Smart Materials

Smart Materials 43722_C000.indd i 10/18/2008 7:26:39 AM 43722_C000.indd ii 10/18/2008 7:26:40 AM Smart Materials...

Author: Mel Schwartz

221 downloads 1858 Views 24MB 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

Smart Materials

43722_C000.indd i

10/18/2008 7:26:39 AM

43722_C000.indd ii

10/18/2008 7:26:40 AM

Smart Materials Edited by

Mel Schwartz

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

43722_C000.indd iii

10/18/2008 7:26:40 AM

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2009 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-13: 978-1-4200-4372-3 (Hardcover) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Smart materials / [edited by] Mel Schwartz. p. cm. Includes bibliographical references and index. ISBN 978-1-4200-4372-3 (alk. paper) 1. Smart materials. 2. Smart structures. I. Schwartz, Mel M. II. Title. TA418.9.S62S48 2008 620.1’1--dc22

2008018721

Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com

43722_C000.indd iv

10/18/2008 7:26:40 AM

Dedication To Carolyn and Anne-Marie who light up my life every day of the year

43722_C000.indd v

10/18/2008 7:26:40 AM

43722_C000.indd vi

10/18/2008 7:26:40 AM

Contents Preface ......................................................................................................................................................... xi Editor ........................................................................................................................................................... xv Contributors .............................................................................................................................................. xvii

1

Residual Stress in Thin Films ............................................................................................................ 1-1

2

Intelligent Synthesis of Smart Ceramic Materials ............................................................................ 2-1

3

A.G. Vedeshwar

Wojciech L. Suchanek and Richard E. Riman

Functionally Graded Polymer Blend ................................................................................................. 3-1

Yasuyuki Agari

4

Structural Application of Smart Materials ........................................................................................ 4-1

5

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites ..................................................................................................................... 5-1 5.1 H ybrid Composites S. Padma Priya .........................................................................................................................5-1

R. Sreekala and K. Muthumani

5.2 Design of an Active Composite Wing Spar with Bending–Torsion Coupling Carlos Silva, Bruno Rocha, and Afzal Suleman .......................................................................................................... 5-7

6

Ferromagnetic Shape Memory Alloy Actuators ................................................................................ 6-1

7

Aircraft Applications of Smart Structures ........................................................................................ 7-1

8

Smart Battery Materials .....................................................................................................................8-1

9

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators ........................................ 9-1

10

Yuanchang Liang and Minoru Taya Johannes Schweiger

Arumugam Manthiram

9.1 Smart Ferroelectric Ceramics for Transducer Applications A.L. Kholkin, D.A. Kiselev, L.A. Kholkine, and A. Safari ........................................................................................... 9-1 9.2 Smart Ceramics: Transducers, Sensors, and Actuators Kenji Uchino and Yukio Ito ................................................9-12 9.3 Noncontact Ultrasonic Testing and Analysis of Materials Mahesh C. Bhardwaj ...........................................................9-27

Chitosan-Based Gels and Hydrogels................................................................................................ 10-1 10.1 C hitosan-Based Gels Kang De Yao, Fang Lian Yao, Jun Jie Li, and Yu Ji Yin ......................................................... 10-1

10.2 Chitosan-Based Hydrogels in Biomedical and Pharmaceutical Sciences Claire Jarry and Matthew S. Shive ....................................................................................................................... 10-13

vii

43722_C000.indd vii

10/18/2008 7:26:40 AM

viii

11

Contents

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials .......................................... 11-1

11.1 S mart Adhesives James A. Harvey and Susan Williams ........................................................................................11-1 11.2 Oxides as Potential Ā er moelectric Materials S. Hébert and A. Maignan ............................................................ 11-4 11.3 Electrically Conductive Adhesives Yi Li, Myung Jin Yim, Kyoung-sik Moon, and C.P. Wong ....................................................................................................................... 11-12 11.4 Electrochromic Sol–Gel Coatings L.C. Klein ..................................................................................................... 11-25

12

Cure and Health Monitoring .......................................................................................................... 12-1

13

Drug Delivery Systems ..................................................................................................................... 13-1

14

Tatsuro Kosaka

13.1 Smart Drug Delivery Systems Il Keun Kwon, Sung Won Kim, Somali Chaterji, Kumar Vedantham, and Kinam Park ..................................................................................................................... 13-1 13.2 Drug Delivery Systems: Smart Polymers Joseph Kost ..........................................................................................13-10

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows .......................................... 14-1

14.1 Introduction and Application of Fiber Optic Sensors Nezih Mrad and Henry C.H. Li ...........................................14-1 14.2 S mart Windows John Bell ...................................................................................................................................14-17

15

Flip-Chip Underfill: Materials, Process, and Reliability ................................................................ 15-1

16

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging ......................................................................................... 16-1

Zhuqing Zhang and C.P. Wong

David Kranbuehl

17

Magnetorheological Fluids ...............................................................................................................17-1

18

Intelligent Processing of Materials .................................................................................................. 18-1

19

Magnets, Magnetic, and Magnetostrictive Materials ..................................................................... 19-1 19.1 M agnets, Organic/Polymer Joel S. Miller and Arthur J. Epstein ................................................................ 19-1

J. David Carlson

J.A. Güemes and J.M. Menéndez

19.2

20

Powder Metallurgy Used for a Giant Magnetostrictive Material Actuator Sensor Hiroshi Eda and Hirotaka Ojima .......................................................................................................................... 19-8

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications ..................... 20-1

20.1 Magnetically Controlled Shape Memory Alloys Ilkka Aaltio, Oleg Heczko, Outi Söderberg, and Simo-Pekka Hannula ..................................................................................................................................... 20-1 20.2 Shape Memory Alloys in Micro- and Nanoscale Engineering Applications Yves Bellouard ................................ 20-8 20.3 Mathematical Models for Shape Memory Materials Davide Bernardini and Ā omas J. Pence ............................20-17 20.4 Shape Memory Alloys Jan Van Humbeeck ........................................................................................................ 20-28 20.5 Smart Materials Modeling Manuel Laso ...........................................................................................................20-36 20.6 On the Microstructural Mechanisms of SMEs Monica Barney and Michelle Bartning....................................... 20-41

21

Current Developments in Electrorheological Materials ................................................................. 21-1

22

Carbon Microtubes and Conical Carbon Nanotubes ..................................................................... 22-1

23

“Smart” Corrosion Protective Coatings ..........................................................................................23-1

24

Smart Polymers for Biotechnology and Elastomers ....................................................................... 24-1 24.1 C onducting Polymers Gordon G. Wallace and C.O. Too .................................................................................... 24-1 24.2 Piezoelectricity in Polymers Aleksandra M. Vinogradov ................................................................................... 24-10 24.3 Polymers, Biotechnology, and Medical Applications Igor Yu Galaev and B. Mattiasson .....................................24-13

43722_C000.indd viii

Frank E. Filisko

Santoshrupa Dumpala, Gopinath Bhimarasetti, Suresh Gubbala, Praveen Meduri, Silpa Kona, and Mahendra K. Sunkara Patrick J. Kinlen and Martin Kendig

10/18/2008 7:26:40 AM

Contents

25

ix

Vibration Control for Smart Structures ......................................................................................... 25-1 25.1 Vib ration Control Seung-Bok Choi and Young-Min Han ................................................................................... 25-1

25.2

Smart Materials for Sound and Vibration Control Cai Chao, Lu Chun, Tan Xiaoming, and Zheng Hui .............................................................................................. 25-16

26

Active Truss Structures ................................................................................................................... 26-1

27

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals ............ 27-1

28

Molecular Imprinting Technology ................................................................................................. 28-1

29

Biomedical Sensing ......................................................................................................................... 29-1

30

Intelligent Chemical Indicators ...................................................................................................... 30-1

31

Piezoelectric Polymer PVDF Microactuators ................................................................................. 31-1

32

Ultrasonic Nondestructive Testing and Materials Characterization ............................................ 32-1

33

Lipid Membranes on Highly Ordered Porous Alumina Substrates ............................................... 33-1

B. de Marneffe and André Preumont

Jesse E. Purdy and Alison Roberts Cohan David A. Spivak

Dora Klara Makai and Gabor Harsanyi Christopher O. Oriakhi

Yao Fu, Erol C. Harvey, and Muralidhar K. Ghantasala John A. Brunk

Andreas Janshoff and Claudia Steinem

Index ............................................................................................................................... .............................I-1

43722_C000.indd ix

10/18/2008 7:26:41 AM

43722_C000.indd x

10/18/2008 7:26:41 AM

Preface Many “ smart” m aterials w ere i nvented mo re t han 3 0 ye ars a go, b ut t heir de velopment a nd i mprovement o ver t he pa st t hree decades has led to new, more varied uses of these adaptable materials. Smart materials have properties that can be altered by temperature, moisture, electric or magnetic elds, pH, and stress. Ā ey can change shape and color, become stronger, or produce voltage as a result of external stimuli. Magnetostrictive materials—materials that expand when exposed to a magnetic eld—were discovered in the 1800s, but have uses today as varied as automotive sensors for collision avoidance and vibration dampening for surgical tools. Materials such as brass, nickel-titanium, and gold-cadmium are shape memory alloys (SMAs), which alter their shape in response to changes in temperature and then return to their original shape. In the past 30 years, SMAs have seen widespread use in applications such as miniature surgical tools that can twist and pull when a small amount of heat is applied, wires that expand and contract for use by dentists in straightening teeth, cell phone antennas that resist breaking, stents and bone plates that must fuse together or expand in order to stay in place, and exible eyeglass frames. “Smart m aterials,” w hich  nd widespread applications today, should be recognized along with the familiar metals, plastics, ceramics, composites, powder metals, a nd specialty-type a nd multifunctional materials (SMAs, microelectromechanical systems [MEMS], f unctionally g raded m aterials [FGMs], a nd na nomaterials). Ā ree de cades a go, i t app eared a s t hough sm art m aterials would be the next step in engineering design. By using smart materials instead of adding mass, engineers can endow structures with built-in responses to a myriad of contingencies. In their various forms, these materials can perform as actuators, which can adapt to their environments by changing characteristics such as shape and stiff ness, or as sensors, which provide the actuators with information about structural and environmental changes. All aircraft, whether military commercial, or privately owned, should be and are inspected and maintained on a re gular basis. Intelligent research programs benet t his segment of t he aerospace industry. Sensors introduced into current a nd f uture aircraft designs will help technicians and mechanics to provide a more sophisticated inspection technique for maintenance and repairs. Smart materials are beginning to play an important role in civil engineering designs for dams, bridges, highways, and buildings. An example of a smart materials project sponsored by the U.S. Navy and Army Corps of Engineers is to remove corrosion that has damaged a Navy pier and install sheets of composite materials containing sensors. Ā ese sensors embedded throughout a concrete and c omposite s tructure c an ac t l ike ner ves, s ensing w hen a reas o f t he s tructure b egin to de grade a nd a lerting m aintenance engineers to the need for repairs. Ā e automotive industry is also eager to i ncorporate intelligent materials technology. Currently, researchers are working on an industry-sponsored project to develop smart car seats that can identify primary occupants and adapt to their preferences for height, leg room, back support, and so forth. More profound changes are looming in automotive design based on smart materials. For instance, the technology to enable cars to tell owners how much air pressure tires have, when oil changes are needed, and other maintenance information exists as of today. However, this technology is expensive, but developing solid state and smart materials technologies will bring costs down. Ā ere is a shift in the culture toward consumers being given more information and taking more responsibility for knowing when maintenance and repairs are needed. Ā e future of stereophonic sound will be altered by another facet of smart materials research. Ā e development of u ltrahigh-delity stereo speakers using piezoelectric actuators, which expand and contract in thousandths of a s econd in response to applied voltage, is aimed at t urning whole house walls or car interiors into speakers by imbedding them with the tiny actuators. Ā us, 50 years from now, we would not need to install separate speakers in our homes and cars in an attempt to achieve maximum musical effects. Our cars and houses will offer built-in surround sound. A new technology for implants that may improve construction or repair of bones in the face, skull, and jaw has been developed by researchers from the American Dental Association Foundation and the National Institute of Standards and Technology. Ā e new

xi

43722_C000.indd xi

10/18/2008 7:26:41 AM

xii

Preface

technology provides a method for making scaffolds for bone tissue. Ā e scaffold is seeded with the patient’s own cells and is formed with a cement paste made of minerals also found in natural bone. Ā e paste is mixed with beads of a natural polymer (made from seaweed)  lled with bone cells. Ā e paste is shaped or injected into a bone cavity and then allowed to harden with the encapsulated cells dispersed throughout the structure. Ā e natural polymer beads gradually dissolve when exposed to the body’s uids, creating a scaffold that is  lled by the now released bone cells. Ā e cement, a c alcium phosphate material, is strengthened by adding chitosan, a b iopolymer extracted from crustacean shells. Ā e implant is further reinforced to about the same strength as spongy natural bone by covering it with several layers of a biodegradable ber mesh already used in clinical practice. Bone cells are very smart. Ā ey can tell the difference between materials that are bioactive compared to bioinert polymers. Ā e material is so designed to be similar to mineral in bone so that cells readily attach to the scaffold. In addition to creating pores in the hardened cement, the natural polymer beads protect the cells during the 30 min required for the cement to harden. Future experiments will develop methods for improving the material’s mechanical properties by using smaller encapsulating beads that biodegrade at a predictable rate. Other developments include a smart bandage, which has an in situ programmable medical device to treat wounds and burns, a smart pill for vitamin and nutraceuticals formulation that prevents cognitive decline in aging and amphiphilic polymeric materials where the polymers can be used in the development of physiologically stable, nonleaking, nonimmunogenic, safe, and efficacious targeted drug delivery systems. Recent advances in nondestructive evaluation (NDE) sensor technologies, health monitoring, and life prediction models are revolutionizing component inspections and life management, and will signicantly improve the reliability and airworthiness of aerospace s ystems. S everal ke y adv ances i n N DE, health, monitoring, a nd prognostics programs i nclude emerg ing s tructural health monitoring technologies for aerospace applications, high-temperature health monitoring, advances in NDE technologies for measurement of subsurface residual stresses, computational methods and advanced NDE techniques, and materials damage assessment techniques a nd p rognosis mo dels f or p rediction o f rem aining u seful l ife. Ā e adv ancements de scribed p rovide t he te chnologies required for lowering m aintenance c osts, i ntegrated d amage prognosis a nd l ife prediction, en hancing rel iability a nd s afety, a nd improving the performance and operational efficiency of current and future aerospace systems. Another area of focus is smart machines. For example, moving composite manufacturers out of hand layup and open molding into more environmentally acceptable closed-molding alternatives where a reliable source of preforms is a necessity for volume production in closed-mold processes, such as resin transfer molding (RTM), vacuum-assisted RTM (VARTM), and vacuum-assisted resin infusion molding (VARIM). Large preforms are facilitated by the large-scale preformer (LSP). Ā e LSP’s computerized spray-up process requires no human contact from the time the tooling enters the preform manufacturing cell until it is ready to be demolded: spray-up, compression, cure of the binder, cooling, and demold operations are all accomplished robotically according to i nstructions preprogrammed into the system’s soft ware. LSP is a smart system that tells the user where every part is in the process at all times. Ā e LSP’s inaugural application, and the largest preform produced thus far, was for fuel containment vessels, which were resin transfer molded. Ā ese vessels have been installed beneath gasoline pumps in gas stations to prevent contamination of soil if there is a leak in the pump-to-tank plumbing. Composites a re now used i n monitoring systems where t hey a re combined w ith ot her materials a nd sensors. A n example is a composite “shape sensing mat” for use with a metallic riser system. Ā e exible mat, which incorporates ber optics, wraps around a steel riser, and enables operators to monitor excessive bending and fatigue life during riser deployment on a dynamic positioned oil drilling ship moored in the Gulf of Mexico as well as smart downhole coiled tubing complete with power and data transmission capabilities for drilling or workover applications and oileld pipelines. Finally, machines can process simple commands, but they are not very good at guring out complex orders or unstated common sense. Command a machine to “paint the computer case before you box it,” or “provide power to the computer before you switch it on” and the machine may box the product before the case is dry, or plug and then unplug a c omputer before switching it on. Ā e meaning of the word “before” is quite different in these two cases. Ontologists, who study and understand the thought process, hope to end t he a ge o f s tupid m achines. O ntologists, w ho h ave c reated s ome o f t he mo st adv anced log ic s ystems, p lan to s hare t heir leading-edge concepts on such comprehensive ideas as time, space, and process. Ā e promise to cooperate eventually could lead to soft ware that will enable machines to interpret and act on commands with near human common sense. Efforts to equip machines with articial intelligence capacity have, up to now, been relatively rudimentary. Softwa re programs might, for instance, give machines used to m ake furniture considerable “understanding” of terms and frames of reference used in the furniture business. But such collected knowledge is of limited use, and human operation is necessary at virtually every step in the manufacturing process. A machine that incorporates expanded frames of reference of such “higher ontologies” as space

43722_C000.indd xii

10/18/2008 7:26:41 AM

Preface

xiii

and cost might be able to make design and shipping decisions virtually on its own. Ā e future is bright and with optimism will enable the leading ontologists throughout the world to continue this promising work. Ā e book contains many of the examples and of the aspects that I have discussed and the contributions and efforts of 60 experts in the various  elds of smart materials and smart material systems. I h ope the readers will appreciate the work of a m ultitude of scientists, educators, researchers, academia, and industry people who have made considerable innovative progress in bringing forth their endeavors. Mel Schwartz

43722_C000.indd xiii

10/18/2008 7:26:41 AM

43722_C000.indd xiv

10/18/2008 7:26:41 AM

Editor Mel S chwartz h as de grees i n me tallurgy a nd eng ineering m anagement a nd h as s tudied l aw, me tallurgical eng ineering, a nd education. His professional experience extends over 51 years. He has served as a metallurgist in the U.S. Bureau of Mines; as a metallurgist and producibility engineer in the U.S. Chemical Corps; as a technical manufacturing manager, chief R&D laboratory, research manufacturing engineering, and senior staff engineer in Martin-Marietta Corporation for 16 years; as a program director, manager and director of manufacturing for R&D, and chief metals researcher in Rohr Corp for 8 years; and as a staff engineer and specication specialist, chief metals and metal processes, and manager of manufacturing technology in Sikorsky Aircraft for 21 years. After retirement, Schwartz served as a consultant for many companies including Intel and Foster Wheeler, and is currently editor of SAMPE Journal of Advanced Materials. Schwartz’s professional awards and honors include Inventor Achievement Awards and Inventor of the Year at Martin-Marietta; C. Adams Award and Lecture and R.D. Ā omas Memorial Award from AWS;  rst recipient of the G. Lubin Award and an elected fellow from SAMPE; an elected fellow and Engineer of the Year in CT from ASM; and the Jud Hall Award from SME. Schwartz’s other professional activities include his appointment to ASM technical committees (joining, composites and technical books, ceramics); manuscript board of review, Journal of Metals Engineering as peer reviewer; the Institute of Metals as well as Welding Journal as peer reviewer; U.S. leader of International Institute of Welding (IIW) Commission I ( brazing and related processes) for 20 years and leader of IIW Commission IV (electron beam/laser and other specialized processes) for 18 years. Schwartz owns ve patents, the notable one being aluminum dip brazing paste commercially sold as Alumibraze. He has authored 16 books and over 100 technical papers and articles and is an internationally known lecturer in Europe, the Far East, and Canada. He has taught i n U.S. Universities (San Diego State University, Yale University), ASM i nstitutes, McGraw-Hill seminars, a nd i nhouse company courses.

xv

43722_C000.indd xv

10/18/2008 7:26:41 AM

43722_C000.indd xvi

10/18/2008 7:26:42 AM

Contributors Ilkka Aaltio Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland

John A. Brunk National Nuclear Security Agency’s Kansas City Plant Kansas City, Missouri

Hiroshi Eda Intelligent Systems Engineering Department Ibaraki University Hitachi, Japan

Yasuyuki Agari Osaka Municipal Technical Research Institute Osaka, Japan

J. David Carlson Lord Corporation Cary, North Carolina

Monica Barney Nitinol Devices and Components Freemont, California

Cai Chao Institute of High Performance Computing Singapore

Arthur J. Epstein Department of Physics and Department of Chemistry Ohio State University Columbus, Ohio

Michelle Bartning Cordis Advanced Medical Ventures Freemont, California John Bell Queensland University of Technology Brisbane, Queensland, Australia

Somali Chaterji Weldon School of Biomedical Engineering Purdue University West Lafayette, Indiana

Frank E. Filisko Ā e University of Michigan Ann Arbor, Michigan Yao Fu Silverbrook Research Pty. Ltd. St. Balmain, Sydney, Australia Igor Yu Galaev Lund University Lund, Sweden

Yves Bellouard Mechanical Engineering Department Eindhoven University of Technology Eindhoven, Ā e Netherlands

Seung-Bok Choi Department of Mechanical Engineering Inha University Inchon, South Korea

Davide Bernardini Department of Structural and Geotechnical Engineering University of Rome Rome, Italy

Lu Chun Institute of High Performance Computing Singapore

Mahesh C. Bhardwaj Ā e Ultran Group State College, Pennsylvania

Alison Roberts Cohan Pacic Whale Foundation Maui, Hawaii

Suresh Gubbala Department of Chemical Engineering University of Louisville Louisville, Kentucky

Gopinath Bhimarasetti Department of Chemical Engineering University of Louisville Louisville, Kentucky

Santoshrupa Dumpala Department of Chemical Engineering University of Louisville Louisville, Kentucky

J.A. Güemes Department of Aeronautics Universidad Politécnica de Madrid Madrid, Spain

Muralidhar K. Ghantasala Department of Mechanical and Aeronautical Engineering Western Michigan University Kalamazoo, Michigan

xvii

43722_C000.indd xvii

10/18/2008 7:26:42 AM

xviii

Young-Min Han Department of Mechanical Engineering Inha University Inchon, South Korea Simo-Pekka Hannula Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland Gabor Harsanyi Department of Electronics Technology Budapest University of Technology and Economics Budapest, Hungary Erol C. Harvey Industrial Research Institute Swinburne University of Technology Hawthorne, Melbourne, Australia James A. Harvey Under the Bridge Consulting Inc. Corvallis, Oregon S. Hébert Laboratoire Crismat CNRS ENSICaen Caen, France

Contributors

Claire Jarry Bio Syntech Canada, Inc. Laval, Quebec, Canada Martin Kendig Teledyne Scientic Company Ā ousand Oaks, California A.L. Kholkin Department of Ceramics and Glass Engineering and Center for Research in Ceramic and Composite Materials University of Aveiro Aveiro, Portugal L.A. Kholkine Department of Electrical and Computer Engineering University of Porto Porto, Portugal Sung Won Kim Department of Pharmaceutics Purdue University West Lafayette, Indiana Patrick J. Kinlen Crosslink Fenton, Missouri

Joseph Kost Department of Chemical Engineering Ben-Gurion University Beer Sheva, Israel David Kranbuehl Chemistry and Applied Science Departments College of William and Mary Williamsburg, Virginia Il Keun Kwon Department of Pharmaceutics Purdue University West Lafayette, Indiana Manuel Laso Laboratory of Non-Metallic Materials Universidad Politécnica de Madrid Madrid, Spain Henry C.H. Li School of Aerospace, Mechanical and Manufacturing Engineering RMIT University Fishermans Bend, Victoria, Australia Jun Jie Li School of Chemical Engineering and Technology Tianjin University Tianjin, China

Oleg Heczko Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland

D.A. Kiselev Department of Ceramics and Glass Engineering and Center for Research in Ceramic and Composite Materials. University of Aveiro Aveiro, Portugal

Zheng Hui Institute of High Performance Computing Singapore

L.C. Klein Rutgers University Piscataway, New Jersey

Yuanchang Liang Department of Mechanical Engineering University of Washington Seattle, Washington

Silpa Kona Department of Electrical and Computer Engineering University of Louisville Louisville, Kentucky

A. Maignan Laboratoire Crismat CNRS ENSICaen Caen, France

Jan Van Humbeeck Catholic University of Leuven Leuven, Belgium Yukio Ito Ā e Pennsylvania State University University Park, Pennsylvania Andreas Janshoff Johannes Gutenberg Universität Mainz, Germany

43722_C000.indd xviii

Tatsuro Kosaka Graduate School of Engineering Osaka City University Osaka, Japan

Yi Li School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia

Dora Klara Makai Department of Electronics Technology Budapest University of Technology and Economics Budapest, Hungary

10/18/2008 7:26:42 AM

Contributors

Arumugam Manthiram Materials Science and Engineering Program University of Texas at Austin Austin, Texas B. de Marneffe Active Structures Laboratory Université Libre de Bruxelles Brussels, Belgium B. Mattiasson Lund University Lund, Sweden Praveen Meduri Department of Chemical Engineering University of Louisville Louisville, Kentucky J.M. Menéndez Composite Materials Technology Department Airbus Getafe (Madrid), Spain Joel S. Miller Department of Chemistry University of Utah Salt Lake City, Utah Kyoung-sik Moon School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia Nezih Mrad Department of National Defence National Defence Headquarters Ottowa, Ontario, Canada K. Muthumani Structural Dynamics Laboratory Structural Engineering Research Centre, CSIR Chennai, India Hirotaka Ojima Intelligent Systems Engineering Department Ibaraki University Hitachi, Japan Christopher O. Oriakhi Imaging and Printing Supplies Hewlett-Packard Company Corvallis, Orgeon

43722_C000.indd xix

xix

Kinam Park Department of Biomedical Engineering Purdue University West Lafayette, Indiana

David A. Spivak Department of Chemistry Louisiana State University Baton Rouge, Louisiana

Thomas J. Pence Department of Mechanical Engineering Michigan State University East Lansing, Michigan

R. Sreekala Structural Dynamics Laboratory Structural Engineering Research Centre, CSIR Chennai, India

André Preumont Active Structures Laboratory Université Libre de Bruxelles Brussels, Belgium

Claudia Steinem Georg-August Universität Göttingen, Germany

S. Padma Priya University of Mysore Mandya, India

Wojciech L. Suchanek Sawyer Technical Materials, LLC Eastlake, Ohio

Jesse E. Purdy Department of Psychology Southwestern University Georgetown, Texas

Afzal Suleman Department of Mechanical Engineering University of Victoria Victoria, British Columbia, Canada

Richard E. Riman Department of Materials Science and Engineering Rutgers University Piscataway, New Jersey Bruno Rocha Department of Mechanical Engineering Instituto Superior Técnico Lisbon, Portugal A. Safari Rutgers University Piscataway, New Jersey Johannes Schweiger German Aerospace Society Ban Heilbrunn, Germany Matthew S. Shive Bio Syntech Canada, Inc. Laval, Quebec, Canada Carlos Silva Laboratory of Aeronautics Portuguese Air Force Academy Sintra, Portugal Outi Söderberg Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland

Mahendra K. Sunkara Department of Chemical Engineering University of Louisville Louisville, Kentucky Minoru Taya Department of Mechanical Engineering University of Washington Seattle, Washington C.O. Too Intelligent Polymer Research Institute University of Wollongong Wollongong, New South Wales, Australia Kenji Uchino Department of Electrical Engineering Ā e Pennsylvania State University University Park, Pennsylvania and Micromechatronics Inc State College, Pennsylvania Kumar Vedantham Department of Pharmaceutics Purdue University West Lafayette, Indiana

10/18/2008 7:26:42 AM

xx

Contributors

A.G. Vedeshwar Department of Physics and Astrophysics University of Delhi New Delhi, India

C.P. Wong School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia

Aleksandra M. Vinogradov Department of Mechanical and Industrial Engineering Montana State University Bozeman, Montana

Tan Xiaoming Institute of High Performance Computing Singapore

Gordon G. Wallace Intelligent Polymer Research Institute University of Wollongong Wollongong, New South Wales, Australia

Fang Lian Yao School of Chemical Engineering and Technology Tianjin University Tianjin, China

Susan Williams Hewlett Packard Company Corvallis, Oregon

Kang De Yao Research Institute of Polymeric Materials Tianjin University Tianjin, China

43722_C000.indd xx

Myung Jin Yim School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia Yu Ji Yin Research Institute of Polymeric Materials Tianjin University Tianjin, China Zhuqing Zhang Hewlett-Packard Company Corvallis, Oregon

10/18/2008 7:26:42 AM

1 Residual Stress in Thin Films ntroduction .............................................................................................................................. 1-1 Models and The oretical Background..................................................................................... 1-1 Experimental Methods for the Measurement of Residual Stress ...................................... 1-2 Residual Stress-Dependent Optical Properties of Some Layered Structured Semiconductors .........................................................................................................................1-4 1.5 Summary and Future Direction ........................................................................................... 1-12 References ........................................................................................................................................... 1-12

1.1 I 1.2 1.3 1.4

A.G. Vedeshwar University of Delhi

1.1 Introduction One o f t he mo st c ommon a nd ye t a n i mportant i ssue i n t hin  lms is the persistence of stress aĀer t he  lm growth, which is termed as t he internal residual st ress. A lthough t he attempt to measure a nd u nderstand t he s tress i n  lms s tarted a s e arly a s 1877 [1], the topic continues to be interesting and important even today with various innovative analyses and measurement techniques. A few early authoritative reviews [2,3] on this topic have been excellent sources of information, which has assisted in the further g rowth o f k nowledge. The re sidual s tress de pends o n various factors like t he method of growth, growth pa rameters, nature of substrates, and starting material and then aĀer growth  lm processing, etc. Development of stress in  lms c ould b e both adv antageous a nd d isadvantageous. A v ery w ell-known, advantageous effect is the formation of self-assembled quantum dots i n h eteroepitaxial g rowths c aused b y t he de velopment o f stress due to the lattice mismatch between lm and the substrate [4,5]. Thus, the quantum dot growth can be tailored or controlled in a de sirable w ay b y a p roper c hoice o f t he e xtent o f  lm– substrate lattice mismatch. This has already led to the successful development of various devices based on quantum dots. However, residual stress may be quite undesirable and disadvantageous in the f abrication o f a lmost a ll ot her p lanar t hin  lm electronic devices as it could lead to failure of the device, or else can modify the de vice p erformance u ndesirably, e .g., m icroelectronic o r microelectromechanical s ystems ( MEMS). Sub sequently, t he study of the residual stress effect and its elimination or minimization b ecomes ne cessary f or suc h app lications. M ost o f t he properties of the materials are affected by the stress, either externally applied or the internal residual stress. Therefore, the study of t he residual stress effect in thin  lms could be a nalogous to the e xternally app lied p ressure e ffect. This fac t faci litates t he

possibility of studying the pressure-dependent physical properties without any actual externally applied high-pressure experiments. As the band structure depends crucially on the structure, the electronic properties of the semiconductors can be expected to e xhibit i nteresting s tress–strain de pendence. The s tudy o f residual stress therefore seems to be of signicant importance, although t he f undamental me chanisms f or i ts o rigin a re f ar from being fully understood. Therefore, I will try to present the necessary and sufficient information on the topic in this limited review.

