Vibrational Spectroscopy of Biological and Polymeric Materials

Transreactions in Condensation Polymers Stoyko Fakirov (Ed.) Transreactions in Condensation Polymers Stoyko Fakirov (...

Author: Stoyko Fakirov

170 downloads 1647 Views 15MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

Transreactions in Condensation Polymers Stoyko Fakirov (Ed.)

Transreactions in Condensation Polymers Stoyko Fakirov (Ed.)

BWILEY-VCH Weinhehn .New York Chichester .Brisbane Singapore .Toronto

Prof. Dr. Stoyko Fakirov

Lab. Structure and Properties of Polymers University of Sofia 1 James Bourchier Blvd. 1126 Sofia Bulgaria

This book was carefully produced. Nevertheless, authors, editor, and publisher do not warrant the information contained therein to be free of errors Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data: applied for Deutsche Bibliothek Cataloguing-in-Publication Data: Transreactions in Condensation Polymers / Stoyko Fakirov (ed.) - Weinheim ;New York ;Chichester ;Brisbane ; Singapore ;Toronto :WILEY-VCH, 1999 ISBN 3-527-29790-1

0 WILEY-VCH Verlag GmbH, D-69469 Weinheim (Federal Republic of Germany), 1999

Printed an acid-free and chlorine-free paper. All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form - by photoprinting,microfilm,or any other means - nor transmitted or translated into machine language without written permission from the publishers Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Printing: Strauss Offsetdruck, D-69509 Morlenbach Bookbinding: Buchbinderei Osswald & Co., D-67433 Neustadt (WeinstraOe) Cover Design: Angelika Kilian Printed in the Federal Republic of Germany.

Contributors

Aerdts, A. M. Eindhoven University of Technology, Laboratory of Polymer Chemistry and Technology, P. 0. Box 513, 5600 MB Eindhoven, The Netherlands

BaltbCalleja, F. J. Institute of Structure of the Matter, CSIC, Serrano 119-123, 28006 Madrid, Spain

Berti, C. Department of Applied Chemistry and Materials Science, University of Bologna, Wale Risorgimento 2, 40136 Bologna, Italy Blackwell, J. Department of Macromolecular Science, Case Western Reserve University, Cleveland, OH 44106-7202, USA Denchev, Z. Faculty of Chemistry, Laboratory on Polymers, Sofia University, 1 James Bourchier Ave., 1126 Sofia, Bulgaria Devaux, J. Catholic University of Louvain, Faculty of Applied Sciences, Laboratory of Chemistry and Physics of High Polymers, Croix du Sud 1, B-1348 Louvain-la-Neuve, Belgium

Economy, J. Department of Materials Science and Engineering, University of Illinois at Urbana-Champaign, 1304 West Green Street, Urbana, IL 61801, USA

VI

Contributors

Eersels, K. L. L. Catholic University of Leuven (KULeuven), Department of Chemistry, Laboratory for Macromolecular Structural Chemistry, Celestijnenlaan 200F, B-3001 Heverlee, Belgium Fakirov, S. Faculty of Chemistry, Laboratory on Polymers, Sofia University, 1 James Bourchier Ave., 1126 Sofia, Bulgaria

Fiorini, M. Department of Management, &search Unit on Technology and Resources, University of Bologna, Piazza Scaravilli 2, 40126 Bologna, Italy F'rich, D. ARC0 Chemical Co., Technical Center, South Charleston, WV 25303, USA Groeninckx, G. Catholic University of Leuven (KULeuven), Department of Chemistry, Laboratory for Macromolecular Structural Chemistry, Celestijnenlaan 200F, B-3001 Heverlee, Belgium

James, N. R. Division of Polymer Chemistry, National Chemical Laboratory Pune 411008, India Kricheldorf, H. R. Institute for Technical and Macromolecular Chemistry, Hamburg University, Bundesstrasse 45, 20146 Hamburg, Germany Mahajan, S. Division of Polymer Chemistry, National Chemical Laboratory Pune 411008, India McCullagh, C. M. Department of Macromolecular Science, Case Western Reserve University, Cleveland, OH 44106-7202, USA

Contributors

VII

Montaudo, G . Department of Chemistry, University of Catania, Viale A. Doria 6, 95125 Catania, Italy

Pilati, F. Department of Chemistry, University of Modena, Via Campi 183, 41100 Modena, Italy Puglisi, C . Institute for Chemistry and Technology of Polymeric Materials, National Council of Research, Viale A. Doria 6, 95125 Catania, Italy

Samperi, F. Institute for Chemistry and Technology of Polymeric Materials, National Council of Research, Viale A. Doria 6, 95125 Catania, Italy Schneggenburger, L. A. Department of Materials Science and Engineering, University of Illinois at Urbana-Champaign, 1304 West Green Street, Urbana, IL 61801, USA

Sivaram, S . Division of Polymer Chemistry, National Chemical Laboratory, Pune 411008, India

Warth, H. Bayer AG, Plastics Business Group, D-41538 Dormagen, Germany Xanthos, M. Polymer Processing Institute, Hoboken, N J 07030, USA and NJ Institute of Technology, Department of Chemical Engineering, Chemistry and Environmental Science, Newark, NJ 07102, USA

CONTENTS Chapter 1

Interchange Reactions in Condensation Polymers and Their Analysis by NMR Spectroscopy H. R. Kricheldorf, Z. Denchev

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . .. 2. Nuclear magnetic resonance as an analytical tool (‘H, I3C, “N and 2gSiNMR) . . . , . . . . . . . .. . . . . . . . . . . . . . . . . . 2.1. Basics of the method . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . .. , . . 2.2. High resolution NMR of polymers . . .. . . . . .. . . .. . . . . . . . . .. . 3. Interchange reactions involving different functional groups . . . . . . 3.1. Reactions taking place in polyesters involving ester groups . 3.2. Reactions taking place in polyamides involving amine and amide groups . . . . . .. . . . . . . . . . . . . . . . .. .. .. . . . .. . . . . . . . . .... 3.3. Interchange reactions involving Si-0 bonds . . . . . . . . . . . . . . . . . 3.4. Interchange reactions involving urethane and urea groups . . 4. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . ... . . . .. . . . . .. ... . . . References . . . . . . . . .. . . . . . . . . . .. . . . . . . .. . . . . . . . . . .. . . . . . . . .. . . . .

.

Chapter 2

1 3 3 16 32 32 50 57 66

70 71

Effects of Catalysts in the Reactive Blending of Bisphenol A Polycarbonate with Poly (alkylene terephthalate)s F. Pilati, M. Fiorini, C. Berti

1. Introduction .... .... . _ . . . ................ . ..................... 1.1. An outlook on reactive blending of polyesters and polycarbonates ....... ..................................... 1.2. Bisphenol A polycarbonate: an overview .... .. ..... .. ...,.. 1.3. Poly(alky1ene terephthalate)s: an overview .. . ... . . . . . . . . 1.4. Blends of PC and poly(alky1ene terephtha1ate)s: a literature survey . . . . . . . . . . . . . . .. . . . .. . .. . . . . . . . . . . . . . ... .

. . .

79 79 80 82

83

X

Contents

2. Possible reactions occurring during melt-mixing of polyesters and P C .............................................. 2.1. Exchange reactions ........................................ 2.2. Side reactions ............................................. 2.3. Catalyst inhibitors ........................................ 3 . Evolution of the chemical structure during melt-mixing ......... 3.1. Approaches to the investigation of the resulting chemical structure .................................................. 3.2. Effects of catalysts in the reactive blending of PC/polyester systems ..................................... 4. Conclusions .................................................... References .....................................................

Chapter 3

85 86 90 91 92 93 100 120 121

Model Studies of Transreactions in Condensation Polymers J . Devaux

.

1 Introduction ................................................... 125 2 . Theoretical ..................................................... 127 2.1. Microstructure of copolycondensates from transreactions ... 127 133 2.2. Kinetics of transreaction ................................... 3. Application to the PC/PBT system ............................ 136 136 3.1. Microstructural study ..................................... 3.2. Kinetic study .............................................. 139 3.3. Mechanism of the PC/PBT transcondensation ............. 143 4 . General discussion and conclusion .............................. 155 References ..................................................... 157

Chapter 4

Copolymer Composition: a Key to the Mechanisms of Exchange in Reactive Polymer Blending G . Montaudo. C. Puglisi. F . Samperi

1. Introduction ................................................... 2 . Exchange mechanisms of reactive polymers in the melt .......... 161 3 . Exchange reactions occurring by inner-inner mechanism (Case 1) 165 3.1. Capped PBT/PC blends .................................. 165 170 3.2. Capped PET/PC blends ................................... 4 . Exchange reactions occurring by outer-inner mechanisms ....... 173

Contents

XI

4.1. PET/PTX blends (Case 2) ................................. 4.2. PET/PEA blends (Case3) ................................. 4.3. PBT/PC blends (Case4) .................................. 4.4. P E T/PC blends (Case4) .................................. 4.5. Nylon 6 /PC blends (Case5) ............................... 5. Conclusions .................................................... Appendix ...................................................... Monte Carlo modelling of exchange reactions .............. References .....................................................

Chapter 5

173 174 180 182 183 189 190 191

Interchain Transesterification Reactions in Copolyesters J . Economy. L . A . Schneggenburger. D. Rich

1. Introduction and background ................................... 2 . Synthesis and microstructure ................................... 3. Randomisation processes ....................................... 4. Sequence ordering .............................................. 5. Adhesive bonds in polyesters formed by ITR .................... 5.1. Liquid-crystalline copolyesters ............................. 5.2. I T R in thermosetting polyesters ........................... 6. Mechanism of adhesive bond formation ......................... References .....................................................

Chapter 6

195 196 201 205 209 209 211 214 216

Inhibition of Transreactions in Condensation Polymers N . R . James. S. S. Mahajan. S. Sivaram

1. Introduction ................................................... 219 1.1. Polymer blends ............................................ 219 1.2. Reactive compatibilisation ................................. 221 1.3. Transreaction during melt-blending ........................ 221 2. Control of transesterification in polyester blends ................ 226 226 2.1. Introduction ............................................... 2.2. Inhibitors for transreaction in polyester and 227 polycarbonate blends ...................................... 3. Methods of analysing transreactions in polymer blends .......... 233 3.1. IR spectroscopy ........................................... 233

XI1

Contents

3.2. NMR spectroscopy . . . . . . . . . . . .. .. . . .. . . . . . .. . . .. . . . . ... . . . 239 3.3. Differential scanning calorimetry . . . . . . . . . . . . . . . .. ... ... 250 3.4. Size-exclusion chromatography . . . . . . .. .,. .. . . .. . . . . .. . . . .. . 262 4. Conclusions ..... ........ ....................................... 263 References . . . . . . .. . . . .. . . . . . . . . . . . . .. . . . . . .. .. . . . . . . . . . .. .. . . . . 263

..

Chapter 7

..

Reactive Melt Processing of Aliphatic/ Aromatic Polyamide Blends: Effect on Molecular Structure, Semicrystalline Morphology and Thermal Properties K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

1. General introduction . .., ........ . . . .... . ....... . ....... . .... . . . 267 2. Influence of the processing conditions on the thermal behaviour of PA 46/PA 61 blends .. . . . . . . . . . .. . . . . . . . . . . . .. . . .. . . . . ,. . . .. . 269 2.1. Introduction . . . .. . . . . . . .. . . . .. . . . .. . . .. . . . .. . . . .... .. ... ... 269 2.2. Coprecipitation versus melt-mixing . . .. . . . . . . . .. . . . .. . . . . . 273 2.3. Influence of processing conditions .. . . . ... ... . . . . .. . . .. . .. . 275 3. Influence of the blend composition on the thermal behaviour of PA 46/PA 61 blends . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . ,. .. . ... . . .. 278 3.1. Crystallisation and melting behaviour ... . .. . . .. . . . ... . . .. .. 278 4. Molecular characterisation of PA 46/PA 61 blends by means of 13C NMR ...... ......... ........................ ..... ... ....... 281 4.1. Theoretical considerations . . . .... . . . . . . . . . . . . . . .. . .. . . . . . . . 283 4.2. Crystallisation behaviour of PA 46/PA 61 copolymers, prepared by melt-blending, as a function of the extrusion temperature, extrusion time, and blend composition . . . . . . 287 4.3. Molecular structure of PA 46/PA 61 copolymers, prepared by melt-blending, as a function of the extrusion temperature, extrusion time, and blend composition . ..... .. . .. .. .... . . 289 5. Characterisation of transamidation reactions in PA 46/PA 61 blends using gradient elution chromatography . . . . .. . . . .. . . . . .. . 293 5.1. Influence of melt-blending conditions on the degree of transamidation . . . . . .. . . .. . . ,. . . . . . . . . . . . . . . . . . . . . . .. . . . . . 295 5.2. Influence of end-groups on the degree of transamidation .. .. 301 6. Morphological structure of melt processed PA 46/PA 61 blends .. 302 6.1. Semicrystalline morphology of melt processed PA 46/PA 61 blends .............. ....................................... 302 6.2. Relation between the crystalline morphology and the melting behaviour of the PA 46/PA 61 copolymers ... . 311

.

.

.

.

.

.

.

.

.

Contents

XI11

7. General conclusions . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 312 References . .. . . . . .... . .. .... . . ... . .. . . . . .. ..... ... .. . .. .. . . .. .. 315

Chapter 8

Sequential Reordering in Condensation Copolymers S. Fakirov, Z. Denchev

1. Evidence of the occurrence of chemical interactions in blends of condensation polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 1.1. Evidence derived from the behaviour of the crystalline phase 321 1.2. Evidence derived from the behaviour of the amorphous phases ....... ... ..... . .... ... .. ... .. . .. . ..... . ....... ..... 324 1.3. Evidence derived from the behaviour of crystalline and amorphous phases . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . .. . . .. . . . . 324 1.4. Evidence derived from chromatographic methods . . .. . .. .. . 329 2. Melting-induced sequential reordering in condensation copolymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 2.1. Melting-induced sequential reordering in condensation copolymers obtained from blends of immiscible partners . . . . 333 2.2. Melting-induced sequential reordering in condensation copolymers obtained from blends of miscible partners . . . . . . 342 3. Crystallisation-induced sequential reordering in condensation copolymers . . .. . . . . . .. . . .. .. . . . . .. .. . . .... . . .. .. . . .. ... . . 347 3.1. Evidence of crystallisation-induced reordering derived from the crystalline phase behaviour . . . . . .. . . . . . . . . . . .. . .. . . . . . . 348 3.2. Evidence of crystallisation-induced reordering derived from the amorphous phase behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 4. Miscibility-induced sequential reordering in condensation copolymers obtained from miscible and immiscible partners . . . . . 364 4.1. Background . . . . . ... . . . . . . . . .. . . . . . . . . . . . . . . . . . . .. . . . . .. . . . 364 4.2. Experimental observations ... . . .. ... . .. ... ..... ... . . . .. ... . 365 4.3. Models and thermodynamic considerations . . . . . . . . .. . . . . . . . 369 5. Study of the sequential order in condensation copolymers by means of size exclusion chromatography after selective degradation 373 5.1. Basics of the SEC technique . ... . . . . . .. . . . . . . . . . . . . . . . . . . . . 374 5.2. Selective degradation of PC-containing condensation copolymers . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . .. . . . .. . . . .. . . . . . . 377 5.3. Sequence length determination in poly(ethy1ene terephthalate) - bisphenol A polycarbonate random copolymers as revealed by combined NMR and SEC studies 378 6. Conclusions .... ... ..... ... , . . ... , ............. ... ....... . ...... 385 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . ... . . . . . . . .. . . . . .. . 386

Contents

XIV

Chapter 9

X-ray Analysis of Transesteriflcation in Blends of Thermotropic Copolyesters J. Blackwell, C. M. McCullagh

Introduction .... ......... .................... .................. 391 Scattering by aperiodic polymer chains . . . .. . . . . .. .. . . , . .. . .. . . . 395 X-ray analysis of copolyester blends . . . . . . . . . .. . . ... . .. . . . .. . . . 396 Kinetics of transesterification . ... . .. . ... - .. . . ..... . ... . . ..... . 402 4.1. Random transesterification . . . . . . .. . . . .. . . . .. . . . . . . .. .. . ... 403 4.2. Transesterification kinetics ... . .. . . . . . . . . . . . .. . . .. . . . . ... .. . 405 5. Conclusions ..... ................ ............................... 408 References . . . .. . . .. . . . . . . , . . . . . . . .. . . .. . . .. .. .. . . . . . . .. . . . . .. 409

1. 2. 3. 4.

.

..

. .

Chapter 10

Effects of Transreactions on the Compatibility and Miscibility of Blends of Condensation Polymers M. Xanthos, H. Warth

. . .

1. Principles of blend compatibilisation . .. . . .. .. .... . .... . .. . , ... 411 2. Transreactions applied to blend compatibilisation . . .. . . .. . .. 412 3. Transreactions applied to specific binary blends . . . . . . .. . .. . . . .. . 416 3.1. Polyester/polyester blends . . .. . . . .. . . . . . . . . . . . . . .. .. . . . . . 416 3.2. Polyamide/polyamide blends . . . .. . . . . . . . ... . . .. . . . . . . . .. . . 422 3.3. Polyamide/polyester blends . .. . .. . . . ... . . . ..... ... . .... . . .. 423 References . . . .. . . . . . . . .... . . .. . . . . . . . . . . . . . .. .. . . . . . ..... . . . .. 424

.

.

Chapter 11 Effect of Transreactions and Additional Condensation on Structure Formation and Properties of Condensation Polymers F. J. Baltb Calleja, S. Fakirov, H. G. Zachmann 1. Relationship between interchain reactions and structure of condensation polymers . . . . . . . . . . ..... . . . . . . . . . . . . . .. . . . . . . . .. .. 429 1.1. Effect of interchain reactions on structure formation and properties of condensation polymers . . . ... . . ... .. . . . . . . . .. . 429 1.2. Structure formation in blends of condensation polymers with interchain reactions occurring to various extents . . . . . . . . . 433

. .

xv

Contents

1.3. Effect of polymer structure and morphology on chemical 443 interactions in condensation polymers ...................... 2 . Chemical interactions on the interfaces and interphases of condensation polymers ......................................... 445 2.1. Homochemical healing ..................................... 445 2.2. Heterochemical healing and healing with coupling agents ... 448 2.3. Chemical healing in crosslinked polyamides ................ 449 2.4. Transreactions at the phase boundary of semisolid blends of condensation polymers .................................... 450 2.5. Chemically released diffusion via transreactions in condensation polymers .................................... 452 3 . Effect of interchain reactions in condensation polymers on their mechanical properties ..................................... 453 3.1. Mechanical properties as revealed by tensile experiments ... 453 3.2. Mechanical properties as revealed by microhardness ........ 455 4 . Some practical aspects of the chemical interactions in 467 condensation polymers ......................................... 4.1. Copolycondensates resulting from solid-state additional condensation .............................................. 467 4.2. Copolycondensates resulting from transreactions in the melt ................................................... 467 4.3. Compatibilisation by means of interchange reactions ....... 468 4.4. Preparation of laminates from films of condensation polymers by means of interchain reactions ................. 468 4.5. Upgrading of molecular weight of condensation polymers by means of additional condensation in the solid state ......... 469 4.6. Recycling of condensation polymers by means of interchain 470 reactions .................................................. 4.7. Improvements of the finished-product properties ........... 472 References ..................................................... 474 Author index

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

481

Subject index

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

483

Preface

An inherent property of condensation polymers, in contrast to polyolefins, is their ability to react with each other. Regardless of the mechanism of the chemical reactions during polymer synthesis, the presence of groups of ester, amide, urethane, and other similar types, as well as carboxylic, amine, etc., groups in the macromolecules makes the post reactions possible. The practical importance of these reactions was first recognised by Flory, who got his patent in 1939 for molecular weight upgrading of polyamide 6 by means of solid state post-condensation. In his fundamental book, he analysed for the first time the constructive and destructive reactions in condensation polymers and their blends. Later, the attention of polymer chemists and physicists was focused on much more attractive problems of polymer science. For the chemists the additional condensation and transreactions seemed to be rather primitive, while physicists hardly knew about their existence, although they always took place during such a “purely” physical treatment as annealing at temperatures close to melting. These reactions attracted again the attention of polymer scientists by the start of intensive studies of polymer blends. It turned out that immiscibility and incompatibility could be overcome to a great extent by producing thin copolymer layers at the interface of blends of condensation polymers, mostly via transreactions. The ability of condensation polymers to undergo additional chemical reactions is fascinating. These reactions allow one (i) to prepare novel copolymers with desired composition and sequential order, as well as to enhance compatibility, (ii) to obtain more uniform polymers by minimising molecular weight fluctuations in a melt stream during polycondensation and processing and (iii) to provide for chemical healing of laminates of condensation polymers. A good friend of mine, who is a polymer physicist with world-wide reputation, does not believe that transreactions really exist. When, by chance, I told this to another common friend, who is a famous polymer chemist, his reaction was, “Tell him that Professor Flory, with whom he worked so many years, would be very unhappy if he could hear his statement”. As a matter of fact, this was the first very strong impulse for starting this project.

In the present book, the term condensation polymers is used as solely referring to the type of heterochain macromolecules regardless of the chemical mechanism and of the way of their preparation, e.g., by means of additional polymerisation or condensation polymerisation (polycondensation). Although attention is focused mainly on transreactions, additional condensation is also discussed in some chapters since both types of reactions take place simultaneously, particularly under vacuum. Chemical reactions of the destructive type are beyond the scope of this book. It should be mentioned also that the good intention of the Editor to introduce more or less uniform terminology, in the description of the reactions under consideration, failed because the preference of one or another term for the same reaction, expressed by some authors, was too strong. An international team of polymer chemists and physicists experienced in the field tried to cover the main topics related to the chemical interactions in condensation polymers. As Editor, I enjoyed the work with the individual contributors and gratefully appreciate their support, prompt response and patience. My thanks are also extended to the Institute for Structure of the Matter, Madrid, for the hospitality during my sabbatical tenure offered by DGICYT, Spain, where this project was finalised. I am greatly indebted to my coworker Mrs. S. Petrovich for her everyday help. Madrid, Dezember 1998

S. Fakarov Editor

SYMBOLS USED IN THIS BOOK Symbol

Definition

a,b A [A1

initial molar fractions of polycondensates 1 and 2 absorbance concentration of low molecular weight species degree of randomness term accounting for adsorption and other enthalpic processes in SEC column molar fraction of comonomer A or B in copolymer constant difference in specific heats between the liquid and glassy states of component i spacing of crystal planes diagonal length of indentation spacing of (hkl) planes polydispersity index degree of polymerisation molar fraction of homopolymers energy elastic modulus storage modulus loss modulus atomic scattering factor with Bj as the central unit molar fraction of triad molar fraction of BB or BN dyad SEC mobile phase volume flow rate force applied by indenter molar fraction of structural unit Ai molar fraction of an AiBj dyad in a polymer Fourier transform of cross-convolution of monomer j with monomer k frequency of nucleus n frequency of block length n crystal growth; modified Avrami parameter for microhardness

B

B CA, CB

C1

A%, d d d(hk1) D DP eHOM

E E E'

E"

f

fA,BjAk fBBi fBN

F

F

FAi

FA^ B~

Fjk

Fn

Fn

G

(z)

xx

Symbols used in this Book

Definition

H& HBL HPC

k k k, ~

A B

k2

kn

K K K

KS

KSEC

loss modulus probability of formation of a polymer having ncycunits Planck's constant microhardness microhardness (or hardness) of crystallised sample transition enthalpy reaction heat enthalpy of melting microhardness of quenched amorphous sample microhardness of fully amorphous powder enthalpy of crystallisation hardness of crystals enthalpy of crystallisation molar heat of fusion microhardness of fully crystallised powder local magnetic field experienced by a nucleus enthalpy of melting microhardness of minor and major components of blend maximum microhardness melting enthalpy of a perfect infinite crystal microhardness of coreactive PC/PET blend microhardness of homo PC molar heat of melting of crystalline units A at Tk static magnetic field rotating magnetic field orthogonal to Ho spin angular momentum quantum number signal intensity of A-A bond signal intensity of A-B bond signal intensity of B-B bond scattering along chain axis direction strength of scalar coupling between coupled nuclei with 2 intervening bonds Boltxmann's constant monomeric transesterification rate monomer ratio A/B in A-B copolymer transreaction rate constant reaction rate equilibrium constant geometrical factor Mark-Houwink coefficient of PS standard Mark-Houwink coefficient of sample ratio of average solute concentration inside pores to that outside the pores (SEC)

Symbols used in this Book

XXI

Definition lamellar thickness total number of units in sample long period or spacing (SAXS measurements) torque average length of block A number-average sequence length of an S unit number-average sequence length of a T unit magnetic quantum number average length of polymer unit, or aliphatic polymer unit parameter describing internal mobility of groups in a single chain molecular weight molar fraction of k in completely random copolymer of monomers j and k, modified for non-randomness if necessary number-average molecular weight number-average molecular weight weight-average molecular weight weight-average molecular weight z-average molecular weight repetition factor; n(x y) = degree of polymerisation average degree of polymerisation at time t average length of aromatic polymer unit degree of polymerisation number of ester linkages Avrami exponent for microhardness average degree of polymerisation at time 0 slopes of microhardness us. time curves number-average sequence lengths of comonomers A and B in copolymer number of units in a cycle population of upper and lower energy states number of times an oligomer A,Bn appears in simulated sample number of polymer reactive groups per unit volume number of AlBl units on polyamide 46 upper limit of p at infinite concentration probability of crystallisable homopolymer unit A being followed by another A unit molar fraction of monomer j probability of reaction at a given site probability of A, unit being followed by Aj unit, or Bj unit, in copolymer

+

Symbols used in this Book

XXII

Symbol

Definition

PS

molar fraction of sebacate probability of finding a T unit next to an S unit molar fraction of terephthalate element in probability matrix for sequence distribution of units x,y number of AzBz units in polyamide 46 point at which exchange takes place distance between nuclei transcondensation ratio cross reaction parameter gas constant molar fraction of comonomer A or B in sample (mixture of comonomers and copolymer) number of exchange events per initial molecule reaction entropy entropy increase for randomisation of initial copolymer blend or diblock entropy increase for randomisation of restored block copolymer extrusion time evolution period acquisition (detection) time annealing time crystallisation time time at which hardening kinetics change melt-pressing time SEC peak elution time temperature spin-lattice relaxation time (longitudinal relaxation time) spin-spin relaxation time (transverse relaxation time) annealing temperature Brill transition temperature Brill transition temperature during cooling from the melt crystallisation temperature crystal-nematic transition temperature glass transition temperature glass transition temperature of coreactive PC/PET blend glass transition temperature of component i glass transition temperature of homo P C healing temperature melting point melting points of less and more perfect crystallites in randomised equimolar PET/PC copolymer, respectively

PST

pr

Pb

S

AS AS AS''

Tb,c rn

TC,

Tm TA,Tg

Symbols used in this Book

XXIII

Definition equilibrium melting temperature of random copolymer equilibrium melting temperature of semicrystalline homopolymer thermodynamic equilibrium melting point interstitial liquid volume between SEC packing particles crystal-nematic transition temperature elution volume internal pore volume (SEC) vector in MOSES program SEC elution (retention) volume molar fraction of one coponent in copolyester blend average lengths of polymer sequences average lengths of polymer sequences degree of crystallinity weight fraction of co-oligomers weight fraction of homopolymers A or B mass fraction of component i total concentration of PEN units concentration of P E T units weight fraction of component i vector in MOSES program probability of block length n distance from chain ends number-average sequence length of AlBl groups number of transesterification events number-average sequence lengths of A1B and AzB fractional contribution of z- or y-mer to molar volume of X Y polymer average lengths of ethylene terephthalate and bisphenol A terephthalate sequences nominal concentration of starting material i concentration of starting material m of minor nominal concentration molar fraction of crystallisable units A concentration of component i in blend actual volume fraction of component i in blend molar fraction of A1Bz or AzBl dyads a t time t molar fraction of AlBz or AzBl dyads at equilibrium molar fraction of component i initial number-average degree of polymerisation of A or B distance along chain axis; atomic coordinate along the chain axis direction monomer j separation reciprocal space coordinate of z

XXIV

Symbols used in this Book

Symbol

Definition

a

fraction of minor component m Mark-Houwink coefficient of PS standard linear crystallinity, t , / L Mark-Houwink coefficient of sample magnetogyric ratio chemical shift cohesive energy density loss factor strain initial number of chain ends relative elongation at break viscosity intrinsic viscosity invariant in Chapter 7, Eq. (15) intrinsic viscosity angle between magnetic moment p and applied field Ho scattering angle draw ratio magnetic moment frequency of energy quantum absorbed or emitted by nucleus number of transesterication events per monomer precessional or Larmor frequency electron density of amorphous phase electron density of crystalline phase screening constant normal stress stress at break fold surface free energy degree of crystallinity degree of randomness percentage of transamidation

(Y

(YL a s

Y 6

62

tan 6 & EO

&b

rlint

8

28

x

P Y

X

&!

