Modern Styrenic Polymers: Polystyrenes and Styrenic Copolymers

Modern Styrenic Polymers: Polystyrenes and Styrenic Copolymers

Wiley Series in Polymer Science Series Editor: Dr John Scheirs Excel Plas PO Box 2080 Edithvale VIC 3196 AUSTRALIA scheirs.john @ pacific.net.au Modern Fluoropolymers High Performance Polymers for Diverse Applications Polymer Recycling Science, Technology and Applications Metallocene-based Polyolefins Preparations, Properties and Technology Polymer-Clay Nanocomposites Dendrimers and Other Dendritic Polymers Forthcoming titles: Modern Polyesters Environmentally Degradable Polymers

Modern Styrenic Polymers: Polystyrenes and Styrenic Copolymers

Edited by

JOHN SCHEIRS ExcelPlas Australia, Edithvale, VIC, Australia and

DUANE B. PRIDDY 6004 Camelot Ct, Midland, Ml, USA

WILEY SERIES IN POLYMER SCIENCE

John Wiley & Sons, Ltd

Copyright © 2003 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England Telephone (+44) 1243 779777 Email (for orders and customer service enquiries): [email protected] Visit our Home Page on www.wileyeurope.com or www.wiley.com All Rights Reserved. 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 under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London WIT 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to [email protected], or faxed to (+44) 1243 770620. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Other Wiley Editorial Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103–1741, USA Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 22 Worcester Road, Etobicoke, Ontario, Canada M9W 1L1

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 0 471 49752 5 Typeset in 10/12pt Times by Kolam Information Services Pvt Ltd, Pondicherry, India. Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire. This book is printed on acid-free paper responsibly manufactured from sustainable forestry in which at least two trees are planted for each one used for paper production.

Contents Contributors Series Preface Preface About the Editors I

xxi xxvii xxix xxxiii

INTRODUCTION TO STYRENIC POLYMERS 1

Historical Overview of Styrenic Polymers John Scheirs

1 2 3 4 5 6 7 8 9 10 11

Introduction General-purpose Polystyrene Foamed Polystyrene Rubber-modified Polystyrene ABS ASA Early Styrene Copolymers Styrenic Block Copolymers Syndiotactic Polystyrene Modern Polystyrene Production The Future References

2 Polystyrene and Styrene Copolymers - An Overview Norbert Niessner and Hermann Gausepohl

1 2 3 4

Introduction Polymerization Processes Structure and Morphologies

3

3 4 13 18 18 20 21 21 22 22 22 23 25

25 27 29 29

vi

CONTENTS

5 Properties 6 Properties, Range and Applications of MABS Products References H

34 38 41

PREPARATION OF STYRENIC POLYMERS

43

3 Commercial Processes for the Manufacture of Polystyrene Bernard J. Meister and Clark J. Cummings

45

1 Introduction 2 Technical Constraints that Influence Reactor Selection 2.1 Temperature Control 2.2 Chemistry-related Constraints 2.3 Constraints Due to Reactor Mixing 2.4 Constraints Related to the Rubber Modification of Polystyrene 2.5 Reactor Requirements for Producing Copolymers 3 Polystyrene Devolatilization 3.1 Devolatilization Concepts 3.2 Devolatilization Equipment 3.3 Steam Stripping 4 Current Polystyrene Polymerization Processes 5 Process Simulation and Control References 4 Approaches to Low Residual Polystyrene Duane B. Priddy 1 Introduction 2 Summary of R&D Approaches 2.1 Devolatilizer Design 2.2 Assisted Devolatilization 2.3 Scavengers 2.4 Absorbers 2.5 High Monomer Conversion Polymerization 2.6 Solid Polymer Treatment 3 Friedel-Crafts Catalyst 3.1 Addition of Friedel-Crafts Catalyst to Monomer 4 Latent Acid Catalysts 4.1 Tosylates 5 Monomer Regeneration Upon Heating 6 Epilog References

45 46 46 48 52 54 57 59 59 60 65 66 69 71 73 73 75 75 76 78 80 80 82 83 85 85 86 88 91 91

CONTENTS

5 Process Modelling and Optimization of Styrene Polymerization J. Gao, K. D. Hungenberg and A. Penlidis 1 Introduction 1.1 General Kinetic Scheme of Styrene Homopolymerization 1.2 Treatment of Gel Effect 2 Process Simulation and Optimization of Styrene Homopolymerization 2.1 Using Initiator Combinations with Designed Temperature Profile 2.2 Using Bifunctional Initiators 2.3 Using Reactor Combinations 3 Conclusion 4 Symbols References 6 Living Free Radical Polymerization of Styrene Alessandro Butte, Giuseppe Storti and Massimo Morbidelli 1 Introduction 2 LRP Overview 2.1 Nitroxide-mediated Polymerization (NMP) 2.2 LRP by ATRP 2.3 LRP by Degenerative Transfer 2.4 LRP by RAFT 3 Kinetics of LRP 3.1 Main Features of the Different LRP Processes 3.2 Homogeneous vs Heterogeneous LRP Processes 4 Applications to Styrenic Polymers References

vii

93 93 94 98 100 100 101 105 107 107 108 111 111 112 115 116 118 118 120 120 122 125 127

7 Increasing Production Rates of High MW Polystyrene Bryan Matthews and Duane B. Priddy

129

\ Introduction 2 Speeding Up the Rate of Polystyrene Production Using Chemical Initiators 3 Speeding Up the Rate of Polystyrene Production Using Acid Mediation 4 Use of Acid to Tailor the Molecular Weight Distribution 5 Modeling Acid-mediated Styrene Polymerization

129 130 133 139 140

viii

CONTENTS

5.1 Styrene Auto-initiation Model 5.2 Acid Model Development 5.3 Model Results 6 Conclusions References

141 141 143 145 146

8 Preparation of Styrene Block Copolymers Using Nitroxide Mediated Polymerization Duane B. Priddy

147

1 Introduction 2 Mechanism and Limitations 3 How Living is NMRP? The Results of Model Studies 4 Block Copolymers via the Macroinitiator Approach 5 Preparation of Block Copolymers Using Alkoxyamines as Chain-stoppers in Step-growth Polymerization 6 Preparation of Block Copolymers via Sequential Addition of Monomers (SAM) 7 Preparation of Block Copolymers Using Multiple-headed Initiators References III MAJOR CLASSES OF STYRENIC POLYMERS

147 148 149 152 155 156 159 162 163

9 Particle Foam Based on Expandable Polystyrene (EPS) Rolf-Dieter Klodt and Brad Gougeon

165

1 Introduction 2 EPS Based on Suspension Polymerization 2.1 Production of EPS Raw Material 2.2 From Raw Material to Foam 2.3 Physical and Mechanical Properties 2.4 Applications 3 EPS Based on Extrusion Process 3.1 Extrusion 3.2 Post Extrusion 3.3 Steam Expansion of EPS Loose-fill Resin: Theory and Practice References

165 166 166 182 185 188 190 191 192 194 197

CONTENTS

ix

10 Rigid Polystyrene Foams and Alternative Blowing Agents Kyung Won Suh and Andrew N. Paquet

203

1 Introduction and General Description 2 Nomenclature 3 Theory of the Expansion Process 3.1 Bubble Initiation 3.2 Bubble Growth 3.3 Bubble Stabilization 4 Properties and Their Relation to Structure 4.1 Test Methods 4.2 Properties of Commercial Products 4.3 Cells 4.4 Gas Composition 4.5 Rigid Cellular Polymers 4.6 Creep 4.7 Structural Foams 5 Thermal Properties 5.1 Thermal Conductivity 5.2 Coefficient of Linear Thermal Expansion 5.3 Maximum Service Temperature 5.4 Moisture Resistance 5.5 Environmental Aging 5.6 Other Properties 6 Commercial Production and Processing 6.1 Manufacturing Process 6.2 Decompression Expansion Processes, Physical Stabilization 7 Applications 7.1 Thermal Insulation 7.2 Refrigeration 7.3 Construction 7.4 Structural Components 7.5 Marine Applications 7.6 Other Uses 7.7 Energy Considerations in Foam Insulation 8 Environmental, Health and Safety Considerations 8.1 Flammability 8.2 Blowing Agents and Environmental Issues References

203 204 205 205 206 207 207 207 207 209 210 210 211 212 213 213 216 216 216 217 217 218 218 218 221 221 223 223 223 224 224 224 225 225 226 228

CONTENTS

11 Polystyrene Packaging Applications: Foam Sheet and Oriented Sheet Gary C. Welsh 1 Introduction 2 Oriented Polystyrene Sheet 3 Extruded Polystyrene Foam Sheet References 12 Preparation, Properties and Applications of High-impact Polystyrene M. F. Martin, J. P. Viola and J. R. Wuensch 1 Introduction 2 Properties 2.1 General Properties 2.2 Mechanical Properties 2.3 Impact Properties 2.4 Thermal Properties 2.5 Electrical Properties 2.6 Rheological Properties 2.7 Resistance to Solvents 3 Basic Chemistry 3.1 Matrix Molecular Weight 3.2 Elastomer Considerations 3.3 Environmental Stress Crack Resistance (ESCR) 3.4 Thermal and Oxidative Stability 4 Manufacture 4.1 Process Evolution 4.2 Modern Commercial Process 5 Fabrication 5.1 Fabrication Process and Part Properties 6 Application 7 Acknowledgements References 13 Key Structural Features Impacting SAN Copolymer Performance R. P. Dion and R. L. Sammler 1 Introduction 2 Characterization 2.1 Chromophores 2.2 Sequence Distributions

233 233 233 239 245 247 247 248 248 248 250 252 252 253 253 256 256 256 261 264 266 266 268 271 273 275 279 279 281 281 283 283 284

CONTENTS

3

4 5 6

xi

2.3 AN Levels 2.4 MWD 2.5 Composition Distribution 2.6 Multidimensional Analysis Fabrication Performance 3.1 Shear Flow 3.2 Entangled Chains 3.3 Time-Temperature Superposition 3.4 Cross Model 3.5 Nonlinear Shear Flows 3.6 Relaxation Spectra 3.7 Extensional Flow 3.8 Break Points 3.9 Brittle Breaks 3.10 Flow Birefringence Multiphase Systems Solid-phase Behavior Conclusion References

14 Rubber Particle Formation in Mass ABS Gilbert Bouquet 1

2 3 4 5

6

7 8 9

Manufacture of ABS 1.1 Emulsion Process 1.2 Mass Process Phase Separation Phase Inversion Phase Diagram Rubber Particle Sizing 5.1 Shear 5.2 Viscosity 5.3 Interfacial Tension Grafting 6.1 Graft Analysis 6.2 Effect of Process Parameters 6.3 Master Curve 6.4 Graft Model Crosslinking Sizing Window Rubber Particle Morphology References

285 285 285 286 287 287 287 288 289 289 290 291 293 293 293 294 296 297 298 305 305 305 306 306 307 307 308 308 308 310 311 311 311 313 313 314 316 317 318

xii

CONTENTS

15 High Heat Resistant ABS Technology Rony Vanspeybroeck, Robert P. Dion and Joseph M. Ceraso 1 2 3 4 5 6

Introduction Substituted Styrenes Imides Maleic Anhydride Modified Nitriles Various High Heat-resistant ABS Grades References

321 321 324 326 330 333 334 338

16 Synthesis, Properties and Applications of Acrylonitrile-Styrene-Acrylate Polymers 341 G. E. McKee, A. Kistenmacher, H. Goerrissen and M. Breulmann 1 Introduction 2 ASA Market 3 Production of ASA 3.1 Early Developments 3.2 Emulsion Polymerization Process 3.3 Bulk Polymerization Process 3.4 Microsuspension Polymerization Process 4 Properties of ASA 4.1 Ageing Properties 4.2 Impact Behaviour 5 Additional Areas of Investigation 6 ASA Blends 7 Applications of ASA 7.1 General 7.2 Solar Technology 7.3 Safety in the House and in the Office 7.4 ASA for Automotive Body Panels with PFM Technology 8 Future Perspectives References IV SYNDIOTACTIC POLYSTYRENE 17 Synthesis of Syndiotactic Polystyrene Norio Tomotsu, Michael Malanga and Juergen Schellenberg 1 Introduction 2 Catalytic Systems for SPS 2.1 Transition Metal Complexes

341 342 343 343 343 345 347 348 348 351 352 352 355 355 356 357 357 359 359 363 365 365 366 366

CONTENTS

xiii

2.2 Co-catalysts 3 Copolymerization 3.1 Polymerization of Substituted Styrenes 3.2 Copolymerization of Styrene and Ethylene 3.3 Copolymerization of Styrene and Dienes 4 Mechanisms of Polymerization of Styrene 4.1 Active Site Species 4.2 Kinetic Analysis of Styrene Polymerization 4.3 Effects of Hydrogen 5 Conclusion References

370 375 375 377 377 378 378 382 385 386 386

18 Characterization, Properties and Applications of Syndiotactic Polystyrene Komei Yamasaki, Norio Tomotsu and Michael Malanga

389

1 Introduction 2 Characterization 2.1 Structure 2.2 Crystal Form 3 Physical Properties 3.1 Thermal Properties 3.2 Crystallization Behavior of SPS 3.3 Comparison of Crystallization Properties of SPS with IPS 3.4 Solvent Resistance 3.5 Rheological Properties 3.6 Mechanical Properties of Neat SPS 4 Properties of Commercialized SPS and Its Applications 4.1 Mechanical and Flow Properties 4.2 Electrical Properties 4.3 Chemical Resistance 4.4 Improvement of Polystyrene by Blending SPS 5 Summary References

389 390 390 390 392 392 393 395 396 397 399 401 402 402 404 405 408 408

19 Rubber Modification of Syndiotactic Polystyrene G. E. McKee, F. Ramsteiner and W. Heckmann

411

1 Introduction 2 Energy Dissipation in Polystyrene Polymers 3 Impact Behaviour of Rubber-modified sPS 4 Rubber Modification

411 412 415 417

xiv

CONTENTS

4.1 Styrene Block Copolymers as Impact Modifiers 4.2 Core-Shell Impact Modifiers 4.3 Preparation of sPS in the Presence of a Rubber 5 Present Situation and Future Perspectives References 20 Polymeric Blends Based on Syndiotactic Polystyrene L. Abis, R. Braglia, G. Giannotta and R. Po 1 2 3 4

Introduction Overview of sPS Properties Patent Literature on sPS Blends Microscopic, Thermal and Mechanical Properties of sPS Blends 4.1 Miscible Blends 4.2 Immiscible Blends 5 Conclusions 6 List of Abbreviations References V

418 423 428 428 429 431 431 431 433 438 439 447 458 459 460

STYRENIC BLOCK COPOLYMERS

463

21 Styrenic Block Copolymer Elastomers R. C. Bering, W. H. Korcz and D. L. Handlin, Jr

465

1 2 3 4

Introduction Synthesis of Styrenic Block Copolymer Elastomers Properties of Styrenic Block Copolymer Elastomers Applications of Styrenic Block Copolymer Elastomers 4.1 Commercial Styrenic Block Copolymers 4.2 Adhesives and Sealants 4.3 Bitumen Modification 4.4 Footwear 4.5 Polymer Modification 4.6 Viscosity Index Improvers and Other Applications References

22 Preparation, Properties and Applications of High Styrene Content Styrene-Butadiene Copolymers David L. Hartsock and Nathan E. Stacy 1 History 2 SBC Synthesis and Manufacture 3 Key Features, Properties and Grades

465 465 474 487 487 489 492 493 493 496 497 501 501 502 504

CONTENTS

xv

4 Current Commercial Applications 4.1 Major Markets 4.2 Single Service 4.3 Rigid Packaging 4.4 Garment Hangers 4.5 Flexible Packaging 4.6 Medical 4.7 Consumer Goods 4.8 Toys 4.9 Displays 5 SBC Blends 5.1 Clear Blends 5.2 Opaque Blends 5.3 Others 6 Future Applications References

507 507 508 508 511 514 515 518 519 519 520 520 525 528 529 529

VI NOVEL POLYSTYRENES

531

23 Hydrogenated Polystyrene: Preparation and Properties Stephen F. Hahn 1 Introduction 2 Synthesis of Polycyclohexylethylene (PCHE) 3 Catalytic Hydrogenation 3.1 Catalysis and Conditions 3.2 Hydrogenation Mechanism 4 Polymerization of Vinylcyclohexane to PCHE 5 Characterization of PCHE 5.1 Atactic PCHE 5.2 Isotactic PCHE 5.3 Syndiotactic PCHE 6 Copolymers Containing PCHE 6.1 Random Copolymers 6.2 Block and Graft Copolymers 6.3 Graft Copolymers 7 Proposed Applications of PCHE-based Materials 8 Acknowledgment References 24 Branched Polystyrene Kurt A. Koppi and Duane B. Priddy 1

Introduction

533

,

533 533 534 534 536 539 539 539 545 546 547 547 547 551 551 553 553 557 557

xvi

CONTENTS

2 Preparation of Branched Polystyrene 2.1 Radical Polymerization 2.2 Anionic Polymerization 3 Rheology of Branched Polystyrenes 3.1 Star Branched Polymers 3.2 Comb Branched Polymers 3.3 Randomly Branched Polymers 3.4 Extensional Rheology 4 Conclusions References 25 'Super Polystyrene' - Sryrene-Diphenylethylene Copolymers G. E. McKee, F. Ramsteiner, W. Heckmann and H. Gausepohl 1 Introduction 2 Preparation of DPE Monomers and Polymers 2.1 1,1 -Diphenylethylene Monomer Synthesis 2.2 S/DPE Polymer Synthesis 3 Properties of Styrene-Diphenylethylene Polymers 4 Blends of S/DPE Polymers 5 Rubber Modification of S/DPE Polymers 5.1 Modified High-impact Polystyrene (HIPS) Process 5.2 Core-Shell Impact Modifiers 5.3 Tri-block Copolymer of StyreneHydrogenated Butadiene-Styrene [S-B(H)-S] 5.4 Tri-block Copolymers of S/DPEHydrogenated Butadiene-S/DPE 6 Thermoplastic Elastomers 7 Summary References 26 Ethylene-Styrene Copolymers Y. W. Cheung and M. J. Guest 1 Introduction to Ethylene-Styrene Copolymers 2 Copolymerizations of Ethylene and Vinyl Aromatic Monomers 3 Structure-Property Relationships for Ethylene-Styrene Interpolymers 3.1 Thermal Transitions/Viscoelastic Behavior 3.2 Mechanical Properties 3.3 Melt Rheology and Processability 4 Materials Engineering Aspects

557 557 564 565 566 569 571 573 577 577 581 581 582 582 582 583 585 586 587 588 596 599 600 601 603 605 605 606 608 609 613 614 616

CONTENTS

4.1 Interpolymer Blends 4.2 Blends of Ethylene-Styrene Interpolymers: Miscibility Considerations 4.3 Filler Composites 4.4 Terpolymers 5 Attributes and Applications 6 Summary 7 Acknowledgments References VII PROPERTIES OF STYRENIC POLYMERS 27 Fracture Behaviour of High-impact Polystyrene and Acrylonitrile-Butadiene-Styrene T. Vu-Khanh 1 Introduction 2 Quantitative Characterization of Fracture 2.1 Brittle Fracture 2.2 Semi-ductile Fracture 2.3 Ductile Fracture 3 High-impact Polystyrene 3.1 Effect of Temperature 3.2 Effects of Loading Rate 3.3 Dynamic Effect and Adiabatic Heating 4 Acrylonitrile-Butadiene-Styrene 5 Conclusion References 28 Dynamic Mechanical Behaviour of Atactic Polystyrene, High-impact Polystyrene and Other Styrenic Polymers S. N. Goyanes and G. H. Rubiolo 1 Introduction 2 Polystyrene 2.1 Effect of Polymer Structure and Additives on the Dynamic Mechanical Spectroscopy of Polystyrene 3 Copolymers of Styrene 4 Rubber-modified Polystyrene (HIPS) and SAN Copolymers (ABS) References

xvii

617 617 620 623 625 626 627 627 631 633 633 635 635 637 639 643 645 648 653 654 661 662

665 665 666 667 676 678 681

xviii

CONTENTS

29 Flame-retardant Polystyrene: Theory and Practice Bruce King 1 Introduction 2 Applications of Flame-retardant Styrenic Polymers 3 Flammability Requirements and Tests 3.1 Regulatory Test Methods 3.2 Research Methods 4 Mechanisms of Flame Retardation 4.1 Vapor-phase Mechanisms 4.2 Condensed-phase Mechanisms 5 Halogen-based Flame Retardants for Styrenics 6 Styrenic Blends 7 Environmental Concerns 8 Summary References 30 Photochemical Degradation of Styrenic Polymers B. Mailhot, A. Rivaton and J. L. Gardette 1 Introduction 2 Photooxidation of the Homopolymer Polystyrene (PS) Under Irradiation at A > 300 nm 2.1 Experimental Results 2.2 Discussion 3 Photooxidation of Poly(styrene-coacrylonitrile) (SAN) 3.1 Experimental Results 3.2 Discussion 4 Photooxidation of Acrylonitrile—ButadieneStyrene (ABS) 4.1 Analysis of the Photooxidation 4.2 Photooxidation Rate 4.3 Discussion 5 Photooxidation of a Blend of SAN and EPDM (AES) 5.1 FTIR Analysis of AES Films During the First Stages of Photooxidation 5.2 FTIR Analysis of AES Films for Longer Irradiation Periods 5.3 Discussion 6 Photooxidation of Blends of Polystyrene and Poly(vinyl methyl ether) (PVME-PS) 6.1 Introduction

685 685 686 687 687 689 690 690 692 692 699 699 700 701 703 703 704 704 707 709 709 710 712 712 713 715 716 717 718 718 720 720

CONTENTS

xix

6.2 Experimental Results 6.3 Surface Analysis 6.4 Discussion 7 Conclusion References 31 Analysis and Levels of Styrene Dimers and Trimers in Polystyrene Food Containers Hiromi Sakamoto 1 2 3 4 5 6 Index

Introduction Structure and Analysis of Styrene Dimers and Trimers Content of SDs and STs in PS Food Containers Migration of SDs and STs from PS Food Containers Biological Evaluation of SDs and STs Conclusion References

720 721 722 723 725 727 727 728 730 731 737 742 743 745

This page intentionally left blank

Contributors L. Abis Polimeri Europa Centre Ricerche Novara 'Istituto G. Donegani' Via G. Fauser 4 1-28100 Novara Italy

R. Braglia Polimeri Europa Centre Recerche Novara 'Istituto G. Donegani' Via G. Fauser 4 1-28100 Novara Italy

R. C. Bening Kraton Polymers LLC Westhollow Technology Center 3333 Highway 6 South Houston, TX 77082 USA

A. Butte Federal Institute of Technology Zurich Laboratorium fur Technische Chemie ETH Honggerberg, HCI F125 CH-8093 Zurich Switzerland

M. Beulmann BASF AG D-67056 Ludwigshafen Germany

J. M. Ceraso 438 Bldg The Dow Chemical Company Midland, MI 48667, USA

G. Bouquet The Dow Chemical Company 200 Larkin Building 1605 Joseph Drive Midland, MI 48674 USA

Y. W. Cheung INSITE, Technology R&D The Dow Chemical Company Dow Texas Operations Freeport, TX 77541-3257 USA

XXII

CONTRIBUTORS

C. J. Cummings Polystyrene Research and Development 438 Bldg, Dow Chemical Company Midland, MI 48667, USA

H. Goerrissen BASF AG D-67056 Ludwigshafen Germany

R. P. Dion Materials Sciences The Dow Chemical Company 1707 Building Midland, MI 48674 USA

B. Gougeon The Dow Chemical Company 200 Larkin Building 1605 Joseph Drive Midland, MI 48674 USA

J. Gao Polymer Technology and Process Development BASF AG D-67056 Ludwigshafen Germany

S. N. Goyanes Departmento de Fisica Facultad de Ciencias Exactas y Naturales Universidad Nacional de Buenos Aires Ciudad Universitaria, Pabellon I 1428 Buenos Aires, Argentina

J. L. Gardette Laboratoire de Photochimie Moleculaire et Macromoleculaire UMR CNRS 6505 Universite Blaise Pascal (Clermont-Ferrand) F-63177 Aubiere Cedex, France

M. J. Guest Polythylene and INSITE Technology R&D The Dow Chemical Company 1707 Building Midland, MI 48674 USA

H. Gausepohl BASF AG D-67056 Ludwigshafen Germany

S. F. Hahn Polymer Chemistry Discipline, Corporate Research and Development The Dow Chemical Company 1707 Building, Midland, MI 48667 USA

G. Giannotta Polimeri Europa Centro Ricerche Novara 'Istituto G. Donegani' Via G. Fauser 4 1-28100 Novara, Italy

D. L. Handlin, Jr Kraton Polymers LLC Westhollow Technology Center 3333 Highway 6 South Houston, TX 77082 USA

CONTRIBUTORS

XXIII

D. L. Hartsock Chevron Philips Chemical Co. 201A ARE Bartlesville, OK 74004 USA

W. H. Korcz Kraton Polymers LLC Westhollow Technology Center 3333 Highway 6 South Houston, TX 77082, USA

W. Heckmann BASF AG ZK/Z B001 D-67056 Ludwigshafen Germany

B. Mailhot Laboratoire de Photochimie Moleculaire et Macromoleculaire UMR CNRS 6505 Universite Blaise Pascal (Clermont-Ferrand) F-63177 Aubiere Cedex, France

K.-D. Hungenberg Polymer Technology and Process Development BASF AG D-67056 Ludwigshafen, Germany

M. Malanga R&D Engineering Plastics The Dow Chemical Company Midland, MI 48667 USA

B. King The Dow Chemical Company Midland, MI 48667 USA

M. F. Martin BASF AG D-67056 Ludwigshafen Germany

A. Kistenmacher BASF AG D-67056 Ludwigshafen Germany

B. Matthews Dow Polystyrene R&D Midland, MI 48667 USA

R.-D. Klodt Dow Central Germany Building H 108 D-06258 Schkopau Germany

G. E. McKee BASF AG ZK/Z B001 D-67056 Ludwigshafen Germany

K. Koppi Dow Polystyrene R&D Midland, MI 48667 USA

B. J. Meister 2925 Chippewa Ln. Midland, MI 48640 USA

xxiv

CONTRIBUTORS

M. Morbidelli Swiss Federal Institute of Technology Zurich Laboratorium fur Technische Chemie, ETH Honggerberg, HCI F125 CH-8093 Zurich, Switzerland

F. Ramsteiner BASF AG ZK/Z B001 D-67056 Ludwigshafen Germany

N. Niessner BASF AG Styrene Copolymers and Ultraform ZKT/C-B1 D-67056 Ludwigshafen Germany

A. Rivaton Laboratoire de Photchemie Moleculaire et Macromoleculaire UMR CNRS 6505 Universite Blaise Pascal (Clermont-Ferrand) F-63177 Aubiere Cedex, France

A. N. Paquet The Dow Chemical Company 200 Larkin Building 1605 Joseph Drive Midland MI 48674 USA

G. H. Rubiolo Departmento de Fisica Facultad de Ciencias Exactas y Naturales Universidad Nacional de Buenos Aires Ciudad Universitaria, Pabellon I 1428 Buenos Aires, Argentina

A. Penlidis Chemical Engineering University of Waterloo Waterloo Ontario Canada, N2L 3G1

H. Sakamoto Kanagawa Environmental Research Center 1-3-39 Yonnomiya (Shinomiya), Hiratsuka Kanagawa, 2540014, Japan

R. P6 Polimeri Europa Centro Recerche Novara 'Istituto G. Donegani' Via G. Fauser 4, 1–28100 Novara Italy

R. L. Sammler Materials Sciences The Dow Chemical Company 1702 Building Midland, MI 48672 USA

D. B. Priddy Priddy & Associates LLC 6004 Camelot Ct. Midland, MI 48640, USA

J. Scheirs ExcelPlas PO Box 2080, Edithvale, VIC 3196 Australia

CONTRIBUTORS

XXV

J. Schellenberg R&D Engineering Plastics Dow Central Germany D-06258 Schkopau, Germany

J. P. Viola BASF AG D-67056 Ludwigshafen Germany

N. E. Stacy Chevron Philips Chemical Co. 201A ARB Bartlesville OK 740004 USA

T. Vu-Khanh Universite de Sherbrooke Faculte de Genie/Department de Genie Mecanique 2500 boul. De 1'Universite Sherbrooke, Quebec Canada, J1K 2R1

G. Storti Federal Institute of Technology Zurich Laboratorium fur Technische Chemie, ETH Hoggerberg, HCI F125 CH-8093 Zurich Switzerland

G. C. Welsh The Dow Chemical Company 200 Larkin Center 1605 Joseph Drive Midland MI 48674 USA

K.W. Suh The Dow Chemical Company 200 Larkin Building 1605 Joseph Drive Midland, MI 48674, USA

J. R. Wuensch BASF AG D-67056 Ludwigshafen Germany

N. Tomotsu Polymer Research Laboratory Idemitsu Petrochemical Co., Ltd Anesaki-kaigan, Ichihara Chiba, 229-0193 Japan

K.Yamasaki Plastics Technical Center Idemitsu Petrochemical Co. Ltd Anesaki-kaigan, Ichihara Chiba, 229-0193 Japan

R. Vanspeybroeck 438 Building Dow Chemical Company Midland, MI 48667, USA

This page intentionally left blank

Series Preface The Wiley Series in Polymer Science aims to cover topics in polymer science where significant advances have been made over the past decade. Key features of the series will be developing areas and new frontiers in polymer science and technology. Emerging fields with strong growth potential for the twenty-first century such as nanotechnology, photopolymers, electro-optic polymers, etc., will be covered. Additionally, those polymer classes in which important new members have appeared in recent years will be revisited to provide a comprehensive update. Written by foremost experts in the field from industry and academia, these books have particular emphasis on structure-property relationships of polymers and manufacturing technologies as well as their practical and novel applications. The aim of each book in the series is to provide readers with an in-depth treatment of the state-of-the-art in that field of polymer technology. Collectively, the series will provide a definitive library of the latest advances in the major polymer families as well as significant new fields of development in polymer science. This approach will lead to a better understanding and improve the cross fertilization of ideas between scientists and engineers of many disciplines. The series will be of interest to all polymer scientists and engineers, providing excellent up-to-date coverage of diverse topics in polymer science, and thus will serve as an invaluable ongoing reference collection for any technical library. John Scheirs June 1997

This page intentionally left blank

Polystyrene was the first synthetic polymer to be prepared. In fact there are reports of its existence as early as 1839 (see Chapter 1). Polystyrene was first produced for commercial sale in 1931 by BASF and in the US by Dow in 1938. It is well known that polystyrene is a glassy, amorphous polymer with outstanding clarity, gloss and processability. Unfortunately it is also inherently brittle and suffers from poor chemical resistance (see Chapter 2). These deficiencies were remedied early on by the development of high-impact polystyrene (see Chapter 12 on HIPS) and styrene-acrylonitrile copolymers (see Chapter 13 on SAN copolymers and Chapters 14 and 15 on ABS terpolymers,* as well as Chapter 16 on ASA terpolymers). Styrenic block copolymers have also considerably widened the scope of these polymers from elastomeric materials (Chapter 21) to high-clarity, high impact strength resins (Chapter 22). These latter durable copolymers offer a balance of performance and economics that bridges the gap between high cost, clear engineering polymers and low cost, brittle plastics like general purpose polystyrene. Since polystyrene is one of the oldest commercial polymers with over 9 million tonnes/yr of sales, there have been thousands of patents issued covering all aspects of its manufacture and property enhancement. The styrene monomer readily polymerizes to polystyrene either thermally or with free-radical initiators (see Chapter 6 on free-radical polymerization and Chapter 8 on nitroxidemediated polymerization). Commercial processes for the manufacture of polystyrene are described in Chapter 3 while process modelling and optimization of styrene polymerization is examined in Chapter 5. Styrene also can be polymerized via anionic and Ziegler-Natta chemistries using organometallic initiators. Using free radical and anionic polymerization chemistries, the Strictly speaking, ABS is a rubber-tonghened SAN copolymer.

xxx

PREFACE

specific position of the benzene ring in the monomer units of regular polystyrene is somewhat random and hence inhibits crystallization. Advances in the development of new metallocene polymerization catalyst technology however has enabled the development of syndiotactic polystyrene which is semicrystalline, has a melting point of 270°C and has excellent environmental stress crack resistance (see Chapters 17–20). New metallocene catalyst technology has also enabled the development of novel ethylene-styrene interpolymers (see Chapter 26). Modified variants of polystyrene have also been developed with the advent of hydrogenated PS (Chapter 23), branched PS (Chapter 24) and 'super' PS (Chapter 25). Since the strength, flammability and photodegradation of styrenic polymers have major end-use implications, these properties are covered in detail in Chapters 27, 28, 29 and 30 respectively. The high melt strength of polystyrene enables it to be easily foamed (see Chapters 9 and 10 on PS foam), blown into films, and drawn into sheets (see Chapter 11 on OPS). Polystyrene foams find a variety of uses including insulation and packaging. The family of styrenic polymers now span the breath from commodity plastics to high-grade engineering polymers. Ongoing advances in new catalyst technology and 'controlled radical polymerisation' will undoubtedly yield new styrenic polymers with well-defined architecture (as we have recently seen with the introduction of syndiotactic PS and ethylene-styrene interpolymers). Advances in the synthesis of dendritic and hyperbranched styrenic polymers will also contribute to the state of new polystyrenic products. The key attribute of polystyrene that has led to its huge commercial success is its low cost. Resistance of polystyrene fabricators to pay extra for improved performance and intense competition of polystyrene producers for increased market share have led to highly optimized and huge polystyrene production facilities (a typical 'world-scale' polystyrene plant produces about 230000 tonnes/yr of product). The costs associated with the introduction of new and improved polystyrene products must be low enough that profit can be realized by the manufacturer without raising the sales price. This limitation, and the ongoing effort of the chemical industry to scrutinize/justify R&D budgets, places an intense challenge on industrial polystyrene researchers. Other pressures on the polystyrene industry include environmental and regulatory issues (i.e., litter, migration of residual small molecules into food products, evolution of volatile organics during manufacture and processing, etc.) - see Chapters 4 and 31. These issues will undoubtedly dominate much of the research efforts devoted to polystyrene. Academic researchers are not under such focused cost constraints and therefore they will likely continue working on the development of new chemistries for making new styrenic polymers having novel controlled architectures.

PREFACE

xxxi

The future direction of polystyrene R&D efforts is uncertain but it is likely that it will continue focusing on lowering manufacturing costs, improving product performance/properties (especially flow/strength balance), reducing the level of residual small molecules left in the product, and developing new applications. This book provides the reader with comprehensive information about polystyrene, and a historical overview of its development, as well as reviews describing the latest new technological developments. J. Scheirs and D. B. Priddy June 2002

This page intentionally left blank

About the Editors JOHN SCHEIRS (PhD) John Scheirs is a polymer research specialist with broad interests in polystyrenes and styrenic copolymers. He is the principal consultant with ExcelPlas, a polymer consulting company. John was born in 1965 in Melbourne and studied applied chemistry at the University of Melbourne and obtained a PhD in polymer science. He has worked on projects concerning the fracture, stress cracking, processing, characterization and recycling of styrenic polymers. John has authored over 50 scientific papers, including eight encyclopedia chapters, and a number of books on polymer analysis and polymer recycling.

DUANE PRIDDY (PhD) Duane B. Priddy has worked for the Dow Chemical Company for 33 years, having retired at the end of 2001 from his role as Research Scientist in Polystyrene R&D. Duane began his career in Dow in the Benzene Research Laboratory in 1966. In 1968 he took a two-year leave of absence to attend Michigan State University where he obtained a PhD in Organic Chemistry. In 1970, he returned to Dow. In 1972, Duane joined Polystyrene R&D where he developed the initiator DP275 for polystyrene manufacture. DP275 is currently utilized globally in all Dow polystyrene production plants and is recognized as the industry standard. Duane holds more than 65 patents and has published more than 100 technical papers outside of Dow. He is an adjunct professor at Central Michigan University and at Michigan Technological University. In 2001, Duane was named a Fellow of the American Chemical Society and was awarded the Lifetime Achievement Award by Dow's Polystyrene Business. Most recently he was awarded the Excellence in Science Award by the Midland Section of the American Chemical Society during their annual Fall Scientific Meeting.

This page intentionally left blank

PART I

This page intentionally left blank

1 Historical Overview of Styrenic Polymers JOHN SCHEIRS ExcelPlas, Edithvale, VIC, Australia

1 INTRODUCTION Styrene has been known from the mid-nineteenth century as a clear organic liquid of characteristic pungent sweet odour. It was also known to have the ability to convert itself under certain conditions into a clear resinous solid that is almost odour-free, this resin then being referred to as 'metastyrol'. Styrene readily polymerizes in air and it is therefore not surprising that there are a number of early obscure reports referring to its 'polymerization' predating 1900. However, because the concept of polymerization had not yet been proposed (until Staudinger in 1920), many of these early reports referred to the 'oxidation' or 'hardening' of the styrene monomer. In 1839, Eduard Simon, an apothecary in Berlin, distilled storax resin obtained from the 'Tree of Turkey', (liquid ambar orientalis) with a sodium carbonate solution and obtained an oil which he analysed and named styrol (what we now call styrene) [1]. He recorded the following observation: 'that with old oil the residue which cannot be vaporised without decomposition is greater than with fresh oil, undoubtedly due to a steady conversion of the oil by air, light and heat to a rubberlike substance'. Simon believed he had oxidised the material and called the product styrol oxide. Later, when he realised that it contained no oxygen, the product became known as metastyrol. This puzzled the early chemists as there was no change in empirical formula despite the very pronounced alteration in chemical and physical properties. Unknowingly, this was the first recorded instance of polymerization.

Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy CO 2003 John Wiley & Sons Ltd

4

J. SCHEIRS

A few years later, in 1845, Blyth and Hofmann [2] observed that 'metastyroF was formed when styrene was exposed to sunlight, while it remained unchanged in the dark. This is the first report of photopolymerization. It was established by Blyth and Hofmann in 1845 [2] that 'styrol and its conversion product 'metastyroF had the same elemental composition with an equal number of carbon and hydrogen atoms. After nitration, the N:C ratio was 1:8 for styrol and 1:7 for metastyrol, leading to the conclusion that C8H8 had been converted to C 7 H 7 . However, the conclusion that styrol had polymerized was not reached until 75 years later. The first samples of polystyrene were characterized by the German organic chemist Staudinger [3]. It was observed that the polystyrene could be fractionated into samples with different solution viscosities and this observation was incompatible with the notion that the substance was a colloidal aggregate. Staudinger challenged the notion that polymeric substances are held together by 'association forces'. It was Staudinger who first realised that the solid that Simon had isolated from natural resin was in fact composed of long chains of styrene molecules. Staudinger postulated that polystyrene was a high molecular weight polymer. His critics argued that it could not be a high molecular weight polymer because of its solubility in common solvents. He introduced the term 'macromolecules' to describe these long-chain compounds. Fierce controversy with his colleagues caused Staudinger to move from the Swiss Federal Institute of Technology in Zurich (ETH) to the University of Freiburg. In 1929, Staudinger and co-workers also synthesized hexahydropolystyrene by the nickel-catalysed hydrogenation of polystyrene [4,5]. The hydrogenated polystyrene, also known as poly(cyclohexylethylene), had improved oxidative and radiation stability relative to conventional polystyrene. It was also Staudinger in 1932 who first proposed that the inability of polystyrene to crystallize was due to its lack of stereoregularity which rendered it amorphous. It is its amorphous nature that is responsible for its solubility though others claimed that polymer solubility was incompatible with very high molecular weight [3].

2 GENERAL-PURPOSE POLYSTYRENE (GPPS) Styrene readily polymerizes to polystyrene (PS) either thermally or with freeradical initiators. A limiting factor in the commercial exploitation of polystyrene was the high reactivity and considerable heat of polymerization of styrene. The polymerization rate of styrene is exceedingly fast and considerable heat is generated. This was an intimidating obstacle to commercial production of PS since many in the industry were concerned that the large-scale polymerization of styrene may result in a dangerous uncontrolled reaction. The process

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

5

involved heating styrene monomer in bulk containers. A major limitation of this approach was the need for heat removal from the highly viscous melt. High temperatures can be reached in large mass reactors (>300°C) and thermal degradation of the resultant PS can occur. This problem was later solved by installing heat exchanger tubes in the reaction medium. The first commercial production of PS was in 1931 by BASF. To prevent premature polymerization of styrene monomer, special inhibitors had to be added so that it could be stored until needed. Polymerization inhibitors were also required to prevent polymer formation during distillation of styrene monomer from ethylbenzene. Before the large-scale production of PS could occur, a consistent supply of styrene monomer was required. In 1930, Dow started to produce styrene monomer by cracking its ethylbenezene precursor. In 1938, Dow began manufacturing commercial quantities of PS. The propensity for styrene monomer to polymerize allowed extremely simple and crude polymerization techniques. The first technique used by Dow was known as the 'can process' since it basically involved filling 10 gallon metal cans (Figure 1.1) with styrene monomer followed by heating the cans in a heating bath at progressively higher temperatures for a number of days. After this time the polystyrene (polymerized to approximately 99 % conversion) was removed from the can and crushed to a free-flowing powder. The development of styrene and PS manufacturing technology was spurred on by the advent of World War II. During this time the supply of natural

Figure 1.1 Early photograph of the 'can' process for the commercial production of polystyrene. This simple process involved filling 10 gallon metal cans with styrene monomer, thermally polymerizing it in heated baths and then grinding the polystyrene cylinders that formed, (courtesy of Dow Chemical Company)

6

J. SCHEIRS

rubber from the Far East was terminated. The acute rubber shortage accelerated the development of styrene-based synthetic rubber. With the outbreak of the war, the United States embarked on a scientific programme that rivalled the Manhattan Project in its scope and significance. Nearly a billion dollars was spent on research and development of synthetic rubber needed to keep the Allied war effort in motion. The key players such as Dow, Monsanto and Koppers Chemical cooperatively produced record quantities of styrene monomer for the preparation of styrene-butadiene rubber. A staggering 180 0001 of styrene monomer were produced per year towards the end of the War with most being used for the production of the synthetic rubber Buna-S (also known as GRS rubber, where GR stands for government rubber and S for styrene). There was also cooperation between the main rubber-producing companies, Goodyear, B. F. Goodrich, Standard Oil, Firestone and US Rubber. In early 1942, the American Synthetic Rubber Research Program commenced. Along with the major rubber-producing companies, 11 university research groups, including Carl 'Speed' Marvel at the University of Illinois, Izaak 'Piet' Kolthoff at the University of Minnesota and W. D. Harkins and Morris Kharasch of the University of Chicago, joined the effort to make synthetic rubber work. Their objective was to set up four plants that would produce 30 000t each of synthetic rubber per year. By the end of 1942, four plants were established but their output was under the target. By the end of 1943, 15 plants were in operation, and supply had begun to meet demand. The research focus during the War was on refinement, enhancement and incremental improvement of existing processes. For example, if the rubber is allowed to polymerize until no monomer is left then long, branched molecules are produced, which gel and make the rubber difficult to process. To solve this problem, the reaction is only allowed to proceed to 72 % conversion and a thiol modifier, a chain-transfer agent, was used to control molecular weight. It was also observed that the polymerizations have an 'induction period' which varied from batch to batch. During the induction period nothing seems to be happening, then, all of the sudden, the reaction takes off. The researchers at the University of Illinois found that this is due to different fatty acids present in the different soaps needed for the emulsion process. These soaps also cause the solution to foam during the recovery of the remaining monomer. This problem leads to the development of silicone defoamers. The properties of the Buna-S type rubber are highly dependent on the amount of styrene in the rubber. To determine properties, it is important to know how much styrene had been incorporated. William O. Baker of Bell Telephone Laboratories solved this problem by developing a procedure for determining the amount of styrene using the refractive index of a solution of the rubber. It was not until after World War II, when styrene monomer capacity could be diverted from its essential wartime use for styrene-butadiene synthetic rubber, that polystyrene became an important commercial plastic. When the War fin-

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

7

ished the supply route for natural rubber was re-established and there was an oversupply of styrene monomer. The extensive infrastructure for styrene production and the enormous body of process and technical knowledge laid the foundation for the post-War development of polystyrene and styrenic copolymers. Prior to 1941, Germany had a major technical and industrial lead over the USA, having already established an industrial styrene monomer production process, a styrene-butadiene elastomer process and a mass styrene polymerization process [6]. Figure 1.2 shows the polymerization vessels at I. G. Farben in 1940. Figure 1.3 shows a bank of polymerization kettles. The Germans began the first technical production of polystyrene in 1930 while the first production of polystyrene in the USA was some 8 years later by Dow in 1938. Interestingly, at the inception Dow did not have a strategic objective to enter into the polystyrene business. Rather, Dow believed that ethylcellulose and poly (vinylene chloride) (Saran) were the commodity polymers of the future. Dow was producing ethylbenzene as a solvent and electrical fluid. However, when the markets for ethylbenzene did not develop it decided to crack the ethylbenzene and produce styrene. After it had stockpiled large quantities of unstable styrene, Dow initiated a 'crash' programme to develop polystyrene. Thus even though Dow did not initially intend to produce polystyrene commercially when its petrochemical programme was initiated it became a logical business decision to do so. Early on there were numerous technical barriers that made polystyrene difficult to produce and to process. For example, it was made by an extremely slow production process and its high average molecular weight and broad molecular weight distribution made it difficult to injection mould [6]. Dow researchers ultimately developed ways to lower the average molecular weight and added certain lubricants to improve processability, thus making general-purpose polystyrene which fast acquired the reputation of being the easiest thermoplastic to mould. Other technical barriers were the need to control the exotherm of polymerization and to produce colour-free polystyrene. While the manufacture of styrene seems simple and straightforward, in the early days at Dow there were three major impurities in the styrene monomer apart from residual ethylbenzene. These were phenylacetylene (which acted as an inhibitor for styrene polymerization), divinylbenzene (which caused plugging and fouling of the distillation column for separating styrene from its precursor, ethyl benzene) and sulphur (which caused discoloration of the polystyrene). Finally, in 1938, the 'crash' programme resulted in the first saleable polystyrene batches. This was produced in metal cans lined with tin to yield high-purity polystyrene. These cans were filled with styrene and immersed in heated waterbaths where the styrene would thermally polymerize. The process was very slow and labour intensive. Further, the exotherm of polymerization was greatest in the centre of each can, which led to a core of lower than average molecular weight. After polymerization was complete, the polystyrene billet was ground up and mixed to distribute the different molecular weight regions [6].

J. SCHEIRS

Figure 1.2 Photograph taken in 1940 of a styrene polymerization vessel inside the I. G. Farben plant in Ludwigshafen, Germany (courtesy of BASF, Ludwigshafen)

Figure 1.3 Reaction kettles in the BASF polystyrene production plant (courtesy of BASF, Ludwigshafen)

10

J. SCHEIRS

Although the can process was very slow, it lent itself to easy expansion of production output by simply adding more cans and heating baths. It was also discovered that by adding some peroxide catalyst to the styrene monomer the production throughput could be increased significantly. In fact, this innovation led to a doubling of Dow's plant capacity since faster polymerization rates could be achieved while still controlling the exotherm [6]. While the USA was progressing with the can process, Germany had already developed a continuous process for the mass polymerization of styrene. After World War II, researchers from Dow visited the German polystyrene plants and were surprised to learn of their scale and sophistication. One of the key people on the investigating team that went to Germany was Dr Goggin, founder of Dow's Plastics Technical Service Department. The American teams that visited I. G. Farben after the War recorded their findings in a historic report [7]. This report clearly showed the advantages of a continuous production process for polystyrene. Further, the first industrial production of SAN was in 1936 also by I. G. Farben in Ludwigshafen. Central to Germany's development of polystyrene technology was Herman F. Mark (Figure 1.4). Mark worked at I. G. Farben Industrie for 6 years from 1927 to 1932, first as a research chemist (1927-28), then as Group Leader (1928–30) and finally as Assistant Research Director (1930-32). Because of the changing political climate, Mark moved to the University of Vienna, where he became Professor of Chemistry and Director of the First Chemical Institute (1932-38). While at I. G. Farben Industrie, Mark played a major role in the development of styrene monomer and PS. Mark patented a process in 1929 for the production of styrene from ethylbenzene via catalytic dehydrogenation [8]. The German Chemical giant I. G. Farben developed the continuous tower process for PS in the 1930s. The German PS polymerization plant shown in Figure 1.5 overcame the problem of the polymerization exotherm and thermal runaway by using a tank reactor with heat-transfer tubes criss-crossed through it. The reaction temperature was gradually increased and controlled, and polystyrene was removed via an auger. This design was later improved by prepolymerizing in stirred kettles prior to the tower process (Figure 1.6). After the War, Dow began to focus on constructing its own continuous mass polymerization plants for PS. Known as the tube tank process, it consisted of two nonagitated horizontal tube tanks containing arrays of tubes through which a heat-transfer fluid (Dowtherm™) flowed in order to control the exotherm of polymerization (Figure 1.7). Each tank had a capacity of 18000kg of styrene monomer and represented a batch process, but when alternately sequenced the process was continuous. When the styrene in tank 1 had reached high conversions, a special polymer pump pumped the molten polystyrene at 220240 °C to the bottom receiving tank. Polymerization was then started in tank 2. The bottom receiving tank was under vacuum to remove volatiles such as unreacted monomer and also dimers, trimers and other oligomers. There was

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

11

always polystyrene in the receiving tank so that the extruder and pelletization process could operate on a continuous basis [6]. The exotherm in the Dow process supplied most of the heat needed to produce a molten PS ready for pelletization. These units were extremely successful because of the very large heat transfer surface and the efficiency of the Dowtherm™ heat transfer fluid. Twelve such tube tank plants were installed and operated at the Dow site in Midland, Michigan and they produced prodigious quantities of polystyrene over many years [6].

Figure 1.4 Photograph of Herman F. Mark taken in 1936. Mark worked at I. G. Farben Industries in Germany for 6 years, from 1927 to 1932, and played a major role in the industrial development of styrene monomer and polystyrene (courtesy of BASF, Ludwigshafen)

12

J. SCHEIRS Styrene Reaction Temperature

!00°C

!50°C

BASF's Continous Styrene Polymerization Plant - circa 1932

Figure 1.5 Schematic of BASF's early tower process for the continuous polymerization of styrene. This configuration was designed by C. Wulff and E. Dorrer in the early 1930s. Polymerization was thermally initiated and the exotherm controlled by heat transfer tubes (courtesy of BASF, Ludwigshafen)

In the following years, other methods of polymerizing styrene were developed such as suspension polymerization by Koppers Chemical, which was first introduced in the 1940s and which showed rapid development in the 1950s. The suspension polymerization process is still in use for the production of PS [9], although it has been largely replaced by more economical techniques such as continuous mass polymerization. It is interesting that the polystyrene produced by suspension polymerization, particularly the Koppers material, had a heat distortion temperature superior to that of the Dow polystyrene [6]. This was attributed to the measurable levels of residual dimers and trimers in the Dow product due to its thermal initiation and which were absent in the peroxide-initiated suspension process. The suspension polymerization process has a number of distinct advantages over competitive processes. It allows excellent control over the polymerization temperature and a lower viscosity reaction medium. Furthermore,

HISTORICAL OVERVIEW OF STYRENIC POLYMERS styrene

13

styrene reaction temperature

prepolymerization

polymerization

extrusion BASF's Continuous Styrene Polymerization Plant - circa 1936 Figure 1.6 Schematic of BASF's improved tower process for the continuous polymerization of styrene. In this design (dated 1936) the styrene was first polymerized up to 30–35% conversion in a stirred kettle and then transferred to the tower reactor for polymerization up to 97% completion (courtesy of BASF, Ludwigshafen)

expandable polystyrene and high-impact polystyrene are also produced by this technique.

3

FOAMED POLYSTYRENE

The concept of cellular polystyrene was first reported in 1935 by the Swedish inventors Munters and Tandberg [10], who filed a patent entitled 'Foamed Polystyrene'. Ray Mclntire, a young researcher at Dow Chemical, is credited with inventing Styrofoam. Mclntire said his invention of foamed polystyrene was accidental. His invention came as he was trying to find a flexible electrical insulator around the time of World War II. He worked at developing a rubberlike substance that could serve as an electrical insulator. Although polystyrene was a good insulator it was far too brittle. Mclntire tried to make a new rubberlike polymer by combining styrene with isobutene, a volatile liquid, under pressure. He tried combining styrene with isobutene, but he accidentally

14

J. SCHEIRS MONOMER TUBE TANK I o oo o o ooo o o o o oo o oo o oo o oooooo 0 0 0 0 0 0 0

o ooo o ooo oooo ooo oo o

o o o o oo oO oo

o o o o

o o o o

o o o o

o o o o

oo oo o o

O O O 0

o oo

MOLTEN POLYMER PUMPS

TUBE TANK 2 o oo o o o o o oo 0000 o o o o o oooo o o o o o ooooo o o o o o ooooo o o o o o ooooo o o o o o ooooo o o o o o oooo o o o o o oooo o o o o o ooo O O O 0 o o

o o o o o o o o

o o o o

X^

VACUUM

MOLTEN V. ••'/ POLYMER ••// • 's

COLORANTS I I

z

1

t

EXTRUDER

Figure 1.7 Schematic of Dow's 'tube tank' process which represents the first commercial continuous polymerization process for polystyrene in the USA. The figure shows a cross-section through the centre of three longitudinal unagitated tanks. Styrene was thermally polymerized in tube tanks 1 and 2 and then devolatilized in the bottom receiving tank, which was always about half full and under vacuum [adapted from Boyer, R. F., J. Macromol. Sci. Chem., A15, 1411 (1981)] added too much of the latter - and was surprised to see that the isobutene formed tiny bubbles. The result was a foamed polystyrene with a cellular microstructure, 30 times lighter than regular polystyrene. Mclntire stayed with Dow Chemical until his retirement in 1981. The word Styrofoam is still trademarked by Dow, and it technically only applies to a kind of insulation

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

15

used for building materials. Today, however, many companies produce products made of PS foam, and the brand name Styrofoam is commonly used to describe them all. It was only in the early 1940s that commercial production of cellular polystyrene began. In 1942, Dow began research on an extrusion process for the production of PS foam using a low-boiling chlorocarbon (methylene chloride) as the blowing agent. The product was extruded into large foam logs, which were then cut into boards and planks. This material was given the trademark Styrofoam™ in 1943 [11]. This foam was rapidly adopted by the US Coast Guard and US Navy as a buoyancy medium and insulation material. BASF developed its own process for foaming polystyrene in the early 1940s (Figure 1.8). This process was later refined by the improved suspension polymerization process which produces foamable polystyrene beads. A blowing agent (typically pentane) can be introduced during the polymerization of styrene or introduced later in a separate impregnation step under pressure and heat [11]. All major PS foam bead producers took out a license from BASF for this patented technology [12–14]. Figure 1.9 shows a promotional illustration from 1952. The single most important factor responsible for the rapid commercial growth of expandable PS is its ability to be steam-moulded into lightweight, closed-cell, low-cost foams suitable for beverage cups, packages, ice buckets, picnic chests, insulation board, etc [11]. (Figure 1.10).

Figure 1.8 Early photograph (ca 1948) showing some of the earliest polystyrene foam (Styropor™). Foamed polystyrene has unrivalled low-density and thermal insulating properties (courtesy of BASF, Ludwigshafen)

16

J. SCHEIRS ,,Das fekhteste Schiff der Welt ist das STYROPOR-Schiffchen

Figure 1.9 A promotional photograph highlighting showcasing polystyrene foam for the BASF stand at the 1952 Kunststoffemesse show in Dusseldorf, Germany (courtesy of BASF, Ludwigshafen)

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

17

Figure 1.10 Polystyrene foam rapidly became the packaging material of choice for everything from medical instruments to engines (courtesy of BASF, Ludwigshafen)

In the late 1960s, demand for foamed PS increased dramatically owing to its increasing use in meat trays, fruit boxes and egg cartons. The expansion was stimulated by the design of extruders with provision for introducing the blowing agent into the barrel, thereby obviating the need to buy the more expensive expandable PS pellets as a raw material [11]. In 1969, the market for PS foam grew enormously especially for self-extinguishing grades. The main applications were for insulation board in cold-storage refrigeration, housing insulation and cut ceiling tiles.

18

4

J. SCHEIRS

RUBBER-MODIFIED POLYSTYRENE

Rubber-modified polystyrene was the next logical evolution after generalpurpose polystyrene. Very early on it was apparent that the Achilles heel of polystyrene was its inherent brittleness. Rubber-modified polystyrene is a twophase system consisting of a dispersed rubber phase and a continuous polystyrene phase (or matrix). Impact-modified polystyrene was invented as early as 1927 by Ostromislensky [15] by addition of natural rubber either polymerized with styrene or blended in polystyrene. In the early 1940s, researchers at Dow produced interpolymer blends of styrene and butadiene by an emulsion process. The polymer, called Styralloy™ 22, was used as insulation for radar cables until it was displaced by low-density polyethylene produced by ICI. Later, Dow experimented with soluble GRS copolymerized with styrene to make high-impact polystyrene. In 1954, Haward of Shell obtained a patent for rubber-modified PS made by suspension polymerization [16]. This early product, however, contained 'fisheyes' - small crosslinked gel particles - since each suspension particle was crosslinked by styrene-butadiene rubber. Researchers at Monsanto overcame this problem [17] by including a prepolymerization step with shearing agitation. In 1954, Dow finally perfected a 'can' process to make high-impact polystyrene (HIPS). The secret was that the traditional 'can' process could not simply be used since the product would be full of gel particles of rubber ('fish-eyes'); instead, the styrene-rubber mixture was first carried out to 30% conversion with shearing agitation. Then the mixture was transferred to 10 gallon cans where the reaction was completed. This process was documented in the now famous Amos patent [18]. In the late 1960s, Dow initiated patent infringement suits against its major competitors (Monsanto, Standard Oil, Amoco Chemical, Dart Industries) over the patent for high-impact polystyrene by Amos [6].

5 ABS While Dow had experimented extensively with ABS-type polymers and even produced ABS in their commercial HIPS plant, they lost the lead in the development of commercial ABS resins. Dow sued Monsanto in 1969 for infringement of the ABS claim in the patent by Amos. The judge ruled in Monsanto's favour on the basis that Monsanto's product was outside the claimed composition. In another case, where Dow sued Dart Industries, the judge ruled that the Amos patent was invalid for reasons of obviousness [6].

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

19

As early as 1948 it was known that poly(styrene-co-acrylonitrile) or SAN could be blended with Buna-N (a copolymer of butadiene and acrylonitrile) or Buna-S (a copolymer of butadiene and styrene) to produce useful thermoplastics [19]. The commercial introduction of these polymers, however, was restricted by their poor low-temperature impact properties. Researchers at Marbon (a division of Marsene Corp., later renamed Borg Warner) knew that polybutadiene remained rubbery at temperatures lower than the earlier cited copolymers; however, polybutadiene and SAN were incompatible [20]. The trick was first to produce polybutadiene by emulsion polymerization and then to use this latex as the medium for the emulsion polymerization of styrene and acrylonitrile (Figure 1.11). In 1959, Borg Warner patented this ABS produced by grafting styrene plus acrylonitrile into polybutadiene. The Cyclolac™ brand of grafted ABS produced by Borg Warner rapidly became the market leader in ABS. In 1988, GE Plastics acquired Borg Warner Chemicals and its ABS technology. The chief researcher responsible for this discovery was Calvert [21,22]. The material he produced from the emulsion copolymerization had uniformly dispersed domains of rubber in a continuous phase of SAN. The rubber dispersion was stabilized by the SAN that was grafted to the polybutadiene emulsion particles [20]. It is interesting to note that Borg Warner conducted more than 10000 laboratory trials before the optimum ABS composition was commercially produced in Styrene Acrylonitrile

Emulsion Polymerization of Styrene and Acrylonitrile Figure 1.11 Schematic of BASF's stirred tank emulsion polymerization reactors (dated 1940) for the production of styrenic copolymers (courtesy of BASF, Ludwigshafen)

20

J. SCHEIRS

1954. In 1957, Cyclolac™ T was commercialized. The T designated toughness and later it was thought to stand for telephone as this grade of ABS became the industry standard for making telephone housings. The ability of ABS sheets to be thermoformed opened the door to a range of markets such as luggage (Samsonite™), machine housings, refrigerator liners and boat hulls. When ABS was first commercialized, there was much confusion in the plastics industry referring to it as a terpolymer. The system is not a terpolymer as butadiene is added to the reactor as a polymer along with styrene and acrylonitrile monomers. Polymerization causes SAN to be grafted to the rubber to produce a dispersible domain. It is indeed a requirement that the polybutadiene regions exist as a separate phase of a specified size. Since the domain size is critical to its impact properties, it is important that it is stable through compounding and processing steps [20]. High-heat versions of ABS were subsequently introduced by using a-methylstyrene as a partial replacement of styrene, in both the graft and the matrix. This resulted in an ABS polymer with a higher heat deflection temperature. The heat deflection temperature could be tailored by varying the level of a-methylstyrene. These new high-heat versions of Cyclolac™ ABS were introduced to the automotive market in 1958 and won widespread acceptance [20]. The high heat resistance grades, however, had higher melt viscosities, making them more difficult to process owing to the stiffer molecule that results from the addition of methyl groups to the backbone. The next evolution in ABS technology was the need to produce a transparent ABS. Existing ABS was opaque owing to the scattering of light by the rubber domains. While producing smaller domains would make the system clear, it led to a loss of impact strength. The answer was to modify the refractive index of the components so that the various phases were less optically different. A fourth 'monomer', methyl methacrylate, was used to minimize the refractive index variation in the ABS and a clear impact-resistant thermoplastic named Cyclolac™ CIT was achieved [20].

6 ASA

Acrylonitrile-styrene-acrylate (ASA) polymers share obvious similarities with ABS but ASA was only developed in the 1960s. ASA polymers are essentially SAN polymers impact modified with an acrylate rubber. The earliest attempt to make ASA was by Herbig and Salyer of Monsanto [23] using butyl acrylate as the rubber phase. This work was then refined by Otto [24] and Siebel [25], both of BASF, who copolymerized butyl acrylate with butadiene to prepare the rubber phase.

HISTORICAL OVERVIEW OF STYRENIC POLYMERS

7

21

EARLY STYRENE COPOLYMERS

The first styrene copolymer was reported in 1930 by Wagner-Juaregg and was a copolymer of styrene and maleic anhydride [26]. This copolymer (SMA), which was called a heteropolymer by its inventor, has excellent resistance to continuous exposure in boiling water. Koppers produced SMA moulding powders under the tradename Dylark™. Arco has since acquired this business and continues to produce these SMA resins today under the Dylark tradename. Another styrene copolymer with better heat resistance than regular polystyrene is the copolymer of styrene and fumaronitrile which was reported in 1948 [27]. Both of these styrene copolymers are based on nonpolymerizable monomers - that is, fumaronitrile, like its corresponding anhydride (maleic anhydride), does not form homopolymers but readily copolymerizes with styrene at levels of up to 40%. Monsanto attempted to commercialize the styrene-fumaronitrile copolymer under the tradename Cerex™, but residual fumaronitrile was a powerful vesicant (an irritant which causes blisters) and the project was shelved [28].

8

STYRENIC BLOCK COPOLYMERS

The advent of alkyllithium-initiated anionic polymerization based on the fundamental work by Szwarc in 1956 [29,30] opened the door to the commercial development of styrene-butadiene copolymers. Styrenic block copolymers were first produced in the late 1950s after the discovery by Szwarc of living anionic polymerization [31]. In the late 1950s, Shell, Phillips and Firestone manufactured styrene-butadiene copolymers produced by anionic polymerization also referred to as 'living polymerization'. The styrene-butadiene copolymer named K-resin with a high styrene content was invented in the early 1960s at Phillips Petroleum by Alonzo Kitchen and is purportedly named after his surname initial. The alkyllithium initiators such as n-butyllithium allowed the production of copolymers with tightly defined and controlled polymer microstructures with highly regular block structures. To this day, living anionic polymerization continues to be the main commercial route for the production of styrenic block copolymers [31]. Most of these polymers are made using butyllithium catalysts. Today the following styrene–butadiene copolymers are well known under the tradenames Kraton™, K-Resin™, Styroflex™ and Styrolux™.

22

J. SCHEIRS

9 SYNDIOTACTIC POLYSTYRENE Syndiotactic polystyrene (SPS) with its melting point of 270 °C has been claimed as the first styrenic engineering plastic. Syndiotactic polystyrene was first synthesized in 1985 by Ishihara of Idemitsu Kosan by using titanium complexes/methylaluminoxane catalyst [32–34]. SPS displays entirely different properties to conventional polystyrene such as high chemical resistance and excellent environmental stress crack resistance. SPS does share one major property with conventional PS, however, namely inherent brittleness. For this reason, SPS is generally modified with rubber tougheners or glass fibre reinforcement. When SPS is reinforced with glass fibres it has comparable toughness to glass-filled PBT and glass-filled nylon 66. Blending of SPS with other polymers is another strategy for improving performance properties (see Chapter 20). Dow and the Idemitsu Petrochemical entered into a cooperation in 1988 and are now commercially producing SPS under the tradenames Questra™ and Xarec™, respectively.

10 MODERN POLYSTYRENE PRODUCTION In the late 1970s, Dow undertook a major programme to change the way polystyrene was commercially produced. The conversion involved moving from thermally initiated polystyrene polymerization to polymerization initiated by a bifunctional initiator (Dow calls the initiator DP275, named after its inventor Duane Priddy). Today the majority of general-purpose polystyrene is produced by solution polymerization in a continuous process. The solution process allows for the easy removal of heat from the polymerization medium. This technique, however, necessitates the use of exhaustive post-polymerization devolatilization equipment employing high temperatures and high vacuum to ensure the removal of volatiles and oligomers.

11 THE FUTURE Figure 1.12 shows the timeline of discovery of various styrenic polymers and copolymers. It would be naive to suggest that the rate of invention and innovation will level off in this century. Rather, the pace of discovery of new styrenic polymers will probably increase. Advances in new catalyst technology and 'controlled radical polymerisation' technology will undoubtedly yield new styrenic polymers with well-defined architecture, as we have recently seen with the introduction of Syndiotactic PS and ethylene-styrene interpolymers.

23

HISTORICAL OVERVIEW OF STYRENIC POLYMERS DOW 1 ES

IDEMITSU

SYNDIOTACTIC PS

BASF

LOCK COPOLYMERS ASA ABS

HIGH IMPACT PS

BASF STYRENE^MONOMER

BASF I 1930

1940

1950

1960

1970

1980

1990

2000

Figure 1.12 Timeline of the development of styrenic polymers (adapted from a BASF document by Franz Haaf, entitled '50 Jahre Polystyrol - Entwicklung', BASF, Ludwigshafen)

Advances in the synthesis of dendritic and hyperbranched styrenic polymers will also contribute to the slate of new products. Hitherto uncommerciallized are a range of polymers in which styrene is copolymerized with three or more other monomers. This is because styrene readily copolymerizes with other monomers, even nonpolymerizable monomers such as anhydrides and nitriles. As far as process technology is concerned, all commercial polystyrene has until now been produced using free radical chemistry. BASF and Asahi have recently been aggressively developing 'retarded' anionic polymerization process technology which will allow them to manufacture polystyrene (both general-purpose polystyrene and high-impact polystyrene) in continuous bulk polymerization processes. The advantage of these new anionic processes will be the production of highly pure PS virtually free of traces of volatile impurities such as unreacted styrene monomer. Also, anionic polymerization processes allow the production of PS having a more narrow polydispersity, resulting in an improved flow/ strength property balance.

REFERENCES 1. Simon, E., Ann. Chem., 31, 265 (1839). 2. Blyth, J. and Hofmann, A. W., Ann. Chem., 53, 289 (1845).

24

J. SCHEIRS

3. Staudinger, H., Die Hochmolekularen Organischen Verbindungen, Kautschuk und Cellulose, Springer, Berlin, 1932. 4. Staudinger, H., Geiger, E. and Huber, E., Chem. Ber., 62, 263 (1929). 5. Staudinger, H. and Wiedersham, V., Chem. Ber., 62, 2406 (1929). 6. Boyer, R. F., J. Macromol. Sci. Chem., A15, 1411 (1981). 7. DeBell, J. M., Goggin, W. C. and Gloor, W. E., German Plastics Practice, Debell and Richardson, Springfield, MA, 1946. 8. Mark, H. and Wulff, C., German Patent DRP 550055 (to I. G. Farben Industrie) (1929). 9. Meister, B. J. and Malanga, M. T., 'Styrene Polymers', in Encyclopedia of Polymer Science and Engineering, ed. Moore, E. R.), Wiley, New York, vol. 16, p. 21 (1989). 10. Munters, C. G. and Tandberg, J. G., US Patent 2023204 (1935). 11. Frisch, K. C., 'History of Science and Technology of Polymeric Foams', in 'History of Polymer Science and Technology', eds. Seymour, R. B., Marcel Dekker, New York (1982). 12. Stastney, F. and Goeth, R., US Patent 2681 321 (to BASF) (1956). 13. Stastney, F. and Buchholz, K., US Patent 2744291 (to BASF) (1956). 14. Stastney, F., US Patent 2787809 (to BASF) (1957). 15. Ostromislensky, J. J., US Patent 1 613673 (1927). 16. Haward, R. N. and Elly, J., US Patent 2668806 (1954). 17. Stein, A. and Walter, R. L., US Patent 2862909 (to Monsanto) (1958). 18. Amos, J. L., Mclntire, O. R. and McCurdy, J. L., US Patent 2694692 (to Dow) (1954); Amos, J. L., Polym. Eng. Sci., 14, 1 (1974). 19. Daly, L. E., US Patent 2435202 (to US Rubber Co.) (1942). 20. Pavelich, W. A., 'A Path to ABS Thermoplastics', in High Performance Polymers: Their Origin and Development, ed. Seymour, R. B. and Kirshenbaum, G. S., Elsevier, New York, p. 125 (1986). 21. Calvert, W. C., US Patent 2908661 (to Borg-Warner) (1959). 22. Calvert, W. C., US Patent 3238275 (to Borg-Warner) (1966). 23. Herbig, J. A. and Salyer, I. O., US Patent 3 118855 (to Monsanto) (1964). 24. Otto, H.-W., German Patent DE 1 182811 (to BASF) (1965). 25. Siebel, H. P. and Otto, H.-W., German Patent DE 1 238207 (to BASF) (1967). 26. Wagner-Juaregg, T., Chem. Ber., 63, 3213 (1930). 27. Fordyce, R. G., Chapin, E. C. and Ham, G. E., J. Am. Chem. Soc., 10, 2489 (1948). 28. Seymour, R. B., 'Styrene-Maleic Anhydride-Vinyl Monomer Terpolymers and Blends', in High Performance Polymers: Their Origin and Development, ed. Seymour, R. B. and Kirshenbaum, G. S., Elsevier, New York, p. 125 (1986). 29. Swarc, M., Levy, M. and Milkovich, R., J. Am. Chem. Soc., 78, 2656 (1956). 30. Swarc, M., Nature (London), 178, 1168 (1956). 31. Szwarc, M.,J. Polym. Sci., Part A, Polym. Chem., 36, ix (1998). 32. Ishihara, N., Kuramoto, M. and Uoi, M., Japanese Patent JP 62 187708 (to Idemitsu Kosan Company) (1985). 33. Ishihara, N., Kuramoto, M. and Uoi, M., European Patent JP 210615 (to Idemitsu Kosan Company) (1986). 34. Ishihara, N., Kuramoto, M. and Uoi, M., Macromolecules, 21, 2464 (1986).

Polystyrenes and Styrene Copolymers - An Overview NORBERT NIESSNER AND HERMANN GAUSEPOHL BASF AG, Ludwigshafen, Germany

1

INTRODUCTION

Polystyrene is one of the most widely used thermoplastic materials ranking behind polyolefins and PVC. Owing to their special property profile, styrene polymers are placed between commodity and speciality polymers. Since its commercial introduction in the 1930s until the present day, polystyrene has been subjected to numerous improvements. The main development directions were aimed at copolymerization of styrene with polar comonomers such as acrylonitrile, (meth)acrylates or maleic anhydride, at impact modification with different rubbers or styrene-butadiene block copolymers and at blending with other polymers such as polyphenylene ether (PPE) or polyolefins. Polystyrene ('general-purpose polystyrene', GPPS) can be understood as a linear polyethylene chain with laterally attached phenyl rings, being responsible for the enhanced glass transition temperature and high refractive index. Stiffness, brilliance, gloss and hardness are the main characteristics of this material. Consequently, applications such as audio/video cassette packs, beakers, transparent food packagings, shower cabinets, lamp covers, etc., are dominant. To overcome the brittleness of GPPS, the material was modified by incorporation of polybutadiene. Impact-modified polystyrene (IPS) was invented by Ostromislensky [1] and has been commercialized since the 1950s. IPS consists of a polystyrene matrix with embedded cellular rubber particles. By rubber

Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers, Edited by J. Scheirs and D. B. Priddy © 2003 John Wiley & Sons Ltd

26

N. NIESSNER AND H. GAUSEPOHL

toughening, however, transparency is lost. The use of styrene—butadiene block copolymers instead of polybutadiene results in translucent impact polystyrene with a core/shell particle morphology. These particles are small and thus reduce scattering of visible light. Enhanced property demands in the packaging sector and also in the electric/ electronic and automotive sectors require improved product properties. Homogeneously miscible blends with, e.g., polyphenylene ether (PPE) combine the excellent processability of the amorphous polystyrene with the thermal stabilty of its blend partners. Styrenic copolymers are materials capable of thermoplastic processing which, in addition to styrene (S), also contain at least one other monomer in the main polymer chain. Styrene—acrylonitrile (SAN) copolymers are the most important representative and basic building blocks of the entire class of products. By adding rubbers to SAN either ABS (acrylonitrile-butadiene-styrene) or ASA (acrylate—styrene—acrylonitrile) polymers are obtained depending on the type of rubber component employed. These two classes of products yield blends composed of ASA and polycarbonate (ASA + PC) or ABS and polyamide (ABS + PA). MABS polymers (methyl methacrylate—acrylonitrile—butadiene—styrene) together with blends composed of polyphenylene ether and impact-resistant polystyrene (PPE/PS-I) also form part of the styrenic copolymer product range. Figure 2.1 provides an overview of the different classes of products and trade names. A characteristic property is their amorphous nature, i.e. high dimensional stability and largely constant mechanical properties to just below the glass transition temperature, Tg.

I

Transparent base polymer | I

I

ps | I

*-v

Addition of rubber 1

• • ' Impact-modified polymer |m «l I • I

Blend PPE/PS-I

ABS

Figure 2.1 Overview of the different classes of styrene polymers

27

POLYSTYRENES AND STYRENE COPOLYMERS

2

POLYMERIZATION

Styrene is one of the few monomers able to be polymerized under free radical, anionic, cationic and metal catalysed conditions. This is due to low polarity of the styrene molecule and to the resonance stabilization of the growing polystyryl species in the transition state. According to Mayo, [2,3] it is assumed that styrene forms a Diels—Alder adduct which isomerizes to phenyltetralin (a) or transfers a hydrogen atom to a further styrene molecule forming two radicals in a solvent cage. These radicals are stabilized by disproportionation (b, c) or recombination (d) to cyclic dimers and trimers. Diffusion of the radicals from the cage and subsequent polymerization to polystyrene is in fact only a side reaction in this scheme (Figure 2.2). Recently, strong evidence in support of this mechanism was obtained by Buzanowski et al. [4]. In addition to the thermal initiation, the use of peroxides or azo components is a common and well established method to start the chain reaction. Peroxides increase the rate of the polymerization process and improve the grafting efficiency in the case of IPS. More recently, multifunctional peroxides have also been used in order to obtain products with special molecular weight distributions. Up to relatively high conversions, the rate of polymerization can be satisfactorily represented as a first-order reaction [5] according to the following equation: -d[M]/df = rp = *p[P'][M] where rp= polymerization rate, kp= rate constant, [P*] = stationary radical concentration, and [M] = monomer concentration. This equation is valid under the assumption that the number of polymerizing monomer units is much larger than the monomer consumption which is needed for the initiation. A set of kp

+s

PS

Figure 2.2 Reactions during styrene oligomer formation in the chain initiation phase

28

N. NIESSNER AND H. GAUSEPOHL

values is listed in Ref. 6. A state-of-the-art method is pulsed laser initiated polymerization [7,8]. The polymerization reaction is terminated by disproportionation and recombination, with the ratio depending mainly on the temperature. Chain termination is represented by

where kt is the termination coefficient, values also being listed in Ref. 6. The mode of termination determines the dispersion index D: D = Pw/Pn = 2 for disproportionation and

for recombination, where Pw, and Pn are the weight- and number-average degree of polymerization, respectively. In addition to termination, transfer reactions to all other components of the polymerization occur. These transfer reactions do not affect the polymerization rate. However, they diminish the average molecular weight. The equation for transfer reactions is Rtr = Ktr[P*][TH]

where kp = transfer rate constant, and [TH] = concentration of transfer agent. The main disadvantage of radical polymerization reactions is their low selectivity, i.e. neither the molecular weight, the molecular weight distribution nor the molecular structure can be precisely controlled. In the literature, many experiments have been described to overcome this drawback by applying so called iniferters (initiator—transfer agent—termination). These substances act as initiators, transfer agents and termination agents. Typical examples for controllers are nitroxyl compounds such as 2,2,6,6-tetramethyl-l-piperidinyl oxide (TEMPO) and 2,2, 5,5-tetramethyl-l-pyrrolidinyl oxide (PROXYL) [9,10], thiocarbamate [11] and tetraphenylethane derivatives [12]. All these substances react with the growing macromolecular radicals by forming temporarily dormant species, minimizing termination by recombination or disproportionation [13,14]. The introduction of a dynamic equilibrium between dormant and active species leads to a low stationary concentration of free radicals and a relatively high concentration of reactive polymer chains. This results in a low overall polymerization rate. By this method, ordered copolymers from styrene and butadiene were synthesized with only limited success [15] owing to the formation of homopolymers.

POLYSTYRENES AND STYRENE COPOLYMERS

3

29

PROCESSES

Polystyrene was first commercially produced by BASF in 1931. The inventors of this mass polymerization process had to resolve two conflicting problems: heat removal from high-viscosity melts and development of an appropriate workup scheme. Heat removal was accomplished by heat-exchange tubes in the polymer melt. Polystyrene was obtained after exit of the melt from a hightemperature polymerization zone, using an extrusion screw. Today's processes are characterized by a fully continuous polymerization with heat removal by evaporation of styrene and solvent. The main advantage - consistent temperature — results in a high product quality with a narrow molecular weight distribution (Mw/Mn = 2.2–2.4) and high transparency. Impact-modified polystyrene is mainly produced by mass polymerization, either in tower cascades or tank/tower cascades. In the latter case, particle size and morphology can be defined by variation of the viscosity ratio between the continuous and the discontinuous phases, the stirrer velocity, the molecular weight of the polybutadiene rubber and the amount of rubber. Typical particles sizes are 2–20 u,m, this being the optimum for effectively dissipating impact energy. Styrene copolymers such as SAN and ABS are basically produced according to GPPS and IPS technology. However, effective impact modification in ABS is accomplished with smaller particles. Particles around or smaller than 1 fxm can also be produced by emulsion technology and thus emulsion polymerized ABS dominates by far in today's global ABS markets. In order to guarantee optimum impact modification, the rubber must be adapted to the base polymer (matrix). This is done by grafting with monomers which are miscible with the matrix. Thus, for example, the polybutadiene rubber for ABS is grafted with a mixture of styrene and acrylonitrile. The graft rubbers produced in this way consist of a flexible rubber core surrounded by a graft shell which provides linkages to the matrix in question (Figure 2.3). This process is performed in emulsion. The particles of rubber are finely dispersed in the rigid phase, i.e. the impact-modified styrenic copolymers thus have a multiphase structure. Figure 2.4 gives a brief overview of mass and emulsion polymerization processes.

4 STRUCTURE AND MORPHOLOGIES Commercial polystyrene manufacturing techniques are based either on a suspension process if the material is to be foamed or on a bulk polymerization process for GPPS and IPS. ABS-type polymers can also be produced via emulsion polymerization. Figure 2.5 shows the differences in emulsion and mass polymerization processes and the resulting morphology. Typically,

30

N. NIESSNER AND H. GAUSEPOHL

Elastomer core (rubber) Graft envelope Matrix Modification of properties due to • the chemical nature and quantity of the rubber • the structure morphology of the rubber • the particle size and particle size distribution

Figure 2.3 particles

Typical impact modification of styrene copolymer via emulsion core/shell

Mass ABS

Emulsion ABS Butadiene

H,O

S/AN Initiator

S/AN

Degassing 1r

H,O

Precipitation /Drying

Extruder

h

(residual

ABS

Features

Features Graft shell can be produced in a controlled manner independently from matrix polymer

•

High rubber content possible

•

High rubber efficiency

Monodisperse particles with defined particle size

•

Special HI PS-like morphology

Figure 2.4

Only few process steps

• Light color

Differences between mass and emulsion polymerization processes

POLYSTYRENES AND STYRENE COPOLYMERS Emulsion Polymerization Glossy surface

Figure 2.5

31

Mass Polymerization Lower Gloss Light and constant color

Differences between mass and emulsion polymerization: morphology

particles with diameters in the range of the wavelength of visible light can be produced by emulsion polymerization. As a consequence, glossy products with high-quality surface appearance are based on this technique. Larger particles often with inclusions of matrix polymer - result from mass polymerization. These morphologies typically cause a less glossy surface appearance, and are characterized by a high rubber efficiency, i.e. impact strength per unit amount of rubber. Rubber particles dissipate impact energy only if they can effectively initiate and terminate 'crazes'. By this 'crazing', energy is transformed into deformation of rubber particles, eventually accompanied by the formation of voids in the rubber particle itself. Deformation, however, initiates crazes, that can effectively be stopped by other rubber particles (Figure 2.6). A typical electromicrograph shows the formation of these micro-cracks after absorption of impact energy, and further magnification reveals the high inner surface caused by the formation of fibrils perpendicular of the craze direction (Figure 2.7). The predominant fracture mechanism in polymers with low polarity (such as IPS or ABS with low acrylonitrile content) is the craze mechanism (with an optimum particle size of approximately 2–6 |xm). With increasing polarity, the dissipation of energy by formation of shear bands ('shear yielding') becomes more pronounced. This second mechanism is facilitated by very small particles in the sub–100 nm range. As a rule of thumb, it is accepted that - starting with IPS an increase in the acrylonitrile content in rubber-modified styrene polymers causes a shift towards lower optimal particle size. For this reason, mass polymerized ABS typically has significantly smaller cellular particles than mass polymerized IPS. Nevertheless, anionic polymerization will be the pacemaking tool for the synthesis of innovative materials which shift the properties of the polystyrene

32

Figure 2.6

N. NIESSNER AND H. GAUSEPOHL

Formation of crazes after mechanical impact

Figure 2.7 TEM picture of crazes in ABS

family into the domain of thermoplastic elastomers and engineering resins [16]. The reason for this is that only anionic polymerization permits the control of the molecular architecture of the molecules, allowing the preparation of tailormade polymeric materials. Figure 2.8 gives an overview on the variety of the structural possibilities.

33

POLYSTYRENES AND STYRENE COPOLYMERS molecular composition 2-block

to

block and tapered block transition

multiblock D-M-D-i-0 D-1-D-l-D-i statistical to alternating monomer insertion

molecular structure linear

branched

star

comb

functionality monofunctionality

bifunctionality

molecular weight distribution mono-modal mono-modal narrow distribution broad distribution Figure 2.8

multimodal

Control of the molecular design by anionic polymerization

The molecular weight can be controlled by the ratio of the initator to the monomer, the molecular weight distribution by the type of polymerization (discontinuous or continuous) and the modality by single or multiple initiation. Sequential addition of different monomers leads to block copolymers with sharp or tapered transitions. In the presence of Lewis bases, statistical or alternating copolymers can be obtained. The molecular structure can be selected between linear, branched, star- and comb-like or dendritic molecules. Finally, functionalities can be introduced for changing the polarity of the resulting product or for extending the reaction profile. Styrene—butadiene block copolymers such as Kraton™, K-Resin™, Styrolux® and Styroflex® are examples of the versatility of this method and have long been manufactured via anionic polymerization on a commercial scale.

34

N. NIESSNER AND H. GAUSEPOHL

Most recently, block copolymers of the type ABC with completely new and previously unknown morphologies were synthesized by Stadler et al. [16]. By molecular self-assembly he obtained products with supramolecular structures and new property profiles, e.g. the block copolymer with 24% polystyrene, 7% polybutadiene and 69% poly(methyl methacrylate) (PMMA) exhibits higher toughness and higher tensile strength than PMMA. TEM pictures reveal polystyrene cylinders surrounded by a helical string of polybutadiene and embedded in a PMMA matrix (Figure 2.9a). This type of block copolymer but with a higher proportion of polybutadiene is very suitable also as elastomers because these products exhibit high tensile strength compared with classical ABA systems (Figure 2.9b). In order to upgrade polystyrene towards an engineering plastic material such as ABS, the long-term service temperature needs to be raised above 100 CC. Consequently, the polymer chain must be stiffened by radical copolymerization of styrene with a variety of ring-substituted styrenes. However, in these cases the influence on the glass transition temperature of the polymer is not very strong, but by using anionic polymerization technology a new polymer class with 1,1-diphenylethylene as comonomer could be introduced. It exhibits an overall advanced performance compared with GPPS without sacrificing the typical polystyrene property profile. The incorporation of additional phenyl rings into the polystyrene backbone results in a stiffened polymer chain. Conventional, butyllithium-initiated polymerization faces a major drawback. Owing to the living nature of anionic polymerization, all chains grow at the same time. As a consequence, turnover rates are exceedingly fast at high monomer concentrations, and removal of the heat of polymerization is difficult. Hence anionic polymerization has been restricted to dilute solutions and low temperatures. Under these conditions, polystyrene cannot be produced economically. However, with certain Lewis acids such as dibutylmagnesium this problem could be solved [18,19].

5

PROPERTIES

Commercial polystyrene is an amorphous material with a molecular weight Mw = 100000–400000. Low specific gravity, transparency and brilliance, absence of colour, low shrinkage and ease of fabrication are its most characteristic features. At temperatures sufficiently below the glass transition temperature (Tg) and at low deformations the material obeys Hooke's law of elasticity under external stress. The tensile strength and flexural strength are strongly dependent on the molecular weight of the polymer. Below a certain molecular weight (Mw « 120000) the material fails completely, whilst at higher values the tensile strength and the elongation at break increase until an upper limit is reached. At

POLYSTYRENES AND STYRENE COPOLYMERS

Figure 2.9

35

Morphology of special ABC block copolymer. Courtesy of BASF

higher impact and deformation rates, polystyrene tries to withstand the external stress by forming crazes, thus dissipating the impact energy over a broader area of the specimen. Crazes are voids which are crossed by orientated fibrilles.

36

N. NIESSNER AND H. GAUSEPOHL

They are the precursors of cracks and mainly determine the tensile strength of the material. The reason for the formation of crazes lies in the insufficient segment mobility of the molecular chain. Rubber-modified polystyrene exhibits a much higher toughness than crystal polystyrene because small dispersed rubber particles enhance the stress concentration and the larger ones stop their growth, thus preventing crazes from developing into cracks. The prerequisite for this mechanism is a well defined adhesion of the rubber particles to the matrix, which in turn depends on the grafting efficiency of the rubber [20,21]. Table 2.1 shows some important properties of polystyrene. Above its Tg, polystyrene is a viscoelastic melt. It is called viscoelastic because the polymeric material displays both a viscous and an elastic response to shear stress, depending on the rate and the temperature of the test. Two main factors influence the viscous and the elastic behaviour of the product, namely its molecular weight and the molecular weight distribution. For mondisperse polystyrenes the zero-state viscosity is proportional to M3.4. The ratio Mw/Mn has only a minor influence on the viscosity, but it greatly influences the steadystate shear compliance, i.e. small amounts of high molecular weight polystyrene blended into a medium molecular weight polystyrene enhance the elasticity and melt strength. Details concerning the rheology of polystyrene are given elsewhere [22,23]. The development of SAN was triggered by the idea of building a polar comonomer into polystyrene to improve its resistance to chemicals and to stress cracking. The relatively polar acrylonitrile presented itself as a suitable comonomer in this case. Styrene—acrylonitrile copolymers are further characterized by high rigidity and thermal shock resistance. Two parameters substantially determine the properties of SAN: molecular weight and the proportion of acrylonitrile. On account of its combination of properties, SAN is principally employed in a wide range of household articles. Other examples include transparent parts for kitchen appliances and coffee machines. In the field of household goods, its high stability in dishwasher conditions is the principal argument in favor of SAN. Other areas of application are packaging for cosmetics, toothbrushes and lamp covers. Relatively large industrial parts such as large industrial batteries (Figure 2.10) are a remarkable field of application, since the housings weigh up to 14kg. Fitted with lead electrodes and filled with sulfuric acid they have a service life of more than 10 years, which is a highly impressive testimony to SAN's durability and resistance to chemicals.

Table 2.1

Properties of atactic polystyrene

Property

Unit

Test method

Value (GPPS)a

Value (IPS)b

Density Refraction index Coefficient of expansion:

g/cm3 —

ISO 1183 ISO 489

1.050 1.590

1.050 —

cm3 /cm3 K cm3 /cm3 K

— — ISO 75-2

2.805 x 10–4 5.650 x 10–4 100 86

— — 73-85

ISO 75-2

98

81-96

W/mK — — Om MPa MPa %

ISO 306 DIN 52 612 IEC 60 250 IEC 60 250 IEC 6093 ISO 527-2 ISO 527-2 ISO 527-2

101 0.17 2.5 / 2.5 0.9 x 10–4/0.7 x 10–4 >1014 3300 42-59 1.5-3.0

91-101 0.17 2.5/2.5 1.0 x 10 - 4 /4x 10-4 >1014 1400–2800 21-43 1.4–2.0

kJ/m2 kJ/m2

ISO 179/leU ISO 179/leU

10-28

70-NBC 50–160

Tg

Glass transission temperature (Tg) Heat deflection temperature under 1.8MPa load(HDT/A) Heat deflection temperature under 0.45 MPa load (HDT/B) Vicat softening point A50 Thermal conductivity Dielectric constant at 100 Hz/1 MHz Dissipation factor at 100Hz/l MHz Volume resistivity Tensile modulus of elasticity Tensile stress at yield Elongation at yield Charpy impact strength: At +23 °C At-30°C GPPS: general-purpose polystyrene. IPS: impact polystyrene. NB: no break.

°c °c °c O/"~l

2

-4

38

N. NIESSNER AND H. GAUSEPOHL

Figure 2.10

6

Industry batteries made from SAN

PROPERTIES, RANGE AND APPLICATIONS OF MABS PRODUCTS

While amorphous, single-phase materials such as SAN or PMMA are transparent, rubber-modified, multiphase materials are opaque as a general rule. The reason for the loss of transparency is the scattering of light by the particles of rubber dispersed in the amorphous matrix. However, as with every rule there

POLYSTYRENES AND STYRENE COPOLYMERS

39

are also exceptions here. In the case of MABS the following concept is used: the rubber for MABS is built up in such a way that it has exactly the same refractive index as the matrix. In this way the rubber is rendered 'invisible', the light is not scattered and the material remains transparent (Figure 2.11). MABS is being increasingly used also for transparent parts for computer housings and household appliances (Figure 2.12), offering a high degree of freedom to accomplish even the most sophisticated designs. Graft rubbers based on polybutadiene (ABS) or acrylate esters (ASA) are generally used for the impact modification of styrene copolymers. Depending on the rubber component used, special features emerge in the final properties (Figure 2.13). Polybutadiene yields very high levels of toughness even at low temperatures but it contains numerous double bonds which can be relatively easily attacked by UV radiation and oxygen. On exposure to outdoor conditions or under the action of heat this attack results in yellowing and embrittlement of ABS even after a short time. Acrylate rubbers which are employed in ASA contain no double bonds. For that reason, ASA is substantially more resistant to weathering than ABS. Owing to the polar acrylate component, ASA is also more resistant to stress cracking than ABS. ABS in turn has significant advantages in low-temperature impact resistance on account of the very low glass transition temperature of the polybutadiene rubber.

Figure 2.11 Transparent impact-modified styrene copolymers by isorefractive phases

40

N. NIESSNER AND H. GAUSEPOHL

Figure 2.12 Transparent plastics (including styrene copolymers) for state-of-the-art household equipment

Figure 2.13

Differences between ASA and ABS polymers

POLYSTYRENES AND STYRENE COPOLYMERS

41

Figure 2.14 Electomicrographs of a cross-section perpendicular to the surface of ABS and ASA after 500 h of UV exposure (Xenotest, ISO 4892-2A) and subsequent treatment with soap

On account of its acrylate rubber content, ASA is superior to ABS with regard to its weathering resistance and thermostability. In outdoor applications ASA yellows to a significantly lower degree and also retains its impact resistance over a substantially longer period of time. The tendency to yellowing in both ABS and ASA can be reduced further by the addition of UV stabilizers. However, even stabilized ABS does not reach the level of unstabilized ASA (Figure 2.14).

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Ostromislensky, I., US Patent 1613673 (1927). Mayo, F. R., J. Am. Chem. Soc., 15 (1953) 6133. Hiatt, P. D. and Bartelt, J., J. Am. Chem. Soc., 81 (1959) 1149. Buzanowski, W. C, Graham, J. D., Priddy, D. B. and Shero, E., Polymer, 33 (1992) 3055. Husain, A. and Hamielec, A. E., J. Appl. Polym. Sci, 22 (1978) 1207. Berger, K. C. and Meyerhoff, G. in Polymer Handbook, Vol. 3, Wiley, New York (1989). Olaj, O. F., Schnoll-Bitaj, I. and Hinkelmann, F., Makromol. Chem., 188 (1987) 1689. Deady, M., Mau, A. W. H., Moad, G. and Spurling, T. H., Makromol. Chem., 194 (1993) 1691. Georges, M. K., Veregin, R. P. N., Kazmaier, P. M. and Hamer, G. K., Macromolecules, 26(1993)2987. Brinkmann-Rengel, S. and Niessner, N., ACS Symp. Ser., 768 (2000) 394. Otsu, T., Yamashita, K. and Tsuda, K., Macromolecules, 19 (1987) 287.

42

N. NIESSNER AND H. GAUSEPOHL

12. 13. 14. 15.

Bledzki, A. and Braun, D., Polym. Bull., 16 (1986) 19. Webster, O. E., Science, 251 (1991) 887. Gretza, M. K., Madereu, D. and Matyjaszewski, K., Macromolecules, 27 (1994) 638. Georges, M. K., Veregin, R. P. N., Kazmaier, P. M. and Hamer, G. K., Polym. Prepr., 35 (1994) 582. Knoll, K. and Niessner, N., ACS Symp. Ser., 696 (1998) 112. Karppe, U., Stadler, R. and Voigt-Martin, I., Macromolecules, 28 (1995) 4558. Desbois, P., Fontanille, M., Deffieux, A., Warzelhan, V., Latsch, C. and Schade, C. H., Macromol. Chem. Phys., 200 (1999) 621. Ebara, K., Tanji, S. and Sawamoto, M., Patent WO/97 33923 (1997). Bucknall, C. B., Toughned Plastics, Applied Science Publishers, London (1977). Kramer, E. J. and Krauch, H. H., Crazing in Polymers, Springer, Berlin (1983). Ferry, J.D., Viscoelastic Properties of Polymers, Wiley, New York, 3rd edn, 1980. Carreau, P. J., Rheology of Polymeric Systems, Hanser, Munich (1997).

16. 17. 18. 19. 20. 21. 22. 23.

Preparation of Styrenic Polymers

This page intentionally left blank

BERNARD J. MEISTER AND CLARK J. CUMMINGS The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Styrene is one of the easiest monomers to polymerize and of course this led to the early discovery and commercialization of polystyrene. Solid polystyrene was probably first prepared in 1845 [1] by heating the monomer in air. At first the reaction was considered to be an oxidation until Staudinger [2] first proposed the long-chain structure of polymer molecules. This proposal and a host of ensuing studies to understand the chemistry of polymerization led to commercial interest in producing this new type of material. The Dow Chemical Company had a major effort in development of a commercial process for polystyrene in the 1930s leading to the can process documented by Boyer [3]. In the same time frame, I. G. Farben in Germany developed the well known continuous tower process [4]. When the details of the I. G. Farben process became known to the US chemical companies after World War II, effort was placed on developing continuous processes due to the much lower labor costs and improved uniformity of the product. The reactors for these continuous processes took many forms, from the initial towers and tube tanks to agitated versions to stirred tank reactors with ebullient cooling, pipelines filled with static mixers and recirculated tube banks. Over the same time period, batch stirred tank reactors utilizing solution, suspension and emulsion polymerization were also developed. Because most of these processes did not go to complete conversion, devolatilization processes were developed that continue to evolve because of taste, odor and health related concerns with residual monomers and solvents. For the practitioner, the choice among all Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy >( 2003 John Wiley & Sons Ltd

46

B. J. MEISTER AND C. J. CUMMINGS

these possibilities depend somewhat on the technical base of the individual company (knowledge, patents and installed capital) and on the competing desires of product property range, product quality and uniformity, low manufacturing cost including maintenance and any copolymers or composites such as rubber-modified polymers that are included in the product mix. There are a number of prior works that tackle this same general subject [5–10]. The most important of these is the detailed review by Simon and Chappelear [5]-

2 TECHNICAL CONSTRAINTS THAT INFLUENCE REACTOR SELECTION 2.1

TEMPERATURE CONTROL

The first constraint that the engineer faces when the task is to develop a commercial process, is that a large amount of heat is generated when one converts styrene to polystyrene. The heat of polymerization is approximately 300 BTU/lb (700 J/g) at 100 °C and decreases with increasing temperature. This leads to a temperature increase of approximately 350°C for pure styrene if it is polymerized to completion and no heat is removed from the process. This alone does not rule out an adiabatic process; however, polystyrene degrades rapidly at temperatures above 250 °C and the molecular weight of the polystyrene produced decreases rapidly with temperature (see Figure 3.1). Therefore, no practical products could be produced with an adiabatic process without at least 50 % diluent to limit the temperature rise to 175 °C. The chemist in the laboratory does not have to deal with the heat generation issue. Ten grams of styrene are sealed in an ampoule and placed in a bath at the reaction temperature and the heat transfer is sufficient to maintain the temperature of the reactants under most conditions. The Dow can process [3] mentioned in the Introduction was a direct scaleup of the chemists ampoule to 10 gallon cans. There was a considerable rise in temperature in the middle of the 10 gallon can leading to high molecular weight on the outside and low molecular weight on the inside that was then ground up and blended. This lack of temperature control severely limits the product mix. A natural evolution from the batch bulk process was to the batch suspension process where water as the continuous phase supplies a large heat sink that allows control of the temperature of the polymerizing mass, yielding control of the molecular weight and the molecular weight distribution of the polystyrene. The alternative path of evolution was to the continuous solution process, first demonstrated with the tower process by I. G. Farben and implemented by Dow and others as either towers or tanks filled with heat transfer tubes. These

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE 47 50

20 10

0.50

0.20 0.15 2X10 6 1X10 6 5X10 5 2X10 5 1Xl05 Number-average molecular weight

Figure 3.1 Relationship between initial rate (wt%/h) of styrene polymerization and molecular weight at various temperatures. Reproduced with permission from Encyclopedia of Polymer Science and Engineering, Mark (Ed.), John Wiley & Sons, NY. Copyright John Wiley & Sons

towers or tanks were filled with as many heat transfer tubes as possible but they were severely limited in the rate of polymerization that they could control. The tube tanks that were used by Dow were limited to conversion rates of less than 5 %/h and therefore about 15 h of residence time. In order to produce 4000 kg/h of product at 75% conversion, 80000kg of reactor inventory were required. This large amount of inventory leads to slow product changeovers and large amounts of off-specification material. In addition, high solids material would collect on the tubes, leading to decreasing heat transfer with time so that periodic cleanouts were required. Because of the rate limitations of the tower and tube-tank processes that were primarily heat transfer constraints, further developments in the continuous solution process for crystal polystyrene (GP) were aimed at improving heat transfer. One obvious solution was to incorporate agitation of some type in the reactor. Although at Dow the incorporation of agitation in the reactors came about with the development of rubber-modified polystyrene [11], and this aspect will be discussed in a later section, agitation also significantly raises the heat transfer

48

B. J. MEISTER AND C. J. CUMMINGS

coefficient and allows a threefold increase in polymerization rate. This is extremely important in the economics of the process because the in-process inventory and the offgrade during transitions can be reduced by a factor of three. This also means less time in the finishing lines for any degradation to take place. The second major approach to improving heat transfer in the continuous solution process is boiling heat transfer. The first report of the utilization of an ebullient stirred tank reactor to produce polystyrene was a patent assigned to Union Carbide in 1950 [12]. The implications of a well mixed reactor on the product will be discussed in the section on mixing. An alternative approach to the utilization of boiling heat transfer is embodied in a patent assigned to Monsanto [13]. In this approach, the reactor is horizontal with reactor internals that segregate the reactor into a series of chambers that have a common vapor space. As the boiling point of a polystyrene solution increases with the concentration of polystyrene, the temperature in the chambers increases with conversion. This allows a temperature profile that is not possible in the single well mixed reactor. The use of boiling heat transfer raises the maximum conversion rates that can be controlled significantly beyond that of the agitated towers filled with heat transfer tubes. The main limitation that occurs is the removal of the vapor bubbles from the polymer solution. As the viscosity of the polystyrene solution increases rapidly with conversion, this becomes most limiting when the viscosity of the polystyrene solution exceeds 1000P (dPa/s). Below this viscosity, conversion rates of 40%/h can be controlled, but above this viscosity, the polymer mass foams up into the condenser and temperature control is lost, so that the maximum conversion rate decreases rapidly at high polymer concentrations.

2.2

CHEMISTRY-RELATED CONSTRAINTS

Thus far, all of the discussion has been related to the production of polystyrene that is initiated thermally. The main reactions that one must be concerned with that determine the rate of reaction and the molecular weight of the polymer chains that are produced are initiation, propagation, termination and chain transfer. The thermal initiation reaction has a high activation energy with the number of chains initiated per unit time increasing by a factor of seven on raising the reaction temperature from 120 to 140CC. The rate of reaction increases roughly by a factor of four so that the average chain molecular weight decreases by a factor of 1.75 over this same temperature range. The engineering consequence of this is that low molecular weight polystyrene can be produced at a conversion rate much greater than the high molecular weight polystyrene. If the same reactor is to be used to produce a molding resin of molecular weight 150000 and a high-grade extrusion resin of molecular weight 350000, then the

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

49

molding resin could be produced at 140 °C at a rate of 40%/h and the extrusion resin could be produced at 110 °C at a rate of only 5%/h. Fortunately, this is about the full range of commercial products. However, it leaves the design engineer specifying the size of equipment with a difficult choice. This is what was meant when we titled this section 'chemistry-related constraints' as the reactor design is heavily influenced by inherent chemistry and the product mix. One possible solution to the product mix problem is to design the production unit to operate at the product rates of the highest molecular weight product and then use a chain transfer agent such as a thiol or high levels of a solvent such as ethylbenzene to produce the low molecular weight product. At this point we also need to mention the molecular weight distribution of the product. The lifetime of a growing polystyrene chain is fairly short. This can be calculated from T = DP/kp(M] where kp is the propagation rate constant (1420L/mol s at 140°C), DP is the number-average degree of polymerization (1200 at 140°C) and [M] is the monomer concentration in mol/L(7.68 mol/L at 140°C) . Therefore, at 140°C the lifetime of the growing chain is only 0.1 1 s. If a reactor in batch is operated at a time-temperature profile or a series of continuous reactors are operated at a temperature sequence, some chains will be produced at each temperature and a distribution of molecular weights will be obtained. The normal method of representing this distribution is to use the ratio of the weight-average molecular weight, Mw, to the number-average molecular weight, Mn, where

and

where ni is the number of chains of molecular weight Mi. Even at a fixed temperature, a distribution of molecular weights is produced because of the random nature of the chain termination process. If the chains all terminate by chain transfer, the ratio Mw/Mn is approximately 2.0, and if they all terminate by coupling, the ratio Mw/Mn is approximately 1.50. Under conditions where thermal polystyrene is produced, the ratio is approximately 1.90 because most chains are terminated by chain transfer to monomer, solvent and other agents. The use of the particular product may dictate an optimum molecular weight distribution. If the product is used to make an extruded foam, a relatively narrow distribution may be desirable to retain orientation and strength in the

50

B. J. MEISTER AND C. J. CUMMINGS

cell walls. However, a molding product might require a broad distribution because of the shear sensitivity and easier processing that results. The main point here being that the ability to adjust the molecular weight distribution is a desirable if not essential feature of a commercial reactor to produce polystyrene. This generally means the capability to react at sequentially higher temperatures or to add fast-acting chain transfer agents during the process is an important feature. One method commonly utilized commercially to surmount the chemistry constraints of thermally initiated polystyrene is the use of free radical initiators of either the peroxide or azo type. Initiators that have 1 h half-lives slightly below the thermal temperatures are most effective so that initiators with 1 h half-lives in the range 100–140°C are most often used. The utilization of chemical initiators under conditions where the thermal initiation reaction is suppressed generally leads to a boost in the rate at which a given molecular weight polystyrene can be produced. The main reason for this is that the Diels– Alder reactive dimer formed in the thermal initiation step is a chain transfer agent and plays a significant role in controlling the molecular weight. When the thermal initiation reaction is suppressed by lowering temperature or monomer concentration, but the radical concentration is maintained with an initiator, the resulting polystyrene has a higher molecular weight. During the 1980s, bifunctional initiators became widely used among the commercial producers of polystyrene. The reason for this is a larger boost in the rate-molecular weight relationship because a difunctional initiator with two peroxide groups in the same molecule can lead to two chains of average length DP and one chain of average length 2DP. Figure 3.2, taken from the review by Priddy [14], shows an illustration of the significance of this effect for a particular case. Each individual initiator will have somewhat different curves based on the half-lives and efficiency of generating free radicals of the specific monofunctional or bifunctional peroxide. Certainly, one can think of possible methods to extend this approach such as polymeric or cyclic peroxides that yield only diradicals or multifunctional peroxides that yield branched polystyrene. We do not know if any of these concepts are utilized commercially. There are other ways of chemically shifting the rate-molecular weight relationship. One of these is to add a small amount of divinylbenzene to the styrene, which leads to the incorporation of branch points in the polymer chain. This works well in a well mixed batch reactor in the laboratory but often leads to gels and crosslinked polymer in large-scale continuous reactors. A second approach investigated by Priddy [14] is the use of an acid to suppress thermal initiation. This results in a shift in the ratio of propagation to initiation that results in increased molecular weight. This also has not made it to the commercial realm. The final method of shifting free radical kinetics is emulsion polymerization. Unlike suspension polymerization where the reaction kinetics are identical with those of mass polymerization, in emulsion polymerization the droplets are small

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

51

350

300

Spontaneous Monofunctional ' initiator

250

Difunctional initiator

200 10

20 30 40 50 Polymerization rate, %/h

60

Figure 3.2 Polymerization rate advantage of difunctional initiator. Reproduced with permission from Encyclopedia of Chemical Technology, Kirk (Ed.), John Wiley & Sons, NY (1997). Copyright John Wiley & Sons

enough so that only one polymer chain is growing within the droplet at a time. This essentially eliminates the termination mechanisms that involve two active radicals and therefore allows the production of higher molecular weight polystyrene at a given temperature. Back in the 1960s this was a competitive way to produce high molecular weight crystal polystyrene, but it is now used only for copolymers intended for latex applications and for small particle ABS. The second major way to alter kinetics is to utilize ionic polymerization instead of free radical polymerization. Styrene polymerization can proceed through a positively charged species (cationic polymerization) or a negatively charged species (anionic polymerization). These polymerizations are very sensitive to impurities so that extensive pretreatment of the monomer is required. Even with pretreatment, there is much chain transfer taking place during cationic polymerization so that the molecular weight is low and molecular weight distributions are similar to free radical polymerizations. Anionic polymerization, however, can be used to produce high molecular weight narrow distribution polystyrene. If all the chains are initiated at the same time and the temperature is kept low to minimize chain transfer, molecular weight distributions very close to monodisperse can be produced. The commercial uses of these polymers seem to be limited to instrument calibrations and laboratory studies of the effects of molecular weight on rheology and physical properties. However, anionic polymerization as a potential commercial method for producing polystyrene has been extensively studied by Dow and others. The potential for high polymerization rates, complete conversion of

52

B. J. MEISTER AND C. J. CUMMINGS

styrene, low oligomers, and the possibility of tailoring molecular weight distributions over a wider range was the driver behind this work. Much of this effort has been focused on backmixed reactors where the rate of addition of monomer could be used to control the rate of reaction [15]. Although anionic polymerization has been shown to be a serious contender for the production of crystal polystyrene, no commercial production facilities have been built. All of the above and later sections will deal only with atactic polystyrene. Styrene monomer can also be polymerized to stereoregular structure through the use of coordination catalysts. Until 1986, only isotactic (phenyl rings cis to each other) had been produced. Isotactic polystyrene crystallizes slowly, has a melting point of 230 °C and has found no commercial uses. Since 1986, considerable industrial research and development has been focused on syndiotactic polystyrene (phenyl rings trans to each other). This material crystallizes more rapidly and has a higher melting point of 270 °C. However, because it crystallizes during polymerization, processing is more complicated than for atactic polystyrene, and the specifics are best reviewed in a discussion of the manufacture of crystalline polymers.

2.3

CONSTRAINTS DUE TO REACTOR MIXING

An important characteristic of a polymerization reactor is the degree of mixing. Mixing is important to have control of the temperature distribution and the composition distribution in the reactor. In most cases, we can separate into two components of mixing, axial mixing in the direction of flow and radial mixing across the reactor perpendicular to flow. In a batch reactor, one might change this separation to mixing parallel to and perpendicular to the agitator shaft. The degree of axial mixing can vary from effectively zero in a perfect plug flow reactor to infinite in a perfectly backmixed reactor. In the perfect plug flow reactor with perfect radial mixing, every fluid element that enters the reactor at time t leaves the reactor at time (t + 0), where 9 is the residence time defined by the reactor volume divided by the flow rate. This could be accomplished by connecting a large number of stirred tanks in series or possibly using a pipeline filled with effective static mixers. However, if an empty pipe is used, because of the viscosities involved, laminar flow will take place, the velocity profile will be parabolic and material near the wall will have a much longer residence time than material in the center and will polymerize to much higher conversion. This was learned in a dramatic fashion in the early 1960s when a 1 inch diameter pilot pipeline was scaled up to commercial scale. The commercial unit operated for less than a day before it progressively filled with solid polymer with monomer squirting down the middle. In the 1 inch diameter pipe, radial diffusion was sufficient to accomplish elimination of the radial concentration gradient. In the large-scale pipeline, this was not the case. Once

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

53

a wall layer has built up, it continues to grow because monomer diffuses into the layer faster than polymer diffuses out. This monomer polymerizes and adds to the layer. If any branching reactions can take place, the effect is amplified, because the molecular weight of existing polymer in the layer increases with time. This general phenomenon of buildup of solid polymer in static areas of the reactor is common to all continuous solution reactors. A major component of the design is to eliminate regions where this may occur. This is one of the drivers for incorporating agitators in the tower reactors which will be discussed in the next section and has also been a driver for removing heat transfer tubes from the reactor which add wall surface. The opposite of the large diameter pipeline with little axial or radial mixing is the perfect backmixed reactor with instantaneous mixing and uniformity. For polystyrene reactors with several hours of residence time, complete mixing in l–2min is usually adequate to satisfy a practical definition of perfectly mixed. The probability of exit of any fluid element from this type of reactor is independent of when it entered. The residence time distribution is exponential and the molecular weight distribution in the case of no termination is Mw/Mn = 2.0, which will spread out to 2.3 when chain transfer controls. If product requirements necessitate a narrower residence time distribution, one can utilize several of these reactors in series. This becomes necessary to control the grafting distribution in rubber modified polystyrene. The choice of agitator for the stirred tank reactor depends primarily on the viscosity range of the polymer solution in the reactor, ranging from turbines on the low end to helical agitators that function more like extruders on the high end. The details of the agitators are specific to the individual companies involved. These reactors have little problem with wall buildup, particularly when the walls are not used for heat transfer. Cooling is primarily from the cold feed and by operating the reactor at the pressure at which the polymer solution boils. Because the termination reaction is strongly diffusion controlled, these reactors are susceptible to a two-phase solution, particularly when operated at low levels of solvent. Also, as the percentage conversion is increased, the viscosity increases and mixing becomes less and less perfect. As this occurs, temperature gradients within the reactor become more significant. Increased temperature at the same pressure means increased conversion and product uniformity may suffer. As with most polymer process work, detailed modeling, pilot plant range finding, and plant-scale trial and error lead to optimized conditions. Figure 3.3 summarizes mixing for the various types of continuous solution reactors. Eliminating the wall buildup prevalent in nonradial mixed reactors drives the reactor designer from left to right. The choice from top to bottom depends on the product requirements and the degree of temperature and composition dispersion needed to accomplish them, and the capital cost of multiple control zones in series.

54

B. J. MEISTER AND C. J. CUMMINGS Instantaneous perfect mixing

Radial mixing

.With internal mixing Higher agitation

.

Recirculation loop

D

C

Perfect CSTR

C

Recirculation

_

c Agitated tower

Unagitated tower Empty pipe

Axially segregated agitated reactor

CSTRs in series With internals Pipe with static mixers

' No mixing Figure 3.3 Solution process polystyrene reactors as a function of axial and radial mixing. Reproduced with permission from Encyclopedia of Polymer Science and Engineering, Mark (Ed.), John Wiley & Sons, NY. Copyright John Wiley & Sons

2.4

CONSTRAINTS RELATED TO THE RUBBER MODIFICATION OF POLYSTYRENE

Rubber is incorporated with polystyrene in commercial high-impact polystyrene (HIPS). The rubber ends up as domains in the size range 0.5-5.0 um with the range 1.0–2.0 um being the most effective at toughening the polystyrene. It was learned early that compounding rubber into finished polystyrene was relatively ineffective. There must be some connecting links between the rubber phase and the polystyrene phase. This is usually accomplished by grafting some of the polystyrene chains to the rubber during the polymerization. The current process for

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

55

rubber-modified polystyrene evolved from early work at Dow [16] that demonstrated that if rubber is dissolved in the feed and the polymerization was carried out in tower-type reactors with sufficient agitation, the final polystyrene product contained small rubber domains that contributed significant toughness to the product. The initial process development work is detailed in a paper by Amos [11]. In the continuous solution process for rubber-modified polystyrene, the rubber is first dissolved in the styrene feed. This rubber solution is then fed to the reactor. A complex sequence of morphology development takes place in the polymer solution as the styrene polymerizes. The phase diagram in Figure 3.4 illustrates that as soon as point A a small amount of polystyrene is formed in a rubberized feed, illustrated here as 8 % rubber, and two phases are formed, one phase containing polystyrene and styrene and the other containing rubber and styrene. The tie lines in Figure 3.4 such as line B-C show that the rubber phase is more dilute than the polystyrene phase. The polystyrene phase exists as small domains in the continuous rubber phase. The polymerization proceeds along the line A-E with increasing amounts of the polystyrene phase suspended as small droplets in the rubber phase. When point F is reached, the volume of the polystyrene phase is equal to the volume of the rubber phase. With sufficient agitation at this point, phase inversion starts to take place and polystyrene Styrene 100 wt%

40 wt%

40 wt%

Polystyrene

Equal phase volume line

Rubber

Figure 3.4 Ternary phase diagram for the system styrene-polystyrene-polybutadiene rubber. Reproduced with permission from Encyclopedia of Polymer Science and Engineering, Mark (Ed.), John Wiley & Sons, NY. Copyright John Wiley & Sons

56

B. J. MEISTER AND C. J. CUMMINGS

gradually becomes the continuous phase. Soon after phase inversion is complete, the rubber domains become fixed in size and these particles maintain their boundaries as polymerization continues in both phases. A typical set of rubber particles formed in this sequence of events, that has been polymerized to relatively high conversion, is shown in Figure 3.5. This material has particles 2-3 um in diameter with large internal occlusions of polystyrene. These occlusions tend to grow larger the more polymerization takes place after the boundaries are set. The size of the rubber particles and the size of the occlusions within the particle are key parameters determining the final properties of HIPS, in addition to the amount of rubber and the crosslink density of the rubber phase. The variables that determine the ultimate morphology include the reactor agitation, the amount of backmixing, and the variables that affect the grafting and crosslinking reactions between the two phases. These include the type and molecular weight of the rubber, any peroxide that might be used as an initiator or grafting agent and the molecular weight of the polystyrene being produced. Echte [17] has demonstrated how the morphologies developed during polymerization relate to the structures of block copolymers in solution. When a polystyrene chain is grafted to a rubber chain, the polystyrene part is compatible with the polystyrene phase and the rubber part is compatible with the rubber phase. This molecule migrates to the interface between the phases. If the

Figure 3.5 Typical HIPS particles

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

57

polystyrene part is much shorter than the rubber part, the molecule can conform better to a small polystyrene occlusion in the rubber phase. Similarly, if the rubber part is much smaller than the polystyrene part, the molecule can conform better to a small rubber domain in the polystyrene phase. This would be the exterior shell of the rubber particle. Turley and Keskkula [18] showed clearly how increasing agitation decreases the particle size, the amount of occluded polystyrene within the particle and the rubber phase volume fraction when graft and graft molecular weight are held constant. How does all this physical chemistry affect the reactor design? It becomes very important because many reactor types used for crystal polystyrene are unsuitable for rubbermodified polystyrene. If one attempted to use a rubberized feed in the early unagitated tower or tube tank reactors, a very disappointing result would be obtained as phase inversion would not take place. If one feeds a rubber solution to one of the more modern single-stage backmixed boiling reactors operating at high conversion used for GP polystyrene, another relatively undesirable result is obtained. Because the reactor is operating well past the phase inversion point, the rubberized feed is immediately broken up and the monomer diffuses out, leaving large dense particles. Some of the particles stay in the reactor a long time and become overgrafted and others leave the reactor rapidly with no graft, leading to a rubber-modified polystyrene with poor properties. Using chemistry that promotes high rapid grafting can surmount most of these problems [19]. However, if the typical highly occluded uniform particles are desired and backmixed reactors are to be utilized, then four reactors in series are usually required. The first reactor operates with rubber continuous to graft the rubber, the second reactor operates just above phase inversion to size the particles and two finishing reactors are used to build the occlusions. The alternative to this approach is the original Dow process. [16,20,21] Three agitated tower reactors are operated in series with phase inversion and particle sizing occurring in the first reactor, and occlusion building occurring in the second and third reactors which are agitated at a slower rate.

2.5

REACTOR REQUIREMENTS FOR PRODUCING

COPOLYMERS

The production of copolymers leads to some additional constraints to reactor design beyond what is required for homopolymer. The most important of these is composition drift. The reactivity ratios of a monomer mixture define the composition of a copolymer that is instantaneously produced from a given monomer mixture. This is true in a plug flow reactor or a backmixed reactor. However, in the plug flow reactor, the copolymer composition drifts from that produced from the initial monomer composition to that produced by the monomer composition at the end of the polymerization. In contrast, in the backmixed reactor, all copolymer produced is of the same composition, which

58

B. J. MEISTER AND C. J. CUMMINGS

is determined by the ratio of unreacted monomers in the reactor, not the feed composition. The largest volume copolymer of styrene is styrene-co-acrylonitrile (SAN). The reactivity ratios for this system with styrene defined as monomer 1 are approximately r\ = 0.40 and r2 = 0.04. This set of values makes this polymer subject to a high degree of composition drift if the shift is not compensated by reactor design. Composition drift is undesirable in SAN copolymers because changes of only a few weight percent acrylonitrile can make two copolymer chains of high molecular weight incompatible and lead to haze in the final product. This is particularly disadvantageous if the copolymer is being sold as a clear molding resin. This is of less concern if the SAN is being blended to produce rubbermodified polymer ABS. Fortunately, in the styrene-acrylonitrile system, there is an azeotropic composition where the equilibrium monomer composition and polymer composition have the same ratio. This is illustrated in Figure 3.6, which shows the azeotropic point to be 76 wt% styrene and 24 wt% acrylonitrile. Polymers produced in linear flow reactors or axially segregated reactors usually operate close to the azeotropic composition. If not, they must have many addition ports to add the monomer that is being depleted in order to prevent a large composition drift. Reactors specifically designed for copolymers are usually of the backmixed type. If ebullient heat transfer is utilized, much care must be taken with the handling of the vapor stream, which will be rich in acrylonitrile owing to its low boiling point. Once this is condensed, it must be returned to the reactor and mixed in rapidly, so as not to produce the high-percentage acrylonitrile polymer chains that would be incompatible with the main polymer being produced.

'

I

I

1 . 1

0 0.2 0.4 0.6 0.8 1.0 Mole fraction styrene in feed monomers

Figure 3.6 Relationship between feed composition and copolymer composition of syrene-acrylonitrile. Point A indicates the azeotropic composition. Reproduced with permission from Encyclopedia of Polymer Science and Engineering, Mark (Ed.), John Wiley & Sons, NY. Copyright John Wiley & Sons

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE 59 3 3.1

POLYSTYRENE DEVOLATILIZATION DEVOLATILIZATION

CONCEPTS

Because the reaction processes described previously do not take the reaction to completion, a separate unit operation is required to remove monomer(s) and solvent from the polymer product. This is typically completed by heating the polymer solution and flashing off the unwanted monomer and solvent. There are several concerns such as equilibrium levels, polymer degradation, and mass transfer that must be considered. When completing a flash devolatilization, it is usually desirable to heat the polymer as much as practical to ensure the best separation. However, not unlike other polymers, polystyrene will degrade under typical devolatilization temperatures, 210-250 °C. In addition to loss of polymer molecular weight, styrene is formed when the polymer and/or oligomers degrade. When designing a devolatilization unit operation, it is important to minimize residence time at these elevated temperatures. Modern commercial devolatilization designs have focused on maximizing heat transfer and minimizing residence time. Devolatilization performance is usually measured against the equilibrium amount of volatile in the final polymer. The equilibrium level for the devolatilization conditions used can be calculated using a simplified Flory-Huggins equation for monomer activity in the polymer melt [6]. By equating the partial pressure of the monomer solution to the flash tank partial pressure, the following results: P = (j)P0e(l+'/J where P is the flash tank pressure, 4> is the volume fraction of monomer in the devolatilized product, p° is the vapor pressure of pure styrene at the flash temperature and X is the Flory-Huggins interaction parameter. For a polystyrene-styrene system, X — 0.34. Therefore,

^

3.8/>c

and under operating conditions (230 °C): W

P

where W is the weight fraction of styrene in the devolatilized polystyrene. In a commercial polystyrene manufacturing plant, the final residual levels can deviate significantly from the equilibrium levels. This deviation can be attributed to

60

B. J. MEISTER AND C. J. CUMMINGS

two factors. The first is the generation of monomer from the degradation of polymer and oligomers. Even if moderate temperatures are used, if there is significant residence time in the system after the flash devolatilization, there will be monomer formation. In a polymer solution containing 300 ppm styrene dimer held for 1 h at 230 °C, an additional 95 ppm of styrene will be generated [6]. The second factor causing a deviation between equilibrium and actual volatile concentrations is mass transfer. In the flash devolatilization of polystyrene, the generation of thin polymer films is critical. However, even with the high surface area created by bubble formation and growth, mass transfer significantly impacts the end results. Measuring the mass transfer rate in a dynamic foam is extremely difficult because the interfacial area is both unknown and constantly changing. However, in comparing commercial data to equilibrium calculations, Meister and Platt have developed correlations for the mass transfer effect [22]. In 1989, Meister and Platt published data comparing equilibrium volatile levels in polystyrene with actual commercial performance [22]. Figure 3.7 is a plot of styrene concentration versus devolatilizer vacuum. The two products shown differ in their molecular weight: product 1 has Mw = 304000 and product 2 has Mw = 196000. It is important to note that the offset between the data and the equilibrium levels is independent of both viscosity and devolatilizer pressure. Figures 3.8 and 3.9 are plots of styrene dimer and trimer levels as a function of devolatilizer pressure. Because these oligomers are of much higher molecular weight than styrene, their diffusion rates are much lower and so the departure from equilibrium is more significant. 3.2

DEVOLATILIZATION

EQUIPMENT

There are three basic types of devolatilization equipment that have been used for the commercial manufacture of polystyrene: wiped film evaporators, devolatilizing extruders and flash evaporators. In wiped film evaporators, the polymer solution is fed into a vessel under vacuum. The solution is moved into thin films along the vessel walls by a set of rotating blades. These blades continue to move the polymer through the vessel while continually renewing the surface area. The tank walls are heated to supply the required energy for devolatilization. These units are typically mounted vertically with the polymer solution fed at the top. At the bottom is a melt pool where a gear pump transfers the melt to the next unit operation, typically pelletization. Advantages of the wiped film evaporator include: • Because the devolatilization occurs in thin films on the heated vessel wall, it is possible to input heat during devolatilization. This can be important because for many heat-sensitive polymers, the temperature required for an adiabatic flash would cause excessive polymer degradation.

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

1000

61

DATA-PRODUCT 2DATA-PRODUCT 1

750

500

EQUILIBRIUMPRODUCT 2

250

A O D O

10

PRODUCT PRODUCT PRODUCT PRODUCT

1 1 2 2

STREAM STREAM STREAM STREAM

15 20 25 30 VACUUM (TORR)

A B A B

35

Figure 3.7 Residual styrene as a function of devolatilizer vaccum. Reprinted with permission from B. J. Meister and A. E. Platt, Ind. Eng. Chem. Res., 28, 1662 (1989). Copyright 1989 American Chemical Society

• Wiped film evaporators are more effective at removal of components with lower diffusion constants such as styrene dimer and trimer. Although there are clearly some specific advantages with the wiped film evaporators, they have not been widely applied for commercial polystyrene production. Reasons for this are most likely the high equipment and maintenance costs associated with these types of units. The second type of equipment used for volatile removal from polystyrene is the devolatilizing extruder. In these devices, an extruder is equipped with one or more pressure let-down sequences where vacuum is applied. In these devices, polymer surfaces are constantly being renewed, giving excellent mass transfer. Another advantage with the devolatilizing extruder is the ability to add and mix additives after devolatilization. This is especially useful if the additive has a

62

B. J. MEISTER AND C. J. CUMMINGS

500 DATA-PRODUCT 2

400

02 UJ

^ 300 Q

DATA-PRODUCT 1 1

200

100 -

A O d O 10

I 15

PRODUCT PRODUCT PRODUCT PRODUCT I I 20 25

1 1 2 2

STREAM STREAM STREAM STREAM I I 30 35

A B A B

VACUUM (TORR)

Figure 3.8 Residual styrene dimer as a function of devolatilizer vacuum. Reprinted with permission from B. J. Meister and A. E. Platt, Ind. Eng. Chem. Res., 28, 1662 (1989). Copyright 1989 American Chemical Society

vapor pressure that would cause it to be removed under vacuum conditions. While these advantages are important, extruders are probably the most expensive of the three equipment types discussed here. The final type of equipment to be discussed is what has been termed the flash evaporator. While there are many variations of this type of equipment, they all have three main components. The polymer solution is first heated via some type of heat exchanger (Figure 3.10). It is then forwarded to a separation vessel, or flash tank, where the vapors formed are disengaged from the polymer melt. The polymer then collects at the bottom of this vessel until it is forwarded via a gear pump to the next unit operation, typically pelletization. The types of heat exchangers used in these processes can vary widely. In probably the simplest form, a standard shell and tube type exchanger can be

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE 63

5000 DATA-PRODUCT 2

4000

EQUILIBRIUM-PRODUCT 2

w I

3000 DATA-PRODUCT 1 2000

EQUILIBRIUM-PRODUCT 1

1000

A O d O

05

10

PRODUCT 1 PRODUCT 1 PRODUCT 2 PRODUCT 2

STREAM A STREAM B STREAM A STREAM B

15 20 25 30 VACUUM (TORR)

35

Figure 3.9 Residual styrene trimer as a function of devolatilizer vacuum. Reprinted with permission from B. J. Meister and A. E. Platt, Ind. Eng. Chem. Res., 28, 1662 (1989). Copyright 1989 American Chemical Society

used to heat the polymer solution. A pressure control valve may be installed after the heat exchanger to eliminate flashing before entering the flash tank. In this type of operation, the flash is completely adiabatic and care must be taken so that the polymer is not cooled to a point where the viscosity is too great. Conversely, if the temperature is excessive, the polymer will degrade. This type of process has been used very broadly within the polystyrene industry. Over the years, there have been many improvements made to the basic process. One basic improvement is to install a die plate at the top of the flash vessel to create strands of foaming polymer. The falling strand devolatilizer, as it has been called, is an effort to increase the polymer surface area and come closer to reaching equilibrium. In addition to having falling strands, other obstructions have been placed in the flash tank to increase the surface area [23–25],

64

B. J. MEISTER AND C. J. CUMMINGS Pressure Control Valve

Flash tank

Heat exchanger

Polymer melt pool Polymer solution from reactors

TLUJ

Polymer gear pump

Polymer forwarded to pelletization unit Figure 3.10

Basic flash devolatilization process

As mentioned previously, the process shown in Figure 3.10 involves an adiabatic flash. Another improvement to this process is to input heat while the flash is occurring. While this can be done by simply removing the control valve shown in the diagram, reducing this to practice is more difficult. When the back-pressure valve is eliminated, vaporization will begin in the heat exchanger and the two-phase mixture will be carried forward into the flash tank. Ensuring adequate heat transfer into the solution and proper vapor-melt disengagement are two important engineering aspects that must be addressed. Another variation of this process is to mount the shell and tube heat exchanger directly on the flash tank so it will discharge as falling strands directly into the vessel [26,27]. There have been other variations and improvements to this basic process. A common improvement is to insert static mixing elements within the heat exchanger tubes. This can increase heat transfer significantly. Also, by inserting these mixing elements into the polymer flow path, the pressure drop across the tubes is increased and the point of flashing and subsequent polymer-vapor disengagement can be better controlled. Another variation of the flash devolatilization process is the use of plate-type heat exchangers [2831]. These exchangers can be constructed such that they can be inserted directly into a flash vessel. The flowpath inside such an exchanger is typically a long, thin slot. Advantages of these exchangers include high surface area and the ability to design the flow channel to optimize where flashing occurs.

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE 3.3

STEAM

65

STRIPPING

Because of equilibrium and mass transfer limits, the conventional flash devolatilization processes cannot reach monomer/solvent levels significantly below 200 ppm. If levels below 200 ppm are required, an alternative process using a stripping agent is required [32,33]. Although water has been used widely because of its low cost, alternative stripping aids can also be used [34]. Alcohols and light hydrocarbons are potential alternatives. Supercritical carbon dioxide has also been explored [35]. A basic version of this steam stripping process is shown in Figure 3.11. The starting point for this process is a polystyrene melt that has already been nearly completely devolatilized. The stripping agent is injected and mixed into the system and the polymer-stripping agent solution is flashed in another flash vessel. The final product is then pumped from the flash vessel to a finishing unit operation. The purpose of the water injection is twofold: first, the presence of water during the flash significantly reduces the partial pressure of monomer and/or solvent, and second, the flashing water creates more surface area to aid mass transfer. Although steam stripping is used commercially, it is not without drawbacks. The primary drawback is the additional capital and operational costs to install such a system. In addition, there are other process issues that must also be addressed carefully. Water injection Mixer Heat Exchanger

First Devolatilizer Polymer solution from reactors

Second Devolatilizer

Polymer forwarded to pelletization unit Figure 3.11

Basic polystyrene steam stripping process

66

4

B. J. MEISTER AND C. J. CUMMINGS

CURRENT POLYSTYRENE POLYMERIZATION PROCESSES

As one examines the evolution of polystyrene production processes over the last 60 years, one finds that economics, quality of the products and the range of products that can be produced are the main drivers for this evolution. Processes that do not make money for their owners eventually get shut down and the business evolves. Because this is a high-capital business and growth is relatively modest, evolution is slow. However, if one compares the polystyrene business today with that in the 1960s, one finds tremendous changes in the size and type of process used, the companies that are competitive producers, and the global locations of the production plants. Table 3.1 summarizes some of the advantages and disadvantages of the various process choices already described. However, all of the styrenic plants being built today, with the exception of emulsion-based ABS, are continuous free radical polymerizations. If one were to construct a crystal polystyrene (GP) plant today, one would likely choose a backmixed reactor (CSTR) as the principal reactor. This reactor would operate in the 50–70 % solids range and would discharge by means of a gear pump to a devolatilizer system. In this case there would probably be two devolatilizers, the first raising the solids to the range 80–85 % and the second producing the final product. This final devolatilizer would probably combine a heater, a distributor, and a flash tank operated below l0 mmHg absolute pressure and a temperature above 230 °C. The heater might well serve as the distributor with tubes or channels open to the flash tank so vaporization begins in the heater and the flash is not completely adiabatic. From Figure 3.7, the final polymer is probably below 500 ppm residuals under these conditions. The vapor may be cooled in a desuperheater and then passed through a final condenser and recycled back to the feed. As ethylbenzene, cumene and other hydrocarbons that do not polymerize are present in the styrene, they build up in the recycle, causing this to be a solution process whether one wanted it or not. Some of these hydrocarbons leave in the product, but when operating below 10 mmHg absolute pressure in the devolatilizer, there generally is excess recycle generated in the process. As shown in Figure 3.12, the product recovered in the devolatilizer is pumped to a die and cut with a cutter into pellets. The residence time from the devolatilizer to the die is minimized to control the regeneration of styrene at elevated temperature. Plasticizer and other additives are often added at this point to avoid vaporizing them in the devolatilizer. Either a mechanical or a static mixer might be used to ensure the additives are uniformly mixed prior to pelletization. A world-scale plant today may produce as much as 10000 kg/h of polystyrene. Figure 3.13 illustrates what one might build today for a high-impact polystyrene (HIPS) plant today. The individual companies involved have patented a number of specialized reactor configurations. These are illustrated in Figure 27 in a review by Echte [36]. One common configuration is shown here in Figure 3.13 with two backmixed reactors (CSTR) followed by two linear flow

Table 3.1

Polymerization methods for manufacturing polystyrene.

Reactor type

Polymerizing system

Advantages

Continuous solution Free radical (backmixed reactor)

Styrene monomer Recycled solvent W or W/O initiator

Good Temperature Control Limited in final conversion Limited in product range Good for copolymers Good clarity and color Pumping difficulties Uniform product

Continuous solution Free radical (linear flow reactor)

Styrene monomer Recycled solvent W or W/O initiator

Good range of products Good for rubber extension Good clarity and color

Large number of control zones High capital Pumping difficulties Low-cost process for HIPS

Excellent heat control High conversion No devolatilization Good range of products

Need prereactor for HIPS Poorer clarity Poor uniformity Round beads are hazard

High operating costs Better for lowvolume products

Sensitivity to impurities Initiator cost Color of product Cannot produce HIPS

Not proven for high-volume GP

Batch or continuous Styrene monomer Suspension free radical Water carrier Stabilizing agent Several initiators Continuous solution Anionic

Pure styrene monomer Polymerize to completion Much recycled solvent Low residual monomer Anionic initiators High polymerization rate Good for spec, copolymer

Disadvantages

Economics High capital Low-cost process for high-volume GP

68

B. J. MEISTER AND C. J. CUMMINGS Recycle

Desuperheater

Excess Recycle

Feed

Additives

Product Figure 3.12 reactor)

General-purpose polystyrene plant (CSTR=continuous stirred tank

reactors (LFR), all in series, which follows the original patent by BASF [37]. The rubber, which usually arrives at the plant in bales, is ground up in a grinder and added to a large tank with the styrene feed. The rubber is dissolved in this agitated tank over a period of about 12hs. Two tanks are often used in parallel so that one tank feeds the first reactor while the other is dissolving more rubber. The first backmixed reactor is operated with rubber continuously and is used to graft the rubber prior to feeding the second backmixed reactor which operates above phase inversion and is used primarily to size the rubber particles and set the morphology for the two finishing reactors. The solids leaving the first reactor might be 12%, the second reactor might well operate at 30% solids and this conversion is carried to about 80% solids in the two linear flow finishing reactors. The finishing reactors are usually not recirculated or backmixed in order to generate a high level of rubber extension or high toughness for the amount of rubber used. Once the HIPS leaves the reactor system, the finishing and recycle systems may well be identical with those used for generalpurpose polystyrene illustrated in Figure 3.12.

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE Excess Recycle

69

Desuperheater

Product Figure 3.13 High-impact polystyrene plant (CSTR = continuous stirred tank veactor; LFR = linear flow reactor)

The system utilized in Figure 3.13 for HIPS can also be used to produce a solution polymerization ABS. This type of ABS is used in non-glossy applications. The glossy ABS is usually produced in an emulsion process in which emulsified polybutadiene latex is grafted and agglomerated and blended with a continuous phase of SAN. This blended material is then dried and pelletized. This process is not cost competitive with the continuous solution polymeriza tion, but it produces a product with a superior balance of properties that commands a premium price.

5

PROCESS SIMULATION AND CONTROL

Probably no part of the polystyrene production plant has changed as much over the last 30 years as the methods of process control. The early polystyrene processes required little process control because they were operated at reaction rates that were inherently stable. For polystyrene, a rule of thumb is that the reaction rate doubles with every increase in temperature of 10°C. If the reaction is conducted at rates that evolve heat at a rate that requires a temperature

70

B. J. MEISTER AND C. J. CUMMINGS

difference between the polymer and the heat transfer fluid of less than 10°C, then increases in polymer temperature lead to larger increases in heat transfer than increases in the heat evolved and the reactor is stable. However, the development of faster responding process control systems and the evolution to computer control have made possible the move into operational regimes that are not inherently stable, but are controlled by systems that have a faster response to an upset than the process. The other two parts of the major change in process control are first the implementation of gel permeation chromatography as a process control technique for molecular weight and molecular weight distribution. Second is the evolution of kinetic models for styrene polymerization for use as off-line and on-line computerized simulations of the reaction. Fundamental polymerization kinetics are the link between the operational parameters of temperature, monomer concentration and residence time and the resulting rates of reaction and molecular weight produced at any point in the reactor. Off-line simulations for the polymerization of polystyrene and the copolymerization of styrene and acrylonitrile based on fundamental kinetics were developed by Meister [38] at Dow Chemical in the early 1970s. The recognition that both the termination and propagation reactions were diffusion controlled and that the magnitude of diffusion control was primarily dependent on the polymer concentration was the key to obtaining accurate simulations that could be used to guide the operating conditions of the production plants. Models that are similar to the Dow models have been published in the literature and are widely used today. In particular, Hamielec and a series of students [39–41] developed models that are similar in format and predictions, although very different in some of the details. In the late 1970s, the Dow polystyrene production facilities made a major conversion from thermally initiated polystyrene to polymerization initiated by a bifunctional initiator. This brought about a new series of products that could be produced at higher rates and had improved grafting and a lower level of oligomers. The use of the simulation programs was a major factor in making possible the conversion to the new operating conditions and the development of new rules of thumb for adjusting operating conditions to maintain product properties. A method for the computation of the molecular weight of polystyrene formed when using a bifunctional initiator was a key to the success of the model. This was true even though some of the mechanisms involved only became understood later. Models for polystyrene polymerization using bifunctional initiators are now available in the literature [42,43]. Improving the predictions of the diffusion-controlled termination and propagation reactions is a subject of continuing interest [44]. This is particularly true because of the increasing use of backmixed reactors for the production of GP polystyrene and the increasing use of bifunctional initiators in these reactors at high conversions where the diffusion control is amplified. This search for polymerization conditions where high molecular weight polymers can be produced

COMMERCIAL PROCESSES FOR THE MANUFACTURE OF POLYSTYRENE

71

at high production rates often leads to situations where multiple steady states occur [45]. This can lead to oscillatory behavior [46] and it can also lead to the existence of two phases of differing concentration and temperature within the same reactor. Computer control of the operating parameters to maintain setpoints, the use of off-line kinetic models to determine operating conditions, and the use of online models to make predictions of the intermediate parameters and the final product properties as a function of the operating history have all become commonplace in industry over the last 30 years. This certainly has had a strong influence on the improvements in product quality and uniformity and the large increases in production rate that have occurred over this period of time.

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

Blyth, J., and Hofmann, A. W., Ann. Chem., 53, 289 (1845). Staudinger, H., Ber. Dtsch. Chem. Ges., 53, 1073 (1920). Boyer, R. F., J. Macromol. Sci. Chem., A15, 1411 (1981). Debell, J. M., et l., German Plastics Practice, ebell and Richardson, Springfield, MA (1946). Simon, R. H. M., and Chappelear, D. C., Technology of Styrente Polymerization Reactors and Processes, ACS Symposium Series 104, American Chemical Society, Washington, DC (1979). Meister, B. J., and Malanga, M. T., in Styrene Polymers, Encyclopedia of Polymer Science and Engineering, Vol. 16, Wiley, New York, pp. 21–61 (1989). Albright, L. F., Processes for Major Addition Type Plastics and Their Monomers, McGraw-Hill, New York (1974). Boundy, R. H., and Boyer, R. F., Styrene, Its Polymers, Copolymers, and Derivatives, ACS Monograph No. 115, Reinhold, New York (1952). Svec, P., et al., Styrene-based Plastics and Their Modification, Ellis Horwood, New York (1989). Albalak, R. J., Polymer Devolatilization, Marcel Dekker, New York (1996). Amos, J. L., Polym. Eng. Sci., 14, 1 (1974). Allen, L, et al, US Patent 2496653 (to Union Carbide) (1950). Latinen, G. A., US Patent 3794471 (to Monsanto) (1971). Priddy, D. B., in Styrene Plastics, Kirk Othmer Encyclopedia of Chemical Technology, 4th edn, Vol. 22, pp. 1015-1073, Wiley, New York (1997). Priddy, D. B., Pirc, M., and Meister, B. J., Polym. React. Eng., 1, 343 (1993). Amos, J. L., et al., US Patent 2694692 (to the Dow Chemical) (1954). Echte, A., Angew. Makromol. Chem., 58, 175 (1977). Turley, S. G., and Keskkula, H., Polymer, 21, 466 (1960). Meister, B. J., et al, US Patent 4 876 372 (to Dow Chemical) (1989). Platzer, N. Ind. Eng. Chem., 62(1), 6 (1970). Mott C. L., and Kozakiewicz, B. A., US Patent 4221883 (to Dow Chemical) (1981). Meister, B. J., and Platt, A. E., Ind. Eng. Chem. Res., 28, 1659 (1989). Aboul-Nasr, O. T., US Patent 4934433 (to Polysar Financial Services) (1990).

72

B. J. MEISTER AND C. J. CUMMINGS

24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

Aboul-Nasr, O. T., US Patent 5069750 (to Polysar Financial Services) (1991). Desroches, D., and Krupinski, S., US Patent 5 874 525 (to Nova Chemicals) (1999). Gordon, R. E., and McNeill, G. A., US Patent 3 853 672 (to Monsanto) (1973). Hagberg, C. G., US Patent 3966538 (to Monsanto) (1976). Fink, P., et al., US Patent 4153 501 (to BASF) (1979). Aneja, V. P., and Skillbeck, J. P., US Patent 4 808 262 (to General Electric) (1989). Mattiussi, A., et al., US Patent 5084134 (to Montedipe) (1992). Cummings, C. J., and Meister, B. J., US Patent 5453 158 (to Dow Chemical) (1995). Szabo, T. T., US Patent 3 773 740 (to Union Carbide) (1971). Darribere, C., et al., Presented at the 6th International Workshop on Polymer Reaction Engineering (1998). Skilbeck, J. P., US Patent 5350813 (to Novacor Chemicals) (1994). Sacide, A., and Duda, J. L., AIChE J., 44, 582 (1998). Echte, A., in Rubber Toughened Plastics, ed. Riew, C. K. Advances in Chemistry Series 222, American Chemical Society, Washington, DC (1989). Bronstert, K., et al., US Patent 3658946 (to BASF) (1972). Meister, B. J., Dow Internal Documents (1974, 1975). Hui, A. W., and Hamielec, A. E., /. Appl. Polym. Sci., 16, 749 (1972). Friis, N., and Hamielec, A. E., /. Appl. Polym. Sci., 19, 97 (1975). Marten, F. L., and Hamielec, A. E., J. Appl. Polym. Sci., 27, 489 (1982). Villalobos, M. A., et al., J. Appl. Polym Sci., 42, 629 (1991). Dhib, R. et al., Polym. React. Eng., 8, 209 (2000). Vivaldo-Lima, E., et al., Polym. React. Eng. 2, 17 (1994). Henderson, L. S., Chem. Eng. Prog., 83, 42 (1987). Villa, C. M., et al., Polym. React. Eng., 7, 151 (1999).

34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46.

DUANE B. PRIDDY Dow Polystyrene R&D, Midland, Ml, USA

1

INTRODUCTION

Production of polystyrene (PS) in North America began in 1938 by The Dow Chemical Company. The first Dow process involved simply immersing metal cans (full of styrene monomer) in a heated oil bath until the monomer conversion reached 99+ %. Then the cans were opened and the PS was crushed to form granules. The PS prepared and isolated in this manner contained high levels of unreacted styrene monomer which ended up in the final fabricated articles. For the next several years, R&D efforts were focused upon the development of improved manufacturing processes and devolatilization processes to remove the unreacted monomer. The batch 'can process' finally gave way to continuous bulk polymerization technology (Figure 4.1) and suspension polymerization processes. Suspension polymerization had the advantage of yielding polymer granules direct from the polymerization reactor and utilized organic peroxide initiators which led to faster polymerization rates and very high monomer conversion (>99.5 %). The continuous bulk process, on the other hand, only allowed conversion of monomer to polymer of ~80 %, resulting in the need to remove the unreacted styrene monomer. The process developed by Dow to devolatilize the polymer syrup was to pass it through a heat exchanger to take the temperature up to 230-260 °C, and then forwarding it into a vacuum chamber at < 10 mmHg. The unreacted styrene monomer volatilizes from the polymer melt and is then condensed and recycled back into the polymerization process. However, this devolatilization process does not remove all of the monomer. Depending upon the temperature, vacuum, surface area of the polymer melt, and residence time in the Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy 2003 John Wiley & Sons Ltd

74

D. B. PRIDDY

vacuum tank, the level of monomer left in the PS can vary significantly. There is a theoretical minimum that can be achieved based upon the equilibrium partitioning of the monomer vapor with the molten polymer. This minimum varies with temperature and pressure inside the vacuum chamber as shown in Figure 4.2. There has been continuous improvement in PS purification technology over the past 60 years as efforts continued to focus on getting the residual monomer levels lower and lower. These efforts continue even though it is generally Styrene EB Initiator

75%PS

240°C <10mm

100–130°C

30–145°C

145–170°C cutter

Figure 4.1 The most widely utilized PS process. EB = ethylbenzene

—©— ImmHg -Q— 5mmHg -O — lOmmHg - -• - - 5mmHg+10%water

jf, 300c '« 200i) J», ]

D

100-

•

220

-a

I 225

T 230

I I I 235 240 245 Temperature (°C)

I 250

I 255

I 260

Figure 4.2 Vapor-polymer equilibrium partitioning data for styrene in PS vs temperature and pressure [1 ]

APPROACHES TO LOW RESIDUAL POLYSTYRENE

75

accepted that there are no health risks associated with PS containing <1000ppm of residual styrene monomer. However, owing to the relatively low taste and odor threshold of styrene, a taste has been noted in certain foods packaged in PS containing >400 ppm residual monomer. Another problem is that PS is not thermally stable. Therefore, even if special techniques for monomer removal are implemented by the manufacturer, the level of monomer in the final fabricated article is oftentimes higher than that in the raw polymer granules. This chapter summarizes efforts aimed at removing styrene monomer from PS to low levels, i.e. <200ppm. Also discussed is the current state of understanding regarding the mechanism of monomer regeneration by thermal degradation of PS during fabrication.

2

SUMMARY OF R&D APPROACHES

Research can be organized into the following categories: 1. 2. 3. 4. 5. 6.

2.1

devolatilizer design; assisted devolatilization; scavengers; absorbers; high monomer conversion polymerization; solid polymer treatment.

DEVOLATILIZER DESIGN

Most commercial PS producers operate their vacuum devolatilizers at 230260 °C and 5-10 mmHg. From Figure 4.2 it can be seen that the lowest level of residuals possible with conventional vacuum devolatilization operating under these conditions is ~200 ppm of styrene. In actual practice, the level of styrene monomer in PS products falls in the 200-800 ppm range depending upon throughput rate. If the polymer is heated to temperatures higher than 260 °C to drive the volatiles out, the PS begins to decompose (unzip), resulting in the formation of monomer. A way to achieve greater throughput and approach close to equilibrium is the use of wipe film evaporators [2,3]. These devices mechanically deposit the polymer as a thin film on a metal surface. In the 1990s, Dow researchers [4,5] developed a unique device (Figure 4.3) that creates substantial surface area within the polymer melt at very short residence times, which allows the achievement of equilibrium residual levels. The PS melt is passed into a centrifuge equipped with a high surface area porous metal filter. The G-force quickly forces the molten polymer through the

76

D. B. PRIDDY N2 PURGE VACUUM

PACKING

SEAL SEAL HOUSING

DRIVE BASE

Figure 4.3

Drawing of centrifugal devolatilizer

filter, where it forms a thin film on the metal surfaces. The centrifuge chamber is under high vacuum (1 mmHg). The combination of the high vacuum and surface area results in the production of PS having <100ppm residual styrene monomer. After the polymer exits the filter, it is thrown against the perforated exterior surface of the centrifuge where the pressure extrudes it through the holes. A stationary knife blade, that resides a few millimeters from the holes, cuts the extrudate into granules which are conveyed away as a slurry in water. This novel devolatilizer design has been extensively patented. 2.2 2.2.1

ASSISTED DEVOLATILIZATION Strippers

An effective way to get below equilibrium is to replace the styrene in the vapor phase with something else [6]. Generally, it is preferred that the material in the

APPROACHES TO LOW RESIDUAL POLYSTYRENE

77

vapor phase be condensable, nontoxic, low cost, easily separable from the recycle, and inert (e.g. water). However, simply bleeding the replacement material into the vapor space of the devolatilizer would still require significant residence time to reach equilibrium. A way to expedite the equilibrium process is to mix the replacement material with the molten polymer [7]. Then as the material vaporizes in the molten polymer, it nucleates bubbles stripping styrene and polymerization solvent (typically ethylbenzene) residues from the polymer while creating a vapor phase composed mainly of the stripping material. Stripping materials that have been used include nitrogen [7], supercritical fluids [8], ketones [9], methanol [10], and water [11–13].

2.2.2

Bubble Nucieators

Another technique to expedite the transport of the volatile components from the molten polymer is to increase the number and rate of bubbles formed [14]. Techniques that have been used to increase the number of bubbles and their rate of formation (nucleation) are the addition of chemical nucleating agents [15] and ultrasound [16]. Nucleation of bubbles in the molten polymer can help expedite the achievement of equilibrium in conventional falling strand devolatilizers. However, this facilitation mechanism cannot get below equilibrium and thus has minimal value.

2.2.3

Polymer Stabilization

Another situation effecting residual levels existing during PS devolatilization is polymer decomposition or unzipping, which limits the devolatilization temperature to <260 °C. If one tries to go to higher temperature to achieve a more favorable vapor-polymer equilibrium concentration, polymer decomposition begins to dominate. The rate of polymer decomposition can be affected by stabilizing the polymer by the addition of phenolic antioxidants, e.g. 2,6-di-tertbutyl-4-methylphenol [17]. Several Asahi patents indicate the superior

2,4-Di-/m-amyl-6-[H3,5-di-?m-amyl-2-hydroxyphenyl)ethyl]phenyl acrylate

78

D. B. PRIDDY

performance of 2,4-di-terl-amyl-6-[l-(3,5-di-/er/-amyl-2-hydroxyphenyl)ethyl]phenyl acrylate as a stabilizer against monomer regeneration, allowing higher temperature devolatilization to be carried out resulting in the formation of low residual PS [18]. The subject of monomer regeneration during themal degradation of PS will be discussed at greater length later.

2.3

SCAVENGERS

The idea of adding something to molten PS that scavenges residual styrene monomer has been the subject of several R&D efforts over the past several decades. However, there is an inherent difficulty in this concept, i.e. small molecules at very low concentration must diffuse together, find each other and couple together in a viscous polymer. The obvious way to overcome this difficulty is to use the scavenger in higher than stoicheometric concentration. This introduces several new problems, including contamination of the polymer with a new material and increased raw material costs. Figure 4.4 shows the impact of scavenger cost per pound on the increased raw material cost for PS production over a reasonable concentration range for the scavenger. For example, if one places a limit of 1 cent/lb increase to raw material costs for adding a styrene scavenger to the polymer and the scavenger costs $10/lb, the scavenger would have to be effective at a loading in the polymer of <1000ppm. If, however, the scavenger only cost $5/lb, the loading could be increased to 2000 ppm. 2.5-1

u ts !- 5 H o U

'B

H

I °-5H

500

1000

1500

2000

Scavenger Amt. Required (ppm) Figure 4.4

Effect of scavenger costs on PS raw material costs

2500

APPROACHES TO LOW RESIDUAL POLYSTYRENE

79

Scavenger compounds that have been patented include unsaturated fatty acids [19], sulfonyl hydrazides [20], cyclopentadienes [21], styrene-butadiene block copolymers [22], peroxides [23], and terpenes [24]. It would be costly to compound a monomer scavenger into the molten polymer after it exits the devolatilizer. Thus the ideal scavenger has a low molecular weight, is low cost, nontoxic, inert under polymerization conditions so that it can be added to the polymerization feed, and very reactive with styrene once the temperature is raised to >230°C. The problem with peroxides is that they are thermally unstable and do not survive the polymerization to reach the devolatilizer. Blakemore [23] solved this problem by using cyclic peroxides which have very high decomposition temperatures. Their thermal stability is due to the peroxide bond reforming once broken, because the two oxy radicals cannot escape each other so they recouple. If styrene happens to be in the vicinity of the cyclic peroxide while it is a dioxyradical, the diradical adds across the styrene double bond (Scheme 4.1).

Scheme 4.1 styrene

Hypothesized chemistry of how cyclic peroxides lower residual

Rather than adding peroxides that are very thermally stable, ICI researchers added peroxide generating enzymes to PS [25]. These enzymes (oxidases) constantly convert absorbed air to peroxides, which decompose forming radicals that allegedly scavenge the styrene. A class of scavengers developed by Dow researchers is benzocyclobutene (BCB) [26]. In the 1980s, Dow researchers began developing BCB derivatives for the electronics industry. BCB has the same molecular formula as styrene (C8H8), is inert in styrene polymerization, and becomes very reactive toward styrene when heated to >200°C. The chemistry of BCB is shown in Scheme 4.2. Above 200 °C, the strained ring of BCB opens to form o-xylylene, which Diels-Alder couples with the double bond of styrene to form phenyltetralin. Any residual BCB that does not react with styrene continues to react with itself and eventually is converted to C8H8 oligomers.

80

D. B. PRIDDY >200°C Ph

R

Scheme 4.2 Reaction of BCB with styrene to form a dimer. BCB also reacts with itself to form oligomers

2.4

ABSORBERS

It is well known that small molecules can be removed from larger molecules using inorganic materials such as molecular sieves. However, this approach contaminates the polymer with inorganic particulates. The only literature teaching this approach are Russian patents claiming the addition of silica gel [27] and montmorillonite clay [28] to absorb styrene from PS. The advances in nanocomposite technology in recent years may allow further development of this approach. 2.5 2.5.1

HIGH MONOMER CONVERSION

POLYMERIZATION

Free Radical

Typically, continuous bulk free radical polymerization processes produce partial polymer syrups at about 65-80 % solids. The unreacted styrene that remains is removed by evaporation. If the solids content of the polymer could be taken higher, the level of residual styrene in the polymer would be lower, especially when using a one-stage devolatilizer. In suspension polymerization, devolatilization is not even required because styrene is polymerized to >99.9 % monomer conversion. If it were possible to polymerize styrene to very high conversion in bulk polymerization processes, one should be able to achieve significantly lower residual styrene monomer. Owing to viscosity constraints, bulk polymerization reactors cannot operate at >80 % solids. Kelley patented a high-conversion bulk styrene polymerization process resulting in the formation of low residual PS (LRPS), which solves the viscosity problem and allows >99.9% solids to be achieved by finishing the polymerization off in an extruder [29]. Extruders are not very effective heat exchangers yet are designed for handling high-viscosity materials. Thus Kelley carried out the polymerization of pure styrene monomer without the use of a solvent in a conventional polymerizer to normal solids levels and then fed the partial polymer into an extruder where he finished off the polymerization. He used a mixture of initiators having different half-lifes so that radicals were continuously generated. More recently van der Goot and Janssen

APPROACHES TO LOW RESIDUAL POLYSTYRENE

81

[30] also looked at free radical polymerization of styrene directly in an extruder using 1 wt% peroxide initiator. They were not successful in making PS having a molecular weight > 100 000. They got around this problem by attaching a prepolymerizer to the front end of the extruder to polymerize to 25 % solids to make high-MW PS. The low-MW PS then made in the extruder ended up giving them a bimodal PS. The conversion they that achieved in the extruder was 98-99 %. By pulling a vacuum on the last zone of the extruder one could possibly get close to equilibrium styrene level. Injection of steam into the last zone might yield truly LRPS. However, the work published thus far on high conversion bulk polymerization does not appear to be aimed at achieving LRPS.

2.5.2

Anionic

Anionic polymerization chemistry is much better suited for polymerizing styrene to high conversion because of slower rates of termination of growing polymer chains. Therefore, during the polymerization, the concentration of active growing chains is much higher and the MW generally increases with monomer conversion whereas with free radical polymerization, the MW generally decreases rapidly at very high monomer conversions resulting in broadening of the overall polydispersity. Over the past 15 years, there have been significant advances aimed at developing anionic polymerization technology for industrial production of PS. Dow Chemical researchers [31] focused their efforts on solution polymerization in continuous stirred tank reactor (CSTR) processes with heat removal by ebullient cooling whereas BASF and Asahi researchers have focused on continuous plug flow reactors (CPFR) [32]. Since anionic polymerization in CSTR reactors operates at 99.9+ % monomer conversion at steady state and boiling is required to achieve ebullient cooling, a solvent is needed to lower the viscosity of the polymerization mass. The CPFR process, on the other hand, only reaches high monomer conversion in the final stage of the polymerization process where viscosity can be managed by increasing the temperature. The key problem that researchers faced with the CPFR process is the rapid rate of anionic polymerization. The heat removal capacity of CPFR is not sufficient to control the polymerization, resulting in runaway kinetics. The advance that BASF and Asahi researchers discovered is that the propagation rate can be slowed by the addition of certain electron-deficient organometallic reagents (e.g. dibutylmagnesium) [33,34]. Depending on the Mg:Li stoichiometric ratio, the polymerization rate can be adjusted to whatever is necessary to match the heat removal capability of the polymerization reactor being used (Figure 4.5). Anionic polymerization not only allows the production of low residual PS (i.e. typically PS produced using continuous anionic polymerization contains <20ppm residual styrene) but also anionically produced PS is more thermally

82

D. B. PRIDDY 1.4

—e— -a— -0--•• v--

1.2-

1-

Lithium alone Mg/Li = 0.8 Mg/Li = 2 Mg/Li = 4 Mg/Li = 20

0.69-0.4-

0.2

,0-"

50

100

150

200

Polymerization Time (min)

Figure 4.5 Propagation rate-retarding effect of adding dibutylmagnesium to butyllithium-initiated anionic polymerization of styrene

stable. Thus less monomer is regenerated during fabrication of the granules into molded parts. The improved thermal stability of anionic PS is thought to be due to fewer 'weak links' in the polymer backbone. The nature of the weak links continues to be debated. This topic will be discussed later in Section 5. 2.6

SOLID POLYMER TREATMENT

Treatment of PS after it is produced to react out the monomer has been investigated. Both 3 (electron-beam) and -/-radiation have been tried [35], and electron beam appears to be the most effective form of radiation and is most suitable for continuous use [36,37]. The electron beam ruptures C—H bonds, resulting in the formation of PS radicals. These radicals are very reactive and scavenge unreacted monomer. However, if no styrene is in the vicinity of the PS radical, it can do other things such as couple with another PS radical or react with oxygen. Reaction of the radicals with oxygen results in yellowing. Therefore, the irradiation must be carried out on oxygen-free polymer under an inert atmosphere. The optimum temperature and dose for reduction of monomer in polymer granules by a factor of 10 are 4 Mrad at 80 °C (Figure 4.6). Currently electron beam treatment of polymers is used commercially for crosslinking but not for monomer reduction.

83

APPROACHES TO LOW RESIDUAL POLYSTYRENE 700-

600-

500 -3

1

400

2

300 -I

|

—9— 1 Mra< -B — 2 Mrac -O — 3 Mra( - - • - - 4 Mrac

200 H

'

/

*

I 50

100

150

200

250

Irradiation Temperature (°C)

Figure 4.6 granules

3

Effect of temperature and dose on efficiency of monomer reduction in PS

FRIEDEL-CRAFTS CATALYST

The idea is to add a latent Friedel-Crafts catalyst to the polymerization feed that will be inert during the polymerization. Then at >200 °C it becomes activated and converts the residual styrene monomer to phenethyl carbocations. These carbocations can either initiate cationic polymerization of the styrene or can alkylate the benzene rings of PS to become attached. With this approach PS itself is the scavenger for residual styrene monomer in PS (Scheme 4.3). This approach has the advantage over a conventional scavenger approach in that PS units are present in a huge stoichiometric excess relative to residual Catalyst

Scheme 4.3

Alkylation of PS by styrene monomer

84

D. B. PRIDDY

styrene. It was therefore hoped that only a trace amount of catalyst would be required for the process. In evaluating this approach, the question of how and when to introduce the catalyst to the polymerization mixture arose. The simplest method would be to put the catalyst in the styrene monomer being fed to a continuous bulk polymerization system. Then the polymer would be produced with the catalyst molecularly dispersed in it. Priddy et al. evaluated both a sulfonic acid catalyst and also thermally labile acid esters that generate acids during high-temperature devolatilization [38]. Polymerizations and subsequent heat treatments were conducted in glass ampoules. The ampoules were charged with styrene containing either 1000 or 1400ppm of di-tert-butyl peroxide (D/BP). The ampoules were sealed using the freeze-thaw technique to remove dissolved gases and placed in an oven at 135°C for 6h. The polymerization mixtures were subsequently heated further at either 240 or 260 °C for various times to simulate high-temperature devolatilization. Control polymerizations of styrene with DtBP as initiator were also conducted. The levels of residual styrene monomer in the polymer and the polymer molecular weight and polydispersity (PD) were determined using headspace gas chromatography (relative standard deviation = 6%; detection limit = lOppm) and gel permeation chromatography (relative standard deviation = 2 %), respectively. Table 4.1 provides information about the control polymerizations. The results indicate that the residual styrene monomer content increases while heating the polymers at 240 or 260 °C. This is more apparent with polymers heated at 260 °C than with those heated at 240 °C. This result is likely due to the presence of weak links in the polymer chain. When the weak links break, some depolymerization takes place resulting in the formation of monomer and loss of molecular weight (see Section 5). Table 4.1

Polymerization followed by heating styrene using D/BP as an initiator

Reaction conditions Polymer color

Residual styrene (ppm)

A/w

Mn

PD

225600 178100 154 200 122000

114100 89900 82400 61000

2.0 2.0 1.9 2.0

[DtEP] = 1400 ppm 135CC, 6h Then 260 °C, 0.5 h Then 260 °C, 1 h Then 260 °C, 2h

Clear/colorless Clear/colorless Clear/colorless Clear/colorless

74 990 1480 1700

[DtBP] = 1000 ppm 135°C, 6h Then 240 °C, l0 min Then 240 °C, 20 min Then 240 °C, 30 min Then 240°C, Ih

Clear/colorless Clear/colorless Clear/colorless Clear/colorless Clear/colorless

210 690 890 940 920

264000 117000 2.3 165000 64000 2.6 169000 70000 2.4 154000 59000 2.6 164000 70000 2.3

APPROACHES TO LOW RESIDUAL POLYSTYRENE

3.1

85

ADDITION OF FRIEDEL-CRAFTS CATALYST TO MONOMER

It is extremely difficult to distribute trace levels of a catalyst equally in a viscous polymer melt. Therefore, the catalyst was added to the monomer. However, this also presents a problem because most Friedel-Crafts catalysts also initiate cationic polymerization of styrene. The weight-average molecular weight (M w ) range of most commercial PS products is between 180 000 and 350 000. The addition of a Friedel-Crafts catalyst to the polymerization process would certainly have a profound effect on the ability to make PS in this molecular weight range. This is because, unless cationic polymerization of styrene is carried out at sub-ambient temperatures, only oligomeric products are formed owing to the very high rate of termination. An interesting exception to this generalization makes it worthwhile to consider adding acids at the start of the process; e.g. Priddy and co-workers discovered that low concentrations (i.e. 10–500 ppm) of certain sulfonic acids (e.g. camphorsulfonic acid [39] and 2-sulfoethyl methacrylate [40]) catalyze the decomposition of the 'Mayo intermediate' that is involved in initiation of spontaneous styrene polymerization. Since the Mayo intermediate is also a powerful chain transfer agent [41], the reduction of its concentration during styrene polymerization results in a significant increase in polymer chain length. Accordingly, it seemed worthwhile to investigate the effect of adding small amounts of 2-sulfoethyl methacrylate (SEM) on residual styrene levels. It was hoped that the level of sulfonic acid needed to catalyze isomerization of the Mayo intermediate during the polymerization might also be enough to catalyze the Friedel-Crafts alkylation of PS by the residual styrene monomer, resulting in a significant decrease in the residual monomer concentration. However, there is a downside to destroying the Mayo intermediate because it enhances the rate of polymerization by reacting with styrene to form initiating radicals. Therefore, when the Mayo dimer is present, the styrene monomer conversion is higher at a given set of polymerization conditions. This effect will be apparent in some of the data to be presented. Polymerizations of styrene containing 50 ppm of SEM were conducted using the same conditions employed in the control experiments. The results (Table 4.2) indicate that the presence of SEM results in lower styrene conversions, making it impossible to make a good comparison of the effect (relative to the control) of the acid during subsequent high-temperature treatment. However, the data indicate that SEM does not act as a catalyst to reduce significantly the level of residual styrene in the polymer.

4

LATENT ACID CATALYSTS

A way to circumvent the problems encountered by adding acid catalysts to the polymerization feed (i.e. cationic polymerization and interaction with Mayo

86

D. B. PRIDDY

Table 4.2 Polymerization followed by heating styrene using DtBP as an initiator in the presence of 50 ppm of SEM Reaction conditions Polymer color Residual styrene (ppm)

Mv

Mn

PD

[DtBP]= 1400 ppm 135 °C, 6h Then 260 °C, 15min Then 260 °C, 30min Then 260 °C, 60min

Colorless Colorless Colorless Colorless

1387 1954 1538

255600 218 100 194 200 162000

130000 112000 102 200 81000

2.0 1.9 1.9 2.0

[DtBP] = 1000 ppm 135 °C, 6h Then 240 °C, 15min Then 240 °C, 30min Then 240 °C, 60min

Colorless Colorless Colorless Colorless

2222 2847 3121 2578

304000 144700 265000 120 500 249000 113200 224000 97400

2.1 2.2 2.2 2.3

435

dimer) while still gaining the advantage of achieving a molecular dispersion of the catalyst in the polymer is to add a latent acid catalyst to the polymerization feed. Thermally labile acid esters which should de-esterify at elevated temperatures to form the corresponding acid. 4.1

TOSYLATES

Ethyl tosylate and isopropyl tosylate were evaluated as precursors of /Holuenesulfonic acid [38]. It was presumed that these esters would decompose at elevated temperatures as shown in Scheme 4.4.

\ /

SO3-CH2CH3

>200°C

>200°C

Scheme 4.4

SO,-H

CH,=(

SO3-H + CH3-CH=CH2l

Formation of sulfonic acids by de-esterification at high temperature

The residual styrene contents in PS prepared by the polymerization of styrene in the presence of ethyl tosylate (1 mol%) initially increased upon heating to 260 °C but then decreased with time (Table 4.3). After 2h, residual styrene levels were reduced to an undetectable level. Isopropyl tosylate is much

87

APPROACHES TO LOW RESIDUAL POLYSTYRENE

Table 4.3 Polymerization followed by heating styrene with DtBP in the presence of tosylate esters [38]. Reproduced from Y. Kim, H. J. Harwood and D. B. Priddy, J. Appl. Polym. Sci., 83, 1786 (2002), with permission of John Wiley & Sons, Inc.

PD

Reaction conditions Polymer color Residual styrene (ppm) [DtBP]= 1400 ppm, 135°C, 6h Then 260°c, 0.5 h Then 260° c, I h Then 260 °C, 2h

[ethyl tosylate] == 1.0 mol% 360 Colorless 540 Colorless Light yellow 180 None detected Brown

239 700 113900 222900 110500 167 100 80000 160900 80900

[DtBP] = 1400 ppm, 135°C, 6h Then 260 °c, 0.5 h Then 260 ° c, I h Then 260°c, 2h

[isopropyl tosylate] = 1 .0 mol% Light yellow None detected None detected Light brown Brown None detected Brown None detected

273000 142600 1.9 286900 150200 1..9 231200 116400 2.0 258 200 133 900 1..9

[DtBP] = 1000 ppm, 135°C, 6h Then 260 °C, 10 min Then 260 °C, 20 min Then 260 °C, 30 min Then 260 °C, 1 h

[isopropyl tosylate] = 0.15mol% 83 Colorless 280 Light yellow Light yellow 380 460 Light yellow 500 Light yellow

286 500 114800 2.5 211 600 85500 2.5 205 800 83 500 2.5 196 100 77 800 2.5 193 000 76600 2.5

2,.1 2..0 2,.1 2,,0

more effective than ethyl tosylate. The level of residual styrene was undetectable in the PS at the end of the polymerization at 135 °C and remained undetectable after heating at 260 °C while polymer molecular weights were higher than those prepared in the control reaction. This indicates that isopropyl tosylate forms ptoluenesulfonic acid during the polymerization at 135°C. The decrease in molecular weight with heating time at elevated temperatures observed in the control reaction was not observed in this reaction. However, the polymer turned brown during the heating process. When the concentration of isopropyl tosylate in the styrene polymerization was reduced to 0.15 mol% with 1000 ppm DtBP initiator, residual styrene contents were higher than those obtained from a 1 mol% concentration of catalyst but were lower than those obtained in the control reaction. These results demonstrate that although low levels of sulfonic acids (i.e. < 100 ppm) do not act to catalyze the reduction of residual styrene in PS, high levels (i.e. > 1000 ppm) significantly reduce the level of residual styrene monomer in PS at 260 °C. Unfortunately, the high level of sulfonic acid required to remove most of the residual styrene leads to significant polymer discoloration. Furthermore, the high residual level of sulfonic acid left in PS may lead to mold corrosion and other problems.

