Poly Lactic Acid E-book
Upcoming SlideShare
Loading in...5
×
 

Poly Lactic Acid E-book

on

  • 3,818 views

 

Statistics

Views

Total Views
3,818
Views on SlideShare
3,818
Embed Views
0

Actions

Likes
0
Downloads
191
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

Poly Lactic Acid E-book Poly Lactic Acid E-book Document Transcript

  • POLY(LACTIC ACID)
  • Wiley Series on Polymer Engineering and Technology Richard F. Grossman and Domasius Nwabunma, Series Editors Polyolefin Blends Edited by Domasius Nwabunma and Thein Kyu Polyolefin Composites Edited by Domasius Nwabunma and Thein Kyu Handbook of Vinyl Formulating, Second Edition Edited by Richard F. Grossman Total Quality Process Control for Injection Molding, Second Edition M. Joseph Gordon, Jr. Microcellular Injection Molding Jingyi Xu Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications Edited by Rafael Auras, Loong-Tak Lim, Susan E.M. Selke, and Hideto Tsuji
  • POLY(LACTIC ACID) Synthesis, Structures, Properties, Processing, and Applications Edited by RAFAEL AURAS LOONG-TAK LIM SUSAN E. M. SELKE HIDETO TSUJI
  • Copyright Ó 2010 John Wiley & Sons, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. Library of Congress Cataloging-in-Publication Data: Poly(lactic acid) : synthesis, structures, properties, processing, and applications/ edited by Rafael Auras ... [et al.]. p. cm. Includes index. ISBN 978-0-470-29366-9 (cloth) 1. Biodegradable plastics. 2. Lactic acid. 3. Polymers. I. Auras, Rafael. TP1180.B55P65 2010 620.1’92323–dc22 2010006494 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1
  • CONTENTS PREFACE xvii CONTRIBUTORS xxi PART I 1 CHEMISTRY AND PRODUCTION OF LACTIC ACID, LACTIDE, AND POLY(LACTIC ACID) Production and Purification of Lactic Acid and Lactide 1 3 Wim Groot, Jan van Krieken, Olav Sliekersl, and Sicco de Vos 1.1 Introduction 1.2 Lactic Acid 1.2.1 History of Lactic Acid 1.2.2 Physical Properties of Lactic Acid 1.2.3 Chemistry of Lactic Acid 1.2.4 Production of Lactic Acid by Fermentation 1.2.5 Downstream Processing/Purification of Lactic Acid 1.2.6 Quality/Specifications of Lactic Acid 1.3 Lactide 1.3.1 Physical Properties of Lactide 1.3.2 Production of Lactide 1.3.3 Purification of Lactide 1.3.4 Quality and Specifications of Polymer-Grade Lactide 1.3.5 Concluding Remarks on Polymer-Grade Lactide References 2 3 4 4 4 4 5 8 10 10 10 11 13 14 16 16 Chemistry and Thermodynamic Properties of Lactic Acid and Lactide and Solvent Miscibility 19 Zhengyu Jin, Yaoqi Tian, and Jinpeng Wang 2.1 General Properties 2.1.1 Physical and Chemistry Properties of Lactic Acid 2.1.2 Physical and Chemical Properties of Lactide 2.2 Thermodynamic Properties 2.2.1 Vapor Pressures of Lactic Acids at Different Temperatures 19 19 19 19 19 v
  • vi CONTENTS 2.2.2 Temperature Dependence of Densities of Lactic Acid 2.2.3 Temperature Dependence of Viscosity of Lactic Acid 2.2.4 Thermodynamic Properties 2.3 Miscibility Properties of Lactic Acid and Lactide 2.3.1 Miscibility of Lactic Acid with Different Solvents 2.3.2 Miscibility of Lactic Acid with Modifiers in Diluents 2.3.3 Physical and Chemical Equilibrium of Lactic Acid 2.3.4 Miscibility of Lactide with Solvents References 3 20 20 21 21 21 21 22 25 25 Industrial Production of High Molecular Weight Poly(Lactic Acid) 27  Anders S€dergard and Mikael Stolt o 3.1 Introduction 3.2 Lactic Acid Based Polymers by Polycondensation 3.2.1 Direct Condensation 3.2.2 Solid-State Polycondensation 3.2.3 Azeotropic Dehydration 3.3 Lactic Acid Based Polymers by Chain Extension 3.3.1 Chain Extension with Diisocyanates 3.3.2 Chain Extension with Bis-2-oxazoline 3.3.3 Dual Linking Processes 3.3.4 Chain Extension with Bis-epoxies 3.4 Lactic Acid Based Polymers by Ring-Opening Polymerization 3.4.1 Polycondensation Processes 3.4.2 Lactide Manufacturing 3.4.3 Ring-Opening Polymerization References 4 27 28 29 30 31 32 32 33 34 34 34 35 35 36 37 Design and Synthesis of Different Types of Poly(Lactic Acid) 43 Ann-Christine Albertsson, Indra Kumari Varma, Bimlesh Lochab, Anna Finne-Wistrand, and Kamlesh Kumar 4.1 Introduction 4.2 Copolymerization 4.2.1 Synthesis of Copolymers of Lactic Acid: Glycolic Acid 4.2.2 Synthesis of Copolymers of Lactic Acid: Poly(ethylene glycol) 4.2.3 Synthesis of Copolymers of Lactic Acid: d-Valerolactone and Lactic Acid: b-Butyrolactone 4.2.4 Synthesis of Copolymers of Lactic Acid: e-Caprolactone 4.2.5 Synthesis of Copolymers of Lactic Acid: 1,5-Dioxepan-2-one 4.2.6 Synthesis of Copolymers of Lactic Acid: Trimethylene Carbonate 4.2.7 Synthesis of Copolymers of Lactic Acid: Poly(N-isopropylacrylamide) 4.2.8 Synthesis of LA: Alkylthiophene (P3AT) Copolymers 4.2.9 Functional Poly(lactic acid) 4.2.10 Branched Copolymers 4.3 Properties of Copolymers 4.3.1 Degradation of Homo- and Copolymers 4.3.2 Drug Delivery from PLLA Copolymers 4.3.3 Radiation Effects References 43 43 44 44 45 46 46 46 47 47 47 48 53 54 54 55 55
  • CONTENTS 5 Structure and Properties of Stereocomplex-Type Poly(lactic acid) vii 59 Masayuki Hirata and Yoshiharu Kimura 5.1 5.2 5.3 5.4 Introduction Formation of Stereocomplex Crystals Thermal Properties of sc-PLA Crystal Structure of sc-PLA 5.4.1 Unit Cell Parameters and Molecular Conformation of sc-PLA 5.4.2 Density and Heat of Fusion of Sccrystals 5.5 Formation of sb-PLA 5.5.1 ROP Routes to Diblock and Multiblock sb-PLA 5.5.2 SSP Routes to Multiblock sb-PLA 5.6 Applications of sc-PLA References PART II 6 PROPERTIES OF POLY(LACTIC ACID) Chemical Structure of Poly(lactic acid) 59 59 60 60 60 60 61 61 63 64 65 67 69 Xue Jiang, Yan Luo, Xiuzhi Tian, Dan Huang, Narendra Reddy, and Yiqi Yang 6.1 Introduction 6.2 Chain Structure and Configuration 6.2.1 Chain Structure 6.2.2 Configuration 6.2.3 Interlocked Structure, Polymer Blend, and Resistance to Hydrolysis 6.3 Syndiotactic Polymerization and Syndiotacticity 6.4 Conformation 6.5 Amorphous Structure and Thermal Properties 6.5.1 Amorphous and Three-Phase Models 6.5.2 Glass Transition 6.6 Orientation Structure of PLA 6.6.1 Mechanical Orientation by Stretching or Compression 6.6.2 Thermal Orientation (Phase Transition) 6.7 Semicrystalline Structure 6.7.1 General 6.7.2 Three Forms (a, b and c) of the Crystal Structure 6.8 Frustrated Structure 6.9 Molecular Weight 6.10 Summary References 7 69 69 69 70 71 72 74 74 74 75 77 77 77 78 78 78 79 80 80 80 Chemical Compatibility of Poly(lactic acid): A Practical Framework Using Hansen Solubility Parameters 83 Steven Abbott 7.1 A Practical Framework 7.1.1 Thermodynamics Versus Kinetics 7.1.2 Hansen Solubility Parameters 7.2 Solvent Compatibility 7.3 Plasticizers 83 83 83 84 86
  • viii CONTENTS 7.4 7.5 7.6 7.7 Polymer Compatibility Environmental Stress Cracking Rational Composite Design Diffusion and Barrier Properties 7.7.1 Gases 7.7.2 Water 7.8 Pharmacological Transport 7.9 Summary References 8 87 89 89 90 91 91 93 93 94 Optical Properties 97 ˜ Carla M. B. Goncalves, Joao A. P. Coutinho, and Isabel M. Marrucho ¸ 8.1 8.2 8.3 8.4 8.5 Introduction Absorption and Transmission of UV-Vis Radiation Refractive Index Specific Optical Rotation Infrared and Raman Spectroscopy 8.5.1 Infrared Spectroscopy 8.5.2 Raman Spectroscopy 8.6 1H and 13C NMR Spectroscopy References 9 97 97 99 100 100 101 105 108 111 Crystallization and Thermal Properties 113 Luca Fambri and Claudio Migliaresi 9.1 Introduction 9.2 Crystallinity and Crystallization 9.3 Crystallization Regime 9.4 Fibers 9.5 Hydrolytic Degradation References 10 113 114 118 119 121 123 Rheology of Poly(lactic acid) 125 John R. Dorgan 10.1 Introduction 10.2 Fundamental Chain Properties from Dilute Solution Viscometry 10.2.1 Unperturbed Chain Dimensions 10.2.2 Real Chains 10.2.3 Solution Viscometry 10.2.4 Viscometry of PLA 10.3 Processing of PLA: General Considerations 10.4 Melt Rheology: An Overview 10.5 Processing of PLA: Rheological Properties 10.6 Conclusions Appendix 10.A Description of the Software References 125 126 126 126 127 128 130 131 132 137 138 138
  • CONTENTS 11 Mechanical Properties ix 141 Gabriele Perego and Gian Domenico Cella 11.1 Introduction 11.2 General Mechanical Properties and Molecular Weight Effect 11.2.1 Tensile and Flexural Properties 11.2.2 Impact Resistance 11.2.3 Hardness 11.3 Temperature Effect 11.4 Annealing 11.5 Orientation 11.6 Stereoregularity 11.7 Plasticization 11.8 Relaxation and Aging 11.9 Conclusions References 12 141 141 141 143 143 143 144 147 148 149 151 152 152 Permeation, Sorption, and Diffusion in Poly(lactic acid) 155 Eva Almenar and Rafael Auras 12.1 Introduction 12.2 Factors Affecting Permeability, Sorption, and Diffusion in PLA 12.2.1 L-Lactide Unit Content 12.2.2 Plasticizers 12.2.3 Crystallization 12.2.4 Orientation 12.2.5 Free volume 12.2.6 Branching 12.2.7 Environmental Factors 12.3 Permeability, Sorption, and Diffusion of Pure PLA 12.3.1 Gases 12.3.2 Water Vapor 12.3.3 Organics 12.4 Copolymers 12.5 PLA Blends 12.5.1 PLA/Starch Blends 12.5.2 PLA/Poly(e-caprolactone) Blends 12.5.3 PLA/Chitosan Blends 12.5.4 PLA/Poly((R)-3-hydroxybutyrate) Blends 12.6 PLA Laminations 12.7 Coated PLA 12.8 PLA Composites and Fibers 12.8.1 PLA Composites 12.8.2 Fiber-Reinforced PLA 12.9 PLA Nanocomposites 12.10 Future of PLA Membranes References 155 157 157 158 159 160 160 160 161 163 163 165 167 169 169 169 170 171 171 171 171 171 172 172 172 176 176
  • x 13 CONTENTS Migration 181 Herlinda Soto-Valdez 13.1 Migration Principles 13.2 Legislation 13.3 Migration and Toxicological Data of Lactic Acid, Lactide, Dimers and Oligomers 13.3.1 Lactic Acid 13.3.2 Lactide 13.3.3 Oligomers 13.4 EDI of Lactic Acid 13.5 Other Potential Migrants from PLA 13.6 Conclusions References PART III 14 PROCESSING AND CONVERSION OF POLY(LACTIC ACID) Processing of Poly(lactic acid) 181 182 182 182 186 186 187 187 187 188 189 191 Loong-Tak Lim, Kevin Cink, and Tim Vanyo 14.1 Introduction 14.2 Properties of PLA Relevant to Processing 14.3 Modification of PLA Properties by Process Aids and Other Additives 14.4 Drying 14.5 Extrusion 14.6 Injection Molding 14.7 Film and Sheet Casting 14.8 Stretch Blow Molding 14.9 Extrusion Blown Film 14.10 Thermoforming 14.11 Electrospinning 14.12 Conclusion: Prospects of PLA Polymers References 15 191 191 193 194 195 196 201 204 207 208 209 211 213 Poly(lactic acid)/Starch Blends 217 Long Yu, Eustathios Petinakis, Katherine Dean, and Hongshen Liu 15.1 Introduction 15.2 Blending Hydrophobic PLA with Hydrophilic Starch 15.3 Compatibilizers Used for Starch/PLA Blends 15.4 Enhancing Function of Compatibilizer by Controlling Compatibilizer Distribution 15.5 Reactive Blending 15.6 Summary References 16 217 218 219 220 223 225 225 Poly(lactic acid) Blends 227 Sukeewan Detyothin, Ajay Kathuria, Waree Jaruwattanayon, Susan E. M. Selke, and Rafael Auras 16.1 Introduction 16.2 PLA/Nonbiodegradable Polymer Blends 16.2.1 Polyolefins 16.2.2 Vinyl and Vinylidene Polymers and Copolymers 227 227 228 229
  • CONTENTS xi 16.2.3 Elastomers and Rubbers 16.2.4 PMMA/PLA Blends 16.3 PLA/Biodegradable Polymer Blends 16.3.1 Polyanhydrides 16.3.2 Vinyl and Vinylidene Polymers and Copolymers 16.3.3 Aliphatic Polyesters and Copolyesters 16.3.4 Aliphatic-Aromatic Copolyester 16.3.5 Elastomers and Rubbers 16.3.6 Poly(ester amide) 16.3.7 Polyethers and Copolymer 16.3.8 Annually Renewable Biodegradable Materials 16.4 Plasticization of PLA 16.5 Conclusion References 17 238 239 240 240 242 244 255 257 258 258 261 264 266 266 Foaming 273 Laurent M. Matuana 17.1 Introduction 17.2 Plastic Foams 17.3 Foaming Agents 17.3.1 Physical Foaming Agents 17.3.2 Chemical Foaming Agents 17.4 Formation of Cellular Plastics 17.4.1 Dissolution of Blowing Agent in Polymer 17.4.2 Bubble Formation 17.4.3 Bubble Growth and Stabilization 17.5 Plastic Foams Expanded with Physical Foaming Agents 17.5.1 Microcellular Foamed Polymers 17.5.2 Solid State Batch Microcellular Foaming Process 17.5.3 Microcellular Foaming in a Continuous Process 17.6 PLA Foamed with Chemical Foaming Agents 17.6.1 Effect of CFA Content 17.6.2 Effect of Processing Conditions 17.7 Mechanical Properties of PLA Foams 17.7.1 Batch Microcellular Foamed PLA 17.7.2 Microcellular Extrusion of PLA 17.7.3 Microcellular Injection Molding of PLA 17.8 Foaming of PLA/starch Blends References 18 273 273 274 274 274 275 275 275 276 276 276 277 282 286 286 287 288 288 288 288 289 289 Composites 293 Subrata Bandhu Ghosh, Sanchita Bandyopadhyay-Ghosh, and Mohini Sain 18.1 Introduction 18.2 PLA Matrix 18.3 Reinforcements 18.3.1 Natural Fiber Reinforcement 18.3.2 Synthetic Fiber Reinforcement 18.3.3 Organic Filler Reinforcement 18.3.4 Inorganic Filler Reinforcement 18.4 Fiber/Matrix Adhesion 293 293 294 294 295 296 298 298
  • xii CONTENTS 18.4.1 Surface Modification 18.4.2 Compatibilizing Agent 18.5 PLA Nanocomposites 18.6 Processing 18.7 Properties 18.7.1 Mechanical Properties 18.7.2 Thermal Properties 18.7.3 Degradation 18.8 Applications 18.8.1 Biomedical Applications 18.8.2 Packaging Applications 18.8.3 Automotive Applications 18.8.4 Electronic Applications 18.9 Future Developments and Concluding Remarks References 19 298 299 299 300 300 300 303 304 305 305 306 306 307 307 307 Nanocomposites 311 Suprakas Sinha Ray 19.1 Introduction 19.2 PLA Nanocomposites Based on Clay 19.2.1 Structure and Properties of Clay 19.2.2 Preparation and Characterization of PLA/Clay Nanocomposites 19.3 PLA Nanocomposites Based on Carbon Nanotubes 19.4 PLA Nanocomposites Based on Various Other Nanoparticles 19.5 Properties of PLA-Based Nanocomposites 19.6 Biodegradability 19.7 Melt Rheology 19.8 Foam Processing 19.9 Possible Applications and Future Prospects Acknowledgments References 20 311 312 312 312 314 315 316 317 318 319 320 321 321 Spinning of Poly(lactic acid) Fibers 323 Ashwini K. Agrawal 20.1 Defining Fiber And Fiber Spinning 20.2 Melt Spinning Line 20.3 Fluid Dynamics During Spinning 20.3.1 Instabilities During Flow Through Spinneret 20.3.2 Instabilities in the Spinning Zone: Draw Resonance 20.4 Structure Development During Melt Spinning 20.4.1 Spinning at Low Speeds 20.4.2 Spinning at High Speeds 20.5 Post-Spinning Operation 20.5.1 Drawing 20.5.2 Finish Application 20.5.3 Crimping 20.5.4 Heat Setting 20.6 Structure Development During Drawing 20.7 Solution Spinning of PLLA 323 323 326 326 327 328 328 328 330 330 330 330 330 331 333
  • CONTENTS 20.7.1 Wet Spinning 20.7.2 Dry Spinning 20.7.3 Factors Affecting Solution Spinning 20.7.4 Dry-Jet Wet Spinning 20.7.5 Solution Spinning of Stereocomplex Fiber 20.8 Mechanical Properties 20.8.1 Melt-Spun PLLA 20.8.2 Solution-Spun PLLA 20.8.3 Stereocomplex Fiber References PART IV 21 DEGRADATION AND ENVIRONMENTAL ISSUES Hydrolytic Degradation xiii 334 335 335 337 337 338 338 339 341 341 343 345 Hideto Tsuji 21.1 Introduction 21.2 Degradation Mechanism 21.2.1 Molecular Degradation Mechanism 21.2.2 Material Degradation Mechanism 21.2.3 Degradation of Crystalline Residues 21.3 Parameters for Hydrolytic Degradation 21.3.1 Effects of Surrounding Media 21.3.2 Effects of Material Parameters 21.4 Structural and Property Changes During Hydrolytic Degradation 21.4.1 Fractions of Components 21.4.2 Crystallization 21.4.3 Mechanical Properties 21.4.4 Thermal Properties 21.4.5 Surface Properties 21.4.6 Morphology 21.5 Applications of Hydrolytic Degradation 21.5.1 Material Preparation 21.5.2 Recycling of PLA to Its Monomer 21.5.3 Miscellaneous Applications 21.6 Conclusions References 22 345 345 346 355 360 362 362 365 371 371 371 372 372 372 373 373 373 375 376 376 376 Enzymatic Degradation 383 Tadahisa Iwata, Hideki Abe, and Yoshihiro Kikkawa 22.1 Introduction 22.1.1 Definition of Biodegradable Plastics 22.1.2 Enzymatic Degradation 22.2 Enzymatic Degradation of PLA Films 22.2.1 Structure and Substrate Specificity of Proteinase K 22.2.2 Enzymatic Degradability of PLLA Films 22.2.3 Enzymatic Degradability of PLA Stereoisomers and Their Blends 22.2.4 Effects of Surface Properties on Enzymatic Degradability of PLLA Films 22.3 Enzymatic Degradation of Thin Films 22.3.1 Thin Films and Analytical Techniques 22.3.2 Crystalline Morphologies of Thin Films 22.3.3 Enzymatic Adsorption and Degradation Rate of Thin Films 383 383 383 384 385 385 386 388 390 390 391 391
  • xiv CONTENTS 22.3.4 Enzymatic Degradation of LB Film 22.3.5 Application of Selective Enzymatic Degradation 22.4 Enzymatic Degradation of Lamellar Crystals 22.4.1 Enzymatic Degradation of PLLA Single Crystals 22.4.2 Thermal Treatment and Enzymatic Degradation of PLLA Single Crystals 22.4.3 Single Crystals of PLA Stereocomplex 22.5 Future Perspectives References 23 394 394 395 395 396 397 397 398 Thermal Degradation 401 Haruo Nishida 23.1 Introduction 23.2 Kinetic Analysis of Thermal Degradation 23.2.1 Changes in Molecular Weight 23.2.2 Thermogravimetry 23.3 Thermal Degradation Behavior of PLA Based on Molecular Weight Change 23.4 Thermal Degradation Behavior of PLA Based on Weight Loss 23.4.1 Diverse Mechanisms of PLA Pyrolysis 23.4.2 Effects of Polymerization Catalyst Residues 23.4.3 Effects of Chain-End Structures 23.4.4 Thermal Degradation Catalysts 23.4.5 Thermal Degradation Behavior of PLA Stereocomplex: sc-PLA 23.4.6 Control of Racemization 23.4.7 Selective Depolymerization of PLA in Blends 23.5 Conclusions References 24 401 401 401 402 403 403 403 404 406 406 408 409 409 410 410 Photodegradation and Radiation Degradation 413 Wataru Sakai and Naoto Tsutsumi 24.1 Introduction 24.2 Mechanisms of Photodegradation 24.2.1 Photon 24.2.2 Photon Absorption 24.2.3 Photochemical Reaction of Carbonyl Groups 24.3 Mechanism of Radiation Degradation 24.3.1 High Energy Radiation 24.3.2 Basic Mechanism of Radiation Degradation 24.4 Photodegradation of PLA 24.5 Photosensitized Degradation of PLA 24.6 Radiation Effects on PLA 24.7 Modification of PLA by Irradiation References 25 413 413 413 414 415 415 415 415 416 418 419 420 420 Biodegradation 423 Buenaventurada P. Calabia, Yutaka Tokiwa, Charles U. Ugwu, and Seiichi Aiba 25.1 Introduction 25.2 Microbial Degradation 25.2.1 Field Test 25.2.2 ISO and ASTM Standards 423 423 423 424
  • CONTENTS xv 25.2.3 PLLA Degrading Microorganisms 25.3 Poly(L-Lactide) Degrading Enzymes 25.3.1 Proteinase K 25.3.2 PLLA-Degrading Enzyme of Amycolatopsis sp. 25.3.3 Induction of PLLA Degrading Enzymes with Natural Substrates 25.3.4 Other Serine Proteases 25.3.5 Lipase 25.3.6 Enzymatic Degradation in Organic Solvents 25.3.7 Evolution of PLA Degrading Enzymes 25.4 Conclusion and Future Prospects References 26 424 426 426 426 426 426 426 427 428 428 429 Cradle to Gate Environmental Footprint and Life Cycle Assessment of Poly(lactic acid) 431 Amy E. Landis 26.1 Introduction to LCA and Environmental Footprints 26.1.1 Life Cycle Assessment 26.1.2 Uncertainty in LCA 26.2 Life Cycle Considerations for PLA 26.2.1 The Life Cycle of PLA 26.2.2 Energy USE and Global Warming 26.2.3 Environmental Trade-Offs 26.3 Review of Biopolymer LCA Studies 26.4 Improving PLA’s Environmental Footprint 26.4.1 Agricultural Management 26.4.2 Feedstock Choice 26.4.3 Energy Further Reading on LCA References PART V 27 APPLICATIONS Medical Applications 431 431 432 432 432 433 434 434 438 438 439 439 440 440 443 445 Shuko Suzuki and Yoshito Ikada 27.1 Introduction 27.2 Minimal Requirements for Medical Devices 27.2.1 General 27.2.2 PLA as Medical Implants 27.3 Preclinical and Clinical Applications of PLA Devices 27.3.1 Fibers 27.3.2 Meshes 27.3.3 Bone Fixation Devices 27.3.4 Microspheres, Microcapsules, and Thin Coatings 27.4 Conclusions References 28 445 445 445 446 447 447 448 448 453 454 454 Packaging and Other Commercial Applications 457 Shoji Obuchi and Shinji Ogawa 28.1 Introduction 28.2 Applications in Packaging and Containers 28.2.1 Oriented Film 457 457 457
  • xvi CONTENTS 28.2.2 Flexible Film 28.2.3 Shrink Labels and Film 28.2.4 Thermoforming 28.2.5 Lamination 28.3 Other Commercial Applications 28.3.1 Agricultural and Engineering Materials 28.3.2 Electrical Appliances 28.3.3 Automotive Materials 28.4 Conclusions References 29 458 461 462 462 462 462 463 466 467 467 Textile Applications 469 Masatsugu Mochizuki 29.1 Introduction 29.2 Manufacturing, Properties, and Structure of PLA Fibers 29.2.1 PLA Fiber Manufacture 29.2.2 PLA Fibers and Textile Properties 29.2.3 Effects of Structure on Properties 29.2.4 PLA Stereocomplex Fibers 29.3 Key Performance Features of PLA Fibers 29.3.1 Biodegradability and the Biodegradation Mechanism 29.3.2 Moisture Management 29.3.3 Antibacterial/Antifungal Properties 29.3.4 Low Flammability 29.3.5 Weathering Stability 29.4 Potential Applications 29.4.1 Geotextiles 29.4.2 Industrial Fabrics 29.4.3 Filters 29.4.4 Towels and Wipes 29.4.5 Home Furnishings 29.4.6 Clothing and Personal Belongings 29.5 Conclusions References 30 469 469 469 469 470 471 471 471 472 472 473 474 474 474 475 475 475 475 475 475 476 Environmental Applications 477 Akira Hiraishi 30.1 Introduction 30.2 Application to Water and Wastewater Treatment 30.2.1 Application as Sorbents 30.2.2 Application to Nitrogen Removal 30.3 Application to Bioremediation 30.3.1 Significance of PLA Use 30.3.2 Bioremediation of Organohalogen Pollution 30.3.3 Other Applications 30.4 Concluding Remarks and Prospects Acknowledgments References INDEX 477 477 477 479 482 482 482 483 484 484 484 487
  • PREFACE Poly(lactic acid) (PLA) cannot be considered as a new polymer. As early as 1845, PLA was synthesized by Thophile-Jules Pelouze by the condensation of lactic acid e [1]. In 1932, Wallace Hume Carothers et al. developed a method to polymerize lactide to produce PLA that was later patented by DuPont in 1954 [2]. Although PLA existed for several decades, its use was limited to biomedical applications (e.g., biocompatible sutures, implants, biologically active controlled release devices) due to its high cost. The low molecular weight PLA polymers obtained also hampered their wide-ranging applications. The breakthrough occurred in the early 1990s when Cargill Inc. succeeded in polymerizing high molecular weight PLA using a commercially viable lactide ring opening reaction [3, 4]. In 1997, Cargill Dow LLC, a joint venture between Cargill Inc. and The Dow Chemical Company, was formed to begin truly commercially significant production of PLA resins under the trade name NatureWorksÔ. This is a major landmark in PLA’s history because it signifies the beginning of a large-scale use of this bio-based polymer, transforming PLA from a specialty material to a commodity thermoplastic. The increased availability of PLA stimulated an increased in its research and development activities. A survey of the literature revealed that the number of published articles related to PLA increased exponentially over the past decade, which can be also partly attributed to the escalating “green” movement that is stimulating the use of bio-based polymers. To date, the major PLA resin suppliers have been Cargill (in the United States known as IngeoÔ), Mitsui Chemicals, Inc. (in Japan known as LACEAÔ), Purac (The Netherlands), and Teijin Limited (in Japan known as BiofrontÒ ). Other important events that took place pertaining to PLA are summarized in Figure P.1 and Table P.1. While the information available in the literature is massive, at the inception of this volume, no reference book could be found that coherently assembled the scientific and technological knowledge about PLA. Our main motive for editing this book was to consolidate the most relevant information on PLA into a volume that serves as a one-source reference for readers who are keen on this unique biodegradable polymer. Organized in five parts, Part I of this book covers several important topics, including chemistry and production of lactic acid (Chapter 1) and lactide (Chapter 2), which are the essential building blocks of PLA. Different polymerization reactions for the production of PLA are covered in Chapter 3. In view of certain shortcomings of PLA, copolymerization of PLA with other monomers and stereocomplexation with optimal enantiomer lactide ratios are gaining increased popularity as ways to enhance the material properties of the resulting polymer. These topics are covered in detail in Chapters 4 and 5, respectively. These chapters set the stage for discussions in Part II of this book, in which different material properties of PLA are covered in eight separate chapters. Chain configuration, tacticity, and crystal structure are discussed in Chapter 6 to illuminate how chain structure affects the material properties of PLA and its copolymers. Chapter 7 investigates the compatibility of PLAwith solvents and other polymers, an important aspect that should be considered during end-use applications. The interaction of PLA with electromagnetic radiation for probing the molecular structure and interactions are discussed in Chapter 8. The essence of spectroscopy techniques for PLA analysis, including UV–VIS, FTIR, Raman, and NMR, is reviewed in this chapter. Crystallization, thermal, and rheological properties of PLA are discussed in Chapters 9 and 10, and these xvii
  • xviii PREFACE FIGURE P.1 Number of works published since 1960 based on Scholars Portal Search (accessed on 28 Dec 09) using keywords “polylactide”, “poly(lactic acid)”, and “polylactic acid”. TABLE P.1 Significant Events Related to PLA Production that Occurred over the Past Few Decades 2010 2009 2009 2009 2008 2008 2007 2007 2005 2003 1997 1997 1996 1994 Jung et al. employed recombinant Escherichia coli to produce PLA [5] PURAC, Sulzer, and Synbra announced production of PLA from solid lactide for foamed products Galactic and Total Petrochemicals from Belgium created a joint venture, Futerro, to begin PLA production Cargill Inc. acquired full NatureWorks ownership from Teijin Limited Uhde Inventa Fischer and Pyramide Bioplastics announced large-scale production of PLA in Guben, Germany PURAC started to commercialize solid lactide monomers under PURALACTÔ Teijin launched heat-resistant stereocomplex PLA under BiofrontÔ NatureWorks LLC and Teijin Limited formed 50–50 joint venture to market IngeoÔ bio-based thermoplastic resins Cargill Inc. acquired The Dow Chemical Company’s share in Cargill Dow LLC 50–50 joint venture Toyota produced and developed PLA for automotive applications Formation of Cargill Dow LLC, a 50–50 joint venture of Cargill Inc. and The Dow Chemical Company, to commercialize PLA under the trade name NatureWorksÔ Fiberweb (now BBA, France) introduced melt-blown and spun-laid PLA fabrics under DeposaÔ brand name Mitsui Chemicals, Inc. commercialized PLA produced by polycondensation route Kanebo Ltd. introduced LactronÒ PLLA fiber and spunlaid nonwovens are important to elucidate the melt processing phenomena of PLA. In the remainder of Part II, Chapters 11, 12, and 13 deal with the mechanical, permeability, and migration behaviors of PLA, respectively, and will serve as handy references for designing and engineering PLA products for various end-use applications. In Part III, seven chapters are devoted to summarizing the state of the art of processing and conversion technologies for PLA, covering topics such as extrusion and molding (Chapter 14), polymer blending (Chapters 15 and 16), foaming (Chapter 17), preparation of micro- and nanocomposites (Chapters 18 and 19), and fiber spinning (Chapter 20). One of the hallmarks of PLA polymers is that they are degradable, which has been viewed as an attractive feature for certain applications. In Part IV, six chapters are included to discuss in great detail the various degradation modes of PLA, including hydrolytic degradation (Chapter 21), enzymatic degradation (Chapter 22), thermal degradation (Chapter 23), photodegradation (Chapter 25), and biodegradation (Chapter 25). This part ends with Chapter 26 in which the life cycle assessment and the environmental footprint of PLA are objectively discussed. Finally, in Part V, various applications for PLA are discussed, including medical items (Chapter 27), packaging (Chapter 28), textiles (Chapter 29), and environment-related applications (Chapter 30). Rather than eliminating all duplicate materials between chapters, we deliberately allowed some overlap in discussions to enable the chapters to stand alone to some extent. This volume skillfully brings together the work of many contributors who are experts in their respective research areas. This volume would not have been possible without
  • PREFACE their help and contributions. We are indebted to them for their participation and patience during the preparation of this book and are grateful that they have entrusted us to edit their contributions as per the requirements of each chapter. We hope that readers will find this book useful. We are looking forward to receiving comments and constructive feedback regarding the content of this book [5]. Finally, we are indebted to our three academic institutions, Michigan State University, University of Guelph, and Toyohashi University of Technology, for allowing us to dedicate our effort and time to the completion of this edited book. Our most grateful thanks are to our colleagues for providing a sounding board to discuss ideas and explore new concepts about biodegradable polymers and materials in general; to our editor at John Wiley & Sons, Inc., Jonathan T. Rose, for supporting this proposal, and walking us through its completion; to Lisa Van Horn for coordinating the production of the book; and to Sanchari Sil, our project manager at Thomson Digital, for her invaluable patience to in answering our endless questions about the final proofing of the book. Overall, we could not put our effort into this task without the unconditional support of our families, so that our most special thanks go to all of them. xix REFERENCES 1. H. Benninga, A History of Lactic Acid Making, Springer, New York, 1990. 2. W. H. Carothers, G. L. Dorough, F. J. van Natta. J. Am. Chem. Soc.1932, 54, 761–772. 3. P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchardt, U.S. Patent 6,326,458, 2001. 4. P. R. Gruber, E. S. Hall, J. J. Kolstad, M. L. Iwen, R. D. Benson, R. L. Borchardt, U.S. Patent 5,357,035, 1994. 5. Y. K. Jung, T. Y. Kim, S. J. Park, S. Y. Lee, Biotechnol. Bioeng. 2010, 105, 161–171. RAFAEL AURAS East Lansing, Michigan LOONG-TAK LIM Guelph, Ontario, Canada SUSAN E. M. SELKE East Lansing, Michigan HIDETO TSUJI Toyohashi, Aichi, Japan May 2010
  • CONTRIBUTORS Steven Abbott, Steven Abbott TCNF Ltd., Ipswich, England Hideki Abe, Chemical Analysis Team, RIKEN Institute, Saitama, Japan Seiichi Aiba, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan Ashwini K. Agrawal, SMITA Labs, Department of Textile Technology, Indian Institute of Technology, New Delhi, India Ann-Christine Albertsson, Department of Fiber and Polymer Technology, Royal Institute of Technology, Stockholm, Sweden Sukeewan Detyothin, School of Packaging, Michigan State University, East Lansing, Michigan; Department of Agroindustry, Faculty of Agriculture, Natural Resources and Environment, Naresuan University, Phitsanulok, Thailand John R. Dorgan, Colorado School of Mines, Golden, Colorado Luca Fambri, Department of Materials Engineering and Industrial Technologies and BIOtech Research Center, University of Trento, Trento, Italy Anna Finne-Wistrand, Department of Fiber and Polymer Technology, Royal Institute of Technology, Stockholm, Sweden Eva Almenar, School of Packaging, Michigan State University, East Lansing, Michigan Subrata Bandhu Ghosh, Center for Biocomposites and Biomaterials Processing, Faculty of Forestry, University of Toronto, Toronto, Ontario, Canada Rafael Auras, School of Packaging, Michigan State University, East Lansing, Michigan Carla M. B. Gon¸ alves, CICECO, Chemistry Department, c University of Aveiro, Aveiro, Portugal Sanchita Bandyopadhyay-Ghosh, Center for Biocomposites and Biomaterials Processing, Faculty of Forestry, University of Toronto, Toronto, Ontario, Canada Buenaventurada P. Calabia, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan Gian Domenico Cella, Novamont SpA, Novara, Italy Kevin Cink, NatureWorks LLC, Minnetonka, Minnesota Wim Groot, PURAC, Gorinchem, The Netherlands Akira Hiraishi, Department of Environmental and Life Sciences, Graduate School of Engineering, Toyohashi University of Technology, Toyohashi, Aichi, Japan Masayuki Hirata, Department of Biomolecular Engineering, Center for Fiber and Textile Science, Kyoto Institute of Technology, Kyoto, Japan Jo~o A. P. Coutinho, CICECO, Chemistry Department, a University of Aveiro, Aveiro, Portugal Dan Huang, Key Laboratory of Eco-textiles of Ministry of Education, College of Textile and Clothing, Jiangnan University, Wuxi, China Katherine Dean, CSIRO, Materials Science and Engineering, Melbourne, Australia Yoshito Ikada, Faculty of Medicine, Nara Medical University, Nara, Japan xxi
  • xxii CONTRIBUTORS Tadahisa Iwata, Department of Biomaterial Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Tokyo, Japan Haruo Nishida, Eco-Town Collaborative R&D Center for the Environment and Recycling, Kyushu Institute of Technology, Fukuoka, Japan Waree Jaruwattanayon, School of Packaging, Michigan State University, East Lansing, Michigan; Thai Packaging Center, Thailand Institute of Scientific and Technological Research, Bangkok, Thailand Shoji Obuchi, Materials Laboratory, Mitsui Chemicals, Inc., Chiba, Japan Xue Jiang, Key Laboratory of Eco-textiles of Ministry of Education, College of Textile and Clothing, Jiangnan University, Wuxi, China Zhengyu Jin, School of Food Science and Technology, Jiangnan University, Wuxi, China Ajay Kathuria, School of Packaging, Michigan State University, East Lansing, Michigan Yoshihiro Kikkawa, Photonics Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan Shinji Ogawa, Process Technology Center, Mitsui Chemicals, Inc., Chiba, Japan Gabriele Perego, Vanadis Srl, Colleretto Giacosa, Italy Eustathios Petinakis, CSIRO, Materials Science and Engineering, Melbourne, Australia Narendra Reddy, Department of Textiles, Clothing and Design, University of Nebraska-Lincoln, Lincoln, Nebraska Mohini Sain, Center for Biocomposites and Biomaterials Processing, Faculty of Forestry, University of Toronto, Toronto, Ontario, Canada Yoshiharu Kimura, Department of Biobased Materials Science, Center for Fiber and Textile Science, Kyoto Institute of Technology, Kyoto, Japan Wataru Sakai, Macromolecular Science and Engineering, Kyoto Institute of Technology, Kyoto, Japan Kamlesh Kumar, Center for Polymer Science and Engineering, Indian Institute of Technology, New Delhi, India Susan E. M. Selke, School of Packaging, Michigan State University, East Lansing, Michigan Amy E. Landis, Department of Civil and Environmental Engineering, Swanson School of Engineering, University of Pittsburgh, Pittsburgh, Pennsylvania Suprakas Sinha Ray, DST/CSIR Nanotechnology Innovation Center, National Center for Nano-Structured Materials, Council for Scientific and Industrial Research, Pretoria, Gauteng, Republic of South Africa Loong-Tak Lim, Department of Food Science, University of Guelph, Guelph, Ontario, Canada Hongshen Liu, CSIRO, Materials Science and Engineering, Melbourne, Australia; Center for Polymer from Renewable Resources, SCUT, Guangzhou, China Bimlesh Lochab, Center for Polymer Science and Engineering, Indian Institute of Technology, New Delhi, India Yan Luo, Department of Applied Chemistry, College of Chemistry, Chemical Engineering and Biotechnology, Donghua University, Shanghai, China Isabel M. Marrucho, CICECO, Chemistry Department, University of Aveiro, Aveiro, Portugal; Instituto de Tec´ nologia Qu ımica e Biologica, ITQB2, Universidade Nova de Lisboa, Oeiras, Portugal Laurent M. Matuana, School of Packaging, Michigan State University, East Lansing, Michigan Claudio Migliaresi, Department of Materials Engineering and Industrial Technologies and BIOtech Research Center, University of Trento, Trento, Italy Masatsugu Mochizuki, Center for Fiber and Textile Science, Kyoto Institute of Technology, Kyoto, Japan Olav Sliekers, PURAC, Gorinchem, The Netherlands ´ Herlinda Soto-Valdez, Centro de Investigacion en Alimen´ tacion y Desarrollo A.C., Hermosillo, Sonora, Mxico e  Anders S€dergard, Laboratory of Polymer Technology, o  Abo Akademi University, Turku, Finland  Mikael Stolt, Laboratory of Polymer Technology, Abo Akademi University, Turku, Finland Shuko Suzuki, Faculty of Medicine, Nara Medical University, Nara, Japan Yaoqi Tian, School of Food Science and Technology, Jiangnan University, Wuxi, China Xiuzhi Tian, Key Laboratory of Eco-textiles of Ministry of Education, College of Textile and Clothing, Jiangnan University, Wuxi, China Yutaka Tokiwa, Okinawa Industrial Technology Center, Uruma, Okinawa, Japan Hideto Tsuji, Department of Environmental and Life Sciences, Graduate School of Engineering, Toyohashi University of Technology, Toyohashi, Aichi, Japan
  • CONTRIBUTORS Naoto Tsutsumi, Macromolecular Science and Engineering, Kyoto Institute of Technology, Kyoto, Japan Charles U. Ugwu, Okinawa Industrial Technology Center, Uruma, Okinawa, Japan Jan van Krieken, PURAC, Gorinchem, The Netherlands Tim Vanyo, NatureWorks LLC, Minnetonka, Minnesota Indra Kumari Varma, Center for Polymer Science and Engineering, Indian Institute of Technology, New Delhi, India xxiii Sicco de Vos, PURAC, Gorinchem, The Netherlands Jinpeng Wang, School of Food Science and Technology, Jiangnan University, Wuxi, China Yiqi Yang, Department of Textiles, Clothing and Design and Department of Biological Systems Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska Long Yu, CSIRO, Materials Science and Engineering, Melbourne, Australia; Center for Polymer from Renewable Resources, SCUT, Guangzhou, China
  • PART I CHEMISTRY AND PRODUCTION OF LACTIC ACID, LACTIDE, AND POLY(LACTIC ACID)
  • 1 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE WIM GROOT, JAN VAN KRIEKEN, OLAV SLIEKERSL, 1.1 AND INTRODUCTION Natural polymers, biopolymers, and synthetic polymers based on annually renewable resources are the basis for the twenty-first-century portfolio of sustainable, eco-efficient plastics [1]. These biosourced materials will gradually replace the currently existing family of oil-based polymers as they become cost- and performance-wise competitive. Polylactide or poly(lactic acid) (PLA) is the front runner in the emerging bioplastics market with the best availability and the most attractive cost structure. The production of the aliphatic polyester from lactic acid, a naturally occurring acid and bulk produced food additive, is relatively straightforward. PLA is a thermoplastic material with rigidity and clarity similar to polystyrene (PS) or poly(ethylene terephthalate) (PET). End uses of PLA are in rigid packaging, flexible film packaging, cold drink cups, cutlery, apparel and staple fiber, bottles, injection molded products, extrusion coating, and so on [2]. PLA is bio-based, resorbable, and biodegradable under industrial composting conditions [1, 3, 4]. PLA can be produced by condensation polymerization directly from its basic building block lactic acid, which is derived by fermentation of sugars from carbohydrate sources such as corn, sugarcane, or tapioca, as will be discussed later in this chapter. Most commercial routes, however, utilize the more efficient conversion of lactide—the cyclic dimer of lactic acid—to PLA via ring-opening polymerization (ROP) catalyzed by a Sn(II)-based catalyst rather than polycondensation [2–6]. Both polymerization concepts rely on highly concentrated polymer-grade lactic acid of excellent quality SICCO DE VOS for the production of high molecular weight polymers in high yield [2–4, 7]. Purification of lactic acid produced by industrial bacterial fermentation is therefore of decisive importance because crude lactic acid contains many impurities such as acids, alcohols, esters, metals, and traces of sugars and nutrients [4]. The lactide monomer for PLA is obtained from catalytic depolymerization of short PLA chains under reduced pressure [4]. This prepolymer is produced by dehydration and polycondensation of lactic acid under vacuum at high temperature. After purification, lactide is used for the production of PLA and lactide copolymers by ROP, which is conducted in bulk at temperatures above the melting point of the lactides and below temperatures that cause degradation of the formed PLA [4]. Processing, crystallization, and degradation behavior of PLA all depend on the structure and composition of the polymer chains, in particular the ratio of the L- to the D-isomer of lactic acid [2, 4, 6, 8, 9]. This stereochemical structure of PLA can be modified by copolymerization of mixtures of L-lactide and meso-, D-, or rac-lactide resulting in high molecular weight amorphous or semicrystalline polymers with a melting point in the range from 130 to 185 C [3, 4, 6–10]. Isotactic PLLA homopolymer—comprising L-lactide only—is a semicrystalline material with the highest melting point, while PLA copolymers with higher D-isomer content exhibit lower melting points and dramatically slower crystallization behavior, until they finally become amorphous at D-contents higher than 12–15% [8–10]. Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 3
  • 4 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE For decades, ROP has been the preferred route to PLA for biomedical applications with small production volumes. PLLA and copolymers with rac-lactide, glycolide, and ecaprolactone for resorbable biomedical applications have been produced by, for example, PURAC, previously known as CCA, since the 1970s [5]. Since the 1990s, the ROP concept is also used for high-volume production of PLA grades for other end uses. Large-scale production of PLA, copolymers of L- and meso-lactide, was started in 2002 by a joint venture of Cargill and Dow under the name NatureWorks LLC. Nowadays, since July 1, 2009, NatureWorks LLC is again wholly owned by Cargill and has a production capacity of 140 ktpa for its Ingeo PLA grades in Blair, Nebraska [11]. The attractive price and commercial availability of lactic acid were important reasons why PLA became the first massproduced bio-based polyester. The critical success factor for a final breakthrough of all green chemicals and plastics based on annually renewable materials is economic sustainability. Thus, the very basis of cost-competitive PLA is an industrial fermentative production process for lactic acid with efficient use of carbohydrates followed by excellent purification technology with minimum generation of by-products. An important impulse for the expanding bioplastics market is the commercialization of lactide monomers for PLA by PURAC in 2008. Solid D- and L-lactides are now available in bulk quantities and can be polymerized into a whole range of tailor-made polylactides by continuous melt polymerization processes, like the technology based on static mixing reactors that was jointly developed by Sulzer and PURAC. PLA offers an unprecedented market potential to lactic acid producers all over the world, but not all potential players can succeed, because PLA production poses stringent demands to lactic acid quality and price. The chemistry and physics of today’s fermentative production and industrialscale purification of lactic acid and lactide are the subject of this chapter. 1.2 1.2.1 LACTIC ACID History of Lactic Acid Lactic acid was discovered in 1780 by the experimental chemist Carl Wilhelm Scheele, who isolated ‘‘acid of milk’’ from sour whey [12, 13]. A further description of the history of lactic acid by Holten and Benninga shows that industrial production of lactic acid started in the United States in the 1880s [14, 15]. Avery patented and applied a process of fermentation of vegetable sugars [16]. The actual application was the use of a mixture of calcium lactate and lactic acid as baking powder. Unfortunately, this application was not a big success, but other applications in food and textile dyeing were developed. FIGURE 1.1 Two enantiomeric forms of lactic acid: (S)- and (R)2-hydroxypropionic acid. In 1950, the first commercial production of synthetic lactic acid started in Japan [15]. Lactonitrile was produced from acetaldehyde and hydrogen cyanide and hydrolyzed in the second stage to lactic acid. For some decades, synthetic lactic acid competed with lactic acid obtained by fermentation, but currently almost all lactic acid is produced by fermentation. 1.2.2 Physical Properties of Lactic Acid Lactic acid (2-hydroxypropanoic acid) is the simplest 2-hydroxycarboxylic acid (or a-hydroxy acid) with a chiral carbon atom and exists in two enantiomeric forms (Figure 1.1). The chirality of lactic acid often results in confusion regarding nomenclature. A number of different names are used in the literature. This confusion is the result of mixing the molecular structure and a physical property (optical rotation). (S)-Lactic acid (or L-lactic acid) has a slightly positive specific optical rotation and is frequently named L-( þ )-lactic acid [14]. However, a concentrated solution of (S)-lactic acid at equilibrium contains lactic acid oligomers, which results in an overall negative optical rotation. Therefore, it is advised to use the structural R/S notation or the older notation of L and D and avoid the þ and À of the optical rotation (Table 1.1). 1.2.3 Chemistry of Lactic Acid The lactic acid molecule has a hydroxyl and an acid functional group, which may result in intermolecular and intramolecular esterification reactions. The first step is the formation of a linear dimer (lactoyl lactic acid). This condensation reaction can proceed to higher oligomers and is promoted by removal of water. Also a cyclic dimer, lactide, is formed in small amounts. Lactide can be formed by intramolecular esterification of lactoyl lactic acid or by breakdown of higher oligomers. All reactions are equilibrium reactions (Figure 1.2). Due to these reactions, a solution of lactic acid at equilibrium consists of monomeric lactic acid, dimeric lactic acid or lactoyl lactic acid, higher oligomers of lactic acid, and
  • LACTIC ACID TABLE 1.1 Physical Properties of Lactic Acid [14] Property Value CAS number General: 50-21-5 (S)-Lactic acid: 79-33-4 (R)-Lactic acid: 10326-41-7 90.08 Molecular weight (g/mol) Formula Melting point ( C) Crystal structure Solid density (g/mL) Solubility in water (wt%) Heat of fusion (kJ/mol) Boiling point ( C) Liquid density (g/mL, 20 C) Viscosity (mPa s) pKa Specific heat (J/(g K) at 25 C) C3H6O3 18 (racemic) 53 (chiral pure) (S)-Lactic acid: orthorhombic, space group P212121 1.33 (solid, 20 C) 86 (20 C, monomeric (S)-lactic acid) (S)-Lactic acid: 16.8 122 (at 14 mmHg) 1.224 (100% undercooled liquid) 1.186 (80.8% solution in water) 28.5 (85.3% solution in water, 25 C) 3.86 Crystalline (S)-lactic acid: 1.41 Liquid lactic acid: 2.34 Reference [20] [20] [21] [22] [22] [23] [23] [24] [25] [26] lactide. The ratios between all substances depend on the amount of water present; for example, a 90.1% lactic acid solution (total acidity) contains about 59.3% of monomeric lactic acid and 27.3% of lactoyl lactic acid and higher oligomers [14]. FIGURE 1.2 The condensation reactions are also the reason that it is quite difficult to obtain pure, solid, and enantiopure lactic acid. This can only be achieved by crystallization [27, 28]. The kinetics of the condensation reactions determine the stability of a solution of monomeric lactic acid and have a large influence on the stability of solid lactic acid. 1.2.4 [17] [18] [19] 5 Production of Lactic Acid by Fermentation Almost all lactic acid available on the market is produced by fermentation. During fermentation, a suitable carbohydrate is converted to lactic acid by microorganisms. Although some of the microorganisms used, such as the mold Rhizopus, need oxygen for growth, the actual conversion of sugars to lactic acid is carried out without oxygen. As a matter of fact, the complete oxidation of a sugar to carbon dioxide and water is energetically much more favorable, so lactic acid is mainly formed under anaerobic conditions. Indeed, most lactic acid producing microorganisms are inactive when oxygen is continuously present in high amounts [29]. Upon entering the cell, the sugar is first converted to pyruvate by several enzymatic steps. This conversion yields chemical energy in the form of ATP (adenosine triphosphate) and reducing equivalents (NADH); see the reaction in Figure 1.3. In order to recycle these reducing equivalents, microorganisms convert the pyruvate into the more reduced lactic acid; see the reaction in Figure 1.4. In other words, lactic acid is mainly produced to keep the cellular processes going [30]. The chemical energy obtained is used by several processes elsewhere in the cell, for example, cell growth, maintenance, and sometimes even motility. The reaction in Figure 1.3 takes place in the so-called homofermentative lactic acid bacteria (LAB). Homofermentative bacteria have almost exclusively lactic acid as a Lactic acid condensation reactions: interchange between lactide, oligomers, and poly(lactic acid).
  • 6 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE FIGURE 1.3 During conversion of glucose to pyruvate, chemical energy (ATP) is generated as well as reducing equivalents (NADH). fermentation product, in contrast to heterofermentative bacteria that produce a mixture of lactic acid, acetate, CO2, and acetate or ethanol [31]. Heterofermentative bacteria were believed to use exclusively the so-called phosphoketolase pathway, and homofermentative bacteria were believed to use exclusively the glycolysis (Figure 1.3) that splits C6 into two C3 molecules [31]. The phosphoketolase pathway is a route where a C6 is transformed to a C5 sugar (and CO2) and split into a C2 and a C3 molecule. The C3 molecule is then converted to lactic acid whereas the C2 molecule is converted to acetate or ethanol. In the same traditional view, C5 sugars were regarded as leading to this heterofermentative metabolism, which is less interesting from the point of view of industrial production as a lot of acetic acid or ethanol is produced simultaneously. Although some bacteria seem to fit well in this paradigm, more recent literature has shown that this view is oversimplified and somewhat obsolete for a number of reasons. . . . Some heterofermentative bacteria are shown to have both pathways active at the same moment and produce mostly lactic acid under certain circumstances [32, 33]. Pentoses can lead exclusively to lactic acid as a fermentation product [34]. Lactic acid producing organisms that do not have a phosphoketolase pathway can still produce acetate or ethanol, formed by the usual mixed acid fermentation, via pyruvate. This is the case for, for example, Lactococcus lactis [35]. FIGURE 1.4 Lactic acid formation from pyruvate: reoxidation of NADH and NAD takes place; NAD can be used again in the reaction of Figure 1.3. The reason why even heterofermentative bacteria prefer to produce mostly lactic acid is related to the fast generation of chemical energy and thus fast growth and acidification of the environment [32, 36, 37]. The fast growth and acidification gives lactic acid bacteria a competitive advantage and that is exactly why lactic acid bacteria are so troublesome in ethanol fermentations [38]. Lactic acid production is certainly not restricted to bacteria or fungi. Higher organisms, including humans, also use lactic acid formation for fast supply of energy in muscles when needed [39]. The uniformity in this biochemistry is in sharp contrast with the degrees of freedom one has in choosing the microbes, the acid-neutralizing agent, nutrients, and carbohydrates needed for industrial lactic acid fermentation. Only delicate weighing of the pros and cons of every possibility leads to an economically feasible fermentation. 1.2.4.1 The Microbes There are several important features a microorganism used for the production of lactic acid must have in order to be industrially attractive: . . . . . . high productivity to reduce fermentation time, high conversion yield to reduce carbohydrate costs, ability to use cheap sources of nutrients to reduce nutrient costs, high end concentration to reduce evaporation costs, low amount of by-products to increase purification yield, and, of course, the organisms must be robust with regard to contamination and infections. Every microorganism has its own benefits and drawbacks, but lactobacilli (present in many food fermentations) and Rhizopus (a fungus) are the most reported [40]. Besides lactobacilli and Rhizopus, Streptococcus, Pediococcus, Sporolactobacillus inulinus, Bacillus coagulans, and several yeasts are mentioned in the excellent overview by Vaidya et al. [41]. Lactobacilli generally have high productivity, but special and often expensive nutrient requirements. Rhizopus needs much less nutrients, but has a lower yield, needs oxygen, and its morphology is sometimes difficult to handle. Of course, via genetic manipulation, researchers have tried to make an ideal lactic acid producing microorganism. 1.2.4.2 Stereochemical Purity In order to make semicrystalline, high-melting PLA, stereochemically pure lactic acid is needed. Not all microorganisms yield such stereochemically pure lactic acid and some even produce a racemic mixture [29]. Therefore, a strain must be chosen that meets the quality demands. Finding such a strain that produces L-lactic acid in an economically feasible manner is relatively easy. Producing D-lactic acid by bacterial fermentation on an industrial scale is far more difficult.
  • LACTIC ACID Several natural D-lactic acid producing bacterial species exist; Sporolactobacillus inulinus, Sporolactobacillus laevolacticus (previously Bacillus laevolacticus), and Lactobacillus delbrueckii are among these bacteria [29, 42, 43]. Also, patents have been filed claiming the production of D-lactic acid by a genetically modified microorganism. Several different species such as Kluyveromyces and Escherichia coli have been claimed so far [44, 45]. 1.2.4.3 Nutrients The most well-known lactic acid producing organisms, such as Lactobacillus and Lactococcus species, are members of the taxonomic order of Lactobacillales, also commonly referred to as lactic acid bacteria. These lactic acid bacteria have their really complex nutrient need in common [29]. Vitamins and peptides need to be added to the medium to enable growth. This can be done by adding peptones, yeast extract, or corn steep liquor, but this is expensive. Nutrients for lactic acid production can also be derived from nutrient-rich waste streams such as rice bran, fish waste, or vinification lees [46–48]. 1.2.4.4 Neutralization Lactic acid fermentation inevitably leads to a drop in pH, and without neutralization the microorganism is quickly unable to continue the fermentation, as the environment becomes too acidic. Several bases can be used to neutralize the acidity during fermentation, and the choice of the base will determine the nature of the downstream processing (DSP). Most industrial lactic acid plants use Ca(OH)2 or CaCO3, which results in the production of a large amount of gypsum as a by-product. A major challenge in lactic acid production is to find or construct an efficient microorganism that can produce at such a low pH that the fermentation does not require neutralization. Lactic acid bacteria are usually able to grow at low pH, but it is difficult to find an organism capable of producing lactic acid in reasonable amounts at pH close to the pKa of lactic acid [49]. Another solution is to construct a lactic acid producing yeast but organisms like this still suffer from low productivities (amount of lactic acid produced per hour) and low final concentrations, leading to the requirement for large fermenter volumes and high amounts of water evaporation [50]. Some basic hurdles have to be overcome in order to improve the low-pH fermentation by yeasts. Although yeasts are very resistant to low pH, the export of lactate from the yeast cell to the outside medium costs them as much energy as they get from lactic acid production by fermentation. For this reason, lactic acid producing yeasts need reasonable amounts of oxygen in order to generate enough energy to survive [51]. In contrast, traditional lactic acid bacteria use another way to transport lactic acid across the membrane and even gain extra energy by exporting lactic acid to the medium [52]. 7 1.2.4.5 Carbohydrates for Lactic Acid Production In principle, any carbohydrate source containing pentoses (C5 sugars) or hexoses (C6 sugars) can be used for the production of lactic acid, although it is very rare that any particular microorganism is able to use all possible and available C5 and C6 sugars. Pure sucrose from sugarcane or sugar beets and glucose from starch are available in large amounts and readily fermentable. Polysaccharides such as cellulose or starch are more complex and need special pretreatment. When using less pure sources such as raw sugar beet juice, the impurities must be removed somewhere in the total lactic acid production process [53]. This can be done before, during, or after the fermentation. This often leads to special adaptations in the production plant. Last but not least, the local price and availability of the carbohydrate source determine the raw material of choice for industrial fermentation. Another usable disaccharide is lactose present in whey, as was used by Scheele when he discovered lactic acid in 1780 [12]. 1.2.4.6 Starch Starch occurs in discrete granules and is usually a mixture of two homopolymers of glucose, amylopectin and amylose. Starch can be derived from corn, wheat, potato, or tapioca [54]. Although some microorganisms are able to degrade and ferment starch directly to lactic acid, most lactic acid producing microorganisms cannot hydrolyze starch themselves. A solution is to hydrolyze the starch to glucose prior to fermentation with the commercially available enzymes, a-amylase and glucoamylase. This can be done in a separate process, so no incompatibilities are present between the optimal pH and temperatures of the enzymes on one hand and the optimal pH and temperature of the microbes on the other. However, if the right combination of enzymes, microorganisms, pH, and temperature is carefully chosen, the hydrolysis and fermentation can be carried out in one reactor. This process is generally called SSF (simultaneous saccharification and fermentation) [55]. Prior to SSF, the starch granules usually must be gelatinized at high temperature by cooking. However, even a cooker is optional nowadays as commercial enzymes are becoming available that are able to attack and hydrolyze the granules efficiently and fast enough at relatively low temperatures. 1.2.4.7 Lignocellulose Sucrose and starch have in common that they are used for food and nowadays, with oil wells drying out and prices rising, also for biofuels. A decrease in the availability of fossil fuels is envisaged for the future, and with increasing population, more food is needed at reasonable prices. Therefore, the ideal raw material for biofuels and bioplastics is carbohydrates that are not edible. Such material is abundantly available around the globe as lignocellulose, like in corn stover or wheat straw. Lignocellulose consists of the glucose homopolymer cellulose, the heteropolymer hemicellulose, and lignin. Hemicellulose consists of hexoses
  • 8 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE and pentoses. In all, lignocellulose contains roughly 80% fermentable sugars, but this largely depends on the source [54]. The remainder, lignin, is a phenolic polymer that is difficult to degrade and is not directly usable for lactic acid production. It may be used for energy production though, which can be returned to the lactic acid plant. A purer source of cellulose without lignin is waste paper that can be used for lactic acid production at lab scale [56]. Thus, even this book can eventually be converted into PLA! Complete utilization of cellulose and hemicellulose requires selection or genetic modification of an organism that is able to ferment pentoses. In order to obtain monosaccharides from the raw material, several pretreatments and/or separations are required. First, the lignocellulosic material is mechanically treated and then delignified (pulped) by strong alkali or acid treatment. The (hemi)cellulose part becomes more accessible for enzymes at the same time. Subsequent enzymatic treatment mainly yields glucose and xylose and some arabinose. The enzymatic treatment and subsequent fermentation can be done in separate reactors or in one fermenter, in an SSF concept similar to starch SSF [57]. 1.2.4.8 Batch versus Continuous Fermentation A process can be run in batch or continuous mode. In continuous mode, there is a constant flow of fermented sugar out of the reactor that is equal to a continuous flow of fermentation medium into the reactor. During batch fermentation, there can be an inflow of medium, but there is no outflow [58]. Batch fermentation needs to be inoculated with a starter culture every time, whereas this is not needed in a continuous fermentation setup. However, in case of problems, the continuous fermentation needs to be restarted, so an infrastructure for starter cultures is needed anyway. A high volumetric production rate can be achieved when combining continuous FIGURE 1.5 fermentation with biomass retention, leading to smaller fermenter size [59]. It must be stated that the lactic acid concentration is lower compared to batch culture [58]. The concentration of lactic acid influences the water balance in the production plant. In all scenarios, microorganisms produce an aqueous lactic acid solution, comprising mainly lactate and counterions from the base, impurities from raw materials or fermentation by-products, residual sugars and polysaccharides, and the microorganism itself. 1.2.5 Downstream Processing/Purification of Lactic Acid When Scheele discovered lactic acid, he recovered and purified the lactic acid from sour whey by saturation with lime, filtering off the crude calcium lactate, acidifying the crystal mass with ‘‘acid of sugar’’ (oxalic acid), filtering off the calcium oxalate, and evaporating to obtain a crude viscous lactic acid [12, 13]. Basically, this process with a calcium-based neutralized fermentation and sulfuric acid instead of oxalic acid is the same process used in industry today for the production of crude lactic acid. Drawbacks are the continuously rising costs of lime/chalk, sulfuric acid, and other chemicals and the disposal of large quantities of gypsum (CaSO4Á2H2O), as an unavoidable side product of this technology. In such a process also the first down stream processing (DSP) step, biomass removal by filtration, can be accomplished relatively easily in a (mild) liming step, in essence quite similar to the traditional liming step to remove protein in sugar beet or sugarcane processing in sugar mills. A simplified block scheme of the traditional lactic acid production process including fermentation is shown in Figure 1.5. Simplified block scheme of traditional lactic acid production process.
  • LACTIC ACID TABLE 1.2 9 Summary of Lactic Acid Purification Methods Lactic Acid Purification Method Advantages Disadvantages Crystallization [27, 28] Esterification/distillation [52] Lactic acid distillation [27, 28, 53] Highly pure lactic acid product Highly pure acid, scale-up Good splitting for heavy compounds Potentially high yield Amount of mother liquor by-product, scalability Relatively high utility cost, amount of residue as by-product Amount of residue as by-product Extraction [54, 55] 1.2.5.1 Purification Methods for Lactic Acid Crude lactic acid, which may be upgraded by simple active carbon treatment and/or ion exchange to remove impurities and salts, can be directly used in a large number of food applications. Traditionally, taste, smell, and heat stability for color formation have been used to express lactic acid quality. The presence of acids (e.g., acetic acid and pyruvic acid), alcohols (e.g., methanol and ethanol), and esters can directly influence taste and smell [4]. The presence of residual sugar and nitrogen compounds greatly influences heated color, that is, browning of the liquid upon heating. The formation of color upon heating prohibits the use of crude acid in foods that need to undergo pasteurization/sterilization. Over the decades, the demand for purer lactic acid with improved color stability upon heating has increased, as exemplified by the need for ultrapure lactic acid as a sodium lactate base in pharmaceutical infusion products. At present, a chemical engineer can choose from a number of mature industrial methods to purify lactic acid. Table 1.2 lists their relative advantages and disadvantages. Choices in an overall process are governed by raw material costs, utility costs, and, last but not least, outlets for by-products. The purification methods described above each involve considerable technological know-how: . . . Esterification/Saponification. Esterification of lactic acid with methanol/ethanol yields systems with good separation characteristics to separate many impurities with different boiling points [60]. However, the energy demand of a full reaction/distillation route from crude acid to pure acid is high. Crystallization. Crystallization can yield an excellent lactic acid grade, but the yield is low. Lactic Acid Distillation. Industrial equipment is available to distill lactic acid at low vacuum. Higher molecular weight components such as sugar and protein will leave the system as a residue. Heat-stable lactic acid is obtained as the top product. In the stages of dewatering the crude lactic acid prior to distillation, the formation of oligomers will limit an overall high distillation yield. Complex (e.g., for emulsion, entrainment issues), extractant cost . Extraction. An extraction/back-extraction process, for example, with the well-described tertiary amine systems, is a suitable way to purify lactic acid [61, 62]. The possible combination of extraction with low-pH fermentation yields an elegant concept to arrive at a gypsum-free process. For future large-scale, low-cost lactide/PLA production, lactic acid DSP will need to meet new challenges: . . Use of Low-Cost and Nonedible Substrates. Whereas production of lactic acid from sucrose or glucose syrup is well established, crude sources (starches, sugars, or future lignocellulose hydrolysates) will form the next hurdle as they contain much more impurities and possible fermentation inhibitors. Gypsum-Free Processing. For large-scale, sustainable PLA production, a fermentation process that does not coproduce a mineral salt is a must. 1.2.5.2 Gypsum-Free Lactic Acid Production Gypsumfree lactic acid production can be briefly categorized as follows: . . Low-pH Fermentations Coupled to In Situ Product Removal. As discussed in Section 1.2.4, fermentations can be carried out without neutralization at pH 2–3 with genetically modified yeast or at pH 4 with LAB with partial neutralization [50]. When a separation method to recover the undissociated acid is integrated with fermentation, a process route can be designed in which no gypsum is produced. In the literature, a number of separation methods are described with an emphasis on extraction [63]. Cost efficiency in the fermentation (e.g., nutrients, yield) and the practical processing of large dilute streams need breakthroughs for economical processing. Electrochemical Splitting of a Neutral Lactate Salt. Numerous articles have described the splitting of a lactate salt, notably sodium lactate, into lactic acid and the original base [64]. With this principle, a gypsumfree process can be designed, with electrodialysis
  • 10 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE . separate from or integrated with fermentation. The use of electrodialysis with new bipolar membranes is straightforward, but a large-scale commercial breakthrough as in the 1980s and 1990s with monopolar membranes for the chloro-alkali process is still pending. Electrodialysis involves relatively high electricity costs and a huge membrane area, but these costs may be managed in biorefinery concepts with integrated energy production. Chemical Salt Splitting of a Lactate Salt. Lactate salts can be split with the help of auxiliary chemicals and the regeneration of these chemicals. A patent by Baniel et al., for example, describes a method in which a sodium lactate solution is acidified with CO2 under pressure, and simultaneously undissociated lactic acid is extracted and insoluble sodium bicarbonate (NaHCO3) is formed [65]. Another patent describes the splitting of ammonium lactate by esterification with butanol while liberating ammonia [66]. In the distillation process, the butyl lactate can be hydrolyzed with water to liberate lactic acid. This is an interesting option, but the energy consumption and side reactions such as the formation of lactamide and racemization require attention. Chemical salt splitting processes with the recycle of chemicals can be complex, but it is a challenge to develop a system with straightforward chemistry, high yield, low energy consumption, and good scaleability. 1.2.5.3 Modern Industrial Methods In overall process development, knowledge about dealing with impurities will be important. Residual sugar in the broth and sugar degradation products play a role throughout the process at the various levels of temperature and acidity. Color may be formed at any step from low- to high-boiling color precursors. Volatile acids such as acetic acid and formic acid will partition throughout DSP and their concentration in recycle streams must be prevented. In the design of a modern lactic acid plant, mathematical models are indispensable. For example, the kinetic model of oligomerization of lactic acid and the right thermodynamic model for the gas/liquid equilibria are important in design for the concentration of lactic acid by evaporation as well as for prepolymerization in the lactide route. Lactic acid solutions and vapors are quite corrosive and knowledge of the material of construction is a must for a lowmaintenance plant. Also, wastewater treatment is an integral part of a lactic acid plant. Aerobic systems are state of the art, but anaerobic systems are increasingly used to treat acidcontaining wastewater streams. The biogas can then be profitably used for steam production. While a plant using sucrose has a net intake and net purge of water, future plants using crude, low-cost, water-rich substrates will need to pay more attention to the water balance and wastewater treatment. Although the fermentation industry can be considered traditional, new technologies may quickly find uses. The rapid commercial application of filtration techniques such as in membrane bioreactors in wastewater treatment and the fast introduction of nanofiltration for making process water from river water are examples. The discovery of ionic liquids with high distribution coefficients for lactic acid in dilute solutions may lead to breakthroughs [67]. New steam boiler concepts that can handle residues can drastically change DSP layout in energy-efficient integrated biorefineries. 1.2.6 Quality/Specifications of Lactic Acid The dehydration of lactic acid to make the prepolymer should start with an ÀOH to ÀCOOH ratio of 1:1. All other components with ÀOH and ÀCOOH functionality disrupt the stoichiometric balance and may be incorporated as comonomers during prepolymerization, which limits the final lactide production yield from lactic acid. Little public information is available on the technical and economic relationship between lactic acid quality and lactide synthesis. Only a few patents mention the effect of metal impurities on racemization [68, 69]. Stereochemical purity is one of the key parameters determining lactic acid purity. Lactic acid purified by crystallization may be taken as the benchmark in lactide manufacture, but the expected unfavorableeconomicsofmakingcrystalline acid inrelationto mother liquor processing may prevent its commercial use for lactide/ PLA. The next level of quality with the right commercial relevance is heat-stable lactic acid. Heat stability puts constraints on the content of sugar, and thus on the DSP method used in the process. It is unlikely that suitable acid for making lactide will contain sugar because of the high temperatures involved (see the next section) and the well-known practical decomposition problems when sugars are cracked. In practice, this means that color, or actually heated color (color after heating of the acid), is an important indicator for the suitability of the acid for lactide/PLA production [6, 70]. The appeal for lactic acid with little or no sugar and the DSP methods mentioned in practice lead to demands for separation methods that are similar for sugar and other heavy components such as proteins, amino acids, and polysaccharides. It is expected that the desired quality of lactic acid for making lactide/PLA will evolve, with overall process yields and economics as the criteria. 1.3 1.3.1 LACTIDE Physical Properties of Lactide The dehydrated, cyclic dimer of lactic acid is commonly called lactide (3,6-dimethyl-1,4-dioxane-2,5-dione). Due to
  • LACTIDE described in a patent by Gruter and Pohl in 1914 [72]. Lactic acid was self-esterified at 120–135 C, and air was drawn in to remove the water. Next, zinc oxide was added as a catalyst and lactide was distilled off under vacuum at 200 C. In practice, modern industry cannot dispense with this concept of thermal catalytic depolymerization for lactide production. A major step forward was the use of a tin catalyst, a frequently used coordinating catalyst in polymerizations, in the process. The general scheme of lactide manufacture including the purification is shown in Figure 1.7. In the past two decades, several papers have appeared on lactide manufacture [73, 74]. A main underlying problem in understanding all information is that the reaction from oligomer to lactide is an equilibrium reaction. In order to pull the reaction toward the right, lactide must be withdrawn from the system. In reaction engineering terms, this means that the chemical kinetics of the reaction cannot be understood without consideration of the method and efficiency of lactide removal. In terms of know-how described in patents, this means that reported lactide production rates depend to a large extent on the geometry of the equipment in which lactide synthesis is performed and that provides for removal of lactide vapor from the reaction zone. In modern chemical technology, one of the goals is to fully understand a given system, capture the knowledge in models to describe experimental work, and ultimately use these models to design, optimize, and debottleneck large-scale FIGURE 1.6 The three diastereomeric structures of lactide (3,6dimethyl-1,4-dioxane-2,5-dione). the two asymmetric carbon atoms in the molecule, lactide exists in three different forms (Figure 1.6). In addition to the three diastereomeric structures mentioned above, also a racemate of D-lactide and L-lactide exists: rac-lactide or DL-lactide (Table 1.3). 1.3.2 Production of Lactide The synthesis of lactide was first described by Pelouze in 1845 [71]. He investigated the self-esterification of lactic acid by heating and driving off water and obtained a prepolymer that was no longer fully miscible with water. Upon continued heating of the prepolymer, he noticed that in a certain distillate fraction nice crystals were formed. He was able to deduce the chemical formula and gave the name ‘‘lactid’’ to the substance. An improved procedure was TABLE 1.3 11 Physical Properties of the Lactides Unit CAS number Molecular weight Melting point Boiling point Heat of fusion Heat of vaporization Solid density Liquid viscosity FIGURE 1.7 g/mol  C  C J/g kJ/mol g/mL mPa s D-Lactide 13076-17-0 144.12 96––97 L-Lactide [6] meso-Lactide 4511-42-6 144.12 96 13076-19-2 144.12 53 [64] 146 63 1.32–1.38 2.71 (110 C); 2.23 (120 C); 1.88 (130 C) 128 [64]; 118 [6] rac-Lactide 116559-43-4 125 [6] 142 (20 mbar) [64] 185 [6] 1.32–1.38 [6] Schematic illustration of lactide manufacture by thermal catalytic depolymerization of lactic acid oligomers.
  • 12 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE equipment. For the present system, this means that one must develop process know-how on chemical kinetics and thermodynamics of lactide and HL oligomers, and on physical phenomena related to equipment design. These aspects will be relevant for both the prepolymerization and the synthesis of lactide, as these chemical systems are highly similar. In practice, however, lactide synthesis is more complex as chemistry, recovery and type of equipment are intertwined, and the viscous nature of reaction mixtures requires special attention. With these aspects in mind, the information on the lactide synthesis that can be found in the literature is summarized below. 1.3.2.1 Prepolymerization A general procedure for batch prepolymerization is described in a patent by O’Brien et al. [75]. Typically, vacuum pressures of 70–250 mbar and temperatures up to 190 C are used to dewater lactic acid to a prepolymer with an average degree of polymerization (DP) of around 10 in a batch process time of 6 h. For lab-scale equipment, it was also found that thin film and rotating flask vacuum equipment showed faster reaction times than a stirred tank, indicating the importance of mass transfer of water in the already viscous prepolymer. Continuous prepolymerization has also been described in a number of patents, for example, in stirred tanks in series or in evaporator-type equipment [68, 76, 77]. Usually patents describe prepolymers with a DP of 7–20 as feed to the lactide synthesis. Using modern HPLC methods, it has been shown that in oligomeric systems up to DP 10, an equilibrium is present with constant equilibrium constants between the oligomers [6, 72]. 1.3.2.2 Lactide Synthesis During Prepolymerization Because the composition of a mixture comprising lactic acid oligomers and lactide is governed by chemical equilibria, a prepolymerization exhibits relatively high concentrations of lactide (HL2–H2O–L2 equilibrium) around DP 2. Sinclair et al. distilled these fractions to recover lactide, but the crude lactide was quite impure, which may prevent economical processing [73]. In hindsight, the patent describes trials to optimize Pelouze’s original lactide synthesis without catalyst [71]. (stannous 2-ethylhexanoate) is a liquid catalyst that can be handled easily, is food grade, and is widely available. Thinking in terms of mechanisms, the equilibrium concentration of lactide in an oligomer mixture is 5% or less, and it will boil off at low vacuum [6, 68]. The catalyst increases the rate of lactide formation by facilitating lactide formation by backbiting from hydroxyl chain ends of oligomers [4, 74]. In a batch experiment, the rate is initially constant, but during synthesis esterification also occurs, and the DP of the polyester rises concomitantly. The melt viscosity of the reaction mixture increases accordingly and at the end of a batch process, mixing the highly viscous residue becomes very difficult, which limits the extent to which the residue can be depleted of lactide. In engineering terms, this means that mass transfer of lactide from the liquid to the gas phase decreases as viscosity increases. The balance between lactide production and lactide removal plays a role in all experiments that one might want to investigate on lab scale. For example, catalyst concentrations of 0.05–0.2 wt% tin(II) octoate are mentioned in the literature, but traditional experiments to verify the order of the reaction for the catalyst are difficult because of the influence of mass transfer limitations. 1.3.2.4 Continuous Synthesis In 1992, Gruber et al. [68] described a continuous lactide synthesis in which prepolymer is fed continuously to a reactor, crude lactide is evaporated under vacuum, and residue is removed. Typical operating conditions for the reactor were residence time around 1 h, vacuum pressure 4 mbar, temperature 213 C, and catalyst amount 0.05 wt% tin(II) octoate on feed. The conversion per pass was around 70%, and the overall yield was increased by recycling the residue to the lactic acid section of the process, where the oligomers are hydrolyzed again. Especially in the patent literature, several different reactor types are described for continuous lactide synthesis: . . . . 1.3.2.3 Basic Research on Batch Lactide Synthesis and the Catalysts Used Noda and Okuyama reported on the batch synthesis of lactide from DP 15 prepolymer with various catalysts at 4–5 mbar and 190–245 C [74]. In a batch synthesis with 50 g of oligomer in a stirred flask, the evolution rate of crude lactide is rather constant and then starts to decline and the conversion levels off at 80–90%. The tin catalyst performed best compared to other catalysts and showed the lowest levels of racemization. Tin octoate . Stirred tank reactor with different stirrer types [76]. On a bench scale, the reactor is jacketed for heating. Stirred reactor with a distillation section on top of the reactor to fractionate the product [50]. Thin film evaporator with a typical conversion of 80% on pilot scale [70]. Horizontal wiped film evaporator. In a patent by Kamikawa et al. [77], the use of horizontal wiped film is described. In the horizontal mode, the residence time of the reaction mixture can be controlled and a conical form is used in which wipers transport the viscous residue. Distillation column. In a patent by O’Brien et al. [75], a distillation column with perforated plates and optional use of packing material and heating on the stage are described. In an experiment with a single tray, a DP 10 feed was fed to the top, and N2 was used to strip the
  • LACTIDE lactide from the liquid. At different residence times, the conversion on the tray could be as high as 93% at 210–215 C. In other patents, the use of N2 gas as a stripping agent is mentioned, but it is to be expected that in large-scale equipment the processing of large amounts of inert gases will be less economical compared to the use of vacuum systems. Some metal cations such as sodium and potassium in the feed increase racemization risk, while other metals (Al, Fe) are catalytically active in transesterification, resulting in competitive polylactide formation [68, 69]. Through corrosion, metals may be released in the residue and will build up there [6, 75]. Some patents discuss the presence of acid impurities in the process [6, 7, 67, 78]. Mono- and dicarboxylic fermentation acids are responsible for stoichiometric imbalance in the lactic acid polycondensation reaction. Consequently, the composition of the obtained lactic acid oligomer chains can differ from pure PLA, resulting in impeded and incomplete catalytic depolymerization of the oligomers into lactide. In PLA manufacture, degradation reactions play a role, mainly via intramolecular chain scission, and this may also affect lactide synthesis. Reviewing the literature provides a list of process aspects that need consideration in the design of a solventless synthesis operated with vacuum equipment. . . . . . . Temperature. Intrinsic reaction rates increase with temperature. At higher temperature also, the vapor pressure of lactide above the reaction mixtures increases. The reaction rate of racemization will also increase with temperature. In Witzke’s Ph.D. study, information on activation energies can be found [6]. Pressure. Pressures of 10 mbar or less are used. At higher pressures, the driving force for lactide evaporation will be lower, and the overall reaction rate will be lower. Low pressures will require detailed considerations of equipment size, vacuum systems, condensers, and so on. Feed DP. The feed DP has two effects. First, a low DP feed will contain more monomer lactic acid that boils at a lower temperature than lactide, and this will contaminate the crude lactide distilled off from the reactor. Also, monomer lactic acid can be released from DP 3 with the catalyst, leading to more acidity in the crude lactide. Second, it is to be expected that at a higher feed DP the residue in the reactor will have a higher DP and viscosity with consequences for equipment design. The influence of prepolymer DP on the meso-lactide level formed during lactide synthesis was discussed by Gruber et al. [69]. Increasing feed DP clearly resulted in a decrease in the lactic acid concentration in the crude lactide. A drawback is that the meso-lactide concentration also increased significantly. Catalyst Concentration. More catalyst will increase the overall reaction rate. In practice, this effect may not be linear, since next to kinetics mass transfer in the equipment will play a role. Racemization. In the production of stereochemically pure lactide, formation of the other lactic acid enantiomer and meso-lactide is unwanted. Higher temperatures, longer reaction times, and increased catalyst levels result in increased rates of racemization [4, 6, 69]. Since temperature and catalyst influence the rate of lactide formation as well, controlling the racemization rate can become quite complex. Impurities. Data in the literature on the role and fate of impurities from the feed in the synthesis are scarce. 13 On the one hand, it can be concluded that the lactide synthesis is straightforward in the sense of making a prepolymer and releasing lactide by thermal catalytic depolymerization at low pressure. On the other hand, it can be concluded that the scale-up from a lab-scale process to an economical, large-scale process with high yield and no compromises on stereochemical purity is a complex multifaceted task. 1.3.3 Purification of Lactide A lactide synthesis reactor invariably produces a crude lactide stream that contains lactic acid, lactic acid oligomers, water, meso-lactide, and further impurities. The specifications for lactide are stringent mainly for free acid content, water, and stereochemical purity. Basically, two main separation methods, distillation and crystallization, are currently employed for lactide purification: . Distillation. Splitting the multicomponent mixture consisting of lactide, water, lactic acid, and its oligomers into pure fractions requires considerable knowhow on kinetics and operation of vacuum equipment. Distillates and bottoms may be recycled, but the accumulation of impurities from the feed or the production of meso-lactide during the process requires careful finetuning of temperatures and residence times. Distillation is well described in the patent by Gruber et al. in 1993 [68]. The crude lactide from the synthesis is distilled in the first column to remove the acids and water, and then meso-lactide is separated from lactide in the second column. As the boiling points of all compounds are in the range of 200–300 C, low pressures are used. Since the difference in boiling temperature of lactide and meso-lactide is quite small, this distillation requires a lot of theoretical stages (>30). The Cargill/ NatureWorks distillation uses a series of distillation
  • 14 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE . . columns and is performed continuously [4]. Part of the distillation can also be integrated with the reaction [79]. Solvent Crystallization. A commonly used laboratory method for lactide purification is recrystallization from mixtures of toluene and ethyl acetate [4]. Lactide of extremely high purity can be obtained by repeated crystallization with different toluene/ethyl acetate ratios. Several patents also mention the use of solvents for the crystallization of lactide, but for large scale, melt crystallization without the use of solvents is preferred. Melt Crystallization. Lactide crystallizes easily and several patents describe how crystallization can yield lactide with required specifications regarding lactic acid content, oligomers, meso-lactide, and water. An early patent describes such a crystallization method and includes some information on the thermodynamic equilibria (eutectica) of the lactide/lactic and the lactide–meso-lactide system, which define the maximum yield as a function of these impurities in the feed [80]. In patents, the use of different types of equipment is mentioned: static equipment, falling film crystallizers, vertical column with scraper to remove crystal mass from the cooled wall, and scraped heat exchanger coupled to a wash column [70, 80, 81]. For large scale, it is a challenge to design and scale-up the crystallization equipment with respect to the needed heat transfer areas and hydrodynamics, and the possible increase of viscosity of mother liquor by oligomerization of lactide and residual acid. The choice between distillation, crystallization, or novel separation methods such as absorption or membrane separation is determined by the desired stereochemical purity of the product. Crystallization yields highly pure lactide, suitable, for example, for high-melting PLLA homopolymer of high molecular weight. Affordable distillation equipment does not fully remove all meso-lactide, and consequently, a lactide monomer mixture for PLA copolymers with other thermal properties is obtained upon ring-opening polymerization. The design of the separation system relies on detailed knowledge of the thermodynamic properties of the compounds and the kinetics of the reactive system. Obtaining FIGURE 1.8 such know-how requires sophisticated analytical methods for lactic acid and its oligomers, lactides, and residues. Impurities can also be formed in lactide synthesis, similar to PLA degradation reactions, and gas chromatography (GC) methods are needed to identify these compounds and determine their fate in the process. 1.3.4 Quality and Specifications of Polymer-Grade Lactide The specifications and allowed impurity levels of lactide monomer for PLA are defined by the polymerization mechanism and the applied catalyst. PLA is commercially produced by ROP of lactides in bulk. The tin(II)-catalyzed process offers good control over molecular weight and reaction rate provided that it is performed in the absence of impurities such as water, metal ions, lactic acid, or other organic acids. Purification of crude lactides is therefore indispensable for the industrial manufacture of high molecular weight PLA (Mw > 100 kg/mol). In fact, lactide is the ultimate form of lactic acid, in its dehydrated and purest form. 1.3.4.1 Role of the Catalyst and Initiator in Lactide Polymerization The theoretical description of the Sn(Oct)2-catalyzed ROP of cyclic esters has been studied by many authors, but there does not appear to be a theory that consistently explains all experimental results of the coordination–insertion polymerization [3, 4, 82–84]. Different polymerization mechanisms may dominate, depending on polymerization conditions, catalyst and initiator concentration, and the presence of a solvent. Here it is assumed that lactide is polymerized in bulk with Sn(Oct)2—a Lewis acid—and that the mechanism follows the model proposed by Kowalski et al. [84]. Since lactide is a cyclic ester, its ring can be opened by nucleophilic attack on the ester bond to start polymerization. Suitable initiators (nucleophiles) are water and alcohols, including the hydroxyl group of lactic acid. One ester linkage of a lactide ring is cleaved by reaction of the OH group of the initiator R-OH, creating a new R-O-C(O)- ester end group and an OH end group (Figure 1.8). Every initiating molecule is covalently bonded as an end group to each polymer chain [84]. Via transesterification Ring-opening polymerization of lactide to PLA initiated by an alcohol.