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 1

Interchange Reactions in Condensation Polymers and Their Analysis by NMR Spectroscopy

H. R. Kricheldorf, Z. Denchev

1. Introduction

Interchange reactions* are a phenomenon that concerns numerous classes of polymers. Recently, these interactions have been subject to extensive research due to the fact that they open the route to some new methods of polymer modification and even the preparation of novel polymer materials. Interchange reactions take place at elevated temperatures (most frequently in the melt) between functional groups belonging to molecules with different degrees of polymerisation or different chemical compositions. As a rule, they are reversible equilibrium interactions, typical of polycondensates, and have been recognised since these polymers were first made. Most prominent examples are polyesters and polyamides, where interchange reactions are best studied and understood. However, during recent decades, a number of publications have dealt with interchange reactions that involve urethane and urea groups, Si-0 bonds, etc.; these also deserve special attention. *There is a great variety of different terms used in the literature when addressing the interchange reactions, e.g., transreactions, transesterification, ester-ester interchange, etc. In this chapter, the general term “interchange reactions” is used consistently. It is classical English, widely accepted and highly versatile.

2

H. R. Kricheldorf, Z. Denchev

1 Interchange reactions I

Low molecular weight (monomer) systems resulting in monomer lproducts (model systems)

Molecular weight decrease (degradation)

High molecular weight (polymer) systems resulting in:

Molecular weight retention (copolymer formation)

Monomer systems but resulting in polymer products (equilibrium polycondensation)

Molecular weight increase (additional polycondensation)

Figure 1. Relationships between interchange reactions This chapter covers the characteristics of some significant types of interchange reactions, such as acidolysis, alcoholysis, aminolysis, esterolysis, taking place in low or high molecular weight systems and resulting in different products - low molecular weight compounds, homo- or copolymers. The scheme in Figure1 depicts the mutual connections and relations between all the types of interchange reaction. However, this classification is quite superficial: for instance, when discussing the interchange reactions in polymer systems, attention is focused on copolymer formation, although it is clear that, depending on the conditions of treatment and chemical composition of the blend constituents, the process should be accompanied by either degradation or additional polycondensation. These three processes are closely connected and should be considered as inevitable parts of the condensation equilibrium. It is worth noting that there had been some indications that interchange reactions might be possible in some carbochain polymers (ie., with allcarbon backbones). These also result in polymer modification, but occur to a much lesser degree. For this reason they are treated as secondary

NMR Analysis of Interchange Reactions in Condensation Polymers

3

reactions, taking place during the polyaddition [1,2],and are beyond the scope of this chapter. Very often it is of prime importance to discover the effect of the interchange reactions on the microstructure of the respective system - for instance, t o find out whether or not a copolymer is formed as a result of interchange reactions in monomer or polymer systems, or to determine the sequence length distribution, etc. High resolution nuclear magnetic resonance (NMR) has proved to be the most useful method for the direct experimental determination of the polymer microstructure. Of the two nuclei 'H and 13C, which possess spin and are common in synthetic polymers, 'H initially served as the spin probe in NMR polymer studies. However, though l H is more abundant than 13C, proton NMR spectra of synthetic polymers suffer from a narrow dispersion of chemical shifts and extensive 'H-'H spin coupling. 13C NMR, as currently practiced, does not suffer from these difficulties, of which the latter has recently been turned to advantage for 'H NMR by 2D techniques. The advent of proton-decoupled spectra recorded in Fourier-transform mode has quickly made 13C NMR spectroscopy the method of choice for determining polymer microstructures. Other methods, such as 15N and 29Si NMR, are rapidly gaining importance as irreplaceable tools for the characterisation of siloxanes and N-containing polycondensates. For all these reasons, the basic principles and importance of modern NMR techniques in view of their application for interchange reaction characterisation are discussed in this chapter. 2. Nuclear magnetic resonance as an analytical tool (lH, 13C, 15N and 29Si NMR) 2.1. Basics of the method NMR spectroscopy belongs among the radiospectroscopic methods, where the basic transitions are those between spin (or magnetic) energy levels of the nuclei. In contrast to the optical transitions (e.g., vibrational, rotational, electronic), the nucleus can absorb radiofrequencies only if the molecules are placed in a strong, external magnetic field. This is because in the absence of magnetic field, the different spin states of the nuclei have the same energy, i e . , they are degenerate. 2.1.1. Magnetic properties of the nucleus While the nuclei of all atoms possess charge and mass, not every nucleus has angular momentum and a magnetic moment. Nuclei with odd mass numbers have spin angular momentum quantum numbers I, with values that are odd-integral multiples of 1/2. Nuclei with even mass numbers are spinless if their nuclear charge is even, and have integral spin I if their nuclear charge is odd.

4

H. R. Kricheldorf, Z. Denchev

The angular momentum of a nucleus with spin I is simply I(h/2n), where h is Planck's constant. If I # 0, the nucleus will possess a magnetic moment, p, which is taken parallel to the angular-momentum vector. A set of magnetic quantum numbers m, given by the series

m=I, I-1,1-2

,... ,-I

(1)

describes the values of the magnetic moment vector which are permitted along any chosen axis. For nuclei of interest here (IH, 13C, 15N, IgF, 29Si, 31P), I = 1/2, and thus m = +1/2 and -1/2. In general, there are 21 + 1 possible orientations of p, or magnetic states of the nucleus. The ratio of the magnetic moment and the angular momentum is called the magnetogyric ratio, y: y = 2np/hI

(2)

and is characteristic of a given nucleus. The nuclei commonly observed in NMR studies of polymers usually have spin I = 1/2, and are characterised by 21 + 1 = 2 magnetic states, m = +1/2 and -1/2. Both nuclear magnetic states have the same energy in the absence of a magnetic field, but they correspond to states of different potential energy upon application of a uniform magnetic field Ho. The magnetic moment p is either aligned along (m = +1/2) or against (m = -1/2) the field W O with , the latter state corresponding to a higher energy. Detection of the transitions of the magnetic nuclei between these spin states [m = +1/2 (parallel), m = -1/2 (antiparallel)] are made possible by the NMR phenomenon. Table 1. Magnetic characteristics of some atomic nuclei [3] Nucleus Natural Atomic Magneto- Magnetic abundance number gyric moment p I ratio? (magne(%I (rad.s.Oe) tons) 'H' 99.98 112 26753 2.79270 1 4 107 0.85738 2 ~ 1 ( ~ 0.016 ) llg5 81.17 3/2 8583 2.6880 lZc6 98.89 0 13c6 0.70216 1.11 1/2 6728 1 4 ~ 7 0.40357 1 1934 99.64 1 5 ~ 7 -0.28304 0.36 1/2 -2712 1608 0 99.76 1708 0.037 5/2 -3628 -1.8930 19F9 100 1/2 25179 2.6278 2Ssi14 92.28 0 29si14 -0.55477 4.67 112 -5319 3 1 ~ 1 5 100 1.1305 112 10840 32~16 95.06 0 33~16 0.64274 0.74 312 2054

Quadropole Relative Resonance moment Q amplitude frequency cm2) of the (MHz) signal 1.000 100 0.010 15.4 0.00274 0.165 32.2 0.0355 0.016 25.1 0.001 7.2 0.02 0.001 10.1 0.029 13.5 -0.004 0.834 94.0

-

-0.064

0.078 0.066

19.9 40.5

0.002

7.67

NMR Analysis of Interchange Reactions in Condensation Polymers

5

The magnetic characteristics of some atomic nuclei which are of interest in organic chemistry and more or less appropriate for experimental NMR studies, are given in Table 1. It is seen that the hydrogen nucleus (the proton) combines all the properties that are favourable for NMR analysis: a spin number of 1/2 (ie., lack of quadrupole momentum), high natural abundance of the isotope and the largest magnetic moment. All these properties, together with the presence of hydrogen in the majority of organic compounds, explain the exceptional role and significance of this nucleus in NMR spectroscopy. After the protons, the nuclei of fluorine and phosphorus should be mentioned as convenient objects for NMR studies. It is seen in Table 1 that the most widespread isotopes of elements that are significant in organic chemistry, such as 12C, l60,28Si and 32S,cannot be studied by the NMR technique (I = 0). The 14N investigations are strongly hampered by the presence of a quadrupole momentum. For these reasons 13C, 15N and 29Si are most appropriate for NMR studies despite the fact that their natural abundance is low. 2.1.2. Resonance Let us discuss the interactions of magnetic fields applied to the magnetic moments of nuclei with spin I = 1/2. Figure2 is a schematic of the nuclear magnetic moment p in the presence of an applied magnetic field Ho, acting along the z-axis of the coordinate system. The angle B between the magnetic moment and the applied field does not change, because the torque

Figure 2. Nuclear magnetic moment in a magnetic field [4]

H. R. Kricheldorf, Z. Denchev

6

tending to tip p toward Ho is exactly balanced by the spinning of the magnetic moment, resulting in nuclear precession about the z-axis. Increasing Ho, in an attempt to force the alignment of p along the z-axis, results in faster precession only. A good analogy is provided by the precession of a spinning top in the Earth's gravitational field. The precessional or Larmor frequency, VO, of the spinning nucleus is given by 7 vo = -Ho 27T

(4)

and is independent of 8. However, the energy of the spin system does depend on the angle between p and Ho:

E

= -p.

H~ = -

p

case ~

~

(5)

We may change the orientation 8 between p and H O by application of a weak rotating magnetic field W1 orthogonal to Ho (Figure2). Now p will experience the combined effects of H1 and Ho if the angular frequency of H1 coincides with V O , the precessional frequency of the spin. The nucleus absorbs energy from H I in this situation and 8 changes; otherwise H I and p would not remain in phase and no energy would be transferred between them. If the rotational rate of H1 is varied through the Larmor frequency of the nucleus, a resonance condition is achieved, accompanied by a transfer of energy from H1 to the spinning nucleus and an oscillation of the angle 0 between Ho and p. At HO = 2.34 T (1T = 1 tesla= 10 kilogauss), the resonant frequencies of the ' H , 19F, 31P, 13C, 29Si,and 15N nuclei are vo = 100, 94, 40.5, 25.1, 19.9, and 10.1 MHz, respectively [4].

2.1.3. Interactions and relasations of nuclear spins Figure3 illustrates the magnetic energy levels for a spin -1/2 nucleus in a magnetic field Ho.The energy distribution between nuclear spin states is

AE = 2pHo

(6)

and the relative populations of the upper (+) and lower (-) states is given by the Boltzmann expression

N+

N- = exp

AE 2PHO ( - m) = exp (-

The excess population of the lower energy state is

7

NMR Analysis of Interchange Reactions in Condensation Polymers

where the approximation e-” = 1 - x, for small z,has been adopted. Figure 3 shows the creation of two energy levels when a nucleus is placed in a strong external magnetic field. This effect is called Zeeman’s splitting of energy levels. If one considers the well-known expression

AE = hu

(9)

where u is the frequency of the energy quantum absorbed or emitted by the nucleus, one obtains

The last expression is a basic relationship in NMR, showing that for a nucleus of a given type, characterised by a magnetogyric ratio y, the resonant frequency u is proportional to the applied external magnetic field

HO .

At a field strength Ho = 2.34 T, the difference between magnetic energy levels for proton nuclei is N call which results in an excess populaspins of lower energy. For an assemblage of nuclei, this tion of 2 x small spin population difference leads to a correspondingly small macroscopic moment directed along No.Removal of Ho results in a loss of the macroscopic moment , because the magnetic energy levels are degenerate in the absence of the field [4]. At resonance, the nuclei pass to upper energy levels by energy absorption or vice versa, to lower energy levels by energy emission. Since the absorptive transitions are prevailing, a tendency exists toward equalisation of the populations of the levels with time. When such a state (called saturation) is achieved, the energy absorbed by the sample becomes equal to that emitted and the NMR signal disappears. However, under appropriate experimental conditions, the NMR observations could be infinitely long. This is due to the fact that emissionless processes (called relaxation) are

-

Ho = 0

Ho > 0

E

m

Figure 3. Energy levels for a spin -1/2 nucleus in a magnetic field HO [5]

8

H. R. Kricheldorf, Z. Denchev

taking place in the sample; as a result the energy absorbed by the nuclei decreases and the system is kept at a state of Boltzmann equilibrium. The question arises, what mechanisms are responsible for relaxing upper-level spins to the lower level after application of Ho, thereby maintaining parity between the spin and sample temperatures? Such a relaxation is possible because each spin is not completely isolated from the rest of the molecules in the sample, called the lattice. The spins and the lattice may be considered to be separate coexisting systems which are weakly coupled through an inefficient yet very important link, by which thermal energy may be exchanged. The molecular motions of the neighbouring nuclei, which constitute the lattice, provide the mechanism for transferring thermal energy between the spins and their surroundings. The relative motions of neighbouring nuclei generate fluctuating magnetic fields which are experienced by the observed nucleus as it precesses about the direction of the applied field H o . A broad range of frequencies will be associated with the fluctuating fields produced by the lattice motions, because these motions are almost random in respect to the observed nucleus. Components of the fluctuating magnetic fields, generated by the lattice motions, which lie along Ho (Figure2) and have frequency vo will, like H I , induce transitions between the magnetic energy levels of the observed nuclei. The rates of this spin-lattice relaxation must therefore be directly connected to the rates of molecular motions in the lattice. The spin-lattice relaxation time, T I ,is the time required for the difference between the excess and equilibrium spin populations to be relaxed by a factor of e. For liquids, 21' is usually in the range of 10-2-102 s, while in solid samples 2'1 may be as long as hours. Spin-lattice relaxation re-

Ho

Figure 4. A pair of precessing nuclear moments with static (a) and rotating (b) component (41

9

NMR Analysis of Interchange Reactions in Condensation Polymers

sults in a change in energy via a redistribution of magnetic moments with components along the applied field Ho.As a result, TI is often termed the longitudinal relaxation time: it is associated with a decay of the macroscopic nuclear moment along the direction of the applied field Ho (the z-direction; see Figure 2). There is a second mode by which nuclear magnetic moments may interact. This interaction is illustrated in Figure4. Here a pair of nuclear moments are precessing about the Ho-axis, and each is decomposed into a static component along Ho ( a ) and a rotating component in the xy plane ( b ) . If the rotating component precesses at the Larmor frequency YO, a neighbouring nucleus may be induced to undergo a spin transition, resulting in a spin exchange. No net change in the total energy is produced by the exchange of neighbouring nuclear spins, but the lifetimes of the interacting spins are clearly affected. This exchange of neighbouring nuclear spins is called spin-spin relaxation and is characterised by Tz, the spin-spin relaxation time. Tz is also called the transverse relaxation time, because it is related to the rate of change of magnetization in the xy plane, which is transverse to the Ho field direction. 2.1.4. Chemical shift

We have seen that by application of a rotating magnetic field HI t r a n s verse to the static field Ho, about which a spinning nuclear magnet is precessing, we can flip the nuclear spin by rotating HI at the precession or Larmor frequency Y O . If all nuclei of the same type, e.g., all protons were to resonate at the same field strength Ho, NMR would not be a spectroscopic tool useful for the study of molecular structures. Fortunately, soon after the application of NMR to condensed phases it was observed that the characteristic resonant frequency of a nucleus depends on its chemical or structural environment. The cloud of electrons about each nucleus produces orbital currents when placed in a magnetic field Ho. These currents produce small local magnetic fields which are proportional to Ho but are opposite in direction, thereby effectively shielding the nucleus from Ho. Consequently, a slightly higher value of Ho is needed to achieve resonance. The actual local field, HI,,, experienced by a nucleus, can be expressed as Hloc = HO(1 - 0)

(12) where 0 is the screening constant. It is highly sensitive to chemical structure but independent of Ho. The resonant Larmor frequency becomes

and the difference between magnetic energy levels is now (see Eq. (6))

A E = 2pHloC = 2pHo(l-

0)

(14)

10

H. R. Kricheidorf, 2. Denchev

Nuclear screening decreases the spacing of nuclear magnetic energy levels. An increase in the magnetic shielding requires an increase in Ho at constant vo and a decrease in vo at constant Ho to achieve resonance. Nuclear shielding is influenced by the number and types of atoms and groups attached to or near the observed nucleus. The dependence of u on the molecular structure is of major significance to NMR as a tool in the study of molecular structures. There is no natural fundamental scale unit in NMR spectroscopy. Both the energies of transition between spin quantum levels and the nuclear shielding produced by the screening constant u are proportional to the applied field Ho. In addition, there is no natural zero of reference in NMR. These difficulties are overcome by (i) expressing the resonant frequencies of nuclei in parts per million (ppm) relative changes in Ho and (ii) referring the observed changes or displacements in resonance, called chemical shifts, to the ppm relative change in the resonant frequency of an arbitrary reference substance added to the sample. In lH and 13C NMR spectroscopy, it is customary to use tetramethylsilane (TMS) as the reference compound, where the chemical shifts 6 of both the 'H and I3C nuclei are taken as b = 0 ppm. It is beyond the scope of this chapter to discuss the factors affecting the chemical shift. For this reason, we could encourage the reader to consult the general NMR texts [5-81. However, it is worth mentioning here the most important factors, namely: electronegativity of the substituents, magnetic anisotropy, electric field effect, hydrogen bond formation, repulsive electron effects. The chemical shifts of carbon-bonded protons are almost insensitive to temperature and sample concentrations. However, protons bonded to heteroatoms (e.g., 0-H, N-H and S-H) display strong dependence on the above factors. These effects are caused mostly by the tendency of OH, SH and NH protons to form hydrogen bonds. In general, this leads to screening effects, i e . , to downfield shifts, although the reasons are not always well understood. It is always strongly recommended to check the possibility of hydrogen bond formation between the sample and the solvent. For instance, dimethyl sulfoxide (DMSO) forms strong hydrogen bonds with OH and NH groups, resulting in significant changes in the chemical shifts.

2.1.5. Spin-spin coupling

It is well established that in numerous cases the NMR signals are split into several components. This is illustrated in Figure5, where the proton spectrum of 1,1,1,2,3,3-hexachloropropaneis depicted. It is seen that the signals of the two protons are split into doublets. This observed fact is due to the so-called spin-spin interaction. It is established that the magnetic moments of the nuclei interact both

NMR Analysis of Interchange Reactions in Condensation Polymers

-CHC12

7

11

-CHCl-

6

5

6, PPm

Figure 5. NMR spectrum of CCbCHClCHClz in solution [3]

directly (through the space) and indirectly - by the valence electrons. The former effect is called direct spin-spin (or dipole) coupling and does not influence the NMR spectra taken in liquid or gaseous media since it is nullified due to the fast molecular motion. In the case of solids, the direct spin-spin interaction leads to a quite significant broadening of the resonance signals. From now on, we will discuss only the indirect (or scalar) coupling. It is observed in the NMR spectra of liquids and gases, and is of prime importance in their interpretation. Two neighbouring nuclear spins will feel, in addition to H o , the local magnetic field Hloc they are producing. Ellocis given by

H ~ ,= , ~* , ~ r - ~ ( 3 c oes ~ 1)

(15)

where r is the distance between nuclei, and 8 is the angle between H O and the line joining the nuclei (see Figure4). The fact that HlOc may add t o or subtract from Ho, depending on whether the neighbouring magnetic dipole is aligned with or against Ho, is reflected by the f sign. This type of spin-spin coupling is called dipolar coupling and serves to broaden the resonance line of a nucleus. There are two important situations where dipolar coupling does not contribute t o line broadening. The first is when all neighbouring nuclei are rigidly oriented at the magic angle of 8 = 54.7O, where cos26 = 113 and HlOc = 0 (see Eq. (15)). If the relative orientations of neighbouring spins vary rapidly with respect to the time a nucleus spends in a given spin state, i e . , with respect to the spin-spin relaxation time T2, than HlOcis given by its space average,

Hloc= ,w-~J(3 cos2 0 - 1)sin Ode

(16)

0

which also vanishes. Both these circumstances are important for observing high-resolution NMR spectra of polymers. Nuclear spins may also be coupled by orbital motions of their valence electrons or polarisation of their spins occurring indirectly through the intervening chemical bonds. Unlike dipolar coupling of nuclear spins, this

H. R. Kricheldorf, Z. Denchev

12

(b)

(8) trans

gauche

I

I

H

H

H

'J w 12 Hz

3J

FY

2 Hz

Figure 6 . Vicinal J IH-'H scalar coupling in a saturated hydrocarbon [4] indirect or scalar coupling is not affected by molecular tumbling and is also independent of Ho. Two spin -1/2 nuclei so coupled will each split the other's resonance into a doublet, because in a large collection of such nuclear pairs the probabilities of each finding the other's spin along (+1/2) or against (-1/2) HO are almost equal. If one nucleus of the pair is further coupled to a second group of two identical nuclei with +, - (- +), and - - spin orientations, then the resonance of the first nucleus will appear as a 1:2:1 triplet. The resonance of the identical pair will be a doublet. A single nucleus coupled to three equivalent neighbouring spins with -, - (+ - -, - -, - - +); and - - - orientations would exhibit a 1:3:3:1 quartet resonance. A spin -1/2 nucleus with n equivalently coupled neighbours also of spin 1/2 will have its resonance split into n 1 peaks. In the NMR spectra of polymers, only 'H-'H, 13C--'H, 13C-19F, 15N'H, lgF-lgF, 19F-'H, 29Si-1H, and 31P-1H scalar couplings are important. The magnitude and sign of the scalar coupling of two magnetic nuclei depend on substituents and geometry. The strength of the coupling in Hertz is designated J , where the superscript z denotes the number of intervening chemical bonds between the coupled nuclei. A particularly useful relation is based on the observed geometry-dependent vicinal 'H-'H coupling J illustrated in Figure6. Here it is observed that when the vicinal protons are trans (a) the scalar coupling is large (about 12 Hz), but it is markedly reduced to about 2 Hz in their gauche arrangement (b).