1.2

Models and Theoretical Background

Almost all lms have stress and conceptually the total stress can be thought of having three major contributions, external, thermal and the intrinsic. It can be represented as [6] σ = σ external + σ thermal + σ intrinsic (1

.1)

as o Āen d istinguished i n t he l iterature. I ntrinsic s tresses a re developed d uring t he de position p rocess o f t he  lms. Ther mal stress arises due to the difference between the thermal expansion coefficients o f t he sub strate a nd t he  lm m aterial, a nd o ccurs especially during the cooling phase. External stresses are mainly considered as due to, for instance, oxidation and incorporation of impurities etc. Most of the formulas used in experimental determinations of  lm stress sf are the modications of an equation rst derived by Stoney in 1909 [7], which is given by σf =

E Sd S2 (1 6R (1 − ν S )d f

.2) 1-1

43722_C001.indd 1

10/3/2008 8:42:10 PM

1-2

Smart Materials

where ES and νS are Young’s modulus and the Poisson ratio of the substrate, respectively d S and d f are the thickness of substrate and film, respectively R i s t he r adius o f c urvature o f t he b ending i n t he  lm– substrate system caused by the stress The me asurement of R w ill de termine sf. I n t he l iterature, t he (1 − νS) correction is frequently omitted. Most of the optical interferometric methods devised to measure R (or change in R) are based on the dependence of fringe-width on R. The origin of intrinsic stress is not yet fully explained by any universal mechanism. However, Buckel [8] suggested and classied the processes leading to the generation of internal stress into t he c ategories a s: ( 1) d ifferences i n t hermal e xpansion coefficient o f  lm a nd sub strate, (2) d ifferences i n t he l attice spacing of single crystal substrates a nd t he  lm during epitaxial growth, (3) variation of the interatomic spacing with crystal size, (4) adsorption or incorporation of atoms from residual gases o r c hemical re actions, ( 5) re crystallization p rocesses, (6) microscopic voids and special arrangements of dislocation, and (7) phase transformations. Large i ntrinsic ten sile s tresses ob served i n me tal  lms has been explained by a model [9], which suggests that the annealing and shrinkage of the layer buried behind the advancing surface of t he g rowing  lm le ads to s tress. Hoff man [10] t ried to e stimate the intrinsic residual stress using a model in which the isolated g rains c oalesce a nd f orm g rain b oundaries. Gr ains o f radius R are assumed to be separated by a distance d (constrained relaxation leng th), o f t he o rder o f i nteratomic d istances. This model u ses a g rain-boundary p otential ( similar to M oorse potential) having a minimum at equilibrium atomic separation. The tensile or compressive strain developed in the  lm is attributed to the situation of growth whether the interatomic distance is smaller or greater than the equilibrium distance, respectively. The intrinsic residual stress deduced by this model is given by σ intrinsic =

Efδ P (1 (1 − ν f )2R

.3)

where P is the packing density of the lm Ef an d νf a re Young’s mo dulus a nd P oisson r atio o f t he lm, respectively It should be noted that the term Ef /(1 − νf ) replaces Ef because of the assumed biaxial nature of the stress. The parameter d can be determined from the interaction potential between the concerned atoms, which is oĀen not so easy. Recently, Nix and Clemens [11] attempted to i mprove t his mo del o f c oalescence me chanism using a thermodynamical approach. They derived a relationship for the maximum value of the stress as 1/ 2

⎡ E f (2γ g − γ gb ) ⎤ σ intrinsic (max) = ⎢ ⎥ (1 ⎣ (1 − ν f )R ⎦

43722_C001.indd 2

.4)

where gg and ggb are the surface tensions per unit area of the isolated g rains a nd t he g rain b oundary, re spectively. A gain, h ere too, the determination of the parameter ggb is quite difficult in this improved model. Another very recent report [12] attempted to explain t he i ntrinsic residual stress i n metal  lms using a mo del based on t he size-dependent phase t ransformations of t he na nograins via the size dependence of the melting temperature of nanoparticles assumed to be present at the early stage of lm deposition. Similarly, a model for compressive stress generation in polycrystalline  lms during t hin  lm growth is reported [13] in which the driving force is an increase in the surface chemical potential caused by the deposition of atoms from the vapor. The increase in surface chemical potential induces atoms to  ow into the grain boundary and hence c reates a c ompressive s tress i n t he  lm. A n umber of other s tress mo dels [14–18] der ived f rom t hese ide as h ave b een invoked in t he literature to e xplain both tensile a nd compressive stress in thin lms, which will not be discussed further. The magnitude of intrinsic stresses in lms may also be related to the microstructure of lms, i.e., morphology, texture, grain size, etc. Ther mal effects can also contribute to  lm stress signicantly. Films grown at elevated substrate temperatures and then cooled to a mbient tem perature w ill de velop t hermal s tress. Si milarly, the lms either thermally cycled or cooled to cryogenic temperatures will also develop thermal stress. A biaxial stress can appear in lms grown on substrates having different thermal expansion coefficients th an th e  lm, at a tem perature T higher or lower than t he subst rate o r de position tem perature TS. The strain eT developed under such conditions is given by [3] ε T = (α f − α s )∆T (1

.5)

where ∆T = (T – TS), a s, and a f are the thermal expansion coefficients of the substrate and  lm, respectively. Using Hooke’s law, the relation between thermal stress and the elastic strain in the absence of any plastic deformation in t he  lm–substrate structure during temperature change can be obtained as σ f (T ) =

E f (α s − α f )∆T (1 (1 − ν f )

.6)

A t hird k ind, ter med a s e xtrinsic s tress, c an b e d istinguished apart f rom t hermal s tress a nd i ntrinsic s tress i n t he o verall residual s tresses i n  lms. Various mole cules c an p enetrate t he open voids and pores present in a not so fully dense lm and are adsorbed on pore walls. The interaction forces between adsorbed species, e specially b etween p olar s pecies suc h a s w ater mole cules, can modify residual stresses. Hirsch [19] proposed a model to explain the origin of extrinsic stress based on the adsorption of polar species on pore walls.

1.3 Experimental Methods for the Measurement of Residual Stress A wide variety of methods have been explored for measuring the stress in thin  lms [6,20]. They may broadly be categorized on

10/3/2008 8:42:12 PM

1-3

Residual Stress in Thin Films

the basis of the physical phenomena used as the technique such as diff raction (x-ray or electron diff raction), optical interferometry, electrical or electromechanical etc. However, a c onvenient and app ropriate me thod m ay b e c hosen de pending o n t he requirements like measurements in situ or ex situ, types of lm– substrate m aterial, le vel o f ac curacy e tc. The interferometric methods are normally used to measure the extent of bending or deformation of a thin substrate (cantilever) from its equilibrium position caused by the  lm h aving s tress de posited o n i t. The fringe width is proportional to the extent of bending or deformation and hence the stress can be determined. The nature of the biaxial stress such as balanced, unbalanced, or the different type of stress components, etc. can also be analyzed by the shape of the f ringe pat tern a s has been reported for sputtered Mo  lms [21]. A tensile stress will bend the substrate so that the lm surface is concave and the opposite (convex) happens with the compressive stress. A thin cantilever is most widely used as the substrate. A v ariety of i nterferometric techniques such as laser reective interferometry [22–24], laser spot scanning interferometry [25], optical leverage with a laser beam [26,27], etc. have been developed. However, it should be noted that the stresses measured by cantilever techniques and x-ray diffraction (XRD) need not ne cessarily the same because X RD gives the strain in the crystal lattice of the grain (intragranular) while the cantilever me thod de termines t he s tress d ue to g rain b oundaries between the grains (intergranular). There are a few other innovative modern techniques developed to meet specic requirements such a s i ndentation [ 28,29] a nd l aser s pallation [ 30], s train gauges [31–33], Raman measurements [34–36], and so on. X-ray or electron diff raction i s one of t he si mplest ye t mo st powerful me thods f or m icrostructural c haracterization. Therefore, more emphasis will be given to this technique here. The details can be found from the standard and widely referred books [37,38]. XRD data can also be used to determine either uniform or nonuniform residual stress in the lms. The effect of strain, both uniform and nonuniform, on the direction of x-ray reection causes the diff raction peak in a denite way. The uniform strain shiĀs the peak on either side of the unstrained 2q to the lower side for tensile and to the higher side for compression. If a g rain is given a u niform tensile strain at r ight angles to t he reecting p lanes ( parallel to sub strate p lane), t heir s pacing becomes larger than unstrained equilibrium spacing d 0 and the corresponding diffraction line shiĀs to lower angles but does not otherwise change. This line shiĀ for the given (hkl) can be used to calculate the strain e z present on (hkl) along the z-axis in the lm as εz =

0 ∆d hkl d hkl − d hkl (1 = 0 0 d hkl d hkl

.7)

The r ight h and side o f E quation 1.7 w ill b e p ositive for ten sile stress and negative for a c ompressive stress. However, it should be noted that normally the d0 value for the desired (hkl) is taken from powder data (ASTM or JCPDS) of the concerned material because f or a go od app roximation, p owder m ay b e t reated a s

43722_C001.indd 3

stress free or having a negligible stress. Thus, strain for all (hkl) peaks o f c onsiderable i ntensity i n t he d iffractogram c an b e determined and their average ∆d/d can be calculated for a r andomly o riented p olycrystalline  lm. The m acrostress t hus c an be de termined b y m ultiplying t he a verage s train ∆d/d by t he elastic constant of the material. However, quite oĀen, there will be a single most intense peak (hkl) in the diff ractogram for (hkl)oriented parallel to the substrate plane and the determination of strain is only for that (hkl). Similarly, using electron diffraction data, the d-spacings can be determined [39] and hence the strain. For a biaxial stress along the x- and y-axes (sx ≠ 0, sy ≠ 0), sz = 0 we have strain normal to the  lm surface given by [37,38] εz = −

νf (σ x + σ y ) (1 Ef

.8)

The above equation represents a contraction if sx and sy are tensile. Therefore, a p ositive e z determined by Equation 1.7 would imply t he c ompressive b iaxial s tress. A c omponent o f b iaxial stress in any desired direction on the xy-plane can be measured by tilting the sample by an angle y, the details of which may be found in Refs. [37,38]. This method is known as the sin2y method in t he mo dern l iterature. A no nuniform m icrostrain c auses a broadening of t he corresponding diffraction line. The relationship between t he broadening produced and t he nonuniformity of the strain can be obtained by differentiating Bragg’s law. We therefore obtain b = ∆2q = −2

∆d tan q (1 d

.9)

where b is the extra broadening, over and above the instrumental breadth of the line, due to a fractional variation in plane spacing, ∆d/d. This equation a llows the variation in strain, ∆d/d, to be c alculated f rom t he ob served b roadening. The maximum strain thus found can be multiplied by the elastic modulus Ef to give the maximum stress present. Although considerable progress has taken place both in theory and in experimental methods, the residual stress data in the literature show wide variations for various materials measured by different me thods. Bro ad t rends a re v isible i n t he p ublished results and can be summarized as 1. Most o f t he me tal  lms e xhibit i nvariably ten sile s tress with a l arge magnitude (200–1100 MPa or 106 N/m 2) and less dependence on the nature of the substrate. 2. Refractory metals and metals having high melting points generally e xhibit h igher s tresses ( 600–1200 MPa) t han soĀer and lower melting point metals (20–100 MPa). 3. Normally, b oth c ompressive a nd ten sile s tresses a re observed i n no nmetallic  lms f requently w ith sm aller magnitudes (7–700 MPa). The crossover between the two types of stress is also observed. Data on some metals can be found in Refs. [3,9]. Similarly, early investigations on dielectric and optical coatings have been

10/3/2008 8:42:14 PM

1-4

Smart Materials

described in Refs. [3,40,41]. A re cent report shows the relationship b etween t he p referred o rientations a nd s tresses f or s ome metals [42]. The stress behavior in a very important technological material, GaN, has been studied [43–45] and ot her nitrides also have been investigated [46–48]. These are only a few informative examples. However, there are a large numbers of reports on v arious m aterials, s pecic s tudies, e xperimental me thods, analyses e tc., w hich a re b eyond t he s cope a nd p urpose o f t his chapter.

1.4

Residual Stress-Dependent Optical Properties of Some Layered Structured Semiconductors

It may be i nteresting to u nderstand t he origin a nd behavior of residual stress i n  lms; however, it is more important to know how the residual stress affects the various properties of the lm. In this section, some selected data will be presented for illustrating some of the points so far discussed. The layered structured semiconductors (like metal dihalides, chalcogenides, etc.) were chosen for t his purpose b ecause of t heir a nisotropy a long a nd across the layer. In a layered structure, the constituent atoms are bound by a strong covalent bonding within the ab-plane forming

a sheet-like structure (layer) and these sheets are stacked along the t hird d irection (c-axis) b y a w eak v an der W aals b onding between t hese s heets to f orm a t hree-dimensional s olid, t hus leading to a nisotropy a long pa rallel a nd p erpendicular to t he c-axis. A ll t he  lms p resented i n t his s ection w ere g rown b y thermal evaporation using a molybdenum boat onto a glass substrate at room temperature (RT) or liquid nitrogen temperature (LNT) at a v acuum of about 10−6 Torr a nd t heir stoichiometry was well characterized and conrmed by XPS and EDAX. XRD in Bragg–Brentano focusing geometry was used for structural and residual stress analyses. Apart f rom t he i nherent a nisotropy, m any me tal io dides prefer certain (hkl)-oriented growth and a great affin ity for crystallinity (even at LNT) in thin lms. A typical case of lead iodide is illustrated in Figure 1.1 where the structural and strain analyses are su mmarized. Without a ny exception, PbI2  lms g row w ith (00l)-orientations (that is (00l) planes parallel or the c-axis perpendicular to the substrate plane) under any growth conditions including the low temperature. Th is is manifested by the appearance of only (00l) peaks in XRD as shown in Figure 1.1 even for ultrathin  lms. Films thinner than 20 nm were analyzed by transmission electron microscopy (TEM). Only representative micrographs are shown in the insets (a) and (b) for a 5 nm thick lm. The d-spacings (with t he M iller i ndices i ndicated o n t he

400

(106) (105)

(00l)-Oriented PbI2 Films

(001)

Film thickness 61 nm 93 nm

(201) (111) COUNTS (arb. units)

Counts

300

200

12

100

(b) (001)

12.5

13

13.5

2θ (degrees) 334 nm

(002)

(003)

(004)

0

(c)

Counts (arb. units)

254 nm 334 nm

5 nm thick film (a)

133 nm 173 nm

10

20

30

40

50

60

2q (⬚) 15 nm

(003)

(004)

30 nm 10

20

30

40

50

60

2q (⬚)

FIGURE 1.1 Typical XRD scan of (00l)-oriented ultrathin PbI2  lms for two  lm thicknesses as labeled in the gure. The inset (a) shows TEM for a 5 nm thick  lm and (b) is its electron diff raction. Inset (c) displays the XRD for very thick  lm and the inset within shows the shiĀ of (001) peak from unstrained value (ASTM data indicated by the broken line) for various  lm thicknesses.

43722_C001.indd 4

10/3/2008 8:42:16 PM

1-5


0.0075

18

Grain size (nm)

Ultrathin PbI2 films

0.005

0.0025

by TEM 13

by XRD

8 3

∆d/d

0

5

10

15

20

25

30

Film thickness (nm)

0

−0.0025

by TEM by XRD

−0.005

2

6

10

14 18 22 Film thickness (nm)

26

30

FIGURE 1.2 Development of s train e z in ultrathin PbI2  lms as a f unction of  lm thickness. Inset shows a l inear grain size growth with  lm thickness. The method of strain and grain size determination is identied in the gure.

micrograph) calculated from the radius of the rings match quite well with ASTM card No. 07-0235 for 2H polytype of PbI2. The unstrained value of dhkl is taken from the above ASTM data for determining the strain. The (001) peak for thicker  lms are displayed i n t he i nset w ithin i nset (c) de picting t he s train, ei ther tensile or compressive, in comparison with the ASTM value indicated by the broken vertical line. Strain is almost negligible in  lms thinner than 12 nm as depicted in Figure 1.2, which shows the development of strain with  lm thickness. The functional dependence of strain on grain size can be expected to be similar to that on lm thickness because of the linear relationship b etween t he t wo a s shown i n t he i nset. Q ualitatively, t his seems to b e c onsistent w ith t he c oalescence mo del men tioned earlier a s g rains a re well s eparated i n u ltrathin  lms (T EM i n inset (a)) and start coalescing in thicker lms, causing the biaxial stress responsible for the observed strain. However, quantitative estimation of t he s tress re quires more i nformation. Ther efore, the development of biaxial stress can be expected, in general, for any  lm t hicker t han 1 0 nm. The i mportant c onsequence a nd necessity o f t he s train a nalysis i n u ltrathin Pb I2  lms i s h ighlighted and demonstrated [49], especially while determining the quantum connement contribution to the change in optical features. Results could lead to w rong conclusions if one ignores or overlooks t he e ffect of strain on t he opt ical properties. Re sults for t hicker Pb I2  lms a re d isplayed i n F igure 1 .3 f or v arious experimental parameters. The data shown are for lms grown at RT and LNT. Interestingly, the behavior of strain is exactly opposite to e ach ot her i n t he t wo c ases, most l ikely due to t he different thermal stress contribution to the total stress. However, the variation of strain w ith  lm thickness, annealing temperature, a nd t ime a s s hown i n F igure 1 .3a, b , a nd c , re spectively, may not be explained in a simple manner. Again, the qualitative explanations may have to b e der ived f rom t he coalescence a nd grain boundary models with required modications and inputs.

43722_C001.indd 5

For i nstance, t he t hickness de pendence o f s train re quires t he knowledge of g rain si ze, g rain den sity, a nd de fect den sity, e tc. for understanding the grain boundary structure. Switching from one type of biaxial stress to t he other with  lm thickness could possibly b e a n e xplanation f rom t heir m icrostructure. That is, different sx and sy could lead to this kind of strain variation with  lm thickness. For example, strain is negligible for a 63 nm thick  lm, which indicates very small biaxial stress, possibly resulting from the opposite type of the two components. Similarly, it may be p ossible to ac count f or t he ob served s train b y t he h elp o f microstructural analyses. Annealing is normally carried out to facilitate t he rel axation o f t he re sidual s tress. H owever, i t m ay not ne cessarily re duce t he s tress b ecause t he c hanging g rain boundary, microstructure, defect density, etc. could increase the stress as well in some cases. Figure 1.3b displays the effect of annealing temperature on the strain. The wavy nature is difficult to explain just like the one with annealing duration as shown in Figure 1.3c. Nevertheless, data need to be analyzed more carefully u sing ot her p ossible m acrostructural de tails ob tained o n the s ame s ample f or a b etter u nderstanding o f suc h b ehavior. The above discussion illustrates the t ypical intricacies inherent in the analyses of the residual stress origin. Such a detailed study is difficult to nd in the literature. The s tress b ehavior diff ers f or d ifferent m aterials. A nother material, t he z inc io dide h as a d ifferent st rain de pendence o n  lm thickness as shown in Figure 1.4. The strain in d104 is quite large c ompared to f ew ot her p eaks o f c onsiderable i ntensity using XRD. ZnI2  lms are also not well oriented. Ther efore, the average of strains of a few intense peaks is determined and plotted along with that of (104) in the gure. Although the comparative qualitative behavior of the two materials with lm thickness is not very much different, t heir e ffect o n opt ical p roperties i s entirely different [50], which we will see little later. The increasing and s aturating s train w ith t he  lm t hickness c ould si mply b e

10/3/2008 8:42:17 PM

1-6

Smart Materials 0.006 (00l)-Oriented PbI2 films

Grown at LNT

0.004 0.002

∆d/d

0 −0.002 −0.004 −0.006 −0.008

Grown at RT 0

100

200

300

400

Film thickness (nm)

(a)

93 nm Thick Pbl2 film annealed in vacuum for 2h

0.006 0.004

Grown at RT

∆d/d

0.002 0 −0.002 −0.004 −0.006 Grown at LNT −0.008 20

70

120 170 220 Annealing temperature (⬚C)

(b)

270

320

0.008 174 nm Thick Pbl2 film annealed at 200⬚C in vacuum

0.006

Grown at LNT

0.004

∆d/d

0.002 0 −0.002 −0.004 Grown at RT

−0.006 −0.008 0 (c)

1

2 3 Annealing time (h)

4

FIGURE 1.3 Residual strain in thicker PbI2  lms grown at R T and LNT as the function of (a)  lm thickness, (b) annealing temperature, and (c) annealing time.

43722_C001.indd 6

10/3/2008 8:42:17 PM

1-7


Znl2 films

0.018 Znl2 films

0.028

∆d/d

∆d/d

0.022

0.014

0.024

(Average) (104)

0.02 0

100

200

Treatment temperature (⬚C)

0.01 200

300

400 500 Film thickness (nm)

600

700

FIGURE 1.4 Average strain (average of fe w i ntense peaks of X RD) a nd strain i n d104 a s t he f unction of t hickness for Z nI 2 lms. The inset shows the dependence of these quantities on annealing temperature (annealed in vacuum for 1 h).

understood to mimic the grain density behavior with lm thickness [51,52]. The inset of Figure 1.4 shows the effect of vacuum annealing or heat treatment temperatures on the strain. Strain is found to i ncrease w ith t reatment i n b oth t he qu antities, h owever, more dramatically for strain in (104). This increase could be due to t he ad sorption o f re sidual ga ses p resent i n t he v acuum chamber maintained in the range 10−4–10−5 Torr during annealing. The uctuation i n t he v acuum le vel i s d ue to de gassing caused by the prolonged heating in the chamber. Another different type of strain dependence on  lm t hickness i s s hown i n Figure 1 .5 f or a ( 00l)-oriented C dI2  lm as w ell as (102) a nd

(002)-oriented HgI2 lms. The t ypical t hickness dependence of strain in CdI 2 i s not su rprising a nd c an b e e xplained b y t he increasing grain density with  lm thickness. Even the decreasing strain in CdI2  lms with annealing temperature as shown in the inset is very much like that one can ideally expect [53]. This opposite behavior compared to that of ZnI2 in the inset of Figure 1.4 is mainly responsible for t heir basic microstructural d ifference, t hat i s, C dI2  lms a re c ompletely oriented a s opp osed to randomly or iented Z nI2  lms. H owever, t he in creasing s train with substrate temperature as depicted in the inset is hard to explain by thermal stress alone because the substrate temperature

0.0035 Cdl2 films

∆d/d

0.002

Substrate

0.001

Hgl2 films (002)

Annealing

∆d/d

0 100

0

0.0025

200

Temperature (⬚C)

Cdl2 films Hgl2 films (102)

0.0015 0

500

1000 1500 Film thickness (nm)

2000

2500

FIGURE 1.5 Film thickness dependence of strain for (00l)-oriented CdI2  lms and (102), (002)-oriented HgI2 lms. The inset shows the dependence of strain on annealing and substrate temperatures for (00l)-oriented CdI2 lms.

43722_C001.indd 7

10/3/2008 8:42:17 PM

1-8

Smart Materials 0.0025

0.03 Znl2 films

∆d/d

∆d/d

(104) 0.0015

0.02

Average 0.01 50

100 150 200 Grain size (nm)

0.0005 0

4

8

12

Grain size (µm)

FIGURE 1.6 Dependence of strain on grain size for (00l)-oriented CdI2 lms. The inset shows the average strain and strain in d104 as a function of grain size for ZnI2 lms.

is known to affect the grain boundary structure as well. Another example i n t he  gure, HgI2, h as a si mple l inear s train de pendence on lm thickness, although it differs substantially both in magnitude a nd slope for t he t wo d ifferent g rowth or ientations [54]. It will still be difficult to speculate or predict the origin and behavior of the stress with certainty in either completely or randomly o riented  lms o f a ny m aterial, e specially, t he d ielectric and compound semiconductors etc. Nevertheless, the strain data will still be very useful in understanding its effect on the various physical properties of t he m aterial. The e xplicit d ependence of strain on the grain size is illustrated for CdI2  lms in Figure 1.6 and for Z nI2  lms i n t he i nset of t he  gure. The strain dependence of C dI2 on t he g rain si ze s eems to b e qu alitatively ide al and as expected according to the model of increasing grain density, especially in oriented lms [53]. Normally, in such cases, the grain s ize v aries lin early w ith  lm t hickness a nd t herefore a similar strain dependence on  lm t hickness c an b e e xpected. Actually, t his t rend i s i ndicated i n t he l imited re gion o f  lm thickness o f C dI2 in Figure 1.5. However, the grain size data shown i n F igure 1 .6 i ncludes t he re sults f rom  lm thickness, annealing, a nd sub strate h eating e xperiments. The mo re p uzzling and strange result is the strain dependence of ZnI2 shown in the inset where data for both average and the (104) plane are shown a s d iscussed a bove. A ny re asonable e xplanation o f t his trend is difficult. The unexpected behavior could well originate from the random orientation of grains and the minimum may be occurring f or opt imal c lose-packing g rain si ze. I t i s o nly o ne more e xample i llustrating t he v ariation i n d iversied strain behavior. A s we have s een s o f ar f rom t hese l imited e xamples, the origin and behavior of stress is fairly simple in some materials a nd a re e qually d ifficult a nd u npredictable f or s ome ot her materials.

43722_C001.indd 8

We will see how much essential and useful the stress analysis is in t hin  lm s tudy f or u nderstanding t he a ffected physical properties. The a bove e xamples o f s train f or v arious m aterials will be extended to i nterpret the optical properties of a re spective material. Analysis becomes fairly simple for oriented  lms like P bI2. I n c-axis o riented  lms, t he s train a long t he z-axis resulting from the xy-plane biaxial stress can be treated as equivalent to the effect of uniaxial stress along the z-axis. This would certainly aff ect t he ba nd s tructure to a me asurable e xtent i n optical properties, at least in layer structured materials. Such an analysis is illustrated in Figure 1.7 for PbI2. The various known excitonic peaks of PbI2 observed in the optical absorption spectra a re l abeled i n t he i nset of t he  gure for c onvenience. B oth XRD a nd opt ical a bsorption me asurements w ere do ne o n t he same sample for t he reliability of t he analyses. Excitonic peaks show a reasonably good and consistent linear dependence on strain, either positive or negative. Slightly different slopes of the peaks can be attributed to the extent of distortion of the atoms in the u nit c ell, w hich a re re sponsible f or t he o rigin o f t he c oncerned p eak. The f urther de tails o f t he a nalysis i n v iew o f t he band structure of PbI2 can be found in the recent report [55]. The important p oint h ere i s t he s train i nduced energ y s hift o f the optical peaks and the magnitude of the shiĀ compares with that due to quantum connement. Therefore, the true picture of quantum connement effect in PbI2 is possible only if the strain contribution is separated [55]. Many such minute and important details could be obtained through a very careful stress analyses. Similar analyses for oriented CdI2 and HgI2 lms are shown in Figure 1.8. B oth t hese c ompounds d id not s how a ny e xcitonic peaks at R T a nd t herefore, t he opt ical ba nd gap Eg wa s de termined u sing s tandard me thod f rom t heir a bsorption s pectra [53,54]. The s train de pendence of Eg for C dI2 i s not l inear a nd

10/3/2008 8:42:18 PM

1-9


Peak C

3

Pbl2 films Peak C

2.2

Peak B Peak B

Absorption

Exciton peaks’ energy (eV)

3.3

1.7 Peak A

1.2

2.7

93nm 133nm 174nm Film thickness

0.7 330

2.4 −0.016

430

Wavelength λ (nm)

−0.012

530

−0.008

Peak A

−0.004 ∆d/d

0

0.004

0.008

FIGURE 1.7 Strain dependence of three excitonic peaks of PbI2 lms. The three excitonic peaks are identied on the optical absorption spectra shown in the inset. The inset also depicts the shiĀ in peaks for few  lm thicknesses due to strain.

unusual as can be seen from the gure. It suggests an important fact that there exists a threshold strain up to which Eg is slightly affected a nd de creases s harply b eyond t he t hreshold [56]. This threshold lies on the side of tensile strain and it increases with a small slope on the compressive side [57]. The compressive strain was produced in the sample by energetic ion irradiation [57]. The effect o f s train o n (102) a nd (002) o riented Hg I2  lms i s qu ite opposite both in the magnitude and the slope [54]. A f ew more examples of oriented growths are illustrated by XRD in Figure 1.9 for BaI2 a nd SrI2  lms. As mentioned earlier, t he presence of a very intense single peak indicates the (hkl)-orientation parallel

to the substrate plane. Furthermore, such different orientations exhibit a different texture or morphology, which can be observed by scanning electron microscopy (SEM). Thus, bo th Ba I2 an d SrI2  lms tend to grow in three different orientations under certain g rowth c onditions a s s hown i n F igure 1 .9. F or t hese t wo materials too, the strain dependence of Eg of respective orientation differs signicantly from each other as shown in Figure 1.10. Here also, Eg was determined from their absorption spectra and no e xcitonic p eak w as v isible at R T. These c ould b e i mportant results in understanding the band structure and probably can be veried by the density functional (DF) calculations. Finally, the

3.6 Cdl2 films

E g (eV)

3

Hgl2 films (002)

2.4 Hgl2 films (102) 1.8 0.0005

0.0015

∆d/d

0.0025

0.0035

FIGURE 1.8 Strain-dependent optical band gap Eg for (00l)-oriented CdI2  lms and (102), (002)-oriented HgI2 lms.

43722_C001.indd 9

10/3/2008 8:42:18 PM

1-10

Smart Materials

(hkl)-Oriented Bal2 films

Counts (arb. units)

(015)

(113)

(301)

5

15

25 2q (⬚)

(a)

35

45

(022)

Counts (arb. units)

(hkl)-Oriented Srl2 films

(202)

(210)

10 (b)

15

20

25

30

35

2q (⬚)

FIGURE 1.9 XRD scans depicting the various (hkl)-oriented lm growth for (a) BaI2 lms and (b) SrI2 lms. The growth orientations are labeled on each scan.

example of randomly oriented polycrystalline ZnI2  lms is summarized in F igure 1.11. The a bsorption s pectra o f t he v arious heat-treated samples are displayed in Figure 1.11a, which denotes various optical features. Peaks B and C are identied as excitonic peaks a nd t he de termination o f Eg i n suc h a c ase i s de scribed

[50]. All these three quantities are to scale with the strain in d104 as shown in Figure 1.11b and do not show any systematic dependence on the average strain. In contrast, peak A does the opposite as shown in Figure 1.11c. This is the reason why the strain in (104) and the average strain are separately determined as shown

(hkl )-Oriented Bal2 films

(015) 3.5

Eg (eV)

(113)

(301)

3.3

(301) 3.1 (015) (113) 2.9 −0.008 (a)

FIGURE 1.10

43722_C001.indd 10

−0.006

−0.004

−0.002 ∆dhkl /dhkl

0

0.002

Strain dependence of optical band gap Eg for various (hkl)-orientations in (a) BaI2 lms

10/3/2008 8:42:18 PM

1-11

Residual Stress in Thin Films 3.8 (hkl)-Oriented Srl2 films

3.6

Eg (eV)

(202) 3.4 (210) (022) (210)

3.2

(022) (202) 3 −0.02

−0.015

−0.01

−0.005 ∆dhkl /dhkl

(b)

0.005

0.01

(b) SrI2  lms. The growth orientations are labeled in the gure.

Znl2 films B

A

C

250⬚C

a (cm−1)

150⬚C 200⬚C

Exc. peak B, exc. peak C & Eg (eV)

FIGURE 1.10 (continued)

0

4 Znl2 films 3.6 Exciton peak B Exciton peak C Eg 3.2

2.8 0.02

0.025 ∆d104 /d104

(b) 4.3

50⬚C

Eg RT

3.1

3.4 3.7 4.0 4.3 h ν (eV)

Peak A (eV)

100⬚C

2.5 2.8 (a)

0.03

4.2

4.1 0.01

(c)

0.014

0.018

0.022

∆d/d (average)

FIGURE 1.11 Summary of strain-dependent optical properties of ZnI2 lms: (a) Optical absorption spectra of 470 nm thick lm for various annealing (in vacuum) temperatures as labeled on e ach spectrum. The various optical features like excitonic peaks B a nd C, peak A, and the band gap Eg are identied on the spectra, (b) dependence of peaks B, C, and Eg on the strain in d104, and (c) the dependence of peak A on the average strain.

in Figure 1.4. Thus, the carefully carried out residual stress analysis a long w ith ot her c haracterizations c ould h elp i n b etter understanding the material. These f ew d iversied e xamples of opt ical prop erties a nalyzed in view of the stress demonstrate the importance of stress analyses in lms. In fact, a number of other physical properties also depend on stress and their stress dependence can reveal many microscopic details. For instance, the Hall resistance of

43722_C001.indd 11

semiconductors, critical temperature of superconducting lms, ferromagnetic re sonance i n f erromagnetic  lms, d ielectric constant o f d ielectric  lms, a nd t he l ikes c an b e e xpected to exhibit t he s train de pendence. C onversely, t he s hiĀ in t hese relevant quantities may be used to estimate the order of magnitude of the stress present in the  lm provided the exact stress dependence of the concerned quantity is known. The linear stress dependence of an easily and conveniently measurable

10/3/2008 8:42:19 PM

1-12

Smart Materials

physical quantity of any material can qualify it as a smart material for strain gauge applications. The main reason for the wide data range of various physical quantities reported in the literature for the lms of the same material could be due to ignoring the stress in the lm. One such example is the optical band gap of a p articular s emiconductor, which mostly d iffers f rom one report to t he ot her for t hick  lms a s much a s i n t he r ange of 1 eV where stress is normally ignored. Without the knowledge of t he in uence o f s tress, i t i s d ifficult to de termine t he t rue intrinsic value of t he ba nd gap w hich is normally used as t he reference f or ba nd s tructure c alculations. Si milar si tuations exist f or a f ew o f t he ot her ph ysical qu antities a s w ell a s f or which stress effect is quite signicant.

1.5

Summary and Future Direction

It i s a h ard t ruth t hat a ny t hin  lm i s not c ompletely f ree of residual s tress. This fact makes us learn to live with stress by understanding its nature and behavior. Therefore, we practically we c annot a fford to a void o r ig nore t he s tress c ompletely. A n appropriate, affordable, and convenient method of stress analysis can be adopted out of several available choices from t he literature. The i nformation a bout t he s tress i s a lready t here i n many o f t he ro utine s tructural c haracterization te chniques, which have to b e extracted by careful analyses as illustrated in this b rief ac count. I ndeed, a t ruly u niversal a nd qu antitative theory for the internal residual stress in  lms is perhaps yet to be developed if there could be any. The entire data in the literature and t heir d ifferent possible explanations make it doubtful that such a unique theory valid for different  lm materials and different methods of deposition will ever emerge. The main stumbling blocks f or suc h a n at tempt a re s ome o f t he i nherent p roblems and uncertainties of lms and their growth processes. Film stoichiometry ( or atom ic c omposition), s tructural d efects, g rainboundary st ructure,  lm–substrate i nterfacial p roperties e tc. are some of the difficulties in describing them in terms of macroscopic elastic models. Nevertheless, the study of residual stress in t hin  lms w ill c ontinue to b e i mportant a nd ne cessary f or exploiting them in various applications even if the origin cannot be traced convincingly. Sooner or later, the residual stress analyses in lms will draw more attention and be consolidated in nding further importance and priority in materials research.

References 1. E. J . Mills, On ele ctrostriction, Proc. Ro y. S oc. Lo ndon 26, 504–512, 1877. 2. R. W. Hoffman, The me chanical p roperties o f t hin  lms, in Ā e U se o f Ā in F ilms i n Ph ysical I nvestigations, J . C. Anderson (ed.), Academic Press, New York, p. 261, 1966. 3. R. W. Hoffman, M echanical p roperties o f t hin co ndensed lms, in Physics of Ā in Films, Vol. 3, G. Hass and R. E. Thun (eds.), Academic Press, New York, p. 211, 1966. 4. W. J . P arak, L. M anna, F . C. S immel, D . G erian, a nd P. Alivisatos, Quantum dots, in Nanoparticles: From Āeor y to

43722_C001.indd 12

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21.