88

5

D. B. PRIDDY

MONOMER REGENERATION UPON HEATING

It is well known that the PS backbone contains weak links and that fabrication of the polymer generally leads to an increase in the level of residual monomer. The nature of the weak links in PS has been an on-going debate for decades. Previous studies have revealed the following: 1. initiator-derived residues left in the polymer affect thermal stability [42,43]; 2. PS produced by anionic polymerization contains fewer weak links than PS prepared using free radical chemistry [44]; 3. styrene dimers formed during the free radical polymerization are thermally unstable, forming styrene monomer when the polymer is heated to >200 °C; 4. exclusion of oxygen from styrene prior to and during the free radical polymerization of styrene improves thermal stability; 5. the addition of phenolic antioxidants to PS improves its thermal stability [17]. One of the remaining issues still needing resolution is the contribution (to PS instability) of the head-to-head (H-H) links formed in the PS backbone by termination by radical coupling. Previous studies indicate that the H-H links may not be the weak links in PS [45,46]. However, the thermolysis studies were carried out at very high temperatures well above typical PS fabrication temperatures. Recently Priddy et al. prepared H-H PS and also head-to-tail (H–T) PS and compared their relative thermal stabilities [47]. The thermal degradation characteristics of the two polymers over a wide range of temperatures were compared using thermogravimetry. Kinetic plots [–In (wt%) vs time] for the degradation of both H-H and H-T PS are displayed in Figure 4.7. As may be noted, at the outset these plots are essentially identical for the two polymers (see Figure 4.8, which is an expansion of the boxed area at the start of Figure 4.7). The rate constants determined from the slopes of the plots in Figures 4.7 and 4.8 are 3.49 x 10-6 s-1 for H-H PS and 4.15 x 10-6 s-1 for H-T PS. Although these values are not identical they are similar and suggest that the initial degradation process is the same for the two polymers. Degradation of H-H PS is well behaved over the entire range of degradation and presumably reflects main-chain cleavage between H-H units. The same is apparently the case for H-T PS degradation during the initial stages of decomposition. However, the number of H-H units in the H-T PS is no greater than about one per chain, assuming that termination of polymerization occurs exclusively, or nearly exclusively, by radical coupling. After these units have been removed, degradation of the H-T polymer is much more rapid than that for the H-H polymer. The degradation for the H-T polymer is well behaved at this temperature and probably reflects chain unzipping to generate styrene monomer. When degradation of the two polymers is carried out at 320 °C, the kinetic plots deviate earlier, indicating that random events for decomposition of the H-T polymer

89

APPROACHES TO LOW RESIDUAL POLYSTYRENE 0.45 0.40.350.3 0.250.20.150.

HHPS > HTPS

0.05-1 0 0

20

40

60

80

100

120

140

-3

Time (sx 10

Figure 4.7 Kinetic plots for thermal degradation of H-H and H-T PS at 280 °C. The boxed area at the origin is expanded to produce Figure 4.8 0.014 0.012 0.01 ^

0.008 -

*HHPS • HTPS

5 0.006 0.004 0.002 0 0

0.5

1

1.5

2

2.5

-3

Time (sX 10 ) Figure 4.8 at 280°C

Early portion of kinetic plots for thermal degradation of H-H and H-T PS

begin to occur at this temperature. When the decomposition temperature is raised to 350 °C, the degradation characteristics of the H-H polymer remains well behaved. The degradation of the H–T polymer is more complex and is probably reflective of several different processes occurring simultaneously. The degradation of H–H PS is well-behaved at all temperatures utilized. The degradation is clearly first order over a range of temperatures for which the reaction was examined. The rate constants for degradation at the various temperatures are shown in Table 4.4. A plot of In(klT) versus 1/r, where ris the absolute temperature and k is the corresponding rate constant, permits the extraction of the enthalpy of activation of 50.5kcal/mol for this process.

90

D. B. PRIDDY

At 280 °C the degradation of H-T PS is well behaved beyond the initial decomposition. A rate constant of 7.52 x 10 -6 s -1 characterizes this process. If what is occurring here is chain unzipping to generate styrene monomer and it is uncomplicated by competing processes, this provides a good reflection of the facility of that process. It might be noted that this value is only marginal greater than that (4.15 x 10 - 6 s - 1 ) attributed to the cleavage of H-H linkages in this polymer. However, the one is largely complete before the other begins. Based on all the foregoing, it is possible to suggest that the initial event in the thermal degradation of H-T PS is the cleavage of H-H linkages largely present in the polymer as a consequence of polymerization termination by combination of propagating species. This early stage degradation is very similar to that observed for H-H PS over the entire range of degradation. After all the H-H units have been removed, more complex degradation occurs. At low temperature, i.e. 280 °C, the predominant process would seem to be chain unzipping to form monomer. At higher temperature (>300°C), degradation probably involves significant random chain scission. Evolved gas analysis for the degrading polymers was conducted by thermogravimetry-mass spectrometry (TG-MS) and thermogravimetry-gas chromatography-mass spectrometry (TG-GCMS). The H-H sample lost very little mass at 280 °C. However, evolution of some volatile fragments, albeit in very small amounts, was detected. It might be noted that the mixture of volatile products contains few components and no styrene monomer. The major component of the mixture is a C11H16 isomer, tentatively identified a s 3-methylbutylbenzene. A l l other components a r e The behavior of the H-T polymer stands in sharp contrast to that of the H-H polymer. The only volatile compound formed from initial degradation of H-T at 280 °C is styrene monomer. The results of the evolved gas analysis are fully consistent with those generated by thermogravimetry. They suggest that the thermal degradation of H-T PS, particularly at low temperature (<300 °C), is initiated at H-H linkages present in the polymer as a consequence of polymerization termination by radical coupling. The macroradicals formed then undergo sequential unzipping to evolve styrene monomer. At low temperature (<300 °C) the initial degradation event is clearly scission of H-H linkages. At 280 °C, both processes are first order with rate constants of 4.15 x 10 -6 and 7.52 x 10 - 6 s - 1 , respectively. At Table 4.4 Rate constants for the thermal degradation of H-H PS Temperature (°C)

k ( s - 1 ) x 105

280 320 350

0.153 3.02 30.9

APPROACHES TO LOW RESIDUAL POLYSTYRENE

91

higher temperatures, the degradation is more complex and involves random chain scission and subsequent transformations in addition to H-H scission.

6

EPILOG

It is likely that research will continue aimed at lowering the level of residual styrene monomer in PS. However, since PS is a high-volume commodity plastic and producers are struggling to minimize their production costs, any technology for lowering the residual monomer level must have low capital and operating costs. If, for example, the PS industry decided to change from free radicalbased process chemistry to anionic chemistry, the total conversion cost of monomer to polymer would increase significantly owing to the high costs of organometallic initiators and monomer and solvent purification (not required for free radical chemistry). Currently the market for PS undergoes substitution with polypropylene during high points in the PS pricing cycle. If the manufacturing cost of PS increases significantly, substitution pressure from polypropylene becomes greater. This factor must be kept in mind as companies consider implementation of new low residual PS technology.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Chen, C. C., Polym.-Plast. Technol. Eng., 1994, 33, 55. Carter, L. F., Plast. Eng. (N.Y.), 1996, 33, 469. McKenna, T. F., Rec. Prog. Genie Proced., 1995, 9, 19. Moore, E. R. and R. A. Hay, II, US Patent 4 952 672 (to Dow), 1990. Hay, R. A. and A. C. Dowell, US Patent 4940472 (to Dow), 1990. Ravindrassath, K. and R. A. Mashelkar, Chem. Eng. Sci., 1988, 43, 429. Yang, C. T., AIChE J., 1997, 43, 1874. Krupinski, S. and J. T. McQueen, Eur. Pat. Appl. 798314 (to Nova), 1997. Skilbeck, J. P., US Patent 5350813 (to Novacor), 1994. Fujimoto, S., US Patent 3 987 235 (to Dow ), 1976. Street, B., US Patent 3 536 787 (to Shell), 1970. Wang, N. H. and N. Hashimoto, German Patent 19 521 713 (to Japan Steel Works), 1996. Skilbeck, J. P., US Patent 5 380 822 (to Novacor), 1995. Yoo, H. Y. and C. D. Han, Polym. Process Eng., 1984, 2, 129. Albalak, R. J., Z. Tadmor, and Y. Talmon, Mater. Res. Soc. Symp. Proc., 1992, 237, 181. Li, T. L., I. C. Tung, and D. Yu, US Patent 5 468 429, 1995. Schwaben, H.-D., Eur. Pat. Appl. 678 550 (to BASF), 1995. Imayoshi, M., Japanese Patent 07 149 817 (to Asahi), 1995. Gomez, I. L. and E. F. Tokas, US Patent 4215024 (to Monsanto), 1980.

92

D. B. PRIDDY

20. Camerman, P. J. A., US Patent 4124655 (to Cosden), 1978. 21. Nagai, S., A. Ueda, and H. Ikezawa, Makromol. Chem., 1981, 182, 1669. 22. Farrar, R. C., D. L. Hartsock, and F. X. Mueller, US Patent 5 185 400 (to Phillips Petroleum), 1993. 23. Blakemore, P. N., British Patent 1 330 896 (to Monsanto), 1973 24. Tokas, E. F., US Patent 4221 905 (to Monsanto), 1980. 25. Overbeek, G. C. and Y. W. Smak, US Patent 5 292 660 (to ICI), 1994. 26. Warakomski, J. M., W. C. Pike, and R. A. Devries, J Appl. Polym. Sci., 2000, 78, 2008. 27. Mironenko, N. I. and A. G. Demidenko, Russian Patent 215 488 (to Izobret), 1968. 28. Khomutov, A. I., Russian Patent 328 116 (to Institute of Physical Chemistry), 1975. 29. Kelley, J. M., US Patent 5274029, 1993. 30. van der Goot, A. J. and L. P. B. M. Janssen, Adv. Polym. Technol., 1997, 16, 85. 31. Priddy, D. B., M. Pirc, and B. J. Meister, Polym. React. Eng., 1993, 1, 343. 32. Gausepohl, H., W. Fischer, W. Loth, and R. Thiele, German Patent 19618678 (to BASF), 1997. 33. Desbois, P., Macromol. Chem. Phys., 1999, 200, 621. 34. Menoret, S., Macromol. Chem. Phys., 2001, 202, 3219. 35. Stannett, V. T., ACS Polym. Prepr., 1987, 28, 301. 36. Derbyshire, R. L., Radial. Phys. Chem., 1979, 14, 333. 37. Washio, M. and M. Shiraishi, Japanese Patent 04072 331 (to Sumitomo), 1992. 38. Kim, Y., H. J. Harwood, and D. B. Priddy, J. Appl. Polym. Sci., 2002, 83, 1786. 39. Buzanowski, W. C., E. Shero and D. B. Priddy, Polymer, 1992, 33, 3055. 40. Priddy, D. and V. A. Dais, US Patent, 5962605 (to Dow), 1999. 41. Pryor, W. and J. Coco, Macromolecules, 1970, 3, 500. 42. Cameron, G. G., W. A. J. Bryce, and I. T. McWalter, Eur. Polym. J., 1984, 20, 563. 43. Kristina, J., G. Moad, and D. H. Solomon, Eur. Polym. J., 1989, 25, 767. 44. Chiantore, O., M. Guaita, and N. Grassie, Polym. Degrad. Stab., 1985, 12, 141. 45. Luederwald, I. and O. Vogl, Macromol. Chem., 1979, 180, 2295. 46. Inoue, H., M. Helbig, and O. Vogl, Macromolecules, 1977, 10, 1331. 47. Howell, B. A., Y. Gui and D. B. Priddy, ACS Polym. Prep., 2002, 43, 360

Process Modelling and Optimization of Styrene Polymerization J. GAO AND K. D. HUNGENBERG Polymer Technology and Process Development, BASF AG, Ludwigshafen, Germany

A. PENLIDIS Chemical Engineering, University of Waterloo, Waterloo, Ontario, Canada

1

INTRODUCTION

In polystyrene production, it is desirable to produce polystyrene with the best properties at the lowest cost. In a batch process, the productivity can be increased by decreasing the batch time while still maintaining the molecular weight distribution of the final polymer in a desired range. Optimization policies can be designed based on a good understanding of styrene reaction kinetics. Tieu et al. [1] reviewed optimization policies for polymerization reactors. The surveyed papers covering batch, semi-batch and continuous reactors, and five typical case studies are discussed (temperature programming and initiator policies). Related numerical aspects of the optimization problem were discussed in a subsequent paper by Tieu et al, [2]. Among optimization efforts in the last 20 years, noteworthy is the paper by Wu et al. [3]. They studied, both theoretically and experimentally, the bulk polymerization of styrene and derived optimal temperature policies which minimize the time to reach a desired final conversion and number-average molecular weight. A time saving of about 18% was realized when compared with isothermal operation. The molecular weight and polydispersity of the polymer obtained during the experimental verification were in good agreement with the predictions of a mathematical model. Modern Styrenic Polymers: Polystyrene and Styrenic Copolvmers. Edited by J. Scheirs and D. B. Priddy f; 2003 John Wiley & Sons Ltd

94

J. GAO ETAL.

More recent efforts (primarily at the simulation level) on the optimization of styrene-related systems include Cavalcanti and Pinto [4], suspension reactor for styrene-acrylonitrile, and Hwang et al. [5], thermal copolymerization in a continuously stirred tank reactor (CSTR). In order to develop a sound optimization policy, a good understanding of styrene polymerization kinetics is necessary. In the following section the general kinetic scheme of styrene homopolymerization is introduced. 1.1

GENERAL KINETIC SCHEME OF STYRENE HOMOPOLYMERIZATION

The basic reaction scheme for free-radical bulk/solution styrene homopolymerization is described below. A complete description of copolymerization kinetics involving styrene is not given here; however, the homopolymerization kinetic scheme can be easily extended to describe copolymerization using the pseudokinetic rate constant method [6]. Such practice has been used by many research groups [7-10] and has been used extensively for modelling of copolymerization involving styrene by Gao and Penlidis [11]. In this section, all rate constants are defined as chemically controlled, i.e. they are only a function of temperature.

1.1.1

Initiation (by Monofunctional Initiator) I

2RJ

Ro* + Af —>• R*

(1)

(2)

where I is the monofunctional chemical initiator, R0 is the primary radical from initiator decomposition, M is the monomer and R is a radical with a monomer unit attached. The rate of initiation through a monofunctional initiator RIM (mol/L s) is RIM = 2fkd[l]

(3)

A distinctive characteristic of styrene polymerization is its thermal selfinitiation at high temperatures (without the presence of a chemical initiator). The mechanism of styrene thermal initiation was first described by Mayo [12]. The kinetics of thermal initiation were described by Weickert and Thiele [13] as a second-order reaction, while Hui and Hamielec [14], Husain and Hamielec

PROCESS MODELLING OF STYRENE POLYMERIZATION

95

[15] and Ito [16] proposed that the thermal initiation is a third-order reaction which is more widely used today. The initiation rate is believed to be of third order with respect to monomer concentration. The kinetics of styrene thermal initiation are as follows: M + M, AH k^\

M + AH

(4)

M'(PhCHCH3) + A M+ M

-

M + Ai-

M + AH i M+AH=

(5) R1 R1

(6) (7)

dinner

(8)

trimer

(9)

where AH denotes the Diels-Alder adduct, M* and A* are initial monomeric and Diels-Alder adduct radicals generated in reaction (5), these radicals can further react with styrene monomer to start the polymer chain growth. The rate-determining step in styrene thermal initiation is reaction (5). The rate of thermal initiation RIT (mol/Ls) is expressed as RIT = 2ki{[M]3

(10)

The total rate of initiation RI (mol/Ls) is the summation of styrene thermal initiation and chemical initiator decomposition, i.e. RI = 2fkd[I] +

2ki[M]3

(11)

It is worth noting that the dimer and trimer generated in reactions (8) and (9) can react with polymeric radicals as a chain transfer agent, and therefore their effect on the polymer molecular weight should not be neglected; the quantitative estimation of the concentration of these byproducts depends on the fact that whether the rate of thermal initiation is a second- or third-order reaction of monomer concentration. More kinetic information for such transfer reactions can be found in a number of publications [14–19]. Nevertheless, detailed kinetic studies on such Diels-Alder byproducts remain scarce. Katzenmayer [20], Olaj et al. [21,22], and Kirchner and Riederle [23] have published some quantitative results on this matter.

96

1.1.2

J. GAO ETAL

Initiation (by Bifunctional Initiator)

Bifunctional initiators have already been used in research laboratories and industrial production. They have the general structure Ri-O-O-R2-O-O-Ri, where R1 and R2 are hydrocarbon ligands. The oxygen bonds O—O, when heated, can thermally break and generate two radicals: R1—OO—R2—OO—R1 ^U R1—O* + R1—OO—R2—O*

(12)

The above reaction can be generalized as

I R0 +

R0

(13)

According to the bifunctional initiation mechanism, both decomposed initiator radical R0 and undecomposed initiator radical R0 participate in the polymerization process. This means that each initial initiator radical R0 or R0 may generate a different polymer chain:

R0 + R0 + M

M-^+R 1

(14)

^R1

(15)

R1 contains an undecomposed peroxide radical and is called a 'macro' peroxide. It can further decompose like an initiator: R1—OO—R2—O* -H R1—O- +• O—R2—O*

(16)

If the macro peroxide radical terminates with another polymeric radical, it will generate a dead polymer chain which contains one undecomposed peroxide, Pr [see reaction (24)]. If it terminates with another macro peroxide radical, a dead polymer with two undecomposed peroxides Pr will be generated [see reaction (25)]. These two decomposable dead polymers can undergo decomposition to generate radicals, hence they should be included in the initiation step: Pr^»R0 + p

r

*R

0

+R

Rr r

(17) (18)

The total rate of initiation (mol/Ls) involving bifunctional initiator as well as styrene thermal initiation RI can thus be expressed as

PROCESS MODELLING OF STYRENE POLYMERIZATION

R1 = 2k{[M]3 + 2fkdI[I] +f'kd2(Pr + 2P r )

97

(19)

In early work, Prisyazhnyuk and Ivanchev [24] studied the fundamentals of the polymerization using bifunctional initiators. Recent studies on modelling and simulation can be found in a number of publications [25–30]. Model testing results of styrene homopolymerization using an extensive list of homo- and bifunctional initiators can be found in Dhib et a/. [31]. 1.1.3

Propagation

The general propagation reaction is

Rr+1

R +M

(20)

and the propagation reaction for a macro peroxide radical is

Rr + M-^R; +I

(21)

where Rr is the macro polymeric radical with chain length r and Rr is the polymeric radical of chain length r. When a bifunctional initiator is used, the rate of propagation, which is also the rate of polymerization Rp (mol/Ls), is

1.1.4

Termination

As stated before, when a bifunctional initiator is used, a macro peroxide radical containing undecomposed peroxide is generated, which will lead to additional termination reactions (24) and (25), assuming that styrene polymeric radicals terminate only through combination: Rr* + Rs*

> Pr+s

Rr + Rs —> P r+s ~•

"•*-' •

k

^^

Rr + Rs —>• initiator present. The value of kt was recently measured by Buback et al. [32].

98

1.1.5

J. GAO ETAL

Chain Transfer Reactions

Transfer reactions to small molecules (T) including monomer, solvent, chain transfer agent (CTA), etc., are all very similar and are generalized as follows:

1.1.6

Rr + T-^P r + T*

(26)

R* + T -^ Pr + r

(27)

Reactions with Inhibitor R* + Z + P R; +

(28)

r

Z^Pr

(29)

where Z is a inhibitor. Detailed equations describing the rate of polymerization and molecular weight development can be found in Dhib et al. [31]. Yoon et al. [33] also calculated the molecular weight distribution in styrene batch thermal polymerization. 1.2

TREATMENT OF GEL EFFECT

The gel effect and limiting conversion are commonly observed in styrene homo-/copolymerization. Great advances have been made regarding polymerization kinetics at high conversion by different groups [34–39]. In a study by Vivaldo-Lima et al. [40], the models from Chiu et al. [39] and Marten and Hamielec [34] were compared. They concluded that the approach taken by Marten and Hamielec [34] based on free-volume theory seems to be able to describe successfully the whole course of polymerization up to the limiting conversion. Son and Sundberg [35-38] also used and refined further the freevolume theory to investigate the polymerization kinetics of styrene and obtained very good results. According to the free volume theory, all species in the reaction mixture occupy a certain amount of free volume. A small monomer molecule provides more free volume than a long-chain polymer molecule. The total amount of free volume of the system Vf is the summation of the free volumes provided by all species in the reaction mixture, and can calculated by the following equation:

PROCESS MODELLING OF STYRENE POLYMERIZATION

- ,°.\-n-(T -fOMJ

T -\\-Jrr

— J-e.i)\

99

nrvi

V-JW

Based on the free volume theory, Ff,/ and a, (thermal expansion coefficient) are assumed to be constant for different molecules. Vfj° has a value of 0.025 and <x.j has a value of 0.001 for small molecules such as solvent or monomer and 0.00048 for polymer long chains. The subscript j stands for species in the reaction mixture such as monomer, solvent, polymer, etc.. After polymerization has started, the free volume of the reaction system will continuously decrease. The smaller the amount of free volume, the lower is the mobility of the long-chain molecules. At a certain point, the rate constant of termination becomes diffusion-controlled and autoacceleration takes place. The diffusion-controlled rate constant kt will become smaller than the original chemically controlled rate constant kt. To keep track of the diffusion-controlled kt, one can use the following equation: \

P ~A

ex

(\

~

As the reaction continues, the system becomes more viscous and hence the free volume of the system decreases further. Depending on the glass transition temperature of the reaction mixture, when the free volume is less than a critical value the rate constant of propagation kp will also become diffusion controlled and it could be calculated by the following equation:

(32) By using the free-volume theory, one can successfully monitor the whole course of free-radical polymerization. It is commonly accepted that the presence of autoacceleration and limiting conversion in the polymerization reaction are due to the diffusion-controlled kt and kp, respectively. The initiation efficiency, f, will behave in a way similar to kp. 1.2.1

'Reaction Diffusion' Control Model

In the very late stages of the polymerization, as the reaction mixture becomes too viscous for polymeric radicals to move, the radical center can only move by addition of monomer molecules via a propagation reaction. The so-called 'reaction diffusion' controlled kt at this stage can be estimated by the equation proposed by Stickler et al. [41], and it is considered to be proportional to the rate constant of propagation and monomer concentration.

100

J. GAO ETAL

The 'reaction diffusion' regime was further clarified by Russell et al. [42] According to their model, the actual 'residual termination' rate constant lie! between two limiting values, a minimum, corresponding to a rigid chain, suet as polystyrene, and a maximum, corresponding to a flexible chain. It has beer found that the expression of the reaction diffusion controlled kt from Stickleretal. al. [41] is the same as the minimum value proposed by Russell et al. [42]. Both approaches share some common characteristics. Reaction diffusion control plays an important role in styrene homopolymerization since it is the main method of termination in later stages of the polymerization. It must be stated that up to now models that can describe detailed styrene polymerization including all kinds of initiation step are rare. The work of Dhib el al. [31] is so far the most comprehensive in this respect. It is a common practice to fit the model to experimental data under different reaction process conditions.

2 2.1

PROCESS SIMULATION AND OPTIMIZATION OF STYRENE HOMOPOLYMERIZATION USING INITIATOR COMBINATIONS WITH DESIGNED TEMPERATURE PROFILE

There are several ways to minimize batch time. One can either raise the reaction temperature or simply add more initiator to the system. However, both methods will reduce the molecular weight. High levels of expensive initiators result in lower molecular weight and leave a larger amount of undecomposed initiator, which is costly to remove and also an environmental hazard. Using more initiators also increases the product cost. Hence it is preferable to run the reaction initiated by the right type of initiator or even multiple initiators at a temperature or a customized temperature profile to produce the desired polymer at low cost. Butala et al. [43] applied optimization techniques to styrene polymerization initiated by BPO (dibenzoyl peroxide, 1 h half-life time, 91°C) and TBPB (tertbutyl perbenzoate, 1 h half-life time, 124°C). As mentioned before, the batch time can be minimized by using nonisothermal temperature profiles. Three independent runs with different optimization policies were performed. The detailed control policies are listed in Table 5.1. Figures 5.1-5.3 show the temperature policies in the three runs. It can be seen from the results in Table 5.1 that by using the right combination of initiators combined with a rising temperature profile, one can reduce the batch time significantly while at the same time increasing the molecular weight (comparison of policies 3 and 1). The final residual initiator amount is also greatly reduced. Their experimental results confirm that the optimization goal was achieved by employing the proper control policy.

PROCESS MODELLING OF STYRENE POLYMERIZATION

101

Table 5.1 Optimization policies in styrene homopolymerization. Reprinted from J. Gao and A. Penlidis, J. Macromolecular ScL, Reviews in Macromol. Chem. and Phys., C36(2), 199(1996), by courtesy of Marcel Dekker, Inc.

BPO (mol%) TBPB (mol%) Temperature (°C) Batch time (h) Conversion Xn Residual initiator (ppm)

Policy 1

Policy 2

Policy 3

80 20 87-103 6.04 98.4 780 680

69.2 30.8 94–107

39.4 60.6 119-133 3.04 98.4 930

5.13

98.5 782 660

380

375-

370-

365-

experimental temperature profile simulation temperature profile 360

0

100

200 300 Time (min)

400

500

Figure 5.1 Temperature profile used in optimization policy 1. Reprinted from J. Gao and A. Penlidis, J. Macromolecular Sci., Reviews in Macromol. Chem. and Phys., C36(2), 199(1996), by courtesy of Marcel Dekker, Inc.

2.2

USING BIFUNCTIONAL

INITIATORS

The success of obtaining polystyrene products by free radical processes is affected to a significant extent by the quality and performance of initiators. Monofunctional initiators such as benzoyl peroxide or azobisisobutyronitrile have been utilized in bulk and solution polymerizations for theoretical studies

102

J. GAO ET AL. 385

380-

experimental temperature simulation temperature

365 100

200 Time (min)

300

400

Figure 5.2 Temperature profile used in optimization policy 2. Reprinted from J. Gao and A. Penlidis, J. Macromolecular Sci., Reviews in Macromol. Chem. and Phys., C36(2), 199(1996), by courtesy of Marcel Dekker, Inc. 410

experimental temperature profile simulation temperature profile 390 100 150 Time (min)

200

250

Figure 5.3 Temperature profile used in optimization policy 3. Reprinted from J. Gao and A. Penlidis, J. Macromolecular Sci., Reviews in Macromol. Chem. and Phys., C36(2), 199(1996), by courtesy of Marcel Dekker, Inc.

PROCESS MODELLING OF STYRENE POLYMERIZATION

103

and practical applications. Gao and Penlidis [44] reviewed the wide range of research carried out on polymerizations of various monomers, including styrene. They incorporated most previous work on monofunctionally initiated free radical polymerization in a comprehensive simulation package. The market requirements regarding the properties and varieties of polymer grades are very specific. Consequently, there is a growing need to produce polymers with certain quality and to seek processes with high rates of polymerization and controlled polymer molecular weights. Organic peroxides, existing in large numbers, are widely used in the polymerization industry. It has been demonstrated that peroxides play a major role in controlling molecular weights, polymer properties and rate of polymerization. [45] For economic and safety reasons, the selection of the appropriate (initiator) peroxide is of great importance. Polystyrene reactors are usually designed for producing polystyrene in the temperature range 90–180 °C. If the peroxide decomposes at a very slow rate, the overall reactor efficiency becomes poor, and undecomposed peroxides may exit the reactor. Recycling impurities to the reactor causes operation problems that are not easy to overcome even by means of good control schemes. On the other hand, if the decomposition is very fast, a runaway polymerization is likely to occur. The increase in polymerization rate observed when using bifunctional initiators is one of the main motivations for both industry and academia to understand peroxide chemistry better. In general, an increase in reaction temperature for boosting productivity lowers polymer molecular weights. This problem of obtaining lower molecular weight averages can be circumvented by the use of a proper bifunctional initiator. An extensive review of bulk/ solution polymerization of styrene with both mono- and bifunctional initiators, with literature and industrial (Elf Atochem) data, was recently published by Dhib et al. [31]. Over 80 polystyrene-related references were discussed along with a general and flexible mathematical model. Additional literature sources can be found in Krispin et al. [46] and Cavin et al. [47]. An optimization exercise, similar to that of Wu et al., [3] but now with a bifunctional initiator, Luperox 101 [2,5-dimethyl-2,5-bis (tert-butylperoxy)hexane, 0.002 mol/L], would yield similar beneficial results from temperature programming. The designed temperature programming profile is displayed in Figure 5.4. One can obtain shorter batch times (productivity enhancements, see Figure 5.5) and stable molecular weight averages (product quality improvements, see Figure 5.6), compared to typical isothermal operation. Recently, tetrafunctional initiators have also been introduced for styrenics. In 2001, Atofina Chemicals introduced a novel tetrafunctional initiator, Luperox JWEB50, developed specifically for the styrenics industry to produce high molecular weight, high-heat, crystal polystyrene with improved productivity in a cost-effective manner. JWEB50 is a room temperature stable, liquid peroxide with a half-life similar to those of currently used cyclic perketals, appropriate for use in mass polystyrene processes. A unique aspect of

104

J. GAO ET AL. 170.0

160.0-

u 150.0 g. 140.0 u

o

130.0-

1 oi

— Isothermal Operation

120.0-

- - • Temperature Programming Operation 110.0

10

20

30

40 50 60 Time, t(min)

70

80

90

100

Figure 5.4 Reactor temperature profiles for the bulk polymerization of styrene initiated by Luperox 101 (0.002 M) for isothermal and temperature programming operation

Isothermal Operation - - - Temperature Programming Operation 10

40

50 60 Time, t(min)

Figure 5.5 Conversion as a function of time for the bulk polymerization of styrene initiated by Luperox 101 (0.002 M) for isothermal operation at 140 °C and temperature programming operation

105

PROCESS MODELLING OF STYRENE POLYMERIZATION 325 000300 000•

Weight Average Molecular Weight, M

Isothermal Operation - - - Temperature Programming Operation

•J3 200000-

jg

"3

175000-

S

150 000- - Number Average Molecular Weight, Mn

i

125 000

0.00

0.10

0.20

0.30

0.40 0.50 0.60 Conversion, X

0.70

0.80

0.90

1.00

Figure 5.6 Molecular weight averages for the bulk polymerization of styrene initiated by Luperox 101 (0.002 M) for isothermal operation at 140°C and temperature programming operation

JWEB50 is its ability to produce long-chain branching, in addition to providing higher molecular weight resins than those obtainable from a bifunctional initiator alone. Krupinski [48] used a combination of thermal and tetrafunctional peroxide initiation to produce polystyrenics with improved melt strength. An extensive (and the first academic) kinetic evaluation of JWEB50 (with studies of temperature, initiator concentration and initiator functionality effects) was given by Fityani-Trimm and Penlidis [49].

2.3

USING REACTOR COMBINATIONS

It is known that general-purpose polystyrene (GPPS) can be produced in a CSTR reactor. However, owing to the limitation of mixing and heat removal capacity of a CSTR, polystyrene can only be produced at relative low temperatures and conversion levels. Efforts have been made for many years to increase the heat removal capacity, e.g. by the use of an additional cooling coil, but such apparatus together with a predetermined temperature profile may lead to undesirable product properties. The high viscosity of polystyrene also presents a processing problem for a CSTR reactor. Even though it is common industrial practice to add a small percentage of a solvent such as toluene or ethylbenzene to the reaction mixture, polystyrene still cannot be produced in a CSTR up to

106

J. GAO ET AL.

high conversion. In order to increase productivity and reduce residual monomer, it is highly desirable to achieve high conversion. An alternative is to combine the CSTR with a plug flow reactor or, sometimes called, a tower reactor. CSTR is suitable for polymerization at low conversion level where the viscosity is relatively low. The reaction mixture is subsequently transferred to a plug flow reactor. The unagitated plug flow reactor can have several temperature zones and additional reactant can be added to each zone. This provides greater flexibility in comparison with a CSTR. It is also common to use a loop reactor with a static mixer instead of a CSTR reactor. Such a loop reactor produces better mixing for more viscous reaction mixtures. A pilot plant which is similar to the process described above was used by Cavin et al. [50,51] in their work on online conversion monitoring using an ultrasound velocity measuring device. Figure 5.7 displays a sketch of such a pilot plant. I. G. Farbenindustrie in Germany implemented such a concept to produce polystyrene commercially in the 1930s. Two CSTRs in parallel followed by a plug flow reactor were used in their process. During World War II, Union Carbide applied for a patent (US Patent 2496653, 1950) for a continuous polystyrene process. Their process consisted of three cascade CSTR reactors followed by a plug flow reactor. The temperature in the three CSTR reactors is 100, 115–120 and 140 °C, respectively. The conversion at the outflow of the third CSTR reactor is around 85 %. The temperature in the plug flow reactor is between 210 and 215°C. The final conversion at the plug flow reactor was claimed to be 97 %. loop reactor

Monomer feed

Q

degassing unit

Initiator feed

solvent feed

ultrasound probe Static mixer Figure 5.7 Pilot plant (a loop reactor and a plug flow reactor) for styrene polymerization. Reprinted from Cavin, Meyer and Renken, Polymer Reaction Engineering 8(3), 201(2000), by courtesy of Marcel Dekker, Inc.

PROCESS MODELLING OF STYRENE POLYMERIZATION

107

Other chemical companies have also designed their own continuous process to produce high-impact polystyrene (HIPS), such as the Dow process, which consists of three elongated reactors in series (US Patent 2727884, 1955); the BASF process, which consists of a prepolymerization CSTR followed by cascade of three CSTRs (US Patent 3658946, 1972); the Shell process, which consists of three CSTRs followed by a plug flow reactor (US Patent 4011 284, 1977); and the Monsanto process, which consists of a CSTR followed by a horizontal plug flow reactor (US Patent 3 903 202, 1975). More recently, a patent granted to BASF (German Patent 4236058, 1994) claimed that using a CSTR followed by an adiabatic plug flow reactor can have certain advantages such as higher space-time yield, lower energy consumption, and lower investment and maintenance costs. Under conditions suggested in the patent, i.e. 94 wt% styrene and 6 wt% ethylbenzene in the feed, the CSTR has a 2 h residence time, the temperature is 167 °C, the conversion at the exit of the CSTR reactor is 63 % and the weight-average degree of polymerization is 2480. The plug flow reactor is operated adiabatically with temperature at the exit reaching as high as 237 °C. The final conversion is about 84% with a weight-average degree of polymerization of 2280.

3

CONCLUSION

It is known that for a commodity product such as polystyrene, the production costs can be reduced by optimizing the process and polystyrene with desired properties can be produced by regulating the reaction conditions. This chapter has discussed briefly the kinetics of styrene polymerization and several commonly used process optimization policies. It has been demonstrated that by using a temperature profile combined with selective mono- or bifunctional initiators, one can decrease the bach time while still maintaining the desired molecular weight averages. The more recently developed tetrafunctional initiator by Elf-Atochem provides a new tool for control/optimization purposes. Additional optimization methods not discussed in this chapter are related to advanced online control theory combined with a kinetic model to control the polystyrene process. Mankar et al. [52] gave a review of online optimizing control of polymerization reactors.

4 A

SYMBOLS a free volume parameter estimated from experimental data [Equation

J. GAO ET AL.

108

B / /' kd kd\ kd2 ki kp kpQ kt kt0 kft kz Mw Mwcr n T Tg,j V Vj Vf KfCT Ffcrp Ffjo Xn [/]

a free volume parameter estimated from experimental results [Equation (32)] initiator decomposition efficiency bifunctional initiator decomposition efficiency rate constant for decomposition (s -1 ) rate constant for decomposition of bifunctional initiator (s -1 ) rate constant for decomposition of macro peroxide initiator (s - ') thermal initiation efficiency (L3/mol3 s) diffusion-controlled rate constant for propagation (L/mol s) chemically controlled rate constant for propagation (L/mol s) diffusion-controlled rate constant for termination (L/mols) chemically controlled rate constant for termination (L/mols) rate constant for chain transfer (L/mols) rate constant for inhibition (L/mols) accumulated weight-average molecular weight acccumulated critical weight-average molecular weight when kt becomes diffusion controlled a parameter that has a universal value of 1.75 reaction temperature glass transition temperature of species j total volume of the reaction mixture volume of the species j free volume of the system critical free volume of the system when &t becomes diffusion controlled critical free volume when kp becomes diffusion controlled free volume of species j at its glass transition temperature number-average degree of polymerization concentration of species j (mol/L)

REFERENCES 1. 2. 3. 4. 5. 6.

Tieu, D., Cluett, W. R. and Penlidis, A., Polym. React. Eng. J., 2, 275 (1994). Tieu, D., Cluett, W. R. and Penlidis, A., Comput. Chem. Eng., 19, 375 (1995). Wu, G. Z. A., Denton, L. A. and Laurence, R. L., Polym. Eng. Sci., 22, 1 (1982). Cavalcanti, M. J. R. and Pinto, J. C, J. Appl. Polym. Sci., 65, 1683 (1997). Hwang, W. H., Chey, J. I. and Rhee, H. K., J. Appl. Polym. Sci., 67, 921 (1998). Hamielec, A. E. and MacGregor, J. F., in Polymer Reaction Engineering, ed. Reichert K. H. and Geiseler W., Hanser, New York, p. 21 (1983). 7. Broadhead, T. D., Hamielec, A. E. and MacGregor, J. F., Makromol. Chem. Suppl., 10/11, 105 (1985). 8. Hamielec, A. E., MacGregor, J. F. and Penlidis A., Makromol. Chem. Macromol. Symp., 10/11, 521 (1987). 9. Xie, T. Y. and Hamielec, A. E., Macromol. Chem. Theory Simul., 2, 421 (1993).

PROCESS MODELLING OF STYRENE POLYMERIZATION

109

10. Dube, M. A., Scares, J. B. P., Penlidis, A. and Hamielec, A. E., Ind. Eng. Chem. Res., 36, 966 (1977). 11. Gao, J. and Penlidis, A., J. Macromol. Sci. Rev. Chem. Phys., C38, 651 (1998). 12. Mayo, F. R., J. Am. Chem. Soc., 90, 1289 (1968). 13. Weickert, G. and Thiele R., Plaste Kautschuk, 30, 432 (1983). 14. Hui, A. W. and Hamielec, A. E., J. Appl. Polym. Sci., 16, 749 (1972). 15. Husain, A. and Hamielec, A. E., J. Appl Polym. Sci., 22, 1207, (1978). 16. I to, K., Polym. J., 18, 877 (1986). 17. Pryor, W. A., ACS Symp. Ser., 69 (1978). 18. Pryor, W. A. and Lasswell, L. D., Advances in Free Radical Chemistry, Academic Press, New York, 1975. 19. Wiesner, J. and Mehnert, P., Chem. Ing. Tech., 45, 1269 (1974). 20. Katzenmayer T., PhD Thesis, Deutsche Gesellschaft fuer Chemisches Apparatewese, Chemische Technik und Biotechnologie, Frankfurt am Main, 1987. 21. Olaj, O. F., Kauffmann, H. F. and Breitenbach, J. W., Makromol. Chem., 177, 3065 (1976). 22. Olaj, O. F., Kauffmann, H. F. and Breitenbach, J. W., Makromol. Chem., 178, 2707 (1977). 23. Kirchner, K. and Riederle, K. Angew. Makromol. Chemi, 111, 1 (1983). 24. Prisyazhnyuk, A. I. and Ivanchev, S. S., Polym. Sci. USSR, 12, 514 (1970). 25. Choi, K. Y. and Lei, G. D., AIChEJ., 33, 2067 (1987). 26. Wittmer, P., Angew. Makromol. Chem., 170, 1 (1989). 27. Villalobos, M. A., Hamielec, A. E. and Wood, P. E., J. Appl. Polym. Sci., 42, 629 (1991). 28. Estenoz, D. A., Leal, G. P., Lopez, Y. R., Oliva H. M. and Meira, G. R., J. Appl. Polym. Sci., 62, 917 (1996). 29. Makwana, Y., Moudgalya, K. M. and Khakhar, D. V. Polym. Eng. Sci., 37, 1073 (1997). 30. Gonzalez, I. M., Meira, G. R. and Oliva, H. M., J. Appl. Polym. Sci., 59,1015 (1996). 31. Dhib, R., Gao, J. and Penlidis, A., Polym. React. Eng. J., 8, 299 (2000). 32. Buback, M., Kowollik, C., Kurz, C. and Wahl, A., Macromol. Chem. Phys., 201, 464 (2000). 33. Yoon, W. J., Ryu, J. H., Cheong, C. and Choi, K. Y., Macromol. Theory Simul.. 7, 327 (1998). 34. Marten, F. L. and Hamielec, A. E., J. Appl. Polym. Sci., 27, 489 (1982). 35. Soh, S. K., and Sundberg, D. C., J. Polym. Sci., Polym. Chem. Ed., 20, 1299 (1982a). 36. Soh, S. K., and Sundberg, D. C., J. Polym. Sci., Polym. Chem. Ed., 20,1315 (1982b). 37. Soh, S. K., and Sundberg, D. C., J. Polym. Sci., Polym. Chem. Ed., 20, 1331 (1982c). 38. Soh, S. K., and Sundberg, D. C., J. Polym. Sci., Polym. Chem. Ed., 20,1345 (1982d). 39. Chiu, W. Y., Carratt, G. M. and Soong, D. S., Macromolecules, 16, 348 (1983). 40. Vivaldo-Lima, E., Hamielec, A. and Wood P. E., Polym. React. Eng., 2, 17 (1994). 41. Stickler, M., Panke., D. and Hamielec, A. E., J. Polym. Sci., Polym. Chem. Ed., 22, 2243 (1984). 42. Russell, G. T., Napper, D. H., and Gilbert, R. G., Macromolecules., 21, 2133 (1988). 43. Butala, D. N., Liang, W. R. and Choi, K. Y., J. Appl. Polym. Sci., 44, 1759 (1992). 44. Gao, J. and Penlidis, A., J. Macromol. Sci. Rev. Macromol. Chem. Phvs., C36, 199, (1996). 45. Suyama, S., Ighigaki, H., Nakamura, T., Sugihara, Y., Kumura, H. and Watanabe, Y., Polym. J., 26, 273 (1994). 46. Krispin, I, Pauer, W., Fresen, I. and Moritz, H. U., DECHEMA Monograph, 5th International Workshop on Polymer Reaction Engineering, p. 259 (1995).

110

J. GAOETAL.

47. Cavin, L., Rouge, A., Meyer, Th. and Renken, A. Polymer, 41, 3925 (2000). 48. Krupinski, S. M., US Patent 6166099 (2000). 49. Fityani-Trimm, S. and Penlidis, A., Unpublished results from MASc thesis by Fityani-Trimm, S., Department of Chemical Engineering, University of Waterloo, Canada (2001). 50. Cavin, L., Meyer, Th. and Renken, A. Polymer, 8, 201 (2000b). 51. Cavin, L., Renken, A. and Meyer, Th., Polym. React. Eng. J., 8, 225 (2000c). 52. Mankar, R. B., Saraf, D. N. and Gupta, S. K., J. Polym. Eng., 18, 371 (1998).

Living Free Radical Polymerization of Styrene ALESSANDRO BUTTE, GIUSEPPE STORTI AND MASSIMO MORBIDELLI Laboratorium furTechnische Chemie, ETH Zurich, Zurich, Switzerland

1 INTRODUCTION Despite its many benefits and widespread use (about half of the total production of polymers in the United States in 1995 was carried out by this process [1]), a significant drawback of free radical polymerization (FRP) is the lack of control over the polymer structure due to the different termination mechanisms, mainly radical coupling and disproportionation. In FRP it is possible to control one among the different properties of the process or of the product, such as polymerization rate (by changing initiator concentration and/or temperature), degree of polymerization (by addition of chain transfer agents), end functionalities (by using specific functional initiators or transfer agents) and copolymer composition (by suitable monomer feed policies). However, it is not possible to control simultaneously all these properties and more complex structures and block copolymers are impractical. Therefore, it is not surprising that the development of a polymerization process able to meet such requirements, living free radical polymerization (LRP), has been a long-standing goal in polymer science. This is witnessed by the explosive expansion that the synthesis and the development of this technique have undergone since its first accomplishment, making LRP one of the leading topics in polymer engineering. In living polymerization processes it is in fact possible to adjust the final degree of polymerization by simply tuning the initiator amount while keeping narrow the chain length distribution (CLD), i.e. with polydispersity values lower than 1.3 (this quantity reflects the broadness of the final CLD; typical minimum values in FRP range from 1.5 to 2). Also, by suitable selection of chemistry and structure Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy ;r: 2003 John Wiley & Sons Ltd

112

A. BUTTE ET AL.

of the initiator, one can select the polymer chain-ends and block-polymerize multiple monomers. In other words, using only a few monomers, by living polymerization one can create an infinite range of new materials with vastly different chemical and mechanical properties by suitable design of the polymer structure, composition and chain-ends, as shown schematically in Figure 6.1 [2]. The goal of this chapter is to provide an overview of LRP, with emphasis on the applications to styrene homo- and copolymers. The chapter is structured as follows: a general introduction to LRP and the main mechanisms underlying the reaction is first presented. Then, we introduce some basics about the process kinetics when operated under homogenous (bulk/solution) and heterogeneous (emulsion/mini-emulsion) conditions. The main advantages/disadvantages are mentioned and corroborated by simplified modeling. Finally, the main applications of LRP to the production of highly homogeneous homo- and block copolymers involving styrene are reviewed.

2

LRP OVERVIEW

The only way to control the polymer structure properly during its synthesis is by a living process. In conventional FRP, in fact, bimolecular combination limits the chain lifetime to a small fraction of the entire process time and, therefore, changes in the operating conditions (monomer concentration and Structure

Linear

Star

Comb

Net

Composition

Home

Block

Random/ Gradient Alternating

Graft

Chain Ends

End-functional

X X Multifunctional

x

Figure 6.1 Schematic of the microstructure control potential offered by LRP (adapted from Ref. 2)

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

113

composition, viscosity, temperature, etc.) affect the structure of the polymer chains produced at different stages of the process. In a living polymerization, instead, these changes are equally distributed over all the polymer chains, which grow uniformly during the whole duration of the reaction. Moreover, the living chains being still able to restart propagating when the monomer is completely depleted, living polymerization represents the route to the production of block copolymers by simple addition of the comonomer. This is clearly not possible in FRP. So far, the only living processes industrially available were anionic and cationic polymerization [3], which generally surfer of little or no termination. In these processes, the initiation step is very fast compared with the process time, hence all the chains start to grow almost simultaneously. The degree of polymerization, DP, increases linearly with monomer conversion and it is inversely proportional to the initiator concentration. At the same time, Poisson-like distributions of the polymer chain length are obtained with final polydispersity values close to the ideal value of (1+1/DP). Finally, the polymer retains the ionic end groups until the end of the polymerization and, thus, the polymerization can be restarted by further addition of monomer. However, this kind of polymerization is often impractical from the industrial viewpoint, since it requires high degrees of purity of the reactants, very low temperatures and the use of solvents and, most of all, it is not applicable to several widely used monomers, such as styrene. On the other hand, FRP can be applied to a broad range of monomers (nearly all vinyl and vinylidene monomers) and requires mild reaction conditions [4]. FRP can in fact easily operate in the presence of impurities, such as residuals of inhibitor and oxygen traces, and over a wide temperature range. However, when dealing with FRP, it is not practical to approach living conditions by simply reducing the radical concentration so as to minimize the rate of bimolecular termination (second-order reaction with respect to radical concentration, the propagation being a first-order reaction). Besides the drawback of a corresponding reduction in the polymerization rate, this would lead to the production of polymer chains with extremely high molecular weight, since the instantaneous DP is given by the ratio between the frequencies of propagation and termination (kpM/ktR*, where kp and kt are the rate constants of the propagation and termination reactions, M is the monomer concentration and R* is the active chain concentration). Therefore, no control over the polymer microstructure is actually achieved. Using LRP, this control is instead restored by introducing an additional but reversible termination reaction with a 'capping' species, X. In this case, the instantaneous DP grown during a single active step is still given by the ratio between the rates of propagation and reversible termination (kpM/k*X, where k* is the rate constant of the reversible termination and X is the concentration of species X). If the rate of this new termination is so large as to be dominant with respect to that of the irreversible termination reactions and comparable to that of propagation, the

114

A. BUTTEETAL

polymer growth will be distributed all over the process. The main reactions involving the active chains are shown in Scheme 6.1. -X R-X —.

+RT

+X R* —-P

(J +M

Scheme 6.1 Simplified scheme of the reactions involving the active chains in LRP

Living polymerization is typically started by introducing in the system an 'initiator' providing the capping species, indicated as R-X in Scheme 1. This initiator can reactivate itself many times and add some monomer units before going back to the so-called dormant state in the form R-(M)n-X, where (M)M indicates a polymer chain made of n monomer units. In other words, the living process can be regarded as the insertion of a well-defined number of monomer units in between the groups R and X, which are always acting as polymer chainends and, thus, defined a priori. The average DP of the polymer is given by the ratio of the converted monomer, M0x (where x is the monomer conversion), to the total number of polymer chains. Note that some irreversibly terminated chains are produced, since these reactions are always taking place in competition with the activation/deactivation reactions, as shown in Scheme 6.1. If the fraction of terminated chains remains negligible compared with the initial amount of initiator, RX, this last value corresponds also to the concentration of the dormant chains in the system and the DP grows linearly with conversion (DP = M 0 x / R X } , as shown in Figure 6.2a. Moreover, if the number of active periods each chain experiences is large enough (i.e. if the number of monomer units added per active period is small enough), the polymer growth is distributed throughout the process duration and all the chains grow uniformly, leading to (i) a continuous reduction in the polydispersity value of the CLD (see Figure 6.2a) and (ii) a shift of the polymer peak with conversion (see Figure 6.2b). These features are usually considered a clear indication of the 'livingness' of the process. It is now worthwhile repeating that, even if the rate of the reversible termination is dominant with respect to that of the irreversible termination, terminations occur anyhow in the system to some extent. Some authors have therefore proposed to address the process as 'controlled' radical polymerization, being a true living polymerization a process without terminations. Since a certain degree of termination is found also in ionic polymerization [3b], we fully agree with the definition of living process given by Szwarc [3a]: 'a process where the polymers retain their ability to propagate for a long time and grow to a desired maximum size while their degree of termination is still negligible.' Therefore, the process will be simply addressed as living radical polymerization in the rest of this chapter.

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

115

80

60

40

20

20

(a)

Conversion (%)

40

l.0

100 (b)

1000

10000

100000

Molecular Weight (g/mol)

Figure 6.2 (a) Average DP and polydispersity value versus conversion for bulk living polymerization of styrene by RAFT; (b) corresponding gel permeation chromatogram. Reaction conditions: 7 = 80 °C; initiator, AIBN; RAFT agent, Z = Ph, R = CH(CH3)Ph (cf. Scheme 6.4); styrene: RAFT agent: initiator = 600:4:1 (w/w)

2.1

NITROXIDE-MEDIATED POLYMERIZATION

(NMP)

The ability of nitroxide stable radicals to react with carbon-centered radicals and to act as radical inhibitors [5a] has been known since the beginning of the 1980s, when Solomon and co-workers showed that the ability of nitroxides to react reversibly with growing polymer chains can be used to produce low-/)/* polymers [5b]. However, it was only in the 1990s with the work of Georges and co-workers [5c,d] that this novel polymerization technique, and in general LRP, received the attention it deserved. This living mechanism consists in the reversible combination of the growing radical chains, Rn*, and the so-called 'persistent radical species', X* (the nitroxide radical group), to form dormant polymer chains, Rn-X: R' + X*^ Rn-X

(1)

*a

The reaction can be initiated by using an alkoxyamine, R0X, where RQ is an alkyl and X a nitroxide group. As an alternative option, the polymerization can be also started by using a conventional chemical initiator (such as AIBN or BPO) in the presence of stable nitroxides [5e]. Today, many different routes are known which use different persistent radicals [5f–l]. Among these, TEMPO (1 in Scheme 2) is by far the most widely used, even though it suffers from very limited applicability to monomers different from styrene and high operating temperatures (about 120–140°C). Recent studies were aimed at reducing the operating temperature and broadening the monomer applicability so as to enlarge the range of block copolymers accessible by this technique (see Table 6.1). Despite these efforts, the application range remains limited.

116

A. BUTTE ET AL.

-Y

N-O

N-O.

N-O

,/\

EtO

OEt

Scheme 6.2 Examples of nitroxides: 1, TEMPO [5k]; 2, DDPO [5f]; 3, DEPN [5g,i,j]; 4, TIPNO [5h]

2.2

LRP BY ATRP

Atom transfer radical polymerization (ATRP) was first reported by Kato et al. [6a] and Wang et al. in 1995 [6b,c]. This mechanism is based on the so-called atom transfer radical addition reaction [7]. This reaction is catalyzed by a metal: the homolytic cleavage of the bond in an organic halide occurs through transfer of the halogen to the metal complex accompanied by oxidation of the metal atom. The catalytic cycle is closed by back transfer from the transition metal to the final adduct of the halogen. It is clear that, if the produced radical can undergo a few propagation steps before giving the back transfer and if this product is still able to undergo another transfer cycle, this reaction can be used to produce the same exchange between active and dormant state found in NMP. The resulting reversible reaction is R* + X-Me(w+1)/L ±i RW-X + Me(n)/L

(2)

where X indicates the halogen atom, Me the metal and Li the ligand. Table 6.1 Examples of living polymerizations by NMP using the nitroxides shown in Scheme 6.2 Monomer

Nitroxide Temperature (°C) Mn (g/mol) Polydispersity Ref.

Styrene Styrene Styrene Styrene Styrene n-Butyl acrylate n-Butyl acrylate n-Butyl acrylate-6/ocA:-styrene Acrylonitrile N,N-Dimethylacrylamide

1 2 3 4 4 3 4 4

130 130 120 120 85 120 120 120

12900 14600 16000 21 500 24000 17500 23 500 -50000

.16 .30 .11 .14 .13 .09 .09 .27

5f 5f 5j 5h 5h 5j 5h 51

4 4

120 120

55000 48 000

.13 .21

5h 5h

117

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

ATRP has been widely applied to the polymerization of styrene [2]. Nonetheless, ATRP owes most of its success to the large compatibility with many different monomers [6d], such as acrylates, methacrylates, (meth)acrylamides and acrylonitrile, which readily made this technique available for the production of several new block copolymers (see Table 6.2). Even though the majority of the work has been done with copper as transition metal, styrene ATRP has been carried out using Fe-, Ru-, Ni-, Pd- and Co-based systems [6d]. Note that ATRP does not require the high reaction temperature typical of NMP and this is also part of the success of this polymerization technique. Different ligands have been used to solubilize the copper atom (see Scheme 3) and it has been noticed that they not only make the copper ready for the reaction but can also modify the reactivity of the metal towards both the activation and deactivation reactions. Actually, the presence in the system of a metal, the need for complex ligands to solubilize them and the deep color typically imparted by this

Scheme 6.3 Examples of ligands used to carry out a Cu-based ATRP: 1, 2,2'bipyridine (bipy) [6b,c]; 2, 4,4'-di(5-nonyl)-2,2'-bipyridine (dNbipy) [6g,h]; 3, TPMA [6e]; 4, PMDETA [6f,i]

Table 6.2 Examples of living polymerizations by Cu-based ATRP using the ligands reported in Scheme 3 Monomer

Ligand

Temperature (°C)

Styrene Methyl acrylate Methyl methacrylate n-Butyl acrylateblock-styrenQ Styrene n- Butyl acrylate6/ocA>styrene Methyl acrylate Methyl methacrylate Acrylonitrile

TPMA TPMA TPMA PMDETA

110 50 50 60/100

8 600 15 200 10 500 9 500

1.07 1.05 1.12 1.14

6e 6e 6e 6f

PMDETA PMDETA

100 60/100

2 800 21 720

1.15 1.27

6f 6f

-18 000 -18 000 3 260

< 1.2 < 1.2 1.04

6i 6i 6j

PMDETA PMDETA bipy

90 90 44

Mn (g/mol)

Polydispersity

Ref.

118

A. BUTTE ET AL.

complex to the final polymer if not removed represent the major drawbacks of this process. Thus far, chlorine and bromine have been successfully used as halogen atoms, while iodine gives rise to side reactions [2]. Occasionally, the use of pseudo-halogen groups such as thiocyanates has been reported [2].

2.3

LRP BY DEGENERATIVE TRANSFER

As mentioned above, in both NMP and ATRP the exchange between the active and the dormant states is based on a reversible (although different) termination mechanism. Therefore, the exchange directly affects the radical concentration. In LRP by degenerative transfer, instead, this exchange is carried out by direct transfer of the w-end group between an active and a dormant chain. When an iodine atom is used as end group, the reaction can be expressed as follows:

R: + R W -I^>R^ + RW-I

0)

Therefore, the main difference from the previous two systems is that this living mechanism does not form new radicals and a conventional initiator is needed to start and 'sustain' the reaction. The initial amount of this species has to be properly selected. In fact, since the living reaction (3) does not affect the radical concentration, the final concentration of the chains terminated by bimolecular combination will be half of the initial concentration of the initiator. Therefore, the initial concentration of the species carrying the iodine group (in the following simply called the 'transfer agent') determines the final DP of the polymer provided that the initiator concentration is small compared with that of the transfer agent. Only a few papers have appeared dealing with LRP by DT [8], and the applications are almost completely limited to the homopolymerization of styrene. In this case, it was possible to obtain good control of the final CLD, with polydispersity values as low as 1.3-1.4. Better performances are difficult with styrene, mainly because of the limited transfer activity of the iodine atoms. This is the main reason for the very poor results obtained when applying this process to the polymerization of acrylates (e.g. n-butyl acrylate) and for the complete lack of control reported for other monomers [8].

2.4

LRP BY RAFT

The RAFT process can be regarded as a special case of degenerative transfer. As shown in Scheme 6.4, the reaction proceeds through the direct interaction of an active and a dormant chain with the formation of a reaction intermediate involving both chains [9a,b]. At this stage, the reaction can either go back,

119

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

•R"

X R»

/R "

O

R'

z

5=^

°

O

JL /\c

z

x

R'

b

"

I

^R'

A^c s

z

Z = aryl, alkyl, OR' , SR' R = CH(CH3)Ph, C(CH3)2Ph, C(CH3)2CN Scheme 6.4 Kinetics of the RAFT polymerization and examples of transfer agents commonly used in the literature

forming the initial radical again, or proceed forward, with the transfer of the Y=C(Z)Y moiety from the dormant to the active chain, which is now identified as the transfer species. The best results have been reported when using a sulfur atom as the Y group [9a-d], while less convincing results have been obtained by using CH2 [9e]. In fact, in the latter case the intermediate can easily propagate originating a branched chain and this will result in the broadening of the final CLD. Examples of sulfur-based transfer agents commonly used in RAFT polymerization are shown in Scheme 6.4. Satisfactory results have been obtained by RAFT polymerization of styrene, but the process is also effective for many other monomers, such as acrylates and methacrylates (see Table 6.3). As reported in the same table, the values of the operating temperature used to carry out this polymerization are close to those typical of conventional radical polymerization. The low temperature, along with the wide application range, place this mechanism among the most promising techniques to be applied on an industrial scale for the production of new materials, in competition with ATRP. Once again, a significant drawback is the need to remove from the product the sulfur atoms, which confer on the final polymer a deep color ranging from yellow to red. Table 6.3 Examples of living polymerizations by RAFT using the transfer agents reported in Scheme 6.4 Temperature

Conversion

Monomer Styrene 110 60 Methyl acrylate 80 Methyl methacrylate — Styrene–DMA — Ethylene oxide-styrene Styrene-block-n-butyl 110/60 acrylate-block-styrene Acrylic acid 60 60 Vinyl acetate

Mn (g/mol) Polydispersity Ref.

91 55 95 — — —

27 800 65400 59 300 43 000 7 800 161 500

1.09 1.06 1.13 1.24 1.07 1.16

9c 9c 9c 9d 9d 9c

53 96

66 800 22 700

1.13 1.24

9f 9f

120

3 3.1

A. BUTTE ET AL.

KINETICS OF LRP MAIN FEATURES OF THE DIFFERENT LRP PROCESSES

As already pointed out, the final aim of LRP is to control strictly the architecture of the polymer chain, i.e. properties such as degree of polymerization and monomer sequences. This is achieved by minimizing the fraction of dead chains in the system while having a uniform growth of the whole population of polymer chains. The degree of homogeneity in the polymer chain length is the key factor in homopolymerization and it is usually expressed in terms of polydispersity ratio, Pd. The fraction of dead chains becomes even more important when the process is intended to produce block copolymers, where terminated homopolymer chains represent a significant drawback with respect to the product quality. Actually to minimize terminations, different strategies are effective for the different living mechanisms, as briefly reviewed below. For the sake of clarity, let us start discussing NMP in bulk. As pointed out in Section 2, a successful LRP requires a deactivation reaction dominant with respect to bimolecular termination (cf. Scheme 6.1). Since these bimolecular reactions, deactivation and termination, are very fast and controlled by diffusion, it is worth showing how deactivation becomes the favored reaction path anyhow. Let us consider a batch reactor. For this system, the material balances of all species are readily written down and parametric simulations easily performed [10]. By solving such a set of equations while using parameter values typical of styrene homopolymerization in the presence of TEMPO at 125°C (ka = ID"3 s~\ fcd = 5 x 107 L/mol/s, kp = 2 x 103 L/mol/s, kt = 2 x 108 L/mol/s, MQ = 8mol/L,RX 0 = 4 x 10-2 mol/L), [10] the concentration profiles shown in Figure 6.3a and the reaction rates in Figure 6.3b were calculated. If a defined amount of RX is initially charged (usually a small molecule or oligomer carrying the nitroxide species), transient species, R*, and persistent species, X*, are produced in stoichiometric amounts by activation and their concentrations start growing at the same rate. At such a low conversion, in fact, the termination and deactivation rates are too small to compete with activation (see the short-time region in Figure 6.3a and b). Let us now suppose that bimolecular termination is not taking place and the only bimolecular event in the system is deactivation, i.e. the reversible recombination of R* and X*. As soon as the concentration of these two species grows, the deactivation becomes more and more important and equilibrium between activation and deactivation rates is finally established. This is clearly shown in Figure 6.3a (dashed line), where R* and X* are shown to reach a plateau at the same concentration value. However, in the presence of bimolecular termination, the concentration of R* is reduced by termination, thus resulting in the build-up of irreversibly terminated chains, Pc, and persistent radicals, X*. As the process goes on, the excess of X*

121

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

makes the deactivation more and more favored over bimolecular termination, the ratio between the two reaction rates, k^X*/ktR*, being larger and larger (see Figure 6.3a and 3b, solid lines). This phenomenon is usually referred to as the 'persistent radical effect' [11] and it has several important consequences. First, the lifetime of each polymer chain is much longer than that typical of conventional polymerization processes and, although bimolecular termination is always active, the concentration of the dormant chains will be much larger than that of the terminated chains. Second, the number of monomer units that every chain succeeds in adding during an active period (given by the ratio between propagation and deactivation rates, k p M / k d X * } decreases. Since more active periods are needed to reach the desired final degree of polymerization, the growth process becomes more homogeneous. This effect becomes evident when looking at the expression of the CLD polydispersity for the system without bimolecular termination:

PA =1 +

add

(4)

X

where DP0 = Mo/RXo and Nf^- = kpMQ/k&X*. For the specific polymerization under examination, a final polydispersity value of about 1.5 is calculated by Equation (4), while much smaller polydispersity values have been found experimentally under similar conditions (cf. Table 6.1). This result is even more significant by noting that the real process involves additional side reactions negatively affecting the product quality, such as thermal initiation. The theoretical overestimation of the final Pd is explained by the absence of the build-up of persistent

10-4

10"

1010-6

•2 10-

-8 10

10(a)

10°

102

Time (s)

104

10-

106

(b)

10-2

10°

102

104

Time (s)

Figure 6.3 (a) Concentrations of the species involved in the reaction and (b) reaction rates as a function of time. Kinetics in the absence (dashed lines) and presence (solid lines) of bimolecular termination. Concentrations: R* = active chains; X* = active chains; Pc = terminated chains. Reaction rates: Ra = activation; Rd = deactivation; ftt = bimolecular termination

122

A. BUTTE ET AL.

radicals when bimolecular terminations are not taking place: the concentration of X* remains very small and, therefore, the number of monomer units added per active period is large. Accordingly, each chain undergoes to just few active periods, making the polymer growth quite different from chain to chain. Fischer [1 la] derived equations for the concentration of radicals in the shortand long-time intervals of LRP (cf. Figure 6.3): R*=X* = ka(RX0)t

short-time

(5a)

1/3

r1/3

long-time

(5b)

Note that the duration of the first polymerization interval, when the rate of activation is far larger than the rate of all bimolecular processes, is usually a few seconds or less [t w (2kaktRXo)~l/2]. On the other hand, Equations (5a) and (5b) not only confirm that termination plays a key role in determining the kinetics of the process, but also indicate that the second interval of the reaction is characterized by the equilibrium between activation and deactivation, i.e. R* = kaRX/(kdX*). Although the living mechanism in ATRP takes place by a different reaction mechanism, the same arguments as presented above for NMP can be used. It is, in fact, clear that, after defining a pseudo-activation rate constant [A^ = k&Me(n\ cf. Equation (2)], the same equations as for NMP apply to ATRP. On the other hand, degenerative transfer and RAFT are characterized by completely different kinetics. In these two systems the persistent radical effect is totally absent and this makes the kinetics of the two processes identical with those of a conventional nonliving system. Accordingly, we have already pointed out that an initiator is necessary to sustain the reaction and the concentration of R* is set by the equilibrium between initiation and bimolecular termination. The radical concentration must be kept as small as possible to guarantee a final fraction of dead chains negligible with respect to the dormant polymer. At the same time, the rate of the transfer reaction must be kept as large as possible to ensure a large number of growth periods to all the dormant chains. In fact, Equation (4) is still valid in a system without terminations on the condition that Nfl£ is now defined as kpMo/kn(RX). 3.2

HOMOGENEOUS

HETEROGENEOUS LRP PROCESSES

Whatever the living mechanism, an essential requirement for a successful LRP is the minimization of the fraction of dead chains. In a bulk or solution reaction, the final amount of dead chains is a function of the radical concentration only: large polymerization rates correspond to high dead chain concentrations.

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

123

According to the same bulk model as mentioned above, it can be shown that the time needed to obtain 90% conversion after proper tuning of the process parameters so as to obtain a defined fraction of dead chains, , is given by [11 a]

where the factor C is different for the different living mechanisms (NMP/ ATRP: C = 4/3; RAFT/DT: C = 1). Using typical parameter values for styrene homopolymerization at 80 °C [10], reaction times of the order of l00 h are needed to have <£ = 0.05. Even if it can be shown that low polydispersity values can be achieved also at larger 0 values (> 0.2-0.3) [11], these values become unacceptable in copolymerization. While increased reaction temperatures and thus propagation rates generally should promote smaller dead chain contents [cf. Equation (6)], in the case of styrenic copolymers this leads to limited improvements only since undesirable side reactions negatively affecting the polymer quality (such as chain transfer and thermal initiation) become more and more important. Of course, when block copolymerizing different fastpropagating monomers, this problem may become less crucial. A possible solution to this problem comes from the heterogeneous processes, namely emulsion polymerization. It is well known that styrene polymerization in small colloidal particles follows the so-called zero-one kinetics [4b]: after being transported from the aqueous solution to a particle, the radical continues to propagate until a second one enters the same particle. At this point, an instantaneous termination between the two takes place and the reaction will restart in this particle only when another radical comes in. Neglecting the time period spent by the particle with two radicals, the particle spends half of its lifetime with either one or no radicals in it and the average number of radicals in the particle, i, is equal to 0.5 (or the average radical concentration is 0.5/V p N A , where vp is the particle volume and NA is Avogadro's number) whatever the frequency of entry, p. On the other hand, the rate of termination is independent of the radical concentration and is determined by the entry rate only. This is also the case for the instantaneous degree of polymerization, which is now given by kpM/p. Similar kinetics are readily obtained with DT and RAFT as living mechanisms. Since a transfer reaction is taking place in both cases, the same kinetics as in a nonliving process are established. Therefore, the fraction of dead chains in the system can be adjusted by properly tuning the frequency of entry, while the transfer reaction rate independently controls the homogeneity of polymer growth. In Figure 6.4a, the expected kinetics of the process are sketched in terms of number of active chains per particle. The picture is exactly the same as in a conventional nonliving zero-one system and, therefore, the high polymerization rates typical of this kind of heterogeneous system are achieved. On the

A.BUTTE ET AL.

124 i

ki

2-

2,

^

-

— > _L/fc

i/f,

1."P,

—.— fe. (a)

0

50

100

150

(b) 0

0.25

0.5

0.75

Figure 6.4 Expected kinetics for an emulsion RAFT/DT (a) and NMP/ATRP (b) i = number of radicals per particle

other hand, during the time spent with one radical, the degenerative transfer is working (about p/(kexRX) degenerative transfer events occur during this period), so that all polymer chains grow homogeneously. Finally, during each active period each chain adds an average of kpM/(knRX) monomer units, exactly as in the bulk case. The same kind of increased propagation rate is not found when using NMP or ATRP. Once again with reference to NMP for simplicity, it is not possible to have one radical per particle for a significant time period, since each activation event will produce a transient and a persistent radical: keeping in mind the extremely small size of a typical polymer particle produced in an emulsion, they recombine almost immediately, giving the kinetic behavior depicted in Figure 6.4b. Note that the average time that particles spend with zero and one radical is l/fa and l/fd, respectively, where fd = k&X* and/a = kaRX, the frequencies of activation and deactivation, respectively. This reduction in polymerization rate cannot be reversed by increasing the rate of the activation reaction: when fa approaches fd, a second activation is more likely to take place instead of a deactivation, the probability of this event being equal to fa/(fa +fd)- Accordingly, the two transient radicals terminate with each other and accumulate two persistent radicals in the particle, thus increasing the frequency of deactivation. In other words, each particle accumulates persistent radicals until the rate of the second activation becomes small enough (i.e. when fa «fd), and the average number of radicals per particle has a value much smaller than 0.5, that typical of DT and RAFT in emulsion. Note that when fa «fd, this average number of radicals can be readily estimated as/a//d. Actually, it has been shown that the process kinetics approach that of the corresponding bulk process, thus canceling the advantages of operating in emulsion from a kinetic point of view [12]. In addition to the kinetic considerations illustrated above, operating a living polymerization in emulsion is still the subject of research efforts by many groups [13]. The multi-phase environment complicates the global kinetics of the process and this is particularly evident in an ab initio emulsion polymeriza-

125

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

tion. The need to have the R-X species, generally very hydrophobia, inside the polymeric reaction locus requires a fast material transport of this species across the water phase out of the monomer droplets, where it is initially stored, to the polymer particles, where the reaction actually takes place. This transport must satisfy two fundamental requirements [13a]: (i) R-X must be readily available in the particles so that all the chains can start growing since the beginning of the process and (ii) R-X must be uniformly distributed among all particles. Among different alternatives, the best way to fulfill both requirements at the same time is to operate the process in miniemulsion [13a–d]. In this case, small monomer droplets are the primary locus of reaction. Since monomer and transfer agent are already there from the reaction start, the need for reactant interphase transfer vanishes. However, further difficulties have been reported, such as incomplete droplet nucleation [13a,b] and latex stability problems [13b]. Recent attempts to carry out LRP in standard and mini-emulsions are summarized in Table 6.4. Note that good quality styrene homo- and block copolymers have been produced in this way, thus confirming the potential of the process in emulsion.

4

APPLICATIONS TO STYRENIC POLYMERS

As already mentioned, the production of new polymeric materials represents the most attractive feature of LRP, in particular with reference to new block copolymers. A possible field of application is the production of amphiphilic Table 6.4 Examples of living processes carried out in ab initio emulsion (E) or miniemulsion (mE) polymerization using different living mechanisms Monomera

Sty MMA-block-Sty Sty-block-(BA-co-Sty) (2-Ethyl)hexyl methacrylate-Woc/:-

Mechanism Process

Temperature Mn (°C) (g/mol) Polydispersity Ref.

RAFT RAFT RAFT RAFT

mE mE mE mE

65/75

NMP NMP

mE mE E E E

80 70 70

37860 81 450 29 110 12 000

1.36 2.44 1.92 1.38

13a 13a 13a 13b

135 115 80 90 90

17396 -50000 35580 41 300 25500

1.12 1.27 1.38 1.57 1.16

13c 13d 9b 9b 13e

Sty

Sty BA-Wodt-(BA-fo-Sty)

Sty MMA-block-Sty BA-block-Sty

RAFT RAFT ATRP

Sty = styrene; MMA = methyl methacrylate; BA = n-butyl acrylate.

126

A. BUTTEETAL

polymers, i.e. polymers exhibiting emulsifier-like properties in aqueous solution, such as the ability to form micelles and to stabilize emulsion polymerization. Their production was particularly difficult with anionic and cationic living polymerization since the process cannot be carried out in aqueous solution. Thanks to LRP, a number of direct syntheses of different classes of amphiphilic polymers have been reported recently, where the polystyrene is the hydrophobic block and the hydrophilic block is acrylic acid [14a-c], styrene sulfonate [14d] and vinylbenzyltriethylammonium chloride [14c]. New classes of polymers can be also obtained with chain architectures impossible to obtain by FRP. Examples are star polymers [15a], which generally exhibit different hydrodynamic properties and higher degrees of chain end functionality compared with linear polymers of comparable composition, and graft copolymers with well defined structures [15b]. With reference to styrene polymerization, the production of molecular brushes, also termed bottle brushes, has been reported [15c]. These materials act as nanoscopic channels or worm-like micelles and they consist of a polymer backbone on which a copolymer is grafted, thus resulting in a core-shell cylindrical structure, the shell being made of rigid polystyrene and the core of a soft polymer (e.g. n-butyl acrylate), or vice versa. Another interesting application of LRP is represented by the possibility of introducing highly specialized functional groups at the polymer chain-ends, which can be polymerized directly without requiring any protecting agent [16a,b]. In particular, in ATRP a variety of functional groups have been introduced through the initiator, such as hydroxy, cyano, epoxy, allyl, vinyl, acetate, lactone and amide [16b-d]. It is also possible to start the polymerization using macroinitiators based on mono- and difunctional polymers, such as ethylene and butylene [16c]. Note that under these conditions the polymer is conserving the halogen end functionality, which may be subject to further transformation processes such as chain extension and end-group displacement reactions. Accordingly, it is possible to obtain polymers with different functionalities at the chain-ends [6d]. Finally, when dealing with emulsion systems, it is important to couple the control capabilities of LFRP with the control upon the particle size distribution and particle morphology. These, in fact, have a major impact on the properties of the final product. As an example of application, the block copolymerization of styrene with butyl acrylate and acetoacetoxyethyl methacrylate (AAEMA) has been reported [17]. The resulting polymer particles exhibit a core-shell structure, where styrene forms the rigid core. The possibility of cross-linking AAEMA units in the presence of rigid spots (styrene) chemically bonded to the soft phase is expected to enhance the mechanical tensile properties of the resulting polymeric films.

LIVING FREE RADICAL POLYMERIZATION OF STYRENE

127

REFERENCES 1. Chem. Eng. News, 1997, 75(25), 48. 2. Patten TE, Matyjaszewski K, Adv. Mater., 1998, 10, 901. 3. (a) Szwarc M, Nature (London), 1956, 178, 1168; (b) Szwarc M, J. Polym. Sci. A: Polym. Chem., 1998, 36(1), ix. 4. (a) Moad G, Solomon DH, The Chemistry of Free-Radical Polymerization, Pergarnon Press, Oxford, 1995; (b) Gilbert R. G, Emulsion Polymerization. A Mechanistic Approach, Academic Press, London, 1995. 5. (a) Moad G, Rizzardo E, Solomon DH, Polym. Bull, 1982, 6, 589; (b) Solomon DH, Rizzardo E, Cacioli P, US Patent 4581429, 1986; (c) Georges MK, Veregin RPN, Kazmaier PM, Hamer GK, Macromolecules, 1993,26,2987; (d) Georges MK, Veregin RPN, Kazmaier PM, Hamer GK, Saban M, Macromolecules 1994, 27, 7228; (e) MacLeod PJ, Veregin RPN, Odell PG, Georges MK, Macromolecules, 1997, 30, 2207; (f) Puts RD, Sogah DY, Macromolecules, 1996, 29, 3323; (g) Grimaldi S, Finet JP, Le Moigne F, Zeghdaoui A, Tordo P, Benoit D, Fontanille M, Gnanou Y, Macromolecules, 2000,33,1141; (h) Benoit D, Chaplinski V, Braslau R, Hawker CJ, J. Am. Chem. Soc, 1999,121, 5929; (i) Benoit D, Grimaldi S, Robin S, Finet JP, Tordo P, Gnanou Y, J. Am. Chem. Soc, 2000,122, 3904; (j) Benoit D, Grimaldi S, Finet JP, Tordo P, Fontanille M, Gnanou Y, ACS Symp. Ser., 1997, 685, Chapt. 14; (k) Greszta D, Matyjaszewski K, Macromolecules, 1996,29, 7661; (1) Farcet C, Charleux B, Pirri R, Macromolecules, 2001, 34, 3823. 6. (a) Kato M, Kamigaito M, Sawamoto M, Higashimura T, Macromolecules, 1995,28, 2093; (b) Wang JS, Matyjaszewski K, J. Am. Chem. Soc., 1995,117, 5614; (c) Wang JS, Matyjaszewski K, Macromolecules, 1995, 28, 7901; (d) Coessens V, Pintauer T, Matyjaszewski K, Prog. Polym. Sci., 2001, 26, 337; (e) Xia J, Matyjaszewski K, Macromolecules, 1999, 32, 2434; (f) Davis KA, Charleux B, Matyjaszewski K, J. Polym. Sci. A: Polym. Chem., 2000, 38, 2274; (g) Arehart SV, Matyjaszewski K, Macromolecules, 1999, 32, 2231; (h) Matyjaszewski K, Patten TE, Xia J, J. Am. Chem. Soc., 1997, 119, 674; (i) Xia J, Matyjaszewski K, Macromolecules, 1997, 30, 7697; (j) Matyjaszewski K, Jo SM, Paik HJ, Gaynor SG, Macromolecules, 1997, 30, 6398. 7. (a) Curran DP, Synthesis, 1988, 489; (b) Curran DP, in Comprehensive Organic Synthesis, vol. 4, ed. Semmelhack MF, Perfamon Press, Oxford, 1991, p.768. 8. (a) Matyjaszewski K, Gaynor SG, Wang JS, Macromolecules, 1995, 28, 2093; (b) Gaynor SG, Wang JS, Matyjaszewski K, Macromolecules, 1995, 28, 8051; (c) Goto A, Ohno K, Fukuda T, Macromolecules, 1998, 31, 2809. 9. (a) Goto A, Sato K, Tsujii Y, Fukuda T, Moad G, Rizzardo E, Thang SH, Macromolecules, 2001,34,402; (b) Moad G, Chiefari J, Chong BYK, Krstina J, Mayadunne RTA, Postma A, Rizzardo E, Thang SH, Polym. Int, 2000, 49, 993; (c) Mayadunne RTA, Rizzardo E, Chiefari J, Krstina J, Moad G, Postma A, Thang SH, Macromolecules, 2000, 33, 243; (d) Chong BYK, Le TPT, Moad G, Rizzardo E, Thang SH, Macromolecules, 1999, 32, 2071; (e) Krstina J, Moad G, Rizzardo E, Winzor CL, Berge CT, Fryd M, Macromolecules, 1995, 28, 5381; (0 Rizzardo E, Chiefari J, Mayadunne RTA, Moad G, Thang SH, ACS Symp. Seri, 2000, 768, 168. 10. Butte A, Storti G, Morbidelli M, Chem. Eng. Sci., 1999, 54, 3225. 11. (a) Fischer H, J. Polym. Sci. A: Polym. Chem., 1999, 37, 1885; (b) Fischer H, Macromolecules, 1991, 30, 5666. 12. Charleux B, Macromolecules, 2000, 33, 5358. 13. (a) Butte A, Storti G, Morbidelli, Macromolecules, 2001, 34, 5885; (b) de Brouwer H, Tsavalas JG, Schork JF, Monteiro MJ, Macromolecules, 2000, 33, 9239;

128

14.

15.

16.

17.

A. BUTTE ETAL

(c) Keoshkerian B, MacLeod PJ, Georges MK, Macromolecules, 2001,34, 3594; (d) Parcel C, Charleux B, Pirri R, Macromolecules, 2001, 34, 3823; (e) Matyjaszewski K, Shipp DA, Qiu J, Gaynor SG, Macromolecules, 2000, 33, 2296. (a) Burguiere C, Pascual S, Bui C, Vairon JP, Charleux B, Davis KA, Matyjaszewski K, Betremieux I, Macromolecules, 2001, 34, 4439; (b) Davis KA, Charleux B, Matyjaszewski K, J. Polym. Sci. A: Polym. Chem, 2000, 38, 2274; (c) Burguiere C, Pascual S, Coutin B, Polton A, Tardi M, Charleux B, Matyjaszewski K, Vairon JP, Macromol. Symp., 2000,150, 39; (d) Bouix M, Gouzi J, Charleux B, Vairon JP, Guinot P, Macromol. Rapid Commun., 1998, 19, 209. (a) Matyjaszewski K, Miller PJ, Pyun J, Kickelbick G, Diamanti S, Macromolecules, 1999, 32, 6526; (b) Shinoda H, Miller PJ, Matyjaszewski K, Macromolecules, 2001, 34, 3186; (c) Borner HG, Beers K, Matyjaszewski K, Sheiko SS, M oiler M, Macromolecules, 2001,34,4375. (a) Coessens V, Pyun J, Miller PJ, Gaynor SG, Matyjaszewski K, Macromol. Rapid Commun., 2000, 21, 103; (b) Gaynor SG, Matyjaszewski K, ACS Symp. Seri., 2000, 768, 347; (c) Matyjaszewski K, Teodorescu M, Miller PJ, Peterson ML, J. Polym. Sci. A: Polym. Chem., 2000, 38, 2440. Monteiro MJ, de Barbeyrac J, Macromolecules, 2001, 34, 4416.

Increasing Production Rates of High MW Polystyrene BRYAN MATTHEWS AND DUANE B. PRIDDY Dow Polystyrene R&D, Midland, Ml, USA

1

INTRODUCTION

Over 20 billion pounds of polystyrene are produced annually and most is produced utilizing continuous bulk free radical polymerization. Styrene polymers possess a combination of useful properties that, combined with their relative low cost, contribute to their widespread use in industry. Applications for polystyrene include electronics, medical, food packaging, optical, appliance, and automotive. These industries turn to polystyrene for its general features such as high modulus, hydrophobicity, chemical inertness, and ease of fabrication. Each application also has various specific requirements such as strength, heat resistance, and processability. A key parameter in achieving this variety in polystyrene properties is the average length of the polymer chains. In general, higher molecular weight (longer chains) will increase toughness properties but decrease processability by lowering the melt flow rate and increasing the required processing temperature/time. Styrene monomer will spontaneously or auto-polymerize and must be inhibited to prevent reaction during transport and storage. Polymerization is initiated by the generation of free radicals either by the reaction of the styrene with itself ('auto-initiation') or by means of a peroxide initiator ('chemical initiation'). Radicals rapidly propagate by reaction with monomer and ultimately terminate by coupling with another growing radical or by transferring the radical to a small molecule to start a new chain (chain transfer). Two key variables to be controlled in this very simple process are production rate and product molecular weight. Polystyrene is commercially produced in Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy 0 2003 John Wiley & Sons Ltd

130

B. MATTHEWS AND D. B. PRIDDY

very large-scale (generally >200001b/h) processes, so any improvement in reaction rate typically leads to a reduction in production cost. The rate is defined as the conversion of monomer to polymer and is essentially the same as the rate of propagation (monomer consumed by initiation and chain transfer are negligible as per the traditional 'long-chain assumption') [1]: production rate = propagation rate constant x radical concentration x monomer concentration Temperature is a key control parameter for adjusting production rate. Increasing the temperature increases both the propagation rate constant (according to an Arrhenius correlation) and the concentration of radicals due to increased rates of initiation. Another means of increasing rate is by the addition of chemical initiators to increase the radical concentration. Heat removal capabilities of the system provide a practical limitation to production rate; since the reaction is exothermic and the rate increases with temperature, the reaction will run away if the cooling system is inadequate. The other variable to be controlled is the molecular weight (MW) of the polymer. As indicated above, the MW determines several important product properties that are tightly specified by the customer for each application. One measure of MW is the 'number-average' molecular weight (A/n), the total weight of polymer divided by the number of polymer chains. Expanding on this definition, the instantaneous Mn of the polymer being made is determined by the rate of propagation divided by the sum of the termination and chain transfer rates: Mn(instantaneous) = MWstyrene [R P /(R t + /Rr)] As before, temperature increases all of the rates and the radical concentration (note that termination rate is proportional to the square of the radical concentration). The end result is that product MW is inversely proportional to temperature and therefore inversely proportional to production rate. MW can also be decreased by the addition of chain transfer agents or chemical initiators.

2 SPEEDING UP THE RATE OF POLYSTYRENE PRODUCTION USING CHEMICAL INITIATORS The inverse relationship between rate and MW traditionally presents a problem for the economic production of high MW polystyrene products owing to their slow production rates. The production rate of high MW products is generally increased by the use of peroxides. The addition of a simple monofunctional peroxide such as tert-butyl perbenzoate results in about a 15 % production rate increase over the use of auto-initiation. The use of difunctional peresters [2] and perketals [3] results in >30% rate increases over auto-initiation. However, these

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

131

initiators decompose sequentially, resulting in the formation of polystyrene containing peroxide functionality (Figure 7.1). When the second peroxide group homolytically cleaves to form a pair of radicals, the polystyrene bound radical is slow to diffuse from the other radical, resulting in some in-cage decomposition. The net result is that the 'diradical efficiency' of these types of difunctional initiators is low [4,5]. Theoretically, initiators that instantaneously form only diradicals should initiate 'double-ended' polystyrene chains that propagate from both ends simultaneously resulting in the formation of high MW polystyrene at very fast rates. This is because termination by radical coupling only leads to higher MW polystyrene that is still propagating on both of the chain-ends. However, the formation of diradicals in styrene that efficiently initiate polymerization has been a challenge [6]. Cyclic peroxides form diradicals upon thermolysis. However, their half-lives are much higher than those of their noncyclic homologues (due to recombination of the radicals) and they have very poor efficiency toward styrene initiation [7]. This poor efficiency is likely due to the fact that the radicals are bound to each other. Since they cannot get away from each other, they end up self-destructing before they escape each other and add to styrene. Even when styrene does add to one of the radical centers, efficiency is low because of ring closure of the new styrene adduct diradical (Figure 7.2). Dow researchers felt that a solution this problem would be to generate diradical species where the two radical centers are isolated from each other. They therefore generated /7-xylylene and p-phenylene diradicals (Figure 7.3) in

R-O—O—R'—O—O—R

»~

JR-O'

+ 'O—R'—O—O—R

Difunctional Peroxide S

ROPS-PSO—R' — O — O—R

~*

ROPS.

+

. PSO — R'—O—O—R

Polystyrene containing peroxide

ROPS-PSO —R t —OPS -PSO —R Figure 7.1 Simplistic depiction of styrene polymerization initiated using a difunctional peroxide (two peroxide groups sequentially decompose)

132

B. MATTHEWS AND D. B. PRIDDY

Ph

Figure 7.2 Depiction of the addition of a diradical species adding to styrene to form small cyclic molecule rather than polymer

600°C

paracyclophane

p-Xylylene

Poly-p-Xylylene

>100°C Bergman Cyclization Enediyne Figure 7.3

p-Phenylene Diradical

Poly-/?-Phenylene

Generation of diradicals that homopolymerize in styrene as solvent

styrene solution. However, in both cases the radicals only couple with each other, resulting in the formation of poly-/>-xylylene and poly-p-phenylene, respectively, with no inclusion of styrene units in the polymer. This total exclusion of styrene is intriguing and is likely due to electronic factors,

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

133 o

o

o

if

if

^

-O-CL/O

o

o-o-

0.0_

IV

Figure 7.4 Multifunctional peroxides that have been evaluated for the fast production of polystyrene

i.e., both /»-xylylene and /7-phenylene radicals and the styrene double bond are electronegative. When these diradicals are generated in an electropositive monomer, such as an acrylate, polymerization is initiated very efficiently [8]. Another type of initiator that has been evaluated for increasing polystyrene production rates are the multifunctional peroxides. Examples include 2,2-bis [4,4-bis(ter?-butylperoxy)cyclohexyl]propane (I) [9], peroxyfumaric acid, 0,0tert-butyl 0-butyl ester (II) [10], tert-butyl peritaconate (III) [11], and poly (monopercarbonates) (IV) (Figure 7.4) [12]. Although all of these initiators indeed show extremely fast production rates of high MW polystyrene, they all suffer from a flaw, i.e. the polystyrene produced is branched and special precautions must be taken to keep the continuous bulk polymerization reactors from fouling [13]. This is likely why none are currently used commercially for polystyrene manufacture.

3

SPEEDING UP THE RATE OF POLYSTYRENE PRODUCTION USING ACID MEDIATION

A very unique approach to increasing the production rate of high MW polystyrene was recently developed by Dow researchers. They discovered that the rateMW curve for auto-initiation polymerization of styrene can be significantly

134

B. MATTHEWS AND D. B. PRIDDY

shifted by performing the polymerization in the presence of an acid [14]. Figures 7.5 and 7.6 show the relationship between conversion rate (amount polymerized after 1 h) and 'weight-average' molecular weight (Mw). The addition of acid decreases production rate but greatly increases the Mw produced. By increasing the temperature, the rate can be recovered while maintaining a net increase in Mw. For example, production of an Mw 500000 polymer normally requires a temperature of around 113 °C and is thus limited to a conversion rate of 4%/h. By adding 250 ppm of 10-camphorsulfonic acid (CSA) and running at 140 °C, the rate can be increased to over 15%/h (Figure 7.5). To understand this effect requires a discussion of the mechanism of autoinitiation. Of the mechanisms proposed, two receive most of the discussion in the literature. A mechanism proposed by Flory [15] involves a 1,4-cyclobutane diradical intermediate and has been supported experimentally by polymerization in the presence of a free radical scavenger. Several of the dimers found in thermally initiated polystyrene are also attributed to a Flory initiation mechanism.

0.300

140°C

ppm CSA 100 ppm CSA

0.200

500 ppm CSA

0.100 -

0.000 300000

400000

500000

600000

700000

Mw

• With Acid

Data w/o Acid

Figure 7.5 Rate-/Ww curves for styrene polymerization after 1 hour with acid at 140°C and without acid at various temperatures. Reprinted from W. C. Buzanowski, J. D. Graham, D. B. Priddy and E. Shero, Polymer, 33, 3055(1992) with permission of Elsevier Science

135

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE 2500

2000-

- 1500I

1000-—a-.

500 4

6

10

Polymerization Rate (%/h)

Figure 7.6 Styrene polymerized both in acidic and neutral environments under a variety of conditions to generate rate-Mw curves

Mayo [16] proposed an alternative mechanism that is currently widely supported. Figure 7.7 shows a schematic of the Mayo mechanism. A Diels-Alder reaction between two styrene molecules produces an intermediate dimer (DH), also referred to as 'Mayo dimer'. DH is highly reactive and has never been isolated. To complete the auto-initiation, DH reacts with a third styrene molecule via 'molecular assisted homolysis' [17] to form a phenyltetraline radical (D*) and a phenethyl radical (SH"). A second reaction involving DH is to undergo chain transfer with a growing radical chain to produce a dead polymer chain (PS-H) and a new growing radical. The chain transfer constant (Kct) of DH has been estimated at 10, which is the highest Kci ever reported for a molecule that contains no heteroatoms [18,19]. Also indicated in Figure 7.7 is the possibility of acid-catalyzed aromatization of DH to an unreactive dimer (DA). Under neutral conditions, only traces of DA are found. However, when a small amount of CSA is added to styrene undergoing polymerization by auto-initiation, significant levels of DA are formed along with polystyrene of higher than expected MW. We believe this is strong support for the Mayo mechanism since acid would have little affect on the Flory diradical intermediate.

136

B. MATTHEWS AND D. B. PRIDDY Ph

DA

D-

P

Figure 7.7 Auto-initiation of styrene monomer by the Mayo mechanism and acidcatalyzed aromatization of the intermediate reactive dimer

Examining the mechanism, the effect of acid on styrene polymerization can be qualitatively explained. The acid very rapidly catalyzes aromatization of DH to form DA, thus significantly reducing the concentration of DH. This slows the radical-forming reaction and reduces the number of radicals available for propagation. Although the propagation rate constant is not affected by the presence of CSA, the actual rate of propagation drops owing to the reduced radical concentration. This is what causes the decrease in conversion rates observed in the presence of acid. The effect on MW is twofold. First, the same reduction in radical concentration that lowers propagation rate also reduces the rate of chain termination. Since chain termination is proportional to the square of radical concentration (propagation is directly proportional), the ratio of chain growth to chain termination is shifted in favor of higher molecular weight. Second, chain transfer to DH is believed to be significant in decreasing the MW of the polymer. The acid-catalyzed aromatization reaction reduces the concentration of DH available for chain transfer, thereby contributing to the MW increase. The aromatization reaction in this case is an example of homogeneous catalysis. The catalyst, reactant, and product are all dissolved in liquid styrene monomer. As such, traditional heterogenous catalysis considerations such as

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

137

surface area and pore diffusion are of less concern than factors such as solubility, acid strength, and concentration. The aromatization of DH is readily catalyzed by strong acids and is very fast. However, strong acids can have the undesirable effect of initiating polymerization via a cationic (rather than free radical) mechanism. Cationic polymerizations proceed rapidly and produce very low molecular weight products owing to fast termination by chain transfer to monomer. A key consideration then is that the acid must be strong enough to catalyze the aromatization without initiating cationic polymerization. In their work, Dow researchers tested five soluble organic acids (acetic, benzoic, pyruvic, and phenylphosphonic acids, in addition to CSA) having a variety of acid strengths (Figure 7.8) [14]. The effects on polymerization rate and Mw (140 °C/2h) were not significant for the weaker acids (pKa> 2). CSA, the strongest acid available that did not initiate cationic polymerization, had the most impact on the results and was selected as the catalyst for further study. As would be expected, the concentration of catalyst determines the degree to which the polymerization was affected. The data displayed in Figure 7.5 show how Mw increases with increasing acid levels. By choosing the strongest acceptable acid, the concentration required to achieve a desired effect can be minimized. The solubility characteristics of the acid are also important. Acids that are insoluble in styrene would have difficulty contacting the DH dimer to catalyze the aromatization reaction. Acids that are slow to go into solution could cause practical problems for the design and operation of the process. For their study, the authors chose only acids that readily dissolved in styrene. -- 420000 0.45--

-- 320000

0.35--

220000

0.25 4.5

Conversion

Mw

Figure 7.8 Effect of acid strength on styrene conversion (140 °C/2 h) [14]. Reprinted from W. C. Buzanowski, J. D. Graham, D. B. Priddy and E. Shero, Polymer, 33, 3055(1992) with permission of Elsevier Science

138

B. MATTHEWS AND D. B. PRIDDY

While the promise of greatly increased rates for producing high MW resins is exciting, there are naturally several practical considerations that must be considered when implementing the process in the real world. Among these are corrosion, oligomer levels, catalyst addition, and recovery or purge of acid after the reaction is complete. Certainly the introduction of acid to a previously acid-free process requires a fresh look at the materials of construction from a corrosion standpoint. In addition to the polymerization reactor, unit operations such as solvent devolatilization, condensation and storage for recycle are potentially affected, greatly increasing the number of vessels, exchangers, and piping involved. The relatively low concentration of acid used is a plus, but corrosion problems are aggravated by the presence of water impurities in the styrene and the high temperatures of the reaction and devolatilization processes. Exposure to acid could be minimized by introducing a neutralizing agent late in the reaction process (at the expense of adding yet another new component and its associated storage and feed control equipment to the process) [20]. Another concern is what to do with the acid catalyst at the end of the process. Typically, styrene polymerization is carried out to 50-90% conversion. The unreacted styrene is flashed from the polymer product, condensed, and recycled into the reactor feed. If the materials of construction allow, a volatile acid catalyst could also be continuously recycled through the system, theoretically reaching a point where no fresh catalyst feed would be required. If corrosion in the devolatilization and recycle system is a concern, then the neutralization option mentioned above might be considered. A solution to minimize corrosion potential and eliminate concerns about migration of acid residues left in the PS is to use a vinyl-functional acid that becomes locked into the polymer chain. An example is 2-sulfoethyl methacrylate (SEM). SEM has been found to be even more effective (~ l O x ) on a weight basis than CSA, i.e. 50 ppm of SEM gives about the same effect as 500 ppm of CSA [21]. Even at 10 ppm, SEM results in a significant effect [22]. A problem with SEM is that it is a viscous oil that is only sparingly soluble in styrene. However, the solubility of SEM in styrene can be greatly enhanced by diluting it with an equal volume of methacrylic acid (MAA), thus significantly facilitating dissolution [23]. The means of introducing the catalyst to a large continuous polystyrene process should also receive attention. The low concentrations (and resulting low flow rates) and sensitivity to concentration require an accurate metering system for the acid catalyst. Also of importance is how the acid is mixed into the stream, particularly if the acid is slow to dissolve. Poor mixing of acid could lead to local areas of high concentration, resulting in possible cationic polymerization and the unintended broadening of the molecular weight distribution curve. There is also a potential to create high molecular weight polymer gels that lead to a plant shut-down and clean-out.

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

139

Finally, there is the question of oligomer (dimer and trimer) levels in the product. Returning to the mechanism (Figure 7.7), we see that the desired effect of increased molecular weight is achieved by the catalyzed conversion of the reactive dimer DH to the unreactive dimer DA. This has an additional positive effect of reducing trimer levels in the product, but overall there is an increase in total oligomer level due to the high levels of DA formed (Figure 7.9). Oligomers can be both positive (acting as plasticizers to improve processability) and negative (decreasing strength properties, for example). These effects must be considered to determine whether the desired product is a candidate for rate increases using acid catalysis.

In the mid-1980s, Dow researcher Wesselmann [24] showed that for broad polydispersity (i.e. Mw/Mn > 2.5) polystyrene, the flow-strength balance could be improved by the addition of a few wt% of very high MW polystyrene chains (i.e. chains having MW > 106). The polystyrene could be described as having a bimodal molecular weight distribution (BMWD). Wesselmann prepared the BMWD polystyrene by warming unstabilized styrene at 60 °C for five days to form a syrup containing 10 % of polystyrene having an Mw of 1.4 x 106. This syrup was then fed to a continuous bulk polymerization reactor along with a peroxide initiator to produce BMWD. However, Dow did not commercialize 9000" 7000 - 5000 - 3000 1000 100

200

300

CSA cone, (ppm)

DA

Figure 7.9

Trimer

Effect of acid level on oligomer formation

B. MATTHEWS AND D. B. PRIDDY

140

Table 7.1 Results of styrene polymerization at 140°C in the presence of lOO ppm SEM and lOOOppm BPO

Polymerization time (h)

Conversion 38 55 62

2 4 6

I 12

I 16

I 12

I 16

M/w/1000

Mw/Mn

350 540 525

4.0 4.2 3.9

12

I 16

GPC Retention Time (min)

Figure 7.10 GPC curves of BMWD polystyrene produced at 140 °C in the presence of 100ppm SEM and lOOOppm BPO

BMWD at the time because of the very slow polymerization rate needed to produce the Afw1.4 x 106 polystyrene. However, in recent years, through implementation of acid mediation technology, Dow successfully commercialized BMWD polystyrene used for Styrofoam production [25]. Gel permeation chromatography (GPC) is a powerful tool for looking at the molecular weight distribution of polymers. Table 7.1 and Figure 7.10 show the evolution of the GPC during polymerization (at 140°C) of styrene containing lOOOppm benzoyl peroxide (BPO) and 100ppm SEM [26]. At 140CC, the BPO quickly decomposes, forming low MW polystyrene. Once the BPO is gone, only spontaneous polymerization takes place, which, in the presence of SEM, leads to the formation of ultra-high MW polystyrene. As the polymerization progresses, the ratio of the high MW component in the BMWD blend increases relative to the low MW component.

5

MODELING ACID-MEDIATED STYRENE POLYMERIZATION

In this part of the chapter, a model of acid-mediated styrene polymerization will be developed and discussed. The purpose of the modeling work was twofold: to provide further analysis to support or contradict the proposed mechanisms (Diels-Alder initiation, chain transfer to Mayo dimer, etc.);

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

141

• to provide a useful tool for analyzing, designing, and operating an acidcatalyzed styrene polymerization process.

5.1

STYRENE AUTO-INITIATION MODEL

The kinetic model of styrene auto-initiation proposed by Hui and Hameilec [27] was used as a starting point for this work. The Mayo initiation mechanism was assumed (Figure 7.2) but the acid reaction was of course omitted. After invoking the quasi-steady-state assumption (QSSA) to approximate the reactive dimer concentration, Hui and Hameilec used different simplifying assumptions to derive initiation rate equations that are second and third order in monomer concentration.

5.2

ACID MODEL DEVELOPMENT

Addition of acid to the process adds only one new reaction: DH + S — »DA Acid-catalyzed aromatization Rs = ksS • DH where S = acid concentration and DA = aromatized (unreactive) Mayo dimer; compound abbreviations in bold indicate concentrations. Unfortunately, this reaction requires that the concentration of DH be known. The addition of this reaction also negates the simplified initiation rate equation used in Hui and Hameilec's model [27]. One approach to modeling the acid-modified polymerization would be to start from scratch, re-do the QSSA on DH and try to find a new simplified initiation rate equation as a function of monomer and acid. The same would then have to be done for the rate of chain transfer to dimer. These new expressions would then need to be fitted to the experimental data. A (presumably) simpler approach is to estimate the change in DH concentration due to the addition of the acid: .

(h —

DH concentration with acid present DH concentration without acid

All of the other rate equations could then be expressed in terms of to estimate their rate in the presence of acid (quantities in the presence of acid will be denoted with a prime symbol):

142

B. MATTHEWS AND D. B. PRIDDY

Reaction 2M-+DH DH^2M DH + M -+ 2R DH + R -» R + P M + R -» R + P R + M -> R R 4- R -> P

Diels-Alder forward Diels-Alder reverse Initiation Transfer to dimer Transfer to monomer Propagation Termination tft

Rate without acid /?i = K\M2 R^ =K-\DH R, = k,M^ Rm = kmMR Rm = kmMR Rp = kpMR = A:t/?2

Rate with acid /?, = KXM2 /?_, = R'- = Ri R'm = R^ = 0 l/2 /? R' = 01/2/?p /?[ = ^/?,

In order to estimate , the QSSA is again applied to DH, with and without acid: DH =

Ri-R>-Rm

where

DH' _ /?i - (j>Rt - 4>3/2Rm ( r\ TJ

JLJn

n

l\\

n

\

\

R

n

— Y\j — -f^m

If the rate of chain transfer to DH is significant, then is calculated by the model using successive substitution. Starting with = 1 on the right-hand side, the calculation converges sufficiently with three iterations. If Rm is negligible (for example, when considering chain transfer to monomer only), then can evaluated directly. This expression for (/> contains two unknown rate constants, k\ and ks. The Arrhenius coefficients for these rate constants were determined using the 140 °C conversion data from Figure 7.6. The parameters were estimated for both the case of chain transfer to monomer, and again for chain transfer to DH. Given that the two chain transfer models differ only in their predictions of A/w and that the fit was against conversion data, the optimum Arrhenius constants for both cases were the same: L = 6.91 x 102e758/;r

These model equations were implemented in a single computer program that does both standard autopolymerization and acid-mediated polymerization:

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

143

by setting initial acid concentration to zero, (f> —* 1 and all the rates are calculated according to the Hui and Hamielec model.

5.3

MODEL RESULTS

During the parameter estimation, it was noted that the model is relatively insensitive to changes in k\, and therefore it should not be taken as a reliable prediction of the Diels-Alder reaction rate constant. Effectively, the rate R\ is very large compared with the rates of initiation and chain transfer, and therefore the DH ratio calculation could be simplified without significantly changing the model results: ' - l ~DH ~ l+ksS DH

There are two other implications: (1) the relative magnitude of the rates validates Hui and Hameilec's simplifying assumption that lead to a thirdorder rate equation for thermal initiation; (2) the uncertainty in the prediction of k\ precludes any argument for or against the possibility that acid catalyzes the Diels-Alder reaction as well as the aromatization of DH. Turning to model results, Figure 7.11 shows the predicted monomer conversion after 1 h (after 0.5 h for the 160 °C case). The effect of acid on conversion

0.4 - •

.2

0.3 ••

200

T = 120°C Figure 7.11

400 600 CSA (ppm)

T = 140°C

800

-T = 160°C

Effect of acid level on conversion and Mw at 140 °C

1000 • Data

144

B. MATTHEWS AND D. B. PRIDDY

appears to be well predicted, lending some credence to the mechanism and structure of the model. It should be noted, however, that what few data were available were used to make the fit. More data are required to validate the conversion predictions thoroughly. Molecular weight predictions are compared with data from Buzanowski et al. [14] in Figures 7.12 and 7.13. In both figures, plot (a) shows results assuming chain transfer to monomer only (Model A), while Model B assumes transfer to DH only and is represented in plot (b). Neither chain transfer model is adequate. With the exception of the 160°C run, Model A severely under-predicts Afw. Under-prediction can have severe consequences in the plant: run conditions predicted to be safe by the model would actually generate much higher Afw and therefore high viscosity and possibly gels leading to shut-down. The Mw predictions in Figure 7.12 show the correct overall trend (A/w increases with increasing acid concentration), but the magnitude is off by a factor of 2-3. Several possible explanations for the poor Mw predictions can be considered. Since Afw is the problem, a review of the moments equations contained in the base model is in order to see if they are compatible with the way in which $ is implemented. All of the moment concentration rate equations are based on either the rate of propagation or ratios of termination and chain transfer to the propagation rate. These ratios are very straightforward to correct using the Mayo dimer factor, . The zeroth- and first-order moment rates appear correct, as they are the concentration of terminated chains and the concentration of monomer that has been polymerized, respectively. Mn data are not available to validate the ratio of first to zeroth moment. Calculation of Mw requires the second-order moment, but re-deriving it as a check is beyond the scope of this work. Another source of model error could be the Cm equation shown earlier. A linear correction factor is applied to compensate for the gel effect on propagation rate. Since the datapoints available are at relatively low conversion where the Comftrbom whfc Bnwomki's Figore 3 u§iDg dnia transfer to noMaer model

0

(a)

|

200

T = UCfC

Figure 7.12

400 600 CSA (ppn)

800

i wit* BuMomfci't FigBt 3 r to M«yo 41

1000

T=I40°C ——T=160°C • Dau|

0

(b) |

T = I20°C

T = 140'C ——T = 160T

Effect of acid level on Mw at various temperatures

»Dau|

145

INCREASING PRODUCTION RATES OF HIGH MW POLYSTYRENE

gel effect is low, this correction might be interfering with the acid effects. The Cm calculation for chain transfer to dimer contains a significant correction for dimer concentration which could be duplicating the correction applied by the (j> factor. Assuming that the equations are reasonably accurate, the chain transfer mechanism could be called into question next. The model as written requires chain transfer to monomer or dimer, but not both. Since the two cases fall on either side of the experimental results, a modification to allow both to occur simultaneously might lead to a better fit. Although neither of the chain transfer models lead to a suitable tool for production planning, they both exhibited the correct trends for conversion and Mw relative to the addition of acid (Figure 7.13). Since the model modifications are based solely on altering the amount of DH available for initiation and chain transfer, the model results are consistent with the Mayo mechanism.

6

CONCLUSIONS

Significant improvements in the rate of manufacturing high molecular weight polystyrene can be achieved by adding small amounts of strong acid to the process. The acid should be soluble in styrene and must not initiate cationic polymerization. Other concerns such as corrosion, acid recovery, and allowable oligomer levels in the product must be considered before implementing the acid process. The acid works by catalytically aromatizing the Diels-Alder intermediate produced by the Mayo auto-initiation mechanism. A program written to simulate the Diels-Alder and aromatization mechanisms was successful in fitting the experimental conversion data. Molecular weight was underpredicted then over-predicted by changing the chain transfer assumptions. The Comparison with Buzanowski's Figure 5 using chain transfer to monomer model

300000 (a)

400000

500000

600000 Mw

700000

Comparison with Buzanowski's Figure 5 using chain transfer to Mayo dimer model

800000

| — w/o Acid — Sim with Acid OData with Acid > Data w.'o Acid|

Figure 7.13

300000

400000

500000

600000 Mw

700000

800000

(fc>) |—- w/o Acid — Sim with Acid» Data with Acid »Data w/o Acid|

Comparison of Rate-Mw curves with and without acid

146

B. MATTHEWS AND D. B. PRIDDY

general trend for Mw was correct in both cases, lending additional support to the choice of the Mayo mechanism with acid-catalyzed deactivation of the reactive dimer intermediate.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Cao, G.; Zhu, Z.; Le, H.; Zhang, M.; Yuan, W., /. Polym. Eng., 1999, 19, 135. Benbachir, M.; Benjelloun, D., Polymer, 2001, 42, 7727. Priddy, D. B., Adv. Polym. Sci., 1994, 111, 67. Drumright, R. E.; Ellington, E.; Kastl, P. E.; Priddy, D. B., Macromolecules, 1993, 26, 2253. Drumright, R. E.; Kastl, P. E.; Priddy, D. B., Macromolecules, 1993, 26, 2246. Toplikar, E. G.; Herman, M. S.; Buyle Padias, A.; Hall, H. K., Jr; Priddy, D. B., Polym. Bull. (Berlin), 1997, 39, 37. Dais, V. A.; Drumright, R. E.; Ellington, E.; Kastl, P. E.; Priddy, D. B., Macromolecules, 1993, 26, 2259. Drumright, R.; Terbruggen, R.; Priddy, D.; Koster, R., US Patent 5618900 (to Dow Chemical), 1997. Morioka, I.; Yamada, K., Japanese Patent 05 178914 (to Sekisui Plastics), 1993. Fuku, M.; Okada, Y.; Aoshima, K., Japanese Patent 62 197 407 (to Nippon Oils and Fats), 1987. Pike, W.; Priddy, D.; Vollenberg, P., US Patent 5 663 252 (to Dow Chemical), 1997. Sanchez, J.; Yormick, J.; Wicher, J.; Malone, K., World Patent 9807684 (to Elf Atochem North America), 1998. Cummings, C.; Hathaway, P. US Patent 5455321 (to Dow Chemical), 1995. Buzanowski, W. C.; Graham, J. D.; Priddy, D. B.; Shero, E., Polymer, 1992, 33, 3055. Flory, P. J., J. Am. Chem. Soc., 1937, 59, 241. Mayo, F. R., J. Am. Chem. Soc., 1968, 90, 1289. Pryor, W. A.; Coco, J. H.; Daly, W. H.; Houk, K. N., J. Am. Chem. Soc., 1974, 96, 5591. Pryor, W. A.; Graham, W. D.; Green, J. G., J. Org. Chem., 1978, 43, 526. Graham, W. D.; Green, J. G.; Pryor, W. A., J. Org. Chem., 1979, 44, 907. Pike, W. C.; Priddy, D. B.; Roe, J. M.; Rego, J. M., World Patent 0068281 (to Dow Chemical), 2000. Priddy, D. B.; Dais, V. A., World Patent 9618663 (to Dow Chemical), 1996. Roe, J. M.; Rego, J. M.; Priddy, D. B., US Patent 6084044, 2000. Pike, W. C.; Priddy, D. B., World Patent 9900432 (to Dow Chemical), 1999. Wesselmann, M. A., US Patent 4 585 825 (to Dow Chemical), 1986. Paquet, A.; Priddy, D.; Vo, C. V.; Pike, W.; Hahnfeld, J., US Patent 5650 106 (to Dow Chemical), 1997. Matthews, B.; Pike, W.; Rego, J.; Kuch, P.; Priddy, D., J. Appl. Polym. Sci., in press. Hui, A. W.; Hamielec, A. E., J. Appl. Polym. Sci., 1972, 16, 749.

8

Preparation of Styrene Block Copolymers Using Nitroxide Mediated Polymerization DUANE B. PRIDDY Dow Polystyrene R&D, Midland, Ml, USA

1

INTRODUCTION

Styrene is a very versatile monomer. It can be polymerized by most types of polymerization mechanisms, e.g. free radical (FR), Ziegler-Natta (ZN), anionic, and cationic. Classical ZN polymerization of styrene yields isotactic polystyrene. However, if methylalumoxane (MAO) is added as a co-catalyst, syndiotactic polystyrene is formed. The resulting polymers formed using the various mechanisms of polymerization are summarized in Scheme 8.1. Styrene-containing block copolymers are commercially very important materials. Over a billion pounds of these resins are produced annually. They have found many uses, including reinforcement of plastics and asphalt, adhesives, and compatibilizers for polymer blends, and they are directly fabricated into articles. Most styrene-containing block copolymers are manufactured using anionic polymerization chemistry. However, anionic polymerization is one of the more costly polymerization chemistries because of the stringent requirements for monomer and solvent purity. It would be preferred, from an economic cost perspective, to have the capability to utilize free radical chemistry to make block polymers because it is the lowest cost mode of polymerization. The main reasons for the low cost of FR chemistry are that minimal monomer purification is required and it can be carried out in continuous bulk polymerization processes.

Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy O 2003 John Wiley & Sons Ltd

148

D. B. PRIDDY Block Polym.

Low MW (<20,000) Hercules Piccolastic

Scheme 8.1

Broad PD (>2) Low cost (forgiving) No initiator required

Modes of polymerization of styrene

Nitroxide mediated radical polymerization (NMRP) was pioneered by Rizzardo and Solomon in the mid-1980s [1]. Their work went unnoticed for almost a decade until Georges et al. reported the preparation of narrow polydispersity (PD) (<1.2) polystyrene using NMRP [2]. This report initiated an explosion of research aimed at both understanding the mechanism of NMRP and also utilizing it to prepare block copolymers. This chapter describes the application and limitations of NMRP for making styrene-containing block copolymers.

2

MECHANISM AND LIMITATIONS

NMRP is not a true living polymerization but it has some attributes of a living polymerization, e.g. rather narrow polydispersity polymers can be produced, polymer molecular weight increases linearly with monomer conversion, and sequential addition of monomers leads to block copolymers. However, no one has yet produced truly monodisperse polymer using NMRP. Therefore, there are likely side reactions going on during chain growth that lead to adventitious termination. The reason why FR polymerizations are not living is that growing polymer radicals interact with each other, resulting in chain termination. There are several modes whereby polystyryl radicals become terminated, including radical coupling, disproportionation, and chain transfer. Total elimination of these bimolecular processes from an FR polymerization is impossible. However, if one can keep the FR concentration very low, the rate of these termination

149

STYRENE BLOCK COPOLYMERS USING NITROXIDE

reactions can be significantly slowed to a level where a virtual living polymerization can be carried out. The way in which the FR concentration is maintained at a very low level in NMRP is by the addition of stable free radicals (e.g. nitroxyl) to the process. Nitroxyl radicals couple with carbon-centered radicals at diffusion-controlled rates. The resulting C—O bond is thermally labile and readily undergoes reversible homolytic cleavage to rebirth the polystyryl radical. This dormant-living equilibrium is what causes NMRP to have the attributes of a living polymerization. However, even though the addition of nitroxides to styrene polymerization lowers the concentration of carboncentered FR and thus slows the termination reactions, there still are enough FRs for some termination to take place.

3

HOW LIVING IS NMRP? THE RESULTS OF MODEL STUDIES

The most commonly studied unimolecular initiator for NMRP has been (see Scheme 8.2) I. This initiator was first synthesized by Priddy et al. by abstracting an H-atom from ethylbenzene in the presence of TEMPO [3]. They then studied I as a model of the chain-end in polystyrene made in the presence of TEMPO. They found that I decomposes upon heating under anaerobic conditions to form products resulting from both radical coupling and disproportionation (Scheme 8.2). Georges et al. also studied the thermolysis of I but instead of the formation of diphenylbutane as the minor product, they observed acetophenone [4]. Acetophenone is likely formed because of the presence of dissolved oxygen during their experiment.

+ HO—N

>

7

+

V Major

Major

Scheme 8.2 Reactions of I upon thermolysis

Minor

150

D. B. PRIDDY

There are several ways in which block copolymers can be made. The three main methods are (1) sequential addition of monomers, (2) the preparation of a functionalized polymer followed by the use of the functionalized polymer as a macroinitiator or chain-stopper for initiation or termination of polymerization of the second monomer, and (3) use of a multiple-headed initiator. The purity of the block copolymers produced in these processes is dependent upon the livingness (lack of side reactions that lead to termination) of the chemistry used to make them. If the integrity of the chain-ends is maintained throughout the polymerization because all possible termination mechanisms are absent or eliminated, then pure block copolymers can be produced. If, however, impurities get into the process or if there are side reactions that lead to chain termination, the resulting block copolymers are contaminated with some homopolymer. Depending upon the application, some contamination of homopolymer in the block copolymer may be acceptable. In an effort to understand the limitations of NMRP for making functionalized and block copolymers, Priddy et al. carried out NMRP polymerization of styrene using model alkoxyamine I having a high extinction coefficient phenylazo chromophore attached to it, either on the initiating phenethyl radical (II) or else on the terminating TEMPO radical (III) [5]. This allowed quantification of the amount of the functionalized chain-ends during the polymerization of styrene using GPC-UV/VIS analysis. The results showed that a much higher percentage of polymer chain-ends have an attached chromophore group when using II versus III as the initiator (Figure 8.1). This finding suggests that there are more competing side reactions leading to termination than competing reactions leading to initiation of new chains. These data clearly show that highly pure block copolymers cannot be prepared using NMRP and that the purity of the block copolymer is inversely proportional to the molecular weight of the polymer segment formed using NMRP. The poor chain-end purity achievable using TEMPO-based alkoxyamine NMRP initiators led researchers to develop new nitroxyl radicals that will mediate vinyl polymerization more effectively. Hawker et al. utilized combinatorial techniques to synthesize and screen many different nitroxyl radical structures [6]. Their work led to the development of nitroxyl IV having a /^-hydrogen. Whereas TEMPO only works well for styrenic monomers, IV is claimed to work well for acrylates and diene monomers. Also, the polymerization rates achievable using IV are much faster than when using TEMPO. The alkoxyamine unimolecular initiator V, made from IV, has been successfully used to make a variety of block copolymers [7-9]. Recently, Hawker et al. synthesized V having a chromophore attached to either the initiating (VI) or the terminating (VII) radical (similar to the work of Priddy et al. on chromophore-labeled TEMPO mentioned previously) [10]. The results of the study showed that mediation of polymerizations using V yields polymers having greater end-group purity than polymerizations mediated using alkoxyamines based on TEMPO (Figure 8.2).

151

STYRENE BLOCK COPOLYMERS USING NITROXIDE l00-i "§95-

n

oe

o~\— -——— _

90-

^

4>

O

85-

n €

75-

~~"— '—

D^

a.

80-

\ \ .

\ \ DX

"?

.1 w

'~—~

\

o

o = k.

"""

O

0

70-

\

.c

—©— Initiated with II -a— Initiated with III

65^

\ \ \q

605

1

1

1

i

1

1

10

15

20

25

30

35

Mn/1000

Figure8.1 Comparison of PSchains having initiating versus terminating fragment derived from TEMPO-based alkoxyamine initiators. Polymerizations were carried out at 120°C 100 -i 98-

94o

92 -

86-

—e— Initiated using VI -B— Initiated using VII

84 100

50

150

Mn/1000

Figure 8.2 Comparison of PS chains having initiating versus terminating fragment derived from V-based alkoxyamine initiators. Polymerizations were carried out at 120 °C

152

D. B. PRIDDY

O

rv

VII

Chromophore

Chromophore

Another /?-hydrogen-bearing nitroxide having improved performance in NMRP is N-ter/-butyl-./V-( 1 -diethylphosphono-2,2-dimethyl)propylnitroxyl (VIII) [11,12]. A class of improved nitroxides (compared with TEMPO) for NMRP without a ^-hydrogen are the imidazolidone nitroxides (IX) developed by CSIRO researchers [13]. Both VIII and IX are claimed to give improved performance and higher end-group fidelity in NMRP, especially for acylates, compared with TEMPO.

O

N

•O—N O

I OEt

VIII 4

OEt

R

R

IX

BLOCK COPOLYMERS VIA THE MACROINITIATOR APPROACH

The highest volume commercial block copolymers are the styrene—butadiene (S-B) block copolymers. S–B block copolymers are manufactured using anionic polymerization with sequential addition of monomer (SAM) techniques. Attempts to make S—Bpolymers using NMRP via SAM have been limited because NMRP does not generally work well for diene monomers. Therefore, Priddy et al.

STYRENE BLOCK COPOLYMERS USING NITROXIDE

153

utilized the macromonomer approach. Butadiene was polymerized using traditional ani6nic chemistry to produce polybutadienyllithium (PBD-Li) [14]. The PBD-Li was then terminated with an alkoxyamine functional epoxide (X) to produce an alkoxyamine-functional macromonomer (XI). Addition of XI to bulk styrene polymerization at 120°C led to the formation of S—B copolymer (XII) (Scheme 3). The S—Bblock copolymer produced (XII) was characterized by a variety of techniques including NMR, gel permeation chromatography (GPC), thin-layer chromatography (TLC), and transmission electron microscopy (TEM). The results of these analyses clearly showed that the block copolymer was fairly pure with very little homopolymer contamination. One potential commercial application of this technology is the preparation of S—Bblock copolymers in situ during the manufacture of high-impact polystyrene (HIPS) and also acrylonitrile—butadiene—styrene (ABS) copolymers [15]. A significant proportion of HIPS and ABS polymers are manufactured in continuous bulk polymerization processes where an S–B block copolymer is dissolved in the monomer being fed to the reactor. Dow Chemical researchers demonstrated that addition of TEMPO-functionalized PBD can be added to the monomer feed instead of S–B block rubber, resulting in the formation of HIPS and ABS resins having similar properties to resins produced using block rubbers. Under appropriate conditions, transparent HIPS and ABS resins are formed. The transparency is achieved because the rubber and polystyrene or SAN domains are too small to scatter light. The TEMs of an S—Bblock copolymer XII and a transparent HIPS made using in situ-formed S—Bblock copolymer is shown in Figure 8.3.

PBD

Scheme 8.3 Synthesis of S—B block copolymers using sequential anionic/NMRP polymerization techniques

154

D. B. PRIDDY

Figure 8.3 TEMs of XII and TIPS made using /ns/fu-formed S—B by addition of XI to a continuous bulk styrene polymerization

Another example of the macroinitiator approach to making block copolymers is shown in Scheme 8.4. Since methyl methacrylate (MMA) polymerization cannot effectively be initiated by TEMPO-based alkoxyamine initiators, a poly(methyl methacrylate) macroinitiator (XIII) was prepared using conventional free radical polymerization [16]. However, the azo initiator was functionalized with a TEMPO-based alkoxyamine. Since the main mechanism of termination during bulk MMA polymerization is by radical coupling, most of the MMA polymer chain-ends are functionalized with alkoxyamine groups.

OH

CN

CN

+

Cl

o

N=

MMA/normal free radical polym. s I

Scheme 8.4 Synthetic approach to block copolymers using sequential normal/living radical polymerization

STYRENE BLOCK COPOLYMERS USING NITROXIDE

155

Addition of the XIII as a macroinitiator to styrene polymerization resulted in the formation of S—MMA—S triblock copolymer. The same procedure was also used to make styrene—butyl acrylate block copolymers. The success of this chemistry versus a control was demonstrated by conducting a parallel experiment where PMMA was prepared under the same conditions as the preparation of XIII, except using azobisobutyronitrile (AIBN) as the initiator instead of the alkoxyamine functional azo initiator. The AIBNand alkoxyamine functional azo-initiated PMMA were dissolved in styrene and heated at 130°C. A film of the resulting block copolymer made using alkoxyamine-functionalized XIII was translucent and flexible whereas a film of the polymer formed by polymerizing styrene in the presence of unfunctionalized PMMA control was opaque and very brittle. A final example of the use of the macroinitiator approach for block copolymer synthesis is the use of a macro azo initiator to initiate NMRP of styrene (Scheme 8.5). This process was used to make styrene-bl-siloxanes [17].

Scheme 8.5 azo initiator

5

Preparation of styrene-bl-siloxane using NMRP initiated with a macro

PREPARATION OF BLOCK COPOLYMERS USING ALKOXYAMINES AS CHAIN-STOPPERS IN STEP-GROWTH POLYMERIZATION

The molecular weights of polymers made using step-growth polymerization are typically controlled by the addition of a chain-stopper to the process. The chainstopper to monomer ratio determines the final molecular weight of the polymer. If functionalized chain-stoppers are used, functionalized polymers are produced. If macro chain-stoppers are used, triblock copolymers are formed during the step-growth polymerization. The highest volume commercial step-growth polymer is polycarbonate. There has been considerable interest in preparing block copolymers containing

156

D. B. PRIDDY

polycarbonate blocks. These materials have potential utility as compatibilizers for blends of polycarbonate with other polymers. Therefore, Priddy et al. set out to prepare polycarbonate-containing block copolymers utilizing NMRP techniques [18]. The approach is shown in Scheme 8.6. Polystyrene was prepared using the difunctional alkoxyamine initiator XIV. Since XIV contains a carbonate linkage at its center, the resulting polystyrene has a carbonate linkage in the center of the chain. Hydrolysis of the carbonate linkage yields polystyrene having a phenolic group on one end. Phenols are the most common chain-stoppers for the manufacture of polycarbonate. Addition of the phenolterminated polystyrene as a macro chain-stopper to the polycarbonate (PC) process led to the formation of S-PC-S block copolymer.

6

PREPARATION OF BLOCK COPOLYMERS VIA SEQUENTIAL ADDITION OF MONOMERS (SAM)

Prior to the development of the 'universal nitroxide' IV, the SAM technique was mainly limited to the preparation of block copolymers containing only vinylaromatic monomers. This is because TEMPO-based NMRP does not work well for other monomers. Although several papers appeared claiming to have successfully prepared block copolymers with acrylates using TEMPO chemistry, it is doubtful that they were very pure.

XIV

Styrene

N-O-PS.

120°C

PS—O—N

O—PC—O

PS—O—N

Scheme 8.6 Preparation of S-PC-S triblock copolymer using phenoxy functional PS prepared using NMRP as a chain-stopper in the PC process

STYRENE BLOCK COPOLYMERS USING NITROXIDE

157

There has been considerable debate over the reason why TEMPO does not work well for acrylates. CNRS researchers utilized NMR and matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) techniques to look at the terminal end-groups on poly(n-butyl acrylate) produced using TEMPO mediation [19]. They found that most of the chains were terminated by to-unsaturation resulting from either elimination or O—H TEMP from the chain-end or disproportionation between TEMPO and the polyradical (Scheme 8,7). The discovery of improved nitroxides IV, VIII, and IX led to the demonstration of a number of block copolymers via SAM. Table 8.1 list examples of block copolymers made using SAM and the type of mediating nitroxide used. The reactivity ratios of comonomer pairs in free radical polymerization are not changed during NMRP [20]. It is interesting that monomers such as MM A that do not give narrow PD polymer during NMRP mediated using TEMPO yield narrow PD copolymers when styrene is added. Examples of monomers that have been statistically copolymerized with styrene using NMRP include acrylonitrile, N, N-dimethylacrylamide, acrylic acid, methyl methacrylate, 2-hydroxyethyl acrylate, and glycidyl acrylate [8]. The ability of free radical chemistry to yield an infinite number of statistically random copolymers with a variety of comonomer pairs, coupled with the living nature of NMRP, provides the opportunity to synthesize a multitude of unique block copolymer compositions, for example block copolymers containing random and gradient copolymer blocks. NMRP can be used to make block copolymers by first polymerizing monomer 1, isolating the nitroxide-terminal polymer and then using it

^. . . Disproportionation

ao/wx,—CH2—f

+

o4

O-nBu

HO-N

)

V

TEMPO-H

Scheme 8.7 Proposed mechanism of termination during NMRP of n-butyl methacrylate using TEMPO

158 Table 8.1

D. B. PRIDDY Examples of block copolymers prepared using the SAM technique

Monomer 1

Monomer 2

Styrene w-Butyl acrylate Styrene

n-Butyl acrylate Styrene n-Butyl acrylate

Styrene Styrene Styrene Styrene Styrene Styrene Styrene Styrene Styrene Styrene Styrene Styrene tert-Eutyl acrylate Isoprene 4-Chloromethylstyrene

Nitroxide mediator

TEMPO [22], IX [13] IV[8], VIII [12] 4-Oxo-TEMPO [22], IV [8], VIII [12] Methyl methacrylate TEMPO [23] Ethyl methacrylate TEMPO [23] TEMPO [23] Octyl methacrylate Vinyl acetate TEMPO [23] N,N-Dimethylacrylamide TEMPO [23] 2-(Dimethylamino)ethyl acrylate TEMPO [23] 4-tert-Butylstyrene Di-tert-Butyl nitroxide [24] 4-Methylstyrene TEMPO [25], IX [13] TEMPO [26] Butadiene TEMPO [26], IV [9] Isoprene TEMPO [27] 4-Acetoxystyrene TEMPO 4-Chloromethylstyrene IV [9] Isoprene IV [9] Styrene TEMPO [28] Styrene

as a macroinitiator to carry out a random or gradient NMRP copolymerization of a binary monomer mixture. If the macroinitiated copolymerization of the binary monomer mixture is stopped at low monomer conversion, the second block is a random copolymer. However, if the monomer conversion is carried to high conversion, the second block is a gradient copolymer. Typically, normal free radical copolymerizations are stopped at low monomer conversion to eliminate copolymer composition drift. The composition drift occurs because the ratio of unreacted monomers is changing with conversion and new chains are always being initiated. Therefore, chains formed at low monomer conversion have a different monomer ratio than chains formed at high monomer conversion. However, since NMRP virtually eliminates termination processes, the second copolymer block on all of the macroinitiated chains have the same composition. As the ratio of unreacted monomers changes with conversion, a copolymer composition gradient from initiated to terminal end takes place. For example, if NMRP of a mixture of 10mol% acrylonitrile and 90mol% Styrene

STYRENE BLOCK COPOLYMERS USING NITROXIDE

159

is macroinitiated with polystyryl-TEMPO, and the monomer conversion is only carried to 10%, an S-bl-S-stat-AN is formed. However, if the monomer conversion is carried to 90%, a S-bl-S-grad-AN is formed. If the reactivity ratios of the two monomers are very low (i.e. « 1) such that an alternating copolymer block is formed (e.g. styrene and maleic anhydride), a triblock polymer can be formed. For example, if NMRP of a mixture of 90mol% styrene and 10mol% maleic anhydride is macroinitiated using polystyrylTEMPO, an alternating block consisting of 1:1 styrene and maleic anhydride is formed. Once all of the maleic anhydride has been consumed, a pure polystyrene block is formed. The result is an S-bl-S-alt-MA-bl-S triblock copolymer.

7

PREPARATION OF BLOCK COPOLYMERS USING MULTIPLE-HEADED INITIATORS

There is a lot of similarity between this concept and the macroinitiator approach in that a macromonomer is first formed by either (1) forming an alkoxyamine end-functional polymer by initiating a polymerization using an alkoxyamine functional initiator, or (2) using a functionalized alkoxyamine to initiate an NMRP resulting in the formation of a polymer that is capable of initiating a polymerization of another monomer by a mechanism other than NMRP. Hawker et al. developed a dual-functional molecule capable of initiating both NMRP and anionic ring-opening polymerization and then used it to demonstrate both approaches just described by using it to make block copolymers building the NMRP block either first or last (Scheme 8.8) [21]. In another example, Yildirim et al. photochemically generated anthracene radical cations in the presence of TEMPO [29]. TEMPO immediately trapped the radical to form the TEMPO-anthracene cation, which was subsequently used to initiate cationic polymerization of cyclohexene oxide (CHOX). The resulting alkoxyamine-functional polycyclohexene oxide was used to macroinitiate styrene polymerization, resulting in the formation of S-bl-CHOX (Scheme 8.9). Puts and Sogah were the first to propose the concept of making block copolymers via NMRP using multiple-headed initiators (Scheme 8.10). The initiator they developed (XV) is capable initiating blocks via four different polymerization mechanisms [30].

D. B. PRIDDY

160

HO

n Caprolactone

Scheme 8.8

Use of double-headed initiator to make block copolymers

161

STYRENE BLOCK COPOLYMERS USING NITROXIDE

Scheme 8.9 NMRP

Double-headed initiator that initiates both cationic polymerization and

Anionic Ring-opening Polymerization

HO

Cationic Ring-opening Polymerization of Oxazolines

Scheme 8.10

Anionic Vinyl Polymerization

Multiple-headed initiator to make copolymers via NMRP

162

D. B. PRIDDY

REFERENCES 1. Solomon, D. H.; Rizzardo, E.; Cacioli, P., Eur. Pat. Appl. 135280, 1985. 2. Georges, M. K.; Veregin, R. P. N.; Hamer, G. K.; Kazmaier, P. M, Macromol. Symp., 1994, 88, 89. 3. Li, I.; Howell, B. A.; Matyjaszewski, K.; Shigemoto, T.; Smith, P. B.; Priddy, D. B., Macromolecules, 1995, 28, 6692. 4. Moffat, K. A.; Hamer, G. K.; Georges, M. K., Macromolecules, 1999, 32, 1004. 5. Zhu, Y.; Li, I. Q.; Howell, B. A.; Priddy, D. B., ACS Symp. Ser., 1998, 685, 214. 6. Hawker, C. J.; Benoil, D.; Rivera, F., Jr; Chaplinski, V.; Nilsen, A.; Braslau, R., Polym. Mater. Sci. Eng., 1999, 80, 90. 7. Benoit, D.; Harth, E.; Helms, B.; Rees, I.; Vestberg, R.; Rodlert, M.; Hawker, C. J., ACS Symp. Ser., 2000, 768, 123. 8. Benoit, D.; Chaplinski, V.; Braslau, R.; Hawker, C. J., J. Am. Chem. Soc., 1999,121, 3904. 9. Benoit, D.; Harth, E.; Fox, P.; Waymouth, R. M.; Hawker, C. J., Macromolecules, 2000, 33, 363. 10. Rodlert, M.; Harth, E.; Rees, I.; Hawker, C., J. Polym. Sci., Polym. Chem. Ed., 2000, 38, 4749. 11. Benoit, D.; Grimaldi, S.; Robin, S.; Finet, J.; Turdo, P.; Gnanou, Y., J. Am. Chem. Soc., 2000, 122, 5929. 12. Robin, S.; Gnanou, Y., ACS Symp. Ser., 2000, 768, 334. 13. Chong, Y. K.; Ercole, F.; Moad, G.; Rizzardo, E.; Thang, S. H.; Anderson, A. G., Macromolecules, 1999, 32, 6895. 14. Kobatake, S.; Harwood, H. J.; Quirk, R. P.; Priddy, D. B., Macromolecules, 1998, 31, 3735. 15. Priddy, D. B.; Li, I. Q., US Patent 5721 320, 1998. 16. Li, I. Q.; Howell, B. A.; Dineen, M. T.; Kastl, P. E.; Lyons, J. W.; Meunier, D. M.; Smith, P. B.; Priddy, D. B., Macromolecules, 1997, 30, 5195. 17. Yoshida, E.; Tanimoto, S., Macromolecules, 1997, 30, 4018. 18. Li, I. Q.; Knauss, D. M.; Gong, Y.; Howell, B. A.; Priddy, D. B., Polym. Prepr. (Am. Chem. Soc, Div. Polym. Chem.), 1999, 40, 383. 19. Burguiere, C.; Dourges, M.-A.; Charleux, B.; Vairon, J.-P., Macromolecules, 1999, 32, 3883. 20. Baumann, M.; Roland, A.-I; Schmidt-Naake, G.; Fischer, H., Macromol. Mater. Eng., 2000, 280, 1. 21. Hawker, C. J.; Hedrick, J. L.; Malmstroem, E. E.; Trollss, M.; Mecerreyes, D.; Moineau, G.; Dubois, P.; Jerome, R., Macromolecules, 1998, 31, 213. 22. Listigovers, N. A.; Georges, M. K.; Odell, P. G.; Keoshkerian, B., Macromolecules, 1996, 29, 8992. 23. Yousi, Z.; Jian, L.; Rongchuan, Z.; Jianliang, Y.; Lizong, D.; Lansun, Z., Macromolecules, 2000, 33, 4745. 24. Jousset, S.; Hammouch, S. O.; Catala, J. M., Macromolecules, 1997, 30, 6685. 25. Li, I. Q.; Howell, B. A.; Koster, R. A.; Priddy, D. B., Macromolecules, 1996, 29, 8554. 26. Georges, M. K.; Hamer, G. K.; Listigovers, N. A., Macromolecules, 1998, 31, 9087. 27. Listigovers, N. A.; Georges, M. K.; Honeyman, C. H., Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.), 1997, 38, 410. 28. Tsoukatos, T.; Pispas, S.; Hadjichristidis, N., Macromolecules, 2000, 33, 9504. 29. Yildirim, T. G.; Hepuzer, Y.; Hizal, G.; Yagci, Y., Polymer, 1999, 40, 3885. 30. Puts, R.; Sogah, D., Macromolecules, 1997, 30, 7050.

P A R T III

This page intentionally left blank

(EPS) ROLF-DIETER KLODT Dow Central Germany, Schkopau, Germany

BRAD GOUGEON The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Foamable polystyrene beads are generally produced by two basic processes: 1. polymerization of styrene in suspension into spherical beads containing a blowing agent, and its finishing in a multi-step process (Figure 9.1); 2. incorporation of a blowing agent during the extrusion process of bulk polystyrene, with the polymer strands quenched in a water bath to avoid foaming and consequent strand cutting. The resulting raw material, in the form of beads or granules, is called expandable polystyrene (EPS). The usual procedure is for the EPS beads or granules to be prepared in one location and transported to other locations, where they are expanded and/or molded into their final forms. This process has inherent advantages: the costs of shipping the voluminous foam is minimized and intricate molding shapes can be molded directly without postprocessing. The expandable polystyrene particles, based on suspension polymerization, are converted into foam in three steps: pre-foaming, temporary storage and final Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy © 2003 John Wiley & Sons Ltd

166

R.-D. KLODT AND B. GOUGEON

solid-liquid separation drying •

Expandable PolyStyrene beads

Figure 9.1 Schematic Diagram of the production of EPS and particle foam

foaming (Figure 9.1). The main applications for EPS particle foam based on suspension polymerization are in thermal insulation and in the packaging sector. Extruded EPS is normally processed directly into loose-fill material for packaging applications. The world-wide consumption of EPS is currently around 2.35 million tonnes per year (based on 1999). An annual increase in production of around 4% is forecast until 2005.

2

EPS BASED ON SUSPENSION POLYMERIZATION

2.1

2.1.1

PRODUCTION OF EPS RAW MATERIAL

General Description of Suspension Polymerization

Spherical beads that can be expanded into foam under the influence of heat or steam are produced directly by suspension polymerization in the presence of blowing agent. The term suspension polymerization describes a process in which water-insoluble monomers are dispersed as liquid droplets with suspension stabilizer and vigorous stirring to produce polymer particles as a dispersed solid phase. Initiators used in suspension polymerization are oil-soluble. The polymerization takes place within the monomer droplets. The kinetic mechanism of the suspension process is considered to be a free radical, water-cooled 'microbulk' polymerization [1]. Suspension stabilizing agents are present in the suspension to obtain and stabilize a desired droplet distribution of the dispersed phase. The suspension stabilizer has to be soluble or wetted in/by water. The particle size can cover

167

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

diameters between 10 |xm and several millimeters. Commercial EPS is focused on a particle size range of 0.1–2 mm, preferably 0.4-1.6 mm. Suspension polymerization has the following characteristics: relatively easy heat removal and temperature control, low dispersion viscosity over the whole conversion, relatively low level of impurities in the finished product and fairly low separation costs. The greatest advantage, however, is to achieve a final expandable product in particle form, with special requirements for the morphology, which is simple to isolate, something which cannot be achieved by the other polymerization technologies.

2.1.2

The Technological Steps for the Manufacture of EPS

The suspension polymerization for the production of EPS is carried out in a jacket batch reactor with a stirrer and usually two or more baffles. The volume of the vessels is normally between 20 and 100m3. At the start of the production process, the water phase and the monomer phase are placed in the vessel, and additives are added either before or during the polymerization. These are typically water- and monomer-soluble components dissolved or dispersed in separate vessels before the start of reaction (Figure 9.2). The monomer/water phase ratio usually lies between 40:60 and 60:40. The filled reactor is heated to reaction temperature. In general, the polymerization is carried out in more temperature steps, gradually increasing the temperature [2]. During the free-radical polymerization a blowing agent (pentane) is added under pressure. Suspending agents Styrene

Water

^ Screening

Additives

tion

\/

hopper sr

X*" ~f

Hopper

, YY Y

ating

1

1

1

Reactor

UD OJ

Figure 9.2

Technological steps for the manufacture of EPS beads

168

R.-D. KLODTAND B. GOUGEON

After conversion of the styrene monomer droplets to EPS beads, the reactor is cooled and the suspension is transferred, usually into a stirred mixing vessel. The further finishing process is normally continuous. The expandable beads are separated from water by centrifuges or rotating sieves. In different methods, EPS beads are then washed and partly pretreated. They are then dried using flash, fluid bed-, or/and silo dryers [3], sieved, screened in various bead size fractions and coated, depending on their size and application. The coating is normally applied in batch or continuous blenders using solid or fluid coating materials. The sizes of the beads and the coatings used affect the processing properties of the EPS foams produced from EPS beads. The different EPS types are separately packed, usually in octabins, IBCs or silo trucks for use by EPS foam manufacturers. 2.1.3

Polymerization and Impregnation Process

2.1.3.1

Polymerization

Usually the polymerization process is carried out in two or more stages. During the first stage, the final particles are formed, and in the second stage blowing agent is added and this penetrates into the beads. The duration of the second step is preferably determined by the residual monomer concentration that needs to be achieved (normally <1000ppm). For both stages of the process, initiators with different decomposition characteristics have to be applied to initiate the polymerization. Two peroxides are preferably used for the production of EPS: (a) dibenzoyl peroxide (BPO), to initiate the first stage of the polymerization at a temperature of about 90 °C, and (b) tert-butyl peroxybenzoate (ter/BuPB), for the second stage in a temperature range of 115–130 °C. A typical time-temperature regime for the production of EPS using both of these initiators is given in Figure 9.3.

20

0:00

2:00

4:00

6:00

8:00 10:00 cycle time (h)

12:00

14:00

16:00

Figure 9.3 Typical temperature—time regime for a two—stage polymerization process

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

169

However, the probability for the reaction progression greatly depends on the monomer conversion. Because the viscosity of the dispersed phase, in the first stage, is fairly low and the quantity of styrene is sufficiently high, the decomposition process (Figure 9.4) occurs only up to the benzoyloxy radical, which can directly start the kinetic chain. The purely thermal start of chains with reactive dimers of styrene, as a result of Diels—Alder reaction, can be ignored at fairly low temperatures of suspension polymerization, in contrast to the conditions for the bulk styrene process [4—7]. The kinetic scheme with constant reaction of the polymer/monomer droplet increases fairly quickly with conversion, and the mobility of the polymer chains rapidly falls below the mobility of the monomer. The reduced diffusion of live polymer chains in the droplet will reduce the rate of termination of polymerization. The associated increase in the number of radicals will cause a rapid increase in the polymerization rate. This phenomenon is well known as the Trommsdorf or gel effect [8,9]. The gel effect causes a growth of the polymer chain length and widening of the molecular weight distribution (Figure 9.5). To decrease the viscosity at this stage of polymerization requires a continuous or stepped increase in the polymerization temperature [10–15]. The simultaneous addition of small amounts of bifunctional monomers and chain transfer agents (ktr > 2) also results in a more desirable molecular weight distribution with a flat, low Mw side and a sharply decreasing high Mw flank. The procedure of later addition of the chain transfer agent before the gel effect takes off is preferred. A low concentration of branched macromolecules should result in a higher melt flow rate during the pre-foaming step of the expandable beads [12,13]. In the literature, the procedure of earlier addition of blowing agent is described to reduce the viscosity of the droplets. Hamielec and co-workers [16] pointed to the fact that temperature increase and, simultaneously, earlier addition of

M-

CH3 ^ ^ CH ru C— 3 CH,3 \^

*-°TrnrO~».H—O -C02 Q II -,CH3 — C — CH3+CH3 « CH I

\^ CH 3 -c-0« CH3

-<^ scission CH

i^h^

I H O - f - CCR H 3 +RR . CH3

Figure 9.4

Decomposition process for (a) BPO and (b) tert BuPB

170

R.-D. KLODT AND B. GOUGEON 1.40

0.00

3.00

3.50

4.00

4.50 5.00 5.50 Log(Molecular Weight)

6.00

6.50

Figure 9.5 Change of Mw distribution with polymerization time (h); 0.32% BPO, T = 90°C.

blowing agent (n-pentane) with monofunctional initiators lead to a limited conversion and plasticizing effect by pentane, causing enhanced coalescence and resulting in total instability of the suspension. Hamielec and co-workers [16] and Moritz [17] suggested using, preferably, bifunctional initiators in which enhanced coalescence is completely overcome by the very short duration of the particle growth, owing to high polymerization rates. Towards the end of the polymerization, however, the chance that a primer radical initiates polymerization is reduced. The amount of very reactive benoyloxy radicals which can react with the polymer chain under abstraction of hydrogen will increase with consequent formation of benzene. This is why the exchange of the secondstage initiator tertEuPB is with an initiator that does not contain aromatic groups, such as tert-butylperoxy-2-ethylhexyl carbonate (TBEC), which is currently used by most EPS producers, particularly in western Europe and in the USA [18,19]. An interesting replacement initiator for tertEuPE is also tertamylperoxy-2-ethyl hexylcarbonate (e.g. Luperox TAEC). Its radicals have slightly faster decomposition rates, diffuse better into the high conversion matrix of an EPS particle and are more selective for adding styrene monomer [20]. The possibility of simultaneously using the synergist of the flame retardant agent, mostly dicumyl peroxide (DCP), for the reduction of the residual monomer and benzene content is described in a patent [21].

2.7.3.2

Impregnation

In general, the blowing agent can be added to the batch before or during the suspension polymerization [22]. It can also be incorporated into polystyrene that

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

171

has been re-suspended in water. The procedure of addition of blowing agent is referred to as impregnation. Too early an addition of blowing agent has the disadvantage of retarding polymerization because of a dilution effect [16]. The addition of the blowing agent after the polymerization is, in practice, only used for the recycling of waste materials and side bead fraction, which are not for sale [23]. The impregnation during the polymerization at a conversion of about 6585 % is currently the one technological process of practical importance. Procedures with addition of blowing agent at the point where the density of PS particles overcomes the density of the water phase (at about 66 % conversion) are preferred [24]. Blowing agents used are halogen-free CnHm compounds with n = 4–6, such as n-butane, isopentane, n-pentane, neopentane and hexane [25]. However, mix– tures of iso- and n-pentane are currently preferred for the production of standard EPS types. The diffusion rates for both pentane isomers are significantly different at given temperature. In the case of a high content of isopentane, the pre-foamed beads have the ability to expand over a long period during temporary storage but, on the other hand, longer cooling times are needed for the demolding process. Finally, the correct selection of the composition of a pentane mixture is a compromise between good foamability and fast diffusion behavior. The increase in pressure in the reactor and its progression during the impregnation step depend on the time (conversion) and the temperatures of the addition of pentane and the composition of the pentane mixture. With later addition of pentane, an asymptotic decrease of the pressure is observed, which conforms with the Fickian diffusion. It has been shown in the case of n-pentane that when the blowing agent is added to the suspension, it first diffuses very rapidly into the periphery of the beads (diffusion coefficient D = 10 -6 cm 2 /s), but after almost complete pentane uptake, the further diffusion toward the core of the beads occurs very slowly (D « 2 x 10 -7 -10 -10 , depending on the degree of conversion) [26]. If the addition of pentane occurs at 60–70% conversion, two influences result: an increase in pressure due to the arranged loss of soluble styrene with increasing conversion, and a decrease in pressure because of increasing diffusion of pentane into the beads. The equilibrium pressure for the quaternary system styrene—polystyrene—isopentane—n—pentane has been calculated by Wolfahrt [27] for different conversions, temperatures andn-/isopentaneratios using a thermodynamic sorption model based on chain-of-rotators equationof-state. The expansion capacity up to low foam densities can be influenced by the concentration of the pentane in the beads. Normally, 6–7 wt % pentane is used in standard EPS types and less than 3.5 % in so-called low-pentane types [28]. The thermogram from the differential scanning calorimetry (DSC) of EPS beads shows only one glass transition temperatures (Tg), indicating the state of chemical solution of polystyrene in pentane. Tg falls by around 7 °C per wt % pentane [29].

172

R.-D. KLODT AND B. GOUGEON

Although the pentane is homogeneously distributed in the dispersed phase during the impregnation process, which requires 2–3 h at 115–130°C, the fraction with the finest particle size range contains significantly lower pentane concentration after completion of the process than the largest sizes. The reason for this is the shorter diffusion paths over the particle diameter, which leads, during the finishing process, to the differential diffusion of pentane out of the particles, depending on its diameter. Maturation hoppers with constant residence time are frequently used to make the diffusion process controllable and to provide a homogeneous pentane distribution over the EPS bead diameter [130]. At the time of packaging, the diffusion process is slowed by using pentanebarrier film in all appropriate packaging. In octabins with film liners, it is possible to store EPS beads for several months without significant loss of potential pre-foaming power.

2.1.4

Particle Formation and Stabilizing Process

In principle, the formation of a particle begins with the dispersion of the styrene phase in water. Fine drops with a definite distribution are formed under the influence of agitation. It is assumed that the hydrodynamic forces in the turbulent field of the reactor - strongest in the stirrer region - cause droplet break-up as a result of turbulent fluctuations. The drops break up if the available energy dissipation is sufficient to overcome the surface tension and viscoelastic resistance of the droplets. Coalescence occurs if drops collide with each other with sufficient energy to distort the drops to a degree that the liquid film between the drops can drain away before the flow field separates the drops again. The average drop size decreases with increase in stirrer speed, decrease in hold-up fraction and stabilizer concentration. The drop size distribution is found to be dependent on impeller type. It is possible to 'freeze' the initial particle size and size distribution up to the end of the polymerization by different procedures [30–33], but these are not applicable to reach the target particle size range of EPS. In general, with increasing conversion in the dispersed phase, the viscosity increases sharply, as do the properties of the interfacial layers, and the density changes slightly, causing a change in the equilibrium between breakage and coalescence towards coalescence. The particle distribution shifts to a wider distribution with a larger particle size (Figure 9.6). This phase is normally identified as the sticky phase. Lack of suspension stabilizer in this phase results in total instability. Finding a suitable stabilizer system, flow conditions and procedures is the way to solve the particle formation problem and to shift the particle size in the direction of market requirements. Finally, the aim is to achieve partial or controlled instability.

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

173

Figure 9.6 Particle size distribution of PS depends on the conversion, using pickering stabilizer system: (1) 24.3%; (2) 44%; (3) 51 %; (4) 57%; (5) 70% conversion

2.1.4.1

Applied Stabilizer Systems and Suspension Stabilizing Mechanism

Typical suspension stabilizers for the production of EPS are water-soluble, surface-active macromolecules, such as poly(vinyl alcohol) (PVA), hydroxyethylcellulose (HEC) and polyvinylpyrrolidone (PVP), or natural products, such as gelatin [36–40], and insoluble inorganic powders, such as tricalcium phosphate (TCP), also called 'pickering stabilizer', mostly in combination with surfactants called extenders [33–35,44], or a combination of these [129]. The differences and specialties of these stabilizing mechanism are described briefly below:

2.7.4.2

Mechanism of Steric Stabilization with Macromolecules

Surface-active macromolecules are primarily suitable for the stabilization of large particles because of the large thickness of the adsorption layer and good

174

R.-D. KLODT AND B. GOUGEON

attachment (multiple attachment points) of the hydrophobic chains in the monomer drops. The range of the macromolecular adsoption layer is generally much wider than the effect of the electrical interactive forces. In general, macromolecules are preferred with a more or less blocky structure in which one block or group has a hydrophobic nature which acts as an anchoring component on the droplet surface and the other block or group has to be a good solvate with water and provide the steric repulsion together giving a train-loop chain arrangement [41,46]. Using the HVO theory [135], the steric repulsive force penetrating adsorption layers during anchoring of two particles can essentially be attributed to the change in the volume restriction effect of macromolecules in the area of tails and the osmotic effect at its additional anchoring. PVA molecules show a typical energy—distance profile of adsorbed amphiphilic macromolecules (Figure 9.7) with a train, loop, tail arrangement. The hydrophobic charge on the interface, which is fundamental for the polymer adsorption, increases with increase in the residual acetate content and the blocky character of acetate groups. This fact is contrary to the aim of wanting large particles because of the simultaneous reduction in interfacial tension [42]. It has been suggested [43] that high molecular weight PVA should be used with fairly low proportion of acetate groups for the production of EPS. Gotze and Sonntag [45] found that the extension of the PVA molecules on the interface increases with increase in molecular weight. A high viscosity in the surface also results in a slow drainage of water out of the area between the drops. A good alternative is the use of PVP, which is not so strong a surfactant, but shows a range of interaction forces comparable with PVA [45]. PVP is normally added to the reactor first during the polymerization [47]. 0.00025

Polyformal M= 50x103

distance (nm)

Figure 9.7 Energy—distance curve for different surface-active macromolecules: experiments with crossed tubes

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

175

High molecular weight chains giving high viscosity in the water phase are also a good protection against coalescence. Examples of this type are gelatin, alginate or/and carboxy- or hydroxyethylcellulose, also well known as protective colloids (the terms 'steric stabilizing' and 'protective colloids' are often used in the same context). Owing to the lack of large hydrophobic groups, the surface tension reduction effect is comparatively low. These types of stabilization are used for the manufacture of ion-exchange resins through copolymerization of styrene and divinylbenzene. Patents also describe combinations of HEC with pickering stabilizer for the production of EPS [48]. Cellulose ethers generally have a large adsorption layer with a compact structure (Figure 9.7) and high interfacial shear viscosity [49].

2.1.4.3

The Effect of Inorganic Pickering Stabilizer

The pickering system consists of water-insoluble inorganic solid particles and, mostly, an amphiphilic cosurfactant, also called an 'extender'. Tertiary calcium phosphate (TCP) with an average diameter of about 0.2–0.4 is normally used as the inorganic powder. Sodium alkylbenzenesulfonate (ABS) is preferred as an extender [53]. The stabilizing mechanism of the pickering system is that the solid particles are able to form a mechanical barrier between the monomer drops, which reduces the probability of particles approaching each other to a critical distance, which would then result in coalescence. In principle, it is possible to achieve a stabilizing effect with the inorganic particles alone. The TCP particles do not, however, have a charge to attach to the organic phase and are, as a result, only loosely bonded to the surface of styrene drops. The extender molecule helps to fix the inorganic particle into the surface of monomer drops [54] (Figure 9.8). Because of its amphiphilic character, it simultaneously achieves the wetting of the inorganic particles in the water and organic phases. The bonding of TCP on to ABS is based on an ionic attraction mechanism between the negatively charged ABS sulfonic group and the positive charge of the Ca. The tails of the —(CH) M — chain (n = 14–20) of ABS are bonded by hydrophobic inter-action on/into the styrene surface (Figure 9.8). Deslandes [55] found that the layer of TCP that surrounded the polymer beads was actually composed of two distinct parts: a thin layer made of uniformly distributed TCP primary particles and a second layer, normally much thicker, which consisted of loosely packed agglomerates of TCP.

176

R.-D. KLODT AND B. GOUGEON

Figure 9.8 The mechanism of the pickering stabilizer

2.1.4.4

Particle Size Control

The result of the suspension polymerization is a, more or less, wide particle size distribution (PSD). The interest lies in being able to shift the various proportions of particle size ranges towards the actual market requirement. The PSD has to be as narrow as possible. The average diameter using the pickering stabilizer systems, such as TCP, in combination with an extender, such as ABS, can be controlled, to some extent, by varying the concentration of TCP, the ratio of the amount of extender to TCP, and also the addition time of the extender [53,54,56– 58]. With the use of macromolecular suspension stabilizers, such as PVP and PVA, different times for the addition of the suspension agent can also be used for particle size control [59]. The final PSD is controlled by the geometric factors of the reactor and stirrer systems and the hydrodynamic conditions which will affect the drop break-up and coalescence during the suspension polymerization. Specifically, these are the agitator type and dimensions, vessel geometry, operating parameters and the stirrer speed [61–65, 67–71,129]. Attempts to produce large particles with narrow bead distribution ranges have been suggested. In a patent [72], the application of hydrodynamic conditions is described, using tall reactors with volumes > 60 m3 to achieve a standard deviation of particle size distributions of < 25 %. An independent claim is included for a hydrodynamic system, where the ratios of stirrer diameter to

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

177

reactor diameter and of stirrer surface blade height to stirrer and reactor diameter have to have definite dimensions. The aim is to reach ratios as low as possible between surface velocity, shear rate and pumping rate. Through sudden or stepwise reductions of stirrer speed during the polymerization, it should be possible to obtain narrower PSDs using the pickering stabilizer system [66]. In addition, a number of co-modifiers have been developed as replacements for, or in addition to, the standard extender of ABS type, to improve the particle size distribution. Sodium or potassium persulfate, sodium bisulfite, partly combined with hydroxy ceto compounds, unsaturated carboxy acids [73–75], small quantities of water-soluble sodium polystyrene sulfate [76] and/or copolymers of acrylic acid and 2-ethylhexyl acrylate [10] have been suggested as extenders for narrower PSD. Marclay [77] discussed the possible mechanism of extender action by potassium persulfate in suspension polymerization. Further suggestions for narrower particle size distribution, improved suspension stability and more accurate particle size control are combinations of water-insoluble pickering stabilizer of TCP type with compounds such as ethylenediaminetetraacetate (EDTA) and its salts [78], EDTA, CaCO3 and hydroxyethylcellulose [79], and sodium ^-naphthalene sulfonate and sodium polyacrylate [80]. A narrower PSD should also be achieved if the inorganic pickering salt is precipitated immediately before its use [2,128]. Vilchis et al. [81] presented a new idea to achieve better control of the particle size distribution by the synthesis in situ of a water-soluble copolymer of acrylic acid–styrene as suspension stabilizer without additional inorganic phosphate. Publications describe increasing the particle formation by using a physical (population balance, Maxwell fluid, power law viscosity, compartment mixing) modeling approach [22,60,98,105].

2.1.5

Additives

During or at the end of the EPS production, a number of additives can be incorporated to improve process and application properties. Additives can include nucleation agents, flame retardants, fast-cool agents, anti-lump and anti-static agents, stabilizers, plasticizers, pigments, etc. Some of the more important additives and their functions are described in more detail below.

2.1.5.1

Nucleation Agents

Nucleation agents are substances that are able to initiate and control cell formation and growth. They are purposely incorporated into the polymeric structure. At their location, the blowing agent (pentane) preferably evaporates when the EPS beads are exposed to saturated steam. However, unwanted

R.-D. KLODT AND B. GOUGEON

178

nucleation effects can arise from incorporated water, suspension stabilizer or hexabromocyclodecane (HBCD) [82–85]. Nucleation agents must be able to control nucleation independent of other effects and must have the ability to mask the bubble initiation of other sources. Waxes such as paraffins, chloroparaffins, and Fischer–Tropsch waxes and also esters and amides of fatty acids have been described [86–89] for this function. Finely divided phase-incompatible polymers, such as low molecular weight polyethylene (PE), are normally used in industry for cell size control [90,92–95]. Above its melting point, low molecular weight PE will dissolve in styrene. With growing conversion phase, inversion occurs and the domain size of polyethylene is set after the cooling step. The PE domains operate as gathering sites for the blowing agents, which are be able to form bubbles when treatment with heated steam occurs. The molecular weight, the branching degree, the concentration and the degree of distribution of PE determine the weak point of critical size in the polymer matrix resulting in the maximum diffusion path length. In general, the unbranched PE types with a degree of crystallinity >80% are, as previously described, suitable because of their higher incompatibility with polystyrene [90] (Figure 9.9). Investigations by Klodt et al. [91] showed that branched PE waxes with degrees of branching up to 1.5 and crystallinity above 40% can also be successfully used for the production of uniform cell size distribution (Figure 9.10). If the compatibility between PE and PS is further increased, for instance by increasing the degree of branching of PE or by

>/•>

vD

OO

O

—

«N

(N



fN


Cell Size (um) Figure 9.9 Cell size distribution depends on the quantity of unbranched PE (Mw, 1500; dispersity, 0.18; degree of crystallinity, 87%)

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

179

Figure 9.10 Nucleation effect of branched polyethylene Mw, 6770; dispersity, 0.89; branches per molecule, 1.21; degree of crystallinity, 48%; (a) 0%, (b) 0.05%, (c) 0.2% polyethylene

copolymerization of ethylene with octene or styrene, the nucleation effect diminishes more and more [91].

2.7.5.2

Flame Retardant Agents

EPS flame retardants are usually bromine compounds, in general with at least two bromine atoms and with a bromine content higher than 40 wt%, principally a suitable linear aliphatic brominated organic compound and bromo-substituted cycloalkane. The final product should have a bromine content of 0.6–3 wt% [97]. Hexabromocyclododecane (HBCD) is commonly used [97,99,100]. The flame retardant effect of HBCD is intensified by adding small amounts of so-called synergists, resulting in satisfactory flame retardancy achieved with much smaller amounts of the bromine-containing compound. DCP is a preferred synergist or an alternative may be a less volatile and more temperature stable —C—C initiator, such as 2,3-dimethyl-2,3-diphenylbutane [101]. The cyclic aliphatic bromine compound HBCD is, in contrast to polybrominated compounds, compatible with most regulations for protection of health and the environment, but is considered by the EU as a substance which should not be used in plastic materials in contact with food [102]. During combustion, halogenated fire retardants split from hydrogen bromide (HBr). The inhibition is controlled by the thermally initiated release of hydrogen halides, which, in reaction with the high-energy OH* and H* radicals, produce lower energy halogen (X*) radicals, which then inhibit the described mechanism [103]. The effect of the radical-forming synergist can be described by (a) acceleration of polymer degradation under thermal stress, (b) the resultant reduction in the viscosity of the PS melt combined with a faster flowing away of the PS melt from the flame front and (c) the faster formation of HBr as gas-phase flame

180

R.-D. KLODT AND B. GOUGEON

inhibitor [104]. Investigations by Hamann et al. [106] have shown that all requirements for building materials are also passed after a long period of storage of foam parts using HBCD and DCP as a fire retardant system because of their long half-lives (> 67 and >1000 year, respectively) at normal temperatures. Fire performance of foam parts is regulated in national or industry standards such as DIN 4102 and ASTM E84 via the model building codes in the USA.

2.1.5.3

Coatings

The untreated surface of EPS particles would provide poor processing characteristics. Because of this, a number of coating components with different effects are required to improve the foaming, aging, blocking performance and the final application properties of the foam. Polystyrene beads and pre-foamed particles normally tend to acquire various quantities of static charge which are normally long lasting and cause agglomeration. Therefore, anti-static agents are added before screening. In addition to flowability problems with the EPS beads during the finishing steps, static charges on the foamed beads also negatively influence the homogeneous filling of block forms and, in particular, molding equipment. Effective anti-static agents used include small quantities of esters of fatty acids and amines, quaternary ammonium salts, alkylphosphates, and fatty alcohol condensed with ethylene oxide on to propylene oxide [108, 109]. An anti-lumping agent protects against the propensity of particles to fuse together during the pre-foaming process at temperatures above the transition temperature. Anti-lumping agents used are, for example, metal stearates, inorganic powders, e.g. SiO2 and CaCO3, and also powders of polymeric materials, such as polyamide waxes [110–114, 131]. As anti-lumping agents often impact the fusion quality during molding, their concentration and composition should conform with the contradictory demands in both steps of foam production. Modern EPS types have been developed to improve the molding operation capacity providing short aging and cooling times. External fast-cool agents usually used are various mixtures of glyceride esters of higher fatty acids, preferably with carbon chain lengths of 14–20 [115–118]. The effect of cooling time reduction is explained by the formation of fine canals in the bead surface which support faster diffusion of pentane out of the beads [119] (Figure 9.11e and f). Because of a more rapid loss of pentane during the pre-foaming process, the foamability generally falls with a reduction in cooling time. Consequently, the balance of both properties is matched to suit the kind of application. In fact, the cooling time has more significance as the density and thickness of the foam increase.

181

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

The variation of composition and concentration of components in a mixture provide the opportunity to adapt the properties of one particle size to a different application (Figure 9.12).

Figure 9.11 Surface of different treated beads: (a, b) EPS raw material bead surface without coating; (c, d) EPS raw material bead surface with coating; (e, f) EPS prefoamed material surface with coating 4.5

T C £*J 3.J

_ 2.5

1.5 20

40

60 80 100 120 storage time of foam beads (h)

140

160

Figure 9.12 Influence of EPS particle size and external 'fast cool agent' on the pentane diffusion behavior: (1) 0.55 mm, standard; (2) 0.85 mm, fast cool agent; (3) 0.85mm, standard

182

2.2

R.-D. KLODT AND B. GOUGEON

FROM RAW MATERIAL TO FOAM

The conversion of EPS beads into foam consists of three major steps [124]: • pre-expansion (prefearming) of EPS particles; • maturation or temporary storage of pre-expanded beads; • final foaming (molding or blocking process).

2.2.1

Pre-Foaming Process

In the pre-foaming stage, the EPS raw material particles are heated by saturated steam to above the glass transition temperature of the PS in a pre-foamer. The blowing agent dissolved in the PS matrix expands and a system of spherical cells is formed. During this pre-expansion step, the bead volume can increase by a factor of 40–80 [120–122]. In general, two types of pre-foamer, continuous and discontinuous working units, are used for the pre-expansion process. The continuous pre-foamer normally consists of an open stirred tank with baffles. The raw material is fed by an adjustable feed screw near the container bottom. Steam is introduced at the bottom of the pre-foamer. The EPS beads move slowly upwards during pre-foaming and discharge into the fluidized bed via the overflow. The foam density attained depends on the residence time in the pre-foamer, which is adjusted by the feed screw or the weir height. High foam densities (>30g/l) require temperature control by air addition. For low bulk densities (special, <15g/l), the material pre-foamed once is pre-foamed again, after temporary storage. Discontinuous pre-foaming in a closed system allows higher temperatures at elevated pressure. The EPS material is loaded batchwise from above. Steaming can be accomplished using nozzles or screens at the bottom. Upon reaching the target foam density, often determined by light barriers in the vessel, the foam beads are discharged at the bottom into the fluid bed for drying (Figure 9.13). The EPS processing equipment industry offers machines that regulate the foam density automatically. Pre-foamers, with a volume of 1–4 m3, are currently the preferred size, with usual capacities of 500–2500 kg EPS per hour [123]. There are advantages and disadvantages of both types of pre-foamer. Continuous pre-foaming allows for a very simple design and has a lower energy requirement and steam consumption than the batch process. At foam densities of 20 g/1 and higher, the continuous pre-foamer allows a higher material flow rate per unit volume and time. Discontinuous pre-foamers are more costly, but because of the higher steam pressure, lower foam densities can be achieved. In Europe, more and more discontinuous pre-foamers are being used, the reason being the availability of low pentane grades for reduced pentane emission.

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

183

Expanded bead outlet

Fluid bed dryer

Figure 9.13

2.2.2

Schematic diagram of a discontinuous operating pre-foamer

Maturation

After the pre-expansion process, the pre-foamed beads are cooled. Consequently, the remaining pentane in the cells condenses, creating a vacuum in the beads. In this state, the pre-foamed particles are very friable. A definite residence time of maturation of pre-expanded beads is required before the molding process can be carried out. The pre-foamed beads are typically stored in fabric silos (large fabric storage bags). During the maturation, process air penetrates into the cells until the internal pressure approaches atmospheric pressure. If, at this point, the foam is heated again, it will expand further, and can be used for a second or third expansion process in a continuous prefoamer or for the final foaming process in a mold [123]. In the event of a longer temporary storage time, the volatile pentane diffuses through the cell walls to the outside and the pre-foaming power slowly reduces.

2.2.3

Final Foaming Procedures

After temporary storage, the pre-foamed beads are welded into a homogeneous foam by renewed steam supply, in a perforated mold. Foaming takes place in molds, molding machines (Figure 9.14) or special equipment (slab units, belt units). The characteristic processes are mold filling, steaming, cooling and demolding [123]. As example the process of block molding is briefly described below.