+ +

+ + + +, + +

+

+ + +;

+

2.1.6. Ezperimental observation of N M R In an NMR experiment, one investigates an assemblage of nuclei of a given type (e.g., protons) rather than the separate nucleus. In general, nuclei always take the level (or the state) of lower energy. In the case of nuclei with J = 1/2, this is the state with m = f 1 / 2 (see Figure3). However, the molecular thermal motion opposes this trend since its energy is several

NMR Analysis of Interchange Reactions in Condensation Polymers

13

NMR sample tube , Radio-frequency coil Tuned radio-frequency circuit

Signal to NMR system’s . electronics

Figure 7. Cross-section of a superconducting NMR magnet IS]

times higher. As a result, an equilibrium state is reached, characterised by the Boltzmann expression (Eq. (7)). The excess population of the lower energy state (Eq. (8)) is rather small. For magnetic fields between 2.0 and 2.5 T at room temperature, calculations according to Eq. (8) show that the population of spins at the lower energy level is higher by only 0.001% than that at the higher level. Nevertheless, it is precisely this minor difference that makes absorption transitions more probable than the emission ones. As a result, the sample absorbs external energy at its Larmor frequency. The N M R experiment is to register this absorption with suitable inst rumentation. Figure7 is a drawing of a superconducting magnet used in a modern high-field N M R spectrometer. The magnet is placed in a liquid helium bath to maintain its superconductivity. The radio-frequency (rf) coil provides the energy appropriate to excite the nuclei in the sample to resonance. The degeneracy of the nuclear magnetic spin energy levels is removed by the static magnetic field H o . Application of the rotating magnetic or electromagnetic field H I excites transitions between these energy levels. When the frequency of the H I field (rf, in MHz) is equal to the Larmor frequency of the observed nucleus, the resonance condition occurs, i. e.,

Most samples will have multiple Larmor frequencies, because most

14

H. R. Kricheldorf, Z. Denchev

molecules have more than a single magnetically equivalent group (e.g., CH, CH2, CH3), leading to several resonance frequencies or chemical shifts. The method used to excite the nuclei and achieve resonance must clearly be capable of covering all of the Larmor frequencies in the sample. Two principal methods have been developed to achieve the resonance condition in NMR spectroscopy - continuous wave (CW) and Fourier transform (FT). In the CW method each magnetically equivalent nucleus is successively made to resonate by sweeping either of the magnetic fields, the rf H I or the static Ho field. As each nucleus is brought into resonance by the field-sweeping process, a voltage is induced in the rf pick-up coil (see Figure7). After amplification, this signal is detected directly in the frequency domain and recorded in a plot of voltage (intensity) us. frequency. In contrast, the FT method employs signal detection in the time domain, followed by a Fourier transformation into the frequency domain. Simultaneous excitation of all the Larmor frequencies is provided by applying a pulse (short burst) of rf signal at or near vo, resulting in an equalisation of the populations of the nuclear spin energy levels. Equilibrium spin p o p ulations are reestablished in a free induction decay (FID)process following the rf pulse. The vector diagram in Figure8 can be used to visualise the

Figure 8. Pulsed NMR experiment in the rotating frame. (a) Net magnetisation Mo along Ho; (b,c) rf field H I applied perpendicular to No for a duration sufficient to tip Mo by 90" into the x'y' plane; (d,e) spins begin to relax in the x'y' plane by spin-spin ("2) processes, and in the z'-direction by spin-lattice (TI) processes; (f) equilibrium M Ois reestablished along Ho [4]

NMR Analysis of Interchange Reactions in Condensation Polymers

15

effect of the rf pulse ( H I )on the nuclear spins and their subsequent FID to equilibrium. At equilibrium in the presence of Ho, more spins will be aligned along Ho than against it, and this is indicated by the net magnetic moment Mo, drawn along the field direction z’ in Figure 8a (note that the primes indicate that the reference frame z‘y’z’ is rotating at the Larmor frequency). The net magnetisation has been tipped 90” into the d y ’ plane (Figure8b,c) by application of an rf pulse H1 with duration that is just sufficient to equalise the magnetic energy levels, ie., Mo = 0 along z’. Following the rf pulse (Figure 8d,e), the spins begin to reestablish their initial state through TIand T2 relaxation processes. Spin-spin interactions in the transverse (z’y‘)plane cause dephasing of the spins in this plane (Tzprocess), while spin-lattice interactions cause the spins to relax along the 2’-direction (TIprocess). Usually many signals must be accumulated before a spectrum with adequate signal-to-noise ratio can be obtained, particularly for nuclei with low natural abundance, such as 13C, 15N and ”Si. The pulse repetition rate is governed by the relaxation time TI. Figure 9 gives a pulse sequence representation of the vector diagram in Figure& The detected signal, or FID, is obtained as a voltage in the

v

Time

-

Figure 9. Repetitive pulse sequence (a) and Fourier transformation of the timedomain FID into the frequency-domain NMR signal (b) [4]

16

H. R. Kricheldorf, Z. Denchev

Plotter

Computer

conversion to

Figure 10. Block diagram of a pulsed FT Nh4.R spectrometer [6] time domain. The pulse is sent out repeatedly to improve the signal-tonoise ratio, and the delay time between pulses must be long enough for 21' relaxation processes to complete. Fourier transformation of the timedomain signal results in the usual frequency-domain spectrum. The FT method saves time by collecting data all at once, rather than from the slow CW sweep of the field, and is well suited to signal averaging by collecting many FIDs from weak signals before Fourier-transforming them. Finally, Figure10 shows the block diagram of a modern pulsed NMR spectrometer. The sample is first positioned in the most homogeneous part of the magnetic field (see Figure7). Then the computer activates the pulse programmer and precisely timed digital pulses are sent out. The rf pulses are generated by superimposing rf signals on these pulsed digital signals. The rf pulses are amplified and sent to the sample, where they produce FIDs. Upon amplification and detection by audio conversion, these signals are filtered and converted into a digital representation using an analogueto-digital (A-D) converter. These digital signals are finally stored in the computer for further processing or plotting. 2.2. High resolution

NMR of polymers

Though NMR spectroscopy is over 40 years old, it still remains in a state of rapid development. Magnetic field strengths of current superconducting NMR spectrometers are almost 40-fold stronger than those employed in the first permanent-magnet prototype spectrometers. Pulse programmable FT spectrometers allow the selective observation of nuclei based on their unique structural (chemical) and motional characteristics. Almost daily, new concepts (2D NMR, cross-polarisation, etc.) and new techniques (INEPT, DEPT, magic-angle sample spinning, etc.) are reported and applied to a variety of molecular systems, including both synthetic and biopolymers.

NMR Analysis of Interchange Reactions in Condensation Polymers

17

Though the first report of polymer NMR spectra (a wide-line 'H NMR study of natural rubber [9]) appeared the year after the discovery of the NMR phenomenon in bulk matter [10,11],it was only in the late 1950s that high-resolution NMR spectra were recorded for polymers. Even from polymer solutions, which are often very viscous, reasonably well resolved NMR spectra can be obtained, such as those reported for polystyrene [12,13]. The rapid local motions of polymer chain segments (nanosecond to picosecond range) produce high-resolution NMR spectra of dissolved polymers. The volumes pervaded by dissolved macromolecules are much larger than their molecular volumes and produce highly viscous solutions through polymer-polymer entanglements and entrapment of surrounding solvent molecules. However, as already mentioned, both the frequency at which a magnetic nucleus resonates and the width of the resulting resonance peak depend on the local structure and its motional dynamics in the immediate vicinity of the observed nucleus. Thus, NMR serves as a local microscopic probe of molecular structure and its motions, and can even provide highly resolved spectra of dissolved polymers, regardless of the fact that their overall motion may be sluggish since their local segmental motions are rapid.

2.2.1. ' H N M R in the study of interchange reactions Historically, the proton ('H) was the first nucleus observed in the magnetic resonance of polymers. The average sequence length in vinyl copolymers and the stereosequence in stereoregular polymers have been initially studied by high resolution NMR spectroscopy [14]. At that time, there had been very few reports on the determination of sequence length in condensation polymers [15]. The first noteworthy work on this subject was performed by Yamadera et aZ. [16]. In their paper, the 'H spectra of the following homo- and copolyesters were measured: poly(ethy1ene terephthalate) (PET) , poly(ethy1ene isophthalate) (PEI), poly(ethy1ene ort hophthalate) (PEO) , poly (ethylene sebacate) (PES) , poly(ethy1ene transhexahydroterephthalate) (PEH), poly(resorcino1 terepht halate/isopht halate) terepht halate) (PRT) , poly (ethylene (PET-I, 50:50), poly(ethy1ene terephthalate/orthophthalate) (PET0, 50:50), poly(ethy1ene terephthalate/sebacate) (PET-S, 50:50), poly(ethy1ene terephthalate/hexahydroterephthalate) (PET-H, 50:50), poly(ethylene/resorcinol terephthalate) (PE-RT, 50:50). These Samples were prepared from glycols and dimethyl esters of the acids by polycondensation in the melt under reduced pressure. The transesterification reaction was carried out between PET and PES, mixed in various proportions at intervals of 10% and stirred under a nitrogen atmosphere at 276°C for 10, 30, 60, 120, and 180 min. The NMR spectra were measured in a trifluoroacetic acid solution (0.05 g/ml) at 70°C with a Varian A-60 spectrometer. In the NMR spectra of copolyesters, such as PET-0, PET-S and PET-

H. R. Kricheldorf, 2.Denchev

18

2

4

6

8

10

Figure 11. NMR spectra of copolyesters (50:50): (a)PET-I; (b) PET-0; (c) PETS; (d) PET-H, and (e) PERT [16]

H, three peaks are observed in the region of 5 . 5 4 . 0 ppm (Figure 11). Two peaks (one at each side) are assigned to the ethylene groups placed between the same acid groups, because they coincide with the ethylene proton signals of the homopolyesters. A central peak, which is not observed in the spectra of the respective homopolyesters, is due to the ethylene glycol residue bonded to the different acid groups at both sides. The intensity of this peak represents the amount of the “heterolinkage”. For instance, the ethylene glycol residue (-G-) in PET-S can be adjacent to terephthalate at both sides (T-GT), to sebacate at both sides (S-GS) or to terephthalate and sebacate (T-GS). The respective proton signals appear at 5.53, 5.93 and 5.73 ppm. No fine splitting is observed in the proton signal of the ethylene glycol residue in the heterolinkage because of the small difference in the chemical shifts between the two methylene protons. PET-S is taken as an example. Molar fractions of terephthalate (Pr) and sebacate (Ps) are obtained from the intensities of the three kinds of signals in the NMR spectrum:

& = fi-G-S/2 + h - G - T PS = &-G-S/2

+ PS-G-S

(18) (19)

where &-G-T, PS-G-Sand &-G-s represent the ratios of the integrated intensities of T-G-T, S-GS, and T - G S signals, respectively, to the total intensity of the ethylene glycol residues. If one could inspect the units along the PET-S copolymer chain from one end to the other, the probability of finding a T unit next to an S unit would be

19

NMR Analysis of Interchange Reactions in Condensation Polymers

Similarly, an S unit exists next to a T unit with a probability of f i S = PT-GS12fi

(211

The degree of randomness is defined by

B

(22)

PST -t- f i S

When B = 1, the T and S units take a random distribution and the probability of finding a copolymer unit obeys Bernoulli statistics. If B < 1, these units tend to cluster in blocks of each unit, and finally B = 0 in a homopolymer mixture, whereas if B > 1, the sequence length becomes shorter, and B = 2 in an alternating polymer. The number-average sequence length of a T unit (En,) and an S unit ( E n s ) are given by

E,T

= 2 P T / f i - ~ - s = 1/&s

Ens = 2PS/fi-G-S

(23) (24)

= l/PST

Table 2 shows the result of transesterification between PET and PES. For all feed ratios examined, the value of B becomes almost 1 and the number-average sequence lengths Ens and L,T approach the theoretical values of a random copolymer after transesterification for 3 h. The above results illustrate the implementation of the IH N M R technique for sequential analysis of copolymers obtained in a molten binary homopolymer blend, where interchange reactions are possible. Obviously, as these progress, the polymer blend transforms into a block copolymer with the length of both PET and PES units decreasing to reach the values of a random copolymer, obeying the Bernoulli statistics. In the studies of Devaux et al. [17-191, the more complex system comprising poly(buty1ene terephthalate) (PBT) and bisphenol A polycarbonate (PC) was analysed. Employing again the 'H NMR technique and bearing in mind the mathematical approach of Yamadera and Murano [IS], the authors determine the sequence length of PBT and PC units in a copolymer obtained from the starting physical blend by means of exchange interaction Table 2. Changes in sequence distribution during transesterification between PET and PES at various feed ratios (161 PET/PES 50/50

Time (min)

Ens

.i;nT

B

10 30 60 120 180

10.905 3.289 2.404 2.222 2.141

10.638 3.236 2.353 2.165 2.092

0.186 0.613 0.841 0.912 0.945

H. R. Kricheldorf, 2.Denchev

20

(i.e., transesterification). This method is discussed in detail in Chapter 3 and for this reason it will not be given here. It is worth mentioning the same trend toward copolymer randomisation after application of appropriate annealing conditions - T, > 250°C and t , > 200 min. Recently, a new ‘H NMR technique has been developed, allowing the assignment of resonances of very complex molecules - the s+called chemical shift correlated (or COSY) spectrum. The basic principle of this technique may be seen in Figure 12. If, instead of transforming (FT) the free induction decay (FID)immediately after the 90’ rf pulse in the usual way (see Section 2.1.6), one allows a time interval for the nuclear spins to precess in the transverse plane and for the evolution of interactions between them, it is possible to obtain important information concerning the nuclear spin system. One may divide such an experiment into three time domains, as indicated in Figure12. The nuclear spins are allowed to equilibrate with their surroundings via spin-lattice relaxation during the preparation period. Following the 90; rf m

Preparation Evolution 0-

Transpose

Contour

Figure 12. Schematic representation of a two-dimensional (2D) correlated (COSY) experiment and spectrum. The correlated influence of the J-coupling between nuclei of different chemical shifts is shown [4]

NMR Analysis of Interchange Reactions in Condensation Polymers

21

pulse, the 2, y, and z components of the nuclear spins evolve under all the forces acting on them, including their direct through-space dipoledipole and through-bonds scalar (J)couplings. This time, tl, is termed the evolution period and provides, along with the acquisition or detection time, t 2 , common to all pulse experiments, the two-dimensional (2D) character of this experiment. Systematic incrementation of the evolution time tl (see Figure 12) provides the second time dependence. After each tl period, a second 90; rf pulse is applied and the exchange of nuclear spin magnetisation may occur. The FID is acquired during t 2 and transformed. The pulse sequence shown in Figure 12 is appropriate for the observation of a COSY spectrum, where the correlating influence between nuclear spins is their scalar J-coupling. The FID following each tl is different because the interacting spins modulate each other’s response. Each FID detected in t 2 is transformed, producing a series of 1024 matrix rows, one for each tl-value. Each row may consist of 1024 points (square data matrix), representing the frequency-domain spectrum for a particular value of tl , while the columns provide information about how the FIDs were modulated as a function of

tl.

By looking down the columns of the data matrix, in an operation called

9.0

-

8.8-

8.68.48.2.

g 8.0a 7.8rz 7.6n

!i(,

,

,

,

,

,

. .,;

8.0

7.6

,

I

6.8

6.6 9.2

8.8

8.4

7.2

6.8

Fa ( P P ~ ) Figure 13. COSY spectra of the soluble fraction of a PBT/PAr blend after transesterification [20]

22

H. R. Kricheldorf, Z. Denchev

the “transpose” in Figure 12,1024 new FIDs are constructed. (Note that at this stage the spectrum is represented, for the sake of simplicity, as a single resonance.) A second Fourier transformation is performed on the newly transposed FIDs, leading to a 2D data matrix which is actually a surface in three-dimensional space. The surface may be represented as either a stacked plot or a contour plot. The latter is usually preferred, since the stacked plot does not clearly show complex relationships and is very timeconsuming to record. Nuclei which do not exchange magnetisation have the same frequencies, F1 and Fz, respectively, during tl and t 2 (ie., F1 = F2) and yield the normal spectrum along the diagonals of the contour plot. Scalar-coupled nuclei exchange their magnetisation and have a final frequency differing from the initial one, i e . , F1 # Fz. These coupled nuclei give rise to the off-diagonal contours or cross peaks shown in Figure 12. If one considers the particular case of an equimolar PBT/polyarylate (PAr) system after intensive transesterification (240 min at 250°C), the COSY spectra of Figure13 are obtained (201. This figure makes evident the scalar coupling of the protons absorbing at 7.71 ppm and 8.56 ppm (a triplet and a doublet, respectively). As indicated by the authors [20],these peaks should be ascribed to the aromatic protons of an asymmetrically substit uted isopht halyl unit

Because of the extreme complexity of the PBT/PAr system, at this point there is no method developed for sequence analysis of these copolymers. 2.2.2. l S C N M R in the study of interchange reactions

The 13C nucleus occurs at a natural abundance of only 1.1% and has a small magnetic moment - about one-quarter that of the proton. Both factors tend to mitigate against the observation of high-resolution 13C NMR spectra. However, the decrease in observational sensitivity of the 13C nucleus can be compensated by employing the pulsed FT technique, combined with spectrum accumulation, as described in Section 2.1.6. The time saved by the pulsed FT recording of spectra makes possible the accumulation of a sufficient number of spectra to produce a suitable signal-to-noise ratio. Further increase in 13C signal intensity is obtained by removing the nuclear spin coupling between l3C nuclei and their directly bonded protons, and from the accompanying nuclear Overhauser enhancement (NOE). Removal of the strong (125-250 Hz) 13C-lH nuclear coupling, by providing a second rf field at the proton resonance frequency, results in the collapse of 13Cmultiplets and an improved signal-to-noise ratio. Saturation of nearby protons

NMR Analysis of Interchange Reactions in Condensation Polymers

23

produces a non-equilibrium polarisation of the 13C nuclei, which exceeds the thermal value and yields an increase in the observed signal strength. It has been demonstrated [21] that the dipoledipole coupling mechanism dominates for the 13C isotope and a maximum NOE factor of 3 is produced by a directly bonded proton. Having discussed several of the means utilised to overcome the inherent insensitivity of the 13C nucleus, let us now mention the principal advantage of 13C NMR spectroscopy of organic molecules, including polymers. It is the increased sensitivity of 13C shieldings to molecular structure, conformation and solvation (in the 200 ppm range for neutral organics, compared t o 10-12 ppm for 'H shieldings), which has resulted in the replacement of 'H NMR by 13C NMR as the method of choice in molecular structure investigations. It is important to mention here that in 13C spectra, the intensities of the signals do not always correlate well with the number of corresponding C atoms. This effect is explained by the different relaxation rates of the C atoms, which depend very strongly on their surroundings. For instance, the C atoms which are not bonded to H-atoms usually give weaker signals than those of CH3, CH2, and CH groups. However, by employing proper pulse sequences and times, this problem can be eliminated, so that reliable quantification of the 13C NMR signals is possible. Nevertheless, the 13C NMR spectra are much simpler than the proton ones. Generally, their interpretation is based on the chemical shifts of the signals. Peak intensity or multiplicity is rarely used. As in the proton spectra, the chemical shifts are usually referenced to a n internal TMS standard. Let us now consider a particular example of how 13C NMR can be used for the characterisation of the microstructure of condensation homo- and copolymers [22]. In Figure 14, the CO signals of different Nylon 6-Nylon 6,6 copolyamides are compared. The copolyamides are prepared in two basic ways: (A) from the corresponding neat polyamides via transamidation (2. e., by interchange reaction) or (B-D) from monomers. Obviously, the appearance of four types of CO groups is a n indication of copolymer formation, this process being strongly dependent on the starting system (homopolymer or monomer mixture) and in the second case on the AH salt/&-caprolactam monomer ratio. In the case of an equimolar ratio of the monomer units, four amide groups should be expected, namely: (a) characteristic of neat Nylon 6; (d) characteristic of neat Nylon 6,6; (b) and (c) structures obtained by the interchange reactions. The first two amide groups should be attributed to the corresponding homopolyamides, while the last two belong to transitional Nylon 6-Nylon 6,6 structures that do not exist in the starting neat polymers or monomers. Let us define the signal intensities of the homolinkages A-A and BB, and of the corresponding transitional structures A-B and B-A in the following way: I A = signal intensity of the A-A bond; I A ~ = signal intensity

H.R. Kricheldorf, Z.Denchev

24

1 a

A

I B

a

w b

*.-NH-( CHz)s-CO-NH-(CH2)s-NH-**

C

C

..-CO-(CH2)r-CO-NH-(CH2)s-CO-**

D

d

a

*.-CO-( CH2)r-CO-NH-(CH2)s-NH-** 1

-

-\

M

N'

30 Hz

Figure 14. CO signals in the 13C NMR spectra of various Nylon 6-Nylon 6,6 copolyamides measured in sulfuric acid (98%):A - prepared by transamidation from a Nylon 6/Nylon 6,6 blend for 2 h at 28O-29O0C; B - by copolymerisation C - by copolymerisation of AH salt and E-caprolactam (1:l) for 8 h at 26OoC; of AH salt and E-caprolactam (1:6); D - by copolymerisation of AH salt and E-caprolactam (3:2)[22]

NMR Analysis of Interchange Reactions in Condensation Polymers

25

of the A-B bond; IB = signal intensity of the B-B bond; IBI = signal intensity of the B-A bond, and k - monomer ratio A/B in the copolymer. The average length of both Nylon 6 and Nylon 6,6 blocks in the resulting copolymers is given by

Since in a binary system the number of A-B and B-A units should be the same,

it follows that

IA -k I A = ~ k(IB + I B I )

(27)

On reaching the thermodynamic equilibrium, one may use Eqs. (28)-(30) instead of Eq. (27):

I A = k I ~ 1; I A = ~ I A / k ; IA/IAt = k IB = h / k ; I B = ~ kIB ; I B , / I B= k

(28) (29)

The validity of the above equations is clearly demonstrated by the peak intensity ratios in Figure 14B-D, where k is known for all polymer samples. For instance, when k = 1:6 (C), this gives for IA/IB a value of 1:36, which is close t o the experimentally established one. Also very interesting are the results for L A and LB in Figure 14A: the Nylon 6-Nylon 6,6 copolymer prepared via interchange reactions from neat polyamides by annealing for 2 h at 280-290°C produces four CO peaks. Although they are not completely resolved, a rough estimate of the block lengths according to Eq. (25) gives an average length of about four Nylon 6 and four Nylon 6,6 units. In the copolymer produced from a n equimolar monomer blend subjected to polycondensation for 8 h at 260°C (Figure14B), the average length of each block type is of about 2.0. Similar results are obtained in a polyester blend comprising poly(hexamethy1ene terephthalate) and poly(ethy1ene adipate). Again, LA = L B 4 for copolymers obtained via interchange reactions in a binary blend of the neat polyesters, and LA = LB 2 for the copolymer produced from the corresponding diacids and diols. More recently, Backson et al. [23] got the same results for the PETIPBT system. A possible explanation for not reaching a completely random copolymer in a binary homopolymer system by interchange reactions is given in Chapter 8. Let us now consider an example of 13C NMR sequence analysis in ternary aliphatic copolyamides [24]. Copolymerisation of three suitable N

N

26

H.R.Kricheldorf, Z. Denchev

monomers should lead to ternary copolyamides containing nine different amide groups, if each monomer has reacted with itself and with both other monomers. A product resulting from co-condensation of AH salt and two lactams is considered. A mixture of 50 mmol portions of c-caprolactam, of AH salt and of 12dodecanelactam was first heated for 30 min at 200°C under a slow stream of nitrogen and then for 8 h at 260°C. A similar experiment was carried out with 100 mmol e-caprolactam, 50 mmol AH salt and 50 mmol 12-dodecanelactam. It was found that the copolyamide contains nine different amide groups which can be identified by their CO signals in the 90.5 MHz 13C NMR spectra. From the intensities of the CO signals, the ratio of monomer units in the copolymers, and hence the reactivity of the monomers, can be estimated. Furthermore, the average length of the homogeneous blocks can be calculated. As already mentioned, a maximum of nine CO signals is expected in the product of the above co-condensation (Nylon 6-6,6-12), and this is precisely the case visualised by the spectra shown in Figure 15. If the signals of all the different amide groups of a copolyamide are resolved, as in the case studied, two kinds of information can be obtained from a quantitative evaluation of the 13C NMR spectra: (i) the ratio of monomer units in the isolated copolymer allows one to estimate the relative reactivities of the monomers under the polymerisation conditions chosen and (ii) the signal intensities of the homogeneous amide bonds (A-A, B-B, C-C) compared to the heterogeneous ones allow the calculation of the average length of the homogeneous blocks. Thus, one can determine whether a block copolymer, an alternating sequence, or a random copolymer is formed. The following definitions are used in the discussion: A, B, C denote the three IAB, IAC = intensities of the A-A, A-B and different monomer units; IAA, A-C bonds, respectively (signals z,z’,x” in Figure15); IBB, IBA, IBC = intensities of the B-B, B-A and B-C bonds, respectively (signals y, y’, y”); Icc, ICA, ICB = intensities of the C-C, C-A and C-B bonds, respectively (signals I, z’, z” in the same figure). The ratios of monomer units in the copolymer are described as k-values according to the following equations:

The ratio of monomer units in a copolymer is given by the concentration of monomers in the reaction mixture if polymerisation is quantitative. If not, the ratio also depends on the reactivity of the monomers and must be calculated from signal intensities or by means of other methods. For such a calculation, one may use all the signals that fulfill two conditions. The signal must stem from structurally identical groups in the various monomer units, as is the case CO-a-CH2 and w-CH2 groups in

I

I 179

.

I

2''

z' -NH-( CHz )1l-CO-NH-( CH2)s-CO-

-NH-( CHz) 1 I -CO-NH-( CHz)11-CO-

z

I 178

I

6 (PP4

179

178

-CO-( CHZ).I-CO-NH-( CHz) 1I 40- -NH-( CH2) 1l-CO-NH-( CHZ)~-NH-

Y"

-CO-( CHZ)~-CO-NH-( CHz )s-CO-

Y'

-CO-( CHZ)~-CO-NH-(CH2)c-NH-

Y

Figure 15. C 6 signals in the 13C NMR spectra (90.5 MHz, FSOJH) of a Nylon6-6,612 copolymer prepared from a mixture of e-capdactam, AH salt and 12-dodecanelactam, and (B) a 1:l:l mixture of the same monomers [24]

180

I

-NH-( CHZ)s-CO-NH-( CHz)11-CO-

XI'

-NH-( CHZ)~-CO-NH-( CH2)c-W-

2'

-NH-( CHz)s-CO-NH-( CHz)s-CO-

X

R

te 2.

5

z

28

H. R. Kricheldorf, Z. Denchev a

-NH-(CHz)r-CHz-CO-NH-

a

b

b

-CO-(CH~)~-CHZ-CO-NHC

-NH-( CHz )lo-C&-CO-NH-

J

I

d(ppm)45

*

I

43

,

I

41

,

I

39

,

I

37

.

1

35

.

I

33

.

