–1188. van Noort, P. C. M., Cornelissen, G., Ten Hulscher, T. E. M., and Belfoid, A. (2002), Influence of sorbate planarity on the magnitude of rapidly desorbing fractions of organic compounds in sediment, Environ. Toxicol. Chem. 21, 2326–2330. Vrbancich, J. and Ritchie, G. L. D. (1980), Quadrapole moments of benzene, hexafluorobenzene and other non-dipolar aromatic molecules, J. Chem. Soc. Faraday Trans. 2 76, 648–659. Walters, R. W. and Luthy, R. G. (1984), Equilibrium adsorption of polycyclic aromatic hydrocarbons from water onto activated carbon, Environ. Sci. Technol. 18, 395–403. Wang, J.-S., Kamiya, Y. and Naito, Y. (1998), Effects of CO2 conditioning on sorption, dilation, and transport properties of polysulfone, J. Polym. Sci., Part B: Polym. Phys. 36, 1695–1702. Weber, W. J. Jr., Kim, S. H., and Johnson, M. D. (2002), Distributed reactivity model for sorption by soils and sediments.15. Highconcentration co-contaminant effects on phenanthrene sorption and desorption, Environ. Sci. Technol. 36, 3625–3634. Weber, W. J. J., McGinley, P. M., and Katz, L. E. (1992), A distributed reactivity model for sorption by soils and sediments. 1. Conceptual basis and equilibrium assessments. Environ. Sci. Technol. 26, 1955–1962. Welhouse, G. J. and Bleam, W. F. (1993a), Atrazine hydrogenbonding potentials, Environ. Sci. Technol. 27, 494–500. Welhouse, G. J. and Bleam, W. F. (1993b), Cooperative hydrogen bonding of atrazine, Environ. Sci. Technol. 27, 500–505.

49

Wen, B., Huang, R. X., Li, R. J., Gong, P., Zhang, S., Pei, Z. G., Fang, J., Shan, X. Q., and Khan, S. U. (2009), Effects of humic acid and lipid on the sorption of phenanthrene on char, Geoderma 150, 202–208. Wen, B., Zhang, J. J., Zhang, S. Z., Shan, X. Q., Khan, S. U., and Xing, B. S. (2007), Phenanthrene sorption to soil humic acid and different humin fractions, Environ. Sci. Technol. 41, 3165–3171. Werth, C. J. and Hansen, K. M. (2002), Modeling the effects of concentration history on the slow desorption of trichloroethene from a soil at 100% relative humidity, J. Contam. Hydrol. 54, 307–327. Werth, C. J. and Reinhard, M. (1997a), Effects of temperature on trichloroethylene desorption from silica gel and natural sediments. 1. Isotherms, Environ. Sci. Technol. 31, 689–696. Werth, C. J. and Reinhard, M. (1997b), Effects of temperature on trichloroethylene desorption from silica gel and natural sediments. 2. Kinetics, Environ. Sci. Technol. 31, 697–703. White, J. C., Hunter, M., Nam, K., Pignatello, J. J., and Alexander, M. (1999a), Correlation between the biological and physical availabilities of phenanthrene in soils and soil humin in aging experiments, Environ. Toxicol. Chem. 18, 1720–1727. White, J. C., Hunter, M., Nam, K., Pignatello, J. J., and Alexander, M. (1999b), Correlation between the biologican and physical availabilities of phenanthrene in soils and soil humin in aging experiments, Environ. Toxicol. Chem. 18, 1720–1727. White, J. C., Hunter, M., Pignatello, J. J., and Alexander, M. (1999c), Increase in the bioavailability of aged phenanthrene in soils by competitive displacement with pyrene, Environ. Toxicol. Chem. 18, 1728–1732. Wijnja, H., Pignatello, J. J., and Malekani, K. (2004), Formation of pi-pi complexes with phenanthrene and model pi-acceptor humic subunits, J. Environ. Qual. 33, 265–275. Williams, J. H. (1993), The molecular quadrapole moment and solid-state architecture, Acc. Chem. Res. 26, 593–598. Wu, J. Z. and Prausnitz, J. M. (2008), Pairwise-additive hydrophobic effect for alkanes in water, Proc. Natl. Acad. Sci. USA 105, 9512–9515. Wu, S. and Gschwend, P. M. (1986), Sorption kinetics of hydrophobic organic compounds to natural sediments and soils, Environ. Sci. Technol. 20, 717–725. Xia, G. and Ball, W. (2000), Polanyi-based models for the competitive sorption of low-polarity organic contaminants on a natural sorbent, Environ. Sci. Technol. 34, 1246–1253. Xia, G. and Ball, W. P. (1999), Adsorption-partitioning uptake of nine low-polarity organic chemicals on a natural sorbent, Environ. Sci. Technol. 33, 262–269. Xia, G. and Pignatello, J. J. (2001), Detailed sorption isotherms of polar and apolar compounds in a high-organic soil, Environ. Sci. Technol. 35, 84–94. Xing, B., Gigliotti, B., and Pignatello, J. J. (1996), Competitive sorption between atrazine and other organic compounds in soils and model sorbents, Environ. Sci. Technol. 30, 2432–2440. Xing, B. and Pignatello, J. J. (1997), Dual-mode sorption of lowpolarity compounds in glassy poly(vinyl chloride) and soil organic matter, Environ. Sci. Technol. 31, 792–799.

50


Xing, B. and Pignatello, J. J. (1998), Competitive sorption between 1,3-dichlorobenzene or 2,4-dichlorophenol and natural aromatic acids in soil organic matter, Environ. Sci. Technol. 32, 614–619. Xiong, J. C. and Maciel, G. E. (2002), Interactions between pyridine and coal at the molecular level: Insights from variable-temperature H-1 NMR studies of pyridine-saturated coal, Energy Fuels 16, 497–509. Yamamoto, H. and Liljestrand, H. M. (2004), Partitioning of selected estrogenic compounds between synthetic membrane vesicles and water: Effects of lipid components, Environ. Sci. Technol. 38, 1139–1147. Young, T. M. and Weber, W. J., Jr. (1995), A distributed reactivity model for sorption by soils and sediments. 3. Effects of diagenetic processes on sorption energetics, Environ. Sci. Technol. 29, 92–97. Yun, Y. and Suuberg, E. M. (1993), New applications of differential scanning calorimetry and solvent swelling for studies of coal structure: Prepyrolysis structural relaxation, Fuel 72, 1245–1254. Zhang, L., Leboeuf, E. J., and Xing, B. S. (2007), Thermal analytical investigation of biopolymers and humic- and carbonaceousbased soil and sediment organic matter, Environ. Sci. Technol. 41, 4888–4894.

Zhao, D. and Pignatello, J. J. (2004), Model-aided characterization of tenax-TA for aromatic compound uptake from water, Environ. Toxicol. Chem. 23, 1592–1599. Zhao, D., Pignatello, J. J., White, J. C., Braida, W., and Ferrandino, F. (2001), Dual-mode modeling of competitive and concentrationdependent sorption and desorption kinetics of polycyclic aromatic hydrocarbons in soils, Water Resour. Res. 37, 2205–2212. Zhu, D., Hyun, S., Pignatello, J. J., and Lee, L. S. (2004), Evidence for pi-pi electron donor-acceptor interactions between pi-donor aromatic compounds and pi-acceptor sites in soil organic matter, Environ. Sci. Technol. 38, 4361–4368. Zhu, D., Kwon, S., and Pignatello, J. J. (2005), Adsorption of singlering organic compounds to wood charcoals prepared under different thermochemical conditions, Environ. Sci. Technol. 39, 3990–3998. Zhu, D. and Pignatello, J. J. (2005a), Characterization of aromatic compound sorptive interactions with black carbon (charcoal) assisted by graphite as a model, Environ. Sci. Technol. 39, 2033–2041. Zhu, D. and Pignatello, J. J. (2005b), A concentration-dependent multi-term linear free energy relationship for sorption of organic compounds to soils based on the hexadecane dilute-solution reference state, Environ. Sci. Technol. 39, 8817–8828.

2 COMPREHENSIVE STUDY OF ORGANIC CONTAMINANT ADSORPTION BY CLAYS: METHODOLOGIES, MECHANISMS, AND ENVIRONMENTAL IMPLICATIONS STEPHEN A. BOYD, CLIFF T. JOHNSTON, DAVID A. LAIRD, BRIAN J. TEPPEN, 2.1. Introduction 2.2. Contributions of Clays Versus Soil Organic Matter to Sorption 2.3. Macroscopic Sorption Studies 2.4. X-Ray Diffraction 2.5. Vibrational Spectroscopic Studies 2.6. Molecular and Quantum Mechanical Simulations 2.7. Environmental Implications 2.8. Summary

2.1. INTRODUCTION The publication of two papers in 1979 (Chiou et al. 1979; Karickhoff et al. 1979) marked the beginning of a major redirection of thinking regarding the sorption of neutral organic contaminants (NOCs) and pesticides in soil/sediment–water systems. Prior to these important papers, soil was generally viewed as an adsorbent containing two highsurface-area components, soil organic matter (SOM) and soil clays, that were responsible for the adsorption of NOCs and pesticides by a variety of mechanisms (e.g., H bonding, van der Waals forces) unique to the particular combination of soil and the organic solute of concern. The 1979 papers set forth a new conceptualization of NOC sorption by soils, which has come to be known as partition theory. With the development of partition theory, the idea that SOM played a dominant, if not singular, role in the sorption of NOCs and pesticides by soils also gained acceptance. Soil organic matter was viewed as an organic partition

AND

HUI LI

phase rather than a high-surface-area adsorbent (Pennell et al. 1995). Accordingly, organic solutes in water would essentially dissolve in amorphous SOM in much the same way that they partition from water into an immiscible organic phase such as octanol or hexane, or into organic polymers/solids such as polyurethane or rubber. The magnitude of sorption would depend directly on the amount of SOM as well as the solubility of the NOC or pesticide in water and its solubility in SOM. Solutes with low water solubility partition to a greater degree into SOM manifesting a larger sorption or partition coefficient (often denoted Kp or Kd). Sorption is quantified by this coefficient, which is in essence a simple ratio of the two concentrations: Kp ¼ Csoil/Cwater, where Csoil and Cwater are the solute concentrations in SOM and water, respectively. Furthermore, when the individual Kp values (for a given NOC) among a series of soils are normalized to the corresponding fractional SOM content (fom) of the soil, the normalized values tend to converge within a factor of 2 (Kile et al. 1995), especially for NOCs devoid of polar functional groups. The SOMnormalized value, Kom ¼ Kp/fom, represents sorption per unit mass of SOM, and can for all practical purposes be considered a “constant.” Once the Kom of a NOC is known, Kp for soils can be easily estimated by Kp ¼ Komfom, that is, simply by knowing the fom of a soil. For context, owing to the semipolar nature of SOM (which has an O content of 40%–50% derived from polar functional groups such as phenolic ---OH and ---COOH groups), the Kom value of a specific NOC/pesticide is generally smaller than the corresponding octanol–water partition coefficient (Kow) by approximately an order of magnitude. Today, Kom is routinely used to predict Kp in NOC/pesticide leaching models; Kp

Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems. Edited by Baoshan Xing, Nicola Senesi, and Pan Ming Huang. Ó 2011 John Wiley & Sons, Inc. Published 2011 by John Wiley & Sons, Inc.

51

COMPREHENSIVE STUDY OF ORGANIC CONTAMINANT ADSORPTION BY CLAYS

defines site-specific soil–water distribution for a given NOC–soil combination. Solute partitioning, which has now gained broad acceptance as a sorption mechanism, manifests certain sorption characteristics that distinguish it from adsorption (Chiou 2002; Chiou et al. 1983). Partitioning is a process of dissolution into the partition phase, whereas adsorption involves condensation of a solute on the surface of the adsorbent. Solute partitioning manifests linear sorption isotherms, noncompetitive sorption effects in multisolute systems (two or more NOCs sorbing simultaneously), and low heat effects on sorption (i.e., low and constant enthalpies). Solute adsorption manifests (generally) nonlinear isotherms, competitive effects (negative) on sorption in multisolute systems, and comparatively large negative enthalpies (or large heat effects on sorption). Considerable evidence has been provided to demonstrate the characteristics of solute partitioning in numerous studies of sorption of NOCs from water by soils and sediments (Chiou 2002). Further examination of NOC sorption by soil has revealed some evidence of sorption isotherm nonlinearity, and competitive effects on sorption in multisolute systems, especially at low relative solute concentrations (the ratio of concentration of solute in water to aqueous solubility of the solute). These effects, which are typically of low magnitude, have been reconciled by several theories that augment the fundamental process of solute partitioning. One such theory suggests that small amounts of high-surface-area carbonaceous materials (e.g., chars, black carbon) are responsible for the observed deviations in sorption behavior predicted from partition theory (Chiou et al. 2000). Also, SOM has been viewed as a dual sorbent containing both “hard” and “soft” components with the former responsible for the deviations from sorption characteristics predicted by strict solute partitioning (e.g., Pignatello and Xing 1996; Leboeuf and Weber 1997; Xia and Ball 2000). These ideas are best viewed as further refinements that augment the concept of solute partitioning as it pertains to NOC sorption by soils and sediments rather than a challenge to the basic idea of NOC partitioning into SOM as the dominant sorptive mechanism. One very important manifestation of the widespread acceptance of partition theory to describe NOC sorption by soils is that most NOC/pesticide leaching models now use simple linear partition coefficients (often estimated by Kp ¼ Komfom) to predict soil–water distribution of NOCs and pesticides. This approach implicitly ignores or discounts any possible contribution of soil clays to sorption. In this context some proponents of partition theory have suggested that soil mineral surfaces are in general polar or hydrophilic in nature, and that in the presence of ambient moisture, water molecules are preferentially adsorbed on these surfaces (Chiou 2002; Chiou and Shoup 1985). Accordingly, since NOCs lack the ability to displace strongly adsorbed water molecules from the mineral surfaces, the mineral compo-

nents are in essence deactivated as adsorbent surfaces for NOCs and pesticides. Evidence has been provided to demonstrate that the addition of water to anhydrous soils results in a diminution in the overall uptake of certain NOCs and pesticides by soils. These data were interpreted to mean that mineral surfaces (e.g., clays) are active solid adsorbents for NOCs in the absence of water, and that under such conditions both mineral phase adsorption and organic matter partitioning are operative sorption processes. However, as water is added, it displaces adsorbed NOCs and pesticides from mineral surfaces (which preferentially bind water) hence sorption overall is reduced and partitioning into SOM is left as the dominant process responsible for the sequestration of NOCs and pesticides in soil–water systems. At this point in the development of our understanding of NOC/pesticide sorption by soils, dogma held that minor “unsuppressed” sorption of NOCs to clays may occur, but that it was dwarfed by sorption to SOM. Then, in 1992 another important paper was published (Laird et al. 1992) showing that relatively pure smectite clays could effectively adsorb atrazine from water. Interestingly, atrazine adsorption by these clays ranged widely, from essentially complete removal to comparatively minimal uptake (Fig. 2.1). This variability was observed by using many different members of the group of smectite clays. These clays possess subtle differences in composition owing to differences in isomorphic substitution in the tetrahedral Si---O and/or octahedral 200

150 x/m (µmol kg-1)

52

100

50

0 0

10

20

30

40

50

-1)

CE (µmol L

Figure 2.1. Adsorption of atrazine from water by several different reference and soil smectite clays [from Laird et al. (1992)]. Plots are of concentrations of sorbed atrazine (x/m) versus equlibrium aqueous atrazine concentration (CE). Sorbents (in descending order) are as follows: Panther Creek beidellite, hectorite, IMV saponite, Amory montmorillonite, Wyoming bentonite, Belle Fourche montmorillonite, Upton montmorillonite, Webster smectite, Polkville montmorillonite, IMV bentonite, Chambers montmorillonite, Camp Bertean montmorillonite, Carmeron smectite/illite, and Otay montmorillonite.

CONTRIBUTIONS OF CLAYS VERSUS SOIL ORGANIC MATTER TO SORPTION

53

TABLE 2.1. Structural Formulas for Dioctahedral and Trioctahedral Clays with either Tetrahedral or Octahedral Isomorphic Substitution Substitution

Dioctahedral

Trioctahedral

Tetrahedral

Beidellite: Na0:33 Al2 ðSi3:67 Al0:33 ÞO10 ðOHÞ2 Nontronite Na0:33 Fe2 ðSi3:67 Al0:33 ÞO10 ðOHÞ2 Montmorillonite Na0:33 ðAl1:67 Mg0:33 ÞSi4 O10 ðOHÞ2

Saponite: Na0:33 Mg3 ðSi3:67 Al0:33 ÞO10 ðOHÞ2

Octahedral

Hectorite Na0:33 ðMg2:67 Li0:33 ÞSi4 O10 ðOHÞ2

Source: Adapted from Gaines et al. (1997).

Al---O sheets of the 2 : 1 clays, and the type (dioctahedral vs. trioctahedral) of octahedral sheet. Specific structures of important members of the smectite group are given in Table 2.1. Remarkably, these subtle differences manifested very large differences in the affinity of particular smectite clays for aqueous-phase atrazine. Soon thereafter, an important series of papers (Haderlein and Schwarzenbach 1993; Haderlein et al. 1996; Weissmahr et al. 1997) similarly showed that clay minerals common to soils, namely, kaolinite, illite, and smectite, could also effectively adsorb nitroaromatic compounds (NACs) from water. Among these, smectite clays were by far the most effective NAC adsorbents. From these and similar studies several questions arise: whether (1) the role of soil clays as effective adsorbents of NOCs/pesticides in soil–water systems been erroneously neglected; (2) when and under what conditions clay minerals are effective adsorbents for NOCs and pesticides in soils; (3) the specific underlying forces and mechanisms by which clays, especially smectite clays, functioned as highly effective adsorbents for NOCs; and (4) how one goes about revealing these molecular-scale forces and mechanisms?

2.2. CONTRIBUTIONS OF CLAYS VERSUS SOIL ORGANIC MATTER TO SORPTION In an attempt to address the first question, we conducted a study that directly compared the potential contributions of SOM and montmorillonite clay as sorbents for several important examples of NOCs and pesticides (Sheng et al. 2001). We used an organic soil (Houghton muck) essentially devoid of clay minerals (having an organic carbon content of 49.5%) as a “surrogate” for pure SOM. This soil, where SOM was the singular (organic) sorbent for NOCs/pesticides, was compared to potassium-saturated montmorillonite, which represented a purely clay mineral adsorbent. On a unit mass basis, SOM was a more effective (to varying extents) sorbent than was K-montmorillonite for biphenyl, diuron, and parathion. However, for atrazine, carbaryl, dichlobenil, and dinitro-ocresol, K-montmorillonite was more effective than SOM as a sorbent phase. Since the clay content of most mineral soils significantly exceeds the SOM content, overall sorption of certain NOCs and pesticides could plausibly be controlled by their interactions with clays, especially the expandable smectites, which possess particularly high surface areas of

800 m2/g. Among these pesticides, specific interactions between pesticide substituents and exchangeable cations of clay, hydrophobic and/or electron donor–acceptor interactions between the aromatic rings of pesticide molecule and the siloxane surfaces of clays, and steric hindrance due to bulky substituents on the pesticide structure seemed to be important determinants of the extents of pesticide adsorption by K-montmorillonite. The occurrence of significant clay–organic interactions observed in the study by Sheng et al. (2001) was consistent with earlier studies that had shown appreciable adsorption of certain pesticides by clays (Bailey and White 1970; Green 1974; Mortland 1986). Also, previous investigations had sought to define conditions under which clay minerals contributed significantly to NOC/pesticide retention in soils, and concluded that this could occur when clay to SOM ratios in soil were greater than 5 to 30 (Hassett et al. 1981; Karickhoff 1984). In another study, the critical clay to organic carbon ratios at which mineral phase sorption accounted for 50% of overall sorption by soils and sediments were ca. 62 for atrazine and 84 for alachlor (Grundl and Small 1993). In retrospect, these seem like fairly gross estimates of soil properties that could be used to benchmark the important role of clays in pesticide retention by soils. More recent studies clearly indicate a much more specific set of conditions that manifest high-affinity clay adsorbents that are unique to the particular combination of solute properties and structure as well as clay type and type of exchangeable cation associated with the clay. Following demonstration that for several classes of important NOCs and pesticides smectite clays are equally or more effective adsorbents than SOM (compared as isolated components), the next logical line of investigation was to reveal the underlying operative mechanisms and forces responsible for adsorption by smectite clays. These clays have been the primary focus of this body of research because of their widespread occurrence, large surface areas, and reversible expandability. These clays consist of a basic 2 : 1 layer structure with an octahedrally coordinated AlO sheet sandwiched between two tetrahedrally coordinated SiO sheets. The specific members of this group of clays differ in being either dioctahedral or trioctahedral, and on the type, location, and extent of isomorphic substitution. For example, saponite is a trioctahedral clay (MgO central sheets) with nearly 100% of the isomorphic substitution

54


TABLE 2.2. Characteristics of Smectite Clay Minerals Used in More Recent Studies as Adsorbents for NOCs and Pesticides

Clay

Smectite Clay Type and Octahedral Sheet Type

SWy-2 SAz-1 Upton SHCa-1 SapCa-2

Montmorillonite, dioctahedral Montmorillonite, dioctahedral Montmorillonite, dioctahedral Hectorite, trioctahedral Saponite, trioctahedral

Tetrahedral Charge, %

Cation Exchange Capacity cmol/kg

Surface Area m2/g

BET N2 Surface Area, m2/g

OC %

Surface Charge Density, mmol/m2

3.6 12 1.6 14 100

83.6 130 113.3 86.4 97.4

766 768 730 743 750

31.82 — 39.8 63.19 —

0.07 — 0.06 K þ Al3þ > Ba2þ > Na þ > Ca2þ (Fig. 2.2). Calculation based on sorption isotherms showed that at a relative concentration of 0.1, Cs-SWy-2 was 1.3 times more effective than K-SWy-2, 5.5 times more effective than Al-SWy-2, 12 times more effective than Ba-SWy-2, 28 times more effective than Na-SWy-2, and 197 times more effective than Ca-SWy-2 for DNOC adsorption from water. Haderlein and Schwarzenbach (1993) and Haderlein et al. (1996) observed qualitatively similar effects of exchangeable cations on adsorption of NACs. Haderlein and co-workers attributed the strong adsorption of NACs by smectites to the formation of electron donor (negative charge sites in clays)–electron acceptor (NACs where the ---NO2 groups attract electron density from the aromatic p-ring system) (EDA) complexes. In a study by Boyd et al. (2001) adsorption of a series of substituted nitrobenzenes by K-SAz-1 revealed that electron-withdrawing groups did enhance sorption in a predictable manner on the basis of the Hammett substituent constant s, and this appeared to be consistent with the proposed formation of

MACROSCOPIC SORPTION STUDIES

75 Cs-SWy-2

Adsorption (mg/g)

60 K-SWy-2 45

30

15

Al-SWy-2 Na-SWy-2

0 0

10

20 Concentration (mg/L)

30

40 Ca-SWy-2

Figure 2.2. Adsorption of dinitro-o-cresol (DNOC) from water by SWy2 montmorillonite saturated with different exchangeable cations [from Sheng et al. (2002)]. Plots are of concentration of adsorbed DNOC versus equilibrium aqueous concentration of DNOC.

EDA complexes. However, quantum calculations of the gas-phase NACs revealed that the electron density of the aromatic p–ring system was unchanged among the substituted nitrobenzenes. Rather, electron density donated by a second substituent on nitrobenzene seemed to be appropriated by the ---NO2 group, leaving the aromatic p ring relatively unaffected. Thus, the quantum calculations did not support the dominance of an EDA mechanism since there was no evidence that the aromatic ring was any more or less electrondeficient regardless of the nature of the second substituent. This observation suggested the predominance of other mechanisms and forces in the adsorption of NACs by smectite clays. It was clear, from a cursory look at exchangeable cation ordering relative to the affinity of DNOC and other NACs, that cations with lower hydration energies, namely, Cs þ and K þ , produced higher-NAC-affinity smectites than did those saturated with cations that had higher hydration energies. This led us to hypothesize that ---NO2 groups, which possess partial negative charge, might form complexes with the exchangeable cations, either directly or through the intermediation of water. Weaker cation hydration would logically favor such interactions. Furthermore, the magnitudes for adsorption of various substituted NACs by a given K-saturated smectite (Boyd et al. 2001) followed an ordering that seemed plausibly consistent with the ability of the functional groups to form complexes with interlayer K þ . In a more recent study (Liu et al. 2009), the adsorption of dibenzo-pdioxin (dioxin) by saponite exchanged with different inorganic cations revealed that Cs-saponite effectively adsorbed dioxin from water, reaching nearly 1% wt/wt. In this instance there was a much larger difference in adsorption (of dioxin) by Cs- versus K-saponite, as compared to sorption of NACs.

55

Adsorption appeared to involve one or both of the oxygens in the dioxin ring. Dioxin adopted at least two orientations on the saponite interlayer. In one, dioxin is essentially dehydrated as it interacts with the opposing siloxane sheets and with coplanar Cs þ via one of the dioxin ring oxygens, analogous to adsorption of NACs. At higher loadings dioxin is intercalated between Cs þ and the clay surface in a tilted orientation when both oxygens of the dioxin ring interact with Cs þ . Because the negative charge character of the dioxin ring oxygens is less than that of the oxygens of ---NO2 groups, the adsorptive requirement for a weakly hydrated cation, namely, Cs þ , is greater than that observed for NACs. Hence, dioxin adsorption by Cs-saponite is much higher than that by K-saponite, whereas this difference is rather smaller for NAC adsorption (Fig. 2.2). A second important finding from macroscopic sorption studies was that smectite clays with lower charge densities were more effective adsorbents for NACs and other NOCs compared to clays with higher charge densities. This had first been observed in studies of adsorption of vapor-phase aromatic hydrocarbons (e.g., benzene, toluene) by smectites exchanged with tetramethylammonium (Lee et al. 1990). Comparison of a relatively “low-charge” smectite (SWy-2, surface charge density 1.09 mmol/m2) to a “high charge” smectite (SAz, surface charge density 1.69 mmol/m2) revealed that the lower-charge SWy-2 clay was a more effective adsorbent for gas-phase aromatic hydrocarbons. Also, uptake of the NOC vapor by dry clay was higher than uptake of the corresponding solute from water. Subsequent studies with similar trimethylphenylammonium clays showed the same trends for adsorption from water of a larger group of aromatic hydrocarbons (Jaynes and Boyd 1990, 1991a). Also, when the charge density of the high-charge SAz clay was chemically reduced (by the Li reduction method), adsorption was directly related to the degree of charge reduction (Jaynes and Boyd 1991b). In our study of DNOC adsorption by K-exchanged smectites we found that the lower charged SWy-2 adsorbed more DNOC from water than the higher-charge SAz, and charge reduction of the SAz clay resulted in proportionately higher adsorption of DNOC (Sheng et al. 2002), (Fig. 2.3). So, from these macroscopic sorption studies, three key observations were made that led to the beginnings of hypotheses regarding the operative adsorption mechanisms: (1) adsorption was higher in smectites exchanged with weakly hydrated exchangeable cations, (2) adsorption of the organic vapor by the dry clay was somewhat higher than adsorption of the corresponding solute from water (i.e., water reduced uptake) and (3) adsorption by lower-charge-density clays was higher than adsorption by higher-charge-density clays. From these observations we proposed that (1) adsorption could involve the exchangeable cation via complexation with polar functional groups or structural units (e.g., with ---NO2 groups of adsorbed NACs), (2) siloxane surfaces between

56


60

(a)

45

K-SWy-2

30

Amount absorbed (mg/g)

K-SAz-1

15

0

(b) 0.40 CEC

60 0.55 CEC

45 0.70 CEC

30

15 0.85 CEC 1.0 CEC K-SAz-1

0 0

5

10

15

20

25

Equilibrium concentration (mg/L)

Figure 2.3. (a) Adsorption of dinitro-o-cresol (DNOC) by homoionic lower K-smectite (K-SWy-2) and higher (K-SAz-1) surface charge density; (b) sorption of dinitro-o-sec-butylphenol (Dinoseb) by charge-reduced K-SAz-1 (from 15% to 60% reduction in cation exchange capacity—denoted 0.85 CEC to 0.4 CEC) [from Sheng et al. (2002)].

exchangeable cations were nanoscale adsorption domains, and (3) water tended to decrease adsorption by hydrating the exchangeable cation—thereby reducing the potential for complex formation and by obscuring the siloxane surface; in both instances more water (i.e., more cation hydration) meant lower adsorption overall. The effects are illustrated schematically in Figure 2.4. 2.4. X-RAY DIFFRACTION X-ray diffraction (XRD) has provided molecule-scale information regarding the separation of the 2 : 1 smectite layers (i.e., the interlayer distance) in clays with and without adsorbed NOCs and pesticides. The interlayer distance of smectites is determined largely by the hydration propensities of the exchangeable cations, and to a leaser extent the layer charge. For example, homoionic K-smectites at 100% relative humidity (RH) typically exhibit interlayer spacings of 12.5–15 A, with lower-charged smectites (e.g., SWy-2)

tending toward the higher spacings (MacEwan and Wilson 1980). Cesium-smectites have a strong tendency to equilibrate with 12.5 A spacings, even in bulk water, due to the lower hydration energy of Cs þ compared to K þ . Smec- tites saturated with Ca2þ , Ba2þ ,or Al3þ always swell to 15 A (d001) in aqueous suspension. Hence, exchangeable cation hydration also influences NOC/pesticide adsorption by smectite clay through its influence on the interlayer spacing of clay sheets in the clay tactoids. In a particularly revealing experiment, Sheng et al. (2002) used XRD to quantify the swelling behavior of K-SWy-2 films in the presence and absence of water vapor, and with varying amounts of adsorbed DNOC (Fig. 2.5). As expected, exposure of previously air-dried K-SWy-2 films to 100% RH caused an increase in the interlayer spacing from 11 to 15 A. Interestingly, the presence of adsorbed DNOC restricted swelling of the rewetted clay. At higher DNOC loadings the interlayer spacings for the air-dried and rewetted (100% RH) clays were nearly equal (12.2 A). The presence of adsorbed DNOC apparently caused the monolayer (i.e., one layer of intercalated water) to be retained even though the same KSWy-2 smectite would swell further (i.e., to 15 A) in the absence of DNOC. Thus, the 12 A spacing appeared optimal for DNOC adsorption. In this configuration, the DNOC molecules would be flat on, and parallel to, the siloxane surfaces of K-SWy-2. Furthermore, the 12 A spacing would allow adsorbed DNOC to interact simultaneously with the opposing clay siloxane surfaces, thereby minimizing its contact with water. Considering that the free energy of hydration of many small NOCs is in the range of 10–30 kJ/ mol1, removal of DNOC from bulk water on intercalation may provide sufficient energy to prevent K-SWy-2 smectite from swelling beyond the measured spacing of 12–12.5 A. This spacing provides an interlayer distance of 3 A, which corresponds to the approximate thickness of DNOC. Since Cs-smectites tend to maintain a 12.5 A spacing in water, due to the lower hydration energy of Cs þ versus K þ , the Cs-smectite is particularly well suited for intercalation of DNOC, and is a somewhat more effective adsorbent for DNOC (Fig. 2.2). It should be noted that these experiments were conducted using self-supporting clay films rather than aqueous suspensions of the K-SWy-2 clay, which are assuredly present under ambient environmental conditions, along with the particles in lesser-hydrated states. From the data of Sheng et al. (2002), it was unknown whether the clays adopt the 12 A spacing in suspension as a result of DNOC adsorption. However, more recently developed, novel XRD techniques now enable XRD measurement of interlayer spacings in actual clay suspensions (Chappell et al. 2005; Li et al. 2007) as well as in air-dried forms. Smectites exist as quasicrystals, which are stacks of subparallel 2 : 1 phyllosilicate layers with parallel c axes and randomly oriented a- and b-axes. In aqueous suspensions, smectite quasicrystals are dynamic in the sense that

X-RAY DIFFRACTION

-

•Low hydration exchangeable cation

Clay layer

EC Adsorptive

-

•Low charged clay

EC

domain

-

•Highest NOC adsorption

•Low hydration exchangeable cation

-

EC EC

EC

•High charged clay

EC

-

-

•Intermediate NOC adsorption

-

•High hydration exchangeable cation

EC

EC

•Low charged clay

EC

57

•Intermediate NOC adsorption

EC

•High hydration exchangeable cation

EC

-

EC

•High charged clay

-

•Lowest NOC adsorption

Figure 2.4. Availability of smectite clay interlayer domains for neutral organic contaminant (NOC) adsorption. Circles around inorganic exchangeable cation (EC) represent the hydration sphere. Lower hydration manifests larger adsorptive domains and greater potential for direct NOC–EC interactions as well as NOC interactions with the siloxane surface.

they are capable of swelling by imbibing water and/or organic molecules between layers within quasicrystals (crystalline swelling) and in the sense that large quasicrystals may cleave, producing two or more smaller quasicrystals, and conversely several smaller quasicrystals may coalesce forming a single large quasicrystal (Laird 2006). In aqueous suspensions some smectites, principally those saturated with Na þ and Li þ in dilute electrolyte solutions, are capable of 15.0

14.5

d 001 (Å)

12.5 Air-dried and re-wetted

12.0 Air-dried

11.5

11.0 0

10

20

30

40

50

Amount of DNOC adsorbe (mg/g)

Figure 2.5. Basal spacings of K-smectite (K-SWy-2) clays with varying levels dinitro-o-cresol (DNOC) sorption in the presence and absence of water [from Sheng et al. (2001)].

60

complete delamination such that the individual lamella are separated by distances 30 A and behave as semi-independent colloids (colloids are not truly independent until the suspension is so dilute that colloid concomitant volumes do not interact). Both crystalline swelling/shrinking and the breakup/reformation of quasicrystals are inherently hysteretic processes, as energy is necessary to effect the physical rearrangements of matter that are required for both directions of both process [see Laird (2006) for a more complete discussion of smectite swelling processes]. Any change in the physical state of a smectite quasicrystal in an aqueous system inherently changes the affinity of the smectite for an organic solute. Crystalline swelling, for example, can change the size of interlayer domains from 12.5 A spacing, which is optimal for retention of many organic molecules, to a more hydrated and hence less optimal 15 or 18 A spacings. Similarly the breakup of a large smectite quasicrystal into two or more smaller quasicrystals diminishes the number of interlayer adsorption sites and increases the number of external surface adsorption sites. Within interlayers organic molecules may interact simultaneously with the basal oxygens of two opposing siloxane surfaces, whereas organic molecules interact with only one siloxane surface on an exposed external surface. Furthermore, the permittivity of the interlayer water is lower than the permittivity of water adjacent to an external surface. Weakly polar organic molecules are partitioned from a high-permittivity aqueous phase into a low permittivity aqueous phase with the driving force being an increase in system entropy.

58


20 Å

Relative intensity

17.5 Å 15.0 Å

12.5 Å

Ca-ND Ca-AD

K-ND K-AD

3

4

5 6 7 Degrees two theta

8

9

Figure 2.6. Transmission X-ray diffraction patterns for Panther Creek smectite suspensions. The smectite was saturated with Kþ or Ca2þ and either air-dried and resuspended (AD) or never dried (ND). The patterns have been vertically offset. [The figure is adapted from data originally presented by Chappell et al. (2005)].

4000 Ca-ND Ca-AD

Sorbed Atrazine (µmol kg-1)

The impact of crystalline swelling on adsorption of organic molecules was clearly demonstrated by Chappell et al. (2005). For the study, two suspensions of K þ -saturated Panther Creek smectite were prepared by dialysis from an original highly swollen Na þ Panther Creek biochar suspension. The two K-smectite suspensions were identical in every way except that one had been air-dried and then resuspended while K þ -saturated and the other was maintained as an aqueous suspension the entire time it was in the K form. Quasicrystals in the never-dried K-smectite (K-ND) suspension were swollen (Fig. 2.6: broad XRD peak between 15 and 17.5 A indicating interstratification of lamellae with two or three layers of interlayer water molecules) to a greater extent than were the quasicrystals in the air-dried K-smectite (KAD) suspension (Fig. 2.6: broad XRD peak between 12.5 and 15 A indicating interstratification of layers with one or two layers of interlayer water molecules). This legacy in the extent of crystalline swelling from having been air-dried and then resuspended caused an order of magnitude increase in the affinity of the K-AD sample for atrazine relative to the K-ND sample (Fig. 2.7). By contrast, Ca-smectite in air-dried and never-dried (Ca-AD and Ca-ND, respectively) suspensions adsorbed similar amounts of atrazine and exhibited similar broad 20-A XRD peaks indicating that most interlayers held four layers of interlayer water molecules regardless of the air-dried or never-dried treatments. The results (Chappell et al. 2005) demonstrate that the impact of saturating cation on the affinity of a smectite for organic molecules comes primarily from the impact of the cation on the extent of interlayer hydration, which for K-smectites is strongly influenced by the history of sample handling, or by extension, wetting–drying cycles in nature. Comparison of XRD patterns for randomly oriented K- and Ca-smectite quasicrystals in aqueous suspensions

K-ND

3000

K-AD

2000

1000

0 0.00

0.05

0.10

0.15

0.20

0.25

0.30

Reduced concentration

Figure 2.7. Isotherms for sorption of atrazine on Panther Creek smectite saturated with Kþ or Ca2þ and either air-dried and resuspended (AD) or never dried (ND). Reduced concentration refers to the ratio of the concentration of atrazine in water to aqueous solubility of atrazine. [The figure is adapted from data originally presented by Chappell et al. (2005)].

(Shang et al. 1995) with XRD patterns for air-dried- and oven-dried-oriented films of the same smectites provided insight into the impact of the breakup and reformation of smectite quasicrystals on sorption of DNOC (Pereira et al. 2007, 2008). The phenolate form of DNOC, which is the dominant form of DNOC in solutions with pH 4.4, was adsorbed primarily on external surfaces of K-smectite quasicrystals in aqueous suspensions and entered the interlayers as K-DNOC complexes when individual K-smectite layers and small quasicrystals coalesced to form large quasicrystals on drying. By contrast, in Ca-smectite suspensions (pH 4.4) the phenolate form of DNOC was adsorbed only on external surfaces quasicrystals and there was no evidence for the formation of Ca-DNOC complexes. However, some of the phenolate DNOC became entrapped between substacks within Ca-smectite quasicrystals as the suspensions were dried to form oriented clay films. The neutral form of DNOC (pH 4.4) was adsorbed in the interlayers of a low-charge-density Ca-smectite but not in the interlayers of a high-chargedensity Ca smectite, apparently due to steric restrictions. Studies of homoionic smectite suspensions provide insight into mechanisms controlling smectite–organic interactions. However, soils and sediments invariably contain multiple types of cations, with Ca2þ , Ma2þ , and K þ typically being the most abundant cations in temperate region soils. Thus, in natural systems, cation exchange reactions and the adsorption/desorption of organic solutes occur simultaneously. As discussed above, both the type of cation adsorbed on the exchange complex of a smectite and the physical state of the smectite quasicrystals have a large influence on the

VIBRATIONAL SPECTROSCOPIC STUDIES

affinity of smectites for organic solutes. Similarly, the extent of crystalline swelling of smectites influences cation ex change selectivity, such that interlayers with 12.5 A spacings have a distinct preference for weakly hydrated monovalent cations (e.g., K þ ) and more expanded interlayers (d spacings of 15–20 A) prefer more strongly hydrated divalent cations such as Ca2þ and Ma2þ (Laird and Shang 1997). Thus quasicrystal dynamics regulates interactions between cation exchange reactions and the adsorption/desorption of organic solutes, and conversely the loading of inorganic cations and organic molecules in the interlayers regulates quasicrystal dynamics (Li et al. 2004b; Chatterjee et al. 2008). Furthermore, the hysteresis that is inherent in crystalline swelling/ collapse and quasicrystal formation/breakup (Laird et al. 1995) causes hysteresis in both cation exchange reactions (Laird and Shang 1997) and hysteresis in the adsorption/desorption of organic solutes (Li et al. 2004b; Chatterjee et al. 2008). Demixing of both inorganic cations and organic solutes, such that K þ and organic solutes are selectively coadsorbed in collapsed (d ¼ 12.5 A) interlayer domains while Ca2þ and water molecules are selectively coadsorbed in expanded (d ¼ 15–20 A) interlayer domains of the same quasicrystal (Pils et al. 2007), is another implication of the interaction between quasicrystal dynamics, cation exchange reactions, and the adsorption/desorption of organic solutes.

2.5. VIBRATIONAL SPECTROSCOPIC STUDIES In order to achieve a mechanistic understanding of NOC interactions with soil constituents, adapted spectroscopic methods sensitive to molecular interactions are required. Sorption data are macroscopic in nature and therefore, are fundamentally insensitive to molecular phenomena (Johnston and Sposito 1987). In the case of NAC sorption to clay minerals, vibrational spectroscopy, in particular, has provided molecule-level insights about the sorption mechanisms on several different aspects. In general, NACs are well suited for clay sorption studies because selected NACs show a high affinity for smectite, and because NACs have strong IR-active vibrational modes that are sensitive to short-range intermolecular interactions (Saltzman and Yariv 1975; Weissmahr et al. 1997; Johnston et al. 2002; Sheng et al. 2002). In the case of 1,3,5-TNB on K-SWy-2 smectite, for example, sorption values approach 400 mmol/g (Johnston et al. 2001). The vibrational bands associated with the ---NO2 groups are of particular interest in NAC surface studies. In earlier work, Yariv and co-workers (Yariv et al. 1966; Saltzman and Yariv 1975) used IR spectroscopy to study the interactions of nitrophenol and nitrobenzene with smectites. The positions of the ---NO2 vibrational bands were affected by the clay surface and by the nature of the exchangeable cation when compared to the neat compound. Spectra in these

59

earlier studies were collected from air-dried and heated clay films to minimize the interference from water and utilized high surface loadings of the organic solute of interest. No clear spectral trends were observed from the air-dried films and evidence for inner-sphere complexation was observed for the heated films exchanged with different cations (Saltzman and Yariv 1975). The principal bands of interest are the nasym(NO) and nsym(NO) bands along with the ---NO2 deformation bands. The ---NO2 group is highly electronegative, and the vibrational motions associated with nasym(NO) and nsym(NO) modes induce a large change in the induced dipole moment, which translates to intense/strong IR-active modes. Of particular benefit to clay surface studies is the fact that these vibrational modes occur in accessible spectral regions that are not obfuscated by the intense clay bands or by the presence of water (Fig. 2.8) (Johnston et al. 2001). The NACs can be considered to function as molecular probes that have diagnostic properties that are sensitive to changes in their local environment (Johnston et al. 1993). In a related way, the chemical shift of both 15 N and 17 O nuclei have been used to probe changes in the chemical environment around the ---NO2 group resulting from intermolecular interactions. Sorption of NAC is favored on low-charge-density clay minerals that are exchanged with less hydrated exchangeable cations (Haderlein et al. 1996; Boyd et al. 2001). On the basis of earlier studies (Jaynes and Boyd 1991b; Laird et al. 1992, 1994; Barriuso et al. 1994), increased sorption was attributed to the presence of neutral siloxane surface sites, which are regions on the smectite surface not in the near vicinity of exchangeable cations, waters of hydration surrounding exchangeable cations, or the isomorphic substitution sites themselves (Boyd et al. 2001; Sheng et al. 2002; Johnston et al. 2004; Li et al. 2004a). These portions of the clay surface are less hydrated and are considered to provide favorable sorption domains for semi-polar organic solutes such as NACs, atrazine, and related compounds. Haderlein and coworkers were among the first to show that a wide range of NACs showed a high affinity for smectites in sorption studies from aqueous suspension (Haderlein and Schwarzenbach 1993; Haderlein et al. 1996). Nitroaromatic compounds that show high affinities for smectites were generally found to be planar structures containing more than one nitro group (Haderlein et al. 1996). This work was extended using sorption and molecular spectroscopy [NMR, UV–visible, and in situ attenuated total veflection (ATR)–FTIR spectroscopy] to the study of NAC sorption mechanisms on smectite (Weissmahr et al. 1997). Although direct coordination mechanisms had been proposed in earlier NAC-smectite sorption studies (Yariv et al. 1966; Saltzman and Yariv 1975; Fusi et al. 1982) on the basis of air-dried clay films, they observed similar in situ ATR-FTIR results between Cs þ - and K þ exchanged smectites. Based on the spectral similarities they concluded that the exchangeable cation did not play a

60


Figure 2.8. FTIR spectra of 1,3-dinitrobenzene sorbed to K-SWy-1 smectite. Individual band assignments are shown as well as spectral regions where IR absorption resulting from the smectite, sorbed water, and 1,3-dinitrobenzene occur [from Johnston et al. (2001)].

significant role and attributed the high affinity of certain NACs for smectites to the formation of a electron-donor– acceptor (EDA) complex discussed earlier between the p electrons of the NAC and the electrons of the siloxane oxygen atoms, which was not entirely consistent with the earlier 1550

ex situ IR studies (Yariv et al. 1966; Saltzman and Yariv 1975; Fusi et al. 1982). In order to address this apparent inconsistency further, we investigated the spectroscopic properties in the study of NACs sorption on three different smectites exchanged

(a)

Na+ K+

1548

νasymm (NO) Mg+2

Band position (cm–1)

1546

Ba+2 Ca+2

1544 SWy-1 SAz-1 SHCa-1

1358 (b) 1356 1354

νsymm (NO)

1352 1350 1348 -2500

-2000

-1500

-1000

-500

0

Enthalpy of Hydration (kJ/mol)

Figure 2.9. Positions of the nasymm(NO) (a) and nsymm(NO) (b) bands of 1,3,5-trinitrobenzene sorbed to SAz-1, SWy-1, and SHCa-1 smectite exchanged with Mg, Ca, Ba, Na, and Cs; the data are plotted as a function of the enthalpy of hydration of the exchangeable cation [from Johnston et al. (2001)].


61

exchanged with strongly hydrated cations (Mg2þ , Ca2þ , Ba2þ , Na þ). The splitting decreased to 188 cm-1 for the weakly hydrated cations (K þ and Cs þ ) (Fig. 2.9). A similar decrease in the splitting between the nasym(NO) and nsym(NO) bands was observed for 1,3-dinitrobenzene (Fig. 2.10). In order to investigate this splitting further, quantum mechanical calculations of gas-phase 1,3,5-TNB, 1,3-dinitrobenzene, and inner-sphere K complexes of both NACs were undertaken (Johnston et al. 2001), using the B3LYP level of theory and the 6–311G basis set (Frisch et al. 1998). Splittings between the computed vibrational frequencies for the K complexes relative to the uncomplexed NACs showed the same shifts in direction, although of somewhat larger magnitude, that had been observed experimentally. These computational results showed that the decrease in splitting resulted from strengthened interactions between the ---NO2 oxygen atoms and K þ , thus supporting the hypothesis that adsorbed NACs formed complexes with interlayer K þ ions. For more strongly hydrated exchangeable cations, the effective distance between the ---NO2 group

with cations whose enthalpies of hydration varied from 315 kJ/mol (Cs þ) to 1960 kJ/mol (Mg2þ ) (Evangelou 1998). Although Weissmahr et al. (1997) concluded, in light of spectral similarities between Cs- and K-exchanged smectites, that ---NO2 groups were not involved in direct coordination to surface sites or exchangeable cations, the difference in the enthalpy of hydration between Cs þ (315 kJ/mol) and K þ (360 kJ/mol) is almost negligible relative to the difference of more strongly hydrated alkali-metal cations (Li þ and Na þ ) or the alkaline-earth cations (Mg2þ , Ca2þ and Ba2þ ) (Friedman and Krishnan 1973). We found that the nature of the exchangeable cation influenced both the position and the relative intensities of the ---NO2 symmetric and asymmetric stretching bands compared to their spectra obtained in aqueous solution (Fig. 2.9, Table 2.3) (Johnston et al. 2001a, 2002, 2004; Sheng et al. 2002). In the case of 1,3,5-trinitrobenzene, for example, the splitting between the asymmetric and symmetric NO2 stretching bands remained relatively constant at 200 cm1 for 1,3,5-TNB in aqueous solution and when sorbed to smectites

TABLE 2.3. Influence of Exchangeable Metal Cations on Positions (n, cm-1) and Relative Intensities (I) of ---NO2 Stretching Bands of 1,3,5-Trinitrobenzene Sorbed by Different Smectite Clays (SWy-1, SAz-1, SHCa-1) Element(s)

nasym(NO)

Isym

Frequency Difference

Iasym/Isym

0.98 1.12 1.18 2.48 3.40 18.79 22.16

200 200 200 200 196 196 191

0.79 0.83 0.84 0.83 1.12 1.05 1.15

1350 1350 1349 1348 1352 1355 — —

— — — — — 6.17 8.56 21.65

— — — — — 187 197 193

— — — — — 1.00 1.05 1.27

1349 1349 1349 1349 1352 1353 1350 1356

2.96 3.49 2.94 2.56 2.51 18.23 12.65 10.63

200 200 200 200 197 190 196 188

0.83 0.89 0.83 0.73 0.94 1.21 1.29 1.40

nsym(NO)

Iasym

SWy-1 Mg Ca Ba Al Na NH4 K

1549 1549 1549 1549 1548 1545 1546

0.78 0.93 0.99 2.07 3.81 19.65 25.39

1349 1349 1349 1348 1352 1349 1355 SAz-1

Mg Ca Ba Al Na Cs NH4 K

— — — — — 1542 1547 1547

— — — — — 6.18 — —

Mg Ca Ba Al Na Cs NH4 K

1549 1549 1549 1549 1549 1544 1546 1544

2.47 3.09 2.43 1.88 2.36 22.10 16.36 14.87

SHCa-1

Source: Adapted from Johnston et al. (2001).