  • LACTIDE FIGURE 1.9 Equilibrium reaction of tin octoate with alcohol initiator or impurities to form catalytically active tin alkoxide bonds Sn-O-R [76]. reactions, the 2-ethylhexanoate ligands of the SnOct2 catalyst will also end up as octanoic ester groups in the polymer. In some papers, the Sn(II) catalyst is indicated as the initiator, presumably because lactide also polymerizes upon addition of that substance, and the effect of impurities is overlooked. An initiator—or coinitiator—is a substance that can start polymerization, in the case of lactide by opening the lactide ring, and thus offers control over molecular weight. This has to be a nucleophile and cannot be the Sn catalyst itself, as supported by the excellent work of Kowalski et al. who proved that SnOct2 needs activation with R-OH (Figure 1.9) [84]. In a nutshell, the total hydroxyl content, including R-OH initiator and lactic acid impurities, determines the maximum attainable Mn (number-average molecular weight) [4, 6]. The rate of polymerization is controlled by factors such as temperature and catalyst content, with the remark that a tin (II) octoate catalyst requires traces of the initiator to become active. 1.3.4.2 Alcohols If water is the initiator, R equals H and hydrolysis of lactide produces lactoyl lactic acid (HL2). Propagation with lactide in the presence of a polymerization catalyst produces PLA with a hydroxyl and one carboxylic acid end group, as if the PLA was obtained by polycondensation of lactic acid. If the hydroxyl group of lactic acid acts as an initiator, PLA with one hydroxyl end group and a lactic acid end group (HOOC-CH(CH3)-O-C(O)-) is obtained. If the initiator itself is polymeric in nature, for example, polyethylene glycol (PEG), lactide can polymerize from the hydroxyl end group(s) of PEG resulting in PEG–PLLA diblock or triblock copolymers. The molar ratio of monomer to initiator (M/I)—where initiator can also be read as total hydroxyl content—basically controls the final, average molecular weight (Mn) of the PLA. A high amount of initiator produces short polymer chains, and a low amount of initiator produces high molecular weight polymer. The lower the amount of potentially initiating hydroxyls in the lactide monomer, the higher the maximum attainable degree of polymerization [69]. Since water and lactic acid can both cause ring scission of the lactide and initiate polymerization, their amounts in the lactide must be low and should be specified. 1.3.4.3 Carboxylic Acids Carboxylic acids are poor initiators, but they are believed to interfere with the commonly used Sn(II) polymerization catalyst. According to Kowalski, 15 carboxylic acids may suppress the rate of polymerization by shifting the equilibrium between ROH and Sn(Oct)2 to the inactive Sn(Oct)2 side [83, 84]. Consequently, longer polymerization times are needed to achieve the desired molecular weight, accompanied by unavoidable degradation caused by the extra residence time at high temperature in the presence of a catalyst [84]. The effect of carboxylic acids on lactide polymerization rate was published in 1993 in patents by Ford and O’Brien [78, 85]. The results clearly show the dramatic rate-decreasing effect of organic acids: according to O’Brien, melt polymerization slows down by a factor of 2 upon increasing free acidity from less than 2 to between 2 and 4 meq/kg [85]. Witzke, however, states that the presence of lactic acid did not negatively influence polymerization rate [4, 6]. Lactic acid is therefore a practically used initiator that is already present in lactide as an impurity. Lactic acid and its oligomers have a hydroxyl group and a carboxylic acid group. Consequently, a free acidity of 10 meq/kg—that is, 900 ppm expressed as lactic acid equivalents—in lactide corresponds to a hydroxyl concentration that limits Mn to 100 kg/mol. Free acidity of 4 meq/kg sets a theoretical limit of 250 kg/mol to Mn. Free acid and water content specifications are essential for any lactide grade; the lower the amount of hydroxyl impurities, the better the storage stability and product properties of the lactide. 1.3.4.4 Metals Metal cations such as Sn, Zn, Fe, Al, and Ti not only accelerate polymerization, but can also affect hydrolysis, oxidation, racemization, or other degradation mechanisms of PLA and lactides [4, 6]. Consequently, the lactic acid used for lactide preparation should be very low (ppm) in metal cations in order to avoid considerable racemization during lactide synthesis. O’Brien has shown that the formation of dark color of lactide was a direct function of the iron content of the material in which the lactide was in contact [86]. Other examples in the patent (Examples 7 and 8) demonstrate the desirability of having low alkali (e.g., sodium) content and minimizing the depolymerization temperature. Cationic impurities such as sodium ions have no direct effect on lactide production rate, but the sodium content has a direct correlation with the meso-lactide content in the crude lactide [67, 87]. 1.3.4.5 Stereochemical Purity The higher the stereochemical purity of the lactide monomer, the higher the stereochemical purity of the obtained PLA, which controls material properties such as melting point, crystallinity and crystallization rate, and mechanical strength [8, 9, 88]. The strong dependence on D-isomer content presents an opportunity to control polymer properties. NatureWorks
  • 16 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE Ingeo PLA is easily processable and suitable as amorphous biopackaging material as a result of its relatively high mesolactide content. The downside is the poor resistance to elevated temperatures (low heat distortion temperature, HDT) during transportation, storage, and use of articles produced from this bioplastic. meso-Lactide—which contains an L- and a D-isomer—is an unavoidable side product of lactide production and must be separated from L- and D-lactides of high stereochemical purity. Kolstad [9] investigated the crystallization behavior of copolymers of L-lactide and meso-lactide. He found that every 1% of meso-lactide comonomer—or D-isomer— causes a 3 C reduction in the melting point of the PLA copolymer. With 3% meso-lactide in PLA, crystallization is more than two times slower than PLLA under the same conditions. With 6% meso-lactide incorporation, the difference can be up to 10 times! This underlines the need for a low meso-lactide content in the monomer mixture for semicrystalline PLA, because meso-lactide formation by racemization cannot be avoided during melt polymerization of lactides. According to Gruber and coworkers, racemization, which lowers the stereochemical purity of the PLA, is believed to be driven by factors such as temperature, pressure, time at a given temperature or pressure, the presence of catalysts or impurities, and relative concentrations of the two enantiomers at any given time during the polymerization process [88]. PLA grades for more demanding applications that require better heat resistance are achievable by stereocomplexation with PDLA [89]. This is only effective with PLA grades of high stereochemical purity. In order to prepare high-quality PLA, it is necessary to start with lactide monomers with the highest possible stereochemical purity, that is, the lowest meso-lactide content that is technically and economically achievable by purification. D-Lactide can be obtained if one has the appropriate biochemistry to produce the D-enantiomer of lactic acid by fermentation of carbohydrates. Copolymerization of controlled mixtures of L- and D-lactides subsequently offers the advantage of precise control over PLA properties. Moreover, D-lactide is the monomer for the production of poly(D-lactide), which is able to form high-melting stereocomplex PLA via 1:1 racemic cocrystallization with P(L)LA, as will be discussed in Chapter 5 [89]. 1.3.5 Concluding Remarks on Polymer-Grade Lactide In conclusion, the most important quality specifications for lactide monomers are those of free acidity, water, metal ion content, and stereochemical purity. . Free acidity, for example, lactic acid or lactoyl lactic acid, slows down the rate of polymerization and limits the achievable degree of polymerization. According to . . . the patent literature, free acidity of polymer grade lactide should be <10 meq/kg, and preferably no more than 5 meq/kg. Water causes hydrolysis of lactide and also limits the attainable degree of polymerization of PLA. Metal ions need to be specified in low quantities, because Sn, Zn, Fe, and Al cations accelerate polymerization, but may also affect hydrolysis, oxidation, or other degradation mechanisms. Sodium in particular causes racemization even in ppm amounts. Stereochemical purity expresses the sum of meso-lactide and D-lactide in L-lactide and vice versa. The higher the stereochemical purity of the lactide monomer, the higher the stereochemical purity of the obtained PLA, which controls material properties such as melting point, crystallinity, and mechanical strength. REFERENCES 1. R. Narayan, Drivers & rationale for use of biobased materials based on life cycle assessment (LCA), GPC 2004 Paper Abstract #18, Michigan State University, 2004. 2. D. Garlotta, J. Polym. Environ. 2001, 9(2), 63–84. 3. A. P. Gupta, V. Kumar, Eur. Polym. J. 2007, 43, 4053–4074. 4. M. H. Hartmann, High molecular weight polylactic acid polymer, in: D. L. Kaplan (Ed.), Biopolymers from Renewable Resources, Springer, Berlin, 1998, Chapter 15, pp. 367–411. 5. J. Nieuwenhuis, Clin. Mater. 1992, 10, 59–67. 6. D. R. Witzke, Introduction to properties, engineering, and prospects of polylactide polymers, Ph.D. thesis, Department of Chemical Engineering, Michigan State University, East Lansing, MI, 1997. 7. D. E. Henton, P. Gruber, J. Lunt, J. Randall, Polylactic acid technology, in: A. K. Mohanty, M. Misra, L. T. Drzal (Eds.), Natural Fibers, Biopolymers, and Biocomposites, CRC Press, 2005, Chapter 16, pp. 527–577. 8. D. W. Grijpma, A. J. Pennings, Macromol. Chem. Phys. 1994, 196, 1649–1663. 9. J. J. Kolstad, J. Appl. Polym. Sci. 1996, 62, 1079–1091. 10. R. G. Sinclair, E. S. Lipinsky, U.S. Patent PATN 5,502,158, 1996 (to Ecopol LLC). 11. Cargill/NatureWorks LLC Press Release, July 1, 2009, available at http://www.natureworksllc.com (accessed date: July 24, 2009). 12. C. W. Scheele, Kgl. Vetenskaps-Academiens nya Handlingar (Stockholm) 1780, 1, 116–124. 13. L. Dobbin, The Collected Papers of Carl Wilhelm Scheele, G. Bell & Sons Ltd, London, 1931. 14. C. H. Holten, A. M€ller, D. Rehbinder, Lactic Acid, Verlag u Chemie, Weinheim, 1971. 15. H. Benninga, A History of Lactic Acid Making, Kluwer Academic Publishers, Dordrecht, 1990.
  • REFERENCES 16. C. Avery, U.S. Patent 243,827, 1881 (to Avery Lactate Company). 17. F. Kraft, W. A. Dyes, Ber. Dtsch. Chem. Ges. 1895, 28, 2589–2597. 18. H. Borsook, H. M. Huffman, Y. P. Liu, J. Biol. Chem. 1993, 102, 449–460. 19. A. Schouten, J. A. Kanters, J. van Krieken, J. Mol. Struct. 1994, 323, 165–168. 20. PURAC internal data. 21. G. Saville, H. A. Gundry, Trans. Faraday Soc. 1959, 55, 2036–2038.  ˇ 22. A. Sepitka, Prumysl Potravin 1961, 13, 661–665. 23. R. A. Troupe, W. L. Aspy, P. R. Schrodt, Ind. Eng. Chem. 1951, 43, 1143–1146. 24. W. Ostwald, Z. Phys. Chem. 1889, 3, 170–197, 241–288, 369–22. 25. H. M. Huffman, E. L. Ellis, H. Borsook, J. A. Chem. Soc. 1940, 62, 297–299. 26. G. S. Parks, S. B. Thomas, D. W. Light, J. Chem. Phys. 1936, 4, 64–69. 27. J. van Breugel, J. van Krieken, A. Cerda Baro, J. M. Vidal Lancis, M. Camprubi Vila, WO 00/56693, 2000 (to PURAC). 28. J. van Krieken, WO 02/022546, 2002 (to PURAC). 29. W. P. Hammes, C. Hertel, The genera Lactobaccilus and Carnobacterium, in: M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, E. Stackebrandt (Eds.), The Prokaryotes, 3rd edition, Springer, New York, 2006, pp. IV/320 ff. 30. M. T. Madigan, J. M. Martinko, J. Parker, Brock Biology of Microorganisms, 9th edition, Prentice Hall, New Jersey, 2000, pp. 118–121. 31. O. Kandler, Antonie van Leeuwenhoek 1983, 49, 209–224. 32. C. Plumed-Ferrer, K. M. Koistinen, et al., Appl. Environ. Microbiol. 2008, 74, 5349–5358. 33. M. H. Saier, J. J. Ye, et al., J. Bacteriol. 1996, 178, 314–316. 34. A. M. Rodas, et al., Int. J. Syst. Evol. Microbiol. 2006, 56, 513–517. 35. M. Cocaign-Bousquet, C. Garrigues, P. Loubiere, N. D. Lindley, Antonie van Leeuwenhoek 1996, 70, 253–267. 36. A. P. Oliveira, J. Nielsen, J. F€rster, BMC Microbiol. 2005, 5, 39. o 37. B. Teusink, et al., J. Biol. Chem. 2006, 281, 40041–40048. 38. N. V. Narendranath, et al., Appl. Environ. Microbiol. 1997, 63, 4158–4163. 39. W. F. Kemper, et al., Proc. Natl. Acad. Sci. USA 2001, 98, 723–728. 40. Z. Ying Zhang, B. Jin, J. M. Kelly, Biochem. Eng. J. 2007, 35, 251–263. 41. A. Vaidya, R. Pandey, S. Mudliar, M. Kumar, T. Chakrabarti, S. Devotta, Crit. Rev. Environ. Sci. Technol. 2005, 35, 429–467. 42. O. Michio, K. Kimitoshi, Jpn. Patent 61293388, 1986 (to Daicel Chemical Industries). 43. J. P. de Boer, et al., Appl. Environ. Microbiol. 1993, 59, 2474–2478. 44. V. Rajgarhia, et al., WO 03/102201, 2003 (to Cargill Dow LCC). 17 45. M. Wada, et al., WO 05/033324, 2005 (to Mitsui Chemicals). 46. G. Bustos, A. B. Moldes, J. M. Cruz, J. M. Dominquez, J. Agric. Food Chem. 2004, 52, 801–808. 47. M.-T. Gao, et al., Bioresour. Technol. 2008, 99, 3659–3664. 48. M.-T. Gao, M. Hirata, E. Toorisaka, T. Hano, Bioresour. Technol. 2007, 97, 2414–2420. 49. T. J. Carlson, E. M. Peters, U.S. Patent 6,475,759, 2002 (to Cargill Inc.). 50. S. Saitoh, et al., Appl. Environ. Microbiol. 2005, 71, 2789–2792. 51. A. J. A. van Maris, et al., Appl. Environ. Microbiol. 2004, 70, 2898–2905. 52. W. N. Konings, et al., Antonie van Leeuwenhoek Int. J. Gen. Mol. Microbiol. 1997, 71(1–2), 117–128. 53. D. Visser, J. van Breugel, J. M. de Bruijn, P. A’Campo, WO 08/000699, 2008 (to PURAC Biochem BV). 54. R. L. Whistler, J. N. BeMiller, Carbohydrate Chemistry for Food Scientists, 1st edition, American Association of Cereal Chemist Inc., St. Paul, 1997, 117 pp. 55. R. Anuradha, A. K. Suresh, K. V. Venkatesh, Process Biochem. 1999, 35, 367–375. 56. E. Y. Park, P. Ngoc Anh, N. Okuda, Bioresour. Technol. 2004, 93, 77–83. 57. S. I. Abe, M. Tagaki, Biotechnol. Bioeng. 1991, 37, 93–96. 58. S. Ding, T. Tan, Process Biochem. 2006, 41, 1451–1454. 59. E. Ohleyer, H. W. Blanch, C. R. Wilke, Appl. Biochem. Biotechnol. 1985, 11, 317–332. 60. A. A. Dietz, E. F. Degering, H. H. Shopmeyer, Ind. Eng. Chem. 1947, 39, 82–85. 61. B. I. Veldhuis-Stribos, et al., WO 0127064 A1, 2000 (to PURAC). 62. J. van Krieken, et al., WO 05123647A1, 2005 (to PURAC). 63. A. M. Eyal, et al., WO 9919290, 1999 (to Cargill Inc.) 64. M. Bailly, Desalination 2002, 144, 157–162. 65. A. M. Baniel, et al., U.S. Patent 5,510,526, 1994 (to Cargill Inc.). 66. A. Kumagai, et al., EP 0614983A3, 1996 (to Musashino). 67. J. Martak, S. Schlosser, Sep. Purif. Technol. 2007, 57, 483–494. 68. P. Gruber, J. Kolstad, et al., U.S. Patent 5,338,822, 1992 (to Cargill Inc.); U.S. Patent 5,258,488, 1993 (to Cargill Inc.). 69. P. Gruber, J. Kolstad, et al., U.S. Patent 6277951B1, 1999 (to Cargill Inc.). 70. P. Coszach, J. C. Bogaert, F. van Gansberghe, U.S. Patent 0014975A1, 2006 (to Galactic). 71. J. Pelouze, J. Chem. Pharm. 1845, 53, 112–124. 72. H. Pohl, U.S. Patent 1,095,205, 1914 (to Gruter). 73. R. G. Sinclair, R. A. Markle, R. K. Smith, U.S. Patent 9,205,167, 1992 (to Battelle). 74. M. Noda, H. Okuyama, Chem. Pharm. Bull. 1999, 47, 467–471. 75. W. O’Brien, L. A. Cariello, T. F. Wells, WO 9606092, 1995 (to Ecological Chemical Products Company). 76. T. Matsuo, et al., U.S. Patent 25222379A1, 2005 (to Hitachi/ Toyota).
  • 18 PRODUCTION AND PURIFICATION OF LACTIC ACID AND LACTIDE 77. M. Kamikawa, et al., EP 1873185A1, 2006 (to Hitachi). 78. Th. M. Ford, U.S. Patent 5,310,599, 1993 (to Du Pont de Nemours & Co.). 79. J. Meerdink, A. S€dergard, WO 05056509A1, 2003 (to Hycail). o 80. I. D. Fridman, J. Kwok, U.S. Patent 5,264,592, 1993 (to Camelot Industries). 81. R. U. Scholz, R. P. M. van der Steen, (2007), WO/2007/148975, Purification of Lactide Streams. 82. H. R. Kricheldorf, et al., Macromolecules 2000, 33, 702–709. 83. X. Zhang, et al., J. Polym. Sci. Part A 1994, 32, 2965–2970. 84. A. Kowalski, et al., Macromolecules 2000, 33, 7359–7370. 85. W. G. O’Brien, et al., DE 44044838A1, 1993 (to Du Pont de Nemours & Co.). 86. W. G. O’Brien, et al., U.S. Patent 5,521,278, 1994 (to Ecological Chemical Products). 87. P. Kruger, J. Hall, J. J. Kolstad, et al., U.S. Patent 6,005,067, 1998 (to Cargill Inc.). 88. P. Gruber, J. Hall, J. J. Kolstad, et al., WO 9509879A1, 1993 (to Cargill Inc.). 89. H. Tsuji, Macromol. Biosci. 2005, 5, 569–597.
  • 2 CHEMISTRY AND THERMODYNAMIC PROPERTIES OF LACTIC ACID AND LACTIDE AND SOLVENT MISCIBILITY ZHENGYU JIN, YAOQI TIAN, 2.1 2.1.1 AND JINPENG WANG GENERAL PROPERTIES Physical and Chemical Properties of Lactic Acid Lactic acid is a three-carbon organic acid: one terminal carbon atom is part of an acid or a carboxyl group; the other terminal carbon atom is part of a methyl or a hydrocarbon group; and the central carbon atom is part of an alcohol group. Pure anhydrous lactic acid is a white crystalline solid with a low melting point. In contrast to other acids, lactic acid is nonvolatile, odorless, and colorless and has a mild acidic taste. The carbon, hydrogen, and oxygen contents are 40%, 6.71%, and 53.29%, respectively [1]. The two forms of lactic acid, L and D, are isomers with different characteristics. Hydroxyl and carboxyl groups of lactic acid permit a wide variety of chemical reactions. Lactic acid has a pKa of about 3.86 at 25 C. It is a moderate acid that can react with active metals to form gaseous hydrogen and a metal salt, with cyanide salts to generate gaseous hydrogen cyanide, and with diazo compounds, dithiocarbamates, isocyanates, mercaptans, nitrides, sulfites, nitrites, thiosulfates, sulfur dioxide (SO2), carbonates, and bicarbonates to release flammable and/or toxic gases. Lactic acid can form dilactide by esterification with various alcohols and catalytic distillation, form glycol by hydrogenolysis, and produce lactate ester by catalytic dehydration [2–4]. Using such reactions, lactic acid has many large-volume industrial and consumer product applications. Some physical and chemical properties of lactic acid are shown in Table 2.1. 2.1.2 Physical and Chemical Properties of Lactide Lactide with its two asymmetric carbon atoms has three steroisomeric forms: L-lactide in which both asymmetric carbon atoms possess the L (or S) configuration; D-lactide in which both asymmetric carbon atoms possess the D (or R) configuration; and meso-lactide in which one asymmetric atom has the L configuration and the other has the D configuration [7]. Some physical and chemical properties of lactide are shown in Table 2.2. 2.2 THERMODYNAMIC PROPERTIES 2.2.1 Vapor Pressures of Lactic Acids at Different Temperatures For a given mixture, vapor–liquid equilibrium at constant pressure (P) and temperature (T) is generally described by the following equation: wi yi Pi ¼ ci xi Pi sat wi sat ð2:1Þ where ci is the activity coefficient of component i, wi is the fugacity coefficient, and xi and yi are the compositions of the liquid and vapor phases, respectively. Psat is the vapor i pressure at temperature T and wsat is the fugacity coefficient i of pure saturated vapor i at temperature T and pressure Psat . i The vapor pressure for pure components in Equation 2.1 can be obtained using Antoine’s equation (2.2) that Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 19
  • 20 CHEMISTRY AND THERMODYNAMIC PROPERTIES OF LACTIC ACID AND LACTIDE AND SOLVENT MISCIBILITY TABLE 2.1 Physical and Chemical Properties of Lactic Acid Properties Value Molecular weight Melting point ( C) Boiling point ( C) Dissociation constant Ka at 25 C  pKa (25 C) Optical rotation in degrees Isomer Reference 90.08 52.8 53.0 16.8 103 122 (14 mmHg) 1.90 Â 10À4 D, L, DL DL [5] [5] [5] [5] [6] [5] D [6] DL D [6] [5] [5] [6] [6] L [6] 1.38 Â 10À4 3.83 3.79 3.73 À2.5 þ 2.5 TABLE 2.2 D L DL D D L DL 2.2.2 Acid Value Molecular weight Melting point ( C) Boiling point ( C) Optical rotation in degrees Isomer 144.13 144.13 144.13 97 97 125–127 255 À260 þ 260 [8] [8] [8] [8] [8] [8] [8] [8] [8] L meso D L meso L L D correlates empirically the temperatures and vapor pressures of solutions: logðPi sat Þ ¼ AÀ d ¼ ð1:00395À0:000049367TÞ Â d 0 Reference D B CþT Temperature Dependence of Densities of Lactic Density data for aqueous lactic acid solutions have been reported previously (Table 2.3). These density data are in good agreement with a recent report showing that the density increases almost linearly with increasing concentration and decreasing temperature. Densities of reaction mixtures of lactic acid and methanol have also been investigated. Troupe and Kobe [12] used a calibrated thin-walled 100 mL cassia flask to determine the densities of methanol/lactic acid mixtures. An equation based on this calibration was derived for the relationship between true and apparent densities of the reaction mixtures: Physical and Chemical Properties of Lactide Properties ð2:2Þ where T is the absolute temperature, P is the vapor pressure, and A, B, and C are experimental constants. For pure lactic TABLE 2.3 acid, parameters A, B, and C are reported in a previous work as 7.51, 1965.70, and À91.02, respectively, between 273.15 and 500.15 K [9]. To the best of our knowledge, detailed temperature and vapor pressure data for lactide have not been published. ð2:3Þ where d is the true density (g/mL), d0 is the apparent density (g/mL), and T is the temperature ( C). Using the calibration, the true density values were determined, assuming that the volume change during reaction was negligible. Some of the density data for 85% lactic acid/methanol and 44% lactic acid/methanol reaction mixtures are presented in Tables 2.4 and 2.5, respectively. As shown in Tables 2.4 and 2.5, the true densities of the reaction mixtures decrease linearly with increasing temperature. This linear correlation was also confirmed in another study conducted by the same authors [13]. 2.2.3 Acid Temperature Dependence of Viscosity of Lactic As shown in Table 2.6, the viscosity of lactic acid solutions increases with increasing concentration but decreases with increasing temperature. Densities of Lactic Acid/Water Solutions at Different Temperatures [10, 11] Densities of Lactic Acid Solutions (g/mL) Temperature ( C) 20 25 30 40 50 60 70 80 9.16% 24.35% 45.48% 64.89% 75.33% 85.32% 1.01955 1.01811 1.01585 1.01138 1.00674 1.00076 0.99504 0.98899 1.05678 1.05446 1.05183 1.04715 1.04146 1.03513 1.02958 1.02260 1.10980 1.10536 1.10182 1.09427 1.08703 1.07925 1.07219 1.06399 1.15526 1.15181 1.14723 1.13987 1.13205 1.12357 1.11532 1.10762 1.17860 1.17182 1.17013 1.16132 1.15262 1.14250 1.13407 1.12511 1.1989 1.1918 1.1901 1.1813 1.1718 1.1631 1.1536 1.1443
  • MISCIBILITY PROPERTIES OF LACTIC ACID AND LACTIDE TABLE 2.4 Some of the Density Data for 85% Lactic Acid/ Methanol Reaction Mixtures at Different Temperatures L/M Mixtures and Used Weight L/M ¼ 0.125, 88.7963 g L/M ¼ 0.25, 94.7205 g L/M ¼ 0.5, 105.3615 g L/M ¼ 1.0, 109.1140 g 2.2.4 Apparent Density, d0 True Density, d 0.8880 0.8704 0.8530 0.8356 0.8180 0.9472 0.9305 0.9127 0.8946 0.8787 1.0172 1.0002 0.9828 0.9651 0.9588 1.0817 1.0633 1.0472 1.0283 1.0122 Temperature ( C) 0.8906 0.8721 0.8538 0.8356 0.8172 0.9501 0.9323 0.9136 0.8946 0.8778 1.0202 1.0022 0.9838 0.9651 0.9479 1.0849 1.0654 1.0482 1.0283 1.0122 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 References [12, 13] TABLE 2.5 Some of the Density Data for 44% Technical Lactic Acid/Methanol Reaction Mixtures at Different Temperatures L/M Mixtures and Weight Used L/M ¼ 0.125, 93.694 g [12, 13] L/M ¼ 0.25, 99.8330 g [12, 13] L/M ¼ 0.5, 105.3615 g [12, 13] L/M ¼ 1.0, 109.1140 g 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 Apparent Density, d0 True Density, d 0.93694 0.92146 0.90421 0.88683 0.86813 0.99753 0.98260 0.96625 0.94898 0.93087 1.05308 1.03988 1.02292 1.00631 0.98880 1.09168 1.07910 1.07079 1.04715 1.03050 Temperature ( C) 0.93813 0.92182 0.90378 0.88562 0.86620 0.99879 0.98299 0.96579 0.94770 0.92880 1.05440 1.04029 1.02245 1.00495 0.98660 1.09306 1.07950 1.06297 1.04573 1.02820 Reference [14] [14] [14] [14] Thermodynamic Properties Thermodynamic properties of lactic acids and lactide, including entropy of fusion (DS), heat of formation (DHf ), and heat capacity (Cp), have been determined calorimetrically. Table 2.7 summarizes some of the thermodynamic data at 25 C. 2.3 MISCIBILITY PROPERTIES OF LACTIC ACID AND LACTIDE 2.3.1 21 The physical equilibrium distributions of lactic acid in several organic phases are tabulated in Table 2.8. 2.3.2 Miscibility of Lactic Acid with Modifiers in Diluents Diluents are often used to modify the physical properties of solvents (e.g., viscosity, specific gravity, and surface tension). The nature of the diluent and the modifier may Miscibility of Lactic Acid with Different Solvents The distribution coefficient (KD) has been investigated as a function of various process variables. It is defined as a ratio of the total lactic acid concentration in the organic and the aqueous phases (Equation 2.4). Lactic acid can exist as two forms in aqueous solution, namely, the undissociated and dissociated forms. Lactic acid exists as an undissociated form (HLA) in the aqueous solution when pH < pKa (pKa ¼ 3.86, 25 C) [2]. But in the organic phase, lactic acid is always present in the undissociated form. KD ¼ ½HLAŠo ½HLAŠw ð2:4Þ TABLE 2.6 Acids [11] Viscosity of Aqueous Solution of Various Lactic Viscosity (10À3 Pa s) Temperature ( C) 9.16% 24.35% 45.48% 64.89% 75.33% 25 30 40 50 60 70 80 1.15 1.03 0.809 0.671 0.572 0.473 0.416 1.67 1.46 1.13 0.918 0.746 0.632 0.532 3.09 2.74 2.03 1.59 1.26 1.02 0.843 6.96 6.01 4.22 3.12 2.38 1.85 1.47 13.03 10.55 7.08 4.98 3.57 2.73 2.08
  • 22 CHEMISTRY AND THERMODYNAMIC PROPERTIES OF LACTIC ACID AND LACTIDE AND SOLVENT MISCIBILITY TABLE 2.7 Some of the Thermodynamic Parameters of Lactic Acid at 25 C Crystalline L-( þ )Lactic Acid Racemic Lactic Acid Reference À63 þ 1868 þ 4030 þ 12.2 À321,220 À165,890 þ 0.338 À137.2 À124,980 Thermodynamic Parameters À63 – þ 2710 þ 9.4 À325,600 À163,000 þ 0.559 À125.3 À126,500 [15] [15] [15] [15] [15] [15] [15] [15] [15] Heat of dissociation (DH, cal/mol) Heat of solution (DH, cal/mol) Heat of fusion (DH, cal/mol) Entropy of fusion (DS, cal/(mol  C))  Heat of combustion (DHc , cal/mol)  Heat of formation (DHf , cal/mol) Heat capacity (Cp, cal/(g  C)) Entropy of formation (DS , cal/(mol  C)) f Free energy of formation (DFf , cal/mol) TABLE 2.8 Distribution Coefficients of Lactic Acid in Different Solvents Solvents Isobutanol Cyclohexanol Isoamylol Pentanol Hexanol 2-Ethylhexanol Phenylcarbinol Octanol 1-Decanol Ethyl ether Isopropyl ether Isophorone Methyl isobutyl ketone Diisobutyl ketone Cyclohexanone Nitroethane Ethoxypropane Xylene Ether Ethyl carbonate Chloroform Chloroform Furan Ethyl acetate MIBK Butyl acetate Temperature ( C) 31 25 25 31 28 28 25 25 25 28 28 28 25 25 25 25 25 25 25 25 25 28 25 25 25 25 KD 1.37 1.74 2.22 2.45 3.40 6.8 2.33 0.29 0.31 9.8 35 2.33 7.9 36 1.87 11.0 35 1.97 9.8 21.0 106 100 3.32 4.07 0.31 0.4 Reference [16] [11] [11] [16] [16] [16] [11] [17] [17] [16] [16] [16] [11] [11] [11] [11] [11] [18] [11] [11] [11] [16] [11] [11] [17] [19] significantly affect the miscibility of lactic acid in the solvent. Table 2.9 shows the distribution coefficients of lactic acid in different diluents with modifiers. Overall, the distribution coefficient of lactic acid increases with increasing concentration of modifier in the organic phase. 2.3.3 Physical and Chemical Equilibrium of Lactic Acid In the case of undissociated lactic acid molecules, the equilibrium can be represented by a system of equations TABLE 2.9 Distribution Coefficient of Lactic Acid in Diluent with Modifiers KD Solvents 70% (v/v) TBP, 30% (v/v) dodecane 30% (v/v) Aliquat 336, 70% (v/v) 1-decanol 30% (v/v) TOA, 70% (v/v) 1decanol 15% (v/v) Aliquat 336, 15% (v/v) TOA, 70% (v/v) 1decanol 15% (v/v) Aliquat 336, 15% (v/v) 1-decanol, 70% (v/v) dodecane 30% (v/v) Aliquat 336, 40% (v/v) TBP, 30% (v/v) dodecane 15% (v/v) TOA, 15% (v/v) TBP, 70% (v/v) dodecane 15% (v/v) TOA, 15% (v/v) 1decanol, 70% (v/v) dodecane 10% Alamine 336, octanol 20% Alamine 336, octanol 30% Alamine 336, octanol 20% Alamine 336, MIBK 30% Alamine 336, MIBK 40% Alamine 336, MIBK 20% Alamine 336, decanol 30% Alamine 336, decanol 40% Alamine 336, decanol Reference 40.29 [20] 39.40 [20] 94.78 [20] 97.11 [20] 14.72 [20] 48.71 [20] 19.19 [20] 53.42 [20] 15.35 19.69 25.96 0.72 2.68 4.24 12.57 16.44 23.37 [21] [21] [21] [21] [21] [21] [21] [21] [21] due to the formation of different complexes. The reaction of lactic acid with solute (B) produces a reaction complex (BHLA) that remains largely in the organic phase and may be represented by HLAw þ Bo $ BHLAo ð2:5Þ
  • 23 MISCIBILITY PROPERTIES OF LACTIC ACID AND LACTIDE A quantitative interpretation of the equilibrium for the HLA-B extraction can be defined by an equilibrium complexation constant, KE: KE ¼ ½BHLAŠo ½BHLAŠw ½BŠo ð2:6Þ The concentration of B in the organic phase would be ½BŠo ¼ ½BŠi;o À½BHLAŠo KD ¼ A plot of log KD versus log[B] will give a straight line with a slope of unity. However, if p:q HLA-B complex is formed, then K11 HLAw þ Bo ! BHLAo ð2:9Þ ½BŠq CHLApÀ1 o ¼ pKE ½BŠq o ð1 þ Ka =½H þ ŠÞ ð1 þ Ka =½H þ ŠÞp   CHLA q CHLApÀ1 ¼ pKE ½BŠo ÀKD q p ð1 þ Ka =½H þ ŠÞp ð2:17Þ Miscibility of lactic acid depends on the concentration of lactic acid in the extraction regime. The value of the physical mass transfer coefficient kL is required for confirming the regime of extraction. This coefficient can be obtained by conducting physical extraction of lactic acid from water. For a batch process, a differential mass balance yields the following equation: Vw K21 HLAw þ BHLAo ! BHLA2;o K12 HLAw þ 2Bo ! B2 HLAo ð2:10Þ ð2:11Þ . . . Kpq pHLAw þ qBo ! Bq ðHLAÞp;o ð2:12Þ Thus, the equilibrium complexation constant for the reaction represented by the above equations is KEpq ¼ ½Bq ðHLAÞp Šo ð2:13Þ ½BŠd ½HLAŠp o w The concentration of undissociated molecules in the aqueous phase [HLA] can be calculated from the total concentration CHLA and the dissociation constant (pKa ¼ 3.86, 25 C): HLAw ( H þ þ LAÀ ) CHLA ¼ ½HLAŠ þ ½LAÀ Š Ka ¼ ½H þ Š½LAÀ Š ½HLAŠ ½HLAŠ ¼ ð2:14Þ CHLA ð1 þ Ka =½H þ ŠÞ ð2:15Þ ð2:16Þ KD ¼ KE If 1:1 HLA-B complex is formed, then ½HLAŠo ¼ ½BHLAŠo : ð2:8Þ ½HLAŠw Substituting KE and CHLA from Equations 2.14–2.16, ð2:7Þ log KD ¼ log KE þ log½BŠo ½Bq ðHLAÞp Šo d½HLAŠo ¼ kL Ac ð½HLAŠEo À½HLAŠIno Þ dt ð2:18Þ Integration of this equation yields Vw kL ¼ Ac t ð ½HLAŠo 0 d½HLAŠo ð½HLAŠEo À½HLAŠIno Þ ð2:19Þ To avoid problems due to the reversible reaction, only initial rates are considered for evaluating the reaction kinetics. The equilibrium complexation constants of lactic acid in several solvents with different diluents are shown in Table 2.10. In the case of reactive extraction, the equilibrium concentration of lactic acid in the organic phase is much higher compared to physical extraction with pure solvents. With increasing equilibrium concentration, the polarity increases, which improves the lactic acid extraction. At high equilibrium concentrations, the limited solubility of the complex and the high viscosity of the organic phase are responsible for the limitation of the distribution. The miscibility of lactic acid and lactide with supercritical CO2 was studied by Gregorowicz [24]. The solubility of the monomer in a supercritical solvent depends on temperature, pressure, and overall composition of the reacting mixture equilibrium between different phases. It is possible that the concentration of the monomer in the supercritical fluid phase may vary over a wide range. In general, the solute solubility in compressed carbon dioxide increases as pressure increases from 50 to 200 bar (Figure 2.1). At pressures below 130 bar, the solubility decreases as temperature increases. Thus, it is
  • 24 3 3 – 3 3 3 – 30 C – – – 25 C 30 C 30 C 25 C TOA TOA Alamine 336 Alamine 336 TMA IL Regime 20 C Temperature – 1 1 1 – – – Order – 16.67 sÀ1 1.38 sÀ1 – 5.89 Â 10À9 m4/ (mol s) – Rate Constant 20% Alamine 336 in octanol IL-104 Octanol MIBK 20% Decanol and 50% dodecane Xylene Xylene Organic Phase Diluents – – – 0.8–1.2 m/s 2 Â 10À6 N0.4 – – 3.5 Â 10À5 N4.7 – – – – 6.54 Â 10À5 m/s Mass Transfer Coefficient, kL The Equilibrium Complexation Constants of Lactic Acid with Several Solutes TOA Solute TABLE 2.10 1) 2) 1) 1) 2) (m3/mol)2 1) (m3/mol)2 1) 1) 1) 1) (1, 1) (2, 1) (3, 1) (1, (2, (1, (2, (1, (2, – (1, 2) (3, 1) (1, 1) m3/mol (1, (1, (3, (1, Species (p, q) 40.9 dm3/mol 123 (dm3/mol)2 38.4 (dm3/mol)3 [23] [17] [21] [21] [22] 3.18 Â 10À8 (m3/mol)2 5.48 Â 10À10 (m3/mol)3 2.4671 dm3/mol 3.1851 (dm3/mol)2 1.2811 dm3/mol 10.15 dm3/mol 10.13 (dm3/mol)2 73.55 dm3/mol 10.95 (dm3/mol)2 – [18] [18] Reference 1.31 dm3/mol 8.08 Â 10À2 (dm3/mol)2 0.95 (dm3/mol)3 3.41 Â 10À4 m3/mol Equilibrium Complexation Constant, KEpq
  • REFERENCES log(y × 104) 10 1 , 313 Κ; 50 FIGURE 2.1 80 , 318 Κ; 110 140 P (bar) , 328 Κ 170 200 Solubility of lactic acid in carbon dioxide. possible to control the concentration of the monomer by appropriate adjustment of pressure and temperature. 2.3.4 Miscibility of Lactide with Solvents Lactide is miscible with benzene, toluene, xylene, methylene chloride, chloroform, tetrahydrofuran, ethyl acetate, methanol, isopropanol, acetone, and butanone. The solubility increases with increasing temperature [25]. Lactide will hydrolyze to lactic acid in water at room temperature, and the rate of hydrolysis of meso-lactide is much higher than that of D,L-lactide [26]. REFERENCES 1. N. Thongwai, Doctoral thesis, Louisiana State University, 1999, pp. 4–5. 2. R. Datta, M. Henry, J. Chem. Technol. Biotechnol. 2006, 8, 1119–1129. 3. A. S€dergard, M. Stolt, Prog. Polym. Sci. 2002, 27, 123–1163. o 4. N. Narayanan, Electron. J. Biotechnol. 2004, 7, 167–179. 5. H. I. Aljundi, Doctoral thesis, Michigan State University, 2000, pp. 6–10. 25 6. M. Dworkin, S. Falkow, E. Rosenberg, K. H. Schleifer, E. Stackebrandt, Organic Acid and Solvent Production, Springer, New York, 2006, pp. 511–755. 7. H. P. Benecke, R. A. Markle, R. G. Sinclair,U.S. Patent 5,332,839, 1994 (to BioPak Technology, Ltd.). 8. J. J. Zhu, Y. L. Wang, X. J. Quan, J. Xin, R. Wang, Chem. World 2005, 3, 173–176. 9. PRO/II Library, Simulation Sciences, Stockport, Cheshire, UK, 1999. 10. R. A. Troupe, W. L. Aspy, P. R. Schrodt, Ind. Eng. Chem. 1951, 43, 1143–1146. 11. D. Xu, M. Sha, Y. Li, Organic Acid Fermentation Technology, Chemistry Industry Press, Beijing, 1991, 146 pp. 12. R. A. Troupe, K. A. Kobe, Ind. Eng. Chem. 1950, 42, 801–810. 13. R. A. Troupe, E. Y. Dimilla, Chem. Process 1957, 49, 847–855. 14. R. A. Troupe, K. A. Kobe, Ind. Eng. Chem. 1950, 42, 1403–1409. 15. C. H. Holten, A. Mueller, D. Rehbinder, Lactic Acid: Properties and Chemistry of Lactic Acid and Derivatives, Verlag Chemie, Copenhagen, 1971. 16. Q. R. Jin, J. M. Zhang, Q. Xu, Organic Acid Fermentation Process Engineering, China Light Industry Press, Beijing, 1989, pp. 341–342. 17. L. K. Wasewar, A. B. M. Heesink, G. F. Versteeg, V. G. Pangarkar, Chem. Eng. Sci. 2004, 59, 2315–2320. 18. R. S. Juang, R. H. Huang, Chem. Eng. J. Biochem. Eng. J. 1997, 65, 47–53. 19. P. V. Frieling, K. Sch€gerl, Process Biochem. 1999, 34, u 685–696. 20. M. Marinova, G. Kyuchoukov, J. Albet, J. Molinier, G. Malmary, Sep. Purif. Technol. 2004, 37, 199–207. 21. L. K. Wasewar, A. B. M. Heesink, G. F. Versteeg, V. G. Pangarkar, J. Biotechnol. 2002, 97, 59–68. 22. D. Yankov, G. Kyuchoukov, J. Molinier, J. Albet, G. Malmary, Biochem. Eng. J. 2004, 21, 63–71. 23. J. Martk, S. Schlosser, Sep. Purif. Technol. 2007, 57, 483–494. a ˇ 24. J. Gregorowicz, J. Supercrit. Fluid. 2008, 46, 105–111. 25. S. X. Shi, Y. Z. Xia, Z. P. Guo, S. K. Jiao, X. Y. Li, J. Beijing Univ. Chem. Technol. 2003, 30, 32–34. 26. L. X. Fang, Q. S. Jing, Chem. Eng. 2008, 148, 4–6.
  • 3 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID)  ANDERS S€DERGARD AND MIKAEL STOLT o 3.1 INTRODUCTION Polymers based on lactic acid (PLA) are a most promising category of polymers made from renewable resources. They are not only compostable and biocompatible, but also processable with most standard processing equipment. The properties of lactic acid based polymers vary to a large extent depending on the ratio between, and the distribution of, the two stereoisomers or other comonomers [1–3]. The polymers can be manufactured by different polymerization routes, which are schematically described below (Figure 3.1). PLA of high molecular weight is most commonly made by ring-opening polymerization (ROP) of the ring-formed dimer, dilactide (lactide; 3,6-dimethyl-1,4-dioxane-2,5dione), which is made by depolymerization of the polycondensed lactic acid (LA; 2-hydroxypropanoic acid). This route is a two-step reaction that usually involves additional purification steps and is therefore related to significant costs. It is often stated in the art that the preparation of a high molecular weight PLA by a direct dehydration condensation reaction is not feasible due to the equilibrium not favoring a high molecular weight polymer. PLA prepared from polycondensation has low molecular weight and poor mechanical properties and therefore is not suitable for many applications. The commercial interest for solving this problem has increased because of the need of cost-effective approaches in the manufacturing of lactic acid based polymers with a high molecular weight. Solvent-assisted polycondensation is one way to overcome this problem [4] and melt polycondensation followed by solid-state polycondensation is another one [5]. The third approach to achieve high molecular weight LA based polymers is to utilize the terminal groups of the prepolymer in linking processes where a linking agent is employed [6]. Such prepolymers can be composed of solely one stereoisomer, combinations of D- and L-lactoyl units in various ratios, or lactic acid in combination with other hydroxy acids or di- or multifunctional comonomers. If the lactic acid is polycondensated in the presence of difunctional monomers (e.g., diols or diacids), the resulting prepolymer will have the same end groups in both chain ends; that is, the prepolymer is a telechelic macromer [6]. Since the first commercial products of lactic acid based polymers were introduced in the market, several attempts have been made to manufacture lactic acid based polymers or convert the polymer into other products [7]. The main use was initially in medical applications, and this product field is still an important one with a number of companies still being active [8–10]. The research on PLA and its copolymers intended for medical applications has accelerated since then. Over the past two decades, there has also been an increase in large-scale industrial production for commodity use, for example, in the packaging and paper industries [11, 12]. During the last few years, several companies have become active in manufacturing of large volumes of polylactide. Since 2005, the business of bio-based material production has grown significantly. This growth is driven by three main factors: (i) oil prices—a high oil price is reflected in the pricing of traditional polymers and in energy costs; (ii) legislation—favorable legislation and government support make the use of bio-based materials an attractive option to traditional polymers; and (iii) image—an increased awareness among brand owners and retailers makes compostable and renewable products real alternatives with added value in terms of sustainability. PLA has a unique position in the field Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 27
  • 28 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) O O COOH HO poly(lactic acid) O O Cat., Δ –H2O HO COOH Structural comonomers Difunctional comonomers Telechelic poly(lactic acid) HO Cat., Δ initiator Cat., Δ , –H2O Lactic acid Cat., Δ –H2O ROP Chain extension OH FIGURE 3.1 Manufacturing routes for lactic acid based polymers. of bio-based products that is related to its processability, appearance, and properties such as mechanical strength and barrier. Several commercial manufacturing operations have been attempted over the years. Some of these operations have been terminated or reorganized [3], while quite a number of commercial operations have been notified more recently [13–18]. 3.2 LACTIC ACID BASED POLYMERS BY POLYCONDENSATION From a chemistry point of view, lactic acid can form PLA by means of the reaction of the hydroxyl and carboxylic acid groups of lactic acid. By removing the water formed during this condensation reaction, the reaction proceeds toward the product side, PLA: HO-ðCHðCH3 ÞCOOÞn -H þ HO-ðCHðCH3 ÞCOOÞm -H ! HO-ðCHðCH3 ÞCOOÞn þ m -H þ H2 O where n and m ! 1. The removal of water becomes more difficult and can be rate determining when producing a higher molecular weight PLA due to the increased viscosity of the reaction mixture. By means of applying vacuum, the water removal can be enhanced. However, during the polycondensation of lactic acid, other side reactions also occur, such as transesterification, resulting in the formation of ring structures of different sizes [19]. These side reactions have a negative influence on the properties of the polymer. The formation of ring struc- tures, such as lactide, lowers the overall molecular weight and the removal of the lactide formed reduces the first pass yield of the process. The formation of lactide cannot be excluded, but to suppress the lactide formation and increase the first pass yield of the polycondensation reaction of lactic acid, the lactide can be returned back to the reaction mixture. A partial condenser (reflux condenser) or a rectification column placed on top of the polycondensation reaction can be used to recycle lactide back to the reaction mixture. The addition of stabilizers, such as antioxidants or phosphorous compounds, can reduce the color formation [20]. The lactide formation becomes substantial at high reaction temperatures (>200 C) [21]. To suppress the lactide formation, the polycondensation reaction should thus be carried out at temperatures below 200 C. Conducting the polycondensation at low temperatures again has a negative effect on the removal of water due to the relatively high viscosity of the reaction mixture in addition to a lowered reaction rate. Since polycondensation should be the main reaction, the removal of water should be as high as possible without allowing the reaction mixture to undergo transesterification reactions. Besides the removal of reaction water from the viscous reaction mixture, the quality of the monomer (lactic acid) with respect to chain terminators such as monocarboxylic acids (formic acid, acetic acid, propionic acid, etc.) or monohydroxy alcohols (methanol, ethanol, propanol, etc.) is important. To obtain a desired high molecular weight, the amount of end terminators must be limited. In Figure 3.2, the theoretical relationship between the concentration of chain terminator in mol% and the corresponding number-average molecular weight (Mn) at different conversions of the func-
  • LACTIC ACID BASED POLYMERS BY POLYCONDENSATION 29 FIGURE 3.2 Theoretical relation between concentration of chain terminator and molecular weight (Mn) at different conversions of the functional groups. tional groups is shown. With a chain terminator content of 0.1 mol%, a maximum Mn of 72.000 g/mol can be obtained at 100% conversion of the functional groups. This clearly demonstrates the importance of using high-purity lactic acid (often also called polymer-grade lactic acid) during the polycondensation reaction. 3.2.1 Direct Condensation The preparation of PLA from lactic acid by direct condensation can be divided into three principal stages: (a) removal of the free water content, (b) oligomer polycondensation, and (c) melt polycondensation of high molecular weight PLA: (a) Besides lactic acid, the feedstock also contains the socalled free water. Due to the equilibrium of lactic acid and water, some low amount of oligomers of lactic acid (linear dimer, linear trimer, etc.) can already be formed in this stage. To convert lactic acid to PLA, first the free water has to be removed. The evaporation of the free water requires a system having good heat transfer and can be carried out in commonly known evaporators, such as falling film evaporators. Flash evaporation can also be used to remove the free water in lactic acid feedstock. (b) In the second stage, the lactic acid is converted into low molecular weight PLA or oligo(lactic acid). In this step, the removal of water is not critical because of the low viscosity of the reaction mixture. The ratedetermining step in this stage is usually the chemical reaction, which is significantly affected by the catalyst used [22]. Traditional polycondensation catalysts are strong acids, and organometallic compounds are also commonly used catalysts. The low molecular weight PLA polycondensation can also be carried out in an evaporator or alternatively in a stirred reactor having an agitator that generates good radial and axial mixing. The loss of lactic acid due to entrainment can be overcome by using a reflux condenser, a demister package, or a rectification column. Preferably, this stage is carried out in a system having a narrow residence time distribution (plug-flow behavior) to obtain a prepolymer of lactic acid of narrow molecular weight distribution (small dispersion). (c) The third stage is the melt polycondensation in which the removal of water becomes critical. To enhance the polycondensation reaction, and not the transesterification reactions, the water formed in the reaction mixture should be removed efficiently. The rate-determining step in this phase is the mass transfer of water. To enhance both mass and heat transfer, the melt polycondensation reaction should be applied in an apparatus having an efficient renewal of phase boundary layers. The apparatus should have intensive mixing and kneading in order to homogenize the reaction mixture. The removal of water from the viscous PLA mass can be further enhanced by carrying out the reaction under vacuum conditions in an inert atmosphere. A mathematical model for the polycondensation of lactic acid accounting for water removal by diffusion has been developed [23]. The increasing molecular weight of the PLA requires a system that can handle high-viscosity mass. Such an apparatus could be a rotating disk type of reactor, generating a good surface renewal to enhance the mass transfer of the water formed. Such an apparatus
  • 30 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) should also have very good heat transfer to have a homogeneous temperature profile in the reaction mixture. Especially the mechanical heat formed due to mixing and kneading of the highly viscous PLA should be controlled. In this stage also a plug-flow behavior is preferred to obtain a narrow molecular weight distribution. Only a few studies have dealt with the influence of the catalyst when preparing PLA of high molecular weight through the direct bulk condensation reaction. In most studies with regard to catalysts, the polycondensations are carried out only to obtain low molecular weight polymers with an Mw of a few thousands, before they are stopped. PLA having a molecular weight of as high as 130.000 g/mol (gel permeation chromatography (GPC) relative to PS standards) was synthesized by direct bulk condensation polymerization at 180 C using titanium(IV) butoxide as catalyst [24]. In another study, several metal catalysts based on Ge, Sb, Zn, Fe, Al, Ti, and Sn were employed in the melt polycondensation reaction [25]. The most efficient catalyst was found to be SnO with regard to molecular weight of the PLA, but the yield was below 40% when using this catalyst at 180 C (20 h). However, when using p-toluenesulfonic acid as a co-catalyst with SnCl2, the efficiency was drastically improved and molecular weights above 100.000 g/mol (GPC relative to PS standards in chloroform, 35 C) were achieved within 15 h of polycondensation. Sodium carbonate, calcium carbonate, and lanthanum oxide have also been used as catalysts when preparing PLA of high molecular weight [26]. Weight-average molecular weights ranging from 63.000 to 79.000 g/mol (GPC relative to PS standards in chloroform at 40 C) were obtained by melt polycondensation but in a poor yield (33–52%). To achieve an increased molecular weight of the PLA, comonomers with functionality higher than two have been used. A process for making a star-shaped PLAwas described, where the lactic acid is polycondensated in the presence of a polyhydroxyl compound having at least four hydroxyl groups [27]. The PLA obtained has a higher molecular weight than a polymer prepared without the use of comonomer, but the invention possesses a clear limit in obtainable molecular weight. If the polyhydroxyl compound is used in large amounts, the polymer will be hydroxyl terminated and the condensation reaction cannot continue, thus yielding a low molecular weight polymer. On the other hand, if the polyhydroxyl compound is used in small amounts, the effect of the polyhydroxyl compound will diminish and the polycondensation reaction will be a blend of star-shaped PLA and linear PLA. A hyperbranched PLA of high molecular weight was also manufactured by coupling a first prepolymer having at least three functional end groups with a second prepolymer having at least two functional end groups by a condensation reaction between the end groups in the prepolymers [28]. The improvement of the process was that the number of arms and/ or molecular weight of the functionalized prepolymers can be accurately adjusted, thus affecting the properties of the resulting hyperbranched polymer in a desired way. Molecular weight in excess of 200.000 g/mol (GPC relative to PS standards in chloroform at 30 C) was obtained for the hyperbranched PLA. Lactide has been used as a coreactant and yield enhancer in the polycondensation reaction of lactic acid [29]. Mw’s in the range of 65.000–83.000 g/mol were obtained in 17–42 h (GPC, 40 C, chloroform), starting from 90 wt% lactic acid, when an inorganic solid acid catalyst (aluminum silicate) was used. Copolymers with high enough molecular weight for practical use were prepared from succinic acid and 1,4butanediol and minor amounts of lactic acid [30]. An increase in reaction rate was claimed when the aliphatic diol and the aliphatic dicarboxylic acid were polycondensated using a few mole percent lactic acid and a germanium oxide catalyst. 3.2.2 Solid-State Polycondensation The disadvantage of the PLA prepared by the direct polycondensation is often a limited molecular weight in combination with a low yield. Some progress in increasing the molecular weight of the PLA has recently been achieved, though, by sequential melt/solid polycondensation [5, 31]. In the sequential melt/solid-state polycondensation, the three first stages as described for direct polycondensation (i. e., removal of the free water content, oligomer polycondensation, and melt polycondensation) are utilized with an additional fourth stage. In the fourth stage, the melt-polycondensated PLA is cooled below its melting temperature, often followed by particle formation as it solidifies. The solid particles are then subjected to a crystallization process, where two phases can be identified: a crystalline phase and an amorphous phase. It is believed that the reactive end groups, as well as the catalyst, are concentrated in the amorphous phase in between the crystals (Figure 3.3), thus yielding an apparent enhancement of the polycondensation rate although the polycondensation is performed in the solid state at a low temperature (i.e., below the melting temperature of the polymer). A metal catalyst can catalyze the solidstate polycondensation in the amorphous phase as well as the melt polycondensation. These catalysts can be different metals or metal salts, from metals such as Sn, Ti, and Zn. The rate-determining step in the solid-state polycondensation is the mass transport of the reaction water by molecular diffusion. The removal of water can be further enhanced by carrying out the reaction under vacuum conditions in an inert atmosphere.
  • LACTIC ACID BASED POLYMERS BY POLYCONDENSATION 31 obtained for the stereoblock PLA (GPC relative to PMMA standards with hexafluoroisopropanol (HFIP) as the eluent). In another study, it was found that the weight-average molecular weight of the resultant stereoblock PLA was strongly influenced by the lactide/oligomer content in the melt blend, which is determined by the melt-blending conditions because it is directly correlated with the crystallinity of the polycondensation products [35]. The effect of crystallization on the solid-state polycondensation of PLLA has also been investigated [36]. The results showed that the Mw of the PLA reached a maximum value when a crystallization time of 30 min (105 C) and solid-state polycondensation of 35 h (135 C) were used. FIGURE 3.3 Schematic polycondensation. description of the solid-state A process for preparing PLA by the sequential melt/solidstate polycondensation has been described [32]. The process comprises a liquid-phase polycondensation reaction step, subsequently followed by a solidification and particle formation step of the prepolymer formed, by crystallization of the prepolymer particles, and finally a solid-phase polymerization step. The weight-average molecular weight of linear PLA obtained by this process was above 100.000 g/mol that in many cases was a 10-fold increase when compared to the prepolymer. The total process time to prepare the mentioned PLA was about 100 h, starting from 88% lactic acid. The weight-average molecular weight was determined by GPC in 40 C in chloroform in comparison to polystyrene standards. A similar process for making poly(hydroxycarboxylic acid) is also described where a low molecular weight polycondensate is pelletized and crystallized, and a solid-phase polycondensation reaction step is performed by heating the pellets to a temperature not lower than the crystallization temperature [33]. According to the invention, pellets of poly (hydroxycarboxylic acid) of low molecular weight cause no blocking in the equipment, and it is possible to prepare poly (hydroxycarboxylic acid) of high molecular weight efficiently. The weight-average molecular weight obtained by this process is in the range of 128.000–152.000 g/mol (GPC, 40 C, chloroform) requiring a minimum solid-phase polycondensation reaction time of 40 h. Stereoblock PLA was synthesized by solid-state polycondensation of a 1:1 mixture of PLLA and PDLA [34]. In the first step, PLLA and PDLA having a medium molecular weight were melt polycondensated. The PLLA and PDLA were then melt blended in a 1:1 weight ratio to allow formation of their stereocomplex, and the blend was subjected to solid-state polycondensation. Some process optimization with regard to polymerization conditions was done and molecular weights exceeding 100.000 g/mol were 3.2.3 Azeotropic Dehydration In azeotropic dehydration, the same principal stages as in direct melt condensation of lactic acid are present, with the exception that the last high viscosity melt-polycondensation stage is eliminated because the polycondensation is performed in solution. The removal of the reaction water from the reaction medium thus becomes easier and a higher molecular weight of the PLA is achievable. The solvent, on the other hand, has to be dried from the water produced in the reaction using a drying agent (e.g., molecular sieve). Alternatively fresh, dry organic solvent can be added during the reaction, which is undesired from both an environmental and an economical point of view. Another disadvantage when using organic solvents in the dehydration reaction is that the prepared polymer has to be collected from the solvent, typically by using a nonsolvent for the polymer, and dried. These steps cause excess labor, are time-consuming, and usually lower the yield of the raw material used. The boiling point of the solvent also sets a restriction on the polycondensation temperature that can be used. However, the optical purity of the PLA can be retained because of the lower temperature used. Several patent applications have been filed on the azeotropic dehydration of PLA. A process was claimed wherein the organic solvent is removed from the reaction mixture and an additional solvent, that has a water content less than the water content of the solvent removed from the reaction mixture, is added to the reaction mixture [37]. The removed solvent was dried using, for example, molecular sieves, phosphorus pentaoxide, or metal hydrides and added back to the reaction mixture. In another similar application, the drying agent used was an ion exchange resin [38]. Examples of solvents that were claimed included anisole or diphenyl ether. Azeotropic dehydration of lactic acid containing impurities (e.g., chain terminators such as methanol, ethanol, acetic acid, and pyruvic acid) in a total amount of 0.3 mol% has also been reported [39]. When the lactic acid contained 0.16 mol% methanol, a molecular weight of 50.000 g/mol
  • 32 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) (viscometry, dichloromethane, 20 C) was obtained in diphenyl ether at 130 C using tin powder as a catalyst. A methanol content of 0.02 mol% yielded a PLA with Mw of 320.000 g/ mol using the same polycondensation procedure. The effect of several different catalysts on the azeotropic dehydration of lactic acid in diphenyl ether has been studied [40]. The most effective catalysts were found to be Sn compounds (Sn powder, SnO, and SnCl2), Ni(OAc)2, and CH3-Ph-SO3H. When using these catalysts, weight-average molecular weights exceeding 100.000 g/mol according to GPC results relative to polystyrene standards (chloroform, 40 C) were obtained for the PLA. Haloiminium salts have also been utilized as polycondensation agents in azeotropic dehydration of hydroxycarboxylic acids, including lactic acid [41]. A process to further increase the Mw of the hydroxycarboxylic acid copolymerization with polyfunctional compounds was described [42]. The polyfunctional compounds were those having three or more carboxylic end groups or hydroxyl groups. In addition to this, a second compound having two or more functional end groups was present in the reaction mix. A disadvantage of the invention is that all compounds are preferably added at the same time in the beginning of the reaction, thus giving an uncontrollable reaction and therefore also reproducibility problems. 3.3 LACTIC ACID BASED POLYMERS BY CHAIN EXTENSION 3.3.1 Chain Extension with Diisocyanates Chain extension with diisocyanates is undoubtedly the most commonly applied chain-extension approach for aliphatic polyesters (Figure 3.4). A large number of different diisocyanates have been used as linking molecules in the preparation of aliphatic poly(ester-urethane)s. Various isocyanates are listed in Table 3.1 together with the type of prepolymer used. The most frequently used diisocyanate in the preparation of aliphatic poly(ester-urethane)s is 1,6-hexamethylene diisocyanate because of its low toxicity, and the use of this isocyanate will accordingly be discussed in more detail. Numerous examples of the use of diisocyanate chain extension of aliphatic polyesters can be found in the scientific publications and in the patent literature [52–54]. FIGURE 3.4 Chain-extension reactions of lactic acid based prepolymers using diisocyanates. TABLE 3.1 Diisocyanates Used in Preparation of Aliphatic Poly(ester-urethane)s Name 1,6-Hexamethylene diisocyanate 1,4-Butanediisocyanate 1,6-Hexamethylene diisocyanate 1,4-Butanediisocyanate Methylenediphenyl diisocyanate 4,40 -Dicyclohexylmethane diisocyanate Isophorone diisocyanate Ethyl 2, 6-diisocyanohexanoate 1,6-Hexamethylene diisocyanate 1,6-Hexamethylene diisocyanate Methylene diphenyl isocyanate Prepolymer Composition L-LA, Reference butanediol [43] e-CL, butanediol 1,3-propanediol, succinic acid PLA, butanediol LA [44] [45] L-LA, [48] butanediol [46] [47] L-LA, butanediol e-CL, glycolide, inositol L-LA, mandelic acid [48] [49] L-LA, malic acid [50] butyl glycidyl ether [51] L-LA, [50] 3.3.1.1 Chain-Extension Reaction Parameters The amount of diisocyanate determines to a large extent how the linking reaction proceeds. For a hydroxyl-terminated prepolymer and an equimolar amount of diisocyanate, the chainextension reaction proceeds rapidly and the molecular weight reaches its maximum in a few minutes, after which the molecular weight starts to decrease due to thermal degradation [55]. Side reactions become more evident if the amount of diisocyanate is increased, which can be seen from an increase in weight-average molecular weight but not in the number-average molecular weight. This is caused by the formation of isocyanate-terminated prepolymers in excess of chain extender and the further reaction with urethane bonds yielding allophanates, branching and eventually cross-linking. The terminal groups of the prepolymers are also important for the progress of the chain extension. Carboxylic acid end groups can react with isocyanates and yield amides, carboxylic anhydrides, or ureas even if the reactivity with hydroxyl groups is significantly higher than with carboxylic acid groups. The presence of carboxylic acid end groups has furthermore been suggested to retard the reaction between hydroxyls and isocyanates [56]. The acid number of the prepolymers is consequently an important parameter and useful as a tool for controlling the chain-extension reaction [55, 57]. Molecular weight, molecular weight distribution, and long-chain branching can be varied by changing reaction conditions and acid number of the prepolymer. This gives a possibility for tailoring the viscoelastic properties of the polymer and tailoring lactic acid based polymers for
  • LACTIC ACID BASED POLYMERS BY CHAIN EXTENSION specific processing equipment and conditions [57]. The third important parameter in the linking process is the catalyst, which will affect not only the reaction rate, but also the racemization. The racemization, that is, degree of lost tacticity, determines the ability of the poly(ester-urethane) to crystallize, which in turn will affect the mechanical and thermomechanical properties of the polymer. It can be concluded that the preparation of poly(ester-urethane)s can be best controlled by the following: (a) the presence of more than one type of end groups in the prepolymer being as low as possible, (b) the semicrystalline nature being retained during both the prepolymer preparation and the linking step, (c) keeping the molecular weight of the prepolymer below a certain level to perform successful linking, and (d) using a catalyst that preferably is also nontoxic, with retained activity in both reaction steps [58]. 3.3.1.2 Properties of Poly(ester-urethane)s The thermal and mechanical properties of poly(ester-urethane)s are similar to those of polylactide prepared by ring-opening polymerization, but most of the poly(ester-urethane)s described in the literature are amorphous, with a few exceptions [58]. This means that some of the properties need to be improved to make useful end products. For many applications, the brittleness is an issue and for others the low heat resistance. Different approaches have been suggested for reducing the brittleness of PLA, for example, by copolymerization [59], blending [60], or adding plasticizing compounds [61]. The copolymerization approach has successfully been applied for poly(ester-urethane)s by equipping the prepolymers with elastomeric properties by copolymerization. e-Caprolactone–lactic acid copolymers have been reported to result in a significant increase in the strain and the flexibility [62]. Table 3.2 shows the changes in the material properties that have been achieved by varying the prepolymer composition [50, 62]. The softening point of poly(ester-urethane)s based on e-caprolactone–LA prepolymers can be varied to a large extent by changing the e-caprolactone (CL) content. The properties of thermoplastic poly(L-lactic acid-co-e-caprolactone-urethane)s changed according to the molar ratio of the monomers in the copolymer. Small amounts of CL increased the strain of the poly(ester-urethane)s, while at higher CL content the poly(ester-urethane)s exhibited lower strength but high elongation [50, 62]. TABLE 3.2 33 The low heat deflection temperature of PLA limits its use for several application fields, such as in packaging materials and electronic components. The introduction of rigid building blocks [63] or cross-links [64] is known, for instance, to increase the glass transition temperature and/or heat resistance of lactic acid based polymers. The effect of different amounts of comonomers in the prepolymers on the Tg and mechanical properties of poly(ester-urethane)s is demonstrated in Table 3.2. The heat resistance of poly(ester-urethane)s can be improved by the copolymerization of lactic acid with D,L-mandelic acid. This broadening of the operating temperature range is of clear practical importance. The incorporation of other comonomers that impede rotation and make polymer chains less mobile also causes an increase in Tg, even if the same comonomers can depress the rate of polycondensation [50]. The hydrolysis behavior of amorphous lactic acid based poly(ester-urethane)s is similar to that of regular PLA, with a typical water absorption and decrease in molecular weight followed by weight loss at a later stage [65]. The biodegradation of poly(ester-urethane)s has been evaluated in several studies [66]. It has been found that increasing the amount of diisocyanate used as a linking agent increases the biodegradation rate to some extent, which has been explained by an activating effect of a degradation product attributed to the linking agent. All the poly(ester-urethane)s in this study did biodegrade; that is, 90% of the theoretical CO2 was produced during 6 months, as stipulated in the CEN standard for biodegradability of packaging materials [67]. In a further part of the study, the Flash test, which is based on the kinetic measurement of bioluminescence of Vibrio fischeri, was applied to evaluate the formation of potentially toxic metabolites in the compost matrix during the biodegradation. The poly(ester-urethane) based on 1,6-hexamethylene diisocyanate produced a toxic response in the test. The poly(esterurethane) prepared by using 1,4-butane diisocyanate, on the other hand, did not show any toxic effects [66]. 3.3.2 Chain Extension with Bis-2-oxazoline Bis-2-oxazolines were described in the 1960s as useful in the preparation of poly(ester-amide)s and manufacturing processes were later developed, for instance, for chain extension of aromatic polyesters [68, 69]. Recently, bis-2-oxazolines have also been applied in the linking of LA based prepoly- Thermal and Mechanical Properties of Poly(ester-urethane)s Composition LA:butanediol LA:e-caprolactone LA:e-caprolactone LA:DL-mandelic acid:butanediol LA:DL-mandelic acid:butanediol Ratio 98:2 93:7 63.7:36.3 89.1:8.9:2 78.9:19.1:2 Tg ( C) 53 35 À5 58 60 Tensile Strength (MPa) 47 Æ 2 23 Æ 3 1.6 Æ 30.1 34 Æ 38 49 Æ 31 Strain (%) 3.7/0.3 420/20 900/50 1.8/0.4 3.1/0.1
  • 34 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) mers. Prepolymers with predominantly carboxylic acid termination were linked using 2,20 -bis(2-oxazoline) as chain extender [70]. 2-Oxazolines are inert toward aliphatic alcohols [71] and accordingly react selectively with the carboxyl end group of the prepolyester through ring opening between positions 1 and 5 of the oxazoline, yielding compounds possessing both amide and ester bonds (Figure 3.5). 3.3.2.1 Chain-Extension Reaction Parameters The molecular weight of the poly(ester-amide) strongly depends on the polymerization temperature and the molar ratio of oxazoline and carboxylic acid end groups [70]. High molecular weight polymers can be produced only within a narrow range of polymerization parameters. The amount of 2-oxazoline must be optimized because too high an excess of oxazoline results in a dominant blocking reaction and the hydroxyl end group concentration becomes too high, leading to faster degradation than polymerization. An optimal polymerization was achieved when a molar ratio of end groups of 1.2:1.0 (Ox/COOH) at 200 C was used. At lower temperature, the linking reaction is insufficient and at higher temperatures significant thermal degradation takes place. At optimal conditions, the linking process can be completed in a few minutes [70]. 3.3.2.2 Properties of Poly(ester-amide)s Poly(ester-amide)s are, like poly(ester-urethane)s, amorphous polymers, but provide an interesting alternative to other biodegradable polyesters due to the incorporated oxamide linkage in the polyester backbone. This feature equips the polymer with different mechanical properties as well as stability when compared to poly(ester-urethane)s. A slightly higher mechanical strength and lower elongation have their origin in the rigid configuration of the linking agent. The presence of the oxamide linkage affects both the hydrolytic and the thermal stability. The blocking of the terminal groups reduces the melt degradation and the increased hydrophilicity speeds up the hydrolytic degradation [55, 70, 72]. Poly(esteramide)s undergo biodegradation well in the time framework stipulated in the norms. The ecotoxicity issue observed for the poly(ester-urethane)s prepared by 1,6-hexamethylenediisocyanate can also be avoided if 2,20 -bis(2-oxazoline) is used as the chain extender [66]. FIGURE 3.6 Chain-extension reactions of lactic acid based prepolymers using bis-epoxies. 3.3.3 Dual Linking Processes The selective reactivity of oxazolines provides a possibility of performing dual linking processes with both diisocyanates and oxazolines. It has been shown that the order of addition of the linking agent affects both the reaction and the structure of the polymers, for example, the degree of branching. Simultaneous addition of 2,20 -bis(2-oxazoline) and 1,6-hexamethylene diisocyanate results in a slower increase of the molecular weight than does sequential addition. This approach is of particular benefit in the linking of lactic acid prepolymers prepared by polycondensation of solely LA. Prepolymers with a higher molecular weight and lower acid number can be prepared without the addition of a diol. Linking with oxazoline will, in this way, both increase the molecular weight and further reduce the acid number, which will make the subsequent diisocyanate linking more successful and result in higher molecular weight, less thermal degradation, and shorter reaction times in total [55]. The lactic acid based polymers prepared by dual linking with diisocyanates and 2-oxazolines exhibit properties of both poly(ester-urethane)s and poly(ester-amide)s. This, together with the fact that the dual linking process can be used for controlling the branching, opens a wider field of applications, for instance, in applications where the melt flow behavior or the hydrolytic degradation need to be tailored. 3.3.4 Chain Extension with Bis-epoxies Bis-epoxies have been reported to be useful for chain extending PLA. The epoxy groups can react both with hydroxyl groups and with carboxylic acid groups (Figure 3.6). However, the latter reaction has been found to be the more rapid one. The opening of the epoxy ring yields a secondary hydroxyl, but this one does not readily react further with remaining epoxy groups [51]. 3.4 LACTIC ACID BASED POLYMERS BY RINGOPENING POLYMERIZATION FIGURE 3.5 Chain-extension reactions of lactic acid based prepolymers using bis-2-oxazolines. The generally applied ROP process for polylactides involves three separate steps: polycondensation, lactide manufacturing, and ring-opening polymerization (Figure 3.1). All three chemical processes have basically been known for a long
  • LACTIC ACID BASED POLYMERS BY RING-OPENING POLYMERIZATION time. Carothers et al. [73] did the first observations on the reversible formation of the ring-formed dimer of a-hydroxy acids and the self-condensation ability of LA was discovered even longer back in time [74]. The results of these pioneering works have later been utilized in further scientific studies as well as in making and improving technically and economically feasible processes. The three different processes deal with a number of critical steps. Some of these issues are intrinsically present in all the steps of the manufacturing process due to the nature of the lactic acid molecule, while other issues are generated in the separate process steps as a result of the process conditions. The most crucial parameters are summarized below along with a summary of the harmful effects that can be seen in the PLA: (i) Racemization: The racemization may have its origin in the optical purity of LA or be generated and pronounced in any of the process steps. An increased amount of the antipodal structure of the repeating unit will result in drastic changes in the crystallization behavior and eventually affect many other properties of the end products [75, 76]. (ii) Lactide Purity: The lactide can contain impurities such as acids or oligomers formed during the depolymerization or purification step. The presence of impurities in the lactide and the amount thereof will affect polymerization rate, molecular weight, or both [77]. (iii) Residual Monomer Content: The presence of residual lactide in the polymer and the amount thereof will have harmful effects onthe performance of the polymer during processing and may also cause undesired property changes in the end products [78]. A few complete process descriptions going from LA to polylactide can be found in the literature, for instance, the Cargill process [79], the Inventa Fisher process [80], and the Boehringer process [81]. However, most references are found on new scientific results and detailed process improvements, which will be discussed in the following sections. 3.4.1 Polycondensation Processes Many technical processes involve esterification reactions and these have traditionally been of main importance in the preparation of polyesters. Previous chapters have dealt with the preparation of mainly high molecular weight LA polymers involving polycondensation. This chapter focuses on processes where the polycondensation is a process step in the ROP polylactide manufacturing chain and the PLA prepared is generally of low molecular weight. The lactide manufacturing is done by depolymerization of PLA that preferably is in the Mw range of 400–2500 g/ 35 mol [82]. Both catalyzed and uncatalyzed polycondensation reactions of lactic acid as such have been known as an industrial process since the 1940s and can be considered to be common knowledge [83]. This explains the fact that not more than a few relevant patents can be found for the polycondensation of LA into low molecular weight polymer. The patent literature found is mainly related to other inventions in connection with the polycondensation process, for instance, the use of different catalyst systems, such as solid inorganic catalysts containing alumina silicate [84] or alkali metal compounds [26]. The former reference also provides the conditions for the polycondensation processes in terms of temperature increase (from 105 to 150 C) along with pressure reduction (350–30 mmHg). Further references are related to technical solutions for the polycondensation process. One invention describes the production of a lactic acid polycondensate with a degree of polymerization of 1.59–2.63 and immediately separation of the lactide contained in the polycondensate [85]. Another uses an adiabatic reactor at 120–180 C and has a recycling loop, and it eliminates the water vapor [86]. Two additional references for polycondensation processes focus on improving water removal [87, 88]. 3.4.2 Lactide Manufacturing It was earlier mentioned that the reversible lactide formation from polycondensated lactic acid was initially explored by Carothers. He furthermore observed that manipulation of the temperature and pressure could be utilized for pushing the equilibrium toward the lactide product. This was utilized later for the preparation of lactide, but the presence of other species (e.g., lactic acid, water, lactoyllactic acid, lactoyllactoyllactic acid, and higher oligomers) necessitates further purification of the crude lactide to make it useful for polymerization purposes. Various technologies for lactide manufacturing are found in the literature. Batch-wise or continuous manufacturing processes have been described, as well as the use of different catalysts [89, 90]. A typical manufacturing process on an industrial scale involves heating PLA to 130–230 C at reduced pressure in the presence of 0.05–1.0 wt% of tin dust, or an organic tin compound derived from a carboxylic acid having up to 20 carbon atoms, in such a way that the produced lactide is distilled off and the PLA is continually or batchwise replenished [91]. In some processes, a fluid is used to make the separation of lactide more efficient. This can, for instance, be done by stripping off and recovering lactide from a gaseous nonreactive feed containing lactic acid polycondensate [92]. The crude lactide will in most cases contain different impurities that will make the monomer mixture unsuitable for direct ring-opening polymerization as such. The optical purity, acid number, and yield of the lactide will accordingly affect the economy of the manufacturing to a large extent. There are mainly three purification approaches
  • 36 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) suggested in the literature: solvent-assisted purification, crystallization from the melt, and purification in the gas phase. 3.4.2.1 Solvent-Assisted Purification Solvent-assisted purification has been described in both the scientific and the patent literature. The most commonly used method for purifying lactide is by crystallization from ethyl acetate or from toluene and subsequent drying of the lactide under vacuum [93, 94]. A few industrial processes for manufacturing and purifying lactide from solvents have been described in the patent literature. In one reference, lactide is purified by washing with C4–12 ethers [95]. In another approach, lactide is recrystallized in a suitable solvent followed by azeotropic distillation of the solvent [96]. In the third approach, lactide is dissolved in an organic solvent that is immiscible with water, the solution is extracted with water, and the lactide is isolated from the organic solvent [97]. One method utilizes the difference in the hydrolytic degradation rate of meso- and L-lactide. The crude lactide is purified in water whereby the meso-lactide is hydrolyzed into lactoyllactic acid and solubilized, whereas the pure L-lactide is crystallized and can be separated [98]. 3.4.2.2 Melt Crystallization The differences in the melting points for D- or L-lactide and the meso-lactide can be used for separating the different lactides from each other and from other impurities. Crystallization of lactide from the melt has recently been described as a three-step industrial process including crystallization–sweating–melting [99]. The process can also be combined with other lactide manufacturing techniques to improve the lactide quality. Another process is described as an integrated process for the manufacture of purified lactide, where the final step requires that concentrated lactide is subjected to melt crystallization to separate lactide fractions [100]. 3.4.2.3 Separation in the Gas Phase The differences in the boiling point of the different lactic acid species can be utilized in the purification of lactide. The most volatile compounds are water and lactic acid; meso- and racemic pure lactides are less volatile, and lactic acid oligomers are often in the liquid phase. One purification method employs a gas stream for purification of an impure cyclic ester by passing a gaseous inert substance through the impure cyclic ester in a molten state and removing the gas stream, whereby the purified cyclic ester is recovered from the gas stream [101]. There are several references on lactide distillation processes applied on an industrial scale, and even if the processes are similar, they all have different detailed technical solutions. Some basic differences between the processes described are in many cases found in how the outlet for the purified lactide is arranged. In one method, the liquefied gaseous impurities are separated from the solidified lactide, after which the liquefied impurities are returned to the lactide synthesis step [102]. Another process involves producing pure lactide using a final step of purifying meso-lactide from L-lactide and/or D-lactide by distillation to give one mesoenriched purified lactide stream and one meso-depleted purified lactide stream [103]. A further and improved process uses a partial condensation of the rectified lactide, whereby the low boiling point gaseous fraction remains as a vapor and is discarded. The lactide fraction is condensed and passed in the liquid phase to a distillation column [104]. 3.4.3 Ring-Opening Polymerization ROP of L-lactide is generally the most preferred route for preparing high molecular weight polylactide due to the possibility of an accurate control of the chemistry, and thus varying the properties of the resulting polymers in a more controlled manner. This makes ROP well suited for a largescale process. Polymerizations of lactide have successfully been carried out by using melt polymerization, bulk polymerization, solution polymerization, and suspension polymerization techniques. Each of these methods has its own advantages and disadvantages, but melt polymerization is generally considered the most simple and reproducible method and will be discussed later in more detail [105]. 3.4.3.1 Reactor Design The simplest type of reactor system is a reaction vessel with an agitator. The number of vessels can vary depending on the desired polymerization conditions [106]. A combination of this type of reactor and a static mixer has also been developed for a continuous polymerization process for preparing polyesters from glycolide, lactide, or CL. The column type of plug-flow reactor is preferably equipped with agitation blades in order to ensure appropriate mixing [107]. A similar concept is described in another patent but the static mixer here can optionally be linked to an extruder as the final process step [108]. A static mixer for continuous ROP of lactide is described in another U.S. patent. The mixer is equipped with mixing elements designed to enable mixing in both axial and crosswise directions [109]. ROP can also be performed by reactive extrusion, provided that the residence time and catalyst efficiency match [110]. 3.4.3.2 Catalyst Systems A vast number of catalysts have been utilized in the ROP of lactide, of which the most studied are the carboxylates and alkoxides of Sn [111–120] and Al [121–127]. Of these, stannous 2-ethylhexanoate (tin octanoate) is the most intensively studied. The polymerization mechanism is suggested to involve a preinitiation step, in which stannous 2-ethylhexanoate is converted to a stannous alkoxide by reaction with a hydroxyl-bearing compound. Then, the polymerization proceeds on the tin–oxygen bond of the alkoxide ligand, whereas the carboxylate itself is
  • REFERENCES inactive in the polymerization [114]. A review with emphasis on Sn- and Al-catalyzed ring-opening polymerization has been published by Stridsberg et al. [128]. Some kinetic studies were also included in the review. However, the highly active catalysts based on, for example, tin compounds are toxic [129], and efficient catalysts showing less toxicity based on Ca [130–136], Fe [93, 137–147], Mg [148–151], and Zn [151–158] have, therefore, been developed for lactide and lactone polymerization. Many of these, however, tend to cause racemization of PLA, especially when polymerizing at high temperatures. In addition to the aforementioned metals, Kricheldorf et al. [159] used other salts prepared from cations and anions belonging to the human metabolism in the ROP. Zinc lactate was found to be the most efficient of the tested catalysts with regard to reactivity and obtaining high molecular weight PLA. Recently, however, a potassium-based catalyst been reported to be more efficient in the ROP of polylactide to high molecular weight [160]. Other catalyst/ initiator systems of low-toxicity metals for ROP have been discussed in a study by Okada [161]. Catalysts have been developed for the stereoselective ROP of lactides. Semicrystalline PLAs were prepared from both meso-lactide (yielding syndiotactic PLA) and racemic lactide (yielding stereoblock isotactic PLA) using chiral aluminum catalysts containing bulky ligands [125, 126, 162]. After annealing, a Tm of 152 C was obtained for the syndiotactic PLA and the racemic PLA was reported to have a Tm of 191 C. The high Tm for the latter PLA was believed to result from stereocomplex formation of synthesized stereoblock PLA. The work using aluminum catalysts in stereoselective polymerization has continued [127, 163– 165], and other metal complexes have been utilized as well [151, 157, 166–168]. Many of the studies though were conducted only in solution; therefore, the selectivity of the catalyst in, for example, melt polymerization remains unclear. Metal-free catalysis of ROP was recently reviewed [169]. Both organocatalytic (nucleophilic, cationic, and bifunctional) and enzymatic approaches were discussed. 3.4.3.3 Post-Polymerization Treatments Post-polymerization treatments for PLA prepared by ROP are much related to the processing and processability of the polymer. The processing of PLA is more demanding than that of commodity plastics due to the hygroscopic nature and the limited melt stability that can lead to hydrolytic degradation. The post-polymerization treatments can mainly be divided into those performed in the melt as a finishing process or those done as a subsequent and independent processing step. Excluded from this review are post-polymerization treatments involving simple polymer modification by the use of processing aids and other additives. The processes performed in the melt that are described in the literature are mainly focusing on improving the melt stability and the processabil- 37 ity. Catalyst deactivation is one important feature that has been applied to PLA. Deactivators used include phosphorous containing compounds [170, 171], antioxidants [172], acrylic acid derivatives [173], and organic peroxides [174]. The catalyst deactivation is generally performed in combination with a lactide removal process, which can be done by removing the low Mw material at low pressures and at a temperature sufficiently high for distillation (devolatilization) [175, 176]. This process has been further developed by applying an inert gas flow in addition to the reduced pressure, which enables improved removal of the unreacted lactide [177]. The recovery of lactide has also been integrated in the polymerization process of new PLA as a means for improving the efficiency in the manufacturing chain [178]. Another way of reducing the lactide content for PLA is to apply solid-state polymerization of the residual lactide containing PLA below its Tm, which besides reducing the residual lactide content also increases the molecular weight of the polymer [179]. Separate post-polymerization treatments of PLA have also been described in the literature. Drying of the polymer is generally done before processing to minimize the thermohydrolysis and molecular weight reduction during the melt processing. Suggested drying conditions for PLA are 60 C under vacuum and the use of hot dry air [180]. More recently, the end-of-life options of bio-based polymers have been brought into sustainability discussions. For PLA, this can be seen in the form of a number of suggested approaches on how to deal with waste materials from the polymerization process, the manufacturing process of end products, or the end product after its use. Converting of PLA into lower molecular weight polymers has been described, as well as the complete hydrolysis of the polymer into lactic acid for use as new building blocks for either biosolvents or polymers [181, 182]. REFERENCES 1. H. Tsuji, Y. Ikada, Macromol. Chem. Phys. 1996, 197, 3483–3499. 2. J. Huang, M. S. Lisowski, J. Runt, E. S. Hall, R. T. Kean, N. Buehler, J. S. Lin, Macromolecules 1998, 31, 2593–2599.  3. A. S€dergard, M. Stolt, Prog. Polym. Sci. 2002, 27, o 1123–1163. 4. I. Ajioka, K. Enomoto, K. Suzuki, A. Yamaguchi, J. Environ. Polym. Degrad. 1995, 3, 225–234. 5. S.-I. Moon, C. W. Lee, I. Taniguchi, M. Miyamoto, Y. Kimura, Polymer 2001, 42, 5059–5062. 6. K. Hiltunen, M. H€rk€nen, J. V. Sepp€l€, T. V€€n€nen, a o aa aa a Macromolecules 1996, 29, 8677–8682. 7. G. M. Bohlmann, Y. Yoshida,CEH Marketing Research Report; Biodegradable Polymers, SRI International, February 2000.