I

31

Figure 16. cu-CHz and w-CHz signals in the NMR spectrum (90.5 M H z , FSOsH) of a Nylon 6 4 , 6 1 2 copolymer prepared from a 1:l:l mixture of Ecaprolactam, AH salt and 1Zdodecanelactam(see Figure 15B) [24] most aliphatic polyamides. Thus the signal intensities are not influenced by different nuclear Overhauser effects or different segmental motions. The second condition is that the signals of different monomer units must be resolved to such an extent that unambiguous assignment and quantitative evaluation is possible. In the case of Nylon 6-6,6-12, considered here, both CO and a-CH2 signals obey these requirements. If clearly separated, as shown in Figure16, the a-CH2 signals are advantageous for a quantitative evaluation for two reasons: the absence of splitting provides a better signal-tenoise ratio, and in the case of the CO signals, the intensities of more peaks are to be measured according to Eqs. (34)-(36):

Since all the interesting signals are not always well resolved, it is important to keep in mind that in all kinds of copolymers, corresponding heterogeneous bonds should be present in equal concentrations. Hence for ternary copolyamides, the following equations must be fulfilled:

From Figures 15B and 16, the following values were estimated:

~ A = B 0.9;

NMR Analysis of Interchange Reactions in Condensation Polymers

29

kAC = 4.5; kBC = 5.0; A:B:C = 10:9:2, indicating that the reactivity of the monomers decreases in the order:

AH salt > E-caprolactam >> 12-dodecanelactam

For the calculation of the average length of the homogeneous blocks in random copolyamides, only the CO signals are useful. The average block length of each monomer unit is given as follows:

Thus the block length of Nylon 6 in Nylon 6-6,6-12, prepared from a 1:l:l monomer mixture (Figurel5B) is 1.75, while it is 2.75 in the analogous copolyamide prepared from a 2:l:l monomer mixture (Figurel5A). On the other hand, the average block length of Nylon 12 is in the range of 1.1-1.2 in these copolyamides. Since the block length of Nylon 6,6 is also below 3, the conclusion can be drawn that these ternary copolymers possess a structure that is more random than blocky in character. Another point of interest is to determine whether the sequence of a copolyamide is the result of a thermodynamically controlled equilibrium caused by transamidation reactions or it originates from a kinetically controlled polymerisation. In the former case, the intensity ratios of the CO signals obey Eqs. (43)-(45), since the concentrations of all the kinds of amide groups depend exclusively on the ratios of the respective monomer units under these conditions: 2

= I A c kAC = IBk i B = Ic k A c

(43)

IBB= I B A k B A = IBCkBC = I A k i A = Ickgc ICC = I C A k C A = I ~ kCB B = I A k & = IBk 2c B

(44)

IAA = I A B kAB

(45)

In order to test whether or not the sequence is thermodynamically controlled, the signal intensities of the homogeneous and heterogeneous bonds must be compared. If the CO signals of the homogeneous bonds are less intense than given by EQs. (43)-(45), the formation of an alternating sequence is favoured. If the signals of the homogeneous bonds are more intense, a tendency to block formation exists. This investigation reveals that the copolyamides of Figure 15 possess a thermodynamically controlled primary structure, as expected for condensation at 250OC. A problem, which cannot be solved immediately by NMR spectroscopy, is the block length distribution within one chain and between different chains of one sample. However, if the copolyamide sequence is thermodynamically controlled, it is expected that the block length distribution

30

H. R. Kricheldorf, 2.Denchev

is likewise thermodynamically controlled, by analogy with the molecular weight distribution. In other words, the average block length is also the most probable block length in every part of the chain and in all chains of one sample. Finally, it should be pointed out that the quantitative evaluation of 13C NMR spectra has a limited accuracy for several reasons. If the copolymer chain is built up of monomer units that are very different in structure, their segmental mobilities (and hence 21' values) may be different. In this case, appropriate pulse width and repetition time should be selected to provide accurate intensity ratios. Furthermore, possible differences in the NOE must be taken into account. Limiting factors are, of course, the signalto-noise ratio and the resolution of signals. In most cases, only an FT NMR spectrometer working at a high field strength can provide 13CNMR spectra of good quality.

2.2.3. Other nuclei - "N and 20Si Because 15N and 29Si nuclei can be found in several important classes of polymers, here their NMR characteristics are briefly outlined in relation to the more commonly observed 'H and 13C nuclei. 15N and 29Si are spin -1/2 nuclei, and occur in 0.37 and 4.7% natural abundance, respectively. They each exhibit a range of chemical shifts at least as broad as observed for 13C nuclei. Though more abundant than 13C (4.7 us. l.l%),the "Si nucleus has even smaller nuclear dipole and magnetogyric ratio than I3C, so 29Si resonances about twice as sensitive as 13C resonances should be expected. However, since the magnetic moment and spin of the 29Sinucleus are antiparallel, y is negative. When broad-band proton decoupling is used to remove 29Si-1H scalar coupling, instead of a signal enhancement, as observed in 13C NMR, the 29Sisignal may be reduced in intensity. In addition, the spin-lattice relaxation times for 29Si nuclei in the dissolved state are t y p ically rather long, much like those of 13C nuclei. Nevertheless, pulsed FT NMR techniques have made the 29Si nucleus a valuable probe of silicon polymer microstructure. There are two special features in measuring silicon NMR. The first concerns the fact that silicon-containing materials constitute a major part of the construction materials of the probe head, resulting in a broad background signal at about -110 ppm. There are three ways to alleviate the problem: (i) if the signals are narrow, the smallest sweep possible should be used; (ii) if there are couplings to protons, population transfer pulse programs can be used; and (iii) if the lines are broad, subtract from a blank spectrum obtained under otherwise identical conditions. The other peculiarity concerns spectra of organosilicon compounds obtained with broad band decoupling of the protons. The NOE can then lead to null signals, if the (29Si, lH) dipole-dipole contribution to the other

NMR Analysis of Interchange Reactions in Condensation Polymers

31

longitudinal relaxation paths of the silicon is close to 2.52. Because relaxation times depend on the correlation time of the molecule, the signal intensity of a 29Si spectrum with an NOE varies with temperature. Again, there are three ways to remove this problem: (i) Doping the sample with a shiftless relaxation reagent, e.g., chromium acetylacetonate (Cr(acac)s) (ca. mol.l-'), which also gives shorter relaxation times as a side benefit. There are several disadvantages of this approach. The silicon compound can interact strongly with the chromium complex, the purity of the sample is impaired and population transfer experiments are not effective any more. (ii) Inverse gated decoupling. Here proton decoupling is only active during acquisition with long waiting times (3 to 5 times the relaxation time TI) between scans. The advantage of not polluting the sample is offset by a n ineffective use of spectrometer time, which can be somewhat alleviated by using shorter pulses (40")and shorter recovery times (20 s). (iii) Using population transfer pulse programs [25]. Nitrogen has two useful nuclei for NMR spectroscopy, 14N and 15N,each with advantages and drawbacks; the latter restricted the use of nitrogen NMR in the early decades of NMR spectroscopy, but FT spectroscopy, higher field magnets, larger samples, clever pulse techniques, and multidimensional spectroscopy have greatly expanded the use of nitrogen NMR P61. Both nuclei have rather low magnetogyric ratios y, so their sensitivity t o NMR detection is rather low, relaxation processes are rather slow, coupling constants J(N,X) are small and J(N,N) values very small. The spin -1/2 nucleus 15N is often used in high resolution work but its natural abundance is low, 0.36%. Another disadvantage is that y (15N) is negative, so NOE factors are negative. 15N signals become more negative with proton decoupling. The maximal proton-induced NOE factor for 15N is -4.93. A disadvantageous NOE can be improved by the use of paramagnetic additives, such as Cr(acac)3. In numerous studies enrichment of 15N is required, which may not be expensive if nitric acid, ammonium salts, or nitrites can be used as starting materials. With enrichment to 99%, the NMR receptivity is six times that of 13C in natural abundance. Sensitivity enhancement by polarisation transfer, by INEPT, or related methods is helpful, particularly if there are protons directly attached to the 15N. The gain, compared to 13C, is now greater, since y ("N) is smaller than y (13C). The highly abundant 14N nucleus (99.64%) has almost six times the receptivity of I3C in natural abundance (1%).However, I4N is quadrupolar (I = l), so that 14N NMR spectroscopy commonly suffers from line broadening and loss of spinspin coupling due to relaxation being too fast. The 14N quadrupole moment is relatively small, however, and if the local symmetry is high or the sample viscosity is low, 14N studies can be performed in high resolution. This possibility is often overlooked nowadays, although the phenomenon of the chemical shift was discovered long ago

32

H. R. Kricheldorf, Z. Denchev

in 14N resonance in an aqueous solution of ammonium nitrate, giving two sharp lines. Quantitative work on 14N or 15N resonance is difficult because of the multiplicity of factors affecting the signal intensities, and on 14N spectroscopy because of the line width. Nitrogen NMR spectroscopy affords a variety of information, with the choice of a spin 1/2 or a quadrupolar nucleus, and an unusual variety of both types, giving a range of 1350 ppm in chemical shift. Nitrogen forms bonds with all the elements except for the completely inert ones. Nitrogen can be found in nine stable oxidation states with bond orders up to three and coordination numbers up to six, the highest being in metal clusters. Furthermore, the shifts, coupling constants, and 14N line widths can be interpreted in terms of bond type, because of the characteristic influences of lone pair and 7c electrons associated with the observed nucleus, these electrons being of great importance to chemical structure and reactivity.

3. Interchange reactions involving different functional groups

As seen in the previous section, exchange reactions in both monomer and polymer systems have a significant effect on the structure of the materials

obtained, and therefore on their properties. In this section, we consider the most important interchange reactions, as shown schematically in Figure 1, namely: alcoholysis, acidolysis, esterolysis, aminolysis, and some special types of exchange interactions, all taking place in low and high molecular weight systems. The pattern that will be followed for each interaction type is given below: - Definition,

peculiarities in monomer (model) systems (where applicable); - Possible application for polymer synthesis; - Occurrence in polymer systems and accompanying effects on their microstructure; - Recent developments involving the corresponding exchange interaction.

3.1. Reactions taking place in polyesters involving ester groups By analogy with classical organic chemistry, most interchange reactions taking place or resulting in polyesters might be generalked as substitutions at carbonyl carbon atoms through an addition-elimination process as illustrated in Scheme 1 [27], where X = OH, OR”, OCOR”’ and C1, Y can be a neutral or negatively charged nucleophilic agent (R’OH and R’O or R’CO2, respectively), and R, R‘, R”,and R‘” are alkyl or aryl groups.

33

NMR Analysis of Interchange Reactions in Condensation Polymers

R-C(

/o + Y X

0"

2 [Ft-y-x] li

Y6+ la

0" or [R-?-XI li

Y" lb

% R-CdoY' ki1

+x

Scheme 1

The addition intermediates la and lb, for neutral and negatively charged nucleophilic agents, respectively, cannot be isolated or detected in any direct manner, and are therefore postulated on the basis of isotopic oxygen-exchange reactions and by extrapolating evidence of stable addition compounds observed for anhydrides, amides, etc. Experimental evidence also suggests that formation of the addition intermediate probably occurs through a perpendicular approach to the carbonyl carbon atom by the attacking nucleophile. The overall reaction is stepwise in nature and the relative rates of formation and partition of the addition intermediate determine the overall rate and the equilibrium between reactants and products. In general, the formation of the addition intermediate 1 has been found to be the slow step, and a catalyst is often used to increase the rate of this stage. The chemical structures of R, R', R", R"', XIand Y may influence both the rate of formation of 1 and its partition and, consequently, the overall rate and equilibrium. According to the mechanism proposed, an increase in the electronwithdrawing power of R will result in easier formation of 1. However, the resonance interaction of R with the carbonyl group tends to stabilise the ground state with respect to the transition state, which must be similar to the tetrahedral intermediate, and hence reduces the rate of formation of 1. Bulky groups on R can hinder the nucleophilic attack and therefore reduce the rate of the first step. Structural changes in X are related to both inductive and resonance effects and are more difficult to interpret. The eIectron-withdrawing power of X increases both the rate of formation of 1 and its partition toward products. Increased resonance of X with the carbonyl group increases the stability of the ground state and results in a lower rate; resonance interaction increases in the order C1< 02CR' < OR' < N%. The effectiveness of the nucleophile Y is, of course, related to the reactivity of the carboxylic acid derivatives; the more reactive the latter, the wider the range of nucleophiles which can be used effectively. The overall rate of the reaction depends on the rate of formation of 1 and on its partition between reactants and products; when X and the attacking nucleophile Y are equally good leaving groups, the increase in nucleophilicity parallels the increase in the overall rate.

34

H. R. Kricheldorf, Z. Denchev

The most widely used reactions for the preparation of polyesters are direct esterification (X = OH, Y = R’OH) and alcoholysis (X = OR”, Y = R’OH), usually performed at high temperature in the melt, and reactions of acyl chlorides (X = Cl) with hydroxy compounds (R’OH) or phenolates (R‘O-), generally carried out at low/medium temperature in solution or by interfacial synthesis. Exchange reactions, such as acidolysis (X=OH, Y = R’CO2R”’) and ester-ester exchange (X = OR”, Y = R’COzR”‘) can also take place at high temperature in the presence of suitable catalysts [27].

3.1 .I. Alcoholysis The interchange reaction between ester and hydroxyl groups is usually called alcoholysis. This is a reaction of great commercial importance, since most of the industrial processes of polyester manufacture make use exclusively or in large part of alcoholysis, which is represented in Scheme 2 for (-A-B-) and (-D-) type monomers, respectively. nROzC-B-C02-R+nHO-A-OH=i(-OC-B-CO-0-A-0-)n+2nROH nHO-D-CO2R (-0-D-CO-)nOR+ (n- 1)ROH Scheme 2 Ester derivatives have lower melting points, higher solubility in diols and can usually be obtained at a higher purity grade than the corresponding acids. Therefore, they often lead to better-quality products with easier process control. However, higher costs, resulting from more expensive raw materials, higher energy consumption and more expensive plants, sometimes make direct esterification more convenient, as in the case of PET

PI.

According to the general Scheme 1 for addition-elimination reactions, alcoholysis is assumed to occur through nucleophilic attack of a hydroxy compound on a carbonyl carbon atom. The reaction rate is therefore determined by both k l , the rate of nucleophilic attack, and k q / k i , the partitioning of the addition intermediate, and the equilibrium constant is given by K = k l k i / k i , l k i , l . In this way, reaction rates and equilibrium constants depend on the chemical structure of A and R. It is found that aliphatic diols can react with both alkyl and aryl esters, while phenolic compounds, which are poorer nucleophiles and better leaving groups than aliphatic hydroxy compounds, require R = aryl to yield polyesters. As found for direct esterification, compounds with tertiary hydroxyl groups are generally not suitable for polyesterification via alcoholysis. When RO is a better leaving group than A 0 and HO-A-OH is more nucleophilic than ROH, the equlibria in Scheme 2 are shifted toward the products and it could be assumed that polyester should be obtained under mild conditions and without removal of by-products, analogously to

NMR Analysis of Interchange Reactions in Condensation Polymers

35

the situation with polyamides [29]. Ogata et al. [30] attempted to prepare polyesters starting from “active diesters”, i.e., from diesters where R is a good leaving group. They actually observed an increase in the equilibrium constants from values lower than unity, usually observed for both A and R aliphatic radicals [31], to values of 6-17 for phenoxy, thiophenyl, and 3-oxypyridyl leaving groups. However, the reactions of ‘Lactive”adipates, terephthalates and isophthalates with ethylene glycol and butane diol occur only in the presence of suitable catalysts and, when performed in a closed system, yield only low molecular weight polyesters. The authors concluded that application of vacuum to remove ROH is also required in order to achieve high molecular weights. In contrast to direct esterification, alcoholysis proceeds very slowly in the absence of catalysts [32,33], even at high temperature. Strong protic acids, such as ptoluenesulphonic acid, sulfuric acid, etc., catalyse alccholysis, but they are not as effective as they are in direct esterification [33]. Furthermore, since they catalyse side reactions as well, the resulting polyesters are generally of poorer quality and are more prone to hydrolysis, compared to those prepared in the presence of metal derivatives, which are therefore the preferred catalysts. Due to the economic relevance of alcoholysis, an enormous number of patents have been published claiming, however, catalysts of uncertain value or novelty. Acetates of lead(II), lead(I), zinc, manganese, calcium, cobalt, and cadmium, and oxides such as S b 0 3 and GeO:!, for the first and second stages of reaction, respectively, and titanium alkoxides for both stages, have been found to be the most effective catalysts [34].Their overall activity is probably the result of various factors; solubility in the reaction medium [32,35],exchange-reaction capability of the original ligands with reactants, and effects on concomitant reactions are probably the most relevant. All these factors can obviously be affected by reaction conditions, such as type and concentration of functional groups, and temperature. The effects of catalysts on the exchange interactions in polyesters are considered in detail in Chapter 2. The role of alcoholysis is important in the preparation of polyesters (both neat and copolymers) by ring-opening polymerisation of cyclic esters [27]. Equilibria occurring in ring-opening polymerisation proceeding via alcoholysis are shown in Scheme 3. Active hydrogen donors, such as water, alcohols, amines, and similar substances, can be conveniently used to start hydrolytic polymerisation; a proper choice of initiator may provide a useful method of controlling the nature of the end-groups. Alcohols, amines, and similar monofunctional initiators lead to macromolecules with a functional hydroxyl group at one end only; water, aliphatic diols, and other similar difunctional initiators give macromolecules with functional groups, and the chains grow at both ends. In principle, the molecular weight of the resulting polyesters may be controlled by the ratio of lactone concentration to that of initiator [36].

36

H. R. Kricheldorf, Z. Denchev XH

+ 0-C-A

IJ

* XCO-A-OH

0 XCO-A-OH+

nO=C-

I)

A

* XCO(ACOz),AOH

0

Scheme 3

It has been reported that non-catalysed polymerisation initiated with hydrogen donors occurs at a relatively slow rate and gives only low molecular weight polyesters. The polymerisation is probably started by nucleophilic attack of the initiator on the carbonyl group of the monomer and proceeds by subsequent nucleophilic attack of the resulting hydroxyl endgroups. Metal ion salts or titanium and tin alkoxides are effective catalysts for this reaction and lead to an increase in both reaction rate and molecular weight. These catalysts, however, usually lead to broader molecular weight distributions because redistribution by interchange reactions occurs simultaneously with the stepwise chain growth [27]. When water is used as initiator, direct esterification may occur during polymerisation at high temperature and contribute to molecular weight broadening and, provided that volatile products are removed, to molecular weight increase, along with nucleophilic propagation. When acids or bases are used as catalysts, an ionic mechanism may become the predominant one [36]. In general, polymerisation of lactones and cyclic esters depends to a considerable degree on their chemical structure and, in particular, on their ring size and on the type and position of the substituents. High ring strain, originating from both angle distortion (three- and four-membered rings) and hydrogen-atom crowding within the ring (rings with more that seven or eight atoms), favours polymerisation, while substituents diminish the polymerisability of these monomers by increasing the ring stability with respect to the open chain. Consequently, most of the four-, seven- and eight-membered ring cyclic esters and carbonates are polymerisable, although some substituted ones can resist polymerisation [37]. Accordingly, y-butyrolactone, a cyclic ester with a fivemembered ring, does not polymerise under the usual reaction conditions, even though it has been reported that it polymerises (20%yield) to a low molecular weight polyester at 160°C under pressure of 2000 MPa. Lactones with a greater number of atoms in the ring can polymerise readily (6-valerolactone and Ecaprolactone) or with more difficulty (3-n-propyl-6-valerolactone2 and 6,6dimethyl-6-valerolactone3) or do not polymerise at all (pentadecanolide). A similar behaviour is found for cyclic diesters (polymerisability of diglycolide 4 > dilactide 5>> tetraphenyldiglycolide 6 and tetramethyldiglycolide 7) 1371.

6 b

37

NMR Analysis of Interchange Reactions in Condensation Polymers

&Me

n-Pr

2

3

Me

M$

0

0

0

4

5

6

h

eM$:

0

0

Me

7

The ring size also has a very important effect on the extent to which the cyclic monomers can be converted into polyesters. For instance, a substantial amount of 6-valerolactonewas reported to exist in equilibrium with the polymer at temperatures exceeding 150°C [38]. Along with the above-mentioned method for polymer preparation including alcoholysis, the latter reaction can occur in polymers, too, as represented in Scheme 4.

+

---C-B-CO-A-O----

II

0

II

0

to

__f

to

R-OH

-OO-.+HO-AAOO--

---C-B-CO-R

II

0

II

0

Scheme 4

This exchange normally leads to a decrease in molecular weight of the polymer. The kinetics of alcoholysis in a poly(decamethy1ene glycol adipate) polymer was studied by Flory [39]. Korshak et al. have studied the alcoholysis of poly(hexamethy1ene sebacate) by cetyl alcohol [40]. It has been shown that this type of interchange reaction takes place in the presence of basic and, especially, of acidic catalysts. The non-catalysed process requires high temperatures and longer treatment times. Wichterle and Eksner [41] have demonstrated for the first time that the amide bond can also be subject to alcoholysis. A relatively new method for the "environment-friendly" preparation of polycarbonate also involves an interchange reaction [42]:

nHo&eoH CH3

+n

~

o

i

o

*a

0

This could be considered as alcoholysis of diphenyl carbonate by bispheno1 A. Non-toxic solvents are employed and the by-product (PhOH) can

H. R. Kricheldorf, Z. Denchev

38

be recycled [42]. The reaction temperature range is 150-320°C, while the pressure applied can be from atmospheric to less than 1 mm Hg. Typical catalysts are bases, such as alkali metals (Li, Na) or their hydroxides. A number of carbonate esters successfully produced via interchange reactions of diphenyl carbonate and aromatic diols are reported in the literature. In general, aliphatic diols do not possess the thermal stability required to survive the polymerisation reaction. Attempts have been made to produce diaryl carbonates by the direct oxidative coupling of phenols to carbon monoxide or by direct condensation with carbon dioxide [43]. The expensive and relatively inefficient catalysts (Group VIII metals, Pd being favoured) and low yields have precluded a commercial process. It is generally acknowledged that the discovery of a convenient and cheap direct production of diaryl carbonates, without the agency of phosgene, would be a revolutionary development.

3.1.2. Acidolysis The exchange reaction between carboxyl and ester groups, commonly called acidolysis, is schematically represented below for (-A-B-) type monomers, and is also valid for (-D-) type monomers, where A, B and R can be aliphatic or aromatic (Scheme 5 ) [44].

+

nR-C02-A-02C-R nHOzC-B-COzH 2nRCOzH -(O-A-O-CO-B-CO-), Scheme 5

+

As for other ester-exchange reactions, it is generally accepted that the above interaction is an “equilibrium” one. The value of the respective equilibrium constant can be calculated from the equilibrium constants of hydrolysis of the esters, i.e., K = Kh/Kh, where Kh and Kht are the equilibrium constants of the hydrolysis of RCOzA- and -BCOzA-, respectively. If the chemical structures of the esters are not too different, it can be assumed that Kh and Kht have similar values and consequently K could be approximated to unity. In accordance with the expected low K-value, acidolysis can be successfully applied for the synthesis of polyesters, provided that RCOzH can be easily removed from the reacting system. The lower volatility of benzoic acid is probably among the reasons why benzoates are less appropriate than acetates (451. Due to volatility and monomer cost, acidolysis has been used in practice only when RCO- is an acetyl group. The reaction can be carried out by heating a bisphenol diacetate with a dicarboxylic acid [46]at high temperature in the melt, or in solution in a high-boiling solvent or even in the solid state, and acetic acid is removed under reduced pressure or distilled off with the solvent. It has been found that bulky alkyl groups, such as tert-butyl groups, adjacent to the phenolic function, can sterically hinder the polyesterification by acidolysis and,

NMR Analysis of Interchange Reactions in Condensation Polymers

39

when diacetates cannot undergo polymerisation, the corresponding bisphenols are insoluble in alkali [46]. Analogously to (-A-B-) type monomers, acetoxyarenecarboxylic acids have been converted to (-D-) polyesters or included, as comonomers, in the chains of previously prepared polyesters [47,48].When poly(oxy-1,Cphenylenecarbonyl) was obtained by heating p acetoxybenzoic acid in the presence of magnesium turnings at 220-280°C under reduced pressure in an argon flow, it was found that etherification and decarboxylation side reactions occur also [49]. For the preparation of copolymers, a polyester is melted and heated at high temperature with the acetoxyarenecarboxylic acid [47,48] (or with an equimolar mixture of a dicarboxylic acid and a bisphenyl diacetate). After an initial decrease in the molecular weight of the starting polyester, as a result of acidolysis and ester-ester exchange reactions, the molecular weight increases again in the subsequent polymerisation step where acetic acid is removed under reduced pressure. Acidolysis reactions can also be performed on acetoxy derivatives of aliphatic diols but since they react with more difficulty than diacetates or aromatic compounds, and since the corresponding diols are less expensive and can easily react with either dicarboxylic acids or their esters, acidolysis is not often employed for aliphatic diols. It has also been found that, for the acetoxy derivatives of aliphatic hydroxyl groups) substituents may hinder polymerisation by acidolysis, while the non-acylated monomer can polymerise by direct esterification and alcoholysis [50]. Acidolysis can take place in both polyesters and polyamides. In the case of polyesters, it proceeds according to Scheme 6 and should play an important role in redistribution (of molecular weight and of comonomeric units) during melt-blending at high temperature. Evidence of an abrupt decrease in molecular weight when polyesters are heated with low molecular weight compounds bearing carboxyl groups suggests that this reaction takes place to a considerable extent under the conditions commonly encountered in high-temperature polyesterification and, in certain cases, contributes significantly to the overall rate of molecular weight increase [45,51]. -O(CH2)n-O-C0

- (CH2)m-0-

+

RCO-OH

to

0

Scheme 6 Many different substances have been reported to be effective catalysts for this reaction: H2S04, BF3, magnesium, and BuzSnO are some examples; however, this aspect has not been extensively and systematically studied and a comparison of the catalytic activity of various compounds is therefore

40

H. R. Kricheldorf, Z. Denchev

impossible at present. The mechanism of acidolysis has been studied for low molecular weight compounds in a temperature range well below that commonly used for polymerisation; evidence of both alkyl-oxygen and acyl-oxygen fission has been reported [52]. For polymerisation by acidolysis, two mechanisms have been proposed: according to the first one [46], a reaction occurs by acylhxygen fission, involving an intermediate anhydride, analogous to that reported for low molecular weight compounds; the second mechanism [53], rejecting the first one, postulates that the reaction proceeds via a four-memberedring transition state. Both mechanisms are, however, based on uncertain evidence. As mentioned above for equilibrium constants, very few data have been published about the kinetics; Korshak et al. [51] suggested that the reaction between ethyl stearate and acetic acid in trioxane solution at 164°C was first-order overall. The reaction of pbutylbenzoic acid with PC was assumed to be second order, and an activation energy of 98 kJ/mol was calculated [19].The scarcity of kinetic data could be one of the reasons why the contribution of acidolysis is usually disregarded in polymerisation when direct esterification and alcoholysis can also occur. However, the presence of dicarboxylic acids has often been observed among the volatile products [54], suggesting that a contribution to the overall molecular weight increase originates from acidolysis. This contribution, which is particularly relevant when the hydroxyl-to-carboxyl end-group ratio becomes very low, should be taken into account for the correct interpretation of the experimental results of solid state polymerisation of PBT [55].