62


Figure 2.10. Comparison of the FTIR spectra of 1,3-dinitrobenzene sorbed to K-SWy-1 montmorillonite (upper spectrum) to that of 1,3-dinitrobenzene in aqueous solution (lower spectrum, dashed line); the splitting between the nasymm(NO) and nsymm(NO) bands is shown as well as the intensity ratio Iasym/Isym [from Johnston et al. (2001)].

and the positively charged cation is increased, essentially turning off the site-specific interaction. In addition to the cation-induced change in band positions of the nasym(NO) and nsym(NO) bands, the intensity ratio of Iasym(NO)/Insym(NO) was also perturbed and varied as a function of enthalpy of hydration of the exchangeable cation (Fig. 2.11). In prior NAC surface studies, this intensity ratio was sensitive to different types of intermolecular interactions resulting from changes in the O---N---O angle and hydrogen bonding to oxide surfaces (Ahmad et al. 1996). Saltzman and Yariv (1975) observed that the position of the nasym(NO) band decreased and its relative intensity

Intensity ratio iasymm/isymm

1.45 1.35 1.25

SWy-1 SAz-1 SHCa-1 K

+

1.15 1.05 +2

0.95 0.85 0.75 -2500

Ca Mg

+2

+2

Ba

Na+

-1500 -1000 -500 -2000 Enthalpy of hydration (kJ/mol)

0

Figure 2.11. Intensity ratio (integrated area) of the nasymm(NO) and nsymm(NO) bands as a function of the enthalpy of hydration of the exchangeable cation on three smectite clays (SWy-1, SAz-1 and SHCa-1) [from Johnston et al. (2001)].

increased for p-nitrophenol sorbed on smectite when heated, and this was attributed to direct complexation with the exchangeable cation. Furthermore, the spectroscopic results obtained in these studies agree with the results obtained by Weissmahr et al. (1997) in that the IR data obtained for K- and Cs-exchanged smectites are similar. The vibrational bands associated with the ---NO2 group are relatively sharp and well resolved with full-width at halfmaximum (FWHM) bandwidths of 10 cm-1, which is quite narrow for solutes sorbed on surfaces in aqueous suspension. Although highly resolved, the band positions of the NO2 groups do not experience large shifts in position resulting from intermolecular interactions (Nyquist and Settineri 1990; Ahmad et al. 1996). In the context of prior NAC surface studies, the shift in band positions of the n(NO) bands for the NACsmectite complexes are some of the largest shifts reported (Conduit1959; Urbanski and Dabrowska 1959; Jonathan 1960; Borek 1963; Baitinger et al. 1964; Green and Lauwers 1971; Nyquist and Settineri 1990; Ahmad et al. 1996). In spectroscopic studies of environmentally relevant solutes such as NACs, it is useful to combine sorption and spectroscopic methodologies such that the amount of NAC sorbed is known. This type of information provides a direct link between the macroscopic sorption data and the molecular insights gained from spectroscopy. We conducted a series of parallel sorption and spectroscopic measurements and found that intensity of the NAC vibrational bands [nsym(NO) of 1,3-dinitrobenzene is shown in Fig. 2.12) increased linearly with increasing surface coverage as determined using HPLC (Fig. 2.12a). This confirmed the identity of the surface species as 1,3-DNB and not some


(c)

63

1

Absorbance@ 1353 cm-1

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

5

10

15

20

25

30

35

Amount of 1,3-DNB sorbed (mg/g)

Figure 2.12. (a,b) Absorbance of the nsymm(NO) band of 1,3-dinitrobenzene (DNB) at 1353 cm1 at different surface coverages (a) Each spectrum shown in (a) corresponds to a point on the HPLCderived sorption isotherm shown in (b). Combinations of HPLC- and FTIR-derived sorption isotherms are shown in the (b). Adsorption isotherms (b) are plots of concentration of sorbed DNB (Cs) versus equilibrium aqueous concentration of DNB (Ce) [from Johnston et al. (2001)]. Plot of the absorbance of the nsymm(NO) band of 1,3-dinitrobenzene (1,3-DNB) at 1353 cm-1 as a function amount of 1,3DNB sorbed by K-smectite (K-SWy-1) montmorillonite determined using HPLC methods [from Johnston et al. 2001)].

degradation product. In addition, unlike traditional batch sorption methods, the surface solute is directly observed using FTIR as opposed to batch-sorption-derived data, which are based on difference in aqueous-phase concentrations measured using HPLC. In addition, the quantitative analysis of the spectroscopic data provides a measure of the limit of detection of NACs on smectite surfaces. For the data shown in Figure 2.12b, based on a minimum absorbance value of 1 milliabsorbance unit (mAU), the detection limit of 1,3-DNB would be approximately 150 nmol/g. Polarized infrared spectroscopy can provide a direct means of determining the molecular orientation of sorbed species on oriented self-supporting film clay films (Johnston et al. 2002; Ras et al. 2003, 2007). In the case of planar organic solutes, such as many NACs, the vibrational modes

can be divided into in-plane and out- of-plane vibrational modes. The atomic motions corresponding to the nasym(NO) and nsym(NO) modes of 1,3,5-TNT, 1,3-DNB, and DNOC are aligned within same plane as the aromatic ring. In contrast, the vibrational motion of the atoms involved in the out-ofplane ---NO2 deformation are oriented (as their name implies) out of plane. Because the smectite particles have a high aspect ratio, highly oriented self-supporting films can be made (Johnston and Premachandra 2001). When planar organic solutes are sorbed on smectites, linear dichroism techniques can be used to determine the molecular orientation of the sorbed species (Margulies et al. 1988). If the dichroic ratio (DR) for a particular vibrational mode is >1, the orientation of that particular vibrational mode is aligned parallel to the clay surface. The DR for the nsym(NO) band of

64


Figure 2.13. FTIR spectra of dinitro-o-cresol (DNOC) adsorbed by K-smectite (K-SWy-2) at 0 and 45 of beam incidence [from Sheng et al. 2002)].

DNOC on K-SWy-2 is 1.32. In support of this value, the measured DR value for the out-of-plane ---NO2 rock is 0.52 (Fig. 2.13). Together, these linear dichroism values indicate that the molecular plane of DNOC is parallel to the [001] plane of the clay surface (parallel to the siloxane surface as illustrated in Fig. 2.14). Interestingly, similar studies with

dioxin sorbed on Cs-saponite revealed that intercalated dioxin was present in orientations that were not parallel to the siloxane surface, in agreement with XRD measurements showing expanded interlayer distance of 15.4 A at high loadings of 0.8% wt/wt (Liu et al. 2009).

2.6. MOLECULAR AND QUANTUM MECHANICAL SIMULATIONS

Figure 2.14. Illustration of sorbed dinitro-o-cresol (DNOC) molecule laying flat on the siloxane surface showing the siloxane surface and water molecules surrounding the exchangeable cations. The dimensions of the DNOC molecule and the distance between the exchangeable cations are drawn to scale. The DNOC is shown as a representative nitroaromatic compound (NAC). [From Johnston et al. 2002)]. (See insert for color representation of this figure.)

In addition to the quantum chemical studies described above, molecular simulations using classical dynamics have been used (Teppen et al. 1997, 1998) to integrate experimental data and to explore the molecular mechanisms of NOC interactions with clay mineral surfaces. Clay minerals with compositions similar to various smectites were constructed (Boyd et al. 2001; Sheng et al. 2002; Chappell et al. 2005; Aggarwal et al. 2006a). Our experimental adsorption isotherms were used to choose realistic loading rates for the NOCs, and interlayer water contents were derived by inference from X-ray diffraction patterns gathered for both airdried clay films (Sheng et al. 2002; Li et al. 2004a) and suspension-phase (Chappell et al. 2005) NOC–water–clay complexes. Water molecules and NOCs were inserted into the simulation cell at random initial positions and orientations, and the water contents were adjusted until the simulated equilibrium d(001)-spacing agreed with experimental values, so that the simulated interlayer environment corre-

ENVIRONMENTAL IMPLICATIONS

65

2.7. ENVIRONMENTAL IMPLICATIONS

sponded as closely as possible to that of the adsorption experiments. After lengthy equilibrations, the interlayer structures were sampled. An example is shown in Figure 2.15, for which the clay layers have been removed to reveal the NAC–cation–water interactions in the interlayer region. The results are consistent with our hypotheses derived from FTIR spectroscopy in that a variety of inner- and outersphere NAC-K þ complexes are evident. One general observation from the simulations is that such complexes are almost forced by the sterically crowded nature of the smectite interlayer region. Specifically, in a smectite interlayer of 12.5-A d001 spacing as in Figure 2.5, each monovalent cation occupies only 0.5–1 nm2, while single-ring aromatic NOC molecules are at least 0.5 nm2 in size. Thus, the adsorbed NOCs must be forced into close proximity to cations, and this situation will be energetically favorable only if the NOC contains O- or N-based functional groups that are at least polar enough to plausibly coordinate cations. Indeed, two or more polar functional groups are even more favorable since the NOC must fit between several cations (Fig. 2.15). For example, the close proximity of many monovalent cations can be used to rationalize why the magnitudes of nitrobenzene sorption by K-smectites follow the order 1,3,5-TNB 1,4-DNB 1,3-DNB NB. Also, dibenzo-p-dioxin sorbs up to 8000 mg/kg on certain Cs-smectites (Liu et al. 2009), while similar-sized PAHs that lack the oxygen functionality sorb to a much smaller extent to the same clay.

Webster A (K)

p-NCB sorbed (mg g-1)

Figure 2.15. Molecular dynamics snapshot of dinitro-o-cresol (DNOC) in K-smectite (K-SWy-2) clay interlayer, (Kþ ¼ green, O ¼ red, C ¼ gray, N ¼ blue, H ¼ white). (See insert for color representation of this figure.)

The results of our studies and those others (referenced above) clearly demonstrate that NOCs and pesticides often display strong affinities for expandable 2 : 1–layer smectite clays, especially those saturated with weakly hydrated cation (e.g., K þ , Cs þ ). Most of these studies have been conducted using relatively pure reference clay specimens. In the environment, soil clay minerals and SOM are usually associated with each other. Soil organic matter might obscure clay surfaces, thereby reducing the availability and hence efficacy of soil mineral fractions for adsorption of organic compounds. Until 2005 or so, there were no studies that directly assessed the effectiveness of unisolated clay minerals in soils as adsorbents for NOCs, and specifically the extent to which mineral components are available for adsorption of NOCs. As an initial step to approach this question, Charles et al. (2006a,b) measured sorption of several NACs by K þ and Mg2þ -saturated soils and SOM-removed soils. The results showed that extraction of SOM caused an increased sorption by K þ -saturated soils (Fig. 2.16) demonstrating that SOM posed an overall negative effect on sorption of NACs (e.g. p-nitrocyanobenzene, p-NCB), due to obscuration of p-NCB binding sites on soil clays. In contrast, sorption of p-NCB by Mg2þ -saturated Webster soil was greater than that by the Mg2þ -saturated SOM-removed soil (Fig. 2.16). Removal of SOM might liberate some clay surface sites; however, Mg-clays have low affinities for aqueous-phase NACs. Because clays in the Mg2þ -saturated SOM-removed soil were relatively inefficient sorbents for p-NCB, SOM was left as the principal sorbent phase, and its

Webster A (Mg) Webster A (OM-removed) (K)

0.6

Webster A (OM removed) (Mg) a

0.4

b

0.2

c d

0.0 0

10

20

30

Aqueous p-NCB concentration (mg L-1)

Figure 2.16. Adsorption isotherms for p-nitrocyanobenzene (pNCB) sorption by Kþ - and Mg2þ -saturated Webster soil (whole soil, and soil from which SOM was removed); the four isotherms (labeled a–b) are statistically different at p < 0.05 [from Charles et al. (2006a)].

66


removal decreased sorption overall. Results such as these demonstrate that SOM and clay minerals can contribute to the sequestration of NOCs in soils, and that these contributions are interrelated. To estimate contributions of sorption by minerals, several previous studies utilized the simple additive product of isolated SOM (e.g., humic acids) and mineral components (Pusina et al. 1992; Celis et al. 1997; Onken and Traina 1997; Li et al. 2003). However, this approach is inadequate to describe sorption of organic species by soils because it unrealistically assumes that these soil components act independently even though it is known that they are intimately associated in soils. Recognizing this, Karickhoff (1984) proposed the use of a mineral phase availability factor ( fa) to assess the fraction of mineral surfaces (mostly clays) available for sorption of NOCs by soils, although the approach was not evaluated experimentally. This equation incorporated potential SOM blockage of sorption sites on clays by summing NOC sorption to clay and SOM: Qsoil ¼ fa Qmin fmin þ Qsom fsom

ð2:1Þ

where Qsoil is the NOC mass sorbed per unit mass of soil, Qsom and Qmin are the SOM-sorbed and mineral-sorbed NOC per unit mass of the respective sorbent phase, and fmin and fsom are the fractional mineral and SOM contents of soil. The term fa represents the fractional availability of sorption sites on the clay components of whole soil, that is, the fraction of mineral sorptive surfaces available in whole soil. The term fa is a plausible refinement of simple addition of individual soil component contributions. Fractional availability ranges from 0 (unavailable) to 1 (100% available). Karickhoff (1984) equated the mineral fraction to the clay fraction owing to the high surface area of clays in general and of smectites in particular. Using Equation (2.1) to experimentally estimate the fa values requires knowledge of sorption by isolated soil minerals and SOM, but it is unachievable to fully separate unaltered soil components for sorption measurements. To address this deficiency, Charles et al. (2006a) developed a novel approach to determine fa values that involves measuring the difference in sorption by soil whose cation exchange sites are saturated with different cations (i.e., K þ vs. Mg2þ ) that render the clay surfaces adsorptive or nonadsorptive for NOC probe molecules. Specifically, the equation is expressed as QK-soil QMg-soil ¼ fa ðQK-min fmin QMg-min fmin Þ þ ðQK-som fsom QMg-som fsom Þ ð2:2Þ where the terms are the same as those in Equation (2.1), and subscripts K- and Mg- are added to differentiate K þ - and Mg2 þ -saturated soils or soil components.

Since QK-som QMg-som for NOCs (Charles et al. 2006a), it follows that QK-soil QMg-soil fa ðQK-min fmin QMg-min fmin Þ

ð2:3Þ

Equation (2.3) eliminates the Qsom and fsom terms, which are difficult to obtain experimentally but needed in Equation (2.1). For instance, reliance on Qsom introduces significant errors in Equation (2.1) since Qsom cannot be obtained directly from sorption by isolated and unaltered SOM. Using several NAC probe molecules, Charles et al. (2008) estimated the soil mineral availabilities of several smectitic soils. The results showed that 46%–100% of mineral surfaces in the Webster soil horizons and 35%–96% of mineral surfaces in the Clarion soil horizons were available for NOC adsorption. The fa values were negatively correlated with the ratio of SOM/smectite contents in soils. Thus, SOM can reduce the availability of clay mineral surfaces for NOC adsorption. In soils SOM may coat clay surfaces, bridge clay packets resulting in soil aggregate formation, inhibit smectite shrinking and swelling, and partially block the entrance of NACs into interlayer regions. Our studies have revealed that, in fact, fa is not a fixed intrinsic value for a given soil. Rather, the value depends on the type of the probe molecule used. Probe molecules with higher affinity for smectite surface adsorption sites (e.g., NACs) are more effective in accessing these sites, thereby manifesting higher fa values. This is attributed to the more effective competition for mineral adsorptive sites by NACs than by SOM. It is apparent that a large portion of soil mineral fractions are available for adsorption of NOCs and pesticides in whole soils, particularly for adsorbates (e.g., NACs) with strong affinities for soil smectites whose cation exchange sites are partially or fully saturated with weakly hydrated cations (e.g., K þ , Cs þ , NH4 þ ). It is not necessary to create fully K þ -, Cs þ -, or NH4 þ saturated soils or sediments for significant clay mineral adsorption. The amount of weakly hydrated inorganic cations could be just sufficient to produce clay demixing of exchangeable cations. In other words, some (but not all) clay layers or regions of clay tactoids may be fully compensated with K þ , Cs þ or NH4 þ leading to high-affinity NOC adsorption sites, while other regions are saturated with different exchangeable cations (e.g. Ca2þ , Mg2þ , or Na þ ) that do not induce strong affinities for NOCs (Chatterjee et al. 2008; Li et al. 2004b). To demonstrate this principle, adsorption of pesticides with different polarities (i.e., dichlobenil, monuron, biphenyl) was measured by homoionic K- and Ca-SWy-2 in KCl/ CaCl2 aqueous solutions (Li et al. 2004b). The presence of different amounts of KCl and CaCl2 in solution resulted in varying populations of K þ and Ca2þ on the clay exchange sites as a result of cation exchange. Pesticide sorption coefficients were calculated at a relative concentration of

1

K-SWy-2 Ca-SWy-2

0.1

0.01 0.0

dichlobenil monuron biphenyl

0.2

0.4

0.6

0.8

1.0

Figure 2.17. Sorption coefficients of pesticides normalized to the sorption coefficient by K-SWy-2 at the aqueous relative concentration (ratio of the solute concentration in water to the water solubility of solute) of 0.1 as a function of fractional Kþ populations on mineral surfaces. The open (hollow) symbols represent the sorption in which Kþ on homoionic K-SWy-2 was replaced by Ca2þ , and the solid symbols indicate the sorption in which Ca2þ on homoionic Ca-SWy-2 was replaced with Kþ from aqueous solutions. [from Li et al. (2004b)].

0.1 (aqueous equilibrium concentration/aqueous solubility) and normalized to the corresponding sorption coefficients by homoionic K-SWy-2 (i.e., 580 L/kg for dichlobenil, 27 L/kg for monuron, and 6.4 L/kg for biphenyl). When the normalized sorption coefficients were plotted against K þ fractions (fK) on mineral surfaces (Fig. 2.17), sorption of the least polar biphenyl remained nearly constant (and low) across the variation of fK on minerals from zero to one. No apparent enhancement was observed for monuron sorption on CaSWy-2 exchanged with KCl up to fK ¼ 0.71, whereas when K-SWy-2 underwent exchange with CaCl2, sorption was reduced by about half when the fractional K þ saturation decreased from 1 to 0.81. For dichlobenil, sorption by sorbents derived from Ca-SWy-2 increased by approximately 4 times as fK increased from zero to 0.71. Replacement of K þ from K-SWy-2 by Ca2 þ (fK ¼ 1–0.66) manifested gradually diminishing sorption of dichlobenil to 40% of that by homoionic K-SWy-2. Interestingly, dichlobenil sorption for clay derived from Ca-SWy-2 with fK ¼ 0.71 was substantially lower than the corresponding clay derived from K-SWy-2 with fK ¼ 0.66. The reason for such phenomena is a tendency for preservation of the original clay structures, namely, the smectite interlayer spacings associated with K- versus Casmectite as cation exchange (K þ ! Ca2 þ vs. Ca2 þ ! K þ ) proceeds, which can be viewed as a type of cation exchange hysteresis (Laird and Shang 1997; Laird 2006). Dichlobenil adsorbed much more strongly to K þ - rather than Ca2 þ -saturated smectite, so that as the K þ $ Ca2 þ exchange process occurs more favorable adsorption domains persist at a given fK when starting from the K þ -saturated end member.

67

Just as cation exchange processes on smectites alter adsorption of NOCs, changes in aqueous solution conditions such as ionic strength could also influence adsorption via their effects on clay quasicrystal structures. In aqueous solution, smectite clays are often present as quasicrystals consisting of stacks of clay platelets separated by interlayers filled with exchangeable cations and water molecules. Sorption of NOCs must effectively compete with interlayer water molecules to access adsorption domains on clay siloxane surfaces. Increasing aqueous salt concentration can promote the formation of clay aggregates, and reduce the amount of water between clay layers, thereby manifesting smaller interlayer distances in the formed quasicrystals and facilitating intercalation of NOCs. For example, sorption of 1,3DNB by K-SWy-2 increased with increasing KCl aqueous concentration (Fig. 2.18). At a relative 1,3-DNB aqueous concentration of 0.05, sorption by K-SWy-2 increased approximately 1.4, 1.7, 2.0 and 2.2 times as the KCl concentration increased from 0.01 M to 0.05, 0.10, 0.20, and 0.30 M, respectively (Li et al. 2007). Similar results were also observed for pesticide adsorption (Li et al. 2006). The small reduction in 1,3-DNB solubility due to the “salting out” effect was shown to be incapable of causing such a large increase in 1,3-DNB sorption by K-SWy-2. X-ray diffraction patterns and light absorbance of K-clay suspensions indicated the aggregation of clay particles and the formation of quasicrystal structures as KCl ionic strength increased, which is believed to be responsible for the enhanced adsorption of NACs and pesticides at higher ionic strength. For adsorption of NOCs by smectites, the type of exchangeable cation is the primary determinant of the size of sorptive domains in the clay galleries, clay interlayer distances, and the formation of complexes with NOCs. Cation exchange reaction (K þ $ Ca2 þ ) on smectite generates a range of K þ -saturated fractions or domains in the clay, 50000 Sorbed Concentration (mg kg-1)

Relative Sorption Coefficient

ENVIRONMENTAL IMPLICATIONS

0.30 M KCl 0.20 M KCl

40000

0.10 M KCl 0.05 M KCl

30000 0.01 M KCl

20000

10000

0 0

10

20 30 40 50 60 70 Aqueous Equilibrium Concentration (mg L-1)

Figure 2.18. Sorption isotherms of 1,3-dinitrobenzene (1,3-DNB) by K-SWy-2 from 0.01, 0.05, 0.10, 0.20, and 0.30 M KCl aqueous solutions [from Li et al. (2007)].

68


resulting in enhanced or reduced NOC adsorption (Chatterjee et al. 2008; Li et al. 2004b). Such simple cation exchange processes could be used as geochemical modulator in the development of environmentally friendly protocols to control the sorption, mobility, and bioavailability of NOCs in smectitic soils or soils amended with smectite clays. For example, Roberts et al. (2007a,2007b) demonstrated the successful application of K þ /Ca2 þ exchange reactions on smectite clays as a means to modulate the retention and release of NACs, and hence their toxicity to plants. Ionic strength is another plausible way to control clay interlayer environment, thereby modulating the degree of NOCs sorption/desorption in environmental systems. These simple geochemical controls on the adsorption/desorption of toxic NOCs could be used in bioremediation/phytoremediation to modulate the bioavailability and toxicity of NOCs to microorganisms and plants. 2.8. SUMMARY Sorption of NOCs by smectites can assume values along a continuum from zero up to 100 g NOC per kg clay. The critical factors that control sorption are a complex interplay among the functional groups of the NOC, the layer charge of the clay mineral, and the hydration of interlayer cations (Fig. 2.14): 1. The clay layer charge plays an important role because it controls the lateral distance between strongly hydrated interlayer cations. While “hydrated radii” are poorly defined, typical estimates of hydrated radii are 9.6 A for Ca2þ , 5.3 A for K þ , and 3.6 A for Cs þ (Evangelou 1998). Thus the cross-sectional area of hydrated Ca2þ is about 2.9 nm2, more than 7 times that of Cs þ (0.4 nm2), so one hydrated Ca2þ takes up 3.5 times as much surface area in the clay interlayer as do the two Cs þ ions that Ca2þ replaces. Using these hydrated radii along with an idealized smectite basal surface area of 750 m2/g and a 95 cmol/kg CEC, we compute that the hydrated radii of Ca2þ or K þ must overlap in the interlayer regions, meaning that the lateral adsorption domains (Fig. 2.4) are very small and the NOC would have to compete with strongly bound water for interlayer sorption sites. On the other hand, Cs þ and its hydration shells are projected to occupy only about 67% of the interlayer space; that is, about one-third of the surface area in a Cs-smectite may consist of lateral sorption domains (Fig. 2.4), implying that NOCs need compete only with weakly bound water in the Cs-smectite interlayers. This may explain why the Cs-smectite is the more effective adsorbent for essentially all NOCs studied. This idea of lateral domains is supported by several observations

(Lee et al. 1990; Jaynes and Boyd 1991b; Laird et al. 1992; Sheng et al. 2001, 2002) of an inverse relationship between the CEC of clays and the amount of the organic solute adsorbed, since fewer interlayer cations per unit surface area mean larger lateral adsorption domains available for organic solutes (Fig. 2.4). When layer charge is quite large, adsorption of even polar NOCs by smectites can be sharply reduced, because hydration radii for the greater number of exchangeable cations begin to overlap. In contrast, even Ca-smectites with low layer charges exhibit a strong sorption of atrazine (Laird et al. 1992). 2. As the organic functional group becomes more polar, the ability of the NOC to displace water from clay interlayers becomes stronger, apparently due to enhanced inner- and/or outer-sphere complexation between the NOC and interlayer cations, resulting in increased adsorption. Thus, pesticides with multiple strongly polar functional groups like ---NO2, ---C¼O or ---C N exhibit a strong sorption to K-smectites (Sheng et al. 2001; Boyd et al. 2001), even though the hydrated radii of the K þ ions are expected to overlap. Such polar functional groups apparently allow the NOC to form pesticide–cation complexes that are strong enough to displace water from the hydration shells of interlayer cations. 3. Compounds with functional groups that are less polar (e.g., atrazine, TCE, dioxin) are less strongly bound by K-smectites but can still be strongly adsorbed by Cssmectites, apparently because these NOCs are still able to displace the more weakly bound interlayer water found in the adsorptive domains (Fig. 2.4) between the hydrated Cs ions. The propensity for less polar NOCs to occupy these sites is further enhanced when the clay basal spacing is optimized at approximately 12.5 A, as in the cases of Cs- and certain K-smectites, as discussed above. In these cases, the NOCs can be mostly dehydrated, and a hydrophobic component contributes to NOC adsorption by smectites (Li et al. 2004a; Chappell et al. 2005). In this context, the site of negative charge on the clay can influence the retention of pesticide, presumably because tetrahedrally substituted smectites swell less in water and thereby contain narrower-slit pores that favor hydrophobic sorption (Aggarwal et al. 2006a,b). Additionally, the negative charges are more localized in tetrahedrally substituted smectites affording more neutral (hydrophobic) siloxane surface area for NOC adsorption. Apparently, then, an optimal inorganic sorbent for NOCs should be a Cs þ -saturated smectite with a low layer charge resulting from tetrahedral substitution. These criteria maximize adsorption domains parallel to the clay surfaces while

REFERENCES

optimizing (near 12.5 A) the adsorption domains perpendicular to the clay surfaces. Such clays may adsorb 10% of their weight for an NOC such as TNB with multiple, strongly complexing functional groups. In contrast, such clays adsorb some 1% by weight of more hydrophobic NOCs of lesser ability to form complexes with interlayer cations.

ACKNOWLEDGMENTS This project was supported by grant P42 ES004911 from the National Institute of Environmental Health Science (NIEHS), National Institute of Health (NIH), and National Research Initiative Competitive Grants from the USDA National Institute of Food and Agriculture. The contents are solely the responsibility of the authors and do not necessarily represent the official views of these federal agencies. REFERENCES Aggarwal, V., Li, H., and Teppen, B. J. (2006a), Triazine adsorption by saponite and beidellite clay minerals, Environ. Toxic. Chem. 25, 392–399. Aggarwal, V., Li, H., Boyd, S. A., and Teppen, B. J. (2006b), Enhanced sorption of trichloroethene by smectite clay exchanged with Cs þ , Environ. Sci. Technol. 40, 894–899. Ahmad, I., Dines, T. J., Rochester, C. H., and Anderson, J. A. (1996), IR study of nitrotoluene adsorption on oxide surfaces, J. Chem. Soc. Faraday Trans. 92, 3225–3231. Bailey, G. W. and White, J. L. (1970), Factors influencing the adsorption, desorption, and movement of pesticides in soil, in Residue Reviews: Residues of Pesticides and Other Foreign Chemicals in Foods and Feeds, Gunther, F. A. and Gunthereds J. D. eds., Vol. 32, Springer-Verlag, , New York, PP. 29–92. Baitinger, W., Schleyer, P. V. R., Murty, T. S. S. R., and Robinson, L. (1964), Nitro groups as proton acceptors in hydrogen bonding, Tetrahedron 20, 1635–1647. Barriuso, E., Laird, D. A., Koskinen, W. C., and Dowdy, R. H. (1994), Atrazine desorption from smectites, Soil Sci. Soc. Am. J. 58, 1632–1638. Borek, F. (1963), Effect of p-CH2X substituents on vibrational frequencies of the aromatic nitro group, Naturwissenschaften 50, 471–472. Boyd, S. A., Mortland, M. M., and Chiou, C. T. (1988b), Sorption characteristics of organic compounds on hexadecyltrimethylammonium smectite, Soil Sci. Soc. Am. J. 52, 652–657. Boyd, S. A., Lee, J. F., and Mortland, M. M. (1988c), Attenuating organic contaminant mobility by soil modification, Nature 333, 345–347. Boyd, S. A., Sun, S., Lee, J. F., and Mortland, M. M. (1988a), Pentachlorophenol sorption by organoclays, Clays Clay Miner. 36, 125–130. Boyd, S. A., Sheng, G., Teppen, B. J., and Johnston, C. T. (2001), Mechanisms for the adsorption of substituted nitro-

69

benzenes by smectite clays, Environ Sci. Technol. 35, 4227–4234. Celis, R., Cox, M. C., Hermosin, M. C., and Cornejo, J. (1997), Sorption of thiazafluron by iron - and humic acid-coated montmorillonite, J. Environ. Qual. 26, 472–479. Chappell, M. A., Laird, D. A., Thompson, M. L., Li, H., Teppen, B. J., Aggarwal, V., Johnston, C. T., and Boyd, S. A. (2005), Influence of smectite hydration and swelling on atrazine sorption behavior, Environ. Sci. Technol. 39, 3150–3156. Charles, S. M., Li, H., Teppen, B. J., and Boyd, S. A. (2006a), Quantifying the availability of clay surfaces in soils for adsorption of organic contaminants and pesticides, Environ. Sci. Technol. 40, 7751–7756. Charles, S. M., Teppen, B. J., Li, H., Laird, D. A., and Boyd, S. A. (2006b), Exchangeable cation hydration properties strongly influence soil sorption of nitroaromatic compounds, Soil Sci. Soc. Am. J. 70, 1470–1479. Charles, S. M., Teppen, B. J., Li, H., and Boyd, S. A. (2008), Fractional availability of smectite surfaces in soils for adsorption of nitroaromatic compounds in relation to soil and solute properties, Soil Sci. Soc. Am. J. 72, 586–594. Chatterjee, R., Laird, D. A., and Thompson, M. L. (2008), Interactions among K þ -Ca2 þ exchange, sorption of m-dinitrobenzene, and smectite quasicrystal dynamics, Environ. Sci. Technol. 42, 9099–9103. Chiou, C. T. (2002), Partition and Adsorption of Organic Contaminants in Environmental Systems, Wiley, Hoboken, NJ. Chiou, C. T. and Shoup, T. D. (1985), Soil sorption of organic vapors and effects of humidity on sorptive mechanism and capacity, Environ. Sci. Technol. 19, 1196–1200. Chiou, C. T., Peters, L. J., and Freed, V. H. (1979), A physical concept of soil-water equilibria for nonionic organic compounds, Science 206, 831–832. Chiou, C. T., Porter, P. E., and Schmedding, D. W. (1983), Partition equilibria of nonionic organic compounds between soil organic matter and water, Environ. Sci. Technol. 17, 227–231. Chiou, C. T., Kile, D. E., Rutherford, D. W., Sheng, G., and Boyd, S. A. (2000), Sorption of selected organic compounds from water to a peat soil and its humic-acid and humin fractions: Potential sources of the sorption nonlinearity, Environ. Sci. Technol. 34, 1254–1258. Conduit, C. P. (1959), Ultraviolet and infrared spectra of some aromatic nitro-compounds, J. Chem. Soc. 3273–3277. Evangelou, V. P. (1998). Environmental Soil and Water Chemistry: Principles and Applications, Wiley, New York. Friedman, H. L. and Krishnan, C. V. (1973), Thermodynamics of ionic hydration, in Water: A Comprehensive Treatise, Vol. 3, Aqueous Solutions of Simple Electrolytes, Franks, F., ed, Plenum, New York, PP. 1–118. Frisch, M. J., Trucks, G. W., et al. (1998), Gaussian 98, Revision A.9, Gaussian, Inc., Pittsburgh, PA. Fusi, P., Ristori, G. G., and Franci, M. (1982), Adsorption and catalytic decomposition of 4-nitrobenzenesulphonylmethlycarbamate by smectite, Clays Clay Miner. 30, 306–309.

70


Gaines, R. V., Skinner, H. C. W., Foord, E. E., Mason, B., and Rosenzweig, A. (1997), Dana’s New Mineralogy, Wiley, New York. Green, J. H. S. and Lauwers, H. A. (1971), Vibrational spectra of benzene derivatives — XIII The nitrobenzenes, Spectrochim. Acta. 27A, 817–824. Green, R. E. (1974), Pesticide-clay-water interactions, in Pesticides in Soil and Water, Guenzi, W. D. ed., Soil Science Society of America, Madison, WI PP. 3–37. Grundl, T. and Small, G. (1993), Mineral contributions to atrazine and alachlor sorption in soil mixtures of variable organic carbon and clay content, J. Contam. Hydrol. 14, 117–128. Haderlein, S. B. and Schwarzenbach, R. P. (1993), Adsorption of substituted nitrobenzenes and nitrophenols to mineral surface, Environ. Sci. Technol. 27, 316–326. Haderlein, S. B., Weissmahr, K. W., and Schwarzenbach, R. P. (1996), Specific adsorption of nitroaromatic explosives and pesticides to clay minerals, Environ. Sci. Technol. 30, 612–622. Hassett, J. J., Banwart, W. L., Wood, S. G., and Means, J. C. (1981), Sorption of a-naphthol: Implications concerning the limits of hydrophobic sorption, Soil Sci. Soc. Am. J. 45, 38–42. Jaynes, W. F. and Boyd, S. A. (1990), Trimethylphenylammoniumsmectite as an effective adsorbent of water soluble aromatic hydrocarbons, J. Air Waste Manage. Assoc. 40, 1649–1653. Jaynes, W. F. and Boyd, S. A. (1991a), Clay mineral type and organic compound sorption by hexadecyltrimethylammoniumexchanged clays, Soil Sci. Soc. Am. J. 55, 43–48. Jaynes, W. F. and Boyd, S. A. (1991b), Hydrophobicity of siloxane surface in smectites as revealed by aromatic hydrocarbon adsorption from water, Clays Clay Miner. 39, 428–436. Johnston, C. T. and Sposito, G. (1987), Disorder and early sorrow: Progress in the chemical speciation of soil surfaces, in Future Developments in Soil Science Research. Boersma, L. L. ed., Soil Science Society of America, Madison, WI, PP. 89–100. Johnston, C. T. and Premachandra, G. S. (2001), Polarized ATRFTIR study of smectite in aqueous suspension, Langmuir 17, 3712–3718. Johnston, C. T., Sposito, G., and Earl, W. L. (1993), Surface spectroscopy of environmental particles by Fourier transform infrared and nuclear magnetic resonance spectroscopy, in Environmental Particles, Vol. 2 in Environmental Analytical and Physical Chemistry Series, Buffle J. and van Leeuwen, H. P., eds., Lewis, Boca Raton, FL, PP. 1–36. Johnston, C. T., Boyd, S. A., Teppen, B. J., and Sheng, G. (2004), Sorption of nitroaromatic compounds on clay surfaces, in Handbook of Layered Materials, Auerbach, S. M., Carrado, K. A., and Dutta, P. K., eds., Marcel Dekker, New York, PP. 155–189. Johnston, C. T., De Oliveira, M. F., Teppen, B. J., Sheng, G., and Boyd, S. A. (2001), Spectroscopic study of nitroaromatic-smectite sorption mechanisms, Environ. Sci. Technol. 35, 4767–4772. Johnston, C. T., Sheng, G., Teppen, B. J., Boyd, S. A., and de Oliveira, M. F. (2002), Spectroscopic study of dinitrophenol herbicide sorption on smectite, Environ. Sci. Technol. 36, 5067–5074. Jonathan, N. B. H. (1960), Relations between force constants, bond orders, bond lengths, and bond frequencies for some nitrogenoxygen bonds, J. Molec. Spectrosc. 4, 75–83.

Karickhoff, S. W. (1984), Organic pollutant sorption in aquatic systems, J. Hydraul. Eng. 110, 707–735. Karickhoff, S. W., Brown, D. S., and Scott, T. A. (1979), Sorption of hydrophobic pollutants on natural sediments, Water Resour. Res. 13, 241–248. Kile, D. E., Chiou, C. T., Zhou, H., Li., H., and Xu, O. (1995), Partition of nonpolar organic pollutants from water to soil and sediment organic matters, Environ. Sci. Technol. 29, 1401–1406. Kukkadapu, R. K. and Boyd, S. A. (1995), Tetramethylphosphonium- and tetramethylammonium-smectites as adsorbents of aromatic and chlorinated hydrocarbons: Effect of water on adsorption efficiency, Clays Clay Miner. 43, 318–323. Laird, D. A. (2006), Influence of layer charge on swelling of smectites, Appl. Clay Sci. 34, 74–87. Laird, D. A. and Shang, C. (1997), Relationship between cation exchange selectivity and crystalline swelling in expanding 2:1 phyllosilicates, Clays Clay Miner. 45, 681–689. Laird, D. A., Shang, C., and Thompson, M. L. (1995), Hysteresis in crystalline swelling of smectites, J. Colloid Interface Sci. 171, 240–245. Laird, D. A., Barriuso, E., Dowdy, R. H., and Koskinen, W. C. (1992), Adsorption of atrazine on smectites, Soil Sci. Soc. Am. J. 56, 62–67. Laird, D. A., Yen, P. Y., Koskinen, W. C., Steinheimer, T. R., and Dowdy, R. H. (1994), Sorption of atrazine on soil clay components, Environ. Sci. Technol. 28, 1054–1061. Lawrence, M. A. M., Kukkadapu, R. K., and Boyd, S. A. (1998), Adsorption of phenol and chlorophenols by tetramethylammonium- and tetramethylphosphonium-exchanged montmorillonite, Appl. Clay Sci. 13, 13–20. Leboeuf, E. J. and Weber, W. J. (1997), A distributed reactivity model for sorption by soils and sediments. 8. Sorbent organic domains: Discovery of a humic acid glass transition and an argument for a polymer-based model, Environ. Sci. Technol. 31, 1697–1702. Lee, J. F., Crum, J., and Boyd, S. A. (1989), Enhanced retention of organic contaminants by soils exchanged with organic cations, Environ. Sci. Technol. 23, 1365–1372. Lee, J. F., Mortland, M. M., Chiou, C. T., Kile, D. E., and Boyd, S. A. (1990), Adsorption of benzene, toluene, and xylene by 2 tetramethylammonium-smectites having different charge-densities, Clay Clay Miner. 38, 113–120. Li, H., Teppen, B. J., Johnston, C. T., and Boyd, S. A. (2004a), Thermodynamics of nitroaromatic compound adsorption from water by smectite clay, Environ. Sci. Technol. 38, 5433–5442. Li, H., Teppen, B. J., Laird, D. A., Johnston, C. T., and Boyd, S. A. (2004b), Geochemical modulation of pesticide sorption on smectite clay, Environ. Sci. Technol. 38, 5393–5399. Li, H., Sheng, G., Teppen, B. J., Johnston, C. T., and Boyd, S. A. (2003), Sorption and desorption of pesticides by clay minerals and humic acid-clay complexes, Soil Sci. Soc. Am. J. 67, 122–131. Li, H., Teppen, B. J., Laird, D. A., Johnston, C. T., and Boyd, S. A. (2006), Effects of increasing potassium chloride and calcium chloride ionic strength on pesticide sorption by K- and Casmectite, Soil Sci. Soc. Am. J. 70, 1889–1895.

REFERENCES

Li, H., Pereira, T. R., Teppen, B. J., Laird, D. A., Johnston, C. T., and Boyd, S. A. (2007), Ionic strength-induced formation of smectite quasicrystals enhances nitroaromatic compound sorption, Environ. Sci. Technol. 41, 1251–1256. Liu, C., Li, H., Teppen, B. J., Johnston, C. T., and Boyd, S. A. (2009), Mechanisms associated with the high adsorption of dibenzo-pdioxin from water by smectite clays, Environ. Sci. Technol. 43, 2777–2783. MacEwan, D. M. C. and Wilson, M. J. (1980), Interlayer and intercalation complexes of clay minerals, in Crystal Structures of Clay Minerals and Their X-Ray Identification, Brindley, G. W. and Brown, G., eds., Mineralogical Society, London, PP. 197–248. Margulies, L., Rozen, H., and Banin, A. (1988), Use of X-ray powder diffraction and linear dichroism methods to study the orientation of montmorillonite clay particles, Clays Clay Miner. 36, 476–479. Mortland, M. M. (1970), Clay-oragnic complexes and interactions, Adv. Agron. 22, 75–117. Mortland, M. M. (1986), Mechanisms of adsorption of non-humic organic species by clay, in Interactions of Soil Minerals with Natural Organics and Microbes (special publication 17.), Huang, P. M. and Schritzer, M., eds., Soil Science Society of America, Madison, WI, PP. 59–76. Nyquist, R. A. and Settineri, S. E. (1990), Infrared study of substituted nitrobenzenes in carbon tetrachloride and chloroform solutions, Appl. Spectrosc. 44, 1552–1557. Onken, B. M. and Traina, S. J. (1997), The sorption of pyrene and anthracene to humic acid-mineral complexes: Effect of cosolute, J. Environ. Qual. 26, 132–138. Pennell, K. D., Boyd, S. A., and Abriola, L. M. (1995), Surface area of soil organic matter reexamined, Soil Sci. Soc. Am. J. 59, 1012–1018. Pereira, T. R., Laird, D. A., Johnston, C. T., Teppen, B. J., Li, H., and Boyd, S. A. (2007), Mechanism of dinitrophenol herbicide sorption on smectites in aqueous suspensions at varying pH, Soil Sci. Soc. Am. J. 71, 1476–1481. Pereira, T. R., Laird, D. A., Thompson, M. L., Johnston, C. T., Teppen, B. J., Li, H., and Boyd, S. A. (2008), Role of smectite quasicrystal dynamics in adsorption of dinitrophenol, Soil Sci. Soc. Am. J. 72, 347–354. Pignatello, J. J. and Xing, B. (1996), Mechanisms of slow sorption of organic chemicals to natural particles, Environ. Sci. Technol. 30, 1–11. Pils, J. R. V., Laird, D. A., and Evangelou, V. P. (2007), Role of cation demixing and quasicrystal formation and breakup on the stability of smectitic colloids, Appl. Clay Sci. 35, 201–211. Pusina, A., Liu, W., and Gessa, C. (1992). Influence of organic matter and its clay complexes on metolachlor adsorption on soil, Pesticide Sci. 36, 283–286. Ras, R. H. A., Schoonheydt, R. A., and Johnston, C. T. (2007), Relation between s-polarized and p-polarized internal reflection spectra: Application for the spectral resolution of perpendicular vibrational modes, J. Phys. Chem. A 111, 8787–8791.

71

Ras, R. H. A., Johnston, C. T., Franses, E. I., Ramaekers, R., Maes, G., Foubert, P., de Schryver, F. C., and Schoonheydt, R. A. (2003), Polarized infrared study of hybrid Langmuir-Blodgett monolayers containing clay mineral nanoparticles, Langmuir 19, 4295–4302. Roberts, M. G., Rugh, C. L., Li, H., Teppen, B. J., and Boyd, S. A. (2007a), Reducing bioavailability and phytotoxicity of 2,4dinitrotoluene by sorption on K-smectite clay, Environ. Toxic. Chem. 26, 358–360. Roberts, M. G., Rugh, C. L., Li, H., Teppen, B. J., and Boyd, S. A. (2007b), Geochemical modulation of bioavailability and toxicity of nitroaromatic compounds to aquatic plants, Environ. Sci. Technol. 41, 1641–1645. Saltzman, S. and Yariv, S. (1975), Infrared study of the sorption of phenol and p-nitrophenol by montmorillonite, Soil Sci. Soc. Am. J. 39, 474–479. Shang, C., Laird, D. A., and Thompson, M. L. (1995), Transmission X-ray diffraction technique for measuring crystalline swelling of smectites in electrolyte solutions, Clays Clay Miner. 43, 128–130. Sheng, G. and Boyd, S. A. (2000), Polarity effect on dichlorobenzene sorption by hexadecyltrimethylammonium-clays, Clays Clay Miner. 48, 43–50. Sheng, G., Johnston, C. T., Teppen, B. J., and Boyd, S. A. (2001), Potential contributions of smectite clays and organic matter to pesticide retention in soils, J. Agric. Food Chem. 49, 2899–2907. Sheng, G., Johnston, C. T., Teppen, B. J., and Boyd, S. A. (2002), Adsorption of dinitrophenol herbicides from water by montmorillonites, Clays Clay Miner. 50, 25–34. Teppen, B. J., Rasmussen, K., Bertsch, P. M., Miller, D. M., and Sch€afer, L. (1997), Molecular dynamics modeling of clay minerals. 1. Gibbsite, kaolinite, pyrophyllite, and beidellite, J. Phys. Chem. B 101, 1579–1587. Teppen, B. J., Yu, C.-H., Miller, D. M., and Sch€afer, L. (1998), Molecular dynamics simulations of the sorption of organic compounds at the clay mineral/aqueous solution interface, J. Comput. Chem. 19, 144–153. Theng, B. K. G. (1974), The Chemistry of Clay-Organic Reactions, Wiley, New York. Urbanski, T. and Dabrowska, U. (1959), The influence of the conjugation on the position of the infra-red band of the nitro group in some aromatic nitro-compounds, Bull. Acad. Polonaise Sci. 7, 235–237. Weissmahr, K. W., Haderlein, S. B., Schwarzenbach, R. P., Hany, R. and Nuesch, R. (1997), In situ Spectroscopic investigations of adsorption mechanisms of nitroaromatic compounds at clay minerals, Environ. Sci. Technol. 31, 240–247. Xia, G. S. and Ball, W. P. (2000), Polanyi-based models for the competitive sorption of low-polarity organic contaminants on a natural sorbent, Environ. Sci. Technol. 34, 1246–1254. Yariv, S., Russell, J. D., and Farmer, V. C. (1966), Infrared study of the adsorption of benzoic acid and nitrobenzene in montmorillonite, Israel. J. Chem. 4, 201–213.