22.

23.

Application, G. Schmid (ed.), Wiley-VCH Verlag, Weinheim, p. 27, 2004. R. Notzel, Self-organized growth of quantum-dot structures, Semicond. Sci. Technol. 11, 1365–1379, 1996. D. S. C ampbell, M echanical p roperties o f t hin  lms, in Handbook of Ā in Film Technology, L. I. Maissel and R. Glang (eds.), McGraw-Hill, New York, 1970. G. G. Stoney, The tension of metallic lms deposited by electrolysis, Proc. Roy. Soc. London A82, 172–175, 1909. W. Buckel, Internal stresses, J. Vac. Sci. Tech. 6, 606–609, 1969. E. Klokholm and B. S. Berry, Intrinsic stress in e vaporated metal lms, J. Electrochem. Soc. 115, 823–826, 1968. R. W. Hoffman, Stresses in thin lms: The relevance of grain boundaries a nd im purities, Ā in S olid Fi lms 34, 185–190, 1976. W. D . N ix a nd B . M. Clemen s, Cr ystallite coales cence: A mechanism for intrinsic tensile stresses in thin lms, J. Mater. Res. 14, 3467–3473, 1999. G. Guisbiers, O. Van Overschelde, and M. Wautelet, Nanoparticulate o rigin o f intrinsic r esidual st ress in t hin  lms, Acta Mater. 55, 3541–3546, 2007. E. Chason, B. W. Sheldon, L. B. Freund, J. A. Floro, and S. J. Hearne, Origin of compressive residual stress in p olycrystalline thin lms, Phys. Rev. Lett. 88, 156103(1–4), 2002. H. Windischmann, Intrinsic stress in sputter-deposited thin lms, Crit. Rev. Solid State Mater. Sci. 17, 547–596, 1992. Y. Pauleau, Generation and evolution of residual stresses in physical vapour-deposited thin lms, Vacuum 61, 175–181, 2001. C. A. Davis, A simple model for the formation of compressive st ress in t hin  lms b y io n b ombardment, Ā in Solid Films 226, 30–34, 1993. M. Itoh, M. Hori, and S. Nadahara, The origin of st ress in sputter-deposited tungsten lms for x-ray masks, J. Vac. Sci. Technol. B9, 149–153, 1991. J. -D . K amminga, T . H. de K eijser, R . D elhez, a nd E. J . Mittemeijer, A mo del f or st ress in thin la yers ind uced b y mistting particles: An origin for growth stress, Ā in Solid Films 317, 169–172, 1998. E. H. Hirsch, Stress in porous thin lms through adsorption of polar molecules, J. Phys. D13, 2081–2094, 1980. D. A. Glecker and S. I. Shah, Handbook of Ā in Film Process Technology, IOP, Bristol and Philadelphia, 1995. R. E. Cuthrell, D. M. Mattox, C. R. Peeples, P. L. Dreike, and K. P. Lamppa, Residual stress anisotropy, stress control, and resistivity in p ost ca thode ma gnetron sp utter dep osited molybdenum  lms, J. Vac. S ci. T echnol. A6, 2914–2920, 1988. C. H. Wu, W. H. Weber, T. J. Poter, and M. A. Tamor, Laser reective interferometry for in situ monitoring of diamond lm growth by chemical vapor deposition, J. Appl. Phys. 73, 2977–2982, 1993. M. Sternheim, W. van G elder, and A. W. Hartman, A las er interferometer system to mo nitor dry etching of patterned silicon, J. Electrochem. Soc. 130, 655–658, 1983.

10/3/2008 8:42:19 PM


24. J. K empf, M. N onnenmacher, a nd H. H. Wagner, Ele ctron and ion beam induced heating effects in solids measured by laser interferometry, Appl. Phys. A56, 385–390, 1993. 25. J. Kempf, Optical in situ sputter rate measurements during ion sputtering, Surf. Interf. Anal. 4, 116–119, 1982. 26. G. J. Leusink, T. G. M. Oosterlaken, G. C. A. M. Janssen, and S. Radelaar, In situ sensitive measurement of stress in thin lms, Rev. Sci. Instrum. 63, 3143–3146, 1992. 27. T. Aoki, Y. Nishikawa, and S. Kato, An improved optical lever technique f or me asuring  lm stress, Jpn. J. Appl. Phys. 28, 299–300, 1989. 28. S. Suresh and A. E. Giannakopoulos, A new method for estimating residual stresses by instrumented sharp indentation, Acta Mater. 46, 5755–5767, 1998. 29. M. Herrmann, N. Schwarzer, F. Richter, S. Frühauf, and S.E. Schulz, Determination of Young’s modulus and yield stress of p orous low-k materials by na noindentation, Surf. C oat. Technol. 201, 4305–4310, 2006. 30. R. Ikeda, T. Uchiyama, H. Cho, T. Ogawa, and M. Takemoto, An advanced met hod for me asuring t he residual st ress of deposited  lm u tilizing las er sp allation te chnique, Sci. Technol. Adv. Mater. 7, 90–96, 2006. 31. K. T. V. G rattan, B . T. M eggitt (e ds.), Optical F iber S ensor Technology, Chapman and Hall, London, 1995. 32. S. M. M. Quintero, W. G. Quirino, A. L. C. Triques, L. C. G. Valente, A. M. B. Braga, C. A. Achete, and M. Cremona, Thin lm st ress me asurement b y  ber o ptic stra in ga ge, Ā in Solid Films 494, 141–145, 2006. 33. A. Rizzo, L. C apodieci, D. Rizzo, and U. Galietti, L ow cost technique f or me asuring in si tu st rain o f na nostructures, Mater. Sci. Eng. C 25, 820–825, 2005. 34. J. K. Shin, C. S. Lee, K. R. Lee, and K. Y. Eun, Effect of residual stress o n the R aman-spectrum a nalysis o f t etrahedral amorphous carbon lms, Appl. Phys. Lett. 78, 631–633, 2001. 35. A. C. F errari a nd J . Rob ertson, I nterpretation o f R aman spectra of disordered and amorphous carbon, Phys. Rev. B61, 14095–14107, 2000. 36. J. Zh u, J . H an, A. Li u, S. M eng, a nd C. J iang, M echanical properties and Raman characterization of amorphous diamond lms as a function of lm thickness, Surf. Coat. Technol. 201, 6667–6669, 2007. 37. B. D . C ullity, Elements o f X-r ay Di ffraction, 2nd e dn., Addison-Wesley, Reading, MA, 1978. 38. H. P. Klug and L. E. Alexander, X-ray Diffraction Procedures for Polycrystalline and Amorphous Materials, 2nd edn., John Wiley, New York, 1974. 39. D. Shindo and T. Oikawa, Analytical Electron Microscopy for Materials Science, Springer-Verlag, Tokyo, 2002. 40. A. E. Ennos, S tresses de veloped in o ptical  lm coatings, Appl. Opt. 5, 51–61, 1966. 41. H. K. Pulker, Coatings on Glass, Elsevier, Amsterdam, 1984.

43722_C001.indd 13

1-13

42. J. Zhang, K. Xu, and V. Ji, Dependence of stresses on grain orientations in thin p olycrystalline  lms on substrates: an explanation of the r elationship between preferred orientations and stresses, Appl. Surf. Sci. 180, 1–5, 2001. 43. C. Kisielowski, J. Kruger, S. Ruvimov, T. Suski, J. W. Ager III, E. J ones, Z. Lilien tal-Weber, M. R ubin, E. R . Weber, M. D. Bremser, and R. F. Davis, Strain-related phenomena in GaN thin lms, Phys. Rev. B54, 17745–17753, 1996. 44. R. F. D avis, S. Einfeldt, E. A. Preble, A. M. Roskowski, Z. J. Reitmeier, a nd P. Q. Mira glia, Galli um ni tride a nd r elated materials: c hallenges in m aterials p rocessing, Acta M ater. 51, 5961–5979, 2003. 45. E. Eiper, A. Hofmann, J. W. Gerlach, B. Rauschenbach, and J. Keckes, Anisotropic intrinsic and extrinsic stresses in epitaxial wurtzitic GaN thin  lm on γ-LiAlO2 (1 0 0), J. Cryst. Growth 284, 561–566, 2005. 46. C. Sarioglu, The effect of anisotropy on residual stress values and mo dication o f S erruys a pproach to r esidual st ress calculations for coatings such as T iN, Z rN and HfN, Surf. Coat. Technol. 201, 707–717, 2006. 47. M. A. Moram, Z. H. Barber, C. J. Humphreys, T. B. Joyce, and P. R. Chalker, Young’s modulus, Poisson’s ratio, and residual stress a nd stra in in (111)-o riented s candium ni tride thin lms on silicon, J. Appl. Phys. 100, 023514(1–6), 2006. 48. W. J. Zhang and S. Matsumoto, Investigations of cr ystallinity a nd r esidual st ress o f c ubic b oron ni tride  lms by Raman spectroscopy, Phys. Rev. B63, 073201(1–4), 2001. 49. V. G ulia, A. G. Vedeshwar, a nd N. C. M ehra, Qua ntum dot-like b ehavior o f ul trathin PbI 2 lms, Acta M ater. 54, 3899–3905, 2006. 50. A. G. Vedeshwar and P. Tyagi, Excitonic absorption in ZnI2 lms, J. Appl. Phys. 100, 083522(1–6), 2006. 51. P. Tyagi a nd A. G. Vedeshwar, Op tical p roperties o f Z nI2 lms, Phys. Rev. B64, 245406(1–7), 2001. 52. P. T yagi a nd A. G. Vedeshwar, Resid ual st ress dep endent optical properties of ZnI2 lms, Phys. Stat. Sol. (a) 191, 633– 642, 2002. 53. P. Tyagi and A. G. Vedeshwar, Grain size dependent optical properties of CdI2 lms, Eur. Phys. J. AP19, 3–13, 2002. 54. P. Tyagi and A. G. Vedeshwar, Anisotropic optical band gap of (102) and (002) oriented lms of red HgI2, Phys. Rev. B63, 245315(1–6), 2001. 55. V. G ulia a nd A. G. Vedeshwar, Op tical p roperties o f PbI 2 lms: Quantum connement and residual stress effect, Phys. Rev. B75, 045409(1–6), 2007. 56. P. Tyagi and A. G. Vedeshwar, Effect of residual stress on the Optical properties of CdI2 lms, Phys. Rev. B66, 075422(1–8), 2002. 57. R. S. R awat, P. Arun, A. G. Vedeshwar, P. L ee, a nd S. L ee, Effect o f ener getic io n irradia tion o n C dI2 lms, J. Ap pl. Phys. 95, 7725–7730, 2004.

10/3/2008 8:42:19 PM

43722_C001.indd 14

10/3/2008 8:42:20 PM

2 Intelligent Synthesis of Smart Ceramic Materials 2.1 I ntroduction ............................................................................................................................... 2-1 2.2 H ydrothermal Synthesis of Smart Ceramic Materials—An Overview ............................... 2-1

2.3

Wojciech L. Suchanek Sawyer Technical Materials

Richard E. Riman Rutgers University

2.1

Process De nition • Merits of Hydrothermal Synthesis of Ceramics • Chemical Compositions and Morphologies of Smart Ceramics • Hydrothermal Hybrid Techniques • Industrial Production of Smart Ceramic Materials

Intelligent Control of Phase Assemblage ................................................................................2-4 Construction of a Ā ermodynamic Model • Methodology for Generating Stability and Yield Diagrams • Utilization of Ā er modynamic Modeling

2.4 Intelligent Control of Crystal Size and Morphology............................................................. 2-7 Ā ermodynamic Variables • Non thermodynamic Variables

2.5 S ummary .....................................................................................................................................2-8 Acknowledgments ...............................................................................................................................2-8 References .............................................................................................................................................2-8

Introduction

Ceramic materials are being used in nearly all advanced materials applications either as pure ceramics or as ceramic components of composites or devices w ith metallic or organic constituents. Growing dem and f rom v arious i ndustries to s ynthesize sm art materials w ith mo re a nd mo re s ophisticated f eatures re quires the use of smart starting materials, w ith very well-dened and controlled specic properties, such as size, shape (morphology), chemical composition and defect structure, surface functionalization, e tc. Ā is in turn requires use of advanced synthetic methods, which, in addition to producing superior product, should b e a lso i nexpensive a nd en vironmentally f riendly. Ā e low-temperature hydrothermal synthesis meets all of the above requirements; mo reover, i t i s v ery w ell su ited to p roduce b oth smart ceramic starting materials, such as powders or bers, and already shaped products, such a s bulk ceramic pieces,  lms or coatings a nd s ingle cr ystals [ 1–6]. Ā e n umber o f s cientic papers on hydrothermal synthesis of materials has been steadily increasing since 1989 a nd is c urrently on the le vel of hundreds per year [2,3]. Ā e number of papers on hydrothermal synthesis of ceramic powders alone has nearly quadrupled between 2000 and 2004. Consequently, a l arge family of smart ceramics, primarily powders and coatings, has emerged that can be prepared under very mild hydrothermal conditions (T < 200°C, P < 1.5 MPa). Ā is has generated lots of c ommercial interest in hydrothermal technology [1–6].

2.2 2.2.1

Hydrothermal Synthesis of Smart Ceramic Materials—An Overview Process Defi nition

Hydrothermal synthesis is a process that utilizes single or heterogeneous phase reactions in aqueous media at elevated temperature (T > 25°C) and pressure (P > 100 kPa) to crystallize ceramic materials directly from solution [1,6]. Reactants used in hydrothermal s ynthesis a re u sually c alled p recursors, w hich a re administered in forms of solutions, gels, or suspensions. However, hydrothermal growth of single crystals requires in most cases use o f s olid n utrient, w hich re crystallizes d uring t he g rowth process. Mineralizers are organic or inorganic additives that are used to c ontrol pH or en hance solubility. Other add itives, a lso organic or inorganic, are used to s erve other functions, such as controlling crystal morphology, chemical composition, particle dispersion, etc. Syntheses are usually conducted at autogeneous pressure, which corresponds to t he saturated vapor pressure of the solution at the specied temperature and composition of the hydrothermal s olution. H igher p ressures up to o ver 5 00 MPa and temperatures over 1000°C [5] may be necessary to facilitate reactant d issolution a nd g rowth o f c ertain t ypes o f c eramic materials, a nd a re u sually app lied i n si ngle c rystal g rowth. Nevertheless, mild conditions are preferred for commercial processes where temperatures are less than 350°C and pressures less than approximately 50 MPa. Intensive research has led to a better 2-1

43722_C002.indd 1

10/3/2008 8:42:53 PM

2-2

Smart Materials

understanding o f h ydrothermal c hemistry, w hich h as sig nicantly reduced the reaction time, temperature, and pressure for hydrothermal c rystallization of c eramic m aterials, pre dominantly p owders a nd c oatings (T < 200°C, P < 1.5 MPa) [2,4,5,7]. Ā is b reakthrough h as m ade h ydrothermal s ynthesis mo re economical since processes can be engineered using cost-effective and p roven p ressure re actor te chnology a nd me thodologies already established by the chemical process industry.

2.2.2

Merits of Hydrothermal Synthesis of Ceramics

Hydrothermal s ynthesis o ffers ma ny adv antages o ver co nventional a nd no nconventional c eramic s ynthetic me thods. A ll forms of ceramics can be prepared with hydrothermal synthesis,

(a)

(c)

(e)

5 cm

5 µm

1 µm

such as powders,  bers, and single crystals, monolithic ceramic bodies, and coatings on metals, polymers, and ceramics (Figure 2.1). From the standpoint of ceramic powder production, there are far fewer time- and energy-consuming processing steps since hightemperature c alcination, m ixing, a nd m illing s teps a re ei ther not necessary or minimized. Moreover, the ability to precipitate already crystallized powders directly from solution regulates the rate a nd u niformity o f n ucleation, g rowth, a nd a ging, w hich results i n i mproved control of si ze a nd morphology of crystallites and signicantly reduced aggregation levels, that is not possible with many other synthesis processes [8]. Figure 2.2 shows several e xamples o f t he v arieties o f mo rphologies a nd pa rticle sizes p ossible w ith h ydrothermal p rocessing. Ā e elimination or re duction of a ggregates combined w ith na rrow pa rticle si ze distributions i n t he s tarting p owders le ads to opt imized a nd

(b)

(d)

(f)

10 nm

3 µm

6 µm

FIGURE 2 .1 Examples of v arious for ms of c eramic m aterials s ynthesized h ydrothermally: (a) α-quartz si ngle c rystal, ( b) c arbon n anotube, (c) PZT powder, (d) carbon bers, (e) epitaxial KNbO3  lm on SrTiO3 wafer, and (f) epitaxial KNbO3  lm on LiTaO3 wafer.

43722_C002.indd 2

10/3/2008 8:42:54 PM

2-3

Intelligent Synthesis of Smart Ceramic Materials

(a)

(b)

300 nm

(c)

200 nm

(d)

2 µm

5 µm

(e)

(f)

10 µm

10 µm

FIGURE 2 .2 Examples of v arious si zes a nd mor phologies of c eramic p owders s ynthesized h ydrothermally: ( a) e quiaxed n anosized Z nO, (b) nanosized hydroxyapatite needles, (c) LiMnO3 platelets, (d) carbon nanotubes, 50–100 nm in diameter, several microns in length, (e) hydroxyapatite whiskers, and (f) equiaxed α-Al2O3.

reproducible properties of ceramics because of better microstructure control. From the standpoint of thin lms (coatings), other methods such as physical vapor deposition, chemical vapor deposition, and sol–gel suffer from the disadvantage that they all require h igh-temperature processing to c rystallize t he ceramic phase. Ā ermally induced defects result, such as cracking, peeling, undesired reactions between the substrate and coating, and decomposition of the substrate material. In contrast, hydrothermal synthesis can be used to directly crystallize  lms on to substrate s urfaces a t l ow t emperatures, t hereby e nabling n ew combinations of materials such as ceramic coatings on polymer substrates. Hydrothermal processing can take place in a w ide variety of combinations o f aque ous a nd s olvent m ixture-based s ystems. Relative t o s olid-state p rocesses, li quids a ccelerate diff usion,

43722_C002.indd 3

adsorption, re action r ate, a nd c rystallization, e specially u nder hydrothermal conditions [3,7]. However, unlike many advanced methods that can prepare a large variety of forms and chemical compounds, such as chemical vapor-based methods, the respective costs for instrumentation, energy, and precursors are far less for h ydrothermal me thods. H ydrothermal me thods a re mo re environmentally b enign t han m any ot her s ynthesis me thods, which c an b e at tributed i n pa rt to energ y-conserving lo wprocessing temperatures, a bility to re cycle w aste, a nd s afe a nd convenient disposal of waste that cannot be recycled [3]. Ā e low reaction tem peratures a lso a void ot her p roblems en countered with high-temperature processes, particularly during single crystal growth, for example, poor stoichiometry control due to volatilization o f c omponents o r c rystal c racking d ue to ph ase transformations taking place during cooling.

10/3/2008 8:42:55 PM

2-4

Smart Materials

An important advantage of the hydrothermal synthesis is that the purity of hydrothermally synthesized materials signicantly exceeds the purity of the starting materials. Ā is is bec ause hydrothermal crystallization is a purication process in itself, in which the growing crystals or crystallites reject impurities present i n t he g rowth en vironment. M aterials s ynthesized u nder hydrothermal c onditions o ften e xhibit diff erences in p oint defects when compared to materials prepared by high-temperature s ynthesis me thods. F or i nstance, i n ba rium t itanate, hydroxyapatite, o r α-quartz, w ater-related l attice de fects a re among t he mo st c ommon i mpurities a nd t heir c oncentration determines essential properties of these materials. Ā e problem of water incorporation can be overcome by properly adjusting the synthesis c onditions, u se o f w ater-blocking add itives, o r e ven nonaqueous solvents (solvothermal processing). Another important technological advantage of the hydrothermal technique is its capability for continuous materials production, which can be particularly useful in continuous fabrication of ceramic powders [9].

2.2.3

Chemical Compositions and Morphologies of Smart Ceramics

A g reat v ariety o f c eramic m aterials h ave b een s ynthesized b y hydrothermal methods. Most common are oxide materials, both simple o xides, suc h a s Z rO2, T iO2, SiO 2, Z nO, F e2O3, A l2O3, CeO2, SnO2, Sb2O5, Co3O4, HfO2, etc., and complex oxides, such as B aTiO3, SrTiO3, PZ T, PbTiO3, K NbO3, K TaO3, L iNbO3, f errites, apatites, t ungstates, v anadates, molybdates, zeolites, e tc., some o f w hich a re me tastable c ompounds, w hich c annot b e obtained using classical synthesis methods at high temperatures. Hydrothermal synthesis of a variety of oxide solid solutions and doped compositions is common. Ā e hydrothermal technique is also well suited for nonoxides, such as pure elements (for example Si, Ge, Te, Ni, diamond, carbon nanotubes), selenides (CdSe, HgSe, CoSe2, NiSe2, CsCuSe4), tellurides (CdTe, Bi2Te3, Cu xTey, AgxTey), suldes (CuS, ZnS, CdS, PbS, PbSnS3),  uorides, nitrides (cubic BN, hexagonal BN), aresenides (InAs, GaAs), etc. [2,4,10–12]. Crystalline products with a specic chemical or ph ase composition c an b e u sually s ynthesized h ydrothermally i n s everal different forms, such as single crystals, coatings, ceramic monoliths, or powders. Among them, the powders exhibit the largest variety of morphologies, suc h a s e quiaxed (for e xample c ubes, spherical), elo ngated ( bers, w hiskers, na norods, na notubes), platelets, nanoribbons, nanobelts, etc., with sizes ranging from a few nanometers to tens of microns (Figure 2.2). Core–shell particles and composite powders consisting of a mix of at least two different p owders c an b e a lso p repared i n o ne s ynthesis s tep. Some of the powders can even adopt nonequilibrium morphologies (Figure 2.4c and d).

2.2.4 Hydrothermal Hybrid Techniques In o rder to add itionally en hance t he re action k inetics o r t he ability to m ake ne w m aterials, a g reat a mount o f w ork h as been do ne to h ybridize t he h ydrothermal te chnique w ith

43722_C002.indd 4

microwaves ( microwave–hydrothermal p rocessing), e lectrochemistry (hydrothermal–electrochemical synthesis), ultrasound (hydrothermal–sonochemical s ynthesis), m echanochemistry (mechanochemical–hydrothermal synthesis), optical radiation (hydrothermal–photochemical s ynthesis), a nd h ot-pressing (hydrothermal hot pressing), as reviewed elsewhere [1,3,5–7].

2.2.5

Industrial Production of Smart Ceramic Materials

Several h ydrothermal te chnologies, p rimarily f or t he p roduction o f si ngle c rystals, suc h a s α-quartz f or f requency c ontrol and opt ical app lications ( Sawyer T echnical M aterials, T okyo Denpa, N DK), Z nO f or U V- a nd bl ue l ight-emitting de vices (Tokyo Denpa), and KTiOPO4 for nonlinear optical applications (Northrop Grumman-Synoptics), have a lready b een developed that demonstrate the commercial potential of the hydrothermal method. Ā e volume of the hydrothermal production of α-quartz single crystals is estimated at 3000 tons/year [2]. However, the largest potential g rowth a rea for commercialization is ceramic powder production. Ā e widely used Bayer process uses hydrothermal methods to dissolve bauxite and subsequently precipitate a luminum h ydroxide, w hich i s l ater he at-treated at h igh temperature to crystallize as α-alumina. In 1989, the worldwide production rate was about 43 million tons/year. Ā e production of pe rovskite-based d ielectrics a nd zi rconia-based st ructural ceramics is a promising growth area for hydrothermal methods [9]. C orporations suc h a s C abot C orporation, S akai C hemical Company, M urata I ndustries, Fe rro C orporation, S awyer Technical M aterials, a nd ot hers h ave e stablished c ommercial hydrothermal pr oduction pr ocesses for pre paring c eramic powders.

2.3

Intelligent Control of Phase Assemblage

Ā ermodynamic modeling can be used to i ntelligently design a process to b e t hermodynamically f avored u sing f undamental principles i nstead o f t he t ime-consuming E disonian me thods [1,13]. For a given hydrothermal synthesis of ceramic, the effects of precursor or additive concentrations, temperature, and pressure can be modeled to dene the processing variable space over which t he ph ases o f i nterest a re s table. Ā e thermodynamic modeling c an b e ac complished u sing c ommercially a vailable OLI computer soft ware, which also contains a database of thermodynamic properties for many common systems [1,13].

2.3.1

Construction of a Thermodynamic Model

In order to construct a thermodynamic model, one has to know all possible species a nd i ndependent chemical reactions occurring in the system being investigated. Even a simple hydrothermal system such as Ba–Ti–K–H2O, which is used to synthesize BaTiO3

10/3/2008 8:42:57 PM

2-5


ceramics, contains a signicant number of species and independent chemical reactions. Determining t he equilibrium concentrations f or e ach o f t he s pecies re quires t he u se o f a utomated computer-based s olution a lgorithms. St ability d iagrams c oncisely re present t he t hermodynamic s tate o f m ulticomponent,

log10 [m(Pb(acet)2) = 1.5m(NbCl5) = 3m(Mg(NO3)2)]

0

2

0

4

T = 200⬚C (473 K) 6 8

Pb3Nb4O13 stable

−0.8

12

0

−0.4

KNbO3 + Mg(OH)2 + Pb3Nb4O13

Mg(OH)2 stable Pb3Nb4O13 stable

KNbO3 stable Mg(OH)2 stable

All aqueous species

−1.2

10

Nb2O5 + Pb3Nb4O13

Nb2O5 −0.4

multiphase aqueous systems in wide ranges of temperature and reagent (precursor) concentrations (Figure 2.3). Yield d iagrams specify the synthesis conditions suitable for quantitative precipitation of the phase of interest (Figures 2.3b and 2.4a). Calculations of stability a nd y ield d iagrams a re ba sed on a t hermodynamic

−1.6

−2 14

−2 (a)

4

2

6 m (KOH)

8

10

T = 160⬚C (433 K)

log10 [m(TiO2) = m(Ba(OH2) = 10m (CO2)]

0

0

2

4

6

8

−1.6

−1.2

−1.6

Mg(OH)2

0

−0.8

10

12

Yield = 0.99 12

14 0

−1.6

BaCO3{s}

−3.2

−3.2 BaTiO3{s}

−4.8

−4.8

Aqueous species

−6.4

−8 (b)

−6.4

−8 0

2

4

6

8

10

12

14

pH

FIGURE 2.3 (a) Calculated stability diagram for t he Pb–Mg–Nb–K–H2O system at 2 00°C where input precursor concentration is plotted as a function of m ineralizer (KOH) c oncentration. Ā e s ymbols d enote e xperimentally o btained ph ase a ssemblages c orresponding to t he re action conditions specied by t he equilibrium diagram: Mg(OH)2 + Pb 3Nb4O13, KNbO3 + Mg(OH)2. (b) Calculated stability/yield diagram for t he Ba–Ti–CO2–H 2O system at 160°C using Ba(OH)2 and TiO2 (rutile) when the amount of CO2 in precursors is 0.1 times the amount of TiO2 and the Ba/Ti ratio is equal to 1. Ā e solid lines denote the incipient precipitation boundaries for BaTiO3 and BaCO3. Ā e shaded area corresponds to the BaTiO3 yield >99%.

43722_C002.indd 5

10/3/2008 8:42:57 PM

2-6

Smart Materials All aqueous species 0

PbO + PZT

12

Average particle size (µm)

PZT 70/30, yield > 99% -0.8 • log (Ti)

TiO2 precursor 6m KOH, 0.33m PZT, 150⬚C, 24h

11

PbO

−1.6

• 0% < Yield < 99%

−2.4

No PZT

10

ZTO precursor 1m TMAH, 0.11m PZT, 150⬚C, 24h

9 8 7 6 5 4 3 2 1

−3.2 (a)

0

0

2

4

6 8 10 [KOH] (mol/k g H2O)

(c)

12

14

0

(b)

5 µm

200

400

600

800

1000 1200 1400 1600 1800

Stirring speed (rpm)

(d)

5 µm

FIGURE 2.4 Example of intelligent control of chemical composition, size, and morphology of PZT crystallites using thermodynamic and nonthermodynamic variables. (a) Calculated stability/yield diagram in Pb–Ti–Zr–K–H 2O system at 150°C showing stability eld of the PZT phase with over 99% yield. (b) Control of crystallite size between 250 nm and 10 µm using simple agitation speed during hydrothermal synthesis and type of the precursor used. (c) and (d) Control of mor phology during hydrothermal synthesis at 1 50°C using concentration of t he TMAH mineralizer, which was (c) 0.5 m and (d) 1.0 m.

model t hat c ombines t he H elgeson–Kirkham–Flowers–Tanger (HKFT) equation of state for standard-state properties of aqueous species with a nonideal solution model based on the activity coefficient e xpressions de veloped b y Bro mley a nd P itzer, a nd modied b y Z emaitis e t a l. F or s olid s pecies, s tandard-state properties a re u sed i n c onjunction w ith ba sic t hermodynamic relationships. Fugacities of components in the gas phase are calculated f rom t he Re dlich–Kwong–Soave e quation o f s tate. A more detailed description of the thermodynamic model as well as citations t hat cover more of t he f undamentals c an be found elsewhere [1,13].

2.3.2 Methodology for Generating Stability and Yield Diagrams First, the desired product and components of the hydrothermal system have to be dened. Ā us, the identities of the precursors, mineralizers, a nd ot her add itives needed for t he s ynthesis of a

43722_C002.indd 6

required so lid p hase n eed t o be speci ed. Ā is i nformation is used a s i nput d ata, a long w ith t he r ange of re agent concentrations, temperature, and pressure specied by the user. Ā er efore, it is important that there is a data bank that is relevant for all the components in the system, which contains standard-state thermochemical p roperties a nd i ndependent re action s ets f or a ll species, HKFT equation of state parameters for aqueous species, and Redlich–Kwong–Soave equation of state parameters for gaseous species. Ā e OLI soft ware also stores binary parameters for ion–ion, ion–ne utral, a nd ne utral–neutral s pecies i nteractions. It should be noted, however, that the standard-state properties and parameters are frequently obtained by regressing numerous kinds o f t hermodynamic d ata. Ā ese d ata i nclude v apor p ressures, osmotic coefficients, activity coefficients, enthalpies, and heat capacities of solutions, and solubilities and heat capacities of solids. When data are not a vailable in the OLI data bank for the specied hydrothermal system, a private data bank must be constructed. Ā e literature can be consulted for data, as well as

10/3/2008 8:42:58 PM

2-7


several methods for estimating thermochemical data. Ā e chemistry model generation step creates the species and reactions that are possible with the given components of the system. All possible combinations of ions, neutral complexes, and solids are considered i n t his s tep. Ā e c hemical s peciation mo del i s a s et o f equations, wh ich co ntains ch emical eq uilibrium eq uations, phase equilibrium equations, mass balance, and electroneutrality equations. O nce t he t hermodynamic c onditions a re s pecied, the chemical speciation model is solved. Equilibrium concentrations of all species are calculated as a function of variables such as mineralizer concentration. Ā is gives the equilibrium composition of a specic set of reaction conditions, but to u nderstand the o verall b ehavior o f t he s ystem c omputations m ust b e performed over a wide range of conditions. For this purpose, we specify t he p rocessing v ariables o f i nterest. Ā e O LI s oftwa re offers a  exible choice of independent variables as x- and y-axes of s tability a nd y ield d iagrams, w hich i nclude p recursor a nd mineralizer concentrations (Figures 2.3a and 2.4a), solution pH (Figure 2.3b) and temperature, in addition to the electrochemical potential, t he latter being used for t he simulation of corrosion. In the case of y ield diagrams, the y ield value (e.g., 99%, 99.9%, and 99.95%) of the desired material must be chosen. Ā e upper temperature l imit o f t he O LI s oft ware for creation of both stability and yield diagrams is around 300°C, which covers mild hydrothermal synthesis conditions for most ceramic materials.

2.3.3 Utilization of Thermodynamic Modeling Calculated phase diagrams can perform many functions during the c ourse o f i ntelligent s ynthesis o f c eramics. I n add ition to precisely dening the concentration–temperature–pressure processing variable space over which the phases of interest are stable, many different types of precursor systems and additives can be compared, and experiments can be designed to make materials that have never been previously prepared in hydrothermal solution [1,13]. Moreover, t he t hermodynamic modeling enables to compute chemical supersaturation in a g iven system. Ā u s, one can evaluate how the processing variables on the phase diagram inuence supersaturation within a ph ase stability region. Once the sup ersaturation i s k nown, b y u sing c onventional c rystal growth models, one can identify when the process is dominated by nucleation and when by the growth rate, which can be utilized in morphology and size control, as demonstrated in the following section.