  • 38 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) 8. R. K. Kulkarni, K. C. Pani, C. Neuman, F. Leonard, Arch. Surg. 1966, 93, 839–843. 9. Purac Biomaterials, available at http://www.puracbiomaterials.com (access date: November 22, 2008). 10. Boehringer Ingelheim, Biodegradable monomers and polymers, available at www.boehringer-ingelheim.com (access date: October 21, 2008). 11. J. C. Bogaert, P. Coszach, Macromol. Symp. 2000, 153, 287–303. 12. R. Auras, B. Harte, S. Selke, Macromol. Biosci. 2004, 4, 835–864. 13. M. Crank, M. Patel, F. Marscheider-Weidemann, J. Schleich, B. Heusing, G. Angerer,in: O. Wolf (Ed.), Tehno-Economic Feasibility of Large-Scale Production of Bio-Based Polymers in Europe, European Commission, EUR 22103 EN, 2005. 14. NatureWorks LLC, available at http://www.natureworksllc. com (access date: November 23, 2008). 15. Trends in Japan, Plastic from plants, available at http://webjapan.org/trends/business/bus050107.html (access date: November 24, 2008). 16. Total and Packaging, PLA (PolyLactic Acid): an ambitious research project, available at http://www.packaging.chemicals.total.com/content/documents/37_ato%20packscope% 2031%20en%20-%20pla.pdf (access date: November 18, 2008). 17. B. Merzenich,PLA: developments and perspectives, 1st PLA World Congress, Munich, September 9–10, 2008. 18. H. van der Pol,Lactides: the missing link for PLA, 1st PLA World Congress, Munich, Sept. 9–10, 2008. 19. S. Kki, I. Bodnr, J. Borda, G. Dek, M. Zsuga, J. Phys. e a a Chem. B 2001, 105, 2833–2836. 20. N. M. Qureshi, B. Woodfine, EP 0937743, 1999 (to Kobe Steel). 21. A. Duda, S. Penczek, Macromolecules 1990, 23, 1636–1639. 22. D. K. Yoo, D. Kim, D. S. Lee, Macromol. Res. 2005, 13, 68–72. 23. Y. M. Harshe, G. Storti, M. Morbidelli, S. Gelosa, D. Moscatelli, Macromol. Symp. 2007, 259, 116–123. 24. G. X. Chen, H. S. Kim, E. S. Kim, J. S. Yoon, Eur. Polym. J. 2006, 42, 468–472. 25. S-I. Moon, C. W. Lee, M. Miyamoto, Y. Kimura, J. Polym. Sci. Part A 2000, 38, 1673–1679. 26. H. Maruyama, T. Murayama, N. Yanagisawa, N. Tsuzaki, EP 0848026, 1998 (to Kyowa Yuka). 27. Y. H. Kim, K. D. Ahn, Y. K. Han, S. H. Kim, J. B. Kim,U.S. Patent 5,434,241, 1995 (to Korea Institute of Science and Technology).  28. A. S€dergard, M. Stolt, EP 1525249, 2005 (to Tate & Lyle o Public Limited Co.). 29. N. Yanagisawa, T. Nezu, K. Hotta, T. Murayama, N. Tsuzaki, S. Kodama, H. Maruyama, Y. Yokomori, EP 0823448, 1998 (to Kyowa Hakko Kogyo). 30. K. Miyazaki, H. Noguchi, T. Ota, A. Kasai, H. Yamaoka, EP 0792901, 1997 (to Mitsubishi). 31. S.-I. Moon, I. Taniguchi, M. Miyamoto, Y. Kimura, C. W. Lee, High Perform. Polym. 2001, 13, 189–196. 32. Y. Terado, H. Suizu, M. Takagi, M. Ajioka, S. Hiraoka, M. Sakai, H. Suzuki, R. Shinagawa, S. Ogawa, Y. Kotaki, EP 0953589, 1999 (to Mitsui Chemicals). 33. S. Hiraoka, H. Tsuboi, M. Sakai, EP 1114840, 2001 (to Mitsui Chemicals). 34. K. Fukushima, Y. Furuhashi, K. Sogo, S. Miura, Y. Kimura, Macromol. Biosci. 2005, 5, 21–29. 35. K. Fukushima, Y. Kimura, Macromol. Symp. 2005, 224, 133–143. 36. H. Xu, M. Luo, M. Yu, C. Teng, S. J. Xie, Macromol. Sci. Part B 2006, 45, 681–687. 37. K. Enomoto, M. Ajioka, A. Yamaguchi, EP 0572675, 1993 (to Mitsui Chemicals). 38. K. Goto, Y. Mori, M. Miyamoto, EP 0826711, 1998 (to Mitsui Chemicals). 39. M. Ohta, S. Obuchi, Y. Yoshida, EP 0603889, 1994 (to Mitsui Chemicals). 40. M. Ajioka, K. Enomoto, K. Suzuki, A. Yamaguchi, Bull. Chem. Soc. Jpn. 1995, 68, 2125–2131. 41. S. Tamai, Y. Mori, K. Goto, K. Watanabe, O. Kohgo, K. Shimizu, T. Kataoka, T. Kuroki, W. Yamashita, H. Mizuta, T. Nagata, EP 0769512, 1997 (to Mitsui Chemicals). 42. Y. Terado, H. Suizu, C. Higuchi, M. Ajioka, EP 0829503, 1998 (to Mitsui Chemicals). 43. K. Hiltunen, J. V. Sepp€l€, M. H€rk€nen, J. Appl. Polym. Sci. aa a o 1997, 63, 1091–1100. 44. C. J. Spaans, J. H. de Groot, F. G. Dekens, A. J. Pennings, Polym. Bull. 1998, 41, 131–138. 45. Y. Liu, M. S€derqvist-Lindblad, E. Ranucci, A. C. Albertsson, o J. Polym. Sci. Part A 2001, 39, 630–639. 46. Y. Di, S. Iannace, E. Di Maio, L. Nicolais, Macromol. Mater. Eng. 2005, 290, 1083–1090. 47. W. Zhong, J. Ge, Z. Gu, W. Li, X. Chen, Y. Zang, Y. Yang, J. Appl. Polym. Sci. 1999, 74, 2546–2551. 48. K. Hiltunen, J. V. Sepp€l€, M. H€rk€nen, J. Appl. Polym. Sci. aa a o 1997, 64, 865–873. 49. P. Bruin, G. J. Veenstra, A. J. Nijenhuis, A. J. Pennings, Macromol. Chem. Rapid Commun. 1988, 9, 589–594. 50. J. Kylm€, M. H€rk€nen, J. V. Sepp€l€, J. Appl. Polym. Sci. a a o aa 1997, 63, 1865–1872. 51. P. Bonsignore,U.S. Patent 5,470,944, 1995 (to Arch Dev Corp.). 52. E. Ranucci, Y. Liu, M. S€derqvist-Lindblad, A. C. Albertsson, o Macromol. Rapid. Commun. 2000, 21, 680–684. 53. C. J. Spaans, V. W. Belgraver, O. Rienstra, J. H. de Groot, R. P. H. Veth, A. J. Pennings, Biomaterials 2000, 21, 2453–2460. 54. J. V. Sepp€l€, M. H€rk€nen, K. Hiltunen, M. Malin, J. Kylm€, aa a o a PCT/WO 9601863, 1996 (to Alko Group Ltd.; Neste Oy). 55. J. Kylm€, J. Tuominen, A. Helminen, J. V. Sepp€l€, Polymer a aa 2001, 42, 3333–3343. 56. A. C. Draye, J. J. Tondeur, J. Mol. Catal. A: Chem. 1999, 138, 135–144.
  • REFERENCES 57. A. Helminen, J. Kylm€, J. Tuominen, J. V. Sepp€l€, Polym. a aa Eng. Sci. 2000, 40, 1655–1662.  58. M. Stolt, K. Hiltunen, A. S€dergard, Biomacromolecules o 2001, 2, 1243–1248. 59. D. W. Grijpma, G. J. Zondervan, A. J. Pennings, Polym. Bull. 1991, 25, 327–333. 60. M. E. Broz, D. L. var der Hart, N. L. Washburn, Biomaterials 2003, 24, 4181–4190. 61. N. Ljungberg, B. Wesslen, J. Appl. Polym. Sci. 2002, 86, 1227–1234. 62. J. Kylm€, J. V. Sepp€l€, Macromolecules 1997, 30, a aa 2876–2882.  63. S. Inkinen, A. S€dergard, M. Stolt, PCT/WO 2008056136, o 2008 (to Tate & Lyle Public Limited Co.). 64. D. W. Grijpma, E. Kroeze, A. J. Nijenhuis, A. J. Pennings, PCT/WO 9313154, 1993 (to DSM N.V.). 65. K. Hiltunen, J. Tuominen, J. V. Sepp€l€, Polym. Int. 1998, 47, aa 186–192. 66. J. Tuominen, J. Kylm€, A. Kapanen, O. Venelampi, M. a It€vaara, J. V. Sepp€l€, Biomacromolecules 2002, 3, a aa 445–455. 67. CEN 13432; European Standardization Committee, 2000. 68. K. Fukui, T. Kagiya, S. Narisawa, T. Maeda,GB 1117798, 1969 (to Sunimoto Chemical Co. Ltd.). 69. H. Inata, S. Matsumura, M. Ogasawara, EP 0020944, 1981 (to Teijin Ltd.). 70. J. Tuominen, J. V. Sepp€l€, Macromolecules 2000, 33, aa 3530–3535. 71. R. Po, L. Abis, L. Fiocca, R. Masani, Macromolecules 1995, 28, 5699–5705. 72. J. Tuominen, J. Kylm€, J. V. Sepp€l€, Polymer 2002, 43, 3–10. a aa 73. W. H. Carothers, G. L. Dorough, F. J. van Natta, J. Am. Chem. Soc. 1932, 54, 761–772. 74. J. Pelouze, Ann. Chemie 1845, 8, 112–124. 75. D. W. Grijpma, A. J. Pennings, Macromol. Chem. Phys. 1994, 195, 1649–1663. 76. J. J. Kolstad, J. Appl. Polym. Sci., 1996, 62, 1079–1091. 77. X. Zhang, D. A. Macdonald, M. F. A. Goosen, K. B. Mcauley, J. Polym. Sci.Part A 1994, 32, 2965–2970. 78. S. Jacobsen, H. G. Fritz, P. Dege, P. Dubois, R. Jrome, e e ˆ Polymer 2000, 41, 3395–3403. 79. J. J. Kolstad, D. R. Witzke, M. H. Hartmann, E. S. Hall, J. F. Nageroni, PCT/WO 9950345, 1999 (to Cargill Inc.). 80. L. Gerking, R. Hagen, K. Richter, F. Idler, W. Reimann, B. Hanzsch, PCT/WO 0181610, 2001 (to Inventa-Fisher GmbH). 81. D. Reichert, F-D. Klingler, H. Schwall, A. Christmann, B. Buchholz, EP 0372221, 1990 (to Boehringer Ingelheim Pharma KG). 82. P. Gruber, E. Hall, J. Kolstad, M. Iwen, R. Benson, R. Borchardt, PCT/WO 9509879, 1995 (to Cargill Inc.). 83. P. D. Watson, Ind. Eng. Chem. 1948, 40, 1393–1397. 84. M. Tanaka, Y. Ogawa, T. Miyagawa, T. Watanabe, EP 0172636, 1986 (to Wako Pure Chem. Ind. Ltd.). 39 85. R. Sinclair, R. Markle, R. Smith, EP 0552226, 1993 (to Biopak Technology Ltd.). 86. R. Teissier, S. Tretjak, E. Burtin, PCT/WO 03066187, 2003 (to Atofina). 87. H. Maeda, K. Shimizu, Y. Kurishiru, E. Kawada, K. Fujisawa, EP 0808861, 1997 (to Kobe Steel Ltd.). 88. K. K. Bhatia, N. E. Drysdale, K. Lin, R. S. Nash, T. W. Stambaugh, PCT/WO 9317993, 1993 (to Du Pont). 89. M. Mueller, DE 3632103, 1988 (to Boehringer Ingelheim). 90. H. E. Bellis, K. K. Bhatia, WO 92/00292, 1992 (to Du Pont). 91. M. Mueller, DE 3708915, 1988 (to Boehringer Ingelheim). 92. K. K. Bhatia, EP 0531458, 1993 (to Du Pont).  93. M. Stolt, A. S€dergard, Macromolecules 1999, 32, o 6412–6417. 94. J. W. Leenslag, A. J. Pennings, Makromol. Chem. 1987, 188, 1809–1814. 95. P. Hoessel, G. Hoffmann, F. Braun, D. Kratz, M. Brudermueller, T. Witzel, EP 0588222, 1994 (BASF AG). 96. P. Degee, P. Coszack, BE 1008099, 1996 (to Brussels Biotech SA). 97. K. S. de Vries, EP 0275581, 1988 (to Akzo NV). 98. K. Hiltunen, J. V. Sepp€l€, FI 105185, 1999 (to JVS Polymers aa Oy). 99. M. Stepanski,1st PLA World Congress, Munich, September 9–10, 2008. 100. W. G. O’Brian, L. A. Cariello, T. F. Wells, EP 0777664, 1997 (to Ecological Chem. Prod.). 101. K. K. Bhatia, EP 0531462, 1993 (to Du Pont). 102. H. Ohara, M. Ogaito, DE 19631633, 1998 (to Shimadzu Corp.). 103. P. Gruber, E. Hall, J. Kolstad, M. Iwen, R. Benson, R. Borchardt, EP 1247808, 2002 (to Cargill Inc.).  104. M. Meerdink, A. S€dergard, EP 1694626, 2005 (to Tate & o Lyle Public Limited Co.). 105. A. J. Nieuwenhuis, J. Clin. Mater. 1992, 10, 59–67. 106. H. Ebato, S. Imamura, EP 0661325, 1995 (to Dainippon Ink and Chem.). 107. T. Iiyama, T. Sato, M. Yamada, EP 0916684, 1999 (to Daicel Chem.). 108. H. Ebato, S. Oya, Y. Kakizawa, H. Furuta, K. Arai, EP 0618250, 1994 (to Dainippon Ink and Chem.). 109. T. Morio, K. Yasutoshi,U.S. Patent 5,484,882, 1996 (Dainippon Ink and Chem.). 110. S. Jacobsen,Darstellung von Polylactiden mittels reaktiver Extrusion, Ph.D. thesis, Universit€t Stuttgart, Germany, a 2000. 111. A. J. Nijenhuis, D. W. Grijpma, A. J. Pennings, Macromolecules 1992, 25, 6419–6424. 112. J. Dahlmann, G. Rafler, Acta Polym. 1993, 44, 103–107. 113. H. R. Kricheldorf, I. Kreiser-Saunders, C. Boettcher, Polymer 1995, 36, 1253–1259. 114. A. Kowalski, A. Duda, S. Penczek, Macromolecules 2000, 33, 7359–7370.
  • 40 INDUSTRIAL PRODUCTION OF HIGH MOLECULAR WEIGHT POLY(LACTIC ACID) 115. A. Kowalski, J. Libiszowski, A. Duda, S. Penczek, Macromolecules 2000, 33, 1964–1971. 116. H. R. Kricheldorf, I. Kreiser-Saunders, A. Stricker, Macromolecules 2000, 33, 702–709. 117. K. Stridsberg, M. Ryner, A. C. Albertsson, Macromolecules 2000, 33, 2862–2869. 118. M. Ryner, A. Finne, A. C. Albertsson, H. R. Kricheldorf, Macromolecules 2001, 34, 7281–7287. 119. K. B. Aubrecht, M. A. Hillmyer, W. B. Tolman, Macromolecules 2002, 35, 644–650. 120. H. R. Kricheldorf, B. Fechner, Biomacromolecules 2002, 3, 691–695. 121. P. Dubois, C. Jacobs, R. Jrome, P. Teyssi, Macromolecules e ˆ e 1991, 24, 2266–2270. 122. A. Kowalski, A. Duda, S. Penczek, Macromolecules 1998, 31, 2114–2122. 123. P. Dege, P. Dubois, R. Jrome, S. Jacobsen, H.-G. Fritz, e e ˆ Macromol. Symp. 1999, 144, 289–302. 124. J. L. Eguiburu, M. J. Fernandez-Berridi, J. San Roman, Polymer 2000, 41, 6439–6445. 125. T. M. Ovitt, G. W. Coates, J. Polym. Sci. Part A 2000, 38, 4686–4692. 126. C. P. Radano, G. L. Baker, M. R. Smith, J. Am. Chem. Soc. 2000, 122, 1552–1553. 127. Z. Zhong, P. J. Dijkstra, J. Feijen, J. Am. Chem. Soc. 2003, 125, 11291–11298. 128. K. M. Stridsberg, M. Ryner, A. C. Albertsson, Adv. Polym. Sci. 2002, 157, 41–65. 129. M. C. Tanzi, P. Verderio, M. G. Lampugnani, M. Resnati, E. Dejana, E. Sturani, J. Mater. Sci.: Mater. Med. 1994, 5, 393–396. 130. P. Dobrzynski, J. Kasperczyk, M. Bero, Macromolecules 1999, 32, 4735–4737. 131. Z. Zhong, M. J. K. Ankon, P. J. Dijkstra, C. Birg, M. e Westerhausen, J. Feijen, Polym. Bull. 2001, 46, 51–57. 132. Z. Zhong, P. J. Dijkstra, C. Birg, M. Westerhausen, J. Feijen, Macromolecules 2001, 34, 3863–3868. 133. Z. Zhong, S. Schneiderbauer, P. J. Dijkstra, M. Westerhausen, J. Feijen, J. Polym. Environ. 2001, 9, 31–38. 134. L. Piao, M. Deng, X. Chen, L. Jiang, X. Jing, Polymer 2003, 44, 2331–2336. 135. Z. Zhong, S. Schneiderbauer, P. J. Dijkstra, M. Westerhausen, J. Feijen, Polym. Bull. 2003, 51, 175–182. 136. D. J. Darensbourg, W. Choi, O. Karroonnirun, N. Bhuvanesh, Macromolecules 2008, 41, 3493–3502. 137. H. R. Kricheldorf, A. Serra, Polym. Bull. 1985, 14, 497–502. 138. H. R. Kricheldorf, C. Boettcher, Macromol. Chem. 1993, 194, 463–473. 139. A. Nakayama, N. Kawasaki, I. Arvanitoyannis, J. Iyoda, N. Yamamoto, Polymer 1995, 36, 1295–1301. 140. I. Arvanitoyannis, A. Nakayama, E. Psomiadou, N. Kawasaki, N. Yamamoto, Polymer 1996, 37, 651–660. 141. H. R. Kricheldorf, D. O. Damrau, Macromol. Chem. Phys. 1997, 198, 1767–1774.  142. A. S€dergard, M. Stolt, Macromol. Symp. 1998, 130, o 393–402.  143. A. S€dergard, M. Stolt, EP 0937116, 1998 (to Fortum Oyj). o  144. M. Stolt, K. Jalander, A. S€dergard, Polym. Prepr. 1999, 40 o (2), 675–676. 145. B. J. O’Keefe, S. M. Monnier, M. A. Hillmyer, W. B. Tolman, J. Am. Chem. Soc. 2001, 123, 339–340. 146. P. Dobrzynski, J. Kasperczyk, H. Janeczek, M. Bero, Polymer 2002, 43, 2595–2601. 147. B. J. O’Keefe, L. E. Breyfogle, M. A. Hillmyer, W. B. Tolman, J. Am. Chem. Soc. 2002, 124, 4384–4393. 148. R. Dunsing, H. R. Kricheldorf, Polym. Bull. 1985, 14, 491–495. 149. H. R. Kricheldorf, S.-R. Lee, Polymer 1995, 36, 2995– 3003. 150. J. Kasperczyk, M. Bero, Polymer 2000, 41, 391–395. 151. B. M. Chamberlain, M. Cheng, D. R. Moore, T. M. Ovitt, E. B. Lobkovsky, G. W. Coates, J. Am. Chem. Soc. 2001, 123, 3229–3238. 152. M. Bero, J. Kasperczyk, Z. J. Jedlinski, Makromol. Chem. 1990, 191, 2287–2296. 153. M. Bero, J. Kasperczyk, G. Adamus, Makromol. Chem. 1993, 194, 907–912. 154. G. Schwach, J. Coudane, R. Engel, M. Vert, Polym. Bull. 1994, 32, 617–623. 155. G. Schwach, J. Coudane, R. Engel, M. Vert, Polym. Int. 1998, 46, 177–182. 156. E. Bukhaltsev, L. Frish, Y. Cohen, A. Vigalok, Org. Lett. 2005, 7, 5123–5126. 157. T. R. Jensen, C. P. Schaller, M. A. Hillmyer, W. B. Tolman, J. Organomet. Chem. 2005, 690, 5881–5891. 158. H. Y. Chen, H. Y. Tang, C. C. Lin, Macromolecules 2006, 39, 3745–3752. 159. H. R. Kricheldorf, I. Kreiser-Saunders, D.-O. Damrau, Macromol. Symp. 1999, 144, 269–276. 160. Y. Lemmouchi, M. C. Perry, A. J. Amass, K. Chakraborty, E. Schacht, J. Polym. Sci. Part A 2008, 46, 5348–5362. 161. M. Okada, Prog. Polym. Sci. 2002, 27, 87–133. 162. T. M. Ovitt, G. W. Coates, J. Am. Chem. Soc. 1999, 121, 4072–4073. 163. K. Majerska, A. Duda, J. Am. Chem. Soc. 2004, 126, 1026–1027. 164. P. Hormnirun, E. L. Marshall, V. C. Gibson, A. J. P. White, D. J. Williams, J. Am. Chem. Soc. 2004, 126, 2688– 2689. 165. Z. Tang, X. Chen, Y. Yang, X. Pang, J. Sun, X. Zhang, X. Jing, J. Polym. Sci. Part A 2004, 42, 5974–5982. 166. M. H. Chisholm, J. C. Gallucci, K. Phomphrai, Inorg. Chem. 2004, 43, 6717–6725. 167. M. H. Chisholm, J. C. Gallucci, K. Phomphrai, Inorg. Chem. 2005, 44, 8004–8010. 168. S. K. Russell, C. L. Gamble, K. J. Gibbins, K. C. S. Juhl, W. S. MitchellIII, A. J Tumas, G. E. Hofmeister, Macromolecules 2005, 38, 10336–10340.
  • REFERENCES 169. D. Bourissou, S. Moebs-Sanchez, B. Martin-Vaca, C. R. Chimie 2007, 10, 775–794. 170. GB 588834, 1947 (to Resinous Prod. & Chemical Co.) 171. N.A. Higgins,U.S. Patent 2,683,136, 1954 (to Du Pont). 172. P. Gruber, J. J. Kolstad, E. S. Hall, R. S. E. Conn, EP 0615532, 2000 (to Cargill Inc.). 173. J. J. Kolstad, D. R. Witzke, M. H. Hartmann, E. S. Hall, J. F. Nageroni, EP 1070097, 2001 (to Cargill Inc.).  174. A. S€dergard, J.-F. Selin, M. Niemi, C.-J. Johansson, K. o Meinander, EP 0737219, 1999 (to Fortum Oyj). 175. D. G. Lowrance,U.S. Patent 1,995,970, 1935 (to Du Pont). 176. P. Gruber, J. J. Kolstad, C. Ryan, EP 0615529, 1994 (to Cargill Inc.). 177. H. Ohara, S. Sawa, M. Ito, Y. Fujii, M. Oota, H. Yamaguchi, U.S. Patent 5,770,682, 1998 (to Shimadzu Corp.). 41 178. H. Ohara, T. Okamoto,U.S. Patent 5,728,847, 1998 (to Shimadzu Corp.). 179. H. Ohara, S. Sawa, T. Kawamoto, EP 0664309, 1995 (to Shimadzu Corp.). 180. Crystallizing and Drying of PLA, NatureWorks LLC, available at http://www.natureworksllc.com/product-andapplications/ingeo-biopolymer/technicalresources/$/media/ Product%20and%20Applications/Ingeo%20Biopolymer/ Technical%20Resources/Processing%20Guides/ProcessingGuides_CrystallizingandDryingPLA_pdf.ashx (access date: November 10, 2008). 181. F. G. Hutchinson, EP 0244114, 1987 (to ICI PLC). 182. K. Yamamoto, T. Tani, T. Aoki, Y. Hata, EP 1310517, 2003 (to Wako Pure Chem. Ind. Ltd.; Tadeka Chem. Ind. Ltd.).
  • 4 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) ANN-CHRISTINE ALBERTSSON, INDRA KUMARI VARMA, BIMLESH LOCHAB, ANNA FINNE-WISTRAND, AND KAMLESH KUMAR 4.1 INTRODUCTION High molecular mass poly(lactic acid) (PLA) is obtained either by the polycondensation of lactic acid (LA) or by ring-opening polymerization (ROP) of the cyclic dimer 2,6-dimethyl-1,4-dioxane-2,5-dione commonly referred to as dilactide or lactide. The structures and stereochemistry of lactide monomers have already been described in earlier chapters (Chapters 1 and 3). Commercial PLA is usually prepared from L-lactide (LLA), because the resulting polymer poly(L-lactic acid) (PLLA) is semicrystalline with a relatively high melting and glass transition temperature (Tg). The mechanical properties are retained very near to its melting temperature. The monomer LLA can be prepared with relatively high enantiopurity from corn starch fermentation. PLLA is a versatile, semicrystalline, degradable polymer having excellent mechanical properties, good biocompatibility, and low toxicity. It has been used in a variety of applications in the pharmaceutical and biomedical fields, as well as used as a degradable plastic for disposable consumer products. In tissue engineering, PLLA has been used as biodegradable scaffold where the transplanted cells can remold their intrinsic tissue superstructural organization and thereby lead to the desirable three-dimensional structure and physiological functionality of a regenerated organ [1]. However, PLLA has certain shortcomings that limit its applications. It is a brittle, hard polymer with a very low elongation at break, is hydrophobic, and takes a long time to degrade. The properties of PLLA can be tailor made by copolymerization (random, block, and graft), change in molecular architecture (hyperbranched polymers, star shaped, or dendrimers), functionalization (end group functionalization or pendant groups such as carboxyl, amino, or thiol), or blending with other polymers. Physical properties, such as glass transition temperature (Tg), melting temperature (Tm), crystallinity, and hydrophobicity, and mechanical properties can be significantly affected by such modifications. The functionalization of PLLA can provide specific biointeractions with cells, which is specifically needed in tissue engineering. Several reviews have been published in the past correlating structure of the lactic acid and other hydroxyl acid polymers and copolymers with their properties and applications [1–10]. In this chapter, the preparation of polymers and copolymers of lactic acids with different structures, using polycondensation and ring-opening polymerization, is described. The effect of structure and composition on the properties of structurally modified polymers is also discussed. 4.2 COPOLYMERIZATION The stereocopolymers of lactic acid, prepared by the polymerization of various stereoisomers, are discussed in a subsequent section in this book and will not be discussed here. Typical comonomers that have been used for lactic acid or lactide copolymerization are glycolic acid or glycolide (GA) [11–17], poly(ethylene glycol) (PEG) or poly(ethylene oxide) (PEO) [15–43], poly(propylene oxide) (PPO) [16–18], (R)-b-butyrolactone (BL), d-valerolactone (VL) [44–46], e-caprolactone (CL) [47–54], 1,5-dioxepan2-one (DXO) [55–60], trimethylene carbonate (TMC) [61], Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, Edited by Rafael Auras, Loong-Tak Lim, Susan E. M. Selke, and Hideto Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 43
  • 44 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) FIGURE 4.1 Structure of various comonomers. N-isopropylacrylamide (NIPAAm) [62–65], and so on. The structures of some of these comonomers are given in Figure 4.1. The monomer distribution (random or block) in the copolymers depends on the monomer pairs, the nature of the catalysts, and the polymerization conditions. 4.2.1 Synthesis of Copolymers of Lactic Acid: Glycolic Acid Low molecular weight poly(lactide-co-glycolide) (PLA-coGA) copolymers have been prepared by the step-growth polycondensation of lactic acid and glycolic acid. Such copolymers are obtained by heating the desired composition of monomers and removing the water evolved during the polycondensation by atmospheric or vacuum distillation. The copolymers thus obtained are brittle and glassy or waxy and sticky, depending on the feed composition and molecular weight. A copolymer with a weight-average molecular weight of 160,000 g/mol and a Tg of 55 C was prepared by dehydration condensation of lactic acid and glycolic acid in diphenyl ether in the presence of tin powder. Films with a tensile strength of 570 kg/cm2 and an elongation of 14% were obtained [11]. High molecular weight copolymers, on the other hand, are prepared by ring-opening polymerization of L- or D-lactide and 1,4-dioxane-2,5-dione or diglycolide (generally referred to as glycolide) under inert atmosphere or in vacuum. The polymerization can be carried out in bulk, in solution (tetrahydrofuran (THF), toluene, dioxane, etc.), or in suspension or emulsion. The temperature of bulk polymerization is generally in the range of 100–150 C, whereas in solution polymerization, low temperatures (0–25 C) have been used to minimize the side reactions (inter- and intramolecular transesterifications). Melt polymerization at temperatures higher than the melting point of the monomers has also been reported [1]. Many organometallic compounds such as oxides, carboxylates, and alkoxides are effective initiators for the ROP of lactones. Metallic zinc, zinc lactate, and stannous-2-ethyl hexanoate (also referred to as stannous(II) octanoate [Sn(Oct)2]) either alone or with a coinitiator (water or alcohol) have been used as initiators for the polymerization of lactide. Both zinc lactate and stannous octanoate are approved by the American Food and Drug Administration (FDA) for surgical and pharmacological applications. Stannous octanoate is a preferred catalyst because it gives a faster reaction rate and higher molecular weight than zinc metal or zinc lactate. Zn-catalyzed PLA polymers exhibited more hydrophilicity and degradation susceptibility than the Sninitiated polymers [3, 5, 13]. However, it has been shown that stannous octanoate is slightly cytotoxic [3, 5, 14]. When polymerization was done using high initiator to monomer ratios, residues such as ethyl-2-hexanoic acid or hydroxy tin octanoate have been detected. The repeated precipitation of polymer from organic solvent in methanol or ethanol was not sufficient to remove such hydrophobic residues. PLGA copolymers having lactyl content of 70–90% have been prepared by copolymerization at 160 C for 20 h using a desired ratio of monomers and [Sn(Oct)2] as a catalyst [15]. The weight-average molecular weight of the copolymers ranged from 9.07 to 7.95 Â 104 g/mol. Bulk polymerization of LLA and GA (75:25) using [Sn(Oct)2] at 60 C for 2 h (for removal of traces of water) and at 165 C for 4.5 h gave a polymer with weight-average molecular weight of 50,000–70,000 g/mol [16]. Block copolymers of lactide with glycolide have been prepared by sequential addition of monomer onto the reactive chain end of polymer produced from another monomer, or by using a hydroxyl-terminated homopolymer as a chain transfer agent [17]. The polymerization of lactide and glycolide can proceed by anionic, carbocationic, or coordination insertion mechanisms that have been described in previous sections. PLGA copolymers of varying molar masses and compositions are also available commercially. 4.2.2 Synthesis of Copolymers of Lactic Acid:Poly (ethylene glycol) The unique properties of PEG also referred to as PEO, such as solubility in water and polar organic solvents and its insolubility in nonpolar solvents such as ethyl ether and heptane, lack of toxicity, rapid clearance from the body [18], high mobility, and FDA approval for internal consumption, make it a suitable comonomer for the preparation of block copolymers of LA or LA–GA copolymers. The copolymerization of hydrophobic LLA with hydrophilic poly(ethylene oxide) or poly(propylene oxide) or their block copolymers has been used to prepare diblock or triblock copolymers [19–21]. Several triblock copolymers of LLA, D,L-lactide (DLLA), and PEO in which PEO forms the central block have been reported in the literature [22–27]. These copolymers are more hydrophilic, flexible, and biodegradable than PLLA homopolymer [28]. The hydrophilic domains of PEO act as a
  • COPOLYMERIZATION surface modifier of hydrophobic PLA domains of the microspheres and can promote the stability of water-soluble molecules (e.g., l-DNA) and increase their loading efficiency. The biodegradability and biocompatibility of these copolymers make them suitable candidates for controlled delivery of water-soluble molecules [29]. The diblock and triblock polymers have been prepared by bulk or solution polymerization using stannous chloride [29] or [Sn(Oct)2] at high temperature (100–150 C) for several hours [30–32], potassium tert-butoxide [33], sodium hydride [34], calcium hydride/Zn [35], or zinc metal [36]. Block copolymers have also been prepared in the absence of added catalyst [37]. High polymerization temperatures generally reduce the molecular weight of the PLLA block [32]. A wide range of products has been prepared by changing the molecular weight of PEG (1000–30,000 g/mol), or the composition of LLA monomer or LLA–GA copolymer and PEG in the initial feed. A representative structure of such triblock copolymers is depicted in Figure 4.2. One- and two-phase triblock copolymers can be prepared by controlling these parameters. Synthesis and applications of copolymers of lactide and PEG have been reviewed [38, 39]. Melt polymerization of PLA (Mw 5000–400,000 g/mol) and poly(alkylene ethers) (Mw 500–50,000 g/mol) with !1 OH per polymer unit has been carried out using transesterification catalysts such as Ti (OBu)4 at 200 C [40]. ABA triblock copolymers, consisting of L-( þ )-lactide or L-( þ )-lactide–glycolide copolymer and poly(oxyethylene), have been synthesized by bulk polymerization at 110 C for several hours using aluminum triisopropoxide as a catalyst [41], as shown in Figure 4.3. Cholesterol-tethered polymers found utility for attachment of modified polymers to the cells. Cholesterol–PEG–PDLA copolymer (CPEG–PDLA) has been reported for pro- 45 FIGURE 4.3 Schematic diagram of the synthesis of ABA triblock copolymer using aluminum triisopropoxide as a catalyst [41]. moting osteoblast attachment and proliferation [42]. CPEG–PDLA with 5 and 15 ethylene glycol units promoted osteoblast attachment and growth, while with 30 ethylene glycol units prevented adhesion and proliferation. In a typical polymerization, D,L-lactide was heated under vacuum with Sn(Oct)2 as a catalyst and PEG cholesterol ethers (oxymethylene repeat units of 5, 15, and 30) as initiators. The reaction was carried out at 130 C for 24 h (Figure 4.4). Low molecular weight oligomers of cholesteryl-(L-lactic) (CLAn) with n 20 have been synthesized in bulk at 150 C [43]. 4.2.3 Synthesis of Copolymers of Lactic Acid:dValerolactone and Lactic Acid:b-Butyrolactone FIGURE 4.2 Representative structure of triblock copolymer of P(LLA-b-PEG-b-LLA) [32]. Very few studies have been reported on the copolymerization of LA with d-valerolactone (VL) and b-butyrolactone (BL). Anionic block copolymerization of VL and LLA in the presence of potassium methoxide in THF at 20 C gave diblock copolymers with expected compositions and molecular weights [44]. Slight racemization of LLA was observed during polymerization due to transesterification reactions. The block copolymers of LA and BL have been prepared by first preparing a hydroxyl-terminated poly(b-butyrolactone) (PBL). The ring-opening polymerization of (R)-BL or
  • 46 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) (r1 ¼ 10.8) and CL (r2 ¼ 0.37) have been prepared by using lanthanide halides as initiators [50]. High molecular weight copolymers of LLA and CL using Al and Zn compounds (e. g., Et2AlOEt, aluminum acetyl acetonate, etc.) as catalysts and with microstructure varying from random to diblock copolymers have been reported [54]. In the case of aluminum acetyl acetonate, the reactivity ratio of LLA was 44 and CL was 0.25. The microstructure depended on the temperature and the kind of initiator used. A series of copolymers of D,L-lactide and CL were synthesized by ROP using zinc lactate as a catalyst and carrying out the reaction at 145 C for 8 days. The characterization of PCL and its copolymers with lactides is often done by size exclusion chromatography, DSC, NMR, and stress–strain analysis. Kister et al. used vibrational spectroscopy, particularly Raman spectroscopy, for determination of morphology, conformation, configuration, and composition of the copolymers [49]. Raman spectroscopy thus appeared to be a suitable method for the identification of P(DLA-co-CL) samples directly from solid samples without any special preparation. FIGURE 4.4 Synthetic route for synthesis of cholesterol– PEG–PDLA [42]. 4.2.5 Synthesis of Copolymers of Lactic Acid:1,5Dioxepan-2-one (RS)-BL with distannoxanes as catalyst in the presence of 1,4-butanediol as initiator gave optically active poly[(R)-BL] or atactic poly[(RS)-BL] with secondary hydroxyl chain ends and oxytetramethylene units in the backbone. These polymers may be used to initiate the copolymerization of LA at the chain ends and form block copolymers. The optically active P[(R)-BL] is brittle whereas atactic P[(RS)-BL] is elastomeric in nature and has been used to prepare elastomeric polymers with brittle PLLA. However, Sn(IV) compounds are active transesterification catalysts and may cause unit scrambling when LLA is used as a monomer. Therefore, a two-stage polymerization is carried out. In the first step, telechelic poly[(RS)-BL] in the molecular weight range of 5000–12,000 g/mol is prepared at 100 C by maintaining the desired molar ratios of (RS)-BL and 1,4butanediol and using Sn(IV) as catalyst. In the second stage, the desired ratio of hydroxyl-terminated poly[(RS)-BL] and LLA monomer is added and Sn(Oct)2 is used as a catalyst and polymerization is carried out at 160 C [45]. Hori et al. [46] have earlier reported random copolymers of LLA and (R)BL. Albertsson and coworkers [55–60] have done extensive studies on copolymers of lactic acid (lactide) and 1,5-dioxepan-2-one. The poly(1,5-dioxepan-2-one) (PDXO) is a completely amorphous, hydrophilic, and biocompatible wax-like polymer with a Tg of À39 C. When DXO is used as a comonomer with lactides, it increases the hydrophilicity and rate of degradation of the copolymers. The copolymers show characteristics of thermoplastic elastomers that are suitable for biomedical applications such as slowly degrading sutures, temporary implants, and drug vehicles. Although the synthesis of DXO was reported as early as 1972, Mathisen et al. [55] in 1989 described an improved reaction scheme with a high yield. PDXO forms a soft amorphous block while PLLA forms a hard semicrystalline segment in the triblock copolymer of P(LLA-b-DXO-b-LLA) [56–59]. Blocks of controlled lengths have been synthesized using a tin oxide initiator. Degradable polyesters with strictly defined structure, unique mechanical properties, and modified degradation profiles have been prepared and characterized [59]. The morphology of spin-coated films of triblock copolymers of PLLA and PDXO was investigated by AFM. These studies revealed the absence of nanoscale morphology in these films [60]. 4.2.4 Synthesis of Copolymers of Lactic Acid: e-Caprolactone 4.2.6 Synthesis of Copolymers of Lactic Acid: Trimethylene Carbonate Copolymerization of LA and CL has been extensively investigated [47–54]. Random copolymers of DLLA Thermoplastic elastomers (TPEs), having unusual physical and chemical properties, have been prepared by copolymer-
  • COPOLYMERIZATION ization of LLAwith TMC. Recently, a batch procedure for the preparation of biodegradable TPEs based on multiblock copolymers of LLA with TMC has been reported [61] using a combination of ring expansion polymerization and ROP. The initiator used was 2,2-dibutyl-2-stanna-1,3-oxepane. The block lengths were varied via the monomer/initiator and TMC/LLA ratio. These copolymers were transformed in situ into multiblock copolymers by ring-opening condensation with sebacoyl chloride (Figure 4.5). 4.2.7 Synthesis of Copolymers of Lactic Acid:Poly(Nisopropylacrylamide) A block copolymer of poly(N-isopropylacrylamide) (PNIPAAm) and PLA may combine the thermosensitive property of PNIPAAm and the biodegradable property of PLA. Polymeric micelles from such polymers can improve protein release properties. Temperature change can alter the hydrophilicity and conformation of PNIPAAm, which may affect the physicochemical properties of microspheres of the polymer. Amphiphilic block copolymers of P(NIPAAm-b-LA) have been prepared by ROP of LA in toluene using Sn(Oct)2 as a catalyst and hydroxyl-terminated PNIPAAm, as depicted in Figure 4.6 [62]. Such copolymers have recently been synthesized by ROP of lactide using the two hydroxyl groups of S,S0 -bis(2-hydroxyethyl-20 -butyrate)trithiocarbonate (BHBT). PLA-b-PNIPAAm-b-PLA was prepared by reversible addition–fragmentation chain transfer (RAFT) polymer- 47 ization of NIPAAm using PLA with a centered trithiocarbonate unit as a RAFT agent [63]. Self-organization of such amphiphilic block copolymers in aqueous solutions indicated the formation of vesicles. Stabilization of vesicles was attained by cross-linking chain extension of the PNIPAAm block using hexamethylene diacrylate [64]. Multifunctional micelles for cancer cell targeting, distribution, and anticancer drug delivery have been prepared using PNIPAAm-co-methacrylic acid-gPDLA and diblock copolymers [65]. 4.2.8 Synthesis of LA:Alkylthiophene (P3AT) Copolymers End-functionalized poly(3-alkylthiophene) (P3AT), where the alkyl side chain of thiophene moiety contains either 6 or 12 carbons in length, has been used as a macroinitiator for the ROP of lactide, thereby yielding rod-coil block copolymers (Figure 4.7) [66]. 4.2.9 Functional Poly(lactic acid) Polylactides having amino, carboxyl, or other functional (pendant or chain end) groups have been reported in the literature. These functional groups can be utilized for chemical modification or as binding sites for biomolecules to impart selective binding and adhesion. ROP of L- or D,Llactide using bis(hydroxymethyl) butyric acid (BHMBA) as FIGURE 4.5 Representative structure of multiblock copolymers of P(L-LA–TMC) [61].
  • 48 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) FIGURE 4.6 Reaction sequence for preparation of block copolymer of PNIPAAm and PLA [62]. an initiator and Sn(Oct)2 as a catalyst at 130 C yielded PLA with pendant carboxyl groups. The chain extension of this polymer with diisocyanate yielded poly(ester–urethane) containing carboxyl groups as pendant functional groups [67]. Thiol-functionalized PEG-b-PLA has been prepared by ROP of D,L-lactide using PEG disulfide as the macroinitiator. The disulfide bond was cleaved using tributylphosphine to generate a block copolymer having a thiol unit at the PEG end [68]. Functionalization of PLA by grafting of maleic anhydride (MAn) by reaction extrusion in the presence of free radical initiators (tert-butyl peroxide) has also been carried out. The presence of high succinic anhydride units in the grafts was confirmed by FTIR and NMR studies. Low percentage grafting was observed in PLA due to presence of limited free radical sites [69, 70]. Finne and Albertsson have introduced a double bond in the polylactide by using 1,1-di-n-butyl-stanna-2,7-dixacyclo-4-heptene as the initiator [71, 72]. The presence of a double bond in the lactide macromonomer provided a variety of opportunities for further modification. For example, epoxidation was carried out with m-chloroperoxybenzoic acid (mCPBA) and a quantitative conversion of the double bond to epoxide was observed. PLA-functionalized polyoxanorbornenes with one or two exo-PLA chains as well as two endo-, exo-chains were prepared using Sn(Oct)2 as a catalyst in the presence of mono- or dialcohol derivatives of oxanorbornenes [73]. These macromonomers are then subjected to ring-opening metathesis polymerization (ROMP) to yield graft copolymers (Figure 4.8). Low molecular weight oligo(lactide)s end capped with fumarate groups have been used for in situ cross-linkable scaffolds for tissue engineering [74]. Side chain functionalized diastereomeric lactides were synthesized from commercially available amino acids and their subsequent polymerization or copolymerization [75]. This approach allows the incorporation of any protected amino acid for the preparation of functionalized cyclic monomers. The quantitative deprotection of amino acids leads to the formation of new functionalized PLA-based polymers. Protected functional lactide copolymers can be synthesized by the copolymerization of dibenzyloxy-substituted monomers with lactide. Deprotection followed by modification with succinic anhydride with carboxyl side chains is suitable for peptide coupling. Such a modification can control the attachment of cells in tissue engineering and other biomedical applications [76]. 4.2.10 FIGURE 4.7 Synthesis of poly(3-alkylthiophene)-b-PLA [66]. Branched Copolymers Studies on copolymers of PLA having a core of PLA (or another comonomer) and branches of another monomer (or lactide comonomer) have been extensively reported in the literature. Graft copolymers having different architec-
  • COPOLYMERIZATION 49 O O OH OMe + n O Sn(Oct) 2 O O O O O O n O OMe O H ROMP O m MeO O O O O O n O O O OH OH + n Sn(Oct) 2 O O O O O O exo, exo exo, endo FIGURE 4.8 H O O O H n O O O O n H Synthesis of polylactide PLA-functionalized polyoxanorbornenes and subsequent polymerization (ROMP) [73]. tures (linear branches, hyperbranched, star-like, brush-like, and comb-like) have been synthesized with an aim to modify the properties of the PLA. The hydrophilicity or crystallinity of these copolymers can be significantly varied and controlled by preparation of these architectures. A general reaction for the preparation of such copolymers is depicted in Figure 4.9. Branched PLA is different from linear PLA in physical, thermal, and mechanical properties. Such polymers have been prepared by using multifunctional alcohols, for example, inositol, pentaerythritol, glycerol, and so on [77–82]. FIGURE 4.9 Finne and Albertsson prepared four-arm star-shaped PLLA using novel spirocyclic tin initiators [77]. Kricheldorf et al. [78] polymerized LLA using bismuth triacetate and pentaerythritol as initiator and coinitiator, respectively. Kim et al. [79] and Arvanitoyannis et al. [80] used Sn(Oct)2 or tetraphenyl tin and pentaerythritol or glycerol, respectively, as the initiator and coinitiator system for LLA polymerization. Similar star-shaped polymers using multifunctional alcohols were reported by Korhonen et al. [81]. As can be seen, many types of branched PLA have been prepared by using organometallic catalysts and multifunc- Synthetic route for the preparation of branched PLLA.
  • 50 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) FIGURE 4.10 Reaction scheme of enzymatic polymerization [82]. tional alcohols. Recently, branched PLA has been synthesized using lipase PS (Pseudomonas fluorescens) catalyzed ring-opening polymerization of lactide monomers (LLA, DLA, DLLA) (Figure 4.10) [82]. Branched copolymers have also been synthesized by the preparation of macromonomers. Various types of methacrylate-functionalized macromonomers have been reported in the literature for the preparation of graft and star copolyesters. The reaction scheme used for the preparation of the macromonomers is depicted in Figure 4.11. Segmented terpolymers of poly(alkyl methacrylate)-gpoly(D-lactide)/poly(dimethylsiloxane) (PLA/PDMS) were prepared by combination of a ‘‘grafting through’’ technique (macromonomer method) and controlled/living radical polymerization such as atom transfer radical polymerization (ATRP) or reversible addition–fragmentation chain transfer polymerization. In a single-step approach, the low molecular weight methacrylate monomer (methyl methacrylate (MMA), butyl methacrylate (BuMA)) (Figure 4.12) was polymerized with a PLA or PDMS macromonomer. The second strategy was a two-step approach in which a graft copolymer containing one macromonomer was polymerized with PLA macromonomer (Figure 4.12) [83]. 4.2.10.1 Graft Copolymers As mentioned earlier, the macromolecular design of a polymer regulates its physicochemical properties. Advanced structures such as combs, brushes, ladders, and so on have been synthesized to meet the vast demands from different targeted applications of such polymers. Several graft copolymers based on LA have been prepared to modify the properties such as degradability, transition temperatures (Tg and Tm), morphology, mechanical properties, and solubility. Surface characteristics of PLA films have also been modified by grafting. Micelle structures, having a multifunctional core and hydrophobic shell, have been developed with higher drug activity and lower material FIGURE 4.11 Synthesis of macromonomers.
  • COPOLYMERIZATION FIGURE 4.12 51 Structure of poly(lactide) and poly(dimethylsiloxane) macromonomers and macroinitiators [83]. toxicity. Some of these modifications are described in the following text. The star-shaped highly branched polymers are discussed separately in Section 4.2.10.2. To prepare degradable polymers, graft copolymers of PLA macromonomer and tert-butyl acrylate were prepared by free radical polymerization. An increase in lactic acid units resulted in an increase in degradation rate [84]. ATRP of MMA (96.5%) and (meth)acrylate-terminated PLA macromonomer (Mn 2800 g/mol, 3.5%) yielded a homogeneously branched PMMA-g-PLA of low polydispersity index (PDI ¼ 1.15) [85]. The reactivity ratio of MMA for conventional radical polymerization is 1.09 while with ATRP is 0.57. This accounts for the lower PDI of ATRP synthesized PMMA-gPLA. Biodegradable comb-like polymer prepared by free radical copolymerization of PLA macromonomer with vinyl (N-vinylpyrrolidone) and acrylic monomers (MMA, methacrylic acid (MA)) has been reported [86]. The hydroxyl group of poly(2-hydroxyethyl methacrylate) (HEMA) anchored to a gold surface-initiated ROP of lactide using Sn (Oct)2 as a catalyst. The final structure of the polymer is a surface-anchored poly(hydroxyethyl methacrylate)-g-polylactide, to yield an overall structure of a ‘‘bottlebrush’’ [87]. PLA and its random copolymer with glycolide have also been grafted onto poly(vinyl alcohol) to increase hydrophilicity and manipulate the structure [88]. A novel comb-type PLA was prepared using a depsipeptide–lactide random copolymer having pendant hydroxyl groups as macroinitiator for graft polymerization of LA. The comb-type polymer had a lower Tg, Tm, and crystallinity than linear PLA [89]. A graft copolymer of poly(NIPAAm-co-methacrylic acid)-g-DLLA along with diblock copolymers of PDLLA with PEG and poly(2-ethyl-2-oxazoline) has been used for formation of mixed micelles with a multifunctional core and shell morphology. These micelles exhibited higher drug activity and lower material cytotoxicity [90]. New thermoresponsive, pH-responsive, and biodegradable nanoparticles comprising PDLA-g-(NIPAAm-comethacrylic acid) were prepared by grafting PDLA onto NIPAAm-co-methacrylic acid copolymer. A core–shell structure was formed with a hydrophilic outer shell and a hydrophobic inner core that exhibited a phase transition temperature above 37 C. The drug loading level of 5-fluorouracil (5-FU) as encapsulated nanoparticles from these copolymers could be as high as 20%. The release of 5-FU was controlled by the pH in the aqueous medium. These studies indicated that these nanoparticles can be used as a drug carrier for intracellular delivery of anticancer drugs [91]. The surface characteristics of solvent cast PLA films have been modified by grafting with vinyl acetate, acrylic acid, and acrylamide by a UV-induced photopolymerization process [92]. In biological systems, an organism is capable of creating the proper organic matrix as a substrate for the nucleation and growth of inorganic crystals due to the interfacial interaction between inorganic and organic phases. In analogy, in vitro fabrication of novel inorganic/organic composites of inorganic crystals could be controlled by organic polymer. For this purpose, PLA surfaces have been modified with poly(methacrylic acid) via photooxidation and UV-induced polymerization. Thus, the introduced carboxyl groups onto PLA surfaces act as the nucleation sites of hydroxyapatite crystals. Nanohydroxyapatite/PLA composites with interfacial interaction between the two phases have been prepared using these graft copolymers [93]. FTIR, XRD, and SEM studies supported that the modified PLA could act as a template to control the nucleation, growth, morphology, size, and distribution of hydroxyapatite crystals over the organic phase. A thermoplastic polyolefin elastomer TPO-g-PLA was prepared by grafting PLA onto maleic anhydride-functionalized TPO in the presence of 4-dimethyl aminopyridine (DMAP). A high reaction temperature and a high DMAP concentration resulted in the polymerization of PLA. These copolymers were used as a compatibilizer for PLA/TPO blends. An increase in concentration of this copolymer from 0% to 2.5% resulted in an increase in elongation at break and tensile toughness of the blends [94]. Butanediamine (BDA)-g-PDLLA was synthesized by grafting maleic anhydride onto the side chains of PDLLA via melt-free radical polymerization using benzoyl peroxide as initiator. BDA was then grafted via a N-acylation reaction. The biodegradation behavior of these graft copolymers could
  • 52 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) be controlled by the content of BDA. Uniform degradation was observed in these copolymers with no acidity-induced autoacceleration [95]. New amphiphilic graft copolymers of hyaluronic acid (HA) have been prepared by grafting both hydrophobic (PLA) and hydrophilic branches (PEG) on the backbone. The copolymers (PLA-g-HA-g-PEG) were characterized by spectroscopic techniques. Branched PLA with various lengths of graft chains were synthesized by ROP of L- or D-lactide with polyglycidol as an initiator [96]. The branched PLLA had a lower Tg, Tm, crystallinity, and Young’s modulus and higher strain at break than the corresponding linear PLLA or PDLA film. The PLA surface was chemically modified by a singlestep, nondestructive grafting technique using vinyl monomers such as acrylamide, maleic anhydride, and N-vinylpyrrolidone in the vapor phase. Benzophenone was used as a photoinitiator under solvent-free conditions. The modified surfaces exhibit higher wettability, and the grafting was verified by X-ray photoelectron spectroscopy, attenuated total reflection, FTIR, contact-angle measurements, and scanning electron microscopy. The graft chain pendant groups remain functional and can subsequently be modified so that a tailor-made surface with desired properties may be achieved [97]. Acrylic acid-grafted polylactide (PLA-g-AA) and multihydroxyl-functionalized multiwalled carbon nanotubes have been melt blended to improve thermal stability and mechanical properties of the composite. The formation of a covalent bond (ester linkage) resulted in a significant improvement in compatibility [98]. Alternatively, carboxylic acid-functionalized multiwalled carbon nanotubes were grafted onto PLLA by a one-step in situ polycondensation reaction [99]. Acrylic acid-grafted PLA, titanium tetraisopropylate, and starch blends have been prepared by an in situ sol–gel and melt blending processes. The carboxylic acid groups of acrylic acid act as a coordination site for the titania phase to form a Ti–O–C chemical bond. This resulted in a nanoscale dispersion of TiO2 in the polymer matrix [100]. PLA-g-dextrans having various lengths and number of grafted chains and sugar units were synthesized using the FIGURE 4.13 trimethylsilyl protection method. The surface of these films is believed to be covered with hydrophilic dextran segments, which leads to the suppression of cell attachment and protein absorption onto the film [101–103]. Poly(lactide-g-dextran) copolymers were synthesized by a three-step process: partial silylation of the dextran hydroxyl groups, ROP of D,L-lactide initiated by the remaining hydroxyl groups of dextran, followed by silyl ether deprotection under mild conditions. The emulsifying properties of these glycopolymers depend on the PLA/dextran ratio [104]. PLAg-dextran and PLA-g-silylated dextran adopt a core–shell conformation in various solvents [105]. Studies on encapsulation and release behavior of bovine serum albumin from PLA-g-dextran revealed a higher loading than in PLLA microspheres [106]. Studies on PLA-g-gelatin have been extensively reported in the literature. These bioabsorbable amphiphilic polymers are useful for parenteral drug delivery systems and tissue engineering. These copolymers have been prepared by the ROP of LLA onto functionalized gelatin using bulk copolymerization at 140 C or solution copolymerization at 80 C with Sn(Oct)2 as the catalyst. The number of grafting sites on the gelatin chain could be adjusted by partial trimethyl silylation of pendant hydroxyl, amino, and carboxylic acid groups [107]. 4.2.10.2 Star Copolymers A block copolyester of PLA and hydroxyl-terminated four-armed poly(e-caprolactone) was prepared by a condensation reaction using Sn(Oct)2 or Fe(OAc)2 as an initiator. The formation of a block copolyester as well as a blend was indicated, as shown in Figure 4.13 [108]. Precision synthesis of microstructures in star-shaped copolymers of e-caprolactone, LLA, and DXO was accomplished using a spirocyclic tin initiator and stannous(II) octoate (cocatalyst) together with pentaerythritol ethoxylate (coinitiator) [109]. Four-arm star-shaped PDLLA oligomers of controlled molar mass and narrow molar mass distribution were synthesized by using ethoxylated pentaerythritol initiator. The terminal hydroxyl group was converted to methacrylate Condensation of PCL and PLA to yield star-shaped branched copolyesters [108].
  • (methacrylic anhydride) or 2-isocyanatoethyl methacrylate end groups. Photo-cross-linking of these functional oligomers yielded networks with high gel contents. The Tg of the copolymers depended on the prepolymer molar mass [110]. The four-arm star-branched block copolymer of L-PLA and PEO [111] was investigated for the release of anticancer drugs 5-FU and paclitaxel. The drug release of paclitaxel from the micellar nanoparticles could be controlled over 2 weeks and completely released. The rapid and complete release of drug was accounted due to the rapid degradation of micelles from star-shaped copolymer, compared to linear block copolymers. Star copolymers have been prepared by using tosylated b-cyclodextrin (b-CD) as a core and lactide 2ethyl-2-oxazoline copolymers as branches. The hydroxyl functional group of (Tosyl)7-b-CD was used as the initiator for the ring-opening polymerization of lactide. The hydroxyl chain end of PLA chain was later used for ROP of 2-ethyl-2oxazoline. A seven-arm copolymer having seven arms of poly(lactide-co-2-ethyl-2-oxazoline) was successfully prepared [112]. A novel biodegradable chestnut-shaped polymer having a PLA shell and hyperbranched D-mannan (HBM) was synthesized by polymerization of LLA and HBM with DMAP as catalyst. The number of PLA chains on PLA-HBM could be controlled by the ratio of DMAP to sugar [113]. The star-shaped PEO–PLA copolymers have shorter degradation times and exceptional amphiphilic characteristics and may be excellent candidates for drug release [114]. The star-shaped PLA attached to poly(amidoamine) dendrimer (PAMAM) was prepared by bulk polymerization of lactide with PAMAM. Unlike the linear PLA with similar molar mass, PAMAM-g-PLA had a higher hydrophilicity and a faster degradation rate. The highly branched structure significantly accelerated the release of water-soluble bovine serum albumin from these graft copolymers, whereas a time lag was observed in linear PLLA of similar molecular weight [115]. 4.3 PROPERTIES OF COPOLYMERS The physical properties and biodegradability of PLA copolymers can be easily controlled by changes in the structure and composition of the repeat units, flexibility of the chain, presence of polar groups, molecular mass, crystallinity, and orientation of the backbone chains. The properties of PLA depend on the stereoisomers used for their preparation. PLLA and PDLA are semicrystalline hard materials with modulus of 2.7 GPa, tensile strength of 50–70 MPa, elongation at break of 4%, flexural modulus of 5 GPa, and flexural strength of 100 MPa [116–119]. The melting point is around 180 C and Tg is 60–65 C. The molar mass of the polymer as well as degree of crystallinity have a significant influence on the mechanical properties [120– 124]. Polymerization of a racemic mixture of 1:1 D,D-LA and L,L-LA or meso-lactide gave an amorphous polymer with a Tg of 55–60 C and a tensile modulus of 1.9 GPa. The in vitro degradation of PLLA is much slower than PMLA due to its crystalline nature, and it takes 2 years for complete degradation of the former polymer. The stereoselective po-
  • 54 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) most amenable to morphological changes during preparation of porous PLGA microparticles using a supercritical carbon dioxide pressure quench treatment of particles prepared using the conventional emulsion–solvent evaporation method [132]. In poly(LA–PEG) or poly(LAGA-b-PEG) block copolymers, the LA blocks are hydrophobic while the PEG blocks are soluble in water. As a consequence, such copolymers may form a micellar structure in water and are thus potential candidates for controlled drug delivery applications. The introduction of PEG blocks in PLLA or P(LA-co-GA) copolymers increases flexibility and toughness. The Tg of PLLA–PEG block copolymers showed a strong dependence on the composition and molecular weight of PEG [34]. A significant reduction in Tg was observed by using high molecular weight PEG or a high weight percent of PEG in the initial feed composition [133]. Microspheres based on poly(DL-lactide) and triblock copolymers of P(DL-LA-bPEG-b-DL-LA) and loaded with l-DNA have been prepared by a conventional solvent evaporation method based on formation of multiple w1/o/w2 emulsion. The degradation profile of these microspheres was quite different because of more swelling in the triblock copolymer due to the presence of the PEG block. This swelling helped in maintaining a more stable condition for DNA and minimized initial burst release [29]. The poly(LLA-b-VL) copolymers having the monomers in the ratio 57:43 showed two endothermic transitions in the differential scanning calorimetry (DSC), representing the Tm of PVL and PLLA, around 52 and 156 C, respectively. However, only one Tm was observed in the block copolymers having higher ratio of one comonomer (e.g., LLA:VL ¼ 19:81 and 81:19) [44]. TPE block copolymers having semicrystalline LLA terminal blocks and an amorphous heterogeneous middle block were prepared from DXO and TMC using a five-membered tin oxide initiator. All the copolymers exhibited highly elastic behavior with a maximum stress at break of 35.6 MPa for a copolymer having no DXO and maximum strain at break 1089% when the ratio DXO:TMC:LLA was 200:200: 200 [133]. The mechanical properties of triblock (LLA-co-DXO) depend on the composition of the backbone [134] and stress at break ranged from 4 to 42 MPa and elongation at break from 463% to 828% [135]. 4.3.1 Degradation of Homo- and Copolymers The hydrolysis of copolymers of P(LA-co-GA) was investigated at 37 and 60 C for 80 days. A three-stage degradation was observed: during the first stage, the molecular weight decreased rapidly with little weight loss; in the second stage, a severe weight loss was observed and monomer formation was initiated; and in the third stage, via hydrolysis the oligomers were transformed to lactic acid and glycolic acid [136]. Hydrolysis of the triblock copolymer poly(LLAb-DXO-b-LLA) of different compositions was studied in a buffered salt solution at 37 C and pH 7.4. The rate of degradation was influenced by the original molecular weight of the sample, and the copolymer composition had no effect on the degradation [137]. During in vitro degradation carried out for 59 days for the elastic copolymers of DXO and LLA, both exhibited good retention of mechanical properties, with elongation at break 600–800% and elastic modulus 8–20 MPa. The rate of degradation of PLA has been controlled by copolymerization with monomers such as CL, glycolic acid, DXO, a-malate, glycine, HEMA, and ethylene glycol. Blending of PLLA with other polymers has also been attempted [138]. Recently, covalent grafting of PLA to tune the in vitro degradation rate has been reported. Grafting was done with acrylamide, N-vinylpyrrolidone, or acrylic acid. The in vitro rate of degradation was enhanced and the grafted surface layer was found to be covalently attached to the surface [139]. The rates of enzymatic hydrolysis (proteinase K) for branched PLLA (prepared from pentaerythritol with four branches and from polyglycerin with 22 branches) were found to be dependent on the average molecular weight of the PLLA segment in the branched molecules, not on the overall molecular weight of the samples [140]. The biodegradability of PLA has been extensively investigated [141–146]. PLLA and its copolymers have been degraded in the presence of different types of enzymes such as pronase, bromelain, Rhizopus delemar lipase, lipase from Rhizopus arrhizus, and proteinase K from Tritirachium album [141]. The enzymatic degradation by proteinase K has been the subject of several reports [142–145]. Reeve et al. [142] carried out the degradation of a series of PLA stereocopolymers by proteinase K and observed that the enzyme preferentially attacks L-lactyl units. The amorphous region of the polymer is more susceptible to degradation. The degradation of PLA stereocopolymers by proteinase K increased with a decrease in crystallinity and an increase in hydrophilicity of the polymers. The alkaline and acid hydrolysis of PLA has also been investigated. The presence of D-lactyl units reduces the hydrolysis rate [146]. 4.3.2 Drug Delivery from PLLA Copolymers The successful utilization of polymeric materials within the living body is highly dependent on the structural architecture and monomer unit distribution in the polymer. Nanoparticles with a hydrophobic surface (e.g., PLA and PLGA) are rapidly taken up by the cells of the reticuloendothelial systems (RES) [147]. Polymeric particles with a hydrophilic surface can avoid this uptake to a greater extent, thereby prolonging the lifetime in the circulation, which may help in efficient
  • REFERENCES delivery of the therapeutic agent. Self-organizing block copolymers have the possibility of entrapping a hydrophobic drug in the micelle core while the micelle’s hydrophilic shell confers water solubility. Intelligent drug delivery vehicles can be designed by utilizing shell forming polymers that exhibit stimuli-responsive behavior. Block copolymers of PDLA and PNIPAAm have been widely investigated as potentially useful carriers for targeted delivery [62–65]. Particles prepared from an amphiphilic polymer have functional sites where chemical modifications can be carried out that may provide specific surface characteristics such as charge, hydrophilicity, and targeting capabilities. Degradable graft copolymers with amino acids (lysine (Lys), aspartic acid (Asp), alanine (Ala), etc.) have therefore been prepared where the side chains could be at neutral pH (PLA-g-Ala), positively charged (PLA-g-Lys), or negatively charged (PLA-g-Asp). In such copolymers, polyamino side chains tend to concentrate at the surface of the particles [148]. The capabilities of microparticles to serve as carriers in controlled drug release and delivery devices were demonstrated by encapsulation and release of rhodamine B, a low molecular weight model. The effect of morphology on the drug release in blends as well as copolymers of PLLA and PDXO was investigated by Albertsson and coworkers. The microspheres obtained from blends were more compact and crystalline, while the copolymer microspheres had an amorphous structure that affected the hydrolysis under humid conditions. The storage stability of copolymers was studied for 5 months and was found to be less than that of blends [149]. Albertsson and coworkers [150] have recently reported a nondestructive preparation of resorbable polymer scaffolds with heparin covalently bonded to the PLA surface and an osteoinductive growth factor. This was achieved by photochemical vapor-phase grafting of acrylamide on PLA film and subsequent reduction of amide groups of polyacrylamide to amino groups for covalently linking heparin and immobilization of osteoinductive growth factor, recombinant human bone morphogenetic protein-2, in the heparin layer. A functionalized triblock copolymer HOOC–PLA–PEG–PLA–COOH with polybasic carboxylic end groups was found to have higher drug encapsulation efficiency [151]. Porous scaffolds were successfully fabricated from copolymers of DXO, LLA, and e-CL through a solvent casting and particulate leaching technique, in which methanol was used to wet and swell the composite before leaching, thereby leading to an interconnected porous network. In the DSC thermograms of these copolymers, only a single Tg located between corresponding copolymers was observed, indicating thereby a continuous amorphous phase due to the randomness of the copolymers [152]. Surface functionalization of porous resorbable scaffolds by covalent grafting yielded products with better hydrophilicity [153]. 55 The morphology and structural architecture of polymers has significant effect on drug release. The release rates of 5FU and paclitaxel, widely used chemotherapeutics, have been investigated in di-, tri-, and four-arm (star-branched) block copolymers of PLA and PEO. Micellar aggregates were prepared from these block copolymers and release rates were studied over 3 weeks. More complete drug release was observed in star-shaped polymers [154]. 4.3.3 Radiation Effects The effect of radiation (c- and electron beam) on the degradation of PLA and its copolymers has received considerable attention in the past [152, 155, 156]. Irradiation of polymers generates free radicals that induce chemical changes such as chain scission and cross-linking. The atmosphere of the surroundings, irradiation dose, chemical composition, and morphology of the polymer influence the degradation mechanism. The type of end groups, pendant units, and copolymer structure (such as aromatic or aliphatic units) has significant effect on the stability of the polymers toward irradiation. In aliphatic polyesters, the ester linkage and the tertiary carbons in the branched polyesters are the preferred site for the degradation. Mechanical properties and molecular weight are significantly affected by radiation. Electron beams and c-rays are used for the sterilization of temporary implants used to regenerate new tissues [157]. Generally, a dose of 2.5 Mrad is used for such purposes [158]. Albertsson and coworkers [159] have recently reported that copolymerization of LLAwith a small amount of CL or DXO increased the stability of PLLA. The most abundant low molecular weight degradation product was identified as DXO. It is obvious from the above that considerable effort has been expended in the past on the design and synthesis of lactide copolymers. Products with tailored properties can be obtained to suit the requirements of biomedical fields and pharmaceutical applications. REFERENCES 1. J. J. Grodzinski, Polym. Adv. Technol. 2006, 17, 395. 2. A. C. Albertsson, I. K. Varma, Aliphatic polyesters: synthesis, properties and applications, in: A. C. Albertsson (Ed.), Advances in Polymer Science, Vol. 157, Springer, Berlin, 2002, p. 1. 3. A. C. Albertsson, I. K. Varma, Biomacromolecules 2003, 4, 1466. 4. A. C. Albertsson, I. K. Varma, Aliphatic polyesters, in: Y. Doi, A. Steinbuchel (Eds.), Biopolymers, Vol. 4, Wiley-VCH, Weinheim, 2002, p. 25. 5. I. K. Varma, A. -C. Albertsson, R. Rajkohwa, R. K. Srivastava, Prog. Polym. Sci. 2005, 30, 949.
  • 56 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) 6. A. S€derg€rd, A. Stolt, Prog. Polym. Sci. 2002, 27, 1123. o a 7. M. Spinu, C. Jackson, M. Y. Keating, K. H. Gardener, J. Macromol. Sci., Pure Appl. Chem. 1996, A33 (10), 1497. 8. W. Amass, B. Tighe, Polym. Int. 1998, 47, 89. 9. S. Vainiop€€, P. Rokkanan, P. T€rm€l€, Prog. Polym. Sci. aa o aa 1989, 14, 679. 10. O. Tatsuro, O. Yuichi, J. Polym. Sci. Part A 2004, 42, 453. 11. M. Ajioka, C. Higuchi, T. Yamaguchi,Jpn. Kokai Tokkyo Koho JP 93-160686, 1995, CAN 123:113181. 12. Y. M. Zhao, Z. Y. Wang, J. Wang, H. Z. Mai, B. Yan, F. Yang, J. Appl. Polym. Sci., 2004, 91, 2143. 13. G. Scwach, J. Coudane, R. Engel, M. Vert, Biomaterials 2002, 23, 993. 14. M. C. Tanzi, P. Verderio, M. G. Lampugnani, M. Resnati, E. Dejana, E. Sturani, J. Mater. Sci. Mater. Med., 1994, 5, 393. 15. C. Min, W. Cui, J. Bei, S. Wang, Polym. Adv. Technol. 2007, 18 (4), 299. 16. G. Khang, J. H. Choee, J. M. Rhee, H. B. Le, J. Appl. Polym. Sci. 2002, 85, 1253. 17. I. Barakat, P. Dubois, C. Grandfils, R. Jerome, J. Polym. Sci. Part A 2001, 39, 294. 18. T. Yamoka, Y. Tabata, Y. Ikada, J. Pharm. Sci. 1994, 83, 601. 19. H. Younes, D. Cohn, Eur. Polym. J. 1988, 24, 765. 20. Y. Kimura, Y. Matsuzaki, H. Yamane, T. Kitao, Polymer 1989, 30, 1342. 21. T. Yamaoka, Y. Takahashi, T. Ohta, M. Miyamoto, A. Murakami, Y. Kimura, J. Polym. Sci. Part A 1999, 37, 1513. 22. D. Cohen, H. Younes, G. Maron, Polymer 1987, 20, 2018. 23. H. Younes, D. Cohen, J. Biomed. Mater. Res. 1987, 21, 1301. 24. X. M. Deng, C. D. Xiong, L. M. Cheng, R. P. Xu, J. Polym. Sci. Part C 1990, 28, 28. 25. Z. Jedlinski, P. Kurock, W. Walach, I. Radecka, Makromol. Chem. 1993, 194, 1681. 26. K. J. Zhu, X. Lin, S. Z. Yang, J. Appl. Polym. Sci. 1990, 39, 1. 27. H. R. Kricheldorf, C. Boettcher, Makromol. Chem. 1993, 194, 1653. 28. H. Tsuji, in: Y. Doi, A. Steinbuchel (Eds.), Biopolymers, Vol. 4, Wiley-VCH Verlag GmbH, Weinheim, 2002, p. 129. 29. X. Deng, X., M. Yuan, X. Li, L. Liu, W. X. Jia, J. Appl. Polym. Sci. 2002, 86, 2557. 30. H. R. Kricheldorf, J. M. Haack, Macromol. Chem. 1993, 194, 715. 31. J. R. Mohammadi, S. M. F. Farnia, M. N. Sarbolouki, J. Appl. Polym. Chem. 1999, 74 (8), 2004. 32. J. M. Onyari, S. J. Huang, Macromol. Symp. 2003, 193, 143. 33. H. R. Kricheldorf, C. Boettcher, Macromol. Chem. Macromol. Symp. 1993, 73, 47. 34. Z. Jedlinski, P. Kurock, W. Walach, H. Janeczek, I. Radeka, Macromol. Chem. 1993, 194, 1681. 35. S. Li, S. Anjard, I. Rashkov, M. Vert, Polymer 1998, 39 (22), 5421. 36. S. Li, M. Vert, Macromolecules 2003, 36 (21), 8008. 37. P. Cerrai, M. Tricoli, Macromol. Chem. Rapid Commun. 1993, 14, 529. 38. F. Wang, Z. Wang, Y. Zhao, Hecheng Shuzhi Ji Sulia 20 (6), 58, 2003, CAN 141:226157. 39. X. Li, X. Yuan, Huaxue Jinzhan 19 (6), 973, 2007, CAN 148:175013. 40. T. Mihara, M. Takahashi,Jpn. Kokai Tokkyo Koho JP 2004231773, 2004, CAN 141:174937. 41. Y. Li, T. Kissel, J. Control. Release 1993, 27 (3), 247. 42. G. Yu, J. Ji, J. Shen, J. Mater. Sci. Mater. Med. 2006, 17 (10), 899. 43. H. A. Klok, J. J. Hwangh, S. N. Iyer, S. I. Stupp, Macromolecules 2002, 35, 746. 44. P. Kurock, J. Penczek, J. Franek, Z. Jedlinski, Macromolecules 1992, 25(9), 2285. 45. S. Hiki, M. Miyamoto, Y. Kimura, Polymer 2000, 41, 7369. 46. Y. Hori, Y. Takahashi, A. Yamaguchi, T. Nishishita, Macromolecules 1993, 26, 4388. 47. M. Stolt, M. Viljmaa, A. S€derg€rd, P. T€rm€l€, J. Appl. o a o aa Polym. Sci. 2004, 91, 196. 48. T. Tsuji, Y. Tezuka, Macromol. Biosci. 2005, 5, 135. 49. G. Kister, G. Cassanas, M. Bergounhon, D. Hoarau, M. Vert, Polymer 1999, 41 (3), 925. 50. J. Brostr€m, A. Boss, I. S. Chronakis, Biomacromolecules o 2004, 5, 1124. 51. Y. Shen, K. J. Zhu, Z. Shen, K. M. M. Yao, J. Polym. Sci. Part A 1996, 34, 1799. 52. T. Tsuji, T. Yamada, J. Appl. Polym. Sci. 2003, 87, 412. 53. H. A. Declerco, M. J. Cornelissen, T. L. Gorski, E. H. Schacht, J. Mater. Sci. Mater. Med. 2006, 17, 113. 54. M. Bero, J. Kasperrczyk, G. Adamus, Makromol. Chem. 1993, 194, 907. 55. T. Mathisen, K. Masus, A. C. Albertsson, Macromolecules 1989, 22, 3842. 56. A. C. Albertsson, A. Loefgren, J. Macromol. Sci. Pure Appl. Chem. 1995, A-32, 41. 57. K. Stridsberg, A. C. Albertsson, J. Polym. Sci. Part A 2000, 38, 1774. 58. M. Ryner, A. C. Albertsson, Biomacromolecules 2002, 3, 601. 59. K. M. Stridsberg, M. Ryner, A. C. Albertsson, Adv. Polym. Sci. 2002, 157, 41. 60. A. Finne, N. Andronova, A. C. Albertsson, Biomacromolecules 2003, 4, 1451. 61. D. Pospiech, H. Komber, D. Jehnichen, L. Haussler, K. Eckstein, H. Scheibner, A. Janke, H. R. Kricheldorf, O. Petermann, Biomacromolecules 2005, 6, 439. 62. F. Kohori, K. Sakai, T. Aoyagi, M. Yokoyama, Y. Sakurai, T. Okano, J. Control. Release 1998, 55, 87. 63. Y. You, C. Hong, W. Wang, W. Lu, C. Pan, Macromolecules 2004, 37, 9761. 64. M. Hales, C. K. Barner, T. P. Davis, M. M. Stenzel, Langmiur 2004, 20, 10809.