3 THE ROLE OF ORGANIC MATTER–MINERAL INTERACTIONS IN THE SORPTION OF ORGANIC CONTAMINANTS MYRNA J. SIMPSON AND ANDRE J. SIMPSON 3.1. Introduction 3.2. Organic Matter–Mineral Interactions 3.3. The Role of Organic Matter–Mineral Interactions in Organic Contaminant Sorption 3.4. Summary and Synthesis of Future Research Directions

3.1. INTRODUCTION Organic matter (OM) is ubiquitously found in the environment and plays several, critical roles in environmental processes, such as OM cycling and turnover and the sorption of problematic organic chemicals (Feng et al. 2005; Kleber et al. 2007; Kang and Xing 2008). The role of OM in the sorption of organic chemicals has been an active area of research since the early 1960s. Early studies on the sorption of organic chemicals to soil recognized that OM is an important soil characteristic, especially for nonionic hydrophobic chemicals. For example, Lambert et al. (1965) proposed that there was an “active” fraction of soil OM that was responsible for sorption of chemicals to soil. Other notable studies suggested that the variation in diuron sorption coefficients was due to the accessibility of soil OM groups on colloidal surfaces (Hance 1965). Doherty and Warren (1969) later hypothesized that the relationship between herbicide binding and soil OM was governed by some other physical or chemical factor that soil OM itself was correlated. These studies (Doherty and Warren 1969; Hance 1965; Lambert et al. 1965), followed by subsequent studies (Bailey and White 1970; Hance 1969; Lambert 1967), concluded that soil

OM was playing a principal role in the sorption of chemicals to soil and defined future research directions (Fig. 3.1). Although these types of studies continued in the 1970s, studies on soil fractions became more prevalent (Fig. 3.1) with emphasis on defining the role of OM and later, OM structure in the sorption of organic chemicals. For example, Hayes (1970) demonstrated the value of employing chemical fractions in herbicide sorption studies and showed that the use of OM fractions had not yet been explored to its full potential. The work of Karickhoff et al. (1979) established a linear relationship between sorption partition coefficients (Kd values) with soil organic carbon contents. Hence, the practice arose of reporting organic carbon normalized (KOC) sorption coefficients in addition to other experimental parameters emerged. Empirical correlations between KOC values and the octanol–water partition coefficient (KOW) were also proposed (Chiou et al. 1979; 1982; 1983; 1985; Chiou 2002) and suggested that a soil or sediment Koc value could be predicted from the KOW value because sorption mechanisms were due mainly to phase partitioning. However, research showed that octanol is a poor surrogate for soil OM (Mingelgrin and Gerstl 1983; Xing et al. 1994a), and the emphasis to further understand how soil OM chemical structure governed the sorption of organic contaminants in soil was reinforced by these studies. Figure 3.1 depicts an overview of three main research “stages”; as mentioned previously, early studies with whole soils identified the importance of organic matter content in organic chemical sorption processes. Several studies showed that clay minerals, especially in the presence of water, did not sorb significant amounts of organic chemicals, thus, a strong focus on OM and OM fractions ensued. Researchers used


73

74

THE ROLE OF ORGANIC MATTER–MINERAL INTERACTIONS IN THE SORPTION OF ORGANIC CONTAMINANTS

Figure 3.1. An illustration of the progression of organic chemical sorption research since 1965. Early studies focused on the sorption of chemicals to soils (stage I) and propose mechanisms of sorption. With the onset of analytical techniques, the focus of the research shifted to interactions with soil components (stage II), which enabled researchers to link sorption behavior to structure. More recent studies have emphasized the importance of the relationship between OM and minerals on chemical sorption (stage III).

soil OM fractions [fulvic acid (FA), humic acid (HA), and humin], which are operationally defined and isolated using chemical fractionation, to gain insight into soil OM and contaminant interactions with the main objective of elucidating sorption mechanisms. Analytical instrumentation that were typically designed for simple mixtures and small molecules, often limited their application to “whole” soils. Therefore, using soil OM fractions as sorbents facilitated the study of macroscopic sorption mechanisms to individual soil OM fractions. Consequently, macroscopic studies with only soil OM chemical fractions began to emerge with the specific goal to develop soil OM structure–contaminant relationships for nonionic hydrophobic organic contaminants (Fig. 3.1). Interaction studies with charged and/or water-soluble organic chemicals were also conducted such that mechanisms could be defined and eventually predicted. Although studies with isolated OM or model compounds were critical for identifying and confirming the soil characteristics that governed organic chemical sorption processes, reports began to emerge showing that sorption processes were not regulated by OM content alone. For example, Garbarini and Lion (1986) reported that toluene and trichloroethylene sorption to FA, HA, and humin could not be explained by the organic carbon content alone. Their results suggested that oxygen content in addition to carbon content provides a more accurate prediction of toluene and trichloroethylene sorption (Garbarini and Lion 1986). In the late 1980s, researchers began to focus more on soil OM structure. Grathwohl (1990) demonstrated that samples of varying degrees of digenesis, namely, coals versus HA, produced varying Koc values for a series of chlorinated aliphatic hydrocarbons. A linear

relationship between log Koc values and the H/O atomic ratio further emphasized the importance of OM composition in addition to OM content. Numerous other studies have provided strong evidence for relationships between soil OM fraction structural parameters and Koc values (Chen et al. 1996; Xing 2001; Kang and Xing 2005). Studies with constructed OM–mineral complexes suggested that the mineral phase plays an indirect role by governing organic matter accessibility at the soil–water interface (Murphy et al. 1994; Jones and Tiller 1999). The evidence from these results combined with studies with soil OM fractions, which could not be explained by OM structure alone, resulted in more recent studies (represented by stage III in Fig. 3.1) whose sole focus is on examining the precise and indirect role of clay minerals and other inorganic soil constituents on controlling OM accessibility and reactivity at the soil–water interface. A large portion of OM in soil is associated with inorganic mineral phases, but only since the 1990s has it been acknowledged that this relationship can alter the environmental reactivity of OM (Baldock and Skjemstad 2000; Feng et al. 2005; Kleber et al. 2007; Kang and Xing 2008). This chapter focuses on recent research on OM–mineral interactions and their role in the sorption of nonionic hydrophobic organic contaminants. As indicated previously, studies of OM–mineral interactions have increased considerably since the 1990s but have primarily focused primarily on interactions with nonionic hydrophobic organic contaminants. This increase in research activity is not limited to organic chemicals but is conducted in parallel with research that aims to define the role of OM–mineral associations that may limit or slow the turnover of easily degradable OM

75

ORGANIC MATTER–MINERAL INTERACTIONS

(Baldock and Skjemstad 2000; Chenu and Plante 2006; Mikutta et al. 2007). Consequently, OM–mineral interaction studies exist well beyond the context of organic contaminant interactions and have been studied within the framework of soil biogeochemistry and soil OM responses to climate change. This chapter will first review OM–mineral studies followed by the role that OM–mineral complexes play in organic chemical sorption processes. The emphasis of the review will be on the use of molecule-level techniques, such as nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS), which are both used to investigate OM–mineral and organic contaminant interactions in detail. Readers who are interested in OM–mineral studies and their role in soil OM turnover are referred to several review papers [Baldock and Skjemstad (2000); Chenu and Plante (2006); Kleber et al. (2007); K€ ogel-Knabner et al. (2008), and references cited therein].

3.2. ORGANIC MATTER–MINERAL INTERACTIONS Organic matter–mineral interactions have been studied intensively since the 1990s because of their role in soil OM turnover, preservation of labile OM in sediments, and role in contaminant sorption processes (Arnarson and Keil 2000; Feng et al. 2005; Kleber et al. 2007; Kang and Xing 2008; K€ ogel-Knabner et al. 2008). Organic matter associations with mineral phases have resulted in reduced biodegradation of OM, and it is hypothesized that minerals provide OM both physical and chemical protection from biological attack (Baldock and Skjemstad 2000; Chenu and Plante 2006; Mikutta et al. 2007). It is believed that OM may be adsorbed to mineral surfaces via six mechanisms: ligand exchange, cation bridges (including water bridges), anion exchange, cation exchange, van der Waals interactions, and hydrophobic bonding (Arnarson and Keil 2000; Tombacz et al. 2004; Feng et al. 2005). Several researchers have shown that solution chemistry (ionic strength, pH, and dominant cation) determines the conformation of the OM prior to sorption on the mineral surface. Murphy et al. (1994) illustrated how solution conditions can promote OM to adopt a coiled or stretched conformation prior to sorption onto mineral surfaces (Fig. 3.2). High ionic strength and low pH values encourage dissolved OM to adopt a coiled or condensed configuration because negative charges of OM functional groups are neutralized by solution cations. A stretched or open structure results from charge repulsion under conditions of low ionic strength and high pH solution. Consequently, dissolved OM can acquire an open or closed geometry on the basis of the solution conditions prior to sorption onto the mineral surface, which, in turn, governs the physical accessibility to OM sorption domains (Jones and Tiller 1999; Murphy et al. 1994; Schlautman and Morgan 1993). Other

Humic substance in solution

Ligandexchange? Humic substance sorbed to surface

pH 6.5 Low I

pH 4.5 High I Low I Ca Na Ca Na Ca

Ligandexchange?

Na

Ca Ca Ca

Ternary Ligandsurface exchange? complex?

Ca

Ca Ca

Al- or Fe-Oxide Surface

Figure 3.2. Illustration of organic matter conformation with varying solution conditions; solution conditions (ionic strength) have been employed to induce a coiled or stretched OM structure prior to sorption to mineral surfaces [Reprinted with permission from Murphy et al. (1994)].

studies, which have explored the extent of OM sorption on mineral surfaces, have found OM sorption to increase with increasing ionic strength and decrease with increasing pH (Arnarson and Keil 2000; Baham and Sposito 1994; Satterberg et al. 2003). It has also been reported that di- and trivalent cations, namely, Ca2þ and Al3þ , enhance OM sorption by increasing the number of cation bridges (Arnarson and Keil 2000; Murphy et al. 1994; Schlautman and Morgan 1993). The concentration of dissolved OM also determines the mechanism by which OM–mineral phases form (Feng et al. 2005; Reiller et al. 2006). Mineral charge, surface area, and morphology will determine points of contact for OM and promote varying modes of OM sorption (Hur and Schlautman 2003; Meier et al. 1999; Feng et al. 2005; Kleber et al. 2007). For example, dissolved OM sorption by iron and aluminum oxides is believed to occur via ligand exchange, whereas cation bridges are believed to dominate in OM sorption to aluminosilicates (Chorover and Amistadi 2001; Zhou et al. 1994). Collectively, the aforementioned studies on OM–mineral interactions concluded that (1) sorption of dissolved OM to mineral surfaces is competitive and high-molecular-weight compounds are preferentially sorbed over low-molecularweight compounds, (2) aromatic moieties are preferentially adsorbed over aliphatic groups, (3) adsorption of OM increases inversely with pH, (4) the thickness of OM coatings on minerals varies with concentration, (5) adsorption to mineral surfaces is reversible, and (6) OM sorption is dependent on mineralogy (Arnarson and Keil 2000; Collins et al. 1995; Gu et al. 1996; Meier et al. 1999; Fein et al. 1999; Ochs et al. 1994; Theng 1982; Vermeer and Koopal 1998; Vermeer et al. 1998; Wershaw et al. 1996a,b). These conclusions are based on quantitative descriptors, such as thermodynamic data and isotherm shape, which were used to yield empirical relationships from which mechanistic information was inferred. Consequently, more recent studies have

76

THE ROLE OF ORGANIC MATTER–MINERAL INTERACTIONS IN THE SORPTION OF ORGANIC CONTAMINANTS

focused on OM–mineral interactions using advanced OM characterization methods in attempts to define the underlying fundamental chemical processes (Wattel-Koekkoek et al. 2001; Feng et al. 2005; 2006; Kang and Xing 2005; Wang and Xing 2005; Reiller et al. 2006, Simpson et al. 2006; Joo et al. 2008a; Kang and Xing 2008). For example, WattelKoekkoek et al. (2001) employed 13C NMR and pyrolysis– gas chromatography (GC)-MS to characterize OM associated with different mineral types and found that kaoliniteassociated OM was rich in polysaccharide products while smectite-associated OM contained more aromatic compounds, suggesting that binding mechanisms and the quality of sorbed OM are mineral-dependent. Advanced solution-state and high-resolution magic-angle spinning (HR-MAS) NMR methods have been used to study the competitive sorption of model OM mixtures (Simpson et al. 2006). High-resolution MAS NMR is a semisolid NMR technique that probes structures at the soil–water interface, and only structures in contact with the NMR solvent are observed (i.e., pure solid domains are not observed). The model compound mixture used was comprised of four main structural groups that are found in OM: 1-palmitoyl-3-stearoyl-rac-glycerol (to mimic large aliphatic molecules with functionalities similar to those present in plant cuticles), a small peptide (Arg–Pro–Leu–Glc–NH2), maltohexose (to represent carbohydrates), and the 500–1000-molecularweight fraction of commercially prepared lignin. The model compound mixture was prepared using an equal mass of each model compound and the mixture reacted with mont2-D NMR (TOCSY) of unsorbed compounds in solution

morillonite for 48 h. The unsorbed compounds in the mixture contained signals from all compounds; however, the HRMAS NMR results showed that only the aliphatic structures sorbed to the clay surface (Fig. 3.3). These results indicate that aliphatic structures, such as those found in plant cuticles, are preferentially sorbed to montmorillonite surfaces. The other OM model compounds displayed a low affinity for the montmorillonite surface and remained in solution (Fig. 3.3). A mixture of FA and HA was also sorbed to montmorillonite and similarly, it was reported that aliphatic components had a stronger affinity for the mineral surface than did other OM components. Although this study provided detailed, NMR evidence for the preferential sorption of OM components to montmorillonite, the emphasis on the novel development of NMR did not allow for other experimental variables (mineral type and/or solution conditions) to be explored. Feng et al. (2005) employed batch experiments with peat HA and kaolinite and montmorillonite under varying ionic strength, solution cation, and pH values. The authors employed both solution-state and HR-MAS NMR to study the composition of unsorbed and sorbed OM, respectively. Parameters of the OM sorption isotherms were also determined and used to quantify the variation in binding mechanisms (i.e., ligand exchange, cation bridging, van der Waals, and hydrophobic bonding) with varying solution conditions. Ionic strength, pH, and the dominance of either Ca2þ or Na þ determined the extent of OM sorption to mineral surfaces (kaolinite and montmorillonite) in this study. The sorption isotherms were modeled using the Freundlich HR-MAS NMR (TOCSY) of sorbed compounds on montmorillonite

Carbohydrate Aliphatic

Lignin Peptide Figure 3.3. Two-dimensional solution-state and HR-MAS NMR spectra of unsorbed and montmorillonite-sorbed OM components. The NMR results show that aliphatic structures are preferentially sorbed to the mineral surface. [Data from Simpson et al. (unpublished) with experimental details in Simpson et al. (2006)].

ORGANIC MATTER–MINERAL INTERACTIONS

2.50 Hydrophobic interactions

Freundlich Kf (mgC1-N lN g-1)

2.25 2.00

Ligand exchange van der Waals forces Cation bridging

1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00

4 7 4 7 4 7 7 4 pH pH pH pH pH pH a, pH a, pH a, a, a, a, a, a, C C N N C C N N 1 1 01 0.01 01 01 0.01 01 0.0 0.0 0.0 0.0 0.0 0.0

Equilibrium Solution Conditions

Figure 3.4. Freundlich coefficients and quantification of different binding mechanisms under varying equilibrium solution conditions (ionic strength, pH, and dominant cation) during the sorption of dissolved humic acid to mineral surfaces [modified from Feng et al. (2005)].

equation, and the resulting Kf values varied significantly with solution conditions (Fig. 3.4). In addition, solution conditions were also found to vary the mode of binding. For example, Na þ (I ¼ 0.001 M) weakens cation bridging and decreases in pH-promoted ligand exchange between dissolved OM and mineral surfaces (Fig. 3.4). As the ionic strength increases in the presence of Ca2þ , both cation bridging and van der Waals interactions were enhanced by the compression of the electric double layer. Ligand exchange was estimated to account for 32% of HA sorption on clay surfaces, van der Waals 22%, and cation bridges 41% when Ca2þ was the background electrolyte. Kaolinite displayed higher or similar adsorption for OM than montmorillonite when Na þ was the dominant cation present. Although it was expected that the higher-surface-area mineral (montmorillonite) would sorb more HA, kaolinite showed a higher affinity for HA when compared to montmorillonite (when normalized to external surface area only). However, it has been suggested that OM compounds do not intercalate with expanding minerals (Baham and Sposito 1994), although intercalation has been observed for a 625-Da surfactant and Ca-montmorillonite (Salloum et al. 2000). Zhou et al. (1994) also reported increased sorption of HA with kaolinite than montmorillonite in NaCl solutions. Montmorillonite has been observed to sorb highmolecular-weight OM compounds because of its high CEC and surface area (Satterberg et al. 2003), but Chorover and Amistadi (2001) showed that montmorillonite selectively sorbed low-molecular-weight OM in comparison with kao-

77

linite. The discrepancies in these results suggest that either different types of OM are sorbed under varying solution conditions and/or multiple layers of OM may be sorbed to varying extents, depending on the mineral type. Solution-state 1 H NMR spectra of the unbound HA and 1 H HR-MAS NMR spectra of HA sorbed to kaolinite and montmorillonite 0.01 M Na þ (pH 7) are shown in Figure 3.5 (Feng et al. 2005). Sorbed OM to both kaolinite and montmorillonite are rich in CH3 groups from short side-chains found in amino acids. In addition, the presence of aromatic signals (6.5–8 ppm) suggests the presence of peptide material in both samples, but the peptide signature is more prevalent on the surface of montmorillonite (Fig. 3.5). An estimate based on the intensity of the CH3 versus CH2 signals suggests that there is more peptide material sorbed to montmorillonite than to kaolinite. The CH2 groups in the 1 H HR-MAS spectra suggest that long-chain aliphatic structures preferentially bind to both minerals, but carbohydrates largely remained in the unbound fraction. The spectrum of HA-kaolinite is clearly dominated by CH2 signals (Fig. 3.5), suggesting that long-chain polymethylene structures are preferentially sorbed to kaolinite rather than montmorillonite. A study by Wang and Xing (2005) that also utilized NMR spectroscopy corroborated these findings, and the authors concluded that aliphatic components of HA are preferentially sorbed over aromatic structures on both kaolinite and montmorillonite surfaces. Although the studies by Feng et al. (2005) and Wang and Xing (2005) provided insight into the nature of sorbed OM on kaolinite and montmorillonite mineral surfaces, these studies did not address OM molecular-weight-fractionation with sorption. Moreover, neither of these studies addressed interactions with iron or aluminum oxides, which can dominate in some soils, and hence, the results with kaolinite and montmorillonite may not apply to other mineral types. As mentioned previously, some studies have shown that high-molecular-weight OM is preferentially sorbed to some mineral surfaces (Chorover and Amistadi 2001; Collins et al. 1995; Vermeer and Koopal 1998). For example, Chorover and Amistadi (2001) found that high-molecularweight OM preferentially sorbed onto goethite, whereas lowmolecular-weight OM preferentially sorbed to montmorillonite. Reiller et al. (2006) further examined HA sorption to hematite to test whether high- or low-molecular-weight OM constituents were preferentially sorbed. Compositional shifts of HA with sorption were determined using electrospray ionization (ESI) quadrupole time-of-flight (QTOF) MS, which provides accurate molecular mass information. The sorption of HA to hematite resulted in a decline of the lowmolecular-weight ( 2000

(Quantum dot)

(bulk)

LUMO CB Energy

Band gap

HOMO

The electrons in the filled VB can be promoted to the CB on excitation with photons carrying energy greater than Eg. Such a process yields CB electrons (eCB) and VB holes (hVB þ ) in equal amounts. Accordingly, the absorption of a semiconductor starts at the point where the photon energy is equal to that of the bandgap, and after that, rises steeply with increasing photon energies. This excitation process is known as band-to-band transition, and usually provides the most effective means of adsorption for the semiconductor. 4.2.2.4. Absorption of Other Species in the Environment 4.2.2.4.1. Color Centers. Most metal oxides (known as insulators) have very wide bandgaps. Their intrinsic band–band transition requires irradiation with far-UV light, and it is impossible for them to be excited by the incoming solar light. In many cases, however, the light can be absorbed by some intrinsic surface defects (Volodin 2000). Typically, primary light absorption in such systems is related to the O ! M charge transfer of an M¼O bond:

M

VB

Figure 4.2. A schematic representation of the formation of an energy band during the assembly of atoms into a crystal lattice, assuming that one atom contributes one orbital; N denotes the number of atoms (adapted from Hoffmann et al. (1995)].

O

hυ

δ−

M

O

δ+

ð4:1Þ

Because of the high energy necessary for the excitation of M¼O bonds, these systems exhibit absorption and photocatalytic activity only in the UV region. Many surface defects and coordinatively unsaturated structures can have primary light absorption in the visible region. It has been shown that irradiation of a wide-bandgap insulator material, (e.g., alkaline-earth metal oxides) with highly energetic radiation generates relatively stable surface

IMPORTANT CONCEPTS AND PROCESSES OF THE ENVIRONMENTAL PHOTOCHEMISTRY OF ORGANIC CONTAMINANTS

defects (the so-called electron F-type color centers and hole V-type color centers, corresponding to anion and cation vacancies, respectively), among which the F centers are more important for initiating the photochemical transformations. For example, on the loss of an O atom in a metal oxide, the 0 ~ 2 electrons that remain trapped in the oxygen vacancies can give rise to an F center (Serpone 2006). Many of the color centers are able to absorb visible light and initiate photochemical transformations of adsorbed molecules (Emeline et al. 1998a,b, 1999). 4.2.2.4.2. Charge Transfer Complex. Photoprocesses can be also induced from absorption of light by adsorption complexes formed on the surface of wide-bandgap oxides. These complexes are usually generated by the adsorption of donor molecules on strong surface acceptor sites, and the absorption of light by such complexes is attributed to the intermolecular electron transfer from donor to acceptor. This class of absorption can extend the photoresponse of the oxide in the visible or near-visible region (Volodin 2000). Frei and co-workers proposed that the hydrocarbon–O2 charge transfer complex generated inside the cavities of alkali or alkalineearth ion-exchanged zeolites can absorb the light in the region of red and near-IRlight [see Eq. (4.2)]. The highly polar charge transfer states from the hydrocarbon to the oxygen molecule, after stabilization by the large electrostatic field present in the zeolite channels, can lead to selective oxidation of olefins, alkyl-substituted benzenes and alkanes under visible light irradiation at room temperature (Blatter and Frei 1994; Blatter et al. 1998). Similar photoprocesses initiated by the visible light have been observed on Ba-exchanged zeolites of different structures (Myli et al. 1997). In close relation to this finding, Seo et al. (2005) revealed that polycyclic aromatic compounds also form visible CT complexes with dry TiO2: visible light

½RH--O2 Ð ½RH þ --O2

ð4:2Þ

A common characteristic for absorption of the abovementioned direct charge transfer complex is the existence of an electron donor (reductant) and an acceptor (oxidant), which are electronically coupled by a close contact. This close contact is frequently facilitated by the electrostatic attraction within an ion pair. The formation of a charge transfer complex has been shown to constitute a new and very promising class of photochemical system under light illumination. 4.2.2.4.3. Metal-Metal Charge Transfer. When widebandgap oxides contain transition metal ions, which are good candidates for environmentally applicable photocatalysts, the oxides usually exhibit absorption bands of transition metal ions. However, when two transition metals with different donor and acceptor abilities are

95

codoped into these oxides, metal-to-metal charge transfer (MMCT) may occur under visible irradiation, which has been shown to be a new class of visible-light-absorbing chromophores for photochemical reactions (Nakamura and Frei 2006; Weare et al. 2008) The MMCT of interest for the photocatalytic reaction included Zr(IV)--O--Cu(I) (Lin and Frei 2005b), Ti(IV)--O--Cu(I), and Ti(IV)--O--Sn(II) (Lin and Frei 2005a), Ti(IV)--O--Fe(II) (Xie et al. 2008), and Ti(IV)-O--Ce(III) (Nakamura et al. 2007). The visible absorption of these systems originates from the charge transition from the reducing metal sites (donors) to the adjacent oxidizing sites (acceptors), where they exhibit strong electronic coupling. The use of oxobridged heterobimetallic assembly as absorbing chromophores is a promising strategy for developing the visible-light photocatalysts, not only because of its flexible design and control of oxidation/ reduction (redox) potential and absorption wavelength but also because of its high durability for complete inorganic molecule-based photocatalysis (Nakamura et al. 2007). 4.2.3. Basic Photophysical and Photochemical Processes The absorption of a single photon is the only initiation event in photochemistry. After that, most of the absorbed photon energy is released as heat and light by the photophysical processes in all the abovementioned absorbing systems. Only a very small fraction of excited species can undergo chemical reactions. Considering that the reader can find a detailed description about these basic photophysical and photochemical processes in textbooks on photochemistry (Coyle 1991; Klessinger and Michl 1995), we give only a very brief introduction to them, focusing mainly on these relative processes in the following sections, such as emission of fluorescence or phosphorescence, the intersystem crossing process. 4.2.3.1. Photophysical Processes. Photophysical processes do not result in a chemical transformation of the substrates. Normally, the first event after absorption is to reach the zerovibration electronically first excited state S1 by the vibrational relaxation and/or by internal conversion from one of the vibrational levels of a higher electronic state such as S2. Figure 4.3 is a schematic representation of subsequent processes after the zero-vibration first excited state (1 AB ). The excited species (e.g., organic molecule, semiconductor, or metal complex) can return to its ground state AB by emission of fluorescence (EF), or it can reach the triplet state 3 AB by intersystem crossing (ISC), and after loss of excess vibrational energy it can return to the ground state AB by emission of phosphorescence (EP). Radiationless deactivation from 1 AB to ground state can occur via internal conversion (IC). Radiationless deactivation from 3AB to ground states can occur by intersystem crossing (ISC). It should be noted that the abovementioned processes usually are followed by a vibrational relaxation to the corresponding zero-vibration

96

PHOTOCATALYTIC DEGRADATION OF ORGANIC CONTAMINANTS ON MINERAL SURFACES

1

AB*

Photochemical process

Photophysical process EF

IC

C

ISC EP

AB

3

A++ B- (Heterolytic Cleavage)

AB+C* (Energy Transfer)

A•+B• (Homolytic Cleavage)

AB+•+ C-• or AB-• + C+• (Charge Transfer)

AB

ISC

AB++e- (Ionization) IC: Internal Conversion ISC: Intersystem Crossing EF: Emission of Fluorescence

ABC (Addition)

BA (Isomerization) AB-AB (Dimerization) Indirect Photoreaction

EP: Emission of Phosphorescence Direct Photoreaction

Figure 4.3. Primary photophysical and photochemical processes immediately after a species (AB) absorbs radiation to its first excited state (AB). The asterisk (‘ ’) indicates excited state. All physical and chemical processes are usually accompanied by vibrational relaxation, but are not labeled as such here.

states. Although the photophysical process cannot cause the degradation of the organic compounds, it provides many useful ways to observe and measure the excited-state species. For example, the most effective detection method of 1 O2 is to measure the phosphorescence at 1280 nm. 4.2.3.2. Primary Photochemical Process. Competing with the photophysical processes, the photochemical reaction may occur, in which new chemicals are formed. Only the first excited singlet states and the lowest triplet states have a chance of living long enough to participate in photochemical processes. The photochemical reaction can be divided into direct photolysis and indirect photolysis according to the species that absorb the light to initiate the reaction. Direct reaction occurs when the initially excited molecule undergoes a chemical reaction. In indirect photolysis a reaction is initiated through light absorption by some chromophores other than the substrate itself. In general, the photochemical reaction can be divided into two steps: 1. The first step includes photoinduced energy transfer, electron transfer, bond cleavage, and ionization. This step leads to the formation of highly reactive species (e.g., free radicals, singlet oxygen), which can initiate further reactions, which can be termed primary reactions. 2. The second step consists of secondary or “dark” reactions. In this process, a series of reactions (usually free-radical chain reactions) are involved, and these reactions end in the formation of stable products. It is worthwhile to bear in mind that primary reaction

products always undergo a rapid backward electron transfer. The formation of stable photoproducts depends on the competition between this backward electron transfer and secondary processes. A permanent chemical change takes place if these secondary processes are faster than the backward electron transfer. 4.2.3.2.1. Direct Photolysis. As mentioned above, the absorption of light by organic molecules usually is caused by the transition of the electron from the bonding orbital (such as p) or nonbinding orbital (such as n) to the antibonding orbital (such as p ). The result of this transition is that the molecule will become less stable, thus facilitating transformation into other compounds. From the energy perspective, absorption of photons at the UV–visible range increases the energy content of a molecule by 150–400 kJ/mol, which yields enough energy to produce homolytic or heterolytic breakages in the molecules. Depending on the properties of molecules in excited states (e.g., geometries, dipole moments, acidity and basicity), the excess energy can drive the molecule to undergo various chemical transformations, such as photodissociations (homolysis or heterolysis), photoionization, intramolecular rearrangement, isomerization, abstraction of a hydrogen atom, or dimerization (Fig. 4.3). All these reactions can be initiated for the singlet or triplet excited states of the molecule. The importance of each pathway in the photolysis of the organic pollutants depends on their capacity to absorb photons and to undergo chemical changes after light absorption (Edhlund et al. 2006). In the environment, the direct photodegradation of organic pollutants with short-wavelength UV radiation (e.g., most

IMPORTANT CONCEPTS AND PROCESSES OF THE ENVIRONMENTAL PHOTOCHEMISTRY OF ORGANIC CONTAMINANTS

pesticides) is expected to be, in general, of limited importance (Burrows et al. 2002). In the artificial system, the direct photolytic process has been employed for the degradation of organic contaminants in some studies (Legrini et al. 1993). 4.2.3.2.2. Indirect Photochemical Reactions. An indirect photochemical reaction occurs when energy or electron from the initially excited molecule (photosensitizer) is transferred to another molecule, and cause the latter to undergo a chemical reaction. Indirect processes are common in the natural environment and are especially important because they can alter molecules that resist photolysis or have poor absorption of sunlight (Zafiriou et al. 1984). The efficiency of these reactions depends on the number of reactive species produced by excitation of the chromophores as well as on the ability of these species reacting with the contaminant. Photosensitization can occur through either energy transfer or electron transfer pathways. The most widely studied and important energy transfer photosensitization pathway in environmental photochemistry is the activation of the triplet ground state 3 O2 to more reactive singlet oxygen (1 O2 ). Followed the adsorption of light by sensitizer, its triplet states may be generated by intersystem crossing. If the triplet sensitizer has enough energy and lifetime, it can interact with dissolved oxygen to form singlet oxygen. In natural waters, dissolved organic matter (i.e., humic substances or the synthetic or natural dyes) are the most dominant sensitizers to produce singlet oxygen (Boreen et al. 2004, 2005, 2008; Latch et al. 2003). The singlet oxygen, despite its rapid deactivation in water solution (due mostly to intersystem crossing to the triplet ground state), is able to degrade a variety of organic compounds. Photosensitization may also occur through an electron transfer pathway. In an excited state, a molecule can be a better oxidant or reductant than in its ground state. For instance, after highest occupied–lowest unoccupied molecular orbital (HOMO-LUMO) excitation, the half-occupied HOMO can readily accept another electron, while the single electron from the LUMO can readily be transferred to an acceptor, that is, the excited molecules become powerful electron donors or acceptors. The photoinduced electron transfer can lead to separation of the charge and can initiate a free-radical reaction. Therefore, it plays an important role in many fields such as photoelectric conversation and the transformation and removal of organic contaminants. 4.2.3.3. Introduction to Photocatalytic Reactions. Photocatalytic reaction is one kind of indirect photolysis. In such a process, a photocatalyst is introduced, which can accelerate the photochemical reaction, but remain chemically unchanged. The most interesting and widely studied photocatalytic systems are those using wideband semiconductors, particularly TiO2 as the photocatalyst, in which the

97

photochemical reaction takes place in the interfacial region between the semiconductor and the solution. According to the chromophores of light absorption in initiating the photocatalytic reaction, heterogeneous photocatalysis can be classified into the semiconductor-initiated photocatalysis and sensitizer-initiated photocatalysis. 4.2.3.3.1. Semiconductor-Initiated Photocatalysis. The initiation step in semiconductor-initiated photocatalytic reactions is absorption of photon with sufficient energy to induce band–band transition (Fig. 4.4a). After that, the photogenerated valence band hole (h þ ) and conduction band (electron) migrate to the semiconductor surface and are trapped in the subsurface and surface states of the semiconductor particle. Charge carrier trapping can suppress recombination and increase the lifetime of the separated electron and hole. The ability of a semiconductor to undergo photoinduced electron transfer to the adsorbed species on the surface adsorbate is thermodynamically governed by the band energy positions of the semiconductor and the redox potentials of the adsorbate. If a reduction of the species in the electrolyte is to be performed, the conduction band position of the semiconductor has to be positioned above the relevant redox level, and for the oxidation of adsorbate by holes, the potential level of the adsorbate must be above (more negative than) the valence band position of the semiconductor. The natural minerals with semiconductor properties usually are the metal oxide and metal sulfide minerals. Most metal oxide semiconductors have valence band edges 1–3 eV below the H2O oxidation potential, but the potentials of their conduction band edges are close to, or lower than, the H2O reduction potentials. Accordingly, these oxides are strong photooxidation catalysts in aqueous solutions, but are limited in their reducing power. For example, the valence band of the anatase TiO2 semiconductor has an oxidation potential of þ 2.7 V, which is a strong oxidizing agent, whereas its conduction band has a reduction potential of 0.5 V (a moderate reductant). For non-transition-metal sulfides, the valence band holes are less oxidative than the metal oxides, but their conduction band electrons are strongly reductive. Most transition metal sulfides, however, are characterized by small bandgaps ( cFe2O3 > c-FeOOH a-Fe2O3 Feaerosol > a-FeOOH. The halogenated acetic acid acts as an electron donor, and the relative rates of photooxidation of these acids with ferrihydrite (am-Fe2O33H2O) follow the order FCH2COOH > ClCH2COOH > BrCH2COOH > ICH2 COOH, FCH2COOH > F2CHCOOH > F3CCOOH. In terms of the observed strong kinetic isotope effects, Pehkonen and colleagues proposed that the bulk iron oxides phase is the principal chromophore, namely, that iron oxides serve as a photocatalyst in the oxidation of halogenated acetic acids. Photogenerated surface-bound hydroxyl radicals abstract hydrogen atoms from mono- and disubstituted haloacetic acids to yield haloacetate radicals, which in turn produce the corresponding halide and glycolic acid. For fully halogenated haloacetic acids, they seem to be oxidized via a photo-Kolbe-type reaction to yield the corresponding halo acids (i.e., HF, HCl, HBr) and CO2. 4.3.1.2.2. Phenols and Substituted Phenols. Several attempts have been made to explain the observed photodegradation of phenols (Chatterjee et al. 1994), aminophenols (Andreozzi et al. 2003), methylphenols (Mazellier and Bolte 2000), and chlorophenols (Bandara et al. 2001) by a band model. Andreozzi et al. (2003) investigated the photooxidation of 2-aminophenol by goethite at different pH values (3.0–8.0) and catalyst loading (100–500 mg/L). A negligible mineralization in the presence of O2 has been observed during the oxidation process. They ruled out the involvement of hydroxyl radicals in oxidation of 2-aminophenol. Mazellier and Bolte (2000) studied the light-induced transformation of 2,6-

dimethylphenol (DMP) in the system of UV/goethite/O2 and identified 2,6-dimethylbenzoquinone and 4,40 dihydroxy-3,30 ,5,50 -tetramethylbiphenyl (dimmer of DMP) as main products. They concluded that positive holes, rather than hydroxyl radicals, are responsible for DMP degradation, whereas conduction band electrons are trapped by oxygen adsorbed at surfaces. Bandara et al. (2001) observed that mono-, di-, and trichlorophenols can be efficiently degraded but partially mineralized on a-Fe2O3, and that degradation follows with pseudo-first-order kinetics. The overall photocatalytic degradation increases in the order 2,4,6trichlorophenol (2,4,6-TCP) < 2,3-dichlorophenol (2,3DCP) < 2-chlorophenol (2-CP) < 2,4-DCP. However, a-FeOOH is found to be active only for 2,4dichlorophenol (2,4-DCP). Like the process with TiO2 photocatalyst, photodegradation of chlorophenols on a-Fe2O3 is initiated from the para-hydroxylation of the parent compounds as suggested by the identified intermediates. 4.3.1.2.3. Dyes. Despite intensive investigations, there is still controversy concerning the mechanisms of dye degradation on iron oxides under visible-light irradiation. For instance, Orange II (Org II), a common toxic azo dye, is thought to be degraded through photosensitized and semiconductor photocatalysis pathways. The inherent difference between the two pathways (as described above) should lie in their excited chromophores; the former are dyes or dye–Fe surface complexes, but bulk catalyst for the latter one. Bandara et al. (1999) presented a series of spectroscopic evidences to support their photosensitized degradation concept. Org II can form a bridged bidentate complex with a-Fe2O3 via its sulfonic group, especially at acidic pH values (Bandara et al. 1999a). The complex [Org II a-Fe2O3] is the prerequisite for an efficient charge transfer process. The electron injection from excited dye to the conduction band of a-Fe2O3 proceeds within the duration of a laser pulse ( c-FeOOH > d-FeOOH > a-Fe2O3 > a-FeOOH > b-FeOOH. The authors attributed the rate differences to intrinsic diversities in crystal and surface structure rather than to the surface area or bandgap. This type of LMCT mechanism involving carboxylic groups has also been applied to explain photooxidation of sulfurbearing organic pollutants (Johansen and Key 2006). Methanesulfinic acid (MSIA, CH3SO2H) is a dimethylsulfide (DMS) oxidation intermediate. There exists structural similarity between carboxylic acid (--COOH) and the sulfinic acid (--SOOH) as center C and S atoms have almost identical electronegativities (2.55 and 2.58, respectively). Therefore, sulfinic group may be involved in the formation of a surface complex between Fe(III) and MSIA. Borer et al. (2007) reported the photodecomposition of citrate on c-FeOOH by in situ attenuated total reflection fourier transform infrared spectroscopy (ATR-FTIR). The primary photoproduct of citrate is acetonedicarboxylic acid generated from photodecarboxylation. The adsorbed acetonedicarboxylic acid is further decomposed to acetoacetate and even acetone as its final product at pH 4, but at pH 6, no further degradation of acetonedicarboxylic acid was observed. They attribute this type of pH-dependent selective photooxidation of the a-hydroxycarboxylic acid functional group of citrate to the different molar fractions of innersphere citrate surface complexes at pH 4 and pH 6 and their varied photochemical reactivities. Photolysis of the complexes of iron oxides with carboxylates on the addition or absence of H2O2 has been used to initiate the degradation of contaminant species (often via heterogeneous photo-Fenton processes). This method is particularly suitable for decomposition of some organic compounds that have low affinity to iron oxide such as diuron. Mazellier and Sulzberger (2001) observed the degradation of diuron in irradiated goethite/oxalate suspension at 3 pH 6. They attributed the degradation to attack by . OH radicals derived from photoreduced Fe(II) and H2O2. Concentrations of both oxalate and protons affected the rate of light-induced diuron transformation. The formation of Fe(II)(aq) is the rate-determining step for diuron degradation in these heterogeneous photo-Fenton systems. Liu et al. (2006), Li et al. (2007), and Lan et al. (2008) have studied a variety of synthetic iron oxides and organic

pollutants, including c-FeOOH, c-Fe2O3, a-Fe2O3, 2-mercaptobenzothiazole, bisphenol A, and pentachlorophenol (PCP). They reported an optimal initial concentration of oxalate (0.8 mmol L1) for degradation of PCP in UV/ c-Fe2O3/oxalate systems (Lan et al. 2008). Photoactive Fe(C2O4)2 and Fe(C2O4)33 are assigned to the dominant Fe(III)-oxalate species according to Fe(III) speciation calculation. Six intermediates are identified, including tetrachlorocatechol (TeCC), 2,3,5,6-tetrachloro-1,4-hydroquinone (TeCHQ), tetrachloro-o-benzoquinone (o-chloranil), and tetrachloro-p-benzoquinone (p-chloranil) formic and acetic acids. However, why higher concentrations of oxalate inhibited the rates of PCP degradation is still open to debate, although Li et al hypothesized that oxalate also react with hydroxyl radicals, thereby slowing down PCP transformation at higher oxalate concentration. Alternatively, there is another feasible explanation, namely, that Fe(III)-oxalate surface complexes convert from a bidentate, binuclear to a monodentate, binuclear structure at this threshold value of oxalate concentration (0.8 mmol/L) (Mazellier and Sulzberger 2001). He et al. (2002) observed that UV irradiation can significantly accelerate the degradation of an azo dye–Mardant Yellow 10 (MY10) on goethite. Photoreaction intermediates are confirmed, including acetic acid, nitrobenzene, and 4hydroxybenzenesulfonic acid by MS analysis and NO3 and SO42 by IC measurement. The authors proposed that photoinduced LMCT process of the surface complex of H2O2 with the oxide surface metal centers may be responsible for the MY10 degradation (Fig. 4.6). Briefly, the reaction is initiated by the formation of a precursor surface complex of H2O2 with surface Fe(III). It is followed by a cleavage of the O--O bond of the surface complex, leading to generation of a high-valence iron-oxo (ferryl) species and hydroxyl radical. This unstable ferryl species is able to react with H2O to give another . OH or directly oxidize adjacent pollutant molecules. They confirmed the . OH attack mechanism by an ESR trapping experiment. In a heterogeneous photo-Fenton system, the formation of a surface complex of Fe(III) with target pollutants, instead of H2O2, can also assist their decomposition and mineralization (He et al. 2005). Rates of degradation of aromatic compounds are related to these sorption capacities toward iron oxides: salicylic acid m-hydroxylbenzoic acid > p-hydroxylbenzoic acid benzoic acid > p-biphthalic acid > phenol > benzenesulfonic acid. 4.3.2. Iron-Bearing Layer Silicates Layer silicates are perhaps the most important and chemically active components of the clay mineral fraction. These natural minerals possess layered structures, large surface areas, and high cation exchange capacity (CEC). Smectites are hydrated 2 : 1-layer silicates composed of an (Fe, Al, Mg) octahedral sheet linked between two (Si, Al) tetrahedral

PHOTOSENSITIZED DEGRADATION OF ORGANIC POLLUTANTS ON TITANIUM DIOXIDE SURFACE UNDER VISIBLE IRRADIATION

FeIIIOOH UV H2O2

Degradation product

FeIIIOH

O=Fe IV Dye

HO

H2O

Figure 4.6. Cycling of Fe species coupled with dye degradation in the presence of hydrogen peroxide under UV light irradiation [stages: I—surface complex formation; II—light absorption; III— dissociation and decarboxylation of the oxidized oxalate; IV— detachment of surface Fe (II)] (He et al. 2002).

sheets by oxygen ligands (Thomas 2005). Most smectites contain appreciable amounts of octahedral iron in the ferric and/or ferrous form, and total iron content can be up to 25 wt% (Murad and Fischer 1988). On microbial or chemical reduction, Fe(II) complexes associated with clay minerals are of great significance for contaminant transformation. It is reported that the reduced smectites can act as a reducing agent for most of pesticides, chlorinated aliphatics, and nitroaromatics, eliminating either --Cl or --NO2. Stucki and his colleagues have examined the reactivities of reduced ferruginous smectites with respect to a range of pesticides, including atrazine, alachlor, trifluralin, oxamyl, chloropicrin, dicamba, and 2,4-D (Stucki 2005). These pesticides, with the exception of 2,4-D, can extensively react with clay surfaces. Their degradation products are observed but have not yet been fully identified. Dechlorination of chlorinated aliphatics (Lee and Batchelor 2004) and reduction of nitroaromatics (Hofstetter et al. 2003) have also been achieved by reaction with reduced smectites. Iron in clay minerals exists in very different chemical environments within the clay structure and at the mineral surface. Four different types of iron species may be present on clays (Hofstetter et al. 2003): (1) free iron oxides that distribute randomly on the clay surface, (2) structural iron that substitutes aluminum/silicon in the octahedral/tetrahedral lattice, (3) iron complexed by surface hydroxyl groups at edge surfaces, and (4) iron bound by ion exchange at basal siloxane surfaces. Free iron oxides in clays can be removed with citrate/bicarbonate/dithionite (CBD) extraction (Mehra and Jackson 1960). This CBD method makes it possible to compare the photoreactivities of different iron species

103

toward the same organic pollutant. Hofstetter et al. (2003) characterized the various Fe(II) species of reduced clays in terms of their accessibility to and reactivity for nitroaromatic compounds (NACs). Results reveal that interlayer exchanged Fe(II) does not contribute to the NAC reduction, but both edge-complexed and structural Fe(II) are effective in yielding the aniline product. However, on light irradiation, interlayer exchanged Fe(II) exhibit much higher reactivity for catalysis of the mineralization of malachite green (MG) than did structural Fe(II) (Cheng et al. 2008). Song et al. (2006) compared the photoreactivities of free oxides on clay surface and structural iron sandwiched in between two silica tetrahedral sheets. It was found that structural iron in the octahedral lattice inefficiently photocatalyze the decomposition of hydrogen peroxide under UV irradiation, but free oxides exhibit good reactivity. This difference results from the fact that direct photoexcitation cannot lead to the photoreduction of structural iron, which is unlike free oxides: .

FeIII OH þ hn ! FeII þ OH

ð4:12Þ

When photoreactive compounds such as N,N-dimethylaniline (DMA), rhodamine B (RhB), or malachite green (MG) intercalate into the clay layer, structural iron is able to greatly promote the decomposition of H2O2. According to the low redox potential of excited states of these compounds (2.85, 1.09, and 1.08 V vs. NHE for DMA, RhB, and MG, respectively), they can donate electron to structural iron (0.44 V vs. NHE) under light irradiation. Organics without absorption above 300 nm (such as cetyltrimethylammonium bromide) do not accelerate the reduction of H2O2, although it also exhibits high affinity toward clay particles. Therefore, the reduction of clay iron(III) to iron(II) can be achieved either by a light-induced LMCT process (for iron oxides) or electron injection of organic matters on irradiation (for structural iron) (Fig. 4.7).