2.4

Intelligent Control of Crystal Size and Morphology

We will discuss this aspect of intelligent synthesis of ceramics using mo stly p owders a nd c oatings a s e xamples. Ā ese two forms of ceramics constitute the vast majority of ceramic materials synthesized under mild hydrothermal conditions and their size or morphology control is governed by principally the same

43722_C002.indd 7

rules. Control of single crystal growth has been reviewed elsewhere [2]. With t he t hermodynamic v ariables p rocessing s pace w ell de ned for t he phase of i nterest, a r ange of c onditions c an b e then e xplored to c ontrol re action a nd c rystallization k inetics for the purpose of developing a process suitable to produce the desired form of ceramics (e.g., particular size, morphology, aggregation level for powders, and microstructure, i.e., size and morphology of constituting crystallites for coatings). Ā er modynamic p rocessing v ariables, suc h a s tem perature, p H, c oncentrations of reactants, and additives, determine not only the processing s pace f or a g iven m aterial b ut a lso i nuence both reaction a nd c rystallization k inetics. Ā e p henomena t hat underlie the size and morphology or microstructure control using t he t hermodynamic v ariables a re t he overall nucleation and growth rates, which control crystal size and the competitive growth r ates a long p rincipal c rystallographic d irections t hat control morphology. Size of the crystals can be thus controlled by varying temperature and concentration. Crystal morphology and size can be additionally affected by surfactants, which can adsorb o n spe cic c rystallographic f aces a nd s olvents, w hich adsorb similarly as well as regulate solubility. Unfortunately, changing the thermodynamic synthesis variables is constrained by the phase boundaries in a specic phase diagram. Ā er efore, it may or may not be possible to exploit all sizes, morphologies, and m icrostructures for a c ertain m aterial by mo difying only the th ermodynamic v ariables. H owever, n onthermodynamic variables are also extremely important when operating in thermodynamically limited processing variable space. For instance, changing the stirring speed during synthesis of lms or powder can change the crystallite size by orders of magnitude [14].

2.4.1 Thermodynamic Variables Ā e technique of using thermodynamic variables, such as temperature, c oncentrations o f re actant, v arious add itives o r s olvents, is the most widely used method to control size and morphology of c rystals u nder h ydrothermal c onditions. I t i s because w ith i ncreasing tem perature a nd c oncentrations ( i.e., supersaturation), both nucleation and growth rates signicantly increase. It is widely believed that size of the crystals increases with increasing temperature or concentrations. However, this is true only when t he g rowth r ate dominates over t he nucleation rate. If t he nucleation rate dominates over t he growth rate, t he relationship will be opposite, i.e., size of the crystals will decrease with increasing temperature or concentrations, because there are so many particles formed rapidly and they have little time to grow. In extreme cases of very high nucleation rate, even amorphous materials c an b e s ynthesized. N ucleation- a nd g rowth-ratecontrolled s ynthesis r anges d iffer f or e very m aterial a nd h ave to be calculated or determined experimentally in each particular case. Changing re actants c oncentration c an b e u sed not o nly to control si ze o f t he c rystallites b ut a lso to c ontrol t heir s hape. Fibers or cubes of lead zirconate titanate (PZT, chemical formula

10/3/2008 8:42:59 PM

2-8

Pb(ZrxTi1−x)O3, where x = 0 – 1) can form during the hydrothermal s ynthesis de pending up on t he c oncentration o f te tramethylammonium hydroxide (TMAH) mineralizer (Figur e 2 .4c and d) [15]. Ā e morphology of crystals can be furthermore controlled by the use of appropriate chemical additives or solvents present in the crystallization environment. Ā is approach utilizes preferred adsorption of the additives or solvents on particular faces of the growing crystals, thus preventing growth in certain directions. Consequently, crystals of t he same ceramic phase w ith various levels of aspect ratios, from platelets through whiskers, can be synthesized. A good example of using this approach is hydroxyapatite [16,17]. If chelating agents (lactic acid, EDTA, etc.) are used to prevent precipitation of amorphous precursor at ambient temperatures, precipitation at ele vated temperatures u sually y ields larger crystals [16]. Use o f tem plates, suc h a s s elf-assembled o rganic mole cules, emulsions, or nanotubes in order to direct the growth of the ceramic ph ase re sults i n a v ariety o f m aterials a rchitectures including me soporous m aterials, na norods, a nd na nosized monodispersed powders [10]. Ā ese approaches, however, require removal of the template after the synthesis, which often involves high-temperature calcination. Monodisperse c eramic pa rticulate s ystems w ith m inimized aggregation le vels c an a lso b e s ynthesized f rom aque ous s olutions without templates by precipitating  rst a small number of nuclei and subsequently reducing supersaturation level in order to prevent further nucleation and enable diff usion growth of the formed nuclei [18]. While the powder uniformity is excellent, it requires working with dilute reactant concentrations (e.g., ≈0.005 M) so that both supersaturation and nucleation are highly uniform and colloidal stability can be maintained [12].

2.4.2 Nonthermodynamic Variables Nonthermodynamic s ynthesis v ariables, suc h a s s tirring r ate and me thod o f p recursor p reparation, c an b e e ssential i n si ze and morphology control of crystals, while keeping the thermodynamic variables constant. Generally speaking, stirring during crystal growth leads to an increase in the probability of spontaneous nucleation, a decrease in supersaturation inhomogenities, and an increase of the growth rate. Ā e nal size of the crystals is t hen de termined b y a ba lance b etween t he n ucleation a nd growth r ates. A go od e xample o f a v ery s trong i nuence of nonthermodynamic v ariables o n c rystal si ze a nd mo rphology is PZT [14]. Concentration of the chemicals used in hydrothermal PZT synthesis was well within the calculated stability eld of the PZT phase, which is shown in Figure 2.4a. Increasing the stirring speed alone from 200 to 1700 rpm has reduced the particle size of PZT crystals almost two orders of magnitude, from over 10 µm to a submicron range (Figure 2.4b). In this particular case, the nucleation rate seems to dominate over the growth rate, therefore t he c rystal si ze de creased w ith i ncreasing s tirring speed. A t h igher s tirring s peeds, t he PZ T pa rticles e xhibited

43722_C002.indd 8

Smart Materials

better-dened cube habit and were more uniform in size, which is in line with reduction of supersaturation inhomogenities with increasing a gitation r ate. Use o f d ifferent precursors i mpacted the crystal size as well (Figure 2.4b). Stirring speed or velocity of the solution ow a long t he substrate can affect signicantly the thickness, morphology, and chemical c omposition o f  lms g rown u nder h ydrothermal conditions, as demonstrated using BaTiO3, SrTiO3, and PZT thin  lms [14]. Film t hickness a nd morphology i s a lso sig nicantly impacted by t he t ype a nd c rystallographic or ientation of t he substrate used (Figure 2.1e and f).

2.5

Summary

Hydrothermal s ynthesis i s en vironmentally f riendly, energ yconserving low-temperature crystallization of all forms of smart materials directly from aqueous solutions. Hydrothermal crystallization a ffords e xcellent c ontrol o f c hemical c omposition (oxides, nonoxides, native elements, dopants, solid solutions), as well as shape, size, and morphology of ceramics. Understanding the ph ysicochemical p rocesses o ccurring i n t he aque ous s olution is the key for intelligent engineering of hydrothermal synthesis of smart ceramics. Ā ermodynamic modeling i s u sed to predict formation conditions for phases of interest. Subsequently, size, mo rphology, a nd a gglomeration le vel o f t he s ynthesized ceramics can be controlled in wide ranges using both thermodynamic and nonthermodynamic (kinetic) variables. Such intelligent app roach ena bles to ac hieve c ost e ffective sc ale-up a nd commercial p roduction o f sm art c eramics u sing t he h ydrothermal technology.

Acknowledgments Ā e authors w ish to g ratefully acknowledge t he supp ort of t he Office of Naval Research, Ā e National Institute of Health, OLI Systems, I nc., a nd S awyer Technical M aterials, L LC, f or t heir generous support of the research cited in this manuscript.

References 1. Riman, R .E., S uchanek, W.L., a nd L encka, M.M., Hydrothermal cr ystallization o f cera mics, Ann. Ch im. S ci. Mat., 27, 15, 2002. 2. Byrappa, K. and Yoshimura, M., Handbook of Hydrothermal Technology, N oyes Pub lications/William Andrew Publishing LLC, Norwich, NY, 2001. 3. Yoshimura, M., Suchanek, W.L., and Byrappa, K., Soft solution p rocessing: A stra tegy f or o ne-step p rocessing o f advanced inorganic materials, MRS Bull., 25, 17, 2000. 4. Sômiya, S., Ed., Hydrothermal Reactions for Materials Science and Engineering. An Overview of Research in Japan, Elsevier Science, London, 1989. 5. Roy, R . Accelerating the kinetics o f low-temperature inorganic syntheses, J. Solid State Chem., 111, 11, 1994.

10/3/2008 8:42:59 PM


6. Suchanek, W.L., L encka, M.M., a nd Rima n, R .E., H ydrothermal synthesis of ceramic materials, in Aqueous Systems at Elevated Temperatures and Pressures: Physical Chemistry in Water, Steam, and Hydrothermal Solutions, Palmer, D.A., Fernández-Prini, R ., a nd H arvey, A.H., E ds., Els evier, Amsterdam, 2004, Chapter 18. 7. Yoshimura, M. a nd S uchanek, W., I n si tu fa brication o f morphology-controlled advanced ceramic materials by soft solution processing, Solid State Ionics, 98, 197, 1997. 8. Riman, R .E., in High P erformance C eramics: S urface Chemistry in Processing Technology, Pugh, R. and Bergström, L., Eds., Marcel-Dekker, New York, 1993, p. 29. 9. Dawson, W.J., Hydrothermal synthesis of advanced ceramic powders, Ceram. Bull., 67, 1673, 1988. 10. Adair, J.H. and Suvaci, E., Morphological control of particles, Curr. Opin. Colloid Interf. Sci., 5, 160, 2000. 11. Niesen, T .P. a nd D eGuire, M.R ., Re view: D eposition o f ceramic t hin  lms a t lo w tem peratures f rom aq ueous solutions, J. Electroceramics, 6, 169, 2001.

43722_C002.indd 9

2-9

1 2. Sugimoto, T., Fine P articles, S ynthesis, Cha racterization, and Mechanisms of Growth, Marcel-Dekker, Inc., New York, 2000. 13. Lencka, M.M. a nd Rima n, R .E., in Encyclopedia of Smar t Materials, Vol. I, S chwartz, M., E d., John Wiley, New York, 2002, pp. 568–580. 14. Suchanek, W.L. et al ., H ydrothermal dep osition o f oriented epitaxial Pb(Z r,Ti)O3 lms under va rying hydrodynamic conditions, Crystal Growth Design, 5, 1715, 2005. 15. Cho, S.-B ., Ole dzka, M., a nd Rima n, R .E., H ydrothermal synthesis of acicular lead zirconate titanate (PZT), J. Crystal Growth, 226, 313, 2001. 16. Suchanek, W. et al., Biocompatible whiskers with controlled morphology and stoichiometry, J. Mater. Res., 10, 521, 1995. 17. Riman, R .E. et al ., S olution syn thesis o f h ydroxyapatite designer particulates, Solid State Ionics, 151, 393, 2002. 18. Matijevic, E., M onodispersed met al (h ydrous) o xides—a fascinating  eld of colloid science, Acc. Chem. Res., 14, 22, 1981.

10/3/2008 8:42:59 PM

43722_C002.indd 10

10/3/2008 8:42:59 PM

3 Functionally Graded Polymer Blend 3.1 I ntroduction ............................................................................................................................... 3-1 3.2 Mechanism of the Preparation Method ..................................................................................3-2 3.3

Diff usion–Dissolution Method • P olymerization–Diffusion Method

Preparation and Characterization of Several Types of Functionally Graded Polymer Blends ............................................................................................................3-5 Amorphous Polymer/Amorphous Polymer Miscible Blend (Dissolution–Diffusion Method) • Amorphous Polymer/Crystalline Polymer Miscible Blend (Dissolution–Diff usion Method) • Amorphous Polymer/Amorphous Polymer Immiscible Blend (Dissolution–Diff usion Method) • Amorphous Polymer/Crystalline Polymer Immiscible Blend (Polymerization–Diffusion Method)

3.4 Functional and Smart Performances and the Prospect for Application .............................3-8

Yasuyuki Agari Osaka Municipal Technical Research Institute

3.1

Functional and Smart Performances of PVC/PMA Graded Blend • Functional and Smart Performances of PEO/PLLA Graded Blend • Functional and Smart Performances of PEO (or PEO/LiOCl4)/PBMA Graded Blend

3.5 Prospects for Application in the Functionally Graded Blends........................................... 3-13 References ...........................................................................................................................................3-13

Introduction

Many reports have been published on functionally graded materials made of metals and ceramics [1]. Ā ese g raded m aterials have characteristics of improved strength against thermal stress as well as excellent electromagnetic and optical properties. Ā er e have also been many reports on a functionally graded ceramic, which can be called a sm art material. Ā ese ceramics showed a strong t hermoelectric per formance, wh ich sh ifted w ith a n increase in temperature. As a re sult, the thermoelectric performance was very high in a wide temperature gradient. Ā ere have also been some reports on functionally graded polymeric materials [2–43]. Ā ese functionally graded polymeric materials c an be classied i nto four t ypes f rom c urrently u sed materials, as shown in Table 3.1. Ā e graded structures may be classied into six types. However, reports on functionally graded polymer blen ds a re s omewhat l imited [ 4–7,14–28], a lthough there h ave b een p ublished re ports o n v arious t ypes o f blen ds. Ā e f unctionally g raded polymer blend has been considered to have a structure as shown in Figure 3.1. Ā e blend has two different surfaces without an interface and can have both the advantages of a laminate and a homogenous blend.

We have devised a new method called the dissolution–diffusion method f or p reparing f unctionally g raded p olymer blen ds [4–7,24–28]. In t his method, g raded polymer blends a re classied into two types. Graded polymer blends are usually prepared by t hree me thods: su rface i nclination i n t he mel t s tate [14,15], surface inclination in the solution method [16,17], and diffusion in t he melt me thod [19–23]. Ā e dissolution–diffusion method devised by us is only one of the methods that can be used in preparing b oth t ypes o f g raded p olymer blen ds. O ur me thod h as the following advantages in comparison with the other methods. Ā e preparation time in our method is very short, and primarily is 1 00 t imes s horter t han t hat o f d iff usion i n t he mel t s tate method. Ā e opt imum c ondition c an b e e asily de termined because our method has many controllable factors. Furthermore, the chemical decomposition of molecules does not occur in the preparation by our method because the preparation is performed at a lower temperature. Ā erefore, we believe that our method is the most useful method. In further work, we recently found a new preparation method (polymerization–diff usion me thod) b y t he p olymerization o f a monomer during diff usion using a m acroazoinitiator [29,30]. 3-1

43722_C003.indd 1

10/3/2008 8:43:36 PM

3-2

Smart Materials TABLE 3.1

Various Types of Functionally Graded Polymeric Materials

Types of Used Materials

Structure

Metal (or ceramic)/polymer

Composites

Polymer–polymer

Immiscible polymer blend

Miscible polymer blend Atom–atom (intramolecules)

Copolymer (random) Copolymer (tapered) Density of cross-linking

High-order structure (same polymer) Crystal structure

Content of polymer A (%)

Polymer A

Polymer B

100

100

100

0

0

0

Position of measuring point Laminate

FIGURE 3.1

Functionally graded blend

Homogeneous blend

Schematic model of functionally graded blend.

It i s u seful f or t he p reparation o f a n i mmiscible g raded blen d and can form a graded structure of similar molecular weight. Ā is chapter presents a detailed description of the preparation mechanism us ed in diss olution–diffusion a nd p olymerization– diffusion metho ds t o p repare f unctionally grade d p olymer blends. Ā e chapter also explains how to determine the optimum

43722_C003.indd 2

Preparation Method • Laminate method • Electric eld method • Centrifugation method • Flame spraying method • Surface inclination in melt state method • Surface inclination in solution method • Dissolution–diffusion method • Diffusion in melt method • Dissolution–diffusion method • Diffusion method of monomer in polymer gel • Living anion or radical polymerization method • Changing method of cross-linking concentration • Changing method of cross-linking temperature • Injection mold method

Size of Dispersion Phase Big

Molecular order Atom order

Same atoms and molecules

condition f or the s everal typ es o f f unctionally grade d p olymer blends (polyvinyl chloride [PVC]/polymethacrylate, polymethyl methacrylate [PMMA], p olyhexyl methacr ylate [PHMA]) o r polycaprolactone (PCL), polyethylene oxide (PEO)/poly(l-lac tic acid) (PLLA), bisphenol A type polycarbonate (PC)/polystyrene (PS), PEO/polybutyl methacrylate (PBMA), and (PEO/LiClO 4)/ PBMA us ed in c haracterizing grade d str uctures f or b lends b y measuring the Fourier transform infrared (FTIR) spectra, confocal Raman spectra, thermal behaviors around the glass transition temperature ( Tg) b y Diff erential S canning C alorimetry (DSC) methods, b y S canning Ele ctro-Microscopy–Energy Disp ersive X-ray S pectrometry (S EM-EDX) obs ervation, a nd b y n uclear magnetic resonance (NMR), or Gel Permeated Chromatography (GPC) measur ement. F urthermore, s everal typ es o f f unctional properties are discussed especially for smart performance, which were ca used b y the grade d str ucture. Finall y, the p rospects o f functionally g raded p olymer blends a nd t heir a pplications a re discussed.

3.2

Mechanism of the Preparation Method

3.2.1 Diffusion–Dissolution Method Ā e mechanism of formation of a graded structure by the diffusion–dissolution method is considered to be the following [27]. After a polymer B solution is poured onto a polymer A  lm in a glass petri dish, polymer A begins to dissolve and diff use in the solution on t he a ir side ( Figure 3.2), but t he d iff usion is i nterrupted when all the solvent evaporates. Ā us, a blend lm is produced, which consists of a concentration of gradient of polymer A/polymer B in the thickness direction.

10/3/2008 8:43:36 PM

3-3

Functionally Graded Polymer Blend

Evaporation Polymer B solution Dissolution and diffusion

Polymer A film

FIGURE 3.2

rate of p olymer A i n p olymer B s olution, a nd (c) i nterruption time for the diff usion due to the completion of solvent evaporation. F actors f or c ontrolling t he a bove ph enomena a re (1) t he type of s olvent, (2) t he c asting temperature, (3) t he mole cular weight of polymer A, and (4) the amount of polymer B solution. Until polymer A completely dissolves or reaches the surface of polymer B solution, i.e., in the formation of the  rst and second types of structure, the diff usion of polymer A i n the polymer B solution is considered to obey Fick’s second law (Equation 3.1) by the a ssumption of ne glecting t he e vaporation of t he s olvent i n polymer B solution during the diffusion:

Schematic model of dissolution–diffusion method.

Following t he s teps o f d issolution a nd d iff usion o f p olymer A, the graded structures should be classied i nto t hree t ypes (Figure 3.3). First type: Polymer A begins to dissolve and then diffuses, but does not yet reach the air side surface of polymer B solution. Ā e blend has three phases (polymer A, polymer B, and a thin graded structure). Second type: Just when all of polymer A has dissolved, the diffusion f rontier reaches up to t he air side su rface of polymer B s olution. Ā e blen d h as o ne g raded ph ase from t he su rface to t he ot her o ne, w hile t he su rfaces are composed of polymer A only or polymer B only. Ā ir d type: After the dissolution and diff usion of polymer A reached up to the air side surface of polymer B solution, polymer A and polymer B molecules began to mix with each other and became miscibilized. Ā e concentration gradient then began to disappear. Ā e formation of a concentration gradient should depend on (a) dissolution rate of polymer A in polymer B solution, (b) diffusion

Graded structure 1 Polymer B Narrow graded phase Polymer A Graded structure 2

Wide graded blend

⎛ ∂2C A ⎞ ∂C A (3 = DAB ⎜ ∂t ⎝ ∂x 2 ⎟⎠

.1)

where CA is the concentration of polymer A t is the passed time x is the distance from the surface of polymer A sheet DAB is an apparent diffusion coefficient Ā e point where CA approaches 1 and shifts to t he petri glass side, thus carrying forward the dissolution of polymer A. Ā us, by c onsidering t his e ffect a nd re arranging m athematically Equation 3.2 is obtained from Equation 3.1: ⎛ (x − b) ⎞ C A = erfc ⎜ ⎟ ⎝ 2 DAB t ⎠ ∞

⎛ 2 ⎞ erfc(x ) = ⎜ exp −x 2 dx ⎝ p ⎟⎠

∫

( )

(3.2)

x

where b is the distance between the petri glass side su rface and the ot her side o f rem ainder o f p olymer A , w hich h as not d issolved yet. Ā erefore, the gradient pro le in the blend at t can be estimated by Equation 3.2. Adaptability of Equation 3.2 t o t he e xperimental data w as examined in the PVC/PMMA graded blend, which is given a detailed e xplanation in S ection 3 .3.1. Ā e e xperimental d ata agreed approximately with the ones predicted by Equation 3.2, as shown in Figure 3.7. DAB and b were obtained as 6.38 mm 2/s and 57 mm, respectively. Ā e DAB value was much larger t han the v alue ob tained b y d iff usion i n mel t s tate me thod, w hich implies that the dissolution–diff usion method is very useful. Further, a thicker and a more excellent graded blend lm was prepared by the m ultiple-step metho d, as i llustrated in Figur e 3.4. In this case, the graded blend was obtained by repeatedly preparing and changing the composition of the blend in the pouring solution.

Graded structure 3

3.2.2 Gentle graded blend

FIGURE 3.3 Schematic models of various types of graded structures.

43722_C003.indd 3

Polymerization–Diffusion Method

The mechanism of formation of a graded structure by polymerization during the diff usion method is as f ollows [30]. After the polymer A lm containing a macroazoinitiator, i.e., a radical type initiator, wi th o ligomeric p olymer A s egment is cast o n a  lter, the

10/3/2008 8:43:37 PM

3-4

Smart Materials

Two-step method

One-step method Evaporation of solvent First step

Polymer B solution Dissolution–diffusion

Evaporation of solvent

Evaporation of solvent

Polymer A/Polymer B (5/5) solution Dissolution–diffusion

Polymer A/Polymer B (7/3) solution Dissolution–diffusion Polymer A film

Polymer A film

Polymer A film

Evaporation of solvent Polymer B solution

Second step

Four-step method

Dissolution–diffusion Film formed in the first step

Evaporation of solvent Polymer A/Polymer B (5/5) solution Dissolution–diffusion Film formed in the first step Evaporation of solvent Polymer A/Polymer B (3/7) solution Dissolution–diffusion

Third step

Film formed in the second step Evaporation of solvent Polymer B solution

Fourth step

Dissolution–diffusion Film formed in the third step

FIGURE 3.4

Schematic models of multiple steps method.

resulting laminate is p ut on a mo nomer B s olution in a al uminum petri dish kept at a constant temperature (Figure 3.5). Ā e monomer B begins to diffuse into polymer A lm to the air side with polymerization. When the process is interrupted on the way by enough diffusion of the monomer B, a blend lm is produced, which consists of a concentration gradien t o f p olymer A/polymer B in the thic kness direction.

Ā e graded structures could be determined by the bala nce of polymerization a nd diff usion o f mo nomer B . Ā e temperature content o f a macr oazoinitiator, time , i .e., la rgely eff ected the balance. Ā e temperature and the time enha nced both the p olymerization and the diffusion, and the content of the macroazoinitiator enhanced only the polymerization. Ease of evaporation of monomer B enhanced only the diffusion.

Diffusion of Monomer B

Evaporation of Monomer B

Polymer A film

Polymerization

Filter Monomer B Heating

FIGURE 3.5 Schematic model of polymerization–diffusion method.

43722_C003.indd 4

10/3/2008 8:43:38 PM

3-5


Preparation and Characterization of Several Types of Functionally Graded Polymer Blends

3.3.1

Amorphous Polymer/Amorphous Polymer Miscible Blend (Dissolution–Diffusion Method)

In the cas e of the PV C/PMMA system [6,7], s amples were prepared by changing the four controllable conditions: (1) the type of solvent, (2) the castin g temperature, (3) the mo lecular weight of the PVC, and (4) the amount of the PMMA solution. Ā e graded structures o f thes e s amples w ere c haracterized b y FTIR -ATR, Raman microscopic spectroscopy, and DSC methods. Here, FTIRATR was the spectroscopy, by which the IR spectra on the surface layer (a bout 1-10 mm o f thic kness) was me asured. Figur e 3.6 shows the graded structure of the samples in the direction of the thickness, measured by FTIR-ATR. In the similar blend with the graded structure 1, for a la minate, the PMMA co ntent increased at 60% of the distance/thickness and was conrmed to have a thin graded layer (about 10%–20% o f the dist ance/thickness). In the blend with the graded structure 3, the PMMA content was kept to about 50% in all dist ance ranges. However, in the b lend with the graded structure 2, the PMMA content gradually increased in the range f rom 0% t o 100% o f the dist ance/thickness. Ā us, i t was found that this blend had an excellent wide concentration gradient. In this case, the PMMA content was estimated from the ratios of the absorption band intensities at 1728 cm−1 (the stretching of the carbonyl group in PMA) and 615 cm−1 (the stretching of C–Cl bond in the PVC). Ā e change in the PMMA content in the thickness direction of the blended lm was estimated by measuring the FTIR-ATR spectra on a sliced layer of the blended lm. Ā e c hange i n t he PV C c ontent o f t he g raded blen d w as characterized b y t he c onfocal R aman s pectroscopy me thod, similar to t he F TIR-ATR me thod, a s s hown i n F igure 3.7. Ā e measurement of the Raman microscopic spectra was performed by measuring the Raman spectra at the focused point, which was

PMMA content (%)

100 Graded structure 1 Graded structure 2 Graded structure 3

80 60 40 20 0 0

20 40 60 80 100 (Distance from petri glass side)/(Sample thickness) (%)

FIGURE 3.6 Change i n PMMA c ontent i n t he t hickness d irection of several types of the PVC/PMMA graded blends.

43722_C003.indd 5

1.0 Focused area

0.9

FTIR method Raman method Predicted values

0.8

Change

0.7 PVC content

3.3

0.6 0.5 0.4 0.3 0.2 0.1 0.0

0

20 40 60 80 100 120 140 160 180 Distance from petri glass side (µm)

FIGURE 3.7 Change in the PVC content in the thickness direction of the PVC/PMMA blend.

shifted 1 0 b y 1 0 mm, f rom o ne su rface a rea to t he ot her o ne. It could be conrmed that the blend had a c omparatively thick layer of a g raded structure phase. Ā is method is considered to be signicantly useful because an easy and detailed estimation can be made for a graded pro le of a blend. Furthermore, t he g raded s tructure w as c haracterized b y the DSC method. The DSC curve of the blend having a widely graded s tructure ( graded s tructure 2 ), w hich s hows a mo re gradual s tep ar ound Tg t han t he ot hers, i s s hown i n F igure 3.8. Similarly, the structures of the samples, which were prepared in the several types of the conditions mentioned previously, w ere i nvestigated a nd i t w as f ound t hat t he t ypical optimum c ondition ( molecular w eight o f PV C: M n = 35,600, Mw = 60,400, t ype of s olvent: te trahydrofuran (T HF)/toluene (5/1), volume of solvent: 0.23 ml/cm 2 , tem perature: 3 33 K) could be obtained. In the case of PVC/PHMA system [7], the graded structure of the s ample c ould not b e e stimated b y t he F TIR-ATR a nd t he DSC me thods b ecause PHMA w as very s oft at ro om temperature. Ā us, the graded structure was measured by the SEM-EDX method (Figure 3.9). Ā e chlorine content in the sample increased gradually o n t he p etri g lass side a nd t hen i t w as c onrmed to have a wide graded structure. Finally, the structures of the samples, which were prepared in t he several t ypes of t he previously mentioned conditions, were investigated. The typical optimum condition (molecular weight of P VC: M n = 35,600, Mw = 60,400, t ype o f s olvent: MEK, volume of solvent: 0.37 ml/cm 2 , tem perature: 3 13 K) was obtained.

10/3/2008 8:43:39 PM

3-6

Smart Materials

PVC content (%)

Exotherm

Graded structure 1

Graded structure 2

Graded structure 3

100 J 90 E E 80 J 70 60 50 40 30 Solution volume 20 J 0.182 ml/cm2 10 E 0.364 ml/cm2 0 0 50 100

E J

E

EJ EJ E JE

E E E J

150

200

250

Distance from petri glass side (µm) 330

360 390 Temperature (K)

420

FIGURE 3.8 DSC curves of several types of PVC/PMMA blends.

FIGURE 3.10 Graded structures of the PVC/PCL graded blends measured by the FTIR-ATR method.

100 90 E 80 70 60 50 40 30 20 10 0 0

340 J

320 E

300

E J

E

E

280

J E

260

E

J Tg E PVC content

J

E J

240

E

50 100 150 200 Distance from petri glass side (µm)

J

Tg (K)

20 µm

PVC content (%)

Chlorine content

360

E E J

220 200 250

FIGURE 3.11 Graded structures of PVC/PCL graded blends characterized by the DSC method. Thickness direction

FIGURE 3 .9 Chlorine c ontent a long t he t hickness of PV C/PHMA graded blend (×750).

3.3.2

Amorphous Polymer/Crystalline Polymer Miscible Blend (Dissolution–Diffusion Method)

In the case of the PVC/PCL system [25], we obtained the optimum conditions when t he graded polymer blend was prepared w ith a wider compositional gradient, similar to the PVC/PMMA system. Figure 3.10 shows the PVC content of the samples for the direction of thickness, measured by the FTIR-ATR method. PVC decreased at about 70 mm from the petri glass side and decreased gradually until the surface of the air side, that is, about 240 mm apart from the petri glass side in both cases of the solution volume. Ā en, the change of Tg for the thickness direction of the blend  lm was characterized by the DSC method (Figure 3.11) in the case o f 0. 364 ml/cm 2 o f s olution v olume. Tg de creased w ith the increase of the distance from the petri glass side, similar to the PVC content. Ā us, the graded structure in the PVC content was conrmed by the graded prole in Tg.

43722_C003.indd 6

Furthermore, t he change of t he PCL crystalline content was evaluated from that of the amount of the heat diffusion effect on the PCL crystalline, measured by the DSC method. Ā e heat diffusion began to increase on the specimen after it w as i nitially kept at 0 a nd then about 130 mm of the distance to t he air side. Ā en, it i ncreased i mmediately a round a bout 180 mm. Ā u s, it was found that the graded structure in the PCL crystalline was formed i n a d istance o f 1 30–240 mm. Ā is m eans th at th e obtained graded PVC/PCL blend had both a gradient concentration of the PVC and a gradient content of the PCL crystalline, as shown in Figure 3.12. Ā e PCL content was about 30% for a distance o f 1 30 mm. Ā is result corresponded to that of the PCL crystalline i n a h omogeneous PVC/PCL blend, which emerged in the case of the larger than 30% of the PCL [44]. It was further considered that the amorphous phase was made up of a miscible amorphous PVC/amorphous PCL blend. Finally, the PCL crystalline phase decreased the closer it came to the surface of the air side again. Ā is phenomenon occurred because t he forming of the amorphous phase was more thermodynamically stable than that of the crystalline phase. Ā e graded structure of the PVC/PCL graded blend could be schematically illustrated in Figure 3.13.

10/3/2008 8:43:39 PM

3-7

160

100 E

140

90

Type of solution E PC -b-PS copol ymer/

80

J

J J

100

E J J

J PVC content E Heat of diffusion

80

J E

J

60

J J E

40 20 0

E

E

0

E

E

J J J

E

50 100 150 200 Distance from petri glass side (µm)

100 90 80 70 60 50 40 30 20 10 0 250

FIGURE 3 .12 Graded st ructure o f P VC/PCL g raded b lends ( PCL crystalline).

PS(1/ 9) blend PS onl y

70 PC content (%)

120

PVC content (%)

Heat of diffusion (J/g)


60 50 40

J

J J E J JJ E E

30 20

E

10 0

J

E J 0 20 40 60 80 100 120 140 160 180 Distance from petri glass side (µm)

FIGURE 3 .14 Graded s tructure of P C/PS g raded ble nd w ith or without the PS-b-PC block copolymer.

(1) Effect of macro phase separations PCL crystalline phase PCL

PVC

FIGURE 3 .13 Schematic mo del of t he PV C/PCL g raded ble nd (graded structure).

3.3.3 Amorphous Polymer/Amorphous Polymer Immiscible Blend (Dissolution–Diffusion Method) We attempted to p repare a g raded PC/PS blend by the dissolution–diff usion method [24], similar to the PVC/PMMA system. In this case, the PS solution was poured onto the PC lm. However, we d id not ob tain a g raded structure, but obtained a homogeneous t wo-layer system, which was composed of about 50% and 0%–10% of PC, as shown in Figure 3.14. Next, the macrophase separation was observed in the former layer. Ā is re sult w as t hought to o ccur b ecause o f t he f ollowing; only three factors, dissolution rate, diff usion rate, and evaporation time, could affect the process in forming a graded structure of miscible blend. However, in forming the graded structure of an immiscible blend, three additional factors also played a role: macrophase s eparation, su rface ro ughness, a nd g ravimetry, a s shown i n F igure 3 .15. E specially, t he m acrophase s eparation may increase t he size of t he separated phases up to t he size of the thickness in the prepared lm, resulting in a possible break in the formation of a strongly graded structure, when its concentration becomes higher by the evaporation of the solvent. Ā erefore, we attempted to prevent macrophase separation during graded structure formation by adding PS-b-PC block copolymer [45]

43722_C003.indd 7

(2) Effect of surface inclination

(3) Effect of gravitation

FIGURE 3 .15 Other f actors t hat h ave a n e ffect on for ming g raded structure in an immiscible blend.

to the PS solution (PS-b-PC block copolymer/PS = 1/9). Āe copolymer may act as a compatibilizer by decreasing the interface energy of the phases. In this case, the PC segment content in the block copolymer was 46% (NMR measur ement). It was f ound that the wide grade d structure in the ob tained PC/PS b lend was f ormed in the dist ance range of 0–100 µ m from the petri glass side (Figure 3.14). Furthermore, we attempted to prepare a graded PC/PS blend by pouring the PC solution containing the block copolymer onto the PS  lm. Figure 3.16 shows the change of the PC content for the direction of the  lm thickness. A wide graded structure was formed and conrmed not only, in the furtherest distance from, but also in the distance closest to the petri glass side. Ā is result meant that surface roughness signicantly inuenced the forming of the graded structure. Ā erefore, we were able to obtain the graded immiscible PC/PS blend by adding the PC-b-PS block copolymer. Āe graded structure a ssumed to b e f ormed f or t he P C/PS g raded blen d could be schematically illustrated in Figure 3.17.