  • REFERENCES 65. C. K. Huang, C. L. Lo, H. H. Chen, G. H. Hsiieue, Adv. Funct. Mater. 2007, 17, 2291. 66. B. W. Boudouris, C. D. Frisbie, M. A. Hillmyer, Macromolecules 2008, 41, 67. 67. T. Cooper, R. F. Sorey, Macromolecules 2008, 41 (3), 655. 68. N. C. Kalarickel, S. Rimer, P. Sarkar, J. C. Leroux, Macromolecules 2007, 40, 1874. 69. R. Mani, M. Bhattachraya, J. Tang, J. Polym. Sci. Part A 1999, 37, 1693. 70. W. Ma, L. Wang, R. A. Gross, S. P. Mccarthy, Polym. Mater. Sci. Eng. 1997, 76, 527. 71. A. Finne, A. C. Albertsson, J. Polym. Sci. Part A 2004, 42, 444. 72. M. Ryner, A. Finne, A. C. Albertsson, H. R. Kricheldorf, Macromolecules 2001, 34, 7281. 73. I. Czelusnik, E. Khosravi, A. M. Kenwright, C. W. G. Ansell, Macromolecules 2007, 40, 1444. 74. E. Jabari, X. He, Polym. Prepr. (ACS Div. Polym. Chem.) 2006, 47 (2), 353. 75. W. W. Gerhardt, D. E. Noga, K. I. Hardcastle, A. J. Garcia, D. M. Collard, M. Weck, Biomacromolecules 2006, 7, 1735. 76. D. E. Noga, T. A. Petri, N. Wack, A. Garcia, D. M. Collard, Polym. Prepr. (ACS Div. Polym. Chem.) 2008, 49 (1), 844;D. A. Noga, A. Kumar, D. M. Collard, M. Weck, A. Gracia, PMSE Prepr. 2007, 96, C77. 77. A. Finne, A. C. Albertsson, Biomacromolecules 2002, 3, 684. 78. H. R. Kricheldorf, H. T. Hachmann, G. Schwarz, Biomacromolecules 2004, 5, 492. 79. S. H. Kim, Y. K. Han, Y. H. Kim, S. I. Hang, Macromol. Chem. 1992, 193, 1623. 80. I. Arvanitoyannis, A. Nakayama, N. Kawasaki, N. Yamamoto, Polymer 1995, 36, 2947. 81. A. Korhonen, A. Helminen, J. V. Seppala, Polymer 2001, 42, 7541. 82. R. K. Srivastava, A. W. Finne, A. C. Albertsson, Y. Doi, H. Abe, Biomacromolecules 2007, 8(10), 3115. 83. J. F. Lutz, N. Jahed, K. Matyjaszewski, J. Polym. Sci. Part A 2004, 42(8), 1939. 84. V. Langlois, K. R. Vallee, J. J. Peron, A. le Borgne, M. Walls, P. Guerin, Polym. Degrad. Stabil. 2002, 76, 411. 85. H. Shinda, K. Matyaszewski, Macromolecules 2001, 34, 6243. 86. J. L. Eguiburu, M. J. B. Fernandez, J. R. San, Polymer 1996, 37, 3615. 87. J. B. Kim, W. Huang, C. Wang, M. Bruening, G. L. Baker, Polym. Brushes 2004, 105. 88. A. Breitenbach, K. F. Pistel, T. Kissel, Polymer 2000, 41, 4781;A. Breitenbach, D. Mohr, T. Kissel, J. Control. Release 2000, 63, 53. 89. F. Tasaka, Y. Ohya, T. Ouchi, Macromolecules 2001, 34, 5494. 90. C. L. Lo, K. M. Lin, C. K. Huang, G. H. Hsiue, Adv. Funct. Mater. 2006, 16, 2309. 91. C. -L. Lo, K. M. Lin, G. H. Hsiue, J. Control. Release 2005, 104, 477. 57 92. A. V. Janorkar, S. E. Proulx, A. T. Metters, D. E. Hirt, J. Polym. Sci. Part A 2006, 44, 6534. 93. Y. Xiao, D. Li, H. Fan, X. Li, Z. Gu, X. Zhang, Mater. Lett. 2007, 61, 59. 94. C. H. Ho, C. H. Wang, C. I. Lin, Y. D. Lee, Polymer 2008, 49, 3902. 95. Y. Luo, N. -X. Wang, C. Fu, S. Wang, Eur. Polym. J. 2007, 43, 3856. 96. T. Ouchi, S. Icchimura, Y. Ohya, Polymer 2006, 47, 429. 97. U. Edlund, M. Kallrot, A. -C. Albertsson, J. Am. Chem. Soc. 2005, 127, 8865. 98. C. S. Wu, H. T. Liao, Polymer 2007, 48, 4449. 99. W. Song, Z. Zheng, W. Tang, X. Wang, Polymer 2007, 48, 3658. 100. H. T. Lia, C. S. Wu, J. Appl. Polym. Sci. 2008, 108, 2280. 101. T. Ouchi, T. Kontani, T. Saito, Y. Ohya, J. Biomater. Sci. Polym. Ed. 2005, 16, 1035. 102. T. Ouchi, T. Kontani, Y. Ohya, J. Polym. Sci. Part A 2003, 41, 2462. 103. T. Ouchi, T. Kontani, Y. Ohya, Polymer 2003, 44, 3927. 104. J. Raynaud, B. Choquenet, E. Marie, E. Dellacherie, C. Nouvel, J. L. Six, A. Durand, Biomacromolecules 2008, 9, 1014. 105. C. Nouvel, P. Dubois, E. Dellacherie, J. L. Six, J. Polym. Sci. Part A 2004, 42, 2577. 106. T. Ouichi, T. Saito, T. Kontani, Y. Ohya, Macromol. Biosci. 2004, 4, 458. 107. J. Ma, H. Cao, Y. Li, J. Biomater. Sci. Polym. Ed. 2002, 13, 67. 108. M. Stolt, M. Viljanmaa, A. Sodergard, P. Tormala, J. Appl. Polym. Sci. 2004, 91, 196. 109. K. Odelius, A. C. Albertsson, J. Polym. Sci. Part A 2008, 46, 1249. 110. A. S. Karikari, W. F. Edwards, J. B. Macham, T. E. Long, Biomacromolecules 2005 6, 2866. 111. L. E. Salaam, D. Dean, T. L. Bray, Polymer 2006, 47 (1), 310. 112. M. Adeile, Z. Zarnegar, R. Kabiri, Eur. Polym. J. 2008, 44 (7), 1921. 113. W. J. Lin, Y. C. Chen, C. C. Lin, C. F. Chen, G. W. Chen, J. Biomed. Mater. Res. Part A 2006, 77B (1), 188. 114. L. M. Pitet, S. B. Hait, T. J. Lanyk, D. M. Knauss, Macromolecules 2007, 40 (7), 2327. 115. Q. Cai, Y. Zhao, J. Bie, F. Xie, S. Wang, Biomacromolecules 2003, 4, 828. 116. D.W. Grijpma, A. J. Nijenhuis, P. G. T. van Wijk, A. J. Pennings, J. Polym. Bull. 1992, 29, 571. 117. L. Fambri, A. Pegoretti, R. Fenner, S. D. Incardona, C. Migliaresi, Polymer 1997, 38, 79. 118. P. T€rm€l€, Clin. Mater. 1992, 10, 29. o aa 119. S. Jacobsen, H.G. Fritz, Polym. Eng. Sci. 1999, 39, 1303. 120. I. Engelberg, J. Kohn, Biomaterials 1991, 12, 292. 121. G. Perego, G. D. Cella, C. Bastioli, J. Appl. Polym. Sci. 1996, 59, 37. 122. Y. Ikada, H. Tsuji, Macromol. Rapid Commun. 2000, 21, 117.
  • 58 DESIGN AND SYNTHESIS OF DIFFERENT TYPES OF POLY(LACTIC ACID) 123. D. W. Grijpma, A. J. Pennings, J. Macromol. Chem. Phys. 1994, 195, 1633. 124. D. W. Grijpma, C. A. P. Joziasse, A. J. Pennings, J. Macromol. Chem. Rapid Commun. 1993, 14, 155. 125. C. P. Radano, G. L. Baker, M. R. Smith, J. Am. Chem. Soc. 2000, 122, 1552. 126. T. M. Ovitt, G. F. Coats, J. Am. Chem. Soc. 2000, 122, 1552. 127. G. Khang, H. B. Lee, J. B. Park, Biomed. Mater. Eng. 1995, 5, 245, 259, 278. 128. G. Khang, J. H. Jeon, J. W. Lee, S. C. Cho, H. B. Lee, Biomed. Mater. Eng. 1997, 7, 357. 129. J. H. Lee, G. Khang, J. W. Lee, H. B. Lee, J. Colloid Interface Sci. 1998, 205, 323. 130. J. H. Jang, L. D. Shea, J. Control. Release 2003, 86 (1), 157. 131. M. H. Sheridan, L. D. Shea, M. C. Peters, D. J. Mooney, J. Control. Release 2000, 64(1–3), 91. 132. K. Koushik, U. B. Kompella, Pharm. Res. 2004, 21 (3), 524. 133. B. Bunter, M. Nowak, J. Kasperczyk, M. Ryba, P. Grieb, M. Walski, P. Dobrzynski, M. Bero, J. Control. Release 1998, 56, 159. 134. A. Natalia, A. C. Albertsson, Biomacromolecules 2006, 7(5), 1489. 135. M. Ryner, A. C. Albertsson, Macromol. Symp. 2001, 175, 11. 136. M. Hakkarainen, A. C. Albertsson, S. Karlsson, Polym. Degrad. Stabil. 1996, 52, 283. 137. K. Stridsberg, A. C. Albertsson, Polymer 2000, 41, 7321. 138. M. Ryner, A. C. Albertsson, Biomacromolecules 2002, 3, 601. 139. M. Kalrot, U. Edlund, A. C. Albertsson, Biomacromolecules 2007, 8, 2492. 140. K. Numata, A. W. Finne, A. C. Albertsson, Y. Doi, H. Abe, Biomacromolecules, 2008, 9(8), 2180. 141. H. Fukuzaki, M. Yoshida, M. Asano, M. Kumakura, Eur. Polym. J. 1989, 25, 1019. 142. M. S. Reeve, S. P. McCarthy, M. J. Downey, R. A. Cross, Macromolecules 1994, 27, 825. 143. H. Cai, V. Dave, R. A. Gross, S. P. McCarthy, J. Polym. Sci. Part B 1996, 34, 2701. 144. S. Li, S. P. McCarthy, Macromolecules 1999, 32, 4454. 145. S. Li, M. Tenon, H. Garreau, C. Brand, M. Vert, Polym. Degrad. Stabil. 2000, 67, 85. 146. S. Li, A. Girrard, H. Garreau, M. Vert, Polym. Degrad. Stabil. 2001, 71, 61. 147. R. Gref, Y. Minamitake, M. T. Peracchia, V. Trubetskoy, V. Torchilin, R. Langer, Science 1994, 263, 1600. 148. G. Caponeti, J. S. Hrkach, B. Kriwet, M. Poh, N. Lotan, P. Colombo, R. Langer, J. Pharm. Sci. 1999, 88, 136. 149. U. Edlund, A. C. Albertsson, J. Polym. Sci. Part A 2000, 38, 786.  150. U. Edlund, S. Danmark, A. C. Albertsson, Biomacromolecules 2008, 9(3), 901. 151. H. Zhang, L. Chang, B. Zhang, P. Li, M. Wang, D. Y. Kang, J. Wang, F. Wao, J. Biomater. Sci. Polym. Ed. 2008, 19, 99. 152. K. Odelius, P. Plikk, A. C. Albertsson, Biomacromolecules 2005, 6, 2718. 153. M. Kallrot, U. Edlund, A. C. Albertsson, Macromol. Biosci. 2008, 8, 645. 154. P. Jie, S. S. Venkatraman, F. Min, B. Y. C. Freddy, G. L. Huat, J. Control. Release 2005, 110, 20. 155. L. Montanani, M. Costantini E. C. Signoretti, L. Valvo, M. Santucci, M. Bartoomei, J. Control. Release 1998, 56, 219. 156. S. C. J. Loo, C. P. Ooi, Y. C. F. Bocy, Polym. Degrad. Stabil. 2004, 83, 259. 157. L. Montanani, M. Costantini, E. C. Signoretti, L. Valvo, M. Santucci, M. Bartolomei, J. Control. Release 1998, 56, 219. 158. J. S. C. Loo, C. P. Ooi, Y. C. F. Bocy, Biomaterials 2005, 26, 1359. 159. P. Plikk, K. Odelius, M. Hakkarainen, A. C. Albertsson, Biomaterials 2006, 27, 5335.
  • 5 STRUCTURE AND PROPERTIES OF STEREOCOMPLEX-TYPE POLY(LACTIC ACID) MASAYUKI HIRATA AND YOSHIHARU KIMURA 5.1 INTRODUCTION Stereocomplex-type polylactides (sc-PLA), consisting of both enantiomeric poly(L-lactic acid) (PLLA) and poly(D-lactic acid) (PDLA), can potentially act as high performance polymers because their melting temperature (Tm ¼ 230 C) is 50 C higher than that of the single PLLA or PDLA polymer (180 C). The stereocomplex (sc) crystal form of sc-PLA is triclinic or trigonal with both 31 (or 32 and 31) helical PLLA and PDLA chains packed side by side, which is quite different from the homochiral (hc) pseudo-orthorhombic or orthorhombic crystal forms. Since their improved thermal and mechanical properties facilitate a wider application, various trials of the formation of sc-PLA have been done. Simple mixing of high molecular weight PLLA and PDLA, however, is likely to lead to hc crystallization rather than sc crystallization to retard the thermal stability of the resultant polymer blends. To promote sc crystallization, stereoblocktype PLA (sb-PLA), a block copolymer of PLLA and PDLA, has been developed. The covalent bond of the enantiomeric sequences of PLLA and PDLA favor sc crystallization of the copolymer. The sb-PLA can be synthesized by solid-state polycondensation (SSP) of a mixture of PLLA and PDLA having medium molecular weight as well as by stepwise ring-opening polymerization (ROP) of D- and L-lactides. 5.2 FORMATION OF STEREOCOMPLEX CRYSTALS In 1987, Ikada et al. discovered that a mixture of enantiomeric PLLA and PDLA forms sc crystals with melting temperature (Tm) 50 C higher than that of the hc crystals of PLLA or PDLA [1]. This discovery was made from solution and later from melt mixtures. It is believed that van der Waals interaction of enantiomeric polymer chains is responsible for the stereocomplexation. The degree of stereocomplexation is therefore affected by the formation conditions and chain structure, particularly by the molecular weights of PLLA and PDLA. With polymers having high molecular weights, the stereocomplexation is hindered by hc crystallization, while complete stereocomplexation can be attained with polymers having sufficiently low molecular weights [2]. Therefore, a certain threshold molecular weight below which stereocomplexation is favored is assumed for PLLA and PDLA, although it may depend on the mixing state. For the stereocomplexation from the melt mixture, the threshold molecular weight is approximately 6 kDa [3], while it increases to 40 kDa and 400 kDa in the cases of solution casting and precipitation, respectively [2–4]. Stereocomplexation between high molecular weight PLLA and PDLA has been a challenge for a long time. In fact, fast solvent evaporation from the cast films of PLLA/PDLA mixtures usually induces rapid hc crystallization together with stereocomplexation. Therefore, to achieve exclusive stereocomplexation, a dilute solution of PLLA and PDLA needs to be evaporated slowly, or isothermally crystallized in a mixture of good and poor solvents. With these conditions, sc-PLA consisting of high molecular weight polymers (approximate 100 kDa) can be obtained [5]. Application of strong shear to the melt blend of PLLA and PDLA is also effective in inducing predominant stereocomplexation. Under shear, macromolecular chains of PLA are extended, which facilitates interaction between the Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 59
  • 60 STRUCTURE AND PROPERTIES OF STEREOCOMPLEX-TYPE POLY(LACTIC ACID) enantiomeric sequences and enables the subsequent sc formation. The related high-shear extrusion and vigorous drawing are also highly effective for stereocomplexation because the highly oriented structure of PLLA and PDLA chains also facilitates the molecular interaction of PLLA and PDLA. It is known that vigorous drawing of PLLA fibers can give b-form crystals consisting of the 31 or 32 helical hc chains [6–8]. Similar drawing of PLLA/PDLA blend fibers induces stereocomplexation between the highly oriented enantiomeric molecular chains having the identical 31 (or 32 and 31) helical conformations. The most effective technique for obtaining exclusive stereocomplexation is to use stereoblock polylactides (sb-PLA) in which both the enantiomeric chains are chemically connected and the chain mixing is allowed to take place at the molecular level. 5.3 THERMAL PROPERTIES OF sc-PLA As mentioned above, the Tm of sc-PLA is much higher than that of hc crystals of PLLA or PDLA. The highest Tm experimentally determined for sc-PLA is 240 C [2, 4, 9], and the equilibrium Tm is reported to be 279 C [10], much higher than that of a crystals (190 C) [11, 12]. The crystallization temperature (Tc) of sc crystals is uncertain, although their crystallization rate is apparently faster than that of hc crystals. In many cases, sc crystallization is accompanied by hc crystallization. The glass transition temperature (Tg) of sc-PLA is reported to be 65–72 C [13], which seems to be slightly higher than that of the ordinary PLLA samples [11], probably because the chain mobility in the amorphous domain is affected by the crystalline states. 5.4 CRYSTAL STRUCTURE OF sc-PLA 5.4.1 Unit Cell Parameters and Molecular Conformation of sc-PLA In 1991, Okihara et al. first analyzed the sc crystals by X-ray crystallography [14]. They reported that the sc crystals are triclinic (P1 space group with parallel chain orientation), having cell parameters of a ¼ 0.916, b ¼ 0.916, c ¼ 0.870 (nm), a ¼ 109.2, b ¼ 109.2, and c ¼ 109.8 , and that the PLLA and PDLA macromolecular chains form 31 helices having a lamella thickness of 0.87 nm with six chains, or each three enantiomeric hc chains penetrating one unit cell. Brizzolara et al. conducted molecular simulation and proposed two types of stereocomplex structures with parallel and antiparallel orientations of the molecular chains [15]. Both structures were also considered to be triclinic, having different cell parameters of a ¼ 0.912, b ¼ 0.913, c ¼ 0.930 (nm), a ¼ 110, b ¼ 110, and c ¼ 109 (P1 space group) and a ¼ 0.930, b ¼ 0.940, c ¼ 0.930 (nm), a ¼ 111, b ¼ 112, and c ¼ 108 (P/1 space group), respectively. Their calculated interaction energies as well as the experimental data revealed that the parallel structure (P1) is more stable than the antiparallel structure (P/1). This computer simulation well supported the growth mechanism of triangular lamellar crystals in the sc formation. In 1997, Cartier et al. proposed a trigonal form for the sc crystals in which the PLLA and PDLA chains take 32 and 31 helical conformations, respectively [16]. The proposed trigonal crystal form ought to have R3c and R-3c symmetries if the two enantiomeric chains are oriented in parallel and antiparallel manners, respectively. They also assumed that the triclinic cell is just a subcell involved in the larger trigonal cell in which six helices penetrate one unit cell having the cell parameters of a ¼ b ¼ 1.498, c ¼ 0.870 (nm), a ¼ b ¼ 90, and c ¼ 120 . In a later study, Sawai et al. [5] analyzed highly oriented stereocomplex samples prepared by solid-state coextrusion of solvent cast films (extrusion draw ratio ¼ 14) that had been fabricated by slow evaporation of solvent (for a week). The WAXS of the samples showed 20 reflections, which were reasonably indexed with a trigonal unit cell proposed by Cartier et al., although the parameters were slightly different: a ¼ b ¼ 1.50, c ¼ 0.823 (nm), a ¼ b ¼ 90, and c ¼ 120 with R3c symmetry. The crystal structures proposed for sc crystals thus far are compared in Table 5.1. For comparison, the crystal parameters reported for hc crystals of PLLA homopolymer are summarized in Table 5.2. 5.4.2 Density and Heat of Fusion of sc Crystals On the basis of the crystal structure and unit cell parameters, the crystal density rsc (g/cm3) of the stereocomplex crystal can be calculated from Equation 5.1: ½M  ðB=Aފ=ðVÞ Â 1027  10À6 ¼ rsc ðg=cm3 Þ ð5:1Þ where M, A, B, and V are the molecular weight of the repeating unit, Avogadro’s number, the number of monomer units per unit cell, and the volume of the unit cells. The B and V values of trigonal crystals are 18 and (H3/2)abc (nm3), respectively, while those of triclinic crystals are 6 and 0.563 or 0.598 (g/nm3), respectively. The two possible V values of triclinic cells are deduced by the two equations: V ¼ abc(1 À cos2a À cos2b À cos2c þ 2 cos a cos b cos c)0.5 or (1/6) detjB0 –A0 , C0 –A0 , D0 –A0 j, where A0 , B0 , C0 , and D0 are positional coordinates of the triclinic cell in rectangular coordinate system. The densities thus calculated are listed in Table 5.1 [5, 14–16]. Sawai et al. [5] also determined the theoretical heat of fusion of sc crystals DHmc, which is the crucial parameter to determine the sc crystallinity (Xc). They utilized the aforementioned samples and plotted the measured heat of fusion (DHm) as a function of the specific volume of the samples (Vs)
  • FORMATION OF sb-PLA TABLE 5.1 61 Density and Unit Cell Parameters Reported for the sc Crystals Okihara et al. [14] Crystal system Chain conformation Unit cell parameter a (nm) b (nm) c (nm) a (degree) b (degree) c (degree) rcalc (g/cm3) Brizzolara et al. [15] Cartier et al. [16] Sawai et al. [5] Triclinic 31 Triclinic 31 Trigonal 31 and 32 Trigonal 31 and 32 0.916 0.916 0.870 109.2 109.2 109.8 1.27 0.912 0.913 0.930 110 110 109 1.21 1.498 1.498 0.870 90 90 120 1.27 1.50 1.50 0.823 90 90 120 1.342 to estimate DHmc by extrapolating to perfect crystallinity by the following linear relationship (5.2): DHm ¼ DHmc ½Xc Š ¼ DHm ½Va =ðVa ÀVc ÞÀVs =ðVa ÀVc ފ ð5:2Þ where Va and Vc are specific volumes of the amorphous and sc crystal domains that are estimated from the density of the sample (r). The r values used were in the range of 1.268–1.301 g/cm3, being apparently larger than the values previously calculated from the unit cell parameters. As a result, the DHmc finally obtained by extrapolation to 1/1.342 (rsc) was 155 Æ 5 J/g, which is significantly higher than those of the a, b [7, 20, 22, 23], and sc crystals reported previously except when compared with one case [11], as noted in Table 5.3. 5.5 FORMATION OF sb-PLA The sb-PLA that exclusively forms sc crystals can be synthesized by ROP of L- and D-lactides, as well as by polycondensation of PLLA/PDLA mixtures through SSP. 5.5.1 ROP Routes to Diblock and Mutiblock sb-PLA In 1990, Yui et al. [24] revealed that stereocomplex with little hc crystallization can be formed easily from diblock-type sb-PLA (PLLA-b-PDLA) because of the easy interaction of the two enantiomeric block sequences. A unique synthetic method to sb-PLA is the stereoselective ROP of rac-lactide. In 1996, Spassky et al. [25] succeeded in synthesizing multiblock sb-PLA by utilizing an aluminum catalyst with enantiomeric Schiff base ligands. However, the polymerization required long reaction time for completion (200 h) and the molecular weight of the resultant polymer was limited (Mn % 13 kDa, Mw/Mn ¼ 1.27). As a result, the Tm of the product polymers was 187 C, which is almost equivalent to that of the hc crystals. Since then, many researchers have reported using various chiral catalysts for the same stereoselective polymerization, which are basically aluminum alkoxide derivatives bearing different ligands [26–32]. These modified catalysts improved reaction time, polymer yield, and stereoselectivity (isotactic, heterotactic, and atactic as shown in Figure 5.1), although the polymerization control was insufficient in regard to the molecular weight and Tm of the resultant polymers. Two-stage ROP of L- and D-lactides seemed to be the easiest way to obtain diblock sb-PLA. However, an optimum process has not been established for attaining high molecular weight. The aforementioned Yui’s results were for the products having relatively low molecular weights. Recently, working with Musashino Chemical Laboratory Ltd. (Japan), we have succeeded in making high molecular weight diblock-type sb-PLA by bulk polymerization of L- and D-lactides (Figure 5.2) [33]. In this method, the polymer prepared by the first polymerization of L- or D-lactide was subjected to reprecipitation to remove the remaining monomer and used as the macroinitiator for the subsequent polymerization of the enantiomeric lactide. The molecular weight of the first polymer should preferably be less than 50,000 Da to enable complete mixing with the second lactide monomer, which is essential for efficient block copolymerization. With the conventional catalyst of tin octanoate, diblock sb-PLA having different PLLA/PDLA block ratios and high molecular weight, reaching ca. 200 kDa, were obtained. The resultant sb-PLA were shown to exclusively undergo sc crystallization, regardless of its block ratio from 94/6 to 6/94 of PLLA/PDLA. Table 5.4 shows some properties of the injection molded parts prepared from diblock sb-PLA samples. Even at PLLA/PDLA ¼ 80/20, only sc crystals are formed, and the samples show high thermal resistivity. With the use of PLLA-rich ratios, cost reduction may also be effectively achieved by reducing the amount of the more expensive D-lactide. Therefore, this stepwise ROP is a promising method to commercialize sc-PLA materials.
  • 62 Pseudoorthorhombic a 103 1.034 0.597 — — — Kalb et al. [18] b The density obtained by experiment. The density obtained by calculation with the unit cell parameters. c The calculated data based on unit cell parameters. d The value reported in the paper. a Crystal structure Chain conformation a (nm) b (nm) c (nm) robs (g/cm3)a rcalc (g/cm3)b Pseudoorthorhombic a 103 1.07 0.645 2.78 — 1.248c De santis et al. [17] Pseudoorthorhombic a 103 1.06 0.61 2.88 1.275 1.285c Hoogsten et al. [6] Various Crystal Parameters Reported for the hc Crystals of PLLA Crystal system TABLE 5.2 Orthorhombic a 103 1.078 0.604 2.873 1.256 1.285d a 103 1.05 0.61 2.88 — 1.298c Miyata et al. [20, 21] Orthorhombic Kobayashi et al. [19] b 31 1.04 1.82 0.90 — 1.275c Orthorhombic Hoogsten et al. [6] b 31 1.052 1.052 0.88 — 1.278c Trigonal Puiggali et al. [8] b 31 1.04 1.77 0.90 1.27 1.301d Orthorhombic Sawai et al. [7]
  • FORMATION OF sb-PLA TABLE 5.3 Heats of Fusion (DHm in J/g) Reported for Various PLA Crystals (J/g) Fischer et al. [22] Jamshidi et al. [11] Loomis et. al. [23] Miyata et al. [20] Sawai et al. [5, 7] 93 — — 204 — — — — 142 135 — — 99a 124b 155a a Crystal b Crystal sc crystal a b The value reported in 2007. The value reported in 2002. FIGURE 5.1 Synthetic scheme of stereoregular PLA and multiblock-type sb-PLA from rac-lactides and stereoselective catalysts. FIGURE 5.2 5.5.2 Synthetic scheme of diblock-type isotactic sb-PLA by stepwise ROP. SSP Routes to Multiblock sb-PLA By extending the idea of diblock-type sb-PLA, multiblocktype sb-PLA (PLLA-multi-PDLA) consisting of relatively short enantiomeric segments were prepared to demonstrate TABLE 5.4 Properties of Injection Molds of Diblock sb-PLA Prepared by the Improved Lactide Method sb-PLA Tm ( C) HDTa ( C) Tenacity (MPa) Modulus (MPa) 20/80 PDLA/PLLA 30/70 PDLA/PLLA 50/50 PDLA/PLLA 214 214 216 180 197 198 64.4 2030 a 63 Heat deformation temperature (Vicat softening temperature). its exclusive stereocomplexation without hc crystallization occurring [34]. First, PLLA and PDLA prepolymers with medium molecular weight were prepared by melt polycondensation of L- and D-lactic acids, respectively [35]. Both prepolymers were then mixed to form the sc crystals and subjected to SSP to obtain the multiblock-type sb-PLA having a molecular weight around 100 kDa (Figure 5.3) [34, 36, 37]. Although the resultant sb-PLA preferentially formed the sc crystals, their molecular weight was not high enough for processing. The limited increase in molecular weight by the present SSP is attributed to the difficulty for the elongated chains to crystallize. In SSP, the chain coupling can be allowed because the polymer tails and catalysts are concentrated in the amorphous domain with the crystallization of the elongated
  • 64 STRUCTURE AND PROPERTIES OF STEREOCOMPLEX-TYPE POLY(LACTIC ACID) FIGURE 5.3 Synthetic scheme of multiblock-type isotactic sb-PLA by solid-state polycondensation. macromolecular chains [38–41]. In the SSP of the PLLA/ PDLA mixture, the chain coupling of PLLA and PDLA, however, forms a racemic unit that cannot be involved in the crystallization. Accordingly, little increase was attained in the molecular weight during the SSP. In our later study, it was shown that when the SSP is conducted at excess amount of enantiomeric compositions of PLLA or PDLA, high molecular weight sb-PLA can successfully be synthesized. The dehydrative coupling of abundant PLLA (or PDLA) molecules is followed by the crystallization of the polymer chains to increase the hc crystallinity by allowing further molecular coupling and molecular weight increase. The resultant PLLA-rich sbPLA, having high molecular weight up to 300 kDa, can be fabricated into polymer films by solution casting and used for analysis of their crystalline morphology and properties [42]. The sb-PLA with PLLA ratio from 85% to 15% can exclusively form sc crystals when they are melt quenched. Therefore, the stereoblock structure can suppress the hc crystallization and preferentially induce the sc formation even with the PLLA- or PDLA-rich compositions. Dynamic mechanical analysis of these sb-PLA films revealed that their thermal stability can be maintained up to 200 C. 5.6 APPLICATIONS OF sc-PLA The mechanical and thermal properties of sc-PLA are compared with those of other bio-based polymers in Table 5.5. The sc-PLA exhibits the highest level of thermal and mechanical properties even though complete sc crystallinity has not yet been achieved between high molecular weight enantiomeric polymers. We think that the total properties of sc-PLA are comparable to those of poly(butylene terephthalate) (PBT) used as an engineering plastic material. The application of sc-PLA products has just started to replace conventional oil-based polymers. It has been demonstrated that both the sc-PLA and the sb-PLA should have high potential as sustainable structural materials to make fibers, films, rods, and other processed plastics. Particularly, sc-PLA fibers show high thermal stability and can be TABLE 5.5 Mechanical and Thermal Properties of the Representative Bio-Based Polymers PLLA sc-PLA PGA  PHB Tm ( C) 170–190 220–240 225–230 188–197 Tg ( C) 50–65 65–72 40 5 DHm (J/g) 93–203 142–155 180–207 146 1.25–1.30 1.21–1.342 1.50–1.69 1.18–1.26 Density (g/cm3) 120–2260 880 80–980 180–200 Tensile strength (MPa) 6.9–9.8 8.6 3.9–14 4.9–5.9 Young’s modulus (GPa) Elongation 12–26 30 30–40 50–70 at break (%) PGA, poly(glycolic acid); PHB, poly(3-hydroxybutyrate).
  • REFERENCES subjected to dyeing and hot pressing by ordinary techniques. The alternative PLLA fibers are heat labile and damaged by thermosetting above 120 C. Teijin Co. (Japan) has already confirmed that a knit of sc-PLA withstands laundering above 180 C. Polymer films of sc-PLA are also applicable as heatresistive insulators with transparency. The intrinsic incombustible nature of PLA, as shown by its oxygen index of 28 [43], is also favorable for its application in housing of various electric appliances and automobile parts. In the near future, wide industrial applications of these materials will be possible [44], although at present Teijin Co. is the only producer of sc-PLA. REFERENCES 1. Ikada, Y. K. Jamshidi, H. Tsuji, S. H. Hyon, Macromolecules 1987, 20, 904. 2. Tsuji, H. S. H. Hyon, Y. Ikada, Macromolecules 1991, 24, 5651. 3. Tsuji, H. Y. Ikada, Macromolecules 1993, 26, 6918. 4. Tsuji, H. S. H. Hyon, Y. Ikada, Macromolecules 1992 25, 2940. 5. Sawai, D. Y. Tsugane, M. Tamada, T. Kanamoto, M. Sungil, S.-H. Hyon, J. Polym. Sci. Part B 2007, 45, 2632. 6. Hoogsten, W. A. R. Postema, A. J. Pennings, G. T. Brinke, P. Zuggenmier, Macromolecules 1990, 23, 634. 7. Sawai, D. K. Takahashi, T. Imamura, K. Nakamura, T. Kanamoto, S. H. Hyon, J. Polym. Sci., Polym. Phys. Ed. 2002, 40, 95. 8. Puiggali, J. Y. Ikada, H. Tsuji, L. Cartier, T. Okihara, B. Lotz, Polymer 2000, 41, 8921. 9. Fan, Y. H. Nishida, Y. Shirai, Y. Tokiwa, T. Endo, Polym. Degrad. Stab. 2004, 86, 197. 10. Tsuji, H. Y. Ikada, Macromol. Chem. Phys. 1996, 197, 3483. 11. Jamshidi, K. S. H. Hyon, Y. Ikada, Polymer 1988, 29, 2229. 12. de Jong, S. J. W. N. E. Van Dijk-Wolthuis, J. J. Kettenes-van den Bosch, P. J. W. Schuyl, W. E. Hennink, Macromolecules 1998, 31, 6397. 13. Tsuji, H. Y. Ikada, Polymer 1999, 40, 6699. 14. Okihara, T. M. Tsuji, A. Kawaguchi, K. Katayama, H. Tsuji, S. H. Hyon, Y. Ikada, J. Macromol. Sci. Phys. 1991, 30, 119. 15. Brizzolara, D. H.-J. Cantow, K. Diederichs, E. Keller, A. J. Domb, Macromolecules 1996, 29, 191. 16. Cartier, L. T. Okihara, B. Lotz, Macromolecules 1997, 30, 6313. 17. DeSantis, P. A. J. Kovacs, Biopolymers 1968, 6, 299. 65 18. Kalb, B. A. J. Pennings, Polymer 1980, 21, 607. 19. Kobayashi, J. T. Asahi, M. Ichikawa, A. Oikawa, H. Suzuki, T. Watanabe, E. Fukada, Y. Shikinami, J. Appl. Phys. 1995, 2957. 20. Miyata, T. T. Matsuko, Polymer 1998, 39, 5515. 21. Miyata, T. T. Matsuko, Polymer 1997, 38, 4003. 22. Fischer, E. W. H. J. Sterzel, G. Wegner, Kolloid ZZ Polym. 1973, 251, 980. 23. Loomis, G. L. J. R. Murdoc, K. H. Gardner, Polym. Prepr. 1990, 31, 55. 24. Yui, N. P. J. Dijkstra, J. Feijen, Makromol. Chem. 1990, 191, 481. 25. Spassky, N. C. Wisniewski, C. Pluta, A. Le Borgne, Macromol. Chem. Phys. 1996, 197, 2627. 26. Ovitt, T. M. G. W. Coates, J. Polym. Sci. Part A 2000, 38, 4686. 27. Coates, G. W. T. M. Ovitt, Patent Application, PCT US2001/ 027724, 2002 (Cornell Research Foundation, Inc., USA), 13 pp. 28. Ovitt, T. M. G. W. Coates, J. Am. Chem. Soc. 2002, 124, 1316. 29. Ishii, R. N. Nomura, T. Kondo, Polym. J. 2004, 36, 261. 30. Zhong, Z. P. J. Dijkstra, J. Feijen, J. Am. Chem. Soc. 2003, 125, 11291. 31. Tang, Z. X. Chen, X. Pang, Y. Yang, X. Zhang, X. Jing, Biomacromolecules 2004, 5, 965. 32. Majerska, K. A. Duda, J. Am. Chem. Soc. 2004, 126, 1316. 33. Komazawa, Y. Z. Tang, Patent Application, PCT/JP2007/ 065778, 2008 (Musashino Chemical Laboratory, Ltd., Japan; Mutual Corporation), p. 53 pp. 34. Fukushima, K. Y. Kimura, Macromol. Symp. 2005, 224, 133. 35. Fukushima, K. M. Hirata, Y. Kimura, Macromolecules 2007, 40, 3049. 36. Fukushima, K. Y. Kimura, Polym. Prepr. 2005, 46, 250. 37. Fukushima, K. Y. Kimura, Polym. Int. 2006, 55, 626. 38. Moon, S. I. C. W. Lee, M. Miyamoto, Y. Kimura, J. Polym. Sci. A 2000, 38, 1673. 39. Moon, S. I. C. W. Lee, M. Miyamoto, Y. Kimura, Polymer 2001, 42, 5059. 40. Moon, S. I. C. W. Lee, M. Miyamoto, Y. Kimura, High Perform. Polym. 2001, 13, S189. 41. Moon, S. I. Y. Kimura, J. Polym. Int. 2003, 52, 299. 42. Hirata, M. Y. Kimura, Polymer 2008, 49, 2656. 43. Li, S. A. El Ghzaoui, E. Dewinck, Macromol. Symp. 2005, 222, 23. 44. Tsuji, H. Y. Ikada, S. H. Hyon, Y. Kimura, T. Kitao, J. Appl. Polym. Sci. 1994, 51, 337.
  • PART II PROPERTIES OF POLY(LACTIC ACID)
  • 6 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) XUE JIANG, YAN LUO, XIUZHI TIAN, DAN HUANG, NARENDRA REDDY, 6.1 INTRODUCTION Commercial PLA is a blend of PLLA and PDLA or copolymer PDLLA, obtained by the polymerization of LLA and DLLA, respectively [1]. Many important properties of PLA are controlled by the ratio of D- to L-enantiomers used and the sequence of arrangement of the enantiomers in the polymers. PLLA constitutes the main fraction of PLA derived from renewable sources since the majority of lactic acid obtained from biological sources exists as LLA. PLA with PLLA content higher than 90% tends to be crystalline while that with lower optical purity is amorphous. The melting temperature (Tm), glass transition temperature (Tg), and crystallinity of PLA decrease with decreasing amounts of PLLA [2–5]. The properties of PLA such as thermal stability and impact resistance are inferior to those of conventional polymers used for thermoplastic applications. Therefore, PLA is not ideally suited to compete against the conventional polymers [5]. In order to improve the properties of PLA and increase its potential applications, copolymers of lactic acid and other monomers such as derivatives of styrene, acrylate, and poly (ethylene oxide) (PEO) have been developed. PLA has also been formulated and associated with nanosized fillers. Modification of PLA, copolymerization with other monomers, and PLA composites are some approaches that have been used to improve the properties of PLA, such as stiffness, permeability, crystallinity, and thermal stability [1–5]. Considerable research is being done to develop and study modified PLA, PLA-based copolymers, and PLA-based composites. The chemical structures of PLA, including the chain structure, configuration, tacticity, conformation, and the frustrated structure, are important parameters that govern AND YIQI YANG the properties of PLA and will be discussed in this chapter. Since the crystal structure of PLA depends on its configuration, the three forms of crystal PLA will also be reviewed. 6.2 6.2.1 CHAIN STRUCTURE AND CONFIGURATION Chain Structure The constitutional unit (i.e., chemical structure or molecular structure) of PLA is shown in Figure 6.1. PLA is a chiral polymer containing asymmetric carbon atoms with a helical conformation. Two optical isomers, L- and D-lactic acid, are used to synthesize PLA. Lactic acid can be produced either biologically or chemically [6]. The lactic acid that is produced by fermentation exists exclusively as L-lactic acid and condensation polymerization of L-lactic acid leads to PLLA with low molecular weights whereas the lactic acid produced by chemical processes could be obtained with various ratios of L- and D-lactic acid. Chemical reactions that form the cyclic dimer lactic acid as an intermediate step to the production of PLA could result in macromolecular chains with LLA and DLA monomer units [7]. The ring-opening polymerization (ROP) route has the advantage of producing high molecular weight polymers [8, 9] and also allows the control of PLA properties by adjusting the ratios and the sequence of L- and D-lactic acid units. Polymers that have stereocenters in the repeating unit can exhibit two structures of maximum order, that is, isotactic and syndiotactic. Isotactic polymers contain sequential stereocenters of same relative configuration while syndiotactic polymers contain sequential stereocenters of opposite relative configuration. These stereoregular polymers are typically crystalline and used in a wide variety of applica- Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 69
  • CHEMICAL STRUCTURE OF POLY(LACTIC ACID) Configuration As early as 1983, Francois et al. [10] reported their investigations on the configuration of PLA. They studied the 13 C-{1 H} NMR spectra of a series of PLA stereocopolymers prepared by three different synthetic methods: (1) ROP of L-, rac-, and meso-LAs initiated by powdered Zn; (2) copolymerization of L- and rac-LAs in different ratios; and (3) condensation polymerization of rac-lactic acid. Here, rac- is the abbreviation of raceme that represents the mixture of two enantiomers, and meso- is the abbreviation of mesomer that represents the compound of two chiral carbon atoms with different stereochemical configurations. High-resolution 500 MHz 1 H NMR spectroscopy is an effective tool for analyzing and characterizing PLA. It can accurately provide information about the distribution of a few stereosequences in the polymer whose resonances are well resolved in the NMR spectrum. 13 C NMR provides complementary stereosequence information; however, the accuracy of 13 C NMR is not comparable to that of 1 H NMR due to the poor signal-to-noise ratio. Thakur et al. [11] used 1 H and 13 C NMR to determine the structure of PLA. It was found that the splitting of the methine resonance caused by coupling to Me protons in the polymer was removed by homonuclear decoupling. By following the reversible polymerization of D,L-LA (rac-LA), it was shown that the stereospecificity (or reactivity ratio) for syndiotactic addition reduced as degree of polymerization increased. The increasing random polymerization is due to the interplay of kinetic and thermodynamic effects. Kinetic effects control the stereochemistry at the earlier stage of polymerization while equilibrium effects dominate the stereochemistry at the later stage. The change in stereosequence distribution during polymerization can also be monitored by 1 H NMR. The results showed that there was a preference for syndiotactic addition during LA stereocopolymerization. The viscosity changes during melt polymerization can also influence the mrm 6.2.2 rrr tions. Stereoregular polymers such as isotactic polypropylene share a significant fraction of the bulk polymer market. The degree and distribution of stereo-irregularities or stereodefects in these polymers may directly affect their final crystallinity. Usually, LAs have three different stereoisomers (LLA, DLA, and DLLA). Their corresponding polymers (PLLA, PDLA, and PDLLA) may have different chain configurations. The chain stereochemistry can in turn influence the physical and mechanical properties of PLAs. mrr/rrm FIGURE 6.1 Constitution of PLA, a chiral molecule. The carbon atom with an asterisk is a chiral atom. stereochemistry. Munson and coworkers [12] also used 13 C CP/MAS NMR to determine whether the PDLA containing 3% L-LA as the stereodefect was incorporated into crystalline regions or amorphous regions of the polymer. They got three main results. First, about 50% of the L-lactide is incorporated into the crystalline lattice of the polymer. Second, the environment of the crystalline defect sites is well defined. Third, the average chemical shift for the defect sites is upfield from the chemical shift of amorphous PLA, and the average chemical shift of the peaks for highly crystalline PLA is downfield from the chemical shift of amorphous PLA. The stereochemical microstructure of polymers can be controlled by the monomer composition in the feed or the stereochemical preference related to the initiating/ catalytic system [13–17]. Chamberlain et al. [13] prepared a series of zinc alkoxide complexes that act as single-site living initiators for the polymerization of DLA or LLA to isotactic PLA, rac-LA to atactic PLA, and meso-LA to syndiotactic PLA. The synthesis of chiral aluminum and yttrium alkoxides as stereospecific catalysts used for LA polymerization has also been reported by Ovitt and Coates [14]. The use of high-resolution 13 C and 1 H NMR for the study of PLA structures has also been reported [18–22]. The tacticity of the polymers can be determined by inspecting the homonuclear decoupled 1 H NMR spectrum of the methine region. Examples of 1 H and 13 C NMR spectra of PLA are shown in Figures 6.2 and 6.3, respectively [14]. The chemical shifts of typical tetrad peaks are listed in Table 6.1. rmr 70 5.24 5.20 5.16 1 5.12 FIGURE 6.2 Homonuclear decoupled H NMR spectrum of the methine region of PLA (500 MHz, CDCl3). Reproduced from Ref. 14 with permission from American Chemical Society. ppm
  • 71 69.5 69.4 69.3 69.2 69.1 69.0 rmr rmr/rrm mrm mrr/rrm rrr CHAIN STRUCTURE AND CONFIGURATION 68.9 68.8 ppm FIGURE 6.3 13 C NMR spectrum of the methine region of syndiotactic PLA (75 MHz, CDCl3). Reproduced from Ref. 14 with permission from American Chemical Society. An aluminum/Schiff base complex {[2,2-dimethyl-1,3propylenebis(3,5-di-tert-butylsalicylideneiminato)](isopropanolato)aluminum(III) (A)} based on a bulky ligand and aluminum isopropoxide was prepared and employed for the stereoselective ROP of rac-LA [17]. The initiating reaction and monomer polymerization followed first-order kinetics. There was a linear relationship between the rac-LA conversion and the number-average molecular weight of rac-PLA with narrow molecular weight distribution (1.04–1.08). The high melting temperature (196–201 C) is due to stereocomplex formation, and isotacticity of rac-PLA indicating that complex A was a highly stereoselective initiator for the ROP of rac-LA. The stereoselectivity was up to 90%. The configuration of the chiral centers in the polymer chain has great influence on crystallinity. The crystal and TABLE 6.1 Chemical Shifts of Tetrads in PLA [14, 18–21] Chemical Shift (ppm) H NMR 13 rrr 5.148 69.240 mrr/rrm 5.154 69.380, 69.020 mrm 5.142 69.144 rmr 5.214 68.902 Tetrads 1 C NMR amorphous contents of the polymer with chiral centers can be obtained by deconvolution of the 13 C CP/MAS NMR spectrum. For a polymer synthesized from 97% D-LA and 3% 13 C-labeled L-LA as a stereodefect, approximately 50% of the stereodefects are incorporated into the crystal region [18]. 6.2.3 Interlocked Structure, Polymer Blend, and Resistance to Hydrolysis Molecular modeling has been introduced by Karst and Yang [23] to explain the influence of PLLA and PDLA blends on the resistance to hydrolysis of PLA. They calculated the DU of PLLA/PDLA blends by molecular modeling, where DU was the change in internal energy for the hydrolysis of the PLA structure. A 50/50 PLLA/PDLA blend was found to have the least negative DU among the PLA blends used in this study and had the greatest resistance to hydrolysis whereas pure PLLA and PDLA exhibited the most negative DU, as shown in Figure 6.4. The difference of DU values between the 50/50 blend and pure PLLA and PDLA was about 37–38 kcal/mol, which indicated that pure PLLA and PDLA were easier to hydrolyze compared to the 50/50 blend. These results are in agreement with earlier experimental work by Tsuji [24]. The study by Karst and Yang [23] suggested that the internal energies (U) before hydrolysis of the PLA blends were relatively high (9116–9156 kcal/mol), which indicated that the polymer was very unstable, as shown in Figure 6.5. The U value before hydrolysis decreased with the increasing
  • 72 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) –40 –45 ΔU = 0.74(%PLLA) –83 R 2 = 1.00 –45 –46 ΔU (kcal/mol) –50 –53 –54 –55 –53 –55 –60 –65 –67 –68 –70 –68 –69 –75 ΔU = –0.73(%PLLA) – 10 R 2 = 1.00 –80 –83 –85 0 20 40 60 %PLLA L-Cleaved 80 –82 100 D-Cleaved FIGURE 6.4 Change in potential energy for hydrolysis (DU) of various blends of PLLA and PDLA versus % PLLA in the blend and whether a PLLA (L-cleaved) or PDLA (D-cleaved) was cleaved. Reproduced from Ref. 23 with permission from Elsevier. percentage of PLLA or PDLA in the blends from 0 to 50%. The change in potential energy for hydrolysis, DU, was found to decrease linearly (À83 to À45 kcal/mol) with the increasing percentage of PLLA or PDLA from 0 to 50%. The 50/50 blend had the greatest resistance to hydrolysis due to its stronger hydrogen bonding and dipole–dipole interactions compared to pure PLLA or PDLA. Hydrogen bonding was 9170 9155 9156 9150 9141 U (kcal/ mol) 9140 9130 9125 9126 9116 9110 9090 9073 9072 9071 9073 9073 9073 9072 9073 9070 9071 9070 9071 9072 9050 0 20 Uncleaved 40 60 % PLLA L -Cleaved 80 100 D -Cleaved FIGURE 6.5 Potential energies (U) of various blends of PLLA and PDLA before hydrolysis (uncleaved), after hydrolysis of one PLLA (L-cleaved), and after hydrolysis of one PDLA (D-cleaved) versus % PLLA in the blend. Reproduced from Ref. 23 with permission from Elsevier. assumed to have a greater effect than the dipole–dipole interactions on the resistance to hydrolysis [23]. Karst and Yang [25] have also used molecular modeling to determine the influence of percentages of LLA and DLA and their sequences in blocks or random arrangements in the polymer on the resistance of PLDLA to hydrolysis. It was found that PLA copolymers had less resistance to hydrolysis than PLLA/PDLA polyblends at the same percentages of LLA and DLA. It was found that the polyblends can form more stereocomplexes that are the most stable structures of PLA. Blends of PLA/PVOH have also been widely used in biotechnology and membrane separation technology. Sheetal and Tejraj [26] applied molecular modeling simulation to analyze the blend compatibility/incompatibility between PLA and PVA. The Flory–Huggins interaction parameter x was computed for different blends using atomistic simulations to predict blend miscibility. They observed miscibility in a 1:9 blend composition of PLA/PVOH, but immiscibility increased with higher PLA component. They also reported that the miscibility of PLA and PVA polymers was attributed to hydrogen bonding effects. 6.3 SYNDIOTACTIC POLYMERIZATION AND SYNDIOTACTICITY The physical and mechanical properties of a polymeric material critically depend on many factors, one of which is stereochemistry. Polymers that have chiral centers in the repeated unit can exhibit two structures of maximum order, isotactic and syndiotactic [27]. Sequential stereocenters of isotactic polymers are of same relative stereochemistry whereas those of syndiotactic polymers are of opposite relative configuration. Due to their stereoregularity, isotactic and syndiotactic polymers are typically crystalline, which is an important feature for many applications. Isotactic polymers are used in a wide range of applications. Typical examples include isotactic polyolefins and almost all natural polymers. In contrast, syndiotactic polymers have limited applications mainly due to their hard productivity and inherently alternating stereochemistry. The properties of syndiotactic polymers are usually similar to or in some cases better than isotactic counterparts according to the studies on syndiotactic polystyrene and other syndiotactic polyolefins [28]. Syndiotactic PLA is expected to be a versatile polymer with controllable stereochemistry. Syndiotactic PLA has been synthesized by syndiospecific polymerization using meso-LA. Three mechanisms are involved in the synthesis of syndiotactic polymers. One is the solid-phase polycondensation for A–B-type monomer with appropriate stereochemistry; for instance, syndiotactic polypeptides are produced by this approach. The second route to syndiotactic polymerization is the chain-end control
  • 73 SYNDIOTACTIC POLYMERIZATION AND SYNDIOTACTICITY FIGURE 6.6 Syndiotactic polymerization of meso-LA. Reproduced from Ref. 29 with permission from American Chemical Society. mechanism (CEM), where the last stereocenter of the growing polymer controls the stereochemical outcome during the monomer addition process. Despite many examples of CEM, this strategy has limited ability to stereocontrol and lacks generality. The third and latest strategy relates to the application of Cs-symmetric catalysts where regularly alternating monomer insertion on enantiotopic coordination sites results in syndiotactic polymers based on the site control mechanism (SCM) [29]. Ovitt and Coates [29] studied new strategies for syndiospecific polymerization and synthesized the first syndiotactic PLA from meso-LA, a cyclic monomer that contains two stereocenters. The polymerization mechanism of meso-LA is shown in Figure 6.6, where the meso-LA is marked as 1. In TABLE 6.2 theory, syndiotactic PLA can be polymerized with meso-LA by both the CEM and the enantiomorphic SCM. The catalytic systems for syndiotactic polymerization of meso-LA proposed and investigated in recent years are summarized in Table 6.2. Although the approach has only been applied successfully in olefin’s syndiotactic polymerization, the SCM route can be used in some other monomers’ syndiotactic polymerizations with a high degree of stereochemical integrity. Since the chain-end control mechanism cannot accurately explain stereochemistry in metal alkoxide initiating LA polymerizations [18, 19], many studies have focused on the initiators related to the site control mechanism [13, 14, 16, 17, 29–31] as shown in Table 6.2. The chainend control mechanism has also been reported [15]. Progress of the Catalyst Systems in Syndiotactic Polymerization with LA Catalytic Systems Properties Mechanisma Aluminum and yttrium metal alkoxides Trispyrazolylborate and trisindazolylborate, magnesium and zinc alkoxides Highly syndiotactic PLA, synthesized by the stereoselective ring-opening polymerization of meso-LA Exhibit only minor enantioselectivity in their polymerization of rac-LA but show significant diastereoselectivity to preferentially polymerize meso-LA by the mixture of racand meso-LA. meso-PLA shows a modest preference for syndiotactic junctions, RSRSRS (S,S)-LA to isotactic PLA, rac-LA to heterotactic PLA, and meso-LA to syndiotactic PLA meso-LA to syndiotactic PLA with 96% enantiotopic ring-opening selectivity Living polymerization of rac-LA for highly isotactic racPLA without any chiral auxiliaries in the catalyst. The stereoselectivity in this catalysis via CEM was similar to or higher than that of existing polymerization via SCM The resultant rac-PLA had a gradient stereocopolymer structure and exhibited reinforced thermal stability due to a stereocomplex formation (Tm ¼ 210 C) High stereoselectivity and well-controlled molecular weight for the ROP of rac-LA Controlled polymerization of rac-LA, providing monodisperse PLA with a higher degree of syndiotactic addition SCM [29] Chiral C3-symmetric, SCM [30] SCM [13] SCM [14] CEM [15] SCM [16] SCM [17] SCM [31] Zinc alkoxide complexes Chiral aluminum and yttrium alkoxides Substitute Schiff base/aluminum alkyl mixed in situ Chiral Schiff base/aluminum alkoxide Aluminum/Schiff base initiator [Ti(trisphenolate)O-i-Pr]2 initiators a Reference CEM: a route of syndiotactic polymerization that is based on the chain-end control mechanism, where the last stereocenter of the growing polymer dominates the stereochemical outcome during monomer addition process; SCM: a recent strategy related to application of Cs-symmetric catalysts, where regularly alternating monomer insertion on enantiotopic coordination sites forms syndiotactic polymers depending on the site control mechanism.
  • 74 6.4 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) CONFORMATION Due to the uncertain structures of PLA, it is difficult to investigate the structural changes induced by processing, for example, segmental orientation or crystallite size in PLA [32–34]. Aou and Hsu [35] applied infrared and Raman spectra to analyze the a-crystal of PLA. They also studied large two-dimensional PLA crystals using infrared dichroic analysis. By combining the two dichroic analyses, accurate band assignments of a-crystals were obtained for the first time. The analytical result of Aou and Hsu was consistent with a 103 helical chain conformation but was opposed to the more simplistic 31 helix. Infrared and Raman activities of the PLA 103 single-chain helix were derived and correlated to the orthorhombic a-crystal. The orthorhombic structure, previously suggested from X-ray studies, was confirmed by vibration analysis. The study of Aou and Hsu [35] confirmed that the 103 helix provided a better description of the PLA structure. The chain conformation and configuration of PLA play important roles in thermal properties of PLA. The effects of PLA chain conformation and configuration on the enthalpy relaxation kinetics of amorphous PLA were investigated by Aou et al. [36]. Conformational analysis using Raman spectroscopy revealed little or no change in helical tg q conformed population during the physical aging process. Configurational defects led to a lower Tg of PLA and thereby accelerated the enthalpy relaxation kinetics at a given annealing temperature. Epitaxial crystallization and crystalline polymorphism of PLAs have been studied by Cartier et al. [37]. A c-axis projection of the unit cell that is assumed as a regular 107 helix conformation is shown in Figure 6.7. This hypothesis FIGURE 6.7 Schematics of the a-phase of PLLA assuming regular 107 helix conformation and parallel helices. This model is an oversimplification of the actual crystal structure. Reproduced from Ref. 37 with permission from Elsevier. depicted an approximation of the actual structure but did not contain all the features of either the epitaxially crystallized film or single crystal patterns. The stable a-crystal modification of optically active polymer, based on a 103 helix conformation for PDLA (107 for PLLA), resulted in a Tc near 155 C. A new crystal modification was introduced by epitaxial crystallization at slightly lower Tc (<140 C). The structure of this new crystal was determined by electron diffraction and packing energy analysis [37]. The helix conformation of PLA crystal is discussed in Section 6.7.2. 6.5 AMORPHOUS STRUCTURE AND THERMAL PROPERTIES 6.5.1 Amorphous and Three-Phase Models PLA can be amorphous or semicrystalline depending on its stereochemical structure and thermal history. The crystallinity of PLA can be characterized by the weight fraction. The raw material of PLA is lactic acid, which is a chiral molecule and has two optically active isomers: L- and D-lactic acid. Polymerization of a racemic mixture of L- and D-lactic acid usually leads to the synthesis of PDLLA that is not crystalline but amorphous. PDLLA is an amorphous polymer due to the random distribution of L- and D-lactic acid units. Because of its amorphous nature, PDLLA shows a faster degradation rate than its stereoregular counterparts and becomes the preferred candidate for drug delivery vehicles and lowstrength scaffolding material for tissue regeneration [25, 38]. The amorphous structure of PLLA has been widely investigated. PLLA is semicrystalline. The crystallization of PLLA leads to various degrees of crystallinity. Semicrystalline PLLA consists of crystalline phases and amorphous phases, and tie molecules can appear in both phases. Parts of the amorphous phase of PLLA could behave differently property-wise because of the usual incomplete decoupling based on geometrical constraints [39]. A three-phase model has been proposed and the affected part of the amorphous phase was called rigid amorphous fraction (RAF) [39–41]. This fraction can be easily characterized by thermal analysis because it may not participate in the glass transition of the amorphous phase (the so-called mobile amorphous phase, MAP) [42]. Arnoult et al. [39] investigated the coupling between the RAF and the MAP of PLLA and compared PLLA with polyethylene terephthalate (PET). They suggested that the presence of the crystalline phase and RAF cannot significantly modify the fragility parameters of PLLA and hence the polymer remained “fragile.” Androsch and Wunderlich [43] found that in PET annealed to various extents, the amount of RAF can be considered as a measure of the coupling between the crystalline phase and the amorphous phase. The coupling between two phases is much weaker in PLLA than in PET. Wunderlich [44] proposed that
  • AMORPHOUS STRUCTURE AND THERMAL PROPERTIES 75 the RAF may not remain rigid until Tm and may decrease or disappear gradually above the Tg of MAP. 6.5.2 Glass Transition The glass transition of PLA has been widely investigated and reported. Pyda and Wunderlich [45–47] reported their studies on the glass transition of amorphous and semicrystalline PLA. The study was performed with adiabatic calorimetry, differential scanning calorimetry (DSC), and temperature-modulated DSC (TMDSC). The glass transition temperature from the reverse heat capacity and the enthalpy relaxation peaks from the nonreversing component of PLA both shifted to higher temperature with increasing extent of annealing. The relaxation times for aging decreased after cooling down to the glass transition temperature and then increased. The glass transition of the rigid amorphous phase of PLA has also been reported by Magon et al. [48]. The glass transition of the rigid amorphous phase was observed from changes of apparent heat capacity between the mobile glass transition temperature and the melting temperature. Because the heat capacity change of MAP at Tg depends on the fusion heat of semicrystalline PLA that has different thermal history, the RAF was selected to evaluate the nonlinearity of the change in heat capacity. The glass transition of semicrystalline PLLA with different morphologies was studied by Picciochi et al. [49]. A three-phase model was proposed and the relative thicknesses of the three phases, that is, crystalline, mobile amorphous, and rigid amorphous phases, were obtained as presented in Figure 6.8. The decrease in Tg was found to be correlated with the changes in the thickness of the rigid amorphous phase as FIGURE 6.9 Storage modulus (a) and loss factor (b) against temperature for PLLA fully crystallized at different crystallization temperatures (Tc), recorded at 1 Hz and during a heating scan at 3 C/ min. Reproduced from Ref. 49 with permission from Springer. FIGURE 6.8 Scheme of the lamellar morphology considering a three-phase model. The long period as detected by small-angle Xray scattering will combine the thicknesses of the crystalline layer, Lc, the mobile amorphous phase, Lma, and the rigid amorphous phase, Lra. Reproduced from Ref. 49 with permission from Springer. shown in Figure 6.9. In their study, the fully transformed semicrystalline specimens with different lamellar morphologies were produced and tested by dynamic mechanical analysis. With the increase in crystallization temperature, Tg decreased because of the corresponding increase in the thickness of the RAF according to the three-phase model. Dionisio et al. [50] also studied the glass transition dynamics of PLLA by real-time dielectric relaxation spectroscopy during the isothermal crystallization process. Dielectric loss in the frequency domain of PLLA during crystallization at 80 C is shown in Figure 6.10. The continuous shifting of the main loss peak that is the dielectric loss in the frequency domain of PLLA during the crystallization process lower frequencies was modeled considering the evolution of three relaxation processes: the a-process of the bulk-like (nonrestricted) amorphous phase, the a-process of the amorphous fraction influenced by the crystalline structure, and the b-relaxation (see Figure 6.10).
  • 76 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) FIGURE 6.10 Dielectric loss in the frequency domain of PLLA during crystallization process at 80 C (circles). The solid lines are the experimental fit to data obtained at different times with the sum of three Havriliak–Negami (HN) functions. Only the loss curves collected after every 10 min are shown for the first 3 h. The last three curves are collected at 4, 5, and 6 h, respectively. The inset shows the experimental results at 2 h (points) and the fitting lines, with the corresponding three HN individual curves (solid lines). Reproduced from Ref. 50 with permission from Wiley Blackwell. The glass transition of PLA is influenced by many factors, such as physical aging [51–53], crystallinity, morphology, and impurities [54–58]. Physical aging of PLA has been investigated in recent years [51–53]. The enthalpy relaxation behavior of PLLA below Tg and the effects of the enthalpy relaxation on the mechanical properties were studied using DSC by Pan et al. [51]. Effects of aging time and cooling rate on enthalpy relaxation were also investigated using DSC and the results are shown in Figures 6.11 and 6.12, respectively. DSC heating curves for PLLA samples annealed at 40 C (%Tg À 18 C) for various times are illustrated in Figure 6.11a. For PLLA, a shoulder was seen at the high-temperature side of the main endothermic peak when ta was between 3 and 24 h. However, this small shoulder was not detected in the aged PDLLA samples as shown in Figure 6.11b. Pan et al. [51] suggested that the crystalline phase was probably developed during the quenching process due to the existence of small amount of crystalline phase in the polymer. The kinetics of the enthalpy relaxation process were analyzed on the basis of DSC results shown in Figures 6.11 and 6.12, and the calculated relaxation rate of enthalpy at 40 C bH ¼ 1.77 J/g per decade, and apparent activation energy Dhà ¼ 1107 kJ/mol. In addition, Quan et al. [53] adopted TMDSC to investigate the glass transition behavior of PDLLA (a low optical activity PLA without crystallinity) and PLLA after physical aging at 273, 298, and 310K for up to 6 months. The results indicated that the glass transition behavior of PDLLA and LPLLA was sensitive to physical aging. Wang et al. [54] studied morphological contributions to glass transition in PLLA. The PLLA samples with different crystallinity were obtained during cooling from the melting state at different scanning rates. They were subjected to annealing below Tg. Two well-distinguished endothermic peaks were found when both intraspherulitic and interspheru- FIGURE 6.11 DSC curves on heating at 10 C/min in the glass transition region for (a) PLLA and (b) PDLLA samples annealed at 40 C for the time (in hours) indicated on each curve. Reproduced from Ref. 51 with permission from American Chemical Society.
  • ORIENTATION STRUCTURE OF PLA 77 and drying protocols and that the drying may markedly lower the Tg of the spheres. 6.6 ORIENTATION STRUCTURE OF PLA The properties of PLLA materials depend to a large extent on the morphology and volume fractions of the crystalline and amorphous regions. The degree of orientation also influences macroscopic properties of the material such as the mechanical strength and thermal recovery. 6.6.1 Mechanical Orientation by Stretching or Compression FIGURE 6.12 DSC heating scan at 10 C/min in the glass transition region for (a) PLLA and (b) PDLLA samples cooled at the various rates (in  C/min) indicated on each curve. Reproduced from Ref. 51 with permission from American Chemical Society. litic amorphous phases appeared corresponding to the glass transition of a bulk-like process (at relatively low temperatures) and a broad process assigned to the glass transition of a confined MAP. Since the temperature of glass transition process increased even in the early stages of crystallization, they attributed the glass transition of PLLA to the molecular motions of the interlamellar amorphous phase. Moreover, Wang et al. [55] reported that the Tg of PLLA, determined calorimetrically, exhibited a general decrease with an increase in crystallization temperature (Tc) for either cold or melt crystallized specimens. The three-phase model was appropriate for interpreting the structure of semicrystalline PLLA. Reignier et al. [57] investigated the effect of carbon dioxide on the glass transition and crystallization of PLLA. The effects of dissolved carbon dioxide (CO2) on the Tg as well as the crystallization kinetics of PLA have been studied. It was found that the Tg decreased nonlinearly as the CO2 concentration increased while the maximum crystallization rate increased with the addition of CO2. Bras et al. [58] investigated the effect of water on the thermal and molecular dynamic behavior of PLLA. They provided a detailed dielectric characterization of the relaxation modes based on a PLLA film containing 0.4 wt% of water. Dried PLLA exhibited an abnormally broad secondary b-relaxation that probably corresponded to the superimposition of multiple processes. The glass transition relaxation process was altered to high frequencies because of the plasticizing effect of water. The results showed that even small quantities of water may have a profound impact on the relaxation behavior of PLLA. As far as water and morphology were concerned, the effects of water on the glass transition temperature of PLA microspheres have been investigated by Passerini and Craig [59]. The results indicated that the microparticles may retain significant water levels according to standard preparation Wong et al. [60] investigated the orientation and structure development in PLA under uniaxial deformation. They also studied the relationships between molecular orientation and some influencing factors such as drawing temperature and stretch ratio. Higher orientation was obtained at higher drawing temperatures of semicrystalline PLLA due to the development of a highly oriented microfibrillar structure and increased crystallinity in the amorphous regions. At a low stretch ratio (l ¼ 1.5) and moderate deformation temperature (85 C), molecular orientation in the crystalline regions was found to progress more slowly than in the amorphous regions. An important consequence was that the orientation of the crystalline phase was always higher than that of the amorphous phase. The molecular orientation in the amorphous regions was significant at high draw ratios (l ¼ 4) while the molecular orientation in the crystalline regions leveled off. Ou and Cakmak [61] studied the influence of biaxial stretching on the crystalline texture in PLA films. Two different stretching modes, simultaneous and sequential biaxial stretching, were investigated. The results showed that simultaneous biaxial stretching always led to films with in-plane isotropy and poor crystalline order. In the first stage of sequential biaxial stretching, oriented crystallization gradually developed while transverse isotropy was maintained. Lamellar orientation in thin films of a model diblock copolymer with symmetric poly(styrene)-b-PLLA (PSPLLA) was investigated by Chen et al. [62] in the molten state on silicon wafer supported surfaces. Stretching and compression were apt to induce orientation of PLA. Pluta and Galeski [63] studied the plastic deformation of amorphous and thermally noncrystallizable 70/30 PLA/PDLLA induced by plane strain compression in a channel die. The results revealed that plastic deformation transformed an amorphous PLA or PDLLA (thermally noncrystallizable) into a crystalline fibrillar texture oriented in the flow direction. 6.6.2 Thermal Orientation (Phase Transition) Thermal properties of PLA have been thoroughly investigated and widely reported. The orientation of PLA induced
  • 78 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) by heating or cooling has also been studied. Zhang et al. [64] investigated the phase transition behavior of the PLLA/ PDLA blend by choosing a suitable annealing condition and using the highly oriented 1:1 stoichiometric blend of high molecular weight PLLA and medium molecular weight PDLA species. They succeeded in observing the phase transition during the heating and cooling processes. In their study, the as-drawn sample was found to contain two crystalline phases, disordered a-form (a0 ) and stereocomplex (bc). In the heating process, the a0 -form was reorganized into the oriented a-phase above 120 C and melted around 180 C. Some segments of the molten chains were found to crystallize on the surface of the bc crystallites with a high degree of chain orientation when the heating time was prolonged. In the subsequent cooling process from the bc phase, the unoriented a-crystal appeared at around 120 C and coexisted with the highly oriented bc crystals. 6.7 6.7.1 SEMICRYSTALLINE STRUCTURE General As discussed in the previous section, PLA can be chemically synthesized from LA either by condensation polymerization or by ROP. A cyclic dimer of lactic acid is a chiral molecule and has two optically active isomers: LLA and DLA. The polymerization of optically pure monomers leads to the formation of stereoregular PLLA or PDLA. The polymerization of rac-LA or meso-LA, however, results in the formation of amorphous PDLLA. Both stereoregular PLLA and PDLA are semicrystalline, with a Tm about 175 C and a Tg about 60 C. More attention has been given to the crystallization behavior of the PLLA homopolymer than that of PDLA because PLLA can be easily obtained from renewable resources. It was predicted that PLLA and PDLA should have similar crystallization and polymorphic behavior but their chirality of molecular chains in the crystal lattice was opposite. Semicrystalline PLA has high strength and modulus comparable to those of polypropylene and polystyrene [52]. The crystal structure of PLA has been investigated by several research groups using X-ray diffraction (XRD), electron diffraction, NMR, and other analytical techniques [36, 65–83]. 6.7.2 Three Forms (a, b, and c) of the Crystal Structure PLLA can crystallize into three forms (a, b, and c), generally referred to as polymorphism. Polymorphism in materials science refers to the existence of more than one form of crystalline structure in a solid material with the same chemical composition [84]. The crystal structure of the a-form has been investigated by several research groups using XRD and electron diffraction techniques [65–74]. The a-structure with a Tm of 185 C is more stable than the b-structure with a Tm of FIGURE 6.13 Crystal structure of a-form PLLA proposed by Sasaki and Asakura. Upper left: ac projection; upper right: bc projection; lower: ab projection. In the ac and bc projections, the molecular chains are enveloped with the van der Waals radii of the constituent atoms. Reproduced from Ref. 74 with permission from American Chemical Society. 175 C [39]. The optical purity of PLA has many profound effects on the structural, thermal, barrier, and mechanical properties of the polymer [85–89]. The a-form is characterized by two antiparallel chains in a left-handed 103 (or distorted 103) helix conformation packed in an orthorhombic (or pseudoorthorhombic) unit cell (Figure 6.13) [74]. The ratio of a and b axes is close to 31/2, indicating a nearly hexagonal packing of helices. The b-form of PLA has also been widely investigated [75–83] ever since Eling et al. [75] first detected the existence of the b-form upon hot drawing the melt-spun or solution-spun PLLA fibers under a high draw ratio (DR). The b-form crystals are generally prepared by stretching their a-counterparts at high temperature and high DR. The structural modes [67] and a nearly hexagonal packing of molecular chains [76] of the b-form crystals have been proposed. Puiggali et al. [77] suggested that the b-form rested on a frustrated packing of three 31 helix chains in a trigonal unit cell with a space group P32. This frustrated structure that will be discussed in the following section appears to be formed to accommodate the random orientation of neighboring chains associated with the rapid crystallization during stretching. As for the transition between a- and b-crystalline phases, Kanamoto and coworkers [81–83] found that a-to-b crystalline phase transition in PLLA proceeded more rapidly with
  • FRUSTRATED STRUCTURE an increase in the DR or extrusion pressure. The formation of the b-form crystals also depended on the temperature of extrusion or drawing. The a-to-b crystal transition proceeded efficiently at temperatures between 130 and 140 C. They suggested that the draw stress and crystal hardening had opposite effects on the extent of conversion of the a-to-b phase transition with increasing extrusion or drawing temperature [82]. Besides, the proportion of the formation of b-form increased as the molecular weight increased under the same processing conditions [83]. A new crystal modification, c-form, was produced by Cartier and coworkers [37] via epitaxial crystallization on a hexamethylbenzene (HMB) substrate. The structure of this crystal has been confirmed by electron diffraction and packing energy analysis and is shown in Figure 6.14 [37]. Two antiparallel helices were packed in an orthorhombic unit FIGURE 6.14 (a) Chain axis projection and (b) a-axis projection of the crystal structure determined for the c-phase of PLLA produced by epitaxial crystallization on HMB at 140 C. The structure corresponds to a minimum of the packing energy. The two antiparallel helices are linked by a 21 screw axis parallel to b. Reproduced from Ref. 37 with permission from Elsevier. 79 cell of parameters a ¼ 0.995 nm, b ¼ 0.625 nm, and c ¼ 0.880 nm. HMB appeared to be a versatile substrate, but the detailed mechanism of formation and thermal properties of the c-form crystals are yet to be determined. 6.8 FRUSTRATED STRUCTURE Cartier et al. [90] demonstrated that triangular or truncated triangular crystal morphologies were almost systematic for polymers with frustrated structures. These frustrated structures were based on trigonal unit cells that contained three threefold isochiral helices in which the (nearly) similar azimuthal setting of two helices differed distinctly from that of the third helix. The triangular growth morphology was a logical consequence of these different azimuthal settings that created different nucleation sites on opposite sides of any given growth plane. Cartier et al. indicated that, besides relatively straightforward cases of twinned crystals, triangular or truncated triangular polymer single crystals were obtained from trigonal unit cells in which opposite sides of a given growth plane were not equivalent. This nonequivalency might be a genuine crystallographic feature of the unit cell. Lotz and coworkers identified such asymmetries in the frustrated packing scheme and the isochiral helices were assembled in a unit cell with P31 or P32 symmetry. The asymmetry stems from the different azimuthal settings of the three helices that built up the unit cell and resulted in different nucleation sites on opposite sides of any given growth plane [90]. However, the evidence suggested that most, if not all, of the frustrated polymer crystals were bound by {100} growth faces that grew at a slower rate than the more densely packed {110} faces [37]. Experimental results and packing energy analyses showed that the specific mode of packing (i.e., cell symmetry) depended critically on the chain sense of the PLLA helices: antiparallel helices pack regularly in an orthorhombic cell in the epitaxially crystallized films [36]. Results from the packing energy analysis indicated that parallel helices would tend to pack in a one-chain trigonal unit cell. This unit cell, however, remained hypothetical and would be possible only for oligo-LLAs due to the existence of chain folds. Cartier et al. [91] suggested that the structure of PLLA formed on stroking or by stretching of a “standard” crystal structure (based on 103 or 107 helices). Moreover, Puiggali et al. [77] also demonstrated that PLA provided an exceptional opportunity to analyze the origin of frustration after two different crystal modifications that share the same helix conformation were characterized. They confirmed that the frustrated packing scheme was of the type NSS; here NSS was one of the frustrated schemes of the physical frustration associated with the helix setting. NSS means north–south–south frustrated scheme. Another packing was the NWW (north–west–west) frustrated scheme. The frustrated structure (trigonal unit cell,
  • 80 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) three helices per cell) tested by Cartier et al. appeared to be a compromise structure that was able to accommodate random orientation of chains imposed by rapid crystallization conditions (stroking or stretching of the a-phase). 6.9 MOLECULAR WEIGHT Molecular weight is very important for the degradation and other properties of PLA. High molecular weight PLA is usually prepared by ROP of LA, although there are several reports on preparation of high molecular weight PLA by polycondensation of lactic acid [92, 93]. Numata et al. [94] reported that the degradation rates of films of branched PLAs were found to be accelerated with an increase in the branch number as well as a decrease in the molecular weight of molecular branches. Numata et al. [95] evaluated the influence of the molecular weight, number of branches, and the stereochemical structures of PLA on the enzymatic hydrolysis rates by using real-time AFM and polymeric monolayers. The results showed that the rate of hydrolysis of the linear PLLA samples increased with decrease in the molecular weight. In contrast, the erosion rates of branched PLLA monolayers are independent of the molecular weight of the samples. The erosion rate of branched PLLA monolayers was found to depend on the average molecular weight of PLLA segments in branched molecules, but not on the overall molecular weight of the samples. The influence of molecular weight and other microstructures on the degradation of PLA has been investigated and widely reported [96–101]. Besides linear PLA, several new types of PLA such as cyclic PLA with well-defined molecular weights have also been designed. Jeong et al. [102] applied zwitterionic polymerization to generate cyclic PLA of defined molecular weight and molecular weight distribution and they also explored the kinetic features of these reactions. The zwitterionic ring-opening polymerization of LA, initiated by N-heterocyclic carbenes, generated cyclic PLAs with well-defined molecular weights (Mn) between 5000 and 30,000 g/mol with narrow polydispersities (Mw/Mn 1.31). The distinguishing features of zwitterionic polymerizations were that they did not only display extremely rapid polymerization rate with kp ¼ 48.7 MÀ1 sÀ1, but also exhibited exceptional control of molecular weight and molecular weight distribution. 6.10 important role in determining the properties of PLA. Further research is necessary to have a better understanding of some of PLA structures such as crystallization form, the frustrated structure model, and the structure of PLA copolymers or blends. SUMMARY The chemical structures of PLA, including chemical composition, interlocked structure, isomers, syndiotacticity, conformation, amorphous structure, semicrystal structure, frustrated structure, molecular weight, and so on, were reviewed in this chapter. The chemical structures play an REFERENCES 1. R. Auras, B. Harte, S. Selke, Macromol. Biosci. 2004, 4, 835–864. 2. H. Urayama, S. I. Moon, Y. Kimura, Macromol. Mater. Eng. 2003, 288, 137–143. 3. H. Tsuji, Y. Ikada, Macromol. Chem. Phys. 1996, 197, 3483–3499. 4. J. R. Dorgan, J. Jansen, M. P. Clayton, J. Rheol. 2005, 49, 607–619. 5. L. T. Lim, R. Auras, M. Rubino, Prog. Polym. Sci. 2008, 33, 820–852. 6. L. Averous, Polylactic acid: synthesis, properties and applications, in: N. Belgacem, A. Gandini (Eds.), Monomers, Oligomers, Polymers and Composites from Renewable Resources, Elsevier, Oxford, 2008, pp. 433–450. 7. P. Bordes, E. Pollet, L. Avrous, Prog. Polym. Sci. 2009, 34, e 125–155. 8. M. Okada, Prog. Polym. Sci. 2002, 27, 87–133. 9. A. C. Albertsson, I. K. Varma, Adv. Polym. Sci. 2002, 157, 1–40. 10. C. Francois, V. Michel, C. Stella, G. Pierre, Polymer 1983, 24, 53–59. 11. K. A. M. Thakur, R. T. Kean, E. S. Hall, J. J. Kolstad, E. J. Munson, Int. J. Polym. Anal. Charact. 1998, 4(5), 379–391. 12. M. T. Zell, B. E. Padden, A. J. Paterick, M. A. Hillmyer, R. T. Kean, K. A. M. Thakur, E. J. Munson, J. Am. Chem. Soc. 1998, 120(48), 12672–12673. 13. B. M. Chamberlain, M. Cheng, D. R. Moore, T. M. Ovitt, E. B. Lobkovsky, G. W. Coates, J. Am. Chem. Soc. 2001, 123(14), 3229–3238. 14. T. M. Ovitt, G. W. Coates, J. Am. Chem. Soc. 2002, 124(7), 1316–1326. 15. N. Nomura, R. Ishii, M. Akakura, K. Aoi, J. Am. Chem. Soc. 2002, 124(21), 5938–5939. 16. K. Majerska, A. Duda, J. Am. Chem. Soc. 2004, 126(4), 1026–1027. 17. Z. Tang, X. Chen, Y. Yang, X. Pang, J. Sun, X. Zhang, X. Jing, J. Polym. Sci. Part A 2004, 42(23), 5974–5982. 18. K. A. M. Thakur, R. T. Kean, E. S. Hall, J. J. Kolstad, T. A. Lindgren, M. A. Doscotch, J. I. Siepmann, E. J. Munson, Macromolecules 1997, 30(8), 2422–2428. 19. K. A. M. Thakur, R. T. Kean, E. S. Hall, J. J. Kolstad, E. J. Munson, Macromolecules 1998, 31(5), 1487–1494. 20. M. H. Chisholm, S. S. Iyer, D. G. McCollum, M. Pagel, U. Werner- Zwanziger, Macromolecules 1999, 32(4), 963–973.
  • REFERENCES 21. K. A. M. Thakur, R. T. Kean, M. T. Zell, B. E. Padden, E. J. Munson, Chem. Commun. 1998, 1913–1914. 22. E. J. Munson, L. K. Carlson, J. E. Jorvig, M. T. Zell, J. S. Abbott, M. A. Hillmyer, Determining stereodefect locations in polylactide using solid-state NMR spectroscopy, in: Abstracts of Papers, 225th ACS National Meeting, New Orleans, LA, 2003, POLY-685. 23. D. Karst, Y. Yang, Polymer 2006, 47(13), 4845–4850. 24. H. Tsuji, Polymer 2000, 41(10), 3621–3630. 25. D. Karst, Y. Yang, Macromol. Chem. Phys. 2008, 209(2), 168–174. 26. S. J. Sheetal, M. A. Tejraj, Polymer 2006, 47(23), 8061–8071. 27. M. Farina, Top. Stereochem. 1987, 17, 1–111. 28. N. Tomotsu, N. Ishihara, T. H. Newman, M. T. Malanga, J. Mol. Catal. A 1998, 128(1–3), 167–190. 29. T. M. Ovitt, G. W. Coates, J. Am. Chem. Soc. 1999, 121, 4072–4073. 30. M. H. Chisholm, N. W. Eilerts, J. C. Huffman, S. S. Iyer, M. Pacold, K. Phomphrai, J. Am. Chem. Soc. 2000, 122, 11845–11854. 31. S. K. Russell, C. L. Gamble, K. J. Gibbins, K. C. S. Juhl, W. S. MitchellIII, A. J. Tumas, G. E. Hofmeister, Macromolecules 2005, 38(24), 10336–10340. 32. P. B. Smith, A. Leugers, S. Kang, X. Yang, S. L. Hsu, Macromol. Symp. 2001, 175, 81–94. 33. P. B. Smith, A. Leugers, S. Kang, S. L. Hsu, X. Yang, J. Appl. Polym. Sci. 2001, 82, 2497–2505. 34. X. Yang, S. Kang, Y. Yang, K. Aou, S. L. Hsu, Polymer 2004, 45(12), 4241–4248. 35. K. Aou, S. L. Hsu, Macromolecules 2006, 39(9), 3337–3344. 36. K. Aou, S. L. Hsu, L. W. Kleiner, F.-W. Tang, J. Phys. Chem. B. 2007, 111(42), 12322–12327. 37. L. Cartier, T. Okihara, Y. Ikada, H. Tsuji, J. Puiggali, B. Lotz, Polymer 2000, 41(25), 8909–8919. 38. L. S. Nair, C. T. Laurencin, Prog. Polym. Sci. 2007, 32, 762–798. 39. M. Arnoult, E. Dargent, J. F. Mano, Polymer 2007, 48(4), 1012–1019. 40. H. Suzuki, J. Grebowicz, B. Wunderlich, Polym. J. 1985, 17, 1–3. 41. J. Menczel, B. Wunderlich, J. Polym. Sci. Polym. Lett. 1981, 1, 261–264. 42. M. Kattan, E. Dargent, J. Grenet, Polymer 2002, 43(4), 1399–1405. 43. R. Androsch, B. Wunderlich, Polymer 2005, 46(26), 12556–12566. 44. B. Wunderlich, Prog. Polym. Sci. 2003, 28(3), 383–450. 45. M. Pyda, B. Wunderlich, Reversing and nonreversing heat capacity of poly(lactic acid) in the glass transition region by temperature-modulated differential scanning calorimetry, in: Proceedings of the 31st NATAS Annual Conference on Thermal Analysis and Applications, 2003, pp. 143/ 1–143/9. 81 46. M. Pyda, E. Nowak-Pyda, B. Wunderlich, Physical aging of amorphous poly(lactic acid) by temperature-modulated calorimetry, in: Proceedings of the 33rd NATAS Annual Conference on Thermal Analysis and Applications, 2005, pp. 135.47.662/1–135.47.662/10. 47. M. Pyda, B. Wunderlich, Macromolecules 2005, 38(25), 10472–10479. 48. A. Magon, E. Nowak-Pyda, R. C. Bopp, M. Pyda, Glass transition of the rigid amorphous phase of biodegradable poly(lactic acid), in: Proceedings of the 35th NATAS Annual Conference on Thermal Analysis and Applications, 2007. 49. R. Picciochi, Y. Wang, N. M. Alves, J. F. Mano, Colloid Polym. Sci. 2007, 285(5), 575–580. 50. M. Dionisio, M. T. Viciosa, Y. Wang, J. F. Mano, Macromol. Rapid Commun. 2005, 26(17), 1423–1427. 51. P. J. Pan, B. Zhu, Y. Inoue, Macromolecules 2007, 40(26), 9664–9671. 52. M. Pyda, E. Nowak-Pyda, B. Wunderlich, Physical aging of amorphous poly(lactic acid) by temperature-modulated calorimetry, in: Proceedings of the 33rd NATAS Annual Conference on Thermal Analysis and Applications, 2005, pp. 135.47.662/1–135.47.662/10. 53. D. Quan, K. Liao, J. Zhao, Gaofenzi Xuebao 2004, 5, 726–730. 54. Y. Wang, J. L. G. Ribelles, M. S. Snchez, J. F. Mano, a Macromolecules 2005, 38(11), 4712–4718. 55. Y. Wang, S. S. Funari, J. F. Mano, Macromol. Chem. Phys. 2006, 207(14), 1262–1271. 56. J. F. Mano, J. L. G. Ribelles, N. M. Alves, S. M. Salmeron, Polymer 2005, 46(19), 8258–8265. 57. J. Reignier, J. Tatibouet, R. Gendron, J. Appl. Polym. Sci. 2009, 112(3), 1345–1355. 58. A. R. Bras, M. T. Viciosa, M. Dionisio, J. F. Mano, J. Therm. Anal. Calorim. 2007, 88(2), 425–429. 59. N. Passerini, D. Q. M. Craig, J. Control. Release 2001, 73(1), 111–115. 60. Y. S. Wong, Z. H. Stachurski, S. S. Venkatraman, Acta Mater. 2008, 56(18), 5083–5090. 61. X. Ou, M. Cakmak, Polymer 2008, 49(24), 5344–5352. 62. D. Chen, Y. Gong, H. Huang, T. He, F. Zhang, Macromolecules 2007, 40(18), 6631–6637. 63. M. Pluta, A. Galeski, Biomacromolecules 2007, 8(6), 1836–1843. 64. J. Zhang, K. Tashiro, H. Tsuji, A. J. Domb, Macromolecules 2007, 40(4), 1049–1054. 65. P. DeSantis, A. J. Kovacs, Biopolymers 1968, 6, 299–306. 66. B. Kalb, A. J. Pennings, Polymer 1980, 21, 607–612. 67. W. Hoogsteen, A. R. Postema, A. J. Pennings, G. ten Brinke, Macromolecules 1990, 23(2), 634–642. 68. C. Marega, A. Marigo, V. DiNoto, R. Zannetti, A. Martorana, G. Paganetto, Makromol. Chem. 1992, 193, 1599–1606. 69. J. Kobayashi, T. Asahi, M. Ichiki, A. Oikawa, H. Suzuki, T. Watanabe, E. Fukada, Y. Shikinami, J. Appl. Phys. 1995, 77, 2957–2973.