4.4. PHOTOSENSITIZED DEGRADATION OF ORGANIC POLLUTANTS ON TITANIUM DIOXIDE SURFACE UNDER VISIBLE IRRADIATION Textile dyes and other industrial dyestuffs constitute one of the largest groups of anthropogenic organic compounds that represent an increasing environmental risk. There are more than 100,000 commercially available dyes with over 7 105 tons of dyestuff produced annually (Zollinger 1987). It has been estimated that about 10%–20% of the total world production of dyes is released into the environment from textile, paper, printing industries, and dye houses (Claus et al. 2002). These dyes have not only caused aesthetic problems but also exhibited great biotoxicity and possible mutagenic and carcinogenic effects (Chang et al. 2001). Most

104


Cr(VI) Photoreactive species*

Structural

FeIII / FeIII oxides

hv Photoreactive

species

Photoreactive

species+

Cr(III) UV irradiation

Structural

FeII / Surface

OH

FeII

H2O2

Figure 4.7. Proposed route of H2O2 decomposition catalyzed by iron oxides and structural iron in the presence of photoreactive substances [adapted from Song et al. (2006)].

commercial dyes are designed to resist photodegradation, and few of them can be removed by the applied aerobic microbial process of wastewater (Brown et al. 1981; Meehan et al. 2000; Robinson et al. 2001). Other traditional processes such as adsorption, chlorination, ozonation, and flocculation have been proved insufficient for treatment of these effluents. Therefore, it is urgent and important to exploit efficient methods to remove these dye pollutants. In this context, the self-sensitized degradation of dye pollutants in the presence of TiO2 under visible-light irradiation provides an efficient method for removal of these pollutants by using the solar light. In addition, the presence of photosensitized processes on the surface of semiconductor minerals can induce the oxidative or reductive removal of other organic pollutants under visible light. In the environment, the photosensitization reaction is also one of the primary pathways for the transformation of natural and synthetic dyes in the mineral surface. Thus, a detailed understanding of the photosensitized mechanism is expected to present useful information on the environmental fate and effect of these dye pollutants. Accordingly, the main focus of this section is on the photodegradation of organic pollutants by dye-sensitized TiO2 photocatalysis under visible-light irradiation. The conventional TiO2 photocatalytic degradation of organic contaminants (including dye pollutants) under UV irradiation is not included here, since there have been many extensive and excellent reviews on it since the early 1990s (Chen and Mao 2007; Hoffmann et al. 1995; Legrini et al. 1993; Thompson and Yates 2006). 4.4.1. Dye Adsorption Adsorption of the dye to the semiconductor surface is a prerequisite for photosensitization degradation, since the rapid electron injection from the dye to TiO2 requires strong interaction between dye and surface of TiO2. The TiO2assisted photodegradation of dye pollutants is thus a typical interfacial reaction that occurs just on the surface of TiO2 particles, other than in the bulk solution. The interaction extent and mode of the dyes on the surface of TiO2 particles

are important factors that govern the degradation rate and degradation mechanism. The interaction extent can be reflected by an adsorption isotherm of the dye on the TiO2, which can be estimated by measuring the amount of equilibrium adsorption of the dyes at various concentrations on a certain amount of photocatalyst particles. As shown in Equation (4.13), adsorption data can usually be processed in the form of a plot of amount of adsorbed dye (mad) versus the equilibrium concentration (C) according to the Langmuir adsorption isotherm C C 1 ¼ þ max mad mmax K m ad ad ad

ð4:13Þ

where C is the concentration of dye in the supernatant liquid at equilibrium, mad is the amount of dye adsorbed per unit mass of photocatalyst, mmax is the maximum adsorbed ad quantity per unit mass of photocatalyst, and Kad is the association constant of dye on photocatalyst; mmax ad and Kad can be determined from the slope and intercept of the linearity fitting, respectively. The adsorption isotherm can bear information on the total and macroscopic interaction between the surface and the dyes. To understand the interaction between the dye and surface sites at a molecular level, one must consider the adsorption mode of dye on the specific surface sites of photocatalyst. It is more difficult to determine the adsorption modes, which always involved several combined surface-characteristic techniques, such as UV–visible, FT-IR, Raman, and XPS spectroscopies. Complex and electrostatic interactions are the two most important interaction modes between the dye and surface of a photocatalyst. Complex interaction is the strong chemically binding mode between the functional groups of dyes and the surface metal sites of photocatalyst. In many photosensitized systems, the photosensitizers are linked to the surface by forming chemical complex to achieve efficient electron injection. In the Gr€atzel cell, for example, the photosensitizer usually anchors on the surface of TiO2 via carboxyl or phosphatic groups. For self-sensitized degradation, the dyes with strong complex group such as alizarin red generally can

PHOTOSENSITIZED DEGRADATION OF ORGANIC POLLUTANTS ON TITANIUM DIOXIDE SURFACE UNDER VISIBLE IRRADIATION

undergo rapid degradation (Liu et al. 2000b). The surface oxygen sites of TiO2 in aqueous media can undergo a series of protonization and deprotonization (acid–base) equilibria. As a result, a point of zero (surface) charge (PZC) is present with the change of the solution pH, where the surface sites with positive charge are equal to those with negative charge. Accordingly, the electrostatic interaction between the dye and surface depends greatly on the pH, the properties of surface, and the dyes. The PZC of TiO2 particles is at about pH 6.8 (Zhao et al. 1993). In general, the anionic dyes with sulfonate and carboxyl group manifest strong adsorption on the metal oxide at low pH, due to electrostatic interactions between the positive TiO2 surface and dye anions. For example, Bourikas et al. (2005) found that adsorption of Acid Orange 7 (AO7) on the TiO2 surface occurs to a significant extent only at pH values lower than 7, via the sulfonic group of the azo dye. Another anion dye amaranth with two sulfonate groups has been reported to be adsorbed on TiO2 via the sulfonate group located in the ortho position with respect to the OH group at natural pH (5.7) (Karkmaz et al. 2004). Accordingly, the anionic dye can be more rapidly degraded at low pH, whereas the degradation of cationic dyes needs a pH larger than the PZC of the photocatalyst. At pH greater than PZC, the cationic dyes (such as methylene blue and malachite green) are facilely interaction with the catalyst. Apart from the adjustment of pH to control the dye adsorption, the addition of other surface modifiers can also markedly influence the extent of dye adsorption on the photocatalyst surface. For example, addition of anionic surfactant DBS (sodium dodecylbenzene sulfonate) into the acidic TiO2 colloid can significantly enhance the adsorption of the cationic dye RhB on the TiO2 surface, due to the strong mutual interaction among dyes, DBS, and TiO2. As a result,

N

photoinduced electron transfer efficiency from dye molecules to the TiO2 particles was greatly enhanced after modification of the TiO2 surface by an anionic surfactant (Qu et al. 1998). The surface fluorination of TiO2 (F-TiO2) particles was observed to enhance the adsorption of cationic dyes, such as methylene blue, malachite green, and rhodamine (Rhodamine 6G and Rhodamine B) on the surface of photocatalyst, and their photocatalytic degradation rates are consequently greatly promoted after fluorination (Wang et al. 2008). On the contrary, if the dye (phenosafranin) is separated from the surface of TiO2 by capping the TiO2 with poly styrenesulfonate, the dye can be photostabilized by the inhibition of the photoinduced electron injection from the excited dye molecule (Ziolkowski et al. 1997) 4.4.2. General Performance for Photosensitized Degradation of Dye Pollutants Most of the dyes are relatively stable under visible-light irradiation in the absence of TiO2. In the presence of TiO2, the visible-light-induced electron transfer can cause the dyes to undergo significant degradation. Various dyes with different structure and functional groups (see Fig. 4.8 for structures of some representative dyes) have been examined for their selfsensitized degradation in the presence of TiO2 under visible irradiation. They include azo dyes Acid Orange 7 (Chen et al. 2004; Stylidi et al. 2004; Wang et al. 2004), ethyl orange (EO) (Zhao et al. 2003), xanthene dyes (including Eosin Y, rose bengal, Erythrosine B, Rhodamine B (Li et al. 2002; Zhao et al. 1998), Sulforhodamine B (Liu and Zhao 2000), triphenylmethane dye malachite green (Chen et al. 2002a), crystal violet and fuchsin basic (Li et al. 1999), anthraquinone dye alizarin red (Liu et al. 2000b), squarylium cyanine

O

N

N

O

N

Br O

SO3H

Br O

OH

Br

Br

COOH

COOH SO3H

(a)

(b)

(c)

N

N

O

OH

OH OH

N N

SO3H

SO3H O

(d)

105

(e)

(f)

Figure 4.8. Several representative dyes with different functional groups: (a) rhodamine B (RhB); (b) sulforhodamine B (SRB); (c) eosin; (d) malachite green (MG); (e) alizarin red (AR); (f) Acid Orange 7 (AO7).

106


dye (Wu et al. 1999a, 2000), basic blue dyes (Stathatos et al. 2001), diazo dye (naphthol blue-black) (Nasr et al. 1996), phenosafranin dyes (Ziolkowski et al. 1997), Active Brilliant Red Dye X-3B (Xie et al. 2005; Xu and Langford 2001), and Reactive Red 198 (Kaur and Singh 2007). Wu et al. (1999b) investigated the photodegradation of a series of dyes (Rhodamine B, Orange II, Sulfurhodamine B, fluorescein, alizarin red, squarylium cyanine, and eosin) in the presence of TiO2 particles under airequilibrated controlled conditions and visible light illumination. Epling and Lin (2002) examined the photosensitized degradation of 15 dyes with different functionalities. However, the thiazine dye methylene blue has been reported to be resistant to sensitized degradation under visible irradiation. Besides self-sensitized degradation pathways as mentioned above, sensitization processes can also lead to the oxidative degradation of other coexisting colorless organic pollutants via cationic radical of the sensitizer and the activeoxygen species formed during dye sensitization. Bendig and co-workers (Ross et al. 1994) found that irradiation of an aqueous solution of pollutant in the presence of TiO2 and rose bengal with visible light leads to decomposition of the herbicide terbutylazine, along with the photosensitized degradation of rose bengal. They further showed that the sensitized degradation of terbutryne on the TiO2 particles by tris (4,40 -dicarboxy-2,20 -bipyridyl)ruthenium(II) chloride complexes is more efficient (Lobedank et al. 1997). The selfsensitized degradation of xanthene dyes such as Eosin Y, rose bengal, eruthrosine, and Rhodamine B as sensitizers can result in the concomitant degradation of a small organic compound (2,4-dichlorophenol) under visible-light irradiation (Li et al. 2002). The sensitizers can be regenerated if the produced radical cation of the dye obtains an electron from the electron donor (organic pollutants). The dye-sensitized TiO2 can be considered as photocatalysts, which can degrade other organic pollutants under visible irradiation without the dyes being destroyed. Hodak et al. (1996) have reported the degradation of phenols, thiophenols, 4-chlorophenols, hydroquinones, and salicylic acid by a phthalocyanine dye (hydroxyaluminumtricarboxymonoamide phthalocyanine) sensitization of the TiO2 semiconductor particles, while EDTA, oxalic acid, and benzoquinone did not show any changes after irradiation. They proposed that the radical cation of the phthalocyanine produced by the electron injection of the dye into the conduction band of the TiO2 is the species responsible for the oxidation of the substrates (Hodak et al. 1996). Investigation of the photocatalytic activity of polycrystalline TiO2 samples sensitized by Cu(II)- or metal-free porphyrin dyes showed that the presence of the sensitizers is beneficial for the photoactivity of 4-nitrophenol (4-NP) photodegradation in aqueous suspension. A comparison with similar samples modified by sensitizer Cu(II)- and metal-free phthalocyanines showed that the presence of porphyrin is more

beneficial for both the decomposition rate of 4-nitrophenol and the disappearance of nonpurgeable organic carbon (NPOC) (Mele et al. 2003). Photodegradation of organic pollutants, such as phenol, chlorophenol, 1,2-dichloroethane, trichloroethylene, and pesticide (atrazine), in water has been achieved on the surface of TiO2 semiconductor modified with sensitizers thionine, Eosin Y, Rhodamine B, methylene blue, Nile Blue A, and Safranine O by using visible light. After 5 h of irradiation with a 50-W tungsten lamp, over 55%–72% degradation of pollutants is achieved for the thionine- and Eosin Y–sensitized systems (Chatterjee and Mahata 2001, 2004; Chatterjee et al. 2006). The activation of TiO2 photocatalyst for photocatalysis under visible light using Acid Red 44 (Moon et al. 2003) or poly(fluorene-co-thiophene) (PFT) (Song et al. 2007) has been investigated, and the decomposition of phenol can be realized on the dye-sensitized photocatalysis under visible-light irradiation There are also reports on the reduction of other pollutants by dye-sensitized TiO2 under visible-light irradiation. Evidently, the reduction reaction is induced from the conduction band electron injected from the excited dyes. For example, Choi and colleagues have developed several organic-dyesensitized TiO2 systems for the reductive dechlorination of the perchlorinated compounds such as CCl4 and CCl3CO2 and for the reduction of heavy-metal ions under visible irradiation (Cho and Choi 2002; Cho et al. 2001, 2004). They have investigated the photoreductive decomposition of CCl4 using visible light on TiO2 sensitized by tris-(4,40 dicarboxy-2, 20 -bipyridyl) ruthenium(II) complexes. It was found that the sensitized TiO2 could degrade CCl4 under the irradiation of > 420 nm with a quantum yield of about 103. The rate decreases in the presence of O2, due to competition for the conduction band electrons. 2-Propanol was used as an efficient sacrificial donor of electron to regenerate the oxidized sensitizer. They further found that the deposition of platinum nanoparticles on the dye-sensitized TiO2 drastically enhanced the reductive degradation of trichloroacetate and CCl4. In Pt/TiO2/dye system, the water molecules can act as electron donors to regenerate the sensitizer, with a concurrent production of dioxygen. Finally, it should be pointed out that, in order to clarify the role of visible light in the photocatalytic degradation of dyes and to investigate the mechanism of photosensitized degradation, only some of the studies that have employed visible irradiation to photocatalytically degrade the dyes are mentioned here. In these studies, ether the UV irradiation light is filtered out from a mixed light source or a monochromatic visible-light source (with wavelength longer than 420 nm) is used to avoid direct excitation of the semiconductor. There have also been many more studies using UV or UV-visible mixed light sources for the photocatalytic degradation of dye pollutant; these studies are not described here. In these systems, the pathways of semiconductor-initiated degradation and photosensitizer-initiated degradation are operative

REFERENCES

simultaneously. In fact, during the degradation of dye pollutants in the environment or in the photocatalytic reactor, UV and visible-light radiation pathways are usually coexistent. It is difficult to distinguish these two pathways from each other. However, it is certain that the existence of the photosensitizer-initiated pathway may improve the overall efficiency of the degradation of dye pollutants and the transformation of colorless pollutants in both the natural and photoreactors.

107

ACKNOWLEDGMENT Parts of the work described in Sections 3.3 and 3.4 were financially supported by Project 973 (of the Ministry of Science and Technology of China Grants 2007CB613306 and 1010CB933503), the National Science Foundation of China (Grants 20920102034, 20877076, and 20907056), and the Chinese Academy of Sciences.

4.5. CONCLUSIONS REFERENCES This chapter emphasizes the importance of photochemistry (particularly photocatalysis) on the environmental effect, transformation, and fate of anthropogenic organic compounds. Two typical environmentally relevant processes occurring on the surface of natural minerals are reviewed in detail: 1. Organic compounds with carboxylic and/or hydroxyl/ phenolic functional groups can readily adsorb to ironbearing solids and be oxidized through the pathways of semiconductor photocatalysis and/or photoinduced LMCT processes. However, on the basis of the available data, it is not possible yet to unambiguously distinguish which mechanism operates in the photodegradation of organic compounds at the iron oxide surface. The exact effect of these photocatalytic oxidation processes on the environmental fate of organic pollutant still needs further study. Another ubiquitous iron species in layered smectites also plays a critical role in the transformation of organic pollutants; however, its photoredox cycling scheme requires additional investigation. 2. Photosensitized degradation of organic pollutants, especially for dyestuffs, represents an important transformation pathway of anthropogenic organic compounds under visible-light irradiation. Although the laboratory studies have attempted to clarify the detailed mechanism of photosensitized degradation of dyes, there is still a dramatic lack of field data to evaluate its contribution to dye transformation in polluted soils and sediments. Another aspect that warrants further investigation is the coupling process and mechanism between the photocatalytic degradation of dye pollutants (or the natural chromogenic matters) with the transformation of other environmental species (such as colorless organic pollutants, toxic metals). Although the importance of these coupling processes has been recognized for a long time, systemic evaluation has been insufficient. The toxicity for intermediates and products during the photocatalytic degradation of different dye pollutants should also be evaluated critically.

Ahn, M. Y., Filley, T. R., Jafvert, C. T., Nies, L., Hua, I., and BezaresCruz, J. (2006), Photodegradation of decabromodiphenyl ether adsorbed onto clay minerals, metal oxides, and sediment, Environ. Sci. Technol. 40, 215–220. Andreozzi, R., Caprio, V., and Marotta, R. (2003), Iron(III) (hydro) oxide-mediated photooxidation of 2-aminophenol in aqueous solution: A kinetic study, Water Res. 37, 3682–3688. Bakardjieva, S., Stengl, V., Subrt, J., Houskova, V., and Kalenda, P. (2007), Photocatalytic efficiency of iron oxides: Degradation of 4-chlorophenol, J. Phys. Chem. Solids 68, 721–724. Balzani, V., Bergamini, G., Campagna, S., and Puntoriero, F. (2007), Photochemistry and photophysics of coordination compounds: Overview and general concepts, in Photochemistry and Photophysics of Coordination Compounds I, Balzani, V., and Campagna, S., eds., Springer-Verlag, Berlin/Heidelberg, pp. 1–36. Bandara, J., Mielczarski, J. A., and Kiwi, J. (1999a), 1. Molecular mechanism of surface recognition. Azo dyes degradation on Fe, Ti, and Al oxides through metal sulfonate complexes, Langmuir 15, 7670–7679. Bandara, J., Mielczarski, J. A., and Kiwi, J. (1999b). 2. Photosensitized degradation of azo dyes on Fe, Ti, and Al oxides. Mechanism of charge transfer during the degradation, Langmuir 15, 7680–7687. Bandara, J., Mielczarski, J. A., Lopez, A., and Kiwi, J. (2001), 2. Sensitized degradation of chlorophenols on iron oxides induced by visible light comparison with titanium oxide, Appl. Catal. B Environ. 34, 321–333. Blatter, F. and Frei, H. (1994), Selective photooxidation of small alkenes by O2 with red light in zeolite Y, J. Am. Chem. Soc. 116, 1812–1820. Blatter, F., Sun, H., Vasenkov, S., and Frei, H. (1998), Photocatalyzed oxidation in zeolite cages, Catal. Today 41, 297–309. Boreen, A. L., Arnold, W. A., and McNeill, K. (2004), Photochemical fate of sulfa drugs in the aquatic environment: Sulfa drugs containing five-membered heterocyclic groups, Environ. Sci. Technol. 38, 3933–3940. Boreen, A. L., Arnold, W. A., and McNeill, K. (2005), Tripletsensitized photodegradation of sulfa drugs containing six-membered heterocyclic groups: Identification of an SO2 extrusion photoproduct, Environ. Sci. Technol. 39, 3630–3638. Boreen, A. L., Edhlund, B. L., Cotner, J. B., and McNeill, K. (2008), Indirect photodegradation of dissolved free amino acids: The contribution of singlet oxygen and the differential reactivity

108


of DOM from various sources, Environ. Sci. Technol. 42, 5492–5498. Borer, P., Hug, S. J., Sulzberger, B., Kraemer, S. M., and Kretzschmar, R. (2007), Photolysis of citrate on the surface of lepidocrocite: An in situ attenuated total reflection infrared spectroscopy study, J. Phys. Chem. C 111, 10560–10569. Bourikas, K., Stylidi, M., Kondarides, D. I., and Verykios, X. E. (2005), Adsorption of acid orange 7 on the surface of titanium dioxide, Langmuir 21, 9222–9230. Brown, D., Hitz, H. R., and Sch€aer, L. (1981), The assessment of the possible inhibitory effect of dyestuffs on aerobic waste-water bacteria experience with a screening test, Chemosphere 10, 245–261. Brown, G. E., Henrich, V. E., Casey, W. H., Clark, D. L., Eggleston, C., Felmy, A., Goodman, D. W., Gratzel, M., Maciel, G., McCarthy, M. I., Nealson, K. H., Sverjensky, D. A., Toney, M. F., and Zachara, J. M. (1999), Metal oxide surfaces and their interactions with aqueous solutions and microbial organisms, Chem. Rev. 99, 77–174. Burrows, H. D., Canle, L. M., Santaballa, J. A., and Steenken, S. (2002), Reaction pathways and mechanisms of photodegradation of pesticides, J. Photochem. Photobiol. B Biol. 67, 71–108. Chang, J. -S., Chou, C., Lin, Y. -C., Lin, P. -J., Ho, J. -Y., and Hu, T. L. (2001), Kinetic characteristics of bacterial azo-dye decolorization by Pseudomonas luteola, Water Res. 35, 2841–2850. Chatterjee, D. and Mahata, A. (2004), Evidence of superoxide radical formation in the photodegradation of pesticide on the dye modified TiO2 surface using visible light, J. Photochem. Photobiol. A Chem. 165, 19–23. Chatterjee, D., Dasgupta, S., and Rao, N. N. (2006), Visible light assisted photodegradation of halocarbons on the dye modified TiO2 surface using visible light, Solar Energy Mater. Solar Cell 90, 1013–1020. Chatterjee, S., Sarkar, S., and Bhattacharyya, S. N. (1994), Photodegradation of phenol by visible light in the presence of colloidal Fe2O3, J. Photochem. Photobiol. A Chem. 81, 199–203. Chen, C. C., Li, X. Z., Ma, W. H., Zhao, J. C., Hidaka, H., and Serpone, N. (2002a). Effect of transition metal ions on the TiO2assisted photodegradation of dyes under visible irradiation: A probe for the interfacial electron transfer process and reaction mechanism, J. Phys. Chem. B 106, 318–324. Chen, C. C., Zhao, W., Li, J. Y., and Zhao, J. C. (2002b), Formation and identification of intermediates visible-light-assisted photodegradation sulforhodamine-B dye in aqueous TiO2 dispersion, Environ. Sci. Technol. 36, 3604–3611. Chen, X. and Mao, S. S. (2007), Titanium dioxide nanomaterials: Synthesis, properties, modifications, and applications, Chem. Rev. 107, 2891–2959. Chen, Y., Wang, K., and Lou, L. (2004), Photodegradation of dye pollutants on silica gel supported TiO2 particles under visible light irradiation, J. Photochem. Photobiol. A Chem. 163, 281–287. Cheng, M. M., Song, W. J., Ma, W. H., Chen, C. C., Zhao, J. C., Lin, J., and Zhu, H. Y. (2008), Catalytic activity of iron species in layered clays for photodegradation of organic dyes under visible irradiation, Appl. Catal. B Environ. 77, 355–363.

Cho, Y., Choi, W., Lee, C. H., Hyeon, T., and Lee, H. I. (2001), Visible light-induced degradation of carbon tetrachloride on dye-sensitized TiO2, Environ. Sci. Technol. 35, 966–970. Cho, Y. and Choi, W. (2002). Visible light-induced reactions of humic acids on TiO2, J. Photochem. Photobiol. A: Chem. 148, 129–135. Cho, Y., Kyung, H., and Choi, W. (2004), Visible light activity of TiO2 for the photoreduction of CCl4 and Cr(VI) in the presence of nonionic surfactant (Brij), Appl. Catal. B Environ. 52, 23–32. Ciani, A., Goss, K. U., and Schwarzenbach, R. P. (2005a), Light penetration in soil and particulate minerals, Eur. J. Soil Sci. 56, 561–574. Ciani, A., Goss, K. U., and Schwarzenbach, R. P. (2005b), Photodegradation of organic compounds adsorbed in porous mineral layers: Determination of quantum yields, Environ. Sci. Technol. 39, 6712–6720. Claus, H., Faber, G., and Konig, H. (2002), Redox-mediated decolorization of synthetic dyes by fungal laccases, Appl. Microbiol. Biotechnol. 59, 672–678. Cornell, R. M. and Schwertmann, U. (2003), The Iron Oxides: Structure, Properties, Reactions, Occurences and Uses, WileyVCH, Weinheim. Coyle, H. D. (1991), Introduction to Organic Photochemistry, Wiley, New York. Cunningham, K. M., Goldberg, M. C., and Weiner, E. R. (1985), The aqueous photolysis of ethylene glycol adsorbed on goethite, Photochem. Photobiol. 41, 409–416. Cunningham, K. M., Goldberg, M. C., and Weiner, E. R. (1988), Mechanisms for aqueous photolysis of adsorbed benzoate, oxalate, and succinate on iron oxyhydroxide (goethite) surfaces, Environ. Sci. Technol. 22, 1090–1097. Du, W. P., Xu, Y. M., and Wang, Y. S. (2008), Photoinduced degradation of Orange II on different iron (hydro)oxides in aqueous suspension: Rate enhancement on addition of hydrogen peroxide, silver nitrate, and sodium fluoride, Langmuir 24, 175–181. Edhlund, B. L., Arnold, W. A., and McNeill, K. (2006), Aquatic photochemistry of nitrofuran antibiotics, Environ. Sci. Technol. 40, 5422–5427. Emeline, A. V., Kataeva, G. V., Ryabchuk, V. K., and Serpone, N. (1999), Photostimulated generation of defects and surface reactions on a series of wide band gap metal-oxide solids, J. Phys. Chem. B 103, 9190–9199. Emeline, A. V., Kataeva, G. V., Litke, A. S., Rudakova, A. V., Ryabchuk, V. K., and Serpone, N. (1998a), Spectroscopic and photoluminescence studies of a wide band gap insulating material: Powdered and colloidal ZrO2 sols, Langmuir 14, 5011–5022. Emeline, A. V., Petrova, S. V., Ryabchuk, V. K., and Serpone, N. (1998b), Photochemical and photophysical processes on the surface of wide band gap insulator particulates: Gas/solid system involving scandia (Sc2O3) particles, Chem. Mater. 10, 3484–3491. Epling, G. A. and Lin, C. (2002), Photoassisted bleaching of dyes utilizing TiO2 and visible light, Chemosphere 46, 561–570.

REFERENCES

Guillard, C., Hoang-Van, C. F., Pichat, P., and Marme, F. (1995), Laboratory study of the respective roles of ferric oxide and released or added ferric ions in the photodegradation of oxalic acid in aerated liquid water, J. Photochem. Photobiol. A Chem. 89, 221–227. He, J., Ma, W. H., He, J. J., Zhao, J. C., and Yu, J. C. (2002), Photooxidation of azo dye in aqueous dispersions of H2O2/ a-FeOOH, Appl. Catal. B Environ. 39, 211–220. He, J., Ma, W. H., Song, W. J., Zhao, J. C., Qian, X. H., Zhang, S. B., and Yu, J. C. (2005), Photoreaction of aromatic compounds at a-FeOOH/H2O interface in the presence of H2O2: Evidence for organic-goethite surface complex formation, Water Res. 39, 119–128. Hodak, J., Quinteros, C., Litter, M. I., and Roman, E. S. (1996), Sensitization of TiO2 with phthalocyanines. Part 1. Photo-oxidations using hydroxoaluminium tricarboxymonoamidephthalocyanine adsorbed on TiO2, J. Chem. Soc. Faraday Trans. 92, 5081–5088. Hoffmann, M. R., Martin, S. T., Choi, W. Y., and Bahnemann, D. W. (1995), Environmental applications of semiconductor photocatalysis, Chem. Rev. 95, 69–96. Hofstetter, T. B., Schwarzenbach, R. P., and Haderlein, S. B. (2003), Reactivity of Fe(II) species associated with clay minerals, Environ. Sci. Technol. 37, 519–528. Huang, P. (2004), Soil mineral-organic matter-microorganism interactions: Fundamental and impacts, Adv. Agron. 82, 391–472. Johansen, A. M. and Key, J. M. (2006), Photoreductive dissolution of ferrihydrite by methanesulfinic acid: Evidence of a direct link between dimethylsulfide and iron-bioavailability, Geophys. Res. Lett. 33, L14818. Karametaxas, G., Hug, S. J., and Sulzberger, B. (1995), Photodegradation of EDTA in the presence of lepidocrocite, Environ. Sci. Technol. 29, 2992–3000. Karkmaz, M., Puzenat, E., Guillard, C., and Herrmann, J. M. (2004), Photocatalytic degradation of the alimentary azo dye amaranth: Mineralization of the azo group to nitrogen, Appl. Catal. B Environ. 51, 183–194. Kaur, S. and Singh, V. (2007), Visible light induced sonophotocatalytic degradation of Reactive Red dye 198 using dye sensitized TiO2, Ultrason. Sonochem. 14, 531–537. Kiwi, J. and Gr€atzel, M. (1987), Light-induced hydrogen formation and photo-uptake of oxygen in colloidal suspensions of a-Fe2O3, J. Chem. Soc. Faraday Trans. 83, 1101–1108. Klessinger, M. and Michl, J. (1995), Excited States and Photochemisty of Organic Molecule, VCH, Weinheim. Kuramoto, H. (1996), Disulfide oligomer on a-FeOOH adsorption state and photoadsorption, Langmuir 11, 3417–3422. Lan, Q., Li, F. B., Liu, C. S., and Li, X. Z. (2008), Heterogeneous photodegradation of pentachlorophenol with maghemite and oxalate under UV illumination, Environ. Sci. Technol. 42, 7918–7923. Latch, D. E., Stender, B. L., Packer, J. L., Arnold, W. A., and McNeill, K. (2003), Photochemical fate of pharmaceuticals in the environment: Cimetidine and ranitidine, Environ. Sci. Technol. 37, 3342–3350.

109

Lee, W. J. and Batchelor, B. (2004), Abiotic reductive dechlorination of chlorinated ethylenes by iron-bearing phyllosilicates, Chemosphere 56, 999–1009. Legrini, O., Oliveros, E., and Braun, A. M. (1993), Photochemical processes for water treatment, Chem. Rev. 93, 671–698. Leland, J. K. and Bard, A. J. (1987), Photochemistry of colloidal semiconducting iron oxide polymorphs, J. Phys. Chem. 91, 5076–5083. Li, F. B., Li, X. Z., Li, X. M., Liu, T. X., and Dong, J. (2007), Heterogeneous photodegradation of bisphenol A with iron oxides and oxalate in aqueous solution, J. Colloid Interface Sci. 311, 481–490. Li, X. Z., Liu, G. M., and Zhao, J. C. (1999), Two competitive primary processes in the photodegradation of cationic triarylmethane dyes under visible irradiation in TiO2 dispersions, New J. Chem. 23, 1193–1196. Li, X. Z., Zhao, W., and Zhao, J. C. (2002), Visible light-sensitized semiconductor photocatalytic degradation of 2,4-dichlorophenol, Sci. China Ser. B Chem. 45, 421–425. Lin, W. and Frei, H. (2005a), Anchored metal-to-metal chargetransfer chromophores in a mesoporous silicate sieve for visiblelight activation of titanium centers, J. Phys. Chem. B 109, 4929–4935. Lin, W. and Frei, H. (2005b), Photochemical CO2 splitting by metalto-metal charge-transfer excitation in mesoporous ZrCu(I)MCM-41 silicate sieve, J. Am. Chem. Soc. 127, 1610–1611. Litter, M. I., Villegas, M., and Blesa M. A. (1994), Photodissolution of iron oxides in malonic acid, Can. J. Chem. 72, 2037–2043. Liu, C. S., Li, F. B., Li, X. M., Zhang, G., and Kuang, Y. Q. (2006), The effect of iron oxides and oxalate on the photodegradation of 2-mercaptobenzothiazole, J. Molec. Catal. A Chem. 252, 40–48. Liu, G. M. and Zhao, J. C. (2000), Photocatalytic degradation of dye sulforhodamine B: A comparative study of photocatalysis with photosensitization, New J. Chem. 24, 411–417. Liu, G. M., Li, X. Z., Zhao, J. C., Hidaka, H., and Serpone, N. (2000a), Photooxidation pathway of sulforhodamine-B. Dependence on the adsorption mode on TiO2 exposed to visible light radiation. Environ, Sci. Technol. 34, 3982–3990. Liu, G. M., Li, X. Z., Zhao, J. C., Horikoshi, S., and Hidaka, H. (2000b), Photooxidation mechanism of dye alizarin red in TiO2 dispersions under visible illumination: An experimental and theoretical examination, J. Molec. Catal. A Chem. 153, 221–229. Lobedank, J., Bellmann, E., and Bendig, J. (1997), Sensitized photocatalytic oxidation of herbicides using natural sunlight, J. Photochem. Photobiol. A Chem. 108, 89–93. Mao, Y. and Thomas, J. (1993), Photoinduced electron transfer and subsequent chemical reactions of adsorbed thianthrene on clay surfaces, J. Org. Chem. 58, 6641–6649. Matsuzawa, S., Nasser-Ali, L., and Garrigues, P. (2001), Photolytic behavior of polycyclic aromatic hydrocarbons in diesel particulate matter deposited on the ground, Environ. Sci. Technol. 35, 3139–3143. Mazellier, P. and Bolte, M. (2000), Heterogeneous light-induced transformation of 2,6-dimethylphenol in aqueous suspensions

110


containing goethite, J. Photochem. Photobiol. A Chem. 132, 129–135. Mazellier, P. and Sulzberger, B. (2001), Diuron degradation in irradiated, heterogeneous iron/oxalate systems: The rate-determining step, Environ. Sci. Technol. 35, 3314–3320. Meehan, C., Banat, I. M., McMullan, G., Nigam, P., Smyth, F., and Marchant, R. (2000), Decolorization of Remazol Black-B using a thermotolerant yeast, Kluyveromyces marxianus IMB3, Environ. Int. 26, 75–79. Mehra, O. P. and Jackson, M. L. (1960), Iron oxides removal from soils and clays by a dithionite-citrate system buffered with sodium carbonate, Clays Clay Miner. 7, 317–327. Mele, G., Del Sole, R., Vasapollo, G., Garcıa-Lo´pez, E., Palmisano, L., and Schiavello, M. (2003), Photocatalytic degradation of 4-nitrophenol in aqueous suspension by using polycrystalline TiO2 impregnated with functionalized Cu(II)-porphyrin or Cu (II)-phthalocyanine, J. Catal. 217, 334–342. Moon, J., Yun, C. Y., Chung, K. -W., Kang, M. -S., and Yi, J. (2003), Photocatalytic activation of TiO2 under visible light using Acid Red 44, Catal. Today 87, 77–86. Murad, E. and Fischer, W. R. (1988), The geobiochemical cycle of iron, in Iron in Soils and Clay Minerals, Stucki, J. W., Goodman, B. A., and Schwertmann, U., eds., Reidel, Dordrecht. Myli, K. B., Larsen, S. C., and Grassian, V. H. (1997), Selective photooxidation reactions in zeolites X, Y and ZSM-5, Catal. Lett. 48, 199–202. Nakamura, R. and Frei, H. (2006), Visible light-driven water oxidation by Ir oxide clusters coupled to single Cr centers in mesoporous silica, J. Am. Chem. Soc. 128, 10668–10669. Nakamura, R., Okamoto, A., Osawa, H., Irie, H., and Hashimoto, K. (2007), Design of all-inorganic molecular-based photocatalysts sensitive to visible light: Ti (IV)-O-Ce(III) bimetallic assemblies on mesoporous silica, J. Am. Chem. Soc. 129, 9596–9597. Nasr, C., Vinodgopal, K., Fisher, L., Hotchandani, S., Chattopadhyay, A. K., and Kamat, P. V. (1996), Environmental photochemistry on semiconductor surfaces. Visible light induced degradation of a textile diazo dye, Naphthol Blue Black, on TiO2 nanoparticles. J. Phys. Chem. 100, 8436–8442. Pehkonen, S. O., Ron, S., Erel, Y., Webb, S., and Hoffmann, M. R. (1993), Photoreduction of iron oxyhydroxides in the presence of important atmospheric organic compounds, Environ. Sci. Technol. 27, 2056–2062. Pehkonen, S. O., Siefert, R. L., and Hoffmann, M. R. (1995), Photoreduction of iron oxyhydmxides and the photooxidation of halogenated acetic acids, Environ. Sci. Technol. 29, 1215–1222. Qu, P., Zhao, J. C., Zang, L., Shen, T., and Hidaka, H. (1998), Enhancement of the photoinduced electron transfer from cationic dyes to colloidal TiO2 particles by addition of an anionic surfactant in acidic media, Colloid Surf. A Physicochem. Eng. Asp. 138, 39–50. Reyes, C. A., Medina, M., Crespo-Hernandez, C., Cedeno, M. Z., Arce, R., Rosario, O., Steffenson, D. M., Ivanov, I. N., Sigman, M. E., and Dabestani, R. (2000), Photochemistry of pyrene on

unactivated and activated silica surfaces, Environ. Sci. Technol. 34, 415–421. Robinson, T., McMullan, G., Marchant, R., and Nigam, P. (2001), Remediation of dyes in textile effluent: A critical review on current treatment technologies with a proposed alternative, Bioresour. Technol. 77, 247–255. Ross, H., Bendig, J., and Hecht, S. (1994), Sensitized photocatalytical oxidation of terbutylazine, Solar Energy Mater. Solar Cell 33, 475–481. Seo, Y. S., Lee, C., Lee, K. H., and Yoon, K. B. (2005), 1:1 and 2:1 charge-transfer complexes between aromatic hydrocarbons and dry titanium dioxide, Angew. Chem. Int. Ed. 44, 910–913. Serpone, N. (2006), Is the band gap of pristine TiO2 narrowed by anion- and cation-doping of titanium dioxide in second-generation photocatalysts? J. Phys. Chem. B 110, 24287–24293. Sherman, D. M. (2005), Electronic structures of iron(III) and manganese(IV) (hydro)oxide minerals: Thermodynamics of photochemical reductive dissolution in aquatic environments, Geochim. Cosmochim. Acta 69, 3249–3255. Sherman, D. M. and Waite, T. D. (1985), Electronic spectra of Fe3 þ oxides and oxide hydroxides in the near IR to near UV, Am. Miner. 70, 1262–1269. Shichi, T. and Takagi, K. (2000), Clay minerals as photochemical reaction fields, J. Photochem. Photobiol. C Photochem. Rev. 1, 113–130. Siffert, C. and Sulzberger, B. (1991), Light-induced dissolution of hematite in the presence of oxalate. A case study, Langmuir 7, 1627–1634. Song, L., Qiu, R., Mo, Y., Zhang, D., Wei, H., and Xiong, Y. (2007), Photodegradation of phenol in a polymer-modified TiO2 semiconductor particulate system under the irradiation of visible light, Catal. Commun. 8, 429–433. Song, W. J., Cheng, M. M., Ma, J. H., Ma, W. H., Chen, C. C., and Zhao, J. C. (2006), Decomposition of hydrogen peroxide driven by photochemical cycling of iron species in clay, Environ. Sci. Technol. 40, 4782–4787. Stathatos, E., Petrova, T., and Lianos, P. (2001), Study of the efficiency of visible-light photocatalytic degradation of basic blue adsorbed on pure and doped mesoporous titania films, Langmuir 17, 5025–5030. Stucki, J. W. (2005), Properties and behaviour of iron in clay minerals, in Handbook of Clay Science, Bergaya, F., Theng, B. K. G., and Lagaly, G., eds., Elsevier, Amsterdam, pp. 429–482. Stumm, W. and Sulzberger, B. (1992), The cycling of iron in natural environments: Considerations based on laboratory studies of heterogeneous redox processes, Geochim. Cosmochim. Acta 56, 3233–3257. Stylidi, M., Kondarides, D. I., and Verykios, X. E. (2004), Visible light-induced photocatalytic degradation of Acid Orange 7 in aqueous TiO2 suspensions, Appl. Catal. B Environ. 47, 189–201. Sun, C., Zhao, D., Chen, C., Ma, W., and Zhao, J. (2009), TiO2mediated photocatalytic debromination of decabromodiphenyl ether: Kinetics and intermediates, Environ. Sci. Technol. 43, 157–162.

REFERENCES

Thomas, J. K. (2005), Physical aspects of radiation-induced processes on SiO2, c-Al2O3, zeolites, and clays. Chem. Rev. 105, 1683–1734. Thompson, T. L., and Yates, J. T. (2006), Surface science studies of the photoactivation of TiO2—new photochemical processes, Chem. Rev. 106, 4428–4453. Volodin, A. M. (2000), Photoinduced phenomena on the surface of wide-band-gap oxide catalysts, Catal. Today 58, 103–114. Wallington, T. J. and Nielsen, O. J. (2005), Atmospheric Photooxidation of Gas Phase Air Pollutants, Springer-Verlag, Berlin/ Heidelberg. Wang, K., Zhang, J., Lou, L., Yang, S., and Chen, Y. (2004), UV or visible light induced photodegradation of AO7 on TiO2 particles: The influence of inorganic anions, J. Photochem. Photobiol. A Chem. 165, 201–207. Wang, Q., Chen, C., Zhao, D., Ma, W., and Zhao, J. (2008), Change of adsorption modes of dyes on fluorinated TiO2 and its effect on photocatalytic degradation of dyes under visible irradiation, Langmuir 24, 7338–7345. Weare, W. W., Pushkar, Y., Yachandra, V. K., and Frei, H. (2008), Visible light-induced electron transfer from di-m-oxo-bridged dinuclear Mn complexes to Cr centers in silica nanopores, J. Am. Chem. Soc. 130, 11355–11363. Wu, T. X., Lin, T., Zhao, J. C., Hidaka, H., and Serpone, N. (1999a), TiO2-assisted photodegradation of dyes. 9. Photooxidation of a squarylium cyanine dye in aqueous dispersions under visible light irradiation, Environ. Sci. Technol. 33, 1379–1387. Wu, T. X., Liu, G. M., Zhao, J. C., Hidaka, H., and Serpone, N. (1999b), Evidence for H2O2 generation during the TiO2-assisted photodegradation of dyes in aqueous dispersions under visible light illumination, J. Phys. Chem. B 103, 4862–4867. Wu, T. X., Liu, G. M., Zhao, J. C., Hidaka, H., and Serpone, N. (2000), Mechanistic study of the TiO2-assisted photodegradation of squarylium cyanine dye in methanolic suspensions exposed to visible light, New J. Chem. 24, 93–98. Xie, T.-H., Sun, X., and Lin, J. (2008), Enhanced photocatalytic degradation of RhB driven by visible light-induced MMCT of Ti (IV)OFe(II) formed in Fe-doped SrTiO3, J. Phys. Chem. C 112, 9753–9759.

111

Xie, Y., Yuan, C., and Li, X. (2005), Photocatalytic degradation of X-3B dye by visible light using lanthanide ion modified titanium dioxide hydrosol system, Colloid Surf. A Physicochem. Eng. Asp. 252, 87–94. Xu, Y. and Langford, C. H. (2001), UV- or visible-light-induced degradation of X3B on TiO2 nanoparticles: The influence of adsorption, Langmuir 17, 897–902. Xu, Y. and Schoonen, M. A. A. (2000), The absolute energy position of conduction and valence bands of selected semiconducting minerals, Am. Miner. 85, 543–556. Xyla, A. G., Sulzberger, B., LutherIII, G. W., Hering, J. G., Van Cappellen, P., and Stumm, W. (1992), Reductive dissolution of manganese(III,IV) (hydro)oxides by oxalate: The effect of pH and light, Langmuir 8, 95–103. Zafiriou, O. C., Joussot-Dubien, J., Zepp, R. G., and Zika, R. G. (1984), Photochemistry of natural waters, Environ. Sci. Technol. 18, 358A–371A. Zhao, J. C., Hidaka, H., Takamura, A., Pelizzetti, E., and Serpone, N. (1993), Photodegradation of surfactants 11. Zeta-potential measurements in the photocatalytic oxidation of surfactants in aqueous TiO2 dispersions, Langmuir 9, 1646–1650. Zhao, J. C., Wu, T. X., Wu, K. Q., Oikawa, K., Hidaka, H., and Serpone, N. (1998), Photoassisted degradation of dye pollutants. 3. Degradation of the cationic dye rhodamine B in aqueous anionic surfactant/TiO2 dispersions under visible light irradiation: Evidence for the need of substrate adsorption on TiO2 particles, Environ. Sci. Technol. 32, 2394–2400. Zhao, W., Chen, C. C., Ma, W. H., Zhao, J. C., Wang, D. X., Hidaka, H., and Serpone, N. (2003), Efficient photoinduced conversion of an azo dye on hexachloroplatinate(IV)-modified TiO2 surfaces under visible light irradiation—a photosensitization pathway, Chem. Eur. J. 9, 3292–3299. Ziolkowski, L., Vinodgopal, K., and Kamat, P. V. (1997), Photostabilization of organic dyes on poly(styrenesulfonate)-capped TiO2 nanoparticles, Langmuir 13, 3124–3128. Zollinger, H. (1987), Color Chemistry: Synthesis, Properties and Applications of Organic Dyes and Pigments, VCH, New York.

PART II ANTHROPOGENIC ORGANIC COMPOUNDS IN AIR, WATER, AND SOIL, AND THEIR GLOBAL CYCLING

5 SORPTION OF ANTHROPOGENIC ORGANIC COMPOUNDS TO AIRBORNE PARTICLES HANS PETER H. ARP AND KAI-UWE GOSS 5.1. Introduction 5.1.1. Environmental Relevance of Ambient Gas/Particle Partitioning 5.1.2. Definition of Ambient Equilibrium Gas/Particle Partitioning 5.1.3. Aerosol Sources, Composition, and Size Fractions 5.2. Measuring Ambient Gas/Particle Partitioning 5.2.1. Sample-and-Extract Methods 5.2.2. Inverse Gas Chromatography 5.3. Traditional Gas/Particle Partitioning Models for Apolar Semivolatile Organic Compounds 5.3.1. Basic Partitioning Theory 5.3.2. The Junge–Pankow Model 5.3.3. Pankow Absorption–Adsorption Model 5.3.4. The Octanol Absorptive Model 5.3.5. The EC þ OC Model 5.3.6. Criticisms and Applicability of Traditional Models and SP-LFERs 5.4. Sorbent–Sorbate Interactions Involved in Partitioning and PP-LFERs 5.4.1. Overview of Sorbent-Sorbate Interactions 5.4.2. Describing Absorptive Partitioning with PP-LFERs 5.4.3. Describing Adsorptive Partitioning with PP-LFERs 5.4.4. Other Partitioning Models that Account for Sorbate-Sorbent Diversity 5.4.5. Temperature Dependence of Partitioning 5.5. Partitioning to Individual Aerosol Components 5.5.1. Pure Water 5.5.2. Snow and Ice 5.5.3. Minerals and Metal Oxides 5.5.4. Salts 5.5.5. Elemental Carbon 5.5.6. Water-Soluble Organic Matter 5.5.7. Water-Insoluble Organic Matter

5.6. Partitioning to Mixed Particle Phases 5.6.1. Terrestrial Aerosols 5.6.2. Other Ambient Particles 5.7. Conclusions

5.1. INTRODUCTION 5.1.1. Environmental Relevance of Ambient Gas/Particle Partitioning The air you are breathing contains several organic chemicals. Although the majority of these chemicals are benign, some are bioaccumulative and in sufficient concentrations are toxic. A fundamental aspect governing how you are exposed to these chemicals is if they are present as a vapor or sorbed to airborne particles. For instance, some toxic chemicals are highly volatile (such as carbon monoxide and hydrogen cyanide), and exist in the air completely as a vapor. Other toxic chemicals, such as the heavier polychlorinated dibenzodioxins (PCDDs), exhibit low volatility, and thus do not enter the air substantially as a vapor; however, they can enter the air by sorbing to airborne particles, like dust and smoke particulates. Once inhaled, larger particles will be trapped by the mouth and throat, where they can subsequently release chemicals into the body through mucus membranes. Many of the smaller particles ( 200 2 > 200 2 > 200 2 > 200 N/A N/A N/A >1

Deliquescence salts, WSOM Salt, WSOM 1 nm–100 þ mm

Water spray

Aqueous

Permeable Ad/Absorbing Phases

Notation: N/A—not available; VOC—volatile organic compounds. a Excludes biological debris. b Mass fractions are based on dry weights and are presented as approximate ranges based on the cited references; note that the mass fractions given in this table are highly dependent on the sample’s particulate matter (PM) fraction. c Data from Hueglin et al. (2005) with maximum salts including salts þ unknown, along with Sillanpaa et al. (2005) and Yttri et al. (2007) for EC/OC; WSOM and WIOM values estimated from both WSOM: OM and WIOM ratios ranging from 30%–70% (Gysel et al. 2004; Krivacsy et al. 2001; Wang et al. 2005; Yang et al. 2005). d Estimates based on Tables 5.2 and 5.4 in Cavalli et al. (2004). e Ranges based on data from various sources (Burtscher 2005; Hildemann et al. 1991; Norbeck et al. 1998; Roth et al. 2005b; Schauer et al. 1999) with OM assumed to be WIOM. f Ranges based on data from various sources (Hildemann et al. 1991; Norbeck et al. 1998; Schauer et al. 2002b) with OM assumed to be WIOM. g Ranges based on data from various sources (Hildemann et al. 1991; Schauer et al. 2001), with OM assumed to be WIOM. h From Roth et al. (2005b).

Salt, OM 1 nm–1 mm

None

Combustion

EC

Water, WSOM 1 nm > 10

NOx, SO2, NH3, etc.

Oceans, combustion

Salt

Solid Adsorbing Phases Minerals/Metal Oxides

Mixing phases OM Size ranges 1 > 10 mm Mass Fractions (% dry weight)b Terrestrial PM 2.5c 7–12 Terrestrial PM 10c 13–21 Marine PM 1d N/A Marine PM > 1d N/A Diesel soote 8–15 Gasoline sootf N/A Wood smokeg N/A Road tunnel Aerosolsg N/A

Secondary precursors

Primary sources (examples)

Parameter

TABLE 5.1. Overview of Unique Sorbing Components Found in Aerosols

120

SORPTION OF ANTHROPOGENIC ORGANIC COMPOUNDS TO AIRBORNE PARTICLES

Figure 5.1. A SEM scan of a particle sample collected in Berlin, Germany, during spring 2006 indicating a large porous conglomerate of organic particles with some salts (yellow circle), a sharp-edged, smooth-surfaced salt crystal (blue circle), a pollen grain (red circle), and smooth-edged, smooth-surfaced mineral particles (green circles). Where do the contaminants sorb preferentially? (See insert for color representation of this figure.)

blowing off the filter and onto the downstream sorbent, which causes the determined cip to be too low (a negative bias when determining Kip). To minimize these artifacts the other setups depicted in Figure 5.2 were proposed. The principle of the filter–filter–sorbent (FFS) (Fig. 5.2b) system is quite similar to the FS system, as it simply contains two filters instead of one. The additional filter, referred to as the backup filter, is added to correct for blow on artifacts. If both filters have reached sorption equilibrium with the incoming air, then cip can be corrected for by subtracting the amount of sorbate on the backup filter from the front filter, and ci air can be corrected for by adding twice the amount of sorbate collected on the backup filter. Although FFS setups are recommended over FS, it is worthwhile estimating

beforehand whether equilibrium partitioning to the front and backup filters can actually be achieved under the actual sampling conditions and if corrections should be made (Arp et al. 2007; Mader and Pankow 2001; Volckens and Leith 2003b). Partially in an attempt to avoid blow on artifacts completely, the denuder-filter-sorbent (DFS) systems were developed (Fig. 5.2c). The purpose of the denuder is to strip the air of gaseous SVOCs before they land on the filter. However, as the particles on filter are exposed to SVOC-free air, blow off occurs, and a downstream sorbent is placed to correct for this. When using the DFS setup, it must be ensured that the right denuder is chosen for the target analytes, so that particles do not collect in the denuder, and that breakthrough of analytes through the denuder does not occur (Mader et al. 2001; Volckens and Leith 2003a,b). The electrostatic precipitator–sorbent system (EPS) (Fig. 5.2d) (Volckens and Leith 2003a,b) can be thought of as a filter-free FS setup. Here particles are charged and collected on a conducting surface while gaseous compounds pass by to accumulate on the downstream sorbent. One unique drawback of this approach is that certain molecules can be transformed when they pass through the precipitator, thus leading to losses of those molecules in both the particle and gas phases. 5.2.1.2. Considerations When Designing Sample-andExtract Procedures. The artifacts associated with sampleand-extract methods have been reviewed extensively (e.g., Arp et al. 2007; Arp and Goss 2008; Arp et al. 2008c; Galarneau and Bidleman 2006; Volckens and Leith 2003a,b). All methods carry unique biases that need to be accounted for when designing sample-and-extract setups, as there is no method that is suitable for all chemicals and sampling situations. Generic biases to all techniques can come from fluctuating ambient conditions, such as RH,

Figure 5.2. Simultaneous particle and air-phase chemical collectors used in sample-and-extract methods: (a) filter–sorbent; (b) filter–filter–sorbent; (c) denuder–filter–sorbent; (d) electrostatic precipitator–sorbent. (See insert for color representation of this figure.)