10/3/2008 8:43:40 PM

3-8

Smart Materials 100 90 70 E

60 50

E E

40 30 20 E

E

10

E

E

E

0 0

20

40 60 80 100 120 140 160 180 200 Distance from petri glass side (µm)

FIGURE 3.18 PEO content in the thickness direction of PEO/PBMA graded blend measured by NMR and confocal Raman spectroscopy.

Distance in refraction (abs.)

FIGURE 3.16 Graded s tructure of P C/PS g raded ble nds i n t he c ase that the PC solution was poured on a PS  lm.

PEO content (%)

PC content (%)

80

100 NMR measurement 90 No.1 Confocal Raman spectroscopic measurement 80 70 No.2 No.5 No.6 60 No.4 No.7 50 No.3 40 30 20 10 0 0 200 400 600 800 1000 Distance from the surface of the air side (mm)

Peak 1 (Block coplymer) No. 7 No. 5 No. 3 No. 1 Retention time abs. Large

PC

PS

Peak 2 PEO or MAI

Long Molecular weight abs.

FIGURE 3.19 GPC charts in several types of the points, measured by NMR, in PEO/PBMA graded blend.

FIGURE 3.17 Schematic model of PC/PS graded blend.

3.4 3.3.4

Amorphous Polymer/Crystalline Polymer Immiscible Blend (Polymerization–Diffusion Method)

In t he c ase o f t he i mmiscible g raded blen d o f t he P BMA/PEO system [30], we obtained the optimum conditions in the preparation by the polymerization–diff usion method. Figure 3.18 shows the PEO c ontent of t he s amples for t he d irection of t hickness, measured b y t he c onfocal R aman s pectroscopy a nd t he N MR methods. Ā e d ata obtained by t he N MR me thod were a lmost the same as those by the Raman method. It was found that the NMR method was available and so it was used. Figure 3.19 shows the GPC data of the layers around the thickness points (No. 1, 3, 5, 7 in Figure 3.18) as measured by NMR method. Ā e peak corresponding to t he blo ck c opolymer ( EO-b-BMA c opolymer) became larger when the number of the points was larger. Ā en, the mole cular w eight a nd t he c ontent o f t he blo ck c opolymer became larger as the number of points became larger, that is, the BMA content was larger (Table 3.2). It was considered to o ccur because the polymerization rate was larger with a larger amount of the BMA monomer.

43722_C003.indd 8

Functional and Smart Performances and the Prospect for Application

3.4.1 Functional and Smart Performances of PVC/PMA Graded Blend Our s tudy [6,7] found t hat g raded p olymer blends h ad s everal types of functional properties, inclusive of smart performance. Ā us, the functional properties of the PVC/PMA blend containing graded structure 2 (an extremely wide graded concentration) were explained by comparing t hem w ith t hose of a blen d containing graded structure 1 (similar to a laminate system), a perfectly miscible blend (5/5), of PVC and PMA only. 3.4.1.1

Tensile Properties

Ā e re sults o f t he ten sile p roperties o f PVC, P MMA, t he p erfectly miscible blend (5/5), and the blends with graded structure 1 a nd 2 ( blend t ype 1 a nd 2 ) f or t he v ertical d irection o f t he thickness are summarized in Table 3.3. Ā e tensile strength of the homogeneously miscible blend was the highest, followed by blend type 2, surpassing PVC, PMMA, and the blend type 1. Ā is phenomenon means that the formation

10/3/2008 8:43:41 PM

3-9

Functionally Graded Polymer Blend TABLE 3.2 Molecular Weight and Composition of Several Types of Areas in the PEO/PBMA Functionally Graded Alloy in the Ā ic kness Direction Peak 1 No. 1 2 3 4 5 6 7

Peak 2

Mw

Mw/Mn

Mw

Mw/Mn

Ratio of Peak 1 to Peak 2

1.39 × 106 1.40 × 106 1.80 × 106 1.69 × 106 1.79 × 106 1.58 × 106 5.66 × 105

2.10 2.06 2.22 2.21 2.16 2.39 2.29

4.33 × 104 4.61 × 104 3.93 × 104 3.69 × 104 3.80 × 104 3.84 × 104 3.34 × 104

1.15 1.16 1.13 1.17 1.12 1.13 1.11

0.38 0.37 0.94 1.04 1.15 1.27 2.68

of a graded structure not only suppresses a break at the interface but also gives superior properties compared to the source materials. Ā is was considered to occur because the blend phase with the c oncentration g radient h ad a su fficiently high tensile strength. T he elo ngation at b reak o f blen d t ype 2 app eared sufficiently good. Ā e tensile modulus of t he blend t ype 2 w as higher than that of the blend type 1. It was found that the break in t he ten sile s tress c ould b e suppressed by t he formation of a concentration gradient. 3.4.1.2

Thermal Shock Resistance

Ā ermal s hock re sistance te sts were p erformed b y mo ving t he specimens f rom o ne b ox to a nother ( kept at 25 3 a nd 3 73 K) repeatedly (ve times) every 30 min. Ā e s pecimens w ere t hen evaluated f or t heir t hermal s hock re sistance b y me asuring a maximum angle of warp, as illustrated in Figure 3.20, and adhesive strength in shear by tension loading.

Ā ermal s hock re sistance te sts o f t he blen d t ype 2 w ith t he graded structure 2 w ere performed a nd t he re sults (maximum value of w arp a ngle a nd ad hesive s trength i n s hear by ten sion loading) were compared w ith t hose of t he blend t ype 1 h aving the graded structure 1, as shown in Table 3.3. Ā e  lm of the blend type 1 was highly warped, while that of the blend type 2 did not show any warp. Ā e adhesive strength in shear by tension loading of the blend type 2 was higher than that of the blend type 1. Ā e reasoning for the above properties was as follows: Ā e differences i n t he e xpansion o f t he PV C ( rubber state) and the PMMA (glass state) at high temperature (395 K) concentrated the warp stress at t he interface and decreased the strength o f t he i nterface. H owever, i n blen d t ype 2 , t he ph ase containing a n excellent w ide concentration g radient prevented the wa rp st ress f rom co ncentrating. Ā us, t he t hermal s hock resistance of the blend (blend type 2) with excellent wide concentration gradient was found to be superior to that of the similar blend (blend type 1) to a l aminate lm. Ā e formation of an excellent w ide c oncentration g radient w as f ound to h ave improved the strength of the interface. 3.4.1.3 Smart Performance (DMA Properties) Ā e c hange i n ten sile s torage mo dulus a nd t an d of t he PVC/ PMMA blend type 2 with a wide concentration gradient around Tg w as c ompared w ith t he p erfectly m iscible blen d (5/5) u sing the DMA me asurement (temperature i ncreasing r ate: 1 K/min, frequency: 0.2 Hz). Ā e Tg width measurements of storage modulus and half temperature of Tg width of tan d were estimated as shown in Table 3.3. Ā e h alf w idth o f t he tem perature o f t an d f or t he f ormer (16 K) was signicantly larger than that of the latter (10 K). Ā us, the blend type 2 was conrmed to be a continuous phase, having a wide range of Tg.

TABLE 3.3 Properties of PVC/PMMA Functionally Graded Blends PVC/PMMA Blend Unit

Type 2 a

(kgf/mm2) (%) (kgf/mm2)

6.4 4.5 200

4.5c 2.8c 190

7.2 5.2 220

(K) (K)

20 16

8.6, 11c —

(degree) (kgf)

9 98

170d 71d

Properties Tensile properties Tensile strength Elongation at break Tensile modulus of elasticity DMA properties (tensile mode) Tg width of storage modulus Half-temperature width Tg in tan d Ā ermal shock resistance Maximum warp angle Adhesive strength in shear by tension loading

Type 1

PMTb

PVC Only

PMMA Only

5.7 3.9 230

6.1 3.1 230

11 10

— —

— —

— —

— —

— —

Blend containing graded structure 2. Perfectly miscible blend. c Prepared by the hot press method. d Blend containing graded structure 1. a

b

43722_C003.indd 9

10/3/2008 8:43:42 PM

3-10

Smart Materials

Maximum angle

FIGURE 3.20 Measurement method of maximum angle of warp.

As a re sult, t an d o f t he g raded blen ds o f PVC a nd s everal types of polyalkyl methacrylate (PMA) containing graded structure 2 w ere me asured, a s s hown i n F igure 3 .21. T an d of t he graded PV C/PHMA blen d h ad t he w idest tem perature r ange. Ā erefore, it was concluded that the wide temperature range was caused by the greater differences of Tg in the pair polymers of the graded PVC/PHMA blend. Finally, we i nvestigated t he opt imum c onditions for preparing t he g raded PVC/PHMA blend w ith t he w ider temperature range of tan d. We obtained a PVC/PHMA blend containing an excellent graded structure 2, which showed a peak of tan d in a Solution volume 0.455 ml/cm2

much wider temperature range in comparison with those of the blend containing t he graded structure 1 a nd perfectly miscible blend (5/5), as shown in Figure 3.22. In both systems of the PVC/PMMA and the PVC/PHMA blends, we found that the tensile storage modulus of the blend type 2 contained an excellent graded structure 2 t hat began to de crease at a lower temperature t han t hat of t he perfectly m iscible blend (5/5) and did not have a terrace, while that of a similar blend containing the graded structure 1 in a laminate had some terraces. Sandwich steel beams combining a polymer have been used as damping materials [46]. It is known that the damping efficiency is maximum i n t he temperature range at w hich t he used polymer has a peak of tan d. Ā erefore, it is expected that an excellent graded blend with a peak of tan d in a much wider temperature range i s u seful a s a d amping m aterial i n a l arge tem perature range. Ā is is what the graded polymer blend is expected to be for smart materials. Ā e following principle reects the reasoning for this condition for smart materials. An excellent graded blend was used for the polymer combined with the steel plates, as shown on the right of Figure 3.23. Tg of the graded blend decreases when shift ing occurs from the left to the right side. At the highest temperature, that is, the same temperature as the higher Tg of the pair polymers in the blend, the area in the farthest left side shows the highest and best damping performance. As the area shifts to t he r ight side , t here i s a decrease of temperature. Finally, at the lowest temperature, that is, the same temperature as the lower Tg of the pair polymers in the blend, t he a rea on t he farthest right side s hows t he highest

DMA

PVC/PHMA

PVC/PMMA

Graded structure 1

Increase of tan d

PVC/PBMA

Increase of strage modulus

Increase of tan d

Increase of storage modulus

PVC/PMMA

Graded structure 2

PVC/PBMA

Graded structure 3

PVC/PHMA

230

270

310

350

390

Temperature (K)

FIGURE 3.21 DMA data of PVC/PMA graded blends.

43722_C003.indd 10

200

240

280

320

360

Temperature (K)

FIGURE 3.22 DMA data of PVC/PHMA graded blends.

10/3/2008 8:43:42 PM

3-11


Medium temp. Lower temp. Most superior area in damping property at each temperature

Most superior area in damping property

Weight loss of PLLA (wt%)

100

Steel plates

Temperature

Higher temp.

80 60 40 20 0 0

FIGURE 3.23 Schematic model of the so-called smart performance in the d amping prop erty of s teel pl ate c ombined w ith a f unctionally graded blend.

and b est e xcellent d amping p erformance. Ā erefore, t he a rea showing the high damping performance shifts with the changing temperature. Ā is performance is thus considered as one of the so-called “smart performances.”

3.4.2 Functional and Smart Performances of PEO/PLLA Graded Blend PLLA is known as biomass and biodegradable polymer and its biodegradability was enhanced by blending with PEO. Ā us, we prepared the PEO/PLL A graded blend, which was exp ected to have both hig h b iodegradability a nd t ensile str ength in the v ertical direction o f thic kness [26]. Ā e va rious typ es o f grade d PLL A/ PEO blends, homogeneous blend, and PLLA only were subjected to degrada tion b y a p roteinase K enzyme . Ā e biodegradability

2

4

6 8 Past time (day)

10

12

Graded blend type 1

Graded blend type 2

Graded blend type 3

PLLA only

Homogeneous blend

FIGURE 3.24 Changes of net weight loss of PLLA in various types of graded blends, homogeneous blend, and PLLA only, in the biodegradation test.

was evaluated by the net weight loss of PLLA calculated by difference of weightbefore and after testing. Ā e net weight loss in all of the g raded blends was hig her tha n thos e in the ho mogeneous blend a nd the PLL A o nly (Figur e 3.24). Ā e net w eight loss o f graded blend type 2 wi th the b est excellent graded structure was the highest. Ā us, the graded structure was co nsidered to la rgely increase biodegradability. It was conrmed by an SEM observation of thes e ma terials (Figur e 3.25). Ā e net w eight loss o f all the graded blends, w here a p orous str ucture formed b efore the b iodegradation test, did not result in increase to a great extent. It was

(a)

PLLA

(b)

Homogeneous blend

(c)

Graded blend I

(d)

Graded blend II

FIGURE 3.25 SEM photograph of the cross-section of the sample after biodegradation test for 11 days.

43722_C003.indd 11

10/3/2008 8:43:43 PM

3-12

Smart Materials TABLE 3.4 Tensile Strength of Several Types of Materials

Elastic modulus (mPa)

Materials

25

Tensile Strength (MPa)

Graded blend 1 Graded blend 2 Graded blend 3 Homogeneous blend PLLA only

2.46 3.05 3.20 1.78 2.88

15 10 5 0 0

considered as t he f ollowing: PEO, no t o nly diss olved in to wa ter resulting in increasing the surface area attacked by the enzyme, but also, absorbed the PLL A oligomers that acted as acid ca talysts of water decomposition and subsequently, it promoted the decomposition of PLLA. Furthermore, the strengths of all the graded blends were la rger tha n tha t o f the ho mogeneous b lend (T able 3.4). Ā erefore, it was concluded that the graded structure enhanced the biodegradability while maintaining its high tensile strength.

3.4.3 Functional and Smart Performances of PEO (or PEO/LiOCl4)/PBMA Graded Blend PEO/PBMA graded blend was prepared by the polymerization– diff usion method [30]. It was found that the water vapor permeability ac ross t he g raded blen d  lms w as d ifferent i n t he PEO Permeation direction in the graded blend PEO rich

Laminate (2 layers) Graded composite 1 Graded composite 2

20

10

20

30

40

50 60 D/d (%)

70

100

content on the side of the higher pressure of water vapor (humidity: 90%) at 313 K. Ā us, the permeability of the vapor from the PEO-rich side (direction 1) was always higher than that of the vapor f rom t he P BMA-rich side ( direction 2 ) ( Figure 3 .26). Ā erefore, i t w as f ound t hat t he g raded s tructure c ould c ause anisotropy in the permeability of the water vapor. (PEO/LiClO4)/PBMA g raded blen d w as p repared u sing EO-b-BMA blo ck c opolymer b y t he d iffusion–dissolution method [28]. Ā e change i n t he elastic modulus of t he g raded blend i n t he t hickness d irection w as e stimated a s s hown i n Figure 3.27. Ā e modulus of the graded blend decreased gradually around 45%, while that of the laminate began to decrease immediately (Fi gure 3. 28). Ā us, i t w as c oncluded t hat t he graded blend was prevented from its breakaway at the interface

PBMA rich

10−2 10−4

2

10−6 10−8

Test condition: Temp. 313 K, humidity 90%

Laminate (2 layers)

10−10 10−12 10−2

0.18

Direction 1

0.16

Direction 2

Electric conductivity (S/cm)

0.2

Permeability of water vapor (g/mh)

90

FIGURE 3.27 Tensile elastic modulus of several types of materials in the thickness direction (D/d: (Distance from petri glass side)/(thickness of sample) ).

1

0.14 0.12 0.1

10−4 10−6 10−8 Graded blend 1

10−10 10−12 10−2 10−4

0.08 10−6

0.06

10−8

0.04

0

Graded blend 2

10−10

0.02

10−12 0

0

5

10

15

20

25

30

Past time (h)

FIGURE 3 .26 Permeability of w ater v apor of P EO/PBMA g raded blend in both types of the thickness directions.

43722_C003.indd 12

80

20

40

60

80

100

D/d (%)

FIGURE 3 .28 Electric c onductivity of s everal t ypes of m aterials i n the thickness direction (D/d: (Distance from petri glass side)/(thickness of sample) ).

10/3/2008 8:43:44 PM

3-13

Functionally Graded Polymer Blend TABLE 3.5 Possibility of Applications of Functionally Graded Polymer Blend Expected Functional Property Relaxation of thermal stress

Prevention of vibration and sound Electromagnetic materials Photomaterials Medical materials Packing materials Chemical

Application Mechanical device for antiabrasion Sporting goods Construction materials Vibration and sound proof Electromagnetic shield Copy machine device Optic ber Lens Articial internal organs Articial blood vessels and organs Waterproof adhesive Chemical resistance materials

by its graded structure. However, all of the electric conductivities o f t he g raded blen ds a nd t he l aminate b egan to si milarly increase immediately around 40%.

3.5

Prospects for Application in the Functionally Graded Blends

Functionally graded polymer blends are expected to be used in the future as a substitute for laminates because of their superiority in their strength and thermal shock resistance. Ā e superiority, we believe, is caused by the lack of interface, which suppresses the break at t he interface and the thermal stress. Furthermore, the excellent wide compositional gradient for the graded structure results in improvement in several types of physical properties. Ā erefore, some new functional performances are expected, which could be caused by an improved physical property gradients, a nd c an b e app lied i n v arious t ypes o f app lications, a s shown in Table 3.5.

References 1. Society o f Functionally G raded Material, e d., Functionally Graded Materials, Kogyo Chyosakai, Tokyo, 1993. 2. T. Kitano, Kogyo Zairyo, 43(6),112, 1994. 3. M. T akayanagi, Ā e 23r d C olloquium o f S tructure a nd Property of Polymer, Tokyo, 1993. 4. Y. Agari, Func Grad.. Mater., 16(4), 32, 1996. 5. Y. Agari, Koubunshi Kako, 46(6), 251, 1997. 6. Y. Agari, M. S himada, A. U eda, a nd S. N agai, Macromol. Chem. Phys., 2017, 1996. 7. Y. Agari, M. S himada, A. U eda, T. Anan, R . N omura, a nd Y. Kawasaki, Func. Grad. Mater., 1996, 761, 1997. 8. Y. Agari, M. Shimada, M.Ueda, R. Nomura, and Y. Kawasaki, Polym. Preprints, 47, 701, 1998. 9. Y. Agari, M. S himada, H. S hirakawa, R . N omura, a nd Y. Kawasaki, Polym. Preprints, 48, 698, 1999.

43722_C003.indd 13

10. J. Z. Yu, C. L ei, a nd F. K. K o, Annu. Tech. C onf. S PE, 52, 2352, 1994. 11. M. Omori, A. Okubo, K. Gilhwan, and T. Hirai, Func. Grad. Mater., 1996, 764, 1997. 12. M. Funabashi, T. Kitano, Seni-Gakkaishi, 50(12), 573, 1994. 13. C. M. Ā ai, T . K ato, a nd A. Yoshizumu, J. Ā er mosetting Plastics, Japan, 16(3), 126, 1995. 14. X. M. Xie, M. Matsuoka, and K. Takemura, Polymer, 33(9), 1996, 1992. 15. S. Kanayama and T. Umemura, Seikei Kako, 2(4), 216, 1995 16. Y. Kano, S. Akiyama, H. Sano, and H. Yui, J. Electron Microsc., 44(5), 344, 1995. 17. S. Akiyama and Y. Kano, Kagaku To Kogyo, 71, 44, 1997. 18. S. M urayama, S. K uroda, a nd Z. O sawa, Polymer, 34(18), 3893, 1993. 19. E. J abbari a nd N. A. P eppas, Macromolecules, 26, 2175, 1993. 20. P. F. Nealey, R. E. Cohen, and S. Argon, Macromolecules, 27, 4193, 1994. 21. K. C. F arinas, L. D oh, S. Venkatraman, a nd R . O . P otts, Macromolecules, 27, 5220, 1994. 22. T. E. Shearmur, A. S. Clough, D. W. Drew, M. G. D. van der Grinten, and R. A. L. Jones, Macromolecules, 29, 7269, 1996. 23. M. A. Parker and D. Vesely, J. Polym. Sci., Part B Polym. Phys., 24, 1869, 1986. 24. Y. Agari, M. S himada, A. U eda, T. K oga, R . N omura, a nd Y. Kawasaki, Polym. Preprints Jpn., 45, 2241, 1996. 25. Y. Agari, M. S himada, A. U eda, T. K oga, R . N omura, a nd Y. Kawasaki, Polym. Preprints Jpn., 46, 657, 1997. 26. Y. Agari, Y. Kano, K. Sakai, and R. Nomura, Polym. Preprints, Jpn., 51, 1065, 2002. 27. Y. Agari, Y. Anan, R. Nomura, and Y. Kawasaki, Polymer, 48, 1139, 2006. 28. Y. Agari, T. Morita, M. Shimada, and R. Nomura, J. Jpn. Soc. Col. Mater., 75, 474, 2002. 29. Y. Agari, K.Ohishi , M. S himada, a nd R . N omura, Polym. Preprints Jpn., 51, 647, 2002. 30. Y. Agari, T. Yamamoto, a nd R . N omura, Kagaku t o Ko gyo (Japan), 80, 138, 2006. 31. M. Kryszewski and G. Czeremuszkin, Plaste u Kautchuk, 11, 605, 1980. 32. P. Milczarek and M. Kryszewski, Coll. Polym. Sci., 265, 481, 1987. 33. Y. K oike, H. H idaka, a nd Y. Oh tsuka, Appl. O pt., 22, 413, 1983. 34. Y. Koike, N. Tanio, E. Nihei, and Y. Ohtsuka, Polym. Eng. Sci., 29(17), 1200, 1989. 35. C. F. J asso a nd E. M endizabal, Annu. Tech. C onf. S PE, 50, 2352, 1992. 36. S. Ashai, Polym. Preprints, Jpn., 27, 18, 1978. 37. S. Ashai, Ā e 6th Symposium Functionally Graded Materials, 61, 1993. 38. S. Ashai, Kagaku To Kogyo, 71, 50, 1997. 39. Y. Tsukahara, N. N akamura, T. H ashimoto, a nd H. K awai, Polym. J., 12(12), 455, 1980.

10/3/2008 8:43:45 PM

3-14

40. D. G reszta, K. M atsuoka, a nd K. M atyaszewski, Polym. Reprints ACS, 37, 569, 1996. 41. M. Furukawa, T. Okazaki, and T. Yokoyama, Polym. Preprints, Jpn., 45, 2239, 1996. 42. G. B . P ark, M. H irata, Y. K agari, T. M atsunaga, J . G ong, Y. O sada, a nd D. C. L ee, Polym. Pr eprints, Jp n., 45, 1836, 1996.

43722_C003.indd 14

Smart Materials

43. Y. Ulcer, M. Cakmak, and C. M. Hsiung, J. Appl. Polym. Sci., 60(1), 125, 1996. 44. Y. Agari and A. Ueda, J. Polym. Sci., Part B Polym. Phys., 32, 59, 1994. 45. M. S himada, Y. Agari, a nd Y. M akimura, Polym. Pr eprints, Jpn., 45, 1958, 1996. 46. D. J. Mead, J. Sound Vib., 83, 363, 1982.

10/3/2008 8:43:45 PM

4 Structural Application of Smart Materials 4.1 I ntroduction ...............................................................................................................................4-1 4.2 M aterials and Application .........................................................................................................4-1 Materials

R. Sreekala Structural Engineering Research Centre, CSIR

K. Muthumani Structural Engineering Research Centre, CSIR

4.3 S tructural Uses ...........................................................................................................................4-2 Active Control of Structures • Passive Control of Structures • Hybrid Control • Smart Material Tag • R etrotting • S elf-Healing • S elf-Stressing for Active Control • S tructural Health Monitoring • Active Railway Track Support • Active Structural Control against Wind

4.4 C onclusion ..................................................................................................................................4-7 Acknowledgment .................................................................................................................................4-7 References .............................................................................................................................................4-7

4.1 Introduction Ā e development of durable and cost-effective high-performance construction m aterials a nd s ystems i s i mportant f or t he e conomic well-being of a c ountry mainly because t he cost of civil infrastructure constitutes a major portion of the national wealth. To add ress t he p roblems o f de teriorating c ivil i nfrastructure, research is very essential on smart materials. Ā e research a nd development pro jects a iming to appl y a dvanced t echnologies, such as new materials and new structural systems, can improve the performance of the buildings, reduce the expense of maintenance, and eventually ensure the future sustainability of the buildings. Ā is chapter highlights the use of smart materials for the optimal performance and safe design of buildings and other infrastructures particularly those under the threat of earthquake and ot her nat ural h azards. Sm art m aterials w ith emb edded desired f unctions suc h a s s ensing a nd p rocessing o r w ith improved s tructural p erformances suc h a s s trength, d uctility, usability, and low cost are the features that need to b e explored and applied in structures.

4.2

Materials and Application

4.2.1 Materials 4.2.1.1

Shape Memory Alloys

Ā e term shape memory alloys (SMA) refers to the ability of certain a lloys (Ni– Ti, Cu– Al–Zn, e tc.) to u ndergo l arge s trains,

while re covering t heir i nitial conguration at t he en d o f t he deformation process spontaneously or by heating without any residual d eformation. Ā e particular properties of SMAs are strictly a ssociated w ith a so lid–solid p hase t ransformation, which can be t hermal or stress-induced. Currently, SMAs are mainly applied i n me dical s ciences, ele ctrical, aerospace, a nd mechanical eng ineering a nd t he re cent s tudies i ndicate t hat they can open new applications in civil engineering specically in seismic protection of buildings. Properties, w hich ena ble Ni– Ti w ires f or c ivil eng ineering application, are as follows: 1. Repeated absorption of large amounts of strain energy under loading without permanent deformation 2. Possibility to ob tain a w ide r ange o f c yclic b ehavior from supplemental a nd f ully recentering to h ighly d issipating b y si mply v arying t he c haracteristics o f S MA components 3. Usable strain range of 70% 4. Extraordinary fatigue resistance under large strain cycles 4.2.1.1.1 Substitute for Steel? It is reported that the fatigue behavior of Cu–Zn–Al SMAs is comparable with steel [1]. If larger diameter rods can be manufactured, it has a p otential for use in civil eng ineering applications. Use of  ber-reinforced plastics with SMA reinforcements requires future experimental investigations. 4-1

43722_C004.indd 1

10/15/2008 4:11:11 PM

4-2

4.2.1.2

Smart Materials

Piezoelectric Materials

4.2.1.4 Electro- and Magnetorheological Fluids

Piezoelectricity was discovered as early as 1880 by Curie brothers [2]. When integrated into a s tructural member, a p iezoelectric material generates an electric charge or voltage in response to mechanical forces or stresses. Ā is p henomenon i s c alled direct piezoelectric effect. Ā is is useful for sensing. Ā e converse piezoelectric e ffect c an b e u sed f or c ontrol. A mong t he w ide variety of sensing and actuation devices, the advantages of using piezoelectric ac tuators a nd s ensors i nclude t heir e ffectiveness over a wide frequency range, simplicity, reliability, compactness, and lightweight. Even though lot of progress has been made in the laboratory level structural testing, their usage in large-scale civil eng ineering s tructures i s l imited d ue to h igh-voltage requirements.

Electro- and magnetorheological (ER/MR) uids have essential characteristics t hat cha nge f rom f ree-owing, l inear v iscous uid, to a s emisolid w ith a c ontrollable y ield strength in milliseconds when exposed to a n electric a nd magnetic  eld. Ā ese uids are variable contenders for the development of controllable devices. It is intended to develop a structure that controls its stiff ness and damping characteristic to behave adaptively against earthquake or w ind forces a nd ac hieve s afety a nd f unction by using E R/MR de vices w ith le sser energ y. A l arge-scale M R damper of 20 t capacity has been constructed and tested [3] and the experimental results indicate the applicability of MR damper in the semiactive control mode in the real-world applications.

4.2.1.3

Induced strain actuators (ISA) change their own shapes according to external electric and magnetic elds and vice versa. Ā ese materials have been widely used for small and precision machines because of some advantages such as small sizes, rapid reaction, high power, and high accuracy, etc. ISA materials act as sensors because they cause change in electric or magnetic elds under deformation. ISA m aterials c ould b e u sed to de velop sm art members to realize smart, comfortable, and safe structures. ISA materials c an b e a lso ut ilized f or v ibration mo de c ontrol o f structural memb ers a nd s ensor de velopment. Ā ey a re w idely used in active oor vibration control. Possible applications of ISA include

Engineered Cementitious Composites

Engineered cementitious composites (ECC) is mortar or concrete reinforced by chopped ber. Such composite materials have been m icrostructurally de signed u sing m icromechanical p rinciples. ECC exhibits strain hardening with large strain capacity and s hear d uctility, a nd go od d amage-tolerant me chanical behavior. Ā e use of high-performance cementitious structural elements a s energ y d issipation de vices a nd d amage-tolerant elements h elps to ac hieve a b etter p erforming a nd a d amagetolerant s tructural s ystem. D evelopment of c oncrete e ncased steel column elements also comes under this category. 4.2.1.3.1

Carbon Fiber Reinforced Concrete

Its ability to c onduct electricity and most importantly capacity to change its conductivity with mechanical stress makes a promising material for smart structures. It is evolved as a part of densied reinforced composites (DRC) technology. Ā e high density coupled w ith a c hoice o f  bers r anging f rom s tainless s teel to chopped carbon a nd kelvar, applied under high pressure, g ives the product outstanding qualities as per DRC technology. Ā is technology makes it p ossible to pro duce surfaces with strength and durability superior to metals and plastics. 4.2.1.3.2

Smart Concrete

A mere add ition of 0.5% specially treated carbon bers enables the increase of electrical conductivity of concrete. Putting a load on this concrete reduces the effectiveness of the contact between each ber and the surrounding matrix and thus slightly reduces its conductivity. On removing the load, the concrete regains its original c onductivity. B ecause o f t his p eculiar p roperty, t he product is called “smart concrete.” Ā e concrete could serve both as a structural material as well as a sensor. Ā e smart concrete could function as a traffic-sensing recorder when used as road pavements. It has got higher potential and could be exploited to make concrete reective to radio waves and thus suitable for use in electromagnetic shielding. Ā e smart concrete c an b e u sed to l ay sm art h ighways to g uide s elf-steering cars, which at present follow tracks of buried magnets. Āe strainsensitive concrete might even be used to detect earthquakes.

43722_C004.indd 2

4.2.1.5

Induced Strain Actuators

1. Long span structure 2. Axial force and friction control (for base isolator, including trigger application) 3. Active sound transparency (noise control) [4] 4. Wireless sensors of deformation

4.3 4.3.1

Structural Uses Active Control of Structures

Ā e concept of adaptive behavior has been an underlying theme of active control of structures, which are subjected to earthquake and other environmental type of loads. Ā e structure adapts its dynamic c haracteristics to me et t he p erformance objectives at any instant. A f uturistic smart bridge system (an artist’s rendition) is shown in Figure 4.1 [5]. A thermomechanical approach to develop a constitutive relation f or b ending o f a c omposite b eam w ith c ontinuous S MA bers embedded eccentric to neut ral a xis was used by Sun and Sun [6]. Ā e authors c oncluded t hat SMAs c an b e suc cessfully used for the active structural vibration control. Ā ompson et al. [7] also conducted an analytical investigation on the use of SMA wires to d ampen t he dy namic re sponse o f a c antilever b eam constrained by SMA wires. To date, active structural control has been successfully applied to over 20 commercial buildings and more than 10 bridges. One example of active control is the Kurusima bridge in Shikoku area

10/15/2008 4:11:12 PM

4-3

Structural Application of Smart Materials

Epoxy binders (matrices) Multidirectional carbon or other fibers Actuators Feedback control

Cable with sensors Data acquistion (from sensors) and processing

Dampers Pier sensors Wireless sensors

Protective coating Advanced composite materials

Sensors for intelligent vehicles Composite pipes for de-icing bridge deck (with geothermal energy)

Structural controls Stream sensors

Optical fiber sensors

FIGURE 4.1 Futuristic smart bridge system (an artist’s rendition). (Courtesy: USA Today, 3 March 1997.) J. Holnicki-szulc and J. Rodellar (eds.), Smart Structures, Requirements and potential applications in mechanical and civil engineering, NATO Science Series 3. High Technology, Vol. 65, Springer 1999.

in Japan. Ā e bridge was designed with the application of active vibration c ontrol a s i ntegrated s tructural c omponents. S everal modes of the bridge tower, which were anticipated to be excited by w ind vo rtex, w ere c arefully p rotected b y app ropriate controllers during the construction phase. It therefore made it possible for the tower of this bridge to be built much lighter and more s lender t han o ne f ollowing t raditional de sign. A ctive tuned mass dampers have been installed in the 11-story building, t he Kyobashi S eiwa building i n Tokyo (the  rst full-scale implementation of active control technology) a nd t he Nanjing Communication Tower in Nanjing, China. Ā ere a re t wo s erious c hallenges t hat rem ain b efore ac tive control c an ga in gener al ac ceptance b y eng ineering a nd construction pr ofessionals at l arge. Ā ey a re (1) Re duction o f capital c ost a nd m aintenance (2) I ncreasing s ystem rel iability and robustnessActive control systems consist a set of sensors, a controller, an active control system (actuators), and an external power supp ly. A s chematic s ketch o f ac tive c ontrol s ystem i s shown in Figure 4.2. Nowadays, m uch w ork o n s tructural c ontrol i s f ocused o n intelligent structures, developments of ac tuating materials a nd piezoceramics. D ue to i ts l imited f requency ba ndwidth, S MA has traditionally been used for passive strategies such as dampers or other types of energy dissipation devices instead of being used in active control strategies as actuators. On the other hand, it i s w ell k nown t hat a S MA ac tuator i s c apable o f p roducing relatively large control forces despite its slow response time. Ā is unique c haracteristic of SMA m akes it very at tractive for c ivil

43722_C004.indd 3

Controller

Sensors

Sensors

Power Active control

Excitation

Structure

Response

FIGURE 4.2 Schematic sketch of active control system.

engineering c ontrol app lications w here l arge f orces a nd lo wfrequency band width are mostly encountered. Ā ey conducted experimental s tudies o n ac tive c ontrol o f a  ve-story building model with SMA actuators [8]. It was proven that despite its slow response, it is feasible to use SMA for active control of civil engineering s tructures. But w hile s electing t he a lloy t ype, ut most care should be taken in specifying the temperature range or the transition temperature or the process in alloy-making so as to suit our requirements. Dynamics of SMA should be considered in control design.