3.1.3. Esterolysis Another type of ester exchange reaction, also called ester interchange, double ester exchange or esterolysis, can occur between two ester groups, as shown in Scheme 7. It is perhaps even less studied than acidolysis [44,56] and has not found any practical application in the preparation of polyesters since no advantages are expected, compared to the reactions discussed above. Nevertheless, ester-ester exchange reactions may play an important role in determining the chemical structure of copolyesters prepared or processed at high temperature and in influencing the products prepared by melt-blending of different polyesters. Redistribution of chain lengths and randomisation of chemical units are the consequences of intra- and intermolecular ester-ester exchange reactions. The control of these reactions may provide a new method for the preparation of copolymers with a wide variation in microstructure, directly within processing equipment [18,19,56-62]. RC02R”

+ R’C02R”’

+iR’C02R’‘ Scheme 7

+ RCO2R”‘

NMR Analysis of Interchange Reactions in Condensation Polymers

41

Difficulties in separating the contribution of ester-ester exchange from those of other ester exchange reactions make its study problematic and ambiguous results can be obtained unless end-capped chains are used and scission reactions avoided [44,61]. The failure to use end-capped reactants may invalidate results attributed to ester-ester exchange, as the same reaction products can also be formed from consecutive alcoholysis or acidolysis react ions. Few data have been reported for the kinetics and equilibrium of this reaction. A second-order (first-order for both ester groups) is generally assumed and activation energies of 130-150 kJ/mol have been reported [61,64]. A mechanism involving an association complex has been postulated. No data are available for the equilibrium constant; however, it can be expressed as the product of the equilibrium constants of the alcoholysis reaction of the two esters on the left-hand side of the above equilibrium. For similar chemical structures of the esters, the equilibrium constant is expected t o be close to unity [52].

3.1.4. Recent developments i n the field of interchange reactions including ester groups During the last decade, interchange reactions in polymers have been the subject of extensive research. In a series of papers, J.Otton et al. [65681 investigate all reactions taking place during the formation of P E T via alcoholysis. By means of low molecular weight model compounds, the role of the catalyst was thoroughly examined. In the alcoholysis reaction, titanium was again pointed out to be the most active catalyst. A detailed investigation by means of IR and NMR spectroscopy as well as by electroconductivity, revealed two preferential coordinations on the titanium atom: first, the incoming alcohol, through its oxygen, then, the ester through its OR group. With the alkali metal carboxylates (where the order with respect to the ester is 0), it is the carboxylate anion which acts as a nucleophile on the carbonyl carbon atom of the reacting ester, whereas with Co and Mn (first order with respect to the ester), it is the metal which acts as an electrophile on the oxygen atom of the carbonyl group of the ester. More recently, Lei et al. [69] have also studied the kinetics of the interchange reaction in dimethyl terephthalate (DMT) with bis(2-hydroxyethyl terephthalate) in the synthesis of P E T and have drawn similar conclusions about the role of the catalysts. In a series of papers [70-731, the regularities of interchange reactions in lactone-based systems have been studied. The homo- and copolymers of this type have attracted much interest during the last decade because of their usefulness in medicine. Since the copolymeric lactones have shown a wider range of useful properties, attempts have been made to characterise the interchange reactions in both ring-opening lactone polymerisation and in blends of various lactone-based polymers.

42

H. R. Kricheldorf, Z. Denchev

Thus, the copolymerisation of glycolyde (GLY) and e-caprolactone (CL) were studied by the NMR technique (701 in order to determine the sequence lengths of GLY and CL units, and to find out the reaction mechanisms of the interchange reactions. Some of the basic results obtained are shown in Table 3. The sequence analysis of the GLY-CL copolymers shows that, depending on the temperature and nature of the initiator, copolyesters with a broad variety of compositions and sequences may be obtained. As a rule, acidic catalysts initiate cationic copolymerisation, yielding copolyesters richer in CL units. Complex-forming catalysts initiate an insertion mechanism which favours the incorporation of GLY, whereas anionic catalysts exclusively initiate the homopolymerisation of GLY. When the sequences obtained by copolymerisation of lactones are considered in terms of monomer reactivity, the role of transesterification must be elucidated. In order to determine which of the above catalysts cause intermolecular transesterification, the following experiment was carried out. Poly(e-caprolactone) (pCL) with a degree of polymerisation > 100 was dissolved in nitrobenzene and GLY was polymerised in this solution. Interchange reactions should result in the formation of copolyesters, ie., of Table 3. Copolymerisation of glycolyde and E-caprolactone under various conditions [70] Initiatora

Conditionsbsc

Time Yield Average block length (h)

(%)

GLY,EG

CL,Lc

8 44 44 44

75.5 31.5 91.0 58.5

2.8 17.0 9.0 2.4

2.9 2.2 9.2 7.3

44 44 44 44 44

60.5 23.5 77.5 27.9 71.7

2.9 <4.0 7.0 2.0 2.9

2.9 >30.0 3.0 0.0 5.0

ZnClz Bulk, 100°C Al(O-i-Pr)s Bulk, 100°C (n-Bu)zSn(OMe)z Bulk, 100°C

44 44 44

45.5 80.5 99.0

11.9 5.0 1.5

2.7 4.8 1.8

Al(O-i-Pr)a Nitrobenzene, 150°C (n-Bu)zSn(OMe)z Nitrobenzene, 100°C

44 44

50.5 65.0

1.7 1.8

1.8 1.9

FeCl3 A1C13 BF3.EtzO FSO3H

Bulk, Bulk, Bulk, Bulk,

100°C 100°C 100°C 100°C

FeCL AlC13 BF3.EtzO FS03H FS03H

Nitrobenzene, Nitrobenzene, Nitrobenzene, Nitrobenzene, Nitrobenzene,

~

~~

70°C 70°C 70°C 100°C 150°C

aAl(O-i-Pr)3-aluminium isopropylate, (n-Bu)zSn(OMe)z - dibutyltin dimethylate bFeed ratio glycolyl/hydroxycaproyl units = 1:l CMolarratio initiator/sum of both monomers = 1:lOO

NMR Analysis of Interchange Reactions in Condensation Polymers

43

CL-GLY and GLY-CL bonds. Since the concentration of CL was constant all the time, whereas that of GLY depends on the conversion, the molar ratio of CL-GLY to CL-CL bonds determined from NMR [70] can be a measure of transesterification. The results are listed in Table 4. Table 4 suggests the following conclusions about the occurrence of interchange reactions: (i) Intermolecular transreaction is only detectable in the presence of FeC13, BF3.EtnO and FS03H. Its extent increases with reaction time and temperature. (ii) Aluminium isopropylate and dibutyltin dimethylate do not cause any intermolecular transesterification. At first glance, this might contradict the fact that these alcoholates cause, in the case of copolymerisation of monomers, the formation of short sequences, ie., of random copolymers. Therefore, it should be pointed out that these initiators cause intramolecular transesterification. In a subsequent study, the regularities of the copolymerisation of GLY with P-propiolactone, y-butyrolactone, or S-valerolactone were disclosed [71]. The results obtained were exactly as in the previous case - cationic initiators (FeC13, BF3.Et20, FS03H) cause intermolecular transesterification, whereas the complex-forming catalysts (ZnC12, AlC13 and Al(0-iPr)3) do not. The problem of whether or not an initiator causes transesterification during copolymerisation is of significant importance. Initiators with high transesterification activity are useful for the preparation of amorphous copolyesters with random sequences, whereas initiators without transesterification ability are useful for the preparation of block copolymers. As far as metal alkoxides are concerned for the polymerisation of lactones, thorough NMR investigations [72] show that transesterification activity is high for tin alkoxides and low for aluminium alkoxides. Thus, tin alkoxides are useful for initiators when copolylactones with random sequences are to be synthesized, whereas Al(O-i-Pr)s is best suited for the synthesis of block copolyesters. Table 4. Polymerisation of glycolid in nitrobenzene in the presence of poly(Ecaprolxtone) [70] Initiator FeC13 BF3.EtzO BF3.EtzO FS03H FS03H AlC13

Al(O-i-Pr)s (n-Bu)nSn(OMe)z

T ("C)

Yield (%)

Transesterification (%)

60.0 83.5 30.0 79.5 75.5 21-95 87.5

31.0 12.0 84.0

100 100 150 70 150 100 100 100

Copolymerisation for 44 h in nitrobenzene

00

0.0

32.5 0.0 0.0 0.0

H. R. Kricheldorf, Z. Denchev

44

Along with the copolymerisation of the cyclic monomer lactones, another approach for preparation of copolymers might be the transesterification of the corresponding polyesters [73]. In this investigation, poly(Llactide) (pLL) was solution-blended in equimolar (referred to as monomer units) concentrations with pCL, pGLY or poly(propio1actone). After the common solvent had been evaporated, the physical homopolymer blend was thermostated in a nitrogen atmosphere in a closed flask. The annealing temperatures of the homopolymer blends were maintained in the 100-150°C range. Table 5 summarises the results of the cationic transesterification of pLL and pCL at 150°C in the presence of various catalysts. These data indicate that rapid degradation occurs at 150°C when triAic acid or methyl triflate are applied as initiators (Nos.1-7), so that copolyester could never be isolated. Better results were obtained with BF3.EtzO (Nos.8-11). At shorter reaction times of 48 and 73 h, copolyesters with relatively long blocks may be obtained. Longer reaction times, however, caused complete degradation of the polyesters, and copolymers with nearly random sequences were not obtained. An improvement in this direction could be achieved by means of tributyltin methoxide at 150°C (Table 6). In this series of experiments, a continuous decrease of the average block length of the homogeneous blocks was detectable at longer times. The smallest block length of E-oxycaproyl units (& = 3.5), found after 96 h, comes close to the value of a random sequence (Ec = 2.0). The characterisation of these copolyesters is discussed Table 5. Cationic transesterification of poly(G1actide) and poly(ecapro1actone) at 150°C [73] ~

No. 1 2 3 4 5 6 7 8 9 10 11

Catalysta CF3SOjH CF3S03H CFjSO3H CF3S03CH3 CF3S03CH3 CF3S03CH3 CF3S03CH3 BF3.EtzO BF3.EtzO BF3.EtzO BF3.EtzO

Time (h) 8 24

72

8 24 48 72 8 48 72 120

Yield

(%)b

0 0 0 91 88 0 0 95 86 80 0

qint.

(dl/g)' -

0.41 0.36 -

0.36 0.32 0.16 -

"Molar ratio of monomer units to catalyst = 200:l bYield indicates percentage of the homopolyester blend recovered after annealing 'Intrinsic viscosity, measured at 10 g/l in chloroform at 30"C dCalculated from the 13C NMR spectra

Ccd -

-

11 6 0

NMR Analysis of Interchange Reactions in Condensation Polymers

45

below in more detail. Finally, it is noteworthy that transesterification experiments conducted with blends of pLL and poly(P-propiolactone) under the conditions of Tables 5 and 6 yielded either blends of homopolyesters (after short times, < 24 h) or resulted in complete degradation. These results lead to the conclusion that transesterification of poly(Llactide) does not provide an easy access to entirely random copolylactones. Yet copolyesters with a certain degree of randomness and broad chain heterogeneity can be synthesised in this way, and such copolyesters are difficult to obtain by copolymerisation of L,L-lactide and other lactones. Thus, from a preparative point of view, transesterification of polylactones and copolymerisation of lactones are complementary methods rather than alternat ives. It is worth mentioning the more recent study of Tijama et al. [74] concerning the interchange reactions in the melt of polypivalactone (pPVL) with several diols, diacetates and diacids.

These authors have established that interchange of 8 with bisphenol diacetates and 1,4-butanediol occurs readily, particularly in the presence of a titanium catalyst; it is suggested that an initial cleavage of ester bonds in the polymer chain of 8 takes place, followed by a reaction between the newly formed ester end-groups and initially present hydroxyl chain ends. The acidolysis of pPVL with the diacids proved to be less effective; in the case of 10 mol % isophthalic acid, less than 1%of the diacid was incorpoTable 6. 'llausesterification of poly(G1actide) and poly(&-caprolactone)with tributyltin methoxide at 15OoC [73] NO.

M/Ca

1

200

2

200 200 200

Time (h)

Yield (%)

24

86 74 67 64

48

qint.

(dl/g)

0.62 0.55 0.49 0.45

LL/CLc 1.05 1.02 1.05 0.99

3 72 4 96 aMolar ratio of monomer units and catalyst bMeasured at 2 g/l in chloroform at 30°C 'Molar ratio of lactidyl and &-hydroxycaproylunits in the isolated copolyester as determined by 'H NMR spectroscopy dAverage block length of E-hydroxycaproyl units

Lc 13.3 6.0 5.1 3.5

46

H. R. Kricheldorf, Z. Denchev

rated into the polymer chains and a decrease in the logarithmic viscosity number of only 22% was found. Both the high stability of the ester bond in pPVL toward acids in general and the heterogeneity of these systems are supposed to be the reasons for this pPVL behaviour with respect to acidolysis. The synthesis of copolyesters via interchange reactions of 8 with several compounds has also been studied in the melt. In a first stage, ester bonds in the polymer chain are cleaved and new groups are incorporated into the polymer chain, while in a second step condensation of the endgroups formed occurs. Three procedures have been used, with tetrabutyl orthotitanate as a catalyst. pPVL was heated with equimolar mixtures of bisphenol A diacetate (BPAac) and terephthalic acid (TA), but no polymers were formed; instead, polycondensation of BPAac with TA occurred, leaving the pPVL unaffected. From pPVL and mixtures of BPAac and DMT polymers were obtained which contained a significant amount of copolymeric sequences. However, most of the polymer chains consisted of pPVL and poly(bispheno1 A terephthalate) blocks. Random copolymers with thermal stability were obtained after heating pPVL with PC and DMT. The latter process was studied in detail by IR, DSC, solubility, and selective degradation tests. Copolymer composition seems to be closely associated with the mechanism of the interchange reactions, as shown by its Monte-Carlo modelling performed by Montaudo [75] and described in more detail in Chapter 4. Recent studies dealing with preparation of polyesters via interchange interactions are numerous. In this respect, one should mention the vigorous development in the field of non-phosgene preparation of polycarbonates [76-871. Efforts have been made to design continuous methods for PC preparation [88,89]. Modification of PC polymers at the stage of their preparation via transreactions seems to be very promising, too. In such a way UV stabilised [go] and branched [91] polycarbonates have been prepared. There are also communications disclosing the interchange reactions in blends of PC and low molecular weight benzophenone derivatives [92] aiming at improvement of the properties of the starting PC. As mentioned above, the large-scale production of diary1 carbonates, being the starting materials for “environment-friendly”PC synthesis through alcoholysis, is also a rapidly developing field of research [42,93,94]. Another noteworthy issue is the synthesis of polyester homo- and copolymers via alcoholysis [95,96]. All-aromatic copolymers have been obtained by this technique [97] as well as dyeable PET-based copolyesters, namely PET-co-poly(ethy1ene adipate)-co-isophthalate sodium salt (981. An entirely novel technique for polyester preparation is transesterification in supercritical carbon dioxide [99],which could obviously be used not only for PET but also for many other polyesters [loo]. As reported recentiy by Ootoshi et al. [loll, poly(ethy1ene naphthalate) (PEN) can also be produced by alcoholysis of the corresponding diacid ester.

NMR Analysis of Interchange Reactions in Condensation Polymers

47

Interchange reactions of polyesters with low molecular weight reagents (alcohols, acids) is also a possible method of modification of the basic polymer. Al-Haddad et al. [lo21 and Bakmirzaeva et al. [lo31 have recently reported on the reaction kinetics of PET acidolysis with acetoxybenzoic acid. Glycolysis of PET with oligomeric polyethers is reportedly [lo41 a possible way of preparing polyether-esters. Interchange reactions between PET, hydroquinone diacetate and TA have been studied by Matthew et al. [lo51 and proved to be a suitable method of coupling in a copolymer of PET and hydroquinone-terephthalate units. An interesting trend in the field of transreactions in polyesters is the use of enzymes (mostly lipase and proteinase of various origins) both for transesterification of PET with triglyceride oil [106,107]and for preparation of polyesters by alcoholysis of terephthalic acid diesters with 1,4-butane diol in THF [108]. Similarly, poly( 1,4-dibutyl sebacate) was prepared by Linko et al. [log] using substituted diacids and diols, reaching molecular weights of >130,000. Partial transesterification of sucrose [110] and the preparation of its soybean fatty polyester [lll]are also worth mentioning. Other examples of lipase-catalysed interchange reactions are the alcoholysis of triglycerides to yield di- and monoglycerides, reported by Kumar et al. [112], and the lipase-catalysed transesterification of rape seed oil [113]. Interchange reactions in polyesters blends are also subject of extensive research due to the broad range of materials that can be produced in this way. Even in immiscible blends, where exchange interactions are possible, drastic changes in miscibility occur. The mechanical blends transform first into block copolymers, the block lengths gradually decreasing to attain, at equilibrium, random copolymers (see [114]and references therein). A more detailed approach to this issue may be found elsewhere in this book (see Chapter 8). Recently, Guo has studied the relationship of interchange reactions and miscibility [115]. The author examined the formation of a single phase as a consequence of transesterification, or vice versa, in some polyester blends. On the basis of NMR data, the conclusion was drawn that the interchange reaction (transesterification) is not the necessary condition for miscibility. PET, PC, and PEN containing polyester blends are among the most studied systems as far as interchange reactions in polymer blends are concerned. Interesting for a possible industrial application are the copolymers obtained via interchange reactions in PET/PC [116-1181, PEN/PC [119], PEN/pCL [120], and PET/PEN blends [121-1261. The PET/PBT blend reportedly offers some useful advantages in studying the sequence ordering and length of the copolymers produced, by means of NMR [127] and by DSC [128]. Kollodge et al. [129,130] studied a binary homopolymer blend composed of poly(2-ethyl-2-methylpropyleneterephthalate) and PC in which interchange reactions were induced. Special attention was paid to the socalled midchain-midchain (esterolysis) and midchain-end-group reactions

H. R. Kricheldorf, Z. Denchev

48

(alcoholysis or acidolysis) and their impact on the phase behaviour. It was found that the extent of interchange reactions required to shift the phase behaviour from two phases to one was 4% (2.8% alcoholysis and 1.2% esterolysis). It is worth mentioning here the studies of Ha et al. [131] on a PET/PEI blend, indicating a transformation from block into random ordering with the advance of interchange reactions. Kokkalas et al. have studied the catalyst influence in the so-called postpolycondensation of PET, involving interchange reactions between different PET macromolecules causing an increase of the molecular weight (1321. Polyester blends containing polyarylate 9 and interchange reactions therein have also attracted scientific interest [133-1351. Similar trends toward formation of copolymers and blend compatibilisation have been found and proved by thermal and NMR techniques.

k

PAr (tere:iso=l: 1)

9

Interchange reactions are often induced between polyesters (or polycarbonates) and some liquid-crystalline polymers. Again, the most studied aspects are the phase behaviour in such systems and their rheological properties as a function of the interchange reaction completion [136-1471. In this respect it is worth mentioning that the acidolysis of phenol acetates is the most widely used polycondensation method for the preparation of fully aromatic polyesters, in particular for all commercial liquid-crystalline polyesters - Vectrm, Rodrun@, XydarQ, G r a n l a , etc. Numerous papers and patents exist on this subject [105,148,149] and a recent review paper of Gallot [150] summarised the principles and strategies for the synthesis of block and comb-like liquid-crystalline polymers for biological a p plications. Interchange interactions in the process of ring-opening polymerisation and in copolymer systems produced in this way have also been given some attention. Isoda et al. have studied the impact of the Lewis acid catalyst on the occurrence of interchange reactions in the process of 6-valerolactone polymerisation [151]. It was found that some of the aluminium porphyrins used as catalysts cause transesterification, resulting in a broadening of the

NMR Analysis of Interchange Reactions in Condensation Polymers

49

polymer molecular weight distribution. Ma et al. [152] studied the miscibility, interchange reactions and the formation of ringed spherulites in a PBT/pCL blend at 250°C. By means of DSC and polarised light microscopy, the authors proved that an interchange reaction in this system depends on the miscibility of the starting blend. Type (-D-) copolymers of CL and L-lactide could reportedly be prepared in the melt at 110°C [153]. CL is polymerised first to yield a polymer with pendant OH groups. The latter initiates the polymerisation of subsequently added L-lactide. When L-lactide was first polymerised, followed by copolymerisation with CL, random copolymers were produced, their formation being attributed to interchange reaction (transesterification). Thermal decomposition of poly(1actic acid) was studied by Kopinke et al. by several techniques, including DSC, TG, and pyrolysis [154]. The reaction pathway was found to involve intramolecular transesterification, giving rise to the formation of cyclic oligomers. A very recent study by Montaudo et al. [155]discloses the mechanism of interchange reactions in PBT/PC blends. Employing appropriate polymer samples (end-capped or containing reactive middle-chain and end-groups) , the authors showed that the exchange process may proceed via two different mechanisms: (i) a direct interchange between inner functional groups, i.e., located inside the polymer chains or (ii) an attack of reactive chain-ends on the inner groups. When the concentration of the reactive end-groups is lower (e.g., in high molecular weight PBT/PC and PET/PC blends and in the presence of a transesterification catalyst), the first type of exchange interaction takes place which, in fact, represents an ester-carbonate exchange. With lower molecular weight reagents, the second type of interaction prevails. The latter includes either alcoholysis or acidolysis, depending on the type of reactive end-groups involved. The authors showed that by monitoring the composition of the copolymers formed, it is possible to distinguish between the contributions of the above two reaction patterns. This approach makes it possible to control the composition and yield of the copolymer and may be used in other systems where interchange reactions occur. As a concluding remark concerning the exchange interactions in polyesters, it can be inferred that acidolysis is the most useful method for the polycondensation of phenols, but not successful for alcohols (diols). Alcoholysis is, in contrast, useless for diphenols. Therefore, acidolysis and alcoholysis are complementary methods for the synthesis of polyesters, and the so-called direct ester-ester exchange is appropriate for the synthesis of copolyesters.

H. R. Kricheldorf, Z. Denchev

50

3.2. Reactions taking place in polyamides involving amine and amide groups Many excellent reviews have been written on polyamides, their synthesis and possible reactions [156-1601. In this section, only interchange reactions involving amine or amide groups are discussed. The polymerisation reaction in polyamide synthesis is polyamidation. Its elementary step, amidation, corresponds to the reversible reaction of amide hydrolysis (or alcoholysis) and therefore proceeds via the same tetrahedral intermediate as that of the latter reaction [lSl].Scheme 8 sums up the mechanisms of acid-catalysed and non-catalysed amidation of a carboxylic or ester group by an amino group [156]. I

I I I

I I I

-C-ORII 0

L -4 I

I

t.

'.

I I I I -C-N+I l l

I

!

OH I

i 4

I

Tetrahedral intermediates

i

Final

POUP

Scheme 8

51

NMR Analysis of Interchange Reactions in Condensation Polymers

Polyamides of the (-A-B-)n type can be prepared from diamines and diacids (Reaction (a)) and are also called Nylons, whereas polyamides of the (-D-)n type can be prepared from either aminoacids through their self-condensation (Reaction (b)) or cyclic amides (lactams) by an addition process (c). nH2N -A

in

- NH2 + nHOOC - B - COOH N-A

- N-CP -B - 7

0

+ 2nH20

To initiate the hydrolytic polymerisation of lactams, a fraction of the rings has first to be opened (Reaction (d)). These polymers are called polyamides or Nylons (z+l),and important representatives are polyamides 4, 6 , 11, and 12.

(N - C )+ H2O =H2N - (CH2)= - COOH 4 6 (CH2k

(4

Another synthetic route uses interchange reactions (e) and (f) involving at least one amide group. The total number of end-groups remains unchanged by these reactions, but the chain lengths are redistributed continuously. In Reaction (f), a diamine is formed which, under appropriate conditions, can be removed from the system. The lactam ring addition step ( c ) can also be regarded as an interchange reaction.