MEASURING AMBIENT GAS/PARTICLE PARTITIONING

temperature, PM levels, and compound concentrations, which influences partitioning to fiber filters, denuders, and particles alike. In general, it is best to sample during stable weather conditions, and use sampling strategies that minimize the influence of temperature fluctuations (Galarneau and Bidleman 2006). To minimize the influence of fluctuating ambient conditions, so-called controlled field experiments can be carried out, in which particle filters are loaded with aerosols and subsequently exposed to a stream of gas containing a constant concentration of analyte at a controlled RH and temperature (Mader and Pankow 2002). An additional consideration is that filter (or electrostatic precipitator) extracts do not distinguish between those compounds existing in particle components that are exchangeable with the air phase and those that are not. A compound class for which this issue was raised early on is PAHs (Pankow and Bidleman 1991), as these can be trapped within soot particles during combustion processes and be inaccessible to air and other environmental matrixes. Many of the initial studies extracted PAHs from filters using Soxhlet extractions with methylene chloride, which can readily remove PAHs from nonexchangeable (i.e., extremely slowly exchangeable) aerosol components (Jonker and Koelmans 2002; Jonker et al. 2005). As a result, many of the Kip values for PAHs reported in the literature may be erroneously high (Arp et al. 2008a), compared to what is occurring in the atmosphere over the 1–10-day course of an aerosol’s lifetime. Thus, in order to obtain Kip values specifically for the exchangeable components, we recommend gentler extraction procedures (e.g., air stripping, supercritical fluid extraction (Jonker et al. 2005) or IGC techniques (see below). 5.2.2. Inverse Gas Chromatography Inverse gas chromatography (IGC) uses the same equipment and principles as gas chromatography; the main difference is that IGC experiments are conducted to characterize the sorption behavior of the stationary phase, and not to characterize the injected sorbents (i.e., the column packing material itself is what is being studied). The basic design of an IGC suitable for experiments conducted at environmental conditions is shown in Figure 5.3. In IGC experiments the carrier gas (typically nitrogen or synthetic air for environmental studies) passes through a temperature-regulated water saturator to adjust the RH (Fig. 5.3b) before entering the chemical injection port (Fig. 5.3c). The carrier gas then proceeds along capillaries and enters the column packed with the sorbent of interest (Fig. 5.3d), and then finally passes through the detector (Fig. 5.3e). With this setup, both RH and temperature can be easily and separately controlled (Arp et al. 2006a; Conder and Young 1979; Dorris and Gray 1981). To determine the equilibrium sorption coefficient for the material present in the column (Ki sorbent air), the volume of carrier gas required

121

Figure 5.3. Inverse gas chromatography assembly for sorption experiments at environmental conditions, indicating (a) the carrier gas source (generally synthetic air), (b) the gas humidifier for adjustment of the relative humidity (RH), (c) chemical injection port, (d) column in a temperature-controlled water bath, and (e) chemical detector (generally a flame ionization detector is used at ambient RH).

to bring the sorbate through the column, referred to as the retention volume (Vi ret) is measured. The following equation can then be used to calculate Ki sorbent air values: Ki sorbent air ¼

Vi ret Vdead Msorbent

ð5:3Þ

Here, Vdead is the “dead volume” (i.e., empty volume), which is determined by measuring the retention volume of a nonsorptive tracer (typically methane), and Msorbent is the sorbent mass in the case of absorption or sorbent surface area in the case of adsorption. Note that ideally, instead of using Vdead, it is better to measure the retention volume of the compound of interest in the IGC when an empty column is installed, in case of substantial sorption to the capillaries, frits, or column walls. It is also advised to measure Ki sorbent air at varying carrier gas flow rates and injection volume of sorbent, to test whether sorption equilibrium is reached and if sorption is occurring in the linear part of the isotherm respectively (see Section 5.3.1) (Conder and Young 1979; Roth et al. 2005b). Finally, in the case of peak skewing (which is common with IGC techniques due to heterogeneities in the sorbent and column packing), column breakthrough is best quantified using the first moment of the sorbate peak in the detector (Conder and Young 1979). Column packing is the most critical step in designing an IGC experiment. Packing a column too loosely may result in preferential flow paths, which cause a negative Kip bias. On the other hand, packing the column too tightly will cause excess flow resistance or column blockage, and may even alter the packing material (e.g., via particle congealing). Depending on the media, it may be impractical to pack the

122


sorbate itself as a pure phase, and thus an additional support phase is needed. Examples of support phases include silica supports such as porous glass or Chromosorb [used for water surfaces (Dorris and Gray 1981; Roth et al. 2002)], glass beads [used with diesel soot (Roth et al. 2005b) and humic acid (Niederer et al. 2006)], and GFFs [used with terrestrial aerosol samples (Arp et al. 2008c)]. The support phase must be nonreactive and as low-sorbing as possible. When a supporting phase is completely attenuated, Ki sorbent air values can be determined as before [Eq. (5.3)]. If, on the other hand, the supporting phase also sorbs the incoming sorbate, then Ki sorbent air must be determined using the equation. Ki sorbent air ¼

Vi sorbate þ support Vsupport Msorbent

ð5:4Þ

where Vi sorbate þ support is the total retention volume and Vsupport is the retention to the support and capillaries in the system. In Arp et al. (2008c) an IGC method was developed to measure Kip values for terrestrial aerosols using surfacetreated GFFs as a support phase. Such IGC methods have advantages and disadvantages compared to sample-andextract methods. The advantages are: (1) when measuring Vsupport one inherently accounts for filter sorption artifacts, (2) RH and temperature are controlled independently of each other, (3) only partitioning to the air-exchangeable components of the aerosols occurs, and (4) the chemical data-set for which Kip values can be measured is not limited by the compounds present on ambient particles. The notable disadvantages are (1) the IGC can be used to measure only Kip values and not concentrations and (2) currently it remains infeasible to measure large SVOCs and surfactants at ambient conditions and environmentally relevant concentrations (Arp et al. 2008a; Arp and Goss 2009a).

5.3.1. Basic Partitioning Theory The chemical potential of a substance in the ideal gaseous state [mi(g)] and in a sorbed state [mi(sorbed)] can be expressed as in the following set of equations: mi ðgÞ ¼ moi þ RT ln

mi ðsorbedÞ ¼ moi þ RT lnðci sorbent Xi sorbent Þ

Since the mid-1980s or so, the controlling gas/particle partitioning mechanisms relevant for ambient particles have been under debate, and correspondingly so have Kip predictive models. Thus, in the scientific literature, one currently encounters a large variety of partitioning models and postulates. In this section, the traditional models that were developed for apolar SVOCs are presented. Models developed for polar, ionizable and surfactant molecules will be presented in Sections 5.5 and 5.6. Before doing so, we will introduce basic partitioning theory at low sorbate concentrations (i.e., infinite dilution), as this theory forms the basis for all partitioning models.

ð5:5Þ ð5:6Þ

where moi is the chemical potential of species i in its pure liquid state (i.e., the reference chemical potential), R is the ideal gas constant, T is temperature in kelvins, pi is the partial pressure of species i in the gas phase, p*i L is the saturated subcooled liquid vapor pressure of i (Pa), ci sorbent is the activity coefficient of i in the sorbent (unitless), and Xi sorbent is the mole fraction of i relative to the sorbent (moli mol1 sorbent ) [for more information, see Schwarzenbach et al. (2003)]. Partitioning equilibrium between two phases implies that the chemical potential of i in both phases is equal, that is mi(g) ¼ mi(sorbed), and therefore the equilibrium condition is defined as pi ¼ ci sorbent Xi sorbent p*i L

ð5:7Þ

This equation can also be expressed in terms of volumetric air concentration at equilibrium by dividing by RT: ci air ðmol=m3 Þ ¼

pi c Xi sorbent p*i L ¼ i sorbent RT RT

ð5:8Þ

In the case of absorption, Xi sorbent can be converted to a mass concentration by dividing by the molecular weight of the sorbent, MWsorbent (gsorbent=molsorbent): ci sorbent ðmol=gÞ ¼

5.3. TRADITIONAL GAS/PARTICLE PARTITIONING MODELS FOR APOLAR SEMIVOLATILE ORGANIC COMPOUNDS

pi p*iL

Xi sorbent MWsorbent

ð5:9Þ

The absorptive partitioning constant Ki sorbent, air can thus be expressed in terms of either these concentrations or ci sorbent and p*iL : Ki sorbent;air ðm3air =gsorbent Þ ¼

ci sorbent RT ¼ ci air MWsorbent ci sorbent p*i L ð5:10Þ

An analogous approach can be taken for adsorptive partitioning. When we consider that there is just one type of adsorption site available, the adsorbed concentration at equilibrium csurface can be expressed as csurface ¼ Xi surface Ns

ð5:11Þ

TRADITIONAL GAS/PARTICLE PARTITIONING MODELS FOR APOLAR SEMIVOLATILE ORGANIC COMPOUNDS

where Xi surface is the mole fraction of i per mole of suface site (moli/molsurface sites) and Ns is the surface site density (molsurface sites =m2surface ). The equilibrium adsorption constant, Ki surface, air (m3air =m2surface ), is therefore Ki surface; air

ci surface RT Ns ¼ ¼ ci air ci surface p*i L

ð5:12Þ

Note from Equations (5.10) and (5.12) that the only parameter that is both compound- and sorbent-specific is the activity coefficient (ci sorbent or ci surface). Elucidating sorbent and sorbate parameters that govern these activity coefficients has proved challenging. However, one special case exists, and that is for sorbent–sorbate pairs when the sorbate interacts with the sorbent similarly to how the sorbate interacts with itself (e.g., nonane absorbed in decane would lead to similar interactions with nonane absorbed in nonane); in this case, ci sorbent 1. This ideal case is commonly referred to as Raoult’s law. Raoult’s law generally does not apply when the sorbent and sorbing phases have markedly different properties (e.g., with an apolar sorbate and polar sorbent). The sorbent and sorbate identities are not the only variables to influence ci sorbent and ci surface. They are often dependent on the sorbate concentration. This dependence is referred to as nonlinear partitioning. Nonlinear partitioning behavior can take many forms, from ci decreasing with increasing sorbate concentration, to ci increasing with sorbate concentration, to more complex changes. However, at low sorbate concentration ranges, typical for environmental concentrations of AOCs, ci sorbent and ci surface values are typically constant. However, it should be emphasized that under special environmental conditions nonlinear partitioning can occur, such as at high concentrations near contaminant source zones (like chemical spills), or when dealing with very heterogeneous sorbing phases that contain competing sorptive components. 5.3.2. The Junge–Pankow Model The first gas/particle partitioning model to be widely adopted was the Junge–Pankow model (Pankow 1987). In essence, the model assumed that partitioning of apolar SVOCs to aerosols was occurring with equal affinity to the total surface of the aerosol. Thus, to parameterize this model, it was necessary to know the amount of aerosol surface area per cubic meter of air, q (m2aerosol =m3air ) and the total suspended particle concentration TSP (gaerosol =m3air ). The model can be written as follows: Ki p ðm =gÞ ¼ Ki surface; air ðm =m Þ q TSP 3

3

2

ð5:13Þ

Substituting Equation (5.12) into Equation (5.13) above and taking the logarithm of both sides gives

log Ki p ðm3 =gÞ ¼ log p*i L log ci surface þ log ðNS RT q=TSPÞ

123

ð5:14Þ

The original Junge–Pankow model includes extra parameters to empirically estimate ci surface, which are not shown here (Pankow 1987). Indeed, not only ci surface but also many of the parameters in this equation are difficult to measure or model, thus for practical purposes a simplified version of this equation was recommended for use, and became quite popular: log Ki p ðm3 =gÞ ¼ m log p*i L þ b

ð5:15Þ

Here, m and are b are fitted linear regression constants, with m representing any proportionalities between p*i L and ci surface, and b representing any spillover from this proportionality and the sorbate independent terms [i.e., log (NSRT q/TSP)]. Equations like Equation (5.15), where an unknown log K value is linearly related to a known log K value (here log p*i L , which is commonly available for many compounds) are commonly referred to as single-parameter linear free-energy relationships (SP-LFERs). The term single parameter linear refers to the fitting of the unknown m and b values by a singleparameter linear regression, and free energy refers to the log K values, which are directly proportional to the corresponding free energy of phase transfer, DG (i.e., DG ¼ ln K/RT þ constant). An early, simple, version of this SPLFER that was commonly used in fate models was Ki p ¼ ð6 106 Þ=p*i L (Mackay et al. 1986). 5.3.3. Pankow Absorption–Adsorption Model In the early 1990s there were an increasing number of studies reporting correlations between the total organic carbon (TOC) content in aerosols and Kip values for apolar SVOCs. Thus, the hypothesis emerged that OM absorption might be more important than surface adsorption. Pankow (1994b) showed, using derivations from basic partitioning theory, that even if absorption were the dominating mechanism, the resulting SP-LFER would be similar, as can be derived as follows. Assuming that the entire weight fraction OM in the aerosol, fOM (gOM/gaerosol) is available for absorption and is the exclusive sorbing phase, then Ki p ðm3 =gÞ ¼ fOM KOM

ð5:16Þ

Thus, applying the general equation for absorptive partitioning [Eq. (5.10)] for the OM phase, we obtain Ki OM ðm3 =gÞ ¼

RT MWOM ci OM p*i L

ð5:17Þ

124


Substituting this into Equation (5.16) and taking the logarithm gives

5.3.6. Criticisms and Applicability of Traditional Models and SP-LFERs

fOM RT MWsorbent ð5:18Þ

The models presented in Sections 5.3.2–5.3.5 make several simplifying assumptions (e.g., that the homogeneity and availability of OM and EC are similar across all ambient aerosols samples), and their appropriateness for ambient aerosols has been validated only for low-polarity SVOCs (PAHs, PCBs, PBDE, PCDD/Fs). Therefore, these models are open to some scrutiny and criticism, especially when applied to a broad variety of compounds and particle types. In Section 5.1 we discussed how the OM fraction of ambient aerosols consists of WIOM and mixed WSOM–aqueous domains. By the very (inherent) nature of these domains, they must exhibit widely differing sorption properties (Arp et al. 2008b; Erdakos and Pankow 2004; Erdakos et al. 2006; Griffin et al. 2003; Pun et al. 2002), one favoring waterinsoluble molecules; the other water-soluble molecules. Accordingly, organic compounds unique to these individual fractions have been extracted (e.g., Ding et al. 2008; Yang et al. 2005). Thus, simply because of OM heterogeneity, there is no mechanistic basis for correlations between log Kip and fOM (i.e., the sum of fWIOM and fWSOM) to be consistent for all aerosols and compounds classes, despite its practicality. More details on the differentiating sorption behavior of WIOM and WSOM are presented in Sections 5.5 and 5.6. The assumption that adsorption controls gas/particle partitioning processes is also not robust enough to cover all compounds and particle types; for instance, it is unlikely that amorphous and liquid-like aerosol OM phases in the atmosphere form and sorb by adsorptive processes only. The “EC þ OC” model has so far been validated only for PAHs (Lohmann and Lammel 2004). Polycyclic aromatic hydrocarbons are a special type of SVOC, as they are formed and associate with soot/EC particles during combustion processes (Hart et al. 1993); thus, Kip models developed for PAHs are not necessarily applicable to other compound classes (Arp et al. 2008a). Validations of the EC þ OC model for PAHs in the literature have also been criticized because the Ki EC values used to derive them are highly elusive and variable (Galarneau et al. 2006). An important critical aspect about SP-LFERs in general is that the regression coefficients m and b are both sorbate- and sorbent-specific, because of their dependence on ci (Goss and Schwarzenbach 1998, 2001). To illustrate the sorbent-phase dependence, we note that Ki sorbent; air p*i L SP-LFER m values [Eq. (5.15)] for alkanes absorbed in octanol, ethylene glycol, and water are 0.92, 0.55, and 0.21, respectively. To illustrate sorbate dependence for nonpolar compounds, PCBs and alkanes sorbed in octanol give Ki oa p*i L SP-LFER m values are 0.67 and 0.92, respectively (Goss and Schwarzenbach 1998). A similar analogy can be made for Ki sorbent, air Ki oa SP-LFERs, as different organic solvents can have widely different m values for diverse sorbents and

log Ki p ðm3 =gÞ ¼ log p*iL logci OM þ log

Ignoring the role that varying sorbate–sorbent interactions can play on the ci OM term as before, one can derive the same SP-LFER as before [Eq. (5.15)], although m and b are now functions of different parameters. 5.3.4. The Octanol Absorptive Model Finizio et al. (1997) expanded the previous absorption model by saying that not only is Kip proportional to fOM, but that absorption into aerosol OM was proportional to absorption in pure octanol, that is Ki OM ¼ ao Ki oa

ð5:19Þ

where Ki oa is the octanol–air partition coefficient and ao is some proportionality factor. Substituting Equation (5.19) into Equation (5.16) and taking the logarithm gives log Ki p ðm3 =gÞ ¼ log Ki oa þ log fOM þ logao

ð5:20Þ

This can also be related to a SP-LFER log Ki p ðm3 =gÞ ¼ m log Ki oa þ b

ð5:21Þ

where the y intercept b accounts for log fOM þ log ao, and the slope m accounts for sorbate-dependent deviations between the proportionality of Ki OM and Ki oa. 5.3.5. The EC þ OC Model Dachs and Eisenreich (2000) reported that adsorption onto the soot or elemental carbon (EC) fraction of aerosols is the main sorption process for PAHs, and not absorption into the OM. They suggested for PAHs that a dual-phase sorption model should be used that accounts for both OM absorption and EC adsorption. This prompted many researchers (e.g., Mader and Pankow 2002; Lohmann and Lammel 2004) to use equations of the form Ki p ¼ aOM fOM log Ki oa þ aEC fEC Ki EC

ð5:22Þ

where aOM and aEC are proportionality constants, fEC is the fraction of elemental carbon in the aerosol, and Ki EC is the equilibrium partition coefficient between diesel soot and air.

TRADITIONAL GAS/PARTICLE PARTITIONING MODELS FOR APOLAR SEMIVOLATILE ORGANIC COMPOUNDS

sorbates [see Fig. 5.12 of Schwarzenbach et al. (2003)]. The robustness of a Ki sorbent, air Ki oa SP-LFER depends on how similar the sorbent is to octanol (e.g., hexanol is very similar, while octane, water, and glyoxal are not). As there is little evidence to indicate that aerosol OM is structurally similar to octanol (e.g., aerosol OM contains diverse functional groups, heteroatoms, and aromatic structures), the assumption that the Ki oa SP-LFERs performs equally well for all sorbates should be viewed with scrutiny. It should be mentioned here that a widely held fallacy in the gas/particle partitioning literature is that SP-LFER m values should be generally near 1 if equilibrium partitioning is reached, regardless of the sorbent, sorbate, and SP-LFER, based on some early empirical evidence for SVOCs and ambient particles (e.g., Harner and Bidleman 1998; Pankow 1994b). The examples and theory given above and in the literature show that there is no thermodynamic basis to support this for all sorbents and sorbates (Goss and Schwarzenbach 1998; Pankow and Bidleman 1992). Thus, discussions in the literature based on this assumption should be viewed with skepticism. Despite SP-LFERs exhibiting limited sorbate–sorbent applicability domains, they remain highly practical and popular. They are generally suitable for apolar SVOCs, and perform sufficiently for applications in which high accuracy in predicted Kip values is not required. When dealing with a broad variety of sorbates, however, it is of interest to address which of these two SP-LFER paradigms, p*i L -based and Ki oabased, exhibits the most robust performance for Kip values. Therefore, in Figure 5.4a,b, these two types of SP-LFERs were calibrated to a diverse set of Kip data for one specific aerosol sample, covering several apolar and polar sorbates of low water solubility.

125

In this example, the p*i L SP-LFER (Fig. 5.4a) gives the weaker regression statistic (r2 ¼ 0.57) because the highly polar molecules (seen well above the regression line) exhibit different cip than do compounds with low polarity (below the regression line). This is a typical result for p*i L SP-LFER regressions when both polar and apolar molecules are used, which is why this model performs best when only apolar molecules are considered (e.g., Arp et al. 2006a; Goss and Schwarzenbach 1998). The Ki oa SP-LFERs, on the other hand, reduce this deviation from the regression line by likening the sorbing phase cip to that of octanol, which, although not a perfect surrogate, reduces the deviation of the polar and nonpolar compounds from the regression line (at least for the compound classes tended). Thus, the Ki oa SPLFER is more robust than the p*i L SP-LFER. However, as foreshadowed here in Figure 5.4c, even better performing types LFERs exist, which will be introduced later on. To conclude, of the traditional gas/particle partitioning models, calibrated Ki oa SP-LFERs are the best performing; however, when applying this model to broad sets of aerosol types and compounds, the following cautions need to accounted for: (1) the b value in Eqn. (5.22) being dependent on fOM is an arbitrary assumption with low plausibility, as it does not account for WIOM and WSOM domains; (2) Ki oa SPLFERs alone should not be used to predict Ki p values of highly water-soluble, ionizable, or surfactant molecules, as these sorbents prefer aqueous components (Section 5.6.3); (3) the Ki oa values used should be experimentally determined and be validated with as many consistency checks as possible, as Ki oa data, similar to octanol–water partitioning data, are challenging to measure and literature data can vary substantially (Renner 2002); (4) if no experimentally determined Ki oa is available, a more appropriate surrogate than

Figure 5.4. Comparison of three fitted LFERs using the same compound dataset and experimental Ki p (15 C) values for the same aerosol sample [“Duebendorf fall,” from Arp et al. (2008a)], showing (a) the SP-LFER Pankow adsorption/absorption approach [Eq. (5.15)] (n ¼ 29, RMSE ¼ 0.41), (b) the SP-LFER Ki oa approach [Eq. (5.20)], (n ¼ 51, RMSE ¼ 0.28), and (c) the PP-LFER approach [Eq. (5.25)], (n ¼ 65, RMSE ¼ 0.174). Note that the number of compounds changes for the different datasets as experimentally determined p*i L , Ki oa, or PP-LFER descriptors were not available for all 65 compounds (Arp et al. 2008a).

126


octanol can be used to base Kip predictions (Section 5.6.6); and (5) more accurate models exist (as will be presented below) that should be favored when high accuracy is required (as shown in Fig. 5.4c) or for screening large datasets (Section 5.6.6).

5.4. SORBENT–SORBATE INTERACTIONS INVOLVED IN PARTITIONING AND PP-LFERS 5.4.1. Overview of Sorbent-Sorbate Interactions In order to make gas/particle partitioning models that are more robust than those described in Section 5.3 for diverse organic compounds, it is necessary to look at the specific sorbent–sorbate interactions involved in partitioning, and how these influence ci. Sorbent–sorbate interactions can be classified as specific and nonspecific, which are synonymous for polar and apolar interactions. Nonspecific interactions are called such because they do not occur between specific locations of the molecular surfaces of the sorbent and sorbate. An important type of nonspecific interactions are dispersive interactions (i.e., London interactions, which are a type of van der Waals interaction), which occur between random and temporary electron-rich and electron-poor areas that arise from random electron delocalization. These dispersive interactions occur for all molecular surfaces, and can vary vastly in strength (e.g., from Teflon to activated carbon). An example of dispersive interactions between two butane molecules is depicted in Figure 5.5a. An additional nonspecific interaction is referred to as cavity formation, depicted in Figure 5.5b, which occurs in the case of absorptive partitioning only. When

a sorbate enters a sorbent, the sorbent–sorbent interactions must be broken to accommodate a cavity for the sorbate to occupy. The larger the sorbate, the larger the cavity will have to be, and thus the more energy is required. Similarly, the stronger the sorbent–sorbent interaction, the more energy is required to form such a cavity. Specific interactions occur between specific sites of the sorbent and sorbate’s molecular surface. An example is a hydrogen-bond, which occurs between a permanent electron donor region of one molecule and permanent electron acceptor region of another molecule. Permanent electron donor/acceptor regions can arise from covalently bonded atoms of substantially different electronegativity. As a classic example, the hydroxyl functional group contains a highly electronegative oxygen atom (electronegativity ¼ 3.44) bonded to a weakly electronegative hydrogen atom (electronegativity ¼ 2.20), this causes the oxygen to have an electron donor region and the hydrogen, an electron acceptor region. Alternatively, an ether group contains a substantial electron donor region on the oxygen, but no substantial electron acceptor region is present on the neighboring carbons. Two butanol molecules forming two simultaneous specific interactions with each other are illustrated in Figure 5.6a. Electron-donating regions can also arise from well-structured molecular orbitals. For instance the p-electron rings of aromatic molecules are negatively charged, and can thus interact with the positive electron acceptor regions of a neighboring molecule (Fig. 5.6b). 5.4.2. Describing Absorptive Partitioning with PP-LFERs The overall free energy of phase transfer during partitioning is the sum of the individual free energies for the nonspecific and specific interactions that are occurring, that is X X DGi sorbent=air ¼ DGnonspecific DGspecific i sorbent=air þ i sorbent=air ð5:23Þ

Figure 5.5. Illustration of nonspecific interactions showing (a) random dispersive interactions between temporary electron-rich and electron-poor areas and (b) cavity formation energy, in which sorbent–sorbent bonds (shown in blue) are broken in order to accommodate the sorbate. Random dispersive interactions are always thermodynamically favorable and occur between all molecules. Cavity formation is thermodynamically favorable only when the energy cost of breaking sorbent–sorbent interactions is overcompensated by sorbent–sorbate interactions (black lines). (See insert for color representation of this figure.)

where DGi sorbent=air is the free energy of partitioning, P DGnonspecific i sorbent=air is the contribution of all the nonspecific P sorbate–sorbent interactions and DGspecific i sorbent=air is the contribution of all the specific sorbate–sorbent interactions. As free energies are directly proportional to log K values (again, DG ¼ log K/2.303RT þ constant), Equation (5.23) can be equivalently written as X X logKi sorbent=air ¼ logKinonspecific logKispecific sorbent=air þ sorbent=air þ 2:303RT constant ð5:24Þ Equation (5.24) can also be related to the sorbate and sorbent dependence of the ci sorbent. By substituting

127

SORBENT–SORBATE INTERACTIONS INVOLVED IN PARTITIONING AND PP-LFERS

Figure 5.6. Illustration of specific interactions showing (a) two butanol molecules undergoing two simultaneous specific interactions, where the O- atom acts as an electron donor (negative dipole) and the H atom as an electron acceptor (positive dipole), and (b) the p-electron ring of benzene (negative dipole) interacting with the H atom of a water molecule (positive dipole). (See insert for color representation of this figure.)

Equation (5.10) into Equation (5.24), it becomes evident that ci sorbent itself can be subdivided into individual activity coefficients, each responsible for the various sorbate–sorbent specific interactions: ci sorbent ¼ Pcnonspecific i sorbent Pci sorbent . In order to develop a robust and practical approach to solve Equation (5.24) for all sorbent–sorbate combinations, several unique approaches have been developed [for a partial review, see Lei et al. (2008)]. Of these approaches, Cramer (1980), followed by others (Abraham et al. 1994a; Taft et al. 1985), found that the partition properties of a given compound in any absorptive partition system can be fully characterized by five molecular descriptors. This was later corroborated by quantum chemical modeling (Zissimos et al. 2002). The five-molecular-descriptor equation that we recommend for general use in environmental systems is (Goss 2005) log Ki sorbent=air ¼ s Si þ a Ai þ b Bi þ l Li þ v Vi þ c ð5:25Þ where Si, Ai, Bi, Li and Vi are the sorbate–specific Abraham descriptors for the polarizability/dipolarizability, electron acceptor (i.e. H-bond donor) capability, electron donor (H-bond acceptor) capability, logarithm of the hexadecane/air partition coefficient, and McGowan volume, respectively. Corresponding to the sorbate-specific descriptors are the sorbent-specific descriptors s, a, b, l, and v, along with the fitting constant, c. Unlike the SP-LFER above, log K values of an unknown partition system are linearly related to several known descriptor values and not just one; therefore, equations such as (5.25) are referred to as Polyparameter linear free-energy relationships (PP-LFERs). In this PP-LFER, the lLi and vVi account collectively for the nonspecific contributions of dispersive interactions and cavity formation P ( logKinonspecific i, and bBi account collectsorbent ), and sSi, aAP ively for specific interactions ( logKispecific sorbent ). The fitting constant c depends on contributions unrelated to sorbent– sorbate interactions, such as the units of Ki sorbent, the purity of the sorbent, and the heterogenity of sorption components.

The sorbate-specific descriptors in Equation (5.25) can be measured independently in the lab, as outlined in Abraham et al. (2004), except for Vi, which can be calculated on the basis of molecular structure after Abraham and McGowan (1987). Large tables of sorbate descriptors can be found in the literature (Abraham 1993a; Abraham et al. 1994a; Abraham et al. 1994b; Abraham and Al-Hussaini 2001, 2005; Arp et al. 2008a; Goss et al. 2008; Mintz et al. 2007; Tulp et al. 2008). Example sorbate descriptors for selected SVOCs are included in Table 5.5 at the end of this chapter. The sorbent-specific descriptors are determined via multiparameter linear regression with known experimental log Ki sorbent values and known sorbate descriptors (e.g., Goss and Schwarzenbach 2001; Goss et al. 2005). It should be noted here for readers familiar with these PP-LFERs, that Equation (5.25) has some advantages over more traditional PP-LFER equations found in the literature; it does not require the use of the descriptor Ei (i.e., excess molar refraction), whose estimation for solid compounds introduces additional error (Atapattu and Poole 2008), and it works for partitioning between two condensed phases and for gas and condensed phases alike [for more details, see Goss (2005)].

5.4.3. Describing Adsorptive Partitioning with PP-LFERs Analogously to the case of absorptive partitioning, the logarithm of adsorptive partitioning constants, log Ki surface/air, can be subdivided into specific and nonspecific interactions: logKi surface=air ¼

X

logKinonspecific þ surface

X

þ 2:303RT constant

logKispecific surface ð5:26Þ

Examples of such interactions are illustrated in Figure 5.7. The general PP-LFER to describe adsorption is (Goss 2004; Roth et al. 2002)

128


Figure 5.7. Examples of adsorptive interactions at surfaces, showing (a) nonspecific dispersive interactions, (b) a specific interaction where the surface is the electron donor and butanol is the electron acceptor, and (c) the surface hydroxyl group engaging in two simultaneous specific interactions with a water molecule. (See insert for color representation of this figure.)

logKi surface=air ðm3 =m2 Þ ¼ 0:135Li

qffiffiffiffiffiffiffiffiffiffi cvdW surf þ5:11Bi EAsurf

þ3:60Ai EDsurf 8:47 where the adsorbent-specific descriptors

pffiffiffiffiffiffiffi vdW csurf

ð5:27Þ

, EAsurf, and

EDsurf represent the adsorbents’ dispersive force capability, electron-accepting capability, and electron- donating capapffiffiffiffiffiffiffi bility, respectively. The “0.135Li cvdW ” term accounts for surf P nonspecific log Ki surface (the vdW stands for van der Waals forces), and “5.11BiEAsurf þ 3.60AiEDsurf” collectively account P for log Kispecific surface. The constant 8.47 in Equation (5.27) is derived from de Boer for the standard state of adsorption (assuming the molecules appearing on the surface to be due to random collisions with the surface without any intermolecular forces) (de Boer 1968; Roth et al. 2002). The coefficients in front of EAsurf and EDsurf were derived such that EAsurf and EDsurf could equal 1 for the water surface at 15 C, thus facilitating comparisons of the polarity of various surfaces with that of the water surface (Roth et al. 2002). Note, in contrast for the absorption PP-LFER [Eq. (5.25)], that the adsorption PP-LFER does not require a cavity formation term (as this is irrelavent), nor, interestingly, does it require an Si descriptor (Goss et al. 2005). Although this model has been found to apply to a large number of compounds (Arp et al. 2006a) and adsorbents (Goss 2004), there is one identified outlying compound class: highly fluorinated compounds. For these compounds, a slightly different adsorption PP-LFER is needed, due to the failure of Li to describe the dispersive interaction for these compounds (Arp et al. 2006b). 5.4.4. Other Partitioning Models that Account for Sorbate-Sorbent Diversity Equation (4.3) is not the only PP-LFER that can be used to quantify specific and nonspecific sorbent–sorbate interactions; indeed, a variety of alternative and additional

descriptors can be used (Goss 2005; Liu and Oberg 2009). A particular shortcoming of all PP-LFERs is the statistical correlation between parameters used to describe dispersive interactions and cavity formation, such as the Li and Vi parameters (Goss and Bronner 2006), as both nonspecific interactions are highly dependent on the sorbate’s volume. For certain applications, it may be worthwhile or even necessary to use alternative descriptors. Saying this, however, we favor the PP-LFER in Equation (5.25), especially as a starting point, due to its proven robustness for a large variety of compounds and sorbing phases. Further, the thermodynamic basis and shortcoming of most of the descriptors are heavily documented and well understood and are established using several consistency checks (e.g., Abraham 1993b; Arey et al. 2005; Goss and Bronner 2006). Thus, when outliers from this PP-LFER occur, a sound mechanistic explanation for the outlying behavior is thereby facilitated, and novel insight into the occuring partitioning mechanism may result (e.g., Goss and Bronner 2006). Besides PP-LFER approaches, several other modeling approaches attempt to explicitly account for diverse specific and nonspecific interactions [see the review by Lei et al. (2008)], although only a few have been applied to gas/particle partitioning. Variations of UNIFAC (the universal functional activity coefficient), which correlates ci and p*i L with molecular fragment contributions, are popular for p*i L and water activity coefficients (Asher et al. 2002; Chandramouli et al. 2003a; Chang and Pankow 2006; Clegg et al. 2008a, 2008b). Quantum chemical approaches are becoming increasingly popular for predicting partitioning constants, such as the commercial software COSMOtherm (COSMOlogic GmbH, Leverkusen, Germany), which uses density functional quantum chemical continuum solvation calculations with statistical thermodynamics to determine activity coefficients (Eckert and Klamt 2002, 2005). Another popular computer program that has been found to give good predictions for diverse compound classes is SPARC (scalable processor architecture performs automated reasoning in chemistry), and is a free, online Web application (http://

129

PARTITIONING TO INDIVIDUAL AEROSOL COMPONENTS

ibmlc2.chem.uga.edu/sparc/) that explicitly calculates sorbate–sorbent interactions by using various empirical molecular descriptors that are derived from molecular structure (Hilal et al. 2003, 2004). Particular advantages of COSMOtherm and SPARC are that they can allow for any sorbent and sorbate as input, as well as account for the influence of temperature.

partitioning behavior of organic compounds to these individual phases are discussed. In Sections 5.6 and 5.7, partitioning to mixed-particle samples are discussed. To aid in this discussion it is useful to provide PP-LFERs for these individual phases, as they can be used as predictive tools. Further, these PP-LFERs inherently provide insight on the role of specific and nonspecific interactions in these diverse phases.

5.4.5. Temperature Dependence of Partitioning As with any other equilibrium constant, the temperature dependence of Ki sorbent/air can be expressed using the van’t Hoff equation, which has the following solution when the enthalpy of partitioning between a sorbent and the air (DHi sorbent/air) is assumed to be temperature-independent: log Ki sorbent=air ¼

DHi sorbent=air þ RTa 1 þ constant ð5:28Þ T 2:303R

Here, Ta is the average temperature of the temperature range of interest [note that the term DHi sorbent/air þ RTa is the internal energy of partitioning] (Atkinson and Curthoys 1978)]; DHi sorbent/air is dependent on the sorbate’s enthalpy of vaporization (DvapHi) and the enthalpy of sorbate–sorbent interactions in the condensed phase (HiEsorbent ): DHi sorbent=air ðkJ=molÞ ¼ HiEsorbent Dvap Hi

ð5:29Þ

Ideally, DHi sorbent/air should be experimentally derived using Ki sorbent/air values covering an ambient temperature range. When this is not possible, the following empirical estimation techniques may be appropriate. For absorptive partitioning, it is often (but not always) the case that HiEsorbent is negligible, and thus DHi sorbent/air DvapHi. For predicting the enthalpy of adsorptive partitioning (DHi surface/air), HiEsorbent can no longer be considered negligible. The following empirical relationship is recommended, which has been validated for a diverse dataset of compounds and sorbents (Arp et al. 2006a; Goss 2004). DHi surface=air ðkJ=molÞ ¼ 9:83:log Ki surface=air ð15 CÞ90:5 ðn ¼ 182; r2 ¼ 0:89Þ

ð5:30Þ

5.5.1. Pure Water Condensed water, present in clouds, fog, and rain, is the most abundant condensed phase in the atmosphere. Atmospheric water droplets are not necessarily “pure” but often contain several organic compounds and ions. Here sorption to pure water droplets is discussed, and the influence of dissolved ions and organic cosolvents to water droplet sorption is discussed in Section 5.7.1. Pure water droplets can both adsorb and absorb organic chemicals. In general, large apolar compounds and surfactants prefer to adsorb to the surface of the water droplet, while polar and ionic compounds prefer to absorb into the bulk water matrix. Whether adsorption or absorption dominates depends on the compound and the size of the water droplet. Regarding adsorption, Roth et al. (2002) derived the following adsorptive PP-LFER for the water surface [Eq. (5.27)], in which EAsurf and EDsurf were chosen to be pffiffiffiffiffiffiffi 1 at 15 C as a reference state, and cvdW was calibrated at surf 4.69: log Ki water surf=air ðm3 =m2 ; 15 CÞ ¼ 0:635 Li þ 3:60 Ai þ 5:11 Bi 8:47 ðn ¼ 60; r2 ¼ 0:932Þ ð5:31Þ The PP-LFER for absorptive partitioning [Eq. (5.25)] in the water bulk matrix, Ki wa (Goss 2005) is log Ki wa ðm3air =m3water ; 25 CÞ ¼ 2:07 Si þ 3:67 Ai þ 4:87 Bi þ 0:48 Li 2:55 Vi 0:59 ðn ¼ 390; r2 ¼ 0:997Þ

Other estimation techniques for specific sorbents will be presented later on in this chapter.

5.5. PARTITIONING TO INDIVIDUAL AEROSOL COMPONENTS In Table 5.1, six types of sorption components found in the aerosols were introduced: the aqueous component, salts, minerals, EC, WIOM, and WSOM. In this section, the

ð5:32Þ Note that normally in the literature the term Ki aw is used, which is the inverse of Ki wa; Ki aw is also commonly referred to as the “dimensionless” Henry’s law constant, named after William Henry, who first studied the influence of pressure and temperature on water/air partitioning. The cavity formation term (v) in Equation (5.32) (of 2.55) is more negative than those of other organic liquids [e.g., see Goss (2005)], indicating that it costs a lot more energy for

130


a solvent to make a cavity in water than a cavity in organic liquids, which is attributable to the strong intermolecular H-bonding of water molecules. Large apolar molecules, or apolar moieties of surfactants, cannot overcome this cavity energy and, thus, preferentially adsorb on the water surface. However, the ability of water to make strong H-bonds with polar organic molecules (as indicated by the relatively larger a and b values) overcompensates this high cavity formation energy, and makes water a good solvent for small polar molecules (e.g., methanol is totally miscible). Several organic compounds can additionally ionize in the water phase. The degree of this ionization depends on the compound’s acid dissociation constant in water (pKi a), and the aqueous-phase pH. For the water/air partitioning of ionizable compounds, the distribution ratio of both the neutral and ionic form of the compound (Di wa), can be calculated as follows Di wa

Ki wa ¼ ai a

Di wa ¼

Ki wa 1ai a

for organic acids for organic bases

ð5:33Þ ð5:34Þ

where ai a is the fraction of the compound in the protonated form: ai a ¼ ð1 þ 10

pHpKi

a

1

Þ

DHi wa ðkJ=molÞ ¼2:84 Si þ 32:0 Ai þ 41:8 Bi ð5:36Þ

ðn ¼ 368; r ¼ 0:964Þ 2

Although the vast majority of compounds exhibit decreasing Ki wa with increasing temperature (i.e., positive DHi wa), many compounds can exhibit the opposite trend or negligible temperature dependence (Arp and Schmidt 2004), which depends mainly on the value of HiEwater [which can be either positive or negative (Schwarzenbach et al. 2003)] relative to DvapHi (which is always positive, i.e. endothermic). To quantify whether absorption or adsorption dominates for a given compound and water particles of a given size, the enrichment factor EFi can be calculated as follows EFi ¼

Ki water surf=air ðA=VÞ þ Ki wa Ki wa

wi bulk water ¼

ci water Di wa LWC ¼ ci air þ ci water 1 þ Di wa LWC

ð5:37Þ

where A is the surface area of the water droplet and V is the volume of the droplet. From this equation, an EF < 2

ð5:38Þ

where LWC is the dimensionless liquid water content in a unit of air (m3water =m3air ). 5.5.2. Snow and Ice Snow and ice are important atmospheric sinks of organic contaminants in arctic regions and high altitudes. Roth et al. (2004) conducted low-temperature IGC sorption experiments to snow, and calibrated the following “adsorbentlike” PP-LFER: log Ki snow surf=air ðm3 =m2 ; 6:8 CÞ ¼ 0:635 Li þ 3:38 Bi þ 3:53 Ai 6:85 ðn ¼ 57; r2 ¼ 0:90Þ

ð5:35Þ

The temperature dependence of adsorptive partitioning can be estimated using Equation (5.30). To account for the temperature dependence of absorptive partitioning in water, Mintz et al. (2007) derived

þ 6:35 Li 9:91 Ei þ 13:31

indicates that absorption is dominating and > 2, that adsorption is dominating. The equations outlined here are all that are necessary to estimate how much of any given AOC is distributed between air and condensed pure water droplets in a given parcel of air. For instance, if absorption dominates (EF < 2), one can calculate the fraction present in the water droplets (wi bulk water) analogously to Equation (5.2)

ð5:39Þ This above PP-LFER is referred to as “adsorbent-like” because the fitted constant term was significantly different from the universal adsorption constant of 8.47. This could have been due to incorrect determination of the snow surface area or an additional partitioning mechanism, such as to a quasi-liquid layer. Nevertheless, despite this ambiguity, Equation (5.39) is considered a useful predicting tool, although it should be kept in mind that the sorption properties of a snow surface may change depending on the snow’s degree of crystallization (Roth et al. 2004). 5.5.3. Minerals and Metal Oxides Adsorption to minerals and metal oxides is highly dependent on RH. The reason is that water layers accumulate on mineral and metal oxide surfaces as RH increases. As the RH approaches 100%, the water layer thickens exponentially, until adsorption to the mineral surface is ultimately the same as that onto the water surface (Goss 2004). This is depicted in Figure 5.8 using butanol as an example. This water layer only slightly alters the specific interactions, quantified with EAsurf and EDsurf; which is partly attributable to the fact that the EAsurf and EDsurf values of most minerals at low RH are already similar to that of the bulk water surface (Table 5.2). The thickening water layer with

PARTITIONING TO INDIVIDUAL AEROSOL COMPONENTS

131

Figure 5.8. The influence of relative humidity (RH) on adsorption to minerals and metal oxides, showing the increase of the water-layer thickness with increasing RH. (See insert for color representation of this figure.)

increasing RH lowers Ki surf/air values by substantially reducing the nonspecific interactions, which is seen in a pffiffiffiffiffiffiffi decrease of cvdW with increasing RH (Table 5.2). More surf information on partitioning to diverse mineral surfaces can be found elsewhere (Goss and Schwarzenbach 2002; Goss et al. 2003; Goss 2004).

salts found in the atmosphere, NaCl and (NH4)2SO4, sorbent descriptors at RH below the deliquescence RH can be found in Table 5.2. Interestingly, beneath the deliquescence RH pffiffiffiffiffiffiffi neither the Ki surface/air, EAsurf EDsurf, or cvdW term is surf influenced by increasing RH (Table 5.2). Partitioning to salt-rich aqueous aerosol droplets, such as a fully deliquesced salt, is discussed in more detail in Section 5.7.1.

5.5.4. Salts After water, salts are the second most abundant particle phase in the atmosphere by mass. Unlike minerals, many salts exhibit a deliquescence RH (see Section 5.1.3.2). The deliquescence RH depends on the salt, impurities present, and temperature (Seinfeld and Pandis 2006). For two abundant

5.5.5. Elemental Carbon The EC fraction of combustion aerosols has been found to vary widely, and it can be expected that numerous EC structures exist in the atmosphere. Freshly emitted diesel particles (i.e., “diesel soot”) contain EC as the main

TABLE 5.2. Examples of Some PP-LFER Descriptors for Mineral and salt Surfaces qffiffiffiffiffiffiffiffiffiffi 2 0:5 EAsurf () EDsurf () Surface RH % cvdW surf ðmJ=m Þ

Quartz a-Al2O32

NaCla

(NH4)2SO4a

45 70 90 40 70 90 20 40 60 20 40 60

Minerals 6.79 0.11 6.04 0.09 5.31 0.08 5.38 0.08 5.02 0.08 4.84 0.08 Salts 6.41 0.09 6.30 0.09 6.19 0.08 6.63 0.08 6.55 0.08 6.47 0.07

Deliquescence RH (25 C) for NaCl is 75% and (NH4)2SO4 is 80%. [Tang and Munkelwitz (1993)]. a

Reference

1.06 0.04 0.96 0.03 0.85 0.03 1.02 0.03 0.92 0.03 0.89 0.03

0.89 0.10 0.88 0.09 0.88 0.07 1.13 0.08 1.08 0.07 1.05 0.07

Goss and Schwarzenbach (2002)

0.85 0.03 0.80 0.03 0.78 0.05 0.67 0.40 0.67 0.03 0.58 0.03

0.93 0.09 0.94 0.08 0.94 0.08 1.02 0.10 1.02 0.09 1.04 0.08

Goss et al. (2003)

Goss and Schwarzenbach (2002)

Goss et al. (2003)

132


component, whereas EC is a minor component of most other combustion particles compared to OM (see Table 5.1). Pure, unattenuated EC is expected to sorb similarly to graphite, pffiffiffiffiffiffiffi which has the highest known cvdW value of 10.7–11.5, and surf thus would be a very attractive adsorbent for AOCs, especially if there is a large exposed surface area. However, likely because of this strong sorption, EC surfaces become rapidly attenuated with OM during combustion processes and even further with OM and salts on immediate release into ambient air (Jacobson 2001; Johnson et al. 2005; Shiraiwa et al. 2007). As a rough indication of the degree of this attenuation, the surface area of urban aerosols [e.g., 0.2–2.2 m2/g (Roth et al., 2005a), 0.82–2.4 m2/g (Sheffield and Pankow 1994)] and even road tunnel aerosols [ 7.4 m2/g (Roth et al. 2005b)] are one or two orders of magnitude smaller than the values for diesel soot [e.g., 91 m2/g (Roth et al. 2005b)], indicating that exposed EC surfaces in ambient atmospheres could be rare. A sorption study on commercially available diesel soot with a relatively high EC content (National Institute of Standards and Technology, (NIST), standard reference material 2975, diesel particulate matter (industrial forklift)] calibrated the following PP-LFER pffiffiffiffiffiffiffi descriptors: cvdW ¼ 8.08 0.07, EAsurf ¼ 0.48 0.02, and surf EDsurf ¼ 0.75 0.06 (15 C) (Roth et al. 2005b). The fact that pffiffiffiffiffiffiffi cvdW for diesel soot is than for graphite further indicates surf that EC surfaces in diesel soot are attenuated. As these attenuating phases are strongly sorbed, they are unlikely to be displaced from EC surface by airborne organics during an atmospheric particle’s lifetime (1–10 days). Supporting this claim, experimental evidence is presented below that sorption to EC surfaces within terrestrial aerosols is likely negligible. Thus, partitioning to EC surfaces is likely occurring to a substantial degree during combustion and during the immediate release of combustion particles in the atmosphere, but only negligibly for nonfreshly emitted particles.