184

R.-D. KLODT AND B. GOUGEON

Figure 9.14

2.2.3.7

Example of a molding machine. Courtesy of Kurtz Holding GmbH & Co.

Mold Filling

As a rule, the molds are filled pneumatically. This can be accomplished with a blower centered over the filling injectors. The air blown in is removed through perforated walls, or through special devices. It is also possible to suck in the material by applying a vacuum in the mold. It is practical to add to the prefoamed material the foam accumulated during cutting, as reclaim rates up to 20% are possible without quality deterioration.

2.2.3.2

Steaming

The walls of the mold are equipped with borings, slotted plates or slot nozzles. Behind the walls are steam chambers. Using control valves, it is usually possible to steam each side separately. Prior to the actual steaming, a vacuum is applied in the mold, to remove the air in the charge. The first step in steaming is the socalled 'wash phase', in which usually the two largest surfaces have a small amount of steam applied to them. The actual steam injection process usually occurs over several steps, by alternate lateral steaming, to a pre-set steam pressure or foaming pressure. It is also possible to hold the end pressure for a short time (after-steaming), to achieve good sealing of the block surfaces. The mold is then vented. 2.2.3.3

Cooling

Cooling is accomplished by applying a vacuum. Heat is removed intensively from the mold due to vaporization of water under a vacuum.

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

185

The decreasing foaming pressure can be measured via the internal mold pressure.

2.2.3.4

Demolding

The block can be safely demolded at an inner mold pressure of ~0.02 bar. Demolding occurs by hydraulically or pneumatically operated ejectors once the face wall is opened. It is possible also to open slightly a lateral wall, or the lid. The block is here ejected forward on to a roller conveyor. The blocks are next weighed, marked and transported to the block storage. Temporary storage is necessary to cool the blocks completely and to allow water entrained during steaming to diffuse out. A minimum storage time must be respected to prevent subsequent warping or shrinking of the slabs after they are cut.

2.2.4

Cutting of Foam Blocks

The foam blocks that accrue during block pre-foaming are usually cut to standardized slabs of various dimensions. Cutting with electrically heated wires has the advantage that dust-free, sintered surfaces are formed. To this end, resistance wires of chromium-nickel alloy are heated electrically. Additional heat can be generated by moving the wires (oscillation). The forward feed of the blocks can be mechanical. The wires can also accomplish the forward motion, as they are moved in their clamping frame. The slabs cut are usually automatically labeled and packaged. Foamed PS material can also be cut or processed on wood processing machines, such as saws, milling machines or planes.

2.3 2.3.1

PHYSICAL AND MECHANICAL PROPERTIES Thermal Conductivity

The most important property for insulation is thermal conductivity. The following transport types participate in the transmission of heat: heat conduction in PS, heat conduction in the filling gas (air), radiation heat transfer and heat convection by convection flows in the closed cells. The thermal conductivity of the air in the cells contributes the most to the total heat transport. The radiation fraction depends on the diameter of the cells formed. The thermal conductivity depends on the density of the foamed PS material. Thermal conductivity decreases with increasing bulk density, reaches a minimum and then rises again (Figure 9.15). The following processes are responsible for this characteristic.

R.-D. KLODT AND B. GOUGEON

186 0.0500

0.0300 10

15

20

25

30

35

40

45

Foam density (kg/m3)

Figure 9.15 Lambda as function of the foam density: line, master curve; points, measured values with particle foam on basis of SconaporR (trade name of EPS by BSL/Dow Chemical)

Owing to small polystyrene fractions and a correspondingly large gas volume, in the range of low densities, the conditions are favorable for radiation exchange. An increase in the PS fraction leads to increased absorption, reflection and scattering and heat transfer by radiation diminishes. In the PS skeleton, heat transfer increases and becomes determining for the course of the curve in the range of high bulk density. The effect of thermal radiation can be for the most part eliminated by infrared adsorbers or reflectors (e.g. carbon black). An improved insulation effect in the low density range is a result [7]. Neopor, a development of BASF, is produced on this basis.

2.3.2

Mechanical Properties

2.3.2.1

Compressive Strain at 10 % Compression Set

The compressive strain is given at 10 % compression, since for semi-rigid foamed plastics there is no sudden rupture of the cell structure. In the characteristic compressive force–compression set line for foamed PS material, the curve rises linearly at the beginning. Here foamed PS material behaves elastically. As the strain increases, an irreversible deformation occurs (compressive set). The limit of elasticity lies in the range 1.5–3.5% compression. This limit is shifted toward lower compression levels as the temperature increases. The compressive strain at 10% compression depends on the degree of fusion of the pre-foamed beads. Compressive set increases linearly with increasing density (Figure 9.16).

187

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

X;

&>

\

90 •

« £ 85 . c •* "3 "~ 80 •

s.i

r./ s /

75

> £ .£ b 70 .

o

U

y

60

^^

-— m

// ^

s^

s

2 e 200 •y, .g

i I

Jff*

Q. O 1 00

so °

O

,/

S

^/

.2 § 150

/

'_/

5Q

^

?*•* ^ [

/

50

n. 0

(a )

s \s

o

10

20

30 40 50 60 Fusion degree ( %)

70

8

(b)

20 25 30 Foam density (g/1)

Figure 9.16 Compressive strain at 10% compression as function of (a) the degree of fusion (foam density 17 g/l) and (b) foam density at optimal fusion

2.3.2.2

Flexural Strength

The bending test shows a steady course of the compressive force-bending deflection, even though the relationship is not linear. At low loads, the foamed PS material behaves evenly (linear increase in stress-strain) and up to halfrupture there are only small deviations from a linear course. The flexural strength of foamed PS materials increases with increasing density. The degree of heat-sealing affects the flexural strength. The flexural strength is lower for poorly heat-sealed foamed PS materials.

2.3.2.3

Tensile Strength, Transverse Tensile Strength

The tensile test shows a steady course of the tensile force–elongation curve, where a linear relationship exists only at low tensile forces. There are only small deviations from a linear course at approximately half-rupture. The tensile strength increases with increasing density. The degree of heat-sealing has an effect here also.

2.3.4.3

Acoustic Properties

Owing to their dense, closed surface, untreated foamed PS slabs are acoustically ineffective. Small absorption effects can be achieved by slotting, perforating or corrugating the surfaces. Foamed PS slabs are suitable for footfall-sound insulation only if the dynamic E modulus is lowered by elastification.

188

2.4

R.-D. KLODT AND B. GOUGEON

APPLICATIONS

EPS foam has, like XPS, a closed cellular structure (Figure 9.17), but the production of lower densities is possible. The usual density range is 10–80 g/l (0.6–5.0 lb/ft3). Densities of 10g/l, or lower, are possible after several prefoaming operations.

2.4.1

Construction Industry

EPS foam is used in many building projects for thermal insulation and also, more and more often, for soundproofing in new buildings and modernization or renovation work. EPS foam slabs are used for the insulation of walls, roofs, floors and ceilings. Particle sizes between 0.9 and 1.6 mm are preferably used for this application. For the thermal insulation of walls, there is a difference between outside wall insulation, inside wall insulation, and core insulation (European applications). For the outside wall insulation the EPS foam is put directly on to the stone bearing structure. A fabric reinforced plastering or a ventilated facing protects it from the weather. Using sandwich panels of EPS plasterboard, modern heat insulation standards can be achieved on the walls of older buildings. For core insulation, the insulation layer is between the bearing wall and the external weather resistant wall. Finally, sound insulation is becoming increasingly important. Special elasticated slabs combine sound and thermal insulation.

Figure 9.17 scales

Comparison of EPS particle foam and XPS foam structure on different

189

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

Another system of insulation is the use of EPS molded foam parts (insulated concrete forms) for a combination of outer and inner wall insulation. A wall is built with these molded foam parts and then filled with concrete. For the thermal insulation of roofs, there is a difference between flat roof and steep roof insulation. The insulation of a nonventilated flat roof is done with EPS insulation elements, which could, also, be pre-laminated with a sealing strip. Steep roof insulation is used in attic conversions for residential purposes. There is a difference between insulation under, on and between the rafters. For the between-rafter insulation, different systems are used to make installation very easy. Slabs from the Styrotect [125] system, tongued and grooved, are cut to size, fitted together and laid between the rafters. Also special wedge systems or the use of elasticated slabs are possible. For floor and ceiling insulation, different thermal and sound protection requirements have to be met. The priority for the insulation material of flat ceilings is to provide the necessary impact sound insulation protection. Impact sound insulation slabs are produced by a special manufacturing process [126]. In two or three pre-foaming steps the EPS beads are pre-foamed to a density of 10 g/l or lower. After molding, the formed blocks receive a special follow-up treatment. Before the EPS foam blocks are cut into slabs, the blocks are pressed in special presses to more than half their size and then immediately the load is removed. Because of the destruction of the cell structure in one direction, special elastic impact sound insulation properties are achieved (low dynamic stiffness with acceptable compressibility of the slabs, Figure 9.18). For decorative ceiling layouts and for echo regulation, specially produced slabs (pressed in design) or molded foam parts are available.

«s' 17/15 OS' 22/20 AS' 27/25 • s' 32/30 AS' 33/30 os' 38/35 • s'43/40

3 4 Compressibility (mm)

Figure 9.18 Dynamic stiffness as a function of compressibility of impact sound insulation slabs depends on its dimensions

190

R.-D. KLODT AND B. GOUGEON

Foam molding machine-produced perimeter insulation slabs (density 30 g/l or 1.85 lb/m3) are used for the insulation of outside walls and ground with soil contact, pressure load and moisture requirements. The good properties of molded EPS perimeter slabs (normally a domain of XPS) are achieved by means of hydrophobic coated small particles with very good pre-foaming behavior. For drainage purposes, bitumen banded EPS foam slabs are used. These slabs consist of poorly fused large particles resulting in large channels, where the water can drain away. 2.4.2

Packaging Materials

Foaming in foam molding machines produces packaging materials of EPS foam. In general, small EPS particles (0.4–0.9 mm) are used because the foam molding parts produced can have very complicated shapes and structures. Shapes with thin walls and bridges are also possible. Packaging materials using EPS foam protect packaged goods in storage locations and during transport against mechanical and thermal damage. The advantages over other packing materials are their low weight, resistance to water, good shock absorption behavior, excellent heat/cold insulation and easy processing and handling. Application examples are packaging materials for machines, machines parts, glass, china, optical, electrical and electronic equipment, toys and Pharmaceuticals, and also for edge and surface protection. Food packaging is possible, although, in this case, a special type of EPS with food certification is required. 2.4.3

Other Applications

Because of its diverse properties, EPS foam can be used in many other applications. The most important are: • • • • • • • • 3

additions for the production of lightweight building materials; bath support; safety helmets; foam cups; insulation of cooling equipment and cold boxes; life jackets, life buoys, fenders; plant containers, flower tubs; technical applications, for example full form molding. EPS BASED ON EXTRUSION PROCESS

For many years, wood excelsior and paper products dominated the dunnage packaging market. Dunnage here being defined as a packaging material that

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

191

takes up space and resists the movement of packaged items in a box. As early as the 1960s, it was determined that small-sized EPS particles had two properties that would be highly desirable as a dunnage packing material. The low bulk density (product density plus airspace around the pieces) of shaped polystyrene foam material lowered the total package weight and thus shipping costs. The frictional resistance to movement of the foam particles, when tightly packed, greatly reduced the likelihood that packaged items might shift within the box, reducing the chance of breakage during transit. The Dow Chemical Company first entered the loose-fill packaging market in 1962 with a material that resembled spaghetti strands. Eventually the shape evolved to the 'S' shape that characterized the product from the early 1970s to the present. Other competitors in the polystyrene foam loosefill market include Flo-Pak, manufactured by Free-Flow Packaging, InterPac, manufactured by Inter-Pac, WingPac and C-Pac, manufactured by Rapac, and Alta-Pak, manufactured by Storopack. In 1993, Dow sold the trademark rights to the 'S' shape and Pelaspan-Pac to Storopack. The information about foamed polystyrene loose-fill that follows is an overview specific to materials that are formed in hard resin strands, cut to length, boxed and shipped to and later expanded for customer use at a convertor (expander) location. 3.1

3.1.1

EXTRUSION

Raw Materials

The recipe for EPS loose-fill resin is similar to that of extruded polystyrene foam insulation: polystyrene with about a 200 000 weight-average molecular weight, hexabromocyclododecane (HBCD) as fire retardant, magnesium oxide as oxygen scavenger, calcium stearate as lubricant and n-pentane as blowing agent. Different blowing agents of the CFC and HCFC types have been used in the past but were discontinued as a result of the environmental and economic requirements placed on the product.

3.1.2

Extrusion System

The solids are metered into the extruder screw by means of an accurate rate feeder. The blowing agent enters the mixture downstream of the screw but before the initial mixer. The polymer blend is pushed through a series of heat exchangers and mixers, a gel screen,and through a multi-hole die. The die may have up to 96 holes or individual strands but smaller number was found to be optimal owing to consistency in strand temperature at the outer edges. Figure 9.19 shows the equipment layout through the die.

192

R.-D. KLODT AND B. GOUGEON Accurate rate solids feeder

Blowing agent

Extruder screw

Gel screen

In-line mixer

Cooler

Cooler

-*

Cooler

Figure 9.19

3.2

Schematic diagram of extrusion process

POST EXTRUSION

After the strands leave the die, they are run through a cold water bath to solidify the polystyrene. The strands are then run up through a helical rotary cutter that cleaves the strands to a length of approximately 6–9mm. The resultant 'beads' or 'pellets' are passed through an annealing unit and flash cooled to set nucleation sites and cell size. The beads next go through a filtration system to remove chips, fines and oversize pieces (overs/unders) resulting from the cutting operation. After a dewatering step, the beads are air conveyed to holding silos for an 8–12 h period. Prior to packaging, the beads are run through a six-pass crossflow air declassifier which removes over 98% of the remaining fines. The beads are then transferred to a loading hopper where calcium stearate (antidumping agent) and an anti-static agent are applied to the beads prior to packaging. In the USA, the beads are packaged in 1000 lb net weight 'Gaylords', which are large corrugated boxes set on a wood pallet. The boxes are lined with a heavy-duty polyethylene bag, which, after being filled, is sealed at the top to prevent escape of blowing agent from the container and the beads. Work done in the late 1980s [132] showed that there are many variables to achieving a perfect cut on a strand in this process. The following are the equipment and polymer conditions that must be considered:

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

• • • • • • • • • • • • •

193

die land length; die internal land angle; strand temperature at the die; variations in strand temperatures from center die openings to perimeter die openings; number of die openings (optimum); distance of the cooling water bath from the die; composition and hardness of pull rollers; external 'texture' of pull roller; variations in temperature at the cutter; cutter blade (helix) angle; rake angle/relief angle in the cutter; cutter gap distance; ability to discharge pellets cleanly from the cutter.

Failure to optimize one or more of these parameters can result in cuts that produce broken pieces or appendages to the pellet, which, when expanded, break off and create dust and partial pieces. This situation is difficult to rectify at a convertor location or at a customer's facility and results in a continuous, undesirable, housekeeping problem. Figure 9.20 shows an equipment layout from the water bath through the packaging station. Water bath

Helical cutter

Annealer Cold water

Surface lubricant Anti-static agent t

Packaging

Figure 9.20

Schematic diagram of post-extrusion process

194

3.3

R.-D. KLODT AND B. GOUGEON

STEAM EXPANSION OF EPS LOOSE-FILL RESIN: THEORY AND PRACTICE

EPS loose-fill resin expands under much the same conditions as EPS molding bead resin. The major difference is the orientation and size of the expander. Loose-fill expanders are generally horizontal and vary in size from 3 ft (laboratory size) to in excess of 20 ft in length. Unlike vertical bead expanders, loose-fill expanders rotate the entire drum. Loose-fill expanders do not have internal mixing arms attached to a shaft but, instead, have baffles attached to the interior wall of the rotating drum. These baffles are useful in maintaining the proper retention time inside the drum. They can be either fixed or adjustable. In addition to drum length and baffles, four other parameters influence the residence time in the expander: feed rate, drum pitch or angle from horizontal, drum rotation speed, and steam volume and pressure. Even under the best machine and operating conditions, unexpanded pieces of loose-fill entering the expander can be expected to exit at different times rather than moving through the expander at a constant rate. The output of the pieces follows a bell-shaped curve with the peak output occurring at about 4 min with a commercial size expander [133]. Polystyrene loose-fill normally undergoes three to five expansions with four being the norm. This will take the loose-fill from a bulk density of about 38–40 lb/ft 3 (pcf) (600–640 kg/m3) as an unexpanded pellet to about 0.20– 0.25 pcf (3.2–4 kg/m3) after four expansions or passes, as they are commonly called. Between expansions, a recovery or aging period of 16–24 h is required to refill the cells with air. For the first expansion pass, the unexpanded polystyrene loose-fill pellets are metered into one end of the horizontal expander at a rate of about 1000 lb (454 kg) per hour. This time can vary from 50 to 70 min and is dependent on the equipment and conditions listed above and on the bulk density that is achieved. Normally, the first-pass density aim is in the 0.80– 0.90 pcf (12.8–14.4 kg/m3) range with little (< 5%) shrink after cooling. Shrink or shrinkback is caused by the cooling of gases inside the cells of the foam. The cell walls are stretched during expansion by the heat and expansion of gases (blowing agent) inside the cells. When the cell walls are over-expanded and partially collapse during the cooling stage, this is shrink. A certain amount of shrink is acceptable as long as no cell walls are ruptured and the volume of gases filling the cells between expansions (aging period) is equal to the volume of gases that was in the cells at the peak of expansion. At the outfall of the expander, the first-pass loose-fill is air conveyed through ducts to a large ventilated holding bag or hopper. After the first aging is completed, the once expanded material is air conveyed back to an expander equipped with a large throat or opening to feed the expanded particles into the expander. Conditions are set so that the material exiting the expander will have a bulk density of about 0.38–0.42 pcf (6.0–6.7 kg/m3) with no more than 15% shrink after

195

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

1 min. Third-pass fresh bulk densities should range from 0.26 to 0.30 pcf (4.2–4.8 kg/m3) with no more than 20% shrink. Fourth-pass fresh bulk densities should be in the 0.22–0.25 pcf (3.5–4.0 kg/m 3 ) range with no more than 25% shrink. A typical loose-full expansion setup is shown in Figure 9.21. Conditions for expansion and aging will vary significantly between sea level and higher altitudes such as found in Colorado Springs, Colorado (altitude 6035 ft/1839m). Because of the difference in boiling points, more steam and longer expansion times are required to achieve densities equivalent to those easily attainable at sea level to 1000 ft in altitude. First pass aging bag

Second pass aging bags

Return line to expander

I

^ ~4 *t

* i

Fill line to aging bags

/

Expander

Fill line to aging bags

Gaylord of unexpanded resin (first pass)

\

.*-•

Fourth pass assx\

i / a aging g bags \f

\

Figure 9.21

^

To pack stati

Typical loose-fill expansion setup

Third pass aging bags

196

R.-D. KLODT AND B. GOUGEON

Aging conditions will affect the rate at which the cells recover. Temperatures below 60 °F will slow the process so that more time is required or the convertor may have to accept higher densities on subsequent passes. The size of the aging bags may also affect the recovery rate. Larger bags, filled to the top, will restrain the expanded pieces as they recover and attain their former size. This will also inhibit air exchange for pieces at the center of the bag. A good firstpass bag (for 1000 lb of unexpanded beads) should have a volume of at least 1500 ft3. Other things that will affect ultimate density and appearance of the loose-fill are related to the air conveying (airveying) system used to move the loose-fill around the expansion system. Recently expanded loose-fill pieces need to be moved with a certain amount of care as physical abuse will crush cell walls and compress the loose-fill piece to an extent which it will not recover in the aging process. To this end, the airveying system must be engineered and constructed to minimize 90° angles, especially closely downstream from the blowers in the system where the air velocity is the greatest. The airveying system must also be constructed to minimize sharp edges, protrusions into the ducts, and other details that will create fines, broken pieces, and dust that are normally unacceptable to end-use customers. Loose-fill is either bagged or shipped in bulk truck to the customer. Bags vary in size from 10 to 20 ft 3 , depending on the customer need and the particular geographic market. Bulk systems are popular with companies that use large quantities of loose-fill and have many packaging stations. The trucks used to transport the loose-fill are specially equipped with flexible duct that connects to the duct and bagging system at the customer. This way, 2500–3000 ft3 of loosefill can be shipped to the customer at short notice and unloaded with a minimal amount of manpower involved.

3.3.1

Photodegradable Loose-Fill

In the late 1980s, environmental concerns led to the development of photodegradable resins being incorporated in the polymer mix. Methyl isopropenyl ketone (MIPK) at a level of 5 % was found to be effective in making the loosefill pieces photodegrade to polymer 'dust' with 10–12 months of exposure to sunlight [134].

3.3.2

Physical Properties

Polystyrene loose-fill packaging material used in US Federal Government applications must meet US Government Specification PPP-C-1683 (12/5/ 1988), 'Cushioning Material, Expanded Polystyrene Loose-Fill Bulk (for Packaging Application)'. Compliance with eight properties is the basis for this

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

197

specification: flowability, vibrational settling (at two different loadings), electrostatic adhesion, dynamic cushioning, loaded bulk density, compressive creep after 7 days (three different loadings), compressive set after 24 h (same three loadings) and flammability.

3.3.3 End Uses Expanded polystyrene loose-fill is used in many markets today. The most common are dunnage (used to fill up space in a shipping box), cushioning (where the resilient qualities of the polystyrene foam are used), and texturizer, where it is ground into smaller particles and used in making textured ceilings in commercial and residential buildings.

REFERENCES 1. Munzer, M., Trommsdorff, E., Polymerizations in suspension, High Polym. 29 (1977) (Polymer Processes) 106. 2. Bilgic, T., Karali, M., Savasci, O. T., Effect of the particle size of the solid protective agent tricalcium phosphate and in-situ formation on the particle size of suspension polystyrene. Angew. Makromole. Chem. 213 (1993) 33. 3. JP 09059416 A2 (Achilles Corporation), 1995. 4. Riederle, K., Die Oligomerenbildung bei der thermischen Polymerisation von Styrol bis zu hohen Umsatzen des Monomeren, Dissertation, TU Munchen (1981). 5. Mayo, F. R., The dimerization of styrene, J. Am. Chem. Soc., 90 (1968) 1289. 6. Stein, D. J., Mosthaf, H., Oligomer formation in the thermal polymerization of styrene, Angew. Makromol. Chem. 2 (1968) 39. 7. Hohwiller, H., Neopor - a new EPS generation, in Particle Foam 2000, VDI Verlag, Dusseldorf (2000), 375-383. 8. Trommsdorf, E., Kohle, H., Lagally, P., Polymerization of methyl methacrylates, Makromol. Chem. 1 (1948) 169. 9. Mitta, I., Horie, K. J., Diffusion-controlled reactions in polymer systems, Makromol. Sci. Rev. Macromol. Chem. Phys. C27 (1987) 91. 10. Ethapol 1000 – Operating Manuals - Suspending Agent for Styrenic Polymerization Processes, CIRS, Padova, 1996. 11. AU 8661 177 (The Dow Chemical Company), 1985. 12. US 4652 609 (Atlantic Richfield Company), 1985. 13. US 4661 564 (Atlantic Richfield Company), 1985. 14. US 2857 342 (Monsanto Corporation), 1958. 15. US 3 072 581 (Monsanto Corporation), 1963. 16. Villalobos, M. A., Hamielec, A. E., Wood, P. E., Bulk and suspension polymerization of styrene in presence of n-pentane: an evaluation of monofunctional and bifunctional initiation, J. Appl. Polym. Sci. 50 (1993) 327. 17. Moritz, H. U., Influence of bifunctional initiators, DECHEMA Monogr. 131 (1995) 259. 18. EP 0 574 665 B1 (Huls AG), 1992.

198

R.-D. KLODT AND B. GOUGEON

19. Malta, A., More efficient initiators for the production of styrene (co)polymers, Organic Peroxides Symposium, Akzo Nobel, Conference Proceedings 8 (1998) 1. 20. Technical Documentation. Better Performance in Suspension Polystyrene, Elf Atochem, Paris, 1996, 1-22. 21. EP 0 488 040 (BASF), 1990. 22. Chen, Z., Pauer, W., Moritz, H.-U., Pruss, J., Warnecke, H. J., Modeling of the suspension polymerization process using a particle population balance, Chem. Eng. Technol. 22 (1999) 699. 23. EP 0 575 871 (BASF), 1992. 24. Konno, M., Arai, K., Seito, S., The effect of stabilizer on coalescence of dispersed drops in suspension polymerization of styrene, J. Chem. Eng. Jpn., 15 (1982) 131. 25. Ingram, A. R., Wright H. A., Composition of foam structures from expandable polystyrene beads, Mod. Plast. 41 (1963) 152, 156, 200, 203. 26. Villalobos, M. A., Diffusion of blowing agent into PS beads, Report PPR-01 to Plasti Fab Ltd, Calgary (1991). 27. Klodt, R. D., Wolfahrt, Ch., Blase, H., Hamann, B., Study on interaction of n-pentane with polystyrene in the polymerizing system, Conference Proceedings, Fachtagung Polymerwerkstoffe Merseburg (1994). 28. Busse, M., Hahn, O., Production of complex EPS moldings with defined properties, Plastverarbeiter 44 (1993) 46. 29. Ingram, A. R., J. Cell. Plast. 1 (1965) 69. 30. Matsumoto, S., Takeshita, K., Koga, J., Takashima, Y. J., A production process for uniform-size polymer particles, Chem. Eng. Jpn. 22 (1989) 691. 31. Polacco, G., Semino D., Rizzo C., Feasibility of methyl methacrylate polymerization for bone cement by suspension polymerization in a gel phase, J. Mater. Sci.: Mater. Med. 5 (1994) 587. 32. Hatate, Y., Uemura,Y., Ijichi, K., Kato, Y., Hano, T., Baba,Y. Kawano, Y., Preparation of GPC packed polymer beads by SPG membrane emulsifier, J. Chem. Eng. Jpn. 28 (1995) 656. 33. JP 96-340879 (Hitachi Chemical Co.), 1998. 34. DE 19 650 301 (The Dow Chemical Company), 1966. 35. JP 279602 (Hitachi Chemical Company), 1992. 36. Ahmed, S. M., Effects of agitation, and the nature of protective colloid on particle size during suspension polymerization, J. Dispers. Sci. Technol. 5 (1984) 421. 37. DE 75-2548524 (BASF), 1977. 38. EP 0 425 992 (BASF), 1989. 39. DE 3331 570 (Huls AG), 1983. 40. US 4 042 541 (The Dow Chemical Company), 1976. 41. Goodall, A. R., Greenhill-Hopper, M. J., Characterization of partially hydrolyzed poly(vinyl acetates) for use as stabilizers in suspension polymerization, Macromol. Chem., Macromol. Symp. 35–36 (1990) 499. 42. Fabini, M., Bobala, S., Rusina, M., Macho, V., Preparation of poly(vinyl alcohol) as the dispersant for suspension vinyl chloride polymerizations, Polymer 35 (1994) 2201. 43. Mendizabal, E., Castellanos-Ortega, J. R., Puig, J. E., A method for selecting a poly(vinyl) alcohol as stabilizer in suspension polymerization, Colloids Surf. 63 (1992) 209. 44. EP 0 732 343 A2 (Mitsubishi Chemical BASF Company), 1996. 45. Gotze, Th., Sonntag, H., Forces between quartz surfaces bearing adsorbed macromolecules in good solvents, Colloids Surf. 31 (1988) 181. 46. Hesselink, F.Th., On the theory of stabilization of dispersions with adsorbed polymer, J. Polym. Sci. Polym. Symp. 61 (1977) 439.

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE

199

47. EP 96-201906 (Royal Dutch Shell, Ltd), 1998. 48. DE 3 331 570 (Huls AG), 1983. 49. Klodt, R.-D., Wustneck, R., Ruhle, R., Thummler, W., Investigations on the interfacial rheological behavior of methyhydroxypropylcellulose, Abh. Akad. Wiss., N1 (1986) 257. 50. FR 76–32736 (BASF), 1978. 51. GB 97-21603 (Dyno Industries ASA), 1997. 52. DE 4 029 298 (Huls AG), 1991. 53. Witt, M., Dissertation, Technische Universitat Munchen (1989). 54. Wolters, D., Meyer-Zaika, W., Bandermann, F., Suspension of polymerization of styrene with pickering emulsifiers, Macromol. Mater. Eng. 286 (2001) 94. 55. Deslandes, Y., Morphology of hydroxyapatite as suspension stabilizer in the polymerization of poly(styrene-co-butadiene), J. Appl. Polym. Sci. 34 (1987) 2249. BE 0700533 (The Dow Chemical Co.), 1966. US 4303 784 (Atlantic Richfield Company), 1980. EP 0 575 872 (BASF), 1992. CS 70–5854 (Czech), 1978. Vivaldo-Lima, E., Wood, P. E., Hamielec, A. E., An updated review on suspension polymerization. Ind. Eng. Chem. Res. 36 (1997) 939. 61. Apostilidou, C., Stamatousdis, M., On particle distribution in suspension polymerization of styrene, Collect. Czech. Chem. Commun. 55 (1990) 2244. 62. Stamatousdis, M., Apostolidou, C., Characteristics of particle sizes produced by suspension polymerization of styrene, Part. Part. Syst. Charact., 9 (1992) 151. 63. Tanaka, M., Oshima, E., Dispersing behavior of droplets in suspension polymerization of styrene in a loop reactor, Can. J. Chem. Eng. 66 (1988) 29. 64. Leng, D. E., Guarderer, G. J. Drop dispersion in suspension polymerization, Chem. Eng. Commun. 14 (1982) 177. 65. Schroder, R., Piotrowski, B., On particle formation during suspension polymerization of styrene, Ger. Chem. Eng. 5 (1982) 139. 66. Masato, T., Hideyo, T., Isao, K., Natsakaze, S., Kazuhiko, H., Effect of stepwise and continuous reduction in impeller speed on particle size distributions in suspension polymerization of styrene, J. Chem. Eng. (Jpn.) 21 (1995) 118. 67. Yang, B., Kamidate, Y., Takahashi, K., Takeishi, M., Unsteady stirring method used in suspension polymerization of styrene, J. Appl. Polym. Sci., 78 (2000) 1431. 68. Tanaka, M., Izumi, T., Application of stirred tank reactor equipped with draft tube to suspension polymerization of styrene, J. Chem. Eng. Jpn. 18 (1985) 345. 69. Kuzmanic, N., Mitrovic-Kessler, E., Skansi, D., The influence of mixing on the styrene polymerization product, Chem. Biochem. Q. 6 (1992) 1. 70. DE 19 530 765 Al (BASF), 1997. 71. Tanaka, M., Hosogai, K. Suspension polymerization of styrene with circular loop reactor, J. Appl. Polym. Sci. 39 (1990) 955. 72. DE 19 816 461 C1 (Buna Sow Leuna Olefinverbund GmbH), 1998. 73. US 3 631 014 (Sinclair-Koppers), 1971. 74. JP 08301905 A2 (Sekisui Plastics), 1995. 75. US 3 755 282 (Sinclair-Koppers), 1973. 76. US 4 500 692 (Atlantic Richfield Company), 1985. 77. Maclay, W. N., The mechanism of extender action by potassium persulfate in suspension polymerization of styrene, J. Appl. Polym. Sci. 15, (1971) 867. 78. JP 95–118775 (Sekisui Plastics), 1996. 79. DE 3 728 044 A1 (Huls AG), 1987. 80. DE 19 650 301 (The Dow Chemical Company), 1966.

200

R.-D. KLODT AND B. GOUGEON

81. Vilchis, L., Rios, L., Guyot A., Guillot, J., Villalobos, M. A., In-situ formed copolymer of acrylic acid–styrene as stabilizer during suspension polymerization of styrene, DECHEMA Monogr. 134 (1998) 249. 82. DE 0142193 (Buna Schkopau), 1979. 83. Hinselmann, K., Physical laws in the pre-foaming of expandable polystyrene, Kunststoffe 61 (1971) 152. 84. BE 0 834 383 (BASF), 1974. 85. FR 1 493 947 (Monsanto Corporation), 1965. 86. EP 0000 120 (BASF), 1977. 87. BE 0815 185 (MonsantoCorporation), 1973. 88. BE 0819 346 (Cosden Technology Co.), 1973. 89. EP 0 396 046 b1 (Kanegafuchi Kagaku Kogyo), 1990. 90. Technical Release, Polywax Polyethylene, Petrolite Speciality Polymers Group, Sugar Land, Tx. 91. Klodt, R.-D., Nieter, E., Kuhnberger, W., Koller, F., Bunge, F., Nukleierungswirkung niedermolekularer Ethylenhomo-und-copolymerer im PS Partikelschaum, Polymerwerkstoffe 2000, Halle, Tagungsband (2000) 456. 92. Becker, G. W., Braun, D., In Kunststoffhandbuch, ed. Gausepohl, G. H., Gellert R., Polystyrol (4), Hanser, Munich, 1996, 579. 93. DE 2 448 476 (BASF), 1974. 94. US 3 503 908 (Koppers Corp.), 1969. 95. EP 0 050 968 (American Hoechst Corporation), 1981. 96. DE 3 843 536 Al (BASF), 1988. 97. Troitzsch, J., Brandverhalten von Kunststoffen, Hanser, Munich, 1982. 98. Vivaldo-Lima, E., Wood, P. E. Hamielec, A. E., Penlides, A., Calculation of the particle size distribution in suspension polymerization using a compartmentmixing model, Can. J. Chem. Eng., 76 495. 99. Troitzsch, J., Kunsstoffe 77 (1987) 1078. 100. Encyclopedia of Polymer Science and Engineering, vol. 16, 2nd edn, Wiley, New York. 101. EP 0 374 812 (BASF), 1988. 102. Hardy, M. L., Rouge, B., presented at 6th European Meeting on Fire Retardancy of Polymeric Materials, Lille, 1997. 103. Eichhorn, J., Synergism of free radical initiators with self-extinguishing additives in vinyl aromatic polymers, J. Appl. Polym. Sci. 8 (1964) 2497. 104. Fenimore, C. P., Inhibition of polystyrene ignition by tris(2,3-dibromopropyl) phosphate and dicumyl peroxide, Combust. Flame 12 (1968) 155. 105. Kalfas, G., Yuigen, H., Ray, H. W., Modeling and experimental stydies aqueus suspension polymerization processes, 2. Experiments in batch reactors, Ind. Eng. Chem. Res. 32 (1993) 1831. 106. Hamann, B., Klodt, R.-D., Gellert, R., Pelzers, T., Langzeit-Bewahrung von PSHartschaum der Baustoffklassen Bl bzw. B2 nach DIN 4102, Bauphvsik 21 (1999) 29. 107. Regulatory Status and Environmental Properties of Brominated Flame Retardants Undergoing Risk Assesment in the EU. 108. Product Bulletin - Fatty Amines, Akzo (1990). 109. Gachter, G., Muller, H., Taschenbuch der Kunsstoff-Additive, 3.Ausg., Hanser (1989). 110. US 3 389 097 (Koppers Corp.), 1964. 111. Informationsschriften Aerosil, Fallungskieselsauren undo Silikate, Degussa (1991). 112. DE 4 123 252 (BASF), 1991.

PARTICLE FOAM BASED ON EXPANDABLE POLYSTYRENE 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135.

201

DE 19816469 Cl (Buna Sow Leuna Olefinverbund GmbH), 1998. JP 72–53828 (Badische Petrochemical Co.), 1972. DE 2942865 (Hills AG), 1979. US 4286069 (Hoechst AG), 1980. US 4 278 731 (Atlantic Richfield Corporation), 1980. DE 2 226 168 (BASF), 1972. Ingram, A. R., Cobbs, R. R., Couchot, L. C., in Resinography of Cellular Plastics, ASTM STP 414, American Society for Testing and Materials, Philadelphia (1967) 53–67. US 3 139272 (Bloom), 1964. US 3023 175 (Koppers Corporation), 1962. Lindhof, W., Properties and uses of recycled, expanded poystyrene, Recycl. Recov. Plast. NA (1996) 631. Kunststofftechnik, Particle Foam 2000, VDI Verlag, Dusseldorf (2000). Guidlines for Transport and Storage of EPS Raw Beads, APME (1998) 1–9. RYGOL Dammstoffe Dammen leicht gemacht, Part: roof application (1999). Bollmann, W., Verarbeitungshinweise zur Herstellung von Trittschalldammplatten aus EPS, Kunststoffe 79 (1989) 77–2. Olayo, R., Garcia, E., Garcia-Corichi, B., Sanchez-Vazquez, L., Alvarez, J., Poly (vinyl alcokol) as a stabilizer in the suspension polymerization of styrene: the effect of the molecular weight, J. Appl. Polym. Sci. 67 (1998) 71. DE 4220 225 Al (BASF), 1993. Harre, K. H. Untersuchungen zum Einfluss hydrodynamischer Grossen auf die Suspensionspolymerisation von Styrol, Dissertation, Dortmund (1983). EP 781 638 A2 (Buna Sow Leuna Olefinverbund GmbH), 1997. DE 19816460 Cl (Buna Sow Leuna Olefinverbund GmbH), 1998. Russell, P. M., Hitchcock, M. K., et al. (Dow Chemical), unpublished work (1989). Tusim, M. H., Russell, P. M. (Dow Chemical), unpublished work (1992). Hitchcock, M. K. (Dow Chemical), unpublished work (1990). Hesselink, F. Th., Vrij A., Overbeck, J. Th. G., J. Phys. Chem. 75 (1971) 2094.

This page intentionally left blank

10

Rigid Polystyrene Foams and Alternative Blowing Agents KYUNG WON SUM AND ANDREW N. PAQUET The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION AND GENERAL DESCRIPTION

The beginning of cellular polystyrene may be dated back to the time of the discovery of a material called styrene. Polystyrene is a clear, brittle, thermoplastic aromatic resin:

where n = 1900–2900 (general-purpose resin). It is one of the most versatile thermoplastic resins available for the production of low-cost plastic foams. In 1831, it was discovered that the vapors given off when a gummy material from the balsam tree (called storax) was heated contained the chemical substance styrene. Storax has also been found in substances from embalmed Egyptian mummies some 3000 years old. As early as 1831, scientists knew that liquid styrene was one of those unusual substances that could undergo a certain chemical change to become a hard solid. As scientists gained chemical knowledge about the natural styrene, they eventually learned how to produce synthetic styrene. By 1929, scientists at The Dow Chemical Company were able to devise a method to produce synthetic styrene from benzene and ethylene [1]. This Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy r 2003 John Wiley & Sons Ltd

204

K. W. SUH AND A. N. PAQUET

allowed The Dow Chemical Company to start the development of polystyrene as a material for molding and extrusion. The original concept of cellular polystyrene may be due to the Swedish inventors C. G. Munters and J. G. Tandberg, who filed a patent on 'Foamed Polystyrene' on 21 August 1931, and subsequently obtained US Patent 2023204 on 3 December 1935. In 1941, The Dow Chemical Company started research to develop a commercial process for the production of cellular polystyrene. This batch process, known as the tower process, consisted of blending polystyrene and a low-boiling compound, such as butylene or methyl chloride, in a large tower, with subsequent expansion into large foam logs, which were then cut into the desired boards or other shapes [2]. The Styrofoam (trademark of The Dow Chemical Company) was extruded as 30 cm diameter logs, which were cut into 90 cm lengths. This foam was used by the US Coast Guard and Navy as a flotation device for military equipment, and a floating buoy for the antisubmarine net during World War II. After the War, The Dow Chemical Company found new uses for Styrofoam; the material began to be used for decorative and novelty items, and also in those applications that took advantage of its insulating capability or its buoyancy. During these early stages of product development, advances were also made in process research at Dow. The first continuous extrusion process for producing a polystyrene foam was developed in the late 1940s through the early 1950s, and became the basis for the current extrusion process for the manufacture of polystyrene foams [3]. Other styrenic polymer foams were developed in the mid-1950s through the early 1960s. Examples are molded expanded polystyrene foam (MEPS), extruded polystyrene foam sheet, and expanded polystyrene loose-fill packaging material. Styrenic polymer foams have been commercially accepted in a wide variety of applications since the 1940s [1,4]. The total usage of polystyrene foams in the United States rose from about 4.10 x 105 metric tons in 1982 to an estimated 5.35 x 105 metric tons in 1987. It is expected to grow at a rate of 3–4% for the next several years [5]. For example, a recent Fredonia report on foamed plastics estimates that the 2008 volume will be 10.77 x 105 metric tons [6].

2

NOMENCLATURE

A cellular plastic is defined as a plastic, the apparent density of which is decreased substantially by the presence of numerous cells disposed throughout its mass [7]. In this chapter, the terms cellular polymer, foamed plastic, expanded plastic, and plastic foam are used interchangeably to denote all two-phase gas-solid systems in which the solid is continuous and composed of a synthetic polymer or rubber. The gas phase in a cellular polymer is usually distributed in voids or pockets called cells. If these cells are interconnected, the material is termed open-cell. If

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

205

these cells are discrete and the gas phase of each is independent of that of the other cells, the material is termed closed-cell. The nomenclature of cellular polymers is not standardized; classifications have been made according to the properties of the base polymer [8], the methods of manufacture, the cellular structure, or some combination of these. The most comprehensive classification of cellular plastics [9] has not been adopted and is not consistent with some of the current names for the more important commercial products. According to an ASTM test [10], foamed plastics are classified as rigid or flexible. A flexible foam is one that does not rupture when a 20 x 2.5 x 2.5cm piece is wrapped around a 2.5 cm mandrel at a uniform rate of 1 lap per 5 s, at 15–25 °C. Rigid foams rupture under this test. This classification is used here. In the case of cellular rubber, the ASTM uses several classifications based on the method of manufacture [11,12]. Cellular rubber is a general term covering all cellular materials that have an elastomer as the polymer phase. Sponge rubber and expanded rubber are cellular rubbers produced by expanding bulk rubber stocks, and are open-cell and closed- cell, respectively. Latex foam rubber, which is also a cellular rubber, is produced by frothing a rubber latex or liquid rubber, gelling the frothed latex, and then vulcanizing it in the expanded state. The term structural foam has not been exactly defined, but is used here to refer to rigid foams produced with densities greater than 320 kg/m3.

3 THEORY OF THE EXPANSION PROCESS Foamed plastics can be prepared by various methods. The most widely used, called the dispersion process, involves the dispersion of a gaseous phase throughout a fluid polymer phase, and the preservation of the resultant state. Other methods of producing cellular plastics include leaching out solid or liquid materials dispersed in the plastic, sintering small dispersed particles, and dispersing small cellular particles in the plastic. The latter processes are relatively straightforward techniques of lesser commercial importance. The expansion process has been the subject of extensive investigation because it is the foundation of foamed plastics [13–21]. In general, the expansion process consists of three steps: creation of small discontinuities or cells in a fluid or plastic phase, growth of these cells to a desired volume, and stabilization of the resultant cellular structure by physical or chemical means. 3.1

BUBBLE INITIATION

The development of bubbles within a liquid or polymer solution is generally called nucleation, although the term actually refers only to those bubbles that

206

K. W. SUH AND A. N. PAQUET

separate from the supersaturated liquid or polymer solution in the presence of an initiating site such as a surface irregularity, or at the interface of discrete phases. Gas for the bubbles has several sources: (1) dissolved gases that are normally present in the liquid or polymer solution and are forced into supersaturation by increased pressure; (2) low-boiling liquids that are incorporated into the system as blowing agents and are forced into the gas phase by increased temperature or decreased pressure; (3) gases produced as blowing agents, such as by the water–isocyanate reaction used for CO2 production in polyurethane foams; and (4) chemical blowing agents that decompose thermally to form a gas. Bubble nucleation is affected by a number of conditions. Physically, the effects of temperature, pressure, and in some cases humidity are fairly obvious. Other important parameters are surface smoothness of the substrate, surface characteristics of filler particles, presence and concentration of certain surfactants or nucleators, size and amount of second-phase liquid droplets, and the rate of gas generation. In many cases, bubbles of gas and other contaminants are already present in the liquid or polymer solution, and these serve as sites into which the gas may diffuse. The number and size of these gas bubbles may be another important factor in bubble development. 3.2

BUBBLE GROWTH

The initial bubble is ideally a sphere that grows as a result of the interaction of the differential pressure, A/>, between the inside and outside of the cell and the interfacial surface tension, y. The radius, r, of the bubble at equilibrium is related to these factors as follows: A/> = 27/r

(1)

Furthermore, the rate of growth of a cell depends on the viscoelastic nature of the polymer phase, the blowing agent pressure, the external pressure on the foam, the cell size, and the permeation rate of blowing agent through the polymer phase. Bubbles are enlarged by diffusion of gases coming out of solution into existing bubbles, coalescence of smaller bubbles, and thermal effects on internal bubble pressure according to the gas laws. As bubble size increases with decreasing density, the spherical shape of the bubbles is distorted into polyhedra with planar faces of uniform thickness. Low-density foams of cellular polymers tend to favor cell development into pentagonal dodecahedral cells, which have 12 five-sided membranes. Although this is not the configuration with the lowest surface energy, the angular symmetry is apparently more critical.

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS 3.3

207

BUBBLE STABILIZATION

As the cell walls are squeezed into polyhedra, a wall-thinning effect takes place, and liquid is drained from cell-wall faces into the lines of cell intersections to form ribs or struts, which are typically triangular in cross-section. This cell wall membrane thinning can continue to the point where the cell walls collapse and the cells open. This becomes a very important characteristic of most plastic foams, and affects properties such as thermal conductivity, moisture absorption, breathability, and load bearing. If bubbles are sparsely distributed, as in high-density foams, they will occur as spherical cells because that is the stable shape (with the lowest surface energy). If the foam is less dense, cell-wall stability is achieved by careful control of the factors that influence membrane thinning [16]. Capillary action and gravity may cause drainage of material from the membrane into the ribs. Capillary drainage is proportional to the square of the distance between rib junctions. Increasing viscosity of the fluid reduces the drainage effect. Viscosity increase may be caused by chemical reactions that increase molecular weight through polymerization or cross-linking, or by temperature reduction of high molecular weight thermoplastics. Depression of surface tension in local areas of the cell membrane promotes rupture. Ultimate stabilization occurs as a result of the physical effect of cooling below the second-order transition point (which prevents polymer flow). As final solidification is approached, the previously formed bubbles may be distorted by the system flow or by gravity, thereby producing anisotropy in the cellular structure. This effect must be taken into account by obtaining samples oriented in specific directions to the process flow when the physical properties of plastic foams are to be evaluated.

4 4.1

PROPERTIES AND THEIR RELATION TO STRUCTURE TEST METHODS

Several countries have their own test methods for cellular plastics, and the International Organization for Standards (ISO) Technical Committee on Plastics TC-61 has been developing international standards. Information can be obtained from the American Standards Institute. The most complete set of test procedures has been developed by the ASTM and is published in a new edition every year.

4.2

PROPERTIES OF COMMERCIAL PRODUCTS

It is evident that polystyrene foams have a broad range of physical properties (Table 10.1) [22,23]; the manufacturer should be consulted for the properties of

Table 10.1 Physical properties of commercial extruded polystyrene foams: ASTM C578 - Standard for Rigid, Cellular Polystyrene Thermal Insulation Property

Method

Type X

Type IV

Type VI

Type VII

Type V

Density (min. pcf) (kg/m3) Max. use temperature, (°C) Thermal resistance of I in (25.4mm)

ASTM C303 or ASTM D1822

1.30 (21)

1.60 (26)

1.80 (29)

2.20 (35)

3.00 (48)

— ASTM C177 or ASTM C5 18

74

74

74

80

80

5.60 (0.99) 5.40 (0.95) 5.00 (0.88) 4.65 (0.82) 15.0 (104)

5.60 (0.99) 5.40 (0.95) 5.00 (0.88) 4.65 (0.82) 25.0 (173)

5.60 (0.99) 5.40 (0.95) 5.00 (0.88) 4.65 (0.82) 40.0 (276)

5.60 (0.99) 5.40 (0.95) 5.00 (0.88) 4.65 (0.82) 60.0 (414)

5.60 (0.99) 5.40 (0.95) 5.00 (0.88) 4.65 (0.82) 100.0 (690)

40.0 (276)

50.0 (345)

60.0 (414)

75.0 (517)

100.0 (690)

1.1 (63)

1.1 (63)

1.1 (63)

1.1 (63)

1.1 (63)

0.3 2.0 24.0

0.3 2.0 24.0

0.3 2.0 24.0

0.3 2.0 24.0

0.3 2.0 24.0

(min h ft2 F/Btu)(m 2 C/W): @25°F(-3.9°C) @40°F(4.4°C) @75°F (23.9°C) @110°F (43.3°C) Compressive strength (min. psi) (kPa) Flexural strength (min. psi) (kPa) Water vapor permeance of 1 .0 in. thick (25.4mm), max. perm. (ng/Paam) Water absorption (max. % by volume) Dimensional stability (max. % change) Oxygen Index (min. volume %)

ASTM D 1621 or ASTM C165 ASTM C203, Method 1 Proc. A ASTM E9-8 ASTM C272 ASTM D2128 ASTM D2843

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

209

a particular product. The properties of styrenic structural foams are shown in Table 10.2. These values depend upon several structural variables and should be used only as general guidelines [24]. The properties of a foamed plastic depend on the properties of the base polymer and the geometry of the foam structure, often referred to as structural variables. The polymer phase description must include the additives present. The condition or state of the polymer phase, including its orientation, crystallinity, previous thermal history, and chemical composition, determine the properties of that phase. Polymer state and cell geometry are intimately related because they are determined by forces exerted during the expansion and stabilization of the foam. Density has the most important influence on mechanical properties of a foamed plastic of a given composition. Its effect has been extensively studied.

4.3

CELLS

A complete knowledge of the cell structure of a particular polymer would require the size, shape, and location of each cell. Because this is impractical, approximations are employed. Cell size has been characterized by measurements of cell diameter [25] and of average cell volume [26,27]. Mechanical, optical, and thermal foam properties depend on cell size. Cell shape is governed predominantly by final foam density and the external forces exerted on the cellular structure before its stabilization in the expanded state. In a foam prepared without such external forces, the cells tend to be spherical or ellipsoidal at gas volumes less than 70–80% of the total volume, and the shape of packed regular dodecahedra at greater gas volumes. These shapes are consistent with surface chemistry [26,28,29]. In the presence of external forces, the cells may be elongated or flattened. Cell orientation can influence many properties [22,23], An important characteristic of the cell structure is the extent of communication with other cells. This is Table 10.2

Typical physical properties of commercial structural foams

Property Glass-reinforced Density (kg/m 3 ) Tensile strength (kPa) Compressive strength (kPa) at 10% compression Flexural strength (kPa) Flexural modulus (GPa) Max. use temperature (°C)

ASTM test

ABS

D1623 D1621

No 800 18600 6900 25500 0.86 82

D790 D790

NORYL

High-impact polystyrene

Yes 850 48000 34500

No 800 22700

No 700 12400

20% 840 34500

82700 5.2

41400 1.7 96

31000 1.4

58600 5.2

210

K. W. SUH AND A. N. PAQUET

expressed as fraction of open cells. When many cells are interconnected, the foam has a large fraction of open cells and is termed an open-cell foam. In contrast, numerous noninterconnecting cells result in a large fraction of closed cells and a closed-cell foam. The nature of the opening between cells determines how readily gases and liquids can pass from one cell to another. Because of variation of the flow from cell to cell, a single measurement of the fraction of open cells does not fully characterize this structural variable, especially in a dynamic situation. 4.4

GAS COMPOSITION

In closed-cell foams the gas phase in the cells can contain blowing agent (socalled captive blowing agent), air, or other gases generated during foaming. The thermal and electrical conductivity can be profoundly influenced by the cell-gas composition. In open-cell foams the presence of air exerts only a minor influence on the static properties but does affect the dynamic properties such as cushioning. Table 10.3 lists properties of common blowing agents used to make polystyrene foams. 4.5

RIGID CELLULAR POLYMERS

Compressive strength and modulus are readily determined and have been widely used to characterize rigid plastic foam. Rigid cellular polymers generally do not exhibit a definite yield point, but rather an increased deviation from Hooke's law as the compressive load is increased. The compressive strength is usually reported at some definite deflection (5 or 10%); the compressive modulus is extrapolated to 0 % deflection unless stated otherwise. Structural variables that affect the compressive strength and modulus of a rigid plastic foam are, in order of decreasing importance, plastic phase composition, density, cell structure, and plastic state. The effect of gas composition is minor, with a slight effect of gas pressure in some cases. The strong effect of density and polymer composition on compressive strength and modulus are illustrated in Tables 10.1 and 10.2 [8]. The cell shape or geometry also influences compressive properties [8,22,23,27,28]. In fact, the foam cell structure is controlled in some cases to optimize certain physical properties of rigid cellular polymers. In general, compressive strength and modulus of low-density foams may be expressed as Strength or modulus = A(7* where A and a constants and Q represents the foam density. This relationship is illustrated in Figure 10.1.

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS Table 10.3

211

Physical properties of common blowing agents"

Agent

MW

Solubility

Permeability

BP

HV

TC

Crit. T.

MeCl CFC-12 EtCl HCFC-142b CO2 Ethane Butane Pentane Ethanol HCFC-22 HCFC-124 HFC-152a HFC-143a HFC-134a HFC-125 Water Nitrogen Oxygen Argon

50.5 120.9 64.5 100.5 44.0 30.1 58.1 72.2 46.1 86.5 136.5 66.1 84.0 102.0 120.0 18.0 28.0 32.0 39.9

4.6 1.7 20.0 6.7 0.4 0.3 1.9 13.0 24.0 1.6 ~ 1.0 1.8 ~ 1.5 ~1 ~ 1.0

980.0 0.17 84.0 0.1 1430.0 31.0 4.0 3.5 390.0 19.0 0.1 8.4 <2.0 0.3 <0.3 57 343 167

101.3 39.5 91.3 53.3 136.3 116.6 87.5 85.3 200.0 55.9 40.3 79.6 -50.0 50.8 38.0 540.0 47.4 50.9 -40.0

0.076 0.065 0.077 0.079 0.114 0.12 0.114

0 0 0

-24.2 -29.8 12.4 -9.4 -88.2 -88.6 -11.8 36.1 78.3 -40.8 -10.2 -24.9 -46.0 -26.5 -48.5 100.0 -196.0 -183.0 -185.7

143.0 112.0 187.0 137.1 31.0 32.2 134.0 196.6 243.0 96.0 126.6 197.2 73.1 100.6 66.3 374.1 -147.0 -118.4 -122.3

0.079 0.085 0.078 0.091 0.077 0.126 0.178 0.183 0.123

" MW = molecular weight; Solubility = parts per hundred per atmosphere (25 °C); Permeability = cm3 mil/day 100 in2 atm (25 °C); BP = boiling point (°C); HV = latent heat of vaporization at boiling point (cal/g); TC = thermal conductivity of gas at 24 °C (BTU in/h ft2 °F); Crit. T. = critical temperature (°C).

Strengths and moduli of most polymers increase with decreasing temperature [29]. Tensile strength and modulus of rigid foams vary with density in much the same manner as the compressive strength and modulus [8,22,23]. The structural variables most important to the tensile properties are polymer composition, density, and cell shape [30]. Flexural strength and modulus of rigid foams increase with increasing density in the same manner as the compressive and tensile properties. Shear strength and modulus of rigid foams depend on the polymer composition and state, density, and cell shape. Shear properties increase with increasing density and decreasing temperature [30]. 4.6

CREEP

The creep characteristic of plastic foams is important in structural applications. Creep, either short-term or long-term, is the change in dimensions caused by constant stress. The deformation of polystyrene foam under various static loads

K. W. SUM AND A. N. PAQUET

212

300

2000

200 1000

2
700

. •SGO'S§

500

100 2

70

s.

O 60 ~

50

II "

300

o. E o

C.

E o U

200

U

100

20

30

50 70 100 Density (kg/m2)

10 200

Figure 10.1 Compressive strength at yield vs density for extruded polystyrene foam. Unpublished data of J. M. Corbett. To convert kg/m3 to Ib/ft3, multiply by 0.0624

is shown in Figure 10.2 [31]. Short-term creep exists in foams at all stress levels, but there is no long-term creep below a threshold stress level (see Table 10.4). The successful application of time-temperature superposition [32] for polystyrene foam allows the prediction of long-term behavior from short-term measurements. This is valuable in construction applications, where creep is an important consideration.

4.7

STRUCTURAL FOAMS

Structural foams are usually produced as fabricated articles in injection molding or extrusion processes; they vary in properties (see Table 10.2). Again, the most important structural variables are polymer composition, density, and cell size and shape. Structural foams have relatively high densities (more than 320 kg/m3), and cell structures are primarily comprised of holes, in contrast to the pentagonal dodecahedral structure characteristic of lowdensity plastic foams. Because structural foams are generally not uniform in cell structure, they exhibit a considerable variation in properties with particle geometry (see Table 10.5) [24]; these relations can provide valuable guidance.

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

ca

213

4 -

10

20

30 Time (days)

40

50

60

Figure 10.2 Deformation under various loads for 48kg/m3 (Ib/ft3) molded polystyrene foams [31]. A, 224 kPa; B, 193kPa; C, 103kPa; D, 38kPa. To convert kPato psi, multiply by 0.145. Reproduced from W.B. Brown, Plast.Prog., 1960, 1959, 149 Table 10.4 Minimum load to cause long-term creep in molded polystyrene foama Foam density (kg/m3)b

Load (kPa)c

16 96 160

50 165 455

" Ref. 31. h To convert kg/m3 to Ib/ft 3 , multiply by 0.0624. c To convert kPa to psi, multiply by 0.145.

5 5.1

THERMAL PROPERTIES THERMAL CONDUCTIVITY

The thermal conductivity of cellular polymers has been thoroughly studied in heterogeneous materials [33,34] and plastic foams [25,35-38]. Heat transfer can be separated into its component parts as follows: i

2 _i_ 2

i 3

i_ 2

A — As ~t Ag ~\- Ar -f- Ac

(">\ \2.)

where A = total thermal conductivity, and As, Ag, Ar, and Ac represent solid conduction, gaseous conduction, radiation, and convection, respectively.

214 Table 10.5

K. W. SUH AND A. N. PAQUET Design criteria for low-pressure structural foam0 Equationb

Mechanical property

Comments

Tensile strength

Tfoam =

T0(@f0am/0o)

Compressive modulus Shear modulus Flexural modulus

Cfoam — Q((?foamA?o)2 Shfoam = Sh()(Qf0am/Qn)* -Sfoam = So(Q foam /Qo>

2

•Sfoam — •Sotefoam/Po) Impact

/roam = A)(efoam/^o)4(^/<5o)2

Fatigue Creep

Ffoam = Fo(Q{oam/Qo)2 2 Cr(oam = Cr0(Qfoam/Q0)

Coefficient can increase with temperature Most data on short columns Limited data Empirical, uncontrolled skin thickness, low foam density Law of mixtures, thick skins, high foam density Tentative, few data from many test procedures Ver Y tentative, rule-of-thumb Very tentative, rule-of-thumb

a

Ref. 24 data extrapolated from stress-strain curves; apparent values only. Q0 is the density of the polymer; (5, <5o are thicknesses of the foamed and unfoamed part, respectively. b

As a first approximation [33], the heat conduction of low-density foams through the solid and gas phases can be expressed as the product of the thermal conductivity of each phase and its volume fraction. Most rigid polymers have thermal conductivities of 0.07–0.28 W/(mK), and the corresponding conduction through the solid phase of a 32 kg/m3 foam (3 vol.%) ranges from 0.003 to 0.009 W/(mK). In most cellular polymers this value is determined primarily by the density and the polymer-phase composition. Smaller variations result from changes in cell structure. Although conductivity through gases is intrinsically much lower than through solids, the amount of heat transferred through the gas phase in a foam is usually the largest component of the total heat transfer because of the large gas phase volume (approximately 97 vol.% in a 32 kg/m3 foam). Thermal conductivities of gases used as blowing agents are given in Table 10.3. The halocarbons have much lower values than oxygen and nitrogen, and therefore are used to prepare cellular polymers with lower /.. Gases in the cells can become mixed with air, and the resultant / can be estimated from the following equation: Am = «!/!

+/I 2 A 2

(3)

where xm is the A of the gaseous mixture, n\ and ni are the mole fractions of gases 1 and 2, and X\ and A2 are the thermal conductivities of gases 1 and 2. Changes in total 1 calculated by Equations (2) and (3) agree closely with experimental measurements [26,39-41].

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

215

Ordinarily, convection cannot be detected in cells of diameter less than 4mm [25,36,39]. Since most cellular polymers have cell diameters under 4mm, convection can be ignored. Radiant heat transfer through cellular polymers has also been studied [25,36,39,42,43]. The variation in total thermal conductivity with density is similar for all cellular polymers (see Figure 10.3). The increase in A at low densities is due to increased radiant heat transfer and at high densities to an increasing contribution of As. The thermal conductivity of most materials decreases with temperature. When foam structure and gas composition are not influenced by temperature, the A of the cellular material falls with decreasing temperature. When the composition of the gas phase changes, such as upon the condensation of a vapor, the relationship of X to temperature is much more complex [25,39,44]. The thermal conductivity of a cellular polymer can change upon aging under ambient conditions if the gas composition is influenced by aging. This is the case when oxygen or nitrogen diffuses into polystyrene foams containing a hydrochlorocarbon or a hydrochlorofluorocarbon blowing agent in the cells [16,25,38,39,44–53]. Thermal conductivity of foamed plastics varies with thickness [43]. This has been attributed to the boundary effects of the radiant contribution to heat transfer. 0.043

0.30

„

0.26

!a £ 3

0.22 ffl

.~ 0.032

x

0.18 | 3 T3 O

o 0.14 -=

0.020

50

100 Density, (kg/m3)

150

0.10 200

Figure 10.3 Effect of density on thermal conductivity of rigid cellular polymers. A, polystyrene [25]; B, polystyrene [37]; C, polyurethane–air [37]; D, polyurethane–CFC 11 (CCI3F) [70]; E, polyurethane [37]; F, phenol–formaldehyde [37]; G, ebonite [37]. To convert kg/m3 to Ib/fr, multiply by 0.0624. Reproduced from F. O. Guenther, SPE Transactions, 2, 243 (1962), with permission from the society of Plastics Engineers, Brookfield, Connecticut, USA

216

5.2

K. W. SUH AND A. N. PAQUET

COEFFICIENT OF LINEAR THERMAL EXPANSION

The coefficients of linear thermal expansion of polymers are higher than those of most rigid materials at ambient temperatures because of the supercooledliquid nature of the polymeric state. This large coefficient is carried over directly to the cellular state. A variation of this property with density and temperature has been reported for polystyrene foams [54] and foams in general [8]. When cellular polymers are used in large structures, the coefficient of thermal expansion must be considered carefully because of its magnitude compared with that of most nonpolymeric structural materials. Many designs have been successful [55] when this fact has been taken into account.

5.3

MAXIMUM SERVICE TEMPERATURE

The maximum service temperature of cellular polymers cannot be defined precisely, because the cellular materials, like the parent polymers [56], gradually decrease in modulus as the temperature rises, rather than undergoing a sharp change in properties. The upper temperature limit of use for most cellular polymers is determined predominantly by the plastic phase. The act of fabricating a cellular state normally imposes stress on the polymer phase. This stress may tend to relax at a temperature below the heat distortion temperature of the unfoamed polymer. Additives in the polymer phase, or a plasticizing effect of the blowing agent on the polymer, affects the behavior of the cellular material and of the unfoamed polymer in the same way. Typical maximum service temperatures are given in Tables 10.1 and 10.2.

5.4

MOISTURE RESISTANCE

Plastic foams have advantages over other thermal insulation when exposed to moisture, particularly where subjected to a combination of thermal and moisture gradients. In some cases the foams are also exposed to freeze–thaw cycles. The behavior of plastic foams has been studied under laboratory and field conditions. In plastic roof insulations under controlled thermal gradients, the moisture gains found in polyurethane are greater than those of bead polystyrene, and much greater than those of extruded polystyrene [9]. We conclude from findings on moisture absorption and freeze–thaw resistance of various insulations, and the effect of moisture on thermal performance, that in protected-membrane roofing applications the order of resisting moisture pickup is extruded polystyrene > polyurethane > molded polystyrene [57]. Water absorption values for insulation in use for 5 years were 0.2 vol% for extruded polystyrene, 5 vol% for polyurethane without skins, and 8–30 vol%

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

217

for molded polystyrene. These values correspond to increases in /I of 5–265%. For below-grade applications, extruded polystyrene was better than molded polystyrene or polyurethane without skins in terms of moisture and thermal resistance. Increases in water content and thermal conductivity are related [58–62].

5.5

ENVIRONMENTAL AGING

Environmental aging of cellular polymers is important in most applications. The response of cellular materials to light and oxygen is governed almost entirely by the composition and state of the polymer phase [8]. Expansion into a cellular state increases the surface area; reactions of the foam with vapors and liquids are correspondingly faster than those of solid polymer. All cellular polymers deteriorate under the combined effects of light or heat and oxygen; this may be alleviated by additives [63].

5.6

OTHER PROPERTIES

The acoustic properties of polymers are altered in a cellular structure. Sound transmission changes only slightly, because it depends predominantly upon the barrier density, in this case the polymer phase. Therefore, closed-cell cellular polymers by themselves are poor materials for reducing sound transmission. They are, however, effective in absorbing sound waves of certain frequencies [64]. Materials with open cells on the surface are particularly effective in this respect. The combination of other advantageous physical properties with fair acoustic properties has led to the use of plastic foams in soundproofing [65,66]. The sound absorption of a number of cellular polymers has been reported [7,64,65,67]. The permeability of cellular polymers to gases and vapors depends upon the proportion of open cells as well as on the polymer-phase composition and state. The presence of open cells in a foam allows gases and vapors to permeate the cell structure by diffusion and convection (and thus to have high permeation rates). In closed-cell foams the permeation of gases or vapors is governed by the composition of the polymer phase, gas composition, density, and cellular structure [40,45,65,68,69]. The permeability of the gaseous blowing agents is an important property to consider for blowing agent selection. Rodents chew through cellular polymers but do not ingest the foam as a foodstuff. The resistance to rot, mildew, and fungi is related to moisture absorption [64]. Therefore, open-cell foams support such growth better than closed-cell foams. High humidity and temperature are necessary for the growth of microbes on any plastic foam.

218

6

K. W. SUH AND A. N. PAQUET

COMMERCIAL PRODUCTION AND PROCESSING

6.1

MANUFACTURING PROCESS

Cellular plastics and polymers have been prepared by processes involving many methods of cell initiation, growth, and stabilization. The most convenient method of classifying these methods appears to be one based on cell growth and stabilization. According to Equation (1), the growth of the cell depends on the pressure difference between the inside of the cell and the surrounding medium. Pressure differences may be generated by lowering the external pressure (decompressing), or by increasing the internal pressure in the cells (generating pressure). Other methods of creating the cellular structure are by dispersing gas or solid in the fluid state and stabilizing this cellular state, or by sintering polymer particles in a structure that contains a gas phase. Foamable compositions in which the pressure within the cells is increased relative to that of the surroundings are called expandable formulations. Both chemical and physical processes are used to stabilize plastic foams from expandable formulations. There is no single name for the group of cellular plastics produced by the decompression processes. The operations used are extrusion, injection molding, and compression molding. Physical or chemical methods may be used to stabilize the products. Polystyrene foam is a rigid plastic material with a wide range of densities and applications; it is produced in five basic types: extruded board, extruded sheet, expanded bead molding, injection-molded structural foam, and expanded loose-fill packaging. A useful survey of manufacturing methods for polystyrene foams can be found in Throne [15].

6.2

6.2.1

DECOMPRESSION EXPANSION PROCESSES, PHYSICAL STABILIZATION

Polystyrene

Extruded polystyrene board is blown with HCFC 142b or HFC 134a and sometimes auxiliary blowing agents, and is used primarily as an insulating material in building construction [70–75]. The blowing agent CFC 12, HCFC 142b or HFC 134a plays an important role in achieving the unique properties of extruded polystyrene boardstock, such as high insulation value per unit thickness, high dimensional stability, outstanding resistance to moisture and freeze– thaw deterioration, excellent compressive strength, and good handling (nonflammable and nontoxic) and physical properties. Extruded polystyrene boardstock is used in residential and industrial sheathing, and in roofing in commercial buildings. The foam density ranges from 21 to 40 kg/m3. Extruded polystyrene foam insulation was introduced in the early 1940s by the Dow Chemical Company with the trade name Styrofoam [76,77]. A mixture of

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

219

polystyrene and volatile liquid blowing agent is extruded and expanded through a die to form boards in various sizes. Owens Corning manufactures a similar extruded polystyrene foam under the trade name FoamulaR. A single-screw tandem extrusion line is used to produce foam boardstock in a vacuum chamber connected to a barometric leg that acts as a vacuum seal [78]. Owens Corning also licenses their process to a number of international parties. In 1982, Minnesota Diversified Products started to produce extruded polystyrene foam insulation under the trade name Certifoam, by the LMP process [79]. This process uses a co-rotating twin-screw extruder with a single-screw extension as a cooling section, a combination motionless mixer–homogenizer and heat exchanger, a flat die, and finishing equipment for sizing and curing. More recently Pactiv, formerly Tenneco Packaging, has manufactured foamed polystyrene sheet and boardstock. A tandem extrusion line is used to produce polystyrene foam insulation products. In the extrusion process for producing cellular polystyrene [76], a solution of blowing agent in molten polymer, formed in an extruder under pressure, is forced through an orifice on to a moving belt at ambient temperature and pressure. The blowing agent vaporizes and causes the polymer to expand. The polymer simultaneously expands and cools under conditions that give it enough strength to maintain dimensional stability at the time corresponding to optimum expansion. Stabilization is due to cooling of the polymer phase to a temperature below its glass transition temperature. Cooling is effected by vaporization of the blowing agent, gas expansion, and heat loss to the environment. Polystyrene foams produced by the decompression process are commercially available in the density range 21–40kg/m3 and even higher [80]. Low-density polystyrene foam sheet was first produced by the extrusion of expandable polystyrene beads or pellets containing pentane as blowing agent [81,82]. Currently, polystyrene foam is extruded in a single-screw tandem line or in a twin-screw extruder. Cellular polystyrene can also be produced by an injection-molding process. Polystyrene granules containing dissolved liquid or gaseous blowing agents are used as feed in a conventional injection-molding process [15,83]. With close control of time and temperature and use of vented molds, high-density cellular polystyrene moldings can be obtained. Two notable methods to produce microcellular foams include gas supersaturation in combination with an extrusion process developed by MIT/Trexel [84–86] and the continuous extrusion process by Dow [87,88]. Super-insulating materials are made by the encapsulation of a filler material inside a barrier film, aluminum foil, or metallized film. These materials exhibit 5-7 times the R-value of typical nonvacuum insulating materials depending on vacuum level and barrier and filler type. Uses for these VIPs (vacuum insulation panels) include refrigeration and controlled-temperature shipping containers.

220

6.2.2

K. W. SUH AND A. N. PAQUET

Thermoplastic Structural Foams

Structural foams with an integral skin, cellular core, and a high strengthto-weight ratio are formed by injection molding, extrusion, or casting, depending on requirements [89,90]. Most widely used are the Union Carbide lowpressure process [91] and the USM high-pressure process [92]. In the low-pressure process, a resin containing a blowing agent is forced into the mold, where it expands to fill the mold under pressures of 690–4100 kPa (100–600 psi). This produces structural foam products with a characteristic surface-swirl pattern caused by the collapse of cells on the surface of molded articles. In the high-pressure process, a resin melt containing a chemical blowing agent is injected into an expandable mold under high pressure. Foaming begins as the mold cavity expands. This produces structural foam products with very smooth surfaces, because the skin is formed before expansion takes place. Extruded structural foams are produced with conventional extruders and a specially designed die with an inner, fixed torpedo located at the center of its opening. The extrudate from this die is hollow. The outer layer of the extrudate cools and solidifies to form solid skin; the remaining extrudate expands toward the interior of the profile. The Celuka process, developed by Ugine-Kuhlmann [93], is widely used commercially. Large structural foam products are obtained by casting expandable plastic pellets containing a chemical blowing agent in aluminum molds on a chain conveyor. After the mold has been closed and clamped, it is conveyed through a heating zone where the pellets soften, expand, and fuse to form the cellular product. The mold is then passed through a cooling zone. These structural foam products have a uniform, closed-cell structure but no solid skin. Injection-molded structural foam is widely used for high-density articles such as picture frames, furniture appliances, housewares, utensils, toys, pipes, and fittings. Most of these products are produced by injection molding or profile extrusion methods, from impact-modified polystyrene. Almost all high-density foam products are produced with a decomposable chemical blowing agent that releases nitrogen or carbon dioxide. Medium-density products can be produced with a physical or chemical blowing agent or a combination of both. Typical decomposable chemical blowing agents are sodium bicarbonate and azodicarbonamides.

6.2.3

Other Methods

Some plastics cannot be obtained in a low-viscosity melt of solution that can be processed into a cellular state. Sintering of solid plastic particles and leaching of soluble inclusions from the solid plastic phase give the required dispersion.

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

221

Microporous polymer systems consisting of essentially spherical, interconnected voids with a narrow range of pore- and cell-size distribution have been produced from a variety of thermoplastic resins by phase separation [94]. If a polyolefin or polystyrene is insoluble in a solvent at low temperatures but soluble at high temperatures, the solvent can be used to prepare a microporous polymer. When the solution containing 10-70% polymer is cooled to ambient temperature, the polymer separates as a second phase. The remainder can be extracted. These microporous polymers may be used in rnicrofiltration or as controlled-release carriers for chemicals.

7

APPLICATIONS

Concern over energy conservation and safety has stimulated growth in applications for insulation and cushioning in transport. A healthy economy is also expected to increase the demand for cushioning in furniture, bedding, and flooring, as well as for packaging. Structural foams are widely used as substitutes for wood, metal, or unfoamed plastics. Cushioning is the largest single application of cellular polymers. As might be expected, the flexible foams predominate in this field. However, during the past few years the volume of applications in the structural area, in packaging, and in insulation has overtaken and passed the volume of flexible foams (see Table 10.6) [5]. Styrenic polymer foams have been commercially accepted in a wide variety of applications since the 1940s [1,4]. As mentioned in the Introduction, the total usage of polystyrene foams in the United States rose from about 4.10 x 105 metric tons in 1982 to an estimated 5.35 x 105 metric tons in 1987. It now is expected to grow at a rate of 3-4% for the next several years [5]. For example, a recent Fredonia report on foamed plastics estimates 2003 and 2008 volumes to be 9.48 x 105 and 10.77 x 105 metric tons, respectively [6].

7.1

THERMAL INSULATION

Thermal insulation is the second largest application of cellular polymers, and the largest application for the rigid materials, because of their thermal conductivity, ease of application, cost, moisture absorption, and water vapor transmission (or permeance). The thermal conductivities of various commercial insulating materials are given in Table 10.7. Plastic foams containing a captive blowing agent have much lower thermal conductivities than other insulating materials.

222

Table 10.6

K. W. SUH AND A. N. PAQUET US cellular polymer market" 103 metric tons

Annual growth (%)

Applications and resins

1967

1982

1987*

1995*

1967-82

1982-87

1987-95

Insulation Flooring Construction Cushioning Furniture Packaging Transportation Consumer Bedding Appliances Other Total Flexible polyurethane Rigid polyurethane Polystyrene Poly( vinyl chloride) Others Total

58 20 9 52 40 43 76 44 18 14 68 441 181 68 125 61 6 441

261 98 136 195 103 177 140 136 57 40 225 1567 511 248 410 232 165 1567

347 122 197 249 135 236 186 171 77 51 308

472 154 288 336 175 311 238 225 113 61 408 2781 844 449 699 413 376 2781

10.6 11.3 20.2

5.9 4.5 7.8 5.0 5.6 5.9 5.9 4.8 6.3 4.9 6.4 5.8 5.2 6.2 5.5 6.0 7.8 5.8

3.9 3.0 4.8 3.8 3.3 3.5 3.1 3.4 4.9 2.4 3.6 3.7 3.2 3.7 3.4 3.6 5.8 3.7

a b

2080

658 336 535 311 240 2080

9.2 6.4 9.9 4.1 7.9 7.9 7.2 8.3 8.8 7.1 9.0 8.3 9.3 24.9 8.8

Ref. 5. Courtesy of Predicasts, Inc. Projected.

Table 10.7

Thermal conductivities of commercial insulation materials

Material

Density (kg/m3)

Thermal conductivity @ 20 CC [W/(m/K)]

Sawdust Glass foam Fiberboard: Wood Corkboard Rock wool Cellulose Urea-formaldehyde Fiber glass Polystyrene foam: Molded Extruded Polyurethane or

16 144

0.065 0.058

368 112 32 44 13 10

0.058 0.039 0.048 0.039 0.035 0.043

16 32 24

0.041 0.029 0.023

foam

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS 7.2

223

REFRIGERATION

The low thermal conductivity of polyurethanes, plus the ease of application and structural properties of foamed-in-place materials, affords great freedom of design. As a result, rigid polyurethane foams have displaced rock wool and glass wool in freezers and refrigerators. Large institutional and commercial refrigerators, freezers, and cold storage areas, including cryogenic equipment and gas tanks, are insulated with polystyrene or polyurethane foams. Polystyrene foam is popular where cost and moisture resistance are important; polyurethane is used for spray application. Polystyrene foam is also used in load-bearing sandwich panels for lowtemperature applications.

7.3

CONSTRUCTION

The use of cellular plastics for wall and ceiling insulation of residential buildings has increased more than 200 % over the past decade. Extruded polymeric foam is found in residential construction as sheathing, perimeter and floor insulation under concrete, and combined plaster base and insulation for walls. Both polystyrene and polyurethane foams are highly desirable roof insulants. Residential construction applications may be divided into three areas: insulation of new buildings, retrofit insulation of old buildings, and insulation of mobile homes. In residential sheathing insulation, fiberboard is still the most widely used product, although the use of extruded and molded polystyrene foam and of foil-faced isocyanurate foam is increasing. For cavity-wall insulation, mineral wool, polyurethane, urea-formaldehyde, and fiber glass are widely used, although fiber glass batt is the most economical for stud wall construction. In mobile and modular homes, cellular plastics are widely used because of their light weight and insulation value. Cellular polymers, especially polystyrene and polyurethane, are also widely used for pipe and vessel insulation. The use of cellular rubber and cellular poly(vinyl chloride) in insulation for small pipes is attributed to their ease of application, combustion properties, and low thermal conductivity. The insulating value and mechanical properties of rigid plastic foams have led to the development of several novel methods of building construction, including polyurethane foam panels as unit structural components [95] and expanded polystyrene as a concrete base in thin-shell construction [96].

7.4

STRUCTURAL COMPONENTS

In most applications, structural foam parts are used to replace wood, metals, or solid plastics. They are widely used in appliances, automobiles, furniture, and

224

K. W. SUM AND A. N. PAQUET

construction. The building and construction industry accounts for more than half of the total volume [95–97]. High-impact polystyrene is the most widely used structural foam, followed by polypropylene, HDPE (high-density polyethylene), and PVC [poly(vinyl chloride)]. The sandwich-type structure of polyurethanes with a smooth integral skin produced by reaction injection molding provides a high degree of stiffness and excellent thermal and acoustic properties. 7.5

MARINE APPLICATIONS

The moisture resistance, low cost, and low-density closed-cell structure of many cellular polymers resulted in their acceptance for buoyancy in boats, floating docks, and buoys. Because each cell is a separate flotation unit, these materials cannot be destroyed by a single puncture. Foamed-in-place polyurethane between thin skins of high tensile strength is used in pleasure craft [98]. Other cellular polymers that have been used where buoyancy is needed are produced from polystyrene, polyethylene, poly(vinyl chloride), and certain types of rubber. Foams made from styrene-acrylonitrile copolymers are resistant to petroleum products [99,100].

7.6

OTHER USES

Cellular plastics, mainly polystyrene, are used for display and novelty pieces because of ease of fabrication, light weight, attractive appearance, and low cost. Phenolic foam is used in floral displays, where it can hold large amounts of water for long periods. Polyurethane and polystyrene foams are also used in other floral applications. Cellular poly(vinyl chloride) is used in toys and athletic goods. Cellular urea-formaldehyde and phenolic resin foams have been used to some extent in interior sound-absorbing floors [101]. In general, cost, flammability, and cleaning difficulties have prevented extension into the acoustic tile market. Plastic foams are used in anechoic chambers [66].

7.7

ENERGY CONSIDERATIONS IN FOAM INSULATION

As energy costs continue to rise, energy conservation in building becomes increasingly important. The principal heat losses are through ceilings, walls, windows, doors, floors and foundations, and through air infiltration. Insulated sheathing assemblies are widely used in residential construction. They consist of an interior finish, a stud cavity with or without insulation, exterior insulating sheathing, and exterior siding. Because the R value of an insulation increases with decreasing temperatures, an insulation with an R value measured at 24 °C will perform better than

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS

225

predicted by that R value in winter (when the mean temperature is less than 24 °C) and, by similar reasoning, will perform worse than predicted in summer. In addition, some sheathing materials have tongue-and-groove edges, which probably reduce air infiltration. Furthermore, insulation of sheathing material may reduce air infiltration and the AT across the cavity, and thus improve thermal performance by reducing convection. In industrial and commercial construction, numerous building assemblies have been used successfully for many years. As in residential insulation, roof insulation is the most effective. The roof deck is usually flat and requires new waterproofing. In conventional overdeck assembly, the insulation is located above the roof deck structure and is covered by the built-up roof. Other techniques for roof insulation include membrane assemblies (IRMA) [102], underdeck assemblies, and structural panel assemblies. For wall construction, cavity-wall assemblies, interior-wall substrate assemblies, insulated frame construction, stucco-base assemblies, and sandwich-panel curtain-wall assemblies are used [103]. In cavity-wall assemblies, rigid board insulation is attached to the outside of the structural masonry wall before installing the exterior veneer brick. In interior-wall substrate assemblies, a layer of rigid plastic foam or mineral fiber may be attached to the interior surface of the load-bearing masonry wall; an interior finish material is added, typically a 1.27 cm thick gypsum board. The insulated frame assembly with mineral-fiber insulation is commonly used in the frame cavity. Stucco-base assemblies may also be applied to either frame-wall or masonry-wall construction. In sandwich panels, rigid plastic may be foamed in place between two rigid facings, or foam boardstock may be bonded to two facings to form a structural panel. In foundation insulation, rigid plastic foams are widely used. A covering over the foam provides protection against ultraviolet light and physical damage. The properties of extruded polystyrene foam make it a desirable insulation for foundation applications. Historically, most of the State and Federal specifications and building codes were developed to protect occupants from fire. Today, however, many building codes and specifications require energy conservation, and specify permissible heat loss or gain, as well as the use and manner of installation of insulation products.

8 8.1

ENVIRONMENTAL, HEALTH AND SAFETY CONSIDERATIONS FLAMMABILITY

Plastic foams are organic and therefore combustible. They vary in their response to small ignition sources [104]. All plastic foams should be handled, transported, and used according to manufacturers' recommendations and local and national regulations.

226

K. W. SUH AND A. N. PAQUET

The blowing agents used for virtually all plastic foams are mixtures of inert gases (CO2, N2, H2O), chemical blowing agents that release inert gases, hydrocarbons containing 3-5 carbon atoms, chlorinated hydrocarbons, chlorofluorocarbons such as CFC 11, CFC 12, CFC 113, and CFC 114, and hydrochlorofluorocarbons such as HCFC 142 b, HCFC 141 b, and HCFC 123. It is expected that hydrofluorocarbons (HFC 134a, HFC 152a) will be used in the future. The hydrocarbons are flammable and pose a fire hazard during manufacturing. Increased attention is being given by industry, government, and consumers to the behavior of plastic foams in fire. Small-scale laboratory tests are not predictive of other fire situations [105]. All plastic foams are combustible, some more readily than others. Small quantities of certain additives [106,107] markedly improve the behavior of the foam in the presence of small fire sources. This, too, has been an area of increased attention by the same stakeholders. The effort is focused on improving the efficacy of the fire retardant additives. All plastic foams should be used properly, following the manufacturers' recommendations and governmental regulations. 8.2

BLOWING AGENTS AND ENVIRONMENTAL ISSUES

The ability to make a polystyrene foam obviously requires the use of blowing agents, as already described. Also described has been how blowing agents have influenced the performance properties of the foam. Hence foam developments have often been motivated by issues associated with the blowing agents. Since the theory of ozone depletion by chlorofluorocarbons [108–110] was first published in 1974, extensive research has been conducted. Mathematical models were compared with experimental measurements of CFC concentrations in the upper atmosphere. Because of the difficulty of predicting the outcome of this research, industry initiated programs to develop alternative materials [111]. Subsequently, industry specialists, academics, and governmental agents have developed a number of working agreements. The first significant agreement was the Montreal Protocol implemented in 1987 [112]. Dow was the first polystyrene foam producer to develop a formulation that replaced ozone-depleting substances such as CFC 12 with HCFC 142b that has substantially reduced ozone depletion potential (ODP) of the insulating blowing agent used to make the foam [72–74]. The Montreal Protocol was followed by further agreements to eliminate the use of HCFCs (Kyoto Protocol). Many countries have since taken a more aggressive schedule for eliminating blowing agents that have ODP values greater than zero. Also, attention has recently been focused on the Global Warming Potential (GWP) of the blowing agents. This is an area that is unsettled at present, but expected to become a significant consideration in the use of certain blowing agents. An especially important point to consider is that the best alternatives for desirable insulating gases are HFCs [73,75]. The HFC alternatives have a zero ODP, but GWP values that concern some parties.

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS Table 10.8

227

Flammability and environmental considerations for blowing agents" Relative hazard properties

Agent

ILV

Reactivity

LFL(%)

VOC

ODP

MeCl CFC-12 EtCl HCFC-142b CO2 Ethane Butane Pentane Ethanol HCFC-22 HCFC-124 HFC-152a HFC-143a HFC-134a HFC-125 Water Nitrogen Oxygen Argon

50 1000 150 1000 5000 N 1000 600 1000 1000 500* 1000 NE 1000* 1000b N N N N

Al Low Al Low Low Low Low Low Moderate Low Low Low Low Low Low Moderate Low Low Low

8.2 NF 3.6 6.9 NF 3.0 1.8 1.5 3.4 NF NF 3.9 9.2 NF NF NF NF F NF

Yes Exempt Yes Exempt Exempt Exempt Yes Yes Yes Exempt Exempt Exempt Exempt Exempt Exempt No No No No

Low 1.0 low 0.065 0.0 0.0 0.0 0.0 0.0 0.055 0.02 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

GWP

4500

1800 1

1600 150 47 1000 420 860 0 0 0 0

" TLV = threshold limit value (ppm); LFL = Lower flammability limit (%,v/v); VOC = volatile organic compound as defined by the Clean Air Act; ODP = ozone depletion potential (CFC-11 has been assigned a value of 1, all others relative to CFC-11); the potential to contribute chlorine or other ozone depleting elements to the stratosphere; GWP = global warming potential; these values are relative to carbon dioxide which has been assigned a value of 1; these data are less certain; three criteria are used, radiative forcing, atmospheric lifetime, and temporal integrating period; NF = nonflammable; F = flammable; N = none; NE = none established. DuPont Acceptable Exposure Limit.

Dow and BASF have independently developed polystyrene foam formulations that do not contain HCFCs or HFC blowing agents [88,113–118]. Since these foams do not contain captive gases that have desirable insulating properties, the use of infrared attenuators has been employed to improve the thermal conductivity of the foam products. Certain organic compounds generate smog photochemically (VOC). Because halogenated hydrocarbons have low reactivity in the lower atmosphere, substitution of photochemically reactive compounds for the current blowing agents may reduce the ozone depletion in the stratosphere, but could adversely affect indoor air quality. Therefore, interaction with the total environment must be considered in developing environmentally acceptable blowing agents. Table 10.8 lists a number of blowing agents and their respective properties for flammability and environmental consideration.

228

K. W. SUH AND A. N. PAQUET

REFERENCES 1. Karpiuk, R. S. Dow Research Pioneers; Recollections 1888–1949, Pendell Publishing: Midland, MI, 1984. 2. Mclntyre, O. R. US Patent 2515250 (1947) (to The Dow Chemical Company). 3. McCurdy, J. L.; DeLong, C. E. US Patent 2 669 751 (1954) (to The Dow Chemical Company); McCurdy, J. L.; DeLong, C. E. US Patent 2 740157 (1956) (to The Dow Chemical Company). 4. Goggin, W. C.; Mclntyre, O. R. Br. Plast. 1947, 19, 528. 5. Kratzschmer, G. P. Plastic Trends. T44 Plastic Foams, 1977; Weiser, W. P. 771 US Plastic Foam Markets, Part I - Industry Study, 1983, Predicasts: Cleveland, OH. 6. Industry Study 1131: Foamed Plastics, The Freedonia Group: Cleveland, OH, 1999, Sections II, III, IV, VI, IX, and X. 7. ASTM D 883-80C, Definitions of Terms Relating to Plastics, American Society for Testing and Materials: Philadelphia, PA, 1982. 8. Griffin, J. D.; Skochdopole, R. E. In Engineering Design for Plastics, Baer, E.; ed., Reinhold: New York, 1964. 9. Cooper, A. Plast. Ind. London Trans. 1958, 26, 299. 10. ASTM D 1566–82, Definitions of Terms Relating to Rubber, American Society for Testing and Materials: Philadelphia, PA, 1982, Vol. 37. 11. ASTM D 1056-78, Specification for Cellular Materials - Sponge or Expanded Rubber, American Society for Testing and Materials: Philadelphia, PA, 1982. 12. ASTM D 1055–80. Specification for Cellular Materials - Latex Foam, American Society for Testing and Materials: Philadelphia, PA, 1982. 13. Hansen, R. H. SPE J. 1962, 18, 77. 14. Suh, K. W.; In Handbook of Polymeric Foams and Foam Technology, Klempner, D.; Frisch, K. C.; Eds, Hanser: New York, 1991, Chapter 8, pp. 151–186. 15. Throne, J. L. Thermoplastic Foams, Sherwood Publishers: Hinckley, OH, 1996, Chapters 1, 2, 3, 4, 5, 6, and 9. 16. Frisch, K. C.; Saunders, J. H. Plastic Foams, Vol. 1, Part 1, Marcel Dekker: New York, 1972. 17. Hilyard, N. C.; et al. Mechanics of Cellular Plastics, Macmillan: New York, 1982. 18. Han, C. D. Polym. Eng. Sci. 1978, 18, 699. 19. Han, C. D. Polym. Eng. Sci. 1981, 21, 518. 20. Han, C. D. J. Appl. Polym. Sci. 1976, 20, 1583. 21. Oyanagi, Y.; White, J. L. J. Appl. Polym. Sci. 1979, 23, 1013. 22. Frisch, K. C.; Saunders, J. H. Plastic Foams, Vol. 1, Part 2, Marcel Dekker: New York, 1973. 23. Benning, C. J. Plastic Foams, Vol. I, Wiley-Interscience: New York, 1969; Vol. 2, Wiley: New York, 1980, p 82. 24. Throne, J. L. In Engineering Guide to Structural Foams, Wendle, B. C.; Ed., Technomic Publishing: Westport, CT, 1976; Chapter VI, p. 111. 25. Skochdopole, R. E. Chem. Eng. Prog. 1961, 57 (10), 55. 26. Harding, R. H. Mod. Plast. 1960, 37 (10), 156. 27. Harding, R. H. J. Cell. Plast. 1965, 1 (3), 385. 28. Harding, R. H. Resinography of Cellular Materials, ASTM Technical Publication 414, American Society for Testing and Materials: Philadelphia, PA, 1967. 29. Corruccini, R. J. Chem. Eng. Prog. 1957, 53, 397. 30. McClintock, R. M. Adv. Cryog. Eng. 1960, 4, 132. 31. Brown, W. B. Plast. Prog. 1960, 1959, 149. 32. Hart, G. M.; Balazs, C. F.; Clipper, R. B. J. Cell. Plast. 1973, 9 (3), 139.

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

229

Gorring, R. L.; Churchill, S. W. Chem. Eng. Prog. 1961, 57 (7), 53. Stephenson, Jr.; M. E.; Mark, M. ASHRAEJ. 1961, 3(2), 75. Toohy, R. P. Chem. Eng. Prog. 1961, 57 (10), 60. Doherty, D. J.; Kurd, R.; L-ter, G. R. Chem. Ind, London, 1962, 1, 40. Guenther, F. O. SPE Trans. 1962, 2, 243. Knox, R. E. ASHRAE J. 1962, 4 (10), 3. Harding, R. H. Ind. Eng. Chem., Process. Des. Dev. 1964, 3, 117. Norton, F. I. J. Cell. Plast. 1967, 3 (1), 23. Lander, R. M. Refrig. Eng. 1957, 65 (4), 57. Larkin, B. K.; Churchill, S. W. AIChE J. 1959, 5, 467. Lao, B. Y.; Skochdopole, R. E. Paper presented at the Fourth SPI International Cellular Plastics Conference, Montreal, Canada, SPI: New York, 1976. Patten, G. A.; Skochdopole, R. E. Mod. Plast. 1962, 39 (11), 149. Hilado, C. I. J. Cell. Plast. 1967, 3 (4), 161. Dixon, R. R.; Edleman, L. E.; McLain, D. K. J. Cell. Plast. 1970, 6 (1), 44. Ball, G. W.; Hurd, R.; Walker, M. G. J. Cell. Plast. 1970, 6 (2), 66. Hilado, C. I. J. Cell. Plast. 1967, 3 (11), 502. Booth, J. R.; Grimes. J. T.; J. Therm. Insul. Build. Envelopes, 1993, 16 (April), 356. Booth, J. R.; In Conference Proceedings, Cellular Polymers, Int. Conf., 2nd, 1993, Paper 26. Graves, R. S.; Yarbrough, D. W.; McElroy, D. L.; Fine, H. A.; ASTM Spec. Tech. Publ., 1991, No. 1116 (Insul. Mater.: Test. Appl., 2nd vol.), 572. Booth, J. R.; J. Thermal Insul. Build. Envelopes, 1993, 17 (October), 154. Bomberg, M. T.; Kumaran, M. K.; Cell. Polym., 1995, 14, 343. Vahl, L. Dow Low Temperature Systems, Bulletin 179-2086-77, The Dow Chemical Company, Midland, MI, 1977, p. 267. Wheeler, C. H. Foamed Plastics, Presented at Conference, US Army Natick Labs and Committee on Foamed Plastics, US Department of Commerce Office of Technical Services, PB Report 181576, April 1963, p. 164. Patten, G. A. Mater. Des. Eng. 1962, 55 (5), 117. Dechow, F. I.; Epstein, K. A. ASTM STP660, Thermal Transmission Measurements of Insulation, American Society for Testing and Materials: Philadelphia, PA, 1978, p. 234. Levy, M. M. J. Cell. Plast. 1966, 2 (1), 37. Paljak, I. Mater. Construct. Paris 1973, 6, 31. Mittasch, H. Plaste Kautsch. 1969, 16 (4), 268. Achtziger, I. Kunststojfe 1971, 23, 3. Kaplan, C. W. CRREL Internal Report No. 207, US Army Cold Regions Research and Engineering Laboratory: Hanover, NH, 1969. Searle, N. Z.; Hirt, R. C. SPE Trans. 1962, 2, 32. Cooper, A. Plast. Inst. Trans. 1958, 26, 299. Cooper, A. Plastics (London) 1964, 29 (321), 62. Mod. Plast. 1962, 39 (8), 93. Ball, G. L.; II; Schwartz, M.; Long, I. S. Off Dig. Fed. Soc. Paint Technol. 1960,32,817. Cuddihy, E. F.; Moacanin, I. J. Cell. Plast. 1967, 3 (2), 73. Rogers, C. E. In: Engineering Design for Plastics, Baer, E.; Ed.; Reinhold: New York, 1964. Spenadel, L. Rubber World 1964, 105 (5), 69. Nakamura, M. US Patent 3960792 (1976) (to The Dow Chemical Company). Suh, K. W.; Kennedy, J. M. US Patent 4636527 (1987) (to the Dow Chemical Company).

230

K. W. SUH AND A. N. PAQUET

73. Suh, K. W.; Killingbeck, G. W. Can. Patent 1 086450 (1980) (to the Dow Chemical Company). 74. Suh, K. W.; Severson, J. L. US Patent 4916 166 (1990) (to the Dow Chemical Company). 75. Suh, K. W. US Patent 5011 866 (1991) (to the Dow Chemical Company). 76. Mclntire, O. R. US Patent 2515250 (1950) (to The Dow Chemical Company). 77. Munters, C. G., et al. US Patent 2023 204 (1935) (to C. G. Munters). 78. Phipps, A. L. US Patent 3 704083 (1972) (to A. L. Phipps); US Patent 4247276 (1981)(to Condec Corp.). 79. Plast. Technol. 1978, 24 (July), 8. 80. Gliniecki, V. L. Mod. Plast. 1954, 31 (7). 81. Mod. Plast. 1974, 51 (1), 36, 40. 82. Martens, T. P.; et al. Plast. Technol. 1966, 12 (9), 46. 83. Meyer, L. W. SPEJ. 1962, 18, 1341. 84. Martini-Vvedensby, J. E.; Suh, N. P.; Waldman, F. A. US Patent 4473665 (1984) (to MIT). 85. Cha, S. W.; Suh, N. P., Baldwin, D. F.; Park, C. B. US Patent 5 158986 (1992). 86. Colton, J. S.; Suh, N. P. US Patent 5 160674 (1992). 87. Shmidt, C. D.; Imeokparia, D. D.; Suh, K. W. US Patent 5 674 916 (1997) (to the Dow Chemical Company). 88. Suh, K. W.; Yamada, M.; Shmidt, C. D.; Imeokparia, D. D. US Patent 5679718 (1997). 89. Throne, J. L. J. Cell. Plast. 1976, 12 (5), 264. 90. Freund, R. W.; et al. Plast. Technol. 1973, 35 (November). 91. Angell, R. G., Jr.; US Patent 3436446 (1969) (to Union Carbide Corp.). 92. USM Foam Process, Technical Bulletin No. 653-A, Farrell Company Division: Ansonia, CT. 93. Botillier, P. Fr. Patent 1 498 620 (1967) (to Ugine-Kuhlmann). 94. Worthy, W. Chem. Eng. News 1978, December 11, 23. 95. Paraskevopoulos, S. C. A. J. Cell. Plast. 1965, 1 (1), 132. 96. Forming Thin Shells; Bulletin, The Dow Chemical Company: Midland, MI, 1962. 97. STYROFOAM- Sandwich Construction; Bulletin, The Dow Chemical Company: Midland, MI, 1960. 98. Modern Plastics Encyclopedia, McGraw-Hill: New York, 1964, Vol. 42 (IA), p. 294. 99. Ingram, A. R. J. Cell. Plast. 1965, 1 (1), 69. 100. TYRIL Foam 80; Technical Data Sheet No. 2-7, The Dow Chemical Company: Midland, MI, 1964. 101. Stastny, F. Baugewerbe 1957, 19 (April), 648. 102. Best, J. S. US Patent 3411 256 (1968) (to The Dow Chemical Company). 103. Bess, L. Y.; Ed., Insulation Guide for Buildings and Industrial Processes, Energy Technology Review No. 43, Part I, Noyes Data Corpotation: Park Ridge, NJ, 1979. 104. Hilado, C. J.; Murphy, R. W. ASTM Special Technical Publication (STP 685 Des. Build. Fire Safe), No. 16–105 ASTM, American Society for Testing and Materials: Philadelphia, PA, 1979. 105. FTC Consent Agreement, File 7323040, The Dow Chemical Company: Midland, MI, 1974. 106. Eickhorn, J.; Bates, S. I. US Patent 3058928 (1962) (to the Dow Chemical Company). 107. Suh, K. W. US Patent 4386 165 (1983) (to the Dow Chemical Company).

RIGID POLYSTYRENE FOAMS AND ALTERNATIVE BLOWING AGENTS 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118.

231

Molinar, M. J.; Rowland, F. S. Nature (London) 1974, 249, 810. Stolarski, R. S.; Cicerone, R. J. Can. J. Chem 1974, 52, 1610. Crutzen, P. Geophys. Res. Lett. 1974, 1, 205. Palmer, A. R.; et al. Economic Implications of Regulating Chlorofluorocarbon Emissions from Nonaerosol Applications, R-2524-EPA, Report prepared for the US EPA, Rand Corp.: Santa Monica, CA, June 1980. Elf Atochem North America, Inc., The Montreal Protocol: The First 10 Years, Internal Bulletin ADV 0082-97 6/97-5MPP (1997). Paquet, A. N.; Priddy, D. B.; Vo, C. V.; Pike, W. C; Hahnfeld, J. L. US Patent 5650 106 (1997) (to the Dow Chemical Company). Vo, C. V.; Paquet, A. N. US Patent 5389694 (1995) (to the Dow Chemical Company). Voelker, H.; Gerhard, A.; Schuch, H.; Weilbacher, M.; Weber, R. US Patent 5 182 308 (1993) (to BASF). Voelker, H.; Gerhard, A.; Weilbacher, M.; Weilbacher, F.; Weber, R.; Schuch, H. US Patent 5334337 (1994) (to BASF). Paquet, A. N.; Suh, K. W. US Patent 5210105 (1993) (to the Dow Chemical Company). Bartz, A. M.; Hitchcock, M. K. US Patent 5 373026 (1994) (to the Dow Chemical Company).

This page intentionally left blank

11 Polystyrene Packaging Applications: Foam Sheet and Oriented Sheet GARY C. WELSH The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Polystyrene resins are used in a variety of durable and packaging applications. Table 11.1 is a segmentation of the North American polystyrene resin market by injection molded versus extrusion processing, solid versus foamed structures and packaging versus durable applications. Approximately 58 % of the North American annual consumption of polystyrene resins is used for packaging, primarily food packaging. With the attributes of high modulus, high melt strength and food contact approval, polystyrene is well suited for a variety of food packaging applications. Two of the most important food packaging materials using polystyrene resins are oriented polystyrene sheet (OPS) and polystyrene foam sheet. Together, OPS and polystyrene foam sheet packaging account for nearly 40 % of the polystyrene resins used in food packaging applications in North America.

2

ORIENTED POLYSTYRENE SHEET

Where a clear, thermoformed product is desired, many fabricators use OPS. Because of the low elongation of general-purpose polystyrene, it is necessary to orient the sheet biaxially to obtain the required physical properties for satisfactory part trimming and end use. If there is insufficient elongation in the Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy r) 2003 John Wiley & Sons Ltd.

234

G. C. WELSH

Table 11.1 Major North American markets for polystyrene resins, 1998 [1] Market Injection molding: Durables Packaging Extrusion: Durables (solid) Durables (foamed) Packaging (solid) Packaging (solid, oriented) Packaging (foamed) Expandable bead (foamed): Durables Packaging

Applications

1000

1998 usage, mt

Appliances, electronics, furniture, toys, houseware Closures, tumblers, cutlery, dishes

679

Appliances, electronics, toys, houseware building Insulation board Dairy containers, cups, lids, plates Bakery trays, food service containers Trays, hinged containers, cups, plates, egg cartons

230 91 397 272 338

Insulation/construction Cups, shapes for cushioning, loose fill

147 230

325

sheet, undesirable brittle fracture will occur when the thermoformed part is trimmed from the web or in the end use handling of the container. Typical elongation values for unoriented cast sheet are less than 5 % and are typically in the region of 1.5%. Controlled biaxial orientation can produce OPS that has elongation values of up to 60 %, although elongation values for both machine direction (MD) and transverse direction (TD) are typically about 4% [2]. OPS is commercially available in 0.13–0.76 mm gauge. The tenter frame process for manufacturing OPS is the most common process used today. A typical tenter frame process is shown in Figure 11.1. Tentered OPS is distinguished from polystyrene blown films (which are also biaxially oriented) in that the tenter frame process can produce heavier gauge films than the blownfilm process. Blown films (0.025–0.13 mm) are used primarily in window envelopes, lamination, and printed applications; tentered OPS is typically used in thermoformed trays, lids, and containers for rigid food packaging applications. OPS is usually made of a high heat distortion, unplasticized, high molecular weight (Mw = 300000–320000) general-purpose polystyrene resin. For most food packaging applications, the resin is low in residual solvent and styrene monomer so that no detectable taste or odor is transferred to the food product. OPS used in food packaging applications requires that all resins and additives be approved for food contact. Web trim from the thermoforming process resin is added back to the process as either ground sheet or repelletized material. The OPS extrusion process typically uses a two-stage vented extruder. The most common size OPS lines use either a 200 or 220 mm diameter extruder. The lengthto-diameter (L/D) ratios range from about 24 to 32, the higher ratio extruder

235

FOAM SHEET AND ORIENTED SHEET PACKAGING Casting rolls

MDO unit, 2:1 MD stretch

TDO unit (plan view), 2:1 TD stretch Ovens

Winder

Figure 11.1 A typical OPS (oriented polystyrene sheet) tenter frame process with stretching in the machine (MD) and transverse (TD) directions. Rolls 1–8 are heated and driven. Sheet is heated above and below in the ovens. Sheet geometry as a function of position in process: Position Gauge (mm) MD (% of original) TD (% of original)

A 1.5 100 100

B 0.75 200 100

C 0.38 200 200

being capable of more uniform melts at higher rates. Although there are smaller lines in use (115mm), the cost of the ovens, machine, and transverse direction orientation units makes it more economical to purchase the larger lines. A 200 mm line is capable of 2300 kg/h and the larger 220 mm lines are capable of 3000 kg/h. A screen changer is located between the extruder and die to remove any dirt or foreign particles. Gear pumps located between the extruder and die reduce power consumption and provide a more uniform rate of polymer delivery to the die, providing improved gauge control [3]. In a 200mm OPS line, the die is usually 1.0–1.3m wide. These dies are generally of the 'coat hanger' manifold design with separate choker bar and

236

G. C. WELSH

die gap adjustments. The required die gap is a function of the degree of stretch of both machine direction orientation (MDO) unit and the transverse direction orientation (TDO) unit. The newer OPS lines are equipped with automatically adjustable die lips that use heated die bolts. This type of die is used in coordination with a sheet-gauging device with feedback control capability. From the die, the extrudate continues through a three-roll stack (casting rolls) that provides a uniform initial gauge across the web and brings the sheet to uniform temperature. From the bottom roll on the three-roll stack the web travels to the MDO unit. The MDO unit is a series of heated, driven rolls, each roll driven progressively faster to impart MDO to the sheet (Figure 11.2). Stretching must be performed at the proper sheet temperature (95–120 °C). If the sheet temperature is too low, the web will break or equipment power limitations will be reached. If the sheet temperature is too high, little orientation is imparted to the polymer and there will be no physical property improvement. The web, now moving at a line speed proportional to the stretch imparted by the MDO unit, enters the TDO unit. The TDO unit has a large enclosed oven that uniformly heats the sheet for the transverse stretch. The transverse stretch is accomplished by the use of film clips (Figure 11.3). These clips are attached to a chain-drive system on each side of the web. Each chain-clip assembly clamps the web upon entering the oven and stretches (tenters) the sheet because of the increasing width between the chains along the machine direction (Figure 11.4). On exit of the TDO unit, the web has now been oriented in both the machine

Figure 11.2 A cantilever style MDO (machine direction orienter). Courtesy of Marshall and Williams Plastics

FOAM SHEET AND ORIENTED SHEET PACKAGING

237

Figure 11.3 Tenter clips used on OPS tenter frames. Courtesy of Marshall and Williams Plastics

Figure 11.4

Commercial tenter frame. Courtesy of Marshall and Williams Plastics

and transverse directions and is rapidly cooled to retain this orientation [4]. Typically, the amount of stretch in each direction is equal to achieve balanced sheet properties. After machine and transverse direction orientation, the web continues to an edge-trimming station where the undesirable clamped edges of the web are trimmed. The web is often lightly coated with silicone oil at this point to assist in the future stripping of formed parts from thermoforming molds and denesting the final stacked thermoformed parts. After edge trimming and coating, the web is then slit to the desired width for thermoforming and wound into rolls. Slitting can also be performed off-line.

238

G. C. WELSH

OPS is thermoformed with vacuum alone or with plug-assisted vacuum with the additional option of air pressure. During the forming of OPS, it is important that the tooling be designed to clamp the sheet either around the perimeter of the entire mold or around each individual cavity. This preserves the biaxial orientation in the flange region, which must be maintained to permit clean trimming without brittle fractures. Trimming is typically accomplished with trim-in-place tooling in the forming mold. Trimming can also be performed in a separate trim operation; however, the part shrinkage that occurs between the forming station and trimming station must be allowed for in the trimming tool to achieve accurate indexing and cutting. The OPS extrusion process is distinguished from conventional cast extrusion (unoriented) processes by the additional energy requirements of the MDO and TDO units. Cast extrusion processes generally require 0.2–0.7 kWh per kilogram of extruded product. The OPS process requires 0.7–0.9 kW h per kilogram of extruded product, which increases conversion cost. OPS is generally tested at regular intervals during extrusion. The finished sheet is tested for orientation level using a stress-release test. A force gauge is used to measure the stress level as the sheet specimen is heated in either air or oil. This test is performed in both the machine and transverse directions. If stress levels are too low, fracturing may occur in the trimming operation, or the practical toughness of the thermoformed parts may be low. OPS is a particularly desirable packaging material because of its high flexural modulus, excellent clarity, low taste and odor transfer and relatively low cost. A summary of OPS properties is shown in Table 11.2 [5]. Owing to the combination of high modulus, low specific gravity, high clarity and ease of processing, OPS has demonstrated significant cost advantages versus other plastic packaging materials such as poly(vinyl chloride) (PVC), polypropylene (PP) and amorphous polyester (APET) [6]. The OPS market has realized significant growth since its introduction in the late 1950s. In 1998, approximately 35 OPS lines were in operation producing 272 000 metric tons per year [7]. Most of this volume is used in bakery good trays and food service containers, as shown in Figure 11.5. Table 11.2

Typical properties of biaxially oriented polystyrene sheet (OPS)

Property

Value

Tensile strength (MPa) Tensile modulus (MPa) Tensile elongation (%) Light transmission (%) Specific gravity Material yield (m 2 /N at 0.25 mm)

62-83 3100 5–60 92 1.05 3.8°

a

26 200 in 2 /lb at 1 mil.

FOAM SHEET AND ORIENTED SHEET PACKAGING

Figure 11.5

239

Various OPS food packaging applications

3 EXTRUDED POLYSTYRENE FOAM SHEET Owing to the high modulus, high melt strength and amorphous nature of polystyrene resins, they are well suited for the manufacture of foam sheet structures. The primary advantages of polystyrene foam sheet in disposable packaging are reduced weight, high flexural modulus, low water absorption, ease of thermoforming and low cost. Extruded polystyrene foam sheet is a closed cell, 0.13–6.4 mm thick (gauge) sheet with densities ranging from 32 to 160kg/m3 [8]. Table 11.3 gives a summary of polystyrene foam sheet properties. As with OPS, polystyrene foam sheet is a highly oriented structure with significantly higher tensile elongation (5–20 %) versus solid, unoriented, general-purpose polystyrene resin (1.5%). Table 11.3

Typical properties of polystyrene foam sheet

Property Gauge (mm) Density (g/cm3) Cell size: Machine direction (mm) Transverse direction (mm) Vertical direction (mm) Tensile strength, (machine direction) (MPa) Tensile modulus, (machine direction) (MPa) Tensile elongation, (machine direction) (%)

Value 3.8 0.053 0.23 0.33 0.22 1.5 48 15

240

G. C. WELSH

Polystyrene foam sheet can be made by a variety of extrusion processes (single screw, twin screw) but is most commonly made using a tandem extrusion process as shown in Figure 11.6. This process uses two extruders in series, which effectively optimize the melting, mixing, and cooling unit operations. The diameters of the extruders currently used in this process are usually 115 mm for the primary extruder and 150mm for the secondary extruder, with an output capability of 400 kg/h. Raw materials are introduced in the primary extruder where they are then forwarded and subsequently melted. A blowing agent is accurately metered with a high-pressure piston/metal diaphragm pump into the metering section of the primary screw. The blowing agent is thoroughly mixed into the polymer melt, which is then forwarded to the secondary extruder. A screen changer is usually placed in the connecting pipe between the primary and secondary extruder to remove foreign particles that might otherwise become lodged in the die. A cooling profile is selected on the secondary extruder that will optimize line rate but still achieve the necessary temperature and viscosity of the melt for foaming. The screw speed of the secondary extruder is maintained at a low value (10–30 rpm) to minimize shear heating of the polymer. An annular die is used on the end of the secondary extruder. Die diameters range from 76 to 254mm and die gaps range from 0.25 to 0.76mm. Nucleation of the cells occurs near the exit of the die because of the pressure drop in the die lips. Controlled foaming continues outside the die lips as the material is stretched over a forming mandrel (Figure 11.7). The forming mandrel diameter

© Figure 11.6 A tandem extrusion process for the manufacture of polystyrene foam sheet. (1) Primary extruder; (2) blowing agent addition system; (3) screen changer; (4) secondary extruder; (5) annular die; (6) cooling mandrel; (7) S-wrap; (8) winders

FOAM SHEET AND ORIENTED SHEET PACKAGING

241

Figure 11.7 Tandem foam sheet extrusion lines in operation. Courtesy of Battenfeld Gloucester Engineering Co., Inc

is sized in relationship to the die diameter. The ratio of the mandrel diameter to the die diameter is defined as the blow-up ratio and is typically in the range 2.5:1–5:5.1. The forming mandrel determines the overall circumference of the tube and imparts biaxial orientation into the final sheet [9,10]. The tube is then double slit at the end of the forming mandrel to form an upper and lower web. The web continues to be pulled by the S-wrap and is finally wound into rolls at the winding stations. Typical extrusion conditions are given in Table 11.4. Rolls of polystyrene foam sheet are then placed in storage and allowed to age. During aging, blowing agent diffuses out of the cells much slower than air diffuses in resulting in increased cell pressure. After 2-5 days, the proper concentration of blowing agent and air exists in the cells to promote the optimum postexpansion of the sheet during re-heating prior to thermoforming. Longer aging times (> 4 weeks) may result in brittle sheet since the highly oriented cell structure experiences stress relaxation over time [11]. As the properly aged sheet is heated, it expands because of the increased cell gas pressure. This post-expansion of up to 100% is predictable and is used as a basis for the sheet starting-gauge specification and the clearance of the matched molds in the forming station [12]. Temperature history during a typical thermoforming cycle is shown in Figure 11.8. After the part has been formed, the web moves to a trim press where the individual parts are trimmed from the web. The web skeleton, which may constitute 10–50% of the total web, continues under the trim press to a grinder for reprocessing.

242

G. C. WELSH

Table 11.4 Polystyrene foam sheet tandem extrusion conditions using n-butane Condition

Value

n-Butane(wt%) Talc(wt%) Regrind (wt%) Primary extrudera pressure (MPa) Primary extrudera melt temperature (°C) Die pressure* (MPa) Die melt temperatureb( °C) Die gap (mm) Output (kg/h)

4.0 0.8 30 26 220 9.5 135 0.70 380

115 mm single screw at 90 rpm. From secondary extruder (150 mm single screw at 13 rpm).

Cell Collapse

Time (s)

Figure 11.8 Polystyrene foam sheet surface and center temperature versus thermoforming cycle time (3s). Sheet index indicates sheet movement from oven to forming station [13] The cost of manufacturing thermoformed, polystyrene foam sheet parts is less dependent on raw material cost than other extrusion processes. This is largely due to the combined effects of additional energy costs required to operate two extruders, heat removal requirements in the secondary extruder, cost of pelletizing (densifying) regrind and the relatively low output of the process for the equipment scale and cost. Typical cost factors for the manufacture of thermoformed polystyrene foam sheet products include raw materials 35%, labor 27%, sales and administration 16%, depreciation 8%, utilities 7% and other 7 %.

FOAM SHEET AND ORIENTED SHEET PACKAGING

243

The raw materials used in the foam sheet extrusion process are polystyrene resin, blowing agent, nucleating agent, color concentrate (optional) and regrind. A high molecular weight (Mw = 300 000-320 000), high heat distortion (unplasticized) polystyrene resin has the necessary melt strength to achieve lowdensity sheet with good physical properties. Blowing agents are typically selected for polystyrene foam sheet based on their combination of vapor pressure, solubility, permeability, handling and environmental safety and cost [14]. Owing to concerns over atmospheric ozone depletion in the 1980s, chlorofluorocarbon and hydrochlorofluorocarbon blowing agents are no longer used to manufacture polystyrene foam sheet in North America [15,16]. Aliphatic hydrocarbons, carbon dioxide or blends of the two have become the most common blowing agents [11]. Figure 11.9 shows the effect of blowing-agent concentration on foam density. A nucleating agent is normally used to initiate cell growth and control cell size [17]. For higher vapor pressure and lower solubility blowing agents such as isobutane, a fine particle size (1.5 |xm) talc is used. For more soluble blowing agents with lower vapor pressure such as n-pentane, a blend of citric acid and sodium bicarbonate is preferred. Color concentrates are used in many applications to add esthetic value. Regrind is a natural consequence of the thermoforming/trimming operation and must be re-used because of economic considerations. Because the majority of foam sheet produced is thermoformed into containers for food packaging, all raw materials must meet applicable governmental 0.40

0.35 • 0.30

0.20

0.100.050.00 0

Figure 11.9

2

3 4 Weight %

Polystyrene foam sheet density versus blowing agent concentration

244

G. C. WELSH

regulations for food use. This includes the polystyrene resin, blowing agent, nucleating agent, color concentrates and any other additives. Quality control testing is performed on both the extruded sheet and the final thermoformed part. The more common foam sheet quality control tests include gauge, basis weight (i.e. weight/area), cell size, post-expansion and orientation. In the case of the post-expansion and orientation tests, sheet samples are placed in an oven at 116°C for 20min to simulate the heat history experienced in thermoforming. After heating, the sheet dimensions are noted in the machine, transverse and thickness directions. The shrinkage in the machine and transverse directions is recorded as a measure of orientation; the increase in thickness is recorded as the post-expansion and is related to the performance of the sheet when re-heated in the thermoformer. The test results obtained on the extruded sheet are particularly important in that these tests predict the suitability of the sheet for thermoforming. Because the roll stock is aged for 2-5 days before thermoforming, any undetected sheet problem could result in many unusable rolls being manufactured before the problem was discovered. Thus, 2-5 days of production could require regrinding and recycling. Figure 11.10 shows typical cell structure in a cross-section of polystyrene foam sheet. A significant safety consideration when making polystyrene foam sheet is the use of flammable blowing agents, specifically aliphatic hydrocarbons. The use of flammable blowing agents requires proper area electrical classification [18], area ventilation (adequate air changes per hour), equipment grounding, static elimination and installation of a high-level blowing agent detection system.

Figure 11.10

Cross-section of polystyrene foam sheet

FOAM SHEET AND ORIENTED SHEET PACKAGING

245

Figure 11.11 Various polystyrene foam sheet food packaging applications The use of polystyrene foam sheet has grown significantly since its introduction in the mid-1960s. In 1998, approximately 338 000 metric tons of polystyrene foam sheet were produced (Table 11.1). Most of the current use is in disposable thermoformed food packaging such as egg cartons, meat, poultry and produce trays, fast food containers, dinnerware (bowls and plates) and bottle labels (Figure 11.11).

REFERENCES 1. Mod. Plast., January, 73, (1999). 2. Sweeting, O. J., The Science and Technology of Polymer Films, Wiley-Interscience, New York, 1971, pp. 290–291. 3. Woodworth, C. L., Antec SPE Conference, Brookfield, CT, 1984, pp. 122–126. 4. Simmons, R. S., Polymer films: process, structure and performance, presented at Polymer Technology Symposia, Industrial Materials Institute, National Research Council, Boucherville, Quebec, November 6–7, 2000. 5. Benning, C. J. Plastic Films for Packaging, Technomic, Lancaster, PA, 1983, p. 34. 6. Whaley, J. F., Equipment and applications for OPS packaging, presented at Specialty Plastic Films 1998, 14th Annual World Congress, October 19–21, 1998. 7. Leaversuch, R. D., Mod. Plast., August, 64 (1999). 8. Suh, K. W., in Polymeric Foams, Klempner, D. and Frisch, K. C. (eds), Hanser, Munich, 1991, p. 157. 9. Thomas, L. S., and Cleerman, K. J., SPE J., 28, 61 (1972). 10. Collins, F. H., and Kraus, D. A., Technical Papers, Thirty-first Annual Technical Conference, Society of Plastics Engineering, Montreal, May 7-10, 1973, Vol. XIX, pp. 639-642.

246

G. C. WELSH

11. Welsh, G. C., US Patent 5250577 (1993) (to The Dow Chemical Company). 12. Throne, J. L., Technical Papers, Forty-third Annual Technical Conference, Society of Plastics Engineers, Washington, DC, April 29-May 2, 1985, Vol. XXXI, pp. 1328– 1333. 13. Burt, J., Course Notes, Plastic Institute of America, June 15, 1983. 14. Benning, C. Plastic Foams, Vol. 1, John Wiley & Sons, Inc., New York, 1969, pp. 90–100. 15. Montreal Protocol, Handbook: Ozone Secretariat of the United Nations Environment Programme, 1987. 16. Environmental Protection Agency, 40 CFR Part 82, Fed. Regist., 58 (10), January 15(1993). 17. Lee, S. T., in Foam Extrusion, Lee, S. T. (edit.), Technomic, Lancaster, PA, 2000, pp. 81-124. 18. National Electric Code, Article 500, National Fire Protection Association, Battery March Park, Quincy, MA, 1987.

12

Preparation, Properties and Applications of High-impact Polystyrene M. F. MARTIN, J. P. VIOLA AND J. R. WUENSCH BASF AG, Ludwigshafen, Germany

1

INTRODUCTION

Polystyrene is one of the oldest commercially produced thermoplastic polymers. The homopolymer, frequently referred to as GPPS, is a highly versatile product that has found acceptance in widely diverse applications. The strengths of this brilliant, clear, noncrystalline plastic are processing ease, rigidity, dimensional stability and clarity. However, the low impact strength of polystyrene limits its use. High-impact polystyrenes, frequently referred to as HIPS, are an elastomer modified polystyrene thermoplastic. This two-phase system, consisting of a rubber phase and a continuous polystyrene phase, provides a polymer system that has grown into an important world-scale commodity polymer meeting the needs of thousands of applications. Engineering specialties, which enhance one or more performance characteristics, have further pushed the applications envelope. This versatile product can be found in many compositions that offer an exceptional range of impact properties and processability for applications in the automotive, appliance, power tool, furniture, housewares, telecommunication, electronic, computer, disposable, medical, packaging and recreation markets. The multiple applications which are supported by GPPS and HIPS have fueled global consumption of approximately 14000000 metric tons at the turn of this century.

Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers, Edited by J. Scheirs and D. B. Priddy © 2003 John Wiley & Sons Ltd

248

2

M. F. MARTIN ETAL.

PROPERTIES

Polystyrene homopolymer, commonly referred to as general-purpose polystyrene (GPPS), is chosen for its excellent clarity, rigidity and dimensional stability. While HIPS products are also valued for their excellent dimensional stability and rigidity, they provide good impact strength and high rigidity not exhibited by GPPS. Relative disadvantages of HIPS are poor hightemperature properties, poor oxygen barrier properties, relatively low ultraviolet light stability and lower chemical resistance, compared with crystalline materials.

2.1

GENERAL PROPERTIES

Polystyrene is a brittle thermoplastic material. The addition of rubber increases impact strength considerably. The commercial success of HIPS is largely due to the ease of developing polymer systems that meet application needs of toughness, rigidity, heat distortion resistance and flow behavior. Typically, the enhancement of a specific property requires tradeoffs in other attributes. The key is to develop a material with an acceptable balance of performance for a given application. In modern HIPS polymerization processes, this is accomplished with raw material selection, formulation and operating conditions. Additional melt compounding furthers the possibilities. Specialty grades have been developed to offer enhancements, such as rapid fabrication and improved environmental stress crack resistance. Specialty polymer systems are also easily arrived at by post-reactor modification of the HIPS with specialty additives that provide additional properties, such as ignition resistance, scratch resistance and enhanced chemical resistance.

2.2

MECHANICAL PROPERTIES (TABLE 12.1)

The tensile strength of HIPS increases with decreasing temperature and increases with increasing rate of strain, as would be expected of a rubber toughened system. Figure 12.1 presents a typical stress-strain relationship for an extrusion-grade HIPS. HIPS displays a less steep initial rise in the curve and, after reaching the yield point, pronounced deformation until rupture. This behavior is characteristic of tough materials. The stress-strain behavior of HIPS depends on the temperature and on the deformation rate. As can be seen from Figure 12.2, the elongation at break of high-impact polystyrene decreases as the temperature drops and as the deformation rate increases, while the tensile strength displays the opposite relationship. The temperature effect is significantly greater than the influence of the deformation rate.

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

249

Table 12.1 Tensile properties of several HIPS products Nominal strain Tensile modulus Stress @ yield Strain @ yield @ break ISO527-l,-2 ISO527-l,-2 ISO527-1,-2 ISO527-1, (MPa)(N/mm 2 ) (MPa) (N/mm 2 ) (%) -2 (%)

Product Super high-impact extrusion grade Refrigeration liner extrusion grade General extrusion grade General injection molding grade High flow molding grade

1900

32

1.9

25

1500

21

1.7

35

1850

26

1.5

35

2800

44

2

10

2200

26

1.4

40

o.

S

H

10

10

15

20

25

35

40

45 50 Strain (96)

Figure 12.1 Tensile stress-strain relationships for HIPS at various strain rates HIPS behaves like a viscoelastic solid. It undergoes creep when exposed to a constant tensile stress. Figure 12.3 illustrates typical tensile creep properties of a HIPS product. Creep decreases with increasing molecular weight and generally increases with increasing rubber content. The creep behavior of HIPS depends strongly on the average of the molecular weight (MW). Resins with a higher molecular weight will entangle much tighter under tear stress, which results in an increase in elongation of break under stress propagation. The creep behavior is not influenced until tear propagation starts. A visible line of damage appears under stress, indicating an increase in polymer entanglement (craze lines). At that point, the macromolecules experience a high stretch ratio. Under increased stress, the polymer

M. F. MARTIN ETAL

250

Test temperatpre 50

-40' C -20° ,±0

B

c

40 20°C

5 30

.+4PC

20 + 60°C

10

20

0

V8C

10 20

30 40

50 60

70 80

90 100

Strain (%) Figure 12.2 Stress—strain curves of HIPS at various test temperatures. Strain rate = 2 mm/min, determined on injection-molded test specimens

carbon—carbon backbone breaks and tear propagation continues. In principle, the higher the stress, the earlier is craze formation. 2.3

IMPACT PROPERTIES (TABLE 12.2)

In contrast to the relatively low deformation rate in the tensile test and the creep test, the high deformation rate in the impact test leads to a significantly higher dynamic load on the test specimens. The rubber modification imparts substantial toughness by stabilizing crazing. To achieve a high toughness, it is necessary to grow a large volume fraction of crazes throughout the cross section of the article, without permitting crazes to progress into cracks. Most HIPS fail in the uniaxial loading of a cantilever impact (Izod) test only if the test specimens have been notched beforehand. In the notched impact test, the stress direction in injection molded specimens is the same as the preferred orientation direction. This increases the measured impact toughness and the flexural impact test and, therefore, serves primarily for comparing the toughness of different products. In practice, however, the multiaxial stress applied in a falling dart test without a preferred direction, has to be expected. This is addressed by the Gardner test of ASTM, in which a falling dart strikes centrally on a flat surface (e.g. a circular disk). The results of both test methods are important polymer design considerations.

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

2289.0 2150.0 2011.0 "^ 1872.0

1 2 3 4

1

L \v\

251

Molding Grade 2.5 MPa Molding Grade 7.5 MPa Extrusion Grade 2.5 MPa Extrusion Grade 7.5 MPa

| 1733.0 £ 1594.0 ^ 1455.03 1316.0 ~o ^ 1117.0

\V \w

^-^

^1

^•^^

^-^ 2

""•••^

""":::::^-

1A-JO A

*'->..

L ^3

. J~***~~~•

899.0 760.0

1

. — ~-~.

•4_ ^~*-^, ^ — ^^ •^-^. . «*B^.

621.0 10 .0

1.51e4

3.03e4

4.54e4

(a)

6.05e4

7.56e4

9.08e4

1 06e5

Time (%)

1359.4 1 2 3 4

1264.7 1170.0

Molding Grade 2 5 MPa Molding Grade 7 S MPa Extrusion Grade 2.5 MPa Extrusion Grade 7.5 MPa

1075.3980.6885.9791.2696.5 601.8

\

\

507 1 • 412 4-

V V

/

(b)

Figure 12.3

1

/

,

1.51e4

/4

3

/ === ===--

,/

^=

317 7.

9930 n. 10.0

/ \2

3.03e4

1

1

, t/ 1

4.54e4

1

6.05e4

7.56e4

9.08e4

Time (%)

Creep modulus at (a) 23 and (b) 40 °C (ISO 11403-2)

1.06e5

M. R MARTIN ETAL.

252

Table 12.2

Impact properties of several HIPS products

Product

Charpy impact strength ISO 179 (23°C)(kJ/m 2 )

Charpy impact strength ISO 179 (-30°C)(kJ/m 2 )

Charpy notched impact strength ISO 179 (23 °C) (kJ/m 2 )

Super high-impact extrusion grade Refrigeration liner extrusion grade General extrusion grade General injection molding grade High flow molding grade

160 No Break No Break 70 120

80 120 130 50 80

2 10 12 6 10

2.4

THERMAL PROPERTIES (TABLE 12.3)

The heat distortion resistance of finished HIPS parts is dependent on their shape, the production conditions, the type of heat source and the duration of heating, and also on the HIPS grade in question. Parts produced without application of an external load and having low internal stresses can be heated for a short time to about 15 °C below the Vicat softening temperature without undergoing distortion.

2.5

ELECTRICAL PROPERTIES (TABLE 12.4)

HIPS is a nonpolar polymer and is a very good electrical insulator. The dependence of the dielectric properties upon frequency and temperature is virtually nonexistent, as can be seen in Figure 12.4. Table 12.3

Thermal properties of various HIPS grades

Product

Deflection temperature (0.45 MPa) ISO75-1, -2 (°C)

Vicat softening temperature ISO306(B50, 50c C/h50N) (0C)

CLTE parallel to flow IS011359–1, -2 (cm/cm/cC)

Super high-impact extrusion grade Refrigeration liner extrusion grade General extrusion grade General injection molding grade High flow molding grade

87 87 83 96 82

89.0 89.0 90.0 96.0 84.0

0.0001 0.0001 0.0001 0.0001 0.0001

253

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE Table 12.4

Electrical properties of various HIPS grades

Product

Relative permittivity IEC250 (102/Hz)

Dissipation (loss) factor IEC250 (106/Hz)

Volume resistivity IEC93 (ohm cm)

Super high-impact extrusion grade Refrigeration liner extrusion grade General extrusion grade General injection molding grade High flow molding grade

2.5 2.5 2.5 2.5 2.5

0.0004 0.004 0.0004 0.004 0.004

1 1 1 1 1

x x x x x

1015 1015 1015 1015 1015

Surface resistivity IEC93 (ohm)

1 1 1 1 1

x x x x x

1013 1013 1013 1013 1013

Moist

2.5

Dry

2

10-2 10-3 lO-4 10-4

4

5 logf

Figure 12.4 Dependence of dielectric properties on frequency and temperature (solid line, moist; dashed line, dry)

2.6

RHEOLOGICAL PROPERTIES (TABLE 12.5)

Like all thermoplastics, HIPS is a non-Newtonian fluid. This means that the viscosity depends not only on the temperature, but also on the shear rate. As a result, the melt flow rate increases proportionally with increasing pressure. In Figures 12.5-12.7, the temperature is also shown as a parameter. A stress cracking test on HIPS is illustrated in Figure 12.8. 2.7

RESISTANCE TO SOLVENTS

HIPS resists damage to properties from exposure to water, alkalis and dilute mineral acids. It is swollen by some organic solvents and dissolved by others in a relationship governed by the difference between the solubility parameter of the continuous phase and that of the solvent. HIPS is particularly susceptible to damage when exposed to chlorinated and aromatic hydrocarbons.

M. F. MARTIN ETAL.

254

Table 12.5

Single point melt rheology of various HIPS products

Product

Melt volume flow rate ISO1133 (200 °C/5.0 kg) (cm3/10 min)

Super high-impact extrusion grade Refrigeration liner extrusion grade General extrusion grade General injection molding grade High flow molding grade

4 4 5 8 13

1.0e4

1^

"^

1 180 °C 2 230 °C 3 280 °C

\.

a O,

2-

•1" o u

100

>

10.0

\\

-°

3^-",

^ ^^ "^

1.0

^ "^

ni 1.0

10.0

100.0

1000.0 Shear Rate (s-1)

1.0e4

1.0e5

1.0e6

Figure 12.5 Rheological data (ISO 11403–2) for a super high-impact extrusion resin that has a volumetric flow rate (ISO1133, 200°C/5.0kg) of 4

1 190 °C 2 220 °C

3 250 °C

1.0e4

1000.0

3"

100.0

10.0 0.1

10.0

100.0 Shear Rate (s-')

1000.0

1.0e4

1.0e5

Figure 12.6 Rheological data (ISO 11403–2) for a high-impact extrusion resin that has a volumetric flow rate (ISO1133, 200°C/5.0kg) of 4

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

255

1.0e4

1 200 °C 2 240 °C 3 280 °C

1000.0

100.0

10.0:-

1.0 1.0

10.0

100.0 1000.0 Shear Rate (s-1)

1.0e4

1.0e5

Figure 12.7 Rheological data (ISO 11403–2) for a high-impact injection molding resin that has a volumetric flow rate (ISO1133, 200°C/5.0kg) of 13

A \

X X

\ o.i

1.0

10

100

1 day

1 week

1000

1 month Time (h)

Figure 12.8 Stress cracking test on HIPS at 20 °C in various media. The figure shows creep rupture curves for (1) n-heptane, (2) olive oil–oleic acid (1:1), (3) methanol, (4) battery acid, (5) Nekanil (nonionic surfactant) W-Extra solution, (6) distilled water and (7) air

256

3

M. F. MARTIN ETAL

BASIC CHEMISTRY

HIPS is a two-phase polymer system, the properties of which depend upon a complex relationship between • • • • • • •

the continuous phase molecular weight and molecular weight distribution; the type of elastomer employed; the elastomer phase volume ratio; the elastomer particle size and particle size distribution; the elastomer particle structure; the graft polymer crosslink density; and additive concentrations.

3.1

MATRIX MOLECULAR WEIGHT

The polystyrene molecular weight affects mechanical, Theological and thermal properties. Relationships between molecular weight of the polystyrene phase and polymer properties are similar to those for the homopolymer. Melt viscosity and the relationship between polymer architecture and processability are controlled predominantly by molecular weight. Increasing the molecular weight of the continuous polystyrene phase increases the melt viscosity. Characteristics of the graft copolymer, the molecular weight and the molecular weight distribution are optimized during polymer design to provide the desired rheological properties, with special attention given to melt strength and elastic melt flow at the shear rates expected in the targeted application. The relationship between the structure and properties of HIPS is illustrated in Figure 12.9.

3.2

ELASTOMER

CONSIDERATIONS

GPPS is impact modified by the inclusion of an elastomer to yield a product that may exhibit impact properties an order of magnitude greater than the homopolymer. Commercial elastomers used have a molecular weight of 180000–260000 and are long-chain branched to suppress cold flow. The literature teaches many elastomer systems but the most commonly employed rubber is medium-cis polybutadiene and high-cis polybutadiene. High-cis polybutadiene has a relatively high heat resistance, which provides some advantages to a HIPS formulation at the cost of low-temperature toughness. While formulations with rubber compounded into the GPPS are effective, grafting the elastomer into the continuous phase is preferred. Commercial polymerization processes produce a polymer system that not only has an elastomer incorporated, but also a grafted species where short polystyrene side chains have been attached to the rubber domains. This grafting anchors

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE Molecular Weight

Oil Content

Gel Content

Swell Index

257

Particle Size

Stiffness

Toughness

Heat Distortion Temperature Gloss

Melt Flow Rate

Figure 12.9

Relationship between structure and properties of HIPS [8]

the discrete elastomer particles in the continuous polystyrene phase, yielding better compatibility and much higher impact strength. When external forces act on HIPS, the rubber particles have a stress-relaxing action. To do this job, they must be bound sufficiently to the matrix and have a certain elasticity. This means that the rubber used must be capable of grafting and crosslinking. Both properties can be controlled within certain limits via the microstructure. These requirements mean that nowadays use is predominantly made of medium-cis and sometimes also high-cis rubbers, which can be produced using organolithium initiators or coordinate Ziegler catalysts. High-cis polybutadiene has relatively high heat resistance, which is advantageous in the processing of HIPS. On the other hand, this type of polybutadiene crystallizes at about 0 °C, owing to its stereoregular structure, with the consequence that the low-temperature toughness of polystyrene, produced in this way, is reduced. Apart from pure polybutadiene rubbers, styrene—butadiene block copolymers are also used, enabling products having particle sizes of less than 1 (xm to be produced. They have high gloss and high rigidity, but somewhat lower toughness for a comparable polybutadiene content. Other rubber systems have been commercially successful. Styrene block copolymers yield a HIPS product with a small particle size and provide high gloss. A mixed rubber system consisting of styrene—butadiene block rubber and/or ethylene—propylene diene modified (EPDM) rubber can be blended with the polybutadiene to form bimodal rubber particle size distribution for a

258

M. F. MARTIN ETAL.

combination of high gloss and high impact. The block rubber generates the typical capsular small particles, and the EPDM, which is incompatible with the block rubber, yields large particles as a separate rubber phase. EPDM, with higher degree of unsaturation and use of a high concentration of crosslinkingpromoting peroxide (di-tert-butyl peroxide) to stabilize the rubber particles, is required to improve grafting. Another example of an alternative rubber system is the asymmetric radial polymer (ARPS). ARPS has four equal arms of polybutadiene, with a polystyrene segment attached to one of the polybutadiene arms. A HIPS product made with ARPS blends polybutadiene produces two separate rubber phases with different morphologies and particle size distributions. The ARPS produces a capsular morphology and the polybutadiene produces a normal cellular morphology surrounded by a lamellar structure that provides a reactor product with both high gloss and high impact.

3.2.1

Grafting

The grafting is accomplished in the commercial mass polymerization process by polymerizing styrene in the presence of a dissolved rubber. Dissolving the elastomer in the styrene monomer before polymerization produces HIPS grades. Since the two polymer solutions are incompatible, the styrene—rubber system phase separates very early in conversion. Polystyrene forms the continuous phase, with the rubber phase existing as discrete particles having occlusions of polystyrene. Different production techniques and formulations allow the rubber phase to be tailored to a wide range of properties. Typically: • Increasing the amount of rubber added will increase the toughness. • Increasing the rubber particle size increases toughness to a point of diminishing returns. • The gel content (fraction not soluble in toluene) of polystyrene in the rubber, indicating the phase volume of crosslinking of the rubber, increases toughness to a point (around 30%) and then is deleterious. • The swell index (volumetric ratio of a swollen gel to its unswollen state), indicating the strength of the crosslinked domain, follows the same trend as gel content with an optimal toughness achieved typically around 12% swell. The aim of any grafting is to increase the rubber efficiency, i.e. the ratio between the gel content and the rubber content, and to enable the rubber particles to bond to the polystyrene phase in order to ensure the transmission of external forces from the energy-elastic phase to the entropy-elastic phase. The graft polymer acts as emulsifier and stabilizes the dispersed rubber particles in the two-phase system.

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

259

The grafting mechanism has not yet been explained clearly. Addition to vinyl groups (Figure 12.10) and abstraction of allylic hydrogen atoms (Figure 12.11) have been suggested. Which of the reaction mechanisms dominates depends on the type of initiator, the temperature, the solvent and the substrate. For example, the addition reaction increases with increasing 1,2-vinyl content in the polybutadiene, while hydrogen abstraction dominates in the case of EPDM rubber. Owing to the extremely high importance of grafting for the particle size, the particle structure, the rubber efficiency and the phase adhesion, numerous attempts have been made to increase the graft yield.

i'(R') (R) I

(R) I

(R) I

Figure 12.10 Reaction of the C = C double bond with an initiator free radical (I*) or polystyryl free radical [2–4]

IH (RH) +

Figure 12.11 radical (R*)

Hydrogen abstraction by an initiator free radical (I*) or a polystyryl free

260

3.2.2

M. F. MARTIN ETAL.

Crosslinking

Polybutadienes are polyfunctional compounds which, in a free-radical environment, are not only graftable to styrene, but also their molecules react with one another, initially with the formation of long-chain branches. On further reaction, a coherent network forms. Crosslinking of the rubber commences even during the polymerization at a conversion of above 50 % and especially increases during workup of the polymer in the finishing zone. This process is time and temperature dependent. A quantitative description of the Crosslinking density of these heterogeneous gels has been provided by Karam and Tien on the basis of a modified Flory— Rehner equation, provided that the proportion of polystyrene occlusions is known. In order to retain the elasticity of the rubber and to avoid increasing the glass transition temperature excessively, the Crosslinking density must not be too high. The effect of the Crosslinking density on the glass transition temperature has been studied by Nielson.

3.2.3

Phase Volume Ratio

The ratio between the volumes of the soft component and the polystyrene matrix is known as the phase volume ratio (where the soft component is defined as the rubber component including the occlusions). The volume ratio depends on the rubber content, the number and size of the occlusions and the degree of grafting. The importance of the phase volume ratio for the mechanical properties of the polymer was first recognized by Cigna, and was later confirmed in a series of investigations carried out by Bucknall.

3.2.4

Particle Size and Particle Size Distribution

The rubber particle size (RPS) range of HIPS is essentially determined by three factors, namely the shear field of the reacting material during and immediately after phase inversion, the viscosity ratio between the disperse and coherent phases and the degree of grafting, which affects the interfacial tension. A number of papers have dealt with the break-up of drops of an emulsion in various shear fields. Rumscheidt and Masan described the mechanism of drop formation for various viscosity ratios. Karam and Bellinger showed that the critical deformation of the drops, to allow them to be broken down, reaches a minimum when the viscosity ratio approaches 1. A decrease in the viscosity of the coherent phase results in an increase in the RPS and a reduction in the stability of the oil-in-oil emulsion. An increase in grafting initially gives smaller cell particles and ultimately new particle structures.

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

261

The influence of the shear field has been known for some time. With increasing shear, the particle size decreases, i.e. the particle size distribution peak shifts to lower values, and the relative proportion of smaller particles increases.

3.3

ENVIRONMENTAL STRESS CRACK RESISTANCE (ESCR)

In the presence of stress, certain agents, which are normally innocuous to HIPS, can cause catastrophic failure. This phenomenon is environmental stress cracking (ESC). Substances that cause severe ESC towards HIPS include chemicals such as aliphatic hydrocarbons, e.g. heptane, and foodstuffs with a high fat content, e.g. butter. Commercial trials have identified the four structural parameters which strongly influence the ESCR in HIPS, namely matrix molecular weight, rubber particle size, rubber phase volume or gel content and degree of rubber-phase crosslinking (lowering the swell index). Additives such as mineral oil and other plasticizers can affect the HIPS product environmental stress crack resistance. ESCR properties in HIPS can be improved by absorbing the energy driving the crack propagation and slowing the transport of the aggressive liquid. An optimum ESCR-grade HIPS, then, can be developed by adjusting the swell index and gel content while maintaining a relatively high molecular weight and increased rubber particle size to achieve a desired balance of tensile, impact and ESCR properties. A highly grafted rubber particle, with a high volume of high molecular weight occlusions will absorb the energy, driving propagation better than a poorly grafted small particle. Higher molecular weight grafts, formed for instance when the chain transfer agent is added to the process early in conversion, resulted in only minimal increases in gel content, but drastic increases in the ESCR of some HIPS systems. Increasing the crosslink density, as measured by swell index, of the elastomer through various means of thermal input to the final polymer can double the ESCR of some polymer systems, while having no significant effect on other polymer characteristics. Tables 12.6–12.11 demonstrate anecdotal evidence of these relationships that can be extrapolated to industrial practice. The ESCR performance of a resin is not easily modeled. A laboratory technique for the preparation of thin films of HIPS materials for the study of deformation processes by microscopy allows the deformation process to be better understood. The transmission electron microscope (TEM) allows direct visualization of the crazes themselves in thin films. For good contrast between the crazes and the bulk polystyrene, thin, cast films from 0.5 to 2|xm are required, and also staining of the rubber phase with a heavy atomic species to provide contrast between the rubber and the polystyrene. Another intricacy of this method requires a solution of the HIPS material in a 65:35 methyl ethyl ketone—toluene solution to prevent significant swelling of the rubber particles during the preparation process.

262

M. F. MARTIN ETAL.

Table 12.6 Comparison of a gloss grade HIPS to an ESCR grade (the reagent used in the stress-crack test is a blend of cotton-seed oil and oleic acic: CO—OA). Polymer property

Units

Gloss

ESCR

MFR Vicat Izod Gardner Tensile yield Tensile break or rupture Tensile modulus Tensile elongation Gloss

g/l0min °C ft-lb/in in-lb psi psi psi % % reflectance of 60 ° incident light

4 102 1.2 70 4000 3400 303000 23 90

2.5 99 2.5 190 2500 3500 180000 75 15

Polymer characterization

Units

Gloss

ESCR

Rubber level RPS Rubber particle size distribution (RPSD) Gel content Swell index Molecular weight Molecular weight distribution ESCR CO-OA

% jim

1.2

7.8 3.6 2

12 10 90000 2.2

34 12 90000 2.2

9.5 0.14

%

11

min to failure (1000 psi constant stress)

100

Table 12.7 Effect of swell index adjustment through heat aging on the CO-OA environmental crack resistance of HIPS: ESCR is measured as the time (in minutes) to failure when exposed to reagent under 1000 psi stress Melt flow Swell index Control Heat aged (2h, 200 °C, nitrogen-purged vacuum oven)

ESCR (min) 56 113

11.8 8.1

2.8 2.1

Table 12.8 Effect of various solvents on HIPS: samples are exposed for 10 min under a constant bending strain and then the samples tensile elongation is compared to a control; results are reported as percentage retention of tensile elongation

Standard HIPS extrusion grade ESCR grade

Exposed to Rll

Exposed to 141b

Exposed to CO—OA blend

20 85

14 65

40 82

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

263

Table 12.9 Effect of polybutadiene rubber viscosity on ESCR properties of an extrusion-grade HIPS

MFR(g/10min) Rubber (%) Weight-average RPS (m) Gel content (%) Swelling index ESCR

High-viscosity rubber

Low-viscosity rubber

2.2 10 3.2 25.6 13.8 31

2.2 10 4.8 27.9 11.8 51

Table 12.10 Effect of structural properties on ESCR in an extrusion-like HIPS with 7.8 % high-cis polybutadiene rubber Experiment

MFR(g/10min) Weight-average RPS (m) Gel content (%) Swelling index ESCR (l000psi constant stress) (min to failure)

2.4 5.6 28.8 7.9 46

B

C

D

E

F

G

2.5 5.6 28.5 10.2 33

2.9 5.6 29.9 11.8 20

4.8 6.2 38.3 9.2 19

3.7 5.9 34.2 7.8 35

3.4 5.7 30.5 9.8 20

2.8 6.4 36.5 9.4 60

Table 12.11 Study to understand effect of swell index on ESCR (swell index was adjusted by finishing process temperature)

Finishing process temperature Finishing process residence time Swelling index Gel content Rubber content RPS RPSD MFR ESCR (1000 psi tensile)

Units

Sample A

Sample B

C min % % % M

258 7.5 10.8 22.1 7.5 3.14 1.8 3.2 15

210 7.5 15.2 22.2 7.5 3.19 1.8 3.4 24

g/ 10 min min to failure

With this method, the crazing behavior of an ESCR grade and a standard HIPS gloss grade reveals that the ESCR grade has large rubber particles, with numerous PS occlusions. Crazes tended to propagate 'through' the rubber particles in the ESCR grade. The craze, typically being of smaller dimensions than the rubber particle, impinges on the rubber particle, causing deformation inside the particle and eventually leading to crazing of the PS occlusions. At failure, the rubber particle splits as the craze grows in width.

264

M. F. MARTIN ETAL.

A gloss grade HIPS, on the other hand, reveals a completely different type of crazing morphology. In this polymer, the rubber particles are small, and contain a single occlusion, giving rise to the typical 'core-shell' morphology. These samples craze in completely different manner than the ESCR grade. In the gloss grade, the crazes generally had dimensions larger than the rubber particles and it is apparent that the rubber/PS interface was so strong that separation of the rubber from the PS matrices did not take place. Instead, the interaction of the craze and the particles results in the rubber particle 'splitting' immediately, leaving the PS core (surrounded by a layer of rubber) in the middle of the craze, and a 'cavity' at the edge of the craze. Aggression of a liquid upon a rubber-modified polymer under stress is well studied and depends upon capillary transport of the liquid to the crack tip. Crack propagation is resisted by the molecular weight of the continuous phase and aided by the plasticization effect of the aggressive liquid or plasticizer in the polymer matrix. High molecular weight, unplasticized products can be expected to have enhanced ESCR to aggressive liquids. Polymer additives that can be incorporated in the polymerization process or post-compounded into the HIPS product can add significant enhanced functionality. The use of 2% polyisobutene (PIB), for instance, in the HIPS process feed solution in place of the usual plasticizer can dramatically increase ESCR by a factor of 10 for both standard- and ESCR-grade HIPS. It is interesting that even in the presence of PIB, RPS still has a significant effect on ESCR, as seen in a comparison of extrusion-grade with ESCR-grade HIPS in Table 12.12. 3.4

THERMAL AND OXIDATIVE STABILITY

As with other organic polymers, HIPS is susceptible to degradation when heated in the presence of oxygen. Degradation is a broad term that takes on different meanings to the end user. While it is not easily seen at the molecular level, degradation is recognizable when a finished part is put into its intended use. Typical macroscopic evidence is seen in Table 12.12 Incorporation of PIB improves the environmental stress crack resistance of commercial polymers exposed to 50:50 CO—OA under a constant stress of l000psi

Mineral oil PIB (%) RPS Weight-average RPSD MFR(g/10min) ESCR (min to failure) Izod (ft-lb/in)

Extrusion grade

ESCR grade

1.8 0 4.2 2.4 3.5 15 2.0

0 0 5.4 3.5 2.6 62 2.0

0 2.0 4.2 2.7 2.7 173 2.1

1.8 2.0 6.3 3.5 3.0 262 2.0

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

• • • •

265

discoloration; gel formation; loss of mechanical strength; and/or increased flow or decreasing viscosity.

The last two are the result of a reduction in the polymer's molecular weight. It is also interesting that degradation occurs to some degree immediately after the commercial polymer is produced. It can occur during various phases of the polymer's life cycle: • • • • • •

stripping or devolatilization stage of polymerization process; compounding (if special additives are required); storage and shipment of the resin pellets; fabrication of the final part; end-use environment; and/or in a post consumer recycling process.

The mechanism of polystyrene degradation has been extensively studied and the basic process is similar to that of other organic long-chain molecules based on vinyl groups. Even in the absence of environmental oxygen, the polymer backbone will contain some oxygen groups introduced during the polymerization. Typically, the oxygen enters a closed process with the styrene monomer in the form of dissolved oxygen or oxygenated impurities. It is also readily available in polymerization processes aided by organic peroxides. With atmospheric oxygen available, the degradation process is that much more pronounced. In cases where no additional oxygen is present, polystyrene can undergo nearly pure thermal degradation. The two prevalent mechanisms are sequential elimination of monomer units, which is called unzipping or depolymerization. In this case, styrene monomer is formed. Random chain scission can also occur. It is sometimes combined with unzipping at the reactive broken chain ends. At temperatures approaching 300 °C, up to 40 % of a polystyrene molecule can be converted to styrene monomer. At elevated temperatures in the presence of oxygen, a polymer will decompose to form a polymer peroxy radical, which in turn forms a polymer hydroperoxide and another polymer radical. The polymer radical then combines with oxygen to form another peroxy radical, and the process repeats itself many times before termination occurs. The main cause of the decrease in molecular weight is the chain scission originating at hydroperoxide groups. HIPS is particularly vulnerable to oxidative degradation owing to the unsaturated poly butadiene rubber used. The carbon-hydrogen bond adjacent to the carbon—carbon double bond is many times more likely to be attacked than the polystyrene backbone. This makes HIPS resins much more sensitive to degradation than GPPS resins.

266

M. F. MARTIN ETAL.

Fortunately, there are special additives called antioxidants that can protect the polymer to varying degrees through specific mechanisms. For polystyrene, especially HIPS, a combination primary and secondary systems is often used. The final choice of antioxidant package will be determined by a combination of cost and performance required by the application. Sterically hindered phenols, such as butylated hydroxytoluene (BHT), are excellent primary antioxidants that serve to scavenge the polymer or polymer peroxy radical by donating a hydrogen to cap the radical and form a relatively inactive radical. Another compound with a phenol group that has gained acceptance is a-tocopherol, commonly known as vitamin E. Aromatic amines can also perform this function but are not used because they can cause coloration. Primary antioxidants will be found at levels from 500 to 5000 ppm in commercial polystyrene resins. HIPS resins contain the higher end of the range. Interestingly, the degradation of certain hindered phenols can lead to slight color formation in the finished resin. While the primary antioxidant serves a critical role, it cannot stop all polymer peroxy radicals from propagating. This is where a second class of antioxidants, called peroxide decomposers, comes in. These molecules catalyze the decomposition of the peroxides to nonradical species, thus breaking the repetitive cycle of radical formation. Phosphites and thioesters commonly serve as secondary antioxidants. Phosphites are commonly used in HIPS resins, but care must be taken to use hydrolysis-resistant molecules to avoid the degradation of these species into black specks that render the final product unacceptable. Phosphites are usually found at levels between 500 and 2000 ppm.

4

MANUFACTURE

The first patent on HIPS, a blend of synthetic rubber and transparent polystyrene, was granted in Great Britain as early as 1912. The first graft copolymerization of styrene in the presence of rubber was carried out by Ostromislensky [5]. The decline in the demand for styrene monomer and styrene—butadiene rubber and the simultaneous availability of natural rubber on the world market in the late 1940s drove the development of styrene copolymer processes.

4.1

PROCESS EVOLUTION

The commercial manufacture of polystyrene was batch mode through the 1930s and 1940s, with a gradual transition to continuous bulk polymerization beginning in the 1950s. Suspension polymerization was a common early polystyrene production process, where a single reactor produced a polymer slurry that had to be separated from the water and dried. This process was ideal for free radical

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

267

polymerization to a high degree of conversion and provided an excellent means to remove the exothermic heat of reaction. The dried polystyrene beads would often be compounded in a single screw extruder or Banbury mixer/extruder system, with additives and cut into cylindrical pellets for fabrication. The suspension process was successful for polystyrene homopolymer, but highly efficient rubber modified grades could not be made in this way. Typically, adding an elastomer to the product in the compounding step made the impact-modified grades. The resulting HIPS grade lacked the toughness per unit of rubber found in today's grades. As a means to improve the rubber utilization, a bulk/suspension process evolved, whereby polybutadiene rubber was dissolved in styrene monomer and polymerized in bulk beyond phase inversion before being dropped into suspension. The HIPS produced this way had two distinct advantages over the compounded version: styrene to rubber grafting and discrete rubber spheres or particles uniformly dispersed in a polystyrene matrix. This improved the impact strength dramatically per unit of rubber and gave better processing stability, because the rubber phase was dispersed instead of being co-continuous with the polystyrene. Eventually, bulk polymerization was favored, because by installing reactor vessels in series it is possible to increase the production rate of polymer for a given reactor volume and eliminate the water separation and drying stages. Improvements in reactor design, heat exchange and temperature control also aided the switch. The early bulk polymerization systems used 5-10% solvent to provide a safety factor for runaway reactions. These adverse reactions were not uncommon given the available control and heat transfer technology of the day. Typically, ethylbenzene was the solvent of choice, although toluene would also work. Aromatic solvents were chosen because of their relatively low toxicity, excellent polymer solubility and moderate volatility similar to styrene monomer. Before true continuous reactor trains became common, many were operated in a semi-continuous mode. Typically, there were three or four reactors in series and the styrene would be polymerized to a certain degree of conversion and transferred to the next vessel. This would allow reactants to be transferred into the vacated vessel and batch polymerization begun. This scheme was successful in normal operation, but a surge vessel was needed in case there was a problem with any of the reactors in sequence. Eventually, gear pump and melt pump technology reached a point where it was feasible to pump the partial polymer continuously between vessels. Many of the semi-batch systems were simply upgraded to allow for continuous feed and discharge of final polymer. The design of the reactors in the beginning of the train were different to those at the end because of the increasing viscosity and the slow-down of reaction rate as the styrene conversion increased. Typically, the beginning reactors were continuous stirred tanks (CSTR), while the

268

M. F. MARTIN ETAL.

finishing reactors were plug or tube flow. The finishing reactors needed increased surface area per unit volume to remove heat from the highly viscous, but still reacting mass. After the conversion was completed, typically 65–80% based on styrene feed composition, the polymer and unreacted monomer and solvent were stripped or devolatilized under high temperature and low pressure (partial vacuum). Because the styrene polymerization rate decreases with increasing conversion (lower styrene concentration), the temperature profile was usually an increasing one. For purely thermal initiated systems, the beginning temperatures were between 120 and 135°C and increased upto 160–170 °C. Some processes used organic peroxides to assist the thermal initiation, as was practiced previously in the batch suspension process. The choice of peroxide(s) was different because of the dynamics of the continuous reactor train and the fact that each vessel had a unique styrene concentration versus a single vessel with decreasing styrene concentration. Early-stage peroxides, such as benzoyl and terf-butyl peroctoate, were common choices, while finishing peroxides with higher half-life temperatures were chosen. Common finishing peroxides are dicumyl and di-tert-butyl peroxide. In GPPS systems, these peroxides mainly supplemented the free radicals generated by thermal initiation, whereas in the HIPS process, it was found that they could enhance the grafting of styrene to unsaturated rubbers, such as polybutadiene. Additional benefits of organic peroxide initiators were increased production per unit reactor volume, reduction of styrene oligomers and lower reactor temperatures. The instantaneous removal of peroxide feed to a runaway reactor also provides a safety mechanism. Peroxide-initiated systems have higher reaction rates owing to shorter reactor residence times, so the ability to remove one source of radical initiation quickly is important.

4.2

MODERN COMMERCIAL PROCESS

Today, HIPS is produced by two basic variants: the batch process and the continuous process. Pre-polymerization, i.e. the polymerization phase up to completion of phase inversion, is identical in the two process variants. After completion of the pre-polymerization, the polymerization is continued in suspension in the batch process and in solution in the continuous process. The batch process is, therefore, also referred to as the 'bulk suspension process' and the continuous variant as the 'solution process.' The continuous process is a refinement of the original I.G. Farben process for standard polystyrene, which The Dow Chemical Company has adapted to the needs of rubber-containing styrene solutions. A number of modifications are now practiced.

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

4.2.1

269

Continuous Bulk Process [8]

Today, HIPS is produced predominantly by the continuous bulk or solution process. Industrial production processes can be subdivided into batch preparation, pre-polymerization, main polymerization and work-up. During batch preparation, the rubber, which is supplied in bale form, is comminuted by machine to give pieces with a size of about 1 /cm3, and dissolved in styrene containing a diluent (ethylbenzene or toluene) to give a 2–15% solution. Other reactants and polymer additives can also be added to the rubber solution. Typical ingredients in HIPS include white mineral oil plasticizers, metallic stearates and hindered phenol antioxidants. A chain transfer agent (ethanethiol) can also added to the feed solution, instead of a specific reactor, depending on the desired molecular weight and distribution. The solution is filtered and fed to the pre-polymerization step via pre-warmers. In the pre-polymerization vessels, the rubber solution is polymerized to a conversion of 20–30 %. This phase is where the particle structure, the RPS and the RPSD are fixed. In industry, the pre-polymerization is carried out in continuous-flow stirred tank reactors (Shell, Monsanto, Mitsui Toatsu), tower reactors (Dow Chemical), stirred reactor cascades (BASF) or loop reactors with static mixers (Dainippon Ink and Chemicals). Following preparation of the rubber feed solution, the HIPS process will use one or two stirred reactors to form the rubber phase. In a single CSTR system, the reactor operates above phase inversion. Here, the steady-state effluent is a continuous polystyrene matrix, containing unreacted styrene monomer and distinct rubber particles with occluded polystyrene and styrene monomer. In a two-CSTR system, the first vessel operates below phase inversion to provide superior grafting. In this case, the continuous rubber phase and its reactive sites are readily available for styrene radical grafting reactions to occur. The second vessel operates above phase inversion, as in a single CSTR system, and controls the size of the rubber particles. Whereas the pre-polymerization is carried out at temperatures of from 100 to 150 °C, the main polymerization is carried out at up to 180 °C. Its only aim is to increase the conversion and, thus, improve the economic efficiency of the processes. Target conversions are above 90 %. In order to be able to dissipate the heat of reaction from the solutions of exponentially increasing viscosity in a controlled manner, a number of reactors are generally connected in series. The designs vary considerably. For example, conical reactors with helical ribbon stirrers, horizontal tank reactors with paddle stirrers, reactor cascades and tower cascades have been proposed. The finishing reactor designs vary, with conical reactors stirred by helical ribbons, horizontal cylinders with paddle agitators, shell and tube heat exchanger designs and vertical tower reactors with agitators. In both HIPS and GPPS processes, these vessels maximize the conversion of monomer, without lowering

270

M. F. MARTIN ETAL.

the overall molecular weight to impractical levels. The final polymer is then pumped to a stripping or devolatilizing vessel to remove unreacted styrene and the solvent.

4.2.2

Bulk Suspension Process

The bulk suspension process for HIPS was developed by Monsanto. Batch preparation and pre-polymerization are equivalent to the solution process. After completion of phase inversion, the polymer solution is dispersed in a 2–4-fold amount of water. Suspension aids used are water-soluble organic polymers, such as poly(vinyl alcohol) or polyvinylpyrrolidone, or inorganic compounds, such as Pickering systems. In order to achieve a final conversion of 99.5%, initiator combinations with different decomposition times are used, and the polymerization follows a defined temperature-time profile. The suspension is then centrifuged, dried and compounded. The advantages of the bulk suspension process are its high flexibility, reliable heat dissipation and complete conversion of the styrene.

4.2.3

Post-polymerization Processes

The polymer solution from the polymerization process contains the polystyrene in its final molecular structure and morphology. Additives, such as lubricants, stabilizers, etc., have frequently already been incorporated. Completion of the polymer solution to give the final product generally consists of two steps. In the first, the solution is degassed, such that the low molecular weight constituents, such as solvents, styrene and oligomers, are evaporated as far as possible, and the high-viscosity polystyrene melt is compounded in a subsequent step to give the final product and granulated. The HIPS is sold in granular form, which is easy for the final consumer to handle. The size and shape of the particles are matched to modern processing machines; the particle weight is generally between 10 and 40 mg and the bulk density is from 0.5 to 0.7 kg/L. The technology employed for the stripping varies considerably. Common types are thin-film wiped evaporators, specially designed shell and tube heat exchanges above an open chamber, thin-film plate exchangers and vented twinscrew extruders. Typical residual styrene levels are 300–1000 ppm in conventional systems. Some specially designed GPPS stripping sections can produce levels between 100 and 200 ppm. After stripping, the polymer melt is usually pumped with a gear pump or extruder to a strand die where the final granules are formed. Two types of pellets are common, cylindrical and near spherical. The cylindrical are cut on a wheel after cooling in water, whereas the near spherical

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

271

are cut under water with a die submerged in water. The pellets are then classified to provide uniform size and fully dried before storage and shipping.

5

FABRICATION

Whereas all conventional thermoplastic fabrication techniques have been successfully employed to convert pellets of HIPS into useful articles, extrusion (film, sheet, profile and multi-layer) and injection molding (solid, structural foam and gas-assist) are the predominant processing technologies. Innovative hardware technologies, in both extrusion and injection molding, have provided means to combine less expensive materials, such as polystyrene, with polymers or structures offering key performance characteristics. One of the unique features of the optimal extrusion or injection molding process utilized for HIPS, when compared with other styrenics, is simplicity. HIPS has good thermal stability and resists degradation over a wide range of conditions. Since it is a nonpolar polymer, HIPS displays only a slight tendency to absorb water and the natural, uncolored polymer rarely requires drying prior to melt processing. Most thermoplastic processes utilized for HIPS applications use a single screw to melt and pressurize the polymer. Screw extruders in industry used with HIPS can have screw diameters from 20 to 200 mm and can melt process over 2000 kg/h. Two-stage, vented screw designs are predominantly used for extrusion of HIPS, whereas single-stage reciprocating screws, with a nonreturn valve, are common in the HIPS injection molding process. The vent in the extruder used in the HIPS extrusion process allows the removal of volatiles from the melt. The classical vented screw extruder optimized for HIPS has a length-to-diameter ratio of 36 and a compression ratio of 2-2.5. Standard injection molding screws generally have an effective length of 20-23 D, the length of the feed section being approximately half the length of the screw. The compression and metering sections are of approximately the same length. The pitch is usually 0.8–1 D and the flight-to-depth ratio of the feed and metering sections ranges from 2 to 3. A compression ratio of 2-2.5 has proved to be effective for the processing of styrene polymers and copolymers in injection molding applications. Optimized flight depths for three-section screws are shown in Figure 12.12, as a function of the screw diameter [6]. In the flight depths illustrated in Figure 12.12, a distinction is made between standard screws and shallow-flight screws. Shallow-flight screws pick up less material and, as a result, the residence time is the plasticizing unit is shortened. This can be advantageous when utilizing thermally sensitive additives, such as colorants, or when the part production rate is a small compared with extruder capacity.

M. F. MARTIN ETAL.

272

The 'quality' of the polymer melt produced by the injection unit greatly influences the properties of the molded part. Melt quality depends, to a large extent, on the plasticating performance of the injection screw. A screw not suited to the type of material being processed will produce inhomogeneities in the melt, which may take the form of uneven temperature distribution, poor pigment distribution or inconsistent melting. Further effects of poor screw selection might be excessive melt shear, which could lead to overheating and degradation of the polymer and so cause discoloration and worsening of the part's mechanical properties. Three-section screws are suitable for injection molding of styrenic polymers. Some machine makers offer screws as long as 20–23 D for greater plasticating performance. The screw's compression ratio - the ratio of the depth in the feed section to that in the metering section - should be between 1:2 and 1:2.3. Shallow-flight screws convey less material, but have the advantage that the melt spends less time in the barrel and, therefore, is exposed to less heat. This is particularly important for heat-sensitive products, such as styrene–butadiene copolymers and flame-retardant grades. Very deep cut screws can lead to poor quality parts owing to entrapped air.

D

D L LE LK LA hA hE S R

h

Outer diameter of screw Effective length of screw Length of feed section Length of compression section Length of metering section Flight depth in the metering section Flight depth in the feed section Pitch Nonreturn valve

20–23 0.5–0.55 0.25–0.3 0.2

D L L L

0.8–1.0

Figure 12.12 Injection molding machine reciprocating extruder terminology and typical design parameters. The nonreturn value (R) is illustrated in Figure 12.13

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

273

Hamstring

Screw tip

15°

H*

Flow cross sections in these regions must

Figure 12.13 Nonreturn valve, an important design feature of the injection molding screw, terminology and typical design parameters

The injection molding screw diameter chosen should be matched to the part's shot volume, which - as a general rule - should be not more than about 70 % of the screw's theoretical swept volume. If the screw (diameter) chosen is too small, the swept volume will tend towards the shot volume, leaving insufficient leeway for correcting process fluctuations. On the other hand, if the injection unit is too large, only a small part of the swept volume will be employed per shot and the remainder of the melt in the barrel will possibly be exposed to too much heat. Furthermore, even small variations in the screw stroke would cause considerable variation in the shot weight.

5.1

FABRICATION PROCESS AND PART PROPERTIES

The mechanical quality of a finished part from any fabrication process does not depend only on the material. Structural design matched to the material and processing and also processing conditions appropriate for the product are also of considerable importance. Understanding the effect of process conditions on the injection molded part toughness, as measured by Izod or falling-dart impact, provides an example of the importance of these considerations. These properties are strongly affected by the injection molding process fill speed and melt temperature, as seen in Figure 12.14. Polymer molecular orientation and thermal degradation of the rubber phase most commonly explain this dependence upon process conditions.

M. F. MARTIN ETAL

274

Polymer orientation varies through the thickness of the injection-molded part owing to the 'fountain flow' of the melt in the mold cavity. The flow at the center of the cross-section is deformed through extension and the highly stretched flow front rolls up to the cold mold surface, where orientation is frozen in a thin surface layer. The rest of the melt required to fill the cavity flows under this stationary frozen layer in more or less a plug fashion, with minimum orientation. Surface orientation in an injection-molded part can be significantly different from that in the core of the part. Some tests are affected by core orientation and some test properties are more influenced by surface orientation. Since orientation is not uniform, but has a gradient through and along the flow path, it is difficult to predict directly the effects of process conditions on part properties, without a complex model of the part geometry and estimation of flow characteristics in the cavity. Izod impact strength increases with increasing injection speed and lower melt temperatures. Since the fracture plane is normal to the direction of orientation in the Izod sample, the increased toughness with slower injection speed can be explained by the reinforcing effect of the oriented polymer on the surface. This effect is less noticeable at very high melt temperatures, where the surface can actually be annealed by the temperature of the core and subsequent extended molding cycle time required by the hot melt temperature.