I: -B-C-N-Ad

-B-C-

b

H

H

I

I

+ - B ~ ~ - N - A -a -B-~-N-A-

d

I: N-A-NH2 + H2N-A-

0

H

c -B-C - &-A!-

b

H

+ - B ~ ~ - A - A - (e) 0

+ HzN-A-NHz

(f)

52

H. R. Kricheldorf, 2.Denchev

Melt polycondensation of w-aminoacids or equimolar mixtures (salts) of diamines (R = dkyl or aryl group) and diacids was the very first method of polyamide synthesis; nevertheless it is still largely used. Paradoxically, the weak point of melt polycondensation is the use of high temperatures to carry out the reaction, causing thermal decomposition of certain reactants. This can be avoided by the use of monomers having modified functional groups instead of the zwitterionic aminium and carboxylate groups [162].Modification of the carboxylates into esters or amides eliminates the ionic crystallinity of the monomer and avoids the need for excessive heating. However, this is not a real advantage when one has to operate above the melting temperature of the resulting polymer in order to achieve high molecular weight. Polyamides are readily formed in the same way from w-aminocarboxylic esters [163] and diamines [ 1641. The mechanism of condensation of amines and carboxylic esters is essentially the same as that of non-catalysed condensation (see Scheme 8 ) and could be considered as an interchange reaction between ester and amine groups. It should be mentioned that the ease of the aminolytic (nucleophilic) reaction by the attack of a given amine is influenced by the electronic and steric environments of the ester group and by the nature of the leaving group. In the absence of H+ ions, the reaction rate becomes slow in the later phase of polycondensation; for this reason partial hydrolysis of the starting material is profitable in some cases [165,166]. On the other hand, the use of alkyl esters sometimes gives rise to partial N-alkylation, which could limit the molecular weight of the resulting polymer and worsen its properties [167]. Similarly to the polyester synthesis, the interchange reactions in polyamides should be considered taking into account the polycondensation equilibrium. At different stages of this equilibrium, different types of exchange interactions are predominant. In the very first stages (a) and (b) of polyamide formation, exchange reactions between COOH (or COOR) and NH2 groups from the monomers are responsible for the chain growth. This is also valid for the lactam polymerisation (c). At a later moment, when the concentration of the NHCO groups becomes high enough, they are involved in amide interchange (f). Reactions (g)-(m) give a general idea of the equilibria that should be considered in the hydrolytic amidation process. Reaction (g) shows the polyamide salt formation. These ions cannot yield the amide group directly; what is more, they hinder the amidation. Hence, polyamidation proceeds either from the neutral carboxylic group (h) or via acid-catalysed reaction (m). It is obvious that the amount of water and the pH of the system should also have some effect. -NH~ -NHa

+ HOOC-

+ HOOC-

e -NH$-OOC-NH-CO-

+ Ha0

NMR Analysis of Interchange Reactions in Condensation Polymers

+ H2O * -NHi + OH-COOH + H20 * -COO- + HO : 2Hz0 * H30' + OH-COOH + H+ + -C(OH)$ -C(OH)$ + H2N-CO-NH+ H+ + HzO -NH2

53 (9 (j)

(k) (1)

(4

The reactivity of the functional groups has little dependence on either the length of the aliphatic group or the length of the polymer chain [168,169).Therefore, the polyamidation kinetics of the various polyamides are comparable. In contrast, for lactam ring structures, the reaction equilibria depend on the ring size [170]. The main amide equilibria are: (i) condensation and ring opening and (ii) amide interchange and ring addition. As far as exchange reactions are concerned, the second equilibrium seems to be more important. For the ring addition (c) , the equilibrium value under polymerisation conditions determines the minimal final content of lactam, which is 7.8% for Nylon 6 at 250°C and even higher for some other polymers [170]. Unfavourable values for some lactam polymers (Nylons 4 and 5) make them thermally unstable at higher temperatures. When two different polyamides are mixed for 3 min at 260"C,5% of the amide groups (as measured by NMR) have already undergone amide interchange and a block copolymer is formed. For longer reaction times (120 min at 260°C), a completely random copolyamide can be formed [171-1731. The amide interchange reaction is acid-catalysed [172]. In some cases, the residual monomers can be removed by an amide interchange reaction (f). More particularly, volatile diamines are removed in this way and are therefore often added in excess at the beginning of the reaction. Copolymerisation of caprolactam with other lactams, e.g., 12dodecanelactam, by hydrolytic schemes has the same advantages and disadvantages as homopolymerisation. Copolymerisation with Nylon salts can be performed without the addition of water, eliminating the hydrolysis step and retaining only the ring opening and polycondensation steps. Compared to the homopolymer, the random copolymers, showing depressed crystallinity, are produced on a much smaller scale [157]. Small amounts of comonomer are incorporated in order to influence the dyeing behaviour of textile or carpet fibres. Higher levels of comonomer systems (e.g., hexamethylenediamine-isophthalic acid salt) are used to prepare largely amorphous products which are valued for their transparency [157]. At this point, there have been quite a lot of studies on the phenomenon of interchange reactions in polyamides, enough to clarify the mechanism of transamidation. Korshak and hunze [174] referred to the processes as aminolysis and/or acidolysis, but did not report any rate-determining anal-

54

H. R. Kricheldorf, Z. Denchev

ysis. They indicate that polyamides having capped amine and carboxyl groups do undergo exchange reactions, but at a lower rate. Flory [173] also suggested that aminolysis occurs with the possibility of true amide exchange in polyamide melts. These reactions are:

+

*

-NH

-NH-CO-

-

-NH2

-NH-COAcidolysis

Aminolysis

-COOH

+

-NHz

-NH-COAmidolysis

+

-GO-NH-

=

I

-co

+

HOOC-

CO-

+ I

NH-

-NH

I

-CO

+ I

CO-

NH-

Beste and Houtz [172]have shown that when sebacamide and N,N’diacetylhexamethylenediamine are heated together, amide interchange occurs to form acetamide and poly(hexamethy1ene sebacamide). These authors have found that the presence of water and carboxyl groups significantly affects the reaction and suggested a square-root dependence on the Table 7. Amide interchange; 50 min at 245’C8 [172]

PI 0 [HI0 {COZH]O)”~ [A1 kz K2.5 0.363 0.31 30 10.25 5.09 4.89 31 0.32 10.16 0.374 5.00 4.98 31 0.31 10.14 0.360 4.98 5.00 0.701 0.75 33 22.6 4.50 4.99 0.140 0.15 29 7.48 2.50 5.2 0.077 26 3.0 0.072 2.50 7.48 “Concentrations are given in mequiv/g: [S]O = initial concentration of sebacamide; [HI0 = initial concentration of N,N’-diacetylhexamethylenediamine; [A] = concentration of acetamide after 50min; k2 = second-order rate constant (equiv/g)-’; f(2.5 = second-order rate constant (eq~iv/g)-’.~(min)-’ Table 8. Effect of water on rate of interchange [172] Added water (wt %) 0 0.42 1.16

L‘Anhydr~~~”

K2.5

30.2 34.3 47.0 23.6

NMR Analysis of Interchange Reactions in Condensation Polymers

55

carboxyl group concentration. As seen in Tables 7 and 8, the effect of water is consistent with the catalytic effect of hydrogen ions, ie., hydrolysis and transamidation are certainly consistent with their results. They also reported a possible mechanism involving the attachment of the proton to the amide nitrogen, followed by reaction with water to give a carboxyl group and an alkylammonium ion. Salt formation and amidation could then occur. Beste and Houtz [172] also estimated that about half of the amide groups will take part in an interchange within 30min during the second stage of Nylon 6,6 polymerisation to a molecular weight of about 23 000 at 283°C. Ogata [175],in his studies on the polymerisation and depolymerisation of poly(ecapro1actam) polymers, found that no intermolecular reaction occurs on heating c-caprolactam or diketopiperazine with a polyamide in the dry state. The w i d e groups of chain molecules react with c-caprolactam to a significant extent in the melt - the rates of reactions of monosubstituted or disubstituted acid amide with methyl &-caprolactam or Ecaprolactam being slower than that of the monosubstituted acid amide with c-caprolactam - and N-methyl &-caprolactam did not react at all. He also found that the amino cation RNH; reacts with the amide group of E-caprolactam, while the dissociated carboxyl group RCOO does not. Miller [176]found in studies with mixtures of N-ethylcaproamide and Nhexylacetamide containing small concentrations of hexylamine and caproic acid that amide exchange improves acidolysis and aminolysis with no detectable contribution of a direct reaction between amide groups. His results show that the exchange rate is not accelerated by hexylamine but is increased by caproic acid and that the formation of an anhydride intermediate yields kinetic equations that provide a good fit for the amide exchange rate data. In the rate equation derived (Eq. (46)), [Alo, [Clo, [HCIo, [HAIo, and [EC]o are the initial concentrations of acetic acid, caproic acid, hexylcaproamide, hexylacetamide, and et hylcaproamide, respectively. The activation energy in the 231-275°C range was 27.7 kcal/mol. Kotliar [177] showed that the interchange rate for dispersed molecular blends of Nylon 6 is acid-catalysed and is a complex fuction of the water content. However, his results could not definitely show that the interchange rate is a function of the amine ends. More recently, Fakirov et al. have studied the interchange reactions in PET/Nylon 6 blends [178]. The chemical changes occurring in the esteramide exchange or condensation reactions in the solid state lead to the formation of copolymer layers between the two components in the blend. The interphase plays the role of a compatibiliser. Prolonged annealing at higher temperatures (below the melting of PET) results in the transformation of the initially isotropic polyamide matrix into a copolymer which affects the morphology and properties of the blend.

56

H. R. Kricheldorf, Z. Denchev

The changes in the thermomechanical behaviour of the same PET/Nylon 6 blend (1:l by wt) subjected to mechanical and thermal treatments have been examined by means of dynamic mechanical measurements [179]. It was established from previous studies that PET/Nylon blends are incompatible in the isotropic state, but form the so-called microfibrillarreinforced composites upon extrusion, drawing, and suitable annealing. This study focuses mainly on the amorphous component of the blends and thus complements the above results concerning the crystalline phase. The orientation and crystallisat ion of the homopolymersinduced by drawing improve the dispersion of components and impart some compatibility as far as one glass transition is observed. Yet, by annealing the drawn blend below the melting temperatures of both components (e.g., at 220°C), the biphasic character of the composite is enhanced, inasmuch as the microstructures of both the crystalline and the amorphous phases are improved and the reorganisation of species within separate phases is favoured. The components of the heterogeneous blend become compatible, provided that annealing is performed at a sufficiently high temperature (240°C), falling between the melting temperatures of the two components and allowing the isotropisation of the lower-melting component, Nylon 6. The increase of compatibility is attributed to transreactions producing compatibilising layers of PET-Nylon 6 copolymers as phase boundaries between microfibrils and the amorphous matrix. Prolonged annealing (25 h) leads to the randomisation of the original block copolymers and results in the participation of the entire amount of polyamide in the copolymer, which is evidenced by the disappearance of the glass transition peak of Nylon 6 (1791 (see also Chapter 8). In the study of transamidation reactions in the melt of polyamide blends, it is important to know whether the resulting product is actually a block copolymer, a random copolymer, or a mixture of two homopolyamides. In blends composed of a semicrystalline aliphatic polyamide and an amorphous aromatic polyamide, knowledge of the length of the homogeneous blocks of the aliphatic polyamide can provide information about the kinetics and the crystallisability. It is known that at temperatures of about 250°C, copolymer formation occurs by amide exchange in a melt of two polyamides [176,180]. Some attempts have been made to characterise the extent of interchange reactions in mixtures of poly(m-xylene adipamide) and Nylon 6 [181] and also in mixtures of Nylon-z, y polymers [182],where x indicates the number of carbon atoms separating the nitrogen atoms in the diamine and y the number of straight-chain carbon atoms in the dibasic acid. However, these analyses by NMR are not sufficiently convincing. According to the literature, proton NMR has been used as a tool to identify polyamides and copolyamides, but as shown in [183], 'H NMR is not a general tool for discriminating between aliphatic copolyamides and blends of homopolyamides. It is now well known [22,24,184,185]that the shift differences for blends of homopolyamides, alternating copolyamides,

NMR Analysis of Interchange Reactions in Condensation Polymers

57

and random copolyamides are best expressed in the 13C carbonyl region. Another method is 15N NMR, which makes it also possible to determine sequences; however, the advantage of 'H and 13C NMR over 15N NMR is that the quantitative measurements are more accurate [186]. In a recent paper by Aerdts et al. [187],the 13C NMR spectra of a blend and a copolymer of the semicrystalline aliphatic Nylon 4,6 and the amorphous aromatic Nylon 61 were studied. Both polyamides are very suitable for such a n investigation because of their miscibility over the entire composition range. The absence of multiple phases in the melt simplifies the investigation of transamidation reactions. Another advantage is that this system is of practical interest since both polyamide resins are commercially available products. The copolymer is formed after reaction in the melt and subsequent 13C NMR measurements suggest that exchange reactions have taken place. Four kinds of dyad sequences could be determined in the carbony1 resonances in the 13C NMR spectra of the copolyamides and, from their relative peak areas, it was possible to determine the number-average block length as well as the degree of randomness using the theory developed for polyesters by Devaux et al. [17]. It was shown that after extruding a Nylon 4,6/Nylon 61 blend (1:l by wt) for 90min at 315"C, the degree of randomness was 24% and the average homologous block length of Nylon 4,6 was 8. The resulting copolymer can still crystallise, as demonstrated in [188].

3.3. Interchange mactions involving Si-0 bonds Interchange reactions are best studied and understood in polyesters and polyamides, but this does not exhaust their variety in polymers. Similar reactions can also take place, involving Si-0 bonds. Some aspects related t o the exchange reactions in organosilicon chemistry can be found in [189]. In this section, only the relatively new approach to polymer synthesis employing silylated starting monomers is discussed. In some cases, classical polycondensation reactions between COOH (or COOR) and OH- or NHz-containing starting monomers prove inappropriate due t o various problems arising either during the polycondensation itself (e.g., solidification of the reaction mixture, undesirable side reactions) or a t later stages (e.g., impossibility of purification of the final product). These problems can be avoided by modification (silylation) of all or some of the starting monomers. Such an approach was shown to be highly advantageous in the preparation of some thermotropic poly(ester anhydrides) [190,191] and linear, star-shaped, or hyperbranched poly(ester amides) [192]. Thermotropic poly(ester anhydrides) are of certain practical interest because they combine the useful properties of a fully aromatic polyester with a high rate of biodegradation (e.g., by hydrolysis) reached by incorporation of anhydride groups. Scheme 9 depicts the synthetic route via silylated compounds [190].

‘H30‘

‘H30‘

NMR Analysis of Interchange Reactions in Condensation Polymers

59

As one may conclude, the above reactions can be regarded as nonequilibrium interchange (transacylation) between terephthaloyl dichloride and the silylated hydroquinones and silylated 4-hydroxybenzoic acids to form complex and obviously random sequences of several ester and anhydride groups. Some of the advantages of the “silyl method” are that the bulk condensation of silylated monomers does not require solvents and the volatile chlorotrimethylsilane is the only by-product. At the end of polycondensation, by subjecting the reaction mixture to high vacuum at 300”C, all of the residual low molecular weight by-products and starting materials could distill off or sublime. 13C NMR spectra were obtained from freshly prepared solutions of poly(ester anhydrides) in a mixture of CDC13 and CF3COOH (4:l v/v); they are particularly easy to interpret when the molar ratio of hydroquinone and hydroxybenzoic acid is close to 1:l. The best resolution of carbony1 signals, which are most sensitive to sequence effects, was obtained for samples l l e and l l f (containing phenylhydroquinone). As shown in Figure 17, six relatively strong CO signals are observable. Taking into account that poly(pheny1hydroquinoneterephthalate), when measured separately, exhibits only one CO signal (signal a in Figure17), four CO signals are expected according to the above formula of l l e . However, it was demonstrated in a previous paper that the synthesis of polyanhydrides by the “silyl method” involves rapid transacylation [191].Therefore, two of the six CO signals may originate from the corresponding homoanhydrides formed according to Schemelo. Furthermore, three weak CO signals (z, y and z in Figure 17) are detectable and they result either from transesterification or from end-groups. Low molecular weight model compounds of the endgroups, such as 4-acetoxybenzoic acid or monomethyl terephthalate, did not prove to be useful for reliable assignments. Therefore, the synthesis of sample l l e was repeated at a lower maximum reaction temperature of

174

172

170

168

6 bpml

166

164

Figure 17. 75.4-MHz 13C NMR spectrum of the carbonyl signals of poly(ester anhydride) l l f measured in CDCl3/CF3COOH (4:l v/v) (1901

60

H. R. Kricheldorf, Z. Denchev

250°C and a polymer of lower molecular weight was obtained. Under these conditions, the signals of end-groups should be more intense, and signals resulting from transesterification should be weaker or absent. In fact, the intensities of signals 2, y and z were higher by a factor of 2. This finding indicates the absence of transesterification, in agreement with previous results [191]. 2

m

oceco-0-co*om

-

+

~oc-Q-co-o-co+=&o-com m o q & c o - o - c o ~ o ~ Scheme 10

It is worth mentioning here that silylated hydroquinone and 4hydroxybenzoic acid derivatives can easily be prepared by refluxing with hexamethyldisilazane in toluene for 2-4 h, followed by distillation under reduced pressure [1901. r 1 14

X = F,C1 14a, M a , 16a: -Ab: -A-

C:

-A-

16

16

=(cH2)6

=(CH2)8 =(C&)io Scheme 11

The low or even absent solubility of the fully aromatic homo- and copolyanhydrides can be overcome by implementation of aliphatic reagents according to Scheme 11 [191]. Obviously, when the A-radicals of the silylated acid and acid halide differ from each other, copolyanhydridescould be prepared. In Figure 18, the 13C NMR spectrum of an amorphous copolyanhydride, prepared from silylated sebacic acid and isophthaloyl chloride, is shown; it is clear that, as a result of interchange reactions (transacylation, acidolysis) the copolymer contains the following sequences in a nearly random order:

NMR Analysis of Interchange Reactions in Condensation Polymers

€ €

61

O-CO-O-(CH~).s-CO"

O-CO-O-(CH2).s-CO"!-O-CO

170.0

6 in ppm

165.0

160.0

Figure 18. 75.4MHz 13CNMR spectrum of a copolyanhydride 16,prepared from silylated sebacic acid 14 and isophthaloyl chloride 15,measured in CDCl, with internal standard TMS Another example of interchange reactions involving Si-0 groups is the synthesis of the so-called "hyperbranched polymers". They are thought to have a structure that is intermediate between those of linear polymers and dendrimers. As their name implies, these polymers are also highly branched, but their structure is neither regular nor symmetrical, ie., it contains randomly located branched as well as linear units. The synthesis of such polymers represents one-pot and "self governed" polycondensation process of trifunctional (A2B type) monomers and has the advantage of being much simpler and inexpensive, as compared to that of dendrimers. The hyperbranched polymers possess most of the useful properties of dendrimers: numerous reactive functional end-groups, high solubility, and possibly compatibility with other polymers, low viscosity in solution, eliminated crystallinity, etc. [192]. A major problem in the polycondensation of A2B monomers is the occurrence of undesirable side reactions that lead to the formation of crosslinks, thus preventing the preparation of completely soluble hyperbranched

H. R. Kricheldorf. Z. Denchev

62

structures. The first approach to the synthesis of completely soluble randomly branched (hyperbranched) polyesters was reported in [193] and used later by other research groups [194,195].It is based on the polycondensation of 3,5bis(trimethylsiloxy)benzoyl chloride, 17.A more versatile synthetic route was developed later, employing polycondensation of the trimethylsilyl esters 18 of acetylated trifunctional hydroxybenzoic acids. All studies on these systems demonstrated that polycondensation of silylated carboxylic acids represents a cleaner process, because these acids do not contain acidic protons. Thus the occurrence of acid-catalysed decarboxylation or Fries rearrangement leading to the formation of crosslinksis eliminated to a greater extent. It is worth mentioning here that the polycondensation of R-COZSiMe3 with acetylated phenol groups is the only ester-interchange reaction (or esterolysis) that does not need a transesterification catalyst at temperatures higher that 250°C. \

Me3 SiO

p o c 1 Me3SiO

P-1

~

(-ClSiMea)

17

P

CH3C02

CH3C02

COzSiMe3 18

ta..S 19

(-CH3CO$3iMe3) I

hyperbranched polyesters

NMR Analysis of Interchange Reactions in Condensation Polymers

63

MesSiNH 2CHaCO2a

c o c , 20

+

I

+ o y MeaSiNH

-2MeaSiCI

/

-CIIi COa H

do

IF

0-co

I

CO \

I

wfu

,OAc

6 NH \

AcO

24

Scheme 13 22 yields the crystalline monomer 23 which can be polycondensed as well

(two-step procedure, Scheme 13). In order to synthesize star-shaped polymers with dendritic star arms, three tetrafunctional comononers, 27, 28, and 29, were prepared. As illustrated in Scheme 14,the silylated diamine (pipermine 26 in this particular case) was acylated with 3,5-diacetoxybenzoyl chloride, 25. Three polycondensations were conducted with the isolated monomer

H. R. Kricheldorf, Z. Denchev

64

26

25

28

CHsCOz &

OCOCH3 C

O

-

N

H

W

-

C

O

CH3C02

@ OCOCHa

29

30

Scheme 14

23 and the maximum reaction temperature was varied between 250 and 280°C. The resulting poly(ester amides) 24 were completely soluble, and thus non-crosslinked, when the reaction temperature did not exceed 270°C. At a final reaction temperature of 280°C,a polymer forming gel particles was obtained. A similar series of polycondensations was conducted with

NMR Analysis of Interchange Reactions in Condensation Polymers

65

the silylated monomer 22. Although this monomer was not isolated and purified, completely soluble poly (ester amides) 24 were obtained up to temperatures of 270°C, by analogy with the results found for monomer 23. However, the viscosity values of the polymers prepared from 22 were higher. Thus, the polycondensation of 22 is clearly the more attractive approach, the more so as it may be conducted as a “one-pot” procedure. In this connection, it is worth noting that the products resulting from the LLacetate method” on the one hand, and from the “silyl acetate method” on the other hand, differ not only in their viscosities but also in their chemical structures. While their IR and ‘H NMR spectra are almost identical, the 13C NMR spectra exhibit significant differences, mainly in the region of the CO signals, indicating that the products of the acetate method are influenced by more intensive ester-amide interchange reactions and by more side reactions. Several polycondensations were conducted by in situ preparation of the silylated monomer and subsequent heating with one of the comonomers 27, 28, or 29 up to a maximum temperature of 270°C. The monomer/comonomer ratio was varied from 30:l to 60:l and 9O:l. The results suggest that all three comonomers 27-29 were actually incorporated into the poly(ester amides) 24 because the inherent viscosities increase with the monomerfcomonomer ratios. The ‘H NMR spectra confirm this conclusion. This means, in turn, that the copoly(ester amides) considered here possess the structure of a star with hyperbranched arms (30). Structure 31 obtained from 22 and 27 could be an example. Similar results are also reported in [197].

NH -C I

‘a OAc

Ac

Ac

NH ‘ 0 i a 0 A c

Y

31

In conclusion, it should be mentioned that interchange reactions involving Si-0 bonds, i. e., employing silylated starting monomers, might be very useful in stepgrowth polycondensation reactions to produce rather complex polymers that are otherwise more difficult or even impossible to prepare.

66

H. R. Kricheldorf, Z. Denchev

3.4. Interchange reactions involving urethane and urea g mu p s

A great variety of polymers are described by the general term polyurethanes. These materials are typically synthesised by the addition reaction of an alcohol and an isocyanate group (Scheme 15) [198]. R-NCO

+ H-OR’

s R-NH-CO-OR’

Scheme 15 Polyurethanes based on 4,4-diphenylmethanediisocyanate(MDI), 1,4butanediol (BDO) and polyoxytetramethylene have been investigated intensively because of their practical importance. The structure and morphology of the so-called hard segments, formed by the polyaddition of MDI and BDO, has been a topic of major interest. To obtain reliable experimental data from which to derive firm conclusions, it was necessary to synthesise strictly monodisperse oligourethanes as model compounds for the hard segments of segmented polyurethanes and as hard-segment precursors for the synthesis of model elastomers. The following series of oligomers has been synthesised independently in several laboratories [199-2101. The strategy of oligomer synthesis included the use of protecting groups for both stoichiometric and non-stoichiometric ratios of the reactants, as shown below.

with 5

2 n20

and R = -O-(CH2)4-OH,

R’ = H (diol), or 0

H

H

In the synthesis of the above model oligourethanes, the condensation reaction (n) between a primary m i n e and a chloroformate is employed [198]: R-NHz

+ R‘-OCOCI --+ -HCl RNH-CO-0-R’

(4

NMR Analysis of Interchange Reactions in Condensation Polymers

67

In fact, the reaction pathway is rather complex due to the necessity to isolate and characterise the intermediate products. Thus, the 4,4’diaminomethylenebis(benzene) (MDA) is protected by connecting one of its NH2 groups to a carbobenzoyloxy group (Reaction (0))to give 32 and BDO is converted to 33 (Reaction (p)) [211].

H

0 32

COCl2 -HCI’

(CH3)3C -0 (CH2)r-OC-Cl II 33

(PI

6

By reacting 32 and 33 and subjecting the reaction products to a rather complex chemical treatment [211] that falls beyond the scope of this chapter, the model oligomers 34 are obtained.