5.5.6. Water Soluble Organic Matter Water-soluble organic matter (WSOM) contains several polar functional groups, especially carboxylic acid, carbonyl, alcohol, and ester groups (Duarte et al. 2007, 2008). Kiss et al. (2002) reported a C:H:N:O ratio of 24:34:1:14, indicating high oxidation, and the presence of polymers and aromatic groups. The most common surrogates for WSOM in lab and theoretical studies are dicarboxylic acids and polysaccharides, as these compounds are typically found in WSOM (e.g., Koehler et al. 2006; Marcolli and Krieger 2006). Like salts, WSOM phases are hygroscopic, with deliquescence RH values ranging from 65% (e.g., malonic acid) to 99% (oxalic acid) (Koehler et al. 2006; Marcolli and Krieger 2006). Because of the very polar nature of WSOM, it likely consists mainly of secondary OM, formed by the photooxidation of volatile organic precursors (including

AOCs). To our knowledge, no systematic sorption studies of WSOM or surrogates are available in the literature. However, it can be expected, from the general low deliquescence point of many atmospheric particles (often < 50% RH) and from several laboratory and theoretical studies, that WSOM in the atmosphere is largely mixed with salts and water and therefore that partitioning of contaminants to WSOM as a nonmixed, pure phase is likely negligible compared to partitioning to mixed salt–WSOM–water phases (see Sections 5.6 and 5.7.1). 5.5.7. Water-Insoluble Organic Matter As presented in Section 5.1.2, water-insoluble organic matter (WIOM) comprises both primary and secondary OM. The primary sources can be biologically generated (e.g., pollen grains, viruses) or anthropogenically generated (e.g., combustion oil residues). Similarly, SOA components can derive from either biological or anthropogenic VOC precursors. Thus, WIOM comprises many unique phases, which exhibit differing sorption characteristics, ranging from adsorbing hard cellulose fragments to absorbing liquid oils. For partitioning to a given WIOM phase, the general absorptive or adsorptive partitioning equations apply [Eqs. (5.10) and (5.12), respectively]. For instance, for absorptive partitioning into the SOA-WIOM domain, the equation would be Ki SOA-WIOM ðm3air =gSOA-WIOM Þ ¼

RT MWSOA-WIOM ci SOA-WIOM p*iL ð5:40Þ

This equation for the SOA-WIOM subdomain was not chosen arbitrarily. In fact, it is fundamental for understanding and modeling SOA growth (Pankow 1994a; Seinfeld and Pankow 2003). Additionally, as will be discussed in more detail in Section 5.6, a WIOM subdomain, most likely SOAWIOM, was found to be the dominating sorption domain for low-water-soluble organic compounds to ambient terrestrial aerosols.

5.6. PARTITIONING TO MIXED PARTICLEPHASES 5.6.1. Terrestrial Aerosols Terrestrial aerosols refer to airborne particles found in outdoor, terrestrial atmospheres. Thus, this broad term can refer to aerosols found in rural, urban, forest, desert, and other terrestrial environments. Because there are a large variety of different primary and secondary sources of minerals, OM, and salts in different terrestrial atmospheres, the types of particles found can vary widely. From a human and environmental health perspective, understanding the gas/

PARTITIONING TO MIXED PARTICLE-PHASES

particle behavior of terrestrial aerosols is important, as it is terrestrial aerosols that humans breathe and it is in terrestrial environments that most contaminant AOCs are initially released into the atmosphere. A major criticism of previous ambient gas/particle partitioning models presented in Section 5.3 was that they did not explicitly account for the presence of WIOM and mixed WSOM–aqueous domains. The importance of diverse partitioning into these two domains in ambient atmospheres was originally recognized by researchers interested in SOA growth and hygroscopicity (Griffin et al. 2003; Kleeman et al. 1999; Pun et al. 2002). Since then, several conceptual and UNIFAC-based models have been developed to account for sorption into these two phases more robustly (e.g., Clegg et al. 2008a,b; Erdakos and Pankow 2004; Erdakos et al. 2006; Pankow 2003; Zuend et al. 2008). However, these models do not explicitly account for overall gas/particle partitioning to ambient terrestrial aerosols, in which other particle domains than just WSOM and WIOM are present (EC, minerals, etc.). To address this, the authors of this chapter recently conducted a series of IGC experiments to isolate the dominating sorption mechanism of terrestrial aerosols. This study involved the experimental determination of over 1300 Kip values covering diverse ranges of RH (50%–90%), temperature (15 C–55 C), aerosol samples (covering the four seasons, desert dust, rural, coastal, urban, and suburban particles), and compound classes (polar, apolar, ionizable) (Arp et al. 2008a,b; Arp and Goss 2009a,b), the main results of which are summarized below. 5.6.1.1. Identification of Water-Soluble and Insoluble Sorption Domains. Essentially all types of pure phases and mixed phases can be found in terrestrial aerosols (see Table 5.1 and Fig. 5.1). Gotz et al. (2007) reported a gas/ particle partitioning model that accounted for the various mass fractions of the pure sorption components presented in Section 5.5, which assumed that no internal mixing of the individual components occurs, and that mixed-aqueous condensates were absent. One of the conclusions from this study was that minerals (specifically nonattenuated quartz-like surfaces) should dominate sorption of highly polar compounds at low RH, and that because of this Kip values of polar compounds should decrease with increasing RH (as in Fig. 5.8). However, our IGC experiments did not show this. Instead, the measured Kip values for some polar compounds and ionizable compounds actually increased with increasing RH (see Fig. 5.9), whereas for the remaining polar compounds and all the apolar compounds, no substantial change with RH was observed. These results indicate that adsorption to minerals is of little significance for polar compounds, contrary to what is expected when no mixing is occurring. The increase in Kip with increasing RH for some polar and ionizable compounds, however, indicates that the uptake of water with increasing RH (in the mixed-aqueous

133

Figure 5.9. Comparison of Ki p (15 C, PM10) data at 50% and 90% RH for an urban aerosol sample sampled in Berlin, Germany (Arp et al. 2008b).

condensates) increases the overall sorption capacity for these compounds, while it has a negligible influence on the remaining compounds. To further characterize the importance of the terrestrial aerosol’s hygroscopic components (salts and WSOM), we investigated what changes in Kip occur when the water soluble phases are removed from the aerosol sample (by extraction with water). Representative results are presented in Figure 5.10.

Figure 5.10. Comparison of log Ki p values (15 C, PM10) before and after extraction of water soluble material from the same aerosol sample shown in Figure 5.9 (Arp et al. 2008b).

134


As is evident from Figure 5.10, after extraction of the water soluble components, the Kip values of most compounds remained constant, with the exception of some “small polar” compounds and ionizable compounds (“small polar” compounds here refers specifically to the ones tested: ethanol, npropanol, isopropanol, n-propanoic acid, and 1,4-dioxane). Thus, from these two experiments it appears evident that two sorbing domains are present; a water-soluble, hygroscopic phase that is attractive to highly water-soluble compounds (such as small polar and ionizable compounds), and a water-insoluble, nonhygroscopic phase that is attractive to low-water-soluble compounds. 5.6.1.2. Partitioning into the Water-Insoluble Domain. To understand partitioning into the water insoluble domain, it is best to focus on Kip data at dry conditions (50% RH, and water extracted aerosols) for low-water-soluble compounds. When we compared such Kip data for nine diverse aerosol samples, it was found that the aerosol sample that gave the highest Kip values was a rural sample rich in small organic particles (collected in Aspvreten, Sweden) and the sample that gave the lowest Kip values was a mineral-rich sample (collected in Duebendorf, Switzerland, during a Sahara-sand event). Log Kip values for these two aerosol samples are plotted against those of a reference sample (for which the largest Kip dataset is available) in Figure 5.11 below, along with scanning electron microscope images of these two “extreme” samples. As is evident from Figure 5.11, the differences in the log Kip values from the reference sample and the two “extreme” samples appears to be generally systematic. The mineral-rich sample exhibits log Kip values 0.28 0.20 less than the

reference sample, and the Aspverten sample exhibits log Kip values 0.58 0.20 more than the reference sample. All other tested samples also exhibited a similar systematically deviating behavior. This general, direct proportionality in log Kip for the different aerosol samples and polar and apolar organic compounds at 50% RH indicates that the sorbing phase is similar for these different aerosols, it is only the relative amount (or mass fraction) of this sorbing phase that differs. To validate further if the dominant sorbing phase was similar in all terrestrial aerosol samples, the absorbent PPLFER equation [Eq. (5.25)] was calibrated with sample specific Kip data. As is evident from Table 5.3, the sorbent descriptors are quite similar across all terrestrial aerosol samples (although with some exceptions, such as “Duebenordf Winter’s” s and l), indicating that they all exhibit similar sorbent–sorbate interactions. Note that the predictions made with the fitted PP-LFERs are in very close agreement with experimental data (r2 > 0.95, RMSE < 0.204), indicating that the sorbent descriptors are of good quality. Having established that sorption to the water insoluble domain is similar for all tested terrestrial aerosol samples, the question arises as to what its identity is and whether it is WIOM, minerals, EC, or some mixture thereof. A close examination of the data reveals that absorption is occurring, and therefor WIOM alone is the only plausible dominating sorption phase. Evidence against contributions from adsorbing phases can be summarized as follows: (1) even after removing the water-soluble components, which could have potentially cleaned off surface-attenuating phases and made surfaces available for adsorption, sorption did not increase but stayed the same or decreased (Fig. 5.10); (2) as stated above, Kip values of certain polar compounds and

Figure 5.11. Experimental log Ki p (15 C, 50% RH, PM10) values of a terrestrial aerosol sample rich in organic particles and another that is rich in mineral particles, compared to a reference sample [“Duebendorf fall,” from Arp et al. (2008a)], along with corresponding SEM images.

135


TABLE 5.3. Particle-Specific Aerosol Descriptors from Fitting Ki p Data (15 C, 50%RH, PM10) to the PP-LFER [Eq. (5.25)] Sample

s

l

v

b

a

c

r2

RMSE

n

Zurich Berlin Spring Berlin Wintera Duebendorf Spring Duebendorf Summer Duebendorf Falla Duebendorf Winter Aspvreten Roost Solvent Octanolb

1.09 0.15 1.01 0.09 1.38 0.18 1.14 0.11 1.09 0.11 1.19 0.12 1.63 0.15 0.95 0.09 1.45 0.13 0.66

0.75 0.06 0.78 0.03 0.63 0.09 0.61 0.04 0.66 0.04 0.66 0.05 0.51 0.06 0.64 0.03 0.60 0.05 0.91

0.35 0.15 0.51 0.09 0.98 0.28 0.69 0.12 0.71 0.12 0.73 0.16 1.51 0.19 0.49 0.09 0.86 0.15 -0.14

0.64 0.22 0.30 0.15 0.42 0.20 0.45 0.16 0.48 0.15 0.03 0.15 0.28 0.21 0.55 0.14 0.37 0.18 1.42

2.99 0.14 3.17 0.12 3.21 0.15 2.75 0.12 2.91 0.12 3.37 0.11 3.20 0.16 2.52 0.11 3.12 0.14 3.49

6.84 0.20 7.42 0.15 7.24 0.18 6.48 0.17 7.28 0.15 7.08 0.14 7.33 0.20 5.95 0.12 6.99 0.17 0.25

0.974 0.970 0.967 0.957 0.966 0.972 0.963 0.975 0.965

0.172 0.139 0.165 0.164 0.161 0.174 0.186 0.132 0.204

38 53 50 57 55 65 51 57 59

Notation: ¼ standard deviation; r2 ¼ correlation coefficient; RMSE ¼ root mean square error; n ¼ number of data; definition of sorbent descriptors can be found in the main text. a Recommended for generic use as aerosol descriptors for terrestrial aerosols. b For comparative purposes, the sorbent descriptors for octanol are also provided (Goss 2005). Source: From Arp et al. (2008a).

ionizable compounds increase with RH (Fig. 5.9), and did not decrease or remain the same as they would if adsorption to salts and mineral surfaces dominated; (3) the sample richest in minerals (Duebendorf-Sahara) revealed much smaller Kip values than did any other sample collected (Fig. 5.11); (4) in general, if absorption dominates over adsorption then the ratios of K values for n-alkanes: cycloalkanes and of fluorotelomer alcohols: n-alcohols are 50%), are expected to partition favorably to the mixed-aqueous components of aerosols, as indicated by the increase in Kip with RH in Figure 5.9. To relate Kip to both the WIOM and RH-dependent, aqueous phases, assuming that these two phases do not mix with each other and thus sorb additively, an equation of the following form is needed: Ki p ¼ sorption to WIOM ð5:41Þ þ sorption to RHdependent aqueous phase For compounds that only absorb into the RH dependent aqueous phase, Equation (5.41) can be parameterized as Ki p ¼ fWIOM Ki p WIOM þ

VwRH ðDiaw S Mdry Þ

ð5:42Þ

where fWIOM is the fraction of sorbing WIOM (this does not include the low-sorbing WIOM, such as wood particles and pollen grains), Kip WIOM is the WIOM-air partitioning coefficient, VwRH is the RH-dependent volume of the mixedaqueous phase, Di aw is the inverse of Di wa, and S is an empirical factor to account for any salting/cosolvency effects occurring in the mixed-aqueous phase. 5.6.1.4. Modeling Kip for Surfactants. An increasing number of substances of interest are surfactant in nature (e.g., perfluorinated surfactants, alcohol ethoxylates, alkyl

136


phenols). These compounds have an affinity for the water surface, thus it is important to account for these compounds adsorbing to the RH-dependent aqueous phase. Additionally, many surfactants can form dimers, aggregates, and structures such as micelles once they exceed certain concentrations, which should be taken into consideration. Thus, a way to parameterize Equation (5.41) for surfactants would be

Ki p ¼ fWIOM Ki WIOM þ

SAwRH Dm VwRH i ws;a þ Dm Mdry i aw Mdry ð5:43Þ

where Dm i aw is the bulk water–air phase distribution coefficient for all species (i.e., monospecies, aggregates, micelles), SAwRH (m2water surface ) is the RH-dependent water surface area, 3 2 and Dm i ws;a (mair =mwater surface ) is the water surface–air phase distribution coefficient for all species. So far, attempts to quantitatively account for the gas/particle partitioning of surfactants has proved challenging, and more research in this direction is needed (Arp and Goss 2009a). 5.6.1.5. Estimating Terrestrial Aerosol Kip Values Using PP-LFERs. If PP-LFER sorbate descriptors for the compound of interest are available, the combined fWIOMKi p WIOM term in Equations (5.42) and (5.43) can be estimated using either the Duebendorf Fall or Berlin Winter PP-LFER sorbent descriptors in Table 5.3 (or the average of both). These two sets of sorbent descriptors were calibrated for PM10 aerosols exhibiting “average” sorption behavior. To estimate the fWIOMKipWIOM term for temperatures substantially different from 15 C, DHip can be estimated from (a)

DHi p ¼ ð0:97 0:15ÞDvap Hi DHi p ¼ ð0:91 0:15ÞDHi octanol=air

0

0

-1

-1

-2

-2

-2

-3

-3

-3

-1

-4

-4 0 1 -4 -3 -2 -1 Dual phase fitted PP-LFER log K ip

ðn ¼ 101Þ ð5:45Þ

(c) 1

0

ð5:44Þ

5.6.1.6. Estimating Kip on the Basis of Molecular Structure. One practical shortcoming of the above sorption model for terrestrial aerosols is that PP-LFER descriptors, as well as

(b) Polar and apolar Small polar Ionizable 1:1 line

ðn ¼ 73Þ

These equations should give reasonable results over an ambient temperature range (i.e., from 10 C to 50 C), which is sufficient for most environmental modeling purposes. To model the VwRH =ðDSi aw Mdry Þ term in Equation (5.42), one can obtain experimental pKa and Ki aw values for many compounds from the literature, otherwise the PPLFER for Ki aw [Eq. (5.32)] can be used. If experimentally determined Mdry, VwRH, and aerosol pH are not available, predictions can be done with assumed maximum and minimum VwRH/Mdry ratios, typically from 0.1 to 1 (Khlystov et al. 2005), and realistic pH ranges, typically from 1 to 4 (Yao et al. 2006; Zhang et al. 2007). Conveniently we found that an S value of 1 (i.e., no salt/cosolvent correction) consistently results in good predictions for water-miscible polar and ionic compounds, which implies that the mixedaqueous phase sorbs similarly to pure water (Arp et al. 2008b). The authors recommended checking the influence of the various assumptions with a subsequent sensitivity analysis. A comparison of experimental Kip values at different RH for a specific aerosol sample with PP-LFER estimated fWIOMKipWIOM and literature pKa and Ki aw values is presented in Figure 5.12.

1

1 Experimental log Kip (m3/g)

experimental DvapHi values or the enthalpy of octanol/air partitioning, DHi octanol/air (Arp et al. 2008a):

-4 -4 -3 -2 -1 0 1 Dual phase fitted PP-LFER log K ip

-4

-3 -2 -1 0 1 Dual phase fitted PP-LFER log K ip

Figure 5.12. Comparison of experimental log Ki p values (m3/g, 15 C) for the Berlin Spring aerosol sample at various relative humidity levels [(a) 50% RH, (b) 70% RH, (c) 90 RH] with predictions made using the dual-phase equation [Eq. (5.42)]. The PP-LFER approach was used to account for sorption to the water-insoluble components (fWIOMKi pWIOM) and weight measurements, a measured pH value of 3.0, and an assumed S value of 1 to account for sorption to the water-soluble components [VwRH/(Di awSMdry)] [from Arp et al. (2008b)].


H3C

O

CH3 O

O

O

O

CH3

CH3

Figure 5.13. Surrogate for WIOM phase in ambient aerosols, proposed by Kalberer et al. (2004) [SMILES string ¼ c(c(cc1C (OC2(OC3C(¼O)C)C)OC2O3)C)c(c1)C], molecular weight ¼ 278.29 g/mol).

pKa and Ki aw values, are not available for every conceivable sorbate, thus making predictions unfeasible for many emerging (or not yet existing) compounds. To address this, we recently validated the following alternative predictive method, which requires only the molecular structure of the compound of interest as input (Arp and Goss 2009b): (1) assume that fWIOM is 0.1 (which is a typical value, see Table 5.1), (2) model KipWIOM, pKa, and Ki aw using either the quantum chemical program COSMOtherm or the SPARC online calculator (see below), and (3) estimated VwRH, Mdry, and pH ranges as above. To predict Ki p WIOM, SPARC and COSMOtherm first need to predict p*i L and ciWIOM [Eq. (5.10)]. For this, a molecular surrogate is needed to represent the WIOM phase. Out of all the suitable molecular surrogates that we tested for this purpose, the molecule depicted in Figure 5.13 resulted in the best estimations (Arp and Goss 2009b). This molecule is a proposed SOA structural unit (Kalberer et al. 2004), and thus is a good representative for WIOM’s SOA–sorbate interactions. Figure 5.14 compares experimental Kip values at different RH levels for a specific aerosol sample with COSMOtherm estimated fWIOMKipWIOM and Ki aw values. It is evident from this figure that the calibrated PP-LFER approach (Fig. 5.12)

137

works better than the noncalibrated COSMOtherm approach. However, as the COSMOtherm approach used only molecular structure as input and was not fitted, the COSMOtherm model is more robust, especially considering that the accuracy of being within an order of magnitude is sufficient for many practical applications. COSMOtherm and SPARC can also predict Kip values at different temperature. In general, we found that COSMOtherm gives better correlations to experimentally determined DHip than does SPARC (Arp and Goss 2009b). Note that it may be attractive for octanol to be used as a surrogate instead of the WIOM structure presented in Figure 5.13. However, we do not recommend doing so, as it was earlier established that WIOM exhibits different sorbent–sorbate interactions than octanol (Section 5.6.1). Further, this would mean forfeiting one of the primary advantages of using estimation software such as COSMOtherm or SPARC: that the user is not limited in terms of which molecular surrogates can be used to represent the sorbing phase. Essentially, any molecular structure can be tested; thus it is worthwhile to try and find the best-performing molecular surrogate with the broadest possible chemical application domain. This is not only because of the potential of improved predictions, but also because the process of testing different surrogates can lead to additional useful information, such as the types of sorbent–sorbate interactions involved and structural information about the sorbing phase. As an example, the correlation presented here indicates that SOA composed of aromatic moieties (Fig. 5.13) is the dominating sorption component of terrestrial aerosols for sorbents with low water solubility, which supports the hypothesis that terrestrial SOA consists largely of polymers from the methyl glyoxal pathway with incorporated aromatic moieties (Kalberer et al. 2004).

Figure 5.14. Comparison of experimental log Ki p values (m3/g, 15 C) for the Berlin Spring aerosol sample at various relative humidities [(a) 50%RH, (b) 70%RH, (c) 90%RH] with predictions made using the dual-phase equation [Eq. (5.42)] based on COSMOtherm values to determine Ki pWIOM) and weight measurements and a measured pH value 3.0 to determine sorption to the water-soluble components [VwRH =ðDSi aw Mdry ] [from Arp et al. (2008b)].

138


Nevertheless, there are some current shortcomings of using the tested versions of the COSMOtherm and SPARC approaches that should be taken into account. There are compound classes whose Kip values were consistent outliers (i.e., were not predicted within an order of magnitude compared to experimental data), specifically large halogenated SVOCs using COSMOtherm and highly fluorinated compounds using certain versions of SPARC (Arp et al. 2006b, Arp and Goss 2009b). COSMOtherm has known difficulties predicting the p*i L for large multichlorinated and large multibrominated species. Thus for such compounds we recommend combining COSMOtherm-derived activity coefficients with literature-derived p*i L . One notable shortcoming with SPARC compared to COSMOtherm is that SPARC cannot account for all the differences in specific and nonspecific interactions due to 3D conformation, thus, important differences in the partitioning of stereoisomers of AOCs (such as those of hexachlorohexanes and hexabromocyclododecanes) can be overlooked (Goss et al. 2008). In general, when testing emerging compound classes, we strongly advise performing checks regardless of whether the compound class of interest falls in the model’s application domain. Additionally, as predictive programs such as COSMOtherm and SPARC are updated, it is necessary to reestablish the chemical applicability domain as new versions are released. In Figure 5.15, COSMOtherm estimations (with literature corrected p*i L values for PBDEs and PCDD/Fs) and SPARC estimations are compared with the experimental Kip over nine orders of magnitude from various sources in the literature, which were measured for a wide variety of terrestrial aerosols and environmental conditions. Considering this variety, both the COSMOtherm and SPARC estimations

5.6.2. Other Ambient Particles 5.6.2.1. Mixed-Aqueous Droplets. As mentioned previously, the mixed-aqueous phase of individual aerosols can increase several times in volume with increasing RH, due to deliquescence of salts and WSOM. During cloud-seeding events, the air is thick with completely deliquesced mixedaqueous droplets. The influences that dissolved salts and WSOM can have on partitioning to mixed-aqueous droplets are, indeed, multifarious and complex. Here we will first present, with broad general strokes, the influence that salts and WSOM can have individually, before commenting on their combined influence. When certain salts are present in high concentrations, salting-out and salting-in effects can occur. “Salting out” occurs when a salt lowers the sorptive capacity of an aqueous phase for organic compounds, and “salting in” applies when a salt raises the sorptive capacity. Whether salting out or salting in occurs depends on the salts and the organic compound. Salting out is commonly observed for small ion pairs (such as NaCl and KNO3) and apolar compounds. This is explainable using the sorbent–sorbate interactions presented in Section 5.4, in that the presence of negative and positive ions increases the sorbate–sorbate interactions and thus the required cavity formation energy. Salting in can occur, for instance, when organic salts or large inorganic ions are dissolved in solution (such as carboxylates and surfactants), which can alternatively lower the sorbent–sorbent interactions at the air–water interface and also lower the cavity

8

PAHs PCDDS-P*iL corrected PCDDS-P*iL corrected Organochlorines PCBs Alkanes PBDEs-p*iL corrected 1:1 line

7 6 5 4

Average Literature Kip (m3/g, various T)

Average literature Kip (m3/g, various T)

8

show excellent potential as screening tools, especially as they were not calibrated to any experimental data.

3 2 1 0 -1

PAHs PCDDs PCDFs Organochlorines PCBs Alkanes PBDEs 1:1 line

7 6 5 4 3 2 1 0 -1 -2

-2 -2

-1

0

1

2

3

4

5

6

7

8

COSMO therm predicted log Kip (m3/g, 15ºC, fWIOM = 0.1)

-2

-1

0

1

2

3

4

5

6

7

SPARC predicted log Kip (m3/g, 15ºC, fWIOM = 0.1)

Figure 5.15. Comparison of average literature log Ki p values from various literature sources with predictions using COSMOtherm and SPARC v4.2 [figure from Arp and Goss (2009b)].

8


formation energy within the bulk matrix. A popular empirical approach for quantifying the salting-out effect is expressed as ci water; salt ¼ ci water 10Ki ½salttot S

ð5:46Þ

where ci water, salt is the activity coefficient of species i in the salty water, ci water that in pure water, KiS ; the Setschenow or “salting constant” (units M-1); and [salt]tot, the molar concentration of the salt. Some example values of the salt and sorbate-dependent KiS are compiled in Schwarzenbach et al. (2003). One generalization that can be made from this relationship is that because KiS values generally range from 0.1 to 0.3, the salting-out effect is generally small until salt concentrations become quite high. For example, a theoretical salt present at 3 M and with a KiS of 0.2 M-1 will change ci water by a factor of only 4. Nonionic WSOCs and other organics can cause cosolvency effects in the aqueous phase. A somewhat more detailed introduction of this topic can be found in Schwarzenbach et al. (2003); here we will mention only some general trends. Completely water-miscible organic compounds (such as methanol) increase the water solubility exponentially with concentration; this is also true for many polar partially miscible organic compounds (such as pentanol and phenols), but not -necessarily for nonpolar partially miscible organic compounds. This cosolvency effect is attributable to changes in sorbent–sorbent interactions. For instance, the presence of a compound like methanol in the water matrix weakens the overall sorbent–sorbent interactions per volume, and thereby lowers cavity formation energy. Designing thermodynamically explicit approaches to deal with the sorption properties of mixed-aqueous phases, considering all salts and WSOM over broad, nonlinear concentration ranges, would be a complex task, indeed. To present an example of the complexity, the most common atmospheric salts in acidic urban aersols, (NH4)2SO4 can cause both salting-out and salting-in effects with polar compounds, and certain mixtures of salts and organics in solution even result in phase changes (Marcolli and Krieger 2006). A common approach to dealing with such mixtures is to use variations of UNIFAC-based methods (e.g., Erdakos et al. 2006; Tong et al. 2008; Zuend et al. 2008). Although UNIFAC-based methods are generally used to describe hygroscopicity and water activity, they are also being developed to predict partitioning of organic compounds (e.g., Clegg et al. 2008a,b). Currently, a UNIFAC-based approach is the most feasible accurate way to approach the gas/particle partitioning of organic compounds to a broad array of mixed-aqueous aerosols, particularly during cloud nuclei formation. However, this level of sophistication is likely not necessary for predicting Kip at trace sorbate concentrations for the aqueous components of terrestrial aerosols. As shown above

139

and in Arp et al. (2008b), predictions based on no substantial salting or cosolvency effects were reasonable. For the partitioning of surfactants, however, this assumption may not hold (Arp and Goss 2009a). 5.6.2.2. Combustion and Road Tunnel Aerosols. The composition of combustion particles depends on the combustion media, combustion system, and ambient conditions on release into the atmosphere. Most types of combustion particles contain either primary OM or EC as the main component [see Table 5.1 and Hildemann et al. (1991)]. Systematic PP-LFERs for such aerosols are available only for diesel particles (Section 5.5.5), although calibrated SP-LFERs are available for various combustion particle types [e.g., for combustion particles from wood and cooking oils (Schauer et al. 2001, 2002a)]. More research is needed to fully characterize the sorption behavior of primary WIOM in diverse combustion aerosols. Road tunnel and traffic aerosols contain a mixture of diesel and gasoline combustion particles, which, on release into the environment, condense, grow, and mix with each other as well as with the components in the surrounding atmosphere. This initial mixing is quite rapid, and profound effects on aerosol properties are observed within minutes after emission (Jacobson 2001; Johnson et al. 2005; Shiraiwa et al. 2007). Accordingly, the specific surface area of road tunnel aerosols is significantly reduced compared to diesel soot (see Section 5.5.5), although not as much as it is for terrestrial aerosols, indicating that road tunnel aerosols are in an intermediate stage of mixing. Roth et al. (2005b) measured Kip values for a commercial road tunnel aerosol standard [European Union’s Institute for Reference Materials and Measurements (IRMM), Geel, Belgium, sold as CRM 605]. Although the dominating sorption mechanism could not be definitively isolated, the available Kip data collectively pointed to an absorptive mechanism. The PP-LFER sorbent descriptors derived from these data are not significantly different from the range of descriptors that have been derived for terrestrial WIOM (Table 5.4). This gives some indication that the dominant sorption mechanism to road tunnels is WIOM absorption. 5.6.2.3. Marine Aerosols. Marine aerosols are formed by wave action or wind, which blows off small particles and aqueous droplets from the ocean’s surface. In addition to water and oceanic salts, these particles and droplets contain WIOM (such as lipids and apolar compounds), WSOM (dissolved compounds), and surfactants that reside on the surface microlayer (Oppo et al. 1999). The weight fractions of the WIOM, WSOM, and salt fractions are highly dependent on particle size fraction [see Table 5.1 and O’Dowd et al. (2004)]. Little is currently known about the gas/particle partitioning behavior of marine aerosols. From what is known of the

140


TABLE 5.4. Absorbent PP-LFER Sorbent Descriptors (15 C) for Road Tunnel Aerosols and a Rural Aerosol Sample Rich in WIOM Sample Ki Ki

a road tunnel b Aspvreten

s

l

v

b

a

c

r2

n

0.87 0.08 0.95 0.09

0.91 0.04 0.64 0.03

0.23 0.13 0.49 0.09

0.17 0.08 0.55 0.14

2.5 0.10 2.52 0.11

6.2 0.07 5.95 0.12

0.99 0.975

66 57

Notation: ¼ standard deviation; r2 ¼ correlation coefficient; RMSE ¼ root mean square error; n ¼ number of data. a Data from Roth et al. (2005b). b From coastal, organic-rich aerosols sampled in Aspvreten, Sweden (Arp et al. 2008a).

contents of these aerosols, it is expected that submicrometer particles are more attractive to organic sorbents of low water solubility, due to their abundance of WIOM, and that coarse particle size fractions are better sorbents for highly watersoluble compounds and surfactants. An important consideration is that when marine aerosols enter the atmosphere, they can change size and sorbent composition, by either agglomeration or particle breakup (Oppo et al. 1999; Tabazadeh 2005). Thus, the degree of internal mixing in a sample of marine aerosols, as with the release of combustion aerosols, is likely in a constant state of flux near the site of aerosol generation. 5.6.2.4. Indoor Aerosols. Indoor aerosols consist of terrestrial aerosols (i.e., outdoor aerosols that have come indoors) and particles that are generated indoors (Jones et al. 2000). Sources of indoor-generated particles include cooking, heating, smoking, erosion of materials, shedding of skin flakes, and particle emissions from household and office equipment. Previous studies on the gas/particle partitioning of indoor aerosols (e.g., Kavouras and Stephanou 2002; Naumova et al. 2003; Weschler et al. 2008) and cigarette smoke (Pankow et al. 1994) have developed SP-LFERs for several compound classes. Because of the diversity of indoor environments and indoor aerosols, and the potential exposure to toxic chemicals in indoor atmospheres, more research is needed to better understand the diversity of gas/particle partitioning of indoor aerosols before any generalized statements can be made. Future research in this direction is important, particularly considering that in most societies more time is spent indoors than outdoors. Future research should focus on developing innovative, low-volume sampling strategies that prevent an unrepresentative amount of outside air being sampled, and which target particles and chemicals are of concern in indoor environments (e.g., engineered nanoparticles, particles emitted from modern appliences, pharmaceuticals, antibiotics, flame retardants).

5.7. CONCLUSIONS Although the field of gas/particle partitioning has progressed, more work is needed, particularly for indoor environments (e.g., office buildings, hospitals, underground subway

passages, smoking lobbies), for tropical and arctic environments, emerging contaminants (e.g., surfactants, pharmaceuticals), and transformation products. Progress in this area will involve the continuous evolution, modifications and improvements of existing Kip measurement and modeling techniques. An inherent sampling bias of all the Kip measurements mentioned in this chapter that should be addressed is that sampling itself alters particle morphology and composition, and thereby potentially alters the measured gas/particle distribution behavior. Thus, ideally, future measurement methods should aim toward directly measuring particle and vapor concentrations in the atmosphere in situ, without the need for sample collection. In this chapter, several different methods of estimating Kip values are presented. The best method to use ultimately depends on the required accuracy of model predictions, the availability of experimental data, and the properties of the compound of interest (particularly its water solubility/ surfactant nature). For molecules with a high affinity to water, such as surfactants and ionizable compounds, partitioning models need to explicitly account for sorption to the RH dependent aqueous component [Eqs. (5.42) and (5.43)]. Because condensed water is by far the most abundant condensed phase in the atmosphere, the atmospheric fate of highly water–soluble molecules and surfactants is more likely controlled by sequestration and transport by condensed water droplets (e.g., clouds) than by dry aerosols. Partitioning to water droplets can be more complex to account for, due to simultaneously occurring adsorptive, absorptive, transformation, and micellization processes, which are all influenced by a variety of parameters (e.g., temperature, ionic strength, presence of organics, etc.). Future work on understanding how these parameters and processes interplay is needed. In special situations when human health is directly related to exposure to air particles and chemicals, such as in industrial environments where there is a high risk of exposure to toxic chemicals (e.g., waste incinerators, toxic waste landfills) or with the use of drug inhalers, obtaining accurate Kip values may be a necessity. To assure accuracy, Kip values should be directly measured. If this is not possible for the compounds of interest, than the best way of estimating Kip values currently is to first measure Kip values for other compounds with the IGC method, followed by the derivation of particle-specific calibrated PP-LFERs.

CONCLUSIONS

141

TABLE 5.5. PP-LFER Descriptors for Various Apolar Semivolatile AOCs B

Va

L

S

Organochlorines 0.00b 0.00d 0.00d

0.00b 0.47d 0.50d

1.451 1.580 1.580

7.624c 7.340d 7.570d

0.99b 1.20d 1.28d

0.00e 0.00e 0.00e 0.00e

0.00e 0.00e 0.00e 0.00e

2.645 2.927 3.208 3.490

8.725f 9.743f 10.732f 11.759f

0.00e 0.00e 0.00e 0.00e

Acenaphthylene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]fluorene Benzo[b]fluorene Benz[a]anthracene Chrysene Benzo[a]pyrene Benzo[e]pyrene Benzo[k]fluoranthene

0.00c 0.00g 0.00b 0.00b 0.00h 0.00g 0.00h 0.00h 0.00b 0.00b 0.00h 0.00c 0.00c

0.20e 0.20g 0.29b 0.28b 0.20h 0.29g 0.20h 0.20h 0.35b 0.36b 0.44h 0.44e 0.44e

1.216 1.357 1.454 1.454 1.585 1.585 1.726 1.726 1.823 1.823 1.954 1.954 1.954

6.175c 6.922c 7.632c 7.568c 8.827c 8.833c 9.404c 9.520c 10.291c 10.334c 11.715c 11.656c 11.607c

1.14c 1.03g 1.29b 1.34b 1.55h 1.71g 1.59h 1.57h 1.70b 1.73b 1.98h 1.99c 1.91c

PCBs 2,20 ,4,50 -Tetrachloro 3,30 ,4,40 -Tetrachloro 2,20 ,4,5,50 -Pentachloro 2,3,30 ,4,40 -Pentachloro 2,30 ,4,40 ,5-Pentachloro 3,30 ,4,40 ,5-Pentachloro 2,20 ,3,4,40 ,50 -Hexachloro 2,20 ,4,40 ,5,50 -Hexachloro 2,20 ,3,30 ,4,40 ,6-Heptachloro

0.00i 0.00i 0.00i 0.00i 0.00i 0.00i 0.00i 0.00i 0.00i

0.15i 0.11i 0.13i 0.11i 0.11i 0.09i 0.11i 0.11i 0.09i

1.814 1.814 1.936 1.936 1.936 1.936 2.059 2.059 2.181

8.186i 9.205i 8.868i 9.594i 9.396i 9.884i 9.772i 9.587i 10.031i

1.48i 1.44i 1.61i 1.59i 1.59i 1.57i 1.74i 1.74i 1.87i

CAS or Congener Number (CN)

Compounds

118-74-1 319-84-6 58-89-9 593-45-3 112-95-8 629-97-0 646-31-1

HCB a-HCH c-HCH Alkanes n-Octadecane n-Eicosane n-Docosane n-Tetracosane

208-96-8 86-73-7 85-01-8 120-12-7 206-44-0 129-00-0 238-84-6 243-17-4 56-55-3 218-01-9 50-32-8 192-97-2 207-08-9 49 (CN) 77 (CN) 101 (CN) 105 (CN) 118 (CN) 126 (CN) 138 (CN) 153 (CN) 171 (CN)

A

PAHs

a

McGowan volume as calculated in Abraham and McGowan (1987). Torres-Lapasio et al. (2004). c Abraham (1993a). d Goss et al. (2008). e Ai, Bi, and Si values assumed identical to homolog chemicals that differ only by the amount of n-alkyl or cyclo --CH2-- groups; Li values extrapolated from homolog by adding 0.505 to per --CH2-- unit. f Mutelet and Rogalski (2001). g Abraham et al. (1994a). h Abraham et al. (1994b). i Abraham and Al-Hussaini (2005). b

For large-scale environmental models that cover months or even years, there are a wide variety of factors (e.g., weather, diverse particle sources) that cause a large heterogeneity in particle types and sorption properties, which cause variations in Kip values probably up to two orders of magnitude, or even more for certain compound classes. Here, estimated Kip values only have to be representative, such as the median or average Kip value of all measured values. Therefore, whether one predicts Kip values for molecules of low water solubility with the PP-LFER,

COSMOtherm, SPARC, or the Ki oa SP-LFER method is probably not critical. In fact, a more recent systematic study on the influence of using PP-LFERs and Ki oa SP-LFER for determining the fate of a large chemical dataset found little differences between using these two types of model paradigms, and thus concluded the “best” model to use depends on parameter availability (Brown and Wania 2009). Ideally, to better serve large-scale fate models, it would be desirable to have a better understanding on the temporal and geographical distribution of ambient ambient Ki p values. For

142


instance, once ambient Ki p for a given compound become available for hundreds of locations and dates, the median and standard deviations of these naturally occurring Ki p values could be better represented. It should be emphasized here that input parameters used in Kip predictive models need to be verified and of high quality. For instance, Pontolillo and Eganhouse (2001) reported that octanol–water partition coefficients, from which Kioa data are commonly derived, for DDT in the literature have ranged over four orders of magnitude, with recommended values ranging over two orders of magnitude. This range in Ki oa values may very well be larger than the range of naturally occurring Kip values for DDT, thus highly erroneous Ki oa-SP LFER calibrations and resulting Kip estimations can potentially result from employing inaccurate Ki oa values. An advantage with PP-LFER descriptors (listed in Table 5.5) in this regard is that, because these are determined over a large set of equations and experiments (and not just one experiment), large errors in reported PP-LFERs do not go unnoticed for very long. For the large-scale screening of compounds and for Kip predictions of transformation products and potential industrial products in which no experimental data are available, we recommend using the COSMOtherm and SPARC approaches presented here, which require only molecular structures as input. Ideally, both COSMOtherm and SPARC should be used, perhaps in combination with additional models, as comparing output datasets from several models assists in identifying outliers and the limits of chemical application domains. We recommend that future research continue to develop and utilize COSMOtherm/SPARC-type models. Ultimately, such developments will continue to branch and hold the link together between the more fundamental and applied sciences, not only conceptually but via interdisciplinary scientific collaborations. By applying fundamental theory involving quantum chemistry and thermodynamics, and combining this with microscale structural information from imaging and spectroscopic techniques, and further combining this with macroscale environmental observations, reliable and mechanistically satisfying models can be designed that shed simultaneous insight into microand macroscale processes that are of relevance to the environment and human health.

REFERENCES Abraham, M. H. and McGowan, J. C. (1987), The use of characteristic volumes to measure cavity terms in reversed phase liquid-chromatography, Chromatographia 23, 243–246. Abraham, M. H. (1993a), Hydrogen bonding XXVII. solvation parameters for functionally substituted aromatic compounds and heterocyclic compounds, from gas-liquid chromatographic data, J. Chromatogr. 644, 95–139.

Abraham, M. H. (1993b), Scales of solute hydrogen-bonding: Their construction and application to physicochemical and biochemical processes, Chem. Soc. Rev. 22, 73–83. Abraham, M. H., Andonian-Haftvan, J., Whiting, G. S., Leo, A., and Taft, R. S. (1994a), Hydrogen bonding. Part 34. The factors that influence the solubility of gases and vapours in water at 298K, and a new method for its determination, J. Chem. Soc. Perkin Trans. 28, 1777–1791. Abraham, M. H., Chada, H. S., Whiting, G. S., and Mitchell, R. C. (1994b), Hydrogen bonding. 32. An analysis of water-octanol and water-alkane partitioning and the Dlog P parameter of Seiler, J. Pharm. Sci. 83, 1085–1100. Abraham, M. H. and Al-Hussaini, A. J. M. (2001), Solvation descriptors for the polychloronaphthalenes: Estimation of some physicochemical properties, J. Environ. Monit. 3, 377–381. Abraham, M. H., Ibrahim, A., and Zissimos, A. M. (2004), Determination of sets of solute descriptors from chromatographic measurements, J. Chromatogr. A 1037, 29–47. Abraham, M. H. and Al-Hussaini, A. J. M. (2005), Solvation parameters for the 209 PCBs: Calculation of physicochemical properties, J. Environ. Monit. 7, 295–301. Aiken, A. C., DeCarlo, P. F., Kroll, J. H., Worsnop, D. R., Huffman, J. A., Docherty, K. S., Ulbrich, I. M., Mohr, C., Kimmel, J. R., Sueper, D., Sun, Y., Zhang, Q., Trimborn, A., Northway, M., Ziemann, P. J., Canagaratna, M. R., Onasch, T. B., Alfarra, M. R., Prevot, A. S. H., Dommen, J., Duplissy, J., Metzger, A., Baltensperger, U., and Jimenez, J. L. (2008), O/C and OM/OC ratios of primary, secondary, and ambient organic aerosols with highresolution time-of-flight aerosol mass spectrometry, Environ. Sci. Technol. 42, 4478–4485. Arey, J. S., Green, W. H., and Gschwend, P. M. (2005), The electrostatic origin of Abraham’s solute polarity parameter, J. Phys. Chem. B 109, 7564–7573. Arp, H. P. H. and Schmidt, T. C. (2004), Air-water transfer of MTBE, its degradation products, and alternative fuel oxygenates: The role of temperature, Environ. Sci. Technol. 38, 5405–5412. Arp, H. P. H., Goss, K.-U., and Schwarzenbach, R. P. (2006a), Evaluation of a predictive model for air/surface adsorption equilibrium constants and enthalpies, Environ. Toxicol. Chem. 25, 45–51. Arp, H. P. H., Niederer, C., and Goss, K.-U. (2006b), Predicting the partitioning behavior of various highly fluorinated compounds, Environ. Sci. Technol. 40, 7298–7304. Arp, H. P. H., Schwarzenbach, R. P., and Goss, K.-U. (2007), Equilibrium sorption of gaseous organic chemicals to fiber filters used for aerosol studies, Atmos. Environ. 41, 8241–8252. Arp, H. P. H. and Goss, K. U. (2008), Irreversible sorption of trace concentrations of perfluorocarboxylic acids to fiber filters used for air sampling, Atmos. Environ. 42, 6869–6972. Arp, H. P. H., Schwarzenbach, R. P., and Goss, K.-U. (2008a), Ambient gas/particle partitioning. 2: The influence of particle source and temperature on sorption to dry terrestrial aerosols, Environ. Sci. Technol. 42, 5951–5957. Arp, H. P. H., Schwarzenbach, R. P., and Goss, K.-U. (2008b), Ambient gas/particle partitioning. 1. Sorption mechanisms of

REFERENCES

apolar, polar and ionizable organic compounds, Environ. Sci. Technol. 42, 5541–5547. Arp, H. P. H., Schwarzenbach, R. P., and Goss, K.-U. (2008c), Determination of ambient gas-particle partitioning constants of non-polar and polar organic compounds using inverse gas chromatography, Atmos. Environ. 42, 303–312. Arp, H. P. H. and Goss, K.-U. (2009a), The gas/particle partitioning behavior of perfluorocarboxylic acids with terrestrial aerosols, Environ. Sci. Technol. 43, 8542–8547. Arp, H. P. H. and Goss, K.-U. (2009b), Ambient gas/particle partitioning. 3. Estimating partition coefficients of apolar, polar, and ionizable organic compounds by their molecular structure, Environ. Sci. Technol. 43, 1923–1929. Asher, W. E., Pankow, J. F., Erdakos, G. B., and Seinfeld, J. H. (2002), Estimating the vapor pressures of multi-functional oxygen-containing organic compounds using group contribution methods, Atmos. Environ. 36, 1483–1498. Atapattu, S. N. and Poole, C. F. (2008), Solute descriptors for characterizing retention properties of open-tubular columns of different selectivity in gas chromatography at intermediate temperatures, J. Chromatogr. A 1195, 136–145. Atkinson, D. and Curthoys, G. (1978), Determination of heats of adsorption by gas-solid chromatography, J. Chem. Educ. 55, 564–566. Bidleman, T. F. and Olney, C. E. (1974), High-volume collection of atmospheric polychlorinated biphenyls, Bull. Environ. Contam. Toxicol. 11, 442–450. Bidleman, T. F. (1988), Atmospheric processes—wet and dry deposition of organic-compounds are controlled by their vapor particle partitioning, Environ. Sci. Technol. 22, 361–367. Birch, M. E. and Cary, R. A. (1996), Elemental carbon-based method for monitoring occupational exposures to particulate diesel exhaust, Aerosol Sci. Technol. 25, 221–241. Brown, T. N. and Wania, F. (2009), Development and exploration of an organic contaminant fate model using poly-parameter linear free energy relationships, Environ. Sci. Technol. 43, 6676–6683. Burtscher, H. (2005), Physical characterization of particulate emissions from diesel engines: A review, J. Aerosol. Sci. 36, 896–932. Cavalli, F., Facchini, M. C., Decesari, S., Mircea, M., Emblico, L., Fuzzi, S., Ceburnis, D., Yoon, Y. J., O’Dowd, C. D., Putaud, J. P., and Dell’Acqua, A. (2004), Advances in characterization of sizeresolved organic matter in marine aerosol over the North Atlantic, J. Geophys. Res. Atmos. 109, D24215. Chan, M. N., Choi, M. Y., Ng, N. L., and Chan, C. K. (2005), Hygroscopicity of water-soluble organic compounds in atmospheric aerosols: Amino acids and biomass burning derived organic species, Environ. Sci. Technol. 39, 1555–1562. Chandramouli, B., Jang, M., and Kamens, R. M. (2003a), Gasparticle partitioning of semi-volatile organics on organic aerosols using a predictive activity coefficient model: Analysis of the effects of parameter choices on model performance, Atmos. Environ. 37, 853–864. Chandramouli, B., Jang, M. S., and Kamens, R. M. (2003b), Gasparticle partitioning of semivolatile organic compounds (SOCs) on mixtures of aerosols in a smog chamber, Environ. Sci. Technol. 37, 4113–4121.