1 2 3 4

200

220

Fast Injection Moderatelnjection Slowlnjection Compression Molded

260

Melt Temperature (°C)

Figure 12.14 Effect of process conditions on the Izod impact strength of an article molded in an injection molding-grade HIPS

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

275

Figure 12.15 demonstrates an opposite effect for dart impact properties. Here, the impact is normal to the orientation in parallel and perpendicular directions, allowing the impact point to be the crack initiator. The less orientation in the article thus provides greater impact strength. The compression-molded part, by definition, does not have flow-induced orientation. Comparison of compression-molded part properties with those of an injection-molded part can show the effect of melt temperature on properties. In the compression-molded article without flow-induced orientation, the impact strength remains constant until a certain melt temperature is surpassed and then decreases. This thermal degradation effect can be attributed to the polybutadiene component, which acts as an initiation site for oxidative degradation of the matrices. With these considerations and with experience demonstrating that there is no standard design for the injection molding process, one is led to the conclusion that each process must be uniquely optimized.

6

APPLICATION

HIPS is used in most applications and industries owing to its easy fabrication and low cost. Major industries and market segments include packaging and disposables, appliances and consumer electronics, toys and recreation, building products and furnishing (Figure 12.13). The largest single use for HIPS is packaging, specifically for food packaging or food service, accounting for more than 30% of world consumption (see Table 12.13). Specialty grades of HIPS, with improved thermal properties and environmental stress crack resistance, have allowed the displacement of engineering resins in applications such as refrigeration and fiber alternatives in large thermoformed cups. The packaging industry imposes many demands on the polymer architecture. Here, high-impact, low residual monomer (< 500 ppm), good stiffness and Table 12.13

HIPS sales in major market segments.

Market segment Packaging and disposables Appliances and consumer electronics Toys and recreation Houseware Furniture Building and construction Custom sheet Miscellaneous

Estimated global share of HIPS consumption (%) 32 18 14 5 3 9 5 14

M. F. MARTIN ETAL.

276

2900 2700

1 2 3 4

Compression Molded Fast Injection Speed Moderate Injection Speed Slow Injection Speed

2500 00

2300 2100 1900

1700

1500 160

180

200 220 Melt Temperature (°C)

240

260

Figure 12.15 Effect of process conditions on the dart impact strength of an article molded in an injection molding-grade HIPS

ESCR are required of a product that has excellent deep-draw capabilities and the ability to be printed upon. The large stadium or 'cruiser' cup is an example of these challenging applications. This product has a design that requires several changes in diameter, while having the rigidity to contain more than 1/L of fluid with a comfortable grip. The successful thermoforming process for this application, which competes with the economics of paper cups, requires a deep draw (greater than 22 cm) in a fast cycle (less than 2 s). A HIPS formulation, such as that shown in Table 12.14, meets the needs for a specialty polymer in an estimated 30 metric ton global market. Appliances account for a significant segment of HIPS consumption. This large market, estimated to be 1 300 000 metric tons globally, includes applications such as small appliances, refrigerator liners, television and computer cabinets, air conditioner components and audio/visual cassettes or packages. Specialty HIPS products used in this market segment are required to meet special environmental chemical resistance requirements presented by sophisticated foam laminate structures used in refrigeration construction. Here, the polyurethane foam insulation presents the potential for chemical attack of HIPS, by the blowing agent in the foam system. A HIPS formulation such as that shown in Table 12.15 meets the needs for a specialty polymer in an estimated 500000 metric ton global market.

277

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

In the refrigeration laminate application, as with many advanced HIPS design challenges, the polymer chemist trades one important property for another. Here, the desired gloss is sacrificed to provide toughness and chemical resistance. The co-extrusion process allows the fabricator to meet the gloss Table 12.14

Advanced HIPS formulation useful in emerging packaging applications

Polymer property

Units

MFR

g/lOmin

Vicat Izod Gardner Tensile yield Tensile fail Tensile modulus Tensile elongation Gloss

°C ft-lb/in in-lb psi psi psi

Value

2.5 102 2 300

% reflectance of 60 ° incident light

3500 3900 350000 60 50

Polymer characterization

Units

Value

Rubber level RPS RPSD Gel content Swell index

%

Table 1 2.1 5

im

% —

7.3 1.2 2.4

30 12

Specialty HIPSformulation useful in refrigerator liner applications Value

Polymer property

Units

MFR Vicat Izod Gardner Tensile yield Tensile fail Tensile modulus Tensile elongation Gloss

g/10 min °C ft-lb/in in-lb psi psi psi o/

% reflectance of 60 ° incident light

2500 3400 180000 75 15

Polymer characterization

Units

Value

Rubber level RPS RPSD Gel content Swell index

%

7.8

um

3.6 2

% —

2.5 99

2.5 190

34 12

M. F. MARTIN ETAL.

278

need of the final sheet by co-extruding a gloss cap on the chemically resistant specialty liner material. While a polystyrene homopolymer could provide the gloss required, another special class of HIPS is preferred (Figure 12.15), since the crystal-HIPS system embrittles the system. Figure 12.16 shows this negative aspect of a PS-HIPS system, compared with that of a hybrid system formed by co-extruding a polymer such as that described in Table 12.15 on the liner material. Small appliances present a challenge when formulating the optimal HIPS product. Conflicting properties of high gloss and high impact, with good processability for high-speed fabrication, can be met utilizing cascading stirred reactors to produce the formulation shown in Table 12.16.

. o

°

a 6-

1 GPPS Laminant 2 HIPS/HIPS Laminant

-1

9.

n. 50

100

150

Thickness Figure 12.16 Effect of gloss (cap) layer type and thickness on the impact strength of the co-extruded sheet

Table 12.16

Specialty HIPS formulation useful in refrigerator gloss cap applications.

Polymer property

Units

Value

MFR Vicat Izod Gardner Tensile yield Tensile fail Tensile modulus Tensile elongation Gloss

g/10 min

C ft-lb/in in-lb psi psi psi % % reflectance of 60 c incident light

3.5 102 1.2 40 4500 3750 300000 30 95

Polymer characterization

Units

Value

Rubber level RPS RPSD Gel content Swell index

% u,m — % —

7.1 0.47 3.1 20 12

PROPERTIES AND APPLICATIONS OF HIGH-IMPACT POLYSTYRENE

279

Other major market segments that can utilize the specialty HIPS in Table 12.17, frequently referred to as a super high-impact polystyrene, include household hardware (shutters), furniture, coolers, housewares and toys. Table 12.17 Specialty HIPS formulation useful in applications requiring high gloss and toughness. Polymer Property

Units

Value

MFR Vicat Izod Gardner Tensile yield Tensile rupture Tensile modulus Tensile elongation Gloss

g/10 min

8 98 3.0 >320 3000 2500 307000 75 60

Polymer characterization

Units

Rubber level RPS RPSD Gel content Swell index

% jjim — % —

7

ft-lb/in in-lb psi psi psi % % reflectance of 60 ° incident light light

Value 8.8 0.99 1.45 30 12

ACKNOWLEDGEMENTS

The authors are indebted to the Process and Product Technology Group of BASF Corporation, the predecessor organizations we started with, and to our families, with whose support, patience and understanding we grew in the plastics profession where we are honored with the opportunity to present this summary to our colleagues.

REFERENCES 1. A. Echte; F. Haaf; J. Hambrecht, Angew. Chem. 93 (1981) 372. 2. A. Echte, in Houben-Weyl: Methoden der Organischen Chemie, 4. Aufl., Bd. E 20, Tl. 2, Thieme, Stuttgart, 1987, p. 996. 3. A. Echte, Angew. Makromol. Chem. 58/59 (1971) 175. 4. H. Willersinn, Makromol. Chem. 101 (1967) 206. 5. J. J. Ostromislensky, US Patent 1 613673. 6. N. Niessner, D. Bender, G. Skupin, A. Wagenknecht, Kunststoffe 85 (1995) 86.

280

M. F. MARTIN ETAL.

7. M. Kuhlmann, BASF Internal Report: Processing Polystyrene and Styrene Copolymers. 8. H. Gausepohl, R. Gellert, Polystyrol Kunststoff Handbuch, Vol. 4, Hanser, Munich, 1996, p. 117. 9. E. R. Moore, Encyclopedia of Polymer Science and Engineering, Vol. 16, Wiley, New York, 1989, pp. 1-246. 10. J. R. Wunsch, Polystyrene Synthesis, Production and Applications, Rapra, 10 (2000) Report 112.

13

Key Structural Features Impacting SAN Copolymer Performance R. P. DION AND R. L SAMMLER Corporate Research, The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Styrene-co-acrylonitrile (SAN) copolymers are a versatile group of materials used primarily as transparent structural materials or blend components for polymer alloying. They are cost effective in structural applications such as pitchers, tumblers, blender jars, food processors, rigid packaging, food containers and automotive battery cases. Competitive materials include glass, acrylic, polycarbonate and polystyrene. Clarity is the primary attribute in these applications, and the specific material of choice is determined by the relative weightings of application specific requirements such as toughness, cost, color, antistatic properties, heat resistance, viscosity and solvent resistance (Table 13.1). The major use for SAN copolymers in blending applications is the continuous phase in acrylonitrile-butadiene-styrene (ABS) formulations, and in poly(vinyl chloride) modification, where low viscosity, low color and high toughness are desirable. Commercial SAN copolymers are typically amorphous with a weight fraction WAN of acrylonitrile between 0 and 0.6, a number-average molecular weight Mn of 35–70kg/mol, and a polydispersity index Mw/Mn of about 2-3. Typical properties for a low-viscosity blending-grade polymer and a higherstrength molding-grade polymer are summarized in Table 13.2. The relative impact of molecular weight and WA N on selected properties of SAN copolymers is summarized in Table 13.3. The higher molecular weights Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy ;c) 2003 John Wiley & Sons Ltd

282 Table 13.1

R. P. DION AND R. L SAMMLER Attributes of competitive transparent materials" Polystyrene

Clarity Color Cost Toughness Processability Solvent Heat resistance a

+ + + — + — —

Polycarbonate

Glass

+ + 0 — -

0 + + — — +

+ + + — — + +

—, 0, + = worse, same or better than SAN, respectively.

Table 13.2

Properties of representative SAN polymers TYRIL" 125 TYRIL" 880 Test method

Property Melt How rate (230 °C/ 3.8kg) (g/10 min) Izod impact strength (0.125 in thick, notched, 23 °C) (J/m) Ultimate tensile strength (23 °C) (MPa) Tensile modulus (23 °C) (MPa) DTUL (1.8 MPa, annealed) (°C) Vicat softening point (°C) Flammability rating a

Acrylic

25 16

3.3 16

ASTM D 1238 ASTM D 256

55.6 3585

77.9 3560 MPa

ASTM D 638 ASTM D 638 ASTM D 648 ASTM D 1525 UL-94 HB

100 113 HB

98 109 HB

Trademark of The Dow Chemical Company.

Table 13.3 Effect of molecular weight and composition on selected properties of SAN copolymers0

Clarity Color Cost Toughness Processability Solvent resistance Heat resistance a

Higher molecular weight

Higher acrylonitrile content

0 0

0

+ — + 0

+ — + 4-

— ,0, + = worse, same, or better by the change, respectively.

and WA N levels are best for resin toughness and solvent resistance, but poor for fabrication/processability. SAN copolymers are produced in free-radical polymerizations using solution, suspension or emulsion processes. Industrially useful products are single-

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

283

phase transparent materials. However, SAN polymers are fairly complex mixtures. The free-radical polymerization process used to make SAN copolymers results in broad molecular weight polydispersities (M w /M n > 2). There is also a composition distribution in materials that is superimposed on the molecular size distribution. Compositional drift is largest for polymerizations that are run to high conversion and do not have the azeotropic composition (76:24 WAV S/AN). Consequently, a wide variety of methods have been used for characterizing the molecular structure and the resultant thermal, rheological and solid-state properties. The macroscopic properties of these materials are determined by both very small-scale micromolecular factors and larger scale macromolecular features. The ability of specific chromophores along the polymer backbone to absorb selectively portions of the visible light spectrum changes the color of molded articles made of SAN copolymers. Additionally, macromolecular factors, such as the larger scale shape of the SAN molecules and the ability of the molecule to entangle with other molecules, can alter the shear thinning behavior of the SAN copolymers during fabrication processes. Characterization methods will be described that probe the structure of the SAN copolymers on a number of scales. Additionally, the impact of the measured quantities will be related back to physical properties that have value in the marketplace.

2 CHARACTERIZATION The most important structural features of amorphous SAN copolymers are the weight fraction (WAN) of acrylonitrile and the molecular weight distribution (MWD). These features control the solid-state properties and fabrication performance. Also important are the type and level of conjugated chromopores and the monomer sequence distribution. These features control the visual appearance of the SAN copolymer. The chromophores may introduce unwanted yellowness. A nonuniform sequence distribution may cause unwanted haze from phase separation.

2.1

CHROMOPHORES

A number of techniques from classical small-molecule characterization have been used successfully with high molecular weight SAN copolymers to determine the existence of specific functional groups. The property that correlates most directly with the chemistry at this scale is color. SAN copolymers have a tendency to become yellow, both with age and with increasing acrylonitrile content, and substantial effort has been expended to determine the source of the

284

R. P. DION AND R. L. SAMMLER

undesirable yellowness [1]. In order to have SAN copolymers appear yellow, a fairly large, conjugated chromophore must be present. Both functionality in the backbone of the polymer and low molecular weight species have been hypothesized as sources of the yellowness (Figures 13.1 and 13.2). The bulk composition of the SAN copolymer can be determined by ultraviolet spectroscopy. Absorbances consistent with conjugated systems such as Figures 13.1 and 13.2 have been observed. Studies usually compare the UV spectra of model systems with the actual absorbances seen in SAN copolymers. The models represent chemically reasonable species based upon the starting monomers, known reactivity ratios, and oxidation and rearrangement chemistry [2]. The data are self-consistent with these criteria; however, identification of all chromophores responsible for color formation in SAN copolymers is still work in progress. One source of species with extended conjugation is the cyclization of acrylonitrile triads to form heteroaromatic structures (Figure 13.2).

2.2

SEQUENCE DISTRIBUTIONS

The sequence distribution of the styrene and acrylonitrile monomer units in SAN copolymers can be determined by 13C NMR spectroscopy, and the experimental

Figure 13.1

Small-molecule chromophore from SAN [1]

Figure 13.2 Polymeric chromophore from hypothesized condensation of an acrylonitrile triad in an SAN backbone. Reprinted from Advances in Polymer Sciences, Vol. 121, Ed. by D. B. Priddy, pp. 124 and 142 (1996), with permission from Springer-Veriag.

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

285

data agree well with calculations based on reactivity ratios [3]. The use of NMR for characterizing SAN copolymers has been reviewed [4]. Other methods that have been used to detect styrene-styrene diads in styrenic copolymers are fluorescence spectroscopy [5] and Fourier transform IR spectroscopy [6].

2.3

AN LEVELS

Pyrolysis gas chromatography can be used to determine the acrylonitrile content of the SAN copolymer [7–9]. It is a method that heats the polymer above the decomposition temperature, then separates and identifies the low molecular weight compounds formed. The primary decomposition products are styrene, acrylonitrile, and propionitrile, and the styrene content of the copolymer is directly proportional to the styrene yield from pyrolysis [8].

2.4

MWD

The methods described above allow the determination of the average composition of the SAN copolymer. However, the polymers formed are rather complex mixtures containing both a range of molecular weights and a range of chemical compositions. Since the average composition of the polymer does not uniquely define the expected properties of the mixture, several techniques have been developed to measure the breadths of the molecular weight and composition distributions. The molecular weight distribution of SAN copolymers can be determined by gel permeation chromatography [10]. SAN copolymers are soluble in common solvents such as tetrahydrofuran. The solubility characteristics of SAN copolymers, combined with the commercial availability of columns and data analysis software, make gel permeation chromatographic analysis a rapid and routine procedure. The choice of detectors can allow for the determination of absolute molecular weight [11]. Selection of multiple detectors enables characterization of compositional heterogeneity as a function of molecular weight [12], as illustrated in Figure 13.3.

2.5

COMPOSTION DISTRIBUTION

The composition distribution of SAN copolymers can be determined by liquid chromatography. An effective method is to precipitate the polymer on a column, and then successively elute fractions that differ compositionally by increasing the polarity of the solvent [13–18]. Both normal- and re versed-phase chromatography can be used.

286

R. P. DION AND R. L. SAMMLER 100 260 nm 400 nm

12.0

14.0

16.0

18.0

Retention Time (min) Figure 13.3 Compositional heterogeneity of SAN by GPC. Reprinted with permission from D. S. Allen, M. Birchmeier, J. R. Pnbish, D. B. Priddy and P. B. Smith, Macromolecules, 26, 6068.

2.6

MULTIDIMENSIONAL ANALYSIS

A number of multidimensional analyses have been developed that provide powerful methods for characterizing these polymers. Linking a liquid chromatogram to a pyrolysis gas chromatograph [19] can determine the breadth of the composition distribution, as the method fractionates the SAN copolymer before pyrolysis. This information is useful for determining the source of variation in SAN copolymer properties. Composition drift towards high acrylonitrile-containing fractions can lead to undesirable yellow color, and excessively broad composition drift can cause opacity and brittleness in the material due to phase separation Another powerful method combines gel permeation chromatography with gradient elution-precipitation chromatography [20,21]. Simultaneously determining the breadths of both the molecular weight and compositional distributions can provide insights into structure-property relationships and the control of the polymerization process.

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

3

287

FABRICATION PERFORMANCE

There are many ways to fabricate and model viscoelastic polymeric materials [22–32]. Fabrication often involves nonlinear flows that are spatially inhomogeneous, nonisothermal, and temporally complex. The flows also may involve material phase changes, and/or a wide range (1-5 decades) of strains and strain rates. Rheology is often the bridge between resin design and fabrication performance, and remains an active area of research [22]. Rheometers provide key metrics of melt state fabrication and of the subsequent solidification process. However, they are more frequently used to characterize materials undergoing isothermal homogeneous flows, and for a much restricted range of strains and strain rates. They consequently provide only an elementary and/or complementary view of fabrication performance. Both linear (low strain) and nonlinear flows are commonly used in rheometers. The linear experiments are best suited for fingerprinting the long-chain structure of the polymeric materials. The higher strain experiments are best suited for understanding the fabrication performance. Injection molding and extrusion are commonly selected for the fabrication of SAN copolymers, and both are dominated by nonlinear shear flows.

3.1

SHEAR FLOW

An important metric of the melt state is the complex shear viscosity n*(w, T) when measured over a wide range (0.01-100 rad/s) of angular frequencies w and at representative temperatures T in the fabrication window. It may be measured for SAN copolymers in a dynamic rheometer with a thin disk placed between oscillating parallel plates [33]. The resins should always be dried overnight at 105°C in a vacuum oven to minimize development of bubbles when molten. Typical data on double logarithmic axes are illustrated in Figure 13.4 for two SAN copolymers with contrasting AN levels and molecular weights. Data points are marked with individual symbols. Isothermal data points are connected by thin dashed curves. Each curve is based on a nonlinear regression fit to the isothermal data with a simple three-parameter model. The magnitude of |n| typically falls as w increases and/or as temperature increases. The rate of fall is accelerated at high w, and decelerated at high T.

3.2

ENTANGLED CHAINS

The entanglement molecular weight Mee of SAN copolymers at melt temperatures falls from about 18 [30,34] to 5.0kg/mol [34] as WAN rises from 0.0

R. P. DION AND R. L. SAMMLER

288

r,°c r 160 .

polystyrcne M w = 0.00

r.°c

M = I94.6kg/mol, My A/, = 2.53 ... TTS master curve. TO =200°C

191

170 180 190 200 210 • 220 230 240 ; 250 260 •

104

"»v = ° 5I ,M_= 55kgmol. -M...M. = 1.7

161 171 181

]

'ft • • • TTS master curve. TO =202°O

*•.*»

222 232 242 252 262

10'

Cross model fits

10-'

10-'

10°

10'

3

10

co, rad'S

10°

10'

10'

Id. rad's

Figure 13.4 Dependences of magnitude \ti*\ of the dynamic shear viscosity on angular frequency u and temperature T for two SAN resins with contrasting AN levels and molecular weight

to 0.5. Consequently, the chains associated with the melt rheology illustrated in Figure 13.4 are entangled (Mw/Me ^11). The limiting viscosity |//*(o> —»• 0)| is proportional to (M w /M e ) 35±01 for polystyrene and for many entangled flexible chain resins [35]. It is expected to apply similarly to SAN copolymers.

3.3

TIME-TEMPERATURE

SUPERPOSITION

The similarities of the isothermal curves permit the construction of a master curve (thick dashed curve in Figure 13.4) for any reference temperature TO by the well-known time-temperature superposition principle [30,36]. The superposed master curve expands the experimentally accessible w window for these examples by about three decades relative to that measured at any single temperature. This wider window is critical to understanding fabrication performance. Superposition shift factors a are identified numerically. They are designed to transform magnitude |n| and angular frequency w measured at temperature r to a reduced magnitude |n|/aj and a reduced angular frequency war at temperature TQ. Analysis of the T dependence of a allows the shifting of the data to any melt temperature T. This temperature dependence is best described by the WLF relation [30,36] for temperatures (< 1.27 t g ^460K w 190°C) just above the SAN copolymer glass-transition temperature Tg, or by the Arrhenius relation for higher temperatures [37]. Nevertheless, strategies are available to fit styrenic data to the WLF relationship to temperatures as high as 290 °C [38].

289

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE 3.4

CROSS MODEL

Isothermal flow curves are often summarized by simple empirical models to understand fabrication performance. The Cross model [39,40], given by Equation (1), is well-suited for fitting SAN copolymer data as seen in Figure 13.4. (1)

1 + (COT)l-«

The three model parameters (n0, i and n) are often selected with a nonlinear least-squares algorithm which minimizes the squared difference between the measured and modeled ln |n*| for all w at temperature T. Application of a Levenberg-Marquardt algorithm [41,42] to the SAN copolymer data in Figure 13.4 yields fit parameters summarized in Table 13.4. Error bars are reported to two standard deviations. The parameter n 0 is the limiting viscosity at low w. The reciprocal () of relaxation time i marks the midpoint co for the transition from a power-law exponent of 0 at low wco to n — 1 at high co. Interpretation of these low strain amplitude parameters in nonlinear fabrication shear flows is enabled by the Cox-Merz rule [43].

3.5

NONLINEAR SHEAR FLOWS

Cox and Merz illustrated empirically the equivalence of the co dependence of rj* | to the shear rate y dependence of the steady-state shear viscosity n(y). This rule has been tested frequently, and found to be reliable for flexible-chain resins Table 13.4 Temperature dependence of Cross model parameters for a SAN resin (WAN = 0.51, Mw = 55kg/mol, Mw/Mn = 1.70) Temperature

r(°Q

160.97 ±0.11 171. 18 ±0.14 181.48 ±0.14 191.48 ±0.14 20 1.53 ±0.09 21 1.58 ±0.08 22 1.60 ±0.08 231. 70 ±0.07 241. 80 ±0.08 251. 98 ±0.09 262. 18 ±0.09

Zero-shear viscosity, 2021000 ± 94000 11 25000 ±72000 483000 ± 27000 219400 ± 7800 108000 ± 3000 58800 ± 1700 35100± 800 22100 ± 630 14130 ± 280 8800 ± 120 4860 ± 86

T(S)

Relaxation time, (Pa s)

Power-law exponent, n

14.5 ± 1.5 8.9 ± 1.5 4.59 ± 0.79 2.25 ± 0.28 1.14 ±0.12 0.659 ± 0.075 0.456 ± 0.042 0.336 ± 0.039 0.243 ± 0.021 0.1 665 ±0.0096 0.1076 ±0.0084

0.203 0.246 0.281 0.305 0.323 0.343 0.371 0.401 0.422 0.436 0.459

± 0.009 ± 0.015 ± 0.018 ± 0.015 ± 0.014 ± 0.016 ±0.012 ± 0.016 ±0.011 ± 0.008 ±0.011

290

R. P. DION AND R. L SAMMLER

[22]. In this viewpoint, parameter n0 is the limiting steady-state viscosity for creeping shear flows, i~l marks the shear rate for the transition between Newtonian flow [f/(y) = r\Q] and power-law flow [//(y) oc y""1], and (n — 1) is the limiting shear-thinning rate din >/(y)/dln y at high y. Of course, the powerlaw slope (n — 1) must vary from 0 to —1 to be physically meaningful, and correspondingly n must vary from 1 to 0. Fabricators will use //0 as a metric of melt strength for creeping flows. For example, selection of resins with higher values of rj0 should minimize sag in thermoformed parts, and allow a more uniform part wall thickness. Similarly, T and n are metrics of the onset and rate of shear thinning. They are used to identify low-f/(y) conditions (easy flow) in an extruder, which are often critical to fast part fabrication rates. Direct measurement of the y and T dependences of r/(y) for SAN copolymers can be made with a capillary rheometer [44,45]. A capillary rheometer is designed to probe the high shear rate (l-10 4 s -1 ) portion of the flow curve. Shear rates as high as 107 s-1 have been reported for SAN copolymers [44,45]. Capillary experiments also permit the study of thermomechanical degradation of polymer melts which may accompany Theological and fabrication experiments at y > 10 4 s -1 . Relative to dynamic experiments, much more material (20-200 g) and effort are required when isothermal y-sweep data are properly corrected by the prescriptions of Rabinowitsch and co-workers [46,47] and of Bagley [48,49]. Both features often preclude the use of capillary experiments for rapid screening of new materials available in research-scale quantities (5-20 g).

3.6

RELAXATION

SPECTRA

An isothermal shear flow curve may be transformed into a relaxation spectrum to fingerprint the time scales of molecular motion in the polymeric melt [50–53]. The spectrum represents the most fundamental aspects of melt rheology, and can be incorporated into a constitutive model to predict melt rheology in any flow [54–57]. The predictions, of course, are best for strains and strain rates within the experimental windows, but often remain very good just outside (< 1 decade) these windows. Fortunately, there are excellent approaches to improve the predictions. For example, large-strain data, measured during startup of uniaxial extensional flow, may be used to identify damping factors that improve the prediction of several types of nonlinear flows [54–57]. Alternatively, tandem use of isothermal creep recovery and dynamic experiments may be used to lower the lowest measurable shear rate from about 10-2 to 1 0 - 5 s ' if the molten resin has sufficient thermally stability to withstand the long creep/ creep recovery times [58]. Discrete spectra with less than five relaxation times are commonly used to coarse grain the Theological data [51–53]. The coarse graining is often necessary

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

291

to simulate fabrication performance with computation-intensive numerical algorithms [57] (or POLYFLOW™ from Fluent). In contrast, continuous spectra are useful for a better understanding of the long-chain structure. For example, a bimodal relaxation spectrum indicates the presence of two distinct polymer-chain length scales. The length scales may be due to two types of linear chains with much different lengths (molecular weights), or to identical comb-like chains with the backbone lengths much longer than the branch lengths.

3.7

EXTENSIONAL FLOW

Reliable measurement of the rheological properties of polymer melts in purely extensional flow and at constant strain rate has always been more challenging than for shear flow. Two complementary designs are in use for startup of uniaxial extensional flow [59–61]. The Meissner design [60,61] is more common today owing to its commercial availability. It remarkably stretches a strip of material to very large strains (Hencky strain E = 7; linear strains of e7 w 1100) in a small oven. The trajectories of small glass beads sprinkled on top of the strip are analyzed to verify that the stretching is spatially uniform and to determine the true strain. Trajectory analysis is especially important for SAN copolymers at temperatures below about 150 °C when sample slippage between the metal belts may become appreciable. No neutral-density oil bath is needed in this latest design to minimize sag of the sample strip. More details on sample preparation, data acquisition and analysis, and estimation of errors on Meissner design may be found elsewhere [62]. The maximum strain rate (e < 1 s - 1 ) for either extensional rheometer is often very slow compared with those of fabrication. Fortunately, time-temperature superposition approaches work well for SAN copolymers, and permit the elevation of the reduced strain rates saj to those comparable to fabrication. Typical extensional rheology data for a SAN copolymer (WAN = 0.264, Mw = 78kg/mol,Mw/Mn = 2.8) are illustrated in Figure 13.5 after time-temperature superposition to a reference temperature of 170°C [63]. The tensile stress growth coefficient n+E, 0 was measured at discrete times t during the startup of uniaxial extensional flow. Data points are marked with individual symbols (o) and terminate at the tensile break point at longest time t. Isothermal data points are connected by solid curves. Data were collected at selected £ between 0.0167 and 0.0840s-1 and at temperatures between 130 and 180 °C. Also illustrated in Figure 13.5 (dashed line) is a shear flow curve from a dynamic experiment displayed in a special format (3|n| versus a)-1) as suggested by Trouton [64]. The superposition of the low-strain rate data from two types (shear and extensional flow) of rheometers is an important validation of the reliability of both data sets.

292

R. P. DION AND R. L. SAMMLER

AS-2, \7Q°C

13 Cu

TV- 6

3n'(t)

3tl. eaT 0.0818 0.973 2.10 7.00 7.89 20.8 40.0 i

•o-' -3

-2

-1

0

1 log(taT -i/s)

Figure 13.5 Dependences of the reduced tensile stress growth coefficient ^|(e,0/aj at 170°C on reduced time teT and reduced strain rate ear for a SAN resin (WAN = 0.264, Mw = 78 kg/mol, Mw/Mn = 2.8) during the startup of uniaxial extensional flow. Also illustrated (dashed curve) are dynamic shear viscosity data displayed as 3|^*(a>, 770°C)| versus u-1 as suggested by Trouton [64]. Reproduced from L. Li, T. Masuda and M. Takahashi, J.Rheol., 34(1), 103(1990), with permission of the American Institute of Physics

Strain hardening, as indicated by */£(£, 0 > 3|f*l» is clearly evident in this SAN copolymer for the higher reduced strain rates (ea-r > 1 s -1 ). This is a very valuable feature in a material since it enhances its ability to deliver uniformly thick walls during extension. However, use of this isothermal feature may be underutilized in fabrication since large strains (e = tsaj > 2; e2 % 7.4) and high strain rates (e > 8) are needed. A more often used feature for hardening copolymer melts (e.g. thermoforming) arises from cooling since the melt viscosity has such a strong dependence on temperature. Here, the flows are often sufficiently slow that the tensile stress growth coefficient nE Ce, t) at time t may be estimated simply with Trouton's rule [3|>/*(a> = f-1)|]. A few rheometers are available for measurement of equi-biaxial and planar extensional properties polymer melts [62,65,66]. The additional experimental challenges associated with these more complicated flows often preclude their use. In practice, these melt Theological properties are often first estimated from decomposing a shear flow curve into a relaxation spectrum and predicting the properties with a constitutive model appropriate for the extensional flow [54–57]. Predictions may be improved at higher strains with damping factors estimated from either a simple shear or uniaxial extensional flow. The limiting tensile strain or stress at the melt break point are not well predicted by this simple approach.

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

3.8

293

BREAK POINTS

Wall thickness uniformity is often compromised when fabricating near the material break point. Thinner walls are expected in the most stretched regions, but only if the melt is not allowed to recoil after cessation of flow. Often fabrication conditions are selected to be well away from the break point to minimize these issues. Key break point metrics for the startup of data illustratrated in Figure 13.5 are the time /b and tensile stress coefficient ^(fi, ^) at break. Both quantities may be multiplied by the strain rate e to estimate the Hencky break strain (eb = t^e) and the break stress (0b = ^E(£, 0)- Similar metrics can be defined for other startup extensional flows. The break stress crb of entangled high polymer melts is thought to be dominated by the intrinsic strength of the polymer backbone, and offset by chain disentanglements when re < 1 and the type and level of stress concentrators (inhomogeneities, impurities, unmelts, dust, etc.) present in the melt. A zerothorder estimate of the intrinsic strength (1 MPa) of styrenic and olefinic resins has been reported by Laun and Schuch [67] at high strain rates e with a Rheotens setup [68,69]. This intrinsic stress can be used to improve predictions of extensional flow with constitutive models when break point data are not available. One simply truncates all predictions of flow when the tensile stress exceeds this limit. Nevertheless, it remains far more preferable to have break point data since stress concentrators may lower ab significantly and adversely affect fabrication performance. For example, the break stress a^ and strain eb are very low (0.01 MPa and 0.82, respectively) for the low £#T data reported in Figure 13.5.

3.9

BRITTLE BREAKS

The slopes of the tensile stress growth coefficient curves near the break point are also important. The steep slopes observed for the higher sa-r are typical of brittle breaks. The flat slopes observed for the low e«x are typical of ductile breaks. Strong elastic recoils are expected after cessation of flow for brittle breaks. The recoil may be characterized in uniaxial extensional flow in the Meissner design [55,60–63]. Rapid cooling and solidification of the stretched melt may minimize the melt recoil but may introduce warpage in the part as residual stress grows during the cooling to the part service temperature. The residual stress may later be responsible for part shrinkage if the part temperature approaches the glass transition temperature.

3.10

FLOW BIREFRINGENCE

The optical anisotropy of molten flexible-chain polymers is often very small in a quiescent (nonflowing) state owing to their nearly spherical chain configuration

294

R. P. DION AND R. L. SAMMLER

and random orientation. Flow will induce an optical anisotropy in the melt by deforming the entangled chains from their equilibrium configurations and aligning the deformed chains relative to the flow streamlines. This anisotropy can be sensed by measurement of the flow-induced birefringence A« [70,71]. Birefringence setups can be designed to characterize molten materials undergoing isothermal homogeneous flow. The ranges of strains and strain rates also often coincide with those of rheometers, and consequently may be limited relative to those used in fabrication. Similarly, time-temperature superposition approaches may be used to expand the rate window. State-of-the-art setups suitable for rapid screening of new materials with research-scale quantities (5–20 g) are available for shear flow [72] and startup of uniaxial extensional flow [73,74]. Birefringence experiments provide complementary information to rheological experiments. Both probe the response of the molten chains to the flow, and can be used to fingerprint their wide range of relaxation times. The birefringence experiment senses the alignment rather than the mechanical impedance of chains undergoing flow. It is widely used as a relative metric of stress fields in non-homogeneous flows at low strains [75-77] since it can be measured at any point in the fluid. The birefringence is transformed into stress via the stress-optic rule. This rule relates the refractive index tensor n to the extra stress tensor T at low strains by the stress-optical coefficient C. Extensions of the rule at higher strains remain an active research area [70,71].

4

MULTIPHASE SYSTEMS

Useful multiphase SAN copolymer mixtures can be made by the addition of an impact modifier to SAN in a polymer blending operation. Acrylonitrilebutadiene-styrene (ABS), which is an impact modified SAN copolymer, can be produced by melt mixing SAN copolymer grafted rubber with blending grade SAN copolymers. The properties of these immiscible high polymer blends are generally not a simple average of the properties of the components. They depend critically on the morphology of the immiscible domains that develop during blending and fabrication. An impact modifier with the incorrect graft phase may disperse poorly in the SAN continuous phase, and exhibit poor properties [78–81]. Extensive reviews of the rich field of polymer blends may be found elsewhere [26,82–84]. Since most high molecular weight polymers do not mix on a molecular scale, it is desirable to determine compositional ranges of miscibility. The solubility of SAN copolymers with other polymers has been measured by a variety of techniques. The tendency of the materials to mix or phase separate is determined by the enthalpy of interactions between 'mer' units in the polymers and by the molecular weight of the polymers. It was determined experimentally that

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

295

SAN copolymer compositions in solution will phase separate when they differ in composition by around 4 mol% [85]. The thermodynamics of this system has been described in detail, in terms of interaction parameters [86,87], and calculated phase diagrams agree with published experimental data. Blend morphology commonly depends on the weight fraction and viscoelastic properties of each component, the interfacial tension between components, the shape and sizes of the discontinuous phase, and the fabrication conditions and setup. Most rheological experiments applied to homogeneous melts can also be similarly applied to these immiscible blends [55,63,88,89]. The viscoelastic properties arising from these studies should be labeled with a subscript 'apparent' since the equations used to translate rheometer transducer responses to properties incorrectly assume that the material is homogeneous. Nevertheless, these apparent properties are often found to be excellent metrics of fabrication performance. Flow of the blend at melt temperatures generally stretches the discontinuous phase from its initial shape. Interfacial tension between the immiscible components will oppose this process and attempt to drive the system to a lowenergy spherical-morphology state. Studies of these phenomena in a rheometer permit the estimation of the time required for both processes and the interfacial tension [90–93]. Alternatively, one can estimate the time required for the latter process, and the interfacial tension, from the evolving shape of the discontinuous phase using either a fiber break-up [94,95] or fiber-retraction [33,96] experiment. Interfacial tension depends on the molecular weight of each component [96,97] and on temperature, so it is preferable to measure interfacial tension for the materials of interest at their fabrication temperature. The compatibility of SAN copolymers with an assortment of other polymers has been measured by a variety of techniques. Differential scanning calorimetry is used to determine the glass transition temperatures of copolymers. The values increase slightly with increasing acrylonitrile content and range from around 100 to 115 °C [98,99]. If the glass transition of a second polymer differs from that of the SAN copolymer by 20 °C or more, measurements of the composite transitions in blends can easily be used to measure compatibility. Blends with one glass transition are miscible, blends with the two original glass transitions are immiscible, and blends in which the two glass transitions have moved closer together are considered compatible. Compositional windows of miscibility of SAN copolymers with a-methylstyrene-co-acrylonitrile, styrene-co-acrylonitrile-cofumaronitrile and styrene-co-maleic anhydride polymers have been reported [98-100]. The commercial interest in these materials is due to the higher end use temperatures of impact modified styrene co- or terpolymers relative to SAN copolymers. The impact modifiers are elastomer cores, grafted with SAN copolymer shells, so the miscibility of the continuous phase polymer with the graft phase of the impact modifier influences the physical property balance of the blend [80,81]. Complex mixtures of SAN copolymers can be separated and

296

R. P. DION AND R. L. SAMMLER

identified using combinations of solvent precipitation, gel permeation chromatography and liquid chromatography [78]. Solid-state NMR has been used to examine compatibility in SAN copolymer blends with styrene-maleic anhydride copolymers [101]. Spin diffusion experiments indicate that the two polymers mix on a molecular scale, but the data suggest that there is no specific interaction between the nitrile group and the carbonyl groups in the maleic anhydride. This technique can provide some powerful chemical data that cannot be obtained by other methods. Unfortunately, the method requires the preparation of 13C-enriched polymers. 5

SOLID-PHASE BEHAVIOR

The micromechanical deformation behavior of SAN copolymers and rubberreinforced SAN copolymers have been examined in both compression [102] and in tension [103,104]. Both modes are important, as the geometry of the part in a given application and the nature of the deformation can create either stress state. However, the tensile mode is often viewed as more critical since these materials are more brittle in tension. The tensile properties also depend on temperature as illustrated in Figure 13.6 for a typical SAN copolymer [27]. This resin transforms from a brittle to ductile material under a tensile load between 40 and 60CC.

2

4 6 8 Elongation (%)

Figure 13.6 Temperature dependence of the tensile properties of a SAN resin. Reprinted with permission from Throne, J.TechnologyofThermoformina, Manser Publishers, Munich, Copyright 1996

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

297

Failure of unoriented SAN copolymers is dominated by crazing behavior. The total energy absorbed by the polymer during failure has been increased by optimizing both failure modes. Yielding can be enhanced by orienting the polymer [105], and crazing can be optimized by rubber modification. Semi-empirical approaches are available to predict selected solid-state properties of thermoplastics from molecular structure. Example predictions for about 30 types of thermoplastics, including SAN copolymers, may be found elsewhere [34]. The rheological and thermal properties of high molecular weight SAN copolymers can be modified by the addition of styrene plus acrylonitrile oligomers [106]. These low molecular weight species form during the copolymerization process [107], and are effective plasticizers. The changes in macroscopic properties are a mixture of desirable and detrimental effects. The melt viscosity is decreased, the toughness is increased, and the heat resistance is decreased. Several other properties of copolymers that are important in specific applications have also been measured. The surface properties of polymers determine the nature of adhesives that will stick to a substrate, and the nature of solvents that will wet the surface. The surface energy of some styrene and acrylonitrile have been measured, and the surface is rich in polystyrene when the acrylonitrile content of the copolymer is below 50% [108]. The properties of SAN and rubber-modified SAN copolymers have also been evaluated for oriented films. Biaxial orientation can increase the glass transition temperature by up to 15 °C and decrease the water vapor transmission rate by 30% [105]. SAN copolymers may be blended with polycarbonate (PC) to reduce the birefringence of magneto-optical storage disks to acceptable levels [109]. Methods to measure the birefringence in three different disk orientations are summarized elsewhere [109]. This approach takes advantage of the opposite sign of the birefringence of each blend component to annihilate the birefringence. PC-SAN blends perform better than PC-S blends in that the weight fraction WAN of acrylonitrile may be selected to compatibilize the blend. Blend compatabilization is desirable as it reduces the level of light scattering from the immiscible domains. It also reduces the driving force for additional phase separation and haze if the disk is exposed to heat. Other useful metrics of the optical properties of SAN copolymers are color [110], haze [111], gloss [112], and refractive index [113].

6

CONCLUSION

SAN copolymers are complex mixtures possessing heterogeneity in both chemical composition and molecular weight distribution. Consequently, analytical characterization of these materials is complex, but absolutely critical for

R. P. DION AND R. L. SAMMLER

298

Table 13.5 Summary of characterization techniques and associated structure-property relationships Scale

Characterization technique

Molecular

Ultraviolet spectroscopy Nuclear magnetic resonance Thermogravimetric analysis-mass spectrometry Macromolecular Gel permeation chromatography Liquid chromatography Differential scanning calorimetry Interfacial tension Melt flow rate Macroscopic Impact strength Tensile strength

Polymer attribute

Related property

Molecular conjugation Sequence distribution Pyrolysis products

Color

Molecular weight distribution Compositional distribution Thermal transitions

Surface energy Molecular mobility Energy absorption Resistance to flow in tension Resistance to flow Tensile modulus in tension Resistance to flow DTUL at high temperature Vicat softening point Resistance to flow at high temperature Flammability rating Flammability

Color, clarity Composition —» color, toughness, viscosity Viscosity, toughness Color, clarity Thermal resistance Toughness in blends Viscosity Toughness Toughness Stiffness Thermal resistance Thermal resistance Flammablility

understanding properties and developing commercial applications. Methodology exists to characterize SAN copolymers at the molecular, macromolecular and macroscopic scales (Table 13.5), and some of the structure-property relationships have been elucidated. This information can be used to modify physical properties and optimize relevant properties to best meet part fabrication and application needs.

REFERENCES 1. Priddy DB (1995) Thermal discoloration chemistry of styrene-co-acrylonitrile. Adv PolymSci 121:123–54. 2. Garcia-Rubio LH and Ro N (1985) Detailed copolymer characterization using ultraviolet spectroscopy. Can J Chem 63:253–63.

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

299

3. Schaefer J (1971) High-resolution pulsed carbon-13 nuclear magnetic resonance analysis of the monomer distribution in acrylonitrile-styrene copolymers. Macromolecules 4:107–10. 4. Borbely JD, Graham D, Hill DJT, Lang AP, Munro P and O'Donnell JH (1991) NMR studies of copolymerizations of acrylonitrile. Prog Pac Polym Sci, Proc Pac Polym Conf 1:237–48. 5. Zeng W and Shirota Y (1989) Studies on alternating radical copolymerization: analysis of microstructures of styrene-maleic anhydride, styrene-acrylonitrile, and styrene-methyl methacrylate copolymers by fluorescence spectroscopy. Macromolecules 22:4204–8. 6. Sargent M, Koenig JL and Maeker NL (1991) FT-IR analysis of the monomer sequence distribution of styrene-acrylonitrile copolymers. Appl Spectrosc 45:1726– 32. 7. Tsuge S, Kobayashi T, Sugimura Y, Nagaya T and Takeuchi T (1979) Pyrolysis gas chromatographic characterization of highly alternating copolymers containing styrene and tetracyanoquinodimethane, methyl acrylate, acrylonitrile, or methyl methacrylate units. Macromolecules 12:988–92. 8. Tsuge S, Sugimura Y, Kobayashi T and Nagaya T (1981) Microstructural characterization of various copolymers by pyrolysis-glass capillary gas chromatography. Polym Prepr (Am Chem Soc, Div Polym Chem) 22(l):284–5. 9. Vukovic R and Gnjatovic V (1970) Characterization of styrene-acrylonitrile copolymer by pyrolysis gas chromatography. J Polym Sci, Part A-l 8:139–46. 10. Del Rios JK (1988) Polymer characterization using the photodiode array detector. Am Lab 20:78-82. 11. Garcia-Rubio LH, MacGregor JF and Hamielec AE (1981) Size exclusion chromatography-characterization of copolymers. Polym Prepr (Am Chem Soc, Div Polym Chem) 22(l):292–3. 12. Allen DS, Birchmeier M, Pribish JR, Priddy DB and Smith PB (1993) Thermal styrene-co-acrylonitrile discoloration problem: the role of sequence distribution and oligomers. Macromolecules 26:6068-6075. 13. Gloeckner G and Van den Berg JHM (1987) Copolymer fractionation by gradient high-performance liquid chromatography. J Chromatogr. 384:135–44. 14. Gloeckner G, Wolf D and Engelhardt H (1991) Separation of copoly(styrene/ acrylonitrile) samples according to composition under reversed phase conditions. Chromatographia 32:107–12. 15. Gloeckner G, Wolf D and Engelhardt H (1994) Control of adsorption and solubility in gradient high performance liquid chromatography. Part 3. Sudden-transition gradient elution of styrene/acrylonitrile copolymers. Chromatographia 38:749-55. 16. Kuhn R (1983) Characterization of polymer blends, block copolymers, and graft copolymers by fractionation procedures using demixing solvents. Polym Sci Technol 20:45-58. 17. Lovric L and Res INA (1969) Fractionation of styrene-acrylonitrile copolymers. J Polym Sci, Part A-2 7:1357–66. 18. Teramachi S and Fukao T (1974) Cross fractionation of styrene-acrylonitrile copolymer. Polym J 6:532–6. 19. Cortes HJ, Jewett GL, Pfeiffer CD, Martin S and Smith C (1989) Multidimensional chromatography using on-line microcolumn liquid chromatography and pyrolysis gas chromatography for polymer characterization. Anal Chem 61:961–5. 20. Mori S (1996) Characterization of styrene-acrylonitrile copolymers by size exclusion chromatography/stepwise gradient elution-liquid precipitation chromatography. Int J Polym Anal Charact 2:185–92.

300

R. P. DION AND R. L SAMMLER

21. Ogawa T and Sakai M (1981) Column fractionation of acrylonitrile-styrene copolymers. J Polym Sci, Polym Phys Ed 19:1377–83. 22. Dealy JM and Wissbrun KF (1995) Melt Rheology and Its Role in Plastics Processing. Chapman and Hall, New York. 23. Agassant J-F, Avenas P, Sergent J-Ph and Carreau PJ (1991) Polymer Processing, Principles and Modeling. Hanser, New York. 24. Rosato DV and Rosato MG (2000) Injection Molding Handbook, 3rd edn. Kluwer, Dordrecht. 25. Rosato DV and Rosato DV (1988) Blow Molding Handbook. Hanser, New York. 26. Paul DR and Newman S (1978) Polymer Blends, Vols 1 and 2. Academic Press, New York. 27. Throne JL (1996) Technology of Thermoforming. Hanser, Munich. 28. Wilkinson AN and Ryan AJ (1998) Polymer Processing and Structure Development. Kluwer, Boston. 29. Bird RB, Armstrong RC and Hassager O (1977) Dynamics of Polymeric Liquids. Wiley, New York. 30. Ferry JD (1980) Viscoelastic Properties of Polymers, 3rd edn. Wiley, New York. 31. Middleman S (1968) The Flow of High Polymers. Interscience, New York. 32. Lodge AS (1964) Elastic Liquids. Academic Press, New York. 33. Sammler RL, Dion RP, Carriere CJ and Cohen A (1992) Compatibility of high polymers probed by interfacial tension. Rheol Acta 31:554–64. 34. Seitz JT (1993) The estimation of mechanical properties of polymers from molecular structure. J Appl Polym Sci 49:1331–51. 35. Berry GC and Fox TG (1968) The viscosity of polymers and their concentrated solutions. Adv Polym Sci 5:261–457. 36. Williams ML, Landel RF and Ferry JD (1955) The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. J Am Chem Soc 77:3701–7. 37. Van Krevelen DW and Hoftyzer PJ (1976) Newtonian shear viscosity of polymer melts. Angew Makromol Chem 52:101–9. 38. Lomellini P (1992) Williams-Landel-Ferry versus Arrhenius behaviour: polystyrene melt viscoelasticity revised. Polymer 33:4983-89. 39. Cross MM (1965) Rheology of non-newtonican fluids: a new flow equation for pseudoplastic systems. J Colloid Sci 20:417–37. 40. Cross MM (1969) Polymer rheology: influence of molecular weight and polydispersity. J Appl Polym Sci 13:765–74. 41. Shrager R (1970) Nonlinear regression with linear constraints: an extension of the magnified diagonal method. J Assoc Comp Machinery 17:446–52. 42. Press WH, Flannery BP, Teukolsky SA and Vetterling WT (1988) Numerical Redpies in C, Chapt. 14.4. Cambridge University Press, Cambridge. 43. Cox WP and Merz EH (1958) Correlation of dynamic and steady flow viscosities. J Polym Sci 28:619–22. 44. Takahashi H (1990) Flow behavior of polymer melts under extremely high shear rates and flow analysis of injection molding. J Rheol (Jpri) 18:172–9. 45. Takahashi HIY, Yamamoto S and Kamigaito O (1990) Flow curves of mixtures of acrylonitrile-styrene copolymers with different molecular weights at extremely high shear rates. J Rheol (Jpn) 18:125–8. 46. Rabinowitsch BZ (1929) Uber die viskositat und elastizitat von solen. Z Phys Chem AbtA 145:1–26. 47. Eisenschitz R, Rabinowitsch B and Weissenberg K (1929) Mitt Dtsch Mat PrufAnst Sanderheft 9:91.

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

301

48. Bagley EB (1957) End corrections in the capillary flow of polyethylene. J Appl phys 28:624–7. 49. Bagley EB (1961) The separation of elastic and viscous effects in polymer flow. Trans Soc Rheol 5:355-68. 50. Honerkamp J and Weese J (1989) Determination of the relaxation spectrum by a regularization method. Macromolecules 22:4372–7. 51. Baumgartel M and Winter HH (1989) Determination of discrete relaxation and retardation time spectra from dynamic mechanical data. Rheol Acta 28:511–9. 52. Baumgartel M and Winter HH (1992) Interrelation between continuous and discrete relaxation time spectra. J Non-Newtonian Fluid Mech 44:15-36. 53. Winter HH (1997) Analysis of dynamic mechanical data: inversion into a relaxation time spectrum and consistency check. J Non-Newtonian Fluid Mech 68:225-39. 54. Larson RG (1988) Constitutive Equations for Polymer Melts and Solutions. Butterworths, Boston. 55. Solovyov SE, Virkler TL and Scott CE (1999) Rheology of acrylonitrile—butadienestyrene polymer melts and viscoelastic constitutive models. J Rheol 43:977-90. 56. Otsuki Y, Kajiwara T and Funatsu K (1997) Numerical simulations of annular extrudate swell of polymer melts. Polym Eng Sci 37:1171–81. 57. Baaijens FPT (1998) Mixed finite element methods for viscoelastic flow analysis: a review. J Non-Newtonian Fluid Mech 79:361-85. 58. Kraft M, Meissner J and Kaschta J (1999) Linear viscoelastic characterization of polymer melts with long relaxation times. Macromolecules 32:751-7. 59. Munstedt H (1979) New universal extensional rheometer for polymer melts. Measurements on a polystyrene sample. J Rheol 23:421–36. 60. Meissner J and Hosttler J (1994) A new elongational rheometer for polymer melts and other highly viscoelastic fluids. Rheol Acta 33:1–21. 61. Meissner J (1996) Elongation of polymer melts—experimental methods and recent results. Proc XIIth Int Congr Rheol 1:1–10. 62. Schweizer T (2000) The uniaxial elongational rheometer RME - six years of experience. Rheol Acta 39:428–43. 63. Li L, Masuda T and Takahashi M (1990) Elongational flow behavior of ABS polymer melts. J Rheol 34:103–16. 64. Trouton FT (1906) On the coefficient of viscous traction and its relation to that of viscosity. Proc R. Soc London, Ser. A 77:426–40. 65. Meissner J, Stephenson SE, Demarmels A and Portmann P (1982) Multiaxial elongational flows of polymer melts - classification and experimental realization. J Non-Newtonian Fluid Mech 11:221-237. 66. Meissner J (1985) Rheometry of polymer melts. Annu Rev Fluid Mech 17:45–64. 67. Laun HM and Schuch H (1989) Transient elongation viscosities and drawability of polymer melts. J Rheol 33:119–75. 68. Meissner J (1971) Dehnumgsverhalten von Polyathylen-schmelzen. Rheol Acta 10:230–42. 69. Wagner MH, Bernnat A and Schulze V (1998) The rheology of the rheotens test. J Rheol 42:917–28. 70. Janeschitz-Kriegl H (1983) Polymer Melt Rheology and Flow Birefringence. Springer, New York. 71. Fuller GG (1995) Optical Rheometry of Complex Fluids. Oxford University Press, New York. 72. Kumaraswamy G, Verma RK and Kornfield JA (1999) Novel flow apparatus for investigating shear-enhanced crystallization and structure development in semicrystalline polymers. Rev Sci Instrum 70:2097-104.

302

R. P. DION AND R. L. SAMMLER

73. Kotaka T, Kojima A and Okamoto M (1997) Elongational flow opto-rheometry for polymer melts - 1. Construction of an elongational flow opto-rheometer and some preliminary results. Rheol Acta 36:646–56. 74. Venerus DC, Zhu S-H and Ottinger HC (1999) Stress and birefringence measurements during the uniaxial elongation of polystyrene melts. J Rheol 43:795–813. 75. Quinzani LM, Armstrong RC and Brown RA (1994) Birefringence and laser-Doppler velocimetry (LDV) studies of a viscoelastic fluid through a planar contraction. J Non-Newtonian Fluid Mech 52:1–36. 76. Baaijens HPW, Peters GWM, Baaijens FPT and Mejer HEH (1995) Viscoelastic flow past a confined cylinder of a polyisobutylene solution. J Rheol 39:1243–77. 77. Liang RF and Mackley MR (2001) The gas-assisted extrusion of molten polyethylene. J Rheol 45:211–26. 78. Kuhn R, Mueller HG, Bayer G, Kraemer-Lucas H, Kaiser W, Orth P, Eichenauer H and Ott KH (1993) Characterization of bimodal bigraft ABS. Colloid Polym Sci 271:13–42. 79. Dion RP and Billovits GF (1996) Interfacial tension: a quantitative measure of styrenic blend compatibility. Polym Prepr (Am Chem Soc, Div Polvm Chem) 32:529-30. 80. Dion R and Warakomski J (1993) Heat resistant styrenic polymer blends. US Patent 5212240. 81. Henton D, Dion R and Lefevre N (1990) Process for preparing copolymers of alphamethylstyrene and acrylonitrile. US Patent 4972032. 82. Utracki LA (1989) Polymer Alloys and Blends. Hanser, New York. 83. Hobbs SY, Dekkers MEJ and Watkins VH (1988) Effect of interfacial forces on polymer blend morphologies. Polymer 29:1598–602. 84. Wu S (1990) Chain structure, phase morphology, and toughness relationships in polymers and blends. Polym Eng Sci 27:335–43. 85. Molau G (1965) Heterogeneous polymer systems. III. Phase separation in styreneacrylonitrile copolymers. J Polym Sci, Polym Lett Ed 3:1007–15. 86. Koningsveld R and Kleintjens LA (1985) Liquid-liquid phase separation in multicomponent polymer systems 22. Thermodynamics of statistical copolymers. Macromolecules 18:243-52. 87. Vukovic R, Bogdanic G, Karasz FE and MacKnight WJ (1999) Phase behavior and miscibility in binary blends containing polymers and copolymers of styrene, of 2,6dimethyl-l,4–phenylene oxide and of their derivatives. J Phys Chem Ref Data 28:851–68. 88. Takahashi M, Li L and Masuda T (1989) Nonlinear viscoelasticity of ABS polymers in the molten state. J Rheol 33:709–23. 89. Wetton RE, Corish PJ (1980) DMTA studies of polymer blends and compatibility. Polym Test 8:303–12. 90. Gramespacher H and Meissner J (1992) Interfacial tension between polymer melts measured by shear oscillations of their blends. J Rheol 36:1127–41. 91. Gramespacher H and Meissner J (1995) Reversal of recovery direction during creep recovery of polymer blends. J Rheol 39:151–60. 92. Gramespacher H and Meissner J (1997) Melt elongation and recovery of polymer blends, morphology, and influence of interfacial tension. J Rheol 41:27–44. 93. Levitt L and Macosko CW (1997) Extensional rheometry of polymer multilayers: a sensitive probe of interfaces. J Rheol 41:671–85. 94. Chappelear DC (1964) Interfacial tension between molten polymers. ACS Polym Prepr 5:363-72.

FEATURES IMPACTING SAN COPOLYMER PERFORMANCE

303

95. Elemans PHM, Janssen JMH and Meijer HEH (1990) The breaking thread method: the measurement of interfacial tension in polymer systems. J Rheol 34:1311–22. 96. Ellingson PC, Strand DA, Cohen A, Sammler RL and Carriere C (1994) Molecular weight dependence of polystyrene/poly(methyl methacrylate) interfacial tension probed by imbedded-fiber retraction. Macromolecules 27:1643–7. 97. Anastasiadis SH, Gancarz I and Koberstein JT (1988) Interfacial tension of immiscible polymer blends: temperature and molecular weight dependence. Macromolecules 21:2980-7. 98. Cowie, JMG, Elexpuru EM and McEwen IJ (1992) Observation of restricted miscibility in binary blends of poly(styrene-stat-acrylonitrile) and poly(alphamethylstyrene-stat-acrylonitrile). Polymer 33:1993–5. 99. Warakomski, J and Dion R (1992) The effect of chemical composition on the miscibility of styrene/acrylonitrile/fumaronitrile terpolymers with styrene/acrylonitrile copolymers. J Appl Polym Sci 46:1057–63. 100. Hall W, Kruse R, Mendelson R and Trementozzi Q (1983) New styrene—maleic anhydride terpolymer blends. ACS Symp Ser 229:49-64. 101. Heinen W, Wenzel CB, Rosenmoeller CH, Mulder FM, Boender GJ, Lugtenburg J, de Groot HJM, Van Duin M, Klumperman B (1998) Solid-state NMR study of miscibility and phase separation in blends and semi-interpenetrating networks of 13 C-labeled poly(styrene-co-acrylonitrile) and poly(styrene-co-maleic anhydride). Macromolecules 31:7404–12. 102. Kourtesis G, Renwick GM, Fischer-Cripps AC and Swain MV (1997) Mechanical property characterization of a number of polymers using uniaxial compression and spherical tipped indentation tests. J Mater Sci 32:4493-500. 103. Sultan JN and McGarry FJ (1974) Tensile crazing and shear banding of styrene A. Temperature and rate effects. Polym Eng Sci 14:282–7. 104. Donald AM and Kramer EJ (1982) Plastic deformation mechanisms in poly(acrylonitrile-butadiene styrene) [ABS]. J Mater Sci 17:1765-72. 105. Jabarin SA (1991) Orientation and properties of acrylonitrile copolymers. Polvm Eng Sci 31:644–51. 106. Schellenberg J and Hamann B (1993) Influence of styrene-acrylonitrile oligomers on the properties of ABS graft copolymers. Eur Polym J 29:727–30. 107. Hasha DL, Priddy DB, Rudolf PR, Stark EJ, De Footer M and Van Damme F (1992) Oligomer formation and the mechanism of initiation in the spontaneous copolymerization of styrene and acrylonitrile. Macromolecules 25:3046–51. 108. Adao MHVC, Saramago BJV and Fernandes AC (1999) Estimation of the surface properties of styrene—acrylonitrile random copolymers from contact angle measurements. J Colloid Interface Sci 217:94—106. 109. Siebourg W, Schmid H, Rateike FM, Anders S and Lower H (1990) Birefringence An important property of plastic substrates for magneto-optical storage disks. Polym Eng Sci 30:1133–9. 110. ASTM D 4674–89 (1989) Accelerated testing for color stability of plastics exposed to indoor fluorescent lighting and window-filtered daylight. 111. ASTM D 1003-92 (1992) Haze and luminous transmittance of transparent plastics. 112. ASTM D 2457-90 (1990) Specular gloss of plastic films and solid plastics. 113. Nikolov ID and Ivano CD (2000) Optical plastic refractive measurements in the visible and the near-infrared regions. Appl Opt 39:2067-70.

This page intentionally left blank

14

Rubber Particle Formation in Mass ABS GILBERT BOUQUET The Dow Chemical Company, Midland, Ml, USA

1

MANUFACTURE OF ABS

The commercial production of acryloritrile—butadiene—styrene (ABS) formulations is accomplished by a number of different methods based on free radical polymerization. The two main methods are based on emulsion or solution polymerization techniques. The solution polymerization is mostly called mass or bulk polymerization because only a low amount of solvent is used. Most of the ABS (~85%) is made using the emulsion process. Both techniques have been used in combination (emulsion/mass). Other combinations are with suspension polymerization as final step (mass/suspension and emulsion/suspension) [1]. Although the emulsion process is commercially the most important, the mass process cannot be neglected because it has a number of advantages that will become clear from the more detailed description of both processes.

1.1

EMULSION PROCESS

Although more complicated than the mass process, emulsion polymerization is still widely used because of its greater flexibility [2,3]. The first step is the preparation of the rubber latex using emulsifiers. The crosslinking of the rubber occurs simultaneously during polymerization and is controlled by initiator level, chain transfer agent and process conditions. It is common to increase the particle size by agglomeration, thereby achieving a reduction in cycle time. Modern Stvrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy i(.-i 2003 John Wiley & Sons Ltd

306

G. BOUQUET

The next step is the polymerization of styrene and acrylonitrile in the presence of the rubber latex. Part of the polymerized styrene—acrylontrile is grafted on to the rubber. This grafted rubber concentrate is then either mixed with additional emulsion-prepared styrene-co-acrylonitrile (SAN) copolymer and then coagulated or first isolated and then compounded with SAN.

1.2

MASS PROCESS

In this process uncrosslinked rubber is dissolved in a mixture of the monomers and solvent(s). This solution is pumped into the first reactor which is connected to a series of reactors. The polymerization is started by increasing the temperature, eventually in the presence of an initiator. Most of the rubber grafting and particle sizing happen early in the process. Chain transfer agent level, initiator (type/amount) and shear have a great influence in this stage. Crosslinking of the rubber particles occurs later in the process. The final step is the removal of residual monomer and solvent. The advantages of the mass process are the absence of contaminants (emulsifiers) and a lower waste generation (no white water). Possible disadvantages are related to heat removal, viscosity and limitations in rubber particle sizing.

2

PHASE SEPARATION

When the homogeneous mixture of rubber in styrene—acrylonitrile—solvent mixture is heated, eventually in the presence of an initiator, SAN copolymer is formed. At about 2 % conversion the critical point is reached, meaning that phase separation has occurred. The first observation of opacity does not necessarily indicate the critical point, but often it is very close to it. It was reported as early as 1947 that a solution of two polymers in the same solvent, in most cases, separates into two phases [4]. This incompatibility is a direct result of the extremely small entropy of solution for high molecular weight polymer pairs. At the critical point there are two phases. The first is a continuous phase containing the rubber swollen in the monomer-solvent mix. The second is the dispersed or discontinuous phase consisting of SAN swollen in the monomersolvent mix. The system forms a polymeric oil-in-oil emulsion [5,6]. The rate of separation into two layers is greater the lower the temperature, even if the viscosity is higher. Kuhn and co-workers found that phase separation in a system close to its critical point can be reversed by shear and also by an increase in temperature [7–10].

RUBBER PARTICLE FORMATION IN MASS ABS

3

307

PHASE INVERSION

As the polymerization proceeds, more SAN is formed. This results in an increase in the discontinuous phase volume. At a certain degree of conversion the phase volume of each phase will be identical. If agitation is applied, phase inversion will occur. Thermodynamically, phase inversion is an instantaneous process; however, often there is a considerable delay in reaching the new equilibrium condition. The main reason is the necessity for significant mass transfer between the phases, which is retarded by the high viscosity. Furthermore, the graft copolymer has a stabilizing effect at the interface. The SAN phase will gradually become the continuous phase with dispersed rubber droplets. Although agitation is required for phase inversion to occur, it is not the cause of it. 4

PHASE DIAGRAM

The polymerization sequence can be plotted using a phase diagram (Figure 14.1) [11]. This diagram is representative for nearly all polymer (rubber)polymer (glassy polymer)-common solvent systems. Point a is the composition of the feed. The critical point (phase separation) is reached at point b. Further polymerization generates more glassy polymer (SAN) and phase inversion occurs at point c (rubber continuous —> discontinuous). Point d is reached when all monomer is converted. Monomer of Glassy Polymer

Rubber

Glassy Polymer

Figure 14.1 Phase diagram for rubber—glassy polymer-monomer system [11] a, Initial composition; b, phase separation; c, phase inversion; d, complete conversion. Reproduced from W. A. Ludwico and S. L. Rosen, J.Appl.Polym.Sci., 19, 757 (1975) with permission of John Wiley & Sons, Inc.

308

5

G. BOUQUET

RUBBER PARTICLE SIZING

After phase inversion, the dispersed phase is the rubber phase. This is the first time discrete rubber particles are present in the reaction mixture. The particle size in the final product is an important parameter to optimize the physical properties. To be successful in the manufacturing of mass ABS, it is necessary to understand and control which parameters can be used to control the final rubber particle size. The following parameters will be discussed: • shear; • viscosity; • interfacial tension. 5.1

SHEAR

Spontaneous phase inversion (no shear) has been described by Keskkula [12]. This was demonstrated during the quiescent polymerization of styrene-polybutadiene mixtures containing less than 3 wt% polybutadiene. For industrially important systems (higher rubber content), a minimum amount of shear is required [13]. If no adequate agitation is applied, the system will solidify in the emulsion state before the inversion point. The final product will then consist of a continuous phase of a crosslinked polybutadiene network with dispersed SAN particles. Such a material will not have the typical properties of ABS. A level of agitation above the minimum shear can be used to reduce the rubber particle size. However, agitation is not always effective in reducing the rubber particle size (see section 5.2) and there is also an upper limit based on hardware limitations (torque). Another aspect that plays a role related to shear is the feed rate. Increasing the feed rate means that the residence time in the reactor(s) is shorter. A lower amount of shear is transferred to the prepolymer, giving in general larger rubber particles in the end product. 5.2

VISCOSITY

Two factors related to viscosity play a role in the rubber sizing process: • ratio (p) of disperse phase viscosity (//) and continuous phase viscosity (/*); • viscosity of the continuous phase. Rumscheidt and Mason [14] described particle deformation in a shear field as a function of viscosity ratio (p). There is a minimum and a maximum viscosity ratio where it becomes impossible to reduce the droplet size. The limits described by Karam and Bellinger [15] are 0.005 and 4. Breakup of droplets readily occurs when the viscosity ratio is of the order of 0.2:1. Intuitively, a ratio of 1 would be best because in this case there is a maximum transfer of energy between the

309

RUBBER PARTICLE FORMATION IN MASS ABS

continuous and dispersed phase. This assumption is probably only valid if the interfacial tension is nearly zero and there is no viscous or rigid film at the interface. It was also found that the higher the viscosity of the continuous phase, the greater the ease of particle breakup. Both effects are illustrated in Figure 14.2. The rubber (dispersed) phase viscosity is determined by the rubber level and by the solution viscosity of the rubber. Furthermore, the grafting and crosslinking will also influence the viscosity. The SAN (continuous) phase viscosity is controlled by the molecular weight of the copolymer. With standard rubbers, the viscosity ratio tends to be higher than 1. This is not the optimum and to reduce it the rubber viscosity can be decreased or the SAN viscosity can be increased. Decreasing the rubber viscosity can be achieved by using low solution viscosity rubbers. These rubbers typically are star branched rubbers or block copolymers (styrene—butadiene block). Branching or introduction of a polystyrene block is done to control the cold flow of these materials to avoid problems during storage. Table 14.1 contains a short overview of typical rubbers that are used in mass ABS production. The polybutadiene chain contains a lot of double bounds (allylic hydrogens) that will react with radicals during the copolymerization. Possible reactions are grafting (see Section 5.3) and crosslinking. Both reactions, but especially crosslinking have an important 140

POLYGLYCOL-VEHICLE DC-200 DROPLET

120

RADIUS OF DROPLET 0.08 cm //-0.80 Poise

100

SHEAR RATE AT BREAK UP, Gb, s-1

/x-2.4 Poise

VISCOSITY RATIO, //'/// Figure 14.2 Influence of viscosity ratio and continuous phase viscosity (/i) on drop breakup [15]. Reprinted with permission from H. J. Karam and J. C. Bellinger, Ind. Eng. Chem. Fundam., 7, 575 (1968). Copyright 1968 American Chemical Society

310

G. BOUQUET

Table 14.1 Rubbers used in mass ABS process Producer Type

Solution A/w viscosity (mPas) (g/mol) Mw/M/n Type

Bayer Buna CB HX528 Bayer Buna CB HX565 Firestone Stereon 730

160 45 25

260000 205000 140000

1.4 1.2 1.3

Composition

Linear Polybutadiene Branched Polybutadiene Block 30%Styrene

influence on the rubber molecular weight, and hence viscosity. Crosslinking and the consequences it has on the rubber sizing process will be discussed later. To increase the SAN viscosity, the molecular weight generally has to be increased. Tools are acrylonitrile content, polymerization temperature, initiator (type/concentration), solvent level and chain transfer agent (amount and timing of addition). A higher acrylonitrile content also results in a higher matrix viscosity. Increasing polymerization temperature gives lower molecular weight. Increasing the concentration of monofunctional initiators also results in lower molecular weight. Increasing the functionality of the initiator increases the matrix molecular weight. Often the solvent used has some chain transfer activity, hence an increase in solvent level will give lower molecular weight. Furthermore, the solvent and unreacted comonomers are recycled. During the manufacturing process, side reactions occur, generating products that build up in the recycle. These products often show chain transfer activity. 5.3

INTERFACIAL TENSION

During the sizing process, the rubber phase is becoming increasingly finely dispersed in the SAN matrix. During this process, the surface area is increased. This process requires less energy when the interfacial tension is low. A reduction of the interfacial tension can be achieved by: • utilization of block rubbers; • grafting. Utilization of block rubbers is being successfully used for the production of high-impact polystyrene (HIPS). The styrene—butadiene block rubbers act as emulsifier for the polystyrene matrix and the rubber particles. Echte has extensively described the possible morphologies (cellular, core-shell, rods, droplets, etc.) that can be obtained [16,17]. The same principle cannot be used in ABS because the required block rubbers (SAN—butadiene) are not commercially available. The required SAN-polybutadiene copolymers can be generated in situ during the copolymerization of SAN. This grafting process will be discussed in detail in the next section because grafting is considered a major tool to control rubber particle morphology and thus the physical properties of mass ABS.

RUBBER PARTICLE FORMATION IN MASS ABS

6

311

GRAFTING

The mechanisms involved in grafting and the associated kinetics have not been completely resolved. Many articles have been published related to this topic, each describing specific aspects of the process [18–25]. Polybutadiene contains two reactive groups: double bonds and allylic hydrogens. A radical can add to the double bond, thereby creating a radical on the rubber backbone. If an allylic hydrogen is abstracted, again a radical is created on the backbone but the original radical does not become incorporated as is the case if addition occurs. Once a radical is present on the rubber, the addition of monomer units can start. During this propagation step the graft is created nearly instantaneously according to free radical kinetics. 6.1

GRAFT ANALYSIS

One of the problems in the study of the grafting reaction is the separation and quantification of the different species present in the prepolymer. The three polymeric components present in the prepolymer are: • free SAN; • free rubber; • graft rubber. The analysis of the prepolymer at high conversion and end product is even more problematic because additional crosslinking has occurred. Hughes developed a simple test based on the 'emulsifying' power of an artificial prepolymer containing the copolymer to be tested [26]. A more elaborate technique technique based on extraction was developed by Llauro [27] and Riess [28] based on the reversible crosslinking of the rubber (free + graft). This reversible crosslinking was achieved by attaching -COONa groups to the rubber. The dipole—dipole interaction between the -COONa groups in nonpolar solvents gives gelation of the rubber phase. The free SAN remains soluble and can be removed. Adding methanol breaks up the gel. Huang and Sundberg have described a method based on gel permeation chromatography using a dual detector set-up [29]. However, this method can only be used when the rubber and matrix molecular weights are different. Furthermore, block rubbers cannot be used. Bouquet et al. have developed a method to separate and measure the different fractions, but the actual technique is not disclosed [30]. 6.2 6.2.1

EFFECT OF PROCESS

PARAMETERS

Nomenclature

Graft efficiency = ratio of graft SAN to total SAN

312

G. BOUQUET

Yield graft rubber = ratio of graft rubber to total rubber Graft density = average number of graft chains per grafted rubber molecule

6.2.2

Monomer

Locatelli and Riess showed that the molecular weight of the SAN (free and graft) increased as the monomer concentration was raised [31]. This is in agreement with classical free radical kinetics. Increasing the monomer concentration results in a lower graft efficiency, but the graft density is higher [32]. At higher monomer concentration it is possible that an increasing amount of free SAN is formed in the rubber phase, explaining why the graft efficiency is lowered.

6.2.3

Initiator

Some initiators are more effective in creating radicals on the rubber chain compared with others. Initiators yielding phenyl, benzoyloxy and tert-butoxy radicals are described as very efficient for grafting. A possible explanation is that this type of radical can stabilize the negative charge present in the transition state during hydrogen abstraction. According to Allen et al., benzoyloxy radicals preferentially add to the double bond [24]. The phenyl radicals formed from the benzoyloxy radicals by CO2 loss are more reactive towards allylic hydrogens. 2,2'-Azobis(isobutyronitrile) (AIBN), which generates tertiary carbon radicals, reacts only with 1,2-vinyl units according to Locatelli and Riess [24]. The higher reactivity of the 1,2-vinyl groups is caused by the labile tertiary allylic hydrogens. Hayes and Futamara [25] concluded that AIBN and benzoyl peroxide only generate grafting through copolymerization. None of the graft copolymer is generated by hydrogen abstraction. They consider the reaction of polybutadiene with styrene and acrylonitrile as a true terpolymerization.

6.2.4

Rubber

Riess and Locatelli studied the influence of the 1,2-vinyl content of the elastomer on the overall polymerization rate [33]. The rate is increased as the 1,2-vinyl content of the rubber is higher. The presence of 1,2-vinyl units automatically means that allylic tertiary hydrogens are present, which are easily abstractable by a radical. This gives an increase in the radical sites on the polybutadiene backbone, thereby indirectly boosting the initiator efficiency. Supporting the previous statement, Locatelli and Riess showed that the graft efficiency increased as the 1,2-vinyl content of the rubber is higher [32]. The polystyrene fragment in a block rubber is considered inert for the grafting reaction.

RUBBER PARTICLE FORMATION IN MASS ABS

313

The effect of rubber concentration on grafting was also studied [31]. Adding more rubber resulted in an increase in the graft SAN molecular weight. The explanation given is based on the Trommsdorf effect. A higher rubber concentration means that the rubbery phase becomes more viscous, thereby reducing the termination rate, hence higher molecular weights are obtained. When benzoyl peroxide was used, the graft efficiency increased as rubber concentration was higher. This was not the case when AIBN was used [32]. The higher the rubber concentration, the higher is the probability that a graft active site will be created. When AIBN is used, the grafting is minimal and seems to be limited to a plateau value.

6.3

MASTER CURVE

To understand the grafting process better and to be able to control product properties, it is important to know how fast the rubber is converted to grafted rubber. The yield of graft rubber is a measure of this. If the yield of graft rubber is zero, none of the rubber molecules are grafted. If the yield of graft rubber is 1, then all the rubber chains are grafted with at least one SAN graft. The yield of graft rubber has been determined for a number of initiators at different temperatures [30]. In both cases the observed trend was very similar. At the beginning of the copolymerization the yield increases very fast, reaching a plateau around a solids level of 25 %. A higher yield is obtained when the initiator concentration is increased and when the polymerization temperature is higher. The ideal case would be to combine this information, making it possible to determine the yield under any conditions (initiator type/concentration/temperature). This has been attempted based on the idea that mainly the primary radicals initiate the grafting reaction. The initiator decay determines the rate of primary radical formation. The yield at 25 % solids (plateau) is plotted versus the mean of the initiator decay rate during the timeframe to obtain 25 % solids. This is called the master curve. All the data points for one initiator fall on the same line, which is an indication that primary radical formation indeed plays a dominant role in the grafting process.

6.4

GRAFT MODEL [30]

A graft model has been proposed to explain why it is practically impossible to graft all the rubber during the manufacturing of ABS. With a high amount of initiator (l000ppm benzoyl peroxide), only 70% of the rubber molecules are grafted. For AIBN-initiated runs, the yield even drops to 0.3 (70% of the rubber is not grafted). This is unexpected because the rubber used (Bayer Buna CB HX528) contains 14500 active sites (double bonds/allylic hydrogen)

314

G. BOUQUET

per molecule and per rubber molecule 16 primary radicals are in theory available to initiate the grafting. Brydon et al. performed experiments in HIPS under high dilution and showed that a yield close to the theoretical maximum of 1 is obtainable [34]. A partial explanation for the shielding of the free rubber is offered by Rosen [35]. The reacting system places an upper limit on the amount of rubber that can be grafted, regardless of the chemical nature of the chemistry involved in the grafting process. The limit is caused by the inherent compatibility of the two polymeric phases. If the initiator decomposes in the SAN phase, the primary radical will never be able to initiate the graft reaction because there is no rubber in its proximity. Only initiator that decomposes in the rubber phase has the potential to start the graft reaction. This means that when the polymerization is done in a one-phase system, theoretically all the rubber can be grafted (high dilution). Based on this theory it can be concluded that most of the grafting takes place in the early stages of the polymerization (small SAN phase versus rubber phase). The free rubber is also protected by a chemical restriction. The only species that can create a radical on the rubber backbone is the primary radical. An initiator molecule, present in the rubber phase, decomposes in a solvent cage, and in the case of ABS, the cage contains solvent, styrene, acrylonitrile and eventually chain transfer agent. The primary radical has to escape from the cage to reach a polybutadiene fragment before it reacts with the surrounding monomers or chain transfer agent, which are very efficient radical scavengers. Combining the physical and chemical restrictions imposed upon the grafting process can explain why it is impossible to graft all the rubber. This model is summarized in Figure 14.3.

7

CROSSLINKING

During the grafting process, a radical site (addition/abstraction) is created on the rubber backbone. Propagation of styrene—acrylonitrile results in the SAN graft. The final step in the grafting sequence is the termination of the radical site. Figure 14.4 gives an overview of the different possibilities. Chain transfer and disproportionation result in a grafted rubber with 'normal' SAN molecular weight. Termination by combination with a growing SAN chain also gives a grafted rubber but in this case the molecular weight of the graft is higher (double). A final possibility is combination with another growing graft. The result in this case is a crosslink between two rubber chains. This type of termination increases the molecular weight dramatically and will be reflected in the viscosity of the rubber phase. The viscosity of the rubber phase has a large influence on the sizing process that takes place during the production of ABS. As the viscosity ratio (rubber/SAN) increases, sizing will become more difficult.

315

RUBBER PARTICLE FORMATION IN MASS ABS

Figure 14.3

Model for grafting process showing the physical and chemical restrictions

Termination by transfer or disproportionation

rubber backbone SAN graft

Termination by combination with SAN

Termination by combination with graft 'pre-crosslinking' Figure 14.4

Termination processes during grafting

The most common way to measure crosslinking is by swelling. For systems containing only rubber, the Flory—Rehner equation can be applied. For ABS that is not the case because grafted and occluded SAN will interfere with the swelling behavior. Nevertheless, this method is being used frequently for practical reasons. Karam and Tien have developed a quantitative analysis for the swelling behavior of a heterogeneous gel [36]. They described quantitatively the crosslink density when the proportion of occlusions is known. DMA and DSC techniques are typically insufficiently sensitive to detect small changes in crosslink density. Relaxation experiments based in 13C NMR have been performed but is was difficult to obtain absolute data using this technique [37]. The

316

G. BOUQUET

physical properties of ABS are greatly influenced by the degree of crosslinking. Optimization of crosslinking to control product properties has been described in the patent literature [38,39]. Peng has studied the kinetics of the crosslinking process for HIPS in the absence of initiator [40]. In the absence of styrene and oxygen, polybutadiene cannot be crosslinked by heating alone. For HIPS most of the crosslinking takes place at high conversion. In the presence of acrylonitrile (ABS), the onset of crosslinking starts earlier in the polymerization.

8

SIZING WINDOW

The sizing window is defined as the period during the manufacturing of mass ABS where the rubber particle size can be controlled. The sizing window starts at the point of phase inversion. When the required minimum amount of shear is applied, the rubber phase will change from continuous to discontinuous phase. At this point the rubber particles are very large and the particle size distribution is very broad. The rubber particles will become smaller if shear is applied at the right viscosity ratio. The presence of graft (emulsifier) will facilitate the sizing process. At a certain point in time the rubber phase viscosity will increase because the monomer concentration is reduced through copolymerization and crosslinking becomes important. If the viscosity becomes too high, the sizing will cease even if shear is applied. The period between phase inversion and an excessive viscosity ratio is called the sizing window. Only shear applied in this region will be efficiently used to reduce the particle size and make the particle size distribution narrower. If small particles are required it is important to apply high shear where the viscosity ratio is optimal (0.2-1) and interfacial tension is reduced by grafting. Related to particle sizing, Molau and Kesskula described the concept of type I and II occlusion [5]. The prepolymer is viscous and has a retarding effect on the phase inversion. In most cases multiple emulsions are formed after the phase inversion point. If the agitation is not extremely high these multiple emulsions survive the further copolymerization and give SAN occlusions in the rubber particles. These occlusions are called type I. Type II occlusions are formed when monomer dissolved in the rubber phase is copolymerized. Because SAN is not compatible with the rubber, separation occurs within the rubber particle, giving type II occlusions.

317

RUBBER PARTICLE FORMATION IN MASS ABS 9

RUBBER PARTICLE MORPHOLOGY

In HIPS a wide variety of rubber particle morphology is possible. Echte et al. summarized this in an excellent review (see Figure 14.5) [41]. The key to these different structures is the composition of the styrene—butadiene block rubber, which is an emulsifier for the polystyrene—polybutadiene system. Additional grafting can generate a shift from one structure to another. In the case of mass ABS, the variety of rubber particle morphology is less diverse. Typical examples of morphology are shown in Figure 14.6. If polybutadiene rubber is used (linear or star), cellular particles are obtained with SAN occlusions. In the case of styrene—butadiene block rubber (typically 30% styrene) also cellular particles are obtained but besides the SAN occlusions, polystyrene domains are clearly visible in the particles. To be able to make the other morphologies that are possible in HIPS, the interfacial tension has to be manipulated. Controlling the grafting reaction is a way to achieve this but the possibilities are limited with the tools (mainly initiator) that are currently available.


if

Droplets

•j

V *>%?*'• Ml

Rods

Capsules

.

-

. ft

Rod clusters

Figure 14.5 Possible rubber particle morphology in HIPS

Droplet clusters

318

Figure 14.6

G. BOUQUET

Rubber particle morphology in mass ABS as function of rubber type

REFERENCES 1. Adams ME, Buckley DJ, Colborn RE, England WP, Schissel DN (1993) Acrylonitrile-Butadiene-Styrene Polymers. Rapra Review Reports, Vol. 6, No. 10. 2. Calvert W, US Patent 3 238 275 (to Borg Warner). 3. Bovey FA, Kolthoff IM, Medalia IM, Meehan EJ (1955) Emulsion Polymerization, Interscience, New York. 4. Dobry A, Boyer-Kawenoki F (1947) J. Polym. Sci. 2, 90. 5. Molau GE, Keskkula H (1966) J. Polym. Sci. A-l 4, 1595. 6. White JL, Patel RD (1975) J. Appl. Polym. Sci. 19, 1775. 7. Silberberg A, Kuhn W (1952) Nature (London) 170, 450. 8. Silberberg A, Kuhn W (1954) J. Polym. Sci. 13, 21. 9. Burkhardt F, Majer H, Kuhn W (1960) Helv. Chim. Acta 43, 1192. 10. Kuhn W, Majer H, Burkhardt F (1960) Helv. Chim. Acta 43, 1208. 11. Ludwico WA, Rosen SL (1975) J. Appl. Polym. Sci. 19, 757. 12. Keskkula H (1979) Plasti. Rubb.: Mater. Appl. 16, 71. 13. Freeguard GF, Karmarkar MJ (1957) J. Appl. Polym. Sci. 15, 1657. 14. Rumscheidt FD, Mason SG (1961) J. Colloid Sci. 16, 238. 15. Karam HJ, Bellinger JC (1968) Ind. Eng. Chem. Fundam. 7, 576. 16. Echte A, (1977) Angew. Makromol. Chem. 58/59, 175. 17. Echte A, Haaf F, Hambrecht J (1981) Angew. Chem., Int. Ed. Engl. 20, 44.

RUBBER PARTICLE FORMATION IN MASS ABS 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

319

Huang NJ, Sundberg DC (1995) J. Polym. Sci. A: Polym. Chem. 33, 2533. Huang NJ, Sundberg DC (1995) J. Polym. Sci. A: Polym. Chem. 33, 2551. Huang NJ, Sundberg DC (1995) J. Polym. Sci. A: Polym. Chem. 33, 2571. Huang NJ, Sundberg DC (1995) J. Polym. Sci. A: Polym. Chem. 33, 2587. Walling C, McElhill EA (1951) J. Am. Chem. Soc. 73, 2979. Locatelli JL, Riess G (1973) Angew. Makromol. Chem. 32, 161. Allen PW, Ayrey G, Moore CG (1959) J. Polym. Sci. 36, 55. Hayes RA, Futamura S (1981) J. Polym. Sci. Polym. Chem. Ed. 19, 985. Hughes LI, Brown GL (1963) J. Appl. Polym. Sci. 7, 59. Llauro MF (1970) Thesis, Ecole Superieure de Chimie, Mulhouse. Riess G, Locatelli JL (1975) Adv. Chem. Ser. 142, 186. Huang NJ, Sundberg DC (1994) Polymer 35, 5693. Bouquet G, Kentie WC, De Theije PJG, Van Damme F (1996) Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 37, 536. Locatelli JL, Riess G (1972) Angew. Makromol. Chem. 28, 161. Locatelli JL, Riess G (1973) Angew. Makromol. Chem. 32, 117. Riess G, Locatelli JL (1973) Angew. Makromol. Chem. 32, 101. Brydon A, Burnett GM, Cameron GC (1973) J. Polvm. Sci., Polym. Chem. Ed. 11, 3255. Rosen SL (1973) /. Appl. Polym. Sci. 17, 1805. Karam HJ, Tien L (1985) J. Appl. Polym. Sci. 30, 1969. Curran SA, Padwa AR (1987) Macromolecules 20, 625. Vanspeybroeck RS, Galobardes MR, Maes D, Jones MA, Ceraso JM (2001) US Patent 6/211/298 (to The Dow Chemical Company). Iwamoto M, Nakajima A, Takaku M, Morita H, Ando T, Shirafuji T, Uchida M (1996) US Patent 5/552/494 (to Mitsui Toatsu Chemicals). Peng FM (1990) J. Appl. Polym. Sci. 40, 1289. Echte A, Haaf F, Hambrecht J (1981) Angew. Chem., Int. Ed. Eng. 20, 344.

This page intentionally left blank

15

High Heat Resistant ABS Technology RONY VANSPEYBROECK, ROBERT P. DION, AND JOSEPH M. CERASO The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Conventional acrylonitrile—butadiene—styrene terpolymer (ABS) is a thermoplastic polymer, consisting of a rigid styrene-co-acrylonitrile (SAN) continuous phase and a dispersed rubber phase [1,2]. Some SAN copolymer is chemically bonded (grafted) to the rubber, which is usually polybutadiene or a styrene—butadiene copolymer. The rubber reinforcement of SAN increases its ability to withstand high speed impact at the expense of most other properties [3,4]. Compared with rubber-modified high-impact polystyrene (HIPS), the modification of the polystyrene continuous phase by copolymerization with acrylonitrile results in resins that have higher tensile strength, higher toughness, better solvent resistance and improved heat resistance, while maintaining excellent processability [5,6]. The typical physical properties of ABS are affected by the rubber and the acrylonitrile content, the rubber particle size, rubber particle size distribution, rubber morphology, and the rigid phase molecular weight (both grafted and free). Table 15.1 shows typical values for the properties of selected SAN, HIPS and ABS. The heat resistance performance of a resin is the temperature at which a part will start to be dimensionally unstable and to distort under load. There are various laboratory tests to predict this performance. The most commonly used are the measurement of the Vicat softening point (ASTM D1525, DIN 53460, ISO 306) and the deflection temperature under load (DTUL, ASTM D648). Both measurements monitor the modulus change with temperature, and determine an endpoint when a macroscopic change can Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy C) 2003 John Wiley & Sons Ltd

322 Table 15.1

R. VANSPEYBROECK ETAL. Properties of SAN copolymers, ABS and HIPS

Vicat softening point ( °C) Tensile yield (MPa) Tensile modulus (MPa) Notched Charpy impact (kJ/m2) Melt flow rate (cm3/10/min)

SANb

ABSC

HIPSd

103 63

101 48

88 25

3400

2300

1770

2 25

24 16

18 4.0'

a

Vicat: ISO 306; 50 N/ 50°C/hr. Tensile properties: ISO 527-1,2. Charpy: ISO 179. MFR:ISO1133;220°C, 10kg. * TYRIL® 990 (The Dow Chemical Company). c MAGNUM® 9020 (The Dow Chemical Company). d STYRON® A-TECH 1110 (The Dow Chemical Company). e HIPS @ 200 °C, 5kg.

be detected in a part under load. The Vicat is the temperature at which a flatended needle penetrates a test sample to a depth of 1 mm under a particular load and uniform heating rate, while the DTUL is the temperature at which a test specimen on edge in an oil bath deflects 0.025 cm under a stress of 0.46 or 1.82 MPa (66 or 264 psi) at a heating rate of 2 °C/min. The DTUL test result is influenced by the sample thickness, the molding conditions, and whether or not the sample has been annealed. Since these tests are influenced by both fabrication conditions and part geometry, additional evaluations are required for final qualification in an application. Often, molded parts are placed in an oven at a specified heat condition and the parts are measured for dimensional changes. The heat resistance performance strongly correlates with the glass transition temperature (7g) and the stiffness (modulus) of the resin. In this two-phase system, the Tg of the rigid SAN continuous phase is modestly influenced by the acrylonitrile content and increases approximately 0.3–0.4 °C per weight percent of added acrylonitrile. Within the range of commercial ABS resins, the acrylonitrile content is typically 20–30 wt%. The molecular weight of the SAN continuous phase is very important for determining the toughness and flow balance of the resultant ABS, but has no significant influence on the heat resistance. However, low molecular weight SAN oligomers, residual monomers, or additives such as flow promoters that act as plasticizers all have a negative effect on the heat resistance. The thermal properties are also influenced by the rubber phase. A higher rubber content or a higher rubber phase volume results in a lower modulus. The heat performance of conventional ABS correlates in general with the glass transition temperature (Tg) of the rigid phase. Table 15.2 lists some typical rgs of amorphous polymers. Also listed are the crystalline melting points (r m ) for semi-crystalline polymers. Typically, the heat performance of a neat semicrystalline polymer under low load correlates with its Tm.

HIGH HEAT RESISTANT ABS TECHNOLOGY Table 15.2 tures

323

Typical polymer glass transition and semi-crystalline melting tempera-

Amorphous

~r g (°C)

Semi -crystalline

T*

PS HIPS SAN (30%AN) ABS HH ABS PC

100 100 108 105 110-120 151

Syndiotactic-PS Nylon 6,6 PBT Acetal

270 265 222 175

Tm f( O/*"~*\ C)

Although syndiotactic polystyrene (SPS) consists of pure polystyrene, the tacticity has a dramatic effect on chain dynamics and causes crystallization to occur. This will be examined in a separate chapter. High heat ABS resins described in this chapter refer to a family of multiphase polymers which are dimensionally stable at temperatures where conventional, general-purpose ABS deforms. The maximum end-use temperature at which fabricated parts can meet the functional requirements of the specific application determines whether high heat resistant ABS or conventional ABS will be used. Typical applications for high heat ABS are automotive interior components that can be exposed to temperatures of more than 90 °C. Conventional ABS warps unacceptably in that environment. The heat resistance of ABS can be improved by adding a high-Tg polymer, by chemically modifying the SAN copolymer, or by removing low molecular weight plasticizers. This review will not cover the multiphase blend approach, exemplified by extruded mixtures of polycarbonate and ABS. The focus will concentrate on the chemical modification of ABS by the use of a different, continuous-phase styrenic copolymer. This review describes technical solutions for high heat resistance which make use of the incorporation of new monomers to increase the Tg of the continuous phase through chain stiffening and/or modification of the cohesive energy density. The new monomers can be added directly through terpolymerization with styrene and/or acrylonitrile. Alternatively, they can be copolymerized with styrene and post-blended with conventional high rubber content ABS impact modifiers. Some of the process chemistry influence on residuals will be briefly discussed for tx-methylstyrene termonomer high heat ABS. An overview of the product technologies is also described. The processes are designated but their description will be reviewed in other chapters. The most commonly used commercial monomers that increase the Tg of ABS are maleimides (most recent commercial addition), maleic anhydrides (MA), and substituted styrenes. The majority of ABS produced in North America and Europe relies on a-methylstyrene (aMeS) or N-phenylmaleimide (PMI) as

324

R. VANSPEYBROECK ETAL

termonomers to raise the ToR of the continuous *phase. A small fraction of the ABS market contains MA.

2 SUBSTITUTED STYRENES Table 15.3 lists a series of substituted styrenes that can be used to produce more heat-resistant polymers [7]. The substituent results in a higher rotational energy of the C—C bonds in the polymer backbone and consequently a stiffer polymer chain and higher Tg. The most widely practiced technology uses a-methylstyrene. This is mainly due to its favorable price/performance balance. ot-Methylstyrene is produced by cumene oxidation and occurs as a byproduct in the manufacture of phenol and acetone (Figure 15.1). About 40% of the monomer is used to produce heat-resistant ABS. For styrenic polymers containing 20 wt% of AN, the glass transition temperature increases by about 0.4 °C for every 1 % of styrene that is replaced with otMeS. a-Methylstyrene can be incorporated into the continuous ABS phase using the known processes for conventional ABS, including emulsion [8-12], mass [13], suspension [14], and mass/suspension [15]. However, the incorporation of aMeS into SAN is complicated by two factors, which determine the limitations on the use of aMeS for producing heat-resistant ABS at the rapid rates desirable in a manufacturing environment. The reactivity ratios of aMeS and AN are r\ =0.07 and r2 = 0.15, respectively, which results in the polymer tending toward alternating compositions. The azeotropic composition is approximately 50 mol% (70 wt%) aMeS. As one attempts to produce copolymers with very high aMeS content, the polymerization rate decreases. As the ceiling temperature for polymerization of aMeS is 61 °C [16,17], high-temperature polymerization of copolymers containing high levels of aMeS proceeds slowly and favors the production of low molecular weight or even oligomeric species, which results in poor toughness and heat properties. Table 15.3

Glass transition temperatures of substituted styrene homopolymers

Position

Substituent

— 4 2 2,5 4 a

— tert- Butyl Methyl Dimethyl Phenyl Methyl

Tg

("C) 100 128 136 143 161 168

325

HIGH HEAT RESISTANT ABS TECHNOLOGY CH7

CH

Figure 15.1

H2SO4

Phenol-acetone process to produce a-methylstyrene

The consequences of the aMeS kinetics are that a complete mass ABS product is possible but is limited in the heat resistance attainable. At the applied hightemperature range (100–160°C), polymerization rates are slow and oligomer formation is high if the weight percent of acrylonitrile in the monomer feed is too low [14]. The high levels of oligomeric species plasticize the ABS and greatly reduce its heat resistance, decreasing the benefit of incorporating aMeS. The aMeS incorporated in commercial mass ABS resins is as high as 45 wt%. Most of the ABS producers around the world produce HHABS with aMeS technology. Owing to the composition drift and the need for economical rates of reaction, a continuous stirred tank reactor (CSTR) configuration is required. The reaction is run off azeotrope with high acrylonitrile feed around 50 wt% to enhance the rate of polymerization. An approximately 30:70 AN—aMeS copolymer rigid phase is formed by this process. Some of the HHABS is still manufactured by utilizing 100 % of the emulsion polymerization process. After the polymerization of the polybutadiene latex, aMeS is added along with styrene and acrylonitrile to form the grafted and nongrafted rigid phase. Most of the commercial grades consist of an emulsion SAN grafted polybutadiene containing impact modifier blended with a solution polymerized a-methylestyrene-co-acrylonitrile copolymer. In this approach, aMeS-AN copolymer with up to 30wt% AN is produced in a mass (solution) process. Impact modification is achieved by compounding in an emulsion SAN-grafted rubber concentrate. The aMeS—AN copolymer manufacturing rates are significantly lower than the rate of manufacture for SAN copolymer. aMeS—AN copolymers are not miscible with SAN copolymers [18], but they are sufficiently compatible to produce heat-resistant ABS blends with acceptable properties. Replacing part of the styrene by ot-MeS results in S—aMeS—AN terpolymers. The slower aMeS kinetics result in a rate reduction compared with generalpurpose emulsion ABS. A typical emulsion process for S—aMeS—AN terpolymers is described in a patent [12].

326

R. VANSPEYBROECK ET AL.

Separate manufacturing and blending of aMeS–AN copolymers, S–AN copolymers and grafted rubber concentrates is the oldest way for obtaining high heat ABS and it is practiced by a large number of manufacturers [9–12]. By blending the components in different ratios, families of resins can be designed which vary in heat resistance, melt flow, and toughness. Although SAN and aMeS–AN are generally not miscible, they are sufficiently compatible to produce heat resistant ABS with acceptable property balances. In addition to the traditional free radical polymerization techniques, anionic polymerization at high temperature (100 °C) has been described for S–aMeS copolymers containing up to 67 wt% aMeS [19]. The anionic synthesis of S–otMeS copolymers is not plagued by the low ceiling temperature of 61 °C for radical polymerization. The use of aMeS in polymers of PS and HIPS has not been commercially viable. Much of the higher heat-resistant ABS grades commercially available are produced using emulsion polymerization technology and have high gloss esthetics. When low gloss esthetics are required, such as in many of the automotive interior applications, mass polymerization technology has been utilized. Low gloss applications require larger rubber particles that are most easily produced by a mass polymerization process. Low gloss high heat ABS containing a-methylstyrene or N-phenylmaleimide is commercially produced by the mass (solution) process. When polymers such as ethylene glycidylmethacrylate or styrene-acrylonitrile-methacrylate are added to the ABS formulation, low gloss surface aesthetics are attainable. However, the processing window for fabricating parts with uniform low gloss surfaces is narrower than with mass ABS. Also, reducing the gloss of emulsion polymerized ABS with gloss-reducing polymeric additives tends to increase the melt viscosity of the resin.

3

IMIDES

Incorporation in a polymer of imide groups, that are five-membered planar rings which completely hinder the rotation of the imide residues around the backbone chain of the macromolecule, leads to (co)polymers with great structural stiffness and higher thermal stability. Poly(N-n-alkylmaleimide)s exhibit a Tg corresponding to the number of the carbon atoms in the n-alkyl groups [20], e.g. from 97 °C for poly(TV-n-octadecylmaleimide) (n = 18) to 185°C for poly(TV-n-butylmaleimide) (n = 4). A higher Tg can be obtained by incorporation of N-phenylmaleimide (PMI) [21] and with N-(alkyl-substituted phenyl)maleimides [22]. All the corresponding polymaleimides exhibit excellent thermal stability up to at least 370 °C. Poly(N-phenylmaleimide) undergoes decomposition without softening and melting, when heated to above 400 °C. Other substituted poly(N-phenyl maleimides) soften and melt between 400 and

HIGH HEAT RESISTANT ABS TECHNOLOGY

327

450 °C, accompanying their decomposition. Commercial supply of maleimide monomers started in the mid-1980s. Since that time, the use of imides has increased dramatically. A number of applications have been described, making use of N-alkylmaleimides, such as N-isopropylmaleimide (IPMI) [23] and Ncyclohexylmaleimide (CHMI) [24] for heat-resistant transparent polymer resins, useful, for example, for optical lenses, liquid crystal displays and disks [25]. The most important reason for this is that N-substituted maleimides having an aliphatic or alicyclic group are colorless and most suitable for yielding colorless transparent copolymers, whereas N-arylmaleimides are orange or yellow. However, the majority of process and product developments have been carried out with N-phenylmaleimide (PMI). PMI has also been the most widely used imide for ABS modification. Imide modification leads to some unique properties because it both elevates transition temperatures and thermal decomposition temperatures, allowing for easier fabrication than less stable styrene-containing terpolymers [26]. The deformation temperature of ABS has been shown to increase by 2-3 °C on adding 1 % of PMI. By adding 5–10%, the temperature can be elevated to over 125 °C [27]. Furthermore, facile copolymerization rates and resistance to hydrolysis reactions overcome many of the reaction engineering obstacles encountered by other high heat monomers when scaling to mass or emulsion processes. The synthesis of N-substituted maleimides (Figure 15.2) involves two steps. The first step is a quantitative reaction at ambient temperature between maleic anhydride and a primary amine, yielding a maleamic acid. The second step, requiring elevated temperatures, is ring closure and water elimination. A description of a typical process can be found in the patent literature [28]. Figure 15.3 illustrates that the free radical polymerization of PMI and styrene proceeds in an alternating manner [22,29,30]. Over a wide range of monomer ratio, the copolymerization (at low conversions) results in polymers with PMI content between 45 and 55 % and a Tg between 225 and 245 °C. The same type of alternation is observed in the copolymerization of styrene with maleic anhydride [31] and various N-alkyl-substituted maleimides [32–35]. It is 6

NH,

Figure 15.2

PMI synthesis

328

R. VANSPEYBROECK ET AL.

Figure 15.3 Copolymerization of PMI and styrene: diagram for radical copolymerization of PMI (M1) with styrene (M2) in benzene at 35 °C with AIBN as initiator [22]; f(1) = initial molar fraction of PMI in the monomer feed; F(1) = molar fraction of PMI in copolymers at conversion of 5 % (±0.5 %)

generally accepted that this type of polymerization implies the participation of a charge-transfer complex between styrene, being the electron donor, and maleic anhyride or the maleimide, being the electron acceptor [36–40]. The miscibility of PMI-containing polymers with SAN has been studied by different researchers and is of particular interest to understand the possibilities of blending approaches to increase the heat resistance of ABS. There is a surprisingly high degree of miscibility between styrene-co-Af-phenylmaleimide (SPMI) and SAN copolymers, that increases as the PMI volume fraction increases [41]. Various types of copolymerizations of PMI and styrene are described in the literature [42,43] and various grades of SPMI copolymers are commercially available and used to increase the heat resistance of ABS [44,45]. In addition to blending with SPMI copolymers, PMI can be incorporated into ABS using mass, emulsion [46–50] or suspension [42] free radical polymerization techniques. The high heat ABS resin can be completely produced by mass polymerization, or mass polymerized PMI-SAN can be blended with (emulsion polymerized) SAN-grafted rubber concentrates and/or conventional mass ABS. Sumitomo Naugatuck determined an empirical relation for the compatibility of SAN/SAN-PMI blends based on the polar monomers in each component [51]. Figure 15.4 shows that the miscibility window with SANs becomes wider with increasing PMI level in the terpolymer [52]. A complication with mass polymerization is that at the high temperatures, high concentrations of PMI lead to the generation of oligomeric species, so that

HIGH HEAT RESISTANT ABS TECHNOLOGY

329

% PMI in SAN-PMI Figure 15.4

Miscibility of SAN and SAN–PMI [52]

the boost in heat distortion per unit of PMI decreases. Another complication is that severe composition drift can occur since PMI and styrene tend to form an alternating copolymer with only small amounts of AN incorporated [53]. Owing to this composition drift, a continuous stirred tank reactor (CSTR) configuration is preferred to produce mass PMI–SAN. For PMI-modified complete mass ABS, composition drift can be minimized by adding PMI at different stages in the process. In this way, one can still benefit from the advantages of plug flow reactor technology, such as rubber particles sizing capability, and ease of producing low-gloss materials that are required in the primary automotive interior market. By adding a portion of the total PMI after rubber phase inversion, PMI can be divided between the rubber graft and the continuous phase, thus avoiding incompatibility [54,55]. Improved impact strength and fatigue resistance have been claimed when the total maleimide monomer content and the content of the continuous phase differ by <9 % by adding >20 % of the maleimide monomer at a point in the process after phase inversion [56]. Blends containing mass PMI–SAN and emulsion graft rubber concentrates (a mix of mass and emulsion technologies referred to as hybrids) are preferred for high-gloss applications. For automotive applications, where low-gloss esthetics are often required, hybrids containing mass PMI–SAN and both emulsion and bulk-polymerized SAN-grafted rubber result in excellent balances of toughness, heat resistance, and tensile properties [57]. Another advantage for automotive applications of PMI-containing resins produced by the continuous mass-solution process is the interrelated low carbonemission, fogging and odor properties due to the absence of emulsifiers and processing aids [58]. ABS producers with existing emulsion or suspension capital may also produce high heat grades via imide modification. Mitsubishi Monsanto Chemical described methods of preparing high heat ABS by blending 'graft copolymer'

330

R. VANSPEYBROECK ET AL.

PMI-ABS, suspension PMI-SAN, SAN and ABS [59]. They claimed a preferred blend combination of these components which has excellent heat distortion temperatures, superior impact (when mass and emulsion grafted rubber are present), and good stability at processing temperatures. There are a few alternative approaches to imide copolymers that allow the resin producer to make imide-modified high heat ABS without incurring the cost of the synthesized imide monomer. One is by reacting styrene–maleic anhydrides with a primary amine, either during the polymerization reaction with styrene or in a separate step. Mitsubishi Monsanto has practiced imidization on a commercial scale and described a process which follows the formation of S-MA with addition of amine and AN [60]. They described the manufacture of maleimide copolymers by heating the SMA copolymers with aniline in an extruder [61]. The maleimidation of the anhydride function is not complete, as there is unreacted amine or maleic anhydride in the product. The polymer stability and physical properties depend on the mole percent of maleimidation. Another approach is free radical copolymerization of a maleimic acid with styrene. Quantitative ring closure to the imide form was described by Newman of Dow Chemical Company [62,63]. Heat-resistant moldings can be obtained by pellet or melt blending of conventional ABS with maleimide containing master batch thermoplastics or thermoplastics containing colorants, with very high glass transition temperatures. Denki Kagaku Kogyo described master batches of maleimide copolymers and SAN-grafted rubber that has a Ts >140°C [64,65]. Nippon Shokubai Kagaku Kogyo holds a recent patent claiming an extrusion process for SPMI (54:46) copolymer pellets with a Tg of 206 °C, which exhibits no discoloration or thermal deterioration [66]. The excellent thermal stability of PMI-containing high heat ABS resins ensures that the heat resistance of the produced parts is not significantly influenced by thermal exposure applied during processing. Prolonged exposure at high temperatures of aMeS-based resins can result in an increase in the amount of residuals (Figure 15.5), a decrease in polymer molecular weight and a decrease in Vicat softening point of up to 6°C (Figure 15.6).

4

MALEIC ANHYDRIDE

As mentioned in the previous section, maleic anhydride (MA) undergoes copolymerization with styrene, and the five-membered ring anhydride lends rigidity to the resulting polymer chain backbone, resulting in a higher Tg. It has gained commercial utility in several forms: (1) transparent and glass-filled styrene–maleic anhydride copolymers (SMA); (2) filled and unfilled rubber modified SMA (DYLARK11); and (3) blends of SMA with ABS (CADON*).

HIGH HEAT RESISTANT ABS TECHNOLOGY

granules

331

245 degC/3 min. 270 degC/5 min. 295 degC/10 min. Processing Conditions

PMI Mass ABS Figure 15.5

D alphaMeS ABS

Residual monomers of high heat ABS as a function of thermal exposure

granules

245 degC/3 min. 270 degC/5 min. 295 degC/10 min. Processing Conditions

• Figure 15.6

PMI ABS

D alphaMeS ABS

Vicat softening point of high heat ABS as a function of thermal exposure

The predominant commercial synthesis of MA is by vapor-phase oxidation of hydrocarbons, e.g. benzene, n-butane, or a C-4 hydrocarbon mixture, over a solid catalyst [67]. The oxidation of benzene over a supported vanadium oxide catalyst is the preferred procedure. In a typical process, the reactor gas containing low concentrations of MA is passed through a heat exchanger and

332

R. VANSPEYBROECK ETAL

cooled to about 60 °C where most of the crude product is collected as a liquid. The product is sold as briquettes, flakes, or molten monomer in tank cars or trucks. The monomer is a white solid which melts at 53 °C and boils at 202 °C. The MA residue on the polymer backbone is known to promote hydrophilicity and adhesion, improve dyeability, give functionality for crosslinking, promote compatibility with other polymers and fillers, in addition to improving the heat distortion [68]. Early work by Baer at Monsanto showed that a nonhomogeneous resin could be produced very easily in a mass polymerization [69]. Similarly to PMI, the MA monomer tends to form 1:1 alternating copolymers with styrene arising from charge-transfer complexes [68]. Moore and co-workers at Dow Chemical showed that a three stirred tube mass polymerization train could produce a homogeneous rubber-modified SMA copolymer if each reactor was fully recirculated [70]. They employed vigorous recirculation and separate MA feeds to produce resins with about ca 25 wt% MA in the rigid phase. The excellent impact strength attained was indirect evidence of a single Tg, homogeneous rigid phase. Table 15.4 illustrates selected commercialized high heat styrenic resins with MA as the high heat modifier. Rubber-modified SMA resins fulfill specific market needs such as automobile instrument panel substrate materials where low-temperature ductility is not required. However, high tensile and high impact strengths cannot be simultaneously attained. Efforts to terpolymerize styrene, maleic anhydride, and acrylonitrile have revealed crosslinking behavior at high temperatures. Monsanto researchers reported a systematic study where rubber-modified S–MA–AN resins were made via mass polymerization with different AN levels [71]. They defined the degree of crosslinking which occurred during compression molding by testing for solubility in refluxing methyl ethyl ketone (MEK). The data are summarized in Table 15.5. Monsanto claimed that a maximum of 11 wt% AN can be added before severe crosslinking occurs. Melt processing of materials with this tendency to crosslink is challenging. Table 15.4 Selected MA-containing products and properties (all data from manufacturer's published data sheets) Product

Technology

MFR (g/l0 min) Notched Izoda (J/m) Vicar6 ( = C)

DYLARK® 232 SMA-transparent 1.9C DYLARK 8 378 Impact-SMA 1.0r CADON® 152 SMA-ABS 1.0^ a b c d

Notched Izod impact, AST D256. Vicat softening point, ASTM D1525. Melt How rate (230°C/2.16kg), ASTM D1238. Melt How rate (230 °C/3.8 kg), ASTM D1238.

10.7 154.8 133.5

118 127 126

HIGH HEAT RESISTANT ABS TECHNOLOGY

333

Table 15.5 Solubility properties of SMA–AN impact modified terpolymers after compression molding S:MA:AN

Solubility in MEK after devolatilization

After compression molding

68:26:6 68:23:11 63:23:14 56:26:19

Dispersible Dispersible Not Dispersible Not Dispersible

Dispersible Dispersible Not Dispersible Not Dispersible

To design a resin with the property enhancements of AN without the crosslinking problem, it was found that SMA copolymers and terpolymers could be blended with ABS resins to form miscible blends with properties of HHABS. A fundamental look at the miscibility of SMA copolymers with SAN copolymers indicated that the optimum thermodynamic interaction occurs when the AN content of the SAN is nearly equal to the MA content of the SMA [72]. Kim et al. also found low impact strengths at all modifier levels when blending SMA with SAN-g-polybutadiene (GRC = grafted rubber concentrate) [73]. Blends of SMA with SAN and GRC (SAN + GRC = emulsion ABS) exhibited ductility behavior similar to HHABS. The impact strengths of the polymers were 2-5 ftIb/in, in a notched Izod test at ambient temperature. Dow and Monsanto, among others, have investigated the manufacture of SMA resins both with and without rubber modification. Moore at Dow Chemical Company described a method of producing SMA copolymers via a recirculated coil reactor [74]. In general, SMA copolymers, impact-modified SMA copolymers, and glassfilled SMA copolymers have good competitiveness and reasonable processibility in applications which require heat properties greater than general-purpose HIPS and ABS. They provide a low-cost solution where low-temperature ductility is not required. The more ductile SMA–ABS blends have had limited success owing to their poorer flow and tendency to crosslink and decompose at higher temperatures.

5

MODIFIED NITRILES

Heat-resistant ABS resins can also be produced by polymerizing styrene with modified nitriles, such as fumaronitrile and maleonitrile. Fumaronitrile can be produced from acrylonitrile (Figure 15.7) in a two-step process involving the addition of hydrogen cyanide followed by oxydehydrogenation over metal oxide catalysts. The compatibility of styrene-co-acrylonitrile-co-fumaronitrile (SANF) terpolymers and SAN has been studied [75]. High-gloss, heat-resistant

334

R. VANSPEYBROECK ETAL. NC H7C

\

HCN

CN

Figure 15.7

maleonitrile

Synthesis of fumaronitrile

ABS technology has been developed whereby SANF terpolymers produced in a solution process were melt blended with ABS impact modifiers [76]. However, there are no major applications for the monomer and it is commercially available only in small quantities at high cost.

6

VARIOUS HIGH HEAT-RESISTANT ABS GRADES

In the positioning of ABS products against generic competition, the balance of several key properties for the cost is the key. ABS is sold primarily for its toughness, appearance, heat resistance, and flow. The flow relates to processability, particularly for injection molded applications. Essentially, the flow is balanced with the other properties (excluding appearance). In the case of injection molding one would want the highest toughness and highest flow for the best resin. Typically, increasing rubber increases the toughness, but the flow is decreased with the rubber phase being the highest viscosity. The reality is that toughness, flow, and heat are intertwined properties. One is usually maximized at the expense of others. Table 15.6 lists a series of ABS grades, from general-purpose molding grade to high heat grades in which the heat resistance has been increased over conventional ABS. The Vicat softening points have been increased by the incorporation of various levels and types of heat-boosting monomers, and range from 98 to 112oC. Heat-resistant resins can be divided into several classes of heat resistance: 100–103, 104–106, 107–110, and 111 + oC Vicat softening point. The data in Table 15.6 and in Figure 15.8 show that the toughness/melt flow balance decreases with increasing Vicat softening point. Within the 100–112CC range, mass ABS and mass/emulsion blended resins containing PMI exhibit an enhanced property balance when compared with emulsion ABS resins based on aMeS. An alternative route to obtain the equivalent heat/toughness/flow balance seen with PMI-containing ABS is to melt blend polycarbonate into the ABS resin. The blends contain less than 40 wt% PC and the ABS remains the continuous phase. Also included in Table 15.6 for comparison is a PC–ABS blend where PC is the continuous phase and the ABS is less than 40 wt% of the blend (PULSE8 2000EZ).

Table 15.6 Selected HHABS and PG–ABS resins (all data from manufacturer's published data sheets unless noted)a Heat range General-purpose

Grade

CYCOLAC® GPM5500 RONFOLIN® RT-51 RONFOLIN® GG-20 Medium heat MAGNUM 3325 MT CYCOLAC® BDT5510 CYCOLAC®X11 SINKRAL® C442 POLYLAC® PA-777B High heat RONFOLIN® HH-20 XZ-96502.00 PC modified NOVODUR® KU2-5300 RONFOLIN® HX-05 RONFOLIN® HX-03 CYCOLAC® G365 Highest heat CYCOLAC® Z48 RONFOLIN® HX-10 MAGNUM® 3416 SC MAGNUM® 358 HP MAGNUM® 357 HP Super high heat CYCOLAC® X17 PC–ABS(>60%PC) PULSE® 2000EZ

Supplier

Technology

General emulsion GE Plastics BASF Corporation aMS, emulsion aMS, emulsion BASF Corporation Low residuals, mass Dow Chemical Co. GE Plastics aMS, mass aMS, emulsion GE Plastics Low residuals, mass Enichem Chi Mei Industrial Co., Ltd aMS, emulsion aMS, emulsion BASF Corporation PMI, low residuals, mass Dow Chemical Co. Bayer Corporation Emulsion ABS/PC modified BASF Corporation aMS, emulsion aMS, emulsion BASF Corporation aMS, emulsion GE Plastics aMS, emulsion GE Plastics BASF Corporation aMS, emulsion PMI, low residuals, mass Dow Chemical Co. PMI, mass Dow Chemical Co. PMI, mass/emulsion hybrid Dow Chemical Co. aMS, emulsion GE Plastics PC/mass ABS blend Dow Chemical Co.

Charpy Vicat MFR (gVlOmin) (kj/m 2 ) (°C)

24.0 17.0 10.0 10 15.0 6 6.0 6.5 5.0 6.0 5.0 3.5 7.0 6.0 3.0 3.5 6.5 6.5 6.0 5.0 I5h

16 20 25 18 13 28 18 20 20 25 32 17 9 11 14 12 18 18 16 10 40

98 101 101 101 100 103 104 105 106 106 106 107 108 108 109 110 108 108 109 112 124

a

MFR: ISO 1133; 220 °C, 10kg. Charpy, notched @ 23 °C: ISO 179. Vicat: ISO 306; 50N/50°C/h The MFR for PULSE® 2000EZ is estimated for ISO 1133; 220°C, 10kg from the measured MFR of 7.00g3/10min value at 230°C/3.8kg. h

w

CO Ol

336

R. VANSPEYBROECK ETAL 46

PC/ABS

PC > 60 wt % 42

PC/ABS PC < 30 wt %

38

i| 34 3 30

§:

General Purpose ABS

5 26

! n o 18

O A

O

\

A

14 10

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0 20.0

22.0 24.0

26.0

MeltFlow Rate (g/10') x 100 - 103 deg C o 104-106 deg C

• 106 degC ; PC modified ABS A 107 - 110 deg C

» 106degC;PMI-ABS

o 112degC

--•Linear (107- 110 deg C)

• PC/ABS General Purpose ABS

Linear (100- 103 deg C)

Figure 15.8 Scatter plot of the balance of impact strength and melt flow for ABS and PC-ABS resins with different heat resistance (data from Table 15.6)

There are two linear trend lines plotted in Figure 15.8 for the medium heat range (100 – 103 °C) and the highest heat range (107 – 110°C). The trends indicate that as the heat is increased from the medium heat range into the highest heat range, the toughness flow balance of the highest heat resins becomes significantly lower (dotted line) than the toughness flow balance of the medium heat range of resins (solid line). The second major point to note in Figure 15.8 is that within a heat series, the slope of the trend lines is negative. This clearly indicates that as flow is increased, it is at the expense of toughness. One can also note from Table 15.6 that several resins have increased heat due to low residuals and high AN contents (not shown). An illustrative way to visualize the flow/toughness/heat balance is to view the key resin properties in a bubble plot as shown in Figure 15.9. It can easily be

337

HIGH HEAT RESISTANT ABS TECHNOLOGY

OMFR

i

0 10 20 30

k. 01

1 : £1e» _:

I

PC/A

f I

i a*

I

.

.

>M

4>-o

i

i

-'0

!

i

i

'i

"o

z :

! i

O

0

6 J... n-

o '; °

O«MeS

\.

A High AN Low ReisduaL

i

30

20 15

:

i

•

35

25

j

^ ' &

&« : d " 1 : _o -

IPC/AlU —^^^

! o^— ^ j i

O

i /"A iGPABSl

|

;

\

i

i

10 5 n

95

100

105 110 115 VICAT Heat Distortion (degrees Celsius)

120

125

Figure 15.9 Bubble plot of melt flow rate (bubble) versus Vicat heat distortion and notched Charpy

seen that the high heat ABS decreases in flow and toughness as the Vicat heat resistance is increased. Also, as one changes from a high heat ABS to a PCABS, where the ABS becomes the dispersed or co-continuous phase, a dramatic change in the flow/toughness/heat property balance occurs. The improved property balance is also accompanied by an increase in cost. In Figure 15.9, the open circles with solid lines are for high heat ABS made with aMeS, the open circles with horizontal parallel lines represent high heat ABS made with PMI, and the gray solid circles represent high heat ABS made with low residuals and high AN content with no high heat monomer. The dotted black circle is general-purpose grade ABS and the two bold black circles represent PC–ABS blends. The diameter of the bubble represents the flow of the resin as indicated in Table 15.6. As a final note, the Vicat heat distortion, as noted earlier, is an indicator of heat resistance. The actual heat performance of the part is influenced by the molding process conditions and the flow rheology of the resin itself (molded-in stress can noticeably affect the heat performance of an actual part. Thus the heat ranking as noted on the plot can be different to that in the final part owing to actual differences in rheology, which will then impart different levels of molded-in stress.

338

R. VANSPEYBROECK ET AL.

REFERENCES 1. Adam M. E., Buckley D. J., Colborn R. E., England W. P., Schissel D. N., Rapra Rev. Rep., 6, No. 10 (1993). 2. Mark H. F., et al. (Eds), Encyclopedia of Polymer Science and Engineering, Vol. 16, Wiley, New York, pp. 1–246 3. Product Bulletin SR-606E, Borg Warner Chemicals Inc., Parkers burg, WV. 4. MAGNUM ABS Resins, Product Bulletin 301-899-686, The Dow Chemical Company, Midland, MI. 5. Product Bulletin 301-665-687, The Dow Chemical Company, Midland, MI. 6. Product Bulletin 6364C, Monsanto Co., St. Louis, MO. 7. Maul J., Meyer H. H., Makromol. Chem., Macromol Symp. 53, 23 (1992). 8. Henton D., Dion R. P., Lefevre N. A., US Patent 4972032 (to The Dow Chemical Company) (1990). 9. Minematsu H., Matsumoto T., Saeki T., Kishi A., US Patent 4294946 (to Sumitomo Naugatuck Co.) (1980). 10. Abe M., Iwama M., Tsuchikawa S., Morikawa T., US Patent 4306043 (to Japan Synthetic Rubber Co.) (1981). 11. Mathumoto S., Jagoshi F., US Patent 4526 928 (to Kanegafuchi Kagaku Kogyo Kabushiki) (1985). 12. Grabowski T., GB Patent 1253 226 (to Borg-Warner Coporation) (1971). 13. Schwier C. E., Wu W. C, US Patent 4874829 (to Monsanto Co.) (1989). 14. Rinehart M., US Patent 4169195 (to Borg-Warner Corporation) (1979). 15. Sakano Y., Miyaki N., Japanese Patent JP 02051515 (to Denki Kagaku Kohyo KK) (1990). 16. Rudin A., Chiang S., J. Polym. Sci., 12, 2335 (1974). 17. Driscoll K., Dickson J., J. Macromol. Sci. Chem., A2, 49 (1968). 18. Cowle J. M. G., Elexpuru E. M., McEwen I. J., Polymer, 33, 1993 (1992). 19. Priddy D. B., Traugott T. D., Seiss R. H., J. Appl. Polym. Sci., 41 383 (1990). 20. Mastsumoto A., Oki Y., Otsu T., Polym. J., 23, 201 (1991). 21. Aida H., Kimura M., Fukuoka A., Hirobe T., Kobunshi Kagaku, 28, 354 (1971). 22. Barrales-Rienda J. M., Gonzalez de la Campa J. I., Gonzales Ramos J., J. Macromol. Sci. Chem., All, 267 (1977). 23. Kita Y., Kishino K., Nakagawa K., J. Appl. Polym. Sci., 63, 1055 (1997). 24. Sato H., Jpn. Kokai, 87 156115 (1987); Sato H., Matsuo M., Jpn. Kokai, 89 62315 (1989). 25. Maekawa M., Adachi M., Yasuda Y., Japanese Patent JP 06116338 (to Toray Indsutries) (1994). 26. Robinson J. C., Ziegelmeyer T. A., European Patent Application 0415120A2 (to General Electric Company) (1991). 27. Daihachi Chemical Industry, Plas. Ind. News (Jpn.) 30 (11) 163 (1984). 28. Kita Y., Kentaro S., Maseo B., Atsushi O., US Patent 4623734 (to Nippon Shokubai KK) (1986). 29. Iwatsuki S., Kubo M., Wakita M., Matsui Y., Kanoh H., Macromolecules, 24, 5009 (1991). 30. Matsumoto A., Kubota T., Otsu T., Macromolecules, 23, 4508 (1990). 31. Young L. J., in Polymer Handbook, Brandrup J., Immergut E. H. (Eds) Wiley– Interscience, New York, 2nd edn, pp. 11/105–386 (1975). 32. Coleman L. E., Jr, Conrady J. A., J. Polym. Sci., 38, 241 (1959). 33. Van Paesschen G., Timmerman D., Makromol. Chem., 78, 112 (1964). 34. Yamaguchi H., Minoura Y., J. Polym. Sci., A-l, 8, 1467 (1970).

HIGH HEAT RESISTANT ABS TECHNOLOGY 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.

339

Yamada M., Takase I., Kobunshi Kagaku, 23, 348 (1966). Furukawa J., /. Polym. ScL, Polym. Symp. Ed., 51, 105 (1975). Kokubo T., Iwatsuki S., Yamashita Y., Macromolecules, 1, 482 (1968). Shirota Y., Tomono T., Makromol. Chem., 141, 265 (1971). Seiner J. A., Litt M., Macromolecules, 4, 308 (1971). Abayasekara D. R., Ottenbrite R. M., ACS Polym. Prepr., 27, 462 (1986). Aoki, Y. Macrocomolecules, 21, 1277 (1988). He, J., Wang J., Li S., Gongneng Gaofenzi Uuebo, 12 (1), 19 (1999). Dedovets G. S., Kondratovich A. A., Ivanov V. S., Vestn. Leningr. Univ., Fiz. Khim. (4), 132 (1977). Yatagai H., Plast. Compounding, 45 (1993). Denki Kagaku Kogyo, Mod. Plast. Int., 26 (4), 93 (1996). Toyooka Y., Fujii S., Japanes Patent 03097703 (to Mitsubishi Rayon Co.) (1991). Maeda Y., Miyazaki H., Japanese Patent 03143910 (to Monsanto Chemicals) (1990). Byrdina N. A., et al., Plast. Massy, 10, 36 (1990). Jang B., Jung H., Park H., Korean Patent 9605 08 (to Cheil Industries Inc.) (1992). Maeda Y., Myazaki H. (Monsanto Chemicals, Japan), Japan Kokai Tokyo Koho. Masatsune K., Ogura S., Koiti K., US Patent 4877 833 (to Sumitomo Naugatuck Co.) (1989). Ogura S., Mastsune K., Katsuji U., 45th ANTEC Conference Proceedings, pp. 1365 (1987) Florjanczyk Z., Krawiec W., Makromol. Chem., 190, 2141 (1989). Iwamoto I., Ito N., Sugazaki K., Matsubara T., Ando T., US Patent 4808 661 (to Mitsui Toastsu Chemicals) (1989). Iwamoto I., Ito N., Sugazaki K., Matsubara T., Ando T., US Patent 4954517 (to Mitsui Toatsu Chemicals) (1990). Traugott, T. D., Workentine, S. L., US Patent 5412036 (to Dow Chemical Co.) (1995). Shields N., Van de Langkruis G., US Patent 5 270 387 (to Dow Chemical Co.) (1993). Reunis A., Hogg A., Naughton P. Schoppmann T.D. Nickel R., Markhardt A. SAE Technical Paper 1999-01-0853, International Congress and Exposition, Detroit, MI (March 4, 1999). Aoki Y., Hiroaki M., US Patent 4879343 (to Mitsubishi Monsanto Chemical Company) (1989). Ikuma S., US Patent 4381373 (to Mitsubishi Monsanto Chemical Company) (1983). Oshida T., Kajiwara T., Japanese Patent 02004806 (to Mitsubishi Monsanto Chemical Company) (1990). Newman T., US Patent 5015 712 (to The Dow Chemical Company) (1991). Newman T., US Patent 5077343 (to The Dow Chemical Company) (1991). Noguchi A., Shimura T., Miyashita S., Japanese Patent 10036614 (to Denki Kagku Kogyo) (1998). Shinmura, T., Konishi, K., US Patent 5532317 (to Denki Kagaku Kogyo) (1996). Kinoschit F., Yoshikawa K., Fujioka K., Japanese Patent 2000271928 (to Nippon Shokubai Kagaku Kogyo) (2000). Sharma R., Creswell P., Newson P., Am. Inst. Chem. Eng. Symp. Ser., 87 (1984). Trivedi B. C, Culbertson B. M., Maleic Anhydride, Plenum Press, New York (1982). Baer M., US Patent 2971939 (to Monsanto Chemical Company) (1961). Moore E., Lehrer C., Lyons C., McKeever L., US Patent 3919354 (to The Dow Chemical Company) (1975).

340 71. 72. 73. 74. 75. 76.

R. VANSPEYBROECK ET AL Lee Y, Trementozzi Q., US Patent 4262096 (to Monsanto Chemicals) (1981). Kim J., Barlow J., Paul D., J. Polym. Sci., Polym. Phys. Ed., 27, 223 (1989). Kim J., Kesskula H, Paul D., J. Appl. Polym. Sci., 40, 183 (1990). Moore E., Ind. Eng Chem. Prod. Res. Dev., 25, 315 (1986). Warakomski J. M., /. Appl. Polym. Sci., 46, 1057 (1992). Dion R., Warakomski J. M., US Patent 5 212240 (to The Dow Chemical Company) (1993).

16

Synthesis, Properties and Applications of AcrylonitrileStyrene-Acrylate Polymers G. E. McKEE, A. KISTENMACHER, H. GOERRISSEN, M. BREULMANN BASF AG, Ludwigshafen, Germany

1

INTRODUCTION

Acrylonitrile-styrene-acrylate (ASA) constitutes a versatile member of the group of styrenic copolymers used for housings, covers and other applications which require excellent surface appearance and environmental stability combined with high impact resistance and stiffness. It consists of a poly (styrene– acrylonitrile) matrix modified with small rubber particles. From its architecture, ASA is closely related to ABS; however, instead of polybutadiene rubber particles grafted with poly(styrene–acrylonitrile) (PS AN), poly(alkyl acrylate)-based graft rubber particles are used as the impact modifier (Figure 16.1). Generally, the poly(alkyl acrylate) core is slightly crosslinked to exhibit the needed elastomeric properties. Grafted chains, tethered to the acrylate rubber surface, serve as chemically bonded compatibilizers between the rubber particles and the PSAN matrix (Figure 16.2). These grafted chains usually consist of a PSAN, and optionally additional monomers such as methyl methacrylate are added. The density of the grafting points, the average chain length of the graft polymers, the graft comonomer composition and the relative amount of graft shell to poly(butyl acrylate) core play a significant role in the ASA mechanical and processing properties. In comparison with ABS, where the double bonds of polybutadiene are prone to oxidation and crosslinking due to oxygen, ultraviolet (UV) radiation Modern Styrenic Polymers: Polystyrene and Stvrenic Copolymers. Edited by J. Scheirs and D. B. Priddy f- 2003 John Wiley & Sons Ltd

342

Figure 16.1

G. E. McKEE

Architecture of acrylonitrile–styrene-acrylate

9 w

Styrene Acrylonitrile Rubber

Figure 16.2

Grafted PSAN chains tethered on a rubber surface

or heat, ASA is free of double bonds. Therefore, ASA, while having similar basic properties to ABS, has significant advantages in terms of UV stability and long-term heat resistance. Furthermore, the chemical resistance of ASA is significantly improved in comparison with ABS.

2

ASA MARKET

Only a few manufacturers of ABS also produce ASA. The reason for this is that in spite of the obvious similarities, there are also significant differences in the

ETAL..

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

343

production processes, and the know-how needed to produce a competitive ASA exists in only a few companies. Thus, within the different global regions there are only a few major players with significant ASA capacities, e.g. BASF, General Electric, Bayer, Hitachi and LG Chemicals. Since most of the manufacturers of ASA make use of multipurpose production plants that are used for both ABS and ASA production, the ASA output can be adapted to demand. It is estimated, that the annual demand for ASA is in the range 1-5 % of the annual ABS demand (ABS demand 2001: ca4.5 x 106 /t [1]).

3

PRODUCTION OF ASA

3.1

EARLY DEVELOPMENTS

The first attempts to produce an impact-modified PSAN using an acrylatebased rubber date back to the early 1960s when Herbig and Salyer [2] of Monsanto used a butyl acrylate–acrylonitrile rubber with a high acrylate content and Otto [3] of BASF used a rubber composed of an acrylate copolymerized with a crosslinking agent as the impact modifier. Further work at BASF by Siebel and Otto [4] polymerized butyl acrylate or ethylhexyl acrylate with butadiene and a vinyl alkyl ether in emulsion as the base rubber, which was then grafted with styrene and acrylonitile and used as an impact modifier for the brittle but stiff poly(styrene–acrylonitrile). This was followed by investigations of Willersinn et al. [5] of the same company in which the impact modifier was produced by polymerizing butyl acrylate and a crosslinking agent using free radicals in emulsion. The resulting polymer dispersion was grafted with styrene and acrylonitrile and after its isolation blended with PSAN. A similar patent by van der Werth and Lederer of BP Chemicals followed in 1968 [6]. This emulsion polymerization process is the basic process used today for the commercial production of ASA.

3.2

EMULSION POLYMERIZATION PROCESS

The production process for ASA is shown schematically in Figure 16.3 and is briefly discussed below.

3.2.1

Base Rubber Production

The base rubber is produced by polymerization of an acrylate in emulsion. A crosslinker can be employed, e.g. a diol diacrylate such as butanediol diacry late, divinylbenzene allyl(meth)acrylate, dioldiallylcarbonate, acrylic esters of

344

G. E. McKEE ETAL. base poly(butyl acrylate) rubber grafting with styrene and acrylonitrile polymerisation of the graft rubber in emulsion storage tanks

isolation of rubber extruder ASA

Figure 16.3

Preparation of ASA

tricyclodecenyl alcohol, diallyl maleate, diallyl fumarate and diallyl phthalate [5,7–11]. In most of the commercial products butyl acrylate is employed; however, in the majority of the ASA patent literature, other acrylates, especially ethylhexyl acrylate, are mentioned. The advantage of the latter over poly(butyl acrylate) is its lower glass transition temperature (Tg), -65°C, compared with poly(butyl acrylate) with a value of —45°C. This lower T% opens up the way for better lowtemperature properties for the ASA. The low-temperature properties of ASA are touched on briefly in Section 5. The mean weight particle size is in the usual range for the emulsion polymerization and is less than 1 u,m.

3.2.2

Grafting of Base Rubber

The base rubber is then grafted with styrene and acrylonitrile. The purpose of the PSAN graft shell is to anchor the rubber particles in the PSAN matrix and also to ensure their good dispersion in the PSAN matrix. The graft shell is usually not crosslinked; however, this is not always the case [9].

3.2.3

Isolation of the Grafted Rubber

The grafted rubber is usually isolated by either spray drying or by coagulation followed by drying. The blending of the coagulated rubber containing residual water with the PSAN melt in an extruder is also possible [12].

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

3.2.4

345

Preparation of the PSAN

The PSAN matrix can be prepared in suspension, solution or emulsion. The first two are the most commonly used methods. After the PSAN preparation, the unreacted monomers, water and if necessary solvent are removed.

3.2.5

Blending of PSAN and the Impact Modifier to give the End Product

Blending of the PSAN and the impact modifier is usually carried out in an extruder at a temperature range of 220-300 °C to give the end sales product. During this blending step, additives such as lubricants, pigments and antioxidants can also be added. The morphology of an ASA product with a mean weight particle size of approximately 0.1 jxm is shown in Figure 16.4. As can be seen, the rubber particles have agglomerated, which according to Ramsteiner [13] leads to better mechanics.

3.3

BULK POLYMERIZATION PROCESS

The preparation of ASA in bulk or bulk-suspension polymerization processes has been described by McKee et al. [14–18]. The system is similar to that used in the preparation of high-impact polystyrene (HIPS) and in bulk-produced ABS. Thereby the rubber was prepared using free radical polymerization, dissolved in the SAN monomers which were then polymerized using free radicals. Phase separation between the rubber and PSAN occurred, followed by phase inversion. In the preparation of HIPS and ABS prepared in bulk, the polybutadiene rubber is easily grafted at the pendant 1,2-C—C double bond. In the case of

Figure 16.4

TEM of an ASA containing an emulsion-produced impact modifier

346

G. E. McKEEE7V\L.

poly(butyl acrylate) the rubber is not easily grafted; however, grafting of the SAN on the poly (butyl acrylate) rubber can be achieved by: • Copolymerizing the acrylate monomer with a monomer containing two or more polymerizable double bonds. Thereby some of the double bonds should survive the polymerization stage of the butyl acrylate and serve as grafting points in the SAN polymerization [14]. • Copolymerizing the butyl acrylate with a monomer containing a peroxide group. The peroxide group is stable at the polymerization temperature of the butyl acrylate. After the rubber formation the temperature is increased to decompose the peroxide side group in poly(butyl acrylate), leading in the presence of SAN monomers to a grafting reaction [15]. Alternatively, if higher temperatures are not desired, the peroxide side group can be decomposed using a redox system [16]. • Copolymerizing the butyl acrylate with a SAN macromonomer [17]. An elegant variation of this process is the anionic polymerization of ethylhexyl acrylate or butyl acrylate using, e.g., styrene as the solvent [18,19]. With the correct choice of catalyst only acrylate double bonds are polymerized and not the styrene. Grafting points were introduced via Copolymerizing the ethylhexyl acrylate with a monomer containing two double bonds. The first double bond is chosen to be an acrylate and is thus copolymerized into the rubber backbone. The second double bond, which is not an acrylate, remains intact and can be used as a grafting point in the free radical polymerization step of the SAN monomers. The morphology (Figure 16.5) [20] is different to that obtained using the emulsion process (Figure 16.4). In the case of the emulsion-prepared product the particles are smaller than those produced in bulk and they contain less inclusions (salami structure) than the bulk-prepared product.

Figure 16.5

JEM of ASA produced in bulk

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

3.4

347

MICROSUSPENSION POLYMERIZATION PROCESS

A further method for producing ASA uses rubber particles prepared in a microsuspension process [21-24] (for morphology, see Figure 16.6). Thereby the monomer-water mixture containing a protective colloid is intensively sheared to produce monomer droplets which are then polymerized to give polymer particles of approximately the same size. These particles are then grafted with styrene and acrylonitrile and used as the impact modifier for the ASA. The particles may also be used in polymer formulations as delustering agents [22]. Owing to their large size they lead to an uneven surface of the moulded article (Figure 16.7) which leads to light scattering and thus matt surfaces. Brandstetter et al. [25] produced similar particles without the intensive shearing step by polymerizing acrylate monomer in water in the presence of a long-chain alcohol, an emulsifier and a water-insoluble initiator. The particles have a mean particle size of 0.2–6 u>m. After their preparation the particles can be grafted with SAN and then used as impact modifiers for ASA or as delustering agents for thermoplastics [26].

Figure 16.6 ASA containing a rubber produced using the microsuspension process

Figure 16.7 Scanning electron micrograph of ASA containing an ASA impact modifier prepared using the microsuspension process. These large particles give delustered surfaces

348

4 4.1

G. E. McKEE ET AL.

PROPERTIES OF ASA AGEING PROPERTIES

At first glance, ASA possesses a similar chemical structure to ABS, since both consist of a SAN matrix containing a graft rubber. However, while the core of the graft rubber of ABS consists of polybutadiene, that of ASA consists of poly(n-butyl acrylate) (Figure 16.8), and this accounts for important differences in the properties of the two plastics. In ABS, the double bonds of polybutadiene are prone to oxidation and crosslinking due to oxygen, UV radiation or heat [27,28] (Figure 16.9). The result is deterioration of the rubber, leading to loss of impact strength and discoloration. In contrast, the butyl acrylate rubber of ASA is free of C—C double bonds which gives ASA clear advantages in terms of weatherability [29,30] (Figures 16.10 and 16.11) and resistance against heat ageing (Figure 16.12). ABS rubber: Polybutadiene

n

CH2

= CH - CH = CH2

>

... - C H 2 - C H = CH-CH 2 -CH 2 - CH = C H - C H 2 - ... (1,4 Addition) or ... -CH 2 -CH-CH 2 -CH-... I I CH CH II II CH2 CH2

(1,2 Addition)

ASA rubber: Poly(n-butyl acrylate)

n

Figure 16.8

CH2 = CH I O = C - O - C4 H9

>

...-CH2-CH-... I O = C - O - C4 H9

Chemical structure of the rubber components of ABS and ASA

... - CH = CH - CH2 - CH2 - ...

02

+

... - CH - CH - CH2 - ... I

oo«

». ... - CH - CH = CH - ... —*• I OOH Figure 16.9

Alcohols, diols, ketones, esters or chain cleavage, crosslinking

Attack by oxygen and hydroperoxide formation in polybutadiene

349

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

Figure 16.10 Scanning electron micrographs of the surface of ABS (UV stabilized) and ASA after 500 h of xenon arc weathering according to ISO 4892-2A (specimens struck on exposed side)

0 Figure 16.11 Florida

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Hours of sunshine Change of multiaxial impact strength during outdoor weathering in

20

40 Time (weeks)

60

100

Figure 16.12 Multiaxial impact strength (ISO 6603–2) of ASA and ABS after heat ageing at 90 °C

G. E. McKEE ET AL.

350

A further advantage of ASA vs ABS is its higher resistance against environmental stress cracking, especially against alcohols and many cleaning agents [31] (Figure 16.13). ASA also exhibits advantages over other thermoplastic housing materials such as polycarbonate, PBT and polypropylene, as shown in Table 16.1 The low moulding shrinkage of ASA and of PC is advantageous for housings and covers, because warpage problems are almost nonexistent with these products. However, PC has only a limited resistance against environmental stress cracking, for example by alcohols and cleaning agents, and it yellows much more than ASA during outdoor exposure. Compared with polypropylene, a material widely used because of its low price, ASA has advantages in terms of stiffness, impact strength, heat distortion temperature and weatherability. As an example of the last aspect, the change in gloss during outdoor weathering of UV-stabilized, white-coloured ASA and polypropylene is shown in Figure 16.14. Further, owing to its high moulding shrinkage, polypropylene has a high tendency to warp during or after processing, and it is much more sensitive to scratches than ASA. Thus, because of its superior physical properties, and because it can be easily processed by injection moulding or extrusion, ASA has widespread use and can be found in a large variety of applications. 60 ASA

50 40 30 20

X

ABS

10

100

200

300

400

Ball oversize(fim)

Figure 16.13 Resistance against environmental stress cracking by 2-propanol (ball indentation test ISO 4600, 1 h exposure)

351

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA Table 16.1 door use

Comparison of unreinforced thermoplastic housing materials for out-

Property

Units

ASA PC

Moulding shrinkage (typical) Tensile modulus Tensile strength Notched Charpy impact strength HDT/A Chemical resistance/ environmental stress cracking resistance

% MPa MPa kJ/m 2 °C

0.5 2300 48 15 97 High

0.5 2400 63 100 129 Low

PBT

PP homopolymer

1.5 2500 60 6 65 Very high

1.5 1500 35 3 60 Very high

Polypropylene UV-stabilised. white

9

12

15

18

20

24

Time (months) Figure 16.14 Change of gloss of UV-stabilized, white-coloured ASA and polypropylene during outdoor weathering in Germany

4.2

IMPACT BEHAVIOUR

Ramsteiner et al. [32] investigated the rubber toughening of PSAN. In ASAcontaining particles with a diameter of ca 0.5 |xm and at a graft rubber concentration of 50 wt %, the distance between the rubber particles is small enough for

352

G. E. McKEE ET AL.

energy dissipation via stretching and shearing to occur. With large particles of > 1 jxm and low rubber concentrations such as occur in bulk-prepared ABS, the shearing process at large inter-particle distances cannot occur; however, the large particles are able to dissipate the energy via crazes. If the rubber particles are smaller and in the range of 80 nm the impact strength is low; however, if they agglomerate during processing then a large increase in impact strength occurs [13,32]. To compensate for the low toughness in ASA when using small rubber particles, large particles prepared using the microsuspension polymerization process can be added to the products (700 nm to 100|xm) [22]. Further work showed that the use of particles with a diameter of 0.15–0.8 fim brings better toughness than a particle size of less than 0.1 (xm [33].

5

ADDITIONAL AREAS OF INVESTIGATION

Additional areas of investigation include: • To reduce the surface gloss of articles made from ASA, e.g. by copolymerizing additional monomers into the base rubber or the graft shell or by incorporating large particles into the ASA formulations [22,23,25,34–55]. • To develop flame retarded grades [56–58]. A lot of work has been carried out using ASA—polycarbonate blends. However, this field will not be covered here. • To reduce the glass transition temperature of the rubber to improve the lowtemperature properties of the ASA products. Especially preferred is ethylhexyl acrylate [59,60]. A further method is the incorporation of silicone rubber into the ASA particles [55,61–65].

6

ASA BLENDS

Blends of ASA with many of the common thermoplastic materials are state of the art. The properties of the finished products depend to a large extent on the polymer compatibility often induced by the use of reactive polymers. Only a few of these blends have reached a significant commercial status. The largest blend products in this area are ASA-PC blends followed by ASA—PBT and ASA-PC-PMMA blends. Combinations of ASA with high-Tg matrix polymers are also frequently found in commercial products. Although polycarbonate is an engineering thermoplastic material which provides high toughness, flexibility and thermal stability, it suffers from certain limitations due to poor chemical resistance and low flow characteristics in injection moulding. These shortcomings can be circumvented by blending PC

353

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

with ABS or ASA. Generally, the impact strength and heat deflection temperature increase with increasing PC content (Figure 16.15). ASA as a blend component provides improved weatherability, which is a prerequisite for outdoor applications. In blends of polycarbonate with an ASA (PSAN containing 25 wt% of grafted rubber) extruded at 270 °C [66], the notched Izod impact strength surprisingly showed a maximum at 20 wt% ASA. This striking result is probably due to the critical thickness of the test specimen employed. It is known that for polycarbonate above a critical thickness of the test specimen the notched impact strength deteriorates steeply. This unfavourable critical thickness effect can be eliminated by adding amounts of >5% of ASA to PC [67]. The improved flowability of ASA-PC blends was addressed by Yu [68], by suggesting the use of a three-stage impact modifier, consisting of an acrylate core that is covered by a first shell of crosslinked PSAN and a further shell comprised of linear PSAN. These blends, while retaining the impact properties of the material, exhibit an improved flowability by a factor of 3. Another option in improving the flowability of these blends is the use of a branched polycarbonate. Brandstetter et al. [69] found that a dianhydride branched polycarbonate combined with ASA can be injection moulded at temperatures 5–10 K lower than similar blends based on linear polycarbonates. Increased toughness and colourability over conventional ASA-PC blends can be attained by using a two-stage grafted n-butyl acrylate core [70]. In the first grafting step pure styrene is added followed by a styrene—acrylonitrile Vicat softening temperature,

Charpy notched impact strength, [kJ/m2] —*60 -,

30 70 wt-% PC in ASA/PC blend

100

Figure 16.15 Charpy notched impact strength and Vicat temperature versus PC content in ASA-PC blends

354

G. E. McKEE ETAL.

comonomer composition. Core-shell grafted ASA rubbers containing a styrene core are used in ASA-PC blends to yield products with high gloss and excellent colourability [71]. Ternary blends of ASA—PMMA—PC exhibit improved toughness and flexural strength in comparison with ASA—PMMA blends [72]. An optimum in tensile modulus was found for blends containing >50% ASA. To overcome the critical thickness impact deficiencies of polycarbonate blends, a two-phase crosslinked acrylate—crosslinked styrene—acrylonitrile modifier has been suggested [73]. The same two-phase acrylate—crosslinked styrene—acrylonitrile product is used for significantly increasing the notched impact strength of polycarbonate-poly(l,4-butylene terephthalate) blends containing 18– 30 wt% of PC [74]. Inferior processablility and multiaxial impact strength of conventional ASA—polycarbonate blends can be improved by a three-stage grafted acrylate—rubber of average particle size (d50 weight average) of 200– 700 nm, with a first grafting stage comprised of styrene, a second stage of styrene—acrylonitrile and a third stage of methyl methacrylate [75]. A further improvement of impact strength and colourability can be obtained by using an ASA component with a graft rubber composition having a bimodal particle size distribution [76]. A different approach to solve shortcomings in impact strength and colourability of an ASA-PC blend is the addition of a SAN-grafted EPDM rubber [77]. The low-temperature impact strength can be significantly raised by adding a methyl methacrylate grafted polydimethylsiloxane—poly (n-butyl aerylate) interpolymer [78]. Ternary blends of ASA-PC and an amorphous polyester were shown to have improved melt flow and thick section impact strength in comparison with ASA-PC blends [79]. To improve the notched impact strength and to prevent hydrolysis induced embrittlement of PBT, PBT—styrenic blends have been thoroughly investigated. In comparison with ASA-PC, blends of ASA and PBT generally have to be compatibilized in order to yield acceptable mechanical properties. In the case of glass fibre reinforced PBT—ASA blends, an added and important bonus is reduced warpage properties on injection moulding in comparison with reinforced PBT. One option for compatibilizing PBT with ASA is the use of reactive polymeric additives that are compatible with the PSAN matrix and bear groups (e.g. carboxylic, carboxylic anhydride, epoxide or isocyanate functionalities) for forming chemical bonds with the hydroxyl and carboxyl groups in PBT. A recent study reported on the hydrothermal ageing of ASA—PBT blends containing up to 30% ASA [80]. It was revealed that the moisture uptake at 60CC corresponds to the percentage of modifier rubber. Strikingly, the mechanical properties deteriorate with increasing ASA content, probably resulting from insufficient compatibilization between the two phases. An ASA—PBT with improved hydrolysis resistance and reduced warp was reported for a resin composition containing a difunctional epoxy compound such as bis(3,4-epoxycyclohexylmethyl) adipate [81]. To increase the heat distortion temperature of a PBT—ASA blend by 10–20°C, the addition of talc at a

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

355

concentration of 0.5% is recommended [82]. To overcome disadvantages in colorability, a blend of 40 wt% PMMA, 10 wt% PSAN and a graft butyl acrylate rubber blend with a bimodal particle size distribution was developed [83]. In blends of polyamide with styrenics, compatibilization of the phases is generally a prerequisite for acceptable mechanical properties. This can be accomplished by having reactive groups in the PSAN phase that react with the amino groups of the polyamide phase. Blending is used to improve the main weaknesses of the polyamide, namely moisture sensitivity, toughness and warpage on injection moulding. Another motivation for blending ASA is to improve the heat distortion temperature of the material. This can be achieved by increasing the Tg of the continuous PSAN phase, for example by substituting the PSAN matrix partially or completely with poly(a-methylstyrene-acrylonitrile) [84]. This leads to an increase in the Vicat temperature of up to approximately 110°C. On increasing the amount of the a-methylstyrene-acrylonitrile copolymer, the ease of processability decreases, so often a mixture of both copolymers is used. [85] Alternatively, the Tg of the PSAN phase can be increased by using a terpolymer of a-methylstyrene-acrylonitrile-acrylamide [86]. Another approach to increase the Tg of the continuous phase of the ASA products is to blend it with maleic anhydride- or N-phenylmaleimide-containing resins. The use of matrix tetrapolymers of styrene, acrylonitrile, maleic anhydride and Nphenylmaleimide is recommended to improve considerably the heat distortion temperature [87]. Similarly, a matrix consisting of N-phenylmaleimide—acrylonitrile-styrene used in combination with a core-shell ASA rubber with a polyorganosiloxane core was found to have good colourability and excellent lowtemperature impact strength [88]. It must be remembered, however, that it may be necessary to adjust the graft comonomer composition when adding polar high-Tg copolymers to the matrix. For a matrix containing 22–36 wt% of a styrene—maleic anhydride copolymer and 36–50 wt% of PSAN, the resulting material exhibits an optimized impact strength when the acrylonitrile content of the rubber graft shell is in the range 45–55 wt% [89]. Alternatively, a matrix blend consisting of PSAN, a-methylstyrene-acrylonitrile copolymer and poly-N-methylglutarimide can be used, yielding ASA with Tg up to 120°C [90].

7 7.1

APPLICATIONS OF ASA GENERAL

Owing to its favourable combination of properties, ASA has grown substantially from a small niche product to an important plastic material since its

356

G. E. McKEE ET AL.

introduction into the market some 30 years ago, and it is now being used for many interior and exterior applications. The largest area of application for ASA is the automotive sector [91,92]. Almost all automotive manufacturers use ASA for unpainted exterior parts such as mirror housings, radiator grills, cowl vent grills and pillar covers. ASA is also increasingly being used in the transportation industry [93] and for truck and motor scooter segments, for example for truck door steps, radiator grills, scooter fairings, and other, mostly unpainted, exterior parts. A long-standing use for ASA is for housings for various equipment, machines, etc. [94–99] A further important field of application for ASA is the household sector. Typical applications include: • housings for electric toothbrushes and interdental cleaners; • housings and covers for sewing machines and kitchen appliances; • garden equipment, such as lawnmower housings, parts for lawn and garden irrigation devices and housings for garden lamps [100]; • children's toys [101]. In the building sector, ASA has found various applications such as window frames and sanitary equipment [102–105]. The leisure and sports sector constitutes a further important area of application. Thus, ASA is used to manufacture boat hulls, roofs and sills for camper vans and other exterior plastic parts for recreational vehicles. Since ASA has been on the market for over three decades and has thus acquired a certain maturity, it would be easy to conclude that no more major innovations can be expected from such a product. However, several important new fields of application are currently being developed. Among these are solar and safety technology and automotive body panels. 7.2

SOLAR TECHNOLOGY

Alternative sources of energy such as solar power are gaining significance because of the threat of global warming induced by the use of fossil fuels. A number of new applications involving ASA in the use of solar energy, such as housings for solar collectors and carriers for solar cells, have been developed in recent years ASA is also being used for the housings of solar-powered street numbers, which are already being made by several manufacturers. They contain photovoltaic cells in order to deliver electricity to rechargable batteries. Further applications for ASA in the area of solar energy include, for example, sunlight sensors, solar-powered battery chargers and solar-charged flashlights.

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

7.3

357

SAFETY IN THE HOUSE AND IN THE OFFICE

Another relatively new but strongly growing area of application is that of security and safety components. More and more people are relying on electronic systems to protect themselves, their homes and their offices. Such systems are comprised of a variety of components such as sensors, movement detectors and transmitter modules. Since many of these parts are mounted on the exterior of buildings, a main requirement is long-term weather resistance. Examples of applications in which ASA is already in use are [106] • • • • •

sensors and transmitter modules for electronic alarms; movement detectors; safety lighting and alarms; housings for access control units: housings for fire alarms.

7.4

A very promising new field of application for ASA and ASA-PC blends is automotive body panels. Until now, these have been made almost exclusively from painted metal. Since painting is costly, automotive companies are attempting to save costs by developing technologies that dispense with this operation. As in the case of the 'smart' car, bulk coloured thermoplastics (e.g. PBT—PC blends) may be used to make body panels. However, such parts must still be sprayed with a top coat in order to achieve the desired scratch resistance and UV stability. A promising alternative for the production of unpainted body panels is the use of a co-extruded thermoplastic film as a decorative layer which is backmoulded with a thermoplastic material [107]. This technique has received the name 'paintless film moulding' (PFM®) (Figure 16.16). The thermoplastic films for the PFM® technology normally have a total thickness of approximately 1 mm (Figure 16.17) and have an outer layer of PMMA which provides high gloss, scratch resistance and excellent weatherabilty. The base layer of the film is made from coloured ASA, chosen for its good adhesion both to PMMA and the backmoulding material (e.g. ABS) and for its excellent weatherability and long-term heat resistance. If special colour effects, e.g. metallic effects, are required the film may contain an additional intermediate layer made from high-impact PMMA. In order to make parts by means of the PFM® technology, the co-extruded film is thermoformed and then backmoulded with a compatible thermoplastic. Alternatives for the backmoulding material include, for example, ABS (without or with glass fibre reinforcement) and glass fibre reinforced PBT—ASA blends.

358

G. E. McKEE ET AL.

Preform design: 1. PMMA — r2. Coloured PMMA 3. ASA, ASA/PC -

Layer for layer toward resistance 1. Top layer • gloss • scratch resistance • hardness • weatherability

I and 2= top layers

3.-Film base layer • Injection moulded backing material 4a. ABS (trim, mirror housing, etc.) 4b. e.g. ABS-GF, PBT/ASA-GF (large-area body components)

Figure 16.16

2. Color layer • colour • UV resistance

4. Injection moulded backing material • strength • rigidity 3. Film base layer • impact strength • thermal expansion • toughness • resistance to thermal • colour mechan. strength deformation

Process sequence for paintless film moulding

Process sequence Coextrusion Multiple layer film extrusion

Film preforming Thermoforming

Component production Film insertion into mold

Trimming

In-mold decoration

-oFigure 16.17

Structure of the films used for the PFM* technology

The PFM® films may also be backed with a glass fibre reinforced polyurethane foam. This technology is already being used in thermoplastic roof modules and gives rise to parts having low weight, high stiffness and excellent thermal insulation.

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA

8

359

FUTURE PERSPECTIVES

Although ASA has been marketed for more than 30 years, it is still a product with a considerable potential for new applications owing to the well-balanced cost—property relationship. The strength of ASA lies in the unique combination of its good weatherability, toughness, surface properties and resistance to chemicals. To be successful as an ASA supplier in the future the following prerequisites seem to be mandatory: back integration for the main raw materials, innovative process and development, development of new applications coupled with a reliable and skilled technical customer support service.

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

Lasche H., van Roessel R., Kunststoffe 91, 276 (2001). Herbig J., Salyer I. O.; Monsanto Co., US 3 118 855. Otto H.-W.; BASF Aktiengesellschaft, DE 1182811. Siebel H. P., Otto H.-W.; BASF Aktiengesellschaft, DE 1 238207. Willersinn H., Otto H.-W., Paul R., Schuster L.; BASF Aktiengesellschaft, DE 1260135. van der Werth A., Lederer F.; BP Chemiclas, DE 1 960409. Luetje H.; BASF Aktiengesellschaft, DE 2 311129. Hambrecht J., Schmitt B., Rebafka W., Stephan R., Schwaab J.; BASF Aktiengesellschaft, DE 3 134 103. Yu J. R., Gallager R. E.; Stauffer Chemical Company, US 3944631. Swoboda H., Lindenschmitt G., Bernhard C; BASF Aktiengesellschaft, DE 2826925. McKee G. E., Koch J., Fischer W., Rosenau B., Czauderna B.; BASF Aktiengesellschaft, DE 19508312. Laber W., Gottschalk A., Schwaab J., Jeckel G., Mosthaf H.; BASF Aktiengesellschaft, DE 2 037 784. Ramsteiner F., Kunststoffe, 67, 517 (1977). Mckee G. E., Gausepohl H., Moors R., Rosenau B., Heckmann W.; BASF Aktiengesellschaft, DE 4440676. McKee G. E., Moors R., Gauspohl H., Seibring J.; BASF Aktiengesellschaft, EP 792298. McKee G. E., Rosenau B.; BASF Aktiengesellschaft, DE 19623661. McKee G. E., Rosenau B.; BASF Aktiengesellschaft, DE 19614845. McKee G. E., Jimgling S., Warzelhan V, Gausepohl H.; BASF Aktiengesellschaft, DE 19651300. Jungling S., Mckee G. E., Warzelhan V., Gausepohl H., Fischer M.; BASF Aktiengesellschaft, DE 19651299. Unpublished work. Mckee G. E., Renz G., Jahns E., Kastenhuber W.; BASF Aktiengesellschaft, DE 19633626.

360

G. E. McKEE ET AL.

22. McKee G. E., Rosenau B., Goerrissen H., Jahns E.; BASF Aktiengesellschaft, DE 19702733. 23. McKee G. E., Jahns E., Fischer W., Guentherberg N., Rosenau B.; BASF Aktiengesellschaft, DE 4443 886. 24. Kanebuchi Kagaku Kogyo KK, JP 101 2074. 25. Brandstetter F., Hambrecht J., Hildenbrand P., Echte A.; BASF Aktiengesellschaft, DE 3114090. 26. Unpublished work. 27. Mark H. F., et al. (Eds), Encyclopedia of Polymer Science and Engineering, Wiley, New York, (1985), pp. 388–425. 28. Vollmert B., Grundriss der Makromolekularen Chemie, Vol. 2, Vollmert-Verlag, Karlsruhe, 1979, pp. 219-225. 29. Rosenau B. Kunststoffe 85, 805 (1995). 30. Zahn A., Kunststoffe 87, 314, (1997). 31. Gorrissen H., Kunststoffe Plast. Eur. 89, 31 (1999). 32. Ramsteiner F., McKee G. E., Heckmann W., Fischer W., Fischer M., Ada Polym. 48, 553 (1977). 33. Mittnacht H., Priebe E.; BASF Aktiengesellschaft, DOS 1 911 882. 34. Eichenauer H., Doering J., Ott K.-H., Bottenbruch L.; Bayer AG, EP 139271. 35. Eichenauer H., Zabrocki K., Doering J., Ott K.-H., Bottenbuch L.; Bayer AG, DE 3421 353. 36. Ostermayer B., McKee G. E.; BASF Aktiengesellschaft, DE 3620684. 37. Niessner N., Seitz F.; BASF Aktiengesellschaft, DE 4131 729. 38. Niessner, N., Fischer W., Guentherberg N., Ruppmich K., Seitz F; BASF Aktiengesellschaft, DE 4142910. 39. Fischer W., Guntherberg N., Niessner N., Ruppmich K., Seitz F.; BASF Aktiengesellschaft, DE 4216549. 40. Niessner N., Seitz F., Fischer W., Guentherberg N., Ruppmich K., Moors R., Weiss R.; BASF Aktiengesellschaft, DE 4221 293. 41. Wittmann D., Schoeps J., Beicher H., Piejko K.-E., Weirauch K.; Bayer AG, DE 4229642. 42. Fischer W., Deckers A., Guentherberg N., Niessner N.; BASF Aktiengesellschaft, DE 4 234 296. 43. Fischer W., Guntherberg N., Niessner N.; BASF Aktiengesellschaft, DE 4235976. 44. Fischer W., Guentherberg N., Niessner N.; BASF Aktiengesellschaft, DE 4237 640. 45. Fischer W., Guntherberg N.; BASF Aktiengesellschaft, DE 4242485. 46. Fischer W., Guentherberg N.; BASF Aktiengesellschaft, DE 4439969. 47. McKee G. E., Rosenau B., Wendel K.; BASF Aktiengesellschaft, EP 732377. 48. Mckee G. E., Koch J., Fischer W., Gottschalk A., Guentherberg N., Rosenau B.; BASF Aktiengesellschaft, DE 19509514. 49. Rosenau B., McKee G. E., Schweiger C; BASF Aktiengesellschaft, DE 19536892. 50. McKee G. E., Rosenau B., Heckmann W.; BASF Aktiengesellschaft, DE 19614844. 51. McKee G. E., Rosenau B.; BASF Aktiengesellschaft, DE 19614845. 52. McKee G. E., Roseenau B., Heckmann H.; BASF Aktiengesellschaft, DE 19614846. 53. Bennet J. H., Muelbach K., Kogowski G.; BASF Aktiengesellschaft, WO 9933914. 54. Goerrisen H., Morgenstern H., McKee G. E.; BASF Aktiengesellschaft, DE 19837854. 55. Niessner N., Fischer W.; BASF Aktiengesellschaft, DE 4342045. 56. Seitz F, McKee G. E., Buschl R.; BASF Aktiengesellschaft, DE 4000544.

SYNTHESIS, PROPERTIES AND APPLICATIONS OF ASA 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91.

361

Eckel T., Ooms P., Wittmann D., Buysch H., Bayer AG, DE 4231 774. Techno Polymer KK Aktiengesellschaft, JP 11116827. Seitz F., Ruppmich K.; BASF Aktiengesellschaft, DE 4005210. Fischer W., Koch J., McKee G. E.; BASF Aktiengesellschaft, EP 725091. Niessner N.; BASF Aktiengesellschaft, DE 4124251. Niessner N., Seitz F.; BASF Aktiengesellschaft, DE 4342048. Mitsubishi Rayon, JP 05 339 324. Mitsubishi Rayon, JP 08041 149. Craig D., Hu R.; General Electric Co., WO 200034346. Kato T., Izumi M., Hayashiba Y., Suenaga K., Othake H.; Mitsubishi Rayon Co., DE 2 037 419 (1970). Schirmer H., Peilstocker G., Schuster H.; Bayer AG, DE 2259564 (1972). Yu A. J.; Stauffer Chem. Co., US 4148 842 (1978). Brandstetter F., Hambrecht J., Muenstedt H.; BASF Aktiengesellschaft, DE 3149812(1981). Mitulla K., Swoboda J., Schmitt B., Wassmuth G.; BASF Aktiengesellschaft, EP 111260(1982). Chen C., Peng F. M.; Bayer AG, WO 200060007 (1999). Gilles H. F., Sasserath J. N.; General Electric Co., US 4579909 (1984). Peascoe W. J.; General Electric Co., EP 269861 (1986). McHale A. H., Peascoe W. J.; General Electric Co., EP 272425 (1986). Mitulla K., Hambrecht J., Echte A., Swoboda J., Siebel P., Schwaab J., Frank H.; BASF Aktiengesellschaft, EP 164513 (1984). Wassmuth G., Ruppmich K., Seiler E., Gausepohl H., Benker K.; BASF Aktiengeselschaft, EP 244856 (1986). Wassmuth G., Ruppmich K., Seiler E.; BASF Aktiengesellschaft, EP 244857 (1986). Fujiguchi T., Saito A., Itoi H.; General Electric Co., EP 663425 (1994). Udipi K.; Monsanto Co., EP 440007 (1989) Mohd Ishak Z. A., Ishiaku U.S., Karger-Kocsis J. J. Appl. Polym. Sci. 74, 2470 (1999). Talibuddin S. H., Sastri V.R., Mercx F., Cheret E., Gallucci R.; General Electric Co., WO 200046296 (1999). Cheret E., De Vries R., Mercx F., Kwiecinski V.; General Electric Co., WO 200042105 (1999). Niessner N., Ruppmich K., Mosbach N.; BASF Aktiengesellschaft, EP 603674 (1992). Osima, A., Casale, A., Orsatti, E. Dakli, I.; Montecatini Societa Generate, DE 1569194 (1964). Stein D., Haaf F., Priebe E.; BASF Aktiengesellschaft, DE 2140437 (1971). Jansen U, Ott K.-H., Suemmermann K., Frohberg E.; Bayer AG, EP 330038 (1988). Robison J. C., Ziegelmeyer T.A.; General Electric Co., EP 415 120 (1989). Mitsubishi Rayon Co., JP 08 199025 (1995). Lindner C., Braese H.-E.; Bayer AG, DE 4110975 (1991). Guentherberg N., Deckers, A., Niessner N.; BASF Aktiengesellschaft, DE 4 229 913 (1992). Lindenschmidt G., Ruppmich K., in Proceedings of the Conference 'ABS and Related Polymers in the Automotive Industry', Sueddeutsches Kunsststoff-Zentrum, Wurzburg, Sept. 28–29, 1993.

362

G. E. McKEE ET AL.

92. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630099. 93. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630063. 94. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630142. 95. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630143. 96. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630144. 97. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630117. 98. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630120. 99. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630103. 100. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630061. 101. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630135. 102. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630098. 103. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630095. 104. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630118. 105. Naarmann H., McKee G. E., Pirker A., Sterzel H.-J., Brandstetter F., Bernstorff B.-S., Rosenau B., Endemann U., Straube B.; BASF Aktiengesellschaft, DE 19630097. 106. Treede H. J., Euro Security No. 6, 31 (1998). 107. A. Grefenstein, Metalloberflache 53 (Oct.) 2 (1999).

P A R T IV

Syndiotactic Polystyrene

This page intentionally left blank

17

NORIO TOMOTSU Polymer Research Laboratory, Idemitsu Petrochemical Co., Ltd, Chiba, Japan

MICHAEL MALANGA R&D Engineering Plastics, The Dow Chemical Company, Midland, Ml, USA

JUERGEN SCHELLENBERG R&D Engineering Plastics, Dow Central Germany Schkopau, Germany

1

INTRODUCTION

Polystyrene was commercialized by I. G. Farben in 1931 and it has long been used as a commodity plastic. Although polystyrene is endowed with excellent properties not found in other commodity plastics such as polyolefins, its amorphous nature (relatively low heat and solvent resistance) limits its use in some application areas. Karl Ziegler first discovered in 1953 that transition metal compounds could be activated by aluminum alkyls and used as organometallic catalysts to polymerize ethylene. Giulio Natta discovered stereoregular polymers such as isotactic polypropylene and isotactic polystyrene (IPS) were also prepared using this same family of new catalysts [1,2]. IPS is a semi-crystalline polymer with a melting point of ~ 240 °C. Some companies have tried to commercialize IPS with the idea that it should be a plastic with higher heat resistance. Unfortunately, the crystallization rate of IPS is too slow to be practical in injection molding. One of the most important recent achievements in the field Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy r 2003 John Wiley & Sons Ltd

366

T. NORIO ET AL.

of polymerization catalysts has been the introduction of methylaluminoxane (MAO) by Sinn and Kaminsky [3,4]. The controlled reaction between water and trimethylaluminum produces MAO. The polymerization activities of metal compounds with MAO are higher than those attainable with traditional alkylaluminum compounds. Additionally, this soluble catalyst system can be more easily used to control the stereoregularity of produced polymers by varying the ligand structure of the metallocene. Ishihara et al. first succeeded in the synthesis of syndiotactic polystyrene (SPS) in 1985 [5–7]. With a melting point of 270 °C and a crystallization rate much faster than that of IPS, SPS has been commercialized as a new Engineering Plastic. The detailed physical and mechanical properties are described in Chapter 18. The Dow Chemical Company and Idemitsu Petrochemical Co. Ltd started joint research work on SPS in 1988 and have together vastly improved the catalyst system, polymerization technology, manufacturing process and application areas for this new material. The first commercial plants for the production of SPS were built in Japan in 1996 and in Germany in 1999. Both companies supply SPS products to the plastics industry (tradenames: XAREC1 from Idemitsu Petrochemical Co. Ltd and QUESTRA* from Dow Chemical Co.). One of the most important issues for the commercialization was cost reduction for the production. Idemitsu Petrochemical Co. Ltd and Dow Chemical Co. succeeded in reducing the production cost by catalyst activity improvement, polymerization condition optimization and process development.