34

n=Oto7

In the monodisperse oligourethane synthesis with 4,4’-diphenylmethane iminocarbonyloxy units, a special type of urethane interchange reaction observed only for these materials should be mentioned. Peculiar DSC thermograms and size exclusion chromatography (SEC) of the oligourethanes after melting and recrystallisation revealed that the monodispersity had been destroyed and a polydisperse oligouret hane mixture had been formed [199,200,202]. This was demonstrated unambiguously by the diphenylmethane iminocarbonyl-terminated oligourethanes, which do not have a reactive end-group; thus the reactivity is given by the urethane group only. Starting with the first member of this oligomer series (34,n = 1, ie., a monodisperse material) leads to an oligomer mixture containing all the homologues from n = 0 to 7 (the highest homologue resolved by SEC)

68

H. R. Kricheldorf, Z. Denchev

in the most probable distribution [199,210,212].This transurethanisation (Reaction (9))proceeds in the solid state at temperatures well below the melting point and up to about 195°C without any evidence of reformation of isocyanate and hydroxyl groups, which is the usual decomposition reaction of urethanes at elevated temperatures and is also observed for these oligourethanes above 200°C (Reaction (q)). H

I

to

R-N-C-0-R’

II

R-NCO +HO-R’

(9)

0 From a number of model experiments with differently structured monoand bisurethanes, it can be inferred that this reaction is self-catalysed by

0

II

wN-c-0~ I H

+

H

H

H

H Scheme 16

69

NMR Analysis of Interchange Reactions in Condensation Polymers

hydrogen bond formation between urethane groups of adjacent chains. The transurethanisation possibly proceeds by a urethane interchange reaction of parallel chains (Scheme 16) or by a four-centre type of reaction of antiparallel chains or oppositely arranged groups (Scheme 17). The implications of this particular reaction are that the monodispersity of such oligourethanes is easily lost, and therefore any treatment of the sample at elevated temperatures destroys its uniformity. This reaction is non-catalysed. In the presence of a typical catalyst used in urethane formation (such as organic tin compounds), the urethane interchange reaction is shifted to much lower temperatures; this allows the synthesis of a mixture of the lower molecular weight oligomers. Therefore, any catalyst should be carefully excluded when synthesising model urethanes with diphenylmethane units. From the finding discussed above, the question arises of whether the oligourethanes and polyurethane elastomers synthesised in the presence of a catalyst have the tailored structure, that is taken for granted to establish structure-property relationships [213]. O H H O R1-0-C II - NI # - C H 2 # - k - 8 0 - R 2

+

I

1

1 1 O H

H O

R ? - O - 8 - h ~ C H 2 ~ h - 8 0 - R 4

O H Rl-0-C II - NI# - C H 2 a k - b O - R 3

H O

Scheme 17 One of the possible ways to obtain polyureas is via silylated intermediates [198]. Since this interchange reaction does not include formation and breaking of Si-0 bonds, it is discussed here. Polyureas generally exhibit higher melting points, lower solubility, and higher thermal stability than the corresponding polyurethanes. This is mainly a consequence of the bifurcated hydrogen bonds and therefore stronger interactions between neighbouring urea groups [214,215]. These polymer materials have numerous applications of high commercial importance, due to their special properties as well as because of the relatively

70

H. R. Kricheldorf, Z. Denchev

easy and fast formation of the urea linkage [216]. The synthesis of poly( 1,6hexamethyleneurea) 38 by using 1,6hexamethylenediamine 35 and trimethylisocyanatosilane 36 or hexamethyldisilazane 37 as starting materials (Scheme 18) has been reported [217]: HzN(CH2)aNHz+2(CH)s)sSiNCO 35 36 (CH3)3SiNHCNH(CH2)eNHCNHSi(CH3)

II

0

/

II

0

175OC,17 h

Hz NCNH(CH2)aNHCNH2+6(CH3)3SiNHSi(CH3)3

II

0

II

0

37

The silylated intermediate in Scheme18 is thought to be the first reaction product formed. Its further reaction gives polyurea together with various silicon compounds (hexamet hyldisiloxane, bis( trimethylsilyl) carbodiimide, hexamethyldisilazane, trimethylisocyanatosilane)and ammonia as side products of the polycondensation reaction. The molar masses (&) are reported to be in the range from 1650 to 2460. 4. Concluding remarks

An attempt is made here to present an overview on the wide variety of functional groups involving interchange reactions that take place in condensation polymers. Due to the limited length of this chapter, many particular aspects of this issue fall beyond its scope. However, they are discussed elsewhere in this book. Here we would like to only mention some reactions

NMR Analysis of Interchange Reactions in Condensation Polymers

71

occurring in the thermal degradation of condensation polymers that are somehow related to interchange reactions and are not discussed further. For instance, as far as hydrogen atoms in condensation polymers could be regarded as a special case of functionality, it is worth mentioning the hydrogen transfer in its various forms (N-H, C-H and 0-H) taking place via a n ionic mechanism at temperatures above 350°C. On the other hand, upon thermal treatment, intramolecular exchange becomes quite efficient, giving rise to cyclic oligomers, this process being thermodynamically favoured at high temperatures, or to alternating copolymers. These issues are discussed in detail by Montaudo et al. [218] (see Chapter 4). Interchange reactions occurring mostly in polyesters and polyamides in the solid state and the related phenomenon of chemical healing is d i s closed by Fakirov in [219].In most cases, these processes are accompanied by an abrupt increase of the molecular weight and may have a practical application for recycling purposes (see Chapter 11). In conclusion, it is worth mentioning again that interchange reactions in polyesters and polyamides are best studied and understood because of their prime industrial importance. The trend toward the implementation of transesterification and transamidation in the preparation of new homoand copolymers will most probably have a major importance in the future. However, the rapid development of novel polymer systems will require a deeper insight in interchange reactions involving polyurethanes as well as in the mechanisms of some alternative polycondensation pathways, employing silylated monomers, as in the cases of polyureas, polyanhydrides and hyperbranched structures. References 1. V. V. Korshak, Pure Appl. Chem. 1 2 , 101 (1966) 2. V. V. Korshak, Vsp. Khimii 35, 1030 (1966) (in Russian) 3. S. Spasov, M. Arnaudov, “Application of Spectroscopy in Organic Chemistry”, Nauka i Izkustvo Publ. House, Sofia 1978, p. 205 (in Bulgarian) 4. A. E. Tonelly, “NMR Spectroscopy and Polymer Microstructure: the Conformational Connection”, VCH Publishers, New York 1989 5 . F. A. Bovey, “Nuclear Magnetic Resonance”, Academic Press, New York 1988 6. F. A. Bovey, L. W. Jelinski, “Nuclear Magnetic Resonance“, in: Encyclopedia of Polymer Science, Wiley, New York 1987, vol. 10, p. 254 7. J. B. Stothers, “Carbon-13 NMR Spectroscopf, Academic Press, New York 1972 8. D. Canet, “Nuclear Magnetic Resonance - Concepts and Method’, John Wiley & Sons, Chichester, New York, Brisbane, Toronto, Singapore 1996 9. N. L. Alpert, Phys. Rev. 72, 637 (1947) 10. E. M. Purcell, H. C. Torrey, R. V. Pound, Phys. Rev. 69, 37 (1946) 11. F. Bloch, W. W. Hansen, M. E. Packard, Phys. Rev. 69, 127 (1946) 12. A. Odajima, J. Phys. SOC.Japan 14,777 (1959) 13. F. A. Bovey, G. V. D. Tiers, G. Filipovich, J. Polym. Sci. 38, 73 (1959)

72 14. 15. 16. 17.

H. R. Kricheldorf, 2. Denchev

F. A. Bovey, G. V. D. Tiers, Fortschr. Hochpolymer. Forsch. 3, 193 (1963) T. Harada, N. Ueda, Chem. High Polymers (Japan) 22, 685 (1965) R. Yamadera, M. Murano, J. Polym. Sci. 5A-1, 2259 (1967) J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20,

1875 (1982) 18. J. Devaux, P. Godard, J. P. Mercier, R. Touillaux, J. M. Dereppe, J. Polym. Sci., Polym. Phys. Ed. 20, 1881 (1982) 19. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 20. M. Valero, J. J. Iruin, E. Espinosa, M. J. Fernandez-Berridi, Polym. Commun. 31, 127 (1990) 21. K. F. Kuhlmann, D. M. Grant, J. Am. Chem. SOC.90, 7355 (1968) 22. H. R. Kricheldorf, R. Miihlhaupt, Angew. Makmmol. Chem. 65, 169 (1977) 23. S. C. E. Backson, A. M. Kenwright, R. W. Richards, Polymer 36,1991 (1995) 24. H. R. Kricheldorf, W. Hull, J. Macmmol. Sci.-Chem. A l l , 2281 (1977) 25. H. Marsmann, ‘LSilicon-29NMR”, in: Encyclopedia of Nuclear Magnetic Resonance, edited by D. M. Grant, R. K. Harris, John Wiley & Sons,Chichester, New York, Brisbane, Toronto, Singapore 1996, vol. 7, p. 4386 26. J. Mason, “Nitrogen NMR”, in: Encyclopedia of Nuclear Magnetic Resonance, edited by D. M. Grant, R. K. Harris, John Wiley & Sons, Chichester, New York Brisbane, Toronto, Singapore 1996, vol. 5, p. 3221 27. F. Pilati, “Polyesters”, in: Comprehensive Polymer Science, edited by G. Allen, Pergamon, Oxford, New York, Beijing, Frankfurt, Sao Paulo, Sydney, Tokyo, Toronto 1989, vol. 5, p. 297 28. K. Ravindranath, R. A. Mashelkar, in: Developments in Plastics Technology, edited by A. Whelan, J. L. Craft, Elsevier, London 1985, vol. 2, p. 1 29. G. C. Overberger, J. Shebenda, J. Polym. Sci. 7A-1, 2875 (1969) 30. N. Ogata, S. Okamoto, J. Polym. Sci., Polym. Chem. Ed. 11,2537 (1973) 31. G. Challa, Makmmol. Chem. 38, 105 (1960) 32. P. J. Flory, J. Am. Chem. SOC.62, 2261 (1940) 33. F. Pilati, P. Manaresi, B. Fortunato, A. Munari, V. Passalacqua, Polymer 22, 799 (1981) 34. K. Tomita, H. Ida, Polymer 16, 185 (1975) 35. G. Torraka, R. Turriziani, Chim. Ind. (Milan) 44, 483 (1962) 36. K. Smtome, Y . Kodaira, Makmmol. Chem. 82, 41 (1965) 37. H. K. Hall,Jr., A. K. Schneider, J. Am. Chem. SOC.80, 6409 (1958) 38. B. J. Tighe, in: Developments in Polymer Degradation, edited by N. Grassie, Applied Science, London 1984, vol. 5,p. 31 39. P. J. Flory, J. Am. Chem. SOC.62, 2255 (1940) 40. V. V. Korshak, V. V. Golubev, Izv. Akad. Nauk SSSR, Ser. Khim. 1949, 379 (in Russian) 41. 0. Wichterle, 0. Eksner, Coll. Czechosl. Comm. 25, 1371 (1960) 42. Y. Ono, Pure Appl. Chem. 68, 367 (1996) 43. A. J. Chalk, US 4096169, General Electric Co., 1988 44. A. M. Kotliar, Macromol. Rev. 16, 367 (1981) 45. F. I. Hamb, J. Polym. Sci.,Polym. Chem. Ed. 10, 3219 (1972) 46. M. Levine, S. C. Temin, J. Polym. Sci. 28, 179 (1958) 47. W. J. Jackson, H. F. Kuhfuss, J. Polym. Sci., Polym. Chem. Ed. 14, 2049 (1976)

NMR Analysis of Interchange Reactions in Condensation Polymers

73

48. US 3778410 (1973), Eastman Kodak Co., Inv.: H.F. Kuhfuss, W. J. Jackson; Chem. Abstr. 80, 146894t (1974) 49. J. Economy, R. S. Storm, V. I. Matkowitch, S. G. Cottis, B. E. Nowak, J. Polym. Sci., Polym. Chem. Ed. 14,2207 (1976) 50. H. K. Reimschuessel, T. DeBona, Macromolecules 13,1582 (1981) 51. V. V. Korshak, S. V. Vinogradova, in: Polyesters, edited by J. Burdon, Pergamon, Oxford 1965, p. 261 52. J. Koskikallio, in: The Chemistry of the Carboxylic Acids and Esters, edited by S. Patai, Interscience, New York 1969, ch. 3, p. 103 53. D. F. Loncrini, J. Polym. Sci. 4A-1,1531 (1966) 54. B. Fortunato, F. Pilati, P. Manaresi, Polymer 22, 655 (1981) 55. F. Pilati, G. S. Gostoli, G. C. Sarti, Polym. Process. Eng. 4,303 (1986) 56. V. V. Korshak, S. V. Vinogradova, in: Polyesters, edited by J. Burdon, Pergamon, Oxford 1965, p. 266 57. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22,229 (1982) 58. M. Kimura, R. S. Porter, G. Salee, J. Polym. Sci., Polym. Phys. Ed. 21,367 (1983) 59. M. Kimura, G. Salee, R. S. Porter, J. Appl. Polym. Sci. 29,1629 (1984) 60. F. Pilati, E. Marianucci, C. Berti, J. Appl. Polym. Sci. 30,1267 (1985) 61. P. Goddard, J . M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24,3301 (1986) 62. P. Goddard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24,3315 (1986) 63. P. J. Flory, Chem. Rew. 39,137 (1946) 64. H. G. Ramjit, J . Macromol. Sci.-Chem. A19, 41 (1983) 65. J. Otton, S. Ratton, J. Polym. Sci., Polym. Chem. Ed. 26,2183 (1988) 66. J. Otton, S. Ratton, V. A. Vasnev, G. D. Markova, K. M. Nametov, V. I. Bakhmutov, L. I. Komarova, S. V. Vinogradova, V. V. Korshak, J. Polym. Sci., Polym. Chem. Ed. 26,2199 (1988) 67. J. Otton, S. Ratton, V. A. Vasnev, G. D. Markova, K. M. Nametov, V. I. Bakhmutov, S. V. Vinogradova, V. V. Korshak, J . Polym. Sci., Polym. Chem. Ed. 27,3535 (1989) 68. J. Otton, S. Ratton, J. Polym. Sci., Polym. Chem. Ed. 29,377 (1991) 69. G. D. Lei, K. I. Choi, Ind. Eng. Chem. Res. 57,123 (1992) 70. H. R. Kricheldorf, T. Mang, J. M. Jontk, Macromolecules 17,2173 (1984) 71. H. R. Kricheldorf, T. Mang, J. M. Jontk, Makromol. Chem. 186,955 (1985) 72. H. R. Kricheldorf, M. Berl, N. Scharnagl, Macromolecules 21,286 (1988) 73. H. R. Kricheldorf, I. Kreiser, J. Macromol. Sci.-Chem. A24, 1345 (1987) 74. E. J. Tijama, L. v. der Does, A. Bantjes, Makmmol. Chem. 194,2807 (1993) 75. G. Montaudo, Macromolecules 26,2541 (1993) 76. A. G. Shaikh, S. Sivaram, Ind. Eng. Chem. Res. 31,1167 (1992) 77. A. G. Shaikh, S. Sivaram, C. Puglisi, F. Samperi, G. Montaudo, Polym. Bull. 32,427 (1994) 78. M. M. Despande, A. S. Jadhav, A. A. Gunari, J. C. Sehra, S. Sivaram, J. Polym. Sci., Polym. Chem. Ed. 33,701 (1995) 79. Jpn. Kokai Tokkyo Koho JP 09 59501 [97 595011 (1997), Asahi Chem. Ind., Japan, Inv.: H. Yatani, K. Komya, Chem. Abstr. 126,278349~(1997) 80. Jpn. Kokai Tokkyo Koho JP 09 12843 [97 124831 (1997), Daicel Chem. Japan, Inv.: T. Sugano, Y. Okano, Chem. Abstr. 126,172466t (1997)

74

H. R. Kricheldorf, Z. Denchev

81. Jpn. Kokai Tokkyo Koho JP 08 333446 [96 3334461 (1996), Teijin Ltd. Japan,

Inv.: T. Sawaki, S. Hirata, M. Shimonari, K. Sasaki, Chem. Abstr. 126,

157958m (1997) 82. Jpn. Kokai Tokkyo Koho JP 08 302005 [96 3020051 (1996), Idemitzu Kosan Co. Japan, Inv.: K. Tanaka, Chem. Abstr. 126,104584e (1997) 83. Jpn. Kokai Tokkyo Koho JP 08 302004 [96 3020041 (1996), Idemitzu Kosan Co. Japan, Inv.: M. Nakae, Y. Ishikawa, Chem. Abstr. 126, 104583d (1997) 84. Jpn. Kokai Tokkyo Koho JP 08 245782 (96 245782) (1996), Idemitzu Kosan 85. 86. 87. 88.

Co. Japan, Inv.: J. Takeda, K. Hayashi, K. Iura, T. Kawakami, Chem. Abstr. 126,104584e (1997) K. Tanaka, Idemitzu Giho 39,555 (1996) D.-J. Liaw, P. Chang, Polymer 37,2857 (1996) US 5519106 (1996), Mitzubishi Chem. Co. Japan, Inv.: M. Nukui, T. Tayama, T. Kashiwagi, M. Kimura, H. Shoji, Chem. Abstr. 125,87534y (1996) Jpn. Kokai Tokkyo Koho JP 08 337648 [96 3376481 (1996), Teijin Ltd. Japan, Inv.: T. Sawaki, M. Shimonari, K. Sasaki, T. Muraoka, Chern. Abstr. 126,

172068q (1997) 89. Jpn. Kokai Tokkyo Koho JP 08 333447 [96 3334471 (1996), Teijin Ltd. Japan, Inv.: T. Sawaki, K. Sasaki, T. Muraoka, M. Shimonari, Chem. Abstr. 126, 157959x1 (1997) 90. US 5556936 (1996), General Electric Co. USA, Inv.: J. A. King, P. J. McCloskey, A. M. Colley, D. M. Dardaris, L. P. Fontana, J. G. Berndsen, Chem. Abstr. 125,277535~(1997) 91. US 5597887 (1997), General Electric Co. USA, Inv.: J. A. King, P. J. McCloskey, M. Allice, D. M. Dardaris, Chem. Abstr. 126, 186847e (1997) 92. Y.M a k i u r a , E. Yoshida, T. Kitao, Polym. J . (Tokyo) 29,128 (1997) 93. Jpn. Kokai Tokkyo Koho JP 09 59225 [97 592251 (1997), Mitsubishi Chem. Co. Japan, Inv.: M. Inaba, K. Sawa, T. Tanaka, Chem. Abstr. 126,277275h (1997) 94. Y.Ono, Catal. Today 35,15 (1997) 95. Jpn. Kokai Tokkyo Koho JP 09 44838 [97 448381 (1997), Cheil Synthetic Textiles Co., Ltd. S. Korea, Chem. Abstr. 126,25210911 (1997) 96. Russ. Pat. 2048484 (1995), Inv.: V. N. Eliseenkov, B. E. Ivanov, N. G. Galakova, A. P. Gorbatenko, I. N. Zhmyklov, Izobretenia 32, 190 (1995), Chem. Abstr. 124,344483~(1996) 97. D. Frich, K. Goranov, L. Schneggenburger, J. Economy, Macromolecules 29, 7734 (1996) 98. Jpn. Kokai Tokkyo Koho JP 08 269820 [96 2698201 (1996), Asahi Chemical Ind. Japan, Inv.: K. Tsuruoka, A. Yamashita, J. Kato, Y. Yagi, Chem. Abstr. 126,32959f (1997) 99. A. L. Burke, G. Maier, J. M. Desimone, Polym. Mater. Sci. Eng. 74, 248 (1996) 100. PCT Int. Appl. WO 96 39456 (1996), Inv.: J. M. Desimone, G. Maier, Chem. Abstr. 126,104577e (1997) 101. Jpn. Kokai Tokkyo Koho JP 08 283395 [96 2833951 (1996), Mitsubishi Gas Chem. Co. Japan, Inv.: M. Ootoshi, F. Suzuki, K. Ishiguro, Chem. Abstr. 126,60530~(1996) 102. A. A. Al-Haddad, J. Matthew, React. Funct. Polym. 32,83 (1997)

NMR Analysis of Interchange Reactions in Condensation Polymers

75

103. G. Z. Bakmirzaeva, S. V. Zaljalieva, B. J. Kabulov, V. E. Kim, B. A. Yulchbaev, Dokl. Akad. Nauk Resp. Uzg. 1995,36 (in Russian), Chem. Abstr. 125, 276822b (1996) 104. C. C. F. S. Souza, D. M. M. Lima, E. B. Schoor, K. Tabak, Latin Am. Appl. Res. 26, 61 (1996) 105. J. Matthew, R. S. Ghagade, A. Lodha, S. Ponratham, S. D. Prasad, Macromolecules 30, 57 (1997) 106. L. W. Barret, L. H. Sperling, Abstr. Papers Am. Chem. SOC.203,563 (1992) 107. L. W. Barret, G . S. Ferguson, L. H. Sperling, J. Polym. Sci., Polym. Chem. Ed. 31, 1287 (1993) 108. H. G. Park, H. N. Chang, J. S. Dordick, Biocatalysis 11,263 (1994) 109. Y. Linko, Zh.-L. Wang, M. Lamza, J. Seppala, Ann. N Y Acad. Sci. 1996, 799 110. US 5596085 (1997), Krafts Food USA Inv.: R. S. Silver, G. L. Hasenhiittl, Chem. Abstr. 126, 186305b (1997) 111. PCT Int. Appl. WO 96 40701 (1996), Procter and Gamble Co. USA, Inv.: E. P. Grandberg, R. G. Schafermeyer, J. A. Letton, Chem. Abstr. 126, 119342~ (1997) 112. G. S. Kumar, A. Ghogare, D. Mukesh, J. Appl. Polym. Sci. 63, 35 (1997) 113. Y. Linko, Zh.-L. Wang, E. Uosukainen, J. Seppala, M. Lamsa, VTT Symp. 1996, 163, 237 114. R. S. Porter, L. H. Wang, Polymer 33, 2019 (1992) 115. M. Guo, Polym. Mater. Sci. Eng. 75, 12 (1996) 116. W. G. Zheng, Z. H. Wan, Z. N. Qi, F. S. Wang, Polymer 34,4982 (1993) 117. M. Fiorini, C. Berti, V. Ignatov, M. Toselli, F. Pilati, J. Appl. Polym. Sci. 55, 1157 (1995) 118. V. N. Ignatov, C. Carraro, V. Tartari, R. Pippa, M. Scarpin, F. Pilati, C. Berti, M. Toselli, M. Fiorini, Polymer 38, 201 (1997) 119. M. J. Fernandez-Berridi, J. J. Iruin, I. Maiza, Macromol. Chem. Rapid Commun. 16, 483 (1995) 120. K. Han, H.-J. Kang, Pollimo 20, 224 (1996) (in Korean), Chem. Abstr. 124, 31867611 (1996) 121. M. Guo, J. Am. Chem. Soc., Polym. Prepr. 38, 88 (1997) 122. M. Okamoto, T. Kotaka, Polymer 38, 1357 (1997) 123. F. J. Balti Calleja, L. Giri, H. G. Zachmann, J. Mater. Sci. 32, 1117 (1997) 124. Jpn. Kokai Tokkyo Koho JP 08 269179 (1996), Mitzubishi Rayon Co. Japan, Inv.: S . Tanaka, S. Ishiwatari, M. Tokumitzu, J. Yoshida, Chem. Abstr. 126, 32673h (1997) 125. PCT Int. Appl. WO 96 35571 (1996), Hoechst Celanese USA, Inv.: P. N. Chen, K. P. Music, G. W. McNeely, Chem. Abstr. 126, 32463q (1997) 126. D. W. Ihm, S. Y. Park, Ch. G. Chang, Y. S. Kim, H. K. Lee, J. Polym. Sci., Polym. Chem. Ed. 34, 2841 (1996) 127. S. C. E. Backson, A. M. Kenwright, R. W. Richards, Polymer36,1991 (1995) 128. Y. Yu, K.-J. Choi, Polym. Eng. Sci. 37, 91 (1997) 129. 129. J. S. Kollodge, R. S. Porter, Macromolecules 28, 4097 (1995) 130. J . S. Kollodge, R. S. Porter, Macromolecules 28, 4106 (1995) 131. W. S. Ha, Y. K. Chun, S. S. Jang, D. M. Rhee, Ch. R. Park, J. Polym. Sci., Polym. Phys. Ed. 35, 309 (1997) 132. D. E. Kokkalas, D. N. Bikiaris, G. P. Karayannidis, J . Appl. Polym. Sci. 55, 787 (1995)

76

H. R. Kricheldorf, 2.Denchev

133. E. Espinosa, M. J. Fernandez-Berridi, I. Maiza, M. Valero, Polymer 34, 382 (1993) 134. J. M. Martinez, J. Nazabd, J. I. Eguiazabd, J. Appl. Polym. Sci. 51, 223 (1994) 135. M. J. Fernandez-Berridi, J. J. Iruin, I. Maiza, Polymer 36, 1357 (1995) 136. C. I. Hwang, 0. 0. Park, Polym. J. 27, 232 (1995) 137. D. Frich, J. Economy, K. Goranov, Polym. Eng. Sci. 37, 541 (1997) 138. S. M. Hong, S. S. Hwang, Y. Seo, I. J. Chung, K. U. Kim, Polym. Eng. Sci. 37, 646 (1997) 139. D. S. H a , D. H. Kim, S. D. Cho, B. W. Jo, J. Korean Chem. SOC.41, 157 (1997) 140. K. U. Kim, S. M. Hong, S. S. Hwang, Y. Seo, J. Y. Lee, J. Jang, Korea Polym. J . 4, 198 (1996) 141. A. Nakai, T. Shiwaku, W. Wang, H. Hasegawa, T. Hashimoto, Macromolecules 29, 5990 (1996) 142. R. Benavente, Z. Zhu, J. M. Perena, A. Bello, E. Perez, Polymer 37, 2379 (1996) 143. G. D. Choi, S. H. Kim, W. H. Jo, M. S. Fthim, J. Appl. Polym. Sci. 55, 561 (1995) 144. K.-H. Wei, J.-Ch. Ho, Macromolecules 30, 1587 (1997) 145. M. J. Stachovski, A. T. Dibenedetto, Polym. Eng. Sci. 37, 252 (1997) 146. K.-H. Wei, H.-Ch. Ho, J. Appl. Polym. Sci. 63 1527 (1997) 147. E. M. Woo, S. N. Yau, J. Polym. Sci., Polym. Chem. Ed. 35, 97 (1997) 148. E. T.-G. Choi, Y.-K. Yun, J. Jin, Po1ym.-Plast. Technol. Eng. 36,135 (1997) 149. S. Jacob, J. M. Adduci, Polym. Prepr. (Am. Chem. SOC.,Div. Polym. Chem.) 38, 201 (1997) 150. B. Gallot, Progr. Polym. Sci. 21, 1035 (1996) and references therein 151. M. Isoda, H. Sugimoto, T . Aida, S. Inoue, Macromolecules 30, 57 (1997) 152. D. Ma, X. Xiang, X. Luo, T. Nishi, Polymer 38, 1131 (1997) 153. P. J. A. In’t Veld, E. M. Velner, P. Van De Witte, J. Hamhuis, P. J. Dijkstra, J. Feijen, J. Polym. Sci., Polym. Chem. Ed. 35, 219 (1997) 154. F.-D. Kopinke, M. Remmler, K. Mackenzie, M. Moeder, 0.Wachsen, Polym. Degrad. Stab. 53, 329 (1996) 155. G. Montaudo, C. Puglisis, F. Samperi, Macromolecules 31, 650 (1998) 156. H. Sakiguchi, B. Coufin, “Polyamides” in: Handbook of Polymer Synthesis, Part A, edited by H. R. Kricheldorf, Marcel Dekker Inc., New York 1992, p. 807 157. R. Gaymans, D. Sikkema, “Aliphatic Polyamides” in: Comprehensive Poly-

mer Science, edited by G. Allen, Pergamon, Oxford, New York, Beijing, Frankfurt, Sao Paulo, Sydney, Tokyo, Toronto 1989, vol. 5, p. 357 158. J. Zimerman, in: Encyclopedia of Polymer Science and Engineering, 2nd Ed., edited by H. F. Mark, N. M. Bikales, C. G. Overberger, G. Menges, Wiley, New York 1988, vol. 11, p. 315 159. R. E. Putscher, in: Kirk-Othmer Encyclopedia of Chemical Technology, edited by H. F. Mark, D. F. Othmer, C. G. Overberger, C. G. Seaborg, Wiley, New York 1987, vol. 18, p. 328 160. I. K. Miler, J. Zimmerman, Am. Chem. SOC.Symp. Ser. 285, 159 (1985)

NMR Analysis of Interchange Reactions in Condensation Polymers

77

161. B. C. Challis, J. A. Challis, “Amides and Related Compound-s”,in: Compre162. 163. 164. 165.

hensive Organic Chemistry; Synthesis and Reactions of Organic Compoun& edited by I. 0. Sutherland, Pergamon, Oxford 1979, vol. 2, p. 957 S. B. Speck, J. Am. Chem. SOC.74,2876 (1952) T. Curtis, F. Goebel, J. Prakt. Chem. 37,150 (1988) R. Meyer, J w t w Liebigs Ann. Chem. 347,41 (1906) C. F. Horn, B. T. Freure, H. Vineyard, H. J. Decker, J. Appl. Polym. Sci. 7 ,

887 (1963) 166. Ger. 1203466 (1965), BASF, Inv.: M. Heinzeler, H. Linge; Chem. Abstr. 62, 7964d (1965) 167. G. Bier, “Addition and Condensation Polymerisation Processes”, in: Ad168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183.

vances in Chemistry Series, edited by R. F. Could, Am. Chem. SOC.,Washington DC 1969, vol. 91, p. 612 E. Schwarz, in: Kunststoff Handbook, edited by E. H. R. Vieweg, A. Muller, Hanser, Munchen 1966, vol. 6, ch. 2 P. J. Flory, “Principles of Polymer Chemist$, Cornell Univ. Press, Ithaca, N Y 1953, p. 102 H. Sakiguchi, in: Ring-Opening Polymerisation, edited by K. J. Ivin, T. Saegusa, Elsevier, London 1984, vol. 2, ch. 12 C. Ayers, J. Appl. Chem. 4,444 (1954) L. F. Beste and R. C. Houtz, J. Polym. Sci. 8 , 395 (1952) P. J. Flory, J. Am. Chem. SOC.64,2205 (1942) V. V. Korshak, T. M. Frunze, “Synthetic Heterochain Polyamides” , Monsan, Jerusalem 1964, p. 258 N. Ogata, Makromol. Chem. 30,212 (1959) I. K. Miller, J. Polym. Sci., Polym. Chem. Ed. 14,1403 (1976) A. M. Kotliar, J. Am. Chem. SOC.,Polym. Prepr. 19,111 (1978) S. Fakirov, M. Evstatiev, J. M. Schultz, Polymer 34,4669 (1993) T. Serhatkulu, B. Erman, I. Bahar, S. Fakirov, M. Evstatiev, D. Sapundjieva, Polymer 36,2371 (1995) I. Luderwald, H. R. Kricheldorf, Angeur. Makromol. Chem. 56,173 (1976) I. Takeda, D. R. Paul, Polymer 32,2771 (1991) I. Takeda, D. R. Paul, Polymer 33,3899 (1992) H. R. Kricheldorf, E. Leppert, G. Schilling, Makromol. Chem. 175, 1705 (1974)

184. 185. 186. 187.