  • 82 CHEMICAL STRUCTURE OF POLY(LACTIC ACID) 70. D. Brizzolara, H. J. Cantow, K. Diederichs, E. Keller, A. J. Domb, Macromolecules 1996, 29(1), 191–197. 71. T. Miyata, T. Masuko, Polymer 1997, 38(16), 4003–4009. 72. T. Iwata, Y. Doi, Macromolecules 1998, 31(8), 2461–2467. 73. C. Alemn, B. Lotz, J. Puiggali, Macromolecules 2001, 34 a (14), 4795–4801. 74. S. Sasaki, T. Asakura, Macromolecules 2003, 36(22), 8385–8390. 75. B. Eling, S. Gogolewski, A. J. Pennings, Polymer 1982, 23 (11), 1587–1593. 76. D. Sawai, K. Takahashi, T. Imamura, K. Nakamura, T. Kanamoto, S. H. Hyon, J. Polym. Sci. Polym. Phys. 2002, 40, 95–104. 77. J. Puiggali, Y. Ikada, H. Tsuji, L. Cartier, T. Okihara, B. Lotz, Polymer 2000, 41(25), 8921–8930. 78. D. A. Brant, A. E. Tonelli, P. J. Flory, Macromolecules 1969, 2 (3), 228–235. 79. Y. Chatani, Y. Okita, H. Tadokoro, Y. Yamashita, Polym. J. 1970, 1, 555–562. 80. J. W. Leenslag, A. J. Pennings, Polymer 1987, 28(10), 1695–1702. 81. D. Sawai, K. Takahashi, A. Sasashige, T. Kanamoto, S. H. Hyon, Macromolecules 2003, 36(10), 3601–3605. 82. K. Takahashi, D. Sawai, T. Yokoyama, T. Kanamoto, S. H. Hyon, Polymer 2004, 45(14), 4969–4976. 83. D. Sawai, T. Yokoyama, T. Kanamoto, M. Sungil, S. H. Hyon, L. P. Myasnikova, Macromol. Symp. 2006, 242, 93–103. 84. P. Pan, Y. Inoue, Prog. Polym. Sci. 2009, 34, 605–640. 85. H. Tsuji, R. Okino, H. Daimon, K. Fujie, J. Appl. Polym. Sci. 2005, 99, 2245–2252. 86. J. R. Sarasua, A. L. Arraiza, P. Balerdi, I. Maiza, Polym. Eng. Sci. 2005, 45, 745–753. 87. D. M. Bigg, Effect of copolymer ratio on the crystallinity and properties of polylactic acid copolymers, in: Annual Technical 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. Conference of the Society of Plastics Engineers, 1996, pp. 2028–2039. M. Hutchinson, J. Dorgan, D. Knauss, S. Hait, J. Polym. Environ. 2006, 14, 119–124. M. Pyda, R. C. Bopp, B. Wunderlich, J. Chem. Thermodyn. 2004, 36, 731–742. L. Cartier, T. Okihara, B. Lotz, Macromolecules 1997, 30(20), 6313–6322. L. Cartier, N. Spassky, B. Lotz, Macromolecules 1998, 31(9), 3040–3048. S. I. Moon, C. W. Lee, M. Miyamoto, Y. Kimura, J. Polym. Sci. Polym. Chem. 2000, 38, 1673–1679. S. I. Moon, C. W. Lee, I. Taniguchi, M. Miyamoto, Y. Kimura, Polymer 2001, 42(11), 5059–5062. K. Numata, R. K. Srivastava, A. Finne-Wistrand, A. C. Albertsson, Y. Doi, H. Abe, Biomacromolecules 2007, 8, 3115–3125. K. Numata, A. Finne-Wistrand, A. C. Albertsson, Y. Doi, H. Abe, Biomacromolecules 2008, 9(8), 2180–2185. J. F. Bridges, M. Critchlow, M. P. Irving, S. C. Purkiss, D. C. Taylor, J. B. Lloyd, Biomaterials 2000, 21(2), 199–209. T. Schlosser, W. G. Suess, Pharmazie 2005, 60(7), 514–517. D. V. Plackett, V. K. Holm, P. Johansen, S. Ndoni, P. V. Nielsen, T. Sipilainen-Malm, A. Sodergard, S. Verstichel, Packag. Technol. Sci. 2006, 19(1), 1–24. C. L. Wanamaker, W. B. Tolman, M. A. Hillmyer, Biomacromolecules 2009, 10(2), 443–448. C. M. Dong, Y. Z. Guo, K. Y, Qiu, Z. W. Gu, X. D. Feng, J. Control. Release 2005, 107(1), 53–64. S. Gogolewski, P. Mainil-Varlet, Biomaterials 1997, 18(3), 251–255. W. Jeong, E. J. Shin, D. A. Culkin, J. L. Hedrick, R. M. Waymouth, J. Am. Chem. Soc. 2009, 131(13), 4884–4891.
  • 7 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID): A PRACTICAL FRAMEWORK USING HANSEN SOLUBILITY PARAMETERS STEVEN ABBOTT 7.1 A PRACTICAL FRAMEWORK In an ideal world, a chapter such as this would provide the reader with the ability to know how any chemical will interact with PLA. There are two possible ways of doing this. The first is to provide an overwhelming amount of detail over a large range of chemicals, allowing the experienced reader to extrapolate to their own area of interest. For a comparatively young polymer such as PLA, this is simply not possible. The second approach is to provide a perfect theoretical framework from which the reader can calculate whatever they require. No such framework exists. The approach adopted instead is closer to the second than the first. It starts from a rather simple thermodynamic approach and shows that over a broad range of chemicals and applications this approach yields practical insights that are frequently reliable or, at the very least, point the practical scientist in the right direction. Such frameworks are worthless without experimental validation, so the relatively sparse real-world data are used to check the framework. 7.1.1 Thermodynamics Versus Kinetics If one wants a framework for understanding chemical interactions, thermodynamics has a lot going for it: it is a fundamental grounding for science; it is often amenable to calculations; and it has separable terms (enthalpy and entropy) that are deeply insightful into many phenomena. On the other hand, we all know that kinetic effects can dominate results in the short timescale of experiments. This is partic- ularly true for PLA with its complex repertoire of crystalline effects, which can comprehensively skew results away from the thermodynamically favored route. Of course, crystallinity is an aspect of thermodynamics too. Unfortunately, there does not exist a robust methodology for including crystallinity within any practical framework known to the author. 7.1.2 Hansen Solubility Parameters Ideally, all thermodynamic approaches should lead to the same conclusion, so it should not matter too much which is adopted. But within a practical framework, the choice has to be something that is not only thermodynamically justified but also easy to use, and applicable, over a wide range of topics. One possible approach is UNIFAC (universal functional activity coefficient) and its many variants, particularly UNIFAP (UNIFAC for polymers) [1]. This is a classic group contribution methodology that relies on fitting of large data sets of various properties. In areas where these large data sets exist, the methods are most fruitful. But despite the enthusiasm of the proponents of these methods, there is not a reliable body of results applicable to the sorts of questions of interest to this chapter. With large databases of properties, it is possible to apply QSAR (quantitative structure–activity relationship) techniques to derive good fits between molecular properties (lists exist of 1000 þ properties available from ab initio calculations that can then be tested for fitability [2]) and the desired macroproperty. SPARC (SPARC Performs Automated Reasoning in Chemistry) is an excellent example of this approach Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 83
  • 84 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID) and various methods exist for calculating (Kovts) retention a indices, which are particularly relevant for transmission of odors/flavors through packaging films [3]. The advantage of a QSAR is that within the fitting regime one gets good results (by definition). The disadvantage is that it is hard to use QSAR insights from one domain and apply them to another. The Flory–Huggins x parameters [4] have a deserved reputation as being deeply insightful into the thermodynamic complexities of solvent and polymer systems. If one knows all the x parameters of a system (in particular, the x12 interaction parameters), many of the properties of interest can be calculated. Unfortunately, no good way currently exists for calculating the relevant parameters required for the wide range of practical issues for which we want the insights. Hansen solubility parameters (HSP; note that the acronym is a plural) seem to strike a happy balance between exactness and practicality [5]. At their heart, they are the same as the Flory x parameters—the sum of the HSP terms is calculable from x. They both derive their thermodynamic credentials from the free energies of vaporization of solvents, which in turn govern the thermodynamic solubilities of polymers and liquids. What gives HSP their practical edge is that by splitting the parameters into just three terms, dispersion, polar, and hydrogen bonding (dD, dP, dH), a very large array of problems becomes readily calculable. The sum of the three parameters is the same as the Hildebrand solubility parameter that one frequently finds referred to in the literature. However, there is a very simple illustration of how the Hildebrand parameter is of little use for solubility studies. The Hildebrand parameters of ethylene carbonate and methanol are identical (29) but their solvencies are different. From their Hansen parameters (18, 21.7, 5.1) and (14.7, 12.3, 22.3), it is evident that they are very different, with a very large “distance” (see below) of 20 between them. But of course to use HSP one has to know the three values for each polymer and each solvent. In recent years, there have been five key developments that have made it easier to know these values. First, an extensive list of HSP values for $1200 chemicals and $500 polymers has been published [5]. Second, in addition to the well-known Hoy and Van Krevelen methods for calculating HSP, the powerful new Stefanis– Panayiotou method (which, incidentally, uses UNIFAC group definitions) has become available [6]. Third, with modern PC power, the so-called sphere method [5] for measuring HSP of polymers, fillers, pigments, and nanoparticles has become routine. Fourth, a great deal of progress has been made in calculating HSP from first principles using molecular dynamics [7]. Fifth, a fully automated neural network method for estimating HSP from SMILES (simplified molecular input line entry specification) or Molfile inputs has become available [8]. HSP encapsulate the well-known principle that “like dissolves like.” The question is how does one measure “likeness”? The answer is that if one knows the HSP of, say, a solvent (s) and polymer (p), then the greater the “distance” between them, the less alike they are. The “distance” is the geometric distance in 3D space: 1 =2 distance ¼ ½4ðdDs ÀdDp Þ2 þ ðdPs ÀdPp Þ2 þ ðdHs ÀdHp Þ2 Š When the distance ¼ 0, the solvent and polymer are so thermodynamically similar that the polymer will be highly soluble. As the distance gets greater, the solubility declines. Above a certain distance (the so-called polymer radius), the solubility is negligible. It would be ideal if the radius could be calculated from first principles, but in practice it is something that is measured. It is, in general, smaller (i.e., you need greater thermodynamic compatibility for good dissolution) for higher molecular weight and highly crystalline polymers. The famous factor of 4 in the distance equation had once been seen as a fudge factor required for a convenient fit to the real-world data. The fact that it turned out to have a deep theoretical underpinning was something of a vindication for the theory [5]. All HSP diagrams used in this chapter are based on the HSPiPÔ software developed by Abbott and Hansen [8]. It has to be stressed that the same results can be obtained by anyone with modest programming skills using the methodology fully described by Hansen [5] along with the large number of public domain HSP. 7.2 SOLVENT COMPATIBILITY What does PLA dissolve in? This question is answered in HSP terms in Figure 7.1. At the top left is a list of solvents with their HSP followed by a “1” to indicate that PLA is “soluble.” The word “soluble” is in quotes because the definition can be made as tight or as loose as desired. A tight definition means that the polymer dissolves perfectly. A loose definition means that the polymer swells significantly. Because of the complexities of PLA, it is likely that a looser definition is more appropriate. The key is that the test should be fit for the purpose in hand. If genuine solubility is the main interest, then the tight definition is appropriate. If general compatibility is the main interest (as it is in this chapter), then the looser definition is used. Going down the table (not shown in the figure), one reaches solvents such as ether, hexane, and ethanol that score “0”. The sphere method attempts to find a sphere in HSP space that includes all the good solvents and excludes all the bad ones. In this case, there is a perfect sphere. Twenty good solvents are inside it (dark symbols) and seven bad ones outside (light symbols). The calculation says that the HSP for PLA are (18.6, 9.9, 6.0) and that the radius is a fairly relaxed 10.7, which means that quite a wide range of solvents will dissolve or at least swell PLA. In practice, solvents close to the center of the
  • SOLVENT COMPATIBILITY FIGURE 7.1 85 Finding the center and radius of the PLA solubility sphere. sphere are likely to fully dissolve the polymer while those near the edges will merely swell it. The radius at which swelling predominates depends on the molecular weight and the L/D ratios of the PLA. It is important to note that both the center of the sphere and the radius need to be known in order to understand polymer behavior. The center defines the basic chemical compatibility of the polymer, and the radius defines how specific the compatibility is. A small radius means that solvents (and plasticizers, additives, polymers) must be very close in order to be compatible. A large radius means that a much wider range of materials are compatible. To find a good solvent for a polymer, the simplest thing to do is calculate the distances of a list of well-known solvents and order them by their RED (relative energy difference), which is simply the distance divided by radius, where a value of 1 means that the solvent is right on the soluble/insoluble border. Table 7.1 shows a partial list of solvents that I regularly consider. Having such an ordered list of solvents is immensely powerful when one wants to do something practical with PLA. Sometimes very good solubility is desirable (e.g., cyclohexanone, the best in this list), but other times, one may want borderline solubility, for example, for controlled phase separation, as in the following portion of the list (Table 7.2), and sometimes insolubility is required. The “molar volume” column is an important part of the HSP output. In general, smaller molecules (in particular, smaller, unbranched, linear molecules) are far faster at getting into polymers. So for applications where good and fast dissolution of PLA is desirable, 1,3-dioxolane or MEK would be good starting points as their molar volumes are smaller than, say, benzyl benzoate. TABLE 7.1 to PLA Solvents Listed by Their RED Number Relative Solvent dD dP dH RED Cyclohexanone NMP Isophorone Butyl benzoate 1,3-Dioxolane Nitropropane MEK Dimethyl acetamide Benzyl benzoate Caprolactone THF Tributyl phosphate Acetone DMF 17.8 18 17.0 18.3 18.1 16.6 16 16.8 20.0 19.7 16.8 16.3 15.5 17.4 8.4 12.3 8.0 5.6 6.6 12.3 9 11.5 5.1 15 5.7 6.3 10.4 13.7 5.1 7.2 5.0 5.5 9.3 5.5 5.1 10.2 5.2 7.4 8 4.3 7.0 11.3 0.23 0.27 0.36 0.41 0.43 0.44 0.50 0.53 0.53 0.53 0.54 0.57 0.59 0.64 Already it is clear that HSP are providing a wealth of information. Even more powerful is the fact that the HSP of a mixture of solvents is simply the weighted average of their HSP. This means that it is easy to create optimized solvent blends where “optimized” can mean solubility, safety, environmental friendliness, and so on. Suppose, for example, that one wanted a borderline solvent (e.g., for membrane precipitation) based on 2-propanol, which is a nonsolvent. A glance at the HSP of 2propanol (15.8, 6.1, 16.4) compared to the target of (18.6, 9.9, 6.0) immediately suggests that a solvent with a higher dD and lower dH is needed to balance the properties of 2-propanol. From the list of solvents, butyl benzoate does not look a bad
  • 86 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID) TABLE 7.2 Solvents on the Borderline of PLA Solubility dD dP dH RED Molar Volume 5.7 4.5 1 1.4 4.1 1.0 1.03 18.2 18.0 5.7 5.7 5.1 21.7 6.1 10.6 14.3 8.2 3 2.0 13.5 3.1 2.05 10.9 6.1 14.5 15.8 14.7 5.1 16.4 17.7 0.87 0.88 0.89 0.89 0.90 0.90 0.93 0.95 0.98 1.02 1.09 1.11 1.11 1.15 1.15 124.7 117.1 100 106.6 105.7 123.9 157.3 96 52.9 92 92 96 66 76.9 138.5 Solvent Phenoxyethanol Isopropyl acetate Aromatic hydrocarbons Toluene Cyclohexanol Xylene D-Limonene Sulfolane Acetonitrile 2-Butanol 1-Butanol t-Butanol Ethylene carbonate 2-Propanol DPG 17.8 14.9 18 18.0 17.4 17.6 17.66 20.3 15.3 15.8 16.0 15.2 18.0 15.8 16.5 match and indeed an 85:15 mix of 2-propanol and butyl benzoate has a RED number of 1, right on the borderline. A good check for the theory is van de Witte’s chloroform/ methanol mixes used for his extensive work on membranes [9a–f]. The RED number is calculated to be 1 for a 40:60 ratio, not far from the 55:45 ratio typical of the mixes at cloud point (which corresponds to a radius of 8.6). Of course, the issue of crystallinity plays a big role in van de Witte’s work. As mentioned earlier, the radius of crystalline polymers is smaller and the cloud points shift to a 65:35 ratio, corresponding to a radius of 7.4. HSP can also extend van de Witte’s work in a rational manner. Suppose one wanted to follow his chloroform/methanol work while using “chlorine-free chemistry” for environmental purposes. There is no solvent that is a close match for chloroform, but a 50:50 mix of toluene/THF is a very good match. If this mixture were used as the “good” solvent, then a 55:45 ratio of this mix to methanol would be a good starting point. 7.3 PLASTICIZERS The need for a good plasticizer for PLA is well known and an excellent review discusses the numerous attempts to come up with an optimal solution [10]. One problem, in the context of this chapter, is knowing what “optimal” means at the chemical level. Chemical compatibility with PLA would seem highly desirable to reduce the tendency for the plasticizer to bleed out. Yet, there is plenty of evidence that toughness relies on discrete domains within the PLA that can absorb the stress in impact tests or tensile testing. Perhaps the ideal is both—strongly compatible elements that reduce bleed out and a strong interface to the incompatible elements. Or maybe we define a plasticizer as being soluble/miscible and ask for improvements via domains to come from polymer blends, as discussed below. It is therefore important to have a good idea about inherent compatibility before throwing in yet another molecule in the hope that it will work—because the strong evidence is that most attempted plasticizers do not give satisfactory results [10]. In addition, the ideal plasticizer for PLA should be a lowvolatility liquid, derived from green sources and biodegradable. The lactide monomer is obviously highly compatible but is too small a molecule, so it is easily lost from the system [11, 12]. Oligomeric lactic acid is also highly compatible, but the reduction in modulus is unacceptable. Awellknown green plasticizer is acetyl triethyl citrate. With HSP of (16.6, 3.5, 8.6), its distance from PLA is an unfortunately high 8. The other common citrates are no better; the popular acetyl tributyl citrate has HSP of (16.7, 2.5, 7.4) and a distance of 8.4. It is no surprise, therefore, to find that a big problem with the citrates is their tendency to bleed out of the PLA. Similarly, triacetin (16.5, 4.5, 9.1) has a distance of 7.4. The common adipate plasticizers are also poorly compatible. Because PLA has a high dD value, the obvious place to look for a rational plasticizer is in the aromatics. Dipropylene glycol dibenzoate (18.0, 6.6, 5.6) is only 3.5 away from PLA and would be predicted to be a satisfactory plasticizer. However, the nongreen credentials of these benzoate plasticizers seem to have excluded them, along with phthalates, from consideration within the PLA marketplace. The dipropylene glycol core of the benzoate plasticizer takes us to propylene glycol (PPG) oligomers [13] (as well as ethylene glycol oligomers, PEG), which have a distance $4 and would be predicted to be reasonably compatible. It is known that PEG tends to crystallize with time, so it is arguably an unsuitable plasticizer whereas PPG does not crystallize. It is rather difficult to disentangle the various elements in the effect of PPG on PLA because it is not clear whether enhanced spherulite formation (thanks to the plasticization effect) is desirable or not; it is undesirable because it leads to local “pools” of the plasticizer, excluded from the crystals, but it is desirable because the pools act as stress absorbers and enhance elongation to break [14]. The HSP of epoxidized linseed or soybean oils are not in the literature, so it is not possible to put numbers on their inherent suitability [15]. A patent on a plasticizer that seems to give the right thermomechanical properties without significant bleeding indicates a good rational design principle: make an oligomer based on PLA [16]. The patent shows an OLA (oligomeric lactic acid)-capped, short-chain polyethylene glycol (PEG). As noted above, PEG is also a reasonable plasticizer. An estimate of the HSP based on a 50:50 PLA/PEG ratio is (17.6, 8.6, 7.9) giving a distance of 3. The motif of glycol/esters as plasticizers is a natural one. A study done using poly(1,3butylene glycol adipate) (distance $3.5) is worth noting [17]. Unfortunately, being a fairly rigid polymer it is not really a
  • POLYMER COMPATIBILITY on chemical compatibility and only so much can be demanded of thermodynamics. practical plasticizer, so it had to be blended with triacetin. Although the mix seems to give good plasticization, the study does not report whether the triacetin bleeds out. The authors of the paper attempt a solubility parameter approach but unfortunately use the highly misleading Hildebrand total (see Section 7.1.2) solubility parameter that suggested that triacetin was a good match to PLA. Another study reported the PLA plasticization effect of oligomers of tributyl citrate with diethylene glycol produced via transesterification [18]. The HSP of these are expected to be similar to the other citrates, so it is no surprise to find that the bleeding of these plasticizers is still rather unsatisfactory. The hyperbranched polymer used in Ref. 19 had little beneficial effect on the tensile performance. The HSP of this polymer are not known but are likely to be a significant distance from PLA. But by in situ anhydride reactions (which, incidentally, would be expected to reduce the incompatibility with PLA), the hyperbranched polymer becomes cross-linked into a matrix, which gives big increases in toughness and elongation at break. It is interesting that the simple HSP approach seems to capture the essence of the plasticizer literature and provides a comprehensive explanatory and predictive framework. It cannot, of course, explain all the complex details of crystal morphology and mechanical properties, but this is a chapter TABLE 7.3 87 7.4 POLYMER COMPATIBILITY Polymer blends with PLA hold promise to overcome some of its deficiencies such as relatively poor mechanical resilience. To design blends rationally, the principles of solvent/polymer compatibility (or, if required, incompatibility) can be transferred to polymer/polymer compatibility: “like is compatible with like.” As explained in Section 7.2, for compatibility, the centers of the two polymer spheres should be close and their radii should show a large overlap. Calculating the overlap is not quite as simple as calculating the center-to-center distances. Each sphere has a different radius, so although one sphere (small radius) might find most of itself inside the second sphere, the second sphere (large radius) might find only a small percentage of overlap with the first. When automating the calculations of overlap, it is necessary to calculate both percentages of spherical overlap and score them somehow. In the example below, the total overlap is treated as the most important property, but because all data are shown the user is free to interpret the data in other ways. So in Table 7.3, PMMA is scored as 200 because PMMA has Polymer Overlap with PLA Polymer dD dP dH Radius Polylactic acid (PLA) Polyurethane (PU) Polymethylmethacrylate (PMMA) Polyethersulfone Polyethylmethacrylate (PEMA) Polyethylene oxide (PEO, PEG) Polyvinylchloride (PVC) PET Epoxy Polycarbonate (PC) Polypropylene oxide (PPO, PPG) Polyoxymethylene (PON) Polyvinylidenefluoride (PVF) Polycaprolactone Polystyrene (PS) Polysulphone Polyphenyleneoxide (PPO) CyclicOlefinCopolymer (COC) Polysilicone Polyvinylacetate (PVA) Polyvinylbutyral Nylon 66 Polyacrylonitrile (PAN) Polyethylene (PE) Polypropylene (PP) Polyvinylpyrrolidone (PVP) 18.6 18.1 18.6 19 17.6 17 19.2 18.2 17.4 18.2 16.5 17.2 17 17.7 18.5 16 17.9 18 17.2 17.6 18.6 17.4 22.4 16.9 18 21.4 9.9 9.3 10.5 11 9.7 10 7.9 6.4 10.5 5.9 9 9.2 12.1 5 4.5 6 3.1 3 3 2.2 4.4 9.9 14.1 0.8 0 11.6 6 4.5 5.1 8 4 5 3.4 6.6 9 6.9 7 9.8 10.2 8.4 2.9 6.6 8.5 2 3 4 13 14.6 9.1 2.8 1 21.6 10.7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 Overlap Target 200 100/100 200 100/100 200 100/100 142 42/100 140 41/98 139 41/98 138 41/97 135 40/96 133 39/94 130 38/92 126 37/89 115 34/81 114 34/81 108 32/76 104 31/73 93 27/65 84 25/59 84 25/59 82 24/58 73 21/51 73 21/51 69 20/49 57 17/40 46 14/32 4 1/3
  • 88 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID) a 100% overlap with PLA and PLA has a 100% overlap with PMMA. Because it is difficult to grasp these concepts purely from the numbers, as shown in the figures below, the HSPiP software allows the user to see the two spheres in 3D space. It does not matter, in thermodynamic terms, whether the spheres tend to overlap in the dD, dP, or dH directions, but chemically it is important to understand where the polymers are relative to each other as this can suggest the sorts of solvents that would be mutually more compatible—nonpolar solvents if the bias is toward dD, polar solvents if the bias is toward dP, and hydrogen bonding solvents for a dH bias (Table 7.3). The PLA is defined as the target. Below is a small set of polymers selected from the large list produced by the software, ordered in terms of highest mutual overlap. Not surprisingly, a number of acrylates feature as being closely compatible (as is confirmed by the fact that blends show a single Tg [26]), along with, for example, PET, polyurethane (PU), and polycarbonate (PC). The bottom of the list includes polyamide 6,6 (Nylon), polyvinyl alcohol, polyethylene (PE), and polyvinylpyrrolidone (PVP). Although one would have reached these particular conclusions via chemical intuition, the numerical values over a big range of polymers allow the user to work out more intelligently what happens with other polymers. For example, the rather poor match of PLAwith polystyrene (PS) (108, 32/76) emphasizes the point that PLA/PS polymer blends would represent rather a technical challenge. The thermodynamics of these systems allow us to approach questions beyond those of knowing whether the polymers are mutually compatible. What, for example, are the criteria to be used in an attempt to cast a PLA/PS blend from solvent? For simplicity we will restrict ourselves to pure solvents, but the same approach works with solvent blends. Using THF, a plot of the two polymer spheres plus the solvent shows that the THF is suitable for the PLA but at the edge of the PS. This is difficult to see without actively rotating the 3D plot, but the black dot is the THF, sitting within the larger PLA sphere. This means that as the solvent evaporates from the polymer blend, the PS would phase separate first, which is something that may or may not be desirable (Figure 7.2). Similarly, by using toluene as the solvent (black dot now sitting inside the smaller PS sphere), it is the PLA that will phase separate first (Figure 7.3). Although these statements seem rather obvious when plotted on these 3D graphs, it is surprisingly hard to develop optimized systems without a thermodynamic tool such as HSP and, conversely, remarkable productivity is expected to result from using such a tool. Blending biopolymers to overcome their mutual deficiencies is a popular theme. Unfortunately both poly(butylene succinate) (PBS) [20] and poly(3-hydroxybutyrate) FIGURE 7.2 Solvent spheres for PLA and PS showing THF far from PS. (PHB) [21] have a relatively large HSP distance and are experimentally found to be immiscible or at least only somewhat compatible. However [20], PBS can help the PLA to crystallize if a higher crystallinity version is desired. If it is not possible to find a good biopolymer match, HSP can be used to identify compatible polymers that are at least biodegradable. For example, the HSP match with polycaprolactone is excellent and it is reported to be a useful polymer for blending with PLA [22]. A paper on a poly(ether)urethane PLA blend [23] offers plausible arguments about why the polyether and polyester parts of the blend should be a reasonable chemical match to PLA and thereby provide a suitably tough polymer. Without the chemical structure, it is not possible to estimate the HSP. But given that the paper shows that the blend is a phase- FIGURE 7.3 PLA and PS spheres with toluene close to PS.
  • RATIONAL COMPOSITE DESIGN separated structure, which gains its improved toughness from deformations in the PU-rich segments, it is hard to know whether compatibility is desirable or undesirable. A number of papers studied polyethylenes in PLA as deliberately phase-separated areas to provide better toughness [24]. But without compatibility with the PLA, these are likely to be sources of weakness rather than strength. Hence, there is a trend to use block copolymer compatibilizers such as PLA/PE. As long as the PE is at a low level [25], although if the PE is above 1% the polymer cannot claim to be to composting standards, merely that it is “bio-based.” A helpful review of glass transition temperatures in PLA and its blends provides a good summary of many of these issues [26]. 7.5 ENVIRONMENTAL STRESS CRACKING With a relatively young polymer such as PLA, it is possible to put a new product out into the marketplace with a potentially serious technical liability: it might break unexpectedly in use. Over the years, end users for other polymers have learned by trial and (sometimes disastrous) error to avoid the problem, but thermodynamics can help us avoid the error by rational processes. The problem is that a plastic part in contact with a particular liquid might crack for no apparent reason. The crack will appear at a point of stress in the part, even though mechanical tests will have shown that the stresses pose no risk at all. This phenomenon is called environmental stress cracking (ESC). To say that a polymer will not dissolve in a solvent is to say that the free energy of the polymer and solvent system is lower than that of polymer plus solvent. But if some extra energy is supplied, then the balance can change. Physical stress on a part is sufficient to add that extra energy and therefore allow the hitherto stable polymer and solvent system to become the weak polymer plus solvent, which then breaks. It is hard to quantify that statement, but fortunately we do not have to. If a solvent is inside the polymer HSP sphere, then we do not have an ESC problem because it is already obvious that the solvent and polymer are incompatible. In other words, a “good” solvent does not give ESC (i.e., long-term damage under stress) because the short-term damage is readily seen in short-term tests. On the other hand, if a solvent is well outside the sphere, then experience shows that extra stress is insufficient to overcome the basic thermodynamics. As a practical user, one therefore has to worry only about ESC for solvents that are just outside the sphere. Taking a table of 60 common solvents and sorting them by their RED with respect to PLA, it is clear where the dangers lie. Table 7.4 is identical to Table 7.2. This makes the point that the same HSP data can be insightful for different issues. 89 TABLE 7.4 Solvents Near the Danger Zone for Environmental Stress Cracking dD dP dH RED Molar Volume 17.8 14.9 18 18.0 17.4 17.6 17.66 20.3 15.3 15.8 16.0 15.2 18.0 15.8 16.5 5.7 4.5 1 1.4 4.1 1.0 1.03 18.2 18.0 5.7 5.7 5.1 21.7 6.1 10.6 14.3 8.2 3 2.0 13.5 3.1 2.05 10.9 6.1 14.5 15.8 14.7 5.1 16.4 17.7 0.87 0.88 0.89 0.89 0.90 0.90 0.93 0.95 0.98 1.02 1.09 1.11 1.11 1.15 1.15 124.7 117.1 100 106.6 105.7 123.9 157.3 96 52.9 92 92 96 66 76.9 138.5 Solvent Phenoxyethanol Isopropyl acetate Aromatic hydrocarbons Toluene Cyclohexanol Xylene D-Limonene Sulfolane Acetonitrile 2-Butanol 1-Butanol t-Butanol Ethylene carbonate 2-Propanol DPG Although the thermodynamics are important, kinetics also play their part. The mobility of the poor solvent is important, and to a first approximation, a smaller molar volume represents a more aggressive solvent. There is not much to choose between those solvents on the borderline, though DPG (dipropylene glycol) would be less of an ESC threat than 2-propanol, though they have the same RED value. The RED of ethanol is 1.3. It is a marginal call whether the smaller size of the molecule will make it sufficiently aggressive to cause ESC, but it is certainly a cause for concern. It would be good to test these predictions against the PLA literature, but there seems to be little available. 7.6 RATIONAL COMPOSITE DESIGN Instead of blending polymers to overcome some of the weaknesses of PLA, particles, plates, and fiber fillers can be used. It would seem that “solubility parameters” are irrelevant here for composite blends. But it turns out to be very useful to measure the HSP of fillers in order to find one that is optimized for its compatibility with the PLA. If one were designing a new filled system, a good place to start would be one with HSP of (18.6, 9.9. 6.0) because it would be highly likely to blend in easily and form a stable interface. To find the HSP for a filler, take a range of solvents (typically 15–20) and attempt to disperse the filler in the solvent. Fillers that are compatible with the solvent tend to remain in dispersion, and those that are incompatible will fail to get dispersed or crash out of the dispersion very quickly. Clearly, there is some degree of empirical judgment with this approach, which is why one uses multiple solvents to reduce statistical errors. A dense filler will, for example, settle out faster in less dense solvents, so some allowance has to be
  • 90 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID) made for borderline solvents. A sphere optimizer can then be used to find the HSP center and radius for this filler. A typical example of such an experiment is with “carbon black” [8]. There are many varieties of carbon black. A typical “hydrophobic” carbon black has HSP of (16.6, 8.9, 7.3), which gives a distance of 4.3 from PLA. A typical “hydrophilic” carbon black has HSP of (20.5, 11.0, 13.2) with a distance of 8.2. Both are inside PLA’s radius, but clearly the “hydrophobic” carbon black would be the one to choose. However, even better would be to screen a range of carbon blacks to find one even closer to PLA. Given that carbon black is used with PLA for many reasons such as providing a cheap filler, raising the melt temperature, making it conductive, or creating biosensors, the rational choice of the surface properties could be very important [27]. It is possible, for example, to make the components too compatible. If the carbon black were perfectly (nano-) dispersed within the PLA, then there would be no particle–particle contact and the electrical conductivity would be zero. Conversely, if the carbon is intended to be there as a pure filler, having full (nano-) dispersion with no weak carbon–carbon particle contacts will maximize the strength/stability of the blend. It is interesting to note that the HSP of carbon nanotubes (CNT) is measured (using the same technique on multiwalled nanotubes) and found to be (18.5, 7, 5) [28] with a distance from PLA of only 3.1, making PLA a very good polymer when good CNT dispersions are required. Examples from the literature seem to indicate that the compatibility is indeed good, allowing facile dispersion [29, 30]. Clay fillers offer much potential within PLA. The two classic problems are those of exfoliating the clays and ensuring stable dispersions. The clays are typically provided as organoclays, using quaternary ammonium long chains (typically tallows) via ion exchange to swap with the sodium ions in the clay (a paper exists confirming that there is no intercalation of PLA in the sodium-based clay [31]). Because one of the characteristics of clays is their inherent cheapness, it is not always clear that the organoclays are entirely free of sodium ions or of excess quaternary salt. The literature (see Refs. 32–36) is therefore rather confusing, especially as there are many different types of organoclays and at least three methods for dispersing them within PLA (solvent casting, pure blending, polymerization blending). There are also different ways of analyzing the results; some focus on the clays themselves, others on the functionality of the resulting blends. Running through the literature is a general theme of disappointment—usually there is some modest benefit for one property and a loss of benefit for another. The author has attempted to throw light on the subject using HSP but has failed, perhaps for the same reasons of impurities in the organoclays that are used. The data on solvent effects on the clays are confusing, so it is hard to extract meaningful HSP for any particular clay. The solvent swelling data reported by Burgentzl et al. [37] make intue itive sense, but the interplate distance data (arguably the more fundamental parameter) do not. However, it seems clear to me that a clay such as Cloisite 10A that contains a benzyl group or Cloisite 30B containing a hydroxyethyl group (each also with a tallow group) is more likely to be inherently compatible with PLA than the classic Cloisite 15A with two tallow groups. The relatively good dispersion at short times of the 30B supports the idea [38]. Another study [35] uses Flory parameters to conclude that the interaction between the 30B and PLA is “moderate.” The same study [35] provides a word of caution about particulate contamination in 30B (and, presumably, other clays) and is a reminder that users should thoroughly check that “disappointing” results from clays are not simply due to macrodefects from contamination. Finally, a montmorillonite dispersed via a solvent technique using 1,11-aminoundecanoic acid showed complete exfoliation (better than dodecylamine and 1,12-diaminododecane) with remarkably small effects on the mechanical properties [39]. It is not obvious from HSP why this particular clay should be so good at producing exfoliation in PLA and why good exfoliation should not produce significant changes in the mechanical properties. In summary, it seems that an approach such as HSP can provide an intellectual backdrop to a lot of confusing data (for a recent review of PLA and clays, see Ref. 40). At the very least, it provides a numerical approach that seems superior to phrases such as “this clay is more compatible with PLA because it is more polar” that sometime appear in the PLA/clay literature. 7.7 DIFFUSION AND BARRIER PROPERTIES The (Fickian) diffusion coefficient of a chemical within a polymer is governed by two sets of properties. The first comes from the polymer itself: its free volume and its ability to undergo main-chain segmental motion. The second comes from the chemical: its molecular size, its degree of branching, and its general stiffness. The compatibilities of polymer/ chemical are not, surprisingly, reflected directly in the diffusion constant. So for PLA, the single largest factor governing the rate of diffusion should be the degree of crystallinity; the higher it is, the slower the diffusion. The “should be” is there because there is evident surprise in a number of papers that the degree of crystallinity seems to make little or no difference in diffusivity of chemical compounds [41]. A chemist can evaluate two molecules and make estimates of their relative diffusion constants. The linear 1-hexanone, for example, will diffuse relatively quickly, whereas the relatively rigid chair conformation of cyclohexanone makes it notoriously slow to diffuse out of the many polymers for which it is a good solvent. Similarly, 1-butanol will diffuse
  • DIFFUSION AND BARRIER PROPERTIES faster than 2-butanol and 1-butanol will diffuse faster than 1hexanol. But we all know that chemical compatibility plays a significant role in diffusion and barrier properties. Although the above analysis is correct about diffusion constants, diffusion requires a concentration gradient to drive it. If the chemical and PLA are not compatible, then the concentration of the chemical in the surface layer of the polymer will be low and so will be the rate of diffusion. So, solubility (and therefore solubility parameters) is very important. For the packaging industry in particular, the permeability is the most used measure and this neatly combines both issues because permeability ¼ diffusivity  solubility in the large number of cases where Fickian diffusion is the governing phenomenon. The other key factor is that when a compatible chemical is present, it plasticizes the polymer and allows more mainchain segmental motion. So, there can be a large (orders of magnitude) increase in diffusion coefficient as the chemical diffuses into the polymer. If the chemical is sufficiently compatible to disrupt the crystalline (essentially impenetrable) domains, then the increase in rate of diffusion can be dramatic. The details of diffusion are usually not of great interest to the practical PLA technologist. What is of more importance is to know if, for example, a particular flavor molecule will or will not diffuse out of a PLA package, or whether a drug release will be based on diffusion through the PLA capsule or simply on (bio)degradation of the capsule in vivo. HSP provide a robust framework for answering such basic questions and also help in understanding the full complexities of the system. The quick way of predicting permeability is to find out if the chemical is within, on the edge of, or well outside the HSP sphere of PLA. If it is well within the sphere and is not too big/ branched/rigid, then permeability will be relatively high and PLAwill be a poor barrier. If it is well outside the sphere, then even if it is a relatively small molecule its permeability will be low and PLA will be a good barrier. Chemicals near the edge of the sphere will diffuse according to their molecular size/shape. The definitions of “big,” “small,” and branched or rigid are necessarily subjective. A rule of thumb [8] is that the log of the diffusivity is inversely proportional to the molar volume, with a doubling of molar volume reducing diffusivity by a factor of 10–100. The chemist has to use experience to judge how branched or rigid a molecule might be. The key literature on this subject [42] shows that Dlimonene, for example, has a relatively large distance (8.8). Therefore, its solubility in the surface layer is relatively small and the rate of diffusion will be low. It is also a relatively large, branched molecule, giving it a low diffusion coefficient. Ethyl acetate is somewhat closer (7.5) but would still be expected to be a relatively poor diffuser, though being a smaller molecule it would be expected to be somewhat faster than D-limonene. Other work shows, as would be 91 expected, that oils (such as olive oil) and polar molecules such as acetic acid (which have a low D value and are therefore a large distance from PLA) scarcely diffuse/absorb in PLA [43]. An indirect analysis of barrier properties to blueberry aromas [44] such as nonanal shows that PLA retains the aromas, but again the distance from nonanal to PLA is 6.1, still a relatively poor diffuser. There seems to be a lack of examples of (say) aroma molecules that are a short distance (say, <4) from PLA, where barrier properties would be expected to be limited. As a stimulus to research, a glance at HSP of some typical aroma chemicals suggests that PLA would not be a good barrier for cinnamon-flavored goods. The distance between PLA and cinnamaldehyde is <3. For those who really want to understand the diffusion science then there is no substitute for a full Fickian diffusion modeler. By taking into account concentration-dependent diffusivities, PLA thickness, and starting concentration of the chemical, breakthrough times and permeabilities can be calculated. Or, from measuring breakthrough times, the diffusion coefficients can be estimated by fitting. Here is an example that assumes a relatively high concentration of the chemical (20% volume) but a relatively rigid polymer with a low diffusion rate at low concentration, which rises by five orders of magnitude (not untypical) up to a highly plasticized concentration of 33%. The chemical has broken through 0.1 mm in 68 h and would have a permeation rate of 4.5E–12 g/(cm2 s) (Figure 7.4). This example is relevant only to a liquid packaging example. For vapor barriers, the concentrations are small and therefore the concentration dependence terms are irrelevant. As shown in Auras et al. [42], a very good fit to the experimental data of ethyl acetate diffusion is found with a fixed diffusion coefficient of 8.7E–13 cm2/s (Figure 7.5). 7.7.1 Gases HSP work as well for gases as they do for liquids, so it is possible to make a guesstimate of relative permeabilities of gases if the diffusivities of each gas are assumed (wrongly of course) to be the same. The HSP distances for various gases are shown in Table 7.5, ordered by closeness (higher solubility) to PLA. The literature data on solubilities in PLA are so far not extensive enough to check these predictions with any certainty, but the data in Ref. 45 are encouraging as the solubilities are in the order CO2 > O2 > N2, which is what the table predicts (Table 7.5). 7.7.2 Water So far, water has not been mentioned in this chapter. This is not because interactions between water and PLA are unimportant. Rather, it is because water is such an extreme because of its “almost off the scale” dH parameter
  • 92 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID) FIGURE 7.4 Concentration-dependent solvent uptake, typical of liquid absorption. (thermodynamic effects), its tiny molar volume (kinetic effects), and its reactivity with the PLA ester linkage (chemical effects). In the context of diffusion and barrier properties, water has two key effects: it can hydrolyze the PLA and therefore increase the intrinsic diffusion coefficients and it can also plasticize the PLA through its rather large percentage uptake, thereby increasing the diffusivities of other chemicals. Although there is plenty of data on PLA hydrolysis, the effects of plasticization on diffusion coefficients are not well documented. FIGURE 7.5 The rate of hydrolysis is an extremely complex issue with autocatalytic features that even depend on not only the proportions of D- and L-lactide but whether they are DLcopolymers or mixtures of D- and L-polymers. Fortunately, the papers by Tsuji (e.g., IX in the hydrolysis series [46]) cover the issues for all those who wish to explore them in depth. The problem of water becomes more significant in the context of the next section, pharmacological transport, because although, as we have seen, diffusion and permeability Constant diffusion coefficient, typical of (relatively insoluble) vapor uptake.
  • SUMMARY TABLE 7.5 HSP Distances of Gases from PLA Gas Chlorine Hydrogen sulfide Carbon dioxide Sulfur dioxide Ethylene Acetylene Ethane Oxygen Methane Nitrous oxide Carbon monoxide Ammonia Water Nitrogen Nitric oxide Hydrogen Helium HSP Distance from PLA 6.5 6.6 6.8 7.0 10.7 11.7 13.0 14.0 14.8 16.1 16.2 16.4 16.4 17.7 18.4 29.4 37.1 can be highly complex subjects, adding the complexities of water makes it even harder to work out the behavior of molecules of pharmacological interest when used in their aqueous milieu. 7.8 PHARMACOLOGICAL TRANSPORT Without some guiding principles, predicting and understanding the behavior of, say, controlled release drugs within PLA is extraordinarily hard. The pharmaceutical world likes to use QSAR techniques for predicting just about any effect of interest. The approach has many merits, but generalizations from different QSAR analyses of ostensibly the same topic are hard to come by because different QSARs tend to use different parameters for their best fit. It is surprising to many people that the humble HSP approach, with only three parameters, can be highly insightful even in the complex circumstances of pharmaceuticals. Suppose, to take a simplistic example, that we wanted to make a controlled release version of an analgesic. Which would we expect to diffuse faster through PLA—aspirin or ibuprofen? The molar volume of aspirin is about 75% that of ibuprofen and the HSP distances from PLA are 4.7 and 9.4, respectively. On both accounts, therefore, aspirin should release much faster than ibuprofen. Is this simplistic thinking really relevant to real-world pharmacological issues? It has been shown [47] that the breakthrough times of extremely cytotoxic chemicals through different sorts of gloves can be predicted using the same principles. Here, it is literally a matter of life and death that the estimates be accurate, so the predictions have been experimentally validated. 93 A random cross section of the voluminous literature on microencapsulation of drugs within PLA (or, more often, PLA/PGA copolymers—but for this chapter the distinction will not be made as the HSP of PGA are not far from PLA) shows the need for some unifying principles. The first key question is whether the drug is sufficiently insoluble in the PLA that it is present only inside the capsule. Even crude HSP calculations can give some idea of this. In our simple example, one would expect a Fickian diffusion, linear with respect to the square root of time, for aspirin dissolved in the PLA. This is the classic behavior of a diffuser that is soluble in the matrix. The ibuprofen would be expected to show zeroth-order slow release by diffusion through the membrane from the pool in the center, classic behavior for systems where the mutual solubility is poor. Of the seven samples selected at random from the same journal, the HSP of four molecules (acetaminophen [48], retinoic acid [49], vancomycin [50], and ganciclovir [51]) are so far from PLA that they can plausibly only be (and the data show that they are) zeroth order in release; that is, they rely on a pool inside the shell, with release controlled by the permeability. They should also show a strong time lag before any release. Another aspect of such molecules is that during manufacture a significant proportion is simply stuck to the outside of the shell, so release starts with a sharp burst followed by a time lag. One drug, nifedipine [52], has a notably short distance (<4) from PLA and the best data seem to fit well to a Fickian model corresponding to plenty of the drug being dissolved in the membrane. Metoclopramide [53] is of intermediate distance (5.8) and showed intermediate properties. For high molecular weight (smaller HSP radius), the kinetics were fully zeroth order with a long time lag before the drug diffused through the shell. For lower molecular weight formulations, there was a mixture of zeroth-order and Fickian diffusion. Fluorouracil [54] is of a similar distance (6) to metoclopramide but in the particular circumstances of that paper (the materials were prepared by spray drying) gave zeroth-order kinetics after an initial burst from, presumably, externally absorbed drug. 7.9 SUMMARY The thermodynamic approach, based on the relatively simple HSP methodology, consistently provides insights into chemical compatibility issues spanning a large range of important topics in PLA science and technology. Only water seems to fall outside this framework. The numeric thermodynamic approach, for all its limitations (e.g., kinetic effects), seems to be more fruitful than relying on vague phrases such as “polar attraction” or “hydrogen bonding” when one needs to make
  • 94 CHEMICAL COMPATIBILITY OF POLY(LACTIC ACID) critical decisions about where to turn when faced with the challenges and opportunities presented by PLA. REFERENCES 1. T. Oishi, J. M. Prausnitz, Estimation of solvent activities in polymer solutions using a group-contribution method, Ind. Eng. Chem. Process Des. Dev. 1978, 17, 333–339. 2. R. Todeschini, V. Consonni, Handbook of Molecular Descriptors, Wiley-VCH, Weinheim, 2000. 3. S. H. Hilal, L. A. Carreira, S. W. Karickhoff, C. M. Melton, Estimation of gas–liquid chromatographic retention times from molecular structure, J. Chromatogr. A 1994, 662, 269–280. 4. P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, New York, 1953. 5. C. M. Hansen, Hansen Solubility Parameters: A User’s Handbook, CRC Press, Boca Raton, FL, 2007. 6. E. Stefanis, C. Panayiotou, Prediction of Hansen solubility parameters with a new group-contribution method, Int. J. Thermophys. 2008, 29, 568–585. 7. M. Belmares, M. Blanco, W. A. GoddardIII, R. B. Ross, G. Caldwell, S. H. Chou, J. Pham, P. M. Olofson, C. Thomas, Hildebrand and Hansen solubility parameters from molecular dynamics with applications to electronic nose polymer sensors, J. Comput. Chem. 2004, 25, 1814–1826. 8. S. J. Abbott, C. M. Hansen, HSPiP—Hansen Solubility Parameters in Practice, www.hansen-solubility.com 2008. 9. (a) P. van de Witte, A. Boorsma, H. Esselbrugge, P. J. Dijkstra, J. W. A. van den Berg, J. Feijen, Differential scanning calorimetry study of phase transitions in poly(lactide)–chloroform–methanol systems, Macromolecules 1996, 29, 212–219. (b) P. van de Witte, et al., In situ analysis of solvent/nonsolvent exchange and phase separation processes during the membrane formation of polylactides, J. Appl. Polym. Sci. 1996, 61, 685–695.(c) P. van de Witte, et al., Phase separation processes in polymer solutions in relation to membrane formation, J. Membr. Sci. 1996, 117, 1–31.(d) P. van de Witte, et al., Phase transitions during membrane formation of polylactide. I. A morphological study of membranes obtained from the system polylactide–chloroform—methanol, J. Membr. Sci. 1996, 113, 223–236.(e) P. van de Witte, et al., Metastable liquid–liquid and solid–liquid phase boundaries in polymer–solvent–nonsolvent systems, J. Polym. Sci. Part B 1997, 35, 763–770.(f) P. van de Witte, et al., A morphological study of membranes obtained from systems polylactide–dioxane–methanol, polylactide–dioxane–water, and polylactide–N-methyl pyrrolidone–water, J. Polym. Sci. Part B 1996, 34, 2569–2578. (g) P. van de Witte, et al., Phase behavior of polylactides in solvent–nonsolvent mixtures, J. Polym. Sci. Part B 1996, 34, 2553–2568. 10. K. S. Anderson, K. M. Schreck, M. A. Hillmyer, Toughening polylactide, Polym. Rev. 2008, 48, 85–108. 11. R. E. Conn, J. J. Kolstad, J. F. Borzelleca, D. S. Dixler, L. J. FilerJr., B. N. LaDu, M. W. Pariza, Safety assessment of polylactide (PLA) for use as a food-contact polymer, Food Chem. Toxicol. 1995, 33, 273–283. 12. M. Mutsuga, U. Kawamura, K. Tanamoto, Migration of lactic acid, lactide and oligomers from polylactide food-contact materials, Food Addit. Contam. 2008, 25, 1283–1290. 13. E. Piorkowska, Z. Kulinski, A. Galeski, R. Masirek, Plasticization of semicrystalline poly(L-lactide) with poly(propylene glycol), Polymer 2006, 47, 7178–7188. 14. Z. Kulinski, et al., Plasticization of poly(L-lactide) with poly(propylene glycol), Biomacromolecules 2006, 7, 2128–2135. 15. X. Chen, K. H. Schilling, W. E. KellyJr.,U.S. Patent 5,908,918, 1998 (to Chronopol Inc.). 16. T. Matsumoto, H. Shinnumadate, S. Kumazawa, M. Kimura, U.S. Patent 7,351,785, 2008 (to Toray Industries, Inc. (Japan)). 17. Z. Ren, L. Dong, Y. Yang, Dynamic mechanical and thermal properties of plasticized poly(lactic acid), J. Appl. Polym. Sci. 2006, 101, 1583–1590. 18. N. Ljungberg, B. Wessln, Tributyl citrate oligomers as e plasticizers for poly(lactic acid): thermo-mechanical film properties and aging, Polymer 2003, 44, 7679–7688. 19. R. Bhardwaj, A. K. Mohanty, Modification of brittle polylactide by novel hyperbranched polymer-based nanostructures, Biomacromolecules 2007, 8, 2476–2484. 20. T. Yokohara, M. Yamaguchi, Structure and properties for biomass-based polyester blends of PLA and PBS, Eur. Polym. J. 2008, 44, 677–685. 21. B. M. P. Ferreira, C. A. C. Zavaglia, E. A. R. Duek, Films of PLLA/PHBV: thermal, morphological, and mechanical characterization, J. Appl. Polym. Sci. 2002, 86, 2898–2906. 22. R. G. Sinclair, U.S. Patent 4,057,537, 1977, (to Gulf Oil Corporation). 23. Y. Li, H. Shimizu, Toughening of polylactide by melt blending with a biodegradable poly(ether)urethane elastomer, Macromol. Biosci. 2007, 7, 921–928. 24. K. S. Anderson, S. H. Lim, M. A. Hillmyer, Toughening of polylactide by melt blending with linear low-density polyethylene, J. Appl. Polym. Sci. 2003, 89, 3757–3768. 25. K. S. Anderson, M. A. Hillmyer, The influence of block copolymer microstructure on the toughness of compatibilized polylactide/polyethylene blends, Polymer 2004, 45, 8809–8823. 26. G. L. Baker, E. B. Vogel, M. R. SmithIII, Glass transitions in polylactides, Polym. Rev. 2008, 48, 64–84. 27. N. Wang, X. Zhang, X. Ma, J. Fang, Influence of carbon black on the properties of plasticized poly(lactic acid) composites, Polym. Degrad. Stabil. 2008, 93, 1044–1052. 28. S. Detriche, G. Zorzini, J. F. Colomer, A. Fonseca, J. B. Nagy, Application of Hansen solubility parameters theory to carbon nanotubes, J. Nanosci. Nanotechnol. 2008, 8, 1–11. 29. H. S. Kim, B. H. Park, J. S. Yoon, H. J. Jin, Preparation of biodegradable nanocomposites by incorporation of functionalized carbon nanotubes, Key Eng. Mater. 2006, 326–328, 1785–1788. 30. Y. Shimamura, A. Yamamoku, K. Tohgo, S. Tasaka, H. Araki, Mechanical properties of carbon nanofiber reinforced polylactic acid, Key Eng. Mater. 2007, 345–346, 1225–1228.
  • REFERENCES 31. H. J. M. Hanley, C. D. Muzny, D. L. Ho, C. J. Glinka, E. Manias, A SANS study of organoclay dispersions, Int. J. Thermophys. 2001, 22, 1435–1448. 32. J. H. Chang, Y. U. An, G. S. Sur, Poly(lactic acid) nanocomposites with various organoclays. I. Thermomechanical properties, morphology, and gas permeability, J. Polym. Sci. Part B 2003, 41, 94–103. 33. J. H. Chang, Y. U. An, D. Cho, E. P. Giannelis, Poly(lactic acid) nanocomposites: comparison of their properties with montmorillonite and synthetic mica, Polymer 2003, 44, 3715–3720. 34. C. Thellen, C. Orroth, D. E. Froio, D. Ziegler, J. Lucciarini, R. Farrell, N. A. D’Souza, J. A. Ratto, Influence of montmorillonite layered silicate on plasticized poly(L-lactide) blown films, Polymer 2005, 46, 11716–11727. 35. Z. Yua, J. Yin, S. Yan, Y. Xie, J. Ma, X. Chen, Biodegradable poly(L-lactide)/poly(3-caprolactone)-modified montmorillonite nanocomposites: preparation and characterization, Polymer 2007, 48, 6439–6447. 36. H. Tian, H. Tagaya, Preparation, characterization and mechanical properties of the polylactide/perlite and the polylactide/montmorillonite composites, J. Mater. Sci. 2007, 42, 3244–3250. 37. D. Burgentzl, J. Duchet, J. F. Grard, A. Jupin, B. Fillon, e e Solvent-based nanocomposite coatings. I. Dispersion of organophilic montmorillonite in organic solvents, J. Colloid Interface Sci. 2004, 278, 26–39. 38. S. Bourbigot, G. Fontaine, S. Bellayer, R. Delobel, Processing and nanodispersion: a quantitative approach for polylactide nanocomposite, Polym. Test. 2008, 27, 2–10. 39. J. H. Shim, E. S. Kim, J. H. Joo, J. S. Yoon, Properties and morphology of poly(L-lactide)/clay composites according to the clay modification, J. Appl. Polym. Sci. 2006, 102, 4983–4988. 40. P. Bordes, E. Pollet, L. Avrous, Nano-biocomposites: biodee gradable polyester/nanoclay systems, Prog. Polym. Sci. 2009, 34, 125–155. 41. G. Colomines, S. Domenek, V. Ducruet, A. Guinault, Influences of the crystallisation rate on thermal and barrier properties of polylactide acid (PLA) food packaging films, Int. J. Mater. Form. 2008, 1(1), 607–610. 42. R. Auras, B. Harte, S. Selke, Sorption of ethyl acetate and D-limonene in poly(lactide) polymers, J. Sci. Food Agric. 2006, 86, 648–656. 95 43. D. V. Plackett, V. K. Holm, P. Johansen, S. Ndoni, P. V. Nielsen,  T. Sipilainen-Malm, A. S€dergard, S. Verstichel, Characterio zation of L-polylactide and L-polylactide–polycaprolactone co-polymer films for use in cheese-packaging applications, Packag. Technol. Sci. 2006, 19, 1–24. 44. E. Almenar, H. Samsudin, R. Auras, B. Harte, M. Rubino, Postharvest shelf life extension of blueberries using a biodegradable package, Food Chem. 2008, 110, 120–127. 45. L. Bao, et al., Gas permeation properties of poly(lactic acid) revisited, J. Membr. Sci. 2006, 285, 166–172. 46. H. Tsuji, K. Ankara, Poly(L-lactide). IX. Hydrolysis in acid media, J. Appl. Polym. Sci. 2002, 86, 186–194. 47. C. M. Hansen, Polymer science applied to biological problems: prediction of cytotoxic drug interactions with DNA, Eur. Polym. J. 2008, 44, 2741–2748. 48. M. K. Lai, R. C. C. Tsiang, Microencapsulation of acetaminophen into poly(L-lactide) by three different emulsion solvent-evaporation methods, J. Microencapsul. 2005, 22, 261–274. ¨ u 49. Y. Cirpanli, N. Unl€, S. Calis, A. At Hincal, Formulation and ¸ ¸ ılla in vitro characterization of retinoic acid loaded poly(lactic-coglycolic acid) microspheres, J. Microencapsul. 2005, 22, 877–889. ¨ ¨ 50. Y. Ozalp, N. Ozdemir, V. Hasirci, Vancomycin release from poly(D,L-lactide) and poly(lactide-co-glycolide) disks, J. Microencapsul. 2002, 19, 83–94. 51. X. Chen, C. P. Ooi, W. S. Lye, T. H. Lim, Sustained release of ganciclovir from poly(lactide-co-glycolide) microspheres, J. Microencapsul. 2005, 22, 621–631. 52. K. S. Soppimath, T. M. Aminabhavi, Ethyl acetate as a dispersing solvent in the production of poly(DL-lactide-co-glycolide) microspheres: effect of process parameters and polymer type, J. Microencapsul. 2002, 19, 281–292. 53. S. A. Elkheshen, M. A. Radwan, Sustained release microspheres of metoclopramide using poly(D,L-lactide-co-glycolide) copolymers, J. Microencapsul. 2000, 17, 425–435. ´ ´ 54. M. D. Blanco, R. L. Sastre, C. Teijon, R. Olmo, J. M. Teijon, 5Fluorouracil-loaded microspheres prepared by spray-drying poly(D,L-lactide) and poly(lactide-co-glycolide) polymers: characterization and drug release, J. Microencapsul. 2005, 22, 671–682.
  • 8 OPTICAL PROPERTIES ~ CARLA M. B. GONCALVES, JOAO A. P. COUTINHO, ¸ 8.1 AND ISABEL INTRODUCTION Polylactides (PLAs) have been known for several decades but only recently have these polymers gained commercial significance as a leading environmentally benign plastic available from renewable resources. Therefore, it is highly desirable to understand the optical properties of these materials so that they can be manipulated to develop materials of desirable characteristics for a specific objective. Optical properties such as color, clarity, and refractive index are important in dyeing operations for textile and in various packaging applications [1, 2]. Many of the physical properties of PLA are influenced by the amount and distribution of the R- and S-lactic acid stereocenters in the polymer chain, which reflect its history including the stereochemistry RR (D-lactide), SS (L-lactide), RS (meso-lactide), or a mixture of equal amounts of D- and L-lactide referred to as racemic or DL-lactide) of the feed composition, polymerization kinetics, and extent of transesterification and racemization (see Chapters 1–4). As this rich variety of structures leads to materials with different characteristics, from fully amorphous to semicrystalline, a discussion on the PLA optical properties across a broad range of stereooptical compositions is valuable. Since most applications of PLA-based materials are in the solid state, the detailed knowledge of the composition, bulk structure, and conformation of these materials is crucial. For example, it is well known that poly(L-lactide) (PLLA) obtained from LL-lactide is usually molded at 100–120 C in industrial melt processing because of the higher crystallization rate. It has been reported that when it is crystallized within this temperature region a mixture of crystals, a- and b-forms, is formed [2]. Although infrared and Raman spectroscopy is a very promising analytical technique, as will be M. MARRUCHO shown below, many aspects of the optical properties for solidstate PLA remain unexplored. In this chapter, the main optical techniques used to characterize PLA-based polymers are discussed in four sections: (1) absorption and transmission of UV–Vis radiation, (2) index of refraction, (3) specific optical rotation, and (4) infrared and Raman spectroscopy and NMR. 8.2 ABSORPTION AND TRANSMISSION OF UV–Vis RADIATION Since the packaging industry, more precisely food packaging, plays a dominant role in the short term use of cheap nonbiodegradable petroleum-based materials, their replacement with PLA could provide a significant step toward a greener planet. In order to adequately preserve the quality of the food, the packaging materials have to provide efficient barriers against light, water vapor, atmospheric gases, and volatile organic compounds (VOCs), preventing food degradation and oxidation and preserving aromas and flavors [3]. The absorption and transmission of light by polymers is especially important in the food packaging industry where the packaged goods are light sensitive. Another issue in fresh food packaging is the effect of irradiation in the package since ultraviolet light irradiation is a common method used for lowering microbial population in foods [4]. Sensitive components of foods such as lipids, flavors, vitamins, and pigments may undergo degradation reactions when exposed to light. The spectrum and the intensity of the light source, the conditions of light exposure, and the degree of light transmittance of the packaging material are factors that can dramatically affect the food quality. Thus, packaging Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 97
  • 98 OPTICAL PROPERTIES plays a critical role in prevention of photodegradation of food components during storage. For example, it can slow down adverse reactions, namely, oxidation of fats and oils, formation of sensorially unpleasant volatile compounds (methional, aldehydes, and methyl ketones), loss of vitamins (riboflavin, b-carotene, and vitamin C), production or degradation of free amino acids, increase of the peroxide value, as well as discoloration of pigments, by absorption and reflection of the incident light [5, 6]. The design of the packaging for a specific food product involves not only the choice of the appropriate packaging material but also the addition of the right additives or stabilizers to the packaging in order to provide a more efficient UV–Vis light barrier, and thus a significant improvement in protected food quality after storage when compared to nonprotected food after storage, not to mention extended shelf lives. The transmission of visible light (400–700 nm) and of ultraviolet radiation (100–400 nm) are important parameters in designing the right packaging to preserve and protect products until they reach the consumer. The photochemical degradation of plastics is mainly due to radiation known as UV-B (315–280 nm), because its high-energy content is capable of splitting certain chemical bonds [7]. Although this is an extremely important subject in food packaging applications, to our knowledge, only Auras [8] has measured the PLA visible and ultraviolet light barrier properties and compared them with the properties of commercial polymers traditionally used for food packaging. As can be seen in Figure 8.1 [7], at 225 nm PLA shows a significant increase in UV light transmitted when compared to other standard polymers, reaching about 85% at 250 nm and 95% at 300 nm. Thus, most of the UV-B and UV-A radiation passes through the films. No UV radiation transmission was found in the lower range of UV in 190–220 nm wavelength region. Within the group of conventional polymers, PS and cellophane transmit less radiation in the UV range where most foods are more sensitive and PET does not transmit any light in this wavelength range. LDPE is the polymer that shows the highest transmission of UV light followed by PLA [9]. When the subject of visible radiation is debated, its relationship with colors cannot be forgotten. Each wavelength in the visible light band causes a particular sensation of color. The human eye is not equally sensitive to light emitted at all wavelengths. It is most sensitive to the light in the yellow and green areas of the visible spectrum. When visible light of many frequencies strikes a surface of an object, this object will selectively absorb, reflect, or transmit certain frequencies, thus changing the color perception by the human eye. This selectivity is due to the fact that different atoms and molecules have different natural frequencies of vibration, and they will selectively absorb different frequencies of visible light. Reflection and transmission of incident radiation occur because the frequencies of those light waves do not match the natural frequencies of vibration of the objects. When radiation in these frequencies strikes an object, the electrons in the atoms of the object begin to vibrate. If the object is transparent, then the vibrations of the electrons are passed on to neighboring atoms through the bulk of the material and reemitted on the opposite side of the object. Such frequencies of light waves are said to be transmitted. If the object is opaque, then the vibrations of the electrons are not passed from atom to atom through the bulk of the material. Rather the electrons of atoms on the material’s surface vibrate for short periods of time and then reemit the energy as a reflected light wave. Such frequencies of light are said to be reflected. Transparent materials are materials that allow one or more of the frequencies of visible light to be transmitted through them. Several important parameters are currently used to characterize visible light transmission and color of plastics, such as transparency (ASTM D1746-03 100.0 LDPE 80 PLA 60 PET %T 40 PS Cellophane 20 0.0 190.0 250 300 350 400 450 500 550 600.00 nm FIGURE 8.1 Percent transmission versus wavelength for PLA (98% L-lactide), PS, LDPE, PET, and cellophane films. Adapted from Ref. 7 with permission from Wiley-VCH Verlag GmbH & Co.
  • REFRACTIVE INDEX Standard Test Method for Transparency of Plastic Sheeting) and degree of yellowness (ASTM D6290-05 Standard Test Method for Color Determination of Plastic Pellets), which are based on the absorbance or transmission of UV–Vis light. The transparency, commonly known as ‘‘see through,’’ is defined as the transmission of visible light in the 540–560 nm range [6]. The average transparency of food packaging films is around 95%. The degree of yellowness or the change in the degree of yellowness is a number calculated from spectrophotometric data that describes the change in color of a test sample from clear or white to yellow. PLA, PS, and LDPE have the same degree of yellowness, while cellophane and PET have higher values [7]. Since the light yellow color of PLA products could be a drawback in its application to the packaging industry, creating a consumer perception that the package is old, additives are usually used. In particular, the development of color in PLLA when processed above its melting temperature is a limiting phenomenon in the production of PLLA-based products (implants, fabrics, textiles) by melt processing techniques, such as extrusion and injection molding. This problem adversely affects the properties and end use of the final product making it suitable only for low-grade materials or where color is not an issue. As a matter of fact, discoloration has been considered as one of the most critical problems that arises during melt processing of aromatic polyesters, including poly (ethylene terephthalate) [10, 11] and poly(trimethylene terephthalate) [12]. Wang et al. [13] successfully used UV–Vis spectroscopy for monitoring the process-induced degradation of PLLA during extrusion. UV–Vis spectroscopy provides a powerful and nondestructive tool for real-time detection of the thermal degradation of PLLA, which can be used to optimize the processing conditions since it is very sensitive to minute color changes of the PLLA melt. The observation of the spectra made possible the identification of a redshift of the absorption maximum of a polymer due to new chromophoric groups resulting from thermal degradation. It is known that the absorption maximum of a polymer is shifted to higher wavelengths when the number of conjugated double bonds increases [13, 14]. These authors concluded that pyrolytic elimination is the main degradation mechanism for dry PLLA, and is responsible for the color formation and molecular weight reduction; while an additional hydrolysis mechanism in moist PLLA only reduces molar mass but does not contribute to a change in UV–Vis absorption. A similar study was carried out by Gupta and Deshmukh [15] using PLA in a benzene solution by following the n ! pà transition characteristic of nondegraded PLA that occurs at 287 nm. Nevertheless, a blueshift of this absorption band occurs at 280 nm, suggesting that carbonyl carbon–oxygen bond cleavage is more efficient than other cleavages, which would result in the formation of –COOH groups on the polymer chain ends. Lalla and Chugh [16] dissolved PLA in chloroform and measured the maximum absorbance wavelength. They found this to be 99 at 240 nm and attributed it to the ester group present in the polymer. 8.3 REFRACTIVE INDEX Refractive index is a fundamental physical property of a substance that is often used for its identification, to confirm its purity or to measure its concentration. By definition, the refractive index of a medium is a measure of how much the speed of light is reduced inside the medium. It is a fundamental optical property of polymers that is directly related to other optical, electrical, and magnetic properties. Knowledge about this property is valuable due to its application in the design of new optical polymeric materials. The addition of nanosized inorganic or organic dopants to polymers allows the modification of the polymers’ physical properties enabling the realization of functionalized polymers with new application fields, for example, in microoptics. For example, electron-rich organic dopants, dissolved in polymers, cause a pronounced increase in the refractive index. Polymeric materials have refractive indices that depend on their structure. Lactic acid-based polymers can contain in their structure enantiomers of both L- and D-lactic acid. The most common structures have above 90% L-lactic acid enantiomers and are semicrystalline. Below this percentage, PLA becomes fully amorphous. Hutchinson et al. [1] studied the change in the index of refraction for PLAs by ellipsometric measurements. These authors did not find statistically significant differences in the refractive indices between samples of PLA with different enantiomeric compositions. Nevertheless, a decrease in the index of refraction (1.499–1.448) with the wavelength from 300 to 1300 nm was observed (Figure 8.2). This variation of index of refraction of PLA as a function of wavelength, l (nm), can be described using the well-known Cauchy model given by Equation 8.1. nðlÞ ¼ ð1:445 Æ 0:00075Þ þ 4892 Æ 143 l2 ð8:1Þ Also, many semiempirical group contribution methods derived from the refractive indices of liquid organic compounds as well as organic polymers have been established and give reliable predictions. These group contribution calculations are based on the molecular weight and molecular volume of the monomer, the density of the polymer, and the chemical structure of the polymer. The molar refraction values corresponding to group contribution models such as Lorentz–Lorenz, Gladstone–Dale, Vogel, and Looyenga have been collected extensively by VanKrevelen [17]. Auras et al. [7] reported PLA refractive indices calculated by Lorentz–Lorenz (n ¼ 1.482), Gladstone–Dale (n ¼ 1.492), and Vogel methods (n ¼ 1.482).
  • 100 OPTICAL PROPERTIES 1.50 A = 1.445 ± 7.529 × 10−4 B = 4.8916 × 103 ± 1.426 × 102 Index of refraction 1.49 1.48 1.47 1.46 1.45 1.44 200 400 600 800 1000 1200 1400 Wavelength (nm) FIGURE 8.2 Index of refraction for PLA as a function of wavelength from a global determination of the Cauchy parameters across all optical compositions. Adapted from Ref. 1 with permission from American Chemical Society. Malmgren et al. [18] determined the specific refractive index increment (dn/dc) for PLA with 16.4% of D-isomer, which is amorphous. The resulting dn versus concentration curve was measured using a differential refractometer. The authors gave two similar values for the resulting slope, dn/dc, 0.0237 Æ 0.0034 and 0.0240 Æ 0.0049 mL/g, since the experiments were very difficult to carry out due to air bubbles inside the sample. The obtained dn/dc values for PLA in chloroform are fairly small compared to those of other polymers such as polystyrene, which has a dn/dc of 0.169 mL/g in chloroform. 8.4 SPECIFIC OPTICAL ROTATION The specific optical rotation of a pure material is an intrinsic property of that material at a given wavelength and temperature when dissolved in a particular solvent. The specific optical rotation [a] of PLLA and PDLA polymers was measured in chloroform at a concentration of 1 g/dL at 25 C using a polarimeter and wavelength of 589 nm. The values of a for PDLA and PLA were approximately þ 150 and À150 , respectively [19–21]. Optical rotatory power along the c-axis of PLLA was found positive, while negative along the a-axis. This fact means that right-handed circularly polarized light propagates faster than left-handed circularly polarized light along the helix axis in PLLA. The reverse phenomenon takes place in PDLA. A gyration tensor component g33 of PLLA crystal along the c-axis was found to be extremely large, a few orders of magnitude larger than in the usual optically active crystals. Thus, the helical molecular conformations in helical polymers such as PLLA and PDLA produce huge optical activity. This phenomenon is important for the elucidation of gyro-optical properties of solids and promising for new optical applications utilizing their large optical activity [22]. 8.5 INFRARED AND RAMAN SPECTROSCOPY Infrared spectroscopy is a nondestructive analytical technique used to identify mainly organic materials. In general terms, the state of order of a macromolecular system can be defined by its constitution, configuration, conformation, regularity, stereoregularity, conformational regularity, and crystallinity. Vibrational spectroscopy of a polymer can give information about the state of order by analyzing different types of bands caused by different phenomena such as stereoregularity and conformational regularity of the polymer chain and the crystallinity of the polymer. While the conformational regularity depends on the intramolecular interactions between neighboring chemical groups of the same chain, the crystallinity depends on the intermolecular forces between adjacent chains. It is well known that vibrational spectroscopy is sensitive to local molecular environments, so differences can be observed between these two types of interactions. On the other hand, while these two properties, conformational regularity and crystallinity, are strongly influenced by the pretreatment of the polymer sample and the experimental conditions, the stereoregularity is only affected by chemical reactions [23]. IR and Raman spectroscopies are very important tools for characterization of the chemical and physical nature of polymers. Due to the high sensitivity of IR spectroscopy to changes in the dipole moment of a given vibrating group, this technique is intensively used to identify polar groups. In contrast, Raman spectroscopy is especially helpful in the characterization of the homonuclear polymer backbone due
  • INFRARED AND RAMAN SPECTROSCOPY 101 its sensitivity to changes in polarizability [23]. The characterization of polymers using vibrational spectroscopy is based on empirical interpretation of IR or Raman spectra, since the bands are assigned to the independent vibration of atomic groups in the macromolecule and give information about the structural features of the polymer such as chemical composition, configuration, conformation, and crystallinity. However, a complete theoretical treatment in terms of the vibrational behavior of the polymeric system can only be obtained when the spectra data are obtained from isotopesubstituted polymer analogues and polarization measurements on the specimens. 8.5.1 Infrared Spectroscopy Early studies on PLLA mainly focused on the identification of characteristic bands to investigate the polymer crystallinity. Since Fourier transform infrared (FTIR) spectroscopy is sensitive to the conformation and local molecular environment, this technique has also been used to elucidate the structure of the crystalline polymers. More recently, research on PLLA surface characterization using FTIR has been an object of interest. This section is divided into three parts: structural analysis, surface characterization, and crystallization studies. 8.5.1.1 Structural Analysis: Band Assignment The FTIR spectrum of a polymer in the fingerprint region (n 1500 cmÀ1) is used to identify and characterize the material, since the observed peaks can be assigned to different vibration modes of chemical groups by comparison with cataloged FTIR spectra. Many authors [24–30] have used FTIR to characterize the structure of PLA-based materials (from new composite materials to polymer blends, to copolymers, to the effect of addition of plasticizers, just to mention a few) synthesized by different methods for different purposes and to establish differences in the obtained polymer at different experimental conditions. FTIR spectroscopy for measuring lactide concentration in a PLA matrix has been used by several authors who used different signals to normalize the characteristic lactide absorbance [31–34]. Several important articles have been published concerning the infrared and Raman spectra of PLA-based polymers. For example, one reference work is the pioneering study of Kister et al. [34] on the morphology, conformation, and configuration of PDLA and poly(meso-lactide) stereocopolymers using vibrational spectroscopy. Figure 8.3 shows the IR of the PLA polymers. PLAx is the acronym used to represent homopolymers and stereocopolymers, where x is the percentage of Llactyl units. The differences observed in the several bands in the position, shape, or splitting clearly indicate that FTIR is a powerful technique to study PLA-based polymers. In Figures 8.4 and 8.5, the attenuated total reflection FTIR (ATR-FTIR) spectra of semicrystalline PDLA2:98 and amorphous PDLA20:80 films are presented [36]. FIGURE 8.3 Infrared spectra of poly(L-lactic acid)s: PLA 100 (semicrystalline), PLA 100am (amorphous), PLA 50i (isotactic), PLA 50a (atactic), PLA 50s (syndiotactic), and PLA complex (stereocomplex). (- - -) Band sensitive to the tacticity. Adapted from Ref. 34 with permission from Elsevier. The general band assignments for PLA-based polymers are presented in Table 8.1. The strong IR bands at 2997, 2946, and 2877 cmÀ1 are assigned to the CH stretching region (–CH3(asym), CH3(sym), and CH modes). The C¼O stretching region appears in IR spectra at about 1759 cmÀ1 as a broad asymmetric band mainly due to A and E1 active modes. The CH3 is responsible for the appearance of the band at 1456 cmÀ1. The CH deformation and asymmetric bands appear at 1382 and 1365 cmÀ1. Moreover, the CH bending modes result in the bands at 1315 and 1300 cmÀ1. FIGURE 8.4 Infrared spectra of semicrystalline 98:2 (L:D)poly (lactic acid) [36].