2 CATALYTIC SYSTEMS FOR SPS Since 1985, many different transition metal compounds have been examined for their ability to produce syndiotactic polystyrene in combination with counterions based on methylalumoxane, borane, borate and other chemicals. 2.1 2.1.1

TRANSITION METAL COMPLEXES Metal

Typical transition metal complexes investigated are summarized in Table 17.1 together with the polymerization conditions, the polymer properties, and the catalytic activities. Yang et al. examined rare earth coordination catalysts. The Nd(naph)3/ Al(iBu)3 catalyst system was found to produce syndiotactic-rich polystyrene [8]. They proposed that the catalytically active species might be an ionic complex because the addition of CCU increased the catalytic activity.

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

367

Table 17.1

Polymerization of styrene using various transition metal compounds

Compound

Metal MAO Conversion (mmol) (mmol) (%)a

TiCl4 Ti(OMe)4 CpTiCl3 Ti(acac)2Cl2 ZrCl4 Zr(CH2Ph)4 VOC13 Nb(OEt)5 Ta(OEt)5 Cr(acac)3 MoO(acac) Fe(acac)3 Co(acac)3 Ni(acac)2

0.05 0.05 0.05 0.01 0.05

0.2 0.05 0.25 0.25 0.02 0.02 0.02 0.02 0.25

40 40 40 8 10 16 40 20 20 10 10 10 10 20

4.1 3.8 92.3

0.4 0.7 2.0 0.2 0.2 Trace

1.4 0.5 0.5 1.8 80.8

syndiotactic syndiotactic syndiotactic syndiotactic atactic syndiotactic atactic atactic atactic atactic atactic atactic atactic atacticc

Stereospecificity polymerization conditionsb

1 1 2 1 3 5 4 1 1 4 4 4 4 4

a

Conversion from styrene to polymer. Polymerization conditions:styrene:toluene (ml/ml): (1) 180:100; (2) 23:50; (3) 100:50; (4) 50:100; (5) 40:90. Polymerization temperature and time: (1)–(4) 50 °C, 2h; (5) 90 °C, 4h. c Iso-rich polymer. b

Recently, Wakatsuki and co-workers have shown that samarium compounds produce SPS with lower syndiotacticity than titanium compounds [9]. Group 4 transition metal complexes showed higher activity and higher syndiospecificity than the other metal complexes. The ansa-zirconocene compounds show lower activity and lower syndic-directing Stereospecificity than the corresponding ansa-titanocenes. Zambelli and co-workers also found that Zr compounds [e.g. Zr(CH2C6H5)4, Zr(C7H8)2, ansa-Cp2ZrCl2] catalyze the syndiospecific styrene polymerization [10–14]. Among Group 4 transition metals, titanium compounds show the highest performance.

2.1.2

--Ligand

The polymerization activities of bis-cyclopentadienyl titanium compounds are lower than those of bridged bis-cyclopentadienyl titanium compounds. Miyashita et al. reported the polymerization activities of several bridged bis-cyclopentadienyl titanium compounds [15]. They found that the catalytic activity of CH2(Cp)2TiCl2 is the highest among bis-cyclopentadienyl titanocene compounds. The data indicate that the polymerization activities and also syndiospecificity increase with a decreasing angle between the Cp centroid—Ti—Cp centroid in bis-cyclopentadienyl titanocene compounds.

368

T. NORIO ET AL.

Among the titanium complexes producing SPS, monocyclopentadienyltitanium compounds show the highest polymerization activities and highest syndio-directing stereospecificity as compared with non-cyclopentadienyl, biscyclopentadienyl and bridged bis-cyclopentadienyl titanium complexes. A comparison of the polymerization behavior of CpTiCl3 and [Me5C5]TiCl3 (Cp*TiCl3) under equivalent experimental conditions shows that the Cp* derivative gives a higher degree of syndiotacticity in the resulting polymer, much higher molecular weights and predominantly an increased polymerization activity. The structure of the cyclopentadienyl ligand of the metal complex has been extensively varied. Some of the investigated complexes with substituted monocyclopentadienyl ligands are shown in Table 17.2 and Figure 17.1. The data indicate that substituents on the cyclopentadienyl ligand which are electron releasing generally yield higher polymerization activities. This result suggests stabilization of the active site by electron-releasing substituents. The catalytic activities of bulky substituents such as tBu or SiMe3 are lower than is expected from the electron density of titanium. These bulky groups are believed to sterically hinder the coordination of styrene monomer to the metal catalyst. Recently, Aoyama et al. have shown that Cp with a side ring ligand such as 4,5,6,7-tetrahydroindenyltitanium trimethoxide led to good activity [16]. The strain of the side ring increases the electron donation of the Cp ligand and probably improves the catalyst activity of titanocene. 49Ti NMR (49TiCl4 as standard) shows an electron density on titanium of [l,2,3-Me3-5,6,7,8-tetrahydroindenyl]TiCl3 of –95.3ppm, which is higher than that of Cp*TiCl3 (—80.7 ppm). The polymerization results are shown in Figure 17.2. The catalytic activity of [1,2,3-Me3-5,6,7,8-tetrahydroindenyl]TiCl3 is higher than that of Cp*TiCl3. Ready et al. [17] observed that IndTiCl3 is a significantly better catalyst than CpTiCb. Takeuchi et al. [18] also examined similar compounds at almost the same time, but their performance is lower than that of CpTiCl3. The difference in these results might be due to the difference in polymerization conditions. Takeuchi's et al. results suggest that substituents on the five-membered ring generally increase the catalytic activity among indenyl ligands. Table 17.2

Relationship between catalyst structure and catalytic activity

Compound

Catalytic activity (kg/g Ti)

(EtMe4Cp)Ti(OMe)3 Cp*Ti(OMe)3 [(tBu)2Cp]Ti(OMe)3 (Me4Cp)Ti(OMe)3 [(Me3Si)2Cp]Ti(OMe)3 CpTi(OMe)3

270 265 15 150 30 10

1

H

NMR: (OMe) (ppm) 4.045 4.050 4.072 4.081 4.088 4.112

369

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE 300

250 S 200

•>

150

100

U

50 01— 4.04

4.06

4.08

4.10

4.12

Chemical shift of MeO/ppm Figure 17.1 Relationship between chemical shift of methoxy group and catalytic activity 80 70

^ 60 .1 50 £ 40 c o U 30 20 10

0 50

60

70

80

Temp.fC) Figure 17.2 Catalytic activity of (•) [1,2,3-Me3-5,6,7,8-tetrahydroindenyl]TiCl3 and (+)Cp*TiCI3

2.1.3

a-Ligand

The polymerization activities in the presence of Cp*Ti compounds containing different ancillary ligands, i.e. alkoxide and chloride ligands, with MAO are as follows in order of decreasing catalytic activity [19]: Cp*Ti(OiPr)3 and

370

T. NORIO ETAL.

*Ti(OMe)3 > Cp*Ti(OPh)3 > Cp*Ti(OC6H4CH3)3 > Cp*TiCl3 > Cp*Ti(Oi C3HF6)3. Kaminsky showed that the catalytic activity of CpTiF3 is better than that of CpTiCl3 [20]. The chloride ligand and the electron-withdrawing alkoxide, OiC3HF6, decrease the conversion. The catalyst activity of all these systems is increased by the addition of triisobutylaluminum (TIBA) to the MAO—Ti complex mixture. By the addition of TIBA, all a-ligands are found to be substituted and the activity of the final catalyst system is practically the same. The ESR spectra of mixed solutions of Cp*Ti compounds, MAO and TIBA supported this hypothesis (Figure 17.3) [21]. 2.1.4

Other Catalysts

As a catalyst with a noncyclopentadienyl ligand, Kakugo et al. [22] examined the performance of bridged bisphenolato titanium complexes [(OC6H4-4-CH3-6tC4H9)2ZTiX2]. The catalyst activity of [(OC6H4-4-CH3-6-tC4H9)2S]Ti(OiPr)2 is higher than that of [(OC6H4-4-CH3-6-tC4H9)2CH2]Ti(OiPr)2 [22]. Okuda and Masoud showed that the catalytic activity depends on the nature of the bridging group Z and increases in the order—CH2_ < —CH2CH2_ < —S— < —SO—[23]. A titanium complex with a pyrazolylborate ligand was studied by Campbell and Malanga [19]. Similarities between the cyclopentadienyl ligand and the hydridotris(pyrazolyl)borate ligand have been noted for transition metal complexes. Catalyst efficiencies are much lower than those of the analogous pentamethylcyclopentadienyl complexes. 2.2 2.2.1

CO-CATALYSTS MAO

MAO is a useful co-catalyst for titanium metal complexes in the syndiotactic polymerization of styrene. Its complete structure and role are as yet not Cp*Ti(OMe)3

3315

3340

3365 G

Figure 17.3

ESR spectrum of Cp*TiR3 with MAO and TIBA

371

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

clarified. Several types of MAO with different catalytic activities were produced by different synthesis conditions and the performance evaluated. MAO is produced by the reaction between trimethylaluminum (TMA) and water. The reaction is controlled by the reaction temperature. Small amounts of TMA always exist in MAO. The concentration of TMA in MAO solution has been measured by a titration method, but the reproducibility of the measurement is low. An improved method was reported using 1 HNMR of an MAO solution in toluene in the presence of a small amount of dioxane (Figure 17.4) [21]. This method is easier and more reliable than the titration method. The relationship between catalytic activity and TMA content in MAO is shown in Figure 17.5. Residual TMA in the MAO decreases the catalytic activity. The reactivity of TMA is too strong to control the active site formation and decomposes the active sites for polymerization. The MAO solution was 'dried' in many attempts to remove small amounts of TMA. It was found that not all the TMA could be removed under any conditions. Some amount of TMA is thought to coordinate strongly with MAO and seems to effect changes in the reactivity of MAO.

-0.5

-1.0

Shift/ppm

Figure 17.4

1

H NMR spectrum of MAO in dioxane

-1.5

T. NORIO ET AL.

372

10

20

30

40

50

60

TMA contents/% Figure 17.5 Effect of TMA in MAO on the catalytic activity. Polymerization conditions: 70 °C, 1 h polymerization, Cp*TiOMe3/MAO

MAO is an oligomer produced by the reaction between TMA and water. The molecular weight of MAO has been measured by a cryoscopic method. Dioxane can be used as a solvent for this method. The relationship between catalytic activity and molecular weight of MAO is shown in Figure 17.6. The catalytic activity showed the maximum around a molecular weight of 400 g/mol. It seems probable that a rapid alkylation of the metallocene by MAO takes place first, and the active species arises from a methyl transfer reaction between the metallocene alkyls and MAO. The active species formed from these reactions are probably complexed by a bulky MAOcoordinating non-quenching anion.

2.2.2

Borate

Some types of borate compounds act as co-catalysts for the syndiospecific polymerization of styrene in these catalyst systems. The active borate compounds have a tetraphenylborate anion. The effect of anions on the catalytic activity is summarized in Table 17.3. Fluorine substituents at the of 3,4,5positions increase the catalytic activity and tetrapentafluorophenylborate showed the highest performance. The cations formed react with the titanium complex and form active sites. The reaction equations are as follows: Cp*TiR3 + [RjNH][(C6F5)4B] -+ [Cp*TiR2]+[(C6F5)4B]

RH

373

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

100

300

400

600

500

Mw

Figure 17.6 Effect of molecular weight of MAO on catalytic activity. The molecular weight of MAO was measured by a cryoscopic method. Dioxane was used as solvent Table 17.3

Effects of fluorine in borate

Borate

Catalytic activity (kg/g Ti)

[Me2HNPh][(C6H5)4B] [Me2HNPh][(C6H4F)4B] [Me2HNPh][(2,4-F2C6H3)4B] [Me2HNPh][(l, 3,4-F3C6H2)4B] [Me2HNPh][(C6F5)4B] [Me2HNPh][(3-CF3C6F4)4B]

0 0 5 10 250 20

Cp*TiR3 + [RJN][(C6F5)4B] -> [Cp*TiR2]+[(C6F5)4Br + RjN + RR Cp*TiR3 + [R^C][(C6F5)4B] -» [Cp*TiR2]+[(C6F5)4B]- + R3CR The use of borate with a small amount of TIB A as co-catalyst for polymerization of styrene to SPS was examined by Campbell and Malanga [19], Tomotsu [21] and Kucht et al. [24]. TIB A was found to be a good scavenger of impurities in styrene and to increase the syndiotacticity of the resultant polymer. The effects of cations on the catalytic activity are summarized in Table 17.4. Ammonium borate compounds with a lower pK a , such as [2-CN-pyridine(N)Me][B(C6F5)4] (pKa = —0.3), showed higher activity than those with higher pK a . The by-products of the active site formation reaction are thought to coordinate strongly to the active site. On the other hand, the by-product of the reaction between titanium compounds and [Ph3C][B(C6F5)4] is Ph3CH and

T. NORIO ET AL.

374

does not coordinate to the active site. In this case, however, the [Ph3C][B(C6F5)4] reacts with TIBA and is decomposed. [Ph(PhOMe)2C][B(C6F5)4] and [(PhOMe)3C][B(C6F5)4] are more stable compounds and do not react with aluminum alkyls. Lower reactivity against aluminum alkyls results in an increase in the apparent catalytic activity (Figure 17.7). The decomposition of the borate by TIBA is observed by 1H NMR and the excess amount of borate increases the catalyst activity.

2.2.3

Supported and Heterogeneous Catalysts

It has been demonstrated that mixtures of highly isotactic and highly syndiotactic polystyrene are obtained when using titanium compounds such as TiCl3 or TiCl4 supported on Mg compounds in the presence of MAO [25] (Table 17.5). In this situation, the two types of polymer polymerize simultaneously from two Table 17.4

Effects of pyridinium cations in borate

Borate

pKa

Catalytic activity (kg/g Ti)

[PyMe][(C6F5)4B] [4-CN-PyMe] [(C6F5)4B] [3-CN-PyMe] [(C6F5)4B] [2-CN-PyMe] [(C6F5)4B]

5.2 1.9 1.0 -0.3

2.0 2.8 31.8 40.5

[(MeOPh)3C]+

>-^

[(MeOPh)2PhC]+

40

Figure 17.7

60 80 TIBA/Ti

100

120

Catalytic activity of Cp*TiOMe3 with carbenium borate

375

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE Table 17.5

Polymerization by supported catalysts

Catalyst

Ti (mmol)

TiCl3(AA)

1.0 0.2 1.0 0.2 2.0 0.02 0.02 40 5 0.2 2.0 2.0 0.2

TiCl3 (Solvay) Mg(OEt)2/EB/TiCl4 TiCl4 Ti(OEt)4

a

Al/Ti

100 1000

20 1000

50 500 1000 10 40 500 10 50 500

a

Conv. (%)

Stereospecificity

8.2 2.0 1.9 0.9 2.9 1.1 1.4 7.2 0.4 0.7 0.3 2.5 0.9

Mixture of IPS and SPS Mixture of IPS and SPS Mixture of IPS and SPS Mixture of IPS and SPS IPS (84%) + SPS(16%) IPS(12%) + SPS (88%) IPS(10%) + SPS (90%,) IPS Mixture of IPS and SPS SPS Atactic PS SPS SPS

Polymerization conditions: styrene 50 ml, toluene 100ml, 50 °C, 2h.

different stereospecific active sites. Toluene-soluble titanium catalysts with MAO produce the syndiotactic polystyrene and isotactic polystyrene was produced by the heterogeneous titanium catalyst on the Mg support. The amount of titanium soluble in toluene was increased with increasing addition of MAO. Soga and Nakatani [26] examined Ti(OC4H9)4 supported on SiO2 with MAO and a reacted mixture of Ti(OC4H9)4 and MAO supported on SiO2. The syndiotacticity of the polymers with both catalysts was almost 100%. They found that the catalytic activity was independent of Al/Ti molar ratio. The yields in this case were very low. They suggested that SiOTi(OC4H9)3 heterogeneous species are more stable against reduction than the active species in the soluble system.

3 3.1

COPOLYMERIZATION POLYMERIZATION OF SUBSTITUTED STYRENES

When various ring-substituted styrenes were polymerized using CpTiCl3 with an MAO catalyst system, the corresponding syndiotactic polystyrenes were obtained. 13CNMR spectra of the phenyl C-l carbon of the poly(ringsubstituted) styrenes poly(p-methylstyrene), poly(m-methylstyrene), poly (p-tert-butylstyrene), poly(p-chlorostyrene), poly(m-chlorostyrene) and poly(p-fluorostyrene) were examined [27]. The spectra of each atactic poly (ring-substituted) styrene show many main peaks corresponding to their various

376

T. NORIO ETAL

configurational sequences. These spectra are similar to that of atactic PS. The spectrum of each syndiotactic poly(ring-substituted) styrene shows a single sharp peak at a higher magnetic field corresponding to the rrrr pentad configuration. These spectra are similar to that of SPS. In poly(p-fluorostyrene), the peak at high magnetic field is split, owing to coupling to the 19F nucleus. Soga and co-workers examined the relation between the Hammett's a value of each substituent and the reactivities in copolymerization (Figure 17.8). It is observed that the monomer reactivity is enhanced by electron-releasing substituents in the aromatic ring. Even p-ter/-butylstyrene with a substituent of large steric hindrance shows a high reactivity. This indicates that there is a strong polar effect of the substituent on the rate of addition [28]. Similar effects were observed by Ishihara et al. in the homopolymerization of these monomers [27]. In the syndiospecific polymerization of styrene, the monomer addition has been demonstrated as a secondary addition mechanism. Additionally, the monomer reactivity is enhanced by electron-releasing substituents in the aromatic ring. The electronic effect of the substituent of the ring is transmitted more efficiently to the methine than to the methylene carbon. Styrene and alkylstyrenes form co-syndiotactic polymers. Almost all of the syndiotactic polymers produced have a high degree of crystallinity, and all the Tm values of syndiotactic polymers are higher than those of isotactic polymers with the corresponding substituent. These results imply that the syndiotactic polymers might be useful in industry. 0.4

p-MeSt 0.2

p-BrSt o -0.2 m-ClSt -0.4 -0.2

Figure 17.8 merization

-0.1

0.0

0.1 a

0.2

0.3

0.4

Hammett's a value for substituted styrene and reactivity ratio of copoly-

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

3.2

377

COPOLYMERIZATION OF STYRENE AND ETHYLENE

Mani and Burns [29] examined the copolymerization of ethylene and styrene by using TiCl3 with MAO. The polymer obtained by this catalyst system is isotactic polystyrene. There is no ethylene-styrene bond in the polymer. Tazaki [30] reported that a copolymer, which was obtained from ethylene and styrene using monocyclopentadienyl- or tetraalkoxytitanium compounds with MAO, had a small number of ethylene units in the polymer backbone. Seppala and coworkers [31] also examined the copolymerization of styrene and ethylene using titanium compounds with bulky alkoxy ligands and also CpTiCl3 with MAO. The products of the polymerization were a mixture of polyethylene and syndiotactic polystyrene. This suggests that there are more than two kinds of active sites in this catalyst system active for ethylene and styrene independent of each other. Pellecchia et al. [32] examined Ti(benzyl)4 with borane for producing ethylene-styrene copolymer, and it was reported that all sequences showed an alternating copolymer. Inoue and Shiomura [33] examined zirconium complexes for the ethylenestyrene copolymerization. They found alternating copolymer and homopolymer. The stereoregularity of the polymers was atactic. Kakugo et al. [22] reported that a catalyst based on 2,2'-thiobis(4-methyl-6tert-butylphenoxy)titanium dichloride [(TBP) TiCl2 and MAO produced a mixture of syndiotactic polystyrene and the alternating ethylene-styrene copolymer. They stressed the role of sulfur as essential to obtain the alternating copolymer. They reported that the copolymer was not produced using a similar compound having a methylene bridge [2,2/-methylenebis(4-methyl-6-ferf-butylphenoxy)titanium dichloride] instead of a sulfur bridge in the catalyst ligand. Subsequently a patent disclosed ethylene-styrene copolymerization promoted by catalysts based on bridged amidomonocyclopentadienyltitanium complexes such as [dimethylsilyl(phenylamido)(Cp*)]titanium dichloride and MAO. The copolymer obtained was a random copolymer but it did not contain any regioregularity in the arrangement of the styrene-styrene sequences. Suzuki and co-workers copolymerized styrene and ethylene by ansa-zirconocene compounds [34]. The polymer produced was isotactic polystyrene with ethylene. They found that the phenyl group of the monomer coordinates to the active site and decreases the catalytic activity.

3.3

COPOLYMERIZATION OF STYRENE AND DIENES

Pellecchia et al. copolymerized isoprene and styrene [35] and examined the copolymerization rate. They found a value for the product of the reactivity ratios of r1r2 — 2.3. The difference in the catalytic activity of styrene and isoprene may be due to the difference in coordination strength.

378

4

T. NORIO ETAL

MECHANISMS OF POLYMERIZATION OF STYRENE

Zambelli et al. reported on the mechanism of styrene polymerization [36]. They showed that the main chain of the syndiotactic polymer has a statistically transtrans conformation. It was established then the double-bond opening mechanism in the syndiospecific polymerization of styrene involves a cis opening. The details in the control of the monomer coordination for this polymerization mechanism were examined by Newman and Malanga using detailed 13C NMR. It was shown through the analysis of tacticity error (rmrr) that the tacticity in the polymer is chain-end controlled and that the last monomer added directs the orientation and coordination of the incoming monomer unit prior to insertion [37].

4.1

ACTIVE SITE SPECIES

The polymerization activity of a titanium catalyst increases with an increasing molar ratio of MAO to Ti. The amount of cationic Ti(III) species measured by ESR also increases with increasing ratio of MAO to Ti. This suggests that MAO acts as a reducing agent for Ti(IV) to Ti(III). Cationic Ti(III) might be an active species in the synthesis of SPS. Newman and Malanga [38] synthesized Cp*Ti(OMe)2 via reduction of Cp*Ti(OMe)3 and found that the catalytic activity of this complex with smaller amounts of MAO is almost the same as that of Cp*Ti(OMe)3. Metal alkyl compounds are known to reduce titanium compounds. The effect of reductants on the catalytic activity was evaluated and the data are summarized in Table 17.6 [21]. Very strong reducing reagents such as A1(CH3)3 and A1(C3H5)3 have a negative effect on the catalytic activity. In this case the titanium compound is reduced to Ti(II) by these reagents. As discussed earlier, however, the catalytic activity is significantly increased by the addition of TIBA. The details of the effect of TIBA were evaluated and are shown in Figure 17.9 [21]. TIBA increases the catalytic activity and reduces the molecular weight of the polymer. TIBA reacts with the titanium compound and reduces its oxidation state. Moreover, TIBA reacts with metal alkyl bonds and therefore acts as a chain transfer agent during polymerization by insertion and a new polymer end is started. The polydispersity (M w /M n ) produced by the titanocene catalyst with MAO is always found to be about 2.0. These results show that there is only one active site and that chain transfer occurs during the polymerization. The numbers and the kind of active sites were clarified by the stopped flow method. Without a chain transfer reaction, the polymerization proceeds like a 'living polymerization' and is an easy to analyze reaction. /j-Methylstyrene (PMS) was used in this experiment to avoid plugging during polymerization due to crystallization. The

379

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

Table 17.6 Polymerization activities of [Me5Cp]TiCl3 with MAO and alkylation reagents Reagent

Relative activity"

Mw

None A1(CH3)3 A1(C2H5)3 Al(nC4H9)3 A1(iC4H9)3 Al(nC8H17)3 A1(C2H5)7(OC2H5) Zn(C2H5)2

100 13 23 76 560 100 140 48

750000 64000 84 000 570000 580000 670000 870000 130000

a

Catalyst activity without alkylation reagent as 100

300

200

2 3 T1BA/MAO

Figure 17.9

Effect of TIBA on the catalytic activity

results are shown in Figure 17.10. The Mw/Mn of the polymer was about 1.0. This strongly indicates that only one kind of active site is responsible for the polymerization of the styrene. The number of polymer chains indicated about half of the added titanium compound polymerize styrene. By the addition of hydrogen to PMS, the number of chains increased without any broadening of the molecular weight distribution. Hydrogen did not act as a chain transfer reagent but the number of polymer chains showed that almost all Ti compounds turned to active sites. After the polymerization with hydrogen, ethylmethylbenzene was observed as a by-product. The amount of ethylmethylbenzene is almost equimolar to the hydrogen added. These results showed that almost all of the titanium compounds turned to active sites, but about half of the active sites become dormant by some type of coordination with

380

T. NORIO ETAL

monomer. The addition of hydrogen activates this dormant site during polymerization. The structure of the active site is proposed in Figure 17.11 (a). The structure of the dormant sites might be an irregular coordination of the monomer [Figure 17.11(5)] or a change of the direction of the monomer coordination [Figure 17.11(c)]. The dormant site might be reactivated by the hydrogenation of styrene which irregularly coordinated to the active site. The misinsertion of styrene and the stereoirregularity are difficult to observe. We suppose that the polymerization reaction was stopped after an irregular coordination of styrene. This also supports the chain-end controlled mechanism of stereospecificity.

o

1.0

o

0

0

With hydrogen

0.8 o

Mw/Mn=1.05~1.15

H 0.6 c

•6 0.4

Without hydrogen

0.2 0.0

Figure 17.10 MAO

0.2

0.4 0.6 time/sec

0.8

1.0

Stopped flow polymerization of p-methylstyrene by Cp*TiOMe3 with

(c)

Figure 17.11 Proposed structure of active site and dormant site for styrene polymerization (a) Active site; (b) dormant site by the coordination error of monomer; (c) dormant site by the polymer chain rotation

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

381

The active site structure was determinied by XANES or EXAFS because all of the titanium changed to active site. Figure 17.12 shows the results of these measurements. A sharp peak was observed in XANES but it disappeared on the addition of TIBA, MAO or borate. TIBA, MAO or borate changes the ligand coordination structure of the titanium complex. The position of the edge shifted to the left shows that the tetravalent titanium was reduced to trivalent titanium. The electron density of titanium was reduced by TIBA and MAO or borate and the cyclopentadienyl ligand came closer to the titanium. The structures of active sites produced using either MAO or borate are therefore thought to be almost the same. EXAFS shows a new signal outside the cyclopentadienyl ligand. The structure of the active site was examined by molecular modeling calculations. The reaction between styrene and CpTi(CH3)[CH(C6H5)CH3]+, CpTi(CH3)[CH (C6H5)CH3] and CpTi[CH(C6H5)CH3]+ was examined. The coordination of the monomer to Ti(III) is more stable than to Ti(IV) because the coordination of the vinyl group is fortified by the back-donation from the d-orbital to the vinyl group of styrene. We calculated the Ti(III) cation as an active site. This structure was proposed by Zambelli et al. [36] and involves a Ti(III) cation additionally coordinated to the phenyl group existing at the end of the growing polymer chain in a multi-hapto manner and forming a stabilized structure. However, this structure seems to be too stable to insert styrene by our calculation. The activation energy of the reaction of styrene and CpTi(CH3) [CH(C6H5)CH3] was 9kcal/mol and that of styrene and CpTi[CH(C6H5) CH3]+ was 28 kcal/mol. Therefore, we concluded that the active site is a neutral trivalent titanium instead of a trivalent titanium cation with a coordinated phenyl group. We suppose that MAO or borate exists near the active site and neutralizes the active site to prevent the coordination of the phenyl group. OMe or R

+TIBA+MAO or DMAS

Figure 17.12

XANES and EXAFS results for active sites

382

4.2

T. NORIO ETAL

KINETIC ANALYSIS OF STYRENE POLYMERIZATION

We assumed that the polymerization proceeds by a normal coordination polymerization. The effect of the catalyst concentration on the polymerization was examined by polymerization at different ratios of catalyst to styrene (Figure 17.13). The reaction rate increased in proportion to the catalyst ratio. However, the decay of the polymerization reaction was too fast to explain it as a firstorder reaction. We hypothesized that this polymerization proceeds by a single-site catalyst under different morphological conditions and variable monomer concentrations, i.e. polymerization in the crystalline polymer and in the amorphous polymer state. The frequency factor of the polymerization in the crystalline polymer should be lower than that in the amorphous polymer. The effect of monomer concentration on the polymerization rate is shown in Figure 17.14 using toluene as a solvent. The reaction rate is proportional to the monomer concentration. From these results, the polymerization reaction can be described by the following equations:

d[M]_ ~d0~~~*

conv. = 1 -exp^ —^-—— [1 -exp( - &<j,0)] — , 100

60

120

180

20

[ 1 -exp( -

240

Time/min Figure 17.13

Polymerization of styrene with different concentrations of catalyst

383

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE Conv/%

T>

40 30 20 10 0

20

40

60

8(

Monomer concentration/%

Figure 17.14

Effect of monomer concentration on the catalytic activity

where [M] = monomer concentration; 9 polymerization time; iq] = concentration of active site at higher monomer concentration; [qi = concentration of active site at lower monomer concentration; chain propagation rate constant of active site 1 , including effects of monomer concentration around the active site; chain propagation rate constant of active site 2, including effects of monomer concentration around the active site; decay rate constant of active site 1; decay rate constant of active site 2; *P,[C1*]0 =0.502exp( MQo = 0.050 exp ( - 894/f)[Ti]; fcdJCHo = 699000 exp ( - 5300/0; kd2[C*2]0 = 49400 exp ( - 5300/OUi]; t = polymerization temperature; [Ti] = titanium concentration. The polymerization rate constants were measured by adjustment of the equations above to the experimental polymerization results. The activation energy of chain propagation and decay for active site 1 and 2 were the same because this catalyst system is single site. The results of the fitting of calculation and polymerization are shown in Figure 17.15. The chain transfer reaction was also examined by these equations and the comparison of the calculated predictions and experimental results is shown in Figure 17.16.

384

T. NORIO ETAL

60

180

120

240 Time/min

Figure 17.15 Effect of polymerization temperature on catalytic activity. Line, calculated by kinetic equation; symbols, polymerization results Mw/1000 2000 1500 1000 500

0 50

60

70

80

90

Figure 17.16 Effect of polymerization temperature on Mw. Line, calculated by kinetic equation; symbols, polymerization results Mn = M

ka1 [C*][TIBA] + ka2[C*][MAO] where Mn = number average molecular weight; rp = poljonerization rate; rt — chain transfer rate; kt = termination rate constant;

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

385

ka1 = chain transfer reaction rate constant of TIBA; ka2 = chain transfer reaction rate constant of MAO; kt = 1.19x 1010exp (-8157/r); ka1 = 1.64 x 10 7 exp( - 6860/0; ka2 = 1.19 x 108 exp(- 8615/0MAO and TIBA act as chain transfer agents, whereas ß-hydride elimination was not the main reaction for chain transfer with this catalyst system. Murata et al. [39] showed that a living polymerization was observed in low- temperature polymerization using this catalyst system. This result is in agreement with this model.

4.3

EFFECTS OF HYDROGEN

Triisobutylaluminum (TIBA) is a good activation agent for styrene polymerization: Cp*TiR3 + TIBA + MAO -> Cp*TiR(iBu) -> Cp*TiR(H) + (CH3)2C=CH2 As in the previous section, hydrogen increases the concentration of the active sites. Moreover, we think that the polymerization is preceded by the formation of a titanium hydride complex. If this is correct, hydrogen also increases the overall catalytic activity. The effects of hydrogen pressure on catalytic activity and molecular weight of SPS are summarized in Table 17.7. Hydrogen increases the catalytic activity and the polydispersity. The change in the molecular weight distribution might be the effect of the lack of uniformity of the polymerization system or the result of the formation of new types of active sites. Table 17.7

Effects of hydrogen pressure on styrene polymerization"

Hydrogen pressure (MPa)

Relative activity6

Mw/Mn

None 0.010 0.049 0.098

100 160 210 220

2.1 2.5 4.5 12.0

a b

Polymerization conditions: styrene bulk polymerization, 70 °C, 1 h. Catalyst activity without hydrogen as 100.

386

5

T. NORIO ETAL

CONCLUSION

Titanium compounds with MAO or borate as co-catalysts effectively produce syndiotactic polystyrene from styrene monomer. The design of high-performance catalyst systems is now well demonstrated. The basic structure of the active site, the mechanism of coordination and insertion and the kinetics are also now well understood for this new polymerization.

REFERENCES 1. Natta, G., Pino, P., Corradini, P., Danusso, F., Mantica, E., J. Am. Chem. Soc., 77, 1708 (1955). 2. Natta, G., Pino, P., Mantica, E., Danusso, F., Mazzanti, G., Peraldo, M., Chim. Ind. (Milan), 38, 124 (1956). 3. Sinn, H., Kaminsky, W., Vollmer, H.-J., Woldt, R., Angew. Chem., 92, 396 (1980). 4. Sinn, H., Kaminsky, W., Adv. Organomet. Chem., 18, 99 (1980). 5. Ishihara, N., Kuramoto, M., Uoi, M., to Idemitsu Kosan Co. Ltd., JP 62187708 (1985). 6. Ishihara, N., Kuramoto, M., Uoi, M., to Idemitsu Kosan Co. Ltd., EP 210615 (1986). 7. Ishihara, N., Kuramoto, M., Uoi, M., Macromolecules, 21, 3356 (1988). 8. Yang, M., Cha, C., Shen, Z., Polym. J., 22, 919 (1990). 9. Zhang, Y., Hou, Z., Wakatsuki, Y., Macromolecules, 32, 939 (1999). 10. Zambelli, A., Oliva, L., Pellecchia, C., Macromolecules, 22, 2129 (1989). 11. Grassi, A., Pellecchia, C., Longo, P., Zambelli, A., Gazz. Chim. Ital., 117, 249 (1987). 12. Zambelli, A., Pellecchia, C., Oliva, L., Shimin, H., Chin. J.. Polym. Sci., 6, 365 (1988). 13. Pellecchia, C., Longo, P., Grassi, A., Ammendola, P., Zambelli, A., Makromol. Chem., Rapid Commun., 8, 277 (1987). 14. Zambelli, A., Longo, P., Pellecchia, C., Grassi, A., Macromolecules, 20, 2035 (1987). 15. Miyashita, A., Mabika, M., Suzuki, T., presented at International Symposium on Synthetic, Structural and Industrial Aspects of Stereospecific Polymerization Proceeding, Milan, 1994. 16. Aoyama et al., in International Symposium on Future Technology, in press. 17. Ready, T. E., Day, R. O., Chien, J. C. W., Rausch, M. D., Macromolecules, 26, 5822 (1993). 18. Takeuchi, M., Shozaki, H., and Tomotsu, N., to Idemitsu Kosan Co. Ltd., EP0707013 (1993) 19. Campbell, R. E., Malanga, M. T., Metcon '93, (1993) 315. 20. Kaminsky, W., Metallocenes '96, 211 (1996). 21. Tomotsu, N., Polym. Prepr. Jpn., 42, 919 (1993). 22. Kakugo, M., Miyatake, T., Mizunuma, K., Stud. Surf. Sci. Catal., 56, 517 (1990). 23. Okuda, J., Masoud, E., Macromol. Chem. Phys., 199, 543 (1998). 24. Kucht, H., Kucht, A., Chien, J. C. W., Rausch, M. D., Appl. Organomet. Chem., 8, 393 (1994). 25. Soga, K., Monoi, T., Macromolecules, 23, 1558 (1990).

SYNTHESIS OF SYNDIOTACTIC POLYSTYRENE

387

26. Soga, K., Nakatani, H., Macromolecules, 23, 957 (1990). 27. Tomotsu, N., Ishihara, N., Newman, T. H., Malanga, M. T. J., Mol. Catal. A: Chem., US, 167(1998). 28. Monoi, T., Nakatani, H., Soga, K., Polym. Prepr. Jpn., 38, 1726 (1989). 29. Mani, R., Burns, C. M., Macromolecules, 24, 5476 (1991). 30. Tazaki, T., JP 2059871, 3087301. 31. Aaltonen, P., Seppala, J., Matilainen, L., Leskela, M., Macromolecules, 27, 3136 (1994). 32. Pellecchia, C., Pappalardo, D., D'Arco, M., Zambelli, A., Macromolecules, 29, 1158 (1996). 33. Inoue, N., Shiomura, T., Polym. Prepr. Jpn., 42, 2292 (1993). 34. Aral, T., Ohtsu, S., Suzuki, S., Macromol. Rapid Commun., 19, 327 (1998). 35. Pellecchia, C., Proto, A., Zambelli, A., Macromolecules, 25, 4450 (1992). 36. Zambelli, A., Giongo, M., Natta, G., Macromol. Chem., 112, 183 (1968). 37. Newman, T. H., Malanga, M. T., 4th SPSJ International Polymer Conf., (1993)27. 38. Newman, T. H., Malanga, M. T., J. Macromol. Sci. Pure Appl. Chem. A, 34, 1921 (1997). 39. Murata, M., et al., Macromol. Chem. Phys., 202, 1799 (2001).

This page intentionally left blank

18

Characterization, Properties and Applications of Syndiotactic Polystyrene KOMEI YAMASAKI Plastics Technical Center, Idemitsu Petrochemical Co., Ltd., Chiba, Japan

NORIO TOMOTSU Polymer Research Laboratory, Idemitsu Petrochemical Co., Ltd., Chiba, Japan

MICHAEL MALANGA R&D Engineering Plastics, The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Highly syndiotactic polystyrene (SPS) was synthesized using a homogeneous catalytic system using a titanium compound and methylaluminoxane or borate [1]. The detailed syndiospecific polymerization of styrene is described in the previous chapter. Dow Chemical and Idemitsu Petrochemical produce SPS, which is commercialized the trade names Questra® and XAREC®, respectively. SPS is very different from conventional amorphous polystyrene (atactic polystyrene, APS) in chemical and thermal properties. The high stereoregularity allows SPS to obtain a high level of crystallinity, resulting in a high melting temperature (270 °C), fast crystallization rate [2-5] and high solvent resistance. With these characteristics given by crystallization, SPS has a low specific gravity, excellent electrical properties, hydrolytic resistance and good moldability, which were inherited from APS (Figure 18.1). The excellent balances of mechanical, electrical, solvent resistance, heat resistance and dimensional stability properties combined with a relatively low price have led to the common use of this material in engineering plastics. Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy | 2003 John Wiley & Sons Ltd

390

K. YAMASAKI ETAL

Crystallization

New Engineering Plastics SPS Additional Characteristics

Commodity Plastics

• Heat Resistance • Chemical Resistance • Dimensional Stability

Conventional PS

Original Characteristics + Low Specific Gravity ^ Electrical Properties • Hydrolytic Stability • Moldability

Figure 18.1 Additional characteristics brought by crystallization of polystyrene

2 2.1

CHARACTERIZATION STRUCTURE

Syndiotactic polystyrene is produced by well stereo-controlled coordination polymerization by titanium compounds. The polymer has a high percentage of rrrr pentad structure. The 13C NMR chemical shift for the phenyl-1 carbon and backbone methylene carbon are approximately 145.3 and 44.9 ppm, respectively [1]. In general, these polymers are found to be >99% pure in syndiotactic structure as defined by NMR. The most stereoselective catalysts produce a >99.6% pure syndiotactic structure (Figure 18.2)

2.2

CRYSTAL FORM

SPS takes two different conformations in its crystal, TT and TTGG conformations, depending on the crystallization conditions, and exhibits a very complex polymorphic structure [6-14]. One is the TT conformation which appears when SPS crystallizes from the melt. Two different transcrystals are reported when SPS takes the TT conformation in the crystal, a-form [15-19] and ß-form. The identity period is 5.06 A (Figure 18.3) in both crystal forms. The ß-form is more thermally stable than the a-form. Figure 18.4 shows the detailed structure of the ß-form [6,10]. SPS takes the TTGG conformation when it crystallizes in the presence of organic solvents. SPS takes the TTGG conformation in -y-form and 8-form crystals. The 8-form is a complex of SPS and solvent [20–24]. Figure 18.5 shows the complex of SPS and toluene.

391

SYNDIOTACTIC POLYSTYRENE

[rrrrr] >99.6%

416

45

44

4':

Spinning side band PPM

46.0

45.5

45.0

44.5

44.0

43.5

43.0

Figure 18.2 13C NMR of syndiotactic polystyrene produced by highly stereocontrolled catalyst system. Apparatus: JEOL Lambda 500 (13C: 125.65MHz). Frequency: 25 000 Hz. pulse: Q.OJJLS (45° pulse). Repetition time: 4s. Scans: 10000

Figure 18.3 Conformation of a-form of SPS

392

K. YAMASAKI ETAL

a = 8.81 A

Figure 18.4

13/4

13/4

1/4

1/4

p-Form of SPS

On thermal treatment, the -form is transformed into the -/-form and then into the a-form. Tsutsui et al. examined in detail solvent removal from the 8-form and found that the solvent was removed without a change of the conformation of SPS [25].

3

PHYSICAL PROPERTIES

3.1

THERMAL PROPERTIES

By programmable differential scanning calorimetry (DSC), the glass transition temperature (7"g) of SPS is 100 °C. When amorphous SPS is exposed to heat, it crystallizes above the Tg and a crystallization exotherm maximum is observed at about 150°C, then a melting endotherm peak is observed at 270 °C. For a series of SPS samples with different degrees of crystallinity, the heat of fusion was determined as 53.2 J/g. Gvozdic and Meier tried stepwise annealing of SPS in a DSC instrument and found that the melting temperature increased from ca 270 to 287.3 °C [26]. By a modification of the annealing procedure, they attained a higher melting temperature of 291.5 CC [27].

393

SYNDIOTACTIC POLYSTYRENE b = 13.26 A

2b

a = 17.58 A

y= 121.2°

c = 7.71 A

Figure 18.5 8-Form of SPS: solvent-induced crystallized form with incorporation of toluene in the lattice

3.2

CRYSTALLIZATION BEHAVIOR OF SPS

The crystallization rate is dependent on the crystallization temperature, T, and molecular weight of the polymer. It is important to study the effect of T and molecular weight on the crystallization nature of SPS. Takebe et al. [28] studied the effect of temperature and molecular weight on the crystallization rate of SPS by DSC. When SPS was melted at 300 °C, then rapidly cooled to the crystallization temperature, T, the evolution of crystallization showed a sigmoidal curve with reference to crystallization time (Figure 18.6). The crystallization rate becomes slower as T approaches close to melting point. When the crystallization rate of SPS is analyzed based on Avrami theory, the Avrami index, n, is equal to 3, which suggests that crystallization of SPS proceeds via three-dimensional heterogeneous growth [28,29]. Figure 18.7 shows the time evolution of the crystallinity of SPS of various weight-average molecular weight, Mw. The crystallization rate becomes slower as Mw becomes larger at a fixed T.

394

K. YAMASAKI ETAL 100 80

; 60 ' 40 20

OIL

2

345678

0.1

2

345678

1

2

345

10

Time (tnin) Figure 18.6 Time evolution of degree of crystallinity of SPS isothermally crystallized at various temperatures. Mw = 400 000; Mw/Mn = 2.4; Xct = degree of crystallinity at time t; X£° = degree of crystallinity at infinite time

1.0 0.8

0.4

0.2

0.0

56789

2 1

3 4 5 6 7 8 9 10

2

3

4 5 6 7 89 100

Time (min) Figure 18.7 Time evolution of degree of crystallinity of SPS of various Mw. Isothermal crystallization temperature T = 245 °C. Numbers on curves represent Mw of SPS

From these results, the effect of T and Mw on the crystallization rate constant, k:1/3, can be analyzed according to the Lauritzen-Hoffman nucleation theory [30,31], and the results are shown in Figure 18.8.

SYNDIOTACTIC POLYSTYRENE

395

5 8

4

1.4

(a)

1.6

1.8

T
2.0 2

2.2

1.6

(b)

1.8

T o/T AT X10

2.0

2.2

2

Figure 18.8 Analysis of crystallization rate of SPS according to Lauritzen-Hoffman nucleation theory [Equation (1)] (Af = T0* - 7). (a) Data plotted for various Mw: o4.70x 104; 1.26x 105;x, 3.78 x 105;A, 6.91 x 1050;o, 9.23 x 105;, 2.00 x106. (b) Superposed line by horizontal shift with reference toT m~°/TAT where /c'/« — crystallization rate constant, n = Avrami index, and n = 3 (three-dimensional heterogeneous growth) U = activation energy for diffusion T^c = temperature where flow stops Tm ° = equilibrium melting temperature KB — nucleation constant. When the crystallization rate constant k1/3 is plotted against TmQ/(T&T), where AT = Fm0 — T, and Tm0 is the equilibrium melting temperature, 310°C [32], k1^ of SPS of various Mw have the same slope with reference to r m °/(rAr) (Figure 18.8a). This result suggests that the nucleation constant, Kg, does not depend on Mw of SPS. Then all lines can be superposed on to a single line by a horizontal shift with reference to Tm°/(T&T) (Figure 18.8b). Plotting the shift factors against Mw, one can obtain a single line with slope of -0.8 (Figure 18.9). Thus the crystallization rate of SPS follows the dependence M-0.8. This result suggests that the crystallization rate of SPS obeys the diffusion of molecular chain by reptation, as was suggested by Hoffman.

3.3

COMPARISON OF CRYSTALLIZATION PROPERTIES OF SPS WITH IPS

When SPS is compared with isotactic polystyrene (IPS), a substantial difference is seen in the crystallization. IPS is also crystalline PS, but it has not been commercialized because of its very slow crystallization rate. In contrast, SPS

396

K. YAMASAKI ETAL -j^..T^:.j7rj.7ir.T.TT

10 7 6 5 4

2 -I

>—- j-

3 4 5 6 789 105

•i

i

2

3 4 5 6 789 Mu

j....i...j..4.4.}-4..

106

Figure 18.9 Molecular weight dependence of crystallization rate of SPS: shift factor aM plotted against Mw. The crystallization rate of SPS has an M-0.8 dependence

crystallizes fast enough in the conventional fabrication processes, such as injection molding, that it opens up the opportunity for the high-temperature use of polystyrene. The crystallization rate of SPS is roughly 100 times faster than that of IPS (Figure 18.10) [32]. The C—C backbone of SPS takes a planar zig-zag conformation in the crystal whereas IPS takes a 3/1 helical conformation in crystal. This difference in chain conformation in the crystal affects the crystallization rate of these polystyrenes. Table 18.1 [32] compares the crystallization parameters between SPS and IPS. The work of chain folding, Q, of SPS is smaller than that of IPS, which suggests that the chain folding of SPS can take place more easily than that of IPS. Also, SPS has a smaller surface free energy. These differences are related to the crystallization rate difference between SPS and IPS, which have different stereoregularities. 3.4

SOLVENT RESISTANCE

At room temperature there are no known solvents for SPS. Even though the good solvents for atactic polystyrene such as toluene, benzene, tetrahydrofuran and dichloroethane can swell SPS at room temperature. Some of these solvents dissolve SPS at elevated temperatures, close to their boiling point. For example, a 5-10% by weight solution can be prepared by heating the polymer with trichlorobenzene at 160°C with agitation.

397

SYNDIOTACTIC POLYSTYRENE

1000

0.01

Figure 18.10 Comparison of isothermal crystallization rates of SPS and IPS

Table 18.1 Comparison of parameters for crystallization between SPS and IPS Parameter" m

(K)

A//f (J/m 3 ) a(m) h(m) ab (J/m2) cre (J/m 2 ) £> (kJ/mol)

IPS

SPS 583 5.1 8.8 7.2 4.1 27.0 2.0

x x x x x x

107 10 -10 10 -!0 10 -3 10 -3 104

515.2 9.1 12. 8 5.5 7.1 34.8 3.0

x x x x x x

107 10-10 10-'° 10 - 3 10 -3 104

" T0m = equilibrium melting temperature; AHf = heat of fusion; a, b = unit cell of crystal; a = surface free energy of crystal; ae = surface free energy of crystal; Q = energy to form chain folding in crystal. b from Thomas-Stavely relationship: a = 0.1A//f(a6) 172 .

3.5

RHEOLOGICAL PROPERTIES

The melt viscosity of SPS can be superposed on to a single curve according to the time-temperature law (Figure 18.11). The viscosity of SPS has a 3.4 power dependence on Mw (Figure 18.12). When the shift factor, GT, is plotted against

398

K. YAMASAKI ETAL i

t

<— «•• i-i-M-rtv-

l::::::S:::!::!:ti:!31t:::::J::±:±:l5±55l::::::t::fc:£±±S±tt::::::5:::5::S:5:!5fl

+ .-<-. ^4.+.j.j.

1— '.-••.-',• 1-t-M-f

-t— •f-t-t-t'.-t

referer iceTennp.=290°( referer ice V =2.89X1(15

•t-'.-'.-H

103

.....

»<%«b*,"i'l'i'M

*#&

-j-n-m * --'.-'.-'.•i* t A-

ir—(

J

,

•-;-•!-<-;-!,

102

101

_.. Mw=2.89 x 105 Mw/Mn=2.59 :::: O T-290°C ;;;; n 280°C -- A 300°C -- v 310°C i. 10-2

i i i i i iiii

•!-<-<-H .?»< •t-1-t-H -)•;-;-!•! •;.;-M;

^

... ••4-*-H-»»'... —!"^•^^•^^^••

^iiw ! , ..i.;.;^.a|J»»...,

" " "i"' '.',

- •!•;•>; 4 •• -;.;->; 4

k•

*...

r= 290°C

4W-: J.42 x 10 5 M V

k1n=2.47

/^

6.C13 x K)5 7.1 7 x K)5

O

L ;i

..;.;.;.;; 4

2.53 2.39

i i iiiiiii i i i i i iiil -i i i i-iiili i i i iiiiu 1 100 102 10

i i iiiJ 104

waTaM ( Figure 18.11 Superposed viscosity curve of SPS. Shear viscosities measured for several Mw and temperatures were superposed with reference to temperature (reference temperature 290 °C) and Mw (reference Mw 289000). aT = shift factor for temperature; aM — shift factor for Mw 105

104

102

10' 105

•I---1---VT-

4

5 6 7 89

106

Figure 18.12 Relationship between zero-shear viscosity n0 and Mw. »;0 is proportional to

399

SYNDIOTACTIC POLYSTYRENE

temperature (Figure 18.13), the activation energy to flow was calculated to be 21 kcal/mol for SPS. This value is almost equivalent to that of APS (25 kcal/ mol) [33] and IPS (20 kcal/mol) [34].

3.6

MECHANICAL PROPERTIES OF NEAT SPS

SPS has high a melting point almost comparable to those of other engineering plastics (Table 18.2). In comparison with other crystalline engineering thermoplastics, of particular interest is that SPS has the lowest specific gravity and dielectric constant. In a practical fabrication method, such as injection molding, the mold temperature affects the degree of crystallinity, Xc, of SPS and the morphology [35]. When neat SPS is molded at high mold temperatures, a well developed lamellar structure is observed, and macroscopically spherulites are observed (Figure 18.14), and Xc stays constant across the whole cross-section of the molded part (Figure 18.15). In contrast, when SPS is molded at low mold temperatures far below the Tg, crystallization of SPS is restricted at the surface owing to fast cooling, resulting in a lower Xc and a poorly developed crystalline lamellar structure at the surface. However, in the center of the molded parts, spherulites are formed because the slower cooling allows the crystallization of SPS. The Xc profile across the cross-section is very different from what was molded at high mold 0.30 0.20 0.10 0.00

-0.10 -0.20

-0.30 -0.40 -0.50 -0.60 .700

1.750

1.800

1/TxlO3 (K- 1 )

Figure 18.13 Shift factor, aj, plotted against inverse of temperature. The activation energy to flow was calculated to be 21 kcal/mol for SPS

K. YAMASAKI ETAL

400

Table 18.2

Comparison of mechanical properties of neat polymers Units SPS

Property Specific gravity Melting point Glass transition temperature Flexural strength Flexural modulus Notched Izod impact DTUL(1.82MPa) Vicat softening temperature Dielectric constant (23 °C, 1 MHz)

Figure 18.14 (b)50°C

kg/m

3

°C

°c MPa MPa

kJ/m2

°C

°c

GPPS PBT

PA6

PA66 PPS

1040

1040

1310

1140

1140

1340

270 100 75

— 100 65

224 30 80

224 45 100

260 70 110

285 92 95

3000

2900

2400

2600

2800

3800

2..5 5.4 4.4 2.2 4.4 2.0 138 80 64 89 60 96 104 254 270 250 215 215 3.4 3.5 3.2 3. 1 2.6 2.6

Morphology of injection molded SPS. Mold temperature: (a) 160 and

1UU.U

90.0

_

90.0 80.0 -

80.0 70.0 60.0

>• 70.0

3

| 60.0

O

50.0

X

40.0

U

40.0

30.0 -

^

30.0 20.0

20.0 10.0 n ft

1

500

(a)

1

1

1000 1500 Distance (|im)

1

1

O

50.0

10.0 An

\

r

° i

i

i

i

i 3000

(b)

Distance (um)

Figure 18.15 Profile of degree of crystallinity of injection molded SPS in the crosssection. Mold temperature: (a) 160 and (b) 50 0C

401

SYNDIOTACTIC POLYSTYRENE

temperatures; at the surface of the molded part, Xc is low and it gradually increases towards the center of the part. A sufficiently high mold temperature is, therefore required in order to attain a high degree of crystallinity of SPS. The degree of crystallinity is important as long as the heat resistance of crystalline polymer is concerned. Figure 18.16 shows the dependence of the dynamic elastic moduli (E' and E") on temperature. The dynamic storage modulus, E', exhibits a strong dependence on Xc; for high Xc, E' drops at the TB, then it shows a gradual decline near the melting point, whereas for low Xc, E' shows a steep drop above Tg. Thus the degree of crystallinity is very critical to the heat resistance of SPS.

SPS has heat resistance and chemical resistance in addition to the inherent characteristics of conventional polystyrene. Of significant interest, moreover, is that SPS is cost-competitive because it is synthesized from styrene monomer, a well established and widely available raw material. Although SPS has a brittle nature like APS and is not suitable for use alone for structural material, reinforcement with glass fiber or impact modification by elastomers improves the mechanical properties of SPS. 104

101, 1000-

50

100

150

200

250

300

Temperature (°C) Figure 18.16 Temperature dependence of elastic modulus of SPS with various degrees of crystallinity, Xc: +, 58%; O, 46%; •, 10%

402

4.1

K. YAMASAKI ETAL

MECHANICAL AND FLOW PROPERTIES

The mechanical properties of glass fiber-reinforced materials are compared in Table 18.3. Glass fiber-reinforced SPS (GFSPS) has mechanical properties competitive with those of other engineering thermoplastics. Highlights of GFSPS are low specific gravity and a high heat distortion temperature. Moreover, GFSPS has good flowability; almost comparable spiral flow length with liquid crystal polymer (LCP) which is a representative material in connector applications where thin wall flowability is required (Figure 18.17).

4.2

ELECTRICAL PROPERTIES

GFSPS offers a low dielectric constant and low dissipation factor compared with other thermoplastic engineering plastics (Figure 18.18). Also, the low dissipation factor of GFSPS remains almost constant over a wide frequency range up to 10 GHz (Figure 18.19). GFSPS possesses a higher tracking resistance, CTI Class 2 to Class 0 (Figure 18.20), which suggests further reliable electrical properties of GFSPS. These electrical properties of GFSPS, along with heat resistance and flowability, allow GFSPS to be competitive with existing plastics in electrical applications such as various connectors, antennae, and other electrical devices. Table 18.3 polymers

Comparison of mechanical properties of 30% glass fiber-reinforced

Property Specific gravity Water absorption Mold shrinkage (MD) Tensile strength Tensile elongation Flexural strength Flexural modulus Notched Izod impact DTUL(1.82MPa) DTUL (0.45 MPa) CLTE (MD) Dielectric constant (1 MHz) Dissipation factor (1 MHz) Breakdown voltage 40%GF.

Units

GF-SPS

3

kg/m 1250 0.05 % 0.35 % MPa 118 2.5 % MPa 185 MPa 9000 11 kJ/m2 C 251 C 269 °C 2.5 x!05/°C 2.9 <0.001 kV/mm

48

GF-PBT GF-PET GF-PA66 GF-PPS" 1530 0.06 0.35 138 3.1 215 9500 9 210 225 4.5 3.6

0.003 21

1550 0.10 0.30 152 2.5 196 9800 8 245 250 3.0 3.5

0.007 26

1370 0.60 0.35 177 3.5 255 8300 10 250 262 3.5 3.3

0.009 20

1670 0.02 0.25 147 1.5 206 13700 9 >260 >260 2.2 3.9 0.001 16

403

SYNDIOTACTIC POLYSTYRENE

IRGFSPS (GF30%) LCP G30 (Type II)

60

80

100 120 Injection Pressure (kgf/cm2 gauge)

Figure 18.17 Comparison of spiral flow lengths of SPS, LCP and PPS. SPS has high flowability close to LCP (type II)

(1MHz 25°C)

JPAr -PC(GF30%)-

0.0100

}PSF

0.0010

0.0001

Resin

OPBT (GF30%)

~ O PBT (GF30%) O (PPS (GF40%) IRGFSPS(GF30%) PPEGFSPS(GF30%) Impact modified SPS

O pTFE

3

4

Dielectric constant e' Figure 18.18 Map of dielectric properties of engineering plastics. Among engineering plastics, SPS (impact modified and GF-reinforced HB and IR grades) has very low dielectric dissipation factor and dielectric constant following those of fluorocarbon polymers. PSF, polysulfone; PPE, poly(phenylene ether); PES, poly(ether sulfone); PAr, polyarylate

K. YAMASAKI ETAL

404

0.0001

107

108

109

1010

Frequency (Hz) Figure 18.19 Dependence of dissipation factor on frequency. The low dissipation factor of SPS stays almost constant over a wide frequency range

(GF reinforced) 40

35 30

PET|PA46

o 25

LCP 20

15

Tracking resistance (UL Class) Figure 18.20 Tracking resistance (UL Class) and breakdown voltage

4.3

CHEMICAL RESISTANCE

Chemical resistance is also one of the unique features of SPS. APS originally had good resistance to acids, alkalis and hydrolytic environments below Tg. SPS took over these chemical resistances from APS and acquired resistance to higher temperatures above Tg and, furthermore, resistance to organic solvents

SYNDIOTACTIC POLYSTYRENE

405

(Table 18.4) due to crystallization. GFSPS exhibits good steam resistance comparable to PPS, far better than PBT and PA (Figure 18.21). This chemical resistance along with the excellent electrical properties open up opportunities for GFSPS in electrical parts in automotive applications, such as connectors, electrical control unit (ECU), solenoids, and so on.

4.4

IMPROVEMENT OF POLYSTYRENE BY BLENDING SPS

Current polystyrenes such as GPPS and HIPS have good properties and have well established applications in the market; however, they are poor in chemical resistance because they are amorphous polymers. If their chemical resistance were improved, their applications would be widely extended and could replace ABS or SAN, which have better chemical resistance than GPPS and HIPS. Blending a small amount of SPS improves the chemical resistance of GPPS and HIPS. HIPS/SPS blend and GPPS/SPS blend have chemical resistances Table 18.4 Chemical resistance of 30 % glass fiber-reinforced SPS, PA66 and PBT after immersion for 30 days Temperature (°C) GF SPSa GF PA66a GF PBT'

Type

Chemical

Acid Alkali Salt Alcohol Organic solvents

Blow-bye water'' NaOH aq. (10%) CaCl2 aq. (10%) Methanol Ethyl acetate Acetone Methyl ethyl ketone Toluene Gasoline'

100 80 80 60 70 50 70 80 80

E G E E S S S S S

E/D S E/D S E S E E E

E/D E/D E G G G G G E

Gas oil Engine oil Diesel oil Gear oil Brake oil Silicone grease Antifreeze Window washing

80 150 150 150 80 150 120 80

E E E E E E G E

E E E E E E E/D S

E G G S E E E/D G

Automotive chemicals

a h c

E, excellent; G, good; S, swell; E/D, erosion, dissolution. HC1-H2SO4—HNO3 mixed acid (pH = 3). Premium type.

406

K. YAMASAKI ETAL.

100200

1000

2000

Time (h)

Figure 18.21 Comparison of hydrolytic stability. Pressure cooker test under 120 C steam atmosphere. SPS (30% GF), PBT (30% GF), PA66 (30% GF), PPS (40% GF)

comparable to or better than those of ABS and SAN. Figure 18.22 compares the chemical resistances of HIPS/SPS (20% SPS) blend, HIPS alone and ABS when they were immersed in methyl ethyl ketone (MEK) at room temperature. Only 1 h of immersion makes a difference; HIPS alone and ABS start to dissolve when they are immersed in MEK, and are entirely dissolved within 5 h. However, the HIPS/SPS blend keeps its shape even after 5 h of immersion, although swelling is observed.

Figure 18.22 Comparison of chemical resistance of HIPS/SPS (80:20) blend with those of HIPS and ABS, immersed in methyl ethyl ketone at room temperature

SYNDIOTACTIC POLYSTYRENE

407

When chemicals are weak solvents for polystyrene, stress cracks are formed in the products. Chemical resistances to practical chemicals are compared in Table 18.5. HIPS—SPS blend exhibits better resistance to chemicals which are used in the kitchen and bathroom. In HIPS/SPS blend, SPS may work to prevent the formation of crazes and their propagation to cracks initiated from contact with chemicals. The mechanical properties of HIPS/SPS blend surpass those of HIPS (Figure 18.23). Thus HIPS/SPS blend shows improved chemical resistance with better mechanical properties. Recycling is another important aspect; recycling of ABS and SAN is generally difficult because they have acrylonitrile in their molecule, whereas HIPS— SPS blend is recyclable because its chemical component is styrene. With all these characteristics, HIPS/SPS blend will be very competitive in broad applications where HIPS, ABS and SAN have been used but increased chemical resistance is required.

5

SUMMARY

SPS is entirely different from conventional amorphous styrenics in chemical and physical properties. In addition to characteristics such as low specific gravity, excellent electrical properties, hydrolytic resistance and good moldability similar to those of existing styrenics, SPS has heat resistance, chemical resistance and the characteristics inherent to crystalline polymers that make it a new engineering plastic. SPS is finding new applications and is expected to bring out new horizons in the application areas of engineering thermoplastics. Table 18.5 Chemical resistance of HIPS—SPS (80:20) blend to practical chemicals observation of cracks Type

Chemical

HIPSa

HIPS-SPS blenda

Cooking oil

Palm oil Sesame oil Cologne Mildew remover Kitchen cleaner Bath cleaner ZEPRO SJ

M M M M F F F

F F F S N N S

Toiletry Surfactant Engine oil

M, many cracks; F, fair amount; S, small amount; N, no cracks.

K. YAMASAKI ETAL

408

60

HIPS—SPS blend HIPS 40

20

Izod (kJ/m2) Figure 18.23 HIPS

Tensile strength (MPa)

Elongation Flexural strength (%) (MPa)

Comparison of mechanical properties of HIPS/SPS (80:20) blend and

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

Ishihara, N., Seimiya, T., Kuramoto, M., Uoi, M., Polym. Prepr. Jpn. 35 (1986) 240. Govozdic, N. V. Meier, D. J., Polym. Commun. 32 (1991) 183. Govozdic, N. V., Meier, D. J., Polym. Commun. 32 (1991) 493. Arnauts, J., Berghmans, H., Polym. Commun. 32 (1991) 343. Takebe, T., Funaki, K. and Yamasaki, K., Polym. Prepr. Jpn. 42 (1993) 4306. Kobayashi, M., Nakaoki, T., Ishihara, N., Macromolecules 22 (1989) 4377. Doherty, D. C, Hopanger, A. J., Macromolecules 22 (1989) 2472. Greis, O., Xu, Y., Asano, T., Petermann, J., Polymer 30 (1989) 590. Zimba, C. G., Rabolt, J. F., English, A. D., Macromolecules 22 (1989) 2867. Gomez, M. A., Tonelli, A. E., Macromolecules 24 (1991) 3533. Grassi, A., Longo, P., Guerra, G., Makromol. Chem., Rapid Commun. 10 (1989) 687. Gomez, M. A., Tonelli, A. E., Macromolecules 23 (1990) 3385. Capitani, D., De Rosa, C., Ferrando, A., Grassi, A., Segre, A. L., Macromolecules 25 (1992) 3874. Capitani, D., Segre, A. L., Grassi, A., Ferratndo A., Macromolecules 24 (1991) 623. Guerra, G., Vitagliano, V. M., De Rosa C., Petraccone, V., Corradini, P., Macromolecules 23 (1990) 1539. De Rosa, C., Guerra, G., Petraccone, V., Corradini, P., Polym. J. 23 (1991) 1435. Sun, Z., Miller, R. L., Polymer 34 (1993) 1963. Napolitano, R., Pirozzi, B., Macromolecules 26 (1993) 7225. Auriemma, F., Petraccone, V., Poggetto, F. D., De Rosa, C., Guerra, G., Manfredi, C., Corradini, P., Macromolecules 26 (1993) 3772. Chatani, Y., Shimane, T., Inagaki, T., Ijitsu, T., Yukinari, T., Shikuma, H., Polymer 34(1993) 1620.

SYNDIOTACTIC POLYSTYRENE

409

21. Deberdt, F., Berghmans, H., Polymer 34 (1993) 2192. 22. Deberdt, F., Berghmans, H., Polymer 35 (1993) 1694. 23. Kobayashi, M., Yoshioka, Tashiro, K., Suzuki, J., Funahashi, S., Izumi, Y., Macromolecules 27 (1994) 1349. 24. Kobayashi, M., Yoshioka, T., Imai, M., Itho, Y., Macromolecules 28 (1995) 7376. 25. Tsutsui, K., Tsujita, Y., Yoshimizu, H., Kinoshita, T., Polymer 398 (1998) 5177. 26. Gvozdic, N. V., Meier, D. J., Polvm. Commun. 32 (1991) 183. 27. Gvozdic, N. V., Meier, D. J., Polym. Commun. 32 (1991) 493. 28. Takebe, T., Funaki, K., Yamasaki, K., Polym. Prepr. Jpn. 42 (1993) 4306. 29. Tanaka, K., Haneda, T., Soeno, T., Yonetake, K., Koyama, K., Seikei-Kakou 2 (1990) 166. 30. Hoffman, J. D., Miller, R. L., Polymer 38 (1997) 3151. 31. Hoffman, J. D., Miller, R. L., Macromolecules 21 (1988) 3038. 32. Takebe, T., Funaki, K., Yamasaki, K., in 4th SPSJ International Polymer Conference, Yokohama, 1992, p. 175. 33. Nielsen, L. E., Polymer Rheology, Mercel Dekker, New York, 1977. 34. Suzuki, T. and Kovacs, A. J., Polym. J., 1 (1970) 82. 35. Lopez, L. C, Cieslinski, R. C, Putzig, C. L., Wesson, R. D., Polymer 36 (1995) 2331.

This page intentionally left blank

19

Rubber Modification of Syndiotaetic Polystyrene G. E. McKEE, F. RAMSTEINER AND W. HECKMANN BASF AG, Ludwigshafen, Germany

1

INTRODUCTION

Amorphous polystyrene (aPS) is a widely used technical polymer. Its application at high temperatures is limited, however, by softening at the glass transition temperature near 100°C. With the preparation of syndiotactic polystyrene by Ishihara et al., [1] a new semi-crystalline polystyrene (PS) was born, which had a melting point of the crystalline phase near 270 °C. The exact temperature depends on the actual structure of the crystalline phase [2,3]. This second phase opened the way for the use of polystyrene at considerably higher temperatures than before. The crystalline phase acts as physical crosslinks up to the melting point of the polymer and stops material flow above its glass transition temperature. In Figure 19.1 the shear moduli of a rubber-modified syndiotactic polysytrene (sPS) and aPS determined in the nonresonant torsion vibration test are plotted as a function of temperature. The glass transition temperature is roughly the same in aPS and sPS. Above the glass transition temperature, the observed increase in modulus is caused by further crystallization, if the material was cooled from the melt too quickly. Polystyrene (PS) in its atactic and syndiotactic forms is a brittle thermoplastic, even in an orientated state [4]. To improve the toughness of aPS, impact modification has been practised for a long time, either by polymerizing the styrene in the presence of a polybutadiene rubber leading to high-impact polystyrene, commonly called HIPS, or by blending the polystyrene with multi-block copolymers, mainly of the styrene—butadiene—styrene (S—B—S) type. Modern Styrenic Polymers: Polystyrene and Styrenic Copolvmers. Edited by J. Scheirs and D. B. Priddy © 2003 John Wiley & Sons Ltd

G. E. McKEE ETAL

412

G'(Mpa)

109,

108

107

aPS

106

105. -200

-150

-100

-50

50

100

150

200

250

Temperature (°C)

Figure 19.1 Dynamic shear modulus (1 cycles/s) of aPS and sPS as a function of temperature

In the case of sPS, the problem of its brittleness can be even more acute since it has to compete with engineering plastics which possess an inherent toughness superior to that of sPS. For this reason, a good impact modification of this product is of paramount importance and may even be essential for its survival as a commercial thermoplastic. For this reason a chapter of this book has been dedicated to the impact modification of sPS using elastomers. Since rubber modification plays such an important role for styrene polymers, whether atactic or sydiotactic, we will first look at the methods of energy dissipation in these homopolymers on impact. The use of glass fibres to enhance energy dissipation, which only functions if the fibres are oriented perpendicular to the crack propagation, is important for applications of sPS. However, since this chapter is concerned primarily with the rubber modification of sPS, it will not be considered in detail here.

2

ENERGY DISSIPATION IN POLYSTYRENE POLYMERS

PS, in either its atactic or syndiotactic form, is a polymer which shows no segmental mobility of chain segments below its glass transition temperature. Secondary relaxation processes which can be attributed to mobility in the main chain are missing. Therefore, these materials do not exhibit long-range energy

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

413

dissipation deformation processes on impact below the glass transition temperature and thus the main method of energy dissipation is via craze formation. This has a consequence that PS polymers are very susceptible to brittle failure at these temperatures. Above the glass transition temperature, the whole polymer chain is mobile in the amorphous region and as a result stable energy-dissipating shear processes prevail. Figure 19.2 shows as an example the pattern of the sPS matrix after deformation in tension at 110°C [5]. The herring-bone structure of the crystalline lamellae (white), originally running perpendicular to the tensile direction, is observed in a transmission electron microscope after staining with RuO4. This process can be explained by assuming shearing along the molecules within the crystalline lamella, as demonstrated for semi-crystalline polyethylene [6] by Kanig and more generally outlined by Lin and Argon [7], who included screw dislocation within the lamellae to illustrate shearing along the chains within a crystal. The prerequisite for this kind of deformation is high mobility of the molecules above their glass transition temperature in the amorphous region between the crystalline lamellae. Crazes are not observed. More important is the use of sPS at lower temperatures. Below the glass transition temperature the dilatational strain within the bulk material in tension is released by the well known crazes. Crazes are crack-like structures bridged by highly oriented fibrils. These fibrils, after formation during tensile testing, are stress bearing and can therefore contribute to energy dissipation. A prerequisite for stress bearing is that the molecular weight of the sPS is high enough to form sufficient numbers of tie molecules between the crystalline phases and sufficiently strong entanglements between the individual polymer

Figure 19.2

TEM image of sheared lamellae in sPS after deformation at 110°C

414

G. E. McKEE ETAL.

chains. Figure 19.4 shows the bending strength of sPS as a function of its molecular weight. Obviously a molecular weight of l00 000kg/mol fulfills this prerequisite. For aPS, a molecular weight of 70 000 kg/mol is regarded as a critical value [8]. Above these critical molecular weights the tensile strength only slightly increases further since the molecules are sufficiently entangled to break and do not allow disentanglement at high deformation rates. However, at higher temperatures or slow crack propagation as in creep and fatigue, where the molecules have time to disentangle, the degree of entanglement at even higher molecular weights becomes important. This can be seen in Figure 19.3 [5] where for sPS the crack propagation in fatigue at l0cycles/s is shown as a function of the stress intensity factor for sPS polymers with different molecular weights, using the bending test as in Figure 19.4. The molecular weight dependence is much more pronounced in fatigue than in fast tension. In semicrystalline materials the resistance to crack formation of the tie molecules between the crystals must also be considered [9]. In contrast to aPS, the molecules in sPS are partly embedded in the crystals and are consequently more resistant to being pulled out of the matrix by the propagating crack. Short molecules, however, may be completely incorporated in the crystals and do not form stabilizing tie molecules. Additionally, if the crazes are not stabilized by rubber particles, they are commonly regarded to be precursors of premature failure since the cavities between the fibrils are weak points in the polymer matrix.

1,00E-02 - Mw = 66.000 • Mw = 93.000

1,00E-03

o Mw = 137.000 « Mw = 194.000 a Mw = 792.000

I T3

1,00E-07

10 A# (Mpa m0.5)

Figure 19.3 Crack speed in fatigue of sPS as a function of the amplitude of the stress intensity factor for different molecular weights

415

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE 30-i _ 25h. H 20I 15" oo .S 10 -I

0

10

Figure 19.4

3

100 molecular weight (kg/mol)

1000

Bending strength of sPS as a function of the molecular weight

IMPACT BEHAVIOUR OF RUBBER-MODIFIED sPS

For many applications, the toughness of sPS is insufficient, which has thus led to many attempts in the past to increase its toughness significantly compared with HIPS by blending with rubbers. In the stress field of softer or harder particles than the sPS matrix, typical deformation processes inherent to the matrix are initiated. For rubber modification it is important that the application or test temperature is above the glass transition temperature of the rubber, otherwise the stiffnesses of the two components hardly differ from each other and local stress fields around the rubber particles are not formed. The formation of numerous deformation zones round the rubber particles is generally the basis of impact modification [10]. In polystyrene, crazes are the dominant deformation process and are formed within the stress field of the rubber particles. In Figure 19.5, a transmission electron micrograph (TEM) of rubber-modified deformed sPS is shown. Clearly the rubber particles are partly cavitated and interact with the numerous crazes. It is still open for discussion whether voids are first formed in the rubber particles, as elaborated by Lazzeri and Bucknall [11], and then these voids initiate craze formation in the adjacent region, or vice versa, namely that from local shearing and micro voiding in the matrix near the rubber particles crazes are formed [12] and then voiding occurs in the rubber particles or whether both occur simultaneously. Today it is generally accepted that voiding in the rubber particles is an important factor in reducing the dilatational strain field without forming additionally harmful stress intensities in the matrix. It is also believed that voiding in

416

G. E. McKEE ETAL

Figure 19.5 Crazes and voided rubber particles in rubber-modified sPS after deformation

rubber particles is beneficial for the shearing process. Additionally in crazing materials as in styrene polymers the local volume mismatch between the rubber particles and the craze opening is compensated by voiding in the rubber phase, avoiding the energetically more difficult necking of the grafted rubber particles. Owing to the interaction between the matrix and the rubber particles the latter should have an optimal size of about 1 u,m for PS. If the particles are too small, e.g. below 0.1 jxm, the extension of the stress fields around the particles is too small to nucleate and open crazes and/or voids cannot be formed in the small rubber particles leading in both models to the failure of the impact modification. Agglomeration of small particles to effectively large ones can help to overcome this size limit. If, on the other hand, the size of the particles is too large, the distance between the particles increases for a given rubber concentration and the rubber particles cannot stabilize craze propagation at the long inter-particle distances present in the product. Additionally, the large particles themselves become weak points owing to their large internal cavities. The rubber particles must be grafted for compatibility with the matrix and also for their dispersion in it. This helps to avoid detrimental delamination and void formation at the interface of the matrix and the rubber particle during the deformation and increases the stress-bearing capacity of the rubber particles within the propagating crazes owing to their deformation in the stress direction. This elongation is not possible with inorganic fillers, e.g. glass beads, hence these particles can generate crazes but cannot stabilize their growth. Fillers are therefore far less effective than rubber particles as impact modifiers.

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

417

sPS is compatible with aPS in the melt [13]. This compatibility was checked in our experiments by studying interdiffusion at their interface by a welding experiment [5]. Thereby a block of sPS was brought in contact with a block of aPS, the interface was marked with colloidal gold particles and then annealed for 3 min at 300 °C (Figure 19.6). It could be seen from the crystalline lamellae of sPS which formed during cooling to room temperature that the sPS had diffused into the amorphous aPS block, thus confirming that aPS and sPS are compatible in the melt. Since in blends of aPS and sPS crystallization does not occur in isolated regions of separated homopolymers but is spread homogeneously over the whole material (Figure 19.7), this could also indicate that the amorphous phase of sPS and aPS may be compatible at room temperature, indicating that aPS can be used as the compatibilizer between the rubber and sPS phases. This is usually achieved in the form of a graft shell in core-shell rubbers or as a block, e.g. in styrene—hydrogenated butadiene—styrene block copolymers. Since it is difficult to graft syndiotactic polystyrene on to rubbers, polystyrene still remains the compatibilizing phase of choice for the rubber phase. The different methods employed for rubber modification are reviewed below.

4

RUBBER MODIFICATION

An overview of sPS as a matrix is given in a book edited by Gausepohl [14]. One of the points reviewed is its impact modification.

Figure 19.6 TEM images of the interdiffusion of sPS and aPS at the interface after annealing for 3 min at 300 °C. The original interface was marked with colloidal gold particles (arrows)

418

G. E. McKEEETAL

Figure 19.7

4.1

TEM image of crystalline lamellae of sPS in a mixture of 10 % sPS in aPS

STYRENE BLOCK COPOLYMERS AS IMPACT MODIFIERS

The most commonly used impact modifiers are block copolymers, especially those described by Idemitsu [15]. Several types are described but commercially styrene—rubber—styrene block copolymers are preferred. The outer styrene blocks serve to anchor the middle rubber block. Theoretically a butadiene rubber block is possible but owing to the high processing temperatures of practically 300 °C for sPS, degradation and crosslinking of the polybutadiene phase are to be expected. To prevent this, the butadiene block is hydrogenated to give an SEBS polymer. In the literature it is reported [15] that the notched Izod impact strength of an sPS with a molecular weight of 800000 and a degree of syndiotacticity of 96% could be increased 600% by using 30wt% of a hydrogenated S—B—S block copolymer (G 1652®, Shell Chemical). On using an unhydrogenated S—B—Sblock copolymer (TR 1102®, Shell Chemical) or a pure polybutadiene (NF35AS®, Asahi Kasei), the impact modification was not as efficient as with the hydrogenated products. In both cases the reason could be crosslinking of the polybutadiene phase and in the latter case insufficient compatibility between the polybutadiene and the sPS is presumably an important factor. With increasing modifier content the modulus of elasticity is reduced (Table 19.1). Unpublished work at BASF has shown that using these S—EB—S block copolymers the highest notched impact strength is obtained with a high molecular weight product. Thus for the S—EB—S block copolymers of the Kraton*

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE Table 19.1

419

Impact modification of sPS as described in EP 318 793 [15]

Volume average Izod notched impact Modulus in SPS (wt%) Impact modifier particle size (u,m) strength (kg cm/cm) tension (kg/cm2) 90

100

SEBS(10%) SEBS(20%) SEBS (30%) MAS (20%)" SBS(20%) Polybutadiene (20%) —

2.0 2.5 3.0 0.3 2.5 3.5

4.2 7.3 12.1 8.0 5.4 3.8

38000 34000 30000 35000 32000 30000

—

2.2

40000

a

Core—shell impact modifier: methyl methacrylate-butyl acrylate—styrene copolymer (KM330 R, Rohm & Haas).

type from Shell Chemical on increasing the molecular weight of the polymer from 55000g/mol (Kraton G 1652®) to 170000g/mol (Kraton G 1651®) the Charpy notched impact strength could be doubled (20% block copolymer). The glass transition temperature of the rubber phase is ca —60°C (Figure 19.1); the morphology of sPS block copolymer blend is shown in Figure 19.8.

Figure 19.8 TEM image of sPS modified with S—EB—S block copolymer

420

G. E. McKEE ETAL

The reason for the better impact modification with increasing molecular weight of the block copolymer was not investigated. It could depend on the molecular weight of the rubber block (the higher the better for impact modification) or on the rubber particle size in the sPS matrix (the higher the viscosity of the block copolymer the larger is the particle size under constant mixing conditions). A further possible explanation is that the polystyrene block length plays a critical role. As was stated earlier, above a critical molecular weight of 70000g/mol for aPS and l00 000g/mol for sPS, the influence of the molecular weight on the tensile strength of the homopolymer under conditions of rapid deformation is weak. This would indicate that with increasing molecular weight for the styrene blocks up to 70000g/mol the rubber particles will be increasingly better anchored in the amorphous sPS phase and presumably the sPS toughness will increase correspondingly. In Table 19.2 [5], the influence of the processing conditions on the notched impact strengths of pure and rubber-modified sPS are summarized. The test specimens were injection moulded at a mould temperature of 80, 120 or 140 °C. Some of the specimens were subsequently annealed at 180 °C in the mould. The notched impact strength of pure sPS seems to be uninfluenced by the mould temperature and thus by the orientation of the molecules, and remains brittle with values near 0.6 kJ/m 2 . In the rubber-modified polymer, the notched impact strength is highest when the lowest mould temperature is used. In these specimens crystallinity is likely to be the lowest and the orientation of the molecules the highest. On the other hand, injection moulding at a mould temperature of 120°C, which is above the glass transition temperature of sPS, increases the crystallinity of the product. This can be further increased by annealing the specimens. This dependence on the moulding conditions, well known from other semicrystalline polymers [16], is also valid for sPS. It would appear that increasing crystallinity combined with decreasing orientation embrittles the material, probably owing to the reduced amorphous phase where preferentially crazes are formed, and owing to the increased interaction between the propagating craze and the crystalline lamellae. The influence of lamellae on a propagating craze is shown schematically in Figure 19.9 and was verified by TEM images. Clearly the craze propagation is deflected by crystalline lamellae Table 19.2 Notched impact strength (a^) of styrene polymers with 35% Kraton G 1651 under different processing procedures (Inj.m. = injection moulded)

Parameter

Rubber-modified sPS

sPS

Inj.m. Inj.m. Annealed

Inj.m. Inj.m. Annealed

Melt temperature (°C) 290 290 30minat 180 290 °C 290 CC 30 min at 180 Mould temperature (°C)i 80 120 60 140 6.0 3.7 2.7 0.5 flk(kJ/m2) 0.6 0.6

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

421

Figure 19.9 TEM images and the corresponding sketch of the interaction between crazes and crystalline lamellae

or the lamellae are stretched or sheared, the details depending on the direction of the arriving craze relative to the orientation of the lamella. This additional and therefore, in principle, profitable increase in energy dissipation hampers craze propagation so that embrittlement results with increasing crystallinity. This stopping of the crazes can be confirmed by the TEM image shown in Figure 19.10, in which the tips of some crazes can be observed, whose propagation has been stopped along an imaginary line, which is presumably an invisible crystalline lamella. Usually the crazes run from rubber particle to rubber particle more or less perpendicular to the tensile direction. Loading/unloading/reloading experiments very often confirm, especially in PS products, the rubbery character of crazes. Typical is an S-shaped deformation curve in reloading after preloading the specimens to produce crazes. This behaviour is characteristic of crazes [17] and may be caused by the entropic elastic character of the PS fibrils in the crazes. Figure 19.11 shows this behaviour for Kraton-modified sPS for two mould temperatures using the injection moulding process. In both cases the craze structure is reflected by the S-shaped second deformation curve which is obtained after the linear curve of the first run, in which the crazes have been produced. The use of olefin rubbers [18] as good impact modifiers for sPS when used in conjunction with S—B or S—B—Sblock copolymers, which may be hydrogenated in the butadiene phase, has also been described. Instead of butadiene, isoprene can be used. Examples of the olefinic polymers are polyethylene, ethylenepropylene rubbers (EPR) and polypropylene—(ethylene-~propylene rubber) block copolymers. Here the styrene block copolymers presumably function as

422

G. E. McKEE ETAL

Figure 19.10 TEM images of rubber-modified sPS with craze tips along imaginary lines (- - -), presumably lamellae. Rubber modification: 35 % S—EB—S block copolymer tensile stress (Mpa)

tensile stress (Mpa)

Figure 19.11 Loading/unloading/reloading curves of rubber (35% Kraton)-modified sPS with entropic effects in reloading mode: (a) specimen injection moulded into a mould at 120°C; (b) annealed for 30min at 180°C

compatibilizers between the sPS and rubber phase. Thus with 75 parts of sPS, 25 parts of a block polypropylene containing 15 % of EPR and 15 parts of a styrene—isoprene block copolymer, where the isoprene is hydrogenated, an Izod

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

423

notched impact strength of 8.6kg cm/cm is obtained. This value compares favourably with values obtained using 20% S—B—S block copolymer, where the butadiene phase is hydrogenated [15]. Similar blends and also blends containing additionally polyphenylene ether and glass fibres are described in EP 767211 [19]. Instead of block copolymers, the use of pseudo-random linear copolymers of an aliphatic a-olefin and a vinyl aromatic monomer has been reported [20], where the styrene content of the polymer must be higher than 40 wt%. Preferred are styrene and ethylene copolymers. These blends may contain, amongst other things, an elastomeric olefinic impact modifier such as homopolymers and copolymers of a-olefins. Presumably the styrene—ethylene copolymer acts as a polymer emulsifier for the olefinic impact modifier. Using 5 wt% of an ethylene— styrene (30:70) copolymer and 20% of an ethylene—octene impact modifier in sPS, a tensile elongation (ASTM D638) of 25 % was obtained. Blends of syndiotactic styrene—p—methylstyrene copolymers (SPMS) with poly(styrene)-block-poly(ethylene-co-butylene)-block-polystyrene (SEBS) has been reported. [21] No significant effects on the tensile modulus and strength were observed for blends containing less than 10% SEBS. SEM of drawn samples of the blends showed that the dispersed SEBS phase had been extended to about the same extent as the bulk blend, indicating good adhesion between the two phases. EP 978 536 [22] describes the production of a compatibilizer by reactive extrusion. The elastomer contains reactive groups such as epoxy, anhydride and/or carboxylic groups which react during processing with reactive groups in polymers which are compatible or semicompatible with the sPS. Thus the impact resistance of sPS could be increased from 8 to 45 J/m with a blend of 70 parts of sPS, 25 parts of an ethylene—acrylate—glycidyl methacrylate terpolymer (Lotador AX 8900, Atochem) and 5 parts of a styrene—(ethylene— butene)—styrene block copolymer grafted with maleic anhydride (Kraton FG 190IX, Shell Chemicals). If additionally 0.12% of N,N-dimethylstearylamine was added to catalyse the reaction between the maleic anhydride and the epoxy group, the impact strength could be increased to 125 J/m. A similar system is also described in EP 587098 [23].

4.2

CORE-SHELL IMPACT MODIFIERS

The use of core-shell impact modifiers for sPS is also patented in EP 318 793 [15] (see Table 19.1). These impact modifiers are usually prepared using the emulsion polymerization process, although other methods such as the microsuspension polymerization process are possible. The core usually consists of polymers prepared from an acrylate, especially butyl or 2-ethylhexyl acrylate or butadiene. These rubber particles are then grafted with vinyl monomers, where

424

G. E. McKEE ETAL

for compatibility with the sPS matrix an outer graft shell of polystyrene is presumably advantageous. The use of core-shell impact modifiers combined with styrene—hydrogenated poly butadiene block copolymers in sPS is described by Rohm and Haas [24]. The core of the former type is of polybutadiene or its copolymer, the shell consists predominately of polystyrene. Rohm and Haas found that a synergistic effect is present and that the Izod notched impact strength is higher when both rubber types are used instead of only one. As discussed earlier, the energy dissipation on impact is predominantly via crazes and for craze formation large rubber particles (> 0.5 u,m) are usually more efficient than small particles. Such particles are described by BASF [25,26] and are made using the microsuspension polymerization system. Butyl acrylate monomer is dispersed in water to droplets with a diameter from 0.5 to several microns using high shear conditions. These monomer droplets are then polymerized, grafted with styrene and used as impact modifiers for sPS. Figure 19.12 shows the TEM images of the deformation pattern of sPS, modified with 35 % core-shell butyl acrylate rubber particles. As expected, crazes and voiding in the rubber particles dominate the deformation mechanisms. The notched impact strength (ISO 179/eA) of this material was measured to be for the injection moulded specimen (80 °C mould temperature) 4.3, for the injection moulded specimen (140 °C mould temperature) 3.3 and for the annealed specimen 1.9 kJ/m 2 .

Figure 19.12 TEM images of crazes and cavitated rubber particles after deformation of sPS rubber modified with 35 % core—shell BA-S rubber particles (BA:S = 60:40)

425

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

It would appear that these rubbers have a similar toughening effect as the S—EB—S block copolymers. However, a synergistic effect is present when a mixture of both impact modifiers is employed [27]. Figure 19.13 shows the dynamic mechanical properties of such a blend of sPS with a mixture of Kraton G 1651® (15 %) and microsuspension rubber particles (20 %) consisting of 60 % butyl acrylate (BA) core grafted with 40 % styrene shell (S//BA). The glass transition temperatures of the Kraton® (-60 °C) and the butyl acrylate (—45 °C) phases can be easily distinguished from one another. The TEM image of such a product after deformation is shown in Figure 19.14. The annealed specimen is shown since the two rubber types are better discernible than in the nonannealed sample. As expected, crazing and voiding in the rubber particles dominate. The product had the following notched impact strengths (ISO 179/eA): injection moulded (80 °C mould temperature) 6.3, injection moulded (140 °C) 4.0 and annealed 3.7kJ/m 2 . Again, as in the preceding examples, if the sPS in the specimen is given sufficient time to relax after injection moulding and to crystallize, then the notched impact strength is reduced. Above its glass transition temperature, shear processes dominate in sPS as was shown earlier. In Figure 19.15, the TEM images of deformed sPS modified with the core-shell modifier prepared using the microsuspension method are reproduced. At the surface of the specimen highly oriented rubber particle are discernible without voiding. In the inner part of the specimen, however, cavitated highly oriented particles have been formed. Crazes are not seen in the 2.0X1099 10

,10'

o

,5

o 5

10 r

2.0X10 4 -100.0 -50.0

0.0

50.0

100.0 150.0 200.0 250.0 300.0 Temp (°C)

Figure 19.13 Dynamic shear modulus (cycles/s) of sPS, rubber modified with a mixture of 15 % Kraton G 1651 ® and 20 % S//BA particles produced in microsuspension

426

G. E. McKEEET/lL

Figure 19.14 JEM images of the deformation structure in sPS, rubber modified with 15% Kraton® and 20% S//BA particles produced in microsuspension: (a) injection moulded sample; (b) annealed sample

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

427

Figure 19.15 TEM images of the deformation structure in sPS modified with S//BA core–shell (35 %) after deformation at 110 °C: (a) image taken from the region near the surface of the specimen; (b) image taken from the inner part of the specimen

428

G. E. McKEEETAL

internal part or near the surface. It is likely that voiding reduces the dilatational strain of plain strain in the inner part of the specimen, whereas in the outer region of the specimen near the surface plain stress prevails without dilatational strain.

4.3

PREPARATION OF sPS IN THE PRESENCE OF A RUBBER

Idemitsu [28] also described a process for preparing sPS in the presence of a rubber. A large selection of rubbers are claimed to be effective in increasing the impact strength of the sPS using this process; however, in the examples only butadiene-based rubbers are used. The yields of the polymerization were approximately 50 %. Little information was supplied about the mechanical properties of the sPS-rubber blends. This method has the advantage that the blending step for mixing rubber and sPS is eliminated, thus reducing damage to the polybutadiene rubber which occurs during processing at high temperatures. A similar process is described in EP 559 108 [29]. Here styrene is grafted on to a polymer containing polymerizable carbon–carbon double bonds in the side chain. The polymerization in the grafting step is chosen so that the resulting grafted polystyrene chains have a syndiotactic structure. In the examples the graft precursor is a copolymer of ethylene–propylene and p-(3-butenylstyrene), the latter serving as the grafting site for the sPS chains. These grafted rubbers can be used by themselves, as impact modifiers for sPS or as a compatibilizer between sPS and an ethylene–propylene rubber (EPR). In blends of sPS with an EPR, the compatibilizing effect of the graft polymer was shown using electron microscopy of the fractured surface of the blend. When the graft polymer was present, a much finer dispersion of the EPR in the sPS matrix was observed. In comparison with neat sPS, an increase in the notched Izod impact strength of 700% was observed when 20% of the graft polymer was employed. A similar process is described in EP 911 349 [30].

5

PRESENT SITUATION AND FUTURE PERSPECTIVES

SPS was discovered by Idemitsu and is now produced by Idemitsu on a pilotplant scale and by Dow Chemical in a plant in Germany with a capacity of 30 000–40000 t/a. The strong points of sPS are its low water absorption, high environmental stress cracking resistance, chemical resistance, high heat resistance, good form stability due the similar densities of its crystalline and amorphous phases and good flow behaviour [31]. Its Achilles' heel is, and will probably remain so in

RUBBER MODIFICATION OF SYNDIOTACTIC POLYSTYRENE

429

the near future, its inherent brittleness. For this reason, its main applications will probably use the glass-reinforced grades, which bring a certain increase in toughness when the direction of fracture is perpendicular to the orientated glass fibres. These glass fibre grades, however, will definitely need rubber modification to increase their impact resistance further. For some applications, especially where the currently used material is over-engineered and has a higher toughness than is required, substitution by sPS is possible. Work at BASF indicates that sPS in its neat form is more brittle than even standard polystyrene and from a toughness point of view sPS cannot be classified as an engineering plastic. After modification with 30% glass fibres, the situation is more complex. sPS with 30 % glass fibres from Idemitsu would appear to be less tough than similar glass fibre-modified PBT and nylon 66 [14] grades; however, another study found sPS with 30 % glass to be tougher than PBT also with 30% glass fibres [32]. The latter study however, made no comparison with polyamide. Biaxial tests where the crack formation is not perpendicular to the glass fibres may also cause toughness problems for sPS. From this it can be seen that an efficient and inexpensive impact modification of sPS is unavoidable and one of the main goals in its development. Judging by the number of patent applications in this field and owing to its availability, it would appear that a styrene–butadiene-styrene block copolymer, where the butadiene phase is hydrogenated, is the impact modifier of choice. It remains to be seen whether this deficit in toughness, for example in comparison with polyamide, can be compensated by its other advantages. This, together with the price of the impact-modified sPS, will presumably determine its future as a commercial plastic.

REFERENCES 1. Ishihara N., Kuramoto M., Uoi M., Macromolecules 21, 2464 (1986). 2. Vittoria V., Handbook of Thermoplastics, ed. Olabisi O., Marcel Dekker, New York, 1997, Chapt. 4, pp. 81–106. 3. Gvozdic, N. V., Meier, D. J., Polym. Commun. 32, 493 (1991). 4. Yan R. J., Ajji A., Shinozaki D. M., J. Mater. Sci. 34, 2335 (1999). 5. Ramsteiner F., McKee G. E., Heckmann W., Oepen S., Geprags M., Polymer 41, 6635 (2000). 6. Kanig G., Prog. Colloid Sci. 57, 176 (1975). 7. Lin L., Argon A. S., /. Mater. Sci. 29, 294 (1994). 8. Fellers J. F., Kee B. F., J. Appl. Polym. Sci. 18, 2355 (1974). 9. Jones M. A., Carriere C. J., Dineen M. T., Balwinski K. M., Polym. Mater. Sci. Eng., Proc. ACS Div. Polym. Chem. 67, 212 (1997). 10. Bucknall C. B., Toughened Plastics, Applied Science , London, 1977. 11. Lazzeri A., Bucknall C. B., J. Mater. Sci. 28, 6799 (1993). 12. Argon A. S., Pure Appl. Chem. 43, 247 (1975).

430

G. E. McKEEE7/\L

13. Bonnet M., Buhk M, Trogner G., Rogausch K.-D., Petermann J., Acta Polym. 49, 174 (1996). 14. Gausepohl H (ed.), Polystyrene, Hanser, Munich, 1996. 15. EP 318 793, Idemitsu. 16. Ulcer Y., Cakmak M., Miaco J., Hsiung C. M., J. Appl. Polym. Sci. 60, 669 (1996). 17. Kambour R. P., Koop R. W., J. Polym. Sci. Part A2 183 (1969). 18. EP 324 398, Idemitsu. 19. EP 767 211, Idemitsu. 20. US Pat. 6063872, Dow Chemical Company. 21.Yan R. J., Ajji A., Shinozaki D. M., Polym. Eng. Sci. 41, 618 (2001). 22. EP 978 536, Enichem SpA. 23. EP 587 098, Idemitsu. 24. EP 755972, Rohm & Haas Co. 25. WO 000 9608, BASF. 26. McKee G. E., presented at Technology and Markets for Metallocene Plastics Conference, 3–4 May 2000, a Slid Deutsche Kunststoff Zentrum, Wurzburg. 27. Geprags M., Mckee G. E., Wunsch J., presented at the 217th National Meeting of the American Chemical Society, Arnheim, CA, 21 March 21 1999. 28. EP 440 014, Idemitsu. 29. EP 559 108, Idemitsu. 30. EP 911349, Idemitsu. 31. Koch-Reuss U., Kunstsstoffe, 88, 1139 (1998). 32. Carriere C. J., Bank D., Malanga M., J. Appl. Polym. Sci. 67, 1177 (1998).

20

L. ABIS, R. BRAGLIA, G. GIANNOTTA AND R. PO Polimeri Europa SpA Istituto Donegani, Novara, Italy

1

INTRODUCTION

From the point of view of stereochemistry, polystyrene (PS) (see Section 6 for abbreviations) may exist in three different forms: atactic (aPS), obtained for the first time in 1839 (but only recognized in 1920) through a radical mechanism and currently one of the most important commodity materials; isotactic, obtained using Ziegler–Natta catalysts, discovered in the middle of the last century but never developed owing to its unsatisfactory properties; and syndiotactic (sPS), discovered in 1985 and obtained by using titanium complexes/methylaluminoxane or fluorinated borane catalytic systems [1,2]. The properties of sPS appeared to be so promising that it has recently reached the commercial stage after several years of intensive research. However, the inherent brittleness of sPS appeared to be a heavy limitation to many applications since the very beginning. To improve the ultimate properties of sPS, blending with other polymeric materials is a commonly applied technique and the literature on this topic has steadily increased in recent years. In this chapter the physical and mechanical properties of blends based on sPS and the patents relevant to their preparation and applications are reviewed and discussed.

2

OVERVIEW OF sPS PROPERTIES

Some typical physical and mechanical properties of sPS are reported below; some reported values, e.g. the melting temperature [3], can vary slightly for Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy f 2003 John Wiley & Sons Ltd

432

L. ABIS ET AL.

polymers synthesized with different catalysts, owing to the different content of steric defects. Density

1050kg/m3 (a-form 1070 kg/m3; ß-form 1033kg/m3)[4–6] ~ 100°C [7] 270 °C [3] 2 kJ/m2 (ASTM D 256-A) 41 MPa (ASTM D 638) 3.4 GPa (ASTM D 638) 1 % (ASTM D 638) 3 GPa (ASTM D 790-1) 75 MPa (ASTM D 790-1)

Glass transition temperature(rg) Melting temperature (Tm) Izod impact strength (notched) Tensile strength Tensile modulus Elongation at break Flexural modulus Flexural strength Heat deflection temperature under load (182 MPa) 96 °C (ISO 75) Vicat softening temperature 254 °C (ISO 306) Dielectric constant, 23 °C, 1 MHz 2.6 (IEC 250)

Whereas atactic PS is an amorphous polymer with a Tg of 100°C, syndiotactic PS is semicrystalline with a Tg similar to aPS and a Tm in the range 255275 °C. The crystallization rate of sPS is comparable to that of poly(ethylene terephthalate). sPS exhibits a polymorphic crystalline behavior which is relevant for blend properties. In fact, it can crystallize in four main forms, a, (3, -y and 8. Several studies [8] based on FTIR, Raman and solid-state NMR spectroscopy and WAXD, led the a and ß forms to be assigned to a trans-planar zig-zag molecular chain having a (TTTT)W conformation, whereas the y and 8 forms contain a helical chain with (TTG~G~)2 or (G+G+TT)2 conformations. In turn, on the basis of WAXD results, the a form is said to comply with a unitary hexagonal cell [9] or with a rhombohedral cell [10]. Furthermore, two distinct modifications called a' and a" were devised, and assigned to two limiting disordered and ordered forms, respectively [10]. The crystalline structure of the ß form shows an orthorhombic unitary cell [11] and, like the a form, may assume two limiting ordered and disordered structures, named ß' and ß" respectively. Both the crystalline cells of the y and 8 forms are monoclinic [12]. The 8 form always has some solvent molecules trapped in the crystalline lattice, whereas the "y form is empty. As widely reported, the formation of the above crystalline structures and the transitions between them may occur under appropriate thermal conditions either in the solid or from solutions phases [13]. Owing to its partial crystallinity, sPS is strongly resistant to concentrated acid and bases, oils and greases and most organic solvents, except chlorinated and aromatic compounds that cause swelling. Its density (about 1 g/cm3) is significantly low compared with other engineering thermoplastics such as polyamide-6, poly(butylene terephthalate) and poly(phenylene sulfide), and this is

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

433

advantageous from the point of view of cost/volume ratio. Moreover, since the densities of the crystalline and the amorphous phases are very similar, the material has a high dimensional stability and exhibits low warpage. The absence of polar groups along the polymer chain results in a low dielectric constant with respect to other common engineering thermoplastics. This represents a clear advantage in electric and electronic applications, but also has useful consequences in processing operations where a low moisture adsorption leads to short drying cycles (or none). The mechanical properties are similar to those of aPS, but the elastic modulus is enhanced owing to the crystallinity. Brittleness is typical of styrenic polymers, and this is the major drawback of sPS. Whereas the mechanical properties of aPS rapidly decay above Ts, those of sPS remain good. Addition of inorganic fillers (e.g. glass fibers) leads to a further improvement in thermomechanical properties and also impact resistance. The material can be processed through conventional techniques, such as extrusion (shaped profiles, foils, films, fibers), injection molding and thermoforming.

3

PATENT LITERATURE ON sPS BLENDS

To overcome the main drawbacks of sPS (e.g. poor impact resistance) without impairing the other thermo-mechanical properties (e.g. modulus, heat distortion temperature) and solvent resistance, extensive research has been carried out by blending and compounding it with suitable polymeric and inorganic components. Several patents have been issued on polymer blends having sPS as a main component. Table 20.1 reports the most relevant of them published by the US Patent Office in the period 1985-2000. Generally, the impact resistance of a polymeric material can be improved by blending with elastomers or other resilient polymers. However, owing to the poor miscibility often occurring among blend constituents, a suitable compatibilizer must be added to the formulation. Basically, only a limited number of material classes have been used as components in sPS-based blends. The most common toughening polymers are thermoplastic elastomers [styrene–diolefin (hydrogenated) block copolymers], olefinic elastomers (EPR, EPDM) and polyolefins. Whereas the first class shows some compatibility with sPS (Table 20.1, entry 14) owing to the presence of the styrenic block, the latter require the addition of a compatibilizing agent. Thermoplastic elastomers themselves can act as compatibilizers (for instance entry 15): in this case, the olefinic block of SEBS is thought to interact with the olefin (co)polymer, whereas the side styrenic blocks dissolve in the amorphous phase of sPS. According to the patent assignee, the result of such interactions is a tough material exhibiting a high modulus.

Table 20.1 No.

US patents on sPS blends issued in the period 1985-2000

Claimed blend composition

Example of component(s)

US Patent No.

(a)95-50%sPS (b) 5-50 % poly(phenylene ether)

(b) Poly[(2,6-dimethyl)-l,4-phenylene ether]

4946897

Montedipe/ Enichem

(a) 10-98%sPS (b)90-2%poly(phenyleneether) ([tj] > 0.28dl/g)

(b) Poly[(2,6-dimethyl)-l,4-phenylene ether]

5109068

Idemitsu Kosan Co.

(b) Polyethylene terephthalate); polyethylene terephtalate-co-poly(ethylene glycol) terephthalate]) (b) Polyethylene terephthalate)

5017657

Montedipe

5 395 890

Idemitsu Kosan

(a) 0.01-30% sPS (b) 99.99–70% thermoplastic polyester

(a) !-98%sPS (b) Polyester resin (c) Rubber polymer (optional) (d) Inorganic filler (optional) (a)99-l%sPS (b) 0.1-50% polyphenylene ether with polar groups or poly(phenylene ether) with polar groups + styrenic polymer (c) 0.9-98.9% polyamide (optional) (d) Inorganic filler (optional) (a)25-75%sPS (b) 75-25% polyamide (c) 2 30% rubber (d) Inorganic filler (optional) (a)sPS (b) Fumaric acid-grafted poly(phenylene ether) (c) Polyamide and/or rubber elastomer (optional) (d) Inorganic filler (optional)

Assignee

Co. (b) Maleic anhydride-grafted PPE; maleic anhydride-grafted PPE + SBR rubber (c) Nylon 66

5219940 5703 164 5777021

Idemitsu Kosan

(b) Nylon 6; nylon 66 (c) Maleated SEBS; maleated SBS

6013726

Idemitsu Kosan Co.

(b) Fumaric acid-grafted PPE (c) Nylon 66; SEBS

5952431

Idemitsu Kosan Co.

Co.

8

9

(a)20-80%sPS (b) 20–80% polycarbonate (c) Inorganic filler (optional) (a) sPS (b) Polycarbonate and/or poly(phenylene ether)

10

(a)90-10%sPS (b) 10–90% polycarbonate (c) 0.1-10% imidized poly(phenylene ether)

11

(a)95-5%sPS (b) 5–95% saponified ethylene–vinyl acetate copolymer

12

(a) 0.5-99 % homo- or copolymer of a vinylaromatic monomer (b) 0.5-99% aliphatic -olefin homo- or copolymer (c) 0.5-99 % random copolymer of a vinylaromatic monomer and 1 -olefin (a)sPS (b) Rubber (c) Polyphenylene ether or modified poly (phenylene ether) (d) Inorganic filler (optional) (a)5-98%sPS (b) 2-95 % partially hydrogenated styrene–diolefin block copolymer (c) Polyolefin (optional) (d) Polyphenylene ether (optional) (e) Inorganic filler (optional)

13

14

(b) Bisphenol A polycarbonate + PPE; PPE; maleic anhydride-grafted PPE

5889 069

(b) Bisphenol A polycarbonate; bisphenol fluorene polycarbonate (c) Maleic anhydride-grafted PPE reacted with N-cyclohexyl-1,3-propanediamine

5241015

5346 950

Idemitsu Kosan Co. Dow Chemical Co. Dow Chemical Co. Kururay/ Idemitsu Petrochemical Co. Dow Chemical Co.

(a) sPS (b) HDPE; LDPE; LLDPE; polypropylene; ethylene–propylene rubber (c) Ethylene-styrene pseudorandom copolymer

5460818

(a) sPS and/or maleic anhydride-grafted sPS (b) SEBS; SBS; EPR (c) PPE; maleic anhydride-grafted PPE

5391 603 5326 813 5444 126

Idemitsu Kosan Co.

(b) SEBS (c) Polypropylene; ethylene–propylene copolymer (d) PPE; maleic anhydride-grafted PPE

6005 050

Idemitsu Petrochemical Co.

(continues)

CO Ul

Table 20.1 (continued) US Patent No.

No.

Claimed blend composition

Example of component(s)

15

(a)sPS (b) Olefin (co)polymer (c) (Hydrogenated) styrene–diolefin block or graft copolymer (d) Polyphenylene ether (optional) (e) Waxes, fillers (a) Syndiotactic styrene (co)polymer (b) Modified rubber (c) Rubber (optional) (d) Polyphenylene ether or modified poly(phenylene ether) (optional) (e) Inorganic filler (optional) (0 Crosslinking agent (optional) (a)l-99%sPS (b) l-99%polyamide (c) 0.1-10% polymer compatible with (a) modified with polar groups (d) Rubber (e) Additives (a)10-90%sPS (b) 10-90%polyamide (c) 0.1-20% compatibilizing polymer for (a) and (b) (d) 0.1-10% rubbery polyolefin impact modifier (e) 0.1-5% domain forming rubbery polymer (0 0.1-10% polar group functionalized rubbery polyolefin (g) 0-5% compatibilizing polymer for (a) and (d)

(b) Polyethylene; polypropylene; ethylene-propylene rubber; EPDM rubber (c) SB; SEBS; SEPS (d) PPE; fumaric acid-grafted PPE

5 902 850 6031049

Idemitsu Petrochemical Co.

(a) sPS and/or maleic anhydride-grafted sPS (b) Maleic anhydride.grafted SEBS; glycidyl methacrylate-grafted SEBS; maleic anhydride-grafted EPR (c) SEBS; SBS; EPR (d) PPE; maleic anhydride-grafted PPE (0 Epoxy resin (b) Nylon 66 (c) Maleic anhydride-grafted PPE (d) SEBS; maleic anhydride-grafted SEBS

5 352 727 5 436 397 5418275 5 543 462 5 777 028

Idemitsu Kosan Co.

6013709

Idemitsu Petrochemical Co.

(b) Nylon (c) PPE (d) Ethylene-a-olefin copolymer (e.g. ethylene–octene) (e) Hydrogenated styrene-diolefin block copolymer (0 Maleic anhydride grafted ethylene–a-olefin copolymer (g) Carboxylated or maleated PPE

5990244

Dow Chemical Co.

16

17

18

Assignee

19

20

21

22

23

(a)sPS (b) (Hydrogenated) styrene-butadiene block copolymer (c) Thermoplastic resin (d) Inorganic filler (a) 1-95 % of sPS or modified sPS (b) 5–80% of thermoplastic resin (c) 1–50% of rubber or modified rubber (d) 0.1–10% of modified poly(phenylene ether)

(a)5-95%of sPS (b) 95–5% of OH-, COOH- or NH2-terminated thermoplastic polymer (c) 0.01-15% of polystyrene with epoxy or anhydride groups (a)25-90%of sPS (b) 1–50% of rubber polymer (c) 0.1–30% of nucleating agent 0–90%) of inorganic filler (d) 0–90 % of poly(phenylene ether) or modified poly(phenylene ether) (a)65–95%ofsPS (b) 5–35% of a blend containing 10–90% of hydrogenated styrene–butadiene block copolymer and 90– 10% of a core/shell modifier copolymer with a slightly crosslinked rubber core and a polystyrene shell

(b) SEBS; SBS; maleic anhydride-grafted SEBS (c) Nylon 66; polyarylate; polyethylene

5990217

Idemitsu Kosan Co.

(a) sPS or maleic anhydride-grafted sPS (b) Polyamide (nylon 6); polycarbonate; polyester (c) SEBS; SBS; methyl acrylate–butyl acrylate– siloxane–styrene core-shell rubber; maleic anhydride-grafted SEBS (d) Maleic anhydride-grafted PPE (b) Polyamide (c) Styrene-glycidyl methacrylate copolymer

5760 105

Idemitsu Kosan Co.

5270353

Idemitsu kosan Co.

(b) Hydrogenated styrene–diolefin block copolymer (d) PPE; maleic anhydride-grafted PPE

5391 603

Dow Chemical Co.

5654365

Rohm & Haas Co.

W

438

L. ABIS ET AL.

Ethylene-styrene pseudo-random copolymers (known as ethylene-styrene interpolymers) [14] have also been used to improve the compatibility between sPS and polyolefins, mainly polyethylene (entry 12). PPE is a key component of several blends. It is totally miscible with sPS and can interact with polar polymeric components, for instance polyamides (through hydrogen bonding interactions between PPE oxygens and amidic NH) and other condensation polymers; moreover, other important properties are improved by PPE, such as mechanical properties (entry 1), and a better control of crystallinity is obtained (see Section 4.1.1 for discussion). PPE also acts as a processing aid, improving the melt flowability [15]. In the blends discussed above, the interaction among the components and the compatibilizing agent is purely physical. However, to extend the range of possibilities for optimized blends preparation, reactive processing, where covalent chemical bonds are created between the partners, offers great potential. In this regard, PPE can be functionalized with an unsaturated anhydride (e.g. maleic anhydride), to promote the interaction with the polar groups of polymers; this kind of modification is also used on SEBS (entry 6). The anhydride groups may react with the — NH2 end groups of a polyamide, enhancing the dispersion in the polystyrene matrix (e.g. entry 5). As for polyesters (or polycarbonates), the reaction takes place with the — OH end groups; possibly ester internal moieties of the chain can also be involved in the reaction with — COOH groups arising from the anhydride. Hydroxyl and epoxide groups of some acrylic rubbers may also react with the anhydride. A similar reaction is exploited to link together maleated sPS and maleated rubbers by using epoxy resins (entry 16). A limited number of patents concern sPS blend which are not compatible. In these cases the properties that are described are functional ones and not related to the poor toughness of sPS. For instance, blends of sPS and partially saponified ethylene–vinyl acetate copolymers exhibit improved gas barrier properties (entry 11); small amounts of sPS added to poly(ethylene terephthalate) (although the patent actually claims a wide range of compositions) are useful to increase the polyester crystallization rate (entry 5). The scientific literature reports the theoretical aspects and the basic elements that influence the interactions among the blends containing sPS and their effects on the physical properties. In general, simple systems based on two or three components are considered.

4

MICROSCOPIC, THERMAL AND MECHANICAL PROPERTIES OF sPS BLENDS

Notwithstanding the interest found in patents about sPS blends with polymers suitable to improve their end-use properties, few scientific studies that system-

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

439

atically analyzed their physical properties have been reported in the literature. Mainly immiscible blends have been treated only recently and therefore the scientific knowledge acquired up to now should be considered absolutely preliminary, with a wide space remaining for further investigations. In the following the main scientific achievements obtained on both miscible and immiscible blends are illustrated. 4.1

MISCIBLE BLENDS

As fully described below, sPS has been found to be miscible with aPS, PPE, PVME, TMPC and styrene–l,l-diphenylethylene copolymer. Generally the reported investigations deal with the effect of the second component on crystalline features of sPS, such as polymorphic behavior, crystallization kinetics, morphology and growth rate of crystallites. Just one study reports on toughening sPS by adding suitable components. 4.1.1

sPS/PPE and sPS/aPS

4.1.1.1

Miscibility

For sPS/PPE blends, DSC and DMTA measurements give a single Tg value [16–19], intermediate between those of the components and dependent on composition. The Tg values of sPS and PPE being very different from each other (98 and 220 °C, respectively), this result constitutes an unambiguous proof of blend miscibility within the whole composition range. The same techniques cannot be applied to the case of sPS/aPS blends, as the two components have similar Tg values (less than 10°C difference). A higher resolution of close TB values can be derived from the isothermal heat capacity curves, measured in the vicinity of Tg by modulated DSC [20]. Furthermore, sPS and aPS, when annealed separately below Tg, exhibit in both DSC and modulated DSC distinct endothermic transitions owing to the enthalpy recovery [20,21]. Both methods, when applied to the sPS/aPS blends, give a single temperature for all compositions in agreement with the presence of a miscibility between the components. Additional information on the miscibility of sPS/aPS blends was gained by measuring with DSC the crystallization kinetics and the melting and crystallization peaks [20]. 4.1.1.2

Crystalline polymorphism

The addition of miscible components (PPE, aPS, etc.) to sPS may affect the distribution of a and ß structures formed from melt-crystallized samples.

L. ABIS ET AL.

440

Detailed analyses on this aspect, conducted mainly by WAXD and DSC under both nonisothermal and isothermal crystallization conditions, are outlined below. Nonisothermal crystallization Guerra et al. [16] reported on sPS/PPE blends of various compositions, prepared by compression molding (rmax = 290°C and tmax = lOmin) and then cooled to room temperature at a rate of 10°C/min. Their DSC results show that, below 50 wt%, sPS is completely amorphous. Moreover, WAXD spectra of the same samples indicate that the amount of the a form, which is 100 wt% in neat sPS, decreases on increasing PPE and Tmax (Figure 20. la and b) in favor of the ß form. The authors reported that the loss of the memory of the a form, which they suggested for melt-crystallized neat sPS, is more rapid for the same temperature and time when PPE is present [16]. Recently, Park et al. [7] reported on the nonisothermal crystallization behavior of sPS in sPS/aPS blends as studied by DSC, WAXD and OM. After melting at 300 °C for 5 or 10 min, the samples were cooled to 200 °C at variable rates (l–5°C/min) and then quenched in liquid nitrogen for WAXD experiments, or reheated to 300 °C at 10°C/min for DSC measurements. WAXD proved that sPS crystallizes only in the ß' form, which is not altered by the presence of aPS. Further, during the DSC heating run, two melting endotherms are found in the range 263–273 °C. They are assigned in order of increasing temperature to the ß' form and to the melting of more perfect crystals which recrystallize in the melting range. It was also found that the addition of aPS retards the crystallization and the recrystallization of sPS, makes the sPS crystals less perfect, but does not change the ultimate crystallinity. Pa 100

50

k300°C \

- 10

300°C

\, . V0 (a)

10

20

30

PPE (%)

40

50

280 (b)

300

320

Tm

Figure 20.1 Percentage content (Pa) of the a form in the sPS crystalline phase of sPS/PPE blends: (a) vs blend composition for different Tmax values; (b) vs Tmax for different compositions (wt% PPE). Dashed line: degree of crystallinity (/) vs composition at Tmax = 290°C. Reprinted by permission of John Wiley & Sons, Inc. from Ref. 16

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

441

Isothermal crystallization Injection molded sPS/PPE blends having different composition [19], melted at rmax = 300 °C for rmax = 5 min and then crystallized isothermally for 60 min at various rc; (from 232 to 244 °C), were investigated by means of DSC and WAXD and compared with pure sPS. DSC measurements do not show a melting peak for sPS < 40wt%, suggesting the absence of crystallinity. In contrast, at higher contents (80 wt%) three separate melting endotherrns (labeled I, II and III) between 260 and 271 °C are clearly found, as in pure sPS (Figure 20.2a). A comparison between WAXD patterns (Figure 20.2b) and DSC allows the above peaks to be assigned, in order of increasing temperature, to (3, a and an unstable form. The last form arises from a crystallization–remelting (see ref. 22 for pure sPS) process occurring during the DSC measurement. On increasing Tci, the area of (3 increases whereas that of a decreases. Moreover, both the DSC and WAXD patterns show at the same Tci a content of the (3 form higher in the blend than in pure sPS. Similar studies [23,24] performed with WAXD on sPS/aPS and sPS/PPE blends of various compositions, melted at Tmax = 320°C and then crystallized isothermally at various Tci (242–250 °C for 4 h), show instead that only the ß form is present. Correspondingly, DSC exhibits only melting peaks I and III, and peak II, assigned to a, is practically absent. A comparison between the two kinds of blends suggests that the formation of the a form is entropically inhibited by the random distribution of sPS chains in the other component, rather than by a strong sPS–PPE interaction, as suggested by Guerra et al. [16].

4.1.1.3

Phase structure and morphology

The intimate structure of the miscible phases and the morphology of the growing spherulites have been investigated by several authors by means of optical microscopy. Cimmino et al. [25] reported that films of sPS/PPE samples (80:20 wt%) are homogeneous either in the melt (340 °C) or after isothermal crystallization (240 °C). Moreover, since PPE does not show any trace of segregation in the interspherulitic region, it probably remains occluded in the interlamellar region of sPS, thus suggesting miscibility at a molecular level. The crystallites have a spherulitic morphology and are larger than in pure sPS. This behavior is attributed to a decrease in the nucleation capability of sPS, arising from the dilution in PPE. Woo and Wu [23] compared the properties of pure sPS with sPS/aPS and sPS/PPE blends, melted at J"max = 320°C for f max = 5 min and then isothermally crystallized at Tci = 245°C for 4h. Whereas in sPS/aPS the diameter of spherulites is close to that of pure sPS treated under the same conditions (ca 20 |xm at 238 °C), in sPS/PPE it is larger (30–40 fxm at 238 °C), with a maximum effect for PPE contents of 10-20wt%. Such a behavior is attributed to two

442

L. ABIS ET AL.

TC=244°C

240°C

236°C

232°C

I II III 230

250

(a)

12 (b)

270 Temperature, °C

16 20 2 9 (degree)

290

24

310

28

32

Figure 20.2 sPS/PPE blends (75:25 wt%) at different isothermal crystallization temperatures, (a) DSC thermograms; (b) WAXD patterns. Reprinted from Polymer, vol. 39, Hong B. K., Jo W. H., Lee S. C., Kim J., 'Correlation between melting behavior and polymorphism of sPS and its blends with PPE', p. 1793, Copyright 1998, with permission from Elsevier Science

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

443

possible reasons: the higher Tg of the sPS/PPE blend with respect to sPS may decrease the number of crystallizable nuclei and/or the dilution in PPE may reduce the impingement of sPS crystallites. The morphology of sPS/aPS blends was recently investigated by Park et al. [7] with OM to understand the disposition level of aPS in the sPS spherulites. OM micrographs were obtained on samples melted at 300 °C for lOmin, nonisothermally crystallized by cooling at 2 °C/min and then quenched to room temperature. It appears that during crystallization of sPS, aPS molecules are rejected from the growing fibrous crystals and progressively accumulate into the interfibrillar region of growing spherulites. Furthermore, low molecular weight fractions of aPS occupy the interspherulitic region by retarding the radial growth of sPS spherulites. Both phenomena are considered by the authors to be responsible for the slowing of crystallization that occurs on addition of aPS.

4.1.1.4

Kinetics of crystallization

Cimmino et al. [25] reported that the radial growth rates of crystallization G, measured in sPS/PPE blends, decrease strongly with increase in PPE content (Figure 20.3). This effect might arise from an increase in the transport free energy of crystalline segments in the melt, due the larger Tg of the blend compared with pure sPS, or to a decreased capability of sPS to nucleate, induced by its dilution in PPE.

260

Figure 20.3 Spherulite growth rate (G) for sPS/PPE and sPS/PVME blends as a function of the crystallization temperature 7^ (•) sPS; (Q) sPS/PPE 90:10; (•) sPS/ PPE 80:20; (A) sPS/PVME 80:20; (0) sPS/PVME 70:30; (+) sPS/PVME 50:50. Reprinted from Polymer, vol. 34, Cimmino S., Di Pace E., Martuscelli E., Silvestre C., 'sPS based blends: crystallization and phase structure', p. 2799, Copyright 1993, with permission from Elsevier Science.

444

L. ABIS ETAL.

Wu and Woo [26] compared the isothermal kinetics of sPS/aPS or sPS/PPE melt crystallized blends (Tmax = 320°C,tmax = 5min, Td = 238-252°C) with those of neat sPS. Crystallization enthalpies, measured by DSC and fitted to the Avrami equation, provided the kinetic rate constant k and the exponent n. The n value found in pure sPS (2.8) points to a homogeneous nucleation and a three-dimensional pattern of the spherulite growth. In sPS/aPS (75:25 wt%) n is similar (2.7), but it decreases with increase in sPS content, whereas in sPS/PPE n is much lower (2.2) and independent of composition. As the shape of spherulites does not change with composition, the decrease in n suggests that the addition of aPS or PPE to sPS makes the nucleation mechanism of the latter more heterogeneous. In the sPS/PPE (75:25 wt%) blend, the crystallization rate vs Tci shows maximum at 238 °C close to the value for neat sPS, but decreases with increasing the PPE content. Moreover, crystallization starts after an initial delay that increases with increase in Tmax (from 290 to 340 °C). For sPS/aPS and sPS/PPE blends described in Section 4.1.1.2, Woo and Wu reported [23] that by increasing Tci the onset of the crystallization is delayed and that by addition of either PPE or aPS G decreases. However, an interpretation based on the increase in the transport free energy in the melt due to the higher Tg is ruled out by the fact that aPS and sPS have the same Tg. Hong et al. [21] also observed for sPS/aPS blends melted at rmax = 300°C for tmax = 5 min, and then rapidly cooled, a decrease in G with addition of aPS. However, as the Tg values of sPS and blends are very close, the influence of the transport energy is explicitly ruled out, while it is suggested that the dilution of sPS in aPS increases the free energy associated with the formation of a nucleus of critical dimensions. Analogous measurements were performed by Bonnet et al. [20] on an sPS/ aPS (50:50 wt%) blend, crystallized from both the melt and glassy state. From the glassy state and for Tci less than 110 °C, G is proportional to the square root of time, as expected for a diffusion-controlled growth. In contrast, again from the glassy state but at Tci = 120°C, and from the melt at TC1 = 249° C, the growth is constant with time, in agreement with an interface-controlled growth. Duff et al. [21] reported a study made by means of DSC and WAXD on SPS/ PPE blends of various compositions, precipitated from ethylbenzene solutions, compression molded at 330 °C for 2 min and then slowly cooled to room temperature. In particular, the WAXD patterns show that in sPS-rich blends (>50:50 wt%) sPS is in a ß or ß' form, while small amounts of a are present in the 50:50 wt% blend. The kinetics of crystallization and the mechanism of nucleation of sPS were investigated under isothermal and nonisothermal conditions as a function of blend composition and molecular weights of the components. The experimental curves show that the half-time to crystallization, t\/2, increases with increasing content and molecular weight of PPE, but is not influenced by the molecular weight of sPS. The crystallization kinetics were

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

445

also analyzed with the Avrami equation and the Nakamura model to provide, in addition to k and n, the nucleation constant A and the diffusion coefficient U. On the basis of the above analysis, the authors concluded that crystallization proceeds through a heterogeneous nucleation mechanism followed by a threedimensional radial growth. In addition, since the diffusion coefficient U and half-time to crystallization increase at high molecular weights of PPE, it is suggested that the lack of flexibility of PPE chains (high Tg), combined with a large number of chain entanglements present at high molecular weights, inhibits the diffusion of the amorphous PPE chains away from the front of crystallization, and as a consequence the growth rate of sPS crystallites.

4.1.1.5

Toughening of sPS

Choi et al. investigated the toughening effect of rubber in sPS/PPE blends obtained via reactive extrusion [17]. Whereas the impact strength of sPS/PPE blends has a value intermediate with respect to the two components, the addition of a reactive polystyrene containing 5 wt% of oxazoline comonomer and MA-SEBS (SEBS grafted with maleic anhydride) as impact modifiers produces a synergic effect on toughening. In particular, the best results are obtained with a two-step procedure: reactive mixing of the modifier components, followed by reactive blending with sPS. In this way an almost sixfold enhancement of the impact strength (29 kJ/m) was measured for the sPS/PPE 70:30wt% blend added with 20 phr of MA-SEBS (0.4 phr of maleic anhydride) and 20 phr of oxazoline-modified PS (1.0 phr of oxazoline). DMTA measurements (large reduction in G' and a maximum loss tangent shifted to lower temperature) and rheological measurements (lower values of G' and complex viscosity) point to an enhanced interfacial activity and effective compatibilization induced by the additives during the extrusion process. The finer and more uniform dispersion observed by SEM confirms this interpretation.

4.1.2

sPS/PVME

4.1.2.1

Miscibility

Cimmino and co-workers [25,28] investigated by means of solid-state NMR and DSC the dependence of miscibility on composition and temperature in sPS/ PVME blends. The blends, prepared by casting a solution from o-dichlorobenzene at 130 °C, are found to be immiscible for PVME>20 wt%, in contrast with the miscibility found for aPS/PVME blends. In fact, DSC experiments show two Tg values corresponding to an sPS-rich phase (83:17 wt%) and a PVMErich phase (13:87 wt%). The lack of miscibility is also confirmed by the absence,

446

L. ABIS ETAL.

13

in C NMR solid-state experiments, of cross-polarization from sPS protons to PVME methoxy carbons [28]. Later, Mandal and Woo [29] demonstrated that this system is miscible, and exhibits a behavior equivalent to aPS/PVME blends. The previously found immiscibility is due to the relatively low value of the lower critical solution temperature, which in 50:50 wt% blends induces a phase separation already at temperatures of ca 120°C. In fact, OM, SEM and DSC, applied to blends (70:30 and 50:50 wt%) prepared by casting films from 1-2% solutions of chloronaphthalene at about 120°C, or by precipitation from the same solution with «-heptane, show a substantial homogeneity. However, OM measurements, performed at various temperatures on a series of samples, show a cloud point at ca 120 °C and above, indicating the onset of segregation. At higher temperature (samples briefly treated at 300 °C and then quenched), DSC shows two Tgs at -30 °C (attributed to PVME) and at 95 °C (attributed to sPS), shifted with respect to the pure compounds and corresponding to two partially miscible phases, one rich in PVME and the other rich in sPS. Under slow cooling the process appears to be reversible.

4.1.2.2

Phase structure and morphology

According to Cimmino et al. [25], OM measurements show that in the melt (340 °C) demixing of PVME occurs at all compositions, and during crystallization segregated PVME particles remain occluded in the interspherulitic region of sPS crystals. At the end of the crystallization process, an increase in the spherulite dimensions is observed, indicating that PVME addition reduces the nucleation ability of sPS. Measurements carried out at different crystallization temperatures (Figure 20.3) show that, at the same Tci, G increases in the PVME blend with respect to neat sPS and it is concentration independent. This behavior is accounted for by considering that a lower energy is needed to transport the molecules in the melt, the Tg of the blend being lower than in neat sPS. The independence of concentration seems, according to the authors, to arise from the phase separation which keeps the composition of the sPS-rich phase constant.

4.1.3 4.7.3.7

Other Blends sPS/TMPC

Investigations carried out by DSC [30] on a series of sPS/TMPC blends show a single Tg transition at all compositions, as expected for a miscible blend. The addition of TMPC generally retards the crystallization of sPS, and at contents

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

447

greater than 40 wt% the blends become amorphous. On samples crystallized isothermally at various temperatures, Tm values and implicitly lamellae thickness increase with increase in sPS content. The equilibrium values of Tm, calculated with the Hoffman-Weeks method, give an interaction parameter between polymers of —0.92 J/cm3.

4.7.3.2

sPS/S-DPE

Gausepohl et al. [31] investigated the behavior of blends between sPS and random styrene-l,l-diphenylethylene copolymers obtained by anionic synthesis. The blends were miscible for copolymer contents of 1,1-diphenylethylene lower than 15 wt% as indicated by the occurrence of a single Tg (114°C). Tm and crystallization rate were not influenced. 4.2

IMMISCIBLE BLENDS

As discussed in the first part, blends containing immiscible components such as polyolefins could improve the performances of the inherently brittle sPS. Until now the reported investigations have concerned simple binary blends containing a polyolefin and sometimes SEBS as a compatibilizer. In addition, sPS/ polyurethane and sPS/sulfonated sPS blends were also investigated. All these studies tried to correlate the microscopic features of the blends with their mechanical properties.

4.2.1

sPS/Polyolefins

Abis and co-workers [32,33] reported on a multi-technique characterization of immiscible blends between sPS and several polyolefins, prepared in a mixer at 285 °C and thermo-compressed at the same temperature. SEM performed on the cryo-fractured surface of the blends shows a gross phase separation, while the presence of voids due to the easy pull-off of particles from cryo-fractured surfaces points to a lack of adhesion between components (Figure 20.4a). The particles of polyolefins present in 80:20 wt% blends have dimensions decreasing in the following order: PP « LLDPE > HDPE > EPR > PIB. Coherently, as expected for immiscible blends, Tg values measured by DSC show very small variations with respect to the pure components while the mechanical properties degrade with respect to neat sPS. In particular, for minimum polyolefin contents <40 wt%, uniaxial tensile tests revealed a decrease in Young's modulus, elongation at break and energy to break. For higher contents, a phase inversion of the morphology occurs and the blend properties approach progressively those of the pure polyolefins.

448

4.2.2

L. ABIS ETAL.

sPS/HDPE and sPS/HDPE/SEBS

To promote the compatibility between the components, sPS/HDPE binary blends were mixed at 285 °C in a mixer with 5 and 10 wt% of SEBS and then thermo-compressed at the same temperature [33]. Crystallinity values, 7ci and Tm of sPS, as measured by DSC, decrease slightly with increase in SEBS content, in agreement with the expected miscibility of SEBS polystyrene block into the sPS amorphous phase. More significant evidence of the compatibilizing effect of SEBS comes from SEM of cryo-fractured surfaces. In fact the 72:18:10wt% blend (Figure 20.5a), when compared with the corresponding 80:20 wt% blend (Figure 20.4a), shows a noticeable decrease in dispersed particle size, which agrees well with a decrease in the interfacial tension. In addition, the absence of discernible interfacial boundaries and the lack of voids point to an improved adhesion at the interface. In the 18:72:10 wt% blend a different behavior is observed (Figure 20.5b). In fact this sample, although characterized by the absence of voids and sharp boundaries between particles, exhibits dimensions of the dispersed sPS particles not different from those of the 20:80 wt% sample not containing SEBS (Figure 20.4b). These features could be explained by considering that sPS molecules, as they crystallize at a temperature (225 CC) at which the other components are liquid, tend to aggregate into larger particles. Moreover, in the 76:19:5wt% blend, TEM reveals at the interface a black layer due to SEBS stained by reaction with osmium tetraoxide (Figure 20.6a). In the 72:18:10 wt% blend (Figure 20.6b) the excess of SEBS forms inclusions into HDPE. It is likely that crystallization of sPS, which occurs at 225CC, forces SEBS to segregate from the sPS domain, thus favoring its inclusion in the HDPE phase. The presence of SEBS at the interface was confirmed by small angle neutron scattering (SANS) experiments run on blends where the dispersed component, either sPS or HDPE, had been alternatively deuterated [34]. DMTA measurements also give similar results. The storage shear modulus G' for the pure HDPE, sPS and SEBS samples plotted against temperature (Figure 20.7) shows, at –140 °C, that the modulus of HDPE is considerably higher than that of sPS, whereas that of SEBS is comparable to the value for sPS. Accordingly, the addition of HDPE to sPS increase the G' value of sPS/HDPE blends measured at —140 °C (Figure 20.8a) and conversely the addition of sPS to HDPE decreases the modulus of HDPE. The same trend is observed on sPS/ HDPE blends with 10% of SEBS added. However, significant differences occur, which are better highlighted by plotting the G' values taken at —140 °C for both series of blends as a function of the sPS/HDPE weight ratio (Figure 20.8b). On this graph a decrease in G' when SEBS is added to HDPE-rich samples (sPS/HDPE weight ratio <1) is observed, owing to the increased

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

449

Figure 20.4 SEM of sPS/HDPE blend cryo-fractured surfaces: (a) 80:20 wt%, (b) 20:80 wt%. Reprinted from Ref. 33 by permission of Wiley-VCH

450

L.

ABIS ETAL.

Figure 20.5 SEM of sPS/HDPE/SEBS crio-fractured surfaces: (a) 72:18:10 wt%; (b) 18:72:1 Owt%. Reprinted from Ref. 33 by permission of Wiley-VCH

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

451

Figure 20.6 TEM of sPS/HDPE/SEBS blends: (a) 76:19:5 wt%; (b) 72:18:10wt%. Reprinted from Ref. 33 by permission of Wiley-VCH

452

L. ABIS ETAL 1010,

109-.

108-,

107,

106 -150

-100

-50

0 7(C)

50

100

150

Figure 20.7 Real part of the dynamic shear modulus G' plotted against temperature for (1) sPS, (2) HOPE and (3) SEBS. Reprinted from Ref. 33 by permission of Wiley-VCH

volume of the dispersed soft phase (sPS plus SEBS). On the other hand the addition of SEBS to sPS-rich blends (sPS/HDPE weight ratio 4) leads to an increased modulus. In analogy with a study reported by Sjoerdsma et al. [35] for PS/LDPE/SEBS blends, such a behavior could be explained by assuming that in the binary blends, at — 140°C, dispersed HOPE domains shrink more than the surrounding sPS matrix with a reduction in the mechanical coupling between the two phases. However, on blends with SEBS added, where such an effect is counteracted by the improved adhesion at the interface promoted by the copolymer, a higher modulus is obtained. Notwithstanding the favorable morphology observed by electron microscopy, no improvement in the mechanical properties is observed (Table 20.2). In fact, for HOPE contents >40 wt% the elongation values remain lower than 1 %, whereas the tensile strength and the energy to break decrease. Only for higher contents does HOPE become the matrix and a gradual recovery of mechanical parameters, in particular of the elongation at break, is observed. A similar study performed on aPS/HDPE/SEBS blends [33,36] showed an improvement in both the morphological and mechanical properties of the blend. This different behavior is tentatively explained in Figure 20.9 where phase transitions occurring when the blend is cooled from the melt state to room temperature are sketched. When sPS and aPS blends are at 285 °C, each end-block constituting SEBS (PS and EB) interacts and penetrates into the chemically similar blend component (sPS, aPS and HOPE, respectively), by forming a mixed interface with extended

POLYMERIC BLENDS BASED ON SYNDIOTACTIC

POLYSTYRENE

453

2CL,

O

(a)

-60 -160 -140 -120 -100 -8 7TQ

-40

-20

0

2.5-

2.0-

1.5

(b)

2 3 sPS/HDPE, wt.ratio

Figure 20.8 (a) Real part of the dynamic shear modulus G' plotted against temperature for sPS/HDPE/SEBS at different compositions (wt%): (1) 0:90:10; (2) 18:72:10; (3) 36:54:10; (4) 54:36:10; (5) 72:18:10; (6) 90:0:10. (b) Real part of the dynamic shear modulus G' measured at -140°C for sPS/HDPE (solid symbols) and sPS/ HDPE/SEBS (open symbols) plotted against sPS/HDPE weight ratio. Reprinted from Ref. 33 by permission of Wiley-VCH

interchain entanglements. As the aPS blend is cooled from the meltdown to about 100°C, SEBS polystyrene blocks and aPS vitrify together, thus maintaining the entanglements developed during the melt mixing. At almost the same temperature very fast crystallization of HDPE occurs which immobilizes the SEBS aliphatic blocks within the polyethylene semicrystalline structure. Both effects result in a strengthening of the interfacial adhesion between the blend components. In the sPS blends the same phenomena are likely to occur at 100°C. However, when 225 °C is reached from the melt, the polystyrene end-blocks of SEBS disentangle from the fraction of sPS chains which crystallize at the interface.

454

L. ABIS ETAL ATACTIC PS SEES interface

aPS bulk

SYNDIOTACTIC PS HOPE bulk

melt mixing I 285°C|

sPS bulk

SEBS interface

melt

3

HOPE bulk

1

sPS I crystallization

|225°C

liqf,

melt

melt crystalline

liquid

HDPE crystallization 100°c I

V V I N glassy

aPSl vitrification

amorphous (liquid)

IsPS vitrification

' J crystalline

crystalline

amorphous (liquid)

amorphous (glassy)

.

liq..

V

J crystalline + amorphous (liquid)

Figure 20.9 Liquid/melt to solid transitions of sPS/HDPE/SEBS (right) and of aSP/ HDPE/SEBS blends (left), occurring from 285 °C down to room temperature. Reprinted from Ref. 33 by permission of Wiley-VCH

This process seems to be extended enough to produce a significant weakening of the interface bonding on the sPS side. The load applied in the tensile test to aPS/HDPE/SEBS blends generates in the aPS phase a stress which is efficiently transferred to HDPE dispersed particles through the thick interface and delays the brittle failure of aPS phase to higher elongational values. In the case of sPS/ HDPE/SEBS blends, the poor interfacial adhesion inhibits the stress transfer to the HDPE phase and causes the formation and propagation of cracks throughout the sPS domains which lead to brittle failure of the sample. However, the kind of technique used to mold these blends seems to be important in determining their mechanical properties. In fact, mechanical tests carried on injection molded samples [37] show, with respect to compression molded samples, a significant enhancement of the energy to break for all samples (Tables 20.2 and 20.3). Moreover the addition of 10 wt% of SEBS to an 80:20 wt% sPS/HDPE blend involves in the injection molded samples an increase in both the energy at break and the Izod impact strength, whereas in the thermo-compressed samples no improvement is observed. Differences between compression and injection molded samples are widely acknowledged [38] and

455

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

Table 20.2 Tensile tests on compression molded sPS/HDPE and sPS/HDPE/SEBS blends: Young's modulus (E); strength (trb), elongation (ej,) and energy at break (Eb); elongation at yield (ey). Reprinted from Ref. 33 by permission of Wiley-VCH sPS (wt%)

HDPE

SEBS

(wt%)

(wt%)

E (GPa)

Ob

1 1 1 1 60 700

3.52 2.62 1.81 1.48 1.24 0.91

34 24 14 16

5 5 5 5 5 5

2.95 2.42 1.80 2.46 0.98 0.72

44.7 23.3 15.3 9.7 17.4 15.2

90 72 56 36 18 0

0 18 34 54 72 90

10 10 10 10 10 10

2.49 1.89 1.45 1.06 0.80 0.50

29.4 13.1 11.1 15.7 17.1 12.9

aPS 95 76

HDPE

SEBS

E 2.75 2.16

<7b

100 80 60 40 20 0

0 20 40 60 80 100

95 76 57 38 19 0

0 19 38 57 76 95

5 5

0 19

(%)

(MPa)

16

31 ;nd" 29.7 ; 30.2

Eb x 1000

(J/mm1 ) 17 12 6 8 28 64

2.3 1.0 0.9 0.9 5.3

14 10 6 7 43

708.0

2240

1.3 0.7 0.9 2.3 40.0

19 7 5 15 60

825.0 eb;ey

3040

1.2 ;nd 4.0 ; 2.2

Eb 19.5 39

nd = not detected.

Table 20.3 Mechanical tests on injection molded sPS/HDPE/SEBS blends. Symbols as in Table 20.2 sPS (wt%)

HDPE

SEBS

£b

(wt%)

E (GPa)

ffb

(wt%)

(MPa)

(%)

Eb x 1000(//mm2)

80 76 72

20 19 18

0 5 10

4.14 3.53 3.00

40 48 44

1.5 1.9 2.1

43 70 89

Izod (J/m) 13 26 42

are attributed also to orientation and to additional mixing occurring in the injection process. In our case we found that crystallinity, as measured by DSC, is around 50% for compression and around 30% for injection molded samples. It is likely that in the injection molded blends, owing to the higher content of amorphous sPS, stronger entanglements are formed with the polystyrene endblocks of SEBS [39].

456

4.2.3

L. ABIS ETAL

sPS/SEBS

Ramsteiner and co-workers reported on a morphological and mechanical characterization of sPS/SEBS blends prepared in an extruder and injection molded [40] (see also Chapter 19). With respect to neat sPS, they found a remarkable increase in toughness measured with the Izod test (ca 0.6kJ/m2 without SEBS, > 6 kJ/m 2 with SEES) and with energy release experiments. A fine dispersion and good adhesion of the rubber at the interface induced by the compatibilizing effect of SEBS were evidenced by TEM and considered a prerequisite for the increased toughness. Other factors responsible for this improvement are the orientation induced by the injection molding process, a reduced crystallinity due to rapid cooling and high molecular weights. In the same work mechanisms of deformation under tensile strength were explored by means of SEM, TEM, dilatational and hysteresis measurements. Abis et al. [32] also obtained evidence on compression molded blends of sPS/ SEBS of the occurrence of a phase compatibility between the components arising from the solubility of the polystyrene end-block of SEBS with the amorphous phase of sPS. In fact, although immiscible, a very fine dispersion and adhesion of the rubber particles is observed on SEM. However, contrary to the previous case, no improvement in the mechanical properties of sPS measured by tensile tests is observed, probably owing to the poorer performances of thermo-compressed samples than injection molded samples [38,39]. Blends of a syndiotactic styrene-p-methylstyrene copolymer with SEBS have been reported recently [41]. The blends, mixed in the melt at 280 C and then thermo-compressed at the same temperature, were investigated by DSC, SEM, spectroscopic birefringence and tensile tests. The blend morphology was found to be strongly dependent on mixing time and composition. In particular, SEM shows the presence of a two-phase morphology, where the size of SEBS particles is larger at higher contents (10 u-m at 20 wt%). Addition of SEBS in amounts greater than 10 wt% causes a decrease in the crystallinity, the tensile strength and the Young's modulus, but an increase in the elongation at break and therefore the related toughness. Blends uniaxially drawn at 110 =C exhibit a modulus, strength and energy at break higher than in the isotropic samples.

4.2.4

sPS/EPR and sPS/EPR/SEBS

Hong et al. [42] studied the effect of SEBS (LMW-SEBS with molecular weight 50000 and HMW-SEBS with molecular weight 175000) on the morphological, mechanical and rheological properties of sPS/EPR immiscible blends (80:20 wt%), melted at 300 °C for 15min and injection molded. In the binary blend, SEM showed the presence of large domains well separated at the interface. With the compatibilizer, the dispersion of EPR becomes narrow and

457

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

uniform. However, up to 2 wt% of SEBS, the results are better for LMW-SEBS than for HMW-SEBS, whereas at higher contents the effect of the compatibilizers is similar. The impact strength becomes 3.3 times higher on adding LMWSEBS whereas it improves only marginally on adding HMW-SEBS (Figure 20.10). Microscopic analysis of fractured surfaces indicates, in agreement with the above results, better adhesion between the polymeric components in the first case and poor adhesion in the second case. Addition of the compatibilizer also influences the Theological properties of the blend. In fact, complex viscosity, measured at 300 °C at variable frequency, increases upon addition of LMWSEBS, reaching approximately the value for pure EPR, whereas it changes only slightly with HMW-SEBS. Overall, these results show clearly that LMW-SEBS is more effective than HMW-SEBS in improving adhesion. Such a behavior is attributed to better penetration of LMW-SEBS blocks into the correspondent phase, rather than to better migration of the copolymer at the interface.

4.2.5

Other Blends

4.2.5.) sPS/polyurethane Thermoplastic polyurethane containing 4,4'-diphenylmethane diisocyanate, poly(tetramethylene glycol) and 1,4-butanediol has good mechanical strength, wear resistance and tear resistance and low-temperature elasticity. Aiming to improve the performance of sPS by exploiting these properties, Xu et al. [43]

? 60 •£ 50 W)

I

40

tj 30

a

& 20 10

2

4 6 8 10 Content of triblock copolymer (wt%)

12

Figure 20.10 Impact strength vs content of SEBS for sPS/EPR blends compatibilized with LMW-SEBS (circles) and HMW-SEBS (squares). Reprinted from Polymer, vol. 41, Hong, B. K., Jo, W. H., 'Effects of molecular weight of SEBS triblock copolymer on the morphology, impact strength, and rheological properties of sPS/EPRrubber blends' p. 2069, Copyright 2000, with permission from Elsevier Science

458

L. ABIS ETAL

made an investigation on sPS/polyurethane blends compatibilized with block copolymers. SEM shows for the binary sPS/polyurethane blend (70:30wt%) a morphology typical of immiscible phases, whereas after addition of poly(styrene-b-vinylpyridine) a progressive reduction in the dispersed particles and better adhesion at the interface are observed. DSC measurements gave in the case of the compatibilized blend lower Tg values of the polystyrene components and a lower degree of crystallization, which suggest penetration of polyurethane segments into the sPS phase. The mechanical and rheological results reflect the above picture. In fact, addition of the compatibilizer (up to 7 wt%) induces a progressive increase in the tensile strength and elongation at break, but a decrease in the Young's modulus and a lowering of the rheological parameters (complex viscosity and (7). It is thought that compatibilization is driven either by the solubility of the PS block of poly(styrene-67-vinylpyridine) into the amorphous phase of sPS or, as proved by FTIR, by the NH... N interaction between the vinylpyridine block and the polyurethane.

4.2.5.2

sPS/sulfonated aPS

Li et al. [44] found that the addition of the zinc salt of sulfonated atactic polystyrene to sPS in amounts lower than 10wt% decreases both the crystallization rate and the crystallite size and perfection of sPS. On the other hand, above this content, only minor influences on the properties of sPS are observed. The authors concluded that in the latter case interpolymer association of the ionic groups causes greater liquid-liquid separation, so supporting the hypothesis of substantial incompatibility between such polymers.

5

CONCLUSIONS

Because in commercial applications sPS blends surpass neat sPS in importance, in the future increasing attention to these systems is expected. The above discussion highlights that blending of sPS with other thermoplastic or elastomeric polymers is a convenient way to upgrade its properties. Fully miscible blends with atactic PS or PPE have been prepared and widely studied by several research groups, and now these systems are fairly well characterized. On the other hand, binary and multicomponent blends of sPS with immiscible polymers have been much less studied, although they are more interesting from a practical point of view. Indeed, most of the patent literature is focused on multicomponent blends generally containing an elastomer or another suitable tough polymer, and a block copolymers or graft copolymers as a compatibilizing agent. The latter can be either added as a preformed component or gener-

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE

459

ated in situ. In fact, the presence of a compatibilizer generally enhances the mechanical performances of the blends by favoring efficient dissipation of applied stresses towards the tougher component. In fact, at least in the reported cases, the behavior of aPS/HDPE/SEBS blends seems to follow this rule. However, crystallization occurring in the analogous sPS blends during the thermal treatments constitutes a barrier to the solubility of the compatibilizer in the sPS domain and leaves the mechanical properties unaltered. The search for new compatibilizers and /or suitable processing procedures could be a viable way to overcome these limits.

6

LIST OF ABBREVIATIONS

Compounds aPS EPDM EPR HMW-SEBS LDPE LLDPE LMW-SEBS MA-SEBS PIB PP PPE PS PVME SB SEBS sPS TMPC

atactic polystyrene ethylene-propylene-diene rubber ethylene-propylene rubber high molecular weight SEBS low-density polyethylene linear low-density polyethylene low molecular weight SEBS maleic anhydride-grafted SEBS polyisobutylene polypropylene poly(2,6-dimethylphenylene ether) polystyrene poly(vinyl methyl ether) styrene-butadiene diblock copolymer hydrogenated styrene-butadiene-styrene triblock copolymer syndiotactic polystyrene tetramethyl polycarbonate

Techniques WAXD FTIR NMR DSC SEM TEM OM DMTA

wide-angle X-ray diffraction Fourier transform infrared spectroscopy nuclear magnetic resonance spectroscopy differential scanning calorimetry scanning electron microscopy transmission electron microscopy optical microscopy dynamic mechanical thermal analysis

460

L. ABIS ETAL

Parameters

Tg (°C) Tm (°C) Tci (°C) vc (°C/min) Tmax (°C) 'max(min) G' (Pa) (G((j-m/min) X

glass transition temperature melting temperature isothermal crystallization temperature cooling rate in DSC maximum temperature of the melt maximum residence time in the melt dynamic storage modulus radial growth rate of crystallization degree of crystallinity

REFERENCES 1. Po R., Cardi N., Prog. Polym. Sci., 21, 47 (1996), and references cited therein. 2. Tomotsu N., Ishihara N., Newman T.H., Malanga M.T., J. Mol. Catal A: Chem., 128, 167 (1998). 3. Abbondanza L., Abis L., Cardi N., Garbassi F., Po R., presented at the 4th International Symposium on Ionic Polymerization, Crete, October 2001. 4. Yamato H., presented at Styrenics '93, Zurich, December 1993. 5. Ulcer Y., Cakmak M., Miao J., Hsiung M., J. Appl. Polym. Sci., 60, 669 (1996). 6. Handa Y.P., Zhang Z., Wong B., Macromolecules, 30, 8499 (1997). 7. Park J., Kwon M.H., Park O.O., J. Polym. Sci., Part B: Polym. Phvs., 38, 3001 (2000). 8. Lin R.H., Woo E.M., Polymer, 41, 121 (2000), and references cited therein. 9. Greis O., Xu Y., Asano T., Peterman J., Polymer, 30, 590 (1989). 10. De Rosa C, Guerra G., Petraccone V., Corradini P., Polym. J., 23, 1435 (1991). 11. De Rosa C., Rapacciuolo M., Guerra G., Petraccone V., Corradini P., Polymer. 33, 1423 (1992). 12. Chatani Y., Shimane Y., Inagaki T., Ijitsu T., Yukinari T., Shikuma H., Polymer, 34, 1620 (1993). 13. Kellar E.J.C., Galiotis C. and Andrews E.H., Macromolecules, 29, 3515 (1996). 14. Swogger K., presented at the Worldwide Metallocene Conference, MetCon 95, Houston, TX, 1995. 15. Giannotta G., Abbondanza L., Gennaro A., Po' R., Lucchelli E., Braglia R., Garbassi F., presented at the Polymer Processing Society Regional Meeting, Gothenburg, August 1997. 16. Guerra G., De Rosa C., Vitagliano V.M., Petraccone V., Corradini P., J. Polym. Sci., Part. B: Polym. Phys., 29, 265 (1991). 17. Choi S.H., Cho I., Kim K.U., Polym. J., 31, 828 (1999). 18. Li H., Li G., Yang W., Shen J., Polym. Prepr., Am. Chem. Soc. Div. Polym. Chem., 39, 689 (1998). 19. Hong B.K., Jo W.H., Lee S.C., Kim J., Polymer, 39, 1793 (1998). 20. Bonnet M., Buhk M., Trogner G., Rogaush K.D., Petermann J.. Ada Polvm., 49, 174(1998). 21. Hong B.K., Jo W.H., Kim J., Polymer, 39, 3753 (1998). 22. Yuan Z., Song R., Shen D., Polym. Int., 49, 1377 (2000).

POLYMERIC BLENDS BASED ON SYNDIOTACTIC POLYSTYRENE 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

461

Woo E.M., Wu F.S., J. Polym. ScL, Part B: Polym. Phys., 36, 2725 (1998). Woo E.M., Wu F.S., Macromol. Chem. Phys., 199, 2041 (1998). Cimmino S., Di Pace E., Martuscelli E., Silvestre C., Polymer, 34, 2799 (1993). Wu F.S., Woo E.M., Polym. Eng. ScL, 39, 825 (1999). Duff S., Tsuyama S., Iwamoto T., Fujibayashi F., Birkinshaw C., Polymer, 42, 991 (2001). Cimmino S., Di Pace E., Martuscelli E., Silvestre C., Rice D.M., Karasz F.E., Polymer, 34, 214 (1993). Mandal T.K., Woo E.M., Polymer, 40, 2813 (1999). Koh K.A., Kim J.H., Lee D.H., Lee M., Jeong H.M., Eur. Polvm. J., 34, 1229 (1998). Gausepohl H., Oepen F., Knoll K., Schneider M., Mc Kee G., Loth W., Des. Monom. Polym., 3, 299 (2000). Abis L., Abbondanza L., Braglia R., Giannotta G., Marra G., Facchinetti L., Po R., Polym. Prepr., Am. Chem. Soc., Div. Polym. Chem., 40, 391 (1999). Abis L., Abbondanza L., Braglia R., Castellani L., Giannotta G., Po R., Macromol Chem. Phys., 201, 1732 (2000). Briganti G., Giordano R., Melchionna S., Abis L., Marra G., Giannotta G., Gennaro A., Colloids Surf. A: Phys. Chem. Eng. Aspects, 176, 161 (2001). Sjoerdsma S.D., Dalmolen J., Bleijenberg A.C. A.M., Heikens D., Polymer, 21, 1469 (1980). Lindsey C.R., Paul D.R., Barlow J.W., J. Appl. Polym. ScL, 26, 1 (1981). Abis L., Abbondanza L., Braglia R., Castellani L., Giannotta G., Po R., presented at the 14th International Conference on Modified Polymers, Bratislava, October 2000. McKelvey J.M., Polymer Processing, J Wiley, New York, 1962, Chapt.13, p. 340. Schwarz M.C., Barlow J.W., Paul D.R., J. Appl. Polym. Sci, 35, 2053 (1988). Ramsteiner F., McKee G.E., Heckmann W., Oepen S., Geprags M., Polymer, 41, 6635 (2000). Yan R.J., Ajji A., Shinozaki D.M., Polym. Eng. Sci, 41, 618 (2001). Hong B.K., Jo W.H., Polymer, 41, 2069 (2000). Xu S., Chen B., Tang T., Huang B., Polymer, 40, 3399 (1999). Li H., Shen J., Yang W., Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., 40, 169 (1999).

This page intentionally left blank

Styrenic Block Copolymers

This page intentionally left blank

21

Styrenic Block Copolymer Elastomers R. C. BENING, W. H. KORCZ AND D. L HANDLIN, JR Kraton Polymers LLC, Westhollow Technology Center, Houston, TX USA

1

INTRODUCTION

The synthesis of styrenic block copolymers (SBCs) has been discussed in a number of books and review articles concerning block copolymers [1] and anionic polymerization [2]. A comprehensive review of the field is beyond the scope of this chapter, the objective of which is to provide an overview of the technology, with particular emphasis on processes currently used for commercial production.

Styrenic block copolymers derive their useful properties from their ability to form distinct styrene (hard phase) and diene (rubber phase) domains, with well defined morphologies. To achieve this requires an unusual degree of control over the polymerization. The polymerization must yield discrete blocks of a uniform and controlled size, and the interface between the blocks must be sharp. This is best achieved by so-called living polymerization. For a polymerization to be classified as truly living, it is generally accepted that it must meet several criteria [3]: 1. The extent to which termination of growing chains occurs is insignificant on the time-scale of the polymerization and termination reactions are suffiModern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy •:n 2003 John Wiley & Sons Ltd

466

R. C. BENING ETAL

ciently slow so as to allow a second charge of monomer to be added after the first is consumed. 2. All chains grow uniformly. This is commonly evidenced by a linear relationship between molecular weight and conversion. 3. The number-average degree of polymerization (DPn) is given by the ratio of the concentrations of the monomer and initiator, [M]/[I][4]. 4. Further, if the rate of initiation is fast relative to the rate of propagation, i.e. all of the chains start at the same time, the molecular weight distribution (MWD), defined as the weight-average degree of polymerization over the number-average degree of polymerization (Xw/Xn) will be given by the Poisson distribution:

At a reasonably high degree of polymerization, the value of the MWD approaches unity; the product is monodisperse, that is, all of the chains are essentially the same size. To be commercially attractive, it is also important for the polymerization rate to be slow enough to control in a batch process, yet fast enough to allow reasonable productivity, in a temperature range that is readily accessible in a batch process. In practical terms, this means it is preferable for the half-life for polymerization to be on the order of 1 min to tens of minutes in a temperature range from at or near room temperature to about 150°C. Since shortly after its discovery by Szwarc et al. [5] in the mid-1950s, living anionic polymerization has been recognized as an ideal route to styrenic block copolymers [6]. To date, living anionic polymerization remains the only commercially important technology for SBC synthesis. The anionic polymerization of styrene and common dienes such as butadiene and isoprene satisfies the criteria outlined above, particularly when carried out in a hydrocarbon solvent and initiated by an appropriate lithium alkyl. Although a variety of Group I metal alkyl and aryl species have been used as initiators for anionic polymerization, most commercial SBCs are produced using n- or s-butyllithium [7]. There are a number of reasons for this. Hydrocarbon solutions of both of these reagents are available commercially and in large quantities. The alkyls have reasonable solubility and thermal stability, and although pyrophoric, solutions may be handled easily with appropriate precautions. The higher thermal stability also translates to the growing chain. Although it is possible to achieve living polymerization with alkyl sodium and potassium initiators, the polymerizations must generally be carried out well below room temperature [8]. The ability to initiate rapid polymerization in hydrocarbon solvents such as cyclohexane has important ramifications for the synthesis of useful SBC products. Polar (Lewis basic) solvents increase the

STYRENIC BLOCK COPOLYMER ELASTOMERS

467

extent of 1,2 addition in the polymerization of butadiene and 3,4 addition in the polymerization of isoprene. For products in which the diene block is expected to function as a rubber, it is desirable to maximize 1,4 addition, as the glass transition temperature (Tg) is increased by 1,2 and 3,4 repeat units [9]. In pure hydrocarbon solvents, >90% of the diene repeat units are in the 1,4 configuration. Control of the diene microstructure through the use of polar additives will be discussed in a later section. While s-butyllithium initiates polymerization rapidly in hydrocarbon solvents, a small amount of a Lewis base such as tetrahydrofuran (THF) is generally added to insure rapid initiation by «-butyllithium [10]. If the amount of polar additive is kept to a minimum, it is still possible to achieve an acceptable 1,4 content. A substantial body of work also exists on the preparation of so-called diinitiators, species that result in the simultaneous growth of a polymer chain from both ends. There are two basic approaches to this. One is to use the metal in the presence of a conjugated aromatic species such as naphthalene to generate a radical anion capable of transferring an electron to the monomer. Under suitable conditions, the resulting monomer radical anions rapidly dimerize to form a dilithium species that goes on to add monomer in a living fashion [11]. Although this was one of the first examples of a living anionic polymerization, there are a number of drawbacks. A high level of polar solvent must be present, so that a diene block formed by this process will have a fairly high vinyl (low 1,4) content and the polymerization must generally be executed at a low temperature. It is also difficult to determine quantitatively the initiator concentration, so control of the molecular weight is difficult. A second approach involves addition of 2 mol of a lithium alkyl to a nonpolymerizable diolefin [12]. A considerable amount of work has been done to optimize the preparation of the dkv-butyllithium adduct of 1,3-diisopropenylbenzene (w-DIPB). Under the the right conditions, an effective initiator can be prepared with minimal m-DIPB dimerization [13]. The principle drawback of this system is that a polar additive is required, which has implications for the microstructure, as outlined above. The addition of 2mol of butyllithium to l,3-bis(l-phenylethenyl)benzene (DDPE) can be carried out in pure hydrocarbon solvents. However, efficient diinitiation requires the addition of a polar modifier or alkoxide salt [14]. Another drawback to this approach is the high cost of the diolefin. The kinetics and mechanistic details of the lithium alkyl-initiated anionic polymerization of styrene and diene monomers in hydrocarbon solvents have been the subject of numerous investigations [15]. Some of the first investigations revealed that the propagation reaction was first order in monomer, as might be expected, but followed a fractional order in the lithium alkyl [16]. Most investigators have observed a 0.5 order for the polymerization of styrene. Values have been quoted for the polymerization of butadiene and isoprene ranging from about 0.17 to 0.5, with 0.25 being the most commonly quoted value for both monomers. There is some evidence that the order in lithium for diene polymerization

468

R. C. BENING ETAL

approaches 0.5 at very low chain end concentrations [17]. The fractional orders observed in these polymerizations have generally been explained by a model in which chains are in equilibrium between 'free' and associated states, with the equilibrium lying towards the associated state. Propagation (addition of monomer) is assumed to occur mainly through the small population of 'free' chains. The order has been associated with the degree of aggregation, that is, 'free' poly(styryllithium) chains are said to be in equilibrium with dimers and 'free' poly(butadienyllithium) chains with tetramers. Although there is ample evidence that organolithium species aggregate in hydrocarbon solutions, there is substantial controversy surrounding the dominant aggregate structure and the relationship between the degree of aggregation and kinetic order [18]. As yet, no model has been formulated that predicts the relationship between propagation rate and chain end concentration over a wide range of chain end concentrations and temperatures by explicitly accounting for the equilibrium between 'free' and associated chains. However, progress has been made in the related area of the anionic polymerization of methacrylate monomers [19]. To achieve optimal properties in an AB block copolymer, it is important to control the molecular weight of the blocks, and minimize the amount of 'A' homopolymer produced on addition of the second monomer. Termination reactions do occur in these systems [20], but the rate is fairly slow, particularly at temperatures below about 100°C. In a practical sense, protic impurities present a much greater challenge. In a two-reactor system it is common practice to prepare the first block in one reactor, titrate out impurities in the 'B' monomer charge in a second reactor by adding small increments of butyllithium to a solution of the 'B' monomer until the first sign of color or exotherm, and then the transfer poly(A)Li solution to the second reactor. While the majority of SBC products possess discrete styrene and diene blocks, some discussion of the copolymerization of styrene and diene monomers is warranted. While the rate of homopolymerization of styrene in hydrocarbon solvents is known to be substantially faster that of butadiene, when a mixture of butadiene and styrene is polymerized the butadiene is consumed first [21]. Once the cross-propagation rates were determined (ksd and k& in Figure 21.1) the cause of this counterintuitive result became apparent [22]. The rate of addition of butadiene to a growing poly sty ryllithium chain (ksd) was found to be fairly fast, faster in fact than the rate of addition of another styrene monomer. On the other hand, the rate of addition of styrene to a growing polybutadienyllithium chain (kds) was found to be rather slow, comparable to the rate of butadiene homopolymerization. Thus, until the concentration of butadiene becomes low, whenever a chain adds styrene it is converted back to a butadienyllithium chain before it can add more styrene. Similar results were found for the copolymerization of styrene and isoprene. Monomer reactivity ratios have been measured under a variety of conditions [23]. Values for rs are typically <0.2, while values for dienes (r d ) typically range from 7 to 15. Since

STYRENIC BLOCK COPOLYMER ELASTOMERS

.Li

+

469

=1

k..