143

Chang, E. I. and Pankow, J. F. (2006), Prediction of activity coefficients in liquid aerosol particles containing organic compounds, dissolved inorganic salts, and water—Part 2: Consideration of phase separation effects by an X-UNIFAC model, Atmos. Environ. 40, 6422–6436. Chow, J. C., Watson, J. G., Fujita, E. M., Lu, Z. Q., Lawson, D. R., and Ashbaugh, L. L. (1994), Temporal and spatial variations of PM(2.5) and PM(10) aerosol in the southern California airquality study, Atmos. Environ. 28, 2061–2080. Clegg, S. L., Kleeman, M. J., Griffin, R. J., and Seinfeld, J. H. (2008a), Effects of uncertainties in the thermodynamic properties of aerosol components in an air quality model—Part 1: Treatment of inorganic electrolytes and organic compounds in the condensed phase, Atmos. Chem. Phys. 8, 1057–1085. Clegg, S. L., Kleeman, M. J., Griffin, R. J., and Seinfeld, J. H. (2008b), Effects of uncertainties in the thermodynamic properties of aerosol components in an air quality model—Part 2: Predictions of the vapour pressures of organic compounds, Atmos. Chem. Phys. 8, 1087–1103. Conder, J. R. and Young, C. L. (1979), Physicochemical Measurement by Gas Chromatography, Wiley, New York. Cramer, R. D. (1980), Bc(Def) parameters. 1. Intrinsic dimensionality of intermolecular interactions in the liquid-state, J. Am. Chem. Soc. 102, 1837–1849. Dachs, J. and Eisenreich, S. J. (2000), Adsorption onto aerosol soot carbon dominates gas-particle partitioning of polycyclic aromatic hydrocarbons, Environ. Sci. Technol. 34, 3690–3697. de Boer, J. H. (1968), The Dynamic Character of Adsorption, 2nd ed., Clarendon Press, Oxford. Ding, X., Zheng, M., Yu, L. P., Zhang, X. L., Weber, R. J., Yan, B., Russell, A. G., Edgerton, E. S., and Wang, X. M. (2008), Spatial and seasonal trends in biogenic secondary organic aerosol tracers and water-soluble organic carbon in the southeastern United States, Environ. Sci. Technol. 42, 5171–5176. Dorris, G. M. and Gray, D. G. (1981), Adsorption of hydrocarbons on silica-supported water surfaces, J. Phys. Chem. 85, 3628–3635. Duarte, R., Santos, E. B. H., Pio, C. A., and Duarte, A. C. (2007), Comparison of structural features of water-soluble organic matter from atmospheric aerosols with those of aquatic humic substances, Atmos. Environ. 41, 8100–8113. Duarte, R., Silva, A. M. S., and Duarte, A. C. (2008), Twodimensional NMR studies of water-soluble organic matter in atmospheric aerosols, Environ. Sci. Technol. 42, 8224–8230. Eckert, F. and Klamt, A. (2002), Fast solvent screening via quantum chemistry: COSMO-RS approach, Am. Inst. Chem. Eng. J. 48, 369. Eckert, F. and Klamt, A. (2005), COSMOtherm v 2.1, COSMOlogic GmbH KG, Leverkusen, Germany. Endo, S., Grathwohl, P., and Schmidt, T. C. (2008), Absorption or adsorption? Insights from molecular probes n-alkanes and cycloalkanes into modes of sorption by environmental solid matrices, Environ. Sci. Technol. 42, 3989–3995. Erdakos, G. B. and Pankow, J. F. (2004), Gas/particle partitioning of neutral and ionizing compounds to single- and multi-phase aerosol particles. 2. Phase separation in liquid particulate matter

144


containing both polar and low-polarity organic compounds, Atmos. Environ. 38, 1005–1013. Erdakos, G. B., Asher, W. E., Seinfeld, J. H. and Pankow, J. F. (2006), Prediction of activity coefficients in liquid aerosol particles containing organic compounds, dissolved inorganic salts, and water—Part 1: Organic compounds and water by consideration of short- and long-range effects using X-UNIFAC.1, Atmos. Environ. 40, 6410–6421. Finizio, A., Mackay, D., Bidleman, T., and Harner, T. (1997), Octanol-air partition coefficient as a predictor of partitioning of semi-volatile organic chemicals to aerosols, Atmos. Environ. 31, 2289–2296. Fuzzi, S., Andreae, M. O., Huebert, B. J., Kulmala, M., Bond, T. C., Boy, M., Doherty, S. J., Guenther, A., Kanakidou, M., Kawamura, K., Kerminen, V. M., Lohmann, U., Russell, L. M., and Poschl, U. (2006), Critical assessment of the current state of scientific knowledge, terminology, and research needs concerning the role of organic aerosols in the atmosphere, climate, and global change, Atmos. Chem. Phys. 6, 2017–2038. Galarneau, E. and Bidleman, T. F. (2006), Modelling the temperature-induced blow-off and blow-on artefacts in filter-sorbent measurements of semivolatile substances, Atmos. Environ. 40, 4258–4268. Galarneau, E., Bidleman, T. F., and Blanchard, P. (2006), Seasonality and interspecies differences in particle/gas partitioning of PAHs observed by the integrated atmospheric deposition network (IADN), Atmos. Environ. 40, 182–197. Goss, K.-U. and Schwarzenbach, R. P. (1998), Gas/solid and gas/liquid partitioning of organic compounds: Critical evaluation of the interpretation of equilibrium constants, Environ. Sci. Technol. 32, 2025–2032. Goss, K.-U. and Schwarzenbach, R. P. (2001), Linear free energy relatinships used to evaluate equilibrium partitioning of organic compounds, Environ. Sci. Technol. 35, 1–9. Goss, K.-U. and Schwarzenbach, R. P. (2002), Adsorption of a diverse det of organic vapors on quartz, CaCO3 and alpha-Al2O3 at different relative humidities, J. Colloid Interface Sci. 252, 31–41. Goss, K.-U., Buschmann, J., and Schwarzenbach, R. P. (2003), Determination of the surface sorption properties of talc, different salts, and clay minerals at various relative humidities using adsorption data of a diverse set of organic vapors, Environ. Toxicol. Chem. 22, 2667–2672. Goss, K.-U. (2004), The air/surface adsorption equilibrium of organic compounds under ambient conditions, Crit. Rev. Environ. Sci. Technol. 34, 339–389. Goss, K.-U. (2005), Predicting the equilibrium partitioning of organic compounds using just one linear solvation energy relationship (LSER), Fluid Phase Equilib. 233, 19–22. Goss, K.-U., Arp, H. P., and Roth, C. (2005), Comment on “Model for the adsorption of organic compounds at gas-water interfaces” by C. F. Poole, [J. Environ. Monit. 7, 577 (2005)], J. Environ. Monit. 7, 1105–1106. Goss, K.-U. and Bronner, G. (2006), What is so special about the sorption behavior of highly fluorinated compounds? J. Phys. Chem. A 110, 9518–9522.

Goss, K.-U., Arp, H. P. H., Bronner, G., and Niederer, C. (2008), Partition behavior of hexachlorocyclohexane-isomers, J. Chem. Eng. Data 53, 750–754. Gotz, C. W., Scheringer, M., Macleod, M., Roth, C. M., and Hungerbuhler, K. (2007), Alternative approaches for modeling gas-particle partitioning of semivolatile organic chemicals: Model development and comparison, Environ. Sci. Technol. 41, 1272–1278. Griffin, R. J., Nguyen, K., Dabdub, D., and Seinfeld, J. H. (2003), A coupled hydrophobic-hydrophilic model for predicting secondary organic aerosol formation, J. Atmos. Chem. 44, 171–190. Gysel, M., Weingartner, E., Nyeki, S., Paulsen, D., Baltensperger, U., Galambos, I., and Kiss, G. (2004), Hygroscopic properties of water-soluble matter and humic-like organics in atmospheric fine aerosol, Atmos. Chem. Phys. 4, 35–50. Harner, T. and Bidleman, T. F. (1998), Octanol-air partition coefficient for describing particle/gas partitioning of aromatic compounds in urban air, Environ. Sci. Technol. 32, 1494– 1502. Hart, K. M., McDow, S. R., Giger, W., Steiner, D., and Burtscher, H. (1993), The correlation between in-situ, real-time aerosol photoemission intensity and particulate polycyclic aromatic hydrocarbon concentration in combustion aerosols, Water Air Soil Pollut. 68, 75–90. Hilal, S. H., Karickhoff, S. W., and Carreira, L. A. (2003), Prediction of the vapor pressure boiling point, heat of vaporization and diffusion coefficient of organic compounds, Quant. Struct.–Act. Relat. Combin. Comb. Sci. 22, 565–574. Hilal, S. H., Karickhoff, S. W. and Carreira, L. A. (2004), Prediction of the solubility, activity coefficient and liquid/liquid partition coefficient of organic compounds. Quant. Struct.–Act. Relat. Combin. Comb. Sci. 23, 709–720. Hildemann, L. M., Markowski, G. R., and Cass, G. R. (1991), Chemical-composition of emissions from urban sources of fine organic aerosol, Environ. Sci. Technol. 25, 744–759. Hinds, W. C. (1999), Aerosol Technology: Properties, Behavior and Measurement of Airborne Particles, 2nd ed., WileyInterscience, New York. Hueglin, C., Gehrig, R., Baltensperger, U., Gysel, M., Monn, C., and Vonmont, H. (2005), Chemical characterisation of PM2.5, PM10 and coarse particles at urban, near-city and rural sites in Switzerland, Atmos. Environ. 39, 637–651. Jacobson, M. Z. (2001), Strong radiative heating due to the mixing state of black carbon in atmospheric aerosols, Nature 409, 695–697. Johnson, K. S., Zuberi, B., Molina, L. T., Molina, M. J., Iedema, M. J., Cowin, J. P., Gaspar, D. J., Wang, C., and Laskin, A. (2005), Processing of soot in an urban environment: Case study from the Mexico Citymetropolitanarea,Atmos.Chem. Phys.5, 3033–3043. Jones, N. C., Thornton, C. A., Mark, D., and Harrison, R. M. (2000), Indoor/outdoor relationships of particulate matter in domestic homes with roadside, urban and rural locations, Atmos. Environ. 34, 2603–2612. Jonker, M. T. O. and Koelmans, A. A. (2002), Extraction of polycyclic aromatic hydrocarbons from soot and sediment: Solvent evaluation and implications for sorption mechanism, Environ. Sci. Technol. 36, 4107–4113.

REFERENCES

Jonker, M. T. O., Hawthorne, S. B., and Koelmans, A. A. (2005), Extremely slowly desorbing polycyclic aromatic hydrocarbons from soot and soot-like materials: Evidence by supercritical fluid extraction, Environ. Sci. Technol. 39, 7889–7895. Kalberer, M., Paulsen, D., Sax, M., Steinbacher, M., Dommen, J., Prevot, A. S. H., Fisseha, R., Weingartner, E., Frankevich, V., Zenobi, R., and Baltensperger, U. (2004), Identification of polymers as major components of atmospheric organic aerosols, Science 303, 1659–1662. Kavouras, I. G. and Stephanou, E. G. (2002), Gas/particle partitioning and size distribution of primary and secondary carbonaceous aerosols in public buildings, Indoor Air—Int. J. Indoor Air Qual. Clim. 12, 17–32. Khlystov, A., Stanier, C. O., Takahama, S. and Pandis, S. N. (2005), Water content of ambient aerosol during the Pittsburgh air quality study, J. Geophys. Res. Atmos. 110. Kiss, G., Varga, B., Galambos, I., and Ganszky, I. (2002), Characterization of water-soluble organic matter isolated from atmospheric fine aerosol, J. Geophys. Res. Atmos. 107. Kleeman, M. J., Hughes, L. S., Allen, J. O., and Cass, G. R. (1999), Source contributions to the size and composition distribution of atmospheric particles: Southern California in September 1996, Environ. Sci. Technol. 33, 4331–4341. Koehler, K. A., Kreidenweis, S. M., DeMott, P. J., Prenni, A. J., Carrico, C. M., Ervens, B., and Feingold, G. (2006), Water activity and activation diameters from hygroscopicity data—Part II: Application to organic species, Atmos. Chem. Phys. 6, 795–809. Krivacsy, Z., Gelencser, A., Kiss, G., Meszaros, E., Molnar, A., Hoffer, A., Meszaros, T., Sarvari, Z., Temesi, D., Varga, B., Baltensperger, U., Nyeki, S., and Weingartner, E. (2001), Study on the chemical character of water soluble organic compounds in fine atmospheric aerosol at the Jungfraujoch, J. Atmos. Chem. 39, 235–259. Lee, K. W. and Makund, R. (2001), Filter collection, in Aerosol Measurement: Principles, Techniques and Applications, Baron P. A. and K., Willeke, eds., 2nd ed., Wiley, New York, PP. 197–228. Lei, Z. G., Chen, B. H., Li, C. Y., and Liu, H. (2008), Predictive molecular thermodynamic models for liquid solvents, solid salts, polymers, and ionic liquids, Chem. Rev. 108, 1419–1455. Liu, T. and Oberg, T. (2009), Modelling of partition constants: Linear solvation energy relationships or PLS regression? J. Chemom. 23, 254–262. Lohmann, R. and Lammel, G. (2004), Adsorptive and absorptive contributions to the gas-particle partitioning of polycyclic aromatic hydrocarbons: State of knowledge and recommended parametrization for modeling, Environ. Sci. Technol. 38, 3793–3803. Mackay, D., Paterson, S., and Schroeder, W. H. (1986), Model describing the rates of transfer processes of organic-chemicals between atmosphere and water, Environ. Sci. Technol. 20, 810–816. Mader, B. T., Flagan, R. C., and Seinfeld, J. H. (2001), Sampling atmospheric carbonaceous aerosols using a particle trap impactor/denuder sampler, Environ. Sci. Technol. 35, 4857–4867.

145

Mader, B. T. and Pankow, J. F. (2001), Gas/solid partitioning of semivolatile organic compounds (SOCs) to air filters. 3. An analysis of gas adsorption artifacts in measurements of atmospheric SOCs and organic carbon (OC) when using Teflon membrane filters and quartz fiber filters, Environ. Sci. Technol. 35, 3422–3432. Mader, B. T. and Pankow, J. F. (2002), Study of the effects of particle-phase carbon on the gas/particle partitioning of sernivolatile organic compounds in the atmosphere using controlled field experiments, Environ. Sci. Technol. 36, 5218–5228. Marcolli, C. and Krieger, U. K. (2006), Phase changes during hygroscopic cycles of mixed organic/inorganic model systems of tropospheric aerosols, J. Phys. Chem. A 110, 1881–1893. Mintz, C., Clark, M., Acree, W. E., and Abraham, M. H. (2007), Enthalpy of solvation correlations for gaseous solutes dissolved in water and in 1-octanol based on the Abraham model, J. Chem Inf. Model. 47, 115–121. Mutelet, F. and Rogalski, M. (2001), Experimental determination and prediction of the gas-liquid n-hexadecane partition coefficients, J. Chromatogr. A 923, 153–163. Naumova, Y. Y., Offenberg, J. H., Eisenreich, S. J., Meng, Q. Y., Polidori, A., Turpin, B. J., Weisel, C. P., Morandi, M. T., Colome, S. D., Stock, T. H., Winer, A. M., Alimokhtari, S., Kwon, J., Maberti, S., Shendell, D., Jones, J., and Farrar, C. (2003), Gas/particle distribution of polycyclic aromatic hydrocarbons in coupled outdoor/indoor atmospheres, Atmos. Environ. 37, 703–719. Niederer, C., Goss, K.-U., and Schwarzenbach, R. P. (2006), Sorption equilibrium of a wide spectrum of organic vapors in leonardite humic acid: Experimental setup and experimental data, Environ. Sci. Technol. 40, 5368–5373. Norbeck, J. M., Durbin, T. D., and Truex, T. J. (1998), Measurement of Primary Particulate Matter Emissions from Light-Duty Motor Vehicles, Final Report for CRC Project E-24-2. Oberdorster, G., Sharp, Z., Atudorei, V., Elder, A., Gelein, R., Kreyling, W., and Cox, C. (2003), Translocation of inhaled ultrafine particles to the brain, Proc. 4th Colloquium on PM and Human Health, Pittsburgh, PA, PP. 437–445. O’Dowd, C. D., Facchini, M. C., Cavalli, F., Ceburnis, D., Mircea, M., Decesari, S., Fuzzi, S., Yoon, Y. J., and Putaud, J. P. (2004), Biogenically driven organic contribution to marine aerosol, Nature 431, 676–680. Oppo, C., Bellandi, S., Innocenti, N. D., Stortini, A. M., Loglio, G., Schiavuta, E., and Cini, R. (1999), Surfactant components of marine organic matter as agents for biogeochemical fractionation and pollutant transport via marine aerosols, Mar. Chem. 63, 235–253. Pankow, J. F. (1987), Review and comparative-analysis of the theories on partitioning between the gas and aerosol particulate phases in the atmosphere, Atmos. Environ. 21, 2275–2283. Pankow, J. F. and Bidleman, T. F. (1991), Effects of temperature, tsp and per cent nonexchangeable material in determining the gas particle partitioning of organic compounds, Atmos. Environ. Pt. A-General. Topics. 25, 2241–2249. Pankow, J. F. and Bidleman, T. F. (1992), Interdependence of the slopes and intercepts from log-log correlations of measured gas

146


particle partitioning and vapor-pressure.1. Theory and analysis of available data, Atmos. Environ. Pt. A—General Topics. 26, 1071–1080. Pankow, J. F. (1994a), An absorption-model of the gas aerosol partitioning involved in the formation of secondary organic aerosol, Atmos. Environ. 28, 189–193. Pankow, J. F. (1994b), An absorption-model of gas-particle partitioning of organic-compounds in the atmosphere, Atmos. Environ. 28, 185–188. Pankow, J. F. (2003), Gas/particle partitioning of neutral and ionizing compounds to single- and multi-phase aerosol particles. 1. Unified modeling framework, Atmos. Environ. 37, 4993–4993. Pankow, J. F., Isabelle, L. M., Buchholz, D. A., Luo, W. T., and Reeves, B. D. (1994), Gas-particle partitioning of polycyclic aromatic-hydrocarbons and alkanes to environmental tobaccosmoke, Environ. Sci. Technol. 28, 363–365. Pontolillo, J. and Eganhouse, R. P. (2001), The search for reliable aqueous solubility (Sw) and octanol-water partition coefficient (Kow) data for hydrophobic organic compounds: DDT and DDE as a case study, in Water-Resources Investigations Report, U.S. Department of the Interior, U.S. Geological Survey, Reston, VA, p. 50. Pun, B. K., Griffin, R. J., Seigneur, C., and Seinfeld, J. H. (2002), Secondary organic aerosol—2. Thermodynamic model for gas/ particle partitioning of molecular constituents, J. Geophys. Res. Atmos. 107, 1–14. Renner, R. (2002), The K-ow controversy, Environ. Sci. Technol. 36, 410A–413A. Roth, C. M., Goss, K.-U., and Schwarzenbach, R. P. (2002), Adsorption of a diverse set of organic vapors on the bulk water surface, J. Colloid Interface Sci. 252, 21–30. Roth, C. M., Goss, K.-U., and Schwarzenbach, R. P. (2004), Sorption of diverse organic vapors to snow, Environ. Sci. Technol. 38, 4078–4084. Roth, C. M., Goss, K.-U., and Schwarzenbach, R. P. (2005a), Sorption of a diverse set of organic vapors to urban aerosols, Environ. Sci. Technol. 39, 6638–6643. Roth, C. M., Goss, K.-U., and Schwarzenbach, R. P. (2005b), Sorption of a diverse set of organic vapors to diesel soot and road tunnel aerosols, Environ. Sci. Technol. 39, 6632–6637. Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T. (1999), Measurement of emissions from air pollution sources. 2. C-1 through C-30 organic compounds from medium duty diesel trucks, Environ. Sci. Technol. 33, 1578–1587. Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T. (2001), Measurement of emissions from air pollution sources. 3. C-1-C-29 organic compounds from fireplace combustion of wood, Environ. Sci. Technol. 35, 1716–1728. Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T. (2002a), Measurement of emissions from air pollution sources. 4. C-1-C-27 organic compounds from cooking with seed oils, Environ. Sci. Technol. 36, 567–575. Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T. (2002b), Measurement of emissions from air pollution sources. 5. C-1-C-32 organic compounds from gasoline-powered motor vehicles, Environ. Sci. Technol. 36, 1169–1180.

Schauer, J. J., Mader, B. T., Deminter, J. T., Heidemann, G., Bae, M. S., Seinfeld, J. H., Flagan, R. C., Cary, R. A., Smith, D., Huebert, B. J., Bertram, T., Howell, S., Kline, J. T., Quinn, P., Bates, T., Turpin, B., Lim, H. J., Yu, J. Z., Yang, H., and Keywood, M. D. (2003), ACE-Asia intercomparison of a thermal-optical method for the determination of particle-phase organic and elemental carbon, Environ. Sci. Technol. 37, 993–1001. Schmid, H., Laskus, L., Abraham, H. J., Baltensperger, U., Lavanchy, V., Bizjak, M., Burba, P., Cachier, H., Crow, D., Chow, J., Gnauk, T., Even, A., ten Brink, H. M., Giesen, K. P., Hitzenberger, R., Hueglin, C., Maenhaut, W., Pio, C., Carvalho, A., Putaud, J. P., Toom-Sauntry, D., and Puxbaum, H. (2001), Results of the “carbon conference” international aerosol carbon round robin test stage I, Atmos. Environ. 35, 2111–2121. Schwarzenbach, R. P., Gschwend, P. M., and Imboden, D. M. (2003), Environmental Organic Chemistry, 2nd ed., Wiley, Hoboken, NJ. Seinfeld, J. H. and Pandis, S. N. (2006), Atmospheric Chemistry and Physics: From Air Pollution to Climate Change, 2nd ed., Wiley-Interscience, Hoboken, NJ. Seinfeld, J. H. and Pankow, J. F. (2003), Organic atmospheric particulate material, Annu. Rev. Phys. Chem. 54, 121–140. Sheffield, A. E. and Pankow, J. F. (1994), Specific surface-area of urban atmospheric particulate matter in Portland, Oregon, Environ. Sci. Technol. 28, 1759–1766. Shiraiwa, M., Kondo, Y., Moteki, N., Takegawa, N., Miyazaki, Y., and Blake, D. R. (2007), Evolution of mixing state of black carbon in polluted air from Tokyo, Geophys. Res. Lett. 34, 5. Sillanpaa, M., Frey, A., Hillamo, R., Pennanen, A. S., and Salonen, R. O. (2005), Organic, elemental and inorganic carbon in particulate matter of six urban environments in Europe, Atmos. Chem. Phys. 5, 2869–2879. Tabazadeh, A. (2005), Organic aggregate formation in aerosols and its impact on the physicochemical properties of atmospheric particles, Atmos. Environ. 39, 5472–5480. Taft, R. W., Abboud, J. L. M., Kamlet, M. J., and Abraham, M. H. (1985), Linear solvation energy relations, J. Solution Chem. 14, 153–186. Tang, I. N. and Munkelwitz, H. R. (1993), Composition and temperature-dependence of the deliquescence properties of hygroscopic aerosols, Atmos. Environ. Pt. A—General Topics 27, 467–473. Tao, S., Liu, Y., Xu, W., Lang, C., Liu, S., Dou, H. and Liu, W. (2007), Calibration of a passive Sampler for both gaseous and particulate phase polycyclic aromatic hydrocarbons, Environ. Sci. Technol. 41 568–573. Tong, C. H., Clegg, S. L., and Seinfeld, J. H. (2008), Comparison of activity coefficient models for atmospheric aerosols containing mixtures of electrolytes, organics, and water, Atmos. Environ. 42, 5459–5482. Torres-Lapasio, J. R., Garcia-Alvarez-Coque, M. C., Roses, M., Bosch, E., Zissimos, A. M., and Abraham, M. H. (2004), Analysis of a solute polarity parameter in reversed-phase liquid chromatography on a linear solvation relationship basis, Anal. Chim. Acta 515, 209–227.

REFERENCES

Tulp, H. C., Goss, K. U., Schwarzenbach, R. P., and Fenner, K. (2008), Experimental determination of LSER parameters for a set of 76 diverse pesticides and pharmaceuticals, Environ. Sci. Technol. 42, 2034–2040. Volckens, J. and Leith, D. (2003a), Comparison of methods for measuring gas-particle partitioning of semivolatile compounds, Atmos. Environ. 37, 3177–3188. Volckens, J. and Leith, D. (2003b), Effects of sampling bias on gasparticle partitioning of semivolatile compounds, Atmos. Environ. 37, 3385–3393. Wang, H. B., Kawamura, K., and Shooter, D. (2005), Carbonaceous and ionic components in wintertime atmospheric aerosols from two New Zealand cities: Implications for solid fuel combustion, Atmos. Environ. 39, 5865–5875. Weschler, C. J., Salthammer, T., and Fromme, H. (2008), Partitioning of phthalates among the gas phase, airborne particles and settled dust in indoor environments, Atmos. Environ. 42, 1449–1460. Yang, H., Yu, J. Z., Ho, S. S. H., Xu, J. H., Wu, W. S., Wan, C. H., Wang, X. D., Wang, X. R., and Wang, L. S. (2005), The chemical composition of inorganic and carbonaceous materials in PM2.5 in Nanjing, China, Atmos. Environ. 39, 3735–3749.

147

Yao, X. H., Ling, T. Y., Fang, M., and Chan, C. K. (2006), Comparison of thermodynamic predictions for in situ pH in PM2.5, Atmos. Environ. 40, 2835–2844. Yttri, K. E., Aas, W., Bjerke, A., Cape, J. N., Cavalli, F., Ceburnis, D., Dye, C., Emblico, L., Facchini, M. C., Forster, C., Hanssen, J. E., Hansson, H. C., Jennings, S. G., Maenhaut, W., Putaud, J. P., and Torseth, K. (2007), Elemental and organic carbon in PM10: A one year measurement campaign within the European Monitoring and Evaluation Programme EMEP, Atmos. Chem. Phys. 7, 5711–5725. Zhang, Q., Jimenez, J. L., Worsnop, D. R., and Canagaratna, M. (2007), A case study of urban particle acidity and its influence on secondary organic aerosol, Environ. Sci. Technol. 41, 3213–3219. Zielinska, B. (2005), Atmospheric transformation of diesel emissions, Exp. Toxicol. Pathol. 57, 31–42. Zissimos, A. M., Abraham, M. H., Klamt, A., Eckert, F. and Wood, J. (2002), A comparison between the two general sets of linear free energy descriptors of Abraham and Klamt, J. Chem. Inf. Comput. Sci. 42, 1320–1331. Zuend, A., Marcolli, C., Luo, B. P., and Peter, T. (2008), A thermodynamic model of mixed organic-inorganic aerosols to predict activity coefficients, Atmos. Chem. Phys. 8, 4559–4593.

6 MEASUREMENT AND MODELING OF SEMIVOLATILE ORGANIC COMPOUNDS IN LOCAL ATMOSPHERES SONGYAN DU AND LISA A. RODENBURG 6.1. Introduction 6.2. Some Important Semivolatile Organic Compound (SVOC) Classes 6.2.1. Persistent Organic Pollutants 6.2.2. Polycyclic Aromatic Hydrocarbons 6.3. Cycling in the Atmosphere 6.3.1. Atmospheric Deposition 6.3.2. Chemical Reaction of SVOCs in the Atmosphere 6.4. Monitoring Programs 6.5. Sampling and Analysis 6.5.1. Sampling 6.5.2. Analysis 6.5.3. Challenges 6.6. Source Identification 6.6.1. Diagnostic Ratios and Fingerprints 6.6.2. Other Techniques of Source Identification 6.6.3. Receptor Models 6.6.4. Transport Models 6.7. Conclusions

6.1. INTRODUCTION Semivolatile organic compounds (SVOCs) are defined as the substances with vapor pressures roughly between 104 and 1011 atm (101–106 Pa) at ambient temperatures (Bidleman 1988). At these vapor pressures, significant fractions of their masses in the atmosphere are found in both the particle and the gas phases. This partitioning has implications

for the compounds’ transport and reactivity, as well as the techniques used to measure them. With such a broad definition, it is no surprise that SVOCs constitute a diverse class of chemicals. Semivolatile organic compounds include a wide variety of anthropogenic and naturally occurring chemicals; a comprehensive review of the major SVOC classes is provided by Lee and Nicholson (1994). Our discussion focuses mainly on the organochlorine compounds and aromatic hydrocarbons, more specifically, polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs), organochlorine pesticides (OCPs), and polychlorinated dibenzo-p-dioxins and -furans (PCDD/Fs). Semivolatile organic compounds have many environmental impacts. Atmospheric chemists are concerned primarily with SVOCs as sinks for OH radicals and possible precursors for ozone production and particle nucleation (Finlayson-Pitts and Pitts Jr. 1997). Public health professionals are concerned with SVOCs because some of them, such as PAHs, have human health impacts (IARC 1994). Environmental chemists are concerned with SVOCs because many of them are persistent organic pollutants (POPs), or are subject to air or water quality standards (AQSs or WQSs) (Table 6.1). This chapter provides background on the importance of SVOCs, and then focuses on the methods used to understand their sources and fate on a local scale. Many of these methods rely on the analysis of large datasets, so that issues of data comparability across various monitoring networks and programs are important. For this reason, this chapter will also briefly discuss the methods used to measure and monitor SVOCs in the atmosphere.


149

150

MEASUREMENT AND MODELING OF SEMIVOLATILE ORGANIC COMPOUNDS IN LOCAL ATMOSPHERES

6.2. SOME IMPORTANT SEMIVOLATILE ORGANIC COMPOUND (SVOC) CLASSES 6.2.1. Persistent Organic Pollutants International concerns over chlorinated organic pollutants culminated in the signing of the Stockholm Convention on persistent organic pollutants (POPs), a global agreement on POPs that entered force in 2004. In total, 12 substances were identified as POPs under the original Stockholm Convention, including several OCPs [aldrin, dieldrin, dichlorodiphenyltrichloroethane (DDT), endrin, chlordane, heptachlor, mirex, and toxaphene], as well as the industrial chemicals hexachlorobenzene (HCB), PCBs, and PCDD/Fs (UNEP 2001). The vapor pressures of these POPs range from about 100.3 Pa (monochlorobiphenyls) to as low as 1010 Pa (octachlorodibenzo-p-dioxin), such that most of them qualify as “semivolatile.” The criteria that are generally applied in identifying a chemical as a POP are inherent toxicity, bioaccumulation potential, persistence, and susceptibility to longrange atmospheric transport (LRAT). The United Nations Environment Programme (UNEP) has called for actions to limit the production and release of these so-called “dirty dozen” POPs. Because of their LRAT potential, these chemicals may be transported from countries where they are still in use to countries where they are banned, potentially causing transboundary environmental damage and political disputes. In the European Union, the requirements of two international conventions (the 1979 Convention and the Stockholm Convention) have been implemented by Regulation (EC) 850/2004 on POPs. This regulation also amends the EU regulation on POPs and Directive 79/117/EEC, which prohibits the marketing and use of plant protection products containing certain active ingredients. In the United States, the only existing air standards for SVOCs are

typically set by the Occupational Safety and Health Administration (OSHA) and are designed to protect workers from high (generally indoor) acute exposures, so the standards are high. For example, the OSHA limit for the PCB formulation Aroclor 1254 is 500,000,000 pg/m3 (http:// www.osha.gov/pls/oshaweb/owadisp.show_document? p_table¼STANDARDS&p_id¼9992). Polychlorinated biphenyls have never been reported to exceed that level in ambient air. However, in the United States, all the dirty dozen except mirex are subject to water quality standards (WQSs) (http://www.epa.gov/waterscience/criteria/wqctable/index. html; Table 6.1). The federal WQS for the consumption of water and organisms for PCBs is 64 pg/L (http://www. epa.gov/waterscience/criteria/wqctable/index.html). According to Henry’s law constants for PCBs (Schenker et al. 2005; Bamford et al. 2002b), the concentration in air that would be at equilibrium with this water concentration is about 500 pg/m3, a level that is routinely exceeded in urban areas (Simcik et al. 1997; Totten et al. 2004; Sun et al. 2007). Similarly, heptachlor epoxide, dieldrin, and DDTs can sometimes approach or exceed atmospheric concentrations that are in equilibrium with the WQS (Gioia et al. 2005). In such circumstances, the WQS cannot be met without reducing atmospheric emissions, and studying the atmospheric fate and transport of these chemicals is crucial to managing them in aquatic systems. In addition, “new” or “emerging” substances were more recently added to the UNEP’s list of POPs. These new POPs are a- and b-hexachlorocyclohexane (HCH); chlordecone; tetra-, penta-, hexa-, and hepta- bromodiphenyl ethers (BDEs); lindane (c-HCH); pentachlorobenzene; and perfluorooctane sulfonic acid (and its salts and perfluorooctane sulfonyl fluoride). Many of these compounds have vapor pressures that fall in the semivolatile range, including BDEs

TABLE 6.1. National Air and Water Quality Standards and Calculated Air Concentrations in Equilibrium with Water Quality Standards for Some of the SVOCs

Pollutant PCBs Chlordane 4,4’-DDT Dieldrin a-Endosulfan b-Endosulfan Endrin Heptachlor Heptachlor epoxide Toxaphene a

Henry’s Law,a (Pa m3)/mol 4–100 9.0 1.3 4.5 10. 1.9 0.76 150 3.3 490

Molecular Weight, g/mol

U.S. National Water Quality Standard Water þ Organismb mg/L

188.65–464.2 409.8 354.5 380.93 406.95 406.95 380.92 373.3 389.2 414

0.000064 0.00080 0.00022 0.000052 62 62 0.059 0.000079 0.000039 0.00028

Henry’s law constant from Mackay et al. (1999) except for PCBs from Schenker et al. (2005). http://www.epa.gov/waterscience/criteria/wqctable/index.html. c http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table¼STANDARDS&p_id¼9992. b

Concentration in Air at Equilibrium with WQS, pg/m3

U.S. OSHA Standards, pg/m3c

500 2,900 120 94 260,000,000 48,000,000 18,000 4,800 51 55,000

1,000,000,000 500,000,000 1,000,000,000 250,000,000 — — 100,000,000 500,000,000 — —

CYCLING IN THE ATMOSPHERE

(103–107 Pa) (Shoeib et al. 2004) and HCHs (0.1–0.01 Pa) (Monosmith and Hermanson 1996). Many of the POPs noted in the Stockholm Convention are considered potentially toxic for humans. The Agency for Toxic Substances and Disease Registry (ATSDR) provides toxicity profiles for most POPs (http://www.atsdr.cdc.gov). On the basis of these profiles, dioxins, including PCDDs and PCDFs, have been characterized as likely human carcinogens. Some of the non- or mono-ortho-substituted PCBs that share a similar chemical structure and a common mechanism of toxic action (Molina et al. 2000) are also considered to be “dioxin-like.” The health concerns for PCBs include carcinogenicity and a variety of adverse health effects on the immune, reproductive, nervous, and endocrine systems. The adverse health effects associated with the OCPs include the fact that they are probable carcinogens. There is also evidence that they may harm the endocrine, nervous, and digestive systems. Health effects caused by HCB include harm to the liver, immune system, kidneys, and blood. All BDEs are known to have neurological effects and are suspected to affect the immune system. 6.2.2. Polycyclic Aromatic Hydrocarbons Another important class of SVOCs is PAHs, which are produced mainly from combustion processes and petrogenic sources (e.g., coal, oils, fossil fuels). Some PAHs are produced naturally from forest fires and volcanic eruptions, but anthropogenic emissions of PAHs from fossil fuel burning tend to dominate in most areas (Wild and Jones 1995). Petrogenic sources are the main source of PAHs in areas impacted by oil spills and fossil fuel contamination (e.g., by unburned coal), such as shipping ports and areas around oil refineries. The major health concerns for PAHs and substituted PAHs (e.g., nitro-PAHs) are their mutagenicity and carcinogenicity (Atkinson and Arey 1994). Polycyclic aromatic hydrocarbons were the first class of atmospheric pollutants to have been identified as suspected carcinogens. Their carcinogenicity appears to increase with increasing molecular weight. In contrast, acute toxicity apparently decreases with increasing molecular weight (Ravindra et al. 2001, 2008). The International Agency for Research on Cancer (IARC) has classified the carcinogenicity of 49 PAH into a few groups. Three PAHs are considered as probable carcinogens (class 2A): benzo[a]anthracene, benzo[a]pyrene, and dibenzo[a,h] anthracene. Twelve other PAHs are considered as possible carcinogens (class 2B), including benzo[b]fluoranthene, benzo[j]fluoranthene, benzo[k]fluoranthene, and indeno[1,2,3cd]pyrene. Other PAH compounds that can neither be classified as nor excluded as carcinogenic are listed as group 3 (IARC 1994). Polycyclic aromatic hydrocarbons are also major constituents of soot particles and are involved in the production of secondary organic aerosols (Wild and Jones 1995; Dachs and Eisenreich 2000; Vione

151

et al. 2004). Both these types of particles also have human health impacts. For instance, soot can cause tissue irritation and the release of toxic chemicals from scavenger cells (Lighty et al. 2000). Secondary organic aerosols influence human health through the spread of microorganisms and also cause the enhancement of of mutagenicity and allergenicity by nitration of PAHs in soot and proteins in bioparticles, respectively (Fuzzi et al. 2006). Furthermore, both soot and secondary organic aerosols have important implications on climate and climate change, and thus influence human health in this way as well (Fuzzi et al. 2006). Many PAHs undergo oxidation and/or nitration in the atmosphere, generating toxic products (Vione et al. 2006). The PAHs differ from other classic POPs in their shorter atmospheric half-lives (due to faster reactions with OH radical), their strong affinity for sorption to soot carbon, and the influence of ongoing primary sources on their distribution (Nizzetto et al. 2008). They are therefore a complementary group of compounds for investigating the role of atmospheric persistence on environmental partitioning mechanisms of POPs (Nizzetto et al. 2008).

6.3. CYCLING IN THE ATMOSPHERE Atmospheric transport has been identified as the major mode of long-range transport and global dispersal of most legacy SVOCs, such as hexachlorobenzene (HCB), hexachlorocyclohexane (HCH), and PCBs. These chemicals are capable of being transported from source areas to extremely pristine and remote areas, such as the arctic (Risebrough et al. 1976; Bidleman et al. 1989; Simonich and Hites 1995). Chapter 10 deals with the global cycling of POPs in more detail. Localscale atmospheric transport of POPs is also extremely important to understand, and is the focus of this chapter. High atmospheric concentrations lead to large atmospheric deposition fluxes that can be important and sometimes dominant sources of POPs to aquatic systems (Offenberg and Baker 1997; Totten et al. 2004, 2006a). One of the most famous examples is from Chicago, Illinois, where atmospheric deposition of PCBs into the adjacent Lake Michigan is thought to be one of the largest sources of PCBs to the lake (Zhang et al. 1999; Offenberg et al. 2005). The case in Delaware River is another instance where the atmospheric deposition load exceeds the entire total maximum daily load (TMDL) recently established for PCBs (Fikslin and Suk 2003; Totten et al. 2006a). Semivolatile organic compounds that are emitted to ambient air are subsequently removed from the atmosphere by different processes. The atmospheric lifetimes of SVOCs determine their LRAT potential and overall persistence. Their major atmospheric removal processes include deposition and degradation. Even within the same chemical class, the relative importance of these removal processes can be

152


very compound-specific. Chapter 10 discusses these processes in more detail. This section therefore includes only a brief overview of these removal processes, with emphasis on their implications for measurement and modeling of SVOCs. 6.3.1. Atmospheric Deposition Atmospheric deposition occurs via three major processes: wet deposition, dry deposition of fine/coarse particles, and gaseous deposition. Wet deposition occurs via two mechanisms: the gaseous absorption of SVOCs into raindrops and the scavenging of particles containing SVOCs by rain and snow (Bidleman 1988). Semivolatile organic compounds can also ad/absorb to rain or fog droplets, which have a high surface area/volume ratio. Ad/absorption onto fog droplets (i.e., fog sequestration) is important near coastal areas. Typically, the particle scavenging mechanism dominates over gas absorption into the droplet due to the relatively high Henry’s law constants of many apolar to weakly polar SVOCs (Poster and Baker 1996; Offenberg and Baker 2002). Scavenging of particles by rain/fog droplets is non-size-discriminatory. In contrast, dry particle deposition preferentially removes large particles with high settling velocities. Unlike wet and dry particle deposition, gaseous deposition is a reversible process that is more correctly referred to as air–surface exchange. The gas/particle partitioning behavior of the SVOC is the primary factor that determines which mode of atmospheric deposition will dominate its overall removal from the atmosphere (see Chapter 5 for a comprehensive look at gas particle partitioning). Any SVOCs that reside primarily in the gas phase are most likely to be removed from the atmosphere via air–surface exchange, while dry particle deposition is typically more important for chemicals that have a strong affinity for aerosols. For example, the majority of PAHs (70%–90%) are bound to suspended particles at ambient temperature (Ravindra et al. 2008), therefore, particle-associated deposition tends to dominate their atmospheric removal. For more polar chemicals, wet deposition dominates their atmospheric removal (Gotz et al. 2008). Studies comparing the different removal mechanisms for major SVOCs have been conducted. For heavier PCBs, PCDD/Fs, and some of the OCPs such as DDT, their atmospheric fates are controlled by particle-associated deposition (Koester and Hites 1992; Lohmann and Jones 2005; Gotz et al. 2008). Wet deposition has been observed to be the most efficient process for the atmospheric removal of highly brominated BDE congeners and PCDD/Fs (Bidleman 1988; Hoff et al. 1996; Baker and Hites 2000; Raff and Hites 2007). In contrast, gaseous transfer is the dominant transfer process for some OCPs such as the HCHs, dieldrin, and lighter PCBs (Hoff et al. 1996). Atmospheric deposition always competes with atmospheric degradation. For example, the less volatile hexa-, hepta-, and octa-PCDDs are removed from the atmosphere primarily via deposition to terrestrial surfaces, but for

the more volatile tetra- and penta-PCDD/Fs, atmospheric degradation is predicted to outweigh deposition (Lohmann and Jones 2005). Quantifying atmospheric deposition is difficult. Much of the monitoring of SVOCs that has been conducted to date has been performed in order to quantify their atmospheric inputs to aquatic ecosystems for the purposes of building loads for water quality models. Therefore these monitoring programs must measure the target SVOCs in the gas, aerosol, and precipitation phases. The methodology must be able to accurately quantify SVOCs in the gas versus aerosol phases, which can be difficult since gas/particle partitioning can change during sampling (see Chapter 5). Similarly, measuring SVOCs in precipitation can be subject to artifacts related to accurately separating dry particle deposition from wet deposition. These artifacts are discussed in more detail in Section 6.5. Assuming that the monitoring program is able to accurately measure SVOC concentrations in the gas, aerosol, and precipitation phases, these concentrations must then be converted to deposition fluxes in order to estimate total atmospheric loads of SVOCs to water bodies or terrestrial surfaces. 6.3.1.1. Dry Particle Deposition. The dry particle deposition flux (Fdry ) is calculated by multiplying the particle-phase SVOC concentration (Cpart ) by the particle phase deposition velocity (vd): Fdry ¼ vd Cpart

ð6:1Þ

Since the settling rate and therefore deposition velocity of particles depends on their size, Equation (6.1) would be more correctly written as n X Fdry ¼ Ci vd;i ð6:2Þ i¼1

where the concentration of SVOC is measured on each particle size fraction (Ci ), and the appropriate deposition velocity (vd,i) for each size range is employed. In practice, this is seldom done. Measurement of SVOCs is typically expensive, and for long-term monitoring studies it is generally not feasible to measure SVOCs in several particle size ranges. Also, detection limits can be a problem when the total particle phase SVOC burden is split into many fractions. Therefore most studies use Equation (6.1) and attempt to identify a characteristic deposition velocity that will reflect an average of the deposition velocities of the various particle size fractions, weighted by the percentage of the total SVOC mass contained on each fraction. However, because different SVOCs have different sources and different gas/particle partitioning behavior, they may reside in different particle size fractions and therefore have different characteristic deposition velocities. Also, the particle size distribution can


153

TABLE 6.2. Dry Deposition Velocity for SVOCs Associated with Particulate Phase Location

Method

Chicago Tainan City, Taiwan Rural Taiwan Lake Michigan San Francisco Bay

Dry deposition plates Back-calculated from flux Back-calculated from flux Mylar strips, Apezion L grease Estimation based on wind speed, particle size Dry deposition plates Dry deposition plates Bucket Water vessel Dry deposition plates Dry deposition plates Glass jars Dry deposition plates Inverted frisbees and flat-plate samplers Glass jars Derived from measured concentration in vegetation Dry deposition plates

Chicago Japan

Chicago Canadian deciduous forest Taiwan Bloomington and Indianapolis, IN Canadian deciduous forest Canadian spruce needles Turkey (industrial site)

vary in a systematic way, with urban areas having a greater preponderance of large particles, for example (Pirrone et al. 1995a,b; Franz and Eisenreich 1998). This also affects the characteristic deposition velocity. Many studies have attempted to measure this kind of characteristic dry particle deposition velocity for PCBs and PAHs, but the results vary over a wide range from 0.01 to nearly 10 cm/s (Table 6.2) (Lee 1991; Sheu et al. 1996; Franz and Eisenreich 1998; Odabasi et al. 1999; Vardar et al. 2002; Shannigrahi et al. 2005; Rowe 2006). The reported deposition velocities for different chemical classes vary significantly at different sampling locations and with different sampling methods. Despite many attempts to measure dry deposition directly, there is no generally accepted methodology for collecting dry particle deposition. Different kinds of surrogate surfaces including Teflon plates, Petri dishes, dry or diol-coated filters, buckets, pans filled with water, oil-coated glass plates, and greased strips have all been used to measure dry particle deposition (Eisenreich et al. 1981; Bidleman 1988; Lee 1991; Sheu et al. 1996; Franz and Eisenreich 1998; Odabasi et al. 1999; Vardar et al. 2002; Shannigrahi et al. 2005). Semivolatile organic compounds are difficult to sample because they not only partition between the gas and particle phases but can also can revolatilize from the collection surface. The types of surfaces typically used to investigate dry particle deposition are usually designed to be surrogates for water. Very little is known about dry particle deposition of SVOCs to other surfaces. Horstmann and McLachlan (1998) investigated dry deposition of a variety of SVOCs to forest

Deposition Velocity vd, cm/s

Compound

Reference

0.5 0.39 0.68 0.9 0.2

SPCBs SPCBs SPCBs SPCBs SPCBs

Holsen et al. (1991) Lee et al. (1996) Lee et al. (1996) Franz et al. (1998) Tsai et al. (2002)

5.2 2.9 0.98 0.59 1.39 0.82 1.38 1.23 6.7 2.8 0.4–3.7 0.11 0.42 0.2

SPCBs SPAHs SPAHs SPAHs SPAHs SPAHs PAHs Total PCDD/Fs Total PCDD/Fs

Tasdemir et al. (2004) Shannigrahi et al. (2005) Shannigrahi et al. (2005) Shannigrahi et al. (2005) Odabasi et al. (1999) Franz et al. (1998) Su et al. (2007) Shih et al. (2006) Koester and Hites, (1992)

0.8 0.11

PBDEs PBDEs

Su et al. (2007) St-Amand et al. (2007)

4.9 4.1

OCPs

Bozlaker et al. (2009)

canopies and observed that the dry particle deposition velocities were as much as 15 times higher in deciduous versus coniferous forests. Welsch-Pausch et al. (1995) investigated the deposition of PCDD/Fs to fields containing Welsh ray grass and concluded that dry gaseous deposition is the dominant pathway for the less volatile PCDD/Fs to this specific grassland and therefore probably governs the accumulation of these chemicals in the agricultural food chain. Rowe et al. (2007a,b) point out that the reported values of vd for PCBs are log-normally distributed, with most values falling below 0.5 cm/s. These researchers therefore calculated the geometric mean of the available literature values for PCBs, and used this value of 0.5 cm/s to estimate dry deposition loads to forested watersheds. Similarly, Rodenburg et al. (2010) calculated the geometric mean of a series of literature values for dry deposition of PAHs to water to be 0.3 cm/s. Since both the dry deposition velocity and concentrations of pollutants in the environment tend to be lognormally distributed, the uncertainties in the calculated fluxes and loads are not normally distributed. The uncertainties are therefore not symmetric about the mean, and therefore cannot be propagated by standard methods. Nor is the mean necessarily the most appropriate parameter for characterizing the load or flux. The treatment of uncertainty in fate models relying on estimates of dry particle (and as we shall see later, gaseous) deposition is therefore a challenge. Monte Carlo approaches have since been developed to assess uncertainty in these models (Venier and Hites 2008) and other environmental fate models (Schenker et al. 2009).