  • 102 OPTICAL PROPERTIES FIGURE 8.5 Infrared spectra of amorphous 80:20 (L:D)poly(lactic acid) [36]. The C–O stretching modes of the ester group appear at 1225 cmÀ1 and the C–O–C asymmetric mode appears at 1090 cmÀ1. At 956 and 921 cmÀ1, we can find the bands characteristic of the helical backbone vibrations with the CH3 rocking modes. At 871 and 756 cmÀ1, appear two bands that can be attributed to the amorphous and crystalline phases of PLA, respectively. The bands that appear below 300 cmÀ1 are mainly due to the CH3 torsion modes and the skeletal C–C torsions [7, 34–37]. 8.5.1.2 Surface Characterization Although transmission IR spectra contain information from both the bulk and the surface of the films, the majority of the signal arises from the bulk due to the much larger amount of polymer chains located there. However, a thorough knowledge of the surface structure of PLA is of great importance, since the surface constitutes a barrier to the surroundings and provides an adsorption site where chemical reactions can occur. A very suitable technique for surface studies of PLA is vibrational sum frequency generation (VSFG) spectroscopy. This is a nonlinear laser spectroscopy technique, which is inherently surface (and symmetry) sensitive under the electric dipole approximation for centrosymmetric systems. A very interesting work on surface segregation and restructuring in PDLA and PLLA films of various thicknesses was presented by Paragkumar et al. [38] using ATR-FTIR. It was observed that PDLA surface segregation and the surface restructuring of methyl side groups are influenced by the polymer film thickness (e.g., poly(DL-lactide) films with thickness of 1 mm do not exhibit surface segregation of methyl side groups). On the other hand, poly(L-lactide) thin and clear films with thickness 15 mm undergo surface conformational changes upon solvent treatment with organic solvents such as toluene, acetone, tetrahydrofuran, and ethyl acetate. The solvent-treated surface of PLLA becomes hazy and milky white and its hydrophobicity increases compared to untreated surfaces. FTIR spectroscopic analysis indicated that polymer chains at the surface undergo certain conformational changes upon solvent treatment. These changes are identified as the restricted motions of C–O–C segments and more intense and specific vibrations of methyl side groups. Another study [39] using VSFG methodology in the CH stretching region demonstrated that this technique can generate valuable information to track changes in the surface and bulk structure of amorphous PLLA and PDLA films, and crystalline PLLA and PDLA:PLLA stereocomplex films. For crystalline PLLA, strong surface and bulk signals caused by both the order and the symmetry of the space group were observed. Amorphous PLLA and racemically composed PDLA were found to consist of a relatively disordered bulk. VSFG spectroscopy has been widely and successfully employed to probe vibrational modes in the high-frequency region (1500–4000 cmÀ1) of the IR spectrum. However, these vibrations constitute highly localized modes. Therefore, such measurements mostly report on very local structural parts of the interfacial molecules (i.e., presence and orientation of –CH3, –OH, and C¼O groups). Recently, developments have been made to access a wider variety of surface chemical groups, such as the amide group that allows for the identification of a-helices and b-sheets (in molecules where amide bonds are present). This type of structural information still depends on the presence of a single chemical group. In contrast to high-frequency localized modes, lowfrequency (skeletal) modes are often composed of the movement of several chemical groups. Thanks to their delocalized nature, these modes are extremely sensitive to the 3D structure of molecules. Very recently, Sugiharto et al. [40] showed that knowledge of the structure of the first few monolayers of the PLLA/air surface of a biodegradable polymer can be achieved by performing femtosecond VSFG spectroscopy and that the bulk and the surface response appears through different order susceptibility elements. In its crystalline form, the backbone of a PLLA polymer is organized in a helical structure. The crystalline structure of PLLA has been determined to belong to the P212121 space group [41–43]. This particular combination of skeletal mode frequencies can be correlated to a helical structure in the polymer skeleton [43, 44]. The 3D surface structure of amorphous PLLA (L-A), crystalline PLLA (L-C), and racemic PDLLA (R) using femtosecond VSFG spectroscopy on delocalized modes in the fingerprint region for PLA was studied [45]. Figure 8.6 displays IR transmission spectra of the L-C, L-A, and R films. The IR spectra show some changes in the fingerprint region, which are characteristic of the three films. Figure 8.7 displays VSFG spectra taken in the vibrational fingerprint region of the same films. The spectra of these three chemically identical films show a large diversity, in contrast
  • INFRARED AND RAMAN SPECTROSCOPY 103 TABLE 8.1 Infrared Spectroscopy Data: Peak Band Assignments for Semicrystalline and Amorphous PLLA Infrared and Raman Spectra IR v(cmÀ1) Semicrystalline PLA I Amorphous PLA Raman v(cmÀ1) I Semicrystalline PLA I 2995 2970 2960 2943 2901 2877 1773 1763 1749 1452 1384, 1388 1363, 1371 1356 1293, 1302, 1315 S sh sh VS w M S S S S M M sh S M M S S S Vw M VS 3571 2997 2997 M 2947 M 2947 M 2882 1760 w VS 2882 1760 w VS 1452 1348, 1388 1368 1360 1300, 1313 1270 1215 1185 1130 1090 1045 960 925 875 S S S S M S VS VS S VS S w w M 1452 1385 1365 1360 1300, 1315 1270 1211 1185 1130 1090 1045 960 S S sh S M S VS VS S VS S sh 873 M 1216 1179 1128 1092 1042 954 923 873 760 740 715 S sh M 760 740 710 S sh M 760 736 711 sh M M 695 M 690 M 515 415 400 350 300 295 w sh M M M sh 415 395 345 300 295 sh M M M sh 240 a w M M 240 M 675 578 520 411 398 347 308 300 251 238 208 158 117 77 60 M w M S S w S w M M M S S S M Amorphous PLA I 2997 S 2942 VS 2877 1769 1755 M S sh 1455 1386 1365 1355 1296, 1300 1264 1216 1183 1128 1092 1042 953 S M M M S sh M M S S S sh 873 790 VS w 740 M 700 682 591 Vw w w 410 397 sh S 316 300 250 238 194 160 116 77 S S M w w S S S Assignment nOH (free) nasCH3 nasCH3 nasCH3 nsCH3 nCH nCH n(C¼O) n(C¼O) n(C¼O) dasCH3 dsCH3 d1CH þ dsCH3 d1CH þ dsCH3 d2CH dCH þ nCOC nasCOC nasCOC rasCH3 nsCOC nC–CH3 rCH3 þ nCC rCH3 þ nCC nC–COO cC¼O dC¼O dC¼O cC¼O cC¼O cC¼O d1C–CH3 þ dCCO d1C–CH3 þ dCCO dCCO dCCO d2C–CH3 þ dCOC dsC–CH3 þ dCOC COC deformation tCC tCC tCC Skeletal torsion Skeletal torsion Skeletal torsion Skeletal torsion The intensity, I, of each band is classified as VS (very strong), S (strong), M (medium), w (weak), sh (shoulder), s (symmetrical), and as (asymmetric) [7, 34–37]. to the IR spectra in Figure 8.6, which reflects the secondary and tertiary structure of the biopolymer at the interface. Such dramatic changes are not observed in Figure 8.6 because in linear spectroscopy, all atomic groups in the film participate in generating the signal so that it is the average bulk structure that is compared and not the interfacial one. Direct comparison between the VSFG spectra and the data in Figure 8.6 is therefore not very meaningful when it comes to determining
  • 104 OPTICAL PROPERTIES 1 R 0.8 Absorbance disordered helices (L-A interface), or rather consist of heterogeneously composed chains (R interface). L-C L-A 0.6 0.4 0.2 0 800 1000 1200 Frequency 1400 1600 (cm–1) FIGURE 8.6 Infrared spectra of L-crystalline (L-C), L-amorphous (L-A), and racemic (R) PLA films. Since the materials are chemically identical, only small differences displaying the average bulk structure are observed. Adapted from Ref. 45 with permission from American Chemical Society. the interfacial structure. Thus, VSFG experiments in the fingerprint region are extremely sensitive to changes in the backbone structure of the outermost polymer monolayer at the polymer/air surface. This previously largely unexplored frequency region allows the determination of whether the interfacial biopolymers are ordered helices (L-C interface), FIGURE 8.7 VSFG spectra of the delocalized modes of L-crystalline (L-C), L-amorphous (L-A), and racemic (R) PLA films, taken with three different IR pulses, which are displayed in the bottom. The black lines are fits to the data in which all contributions to the reflected electrical sum frequency field are added. The grey Lorentzians display the most prominent vibrational modes. The chemical repeat unit of L-PLA is also shown, as well as a molecular model of a 103 helix. In the top left panel, the VSFG experiment is illustrated. Adapted from Ref. 45 with permission from American Chemical Society. 8.5.1.3 Crystallization Studies It is well known that IR is sensitive to the local molecular environment. Accordingly, it has been widely used to explore variations in the intra- and intermolecular interactions and structural changes in macromolecules during melting, crystallization, and phase transition behavior of polymers. Generalized 2D correlation spectroscopy has been applied extensively to analyze IR spectra of polymers for three major reasons: first, it has powerful deconvolution ability for highly overlapped bands; second, it provides information about inter- and intramolecular interactions by correlating absorption band intensities of different functional groups; and third, the intensity changes of a specific sequence occurring during the measurement can be derived from the analysis of asynchronous spectra. In particular, the IR spectra of PLLA polymers are very sensitive to structural changes taking place during melt/crystallization. Depending on the preparation conditions, three different crystalline modifications (a, b, c) can be attained for PLLA. Recently, the fine details of dynamic processes during the crystallization of PLLA have become a matter of keen interest. The orthorhombic structure of the b-form, previously suggested from X-ray studies, was confirmed by trichroic vibrational analysis [43]. The structural evolution and crystallization dynamics of PLLA polymers during isothermal crystallization have been studied by IR spectroscopy. The two-dimensional (2D) correlation analysis of time-dependent IR spectra collected during the melt crystallization process revealed details about the intermolecular interaction of the CH3 and C¼O groups and the conformational changes in the C–O–C backbone that are not easily detected by conventional one-dimensional spectra. It was found that the intermolecular interaction of the CH3 group appears during both the induction period and the growth period of PLLA melt crystallization, while the intermolecular coupling of the C¼O group can only be observed during the crystallization period. The order formation of the C–O–C backbone during the induction period of PLLA melt crystallization can also be clearly observed in the 2D synchronous spectra. These observations show that the weak interchain interactions play an important role in controlling the nucleation and growth of polymer crystallization [2]. Detailed analysis of the three crystallization-sensitive regions is presented below. (a) The C¼0 stretching band region of 1860–1660 cmÀ1. Despite extensive studies on the vibration spectra of PLLA a-crystal, the origin of spectral splitting is still not well interpreted, especially for the nC¼O band. So far, it has been proposed that this splitting of nC¼O band can be attributed to the intramolecular coupling [46] or correlation field splitting arising from
  • INFRARED AND RAMAN SPECTROSCOPY the interchain interactions, such as C–H. . .O hydrogen bonding [47, 48] or dipole–dipole [49]. Hydrogen bonding can be discarded from the analysis of the C–H stretching spectral region [43]. The intramolecular coupling is sensitive to the chain conformation and the distribution of conformers. The four components observed in the C¼O stretching band of semicrystalline PLLA are attributed to the four possible conformers, gt, gg, tt, and tg, while in amorphous PLLA only bands corresponding to gt, gg, and tt conformers were found [46]. The correlation field splitting, also called factor group splitting or Davydov splitting, occurs due to the lateral interaction between the chains contained in the unit cell, splitting the absorption in a number of components. In the case of the orthorhombic unit cell of PLLA, the transition moments of the two adjacent PLLA chains can couple in phase or out of phase, leading to the splitting in the FTIR absorption [46]. (b) CH3, CH bending and the C–O–C stretching band in the region of 1500–1000 cmÀ1. Although in the range of 1500–1000 cmÀ1 the bands are highly overlapped, the band splittings of the CH3 asymmetric deformation mode and C¼O stretching during melt crystallization mode can be clearly observed in time-dependent IR spectra. It was found that the 1458 cmÀ1 band reflects the structural order of the CH3 group, and the band at 1109 cmÀ1 is related to the C–O–C trans-conformation in the crystalline phase of PLLA. From 2D correlation analysis, it can be concluded that CH3 groups form a close interchain contact during the induction period, causing the distortion of the 103 helix conformation of PLLA in a-crystals. On the other hand, the C¼O groups of different PLLA chains do not come into close contact in the induction period [2]. The band at 1193 cmÀ1 is sensitive not only to the structural adjustment of the C–O–C backbone but also to the structural order of the CH3 group in the crystalline phase. From the analysis of the difference spectra and 2D correlation spectra in the 1500–1000 cmÀ1 region, it is shown that the structural adjustment of the CH3 group unambiguously precedes that of the ester group [49]. (c) The skeletal stretching and CH3 rocking band region of 970–850 cmÀ1. The absorption band at 921 cmÀ1 is found to be characteristic of the a-crystals and the 871 cmÀ1 band is sensitive to the 103 helix conformation. The former corresponds to a shorter critical sequence length than the latter. The bands at 955 and 860 cmÀ1 are proportional to the concentration of crystals in the a-form [2]. For polymorphic polymers, such as PLLA, the characteristic FTIR bands can be correlated to the different crystal modifications and typically stay distinguishable in a certain 105 process. This makes it possible to illustrate the mechanism for a polymorphic transition process from the molecular level. It has been reported that when crystallized at this temperature region (100–120 C), a mixture of a0 - and a-crystals is formed. Since the PLLA a0 - and a-crystals show different FTIR spectra [50, 51], the structural changes during the annealing process could be detected by using FTIR spectroscopy. The disordered crystal (a0 -form) of PLLA was found to transform into the a-form during the annealing process at elevated temperatures. The a0 to a transition is very dependent on the annealing period (ta: 0–1440 min) and annealing temperature (Ta: 120–160 C). With increasing Ta, the polymorphic transition progresses much more rapidly. As shown in Figure 8.8, the changes in FTIR spectra upon annealing are mainly associated with the splitting of nC¼O and nC–CH3 indicating that the a0 to a transition mainly involves the slight rearrangement of the chain conformation (especially related to the side groups) and packing manner in the unit cell to the more energy-favorable state, corresponding to the reduction of unit cell dimensions. It was proposed that the a0 to a transformation mainly proceeds by the direct solid–solid transition mechanism, since the direct solid–solid phase transition band (1500–1320 cmÀ1) was observed during the annealing process (Table 8.2). Moreover, it was found that Mw affects the crystalline phase transition significantly. In low molecular weight LLA samples, the a0 to a transition is much faster, and can proceed prominently even when annealed at relatively lower temperature [51, 52]. 8.5.2 Raman Spectroscopy The Raman spectra of PLA polymers are also characterized by a C¼O stretching region. The C¼O stretching mode in PLLA presents four active modes in the Raman region designated by A, B, E1, and E2, which could be observed at 1749, 1763, 1769, and 1773 cmÀ1, as mentioned in Table 8.1. The PDLA Raman spectrum is characterized by broad and asymmetric lines. Two bands at 1769 and 1749 cmÀ1 TABLE 8.2 Assignments for the FTIR Bands in the 1260–1000 cmÀ1 Region for PLLA a0 - and a-Crystals [52] IR Frequencies (cmÀ1) a0 1213 1183 1134 1092 1107 1045 1053 a 1213 1222 1183 1134 1092 1107 1045 Assignments nas(C–O–C) þ ras(CH3) rs(CH3) ns(C–O–C) n(C–CH3)
  • 106 OPTICAL PROPERTIES FIGURE 8.8 FTIR spectra (a) and corresponding second derivatives (b) in the frequency region 1260–1000 cmÀ1 recorded for the normal PLLA118 a-crystal and annealed (at 150 C for various periods (ta/min)) PLLA118 a0 -crystals. Adapted from Ref. 52 with permission from American Chemical Society. appear in both PDLA and PDLLA stereocopolymer spectra. As this region of the spectra is very sensitive to changes in helical chain structure, any perturbation due to the introduction of (D,D) or (D,L) units causes the appearance or disappearance of these bands. Stereocomplexes can be identified by a sharp peak at 1745 cmÀ1 and a broad diffusion band at 1760–1780 cmÀ1. This region proved to be very sensitive to the morphology and conformation. PLLA, PDLA, PLA complex, and poly(meso-lactic) stereocopolymers present CH3 asymmetric deformation modes at about 1450 cmÀ1 in both IR and Raman spectra. Deconvolution analysis of the range between 1250 and 1400 cmÀ1 of the Raman spectra shows three groups of splitting bands at about 1390, 1360, and 1300 cmÀ1, which are assigned, respectively, to the A, B, E (E1 and E2) modes of the CH3 and CH bending region [34]. Once again, it is possible to observe sharp and splitting peaks that characterize the semicrystalline PLAs, namely, PLLA and the PLA complex. As previously observed in the C¼O stretching region, the amorphous state is characterized by asymmetric broad bands. For the (L,D)-PLA stereocopolymers, the bands at $1390 cmÀ1, due to dsCH3 symmetric deformation, and $1300 cmÀ1, due to dCH, are broad and with similar intensity and do not have significant shifts in the frequency. Next, we described the bands assigned to skeletal stretching and the rCH3 rocking region, which appear between 1216 and 1179 cmÀ1. Unlike the IR bands, the symmetric and asymmetric C–O–C modes of the PLLA present low-intensity bands in the Raman spectrum. At 1128 and 1042 cmÀ1, two bands assigned to rasCH3 and nC–CH3 stretching, respectively, can be found. The nC–COO stretching is responsible for the strong band at 873 cmÀ1, which becomes broad and asymmetric for the (L,D)-PLA stereocopolymers and presents a shift to higher frequencies (880 cmÀ1) for PLA complex. It is important to note the presence of a band at 920 cmÀ1 for the PLLA and at 908 cmÀ1 for the PLA complex (semicrystalline polymers), and the absence of any band near these frequencies in amorphous polymers (copolymers). The shifts in frequencies at which these bands appear for the semicrystalline polymers are mainly due to the different crystalline forms of each PLA. PLLA crystallizes preferably on left-handed helices while the crystalline PDLA (stereocomplex) adopts right-handed helices. Since the crystalline structure from the stereocomplex is racemic, the stereocomplex crystallizes in a triclinic unit cell to form a 31 helical conformation known as the b-form. In contrast, the individual polyenantiomers crystallize in a pseudo-orthorhombic system with two 103 helices, which is known as the a-form. The band at 920 cmÀ1 is assigned to a-forms while the other is due to b-forms [53]. Finally, in the low-frequency region, below 800 cmÀ1, two bands, one of them in the range 736–760 cmÀ1 and the other in the range 650–677–711 cmÀ1, are observed for PLLA. The corresponding bands of the PLA complex are sharp and located at slightly higher frequencies and correspond to dC¼O and cC¼O. The bands of the dCCO mode appear between 398 and 411 cmÀ1. For the PLA complex, the dCOC skeletal chain deformation band appears in the range of frequencies (291–309 cmÀ1) as a split line while bands characteristic of torsion modes are found at 239, 206, and 160 cmÀ1. For PLLA, the associated bands are located at 230, 210, and 160 cmÀ1. Raman spectra are more sensitive to modifications of chain morphology below 600 cmÀ1 [34, 35] as we can see in Figure 8.9. Polarized Raman spectroscopy has been regarded as a powerful tool to quantify molecular orientation distributions,
  • INFRARED AND RAMAN SPECTROSCOPY be determined independently. The PLLA Raman bands are assumed to be cylindrically symmetric owing to its helical molecular structure. A band at 926 cmÀ1 was assigned to the crystalline regions only of PLLA, whereas another band at 875 cmÀ1 was assigned to both crystalline and amorphous regions. The PLLA molecules were biaxially oriented in both amorphous and crystalline regions. The orientation distribution normal to the surface of the film was found to be broader in the amorphous regions than in the crystalline regions. Furthermore, a predominant unidirectional molecular orientation was observed in the crystalline region, whereas an isotropic molecular orientation distribution parallel to the surface was found for the amorphous phase [54]. Smith et al. [55] characterized using Raman the orientation of PLA films from NatureWorks used for food packaging and marketed by Cargill Dow Polymers. However, the study was carried out monitoring the changes in bands assigned to crystalline and amorphous phase as well as its orientation. The uniaxial and biaxial oriented films were characterized by Raman depolarization and Raman band shifts observed, respectively. Figure 8.10 shows a Raman spectrum of an unoriented film. The bands between 415 and 398 cmÀ1 should receive attention due to their relationship with the C–C–O bonds in the backbone. The band at 873 cmÀ1 can be assigned to a stretch of the C–C bond. There are two bands at 397 and 410 cmÀ1 with extreme sensitivity and selectivity to crystallization and orientation, which can be used to follow crystallization and to characterize orientation in the crystalline phase. The first band is assigned to the amorphous phase and the other is a crystalline band. When these films become oriented, these bands have marked shifts depending on the draw ratio, as can be seen in Figure 8.11 [55]. FIGURE 8.9 Raman spectra of poly(L-lactic acid)s: PLA 100 (semicrystalline), PLA 100am (amorphous), PLA 50i (isotactic), PLA 50a (atactic), PLA 50s (syndiotactic), and PLA complex (stereocomplex). (- - -) Band sensitive to the tacticity. Adapted from Ref. 34 with permission from Elsevier. 1452.7 1127.6 1043.6 687.17 740.15 411.45 300.74 197.37 873.94 since it is capable of obtaining the fourth-order molecular orientation distribution coefficients, as well as the second order. This technique enables the molecular orientation distribution of the crystalline and the amorphous region to 0 200 400 600 800 107 1000 1200 1400 1600 1800 FIGURE 8.10 Raman spectrum of a 4.1% D-PLA unoriented cast film. Adapted from Ref. 55 with permission from John Wiley & Sons, Inc.
  • 108 OPTICAL PROPERTIES O Amorphous O CH3 C (a) 2× CH3 C C C O H C terminus H Observed stereocenter O 3× CH3 O CH3 C C C C 4× (b) O H Observed stereocenter 13C 440 420 400 380 FIGURE 8.11 Raman spectrum of a 4.1% D-PLA-oriented cast film. Adapted from Ref. 55 with permission from John Wiley & Sons, Inc. 1 H AND 13 O H O terminus i s i 360 Raman shift (cm–1) 8.6 O C NMR SPECTROSCOPY Nuclear magnetic resonance (NMR) spectroscopy has been used as a fundamental tool to understand the molecular structure of a wide variety of compounds. In polymer science, its applicability has proven to be extremely important for solid samples. The Zeeman interactions that happen between nuclei determine the average resonance frequency of a particular nuclear species, allowing the observation of specific elements without interferences of others present in the sample [56]. In the NMR spectra of PLA, the observed resonances can be assigned to stereosequence distribution in the polymer and reflect its history including the stereochemistry of the feed composition, polymerization kinetics, and extent of transesterification and racemization [57–60]. The assignments are designated as various combinations of ‘‘i’’ isotactic pairwise relationship (RR and SS) and ‘‘s’’ syndiotactic pairwise relationship (RS and SR). In the NMR spectra, the combinations RR and SS are indistinguishable and have identical chemical shifts, as would RS and SR [57, 61]. Zell et al. [61] assigned both the 1 H and 13 C NMR spectrum of the methine proton and carbon in PLA at the tetrad stereosequence level. They used a combination of twodimensional NMR experiments and selective 13 C labeling and showed that the central pairwise relationship in the 1 H NMR spectrum is determined by the stereocenter in the lactic acid unit attached to the O terminus. They also determined that in the 13 C NMR spectrum the central pairwise relationship of the stereocenter in the lactic acid unit attached to the C terminus obtained consistent NMR and statistical data as can be seen in Figure 8.12. The chemical shifts of 13 C and 1 H nuclei in PLA are affected by the stereoconfiguration of two or three adjacent stereogenic centers on either side (hexad stereosensitivity) [62]. (c) ... R R R R S S R R R R ... s i s 1H FIGURE 8.12 Diagram illustrating the alternative interpretation of 1 H and 13 C stereosequences: (a) central pairwise relationship determined by lactic acid connected to C terminus, (b) central pairwise relationship determined by lactic acid connected to O terminus, (c) direction of central pairwise relationship of 1 H and 13 C resonances. Adapted from Ref. 61 with permission from American Chemical Society. Kricheldorf et al. [59] proposed assignments for the stereosequences of the known microstructures of PLA, which were more recently confirmed by Zell et al. [61]. Figure 8.13 presents the 1 H and 13 C spectra of PLA synthesized using 5% L-lactic and 95% D-lactic obtained by Zell et al. [61]. In the 1 H NMR spectrum, the sis resonance, which appears at 5.24 ppm, gives an integrated intensity of 2.64%. Integration of the resonance at 69.21 ppm (the isi resonance) in the 13 C NMR spectrum gives a value of 6.19% [61]. When the PLA is synthesized using primarily 13 D-lactide and small amounts of L-lactide that was C labeled in the carbonyl position, only two peaks can be observed in the carbonyl region due to the 13 C labels, which confirms that almost all the pairs of S stereocenters are surrounded on either side by several R stereocenters [61]. The NMR spectra of PLA synthesized using (a) fully 13 C labeled and (c) unlabeled 5% L-lactide and 95% meso-lactide are shown in Figure 8.14, along with (b) the NMR spectrum of poly (meso-lactide). The spectra in Figure 8.14b and c are almost identical since the stereosequence distributions for 5% Llactide/95% meso-lactide and for 100% meso-lactide are almost identical. The NMR spectrum in Figure 8.14a is very different from the spectra in Figure 8.14b and c. The reason is that the signals from the fully 13 C-labeled S stereocenters dominate the spectrum, which is similar to the situation observed in the one-dimensional 13 C NMR spectrum. As a result of the splitting of the signals for the 13 C-labeled S stereocenters into a doublet of doublets, the spectrum of the labeled polymer spans a larger chemical shift range (68.5–69.5 ppm) than the unlabeled polymer
  • 1 H AND 13C NMR SPECTROSCOPY 109 94.42 iii isi iis/sii 1H sis / 2.64 2.94 5.25 5.30 5.20 5.25 5.10 5.05 ppm 93.81 sis, iii, iis/sii 6.19 isi 13C 69.4 69.2 69.0 68.8 ppm FIGURE 8.13 1 H and 13 C solution NMR spectra of PLA synthesized using 5% L-lactide and 95% D-lactide. Adapted from Ref. 61 with permission from American Chemical Society. (69.0–69.4 ppm). The center of each set of doublet of doublets corresponds to the expected chemical shift value for an unlabeled carbon. The sis stereosequence is expected from both the unlabeled and labeled stereocenters. For the unlabeled meso-lactide component, an isi stereosequence is expected, while for the 13 C-labeled stereocenters, an isi stereosequence is not expected. In the unlabeled meso-lactide region, the sis resonance appears at 69.0 ppm and is not split, while in the 13 C-labeled L-lactide region, the sis resonance is centered at 69.0 ppm, but it is split into a doublet of doublets that extends $0.4 ppm in both directions, spanning the region from 68.6 to 69.4 ppm. In addition, all of the other resonances expected due to the 13 C-labeled L-lactide lie within 0.1 ppm of the sis resonance, making the total span of all resonances due to 13 C-labeled L-lactide from 68.5 to 69.5 ppm (Figure 8.14). This experiment definitively proves that the peak at 69.0 ppm in the 13 C NMR spectrum is due to sis. If the isi and sis stereosequence assignments were reversed, then the sis peak would be centered at 69.2 ppm and would span from 68.8 to 69.6 ppm. The solid line in Figure 8.14 shows where the sis peak would be centered, the short-dashed line shows where the iss/ssi peak would be centered, and the two long-dashed lines show where the isi peak would be centered and where the farthest peak in that doublet of doublets should be. Since there is no resonance at FIGURE 8.14 13 C NMR spectra of PLA synthesized using (a) 5% fully 13 C-labeled L-lactide and 95% meso-lactide, (b) meso-lactide, and (c) 5% L-lactide and 95% meso-lactide. Adapted from Ref. 61 with permission from American Chemical Society. 69.6 ppm, the assignment of the peak at 69.0 ppm to sis and the peak at 69.2 ppm to isi must be correct. It is also possible to definitively assign the iss/ssi stereosequences to specific resonances in the 13 C NMR spectrum. For a polymer that is synthesized with no isotopically labeled material, both the iss and ssi resonances will always have equal probability and therefore cannot be distinguished in the NMR spectrum. For PLA synthesized using 5% fully 13 C-labeled L-lactide and 95% meso-lactide, both the ssi and iss stereosequences are expected from the large amount of meso-lactide, and these resonances will not be split in the NMR spectrum. When PLA is synthesized using 5% fully 13 C-labeled L-lactide and 95% meso-lactide, only the ssi stereosequence will be split into a doublet of doublets because of 13 C–13 CJ coupling to the methyl and carbonyl carbons. No peak due to iss will be observed from the 13 C-labeled stereocenters. If the ssi resonance is centered at 69.1 ppm, the doublet of doublets will overlap the doublet of doublets generated by the iii, iis, sii, and sis stereosequences, and the spectrum will span from about 68.7 to 69.5 ppm (Figure 8.6). If the ssi resonance occurs at 69.4 ppm, the doublet of doublets will span from 69 to 69.8 ppm. The 13 C NMR spectrum in Figure 8.14a has no
  • 110 OPTICAL PROPERTIES FIGURE 8.15 Left: methine resonances in the 13 C NMR spectra of poly(lactide) samples (a) 1, (b) 2, and (c) 3. The peak at about 68.95 ppm comprises resonances from iii, iis, sii, and sis core stereosequences. Right: methine resonances in the homonuclear decoupled 1 H NMR spectra of poly (lactide) samples (a) 1, (b) 2, and (c) 3. Adapted from Ref. 62 with permission from American Chemical Society. resonance above 69.5 ppm, indicating that the resonance at 69.1 ppm in Figure 8.14b is due to ssi, while the resonance at 69.4 ppm is due to iss [61]. Thakur et al. [62] studied PLA samples with various compositions of L-lactide, D-lactide, and meso-lactide: sample 1 had 3% of both L-lactide and D-lactide and 94% of mesolactide; sample 2 had 51.5% of L-lactide, 1.5% of D-lactide, and 47% of meso-lactide; and sample 3 had 70.9% L-lactide, 0.9% D-lactide, and 28.2% meso-lactide. They found that the carbonyl resonance in 13 C spectrum of PDLLA indicates a higher stereoselectivity than hexad and is most likely to be octad [62]. However, due to the high degree of overlap, the assignments of the peaks for octad stereosequences to carbonyl resonance and probably a hexad stereosensitivity to methyl resonance are difficult. Based on hexad stereosequences, the authors [62] interpreted the 13 C methine resonances and the stereosequences sss, isi, ssi, and iss are well resolved and their assignments are consistent with previous studies made by Kricheldorf [59], Schindler and Harper [63], or Kasperczyk [58]. Although 1 H NMR allows better results regarding stereosequence probabilities, there are still some questions regarding the coupling between the methyl protons and the methine (–CH) protons at each of the stereogenic centers in PLA. As was observed by analysis of 13 C spectra, also in 1 H spectra the peaks corresponding to hexad stereosequences can be observed (Figure 8.15). The intensity distribution of the various stereosequence resonances in the NMR spectra indicates a preference for syndiotactic addition during the polymerization process [62]. However, this preference decreases with increasing extent of polymerization. Steric hindrance at the polymer growing site is probably responsible for the syndiotactic stereospecificity, and the increasingly random lactide addition is due to interplay of kinetically and thermodynamically controlled reactions. Changes in viscosity during the melt polymerization additionally influence the stereochemistry. The effect of transesterification on the stereosequence distribution should be related to increasing time and hence increasingly randomize with increasing polymerization time until the stereosequence distribution in the polymer becomes random. The limiting invariant stereosequence intensity in PLA near equilibrium shows that transesterification and racemization are not frequent enough to influence the stereosequence distribution under the polymerization conditions [64].
  • REFERENCES REFERENCES 1. M. H. Hutchinson, J. R. Dorgan, D. M. Knauss, S. B. Hait, J. Polym. Environ. 2006, 14, 119–124. 2. J. M. Zhang, H. Tsuji, I. Noda, Y. Ozaki, J. Phys. Chem. B 2004, 108, 11514–11520. 3. N. S. Oliveira, J. Oliveira, T. Gomes, A. Ferreira, J. Dorgan, I. M. Marrucho, Fluid Phase Equilibr. 2004, 222, 317–324. 4. K. L. G. Ho, A. L. Pometto, J. Environ. Polym. Degrad. 1999, 7, 93–100. 5. K. N. Turhan, F. Sahbaz, Polym. Int. 2001, 50, 1138–1142. 6. M. Mathlouthi, Food Packaging and Preservation, Blackie, 1994, XVI þ 275 pp. 7. R. Auras, B. Harte, S. Selke, Macromol. Biosci. 2004, 4, 835–864. 8. R. Auras, Ph.D. thesis, Michigan State University, East Lansing, MI, 2004, p. 268. 9. J. Lunt, D. W. Farrington, S. Davies, R. S. Blackburn, Poly (lactic acid) fibers, in: Biodegradable and Sustainable Fibres, Woodhead Publishing Ltd., 2005, Chapter 6. 10. C. F. L. Ciolacu, N. R. Choudhury, N. K. Dutta, Polym. Degrad. Stabil. 2006, 91, 875–885. 11. M. Edge, R. Wiles, N. S. Allen, W. A. McDonald, S. V. Mortlock, Polym. Degrad. Stabil. 1996, 53, 141–151. 12. J. Ramiro, J. I. Eguiazabal, J. Nazabal, J. Appl. Polym. Sci. 2002, 86, 2775–2780. 13. Y. M. Wang, B. Steinhoff, C. Brinkmann, I. Alig, Polymer 2008, 49, 1257–1265. 14. F. Sondheimer, D. A. Benefraim, R. Wolovsky, J. Am. Chem. Soc. 1961, 83, 1675–1681. 15. M. C. Gupta, V. G. Deshmukh, Colloid Polym. Sci. 1982, 260, 514–517. 16. Chugh. N. N., Lalla J. K., Indian Drugs 1990, 27, 516–522. 17. D. W. VanKrevelen, Properties of Polymers, 3rd edition, Elsevier Science B.V., Amsterdam, The Netherlands, 1997. 18. T. Malmgren, J. Mays, M. Pyda, J. Therm. Anal. Calorim. 2006, 83, 35–40. 19. H. Tsuji, S. H. Hyon, Y. Ikada, Macromolecules 1991, 24, 5651–5656. 20. N. Yui, P. J. Dijkstra, J. Feijen, Makromol. Chem. 1990, 191, 481–488. 21. H. R. Kricheldorf, R. Dunsing, Makromol. Chem. 1986, 187, 1611–1625. 22. J. Kobayashi, T. Asahi, M. Ichiki, A. Oikawa, H. Suzuki, T. Watanabe, E. Fukada, Y. Shikinami, J. Appl. Phys. 1995, 77, 2957–2973. 23. H. W. Siesler, K. Holland-Moritz, Infrared and Raman Spectroscopy of Polymers, 1980, VIII þ 389 pp. 24. R. Mehta, V. Kumar, H. Bhunia, S. N. Upadhyay, J. Macromol. Sci. Polym. Res. 2005, C45, 325–349. 25. M. Matsusaki, A. Kishida, N. Stainton, C. W. G. Ansell, M. Akashi, J. Appl. Polym. Sci. 2001, 82, 2357–2364. 26. M. H. Gutierrez-Villarreal, M. G. Ulloa-Hinojosa, J. G. GaonaLozano, J. Appl. Polym. Sci. 2008, 110, 163–169. 111 27. W. H. Song, Z. Zheng, W. L. Tang, X. L. Wang, Polymer 2007, 48, 3658–3663. 28. M. T. Khorasani, H. Mirzadeh, S. Irani, Radiat. Phys. Chem. 2008, 77, 280–287. 29. Y. S. Kim, E. S. Gil, T. L. Lowe, Macromolecules 2006, 39, 7805–7811. 30. Z. J. Ren, L. S. Dong, Y. M. Yang, J. Appl. Polym. Sci. 2006, 101, 1583–1590. 31. B. Braun, J. R. Dorgan, S. F. Dec, Macromolecules 2006, 39, 9302–9310. 32. P. Degee, P. Dubois, S. Jacobsen, H. G. Fritz, R. Jerome, J. Polym. Sci. Polym. Chem. 1999, 37, 2413–2420. 33. M. Jalabert, C. Fraschini, R. E. Prud’Homme, J. Polym. Sci. Polym. Chem. 2007, 45, 1944–1955. 34. G. Kister, G. Cassanas, M. Vert, Polymer 1998, 39, 267–273. 35. G. Kister, G. Cassanas, M. Vert, B. Pauvert, A. Terol, J. Raman Spectrosc. 1995, 26, 307–311. ´ 36. C. M. B. Gon¸ alves, L. C. Tome J. A. P. Coutinho, I. M. c Marrucho, accepted in J. Appl. Polym. Sci. 37. D. Garlotta, J. Polym. Environ. 2001, 9, 63–84. 38. N. T. Paragkumar, E. Dellacherie, J. L. Six, Appl. Surf. Sci. 2006, 253, 2758–2764. 39. C. M. Johnson, A. B. Sugiharto, S. Roke, Chem. Phys. Lett. 2007, 449, 191–195. 40. A. B. Sugiharto, C. M. Johnson, H. B. De Aguiar, L. Alloatti, S. Roke, Appl. Phys. Part B 2008, 91, 315–318. 41. C. Aleman, B. Lotz, J. Puiggali, Macromolecules 2001, 34, 4795–4801. 42. S. Sasaki, T. Asakura, Macromolecules 2003, 36, 83858390. 43. K. Aou, S. L. Hsu, Macromolecules 2006, 39, 3337–3344. 44. S. H. Kang, S. L. Hsu, H. D. Stidham, P. B. Smith, M. A. Leugers, X. Z. Yang, Macromolecules 2001, 34, 4542–4548. 45. A. B. Sugiharto, C. M. Johnson, I. E. Dunlop, S. Roke, J. Phys. Chem. C 2008, 112, 7531–7534. 46. E. Meaurio, E. Zuza, N. Lopez-Rodriguez, J. R. Sarasua, J. Phys. Chem. B 2006, 110, 5790–5800. 47. J. R. Sarasua, N. L. Rodriguez, A. L. Arraiza, E. Meaurio, Macromolecules 2005, 38, 8362–8371. 48. J. M. Zhang, H. Sato, H. Tsuji, I. Noda, Y. Ozaki, Macromolecules 2005, 38, 1822–1828. 49. J. M. Zhang, H. Tsuji, I. Noda, Y. Ozaki, Macromolecules 2004, 37, 6433–6439. 50. P. Pan, W. Kai, B. Zhu, T. Dong, Y. Inoue, Macromolecules 2007, 40, 6898–6905. 51. J. M. Zhang, Y. X. Duan, H. Sato, H. Tsuji, I. Noda, S. Yan, Y. Ozaki, Macromolecules 2005, 38, 8012–8021. 52. P. J. Pan, B. Zhu, W. H. Kai, T. Dong, Y. Inoue, Macromolecules 2008, 41, 4296–4304. 53. H. Bourque, I. Laurin, M. Pezolet, J. M. Klass, R. B. Lennox, G. R. Brown, Langmuir 2001, 17, 5842–5849. 54. M. Tanaka, R. J. Young, Biomacromolecules 2006, 7, 2575–2582.
  • 112 OPTICAL PROPERTIES 55. P. B. Smith, A. Leugers, S. H. Kang, X. Z. Yang, S. L. Hsu, Macromol. Symp. 2001, 175, 81–94. 56. Characterization of Solid Polymers New Techniques and Developments, 1994, XII þ 368 pp. 57. K. A. M. Thakur, R. T. Kean, E. S. Hall, M. A. Doscotch, E. J. Munson, Anal. Chem. 1997, 69, 4303–4309. 58. J. E. Kasperczyk, Macromolecules 1995, 28, 3937–3939. 59. H. R. Kricheldorf, C. Boettcher, K. U. Tonnes, Polymer 1992, 33, 2817–2824. 60. H. R. Kricheldorf, I. Kreisersaunders, C. Boettcher, Polymer 1995, 36, 1253–1259. 61. M. T. Zell, B. E. Padden, A. J. Paterick, K. A. M. Thakur, R. T. Kean, M. A. Hillmyer, E. J. Munson, Macromolecules 2002, 35, 7700–7707. 62. K. A. M. Thakur, R. T. Kean, E. S. Hall, J. J. Kolstad, T. A. Lindgren, M. A. Doscotch, J. I. Siepmann, E. J. Munson, Macromolecules 1997, 30, 2422–2428. 63. A. Schindler, D. Harper, J. Polym. Sci. Part C 1976, 14, 729–734. 64. K. A. M. Thakur, R. T. Kean, E. S. Hall, J. J. Kolstad, E. J. Munson, Macromolecules 1998, 31, 1487–1494.
  • 9 CRYSTALLIZATION AND THERMAL PROPERTIES LUCA FAMBRI 9.1 AND CLAUDIO MIGLIARESI INTRODUCTION Poly(lactic acid) (PLA)-based products have been largely considered for both biomedical and environment-friendly applications [1–9]. Processing conditions and final properties of PLAs are strongly determined/influenced by the crystalline or amorphous structure of the polymer [10–12]. Starting from the early work of Kulkarni et al. [13, 14] in 1966 and 1971, many papers addressed the evaluation of thermal properties and crystallinity of PLA or polylactidebased polymers. The chiral nature of lactic acid results in distinct forms of polylactide, namely, poly(L-lactide) (PLLA), poly(D-lactide) (PDLA), and poly(DL-lactide) (PDLLA), which are synthesized from the L-, D-, and DL-lactic acid monomers, respectively, or from the corresponding L,L-lactide, D,D-lactide, and DL-lactide, respectively [15]. Both L- and D-lactic acid stereoisomers are naturally occurring; however, most of the lactic acid in nature is L-type or sometimes racemic. The fact that lactic acid that is produced in the human body is in the L-enantiomeric form and the interest in the biomedical applications of this polymer have led both research and production to concentrate on L-lactide or DL-lactide polymers [16–18]. The D-isomer does not have many applications, except for use in particular medicinal chemicals. Today petrochemical-based production gives a 50/50 mixture of the L- and D-forms, whereas from fermentation mostly L-lactic acid is obtained. Recently, new research has been reported on the production of D-lactic acid from rice fermentation [19]. PLLA and PDLA are crystalline polymers due to the enantiomeric purity of the pristine monomers and the stereoregularity of the polymer chain. Conversely, PDLLA, which is an equimolar relatively random copolymer of L- and D-lactic acid (or L- and D-lactide), is fully amorphous because of its irregular structure. The stereochemical configuration of PLLA (or PDLA) and PDLLA is analogous to that of isotactic polypropylene and atactic polypropylene. Syndiotactic PLA that can be polymerized from a stereoselective polymerization of meso-lactide can be considered as an alternating DL-lactic acid copolymer, analogous to syndiotactic polypropylene [20]. Worth noting is the fact that in the early work of Kulkarni et al. [14] the glass transition peak of PDLLA at about 60 C was misinterpreted and attributed to melting. Pure PLLA and PDLA have the same properties, that is, a glass transition temperature (Tg) between 50 and 70 C, a melting temperature (Tm) between 170 and 190 C, and a crystallinity of around 35% [1, 2, 7, 8, 21–28]. In the case of syndiotactic polylactide, due to the different stereotactic configuration, lower glass transition temperature and melting temperature were found to be 34 and 155 C, respectively [20]. Crystallization, crystallinity degree, and thermal properties of PLLA depend on the polymer molecular weight, polymerization conditions, thermal history, purity, and so on. As reported by Ikada et al. [25], blending of PLLA and PDLA results in the formation of a stereocomplex with a crystalline structure different from that of each homopolymer and melting temperatures that reach 230 C [25]. Contradictory data are reported about PLLA melting enthalpy, ranging in the literature from 40 to 203 J/g. The most common values adopted for the melting enthalpy of PLLA are 93.6 J/g, as reported by Fisher et al. [21], or 140 J/g, as evaluated by Loomis et al. [29]. Many authors have studied the crystalline structure of PLLA [30–45]. PLLA crystals present three different Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 113
  • 114 CRYSTALLIZATION AND THERMAL PROPERTIES structural conformations that can develop under different processing or treatment conditions [42]. In particular, melt or cold crystallization and solution spinning processes at low drawing temperatures and/or low hot draw ratios induce the formation of a-structures characterized by a left-handed 103 helix that packs in an orthorhombic unit cell with parameters a ¼ 1.06 nm, b ¼ 1.737 nm, and c ¼ 2.88 nm [46]. Stretching, solution spinning, or high hot draw ratios induce the formation of b-crystal structures, a unit cell with a ¼ 1.031 nm, b ¼ 1.821 nm, and c ¼ 0.900 nm, and a chain conformation with left-handed 31 helices [47]. More recently, an additional crystalline form named c that develops under epitaxial crystallization has been reported [48]. 9.2 CRYSTALLINITY AND CRYSTALLIZATION Figure 9.1 shows a typical differential scanning calorimetry (DSC) analysis of thermal behavior of two amorphous PLAs, that is, a PDLLA (Mw ¼ 70 kDa) sample that is intrinsically amorphous and PLLA (Mw ¼ 200 kDa) that was quenched to the amorphous state by fast cooling at À100 C/min after melting. In both cases, the Tg is evident and is located at about 65 C. Above Tg, PDLLA exhibits an aging peak in the interval 65–90 C with an enthalpy of 8 J/g, typical of amorphous polymer aging. The longer the aging time and/or the closer the aging temperature to the Tg, the more intense is that peak. PDLLA can reach an enthalpy value of about 10 J/g, as shown by Celli and Scandola [28]. On the other hand, the initial amorphous PLLA presents a broad crystallization peak in the range 100–160 C centered at about 132 C with DHc ¼ 38 J/ g, followed by melting at 182 C with DHm ¼ 38 J/g. As determined by the equal values of the crystallization and melting enthalpies, the fast quenching produced a fully amorphous material. The variation of specific heat (DCp) at Tg was 0.50 J/(g K). FIGURE 9.2 DSC thermal cycles of PLLA 200 kDa (first heating, cooling, and second heating at Æ10 C/min). Figure 9.2 refers to a commercial PLLA (L210 from Boehringer, Ingelheim, Germany) with a molecular weight of 200 kDa, subjected to the following three DSC runs at 10 C/min: (i) heating from 0 to 220 C, (ii) cooling from 220 to 0 C, and finally (iii) heating from 0 to 220 C. The initially crystalline polymer presented a Tg at 73 C and melted at 192 C. During cooling, apparently no crystallization developed and the following thermal run on the initially amorphous materials showed glass transition (at Tg ¼ 64 C), crystallization (at Tc ¼ 127 C), and melting (at Tm ¼ 181 C). Worth noting between the curves of Figures 9.1 and 9.2 is the difference in the crystallization temperature, which was lower for the slowly cooled polymer, probably due to the formation of crystallization nuclei during cooling. Table 9.1 lists Tg, Tm, Tc, DHc, and DHm of various molecular weight PLLA samples, namely, 2, 30, and 200 kDa, respectively, subjected to the same DSC cycles presented in Figure 9.2. The crystallinity content (X) has been evaluated from the DSC data according to the following equation: 0 X% ¼ 100ðDHm ÀDHc Þ=DHm FIGURE 9.1 Differential scanning calorimetry thermograms of amorphous PLLA and PDLLA (heating rate 10 C/min). ð9:1Þ where DHm and DHc are the melting and the crystallization 0 enthalpies, respectively, and DHm is the reference DHm (93.6 J/g) for PLLA crystals having an infinite size [21].
  • CRYSTALLINITY AND CRYSTALLIZATION 115 TABLE 9.1 Glass Transition Temperature (Tg) and Change of Specific Heat at Tg (DCp), Melting Peak Temperature (Tm) and Enthalpy (DHm), and Crystallization Peak Temperature (Tc) and Enthalpy (DHc) for Different Molecular Weight of PLLA Samples During a DSC Cycle of Heating–Cooling–Heating Between 0 and 220 C at Æ10 C/min (i) First Heating Tg ( C) MW (kDa) 2 30 200 Tm ( C) 58 79 73 DHm (J/g) X (%) 147 171 192 DCp (J/(g K)) 48 68 67 51 73 72 0.17 0.06 0.22 (ii) Cooling MW (kDa) Tc ( C) 2 30 200 – 96 – – 21 – (iii) Second Heating DCp (J/(g K)) 39 50 55 0.56 0.32 0.50 Tg ( C) DCp (J/(g K)) Tc ( C) DHc (J/g) Tm ( C) DHm (J/g) X (%) 43 55 64 MW (kDa) 2 30 200 Tg ( C) DHc (J/g) 0.54 0.32 0.51 107 97 127 9 21 25 142 174 181 9 50 26 0 31 1 Due to the favorable crystallization conditions during the polymerization process, all the products in the first scan display a high crystallinity between 50% and 73%, depending on their molecular weight, and melting temperature that is proportional to the molecular weight according to the classical Flory equation * 1=Tm ¼ 1=Tm À2RM0 =DHmu Mn ð9:2Þ * where Tm is the melting temperature, Tm is the melting temperature at infinite molecular weight, R is the gas constant, DHmu is the heat of fusion per mole of the repeating unit, M0 the molecular weight of the repeating unit, and Mn is the number-average molecular weight. The Flory–Fox equation * Tg ¼ Tg ÀK=Mn ð9:3Þ can also be used to predict the Tg of polymers as a function of * the molecular weight, Tg being the Tg at infinite Mn. While the first heating Tg values are higher than expected due to high crystallinity content (crystalline domains hinder the mobility of amorphous chains), as also confirmed from the low value of DCp at Tg, Tg measured during cooling proportionally depend on molecular weight, in accordance with Equation 9.3. It is also interesting to note that during cooling PLLA (30 kDa) developed some crystallinity, whereas in the other cases crystallization was inhibited by the high content of terminal groups that acted as defects in PLLA (2 kDa) and by the low mobility of the long PLLA (200 kDa) chains. During the second heating, all the polymers evidenced glass transition, crystallization, and melting. Tg and Tm directly depend on molecular weight. The different Tg values measured during cooling and the following heating cycles of the amorphous PLLAs, that is, 39 versus 43 C for the 2 kDa PLLA and 55 versus 64 C for the 200 kDa PLLA, are due to a kinetic effect that depends on the molecular weight of the polymer. The lower Tc of PLLA (30 kDa) is attributable to its crystallinity before the second heating cycle that acted as a nucleating agent and promoted further crystallization, up to 53%, as determined from the melting enthalpy. In the case of PLLA (2 kDa) and PLLA (200 kDa), the lower DHm and crystallinity of about 10% and 28%, respectively, once again directly depend on the effect of short and long polymer chains. As a function of molecular weight, however, the same or different thermal treatments can produce different results, as clearly shown by Migliaresi et al. [27, 32] and summarized in Figure 9.3. The crystallinity of the polymers shown in Figure 9.3 increases with annealing time and with decreasing viscometric molecular weight (Mv). Almost slowly and progressively increasing constant values were reached faster by the lower molecular weight PLLAs. This increase, as discussed later in the chapter, could be attributed not only to kinetic effects but also to thermal degradation of Mv during annealing. This effect has not often been considered in the literature.
  • 116 CRYSTALLIZATION AND THERMAL PROPERTIES FIGURE 9.3 Effect of annealing time on crystallinity of compression molded PLLA at 160 C (adapted from Ref. 32). Migliaresi et al. [27] showed that low molecular weight PLLA (Mv ¼ 18 kDa) was able to crystallize regaining almost 65%, 79%, and 87% of its initial crystallinity, during cooling after full melting at À5, À1, and À0.5 C/min, attaining crystallinity contents of 38.5%, 46.8%, and 51.3%, respectively. Higher Mv PLLA (156 and 425 kDa) developed only a little crystallinity when cooled at À5 C/ min. As pointed, mobility, and hence crystallization ability, is determined, at any given molecular weight, by the temperature of the system. Their data confirmed that same thermal histories lead to materials with substantially different crystallinity amounts as a function of the PLLA molecular weight. More important perhaps, Migliaresi et al. [27] evidenced that any treatment always caused molecular weight degradation, as shown in Table 9.2. This also means that, at a given initial molecular weight, thermal cycles at different heating or cooling rates are representative of polymers having actually different viscometric molecular weight Mv. The effect of molar mass on PLLA Tg and Tm and on PDLLA Tg is clearly shown in Figure 9.4. Tg values of both PLLA and PDLLA show similar trend with molar mass, with a glass transition increase up to about 20 kDa molar mass and FIGURE 9.4 Melting temperature of PLLA (!) and glass transition temperature of PLLA ( ) and PDLLA (*) polymer as a function of molecular weight (adapted from Ref. 31 and authors’ data). . an almost threshold value thereinafter. At each Mn, PLLA displays higher Tg, this being attributable to the hindering effect exerted on the amorphous chain mobility by the crystalline regions. Worth noting is the similar trend for melting, with Tm ranging from about 130 to about 180 C as a function of Mn. The effect of the L- or D-content on the Tg of PLA has been discussed by Lim et al. [12] and Dorgan et al. [49]. The Tm of PLA is also a function of its optical purity. While the calculated theoretical Tm values for pure PLLA are reported to be 215, 207, and 205 C by different authors, the maximum practical obtainable Tm for pure PLLA is around 190 C [23, 35, 50]. The Tg is better evidenced in dynamical mechanical thermal analysis (DMTA) as shown in Figures 9.5 and 9.6, where the storage modulus E0 , loss modulus E00 , and the damping or loss factor tan d ¼ E00 /E0 are presented [51, 52]. In TABLE 9.2 Effect of Cooling Rate C on the Residual Molar Mass of Different Viscometric Molecular Weight PLLAs, Previously Heated at 10 C/min to 220 C Initial Mv After Cooling After Cooling After Cooling (kDa) at C À100 C/min at C À20 C/min at C À5 C/min 18 31 156 425 a 15 26 113 364 (À17%) (À16%) (À28%) (À15%) 14 26 111 357 (À22%) (À16%) (À29%) (À16%) 11 25 75 225 (À39%) (À19%) (À52%) (À49%) Note: Reprinted from Ref. 27. Copyright 1991, with permission from Elsevier. FIGURE 9.5 Storage modulus E0 , loss modulus E00 , and damping tan d of PDLLA. Heating rate 2 C/min and frequency 1 Hz.
  • CRYSTALLINITY AND CRYSTALLIZATION FIGURE 9.6 Storage modulus E0 , loss modulus E00 , and damping tan d of PLLA with a crystallinity of 55%. Heating rate 2 C/min and frequency 1 Hz. these plots, the temperatures corresponding to the tan d or E00 peaks are taken as the Tg values. Both PDLLA (Figure 9.5) and PLLA (Figure 9.6) polymers exhibit a storage modulus in the glassy region of about 2 GPa with a slight difference due to the crystalline content of PLLA (X ¼ 55%). Crystallinity is, however, responsible for the much lower decrease of E0 and E00 of PLLA after Tg and for the progressive increase of damping of PDLLA in the temperature range 100–120 C. Again, crystallinity accounts for the higher temperatures and width of the E00 peak of PLLA, both of them being due to the constraints exerted by the PLLA crystalline domains on the amorphous region mobility. Crystallinity, however, can have an important effect on PLLA mechanical properties and can be affected by the molecular weight. As shown by Perego et al. [39], crystallinity of PLLA of different viscometric molecular weights and annealed at 105 C for 90 min increased from 3–13% to 42–65%, depending on the molecular weight. While for low crystalline polymers, properties such as tensile strength or impact resistance almost did not depend on Mv, crystallization induced significant differences, improving these properties markedly for higher Mv polymers. Following the approach of Diez-Gutierrez et al. [53] that studied the effect of filler on the mechanical properties of talc/polypropylene composites, Fambri et al. [54] analyzed the effect of crystallinity on the storage modulus E0 of PLLA through a glass transition intensity factor defined as S ¼ ðE0 g ÀE0 r Þ=E0 r 117 FIGURE 9.7 Effect of crystallinity on the intensity of the transition S of compression molded samples (data from Ref. 32). phase. With increasing crystallinity, the intensity factor S decreases, with an almost proportional dependence on crystallinity (see Figure 9.7) for compression molded materials (data from Ref. 32). Since the glass transition is a kinetic transition, the measured value depends on both heating rate and frequency. This is clearly evidenced by comparing DMTA curves of Figure 9.6 (at 2 C/min and 1 Hz) and the multifrequency DMTA analysis of Figure 9.8 (at 0.5 C/min and 3, 10, and 30 Hz). At increasing deformation frequencies, the storage moduli of the glassy and rubbery plateaus increase and the apparent Tg is shifted to higher temperatures. The apparent activation energy for the relaxation process can be calculated from the following equation: ln f ¼ ln f0 ÀðEa =RTÞ ð9:5Þ ð9:4Þ In Equation 9.4, E0 g and E0 r are the E0 values in the glassy and rubbery states taken at about 30 C below and above the Tg, that is, 30 and 90 C, respectively. Values of SPDLLA ¼ 1173 and SPLLA ¼ 7.5 were calculated from data of Figures 9.5 and 9.6. A higher S value indicates a superior mobility and/or a higher content of the amorphous FIGURE 9.8 Storage modulus E0 and damping factor tan d of PLLA samples at various frequencies: 3 Hz (continuous line), 10 Hz (dashed line), and 30 Hz (dash-dotted line). Heating rate 0.5 C/min (from authors’ data).
  • 118 CRYSTALLIZATION AND THERMAL PROPERTIES for PLA are poly(ethylene glycol), triacetine, and citrate, laurate, and sebacate esters [26, 62–68]. The primary effect is the improvement of the polymer ductility and drawability, but also an increase of crystallization rate, as shown by Kulinski and Piorkowska [66] for PEG added PLLA. 9.3 FIGURE 9.9 Height (~) and width (!) of damping peak of PLLA as a function of crystallinity (data from Ref. 32). CRYSTALLIZATION REGIME One of the first papers discussing the morphology and crystal growth of PLLA as a function of thermal treatment and molecular weight was published by Vasanthakumari and Pennings in 1983 [30]. Following Hoffman and coworkers [69, 70], the growth rate (G) of a linear polymer crystal with chain folding can be expressed by  where f is the frequency in Hz, T is the temperature of the maximum of the loss factor, R is the ideal gas constant, and Ea is the activation energy for the relaxation process. According to Equation 9.5, apparent Ea of 553 and 620 kJ/mol were calculated for low and high molecular weight PLLA, respectively [53]. In dynamic mechanical tests, the reduced chain mobility due to the crystalline domains is detected by the shift of the damping peak to highertemperatures,that is, the increase of Tg, as already stated for DSC analyses. In addition, as shown in Figure 9.9, the DMTA evidences an increase of the elastic component of the viscoelastic behavior of the polymer as a significant reductionof the damping peak (datafrom Ref.32). At the same time, the width of the tan d peak increases as an effect of the larger distribution of amorphous chain mobility induced by the formation of a less mobile distributed interphase [55, 56]. In practical injection molding conditions, there is little crystallization due to the high cooling rate. For this reason, methods to induce crystallization have been studied by using nucleating agents or plasticizers during polymer processing. By using 1% talc, the crystallization half time can be reduced to less than 1 min [57]. Other nucleating agents such as montmorillonite are less effective than talc [58]. Some organic compounds have also been studied, such as calcium lactate, N,N-ethylenebis(12-hydroxystearamide), benzoylhydrazide compounds, or sodium stearate [59–62]. Addition of 1% calcium lactate induced crystallization of a 90:10 L/D, L-PLA copolymer during the injection molding cycle [59]. Kawamoto et al. [61] found that in the presence of a series of hydrazide compounds having a variety of methylene chain numbers, the Tc, and DHc, of PLLA during cooling increased with the methylene chain number. In this study, decamethylene dicarboxylic dibenzoylhydrazide was the most effective in promoting crystal nucleation by increasing the PLA Tc and DHc to 131 C and 46 J/g, respectively. The addition of a plasticizer decreases the Tg. Common plasticizers used G ¼ G0 exp    ÀKg Àu* exp RðTc ÀT¥ Þ Tc DTf ð9:6Þ where Kg is the nucleation constant, uà is the activation energy of chain mobility in the crystallization site, DT ¼ 0 (Tm À Tc) is the difference between the equilibrium melting temperature and the crystallization temperature, f is a factor equal to 2Tc/(Tm þ Tc), R is the gas constant, T¥ is the temperature at which all the motion associated with viscous flow ceases, and G0 is the front factor. Equation 9.6 can be used to predict the crystallization rate of a number of polymers over a wide range of temperature. As reported in Figure 9.10, the spherulite radius growth rate as a function of Tc reached a maximum for all investigated molecular weights at temperatures ranging around 130 C and depending on the specific polymer molecular weight [30]. FIGURE 9.10 Radius growth rate G as function of crystallization temperature Tc for different viscometric molecular weight samples of PLLA, that is, Mv ¼ 150 kDa (~), 260 kDa (&), 350 kDa ( ), and 690 kDa (*). Reprinted from Ref. 30. Copyright 1983, with permission from Elsevier. .
  • FIBERS In particular, the higher the PLLA molecular weight, the lower the G, as generally observed for many polymers. Different cooling rates also induced variation in the crystal morphology, with the formation of regular geometry and defined spherulites at high undercooling and spherulites with irregular shape and a coarse-grained structure at lower DT. According to the Hoffman theory, polymers such as polyethylene present two different crystallization regimes characterized by the ratio of the corresponding nucleation constants KgI/KgII equal to 2 [69, 70]. The presence of two crystallization regimes was confirmed by Vasanthakumari and Pennings [30] at 163 C for the highest of the PLLA molecular weights investigated, that is, for a polymer with viscometric molecular weight Mv ¼ 150 kDa, while lower molecular weight polymers presented only a regime II crystallization type. Tsuji and Ikada [35] studied the effect of different thermal treatments and annealing histories on solution cast PLLA films, which resulted in materials with different morphologies and physical properties. In particular, annealing was performed on films as (A) cast; (B) cast and molten; and (C) cast, molten, and quenched. Differences in crystallization were mainly attributed to the presence of preexisting spherulite nuclei before annealing, with the B film having higher developed total crystallinity content at any annealing temperature, resulting in higher mechanical strength. An increase in crystallinity was observed in the annealing temperature range 100–160 C. In a later work, Mazzullo et al. [34] have also found a regime III governed crystallization starting at 140 C, with KgIII/KgII equal to 3.13. Several authors have then confirmed the presence of three distinct crystallization regimes, starting at different temperatures and with different constants [36, 41–43]. The crystallization mode of PLLA has also been analyzed by using the Avrami equation [71, 72] Xc ¼ 1ÀexpðÀktn Þ 119 fibers can be produced by melt and solution spinning [24, 73– 75]. The two methods generate fibers with different mechanical properties and degradation rate, which can be explained in terms of the different developed crystal morphologies [47, 73, 76]. In both cases, the high shear rate induced orientation and chain alignment that in turn induced crystallization. Wide-angle X-ray scattering revealed the presence of two distinct crystal modifications in solution-spun fibers and in particular the origin of a so-called b-structure in hot drawn fibers. The b-structure seems to correspond to fibrillar morphology and bears most of the load applied to the fibers. Recently, it has been found that the redrawing of highly oriented PLLA film induces the transformation of the a-form crystallites into the b-form [77]. PLLA solution-spun fibers have better mechanical properties and also higher hydrolytic degradation stability [75]. Crystallinity developed in melt- and solution-spun fibers is reported in Figure 9.11 as a function of the applied draw ratio. In both cases, for draw ratios higher than 8–10, a maximum value of crystallinity equal to about 65% and 44% was reached for melt- and solution-spun fibers, respectively [76, 78]. Similar results were also found in the case of films where, after drawing, crystallinity significantly increased up to about 55% [79]. Melt spinning imparts intense thermal and mechanical stress on PLLA, resulting in a large reduction in molar mass that can reach 30–70% [24, 76]. Hence, when considering the crystallizability of polymer fibers, the effects of molecular weight and the thermal and mechanical drawing should be considered. The relatively low molecular weight of melt-spun fiber, between 30 and 150 kDa, results in high crystallizability especially during drawing. In a double-step process of fiber spinning followed by fiber drawing, crystallization from ð9:7Þ By measuring the fraction of crystallized material, Xc, with time t, a value of the constant n equal to 4 was evaluated by Miyata and Masuko [37] who observed, in nonisothermal crystallization experiments from the melt, a remarkable increase in the polymer crystallization at decreasing cooling rates and a maximum value of the isothermal crystallization rate at 105 C. A similar trend was observed by Iannace and Nicolais [36]. However, these results could be affected by the polymer thermal degradation induced at a low cooling rates, as previously described by Migliaresi et al. [27]. 9.4 FIBERS First studies about the fabrication of PLA fibers were made by Kulkarni et al. [13, 14] and Kalb and Pennings [23]. PLA FIGURE 9.11 Crystallinity of melt-spun PLLA fiber drawn at 160 C (*) and crystallinity of solution-spun PLLA fiber drawn at 190 C (!) as a function of draw ratio.
  • 120 CRYSTALLIZATION AND THERMAL PROPERTIES FIGURE 9.12 DSC curves of melt-spun PLLA fibers collected at different rates, as indicated. Reprinted from Ref. 76. Copyright 1997, with permission from Elsevier. FIGURE 9.13 First and second DSC scans of as-spun PLLA fibers produced by solution spinning (5% chloroform solution). the melt should be properly controlled in order to preserve the ability for further orientation and crystallization. It was reported that depending on the fiber collecting rate, the crystallinity of as-spun fiber initially increased up to about 35%, and then decreased as the fiber collecting rate increased, due to the faster cooling rate of the higher surface area of the lower diameter fibers. Figure 9.12 shows the thermograms of PLLA melt-spun fibers collected at 1.8, 5, 10, and 20 m/min. The different amorphous phase content is initially evidenced by the various DCp at the Tg, 0.36, 0.19, 0.14, and 0.30 J/(g K). Moreover, it should be noted that the crystallization peak of these melt-spun fibers occurred during heating in the range of about 80–120 C, significantly lower than that of completely amorphous PLLA reported in Figure 9.1, with direct dependence on the preexistent nucleating crystalline domains and the oriented polymer structure. At higher temperatures, a small exothermal peak of secondary crystallization at about 160 C precedes the melting. As shown in Figure 9.12, curves revealed two melting peaks, the higher temperature one just indicated by a shoulder in the highest rate collected fibers. The melting at about 188 C was due to crystallinity developed during the melt spinning process, whereas the peak at 180 C corresponded to the melting of crystals generated during the DSC analysis. The percentage crystallinity of the fiber can be evaluated as 5%, 30%, 37%, and 10% for spinning process of 1.8, 5, 10, and 20 m/min, respectively. The drawing temperature is selected above the Tg in order to modify the amorphous phase in the rubbery state; the initial crystallinity, the nucleation and growth rate according to the crystallization regimes I–III, the polymer chain mobility, and the polymer molecular weight are important factors for optimization of fiber orientation and crystallization. Typically, PLLA melt-spun fibers after drawing at 160 C reach a crystallinity of about 50–70% [24, 80]. In the case of solution spinning, the crystallinity of undrawn as-spun fiber was typically 30–40%, higher than that of melt-spun fibers (compare data of Figures 9.12 and 9.13), due to the lower spinning and cooling rate, and in general as a consequence of the more favorable conditions of solvent crystallization. Solvent, in fact, acts as a plasticizer, resulting in a higher chain rearrangement capability. At the same time, the relatively high molecular weight (between 300 and 900 kDa) inhibits the development of high crystallinity due to the lower polymer chain mobility [78,81, 82]. Figure 9.14 shows the first and second DSC scans of asspun PLLA fibers produced from dry spinning. Low Tg due to the residual solvent that acts as a plasticizer, and low Tc and Tm were observed in the first scan. In the second DSC scan after a mass loss of about 15% attributable to solvent remotion, a Tg of 65 C and crystallization and melting peaks at 140 and 180 C, respectively, with a crystallinity of about 25%, were observed. These solution as-spun fibers can be drawn at a higher drawing temperature, above the expected Tm (this is possible due to the high polymer molar mass and initial crystallinity). The results of DSC analysis revealed, after drawing at temperatures between 160 and 210 C, the formation of a double crystalline structure, with two melting peaks at about 160–170 and 190–200 C, as summarized in Table 9.3 and shown in selected thermograms of Figure 9.14. The higher the drawing temperature, the higher the total crystallinity and the enthalpy of the first melting peak at about 160–170 C, thus suggesting the formation of the b-phase, as previously indicated by Pennings and coworkers, in coexistence with the a-phase [47, 73, 74, 82]. At the same time, the Tg was found to increase with the drawing temperature, reaching about 100 C, as a consequence of the high orientation of the crystallized fibers. Moreover, the progressive and remarkable reduction of DCp from 0.1 to about 0.05 J/(g K) indicates both high packing
  • HYDROLYTIC DEGRADATION 121 polymer. Reed and Gilding reported in 1979 a sharp reduction of melting point to 120 C with the addition of 20% glycolide, with no crystallization occurring if glycolide content reached 25% [22]. Figure 9.15a and b summarizes some of their results. Analogously, the L/DL-lactide copolymers present an almost linear Tm depression, from 178 C for pure PLLA to 115 C for the 70/30 copolymer, and a Tg decrease from 63 to about 55 C, as shown in Figure 9.16. 9.5 FIGURE 9.14 DSC thermograms of solution-spun fibers after drawing at various temperatures (glass transition temperature and various melting peaks are evident) [78]. and low mobility of the amorphous phase above Tg, once again as a consequence of the high orientation of the high molecular weight drawn PLLA fibers. Copolymerization of L-lactide with other analogous cyclic lactones, such as DL-lactide, D-lactide, glycolide, or e-caprolactone, produces polymers with relatively random distribution of comonomers. The Tg of PLA copolymers decreases proportionally to the content of the glycolide or e-caprolactone comonomer to some extent. Moreover, the presence of stereochemical defects in PLLA reduces Tm, rate of crystallization, and extent of crystallization of the resulting HYDROLYTIC DEGRADATION The degradation kinetics of PLLA is largely affected by its crystallinity. As is known, degradation of PLA proceeds via hydrolysis, which is in turn controlled by the water diffusion in the free volume amorphous phase. In addition to crystallinity, other factors such as molecular weight, surface/volume ratio, purity, and chain orientation can greatly affect degradation kinetics [7]. Migliaresi et al. [83] compared the hydrolytic degradation of different molecular weight and crystallinity PLLAs, obtained by quenching or annealing treatments. As reported in Figure 9.17 for two different molecular weight PLLA polymers, as degradation proceeds, crystallinity increases reaching an almost equal value for the two polymers despite their largely different initial crystallinities, that is, X ¼ 46% for PLLA (153 kDa) and X ¼ 6% for PLLA (103 kDa). This behavior was not attributable to the loss of amorphous degraded materials, because no appreciable mass loss was detected during the experiment. A concomitant phenomenon associated with the molar mass reduction was the plasticizing effect of the water molecules diffusing into the amorphous regions, which provoked an increase of the chain segmental motion. Ultimately, the crystallinity increased when chain length became short enough to enable spatial rearrangement of the original amorphous regions in crystalline domains. This degradation-induced crystallization is a very important drawback for the biomedical use of PLLA as an implant TABLE 9.3 Glass Transition Temperature (Tg) and Change of Specific Heat at Tg (DCp), Melting Peak Temperature (Tm) and Enthalpy (DHm), and Crystallinity of Solution-Spun Fibers After Drawing at Various Temperatures Drawing Temperature ( C) 160 170 180 190 200 210 First Melting Peak Tg ( C) 83 81 87 98 92 97 DCp (J/(g K)) 0.13 0.12 0.11 0.09 0.09 0.04 Second Melting Peak Tm ( C) DHm (J/g) Tm ( C) DHm (J/g) Crystallinity (%) – – 155 169 167 172 – – 5 16 17 26 190 189 199 203 201 195 29 29 27 17 21 18 31 32 34 35 41 47
  • 122 CRYSTALLIZATION AND THERMAL PROPERTIES FIGURE 9.16 Tg and Tm of poly-L-D, L-lactide as a function of D, L-lactide content (adapted from Ref. 31 and authors data). material, which can last in the implant site for a very long time [84, 85]. Pistner et al. [84] observed in vivo an increase of crystallinity that reached 96% after 52 weeks of implantation in the dorsal muscle of rats. Even though they attributed such increase to the preferential degradation and cleavage of the amorphous regions, it is plausible that also in this case degradation induced a rearrangement of part of the amorphous regions into a crystalline structure, as FIGURE 9.15 Effect of composition on Tg, melting temperature (a) , and crystallinity (b) of L-lactide-co-glycolide copolymers as a function of glycolide content (adapted from literature). Reprinted from Ref. 22. Copyright 1979, with permission from Elsevier. FIGURE 9.17 Crystallinity percentage (~, ~) and molecular weight (*, ) as a function of degradation time in Ringer solution at 37 C for PLLA samples of initial low (open symbols) and high (filled symbols) crystallinity, respectively (adapted from Ref. 83). .
  • REFERENCES previously described by Migliaresi et al. [27]. The excessive longevity of PLLA in vivo, which is considered the major drawback of the polymer for biomedical applications, can be in part related to the observed in vivo degradation-induced crystallization that slows down the hydrolytic attack on the compact stable crystalline regions. Also, residues of PLLA in vivo have been observed after 7 years of implantation. Also worth noting for biomedical uses of PLLA is the increase in crystallinity due to gamma ray irradiation with the formation of new thin crystal lamellae [86]. REFERENCES 1. T. H. Barrows, Clin. Mater. 1986, 1, 233–257. 2. I. Engelberg, J. Kohn, Biomaterials 1991, 12, 292–304. 3. M. Vert, G. Schwarch, J. Coudane, J. Macromol. Sci. Pure Appl. Chem. 1995, A32, 787–792. 4. W. Amass, A. Amass, B. Tighe, Polym. Int. 1998, 47, 89–144. 5. Y. Ikada, H. Tsuji, Macromol. Rapid Commun. 2000, 21, 117–132. 6. D. A. Garlotta, J. Polym. Environ. 2001, 9, 63–84. 7. L. Fambri, C. Migliaresi., K. Kesenci, E. Piskin, Biodegradable polymers, in: R. Barbucci (Ed.), Integrated Biomaterials Science, Kluwer Plenum Publications, New York, 2002, Chapter 4, pp. 119–187. 8. H. Tsusj, Polylactide, in: Y. Doi, A. Steinbuchel (Eds.), Biopolymers. Polyesters III. Applications and Commercial Products, Wiley-VCH Verlag GmbH, Weinheim, 2002, Chapter 5, pp. 129–177. 9. N. Kumar, A. Ezra, T. Ehrenfroind, M. Y. Krasko, A. J. Domb, Biodegradable polymers, medical applications, in: H. F. Mark (Ed.), Encyclopedia of Polymer Science & Technology, 3rd edition, Wiley, Hoboken, NJ, 2003, Chapter 5, pp. 263–285. 10. R. von Oepen, W. Michaeli, Clin. Mater. 1992, 10, 21–28. 11. B. Gupta, N. Revagade, J. Hilborn, Prog. Polym. Sci. 2007, 32, 455–482. 12. L. T. Lim, R. Auras, M. Rubino, Prog. Polym. Sci. 2008, 33, 820–852. 13. R. K. Kulkarni, K. C. Pani, C. Neuman, F. Leonard, Arch. Surg. 1966, 93, 839–843. 14. R. K. Kulkarni, E. G. Moore, A. F. Hegyeli, F. Leonard, J. Biomed. Mater. Res. 1971, 5, 169–171. 15. A. P. Gupta, V. Kumar, Eur. Polym. J. 2007, 43, 4053–4074. 16. A. L. Lehninger, Biochemistry, 3rd edition, Worth Publishers Inc., New York, 1977. 17. J. Lunt, Polym. Degrad. Stabil. 1998, 59, 145–152. 18. T. Maharana, B. Mohanty, Y. S. Negi, Prog. Polym. Sci. 2009, 34, 99–124. 19. C. W. Lee, Fibers Polym. 2007, 8, 571–578. 20. T. M. Ovitt, G. W. Coates, J. Am. Chem. Soc. 1999, 121, 4072–4073. 21. E. W. Fisher, H. J. Sterzel, G. Wegner, Kolloid Z. Z. Polym. 1973, 251, 980–990. 123 22. D. K. Gilding, A. M. Reed, Polymer 1979, 20, 1459–1464. 23. B. Kalb, A. J. Pennings, Polymer 1980, 21, 607–612. 24. S. H. Hyon, K. Jamshidi, Y. Ikada, Melt spinning of poly-Llactide and hydrolysis of the fibre in vitro, in: S. W. Shalaby, A. S. Hoffmann, B. D. Ratner, T. A. Horbett (Eds.), Polymers as Biomaterials, Plenum Press, New York, 1984, pp. 51–65. 25. Y. Ikada, K. Jamshidi, H. Tsuji, S. H. Hyon, Macromolecules 1987, 20, 904–906. 26. H. Younes, D. Cohn, Eur. Polym. J. 1988, 24, 765–773. 27. C. Migliaresi, A. De Lollis, L. Fambri, D. Cohn, Clin. Mater. 1991, 8, 111–118. 28. A. Celli, M. Scandola, Polymer 1992, 33, 2699–2703. 29. G. L. Loomis, J. R. Murdoch, K. H. Gardner, Polym. Prepr. 1990, 31(2), 55. 30. R. Vasanthakumari, A. J. Pennings, Polymer 1983, 24, 175–178. 31. K. Jamshidi, S. H. Hyon, Y. Ikada, Polymer 1988, 29, 2229–2234. 32. C. Migliaresi, D. Cohn, A. De Lollis, L. Fambri, J. Appl. Polym. Sci. 1991, 43, 83–95. 33. C. Marega, A. Marigo, V. Di Noto, R. Zannetti, A. Martorana, G. Paganetto, Makromol. Chem. 1992, 193, 1599–1606. 34. S. Mazzullo, G. Paganetto, A. Celli, Prog. Colloid Polym. Sci. 1992, 87, 32–34. 35. H. Tsuji, Y. Ikada, Polymer 1995, 36, 2709–2716. 36. S. Iannace, L. Nicolais, J. Appl. Polym. Sci. 1997, 64, 911–919. 37. T. Miyata, T. Masuko, Polymer 1998, 39, 5515–5521. 38. J. J. Kolstad, J. Appl. Polym. Sci. 1996, 62, 1079–1091. 39. G. Perego, G. D. Cella, C. Bastioli, J. Appl. Polym. Sci. 1996, 59, 37–43. 40. J. R. Sarasua, R. E. Prud’homme, M. Wisniewski, A. Le Borgne, N. Spassky, Macromolecules 1998, 31, 3895–3905. 41. M. L. Di Lorenzo, Polymer 2001, 42, 9441–9446. 42. M. L. Di Lorenzo, Eur. Polym. J. 2005, 41, 569–575. 43. M. L. Di Lorenzo, J. Appl. Polym. Sci. 2006, 100, 3145–3151. 44. Y. Wang, J. F. Mano, Eur. Polym. J. 2005, 41, 2335–2342. 45. Y. Yuryev, P. Wood-Adams, M. C. Heuzey, C. Dubois, J. Brisson, Polymer 2008, 49, 2306–2320. 46. P. De Sanctis, A. Kovacs, Biopolymers 1968, 6, 299–306. 47. W. Hoogsteen, A. R. Postema, A. J. Pennings, G. ten Brinke, P. Zugenmaier, Macromolecules 1990, 23, 634–642. 48. L. Cartier, T. Okihara, Y. Ikada, H. Tsuji, J. Puiggali, B. Lotz, Polymer 2000, 41, 8909–8919. 49. J. R. Dorgan, J. Jansen, M. P. Clayton, J. Rheol. 2005, 49, 607–619. 50. A. J. Ninjenhuis, D. W. Grijpma, A. J. Pennings, Polym. Bull. 1991, 26, 71–77. 51. L. E. Nielsen, R. F. Landel, Mechanical Properties of Polymers and Composites, 2nd edition, Marcel Dekker, New York, 1994. 52. M. P. Seppe, Rapra Rev. Rep. 1997, 8, 22. 53. S. Diez-Gutierrez, M. A. Rodriguez-Perez, J. A. Saja, J. I. Velasco, Polymer 1999, 40, 5345–5353.
  • 124 CRYSTALLIZATION AND THERMAL PROPERTIES 54. L. Fambri, K. Kesenci, C. Migliaresi, Polym. Compos. 2003, 24 (1), 100–108. 55. D. W. van Krevelen, Properties of Polymers, 3rd edition, Elsevier, Amsterdam, 1998, p. 402. 56. L. C. E. Struik, Physical Aging of Amorphous Polymers and Other Materials, Elsevier, Amsterdam, 1978. 57. T. Ke, X. Sun, J. Appl. Polym. Sci. 2003, 89, 1203–1210. 58. J. Y. Nam, S. S. Ray, M. Okamoto, Macromolecules 2003, 36, 7126–7131. 59. D. M. Bigg, Conference Proceedings ANTEC: Annual Technical Conference, Society of Plastics Engineers, Nashville, TN, 2003, pp. 2816–2822. 60. J. Y. Nam, M. Okamoto, H. Okamoto, M. Nakano, A. Usuki, M. Matsuda, Polymer 2006, 47, 1340–1347. 61. N. Kawamoto, A. Sakai, T. Horikoshi, T. Urushihara, E. Tobita, J. Appl. Polym. Sci. 2007, 10, 198–203. 62. H. Li, M. A. Huneault, Polymer 2007, 48, 6855–6866. 63. S. Jacobsen, H. G. Fritz, Polym. Eng. Sci. 1999, 39, 1303–1310. 64. O. Martin, L. Averous, Polymer 2001, 42, 6209–6219. 65. N. Ljungberg, B. Wesselen, J. Appl. Polym. Sci. 2002, 86, 1227–1234. 66. Z. Kulinski, E. Piorkowska, Polymer 2005, 46, 10290–10300. 67. Z. Ren,L. Dong,Y. Yang, J. Appl.Polym.Sci.2006,101, 1583–1590. 68. I. Pillin, N. Montrelay, Y. Grohens, Polymer 2006, 47, 4676–4682. 69. J. I. Lauritzen, J. D. Hoffman, J. Appl. Phys. 1973, 44, 4340–4352. 70. J. D. Hoffman, G. T. Davis, J. I. Lauritzen, in: N. B. Hannay (Ed.), Treatise on Solid State Chemistry, Vol. 3, Plenum Press, New York, 1976, Chapter 7. 71. M. J. Avrami, J. Chem. Phys. 1940, 8, 212–224. 72. M. J. Avrami, J. Chem. Phys. 1941, 9, 177–184. 73. B. Eling, S. Gogolewski, A. J. Pennings, Polymer 1982, 23, 1587–1593. 74. S. Gogolewski, A. J. Pennings, J. Appl. Polym. Sci. 1983, 28, 1045–1061. 75. C. Migliaresi, L. Fambri, Macromol. Symp. 1997, 123, 155–161. 76. L. Fambri, A. Pegoretti, R. Fenner, S. D. Incardona, C. Migliaresi, Polymer 1997, 38, 79–85. 77. K. Takahashi, D. Sawai, T. Yokoyama, T. Kanamoto, S.-H. Hyon, Polymer 2004, 45, 4969–4976. 78. L. Fambri, A. Pegoretti, M. Mazzurana, C. Migliaresi, J. Mater. Sci. Mater. Med. 1995, 5, 679–684. 79. A. Fujimori, N. Ninomiya, T. Masuko, Polym. Adv. Technol. 2008, 19, 1735–1744. 80. A. Pegoretti, L. Fambri, C. Migliaresi, J. Appl. Polym. Sci. 1997, 64, 213–223. 81. I. Horacek, V. Kalisek, J. Appl. Polym. Sci. 1994, 54, 1751–1757. 82. J. W. Leenslag, A. J. Pennings, Polymer 1987, 28, 1695–1702. 83. C. Migliaresi, D. Cohn, L. Fambri, J. Biomater. Sci. Polym. Ed. 1994, 5(6), 591–606. 84. H. Pistner, D. R. Bendix, J. Muhling, J. F. Reuther, Biomaterials 1993, 14, 291–298. 85. J. E. Bergsma, W. C. de Bruijn, F. R. Rozema, R. R. M. Bos, G. Boering, Biomaterials 1995, 16, 25–31. 86. D. Mililevic, S. Trifunovic, S. Galovic, E. Suljiovrujic, Radiat. Phys. Chem. 2007, 76, 1376–1380.