Figure 21.1 Cross-propagation rates for butadiene-styrene polymerizations

these polymerizations are living, the change in monomer composition with conversion is reflected in the composition along the chain. This 'tapering' is so strong that a fair approximation of a sequential SB diblock copolymer can be formed by initiating a mixture of the two monomers. For some applications it is desirable to produce a block containing a significant amount of 'randomly' placed styrene in the diene block [24]. As a result, a substantial amount of research has been done to find 'randomizers' capable of altering these relative reactivity ratios. Copolymerization in the presence of high levels of ethers such as diethyl ether can dramatically reduce tapering, but at the cost of substantially decreasing the extent of 1,4 addition [25]. The addition of salts of other alkali metals, usually potassium alkoxides, has been reported to reduce tapering without substantially increasing the extent of vinyl addition in styrene-butadiene copolymers. This has been attributed to rapid counterion exchange, since polymerizations initiated by arylpotassium species produce less tapered copolymers [26]. Although styrene-diene diblock copolymers are used in some applications, particularly in the area of viscosity index improvement (VII) additives for motor oil, styrenic block copolymers are most often used as thermoplastic elastomers. In these applications the styrene blocks phase separate, 'crosslinking' the rubber blocks in a thermally reversible fashion. The simplest structure capable of exhibiting this behavior is a linear styrene-diene-styrene triblock. The most obvious way to produce such a molecule is by sequential polymeriza-

470

R. C. BENING ETAL

tion. That is, an aliquot of styrene is polymerized to produce the desired block. The diene monomer is added and then, when the polymerization of the diene is complete, a second aliquot of styrene is added. Although conceptually simple, there can be complications. As noted in the previous discussion on copolymerization, the rate of addition of styrene to polybutadienyllithium in hydrocarbon solvents is slow, at best comparable to the rate of styrene homopolymerization. As a result, the first chains that add styrene will tend to grow much larger than chains that add styrene later in the polymerization. In an extreme case, for example, when the molecular weight of the final styrene block is low, some chains may fail to add styrene before it is all consumed. This is generally referred to as a 'crossover' problem. Lewis bases such as TMEDA and glyme ethers are often added prior to the final styrene charge to increase the rate of addition of styrene to the butadienyllithium chain end [27]. The problem is less severe for isoprene; sequential SIS copolymers can be made with relative ease. Under ideal conditions, it is possible to make triblocks with little, if any, detectable diblock [28]. Linear triblock copolymers can also be prepared by diinitiation of the diene block, followed by addition of styrene. This process is less commonly employed, in part owing to the issues involved in preparing a suitable diinitiator. A more subtle complication results from the tendency of the chain ends to aggregate. Living polymer solution viscosities are much higher at a given polymer concentration in a diinitiated system. In practice, diinitiation is mainly used when the anion of the desired endblock monomer is unable to initiate diene polymerization [29], or conditions that are required for endblock polymerization are incompatible with midblock polymerization (result in too high vinyl, etc.) [30]. An alternative approach involves adding a reagent capable of efficiently linking the diblock anions. A great variety of these coupling agents are known in the literature. The use of coupling agents allows the synthesis of more complex architectures. So-called radial polymers are formed by adding coupling agents that link three or more chains. Commonly used linear coupling agents include dibromoethane [31], dichlorodimethylsilane, and esters such as methyl benzoate [32]. Diblock polymer anions can be efficiently coupled with high-purity (low oligomer) bisphenol A-epichlorohydrin resins. At least 80% of the diblock molecules couple to form triblocks. These products have excellent thermal stability (comparable to sequential products), and good color stability owing to the absence of salts such as LiBr that can promote discoloration at processing temperatures [33]. Although coupling reactions generally leave more diblock than sequential polymerization, which is bad for strength properties, there are some advantages. Crossover problems are eliminated. For optimal properties it is desirable to have symmetric styrene endblocks. This is an inherent feature of coupling. Common radial coupling agents include diesters (three and four arms) and di- and trichloro- or -methoxysilanes (three and four arms, respectively) [34]. One can take advantage of the selectivity of addition to silanes to create asymmetric radial polymers. For tetrafunctional silanes such as

STYRENIC BLOCK COPOLYMER ELASTOMERS

471

SiCU, the third and fourth substitution reactions are much slower than the first two, allowing the synthesis of A2SiB2 structures [35]. A2B2 polymers have also been prepared by adding l,3-bis(l-phenylethenyl)benzene at a 0.5:1 molar ratio on chain ends, and using the resulting dilithium species to initiate the growth of two more arms [36]. So-called star polymers constitute another important class of branched block polymers [37]. Star polymers are prepared by adding a diene monomer that is capable of reacting with the living chain ends, and then reacting with its self to form a crosslinked 'core'. The most commonly used core-forming compound is m-divinylbenzene (DVB). A substantial amount of research has been conducted on the relationship between variables such as DVB:PLi ratio, solvent, reaction temperature and time, arm structure (terminal monomer and molecular weight) and the number of arms in the resulting star polymer and the linking efficiency (percentage of arms incorporated into the star) [38]. In general, the number of arms and the linking efficiency increase with increase in the DVB:PLi ratio. These cores retain lithium alkyl sites capable of initiating anionic polymerization; heteroarm stars have been prepared by growing new blocks out from the core [39]. Because polymer viscosity is primarily a function of the end-to-end distance, the viscosity of a star polymer changes only slightly as the number of arms is increased [40]. Thus, these products remain processable at very high molecular weights. Star polymers are particularly advantageous in applications where resistance to shear is important, since substantial degradation must occur before gross changes in physical properties occur. The theromooxidative and UV light stability of styrenic block copolymer elastomers can be greatly improved by selective hydrogenation of the diene block. Often referred to as 'SEBS' polymers, when derived from butadiene, and 'SEPS' polymers, when derived from isoprene, these materials represent an increasingly important class of thermoplastic elastomers. A controlled level of 1,2 (vinyl) addition is required to produce elastomeric products based on butadiene. If too few vinyl repeat units are present the midblock will be semicrystalline, with properties reminiscent of linear low-density polyethylene. If the vinyl content is too high, the Tg will be too high, and the material will be too stiff near room temperature. Vinyl contents in the range 40-50 % give a good balance of properties. A variety of Lewis bases have been used to control microstructure in anionic polymerization, the main requirement being that the Lewis base is sufficiently stable in the presence of the propagating anion to allow living polymerization. The most commonly used modifiers are ethers and tertiary amines. Since amines are poisons for many hydrogenation catalysts, ethers are used more frequently in the production of hydrogenated polymers. A further distinction can be made between monobasic species such as dialkyl ethers and bidentate species that have the potential to coordinate with lithium, such as glyme ethers and TMEDA (N, N, N', jV'-tetramethylethylenediamine). The former must

472

R. C. BENING ETAL

usually be present at high levels (1-10% of the solvent) to have a significant effect on the microstructure, while the latter can have a profound influence on the microstructure at sub-stoichiometric levels (less than 1 mol per mole of Li). Another distinction that has practical consequences involves the effect of reaction temperature on microstructure. In general, increasing the reaction temperature decreases the extent of 1,2 addition, but this effect is much more pronounced with 'chealating modifiers' such as TMEDA. Tetrahydrofuran (THF) represents something of an intermediate case. Although it cannot chealate Li, it can substantially increase the vinyl content when present at levels as low as 5 mol per mole of Li; the temperature response is also more reminiscent of 'chelating modifiers' [41]. There have been attempts to correlate the effect of a given level of modifier on microstructure with various measures of the change in solvent polarity. These approaches have met with some success for monobasic modifiers (diethyl ether, THF), but break down for dibasic modifiers [42]. Few detailed kinetic and mechanistic studies have been published for Lewis base-modified polymerizations. A common trend is for the propagation rate to increase over a certain range of modifier concentrations, and then decrease, while the order in Li increases [43]. The relationship between lithium solvation (in this case by TMEDA), aggregation, and reactivity was examined in a review article [44]. The view that the effect of TMEDA on reactivity is due to solvation (chealation) of lithium, resulting in 'djsaggregation' and thus increasing the population of reactive species was found to be inconsistent with a number of observations. The mechanistic details of the action of microstructure modifiers, particularly those that are often assumed to chealate lithium, appear to be complex, and represent an opportunity for further research. A variety of catalytic and noncatalytic hydrogenation reactions have been used to hydrogenate the diene block selectively [45]. Catalytic hydrogenation in the presence of so called 'Ziegler-type' catalysts or titanium metallocene complexes are the dominant routes for commercial production [46]. Owing to their high cost, noble metal catalysts are used only when functional groups are present that can poison the aforementioned catalysts, as is the case with nitrile rubber. Ziegler-type catalysts are prepared by reaction of an organic salt of a transition metal, usually nickel or cobalt, with a trialkylaluminum compound such as triethylaluminum. The resulting black solution is probably best described as a colloid. Owing to the nature of the catalyst, mechanistic studies are difficult. The prevailing model assumes that hydrogenation occurs by the action of metal clusters embedded in an alumina-like matrix, but there is little evidence for or against this structure [47]. Although hydrogenation of the styrene block generally occurs only at high temperatures, olefin isomerization can be an issue, and there have been studies aimed at minimizing this side reaction. The moresubstituted olefins that result from isomerization are harder to hydrogenate, making it more difficult to achieve an acceptable level of residual unsaturation. The metal levels present as residuals from the catalyst are high enough to cause

STYRENIC BLOCK COPOLYMER ELASTOMERS

473

color and degradation problems if left in the finished product. A number of processes for catalyst removal have been described [48]. Most often this is done by oxidation of the metal in the presence of aqueous acid; the metals are extracted into the aqueous phase. More recently, titanium metallocene catalysts have become important in SBC production [49]. Among the advantages that have been claimed are a lower propensity for isomerization and higher activity. Most systems consist of a titanium metallocene complex, usually Cp2TiCl2, and an activator, usually a lithium or aluminum alkyl. An alkyl metallocene species is probably the active component in these systems; some dialkyl- and diaryltitanium metallocene compounds catalyze hydrogenation without activators [50]. These catalysts are generally less effective for the hydrogenation of more hindered olefins. Even with higher catalyst loadings and longer reaction times, it is difficult to achieve substantially greater than 90 % saturation of isoprene repeat units in SIS polymers [51]. In principle, the ideal process would be hydrogenation over a supported catalyst in a 'fixed-bed' reactor, that is, by essentially passing a solution of the polymer and hydrogen through a column containing the supported catalyst. Provided that leaching is minimal, catalyst residuals are no longer an issue. Owing to the high viscosity of polymer solutions and steric issues, this approach has met with limited success. Unlike small molecules, polymer repeat units are unable to diffuse freely to active sites on the catalyst. It is not uncommon to suffer from both residual olefin unsaturation and excessive hydrogenation of the styrene repeat units. Reasonable success has been achieved with large pore-size supports designed to facilitate access of the polymer repeat units to the active sites [52]. Fully saturated SBC polymers have also been investigated. Vinylcyclohexane-ethylene/propylene-vinylcyclohexane triblock copolymers have been prepared by complete hydrogenation of SIS polymers using a supported palladium catalyst [53]. Under the appropriate conditions, hydrogenation of the styrene blocks can also be accomplished using Ziegler-type catalysts [54]. Although anionic polymerization remains by far the dominant approach to preparing styrenic block copolymers, living cationic polymerization has been used to prepare radial and triblock copolymers of styrene and isobutylene (IB) with properties comparable to styrene-hydrogenated diene copolymers [55]. It is also possible to polymerize monomers that yield higher Tg endblocks, such as amethylstyrene, to produce higher service temperature elastomers [56]. The main impediment to commercialization is the low temperatures required to avoid excessive termination. Even at —80 °C, IB must be added before polymerization of the 'A' block monomer is complete. Crossover from isobutylene chain ends to sytrenic monomers is slow; coupling agents have been developed to help circumvent this problem [57]. Recently, there has been a great deal of interest in the synthesis of block copolymers thermoplastic elastomers via so-called

474

R. C. BENING ETAL

'controlled' free radical polymerization. Nitroxide-mediated radical polymerization has been used to produce block copolymers of styrenic monomers [58], and progress has been made towards controlled polymerization of dienes [59]. To date, it has not been demonstrated that polymers with physical properties comparable to commercial SBC elastomers can be produced via this route-"* A great deal of research has also been carried out aimed at making acrylic or methacrylic block copolymer analogues of styrenic block copolymers. In principle, blocks with a wide range of glass transition temperatures could be prepared from these monomers. Initial efforts focused on the polymerization of tert-butyl methacrylate (tBMA). Fairly good diene-tBMA block copolymers can be prepared in hydrocarbon solvents at reasonable temperatures, but side reactions still limit the utility of this approach [60]. Researchers at the University of Liege have succeeded in preparing MMA-nBuA-MMA triblock copolymers that are structurally comparable to commercial SIS polymers [61], but the tensile strength was disappointing. On further analysis, this was found to result from the high entanglement molecular weight of the acrylic midblock [62]. The tensile strength of all-acrylic TPEs may never be suitable for some applications. Further work from this group aimed at producing analogous polymers by controlled radical polymerization (in this case, atom transfer radical polymerization) illustrates some of the difficulties encountered in controlling the polymerization sufficiently, at the molecular weights required, to produce an effective TPE [63].

3

PROPERTIES OF STYRENIC BLOCK COPOLYMER ELASTOMERS

Albert Einstein said that it is good to make things as simple as possible, but not simpler. Beneath each simple statement about the properties of styrene block copolymers lie volumes of books, thousands of patents and countless pages of paper and electronic files dedicated to describing and understanding these highly useful polymers, and their applications. The task becomes reducing all of this to simple ideas, simple pictures and simple words, 'but not simpler'. One simple idea is that styrenic block copolymers are almost never used as a stand-alone 100% neat polymer for any application or use. We tend to think about polymers in terms of 'this plastic soda bottle is polyester, or this carpet fiber is polyamide, or this house siding is PVC, or this garbage bag film is polyethylene', fully understanding and meaning that virtually 100% of the named object is that polymer. Our brains usefully process the named polymer properties' set (as neat polymer) into the desired and required property set for its application. Life is simple in the 100% world. It is intuitive, and what we seem to know makes sense, looking either way properties wise, to why this polymer is used for this application.

STYRENIC BLOCK COPOLYMER ELASTOMERS

475

When we name thermoplastic elastomer (TPE) parts, and associate 'TPE' with them, we enter the confusing world of not really being able to 'name the polymer' from an observed property set of the finished goods. The same problem exists in looking in the opposite direction: it is not intuitive from the properties' set of a neat styrenic block copolymer what its use might be. In fact, one polymer may have diverse uses. It all depends on 'what else' is in there with it, because these polymers are not a part of the '100% world'. The styrenic block copolymer is, after all, the minor component most of the time. If we were to research and list every commercial styrenic block polymer made today that has 'elastomeric' properties, and run a rough average for molecular weight, styrene content, and the size of styrene end block, we would likely end up with a polymer that contained about 25 % styrene, had an endblock size of about 15 000, and had an overall molecular weight of 130 000. It is no accident that the styrene content is low (the rubber block comprises 75%), the overall molecular weight is rather puny compared with most other polymers, and the styrene endblock length is greater than the entanglement molecular weight for polystyrene. So why does the entire universe of possible structures for styrenic block copolymers end up with this sort of 'average' at its center? The simple answer has only a few important elements: The polymers should have useful rubbery properties as neat polymers, and they should be capable of further possessing. In the real world of their application in thousands of uses, it is the satisfaction of these two elements that has decided (over the last 35 years) what structures would be made, and yield the simple property set of 'rubbery' (also known as soft and strong, low flexural modulus, low hardness) and 'processable' (capable of being formulated with other ingredients to yield useful melt formable, most of the time, compositions). We can examine these elements in more detail by placing our 'average' polymer in the center of a grid, and then exploring the practical properties' boundaries of the three key polymeric handles of styrene content, overall molecular weight and styrene endblock size. At a high level, the entire realm of rubbery styrenic block copolymers can be mapped as shown in Figure 21.2. Along the horizontal axis is a relative molecular weight scale, which represents the overall size of the styrenic block copolymer (rubber block, and styrene blocks taken together). Along the vertical axis we scale styrene content (which also fixes styrene endblock size once we have chosen a molecular weight on the horizontal axis and assume the two styrene endblocks are of equal length in the styrene-rubber-styrene block copolymer). In the center of this universe, we have plotted one data point: our 'average' styrenic block copolymer, labeled as 'A'. There, one simple step has been taken. Now, we can begin to take a look at some 'what ifs'. What if the styrene content is much higher than data point 'A'? What if the endblock size is much smaller than 'A'?; What if the molecular weight is much less than or much greater than 'A'? Are there boundaries, and what is their nature?

476

R. C. BENING ETAL Styrene Content/ End Block Molecular Weight

Total Molecular Weight Figure 21.2 Universe of styrenic block copolymers

If we look at Table 21.1 [64], we can see that there is a lower limit to how small a styrene endblock can be in a styrene—isoprene—styrene (SIS) styrenic block copolymer before the 'strength' is severely reduced, or is reduced to that of chewing gum. There is a lower minimum molecular weight for the styrene block that will not yield a useful polymer or property set, let alone an elastomer. Thus, we have a lower boundary condition on how small the endblock is allowed to be. Similar results are found for styrene—butadiene—styrene (SBS) block copolymers [65]. If we look at Figure 21.3, we can see that there is an upper limit to the overall styrene content in the polymer if making a polymer to have rubbery properties is the desired outcome [66]. As the styrene content increases, the stress-strain response changes dramatically for these neat SBS polymers. At 53 and 65% styrene content, the polystyrene endblocks form the continuous phase in the phase-separated block copolymer, and these polymers behave more like polystyrene than a rubber at low strain. This low strain behavior is also shown at 39% styrene content, but a rubbery plateau begins to show at lower stress. Table 21.1 Effect of polystyrene molecular weight on the tensile strength of S—I—S. Reproduced with permission from Legge, Holden and Schroeder, Thermoplastic Elastomers; A Comprehensive Review, Hanser Verlag, Munich, 1987 Styrene (wt%)

S—I—Spolymer mol. wt. (x 10 -3 )

Tensile strength (MPa) 300% strain At break

20 20 19 11

13.7-100.4–13.7 8.4-63.4-8.4 7.0-60-7.0 5.0–80–5.0

1.8 1.1 1.3 ~0

27.0 16.0 2.2 ~0

477

STYRENIC BLOCK COPOLYMER ELASTOMERS Endblock Content 30-i

39% S

S–B–S Polymers

28% S

2065% S 10-

13% S 200

400

600

800

1000

Elongation (%) Figure 21.3 Stress-strain curves (at 2 in/min.) for S-B-S ene contents

polymers of various styr-

More 'typical' (elastomeric SBS polymer) stress-strain behavior is exhibited by the 28 % styrene content polymer. The low modulus plateau region extends to an elongation of 400 %, and the polymer is strong. This curve resembles that of a vulcanized rubber. Also, as a reminder, if the styrene content is very low (implying that an endblock size may be too small to entangle), the resulting polymer has the properties of chewing gum. In Figure 21.3, the 13% styrene content example is representative of 'too little'. Thus, we have an upper boundary on styrene content if our intention is to make rubbery polymers. If we look at Figure 21.4, we can begin to think about the processability of styrenic block copolymers, and how molecular weight and the rubber block type affects it [67]. The thermodynamic driving force for these polymers to be phase separated (styrene with styrene, and rubber with rubber as two incompatible phases), even in the 'melt' state, is strong. The rheology data plotted in Figure 21.4 testify to that fact. For many styrenic block copolymers the order-disorder temperature, the temperature above which they would behave as a single phase melt (and manifest viscosities typical for their rather low molecular weights), is well above temperatures at which significant oxidative degradation of the polymer will occur. Many styrenic block copolymers, therefore, are 'intractable' in the 'melt' state (at practical processing temperatures governed by the stability of the polymer, but well below the order—disorder temperature).

R. C. BENING ETAL

478

S-EB-S I O 200 °C CJ 220 A 240 °C O 260 °C O 280 300 108

108

108

SHEAR STRESS (Pa) Figure 21.4 Corrected melt viscosity as a function of shear stress and temperature for the three block copolymers studied. Reproduced with permission from Legge, Holden and Schroeder, Thermoplastic Elastomers; A Comprehensive Review, Manser Verlag, Munich, 1987

If we look at the three panels of data in Figure 21.4, we can begin to appreciate the 'intractable' nature of some of the styrene—hydrogenated rubber—styrene block copolymers. The right panel shows an SBS polymer (not hydrogenated) that is 'typical' in its overall size and styrene content for a commercial polymer. Viscosity is plotted versus shear stress at two testing temperatures (200 and 220 °C) for this SBS polymer, and it serves as our reference point for the center panel. The center panel polymer is the hydrogenated version of this SBS polymer (only the rubber has been hydrogenated). The resultant change in viscosity is remarkable (several orders of magnitude at any given shear stress and temperature). In this case the driving force for much stronger phase separation in the melt (greater incompatibility, greater difference in solubility parameters between the hydrogenated rubber and styrene) has increased greatly. Only at very high temperatures (above 250 °C) and at very high shear stresses does this polymer begin to 'flow'. From experiments done in our laboratories, the extrusion of this polymer as a neat material is 'on the edge' as a material that can be pelletized in a twin-screw extruder. The left panel is an example of a higher molecular weight version of a hydrogenated block copolymer with the same chemical structure as in the

STYRENIC BLOCK COPOLYMER ELASTOMERS

479

center panel. At practical shear stresses (and shear rates) in the range of commercial processes for injection molding, this polymer is truly intractable, and cannot be processed in the neat form. Thus, for many styrenic block copolymers, there is an upper limit with respect to molecular weight and mid (rubber) block chemical type that will preclude them from being processable as neat polymers. If we now take the essence of Table 1 and Figures 21.3 and 21.4 and plot these as boundaries, a picture emerges as shown in Figure 21.5. If we move from data point 'A', our grand 'average' commercial styrenic block copolymer having elastic properties, along the axes of the grid, we run into the boundaries discussed. Moving straight 'south' from point 'A', we eventually drop the styrene content (and endblock molecular weight) sufficiently that we do not create enough styrene length or styrene phase to have a strength-bearing styrene domain network, and the polymer behaves as a weak, low molecular weight rubber (chewing gum). Moving straight 'north' from point 'A', we eventually increase the styrene content sufficiently that the polymer begins to behave with polystyrene continuous phase tendencies. The low strain properties (tensile, flexural modulus) are more plastic-like than rubber-like. Moving to the 'east' from point 'A', we begin to worry about how melt processable the polymer will be as we scale it up (keeping the styrene content fixed, the endblocks become larger and the molecule becomes larger) to higher molecular weights. While this scaling up in molecular size (proportionately) does not affect the tensile properties in any large way (Figure 21.6) [68], the viscosity of the neat polymer is affected and becomes higher, especially at lower shear rates (Figure 21.7) [69]. Moving straight 'west' from point 'A', we scale the molecule to a smaller size at fixed styrene content. Much like moving 'south' from point 'A', we eventually enter a region where the end-block becomes too small to yield useful properties, and the rubber block length (also shortened proportionally) has fewer and fewer entanglements, neither of which proves beneficial in providing a strong, rubbery polymer. The ideas expressed in Figure 21.5 were advanced over 30 years ago [70], and this simple picture still provides a powerful template for mapping the rubbery and utile styrenic block copolymers today. Having roughly defined the practical boundaries of commercial styrenic block copolymers with useful elastomeric properties, why do we have so many choices available within those boundaries? What key property change will drive a polymer manufacturer and/or polymer user to head 'southwest' in Figure 21.5? What would drive a customer to ask for something 'more northeasterly, please'? Or, go as far 'east' as possible and 'do not worry about melt processability for my application'. In essence, why is not everyone totally content to make, and use, polymer 'A' for every application?

480

R. C. BENING ETAL Styrene Content/ End Block Molecular Weight

Plastic Behavior

Weak Plastic

Intractable Flow

Weak Rubber Total Molecular Weight

Figure 21.5 Directional and bounded universe for utile styrenic block copolymer elastomers

280 240

POLYMER

MOL WT x 1Q-3

SBS-6 SBS-7 SBS- 1 SBS- 3 SBS- 5 SBS-8 SBS- 8

8.4 - 63.4 - 8.4 13.7– 10.4–13.7 13.7-63.4-13.7 21.2–97.9 21.2 21.1 –63.4–21.1 13.7 41.2 13.7

% P.S.

200 on on tu

te

160

UJ

£! 120 tu 80 40

10

12

ENTENSION RATIO X

Figure 21.6 Effect of composition and block size on tensile properties of styrene butadiene—styrene triblock copolymers (1 kg/cm2 = 0.1 MPa)

There are two ways to have a discussion about the 'properties' of styrenic block copolymers, and they are fundamentally very different. The first way is to make and fully describe every conceivable polymer, and highlight the distinctions found when making this or that change to the neat polymer. Let us call that

STYRENIC BLOCK COPOLYMER ELASTOMERS

Copolymer 20S–110B–20S 14S–74B–14S 10S–52B–IOS

481

Method of Measurement Modified Mooney Instron Double Cone and Capillary Place Viscometer Viscometer

104 10-2

1

10

Shear Rate, seer1 Figure 21.7

Viscosities of S—B—S polymers at 175 °C

the 'here's what influences' approach to discussing properties. The second way to have this discussion is with an eye to how these polymers are actually used, and from that perspective to talk about why certain choices are made in the selection of or change in an architectural feature (with some distinction in mind). In this case the distinction is driven by specific and utile need, or balancing of property set to achieve best performance and highest value in the final application. We shall call this the 'utile' approach to discussing properties. But why have the 'utile' properties' discussion at all? Unlike other polymers, styrenic block copolymers are almost never used in neat form for any application! In fact, in almost all applications, the block copolymer component of a recipe or formulation is often a minor component (less than 10% in applications such as asphalt modification, oil gels, viscosity improvers for motor oil), and usually comprises less than 40% in most others. Styrenic block copolymers are always developed and designed with this view in mind, and it is not always obvious in just looking at a neat polymer structure and its properties why a particular structure has been commercialized for a given application. The neat polymer properties (measured) are those that have been designed in knowing that the polymer will be a minor component in a specific use, and that its compatibility with other polymers, resins, additives, and filler (comprising the major components in a final recipe or formulation) must necessarily be strongly considered.

482

R. C. BENING ETAL

Let us look at 'utile' properties for polymer 'A', and how just a change in the rubber midblock determines properties. If polymer 'A' has an isoprene rubber block, it will find use in hot melt pressure-sensitive adhesives. Three key properties for this isoprene rubber block molecule are (1) it can be tackified readily with lower molecular resins and oils (chemical compatibility of the rubber block with these ingredients is important), (2) it is easily melt processable (the viscosity of the formulated adhesive is low and can be coated as a thin layer on to a substrate as hot melt fluid at rather low temperatures), and (3) it is capable of 'closing a box'. This is an important test where the ability of the adhesive to hold a box closed in a humid warehouse is most important (you might check in your garage, attic or basement to see if you agree). This test is one of the most important properties that must be considered, and the structure of this SIS polymer 'A' must be designed with the capability to close a box. It must have good properties in shear in the final adhesive recipe: if not 'A', then where in the region to strike the best balance among the three key properties demanded in this SIS polymer? It turns out that the polymer should be lower in styrene content (higher in isoprene rubber content) with styrene blocks large enough to handle shear stresses in keeping a box closed. If the same SIS polymer is designed for precisely the same use in a pressuresensitive adhesive, but the objective is to start with the best molecule for a solvent-based adhesive, the molecular design result is not the same. In this case, because hot melt viscosity is no longer a constraint, and because the SIS molecule can be optimized for high solids in a solvent base, the molecule is completely different (it ends up being higher in styrene content, larger in overall size, and with larger endblocks). In just looking at these two polymers side by side for neat polymer properties, we might be puzzled about their co-existence for the same application. In this case, the answer lies in a hot melt molecule design versus a solvent molecule design. If polymer 'A' has a butadiene rubber block, it will not find application in pressure-sensitive adhesives. It is not readily tackified, it is not readily melt processable, and it will not close a box. However, if dissolved in solvent, compounded with filler and certain resins, it will make the world's best construction mastic, very capable of bonding drywall to wood, etc. If one wants to maximize the 'solids' content in this mastic (use less solvent), and if one wants to design the SBS molecule to be soluble in more environmentally friendly solvents, then in what direction should one head? If not 'A', then where in the region does one strike the balance for a highly extended but tough mastic that allows the solvents of choice while achieving maximum solids and a viscosity low enough to squeeze from a tube? If polymer 'A' has a hydrogenated rubber block, it will be nearly impossible to tackify, will have very high hot melt viscosity, and might be sticky at elevated

STYRENIC BLOCK COPOLYMER ELASTOMERS

483

temperatures (but not pressure sensitive). However, if we took this polymer and added 400-500 wt% oil to it, we could create a very nice oil-gel. To improve the molecule for this application, we might move 'southeast' on the grid (styrene blocks become slightly larger or maintain their size when we put more rubber block into the block copolymer). Or, we might want to make a high-solids solvent-based sealant molecule that is weatherable, and fails cohesively (and repairs itself) in joint movement. In this case we would move 'southwest' on the grid, which would make the overall molecule smaller and make the end block smaller, both of which would help with respect to creating something that is 'weaker' and likely to be of lower viscosity in a solvent-based sealant. We might even weaken it a great deal more by making sure that not all the polymer molecules had a styrene endblock on both ends of the polymer. In these three treatments we have considered the very simple case of 'the molecules are only different in their midblock chemistry', and learned that molecular design is driven by how these molecules, as a minor component in an intended use, become reality with utility in the final application strongly in mind. The important neat properties that we design in, therefore, become those in the property set relevant to the intended usage. And also as important, the exact chemical nature of the rubber block itself becomes a key 'property' in the world of usage and application of these styrenic block copolymers. We have looked at the influence that the rubber block has on properties in a utile sense. It provides the soft connecting tissue between the hard phase (the polystyrene endblock network), and it yields three very different chemical choices, which give rise to distinct property sets useful in distinct applications. We shall now look at the polystyrene polymer block, which is common to all styrenic block copolymers. In viewing styrenic block copolymers as a group on a macro level, there are many property similarities, governed by the phase separation of the polystyrene endblocks into distinct segregated regions. One consequence of this phase-separated polystyrene endblock network is a glass transition temperature associated with it. The measured endblock glass transition temperature is dependent on its size and the method used to measure it (about 85-105 °C when measured by DSC for commercial polymers, and about 10 °C higher across the range if measured by DMA). All styrenic block copolymers, irrespective of the rubber block, have this same common thread. As a consequence, important neat polymer functional properties such as tear strength and tensile strength fall rather precipitously as temperature is increased. In Figure 21.8, a 'softening temperature' for the polystyrene endblock is shown as a function of endblock molecular weight (ranging from about 6000 to 30000). The 'softening temperature' is characterized as the onset of test specimen creep (in a small-strain dynamic mechanical test in tensile mode), the creep point occurring when strain extension becomes considerable in order to maintain the appropriate stress level to continue the test. This 'softening temperature' lies below the measured 7g, and it is an indication of the

484

R. C. BENING ETAL Softening Temperature °C 120

100

80

60

Polystyrene End-Block Size Increasing

Figure 21.8 Endblock size effect on softening point of various SEBS block copolymers by L. K. Djiauw, Kraton Polymers LLC (internal communication), Westhollow Technology Center, Houston, TX (1999)

functional property losses that occur for these polymers below their rgs. The polystyrene endblock is also readily dissolved in good solvents for polystyrene. As such, if neat styrenic block copolymers are in contact with aromatic fluids (certain oils, most phthalate PVC plasticizers, higher solubility parameter solvents such as toluene and xylene), the polystyrene phase is solvated and the network, which provides the mechanism for neat polymer strength and toughness, is severely compromised. These two consequences of having this basic polystyrene phase separated network comprise the 'Achilles' heel' set for styrenic block copolymers. Retention of functional properties suffers with increasing temperature, and compatibility with fluids which dissolve the polystyrene network serve to rob the neat polymer of robust physical properties. Every styrenic block copolymer shares in the strengths and weaknesses that the polystyrene brings to the molecule. These deficiencies can be overcome (to a large extent in many applications) by compounding these neat polymers with semicrystalline polymers (polyethylene or polypropylene, for example, which comprise a co-continuous phase with higher heat and solvent resistance) or by using poly(phenylene ether) to raise the glass transition temperature of the polystyrene endblock phase. The final snapshots present a view of the parts taken together: Rubbery midblocks that are running into and out of strongly phase separated and well ordered endblock regions. In Figure 21.9, we find two glass transition temperatures that are characteristic of styrenic block copolymers, one associated with the rubbery midblocks and the other with the styrene endblocks (in this case for styrenic block copolymers with each having about 30 % styrene content and comparable endblocks).

485

STYRENIC BLOCK COPOLYMER ELASTOMERS

The three rubber blocks are chemically different (isoprene, butadiene, and hydrogenated butadiene). Between each of the two Tgs, the storage modulus for these polymers is rather flat and unchanging over a very broad temperature range. Were it not for the phase-separated polystyrene network (which is supporting the rubber chains), the rubber modulus would rapidly decay in the region between the two J"gs (and be more typical of an unvulcanized and rather low molecular weight rubber polymer). Above the Tg of the polystyrene endblock, the magnitude of the storage modulus remains ordered (highest to lowest EB > B > I) with the hydrogenated block rubber polymer displaying the most persistent and highly associated polystyrene network. In the companion to Figure 21.9, Figure 21.10, much above the polystyrene end-block Tg, viscous flow is ordered I > B > EB. Looking at the practical implications of this strongly phase-separated system, the view can be divided into two regions: a temperature region within which the polymer will be used, and a temperature region within which the polymer will be processed. Figure 21.11 illustrates the physical state of the phase-separated styrenic block copolymers in these two regions of temperature. In the 'end use region' the styrenic block copolymer component in a compounded or extended system (as a minor component, which it generally is) functions to deliver integral strength, toughness, rubbery and elastic

10 rads-1

l.E + 09

G 1650: SEBS D 1101: SBS D 1125: SIS

o

l.E + 05 -100

-60

20 60 Temperature (°C)

100

140

180

Figure 21.9 Rubber midblock effect on dynamic properties of comparable block copolymers by I. Kadri, Shell Chemical Company, Kraton Polymers (internal communication), SRTCL Laboratory, Louvain, Belgium (1999)

486

R. C. BENING ETAL

10

-60

-20

20 60 Temperature (°C)

100

140

180

Figure 21.10 Rubber midblock effect on dynamic properties of comparable block copolymers by I. Kadri, Shell Chemical Company, Kraton Polymers (internal communication), SRTCL Laboratory, Louvain, Belgium (1999)

Figure 21.11 Phase separations and thermal transitions for styrenic block copolymers by I. Kadri, Shell Chemical Company, Kraton Polymers (internal communication), SRTCL Laboratory, Louvain, Belgium (1999)

behavior and properties. In the 'processing region', these polymers behave as very viscous and elastic solids. The picture here is one of styrene endblocks being pulled from their persistent phase-separated network in the melt, and regrouping almost as quickly to resist flow when reforming (especially under conditions of low shearing). In most commercial and ordinary styrenic block copolymers, 'one-phase' melt (above the order—disorder temperature to achieve this flow behavior) cannot be attained, as this temperature is well above the thermooxidative stability limit for these molecules. As a postscript, I would like to acknowledge and thank Geoff Holden. Much of what I learned about block copolymers (and have shamelessly paraphrased in this text) came from his teaching while we were both employed in the

STYRENIC BLOCK COPOLYMER ELASTOMERS

487

Kraton® rubber business before he retired some years ago. I picked up the task of presenting 'Kraton 101: the Basics', once he had left, and this material has been shared with hundreds of people along the way, with people new to the business, and with customers. That important legacy continues today, and in a parallel universe I know Geoff is still in the front lines ably explaining, in simple, but not simpler terms, what makes these polymers as fresh and new as they were when they we first understood over 30 years ago.

4

APPLICATIONS OF STYRENIC BLOCK COPOLYMER ELASTOMERS

The thermoplastic nature of block copolymers allows their use in a wide range of elastomeric applications from pressure-sensitive adhesives to paving modification. Traditional elastomers, such as natural rubber, are soft and strong, but require vulcanization to resist flow at use temperatures. Polyolefin thermoplastic rubbers process easily but exhibit low strength and relatively poor elasticity if uncrosslinked. Block copolymers fill a unique niche combining rapid thermoplastic processing with soft, strong elastomeric properties. ABA triblocks where the A block is either a glassy polymer, e.g. polystyrene, or semicrystalline polymer, e.g. polyethylene, and the B block is an elastomer such as polybutadiene or polyisoprene achieve their strength through phase separation of the blocks. This phase separation results in physical crosslinking below the Tg or melting point of the endblocks. Above the Tg or melting point of the endblocks, this physical crosslinking is weakened, allowing triblock copolymers to flow with a rheology amenable to normal thermoplastic processing techniques. Thermoplastic elastomers can be divided into two major classes: ABA-type block copolymers such as poly (styrene—butadiene—styrene) triblocks described in this section, which are composed of well defined blocks, and statistical block copolymers such as polyurethanes, polyethers, and polyamides. ABA-type block copolymers have much cleaner phase separation than segmented block copolymers, and typically are softer, usually less than 70 Shore A. Segmented block copolymers generally exhibit greater strength in the higher hardness ranges above 90 Shore A. More extensive reviews of thermoplastic elastomers can be found in key references [71].

4.1

COMMERCIAL STYRENIC BLOCK COPOLYMERS

The first commercial thermoplastic elastomers deriving their properties from an ABA block copolymer structure were poly(styrene—isoprene—styrene)and poly (styrene—butadiene—styrene) triblocks introduced in 1965 at an ACS Rubber

488

R. C. BENING ETAL

Division Meeting by Bailey et al. [72] of Shell Development Company. Initial applications were in molded shoe soles, hot melt adhesives, and injection molded rubber goods. To further extend the range of applications, these polymers were made more stable by full hydrogenation in the early 1970s. Fully hydrogenated styrene blocks exhibited poor resistance to oils, leading to the introduction of polymers with selectively hydrogenated rubber blocks. Through selective hydrogenation, SBS became styrene—(ethylene—butylene)— styrene (SEBS), and SIS became styrene—(ethylene—propylene)-styrene (SEPS). A wide range of triblock, diblock and star architectures were introduced in the 1970s with styrene contents ranging from 10% to more than 70%. The members of this broad polymer family that contain >50% rubber block are genetically known as styrenic thermoplastic elastomers. In 1998, worldwide production of styrenic block copolymers (SBCs) from more than 20 manufacturers was more than 106tons. The major manufacturers and the trade names of their products are listed in Table 21.2. The primary applications for SBCs are listed in Table 21.3. There are several characteristics that are general to their applications: Table 21.2

Top producers of styrenic block copolymers in order of decreasing volume

Company

Trade name

Polymer type

Shell Chemical

KratonR D KratonR G EuropreneR KibitonK VectorR FinapreneR CalpreneR Taipol TPER StereonR K-ResinR

SBS, SIS SEBS, SEPS SBS, SIS SBS SBS.SIS SBS SBS, SEBS SBS SBS SBS"

Enichem Chi Mei Dexco Fina Repsol Taiwan Synthetic Rubber Bridgestone/Firestone Phillips

SBS star polymers with >50% PS which are not elastomeric.

Table 21.3 Primary applications of styrenic block copolymers and their 1997 global volumes [73] Application

Volume (tons)

Adhesives and sealants Bitumen modification Footwear Polymer modification Viscosity modifiers

142000 170 000 160000 80 000 17 000

STYRENIC BLOCK COPOLYMER ELASTOMERS

489

1. Because of their ability to blend with a wide variety of materials, SBCs are rarely used alone. In the applications described below, it is unusual for an SBC to make up more than 50 % of a product. 2. Because of their low polarity, SBCs have relatively weak adhesion to other materials and limited oil resistance, and are therefore generally used in applications where polarity is not important or is undesirable. 3. Strength generally increases as the driving force for phase separation between the blocks, XN, increases in the order SIS < SBS « SEPS, SEBS, where N is the polymer molecular weight and X is the segmental interaction parameter. 4. Although the equilibrium morphologies of block copolymers are of great academic interest, SBCs are far from equilibrium in most applications. Figure 21.12 shows the irregular domain shapes typical of a melt-processed SEBS triblock copolymer dispersed in a matrix of polycarbonate. While this polymer would have a cylindrical morphology at equilibrium, here the polystyrene domains are irregular in length and retain only short-range order in packing. 4.2

ADHESIVES AND SEALANTS

Adhesives and sealants were some of the first products to be made from block polymer thermoplastic elastomers and remain among the most important

Figure 21.12 TEM of an SEBS polymer dispersed in a matrix of polycarbonate. The styrene domains and the polycarbonate matrix are stained dark with RuO4 making the rubber appear light

490

R. C. BENING ETAL

today. Unlike epoxy- and cyanoacrylate—based adhesives, SBC adhesives have no strong specific interactions with substrates. The van der Waals forces of these nonpolar materials provide very weak molecular level adhesion. The principal characteristics of SBCs that make them useful in adhesives are their ability to conform to the surface and to dissipate energy uniformly throughout the bulk of the adhesive layer during peel deformation. The ability of a material to conform quickly to a surface is generally referred to as the 'tack' of the adhesive and is inversely proportional to its modulus. SIS polymers containing 10—20% styrene fraction are the most commonly used SBCs in adhesives since isoprene, owing to its high entanglement molecular weight, has the lowest rubber modulus of the commercial SBCs. The low styrene content assures that the styrene domains will be predominantly spherical, thus minimizing their contribution to the adhesive's modulus. Typical total molecular weights are in the range 100000–200000. The low styrene block molecular weight also creates a lower order—disorder transition temperature, which can be further reduced to practical melt temperatures of 150–170 °C by formulation. Dilution of the base triblock copolymer with oil, diblock and tackifying resins typically improves the strength of adhesion. Although this may seem counterintuitive, the adhesion strength is more closely tied to the adhesive's ability to absorb energy than its bulk strength. Figure 21.13 shows the elastic and loss modulus of a typical SIS polymer as a function of temperature. Note that the loss modulus, the ability of the material to dissipate energy, peaks near its Tg at -55 °C, but is low near room temperature where the adhesives are 10000 Storage Modulus

Ispprene Phase Tg

1000

Formulated Loss Modulus

100 10

Styrene Phase Tg ""^"^^•"••^^^^w.

0.1

0.01 -150

-100

-50

0

50

100

150

Temperature (°C) Figure 21.13 Storage and loss moduli of an SIS triblock copolymer as a function of temperature. Formulation with tackifying resins brings the peak in loss modulus to near ambient temperatures for adhesives

STYRENIC BLOCK COPOLYMER ELASTOMERS

491

typically used. Most pressure-sensitive adhesives are, therefore, formulated to a Tg near — 10°C by using stiff, polyisoprene–soluble oligomers called tackifying resins to increase the rubber block Tg. To maximize adhesion further, it is also necessary to match the bulk strength of the adhesive to the strength of the surface adhesion to the substrate. An adhesive that is too elastic will peel off of the surface easily since it adheres only by van der Waals forces. On the other hand, an adhesive that is too weak will flow viscously, absorbing some energy in doing so, but will fail cohesively, that is, in the bulk of the adhesive. A properly balanced adhesive will have enough strength to distribute uniformly the stress from the surface throughout the adhesive so that it will flow with maximum viscous dissipation at a stress which is just less than the stress necessary to disbond the adhesive from the surface. A wide variety of additives have been developed to tailor precisely the modulus, strength and Tg of the soft and hard phases. These fall into three categories: 'plasticizers' (hydrocarbon process oils), midblock resins and endblock resins. 'Plasticizers', not to be confused with the family of polar phthalate esters ordinarily used for making flexible PVC, which behave as good solvents for the styrene endblock, are generally diblock copolymers and hydrocarbon oils chosen to match the solubility parameter of the rubber block. Diblocks are typically used in combination with, or in place of, oils because they will not migrate to the surface and form a weak boundary layer. The function of the plasticizer is to reduce the modulus of the rubber phase, the viscosity of the melt during application, the strength of the bulk adhesive, and the cost of the adhesive. Midblock resins are rigid aliphatic oligomers such as oligomerized cyclopentadiene. These also have several functions: to increase the Tg of the midblock, perhaps add specific adhesion to either the backing or the substrate, and to reduce the adhesive's melt viscosity, strength, and cost. Similarly, low molecular weight aromatic polymers and oligomers such as polyphenylene oxide can be used to modify the properties of the endblock. These may be used to weaken the endblock to reduce process viscosity, or increase the Tg of the endblock to increase service temperature; however, their role in the function of the adhesive is much less important than modifiers for the midblock. A typical pressure-sensitive adhesive formulation might contain 40% SIS, 55% tackifying resin and 5% paraffinic/naphthenic oil. Detailed formulating methods are described by Ewins et al. in the Handbook of Pressure Sensitive Adhesives Technology [74]. Many adhesives require special properties, for example, SBS polymers are often used for crosslinkable adhesives because the polybutadiene midblock is readily crosslinked by peroxides, UV light or electron-beam radiation. The vinyl content of the butadiene block of SBS polymers is often increased from the standard 8%) to near 50% to make it more reactive to crosslinking. Star architectures are also used to reduce the number of crosslinks necessary to form a continuous network throughout the adhesive. For adhesives that require

492

R. C. BENING ETAL

long-term stability in the melt, SEBS or SEPS polymers, which have saturated rubber blocks, are the materials of choice. The precise control that anionic synthesis affords even allows star or radial block copolymers, which combine both saturated rubber blocks attached to the styrene blocks for stability and strength, and unsaturated polyisoprene arms for tack and low modulus. The inherent purity required for anionic synthesis of SBCs has allowed FDA clearance of most SBCs, which permits their use in sensitive applications such as diaper and personal care adhesives. One of the most important features of SBC adhesives is their versatility. Because they are thermoplastic materials of carefully controlled molecular weight, they can be applied both as hot melts and from solvent. More importantly, adhesive manufacturers can tailor their properties to match a wide range of applications. The primary competitors for SBCs in the pressure-sensitive adhesive market are acrylic copolymers. While acrylics have better specific adhesion to polar materials, adhesive manufacturers cannot easily formulate them for varied adhesion applications.

4.3

BITUMEN MODIFICATION

The modification of bitumen, divided into roofing and secondarily paving segments, has grown into the second largest market in the USA and the largest single application for SBCs in Europe. SBS block copolymers are almost exclusively chosen for bitumen applications because of polyisoprene's undesirable tendency to undergo chain scission during degradation and the lower cost of butadiene monomer. The optimal SBS polymers usually contain 30 % styrene and are generally two to four arm stars of high molecular weight, typically 100000–300000. The high styrene content and radial architecture help to retain hard segment continuity at low polymer levels. Where high processing temperatures or harsh climates such as deserts require better stability, saturated SEBS polymers are sometimes used. Unlike adhesive formulations, which can be controlled by the manufacturer, the natural composition of the bitumen largely controls the property of the modified product. The composition of bitumen, a by-product of crude oil refining, is complex and varies with the crude oil source. Its primary components are: 1. saturated hydrocarbons, which are compatible with polybutadiene and thus swell the midblocks of SBCs; 2. aromatic hydrocarbons, which, although largely incompatible with both blocks near room temperature, serve to plasticize the styrene blocks at dissolving temperatures; and 3. resins and asphaltenes, which are incompatible with both blocks.

STYRENIC BLOCK COPOLYMER ELASTOMERS

493

The saturate/aromatic balance is particularly important in making good roofing or paving material. If too little aromatic fraction is present, the strong phase separation of the SBC will prevent it from dissolving in the asphalt at processing temperatures of 160—190°C. Too much aromatic fraction will reduce the strength of the styrene phase association at ambient temperatures, thus reducing the upper service temperature of the modified bitumen. During melt mixing a wide range of grafting, scission, and coupling reactions occur which help to compatibilize the SBS and bitumen. At a concentration of 3 % SBS, polymers begin to form a continuous network swollen by the soluble bitumen components. At 6 % or greater, SBS polymers form continuous polymer networks which increase the toughness of the bitumen at low temperature and greatly reduce flow at elevated temperature. Roofing membranes require higher elasticity than paving bitumen; therefore, they typically contain 8–20% SBS polymers and perform more like swollen elastomers than bitumen. At the lower concentrations typically used in paving, 3-6%, the SBS serves as a more dispersed modifier that reduces rutting and fatigue by increasing elasticity and improving low-temperature cracking resistance. The primary competitors for block copolymers in bitumen modification are styrene–butadiene emulsions, atactic polypropylene and ethylene–vinyl acetate copolymers. 4.4

FOOTWEAR

Footwear was the first commercial application for styrenic block copolymers. In the USA, SBCs have been largely replaced by polyurethanes and PVC, but footwear remains a significant application in Europe and is now the largest SBC market in the Far East, with global consumption of over 150 0001 of SBS polymer in 1997. Because shoe soles require relatively stiff elastomers, they are highly filled compounds containing SBS linear or radial block copolymers with styrene contents between 30 and 50 %. A typical footwear compound consists of 30%) SBS, 30% oil, 15% polystyrene, and 25% talc and/or calcium carbonate. SBS-based footwear has excellent skid resistance, low-temperature flexibility and fatigue life. However, it lacks the abrasion resistance under rapid loading that is required for high-performance footwear because of the increase in energy absorption at elevated temperature associated with the polystyrene domain Tg. 4.5

POLYMER MODIFICATION

Polystyrene and ABS are the polymers that are most commonly modified with SBCs. SBS triblocks and SB diblocks with 30–40% styrene are often used in concentrations near 5% either as a toughener alone in combination with polybutadiene in high-impact polystyrene. In addition to the use of elastomeric

494

R. C. BENING ETAL

SBCs, a significant fraction of the more than 150 000 tons of high styrene content (>60%) SBS polymers are also used in compounds with polystyrene for transparent film, cups, and trays. Polymers with processing temperatures above 200 °C such as poly(phenylene oxide) and polycarbonate are commonly toughened with SEBS or SEPS polymers. However, most engineering polymers are too polar to be effectively modified by nonpolar block copolymers. One notable exception is polyamides, which are effectively toughened by SEBS polymers grafted with maleic anhydride. The maleic anhydride reacts with amine end groups in the polyamide to form a graft copolymer. Figure 21.14 shows the diffuse interface between the nylon 6 phase and a maleic anhydride grafted SEBS. Hydrogenated SBCs are often used to modify polyolefins such as polypropylene, polybutylene and polyethylene. One of the unique characteristics of strongly phase-separated block copolymers such as high molecular weight (>50 000) SEBS and SEPS is their response to shear in the melt. These polymers retain their phase-separated structure well above the Tg of the polystyrene because their order—disorder transition temperatures are above processing temperatures. This phase separation strongly inhibits flow in the absence of shear resulting in infinite viscosity at zero shear rates. The application of shear

Figure 21.14 TEM of a maleic anhydride—functionalized polymer in polyamide 66. The styrene domains are stained dark with RuO4. The maleic anhydride in the rubber phase reacts with the polyamide producing a diffuse interface and fine dispersion of the block copolymer

STYRENIC BLOCK COPOLYMER ELASTOMERS

495

produces a 'block flow' which is more similar to the flow of ultra-high molecular weight polymers than to traditional homopolymers. Addition of homopolymers such as polypropylene greatly increases apparent flow and results in the formation of mechanically stabilized interpenetrating networks (IPNs), which have been described by Gergen et al [75]. The bi-continuous nature of IPNs allows better expression of the properties of both components than a dispersed polymer blend. The morphology of an IPN resembles that of a blend undergoing spinodal decomposition, with the exception that the SEBS or SEPS phase is always convex in its interfacial curvature regardless of the volume fraction, as shown in Figure 21.15. This structure is formed in the high shear zones of mixing equipment such as twin-screw extruders and is prevented from 'ripening' after the shear is removed as it would in normal incompatible blends by the infinite viscosity of the SEBS or SEPS phase. These Theologically stabilized IPNs are often used for injection molding parts such as toys, grips for sporting equipment, and overmolded surfaces.

Figure 21.15 TEM of an IPN of polypropylene (light phase) and SEBS where the styrene domains are stained dark withRuO4.The structure of both phases is continuous, although it appears discontinuous in this two-dimensional slice

496 4.6

R. C. BENING ETAL VISCOSITY INDEX IMPROVERS AND OTHER APPLICATIONS

There are a wide variety of smaller applications for SBCs. For example, the solubility of SBCs in hydrocarbons and control of molecular weight make them ideal viscosity modifiers for automotive engine oils. Automotive engine oil manufacturers use S—EP diblocks, S—EP stars and PEP stars to increase the viscosity of oils resulting in what are commonly called 'multigrade oils'. Controlled degradation is another key performance characteristic of oil modifiers. Because engines operate at high temperatures and shear rates in the presence of oxygen and oxygenated hydrocarbons, the ideal polymer backbone is saturated polyisoprene, PEP. PEP-based polymers tend to degrade by chain scission rather than crosslinking, as hydrogenated polybutadiene does and thus do not add appreciably to engine deposits. Figure 21.16 shows the uniform spherical micelles formed by an S—EP diblock in hydrocarbon oil. The formation of

Figure 21.16 TEM of S—EP micelles deposited on a carbon-coated grid. The polystyrene cores are stained dark with RuO4 The shells of the micelles are made visible by shadowing with carbon/platinum from the direction of the arrows. The micelles that appear not to have shells are on the underside of the film

STYRENIC BLOCK COPOLYMER ELASTOMERS

497

micelles depends primarily on the solvency and molecular weight of the coreforming block and has been extensively explored by Frank and co-workers [76], among others. These micelles act as 'dynamic stars' that can sustain repeated shear events without significant viscosity reduction. Anionic PEP star polymers are also commonly used as viscosity modifiers. The denser shape of star polymers and micelles allows them to deform without breaking in shear fields that would degrade linear polymers, thus making them more efficient thickeners over time. Ethylene—propylene copolymers, acrylates and polyisobutylene are the most common competitors to block copolymers in these markets. The narrow molecular weight distributions and star architecture of the SBCs give them a better balance of properties at molecular weights of > 100 000 than the competing polymers. The concentration at which polymers are used as oil modifiers is close to C*, the overlap concentration that marks the boundary between dilute and semidilute solution. Below C*, the lack of entanglements makes polymers rather inefficient viscosity modifiers. Conversely, significantly above C*, polymers are very efficient thickeners, but the entanglements lead to excessive shear degradation. As the concentration of block copolymer micelles is increased in solution the micelles begin to pack into a pseudo-crystalline lattice. At high concentrations the interpenetration of the chains between micelles presents a significant barrier for movement of the micelles relative to their neighbors, resulting in a macroscopic yield stress. These pseudo-crystalline gels are useful for cable filling compounds in optical cables both to prevent water incursion and to lubricate the optical fibers to prevent breakage during flexing. Gels made from triblocks and diblocks have a variety of other smaller applications, such as electrical connection filling compounds and transparent gel candles.

REFERENCES 1. N. R. Legge, G. Holden, H. E. Schroeder, eds, Thermoplastic Elastomers, a Comprehensive Review, Hanser, Munich, 1987, 2nd edn; A. Noshay, J. E. McGrath, Block Copolymers, Overview and Critical Survey, Academic Press, New York, 1977; G. Holden, Thermoplastic Elastomers, Hanser Munich, 2000; R. P. Quirk, D. J. Kinning, L. J. Fetters, in Comprehensive Polymer Science, G. Allen, J. C Bevington, S. L. Aggarwal, eds, Pergamon Press, Oxford, 1989; P. Dreyfuss, L. J. Fetters, D. R. Hansen, Rubb. Chem. Technol, 53 (1980). 2. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996; J. E. McGrath, ed., Anionic Polymerization, Kinetics, Mechanisms, and Synthesis, ACS Symposium Series 166, American Chemical Society, Washington, DC, 1981; M. Morton, Anionic Polymerization: Principles and Practice, Academic Press, New York, 1983.

498

R. C. BENING ETAL

3. G. Odian, Principles of Polymerization, 2nd edn, Wiley—Interscience, New York, 1981, pp. 377378, 388–389. 4. This relationship holds for monofunctional initiators. If an initiator starts n chains, Xn = [M]/n[I]. 5. M. Szwarc, M. Levy, and R. Milkovich, J. Am. Chem. Soc., 78, 2656 (1956); M. Szwarc, J. Polym. Sci., Part A, Polym. Chem., 36, ix (1998). 6. G. Holden and R. Milkovich, US Patent, 3265 765 (1966). 7. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 131-154. 8. This is partly a result of the low solubility of these reagent is hydrocarbon solvents. Polar solvents needed to form the initiator accelerate the rate of termination at elevated temperatures. 9. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, p. 434; A. F. Halasa, W. L. Wu, Polym. Prepr., Am. Chem. Soc., Div. Polym. Chem., 37, 676 (1996). 10. L. Fetters, J Polym. Sci., Part C, 26, 1 (1969). 11. A. Noshay, J. E. McGrath, Block Copolymers, Overview and Critical Survey, Academic Press, New York, 1977, pp. 188-190. 12. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 110–113. 13. Y. S. Yu, R. Jerome, Ph. Teyssie, Macromolecules, 27, 5957 (1994). 14. G,Y-S Lo, E. W. Otterbacher, A. L. Gatzke, L. H. Tung, Macromolecules, 27, 2233 (1994). 15. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, 1996, pp. 155–171. 16. D. J. Worsfold, S. Bywater, Can. J. Chem, 38, 1891 (1960). 17. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 158–163 18. N. P. Balsara, L. J. Fetters, Macromolecules, 32, 5147 (1999); S. Bywater, Macromolecules, 31, 6010 (1998); J. Stellbrink, L. Willner, O. Jucknischke, D. Richter, P. Under, L. J. Fetters, J. S. Huang, Macromolecules, 31, 4189 (1998); A. A. ArestYakubovish, J. Polym. Sci., Part A, 35, 3613 (1997); L. J. Fetters, J. S. Huang, R. N. Young, J. Polym. Sci., Part A, 34, 1517 (1996); S. Bywater, Polym. Int., 38, 325 (1995); R. N. Young, L. J. Fetters, J. S. Huang, R. Krishnamoorti, Polym. Int., 33, 217(1994). 19. D. Baskaran, A. H. E. Muller, S. Siviram, Macromol. Chem. Phys., 201, 1901 (2000). 20. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 173–196. 21. H. L. Hseih, R. C. Farrar, K. Udipi, in Anionic Polymerization, Kinetics, Mechanism and Synthesis, J. E. McGrath, ed., ACS Symposium Series 166, American Chemical Society, Washington, DC, 1981, pp. 394–398. 22. M. Morton, F. R. Ells, J. Polym. Sci., 61, 25 (1962). 23. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 244–245. 24. J. R. Wunsch, C. Beumelberg, G. Jauer, K. Knoll, P. Weinkoetz, in Proceedings of RAPRA Conference 'TPE98', 1998, p. 1. 25. H. L. Hseih, R. C. Farrar, K. Udipi, in Anionic Polymerization, Kinetics, Mechanism and Synthesis, J. E. McGrath, ed., ACS Symposium Series 166, American Chemical Society, Washington, DC, 1981, pp. 397–01; H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 252-255. 26. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996. pp. 252-255.

STYRENIC BLOCK COPOLYMER ELASTOMERS

499

27. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 312–313. 28. Detection of the diblock component in a sequential triblock by common methods such as gel permeation chromatography (GPC) can be difficult, especially if the molecular weight of the diblock is high in comparison with the third styrene block, as is often the case. The molecular weights of the triblock and the diblock may be too similar for size exclusion methods to separate them. 29. J. M. Yu, Ph.Dubois, R. Jerome, Macromolecules, 29, 7316 (1996). 30. P. Dreyfuss, L. J. Fetters, D. R. Hansen, Rubb. Chem. Techno!., 53 (1980), 756. 31. P. Dreyfuss, L. J. Fetters, D. R. Hansen, Rubb. Chem. Technol., 53 (1980), 757. 32. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, p. 315. 33. Y. Sasaki, S. Y. Nakamichi, Japanese Patent 2802663 (1998). 34. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 347-350; A. Noshay and J. E. McGrath, Block Copolymers, Overview and Critical Survey, Academic Press, New York, 1977, p. 194. 35. Quirk, D. J. Kinning, L. J. Fetters, in Comprehensive Polymer Science, G. Allen, J. C Bevington, S. L. Aggarwal, eds, Pergamon, Press, Oxford, 1989, pp. 10–11; H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 350-352. 36. R. P. Quirk, B. Lee, L. E. Schock, Macromol. Chem., Macromol Symp., 53, 201 (1992). 37. B. J. Bauer, L. J. Fetters, Rubb. Chem. Technol., 51, 406 (1978). 38. D. H. Rein, P. Rempp, P. J. Lutz, Macromol. Chem. Phys., 199, 569 (1998); H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 337-342. 39. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 346-347. 40. T.-Y. Wang, R. C.-C. Tsiang, J.-S. Liou, J. Wu, H.-C. Sheu, J. Appl. Polym. Sci., 79, 1838 (2001). 41. H. L. Hsieh, R. P. Quirk, Anionic Polymerization, Marcel Dekker, New York, 1996, pp. 217–223. 42. D. Hesterwirth, D. Beckelmann, F. Bandermann, J. Appl. Polym. Sci., 73, 1521 (1999); D. Beckelmann, F. Bandermann, J. Appl. Polym. Sci., 73, 1533 (1999). 43. H. L. Hseih, W. H. Glaze, Rubb. Chem. Technol., 43, 22 (1970). 44. D. B. Collum, Ace. Chem Res., 25, 448 (1992). 45. N. K. Singha, S. Bhattacharjee, S. Siviram, Rubb. Chem. Technol., Rubb. Revi., 70, 309 (1997). 46. N. T. McManus, G. L. Rempel, J. Macromol. Chem., Rev. Macromol. Chem. Phys., C35, 239(1995). 47. K. A. Johnson, Polym. Prepr., Am. Chem. Soc., Div. Polym. Chem., 41, 1525 (2000); S. J. Lapporte, Ann. N. Y. Acade. Sci., 158, 510 (1969). 48. N. K. Singha, S. Bhattacharjee, S. Siviram, Rubb. Chem. Technol. Rubb. Rev., 70, 309(1997), pp. 342–343. 49. N. T. McManus, G. L. Rempel, J. Macromol. Chem., Rev. Macromol. Chem. Phys., C35, 239 (1995), pp. 244–246; N. K. Singha, S. Bhattacharjee, S. Siviram, Rubb. Chem. Technol., Rubb. Rev., 70, 309 (1997), pp. 319-321. 50. Y. Kishimoto, T. Masabuchi, US Patent 4673414 (1987); L. R. Chamberlain, C. J. Gibler, US Patent 5039755 (1991). 51. C. Viola, A. Valleri, European Patent 914867 (1999); A. Vallei, C. Cavallo, European Patent 816382 (1998).

500

R. C. BENING ETAL

52. D. A. Hucul, US Patent 5028 665 (1991); D. A. Hucul, US Patent 5110 779 (1992). 53. M. D. Gehlsen, F. S. Bates, Macromolecules, 26, 4122 (1993); F. S. Bates, M. D. Gehlsen, V. L. Hughes, P. Brant, US Patent 5352 744 (1994). 54. H. L. Hassell, US Patent 3644588 (1972). 55. G. Kaszas, J. E. Puskas, J. P. Kennedy, W. G. Hager, J. Polym. Sci. Part A: Polym. Chem., 29,427 (1991); K. Koshimura, H. Sato, Polym. Bull., 29 (1992); R. B. Storey, B. J. Chisholm, Macromolecules, 26, 6727 (1993); M. Gyor, Z. Fodor, H.-C. Wang, R. Faust, J. Macromol. Chem., Pure Appl. Chem., A31, 2055 (1994). 56. Y. Tsunogae, J. P. Kennedy, J. Polym. Sci. Part A: Polym. Chem., 32, 403 (1994); J. P. Kennedy, S. Midha, Y. Tsunogae, Macromolecules, 26, 429 (1993); D. Li, R. Faust, Macromolecules, 28, 4893 (1995); Z. Fodor, R. Faust, J. Macromol. Chem., Pure Appl. Chem., A32, 575 (1995). 57. X. Cao, R. Faust, Macromolecules, 32, 5487 (1999); X. Cao, L. Sipos, R. Faust, Polym. Bull., 45, 121 (2000). 58. E. E. Malmstrom, C. J. Hawker, Macromol. Chem. Phys., 199, 923 (1998). 59. D. Benoit, E. Harth, P. Fox, R. Waymouth, C. J. Hawker, Macromolecules, 33, 363 (2000). 60. C. Zune, P. Dubois, R. Jerome, Polym. Int., 48, 565 (1999). 61. J. D.Tong, R. Jerome, Polymer, 41, 2499 (2000). 62. J. D. Tong, R. Jerome, Macromolecules, 33, 1479 (2000). 63. J. D. Tong, G. Moineau, Ph. Leclere, J. D. Bredas, R. Lazzaroni, R. Jerome, Macromolecules, 33, 470 (2000); G. Moineau, M. Minet, Ph. Teyssie, R. Jerome, Macromol. Chem. Phys., 201, 1108 (2000). 64. M. Morton, in N. R. Legge, G. Holden, H. E. Schroeder, eds, Thermoplastic Elastomers, a Comprehensive Review, Hanser, Munich, 1987, p. 75. 65. G. Holden, E. T. Bishop, N. R. Legge, J. Polym. Sci.. Part C, 26, 45 (1969) 66. Reference Missing * 67. W. P. Gergen, R. G Lutz, S. Davison, in N. R. Legge, G. Holden, H. E. Schroeder, eds, Thermoplastic Elastomers, a Comprehensive Review, Hanser, Munich, 1987, p. 520. 68. M. Morton, Encyclopedia of Polymer Science and Technology, John Wiley & Sons Inc., New York, 1971, Vol. 15, p. 508. 69. G. Holden, E. T. Bishop, N. R. Legge, Thermoplastic Elastomers, Proceedings of the International Rubber Conference, 1967, Maclaren, London, 1968, p. 287: J. Polvm. Sci., Part C, 26, 37 (1969). 70. E. T. Bishop, S. Davidson, J. Polym. Sci., Part C, 26, 60 (1969) 71. G. Holden, N. R. Legge, in N. R. Legge, G. Holden, H. E. Schroeder, eds, Thermoplastic Elastomers, Hanser, Munich, 1992; W. M. Halper, G. Holden, in B. M. Wlaker, C. P. Rader, eds, Handbook of Thermoplastic Elastomers, Van Nostrand Reinhold, New York, 1988. 72. J. T. Bailey, E. T. Bishop, W. R. Hendricks, G. Holden, N. R. Legge, presented at the ACS Rubber Division Meeting, Philadelphia, 1965. 73. R. M. Hoag, N. Takei, in Chemical International Handbook, SRI International, Menlo Park, CA, 1998. 74. E. E. Ewins, D. J. St. Clair, J. R. Erickson, W. H. Korcz, in D. Satas, ed., Handbook of Pressure Sensitive Adhesives Technology, Van Nostrand Reinhold, New York, 1989. 75. W. P. Gergen, R. J. Lutz, S. Davison, in N. R. Legge, G. Holden, H. E. Schroeder. eds, Thermoplastic Elastomers, Hanser, Munich, 1987. 76. A. S. Yeung, C. W. Frank, Polymer, 31, 2089, 2101-2111 (1990); D. A. Ylitalo. C. W.Frank, Polymer 37, 4969 (1996).

22

Preparation, Properties and Applications of High Styrene Content Styrene-Butadiene Copolymers DAVID L. HARTSOCK AND NATHAN E. STACY Chevron Phillips Chemical Co., Bartlesville, OK, USA

1

HISTORY

The development of high styrene content styrene-butadiene copolymers (SBCs), such as K-Resin® SBC, is best thought of as a branch off the history of anionic polymerization and rubber. A number of excellent reviews cover this aspect of the subject in great detail, and should be obtained for detailed examination of the history of rubber and anionically synthesized rubber polymers [1–3]. What follows is a brief overview to fit the high styrene content SBC into a historical context. There is evidence that natural rubber was used by early Americans to make rubber balls over 2000 years ago. However, it has only been since the early twentieth century that rubber has become crucial to maintaining our standard of living in our current technology-based society. Synthetic rubbers, or elastomers as any artificial substance with elastic properties is called, have been a subject of intense research since the late 1800s. These materials were critically needed in the first half of the twentieth century to replace natural rubber in the tires for the newly invented automobile, due to shortages of natural rubber caused by wars. Anionic polymerization of vinyl monomers using alkali metals was reported very early in the twentieth century [4] with the first n-butyllithium polymeriza-

Modern Styrenic Polymers: Polystyrene and Stvrenic Copolvmers. Edited by J. Scheirs and D. B. Priddy 2003 John Wiley & Sons Ltd

502

D. L HARTSOCK AND N. E. STACY

tion of vinyl monomers being reported by Ziegler in 1929 [5]. While some limited work with anionic polymerization continued, most production of elastomers moved to the emulsion polymerization process. Anionic polymerization was not examined for commercial production again until the 1950s, with research done by Stavely [6] and Szwarc et al. [7,8]. The ability to make polymers with tightly defined and controlled microstructures [9] triggered a resurgence of interest in the use of alkyllithium-initiated polymers, which has continued to the present day. This control over the microstructure and the block structures gives significarrt control over the morphology of the final polymer. This morphology controls the final properties of the polymer. Firestone, Shell and Phillips built the first commercial production plants for anionically polymerized SBCs in the 1950s. Research into new products based on this technology led to initial runs of high-styrene SBCs at Phillips Petroleum in the 1960s. K-Resin SBC was invented by Alonzo Kitchen, a research chemist at Phillips Petroleum Research and Development laboratories. With inventorship came the opportunity to name the new resin, which he called 'K-Resin'. The first pilot plant resins were made in 1967, and commercial samples were prepared for test marketing in 1968. Commercial production started in October of 1972 at the SBC plant in Borger, Texas, on a 10 million pound per year capacity line. Initially, the solution product was steam stripped to remove the hydrocarbon solvent, but this left a significant haze in the resin. The finishing system was quickly converted to a devolatilizing extruder. Commercial production continued at this plant until 1979, ending with the opening of a new production facility at Adams Terminal (later renamed the Houston Chemical Complex) in Pasadena, Texas. The new plant had a nameplate capacity of 120 million pounds per year. Plant expansions increased the production capacity in 1988 and 1994 to a total nameplate capacity around 300 million pounds per year. Beginning in 1983, additional producers of SBC producers emerged. The first of these was BASF, with the opening of a 22 million pound annual capacity plant in Ludwigshafen, Germany. This was followed by several other companies, including Asahi Chemical Co., Denka (Denki-Kagaku), Firestone Polymers, and Fina Chemical Co. It is estimated that by 2003, worldwide capacity to manufacture SBC may exceed one billion pounds.

2

SBC SYNTHESIS AND MANUFACTURE

K-Resin* SBC synthesis is a batch anionic solution polymerization of styrene and 1,3-butadiene using an n-butyllithium (NBL) initiator in a process referred to as 'living polymerization'. Although often referred to as a catalyst, each NBL gives rise to a distinct polymer chain. Polymer chains grow by adding monomer

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

503

to the living carbanion in the absence of chain transfer and termination. The polymer lithium chains can stay active for long periods of time if protected from water and oxygen and high temperatures. This type of polymerization can also result in excellent control in making a very narrow molecular weight distribution, approaching a 1.0 Mw/Mn ratio. This narrow molecular weight distribution can be enhanced by the presence of a polar additive such as tetrahydrofuran in the polymerization solvent. The THF will increase the rate of initiation to be comparable to the rate of propagation. The polymerization process is by necessity a batch process because of the living nature of the polymerization process and the need to form the blocks required for the controlled polymer structure. A high polymer solids content in the reactor is made possible by sequential charging of the components and the use of coolant coils to extract heat evolved during polymerization. It is desirable to keep the reaction temperatures no higher than 85 °C to prevent polystyryllithium termination. The solvent used for polymerization of K-Resin SBC is cyclohexane. The number, size and composition of the individual polymer chains created in the polymerization are controlled by the addition of initiator and monomers. Monomers can be added in varying levels and sequences, depending on the properties required. The first commercial KR01 grade was a simple single initiator A-B coupled block copolymer giving a final A-B-A structure. Ensuing polymer development has resulted in much greater complexity of polymer architecture. The ability to reproduce the same molecular structure during the polymerization sequence is critical to producing a product with acceptable quality. Consistency in charge ratios, sequence of monomer addition, reactor solids, initiation and peak temperatures are all critical to product quality. Monomer purity is paramount to making high-quality, consistent polymers. Trace levels of moisture and monomer inhibitor act as poisons and must be removed from the monomer. At best, the poisons react to inactivate the initiator. At worst, the poisons react at various stages of polymerization to inactivate and terminate the growing polymer lithium chain to create free polystyrene or diblock copolymers. Free water can be removed by a coalescer or decanter, and dissolved water can be eliminated by either azeotropic distillation or adsorption on anhydrous alumina to a level less than lOppm. Both monomers and recycled solvents can also carry trace levels of chemical poisons that can destroy the initiator or inactivate the growing polymer lithium chains. These should be reduced to a low and consistent level for optimum operation. After polymerization, the polymer may be coupled. A variety of chemicals can be used as coupling agents, including di- or trifunctional chlorosilanes or mutifunctional esters. Any coupling agent will cause some simple termination of the polymer lithiums chains, resulting in a percentage of uncoupled polymer. Following polymerization and coupling, the polymer is treated and transferred to a feed tank and the process becomes continuous from this point. The

504

D. L. HARTSOCK AND N. E. STACY

polymer 'cement' is composed of more solvent than polymer, and must be concentrated and nearly all the solvent removed before the polymer is pelletized. Pellets are typically cut underwater to produce pellets of spherical shape to the desired size. The pellets are then dried before packaging and transport. The polymerized butadiene retains an unsaturated site, which can degrade and crosslink when processed at elevated temperatures. The crosslinking results in a loss of clarity and other physical properties. To protect the polymer in subsequent devolatilization and converting processes against thermal and oxidative degradation, stabilizers are added. The end use application of the grade plays a role in determining the type and amount of stabilization required. Degradation and crosslinking of the polymer can result in the formation of high molecular weight polymers called gels or fisheyes. Gels can be seen in sheet, bottles and film, and besides being esthetically undesirable, may also cause processing problems, such as pulling holes in film, or creating printing problems due to an uneven surface. In addition to stabilizers, other additives can also be added to SBC polymers during the manufacturing process or during subsequent fabrication. It is very common to add an antiblocking additive since it is common for SBC materials to block or adhere to themselves. The antiblock should be effective at preventing any pellet blocking as well as preventing sheet and film from blocking. Other additives may also be incorporated, such as colors or mold releases.

3

KEY FEATURES, PROPERTIES AND GRADES

The uniqueness of these polymers is derived from a combination of performance attributes. The SBC family of polymers offers outstanding clarity and excellent impact strength or shatter resistance, and are easy to process. Primarily, these type polymers fill the gap between low-cost commodity materials and high-cost performance polymers. For example, crystal polystyrene offers excellent clarity, but very poor impact resistance, and polycarbonate offers excellent impact resistance, but at a significant cost premium. Commercial SBC grades generally are 65-85% polystyrene, with the remaining percentage butadiene. The polystyrene and butadiene blocks form into domains, which result in several different polymer morphologies; a transmission electron micrograph of typical lamellar morphology of high-styrene SBC is shown in Figure 22.1. The high percentage of polystyrene creates a continuous or bi-continuous polystyrene domain. These hard or glassy polystyrene domains, with a Tg above room temperature, give the polymer its high

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

505

tensile and flexural strength as well as high gloss and temperature resistance. The butadiene blocks form into elastic rubbery domains and have a Tg below room temperature that enhances impact resistance, high elongation and flexibility. Polymers that have the same styrene-butadiene ratios can have very different morphologies and physical properties. The physical properties of three grades of SBC polymers are shown in Table 22.1. There are distinct differences in the mechanical properties, thermal properties and impact resistance of these different grades. While they are all very near to 75% in polystyrene content, they exhibit characteristics as if they had very different styrene: butadiene ratios. A more detailed analysis of the polymer would reveal that their morphologies are different in spite of their similar composition of styrene and butadiene.

Figure 22.1 Transmission electron micrograph of typical lamellar morphology of high-styrene SBC

506 Table 22.1

D. L HARTSOCK AND N. E. STACY R

Physical properties of K-Resin SBC grades

Property

ASTM

KR01

Density Flow rate, 5kg, 200 °C Tensile yield strength Flexural yield strength, psi 2 in (50mm) min Flexural modulus Flexural yield strength Heat deflection temperature 264 psi (108 M Pa) fiber stress Vicat softening point Instrumented impact energy Izod impact strength, notched Hardness, Shore D

D792 D1238 D6381 2750

1.01

D790 D790 D648 D1525 D37632 D2562 D2240

KR03

1.01

XK40

1.02

8

7.5

10

4400 2850

3700 2900

2210 3150

215000 6400

205000 4900

101000 3220

170

163

102

200 6 0.4 70

188 263 0.75 65

138 333 No Break 65

Unit g/cm3 g/l0min

psi 4000

psi psi F C

C

in-lb in-lb

—

Unlike high-impact polystyrene (HIPS), where the incorporation of polybutadiene in the polymer dramatically reduces clarity, the SBC has excellent clarity, low haze levels and high light transmission. The clarity of these polymers is due to their morphology. While SBCs contain a high percentage of butadiene, the butadiene is in polymer blocks which form small domains of butadiene. These domains are smaller than the wavelength of visible light, so they do not scatter visible light. A feature that helps bring value to these materials is their low specific gravity or density. Depending on the polystyrene: butadiene ratio, the density of SBC are typically 1.00–1.02g/cm3. Many clear polymers have a density 10–30% higher than nominal values for these high-styrene copolymers. While most products are sold in terms of cost per pound, an analysis of cost per unit volume is in many cases a better reflection of the actual cost. The low density of these SBCs provides a significant part yield advantage, allowing the processor to find economic incentives by being able to manufacture more parts per pound. Another thing that sets these polymers apart is their versatility in processing. Clear SBC can be processed on most conventional equipment without significant modification. With both good flow and melt strength, they can be easily run on most blow molding, sheet and film extrusion, and injection molding machines with high throughput. They can be processed at lower temperatures than many polymers and generally do not require drying. The SBC scrap can be easily reprocessed multiple times with only minor degradation to mechanical and optical properties. During the product introduction phase, there were relatively few grades of SBC available for customers to select from. The same grades were used in many different applications and processes. The same grade of SBC had to have enough melt strength to be used to blow mold a large part but also have adequate flow

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

507

properties to injection mold fast cycles. The general SBC grade also had to be elastomeric enough to blend with polystyrene and act as an impact modifier, yet have enough rigidity and scratch resistance to be used in unblended molded parts. While being successful in all these applications, they were certainly not optimized for any. Now many new process or application-specific grades of SBC are being introduced. These include high-flow and mold-release grades for injection molding. Many other grades are also being developed for optimized blending with polystyrene. As the market continues to grow, the number and range of properties of these new grades will continue to expand.

4

CURRENT COMMERCIAL APPLICATIONS

When K-Resin R SBC, the first of the clear high-styrene butadiene copolymers, was introduced in 1972, marketing efforts were initially and primarily heavily focused on blister packaging. It was originally believed that K-Resin® SBC would be an alternative to PVC in many packaging applications. Ironically, there is very little SBC used in blister packaging, an application that continues to be dominated by PVC and also PET. In spite of not capturing the blister packaging market, these polymers soon started to find success in a wide array of applications. One of the first large applications for K-Resin® SBC was in flash shields for cameras. Small cameras used a disposable external flash that contained a series of bulbs. The SBC were used as a clear cover over the bulbs because of their excellent clarity, ease of molding and impact resistance.

4.1

MAJOR MARKETS

The significant growth of SBC is the direct result of its remarkable versatility. The unique combination of high transparency, good economics and outstanding impact resistance of SBC are a starting point for many clear applications. On top of that, its ability to be used on most types of processing equipment with only minor modifications makes it easy to run at most converters. When SBC themselves may not be the optimum choice for an application, their demonstrated ability to be modified in polymer blends offers additional options. The ability to be blended in both clear and opaque blends allows the manufacturer to optimize both performance characteristics and polymer cost. The SBCs are used in a diverse array of applications, but for the purpose of this discussion they are broken into the following primary markets: single service, rigid packaging, garment hangers, flexible packaging, medical, consumer goods, and toys and displays.

508

4.2

D. L. HARTSOCK AND N. E. STACY

SINGLE SERVICE

The largest single application for SBC has been, and continues to be, in single service items, which are typically food service items filled at point of use. Originally, this classification of goods primarily included only drinking cups and lids, but it also applies to a variety of other food service items. The single service or disposable market includes single-use containers such as cups and lids sold in retail outlets, food service and for institutional use. SBC parts can be found worldwide in fast food restaurants, hotels, picnics, and sporting events. Single service use also includes clear plates, bowls, sundae cups, etc. Originally, the typical clear single service cup was made by injection molding crystal or general-purpose polystyrene. While offering outstanding clarity, these cups suffered a high incidence of leaking and broken cups. SBCs offer the ability to thermoform a thin-walled cup with great clarity and with much greater breakage resistance. Most of the single service parts are not made from 100% SBC, but are almost always a blend with crystal polystyrene (see Section 5.1.1). While they could be precompounded, it is more economical to make machine-side pellet blends with SBC and crystal polystyrene pellets. The SBCs provide excellent toughness, and a thin sheet made from 100% SBC is very flexible and tough; 100% polystyrene, is extremely brittle, and although neat polystyrene can be made into sheet on specialized and expensive orientation equipment, it cannot be easily processed into a usable cup using a conventional extrusion line. The marriage of the two materials combines the best features of both polymers, because the addition of crystal polystyrene improves both part rigidity and economics, and allows fabrication on conventional processing equipment. New grades of SBC are being developed and introduced for the single service market. The primary challenge with these new grades is to maximize the blend ratio of crystal polystyrene, while also maintaining high clarity. Because the crystal polystyrene component is less expensive relative to SBC in the blends, the economic driver is addition of higher percentages of crystal polystyrene to the blend. A new grade, such as XK40 from Chevron Phillips Chemical Company, has allowed sheet extruders to decrease significantly the percentage of SBC in the blend by 10–20%. This will continue to allow SBC blends to be more competitive with other clear polymers such as polyester, oriented polystyrene and clarified polypropylene.

4.3

RIGID PACKAGING

Rigid packaging includes any application in which a product is packaged prior to point of sale. It includes SBC manufactured in a variety of processes and packaging a wide range of goods.

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

509

Food packaging is a significant area of demand for SBC. Its ease of processability allows it to be used in barrier co-extruded structures. It can be coextruded using adhesive layers with EVOH or PVDC to produce a sheet structure that is readily thermoformed into containers for aseptic food packaging. The high barrier and excellent clarity make it ideal for showcasing packaged products [10]. Oriented polystyrene (OPS) is widely used for many packaging applications, including deli and bakery packaging. The biaxial orientation of the crystal polystyrene provides a significant improvement in impact strength and makes a very rigid material. Even with the enhancement of impact resistance, usually some level of impact modifier is required. SBCs are an excellent modifier for OPS. They help maintain the excellent clarity while also improving impact and processability of the OPS. A living hinge can be also be integrally designed into thermoformed packaging. A very significant application for SBC in Europe is clear egg cartons. The clear egg cartons are thermoformed with a living hinge to make a bottom protective container for the egg, and a clear lid to display the egg. Perforations can also be added to the carton during thermoforming which enables the consumer to break the conventional dozens egg carton into a six pack of eggs. Similarly, side-by side co-extrusion also makes use of this ability. In side-by-side extrusion, the sheet is extruded with clear and opaque bands running in the machine direction. The clear bands are typically based on SBC and crystal polystyrene blends and the opaque bands may be high-impact polystyrene. By forming a hinge along the knit line of the clear and opaque bands, a container can be produced with an opaque base and clear top. Some of the first applications for these side-by-side co-extruded trays were for airline meal containers. Now they are expanding into many other areas of food packaging. SBC can be processed on most conventional blow molding and injection blow molding equipment designed for many other types of polymers with little or no equipment modifications. This includes continuous, accumulator head, or reciprocating screw extrusion blow molding equipment. Bottles made from SBC have been used widely for applications such as honey bears. The SBC will withstand the moderate hot fill temperatures required for this application. More predominant in the Asian markets is an attempt to broaden SBC usage in many applications that have been dominated in the past by PVC. SBC blends are now being used in both blister packaging and in integrated circuit packaging tubes. Generally, SBCs are not drop-in replacements for PVC. The SBCs do not have the rigidity of PVC and, when blended with crystal polystyrene, may not have adequate impact strength. However, by incorporating blends of other polymers such as styrene-butadiene elastomers, high-impact polystyrene and acrylate copolymers, sheet or profiles of good impact, clarity and rigidity can be produced. Also in Japan, ice cream cups made out of SBC are very

510

D. L. HARTSOCK AND N. E. STACY

popular. By selecting grades of SBC with excellent cold temperature impact and using a combination of polymers, clear cups can be tailored to deliver a clear package with both toughness and cold temperature resistance. SBC can also be molded into overcaps. These are injection molded clear caps that are often used on bottles for food and cosmetic packaging. Overcaps are typically made by injection molding in multicavity molds and designed for a snap fit. They must have high gloss and clarity and can often have a transparent color. Some packaging applications have been developed taking advantage of the ability of these polymers to incorporate a living hinge. With a proper design, highly functional hinges can be molded into the package (Figure 22.2). Typical applications include small boxes for office supplies (Figure 22.3), hardware and sporting goods. More durable applications such as large hinged boxes for toolboxes have been out of the range of capability for these polymers, and are typically the domain for polyolefins, which can have much greater hinge lives. More elastomeric grades of SBC have been produced that do have higher hinge life, and the hinge life of SBC can be extended further by polymer modification with high butadiene content elastomers (Table 22.2). By increasing the hinge life, these polymers have been used for more demanding applications such as video cassette cases [11]. Antitheft packaging for audio- and videotapes come in a wide variety of choices. SBCs are used in both blow molding and injection molded cases to make a tough and clear protective case to both showcase and protect these products. There are some limitations on SBC, especially in blends with polystyrene. These include poor stress crack resistance and odor concerns for sensitive Fill Perpendicular to Hinge

Hinge Open 0.030" to 0.60"

Hinge Closed

Figure 22.2

Hinge design for styrene-butadiene copolymers

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

Figure 22.3

511

Double-hinged boxes made from KR01 grade of K-Resin® SBC

foods. One way to overcome these deficiencies is by adding a thin co-extruded layer of a copolyester (PETG) to the sheet. The PETG adhesion to SBC is good, but it can be delaminated from the underlying sheet. However, once the sheet is thermoformed, it is very difficult to separate the PETG from the sheet. Generally adhesion is good enough that no additional adhesive is required for most applications. The protective layer of PETG also provides excellent stress crack resistance, protecting the substrate from the stress crack medium. The PETG is a good odor barrier, restricting the release of styrene monomer. Generally comingling materials in regrind can be very problematic, especially if clarity is to be maintained. However, in this case the PETG can be re-used in the SBC substrate with relatively little adverse effect. As long as the total PETG content of the regrind is less than 15%, there is little loss of clarity or other properties [12,13].

4.4

GARMENT HANGERS

Garment hangers have traditionally been made of many different materials, including wood, metal and polymers. Until recently, most of the polymer hangers had been made of polystyrene and polypropylene. A cooperative effort among garment industry stakeholders has led to greater standardization of hanger performance requirements for garment hangers. This cooperative effort by hanger manufacturers, garment manufacturers and retailers has resulted in an industry program which uses a high-performance, transparent hanger for both shipping and displaying many types of garments in retail stores.

Table 22.2 Hinge life data for K-Resin® SBC blendsa 100% KR03 75% KR03 50% KR03 25% KR03 100% KR01 100% KR01 95% KR01 90% KR01 95% K.R03 90% KR03 95% KR03 90% KR03 80% KR03 85% KR03 100% KK38 25% KR01 50% KR01 75% KR01 5% KRATON 10% KRATON 5% KRATON 10% KRATON 5% SMMA 10% SMMA 20% SMMA 5% KRATON 10% SMMA Average flexes until break 641 819 837 745 923 750 1398 3416 1380 1718 289 198 104 367 2896 Standard deviation 169 203 219 140 161 157 267 1051 522 413 51 39 22 146 638 " KRATON

Shell Kraton Dl 102 rubber; SMMA = Nova NAS 21 styrene methy methacrylate copolymer.

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

513

Performance standards for clear garment hangers have been developed for this program in North America and have also been developed for similar programs in other regions of the world. Significant benefits are derived by hanging the garment on a clear hanger at the point of manufacture, shipping on that hanger, and ultimately selling the garment to the consumer at the retail store on that same hanger (Figure 22.4). This eliminates the cost of re-hanging the garments on clear display hangers and also expediting the movement of products to market by reducing the cycle time of getting the garments from the shipping containers to the sales floor. This process puts a number of demands on the hanger. It must have adequate strength to withstand the rigors of shipping, and be strong enough to carry several times its intended load while in transport. It must also withstand exposure to high temperatures while the garments are in storage in warehouses, in transit from the garment manufacturers to retail stores, and when passed through steam tunnels designed to take wrinkles out of the garments. While some of the hangers are 100% SBC, many are blends with crystal polystyrene. Addition of crystal polystyrene enhances the temperature resistance, stiffness, and overall economics of the hanger. Another feature that has made clear SBC hangers the preferred material is its hinge characteristics. Pants and many other clothing accessories are often hung from hangers which use hinged retainer clips to hold the clothing item. The SBC hinge characteristics allow the hanger clip to be molded with an integral hinge to hold the garment securely for shipping and display while not damaging the merchandise.

Figure 22.4

Clear garment hanger molded from SBC

514 4.5

D. L. HARTSOCK AND N. E. STACY FLEXIBLE PACKAGING

SBC film is growing into more and more film applications. It is proving to be well suited for fresh cut produce packaging, candy twist wrap, shrink films, flexible medical and decorative films. SBC can be extruded into either cast or blown film for use in flexible packaging applications. While some applications employ a monolayer SBC film, many films are co-extruded with SBC along with several layers of other materials in order to bring a combination of properties to the final application. Depending on the grade of SBC selected, some film properties can vary significantly. Figure 22.5 shows some of the typical range of properties for SBC film. DK13 is a more elastomeric film than the DK11 grade. SBC film grades have high clarity and gloss, high moisture and gas transmission rates, good stiffness and puncture resistance, excellent crease retention or 'deadfold' and formability. They have the potential to replace some traditional cellophane markets. One example of an SBC film application is fresh cut produce packaging. This MAP (modified atmosphere package) application of SBC-based film structures allows a bag to be designed to extend product freshness for over 2 weeks, essentially doubling the shelf life of produce packaged in a standard lowdensity polyethylene bag. Key SBC performance properties that make this application successful are the oxygen transmission rates, clarity, gloss, printability and stiffness. These bags may also have a co-extruded heat seal layer to facilitate sealing of the bag [14,15]. Melt flow OTR

10

Secant Modulus, MD 200

MVTR

/

"V H \7

\

Secant Modulus. TD

1.4I

ElmendorfTear, TD

V

120 Elmendorf Tear, MD

X/\

X

s—^fe w.

14

Puncture

/

Elongation, MD

Elongation, TD

Tx

350

Dart Drop

DK11 — DK13

Figure 22.5

Film properties of K-Resins SBC grades (DK11 and DK13) 1 mil film

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

515

Applications such as candy twist wrap also take advantage of special characteristics of SBC film. SBC films have excellent deadfold or crease retention properties. That is, when twisted or folded, the film retains the twist and stays wrapped rather than opening. On candy packaging lines this can also translate to packaging candy at higher rates. These high-gloss films can be clear, tinted or opaque and highly decorated by both printing and vacuum metallizing. The SBC can be used without modification, or it can be blended or coextruded to optimize performance and economics [16,17]. SBC films can also be used for shrink film applications. A shrink film is produced with either uniaxial or biaxial orientation, and applied over a part to be packaged. Application of heat to the film allows it to shrink around the part and provide either a functional and/or decorative package. Films made on conventional blown film line or cast film line can be made with various levels of orientation. Blown film can be made with orientation in both the machine and transverse direction, which can be used for shrink-wrap. Cast film can be made with high shrinkage in the machine direction. Additionally, SBC-based films can be oriented in a number of other ways, such as double bubble lines. SBC films are also used for shrink labels, have unique shrink properties and can be oriented to have a high degree of shrinkage. A tentering frame is commonly used to set the exact level of orientation in the desired direction. The shrinkage characteristics may be modified as needed by blending in crystal polystyrene, elastomers or styrenic copolymers. Flexible medical packaging is well suited to SBC for use in form, fill and seal machines. SBC have excellent formability, which allows for nearly perfect replication and filling of molds. To make a peelable seal, the SBC would need to be co-extruded with a lower melting substrate to act as the adhesive layer. The SBC structure provides good web formability and good toughness. It then has the versatility to be sterilized by "y-radiation, ethylene oxide or electron beam. Decorative films are a natural application for SBC. These films can be used for gift wrap, flower wrap or food wrap. They require high gloss, transparency, stiffness, deadfold and transparent color. The economics and stiffness of the film can also be easily modified by blending with crystal polystyrene [18]. Highly decorated SBC films are also used as decorative laminated films on foamed polystyrene trays and packages. The unsaturated butadiene in the polymer does leave it susceptible to crosslinking and forming gels. Gels may not only be esthetically undesirable, but also cause holes in the film. For this reason, SBC film grades are generally highly stabilized to minimize gel levels. 4.6

MEDICAL

With the proliferation of concern regarding blood-borne pathogens and the expanding need for medical devices and packages which enable defects in

516

D. L. HARTSOCK AND N. E. STACY

mechanical function to be detected easily, demand for SBC continues to grow in medical applications. SBC has the ability to mimic higher priced engineering resin properties at lower cost with excellent clarity, impact resistance, radiation sterilization and biological compatibility. Some of the first applications of SBC in medical applications were for drainage units. The units are clear to allow ease of reading fluid levels, and breakage resistant to contain fluids if the parts are dropped or impacted. Applications have expanded in other medical devices and diagnostic equipment, including yankaeurs, centrifuge tubes and medical films. Other application areas include safety needles and various respiratory care devices. SBCs have been used for many sensitive applications including direct blood contact. These types of polymers have been subjected a broad array of medical tests, including USP, blood contact and many other screening tests. The results have all been very favorable, as shown in Table 22.3. Most medical devices require sterilization before packaging and use. The most common methods for doing this include y-irradiation, exposure to ethylene oxide (EtO) gas or electron beam (E-beam). Contact with radiation, ethylene oxide or electron beam can affect not only the microorganisms of concern, but also potentially the medical device or package. The physical properties were tested after irradiation to define any deleterious effect Table 22.3

Medical Testing Results Conclusion/observation

Test

KR03

KR01

Non-toxic Non-toxic

Non-toxic Non-toxic

Non-irritant Non-irritant Meets USP requirements

Non-irritant

USP Acute systemic toxicity (USP) Intracutaneous toxicity (USP) Implantation tests (USP) - 5 day - 90 day Physico-chemical (USP) Hemolysis - extract - direct contact Methemoglobin formation Screening tests Cytotoxicity - MEM elution - agar overlay Ames mutagenicity assay Ocular irritation Sensitization

Non-hemolytic Non-hemolytic Negative Non-toxic Non-toxic Non-mutagenic Non-irritant Not a sensitizer

Blood compatibility

517

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

compared with the results for the control. The physical properties were also retested after 1 and 2 years to assess any long-term effects. Although the level of y-irradiation required for guaranteed sterilization is relatively low at 2.5 Mrad (recommended by USP), the influence of higher dosages such as might occur on repeated exposures or a single higher exposure level was considered essential information. The SBCs were exposed to y-irradiation at dosage levels of 2.5, 5.0 and 7.5 Mrad. The results of these tests, presented in Table 22.4 show that at an exposure level of 2.5 Mrad, with the exception of a small decrease in flow rate and small increase in yellowness, there was no loss in physical or optical properties with the SBC or SBC blends with crystal polystyrene. With increasing dosage levels, the only noticeable changes in physical and optical properties were a continuing decrease in flow rate, slight increase in stiffness in conjunction with a slight decrease in elongation and continuing increase in yellowness. In summary, Table 22.4

Gamma irradiation sterilization results

a

Targeted radiation level Actual radiation levela Flow rate (g/l0min) Tensile strength @ Yield (psi) @ Break (psi) Elongation @ Yield (%) @ Break (%) Tensile modulus (psi) Flexural yield (psi) Flexural modulus (psi) 1/8 in Izod impact Notched (ft-lb/in) Unnotched (ft-lb/in) Dynatup impact (in-lb) Failure mode Vicat softening point (°F) Heat deflection temperature @ 264 psi (°F) Shore D hardness Light transmission % Haze Hunter color a Hunter color b a

60/40 KR03/GPPS

KR03

Test

0.0 0.0 8.4

2.5 3.5 4.7

5.0 6.0 1.9

7.5 10.0 0.5

0.0 0.0 10.7

2.5 3.5 8.4

5.0 6.0 6.0

7.5 10.0 2.5

4150 2850

4250 2950

4250 2900

4250 2700

5350 3550

5450 3600

5500 3650

5400 3750

2.0 230 291,000 5900 238,000

2.0 240 305,000 6000 241,000

2.0 230 307,000 6000 242,000

1.8 200 427,000 6000 243,000

1.6 16 438,000 8100 339,000

1.7 18 631,000 8150 340,000

1.8 18 499,000 8250 342,000

1.6 13 703,000 8250 344,000

0.4 NB 254 Ductile 183 142

0.4 NB 233 Ductile 193 144

0.4 NB 236 Ductile 183 145

0.4 NB 325 Ductile 185 140

0.3 2.2 17 Brittle 199 163

0.3 2.2 12 Brittle 199 162

0.3 2.3 19 Brittle 199 163

0.3 2.1 16 Brittle 199 163

67 93 6.4 -1.2 3.1

67 93 5.9 -1.7 4.5

67 92 5.9 -2.1 5.7

67 93 6.0 -2.7 7.2

76 87 4.0 -0.6 5.9

76 87 4.0 -1.5 8.0

76 86 4.4 -2.0 9.4

76 87 3.6 -2.7 11.1

All radiation values are in Mrad.

518

D. L. HARTSOCK AND N. E. STACY

y-irradiation up to a 10.0 Mrad exposure level has little effect on most physical, mechanical and optical properties of SBC or SBC and GPPS blends, but can cause yellow color development and a decrease in melt flow. Testing also confirmed that both sterilization by EtO and by E-beam also had little effect on SBC polymer properties [19]. SBC may be the resin of choice for many medical applications, but there are some applications in which SBCs are certainly not suitable and those applications should be avoided. Some applications demand a greater chemical and/or stress crack resistance than SBC can withstand. For example, many medical devices have fittings which attach to flexible PVC tubing. Plasticizers used in PVC tubing can migrate into SBC parts in which they are in direct contact, causing a wide range of effects, including stress whitening, swelling and complete dissolution of the part. The magnitude of the incompatibility may depend on many factors, including type and amount of PVC plasticizer, conditions of use, length of storage, assembly methods, sterilization, part stresses, etc. Better results may be achieved with flexible tubing containing lower levels of plasticizer. Some more recent environmental and political pressures concerning phthalates leaching from PVC molded products may provide some newer opportunities for SBC in medical applications. Even though scientific opinion remains mixed as to whether phthalates pose major health concerns, the industry is keenly interested in alternatives. Several manufacturers of health care products have already launched programs aimed at developing alternative materials to eliminate molded and extruded PVC products. SBC are generally attacked by many hydrocarbon solutions. Degreasing of parts in many solvents can result in stress whitening, cracking or premature failure of SBC parts. Some bodily fluids contain high levels of lipids and are considered SBC stress cracking agents. Care should be taken any time SBC parts are exposed to any stress crack medium to ensure suitability for use.

4.7

CONSUMER GOODS

There are a wide range of applications that are made from SBC that may show up in the ordinary household. Some SBCs are used in appliances, housewares and office supplies. Owing to limitations of temperature resistance, chemical resistance, and abrasion resistance, the SBC may be blended with polymers such as styrene-methyl methacrylate (SMMA) to improve functionality. SBCs may also be seen in parts such as aquarium filter tanks, sewing machine dust covers, bird feeders, cake pan lids, pens, vacuum cleaners and computer accessories.

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS 4.8

519

TOYS

Toys have been a long-time SBC application (Figure 22.6). Keeping childresafe from shards and splinters of broken pieces has been one reasons why SBCs continue to be used when designers are selecting polymers for their toys. Many grades of SBC are flexible enough to be deformed on impact rather than break and shatter. From board game covers to toy truck windshields, these polymers are used in nearly every conceivable toy application. They can also be easily colored, and fluorescent colors have been popular with toys. SBCs have also been used in toy applications that do not require clarity. Because SBCs can be easily colored, are durable and have very high gloss, they are used for non-clear components. Although often used in the neat form, SBC can also be blended with other polymers such as high-impact polystyrene. 4.9

DISPLAYS

Some grades of SBC are ideal for display applications. To showcase product, shelves or holders for displays must have outstanding clarity and also be durable enough to withstand abuse. Many of the display parts are large, and must be molded with minimal warpage. There are several SBC grades that provide this combination of properties. Examples of SBC display applications include parts to display greeting cards, cigarettes, candy, etc. The hinge characteristics of SBC allow a flat display to be made that can be folded into place for quick set-up. Some displays are also made through the blow molding process. SBC grades with good melt strength facilitate production of either clear, tinted or opaque large blow molded parts.

Figure 22.6

Clear SBC grades are widely used in toy applications

520

D. L. HARTSOCK AND N. E. STACY

5 SBC BLENDS While it is very common for SBC to be used as a stand-alone material for many parts, as referenced many times in this chapter, these polymers are also frequently blended with other polymers. The capability of SBC to blend is in large part due to their butadiene content and morphology. The combination of alternative polymers with SBC can result in significant performance or economic enhancement. The ability to blend allows the converter to maximize the perfect balance of mechanical properties, performance and economics. SBC may be the minor or major components in blends. Some of the blends have major commercial significance, while other blends are very interesting but of little commercial significance. Often the combinations consist of more than two polymers. In the single service market, it is not unusual to have a cup based on a blend of SBC and GPPS with possible addition of HIPS and SBS to improve both economics and impact properties. Most of the blends are clear, but the requirement for clarity does limit the type and amount of various polymers that can be employed. A close match of refractive index increases the possibility that two materials may be able to be blended and maintain good optical clarity. Other polymers are not as close in refractive index, but are miscible so that clarity can be still achieved in blends of the two components. 5.7 CLEAR BLENDS 5.1.1

Crystal Polystyrene

The polymer most commonly blended with SBC is crystal polystyrene. Crystal polystyrene and SBC do have a significant difference in refractive index, but the polystyrene is miscible in the polystyrene domains of the SBC. Hence the blended part, if well mixed, will have good optical properties. Crystal polystyrene is desirable as a blend resin for SBC because it is of lower cost and also offers advantages in temperature resistance, stiffness and surface hardness. The major disadvantage in blends of SBC with crystal polystyrene is a significant decrease in impact strength as the polystyrene content is increased. In injection molding blends, SBCs are used as both major and minor components, depending on the application. Garment hangers typically have a high percentage of SBC to give adequate impact resistance and hinge life. Often, low levels of SBC are added to a crystal polystyrene part to provide just enough toughness to survive mold ejection and shipping. In thermoformed cups, the SBC content can vary significantly. The SBC content is often close to 50 %, but in many cases there is more crystal polystyrene than SBC in the final part. Hence selection of an optimum crystal

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

521

polystyrene grade is critical. One of the primary factors that influence the selection of the polystyrene is odor. The key to odor in the blends lies in the residual volatile components of the virgin resins. Crystal polystyrene contains unreacted styrene monomer as well as impurities in the styrene monomer feedstream such as ethylbenzene and propylbenzenes. It is common for crystal polystyrene to contain between 100 and 800 ppm of styrene monomer as the primary volatile component. Ethylbenzene will show up at low levels in most polystyrenes, but if ethylbenzene is used as a diluent in the process it may be present at higher levels. SBC, by contrast, are usually extremely low in residual styrene monomer. By themselves, neither the crystal polystyrene nor the SBC components have a strong styrenic odor with these levels of styrene monomer. However, the combination of SBC and crystal polystyrene into blends can result in higher odor than either of the original components. SBC that contained as little as 100 ppm of styrene monomer would have a strong enough odor to be considered unsuitable for sale in a food contact application. Odor in blends is minimized by the selection of polystyrene with low residual volatiles. Polystyrene producers have been reducing the levels of styrene monomer in polystyrene recently, and many commercial grades are available that have < 500 ppm styrene monomer [19]. Some packaging applications may be particularly sensitive to styrene or other volatile components. SBC and crystal polystyrene blends, even with low residual monomer levels in the polystyrene, may be unacceptable for these applications. In this case one possible solution may be to co-extrude a thin food contact layer on the surface of the blended sheet. Studies have indicated that by co-extruding a thin layer of crystal polystyrene on the surface of an SBC and crystal polystyrene blend, the styrenic volatiles and the odor can be significantly reduced. However, the downside is that the added layer of crystal polystyrene can serve to make the blended part more brittle. Another option that could reduce odor, but not have a deleterious effect on impact, is a thin coextruded layer of PETG [20]. The molecular weight of crystal polystyrene can also play an important role in blend properties. There is a trade-off between impact and clarity, with higher molecular weight giving better impact and lower molecular weight providing better clarity. The converter can select the proper crystal polystyrene to match specific performance requirements.

5.1.2

Styrene-Methyl Methacrylate (SMMA)

Another polymer that has been very successful in blends with SBC is polystyrene–methyl methacrylate (SMMA), a styrene acrylic copolymer. Compared with SBC, SMMA offers significantly higher stiffness, surface hardness and temperature resistance. The refractive index of these copolymers can be

522

D. L. HARTSOCK AND N. E. STACY

designed to match the refractive index of a given SBC, resulting in blends that have excellent clarity. When blended for injection molding applications, any level of SBC and SMMA can be blended to achieve the performance requirements of the application. In addition, the blending of SMMA with SBC can result in significantly reducing warpage, to facilitate making large flat parts. When certain SBC grades are modified with low levels of SMMA (20–30%), such blends may retain the high impact strength and elongation normally associated with SBC, while benefiting from increased stiffness, surface hardness and heat resistance contributed by SMMA. Mold release characteristics may also be enhanced, as shown in Tables 22.5 and 22.6 [21]. SMMA blends can also be used in extruded sheet and profile extrusion. Depending on the styrene-acrylic copolymer, its blend level and the processing conditions, the clarity of extruded parts may not be as good as the clarity of injection molded parts. Table 22.5 Physical properties of blends of K-ResinR SBC and styrene-methyl methacrylate copolymera Property

80% KR03NW 70% KR03NW 100% KR01 100% KR03NW 20% SMMA 30% SMMA 100% SMMA

Flow rate (g/10 min) 7.7 Flexural modulus 0.212 (Msi) Flexural yield 600 strength (psi) Tensile yield strength 4200 (psi) Break elongation (%) 17 Izod impact (t-lb/in) Notched 0.2 Unnotched 3.5 Gardner impact 7.1 (cm/kg) Hardness, Shore D 72 Vicat softening 200 point (°F) Heat deflection temperature (°F) 66 psi 186 264 psi 167 Hunter color b 2.8 Light transmission 90 Haze (%)

4.9

8.1 0.211

6.6 0.255

5.8 0.262

2.6 0.492

400

580

610

152

3250

3700

3900

9800

132

64

33

4

1.3 No break No break

0.5 No break No break

0.5 No break No break

0.3 4.1 3.1

63 187

70 199

68 202

82 216

184 161 3.8 89

190 162 3.9 89

191 160 3.8 89

206 163 2.0 90

7.3

5.6

4.6

3.0

Injection molded test specimens. SMMA = Nova NAS 21 styrene-methyl methacrylate copolymer.

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

523

Table 22.6 Warpage (in) of blends of K-Resin® SBC and styrene-methyl methacrylate copolymera 80% KR01 80% KR03NW 70% KR03NW Sample weight (g) 100% KR01 100% KRO3NW 20% KR03NW 20% SMMA 30% SMMA

180 182 184 186

0.338 0.329 0.271 0.253

0.579 0.566 0.523 0.510

0.201 0.174 0.104 0.082

0.254 0.232 0.216 0.178

0.166 0.110 0.044 0.042

a

Injection molded 11 /in x 14/in x 70/mil plaques. SMMA = Nova NAS 21 styrene-methyl methacrylate copolymer.

New grades of SMMA are being developed that have shown tremendous potential for blends in SBC. These new grades result in very high impact in the blends with higher loadings of SMMA. Notched Izod impact values of almost 5 ft-lb/in have been reported for injection molded test specimens containing 50 % SMMA [22].

5.1.3

Styrene-Acrylonitrile (SAN)

Another polymer that blends well with SBC to maintain clarity and good physical properties, although not to the same extent as SMMA, is styreneacrylonitrile copolymer (SAN). The properties of injection molded specimens listed in Table 22.7 demonstrate that blends containing a higher percentage of SBC display superior impact properties, while the higher percentage SAN blends possess greater surface hardness, heat resistance and stiffness. To minimize haze in these blends, the acrylonitrile content should be less than 22 %. As an added benefit, as little as 15–20% SAN in the blend can usually eliminate part warpage. Mold release characteristics are also improved. Performance can vary with different grades of SAN.

5.1.4

Thermoplastic Elastomers (SBS)

The addition of an elastomer (typically a high butadiene content SBS) to SBC will serve to enhance further the elastomeric properties of the SBC. One key feature mentioned previously was improvement in hinge life properties. SBS copolymers can also be added to thermoformed sheet in blends of SBC and crystal polystyrene. The SBS does cause some loss of clarity, but gives more impact resistance to the sheet. Selection of the proper SBS can result in minimal loss of clarity, typically at 3–10% loadings. Styrene-isoprene copolymers (SIS) have also been tested with SBC and can give similar results in impact property improvement.

Table 22.7 Properties

Physical properties of blends of K- ResinR SBC and styrene-acrylonitrile copolymera 100% KR03 90% KR03 80% KR03 70% KR03 60% KR03 50% KR03 40% KR03 30% KR03 20% KR03 10% KR03 10% SAN 20% SAN 30% SAN 40% SAN 50% SAN 60% SAN 70% SAN 80% SAN 90% SAN 100% SAN

Density (g/c3) 1.01 Melt flow(g/10min) 8.0 Flexural modulus (Msi) 0.235 Tensile yield strength (psi) 3950 Break elongation (%) 100 Izod impact (ft-lb/in) 0.44 Notched Unnotched No break Gardner impact (cm/kg) 132 67 Hardness, Shore D Vicat softening point ("F) 196 Heat deflection temperature (°F) 182 66 psi 163 264 psi Light transmission (%) 91

1.02

1.02

1.03

1.04

1.05

1.06

1.06

1.06

1.06

1.05

8.8

8.5

9.2

9.0

8.1

6.4

4.7

3.9

2.8

3.0

0.263 4200

0.282 4475

0.305 5025

0.334 5975

0.348 6700

0.366 6750

0.389 7900

0.423 8550

0.467 8975

0.480 9300

44

28

12

7

5

7

4

4

4

4

0.30 No break

0.32 7.81

0.39 4.19

0.20 2.60

0.12 2.50

0.14 2.40

118 70 202

13 71 205

0.33 3.44 3.50

0.33 3.70

144 68 196

0.55 3.82 3.35

72 210

73 212

— 80 —

— 82 —

— 84 —

— 86 —

0.14 2.90 3.50

195 176 89

194 173 88

198 174 88

198 180 87

202 183 87

210 189 —

211 192 —

212 193 —

214 204 —

" Injection molded test specimens. SAN = Monsanto 31-1000 styrene-acrylonitrile copolymer.

84 219 220 202 87

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

5.2 5.2.1

525

OPAQUE BLENDS Polypropylene

At first glance, the likelihood of blending SBC and polypropylene together and obtaining a useful product seems unlikely, owing to the different chemical natures of the polymers. However, blends of SBC and polypropylene can produce attractive pearlescent or marblescent injection molded parts with high gloss, good stiffness and surface hardness, and excellent hinge life, as seen in Table 22.8. Furthermore, a pronounced synergistic increase in notched Izod impact strength is evident for blends with high SBC content (75090 %). Blends containing polypropylene copolymer tend to exhibit higher notched Izod values than those of polypropylene homopolymer of comparable melt flow. As a rule, the lower the melt flow of the polypropylene, the more distinct is the marbled appearance of the molded part, suggesting less efficient mixing of the incompatible blend components. Precompounding, or even thorough mixing of a dry blend before molding, will decrease the marbling effect, and result in a pearlescent part with high gloss. For marblescent parts, the blend should be molded at the lowest practical temperature using a 100–300 psi backpressure, with a standard sprue and large gate instead of pin or high shear gates since restrictive gates will minimize or eliminate the marbling effect. Owing to the evident incompatibility of the two resins, however, caution must be exercised during processing to avoid delamination of the molded part. For injection blow molded applications, blends of SBC and polypropylene also exhibit an attractive combination of performance and pearlescent appearance. Blends containing 10–40% of SBC exhibit excellent bottle-drop impact (10–11ft) but retain a moisture vapor transmission rate similar to polypropylene. Altogether these properties demonstrate that the esthetically pleasing parts made from blends of SBC and polypropylene can withstand a considerable degree of abuse and contain aqueous solutions throughout extended shelf life. Market applications have included applications such as flowerpots, cosmetic containers and shoes.

5.2.2

High-impact Polystyrene (HIPS)

High-impact polystyrene is an economical, opaque resin that offers moderate impact strength, rigidity and good processability. Two shortcomings of typical HIPS, however, are inherent low elongation and poor gloss. Opaque parts produced from blends of SBC and HIPS possess both higher gloss and elongation than HIPS alone, with similar tensile and heat-resistance properties. In addition, injection molded blends containing 30–70% SBC display notched Izod impact strengths well in excess of those of the individual polymers, as displayed in Table 22.9. The synergistic improvement of impact strength is even greater for extruded sheet samples. Blends of HIPS and SBC have been used in

U1

Table 22.8

Physical properties of blends of K-ResinK SBC and polypropylene"

Property

100%PP

90% PP 10% KR03

75% PP 25%, KR03

50% PP 50% KR03

25% PP 75% KR03

10%. PP 90% KR03

100% KR03

Flexural modulus (Msi) Tensile yield strength (psi) Tensile break strength (psi) Break elongation (%) Izod impact (ft-lb/in) Notched Vicat softening point ( F)

0.254 6000 700 56

0.223 5600 4100 534

0.209 5200 1800 463

0.178 4500 4000 354

0.164 3600 4900 388

0.166 3500 4000 274

0.190 3800 3400 241

0.3 312

0.7 307

0.6 296

0.7 219

9.8 195

0.5 201

0.3 180

" Injection molded test specimens. PP = Marlex polypropylene homopolymer (nominal 5 melt flow).

Table 22.9

Physical properties of blends of K-Resin H SBC and high-impact polystyrene0

Property

90% KR03 80% KR03 70% KR03 60% KR03 50% KR03 30% KR03 100% KR03 10% HIPS 20%, HIPS 30% HIPS 40% HIPS 50% HIPS 70% HIPS 100% HIPS

0.233 Flexural modulus (MSi) Tensile break strength (psi) 3800 Break elongation (%) 90 Izod impact (ft-lb/in) Notched 0.36 Unnotched No break Gardner impact (cm kg) >200 Hardness, Shore D 65 Heat deflection temperature ( F) 174 97.7 Gloss (%)

0.236 3750 110

0.238 3900 87

0.245 3900 62

0.249 3900 66

0.265 3900 47

0.272 3800 46

0.288 3800 35

0.25 No break >200 66 169 85.7

0.57 No break >200 68 170 75.7

1.0 No break >200 69 170 68.4

2.4 No break >200 72 171 68.4

2.2 No break >200 74 175 59.3

1.8 No break >200 77 175 47.5

1.6 24.2 182 78 178 24.8

Injection molded test specimens.

01 N> -si

528

D. L. HARTSOCK AND N. E. STACY

housewares and toys, competing with ABS. Blends of SBC and HIPS produce parts with good impact and gloss, and are easier to color and mold than ABS. HIPS is also used as a very effective impact modifier at very low levels for clear blended sheets of crystal polystyrene and SBC. Because the crystal polystyrene component is less expensive than SBC in the blends, the economic driver is to add more crystal polystyrene to the blend. Test results show a substantial increase in impact for sheet samples to which high impact polystyrene was added. At the lowest level of HIPS addition, 0.5%, a significant improvement in sheet strength is achieved. However, the sheet strength values did not increase much as higher levels of HIPS were added. This is highlighted in Table 22.10. Haze values increased significantly as higher levels of HIPS were added. Sheet clarity remained good when 0.5% HIPS was added to the blend. HIPS also demonstrated improving antiblocking characteristics of the blended sheet. This occurred at levels as low as 0.5% HIPS [23].

5.3

OTHERS

SBC blends have been evaluated with other polymer systems, including PPO, polycarbonate, PVC, PETG and PET. Although these have been the subjects of patents and have resulted in some small commercial applications, wide-scale commercial success has eluded these blends. Table 22.10 Physical properties 50:50 K-Resin* SBC—GPPS 20 mil blended sheet modified with Dow XL 8028 HIPS HIPS added (%) Property

0

0.5

1.0

2.0

Haze (%) Light transmission (%) Dart drop (g/26 in) Tensile yield (psi) MD TD Tensile Break (psi) MD TD Elongation (%) MD TD Oil bath shrinkage (%) MD TD Hunter color

2.6 89.3 232

4.4 89.1 506

11.4 87.4 520

11.5 87.9 435

5000 2850

4400 2600

3700 2550

4150 2850

4050 2650

3250 2500

3100 2550

3000 2550

7.0 7.0

14.0 6.0

9.0 2.0

7.7 3.7

26.5 -2.3 2.1

29.1 -2.9 2.0

28.2 -2.1 1.9

27.6 -2.0 1.7

HIGH STYRENE CONTENT STYRENE-BUTADIENE COPOLYMERS

6

529

FUTURE APPLICATIONS

Application diversification of SBCs continues to proliferate as new grades meet new innovative ideas. New higher butadiene content SBCs are being developed that could find additional opportunities in toys and film packaging. Improved blends of SBC with other materials are also showing great promise. Newly developed grades of styrene-methyl methacrylate (SMMA) copolymers are showing excellent properties in blends with SBC. Izod impact strengths of up to 5ft-lb/in have been demonstrated. These products hold the potential for competing with polycarbonate and acrylics more extensively in housewares, medical products, displays and appliances. Greater interest is being shown in styrenic shrink labels and more film manufacturing capacity is being added to develop those markets more fully. Poor UV stability has hampered growth in outdoor exposure applications. Development is under way to identify stabilization systems and structures that will have higher resistance to deterioration caused by outdoor exposure. Successes in any of these areas, as well as continued growth in current markets, should keep the SBC growth rate high in coming years.

REFERENCES 1. H.L. Hsieh, R.P. Quirk, Anionic Polymerization: Principles and Practical Applications, Plastics Engineering Series, D. E. Hudgin, Founding Editor, Marcel Dekker, New York, Chapter 15 (1996). 2. G.S. Whitby, C.C. Davis, R.F. Dunbrook, Synthetic Rubber, John Wiley & Sons Inc., New York, and Chapman & Hall, London (1954). 3. A. Noshay, J. E. McGrath, Block Copolymers: Overview and Critical Survey, Academic Press, London (1977). 4. F.E. Mathews and E.H. Strange, Br. Patent 24790 (1910). 5. K. Ziegler, Liebigs Ann. Chem., 473, 1 (1929). 6. F. W. Stavely, Ind. Eng. Chem., 48, 778 (1956). 7. M. Szwarc, M. Levy, R. Milkovich, J. Am. Chem. Soc., 78, 2656 (1956). 8. M. Szwarc, Nature (London), 178, 1168 (1956). 9. H. Hsieh, A.V. Tobolsky, J. Polym. Sci., 25, 245 (1957). 10. Coextrusion Technology, November 27–30, 1995, Amsterdam; Course Director C.R. Finch, C40. 11. Hinge Life Study of K-Resin® Blends, Plastics Technical Center Report 421, Chevron Phillips Chemical Co., Bartlesville, OK. 12. 'Deli cups with a clear blend of performance and price', Packaging Dig., January (1994). 13. Coextruded K-Resiri* Styrene—Butadiene Copolymer/PETG Sheet, Technical Service Memorandum 317, Chevron Phillips Chemical Co., Bartlesville, OK. 14. 'Bags extend salad shelf time', Packaging Dig., January (1996). 15. 'Polystyrene blow film starts to get some respect', Plast. TechnoL, November (2000). 16. 'New twist for wrapping candy in Mexico', Packaging Dig., August (1997).

530

D. L. HARTSOCK AND N. E. STACY

17. 'A new twist on film', Converting Mag., May (2000). 18. 'Decorative films served by multilayer structure', Paper Film Foil Converter, June (2000). 19. Medical Applications of K-Resin®, SBC Technical Service Memorandum 292, Chevron Phillips Chemical Co., Bartlesville, OK. 20. Odor in Blends of Polystyrene with K-Resin® SB Copolymers, Plastics Technical Center Report 410, Chevron Phillips Chemical Co., Bartlesville, OK. 21. K-Resin® Styrene-Butadiene Copolymer Blends, Technical Service Memorandum 316, Chevron Phillips Chemical Co., Bartlesville, OK. 22. High Performance Styrenics 0242: NAS® 90/KR03 Blends For Tough, Clear Parts, Technical Bulletin, Nova Chemicals, Calgary. 23. K-Resin®SB Copolymer/Crystal Polystyrene Sheet Property Modification with High Impact Polystyrene, Plastics Technical Center Report 409, Chevron Phillips Chemical Co., Bartlesville, OK.

This page intentionally left blank

23

STEPHEN F. HAHN Corporate Research and Development, The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

The synthesis and characterization of the hydrogenation product of polystyrene (PS), poly(cyclohexylethylene) (ACHE-PEP), have been the subject of persistent research efforts. The earliest published report on this material dates to 1929, with new work appearing steadily up to the present. The primary motivation for saturating the phenyl rings in polystyrene is the potential to improve the stability of polystyrene with respect to oxidative and radiation-induced degradation. It has now also been established that the hydrogenation of polystyrene leads to a substantial increase in the glass transition temperature (Ts). An improved understanding of the relationship between repeat unit structure and physical properties has also provided motivation; the hydrogenation process allows for the direct study of the influence of the new cyclohexyl side substituent in comparison with the phenyl group present in the starting material. At the time of this writing,ACHE-PEPis not an item of commerce. A variety of applications have been proposed, however, that would take advantage of the properties of this material.

2

SYNTHESIS OF POLYCYCLOHEXYLETHYLENE (ACHE-PEP)

Two distinct synthetic strategies have successfully been employed to prepare high molecular weightACHE-PEP.The most extensively utilized route, the hydrogenation Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy f 2003 John Wiley & Sons Ltd

534

S. F. HAHN

of polystyrene, is perhaps the more versatile approach (Scheme 23.1) [1]. The synthesis of the polystyrene precursor can be performed using a variety of wellknown polymerization chemistries such that molecular weight, polydispersity, and tacticity can be independently controlled prior to the hydrogenation step. The difficulties presented by this approach are those associated with performing chemical reactions on high molecular weight polymers using heterogeneous catalysts. It has been possible, however, to obtain virtually complete hydrogenation with little or no chain degradation, and characterization of the resultant ACHE-PEP materials has been reported in considerable detail. The polymerization of vinylcyclohexane (and the closely related vinylcyclohexene) toACHE-PEPhas also been performed (Scheme 23.2). This approach has the advantage of providing polymer in a single synthetic step from monomer, avoiding the catalytic hydrogenation step. The challenges associated with this chemistry are largely due to the availability of the monomer, and in the limited range of control over molecular weight and tacticity of polymers that can be produced by this synthesis.

3 3.1

CATALYTIC HYDROGENATION CATALYSIS AND

CONDITIONS

Staudinger and co-workers reported the synthesis of 'hexahydropolystyrene' by the Ni-catalyzed hydrogenation of polystyrene in 1929 [2,3]. From that time forward, the catalytic hydrogenation of PS has been utilized by numerous other research groups. For the most part, the catalysts used have been metals from

H2, catalyst

Scheme 23.1

Hydrogenation of polystyrene to polycyclohexylethylene

n

Ziegler-Natta catalyst

Scheme 23.2

Polymerization of vinylcyclohexane

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

535

group VIII deposited on various support media; these metals exhibit excellent activity for aromatic ring hydrogenation in nonpolymeric systems [4]. Hydrogenation using various activated forms of Ni predominate [5-16] with later reports focusing on supported Rh [17,18], Pd [19–22], and Pt [23–26] catalysts. The overwhelming majority of the reported PS hydrogenations have been carried out in solution, typically using saturated hydrocarbon solvents such as cyclohexane or decalin. The hydrogen pressure has varied from 500 to almost 4000 psig, at temperatures typically between 150 and 300 °C. The substrate most commonly chosen has been atactic PS prepared by free radical or anionic polymerization; limited work on the hydrogenation of the syndiotactic and isotactic forms has also been reported. The conditions used to achieve high levels of hydrogenation have often led to chain degradation. Some workers have attempted to circumvent this by performing the hydrogenation for longer reaction times at low pressures and temperatures. The issue of molecular weight degradation during hydrogenation has been specifically addressed in one study [20]. This research started with narrow polydispersity, anionically prepared PS substrates to allow for the straightforward detection of chain degradation by gel permeation chromatography (GPC). Degradation was observed while hydrogenating PS in cyclohexane over Pd/BaSO4, but it was found that by the addition of small amounts of the co-solvent THF high degrees of hydrogenation could be achieved without degradation. The use of metal catalysts on porous catalytic supports with a narrow pore size distribution has demonstrated markedly improved catalyst efficiency for PS hydrogenation [1,23–25]. The diffusion of solvated polymer into and out of the porous support allows the polymer better access to catalytic sites, providing for rapid PS hydrogenation with far greater efficiency than conventional catalyst substrates. This innovation allows for hydrogenation at lower catalyst loading levels and higher polymer concentrations than those previously described, using temperatures and hydrogen pressures that do not bring about chain degradation. The catalytic hydrogenation of syndiotactic polystyrene with Ru on carbon [ 18] and with Pd on BaSO4 [21] catalysts has been reported. The catalytic hydrogenation of isotactic PS with Ru on carbon has also been performed [18]. The limited solubility of both the starting material and product in solvents that will allow the hydrogenation chemistry required that this hydrogenation be run at high dilution (2.5%, w/v). Despite these challenges, high levels of hydrogenation have been achieved. The stereoregularACHE-PEPmaterials are highly crystalline materials; differences in tacticity are apparent from physical property and spectroscopic analysis. Figure 23.1 shows the high field region of the 13C NMR spectra of atactic ACHE-PEP, isotacticACHE-PEP,and syndiotacticACHE-PEPalong with isotactic ACHE-PEP prepared by the Ziegler–Natta polymerization of vinylcyclohexane [18]. A few reports on the hydrogenation of polystyrenes with heteroatomcontaining substituents have been made. The hydrogenation of poly(styreneco-methyl acrylate) was performed using a silica/alumina-supported Ni catalyst

536

S. F. HAHN

/""

r

45

40

35

30

25

20 45

40

35

(ppm)

(a)

30

uj V 45

40

^__/V

X

35

30 (c)

20

(ppm)

K

f

"— "\ n-—'

25

(b)

[ u

JJ

. 25

'-

~J

^^-w^A-W*Ni*»»«*i>« 20 45

(ppm)

40

35

30 (d)

25

20

(ppm)

Figure 23.1 13C NMR spectra of (a) syndiotacticACHE-PEPfrom hydrogenation of sPS, (b) atacticACHE-PEPfrom hydrogenation of PS, (c) isotacticACHE-PEPfrom Ziegler—Natta polymerization of vinylcyclohexane and (d) isotacticACHE-PEPfrom hydrogenation of iPS (Ref. 18)

in a mixed cyclohexane—ethyl acetate solvent system, followed by hydrolysis of the acrylate to prepare ionomeric polymers [27]. Levels of hydrogenation of >90% were reported as determined by 'H NMR spectroscopy. The hydrogenation of poly(styrene-co-4-hydroxystyrene) has also been reported to provide hydroxy functionalACHE-PEP[28]. This synthesis was performed by copolymerizing styrene with p-tert butoxystyrene to high molecular weight, debutylating with HC1, and hydrogenating the copolymer with Pd/C in tetrahydrofuran. Complete hydrogenation was obtained with minor chain degradation. 3.2

HYDROGENATION

MECHANISM

The detailed mechanism by which the heterogeneous catalysis of reactions on high polymer occurs has not been the subject of directed study. The mechanism of the catalytic hydrogenation of PS has been the subject of much speculation, however, largely based on evidence obtained by the examination of the properties of incompletely reacted species. Of particular interest are details of the process by which high molecular weight polystyrene interacts with the catalyst surface and becomes saturated, and the structure of the polymer during intermediate steps leading up to complete hydrogenation.

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

537

A 1955 patent [6] reported the hydrogenation of polystyrene using a reduced Ni catalyst; these workers then introduced the solution of hydrogenated polymer to propylene oxide (or other cyclic ethers). This treatment resulted in precipitation of the hydrogenated polymer, while unreacted polystyrene remained in solution. This process was motivated by a desire to prepare ACHE-PEP that was completely free of unsaturated styrene repeat units, which were deemed to be detrimental to the resistance of the polymer towards ionizing radiation. This separation requires that the hydrogenation proceeds such that, at intermediate levels of hydrogenation, a mixture of fully hydrogenated chains with unhydrogenated or minimally hydrogenated chains is present. Spectroscopic analysis of the recoveredACHE-PEPshowed very little residual unsaturation; no detailed analysis of the unreacted (soluble) fraction was reported. The implication of this work is that specific polymer—catalyst interactions occur during the hydrogenation reaction that are physically localized such that individual chains react to completion when in proximity to the catalyst particle, while other PS chains are completely (or almost completely) unreacted. In a 1967 publication [8], the characterization of partially hydrogenated polystyrene (prepared using a reduced Ni catalyst) suggested that, at intermediate levels of hydrogenation, the product consisted of a mixture of completely saturated polymer and unsaturated (or lightly saturated) chains, based on the presence of two discrete glass transition temperatures. This work elicited responses from two other groups, who reported on the characterization of partially hydrogenated PS [9,10] (also prepared using an activated Ni catalyst) in which the opposite conclusion was reached. Based on the observation of single glass transition events for a series of partially saturated samples that systematically approached the Ts observed for an independently prepared isotacticACHE-PEPsample, a random copolymer structure was inferred. The glass transition temperature does eventually increase with increasing degrees of hydrogenation, but deviates significantly from the Tg predicted by the Flory-Fox relationship (Figure 23.2). There is virtually no increase in the Tg until the level of hydrogenation approaches 50 %, followed by a rapid increase in Tg at higher levels of hydrogenation. A more recent study has also addressed the issue of the structure of incompletely hydrogenated PS [16]. Partially hydrogenated samples (using Ni/SiO2) were characterized by transmission electron microscopy, and clear evidence of a two-phase system is apparent at intermediate levels of saturation as observed by transmission electron microscopy (TEM). They also found a single glass transition temperature for various incompletely saturated materials that shifted to higher temperatures with increasing degrees of saturation, but which deviates substantially from the Flory—Fox relationship. The implication of this work is that, under these hydrogenation conditions, samples at intermediate levels of hydrogenation are not mixtures of completely saturated and completely unstaturated chains, nor are the repeat units randomly saturated. One possible explanation is the formation of a 'blocky' structure, where a

S. F. HAHN

538 160 -- DSC

—

Dilatometry (up) Dilatometry (down)

140 U 60

0

50 Degree of Hydrogenation (%)

Figure 23.2 Glass transition temperatures as determined by dilatometry and by DSC methods, with respect to the level of hydrogenation. Reproduced from Ref. 10, with permission of John Wiley & Sons

predominant fraction of individual chains are hydrogenated in segments that are long enough to undergo phase separation. The various stages of the polystyrene hydrogenation reaction with a supported catalyst have been enumerated [19]. These include diffusion of the polymer in solution, adsorption on the catalyst surface, the hydrogenation reaction, rearrangement of local chain conformations at the catalyst surface, and desorption of the polymer away from the catalyst. Depending on polymer molecular characteristics and catalyst structure, interaction of the polymer with the catalyst could potentially proceed such that entire chains of the polymer would become saturated, leading to a mixture of saturated and unreacted chains. Alternatively, segments of the polystyrene chain of varying length could become saturated without complete chain saturation. If such a mechanism were operating, the length of hydrogenated segments would be influenced by the previously described factors and others, including the nature of the catalyst, effective surface area of the catalytic site, temperature, hydrogen pressure, and competitive adsorption between solvent and polymer.

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

539

It has been noted [20] that, during deuteration experiments, more than the expected stoichiometric amount of deuterium was added to the cyclohexyl ring (from density measurements). This isotope exchange event had not been described previously, and requires that some mechanism other than simple transfer of HI across the sites of aromatic ring unsaturation must be occurring during the hydrogenation reaction.

4

POLYMERIZATION OF VINYLCYCLOHEXANE TO ACHE-PEP

ACHE-PEP has been prepared by the direct polymerization of vinylcyclohexane using coordination polymerization catalysts. The monomer vinylcyclohexane has been prepared by the hydrogenation of acetophenone, followed by dehydration of the methylcyclohexylcarbinol. The polymerization reaction has usually been performed using Ziegler—Natta-type catalysts [13,29-35], typically a TiCl3-trialkylaluminum mixture; this reaction provides isotacticACHE-PEP.It is also reported that isotacticACHE-PEPcan be prepared from vinylcyclohexane using a CrCV aluminum silicate catalyst. More recently, the polymerization of vinyl cyclohexane using a homogeneous metallocene catalysts has been described [36,37]. This polymerization appears to operate with a high level of chain end control (a 'living polymerization'), and the rates of polymerization are unusually high. This catalyst was used to perform the synthesis of isotacticACHE-PEPwith narrow polydispersity and to make block copolymers of i-PCHE with 1-hexene. Attempts to perform the cationic polymerization of vinylcyclohexane have been reported. While coordination-type polymerization of vinylcyclohexane monomer gives isotacticACHE-PEP,polymerization under conditions that lead to the formation of carbocationic intermediates leads to polymers with a differentiated backbone structure. Instead of propagating via the vinyl carbons, the cationic polymerization proceeds via hydride shift to a tertiary carbocation, which then propagates to provide the polymer shown in Scheme 23.3 [32]. Conditions for these polymerizations typically involved the use of aluminum halide catalysts in halogenated hydrocarbon solvents at low temperatures [32,38]. For the most part, molecular weights are relatively low.

5 5.1

CHARACTERIZATION OF ACHE-PEP ATACTIC ACHE-PEP

The majority of the research groups who synthesizedACHE-PEPhave been motivated by the relationship between the structure of this polymer and its physical

540

Scheme 23.3

S. F. HAHN

Cationic polymerization of vinylcyclohexane

and mechanical properties. There has been considerable inconsistency in the values obtained for some of these properties, especially the glass transition temperature of atacticACHE-PEPand the melting points of the stereoregular polymers. A chronological listing of the reported values is given in Table 23.1. The use ofACHE-PEPsamples prepared by a variety of techniques with a wide variation in molecular weights and polydispersity accounts for some of the lack of agreement. For research in whichACHE-PEPsamples were prepared using hydrogenation chemistry, the inconsistencies appear for the most part to be tied to uncertainties in molecular weight and purity of the final products. The relatively vigorous conditions employed to bring about high levels of saturation in the hydrogenation of PS have commonly been accompanied by chain degradation. Early reports that focused specifically on the glass transition temperature ofACHE-PEPsuggested that there was virtually no change in Tg compared with PS [39,40]. The first detailed study to focus on complete saturation of the aromatic rings and characterization of the hydrogenated product reached the conclusion that hydrogenation led to a drop in Tg, from 100 °C for PS to 80 °C forACHE-PEP[8]. This particular report elicited responses from two separate groups [9,10], who noted that the interpretation of the DSC analysis takes the Tg not from the midpoint of the heat flow, but from the onset. In addition, the reaction was run in decalin, a high-boiling solvent that is difficult to remove completely from the polymer, and the hydrogenation was accompanied by notable molecular weight degradation. The follow-up reports found the Tg to be 120°C for a 99.7% saturated PS [9] and 142 C for a 97% hydrogenated PS [10] by DSC analysis. A 1973 article advanced the characterization ofACHE-PEPby focusing on the physical and bulk mechanical properties ofACHE-PEP[14]. Although the glass transition temperature was not explicitly determined, this report provided a heat distortion temperature of 142.5 °C, further suggesting that the Tg increased upon hydrogenation. They also reported thatACHE-PEPretained its tensile strength to higher temperatures than PS, and had very similar mechanical properties (tensile strength and impact toughness at room temperature).

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES Table 23.1

541

Reported glass transition and melt temperatures of ACHE-PEP

Year of publication

Method

Tacticity

1952 1959 1961 1962 1964 1965 1967 1968 1969 1969 1970 1976 1978 1978 1991 1992 1993 1995 1998 1999 2001

? Birefringence TMA Instron DSC

a i i a i i a a a i a a a i s i a a a s a

9

DSC DSC/dilatometry DSC/DMA DMA Tensile relaxation DMA DSC DSC ? DSC DSC DSC DSC DSC DSC

Tg (°C)

Tm (°C)

99 0 90 — — 80 120 150 133 143 120 138 127

80 140 137 147 — 147

305 320-325 — 383 372 — — 370 ___ _— — Decomp. 303-351 405 — _._ — 273 —

Ref. 39 48 31 40a 32 40b 8 9 10 10 11 41 17 17 18 34 19 59 16 21 26

A detailed study of the glass transition temperatures of PS and of variously substituted poly(alkylstyrenes) and their hydrogenated counterparts (with careful control of molecular weight degradation) provided considerable insight to the differences in structure-property relationships between these two sets of polymers [20]. The presence of the alkyl substituents influenced the properties of both sets of polymers, but the nature of the influence was profoundly different in the PS series compared to theACHE-PEPseries (Table 23.2). In all cases, hydrogenation was found to lead to a decrease in the density of the polymers, which are in the range from 1.02–1.05 g/cm3 for the substituted polystyrenes to 0.91–0.95 g/cm3 for the hydrogenated polymers. Two examples are particularly illustrative of the trends and differences in Tg for these series. The Tg of poly(ter/-butylstyrene) is 141 °C, substantially higher than that for polystyrene, while the Tg of poly(terf-butylcyclohexylethylene) is 131 °C, measurably lower than that ofACHE-PEP.This contrast points to the influence of the ter/-butylphenyl ring on backbone motion in the styrenic polymer. Where the changes to local chain packing brought about by the tert-butyl group leads to inhibition of backbone motion, the p-tert-butyl group on a cyclohexyl ring does not lead to a dramatic change in the steric influence of that substituent on backbone motion. In contrast with this polymer pair, the Tg of poly

542

S. F. HAHN

Table 23.2 Molecular characteristics of hydrogenated substituted Styrenic polymers—(CH2CHR)—[20]a. Reproduced from Gehlsen et al., J. Polym. ScL, Part B: Polym. Phys., 33, 1527(1995) by permission of John Wiley & Sons, Inc. —R

Cyclohexyl 2-Methylcyclohexyl 3-Methylcyclohexyl 4-Methylcyclohexyl 4-te/7-Butylcyclohexyl

Mw 350000 447000 160000 110000 163000

Density (g/cm3)i r g (°C) 6(25 C C)(A) lO^A"1) 0.9469 0.9402 0.9237 0.9249 0.9140

137 186 141 132 131

7.1 7.2 8.5 7.0 7.0

4.4 3.9 5.4 3.7 2.7

a

b is the monomer segment length, b = Rg/(N/6)1//2, where R% is the radius of gyration and N is the degree of polymerization; p is the conformational symmetry parameter, ß2 = R2g/ V', where V is the segment volume.

(2-methylstyrene) is 132 °C, whereas that of poly(2-methylcyclohexylethylene) is 186°C, the highest Tg in either series. The relative influence of the 2-methyl group on the hydrogenated polymer reinforces the proposition that the ability of the cyclohexyl substituent to crowd the backbone is a dominating factor in influencing the properties of the hydrogenated materials. The extent to which the bulkier substituent inhibits local chain motion is much greater in the saturated polymer than in the starting material. The relationship between molecular weight and the glass transition temperature ofACHE-PEPwas examined in greater detail by the preparation of narrow polydispersity PS, hydrogenating under conditions that avoided degradation, and determining Tg using a DSC method. The relationship between Tg and molecular weight thus established showed that Tg approaches 148 °C in limit of high molecular weight (Figure 23.3) [26]. Attempts to account for the observed differences in physical properties between PS andACHE-PEPbased on repeat unit structure have evolved over the last 70 years as the physical properties have become better established. Before the order of glass transition temperatures had been firmly established, attention was focused on the potential for TT-TT interactions between the phenyl rings, which might give rise to 'chain stiffening' that could not be present inACHE-PEP[8]. More recently, interpretation of the properties ofACHE-PEPhave focused more closely on the potential that the greater relative steric bulk of the ACHE-PEP cyclohexyl ring would influence local chain motion and packing, in comparison with the planar phenyl ring in PS. The cyclohexyl ring, with an sp3-hybridized carbon attached directly to the backbone, and which usually occupies a chair conformation, is considerably bulkier than a phenyl ring. The increased bulk of this substituent increases the energy required to bring about local chain motion by reorganization of the cyclohexyl substituents on adjacent monomer residues (leading to increased Tg) and inhibits chain packing, (leading to decreased density) (Figure 23.4).

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

543

150 145 140 135 130

125 120

5 10-5

0.0001

0.00015 1/Mn

0.0002

0.00025

Figure 23.3 Glass transition temperature of monodisperseACHE-PEPsamples plotted with respect to 1/Mn- Reprinted with permission from J. Zhou et al., Macromolecules, 34, 1739. Copyright 2001 American Chemical Society

The relative influence of the cyclohexyl substituent onACHE-PEPcan be compared with other similarly substituted polyethylene derivatives. The entanglement spacing ofACHE-PEPhas been determined from small angle neutron scattering (SANS) [41] and from rheological measurements [26]. Data from both experiments arrive at a value of Me — 39 000 g/mol forACHE-PEP,roughly twice that of PS and 40 times that observed for polyethylene (PE). Another measure of the influence of the cyclohexyl substituent on the properties of ACHE-PEP is the packing length, p [41]. The packing length is a relative measure of the chain thickness given by the occupied volume of the chain, M ( p N a ) - l divided by the mean square end-to-end distance < R2 >o, where M is molecular weight, p is the polymer density, and N.d is Avogadro's number.ACHE-PEPhas a packing length of 6 A, one of the highest measured, compared with 1.69 A for PE and 3.99 A for PS. The influence of the cyclohexane ring on the glass transition temperature and sub-Tg transitions has been studied by dynamic mechanical methods [10,16, 42]. These comparative studies have focused on the influence of the cyclohexyl substituent on observed thermal transitions. A systematic study of cycloalkyl substituents on alternating carbons of a polyethylene backbone was performed to study the influence of these substituents on the dynamic mechanical spectra of the polymers [42]. These workers also prepared materials in which the cycloalkyl substituents were spaced away from the polymer backbone by successively longer methylene chains, testing the influence of ring proximity

544

S. F. HAHN

Figure 23.4 Molecular models representing five monomer units each of PS (top) andACHE-PEP(bottom). Reproduced from Ref. 1, with permission of Wiley-VCH

on the observed dynamic mechanical spectra. This work established the presence of a low-temperature relaxation event inACHE-PEPat about—150°C, in addition to the a (glass) transition at 120°C (Figure 23.5); although more recent measurements place these transitions at higher (by ca 25 °C) temperatures. The dynamic mechanical analysis of partially hydrogenated atactic polystyrene samples has also been studied [10,16]. In addition to the glass transition temperature and the previously reported relaxation phenomona at —125°C (designated -y), an additional broad transition at — 20 °C was identified (ß) [16]. From the position of the -y transition, and an estimate of its activation energy (36 kJ/mol), this transition has been interpreted as arising from chairto-chair conformational reorientation of the cyclohexyl side group. A transition at the same temperature and with a similar activation energy has also been observed in the dynamic mechanical analysis of cyclohexyl acrylate [43], reinforcing this interpretation of the -y transition inACHE-PEP.Based on molecular dynamic simulations, the ß transition (apparent activation energy 49 kJ/mol) was assigned to the oscillatory motion of the cyclohexane rings about the

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

545

9.0

8.0

7.0

9.0

8.0

7.0

9.0

8.0

-200

-100

0

100

200

TEMPEATURE, °C Figure 23.5 Dynamic mechanical analysis ofACHE-PEPand allylcyclohexane. Reproduced from Ref. 42, by permission of John Wiley & Sons

carbon—carbon bond that connects them to the backbone. This stands in contrast with earlier interpretations of the 3 transition in bothACHE-PEPand PS, where this transition has been assigned to the conformational motion of short segments of the polymer backbone that accompanies reorientation of the cyclohexyl or phenyl rings [44]. The surface energy ofACHE-PEPhas been measured using the Johnson-KendallRoberts (JKR) method [45]; the surface energy ofACHE-PEPis considerably less than that of PS, and marginally lower than that of PE. Some of the physical and mechanical properties that have been reported for ACHE-PEP are provided in Table 23.3.

5.2

ISOTACTIC ACHE-PEP

The characterization of isotacticACHE-PEP(iPCHE) has been shown to have two distinct crystalline forms. Form 1, originally reported from analysis of isotactic

S. F. HAHN

546

Table 23.3

Physical properties of atactic ACHE-PEP

Glass transition temperature (°C) 3

Density, 23 °C (g/cm ) Density, 90 °C, (g/cm3) Refractive index, 589 nm Dielectric constant, 1 kHz Heat capacity [J/(gK)] Coefficient of thermal expansion (below rg), (nm/m °C) Entanglement molecular weight, Me (g/mol) Stress-optical coefficient (brewster) Mark—Houwink—Sakurada constant, k (ml/g) Tensile strength, 23 °C (psi) Flexural modulus 23 °C (GPa) Dynastat impact (unnotched) (kg cm/cm2) Rockwell hardness (R scale) A2 (second virial coefficient), toluene, 37 °C (molcm3/g2) Critical surface energy (mN/m)

147 0.947 0.9272 1.506 2.26 1.39 64 40,000 -200 Toluene: k = 0.0257, a = 0.61 Cyclohexane: k = 0.026, a = 0.65 3,900 2.8 2.9 125.5 (3.2-3.5) xlO 4 28-29

ACHE-PEP prepared by Ziegler-Natta polymerization of vinylcyclohexane,[30] has a 4/1 helical conformation packed in a tetragonal unit cell with a suggested space group I4\/a (axes a = 21.9 A and c — 6.5 A) (Figure 23.6). A second crystalline form has also been identified [46]; form II has been observed in fibers drawn below 240 °C, and when annealed at temperatures above 260 °C reverts to form I. Form II has a 24/7 helical conformation, also in a tetragonal unit cell, with axes a = 20.5 A and c = 44.6 A. The conformational energy of these crystalline forms has been the subject of computational studies [35,47]. The crystalline melting point of isotacticACHE-PEPhas been reported by several of the research groups who have prepared this material [48]. Although the precise melting point of iPCHE has not been agreed upon, its position is generally in the range 320–400 °C, slightly higher than that observed for /PS. The growth of single crystals from chlorinated hydrocarbon solutions and microscopic analysis of these crystals has been published [49].

5.3

SYNDIOTACTIC ACHE-PEP

SyndiotacticACHE-PEP(sPCHE) has been prepared by the hydrogenation of syndiotactic PS [18,21]. The dynamic mechanical analysis of sPCHE with different degrees of hydrogenation has been reported. [21] The sub-Tg relaxation transitions observed in the atactic polymer were also observed in the

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

547

Figure 23.6 Depiction of the unit cell of isotacticACHE-PEPForm 1 unit cell Reprinted with permission from Nishikawa et al., Macromolecules, 29, 5558. Copyright 1996 American Chemical Society

syndiotactic version, and the melting temperature is reported to increase upon hydrogenation from about 270 °C for sPS to about 350 °C for sPCHE [18].

6 6.1

COPOLYMERS CONTAINING ACHE-PEP RANDOM COPOLYMERS

A few attempts have been made to copolymerize vinylcyclohexane with other monomers using coordination-type polymerization catalysts. The successful copolymerization of styrene and vinylcyclohexane [50] with a Zeigler-type catalyst has been reported. The copolymerization of the closely related monomer 4-vinylcyclohexene with ethylene using a Zr-based metallocene and methylalumoxane activator has been described [51]. 6.2

BLOCK AND GRAFT COPOLYMERS

The complete hydrogenation of block copolymers of styrene with diene monomers has been reported. Poly(styrene-&/ocA>diene) based block copolymers in which the polydiene block has been hydrogenated have been commercially available since the 1960s. The most common polymeric structures of this type are based on poly(styrene-bl-butadiene-bl-styrene) and poly(styrene-bl-

548

S. F. HAHN

isoprene-bl-styrene) block copolymers that are synthesized by anionic polymerization. For the examples in which saturation of both the polystyrene and the polydiene segments have been performed, conditions that provide saturation of the polystyrene block also lead to saturation of the polydiene block. The identity and repeat unit structure of the polymeric dienes provide a level of differentiation in the properties of the fully hydrogenated block copolymer. The saturation of polybutadiene that is predominantly of the 1,4-microstructure provides the equivalent of poly(ethylene-co-1-butene) (PEB) with low 1-butene levels (Scheme 23.4). The butadiene polymerization can be modified by the addition of polar cosolvents, which serve to increase the proportion of the 1,2-microstructural repeat unit (it can be made the predominant species). Hydrogenation of high 1,2-polybutadiene gives the equivalent of atactic poly (1-butene), and is commonly referred to as polyethylethylene (PEE). Changing the repeat unit structure of the polybutadiene precursor during the polymerization step will modify the physical properties of the EB block [52]. Polybutadiene with low levels of the 1,2-vinyl repeat unit will hydrogenate to give a crystalline polyethylene with low levels of 1-butene incorporation, while the presence of higher 1,2 content provides an amorphous EB copolymer [53,54]. In addition, the relative amounts of the two repeat unit structures provides polymeric segments with different plateau moduli; higher levels of the 1-butene unit lead to an increased entanglement spacing. The hydrogenation of polyisoprene [55] provides the equivalent of poly (ethylene-fl//-propylene) or PEP, typically with low levels of a 3-methyl-lbutene repeat unit due to 3,4 incorporation of isoprene during the anionic polymerization (Scheme 23.5). Hydrogenated polyisoprene is amorphous regardless of the microstructure of the polymer prior to hydrogenation. Syntheses ofACHE-PEP—PEBandACHE-PEP—PEPblock copolymers of varying architectures have been performed using the hydrogenation approach. A 1962 German patent describes the hydrogenation of polystyrene copolymers with butadiene using Rh on alumina and Ni on kieselguhr [7]. A 1967 US patent describes the hydrogenation of SIS block polymer with Ni on kieselguhr, and characterization of the product [56]. These materials contained 31–54% styrene

RLi _

.

cis& trans 1,4 Scheme 23.4 diene

,

, 1,2

catalyst

R

2E

B

Repeat unit structures of polybutadiene and hydrogenated polybuta-

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

549

RLi

cis and trans 1,4 Scheme 23.5 prene

3,4

Repeat unit structures of polyisoprene and hydrogenated polyiso-

by weight prior to hydrogenation with high degrees of styrene hydrogenation (98 % in one example), and were intended for use as thermoplastic elastomers. The hydrogenated materials had higher tensile rupture strengths and higher tensile modulus at 300% elongation compared with the unhydrogenated starting material, and showed marked improvement in the retention of properties after outdoor exposure. Further exploration [57] into the variation in properties available in the fully saturated poly(styrene-bl-butadiene-bl-styrene) materials focused on modifying the vinyl content in the polybutadiene block. Exploring practical elastomeric properties such as rebound and modulus, this work showed plainly that the level of vinyl in the polybutadiene block dominated the room temperature elastomeric properties of the block copolymer with a preferred level of about 40 mol% percent 1,2-microstructure. An extensive study of fully hydrogenated styrene-containing block copolymers was reported in 1971 [15]. This report included hydrogenated triblock copolymers of styrene with butadiene or isoprene, covering a large number of compositions and molecular weights. In addition to discrete block copolymers, where the composition at the chain crossover changes abruptly from one monomer to the next, work was also done to prepare tapered (compositionally mixed) regions between discrete blocks. The compositions ranged from materials with an excess ofACHE-PEP(tough, transparent plastics) to block copolymers with a greater amount of hydrogenated diene. The retention of physical properties (tensile strength, tensile elongation, and Izod impact) after exposure to weathering was noted. Blends of hydrogenated triblock block copolymers of various compositions withACHE-PEPhomopolymer were also characterized. Despite the body of patent literature describing fully hydrogenated block copolymers and their properties, it has been suggested that complete saturation of styrenic block copolymers with butadiene would result in materials that were incapable of microphase separation. This argument was based on the supposition that the difference in solubility parameters of the fully saturated block copolymer would be so slight that they would not have useful mechanical properties at achievable molecular weights [58]. This assumption has since

550

S. F. HAHN

been tested by several research groups, who have prepared PCHE—PEB, PCHE—PEE, and PCHE—PEP block copolymers and studied the physics of the microphase separation process of these materials. [59-62] The position of the order-disorder transition temperature, TODT [60] (the temperature at which the block copolymer transitions from an ordered, microphase-separated morphology to a disordered state) has been documented for several block copolymer structures and compositions (Table 23.4). The position of TODT has been found to be influenced most profoundly by the hydrogenated poly (diene) microstructure. When comparing TODT values of saturated and unsaturated versions of the same block copolymer, disordering occurs at lower temperatures in PCHE-PEE diblocks, but increases for PCHE—PE and PCHE—PEP. This behavior can be attributed to differences in the Flory— Huggins polymer-polymer interaction parameter x, which decreases upon hydrogenation for the transformation from poly(l,2-butadiene) to PEE but increases for poly(l,4-butadiene) to PEB and polyisoprene to PEP. The observed changes in the thermodynamics of mixing have been rationalized in terms of the influence of the local volume-filling nature of the various polymer species on the thermodynamics of the microphase interface [63,64]. A series of reports on the morphology of the crystalline polyethylene phase in PCHE-PEB block copolymers has been published [65–69]. The higher Tg of the PCHE (compared with PS) effectively locks in the block copolymer morphology prior to the nucleation of PE crystallization, forcing the PE crystallites to form within the block copolymer domain. This process restricts crystallite growth and imposes a specific chain folding orientation on the crystallites that do form with the PEB domain. Figure 23.7 shows a depiction of the morphology of the PCHE-PEB diblock copolymer, in which the polyethylene crystallites are forced to exist within the lamellar morphology of the block copolymer. Table 23.4 Order—disorder transition temperatures of hydrogenated poly(styreneblock-diene) copolymers Polymer PCHE-PEP PCHE-PEP PCHE-PEB PCHE-PEE PCHE—PEP PCHE—PEP PCHE—PEP PCHE-PEE PCHE—PEP ACHE-PEP—PEP

PCHE (%)

Mn( x 103)

50 51 50 50 40 38 32 56 13 13

18 15.6 14 50 25 26 25 23 94 64

(CC)

141 71 235 230 164 178 170 <80 233 231

Ref. 19 61 22 22 62 62 62 59 60 60

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

551

dn = 120A

I

am. PE

d = 180A

Figure 23.7 Morphology of PCHE-PEB diblock copolymer. Reprinted from Hamley etal., Polymer, 37 (1996) with permission of Elsevier Science

6.3

GRAFT COPOLYMERS

The synthesis of graft copolymer structures in which PCHE chains are grafted on to a hydrogenated polybutadiene backbone has been described [70,71]. The polybutadiene-graft styrene substrates were prepared by metallation of polybutadiene, which was then used to initiate a graft styrene polymerization. These polymers were then hydrogenated in cyclohexane using a Co(II) 2-ethylhexanoate-triethylaluminum catalyst at 4000 psi hydrogen pressure at 250 °C. Variation of the polybutadiene microstructure, graft level, and PCHE/EB ratio gave rise to materials with physical properties ranging from elastomeric to tough plastics.

7

PROPOSED APPLICATIONS OF PCHE-BASED MATERIALS

Although PCHE based materials are not commercially available as of this writing, a variety of potential applications have been outlined in scientific publications and in the patent literature. Clearly, the increase in Tg upon hydrogenation (and the commensurate improvement in heat distortion characteristics)

552

S. F. HAHN

has provided motivation to develop these polymers. In addition, the improvement in the stability of the saturated forms towards thermal excursions and irradiation provided incentive for development efforts. More recently, the recognition that saturation brings about useful changes in the electrical and optical properties of PS have been the focus of applications work. Various aspects of the use of PCHE and related materials as a dielectric layer in capacitor films has been described and patented by several workers [72,73]. The material properties that are important for this application are low moisture absorption (which allows for the formation of thin films with minimal pinhole formation), a high heat distortion temperature, and high breakdown voltage. The use of PCHE and copolymers containing PCHE in optical media applications as compact disc and digital video disc substrates has received considerable attention [74]. The inherently low stress optical coefficient of PCHE has the potential to allow for latitude in the disc molding process without inducing birefringence in the finished part. The saturated hydrocarbon composition of this polymer also leads to transparency well down into the UV spectrum, allowing for the use of reading lasers at the shorter wavelengths anticipated for higher data-density substrates. The use of PCHE homopolymer, copolymers, and blends of block copolymers and homopolymers at various levels of hydrogenation have been described. The hydrocarbon nature of this substrate also leads to some unique challenges; a series of patents describes methods to improve adhesion between the disc substrates and subsequentially deposited metal layers [75,76]. A description of the development of a block copolymers with a pentablock (PCHE-EB-PCHE-EB-PCHE) structure specifically for optical media applications has been published [77]. The excellent optical transparency of PCHE and its copolymers, along with the high heat distortion temperatures, have led to the development of this material for lenses and other optical devices [78]. In these applications, very low moisture absorption is also important, both with respect to molding high quality parts and in maintaining dimensional stability during use. The preparation of foams of PCHE and copolymers has also been claimed [79]. A combination of a low-boiling hydrocarbon (butane) foaming agent and a higher boiling hydrocarbon plasticizer (toluene) were used to prepare foams. The increased heat distortion temperature of PCHE has been proposed to lead to utility in insulating foams for hot water pipes and similar applications. In addition, the superior weatherability of this material would allow use in applications in which the product was subjected to UV exposure. The use of hydroxyl-substituted PCHE as a positive photoresist substrate has been patented [80,81]. Hydrogenation improves transparency into the deep UV region (200-300 nm) compared with materials containing aromatic rings and allows for smaller dimensions to be patterned. Features with good definition have been reported using partially hydrogenated substrates.

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

553

PCHE-containing block copolymers have been reported to have utility as melt spun elastomeric fibers [82]. The improved melt processability of these materials (compared with partially saturated block copolymers) allows them to be processed into small denier fibers with excellent tenacity and ultimate elongation.

8

ACKNOWLEDGMENT

Dr Joey Storer of the Dow Chemical Company generated the computer representations of PS and PCHE oligomer units depicted in Figure 23.6.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

D.A. Hucul, S.F. Hahn, Adv. Mater., 12, 1885 (2000). H. Staudinger, E. Geiger, E.Huber, Chem. Ber., 62, 263 (1929). H. Staudinger, V. Wiedersham, Chem. Ber. 62, 2406 (1929). P.N. Rylander, Catalytic Hydrogenation in Organic Synthesis, Academic Press, New York (1979). G. D. Jones, in Styrene, Its Polymers, Copolymers, and Derivatives, p. 689, Boundy, R.H., Boyer, R.F. (eds), Reinhold, New York (1952). A.J. Warner, O.K. Keel, US Patent 2726233, (1955). A. Steinhofer, R. Polster, H. Fruederich, German Patent 1131885 (1962); Canadian Patent 718089 (1965). H.G. Elias, O. Etter, /. Macromol. Sci., Al, 943 (1967). G.L. Taylor, S. Davison, /. Polym. Sci., Polym. Lett., 6, 699 (1968). A. Abe, T. Hama, t Polym. Sci., Polym. Lett., 7, 427 (1969). R.E. Kelchner, J.J. Aklonis, /. Polym. Sci., A2, 799 (1970). J.C. Falk, Makromol. Chem., 160, 291 (1972). J.F. Pendleton, D.F. Hoeg, Polym. Prepr., 13, 427 (1972). J.F. Pendleton, D.F. Hoeg, E.P. Goldberg, Adv. Chem. Ser., 129, 27 (1973). D.F. Hoeg, E.P. Goldberg, J.F. Pendleton, US Patent 3598886 (1971). H. Nakatani, K. Nitta, K. Soga, Polymer, 39, 4273 (1998). M. Helbig, H. Inoue, O. Vogl, /. Polym. Sci., Polym. Symp., 63, 329 (1978). N. Ishihara, M. Kuramoto, US Patent 4 997898 (1991). M.D. Gehlsen, F.S. Bates, Macromolecules, 26, 4122 (1993). M.D. Gehlsen, P.A. Weimann, F.S. Bates, S. Harville, J. Mays, G.D. Wignall, J. Polym. Sci., Part B: Polym. Phys., 33, 1527 (1995). H. Nakatani, K. Nitta, K. Soga, Polymer, 40, 1547 (1999). F.S. Bates, M.D. Gehlsen, V. L. Hughes, P. Brant, US Patent 5352 744 (1994). D.A. Hucul, S.F. Hahn, US Patent 5612422 (1997). D.A. Hucul, S.F. Hahn, US Patent 5654253 (1997). D.A. Hucul, S.F. Hahn, US Patent, 5 700878 (1997). J. Zhao, S. F. Hahn, D.A. Hucul, D.M. Meunier, Macromolecules, 34, 1737 (2001). J.-S. Kim, G. Wu, A. Eisenberg, Macromolecules, 27, 814 (1994).

554 28. 29. 30. 31.

S. F. HAHN

E. Tanaka, A. Kao, Japanese Patent Application 3107563 (1992). G. Natta, J. Polym. Sci., 34, 533 (1959). G. Natta, P. Corradini, I. W. Bassi, Makromol. Chem., 33, 247 (1959). E.A. Mushina, A.I. Perel'man, A.V. Topchiev, B.A. Krentsel, J. Polym. Sci., 52,199 (1961). 32. A. . Ketley, R. . Ehrig, J. Polym. Sci., A2, 4461 (1964). 33. V.I. Kleiner, B.A. Krentsel, L.L. Stotskaya, Eur. Polym. J., 7, 1677 (1971). 34. J.S. Grebowicz, Polym. Eng. Sci., 32, 1228 (1992). 35. C. De Rosa, A. Borriello, P. Corradini, Macromolecules, 29, 6323 (1996). 36. R.J. Keaton, K.C. Jayaratne, J.C.; Fettinger, L.R. Sita, J. Am. Chem. Soc., 122,958 (2000). 37. R.J. Keaton, K.C. Jayaratne, D.A. Henningsen, L.A. Koterwas, L.R.; Sita, J. Am. Chem. Soc., 123, 6197 (2001). 38. J.P. Kennedy, J.J. Elliott, W. Naegele, / Polym. Sci., A2, 5029 (1964). 39. L.A. Matheson, J.L. Saunderson, in Styrene: Its Polymers, Copolymers, and Derivatives, pp. 556–557, Boundy, R.H., Boyer, R.F. (eds), Rienhold, New York (1952) 40. (a) P.P. Reding, J. A. Faucher, R.D. Whitman, J. Polym. Sci., 57, 483 (1962); (b) J.A. Fauchner, F. P. Reding, in Crystalline Olefin Polymers, Part 1, p. 677, R.A.V. Raff, K. W. Doak (eds), Interscience, New York (1965). 41. L.J. Fetters, D.J. Lohse, D. Richter, T.A. Witten, A. Zirkel, Macromolecules, 27, 4639 (1994) 42. C.G. Seefried, J. V. Koleske, J. Polym. Sci., Polym. Phys., 14, 663 (1976). 43. J. Heijboer, Kolloid. Z., 1, 123, (1960). 44. C.A. Berglund, in Encyclopedia of Polymer Science and Engineering, 2nd edn, Vol. 16, pp. 142-148, Wiley, New York (1989). 45. M. Tirrell, Langmuir, 12, 4548 (1996). 46. H.D. Noether, J. Polym. Sci., Part C, 16, 725 (1967). 47. Y. Nishikawa, S. Murakawa, K. Shinzou, A. Kawaguchi, Macromolecules, 29, 5558 (1996). 48. T.W. Campbell, A.C. Haven, J. Appl. Polym. Sci., 1, 73 (1959). 49. J.D. Hutchison, J. Polym. Sci., A3, 2710 (1965). 50. G.B. Kharas, Y.V. Kissin, V.I. Kleiner, B.A. Krentsel, L.L. Stotskaya, R.Z. Zakharyan, Eur. Polym. J., 9, 315 (1973). 51. W. Kaminsky, D. Arrowsmith, H.R. Winkelbach, Polym. Bull., 36, 577 (1996). 52. H.L. Hsieh, R.P. Quirk, Anionic Polymerization Principles and Practical Applications, Chapt. 9, pp. 197–235, Marcel Dekker, New York (1996). 53. J.M. Carella, W.W. Graessley, L.J. Fetters, Macromolecules, 17, 2775 (1984). 54. T.M. Krigas, J.M. Carella, M.J. Struglinski, B. Crist, W.W. Graessley, J. Polym. Sci., Polym. Phys. Ed., 23, 509 (1985). 55. J.T. Gotro, W.W. Graessley, Macromolecules, 17, 2767 (1998). 56. W.R. Haefele, C.A. Dallas, M.A. Deisz, US Patent 3333024 (1967). 57. R.C. Jones, US Patent 3431 323 (1969). 58. W.P. Gergen, R.G. Lutz, S. Davison, in Thermoplastic Elastomers, G. Holden, N. R. Legge, R.P. Quirk, H.E. Schroeder, (eds), Chapt. 11, pp. 297–333, Hanser Ps, Cincinnati, OH (1996). 59. M.D. Gehlsen, F. S. Bates, Macromolecules, 27, 3611 (1994) 60. J.L. Adams, D.J. Quiram, W. W. Graessley, R. A. Register, G.R. Marchand, Macromolecules, 31, 201 (1998) 61. N.P. Balsara, C.C. Lin, H.J. Dai, R. Krishnamoorti, Macromolecules, 27, 1216 (1994).

HYDROGENATED POLYSTYRENE: PREPARATION AND PROPERTIES

555

62. C. Lai, W. B. Russel, R.A. Register, G.R. Marchand, D.H. Adamson, Macromolecules, 33, 3461 (2000). 63. F.S. Bates, G.H. Fredrickson, Phys. Today, 52, 32 (1999) 64. G.H. Fredrickson, A.J. Liu, F.S. Bates, Macromolecules, 27, 2503 (1994). 65. I.W. Hamley, J.P.A. Fairclough, N.J. Terrill, A.J. Ryan, P.M. Lipic, F S. Bates, E. Towns-Andrews, Macromolecules, 29, 8835 (1996). 66. I.W. Hamley, J.P.A. Fairclough, A.J. Ryan, F.S. Bates, E. Towns-Andrews, Polymer, 37, 4425 (1996) 67. D.J. Quiram, R.A. Register, G.R. Marchand, D.H. Adamson, Macromolecules, 31, 4891 (1998). 68. P.A. Weimann, D.A. Hajduk, C. Chu, K.A. Chaffin, J.C. Brodil, F.S. Bates, /. Polym. ScL, Part B: Polym. Phys., 37, 2053 (1999). 69. Y.-L., Loo, R.A. Register, D.H. Adamson, J. Polym. Sci, Part B: Polym. Phys., 38, 2564 (2000). 70. J.C. Falk, D.F. Hoeg, R.J. Schlott, J.F. Pendleton, J. Macromol Sci. Chem., A7, 1669, (1978). 71. J.F. Pendleton, R.J. Schlott, D.E. Hoeg, Canadian Patent 912 188 (1972). 72. S. Ruben, US Patent 2 266 812 (1941). 73. Y. Ozaki, Canadian Patent 2072 186 (1992). 74. M. Murayama, K. Kasahara, US Patent 4 911966 (1988). 75. Y. Suga, E. Tanaka, S. Kato, K. Sato, US Patent 5073 427 (1991). 76. S.F. Hahn, L.C. Lopez, M.D. Newsham, W.G. Lutz, US Patent 6030680 (2000). 77. F.S. Bates, G.H. Fredrickson, D.A. Hucul, S.F. Hahn, AIChE J., 47, 762 (2001). 78. T. Nagamune, T. Suzuki, PCT Application WO9855886 A (1999). 79. F. Kaneka, S. Sugawara, T. Okanika, Y. Mukoyama, Japanese Patent Application 9028896(1991). 80. R. Sinta, R.C. Hemond, D.R. Medeiros, M.M. Rajaratnam, J.W. Thackeray, D. Canistro, US Patent 5 258 257 (1993). 81. C.-L. Mertesdorf, H.-T. Schact, N. Muenzel, P.A. Falcigno, US Patent 6 124 405 (2000). 82. R.J. Patel, S.F. Hahn, C. Esneault, S. Bensason, Adv. Mater., 12, 1813 (2000).

This page intentionally left blank

24

Branched Polystyrene KURT A. KOPPI AND DUANE B. PRIDDY Dow Polystyrene R&D, Midland, Ml, USA

1

INTRODUCTION

Most commercial polymers are linear with only traces of branching (the most notable exception being low-density polyethylene). However, certain applications require high melt elasticity, e.g. blown films, extrusion blow molding, and foams. For these applications, polymer manufacturers often intentionally introduce branching into the polymer. Examples of commercial polymers that gain improved melt elasticity by the introduction of branching include polyethylene, polypropylene, and polycarbonate. However, branched polystyrene is currently not produced commercially. The objective of this chapter is to describe the chemistry that can be used to introduce branching in polystyrene and to discuss the melt rheological properties of branched polystyrene. We will propose a theory as to why branching may offer less performance advantage in polystyrene than in other polymers.

2 2.1

PREPARATION OF BRANCHED POLYSTYRENE RADICAL POLYMERIZATION

Most commercial polystyrene is manufactured using radical polymerization because it is the lowest cost chemistry [1]. The low cost is primarily due to the forgiving nature of the chemistry, i.e. monomers and solvents do not need to be highly pure. The economic impact of this attribute is significant. For example, the cost to convert styrene monomer to polymer using anionic polymerization chemistry is about 50% higher than for radical polymerization. This cost Modern Stvrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy r 2003 John Wiley & Sons Ltd

558

K. A. KOPPI AND D. B. PRIDDY

difference is mostly due to the cost of monomer purification. Since radical polymerization is used to manufacture polystyrene, all commercial polystyrene is slightly branched. The source of the branching is generally considered to be chain transfer to polymer [2]. The extent of branching is likely higher when peroxide initiators are utilized, especially when utilizing peroxides that generate potent H-atom abstracting tert-butoxy radicals. In an effort to quantify the extent of branching during radical polymerization of styrene in the presence of tert-butoxy radical generating initiators, Niki and Kamiya dissolved polystyrenes of various molecular weights in benzene [2]. A source of ter/-butoxy radicals (di-tert-butyl peroxylate) was introduced into the solution. If tertbutoxy radicals have ready access to labile H-atoms, tert-butanol is formed. If, however, the availability of H-atoms is poor, the tert-butoxy radicals fragment to form methyl radicals and acetone (Figure 24.1). A small molecule mimic of polystyrene (cumene) was dissolved in benzene at a concentration of 1 M available benzylic H-atoms. When ter/-butoxy radicals were introduced, the ratio of acetone to ter/-butanol was 11. However, when tert-butoxy radicals were introduced into benzene solutions containing polystyrenes of various molecular weights at the same molar concentration of benzylic H-atoms, the acetone: rm-butanol ratio dropped precipitously as the molecular weight of the polystyrene increased (Figure 24.2). Clearly the benzylic H-atoms attached to the polystyrene backbone are not as labile as in cumene. This is likely due to the steric effect of the coil configuration of the polymer chain which blocks access of the tert-butoxy radicals. Nonetheless, some backbone H-atom abstraction from the polystyrene backbone does occur during radical polymerization of styrene. The extent of abstraction is proportional to the concentration of peroxide initiator added to the process. Typically, in commercial continuous bulk polymerization processes the concentration of peroxide initiator is kept below 500 ppm. Also a few percent of a solvent having some chain transfer activity (ethylbenzene) is added to the styrene feed. This is done so that the extent of branching is small. If the concentration of initiator is increased to >500 ppm and/or the chain transfer solvent falls below a certain level, the extent of branching can increase to a level where gels began to appear in the product. The mechanism of

• -4-OO

o o

-

" " OO+-

^

2 —j- O

/

CH 3 .+

Figure 24.1

O ^U^

+

2 CO2

\PS-H -j-OH

+ PS-

Formation of ferf-butoxy radicals and their use to measure H-abstraction

BRANCHED POLYSTYRENE

559

aa

0

20

40

60

80

100

DP

Figure 24.2 Effect of degree of polymerization (DP) on the ability of tert-butoxy radicals to abstract H-atoms from the PS backbone to form tert-butyl alcohol (t-BA) or fragment to form acetone

the gel formation is not certain. However, it appears that a film of polystyrene forms on the metal surfaces inside the polymerization reactor [3]. This film is dynamic and is in equilibrium with the polystyrene in solution. However, the polystyrene coating the reactor walls has a longer residence time inside the reactor than the polystyrene in solution. The longer residence time results in more H-atom abstraction/branching. The polystyrene thus coating the reactor walls grows to a higher molecular weight than the polystyrene in solution, which subsequently slows the dissolution/redeposition equilibrium. Once branching reaches a point where there is a significant difference in viscosity between the polystyrene in solution and the polystyrene attached to the reactor wall, redissolution of the polystyrene from the reactor wall creates what appear to be 'gels' on the surface of extruded sheet or what are commonly referred to as 'fish-eyes' in blown polystyrene films. This problem is not unique to polystyrene made using continuous bulk polymerization processes. Polystyrene made by suspension polymerization is even more heavily branched. The higher levels of branching are due to the use of higher concentrations of initiator, no added solvent, and a higher monomer to polymer conversion in the polymerizer, i.e. 80% and >99% conversion for continuous bulk and suspension polymerizations, respectively. The main reason that the branching which takes place in suspension polymerization does not lead to reactor fouling is primarily due to the fact that each droplet of monomer dispersed in the continuous water phase is like a tiny isolated reactor. The water phase is continuously in contact with the reactor

560

K. A. KOPPI AND D. B. PRIDDY

wall rather than the monomer/polymer. If one adds 100ppm of divinylbenzene to a continuous bulk polymerization, the polymerization reactor eventually becomes totally plugged with crosslinked (insoluble) polystyrene after only a few weeks of continuous operation. However, addition of the same level of divinylbenzene to a suspension polymerization leads to the formation of branched polystyrene without any effect on reactor performance even after months of continuous (batch) operation. This illustrates the significance of the differences between the suspension and continuous bulk polymerization processes with regard to the negative impact of branching upon polymer homogeneity and long-term reactor performance. However, it should be pointed out that no-one any longer manufactures polystyrene using suspension polymerization because it cannot compete economically with the low conversion cost achieved by continuous bulk polymerization processes. For these reasons, branched polystyrene is not commercially available. Many researchers have attempted to make branched polystyrene in continuous bulk radical polymerization processes. Approaches involving the addition of additives to the polymerization process which lead to branching inside the polymerization reactor always lead to gel problems. Examples include addition of divinylmonomer [4], vinyl peroxides (e.g. I) [5,6], branched peroxides (e.g. II) [7], vinyl-functional chain transfer agents (HI) [8], and the use of additionfragmentation chain transfer agents that lead to the formation of polystyrene macromonomers (Figure 24.3) [9]. Recently, it was found that the common CT agent a-methylstyrene dimer operates by an addition-fragmentation mechanism (Figure 24.4). [10] The attribute of radical polymerization that makes it so problematic for making polymers of controlled structure is the complexity of the termination process (Figure 24.5). Since the main mode of termination is radical coupling, coupling of the polystyryl macroradicals produced by H-atom abstraction leads directly the formation of a crosslink. Dow Chemical Company researchers developed a laboratory experiment to quantify the propensity of a branching agent to form gel during continuous bulk radical polymerization of styrene [11]. Styrene is loaded into a feed tank along with the branching agent. The styrene is slowly pumped through a glass condenser containing a steel rod insert. The jacket of the condenser is heated at 100 °C with recirculated oil. After various periods of time the steel rod is removed from the condenser, washed with solvent and weighed. With the recent development of living radical polymerization, the problem of gel formation during radical polymerization possibly can be controlled. This is because termination by radical chain coupling is virtually eliminated. Thus Hawker reported the preparation of soluble hyperbranched polystyrene using alkoxyamine IV as a living radical polymerization initiator [12].