154


6.3.1.2. Wet Deposition. Wet deposition is calculated as Fwet ¼ vr Cr;VWM

ð6:3Þ

where Cr,VWM is the volume-weighted mean concentration of the SVOC in precipitation and vr is the rain depth. In contrast to vd and vg, vr can be measured relatively accurately, resulting in relatively low uncertainties in wet deposition fluxes and loads. Semivolatile organic compound concentrations in rain tend to vary over many orders of magnitude and are often a function of the amount of rain; small rain events tend to contain high SVOC concentrations due to the efficient scavenging of particles at the beginning of the rain event. For this reason, the volume-weighted mean concentration is used in equation 6.3 (Van Ry et al. 2002). 6.3.1.3. Gaseous Exchange. Gaseous exchange is reversible, and the overall gaseous air/surface exchange can be calculated via Fnet

Cs ¼ vg C g KSA

ð6:4Þ

where Fgas is the gaseous absorption flux, Cg is the concentration of the chemical in ambient air, Cs is the concentration of the chemical on the surface (or in the deposition matrix), and KSA is the equilibrium constant for partitioning of the chemical between air and the surface media. Under extreme disequilibrium situation, specifically, strong fugacity gradient toward the surface such as in remote “uncontaminated” areas, gaseous exchange can be calculated as Fgas ¼ Cg vg

ð6:5Þ

where Fgas is the gas absorption flux, vg is the gaseous deposition velocity, and Cg is the concentration of the SVOC in the gas phase. Measurement of Cs can be simple (i.e., when the matrix is water) or complex, as when the matrix is the wax of a leaf cuticle. For this reason, most monitoring networks measure Cg only and apply Equation (6.5) to calculate only the deposition flux. Similar to dry particle deposition velocities, gaseous exchange velocities are highly uncertain (Bohme et al. 1999) and depend on the receptor surface (water, soil, vegetation). Air–water exchange is probably the best understood of the air–surface exchange processes, but even here there are many competing models that calculate vg as a function of meteorological variables such as wind speed and temperature, as well as the diffusivity and Henry’s law constant of the SVOC (Whitman 1923; Mackay and Yeun 1983; Liss and Merlivat 1986; Wanninkhof et al. 1987; Wanninkhof 1992; Achman et al. 1993; Zhang

et al. 1999; Totten et al. 2003; Frew et al. 2004; Blomquist et al. 2006). Other models seek to describe air–water exchange in flowing waters, where turbulence induced by flow of water over the rough river bottom provides energy for mass transfer (O’Connor and Dobbins 1958; Lamont and Scott 1970; Moog and Jirka 1999a, b). Water parameters such as salinity and the presence of surface films can also have an effect on vg (Downing and Truesdale 1955; Broecker et al. 1978; Schwarzenbach et al. 2003). For air–water exchange calculations, vg is typically calculated as a function of resistances to mass transfer in both the air (va) and water (vw) 1 1 1 ¼ þ vg vw va H 0

ð6:6Þ

where H0 is the dimensionless Henry’s law constant. The coefficients va and vw have been empirically defined on the basis of experimental studies using tracer gases such as CO2 and SF6 (Whitman 1923; Wanninkhof et al. 1987; Wanninkhof 1992; Schwarzenbach et al. 2003), because the mass transfer coefficients for air–water exchange have rarely been measured directly for SVOCs. Differences in diffusivity or Schmidt number between these gases and SVOCs are then used to estimate vg and vw for SVOCs. This approach implicitly assumes that differences in solubility between the various gases are unimportant, an assumption that has been called into question. For example, Blomquist et al. (2006) note that increased solubility can reduce the impact of bubbles on the exchange velocity. Thus even the most widely studied air–surface exchange process (air–water exchange) is not well understood for SVOCs. In addition, these models suggest that the dependence of vg on wind speed is nonlinear and increases dramatically at high wind speeds. Wind speed values are typically lognormally distributed, so vg values and fluxes are also lognormally distributed (Rodenburg et al. 2010). A relatively new approach is the use of the water surface sampler (WSS), which has been used to collect both particleand gas-phase deposition of SVOCs, including PAHs (Odabasi et al. 1999; Tasdemir and Esen 2007) and PCBs (Tasdemir and Holsen 2005). The WSS is used in combination with dry deposition plates, so that the dry particle deposition component of total deposition can be subtracted out and gaseous deposition to the water surface can be calculated and then used to derive vg. The WSS is assumed to capture the deposited particles with 100% efficiency since there is no possibility for particles to bounce off the water surface. The WSS has a surface area of about 0.5 m2, so it is unclear whether results from this approach are applicable in the field. Also, the design of the WSS presumably prevents it from being able to investigate the impact of factors such as breaking waves and natural organic matter on air/water exchange.


Sampling artifacts associated with the WSS can include loss of particle-phase SVOCs from the WSS filter due to dissolution into the water, or an increase in the particle-phase SVOC flux caused by adsorption onto the filter itself or onto particles on the filter. For relatively soluble SVOCs such as lower-molecular-weight PCBs and PAHs, the first artifact is thought to dominate and the measured deposition flux will therefore be biased low. More recent measurements using the WSS suggest that the vg values derived by the two-film method described above may be too low by a factor of 2, although it is not clear whether this discrepancy is due to some artifact associated with the geometry of the WSS (Odabasi et al. 2001). Other air–surface exchange processes such as air–soil exchange and air–vegetation exchange also play important roles in the atmospheric fate and transport of SVOCs (Schroder et al. 1997; Horstmann and McLachlan 1998; Bohme et al. 1999; Wania and McLachlan 2001; Ould-Dada 2002). Airborne SVOCs can be trapped by forest canopies, which lead to reduced air concentrations and elevated contaminant levels in forest soils. This effect is also observed in global POP fate models, which predict that the world’s forests markedly decrease the long-range atmospheric transport (LRAT) of some SVOCs (Wegmann et al. 2004; Su and Wania 2005). The appropriate estimation of the deposition velocity to the forest is important for the simulation of LRAT of SVOCs. However, relatively few values for gaseous deposition velocities to surfaces other than water have been published [see Rowe et al. (2007b) for a summary]. The limited number of published values appear to be lognormally distributed. Horstmann and McLachlan (1998) have published gaseous deposition velocities for deciduous and coniferous forests. The average summer gaseous deposition velocity was nearly 5 times higher in the deciduous canopy than in the coniferous forest. Later, Su et al. (2007) further confirmed the significance of forest trapping effect for SVOCs despite differences in local climate, canopy composition, and structure. These longer-term studies did not investigate the effects of wind speed or temperature on deposition velocities. Rowe et al. (2007a) have argued that gaseous exchange between air and forested ecosystems is at or near equilibrium, suggesting that these mass transfer coefficients may be of minor importance in modeling. Thus, as with dry particle deposition, SVOC concentrations in the gas phase can be measured with relatively low uncertainty, while the large uncertainties inherent in gaseous deposition velocities introduce large uncertainties into the calculated atmospheric deposition fluxes and loads. Previous investigations of air–water exchange of PCBs (Nelson et al. 1998; Bamford et al. 2002a) have estimated that the inherent uncertainty in fluxes ranges from 40% to 900%, with the majority (88%) of this uncertainty attributed to the uncertainty in vg (Nelson et al. 1998). Many water quality models use a single value of vg for each chemical, despite its

155

dependence on wind speed. For example, the water quality model for PCBs, PAHs, and PCDD/Fs in the New York/ New Jersey Harbor uses air–water exchange mass transfer coefficients that are independent of wind speed (HydroQual 2007). In contrast, the Delaware River PCB TMDL model uses time-variable mass transfer coefficients that are a function of both wind speed and current velocity (Delaware River Basin Commission 2003). As with dry particle deposition velocities, gaseous deposition velocities are lognormally distributed. As noted above, standard methods of propagating uncertainty in gas absorption fluxes are not applicable, and treatment of uncertainty in fate models for SVOCs requires more sophisticated methods such as Monte Carlo analysis (Venier and Hites 2008; Schenker et al. 2009). 6.3.2. Chemical Reaction of SVOCs in the Atmosphere The extent to which SVOCs react in the atmosphere can dictate their long-term environment fate. If the atmospheric degradation reactions are slow compared to the rates of deposition, then a greater proportion of the emitted SVOCs could reach human and terrestrial food chains. Atmospheric SVOCs can be removed by photodegradation, including direct photolysis and indirect photolysis, that is, reaction with photochemically generated radicals. Direct photolysis in the atmosphere of a given compound is faster in the gas phase than when the compound is sorbed on particles (Koester and Hites 1992). Generally, the OH radical is the most important photochemically generated oxidant, and chemical reactions with the OH radical are the dominant loss processes for most organic compounds in the atmosphere (Atkinson 1990). Gas-phase reactions of OH radical with PCDD/Fs (Lohmann et al. 2006), PCBs (Anderson and Hites 1996; Totten et al. 2002; Mandalakis et al. 2003), and PAHs (Kwok et al. 1994) have been investigated in lab and field studies. The OH radical reaction is expected to be a significant removal pathway in the atmosphere for the lighter and intermediate PCB congeners and two- to four-ring PAHs, as well as PCDD/Fs with less than five chlorines because of their preferential partitioning into the gas phase (Anderson and Hites 1996; Brubaker and Hites 1998). The influence of OH radical reactions on the loss of SVOCs can be illustrated by the diurnal measurements of a given source strength as in studies on PCBs (Anderson and Hites 1996; Totten et al. 2002; Mandalakis et al. 2003; MacLeod et al. 2007) and PAHs (Simcik et al. 1997) or by laboratory chamber studies (Atkinson and Arey 1994; Anderson and Hites 1996). In addition, PAHs react with nitrate radical (NO3) at night (Arey et al. 1989; Atkinson and Arey 1994). This is not a major degradation pathway, but is notable because it leads to the formation of mutagenic nitro-PAHs and other nitropolycyclic aromatic compounds including nitrodibenzopyranones (Atkinson and Arey 1994).

156


Photolysis effectively occurs to PAHs only in the gas phase or when sorbed to particle surfaces. No evidence has been observed for the direct gas-phase photolysis of two- to four-ring PAHs. Photolysis of nitro-PAHs has been observed under ambient outdoor sunlight conditions and in indoor chambers with black light irradiation (Atkinson et al. 1989; Arey et al. 1990). Studies demonstrated that PAHs and PCDD/F sorbed to atmospheric particles had significantly slower rates of direct photolysis than did gas-phase compounds (Behymer and Hites 1985; Koester and Hites 1992). However, it has also been suggested that direct photolysis of PAHs on particles might play a significant role in their transformation, due to the longer residence time of particle bound PAHs, even though the kinetics of reaction are slower than in the gas phase (Vione et al. 2004) The rate of direct photolysis of gas- and particle-phase PBDEs in the atmosphere has also been studied. Raff and Hites (2007) estimated that in the lower troposphere, wet and dry deposition account for > 95% of the removal of BDE-209, while photolysis accounts for 90% of the removal of gas-phase congeners such as BDE-47. Atmospheric half-lives of low-molecular-weight PCBs (di-, tri-, and tetra-PCB) due to direct photolysis have been estimated to be on the order of weeks (Bunce et al. 1989), while the calculated atmospheric lifetimes of PCB congeners based on their gas-phase reactions with OH radicals range from 3 to 120 days (Sinkkonen and Paasivirta 2000). Photochemical reactions of SVOCs can cause their atmospheric concentrations to vary on a diel cycle. Diel variability in atmospheric SVOC concentrations has important implications for monitoring, as SVOCs are typically measured in 24-h integrated sampling to avoid any systematic bias related to the time of day when the samples are collected. Recent studies have studied the diel (24-h) variation of the concentration of SVOCs including PCBs, PBDEs, and chlordane (MacLeod et al. 2007; Moeckel et al. 2008; Gasic et al. 2009). These studies used a combination of carefully designed and timed experiments and modeling to predict the short-term behavior of SVOCs. These researchers argue that diel variations in atmospheric SVOC concentrations are a function of four factors: temperature-driven volatilization sources, atmospheric stability (boundary-layer dynamics), OH radical reactions, and source types. The type of source can affect the diel variability in SVOC concentrations due to diel variations in emissions. For example, emission rates of PAHs from vehicles will be higher during daytime when more vehicles are on the road.

6.4. MONITORING PROGRAMS Many programs measure atmospheric concentrations of POPs for the purpose of estimating atmospheric deposition fluxes and loads to aquatic ecosystems. As such, these

monitoring programs in the United States are motivated primarily by implementation of the Clean Water Act. Considerations in the design of monitoring networks are . . . . .

Where should monitoring sites be located? How often and for what duration should samples be collected? Which analytes should be measured? What types of sampling methods should be employed? What types of analytical methods should be employed?

Different monitoring networks have approached these issues in different ways. In terms of location, some networks have focused on remote areas to measure background atmospheric deposition signals, while others have targeted urban areas to capture the effects of urban or industrial activity on atmospheric SVOC concentrations. Sampling frequency and duration are often limited by cost and availability of funding. The analyte list is typically determined by the specific research goals, although the sampling and analysis methods used for many SVOCs (PCBs, OCPs, PAHs, and PBDEs) are basically the same. As a result, all these classes of SVOCs can typically be measured for low incremental cost once the samples have been processed for the target analytes. In contrast, measuring PCDD/Fs in the atmosphere often requires longer sample collection times and more rigorous cleanup methods. While sampling methodologies have been standardized largely for active air sampling for SVOCs, the analytical methodologies are constantly improving. In particular, many protocols are moving toward the use of high-resolution mass spectrometry. This technique provides greater sensitivity and selectivity, but at much higher cost. The first large SVOC monitoring network in the United States was the Integrated Atmospheric Deposition Network (IADN), which has been in operation since 1990. It includes five master stations located on remote shorelines of the Great Lakes in the United States and Canada (Hoff et al. 1996; Hillery et al. 1998). Urban sites within Chicago were later added to the network (Basu et al. 2004) (http://www.msc.ec. gc.ca/iadn/stations/chicago_e.html). After successfully fulfilling its first and second implementation plans, the third implementation plan is currently ongoing, covering the years 2005–2010. In addition to the monitoring of “core chemicals” (PCBs, OCPs, and PAHs), chemicals of emerging concern such as PBDEs as well as PCDD/Fs have been added to the list of analytes (Environment Canada and the United States Environmental Protection Agency 2005). The IADN operates on a 12-day sampling schedule, collecting one 24-h integrated sample for gaseous and particle-phase SVOCs every twelfth day. Precipitation is usually collected via integrated 24-day samples. This sampling schedule appears to be adequate for SVOCs, and generates about 100 samples per year per site for SVOC analysis [including blanks and

MONITORING PROGRAMS

QA (quality assurance) samples]. The objectives of the IADN study are to determine the loading of persistent toxic contaminants from the atmosphere to the Great Lakes basin from both urban and regional sources. The IADN has been continuously funded via an agreement between the United States and Canada. Results from the IADN network show that the concentrations of most SVOCs are much higher in the urban area of Chicago than at the remote sites (Basu et al. 2004; Sun et al. 2006). Concentrations of PCBs and OCPs display longterm declines at most IADN sites (Cortes et al. 1998; Buehler et al. 2002; Buehler and Hites 2002; Sun et al. 2006, 2007). There is no declining trend observed for PAHs because of the presence of ongoing sources (Buehler and Hites 2002). Concentrations of most SVOCs at most IADN sites are temperature-dependent, with higher concentrations occurring during warmer periods. This temperature dependence is typically modeled using the Clausius–Clapeyron equation (Simcik et al. 1999b; Carlson and Hites 2005). The Chesapeake Bay Atmospheric Deposition Study (CBADS) (Leister and Baker 1994) was conducted at three nonurban sites along the shoreline of the bay during 1990–1993. Ambient air samples were collected biweekly, and air sampling times ranged from 12 to 24 h. The primary objective of the CBADS study was to provide the best possible estimates of total annual atmospheric loading of a variety of trace elements and organic contaminants directly to the surface waters of the Chesapeake Bay. The CBADS study was not designed to capture long-term time trends in atmosphere contaminant concentrations, and thus the network shut down after sufficient short-term data had been collected. The measured atmospheric concentrations of PAHs and PCBs at the northern and southern regions of Chesapeake Bay are similar and exhibit seasonal variability (Dickhut and Gustafson 1995). Though wet deposition concentrations displayed little seasonal variation, spatial variation of the wet deposition fluxes was observed. Atmospheric deposition was found to be an important source of SVOCs to Chesapeake Bay (Dickhut and Gustafson 1995). On the basis of the IADN experience, the New Jersey Atmospheric Deposition Network (NJADN) was designed to capture both the urban and regional signals of SVOCs by locating monitoring sites in a variety of land use types (Gigliotti et al. 2000; Van Ry et al. 2002; Totten et al. 2004). The NJADN project was designed to monitor the loadings of organic pollutants to adjacent water bodies and to assess spatial and temporal trends in SVOC concentrations. The history of the NJADN illustrates one of the major problems confronting SVOC monitoring networks in the United States—these networks are typically designed to address a local or regional water quality issue, and once data to support that issue are generated, the network generally shuts down. The NJADN was originally (1997–1999) funded by the Hudson River Foundation and consisted of three sites focused on the New York/New Jersey Harbor

157

(Eisenreich 1998). Later (1999–2001) it was funded by the New Jersey Department of Environmental Protection and included nine sites throughout New Jersey and focused on water and air quality statewide, measuring SVOCs as well as trace metals (Reinfelder et al. 2004; Totten et al. 2004). Most recently (2002–present) it has been funded by the Delaware River Basin Commission and includes three sites focused on deposition of PCBs to the Delaware River. The NJADN sites have included varying land-use regimes, including urban/ industrial, suburban, coastal, and rural areas. The sampling schedule and methodology are similar to those of the IADN; gas and particle phase samples are collected for 24 h every twelfth day, and integrated precipitation samples are collected over 24-day periods. Analytes have at various times included PAHs, PCBs, OCPs, trace metals, and mercury. Data from the NJADN have been used to estimate atmospheric deposition loads of PCBs and PAHs to the New York/ New Jersey Harbor (Totten et al. 2004; Gigliotti et al. 2005) and loads of PCBs to the Delaware River (Totten et al. 2006a) in support of the TMDL for PCBs. Like the IADN network, data from the NJADN demonstrate that concentrations of most SVOCs are higher in urban areas, display significant temperature dependence via a Clausius–Clapeyron-type relationship, and are undergoing long-term declines (Totten et al. 2004; Gioia et al. 2005). In the United States, dioxin-like compounds are monitored via the National Dioxin Air Monitoring Network (NDAMN) (Cleverly et al. 2000), which is run by the U.S. Environmental Protection Agency (USEPA). The NDAMN now consists of 35 sites situated in rural and remote locations, including many national parks, across the United States. An ambient air sample is collected for a 4-week period, every 3 months, concurrently at each site (Cleverly et al. 2000). Other monitoring programs for SVOCs exist in the United States, but many are run by state governments and do not publish results in peer-reviewed sources. This makes obtaining data and comparing data between networks difficult. Table 6.3 presents a partial list of some other atmospheric SVOC monitoring programs. As with the networks described above, the expense of measuring SVOCs often means that monitoring networks are operated for only short periods. In Europe, the European Monitoring and Evaluation Program (EMEP) was established as a co-operative program for monitoring and evaluation of the long-range transmissions of air pollutants in Europe (http://tarantula.nilu. no/projects/ccc/index.html). Persistent organic pollutants (POPs) were added to the EMEP’s monitoring program in 1999 (Gusev et al. 2008). The measured POPs include PAHs, PCBs, OCPs, HCB, and HCHs. Currently there are 15 sites monitoring POPs in air and precipitation, with most sites located around the North and Baltic Seas in the Arctic and in the Czech Republic. The monitoring in individual countries is financially supported by national environmental agencies. The EMEP service provides a guidance manual on sampling

158


TABLE 6.3. Examples of SVOC Monitoring Programs Monitoring program

Description

California Ambient Dioxin Air Monitoring Program (CADAMP) (http://www.arb.ca.gov/aaqm/qmosopas/ dioxins/dioxins.htm) National Dioxin Air Monitoring Network (NDAMN) (http:// cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid¼54936)

Monitoring for dioxins and PBDEs

2001–2004

California EPA

Monitoring dioxins at rural and nonimpacted locations throughout the United States Heavy metals, POPs, and acidifying compounds from Arctic sites Atmospheric deposition of heavy metals, POPs, and sulfur and nitrogen compounds to the North Sea Investigating airborne contaminants in national park ecosystems using a network of eight primary parks in the western United States Identifying long-range transport and the load of POPs and other organic compounds in remote alpine regions

1998–2002

USEPA

1991–present

Arctic council

1987–present

OSPAR Commission

2002–2008

National Park Service

2005

European Union

Arctic Monitoring and Assessment Programme (AMAP) (http://www.amap.no) Comprehensive Atmospheric Monitoring Programme (CAMP) (http://www.ospar.org)

Western Airborne Contaminants Assessment Project (WACAP) (http://www.nature.nps.gov/air/Studies/air_toxics/wacap. cfm)

Monitoring Network in the Alpine Region for Persistent and other Organic Pollutants (MONARPOP) (http://www.monarpop.at)

and analysis methods. However, individual countries are free to use their own methods if they can be documented to be equally reliable.

6.5. SAMPLING AND ANALYSIS Chapter 12 provides more detail on methods of measuring SVOCs in air. This section therefore summarizes the main methods used in monitoring studies and highlights issues of quality assurance and quality control that can affect the usefulness of the data generated. 6.5.1. Sampling 6.5.1.1. Active Air Sampling. Currently, the high-volume air sampler is the most widely used approach for sampling SVOCs in air. This active sampling approach utilizes a pump that draws air through a filter to retain the particle-phase SVOCs and then through an ad/absorbent media (typically a foam or resin) to retain vapor-phase SVOCs. High-volume air samplers equipped with polyurethane foam (PUF) adsorbents have been used extensively and are recommended by USEPA to study the occurrence and speciation of SVOCs such as pesticides, PCBs, and PAHs in the atmosphere (USEPA 1999a). The EMEP, IADN, CBADS, and NJADN systems all used high-volume air samplers; EMEP, CBADS, and NJADN use PUF as the adsorbent. While IADN switched

Duration

Agency

from PUFs to XAD-2 in 1992 at U.S. sites, Canadian sites still use PUF only. In general, XAD-2 resin has a higher collection efficiency for volatile SVOCs than does PUF, as well as a higher retention efficiency; however, XAD-2 may be associated high blank levels of some compounds, especially PAHs (Franz and Eisenreich 1993). Polyurethane foam cartridges are easier to handle in the field and maintain better flow characteristics during sampling (USEPA 1999a). Some studies use a PUF/XAD/PUF sandwich system that is designed to combine the best features of PUF and XAD sorbents, and can be used in a PUF hi-vol without modifying the sampler (Yao et al. 2006; Primbs et al. 2007). Active sampling with a hi-vol sampler requires expensive equipment, a secure location for sampling, a power source, and a significant manpower investment. Active air sampling typically involves collection of a single sample for 24 h, or at most a few days. Longer sampling times introduce problems of breakthrough of analytes on the polyurethane foam sampling cartridge. This is especially a problem for PCDD/Fs, since their low concentrations in the atmosphere typically necessitate sampling intervals of 48 h or more (Cleverly et al. 2000). Acquiring long-term average SVOC concentrations requires collecting, analyzing, and averaging the data from many samples, thus adding to the cost. Therefore, active sampling is often an impractical and expensive approach for investigating the spatial variability of SVOCs. There are some sampling artifacts associated with the design of the high-volume sampler that can affect gas- versus

SAMPLING AND ANALYSIS

particle-phase measurements (Cotham and Bidleman 1991; Gundel et al. 1995; Peters et al. 2000) One of the notable sampling artifacts is known as “blowoff,” whereby SVOCs that have been collected on particulate matter from ambient air can volatilize off of the particles. This blowoff will lead to enhanced apparent vapor-phase concentrations. Another sampling artifact associated with the high-volume sampler is “blowon,” whereby vapor-phase SVOCs can sorb onto the filter media or the accumulated particulate matter and/or organic matter on the surface of the filter media. This can cause an apparent increase in the particle-phase concentration of SVOCs. A backup filter is sometimes installed in series with the first collection filter to determine the magnitude of the artifact due to gas-phase sorption to the filter collection media (Primbs et al. 2007). This approach is presented in more detail in Chapter 5. Additionally, highvolume samplers can be subject to contamination artifacts. Basu et al. (2000) reported that the foam gasket in the highvolume air samplers can become contaminated with PCBs from the air and then release PCBs back into the sampled airstream. Another active air sampler, the diffusion denuder sampler, has been developed to provide an alternative sampling method to minimize potential sampling artifacts (Gundel et al. 1995). In denuder samplers, vapor-phase SVOCs are removed from the airstream by diffusion onto an adsorbent coating prior to removal of particulate matter by filtration. An adsorbent downstream of the filter then collects any SVOCs that are volatilized from the collected particulate matter. Diffusion denuder samplers (e.g., the high-capacity integrated organic gas and particle sampler (IOGAPS)) can dramatically improve the recovery of lowermolecular-weight gas and particle PAHs such as naphthalene and anthracene (Poor et al. 2004). Potential sampling artifacts with diffusion denuders are loss of fine particulate matter to the surface of the denuder tube, which would result in an apparent higher vapor-phase measurement; desorption of SVOC from particles while in transit through the denuder tube; and breakthrough of volatile analytes to the downstream adsorbent, resulting in an apparent higher particlephase measurement (Peters et al. 2000). Peters et al. (2000) compared the performance of high-volume and diffusion denuder samplers for the measurement of SVOCs. Their results suggested that the difference in the measured gas/ particle partitioning of SVOCs in these two different types of samplers can be viewed as a function of vapor pressures and sampler geometry. The high-volume samplers are suitable for measuring SVOCs with subcooled liquid vapor pressures ( pL ) < 0.2 Pa. Although the diffusion denuder sampler does not appear to suffer from this limitation and could efficiently sample the more volatile SVOCs, it may underestimate the partitioning of the less volatile SVOCs to the particle phase. These issues are presented in more detail in Chapter 5.

159

6.5.1.2. Passive Air Sampling. Passive air sampling (PAS) techniques have been developed more recently to address the need for inexpensive and simple monitoring of SVOCs in the atmosphere. Passive air samplers are chemical accumulators that rely on air currents to deliver chemical to the sampler media, which may consist of semipermeable membrane devices (SPMDs), PUF disks, polymer-coated fibers, polymer-coated glass, or XAD-2 resin. Because passive air samplers do not need a power source, they are more suitable for remote locations away from the electric grid. In general, PAS is a more cost-effective approach, which can provide integrated atmospheric SVOC concentrations over a period of months and assess concentrations in air simultaneously at multiple sites at far lower cost. Polyurethane Foam passive samplers are the most widely used media for the investigation of SVOCs in the atmosphere in PAS studies. The PUF samplers were designed to sorb only gas-phase organics, but particles will also deposit to some extent in the passive sampler. The geometry of the sampler housing generally limits the amount of particles reaching the sampling matrix. The sampler housing can not only help prevent the particle deposition but also reduce the dependence of sampling rate on meterological conditions such as precipitation and wind speed (Pozo et al. 2004). The use of passive sampling methods to monitor atmospheric concentrations has greatly increased. Passive samplers have been used to investigate the vertical (Moreau-Guigon et al. 2007; Li et al. 2009), temporal (Meijer et al. 2003; Motelay-Massei et al. 2005; Moreau-Guigon et al. 2007), and spatial (Meijer et al. 2003; Harner et al. 2006a; Du et al. 2009) distribution of atmospheric POP concentrations. The utility of PASs has been demonstrated not only at the local scale but also at global scale (Jaward et al. 2004, 2005). The theory of passive air sampling is described in detail elsewhere (Muller et al. 2000; Shoeib and Harner 2002; Bartkow et al. 2005), and is briefly summarized here. The passive sampling medium (PSM) is a hydrophobic organic matrix that has a high capacity for organic chemicals. Hydrophobic organic contaminants such as PCBs in the gas phase will preferentially partition into this matrix. The extent to which the organic chemicals are enriched relative to air can be described by a passive sampler medium–air partition coefficient (KSV). It has been shown that KSV is related to the octanol–air partition coefficient (Koa) of the chemical, which is typically very high for SVOCs, leading to a strong driving force for uptake of SVOCs onto the PSM. For example, Koa values for PCBs are on the order of 106–1012 (Harner and Bidleman 1996), 108–1012 for PCDD/Fs (Harner et al. 2000), 109–1012 for BDEs (Harner and Shoeib 2002), and 105–1012 for PAHs (Ma et al. 2010). The exchange of gaseous SVOCs between the ambient air and the passive sampler matrix occurs via diffusion. Molecular diffusion is the exclusive transport mechanism

160


for mass transfer at the air–sampler interface and in the sampler. The exchange of gaseous SVOCs between the atmosphere and the passive sampler matrix occurring via diffusion can be described by the equation VS

dCS CS ¼ kO AS ðCV Þ dt KSV

ð6:7Þ

where VS is the volume of the passive sampling medium, CS and CV are the chemical concentrations in the passive sampler medium and in the bulk air, respectively; AS is the surface area of the passive sampler; KSV is the sampler/air partition coefficient; kO is the overall mass transfer coefficient, and t is time. The mass transfer across the sampler–air interface is simply the net result of resistance in the air boundary layer and in the sampler matrix. According to the Whitman two-film approach, the overall mass transfer coefficient (kO ) can be deduced from the two mass transfer coefficients acting in series, namely, the air-side (kV ) and the passive-sampler-side (kS ) mass transfer coefficients: 1 1 1 ¼ þ kO kV kS KSV

ð6:8Þ

Therefore, substituting Equation (6.8) into Equation (6.7) yields VS

dCS 1 CS ¼ AS ðCV Þ 1=kV þ 1=kS KSV dt KSV

becomes more significant and the uptake has moved into the curvilinear stage. Finally, the uptake will reach quasi-equilibrium as the fugacity of the chemicals in the sampling matrix and the surrounding air stabilize around the KSV value (i.e., CS/KSV ¼ CV), there is no net uptake and CS becomes constant. The time taken to attain equilibrium is influenced by sampler and analyte characteristics. The differential accumulation equation [Eq. (6.8)] can be solved analytically with certain boundary conditions to provide an exact description of the uptake profile. For a deployed sampler in the field, we can assume that the vaporphase concentration (CV ) is constant, the initial sampler concentration (CS0 ) is zero, and all other parameters with the exception of CS are constant with time. The solution of Equation (6.9) is then CS ¼ KSV CV

AS 1exp t KSV VS ð1=kV þ 1=kS KSV Þ ð6:10Þ

This equation is a complete uptake profile of chemicals onto the passive air sampler. The KSV coefficient can be interpreted as the equivalent volume of air that contains the same mass of analyte as one unit volume of passive sampling medium under equilibrium conditions (i.e., KSV ¼ Vair =VS ¼ CS =CV ). Therefore, by analogy and replacing terms in Eq. (6.10), the equivalent air sample volume is

ð6:9Þ

The exchange of chemicals between the passive sampler and the ambient air can be presented in three stages as depicted by Equation (6.9) and illustrated in Figure 6.1. In the beginning of the sampling period when the concentration of chemicals in the PAS is minimal, the elimination rate term (CS /KSV) is negligible and the uptake of organic chemicals is linear. When the chemical builds up in the PAS, CS /KSV

Vair ¼ KSV VS

As k v 1exp t Vs Ksv

ð6:11Þ

When the diffusion transport is limited by the resistance on the sampler side, kV kS KSV , Equation (6.10) can be further simplified as follows: AS kS CS ¼ KSV CV 1exp t VS

ð6:12Þ

Amount sequestered

On the other hand, if the transport is limited by the resistance on the air side, that is, kV kS KSV , Equation (6.10) can be expressed as Curvilinear

AS kV CS ¼ KSV CV 1exp t KSV VS Equilibrium partitioning

Linear

time

Figure 6.1. Three-stage uptake curve for passive sampler.

ð6:13Þ

Air-side limitation to chemical exchange has been observed for a range of PCBs (Shoeib and Harner 2002) and for certain PAHs (Muller et al. 2000; Bartkow et al. 2005). As indicated in the Equation (6.11), the equivalent sample volume is a function of KSV . For heavier and more hydrophobic chemicals (i.e., high KSV values), higher equivalent sample volume is obtained because of longer time required to achieve equilibrium. The time required for a chemical that


has higher KSV to reach equilibrium can be quite long. For instance, for chemicals with KOA greater than 109, the deployment time to reach equilibrium will be greater than 450 days (Shoeib and Harner 2002). Thus it is not necessary, and not always desirable or possible, to deploy passive samplers long enough to achieve equilibrium. The main challenge associated with the use of PAS is the conversion of the mass of analyte on the sample to a concentration in the gas phase, which requires knowledge of the volume of air actually sampled. This volume can be calculated from Equation (6.11) provided that all the necessary parameters are available. The air volume can also be estimated by combining passive and active measurements and thereby calibrating the passive sample mass to the measured concentration from active monitoring. This partially negates the benefits of passive sampling, and it also assumes that all PASs have the same sampled volume, despite different locations and micrometeorology. Another approach to calculating the air concentrations when the passive sampler is not at equilibrium is to assume an arbitrary sampling rate. For instance, a sampling rate of 3–5 m3 of air per day has been assumed in some studies (Pozo et al. 2004; Harner et al. 2006b). Adding depuration compounds [i.e., performance reference compounds (PRCs)] is a more direct way to measure the sampling rate (Pozo et al. 2004). These compounds are typically either isotopically labeled chemicals or unlabeled chemicals that are not present in the atmosphere. These compounds are added to the PAS prior to its deployment to the field, and loss or depuration of the PRCs is then related to the uptake rates of target analytes. Depuration of the initially spiked PRC can be described as CPRC ¼ CPRC;0 expðke tÞ

ð6:14Þ

lnðCPRC =CPRC;0 Þ ¼ ke t

ð6:15Þ

or

where CPRC is the concentration of PRC remaining in the passive sampler, CPRC;0 is the initial concentration of PRC added to the passive sampler, and ke is the depuration rate constant. Because uptake of SVOCs is usually airside-controlled, the rate of chemical uptake will be the same as the rate of loss: kU ¼ ke . According to Equation (6.13), the uptake rate kU for chemicals under air-side control can be calculated as follows: kU ¼

A S kV KSV VS

ð6:16Þ

By combining Equation (6.14) and (6.15), we can calculate the air-side mass transfer coefficient (kV ) from the recovery of PRCs initially spiked into the PSM (i.e.,CPRC =CPRC;0 ):

kV ¼ ln

CPRC VS KSV * * CPRC;0 AS t

161

ð6:17Þ

The term kV AS represents the passive air sampling rate. This is a very useful term since it can be used to calculate how much air is sampled by the passive air sampler during the linear uptake stage. Thus, although there are several methods of converting the mass of analyte in the passive samples to gas-phase concentrations, there is still uncertainty involved in this calculation that can limit the utility of the PAS approach. For this reason, when passive air samplers are deployed simultaneously, they are an excellent method of obtaining relative concentrations, but may not be the best choice when absolute concentrations are needed. They are particularly useful for determining “fingerprints” of SVOCs, since the sampling volume is unimportant in this case. For this reason, PAS can be particularly useful for tracking down specific SVOC sources that have unique fingerprints, because such trackdown studies necessarily involve many samplers concentrated in a specific geographic area. An additional limitation of PAS is the relatively low sampling rates (typically a few cubic meters of air per day) of most PAS, requiring long sampling periods, which restricts the temporal resolution that can be achieved. Some researchers (Harner et al. 2003; Farrar et al. 2005a, b) have attempted to overcome this problem by developing rapidly equilibrating sampling media such as polymer-coated plates that can be deployed for as little at 7 days. Passive air samplers have has demonstrated that spatial variability in SVOC concentrations is significant. Passive sampling campaigns in Birmingham (UK), Toronto, and Philadelphia (Harner et al. 2004; Harrad and Hunter 2006; Du et al. 2009) have demonstrated that SVOC sources in urban areas are more patchy and localized than might be expected from IADN and NJADN data, suggesting that the mix of SVOC sources in urban areas is complex. Xiao et al. (2007, 2008) developed a flowthrough sampler (FTS) to collect gaseous and particle-bound SVOCs from large volumes of air by turning it toward the wind and having the wind blow through a porous sampling medium. This sampler provides substantially faster uptake rates than does the typical PAS, while requiring no access to reliable network power. The shorter sampling time resulted from a higher sampling rate compared to PAS, allowing higher temporal resolution. Quantitative relationships between the wind speed outside the sampler and after passage through the PUF were established with a battery-operated data logger and allow the accurate estimation of sampling volumes under conditions of low and high wind speed. However, this wind speed dependence of sampling rate also implies a sampling bias toward times with higher wind speed. This bias can be alleviated by including an annular bypass in the sampler design that helps stabilize the wind speed through the sampler

162


(Xiao et al. 2007). Additionally, it was confirmed that even relatively volatile SVOCs, such as naphthalene and anthracene, do not experience serious breakthrough (Xiao et al. 2008). Considering the absence of reliable power supply and lower atmospheric SVOC levels at remote sites, the application of FTS could allow measurement of SVOC concentrations at higher temporal resolution (seasonal or monthly basis). 6.5.1.3. Sampling of Precipitation. Sampling of precipitation can be conducted by the use of wet-only precipitation collectors (Poster and Baker 1996; Molina et al. 2000; Van Ry et al. 2002). The sampler is fitted with a moisture sensor that, when activated, causes the cover on the rain sampler to pivot open to reveal a square stainless-steel collection funnel. The rainwater then feeds through the funnel and then through a column containing XAD-2 resin, which extracts the organics. This method does not separate particle-bound versus dissolved SVOCs. The evaporating rainwater can leave SVOCs on the funnel that can be sampled via wiping it with an agent such as glass wool. This wipe can contain as much as half of all the SVOCs in the rain sample (Van Ry et al. 2002). 6.5.2. Analysis In general, analysis of SVOCs after sampling is commonly based on a laborious analytical sample treatment involving extraction and cleanup, followed by instrument analysis. 6.5.2.1. Extraction. Air samples are traditionally extracted by Soxhlet extraction, a robust and well-established solid– liquid extraction technique. The shortcomings associated with the use of Soxhlet extraction are the consumption of large amount of solvents (250 mL or more) and longer extraction time (16–24 h). A relatively new rapid and effective extraction technique is the accelerated solvent extraction (ASE) system, which extracts samples under elevated temperature and pressure conditions with less solvent (50 mL) and time (20 min) compared with Soxhlet extraction. The extracts remain liquid under high-pressure conditions. Accelerated solvent extraction systems have been successfully used for extraction of SVOCs from a variety of environmental matrices, including soils, sediment, and particulate matter (Bautz et al. 1998; Fitzpatrick et al. 2000; Hussen et al. 2006; Ryno et al. 2006). Because of the concerns that ASE will coextract or dissolve the polymers used for collecting gas-phase SVOCs, the use of ASE for air samples has been restricted until quite recently (Primbs et al. 2008; Genualdi et al. 2009). 6.5.2.2. Cleanup. Analysis of SVOCs almost always involves a cleanup step prior to instrumental analysis for the purpose of removing interfering chemicals and for fractionation. Atmospheric samples, especially gas-phase samples,

contain relatively low levels of interfering compounds, especially as compared to more challenging environmental matrices such as biota or sewage sludge, so cleanup of atmospheric samples is usually easier and less aggressive. The cleanup step is typically based on gravity flow solid– liquid chromatography with different stationary phases. The most commonly used stationary phases include silica, alumina, Florisil, and different types of carbon (Carbosphere, Carbopack, Amoco PX-21, Celite 545). Generally, acid, neutral, and basic silica gels are used to remove polar interferences, and Florisil or alumina columns are used to remove nonpolar interferences. Carbon columns are sometimes used to remove nonpolar interferences or to fractionate analytes on the basis of polarity. For instance, Carbopak/ Celite can be used to separate some of the coplanar PCBs from the mono- and di-ortho-substituted PCBs (USEPA 1999b). Under certain circumstances, pressurized flow [i.e., high-pressure liquid chromatography (HPLC)] can be used to provide specificity for certain congeners and congener groups. For example, USEPA Method 1613 (USEPA 1999b) uses HPLC equipped with two Zorbax-ODs columns in series (DuPont Instruments Division, Wilmington, DE, or equivalent) to separate 2,3,7,8-substituted CDD and CDF isomers. It should be noted that some stationary phases are incompatible with some analytes. For example, Florisil adsorbs PAHs and removes them from the sample, so it is not a good choice when PAHs are target analytes. Sample cleanups for the analysis of SVOCs from air samples typically utilize one of the following: multilayer silica (activated silica, sulfuric acid on activated silica, sodium hydroxide on activated silica), basic alumina, or Florisil. In general, more efficient and rigorous cleanup methods are needed for the analysis of PCDD/Fs compared to other SVOCs because of their ultralow concentrations in air samples and the coexistence of much higher levels of interferences such as PCBs (Lohmann and Jones 2005). The National Dioxin Air Monitoring Network uses a modified version of the USEPA Method 1613 cleanup, consisting of a two-step cleanup with Biosil A silica gel and Amoco PX-21 carbon columns as the stationary phases (Cleverly et al. 2007). The first step removes polar compounds, while the second step separates out the 2,3,7,8-substituted isomers from all the other PCDD/Fs. In contrast, IADN and NJADN use a one-step cleanup method with alumina as the stationary phase (Gigliotti et al. 2000; Buehler et al. 2002), since they do not measure PCDD/Fs. Additionally, chemical treatments are sometimes applied as part of the cleanup. For example, concentrated acid is used to destroy all nonchlorinated compounds for the measurement of PCBs at the EMEP sites in Norway (Eckhardt et al. 2009). Most cleanup methods separate the sample into multiple fractions containing different target analytes. For example, the cleanup method used by the NJADN network uses alumina and elutes two fractions (Gigliotti et al. 2000). The


first fraction is eluted in hexane and contains the PCBs. The second fraction is eluted in a combination of hexane and dichloromethane and contains the PAHs. Organochlorine pesticides (OCPs) and PBDEs are present in both fractions. This cleanup method therefore allows the measurement of all four of these compound classes from a single procedure. Surrogate compounds, which are typically added to the sample during extraction, are used to track the performance of the extraction and cleanup steps. There is some confusion in terminology here. In Europe, these may be called recovery standards, while the term surrogate standard refers to the standard used for the determination of chemical concentration. In the United States, the standards used in quantitation are typically called internal standards, and surrogate standards are the compounds used to track the losses of analyte that may occur during extraction and cleanup. Here we will use the U.S. terminology. The surrogate compounds may consist of isotopically labeled compounds or compounds that are not present in the environmental samples. For example, some PCB congeners that are not present in most commercial PCB formulations are often used as surrogate standards for the analysis of PCBs by electron-capture detection (ECD). In PCB analysis using USEPA Method 1668B (USEPA 1999b), 39 13 C-labeled PCB congeners are used as performance checks: 3 field standards that are added during sampling, 28 surrogates that are added before extraction, 3 cleanup standards that are added just prior to cleanup, and 5 internal standards added just prior to instrumental analysis. This number of performance standards allows a rigorous tracking of analytical performance, but comes at a price, since 13 C-labeled standards are typically quite expensive. 6.5.2.3. Instrumental Analysis. For PCBs, high-resolution capillary GC coupled with electron-capture detection (ECD) has been a standard method for their measurement in environmental matrixes for decades. A Method 8082 (USEPA 2000) is the standard method for measurement of PCBs by ECD. The ECD method achieves relatively high sensitivity while using a relatively low-cost, lowmaintenance instrument. These advantages led the ECD technique to become the most widely used detection method for the analysis of low levels of halogenated contaminants including PCBs and OCPs. Gas chromatographic ECD is recommended for the analysis of PCBs and OCPs excluding non-ortho-PCBs and toxaphene at part per billion (ppb) levels (Muir and Sverko 2006). Even though GC-ECD is one of the most widely used methods, USEPA Method 8082 is not followed exactly by any of the monitoring networks noted above (IADN, NJADN, CBADS) except for some EMEP sites. The disadvantage of the ECD method is that identification of analytes relies solely on their retention times. Thus coelution of interfering compound with the target analytes can prevent accurate quantification. In the 1990s, the USEPA developed Method 1668 (USEPA 1999b)

163

for measurement of PCBs by high-resolution GC (HRGC) with high-resolution mass spectrometry (HRMS). A minor revision of 1668A, Method 1668B, was published in November 2008 (USEPA 1999b). The HRGC/HRMS method is a more powerful alternative for the determination of SVOCs, including PCBs, organochlorine pesticides, and PCDD/Fs, especially at part per thousand (ppt) levels. High-resolution MS methods have become the preferred methods in many cases because of their very high sensitivity and powerful identification capability. HRGC and HRMS methods generally avoid the problems of ambiguous identification of PCB congeners inherent in the ECD method, although within homologs, many PCB congeners still coelute and are quantified as a sum in Method 1668B. A major obstacle to the adoption of Method 1668B is its cost. The high-resolution GC/MS instrument is at least 10 times more expensive than a standard ECD instrument and is therefore usually beyond the means of most academic labs. In addition, as noted above, 1668B uses a large number of 13 C-labeled PCB congeners as surrogates and internal standards, adding to the cost. As a result, the typical cost for analysis of a single sample by 1668B is approximately $800. For air-monitoring networks that generate as much as 100 samples per site per year, this cost is prohibitive. Thus there is a need for a more cost-effective approach to PCB measurement that still avoids the problem of coelution and ambiguous identification of PCB congeners. The standard analytical method approved by the USEPA for the analysis of PAHs in ambient air (TO-13A) uses capillary gas chromatography with low-resolution mass spectrometry (USEPA 1999b). This approach offers good sensitivity and selectivity. High-resolution MS is not necessary because the concentrations of PAHs in atmospheric samples are typically orders of magnitude higher than those of PCBs. Method TO-13A (USEPA 1999b) uses the isotope dilution approach, which utilizes isotopically labeled internal standards and surrogate compounds. Some of the sites of the EMEP use HPLC with fluorescence detection to measure PAHs (Mano and Schaug 2003). Analysis of PCDD/Fs in environmental matrixes represents one of the biggest analytical challenges owing to a series of problems, including their ultralow concentrations in the environment, the coexistence of a large number of interfering compounds in much higher concentrations, and the need for congener-specific determination to differentiate the toxic congeners from other structurally similar congeners (Eijarrat and Barcelo 2002). Method 1613 is an approved standard USEPA method for determination of the 2,3,7,8substituted PCDD/Fs, which are the most toxicologically important, using isotope dilution by HRGC/HRMS (USEPA 1999b). The air concentrations of 2,3,7,8-TCDD are typically less than 1 fg/m3. Currently, the most sensitive HRGC/HRMS can have sensitivities of

Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems (Wiley Series Sponsored by IUPAC in Biophysico-Chemical Processes in Environmental Systems)

Biophysical Chemistry of Fractal Structures and Processes in Environmental Systems

Biophysical Chemistry of Fractal Structures and Processes in Environmental Systems

Transport processes in pharmaceutical systems

Transport Processes in Pharmaceutical Systems

transport processes in pharmaceutical systems

Adaptive processes in economic systems

Unit Operations and Processes in Environmental Engineering

Physical Processes in Earth and Environmental Sciences

Environmental Transport Processes

Biophysical Chemistry of Fractal Structures and Processes in Environmental Systems (Iupac Series On Analytical and Physical Chemistry Of Environmental Systems Volume 11)

Partition and Adsorption of Organic Contaminants in Environmental Systems

Biophysico-Chemical Processes Involving Natural Nonliving Organic Matter in Environmental Systems

Unit Processes In Organic Synthesis

Modelling Dynamics in Processes and Systems

Flocculation in Natural and Engineered Environmental Systems

Atmospheric Processes and Systems

Dynamical systems and processes

Towards Environmental Innovation Systems

Towards Environmental Innovation Systems

Controversies in Environmental Policy (SUNY Series in Environmental Public Policy)

Statistical Analysis of Environmental Space-Time Processes

Statistical Analysis of Environmental Space-Time Processes

Modeling Demographic Processes in Marked Populations (Environmental and Ecological Statistics)

Metal-catalysis in Industrial Organic Processes

Recent Advances in Adsorption Processes for Environmental Protection and Security

Environmental Geomorphology (Developments in Earth Surface Processes 4)

Discrete Time Analysis of Batch Processes in Material Flow Systems

Knowledge management: systems and processes

Knowledge management: systems and processes

Reaction mechanisms in environmental organic chemistry

Biophysico-Chemical Processes of Anthropogenic Organic Compounds in Environmental Systems (Wiley Series Sponsored by IUPAC in Biophysico-Chemical Processes in Environmental Systems)

Biophysical Chemistry of Fractal Structures and Processes in Environmental Systems

Biophysical Chemistry of Fractal Structures and Processes in Environmental Systems

Transport processes in pharmaceutical systems

Transport Processes in Pharmaceutical Systems

transport processes in pharmaceutical systems

Adaptive processes in economic systems

Unit Operations and Processes in Environmental Engineering

Physical Processes in Earth and Environmental Sciences

Environmental Transport Processes

Biophysical Chemistry of Fractal Structures and Processes in Environmental Systems (Iupac Series On Analytical and Physical Chemistry Of Environmental Systems Volume 11)

Partition and Adsorption of Organic Contaminants in Environmental Systems

Biophysico-Chemical Processes Involving Natural Nonliving Organic Matter in Environmental Systems

Unit Processes In Organic Synthesis

Modelling Dynamics in Processes and Systems

Flocculation in Natural and Engineered Environmental Systems

Atmospheric Processes and Systems

Dynamical systems and processes

Towards Environmental Innovation Systems

Towards Environmental Innovation Systems

Controversies in Environmental Policy (SUNY Series in Environmental Public Policy)

Statistical Analysis of Environmental Space-Time Processes

Statistical Analysis of Environmental Space-Time Processes

Modeling Demographic Processes in Marked Populations (Environmental and Ecological Statistics)

Metal-catalysis in Industrial Organic Processes

Recent Advances in Adsorption Processes for Environmental Protection and Security

Environmental Geomorphology (Developments in Earth Surface Processes 4)

Discrete Time Analysis of Batch Processes in Material Flow Systems

Knowledge management: systems and processes

Knowledge management: systems and processes

Reaction mechanisms in environmental organic chemistry

Recommend Documents