  • 10 RHEOLOGY OF POLY(LACTIC ACID) JOHN R. DORGAN 10.1 INTRODUCTION The readers of this chapter are expected to come from a wide variety of backgrounds with varying degrees of expertise in polymer materials science. Accordingly, a detailed understanding of the rheological properties of polymers is, for many readers, an unknown subject. With these factors in mind, the first topic to be addressed is, What is rheology and why is it important? Simply stated, rheology is the study of the flow and deformation of matter. Rheology is a subfield of continuum mechanics that combines elements of both solid and fluid mechanics. Polymer rheology is a highly advanced science that reveals very important features of polymeric materials like PLA. In particular, the viscometry of dilute polymer solutions is an effective means of determining fundamental properties of polymer molecules; these include the molecular weight and the characteristic ratio (subjects to be discussed in detail in this chapter). More important, the melt rheology of polymeric materials reflects the relationship between molecular structure and dynamic properties. Of these properties, the viscosity and elasticity of the molten polymer are of primary importance in processing of plastics into useful articles. For example, both the shear thinning viscosity and the strain hardening extensional behavior of PLA are of great importance in plastic processing operations such as extrusion and fiber spinning. With this very cursory discussion of the nature and importance of rheology in mind, the attentive reader will ask, Is there anything particularly unique about the rheology of PLA? Indeed, the melt rheology of PLA is distinct from other commodity plastics in at least one important respect. Namely, the chemical stability of PLA can be adversely affected rather easily compared to most other polymeric molecules. At processing temperatures typically needed for semicrystalline PLA, namely, those above the melting temperature, both hydrolysis and pyrolysis can become important. Fortunately, commercially available PLAs are typically stabilized against thermal degradation. However, it is impossible to overstate the importance of drying PLA sufficiently prior to processing (or measuring its rheological properties!). Rheological measurements can provide direct evidence for the loss in molecular weight of improperly stabilized PLA in the molten state. Accordingly, in examining rheological data on PLAs, it is of particular importance to understand what, if any, chemical changes have taken place during the preparation and testing of the samples. In this regard, there is a clear link to dilute solution viscometry that can provide information about molecular weight in a simple test requiring only elementary equipment. As a final remark on the unique nature of PLA rheology, we can state that rheological measurements are particularly useful for assessing the effectiveness of both chemical stabilization against thermal degradation and of drying procedures meant to preclude hydrolytic degradation. This chapter is organized to provide a short yet detailed understanding of the fundamental rheological properties of PLAs; extensive use of references is employed where appropriate. However, the chapter is meant to be self-contained, and so appropriate introductory material is also provided. Properties of the single chain are discussed and related to measurable quantities available from dilute solution viscometry. This is followed by a discussion of the single chain properties of PLA. Next, a brief discussion of the rheology of polymeric materials is provided. In this context, mathematical models are briefly discussed, including the Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 125
  • 126 RHEOLOGY OF POLY(LACTIC ACID) empirical but useful Havriliak–Negami (HN) model that serves as the basis for the software provided with the text. Appropriate conclusions are drawn regarding the influence of PLA rheology on polymer processing operations. Finally, in an appendix to the chapter, a description of software enabling the prediction of melt rheological properties as a function of molecular weight, temperature, and shear rate is given. 10.2 FUNDAMENTAL CHAIN PROPERTIES FROM DILUTE SOLUTION VISCOMETRY Conformational analysis is the study of how molecules can alter their overall geometry (their arrangement in space) through bond rotations. Such analysis is useful for understanding polymer properties, particularly solution properties. Molecular forces in polymeric materials dictate the conformation adopted and can be classified as either bonded or nonbonded. Bonded forces are strictly intramolecular and typically consist of strong covalent bonds. Nonbonded forces are weaker interactions and may be either intramolecular or intermolecular. Examples of nonbonded forces include hydrogen bonding, dipole and multipole interactions, and van der Waals forces (London dispersion forces). Flexibility of polymer chains should not be confused with individual bond flexibility. Polymers, which are considered flexible, are capable of assuming many different geometric conformations in space. Rigid polymers are those that are capable of assuming relatively few conformations due to the restricted bond rotations. 10.2.1 Unperturbed Chain Dimensions The freely jointed chain is a model that treats the conformation of a polymer molecule as a mathematical random walk. This is a simple model for chain conformation in which 1. There are n segments 2. Each skeletal bond is represented as a vector li 3. No restrictions are placed on the allowable bond angle between two adjoining segments (it is freely jointed) 4. All bond lengths are taken as the same and equal to |l | The freely jointed chain can step back on itself; that is, it does not obey the physical reality of excluded volume. Excluded volume effects are caused by the fact that no two polymer segments may occupy the same position in space. The volume that one chain segment occupies is excluded from occupation by another segment. This effect serves to push segments away from one another and thus increase the overall dimensions of the chain. The solvent in which a polymer is dissolved strongly affects its conformation. Solvent quality is a general termi- nology to describe how well a particular solvent can dissolve a particular polymer. In good solvents, a polymer chain adopts an open conformation and is expanded. In poor solvents, a polymer chain adopts a closed conformation and is collapsed. To be quantitative, a measure of the size of a polymer coil of a given molecular structure, molecular weight, and in a particular solvent environment is needed. The end-to-end vector, r, connects the first and last segments of a polymer chain. For the freely jointed chain, its mean squared value may be calculated within the assumptions of the model from a consideration of a random walk of n vectors, each of length l.  2 r o ¼ nl 2 ð10:1Þ Here, the brackets denote an ensemble average (the average over many chains) and the subscripted ‘‘o’’ denotes conditions referred to as unperturbed. A related measure of polymer size is the radius of gyration, s, which is defined as the average of the distance from a segment to the chain’s center of mass; that is, s2 ¼ n 1 X ðri Àrc:o:m: Þ2 n þ 1 i¼1 ð10:2Þ The radius of gyration can be considered as an average measure of the radius of a polymer coil. Theta conditions are the actual conditions under which a real polymer chain assumes its unperturbed dimensions. The solvent is then called a theta solvent and the temperature at which the unperturbed dimensions are produced is called the theta temperature. The unperturbed dimensions are realized because of the balance of two competing processes, the first being excluded volume effects that favor open conformations and the second being unfavorable energetic contacts between the polymer and the solvent. Theta conditions mean that only short-range interactions dictate the polymer conformation because the inter- and intramolecular interactions are identical in enthalpic nature (i.e., the heat of mixing is zero). 10.2.2 Real Chains The characteristic ratio, C1, is a quantity that can be experimentally determined or calculated using advanced molecular simulation techniques. It is defined by C1  2 r o ¼ lim n ! 1 nl 2 ð10:3Þ Equation 10.3 says that the characteristic ratio is a measure of how large the polymer coil is under unperturbed conditions relative to the freely jointed chain. That is, it is a measure of chain stiffness or how the chain conformation is extended
  • FUNDAMENTAL CHAIN PROPERTIES FROM DILUTE SOLUTION VISCOMETRY compared to the freely jointed chain. Consider the compactness of the freely jointed chain—it can step back on and through itself; real chains all have characteristic ratios greater than unity (in fact, all real chains must be greater than two). For example, the characteristic ratio of poly(ethylene oxide) is 4.0, that of polyethylene is 6.7, and that of polystyrene is 10.0. Note that molecular structure influences the quantity as one would expect; the oxygen in poly(ethylene oxide) has no pendant groups and provides low energy bond rotations, whereas the bulky phenyl groups of polystyrene increase the value of C1. The expansion factor, a, is used to characterize the size of polymer coils when they are not under unperturbed conditions—that is, to measure the effects of the polymer’s environment on its overall dimensions. To do so, a is defined as the ratio of the root mean squared end-to-end vector relative to its value in the unperturbed state according to the following equation: a¼  2 1=2 r hr2 i1=2 o ¼  2 1=2 r ðC1 nl 2 Þ1=2 ð10:4Þ The expansion factor is a function of the particular polymer, solvent, and conditions (primarily temperature) being considered. Fundamental polymer chain characteristics such as the characteristic ratio and the molecular weight can be quantitatively related to the melt rheological response. Because of this fact, it is important to be able to characterize these quantities. Several techniques are available, with multiple angle laser light scattering (MALLS) playing an important role. However, viscometry is the simplest technique for characterizing polymer chains of known molecular architecture. 10.2.3 Solution Viscometry The viscosity of a low molecular weight solvent can be dramatically increased by the addition of even a very small amount of high molecular weight polymer. The reason for this effect is that the large polymer coil disturbs the local flow, causing gradients in fluid velocity that dissipate energy, requiring additional work to cause flow, and therefore increasing viscosity. Because relative increases in viscosity can be measured using simple and inexpensive techniques, such measurements of dilute solution viscosities of polymer solutions have been widely practiced for a long time. A critical feature of such measurements is that they provide information on the fundamental chain properties discussed above; this information is available because the viscometry is a measure of the hydrodynamic radius, Rh, of the polymer coil in the solvent being used. 127 The hydrodynamic radius is defined based on a result from fluid mechanics developed by Stokes long ago, the so-called Stokes flow around a sphere. In Stokes flow, the proportionality constant between the force applied by the flowing fluid on the moving sphere and the velocity of the sphere is called the friction coefficient z. For rigid bodies such as colloidal particles, the friction coefficient is given by ð10:5Þ z ¼ 6pgs R where gs is the viscosity of the dispersing solvent and R is the radius of the sphere. This result is generalized to provide a definition for the hydrodynamic radius of polymer coils in solution as Rh ¼ z 6pgs ð10:6Þ That is, according to Equation 10.6, Rh is the radius of a rigid spherical particle that has the same frictional coefficient as the dissolved polymer coil. In dilute solution viscometry, the concentrations employed (measured in volume fraction, j) are of the order j % 0.001. That is, the concentration is below cà , the socalled overlap concentration, at which polymer coils touch and interact. In dilute solution viscometry, the polymer coils are isolated in solution. In this dilute regime, the polymer molecules can be treated as spheres with diameters roughly equal to the root mean squared end-to-end distance, D ¼ hr2i1/2. If the molecules are considered to behave like hard spheres, which implies that the entrained solvent molecules move with the polymer coil, then the Einstein–Batchelor equation may be applied, which is as follows:  à g ¼ gs 1 þ 2:5j þ 7:6j2 ð10:7Þ where g is the viscosity of the solution (or suspension of spheres), gs is the viscosity of the pure solvent, and j is the volume fraction of spheres. The volume fraction of spheres (polymer coils containing entrained solvent) can be found from  ws ¼ Volume of spheres ¼ c2 Volume of solution NA pD3 6 M2  ð10:8Þ where c2 is polymer concentration expressed in g/cm3, NA is Avagadro’s number, M2 is the molecular weight of the polymer, and D is the sphere diameter. The nomenclature associated with viscosity measurements has changed and both new and old terminologies are encountered. Table 10.1 summarizes the nomenclature.
  • 128 RHEOLOGY OF POLY(LACTIC ACID) TABLE 10.1 Nomenclature of Dilute Solution Viscometry Symbol New Name New Units g gs gr ¼ g/gs gsp ¼ gr À 1 gsp/c2 (ln gr)/c2 [g] ¼ lim(gsp/c2) c2 ! 0 Solution viscosity Solvent viscosity Viscosity ratio – Viscosity number Logarithmic viscosity number Limiting viscosity number Pa s Pa s – – mL/g mL/g mL/g Old Name Solution viscosity Solvent viscosity Relative viscosity Specific viscosity Reduced viscosity Inherent viscosity Intrinsic viscosity Old Units Poise Poise – – dL/g dL/g dL/g For monodisperse polymer samples, a log–log plot of the limiting viscosity number against the polymer molecular weight (M2) gives a straight line. This finding is the basis of the Mark–Houwink (MH) equation,  are known. This average molecular weight, M v , is known as the viscosity average molecular weight. On the basis of the expected range for the Mark–Houwink a parameter, it can be seen that a ½gŠ ¼ KH M2    Mn < Mv < Mw ð10:9Þ where a and KH are known as the Mark–Houwink parameters. These parameters depend on the specific polymer, solvent, and temperature under consideration. Values for these parameters are extremely important and so are tabulated in places like the Polymer Handbook. Substitution of Equation 10.9 for the volume fraction of spheres into the Einstein–Batchelor equation along with the use of the root mean squared end-to-end distance for the sphere diameter gives rise to Equation 10.10. ð10:12Þ The viscosity average molecular weight always lies between the number-average and weight-average molecular weights; in good solvents, its value is close to the value of the weightaverage molecular weight. 10.2.4 Viscometry of PLA ð10:11Þ The molecular characterization of PLA depends on the accurate knowledge of its fundamental properties. These include such parameters as Mark–Houwink constants that relate intrinsic viscosity to molecular weights, theta-condition front factors (KH) used to calculate single chain properties, and characteristic ratios (C1) that give an indication of the bonding structure of polymers. Unfortunately, literature reports on some fundamental properties of PLAs seriously disagree. For example, Table XVIII of a review on PLA [1] lists 15 different sets of MH constants [2], and agreement among the various sets is very poor. MH values tabulated in the above review can be evaluated and used to reconstruct data for a Stockmayer–Fixman plot. This analysis plots the intrinsic viscosity divided by the square root of the viscosity averaged molecular weight against the square root of the viscosity averaged molecular weight; it is used to determine the KH for the MH equation. Figure 10.1 shows the KH values so determined (vertical axis intercepts) to vary over a wide range, from lowest to highest, by a factor of 25! From this analysis [3], it was concluded that the state of the literature regarding Mark–Houwink parameters for PLA was inadequate. The literature inconsistencies can be attributed to the following reasons: or may be used to find the molecular weight if the grouping in the parenthesis of Equation 10.10 and the expansion factor 1. Different definitions of the mean squared ‘‘monomer’’ step length  2 3=2 hr io 1=2 ½gŠ ¼ W a3 M 2 M2 ð10:10Þ where W represents the universal viscometric constant (Flory constant) having a numerical value of 2.1  1023 (cm3/g)(g/(mol cm3)) and a is the expansion factor discussed above. The grouping in the parenthesis of Equation 10.10 can be related to the characteristic ratio and is nearly independent of the polymer molecular weight; the dependence of intrinsic viscosity on solvent quality is therefore proportional to the product aM2. In theta solvents, a is unity (the intrinsic 0:5 viscosity scales with M2 ) and in good solvents a is pro0:1 0:8 portional to M2 (the intrinsic viscosity scales with M2 ). Comparison with Equation 10.1 suggests that the Mark–Houwink parameter should lie in the range 0.5 a 0.8. Equation 10.10 may be used to find the expansion factor if theta conditions for the polymer solution are known,  a¼ ½gŠ ½gŠy 1=3
  • 129 FUNDAMENTAL CHAIN PROPERTIES FROM DILUTE SOLUTION VISCOMETRY room temperature for PLA is 1.258 g/cm3. Refractive indexes were determined over the wavelength range 300–1300 nm for solid PLAs at room temperature using ellipsometry. The refractive indexes for all compositions can be represented to within experimental error by a two-term Cauchy expression: nðlÞ ¼ ð1:445 Æ 0:00075Þ þ ð4892 Æ 143Þnm2 =l2 for 300 nm < l < 1300 nm FIGURE 10.1 Stockmayer–Fixman plots, reconstructed from published Mark–Houwink parameters, illustrating inconsistencies in the PLA literature. Rather than the wide range of values exhibited, the y-axis intercepts should be a common value equal to KH. 2. Different expansion factor models for correcting data obtained in good solvents to Q conditions, or neglecting of such correction altogether [4] 3. Different assumed values for the viscosity function F (or Flory constant) used to relate intrinsic viscosities to molecular dimensions in solution 4. Different methods of measuring molecular weights and subsequently relating them to dilute solution viscosities [4–10]. Because of the confused state of the literature, a careful study [3] was performed to determine appropriate physical constants for PLA under dilute solution conditions. The results of this important study are summarized here with attention focused on the most important results useful for characterizing PLA. PLA homopolymers and optical copolymers spanning wide ranges of molecular weight and stereoisomer proportion were prepared by ring-opening polymerizations of L- and D-lactides using tin octanoate as the catalyst. The weight-average molecular weights produced spanned the range from below 104 to over 106 (g/mol), and the range of enantiomer proportions was from 100:0 to 50:50. As part of the careful characterization, both density and indices of refraction were determined and used to calculate the refractive index increment. The refractive index increment is an important physical property when light scattering measurements are conducted. The density of amorphous PLA was found using the data of Witzke [11] on densities above Tg, namely, rðTÞ ¼ rðT ¼ 0ÞexpðaTÞ ¼ 1:2836 expðÀ7:7  10À4 TÞ ð10:13Þ where density, r, is in g/cm3 and temperature, T, is in  C. It is instructive to note that Equation 10.13 implies the density at ð10:14Þ At the wavelength of the sodium D line, to which most tabulated solvent refractive indexes refer, evaluation of Equation 10.14 gives nPLA ð589:2Þ ¼ 1:459 ð10:15Þ Densities (r) and refractive indexes (n) for various solvents were taken from standard tables [12–14], and the refractive indices of solvent mixtures were calculated according to classical [15] volume additivity of polarizabilities, namely, n2 À1 X n2 À1 ¼ wi 2i n2 þ 2 ni þ 2 i ð10:16Þ where the wi’s are volume fractions. The adoption of Equations 10.15 and 10.16 enables the calculation of refractive index increments (dn/dc) that are within experimental error of measured values. It is important to note that a molecular weight determined with an inaccurate dn/dc value will be in error by a factor that is the square of the ratio of accurate and inaccurate dn/dc values. Samples were characterized by means of dilute solution viscometry in three different solvents: size exclusion chromatography (SEC), static multiangle light scattering, variable-angle spectroscopic ellipsometry, and melt rheology. More important, this study [3] also conclusively demonstrated consistency between solvents in determining the KH used to calculate single chain properties; Figure 10.2 shows the resulting Stockmayer–Fixman plot that should be compared with Figure 10.1. The resulting KH data provided by these experiments produce values of characteristic ratios (C1) in the range 6.1 Æ 1.3. These experimental values are in excellent agreement with molecular simulation [16]. As a result of the careful investigation by Dorgan and coworkers, the Schulz–Blaschke and Mark–Houwink constants for dilute PLA solutions in chloroform and in THF have been determined. For chloroform at 30 C, the correct values are kSB ¼ 0.302, K ¼ 0.0131 (mL/g), and a ¼ 0.759, while for THF at 30 C the correct values are kSB ¼ 0.289, K ¼ 0.0174 (mL/g), and a ¼ 0.736. The importance of having reliable and accurate values for the Mark–Houwink
  • 130 RHEOLOGY OF POLY(LACTIC ACID) FIGURE 10.2 Stockmayer–Fixman plots for three separate solvents from the study of Dorgan et al. showing consistency of the determined KH values (contrast with Figure 10.1). FIGURE 10.3 parameters is that they are needed in relating relative molecular weights determined by SEC to absolute molecular weights. Accordingly, without such information, it is very difficult to evaluate absolute molecular weight distributions for PLA, yet such information is critical in understanding a wide variety of phenomena. In addition, there is a great utility of having reliable values for both Mark–Houwink and Schultz–Blaschke parameters in that the combination enables the measurement of an average molecular weight in a very simple viscometry experiment. Equation 10.17 gives an explicit formula for the intrinsic viscosity based only on the relative flow times for a single concentration experiment. ½ gŠ ¼ g=g0 À1 tN = t0 N À1 ¼ c½1 þ kSB ðg=g0 À1ފ c½1 þ kSB ðtN = t0 N À1ފ ð10:17Þ where c is the concentration in g/ml and overtildes denote corrected flow times for the pure solvent and solution, t0 and t, respectively. Accordingly, Equation 10.17 provides a direct method to obtain the intrinsic viscosity using a simple flow viscometer and a single concentration solution. With the value of intrinsic viscosity and the Mark–Houwink parameters, Equation 10.9 can be employed to find an absolute value for theviscometry averaged molecular weight. The application of Equation 10.17 is employed in the accompanying software program that enables the prediction of linear melt rheology based on the determination of molecular weight. 10.3 PROCESSING OF PLA: GENERAL CONSIDERATIONS The bulk properties of PLA are greatly affected by the molecular weight of the polymer, the chain architecture Stereochemistry of L, D, and D,L-PLA backbones. (branched versus linear), and the degree of crystallinity [17, 18]. The relative proportions of L- and D-lactide in the polymer backbone determine the amount of crystallinity of a PLA sample. A diagram of representative backbones is provided in Figure 10.3; the mirror image stereochemistry of poly(L-lactide) with respect to poly(D-lactide) reflects the same relationship as the respective monomers. The lactide dimer obtained from the condensation of one L- and one D-lactic acid is called LD-lactide (or meso-lactide)— a representative polymer structure obtained from this monomer is also shown in Figure 10.3. The stereochemistry illustrated in Figure 10.3 profoundly affects crystallinity in PLA. Samples containing 87.5% L-lactide are completely amorphous, while samples with 92% L-lactide possess some crystallinity [19]. Polymers from 100% L-lactide can be nearly half-crystalline [20]. The melting point (Tm) range of crystalline PLA is 145–186 C [17], although a blend of 100% D-lactide and 100% L-lactide polymers in a ‘‘stereocomplex’’ forms a closely packed crystalline structure that increases the Tm to 230 C [21]. The appearance of the PLA is also affected by the crystalline content. Amorphous PLA and low-crystalline PLA are clear materials with high gloss, while highly crystalline PLA is an opaque white material. The molecular weight, structure, and crystallinity of PLA play important roles in its mechanical properties as well, including tensile strength, tensile modulus, and percent elongation to break [17, 18, 22]. Mechanical properties of solution-spun [23–27] and melt-spun [28–33] PLA fibers have been thoroughly investigated. It has been found that these properties are roughly equivalent to other polyesters, meaning that PLA can replace textiles based on nonrenewable resources. In addition, scanning electron microscopy (SEM) [23, 26, 29, 33] and wide-angle X-ray scattering (WAXS) [28, 34] have been useful in examining surface structure with respect to roughness and fracture surfaces.
  • MELT RHEOLOGY: AN OVERVIEW Understanding surface properties is important for dying and other textile finishing operations. Cicero et al. have provided a complete characterization of the hierarchical fiber morphology from linear PLAs [35, 36], determining thermal, mechanical, and morphological properties of the fibers and showing that properties can be widely manipulated through a combination of processing temperature and draw ratio (the amount of stretching the fiber undergoes). Maximum tensile strength and modulus of 0.38 and 3.2 GPa, respectively, were obtainable. Again, such studies have been instrumental in establishing the suitability of PLA as a textile fiber. Atomic force microscopy showed that the fiber morphology was found to be highly fibrillar with microfibril diameters of $40 nm. The same researchers have studied the effects of branching on fiber properties and morphology [37], and investigated the improvement of fiber properties specifically when thermally stabilized PLA is used [38]. These research studies provide several routes for optimizing the performance of PLA when used as a textile fiber. 10.4 MELT RHEOLOGY: AN OVERVIEW Plastics are typically fabricated into useful articles in the molten state through melt flow, and therefore rheological properties of a polymer are considered to be of great importance. A polymer melt or solution can deform via shear, tensile (elongation), hydrostatic compression, or some combination of these. For introductory purposes, consider simple shear deformation as depicted in Figure 10.4. Here, a cubical volume element is undergoing a shear deformation due to an applied force, F, that acts parallel to one of the upper faces of the cube having a cross-sectional area of A. The shear stress, t, that results is defined according to Equation 10.18. t ¼ F=A ð10:18Þ 131 Provided the deformation is uniform, the shear strain, c, is given by Equation 10.19 c ¼ X=Y ð10:19Þ _ The velocity gradient, or shear rate, c, is also very important in describing the response of a polymer to deformation, _ c¼   d d X 1 dX V dV ðcÞ ¼ ¼ ¼ ¼ dt dt Y Y dt Y dY ð10:20Þ where V denotes velocity of the fluid. The rate of deformation is characterized by the shear rate and rate dependence of properties is the hallmark of viscoelasticity that so dominates the behavior of polymer melts. Specifically, what is important is how fast is the rate of deformation relative to the rate at which molecular motions can occur. The relationship between these relative rates depends, of course, on the conditions of the flow, with temperature having a dominant effect. The viscosity, g, is often described as the resistance of a material to flow; the greater the viscosity, the greater the force needed to deform the fluid. That is, viscosity relates the measurable shear stress to the imposed shear rate (or visa versa). For Newtonian fluids, the viscosity is independent of shear rate and the stress may be written as follows: _ t ¼ gc ð10:21Þ From Equation 10.21, the dimensions of viscosity are stress multiplied by time, and in the SI system viscosity is measured in units of pascal-seconds (Pa s). For polymer melts and solutions, the fluid behavior is non-Newtonian and Equation 10.21 must be modified to allow the viscosity to become a material function of the shear rate. Similarly, material elements may be deformed by pulling on opposite sides of the cube with an equal force; this constitutes an extensional deformation that may be characterized by an extensional viscosity. For solid materials, the relationship between stress, t, and strain, c, is governed by the shear modulus, G, t ¼ Gc ð10:22Þ For such elastic materials, there is no rate dependence. An elastic material subjected to an extensional deformation is characterized by a Young’s modulus, E. Consider the case of a time-varying shear deformation, specifically the case in which the strain is given by a sine wave of frequency v and amplitude A(v) according to Equation 10.23 FIGURE 10.4 applied force. Element undergoing shear deformation due to an cðvÞ ¼ AðvÞsinðvtÞ ð10:23Þ
  • 132 RHEOLOGY OF POLY(LACTIC ACID) The corresponding shear rate would be given by _ cðvÞ ¼ vAðvÞcosðvtÞ ð10:24Þ For a viscous material, as defined by Equation 10.21, the stress would be _ tðvÞ ¼ gcðvÞ ¼ gvAðvÞcosðvtÞ ð10:25Þ whereas for an elastic material, as defined by Equation 10.22, the stress would be tðvÞ ¼ GcðvÞ ¼ GAðvÞsinðvtÞ ð10:26Þ That is, for an elastic material, the stress would be in phase with the imposed strain; in contrast, Equation 10.25 shows that the stress would be 90 out of phase with the strain for a viscous material. For viscoelastic materials such as polymer melts and solutions, the stress contains both in-phase and out-of-phase components. In the small deformation limit where the stress remains linear with respect to the amplitude of the applied strain, the response may be written as h i 00 tðvÞ ¼ AðvÞ G0 ðvÞsinðvtÞ þ G ðvÞcosðvtÞ ð10:27Þ where the frequency (i.e., rate of deformation)-dependent moduli G0 (v) and G00 (v) are known as the storage and loss moduli, respectively. The storage modulus represents the elastic or in-phase response of the material and the loss modulus reflects the viscous or out-of-phase response. Due to their rate dependence, these are known as the dynamic moduli. From the dynamic moduli, it is possible to construct a complex viscosity, the magnitude of which is given by Equation 10.28. sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  0   00  G ðvÞ 2 G ðvÞ 2 þ jg*ðvÞj ¼ v v ð10:28Þ The empirical Cox–Merz rule simply states that the magnitude of the dynamic viscosity as a function of the frequency is equal to the steady viscosity as a function of shear rate. 10.5 PROCESSING OF PLA: RHEOLOGICAL PROPERTIES As discussed above, PLAs have physical properties useful in fibers, packaging, and other applications traditionally dominated by petroleum-based resins. Although the general literature on polylactides is extensive, only a few articles [39–43] have considered rheological properties. Measurements of dynamic, steady, and transient shear viscosities have been presented and the extensional data on PLA showed a strong strain hardening behavior [43]. The rheology of blends of linear and branched PLA architectures has also been comprehensively investigated [42, 44]. For linear architectures, the Cox–Merz rule relating complex viscosity to shear viscosity is valid for a large range of shear rates and frequencies. The branched architecture deviates from the Cox–Merz equality and blends show intermediate behavior. Both the zero shear viscosity and the elasticity (as measured by the recoverable shear compliance) increase with increasing branched content. For the linear chain, the compliance is independent of temperature, but this behavior is apparently lost for the branched and blended materials. These authors use the Carreau–Yasuda model, Equation 10.29, to describe the viscosity shear rate dependence of both linear and branched PLAs and their blends: h iððC4 À1Þ=C3 Þ _ g ¼ C1 1 þ ðC2 cÞC3 ð10:29Þ _ where g is the viscosity, c is the shear rate, and C1–C4 are material-dependent parameters. C1 determines g0 that decreases with increasing linear content. C2 is the relaxation time corresponding approximately to the reciprocal of frequency for the onset of shear thinning. C3 determines the shear thinning that increases with increasing linear content; that is, branched PLA shear thins more strongly than the linear material. The increase of both g0 and shear thinning with the addition of branching is also reported by other studies on PLA polymers with star polymer chain architectures. Tensile and thermal properties of the linear–branched blends were also measured and found to be equal within the statistical error of the experiments. The results suggest that excellent control over rheological behavior of PLA is possible through blending chain architectures without compromising mechanical properties. The above studies do not, however, capture a systematic description of PLA rheology across a broad range of stereooptical compositions, as the materials studied possessed high (>90%) L-stereochemical center content. However, one study [3] does provide a comprehensive evaluation of the linear viscoelastic properties of PLA across a wide range of molecular weights and stereochemical compositions. The major points of this study are summarized here. This study also serves as the basis for the software included with this text for predicting viscoelastic properties of PLAs. An important consideration in studying the melt rheology of PLA is melt stability; at high enough temperatures, PLA will degrade in molecular weight as time progresses. This makes melt stabilization crucial to good measurement
  • 133 PROCESSING OF PLA: RHEOLOGICAL PROPERTIES science. The protocol put forth by Dorgan and coworkers consists of adding tris(nonylphenyl)phosphite (TNPP) as a stabilizer so that the measurement range could be extended to 200 C with negligible degradation. The TNPP concentration used is 0.35 wt%, and this was found to be suitable for the vicinity of 200 C by Lehermeier and Dorgan [42]. However, direct addition of neat additive to melts in an internal mixer [43] may be inadequate. Instead, weighed quantities of the polymers must be dissolved in and then recovered from fixed volumes of dichloromethane (CH2Cl2) solutions containing an appropriate concentration of TNPP by evaporation of the solvent in a vacuum oven at 40 C. This leaves appreciable CH2Cl2 dissolved in the solid polymer that must be removed by preheating the polymer to above the melting point (180 C) in a vacuum oven for 5 min. Subsequently, samples can be compression molded into disks suitable for rheological testing. To facilitate data reduction, a model of the rheological response is desirable. The Havriliak and Negami (HN) model for complex viscosity is written as g*ðvÞ ¼ g1 þ ðg0 Àg1 Þ½1 þ ðit0 vÞa ŠðnÀ1Þ=a ð10:30Þ This has five easily interpretable parameters, each corresponding to a readily discernible feature in a plot of log |gà | versus log v. The parameter g1 is the often negligible or experimentally inaccessible high-frequency limiting value of |gà |. In case g1 is negligible, Equation 10.30 reduces to g*ðvÞ ¼ g0 ½1 þ ðit0 vÞa ŠðnÀ1Þ=a ð10:31Þ Taking the scalar magnitude of both sides of Equation 10.31 leads to jg*ðvÞj ¼ h g0 1 þ 2ðvt0 Þcosðap=2Þ þ ðvt0 Þ2a ið1ÀnÞ=2a ð10:32Þ The effect of changing the model parameters of Equation 10.32 is presented in Figures 10.5–10.8. Figure 10.5 shows the effect of changing h0, the zero shear viscosity, while keeping all other parameters fixed. This figure demonstrates that g0 is simply a vertical scaling factor that sets the low-frequency limiting value of |gà |. Figure 10.6 shows the effect of changing t0, a characteristic viscous relaxation time. As discussed above, when the molecular motion is fast relative to the timescale for deformation, even a polymer melt will flow like a Newtonian liquid—the viscosity will be independent of the shear rate. However, when the deformation rate (the timescale of which is dictated by the inverse shear rate) exceeds the molecular relaxation rate, the polymers are deformed from their equilibrium conformations and shear thinning is observed. The value of is t0 in the model dictates when this transition takes place on the horizontal scale (frequency) as shown in Figure 10.6. The parameter a is an inverse measure of the breadth of the relaxation spectrum, or a direct measure of the sharpness of the ‘‘knee’’ over which the transition from Newtonian flow to shear thinning takes place. The effect of changing this parameter is illustrated in Figure 10.7. This graph shows that FIGURE 10.5 Effect of changing the zero shear viscosity parameter in the HN model.
  • 134 RHEOLOGY OF POLY(LACTIC ACID) FIGURE 10.6 Effect of changing the characteristic time parameter in the HN model. a useful feature of Equation 10.32 is that the sharpness of the knee in the viscosity data may be easily adjusted independently of the high-frequency asymptotic slope through the parameter a. However, values of a other than 1 cause an incorrect low-frequency scaling (G0 and G00 do not have appropriate slopes of 2 and 1, respectively, on a logarithmic scale), and accordingly the HN model should be considered only as a very useful empirical interpolating function. Finally, as shown in Figure 10.8, n À 1 is the high-frequency limiting slope d[log |gà (v)|]/d(log v) in the shear FIGURE 10.7 thinning regime. Examination of viscosities measured for model linear flexible polymers of narrowly distributed high molecular weights and for the Curtis–Bird theoretical model [45] indicates a value of n ¼ 0.03 Æ 0.01; therefore, n is set equal to 0.03 in the model developed for linear PLAs. Master curves based on the principle of time–temperature superposition can be constructed from measured values of |gà (v)| shifted vertically and horizontally as specified by Ferry [46]. Thus, horizontal shifting to a reference temperature T0 and fitting provides the constants c0 and T1 1 Effect of changing a, the breadth parameter, in the HN model.
  • PROCESSING OF PLA: RHEOLOGICAL PROPERTIES FIGURE 10.8 Effect of changing the high-frequency n parameter in the HN model. (Vogel temperature) for the Williams–Landel–Ferry (WLF) shift factor aT, where ÀlogðaT Þ ¼ 135 C1 ðTÀT0 Þ c0 ðTÀT0 Þ ¼ 1 C2 þ ðTÀT0 Þ ðTÀT1 Þ ð10:33Þ The average WLF parameters for PLA are C1 ¼ 3.24 KÀ1 and C2 ¼ 164.9 K; the later correspond to a Vogel temperature of 288.25 K. Vertical shift factors given by r0T0/rT are first applied to the measured values of G0 and G00 using the known temperature dependence of density (Equation 10.13) before determining the shift factors [11]. This WLF temperature dependence is built into the accompanying software. Figure 10.9 shows master curves (G0 , G00 , and |gà |) for an amorphous PLA of high molecular weight. The symbols represent measurements shifted as described above. The continuous curves represent the best-fit two-mode HN model with frequency shifting factors for all temperatures calculated according to Equation 10.33 with a single pair of WLF constants. It is to be emphasized that the data is fit only to |gà | and the resulting values of G0 and G00 are calculated. Also, the infinite shear rate viscosity (g1) parameter of the terminal mode serves as the zero shear viscosity (g0) for the glassy mode. The adopted phenomenological model is capable of describing the observed viscoelastic behavior over a large FIGURE 10.9 Master curves for an amorphous PLA sample with a weight-average molecular weight Mw ¼ 570k (g/mol) over nearly 14 decades in frequency obtained via time–temperature superposition.
  • 136 RHEOLOGY OF POLY(LACTIC ACID) FIGURE 10.10 Scaling of the zero shear viscosity against molecular weight for PLA samples of variable optical composition. (nearly 14 decades) range of frequencies with a set of seven parameters. Figure 10.10 presents measure of zero shear viscosities for a large collection of PLA samples. For the weight-average molecular weight range of 105–106 (g/mol) and a reference temperature of 180 C, the zero shear viscosity is described well by the relationship logðg0 Þ ¼ À14:26 þ 3:4 logðMw Þ regardless of the stereochemical composition. Likewise, average WLF parameters are c1 ¼ 3.24 KÀ1 and c2 ¼ 164.9 K across stereochemical composition. It is now clear that earlier reports [40, 41] of anomalously high scaling of the zero shear viscosity with molecular weight were artifacts of small data sets and thermal degradation. The present findings combined with an earlier study [43] in which samples have been stabilized against thermal degradation clearly demonstrate that the viscosity of PLA scales with molecular weight within the range is expected for linear flexible polymer chains. These findings [47] are built into the accompanying rheology software application. Consistent values for plateau moduli, G0 , are obtained by N integrations of either the relaxation spectra or the loss modulus [48] as described by Ferry [46]: G0 N ð ln l 2 ¼ HðtÞdðln tÞ ¼ p ln tmin ð Àln tmin Àln l mined using Equation 10.34 for the data of Figure 10.9 is 1.0 MPa; analysis across a wide set of PLA samples provides values for the plateau modulus of 1.0 Æ 0.2 MPa. Plateau modulus results can be used to infer characteristic ratios according to a chain packing model. A packing model for linear chains [49] gives the entanglement molecular weight (or molecular weight between entanglements) as Me ¼ rRT=G0 N ð10:35Þ and the same reference then gives the characteristic ratio in terms of Me: 2 C1 ¼ 10rÀ2=3 ðMe ÞÀ1=3 M1 =L ð10:36Þ Additional characteristics of chains in the melt that are available from the packing scheme are the packing length [50], p¼ M hr2 i0 rNA ð10:37Þ the tube diameter, 00 G ðvÞdðln vÞ ð10:34Þ where, in the limits of integration, ln tmin is the abscissa where the relaxation spectrum function H has its minimum value in the transition zone and l is a longest relaxation time, beyond which H(t) ¼ 0. The plateau modulus value deter- dt ¼   !1=2 M e r2 0 M ð10:38Þ and the critical entanglement molecular weight [49] Mc ¼ 4:24pÀ0:65 Me ð10:39Þ
  • CONCLUSIONS TABLE 10.2 Flexible Linear Polymer Melt Chain Comparisons 10.6 Value Polymer Characteristic M1 (relative molecular mass per backbone bond)  L 2 (A2) r (g/cm3) G0 (MPa) N Me (g/mol) Me/M1 (backbone bonds) Mc (g/mol) M   c/M1 (backbone bonds)  r2 o =M (A2) C1  p (packing length) (A)  dt (tube diameter) (A) dt/p (Ronca–Lin ratio) PLA (140 C) PEO (140 C) i-PP (190 C) 24.02 14.68 21.04 2.05 1.152 1.0 3,959 165 9,211 383 0.574 6.7 2.51 47.7 19.0 2.11 1.034 1.8 1,973 134 5,330 363 0.805 5.6 1.99 39.9 20.0 2.37 0.759 0.43 6,797 323 13,635 648 0.694 6.2 3.15 68.7 21.8 Mc represents the molecular weight that must be reached in order to have the zero shear viscosity follow the scaling of Figure 10.10. Rheological measurements combined with the packing model allow a comparison of various molecular parameters for PLA with other known polymer structures. Table 10.2 collects various PLA properties obtained using Equations 10.35, 10.36. For comparison, values for two other linear polymers are also tabulated: poly(ethylene oxide) (PEO) and isotactic polypropylene (i-PP). These polymers are selected because they share some common molecular features with PLA. It should be noted that a distinction should be made regarding the nature of the oxygen group in the PLA backbone as it is part of a rigid ester linkage. This distinction is evident in the data; the tube diameter and the number of backbone bonds between entanglements (Me/M1) in PLA is greater than in PEO as the latter molecule has greater rotational freedom about the oxygen linkage within its backbone. In this respect, the comparison with i-PP is suitable if the virtual bond model of Flory and Tonelli [5, 6] is adopted (i.e., a virtual bond between the two carbons is envisioned as a result of the rigidity of the ester group). An important result presented in Table 10.2 is the finding that the so-called Ronca–Lin ratio is approximately 20; if a higher value of C1 were accepted, this ratio would change and linear PLA would be distinct from the hundreds of linear polymer chains to which the packing model has been applied [51]. Two other noteworthy physical properties result from the examination of the data. These are the critical molecular weight for entanglement and the molecular weight between entanglements. Table 10.2 shows that these values are about 9000 and 4000 g/mol, respectively. For linear polymer melts, a factor of about 2 is a usual finding in polymer physics. 137 CONCLUSIONS Dilute solution viscometry and melt rheology are fundamental techniques for characterizing polymeric materials. Given PLA’s growing importance, it is important to understand the fundamental chain properties that are reflected in the data obtained using these experimental methods. When performed correctly on high-quality data, Stockmayer–Fixman fitting of intrinsic viscosity data for three different solvents yields a consistent Mark–Houwink scale factor for theta conditions of KH ¼ 0.107 Æ 0.022 mL/g. This corresponds to unperturbed chain dimensions described by a characteristic ratio C1 ¼ 6.5 Æ 0.9. Equations 10.17 and 10.9 can be used to determine calibrated absolute values for the viscometry averaged molecular weight of typical commercial PLAs. The melt rheological properties are consistent with the packing model propounded by Fetters and coworkers. Using the packing model and an experimentally measured plateau modulus G0 ¼ 1.0 MPa Æ 20% leads to an implied C1 in the N range 6.7 Æ 0.7. Simulation methods have also been used to estimate PLA characteristic ratios. The study of Blomqvist [16] may be summarized as indicating characteristic ratios in the range 6.1 Æ 1.3, in excellent agreement with the experimental results. While Blomqvist’s computational study does find a systematic trend with optical composition, the predicted magnitude of the effect is small and lies within the range of uncertainty of the experimental investigations. Appreciably, larger values of C1 or a strong dependence of this ratio on optical composition is not confirmed either experimentally or via molecular simulation. The careful experimental work is consistent not only with recent simulation work but also with the pioneering work of Tonelli and Flory [5, 6], who reported a characteristic ratio of about 4.5 (based on real chemical bonds as opposed to the value of 2.0 for the virtual bond model) and a temperature dependence that leads to a value of about 6 at 180 C. Accordingly, values of C1 determined by solution, melt rheology, and simulation now agree and imply that the polylactides are typical linear flexible polymers. Rheological and thermal measurements on a comprehensive and well-characterized set of homopolymers and copolymers spanning wide ranges of molecular mass and stereoisomer proportions (L-content) have been reported in the literature [47]. Within the weight-average molecular weight range of 105–106 (g/mol) and a reference temperature of 180 C, the zero shear viscosity is described well by logðg0 Þ ¼ À14:26 þ 3:4 logðMw Þ, the plateau modulus is 1.0 Æ 0.2 MPa, and average WLF parameters are c1 ¼ 3.24 KÀ1 and c2 ¼ 164.9 K; the later correspond to a Vogel temperature of 288.25 K. The values of the glass transition temperatures at infinite molecular weight for 100, 80, and 50% L-content are 60.2, 56.4, and 54.6 C, respectively. Based upon a chain packing model, molecular parameters  determined include a packing length of 2.51 A, a tube
  • 138 RHEOLOGY OF POLY(LACTIC ACID)  diameter of 47.7 A, and a characteristic ratio of 6.5 Æ 0.9, independent of stereoisomeric composition. The critical molecular weight for entanglement, Mc, is found to be near 9000 g/mol, while the molecular weight between entanglements, Me, is about 4000 g/mol. APPENDIX 10.A DESCRIPTION OF THE SOFTWARE The software consists of a Microsoft Excel workbook that can be downloaded from ftp://ftp.wiley.com/public/sci_ tech_med/poly(lactic_acid). There are two tabs in the workbook: one is labeled ‘‘plot’’ and the other ‘‘worksheet.’’ The plot tab presents a log–log plot of the magnitude of the dynamic viscosity and the moduli (G0 and G00 ) as a function of frequency (in rad/s). The user can change values for independent variables in the worksheet and see the resulting predictions updated in the plot tab. In addition, tabular data are updated in the worksheet. Only three independent variables may be changed, and these are the relative viscosity, the concentration, and the temperature. These variables appear at the top of the worksheet labeled as ‘‘user input’’ and are in red. The concentration corresponds to the concentration used in performing a 1-point experiment according to Equation 10.17 and is entered in units of g/dL. The relative viscosity is the ratio of corrected flow times for the solution of the entered concentration to that of the pure solvent. The temperature is entered in units of degree Celsius. The reference temperature is fixed at 180 C. The program uses the input value of concentration and relative flow times to calculate a molecular weight according to Equations 10.17 and 10.9. This value is used to calculate a zero shear viscosity, and average values for the other parameters of the HN model discussed in the text are adopted to calculate the frequency dependence (i.e., to calculate the viscoelastic moduli as a function of shear rate). All the values adopted in the model are disclosed on the worksheet page. REFERENCES 1. D. Garlotta, A literature review of poly(lactic acid). J. Polym. Environ. 2001, 9, 63–84. 2. P. Kratochvil, U. W. Suter, Definitions of terms relating to individual macromolecules, their assemblies, and dilute polymer solutions (1988). Pure Appl. Chem. 1989, 61, 211–41. 3. J. R. Dorgan, J. Janzen, S. B. Hait, D. K. Knauss, Fundamental solution and single chain properties of polylactides. J. Polym. Sci. Part B 2005, 43, 3100–3111. 4. J. Ren, O, Urakawa, K, Adachi, Dielectric study on dynamics and conformation of poly(D,L-lactic acid) in dilute and semidilute solutions. Polymer 2003, 44, 847–855. 5. A. E. Tonelli, P. J. Flory, The configuration statistics of random poly(lactic acid) chains. I. Experimental results. Macromolecules 1969, 2, 225–227. 6. A. E. Tonelli, P. J. Flory, D. A. Brant, The configurational statistics of random poly(lactic acid) chains. II. Theory. Macromolecules 1969, 2, 227. 7. A. Schindler, D. Harper, Polylactide. II. Viscosity–molecular weight relationships and unperturbed chain dimensions. J. Polym. Sci., Polym. Chem. Ed. 1979, 17, 2593–9. 8. S. H. Kim, Y. K. Han, Y. H. Kim, S. I. Hong, Multifunctional initiation of lactide polymerization by stannous octoate/pentaerythritol. Makromol. Chem. 1992, 193, 1623–1631. 9. C. A. P. Joziasse, H. Veenstra, D. W. Grijpma, A. J. Pennings, On the chain stiffness of poly(lactide)s. Macromol. Chem. Phys. 1996, 197, 2219–2229. 10. S. Kang, G. Zhang, K. Aou, et al. Raman and light scattering analysis of poly(lactic acid) flexibility. Polym. Prepr. 2002, 43, 896–897. 11. D. R. Witzke, Introduction to properties, engineering, and prospects of polylactide polymers. Ph.D. thesis. Michigan State University, East Lansing, MI, 1997, xxvi þ 389 pp. 12. H. G. Elias, Isorefractive and isopycnic solvent pairs, in: J. Brandrup, E. H. Immergut (Eds.), Polymer Handbook, Wiley, New York, 1989, III/13–23 pp. 13. H. G. Elias, Refractive indices of common solvents, in: J Brandrup, EH Immergut (Eds.), Polymer Handbook, Wiley, New York, 1989, III/25–28 pp. 14. D. Fleischer, Physical constants of the most common solvents, in: J. Brandrup, E. H. Immergut (Eds.), Polymer Handbook, Wiley, New York, 1989, p. III/29. 15. J. R. Partington, An Advanced Treatise on Physical Chemistry, Longmans/Green, London/New York, 1953. 16. J. Blomqvist, RIS Metropolis Monte Carlo studies of poly[Llactic), poly(L,D-lactic) and polyglycolic acids. Polymer 2001, 42, 3515–3521. 17. L. Lu, A. G. Mikos, Poly(lactic acid), in: J. E. Mark (Ed.), Polymer Data Handbook, Oxford University Press, New York, 1999, pp. 627–633. 18. I. Engelberg, J. Kohn, Physico-mechanical properties of degradable polymers used in medical applications: a comparative study. Biomaterials 1991, 12, 292–304. 19. M. Vert, G. Schwarch, J. Coudane, Present and Future of PLA Polymers. JMS. Pure Appl. Chem. 1995, A32, 787–796. 20. M. H. Hartman, High molecular weight polylactic acid polymers, in: D. H. Kaplan (Ed.), Biopolymers from Renewable Resources, Springer, Berlin, 1998, pp. 367–411. 21. Y. Ikada, K. Jamshidi, H. Tsuji, S. H. Hyon, Stereocomplex formation between enantiomeric poly(lactides). Macromolecules 1987, 20, 904–906. 22. M. H. Naitove, Push is on to commercialize biodegradable lactide polymers. Plast. Technol. 1995, March 15–17. 23. L. Fambri, A. Pegoretti, C. Migliaarsesi, Biodegradable fibres. Part I. Poly-L-lactic acid fibres produced by solution spinning. J. Mater. Sci. Mater. Med. 1994, 5, 679–683.
  • REFERENCES 24. S. Gogolewski, P. Mainil-Varlet, The effect of thermal treatment on sterility, molecular and mechanical properties of various polylactide. I. Poly(L-lactide). Biomaterials 1996, 17, 523–528. 25. I. Horcek, V. Kal a ısek, Polylactide. I. Continuous dry spinninghot drawing preparation of fibers. J. Appl. Polym. Sci. 1994, 54, 1751–1757. 26. J. W. Leenslag, S. Gogolewski, A. J. Pennings, Resorbable materials of poly(L-lactide). V. Influence of secondary structure on the mechanical properties and hydrolyzability of poly(Llactide) fibers produced by a dry-spinning method. J. Appl. Polym. Sci. 1984, 29, 2829–2842. 27. J. W. Leenslag, A. J. Pennings, High-strength poly(L-lactide) fibres by a dry-spinning/hot-drawing process. Polymer 1987, 28, 1695–1702. 28. B. Eling, S. Gogolewski, A. J. Pennings, Biodegradable materials of poly(L-lactic acid): 1. melt spun and solution-spun fibers. Polymer 1982, 23, 1587–1593. 29. L. Fambri, A. Pegoretti, R. Fenner, S. D. Incardona, C. Migliaresi, Biodegradable fibres of poly(L-lactic acid) produced by melt spinning. Polymer 1997, 38, 79–85. 30. S. D. Incardona, L. Fambri, C. Migliaresi, Poly-L-lactic acid braided fibresproduced by melt spinning: characterization and in vitro degradation. J. Mater. Sci. Mater. Med. 1996, 7, 387–391. 31. K. Mezghani, J. E. Spruiell, High speed melt spinning of poly(Llacticacid) filaments. J.Polym. Sci. Part. B 1998, 36, 1005–1012. 32. G. Schmack, B. T€ndler, R. Vogel, R. Beyreuther, S. Jacobsen, a H. G. Fritz, Biodegradable fibers of poly(L-lactide) produced by high-speed melt spinning and spin drawing. J. Appl. Polym. Sci. 1999, 73, 2785–2797. 33. X. Yuan, A. F. T. Mak, K. W. Kwok, et al. Characterization of poly(L-lactic acid) fibers produced by melt spinning. J. Appl. Polym. Sci. 2001, 81, 251–260. 34. W. Hoogsteen, A. R. Postema, A. J. Pennings, Crystal structure, conformation, and morphology of solution-spun poly(L-lactide) fibers. Macromolecules 1990, 23, 634–642. 35. J. A. Cicero, J. R. Dorgan, Physical properties and fiber morphology of poly(lactic acid) obtained from continuous two-step melt spinning. J. Polym. Environ. 2001, 1–15. 36. J. A. Cicero, J. R. Dorgan, J. Janzen, et al. Supramolecular morphology of two-step melt-spun poly(lactic acid) fibers. J. Appl. Polym. Sci. 2002, 86, 2828–2838. 37. J. A. Cicero, J. R. Dorgan, J. Garrett, et al. Effects of molecular architecture on two-step melt-spun poly(lactic acid) fibers. J. Appl. Polym. Sci. 2002, 86, 2839–2846. 139 38. J. A. Cicero, J. R. Dorgan, S. F. Dec, D. M. Knauss, Phosphite stabilization effects on two-step melt-spun fibers of polylactide. Polym. Degrad. Stab. 2002, 78, 95–105. 39. G. Biresaw, C. J. Carriere, Interfacial tension of poly(lactic acid)/polystyrene blends. J. Polym. Sci. Part B 2002, 40, 2248–2258. 40. J. J. Cooper-White, M. E. Mackay, Rheological properties of poly(lactides). Effect of molecular weight and temperature on the viscoelasticity of poly(l-lactic acid). J. Polym. Sci. Part B 1999, 37, 1803–1814. 41. J. R. Dorgan, J. S. Williams, Melt rheology of poly(lactic acid): entanglement and chain architecture effects. J. Rheology 1999, 43, 1141–1155. 42. H. J. Lehermeier, J. R. Dorgan, Melt rheology of poly(lactic acid): consequences of blending chain architectures. Polym. Eng. Sci. 2001, 41, 2172–2184. 43. L. I. Palade, H. J. Lehermeier, J. R. Dorgan, Melt rheology of high-L content poly(lactic acid). Macromolecules 2001, 34, 1384–1390. 44. H. J. Lehermeier, Poly(lactic acid): an environmentally benign polymer derived from corn: rheological and permeation properties. M.S. thesis, Colorado School of Mines, Golden, Co, 2000, xvii þ 122 pp. 45. R. B. Bird, R. C. Armstrong, O. E. Hassager, Dynamics of Polymeric Liquids, Vol. 2, 2nd edition, Wiley, New York, 1987. 46. J. D. Ferry, Viscoelastic Properties of Polymers, Wiley, New York, 1980. 47. J. R. Dorgan, J. Janzen, M. P. Clayton, et al. Melt rheology of variable L-content poly(lactic acid). J. Rheol. 2005, 49, 607–619. 48. J. M. Dealy, K. F. Wissbrun, Melt Rheology and Its Role in Plastics Processing, Van Nostrand Reinhold, New York, 1990. 49. L. J. Fetters, D. J. Lohse, W. W. Graessley, Chain dimensions and entanglement spacings in dense macromolecular systems. J. Polym. Sci. Part B 1999, 37, 1023–1033. 50. L. J. Fetters, D. J. Lohse, R. H. Colby, Chain dimensions and entanglement spacings, in: JE Mark (Ed.), Physical Properties of Polymers Handbook, AIP Press, New York, 1996, pp. 335–340. 51. L. J. Fetters, D. J. Lohse, S. T. Milner, W. W. Graessley, Packing length influence in linear polymer melts on the entanglement, critical, and reptation molecular weights. Macromolecules 1999, 32, 6847–51.
  • 11 MECHANICAL PROPERTIES GABRIELE PEREGO AND GIAN DOMENICO CELLA 11.1 INTRODUCTION Poly(lactic acid) (PLA) is a glossy, high modulus thermoplastic polymer with properties comparable to polystyrene (PS). The polymers obtained from optically active monomers (L-lactide and D-lactide) are semicrystalline while the optically inactive monomers (racemic D,L-lactide and mesolactide) give amorphous materials [1]. Amorphous PDLLA is typically used for biomedical devices and in particular for slow drug release, while semicrystalline PLLA is selected for applications where higher mechanical and thermomechanical properties are required. PLA is in general characterized by good mechanical properties, with elastic modulus of 3000–4000 MPa and tensile strength of 50–70 MPa. Therefore, it can in principle substitute conventional polymers in many applications, such as packaging, extruded and thermoformed containers, and so on. However, the low elongation at break that is typical of this polymer limits some of its uses. PLA has a glass transition temperature of about 60–70 C, and it is brittle at room temperature, fracturing through a crazing mechanism. Efforts to improve the properties of PLA comprised copolymerization, blending with other biodegradable polymers, and plasticization. Rubber modification through copolymerization has proven to be very effective in increasing the toughness of PLA. It is worth mentioning that the mechanical properties of PLA are not significantly influenced by its synthesis methods [1, 2]. Mechanical orientation is another method for improving the mechanical properties of PLA due to its morphological change from a-crystallites into b-crystallites upon drawing. This process converts the otherwise fragile polymer into a versatile and useful material for practical tools, packaging, and biomedical applications. In this chapter, the mechanical properties of PLA are introduced as a general framework. Next, a more detailed look takes into consideration the relevant physicochemical effects and process parameters. 11.2 GENERAL MECHANICAL PROPERTIES AND MOLECULAR WEIGHT EFFECT 11.2.1 Tensile and Flexural Properties Semicrystalline PLA has a tensile strength of approximately 50–70 MPa, tensile modulus of 3000–4000 MPa, elongation at break of 2–10%, flexural strength of 100 MPa, and flexural modulus of 4000–5000 MPa [1–6]. PLA specimens obtained by a typical injection molding process are generally almost amorphous, because of the slow crystallization kinetics characterizing this material. In Tables 11.1 and 11.2, the results of physicochemical and mechanical characterization of, respectively, PLLA and PDLLA injection molded specimens with different molecular weights are presented [7]. From the analysis of the mechanical data reported in Tables 11.1 and 11.2, it is evident that in the selected range of molecular weights, the tensile and flexural properties of PDLLA and amorphous PLLA are quite different. Tensile strength for PLLA typically ranges from about 50 to 70 MPa while for PDLLA it ranges from 40 to 53 MPa. This different behavior is mainly related to the stereoregularity of the polymer chains, which are characterized, in the case of PLLA, by the presence of only S(À) chiral centers. It has been observed that after annealing at 105 C PLLA samples showed a crystallinity ranging from 45% to 70%, as revealed Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 141
  • 142 MECHANICAL PROPERTIES TABLE 11.1 Mechanical Properties of PLLA Specimens with Different Molecular Weights [7] Sample PLLA I Annealing at 105 C Molecular weight (Mv, Da) Tm ( C) Crystallinity (%) Tensile properties Yield strength (MPa) Tensile strength (MPa) Yield elongation (%) Elongation at break (%) Elastic modulus (MPa) Flexural properties Flexural strength (MPa) Maximum strain (%) Elastic modulus (MPa) Impact resistance Izod, notched (kJ/m2) Izod, unnotched (kJ/m2) Heat resistance HDT ( C) Vicat penetration ( C) Hardness Rockwell hardness (scale H) PLLA II No 23,000 178 9 Yes 20,000 178 70 No 58,000 179 9 Yes 47,000 180 52 No 67,000 181 3 Yes 71,000 178 45 – 59 – 1.5 3550 – 47 – 1.3 4100 68 58 2.3 5.0 3750 68 59 2.2 3.5 4050 70 59 2.2 7.0 3750 70 66 2.0 4.0 4150 64 2.0 3650 51 1.6 4200 100 4.1 3600 113 4.8 4150 106 4.7 3650 119 4.6 4150 1.9 13.5 3.2 18.0 2.5 18.5 7.0 34.0 2.6 19.5 6.6 35.0 57 60 66 157 – 59 – 163 55 59 61 165 85 84 83 84 88 88 by DSC analysis (100% crystallinity ¼ 93 J/g [1]). Tensile strength of these PLLA samples ranged from 47 to 70 MPa in the same range of molecular weights. The effect of molecular TABLE 11.2 Mechanical Properties of PDLLA Specimens with Different Molecular Weights [7] Sample Molecular weight (Mv, Da) Tensile properties Yield strength (MPa) Tensile strength (MPa) Yield elongation (%) Elongation at break (%) Elastic modulus (MPa) Flexural properties Flexural strength (MPa) Maximum strain (%) Elastic modulus (MPa) Impact resistance Izod, notched (kJ/m2) Izod, unnotched (kJ/m2) Heat resistance HDT ( C) Vicat penetration ( C) Hardness Rockwell hardness (scale H) PLLA III PDLLA I PDLLA II PDLLA III 47,500 75,000 114,000 49 40 1.7 7.5 3650 53 44 1.4 4.8 4050 53 44 1.5 5.4 3900 84 4.8 3500 86 4.1 3550 88 4.2 3600 1.8 13.5 1.7 14.0 1.8 15.0 51 52 50 53 50 52 78 72 76 weight on tensile and flexural properties is more evident in annealed PLLA than nonannealed PLLA and PDLLA specimens, with an increase of tensile strength from 47 to 66 MPa in the range of molecular weight between 20,000 and 70,000 g/mol (viscosity-average molecular weight (Mv)). Above Mv ¼ 30,000–40,000, the tensile and flexural properties of amorphous PLLA increase more gradually, while for annealed PLLA the same behavior is reached at higher molecular weight, near Mv ¼ 50,000–60,000. In the case of PDLLA, the relationship between the increase in tensile and flexural properties with molecular weight becomes less pronounced when the molecular weight is higher than 45,000–50,000. A similar phenomenon is observed for amorphous PLLA. The dependence of PLA mechanical properties on its molecular weight has been investigated by many authors. According to the work of Engelberg and Kohn [6], the increase of molecular weight from 107,000 to 550,000 g/mol results in a 20% increase in tensile strength. In general it seems that at high molecular weights, the variation of mechanical properties becomes less pronounced. Comparable values of tensile properties were observed by Grijpma et al. [8] who reported tensile strength of 47 MPa, elongation at break of 1.5%, and elastic modulus of 3650 MPa for unoriented PDLA with weight-average molecular weight of 241,000 g/mol. Slightly different tensile properties of commercial grade PLAs are also described by several authors [9, 10]. Extruded sheets of NatureWorks amorphous PLA (grade 4060, $10% D-isomer) show tensile
  • TEMPERATURE EFFECT strength of 44 MPa, elongation at break of 7%, and elastic modulus of 2500 MPa. Semicrystalline PLA from NatureWorks (grade 4032, $1.5% D-isomer) is characterized by tensile strength of 54 MPa, elongation at break of 9%, and elastic modulus of 2600 MPa. The differences among the values reported in the literature are quite small; however, it is useful to highlight that commercial PLA is manufactured by copolymerizing L-Lactide with small amounts of meso-lactide, to reduce material crystallinity and decrease brittleness. Another study reported tensile properties for commercial PLLA with Mw ¼ 84,000 g/mol manufactured by Mitsui Fine Chemicals (grade Lacea H.100-E), which showed tensile strength of 66 MPa, elongation at break of 1.8%, and elastic modulus of 3300 MPa [11]. A lower value of elastic modulus is reported for PLLA manufactured by Purac of 1151 MPa, with a yield stress of 62 MPa and strain at break of 14.5% [12]. It has been suggested that different methods used for specimen preparation and testing can contribute to different mechanical properties. Nevertheless, data reported in the literature are rather consistent in reporting PLA as having high elastic modulus, high tensile strength, and low deformability. 11.2.2 Impact Resistance The impact resistance of PLA is significantly affected by notching. The values of Izod impact resistance of PLLA fall in the range of 2.0–3.0 kJ/m2 for low-crystallinity (3–9%) samples. The values obtained with more crystalline PLLA (45–70%) are in the range of 3.0–7.0 kJ/m2 [7]. PDLLA is characterized by lower values, with notched Izod impact resistance ranging from 1.5 to 2.0 kJ/m2. In the case of unnotched impact strength, the difference between PLLA samples of different crystallinity is even higher: samples of PLLA with low crystallinity (3–9%) show values ranging between 13 and 20 kJ/m2, while samples with higher crystallinity (32–75%) are characterized by unnotched impact strength of 18–35 kJ/m2. Other literature data show values of notched impact strength of PLLA in the range of 2.0–2.6 kJ/ m2 [13] and unnotched impact strength of 12 kJ/m2 [8]. A dependence of notched impact strength on notch radius was observed by Grijpma et al. [8], who reported that the strength values decreased from 2.2 to 1.2 kJ/m2 when the radius decreased from 1.00 to 0.10 mm. It is interesting to note that material processing has considerable effect on the impact resistance of PLA. Very high molecular weight PLLA (Mv ¼ 780,000 g/mol) shows impact strength values that decrease significantly from 47 kJ/ m2 for as-polymerized polymer to 12 kJ/m2 for compression molded samples of the same material [14]. This behavior is clearly due to the loss of crystallinity associated with the fast cooling during the compression molding process. PLA impact resistance also depends on molecular weight. Charpy impact strength values of 8 and 15 kJ/m2 have been 143 reported for PDLLA having molecular weight of 78,000 and 108,000 g/mol, respectively [8, 14]. More generally, Charpy impact strength values obtained with PDLLA range from 3.5 to 22 kJ/m2, related to different molecular weights, while PLLA is characterized by values ranging from 10 to 35 kJ/ m2 [5, 17]. Therefore, since PLA is a material that is characterized by relatively low values of impact resistance, the effect of crystallinity and molecular weight has to be taken into consideration in practical applications. 11.2.3 Hardness Rockwell hardness of PLA generally ranges between 70 and 90 according to scale H [2, 7] and around 120 according to scale L [2]. The Rockwell hardness of PLLA is affected very little by crystallinity, as evidenced by literature values that range from 83 to 88 H for amorphous PLLA and from 82 to 88 H for semicrystalline PLLA. Also the dependence of Rockwell hardness on molecular weight appears quite small, as shown by data summarized in Tables 11.1 and 11.2. However, the effect of glass transition temperature (Tg) is more evident. PDLLA, in fact, is characterized by lower hardness values, ranging from 72 to 78 H. The lower hardness values for PDLLA as compared to PLLA can be explained in terms of the lower Tg of PDLLA [7]. Among biodegradable polymers, PLA is characterized by high elastic modulus and high hardness. These characteristics, which influence the applications of this material, are strictly related to its chemical composition. The presence of vicinal and regularly distributed polar ester groups, in fact, affects physicochemical interactions between polymer chains, reducing the possibility of chain shear during physical deformation. Polar interchain interactions and the consequent high Tg, therefore, are the origin of the high hardness of PLA. 11.3 TEMPERATURE EFFECT Enantiomerically pure PLLA is a semicrystalline polymer with a Tg of 60–70 C and a melting point of about 180 C. Dynamic mechanical analysis (DMA) of PLLA reveals the a-relaxation associated with Tg of the amorphous phase, as a maximum of tan d peak and by loss modulus (E00 ) curves. The values reported in the literature for the maximum of the tan d peak range from 65 to 72 C, measured at 1 Hz [15–17]. As expected, the Ta values associated with glass transition of PLLA differ from the data obtained by DSC measurements. Lower values of PLLA Tg, are obtained by loss modulus, E00 curves, ranging from 52 to 60 C [16, 18–21]. The E00 peak temperature is 58 C for amorphous PLA and  60 C for semicrystalline PLLA, while the tan d peak is at 65 C for both materials [16]. The storage modulus of PLLA
  • 144 MECHANICAL PROPERTIES shows a rapid decrease at the temperature corresponding to a-relaxation; however, it is interesting to note that before this decrease there is a slight E0 increase. This behavior can be interpreted as the rearrangement of macromolecules, relieving the stress generated during processing. This cold crystallization process takes place at a temperature of about 91 C [19]. Therefore, DMA can be easily used to characterize the nonequilibrium crystallization of PLLA. As shown in Tables 11.1 and 11.2, the heat distortion temperatures (HDT) for PLA do not change much with molecular weight. Amorphous and crystalline PLLA show HDT values of 55–57 C and 60–66 C, respectively, therefore PLA heat deflection temperature seems little influenced by its crystallinity: crystalline PLA reaches slightly higher heat resistance than amorphous PLA. This is due to main effect of glass transition temperature onto HDT, effect that is very similar in both cases. Heat distortion temperature value for PDLLA is around 50 C and this difference is easily understandable if we take into consideration the physicochemical properties of this material; in fact, PDLLA exhibits the lowest Tg, around 50 C, approximately corresponding to its HDT value. Vicat penetration is much more influenced by PLA crystallinity. In the case of PDLLA and amorphous PLLA, Vicat penetration values of 52–53 C and 59–60 C, respectively, were reported. Also these values are very near to the Tg of the polymers. On the other hand, crystalline PLLA presents a very different behavior, with values of 157–165 C. This marked difference in Vicat penetration measurements is related to the contribution of crystallinity to thermomechanical properties of this material at a microscopic level [2, 7]. Available literature data about PLA mechanical properties at different temperatures are reported in Table 11.3 [7]. As shown, the flexural properties of PDLLA, amorphous PLLA, and semicrystalline PLLA at 23, 36, and 56 C are quite different. At 36 C, the flexural modulus decreased by 30%, 23%, and 26% as compared to the 23 C data for PDLLA, PLLA quenched, and PLLA annealed, respectively. However, increasing the temperature to 56 C results in a marked drop of the flexural strength of PDLLA to 0.2 MPa. This result is expected due to glass transition at 50 C together with the lack of crystallinity for the PDLLA polymer. Similarly, the quenched amorphous PLLA shows a substantial reduction of flexural strength to 0.4 MPa at 56 C. Only annealed semicrystalline PLLA maintains useful mechanical properties of 28 MPa flexural strength. Evidently, a small amount of crystallinity (9%) for the quenched PLLA is not sufficient to provide thermomechanical properties that are useful in practice. Therefore, solely PLLA that has been subjected to proper crystallization, by means of annealing, orientation, or nucleation, can offer useful properties for end-use applications where the temperature is greater than 50 C. TABLE 11.3 Influence of Temperature on PLA Flexural Properties [7] Sample PDLLA PLLA Quenched Molecular weight (Mv, Da) Tg ( C) Crystallinity (%) Properties at 23 C Flexural strength (MPa) Maximum strain (%) Elastic modulus (MPa) Properties at 36 C Flexural strength (MPa) Maximum strain (%) Elastic modulus (MPa) Properties at 56 C Flexural strength (MPa) Maximum strain (%) Elastic modulus (MPa) 75,000 58,000 47,000 50 0 58 9 59 52 86 4.1 3550 100 4.1 3600 113 4.8 4150 60 3.3 2800 77 3.9 3400 83 4.3 3600 0.2 n.d. 25 0.4 8 50 28 7.5 950 11.4 PLLA Annealed ANNEALING As already stated, by injection molding PLLA, the manufactured articles are generally nearly amorphous, because of the slow crystallization kinetics of PLLA. In general, the process of crystallization in semicrystalline polymers results in embrittlement of the material and hence results in a decrease of fracture toughness. The effect of annealing on the fracture toughness of PLLA has been investigated in material samples quenched and annealed in the work of Park et al. [22]. PLLA samples prepared by a quenching procedure show Tg of 64 C, Tm of 168 C, and crystallinity (Xc) of 2.7%, while the annealed samples are characterized by 66 C, 169 C, and 48% [22]. As evidenced by these data, Tg and Tm slightly increased due to the annealing treatment, which results in crystallization. Generally, the annealed material has higher Xc ranging from 45% to 70%, as determined by DSC analysis, depending on its molecular weight. In impact tests, different types of crack growth behavior were observed by polarized optical microscopy of annealed PLLA specimens, as illustrated in Figure 11.1 from the work of Park et al. [22]. The quenched sample that has been subjected to quasistatic deformation (Figure 11.1a) exhibits multiple crazes in the vicinity of the crack tip, which is typical for amorphous polymers. Under the impact effect, the number of crazes is significantly reduced, as evidenced by Figure 11.1b. The annealed samples show cracks that propagate through spherulites and along spherulite boundaries with few or no crazes formed around the main crack. Also, in the annealed samples it is possible to observe that crack branching occurs at the quasistatic deformation rate, whereas only single cracks are observed at the impact rate [22]. The toughness of PLLA
  • ANNEALING FIGURE 11.1 Polarizing micrographs of crack behavior of PLA: (a) quenched static, (b) quenched impact, (c) annealed static, (d) annealed impact [22]. 145
  • 146 MECHANICAL PROPERTIES FIGURE 11.2 Polarized micrographs of PLLA annealed microstructures: (a) 70 C/3 h, (b) 70 C/ 8 h, (c) 70 C/24 h, (d) 100 C/3 h, (e) 100 C/8 h, (f) 100 C/24 h [23]. under impact loading is therefore significantly increased by annealing. The crystallization temperature peak of PLLA has been estimated at about 123 C, and therefore a proper temperature for annealing is around 100 C. At this thermal condition, crystal growth is faster than at lower temperature. This is illustrated in Figure 11.2, where the microstructures of PLLA specimens annealed at 70 and 100 C at different treatment times are shown with polarized optical microphotography. It is clear that the density and size of spherulites increase with annealing time and temperature [23]. From a quantitative point of view, the Xc of quenched PLLA is generally around 3%, while PLLA annealed at 70 and 100 C can reach Xc of 23% and 56%, respectively [23]. The storage modulus, E0 , of PLLA, both quenched and annealed at 70 C, shows a sudden decrease around 60 C that is close to the Tg of 64 C. This is evidenced by an E0 value that ranges between 5 and 10 MPa at 80 C. PLLA specimens annealed at 100 C maintain a much higher value of storage modulus above 60 C, with E0 around 1000 Pa at 80 C, due to the Xc contribution to the mechanical properties [23]. Samples of PLLA that have been annealed at 100–110 C are characterized by higher values of elastic modulus and
  • ORIENTATION decreased strain at break in comparison to amorphous material (Table 11.1). These data agree well with others reported in the literature [1, 12]. In general, annealing of PLLA is accompanied by increase in tensile and flexural strengths, as well as impact resistance and thermomechanical properties. This increase of mechanical properties seems to reach a maximum around an Xc of 65%. Above this value, material embrittlement becomes predominant [1]. By increasing the annealing temperature from 110 to 160 C, it is possible to increase the elastic modulus while decreasing the strain at break and tensile strength. Samples annealed at 120 C exhibit ductile failure while those annealed at 150 C have transitioned to brittle failure [12]. Even though PLLA deformability is slightly reduced by annealing, its thermomechanical properties are markedly improved by increasing the material Xc. 11.5 ORIENTATION At room temperature, PLA is a brittle polymer and fractures through a crazing mechanism. Research conducted using simultaneous small-angle X-ray scattering (SAXS) and tensile testing showed that the craze features increase in size with the extent of deformation [12]. The submicrocracks increase in height while decreasing in diameter, showing that the cracks elongate in the loading direction. This is consistent with visual observation during deformation, in which samples are observed to thin out and elongate in the loading direction. In dry amorphous PLLA, although new crazes are constantly formed during deformation, the dominating factor during deformation may be considered to be the changing dimensions of crazes that are generated soon after yield. Simultaneous SAXS and tensile testing of annealed PLLA results in a characteristic pattern that has been observed and 147 discussed by several authors and interpreted as combined cavitation and fibrillated shear. This deformation mechanism is not uncommon in polymeric lamellar systems [12]. Crazing involves the extension of polymeric chains forming fibrils that span the craze [8]. The brittle fracture behavior in tensile and impact testing is due to the crazing mechanism through which the polymer fails. Due to the very localized nature of this process, the craze fails and wide crack growth occurs. The crazing behavior of PLA is shown in Figure 11.3, illustrating the slight deformation of a specimen, followed by cracking. It should be noted that when PDLLA is tested in the compression mode, crazing is suppressed and its behavior is quite ductile. Here, the sample yields at a compressive strain of more than 30% and a compressive stress in excess of 100 MPa [8]. Rubber modification through copolymerization is very effective in increasing the toughness of PLA; however, this is accompanied by significant reduction of elastic modulus and tensile strength. Another significant route to improve PLA mechanical properties is drawing and orientation. As reported by Grijpma et al. [8], PLA orientation at a stretch ratio of l ¼ 2.5 increases tensile strength from 47 (unoriented polymer) to 73 MPa (oriented polymer), elongation at break from 1.5% to 48%, E modulus from 3650 to 4490 MPa, and impact strength (Izod, 0.25 mm notch) from 1.6 to 5.9 kJ/m2. However, perpendicular to the direction of orientation the mechanical properties are usually poorer. It is therefore important to have biaxial molecular orientation to minimize anisotropy and obtain strong PLA films with improved mechanical properties. There are two kinds of crystalline structures in PLA, depending on the formation conditions: the a-form (orthorhombic), which is obtained by crystallization from melt or solution, and the b-form (orthorhombic or trigonal). When an FIGURE 11.3 Brittle fracture of amorphous PLA: (1) specimen before deformation, (2) crazes at a strain of 3.5%, (3) break at 5.5% strain. Photos are kindly provided by Novamont SpA.
  • 148 MECHANICAL PROPERTIES amorphous PLLA film is drawn by tensile force above the Tg, an oriented film with a-crystallites is obtained. When a PLLA semicrystalline film with a-crystallites is drawn, some of the a-crystallites are transformed into the oriented b-crystallites, depending on the draw conditions. The b-crystallites are generated upon tensile drawing at a high temperature and higher draw ratio, whereas drawing at a lower temperature and/or a lower draw ratio leads to a-crystallites. Thus, the drawn products of PLLA commonly consist of a-crystallites or a mixture of a- and b-crystallites [24, 25]. Biaxial orientation of PLLA film is effective in increasing the tensile strength, elongation at break, and elastic modulus. Stretching conditions may be selected within the range of 1.5–6 times in both the longitudinal and lateral directions. However, in view of film strength and evenness of thickness, stretching is preferably two times or more in both directions. The stretching is typically carried out at a temperature between the Tg and Tm of the material, between 70 and 90 C, depending on the stretching method [26]. The tensile strength of a semicrystalline PLLA film can be increased from 50–60 MPa (unoriented film) to 100–200 MPa (biaxially oriented film), while elongation at break can change from 10% to 50–150% and the elastic modulus increase from 2500 MPa to around 3300 MPa, in the machine direction [9, 26]. These data are also confirmed by mechanical testing of commercial grade PLA 4042D ($8% D-isomer) from NatureWorks LLC. Biaxial orientation of the extrusion cast film results in MD elongation at break of 160% and CD elongation at break of 100% [27]. The same material, when extruded without biaxial orientation, has much lower elongation at break values of only around 10%. Elmendorf tear strength values for these materials after orientation, however, are 6 and 5 N/mm in MD and CD, respectively. For unoriented samples, the Elmendorf tear strengths in MD and CD are 4 and 8 N/mm, respectively [27]. The contribution of biaxial orientation in improving the mechanical properties of PLA films is therefore quite evident and this route represents an important tool for manufacturing biodegradable PLA films with good characteristics for several practical applications. The development of different microstructures as a result of the thermomechanical conditions during injection molding depends upon the operative parameters involving melt processing temperature, the injection flow rate, and holding pressure. The shear stress predominantly controls the hot recoverable strain applied during processing and the degree of crystallinity, and both parameters increase with increasing PLA shear stress. The work of Ghosh et al. [28, 29] with optically pure PLLA shows that a low melt processing temperature facilitates the attainment of a high level of molecular orientation and higher degree of crystallinity. The elongation at break increases with increasing molecular orientation and decreases with increasing crystallinity. Both the level of molecular orientation and crystallinity contribute to the maximum tensile strength of injection molded specimens. 11.6 STEREOREGULARITY As already described, PLA can be manufactured to give a wide range of properties because of the chiral nature of lactide. The mechanical characteristics of PLA are known to depend on the choice and distribution of stereoisomers within the polymer chains. High-purity L- and D-lactide form stereoregular isotactic PLLA and PDLA, respectively, with equivalent physicochemical and mechanical properties. These are semicrystalline polymers with a high Tm around 175–180 C and a Tg in the 60–70 C range. The racemic D,Llactide and meso-lactide, on the other hand, form atactic PDLLA and meso-poly(lactide), which are amorphous materials [30–32]. It is quite interesting to compare the behavior of commercial grade PLA 4030D, which was prepared with nominally 98% L-lactide, with the characteristics of PLA 4040D, prepared with nominally 94% L-lactide. While tensile strength and elastic modulus are comparable for both materials and only slightly higher for PLA 4040D, the elongation at break appears quite different. PLA 4030D shows elongation at break values of 11% MD and 5% CD, while for PLA 4040D values are 78% MD and 97% CD. These data show that, besides biaxial orientation, film grade PLLA may benefit from a small amount of amorphous PLA, possibly because of lower crystallite dimensions [10]. It has been observed that a 1 : 1 mixture of pure PLLAwith pure PDLA yields a stereocomplex of the two polymers during crystallization or polymerization. The PLA stereocomplex consists of racemic crystalline structures in which LPLA and D-PLA chains are packed side by side, with a 1 : 1 ratio of L:D monomer units [4, 32, 33]. While the melting temperature of a- and b-crystalline forms of PLA falls in the range 170–180 C, the Tm of PLA stereocomplex is between 220 and 230 C [33]. The high Tm of PLLA/PDLA stereocomplex makes it a difficult material for processing; however, it is interesting to note that the comparison between PLLA/PDLA equimolar blends and the starting materials shows mechanical properties that are markedly improved. Figure 11.4 shows the stress–strain curve for a film obtained from a 1 : 1 blend of PLLA:PDLA, both having a molecular weight Mw ¼ 150,000 g/mol, compared to the behavior of the film made with the PLLA component. Tensile strength and elongation at break of this blended PLA are 46 MPa and 3.8%, respectively. These values are higher than the values for PLLA alone, which are 22 MPa and 2.3%, respectively [34]. The difference in mechanical properties between PLLA/PDLA blend film and PLLA is due to the very different morphology of these polymers, as evidenced by Figure 11.5. As can be seen, the PLLA films are composed of
  • PLASTICIZATION 149 chanical and thermomechanical properties in poly(lactic acid) family. 11.7 FIGURE 11.4 Stress–strain curves for blend film and nonblended PLLA film. The blend was a 1 : 1 mixture of PLLA with Mw ¼ 150,000 g/mol and PDLA with Mw ¼ 150,000 g/mol [34]. normal spherulites, whereas for the blend films, microcrystallites are observed [34]. When the orientation of PLA film is performed on equimolar blends of PLLA and PDLA, in which a mixture of stereocomplex crystallites and a-crystallites are formed [31, 34], it exhibits relevant mechanical effects. Sawai et al. studied the mechanical properties of a stereocomplex PLA film prepared by casting from a solution of an equimolar blend of PLLA and PDLA [35]. The film was uniaxially drawn by solid-state coextrusion and characterized by DMA. The optimum draw temperature resulting in the highest draw and mechanical properties was 200 C. The maximum achieved tensile modulus and strength, for the samples with an extrusion draw ratio of 16 and prepared by solid-state coextrusion of a highly crystalline stereocomplex film, were 9500 and 410 MPa, respectively. Furthermore, the PLA stereocomplex films with an extrusion draw ratio of 16 exhibited excellent thermomechanical stability as evaluated by the E0 measured as a function of temperature. The reported E0 values at room temperature, 100 and 200 C, were 9500, 7000, and 3000 MPa, respectively [35]. Equimolar amounts of PLLA and PDLA stereocomplex are therefore characterized, upon orientation, by the most relevant me- PLASTICIZATION PLLA is characterized by good mechanical properties and clarity in addition to a satisfying processability; however, its brittleness is its major drawback for many applications [36]. For PLA, Elmendorf tear test values range from 5 to 15 N/mm, which is quite low compared to conventional film grade polyolefins. As PLLA is similar to polystyrene, it is also a comparatively brittle and stiff polymer with low deformation at break. Moreover, extrusion casting of PLA film for packaging and many other applications is relatively difficult compared to other polymers. In fact, there is no tolerance for PLA film tearing or cracking when subjected to forces during package manufacturing [19]. It therefore becomes important to modify these properties in such a way that PLA is able to compete with other more flexible commodity polymers such as polyethylene, polypropylene, PET, or PVC. ThereareseveralmeansofimprovingtheflexibilityofPLLA by modifying its physical properties, such as through copolymerization. An example of a suitable comonomer is e-caprolactone, which can result in soft copolymers [37, 38]. Also blending with biomaterials, such as thermoplastic starch [17], is a feasible method to prepare PLA-based materials that can be successfully used for film manufacturing. However, an interestingandpracticalroutetomodifythemechanicalpropertiesof PLLA consists of introducing biodegradable plasticizers in this material composition [39]. In general, plasticizers improve the ductile characteristics of PLA. It is also well known that lactide monomeritselfis an effective plasticizingagent forPLLA, butit has the disadvantage of rapid migration due to the small molecularweightofthiscompound.TheplasticizationofPLLA with lactide therefore results, besides accelerate aging, in a stiff polymer with a sludgy surface [2, 5]. The introduction of plasticizers in the formulation of semicrystalline polymers such as PLA can in principle reduce not only the Tg of the amorphous phase but also the FIGURE 11.5 Representative polarizing microscopic photographs of 1:1 PPLA:PDLA blend and PLLA films. The molecular weight of PLLA and PDLA are 120,000 and 100,000 g/mol, respectively [34].
  • 150 MECHANICAL PROPERTIES Tm of the crystalline domains. Even though the Tm of PLLA can be in part reduced without significantly affecting its thermomechanical properties from a practical point of view, it is quite important to also avoid depressing excessively the melting point. Therefore, the choice of a suitable plasticizer for PLA has to take this aspect into consideration, with the aim of plasticizing this polymer without affecting its heat resistance and without modifying its compostability. A second aspect to be considered is the drop of tensile strength, which should be minimized as much as possible. Labrecque et al. demonstrated good miscibility of PLA with different citrate esters at up to 20–30% by weight of plasticizer concentration, depending on the type of citrate employed [36]. Tensile strength and elongation at break of PLA plasticized with citrate esters are reported in Table 11.4, and show a marked effect of plasticizer at concentrations between 10% and 30%. As expected, all of the plasticizers decrease the tensile strength of PLA significantly (by about 50%) even at 10% concentration, and the deterioration is larger at higher concentrations. On the other hand, elongation at break does not show any significant change at the lower percentages but significantly increases at higher concentrations in all cases [36]. Similar data have been published by Ljungberg et al. showing that stress at break of PLA is lowered from the value of 62 MPa for neat material to 30 MPa in the presence of 15% triacetine and to 35 MPa with 15% tributyl citrate, after 1 day of storage. Elongation at break changes from 6% for the selected PLA to 355% and to 350%, respectively, with 15% triacetine and with 15% tributyl citrate [40, 41]. Ljungberg et al. [40, 41] reported data about DMA characteristics of plasticized PLA. Figure 11.6 shows the evolution of viscoelastic storage and loss moduli as a funcTABLE 11.4 tion of temperature for pressed films of neat PLA and blends containing 15 wt% plasticizer. TbC, TbC-3, and TbC-7 are, respectively, tributyl citrate and oligomeric tributyl citrate obtained by transesterification with diethylene glycol, having 3 and 7 repeating units. The thermograms show the a-relaxation (between 30 and 60 C) of the materials and the effects of cold crystallization (between 80 and 100 C). It is clearly shown in Figure 11.6a that the drop in storage modulus following the a-relaxation is found at a lower temperature for the plasticized materials as compared to neat PLA. The same trend can be observed in Figure 11.6b, where Ta is significantly decreased for all the plasticized materials, with the plasticizer having the lowest molecular weight, that is, TbC, being the most effective [21]. Malonate esters, such as diethyl bishydroxymethyl malonate (DBM), have also been tested as plasticizers for PLA. These compounds lowered Tg of PLLA from an initial value of 54 to 30 C with 15 wt% and to a Tg of 5 C with plasticizer at 25 wt%. However, at a concentration of 20–25 wt%, phase separation occurs because of saturation of the amorphous phase of PLA. Aging a blend containing 15 wt% DBM for 4 months at ambient temperature also leads to phase separation and to the migration of the plasticizer to the film surface. Malonate oligomers, prepared by reacting DBM with acid chlorides, lead to more stable materials in aging tests [18, 20]. It is also worth mentioning here the need to select PLA plasticizers that are stable during material aging, to avoid unfavorable property changes during the intended material life. Polyethylene glycol (PEG) is another compound that has been tested as a plasticizer for PLA [11]. According to Martin and Avrous [17], PEG with a molecular weight of 400 g/mol e reduces the Young’s elastic modulus (E) of a pure PLLA from an initial value of 2050 to 1488 MPa at a concentration of Thermal and Mechanical Properties of PLA Plasticized with Different Citrate Esters [36] Tg ( C) PLA Triethyl citrate (wt%) 10 20 30 Tributyl citrate (wt%) 10 20 Acetyl triethyl citrate (wt%) 10 20 30 Acetyl tributyl citrate (wt%) 10 20 Tm ( C) DH (J/g) Tensile Strength (MPa, Yield) Elongation at Break (%) 59.1 145.2 0.79 51.7 7 42.1 32.6 22.0 134.1 130.9 126.8 0.31 2.86 7.57 28.1 12.6 7.2 21.3 382 610 40.4 17.6 143.1 139.0 0.06 19.1 22.4 7.1 6.2 350 50.8 30.0 14.2 141.7 138.1 131.6 0.91 0.91 18.34 34.5 9.6 7.6 10 320 228 25.4 17.0 139.2 138.9 1.4 3.9 17.7 9.2 2.3 420
  • RELAXATION AND AGING 151 of Martin and Avrous shows that plasticization with olige omeric lactic acid (OLA) reduces the elastic modulus of the same PLLA from 2050 to 1256 MPa at 10% plasticizer and to 744 MPa at 20% plasticizer, while elongation at break increases to 32% and 200%, respectively [17]. In general, the Tg values of plasticized PLA can be predicted with Fox’s law [39]. Plasticization of PLA affects quite significantly the mechanical properties of this material; however, the choice of a plasticizer and its concentration requires great attention to avoid losing physicochemical characteristics such as heat of fusion and, consequently, thermomechanical properties. 11.8 RELAXATION AND AGING When PLLA is quenched from the melt and vitrifies, a nonequilibrium glassy state is reached. Even in the glass, short-range mobility produces molecular rearrangements that drive the thermodynamic variables closer to their equilibrium values. The mobility of the polymer chains, that is, the ability to eliminate excess free volume, is directly related to temperature. Figure 11.7 shows the dynamic mechanical spectrum reported by Celli and Scandola [42] for PLLA after heating the sample at 200 C in order to erase the thermal history. The solid line refers to a sample quenched in a water–ice mixture after extrusion, while the broken line depicts an immediate rerun on the same sample, after cooling from 160 C. Below room temperature, no relaxation process is apparent in either curves, that is, the dynamic mechanical loss tangent is as low as 10À2 over the range À150–20 C [42]. The absence of any loss phenomena below Tg capable of mechanical energy dissipation is likely the reason for the observed brittleness of glassy PLLA and induces failure of lower molecular FIGURE 11.6 (a) Storage modulus curves as a function of temperature from DMA runs comparing blends containing 15 wt % TbC, TbC-3, and TbC-7 with neat PLA. (b) Loss modulus curves as a function of temperature from DMA runs comparing blends containing 15 wt% TbC, TbC-3, and TbC-7 with neat PLA. TbC, TbC-3, and TbC-7 are, respectively, tributyl citrate and oligomeric tributyl citrate obtained by transesterification with diethylene glycol, having 3 and 7 repeating units [21]. 10% and to 976 MPa at 20%, while the elongation at break increases from the initial level of 9% to 26% and 160%, respectively [17]. In general, the storage modulus drop associated with the Tg of plasticized PLA decreases almost linearly with increasing plasticizer concentration, from 67 C for pure PLA to 54 and 46 C for a PEG concentration of 10% and 20%. The work FIGURE 11.7 Dynamic mechanical spectrum of poly(L-lactic acid) Resomer L 214, Mv ¼ 690,000, (—) first run, (- - -) second run [42].