4.3.2 Passive Control of Structures Two f amilies o f pa ssive s eismic c ontrol de vices e xploiting t he peculiar properties of SMA kernel components have been

10/15/2008 4:11:12 PM

4-4

Smart Materials

implemented a nd te sted w ithin t he M emory A lloys f or N ew Seismic Isolation a nd Energy Dissipation Devices (MANSIDE) project. Ā ey are special braces for framed structures and isolation devices for buildings and bridges.

4.3.3

Hybrid Control

Ā e term hybrid control generally refers to a combined passive and active control system. Since a portion of the control objective is accomplished by t he pa ssive system, less ac tive control effort, i mplying le ss p ower re source, i s re quired. Si milar control resource savings can be achieved using the semiactive control scheme where the control actuators do not add mechanical e nergy dir ectly t o t he s tructure, h ence b ounded-input/ bounded-output st ability i s g uaranteed. S emiactive co ntrol devices are often viewed as controllable passive devices. A side benet of hybrid and semiactive control systems is that, in the case of a p ower failure, the passive components of the control still offer some degree of protection unlike a fully active control system. Ā e materials described above can be used to make the h ybrid c ontrol s cheme w orkable. I n t he c ase o f E R/MR dampers, hybrid control scheme is a viable option for realistic structural control.

4.3.4 Smart Material Tag Ā ese sm art m aterial t ag c an b e u sed i n c omposite s tructures. Ā ese tags can be monitored externally throughout the life of the structure to r elate t he i nternal ma terial co ndition. S uch m easurements as stress, moisture, voids, cracks, and discontinuities may be interpreted via a remote sensor [6].

4.3.5 Retrofitting SMAs ca n us ed as s elf-stressing  bers a nd t hus t hey c an b e applied for retrotting. Self-stressing bers are the ones in which reinforcement is placed into the composite in a nonstressed state. A prestressing force i s i ntroduced i nto t he s ystem w ithout t he use o f l arge me chanical ac tuators b y p roviding S MAs. Ā ese materials d o n ot n eed specia lized e lectric eq uipments n or d o they create safety problems in the eld. Treatment can be applied at a ny t ime a fter hardening of t he matrix i nstead of during its curing and hardening. Long- or short-term prestressing is introduced by triggering the change in SMA shape using temperature or e lectricity. Ā ey ma ke ac tive la teral co nnement of b eams and columns a more practical solution. Self-stressing jackets can be manufactured for rehabilitation of existing infrastructure or for new construction. 4.3.5.1

Restoration of Cultural Heritage Structures Using SMA Devices

An innovative technique using superelastic SMA devices for the restoration o f a c ultural h eritage s tructure e specially m asonry buildings were implemented under the framework of the European Commission-funded IS TECH P roject. M asonry b uildings a re

43722_C004.indd 4

largely vulnerable to earthquakes because of their low resistance and ductility during earthquake ground motion. To enhance the seismic b ehavior o f c ultural h eritage s tructures, t he mo st common method traditionally used has been the introduction of localized reinforcements. Usually steel bars or cables served this purpose by increasing stability and ductility. But in many cases, these rei nforcement te chniques p rove i nadequate to p revent collapse. Ā e development of the connection technique was based on the ide a o f usin g t he uniq ue p roperties o f N i–Ti a lloys esp ecially i ts su per elastici ty a nd hig h r esistance t o co rrosion [9]. Ā e idea was to connect the external walls to the oors, the perpendicular walls o r the r oof with an SMA Device that should behave as follows: 1. Under lo w-intensity h orizontal ac tions ( wind, sm all intensity earthquakes), the device remains stiff, a s traditional s teel c onnections do, not a llowing sig nicant displacements. 2. Under h igher i ntensity h orizontal ac tions ( i.e., s trong earthquakes), the stiff ness of the device decreases, allowing “ controlled d isplacements,” w hich s hould re duce amplication of a ccelerations (as c ompared t o s tiff connections) and permit the masonry to dissipate part of the transmitted energy, mainly owing to elasticity exploitation and microcracks formation in the brick walls; consequently, the s tructure s hould b e a ble to su stain a h igh-intensity earthquake w ithout c ollapse, t hough u ndergoing s ome minor damage. 3. Under e xtraordinary h orizontal ac tions, t he s tiffness of the device increases and thus prevents instability. 4.3.5.2 SMA for Seismic Retrofit of Bridges Unseating o f supp orts w as t he m ajor c ause o f b ridge f ailures during earthquakes. Retrot measures to re duce the likelihood of collapse due to unseating at the supports have been in place for m any ye ars. Ā e d amage to b ridges i n t he re cent C hi-Chi, Kobe, and Northridge earthquakes indicate the need to provide better methods of reducing the damaging effects of earthquakes in b ridges. Ā e u se o f re strainer c ables a nd re strainer ba rs to limit t he rel ative h inge d isplacement b ecame p opular i n t he United States following the collapse of several bridges due to loss of supp ort d uring t he 1 971 S an F ernando e arthquake. Re cent earthquakes have demonstrated that restrainers were effective in some c ases. However, m any bridges w ith re strainers su stained serious d amage o r c ollapse. Br idges t hat h ad b een re trotted with restrainer cables failed in both the 1989 Loma Prieta and 1994 N orthridge e arthquakes. F ailure o f J apanese r estraining devices a lso o ccurred d uring t he 1 995 K obe e arthquake. Experimental tests of restrainer cables have shown that failure occurs i n th e c onnection e lements o r th e th rough-punching shear in the concrete diaphragm. In addition, restrainers do not dissipate any signicant amount of energy, since they are generally designed to remain elastic. Analytical studies of bridge and restrainer systems have demonstrated that a very large number

10/15/2008 4:11:14 PM

4-5


SMA restrainer bar/damper

Real system

(a)

Stay cable

SMA damping device

(b)

FIGURE 4.3 SMA restrainer bar used in multispan simply supported bridge abutments and intermediate piers conguration.

of re strainers a re often re quired to l imit jo int mo vement to acceptable le vels, pa rticularly f or h igh s eismic lo ads. I n t hose cases, t he e xcessive n umber o f re strainers w ould i nduce l arge forces in other components of the bridge, such as bearings and columns. Ā e shortcomings of traditional restrainers can potentially b e add ressed w ith t he u se of SMA re strainers. Ā e SMA restrainers, in the superelastic phase, act as both restrainers and dampers (Figure 4.3). Energy dissipation and base isolation are found to be optimal candidates for structural control of structures. As far as bracing systems a re c oncerned, u ntil now a ll t he app lications a nd t he research studies on this technique were focused on the energ y dissipation c apability. F or s eismic re trotting p urposes, t he supplemental recentering devices (SRCD) are found to be useful as t hey provide forces to re cover t he u ndeformed shape of t he structure at the end of the action. In existing structures, in fact, particularly when they were designed without any seismic provision, the energy dissipation can turn out to b e insufficient to limit damage to structural elements. It would then be necessary to strengthen some elements to f ully achieve the design objectives. L ocal s trengthening w ould i mply e xpensive w ork, a lso involving nonstructural parts. Retrotting could turn out to be economically in convenient, and, yet some residual displacement c ould o ccur i n c ase el astoplastic de vices a re u sed. A n alternative s trategy c an b e p ursued b y u sing S MA de vices having s upplemental forc e t o re cover t he u ndeformed s tructural conguration, resulting in the elimination of any residual displacement, while accepting y ielding i n structural elements. A comparison of properties of Ni–Ti w ith steel is g iven i n t he Table 4.1.

43722_C004.indd 5

TABLE 4.1 Comparison of Properties of Ni–Ti SMA with Typical Structural Steel Property

Ni–Ti SMA

Steel

Recoverable elongation (%) Modulus of elasticity (MPa)

8 8.7 × 104 (A) 1.4 × 104 (M) 200–700 (A) 70–140 (M) 900 (f.a.) 2000 (w.h.) 25–50 (f.a.) 5–10 (w.h.) Excellent

2 2.07 × 105

Yield strength (MPa) Ultimate tensile strength (MPa) Elongation at failure (%) Corrosion performance

248–517 448–827 20 Fair

Note: f.a. denotes fully annealed and w.h. denotes work hardened, w hich a re tw o typ es o f tr eatment gi ven t o the allo y. A and M deno te the tw o phases of the allo y namely, austenite and martensite.

4.3.6

Self-Healing

Experimentally proved self-healing behavior [10], which can be applied at microlevel of a material, widens their spectrum of use. Here signicant deformation beyond the  rst crack can be fully recovered and cracks can be fully closed.

4.3.7

Self-Stressing for Active Control

Self-stressing for ac tive c ontrol c an b e u sed w ith c ementitious ber composites with some prestress, which impart self-stressing

10/15/2008 4:11:14 PM

4-6

Smart Materials

thus avoiding difficulties due to the provision of large actuators in ac tive c ontrol, w hich re quire c ontinuous m aintenance o f mechanical pa rts a nd r apid mo vement, w hich i n t urn c reated additional inertia forces. In addition to SMAs, some other materials such as polymers can also be temporarily frozen in a prestrained state that have a potential to be used for manufacturing of self-stressing cementitious composites [1].

4.3.8

4.3.9

Active Railway Track Support

Active control system for sleepers is adopted [5] to achieve speed improvements on existing bridges and to maintain the track in a straight and nondeformed conguration as the train passes. With the help of optimal control methodology, the train will pass the bridge w ith re duced t rack de ections a nd v ibrations a nd t hus velocity could be safely increased. Figure 4.5 shows various positions of the train with and without active railway track support.

Structural Health Monitoring 4.3.10

Use o f p iezotransducers, su rface b onded to t he s tructure o r embedded in the walls of the structure can be used for structural health monitoring and local damage detection. Problems of vibration and UPV testing can be avoided here. Jones et al. [11] applied neural networks to  nd t he magnitude a nd location of an impact on isotropic plates and experimented using an array of piezotransducers surface bonded to the plate. Figure 4.4 shows a typical health monitoring setup making use of optic bers.

Active Structural Control against Wind

Aerodynamic control devices to mitigate the bidirectional wind induced v ibrations i n t all b uildings a re energ y e fficient, since the e nergy i n th e  ow i s u sed to p roduce t he de sired c ontrol forces. Aerodynamic ap system (AFS) is an active system driven by a feedback control algorithm based on information obtained from t he v ibration s ensors [ 5]. Ā e a rea o f  aps a nd a ngular amplitude of rotation are the principal design parameters.

Seismometer Vibration meter Optic fibers

Optic fiber

Optic fibers

Damage detection of piles using optic fibers

Vibration meter Hitting

Cracks Damage detection after damage occured

FIGURE 4.4 Health monitoring setup.

43722_C004.indd 6

10/15/2008 4:11:14 PM

4-7


References

Bridge response without active railway track support

Bridge response with active railway track support

FIGURE 4.5 Active railway track support concept.

4.4

Conclusion

Ā e technologies using smart materials are useful for both new and existing constructions. Of the many emerging technologies available, t he f ew de scribed i n t his c hapter ne ed f urther research to e volve t he de sign g uidelines o f s ystems. C odes, standards, and practices are of crucial importance for the further development.

Acknowledgment Ā e a uthors gra tefully ac knowledge the Dir ector, S tructural Engineering Res earch C entre (S ERC), CS IR, I ndia f or the co nstant encouragement and support rendered in preparation of this manuscript a nd als o gi ving p ermission t o p ublish i t. Ā e kind support a nd guida nce o f all the t eam mem bers o f S tructural Dynamics Laboratory deserve acknowledgment.

43722_C004.indd 7

1. N. Krstulovic-Opara and A.E. Naaman, Self Stressing Fiber Composites, ACI Structural Journal, 97, March–April 2000, 335–344. 2. W.G. Cady, Piezo Electricity, Dover, New York, 1964. 3. G. Yang, B.F. Spencer Jr., J.D. C arlson, and M.K. Sain, L arge scale MR uid dampers: Modeling and dynamic performance considerations, Engineering Structures, 24, 2002, 309–323. 4. A. Sampath and B. Balachandran, Active control of multiple tones in an endosure, Ā e Journal of the Acoustical Society of America, 106-1, July 1999, pp. 211–225. 5. J. H olnicki-szulc a nd J . Ro dellar (e ds.), Smart St ructures, Requirements and potential applications in mechanical and civil engineering, NATO Science Series., 3. High Technology, Vol. 65, Springer 1999. 6. G. Sun and C.T. Sun, Bending of shape memory alloy reinforced composite beam, Journal of Materials Science, 30(13), 1995, 5750–5754. 7. P. Ā omson, G.J. Balas, and P.H. Leo, Ā e use of shape memory Alloys f or pa ssive s tructural d amping, Smart M aterials Structure, 4, 1995, 36–42, IOP publishing limited. 8. J. Li, B. Samali, and C. Chapman, Experimental realisation of active co ntrol o f a  ve st orey b uilding mo del usin g S MA actuators, Advances in Mechanics of Structures and Materials, Loo, Chowdhury and Fragomeni (eds.), 699–704. 9. M.G. Castellano and G. Manos, Ā e ISTECH Project: Use of SMA in t he Seismic Protection of Monuments, Monument– 98 (W orkshop o n S eismic P erformation o f M onuments), Lisbon, 1998. 10. D.J. Hannant and J.G. Keer, Autogeneous healing of Ti based sheets, Cement and Concrete Research, 13, 1983, 357–365. 11. R.T. Jones, J.S. Sirkis, and E.J. Friebele, Detection of impact location a nd ma gnitude f or is otropic p lates usin g neural networks, Journal o f I ntelligent M aterial S ystems a nd Structures, 7, 1997, 90–99.

10/15/2008 4:11:15 PM

43722_C004.indd 8

10/15/2008 4:11:16 PM

5 Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites 5.1 H ybrid Composites.................................................................................................................... 5-1 Introduction • Fut ure Directions

Acknowledgment ................................................................................................................................. 5-7 5.2 Design of an Active Composite Wing Spar with Bending–Torsion Coupling ..................5-7 Introduction • M ulticell Cross-Section Spar Design • R esults • C oncluding Remarks

References ........................................................................................................................................... 5-16

5.1 Hybrid Composites S. Padma Priya 5.1.1

Introduction

Composites a re materials made by t he synergistic assembly of two or more constituting materials (matrix and reinforcement), engineered in such a way that they form a single component and yet can be distinguished on a macroscopic level. Matrix encases the rei nforcements a nd ac ts a s a b inder f or t he  bers. Additionally, t he m atrix h olds t he rei nforcements i n a  xed position a nd t he reinforcement enhances t he properties of t he composite s ystem b y i mparting i ts me chanical a nd ph ysical properties. Composite materials can be engineered with desired properties by choosing and varying the concentration of different t ypes o f m atrices a nd rei nforcements. The d urability o f composites is very high; they have very high strength to weight ratios, a re re sistant to en vironmental c orrosion, a nd p rovide ease of use concept. Composites a re m aterials k nown f rom a ncient d ays, w hen bricks were made (mud reinforced with straw). Until today, we have applications for components of aerospace materials. Mother Nature i s t he  rst m aker of c omposites a nd s ome e xamples of natural composites include wood, bone, etc. Composites are classied into several broad categories depending upon the type of matrices and reinforcements used. Depending on the types of matrices used they can be classied as

1. Metal matrix composites (MMC) 2. Ceramic matrix composites 3. Polymer matrix composites Depending up on t he rei nforcement t ypes t hey c an b e c lassied as 1. Fiber reinforced composites 2. Fabric reinforced composites 3. Particulate composites In o rder to f urther en hance t he p roperties o f t he c omposite materials, hy brid c omposites we re d eveloped. The b ehavior of hybrid composites appears to be simply a weighted sum of the individual components in which there exists a more favorable balance between the advantages and disadvantages. The hybrid composites o ffer a dvantages r egarding s tructural in tegrity a nd sustained load under crash and impact conditions. Hybrid composites a re i nuenced b y a l arge n umber o f m icrostructural parameters such as type of reinforcement, volume proportion of reinforcement, weave pattern of the fabric used as reinforcement, volume fraction of the reinforcement, etc. There are several types of hybrid composites and can be characterized as: (1) interply or tow-by-tow, i n w hich to ws o f t wo o r mo re c onstituent t ypes o f ber a re m ixed i n a re gular o r r andom m anner; ( 2) s andwich hybrids, a lso k nown a s c ore-shell, i n w hich o ne m aterial i s sandwiched between t wo layers of a nother; (3) i nterply or laminated, w here a lternate l ayers o f t he t wo (or mo re) m aterials a re stacked in a regular manner; (4) intimately mixed hybrids, where 5-1

43722_C005.indd 1

10/3/2008 8:44:44 PM

5-2

Smart Materials

the constituent bers are made to mix as randomly as possible so that no overconcentration of any one type is present in the material; (5) other kinds, such as those reinforced with ribs, pultruded wires, thin veils of ber, or combinations of the above. Hybridization can be done through many ways as follows:

is its high fracture toughness, which is 50–100 times higher than that of epoxies. Another important advantage of PEEK is its low water absorption, which is less than 0.5% at 23°C compared to 4%–5% for conventional aerospace e poxies. B eing s emicrystalline, it does not dissolve in common solvents.

1. 2.

5.1.1.1.2 Natural Fibers

Reinforcements Matrices 3. Both reinforcements and matrices

5.1.1.1

Reinforcement Hybridized Composites

These are the composites, which consist of two or more types of reinforcements embedded in a single matrix. The reinforcements can be of different types as follows. 5.1.1.1.1 Synthetic (Man-Made) Fibers Th is c omposite i s c omposed o f t wo t ypes o f rei nforcements and both of them are synthetic. Composites of such types have been developed in order to achieve superior mechanical properties a nd h igh h eat re sistance c apabilities. Various t ypes o f synthetic re inforcements h ave b een d eveloped s uch a s g lass ber, carbon ber, Polyetheretherketone (PEEK), rayon, nylon, Kevlar, polyester, acrylic, ole n, vinyl, aramid, etc. The following bers can be hybridized (e.g., carbon ber/PEEK hybrid fabric, p olyethylene [ PE]/glass f abric, c arbon  ber/aramid bers) in diff erent (weight or volume) ratios according to t he desired en d p roperty. F or i nstance, i f le ss w eight h owever, a strong c omposite i s re quired, a h ybrid f abric c an b e c onsidered. Th is would have one fabric composed of different types of bers in it and some examples for this type of reinforcement are carbon/aramid, aramid/glass, and carbon/glass. Of the materials mentioned above the PEEK material is very much in demand. PEEK is a linear aromatic thermoplastic based on the following repeating unit in its molecules. O C

O

Ketone

Ether

O n Ether

Polyether ether ketone

Continuous-carbon  ber rei nforced P EEK c omposites a re known i n t he i ndustry a s a romatic p olymer c omposite (APC). PEEK is a semicrystalline polymer, and amorphous PEEK is produced when the melt is quenched. The presence of  bers in PEEK composites tends to increase the crystallinity to a higher level because the bers act as nucleation sites for crystal formation. Increasing crystallinity increases both the modulus and the yield strength of PEEK but reduces its strain to failure. PEEK’s maximum continuous-use temperature is 250°C. PEEK i s t he f oremost t hermoplastic m atrix t hat m ay re place epoxies in many aerospace structures. Its outstanding property

43722_C005.indd 2

A c omposite i n w hich i ts rei nforcements c ome f rom nat ural resources is considered natural. Natural  bers, because t hey a re naturally rene wable, c an b e i mbedded i nto b iodegradable p olymeric materials, which can produce a new class of biocomposites. Recently, society has realized that unless the environment is protected, it w ill be t hreatened due to t he loss of nat ural resources. Conservation o f f orests a nd opt imal ut ilization o f a gricultural and other renewable resources like solar energy and seawater has become almost mandatory. In this view, many attempts have been made to replace high-strength synthetic bers with natural bers. A si ngle nat ural  ber m ay not i mpart a ll t he p roperties f or t he nal composite, thus it is hybridized with other types of natural bers. S ome o f t he k nown nat ural  bers a re w ool, c otton, si lk, linen, he mp, r amie, jut e, c oconut, pi newood, pi neapple, a ngora, mohair, a lpaca, k apok,  ax, s isal, b ast  ber, k enaf, w ood  ber, bamboo, ba nana, e tc. To t ailor t he p roperties o f a nat ural  ber reinforced composites, it is hybridized with the second type of natural ber. Researchers have conducted studies on thermal conductivity and specic heat of jute/cotton, sisal/cotton, and ramie/ cotton hybrid fabric-reinforced composites. Small inorganic particles have also been used as  llers to prepare pa rticulate c omposites. The mo st re levant re ason to u se ller in the composite is to re duce the overall cost of the material. F illers a re a lso u sed to i ncrease  ame re tardancy, su rface hardness, e sthetic app eal, a nd t he t hermal p roperties o f t he composites. Some of the types of  llers that have been used are y a sh, si lica, m ica, g ranite p owder, w ood  our, aluminum oxide, carbon black, etc. Using such  llers and natural bers or fabric, a new class of hybrid composites has been produced. One of the reinforcements is  ller while the other is the  ber or the fabric, a nd h ere t he  ller pa rticles a re emb edded b etween t he brous rei nforcement a nd t he matrix. Other research developments underway have used sisal/saw dust in conventional composites and they possess an ease of processing, are environment friendly, and are economically affordable. 5.1.1.1.3 Natural and Synthetic Fibers Th is type of composite consists both of natural and synthetic reinforcements i n t he c omposite s ystem. N atural  ber reinforced composites do not always ful ll all the properties required for some technical applications. In such cases, hybridization w ith sm all a mounts o f s ynthetic  bers a llows t hese natural ber composites to be more suitable for technical applications. The m ost co mmon s ynthetic  ber u sed t o h ybridize the natural ber as a rei nforcement is glass  ber. Some examples of these types of bers are sisal/glass, silk/glass, pineapple ber/glass, jute/glass, etc. With the addition of a small amount of g lass f abric to t he nat ural  ber r einforced co mposite, t he

10/3/2008 8:44:44 PM

5-3

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

mechanical strength and the chemical resistance of the composite has been shown to i ncrease considerably (see Table 5.1 and Figure 5.1). 5.1.1.1.4 Miscellaneous Reinforcements Miscellaneous reinforcement is also done to achieve required properties in the composites and some of them are synthetic ber–particulate (glass along with y ash) composites used to increase t he abrasion resistance a nd compression strength of

the composites, blending various levels of glass to wollastonite allows for tailored composites with high-strength properties or go od d imensional s tability. F iber me tal l aminates ( FML) offer signicant improvements over currently available materials f or a ircraft s tructures d ue to t heir e xcellent me chanical characteristics a nd relatively low density a nd some examples for t hese t ypes o f c omposites a re a luminum 2 024 a lloy; carbon ber/epoxy (Ep), and aluminum 2024 alloy/glass fiber/ Ep composites.

TABLE 5.1 Properties of Selected Commercial Reinforcing Fibers

Fiber

Typical Diameter (mm)a

Specic Gravity

Tensile Modulus (GPa) 72.4 86.9

Glass E-glass S-glass

10 10

2.54 2.49

PAN carbon T-300c AS-1d AS-4d T-40c IM-7d HMS-4d GY-70e

7 8 7 5.1 5 8 8.4

1.76 1.80 1.80 1.81 1.78 1.80 1.96

Pitch carbon P-55c P-100c

10 10 11.9 1.47

Aramid Kevlar 49f Kevlar 149f Technorag Extended-chain PE Spectra 900 Spectra 1000 Boron SiC Monolament Nicalon (multilament)h Al2O3 FiberFPf Al2O3–SiO2i Fiberfrax (discontinuous)

Tensile Strength (GPa)

Strain to Failure (%)

Coefficient of Thermal Expansion (10−6/°C)b

Poisson’s Ratio

3.45 4.30

4.8 5.0

5 2.9

0.2 0.22

231 228 248 290 301 345 483

3.65 3.10 4.07 5.65 5.31 2.48 1.52

1.4 1.32 1.65 1.8 1.81 0.7 0.38

−0.6, 7–12

0.2

2.0 2.15

380 758

1.90 2.41

0.5 0.32

−1.3 −1.45

1.45 1.79 1.39

131 3.45 70

3.62 1.9 3.0

2.8

−2, 59

4.4

−6

38 27 140

0.97 0.97 2.7

117 172 393

2.59 3.0 3.1

3.5 2.7 0.79

5

140 14.5

3.08 2.55

400 196

3.44 2.75

0.86 1.4

1.5

20

3.95

379

1.90

0.4

8.3

2–12

2.73

103

1.03–1.72

−1.72

−0.75

0.35

0.2

1 mm = 0.0000393 in. 1 m/m per °C = 0.556 in./in. per °F. c Amoco. d Hercules. e BASF. f DuPont. g Teijin. h Nippon carbon. i Carborundum. a

b

43722_C005.indd 3

10/3/2008 8:44:44 PM

5-4

Smart Materials

Specific stiffness (tensile strength/density, in ⫻ 106)

Goal

(Kevlar 29) (Kevlar 49) aramid aramid 10

S glass Polyethelene

6

Boron Carbon (P-100)

4 E glass SiC (ceramic fiber)

2

0

200

Carbon (cellon GY-70)

Alumina (fiber FP)

600 400 800 1000 1200 Specific stiffness (tensile modulus/density, in ⫻ 106)

1400

Specic properties of advanced bers.

Matrix Hybridized Composites

The m atrix i s a lso o ne o f t he i mportant c onstituents o f t he composites. H ybridizing a m atrix w ith a nother m atrix a lso means toughening of t he matrix. Suc h matrices w hen u sed to prepare t he c omposites a lso c an le ad to a h ybrid c omposite. Toughening of the matrix is carried out in order to re duce the brittleness of the binder matrix and as a consequence to improve the (mechanical and chemical resistance) properties of the composites. An unmodied matrix usually consists of single-phase materials, while the addition of modiers turns the toughened matrix into a multiphase system. W hen modier domains a re correctly d ispersed i n d iscrete f orms t hroughout t he m atrix, the f racture energ y o r to ughness c an b e g reatly i mproved. I n the c ase of me tals, d ifferent types of alloys have been used to produce t he c omposites. P olymer blen ds, b y de nition, are physical m ixtures of s tructurally d ifferent homop olymers or copolymers. In polymer blends or polymer alloys, the mixing of two or more polymers provides a new material with a modied array of properties. For polymers, there are as known there are two types of matrices: thermoplastic (polymethyl methacrylate, polycarbonate [PC], polybutylene terephthalate [PBT], polystyrene [PS], etc.) and thermosets (Ep, unsaturated polyester, etc). Toughening of thermoplastics is usually carried out by blending them with different types of thermoplastic materials and elastomers i n order to i mprove t heir mechanical properties a nd t he exibility of t he matrix. E xamples i nclude blends of PC a nd a

43722_C005.indd 4

Ordered polymer fiber

(Kevlar 49 (dry))

8

0

5.1.1.2

PBO Carbon (T-300)

Titanium Aluminum HSLA Steel

FIGURE 5.1

PBT

Resin impregnated strands

12

thermoplastic copolyether/ester/elastomers, which is shown to have a signicant effect on t he properties. Polyoxymethylene/ elastomer/ ller ter nary co mposites ha ve b een p repared, i n which a thermoplastic polyurethane and inorganic ller, CaCO3, were used to achieve balanced mechanical properties for polyoxymethylene. Two other thermoplastics have also been blended together to achieve composites with high-performance properties as well as improve the mechanical properties of the composite: p olyamide ( PA) d ispersed i n a P E t hermoplastic a s a matrix for glass bers. Researchers have also found that there is a signicant increase in the degradation properties of thermoplastic blen ds w ith t he add ition o f c eramic m aterials ( PE-coethyl ac rylate w ith a p olyisobutyl me thacrylate p olymer reinforced with ceramic oxide powders). Some other examples of h ybrid t hermoplastic m atrices a re P C/acrylonitrile-butadiene-styrene ABS blend, PS/PC, PVA/polymethyl methacrylate, etc. Though a variety of thermoset materials have been used to prepare m atrix h ybrid c omposites, to ughening o f t hese c omposites by using Ep and unsaturated polyester are probably the most widely used combination. Different kinds of modiers have been studied to improve the toughness o r d uctility o f c ured t hermoset re sins. The y can be classied as liquid rubbers, engineered thermoplastics, reactive diluents, a nd i norganic pa rticles. Some of t he following examples illustrate this type of toughening: Ep toughened with polymethylmethacrylate ( PMMA), Ep to ughened w ith P C, Ep toughened with PBT, Ep with reactive liquid rubber such as Ep

10/3/2008 8:44:45 PM

5-5


phenol l iquid w ith c ashew nut s hells, u nsaturated PE w ith Ep, unsaturated P E to ughened w ith P MMA, u nsaturated P E w ith PBT, etc. Signicant increases in some mechanical- and chemical-resistant properties were found for all of the above combinations of cured thermoset resins. 5.1.1.3 Reinforcement and Matrix Hybridized Composites Composites of this type consist of multiphase materials (four phases) i n which both t he matrix a nd t he rei nforcements have been hybridized. In one way, it can be said that it is the combination of t he above t wo t ypes. This a rea of research is a ne w a nd challenging c oncept b ecause a s t he i ncorporation o f m aterials increases, t here a re mo re i nterphases i n t he c omposite s ystem and if there are any differences in the interfaces of the materials, it could lead to the failure of the complete system. Care should be always taken to meet these future challenges in order to achieve a material with superior mechanical properties. It should also be noted t hat t he v iscosity of t he toughened matrix resin be sufficient enough to wet the hybrid reinforcements. One of the examples that has been developed is silk/glass fabric and used as a reinforcement i n t hermoplastic-toughened Ep re sin to ac hieve high mechanical properties. The potential for the usage of hybrid composites far outweighs any ne gative a spects si nce t here a re a v ast n umber o f p lastic/ plastic, p lastic/metal ar amid-reinforced a luminum l aminate (ARALL), metal/metal (MMC), metal/ceramic, ceramic/ceramic (space shuttle outer tile), and ceramic/plastic composite combination systems yet to be explored, investigated, developed, and in use. 5.1.1.4 Hybrid Composite Applications Hybrid composite materials (HCM) represent the newest of the various c omposite m aterials c urrently u nder de velopment. The hybrid composite category covers both the hybridizing of a composite material with other materials (either other composites or base unreinforced materials) and composites using multiple reinforcements. Fu rther, t his c ategory c overs t he u se o f multiple m aterials ( at le ast o ne o f w hich i s a c omposite) i n structural app lications a nd h ighlights t he m ultiple u ses a nd advantages of composite materials. Hybrid composites can be divided into  ve major subcategories: (1) HCM, (2) selective reinforcements, (3) thermal management, (4) sm art s kins a nd s tructures, a nd (5) u ltralightweight materials. 5.1.1.4.1 HCM HCM are dened as a c omposite material system derived from the integrating of dissimilar materials, at least one of which is a basic composite material. A typical example of a HCM is a reinforced polymer composite combined with a conventional unreinforced homo genous me tal. The H CM blen ds t he de sirable properties of two or more types of materials into a single material system, which displays the benecial characteristics of the separate constituents. An existing example of a hybrid composite is

43722_C005.indd 5

7075-T5 Aramid/epoxy

NOM. 1.3 mm

7075-T5

7075-T6 sheet 0.03 mm

Aramid/epoxy

FIGURE 5.2

ARALL is an example of an HCM in production.