BRANCHED POLYSTYRENE

561

nBuO

Leaving group CH2 or O

•-P /

A

Z

Activating Group eg. Ph, CN, CO2Me

X

\

F = functional group Figure 24.3 action

F I

Z —F

Generic addition-fragmentation CT structure and the mechanism of

Another way to make branched polystyrene and still be able to use low-cost bulk polymerization processes is to wait and branch the polymer after the polymerization is complete. This approach has been referred to as postpolymerizer branching. Priddy and co-workers have pioneered several examples of this approach [13,14]. In general, the approach is to utilize either initiators or comonomers that contain latent functional groups. The functional groups are inert during the polymerization. After the polymerization process is complete, the groups are activated so that coupling takes place (Figure 24.6).

562

K. A. KOPPI AND D. B. PRIDDY

Ph

Ph

Ph a-methylstyrene dimer

f

Ph

Figure 24.4 Addition–fragmentation mechanism for the chain-transfer activity of a-methylstyrene dimer

Chain Coupling

n

Disproport.

Chain Transfer R-H

Primary Radical Coupling

Figure 24.5 Termination processes operational during radical polymerization of styrene

>200°C

Figure 24.6

General approach to post-polymerizer branching

An example of an excellent latent functional group for this purpose is benzocyclobutene (BCB) [13]. Polystyrene made using a BCB functional peroxide leads to the formation of linear polystyrene because the BCB moieties are inert below 180°C. Once the polymer exits the polymerization reactor it is

563

BRANCHED POLYSTYRENE

devolatilizes at 240 °C. The half-life of the BCD moieties at 240 °C is only seconds. Once the cyclobutene ring opens to form the highly reactive orthoquinone methide structure, coupling rapidly takes place resulting in both chain extension and branching (Figure 24.7). Figure 24.8 shows the effectiveness of technique. As expected, the weight average molecular weight (Mw) of the polystyrene exiting the polymerizer decreases as the amount of initiator utilized is increased. However, after heating the polymer to 240 °C, the Mw increases with the amount of initiator used. A second example involves the addition of small amounts of two monomers (to the polymerization process) having functional groups which couple with each other at high temperatures. Although there are likely a variety of monomer pairs that could be used, the acrylic acid-glycidyl methacrylate (AAGMA) pair was utilized to demonstrate the concept (Figure 24.9) [14].

chain extension & branching

Figure 24.7 An example of the post-polymerizer branching approach to branch polystyrene using a BCB-functional initiator

140012001000-

- Mw after polymerization Mw after devolatilization at 240°C

800600400200 4

0 0

500

1000

1500

2000

BCBPO (ppm)

Figure 24.8 Mw before and after heating (2 h at 240 °C) BCBPO initiated polystyrene vs the amount of BCBPO in the monomer feed

564

K. A. KOPPI AND D. B. PRIDDY

>200°C

Figure 24.9 Copolymerization of acrylic acid and glycidyl methacrylate with styrene to achieve post-polymerizer branching

A problem encountered when utilizing the AA-GMA pair was that there is a slight reactivity of the carboxylic acid and epoxide functional groups during the continuous bulk polymerization which led to gels and reactor plugging[14]. The success of this approach requires that the functional groups be totally inert inside the polymerization reactor. 2.2

ANIONIC POLYMERIZATION

Unlike radical polymerization, branched polystyrenes having a variety of controlled structures have been synthesized (Figure 24.10). This is because termination can be precisely controlled. The branched polystyrenes synthesized using anionic chemistry have been used to study the effect of branch structure on rheology [15]. As will be discussed in the next section, branch architecture (like those presented in Figure 24.10) can influence the Theological properties of polystyrene resins.

n~n H Star

Comb

Incipient Network* Figure 24.10

Tree

Dendritic

Several possible polystyrene branch architectures

BRANCHED POLYSTYRENE

3

565

RHEOLOGY OF BRANCHED POLYSTYRENES

The most obvious effect of branching in a polymer is to change the overall molecular conformation. Branching will decrease the size of a polymer molecule over that of its linear homologue. Such a decrease in size can be characterized by the following relation: Rg, branched

=

g th -Rg, linear

(.'•)

where Rg is the radius of gyration of the polymer [16] and gth is a scaling parameter first introduced by Zimm and Stockmayer that depends on branch structure [17]. Values of gth vary from unity for linear polymers to zero in the limit of highly branched polymers. A tabulation of mathematical expressions of gth for various branching architectures was given by Roovers [18]; e.g. for an equal arm star polymer: 3/ 2 e ,—- —~P)— #th

r

(2) vA;

where / is the number of arms per molecule. A comparison of theoretically determined values of gth with those measured experimentally has also been given by Roovers [18]. Because Rg, branched < Rg, linear > the intrinsic viscosity [n] of a branched polymer is expected to be lower than that of a linear polymer of identical molecular weight. Experiments have shown that

for non-free draining (high molecular weight) star polymers [19] and

for free draining polymers [17,20]. The exponent on gth is a function of both branching architecture and solvent quality and has been found to vary between 0.5 and 1.0 [18]. Polymer properties that depend on local chain conformations are unlikely to be influenced by long chain branching. For example, the glass transition temperature (Tg) of a branched polymer is almost always identical with that of its linear homologue. An exception to this general rule has been observed in very highly branched systems where the ratio of chain ends to molecular weight is rather high [21]. For such systems, the high number of branch points causes a

566

K. A. KOPPI AND D. B. PRIDDY

shift in free volume of the polymer, which in turn lowers the glass transition temperature with respect to its linear homologue. As can be seen from these general influences of branching on both [n] and Tg, only properties of polymers that are directly related to their overall dimensions are strongly affected by long chain branching. Another polymer property that can be greatly influenced by the presence of branching is melt rheology. As has already been pointed out, the hydrodynamic radius of a polymer coil in solution can be dramatically affected by branching, which in turn alters [n]. The presence of branching can also alter the rheological properties of polymer melts through a manipulation of the degree of entanglement experienced by any given chain. In addition, the presence of branches can also influence the type of relaxation modes that can be activated in a polymer melt, i.e. the presence of a branch will strongly inhibit the reptation of a polymer molecule along its contour [22,23]. A thorough discussion of the influence of long chain branches on the rheological properties of entangled polymer melts was given by McLeish and Milner [24]. It is this influence of branching on the rheological behavior of polymer melts that will be the focus of the remainder of this chapter.

3.1

STAR BRANCHED POLYMERS

A majority of the fundamental work concerning the influence of branching on the rheology of polymer melts has focused on star polymers owing to their relatively simple and well-defined branch architecture. Over 35 years ago, Kraus and Gruver [25] reported the influence of branching on polymer rheology through the study of three- and four-arm polybutadiene stars. They observed that at low molecular weight, the zero shear viscosity (n0) of the stars was lower than that exhibited by linear polymers of the same molecular weight. Upon increasing molecular weight, however, they observed that n0 of the stars far exceeds that of the linear homologues. This transition was observed to occur at molecular weights of roughly 60000 and 100 000 g/mol for the three- and four-arm stars, respectively. Thus, viscosity enhancement due to branching requires a certain level of arm entanglement before it becomes activated. For these particular examples, the levels of arm entanglement for this transition were Ma « 7Me and Afa ~ 9Me, respectively, with the entanglement molecular weight for polybutadiene taken as Me = 2800 g/mol. Upon increasing the shear rate, a different trend was observed. The shearthinning behavior brought about in response to moderate to high shear rates was observed to be more pronounced in the star polymers than the linear polymers. As a result, the moderate to high shear rate viscosities of the star polymers were significantly lower than those exhibited by linear polymers of the same molecular weight. Summing up, it was shown that above a certain critical molecular weight a branched polymer will exhibit a higher viscosity at low

BRANCHED POLYSTYRENE

567

shear rates than a linear polymer of the same molecular weight, but the branched polymer will then cross the linear polymer viscosity curve at intermediate shear rates and continue to remain below the linear curve at high shear rates. Several other studies concerning the rheology of star polymers have been reported [26-34] since the work of Kraus and Gruver [25]. Graessley et al. [26] extended the work to include solutions of linear, four-arm, and six-arm polyisoprenes. Concentrations in tetradecane ranged from 0.02 to 0.33g/ml and polyisoprene molecular weights ranged from 35000 to 2 000 000 g/mol. Both viscosity and first normal stress measurements were made as a function of shear rate. From these measurements, estimates of the zero shear viscosity r]0, the zero shear recoverable compliance J® (a measure of the elastic energy that can be stored by a polymer), and the characteristic shear rate (dy/d/)0 (locating the onset of a shear rate-dependent viscosity) were determined. It was observed that at low concentrations and low molecular weights, rjQ and J® were lower for the branched samples, whereas at high concentrations and high molecular weights the opposite was true and substantial enhancements of rj0 and J® were found. Despite these enhancements, the product ^ 0 /g(dy/dt)o was observed to be essentially the same for all samples irrespective of concentration, molecular weight, or branching, and as a result it was possible to construct a master curve independent of branching. Subsequent to this work, Quack and Fetters [27] made a very interesting discovery while studying the rheological properties of polyisoprene star polymers. Two sets of polyisoprene stars were investigated: 9–22 arms with Mw, arm = 84 000g/mol and 10-56 arms with Mw arm = 149 000 g/mol. They found that the viscosity of these star polymers was a function of the arm molecular weight and not the total molecular weight. In other words, the viscosity was independent of the number of arms per star and was dependent only on the level of arm entanglement. Subsequent work on polybutadiene [28] and polyisoprene [29] star polymers indicated that a slight dependence of rj0 on the number of arms occurs initially between three and four arms. It was observed that the zero shear viscosities of three-arm stars were observed to be approximately 20% lower than that of four-arm stars with equal arm length. Upon increasing the number of arms further, the effect of arm functionality eventually saturates and T/O becomes independent of the number of arms. Masuda et al.[30] reported data collected for a series of polystyrene star polymers that seemingly conflict with the discovery made by Quack and Fetters [27]. They showed that the viscosity of polystyrene star polymers was dependent on the number of arms. Specifically, they showed that viscosity increased with the number of branches for a series of polystyrene stars with AAv, arm = 55000g/mol and the number of arms ranging from 7 to 39. However, the level of arm entanglement for the polystyrene stars was far lower than that of the polyisoprene stars studied by Quack and Fetters [27].

568

K. A. KOPPI AND D. B. PRIDDY

Clearly, in order for the viscosity of star polymers to be independent of the number of branches, a certain level of entanglement needs to be present. Roovers [31] extended this analysis by measuring the Theological properties of star polymers with an extremely large number of arms. He measured the viscosity of polybutadiene stars with up to 270 arms and found that A/O of these highly branched materials were significantly higher than that of four-arm polybutadienes with the same arm molecular weight (Mv, arm = 1 0 000 g/mol). The short-time relaxation of these highly branched polymers appeared to be comparable to that of the four-arm stars, but their long-time relaxation was observed to be considerably retarded. The reason for this discrepancy probably stems from the high degree of extension experienced by the arms of highly branched stars near the junction point that results from packing constraints. Owing to this high level of extension near the junction point, relaxation facilitated through arm retraction is diminished, resulting in a severe increase in the characteristic relaxation time of the polymer. Following the initial discovery [27] that //0 depends on just arm molecular weight for star polymers with sufficiently high levels of branching, this type of dependence was confirmed by others both theoretically [32] and experimentally [33]. Pearson and Helfand [32] predicted that the zero shear viscosity of star polymers should scale with arm molecular weight (M a ) as 1/2

where Me is the entanglement molecular weight and v' is a constant. Roovers [33] then showed that such an expression fits the observed behavior of both four-arm star polybutadienes and four-arm star polystyrenes. As was stated earlier, an in-depth discussion comparing the Theological behavior of linear and star polymers including a description of the molecular origins of these differences was reported by McLeish and Milner [24]. The polystyrene data used by Roovers [33] to test this scaling law was originally reported in a previous publication by Graessley and Roovers [35]. It is interesting that Graessely and Roovers [35] did not comment on the onset of exponential dependence (rjQ vs Ma) that is clearly seen in their four-arm and six-arm polystyrene data. However, because they did not explore sufficiently high molecular weight polystyrene stars, they did not observe an increase in »/0 of the stars over that of linear polystyrenes as was observed by other researchers [25-29,33] investigating polymer systems other than polystyrene. From the data of Roovers [34], it can be seen that rjQ of their highest molecular weight (Mw = 1 300 000 g/mol) four-arm star polystyrene comes very close to that of the linear homologue. If their study had been extended to slightly higher molecular weights, the increase in A/O that results from branching as reported

BRANCHED POLYSTYRENE

569

by several others [25-29,33] would also have been observed. These results from polystyrene stars show that the level of arm entanglement necessary for viscosity enhancement due to branching is greater for polystyrene than the other systems studied. The crossover occurs at Ma w 9Me for a four-arm polybutadiene, whereas for a four-arm polystyrene it would not occur until Ma > 18Me. It is not clear why this transition should occur at such a higher level of arm entanglement for polystyrene stars than for other star polymers. This observation is in direct conflict with the standard assumption that through a proper scaling of plateau modulus (Go) and monomeric friction coefficient (£) that rheological behavior should be dependent only on molecular topology and be independent of molecular chemical structure. This standard assumption was demonstrated to hold fairly well for the linear viscoelastic response of wellentangled monodisperse linear polyisoprene, polybutadiene, and polystyrene melts by McLeish and Milner [24].

3.2

COMB BRANCHED POLYMERS

Knowledge regarding the Theological properties of comb polymers is desirable as a means of bridging the 'structural gap' between the randomly branched commercial polymers and the model star polymers. The smallest member of the comb polymer family is an H-shaped polymer. In a study of the rnelt rheology of H-shaped polystyrenes with equal segment molecular weight, Roovers [24] found that the zero shear viscosity (^0) of low molecular weight H-polymers are the same as those of linear polymers of the same size, but at high molecular weight H-polymers exhibit enhanced viscosities. Exponential dependence of f/0 on Ma, as discussed earlier for star polymers, was also seen in the H-polymers. The crossover where r}0 of polystyrene H-polymers surpasses that of its linear homologues occurs at Mw w 600 000 g/mol. This value is significantly lower than that observed for four-arm star polystyrenes [35]. Furthermore, the average entanglement of each of the five subchains of the H-polymers at this crossover is Ma ~ 7M e , a value far lower than the level of arm entanglement required to reach the crossover in polystyrene star polymers. Roovers [34] demonstrated that the added branch point found in an H-polymer significantly alters its relaxation behavior vis-a-vis a star polymer. He showed that the value of v' [Equation (5)] was approximately twice that of four-arm star polymers. The time-scales for relaxation of the free-end subchains of H-polymers are very different to the relaxation time-scale associated with the 'cross-bar' subchain that exists between the two branch points of the H-polymer. The free-end subchains of H-polymers relax in a manner nearly identical with that of the arms of a star polymer. However, the 'cross-bar'

570

K. A. KOPPI AND D. B. PRIDDY

subchain relaxes using a very different process. A detailed discussion of these differences has been given by McLeish et al. [36]. The rheological properties of polymers with more complicated branch architecture have been investigated, but unfortunately such polymers are seldom of the monodisperse variety as can be found in star and linear polymers. Work on comb polymers [37-39] has shown in general that foo)comb >

ga>o)linear

(6)

where g = (Rg)branched/(Rg)iinear and its exponent is dependent on the level of arm entanglement. Such a relation indicates an enhancement in viscosity over that of linear polymers due to the presence of branches. However, the dependence of viscosity on the number of branches, their spacing along the backbone, and their length is not clear. In two papers, Pannell [37,38] reported investigations on the influence of branch length and branching frequency in polystyrene combs. In the earlier investigation [37], a series of combs were studied. Each comb possessed a branching frequency of either 33 or 71 branches per molecule and the size of the branches was varied systematically from a molecular weight of 5000 to 45000g/mol. In the later investigation [38], the samples were only lightly branched with branching frequencies ranging systematically from 2 to 19 per molecule. In this case, however, the branches were relatively long (46000410000g/mol). In addition to these branched polymers, the Theological properties of the corresponding backbone polymers (i.e. isolated before performing the branching step) were also investigated. Experiments performed on the highly branched combs [37] revealed that the effect of adding short branches was to lower f/0 below that of the backbone polymer. Upon increasing the branch length, r\Q continues to decrease and eventually a minimum is reached. Further increases result in an increase in rjQ over that of the backbone. Experiments performed on the lightly branched combs [38] revealed a different behavior. At a given branching frequency, //0 was observed to increase above that of the backbone polymer as branch length was increased and no diminishment in rj0 due to branching was observed. However, comparing these lightly branched materials with linear polystyrenes of the same overall molecular weight revealed that f/o of the branched materials was almost always lower than that of the corresponding linear homologue. The only exception to this general trend was observed in the samples with the lowest branching frequency of two branches per molecule at very high branch length, Mbr > 800 000 g/mol. For samples with these characteristics, rjQ of the branched material was greater than that of the corresponding linear homologue. In a later study, Roovers and Graessley [39] investigated the role that molecular weight between branches plays in the rheology of polystyrene combs. Two backbone polymers with molecular weights of 275 000 and 860 000 g/mol

BRANCHED POLYSTYRENE

571

were prepared. From these a series of combs were made. Each comb had approximately the same number of branches (~30) but the length of the branches was varied systematically from a molecular weight of 6500 to 98 000g/mol. It was observed that for the materials with the low backbone molecular weight, »?0 remained essentially constant until Mbr > 2M e , whereupon 7/0 increased with increasing branch length. For the materials with the high backbone molecular weight, however, an increase in rjQ over that of the backbone polymer was observed in each sample; even in samples with low branch lengths, Mbr>Me. This indicates that the amount of entanglement in the segments between branches along the backbone plays an important role in the Theological behavior of comb polymers. Further evidence that this is true is given by comparing samples of identical branch length but differing backbone molecular weight. When this is done it is observed that the relative increase in J70 for a long backbone sample (over that of the corresponding backbone polymer) is far greater than that observed for the corresponding short backbone polymer with the same branch length. The average molecular weight between branches along the backbone of the samples with the low molecular weight backbone were approximately 9000g/mol (< Me) whereas that along the backbone of the samples with the high molecular weight backbone were approximately 30000g/mol (> Me). Hence it can be concluded that entanglement of the doubly constrained segments along the backbone has a greater effect on the rheological properties of comb polymers than does entanglement in the less constrained branches. As a result of the added constraint, the relaxation of a segment along the backbone of a comb will be slower than that of a branch that is relatively free to move at one end.

3.3

RANDOMLY BRANCHED POLYMERS

Only a limited number of careful studies have been performed on randomly branched polystyrenes. In general, the observations just summarized for combs seem to also apply to randomly branched polymers. Masuda and co-workers [40,41] reported a decrease in ?/0 for concentrated solutions of randomly branched polystyrenes. These randomly branched polystyrenes were prepared by the copolymerization of styrene monomer with a small amount of divinylbenzene. The resulting polymers were then fractionated by the fractional precipitation method in benzene to create a series of randomly branched polymers with varying Mw and branching level. Three series were prepared with molecular weight between branches of approximately Mb = 46 000, 93000, and 174000g/mol. The overall molecular weight for these series ranged from Mw = 70000 to 1 120000, 37000 to 6100000, and 75000 to 566000g/mol, respectively. Clearly, the low end of each of these series of polymers contains chiefly linear polymers. With increasing molecular weight, the fractionates

572

K. A. KOPPI AND D. B. PRIDDY

are best described as star polymers followed by H-shaped polymers. Eventually, the polymers become more random in nature with the increased possibility of branches on branches. The rheological properties of these polymers were characterized in 50wt% solutions along with 50wt% solutions of linear polystyrenes. In comparing these results with the previously described polymer melt rheology, one has to bear in mind that as a result of solvent dilution, the entanglement molecular weight for these concentrated polystyrene solutions is greater than what is present in polystyrene melts. Masuda [40,41] and co-workers demonstrated that their linear polymer solutions exhibited the classical power law scaling of f/0 with Mw (j/o ~ M1.0w at low Mw followed by //0 ~ M^,5 at high molecular weight). However, the randomly branched polystyrene solutions exhibited the scaling f/0 ~M0.8Af^84across the whole range of molecular weights. The randomly branched curve overlaid with the linear polystyrene curve at low molecular weights as expected because the low molecular weight fractionates of each random polymer series were essentially linear polymers. The reductions in viscosity observed by Masuda and co-workers [40,41] of randomly branched polystyrenes vis-a-vis linear polystyrenes of the same overall molecular weight are consistent with the theory of Bueche [42] that predicts the viscosity of branched molecules should be less than that of linear molecules of the same overall molecular weight. The basis for this prediction is that a randomly branched molecule has a smaller radius of gyration than a linear molecule of the same molecular weight. It should be noted that this theory does not include entanglement effects. However, as the randomly branched polystyrene solutions exhibit an rjQ ~ M0.84w scaling across the whole ranges, studies demonstrate that entanglement coupling effects are not exhibited by these polymers. It should be noted that when r\Q was plotted as a function of g th M w instead of Mw, a master curve independent of branching was observed, further supporting the applicability of the Bueche theory [42]. The rheological properties of randomly branched polystyrenes in the melt were studied by Hempenius et al. [43]. These polymers were synthesized via so-called arborescent grafting. This technique consists of coupling linear polymer chains through a cascading grafting process. This cascading process is carried out in generations akin to a dendritic polymerization with the notable exception that in arborescent grafting the branches are added randomly. The first generation of such arborescent graft polymers are comb polymers but in subsequent generations branches are added to branches. Three series of such randomly branched polymers were prepared by Hempenius et al. [43]. The first was comprised of graft chains with Mb = 5000 g/mol; four generations were prepared with the resulting molecular weights spanning Mw = 80 000-28 500 000 g/mol across the generations. The other two series were synthesized across three generations using graft chains of Mb = 10000 and 20 000 g/mol, resulting in molecular weights spanning Mw = 156 000–15 700 000 and 389 000-54 800 000 g/mol, respectively.

BRANCHED POLYSTYRENE

573

The zero shear viscosities of these randomly branched polystyrenes were measured and compared with those of linear polystyrenes and it was found that r/0 for all of the branched polymers were far lower than that of linear homologues of the same overall molecular weight. In addition, a scaling of TJQ ~ M1.0w was observed for the first two generations of each branched series of polymers. This behavior is similar to that reported by Masuda and co-workers [40,41] for randomly branched polystyrene concentrated solutions and illustrates the lack of entanglement coupling for these particular randomly branched polymers melts. For the subsequent generation of arborescent graft polystyrenes, a dramatic increase in rj0 was observed by Hempenius et al. [43] for each of the three series included in their study. However, despite this increase in viscosity, the rjQ for each of these is still lower than that of the linear homologue polystyrenes of the same overall molecular weight. This jump in viscosity is due to an increase in branch density which in turn results in increase in chain extension similar to that observed by Roovers [31] for highly branched star polymers. A similar reduction in viscosity to that found by Masuda and co-workers [40,41] and Hempenius et al. [43] for randomly branched polystyrenes vis-a-vis linear polystyrenes was also observed by Ferri and Lomellini [44] for another type of randomly branched polystyrene melts prepared via the copolymerization of styrene monomer with very low levels of divinylbenzene. These randomly branched polystyrenes were described as resembling four arm stars with nonuniform arm lengths. A better description would be a mix of such nonuniform stars with linear polymer. The average number of quaternary branching points per chain for the randomly branched polymers was reported to range from 0.165 to 0.48 for the polymers included in the study. Despite the fact that the arms of these randomly branched polymers were nonuniform and that they were essentially mixed with linear polymer, the rheological behavior was similar to that of uniform four-arm and star polymers. It should be noted that in addition to shear viscosity measurements, extensional viscosity measurements were also performed by Ferri and Lomellini [44] for these polystyrenes.

3.4

EXTENSIONAL RHEOLOGY

Very few studies have been performed investigating the effect of branching on the extensional rheological properties of polystyrenes. Such investigations can be valuable because many of the fabrication operations associated with commercial applications of polystyrene include operations in which the polystyrene melt undergoes an extensional deformation. Some examples are extruded foam sheet, blown film, oriented (tentered) sheet, and thermoforming. The types of deformations associated with these processing operations are best described as

574

K. A. KOPPI AND D. B. PRIDDY

nonuniform biaxial deformations. This term is used because the most common deformation associated with the previously listed fabrication operations is the stretching of molten polymer within a plane; however, that stretching is not necessarily axi-symmetric within the plane. Typically, it is desirable for a polymer to have a certain level of melt strength so that it processes well in these types of fabrication operations. Ideally, one would like to characterize the melt strength of a resin using a test method that best captures the type of deformation experienced by a polymer in a commercial fabrication operation. Owing to the complex flows associated with such operations, the approach that is most commonly used is to characterize the melt strength of a polymer in a uniform flow field and then infer from this information its performance in a real fabrication operation. A comparison of the Theological properties of polymer melts in shear, biaxial extension, and uniaxial extension has been reported by Khan et al. [45]. Four types of polymers were investigated: low-density polyethylene (LDPE), highdensity polyethylene (HDPE), linear low-density polyethylene (LLDPE), and polystyrene. It was found that the differences in the properties of these four polymer melts were most pronounced in the uniaxial extension tests. During uniaxial extension, the branched LDPE sample exhibited rather strong strain hardening, even though its shear and biaxial properties were similar to those of the other melts. Based on this work, Khan et al. [45] concluded that for differentiating between polymer melts, biaxial extension is the least discerning, shear is intermediate, and uniaxial extension is the most discerning. Despite the fact that uniaxial extension is the most discerning technique, a relatively small amount of information has been published using this technique as a way to document the influence of branching on polymer rheology, especially for polystyrene. The predominant Theological technique (and the only one discussed thus far in this chapter) for exploring the influence of branching on polymer melt strength has been shear rheology. Typically, one uses the zero shear viscosity (^0) as a measure of polymer melt strength and the higher the value of ^0, the higher is the melt strength. It is for this reason that the main focus of this discussion of branching architecture on rheological properties has been its influence on rj0. As alluded to earlier, Ferri and Lomellini [44] included extensional viscosity measurements in their study of the rheological behavior of randomly branched polystyrenes. Extensional stress growth functions (rj+) were measured as a function of time at T — 140 °C using a uniaxial extension rate of de/dt = 0.005s-1 for two polystyrenes: a linear polymer with Mw = 192000g/mol and a randomly branched polymer with Mw = 391 000 g/mol. The former had a polydispersity Mw/Mn = 1.7 whereas the latter had Mw/Mn = 3.7. In addition, the latter had an average of 0.30 quaternary branching points per chain. At short times, the rj^ curves for these two polymers were very similar. However, at longer times (higher strains), the randomly branched polystyrene exhibited significant strain

BRANCHED POLYSTYRENE

575

hardening, whereas strain hardening was virtually absent for the linear polystyrene. It was suggested by Ferri and Lomellini [44] that this difference is due to branching, but it could just as easily be due to differences in the molecular weight distributions (both peak and breadth) of these two polymers. The influence of branching architecture on r\+ was performed by Ramsey et al. [46] by investigating the rheological properties of linear, three-arm star, and H-shaped polystyrene of similar molecular weight and polydispersity. The polystyrenes included in the investigation included three linear polymers ranging in molecular weight Mw = 175 000–430 000 g/mol, three threearm stars with Mw = 200 000–450 000 g/mol, and an H-shaped polymer with Mw = 430 000 g/mol. (It should be noted that the polydispersities of these polymers were relatively low and existed across the range M w /M n = 1.03-1.21.) Ramsey et al. [46] reported that on comparison of the three topologies with similar molecular weight (Mw w 450 000 g/mol), the H-shaped polymer exhibited a slightly higher value of q+ than did the three-arm star polymer. It was also reported that the H-polymer exhibited a lower value of »/+ than did the linear polymer. These extensional rheology results follow the same shear viscosity trends as those described earlier for polystyrenes of similar molecular weight and branching architecture. Interestingly, none of these polymers exhibited the same level of extensional strain hardening as did the randomly branched polystyrene studied by Ferri and Lomellini [44]. As the polymers investigated by Ramsey et al. [46] had higher molecular weight than the randomly branched polystyrene studied by Ferri and Lomellini [44], one is led to conclude that very high levels of extensional strain hardening is predominately due to polydispersity. Specifically, it appears that significant extensional strain hardening in polystyrene requires the presence of a high molecular weight component within the molecular weight distribution and not the presence of branching. At least this appears to be true across the ranges of molecular weight, molecular weight distributions, and branching architectures included in the various studies discussed in this chapter. Miinstedt [47] performed an extensive investigation exploring the influence of molecular weight and molecular weight distribution on the extensional rheological properties of linear polystyrenes. Four polystyrene samples (referred to as PS-I - PS-IV) were included in his investigation. The molecular weights of these samples were Mw = 74000, 39000, 353000, and 219 000 g/mol, respectively. The polydispersity of these samples were Mw/Mn = 1.2, 1.1, 1.9, and 2.3, respectively, but as the shapes of the molecular weight distributions of these four polystyrenes were so different from each other, polydispersity is not an adequate descriptor of the molecular weight distributions. PS-I possessed a narrow molecular weight distribution but with the inclusion of a high molecular weight tail. PS-II possessed the lowest molecular weight of the four samples studied but it also possessed a bimodal molecular weight distribution with a distinct high molecular weight component. PS-III possessed the highest

576

K. A. KOPPI AND D. B. PRIDDY

molecular weight of the four samples but its molecular weight distribution also possessed a significant low molecular weight component. Finally, PS-IV was a commercial polystyrene resin with a relatively broad molecular weight distribution. Uniaxial extensional measurements of ne+ as a function of time were reported by Munstedt [47] for PS-I and PS-II at T = 130°C across extension rates of de/dt = 0.001-0.4 s"1. Even though PS-II possessed a lower molecular weight than PS-I, PS-II exhibited a significantly greater degree of extensional strain hardening than PS-I. The degree of strain hardening exhibited by PS-II was as large as that reported for LDPE IUPAC A [48,49]. A similar set of measurements were performed by Munstedt [47] for PS-HI and PS-IV. Owing to the higher viscosity of these samples coupled with the limited achievable range of tensile stress associated with the instrument, uniaxial extensional measurements were performed by Munstedt [47] for PS-IH and PS-IV at T = 160°C across extensional rates of de/dt = 0.00075-0.3s"1. Despite the fact that these two materials have much higher molecular weights than PS-II, they do not exhibit the same level of extension strain hardening as PS-II. In fact, PS-III exhibits no extensional strain hardening at all and the level exhibited by PS-IV is not only less than that exhibited by PS-I but also that exhibited by PS-I. It is believed that the PS-HI and PS-IV exhibit little and no extensional strain hardening, respectively, owing to the presence of significant levels of low molecular weight species within the molecular weight distribution. Both of these polystyrenes contain high molecular weight species but their contributions to extensional strain hardening are diluted by the presence of the low molecular weight species. PS-I and PS-II do not contain significant levels of these low molecular weight species and for this reason the high molecular weight species within their molecular weight distributions can contribute to extensional strain hardening. It is interesting that the peak of the PS-I monomodal molecular weight distribution nearly matched the position of the secondary high molecular weight peak of the PS-II bimodal molecular weight distribution, yet it was PS-II that exhibited the high degree of extensional strain hardening - a level similar in magnitude to that exhibited by branched polyolefin materials. Therefore, it appears as if the magnitude of extensional strain hardening is not directly related to the level of high molecular weight species for polystyrenes such as PS-I and PS-II that do not contain significant levels of low molecular weight species. Rather, it appears as if the shape of the distribution plays a significant role in determining the degree of extensional strain hardening.

BRANCHED POLYSTYRENE

4

577

CONCLUSIONS

The first half of this chapter described various chemistries that can be used for the preparation of polystyrene resins with long chain branches and the second half discussed the influence of various branching architectures on the rheological properties of polystyrene. As was discussed near the beginning of this chapter, the polymer processing performance of certain polymers (most notably polyolefins) is improved dramatically by the introduction of long chain branching. One of the common features of these polymers is that they possess a relatively low entanglement molecular weight. As was discussed in depth in this chapter, in order for the presence of branching to increase the melt elasticity of a polymer, those branches have to be sufficiently entangled. It was also discussed that the greatest increase in melt elasticity occurs when the branched polymer possesses more than one branching point per molecule with the polymer segments between branch points being sufficiently entangled as well. These features, combined with the fact that polystyrene has a relatively high entanglement molecular weight, result in the potential benefits of branching lying outside the reach of commercially viable polystyrene resins due to the incredibly high molecular weights that would be required. The distance between the glass transition (Tg) and melting point for semicrystalline polyolefin resins makes the presence of long chain branches a virtual requirement for certain applications as a method for reaching the required level of melt elasticity. Polystyrene, however, with its accessible glass transition temperature, offers another route for reaching such a required level of melt elasticity, namely a reduction in melt temperature close to Tg. Finally, when the work of Munstedt [47] is juxtaposed with the various rheological investigations of branched polystyrenes that have been described within this chapter, it becomes clear that the best route for developing a polystyrene resin with optimal processing performance will not be achieved via branching but rather through the proper tailoring of the molecular weight distribution.

REFERENCES 1. 2. 3. 4.

Priddy, D. B. Adv. Polym. Sci. 1994, 114, 69. Niki, E.; Kamiya, Y. J. Org. Chem. 1973, 38, 1403. Stimming, U.; Durning, C. Macromolecules 1993, 26, 3271. Zhu, S.; Hamielec, A. Makromol. Chem., Macromol Symp. 1993, 69, 247.

578 5. 6. 7. 8.

K. A. KOPPI AND D. B. PRIDDY

Pike, W.; Priddy, D.; Vollenberg, P. US Patent 5663252, 1997. Matsubara, T.; Ito, N.; Ishida, Y.; Iwamoto, M. US Patent 4376847, 1983. Izumida, K.; Okumura, R. US Patent 5191040, 1993. Tung, L. H.; Hu, A.; McKinley, S.; Paul, A. J. Polym. Sci., Polym. Chem. Ed. 1981, 19, 2027. 9. Priddy, D. B. US Patent 6156855, 2000. 10. Watanabe, Y.; Ishigaki, H.; Okada, H.; Suyama, S. Chem. Lett. 1993, 1089. 11. Cummings, C; Hathaway, P. US Patent 5455321, 1995. 12. Hawker, C. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 1456. 13. Delassus, S. L.; Howell, B. A.; Cummings, C. J.; Dais, V. A.; Nelson, R. M.; Priddy, D. B. Macromolecules 1994, 27, 1307. 14. Tinetti, S. M.; Faulkner, B. J.; Nelson, R. M.; Priddy, D. B. J. Appl. Polym. Sci. 1997, 64, 683. 15. Hahnfeld, J. L.; Pike, W. C.; Kirkpatrick, D. E.; Bee, T. G. ACS Symp. Ser. 1998, 696, 167. 16. Flory, P. Principles of Polymer Chemistry, Ithaca University Press, Ithaca, NY, 1953. 17. Zimm, B.; Stockmayer, W. J. Chem. Phys. 1949, 17, 1301. 18. Roovers, J. Branched Polymers, Wiley, New York, 1989, Vol. 2. 19. Zimm, B.; R., K. J. Polym. Sci. 1959, 37, 19. 20. Ham, J. J. Chem. Phys. 1957, 26, 625. 21. Roovers, j.; Toporowski, P. J. Appl. Polym. Sci. 1974, 18, 1685. 22. de Gennes, P. J. Chem. Phys. 1971, 55, 572. 23. Doi, M.; Edwards, S. Theory of Polymer Dynamics, Oxford Press, Oxford, 1986. 24. McLeish, T.; Milner, S. Adv. Polym. Sci. 1999, 143, 195. 25. Kraus, G.; Gruver, J. J. Polym. Sci. 1965, A3, 105. 26. Graessley, W.; Masuda, T.; Roovers, J.; Hadjichristidis, N. Macromolecules 1976,9, 127. 27. Quack, G.; Fetters, L. ACS Polym. Prepr. 1977, 18, 558. 28. Hadjichristidis, N.; Roovers, J. Polymer 1985, 26, 1087. 29. Fetters, L.; Kiss, A.; Pearson, D.; Quack, G.; Vitus, F. Macromolecules 1993, 26, 647. 30. Masuda, T.; Ohta, Y.; Yamauchi, T.; Onogi, S. Polym. J. (Tokyo) 1984, 16, 273. 31. Roovers, J. J. Non-Cryst. Solids 1991, 131-133, 792. 32. Pearson, D.; Helfand, E. Symp. Faraday Soc. 1983, 18, 189. 33. Roovers, J. Polymer 1985, 26, 1091. 34. Roovers, J. Macromolecules 1984, 17, 1196. 35. Graessley, W.; Roovers, J. Macromolecules 1979, 12, 959. 36. McLeish, T.; Allgaier, J.; Bick, D.; Bishko, G.; Biswas, P.; Blackwell, R.; Blottiere, B. Macromolecules 1999, 32, 6734. 37. Pannell, J. Polymer 1971, 12, 558. 38. Pannell, J. Polymer 1972, 13, 2. 39. Roovers, J.; Graessley, W. Macromolecules 1981, 14, 766. 40. Masuda, T.; Nakagawa, Y.; Ohta, Y.; Onogi, S. Polym. J. 1972, 3, 92. 41. Masuda, T.; Ohta, Y.; Onogi, S. Macromolecules 1986, 19, 2524. 42. Bueche, F. J. Chem. Phys. 1964, 40, 484. 43. Hempenius, M. A.; Zoetelief, W. F.; Gauthier, M.; Moeller, M. Macromolecules 1998, 31, 2299. 44. Ferri, D.; Lomellini, P. J. Rheol. (N. Y.) 1999, 43, 1355. 45. Khan, S.; Prudhomme, R.; Larson, R. Rheol. Acta 1987, 26, 144.

BRANCHED POLYSTYRENE

579

46. Ramsey, R.; Hahnfeld, J.; Pike, W.; Welsh, G.; Lewis, C.; Zawisza, M; Quirk, R.; Carriere, C. Annu. Tech. Conf. Soc. Plast. Eng., 54th 1996, 1124. 47. Miinstedt, H. J. Rheol. 1980, 26, 847. 48. Meisner, J. Rheol. Acta 1971, 10, 230. 49. Laun, H.; Munstedt, H. Rheol. Acta 1976, 15, 517.

This page intentionally left blank

25

'Super Polystyrene' Styrene-Diphenylethylene Copolymers G. E. McKEE, F. RAMSTEINER, W. HECKMANN AND H.GAUSEPOHL BASF AG, Ludwigshafen, Germany

1

INTRODUCTION

As can be seen in earlier chapters of this book, one of the main disadvantages of polystyrene is that its long-term service temperature lies below 100 °C. This has been overcome by producing stereoregular polymers in the form of isotactic and syndiotactic polystyrene, the latter being a commercial product. BASF has investigated additionally another way of increasing the service temperature, namely by incorporating bulky groups into the polymer backbone. To this end, the comonomer 1,1-diphenylethylene (DPE) was chosen owing to its similar chemical composition to styrene and also because phenyl groups are effective at raising the glass transition temperature (Tg) of polystyrene. The structure of the S/DPE polymer is shown in Figure 25.1. Before BASF investigated this product, Quirk and Hsieh [1], Yuki and coworkers [2,3] and Fischer [4] carried out investigations with this monomer. The first two used the anionic polymerization mechanism and Fischer tried to copolymerize this monomer using free radical polymerization. In the latter case the yields were very low. The use of S/DPE blocks in thermoplastic elastomers [5] has also been briefly described. Some of the work carried out at BASF has been published in a recent review article [6]. Owing to the enhanced thermal properties of this copolymer in comparison with atactic polystyrene - the glass transition temperature increases up to 180 °C, depending Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy © 2003 John Wiley & Sons Ltd

582

G. E. McKEEETAL Styrene-DPE-Copolymer

Figure 25.1

DPE

Structure of styrene-diphenylethylene (S/DPE) copolymers

on the DPE content of the polymer - the name 'Super Polystyrene' was coined by BASF. In this chapter we cover the preparation, mechanical properties, and applications cauons itions of of S/DPE s/DPE polymers. Polymers Owing totoits brittle its nature iiaiuic this mis product piuuw. should be rubber modified, modified so special embasis willwillbebe placed placed on this this aspect. aspect rubber so special emphasis on

2 2.1

PREPARATION OF DPE MONOMERS AND POLYMERS 1,1-DIPHENYLETHYLENE MONOMER SYNTHESIS

Commercially, the best way to prepare 1,1-DPE is probably to react styrene and benzene with one another and then to dehydrogenate the resulting 1,1-diphenylethane to 1,1-diphenylethylene. This has been developed to the pilot plant stage in BASF [7]. On a laboratory scale, 1,1-diphenylethylene may be prepared using the Grignard reaction between phenylmagnesium bromide and acetophenone [6]. 2.2

S/DPE POLYMER SYNTHESIS

Owing to the stability of the DPE radical, S/DPE polymers cannot be prepared using free radical polymerization, but they can be easily produced using the anionic polymerization technique. The copolymers are, however, limited to a maximum DPE content of 50 mol% because two consecutive DPE units are not possible in the polymer chain for steric reasons. This leads to a reactivity ratio r

DPE(Kdd/Kds) = 0.

The polymer can be produced in either cyclohexane or ethylbenzene using butyllithium, preferably sec-butyllithium. Since the addition of a styrene molecule to a DPE chain end is slow (decrease of resonance stability) and the addition of a DPE monomer to a growing styryl chain is fast (increased resonance stability), the polymerization rate decreases with increasing DPE content in the polymerizing monomer mixture. The reactivity ratio in cyclohexane was found to be between 0.44 (50ºC) and 0.72 (70 °C).

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE

COPOLYMERS 583

S/DPE polymers are at present not commercially available, but in BASF the polymers have been produced in a miniplant. To avoid the production of molecules with varying DPE content along the polymer chain, the polymer was produced in a continuous flow stirred-tank reactor (CSTR) coupled with a plug flow reactor (Figure 25.2) This gives a product with a homogeneous composition but with a broad molecular weight distribution. This results from the different residence times of the growing polymer chains in the CSTR.

3

PROPERTIES OF STYRENE-DIPHENYLETHYLENE POLYMERS

S/DPE polymers are amorphous like polystyrene and, as a result of the bulky DPE groups, the polymer chains are stiffer than the styrene homopolymer, leading to a higher Tg and thus a higher Vicat softening point and continual service temperature. This can be seen in Figure 25.3 where the Tg and Vicat softening temperature are shown as a function of the DPE content. At the maximum possible DPE content of 50mol%, a Vicat temperature of 175 °C is obtained. The increase in the Young's modulus can be up to 30%, depending on the DPE content, in comparison with general-purpose polystyrene (GPPS) (Figure 25.3). Even at 80 °C a Young's modulus as high as that of GPPS at room temperature can be observed. The disadvantage, however, of stiffer polymer chains is reduced mobility. This has as a consequence a higher resistance to deformation processes in the bulk and therefore processing should be carried out at a temperature higher styrene 427 g/h DPE 273 g/h ethyl benzene s-BuLi 0.15 M 20 g/h

turnover 95%

1.5L

104 HP M [g/mol]

700 g/h

turnover > 97% CSTR: 3L

plug flow reactor: 0.15L

degassing unit

Figure 25.2 Continuous production of S/DPE in a miniplant (39 wt% DPE; Tg = 149°C) [6]

584

G. E. McKEEETAL. increase of softening temperature with DPE content

temperature [°C] 180 n

increase of stiffness with DPE content modulus [MPa 4500-1 4000-

160-

3500140-

30002500

2000 10 PS-

20 30 40 50 DPE content (wt%)

PS

20 30 40 50 DPE content (wt%)

60

Figure 25.3 Dependence of softening temperature and stiffness on DPE content. Reproduced with permission from Gausepohl et al., Designed Monomers and Polymers, 3(3), 299 (2000) with permission of VSP, The Netherlands

than for polystyrene. Thermogravimetric analysis (TGA) measurements (Figure 25.4) in air show that the initial degradation of S/DPE polymers is less than for polystyrene at 300 °C; however, when degradation has begun, it proceeds more quickly for S/DPE than for polystyrene. It is therefore recommended that S/DPE polymers should not be processed too long above 300 C. Under nitrogen the degradation temperature of the S/DPE polymer is, however, higher and considerably above 300 ºC (Figure 25.4). The toughnesses of S/DPE and GPPS are similar and often insufficient. An improvement can be obtained either by adding glass fibres or by rubber modification. Adding glass fibres increases toughness by increasing the resistance to crack propagation perpendicular to them; however, the energy-dissipating deformation processes inherent to the material are restricted, thus reducing toughness. Which of these opposite effects dominates depends on the details, such as fibre concentration and orientation, temperature, test method, and the properties of the matrix. In Table 25.1 a comparison between S/DPE(30) (i.e. 30 wt% DPE) and syndiotactic polystyrene (sPS), both products containing 30 % glass fibres, is shown. It can be seen that the mechanical properties of the S/DPE product are slightly superior to those of sPS. Resistance to polar solvents, however, is probably better in sPS than in S/DPE polymers owing to its partially crystalline nature. A more effective toughening of brittle thermoplastics is often achieved using rubber modification. This is discussed in more detail in Section 5.

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 585 under air

under nitrogen

200

300

200

400

temperature [°C]

300

400

500

temperature [°C]

Figure 25.4 Thermal stability of S/DPE copolymers: TGA measurements (heating rate 10°C/min) Table 25.1 (GF)

Comparison of sPS and super polystyrene containing 30% glass fibre

Property

Units

S/DPE(30) (30 % GF)

SPS (30 % GF)

E-modulus

MPa MPa % kJ/m2 kJ/m 2

10500 108 1.3 19.8 7.2

9800 103 1.1 14.7 4.9

Tensile strength at break Elongation at break Charpy(179eU/23°C) Charpy(179eA/23°C)

4

BLENDS OF S/DPE POLYMERS

S/DPE polymers are compatible with one another when their DPE contents differ by less than 15 wt%. S/DPE polymers containing less than 15 wt% of DPE are also compatible with GPPS and with the amorphous phase of sPS. This has been demonstrated using DSC and transmission electron microscopy [6]. In Figure 25.5, the TEM photographs and DSC curves for sPS-S/DPE blends (1:1) are shown for S/DPE copolymers with 15 and 45% DPE. In the case of S/DPE(15) (i.e. 15 % DPE), only one Tg is observed, at 114 °C, but in the case of S/DPE(45) two Tgs, at 101 and 158°C, are present. The melting point and crystallization rate of the sPS at 270 °C are not affected by either S/DPE copolymer.

586

G. E. McKEEE7V\L.

Figure 25.5 Compatibility of sPS and S/DPE copolymers. Reproduced with permission from Gausepohl etal., Designed Monomers and Polymers, 3(3), 299 (2000) with permission of VSP, The Netherlands

5

RUBBER MODIFICATION OF S/DPE POLYMERS

In the chapter on impact modification of sPS, the basic mechanisms of the rubber toughening of styrenics are described. As was explained earlier, in S/DPE polymers the higher Tg and stiffness are a result of the restricted mobility of the S/DPE chains. This, however, is linked to increased restriction of the energydissipative plastic deformation in bulk. This deficiency can be considerably improved, thus leading to a large increase in the product toughness, by incorporating small rubber particles in the S/DPE polymer. The object of rubber modification is twofold, first to increase extensively in the stress fields of the rubber particles the deformation mechanisms inherent to the matrix material, and second to decrease the dilatational stress fields in the product by internal voiding in the rubber particles. The latter effect requires at least semi-compatibility

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 587

between the rubber particles and the matrix to cause voiding in the rubber particles instead of triggering harmful cracks at the rubber/matrix interface. In the following sections the chemical and physical aspects relevant to the compatibility of the rubber particles with the polymer matrix, the deformation modes, and the toughening mechanism at room temperature are discussed. 5.1

MODIFIED HIGH-IMPACT POLYSTYRENE (HIPS) PROCESS

In the HIPS process, polybutadiene is dissolved in styrene and the styrene is polymerized using free radical polymerization. Thereby some of the growing styrene polymer chains graft on to the polybutadiene rubber, leading to the well known salami particles after phase inversion. The same process for S/DPE is not possible since the S/DPE monomer mixture is not polymerizable by free radical polymerization. The solution is to polymerize the S/DPE monomers anionically, but since no grafting reaction occurs the compatibilization of the two phases must be obtained via other means. This can be achieved by using a block copolymer consisting of an S/DPE and a polybutadiene block. The flow diagram for a pilot plant in BASF is shown in Figure 25.6. Thereby the S/DPEbutadiene block copolymer is discontinuously prepared in toluene and then continuously mixed with styrene and DPE monomers, butyllithium and a S/DPE

B

solvent

Figure 25.6 Direct synthesis of impact-modified poly(S/DPE). Reproduced with permission from Gausepohl et al., Designed Monomers and Polymers, 3(3), 299 (2000) with permission of VSP, The Netherlands

588

G. E. McKEEETAL.

retarder [6]. The styrene is then polymerized in a kettle-tower cascade. After termination, the unpolymerized monomers are removed by degassing. The morphology of the impact-modified S/DPE polymer is also shown in Figure 25.6. As can be seen, the rubber-modified S/DPE has a similar morphology to HIPS. The Vicat softening temperature and the viscosity of the S/DPE product are higher and the toughness (notched impact strength) is lower than for the equivalent HIPS product. 5.2

CORE-SHELL IMPACT MODIFIERS

Blends consisting of crosslinked butyl acrylate particles (core) grafted with styrene (shell) have been reported in the literature. The particles were prepared using microsuspension [8] (0.5-2.0 u,m) or emulsion (particle diameter 0.12 |xm) polymerization processes [9]. 5.2.1

Microsuspension (MSP) Rubber Particles [10]

TheTEM images (Figure 25.7) show the rubber particles (a) MSP 1, (b) MSP2, and (c) MSP 3 mixed into the S/DPE matrix containing 15 % DPE. Figure 25.8 shows the morphology of MSP 1 in S/DPE(30). The blends were injection moulded to dumbbell test pieces. In all cases the grafted rubber concentration was 36 wt%. These rubber particles consist of 40 % styrene as the outer shell, for compatibility with the S/DPE copolymer, and 60 % crosslinked butyl acrylate rubber as the

Figure 25.7 TEM of rubber-modified (36%) injection moulded S/DPE(15): (a) MSP 1, (b) MSP 2, and (c) MSP 3. Cryo ultra thin sections stained with RuO4

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS

Figure 25.7

589

(Continued)

particle core. The medium size for the particles in suspension was determined by light scattering to be 1.2 um for MSP1, 0.4 um for MSP2, and 2.2 (um for MSP3. The deformation mechanism in S/DPE is the same as in polystyrene, namely by crazing [11,12]. Figure 25.9 shows an SEM picture of such a single craze on

590

G. E. McKEEETAL

Figure 25.8 TEM of 36 % MSP 1 rubber, mixed in S/DPE(30) and injection moulded. Cryo ultra thin section stained with RuO4

Figure 25.9 SEM of a craze on the surface of an S/DPE(46) specimen

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 591

the surface of the deformed S/DPE(46) product containing no impact modifier. For increased toughness this crazing must be stabilized by rubber particles. In Figure 25.10 a TEM image of the craze structure after deformation is reproduced for S/DPE(30) with 36% MSP1 and 10% of an S-S/DPE block copolymer, whereby the S/DPE block is compatible with the S/DPE(30) and the polystyrene block with the graft shell of the impact modifier. Crazes run from particle to particle and some of the rubber particles are voided. At room temperature, deformation by voiding in the rubber particles and crazes after deformation beyond the yield stress amounts to about 60% of the total deformation, as determined by volume measurements during deformation, as described by Ramsteiner [13]. To demonstrate the effect of rubber modification on the impact toughness, the notched impact strength and the energy release rate were measured [14]. The results are summarized in Table 25.2. Increasing the DPE concentration in the copolymer from 15 to 30 % seems to embrittle the material slightly, caused by the lower mobility of molecules with higher DPE content. The addition of 10 % of an S-S/DPE block copolymer as a compatibilizer between the styrene graft shell of the impact modifier and the S/ DPE matrix brings an increase in toughness. The lowest impact modification was with the rubber MSP2, which had the smallest particle size (Figure 25.7b). The energy release rate was measured according to ISO CD 17281 as described in the ESIS book [14]. Specimens cut from the middle part of the

Figure 25.10 TEM of crazes in rubber-modified (MSP1) S/DPE(30) with a 10% content of S-S/DPE. Cryo ultrathin section stained with OsO4/RuO4

592

G. E. McKEEETAL

Table 25.2 Notched impact strengths, energy release rates at failure and total energy release rates on impact (2.9 m/s) for rubber-modified S/DPE Matrixa 64wt%

Rubber Notched impact Energy release rate Energy release rate 36 wt% strength (kJ/m 2 ) b (kJ/m 2 ) (peak) (kJ/m 2 ) (total)

S/DPE(15) S/DPE(30) S/DPE(15) S/DPE(15) 54% S/DPE(30) + 10%S-S/DPEC

MSP1 MSP1 MSP2 MSP3 MSP1

a b c

12.5 9.7 2.9 10.3 12

11 7 9 11 9

29 22 19 42 22

Mw approximately 200000g/mol. determined according to ISO 179. S-S/DPE block copolymer, S/DPE block compatible with S/DPE(30).

dumbbell test pieces with different notch lengths were fractured in impact bending in the equipment for the Charpy tests. The initial notch tips were sharpened by tapping a razor blade into the notch. The energy release rate, G, is given by the slope of the plot of the deformation energy U versus the geometric quantity BWcfr: G=U/BW(f)

(1)

where B is the thickness of the specimen, W the width and a geometric factor, which takes into account the ratio of notch length and width of the specimen. The values for as a function of the geometrical data are given in a standard (ISO 13586). In Figure 25.11, the deformation curve recorded as force versus deformation time during impact testing of the rubber-modified S/DPE(30) with MSP1 is reproduced. The deformation speed was 2.9 m/s. As expected for tough materials, this product does not show linear behaviour before fracture. Therefore, linear elastic fracture mechanics cannot be strictly applied, and only apparent values are obtained. This method also allows the division of the whole fracture energy into a part responsible for the failure and the remaining part which is the crack propagation energy. This assumes that the peak stress is regarded as the failure point. Using this assumption, the energies up to the peak force (peak) and the total (total) fracture energies are plotted (Figure 25.12) as a function of the geometric parameters according to Equation (1). For this example, the critical apparent energy release rate up to the peak force is 7.3 kJ/m 2 and the total critical energy release rate (breakage and crack propagation) is 21.8kJ/m 2 . On the basis of the previous assumption for analysing the data, one-third of the fracture energy is needed for failure and two-thirds is additionally needed for the nonlinear crack propagation. For = 21.4mm 2 (Figure 25.12), which corresponds to the short crack length

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 593

0

0.2

0.4

0.6

0.8 1 1.2 time [ms]

1.4

1.6

1.8

2

Figure 25.11 Force time record of rubber modified (MSP1) S/DPE (30%) during three-point bending impact (deformation speed 2.9 m/s) 600

Total 500

400

300-

200-

Peak

100 -

10 BW

15

20

25

[mm2]

Figure 25.12 Fracture mechanics plot of the deformation energy U on impact as a function of the geometric parameter BW for rubber-modified (MSP1) S/DPE(30%)

594

G. E. McKEEETAL

with a notch length (a) of 2 mm, the value for the total energy is lower than that extrapolated using the longer crack lengths. (For the Charpy test a 2 mm initial notch length is employed.) This implies that the energy-dissipating deformation processes in the thicker part of the specimen beneath the initial short notch are less activated than in thinner parts beneath longer notches. The values for the other systems are summarized in Table 25.2, which also gives the energy release rates from total energy and the initial values at the peak for the long notch lengths (small BW values). For all the products the total energy for fracture is about three times the energy up to the maximum force peak (Table 25.2). The product containing the smallest particles, MSP2, is less tough than the others. The best impact modifier, as judged by the portion attributed to the deformation energy, seems to be the product with the largest particles, namely MSP3. In this case, however, the toughness is reduced in the thicker parts of the specimen beneath the notch tip. This finding corresponds to the model that larger particles are more effective than smaller ones in toughening materials with tendency to crazing [15].

5.2.2

Emulsion Rubbery Particles [10]

The rubber particles produced by the emulsion process have diameters of about 0.12 um. They consist of crosslinked poly(butyl acrylate) grafted with polystyrene (60:40 weight ratio). After blending with S/DPE(15) and injection moulding, large agglomerates together with many smaller particles are present, as is shown in the TEM in Figure 25.13. In the deformed specimen, after the notched impact test, the agglomerates are highly elongated and dispersed. The small rubber particles are highly orientated, and some crazes are running through the material perpendicular to the tensile direction (Figure 25.14). In some regions, rows of voids in the tensile direction are also discernible. With small particles, rubber toughening in crazing material is not expected. Indeed, with these small emulsion particles, the notched impact strength was 2.3 kJ/m 2 , only about three times the value for the neat matrix, and the energy release rates (Figure 25.15) were as low as 1 kJ/m 2 up to the peak point and 1.8 kJ/m 2 for the total energy. Both failure energy and plastic deformation were low. These results are supported by the observations using the MSP 2 impact modifier (see Section 5.2.1), where the relatively small 0.4 um particles also have a poor toughening effect in S/DPE and also by the general theory that large particles are necessary for energy dissipation via crazes in polymers consisting of stiff chains. It also indicates that shearing processes are not operating to any large extent in the energy dissipation in S/DPE under impact conditions.

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 595

Figure 25.13 TEM of emulsion polymerized rubber blended with S/DPE(15)

Figure 25.14 TEM of crazes in S/DPE(15) modified with emulsion polymerized rubber. Cryo ultra thin section stained with OsO4/RuO4

596

G. E. McKEEEr/lL.

Total

Peak

10

15

BW [mm2]

Figure 25.15 Fracture mechanics plot of the deformation energy in impact as a function of BW for S/DPE(15) modified with rubber particles prepared in emulsion

5.3

TRI-BLOCK COPOLYMER OF STYRENE-HYDROGENATED BUTADIENE-STYRENE [S-BH-S] [6]

S/DPE copolymers can also be impact modified using triblock copolymers having as the centre block the rubber phase [6]. Typical examples are styrenebutadiene-styrene block copolymers where the butadiene phase is preferably hydrogenated S-B(H)-S, to remove the double bonds. This is advisable owing to the high processing temperatures involved in the S/DPE processing. TEM images of blends of S-B(H)-S with S/DPE(15) and S/DPE(30) are shown in Figure 25.16. The largest particles are about 1 um, which seems promising for toughening. Within the particles the typical lamellar structure of block copolymers can be seen. The white lines are the unstained hydrogenated butadiene and the dark regions the stained polystyrene phases.

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS

Figure 25.16 TEM of S/DPE modified with a S-B H -S block (a) S/DPE(15) matrix; (b) S/DPE(30) matrix

597

copolymer:

The glass transition temperature according to dynamic torsional measurements is near — 60 °C (Figure 25.17) for the rubbery phase, so that toughening above this temperature can be expected. The glass transition temperature of the

598

G. E. McKEE ETAL. 102 2.0X109

2.0X I0 4 ' • -100.0

-50.0

0.0

i . . . . i 50.0 100.0

150.0

200.0

I io-3 250.0

Temp [°C]

Figure 25.17 Dynamic mechanical properties of rubber-modified S-B H -S S/DPE(30)

matrix is increased to 150°C by copolymerization of 30% DPE with 70% styrene. On deformation, intensive crazing in the matrix and extensive voiding in the rubber particles occur, as can be seen from the TEM image in Figure 25.18. Volume measurements during deformation indicate that voiding in the rubber particles and crazes accounts for about 60% of the total deformation. In contrast to HIPS, hysteresis measurements (Figure 25.19a) on injection moulded S/DPE specimens modified with S-B(H)-S rubber do not reveal craze deformation by the entropy elastic behaviour in the second run. The crazes seen in the TEM (Figure 25.18) are produced only in the first deformation. Only after tempering at 200 °C for 30min after the first run is crazing observed as indicated by an S-shaped second run (Figure 25.19b), as is often observed for ABS without this tempering step [16]. In already orientated material the entropic elastic deformation of crazes seems to be impeded, presumably owing to the bulky side group in the chain. Using the fracture mechanics method described in Section 5.2.1 for the microsuspension modifiers, the fracture energy U as a function of the geometric parameter BW in three-point bending impact for the injection moulded blends of S-B(H)-S block copolymers with (a) S/DPE(15) and (b) S/DPE (30) is shown in Figure 25.20. From the slope of the lines in Figure 25.20a, the critical energy release rate up to failure at the peak force yields 10.8 kJ/m 2 , and for the total deformation energy, 35.7kJ/m 2 for the S/DPE(15) blend. For the notched energy, 19.5kJ/m 2 was measured. In the case of the stiffer S/DPE(30) matrix, where less compatibility is expected, the critical energy release rates

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 599

Figure 25.18 TEM of crazes in S/DPE(30) modified with S-B H -S rubber. Cryo ultra thin sections stained with OsO4/RuO4 decrease to 8.7kJ/m 2 (peak) and 24.4kJ/m2 (total) for injection moulded specimens (Figure 25.20b). This indicates that the reduced toughness observed with the S/DPE(30) matrix in comparison with the S/DPE(15) matrix is a result of reduced dissipative deformation energy, which in turn results from the lower mobility of the matrix with higher bulky DPE sequences. The notched impact strength also drops to 14 kJ/m 2 for the matrix with the higher DPE content.

5.4

TRI-BLOCK COPOLYMERS OF S/DPE-HYDROGENATED BUTADIENE-S/DPE [6]

As stated earlier, S/DPE copolymers are compatible with GPPS up to a DPE content of about 15wt%. This means that on modifying S/DPE(>15) with S-B(H)-S the impact modification decreases owing to decreased compatibility between the S/DPE matrix and the styrene blocks of the S-B(H)–S impact modifier. S-DPE, however, is prepared anionically and using this polymerization mechanism S/DPE-butadiene-S/DPE block copolymers can be prepared. Thus the S/DPE blocks can be tailor-made to be compatible with the S/DPE polymer matrix. For compatibility, the DPE content of the blocks and the matrix should not differ by more than about 15 %. As with the S-B(H)-S block copolymers, the double bonds of the butadiene phase should be removed by hydrogenation.

600

G. E.McKEEEETAL stress [MPa] 50-

40

(a)

0

stress [MPa] 50-

40

30-

20-

10 strain [%]

(b)

10

Figure 25.19 Tensile stress-strain curves during loading-unloading-reloading cycle for rubber-modified [S-BH-S] S/DPE(15): (a) injection moulded; (b) tempered at 200 °C for SOmin in a dumbbell-type mould

6

THERMOPLASTIC ELASTOMERS [6]

The possibility of preparing S/DPE-hydrogenated butadiene-S/DPE block copolymers opens up the possibility of preparing a new class of high-temperature thermoplastic elastomers (TPE). By coupling the living di-block copolymer S/ DPE-Bu-Li using 1,4-butanediol diglycidyl ether, a block copolymer with the

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS

601

composition 16/68/16 (wt%) has been prepared [6]. Styrene was added to a solution of DPE and sec-BuLi in cyclohexane. The polymerization of the butadiene block was carried out after the addition of THF to the reaction mixture so as to produce a high 1,2-vinyl microstructure of 40 %. Since this polybutadiene is only slightly crystalline, the hydrogenation step can be easily carried out to a high conversion. The resulting ethene-butene block has a Tg of — 60 °C. The resulting TPE can either be used alone or blended with aliphatic oil and polypropylene. In the former case a higher tensile strength and elongation at break are obtained in comparison with the commercially available styrene-hydrogenated butadiene-styrene block copolymers, especially at high temperatures.

7

SUMMARY

S-DPE polymers are a new type of styrene-based polymers which are of interest where a higher heat distortion temperature is required than with polystyrene. 900

Total 800700

600500400300-

Peak 2001000 0

(a)

10

15

20

25

BW [mm2]

Figure 25.20 Fracture mechanics plot of the fracture energy u as a function of the geometric parameter BW4> in three-point bending impact for (a) the injection moulded blend S/DPE(15) with S-B H -S and (b) the injection moulded blend S/DPE(30) with

S-BH-S

602

G. E. McKEE ETAL

Total

Peak

100-

5

(b)

Figure 25.20

10

15

20

25

BW<J> [mm2]

(Continued)

With increasing DPE content the Tg increases and at a maximum DPE content of 50mol% the maximum Tg of 180°C is attained. The polymers are brittle but the impact strength can be considerably increased via rubber modification. Owing to its availability, S-B(H)-S is probably the rubber of choice. At present these polymers are not available commercially. The polymers can be used neat, i.e. as GPPS, or rubber modified by blending with block copolymers or rubber particles prepared in emulsion or microsuspension and also by using a modified HIPS process. Tri-block copolymers with the middle block being a hydrogenated butadiene phase and the end blocks composed of S/DPE are of interest as the base for thermoplastic elastomers or as a blend component with oil and polypropylene. Finally, by the correct choice of block morphology, a transparent TPE is possible, similar to Styrolux* (a BASF styrene block copolymer) (Figure 25.21). The uses of such products are numerous, especially for housings and in the electrical sector.

'SUPER POLYSTYRENE'-STYRENE-DIPHENYLETHYLENE COPOLYMERS 603

Figure 25.21

The world of styrene-1,1 -diphenylethylene copolymers

REFERENCES 1. Quirk, R. P., Hsieh, H. L., Anionic Polymerisation; Marcel Dekker, New York (1996). 2. Yuki, H., Kosai K., Muharshi, S., Hotta, J., /. Polym. Sci, Part C: Polym. Lett. 2, 1121 (1964). 3. Yuki, H., Kosai K., Okamoto, Y., Murahashi, S., Bull. Chem. Soc. Jpn. 40, 2659 (1967). 4. Fischer, J. P., Makromol Chem. 155, 227 (1972). 5. Trepka, W. J., Polym. Lett. 8, 499 (1970). 6. Gausepohl, H., Oepen S., Knoll, K., Schneider, M, McKee, G. E., Loth, W., Des. Monom. Polym., 3, 299 (2000) 7. Knoll, K., BASF AG, personal communication. 8. WO 9846678. 9. DOS 1 99 103 39. 10. McKee, G. E., Oepen S., Heckmann W., unpublished work. 11. Rabinowitz, S., Beardmore, P. CRC Crit. Rev. Macromol. Sci. 1, 1 (1972). 12. Kramer, J. E., Adv. Polym. Sci. 52/53, 1 (1983). 13. Ramsteiner, F., Polym. Test. 18, 401 (1996). 14. Moore, D. R., Pavan, A., Williams, J. G., Fracture Mechanics Testing Methods for Polymers, Adhesives and Composites, ESIS Publication 28, Elsevier Science, Amsterdam (2001). 15. Bucknall, C. B., Toughened Plastics, Applied Science, London (1977). 16. Ramsteiner, F., McKee, G. E., Heckmann, W., Fischer, W., Fischer, M., Acta Polym. 48, 553 (1997).

This page intentionally left blank

26

Ethylene-Styrene Copolymers Y. W. CHEUNG AND M. J. GUEST The Dow Chemical Company, Dow Texas Operations, Freeport, TX, USA

1

INTRODUCTION TO ETHYLENE-STYRENE COPOLYMERS

Polyethylene and polystyrene are two of the most commercially important and ubiquitous polymers, primarily because of their commercial value. Since the early days of polymer research there has been considerable interest to produce copolymers from ethylene (E) and styrene (S) because of both academic and business interests. Depending on the nature and type of polymerization chemistry, a variety of different molecular architectures can be produced. In addition to the different monomer distributions (random, alternating or blocky nature), there are possibilities for chain branching and tacticity in the chain microstructure. These molecular architectures have a profound influence on the melt and solid-state morphology and hence on the processability and material properties of the copolymers. Historically, it has been a formidable challenge to synthesize efficiently random ES copolymers with high levels of styrene incorporation and having homogeneous structures with uniform comonomer distribution and narrow molecular weight distribution. Developments published in the early 1990s demonstrated the ability of metallocene catalysts to polymerize and copolymerize efficiently a wide range of hydrocarbon-based vinyl monomers, including ethylene and vinyl aromatics such as styrene. In particular, INSITETM+ Technology, including developments of single-site, constrained geometry, addition polymerization catalysts [1–3] has provided an efficient route to produce essentially random copolymers based on ethylene and styrene, which are a novel Trademark of The Dow Chemical Company. Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy :r 2003 John Wiley & Sons Ltd

606

Y. W. CHEUNG AND M. J. GUEST

class of materials. Most research efforts to date have been focused primarily on the catalysts which could copolymerize ethylene and vinyl aromatic monomers, and the associated polymerization chemistry; a synopsis of this activity is presented in Section 2. Most of the available information on technology development is for the ethylene/styrene copolymers as produced by INSITE™ Technology, and subsequent sections focus primarily on the technology of these latter polymers. Information is included from a review on the structure, properties and applications of ethylene-styrene interpolymers [4], updated with more recent developments. Section 3 covers the structure-property relationships of ethylene-styrene interpolymers. Key characteristics of this novel class of copolymers include aromatic/olefin functionality, the glass transition temperature (7"g) of the copolymers and excellent processability. These characteristics suggest that the end-use applications for the copolymers will make use of them in formulated or engineered systems, and this is considered in more detail in Section 4. Potential application areas are also identified. The references cited provide more details on the materials and technology.

2 COPOLYMERIZATIONS OF ETHYLENE AND VINYL AROMATIC MONOMERS Historically, the production of ES copolymers was attempted by using the conventional routes to produce either polyethylene or polystyrene with the introduction of the alternate comonomer into the process. Various references can be found where attempts have been made to copolymerize ethylene and styrene together. Many reported copolymerizations produce heterogeneous polymers consisting of mixtures of copolymers and homopolymers; for example, US Patent 3 117945 [5] describes the production of a 'gross' copolymer, composed of various fractions which could be separated based on their solubility parameter. ES copolymers containing low levels (< 4 mol%) of styrene incorporation have been prepared using both conventional Ziegler-Natta catalysts [6–8] and free radical processes [9]. Most of these synthetic routes generally yield heterogeneous polymers consisting of mixtures of copolymers and homopolymers [5], and do not provide commercially viable routes for their production. Around 1990, developments using metallocene catalysts demonstrated their ability to polymerize and copolymerize efficiently a wide range of hydrocarbon-based vinyl monomers, including ethylene, higher a-olefins, dienes and vinyl aromatics such as styrene. In particular, INSITE™ Technology, including developments of single-site, constrained geometry and addition polymerization catalysts, has provided an efficient route to produce ethylene-styrene interpolymers (ESI). The more popular catalyst types for the preparation of ethylene-styrene copolymers have been those based on titanocene complexes,

ETHYLENE-STYRENE COPOLYMERS

607

exemplified by dimethylsilyl(tert-butylamido)tetramethylcyclopentadienyltitanium dichloride. Many reported research and patent publications have focused on the use of these types of constrained geometry catalysts for the preparation of ethylene-styrene copolymers. Other catalysts of interest for copolymerization have been based on fluorenyl and indenyl ligand structures, such as those identified by Mitsui Toatsu [10]. Studies of catalyst systems that produce syndiotactic polystyrene provided the starting point for research at the University of Salerno [11], and much of their subsequent work on the production and characterization of ethylenestyrene copolymers has been referred to in a short review article by Pellecchia and Oliva [12]. A series of publications from the research group of Mulhaupt (University of Freiburg, Germany) documented the results from a BASFsponsored research project on ethylene-styrene copolymers. This program covered the copolymerization of ethylene with styrene using methylaluminoxane-activated half-sandwich complexes [13] and methylaluminoxane-activated bis(phenolate) complexes [14]. The influence of polymerization conditions on the copolymerization of styrene with ethylene using Me2Si(Me4Cp) (N-tBu)TiCl2/methylaluminoxane catalysts was studied [15], at temperatures ranging from 20 to 80 °C and with variation of styrene monomer concentration. The maximum styrene incorporation achieved was 36.6mol%, with styrene concentration found to have a strong influence on polymerization activity and copolymer molecular weight; limited differential scanning calorimetry (DSC) and dynamic mechanical thermal analysis (DMTA) data were reported for the copolymers. The fractionation and comprehensive characterization of an ES copolymer having 13.8 mol% copolymer S produced with the catalyst system of Serretz et al. [14] was described [16]. Analysis of the four copolymer fractions provided strong evidence of a single-site mechanism for the methylaluminoxane-activated metallocene catalyst. Metallocene-catalyzed ethylenestyrene polymerizations with olefinic termonomers [17] (propylene, 1-octene and norbornene) and with 1,5-hexadiene [18] were also part of the overall research program. Another subject which has received some interest has been the synthesis of alternating ethylene-styrene copolymers, or those copolymers which contain alternating (ES)n sequences. Suzuki et al. [19] described the preparation of alternating atactic (ES)n copolymers by the hydrogenation of poly(phenylbutadienes), produced by anionic polymerization. A group at Sumitomo [20] disclosed the production of ethylene-vinyl aromatic monomer copolymers having a high degree of alternating isotactic polymer chain sequences, although according to the examples given the copolymer was recovered from polymerization product by solvent/nonsolvent treatments. The same authors reported [21] that the alternating isotactic copolymer (Mw = 13000, M w /M n = 1.9) had a crystalline melting point of 116 °C. Izzo et al [22] produced and characterized low molecular weight model compounds of alternating ES copolymers with a

608

Y. W. CHEUNG AND M. J. GUEST

range of catalysts, and characterization seemed to confirm the isotactic structure of stereoregular copolymers. Arai et al. [23,24] showed that a copolymer having alternating polymer chain sequences which are isotactic and crystallizable can be produced when copolymerization is carried out at temperatures of 50 °C or below, using for example rac-[isopropylidenebis(l-indenyl)]zirconium dichloride. Although the presence of crystallizable sequences could contribute to properties such as mechanical strength, it appears from the patent literature [25] that annealing or exposure to solvent is desirable to develop the crystallinity more fully. This observation is analogous to the behavior of isotactic polystyrene: the polymer has a slow rate of crystallization that has hindered its commercialization, owing to the restricted ability to develop crystallinity under fabrication conditions [26]. An overview of the synthesis of ethylene-styrene copolymers has been compiled by Pellecchia and Olivia [12]. A short overview of ethylene-styrene interpolymer technology, including an identification of the most widely investigated catalysts cited in basic patents, has been presented [27]. Whilst the knowledge base continues to grow, the interrelationships of catalyst structure, polymerization conditions such as temperature and the chain microstructures of the resulting polymers will still be subjects of interest. Studies of ethylene-vinyl aromatic monomer polymerizations continue to be published. Chung and Lu reported the synthesis of copolymers of ethylene and P-methylstyrene [28] and the same group extended these studies to produce and characterize elastomeric terpolymers which further include propylene and 1-octene as the additional monomers [29,30]. Returning to the subject of alternative molecular architectures for copolymers, Hou et al. [31] has reported the ability of samarium (II) complexes to copolymerize ethylene and styrene into block copolymers.

3

STRUCTURE-PROPERTY RELATIONSHIPS FOR ETHYLENESTYRENE INTERPOLYMERS

Fundamental to the understanding of the structure-property relationships of materials is the elucidation of the effects of molecular structure on the material properties, including thermal transitions/phase behavior, viscoelastic response, mechanical properties and melt rheology. The introduction of aromatic groups into the polyethylene chain structure results in significant changes to the viscoelastic properties, melt rheology, mechanical properties and interfacial behavior of ES copolymers compared with, for example, other ethylene-a-olefin copolymers such as ethylene-1-octene (EO). Certain aspects of the structure-property relationships have been published previously [32–34], with this section summarizing the salient features of the structure-property relationships.

609

ETHYLENE-STYRENE COPOLYMERS

As a result of the catalyst and process conditions used in their manufacture, the particular copolymers of current major interest are atactic, and contain typically up to about 50mol% (~ 80wt%) styrene. These materials have been described as 'pseudo-random', since successive head-to-tail styrene chain insertions have been shown to be absent, even at high levels of styrene incorporation [1,2]. The term 'ethylene-styrene interpolymer' (ESI) is used here to describe the specific ethylene-styrene copolymers produced via INSITE™ Technology. For convenience, all subsequent comonomer contents are expressed in weight percentages, unless otherwise stated. For example, the code ES70 refers to an interpolymer having 70 wt.% comonomer styrene incorporation.

3.1

THERMAL TRANSITIONS/VISCOELASTIC

BEHAVIOR

Figure 26.1 correlates the DSC data for crystallization (Tc), melting (Tm) and glass transition (Tg) temperatures with the styrene content of ESI. The effects of styrene incorporation on the structure of the ESI and hence on their thermal transitions are profound. Crystallization and melting temperatures and crystallinity (area under the DSC melting endotherm) are found to decrease with increasing styrene content. This is manifested in the DSC data as a change in the shape of the melting endotherm. At low styrene contents (<40%), ESI generally exhibit a well-defined melting process. At higher styrene content, there is no discernible melting endotherm and the materials are essentially amorphous. These findings are in accord with the expectation that the 100

30

40

50 60 Weight % Styrene

70

80

Figure 26.1 Thermal transitions of ethylene-styrene interpolymers (ESI). Tm, TC and Tg denote the melting, crystallization and glass transition temperatures, respectively

610

Y. W. CHEUNG AND M. J. GUEST

crystallizability of ethylene chain sequences is suppressed, and ultimately inhibited, by the incorporation of styrene into the crystallizing chains. Based on steric hindrance arguments, the styrene unit is completely excluded from the crystalline region of the copolymers. From DSC data, ESI have Ts generally in the range -20 to +30 °C. Above 50% styrene content, Tg increases with increasing styrene content. Below about 50 % styrene content, Tg appears to be broadly independent of styrene content. This observation may be attributed to the crystalline domains, which restrict molecular mobility in the amorphous region and influence the glass transition process. In addition, the styrene units in the semicrystalline polymers are located in the amorphous phase, and thus give a higher Tg phase than would be found for a fully amorphous, low styrene content ESI. Transition from the semicrystalline state to an essentially amorphous solid-state structure occurs at about 45 wt% (~ 20 mol%) styrene incorporation. Figure 26.15 can be referred to for a plot of percentage crystallinity versus mole percent comonomer. The relaxation properties as probed by dynamic mechanical spectroscopy (DMS) for a series of ESI are shown in Figure 26.2. In accordance with the DSC results, the tan S loss maximum or Tg for the semicrystalline ESI appears to be fairly independent of styrene content. For the essentially amorphous ESI (>45wt% S), Ts increases with increasing styrene content. When compared with the amorphous ESI, the amplitude and width of the Tg loss peak are lower in amplitude and broader, respectively, for the lower styrene ESI, as is characteristic of a semicrystalline material. The amorphous ESI exhibit an intense loss process associated with the amorphous phase rg(ESI). The width of this loss 10 ES61

ES70 ES77

ES42 y-transition

o.l

0.01 -150

-100

50

-50

100

T, °C

Figure 26.2 Dynamic mechanical spectroscopy (DMS) plot of land versus temperature, measured at 10 rad/s, for a series of ESI with styrene incorporation ranging from 24 (ES24) to 77wt% (ES77)

ETHYLENE-STYRENE COPOLYMERS

611

process is comparable to that of the atactic polystyrene, and provides strong evidence that the glass transition process is characteristic of a homogeneous amorphous material. Of further note in Figure 26.2 is the so-called -/-transition. This has been associated with local chain motions [35] in polyethylene; such chain motions are believed to contribute to the toughness found at sub-ambient temperatures. The relaxation behavior of selected semicrystalline ESI is depicted in Figure 26.3. It can be seen that the loss peak evident in the temperature range -50 to +50°C shows increasing breadth of the relaxation process as the styrene content in ESI decreases. The relaxation processes associated with this loss peak are complex in nature. The relaxation behavior of semicrystalline polymers is fundamentally different from that of amorphous polymers. The longrange segmental motions associated with the Tg process become hindered owing to the restrictions imposed by the crystallites. These observations can be rationalized from the behavior of high-density polyethylene (HDPE): two main transitions, termed a and 7, are observed in HDPE [36]. The higher temperature a-transition is not an amorphous TB process but rather has been attributed to the longitudinal chain diffusion through the crystallites [37]. The mechanism of the a-transition is associated with the motions of the interfacial regions (tie molecules, folds and loops) which require chain mobility in the crystal. The ^-transition is associated with localized molecular motions involving short segments (e.g. three to four CH2) in both the amorphous and crystalline phases. As crystallinity decreases resulting from the incorporation of short chain branches or comonomer, the

Figure 26.3 Dynamic mechanical spectroscopy (DMS) plot of tand versus temperature, measured at 10 rad/s, for semicrystalline ESI with styrene incorporation ranging from 7 (ES7) to 29wt% (ES29)

612

Y. W. CHEUNG AND M. J. GUEST

^-relaxation emerges and increases in amplitude with decreasing crystallinity. Although differences exist, the (3-relaxation is commonly identified as the Tg transition [38]. Accordingly, the nature of the Tg process for the semicrystalline ESI may be associated with the p-relaxation found in other ethylene copolymers. On the basis of the frequency-temperature DMS analysis as shown in Figure 26.4, the diffused loss peak observed in the low-styrene ESI appears to be composed of two relaxation processes characteristic of the a- and p-relaxations. Developing an understanding of the viscoelastic properties and timetemperature-rate dependence of properties has been a key element of the materials science of these interpolymers. In particular, the effect of composition was investigated in the range 43–76% styrene in ESI. Master curves of dynamic moduli extending from glassy to terminal zones were obtained by combining torsion rectangular measurements in the glass transition region with parallel plate rheometry in the melt. Data analysis showed that all these ESI exhibited time-temperature superposition; the temperature dependence of the properties for the amorphous copolymers studied could be described by a single pair of Williams-Landel-Ferry (WLF) coefficients in the range Tg < T < Tg + 100°C. The Theological properties of a series of ethylene-styrene copolymers having up to 20.5 mol% styrene have been reported by Lobbrecht et al. [39]. Master curves were generated and the analysis was considered to show that there was a change from Arrhenius behavior to WLF behavior for those copolymers having > 16.5 mol% styrene. Further aspects of the viscoelastic behavior of ESIs which have been reported to date include linear stress relaxation behavior of amorphous ESI [40] and the creep behavior of amorphous ESI in the glass transition region [41]. Chen et al. [42] 0.25

-100

Figure 26.4 Dynamic mechanical spectroscopy (DMS) plot of land versus temperature, measured at 0.1 1, 10, 100rad/s, for ES7

ETHYLENE-STYRENE COPOLYMERS

613

also reported the large strain-stress relaxation and strain recovery of ESIs at temperatures above Tg, and found that the observed behavior could be rationalized in terms of various network models.

3.2

MECHANICAL PROPERTIES

The tensile stress-strain behavior (23 °C, S.Vmin"1) for ESIs, as depicted in Figure 26.5, generally exhibit large strain at rupture, and have been found to show uniform deformation behavior [22]. Tensile stress-strain behavior for ESIs differing in styrene content and molecular weight has been reported. The deformation response of the semicrystalline materials is predominantly controlled by the level of crystallinity and the connectivity between the crystalline and amorphous phases. The large-scale deformation behavior and recovery behavior of semicrystalline ESI have been studied by Chang et al. [43,44] as a function of temperature, comonomer content and crystallinity and compared with the behavior of ethylene-octene (EO) copolymers. Chen et al, [45] provided an in-depth comparison of the morphological structure and properties of ESIs and EO copolymers and confirmed that aspects of deformation which depended on crystallinity, such as yielding and cold drawing, were determined primarily by comonomer content for both sets of copolymers. Clearly a major factor which is different between the ESI and EO copolymers is the respective location of the Tg and the influence of chain mobility on the mechanical response observed.

100

200

300

400 500 Strain, %

600

700

800

Figure 26.5 Engineering stress-strain curves measured at 23 °C. RT denotes room temperature

614

Y. W. CHEUNG AND M. J. GUEST

The materials in the mid-styrene region (40–65 % S copolymers) are characterized by low modulus, and they show some elastomeric characteristics. Information on the deformation and recovery behavior of selected ESIs was reported by Mudrich et al. [46], with an ES45 copolymer of particular note in terms of strain recovery after deformation. The effects of styrene become dominant in the high-styrene region where the modulus and yield stress are seen to increase. ESIs having Tg above ambient temperature show some characteristics of glassy materials. Additionally, high-styrene (>70% S) ESI exhibit interesting shape-reshape functionality [47]. Good low-temperature toughness of ESI has been evident from impact testing and low-temperature tensile testing. As discussed above, DMS provides some evidence that there are sub-Tg chain motions which may contribute to energy dissipation and toughness. In addition, Chen et al. [40] provided estimates of the entanglement molecular weights (Me) for ESI. The low A/e values found suggest a high entanglement density in these polymers, and this is considered to contribute to the ability of the polymers to shear yield at temperatures below T"g rather than undergo brittle fracture.

3.3

MELT RHEOLOGY AND PROCESSABILITY

The melt rheology and associated information from solid-state DMS, melt strength and pressure-volume-temperature (PVT) property of ESIs have been reported by Karjala et al. [48]. The polymers were shown to have (1) good thermal stability at processing temperatures, (2) viscosities which are a strong function of styrene content in addition to molecular weight and (3) excellent processability, particularly regarding shear thinning characteristics at high shear rates and good melt strength. Melt Theological master curves could be generated via time-temperature superposition. In addition to providing fundamental structure-property understanding, melt rheology has further found utility in the design of polymer blends based on ESI [49]. Viscosity-shear rate dependence as a function of styrene content and molecular weight was analyzed with a Cross model [50], with the results for the zero shear viscosity 0/0), critical shear stress (T*), relaxation time (A.) and powerlaw index («) shown in Table 26.1. An example of a fit of the data to this model is shown in Figure 26.6. As can be seen from Table 26.1, at equivalent melt index and with increasing styrene level, the zero shear viscosities decrease, relaxation times increase and the ESI becomes more shear sensitive. Table 26.1 also contains an average relaxation time estimated from the inverse of the frequency at which the storage and loss modulus cross. These values are lower than those based on the Cross model, but show the same general trend of relaxation time decreasing with increasing melt index and with increases in styrene level at a given melt index/molecular weight.

615

ETHYLENE-STYRENE COPOLYMERS

Table 26.1 Cross model parameters: ^0 is the zero shear viscosity, T* critical shear stress, A relaxation time and n power exponent, where ES20 represents 20 wt% styrene ESI and MI denotes the melt index Material

fo (P)

ES20-0.1MI ES20-0.5MI ES20-3MI ES20-11MI ES60-0.1MI ES60-0.5MI ES60-3M1 ES60-10MI

1.63 3.63 4.10 9.33 4.20 1.48 2.98 8.21

x x x x x x x x

T* (dyn/cm2)

106 105 104 103 105 105 104 103

1.31 9.94 2.67 6.83 2.71 1.73 1.88 1.62

x x x x x x x x

105 104 105 10s 106 106 106 106

/(s)

/ cross-over (s)

n

12.4579 3.6561 0.1537 0.0137 0.1550 0.0860 0.0159 0.0051

1.2716 0.0042 0.0006 ND 0.0882 0.0250 0.0040 0.0012

0.423 0.491 0.569 0.550 0.140 0.206 0.214 0.238

The melt strength of the ESIs was measured at several temperatures and was shown to improve dramatically with both increasing styrene content and changes in temperature. Figure 26.7 shows the effect of styrene level at a constant melt index of 0.1. At equivalent melt index, the melt strength increases from approximately 12 to 25 cN as the styrene level increases from ~ 20 to ~ 60wt%. By decreasing the temperature, the melt strength of ESI is also dramatically increased, in part relating to its relatively high activation energy. Activation energies of melt strength are similar to those of viscous activation energy, of the order of 15 kcal/mol.

Reference Temp. = 190°C

1000

ES20-0.1MI ES20-0.5MI ES20-3MI ES20-11MI ES60-0.1MI ES60-0.5MI

0.01

100 104 co aT Frequency (rad/s)

Figure 26.6 indices (Ml)

108

Melt rheology master curve for ES20 and ES60 at various melt

616

Y. W. CHEUNG AND M. J. GUEST T=190°C;MI=0.1

50 100 Velocity (mm/s)

Figure 26.7

150

Melt strength of ESI measured at 190 CC

As a general comment, melt processability of interpolymers is favorable towards most fabrication techniques. These characteristics have allowed the fabrication of ESI articles by a wide range of standard melt processing techniques, including injection molding, film fabrication, blow molding operations and melt extrusion. The potential for ESI to be utilized in calendering operations has been described by Karjala et al. [51]. Selected ESIs were found to demonstrate the requisite Theological properties and thermal stability to be successfully calendered, and this was supported by commercial-scale validation.

4

MATERIALS ENGINEERING ASPECTS

Although most polymers are in fact modified to a greater or lesser degree to optimize their utility in end-use applications, certain basic characteristics of these ethylene-styrene interpolymers make them particularly likely to be modified, or to be used as modifiers. These characteristics are as follows: 1. polymer chain microstructure, aromatic/olefin functionality and inherent compatibility with a wide range of other polymers, fillers and low molecular weight materials, including bitumens [52], plasticizers [53], tackifiers [54] and processing aids; 2. the location of the glass transition temperature (Tg), in the range -20 to +35 °C, and the associated major changes in, e.g., modulus for relatively small changes in copolymer styrene content or temperature; 3. excellent processability.

ETHYLENE-STYRENE COPOLYMERS

617

This section provides more details on selected aspects of blend systems using interpolymers as components, filler composites and terpolymers. A related aspect which warrants mention in relation to multicomponent systems is the interfacial nature and behavior of these interpolymers. Ronesi [55] presented a study of the interfacial adhesion between LDPE and ESI, analyzing the significant effects of ESI copolymer styrene content, layer thickness and test temperature.

4.1

INTERPOLYMER BLENDS

Primarily because of the olefinic and styrenic functionality, ESI generally exhibit good compatibility with a wide range of polymers. Studies on ESI blends with atactic polystyrene show that significant toughening of brittle polystyrene can be achieved with selected ESI, in large part due to good compatibility between these polymers [56]. The olefinic nature of ESI helps to provide compatibility with olefinic polymers and copolymers, including polyethylenes, ethylene-a-olefin copolymers and polypropylene homopolymers and copolymers [57]. As reported by Diehl et al. [58], interpolymers are also compatible with a broader range of polymers, including styrene block copolymers [59], poly(vinyl chloride) (PVC)-based polymers [60], poly(phenylene ethers) [61] and olefinic polymers such as ethylene-acrylic acid copolymer, ethylene-vinyl acetate copolymer and chlorinated polyethylene. Owing to their unique molecular structure, specific ESI have been demonstrated as effective blend compatibilizers for polystyrene-polyethylene blends [62,63]. The development of the miscibility/ compatibility behavior of ESI-ESI blends differing in styrene content will be highlighted below.

4.2

BLENDS OF ETHYLENE-STYRENE INTERPOLYMERS: MISCIBILITY CONSIDERATIONS

Based on the pioneering work of Molau [64], it is evident that phase separation can occur in blends of two or more copolymers produced from the same monomers when the composition difference between the blend components exceeds some critical value. The mean field theory for random copolymercopolymers blends has been applied to ES-ES blends differing in styrene content to determine the miscibility behavior of blends [65,66]. On the basis of the solubility parameter difference between PS and PE, it was predicted that the critical comonomer difference in styrene content at which phase separation occurs is about 10 wt% S for ESI with molecular weight around 105. DMS plots for ES73 and ES66 copolymers and their 1:1 blend are presented in Figure 26.8.

618

Y. W. CHEUNG AND M. J. GUEST

10

20

30

40

50

60

Figure 26.8 DMS of ES73-ES66 (50:50 weight ratio) blend showing Tg intermediate to the pure component 7gs

The blend, having a 7 wt% styrene difference between blend components, can be considered miscible, as indicated by the presence of a single Tg intermediate of the two pure copolymers. The transition width of the 7"g for the blend is almost identical with those of the pure components, which further supports singlephase behavior in the blend. In accordance with copolymer model prediction, a difference in styrene content of greater than about 10wt% (for a 105 MW) between two ES copolymers is sufficient to drive phase separation. This criterion is further supported by the DMS plot shown in Figure 26.9, in which two distinct Tgs are clearly evident for an ES73-ES58 blend (1:1 weight ratio and having a 15wt% styrene difference between blend components), indicating blend immiscibility. Figure 26.10 shows two distinct Tgs for a blend containing five components for which the styrene content difference between successive components lies close to or below the critical composition difference for phase separation. On the basis of the Tg behavior, it may be rationalized that the five components have phase separated into two phases in which the ES66 and ES73 form one miscible phase and ES58, ES52 and ES46 form another miscible phase. It is interesting that the peak amplitude of the blend is much smaller than that of the pure blend components. These findings clearly indicate that the glass transition process, including peak temperature, width and amplitude, can be controlled by combining ESI with different compositions and hence different degrees of compatibility. Such ESI-ESI blends can therefore be designed to be immiscible in nature, and offer the opportunity to engineer materials with broadened Tg loss processes and enhancements in relaxation behavior, mechanical properties and melt rheology and processability.

619

ETHYLENE-STYRENE COPOLYMERS

-20

-10

0

10

20

30

40

50

60

Figure 26.9 DMS of ES73-ES58 (50:50 weight ratio) blend showing two distinct Tgs indicative of phase separation

20%ES73/20%ES66/ 20%ES58/20%ES52/ 20%ES46

T, °C

Figure 26.10 DMS of five-component blend which phase separates into two distinct phases in which ES73 and ES66 form a miscible phase and ES58, ES52 and ES46 form another miscible phase Chen et al. [67,68] further extended the study of binary blends of ESI over the full range of copolymer styrene contents for both amorphous and semicrystalline blend components. The transition from miscible to immiscible blend behavior and the determination of upper critical solution temperature (UCST) for blends could be uniquely evaluated by atomic force microscopy (AFM) techniques via the small but significant modulus differences between the respective ESI used as blend components. The effects of molecular weight and molecular weight distribution on blend miscibility were also described.

620

Y. W. CHEUNG AND M. J. GUEST

Extending the above rationale regarding miscibility to blends of ESI with styrenic polymers, it is evident that ESI having less than 80 wt% styrene content will be immiscible with polystyrene, unless the molecular weights of the respective polymers are very low. Similarly, ESI with more than 10wt% styrene content will be immiscible with polyethylene.

4.3

FILLER COMPOSITES

It is well known that the addition of fillers to polymers can enhance the stiffness, dimensional stability, upper service temperature, tensile strength, damping characteristics and ignition resistance, in addition to lowering the cost of fabricated parts [69]. Additionally, many durable applications of ESI can be pursued by modifying the balance of properties through the addition of inorganic fillers. The phenyl functionality of ESI is postulated to contribute to compatibility with a wide range of fillers via possible polymer-filler interactions. ESI has been reported to have good filler acceptance with a broad range of inorganic fillers, including calcium carbonate, barium sulfate, alumina trihydrate (ATH) and magnesium hydroxide (MgOH2) [70,71]. The resulting composite materials generally exhibit very good mechanical properties, even for relatively high loadings of fillers. Many combinations of filler types and ESIs with different copolymer styrene contents show exceptional tensile elongations at rupture. Impact testing of ESI-filler systems has further shown that these composites have good toughness. The mechanical properties, and particularly failure performance, indicate that good interfacial bonding exists between ESI and fillers. Figure 26. 1 1 shows the stress-strain behavior of ES30 filled with various levels of ATH. It can be seen that the yield stress increases with increasing level of ATH while the ultimate elongation is in excess of several hundred percent even for materials with more than 50wt% ATH. The modulus of composite materials can be modeled by the generalized Kerner equation:

where M\ is the modulus of the polymer, M the modulus of the composite, k^ the Einstein coefficient, m the maximum volume fraction, 2 the volume fraction of filler and A = kt — 1 • It can been from Figure 26. 1 2 that the relative flex modulus of ES30 filled with ATH and MgOH2 can be reasonably described

621

ETHYLENE-STYRENE COPOLYMERS

Figure 26.11

Engineering stress-strain curves (23 °C) for ATH-filled ES30

20

15

10

10

20

30

40

50

VOLUME % FILLER

Figure 26.12 Relative flex modulus of ATH- and MgOH2-filled ES30. The solid line represents the Kerner model [Equation (1)] prediction

by the Kerner model up to about 30 vol.% filler. The deviation from the Kerner model observed at higher filler levels could be related to agglomeration of fillers and/or the onset of percolation. Interestingly, the modulus of MgOH2-filled ES30 is much higher than that of the ATH-filled material when the filler level exceeds 40 vol.%.

622

Y. W. CHEUNG AND M. J. GUEST

In addition to the modulus prediction, the ultimate properties including elongation and ultimate tensile strength, assuming good adhesion between filler and polymer, can be modeled by the following equations: (3)

(4)

where e is the elongation of the composite, £Q the elongation of the polymer, (f>2 the volume fraction of filler and a the ultimate tensile strength. The relative elongation and ultimate tensile strength for ATH/MgOH2-filled ES30 are depicted in Figures 26.13 and 26.14, respectively. Both the elongation and ultimate tensile strength of the composites generally show positive deviations from the model prediction. Interestingly, the ATH-filled materials generally show higher elongation than that of the MgOFb-filled materials. At 50 vol.% filler loading, this difference vanishes and the elongation for both materials is lower than that predicted by the model. In the case of the ultimate tensile strength, a positive deviation from the model prediction is always observed. This analysis provides additional support that good interfacial adhesion is found in most ESI-filler composites The filled ESI show rheological characteristics suggesting that the compositions can be easily fabricated into final parts. Filled ESI compositions are

0.8

0.6

0.4

0.2

10

20 30 40 VOLUME % FILLER

50

60

Figure 26.13 Relative elongation of ATH- and MgOH2-filled ES30. The dotted line represents the model [Equation (3)] prediction

623

ETHYLENE-STYRENE COPOLYMERS

0.8

0.6

"...

f

0.2

10

20

30

40

50

60

VOLUME % FILLER Figure 26.14 Relative tensile strength of ATM- and MgOH2-filled ES30. The dotted line represents the model [Equation (4)] prediction

expected to find broad utility in applications including, for example, wire and cable [72], injection molded articles, film and sheet structures, profile extrusions and flooring systems.

4.4

TERPOLYMERS

INSITE™ Technology also permits the copolymerization of ethylene and styrene with additional monomers [73], including dienes, higher a-olefins and norbornene. Basic structure-property relationships of ESP terpolymers have recently been published [74]. There are limited, mainly patent, references to the preparation and basic characterization of terpolymers based on ethylene, styrene and propylene (ESP). The use of propylene as a comonomer in combination with ethylene and styrene results in a much broader polymer design space for materials development. Depending on the comonomer composition of the terpolymers, these materials exhibit solid-state microstructures ranging from semicrystalline through to essentially amorphous materials. The terpolymers are most differentiated from ESI when there is measurable crystallinity in the solid-state microstructure, in part because the propylene comonomer introduces methyl groups on the polymer chain which modify both the crystalline and amorphous phases of the solid-state microstructure [75]. Figure 26.15 shows the variation in crystallinty of ESP as a function of total comonomer (styrene and propylene). The crystallinity-composition dependence for ESI is also plotted for comparison. The incorporation of monomer

624

Y. W. CHEUNG AND M. J. GUEST

units such as styrene and propylene into a linear ethylene chain is generally accepted as introducing defects which suppress and inhibit the crystallization of —(CH2)W—sequences. For ESI produced with the constrained geometry catalysts of current interest, no detectable crystallinity is typically found for those copolymers containing more than about 18mol% styrene comonomer, as depicted in Figure 26.15. Owing to steric hindrance, the phenyl group introduced into the chain microstructure by the styrene comonomer is excluded from the crystalline regions. For EP copolymers, the critical propylene content above which crystallization is inhibited occurs at about 30mol%. This difference has been attributed to the partial inclusion of the methyl group of propylene into the ethylene crystalline lattice. Similarly to the EP copolymers, ESP terpolymers with more than 20mol% comonomer still exhibit a measurable level of crystallinity. The glass transition temperature, as measured by the tan S maximum from DMS, for the ESP terpolymers is presented in Figure 26.16. It is interesting that the Tg of the terpolymers is not only dependent on the total comonomer content but also strongly on the propylene/styrene ratio. For a given molar percentage of comonomer, Tg generally decreases with increasing propylene/ styrene ratio, primarily because propylene has a much lower Tg than styrene. In addition to perturbing the crystalline phase, propylene also profoundly affects the amorphous phase, as evidenced by the depression in TB of the semicrystalline ESP. An approach to modeling the Tg behavior of the ESP terpolymers has been introduced, with predictions from the model and experimental data showing satisfactory agreement, despite the simplifying assumptions that were made. 15

(P/S)=0.7 £>

(P/S)=4.5

EP

10

U

ESI

10

12

\

14 16 18 Mol% Comonomer

20

22

Figure 26.15 Crystallinity of ESI, EP and ESP. (P/S) denotes the propylene to styrene molar ratio

ETHYLENE-STYRENE COPOLYMERS

625

30

'(P/S)=0.25 20 • 0.29 10

0.4

0

'0.5

Semi-crystalline

0.9

-10 ^

-20'

(P/S)=0.7 n -7 A* A u. / A

-30

Amorphous

^-

1.6

- •^ "•

• 4.0

*4.5

EP -40 10

20

30

40

50

60

Mol% Comonomer Figure 26.16 Glass transition temperature, measured from the tan<5 peak maximum at 10 rad/s, of ESP. (P/S) denots the propylene to styrene molar ratio

The tensile stress-strain behaviors for the ESI, the series of ESP terpolymers and the EP copolymer are presented in Figure 26.17. All copolymers show elongation at rupture in excess of 600 %. The EP copolymer has a low stress at break compared with that of the ES copolymer, with the ESP terpolymers falling intermediate between the two copolymers, depending on the terpolymer monomer composition. It has been shown that this is correlated with the differences in microstructure between the respective copolymers rather than resulting from differences in their respective Tg, by conducting testing at equivalent reduced temperature (Tg + 10°C). The incorporation of methyl groups produces a less well ordered crystalline domain compared with the ES copolymers, reflected in the significantly lower stress at large strain (>300%) when propylene is used as a comonomer.

5

ATTRIBUTES AND APPLICATIONS

Ethylene-styrene interpolymers exhibit a novel balance of properties that are uniquely different from polyethylenes and polystyrenes. In contrast to other ethylene-a-olefin copolymers, ESI display a broad range of material response ranging from semicrystalline, through elastomeric to amorphous. The styrenic functionality and unique molecular architecture of ESI are postulated to be the basis of the versatile material attributes such as processability (shear thinning, melt strength and thermal stability), viscoelastic properties, low-temperature toughness and broad compatibility with other polymers, fillers and low molecular weight materials.

626

Y. W. CHEUNG AND M. J. GUEST 40

ESP-14,5 ES20

/

30

/ 20

/

/

/ .'

/

.'

.' ESP-7,11

10

EP15

0

2

3

4

5

6

Strain Figure 26.17 Engineering stress—strain curves, measured at 23 CC, for ESI, EP and ESP. Compositions are given in wt% comonomer (S and P)

The current range of potential markets and applications for ESI which have been identified now includes paintable injection molded toys [76], wire and cable, footwear, automotive, bitumen modification, packaging, injection and blow molded articles, adhesives, building and construction. Although intermaterial substitutions in existing applications based on thermoplastic elastomers such as styrene—butadiene block copolymers, flexible PVC and other ethylene-a-olefin copolymers, including ethylene-vinyl acetate, are potential application areas for ESI, the novel combination of material attributes suggests that new application areas will emerge. Of particular note to date are developments using ESI-based materials in foam applications. Novel foam structures offer attractive properties and characteristics including softness, esthetics and drape for a wide range of thermoplastic and crosslinked foam applications. Other product technologies of interest are as injection molded structural foams, as foamed layers in multilayer structures and as foamed blends of interpolymers with styrenic and olefinic polymers [77–79]. Interpolymers also have potential for co-extruded film and sheet applications.

6

SUMMARY

Ethylene-styrene interpolymers are a novel class of polymers exhibiting a unique combination of material attributes that are not found in polyethylenes,

ETHYLENE—STYRENE COPOLYMERS

627

polystyrenes or their blends. The effective production of these novel copolymers has been enabled by INSITE™ Technology. INDEXTM interpolymers were introduced by The Dow Chemical Company in December 1998, and this technology includes ethylene-styrene interpolymers. A product development plant to produce Interpolymers (Sarnia, Canada) had a successful start-up in September 1999. This plant, having a nameplate production capacity of around 22 500t (5 x 107lb) per annum, is currently allowing further product and process developments and application validation. These interpolymers based on ethylene and styrene are an integral part of an exciting new generic class of materials, offering unique opportunities for innovative developments in basic polymer chemistry, catalyst and process development, materials science and engineering and application technology.

7

ACKNOWLEDGMENTS

The authors particularly wish to thank Steve Chum, Scott Mudrich and Teresa Karjala for helpful comments and discussions. They further thank researchers at Case Western Reserve University, including Professor Eric Baer, Professor Anne Hiltner, Hong-Yu Chen and Andy Chang, for their contributions to the understanding of structure-property relationships and material classification. The authors also thank many others, especially Joe Huang and Ken Reichek, for their help with providing the data presented in this chapter. Finally, The Dow Chemical Company and the Interpolymer Business Management team are thanked for permission to publish this work.

REFERENCES 1. European Patent Application EP 416 815 Bl; The Dow Chemical Company; Stevens, J.C., Timmers, F.J., Wilson, D.R., Schmidt, G.F., Nickias, P.N., Rosen, R.K., Knight, G.W. and Lai, S. 2. US Patent 5 703 187; The Dow Chemical Company; Timmers, F.J. 3. Stevens, J.C., in Abstracts of the Metcon, Houston, Tx, p. 157 (1993). 4. Guest, M.J., Cheung, Y.W., Diehl. C.F. and Hoenig, S.M. 'Structure, properties and applications of ethylene/styrene interpolymers', in Metallocene-based Polyolefins: Preparation, Properties and Technology, Vol. 2, ed. J. Scheirs and W. Kaminski, Wiley, Chichestes, Chapt. 12, pp. 271–292, (1999). 5. US Patent 3117945; Union Carbide Corporation; Gorham, W.F. and Farnham, A.G. 6. Kobayashi, S. and Nishioka, A., J. Polym. Sci, Part A, 2, 3009 (1964). Trademark of The Dow Chemical Company.

628

Y. W. CHEUNG AND M. J. GUEST

7. Kawai, W. and Katsuta, S., J. Polym. Sci., Part A, 8, 2421 (1970). 8. Soga, K. and Lee, D., Polym. Bull., 20, 237 (1988). 9. US Patent 4748209; Nippon Petrochemicals Inc.; Orikasa, Y. Kojima, S., Inoue, T., Yamamoto, K., Sato, A. and Kawakami, S. 10. US Patent 5652315; Mitsui Toatsu Chemicals, Inc.; Inoue, N., Tetsunosuke, S., Masahiro, K. 11. Longo, P., Grassi, A., and Oliva, L., Makromol. Chem., 191, 2387 (1990). 12. Pellechia, C. and Olivia, L., Rubber Chem. Technol., 72, 553 (1999). 13. Sernetz, F.G., Mulhaupt, R., Amor, F., Eberle, T. and Okuda, J. Polym. Sci. Part A: Polym. Chem., 35, 1571 (1997). 14. Sernetz, F.G., Mulhaupt, R., Fokken, S. and Okuda, J., Macromolecules, 30, 1562 (1997). 15. Sernetz, F.G. and Mulhaupt, R., Macromol. Chem. Phys., 197, 1071 (1996). 16. Thomann, Y., Sernetz, F.G., Thomann, R., Kressler, J. and Mulhaupt, R., Macromol. Chem. Phys., 198, 739 (1997). 17. Sernetz, F.G. and Mulhaupt, R., J. Polym. Sci., 35, 2549 (1997). 18. Sernetz, F.G., Mulhaupt, R. and Waymouth, R.M., Polym. Bull., 38, 141 (1997). 19. Suzuki, T., Tsuji, Y., Watanabe, Y. and Takegami, Y., Macromolecules, 13, 849 (1980). 20. US Patent 5 043 408; Sumitomo Chemical Company; Kakugo, M., Miyatake, T. and Mizunuma, K. 21. Miyatake, T., Mizunuma, K. and Kakugo, M., Makromol. Chem., Macromol. Symp., 66, 203 (1993). 22. Izzo, L., Oliva, L., Proto, A. and Trofa, M., Macromol. Chem. Phys., 200, 1086 (1999). 23. Arai, T., Ohtsu, T. and Suzuki, S., Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., 349, 38 (1997). 24. Arai, T., Ohtsu, T. and Suzuki, S., Macromol. Rapid Commun., 19, 327 (1998). 25. US Patent 5883213; Denki Kagaku KK; Arai, T., Nakamura, A., Suzuki, S., Otsu, T. and Okamoto, A. 26. Pasztor, A.J., 'Styrene polymers', in Encyclopedia of Polymer Science, 2nd edn, J Wiley, Chichester, Vol. 16, pp. 114–117 (1989). 27. 'Special Feature I. Ethylene-Styrene Interpolymers: Technologies, Products and Markets'. The Metallocene and Single Site Catalyst Monitor', The Catalyst Group, Spring House, PA (December 2000). 28. Chung, T.C. and Lu, H.L., J.Polym. Sci. A: Polym. Chem., 35, 575 (1997). 29. Lu, H.L., Hong, S. and Chung, T.C., Macromolecules, 31, 2028 (1998). 30. Chung, T.C. and Lu, H.L., and Hong, S., Rubber Chem. Technol., 72, 283 (1999). 31. Hou, Z., Tezuka, H., Zhang, Y., Yamazaki, H. and Wakatsuki, Y., Macromolecules, 31, 8650(1998). 32. Cheung, Y.W. and Guest, M.J., in Proceedings of SPE ANTEC 54th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1634 (1996). 33. Chen, H., Guest, M.J., Chum, P.S., Hiltner, A. and Baer, E., J. Appl. Polym. Sci., 70, 109 (1998). 34. Guest, M.J., Cheung, Y.W. and Martin, J.M., in Proceedings of ACS Division of Polymeric Materials: Science and Engineering, Fall Meeting, New Orleans, LA, August 22-26, Vol. 81, pp. 371-372 (1999). 35. McCrum, N.G., Read, B.E. and William, G., Anelastic and Dielectric Effects in Polymeric Solids, J Wiley, New York (1967). 36. Flocke, H.A., Kolloid Z., 180. 118 (1962). 37. Boyd, R.H., Polymer, 26, 323 (1985).

ETHYLENE-STYRENE COPOLYMERS

629

38. Khanna, Y., Turi, E.A., Taylor, T.J., Vickroy, V.V. and Abbott, R.F., MacromoleculeslZ, 1302(1985). 39. Lobbrecht, A., Friedrich, C. Sernetz, F.G. and Mulhaupt, R., J. Appl. Polym. Sci., 65 209 (1997). 40. Chen, H.Y., Stepanov, E.V., Chum, S.P., Hiltner, A. and Baer, E., Macromolecules, 330, 8870 (2000). 41. Chen, H.Y., Stepanov, E.V., Chum, S.P., Hiltner, A. and Baer, E., J. Polym. Sci. B, Polym. Phys., 37, 2373 (1999). 42. Chen, H.Y., Stepanov, E.V., Chum, S.P., Hiltner, A. and Baer, E., Macromolecules, 32, 7587 (1999). 43. Chang, A., Stepanov, E.V., Guest. M., Chum, S., Hiltner, A. and Baer, E., in Proceedings of SPE ANTEC 56th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1803 (1998). 44. Chang, A., Stepanov, E.V., Chum, S.P., Hiltner, A. and Baer, E., in Proceedings of SPE ANTEC 57th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 2121 (1999). 45. Chen, H.Y., Chum, S.P., Hiltner, A. and Baer, E., J. Polym. Sci. B, Polym. Phys., 39, 1578 (2001). 46. Mudrich, S.F., Cheung, Y.W. and Guest, M.J., in Proceedings of SPE ANTEC 55th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1783 (1997). 47. US Patent 6156842; The Dow Chemical Company; Hoenig, S.M., Turley, R.R., Cheung, Y.W., Guest, M.J., Diehl, C.F., Stewart, K.B. and Sneddon, J. 48. Karjala, T.P., Cheung, Y.W. and Guest, M.J., in Proceedings of SPE ANTEC 55th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1086 (1997). 49. Karjala, T.P., Hoenig, S.M., Guest, M.J., Cheung, Y.W., Finlayson, M.P., Walther, B.W. and Montanye, J.R., in Proceedings of SPE ANTEC 58th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1833 (2000). 50. Karjala, T.P., Cheung, Y.W. and Guest, M.J., in Proceedings of SPE ANTEC 57th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 2127 (1999). 51. Karjala, T.P., Walther, B.W., Hill, A.S. and Wevers, R., in Proceedings of SPE ANTEC 57th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 2139(1999). 52. US Patent 6 107374; The Dow Chemical Company; Stevens, J.C., Fimmers, F.J., Gatzke, A.I., Bredeweg, C.J., McKay, K.W., Gros, W.A. and Diehl, C.F. 53. US Patent 5739200; The Dow Chemical Company; Cheung, Y.W., Gathers, J.J., Guest, M.J. and Bethea, J.R. 54. European Patent Application EP 0923619 Al: The Dow Chemical Company; Parikh, D., Guest, M.J. and Speth, D.R. 55. Ronesi, V.M., Polym. Mater. Sci. Eng., 84, 440 (2001). 56. Park, C.P. and Clingerman, G.P., in Proceedings of SPE ANTEC 57th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 2134 (1999). 57. US Patent 6 184294 Bl; The Dow Chemical Company; Park, C.P., Thoen, J., Broos, R., Guest, M.J., Cheung, Y.W., Chaudhary, B.I., Gathers, J.J. and Hood, L.S. 58. Diehl, C.F., Guest, M.J., Chaudhary, B.I., Chueng, Y.W., Van Volkenburgh, W.R. and Walther, B.W., in Proceedings of SPE ANTEC 57th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 2149 (1999). 59. US Patent 5741 857; The Dow Chemical Company; Esnault, C.P. and Edmonson, M.S. 60. US Patent 6 136923; The Dow Chemical Company; Cheung, Y.W. and Guest, M.J. 61. US Patent 6201067; The Dow Chemical Company; Cheung, Y.W., Guest, M.J., Chum, P.S. and Kao, C.I.

630

Y. W. CHEUNG AND M. J. GUEST

62. US Patent 5460818; The Dow Chemical Company; Park, C.P., Clingerman, G.P., Timmers, F.J., Stevens, J.C. and Henton, D.E. 63. Park, C.P. and Clingerman, G.P., in Proceedings of SPE ANTEC 54th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1887 (1996). 64. Molau, G.E., Polym. Lett., 3, 1007 (1965). 65. Cheung, Y.W. and Guest, M.J., J. Polym. Sci. B, Polym. Phys., 38, 2976 (2000). 66. Chen, H.Y., Cheung, Y.W., Chum, S.P., Hiltner, A. and Baer, E., in Proceedings of SPE ANTEC 58th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1828 (2000). 67. Chen, H.Y., Cheung, Y.W., Hiltner, A. and Baer, E., Polymer, 42, 7819 (2001). 68. Chen, H.Y., Chum, P.S., Hiltner, A. and Baer, E., Macromolecules, 34,4033 (2001). 69. Nielsen, L.E. and Landel, R.F., Mechanical Properties of Polymers and Composites, Marcel Dekker, New York (1994). 70. Guest, M.J., Cheung, Y.W., Betso, S.R. and Karjala, T.P., in Proceedings of SPE ANTEC 58th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 2116(1999). 71. US Patent 5973049; The Dow Chemical Company; Bieser, J.O., Cheung, Y.W., Guest, M.J., Thoen, J.A. and Gathers, J.J. 72. Betso, S.R., Guest, M.J., Remenar, R.M., Field, A.W. and Keen, F.E., in Proceedings of SPE ANTEC 58th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 3012 (2000). 73. US Patent 5872201; The Dow Chemical Company; Cheung, Y.W., Guest, M.J., Timmers, F.J. and Hahn, S.F. 74. Guest, M.J., Cheung, Y.W., Karjala, T.P., Ruiz, J.M. and Kolthammer, B.W.S., in Proceedings of SPE ANTEC 59th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 3012 (2001). 75. Alamo, R.G. and Mandelkern, L., Thermochimi. Ada, 238, 155 (1994). 76. Chang, D. and Cheung, Y.W., in Proceedings of SPE ANTEC 59th Annual Conference, Society of Plastics Engineers, Brookfield, CT, p. 1704 (2001). 77. US Patent 6048909; The Dow Chemical Company; Chaudhary, B.I., Barry, R.P. and Cirihal, S.C. 78. US Patent 5993707; The Dow Chemical Company; Chaudhary, B.I., Hood, L.S., Barry, R.P. and Park, C.P. 79. Chaudhary, B.I., Barry, R.P. and Tusim, M.H., J. Cell. Plast., 36, 397 (2000).

Properties of Styrenic Polymers

This page intentionally left blank

27

Fracture Behavior of High Impact Polystyrene and Acrylonitrile-ButadieneStyrene T. VU-KHANH Universite de Sherbrooke, Sherbrooke, Quebec, Canada

1

INTRODUCTION

One of the most important weaknesses of polystyrene is its poor impact resistance. To improve this performance, much attention has been paid to adding an elastomeric, dispersed phase to the polymer matrix. The role of rubber toughening in the impact performance of polymers has been extensively investigated. Most plastics can be made tougher by the addition of a small amount of rubbery material, dispersed as second-phase particles on a microscopic scale. The improvements in toughness can be seen in several commercialized systems based on styrenic polymers that are brittle in their unmodified state. Thus, the glassy, brittle polystyrene (PS) is transformed into high-impact polystyrene (HIPS), used successfully in applications such as refrigerator linings, packaging, vacuum cleaners, fans or even shoe heels. Similarly, styrene-acrylonitrile (SAN) has been modified into the tough acrylonitrile-butadiene-styrene (ABS) and is utilized in applications such as telephone sets, luggage, computer casings and interior car fittings. These materials possess good impact properties and have been widely used as engineering materials. In rubber toughened plastics, the rubber particles constitute the dispersed phase in the polymer matrix. Their essential role is to act as a stress concentrator. When the toughened material is Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy f 2003 John Wiley & Sons Ltd

634

T. VU-KHANH

subjected to a uniaxial stress, the localized stress experienced by the matrix in the immediate vicinity of a rubber particle is magnified by the local stress concentration effect. The matrix will yield locally in response to this localized stress field, thus avoiding a global brittle catastrophic failure of the material. The concentration of stress that initiates the local yielding in the matrix is just the first step in a complex process; if the overall applied stress is increased further the matrix then continues to deform by shear yielding or crazing. Shear yielding involves the creation of bands of highly oriented, stretched material at 45c to the direction of the applied stress. Crazes are cracks, spanned by elongated fibrils of the material which can carry load and thus maintain the structural integrity of the material. Dilatometric studies are often used to distinguish shear yielding, which is essentially a constant-volume process, from the dilatational processes of crazing or voiding. These tests are carried out under tension since a compressive or pure shear state of stress does not allow crazing to occur. Measurements of volume changes in rubber-modified plastics, in tensile creep under constant load have been performed to elucidate the mechanisms of local, irreversible deformation in toughened polymers. HIPS has been shown to exhibit the greatest volume change because the PS matrix forms crazes in preference to shear bands. ABS, however, shows a transition from shear yielding to crazing or large-scale voiding when the deformation level becomes important. In addition to the multiplication of energy dissipation by local plastic deformation, the rubber particles can also deform, stretching to bridge an expanding crack opening. The fracture and cavitation of the rubber particles also relieve the triaxial stress at the crack tip. These mechanisms constitute the basic toughening process. When the temperature and/or loading rate change, the fracture mechanisms become more complex. At relatively low temperatures or high loading rates, HIPS and ABS can become brittle. Like most polymers, the fracture behavior of HIPS and ABS is also strongly time and temperature dependent. Depending on the loading rate and temperature, the fracture mode and performance can be very different. HIPS and ABS can break in a brittle, semi-ductile or ductile manner. Brittle fracture usually results in the shattering phenomenon of the part. In this case, the elastic energy stored in the part is much higher than the energy dissipated in the fracture process. The excess of energy is transformed into kinetic energy and transferred to the debris so that it can fly away with a very high velocity. In large structures, the kinetic energy can assist the crack to propagate without external loading. In terms of safety, this fracture behavior is a real concern. Conversely, ductile fracture generally occurs under stable crack propagation, with more plastic deformation. In this case, after initiation, the cracks can only propagate with additional supply of energy by external loads so that the character of the failure is less catastrophic. It is therefore essential to use a quantitative method to characterize the fracture performance of HIPS and ABS. In fact, the field of fracture of materials has received much attention and research effort after several catastrophic failures of major structures, usually

FRACTURE BEHAVIOR OF HIPS AND ABS

635

made of high-strength metallic materials. The investigations carried out during the two decades following the Second World War have clearly demonstrated that brittle, catastrophic fracture is generally initiated by defects that exist in most materials and structures. This has led to the development of fracture mechanics that has brought significant progress in the understanding of failure of materials. Using the theories in fracture mechanics (developed mainly for metals), quantitative methods have been proposed for the characterization of the impact performance of polymers [1–5]. Most of the work in this area was published in the 1970s. The proposed methods provide a new interpretation of the results of the common Charpy and Izod tests, but using samples containing a sharp initial crack instead in order to simulate the presence of a defect in the material. This chapter presents recent results on quantitative analyses of the fracture behavior of HIPS and ABS. The emphasis is put on the prediction of fracture performance over a wide range of deformation rates and temperatures. Different approaches currently used to characterize fracture performance are applied to HIPS and ABS and recommendations on the use of these methods are presented. The fundamental mechanisms controlling the dependence of fracture performance and yield on deformation rate and temperature are discussed.

2

QUANTITATIVE CHARACTERIZATION OF FRACTURE

It has been shown that fracture is a very complex process and the fracture performance depends on both the initiation and the propagation of a defect [6-10] in the material. Under impact, most polymers break in very distinct manners. Several types of fracture have been identified depending on the amount of plastic deformation at the crack tip and the stability of crack propagation. For each type, an appropriate analysis has been developed to determine the impact fracture energy of the material. These methods have also been verified in various plastics [11,12]. The different fracture behaviors in most polymers are illustrated in Figure 27.1, which shows a schematic drawing of the load-deflection diagram of Charpy tests on HIPS [13] under an impact velocity of 2m/s at various temperatures. With increasing temperature, the fracture mode changes from brittle to semiductile at about –50°C and then becomes ductile at temperatures higher than –30 °C.

2.1

BRITTLE FRACTURE

Brittle fracture occurs when the strain energy stored in the sample up to the point of fracture is much larger than the energy dissipated in the creation of the

636

T. VU-KHANH Load T<-50°C -50°C
T>-30C

Displacement Figure 27.1 Load-deflection diagram of Charpy tests on HIPS at 2 m/s

two fracture surfaces. In this case the crack grows in an unstable manner. The energy absorbed by the specimen to fracture is that stored elastically in the sample up to the point of fracture. The elastic strain energy has also been shown to be related to the material's fracture energy via a calibration factor that can be determined from the geometry of the sample [1–4]. A method based on linear elastic fracture mechanics (LEFM) has been proposed for this mode of fracture to determine the fracture energy at crack initiation [1–4]. In the Charpy test, the energy absorbed by the specimen, U, can be related to the toughness of the material, Gc (the critical strain energy release rate or crack extension force) by c

dC/dA

(1)

giving U = GCBD4>

(2)

with 1 C BD(dC/dA)

(3)

where B is the sample thickness, D its width and 0 is a geometric function which can be evaluated for any geometry by considering the appropriate calibration factor established for pre-cracked samples used in fracture tests [14, 15]. In the

FRACTURE BEHAVIOR OF HIPS AND ABS

637

case of the Charpy test (utilizing three-point-bending samples), the geometric function 4> can be determined from the calibration factor established for threepoint-bend samples for the calculation of the stress intensity factor K:

where P is the load applied on the sample and f(a/D) is a calibration factor that can be found in the literature [14, 15]. The strain energy release rate G is defined as =

2

dA

where C is the specimen compliance, A is the crack area and v and E are the material's Poisson ratio and Young's modulus, respectively. From Equations (4) and (5), dC/dA and C can be written as a function of a/D as follows:

dC dC dA ~ d(a/D)BD

2(1 – v2)S2

and

C = Co +

f 2 (a/D)d(a/D)

(7)

where Co is the specimen compliance for a/D = 0. From the established function f(a/D) for a thee-point-bend sample [16, 17], one can numerically calculate dC/dA, C and , should be a straight line, the slope of which gives Gc.

2.2

SEMI-DUCTILE FRACTURE

When the polymer is more ductile, several steps of stable and unstable crack propagation can successively occur during the fracture process of the sample. Figure 27.2 illustrates the observed fracture surface of such fracture behavior. For this semi-ductile fracture mode, a method using two material parameters has been proposed to characterize the impact resistance of the polymer [11]. The method takes account of an average value of the fracture energy during the stable propagation of the crack, and also the fracture energy at instability. In this case, the total energy absorbed by the sample to break, U, has been shown

638

T. VU-KHANH

Figure 27.2 Fracture surface of semi-ductile fracture: (a) PA 11 at room temperature; (b) HIPS at -40 °C; (c) variables considered in semiductile model

to be the sum of all the energies dissipated during the various crack propagation stages occurring in the same sample: (8)

where the subscripts (1), (2), etc. refer to the first, second, etc. step of stable or unstable crack growth, Gst represents the average value of the fracture energy of stable propagation (the overbar indicates an average value of Gst over the stable propagation step of the crack), Ginst is the fracture energy of the material at the onset of instability, A is the surface area of the stable propagation zone, a is the crack length, D is the specimen width and C is the specimen compliance.

FRACTURE BEHAVIOR OF HIPS AND ABS

639

If the energy lost by the hammer is equal to the energy absorbed by the sample, Utot may be replaced by the measurement of energy obtained from the pendulum after the impact. In Equation (2), the kinetic energy due to unstable crack propagation has not been taken into account. In fact, the amount of energy released during an unstable propagation is higher than the energy necessary to create the two surfaces. This excess energy is transformed into kinetic energy stored in the sample and, when the crack stops, it contributes to the additional amount of energy required to re-initiate the crack growth. Thus, the additional energy supplied by the external force (by the hammer of the impact machine) to the specimen after the crack arrest is smaller than the energy necessary to break a specimen having an equivalent crack length, because of the contribution of this kinetic energy. To a first approximation, it is possible to neglect the additional energy supplied to the sample by the hammer after the first crack arrest. This assumption was also verified by an instrumental impact test and the recorded load-displacement curve showed that the additional energy supplied by the hammer is indeed very small in comparison with the energy stored in the sample until the first instability of the crack. The fracture process takes place as if there were only one step of stable crack propagation, and subsequently the remaining fracture process is entirely brittle. The energy balance of the mixed fracture with successive stable and unstable crack propagation periods can thus be reduced to a simpler equation: U = G st A 1 +G inst BD^

(9)

where Gst is the average fracture energy during the first stable crack propagation stage, A\ is the fracture surface area of the first stable crack propagation zone, Ginst is the fracture energy at the onset of unstable crack propagation and 1 /A 1 , Gst and Ginst can be obtained from the intercept and the slope of the straight line, respectively. It is worth noting that this semi-ductile behavior has been found in other polymers: Newmann and Williams [4] observed stable crack propagation before brittle fracture in ABS over the temperature range from —40 to 0°C; Bernal and Frontini also observed this type of behavior in a rubber-modified thermoplastic at room temperature [12].

2.3

DUCTILE FRACTURE

When the material exhibits a ductile behavior in impact fracture, unstable fracture does not occur. The crack propagation is generally completely stable.

640

T. VU-KHANH

For this type of fracture a method based on the assumption of constant fracture energy during the fracture process has been proposed [4]. The energy absorbed by the sample is considered to be proportional to the fracture surface area. The impact fracture energy is therefore determined from the slope of the plot of absorbed energy, U, against fracture surface area, A. For many ductile plastics, this method often gives abnormally high values of impact fracture energy. Furthermore, an inconsistent negative intercept of the absorbed energy versus fracture surface plot is generally observed. To explain this inconsistency in the case of fracture with ductile behavior, another approach taking into account the crack initiation and crack propagation energies in the material is needed and has been proposed [18, 19]. This approach assumes that the fracture energy of the polymer with ductile behavior varies linearly with crack extension and is given by Gr = Gi + TaA

(10)

where Gr is the actual fracture energy, Gi is the fracture energy at crack initiation, Ta represents the rate of change of Gr with crack extension and A is the fracture surface. Since the energy absorbed by the specimen is mainly dissipated in the fracture process, the energy absorbed by the specimen becomes f T f* J A /"* A i T* A 2. U= II G TdA = GlA+-TaA JA ^

f 1 1 \ (11)

giving

r/ ^

r = G.+^A

(12)

From this equation, one can obtain Gi from the intercept and Ta from the slope of the U/A versus A plot. In fact, another approach is currently widely used to characterize the fracture performance with ductile behavior. In this approach, it is suggested [20–22] that ductile fracture is governed by a single parameter that is the work dissipated per unit area in the inner fracture process zone. This parameter is called the specific essential fracture work, we, and is believed to be constant during the crack propagation period. The energy required to break a sample is considered to be the sum of the work dissipated in the inner process zone of the fracture surface (called the essential work) plus the work dissipated by plastic deformation in the outer plastic zone (called the inessential work) in the specimen [20–22]. To support this concept, experimental measurements have been carried out on deep double edge-notched tension (DENT) and deep single edge-notched (DSEN) specimens [20–22]. It has been suggested that the area under the

FRACTURE BEHAVIOR OF HIPS AND ABS

641

load-displacement diagram of the fracture test is composed of two energy quantities: an absorbed essential work that is directly proportional to the ligament length / and an inessential work dissipated in the outer plastic zone that is proportional to the square of the ligament /. The total energy absorbed by the sample to fracture, Wf, is therefore be expressed by Wf = welt + wpl2tp

(13)

where wp is the plastic work dissipation per unit volume, p is the shape factor representing the geometry of the plastic zone and t is the specimen thickness. It follows that -

= We + Wp/ll

(14)

Since Wf = U and A — lt, Equation (14) is similar to Equation (12). It has therefore been argued [23] that the impact fracture energy at crack initiation in polymers with ductile behavior Gi is rather the essential fracture work we. Furthermore, the second parameter representing the variation of the impact fracture energy during stable crack growth, Ta, has been attributed to the work dissipated in the outer plastic zone and is not related to the fracture process [23]. The essential fracture work concept, as expressed in Equation (13), implies that the slope of the Wf/lt versus / plot of Equation (14) must always be positive, since the second term on the right-hand side of Equation (13) is always positive. In this term, wp is the energy of plastic deformation per unit volume and the product f$l represents the total volume of plastic deformation in the outer plastic zone in the sample. This means that, in the limiting case when there is no outer plastic deformation (/? = 0), the slope of the plot should be zero. In order to test this assumption, measurements were carried out at various loading rates and temperatures on different polymers, using a TPB sample. The results revealed that in several cases, especially at high loading rates such as that in an impact test, these plots exhibit negative slopes even though ductile, stable fracture occurs. Figure 27.3 shows such examples for the case of HIPS. In Figure 27.3, Wf/lt is shown to decrease with increasing l. The observed negative slopes in these plots demonstrate the invalidity of the concept. Indeed, as discussed above, from Equation (14) it can be seen that a negative slope implies a negative plastic work per unit volume (wp < 0), since /?/ > 0. This means that the energy required to deform the material plastically is negative, which is physically impossible. This clearly invalidates the essential fracture work as a parameter for characterizing ductile fracture behavior. The observed negative slopes suggest that the second term on the right-hand side of Equation (13) does not represent the energy dissipated in the outer plastic zone only. This

642

T. VU-KHANH ID '

sS1

12IT-

"•>-

s

•

i " 0

_ 0

i 2

i 4

— _"""*•• ^--w# •»

i i 6 8 1 (mm–2)

i 10

_^

i 12

1<

Figure 27.3 Wf/lt versus / plot for HIPS at -10°C and impact velocity 2 m/s. Reprinted from Vu-Khanh, Theor. Appl.Tract.Mech., 21, 83 (1994), with permission of Elsevier Science

term also depends on the energy dissipated by the crack propagation process and should be taken into account in the total energy absorbed by the sample to fracture. The same observation and conclusion were also made by Bernal and Frontini [12] from their experimental measurements on toughened thermoplastics. Further evidence of the invalidity of this approach will be presented later. It should be kept in mind that the current standard methods are only qualitative and could be very misleading, as clearly stated in most standard procedures. Analysis methods based on fracture mechanics have been developed with the objective of providing a more quantitative description of the impact fracture process. The fracture parameters determined from these methods can provide useful information on fracture behavior and are better material characteristics to be used in the impact fracture characterization. However, it must also be noted that impact fracture of polymers is a complex process and care must be taken in applying the different proposed methods. Depending on the type of fracture behavior, brittle, semi-ductile or ductile, a proper analysis method must be used. Furthermore, numerous results obtained from impact tests suggest that a single parameter commonly used in the current fracture mechanics concept does not completely represent the fracture performance of a polymer. One cannot base considerations only on the single value of fracture energy at crack initiation to determine the material's impact performance and the type of crack propagation should also be considered. For instance, with the same value of Gc and Gi, the material exhibiting stable crack propagation would perform better in terms of impact resistance since, after initiation, fracture can only continue with further supply of energy by external loads, whereas in the case of brittle fracture the crack accelerates without any additional supply of energy from the external forces, leading to the phenomenon of shattering of plastic parts frequently observed in practice. To illustrate this

643

FRACTURE BEHAVIOR OF HIPS AND ABS

difference, let us consider a Charpy specimen having the dimensions thickness B — 5 mm, width D = 10 mm, initial crack a0 = 5 mm and span to width ratio S/D = 4. With the same fracture energy at initiation Gc = Gi = 10 kJ/m, depending on the crack propagation behavior, the energy required to break this specimen, U, would be the following: (1) For unstable fracture (brittle behavior), Equation(l) => U = GCBD3> = 10(5 × 10 × 0.233) = 116.5Nmm = 0.116 J (2) For stable fracture (ductile behavior) with Ta = 0.5Nmm –3 , Equation (5) =» U = GiA +1 TaA2 = (10 × 25) + (^ x 252) = 406.25 N mm = 0.406 J (3) Same as in (2) but with Ta = 0, Equation (5) => U = G i A+^A 2 = (10 × 25) + 0 - 250 N mm = 0.250 J It can be seen that with the same fracture energy at crack initiation of 10 kJ/m2, the energy required to break this sample can vary from 0.116 to 0.406 J, depending on the character of crack propagation. This example demonstrates how a single fracture parameter can be misleading in determining the impact performance of a polymer. 3

HIGH-IMPACT POLYSTYRENE

Figure 27.4 shows an example of brittle fracture behavior in the case of HIPS (STYRON 484C from Dow Chemical) at –50°C, with an impact velocity of 2 m/s. From the slope of the U versus BD$ plot, the value of fracture energy was found to be 4.5 kJ/m2. 0.5

0.4-

0.3D 0.2

y = 0.0045x + 0.2204

0.1

0 0

10

20

30

40

50

60

BD(mm 2 )

Figure 27.4

U versus BD& plot of the impact data on HIPS at -50 °C

644

T. VU-KHANH

Figure 27.5a shows the plots of U versus BD4> for the impact data on the same HIPS at -40 °C with an impact velocity of 2 m/s. At this temperature and loading speed, semi-ductile fracture occurs in HIPS. The large scatter in the U versus BD4> plot demonstrates that Equation (2) cannot be used to determine the fracture energy of the material (similar curves have been reported earlier [11]). The good correlation between U/A 1 and BD4> 1 /A 1 , as shown in Figure 27.5b, demonstrates the validity of Equation (9) for the description of fracture with semi-ductile_ behavior. From Figure 27.5b, the two fracture parameters of the material, Gst and Ginst, can be determined. Good correlations between U/A 1 and BD4> 1 /A 1 have also been reported for various polymers [11, 12]. Figure 27.6a shows the variation of absorbed energy, U, as a function of the ligament area, A, of the Charpy sample for HIPS, tested at room temperature with an impact velocity of 2 m/s. The relationship between the energy absorbed by the sample to break and the fracture surface area is relatively linear. This linear correlation is very misleading since it could suggest that the energy absorbed by the sample to break is proportional to the fracture surface area and, consequently, the fracture energy is constant during crack propagation. However, it can be clearly seen from Figure 27.6a that the intercept of such a plot is negative. This negative intercept suggests that the hammer would receive the energy when it strikes a sample containing an initial crack through its width (a sample that had been broken, a = D and A = 0). This is physically impossible and clearly invalidates the concept. In Figure 27.6b the same data as in Figure 27.6a are presented as a plot U/A versus A, according to Equation [12]. It can be seen that the correlation is also linear, validating the concept of varying fracture energy during crack extension. In this approach, the two parameters Gi and Ta characterizing fracture performance shown in Equations (10)–(12) can be determined from the intercept and the slope of the U/A versus A plot. 0.07

0.06-

0.4-

«r o.o5-

0.3-

J

0.04-

^

0.02 -

5 o-03-

0.20.1-

0.01 -

0 (a)

20 40 BD<J> (mm2)

0

60

0

(b)

20 40 BD<J>,/A1

60

Figure 27.5 Impact data on HIPS at -40 °C: (a) U versus BD4> plot; (b) U/A 1 versus BD

645

FRACTURE BEHAVIOR OF HIPS AND ABS u.uz -

I -

^

xr» y* _x» *•

0.80.6-

x

E £

g,

»'

0.4-

s

s

* * * -*->— __ __ ^ Jfc-^" "" "" - ^

0.01 -

'-*"

^*>"^'

•

<

+/ yt*

0.20"

0.015
D

0.005 -

s

A

0'

20

40 A (mm2)

(a)

60

80

0

20

40 A (mm2)

(b)

60

8

Figure 27.6 Impact data on HIPS at room temperature: (a) U versus A plot; (b) U/A versus A1 plot

3.1

EFFECT OF TEMPERATURE

The variation of fracture energy as a function of temperature for HIPS at a loading rate of 2m/s is shown in Figure 27.7. Depending on the type of fracture involved, the fracture energy corresponding to brittle, semi-ductile or ductile behavior was determined according to Equation (2), (9) or (12), respectively. It can be seen that the transition from brittle to ductile behavior occurs with a peak in the value of fracture energy. As the temperature increases, a semiductile behavior is observed before the peak and the onset of ductile behavior of fracture. In the semi-ductile behavior, a certain amount of stable crack i

20 Ih

i i i |

i i i i |

:_

i i i i

ii

1

1

• Gc (brittle) • Ginst (semi-ductil 2) A Gi (ductile)

—

~ _ —

A

1

-

A ± A _

A

8

o

—

*

A

A

-

A

4 n 1 -100

—

1

1

1

1

-50

1

1

1

1

1

1

0 Temperature (°C)

1

~ II

50

1

1

100

Figure 27.7 Variation of fracture energy as a function of temperature for HIPS at impact velocity 2 m/s. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

646

T. VU-KHANH

propagation occurs in the sample before unstable fracture. As discussed above, one cannot base considerations only on these reported values of fracture energy to determine the impact performance of the material and the type of crack propagation should also be considered. With the same value of Gc or Gi the material exhibiting a stable crack propagation performs better in terms of impact resistance since after initiation, fracture can only continue with further supply of energy by external loads. Conversely, in the case of brittle fracture, the crack accelerates without any additional supply of energy from the external forces. The character of crack propagation is therefore also important in determining fracture performance. The parameter reflecting the crack propagation is Ta in Equations (10)–(12). It represents the variation of fracture energy with crack extension. Figure 27.8 shows the variation of Ta as a function of temperature for a loading rate of 2 m/s. It can be seen that as the temperature decreases, Ta drops sharply before the onset of brittle fracture. From the variation of Gi with temperature in Figure 27.7, it can be seen that Ta tends to increase as Gi decreases. Since the yield stress varies with temperature, there seems to be a correlation between these fracture resistance parameters and the plastic zone in front of the crack tip. Furthermore, at an impact speed of 2 m/s, because the thermal conductivity of polymers is relatively low, it could be expected that the heat generated in the plastic zone could not dissipate away. The rise in temperature at the crack tip could therefore result in a localized reduction in the yield stress and a blunting effect of the crack tip [24, 25].

0.1 --

0.05 - 0--

1 -0.05--

o

f ^-0.15--

-0.2-• -0.25 - -I

-0.3

-50

0

h-

50

100

Temperature (°C)

Figure 27.8 Variation of Ta as a function of temperature for HIPS at loading rate 2 m/s. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

647

FRACTURE BEHAVIOR OF HIPS AND ABS

The parameters governing crack propagation have been proposed by Sih and Madenci [8] with the strain energy density theory. The strain energy density function, dW/dV, in front of the crack tip has been expressed in the form dW _S ~dV~~r

(15)

where S is the strain energy density factor and r is the distance from the crack tip. The crack growth condition has been established by = — — constant r^

dV

where (dW/dV)* is the available energy density that is released when a unit of macro-volume fails as (dW/dV)* becomes critical. From the relationship between the strain energy density factor and the stress intensity factor [6], the above condition for crack propagation can be rewritten using the crack growth resistance expressed in Equation (9): 1 -2v

2n

dVJ

Gi A -+Tar r

where A = Br (with B being the sample thickness). This equation gives the dependence between Gi and Ta according to the strain energy density theory. Figure 27.9 shows the plot of Gi as a function of Ta when the temperature

-0.2

-0.15

-0.1

-0.05

Ta (GJ/m4) Figure 27.9 Variation of Gi with Ta for HIPS when the temperature varies, at loading rate 2 m/s. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

648

T. VU-KHANH

varies, at a loading rate of 2 m/s. The linear correlation between Gi and Ta suggests that the crack growth increment at initiation would be a constant and is independent of temperature. The variation of Gi and Ta with temperature is a consequence of the variation in the yield stress and is governed by the critical value of the strain energy density in front of the crack tip. 3.2

EFFECTS OF LOADING RATE

The effects of loading rate on fracture energy at 23 and — 30 °C are shown in Figure 27.10. The crack propagation is stable for all loading rates at 23 °C.

jj 30-


25-

I20-

i

3^

b"

1510-

5- ' n

/

k > 4

,-•••"'•

1

10 (a)

i

1

i

i

100 1000 10000 100000 1000000 Loading speed (mm/min)

30

252015O

10 (b)

100 1000 10000 100000 1000000 Loading speed (mm/min)

Figure 27.10 Effect of loading rate on fracture energy of HIPS at (a) 23 and (b) -30 °C. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

649

FRACTURE BEHAVIOR OF HIPS AND ABS

However, at -30 °C, unstable fracture occurs in some samples for loading speeds above 6000 mm/min. For both temperatures, as strain rate augments, G{ increases slightly before decreasing to a minimum value and then rises sharply in the region of impact loading. It is noteworthy that this variation is similar to that of Gc for unstable crack propagation observed in polypropylene [24]. The sharp increase in Gi might be due to a blunting effect at the crack tip, induced by localized adiabatic heating. Figure 27.11 shows the variation of Ta with loading rate for these temperatures. In this case, like the effect of temperature, Ta also tends to vary inversely with Gi. Plots of Gi as a function of Ta are shown in Figure 27.12. It can be observed that the correlation between these parameters is also linear.

100 1000 10000 100000 1000000 Loading speed (mm/min)

(a) 0.2

0.15 0.1

0.05

0 -0.05 -0.1

-0.15 -0.2 1 (b)

10

100 1000 10000 Loading speed (mm/min)

100000 1000000

Figure 27.11 Variation of Ta with loading rate for HIPS at (a) 23 and (b) -30 °C. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

650

T. VU-KHANH

20-p 18-16-14-12-JO-

O

S'642-0-1 -0.2

1 -0.15

I— -0.1

-0.05

0

0.05

0.1

0.15

4

Ta (GJ/m )

(b)

Figure 27.12 Variation of Gi with Ta as the loading rate varies for HIPS at (a) 23 and (b) -30°C. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

From Figure 27.12, it is interesting that the intercept of Gi at Ta = 0 seems to be the same in these plots. From the strain energy density theory discussed above, Equation (17) can be rewritten as 2n l-2v

(18)

It can be seen from this equation that the same intercept of Gi versus Ta plot suggests a constant value of (dW/dV)* since r is a constant, as discussed above. The results are in agreement with those reported by others [26,27] and suggests that the critical strain energy density controls the fracture behavior of the

651

FRACTURE BEHAVIOR OF HIPS AND ABS

material. In order to verify this assumption further, the variations of Gi against Ta at various temperatures and loading rates are shown in Figure 27.13. Within the scatter of experimental measurements, there seems to be a unique linear relationship between Gi and Ta. This result suggests that the time and temperature dependence in fracture behavior of HIPS can be conveniently predicted by the strain energy density theory. Measurements carried out at various loading rates reveal that, as the temperature decreases, a peak in fracture energy at crack initiation always occurs before the onset of ductile-brittle transition in fracture behavior. However, the location of the peak changes with loading rate. The brittle ductile transition shifts to higher temperature with increasing loading rate. Figure 27.14 shows the variation of fracture energy with temperature for an impact velocity of 6 m/s. The observed peak in fracture energy is similar to that reported in the literature for many polymers [28–32]. In these reported cases, the observed peak has often been suggested to be due to the molecular relaxation process. The brittle behavior for temperatures below the peak implies that the molecular motions are limited. Above the transition temperature, the ductile behavior offracture indicates that movements of certain segments or regions of macromolecules Series 1

Series2

Series3

• Series4

Series5

40

30

20

-0.3

-0.2

-0.1

0

0.1

0.2

Ta (GJ/m4)

Figure 27.13 Variations of Gi against Ta for HIPS at various temperatures and loading speeds: series 1, loading rate 2 m/s, various temperatures; series 2, temperature 23 °C, various loading rates; series 3 temperature -30 °C, various loading rates; series 4, temperature -20 °C, various loading rates; series 5, temperature -50 °C, various loading rates. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

652

T. VU-KHANH o Gc

Gst

Ginst.

Gi

df>

'-~--\\.

35-

*/ /

3025-

\B

/

\

/

c

b" i o"„

o"

20-

\

i

\

1 i

15/

\ \

.

10- ^ 50• -60

X'

i

-40

-20

i

i

0

20

40

60

Temperature (°C) Figure 27.14 Variation of fracture energy with temperature for HIPS at impact velocity 6m/s. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

can take place, due to a thermally activated process. Although HIPS is essentially a multiphase polymer blend system, the observed peak in Gi could also be induced by a relaxation process in the irreversible plastic deformation at the crack tip prior to fracture, in a similar manner to the peak of the loss factor in viscoelasticity. It has been shown [13] that the brittle-ductile transition can be expressed by the Arrhenius equation that is commonly used for most energyactivated processes: /-A/T (19) e = A exp> V RT

where e is the strain rate, A is a constant, H is the activation energy, R is the gas constant and T is the absolute temperature. For three-point-bend samples, by ignoring the effect of the crack, the nominal strain rate can be estimated by

VD

(20)

where Vis the speed of loading, D is the specimen width and S is the span. Figure 27.15 shows the plot of In as a function of the temperature at brittle-ductile transition, T b - d . The observed linear correlation confirms that an energy activation process controls the brittle-ductile transition in the fracture behavior of HIPS. The Arrhenius equation can be used to predict the change in fracture

653

FRACTURE BEHAVIOR OF HIPS AND ABS 10

5

0--

—5

0.002

0.003

0.004

0.005 0.006 l/T h-d (K-1)

0.007

0.008

Figure 27.15 Ln versus 1 / Tb-d plot for HIPS. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

behavior in order to avoid undesirable catastrophic failure in the material. It is worth mentioning that the brittle-ductile transition in the fracture behavior of HIPS has often been explained by the glass transition of the rubbery phase. However, no direct correlation between the activation energy of the glass transition of the rubbery phase and that of the brittle-ductile transition has been reported. Consequently, care must be taken in predicting the time-temperature dependence of the brittle-ductile transition in the fracture performance of HIPS. It will be shown later that for ABS, the activation energy of the brittle-ductile transition is much lower than that controlling the Tg of PBD and is not related to the glass transition of PBD. 3.3

DYNAMIC EFFECT AND ADIABATIC HEATING

The sharp rise in fracture energy with loading rate in the region of impact loading has been attributed to a dynamic effect [33]. Using the mass-spring model, it has been shown that at very short times (high loading rate in impact tests), the higher frequency oscillations would result in a rising value of the energy measured. The sharply rising values of fracture energy at high loading rates (short times to fracture) are therefore interpreted in terms of a dynamic effect in the impact test. This interpretation cannot explain the sharp rise in fracture energy at higher loading rates observed in Figure 27.10. For the impact tests in this work, a small cushion of plasticine was put on the striker to avoid the bouncing effect. With this technique, the impact loading was progressive and all vibrations were eliminated. The rise in fracture energy is thus due to a rate effect on material

654

T. VU-KHANH

behavior, possibly combined with a blunting of the crack tip due to adiabatic heating. The suggested influence of higher order oscillations was further verified by performing measurements at low temperatures as shown in Figure 27.16. The variation of fracture energy with loading rate measured at -50 °C shows that the fracture energy remains relatively constant and does not rise at higher velocities as observed in Figure 27.4 for -30 and 23 °C. It is also observed that, at -50 CC, fracture becomes brittle in the region of high loading rates. The result confirms again the same effect of time and temperature. The increase in strain rate reduces the mobility of polymer segments and leads to a brittle behavior. It is also worth mentioning that the results in Figure 27.16 show a constant value of fracture energy. From the conventional fracture toughness criteria based only on crack initiation, one can conclude that the material exhibits the same fracture performance over the whole range of loading rates. This is not true since fracture behavior changes with loading rate. Under impact loading, fracture becomes unstable, indicating that catastrophic failure of the part occurs after crack initiation. At lower velocities, since fracture is stable, the crack cannot propagate without additional supply of energy from external loads and the part could still perform its structural function in terms of fracture resistance. 4

ACRYLONITRILE-BUTADIENE-STYRENE (ABS)

Figures 27.17 and 27.18 show respectively the plots U versus BD and U versus A for the data from an impact test at 2.5 m/s on Monsanto ABS samples. It can

15~ E

Gc (unstable fracture)

~ 10O

a" •

Gi (stable fracture)

»

\

***

* n

1

i 10

.

*

o ° *

°0

i i i —1 100 1000 10000 1000001000000 Loading speed (mm/min)

Figure 27.16 Variation of fracture energy with loading rate for HIPS at -50 °C. Reprinted from Vu-Khanh, Theor. Appl. Fract. Mech., 29(2), 75(1998) with permission of Elsevier Science

655

FRACTURE BEHAVIOR OF HIPS AND ABS -i

__ „__.„,„.,„,..... ._.....,._„

......

0.80.60.40.2-

0

0 /

5

10

15

20

25

2

BD(|) (mm )

Figure 27.17 U versus BD# plot of impact data (2.5 m/s) for ABS at room temperature

0.6-

0.4-

0.2-

0

0

/ 10

20

30

40

50

A (mm 2 )

Figure 27.18 U versus A plot of impact data (2.5 m/s) for ABS at room temperature

be seen that a relatively good linear correlation is observed for both brittle and ductile models. The fracture energy of ABS obtained from these data is 39.6 kJ/m2 with the model of brittle fracture (Figure 27.17). However, if a ductile fracture behavior with constant fracture energy during crack propagation is considered, the fracture energy of this ABS becomes 20.2 kJ/m2 (Figure 27.18), almost half the value obtained with the brittle fracture model. It should be mentioned that these values are abnormally high in comparison with that of HIPS. Furthermore, these plots show an inconsistent negative intercept. Such a negative intercept suggests that the hammer would receive energy when it

656

T. VU-KHANH

strikes a sample already broken or with an initial crack length through its width ( a = D and A = 0). This is impossible and clearly invalidates the methods. Again, care should be taken in interpreting impact data. Theoretically, as demonstrated previously for unstable crack propagation [1,2], the energy absorbed by the sample is that stored elastically up to the point of fracture. The elastic strain energy was shown to be related to the material's fracture energy via a ubiquitous factor [1,5] or a calibration factor [2]. For three-point-bend samples, these factors are not a linear function of the fracture surface. On the other hand, ductile fracture generally occurs with stable crack propagation and with a continuous supply of energy from the hammer to the sample. The energy absorbed by the sample to break is therefore not related to the elastic strain energy. Experimentally, two techniques could be used to distinguish between stable and unstable fractures. The first technique consists in observing the movement of the sample after fracture. In unstable crack propagation, the energy absorbed by the sample is that stored elastically. Only part of this elastic strain energy is dissipated in the fracture surfaces. The remaining part is transformed into kinetic energy. As a consequence, the sample generally tends to break completely into two halves and fly away after fracture. In contrast to this behavior, ductile fracture occurs with stable crack propagation and with a continuous supply of energy from the hammer to the specimen. After the fracture, the two halves remain attached together by a thin ligament and are pushed away by the hammer with a much lower velocity. The second technique is based on the observation of the fracture surface. For many polymers the fracture surface of an unstable fracture is typical. With the naked eye, the surface appears rough and often shows a branching effect as opposed to that of stable fracture, where the surface exhibits a whitening effect or could become bright, reflecting light, due to craze formation [34]. Figure 27.19 shows the variation of U/A as a function of A, according to the proposed model of crack initiation and propagation energies. From the intercept of this plot, a fracture energy value of 4.7 kJ/m2 can be obtained for ABS. This value is much lower than that determined above, using the method developed for brittle fracture or the constant fracture energy consideration. The fracture test on three-point-bend samples revealed that the fracture behavior of ABS remains ductile for temperatures above —80 °C at a crosshead speed of 5 mm/min. The maximum fracture energy at crack initiation is also observed around — 80 °C. Figure 27.20 shows that above this temperature, Gi decreases continuously with increasing temperature. When the test speed is increased to 100 mm/min, the temperature at brittle ductile transition is shifted to about -40 °C as shown in Figure 27.21. At an impact velocity of 2.5 m/s, the brittle-ductile transition occurs at around — 20 °C as shown in Figure 27.22. The results also confirm that the fracture energy at crack initiation is maximum at the brittle-ductile transition.

657

FRACTURE BEHAVIOR OF HIPS AND ABS 20

16js

12-

s 8' 4-

20

40

60

A (mm2)

Figure 27.19

U/A versus A plot of impact data (2.5 m/s) for ABS at room temperature

14

12-10

8O

642-

-120

-70

-20 Temperature (°C)

30

80

Figure 27.20 Variation of fracture energy as a function of temperature for ABS at crosshead speed 5 mm/min

Figure 27.23 shows the plot of In as a function of the temperature at the brittle-ductile transition, Tb-d, according to Equation (19). The slope of this plot gives a value of the activation energy of brittle-ductile transition of about 40 kJ/ mol. In order to investigate the role of molecular relaxation in the time-temperature dependence of the brittle-ductile transition, dynamic mechanical tests (DMTA) were also performed on the ABS sample at three frequencies, 1, 10 and 30 Hz. The frequency-temperature dependence of the loss peaks corresponding to the glass transition of PBD and SAN are shown in Figures 27.24 and

T. VU-KHANH

658 25

20

15

j=

o IDS'

• Gc • Gi -80

-30 Temperature (°C)

Ginst 20

Figure 27.21 Variation of fracture energy as a function of temperature for ABS at crosshead speed 100mm/min

14

• Gc • Ginst

12108O

642-

-80

-30

20

70

Temperature (°C)

Figure 27.22 Variation of fracture energy as a function of temperature for ABS at impact velocity 2.5 m/s

27.25, respectively, according to the Arrhenius relationship. From the slope of these plots, the activation energies corresponding to the glass transition were roughly estimated as H = 105 and 298 kJ/mol for PBD and SAN, respectively. With these limited experimental data, it can be nevertheless seen that the activation energy of the brittle-ductile transition is significantly lower than that of the glass transition of PBD and SAN. It is important to point out that the fundamental difference between DMTA and fracture tests is that DMTA only

659

FRACTURE BEHAVIOR OF HIPS AND ABS

43-

2 -

0.003

0.004

0.005

0.006

l/T b-d Figure 27.23

Ln versus I/Tb-d plot for ABS

3.5 32.5 21.5 1-

0.50-0.5 0.0042

Figure 27.24

0.0044

0.0046

0.0048

0.005

Ln (frequency) versus 1/T plot at glass transition for PBD

involves small deformations in the elastic region whereas fracture tests involve large plastic deformation at the crack tip. It is also well known that the energy dissipated in fracture is mostly related to the energy of plastic deformation at the crack tip. The yield process of a polymer is usually regarded as a momentary condition of pure viscous flow because it denotes the point at which the change of stress with strain is zero for a given strain rate. It has thus been considered to be a thermally activated process involving inter and intramolecular motion and has been described by Eyring's theory as follows [35,36]:

(21)

660

T. VU-KHANH

3.5 32.5 2-

1.510.5 0-0.5

0.0025

0.00255

0.0026

0.00265

0.0027

1/Tg (1/K)

Figure 27.25

Ln (frequency) versus 7/7 plot at glass transition for SAN

where H is the activation energy of the yield process, T the absolute temperature, ey the constant strain rate (proportional to the crosshead speed), V* the activation volume, eo the pre-exponential factor and R the universal gas constant. Eyring's theory has been successfully used to describe the yield behavior of a number of polymers, e.g. poly(methyl methacrylate) (PMMA) [37], poly (ethyl methacrylate) (PEMA) [38], isotactic polypropylene (iPP) [39], polycarbonate [36,40], polyethylene (PE) [41], poly(ethylene terephthalate) (PET) [42] and high-impact polystyrene (HIPS) [43]. Although the above concept has often been used to analyze the plastic deformation of polymers exhibiting shear yield, in the case of toughened polystyrene with multiple crazing, Eyring's equation has been successfully applied to characterize the kinetics of irreversible plastic deformation [43]. Tensile tests were performed on the ABS to verify Eyring's theory and the plots of oy/T as a function of Ine y are presented in Figure 27.26. The results show a linear correlation for three temperatures, 22, 40 and 60 °C, with loading rates varying from 10 -3 to l 0 - 1 s - 1 . From these plots, the activation volume V* can be estimated from the slope of the straight lines and the activation energy H of the yield process can be deduced from the horizontal distances between these plots [35]. The value of H determined from the experimental tensile data for ABS was about 400 kJ/mol. The result suggests a higher activation energy than that of the impact test. From these results, it could be speculated that a different relaxation process to that controlling stiffness or yielding controls the brittle-ductile transition in the fracture behavior of ABS. The same discrepancy has also been observed for polystyrene/ethylene-propylene rubber blends [44]. This has been explained by the difference in the temperature ranges involved. The tensile tests were performed in the range of

661

FRACTURE BEHAVIOR OF HIPS AND ABS 0.14 0.120.1 -

0.08H

*

0.060.04 -

0.02-

22°C

40°C

A 60°C

0 -4

-3

-2

-1

0

-1

In [strain rate (s )] Figure 27.26

Variation of of
relatively high temperatures and low loading rates, whereas the brittle—ductile transition observed in fracture tests occurs at relatively low temperatures and the deformation rates at the crack tip are much higher. For the case of polystyrene/ethylene-propylene rubber blends, the brittle—ductile transition has been found to be related to a secondary relaxation process in yield [44].

5

CONCLUSION

Fracture of HIPS and ABS is a complex process and care must be taken in characterizing the fracture performance. Depending on the loading rate and temperature, different types of fracture can occur: brittle, semi-ductile or ductile. For each fracture mode, a proper analysis method must be used. It is essential to keep in mind that fracture performance cannot be characterized by a single fracture parameter but the characteristic of crack propagation must also be considered to determine the fracture resistance at various loading rates and temperatures. A peak in the fracture energy at crack initiation always occurs before the onset of ductile-brittle transition in the fracture behavior. The location of this peak changes with loading rate. The brittle—ductile transition in both HIPS and ABS shifts to higher temperature with increasing loading rate as it is controlled by an energy-activated process. Under an impact velocity of 2 m/s, with increasing temperature the fracture mode of HIPS changes from brittle to semi-ductile at about —50°C and then becomes ductile at temperatures higher than -30 °C. However, at 10mm/min, the brittle-ductile transition

662

T. VU-KHANH

is shifted to about —100°C. At high loading rates, a sharp rise in fracture energy occurs, due to a blunting effect of the crack tip, resulting from adiabatic heating. In this case, the heat generated by plastic deformation at the crack tip cannot be dissipated away, resulting in a local softening. The fracture behavior of ABS remains ductile for temperatures above —80 °C at a crosshead speed of 5 mm/min. The maximum fracture energy at crack initiation is also observed around — 80°C. At an impact velocity of 2.5m/s, the brittle—ductile transition occurs at around —20°C. The activation energy of the brittle—ductile transition of ABS is significantly lower than that of the glass transition of PBD and SAN. A different relaxation process to that controlling the viscoelastic properties is found to control the brittle—ductile transition in the fracture behavior of ABS.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Turner, C.E. (1973) Mater. Sci. Eng. 11, 275. Marshall, G.P., Williams, J.G. and Turner, C.E. (1973) J. Mater. Sci. 8, 949. Plati, E. and Williams, J.G. (1975) Polym. Eng. Sci. 15, 470. Newmann, L.V. and Williams, J.G. (1978) Polym. Eng. Sci. 18, 893. Turner, C.E. (1980) in Fracture Mechanics: Twelfth Conference (ASTM STP 700), American Society for Testing and Materials, Philadelphia, PA, pp. 314–337. Sih, G.C. and MacDonald, B. (1974) Eng. Fract. Mech. 6, 361. Sih, G.C. (1983) Eng. Fract. Mech. 5, 365. Sih, G.C. and Madenci, E. (1983) Eng. Fract. Mech. 18, 1159. Gdoutos, E.E. and Sih, G.C. (1984) Theor. Appl. Fract. Mech. 3, 95. Vu-Khanh, T. and Fisa, B. (1990) Theor. Appl. Fract. Mech. 13, 11. Vu-Khanh, T. and de Charentenav, F.X. (1985) Polym. Eng. Sci. 25, 841. Bernal, C.R. and Frontini, P.M. (1995) Polym. Eng. Sci. 35, 1705. Vu-Khanh, T. and Yu, Z. (1997) Theor. Appl. Fract. Mech., 26, 177. Sih, G.C. (1973) Handbook of Stress Intensity Factors, Lehigh University, Bethlehem, PA. Brown, W.F. Jr and Srawley, J.E. (1996) ASTM STP 410, American Society for Testing and Materials, Philadelphia, PA. Vu-Khanh, T. and Fisa, B. (1985) Polym. Compos. (1985) 6, 249. Vu-Khanh, T. and Fisa, B., (1986) Polym. Compos. 7, 219. Vu-Khanh, T. (1988) Polymer 29, 1979. Vu-Khanh, T. (1994) Theor. Appl. Fract. Mech. 21, 83. Mai, Y.-W. and Cotterell, B. (1985) Eng. Fract. Mech. 21, 123. Mai, Y.-W. and Cotterell, B. (1986) Int.J. Fract. 32, 105. Mai, Y.-W., Cotterell, B., Horlyck, R. and Vigna, G. (1987) Polym. Eng. Sci. 27, 804. Mai, Y.-W. (1989) Polym. Communi. 30, 330. Vu-Khanh, T. and Fisa, B. (1986) Polym. Compos. 7, 375. Williams, J.G. and Hodgkinson, J.M. (1981) Proc. R. Soc. London, Ser. A 375, 231. Gdoutos, E.E. and Sih, G.C. (1984) Theor. Appl. Fract. Mech. 3, 95. Matic, P. and Sih, G.C. (1985) Theor. Appl. Fract. Mech. 3, 209.

FRACTURE BEHAVIOR OF HIPS AND ABS 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

663

Heifboer, J. (1968) J. Polym. Scl, Part C 16, 3755. Boyer, R.F. (1968) Polym. Eng. Sci. 8, 161. Vincent, P.I. (1974) Polymer 15, 111. Hartmann, B. and Lee, G.F. (1979) /. Appl. Polym. Sci. 23, 3639. Woo, L. Westphal, S. and Ling, M.T.K. (1994) Polym. Eng. Sci. 34, 420. Williams, J.G. and Adams, G.C. (1987) Int. J. Fract. 33, 209. Bucknall, C.B. (1997) Br. Plast. November, 118. McCrum, N.G. Buckley, C.P. and Bucknall, C.B. (1997) Principles of Polymer Engineering, Oxford University Press, Oxford, p. 189. Bauwens-Crowet, C. Bauwens, J.C. and Hommes, G. (1969) J. Polym. Sci., Part A2 7, 735. Roetling, J.A. (1965) Polymer 6, 311. Roetling, J.A. (1965) Polymer 6, 615. Liu, Y. and Truss, R.W. (1995) J. Polym. Sci., Part B: Polym. Physi. 33, 813. Duckett, R.A., Goswami, B.C., Smith, L.S.A., Ward, I.M. and Zihlif, A.M. (1978) Br. Polym. J. 10, 11. Truss, R.W., Clarke, P.L., Duckett, R.A. and Ward, I.M. (1984) J. Polym. Sci., Part B: Polym. Phys. 22, 191. Foot, J.S., Truss, R.W., Ward, I.M. and Duckett, R.A. (1987) J. Mater. Sci. 4, 1437. Bucknall, C.B., Davies, P. and Partridge, I.K. (1986) J. Mater. Sci. 21, 307. Ha-Anh, T. and Vu-Khanh, T. (2001) Polym. Eng. Sci. 41, 2073.

This page intentionally left blank

28

Dynamic Mechanical Behaviour of Atactic Polystyrene, High-impact Polystyrene and Other Styrenic Polymers S. N. GOYANES AND G. H. RUBIOLO Universidad Nacional de Buenos Aires, Buenos Aires, Argentina

1 INTRODUCTION Dynamic mechanical properties are the mechanical properties of materials as they are deformed in the low-strain range under periodic forces. In a typical dynamic mechanical experiment, the sample is subjected to a sinusoidal varying mechanical deformation that may be extensional, torsional shear, simple shear, bending, or some complex combination of these modes of deformation. Regardless of the type of deformation or the form of the sample, three properties are measured: the dynamic storage modulus, the dynamic loss modulus and a mechanical damping or internal friction usually called tangent delta (tanc>). When the material deformed in a low-strain range behaves as a linear viscoelastic body, the mechanical damping becomes equal to the ratio between the dynamic loss modulus and the storage modulus and then only two of the three dynamic mechanical properties are independent. The storage modulus provides a measure of the effective stiffness of the material under dynamic loading conditions. The mechanical damping indicates the amount of energy dissipated as heat during the deformation of the material. Both properties are strongly dependent on frequency and temperature. Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy (C) 2003 John Wiley & Sons Ltd

666

S. N. GOYANES AND G. H. RUBIOLO

The dynamic mechanical properties are of direct relevance to a range of unique polymer applications concerned with the isolation of vibrations or dissipation of vibrational energy in engineering components (e.g. noise reduction). Applications concerned with creep and stress relaxation under static loading conditions in engineering components can be approximately estimated from the dynamic mechanical properties determined over wide ranges of frequency and temperature. The investigation of the dynamic mechanical properties has proved to be very useful in studying the structure of high polymers and the variations of properties in relation to performance. These properties have been used to determine the glass transition region, relaxation spectra, degree of crystallinity, molecular orientation, crosslinking, phase separation, structural or morphological changes resulting from processing and chemical composition of polymer blends, graft polymer and copolymers. Comprehensive texts on dynamic mechanical spectroscopy in relation to the molecular structure of polymers are given in Refs 1–3. This chapter discusses the dynamic mechanical properties of polystyrene, styrene copolymers, rubber-modified polystyrene and rubber-modified styrene copolymers. In polystyrene, the experimental relaxation spectrum and its probable molecular origins are reviewed; further the effects on the relaxations caused by polymer structure (e.g. tacticity, molecular weight, substituents and crosslinking) and additives (e.g. plasticizers, antioxidants, UV stabilizers, flame retardants and colorants) are assessed. The main relaxation behaviour of styrene copolymers is presented and some of the effects of random copolymerization on secondary mechanical relaxation processes are illustrated on styreneco-acrylonitrile and styrene-co-methacrylic acid. Finally, in rubber-modified polystyrene and styrene copolymers, it is shown how dynamic mechanical spectroscopy can help in the characterization of rubber phase morphology through the analysis of its main relaxation loss peak.

2

POLYSTYRENE

Polystyrene (PS) as normally prepared is essentially linear and atactic. Isotactic polymers can be made but this is not of commercial interest because of increased brittleness and more difficult processing than the atactic product. The major application of polystyrene is in packaging. Specific additives are incorporated to achieve product characteristics that depend on the end usage. Atactic polystyrene (aPS) is clear, transparent and easily fabricated, and has reasonable mechanical and thermal properties but is slightly brittle and softens near 100°C. It is readily attacked by a large variety of solvents, including drycleaning agents. Its stability to outdoor weathering is poor; it turns yellow and

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

667

crazes on exposure. Many of these defects can be overcome by proper formulation or by copolymerization and blending.