  • 152 MECHANICAL PROPERTIES weight samples. In the first run on the quenched PLLA, the loss tangent curve shows a very intense relaxation peak in the range 50–80 C (peak value 70 C), which is associated with a very steep drop in modulus. This relaxation phenomenon corresponds to the glass/rubber transition of PLLA and its intensity indicates that a very large fraction of the sample has been frozen in the glassy amorphous state after extrusion. The abrupt increase in E0 above 80 C reflects a sudden increase in rigidity of the material and indicates that, above Tg, the polymer chains have acquired enough mobility to crystallize during the DMA run. The occurrence of crystallization is clearly demonstrated by the spectrum obtained in an immediate rerun (broken line in Figure 11.7), where the Tg of PLLA appears as a tan d peak of moderate intensity associated with a fairly small decrease in the storage modulus. In this instance, only a small fraction of the sample undergoes the glass transition, owing to the presence of a significant crystalline phase. The crystallites cause the observed broadening of the tan d peak [42]. Figure 11.8 shows how the glass transition region of the viscoelastic spectrum of PLLA is affected by aging. Here, the broken curves refer to a sample aged for 9 days at 48 C, while the solid lines are for a sample without aging, which was obtained after an immediate rerun of the same sample cooled from 100 C. The aging effect exemplifies as decreasing the loss factor and increasing the storage modulus over the whole temperature range tested. As a result of aging, the mechanical characteristics of PLLA change due to the reduced ability for the material to dissipate energy through reduction of molecular motion. The increased rigidity also results in a more brittle material [42]. Hot recoverable strain is also an important parameter directly related to thermomechanical environment applied during processing. PLA hot recoverable increases steadily with increasing shear stress, as a consequence of increased molecular orientation. The tests performed on injection molded PLA specimens show longitudinal shrinkage ranging from 5.7% to 11.6%, depending on the shear rate used during the injection molding process. This relation is, however, quite moderate and it tends to diminish as the shear rate increases [28]. 11.9 CONCLUSIONS PLA has attracted the attention of industry for packaging and biomedical applications because of its biodegradability characteristics, in addition to its good mechanical and physicochemical properties. From a general point of view, PLA is a rigid, semicrystalline material, with tensile strength ranging from 50 to 70 MPa, modulus of elasticity of 3000–4000 MPa, and elongation at break of 2–5%. However, the toughness of PLA polymers can be successfully modified to meet specific applications by changing the formulation and manufacturing process. Thus, PLA is a versatile polymer that can be converted into several materials having useful properties for many different applications. The copolymerization of L-lactide with minor amounts of meso-lactide or D, L-lactide offers materials with lower brittleness and better film properties. In general, annealing improves the thermomechanical properties. Therefore, extrusion or injection molding of PLA must be optimized. Proper crystallization of PLLA significantly improves the thermomechanical behavior, and this is an important requisite for applications at temperatures of 40–50 C or higher. Biaxial orientation of PLLA film is an important method to increase tensile strength, elongation at break, and modulus of elasticity. Here, the stereocomplex formed by equimolar amounts of PLLA and PDLA, particularly when subjected to biaxial orientation, exhibits very interesting thermomechanical properties, and it is therefore a promising material for new applications. Finally, the plasticization of PLA can have a strong effect on the mechanical behavior of this material, by increasing the elongation at break to about 300% with 20% by weight of a plasticizer such as tributyl citrate. REFERENCES  FIGURE 11.8 Influence of aging on the dynamic mechanical spectrum of PLLA. Aging time: (—) zero, (- - -) 9 days at 48 C [42]. 1. A. S€dergard, M. Stolt, Prog. Polym. Sci. 2002, 27, o 1123–1163. 2. M. Ajioka, K. Enomoto, K. Suzuki, A. Yamaguchi, J. Polym. Environ. 1995, 3, 225–234. 3. H. Urayama, T. Kanamori, Y. Kimura, Macromol. Mater. Eng. 2001, 286, 705–713. 4. D. Garlotta, J. Polym. Environ. 2001, 9, 63–84. 5. S. Jacobsen, H. G. Fritz, Polym. Eng. Sci. 1999, 39, 1303–1310. 6. I. Engelberg, J. Kohn, Biomaterials 1991, 12, 292–304.
  • REFERENCES 7. G. Perego, G. D. Cella, C. Bastioli, J. Appl. Polym. Sci. 1996, 59, 37–43. 8. D. W. Grijpma, H. Altpeter, M. J. Bevis, J. Feijen, Polym. Int. 2002, 51, 845–851. 9. L. Yu, H. Liu, F. Xie, L. Chen, X. Li, Polym. Eng. Sci. 2008, 48, 634–641. 10. R. A. Auras, B. Harte, S. Selke, R. J. Hernandez, J. Plast. Film Sheet. 2003, 19, 123–135. 11. M. Baiardo, G. Frisoni, M. Scandola, M. Rimelen, D. Lips, K. Ruffieux, E. Wintermantel, J. Appl. Polym. Sci. 2003, 90, 1731–1738. 12. A. C. Renouf-Glauser, J. Rose, D. F. Farrar, R. E. Cameron, Biomaterials 2005, 26, 5771–5782. 13. K. S. Anderson, K. M. Schreck, M. A. Hillmyer, Polym. Rev. 2008, 48, 85–108. 14. D. W. Grijpma, A. J. Nijenhuis, P. G. T. van Wijk, A. J. Pennings, Polym. Bull. 1992, 29, 571–578. 15. Z. Ren, L. Dong, Y. Yang, J. Appl. Polym. Sci. 2006, 101, 1583–1590. 16. Z. Kulinski, E. Piorkowska, Polymer 2005, 46, 10290–10300. 17. O. Martin, L. Avrous, Polymer 2001, 42, 6209–6219. e 18. N. Ljungberg, D. Colombini, B. Wessln, J. Appl. Polym. Sci. e 2005, 96, 992–1002. 19. N. Ljungberg, B. Wessln, Biomacromolecules 2005, 6, e 1789–1796. 20. N. Ljungberg, B. Wessln, J. Appl. Polym. Sci. 2004, 94, e 2140–2149. 21. N. Ljungberg, B. Wessln, Polymer 2003, 44, 7679–7688. e 22. S. D. Park, M. Todo, K. Arakawa, J. Mater. Sci. 2004, 39, 1113–1116. 23. S. D. Park, M. Todo, K. Arakawa, M. Koganemaru, Polymer 2006, 47, 1357–1363. 24. D. Sawai, K. Takahashi, A. Sasashige, T. Kanamoto, Macromolecules 2003, 36, 3601–3605. 25. K. Takahashi, D. Sawai, T. Yokoyama, T. Kanamoto, S. H. Hyon, Polymer 2004, 45, 4969–4976. 153 26. S. Kumar, S. Akhtar, V. Kumar, Popular Plast. Packag. 2005, 50, 85–89. 27. NatureWorks technical data sheet, available at http://www. natureworksllc.com/product-and-applications/natureworksbiopolymer/technical-resources/$/media/Product%20and% 20Applications/NatureWorks%20Biopolymer/Technical% 20Resources/Technical%20Data%20Sheets/TechnicalDataSheet_4042D_pdf.ashx. 28. S. Ghosh, J. C. Viana, R. L. Reis, J. F. Mano, Polym. Eng. Sci. 2007, 47, 1141–1147. 29. S. Ghosh, J. C. Viana, R. L. Reis, J. F. Mano, Mater. Sci. Eng. 2008, A 490, 81–89. 30. R. Mehta, V. Kumar, H. Bhunia, S. N. Upadhyay, Polym. Rev. 2005, 45, 325–349. 31. C. C. Chen, J. Y. Chueh, H. Tseng, H. M. Huang, S. Y. Lee, Biomaterials 2003, 24, 1167–1173. 32. J. R. Dorgan, H. Lehermeier, M. Mang, J. Polym. Environ. 2000, 8, 1–9. 33. M. Spinu, C. Jackson, M. Y. Keating, K. H. Gardner, J. Macromol. Sci. A, 1996, 33, 1497–1530. 34. H. Tsuji, Y. Ikada, Polymer 1999, 40, 6699–6708. 35. D. Sawai, M. Tamada, T. Kanamoto, Polym. J. 2007, 39, 953–960. 36. L. V. Labrecque, R. A. Kumar, V. Dav, R. A. Gross, e S. P. McCarthy, J. Appl. Polym. Sci. 1997, 66, 1507–1513. 37. D. W. Grijpma, G. J. Zondervan, A. J. Pennings, Polym. Bull. 1991, 25, 327–333. 38. G. Perego, T. Vercellio, G. Balbontin, Macromol. Chem. 1993, 194, 2463–2469. 39. I. Pillin, N. Montrelay, Y. Grohens, Polymer 2006, 47, 4676–4682. 40. N. Ljungberg, T. Andersson, B. Wessln, J. Appl. Polym. Sci. e 2003, 88, 3239–3247. 41. N. Ljungberg, B. Wessln, J. Appl. Polym. Sci. 2002, 86, e 1227–1234. 42. A. Celli, M. Scandola, Polymer 1992, 33, 2699–2703.
  • 12 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) EVA ALMENAR AND RAFAEL AURAS 12.1 INTRODUCTION The permeation, diffusion, and sorption of gases, vapors, and organics in poly(lactic acid) (PLA) are currently being actively studied by research groups around the world. However, to the best of our knowledge, no comprehensive summary and discussion of this important topic has been published to date. Hence, in this chapter, following a brief introduction to mass transfer phenomena, the current research findings about permeation, sorption, and diffusion of gases, vapors, and organics in pure PLA and in its blends, composites, nanocomposites, and fiber reinforcements are presented. Mechanisms and expressions relating solubility and transport of substances are reviewed. The effects of intrinsic and extrinsic factors are also discussed. Permeability is the transfer of small molecules called ‘‘permeants’’ or ‘‘penetrants’’ through a membrane [1]. Large permeability values are attributed to poor barrier membranes, and small permeability values to good barrier membranes. Plastic materials are relatively permeable to small molecules such as gases, water vapor, organic vapors, and liquids. Plastic materials provide a broad range of permeabilities to those molecules and thus can range from being excellent to poor barriers. Permeation processes take place through a membrane that separates two fluid phases (gas, vapor, or liquid) containing low molecular weight species with differing chemical potential. The molecules at the side of the higher chemical potential tend to diffuse through the membrane and equilibrate the chemical potential on both sides. The permeation process in continuous polymeric materials is accomplished by molecular diffusion within the polymer matrix. Figure 12.1 shows a sketch of a diffusion process through a membrane. Diffusion is a phenomenon controlled by temperature. Therefore, above the glass transition temperature (Tg) of a polymer, the molecules (phase I) are first sorbed into the polymer and then diffuse through the polymer by an activated process of random movement induced by the Brownian motion of the polymer chains [2]. Below Tg, the permeant dwells in a cavity of the polymer for a while and then performs a quick jump into an adjacent cavity. The diffusing molecules mainly move in the amorphous regions of a polymer within the free volume between the polymeric molecules. Since diffusion is a phenomenon controlled by temperature, it can be modeled by the Arrhenius equation. Upon diffusion through the membrane, the molecules are desorbed from the membrane to the low chemical potential side (phase II). For noncontinuous polymeric materials, the transport of penetrants involves flow through pores, and, therefore, the solubility–diffusion phenomenon is not the main mechanism dictating the transport process. Permeation mechanisms have been extensively studied by many researchers [3–5]. Adolf Fick, in 1855, derived the general equations of the diffusion process, and in 1975 a detailed mathematical analysis was compiled by John Crank [6]. The diffusion process for the steady state is described by the general Fick’s first law (Equation 12.1, where the left-hand side of the equation is the flux (F), and the right-hand side is the product of the diffusion coefficient (D) and the gradient operator of the chemical potential (m)). F ¼ Drm ð12:1Þ Poly(lactic acid): Synthesis, Structures, Properties, Processing, and Applications, edited by R. Auras, L.-T. Lim, S. E. M. Selke, and H. Tsuji Copyright Ó 2010 John Wiley & Sons, Inc. 155
  • 156 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) Ft ¼ F1   rffiffiffiffiffiffiffiffi! X  2 2 1 4 l2 Àn l pffiffiffi exp 4Dt n¼1;3;5;... 4Dt p ð12:4Þ where Ft is the flow rate of the compound permeating the film in the transient state at time t, F1 is the equilibrium compound transmission rate, l is the film thickness, and D is the compound diffusion coefficient. From the transient flow rate profile, the half-time diffusion coefficient, D, can be estimated from Equation 12.5 [6]: D¼ FIGURE 12.1 Permeation mechanism of permeants from higher (C1) to lower concentration (C2) through the continuous polymeric material of thickness l. Steps: (1) sorption, (2) diffusion, and (3) desorption. The general dynamic process of Fick’s second law can be described by a most general expression (Equation 12.2). qm ¼ r2 Dm qt ð12:2Þ where t is time. In the case of a membrane where the diffusion coefficient D is independent of the concentration (c) and the chemical potential can be expressed by the concentration, the one-dimensional form of Equation 12.2 is equal to qC d2 c ¼D 2 qt dx ð12:3Þ In Equation 12.3, D is normally assumed to be independent of both penetrant concentration and polymer relaxations at low concentration, and x is the thickness of the membrane. This is especially true for gases such as oxygen (O2), carbon dioxide (CO2) at atmospheric pressure, and some organic compounds. Many theories have been proposed and many models have been developed to describe diffusion in polymers; a detailed description of these models can be found elsewhere [7]. The diffusion processes through the membrane can generally be considered unidirectional and perpendicular to the flat surface, and solutions to the diffusion equations are obtained from the boundary conditions where the Henry’s or Langmuir– Henry’s law is applied. When D is independent of c in a flat membrane of thickness (l), and the concentration at one side of the membrane at time zero is higher than the other side of the membrane, the solution of the one-dimensional equation (12.3) can be expressed by Equation 12.4 [6]: l2 7:199t0:5 ð12:5Þ where t0.5 is the time when Ft/F1 ratio is 0.5; using the value of permeant flow at steady state F1, the permeability coefficient P is determined by P¼ F1 l ADp ð12:6Þ where A is the area of the film, and Dp is the partial pressure gradient across the polymer film. Then, if Henry’s law of solubility applies, such as in the case of diffusion of O2 and CO2 in films at low pressure (i.e., up to 1 atm), the solubility coefficient, S, can be calculated from P ¼ SD ð12:7Þ Terms other than permeability (Equation 12.6), such as permeant transmission rate, permeance, and thickness normalized flow, can be used to describe the steady-state permeation of molecules through the polymer films [2]. Permeant transmission rate is the amount of permeant passing through a plane of unit area normal to the direction of the flow during unit time (Equation 12.8). The term permeance is used when differences in partial pressure between both sides of the material are also taken into account (Equation 12.9), whereas thickness normalized flow considers material thickness but not difference in partial pressure (Equation 12.10). Permeant transmission rate : F¼ q At ð12:8Þ Permeance : R¼ q AtDp ð12:9Þ Thickness normalized flow : N¼ ql At ð12:10Þ For all equations, q is the amount of permeant. Under the SI system, the units of P would be at standard temperature and
  • FACTORS AFFECTING PERMEABILITY, SORPTION, AND DIFFUSION IN PLA pressure conditions m/(m2 s Pa) or kg m/(m2 s Pa). However, the unit Barrer (1 Barrer ¼ 1.33  10À4 m3 (STP) m/(m2 s kPa) may be more suitable when small quantities of permeant are measured. Factors for converting permeabilities from various units to the standard units can be found in the literature [8]. The temperature dependence of P, S, and D in Equation 12.7 can be described by the Van’t Hoff–Arrhenius equations (Equations 12.11–12.13) [9] SðTÞ ¼ S0 expðÀDHS =RTÞ ð12:11Þ DðTÞ ¼ D0 expðÀED =RTÞ ð12:12Þ PðTÞ ¼ P0 expðÀEP =RTÞ ð12:13Þ where DHS is the molar heat of sorption, ED is the activation energy of diffusion, and EP is the apparent activation energy of permeation. It follows that 157 where xc is the degree of crystallinity, and Da and Sa are the diffusion and solubility coefficients in the amorphous part of the polymer. The permeability rate for gases, vapors, and organics of a given polymer is a function of many parameters (see Figure 12.2). These parameters are basically the structural and morphological properties of both the polymer matrix and the permeant. In addition, humidity, temperature, additives, and plasticizers are factors that must be taken into consideration. Chemical structure plays a very important role since the permeability value is specific to the particular polymer/ permeant system. The permeation of a chemical compound in a polymer relates to the chemical structure of the pair. Polar polymers can interact in various degrees with polar and nonpolar permeants. Polymer morphology plays a fundamental role in barrier properties since increasing crystallinity, orientation, and amorphous density decrease permeability. Humidity may have an effect by decreasing or increasing permeability (especially in hydrophilic polymers). Plasticizers in a polymer usually result in increased permeability. P ¼ S0 D0 ð12:14Þ 12.2 FACTORS AFFECTING PERMEABILITY, SORPTION, AND DIFFUSION IN PLA EP ¼ DHS þ ED ð12:15Þ 12.2.1 DHS can be expressed as the sum of the heat of condensation, DHC, and the heat of mixing, DHM, as in the following equation: DHS ¼ DHC þ DHM ð12:16Þ Values for P0, S0, D0, EP, DHS, and ED can be derived from isostatic permeation experiments. In this experiment, one side of the membrane is exposed to a higher concentration of the permeant, and at the other side is swept with a carrier gas (N2 or H2) that delivers the diffused permeant to a sensor so that the concentration in this side is always near zero. Finally, as a first approximation and since some permeants such as O2 and CO2 are insoluble in the crystalline part of polymer films [10, 11], the diffusion, solubility, and permeation coefficients for semicrystalline polymers can be estimated as Dsc % Da ð1Àxc Þ ð12:17Þ Ssc % Sa ð1Àxc Þ ð12:18Þ P % Dsc Ssc % Da Sa ð1Àxc Þ2 ð12:19Þ L-Lactide Unit Content The ratio of L-enantiomers to D-enantiomers (L/D ratio or L:D) is known to affect the properties of PLA, such as melting temperature, degree of crystallinity, and barrier properties. Control of the L:D monomer content is an important molecular feature of PLAs that has a large effect on barrier properties. Generally, higher L-lactide content should result in higher barrier properties of the polymer because of greater stereochemical purity [12]. 12.2.1.1 Gases The literature is inconsistent about the effect of D-lactide unit content on the permeation of gases. Lehermeier et al. [12] found that small changes in L:D stereochemical content (96:04 versus 98:02) had no effect on permeation properties of pure gases such as CO2, O2, nitrogen (N2), and methane (CH4). Auras et al. [13] also studied the effect of the L-lactic acid content on the permeability to O2 and CO2 of semicrystalline PLA films with densities around 1250 kg/m3 and L-lactide contents of 94% and 98%. At 0% relative humidity (RH) and temperatures in the range of 25–45 C, the film with the higher L-lactide content showed higher value of P for CO2 but lower EP. The values of P reported for O2 followed the same trend. However, the O2EP was higher for PLA films with lower L-lactide content (41.43 Æ 3.50 kJ/mol for PLA with 98% L-lactide versus 28.43 Æ 2.93 kJ/mol for PLA with 94% L-lactide) in contrast to the data reported for CO2 (15.65 Æ 0.63 kJ/mol for PLA with 98% L-lactide versus 19.44 Æ 0.93 kJ/mol for PLA
  • 158 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) Polymer content Crystallization Free volume Orientation Branching L-lactide Intrinsic factors Permeant Type Size Environment Temperature Relative humidity P, D, and S Extrinsic Processing Plasticizers Blends Laminations Coatings Composites Nanocomposites FIGURE 12.2 Examples of extrinsic and intrinsic factors affecting P, D, and S in polymers such as PLA. with 94% L-lactide). Since the EP values for O2 were roughly twice those for CO2, these results demonstrate that oxygen permeability was more sensitive to temperature. 12.2.1.2 Water Vapor Water vapor permeability (WVP) of amorphous PLA films (obtained by solution casting) with different L-lactide unit contents (PLLA, P(LLA-DLA) (77:23), and P(LLA-DLA) (50:50)) have been reported as 2.18 Â 10À14, 1.90 Â 10À14, and 1.95 Â 10À14 kg m/ (m2 s Pa), respectively, at 25 C [14]. These results indicate that L-lactide unit content had little effect on the water vapor permeability of the PLA films. The values are comparable to those measured by Auras et al. [15] (1.98 Â 10À14 and 1.98 Â 10À14 kg m/(m2 s Pa) at 20 C, and 1.65 Â 10À14 and 1.61 Â 10À14 kg m/(m2 s Pa) at 30 C for P(LLA-DLA) (98:2) and P(LLA-DLA) (94:6), respectively). The authors also found that water permeability values were similar for the PLA films exposed to different temperatures and RH. The EP for the water vapor permeation was À9.8 kJ/mol for films made from 98% L-lactide and À10.1 kJ/mol for films made from 94% L-lactide. Cairncross et al. [16] studied moisture sorption and transport in three different PLA films (high percentage Llactide, a mixture of lactic acid stereoisomers, and a 50:50 blend of PLLA and PDLA) exposed to RH changes from 0 to approximately 25% at 40 C using a quartz crystal microbalance/heat conduction calorimeter (QCM/HCC). In addition, moisture transport, crystallization, and degradation in PLA were measured through a variety of experimental techniques including size exclusion chromatography, differential scanning calorimetry (DSC), and X-ray diffraction. Only small differences in the sorption properties of crystalline and amorphous films were observed, whereas stereocomplex PLA exhibited less moisture sorption and slightly greater sorption enthalpy than the other films. All sorption enthalpies were higher than the heat of vaporization of pure water (approximately 44 kJ/mol), and there was a general trend of decreasing sorption enthalpy with increasing moisture content. 12.2.1.3 Organic Compounds To the authors’ best knowledge, no research studies have been conducted to evaluate the effect of lactic acid content on the permeability, diffusion, or solubility of organic compounds in PLA. 12.2.2 Plasticizers Plasticizers are widely used in the plastics industry to improve the processability, flexibility, and ductility of glassy polymers [17]. Many factors are involved in plasticizer selection, including molecular structure, polarity, desired product properties, and cost. In the case of PLA, plasticizers decrease polymer rigidity and brittleness by reducing the glass transition temperature of the polymer (amorphous regions), and thus improving its mechanical properties and recovery. In addition, plasticizers lower the melting point of the crystalline regions, thus reducing the processing temper-
  • FACTORS AFFECTING PERMEABILITY, SORPTION, AND DIFFUSION IN PLA ature and the thermal decomposition. Ljungberg and Wessln [18] reported that the melting temperature of PLA e with plasticizer was 10 C lower than without, regardless of plasticizer concentration. Plasticizers were miscible with PLA to 25 wt%. A negative aspect of plasticizer use is the potential decrease in barrier properties of PLA. Plasticizers such as water, polyethylene glycol, lactide, lactic acid, nontoxic citrates, glycerol, and sorbitol have been tested and reported as effective for PLA. 12.2.2.1 Water Cairncross et al. [16] demonstrated that water acted effectively as a plasticizer in PLA films exposed to changes in RH (from 0% to approximately 25% RH) at 40 C. This conclusion was based on the motional resistance of the equipment, a measure of the increased damping properties of the film, which indicates the plasticizing process. The plasticization of PLA by water has also been investigated by hydrating samples with saturated salt solutions and then using DSC with heating and cooling rates of 10K/min (T) to determine the Tg [19]. The Tg of PLA was dependent on the water vapor concentration. The concentration dependence of D for water was expressed in terms of the WLF (William–Landel–Ferry) shift parameter (aT) (Equation (12.20)) with values of 16 and 50K for the parameters a and b, respectively. These are the same values as those used to describe all glass forming polymers [20]. At very low mass fractions of water, the Tg was driven below body temperature (37 C). This manifested itself as changes of orders of magnitude in physical properties, such as viscosity and storage modulus of the polymer, when the PLA was placed in biological environments.   TÀTg DðTÞ ¼ exp Àa aT ¼ DðTg Þ b þ TÀTg ð12:20Þ to become sludgy [26]. To the authors’ best knowledge, permeability studies evaluating the effect of different amounts of lactide and lactic acid as plasticizers have not been published to date. 12.2.2.4 Nontoxic Citrates Nontoxic citrate plasticizers derived from natural citric acid, such as triethyl citrate (TC), tributyl citrate (TBC), acetyl triethyl citrate (ATC), acetyl tributyl citrate (ATBC), and triacetine, have been shown to be effective plasticizers for PLA [27–29]. Some gas permeability tests have been performed to assess the potential use of PLA and nontoxic citrate plasticizer blends in food packaging and other applications. The effect of ATBC on PLA barrier properties was studied by Coltelli et al. [30] using PLA mixed with ATBC (10–35 wt%), followed by compression molding. Yu et al. [31] blended PLA/ATBC mixtures with carbon black (CB) to form electrically conductive polymer composites. Fourier transform infrared (FTIR) experiments revealed that the interaction between the PLA/ATBC matrix and the CB filler was increased by the addition of ATBC. Water vapor permeability values decreased with an increase in ATBC content (at constant CB levels). For example, at 30 wt% CB, the WVP of the PLA decreased from 0.66 Â 10À13 kg m/(m s Pa) (at 0% ATBC) to 0.10 Â 10À13 kg m/(m s Pa) with the addition of 30% ATBC. 12.2.2.5 Glycerol and Sorbitol The effect of incorporating glycerol or sorbitol as a plasticizer in lactic acid casein films has been compared by Chick and Ustunol [32]. Films plasticized with sorbitol were significantly more effective (p < 0.05) moisture and O2 barriers than glycerol-plasticized films and also had improved mechanical properties. 12.2.3 12.2.2.2 Poly(ethylene glycol) Poly(ethylene glycol) (PEG) is widely used as a plasticizer for PLA [21, 22] and has been reported to affect mechanical, physical, and barrier properties. A PEG content of 16.6% (w/w) can provide a favorable compromise between PLA barrier and mechanical properties [23]. Higher PEG contents result in a higher WVTR owing to progressive film plasticization that modifies the hydrophilic character of PLA film [24]. The incorporation of PEG may decrease material cohesion by creating intermolecular spaces and thus increasing D for water. Siparsky et al. [25] also reported higher WVP for PEG/PLA blends compared to that for PLA. This phenomenon was attributed to hydrogen bonding between water and the ester group of PEG, which leads to high solubility and P values. 12.2.2.3 Lactide and Lactic Acid The lactide monomer is an effective plasticizer for PLA but has limited use because it tends to migrate to the material surface, causing the surface 159 Crystallization In most studies of biodegradable semicrystalline polymers and their blends, the crystallinity does affect water transport. Lower S and lower D have been correlated with higher crystallinity fractions. For example, Yoon et al. [33] reported that crystallinity profoundly impacts permeation in biodegradable polymers such as poly((R)-3-hydroxybutyrate) (PHB), poly(e-caprolactone) (PCL), and succinic acid/adipic acid-1-4-butanediol/ethylene glycol (SG). Olkhov et al. [34] reported that higher crystallinity was correlated to a decrease in S and D in biodegradable blends (PHB/poly(vinyl acetate)). Reports of the effects of crystallinity on permeation in PLA tend to be more inconsistent. Plackett et al. [35] observed that PLA with higher degrees of crystallinity had significantly (p 0.0001) lower moisture-binding capacity. This result is in agreement with the work of Li et al. [36], who found a higher moisture absorption rate in amorphous PLA compared to crystalline PLA. In contrast, other studies have reported that water diffusion and water solubility are nearly
  • 160 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) independent of the crystallinity of PLA. Values of 8.04 Â 10À12 and 6.61 Â 10À12 m2/s were obtained for D of water in amorphous and crystalline PLLA materials, respectively [33]. Siparsky et al. [25] reported little correlation between crystallinity of PLA and water permeability or solubility. These researchers achieved different degrees of crystallinity in the PLA by quenching and annealing or by copolymerizing with D-lactide units. The transport of water molecules in the PLA matrices was understood to follow the water cluster model rather than the solution diffusion model. In the water cluster model, the water molecules sorbed in the polymer matrix form clusters through preferential binding to each other by hydrogen bonding. The water molecules are not necessarily soluble in the polymer, but they permeate through it in the cluster form. Cairncross et al. [16] also reported that sorption and diffusion in PLA were unaffected by crystallinity. In addition, these authors hypothesized that moisture sorption in PLA is controlled by hydrophilic end groups, whether dynamic sorption follows Fickian kinetics or whether water absorbs as clusters. Some researchers such as Shogren [37] and Yoon et al. [33] have reported that the mass of water absorbed into PLA at equilibrium increases with the percentage of crystallinity, xc, but others disagree with this statement. Tsuji et al. [14] achieved different percentages of crystallinity (0–35%) in PLLA films by annealing at 140 C for various times; as xc increased from 0% to 20%, the WVTR decreased monotonically from 0.23 to 0.13 kg/(m2 day) at 25  C and 95% RH, but WVTR leveled off as xc exceeded 30%. This behavior was attributed to a higher resistance to water vapor permeation of the restricted amorphous regions compared to the free amorphous regions. Free and restricted amorphous regions were the major amorphous components in PLLA films with a xc range of 0–20% and exceeding 30%, respectively, resulting in the xc-dependent WVTR described above. According to Kitamaru et al. [38] and Saito et al., [39] the higher resistance of restricted amorphous regions to water vapor permeation may be due to specific amorphous areas called ‘‘crystalline–amorphous interfacial regions’’ or ‘‘transition regions between crystalline and amorphous regions,’’ where mobility is highly restricted by adjacent crystalline regions. A decrease in the WVTR of PLLA films due to annealing has also been reported by Shogren [37]; the WVTR value for as-cast PLLA film decreased from 1.72 to 0.82 kg/(m2 day) after annealing the PLA at 130 C for 10 min. 12.2.4 Orientation The orientation of the polymer molecules that results from different processing technologies, such as tentering or blowing, affects the polymer permeability. When amorphous films of PLA cast from a solution (5 wt% methylene chloride) were biaxially oriented, they exhibited higher barrier properties TABLE 12.1 Effect of Orientation on the Oxygen Permeability of Some Commonly Used Polymers at 23 C O2 Permeability (10À17 kg m/(m2 s Pa)) Polymer Unoriented Poly(ethylene terephthalate) Polypropylene Polystyrene a 0.06 0.88 2.45 Oriented 0.03 0.47 1.75 Note: Adapted from Ref. 40. due to the increase in crystallinity [12]. For example, methane permeation through a biaxially oriented PLA was 4.5 times lower than that of unoriented PLA. Lower permeation in oriented films has also been reported for other plastics as shown in Table 12.1. 12.2.5 Free Volume The free volume is an intrinsic property of the polymer matrix and is created by the gaps between entangled polymer chains. The absorption and diffusion ofmolecules in polymers depend greatly on the available free volume. Generally, sorption increases as the amount of free volume increases. D of a diffusant/polymer matrix system has been reported to depend strongly both on diffusant concentration and on temperature, and can be predicted by the free volume theory [40–45]. Netramai et al.[46] calculated the free volume of PLA and other commercial polymers by the group contribution method to evaluate the barrier properties to chlorine dioxide. They found the following trend in the free volume of these glassy polymers: PET (poly(ethylene terephthalate)) < PLA < PS (polystyrene) that match the barrier properties of these polymers (i.e., PClO2-PET ¼ 1.26 Æ 0.03 Â 10À17 kg m/(m2 s Pa); PClO2-PLA ¼ 5.40 Æ 0.03 Â 10À17 kg m/(m2 s Pa); PClO2-PS ¼ 41.82 Æ 0.82 Â 10À17 kg m/(m2 s Pa)). 12.2.6 Branching Both the number and the length of polymer branches contribute to permeability. The higher the number or the length of the branches, the more permeable is the polymer, and vice versa, due to the free volume created between main chains and branches [47]. According to Gruber et al. [48], branched PLA can be produced by peroxideinitiated cross-linking of linear material during reactive extrusion. Using this technology, Lehermeier et al. [12] obtained six different branched PLAs (96% L-lactide) with increments of 20% by weight. Studies with different pure gases, such as CO2, O2, N2, and CH4, showed that changes in PLA chain branching have no effect on the permeation properties for these gases.
  • FACTORS AFFECTING PERMEABILITY, SORPTION, AND DIFFUSION IN PLA 11 Environmental Factors Temperature and moisture level affect the stability of PLA [49–51]. The effects of these two parameters on PLA barrier properties are discussed separately below. 10 5ºC 23ºC 40ºC 9 Diffusion × 1014(m2/s) 12.2.7 161 8 7 6 12.2.7.1 Temperature The effect of temperature on the sorption, diffusion, and desorption of water in bio-based materials has been studied extensively. Bell and Labuza [52] reported that most biomaterials display a negative heat of sorption; these materials typically take up more water at lower temperatures. In contrast, PLA has been reported to take up moisture at a higher rate at increased temperatures [53]. Holm et al. [53] observed that up to 12 times more water was absorbed at 25 C than at 5 C when PLA was exposed to low and high temperatures at equal RH and for equal amounts of time. The higher moisture sorption at 25 C became very pronounced at high RH and long exposure times. At 98% RH, the moisture content was 86 g/100 g dry matter after 189 day at 25 C. At the same RH but at 5 C, the moisture sorption was only 7 g/100 g dry matter over the same time. The amount of water available (the water concentration) differs substantially at different temperatures at the same RH. Both Auras et al. [54] and Shogren [37] observed a decrease in the water vapor permeability of PLA with increasing temperature. Auras et al. [54] also studied the effect of the temperature on the permeation of O2 in PLA. In this study, PLA films containing different L-lactide content (94% and 98%) were exposed to 5, 23, and 40 C; PET films were used for comparison. A significant increase in the value of P for O2 with increasing temperature was observed for both types of PLA. The values of P for O2 increased by approximately 10% with each temperature step; PET films showed a similar behavior. Lehermeier et al. [12] earlier reported that O2 permeability in 100% linear PLA with an L:D ratio of 96:4 was temperature dependent (y ¼ 276.43eÀ13.4x, where y and x are permeation and 1/T, respectively). Auras et al. [54] also studied the effect of temperature on the values of D and S for O2 in PLA and PET films. The value of D for O2 in PLA and PET films increased as temperature increased; an example of this behavior for PLA is shown in Figure 12.3. The values of D for O2 were higher in the PLA containing a higher L-lactide content (98% versus 94% Llactide). Both D and P of PLA were less affected by temperature than S of O2. An example of the effect of the temperature on the value of S for O2 of PLA is shown in Figure 12.4. that are caused by moisture sorption can be determined in multiple ways using spectroscopic technologies, including FTIR, near-infrared (NIR), and nuclear magnetic resonance (NMR) [55–58]. It is well known that the presence of water has a significant effect on the barrier properties of most hydrophilic polymers. According to Hernandez [59], the presence of water in a hydrophilic polymeric matrix may change the way in which a gas or vapor is sorbed and diffused through the polymer. In contrast, hydrophobic polymers typically have a water equilibrium concentration of less than 2 wt% in a saturated environment [54]. Table 12.2 shows the 12.2.7.2 Moisture Moisture sorption is a factor that can limit the use of polymers. Water absorbed from the atmosphere can swell, plasticize, or react with the polymer chains. Changes in the interactions between atoms within a polymer FIGURE 12.4 Oxygen solubility coefficient of PLA film containing 98% L-lactide as a function of water activity (reproduced from Ref. 54 with permission from Wiley-VCH Verlag GmbH & Co. KGaA). 5 4 3 2 0 0.0 0.2 0.4 0.6 Water activity 0.8 1.0 Solubility coefficient × 104(kg/(m3 Pa)) FIGURE 12.3 Oxygen diffusion coefficient of PLA film containing 98% L-lactide as a function of water activity (reproduced from Ref. 54 with permission from Wiley-VCH Verlag GmbH & Co. KGaA). 5.0 5ºC 23ºC 40ºC 4.0 3.0 2.0 1.0 0.0 0.0 0.2 0.4 0.6 Water activity 0.8 1.0
  • 162 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) TABLE 12.2 Effect of Relative Humidity on the Oxygen Permeability of Some Conventional Polymers at 23 C O2 Permeability at 23 C (10À17 kg m/(m2 s Pa)) Polymer Amorphous polyamidea Cellophaneb,c Poly(ethylene-co-vinyl alcohol)a High-density polyethyleneb Low-density polyethylenec Nylon 6, PA6b,c Polyamide (70% AN)a Poly(ethylene terephthalate)b Poly(vinyl acetate)b,c Poly(vinyl alcohol)b,c Dry (0% RH) 0.011 0.00076 0.00012 0.64 2.81 0.0059 0.0013 0.040 0.32 0.00006 Wet (100% RH) 0.008 1.2 0.012 0.64 7.81 0.33 0.004 0.035 0.88 0.15 a Adapted from Ref. 60. Adapted from Ref. 40. c Adapted from Ref. 61. b effect of RH on the O2 permeability of some conventional polymers at 23 C as reported by various sources. Under high RH conditions, O2 permeability increases for poly(vinyl alcohol) (PVOH), poly(ethylene-co-vinyl alcohol) (EVOH), poly(vinyl acetate) (PVA), cellophane, amorphous polyamide (aPA), nylon 6, PA6, and polyamide (70% AN), whereas it remains almost constant for high-density polyethylene (HDPE), low-density polyethylene (LDPE), and PET. For PLA, the absorption of water from the atmosphere has also been reported to affect O2 permeability. The effect of moisture on the values of P, S, and D for O2 of PLA films (94% and 98% L-lactide) has been characterized by Auras et al. [54]. In this study, PLA films were exposed to different water activities (aw ¼ 0–0.9) at different temperatures. PET films were studied for comparison. P of the PLA films decreased with increasing water activity, and this reduction was more pronounced at 40 C and aw ¼ 0 (from 11 Â 10À18 to 8.5 Â 10À18 kg m/(m2 s Pa)) than at 23 C. This behavior is in contrast to the increase in P of several conventional plastics under wet conditions mentioned above and shown in Table 12.2. The behavior of those polymers is due to their hydrophilic nature, whereas PLA and PET films are both relatively hydrophobic. The value of D for O2 in PLA and PET films showed an exponential increase as a function of aw for all temperatures (Figure 12.3), which was attributed to the plasticization effect of the water molecules on the amorphous phase. Moreover, the lower value of D for O2 of the PLA with a higher L-lactide content has been attributed to the higher crystallinity and, thus, more tortuous path for the O2 molecules. The value of S for O2 decreased linearly with increasing aw, as illustrated for one example in Figure 12.4, due to the reduction in free volume as it is occupied by water molecules. The sorption of water in a polymer can be explained by the difference between the solubility parameter (d) of the water (48 MPa0.5) [62] and that of the polymer; when this difference increases, the water sorption decreases. For PLA, d values of 19–20.5 MPa0.5 have been reported that are similar to the d values of 19 and 16 MPa0.5 for PS and PET, respectively [62]. Auras et al. [54] found that the water absorption of PLA films (94% and 98% L-lactide) was difficult to measure below 100 ppm, even after these films were exposed to different temperatures and to aw between 0.11 and 0.94 for more than one week. Therefore, these authors concluded that the PLA films did not absorb measurable amounts of water. In contrast, although PS and PET have d values similar to those of PLA, water absorption values of 320 and 60 ppm for PS and PET, respectively, have been measured at 25 C and aw of 0.5 [63]. Moisture sorption in PLA also depends on the molecular weight and the aliphatic content via end group modification [64]. PLAs with varying aliphatic content were created by ring-opening polymerization of L-lactide with aliphatic fatty alcohols of different chain lengths (C16 ¼ palmityl alcohol ¼ P; C10 ¼ decyl alcohol ¼ D; C4 ¼ butyl alcohol ¼ B). The molecular weight was controlled by varying the ratio of L-lactide monomer to alcohol initiator and by using triethylaluminum as a catalyst. PLAs with benzyl ends of different molecular weights were also synthesized to compare the sorption properties of aliphatic and aromatic end groups. Sample solutions were prepared by dissolving the synthesized PLA in chloroform that was then coated onto the quartz crystals. Each polymer film underwent three sorption/desorption cycles. Differences in moisture sorption of the samples were observed, which indicated that the molecular weight and the aliphatic content of the end group affected the sorption properties. Among the samples with the same aliphatic group but varying molecular weight (P4K (4,000 Da), P6K (6,000 Da), and P10K (10,000 Da)), the aliphatic content at the end group increased with decreasing molecular weight from about 2.3 (P10K) to about 5.7 (P4K) by mass. The water uptake for P6K was smaller than that for both higher and lower molecular weight samples. The larger water uptake for P4K was attributed to the lower molecular mass itself, while for P10K the decrease in aliphatic content of the end group (hydrophobic end group) increased the mass uptake of water. The investigation of samples with the same molecular weight but varying aliphatic content (B10K, P10K, and D10K) yielded additional insight into the effect of the end groups on the sorption properties. Along with an increase in the mass percentage of aliphatic content from 0.6% (B10K) to 1.4% (D10K), and then finally to 2.3% (P10K), there was a consistent decrease in the mass uptake
  • PERMEABILITY, SORPTION, AND DIFFUSION OF PURE PLA of water. Therefore, an increase in the hydrophobic content resulted in reduced sorption properties of the samples. A comparison of samples with the same molecular weight but varying aliphatic and aromatic content (P10K, B10K, D10K, and B10K) showed that the amount of mass uptake was higher in the samples with aromatic end contents (Benzyl10k). Vert et al. [65] showed that the use of different catalysts for the ring-opening polymerization of PLA can have a significant impact on the moisture sorption, and this was attributed to the impurities in the catalyst that contribute to the transesterification reactions and lead to more hydrophobic end groups. In contrast to all these results, Cairncross et al. [16] reported that changes in the end groups of PLA have no effect on sorption properties. The literature is inconsistent regarding the dominant mechanism of moisture transport. Sharp et al. [19] studied the water uptake of poly(DL-lactide) using a QCM and proposed that the equilibrium times of glassy polymers such as PLA lead to non-Fickian diffusion effects, based on the observation of thickness-dependent D. The kinetics of swelling was described using a model based on the Thomas–Windle model of penetrant diffusion in glassy polymers. According to this model, the kinetics of penetrant uptake differs from that of a simple Fickian process because the diffusion depends strongly on the solvent concentration. Rapid swelling of the polymer causes the buildup of osmotic stress in the material, and further swelling is prevented until this stress has been dissipated by viscous flow of the stressed polymer chain. As a result, the solvent concentration in the swollen region of the polymer rapidly reaches the equilibrium value, and a swelling front is produced that propagates through the polymer with uniform velocity. The polymer ahead of the front is glassy, but the material behind it is highly plasticized. Siparsky et al. [25] claim that it is not possible to separate the diffusion process from that of hydrolytic reactions and that therefore the measurement of transport properties requires analyzing simultaneous diffusion/reaction mechanisms. These authors further suggested the presence of clusters of water in the film that resulted in delivered water–polymer interaction that was trivial relative to cohesive water interaction. PLA can be depolymerized in the presence of water at elevated temperatures, which results in poor mechanical properties [66]. PLA can also undergo hydrolytic scission of the polyester linkages if it is exposed to moisture, and this leads to molecular fragmentation even in environments that are only slightly moist [67]. As a consequence of the continuous hydrolysis and the generation of more end groups, the polarity and the moisture sorption properties of PLA change with time, favoring a further increase in the moisture content. This behavior leads to nonequilibrium moisture sorption behavior; thus, it is unclear whether an equilibrium moisture sorption isotherm exists [53]. 163 12.3 PERMEABILITY, SORPTION, AND DIFFUSION OF PURE PLA 12.3.1 Gases 12.3.1.1 Carbon Dioxide Different studies on the values of P, S, and D for CO2 of pure PLA have been published in the literature. According to Petersen et al. [68], PLA has lower CO2 barrier properties than equivalent petroleum-based polymers. In a more detailed study, Auras et al. [15] reported that the values of P for CO2 in PLA films (ranging from 1.99 Â 10À17 to 4.18 Â 10À17 kg m/(m2 s Pa) at 0% RH and at temperatures from 25 to 45 C, respectively) are one order of magnitude lower than those of crystal PS (amorphous) (1.55 Â 10À16 kg m/(m2 s Pa) at 25 C and 0% RH) [69] and one order of magnitude higher than those of PET (1.73 Â 10À18 and 3.17 Â 10À18 kg m/(m2 s Pa) at 0% RH and at 25 and 45 C, respectively) [15]. The results were obtained using semicrystalline PLA films with L-lactide contents of 94% and 98% and densities around 1250 kg/ m3. Smaller differences between the permeation values of PLA and conventional polymers have been reported by Lehermeier et al. [12] and Bao et al. [70] These authors calculated a CO2 permeation value of 1.52 Â 10À13 kg m/ (m2 s Pa) at 30 C, which was reported as similar to that of PS, 1.56 Â 10À13 kg m/(m2 s Pa), and LDPE, 4.16 Â 10À13 kg m/ (m2 s Pa), and different from PET, 2.97 Â 10À15 kg m/(m2 s Pa) [11]. The differences between the results obtained by Auras et al. [15, 71] and the results from Lehermeier et al. [12] and Bao et al. [70] may be explained by differences in the crystallinity and processing conditions used to produce the films. Table 12.3 summarizes the values of P for CO2 of PLA and some conventional polymers at 23 and 30 C. TABLE 12.3 CO2 Permeability Coefficients of PLA and Some Conventional Polymers at 23 and 30 C CO2 Permeability (10À17 kg m/(m2 s Pa)) Polymer Poly(lactic acid) High-density polyethylene Low-density polyethylene Nylon 6, PA6 Poly(ethylene terephthalate) Polypropylene Polystyrene Poly(vinyl chloride) Poly(vinylidene chloride) a 23 Ca – 4.44–6.31 6.31–13.5 0.13–0.22 0.18–0.44 9.00–18.00 – 5.4–6.3 – Adapted from Ref. 72. All values adapted from Ref. 73 expect for PLA. c Adapted from Ref. 12. b 30 Cb 1.52c 5.25 52.82 0.03 0.02 13.80 13.20 1.50 0.04
  • 164 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) QCM studies of CO2 sorption in PLA (L:D ratio of 80:20 and Mw ¼ 102,800 Da) at different temperatures (20 and 40 C) and pressures up to 1 atm identified CO2 as the second most soluble solvent after water in the group water, O2, and ethylene (2.18–10.20 and 0.78–1.84 cm3 (STP)/cm3 for water and CO2, respectively) [74]. In this study, larger deviations for CO2 than for water were observed using the Flory–Huggins model to correlate the experimental results at different temperatures. The EP for CO2 permeation in PLA has been calculated as 15.65 Æ 0.63 and 19.44 Æ 0.90 kJ/mol for poly(lactic acid) films with 98% and 94% L-lactide, respectively [15]. Lehermeier et al. [12] reported a value of Ep ¼ 6.1 kJ/mol, which differs from that of 27.6 kJ/mol reported by Pauly [69] for PET. The variation of the values of P for CO2 as a function of factors such as L-lactide and temperature content has also been studied [15] (discussed in Sections 12.2.1 and 12.2.7). 12.3.1.2 Oxygen The values of P for O2 of PLA film have been shown to be comparable to those of equivalent petroleum-based polymers [68]. In a more detailed study by Lehermeier et al. [12], the values of P for O2 for PLA film was determined as 4.90 Â 10À14 kg m/(m2 s Pa) at 30 C. This value is close to that reported for PS (3.86 Â 10À14 kg m/ (m2 s Pa)) and LDPE (1.25 Â 10À13 kg m/(m2 s Pa)) films and is different from that of PET films (5.94 Â 10À16 kg m/(m2 s Pa)) at the same temperature [10, 11]. Auras et al. [15] found that the values of P for O2 for PLA films (98% and 94% Llactide) were about 20 times lower than values previously published for PS (2.7 Â 10À17 kg m/(m2 s Pa) at 25 C) [75] and about one order of magnitude higher than those published for PET (1.88 Â 10À19 kg m/(m2 s Pa) at 75% RH at 25 C) [15]. Cava et al. [76] have also reported higher values of P for O2 of biodegradable materials PCL, amorphous PLA (aPLA), and poly(hydroxybutyrate-co-valerate) (PHBV) (8 mol%) at 21 C and 40% RH than for PET. Table 12.4 compares the values of P for O2 of PLA and some conventional polymers at 23 and 30 C. The EP for O2 permeation in PLA has been determined by Auras et al. [15], who obtained values of 41.4 Æ 3.5 and 28.4 Æ 2.9 kJ/mol for PLA films with 98% and 94% L-lactide, respectively. However, Lehermeier et al. [12] reported that EP ¼ 11.1 kJ/mol and compared this value to that of 27.6 kJ/ mol for PET as reported by Pauly [69]. The higher EP was attributed to the aromatic ring in the PET polymer chain backbone. The O2 permeability of PLA sheets has been determined by Auras et al. [77], who compared the permeability values for thermoformed oriented PLA (OPLA) (94% L-lactide), oriented PS (OPS), and PET containers at 23 C and 0% RH. The OPLA permeability coefficient (4.33 Â 10À18 kg m/(m2 s Pa)) was one order of magnitude lower than that of OPS (3.91 Â 10À17 kg m/(m2 s Pa)) and one order of magnitude TABLE 12.4 O2 Permeability Coefficients of PLA and Some Conventional Polymers at 23 and 30 C O2 permeability (10À17 kg m/(m2 s Pa)) Polymer High barrier Cellophane Poly(vinyl alcohol) Nylon 6, PA6 Poly(ethylene terephthalate) Poly(vinyl chloride) Poly(vinylidene chloride) Low barrier Poly(lactic acid) High-density polyethylene Low-density polyethylene Polypropylene Polystyrene 23 Ca 30 Cb 0.0013c 0.0001c 0.0513d 0.0540d 0.0733d 0.0009d – – 0.5668 0.0333 0.1099 0.0081 – – 4.3183d 1.3490d 3.7780d 0.4948e 1.5924 8.2498 3.4540 1.6464 a Adapted from Ref. 40. (According to these authors, high-barrier polymers have an oxygen transmission of not more than about 10 cc per mil of polymer per 100 in.2 of surface per day per 760 mm of mercury driving force at 23 C and humidity conditions of use.) b Adapted from Ref. 73. c Determined at 0% RH. d Determined at 100% RH. e Adapted from Ref. 12. higher than that of PET (6.95 Â 10À19 kg m/(m2 s Pa)). The values of P for O2 of PLA, OPS, and PET sheets are similar to those reported for PLA, PET, and PS films in the literature. The authors also reported that the addition of regrind material had an effect on the O2 permeability of OPLA sheets, with a higher value of P for O2 for OPLA with 40% regrind content compared to OPLA without regrind. QCM studies of O2 sorption in PLA (L:D ratio of 80:20 and Mw ¼ 102,800 Da) at different temperatures (20 and 40 C) and pressures up to 1 atm showed that O2 is less soluble in PLA than water and CO2 (0.50–1.57 cm3 (STP)/cm3 for O2 versus 2.18–10.20 and 0.78–1.84 cm3 (STP)/cm3 for water and CO2, respectively) [74]. Permselectivity (b) is the ratio of CO2/O2 permeation of a polymeric material. Depending on this ratio, the CO2 and O2 levels reached in a polymeric package containing respiration produce evolve differently and thus also influence the product differently. In continuous film packages, the permeability of CO2 is typically 2–8 times greater than that of O2, so the CO2 concentration gradient for such packages is usually much smaller than the O2 concentration gradient [78]. According to Salame and Steingiser [40], the permselectivity value ranges between 4.1 and 6.1 for most conventional polymeric materials. For perforated film packages, the permeability of CO2 is more similar to that of O2 as indicated by a permselectivity value of only 0.77 [79], and thus the
  • PERMEABILITY, SORPTION, AND DIFFUSION OF PURE PLA concentration gradients for CO2 and O2 are nearly equal. CO2 levels rise to roughly the same extent that O2 levels decline, and as a result the sum of the O2 and CO2 partial pressures usually remains in the range of 18–20%. Thus, for any given O2 level, the perforated package will have a considerably higher level of CO2 (relative concentration) than the continuous film package. For bio-based polymers, the permselectivity value is higher (approx. 30) than for conventional petroleum-based plastics [80]. Permselectivity values reported for PLA show some variation due to the differences in crystallinity and processing conditions used to produce the films. Studies published by Lehermeier et al. [12] report a permselectivity value of 3.1 at 25 C for amorphous films of PLA cast from solution (5 wt% methylene chloride), which is well below the typical values for other bio-based polymers. Auras et al. [15] reported O2 permeability values at 70% RH that were more than 10 times lower than the values for CO2 at 0% RH, corresponding to a permselectivity value of 10. To reduce this high permselectivity and thereby increase the applications for the polymer, PLA has been microperforated [81]. Microperforated PLA exhibits a permselectivity value of around 1, as one would expect for microperforated materials. 12.3.1.3 Nitrogen Information about D, P, and S of nitrogen (N2) in PLA is scarce. According to Lehermeier et al. [12], the value of P for N2 in PLA films obtained by casting is 1.93 Â 10À14 kg m/(m2 s Pa) at 30 C, which is significantly larger than the coefficient in PET (1.19 Â 10À16 kg m/(m2 s Pa) at 30 C). This difference has been attributed to the presence of aromatic rings in the PET chain backbone, which reduce free volume and chain mobility. No effect of Llactide content (96:04 versus 98:02) or other factors, such as branching, on N2 permeability was observed. N2 permeation in PLA films has been reported to be very similar to that in PS films (3.27 Â 10À14 kg m/(m2 s Pa)) and in LDPE films (2.82 Â 10À14 kg m/(m2 s Pa)). Temperature affects N2 permeability in PLA, and 100% linear PLA with an L:D ratio of 96:4 has been shown to be temperature dependent (y ¼ 109.86eÀ1.36x, where y and x are permeation and 1/T, respectively) [12]. The EP for N2 permeation in PLA has been measured to be 11.2 kJ/mol [12], which is lower than the value of 26.4 kJ/mol for PET as determined by Pauly [69]. PET has a higher EP for N2 due to the presence of aromatic rings, as discussed previously. QCM studies of N2 sorption in PLA (L:D ratio of 80:20 and Mw ¼ 102,800 Da) at different temperatures (20 and 40 C) and pressures up to 1 atm revealed that N2 is less soluble in PLA than water (0.15–1.54 versus 2.18–10.20 cm3 (STP)/ cm3 polymer for N2 and water, respectively) [74]. The correlation of experimental results at different temperatures using the Flory–Huggins model showed larger deviations for O2 than for water and N2. 165 12.3.1.4 Methane Limited information about D, P, and S of CH4 in PLA is available in the literature. The value of P for CH4 of PLA cast from solution (5 wt% methylene chloride) is 1.49 Â 10À14 kg m/(m2 s Pa) at 30 C [12], which is lower than that reported for PET (5.94 Â 10À17 kg m/(m2 s Pa)) at 30 C [11] and similar to that for PS (3.42 Â 10À14 kg m/ (m2 s Pa)) and LDPE (5.94 Â 10À14 kg m/(m2 s Pa)). Lehermeier et al. [12] also reported a reduction of CH4 permeation by a factor of 4.5 times if PLA is biaxially oriented. Furthermore, CH4 permeability is temperature dependent, and for 100% linear PLA with an L:D ratio of 96:4, this dependence was exponential (y ¼ 149.95eÀ1.55x, where y and x are permeation and 1/T, respectively). The EP measured for the permeation of CH4 in PLA was 13.0 kJ/mol [12], which was smaller than that reported for PET (24.7 kJ/mol) by Pauly [69]. 12.3.1.5 Noble Gases (Helium and Argon) There is limited information in the literature about the permeation of the noble gases argon (Ar) and helium (He) in PLA. Phillip et al. [82] measured the transport of Ar, He, and other gases in ultrafiltration membranes made of PLA–PDMA (poly (dimethylacrylamide))–PS. The value of D for He (5.1 Â 10À4 m2/s) was significantly larger than that for N2 (1.9 Â 10À4 m2/s), which in turn was slightly larger than the value for Ar (1.7 Â 10À4 m2/s). For both noble gases, these P’s in PLA were smaller than those reported by the same authors for commercial polycarbonate (PC)-based membranes (4.0 Â 10À4 and 12.8 Â 10À4 m2/s for Ar and He, respectively). The transport mechanism of these gases through the pores in all membranes was well described by a Knudsen diffusion model, and these results were consistent with other values obtained for very small pores [82, 83]. 12.3.2 Water Vapor An initial study on the permeability of water in PLA was carried out by Siparsky et al. [25]. PLA films (prepared by adding 2-hexylhexanoate to molten lactide, followed by purification by precipitation using methylene chloride and methanol as the solvent and the nonsolvent) were heat sealed into bags, filled with molecular sieves, and then sealed. Gravimetric measurements were used to determine S and D. At 50 C and 90% RH, the reported S and D ranged from 400 Â 10À6 to 1000 Â 10À6 cm3 (STP)/(cm3 Pa) and 0.20 Â 10À6 to 1.0 Â 10À6 cm2/s, respectively. Auras et al. [15] measured the WVP of semicrystalline PLA films containing different percentages of L-lactide (94% and 98%) and found that in the range of 40–90% RH. The values for both polymers were practically constant (Figures 12.5 and 12.6). P determined for PLA films in this study (1.48–2.20 Â 10À14 kg m/(m2 s Pa) at 10–37.8 C and 40–90% RH) were higher than those determined previously [69] for PET (1.1 Æ 0.1 Â
  • 166 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) 2.50 P×1014 (kg m/(m2/s/pa)) 10ºC 2.25 2.00 20ºC 1.75 30ºC 1.50 1.25 35 37.8ºC 40 45 50 55 60 65 70 75 Relative humidity (%) 80 85 90 95 FIGURE 12.5 Water vapor permeability coefficient of PLA made from 98% L-lactide as a function of % RH. Error bars represent cumulative imprecision in measurements (reproduced from Ref. 15 with permission from SAGE Publications). 10À15 kg m/(m2 s Pa) at 40–90% RH) and PS (6.7 Â 10À15 kg m/(m2 s Pa) at 25 C). More information on P, S, and D of water vapor of PLA and some conventional petroleum-based polymers at 23 C is summarized in Table 12.5. Table 12.6 provides data for water vapor permeance and permeability at 25 and 38 C. PLA has lower water vapor barrier properties than equivalent petroleum-based polymers, and it can be included in the group of low-barrier polymers following the classification of Salame and Steingiser [40]. In this classification, a high-barrier polymer has a water vapor permeation of not more than 7 Â 10À3 per 2.54 Â 10À5 m of polymer per 6.45 Â 10À2 m2 of surface per day when is in direct contact with water at 38 C and with low RH air circulating on the downstream side of the barrier. Auras et al. [77] determined the water vapor permeability of PLA sheets and compared the water vapor transmission rates and P of thermoformed OPLA (94% L-lactide), OPS, and PET containers at 37.8 C and 100% RH. WVP of OPS and PET were one order of magnitude lower (4.18 Â 10À15 kg m/(m2 s Pa) and 2.82 Â 10À15 kg m/(m2 s Pa), respectively) than that of OPLA (1.34 Â 10À14 kg m/(m2 s Pa)). Similar WVP to the container values were reported for PLA, PET, and PS films [15], and WVP of PLA did not vary significantly with changes in RH. The authors also reported a decrease in WVP of PLA with increasing temperature, a behavior opposite to that of most petroleum-based polymers that exhibit lower water barrier properties at higher temperatures. The EP for water vapor permeation has been determined as À9.8 kJ/mol for PLA films with 98% L-lactide and À10.1 kJ/ mol for PLA films with 94% L-lactide [15]. EP values for PLA can differ depending on the processing conditions and crystallinity. For example, the EP values for amorphous and semicrystalline PLA films have been reported as 5 and À0.1 kJ/mol, respectively, at 100% RH [37]. Very few polymers show negative activation energies for water, and usually P×1014 (kg m/(m2s Pa)) 2.25 2.00 1.75 1.50 1.25 35 40 45 50 55 60 65 70 75 Relative humidity (%) 80 85 90 95 FIGURE 12.6 Water vapor permeability coefficient of PLA 96% L-lactide as a function of % RH (reproduced from Ref. 15 with permission from SAGE Publications).
  • PERMEABILITY, SORPTION, AND DIFFUSION OF PURE PLA TABLE 12.5 Water Vapor Permeability, Solubility, and Diffusion Coefficients of PLA and Petroleum-Based Polymers at 23 C Polymer Poly(lactic acid)a Poly(lactic acid)b Poly(ethylmethacrylate) (PEMA)c Cellulose nitratec Poly(ethyl cellulose), (PEC)c Poly(ethylene terephthalate)c Polystyrenec High-density polyethylened Polypropylened Low-density poly(ethylene)d Ionomer, Surlynd P (Â 10À13) (kg m/ S (Â 10À6) D (Â 10À7) 2 (m s Pa)) (kg/(m2 s Pa)) (m2/s) 80–360 0.198 238 60–110 0.074–0.510 223 0.102 472 670 1,800 2,340 0.026 0.029 110 670 225 225 670 120 a Adapted from Ref. 25. Adapted from Ref. 15. c Adapted from Ref. 84. d Adapted from Ref. 40. b their magnitudes are smaller. Cellulose acetate and cellulose acetate propionate also have negative EP values, À8 and À6 kJ/mol, respectively, for water vapor. Moisture sorption in PLA depends on the molecular weight and the aliphatic content via end group modification [64]. The primary petroleum-based polymers are nonpolar, where- TABLE 12.6 Water Vapor Permeance and Permeability of PLA and Some Conventional Polymers at 90% RH and 38 and 25 C Polymer Poly(lactic acid) High-density polyethylene Low-density polyethylene Nylon 6, PA6 Poly(ethylene terephthalate) Polypropylene Polystyrene Poly(vinyl chloride) Poly(vinylidene chloride) a Adapted from Ref. 72. Adapted from Ref. 73. c Adapted from Ref. 25 at 20 C. b Permeance (10À13 kg m/ (m2 s Pa)) at 38 Ca Permeability (10À13 kg m/ (m2 s Pa)) at 25 Cb – 8.50 3.10 23.91 3.98 – 15.94 4.87 0.50–4.42 140–220c 0.98 1.50 52.52 9.75 0.60 9.00 11.70 0.11 167 as PLA is fairly polar, which leads to a reduction in its water resistance. Water has been reported to seep out of water bottles made from PLA [15]. The plasticization of polymers by sorption of ambient vapors or liquids has been shown to result in a decrease of mechanical properties as well as a loss of beverage products [61]. Sorption isotherms obtained (using a QCM) for different gases and vapors in PLA (with an L:D ratio of 80:20 and Mw ¼ 102,800 Da) at different temperatures (20–40 C) and pressures up to 1 atm revealed water as the most soluble solvent in the group water, O2, and ethylene (2.18–10.20 versus 0.78–1.84 cm3 (STP)/cm3 polymer for water and CO2, respectively) [74]. Smaller deviations for water were observed than for O2 or CO2 when the Flory–Huggins interaction parameters were correlated for each temperature. Yoon et al. [33] compared the values of S and D for water in PLLA with those in various polymers such as polyglycolide (PGA), PHB, PCL, and glycol modified copolyester (SG (SkygreenR)). S and D were measured by monitoring the extension of a quartz spring that was due to weight variation caused by water sorption. D of PLLA was higher than that of PHB and PGA, but lower than that of SG and PCL, and the equilibrium solubility of PLLA was higher than that of PHB and PLC, but lower than that of PGA and SG. Shogren [37] evaluated and compared the WVTR of several biodegradable polyester films at 25 C and showed that the normalized WVTR values of the as-cast and annealed PLLA films (9.49 Â 10À7 and 1.99 Â 10À6 kg/(m2 s), respectively) were higher than those of PHBV with 6, 12, and 18 wt% valerate units (1.50–1.85 Â 10À7 kg/(m2 s)), comparable with that of PCL (2.05 Â 10À6 kg/(m2 s)), and lower than those of aliphatic polyesters (Bionolle) (3.82 Â 10À6 kg/(m2 s)). Although the WVTR values of biodegradable polymers were much higher than those of good barrier materials such as LDPE, they were considered sufficient for short-term (hours to days) protection. This could be extended to weeks, as in PLA water bottles. The values of WVTR were positively correlated to higher polymer solubility parameters, lower crystallinities, and higher free volumes (Table 12.7). These results indicated that the kind of monomer unit is a crucial parameter for the WVTR of biodegradable polyester films. 12.3.3 Organics There is little information in the literature about the diffusion, solubility, and permeability of organic compounds in PLA. Auras et al. [85] measured the values of D, S, and P for ethyl acetate at 30 and 45 C in PLA films containing 98% L-lactide by using gravimetric sorption tests. The difference between the solubility parameters of PLA and ethyl acetate was Dd ¼ 7.14 (J/cm3)1/2, which is >5, and, therefore, the interaction between PLA and ethyl acetate is expected to be small. P, D, and S were reported as 1.22 Â 10À17 kg m/(m2 s Pa), 2.63 Â 10À15 m2/s, and 4.62 Â 10À3 kg/(m3 Pa), respectively,
  • 168 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) TABLE 12.7 Water Vapor Transmission Rates of Some Biodegradable Polymeric Materials WVTR (10À8 kg/(m2 s)) 6 C Polymer Poly(lactic acid) (crystal) Poly(lactic acid) Poly(butylene-succinate) Bionolle Poly(e-caprolactone) Poly(3-hydroxybutyrate-co-3-valerate) (PHBV)-6 Poly(3-hydroxybutyrate-co-3-valerate) (PHBV)-12 Poly(3-hydroxybutyrate-co-3-valerate) (PHBV)-18 25 C 49 C (J/cm3)1/2 31.25 62.50 68.28 47.45 2.08 3.59 4.05 95.07 199.07 381.9 204.86 15.05 24.31 30.01 385.42 1273.1 2800.9 1354.2 143.52 236.11 283.56 22.7 22.7 – 20.8 21.5 21.5 21.4 Crystallinity (%) 66 0 41 67 74 69 62 Adapted from Ref. 37. at 45 C and a partial pressure of 12,654 Pa. The solubility and permeability of PLA to ethyl acetate determined in this study were compared with solubility and permeability values obtained from the literature for PET [86], polypropylene (PP) [87], and LDPE [88]. The solubility of ethyl acetate was slightly higher in PLA than in PET, PP, or LDPE, whereas the permeability of ethyl acetate in PLA was similar to that in PET and lower than that in PP and LDPE. Figure 12.7 compares the various values of P for ethyl acetate. No PS data are available for comparison since it has been reported that PS dissolves in ethyl acetate [89]. PLA appears to be a good barrier to ethyl acetate. Lee et al. [90] determined the values of D, S, and P for ethyl acetate at 23 C and 55% RH in PLA sheet. These authors also reported mass transfer values for other major aroma compounds in food, such as acetaldehyde and trans-2hexenal, and compared those with the values in PET sheets under the same conditions. As shown in Table 12.8, the values of P, D, and S of the three organic compounds for PLA were higher than of those for PET by a few orders of magnitude. The authors also found the following trend in both PLA and PET for the transmission rate of the organic compounds: acetaldehyde < ethyl acetate < 2E-hexenal. The values of D, S, and P for ethyl acetate in the PLA sheet differed by several orders of magnitude from those reported by Auras et al. [85] for PLA film. These differences may be explained by differences in the processing, crystallinity, and L-lactide content. D-Limonene is one of the few aroma compounds that have been investigated in this context. PLA appears to be a good barrier to D-limonene, as shown in one study where no trace of D-limonene was detected after 21 days of testing at 45 C and 258 Pa, resulting in a P less than 9.96 Â 10À21 kg m/ (m2 s Pa) [91]. Regular sorption theory (RST) calculations predicted sorption values of D-limonene in PLA to be lower than those in PET, PP, LDPE, and PS [91]. Haugaard et al. [92] found no detectable D-limonene scalping by PLA 18 16 ln(P×1021) (kg m/(m2s Pa)) a Solubility Parameter 14 12 10 8 6 4 2 0 PLA PET PP LDPE  FIGURE 12.7 Ethyl acetate permeability coefficients for PLA at 30 C, p ¼ 7560 Pa; PET at 30 C, p ¼ 9435 Pa; LDPE at 22 C, p ¼ 4348 Pa; and PP at 22 C, p ¼ 6088 Pa (reproduced from Ref. 85 with permission from Wiley-VCH Verlag GmbH & Co. KGaA).
  • PLA BLENDS TABLE 12.8 Sheets 169 Diffusion, Solubility, and Permeability Coefficients of Ethyl Acetate, Acetaldehyde, and 2E-Hexenal in PLA and PET Permeability (10À14 kg m/(m2 s Pa)) Solubility (kg/ (m3 Pa)) Diffusion (10À14 m2/s) Transmission Rate (10À6 kg/(m2 s)) Sample PET PLA PET PLA PET PLA PET Ethyl acetate Acetaldehyde 2E-Hexenal a PLA 53.4 29.0 30.1 0.25 0.21 – 41.3 19.5 55.89 1.08 0.37 – 1.33 1.50 0.42 0.24 0.48 – 21.2 111.0 0.9 0.11 0.88 – Adapted from Ref. 90. and PS, but a high degree of scalping by HDPE, and concluded that PLA provides better protection against limonene scalping than does HDPE. Whiteman et al. [93] reported PLA as having a lower value of P for D-limonene than PET, LDPE, or PP. D for D-limonene has also been studied in other biodegradable polymers, such as PHBV (8 mol% of valerate units) and PCL, and has been compared to that in PET using transmission FTIR experiments [76]. A higher D was obtained for D-limonene in PET and PCL than in PHBV (4.0 Â 10À13, 5.5 Â 10À13, and 4.0 Â 10À14 m2/s, respectively). 12.4 COPOLYMERS Copolymers of lactide and other monomers have been prepared to produce PLA with more desirable properties. However, literature about the effects of the added monomers on the barrier properties of the lactide copolymer is scarce. Sharp et al. [19] compared the water uptake of poly(DLlactide) and copolymer poly(lactide-co-glycolide) (50/50) by means of a QCM and saturated salt solutions to control the RH and found that the solubility of water in the homopolymer was higher than in the copolymer. In addition, slightly increased water absorption was observed in the polymer with an increase in average weight, which could be explained by the lower ratio of hydrophilic to hydrophobic end groups. These researchers determined that the increase in vapor activity caused an increase in the initial rate of sorption and an increase in the time required to reach equilibrium. This increase in moisture sorption due to plasticization can cause a decrease in glass transition temperature of the polymer film. This latter effect should increase the rate of diffusion through the film, thus reducing the increase in sorption rate over time. 12.5 packaging applications. Blending is the easiest and, in general, the most cost-effective way to prepare polymeric materials with desirable properties [94]. PLA can be blended with other polymers obtained from renewable resources, such as chitosan, starch, and PHB, or with petroleum-based polymers, such as PVA, PCL, and PEG in order to improve its properties. Blending is an effective approach to reduce the cost of PLA without sacrificing its excellent biodegradability and while maintaining certain mechanical and thermal properties. In other cases, PLA can be used as an ‘‘additive’’ to improve physical, mechanical, or barrier properties of other polymers from renewable sources. PLA BLENDS The blending of two or more polymers is performed with the objective of achieving certain physical, rheological, or chemical properties as required for some processing or 12.5.1 PLA/Starch Blends Starch is a carbohydrate biopolymer and consists of polysaccharides, amylose, and amylopectin [26]. Since both PLA and starch are biodegradable and derived from renewable sources, PLA/starch blends have excellent biodegradability. Blending starch with PLA is one of the most promising approaches because starch is an abundant and cheap biopolymer and PLA has good mechanical properties [95]. However, hydrophobic PLA, with hydroxyl and carboxyl end groups, and hydrophilic starch, with plenty of hydroxyl groups, are thermodynamically immiscible (poor interfacial adhesion between phases). Starch remains in a separate conglomerate form in a PLA matrix. The size of conglomerates is determined by the volume fraction, the starch molecular weight, the viscosity ratio, the interfacial energy, and the shear history during processing [96]. The use of compatibilizers and other additives has been studied with the goal of improving the interfacial interactions of these blends. The moisture sensitivity of starch can be reduced by blending the polymer with hydrophobic PLA. Ke and Sun [97] characterized blends of starch and PLA in the presence of various water contents and found that the initial moisture content of the starch had a significant effect on the water absorption of the blends. The water absorption increased slowly with increasing starch content up to 60%, but
  • 170 PERMEATION, SORPTION, AND DIFFUSION IN POLY(LACTIC ACID) increased rapidly for starch contents exceeding 70%. Also, Ke and Sun [98] reported that PLA/starch blends containing gelatinized starch (30% moisture content) had greater water absorption than the other blends because the gelatinized starch is more sensitive to water than the granulate state. The authors added PVOH to a 50%/50% (w/w) starch/PLA blend to enhance the compatibility and mechanical properties of the blend, and they tested PVOH of various molecular weights ranging from 6,000 to 125,000 Da. Increasing the molecular weight of PVOH slightly affected water absorption of the blends, but increasing the PVOH concentration between 40% and 50% increased water absorption. Zhang and Sun [99] blended dioctyl maleate (DOM) with PLA/ starch blends to improve their miscibility; DOM acted as a compatibilizer at low concentrations (below 5%) and served as a plasticizer at concentrations above 5%. The water absorption of these blends increased with DOM concentration, and, as a result, DOM leaching in an aqueous environment was inhibited. Cyclodextrins (CDs), a family of cyclic oligosaccharides composed of b-(1,4)-linked glucopyranose subunits, are produced by enzymatic degradation of starch by bacteria. CDs have a cage-like supramolecular structure that allows them to interact with molecules, ions, or radicals [100], and this structure creates the capacity to carry compounds. Almenar et al. [101] used CDs to carry natural antimicrobial compounds for the formation of bio-based antimicrobial sheets. The inclusion complexes (ICs), CDs carrying the natural volatile 2E-hexenal, were melt extruded with PLLA containing 96% L-lactide, injected into a disk-shaped mold, and pressed into films using a hydraulic press. Different mass fractions of ICs (0%, 15%, and 30%) were tested, and control PLA was blended only with CDs (i.e., without volatile). The bio-based antimicrobial sheets showed a higher permeability to water vapor than neat PLA films (Table 12.9). The addition of the ICs to the PLA did not affect the permeability to CO2, but an IC content of 30% (w/ w) did alter the O2 permeability. Of the three permeants, O2 showed the strongest increase in permeability with increas- TABLE 12.9 and O2 ing IC content. However, the same behavior was observed for PLA/CDs blends (without volatile), so the presence of the volatile did not have a significant effect on the barrier properties of PLA. 12.5.2 PLA/Poly(e-caprolactone) Blends PCL is a semicrystalline polymer with low tensile strength, large elongation at break (above 400%), and processing temperatures similar to aPLA. PCL has the potential to act as a plasticizing agent in blends with aPLA, but the poor barrier properties of PCL may cause an increase in the gas permeability of aPLA, which would be a drawback of these blends [102]. Cabedo et al. [102] prepared a melt blend of aPLA (12% L-lactide and Mw ¼ 190,000 Da) with PCL (Mw ¼ 100,000 Da) that was ground and then compression molded into sheets. The O2 permeabilities of the neat polymeric sheets were 1.1 Â 10À18 and 5.8 Â 10À18 m3 m/ (m2 s Pa) for PLA and PCL, respectively. As expected, the addition of PCL resulted in a significant decrease in the PLA barrier properties, and this decrease was proportional to the amount of PCL added. For the blends containing 36% PCL, the O2 permeability increased by 136% compared to the pure aPLA. To enhance the barrier properties of the PLA/PCL blends, kaolinite (Al2Si2O5(OH)4) (4% w/w), a modified clay, was added and this resulted in enhanced gas barrier properties due to the increased tortuosity level across the sheet. The O2 permeability values of the pure polymers containing kaolinite were 8.94 Â 10À19 and 6.84 Â 10À18 kg m/(m2 s Pa) for PLA and PCL, respectively, representing increases in the O2 barrier of 43% and 17%, respectively. The O2 barrier of the blend with kaolinite was enhanced by 27% (2.72 Â 10À18 kg m/(m2 s Pa)). Plackett et al. [35] also blended PLA with PCL and reported a significant absorption of moisture for the PLA–PCL copolymer films. Nanoclays were introduced into these films in an attempt to improve the barrier properties of the material and to reduce moisture loss from the packaged product, but moisture absorption of the films also increased. Effect of the Addition of CDs and ICs (Containing 2E-Hexenal) on PLA Permeability to Water Vapor, CO2, Permeability (10À17 kg m/(m2 s Pa)) Sample CO2 O2 PLA PLA PLA PLA PLA a Water Vapor 2954 Æ 120a 3875 Æ 560bc 4214 Æ 263c 3597 Æ 349b 3879 Æ 630bc 32 Æ 7a 203 Æ 57a 204 Æ 74a 112 Æ 17a 199 Æ 83a 6 Æ 0a 270 Æ 83a 1615 Æ 77b 247 Æ 49a 1808 Æ 676b 15% CD 30% CD 15% CD2EH 30% CD2EH Within columns, different lowercase letters denote significant differences (p < 0.05) among PLA samples due to the content of CDs and the antifungal volatile 2E-hexenal (CD2EH).
  • PLA COMPOSITES AND FIBERS PCL, which is more hydrophobic than PLA, was blended with PLA by Siparsky et al. [25] to achieve and enhance hydrophobicity. PLA homopolymer and blends were exposed to various temperatures and relative humidities, and WVP of the blends were found to be higher than those of neat PLA. 12.5.3 PLA Chitosan Blends Chitosan is a water-insoluble, nontoxic, edible, biodegradable polymer (polysaccharide) that is obtained commercially from chitin by alkaline deacetylation [103]. Chitosan is the second most abundant biopolymer in nature after cellulose. Since chitosan is a polycationic polymer, its high sensitivity to moisture limits its applications. One way to overcome this drawback is to blend the material with humidity resistant polymers such has PLA. Suyatma et al. [104] combined hydrophilic chitosan with hydrophobic PLA (92% L-lactide and 8% mesolactide, Mw ¼ 49,000 Da) by solution and film mixing, resulting in improved water barrier properties and decreased water sensitivity of the chitosan films. However, testing of mechanical and thermal properties revealed that chitosan and PLA blends are incompatible. Sbastien et al. [23] also investigated and characterized e composite chitosan/PLA films, prepared by solution mixing and a film casting procedure, for potential food applications. PLA was mixed with chitosan to reduce the hygroscopic properties of the chitosan by reducing the availability of its hydroxyl groups that are responsible for its weak moisture barrier. Contact angle measurements showed two types of film–liquid water behavior from the composite films. A higher contact angle was observed for films containing higher PLA content; therefore, the introduction of the PLA into a chitosan matrix decreased the absorbent character of the chitosan films. However, moisture barrier properties were not improved in the PLA/chitosan films. Heterogeneity could be the result of a phase separation of the blend, resulting in films with low structural cohesion that would facilitate the diffusion of water vapor molecules. The addition of 16.6% PEG to the films allowed an easy film recovery without a significant decrease in the moisture barrier properties of the materials. However, difficulties were encountered in the production of a miscible PLA and chitosan film forming solution, leading to heterogeneous films with high water sensitivity and, therefore, limited usability as a packaging material. 12.5.4 12.6 171 PLA LAMINATIONS Polyester/plasticized wheat starch (PWS)/polyester laminate films were prepared by Martin et al. [107] using flat film coextrusion and compression molding and then water immersion. Various biodegradable aliphatic polyesters such as PLA, PCL, poly(ester amide) (PEA or BAK400), poly(butylene-succinate) (PBSA or Bionolle 3001), and PHBV were tested. The resulting multilayers showed good moisture resistance; no PWS swell or delamination was observed. The order in decreasing moisture resistance for these polymers was PLA > PCL > PEA, whereas the order was reversed when the polymers were ranked according to adhesion properties. Lamination with biodegradable aliphatic polyesters may be a possible solution to overcome the rapid water absorption problem of the starch, and such multilayer films may be suitable for applications in food packaging [107]. In this study, PWS blends with the aforementioned polymers were also prepared. For all blends, it was observed that the hydrophobic character decreased rapidly with increasing polyester content (from 0% to 10%) and continued to decrease until the value of pure polyester was reached. PWS/ PLA blends showed the highest hydrophobicity. However, the results indicated that starch laminated with moisture resistant polyesters had better barrier properties than the PWS blends. 12.7 COATED PLA PLA provides a low barrier to gases and water for applications such as carbonated beverages, water, and extended shelf life products. Another possible way to improve PLA barrier properties is by coating with thin organic or inorganic layers, such as silicon oxide (SiOx), aluminum oxide (Al2O3), and diamond-like carbon (DLC). SiOx is the most common commercial technology to increase barrier properties of polymers. PLA bottles have been shown to be compatible with c