ARALL, which consists of high-strength aluminum alloy sheets interleafed with layers of aramid ber-reinforced adhesive as illustrated in Figure 5.1. The A RALL h ybrid c omposite (Figure 5.2) is baselined on several secondary structural composites of  xed-wings ubsonica ircraft. A second example of a hybrid composite i s a c arbon–carbon c omposite (CCC) w ith a si ngle side application o f t he re fractory me tal rh enium. This carbon–carbon–rhenium material is being developed for thermal management heat pipes on space-based radiator systems. Other examples include i nterpenetrating p olymer ne tworks (I PNs), w hich a re hybrid resin matrices consisting of thermoset and thermoplastic resin combinations. Still another HCM concept involves multiple rei nforcement t ypes w ithin a c ommon m atrix suc h a s chopped bers and continuous bers within a polymer matrix. Other HCMs i nclude new composite materials such as na nocomposites, functionally gradient materials (FGMs), hybrid materials (Hymats), IPNs, microinltrated macrolaminated composites (MIMLCs), and liquid crystal polymers (LCPs), which may force development of previously uneconomical process routes if they offer the path to a technical solution for advanced system capability. These materials present opportunities for reducing the number of stages in turbine engines and in so doing may be economically benecial even at a higher material cost because the smaller number of stages leads to greater economy of use. HCM technology is in its infancy in comparison to that of the other types of composite materials. Whereas ARALL and IPNs have been used for the past decade, the other types of HCM are truly emb ryonic. F rom ne arly a ll a spects o f re search, t hese HCM offer great potential for structural applications; however, their widespread use remains a decade or more in the future. 5.1.1.4.2 Selective Reinforcement Selective reinforcement is the category of hybrid composites that provide reinforcement to a structural component in a local area or areas by means of adding a composite material. An example of this is t he use of superplastic forming-diff usion bonding (SPF/ DB) as a means of integrating a titanium-reinforced MMC into a base titanium structure. As part of the initial design approach, consideration must b e g iven to t he to oling re quired f or p lacement o f t he r einforcing m aterial w ithin t he s tructure. This is accomplished by building into t he form tooling areas in which the M MC m aterial i s placed t hat p ermit SPF e xpansion of t he

10/3/2008 8:44:45 PM

5-6

Smart Materials

base t itanium to t he M MC d uring p rocessing a nd a llowing diff usion b onding t o o ccur, t hus a ccomplishing in tegral r einforcing of the nal structure. The selective reinforcement design approach allows the aerospace de signer to ut ilize t he mo re c ostly, h igher p erformance materials o nly w here t hey a re re quired a nd t he le ss e xpensive materials in areas where they can perform the job. This approach leads to an optimization of both cost and performance in the most ideal case. Similarly, problems associated with use of the reinforcing c omposite, suc h a s lo w me chanical jo int s trength, can be eliminated by not rei nforcing the area where the mechanical joining occurs. In reality, while this approach can be very efficient, it requires the introduction of multiple materials to solve the design problem and may in turn result in increased fabrication costs and risks. 5.1.1.4.3 Thermal Management In t he  eld o f t hermal m anagement, c omposite m aterials a nd HCM c an b e i nnovatively c onstructed to e ffectively l imit t he maximum tem perature o f s tructural h ardware a nd r apidly transfer heat from hot areas to cool areas. This capability arises from t he u nmatched t hermal c onductivity o f g raphite  ber, which is higher (in the  ber direction) than that of oxygen-free high-conductivity copper (OFHC). The graphite bers act as heat paths and, by suitable arrangement in the structure of interest, can remove heat by transmitting it along its length. In contrast, the matrix materials can act as thermal insulators so that thermal conduction through the thickness is lower by orders of magnitude than in-plane. This allows designers to develop a structure that is a thermal insulator in some directions but a thermal conductor in others. 5.1.1.4.4

Smart Skins and Smart Structures

Smart skins and smart structures are related in that each contains embedded, no nstructural elemen ts. A sm art s tructure c ontains sensors that monitor the health of the structure itself, such as ber optics to de termine temperature and structural deformations or cracks. A sm art s kin c ontains c ircuitry a nd ele ctronic c omponents that enable the skin to double as part of the electronic system of the parent vehicle, be it an aircraft or a missile. Smart s tructure te chnology, l ike t hat o f sm art s kins, i s s till evolving. The present technology consists of the incorporation of sensors into structural elements in the material processing stage to better control cure (or consolidation), and so on. 5.1.1.4.5 Ultralightweight Materials The c ategory of u ltralightweight materials i ncludes t he emerging family of liquid crystal ordered polymers, which by virtue of their mole cular s tructure, e xhibit e xtremely h igh s pecic strength and specic stiffness. These highly directional materials are similar to c omposite materials in t hat t he long, ordered molecular c hains w ithin t he p olymer ac t very much l ike rei nforcing f ibers i n a c omposite m aterial. E xamples o f t hese materials a re p oly-p-phenylene b enzobisthiazole (P BZT) a nd

43722_C005.indd 6

poly-p-phenylene b enzobisoxazole (P BO). A nother e xample i s gel-spun P E. These po lymers ex hibit ex tremely h igh speci c strength a nd speci c stiff ness a nd c an p otentially b e u sed a s reinforcing bers in composite laminates, as rope or cable, or as a self-reinforced thin  lm structure. These ordered polymers, in the form of thin lms, nd use in shear webs and skin applications f or a ircraft. These t hin  lms c an a lso b e p rocessed i nto honeycomb for l ightweight s tructural applications. It h as b een estimated that a s hear web made of PBZT would be one-eighth the weight of a n a luminum web a nd one-sixth t he weight of a graphite Ep web.

5.1.2

Future Directions

A t ype o f me tallic s tructure, w hich h as o nly re cently re ceived attention is nanostructure, in which the microstructural dimensions a re in t he na nometer range (10−8 to 1 0−9 m). The se nanostructures can be tailored to be either basically equiaxed or layered in nature, depending on production conditions. These nanostructures offer novel combinations of strength, ductility, and stiff ness not possible using conventional processing te chniques. I n t he l ayered c onguration w hen the th ickness o f t he l ayers i s b elow app roximately 1 00 nm, s o-called Koehler strengthening occurs in conjunction with an apparent modulus i ncrease, op ening up t he p ossibility o f v arying a nd controlling both strength and stiff ness by varying layer thickness. P ractical te chniques to p roduce na nostructures f rom metals include the electron beam coevaporation process developed b y t he Ro yal A cademy o f E ngineering ( RAE), w hich i s capable o f 18 k g/h de position r ates a nd me chanical a lloying. The r ealms o f poss ibilities h ere see m a lmost bo undless, w ith the caveat of also including ceramic bers in an overall tailored structure. Aluminum-based a lloys h ave b een u sed to d evelop h ybrid engineered materials such as MMC, mechanically alloyed dispersion strengthened a luminum (DISPAL), and AR ALL. The latter two materials offer high-temperature capability and outstanding fatigue performance, respectively. A number of production approaches have been studied, evaluated, and used to produce MMCs including powder metallurgy (P/M), casting, an in situ (XD) precipitation technique, and plasma spraying. The result is increased stiffness (by a f actor of 3) a nd enhanced temperature capability. In the casting approach, a pa rticulate is dispersed uniformly in the matrix to give attractive levels of strength and stiffness, the l atter o f w hich i s c lose to qu asi-isotropic g raphite-Ep. The X D te chnique (X D i s a t rademark o f M artin-Marietta Corporation now Lockheed-Martin) is a process in which an in situ p recipitation te chnique i s u sed f or t he f ormation o f t he reinforcement dir ectly in t he m atrix o f in terest. U sing X D technology, it is possible to simultaneously produce a number of different reinforcements in a given matrix (Al, Ti, Cu, and intermetallics), thus allowing design of the microstructure to meet a specic ne ed, e .g.,  ne pa rticles f or s trength a nd w hiskers f or

10/3/2008 8:44:46 PM


Ultimate strength as a function of temperature

Strength ~ KSI

200

Silicon carbide/ aluminum

Silicon carbide/titanium

150 Unreinforced titanium 100 50 Unreinforced aluminum 0

200

400

600 800 1000 1200 1400 1600 Temperature (⬚F)

FIGURE 5.3 Ultimate tensi le st rength a s a f unction of temperature for monolithic and SiCf reinforced aluminum and titanium.

creep-rate reduction. Thus, XD technology is a very useful technique f or ac hieving de signer o r eng ineered m icrostructures. Such  exibility at t he m icrostructure le vel a llows c oncurrent design, i .e., t he si multaneous de sign o f t he m aterial a nd component. Titanium alloys reinforced with continuous ceramic bers such as SiC enhanced the strength (Figure 5.3) and modulus (by a factor of 2). It i s p rojected t hat r einforced ti tanium a luminides m ay b e useful up to 1000°C. The estimated strength value for unidirectional r einforced ti tanium a luminides w ith S iC  bers (SCS-6/ Ti xAlyM) is nearly twice that of Inconel 718, over twice that of Ti-6Al-4 V at 540°C, and double that of unreinforced titanium at 760°C.

Acknowledgment I thank my parents, brother, and dearest friend Dr. Ramakrishna for their support and encouragement to write this chapter.

5.2 Design of an Active Composite Wing Spar with Bending– Torsion Coupling Carlos Silva, Bruno Rocha, and Afzal Suleman 5.2.1

Introduction

Adaptive materials are nowadays a lready reliable for industrial applications a nd c an b e e asily i ntroduced i n s ystems ne wly designed, w ith t heir i ntroduction a lready t aken i nto c onsideration, or in older s ystems as a retrot. They have the advantage o f t heir app lication b eing e asily ac hieved, i ndependently of t he ba se m aterial, si nce f or i nstance t hey c an b e b onded (surface glued or embedded) to metal, carbon, composite, etc.

43722_C005.indd 7

5-7

Adaptive s ystems c an i mprove a ircraft re liability, re ducing failures through vibration control or reduction. Vibration reduction me ans d amage, w ear, a nd f ailure p robability re duction, which leads to larger intervals between maintenance operations. This directly leads to a high cost reduction. Aeronautic structures are exible, specially the lifting surfaces. As the airplane is ying, s tructural de formation t akes p lace. Th is leads to a c hange i n t he aerodynamic loads produced by them due to t he shape modication. The aerodynamics forces increment o r de crement c onsequently re sults i n a ne w s tructural deformation. Th is i s a t wo-way i nteraction b etween t he structure a nd t he  uid [1]. The b ehavior c an b e s table, i f a n equilibrium c ondition i s at tained, o r u nstable, w here c atastrophic f ailure c an o ccur. N onetheless, t he si mple v ibration phenomena w ill c ause i ts d amage d uring t he l ife o f t he structures. Vibration control, which provides vibration reduction, utter and d ivergence at tenuation, a nd del ay, c an b e ac hieved u sing actuators to s end p ulses to t he s tructure to opp ose i ts s ensed vibration based on a f requency response methodology. In addition, a multicell cross-section for wing spars can be used. These spars when built from  ber composite laminate materials, with the preferable  ber orientation misaligned with span direction, possess a  exure–torsion coupling. Th is means that when these spars are subjected to  exure, they will present some amount of torsion. For certain ber angles and multicellular cross-section congurations, certain t ypes a nd i ntensities of t hat coupling c an be achieved, such that it can be utilized to counteract aeroelastic de formations a nd ph enomena, a s  utter a nd d ivergence attenuation. Recently, at the Portuguese Air Force Aeronautics Laboratory, Suleman et al. have reported several studies on aeroelastic control and utter alleviation in adaptive ight vehicles [2–4]. These studies led to two active wing concepts: the active spar and active skin c oncepts a nd b oth w ind-tunnel a nd  ight tests were conducted u sing a remote p iloted v ehicle ( RPV) [5]. A sig nicant increase in t he  utter s peed a nd de crease i n t he a mplitude o f vibrations w as ob served. These c onclusions w ere v alidated f or realistic ight conditions [6,7] with the successful results, a analogous proportional control system was applied in this study. 5.2.1.1 Bending–Torsion Coupling Spars Piezoelectric (PZT) actuators and sensors are nowadays reliable for i ndustrial app lications. They h ave t he adv antage o f t heir application b eing e asily ac hieved, i ndependently o f t he ba se material. F or i nstance, t hey c an b e b onded ( surface g lued o r embedded) to me tal, c arbon, c omposite, e tc. The coupling between the actuator and base structure can be easily enhanced in composite structures if one uses a m isaligned layer stacking. If a  at b eam is considered, the misalignment is the ber angle deviation from the length direction. The misalignment can vary from 0° (aligned) to 90° (perpendicular alignment). A structure built with this type of internal layout when loaded vertically will

10/3/2008 8:44:46 PM

5-8

Coupling

Smart Materials

0

90 Degrees

FIGURE 5.4

Actuators and Sensors

The ideal actuator would have a high-energy (mechanical stress/ strain) ac tuation c apability w ith lo w energ y i nput ( voltage/ current), wide frequency range, linear and unidirectional behavior, and c onvenient s hape. I deal s ensors w ould h ave a no ise-free response, h igh p recision, a nd  delity ( low m echanical e nergy stimulation s ensitive), l inear b ehavior, w ide f requency s ensing range, and convenient shape. Both would be easily integrated or bonded into the structures. Due to t his t ype o f s tructure d imensions, lo w-frequency requirements, PZ T ac tuators a nd s ensors w ere s elected. F rom the v arious t ypes o f ac tuators a nd s ensors a vailable i n t he market, Q P40W f rom Q uickPack A CX w as s elected a s t he actuator a nd P ZT m ade of m aterial BM 500 f rom S ensor Technology Ltd. were chosen as the sensor.

5.2.2

Multicell Cross-Section Spar Design

One of t he challenges was to de sign, opt imize, manufacture, and te st a c ontinuous  lament m ulticell c ross-section c omposite s par w ith BTC a nd w ith emb edded PZ T ac tuators (Figure 5.5). The resulting multicell spar will have its upper wall composite bers o riented i n t he s ame d irection a s t he lo wer w all  bers (Figure 5. 5). W hen sub jected to b ending lo ads, t his m ulticell spar will present bending and also torsion deformations (BTC). Orienting the upper and lower wall bers from trailing edge to leading edge, from root to wing tip, when the wing spar bends upwards, it twists in a manner that the angle of attack is decreased (leading edge down). With this kind of orientation, a desirable behavior is at tained. Th is counteracts t he dy namical behavior of utter.

43722_C005.indd 8

Semispars and ber layout.

BTC evolution.

deform vertically and twist. The same effect can be attained if a torsion moment is applied. An inherent consequence is that the vertical d isplacement a nd to rsion a re no lo nger l inear w ith respect to the specic load applied. Early studies i n t his  eld were c arried out by G arnkle and Pastore [ 8]. They a nalyzed t he pa ssive a nd a ugmented ac tive behavior o f st ructures ex hibiting be nding–torsion co upling (BTC) as shown in Figure 5.4. 5.2.1.2

FIGURE 5.5

This i s a pa ssive c haracteristic o f t hese t ypes o f s pars. Using an ac tive s ystem, t his cha racteristic c an be augmented by app lying ac tuators o n t he upp er a nd lo wer m ulticell s par walls. In order to determine the misalignment, a global optimization approach through metamodeling was programmed. For the type of spar presented, considering its constraints and dened objective f unction, t he de sign v ariables s et w as c alculated [ 9]. The optimized solution dened the nal design to be manufactured, as follows: span, 1 m; height, 35 mm; wall thickness, 1 mm; and ber orientation, 10°. 5.2.2.1

Actuators and Sensors Positioning

In order to determine the most advantageous positions to place both actuators and sensors, computational analyses were carried out. With respect to vibration reduction, the location of sensors and actuators was determined by nite element method (FEM) analysis of the vibration modes. This methodology revealed the location of points with larger strain values (near the wing root). The advantage provided by this approach is that a small and light actuator manages to change dramatically the vibration parameters of the wing because it is placed near the points of higher strain. Also, the higher strains permit better sensing denition by t he s ensors due to i ts l arger a mplitudes. This positioning is also adv antageous f or t he f requency-based si mple h armonic motion (SHM). Despite t he fact t hat t he main present work i s not o n SH M, related c onstraints w ere t aken i nto ac count. Wi th re spect to wave propagation methods, it is useful to position the sensors as far away as possible from the actuators. Consequently, the spar was divided into two equal zones, from now on called cells, each one dened by a pair of actuators and one of sensors. The cells have the actuators on its root and sensors on its end on the spanwise direction. The upcoming disadvantage for the vibration reduction is only related to t he sensors amplitude output decrease, which can be easily overcome. 5.2.2.2

Active System

Figure 5. 6 s hows t he ac tive s ystem s cheme. St arting f rom t he PZT sensor, it is assumed that the PZT signal is proportional to

10/3/2008 8:44:46 PM

5-9


Piezoelectric actuator

Amplifier

Active wing

Signal conditioning D/A converter

Bending gain

Piezoelectric sensor

control s cheme, t he ac tuators w ill a lways counteract t he spar’s vertical movement.

Signal conditioning

5.2.3

Lowpass filter

Several experiments were conducted during the work progress. So, s par, w ing, a nd R PV  nal a ssembly w ere a ll te sted. The experiments comprehended static and dy namic on passive and active solutions. Results obtained include sensor output analysis, fast F ourier t ransform ( FFT), a nd d amping c alculations ( see Figure 5.8).

A/D converter

Controller

FIGURE 5.6

Proportional control scheme.

5.2.3.1 the spar’s vertical displacement. This signal has to be conditioned because the control, carried out by a MicroAutoBox DSpace control module, has to b e matched to t he I/O requirements. Inside the c ontrol mo dule, a M atlab Si mulink mo del, p reviously uploaded a nd r unning a s a  ash app lication o n t he c ontroller memory, i mplements t he d irect p roportional c ontrol l aw ( see Figure 5.7). A lo w pa ss  lter is included to retain only the frequencies below 50 Hz. The main concern is the low frequency present in the  rst b ending v ibration mo de a lso a voids i nterferences coming from the power current. Afterwards, a gain is applied to the i nput sig nal a nd i t i s  nally sent as an output back to the signal conditioning board. The re ferred ga in h as to b e de termined t hrough e xperiments. O nce a gain t he I /O h as to b e matched to the one required by the amplier, which by its turn, sends t he  nal sig nal to t he PZ T ac tuators. U sing t his si mple RTI Data

DS1401 Flight recorder

Results

Spar

The spar with the actuators and sensors was clamped on its root and an impact was given to t he tip. Figure 5.8 shows the sensor output time evolution. Using the data attained, the rst natural frequency a nd d amping c oefficient w ere de termined f or b oth situations. From the results shown, it is easily seen that the improvement is signicant. Up to 8 0% of d amping i ncrease c an b e achieved and up to 12%, rst natural frequency augment can be attained (see Table 5.2). 5.2.3.2 Wing Design and Fabrication After nishing the spar’s experiments, a wing was built with it to be tunnel tested. This wing was intended to be used in the future as a s pecimen c apable to a llow t he i mplementation o f SH M methods. Thus, t he mo del w as de signed t aking i nto c onsideration t wo SH M me thods, na mely v ibration ba sed a nd w ave

79944

1

.001 0.3999 79

SensorRec Filter on

12:34

(single)

DS1401 flight recorder

Digital clock

Data type conversion3

TimeRec

Input w/filt (single) Data type conversion1

2.18 Coarse gain

ADC

Sensor output1 ADC_type1_M1_con1

0.5

+ − + Clear input

X

+ +

0

Product

1

DAC

Active_gain

Filter off

DAC_type1_M1_C1 Final output

V offset

−1

+ +

−0.0068 Test-offset

Offset_sensor

Pulse generator

Gain1

Test-output + +

0

DAC

0 Test-gain

X

DAC_type1_M1_C2

Active switch + + Pulse generator1 1 Pulse offset

(single) 0 Active off

Multiport switch

Data type conversion2

1

DS1401 flight recorder

Active on

SwitchRec

FIGURE 5.7 SIMULINK control law model.

43722_C005.indd 9

10/3/2008 8:44:47 PM

5-10

Smart Materials

1 0.8

Dimensionalized sensor signal Active Passive

0.6 0.4 0.2 0 −0.2 0

0.05

0.1

−0.4

0.15

0.2 Time (s)

−0.6 −0.8 −1

FIGURE 5.8

Passive and active responses.

TABLE 5.2 Passive and Active Damping Coefficients and First Natural Frequencies Coefficient Passive spar Active spar Increase

Damping Frequency (Hz) 0.054 0.1 80%

First Natural Frequency 28.92 32.46 12%

propagation. One main feature of the wave propagation method application is that an attached wing spar rib or any kind of stiffener limits wave propagation. Thus, i n o rder to h ave s pace to place the actuators and sensors and to a llow the wave to propagate, t he gener al s tructure o f t his sm all w ing mo del f or w ind tunnel testing was divided only into two cells with a continuous wing s par. H owever, to a ssure t he a erodynamic s hape o f t he wing no nstructural r ibs ( not at tached to t he m ain s par) w ere positioned spanwise, 8 cm apart from each other. Beyond these two cells, there was a structural rib at the wing root and another one 4 cm apa rt s panwise, c onnected to e ach ot her by foam (in addition to the wing spar). This w as i ntended to f orm a rei nforced section to guarantee the wing–fuselage connection. Also, the wing skin was not structural and could be made of a t ransparent  lm, w hich a llowed s eeing t he i nternal w ing s tructure, actuators/sensors positioning, systems, etc. Figure 5.9 shows the wind tunnel testing apparatus. For e ach te st, a s ensor sig nal a mplitude o utput v ersus t ime sample was retrieved and an FFT analysis was done. For instance, Figure 5.10 re presents t he F FT re sults ob tained f or t he 2 0 m/s velocity for the passive wing and active wing with a gain of 50% max. A c lear de crease i n t he v ibration a mplitude ne ar f or t he rst natural frequency can be observed. Maximum signal output amplitudes were retrieved and Figure 5.11 shows the results. As shown, the amplitude decreased from passive to ac tive te sts. G enerally, t his a mplitude ten ds to b e smaller when a higher gain is applied. Figure 5.12 shows the percentage of maximum sensor signal output amplitude decrease: a

43722_C005.indd 10

maximum 8.83% decrease is reached for 30 m/s of velocity and a gain of 50% max. As regards the damping analysis, studies were concentrated around t he  rst nat ural f requency. Ha ving t he s ensor sig nal output, FFT a mplitude versus frequency graphs, damping was calculated, re trieving t he f requency v alue for t he p eak a mplitude a nd f requency v alues (around t he previous) obtained for maximum amplitude divided by the square root of 2. With these

FIGURE 5.9

Wind tunnel test setup.

10/3/2008 8:44:47 PM


FFT 20 m/s

0.012 Pass Act G5

Amplitude

0.01 0.008 0.006 0.004 0.002 0 0

20

40

60

80

100

Hz

FIGURE 5.10

The 20 m/s passive and active wing FFT analysis.

three values, it was possible to determine a d amping coefficient proportional v alue (not t he d amping c oefficient itself) by sub tracting the highest and lowest frequency values divided by the value of the central frequency. With this algorithm, all the damping coefficients were calculated and plotted on the same graph (Figure 5.13) along with the

5-11

respective qu adratic p olynomial t rendlines. These l ast o nes, allowed us to predict utter speed (that corresponds to the velocity f or t he z ero d amping c oefficient, i .e., w hen t he t rendlines would ideally cross the x/velocity axis). It is noted that the utter s peed i ncreases f rom pa ssive to active tests, and it reaches its highest value for the 50% max gain. Also, d amp i mprovements a re ob served f rom pa ssive to ac tive tests, and they are higher also for the 50% max gain. Overactuation and delays inuence the active gain results for gains higher than 50%, re sulting i n p oorer outcomes. Figure 5.14 s hows t he p ercentage damping improvements: a maximum 93% improvement is reached for 45 m/s of velocity and using 50% maximum gain. The maximum damp i mprovement, now obtained, is higher than t he o ne ob tained f or t he s par w hen i ts me chanical te sts were performed (80% max improvement). This increase can be explained by the BTC, which had no inuence on the spar behavior when its mechanical tests were done, but which is essential on wind tunnel tests. On these nal tests, this coupling affected the aeroelastic behavior of the wing. When the wing bends and consequently rotates along the spanwise axis due to the coupling, the aerodynamics of the wing will be inuenced, i.e., aerodynamic

Max amplitude 0.7

Signal amplitude

0.6 0.5

Pass

0.4

G25

0.3

G50 G100

0.2 0.1 0 20

30

40

45

m/s

FIGURE 5.11

Wing maximum amplitude.

%

Max amplitude decrease

10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00

G25 G50 G100

20

30

40

45

m/s

FIGURE 5.12

43722_C005.indd 11

Wing maximum amplitude decrease.

10/3/2008 8:44:48 PM

5-12

Smart Materials Damping

0.012 G25 G50 G100 Pass Poly. (G50) Poly. (G100) Poly. (G25) Poly. (Pass)

0.01 Damp

0.008 0.006 0.004 0.002 0 20

30

40

50

60

70

m/s

FIGURE 5.13

Damping for 25%, 50%, and 100% of maximum gain.

%

Damp improvement

100 90 80 70 60 50 40 30 20 10 0

G25 G50 G100

20

30

40

45

m/s

FIGURE 5.14 Damping improvement for 25%, 50%, and 100% of maximum gain.

loads applied and their distribution is altered, affecting the wing structure, cha nging co nsequently t he w ing st ructure deformations. 5.2.3.2.1 0°/10° Wing Performance Comparison In order to truly assess the ber misalignment behavior, a similar 0° ber-oriented spar was manufactured and tested. Figure 5.15 presents t he a mplitude v ariation f or b oth s pars at d ifferent

velocities. It can be inferred that the misalignment on spars (to a certain extent) by itself can lead to an improvement in the wing’s behavior. Afterwards, it can be clearly augmented by implementing active materials. 5.2.3.3

RPV: Flight Tests

After the successful wind-tunnel testing, the wing was prepared to be installed on an RPV for ight testing. It does not compromise

Max amplitude

Max amplitude decrease

35

1

30

0.8

25

10º 0º

0.6

%

Signal amplitude

1.2

20 15

0.4

10

0.2

5 0

0 20

35

40

45

m/s

FIGURE 5.15

43722_C005.indd 12

20

35

40

45

m/s

0° and 10° maximum amplitude and its comparison.

10/3/2008 8:44:49 PM


FIGURE 5.16 Flight testing modied wing.

the desired behavior but induces some mass increase. The wing must also be equipped with an aileron and respective servo action mechanism (Figure 5.16). Its root had to be reinforced in order to receive the clamping pin provided by the fuselage. To c arry o ut t he  ight te sts, t he a mplier also had to be changed. The one used on previous tests was adequate and was replaced by a sm all bat tery-powered one. Figures 5.17 a nd 5.18 show a ll t he ac tive e quipment ge ar o n b oard a nd i n-ground. During t he g round e quipment te sting, a n i nterference w as detected between the engine and the sensor system. Due to this condition, it was opted to do gliding ight testing. The procedure consisted in ying the RPV as high as possible, which w as l imited b y t he p ilot’s ob servation c apability. Afterwards the engine was cut off and a glide descent aimed to a speed of 20 m/s was performed. The airspeed was monitored in the P C u sing t he J et-Tronic s oft ware, w hich re ceived t he p ilot telemetry sig nal f rom onboard a nd t he te sting t ime w as b eing correlated with speed, and both values registered. This operation revealed that it was very difficult to conduct, even with two assistants, because of the difficulty in keeping constant conditions. Two c onsecutive  ights w ere c arried o ut. F igure 5.19 s hows the  ight data recorded, including the altitude gain. Figure 5.20

5-13

is focused on the glide part of both ights. The ramp functions represent the ight time. Flying the RPV must be done not very far from the pilot, and the a ircraft p erformed a s eries o f t urns t hat i nuenced the results. In the turns, effective wing lift increased and the vibrations w ere h igher. This m ay e xplain w hy s ometimes w hen t he control was ON (and was supposed to decrease vibrations), some very large amplitude peaks appeared. Due to this, no maximum or minimum comparison was done because t hey do not re present the general wing’s behavior. If one considers that a w indow represents a s tate of active or inactive, ight 1 had six windows and  ight 2 had ve windows. Selecting joined windows believed to represent the most equivalent  ight c onditions, w e ac hieved t he f ollowing F FT a nalysis, shown in Figures 5.21 and 5.22. A c lear i mprovement c an b e ob served. A nother i nteresting fact is that the improvement seems to be higher in the last part of both  ights, the  nal straight approach. This can be justied by the fact that the wind conditions were more severe near the ground, thus implying more vibrations on the wing. Adding to this, since the airplane was aligned with the runway less pilots, inputs were necessary to guide the RPV to the landing.

5.2.4

Concluding Remarks

The proposed solutions have resulted in a successful implementation of a ne w spar concept for a c onventional w ing. The proposed s olution c onsists o f a m ulticell c ross-section s par w ith misaligned  bers, w hich re sulted i n a n i mproved a eroelastic performance a nd supp ression o f v ibrations a nd c onsequent postponement of utter. The rst step in the process was to design the spar itself. The conceptual design of the spar indicated that it should have more than one section, and the ber orientation should be aligned at

FIGURE 5.17 RPV internal equipment placing.

43722_C005.indd 13

10/3/2008 8:44:50 PM

5-14

Smart Materials

FIGURE 5.18

Flight test setup on runway. 1.4

Sensor output (V)

1.2 1 0.8 0.6 0.4 0.2 0

3 Time

4

1

0.9

0.9

0.8

0.8

0.7

0.7

0.6 0.5 0.4 0.3

0.5 0.4 0.3 0.2

0.1

0.1

FIGURE 5.20

6 ⫻105

0.6

0.2

0

5

Flight test sensor pattern.

1

0

43722_C005.indd 14

2

Sensor output (V)

Sensor output (V)

FIGURE 5.19

1

0

1

2

3

4 Time

5

6

7

8 ⫻104

0

0

1

2

3

4 Time

5

6

7 ⫻104

First and second gliding  ights sensor output.

10/3/2008 8:44:51 PM

5-15


FFT 0.006 0.005

Passive 0.005

Active

Active

0.004 Amplitude

0.004 Amplitude

FFT

0.006

Passive

0.003 0.002

0.003 0.002 0.001

0.001 0

0

10

10

FIGURE 5.21

20

30 Hz

40

20

30

40

Hz

Second/third and  ft h/sixth windows passive/active FFT analysis.

FFT

FFT

0.006

0.006 Passive

Passive 0.005

0.005

Active

Amplitude

Amplitude

0.003

0.003

0.002

0.002

0.001

0.001 0

0 10

20

30 Hz

40

10

50

20

30 Hz

40

50

Second/third and fourth/ ft h windows passive/active FFT analysis.

an angle with respect to the spanwise direction. On a planview, these bers must start in the trailing edge and end in the leading edge such that the leading edge decreases its displacement when the wing experiences a lift load. For a simple constant rectangular cross-section spar with an overall constant thickness and manufactured with one material, this is the only way of achieving the desired BTC. Next, t he issue was to de cide t he  nal d imensions a nd t he right misalignment of the bers. On one hand, a small misalignment c ould re sult i n a p oor BTC b ut go od v ertical s tiffness and h igh nat ural b ending mo des f requencies. O n t he ot her hand, a h igh m isalignment w ould p resent a b etter c oupling but p oor s tiff ness a nd lo w nat ural b ending f requencies. T o determine t he b est s olution, a n opt imization p roblem w as de ned and solved.

43722_C005.indd 15

Active

0.004

0.004

FIGURE 5.22

50

−0.001

50

The u se of ac tive m aterials w as te sted r ight w ay at t his very early s tage of work. S ensors a nd ac tuators p ositions were c arefully studied. Having in mind future objectives such as structural health mo nitoring, ot her t han v ibration re duction a nd  utter suppression, two pairs of actuators were placed, one at t he root and t he ot her at t he m idspan. I n e ach c ase, o ne ac tuator w as glued on the upper and the other on the lower part of the spar. Two pa irs o f s ensors w ere a lso g lued to t he s par, b ut f or t his work, only one was necessary. Its position is in the midspan and both upper and lower positions were suitable for its task. Experimental ac tive te sts w ere t hen r un a nd c ompared to passive ones. The active methodology consisted of reading t he sensors strain and feed them back to the actuators, via a proportional c ontrol s ystem, to opp ose t he mo vement. Sub stantial damping a nd to rsional nat ural f requency i ncreases w ere

10/3/2008 8:44:53 PM

5-16

attained. This presented a go od mot ivation for continuing t he path followed so far. At t his stage, t he next m ilestone was identied. A w ing was built w ith o ne o f t he m anufactured s pars. Si nce t his s par w as intended to b e u sed f or f uture SH M me thods te sting, a w ing composed of two main cells was designed. One of these methods was ba sed o n w ave p ropagation. St arting f rom t he ro ot, t his wing had a reinforcement box for clamping in wind tunnel and the RPV fuselage. Then it had two cells. Each one was dened by two s tructural r ibs (attached to t he s par). B etween t hese t wo, there were “form” ribs, which were not attached and were placed to g uarantee t he a irfoil s hape a nd t ransmit t he a erodynamic loads to the edges. These were then transmitted to the spar by the attached structural ribs. This way, mechanical waves generated by the actuators could progress through the spar in each cell. Next, w ind t unnel te sts were c arried out. These were r un at several stabilized velocities, starting from 20 m/s to a maximum near the utter speed initiation behavior, approximately around 50 m/s. A similar wing was built at the same time, using a 0° spar to test and compare with the native behavior. The results showed that the native performance is clearly improved when applying a misalignment to the  bers. An approximately 30% reduction in the maximum vertical displacement amplitude was found for a test run at 45 m/s. It also presented a stable vertical displacement behavior a s t he v elocity w as i ncreased. The 0° spar tended to increase this parameter as the velocity rose. When applying active control using the piezo actuators, this aeroelastic performance of the wing was enhanced. A structural damping increase around 90% was attained for 45 m/s. The utter was predicted to be delayed by 8 m/s, which corresponds to a 16% increase. The best results were found when using a 50% gain increase in the proportional feedback control system. At t his s tage, t he s econd m ilestone w as c ompleted a nd t he main objective of the work satised, since the wind tunnel tests closely re produced t he s tructure’s re sponse to t he re al a irow inputs. Nevertheless,  ight testing can push t he limits and test other aspects and act as a determinant to assess the systems viability. So, the  nal milestone was to adapt this wing to the RPV and test it in ight. To do so, the control equipment had to be put onboard. The energ y s ource w as c hanged to bat teries a nd a n RPV, w hich i ndicated a irspeed me asuring e quipment, w as installed. The RPV was manually piloted from the ground. The control system was preprogrammed with the control law and set to save all the  ight data for future postprocessing. During the  ight, t he ac tive s ystem w as a utomatically s et o n a nd off in

43722_C005.indd 16

Smart Materials

intervals of 12 s. Pre ight tests revealed engine interferences of an u ndetermined source. Gliding  ights opt ion was t aken a nd spectral frequency analyses were carried out. The results showed a sig nicant v ibration re duction de spite t he re duced a vailable amplication provided by the internal amplier. This  nal work stage closed the loop of the spar development program and on the passive and active system implementation. As a  nal c onclusion, it c an b e s aid t hat b oth t he pa ssive a nd active multicell cross-section spars proved to b e a v iable solution in a ctive a eroelastic c ontrol s olutions in a daptive  ight vehicles.