2.1

EFFECT OF POLYMER STRUCTURE AND ADDITIVES ON THE DYNAMIC MECHANICAL SPECTROSCOPY OF POLYSTYRENE

The dynamic mechanical loss spectrum of polystyrene, in common with the spectra of most polymers, shows a small number of discrete loss peaks which are best resolved by a low-frequency test, preferably at or below 1 Hz. Figure 28.1 is a schematic idealized plot of the dynamic mechanical spectrum of linear, amorphous, atactic polystyrene showing the four glassy-state loss peaks: a, £, y and 8. The a loss peak is associated with the glass transition and its peak temperature with glass transition temperature Tg. Figure 28.1 is an idealized plot for several reasons: (a) the exact location of each glassy-state loss peak along the temperature scale becomes higher as the frequency of the test methods increases. Each peak shifts to higher temperatures with increasing frequency, the a (Tg) relaxation follows a Williams— Landel—Ferry (WLF) [2] relationship while the others shifts according to an Arrhenius equation [4]. (b) Each relaxation process has a different apparent activation energy, which means that each peak shifts to higher temperatures at different rates with increasing frequency and hence the peaks tend to converge into each other at very high frequency. (c) The height and location of these peaks are sensitive to prior thermal history, tacticity, crosslinking, crystallinity, diluent content and method of polymerization.

100

200 300 Temperature [K]

400

Figure 28.1 Schematic idealized plot of the dynamic mechanical spectrum of linear, amorphous, atactic polystyrene

668

S. N. GOYANES AND G. H. RUBIOLO

The four glassy-state relaxations together with the appropriate activation energy and the suggested modes of molecular motion are shown in Table 28.1. The 8 process is the simplest and best understood relaxation in aPS consisting in the molecular motion stated in Table 28.1. There is doubt in the literature about the molecular origin of -y and p relaxation. Moreover, the very existence of-y relaxation is in question [10, 12]. This question arises from the following facts [20,23]: (a) the styrene monomer Table 28.1 Peak temperatures, activation energies and postulated modes of molecular motion for the mechanical relaxations in polystyrene Activation energy Relaxation temperature (K) (kJ/mol) Molecular motion

2.4 x 10 -3 Hz

358 [5]

1Hz 870 Hz 50 x 103 Hz

379 [6] 388 [7] 392 [8] 417 [9]

1 x 10-2Hz

278 [5]

80 [5]

1Hz

303 [6]

71 [6] 146 [12]

350–417 [10, 11]

Cooperative motion of large numbers of chain backbone atoms resulting from rotational freedom of polymer backbone bonds

Local motion of a small section of main chain coupled with motions of phenyl ring

T *i

7 x 10-1 Hz

129 [13]

33 [14]

1Hz

132 [14] 163 [15] 189 [16] 200 [17] 305 [18] 308 [9]

29–42 [19]

1.7 Hz

33 [20]

9.6 [13]

5.6 Hz 6.3 x 103 Hz

38 [21] 48 [22]

12.5 [10]

11 x 103Hz 34 x 103 Hz 50 x 103 Hz

Phenyl group motions which exchange energy with a local small section of backbone chain

34 [9]

Torsional oscillation of the pendant phenyl groups coupled with a wagging motion

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

669

has a Tg value around the Ty value observed in aPS and (b) the conditions of polymerization and sample preparation were such that the presence of a small amount of residual styrene monomer could not be avoided. However, y relaxation is observed for samples with very low residual monomer although the magnitude of the y loss peak is extremely small compared with that of the 8 process [13–17]. Those who believe that aPS does have a y relaxation process discuss at least two different molecular origins. Some consider the process to be associated with specific internal or local molecular motions of the phenyl groups [19], whereas others suggest that torsional oscillation and wagging motion of the pendent phenyl groups exchange energy with a local small section of the backbone chain [17,24]. The last molecular process was also associated with p relaxation [10,12], but other authors consider only the local motion of a small section of main chain [6,24]. The p relaxation is enhanced by quenching from above the Tg and is suppressed by annealing [10]. Its enhancement by quenching, which introduces extra free volume, implies sensitivity to groups on neighbouring chains.

2.1.1

Effects of Polymer Structure

2.1.1.1

Tacticity

The mechanical 8 relaxation is slightly affected by tacticity. For an isotactic material the peak is narrower and not as pronounced as for an atactic sample [8,17]. The y peak intensity for isotactic PS is substantially lower than for atactic PS and also shifts to lower temperatures [10,17,24]. Moreover, some researchers have indicated that it is completely missing from isotactic PS [11,15]. Tacticity of the PS chain seems to have an influence on the height and breadth of the ft peak. A quenched isotactic PS shows essentially no 3 peak whereas an anionic PS (about a 1:1 ratio of syndiotactic to isotactic placements) shows a high, broad peak [10,11]. The fact that the 3 peak is suppressed by isotacticity implies sensitivity to the physical disposition of neighbouring phenyl groups along the chain. The a relaxation in both isotactic and syndiotactic PS is broader than that in atactic PS and the actual location of the peak is slightly shifted to higher temperature. The broadening effect was attributed to restrictions imposed by crystallites on the amorphous phases [12,24,25]. Nakatani et al. [26] showed how the broadness of a syndiotactic PS sample can be represented by the overlapping of two glass relaxation processes arising from one purely amorphous component and the other amorphous component, which is under restrain owing to the proximity of crystallites.

670 2.7.7.2

S. N. GOYANES AND G. H. RUBIOLO Molecular Weight

Molecular weight affects mainly the a relaxation. Although the temperature of the a peak is often taken as a value for the glass transition temperature, Tg, the two temperatures are not equivalent because the first is strongly dependent on frequency. Several authors has shown that the Tg value measured by dilatometric or differential scanning calorimetry (DSC) methods decreased when the molecular weight diminished following the relationship [12,27,28]

rg = r~ - KgM-1

(i)

where Mn is the number-average molecular weight, T^ = 373 K and /s:g = (8-19) x 104. On the other hand, it is known that the peak temperature of a relaxation shifts with a change in frequency from u>\ to 0)2 following the WLF equation [2]:

where C1 and C2 are constants, the values of which depend on the particular polymer and are independent of the molecular weight. Considering that an equivalent frequency can be associated with the DSC method [9,29], the molecular weight effect on the temperature of the a peak at a given fixed frequency can be estimated from Equations (1) and (2) as follows:

2.7.7.3

Substituents

A pictorial summary of the positions where substituents can be placed on the phenyl ring of polystyrene is given in Figure 28.2. The ortho- and meta-substituted polystyrenes do not exhibit a 8 relaxation whereas para-substituted polystyrenes do. These latter materials have a 8 relaxation that is shifted towards higher temperatures as the mass of X is increased and the size of the peak depends on the nature and number of the substituents [16,24]. The effect on the y relaxation was studied by Illers [30] on a series of poly(para-halogenated styrenes) with X = H, F, Cl, Br and I. The author also

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

671

X

'X

—CH—CH—

ortho Figure 28.2

meta

—CH—CH—

— CH—CH—

para

Pictorial summary of polystyrene substituent positions

showed that the intensity and the peak temperature increase as the mass of X is increased. No data are available for ortho- and meta-substituted polystyrenes. Gao and Harmon [6] reported a -y peak at 240 K (1 Hz) for aPS, which shifted to lower temperature with methyl and tert-butyl para substituents, the shift increasing as the size of the pendant group increases. We note that the value of the temperature of the -y peak for aPS does not correspond to that usually found in the literature. The temperature and intensity of the (3 peak scarcely change on going from H, F, Cl, Br to I in Illers' data. However, the data of Boyer and Turley [10] for three isomers of poly(chlorostyrene) show that the £ relaxation intensity decreases on going from para to ortho substituents. Gao and Harmon [6] analysed the methyl and tert-butyl para sustituent effect on the (3 relaxation in PS. Small modifications of the peak position were observed and the (3-activation energy was the same as obtained for PS The a relaxation peak is strongly affected by the sustituents. In poly(parahalogenated styrenes), the peak temperature increases from 379 to 429 K more or less in proportion to the van der Waals radius of the halogen group, H, F, Cl, Br to I [30]. Gao and Harmon [6] showed that for para sustituents TK and the width of the peak at half-height increases with increase in size of the pendant group. The effects of ring and side-chain substitution on Ta were mainly studied by DSC instead of dynamic mechanical analysis. However, as we mentioned before, the results of both methods are closely connected. Of particular interest are the results for the four methyl isomers, which show about a 100K spread in TB between m-methyl and a-methyl. Substitution at the a-position gives the greatest increase in Tg, indicating that substitution in this position restricts most the freedom of rotation around the polymer chain backbone. The o-methyl group, in poly(2-methylstyrene), also gives a high Tg value, as this will also restrict movement of the polymer chain. Substitution in the meta (3) or para (4) position results in a smaller shift of Tg as this has less effect on motions of the polymer chain. A table detailing the effects of ring and chain substituents on the glass transition temperature of polystyrene derivatives can be found in the Ref. 31.

672

S. N. GOYANES AND G. H. RUBIOLO

Abe and Hama [32] on aPS and more recently Nakatani et al. [26,33] on aPS and syndiotactic polystyrene (sPS) studied the effect of hydrogenation on the dynamic mechanical relaxation spectrum. This technique is capable of adding hydrogen to the unsaturated double bond while maintaining the molecular weight, molar mass distribution and stereochemistry of the original polymer. The former authors reported four characteristic loss peaks at 110Hz: 148K, ascribed to a 8 peak; 223 K, considered as the chair-chair transition of the cyclohexyl ring; possibly a weak (3 process at about 323 K; and the Ta peak at 413 K. The latter authors found, at 50 Hz, about a 50K spread in Ta and increasing intensity of the y peak, without a change of Ty (155K), with increasing degree of hydrogenation for both aPS and sPS. They also reported a p peak at 220 K (50 Hz) on aPS, which shows a similar behaviour to the y peak. We note that the relaxation processes ascribed by Abe and Hama to the peak at 148 K and by Nakatani et al. to the peak at 220 K do not correspond to those usually found in the literature for aPS loss peaks.

2.1.1.4

Crosslinking

Crosslinking affects mainly the a relaxation of polystyrene. Crosslink formation inhibits the molecular motion and, therefore, always causes a net increase in Ta. For example, Ishii et al. [34], working on a copolymer of styrene and 4-ethylstyrene (8 mol%), found a shift of 34K (1 Hz) in Ta towards high temperature. The crosslinked polystyrenes of major commercial interest are copolymers of styrene and divinylbenzene. Results of a systematic study of this system have shown that Tg could be represented by the equation [35]

rg = i? + ^n

c

(4)

where T0g is the glass transition temperature for the uncrosslinked polystyrene and nc is the average number of atoms in the polymer backbone between crosslinks. Illers and Jenckel [36] and Boyer and Turley [10] found that aPS crosslinked with divinylbenzene increases the (3 peak height compared with the polystyrene blank. Studies related to modifications of 8 and -y relaxations under the effect of Crosslinking in aPS are scarce and ambiguous in the literature [14,22].

2.1.2

Effects of Additives

Polymer additives are substances placed in polystyrene to enhance or improve specific characteristics. Common additives used in various formulations include

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

673

some combination of the following: plasticizers, antioxidants, UV stabilizers, flame retardants and colorants. Plastizers are miscible with the polymer whereas the others additives can behave as inert filler or as diluents. Diluents affect the free volume of the polymer. As the free volume increases, the glass transition temperature tends to decrease. The depression of Tg with the addition of diluents, including gases such as CO2, can be estimated following Royer [37] by the expression In 1 _| J\l

I

= f[(l - 0)ln(l - 0) + 6\n0]

(5)

V «/

with ._ M p w

I — co zR

where T0g is the glass transition temperature for the undiluted polystyrene, co is the weight fraction of diluent, z is a lattice coordination number for the polymer repeat unit, ACP? js is the change in heat capacity associated with the glass transition temperature, R is the universal gas constant and Mp and Md are the molecular weights of the monomer and diluent, respectively. The effect of additives, others than plasticizers, on the dynamic mechanical properties of styrene polymers have attracted little attention from researchers. Flame retardants such as l,2-bis(tetrabromophthalimide)ethane, crystalline decabromodiphenyl oxide (DBDPO) and antimony trioxide (Sb2O3) do not affect the a relaxation of aPS [38,39]. Illers and Jenckel [36] studied aPS plasticized with diethyl, dibutyl and dioctyl phthalate. The mechanical measurements of tand at 1 Hz showed the a relaxation moving rapidly to lower temperatures with increasing plasticizer content. The 3 relaxation was submerged below the a relaxation in the plasticized systems. A relaxation process was reported with a peak temperature ranging from 180 up to 267 K on addition of diethyl, dibutyl and dioctyl phthalate, respectively. The authors called this relaxation process -y but they were inclined to ascribe it to the motion of low molecular weight polystyrene dissolved in discrete droplets of plasticizer. According to Heijboer [4], this can also be a motion within the plasticizer molecule, e.g. of the n-butyl group in dibutyl phthalate. Morgan and Nielsen [23] have shown that various solvents cause mechanical loss peaks in polystyrene in the 8 and y regions. They visualized these peaks as being related to the glass transition of diluent molecules. However, it was also

674

S. N. GOYANES AND G. H. RUBIOLO

shown that a number of the polymer—diluent systems give damping peaks below the expected glass transition temperature characteristic of the liquid solvent. The lowering of the solvent Tg was ascribed to interactions with the polymer, molecular packing and the size of the diluent clusters, or a combination of these effects. The authors stated that the original aim of their work was to determine the effect of the organic solvents on the polystyrene 8 peak, but they recognized that the phenomena turned out to be so complex that they had no explanation for it. In particular, they found a very prominent damping peak at 115K in the polystyrene—14.5wt% styrene system, from which they suggested that the so-called -y peak in polystyrene could be due to styrene monomer. We doubt that this is a proof of the source of the -y relaxation in polystyrene because, as the authors mentioned, their results with the polystyrene-toluene system showed complex changes with the toluene concentration and they did not measure those changes in the polystyrene—styrene system. Besides, this subject was not studied after Morgan and Nielsen's work and newer literature data [6,26,33,40] on the y relaxation in polystyrene and its derivatives is always interpreted as resulting from phenyl group motions.

2.1.3

Summary of the Anelastic Spectrum

If one considers only the results on para-substituded polystyrenes, some important generalizations can be drawn: 1. The 8 and -y relaxations are shifted towards higher temperatures as the mass of substituents is increased. 2. When the mass of substituents is changed, the activation energy for the 8 relaxation seems to stay constant. There are no data on the behaviour of the 7 activation energy. 3. The temperature, intensity and activation energy of the (3 peak scarcely change with variation in the mass of substituents. 4. The a relaxation is shifted towards higher temperatures as the size of the pendant group increases. We speculate that the 8 and -y relaxations may have their origin in the following model. First, we accept that torsional oscillation of the pendent phenyl groups coupled with a wagging motion with a frozen backbone chain is the origin of 8 relaxation. However, the mass effect of para substituents indicates that the torsional oscillation is dominant because additional mass in a wagging motion, acting like an oscillation of a double pendulum, must shift the eigenfrequency to higher values. Then, at constant external applied frequency and activation energy, the peak temperature must decrease given the inverse observed behaviour.

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

675

Second, as the temperature is increased, the backbone become locally unfrozen and the motions of the phenyl group exchange energy through a local small section of backbone chain. The energy alternates between one phenyl group and the other with a frequency which we ascribe to that of the y relaxation. Independent of the type of oscillation motion developed by the phenyl group, the 'coupling' frequency is inversely proportional to the eigenfrequency of each phenyl group oscillation. Taking into account the behaviour of Ty with the mass of the para substituents and the additional mass effect on the motions of phenyl group explained above, we conclude that only the wagging motion is involved in the y relaxation. Third, torsional and other types of internal oscillation of the phenyl group can be coupled through the backbone chain at higher temperature. We think that this is the case for the peak reported by Gao and Harmon [6] at 240 K (1 Hz) for aPS. They showed that it shifts to lower temperature with methyl and tert-butyl para substituents, the shift increasing with increase in size of the pendant group, so the effect of additional mass is the inverse of that involved with the -y peak. Moreover, this model should explain the origin of the two peaks reported for hydrogenated aPS and sPS around 150 and 220K [26,32,33]. Further, if these data and those of Gao et al. [6] are expressed in an Arrhenius plot together with the peak temperature and frequency values for the y relaxation given in Table 28.1, then two lines result (Figure 28.3). Therefore, we think that the data in Table 28.1 which lie on the higher temperature line should be reconsidered as coming from the third relaxation process described above. Given the facts that the £ peak does not change appreciably in temperature, height or shape for polystyrene and polyCpara-substituted styrene) even though

4

5

6

7

1/T X 103 (K-1)

Figure 28.3 Secondary y relaxation map of aPS. Numbers refer to references. (A) Data for hydrogenated aPS and sPS; •) datum for aPS from Gao and Harmon [6]

676

S. N. GOYANES AND G. H. RUBIOLO

Ta increases in the same series and its intensity decreases on going from para to ortho substituents, we would argue that a major component of the P peak involves the wagging motion of small sections of the backbone chain lightly coupled with the oscillation motions of phenyl groups. The wagging motion is considered as vibrational oscillation of short chain segments about their mean position as described in the local relaxation mode theories reviewed by Saito et al. [41]. The characteristic frequency of the /*th normal mode is given by (6)

where Cu is the vibrational force constant for a group of mass m. The constant Cu is proportional to van der Waals cross-sectional area of the polymer chain. The results for poly(para-substituted styrene) indicated that the mass of the group increases together with the van der Waals cross-sectional area.

3

COPOLYMERS OF STYRENE

The varieties of copolymers that can be prepared with styrene have greatly expanded the use of the monomer. Dramatic improvements or modifications of physical properties can be achieved by choosing the right comonomer. The dynamic mechanical properties of these copolymers are strongly influenced by the characteristics of the comonomer, the copolymer composition and the miscibility parameter of the constituents parts to function as separable identities. Random copolymers should exhibit only one a relaxation the peak temperature of which will be between the Ta temperatures of the constituents homopolymers according to the equation [42,43]

-\ Ta2\

^'

where w1 and w2 are the weight fractions of the two monomers whose homopolymers have transitions a temperature Ta1 and Ta2, respectively, and B is constant which is close to unity. Block copolymers sometimes yield two a relaxations (associated with the a relaxations of the two homopolymers) and sometimes a single a relaxations (as would a random copolymer). This behaviour can be understood as being related to the mutual solubility of the two constituent homopolymers. If the homopolymers have a certain grade of solubility, there is only one a peak, which is broader than the a peak of a random copolymer with the same

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

677

composition; it is caused by the increase in heterogeneity of composition in different chains. If the homopolymers tend to be insoluble in one another, microphase segregation can occur and their a transitions can be resolved as two peaks by measurements of the dynamic loss modulus or the mechanical damping. Some of the effects of random copolymerization on secondary mechanical relaxation processes will be illustrated here on styrene-co-acrylonitrile (SAN) and on styrene-co-methacrylic acid (SMAA). SAN is one of the largest volume copolymers of styrene. Although acrylonitrile homopolymer is widely used in the fibre industry, it is essentially intractable and unsuitable as a moulding resin. However, modification of polystyrene by copolymerization with acrylonitrile results in resins that have higher tensile strength, better toughness and better solvent resistance than polystyrene. In order to avoid composition drift and nonhomogeneity, SAN copolymers are often manufactured with a composition around 0.62 mole fraction or 76 wt% styrene. Around this composition the anelastic spectrum of SAN, measured at 1 Hz, should show the a relaxation at 384 K [44], the p peak at 317 K [10] and the y relaxation at 157 K [40]. In Ref. 40 a loss peak appears at 243 K and the authors ascribed it to a (3 peak in clear contradiction with the data in Ref. 10. The strengths of both peaks reported in Ref. 40 seems to increase with increase in AN content. Murayama [45] showed a dynamic loss modulus spectrum of SAN measured at 138 Hz with the a relaxation peak at 384 K and only one secondary relaxation at about 182 K (y transition). The (3 relaxation is not exhibited at this frequency because it merges into the main relaxation. According to the model that we have proposed for the y relaxation in the aPS homopolymer, the behaviour of this relaxation in SAN can be explained as follows. First, the strength of the coupled motion through a section of backbone chain between two phenyl groups with a AN group in the middle is larger than that between two consecutive phenyl groups due to a local increase in free volume. Second, the AN group involved in the coupled motion has an incomplete transfer of energy with the phenyl groups because it is presumed to be out of tune. Third, and taking into account the above points, we assume that the number of units in the backbone chain involved in the coupled motion is larger than that in the y relaxation of aPS homopolymer and, therefore, the activation energy increases given a shift of the y peak towards higher temperature. This model also holds for the second peak reported in Ref. 40 where, as in the aPS homopolymer, torsional and other types of internal oscillation of the phenyl group can be coupled through the backbone chain. In fact, the model predicts a higher strength of both peaks with increased AN content in the copolymer. The shift to higher temperature of the 3 peak in SAN with respect to aPS homopolymer can be understood by considering Equation (6). It is expected that Cu will take the same value in SAN and aPS because the van der Waals cross-sectional area is not affected. Therefore, owing to the decrease in the mass

678

S. N. GOYANES AND G. H. RUBIOLO

of the polymer chain, the characteristic frequency of the nth normal mode increases and, at constant activation energy, the peak temperature must also increase. lonomers are ion-containing polymers having a small amount (usually up to 10–15mol%) of ionic groups along nonionic backbone chains. Ionic groups have dramatic effects on the thermo-mechanical properties of polymers, leading to a range of new applications. One of the best known ionomers based on aPS homopolymer is SMAA. Alberola et al. [46] studied a series of SMAA with various amounts of methacrylic acid in the temperature range 100–450 K by high-resolution dynamic mechanical spectrometry, showing three mechanical relaxations: a, p and y. With increasing molar fraction of methacrylic acid, the a relaxation measured at 1 Hz is greatly shifted toward higher temperatures, from 399 K for the lowest acid level (10%) to 420 K for the highest level (22%), and its strength decreases. The fi relaxation is a weak but well-defined peak located at about 350 K at 1 Hz. It appears as a shoulder on the main relaxation at lower acid level and it becomes progressively separated with increasing molar fraction of methacrylic acid. The activation energy of p relaxation was not determined. At lower temperatures, the copolymer shows the y relaxation with a well-defined peak located at about 168 K at 1 Hz. An important feature of this relaxation is its high strength compared with the aPS homopolymer and with that of SAN, this strength increasing with increasing molar fraction of methacrylic acid. The activation energy of the y relaxation is about 40 kJ/mol and it is independent of the molar fraction of methacrylic acid. The characteristics of the dynamic mechanical spectrum of SMAA show drastic changes compared with those of the aPS homopolymer even at very low molar fractions of the added comonomer. All the changes observed reflect the ionic interactions. The a relaxation temperature increases with increasing methacrylic acid content as a consequence of a stable network of chemical crosslinks due to anhydride bridge formation. The y relaxation could be related to local motion of methacrylic acid due to the breakdown of the weakest hydrogen bonds. The (3 relaxation could be attributed to local motion of the backbone chain induced by the breakdown of stronger hydrogen bonds than those invoked for the y relaxation.

4

RUBBER-MODIFIED POLYSTYRENE (HIPS) AND SAN COPOLYMERS (ABS)

Rubber is incorporated into polystyrene primarily to impart toughness. The resulting materials consist of a polystyrene matrix with small inclusions of the rubber (usually 5–10wt% polybutadiene or copolymer rubber). They are termed high-impact polystyrene (HIPS). Grafting of the rubber to the polystyr-

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

679

ene may occur as the rubber is present during the styrene polymerization. These materials are the most effective in enhanced impact strength, particularly if the rubber is slightly crosslinked, but mechanical blends are also used. Like HIPS, acrylonitrile—styrene—butadiene (ABS) polymers have polybutadiene rubber incorporated into styrene—acrylonitrile copolymer (SAN), giving a resin consisting of a two-phase system with inclusions of rubber in a continuous glassy matrix. Again, development of the best properties requires grafting between the glassy and rubbery phases. In rubber-modified polystyrenes, the rubber is dispersed in the polystyrene matrix in the form of discrete particles. The two-phase nature of rubbermodified polystyrene was first suggested by Buchdahl and Nielsen [47] based on data on dynamic mechanical properties obtained with a torsion pendulum. The existence of two prominent loss peaks led to this conclusion, one at low temperatures which is due to the a relaxation of the rubber (e.g. 193 K for polybutadiene) and one at high temperatures which is due to the a relaxation of the matrix (e.g. 373 K for polystyrene). Later, microscopy provided proof of the existence of the rubber phase as a discrete dispersed phase in polystyrene [48]. The morphology of the rubber-modified polystyrenes system involves some complex aspects, such as particle size, size distribution, occlusions of polystyrene inside the rubber phase, interfacial bonding between the rubbery particles and the brittle matrix, etc. Many authors have observed that some of the most important factors in controlling the mechanical properties of HIPS and ABS are rubber particle size [49], volume fraction of the rubbery phase (rubber + occluded polystyrene) [50,51] and the degree of graft [52]. Grafting occurs during the polymerization of styrene when some of the free radicals react with the rubber There is a relatively sharp drop in storage modulus corresponding to the a relaxation of the rubber phase. At temperature above the Tg of the rubber, the storage modulus of the rubber becomes negligible compared with that of the rigid matrix, and the modulus of the composite is due solely to the matrix. Under these conditions, the system can be described by the modified Kerner equation [12,53]: 7 – 5v

where G' is the composite storage modulus, G\, v1 and (fr1 are the storage modulus, Poisson's ratio and the volume fraction of the rigid matrix, respectively, v is Poisson's ratio for the rubber and is the volume fraction for the rubber phase. This equation can be used to estimate the rubber phase volume fraction measuring the composite storage modulus around temperatures in the middle of the range between the rubber and rigid matrix glass transition temperatures [9].

680

S. N. GOYANES AND G. H. RUBIOLO

Dynamic mechanical characterization of rubber-modified polystyrenes or styrene copolymers can reveal some of the variables in the structure of the rubber phase. The tan<5 peak at lower temperature, which can be considered as representative of the rubber a relaxation, was shown to be affected by the volume fraction of the rubbery phase, , in both HIPS [54] and ABS [55]. Wagner and Robeson [54] have shown that for $ values lower than 20%, the a relaxation temperature of rubber phase, T apr , is always lower than that for the pure rubber, T a r . The absolute value of the shift between T apr and Tar decreases while the strength of the relaxation increases with increasing at constant rubber weight fraction. It must be pointed out that in this research an increase in , by inclusion of polystyrene sub-particles inside the rubber phase, led to an increase in the particle size of the rubber phase. Giaconi et al. [55] studied dilute samples prepared by extrusion blending between a pure SAN copolymer and ABS materials with given structure and size of the rubber phase particles. This procedure gives samples with various levels of (f> while maintaining constant the structure and size of the rubber phase particles. The ABS samples having bulk rubber particles, with almost no subincluded SAN, show similar features in the loss peak of the rubber phase to those observed by Wagner and Robeson in HIPS. Moreover, when the size of the rubber particles increases, a splitting of the loss peak into two smaller ones was observed at the lowest values of (j>. The ABS samples having a high included SAN content inside the rubber phase, with a 'salami' morphology similar to that of conventional HIPS, show an increase in the relaxation strength without a Tapr shift on increasing . We think that both dynamic mechanical data obtained by Wagner and Robeson and Giaconi et al. can be interpreted based on the considerations of the latter group. They ascribed the Tapr shift to the thermal stress arising inside the rubber due to the thermal expansion mismatch between the rubber and the surrounding matrix. When the material is cooled below the glass transition temperature of SAN, the rubber, which has a higher thermal expansion coefficient, undergoes a hydrostatic dilatation stress. This stress is higher for lower <j> at constant rubber particle size because the relevant thermal expansion mismatch is actually that existing between the rubber inside each particle and the surrounding two-phase material as a whole, whose expansion coefficient is intermediate between those of pure rubber and pure SAN and dependent on (j). It was shown that the thermal stress, which is maximum for a pure rubber particle with no sub-inclusion, decreases strongly when a single spherical rigid sub-inclusion is added inside the rubber particle forming a composite particle like the typical rubber phase [56]. Then, increasing the sub-inclusion content decreases the buildup of thermal stresses and accounts for the decrease in the shift between Tapr and Tar on increasing (j) at constant rubber weight fraction. This effect continues until a lack of peak shifts is reached at a given particle size. Peak splitting, which is observed for rubber particles with almost no sub-

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

681

inclusion, might be due to a weaker interfacial adhesion in a large particle sample or cavitation phenomena inside the rubber particles under the action of the strong thermal stress existing at the low temperatures reached during the tests. The increase in the relaxation strength with increase in 4> was not interpreted by Giaconi et al. Earlier, this effect was ascribed to the storage modulus relaxation strength which increases with increasing 0 [57], but, its magnitude is insufficient to take into account such changes in the tan<5 relaxation strength. We think that this effect in Giaconi et al.'sdata is easily explained because the pure rubber content always increased with increase in <£, but this is not the case for Wagner and Robeson's data. To explain the latter, it is necessary to consider an interfacial component of mechanical loss, which increases as a result of two factors at constant rubber weight fraction: the volume fraction of the rubbery phase, 0, and the size of the sub-inclusions inside the rubbery phase. Choi et al. [58] studied the effect of the degree of graft in HIPS with a constant average diameter of rubber phase particle, rubber volume fraction and rubber weight fraction. It was pointed out that increasing the degree of graft causes a larger size of polystyrene sub-inclusions. They observed that the tan<5> maximum temperature, Tapr, shifts to higher temperatures as the degree of graft increases. The magnitude of the maximum tanc) as a function of the degree of graft goes through a maximum around to 100–120%. In the frame of the model suggested by Giaconi et al. for the dynamic mechanical behaviour of the rubber phase, we think that the modification introduced by grafting has to be understood as follows. First, Tapr shifts to higher temperature as Tar rises by increasing the grafted polystyrene on rubber chains. Second, the maximum tand evolution with grafting is a competition between the increase in the interfacial component of mechanical loss through the chains grafted to the rubber surfaces and the decrease in total area of sub-inclusions due to the size increment of each of them holding constant the average diameter of rubber phase particles.

REFERENCES 1. Nielsen L. E and Landel R. F., Mechanical Properties of Polymers and Composites, Marcel Dekker, New York, 1994, Chapt. 4. 2. Ferry J. D., Viscoelastic Properties of Polymers, John Wiley & Sons Inc., New York, 1980, Chapt. 2 and 11. 3. Read B. E. and Dean G. D., The Determination of Dynamic Properties of Polymers and Composites, Adam Hilger, Bristol, 1978 4. Heijboer J., in Molecular Basis of Transitions and Relaxations, Meier D. J., ed., Gordon and Breach, London, 1978, pp. 75–102.

682 5. 6. 7. 8. 9. 10.

S. N. GOYANES AND G. H. RUBIOLO

Cavaille J. Y., Jourdan C. and Perez J., J. Polym. Sci. B, 25, 1235 (1987). Gao H. and Harmon J. P., Thermochim. Acta, 284, 85 (1996). Simonsen J., Jacobsen R. and Rowell R., J. Appl. Polym. Sci., 68, 1567 (1998). Wall R. A., Sauer J. A. and Woodward A. E., J. Polym. Sci., 35, 281 (1959). Goyanes S. N., J. Appl. Polym Sci., 65, 865 (2000). Boyer R. F. and Turley S. G., in Molecular Basis of Transitions and Relaxations, Meier D. J., ed., Gordon and Breach, London, 1978, pp. 333-358. 11. Boyer R. F., in Encyclopedia of Polymer Science and Technology. Vol. 13: Physical Properties, Mark H. F. and Gaylord N. G., eds, John Wiley & Sons Inc., New York, 1970, pp. 277-289. 12. Berglund C. A., in Encyclopedia of Polymer Science and Engineering. Vol. 16: Styrene Polymers, Mark H. F., Bikales N. M., Overberger C. G. and Menges G., eds, John Wiley & Sons Inc., New York, 1986, pp. 142-148. 13. Schmieder K. and Wolf K., Kolloid Z., 134, 149 (1953). 14. Illers K. H. and Jenckel E. J., J. Polym. Sci., 41, 528 (1959). 15. Turley S. G. and Keskula H., J. Polym Sci., 14, 69 (1966). 16. Baccaredda M., Butta E. and Frosini V., J. Polym. Sci. B, 3, 189 (1965). 17. Yano O., and Wada Y., J. Polym. Sci. A2, 9, 669 (1971). 18. Yamamoto K. and Wada Y., J. Phys. Soc. Jpn., 12, 374 (1957). 19. Reich S. and Eisenberg A., J. Polym. Sci. A2, 10, 1397 (1972). 20. Chung, C. I. and Sauer, J. A., J. Polym. Sci. A2, 9, 1097 (1971). 21. Sinnott K. M., SPE Trans., 2, 65 (1962). 22. Crissman J. M. and McCammon R. D., J. Acoust. Soc. Am., 34, 1703 (1962). 23. Morgan, R. J. and Nielsen, L. E. J. Polym. Sci. A2, 10, 1575 (1972). 24. Roberts G. E. and White E. F. T., in The Physics of Glassy Polymers, Haward R. N., ed., Applied Science, London, 1973, Chapt. 3. 25. McCrum N. G., Read B. E. and Williams G., Anelastic and Dielectric Effects in Polymers Solids, Dover, New York, 1967. 26. Nakatani H., Nitta K. and Soga K., Polymer, 40, 1547 (1999). 27. Yu Z., Yahsi U., Mcgervey J. D., Jamieson A. M. and Simha R., J. Polym. Sci. B, 32, 2637 (1994). 28. Inoue T., Onogi T., Yao M. and Osaki K., J. Polym. Sci. B, 37, 389 (1999). 29. Hagen R., Salmen L., Lavebratt H. and Stenberg B., Polym. Test., 13, 113 (1994). 30. Illers K. H., Z. Elektrochem., 65, 679 (1961). 31. Lee W. A. and Rutherford R. A., in Polymer Handbook, 2nd edn, Brandrup J. and Immergut E. H., eds, John Wiley & Sons Inc., New York, 1975, Sect. III, p. 139 32. Abe A. and Hama T., Polym. Lett., 7, 427 (1969). 33. Nakatani H., Nitta K. H. and Soga K., Polymer, 39, 4273 (1998). 34. Ishii F., Hirahata W., Yokota K., Tuda K., Hirao A. and Kakuchi T., J. Polym. Sci. B, 37, 3319 (1999). 35. Boyer R. F., in Encyclopedia of Polymer Science and Technology. Vol. 13: Physical Properties. Mark H. F. and Gaylord N. G., eds, John Wiley & Sons Inc., New York, 1970, pp. 309-326. 36. Illers K. H. and Jenckel E., Rheol. Acta, 1, 322 (1958). 37. Royer J. R., PhD Thesis, North Carolina State University, 2000. 38. Radloff D., Spiess H. W., Books J. T. and Dowling K. C., J. Appl. Polym. Sci., 60, 715 (1996). 39. Owen S. R. and Harper J. F., Polym. Degrad. Stab., 64, 449 (1999). 40. Capitani D., Segre A. L., Pentimalli M., Ragni P., Ferrando A., Castellani L. and Blicharski J. S., Macromolecules, 31, 3088 (1998).

DYNAMIC MECHANICAL BEHAVIOUR OF STYRENIC POLYMERS

683

41. Saito N., Okano K., Iwayanagi S. and Hideshima T., in Solid State Physics, Ehrenreich H., Seitz F. and Turnbull D., eds, Academic Press, New York, 1963, Vol. 14. 42. Mandelkern L., Martin G. M. and Quinn F. A., J. Res. Nad. Bur. Stand., 58, 137 (1959). 43. Fox T. G. and Loshaek S., J. Polym. Sci., 15, 371 (1955). 44. Robertson C. G. and Wilkes G. L., Polymer, 42, 1581 (2001). 45. Murayama T., in Encyclopedia of Polymer Science and Engineering. Vol. 5: Dynamic Mechanical Properties, Mark H. F., Bikales N. M., Overberger C. G. and Menges G., eds, John Wiley & Sons Inc., New York, 1986, pp. 299-329. 46. Alberola N., Bergeret A., Battesti P. and Revillon A., J. Appl. Polym. Sci., 48, 2041 (1993). 47. Buchdahl R. and Nielsen L. E., J. Polym. Sci., 15, 1 (1955). 48. Kato K., Polym. Eng. Sci., 7, 38 (1967). 49. Donald A. M. and Kramer E. J., J. Appl. Polym. Sci., 27, 3729 (1982). 50. Bucknall C. B., Cote F. F. P. and Partridge I. K., J. Mater. Sci., 21, 301 (1986). 51. Bucknall C. B., Cote F. F. P. and Partridge I. K., J. Mater. Sci., 21, 307 (1986). 52. Hasegawa R., Aoki Y. and Doi M., Macromolecules, 29, 6656 (1996). 53. Nielsen L. E. and Landel R. F., Mechanical Properties of Polymers and Composites, Marcel Dekker, New York, 1994, Chapt. 7. 54. Wagner E. R. and Robeson L. M., Rubber Chem. Technol, 43, 1129 (1970). 55. Giaconi G. F., Castellani L., Maestrini C. and Ricco T., Polymer, 39, 6315 (1998). 56. Pavan A. and Ricco T., J. Mater. Sci. Lett., 11, 1180 (1976). 57. Mann J. and Williamson G. R., in The Physics of Glassy Polymers, Haward R. N., ed., Applied Science, London, 1973, Chapt. 8. 58. Choi J. H., Ahn K. H. and Kim S. Y., Polymer, 41, 5229 (2000).

This page intentionally left blank

29

Flame-retardant Polystyrene: Theory and Practice BRUCE KING The Dow Chemical Company, Midland, Ml, USA

1

INTRODUCTION

Polystyrene is an inherently flammable polymer. When polystyrene is exposed to temperatures encountered in a flame, it depolymerizes almost completely to form flammable monomer and oligomers [1]. For applications that require reduced flammability, it is possible to slow the burning process with the correct choice of flame retarding additives. For many applications, the goal of flameretardant polystyrene is to prevent small ignition sources (such as electrical short-circuits in televisions) from causing a large fire [2]. In this case, the point of interest is the polymer's tendency to extinguish or spread the flame. Tests, such as UL 94 (Underwriter's Laboratory), have been developed to evaluate this tendency (see Section 3). In the United States, organizations such as the US Consumer Product Safety Commission and the National Association of State Fire Marshals have been instrumental in promoting fire safety of plastics in these applications. In many cases, it is not required by law that flame-retardant polymers be used. Rather, the UL seal is necessary for acceptance by consumers that the products meet the necessary requirements of fire safety. There are a number of flame-retardant styrenic polymers that will be covered in this chapter. These include polystyrene itself, rubber-modified polystyrene [high-impact polystyrene (HIPS)] and rubber-modified styrene—acrylonitrile copolymer [acrylonitrile—butadiene—styrene (ABS)]. Blends with styrenic Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy ,c 2003 John Wiley & Sons Ltd

686

B. KING

polymers will also be covered, including HIPS—poly(phenylene oxide) (PPO) and polycarbonate (PC)—ABS.

2 APPLICATIONS OF FLAME-RETARDANT STYRENIC POLYMERS Flame-retardant polystyrene is used primarily in expanded foam for building insulation. Rubber-modified styrenic polymers are flame retarded for use in a number of applications, such as enclosures for electronics and business equipment. By far the largest volume flame-retardant HIPS application is television enclosures (Figure 29.1); these are made primarily from flame-retardant HIPS [3]. Flame-retardant HIPS has an attractive balance of mechanical properties, processability and cost. Flame-retardant styrenic blends such as HIPS—PPO and PC—ABS also find utility in a number of electrical applications such as printers, computers and monitors. These blends have received increasing attention recently because of their ability to be flame retarded with nonhalogen flame retardants (see Section 7).

Figure 29.1

Television enclosure made from flame-retardant HIPS

FLAME-RETARDANT POLYSTYRENE

3

687

FLAMMABILITY REQUIREMENTS AND TESTS

Many tests and methods have been developed to study flammability, but only those which are important for flame-retardant styrenic polymers will be considered here. Some of these tests are regulatory requirements for specific applications, while others are more for research purposes. The flame retarding of styrenic polymers is often done to pass a specific test, and the formulation needed to pass one test may be completely different to that required for another test.

3.1 3.1.1

REGULATORY TEST METHODS Steiner Tunnel

The Steiner Tunnel test (ASTM E 84) is used to classify the fire-spread potential of products used in wall and ceiling linings [4], and is used to classify expanded polystyrene foam. In this method, specimens are placed on the ceiling of a 24 ft long tunnel. An 88 kW natural gas burner is placed at one end of the tunnel and a forced-air draft with a velocity of 1.22 m/s is introduced. The flame spread is recorded as a function of time and an arbitrary index is calculated from the measurements. This test has been criticized because it does not simulate actual building fire conditions [5,6]. An additional problem with foamed samples is that the specimens either retract out of the reach of the flame or drip on to the floor of the tunnel. In Canada this has been addressed by using a downward-facing burner and mounting the specimens on the floor of the tunnel. Despite its limitations, the Steiner Tunnel method continues to be used to test and rate thermoplastic foams.

3.1.2 UL 94 The UL 94 standard includes a number of different tests, all of which use a small flame burner [7]. The most common of these is the UL 94 V vertical burning test; passing this test is a requirement for television enclosures sold in North America. For this test, a 50 W methane flame is used. A 13 mm wide by 125 mm high specimen bar is suspended with the bottom of the bar 10 mm above the burner tube. The bars can range in thickness from 1.6 to as much as 13 mm, depending on the requirements of the particular application. The flame is applied for 10s and then removed; after the bar has extinguished, the flame is immediately reapplied for a further 10s. A total of five specimens are tested. Figure 29.2 shows various stages of the UL 94 V test.

688

3. KING

Figure 29.2 Various stages of the UL 94 vertical burning test: (a) application of flame; (b) observation of burning time; (c) flame extinguishes

There are a number of different 94 V classifications, the best being 94 V-0. To pass this requirement, none of the individual burn times can be greater than 10 s and the total of the 10 burn times must be less than 50 s. There are a number of other requirements for this rating: none of the specimens may burn up to the clamp, none may drip flaming particles that ignite cotton-wool placed under the specimen and for the second flames the total of the burn time and afterglow cannot exceed 30 s. For a UL 94 V-l rating the individual burn times are less than 30s and the total of all burn times is less than 250 s; in addition, there should not be any flaming drips. For a UL 94 V-2 rating the same restrictions apply for burn times, but flaming drips are allowed. For samples that cannot meet the requirements for the UL 94 V test, the UL 94 HB horizontal burn test is a possibility. Flame spread in the horizontal orientation is much slower than for the vertical test. Two marks are made on the bar at 25.4 and 101.6mm from the free end of the specimen. After the flame has been applied for 30 s, the burn rate between the reference marks

FLAME-RETARDANT POLYSTYRENE

689

is observed. For specimens greater than 3 mm in thickness, the burn rate between the reference marks should not be more than 38 mm/min; for specimens thinner than 3 mm, the burn rate between the marks should not be more than 76 mm/min. Styrenic polymers, in the absence of flame retardants, are rated UL 94 HB.

3.1.3

IEC 60065

In Europe, the legislation covering the consumer electronics industry is IEC 60065 [8]. The 6th edition of this standard is due in August 2002. IEC 60065 is more than just a flammability standard, it also covers requirements for items such as mechanical properties. There are different flammability requirements depending on the voltage of any internal potential ignition source and the distance of the plastic part from the voltage source, but the IEC 60065 flammability requirements are roughly similar to those UL of 94 HB.

3.2

RESEARCH METHODS

Many of the regulatory tests date back to the 1970s, and in some cases it is questionable whether the tests accurately predict performance in actual fire situations. These tests are often a pass/fail type, so it is difficult to obtain quantitative data. There are a few methods that do give quantitative information and these are often used in research applications. Attempts have been made to correlate these methods with the regulatory tests, with little success [9]. The limiting oxygen index (LOI) measures the minimum oxygen content that is required to sustain a flame [10]. The specimen in this test is in a vertical orientation, but it is the top of the specimen that is ignited. Because the specimen burns in a candle-like fashion, LOI may not be representative of fire situations. However, LOI does give quantitative information and may be useful for ranking the relative flammability of samples. The cone calorimeter was developed in the early 1980s by NIST [11]. This method uses 10 by 10cm specimens that may be up to 5cm thick. A coneshaped heater applies a heat flux of up to 100 kW/m2 to the top of the sample. Parameters that can be measured include peak and total heat release rate, mass loss and smoke generation. The data obtained from cone calorimetry can be used for engineering purposes. A micro-scale combustion calorimetric method test has been developed by Walter and Lyon, which involves pyrolysis and combustion calorimetry of the volatile products [12]. Using this technique, the heat release capacity can be obtained. The heat release capacity is a material parameter and has been used to correlate polymer structures with flammability [13].

690

4

B. KING

MECHANISMS OF FLAME RETARDATION

As mentioned previously, when polystyrene is subjected to the temperatures of a flame it pyrolyzes by a depolymerization mechanism to give monomer and oligomers [14]. The combustion of these volatile products in the vapor phase above the sample supplies heat back to the solid sample (Figure 29.3). If the energy supplied by combustion is sufficient to maintain the pyrolysis process, the flame is self-sustaining even after the test flame has been removed. In order to make polystyrene more flame retardant, the cycle of pyrolysis and combustion must be broken. Flame retardants may act in either the vapor or solid (condensed) phase.

4.1

VAPOR-PHASE MECHANISMS

Highly reactive hydrogen and hydroxyl radicals are among the radicals that are present in the vapor phase during combustion. These reactive radicals are essential to the propagation of the combustion process [15]. The most effective way to flame retard styrenic polymers is to add a flame retardant that traps or reacts with these radicals, thus breaking the burning cycle by reducing the amount of heat from combustion that is feeding the pyrolysis process [16]. The vapor-phase flame retardants of choice are organohalogen compounds, and the decomposition of these compounds during pyrolysis releases hydrogen halide into the vapor phase. Some halogen-based flame retardants can release HX directly from decomposition, but others require reaction of halogen radicals with C—H bonds of the styrenic polymer. Figure 29.4 illustrates hydogen halide reactions with hydrogen and hydroxyl radicals [16]. In this figure, X represents the halogen atom and P represents the polymer. Halogen halide can be regenerated by reaction of halogen radicals with hydrocarbon species present in the vapor phase, thus hydrogen halide acts catalytically to quench combustion reactions.

Figure 29.3

Schematic diagram of polymer combustion

FLAME-RETARDANT POLYSTYRENE

691

R—X —»~ X* + R* X* + P-H —»-HX + P* H* + HX —*-H 2 + X* HO* + HX —»- H2O+ X' Figure 29.4

Vapor-phase chemistry of halogen flame retardants

Antimony trioxide acts as a synergist with halogen-based flame retardants. When antimony trioxide is present at a level of 3–4 wt%, about 10 wt% of halogen atoms is required to achieve a UL 94 V-0 rating for styrenic polymers. Much higher levels of halogen would be needed in the absence of antimony trioxide. It is thought that antimony trioxide reacts with hydrogen halide to generate volatile antimony trihalide [17]. The antimony trihalide may react with hydrogen radicals to generate hydrogen halide and eventually antimony oxide or hydroxide (Figure 29.5). The antimony oxide or hydroxide formed in the vapor phase may also act to quench radicals. Another vapor-phase mechanism is to use a flame retardant that generates inert volatile products, thus diluting the concentration of flammable species in the flame [16]. The hydrogen halide gas that is released by halogenated flame retardants may have some action by this mechanism in addition to quenching radicals. Another example is metal hydroxides such as aluminum trihydrate (ATH) and magnesium hydroxide, which release water into the vapor phase. The loss of water is an endothermic reaction, so there is some cooling of the solid polymer that is also associated with the metal hydroxides. The limitation of metal hydroxides for styrenic systems is that levels of up to 60 wt% are required to obtain a UL 94 V-0 rating, and as a result the mechanical properties are usually poor. SbX3 + H' —»» SbX2 + H SbX2 + H' —^ SbX + H* —*-

SbX + HX Sb + HX

Sb + O* —*~ SbO

Sb + HO* —*- SbOH SbO + HO' —*~ SbOH SbOH + H' —»- SbO + H2 SbOH + HO* —»~ SbO + H2O Figure 29.5

Vapor-phase chemistry of antimony trihalide

692

4.2

B. KING

CONDENSED-PHASE

MECHANISMS

There are a number of flame-retarding mechanisms that operate in the solid phase of polymers. One is to use additives that absorb some of the heat of combustion by endothermic reactions; this was mentioned in the previous section in connection with metal hydroxides. Formation of a protective char layer is another important condensed-phase mechanism. Unfortunately, polystyrene does not form any appreciable levels of char during burning, even in the presence of charring catalysts. Some progress has been made in enhancing char formation of polystyrene by the use of Friedel—Crafts chemistry [18]. Intumescence is the formation of a foamed char, which is a particularly good heat insulator. Intumescent packages generally contain a source of carbon to build up char (carbonific), a compound which generates an acid upon heating in the flame, and a compound that decomposes to generate blowing gases to generate the foamed char [19]. The acid is required to cause charring of the carbonific component. A limitation of this approach is that relatively high levels (30 wt% or more) of the intumescent package are required to flame retard styrenic polymers. Another mechanism that can be used in the condensed phase is to use an additive that enhances decomposition, and thus dripping, of the polymer during burning. This may be desirable for achieving a UL 94 V-2 rating, but may not be an option for a V-0 requirement. Promoting nonflaming drips is one way to make nylon polymers with a V-0 rating, but this approach is not used for styrenic polymers.

5 HALOGEN-BASED FLAME RETARDANTS FOR STYRENICS As mentioned previously, halogen-based flame retardants are the most widely used for styrenic polymers. Since halogen-based flame retardants act primarily in the vapor phase, the halogen-containing compounds need to decompose and evolve HX in the same temperature range in which polystyrene pyrolyzes (> 300 °C). Another consideration is that the flame retardant needs to be sufficiently thermally stable to be melt compounded with polystyrene. The order of effectiveness of halogens is I > Br > Cl > F [20]. Iodinecontaining organic compounds are too thermally unstable to be melt compounded with polystyrene, and organoflorine comounds are too thermally stable to be effective as flame retardants. This leaves bromine- and chlorinecontaining compounds as the most effective flame retardants. Aliphatic chlorine compounds find some utility as flame retardants for styrenic polymers, but aromatic chlorine compounds are probably too stable to be effective [21]. Aliphatic bromine compounds are too thermally unstable for com-

FLAME-RETARDANT POLYSTYRENE

693

pounding at typical polystyrene processing temperatures [22]. Aromatic bromine compounds are the most widely used flame retardants for styrenic polymers. The most common halogen-based flame retardants used in styrenic polymers are listed in Table 29.1 [23]. The majority of these are brominated aromatic compounds used to flame retard HIPS and ABS. As mentioned in Section 4, roughly 10 wt% of bromine is required to pass UL 94 V-0 requirements. Antimony trioxide is also used in combination with these brominated compounds. Because expanded polystyrene foam is processed at a lower temperature, aliphatic bromine compounds such as hexabromocyclododerane (HBCD) can be used for this application. The flame retardant levels in these systems are family low, typically less than 3 wt%. These levels are sufficient to pass the Steiner Tunnel test, and synergists such as antimony trioxide are not necessary. Other considerations when choosing a flame retardant are cost and the effect of the flame retardant on physical properties. The solubility of the flame retardant in the styrenic polymer roughly parallels the melting temperature of the flame retardant [24,25]. The lower melting flame retardants tend to be more soluble and can have positive effects on lowering the melt viscosity of the flame-retardant styrenic polymer. Soluble flame retardants may cause blooming, where the flame retardant migrates and deposits on the surface of molded parts. Solubility of the flame retardant can also have negative effects on heat resistance of the styrenic polymer. Insoluble flame retardants have a more negative effect on the toughness of ABS, and in this case the soluble flame retardants are preferred. Comments on each of the flame retardants are as follows: • Decabromodiphenyl oxide (DBDPO) is the most widely used flame retardant for HIPS, its low cost and high bromine content making it a popular choice. DBDPO has some solubility in polystyrene, but also acts as an insoluble filler. DBDPO will discolor on exposure to UV light, so it is usually used in painted or dark-colored parts. • Octabromodiphenyl oxide (Octabrom) is a soluble flame retardant. It is most often used in ABS. • Saytex 8010 is an option when there are concerns with DBDPO (see Section 7). It acts as an insoluble filler, and is most often used in HIPS. The UV stability of Saytex 8010 is better than that of DBDPO. • Brominated indans are also a non-DBDPO option. They are soluble flame retardants, but do not lower the heat resistance of the styrenic polymer because the glass transition temperature of, e.g., FR-1808 is close to that of polystyrene [26,27]. • Ethylenebis(tetrabromophthalimide) is an insoluble flame retardant and is used for its very good UV stability. • Tetrabromobisphenol A is the most widely used flame retardant for ABS. It is a soluble flame retardant. Discoloration due to thermal instability can be a problem if processing temperatures greater than 240 °C are used.

Table 29.1

o>

Halogen-based flame retardants used in styrenic polymers.

Chemical name

Structure

Trade name

Decabromodiphenyl oxide (DBDPO)

Octabromodiphenyl oxide (Octabrom)

Brv

Brv

Melting range (°C)

FR-1210 (DSBG)" DE-83 (GLCC)a Saytex 102 (Albemarle)

305

FR-1208 (DSBG) DE-79 (GLCC)

70-150

Saytex 8010

380

x +y=

Decabromodiphenylethane Br.

. Br

CH 2 CH 2

Br

. Br

Brominated indan

FR-1808 (DSBG)

240-255

BT-93 (Albemarle)

450-455

FR-1524 (DSBG) BA-59P (GLCC) RB-100 (Albemarle)

181

m +n=

Ethylenebis (tetrabromophthalimide)

Br

O

O

Br

O

Br

N—CH 2 CH 2 —N Br

Tetrabromobisphenol A (TBBA)

O

Br

Br CH3 Br

CH3 O> to (continues) *"

Table 29.1

(continued)

Chemical name

Structure

Trade name

Melting range (°C)

Brominated epoxy oligomers (BEO)

F-2016 (DSBG)

105-115

Bis(tribromophenoxy) ethane

FF-680 (GLCC)

223-228

OCH2CH2O

Tris(tribromophenyl)cyanurate

FR-245 (DSBG)

230

FR-1206 (DSBG) CD-75P (GLCC) HBCD (Albemarle)

175-197

Br

Br'

Br

T

Br

Br

Br ,O

Br

Br

Hexabromocyclododecane Br Br

Br Br

(continues) o> CO

Table 29.1

(continued)

Chemical name

Structure

Trade name

Melting range (°C)

Tetrabromobisphenol A bis(allyl ether)

BE-51 (GLCC)

115–120

Chlorinated alkane

Chlorez 760

160

" DSBG, Dead Sea Bromine Group; GLCC, Great Lakes Chemical Corporation.

FLAME-RETARDANT POLYSTYRENE

699

• Brominated epoxy oligomers are available in a number of different forms, with variations in molecular weight and endcapping. In general, these are nonblooming soluble flame retardants with good UV stability. • Bis(tribromophenoxy)ethane is a soluble flame retardant used mostly for ABS; blooming can be an issue. • Tris(tribromophenyl) cyanurate is a relatively new soluble flame retardant. It reportedly has good UV stability and does not bloom. • Some grades of chlorinated alkanes have sufficient thermal stability for use with styrenic polymers, but processing temperatures should be limited to 220 °C. • HBCD is the most widely used flame retardant for expanded polystyrene foam. • Tetrabromobisphenol A bis(allyl ether) finds some utility in expanded polystyrene foam. Other additives such as stabilizers, impact modifiers, mold release agents, colorants and anti-drip agents are often added to flame-retardant styrenic polymers.

6

STYRENIC BLENDS

Styrenic blends find application in more demanding applications. These blends are of considerably higher cost than HIPS or ABS, but offer greater toughness and the option of using a nonhalogen flame retardant (see Section 7). Polystyrene and PPO are miscible polymers, and HIPS is typically used in these blends for greater toughness. HIPS—PPO blends can be flame retarded with brominated flame retardants, but perhaps of more interest are blends flame retarded with aromatic phosphates. Phosphates are thought to act through promotion of charring of PPO, but may also have some vapor-phase activity. Typical levels of PPO used to achieve a UL 94 V-0 rating are 30-50 %; these blends also contain 10–20% of an aromatic phosphate [28]. Polycarbonate and ABS are not miscible, but blends are compatible and have excellent toughness. PC—ABS blends are also flame retarded with aromatic phosphates. These blends are typically very high in polycarbonate (70-80%) and also contain at least 10% of an organic phosphate. PC—ABS blends are preferred for unpainted applications because of their excellent UV stability.

7

ENVIRONMENTAL CONCERNS

There are numerous issues that influence the current and future state of flameretardant plastics. This is particularly evident in Europe, where regulatory and

700

B. KING

legislative requirements and also pressure from consumer and environmental groups play a part in shaping the flammability (fire safety) and environmental requirements of plastics. The current flammability standards in Europe are roughly equivalent to UL 94 HB, but recent studies demonstrate the benefits of using plastics with greater fire safety [29, 30]. Over the last decade, concerns have been raised regarding the potential exposure of workers, consumers and the environment to brominated dioxins and furans resulting from the use of brominated flame retardants. Polybrominated diphenyl ethers were the original target of these concerns. One study indicated that only the pentabrominated version is of concern with respect to eco-toxicity [31], but considerable pressure from environmental groups still remains for the octa- and decabromodiphenyl ethers. Of broader concern is a move in Europe to eliminate the use of all halogenated flame retardants. The technology exists for safe incineration of plastics that contain halogenated flame retardants, but in many cases it appears that emotional issues and environmental groups have more influence on decisions than have scientific studies [32]. In North America, DBDPO is still the dominant flame retardant used for HIPS, whereas in Europe alternative flame retardants such as Saytex 8010 and FR-1808 are preferred. Mixtures of these flame retardants are used in the exporting countries of the East, depending on the global region in which the flame-retardant plastic will be used. Styrenic polymers are very difficult to flame retard with nonhalogen flame retardants. Many nonhalogen flame retardants act in the condensed phase, but these mechanisms are not effective for polystyrene. Flame retardants such as magnesium hydroxide or alumina trihydrate (ATH) require very high loadings, resulting in very poor mechanical properties. Blends of styrenic polymers with char-forming polymers, such as PPO or polycarbonate, are potential nonhalogen candidates. Indeed, it appears that a combination of flame-retardant mechanisms may be required for nonhalogen flame-retardant polystyrene [33]. As one can see, the future of flame retardant use in plastics is currently in a state of flux. It will be interesting to see what this industry looks like in 2010.

8

SUMMARY

Flame-retardant styrenic polymers find utility in applications such as building insulation (expanded polystyrene foam) and electronic enclosures (flameretardant HIPS, ABS and styrenic blends). The most effective flame retardants are halogen-(particularly bromine)-containing compounds; these flame retardants act by inhibiting the radical combustion reactions occurring in the vapor phase. Flame-retardant plastics are in a state of flux, due to influences of

FLAME-RETARDANT POLYSTYRENE

701

regulatory and environmental factors. There are movements to discontinue the use of halogen-containing flame retardants, but there are currently few costeffective nonhalogen alternatives for styrenic polymers.

REFERENCES 1. Hirschler, M. M., in Fire Retardancy of Polymeric Materials, Grand, A. F. and Wilkie, C. W. eds, Marcel Dekker, New York, p. 65 (2000). 2. Lutz, J. T., Jr, Thermoplastic Polymer Additives: Theory and Practice, Marcel Dekker, New York, p. 93 (1989). 3. Fire Retardant Plastics III, Skeist Inc., Whippany, NJ (1997). 4. ASTM E 84. Standard Test Method for Surface Burning Characteristics of Building Materials. American Society for Testing and Materials, Philadelphia. 5. Castino, T. G., Beyreis, J. R., Metes, W. S., Flammability Studies of Cellular Plastics and Other Building Materials Used for Interior Finishes, Subject 723, Underwriters Laboratory, Northbrook, IL (1975). 6. Babrauskas, V., White, J. A., Jr, Urbas, J., Build. Stand. 66(2), 13 (1997). 7. UL 94. Tests for Flammability of Plastic Materials for Parts in Devices and Appliances, 5th edn, Underwriters Laboratories, Northbrook, IL (1998). 8. IEC 60065, Audio, Video and Similar Electronic Apparatus—Safety Requirements, 5th edn (1996), 6th edn (2002), IECEE, Geneva. 9. Nalepa, C. J., in 20th International Conference on Fire Safety, Product Safety Corp., Millbrae, CA, pp. 229–240 (1995). 10. Hilado, C. J., Flammability Test Methods Handbook, Technomic Publishing, Westport, CT (1973). 11. Babrauskas, V., in DiNenno, P. J., ed., The SFPE Handbook of Fire Protecion Engineering, 2nd edn, National Fire Protection Association, Quincy, MA, pp. 3–37–3–52 (1995). 12. Walters, R. N., Lyon, R. E., in Proceedings of 42nd International SAMPE Symposium and Exhibition, Haulik, T. E., Bailey, V. E., Burton, R., eds, Society for the Advancement of Material and Process Engineering, Covina, CA, pp. 1335–1344 (1997). 13. Walters, R. N., Lyon, R. E., Proc. Am. Chem. Soc. Div. Polym. Mater. Sci. Eng., 83, 86 (2000). 14. Madorsky, S. L., Thermal Degradation of Organic Polymers, Interscience, New York (1974). 15. Georlette, P., Simons, J., Costa, L., in Fire Retardancy of Polymeric Materials, Grand, A. F., Wilkie, C. W., eds, Marcel Dekker, New York, p. 252 (2000). 16. Troitzsch, J., Plastics Flammability Handbook, Macmillan, New York (1983). 17. Georlette, P., Simons, J., Costa, L., in Fire Retardency of Polymeric Materials, Grand, A. F., Wilkie, C. W., eds, Marcel Dekker, New York, p. 253 (2000). 18. Li, J., Wilkie, C., Polym. Degrad. Stab., 57, 293 (1997). 19. Camino, G., Delobel, R., in Fire Retardancy of Polymeric Materials, Grand, A. F., Wilkie, C. W., eds, Marcel Dekker, New York, pp. 218–219 (2000). 20. Lutz, J. T., Jr, Thermoplastic Polymer Additives: Theory and Practice, Marcel Dekker, New York, p. 95 (2000). 21. Einhorn, I. N., J. Macromol. Sci. Rev. Polym. Technol., 1, 113 (1971).

702

B. KING

22. Georlette, P., Simons, J., Costa, L., in Fire Retardency of Polymeric Materials, Grand, A. F., Wilkie, C. \V., eds, Marcel Dekker, New York, p. 247 (2000). 23. Georlette, P., Simons, J., Costa, L., in Fire Retardency of Polymeric Materials, Grand, A. F., Wilkie, C. W., eds, Marcel Dekker, New York, pp. 257–260 (2000). 24. Sprenkle, W. E., Southern, J. H., J. Appl. Polym. Sci., 26, 2229 (1981). 25. Pettigrew, F. A., Landry, S. D., Reed, J. S., Proc. Flame Retard., 156 (1992). 26. Smith, R., Georlette, P., Finberg, I., Reznick, G., Polym. Degrad. Stab., 54, 167 (1996). 27. Finberg, L, Utveski, L., Kallos, M., Styrene Plastics, IMEC 8, p. 19 (1997). 28. Cooper, G. D., US Patent 3883613 (1975). 29. Corbey, D., The Fire Safety of European Televisions (2000). 30. Simonson, M., Proc. Am. Chem. Soc. Divi. Polym. Mater. Sci. Eng., 83, 90 (2000). 31. Response to Friends of Earth UK Statement, EBFRIP Website, http://www. ebfrip.org/foe09052000.html, May 9 (2000). 32. Weil, E., presented at 10th BCC Conference on Flame Retardancy, Stamford, CT (1999). 33. Weil, E., Polym. Degrad. Stab., 54, 125 (1996).

30

Photochemical Degradation of Styrenic Polymers B. MAILHOT, A. RIVATON AND J. L. GARDETTE Universite Blaise Pascal (Clermont-Ferrand), Aubiere, France

1

INTRODUCTION

Under irradiation with polychromatic light at A > 300 nm and 60 °C, representative of outdoor exposures, polystyrene (PS) homopolymer, copolymers and blends do not directly absorb the incident radiation. It is well known that the photooxidation of these polymers results from light absorption by chromophoric impurities [1,2]. Photooxidation generates modifications of the chemical structure of the material, which results in the formation of oxidized groups, the development of discoloration and the loss of the initial mechanical properties. After the examination of the PS photooxidation mechanism, a comparison of the photochemical behavior of PS with that of some of its copolymers and blends is reported in this chapter. The copolymers studied include styrene-stat-acrylonitrile (SAN) and acrylonitrile—butadiene—styrene (ABS). The blends studied are AES (acrylonitrile—EPDM—styrene) (EPDM = ethylene-propylene-dienemonomer) and a blend of poly(vinyl methyl ether) (PVME) and PS (PVME— PS). The components of the copolymers are chemically bonded. In the case of the blends, PS and one or more polymers are mixed. The copolymers or the blends can be homogeneous (miscible components) or phase separated. The potential interactions occurring during the photodegradation of the various components may be different if they are chemically bonded or not, homogeneously dispersed or spatially separated. Another important aspect is the nature, the proportions and the behavior towards the photooxidation of the components added to PS. How will a component which is less or more photodegradable than PS influence the degradation of the copolymer or the blend ? We show in this chapter how the Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy fr; 2003 John Wiley & Sons Ltd

B. MAILHOT ETAL

704

addition of polymeric constituents to PS modifies the mechanism and the kinetics of photodegradation in the cases of the homogeneous copolymer SAN (PAN is more resistant toward photodegradation than PS [3]), of the heterogeneous copolymer ABS, of the homogeneous blend PVME—PSand of the heterogeneous blend AES (the PVME, the butadiene and the EPDM components are more photosensitive than PS [4–6]).

2 2.1

PHOTOOXIDATION OF THE HOMOPOLYMER POLYSTYRENE (PS) UNDER IRRADIATION AT A > 300 nm EXPERIMENTAL RESULTS

The infrared (IR) spectra recorded throughout irradiation show an increase in absorbance due to the formation of oxidized products. Because PS presents initial absorption bands in the carbonyl vibration region (1900–1500 cm -1 ), a subtraction of spectra has to be carried out in order to observe the shape of the carbonyl envelope due to the formation of the photoproducts (Figure 30.1). Several maxima or shoulders are observed in the carbonyl region at 1515, 1690, 1698, 1732 and 1785 cm -1 . Another band with a maximum around 1605 cm -1 isalsoobserved,even if the initial absorption of PS at 1603cm-1 interfered in the subtraction. In the hydroxyl region, bands and shoulders at 3250,3450 and 3540cm-1 are observed. 0.5T 0.4

1900

1800

1600

1700

1500

-

Wavenumbers (cm ')

Figure 30.1 FTIR spectra of a PS film (89 ^m) photooxidized at A > 300 nm; initial spectrum is subtracted. Irradiation times: A, 25; B, 90; C, 150; D, 182; E, 217; F, 257; G, 294 h

705

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

Several chemical and physical treatments have been used to identify the oxidation photoproducts that generate these IR absorptions [7,8]. The purpose of chemical derivatization reactions (with reactive gases such as NH3 or SF4 or methanol) or of a physical treatment (photolysis or thermolysis in the absence of oxygen) carried out on pre-photooxidized samples is to modify the IR spectrum of the oxidized samples by producing a decrease in some specific absorption bands and eventually creating new ones. This technique is very useful for locating the maxima of convoluted IR bands. Several examples of such treatments are given in this chapter.

2.1.1

SF4 Treatment of a Pre-photooxidized PS Film at A

300 nm

After reaction with Sp4, all OH groups are converted to F groups [9]. The reactions of carboxylic acids with SF4 give acyl fluorides, and alcohols and hydroperoxides give alkyl fluorides. SF4 treatment carried out on a pre-photooxidized film leads to a decrease in absorption in the carbonyl vibration region (1698, 1710, 1732 and 1753 cm -1 ) and to the formation of two new bands at 1813 and 1841 cm -1 (Figure 30.2). In the hydroxyl domain, the absorption is almost completely lost. These data were compared with those obtained for model compounds introduced by permeation in PS films and then submitted to SF4 for 1 h. The results reported in Table 30.1 show the position of the maxima of the carboxylic acids and their

0.04-

0.00"

-0.04 -

-0.08 1900

1800

1700 Wavenumbers (cm'1)

1600

1500

Figure 30.2 Infrared spectra of a PS film (16|xm) irradiated at 1 > 300 nm, 60 °C, 123h, treated with SF4 for 22 h. Subtracted spectrum between the irradiated and the treated film

706

B. MAILHOT ETAL

Table 30.1

Acids identified by SF4 treatment

Compound Benzoic acid Aliphatic acid

Before SF4 treatment (cm-1)

After SF4 treatment (cm -1 )

1698-1732 1710–1753

1813 1841

derivatives. Comparison of the changes in IR spectra after SF4 treatment of a photooxidized film with the model compounds led to the identification of aromatic and aliphatic acids in PS photooxidized films.

2.1.2

NH3 Treatment of a Pre-photooxidized PS Film at A > 300 nm

reacts mainly with acid and ester groups formed in photooxidized films, leading to carboxylate ions and amide groups, respectively. The reaction of a pre-photooxidized film with NH3 led to a decrease of the carbonyl bands (1698, 1710, 1725, 1732, 1753 and 1785cm -1 ), and to an increase of broad bands with maxima at 1553, 1585 and 1668 cm-1. The maxima at 1553 and 1585cm-1 agree with the C=O stretching vibrations of carboxylate ions. These bands correspond to the reaction products of benzoic and aliphatic acid-type structures, respectively, with NH3 (1698–1732, 1710–1753 c m - l ) These assignments were made after comparison with model compounds introduced in a PS film and treated with NH3. The maximum at 1668 cm - 1 matches the C=O stretching vibrations of amide groups. Benzoic anhydride (1725– 1785cm-1) reacts with NH3 to produce an amide and a carboxylate ion. At 1732cm -1 , in addition to the dimer form of benzoic acid, the formation of an ester and of a 8-lactone is suspected. They both react with NH3 to produce amide groups.

2.1.3

Determination of the Photoproduct Profiles

The distribution of the photoproducts in the thickness of the film can be determined by IR micro-spectrometric analysis. Irradiated films were embedded in an epoxy resin and thin slices of thickness ca l00 jim were analyzed. The variations of the absorbance at 1725cm-1 versus the film thickness are plotted Figure 30.3. The middle of the film is almost as photooxidized as the front and rear sides. The film has a relatively high permeability to oxygen and the photoproducts are fairly homogeneously dispersed in the thickness of the film.

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

707

AOD1725 0.18--

0.12--

0.06--

60

120 Film thickness (nm)

180

240

Figure 30.3 Photoproduct concentration profiles of a PS film photooxidized at A > 300 nm for 292 h

2.2

DISCUSSION

The photoproducts formed in the PS films during irradiation at / > 300 nm were identified using various chemical and physical treatments. In addition to the reactions with SF4 and NH3 treatments with methanol, thermolysis and photolysis experiments were carried out on photooxidized samples [7]. In the light of all those tests, it was concluded that numerous photoproducts formed are low molecular weight species. Those products were volatile and may diffuse out of the film. The photodegradation of PS is induced by chromophoric impurities. Once a radical has been formed, it produces macroradicals by abstraction of a hydrogen atom from the polymer structure. The mechanism of PS photooxydation is presented in Scheme 30.1. The abstraction of a hydrogen atom occurs preferentially at the tertiary carbon of the structure, leading to a polystyryl radical. This radical adds to oxygen to form a peroxy radical. By abstraction of another hydrogen atom, the peroxy radical leads to a hydroperoxide. Hydroperoxides have an IR absorption at 3450 cm - 1 . The decomposition of the hydroperoxide either by photolysis or by thermolysis gives an alkoxy radical that may react in several ways: • By abstraction of a hydrogen atom, an alcohol is formed. Hydroxyl groups are detected by their IR absorption at 3540 cm -1 (free OH) and 3450 cm-1 (bonded OH).

708

B. MAILHOT ETAL

H 1 "CH->— C 1 Ph

OOH 1 *- — CH,— C — PS 1 Ph

hv,O2

VMW>CH

•• OOH 1

OOH 1 2

C

1 Ph

£ Hy— C 2

\

Ph

3450 cm-'

h v,A OH

O

0-

ps

— CH2— C — Ph

/

3540 cm-'/3450 cm-1

/

Ph \ p-scission\

1515cm /160 cm O

O II — CH2— C— Ph

— CH2— C — CH2—

+ — CH2' 1

\

hv 1somerization

O II •C— Ph

+ Ph'

1725 cm-'

1690 cmhv

+ — CH/

-.,

\

,

,..- CH2' + *•' -CH3

f^~^\ «C — CH2 —

/ V i /

0

O

CH2— O

II

\

II

0 0

II

H-C-Ph

Ph-C-0-C-Ph

1704 cm~'

1^25 cm-'/ 1785 cm-

'V r>* U

C — CTH3 *~ "

1- 1 O II HO-C— Ph 698 cm-'/ 1732 cm-1

o /V~N\ II (( )V- C— CH3

^

Ph

vrv

1 ,_ 0

/^\

ril ]l I-H (( Y— ))— S/ ~"CH2 — C— CH3 1 / \1725cm-

0 II

O

-OH

1710cm-'/ 1753 cm-1

11690 cm-1

hv

CH3— C-OH 1710cm-'

Scheme 30.1 Polystyrene photooxidation mechanism

O T1 ii C

*

" CH ^

*

O

H-C-OH 1710cm-'

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

709

• The formation of dibenzoylmethane (1515 and 1603cm -1 ) confirms the idea that the oxidation of many polymers involves a zip oxidation of neighboring carbon atoms as suggested for polypropylene [10]. • By ß-scission, the alkoxy radical may give a chain ketone (detected at 1725cm -1 ) and an end-chain aromatic ketone (1690cm -1 ). The ß-scissions are accompanied by the formation of a benzyl and an alkyl radical. The former is the precursor of benzene and the latter isomerizes to a tertiary radical. This last radical is the precursor, after several reactions, of acetophenone (1690 cm - 1 ), end-chain aliphatic ketone (1725 cm -1 ), end-chain aliphatic acid (1710–1753 cm - 1 ) and acetic and formic acids (1710cm -1 ). The end-chain aromatic ketone (1690cm -1 ) may react photochemically by a Norrish type I reaction, leading to the formation of benzaldehyde (1704 cm - 1 ) and benzoic acid (1698–1732cm - 1 ). No detailed mechanism can be proposed for the formation of benzoic anhydride (1725–1785 cm - 1 ) since several plausible routes exist.

3

PHOTOOXIDATION OF POLY(STYRENE-COACRYLONITRILE) (SAN)

SAN is constituted of styrene and acrylonitrile units copolymerized statistically in the ratio 80:20 mol%. Previous studies on the photooxidation of PS [7,8] and PAN [3] have shown that the photooxidation rates of these polymers were very different: PS degrades about 20 times faster than PAN. Consequently, the first steps of photooxidation of the copolymer SAN is presumed to involve mainly the styrene units. SAN samples have been irradiated and analyzed under the same conditions as PS samples.

3.1

EXPERIMENTAL RESULTS

Fourier transform (FT) IR analysis of the photooxidized SAN samples shows that the oxidation products formed in the copolymer may result not only from the oxidation of the styrene units, even in the first few hours of irradiation [11]. Figure 30.4 shows that the absorbance of the carbonylated photoproducts in the photooxidized SAN samples is different compared with PS (Figure 30.1). Substantial evidence for the contribution of the acrylonitrile units in the photooxidation was obtained by chemical and physical treatments carried out on prephotooxidized samples as described above. For example, the SF4 treatment of a SAN photooxidized sample led to a partial decrease in absorbance in the hydroxyl region, corresponding to the disappearance of alcohols, hydroperoxides and acids. The absorbance remaining after treatment may be assigned to

710

B. MAILHOT ETAL 0.62 0.51 0.40 0.30 • 0.19

0.080.02 1900

1800

1700

1600

1500

Wavenumbers (cm-') Figure 30.4 FTIR spectra of a SAN film (lOO^m), photooxidized at A > 300 nm; initial spectrum is subtracted. Irradiation times: A, 25; B, 65; C, 110; D, 139; E, 170h, F, 202h

the N—H stretching vibrations of amides and imides, the formation of which has been reported during PAN photooxidation [3]. Moreover, as previously observed with PS films, no significant profile was observed and the photoproducts were homogeneously dispersed within the film.

3.2

DISCUSSION

Photooxidation yields a polystyryl radical as identified formerly in the study of PS photooxidation (Scheme 30.2). Two different cases may occur. If this radical is formed in a succession of styrene units (1), it reacts in the same way as in PS. If it is formed on a styrene unit linked to an acrylonitrile unit (2), three reaction pathways may be envisaged. The alkoxy radical resulting from the decomposition of the hydroperoxide formed on this polystyryl radical may react by 3-scission. Scissions (a) and (b) yield chain ketones, acetophenone end-groups and phenyl and alkyl radicals as previously observed in the case of PS photooxidation mechanism. Scission (c) leads to the formation of an aromatic ketone and an alkyl radical. This alkyl radical may be the precursor of acrylonitrile units (identified by IR spectroscopy at 2220 cm -1 ), or may react directly with oxygen and after several reactions generates acid groups, or finally this radical may isomerize to a more

711

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

-CH 2 -C-CH 2 -CHI

*"

Ph

i

Oi OOH I -CH,-C-CH2-CH-

SAN

Ph

CN

hv

-CH 2 -C-CH 2 -CH I I Ph Ph

(a) and (b)

p-scission

ACIDS PAN

— CH 2 -CH 1 C '/ -•- ' \\ O NH o

PAN 0 H-C

hv PAN

Scheme 30.2 Photooxidation mechanism of SAN

\\ O

NH

2

712

B. MAILHOT ETAL.

stable tertiary radical. This tertiary radical is oxidized following a classical oxidation mechanism, yielding hydroperoxides, alkoxy radicals, ketones and ketonitrile groups, and finally acids. SF4 treatments have shown that the proportion of aliphatic acids was higher in SAN than in PS. This observation confirmed once again this oxidation pathway of the acrylonitrile units. The study of PAN photooxidation have shown that the formation of acids was the major factor controlling the degradation. In SAN samples, acid groups were easily formed. Some of the acids may react with the nitrile groups and produce imides [3]. The formation of imides was evidenced in SAN after photolysis at A > 300 nm of samples pre-photooxidized under the same conditions. Photolysis or hydrolysis of the imides yielded amides and aldehydes. Aldehydes may be further oxidized into acids. In PS, the low molecular weight products, mainly of acidic nature, formed during the degradation, may escape fairly easily from the film. In the SAN copolymer, the acids react with the nitrile groups and the loss by diffusion out of the polymer is lowered. Moreover, the migration may be reduced as a result of the lower permeability of SAN compared with PS. We conclude that the photooxidative degradation of SAN is initiated by the styrene units. The mechanism of SAN presents three ways of degradation: a 'classical' photooxidation pathway of PS, an oxidation of the acrylonitrile units initiated by the oxidation of the adjacent styrene units and a degradation of the acrylonitrile units due to the formation of acids among the photoproducts.

4

PHOTOOXIDATION OF ACRYLONITRILE-BUTADIENESTYRENE (ABS)

Acrylonitrile—butadiene—styrene (ABS) is a copolymer of polybutadiene-grSAN. The microstructure of the studied ABS was characterized using FTIR and 13C NMR spectra: 25 mol% acrylonitrile, 40 mol% styrene and 35 mol% butadiene (56% cis, 30% trans and 14% vinyl [5]). The photooxidation of ABS at A > 300 nm has been compared with the photooxidation of the homopolymers polystyrene (PS) and polybutadiene (BR). Oxidation products were identified using FTIR and UV—-visible spectroscopy as well as chemical derivatization reactions [5,12,13]. Particular attention was devoted to the decomposition of hydroperoxides formed in the first stage of exposure. HPLC analysis of the low molecular weight fragments extracted from ABS, BR and PS irradiated films has also been carried out. 4.1

ANALYSIS OF THE PHOTOOXIDATION

Irradiation of ABS under polychromatic light in the presence of atmospheric oxygen leads to noticeable evolutions of the UV and FTIR spectra of exposed films.

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

713

In the carbonyl region, at a weak extent of conversion, photooxidation of ABS and BR led to the formation of a thin absorption band with maxima at 1697cm -1 (a, ß-unsaturated acids) and 1683 cm - 1 (a, ß-unsaturated ketone), and to the formation of a broader absorption band with a maximum at 1721 cm - 1 . As photooxidation proceeded, the intensity of this latter band increased and shifted to 1717cm -1 whilst the band at 1697 cm - 1 became hard to observe. The intensity of the band at 1697 cm - 1 ceased to increase after 16 h of irradiation. When the exposure time was longer than 22 h, only one absorption band was observed. Its maximum shifted from 1717 to 1725cm -1 . In parallel, a shoulder was detected in the range 1775–1785 cm - 1 . It was shown that the absorption around 1725 cm-1 resulted from the convolution of various species: saturated carboxylic acid (1717 cm - 1 ), aliphatic ester (1735 cm - 1 ), a, ß-unsaturated anhydride (1724–1782 cm - 1 ), saturated aldehyde (1727–2720cm - 1 ) and saturated ketone (1725 cm - 1 ). The maximum at 1780 cm - 1 has been assigned to three types of structure: a, ß-unsaturated anhydride (1724 and 1782 cm - 1 ), perester (1789 cm - 1 ) and -y-lactone (1775 and 1175 cm - 1 ). The hydroxyl region revealed a broad absorption band with a maximum around 3430 cm - 1 assigned to the convolution of hydroperoxides, alcohols and acids. In the region of the C—O stretching vibrations, an important increase in absorbance was observed with maxima at 1060 cm - 1 (a, ß-unsaturated ether bridges and ot, (ß-unsaturated alcohols), 1094 cm - 1 (saturated secondary alcohols and saturated ethers) and 1175 cm - 1 (lactone). The development of a band with a maximum at 2731 cm - ', assigned to the C—H stretching vibration of a, ß-unsaturated aldehydes, also ceased after 16 h of irradiation. In the C—H deformation vibrations region, a significant decrease in the intensity of the initial 1,2-vinyl band at 912 cm - 1 was observed. These spectral evolutions, concerning both the formation of photoproducts and the decrease of BR unsaturations, were similar to those observed in BR [14–16]; the photooxidation of PS [7,8] has been shown to be completely different. The evolution of the UV spectra of ABS and BR films throughout irradiation resulted in a gradual increase in light absorption below 450 nm without any definite maximum. These absorbing products were shown to be the roots of a photothermal equilibrium between the cisltrans forms of a, 3-unsaturated ketones.

4.2

PHOTOOXIDATION RATE

The comparison of the photooxidation rates of 100 (Jim films of ABS and of the homopolymers BR and PS is reported in Figure 30.5.

B. MAILHOT ETAL

714

2.0-

T

1.5 -J

25

50

75

Irradiation time (hrs) Figure 30.5 Evolution of OD, measured at 1713crrr1, versus irradiation time of ABS experimental, ABS calculated, BR homopolymer and PS homopolymer. Films thickness: 100 um

Under these conditions of irradiation, the oxidation of the acrylonitrile component does not proceed, as shown by the invariance of the band (2237 cm -1 ) of acrylonitrile units. The calculated photooxidation rate of a 'model' ABS is also reported in Figure 30.5. In this reference system, no interaction occurs between the photooxidation of the elastomeric and styrenic phases. The photochemical evolution of the model ABS has been determined for various irradiation times: the increase in absorbance of the BR homopolmer has been added to that of the PS homopolymer and corrected by a multiplicative factor that takes into account the percentage of each component in the experimental ABS blends. The comparison of the photooxidation rates reported in Figure 30.5, along with analysis of FTIR and UV spectra, suggests that the BR component was implicated as the prime site for the photooxidation of ABS and that the photooxidation of the styrenic phase was enhanced by the presence of the polybutadiene in ABS, compared with PS homopolymer. Immersion in methanol of photooxidized ABS films led to a decrease in absorbance of photooxidation products after 24 h of immersion corresponding to low molecular weight photoproducts that have been extracted by the solvent. Extraction by methanol of BR photoproducts was observed to be fairly low. However, immersion in methanol of photooxidized PS films led to an important decrease in absorbance of the photooxidation products identified as benzoic

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

715

acid, acetophenone, benzaldehyde, methyl benzoate and styrene monomer. One of the peaks observed on the chromatogram of the methanol used to extract the oxidation products from ABS film photooxidized for l00 h was assigned to benzoic acid. No oxidation was detected in PS homopolymer photooxidized for 100 h; benzoic acid was detected only after 470 h of exposure of PS homopolymer. These experimental results confirm that the oxidation of the styrenic phase in ABS is enhanced by the presence of polybutadiene, compared with PS homopolymer. 4.3

DISCUSSION

In the first steps of irradiation, the BR moieties have to be considered as the prime reason for the photooxidation of ABS. Primary hydroperoxidation occurs in the a-position to the double bond of BR units leading to the formation of a, ß-unsaturated hydroperoxides [14–16]. The two mechanisms reported below are proposed to account for the photochemical and thermal homolysis of a, ß-unsaturated hydroperoxides: -CH=CH-CH-CH 2 — + OH' — CH=CH-CH-CH2- " I OOH

°* -CH=CH-C'-CH 2 — + rH I OOH

The alkoxy radicals so formed are the precursors of a, 3-unsaturated ketones, aldehydes, carboxylic acids and alcohols. Saturation reactions, Norrish type I reactions of ketonic compounds and oxidation of aldehydic species occur under irradiation. This finally leads to the formation of saturated carboxylic acids as the main oxidation products accompanied by esters, -y-lactones, peresters, anhydrides, alcohols and ether bridges. In secondary steps, radical species formed in BR photooxidation are able to initiate the oxidation of the neighbouring styrenic moieties: the photooxidation of the styrenic phase is enhanced by the presence of polybutadiene in ABS, compared with PS homopolymer. Butadiene grafting sites, containing tertiary allylic carbon atoms (A), are preferentially oxidized in tertiary hydroperoxides in the first stages of ABS photooxidation rather than in secondary allylic carbon atoms (B): (A) _ (B) -CH 2 -CH CH-CH-CH 2 CH 2 -CH-CH 2 -CH9

r.(r0.} ———-

OOH -CH,-CH-CH = CH-CH,'

CH2_CH-CH2-CH'

9 9

I 9

716

B. MAILHOT ETAL.

The thermal or photochemical homolysis of tertiary hydroperoxides leads to the formation of alkoxy macroradicals; ß-scission of alkoxy macroradicals may occur. This leads to a, ß-unsaturated ketones on the butadiene component and induces the scission of the butadiene—SAN grafts. The macroradical so formed on the SAN macrophase is the precursor, after isomerization, of the oxidation of the styrenic component according to OOH

O'

-CH 2 -C-CH=CH-CH 2 -^ CH 2 -CH-CH 2 -CHI I

-CH 2 -C-CH=CH-CH 2 cH2_CH_CH7_CH_ I ' I

O

CH 2 -C-CH=CH-CH 2 -

+

'CH,-CH-CH 2 -CHI I

PS photooxidation

^

_ _ CH3 - C - CH2- CH—

Photoproducts resulting from the oxidation of the styrenic units (such as benzoic acid) accumulate in ABS at long irradiation times.

5 PHOTOOXIDATION OF A BLEND OF SAN AND EPDM (AES) AES (acrylonitrile—EPDM—styrene) is a blend of SAN and EPDM. SAN is a statistic copolymer of styrene and acrylonitrile. EPDM is an elastomeric terpolymer of ethylene, propylene and a nonconjugated diene. The diene studied here was 5-ethylidene-2-norbornene. The total content of EPDM was 34 mole% [6]. The diene represented 8 mole% of the EPDM and the SAN phase was composed of 80 mole% of styrene and 20 mol% of acrylonitrile. The AES films were irradiated at /. > 300 nm at 60 °C in the presence of oxygen. The photoproducts resulting from the photooxidation of each components of AES were identified by FTIR spectroscopy coupled with the same chemical and physical treatments as mentioned above for the previous studies. As pointed out in the literature [17], the EPDM component is more reactive than the copolymer SAN towards photooxidation.

717

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

5.1

FTIR ANALYSIS OF AES FILMS DURING THE FIRST STAGES OF PHOTOOXIDATION

The IR spectra of an AES film recorded in the hydroxyl vibration region during the first 38 h of irradiation showed an increase in a broad absorption band centered around 3450 cm - l attributed to hydroperoxides. The development of a complex band with a maximum at 1713cm -1 and shoulders around 1690, 1730 and 1770 cm - 1 was observed in the carbonyl vibrations region (Figure 30.6). These maxima correspond to carbonylated photoproducts that have been previously identified during photooxidation of EPDM [17] and ethylene—propylene copolymer [18]. The bands at 1713, 1730 and 1770 cm -1 correspond, respectively, to the absorption of saturated acids (dimer form) and ketones, esters and lactones or peresters; the absorption around 1690 cm - 1 is related to the presence of unsaturated carbonyl species. A decrease in the initial band at 807 cm - 1 , attributed to the ethylidene group of the norbornene moieties, was measured. The total disappearance of this band at 807cm -1 was obtained after about 20 h of irradiation. During the same period, the absorbance at 3450 cm - 1 due to the formation of hydroperoxides increased rapidly. Taking into account an absorption coefficient £(3450 cm-1) of 75 kg/mol cm and an absorbance AOD (change in optical density) = 0.17, the concentration of ROOH was estimated to 0.23 mol/kg. This concentration of hydroperoxides formed during the first 20 h of irradiation was close to the concentration of the ethylidene groups. 0.54

0.45-0.35-0.26-0.17-0.08--0.01 1900

1800

1700

1600

1500

1

Wavenumbers (crrr ) Figure 30.6 FTIR spectra of an AES film (99 (xm) photooxidized at /I > 300 nm; initial spectrum is subtracted. Irradiation times: A, 1; B, 5; C, 9; D, 13; E, 17; F, 21; G, 25; H, 29; I, 33; J, 38 h

718 5.2

B. MAILHOT ETAL. FTIR ANALYSIS OF AES FILMS FOR LONGER IRRADIATION PERIODS

When irradiations of AES films were carried out until a high level of degradation of the film, the evolution of the IR spectra was completely different to those reported above. In the hydroxyl vibration region, the maximum at 3450 cm - 1 was progressively shifted to 3350 cm -1 and in the carbonyl vibration region the maximum at 1713 cm - 1 was shifted to 1725 cm -1 . This maximum at 1725 cm - 1 has been observed in the case of the photooxidation of SAN. In the 2290–2200 cm - 1 region a drastic decrease of the VC=N band (2237 cm - 1 ) of acrylonitrile units was measured whereas the formation of a new low band at 2220 cm -1 was noted (acrylonitrile monomer). The increase in absorbance of the band at 1515 cm -1 , characteristic of the degradation of the styrene units, was plotted as a function of irradiation time (Figure 30.7). It can be deduced from this figure that the degradation of the styrene units is faster in an AES film (containing 52 mol% of styrene units) than in a PS film.

5.3

DISCUSSION

The photoproducts identified from the various treatments carried out on photooxidized AES films are listed in Table 30.2. Each of these photoproducts has been formerly identified in the corresponding homopolymer or copolymer explored under the same conditions of photooxidation. The norbornene unit has been shown to be the primary location of 0.3

AES 6 0.2

0 < 0.1

0.0 i

PS

75

150

225

300

Irradiation time (hrs) Figure 30.7 Evolution of absorbance, measured at 1515cm"1, versus irradiation time of an AES film and of a PS film. Films thickness: 90 jjim

719

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

Table 30.2 Characteristic FTIR wavelengths of the principal photoproducts formed in AES samples under photooxidation Photoproducts

v(cm ')

Origin

Identification method

Hydroperoxides Alcohols Aliphatic ketones Aromatic ketones, a, ß-unsaturated ketones and a, ß-unsaturated acids Benzoic anhydride Amides Imides Aliphatic acids Aromatic acids

3450 3450 1725 1690

SAN, EPDM SAN, EPDM SAN, EPDM PS, EPDM

SF4 SF4 [1-3] [3,4]

CH3OH, NH 3 PS 1725, 1785 3450, 3360, 1680, 1606 PAN [5] NH3, SF4 3250, 1790, 1725, 1690 PAN EPDM, SAN NH3, SF4 1710/1750 1698/1732 PS NH3, SF4

the oxidation of AES samples. It has been proposed in the literature that the photosensitivity of norbornene could result from the presence of the tertiary hydrogen atom in an a-position to the double bond [17]. On irradiation in the presence of oxygen, a tertiary hydroperoxide would be formed by oxidation of this carbon atom: OOH -CH,

hv, O2

—CH,

The thermal or photochemical homolysis of the hydroperoxide leads to the formation of an alkoxy radical. The alkoxy radical is the precursor of unsaturated alcohols, acids and ketones. The decrease in intensity of the band at 807 cm - 1 indicated saturation of the double bond, which could result from a radical addition to the double bond (for example, by reaction with the hydroxyl radicals resulting from the decomposition of hydroperoxides). Saturation reactions result in the formation of saturated alcohols, acids and ketones. The species formed by photooxidation of the norbornene moieties can initiate the oxidation of the neighboring units of polyethylene and polypropylene, as shown by the FTIR analysis. The photoproducts formed during the first few hours of irradiation of AES were of the same nature as those observed for EPDM photooxidation [17]. During the first stages of irradiation, the degradation involves mainly the oxidation of the EPDM component. However, the growth of the band at 1515 cm - 1 , characteristic of the oxidation of the styrene units, gave evidence that the SAN component was slightly photooxidized even during the first stages of irradiation. Moreover, the development of this band

720

B. MAILHOT ETAL.

showed that the photooxidation of SAN was enhanced by the presence of the EPDM in AES, compared with SAN or PS. This result may be explained by the important flux of radicals resulting from the oxidation of EPDM. The same remark may be made concerning the discoloration measured at 350 nm. The increase in absorbance was higher for an AES film compared with SAN and PS. The photoproducts resulting from the oxidation of the styrene and acrylonitrile units accumulated at longer irradiation times (aliphatic and aromatic ketones and acids, amides, imides, etc.).

6 6.1

PHOTOOXIDATION OF BLENDS OF POLYSTYRENE AND POLY(VINYL METHYL ETHER) (PVME-PS) INTRODUCTION

The materials analyzed were blends of polystyrene (PS) and poly(vinyl methyl ether) (PVME) in various ratios. The two components are miscible in all proportions at ambient temperature. The photooxidation mechanisms of the homopolymers PS and PVME have been studied previously [4,7,8]. PVME has been shown to be much more sensitive to oxidation than PS and the rate of photooxidation of PVME was found to be approximately 10 times higher than that of PS. The photoproducts formed were identified by spectroscopy combined with chemical and physical treatments. The rate of oxidation of each component in the blend has been compared with the oxidation rate of the homopolymers studied separately. Because photooxidative aging induces modifications of the surface aspect of the material, the spectroscopic analysis of the photochemical behavior of the blend has been completed by an analysis of the surface of the samples by atomic force microscopy (AFM). A tentative correlation between the evolution of the roughness measured by AFM and the chemical changes occurring in the PVME—PS samples throughout irradiation is presented. 6.2

EXPERIMENTAL RESULTS

The IR analysis of PVME—PS blends of 50:50 mol% irradiated at A > 300 nm in the presence of oxygen shows two periods. During the first few hours of irradiation, the absorption bands formed were observed to be very similar to those observed for pure PVME photooxidized under the same conditions [4]. This result indicated that the bands observed in the blend reflected mainly the formation of the photoproducts of PVME oxidation. After less than 20 h of irradiation, under similar conditions, the oxidation of PVME alone reached a maximum and the IR absorbance corresponding to the oxidation products reached a maximum. Therefore, after this first period, the modifications of

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

721

the spectra of the blend should reflect the oxidation of pure PS. In order to verify that the increase of absorbance in this second period could be attributed principally to the formation of photoproducts resulting from the oxidation of PS, the spectrum of the film irradiated for 21 h was subtracted from those of the films irradiated for a longer duration. The absorption bands were found to be very different to those of pure PS. The spectra showed that once formed, the product at 1733 cm - 1 progressively disappeared. The disappearance of this product modified the shape of the carbonyl band compared with that obtained for the polystyrene photooxidation. The same decrease was observed in the case of the PVME homopolymer, and after a first phase of oxidation leading to the accumulation of the oxidation products, a progressive decrease of the carbonyl absorbance for longer irradiation times was observed. This reflected the loss of some of the oxidation photoproducts. The band at 1733 cm -1 has been attributed to ester groups of keto diester compounds such as dimethyl malonate or 1,3-dimethylacetone dicarboxylate. These molecular products can migrate out of the material or can be photolyzed by a Norrish type I reaction of the ketone, leading to products with a lower molecular weight: CH 3 -0 - C - C H 2 - C - C H 2 - C - O - C H 3 1.3-dimethylacetone dicarboxylate

„ low molecular weight products

The absorbance measured at 1733 cm - 1 plotted as a function of irradiation showed that for blends containing 80 and 50 mol% of PVME, the extent of the oxidation was proportional to the PVME content. For the 30 mol% PVME content, the rate of oxidation was lower than expected, in comparison with pure PVME during the first 30 h of irradiation. After this period, the increase of absorbance is related to the oxidation of PS. Therefore, the PS macromolecules may retard or reduce the photooxidation of the PVME macromolecules in PS-rich blends. Moreover, it has been noted that the photooxidation of the PS macromolecules starts earlier in the blends than in pure PS. This result was deduced from the increase in the band at 1515cm -1 , characteristic of a photoproduct formed only in PS [7]. These results indicated that the photooxidation of the PS macromolecules is accelerated by the presence of PVME.

6.3

SURFACE ANALYSIS

Images of the surface were recorded for increasing irradiation times by atomic force microscopy (AFM). Initially, the surface was flat and only some hill-like

722

B. MAILHOT ETAL.

structures were observable. During irradiation, in addition to the growth of these prominences, the whole surface became irregular. The dimensions of the hill-like structures were evaluated as 150–600 nm diameter and 3–20 nm height. Their diameter was of the same order as the droplets observed for phaseseparated blends [19,20]. The root mean square (RMS) roughness value was determined as a function of the irradiation time. Significant changes in the roughness were observed after 5–6 h of irradiation (Figure 30.8). After 12 h of irradiation, the roughness of the sample was too high to be measured by AFM.

6.4

DISCUSSION

The experimental results showed that, during the first 20 h of irradiation, the photochemical behavior of the blend may be related to that of PVME. To understand better the results of the surface analysis and to try to explain the phase separation phenomenon, a correlation should be made with PVME results. The IR analysis of the photooxidation of PVME homopolymer showed that during the first 5–6 h of irradiation an important increase in a band at 3290 cm - 1 was observed [4]. This band has been attributed to tertiary hydroperoxides. For prolonged irradiations, the hydroperoxide absorption band decreased. The decomposition of hydroperoxides yields acetates: OOH i — CH 2 -C —

o*

I — CH 2 -C —

hv,

'OH

OMe

C-CH,—

— CH,

OMe

MeO

The decomposition of hydroperoxides also produces a chain ketone and methanol: 25 20

15 10 50 0

2

.III

4 6 8 Irradiation time (hrs)

10

12

Figure 30.8 RMS roughness as a function of irradiation time for a PVME—PS blend of 64:36 mol%

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

723

+ "OH

OMe

+ R*

During the first 5–6 h of irradiation, the small droplets initially observed at the surface of the blend grew throughout photooxidation. This suggests that photooxidation may progressively induce a phase separation. The formation of the droplets may indicate the early beginnings of the phase separation. Since phase separation is probably not initiated by a chain scission which should favor solubility, the second mode of decomposition of hydroperoxides may be predominant. It has been shown that specific intermolecular interactions between the PS phenyl ring and the PVME methyl group in the solid state could account for the blend miscibility [21]. Since the decomposition of hydroperoxides leads to the loss of the methoxy groups, the oxidized PVME macromolecules may be less miscible with the PS macromolecules. After 5–6 h of irradiation, a maximum concentration of hydroperoxides is reached and phase separation occurs. The size of the droplets probably reaches a critical value and then the roughness increases significantly. The irregularity of the surface may thus be explained by the formation of oxidation photoproducts on the macromolecular backbone and by the formation of gaseous photoproducts that migrate out of the sample, creating craters. All the photoproducts recognized in the PVME—PS blend have already been identified in the corresponding homopolymers. However, the two components interact and the measurement of the oxidation kinetics gave evidence that the PVME macromolecules were slightly stabilized in the PS-rich blends and, conversely, the oxidation of PS macromolecules was enhanced in PVME-rich blends. By AFM analysis, it was shown that the modifications of the surface aspect can be characterized as a function of the irradiation time using the roughness parameters. A correlation between the modifications of the surface and the modifications of the chemical structure of the macromolecules resulting from irradiation showed that the degradation of the surface depends essentially on the decomposition of the hydroperoxides.

7

CONCLUSION

The first step of the mechanism of photooxidation of PS is the abstraction of a hydrogen atom on the tertiary carbon of the macromolecule structure, which

724

B. MAILHOT ETAL

leads to a polystyryl radical. This radical reacts with oxygen, producing numerous products, including low molecular weight compounds. The numerous steps of the oxidation mechanism have been identified for all the styrenic polymers investigated in the present study. However, the importance of the various ways of evolution depends on the chemical structure of the copolymers and the blends. It is worth noting that the analysis of the behavior of the copolymers or blends requires at first a detailed knowledge of the photochemical behavior of each component studied separately. Moreover, it turned out that the study of the photochemical behavior of the copolymers and physical blends was necessary, because the mechanism and/or the kinetics of photodegradation may not be deduced from the study of the separated compounds. In styrene-stat-acrylonitrile, evidence was obtained that interactions exist between the two components, leading to: • the formation of photoproducts specific to the copolymer that were not identified in PS or PAN; • an unexpected reactivity of acids that usually accumulate in the polymer: in SAN the acid groups react with the nitrile units. Finally, even if acrylonitrile can be considered to be less photoreactive than PS, the copolymer presents more oxidizability than the PS homopolymer. In ABS, because the BR units are more photosensitive than the PS units, they are photooxidized in the first steps of the reaction. The radicals which are formed can attack the neighboring PS units. Moreover, the grafted sites of the PS macromolecules are the starting sites of an additional route of photooxidation of the PS units. Therefore, the kinetics of oxidation of the copolymer ABS are twice as fast as expected on the basis of only addition of the photooxidation rates of the two polymers studied separately. In the AES and PS—PVME blends, no interaction in terms of nature or reactivity of the photoproducts formed was detected. All the blends contained an elastomeric phase which has been shown always to be the most oxidizable component. EPDM and PVME degrade in the first few hours of irradiation and the photoproducts resulting from the oxidation of SAN or PS accumulate at longer irradiation times. However, the styrene units are oxidized faster in the blends than in the homopolymer PS or in SAN. In addition, in the PVME—PS blends rich in PS polymer, the PS retarded the photodegradation of PVME. Finally, it can be deduced from these studies that the major difference between the copolymers and the blends is the possibility of new routes of photooxidation generated by the structure of the copolymer itself when the different monomer units are chemically bonded. These additional ways of degradation increase the photosensitivity of the copolymers.

PHOTOCHEMICAL DEGRADATION OF STYRENIC POLYMERS

725

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

Lawrence, J.B., Weir, N.A., J. Polym. ScL, Polym. Chem. Ed., 11, 105 (1973). Weir, N.A., Whiting, K., Eur. Polym. J., 3, 291 (1989). Mailhot, B., Gardette, J.L., Polym. Degrad. Stab., 44, 223 (1994). Mailhot, B., Morel, S., Gardette, J.L., Polym. Degrad. Stab., 62, 117 (1998). Piton, M., Rivaton, A., Polym. Degrad. Stab., 55, 147 (1997). Mailhot, B., Gardette, J.L., Vib. Spectrosc., 11, 69 (1996). Mailhot, B, Gardette, J.L., Macromolecules, 25, 4119 (1992). Mailhot, B, Gardette, J.L., Macromolecules, 25, 4127 (1992). March, J., Advanced Organic Chemistry, McGraw-Hill Kogakusha, Tokyo, 1977. Chien, J.C.W., Jabloner, H. J. Polym. Sci, Part Al, 6, 393 (1968). Mailhot, B., Gardette, J.L., Polym. Degrad. Stab., 44, 237 (1994). Jouan, X., Gardette, J.L., J. Polym. Sci., Part A, 29, 685 (1991). Jouan, X., Gardette, J.L., Polym. Degrad. Stab., 36, 91 (1992). Piton, M., Rivaton, A., Polym. Degrad. Stab., 53, 343 (1996). Adam, C., Lacoste, J., Lemaire, J., Polym. Degrad. Stab., 24, 185 (1989). Adam, C., Lacoste, J., Lemaire, J., Polym. Degrad. Stab., 29, 305 (1990). Coiffier, F., Arnaud, R., Lemaire, J., Makromol. Chem., 185, 1095 (1984). Delprat, P., PhD Thesis, Universite Blaise Pascal, Clermont-Ferrand, 1993. Tanaka K., Yoon J.S., Takahara A., Kajiyama T., Macromolecules, 28, 934 (1995). Karim A, Slawecki T.M., Kumar S.K., Douglas J.F., Satija S.K., Han C.C., Russell T.P., Liu Y., Overney R. Sokolov J., Rafailovich M.H., Macromolecules, 31, 857 (1998). 21. White, J.L., Mirau, P., Macromolecules, 26, 3049 (1993).

This page intentionally left blank

31

Analysis and Levels of Styrene Dimers and Trimers in Polystyrene Food Containers HIROMI SAKAMOTO Kanagawa Environmental Research Center, Japan

1

INTRODUCTION

In Japan, the production of polystyrene (PS) is fairly high; for example, 115 0000t were produced in 1999, involving one of the four major generalpurpose type resins. It has been reported that some kinds of styrene oligomers, which are by-products of polymerization, remained in PS products [1]. In the Wingspread statement announced by Colborn et al. in 1996, it was reported that certain synthetic chemicals have endocrine-disrupting effects on humans or animals. Most of these chemicals are pesticides, but styrene dimers (SDs) and trimers (STs) are listed in this statement along with plasticizers and antioxidants [2]. Accepting this statement and sensational reports by the mass media, the Japan Environmental Protection Agency (JEPA) published 'Strategic Programs on Environmental Endocrine Disrupters in 1998 (SPEED'98)', presented JEPA's basic policy on this subject, initiated concrete countermeasures and listed about 70 substances, including SDs and STs, that are suspected of being endocrine disrupters (EDs) in a review of the literature [3]. Thereafter, researchers affiliated with the National Institute of Health Sciences (NIHS) investigated certain kinds of containers made of PS, and it was clarified that

Modern Styrenic Polymers: Polystyrene and Styrenic Copolymers. Edited by J. Scheirs and D. B. Priddy :<;} 2003 John Wiley & Sons Ltd

728

H. SAKAMOTO

SDs and STs migrated into food simulated-solvent, soup and noodles [4–6]. Following these reports, a controversy called 'dispute about instant noodles' arose: 'are instant noodles safe to eat or not?' However, from a review of the literature, there is no clear basis for determining whether SDs and STs have endocrine-disrupting effects or not. Nevertheless, from studies carried out following this dispute, many good results were obtained, such as the preparation of standard substances, establishing limits of levels of chemical contained in products and the biological evaluation of SDs and STs. These results are discussed in this chapter.

2 STRUCTURE AND ANALYSIS OF STYRENE DIMERS AND TRIMERS SDs and STs were identified by Kawamura et al. at the NIHS [1]. There are three isomers among SDs, 2,4-diphenyl-l-butene, which is a linear SD (D-3) and cis- and trans-l,2-diphenylcyclobutane, which are cyclic SDs (D-2 and D-4, Figure 31.1). Furthermore, 1,3-diphenylpropane is also considered a kind of SD in some cases (D-l, Figure 31.2). As it is reported that this substance is generated by pyrolysis of PS, it is considered to be formed by thermal polymerization or molding. 2,4,6-Triphenyl-l-hexene is a linear ST (T-l, Figure 31.3), and l-phenyl-4-(l'-phenylethyl)tetralin (Figure 31.4), which are cyclic STs, have four stereoisomers (T-2-T-4). Triphenylcyclohexane (Figure 31.5) is an other cyclic ST (T-5), and there are two STs (T-6, T-7) for which the structures have not been identified. SDs and STs are analyzed together by gas chromatography-mass spectrometry (GC-MS). Before injection, PS samples are soaked in cyclohexane-2-propanol (1:1) overnight at 37 °C or dissolved in dichloromethane and methanol is added to precipitate highly polymerized compounds [4,7]. An example of the GC-MS conditions used is as follows [7]: • GC-MS system: HP5973A, Hewlett-Packard • Column 1: DB-1 (0.25mm i.d. x 30m, film thickness 0.25 jim), J & W Scientific • Column 2: Supelcowax 10 (0.25mm i. d. x 30m, film thickness 0.25 jmi), Sigma-Aldrich Japan, Supelco Inc. (this column is used for separation and quantitative analysis of cyclic trimers) • Temperatures: Oven: 40 °C (2min), ramped at 10°C/min to 280 °C (5min) Injection: 250 °C Ion source: 230 °C • Flow-rate: 1 mL/min (constant flow mode) • lonization: 70 eV

STYRENE DIMERS AND TRIMERS IN POLYSTYRENE FOOD CONTAINERS 729

1,2-Diphenylcyclobutane (cis : D-2, trans : D-4) CH, = C-CH 2 -CH 2

2,4- Diphenyl-1-butene (D-3)

Figure 31.1 Structures of styrene dimers

Figure 31.2 Structure of 1,3-diphenylpropane (D-1)

H2C = C-CH 2 -CH-CH 2 -CH 2

2,4,6-Triphenyl-1 -hexene (T-l) Figure 31.3 Structure of linear styrene trimer

Table 31.1 gives the mlz values of standard substances.

730

H. SAKAMOTO

H3C

1 -Phenyl-4-( 1 '-phenylethyl)tetralin (4 stereo isomers: T2-T4) Figure 31.4

Structure of cyclic styrene trimers

Figure 31.5

Structure of triphenylcyclohexane (T-6)

3

CONTENT OF SDs AND STs IN PS FOOD CONTAINERS

According to our investigations, the contents of SD and STs in disposable lunch boxes differ with each product [8]. The analyzed samples were disposable lunch boxes from the market and boxes containing convenience foods and lunches. Previously, most of these lunch boxes were made of PS only, but recently have gradually changed to polypropylene (PP). Table 31.2 shows the levels of compounds found in various kinds of disposable lunch boxes.

STYRENE DIMERS AND TRIMERS IN POLYSTYRENE FOOD CONTAINERS 731 Table 31.1

Peak No.

1 2 3 4 5 6 7 8 9 10

mlz values of SDs and STs

Code

Substance

D-l D-2 D-3 D-4 T-l T-2 T-3

1 , 3 -Diphenylpropane cis- 1 ,2-Diphenylcyclobutane 2,4-Diphenyl- 1 -butene trans-\ ,2-Diphenylcyclobutane 2,4,6-Triphenyl- 1 -hexene 1 e-Phenyl-4e-( 1 '-phenylethyl)tetralin 1 a-Phenyl-4e-( r-phenylethyl)tetralin (T-3a) 1 a-Phenyl-4a-( 1 '-phenylethyl)tetralin (T-3b) 1 e-Phenyl-4a-( 1 '-phenylethyl)tetralin Triphenylcyclohexane

T-4 T-5

mlz 92.105 104.208 91.208 104.208 91.207 91.207 91.207 91.207 91.207 104.208

Because the supply of standard substances was too late for investigation, only D-3 was analyzed among three SDs. The levels of four stereoisomers of l-phenyl-4-(l'-phenylethyl)tetralin were higher than those of the other substances. The sum of STs ranged from 1860 to 6470 |xg/g. Kawamura et al. investigated PS food containers and also concluded that the content of SDs and STs differed in each product [4]. From the results in Table 31.3, the levels of SDs and STs were 90–1030 and 650–20 770 M-g/g, respectively. These data were obtained by GC with flame ionization detection (FID), because SD and ST standard substances have not yet been established. The levels of D-3 and D-4 are higher than those of the other SDs, and the levels of T-l, T-2, T-3 and T-4 are higher than those of the other STs. The sum of trimers is at least 10 times higher than the sum of dimers. In a noodle cup, for example, the content of SDs and STs covered a wide range. There are certain kinds of PS that are used in food containers, such as generalpurpose PS (GPPS), high-impact PS (HIPS), EPS and PS paper (PSP). In these studies, the content of SDs and STs did not always depend on the type of PS. However, in other studies investigating various types of PS containers for instant foods, the contents of SDs and STs in EPS, for example, Chinese noodles Nos 1-3 and soup Nos 1-5, showed were very low (Table 31.4). The sum of SDs and STs ranged from 60 to 130 and from 370 to 680 |Jig/g, respectively. As polymerization of EPS is a low-temperature reaction, it was expected that only traces of by-products of polymerization would remain [6].

4

MIGRATION OF SDs AND STs FROM PS FOOD CONTAINERS

Using certain kinds of solvents called food simulated solvents, migration from PS food containers was investigated [4]. Distilled water, an aqueous solution of

Table 31.2 Contents of styrene, butylated hydroxytoluene (BHT) and styrene, dimer and trimers [8]. Reproduced by permission of the Food Hygiene Society of Japan Sample

Material

Surface area Styrene BHT SD1 (D-3) ST1 (T-l) ST2 (T-2) ST3 (T-3a) ST4 (T-4b) ST5 (T-4) Sum of trimers (cm2) (jig/g) (>ig/g) (jtg/g) (jtg/g) (jJig/g) (n-g/g) (txg/g) (ng/g) (ng/g)

Lunch boxes on the market 341 HIPS0 A Lid Container PP + filler* 230 B Lid GPPSC 225 Container PP + talc* 224 Convenience stores CLid GPPS 338 Container PP 307 GPPS 450 DLid Container PP 350 E Lid GPPS 418 304 Container PP 361 GPPS F Lid 250 Container PP GPPS GLid 376

270 ND' 220 ND

ND 130 ND 40.5

47.3

ND 220 ND

260 ND 680 ND

690 ND 820 ND

960 ND 1130 ND

310 ND 370 ND

410 ND 480 ND

280 ND 140 ND 230 ND 270 ND 150

ND 61.5 ND 135 ND 140 ND 470 ND

150 ND 110 ND 150 ND 74.0 ND 79.3

1110 ND 1050 ND 1000 ND 780 ND 670

1110 ND 1050 ND 960 ND 1160 ND 1310

1610 ND 1520 ND 1410 ND 1680 ND 1880

520 ND 490 ND 440 ND 550 ND 630

710 ND 680 ND 610 ND 770 ND 880

2630

—

3480

— 5060

— 4790

—

4420

—

4940

—

5370

Container PP 315 ND 84.5 ND ND ND HLid GPPS 414 210 ND 67.5 520 1770 Container PP 269 200 31.8 24.8 260 500 Others (department store, shop of livelihood cooperative and supermarket) I Lid PSPrf 423 230 ND 74.0 990 1620 Container PSP 245 11.0 110 1140 260 790 J Lid GPPS 475 ND 110 250 680 1130 Container PP + filler* 378 ND 46.3 ND ND ND K Lid GPPS 449 ND 150 330 820 1030 Container PP 297 ND 20.1 ND ND ND

ND 2270 590

ND 830 260

ND 1010 250

— 6400 1860

2150 1560 1570 ND 1480 ND

770 530 530 ND 480 ND

940 700 730 ND 640 ND

6470 4720 4640 — 4450 —

" HIPS: high-impact polystyrene. * Material was shown on the container. c GPPS: general-purpose polystyrene. d PSP: polystyrene paper. e Styrene, BHT, SD1 and ST1: ND - 0.05 jig/g; ST2, ST3, ST4 and ST5: ND = 0.5 jxg/g.

Table 31.3 Contents of styrene dimers and trimers in polystyrene products [4]a. Reproduced by permission of the Food Hygiene Society of Japan

Content (|xg/g) Type

Sample

GPPS

1 Cup 2 Cup 3 Cup 4 Food case 5 Food case 6 Food case

HIPS

1 Cup 2 Cup 3 Cup 4 Chow mein case 5 Yoghurt case 6 Tofu case

PS foam

1 Cup 2 Food case

D-l

D-2 D-3 D-4 T-l

Sum of contents (fxg/g) T-2 T-3

ND 40 130 470 190 ND 80 200 1800 1240 30 40 220 1200 1820 10 340 210 3500 1420 ND 430 30 3560 3720 ND 110 60 1700 1000 40 220 1390 1910 ND 30 30 220 1320 2080 ND 60 20 30 120 860 2540 1990 10 30 100 200 2370 2330 60 ND 600 ND 5600 4000 ND 30 100 200 1300 1600 ND ND 80 30 340 70 30 20 130 130 1790 1170

20 ND* ND 30 40 10

T-4 T-5 T-6 T-7 Dimers Trimers Total

510 2510 3210 3220 7380 2040

180 ND ND ND 760 30 220 110 960 60 360 140 990 20 210 110 2000 ND 690 30 640 30 190 60

3830 4200 4130 4630 8000 3400

1110 1150 1230 1310 2000 920

180 2540

30 70 90 110 60 50

240 410 450 530 800 360

190 280 290 590 500 180

1350 6670 7750 9470 17380 5660

1540 6950 8040 10060 17880 5840

120 290 180 310 200 1030 240 340 310 660 200 330

8630 9410 10630 11520 20770 7830

8920 9720 11660 11860 21430 8160

110 310

650 6880

760 7190

60 ND ND ND 730 340 170 140

3 Food case 4 Bowl (donburi) 5 Bowl (donburi) 6 Tray 7 Noodle cup 8 Noodle cup 9 Noodle cup 10 Noodle cup 1 1 Noodle (udon) cup 12 Chow mein case 13 Spaghetti cup

20 10 20 30 40 20 10 ND 20 20 60

30 50 20 ND ND ND ND 30 ND ND ND

110 190 170 310 160 90 60 40 150 150 450

Mean S.D.

20 20

10 20

160 190 2070 1820 3700 1050 150 170 1190 1190 2380 640

380 250 140 60 200 20 20 300 180 140 110

2520 2740 2420 2000 1450 560 500 2400 2270 2130 3770

1980 1500 1780 640 3300 80 80 3800 1950 1940 3800

4060 3310 3800 1280 6610 130 140 7900 3940 3890 7780

1200 1120 1100 360 1940 ND ND 2040 1160 1130 2200

170 150 150 40 280 ND ND 300 220 180 300

540 500 350 400 400 130 90 370 350 310 620

10460 9230 9710 4410 14370 770 720 17520 10070 9790 18720

11000 9730 10060 4810 14770 900 810 17890 10420 10100 19340

70 370 150 70 260 100

380 200

9210 5500

9590 5640

50 120 40 ND 110 ND ND 180 90 90 80

480 290 420 90 680 ND ND 900 440 430 790

" Amounts of dimers and trimers were calculated by comparison with the area of 1. 3-diphenylpropane and benzylbutyl phthalate, respectively. b

XTIPV

,f

1A

rr / rv

CO

cn

736

H. SAKAMOTO

Table 31.4 Contents of styrene dimers and trimers in material of containers [6f Sum of contents (|Ag/g)

Content (n-g/g) Sample

D-3 D-4 T-l

1 2 3 4 5 6 7 8 9 10 11 12 13 14-1 14-2 15 1 Japanese noodles 2 3 Buckwheat noodles 1 1 Chow mein 2 3 1 Spaghetti 2 1 Soup 2 3 4 5 1 Rice 2 3

70 70 50 110 110 170 120 120 100 130 160 150 100 190 180 70 130 60 160 100 70 70 80 60 130 100 90 60 60 50 70 70 60

Chinese noodles

10* 20 10 200 90 160 100 130 110 110 90 120 160 250 240 170 130 170 130 210 310 110 210 150 100 30 20 10 10 30 190 180 170

350 380 280 1580 1280 1640 1380 1410 1250 1500 1950 1560 860 2130 2000 1230 1840 1030 1720 1790 1800 960 1000 1460 1420 300 310 390 450 340 1660 1530 1480

T-2 T-3 T-4 Dimers Trimers Total 30 40 30 1650 1360 1880 740 1290 1220 1600 1600 1670 1680 3050 2200 900 1120 920 1720 1600 2350 880 2300 1200 1540 30 30 40 40 90 1710 840 1200

50 60 40 3350 2850 5000 1780 3400 1810 3550 3450 3550 3400 7350 5450 2250 2520 2450 3490 3150 4900 2750 4740 2150 3300 50 60 60 100 180 3450 2130 2660

30 40 20 1090 950 1670 700 1140 670 1150 1090 1190 1260 2250 1670 890 1100 850 1330 970 1920 1010 1900 820 1250 20 30 30 50 70 1410 780 1030

80 90 60 310 200 330 220 250 210 230 250 270 260 440 420 240 260 230 290 310 380 180 290 210 230 130 110 70 70 80 260 250 230

460 520 370 7670 6440 10190 4600 7240 4950 7800 8090 7970 7200 14780 11320 5270 6580 5250 8240 7510 10970 5600 9940 5630 7510 400 430 520 640 680 8230 5280 6370

540 610 430 7980 6640 10520 4820 7490 5160 8030 8340 8240 7460 15220 11740 5510 6840 5480 8530 7820 11350 5780 10230 5840 7740 530 540 590 710 760 8490 5530 6600

a

Amounts of dimers and trimers were calculated by comparison with the area of 1,3-diphenylpropane and benzylbutyl phthalate, respectively. ND< 10u,g/g.

b

20 or 50 % ethanol and w-heptane were used as the food simulated solvents. In Japan, each utensil, package and container made of plastic is regulated by the Food Sanitation Law, and there are individual standards on migration tests [9]. Generally, for utensils, packages and containers for water-soluble food, the migration test is carried out using distilled water or 4% acetic acid. On the

STYRENE DIMERS AND TRIMERS IN POLYSTYRENE FOOD CONTAINERS 737

other hand, 20% ethanol or n-heptane is used for the migration test for containers that hold fatty foods. Detailed conditions, such as temperature and time of test, are described in each standard. In this study, the migration of SDs and STs was determined by the solubility of the solvent to fat (Table 31.5). SDs and STs migrated into «-heptane at a specific level, into 20 and 50% ethanol only slightly and did not migrate into distilled water. Remarkable differences according to the types of PS were recognized, and containers made of HIPS showed the highest concentration of migrant chemicals. It was considered that a fatty solvent was optimum in the case of HIPS, because an elastomer has been added to the PS. From the results of our studies, SD and STs migrated into vegetable oil, which was used as a fatty food simulated solvent, when heated in a microwave oven (Table 31.6) [8]. After heating for 90s, SDs and STs hardly migrated into vegetable oil. However, after heating for 180s, almost all SDs and STs were detected at high concentrations. The levels of SDs and STs were 0.1-8.4 and 3.0–336ng/cm2, respectively. Moreover, when containers were left for l0 min after having been heated, migration of STs increased. After storage for 24 h at 20 °C, the migration level of STs was also relatively high. The sum of STs ranged from 88.1 to 1290 ng/cm2. Therefore, we concluded that migration of STs into vegetable oil was mainly affected by the length of contact time with the containers; however SDs were mainly affected by the heating time. It was also found that the SDs and STs migrated from a container made of HIPS showed a slightly higher level than other types of PS.

5

BIOLOGICAL EVALUATION OF SDs AND STs

As described previously, SDs and STs are listed in EDs, even though little has been reported to provide a definite basis. Subsequently, a Japanese food company using PS for instant noodle cups investigated the endocrine-disrupting effects of styrene oligomers [10,11]. They examined the effects of styrene monomer, SDs (D-2, D-3 and D-4) and STs (T-l, a mixture of T-2-T-5 and T-6) on an estrogen receptor (ER) binding assay, androgen receptor (AR) binding assay, proliferation of MCF-7 human breast cancer cells (Figure 31.6), steroidogenesis in Ley dig cells of rats (Figure 31.7) and uterotrophic assay of immature rats (Table 31.7). It was concluded that the styrene monomer, SDs and STs tested in these experiments had no endocrine-disrupting effects through ER, AR and steroidogenesis mechanisms. They also examined these substances in a thyroid hormone receptor (THR) binding assay, the Hershberger assay (a response test for castrated male rats) and effects on the

Table 31.5

Migration of styrene dimers and trimers from polystyrene products [4]'\a,b Migrant (|Ag/cm2)I

Sum of migrants (|xg/cm2)

Sample

Solvent

D-l

D-2 D-3 D-4 T-l

T-2

T-3

T-4

T-5

T-6 T-7

Dimers

Trimers

Total

GPPS 4

Water 20% Ethanol 50% Ethanol n-Heptane Water 20% Ethanol 50% Ethanol n-Heptane Water 20% Ethanol 50% Ethanol n-Heptane

ND' ND ND ND ND ND ND 0.02 ND ND ND ND

ND ND ND ND ND ND ND 0.05 ND ND ND ND

ND ND ND 0.04 ND ND 0.04 9.5 ND ND ND 0.80

ND ND ND 0.18 ND ND 0.06 18.9 ND ND ND 2.1

ND ND ND ND ND ND ND 5.4 ND ND ND 0.70

ND ND ND ND ND ND ND 0.44 ND ND ND ND

ND ND ND ND ND ND ND 2.4 ND ND ND ND

ND ND ND ND ND ND ND 0.86 ND ND ND 0.26

ND 0.01 0.06 0.38 ND 0.01 0.14 43.9 ND 0.02 0.07 6.1

ND 0.01 0.06 0.38 ND 0.01 0.14 44.8 ND 0.02 0.07 6.4

HIPS 1

PS foam 4

ND ND ND ND ND ND ND 0.19 ND ND ND 0.12

ND ND ND ND ND ND ND 0.60 ND ND ND 0.14

ND 0.01 0.06 0.16 ND 0.01 0.04 6.5 ND 0.02 0.07 2.5

ND ND ND ND ND ND ND 0.78 ND ND ND ND

" Migration condition: n-heptane 25 °C 60min, other solvents 60°C 30min. h Amounts of dimers and trimers were calculated by comparison with the area of 1,3-diphenylpropane and benzylbutyl phthalate, respectively. ' ND < 0.01

Table 31.6

Results of migration test [8]

Conditions

Migration level (ng/cm2)

Heating time Leaving time Substance

A

90s

0 min

90s

10 min

180s

0 min

NDa 1.8 ND 0.5 17.3 23.8 7.9 9.1 58.6 ND 6.3 0.3 1.9 60.3 65.3 32.6 30.9 191 ND 4.3 0.1 1.4 49.3 64.3

Styrene BHT SD1 (D-2) ST1 (T-l) ST2 (T-2) ST3 (T-3a) ST4 (T-3b) ST5 (T-4) Sum of trimers Styrene BHT SD1 (D-3) ST1 (T-l) ST2 (T-2) ST3 (T-3a) ST4 (T-3b) ST5 (T-4) Sum of trimers Styrene BHT SD1 (D-3) ST1 (T-l) ST2 (T-2) ST3 (T-3a)

B

ND 0.6 0.1 0.4 7.4 9.3 4.6 3.6 25.3 ND 3.6 0.3 0.6 14.8 18.8 9.4 7.1 50.7 ND 3.6 0.3 0.8 19.4 23.5

C

D

E

F

G

H

ND 0.2 ND ND ND ND ND ND — ND 2.5 0.1 0.1 ND ND ND ND 0.1 ND 6.2 0.2 0.2 0.5 1.4

6.4 4.1 ND 0.4 4.3 4.0 4.6 3.5 16.8 7.8 14.0 0.1 0.6 4.5 1.6 3.0 5.9 15.6 ND 0.8 0.1 1.7 1.2 4.4

0.2 4.6 0.5 ND ND ND ND ND — 0.4 90.2 2.6 0.5 3.0 34.6 0.6 9.3 48.0 0.5 12.6 0.5 34.3 79.6 155

0.2 74.6 ND 0.7 ND ND ND ND 0.7 1.0 466 ND 0.6 ND ND ND ND 0.6 0.5 707 3.8 1.8 20.5 51.7

1.0 0.2 6.7 5.8 ND ND 0.8 ND ND ND ND ND ND ND ND ND 0.8 — 0.8 ND 36.3 10.1 0.6 ND 0.4 6.3 ND 29.3 ND 52.5 ND 13.4 ND 27.8 0.4 129 1.2 4.5 22.8 29.1 2.3 8.6 2.4 6.6 21.3 63.0 50.5 119

I

J

K

ND 3.0 ND ND ND ND ND ND — ND 3.7 0.2 0.9 3.3 8.0 1.8 2.3 16.3 ND 20.3 6.4 0.7 30.8 10.5

ND 1.7 ND 0.1 ND ND ND ND 0.1 0.1 2.1 ND ND ND ND ND ND — 0.1 3.8 ND ND ND ND

0.1 3.1 ND ND ND ND ND ND — 0.1 21.4 0.6 0.4 ND ND ND ND 0.4 ND 14.6 1.7 10.1 15.7 33.2

(continues) o CO

g Table 31 .6

(continued) Migration level (ng/cm2)

Conditions Heating time

180s

24 hr storage (20°C)

Leaving time Substance

lOmin

A

B

32.1 11.8 ST4 (T-3b) ST5 (T-4) 32.0 8.6 Sum of trimers 179 64.1 ND ND Styrene 19.3 BHT 5.1 SD1 (D-3) 0.8 0.8 ST1 (T-l) 5.1 3.9 17.4 99.9 ST2 (T-2) 23.3 131 ST3 (T-3a) 11.6 65.8 ST4 (T-3b) 9.0 50.6 ST5 (T-4) Sum of trimers 66.4 351 0.3 Styrene 0.3

BHT

2.6

SD1 (D-3) 0.8 8.7 ST1 (T-l) ST2 (T-2) 32.7 ST3 (T-3a) 18.1 ST4 (T-3b) 44.9 ST5 (T-4) 22.3 Sum of trimers 127

4.5 0.4 1.3 29.9 37.7 18.9 13.9 102

C

D

0.5 0.1 0.4 4.4 3.0 11.8 0.2 10.5 7.9 14.5 1.5 1.0 3.7 0.2 4.1 ND 6.7 ND 2.3 ND 3.2 ND 20.0 0.2 ND 9.9 5.8 19.4 ND ND 18.5 37.9 18.7 32.8 27.6 40.0 9.2 12.3 14.1 23.8 88.1 147

E

53.4 14.0 336 1.1 99.1 8.8 1.8 5.6 26.8 0.3 5.9 40.4 0.4 10.8 0.9 69.2 187 301 68.3 116 742

F 9.9 22.8

107 2.1 973 7.3 18.3 146 269 62.5 143 639 1.4 71.8 ND 34.1 306 501 156 294 1290

Styrene, BHT, SD1 and SD2: ND = 0.1 ng/cm2; ST2, ST3, ST4 and ST5: ND = 1 ng/cm2

G 7.6 4.6 86.4 1.7 54.1 8.4 17.4 131 226 62.1 119 556 1.4 6.9 ND 69.7 274 462 135 74.0 1010

H 28.2 68.1

285 1.1 28.9 0.3 9.3 224 424 123 234 1010 1.5 10.0 1.6 59.7 285 449 125 261 1180

I

J

2.9 3.7 48.6 ND 4.7 4.7 64.5 290 406 135 173 1070 ND 11.6 0.8 47.3 209 289 95.8 122 763

ND ND — 0.4 4.8 1.4 4.0 3.5 7.2 2.3 4.2 21.2 0.3 2.2 0.3 27.1 22.8 40.1 12.9 17.6 121

K 8.6 13.5 81.1

0.1 14.9

1.4 33.8 64.5 99.0 30.0 36.9

264 0.1 5.8 ND 30.3 60.9 95.4 28.1 37.6 252

STYRENE DIMERS AND TRIMERS IN POLYSTYRENE FOOD CONTAINERS 741

(A)

0 -14- 13-12-11-10-9 -8 -7 -6 -5 -4 Concentration (Log mol/L)

Estradiol /7-Nonylphenol Bisphenol A SM

(B)

-14-13-12-11-10-9 -8 -7 -6 -5 -4 Concentration (Log mol/L) Estradiol SD-01(D-3) SD-08(D-2) SD-09(D-4) ST-Ol(T-l) ST-02(T-5) ST-03 (Mixture of T-2-T-4)

Figure 31.6 Effects of styrene monomer (SM) (A), styrene dinners (SD-01, SD-08, SD-09) and styrene trimers (ST-01, ST-02, ST-03) (B) on proliferation of MCF-7 cells [10]. The cells were treated with test compounds for 6 days. *,** p < 0.05, 0.01 (vs control). Each value represents the mean ±SD (n = 3 wells)

serum prolactin concentration in ovariectomized female rats, and it appeared that these substances had no endocrine-disrupting effects [11]. JEPA examined the literature on SDs and STs to determine whether risk assessment was really necessary. Except for the estrogenic effects, it was decided that the assessment was not necessary. Additional examinations were performed on ER binding assay, proliferation of MCF-7 human breast cancer cells (E-screen assay), yeast two hybrid assay and yeast estrogen selective (YES) assay [12]. From the results of these examinations, it was judged that estimating the risk of SDs and STs is not necessary at the present time. JEPA officially announced a revised edition of SPEED'98 in 2000, in which SDs and STs had been deleted from the list of EDs along with n-butylbenzene.

H. SAKAMOTO

l(Hmol/L

10-5mol/L

Figure 31.7 Effects of styrene monomer (SM), styrene dimers (SD-01, SD-08, SD09) and styrene trimers (ST-01, ST-02, ST-03) on hCG stimulated testosterone production by Leydig cells of rats [10]. Testicular interstitial cells containing Leydig cells were incubated with test compounds for 1 h followed by the addition of 0.1 ILJ/mL hCG to the assigned wells. After incubation for 3h culture media were removed and analyzed for testosterone concentrations by EIA. ##: p < 0.01 (vs. -hCG), **: p < 0.01 (vs + hCG). Each value represents the mean ± (n = 3 wells)

6

CONCLUSION

PS products contain certain kinds of styrene dimers (SDs), such as 2,4-diphenyl-1-butene and cis- and trans-l,2-diphenylcyclobutane, and styrene trimers (STs), such as 2,4,6-triphenyl-l-hexene and l-phenyl-4-(r-phenylethyl)tetralin. The contents of these substances differ with each product. It has also been found that SDs and STs migrate into fatty solvents, such as n-heptane or vegetable oil, but the migration into aqueous solvents is very slight. SDs tend to be affected by the heating time of containers involving solvents, and STs tend to be affected by the length of contact time with containers. According to the results of biological evaluations in Japan, it was concluded that styrene, SDs and STs have no endocrine-disrupting effects, contrary to expectation. Further studies are expected in this area.

STYRENE DIMERS AND TRIMERS IN POLYSTYRENE FOOD CONTAINERS 743 Table 31.7 Summary of the effects of styrene monomer (SM), styrene dimers (SD-01, SD-08, SD-09) and styrene trimers (ST-01, ST-02, ST-03) on immature rat uterotrophic assay [10] Histopathological evaluation"

Compound Control Estradiol /•-Nonylphenol Bisphenol A

SM SD-01 (D-3) SD-08 (D-2) SD-09 (D-4)

ST-01 (T-l) ST-02 (T-6) ST-03 (Mixture of T-2-T5)

Dose (mg/kg)

Increase in uterine weight"

0.04 20 200 40 400 20 200 20 200 20 200 20 200 20 200 20 200 20

+++

++ ++

Hypertrophy of uterine luminal epithelium

— + + _ + -

Cornification of vaginal mucosa

_ + + +

-

-

-

200 " +: Positive test results; relative levels of effects are indicated by the number of + signs. —: Negative test results.

REFERENCES 1. Kawamura Y., Sugimoto N., Takeda Y. and Yamada T., J. Food Hyg, Soc. Jpn., 39, 110(1998). 2. Colborn T., Dumanoski D. and Myers J.P., Our Stolen Future, Penguin, New York (1996). 3. Japan Environment Agency, Strategic Programs on Environmental Endocrine Disruptors in 1998 (SPEED '98), JEA, Tokyo (1998). 4. Kawamura Y., Kawamura M., Takeda Y. and Yamada T., /. Food Hyg. Soc. Jpn., 39, 199 (1998). 5. Kawamura Y., Nishi K., Sasaki H. and Yamada T.. J. Food Hyg. Soc. Jpn., 39, 310 (1998). 6. Kawamura Y., Nishi K., Maehara T. and Yamada T., /. Food Hyg. Soc. Jpn., 39, 390 (1998).

744

H. SAKAMOTO

7. Sakamoto H., in Proceedings of the Fourth International Conference on ECOMATERIALS, p. 113 (1999); available from the Society of Non-Traditional Technology ECOMATERIALS Forum, Tokyo, Japan. 8. Sakamoto H., Matsuzaka A., Itoh R. and Tohyama Y., J. Food Hyg. Soc. Jpn., 41, 200 (2000). 9. Standard Methods of Analysis for Hygienic Chemists - With Commentary, Kinbara, Tokyo (1990). 10. Nobuhara Y., Hirano S., Azuma Y., Date K., Ohno K., Tanaka K., Matsushiro S., Sakurai T., Shiozawa S., Chiba M. and Yamada T., J. Food Hyg. Soc. Jpn., 40, 36 (1999). 11. Azuma Y., Nobuhara Y., Date K., Ohno K., Tanaka K., Hirano S., Kobayashi K., Sakurai T., Chiba M. and Yamada T., /. Food Hyg. Soc. Jpn., 41, 109 (2000). 12. Japan Environmental Protection Agency, Data for the 2nd Committee on EDs in 2000, JEPA, Tokyo (2000).

Index Note: Page references followed by 'f' represents a figure and 't' represents a table. acetophenone 149, 539, 582, 715 acrylates 157, 343, 497 acrylic acid 126 acrylic acid-glycidyl methacrylate (AA-GMA) 563–564 acrylic acid-styrene 177 acrylic copolymers 492 acrylonitrile 283, 333, 677, 709, 724 acrylonitrile-butadiene-styrene (ABS) 18–20, 26, 40f, 41, 153, 281, 294, 679, 699, 704 bulk-produced 345 crazes 32f fracture behaviour 633-663 heat resistant technology 321-340 mechanical properties control 679 modification 493-495 photooxidation 703, 712–716, 724 recycling 407 rubber particle formation 305-319 acrylonitrile-styrene-acrylate (ASA) 20, 26, 40f, 41,341–362 applications 355–358 base rubber 343–344 blends 352-355 market 342-343 production 343-347 properties 348-352 security and safety components 357 solar power and 356

active site species 378-381 additives 177-182, 672-674 AES (acrylonitrile-EPDM-styrene) 703, 716–720, 724 agitators 47, 53 alkyllithium-initiated polymers 502 alumina trihydrate 620, 700 aluminum trihydrate 691 amorphous polystyrene 411 see also atactic polystyrene amphiphilic polymers 125–126 ter/-amylperoxy-2-ethyl hexylcarbonate 170 anionic living polymerization 126 anionic PEP star polymers 497 anionic polymerization 23, 34–35, 51–52, 81–82, 113, 346, 471, 557 branched polystyrene 564 styrene block copolymers 147, 473 styrene-butadiene block copolymers 33, 152 vinyl monomers 501–502 weak links 88 anti-lumping agents 180 anti-static agents 180 antimony hydroxide 691 antimony oxide 691 antimony trihalide 691 antimony trioxide 673, 691, 693 antioxidants 88, 673

746

arborescent grafting 572 asymmetric radial polymer (ARPS) 258 atactic poly(cyclohexylethylene) (PCHE) 539–545 atactic polystyrene (APS) 52, 404, 431, 535, 617, 666 -/-relaxation 677 properties 37f sulfonated 458 see also amorphous polystyrene ATRP (atom transfer radical polymerization) 116–118 azobisisobutyronitrile 101 2,2'-azobis(isobutyronitrile) (AIBN) 155, 312, 313 backmixed reactors 57, 66, 70 barium sulfate 620 batch process (bulk suspension process) 268, 270 batch reactors 120 benzaldehyde 709 benzocyclobutene (BCB) 79, 562 benzoic acid 709, 714–715 benzoic anhydride 706, 709 benzoyl peroctoate 268 benzoyl peroxide 101, 313 bimodal molecular weight distribution 139–140 1,3-bis( 1 -phenylethenyl)benzene 467 bisphenol A-epichlorohydrin resins 470 1,2-bis(tetrabromophthalimide) ethane 673 bis(tribromophenoxy)ethane 699 bitumen, modification 492–493 blister packaging 507, 509 block copolymers 120, 418, 487, 676 dynamic properties 486f melt viscosity 478f preparation 159-161 SAM preparation 156-159 step-growth polymerization preparation 155 block molding 183-185 blowing agents 18, 166, 171, 177, 240, 241, 243 environment and 226-227

INDEX flammable 244 impregnation 170–172 rigid polystyrene foams and 203–231 blown film 514, 515 borate compounds 372–374, 389 branched polystyrenes 557-579 preparation 557-564 rheology 565–576 brittle fractures 634, 635–637 brominated epoxy oligomers 699 brominated flame retardants 700 brominated indans 693 bromine compounds 692–693 bubble nucleators 77, 206 building insulation 686, 700 bulk polymerizations 101, 305, 345–346 bulk suspension process 268, 270 Buna-N 19 Buna-S 6, 19 butadiene 153, 421, 467 1,4-butanediol 457 1,4-butanediol diglycidyl ether 600 /?-(3-butenylstyrene) 428 tert-buto\y radicals 558 p-tert butoxystyrene 536 butyl acrylate 346 tert-buiy\ methacrylate (tBMA) 474 tert-buiyl perbenzoate 100, 130 tert-buly\ peroctoate 268 tert-buiyl peroxybenzoate (tertBuPB) 168, 170 butylated hydroxytoluene (BHT) 266 n-butylbenzene 741 butyllithium 21, 34 di-s-butyllithium 466, 467 n-butyllithium 466, 467, 501–502 ^-butyllithium 466, 467 terf-butylperoxy-2-ethylhexyl carbonate (TBEC) 170 calcium carbonate 620 10-camphorsulfonic acid 134 can process 45, 46, 73 cast extrusion processes 238 cast film 514, 515 catalysts 85–87, 374–375, 539 Friedel-Crafts 83–85

INDEX metal 535 metallocene 473, 605, 606 palladium 473 titanium 378, 473 Ziegler-type 472, 473 cationic living polymerization 126 cationic polymerization 51, 85, 113, 137, 138, 145, 159, 161f cavity-wall insulation 223 cellular plastics 204, 207-213, 218, 223, 224 cellular polymers 204, 216, 217 applications 221–225 nomenclature 205 preparation 218 thermal conductivity 213–215 cellular polystyrene 203–204, 219 cellular poly(vinyl chloride) (PVC) 223, 224 cellular rubber 205, 223 Certifoam 219 char layers, formation 692 chlorinated alkanes 699 chlorine, aliphatic compounds 692 chloronaphthalene 446 chromophores 283–284 closed-cell foam 210 closed-cell polymers 205 co-catalysts 370–375 colorants 673 comb branched polymers 569-571 composition drift 57, 58, 158, 329 cone calorimeter 689 construction industry 188–190 continuous bulk free radical polymerization 80, 129 continuous bulk polymerization 73, 84, 147, 266, 269-270, 558-560 continuous extrusion process 204, 219 continuous mass polymerization 12, 13f continuous plug flow reactors 81 continuous solution process 46, 55, 268 continuous solution reactors 53 continuous stirred tank reactors (CSTR) 81, 94, 105, 267, 269, 325, 329 continuous tower process 45

747

copolymerization 94, 375–377, 468, 606–608 copolymers, production reactor requirements 57–58 core-shell impact modifiers 423–428, 588–596 coupling agents 470, 503 crack propagation 414, 646, 648 crazes 413, 414, 415, 420, 421, 425, 590f, 591, 634 crazing 589 creep 249, 666 Cross model 289, 614, 615t crosslinked poly(butyl acrylate) 594 crosslinking 309–310, 314–316, 672 crystal polystyrene 36, 508, 513, 515, 520–521 cushioning 197, 221 cyclohexane 503, 543 cyclohexene oxide 159 cyclohexyl acrylate 544 Cyclolac®T 20 decabromodiphenyl oxide (DBDPO) 673, 693, 694t, 700 degenerative transfer 118, 122, 124 devolatilizer design 75–76 di-tert-butyl peroxide 84, 268 dibenzoyl peroxide (BPO) 100,168 dibenzoylmethane 709 dibromoethane 470 dichlorodimethylsilane 470 dicumyl peroxide (DCP) 170, 268 1,3-diisopropenylbenzene 467 A^TV-dimethylstearylamine 423 2,4-diphenyl-l-butene 728, 742 m-l,2-diphenylcyclobutane 728, 742 trans-\,2-diphenylcyclobutane 728, 742 diphenylethylene 582–583 1,1-diphenylethylene 34, 447, 582 4,4'-diphenylmethane diisocyanate 457 1,3-diphenylpropane 728, 729f divinylbenzene 7, 50 m-divinylbenzene 471 ductile fractures 634, 639–643, 656 dunnage packaging 190–191, 197

748

Dylark® 21 dynamic loss modulus 665 dynamic mechanical properties 666 dynamic storage modulus 665 egg cartons 509 emulsion polymerization 30f, 50, 123, 124, 126, 423, 502 ABS manufacture 305–306, 326 ABS-type polymers 29 ASA production 343–345 mass polymerization vs 3If emulsion rubbery particles 594–596 endocrine disrupters 727–728, 742 environmental stress crack resistance 261-264 environmental stress cracking 350 EPDM (ethylene-propylenediene-monomer) 703, 717 ethyl tosylate 86, 87 ethyl-hexyl acrylate 346, 352 ethylbenzene 5, 149, 267, 520 ethylene glycidylmethacrylate 326 ethylene oxide gas 516 ethylene-1-octene 608 ethylene-octene copolymers 613 ethylene-propylene 428, 497, 717 ethylene-propylene diene modified (EPDM) rubber 257 ethylene-propylene rubber (EPR) 421, 428 ethylene-styrene copolymers 377, 605–630 ethylene-styrene interpolymers (ESI) 22, 438 attributes and applications 625-626 blends 617-620 filler composites 620-623 mechanical properties 613–614 melt rheology and processability 614–616 relaxation 610–611 structure-property relationships 608–616 terpolymers 623–625 ethylene-styrene pseudo-random copolymers 438

INDEX

ethylene-styrene-propylene (ESP) terpolymers 624 ethylene-vinyl aromatic monomers, copolymerizations 606–608 ethylenebis(tetrabromophthalimide) 693 5-ethylidene-2-norbornene 716 ethylmethylbenzene 379 expandable polystyrene (EPS) 166, 175, 190–197, 731 coatings 180–181 compressive strain 186, 187f conversion into foam 182–185 foam applications 188–190 manufacturing steps 167–168 maturation 183 mechanical properties 186–187 particle foam 165–201 physical properties 185–186 raw materials 166–181, 191 steam expansion 194–197 suspension polymerization 166–190 thermal conductivity 185–186 expanded foam, building insulation 686 expanded plastics 204 expanded polystyrene 197, 687, 693, 699 expanded polystyrene loose-fill packaging material 204 expanded rubber 205 expansion process 205-207 extruded polymeric foam 223 extruded polystyrene 204, 216, 218, 223, 239-245 extruded structural foams 220 extrusion 219, 271 Eyring's theory 659, 660 fabrication 271–275, 287–294 falling-dart impact 273 Farben, I. G. 45, 46, 268, 365 flame retardants 179–180, 673, 686, 692, 693, 694–698t flame retardation, mechanisms 690-692 flame-retardant polystyrene 685–702 flexible packaging 514–515 flow birefringence 293–294 foam insulation, energy loss 224–225

INDEX foam sheet extrusion processes, raw materials 243 foamed plastics 204, 205, 209 foamed polystyrene 13–17 foaming procedures 182–185 FoamulaR® 219 food packaging 509 food service items 508 food simulated solvents 731 footwear 493 fracture behavior 634 fracture energy, effects of loading rate 648–653 fractures 635-643, 656 free-radical polymerization 51, 81, 88, 99, 111, 266–267, 305, 345, 474 blowing agent 167 HIPS process 587 SAN copolymers 283 free-volume theory 98–99 fumaronitrile 333 garment hangers 511–513,520 gel effect 98–100 gel permeation chromatography 70, 140 general-purpose polystyrene (GPPS) 4-13, 105, 233, 247, 405, 583, 731 brittleness 25 elastomers and 256–261 resin degradation 265 glass fiber-reinforced SPS (GFSPS) 402, 405 glass transition temperatures 295, 352, 413, 483, 484, 533, 537, 538f, 541, 542, 624 graft copolymers, block and 547–550 graft model 313–314 grafting 309, 311–313 H-shaped polymer 569 halogen halides 690 halogen-based flame retardants, styrenic polymers 692, 694–698t heat exchangers 62, 64 heat release capacity 689 n-heptane 446, 742

749

hexabromocyclodecane (HBCD) 178, 179, 699 hexabromocyclododerane 693 hexahydropolystyrene 534 high styrene content styrene-butadiene copolymers 501–530 applications 507–519, 529 blends 520–528 history 501–502 properties 504-507 synthesis and manufacture 502-504 high-density polyethylene (HOPE) 224, 574, 611 high-impact polystyrene (HIPS) 56, 107, 224, 345, 411, 643–654, 660 applications 275–279 basic chemistry 256–266 block rubbers 310 can process 18 chemical resistance 405 effect of temperature 645–648 electrical properties 252, 25 3t environmental stress crack resistance 261–264 fabrication processes 273 food containers 731 fracture behavior 633–663 fracture energy and loading rate 654f gloss grades 264 impact data 644, 645f manufacture 266-271 mechanical properties control 679 modern commercial processes 268 oxidative degradation 265 plant 66, 69 process 587–588 Theological properties 253, 254t rubber and 54 rubber particle morphology 317, 318f rubber particle size (RPS) 260–261 rubber-modified 321 S-B block copolymers 153 SBC blends 525, 527t, 528 solvent resistance 253 styrenic blends 699 thermal and oxidative stability 264–266 thermal properties 252

750

high-impact polystyrene-poly(phenylene oxide) (HIPS-PPO) 686, 699 high-impact polystyrene-syndiotactic polystyrene (HIPS-SPS) 407 hydrogen halides 691 hydrogenated polystyrene, preparation and properties 533–555 hydroperoxides 707, 712, 722 IEC 60065 689 imidazolidone nitroxides 152 imides 326–330 impact behavior 351–352 impact loading 653 impact modification 417 impact modifiers 294, 418-423, 507, 509, 528 impact sound insulation 189 impact-modified polystyrene (IPS) 25, 29 impact-resistant polystyrene 26 initiation 94-97, 129, 133 initiators 100-101, 114, 131, 160f, 161f, 312 bifunctional 50, 70, 96–97, 101–105, 170 free radical 50 monofunctional 94–95, 101 tetrafunctional 103 injection molding 271, 520, 522 insulation materials 188, 189, 221, 222t, 223, 686, 700 internal friction 665 intumescence 692 ionic polymerization 51, 114 ionomers 678 •Y-irradiation 516–518 isopentane 171 isoprene 377, 421, 467, 482 isopropyl tosylate 86, 87 Af-isopropylmaleimide 327 isotactic poly(cyclohexylethylene) (PCHE) 545–546 isotactic polypropylene 365, 660 isotactic polystyrene 52, 147, 365, 377, 395-396, 431, 608 Izod impact strength 274, 353, 529

INDEX K-Resin R SBC 501, 502–504 K-Resin SBC 506t, 507 K-Resin® 21, 33 Kraton® 21, 33 lamellae, influence on crazes 420, 422f latex foam rubber 205 ir-ligands 367–369 ff-ligands 369–370 linear low-density polyethylene 574 linear polystyrenes, rheological properties 575 liquid crystal polymers 402, 403f lithium 467, 472 living anionic polymerization 466 living cationic polymerization 473 living free radical polymerization 111–128 living polymerization 111–128, 126, 465, 466, 502, 560 living radical polymerization 560 loose-fill expanders 194 low-density polyethylene 574 Luperox JWEB50 103, 105 MABS (methyl methacrylateacrylonitrile-butadienestyrene) 26, 38–41 magnesium hydroxide 691, 700 maleic anhydride 330–333, 494 mass polymerization 29, 30f, 3If, 305, 306 Mayo initiation mechanism 135,136f, 141 mechanical damping 665 metal hydroxides 691 metastyrol 3, 4 methacrylic acid 138, 678 methyl benzoate 470, 715 methyl methacrylate (MMA) 154, 354 methyl methacrylate-nBuA-methyl methacrylate (MMA-nBuA-MMA) 474 a-methyl-styrene 20 methylaluminoxane (MAO) 366, 370-372, 373f, 377, 385, 389 methylalumoxane 147 methylcyclohexylcarbinol 539 a-methylstyrene 323, 324 /7-methylstyrene (PMS) 378

INDEX a- methylstyreneacrylonitrile-acrylamide 355 a- methylstyrene-co-acrylonitrile 295, 325 microsuspension (MSP) rubber particles 588–594 microsuspension polymerization 347, 423 migration tests 736–737 molded expanded polystyrene foam (MEPS) 204 molded polystyrene 216, 223 molecular weight distribution 49, 139-140, 466 motor oil, viscosity index improvements 469 multigrade oils 496 nitriles, modified 333–334 nitroxide-mediated polymerization (NMP) 115, 147–162 nitroxide-mediated radical polymerization (NMRP) 148–152,474 norbornene 719 notched impact strengths 420 nucleation 205, 240 nucleation agents 177–179 Octabrom 693 octabromodiphenyl oxide 693 oil modifiers 496 olefin rubbers 421 olefinic elastomers 433 open-cell foam 210 open-cell polymers 204 organohalogen compounds 690 oriented polystyrene 233–239, 509 packaging, polystyrene 666 packaging materials 190 paintless film moulding (PFM) 357, 358f /jara-substituted polystyrenes, anelastic spectrum 674–676 particle size distribution 176–177, 316 peroxides 27, 79, 130, 133, 266 1 -phenyl-4-(r-phenylethyl)tetralin 728, 731,742 phenylacetylene 7

751

/?-phenylene diradicals 131, 133 phenylmagnesium bromide 582 TV-phenylmaleimide 323, 327 phenyltetralin 79 photooxidation 703, 704–709, 712–720, 724

photopolymerization 4 pickering stabilization, inorganic 175 pickering stabilizer systems 176, 177 plastic foams 204 creep 211–212, 213f flammability 225–226 moisture resistance 216–217 plasticizers 673 plug flow reactors 57, 106, 107 poly-p-phenylene 132 poly-p-xylylene 132 poly(2-methylcyclohexylethylene) 542 poly(a-methylstyrene-acrylonitrile) 355 polyamides 494 polybrominated diphenyl ethers 700 polybutadiene 25, 39, 260, 311, 348 polybutadiene-gr-SAN 712 polybutadienyllithiuin 153, 468, 470 polycarbonate 155, 156, 297, 350, 353, 494, 660, 699, 700 polycarbonate-poly(l, 4-butylene terephthalate) blends 354 poly(cyclohexylethylene) (PCHE) 542 applications 551–553 catalytic hydrogenation 534–539 characterization 539–547 copolymers 547–551 physical properties 546t random copolymers 547 surface energy 545 synthesis 533–534 poly(cyclohexylethylene)-PEP (PCHE-PEP) 548, 550 poly(cyclohexylethylene)-polyethylene (PCHE-PEE) 550 poly(cyclohexylethylene)po1y (ethylene-co-1 -butene) (PCHE-PEB) 548, 550, 551f poly(ethyl methacrylate) 660 polyethylene (PEE) 178, 377, 421, 548, 574, 660

752

poly(ethylene terephthalate) 660 poly(ethylene-co-l-butene) (PEB) 548 polyisobutylene 497 polyisoprene 496, 548 polymer-filler interactions 620 polymerization 27–28, 168–170, 269, 503 auto-initiation 133 high monomer conversion 80-82 inhibitors 5 initiation 94-97, 129, 133 kinetics 98 phase diagram 307 substituted styrenes 375–376 temperature control 46–48 polymers acoustic properties 217 amphiphilic 125–126 bubbles 205–207 cell structure 209–210 degradation and crosslinking 504 effects of structure 669–672 impact performance 635 impact resistance 433 linear thermal expansion 216 randomly branched 571–573 stabilization 77–78 thermal degradation 88 poly(methyl methacrylate) (PMMA) 34, 154, 357, 660 poly(n-butyl acrylate) 348 poly(N-n-alkylmaleimides) 326 poly(N-phenyl maleimides) 326 polyolefin thermoplastic rubbers 487 polyolefins 433, 494, 510 poly(p-fluorostyrene) 376 polyphenylene ether 26 poly(phenylene oxide) 494 polypropylene 224, 350, 730 polypropylene-(ethylene-propylene rubber) block copolymers 421 polystyrene 25–42, 218–219, 224, 574, 666–676, 699 acid mediation 133–139 applications 25, 129 blending SPS 405–407 blown films 234 catalysis and conditions 534–536

INDEX

chemical initiators 130-133 combs 570 degradation mechanism 265 devolatilization 59–65 dynamic mechanical spectroscopy 667–676 effects of additives 672 extensional rheology 573–576 flame-retardant 685–702 foamable 165–166 glass transition temperature 541 highMW 129–146 hydrogenation stages 538 impact properties 250, 252t impact resistance 633 low residual 73-92 manufacture 22, 45-72 modification 493-495 monomer regeneration 88–91 odor 520 para-substituted 674 phenyl ring substituents 670, 67If photooxidation 704–709 polymerization 66–69 process simulation and control 69–71 properties 35–37, 248–250 research approaches 75–83 rubber modification constraints 54-57 rubber toughening 633 rubber-modified 18, 36,47, 55,679,680 stereochemistry forms 431 structure and morphologies 29–35 weak links 88 polystyrene foams 165–166, 223, 224 physical properties 207–209 uses 245 polystyrene food containers, styrene dimers and trimers 727–744 polystyrene loose-fill packaging, physical properties 196–197 polystyrene paper 731 polystyrene polymers, energy dissipation 412–414 polystyrene production plant, reaction kettles 9f polystyrene resins 233, 234t poly(styrene-6/-vinylpyridine) 458

INDEX poly(styrene-acrylonitrile) (PSAN) 341, 345, 351 poly(styrene-b-vinylpyridine) 458 poly(styrene-b1—butadiene-bl-styrene) block copolymers 547 poly(styrene-b1 -isoprene-b1 -styrene) block copolymers 547–548 poly(styrene-butadiene-styrene) 487 poly(styrene-co- 4-hydroxystyrene) 536 poly(styrene-co-acrylonitrile) (SAN) 19, 709–712 poly(styrene-co-methyl acrylate) 535 poly(styrene-isoprene-styrene) 487 polystyrene-poly(vinyl methyl ether) (PS-PVME) 703, 704, 720–723, 724 poly( tert-butylcyclohexylethylene) 541 poly(tert-butylstyrene) 541 poly(tetramethylene glycol) 457 polyurethane 216, 223, 224 poly(vinyl chloride) (PVC) 223, 224, 281, 518 poly(vinyl methyl ether) (PVME) 446,703 PPE 438, 441 PPO 699, 700 propylene 623 PROXYL 28 Questra R 389 Questra® 22 radial polymers 470 radical polymerization, branched polystyrene 557–564 RAFT 118–119, 122, 123 reaction diffusion control model 99–100 reactors 46–58, 105–107 recycling, syndiotactic polystyrene 407 relaxation spectra 290 relaxation strength 681 a-relaxation 669, 670, 671, 676 p-relaxation 669 8-relaxation 670, 674 •Y-relaxation 669, 670, 674, 675 rheometers 287 rigid cellular polymers 210–211 rigid foams 205

753

rigid packaging 508–511 rigid plastic foams 223 rigid polystyrene foams, and alternative blowing agents 203–231 rubber 39, 205, 312–313, 415, 483 grafting 678 modification 417 particle sizing 308–310 rubber-modified styrenic polymers 686 rubber-modified syndiotactic polystyrene (SPS), impact behaviour 415–417 samarium compounds 367 Saytex 8010 693, 700 scavengers 78–80 sealants, adhesives and 489–492 SEES 418, 447, 471, 488, 492 semi-ductile fracture 637–639 SEPS 471, 492 shear 308 shear flows 287, 289–290 shear yielding 634 shrink film applications 515 Simon, Eduard 3 single-use containers 508 solution polymerizations 101, 305 solution process 268, 269 soundproofing 188 spherulite growth 444 sponge rubber 205 stabilization, particle formation and 172-177 stabilizers 504 star polymers 126, 471, 497, 566–569 steam stripping 65 steaming 184 Steiner Tunnel 687, 693 steric stabilization, mechanism 173–175 storax 203 strain energy density theory 650–651 strain hardening 292, 575, 576 stress relaxation 666 stripping materials 76–77 structural foams 205, 209t, 214t, 223-224 Styralloy® 18

754

styrene 203, 354, 377, 709 acrylonitrile monomer units 284 borate as co-catalysts 372 copolymers 21, 25–42, 676–678, 680 dienes copolymerization 377 ethylene copolymerization 377 fumaronitrile copolymers 21 grafting 258–259 homopolymerization 94–98, 100–107 isoprene copolymerization 468 living free radical polymerization 111–128 maleic anhydride copolymers 21 styrene dimers 727, 728 biological evaluation 737–742 structures 729f styrene monomers 715 styrene oligomers, endocrine-disrupting effects 737 styrene polymerization 27–28, 93–110, 378–385 acid-mediated model 140–145 effect of acid 136 kinetic models 70 styrene polymerization vessel, photograph 8f styrene polymers classes 26f notched impact strengths 420 styrene sulfonate 126 styrene trimers 727, 728 biological evaluation 737–742 cyclic 730f linear 729f styrene-1,1-diphenylethylene copolymers 447 styrene-acrylic copolymer 522 styrene-acrylonitrile (SAN) 26, 36, 224, 353–354, 523, 524t, 633, 679, 704 industry batteries 38f recycling 407 styrene-acrylonitrile (SAN) copolymers (ABS), rubber-modified polystyrene (HIPS) and 678–681 styrene-acrylonitrile (SAN) and EPDM (AES) blend, photooxidation 716–720

INDEX

styrene-acrylonitrile (SAN)-grafted EPDM rubber 354 styrene-acrylonitrile-methacrylate 326 styrene-6/-siloxanes 155 styrene-butadiene block copolymers 33, 152, 257

styrene-butadiene copolymers (SBCs) 21, 473, 506 adhesives 490, 492 applications 515–520, 529 blends 520–528 commercial 504 consumer goods 518 film 514 polypropylene and 525, 526t TEM of lammellar morphology 505f styrene-butadiene-styrene (SBS) 429, 476, 477, 488, 492 styrene-co-acrylonitrile (SAN), mechanical relaxation 677 styrene-co-acrylonitrile (SAN) copolymers 58, 306, 321, 328 acrylonitrile content 285 break points 293 brittle breaks 293 characterization 283-286 features and performance 281-303 glass transition temperatures 295 molecular weight 287–288,312 molecular weight distribution 285 multidimensional analyses 286 multiphase systems 294-296 production 282 sequence distributions 284-285 solid-phase behavior 296-297 strain hardening 292 viscosity 310 yellowing 283 styrene-co-acrylonitrile-co-fumaronitrile (SANF) 295, 333 styrene-co-maleic anhydride polymer 295 styrene-co-methacrylic acid (SMAA) 678 styrene-co-Af-phenylmaleimide 328 styrene-diene diblock copolymer 469 styrene-diphenylethylene copolymers 581–603

INDEX styrene-diphenylethylene (S/DPE) polymers blends 585–586 rubber modification 586–600 synthesis 582–583 thermal stability 585f thermoplastic elastomers 581 styrene-diphenylethylene (S/ DPE)-hydrogenated butadiene-S-DPE, triblock copolymers 599 styrene-divinylbenzene, copolymers 672 styrene-ethylene, copolymerization 377 styrene-(ethylene-butylene)-styrene (SEBS) 418, 447, 471, 488, 492 styrene-(ethylene-propylene)~styrene (SEPS) 488 styrene-hydrogenated butadiene-styrene, tri-block copolymers 596–599 styrene-hydrogenated polybutadiene block copolymers 424 styrene-hydrogenated rubber-styrene block copolymers 478 styrene-isoprene block copolymers 422 styrene-isoprene-styrene (SIS) 473, 476, 488, 490, 523 styrene-maleic anhydride (SMA) copolymers 330 styrene-maleic anhydride (SMA) resins, rubber-modified 332 styrene-methyl methacrylate (SMMA) 518, 521–523, 529 styrene-rubber-styrene block copolymers 418 styrene-sfatf-acrylonitrile 703, 724 styrenic block copolymer elastomers 465–500 applications 487–497 properties 474–487 synthesis 465-474 styrenic block copolymers 480, 487–489 styrenic polymers applications 125–126, 699 blends 699 halogen-based flame retardants 692 history of 3–24 photochemical degradation 703–725

755

styrenic structural foams, properties 209 styrenic thermoplastic elastomers 488 Styroflex R 33 Styroflex® 21 styrofoam 13–15, 204, 218 Styrolux R 33 Styrolux® 21 JV-substituted maleimides, synthesis 327 substituted styrenes 324–326, 375–376 2-sulfoethyl methacrylate 85, 138, 140 suspension polymerization 12, 73, 165, 266, 305, 559, 560 devolatilization 80 expandable polystyrene (EPS) 166–190 reaction kinetics 50 rubber-modified PS 18 syndiotactic poly(cyclohexylethylene) (PCHE) 546–547 syndiotactic polystyrene (SPS) 22, 52, 147, 322, 377, 428, 431, 535, 584, 607 applications 401-407, 429 bending strength 414, 415f blends 438-458 brittleness 412, 429, 433 catalytic systems 366–375 characterization 390–392 chemical resistance 404–405 commercial 401-407 crystal form 390-392 crystalization parameters 397t crystalization rate 396, 397f crystallization behavior 393–395 effect of rubber 445 immiscible blends 447–455 lamellae 417 mechanical properties 399–401,425 melt viscosity 397 niiscible blends 439–447 patent literature 433–438 physical properties 392–401 polymeric blends 431–461 properties 392–401, 425, 431–433, 438–458 rheological properties 397–399 rubber modification 411–430 solvent resistance 396

756 syndiotactic polystyrene (contd.) structure 390, 39If synthesis 365–387 US patents 434–437t syndiotactic polystyrene (SPS)/polyolefins blends 447 syndiotactic polystyrene (SPS)/ polyphenylene ether (PPE) blends crystalline polymorphism 439–440 crystallization growth rates 443–445 effect of rubber 445 isothermal crystallization 441, 442f, 442t miscibility 439 nonisothermal crystallization 440 phase structure and morphology 441-443 spherulite growth rate 443f syndiotactic polystyrene (SPS)/ polyurethane blends 457-458 syndiotactic polystyrene (SPS)/poly(vinyl methyl ether) (PVME) 443f, 445–446 syndiotactic polystyrene (SPS)/ styrene-1,1 -diphenylethylene (s-DPE) blends 447 syndiotactic polystyrene (SPS)/sulfonated atactic polystyrene (APS) 458 syndiotactic polystyrene (SPS)/TMPC blends 446–447 syndiotactic polystyrene/atactic polystyrene (SPS-APS) blends crystalline polymorphism 439–440 miscibility 439 morphology 443 nonisothermal crystallization 440 phase structure and morphology 441 sPS/polyphenylene ether (PPE) and 439–445 syndiotactic polystyrene/ ethylene-propylene rubber (SPS/ EPR) blends 456–457 syndiotactic polystyrene/ ethylene-propylene rubber/ styrene-(ethylene-butylene)styrene (SPS/EPR/SEBS) blends 456–457

INDEX

syndiotactic polystyrene/high-density polyethylene (SPS-HDPE) blends 448–455 syndiotactic polystyrene/high-density polyethylene/styrene-(ethylenebutylene)-styrene blends (SPS-HDPE-SEBS) 448–455 syndiotactic polystyrene/styrene(ethylene-butylene)-styrene (SPS-SEBS) 456 syndiotactic styrene-/>-methylstyrene copolymers (SPMS) 423, 456 sytrene block copolymers, nitroxide mediated polymerization 147–162 television enclosures 686 TEMPO (2,2,6,6-tetramethyl-lpiperidinyl oxide) 28,115,149,159 termination 97, 120 tertiary calcium phosphate (TCP) 175 tertiary hydroperoxides 716 tetrabromobisphenol A 693 tetrabromobisphenol A bis(allyl ether) 699 tetrahydrofuran (THF) 467, 472, 536 tetrapentafluorophenylborate 372 thermal copolymerization 94 thermal insulation 188, 221 thermoplastic elastomers 433, 469, 475, 523, 600-601 thermoplastic foams 220, 687 thermoplastics, bulk coloured 357 2,2'-thiobis(4-methyl-6-terfbutylphenoxy)titanium dichloride 377 thyroid hormone receptor binding assay 737 titanium compounds 367, 370 TMEDA (N, N, N', N'tetramethylethylenediamine) 471, 472 a-tocopherol 266 p-toluenesulfonic acid 86, 87 tower process 46, 204 tower reactors 106, 269 transition metal complexes 366-370 triisobutylaluminum (TIBA) 370, 385

INDEX

757

trimethylaluminum (TMA) 371 2,4,6-triphenyl-l-hexene 728, 742 triphenylcyclohexane 728, 730f tris(tribromophenyl)cyanurate 699

vinylcyclohexane 534, 539, 546, 547 vinylcyclohexene 534 4-vinylcyclohexene 547 viscosity index improvers 469, 496–497

UL 94 687–689 UL 94 HB 689, 700 a, 3-unsaturated hydroperoxides 715

wipe film evaporators 60–61, 75

vapor-phase mechanisms 690–691 Vicat softening point 321, 322 vinylbenzyltriethylammonium chloride 126

XAREC R 389 Xarec® 22 o-xylylene 79 p-xylylene 131, 133