References 1. 2.

3.

4.

5. 6.

7. 8. 9.

Bisplinghoff, R .L., Ashley, H., a nd H alfman, R .L., Aeroelasticity, Dover, New York, 1996. Suleman, A., Costa, A.P., and Moniz, P., Experimental utter and buffeting suppression using piezoelectric actuators and sensors, Proceedings of the Smart Structures and Materials— 3674, 1999. Amprikidis, M. and Cooper, J., Adaptive internal structures for active aeroelastic control, Proceedings of the International Forum o n Aeroelasticity a nd S tructural D ynamics—IFASD 2003, 2003. Rocha, J., Moniz, P., C osta, A., and Suleman, A., On ac tive aeroelastic control of an adaptive wing using piezoelectric actuators, Journal of AircraĀ, 42(1), pp 272-278, January— February, 2005. Costa, A., N ovel C oncepts f or P iezoelectric Actuated Adaptive A eroelastic Air craft S tructures, P hD thesis, Instituto Superior Técnico, Portugal, 2002. Rocha, J., Suleman, A., Costa, A., Moniz, P., and Santos, D., Research and development of an active aeroelastic adaptive ight demo nstrator, CEAS/AI AA/NVvL IF ASD 2003, Netherlands, 2003. Rocha, J ., An E xperimental S tudy a nd Flig ht T esting o f Active Aeroelastic Aircraft Wing S tructures, M aSC thesis, University of Victoria, Canada, 2005. Garnkle, M. a nd P astore, C., I ntrinsically sma rt co upled box b eams, Retrie ved February 5, 2007, f rom http://www. pages.drexel.edu/∼garnkm/Spar.html. Silva, C., Rocha, B., and Suleman, A., A metamodeling optimization approach to a wing spar design, 49th AIAA/ASME/ ASCE/AHS/ASC S tructures, S tructural D ynamics a nd Materials, Schaumburg IL, USA, 2008.

10/3/2008 8:44:53 PM

6 Ferromagnetic Shape Memory Alloy Actuators Yuanchang Liang University of Washington

Minoru Taya University of Washington

6.1 I ntroduction ...............................................................................................................................6-1 Review of Shape Memory Alloy • Ferromagnetic Shape Memory Alloy • Driving Mechanisms for FSMA-Based Actuators

6.2 FS MA-Based Actuator...............................................................................................................6-6 References ...........................................................................................................................................6-10

6.1 Introduction 6.1.1 Review of Shape Memory Alloy Shape memory alloys (SMAs), such as Ni–Ti (Nitinol) alloy, have a smaller hysteresis of (austenite Û martensite) phase transformation than that in steel [1]. Ā ere a re n umerous pap ers t hat have e xamined S MAs. Ā is is bec ause S MAs a re sci entically attractive a nd can be used as sensors a nd ac tuators. Ā e phase transformation in SMAs is reversible and usually can be characterized by t he latent heat a nd t he change of lattice pa rameters. To obtain reversible deformation, a heating or cooling device is usually incorporated. SMAs also have been used extensively as key actuator materials with applications in aerospace structures (such a s chevron) a nd me dical i mplant de vices (such a s s tent), where tem perature o r s tress i s a d riving f orce to ob tain l arge strain recovery (shape m emory eff ect [SME]) o r l arge “elastic” strain (superelastic behavior). Ā e stress–temperature phase diagram of SMAs, shown schematically in Figure 6.1, is often u sed to c haracterize t he ph ase transformation b etween s tress a nd tem perature. U sually, i t i s constructed by t he stress–strain curves tested at d ifferent temperatures. Ā e intersections with the temperature axis (stress = 0), TMs, TMf, TAs, a nd TAf a re u sually t aken f rom t he re sults b y differential scanning calorimetry (DSC), which are martensite at the s tart, m artensite at t he  nish, a ustenite at t he s tart, a nd austenite at the nish temperatures, respectively. Ā e se transformation temperatures are functions of stress where the larger the stress, t he h igher t he t ransformation tem perature. F igure 6 .2a and b s chematically s how lo ading a nd u nloading c urves of a n SMA tested at temperatures (T) higher than TAf and below TMs, respectively. As the SMA is loaded at T > TAf from the stress-free stage, O, t he stress–strain curve initially follows a s traight line

(OA) w ith t he s lope o f EA ( Young’s mo dulus o f t he a ustenite phase). Ā is corresponds to the elastic behavior of the SMA with the 100% austenite phase until it reaches the stress (sMs) of the onset s tress-induced m artensitic ( SIM) ph ase t ransformation (austenite → martensite), point A. Upon further loading, it follows a gradual straight line until point B where the curve becomes steep again in a slope of EM (Young’s modulus of the martensite phase). B etween p oints B a nd C , t he S MA i s u nder a n el astic deformation with 100% martensite phase until it is yielded at the yield stress (sy), point C. When the stress is over sy, t he 100% martensite phase of the SMA undergoes plastic deformation by a dislocation gliding mechanism. If under a constant temperature, T > T Af, the loading curve continues up to s ome point below sy followed b y unl oading, th e s tress–strain p ath will f ollow th e elastic u nloading a long t he s traight l ine ( CD) u ntil t he s tress reaches t he o nset o f re verse t ransformation ( martensite → austenite) at point D. Further unloading will render the reverse transformation b y f ollowing a g radual s traight l ine u ntil t he curve h its t he el astic lo ading l ine at p oint E . Ā e unloading beyond point E follows the elastic deformation path (EO) to the origin (O). Ā is loop (OABDEO) in Figure 6.2a is called “superelastic” (SE) behavior of the SMA where the strain at p oint B i s usually in the order of 5% for the polycrystal SMA. On the other hand, i f t he lo ading–unloading process i s c onducted at a c onstant temperature, T < T MS (Figure 6.2b), t he curve follows t he path OAB and leaves a residual strain (eres). Ā is residual strain is re coverable when t he temperature r ises over TAf because t he martensite to austenite phase transformation occurs. Ā e curve will follow the strain axis (BO) and return to the origin (O), i.e., zero strain. Ā is loop (OABO) is the so-called SME. eres is usually on the order of 5% for the polycrystal SMA. Unlike the path OA of SE in Figure 6.2a, the path OA in Figure 6.2b is not linear 6-1

43722_C006.indd 1

10/3/2008 8:45:24 PM

6-2 Stress (s)

Smart Materials

6.1.2 Ferromagnetic Shape Memory Alloy Martensite Austenite

TMf

TMs

TAs

TAf

Temperature (T )

FIGURE 6.1 Stress–temperature diagram of an SMA.

s C

sy

EM B A s Ms D E O

EA e

(a) s

A

B

O (b)

As de scribed b efore, t he ph ase t ransformation o f S MAs c an b e controlled by stress and temperature. Their relationship can be described by a stress–temperature phase diagram as shown in Figure 6.1. Recently, attention has been paid to some SMAs accompanying a change in ferromagnetism at phase transformation and having low hysteresis. Ā is is because the transformation characteristics suc h a s t ransition p oint ( martensite s tart tem perature, TMs) and macroscopically observed strain, caused by the transformation, are possibly controlled by an applied magnetic eld (H). Ā e response of transformation is fast in this case because the characteristic t ime i s governed by t he formation a nd g rowth of martensite, w hich i s i nduced b y a n app lied m agnetic  eld. It i s fast, thus, it is plausible to p roduce an actuator having reversible straining w ith qu ick re sponse to a sig nal i mposed o r de tected. Such an alloy with both ferromagnetic property and SMA behaviors is called ferromagnetic shape memory alloy (FSMA) and it is considered a s a s trong c andidate f or f ast re sponsive ac tuator material. Ā e rel ationship o f ph ase t ransformation, s tress ( s), temperature ( T), a nd m agnetic  eld (H) c an b e p resented a s a three-dimensional ph ase t ransformation d iagram a s s hown schematically in Figure 6.3. Currently, o nly a l imited n umbers o f F SMAs h ave b een found as listed in Table 6.1. Many researchers have extensively studied Ni MnGa, Fe–Ni–Co–Ti, Fe–Pt, a nd Fe–Pd FSMAs to examine SME and superelasticity. Ā ese alloys have also been considered as a possible candidate actuator materials. Among these a lloys, Ni MnGa [ 4,5] a nd F e–Ni–Co–Ti [ 20,21] a lloys have been known to exhibit good SMEs. However, the martensitic transformation start temperature (TMs) of the latter alloy is below − 150°C. I t i s to o lo w to b e p ractically u sed. Fu rther, Kakeshita e t a l. e ven re ported m agnetoelastic m artensitic transformation i n F e–Ni–Co–Ti b y u sing a s trongly app lied magnetic  eld up to 2 .3 × 1 07 A/ m. Ā ey also reported that martensitic transformation is not i nduced by an applied magnetic  eld u ntil 4 .8 × 1 06 A /m i s re ached [ 25]. Ā ese results show that a very strong applied magnetic eld is n eeded t o directly induce martensitic transformation. Ā is is not suitable

e res

e

FIGURE 6.2 Stress–strain curve of an SMA with (a) SE loop at T > TAf and (b) SME loop at T < TMs.

Stress (s )

Martensite Austenite

Temperature (T )

because it is not an elastic deformation. Ā e SMA is in the fully martensite phase at t he temperature below TMs and the loading path OA is involved in the martensite variant rearrangement. However, if the SMA is loaded beyond point A in Figure 6.2b, it is under the elastic deformation with a slope of EM similar to BC in Figure 6.2a.

43722_C006.indd 2

Magnetic field (H )

FIGURE 6.3 Schematics of t hree-dimensional phase transformation diagram of FSMAs.

10/3/2008 8:45:25 PM

6-3

Ferromagnetic Shape Memory Alloy Actuators TABLE 6.1 List of Current FSMAs Alloy

TMs

References

Ni2MnGa Fe–Ni–Co–Ti Fe-23–25Pt Fe-28–31Pd Fe–Pd–Ni NiAlCo

−150°C to 150°C 0. 52, t he structure c hanges f rom t he te tragonal 4 mm ph ase to a nother ferroelectric phase of rhombohedral 3 m symmetry. Figure 9.19 shows the dependence of several d constants on the composition near the morphotropic phase boundary between the tetragonal and rhom bohedral ph ases. The d c onstants h ave t heir h ighest values ne ar t he mo rphotropic ph ase b oundary. This enhancement in the piezoelectric effect is attributed to the increased ease of re orientation of t he polarization i n a n electric  eld. Doping the PZT material with donors or acceptor changes the properties dramatically. D onor dop ing w ith io ns suc h a s N b5+ or T a 5+ provides soft PZTs like PZT-5, because of the facility of domain motion due to the charge compensation of the Pb vacancy, which is gener ated d uring si ntering. O n t he ot her h and, ac ceptor doping w ith Fe 3+ or S c3+ le ads to h ard PZ Ts suc h a s PZ T-8 because oxygen vacancies pin the domain wall motion. 9.2.3.2.3 Lead Titanate PbTiO3 has a tetragonal structure at room temperature and has large tetragonality c/a = 1.063. The Curie temperature is 490°C. Densely si ntered P bTiO3 ceramics cannot be obtained easily because t hey b reak up i nto p owders w hen c ooled t hrough t he Curie temperature. This is pa rtly due to t he large spontaneous strain that occurs at the transition. PT ceramics modied by small amounts of additives exhibit high piezoelectric anisotropy. Either (Pb, S m)TiO3 [8] or (Pb, Ca)TiO3 [9] h as e xtremely lo w planar c oupling, t hat i s, a l arge kt/kp ratio. Here, kt and kp are thickness-extensional a nd p lanar ele ctromechanical c oupling

Orthorhombic

Tetragonal

Ps

Ps

Ps

Dielectric constant

Lead Zirconate–Lead Titanate

10,000

Cubic

ea

5,000

ec 0 −150

−100

−50

0

50

100

150

Temperature (⬚C)

FIGURE 9.17 Dielectric constants of BaTiO3 as a function of temperature. (Modied from Encyclopedia of Smart Materials.)

43722_C009.indd 16

10/3/2008 9:23:41 PM

9-17

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators 500 Cubic a

400

a

Temperature (⬚C)

a 300 Tetragonal

Morphotropic phase boundery

200 c

Ps

0

0 10 PbTiO3

a Ps

a

a

100

Rhombohedral

a

20

30

40

50

60

70

80

90

100 PbZrO3

PbZrO3 (Mole %)

Phase diagram of the PZT system. (Modied from Ceramics, transducers, in Encyclopedia of Smart Materials.)

factors, respectively. (Pb, Nd)(Ti, Mn, In)O3 ceramics that have a zero temperature coefficient of SAW delay have been developed as superior substrate materials for SAW devices [10]. 9.2.3.2.4 Relaxor Ferroelectrics Relaxor fe rroelectrics d iffer f rom nor mal fe rroelectrics; t hey have broad phase transitions from the paraelectric to t he ferroelectric state, strong frequency dependence of the dielectric constant (i.e., dielectric relaxation) and weak remanent polarization at temperatures close to t he dielectric maximum. Leadbased rel axor m aterials h ave c omplex d isordered p erovskite structures of the general formula Pb(B1, B 2)O3 (B1 = Mg 2+, Zn2+, Sc3+, B 2 = N b5+, Ta 5+, W 6+). The B si te c ations a re d istributed

randomly in the crystal. The characteristic of a relaxor is a broad and fr equency di spersive di electric m aximum. I n a ddition, relaxor-type m aterials suc h a s le ad m agnesium n iobate Pb(Mg1/3Nb2/3)O3–lead titanate PbTiO3 solid solution [PMN-PT] exhibit electrostrictive phenomena that are suitable for actuator applications. Figure 9.20 shows a n electric-eld-induced strain curve t hat w as ob served f or 0. 9PMN-0.1PT a nd re ported b y Cross et al. in 1980 [11]. Note that a strain of 0.1% can be induced by an electric eld a s sm all a s 1 k V/mm a nd t hat h ysteresis i s negligibly small for this electrostriction. Because ele ctrostriction i s t he s econdary ele ctromechanical coupling observed in cubic structures, in principle, the charge is

1.00 800

dij (⫻ 10−12 C/N)

600 d15

Strain (⫻10−3)

FIGURE 9.18

0.50 400 d33 200 −d31 0 48

50

52 54 56 PbZrO3 (Mole %)

58

60

FIGURE 9.19 Piezoelectric d st rain coeffi c ients versus composition for t he PZ T s ystem. ( Modied f rom Ce ramics, t ransducers, i n Encyclopedia of Smart Materials.)

43722_C009.indd 17

−15

−10

−5 0 5 10 Electric field (kV/cm)

15

FIGURE 9.20 Field-induced electrostrictive strain in 0.9PMN-0.1PT. (Modied f rom Ce ramics, t ransducers, i n Encyclopedia of Sm art Materials.)

10/3/2008 9:23:41 PM

9-18

Smart Materials

not i nduced u nder applied s tress. The converse electrostrictive effect, which can be used for sensor applications, means that the permittivity (rst der ivative of p olarization w ith re spect to a n electric eld) is changed by stress. In relaxor ferroelectrics, the piezoelectric effect can be induced under a bias eld, that is, the electromechanical coupling factor kt varies as the applied DC bias eld changes. As the DC bias eld increases, the coupling increases and saturates. The se materials can be used for ultrasonic transducers that are tunable by a bias eld [12]. The re cent d evelopment of si ngle-crystal pie zoelectrics started in 1981, when Kuwata, Uchino, and Nomura rst reported a n e normously l arge e lectromechanical c oupling f actor k33 = 92%–95% a nd a p iezoelectric constant d33 = 1 500 pC/N i n solid-solution si ngle c rystals b etween rel axor a nd no rmal ferroelectrics , Pb(Zn1/3Nb2/3)O3-PbTiO3 [1 3]. A fter a bout 1 0 years, Y. Yamashita et al. (Toshiba) and T. R. Shrout et al. (Penn State) independently reconrmed that these values are true, and much more improved data were obtained in these several years, aimed at medical acoustic applications [14]. Important data have been ac cumulated f or Pb( Mg1/3Nb2/3)O3 [P MN], P b(Zn1/3Nb2/3) O3 [PZN], and binary systems of these materials combined with PbTiO3 ( PMN-PT a nd PZ N-PT) f or ac tuator app lications. Strains as large as 1.7% can be induced practically for a morphotropic phase boundary composition of the PZN-PT solid-solution single crystals. Figure 9.21 shows the eld i nduced strain curve for [ 001] o riented 0. 92PZN-0.08PT [14]. I t i s not able t hat t he highest v alues a re ob served f or a rh ombohedral c omposition only when the single crystal is poled along the perovskite [001] axis, not along the [111] spontaneous polarization axis. 9.2.3.3 Polymers PVDF or PVF2 is a piezoelectric when stretched during fabrication. Thin sheets of the cast polymer are drawn and stretched in the plane of the sheet in at least one direction and frequently also

in t he p erpendicular d irection to c onvert t he m aterial i nto i ts microscopically polar phase. Cr ystallization from a mel t forms the no npolar α-phase, w hich c an b e c onverted i nto a nother polar β-phase by uniaxial or biaxial drawing; these dipoles are then reoriented by electric poling. Large sheets can be manufactured a nd t hermally f ormed i nto c omplex s hapes. C opolymerization of vinilydene diuoride with triuoroethylene (TrFE) results i n a r andom copolymer (PVDF-TrFE) t hat has a s table, polar β phase. This polymer does not need to be stretched; it can be poled directly as formed. The thickness-mode coupling coefcient of 0.30 has been reported. Such piezoelectric polymers are used for directional microphones and ultrasonic hydrophones. 9.2.3.4 Composites Piezocomposites co mprised o f p iezoelectric ce ramics a nd polymers are promising materials because of excellent tailored properties. The geometry of two-phase composites can be classied ac cording to t he c onnectivity of e ach ph ase (0, 1, 2 , or 3 dimensionality) into 10 structures; 0-0, 0-1, 0-2, 0-3, 1-1, 1-2, 1-3, 2-2, 2-3, and 3-3 [15]. A 1-3 piezocomposite, or PZT–rod/ polymer–matrix composite is a most promising candidate. The advantages of this composite are high coupling factors, low acoustic i mpedance (square ro ot of t he product of its den sity and elastic stiff ness), a go od match to w ater or human t issue, mechanical exibility, a broad bandwidth in combination with a low mechanical quality factor, and the possibility of making undiced a rrays by s tructuring only t he ele ctrodes. The thickness-mode ele ctromechanical c oupling o f t he c omposite c an exceed the kt (0.40–0.50) of the constituent ceramic and almost approaches the value of the rod-mode electromechanical coupling, k33 (0.70–0.80), of that ceramic [16]. The acoustic match to tissue or water (1.5 Mrayls) of typical piezoceramics (20–30 Mrayls) i s s ignicantly i mproved b y f orming a c omposite structure, that is, by replacing a heavy, stiff ceramic by a light, soft p olymer. P iezoelectric c omposite m aterials a re e specially useful for underwater sonar and medical diagnostic ultrasonic transducers.

2.0 d33 ~ 480 pC/N

9.2.3.5

Tetragonal

Strain (%)

1.5

1.0

0.5

0.0

0

20

60 80 40 Electric field (kV/cm)

100

120

FIGURE 9.21 Field-induced strain curve for [001] oriented 0.92PZN0.08PT. (Modied from Encyclopedia of Smart Materials.)

43722_C009.indd 18

Thin Films

Both zinc oxide (ZnO) and aluminum nitride (AlN) are simple binary c ompounds t hat h ave W urtzite-type s tructures, w hich can be sputter-deposited in a c-axis-oriented thin lm on a variety of substrates. ZnO has reasonable piezoelectric coupling and its thin lms are widely used in bulk acoustic and SAW devices. The fabrication of h ighly c-axis oriented ZnO lms has been extensively studied and developed. The performance of ZnO devices is, h owever, l imited d ue to t heir sm all p iezoelectric c oupling (20%–30%). PZT thin lms are expected to exhibit higher piezoelectric p roperties. A t p resent, t he g rowth o f PZ T t hin  lm is being carried out for use in microtransducers and microactuators. A s eries o f t heoretical c alculations o n p erovskite-type ferroelectric crystals suggests that large d and k values of magnitudes similar to t hose of PZN-PT can also be expected in PZT. Crystal orientation de pendence of piezoelectric properties w as

10/3/2008 9:23:42 PM

9-19

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

d33 Ps

E1

Ps

Ps

E1

E E2

Strain

d 33eff

E

d15

Ps

E2

FIGURE 9.22 Materials.)

Principle of the enhancement in electromechanical couplings in a perovskite piezoelectric. (Modied from Encyclopedia of Smart

phenomenologically c alculated for c ompositions a round t he morphotropic phase boundary of PZT [17]. The maximum longitudinal p iezoelectric c onstant d33 ( four to  ve t imes t he enhancement) a nd t he ele ctromechanical c oupling f actor k33 (more than 90%) in the rhombohedral composition were found at angles of 57° and 51°, respectively, canted from the spontaneous p olarization d irection [111], w hich c orrespond ro ughly to the perovskite [100] axis. Figure 9.22 shows the principle of the enhancement in electromechanical couplings. Because the shear coupling d15 is the highest in perovskite piezoelectric crystals, the applied eld should be canted from the spontaneous polarization direction to obtain the maximum s train. Ep itaxially g rown, [ 001] o riented t hin/thick lms using a rhomboherial PZT composition reportedly enhance the effective piezoelectric constant by—four to ve times.

9.2.4

Applications of Piezoelectricity

Piezoelectric materials can provide coupling between electrical and mechanical energy and thus have been extensively used in a variety o f ele ctromechanical de vices. The dir ect p iezoelectric effect is most obviously used to generate charge or high voltage in applications such as the spark ignition of gas in space heaters, cooking stoves, and cigarette lighters. Using the converse effect, small mechanical displacements and vibrations can be produced in actuators by applying an electric eld. Acoustic and ultrasonic vibrations can be generated by an alternating  eld tuned at t he mechanical re sonance f requency of a p iezoelectric de vice, a nd can b e de tected by a mplifying t he  eld gener ated by v ibration incident o n t he m aterial, w hich i s u sually u sed f or u ltrasonic transducers. A nother i mportant app lication o f p iezoelectricity is f requency c ontrol. The application of piezoelectric m aterials ranges o ver m any t echnology f ields, i ncluding u ltrasonic

43722_C009.indd 19

transducers, ac tuators a nd USMs; ele ctronic c omponents such as resonators, wave  lters, delay lines, SAW devices, and transformers a nd h igh-voltage app lications; ga s ig nitors, u ltrasonic cleaning, a nd m achining. P iezoelectric-based s ensors, f or instance, ac celerometers, a utomobile k nock s ensors, v ibration sensors, s train g ages, a nd  ow meters have been developed because pressure and vibration can be directly sensed as electric signals t hrough t he p iezoelectric e ffect. E xamples o f t hese applications are given in the following sections. 9.2.4.1

Pressure Sensor/Accelerometer/Gyroscope

The ga s ig niter i s one of t he ba sic applications of piezoelectric ceramics. Very high voltage generated in a piezoelectric ceramic under applied mechanical stress can cause sparking a nd ig nite the gas. Piezoelectric ceramics can be employed as stress sensors and acceleration sensors, because of their “direct piezoelectric effect.” Kistler (Switzerland) i s m anufacturing a 3 -D s tress s ensor. By combining a n app ropriate n umber o f qu artz c rystal p lates (extensional a nd shear t ypes), t he multilayer de vice c an de tect three-dimensional stresses [18]. Figure 9.23 shows a cylindrical gyroscope commercialized by NEC-Tokin (Japan) [19]. The cylinder has six divided electrodes; one pa ir i s u sed to e xcite t he f undamental b ending v ibration mode and the other two pairs are used to detect the acceleration. When rotation acceleration is applied around the axis of this gyro, the voltage generated on the electrodes is modulated by the Coriolis force. By subtracting the signals between the two sensor electrode pairs, a voltage directly proportional to the acceleration can be obtained. This type of gyroscope has been widely installed in handheld video cameras to monitor the inevitable hand vibration during operation and to compensate for it electronically on a display by using the sensed signal.

10/3/2008 9:23:42 PM

9-20

Smart Materials Piezoelectric element Matching layer Backing

Vibrator

Lead

Holder Ultrasonic beam

Support

FIGURE 9.23 Cylindrical gyroscope commercialized by NEC-Tokin (Japan). (Modied from Encyclopedia of Smart Materials.)

9.2.4.2 Ultrasonic Transducer One of the most important applications of piezoelectric materials is based on ultrasonic echo  eld [20,21]. Ultrasonic transducers convert electrical energy into a mechanical form when generating an acoustic pulse and convert mechanical energy into an electrical sig nal w hen de tecting i ts e cho. N owadays, p iezoelectric transducers a re b eing u sed i n me dical u ltrasound f or c linical applications t hat r ange f rom d iagnosis to t herapy a nd su rgery. They are also used for underwater detection, such as sonars and sh nders, and nondestructive testing. The u ltrasonic tr ansducers o ften op erate i n a p ulse-echo mode. The transducer converts electrical input into an acoustic wave o utput. The t ransmitted w aves p ropagate i nto t he b ody, and echoes are generated that travel back to be received by the same transducer. These echoes vary in intensity according to the type of tissue or body structure, and thereby create images. An ultrasonic i mage re presents t he me chanical p roperties o f t he tissue, such as density and elasticity. We can recognize anatomical structures in an ultrasonic image because the organ boundaries and uid-to-tissue interfaces are easily discerned. The ultrasonic imaging can also be done in real time. This means that we can follow rapidly moving structures such as heart without motional distortion. In addition, ultrasound is one of the safest diagnostic imaging techniques. It does not use ionizing radiation like x-rays and t hus i s ro utinely u sed f or f etal a nd ob stetrical i maging. Useful a reas for u ltrasonic i maging i nclude cardiac structures, the v ascular s ystem, t he f etus, a nd a bdominal o rgans suc h a s liver and kidney. In brief, it is possible to s ee inside the human body by using a beam of ultrasound without breaking the skin. There a re v arious t ypes o f t ransducers u sed i n u ltrasonic imaging. Mechanical s ector t ransducers c onsist of si ngle, rel atively large resonators that provide images by mechanical scanning such as wobbling. Multiple element array transducers permit the imaging s ystems to ac cess d iscrete el ements i ndividually a nd

43722_C009.indd 20

Input pulse

FIGURE 9.24 Geometry of t he fundamental transducer for a coustic imaging. (Modied from Encyclopedia of Smart Materials.)

enable electronic focusing in the scanning plane at various adjustable p enetration de pths b y u sing ph ase del ays. The two basic types of array transducers are linear and phased (or sector). Linear array transducers are used for radiological and obstetrical e xaminations, a nd ph ased a rray t ransducers a re u seful f or cardiological applications where positioning between the ribs is necessary. Figure 9.24 shows t he geometry of t he basic u ltrasonic t ransducer. The transducer is composed mainly of matching, piezoelectric material, and backing layers [22]. One or more matching layers are used to increase sound transmissions into tissues. The backing is attached to the transducer rear to damp the acoustic return wave and to reduce the pulse duration. Piezoelectric materials are used to g enerate a nd de tect u ltrasound. I n g eneral, b roadband transducers should be used for medical u ltrasonic i maging. The broad ba ndwidth re sponse c orresponds to a s hort p ulse leng th that results in better axial resolution. Three factors are important in de signing b road ba ndwidth t ransducers. The rst is acoustic impedance m atching, t hat i s, e ffectively co upling t he aco ustic energy to the body. The second is high electromechanical coupling coefficient o f t he t ransducer. The t hird i s ele ctrical i mpedance matching, t hat is, effectively coupling electrical energy from the driving electronics to the transducer across the frequency range of interest. The op erator of pu lse-echo t ransducers i s b ased on t he thickness mo de re sonance o f t he p iezoelectric t hin p late. The thickness mode coupling coefficient, kt, is related to the efficiency of converting electric energy into acoustic and vice versa. Further, a low planar mode coupling coefficient, kp, is benecial for limiting e nergies fr om b eing e xpended in a n onproductive l ateral mode. A l arge d ielectric c onstant i s ne cessary to ena ble a go od electrical impedance match to the system, especially in tiny piezoelectric sizes. Table 9 .5 c ompares t he p roperties o f u ltrasonic t ransducer materials [7,23] Ferroelectric ceramics, such as PZT and modied PT, a re a lmost u niversally u sed a s u ltrasonic t ransducers. The success of ceramics is due to their very high electromechanical

10/3/2008 9:23:42 PM

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators TABLE 9.5 Comparison of the Properties of Ultrasonic Transducer Materials PZT Ceramic PVDF Polymer kt Z (Mrayls) e33T/e0 tan d (%) Qm r (g/cm3)

PZT–Polymer Composite

ZnO Film 0.20–0.30 35 10

0.45–0.55 20–30 200–5000

0.20–0.30 1.5–4 10

0.60–0.75 4–20 50–2500

Smart Materials

Smart Materials

Smart Materials

Smart Materials

Smart Materials

Encyclopedia of smart materials

Smart electronic materials

Encyclopedia of Smart Materials Set

Smart Materials and New Technologies

Smart Materials And Structures: New Research

Smart Materials and Technologies in Architecture

Smart Electronic Materials: Fundamentals and Applications

Smart Materials and Technologies in Architecture

Nanoengineering of Structural, Functional and Smart Materials

Smart Materials for Energy, Communications and Security

Smart materials and technologies in architecture

Smart

World Forum on Smart Materials and Smart Structures Technology: Proceedings of SMSST'07, World Forum on Smart Materials and Smart Structures Technology ... in Engineering, Water and Earth Sciences)

Smart Light-Responsive Materials: Azobenzene-Containing Polymers and Liquid Crystals

Ambient Intelligence with Microsystems: Augmented Materials and Smart Objects (Microsystems)

Integrating ERP, CRM, Supply Chain Management, and Smart Materials

Smart Light-Responsive Materials: Azobenzene-Containing Polymers and Liquid Crystals

Smart materials and new technologies: for architecture and design professions

Integrating Erp, Crm, Supply Chain Management, and Smart Materials

Smart Textiles for Medicine and Healthcare: Materials, Systems and Applications

Smart Materials in Architecture, Interior Architecture and Design

Smart Business (Smart Things to Know About (Stay Smart!) Series)

Smart Business (Smart Things to Know About (Stay Smart!) Series)

Smart Leadership (Smart Things to Know About (Stay Smart!) Series)

Smart Dragons

Smart Discipline

Smart Materials

Smart Materials

Smart Materials

Smart Materials

Smart Materials

Encyclopedia of smart materials

Smart electronic materials

Encyclopedia of Smart Materials Set

Smart Materials and New Technologies

Smart Materials And Structures: New Research

Smart Materials and Technologies in Architecture

Smart Electronic Materials: Fundamentals and Applications

Smart Materials and Technologies in Architecture

Nanoengineering of Structural, Functional and Smart Materials

Smart Materials for Energy, Communications and Security

Smart materials and technologies in architecture

Smart

World Forum on Smart Materials and Smart Structures Technology: Proceedings of SMSST'07, World Forum on Smart Materials and Smart Structures Technology ... in Engineering, Water and Earth Sciences)

Smart Light-Responsive Materials: Azobenzene-Containing Polymers and Liquid Crystals

Ambient Intelligence with Microsystems: Augmented Materials and Smart Objects (Microsystems)

Integrating ERP, CRM, Supply Chain Management, and Smart Materials

Smart Light-Responsive Materials: Azobenzene-Containing Polymers and Liquid Crystals

Smart materials and new technologies: for architecture and design professions

Integrating Erp, Crm, Supply Chain Management, and Smart Materials

Smart Textiles for Medicine and Healthcare: Materials, Systems and Applications

Smart Materials in Architecture, Interior Architecture and Design

Smart Business (Smart Things to Know About (Stay Smart!) Series)

Smart Business (Smart Things to Know About (Stay Smart!) Series)

Smart Leadership (Smart Things to Know About (Stay Smart!) Series)

Smart Dragons

Smart Discipline

Recommend Documents