H. R. Kricheldorf, E. Leppert, G. Schilling, Makromol. Chem. 176,81 (1975) G. Shilling, H. R. Kricheldorf, Makmmol. Chem. 176,3341 (1975) H. R. Kricheldorf, Pure Appl. Chem. 54,467 (1982) A. M. Aerdts, K. L. L. Eersels, G. Groeninckx, Macromolecules 29, 1041

(1996) 188. K. L. L. Eersels, A. M. Aerdts, G. Groeninckx, Macromolecules 29, 1046 (1996) 189. E. E. Bostic, in: Chemical Reactions of Polymers; High Polymers, edited by E. E. Fettes, Interscience, New York 1964, vol. 19, p. 515 190. H. R. Kricheldorf, D. Lubbers, Macromolecules 25,1377 (1992) 191. H. R. Kricheldorf, D. Lubbers, Makromol. Chem., Rapid Commun. 11, 83 (1990) 192. H. R. Kricheldorf G. Lijhden, Macromol. Chem. Phys. 196,1839 (1995)

78

H. R. Kricheldorf, 2.Denchev

193. H. R.Kricheldorf, &.-Z. Zang, G. Schwarz, Polymer 23, 1821 (1982) 194. C. J. Hawker, J. M. J. FEchet, J. Am. Chem. Sac. 112,7638 (1990) 195. S. R.Turner, B. I. Voigt, T. H. Moureg, Macromolecules 26,4617 (1993) 196. H.R.Kricheldorf, 0. Stober, D. Lubbers, Macromol. Chem. Phys. 196,3549 (1995) 197. H.R.Kricheldorf, G. Liihden, J. Macromol. Sci. - Pure A w l . Chem. A32, 1915 (1995) 198. C. D. Eisenbach, H. Nefiger, “Poly(urethanes) and Related Polymers”, in: Handbook of Polymer Synthesis, Part A, edited by H. R. Kricheldorf, Marcel Dekker Inc., New York 1992,p. 685 199. C1. Giinter, Diploma Thesis, University of Freiburg, 1982 200. Y.Camberlin, J. 0. Pascault, J. Polym. Sci., Polym. Chem. Ed. 20, 383 (1982) 201. C. D. Eisenbach, C1. Giinter, Proc. Div. Polym. Mater. Sci. Eng. 49, 239 (1983) 202. 2. Y.&in, C. W. Macosko, S. T. Wellinghoff, Proc. Div. Polym. Mater. Sci. Eng. 49,475 (1983) 203. K. K. S. Hwang, K. S. Wu, S. B. Lin, S. L. Cooper, J. Polym. Sci., Polym. Chem. Ed. 22, 1677 (1984) 204. J. A.Miller, S. B. Lin, K. K. S. Hwang, K. S. Wu, P. E. Gibson, S. L. Cooper, Macromolecules 18,32 (1985) 205. C. D. Eisenbach, 35 Jahre Fonds der Chem. Indwtrie, Verband der Chem. Polym. Prep. ( A m Chem. SOC.,Indwtrie), Frankfurt 1985,p. 77 206. C. D. Eisenbach, M. Baumgartner, C1. Giinter, J. Am. Chem. SOC.,Polym. Prepr. 26,7 (1985) 207. C. D. Eisenbach, C1. Giinter, U. Struth, Spez. Ber. KFA Jiilich, 316,107 (1985) 208. Z. Y.Qin, C. W. Macosko, S.T. Wellinghoff, Macromolecules 18,553 (1985) 209. C1. Giinter, Dissertation, University of Freiburg, 1986 210. C. D. Eisenbach, C1. Giinter, M. Baumgartner, “Advances in Elastomer and Rubber Elasticity”, edited by J. Lal, Pergamon, New York 1987,p. 51 211. H. Thaufelder, Diploma Thesis, University of Beyreuth, 1988 212. C. D. Eisenbach, H. Nefiger, M. Baumgartner, C1. Giinter, Ber. Bunsenges. Phys. Chem. 89, 1190 (1985) 213. J. A. Miller, S. B. Lin, K. K. S. Hwang, K. S. Wu, P. E. Gibson, S. L. Cooper, Macromolecules 18,553 (1985) 214. L. Bom, H.Hespe, Colloid. Polym. Sci. 263, 335 (1985) 215. R.Hill, E. E. Walker, J. Polym. Sci. 3,609 (1948) 216. D. Dietrich, Houben- Weyl Methoden der Organkhen Chemie, 4th Ed., edited by H. Barth, J. Falbe, Georg Thieme, Stuttgart 1987,,vol.20/2,p. 1721 217. A. L. DiSalvo, J. Polym. Sci., Polym. Lett. Ed. 12,641 (1974) 218. G. Montaudo, C. Puglisi, “Thermal Degradation in Condensation Polymer8 , in: Comprehensive Polymer Science, edited by G. Allen, Pergamon, vol. 11, 1989,pp. 227-251 and references therein 219. S. Fakirov, “Solid State Reactions in Linear Polywndensates”, in: Solid State Behavior of Linear Polyesters and Polyamides, edited by J. M. Schultz, S. Fakirov, Prentice Hall, Englewood Cliffs, NJ 1990,pp. 1-67 and references therein

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 2

Effects of Catalysts in the Reactive Blending of Bisphenol A Polycarbonate with Poly(alky1ene terephthalate)s

F. Pilati, M. Fiorini, C. Berti

1. Introduction 1.1. A n outlook on reactive blending of polyesters and polycarbonates

Since the early 1970s, polymer blends have been recognised as a potential alternative t o copolymers for the preparation of new polymeric materials and a large research effort has been directed toward the understanding of the relationships between the chemical structure of polymers and the properties of their blends [14]. The miscibility of polymers determines to a great extent the blend properties; blends of completely miscible polymers usually behave as random copolymers of the corresponding monomeric units, with properties intermediate between those of the starting components, while immiscible polymers segregate into separate domains and show properties typical of each polymer. Although blends of immiscible polymers may in principle reveal the best properties of the single components, their characteristics (most frequently the mechanical ones) could be worse than those of the single

80

F. Pilati, M. Fiorini, C. Berti

components or than those expected for miscible blends of a given composition. Since most polymer pairs are immiscible [l-61,the poor adhesion, which in turn determines poor impact resistance and elongation at break, is a strong limiting factor to new polymeric materials. It is also recognised that block and graft copolymers may improve adhesion and other properties significantly when their different segments show preferential miscibility with the respective phases [1,3,5,7-91. It was predicted [10,11], and in some cases demonstrated [12,13], that these copolymers can migrate to the interface where they act as emulsifiers by reducing the size of phase domains and improving adhesion. However, block and graft copolymers are usually difficult to prepare, and quite expensive, and it is possible that only a part of them migrates to the interface during the short times typical of polymer processing. Reactive blending proves to be a quite fascinating route to generate in s i k , directly at the interface, block or graft copolymers, which can improve interfacial adhesion more efficiently than deliberately added compatibilisers. For this reason, reactive blending can be successfully applied to numerous polymer pairs [14-171. Among the many classes of polymers, polycondensates are interesting because of the large number of reactive sites in their skeletons. Within this class, polyesters and bisphenol A polycarbonate (PC) are particularly attractive for their good balance of cost/performance. The ester and carbonate groups can usually undergo several different exchange reactions which lead to the formation of block copolymers. However, in the temperature range typically employed for melt blending, most of these reactions proceed at a sufficiently high rate only in the presence of suitable catalysts. The aim of this chapter is to review the effects of catalysts on reactive blending of polyesters, with a particular emphasis on poly(alky1ene terephtha1ate)s and PC. An overview of the characteristics of PC and poly(alky1ene terephthalate)s, and a brief literature survey of reactive blending of polyester/PC systems precede the more detailed discussion of the effects of catalysts.

1.2. Bisphenol A polycarbonate: an overview Polycarbonates are linear thermoplastic polyesters of carbonic acid with aliphatic or aromatic dihydroxy compounds [18,19]. Commercial polycarbonates are mainly aromatic and among them PC is the most important, since over 99% of the world commercial polycarbonates are based on bisphenol A (BPA). P C is characterised by its unique combination of useful properties, such as good mechanical parameters over a wide temperature range (particularly impact strength), resistance on long-term exposure to high temperature, good electrical properties, dimensional stability, transparency, low flammability, and good resistance to aqueous agents, oils, fats, aliphatic hydrocarbons and alcohols. The main disadvantages of PC are its

Effects of Catalysts in Reactive Blending

81

relatively high cost and, mainly, its high solubility or swelling in many common organic solvents, such as benzene, chloroform, dichloromethane, etc. Aromatic polycarbonates may be prepared by many different procedures but only a few are commercially important: polycondensation in homogeneous phase, interfacial polymerisation, and transesterification in the molten state. A great effort is dedicated to the avoidance of phosgene in the respective syntheses; a promising way seems to be the use of low molecular weight aliphatic carbonate, such as dimethyl carbonate. All commercial P C grades can be prepared by interfacial polymerisation and the respective products have a higher thermal stability and a better molecular weight distribution (MWD), compared to those prepared via transesterification. However, interfacial polymerisation has some disadvantages, such as the necessity of a large amount of solvent, namely dichloromethane, and the difficulty of removing all traces of solvent and NaCl from the polymers. Interfacial polymerisation involves a reaction at the boundary between two immiscible solvents, one protic and the other aprotic (water/dichloromethane are typical industrial solvents). BPA, dissolved in the aqueous phase as bisphenate, reacts at the interface with the polymerising resin, growing into and remaining dissolved in the aprotic solvent. A phase transfer catalyst, such as benzyltriethylammonium chloride or 4dimethylaminopyridine, and vigorous stirring are required to achieve satisfactory reaction rates. To avoid the formation of very high molecular weight resins with intractable rheologies, monofunctional monomers are added as chain stoppers. The transesterification process is carried out in the molten state; in this case the carbonyl group is provided by an aromatic carbonate, normally diphenyl carbonate prepared by reacting phosgene and sodium phenolate. Acceptable rates are obtained by the use of alkali metal hydroxides, such as sodium or potassium hydroxide. Transesterification is carried out at temperatures ranging from 150°C to 320°C and pressures from atmospheric to less than 1 mbar; under these conditions the equilibrium process is shifted to polymer formation by removing the phenol formed in the exchange reaction between BPA and diphenyl carbonate. Polycondensation in homogeneous phase is carried out in a n organic solvent, such as dichloromethane, and in the presence of an organic base, such as pyridine. Phosgene is bubbled into the stirred reaction mixture, leading t o the formation of PC and pyridinium chloride. In this case also, a monofunctional monomer is added to control the molecular weight of the final polymer. The removal of all traces of pyridine and pyridinium chloride from the polymer solution and the cost of the pyridine recovery are among the major disadvantages of this process. It is noteworthy that all industrial processes lead to polymers that contain hardly any hydroxyl terminal groups, which, if present, are deleterious to the thermal stability of polycarbonates [20] (see also Chapter 1).

a2

F. Pilati, M. Fiorini, C. Berti

1.3. Poly(allcy1ene terephthalate)s: an overview Polyesters with a wide variety of properties can be obtained through changes in the molecular architecture and modification by copolymerisation. The most important in terms of volume and product value are poly(alky1ene terephthalate)s, and in particular poly(ethy1ene terephthalate) (PET), poly(buty1ene terephthalate) (PBT), and, more recently, copolyesters obtained from 1,4cyclohexane dimethanol and mixtures of iso- and terephthalate derivatives [21,22]. Due to a good balance of thermal and mechanical properties, combined with a relatively low cost, these polyesters are the most attractive also for reactive blending processes with PC, and the following discussion is limited to them and, in particular, to PET. These polyesters are usually obtained either from the polymerisation of terephthalic acid (TA) with glycols (ethylene glycol (EG), butylene glycol (BG), and 1,4cyclohexane dimethanol (CHDM), for PET-, PBTand CHDM-based polyesters, respectively) or from dimethyl terephthalate (DMT) and the same glycols. In both cases, catalysts are needed to achieve high molecular weight polyesters. A large number of catalysts can be used either for the ester interchange of DMT with glycols (derivatives of Ca, Mn, Mg, Zn, Ti, Sn, etc.) or for the polycondensation stage (derivatives of Sb, Ti, Ge, etc.) [21-231; mixtures of several catalysts are frequently employed. Since the polyesters are usually prepared by bulk polymerisation, the catalysts used remain in the products and, being still active, they can play an important role during blending, as discussed below. A further aspect t o be considered is that side reactions take place during polymerisation [21,23], leading to changes in the chemical structure, e.g., formation of diethylene glycol (DEG) moieties in the backbone chains of PET and of carboxyl and vinyl ester end-groups. The relative extent of these reactions depends on the type of polyester and on the reaction conditions. The content of DEG units in PET is usually relatively small (0.5-2 mol%), whereas the nature and relative amounts of terminal groups may be quite different, depending on the nature of the starting monomers, the time of reaction, the thermal history, and the type of catalyst employed. In the case of PET, most of the terminal groups are usually hydroxyl groups, but a relatively large amount of aromatic carboxyl groups can result from prolonged polymerisation at high temperature or in the presence of catalysts, such as Ca, Zn, Mn, and Ti derivatives, or when PET undergoes hydrolysis after polymerisation. V i y l ester end-groups, derived from thermal degradation, and carbomethoxide end-groups, from an incomplete transesterification of DMT, can also be present, but in negligible amounts. Compared to PET, the formation of carboxyl terminal group is easier in PBT, so the carboxyl-to-hydroxyl groups ratio is typically higher and can change within a broader range. In order to reduce the extent of side reactions (thermal degradation, hydrolysis, etc.), phosphorus compounds

Effects of Catalysts in Reactive Blending

83

are frequently added as catalyst deactivators during the polymerisation of these polyesters and their presence can modify the reaction rates during melt-blending. Most of the data reported in the present chapter were obtained using PET samples prepared and characterised in the authors’ laboratory; in this way the types and amounts of catalysts and of phosphorus compounds are known and all factors originating from differences in polymer production are avoided. 1.4. Btends of PC and poly(a1kylene tewphtha1ate)s: a literature survey

During the last two decades, blends of P C and poly(alky1ene terephtha1ate)s have attracted both scientific and commercial interest [24-771. The effect of melt-blending on the properties of the blends has been extensively studied. Fewer efforts have been directed toward the understanding of the reactions occurring during melt-blending. However, this understanding would provide a route to new polymeric materials by performing a controlled reactive blending or processing. Most papers and patents deal with P E T and PBT, but other products containing 1,Ccyclohexane dimethanol have also been studied; Porter and Wang have reviewed this subject in a recent paper [56]. A survey of the literature published on PC/PET and PC/PBT blends is reported below. 1.4.1. PC/PET blends

The early work on PET/PC blends by Paul reported a single glass transifor blends containing at least 60-70 wt% PET, while tion temperature (T,) compositions below this range showed two glass transitions [25,26]. These findings were interpreted as evidence for a complete miscibility in the amorphous state for PET-rich blends, whereas PC-rich blends separate into two amorphous phases which apparently contain both polymers. It was also suggested that very little if any interchange reaction occurred between ester and carbonate groups during melt-mixing in a Brabender apparatus at 290-300OC. In another study [34], blends prepared using a single-screw extruder at a die temperature of 290°C appeared to contain two amorphous phases over the entire composition range. Transreactions were not taken into consideration, although after annealing the blends showed improved Charpy impact strength and solution extraction data suggested copolymer formation. In the absence of significant transesterification, other authors [39] also reported an almost complete immiscibility in PC/PET blends, while in 1471,dealing with blends prepared by solution-casting and extrusion, the conclusion is drawn that PET dissolved to a greater extent in the PC-rich phase than did P C in the PET-rich phase.

84

F. Pilati, M. Fiorini, C. Berti

The most probable reason for these ambiguous conclusions is disregard for the effects of transreactions that occurred to different degrees. The occurrence of transreactions during melt-blending of PC and PET was recognised as early as in 1985 [35), and the effect of the different reactions occurring during melt-mixing has been thoroughly investigated more recently [37,40,41,54-561. Techniquessuch as differentialscanning calorimetry (DSC), infrared (IR) and nuclear magnetic resonance (NMR) spectroscopy, selective solubility, mass spectrometry, selective chemical attacks, etc., were used to identify the effects of transreactions and to investigate the chemical modification induced. Godard et al. [40,41] concluded that the main reaction was the direct ester-carbonate exchange, but at a lower rate with respect to PC/PBT blends, probably because of the different nature of the catalysts used in the polyester synthesis. In addition, they reported that some other reactions may also take place, such as decomposition of the alkylene carbonate. By using a direct pyrolysis-mass spectrometry technique, Montaudo et al. (371 observed the formation of decomposition products such as carbon dioxide, cyclic ethylene carbonate and its cyclic dimer in the temperature range typical of melt-blending. The importance of side reactions, such as COa and ethylene carbonate release, for the control of the chemical structure of the materials obtained after melt-blending was examined by Berti et al. 154,551. The effect of the catalyst on the rate of transreactions was first reported in 1985 [35]; it was confirmed by Porter who compared the different results obtained in the presence or absence of residual catalyst and concluded that no ester exchange reactions occur in the absence of catalyst [48,49]. The effects of different residual catalysts in PET and of catalysts added to PET and P C before blending were thoroughly studied [70-74,771 and are discussed below (see also Chapter 8).

1.4.2. PC/PBT blends Although the first patents covering PC/PBT blends were issued in the early 1970s [24], the first scientific paper on the miscibility of PC and PBT was by Paul et al. in 1978 [27]. They reported that after blending for 10min in a Brabender Plasticorder at 250°C and compression moulding of test specimens at the same temperature, PC/PBT blends still showed multiple glass transitions with some deviation from the values observed for pure h u mopolymers. These authors concluded that the PC/PBT blends were phase separated, with a partial degree of miscibility in the amorphous phases, but reactions which might have taken place during mixing were not mentioned. The importance of exchange reactions during the mixing of P C with PBT at high temperature was first reported by Mercier et al. [28-321 and in several papers these authors investigated the mechanism of exchange reactions also by using model compounds, characterised the products formed

Effects of Catalysts in Reactive Blending

85

by spectroscopic methods, and performed a kinetic study of the exchange reactions. They concluded that the most likely mechanism was a direct reversible ester-ester interchange reaction catalysed by titanium residues present in commercial PBT. Later, the partial miscibility of a commercial PC/PBT blend was investigated; the conclusion was drawn that the microstructure of the blend was sensitive to the processing conditions because of the ongoing transesterification [34]. The effect of exchange reactions on the morphology was also studied [52]. Other authors [53] extended the study to different processing parameters, such as the level of Ti-based catalyst, temperature, and intensity of mixing, and showed that in order to achieve the high reaction rate required in reactive extrusion, Ti levels of 300ppm and temperatures in the range 260-280OC were required. The effect of the amount of added titanate was studied by DSC and scanning electron microscopy (SEM) [67];the size of the bicontinuous morphology became finer and finer with increasing levels of added titanate. Montaudo et al. [51], in a study of PC/PBT blends by direct pyrolysismass spectrometry, reported the formation of cyclic butylene carbonate along with a series of cyclic compounds containing units of PC and PBT in varying ratios (see also Chapter 8). 2. Possible reactions occurring during melt-mixing of polyesters and PC

The high temperatures required for blending polyesters and PC (T 2 270°C) and the presence of catalysts, residual from the polymerisation or added on purpose, create conditions for a number of possible reactions to take place during melt-mixing. Ester and carbonate groups in the chains of polyesters and PC, respectively, and reactive terminal groups in polyesters are mainly responsible for inter- and intramolecular exchange reactions. However, under the severe conditions of blending, several other side reactions can take place, thermal degradation and release of carbon dioxide from carbonate groups being the most important. Many of these reactions are characterised by significantly high rates (compared to the residence time in the melt-mixing process) only in the presence of active catalysts, and the addition of suitable catalyst inhibitors can limit or eliminate completely the effects of these reactions. A brief summary of the main exchange and side reactions that have to be considered in melt-mixing processes of polyesters and PC is given below along with a survey of the main catalyst inhibitors.

86

F. Pilati, M. Fiorini, C. Berti

2.1. Exchange reactions Exchange reactions can occur inter- or intramolecularly, and can involve ester, carbonate, hydroxyl, carboxyl, and (in principle) also phenol groups.

&

43

----.,

l l l l l l l l l l l l l l l l l l l ~

~IIIIIIIIIIIIIIllllllllllllllllll 11111111111111111111

,--a'

!

V

111111111111111

1111111111111111

P 111111111

U

0

a

Figure 1. Schematic representation of exchange reaction between: (a) midchain ester and carbonate groups; (b) PET end-groups and carbonate groups

In respect to the modification of the chemical structure, the most important exchange reactions are those involving reactive groups of both polymers, polyester and PC; they are represented schematically below for a PC/PET system, as well as in Figure 1. 1) Reaction of hydroxyl terminal groups of polyesters with carbonate groups of PC (alcoholysis): 0

@b-OCHzCHzOH+

0PC

0

0-C-0 "

0

-0 PC

+ 0

2) Reaction of carboxyl terminal groups of polyesters with carbonate groups of PC (acidolysis):

-

Effects of Catalysts in Reactive Blending 0

@COOH

+@-o-!-o@ 0

l@-h-O@

87

0

+ @-0-!-OH

3) Reaction of ester groups of polyesters with phenol terminal groups of PC (phenolysis):

0

0

4) Reaction of ester groups of polyesters with carbonate groups of PC: 0

0

+

0

5) Intramolecular exchange reactions with the formation of cyclic alkylene carbonates:

F. Pilati, M. Fiorini, C. Berti

88 0

0

-@+

~ ~ - O C H z C H z OC-0 I'-

0

I1

C

@LO@+

/ \ 0 0

I

I

6) Intermolecular exchange reactions in polyesters (ester-ester, esterhydroxyl, ester-carboxyl) and in polycarbonates (carbonate-carbonate, phenol-carbonate): @

x

a

+

@

a

b

-

t

@@ -' b+

0

0

II

P=PET, X=Y= -C-0

@@!I

or

II

P = P C , X=Y=O-C-0

P= PET, Z = CH2CH2OH or COOH; P = PC, Z = PhOH The above reactions have different effects on the modification of the chemical structure. Reactions 1-3 lead to the formation of one molecule of block copolymer (and one joining point between segments of polyester and PC) along with the formation of one shorter molecule of polycarbonate or polyester. Due to the lower nucleophilicity of the phenol groups with respect to primary hydroxyl groups, Reaction 3 is expected to be slower than Reaction 1. Studies on model compounds have actually shown that this reaction does not occur [28,31] and, taking also into account that a large fraction of the chain ends of commercial PC are usually capped with alkylated aryl groups deriving from the use of stoppers during the polymerisation, it can be neglected or considered of minor importance. Reaction 2 has been demonstrated to occur using low molecular weight analogues [28,31]. For various polyesters, and in particular for PET, the concentration of carboxyl terminal groups is low and Reaction 2 has a minor effect on the modification of the chemical structure. However, because carboxyl terminal groups can be generated by the thermal degradation of polyesters (see below), in some

Effects of Catalysts in Reactive Blending

89

cases it can play a more important role than expected from the initial characteristics of the polyesters. However, it has been demonstrated recently that Reaction 1 can occur faster than the ester-carbonate exchange and its contribution can be very important for the improvement of miscibility in the very first stage of blending [73]. In contrast to Reactions 1-3, Reaction 4 results in two molecules of block copolymers (and two joining points between polyester and P C segments) and no homopolymer. It is interesting to note that if one considers the effect of Reaction 3 to be negligible, the decrease of the average length of the polyester chains due to exchange reactions can occur only by estercarbonate exchange reactions. While the main result of Reactions 1-4 is the formation of block copolymers, intramolecular exchanges, such as those schematically represented as Reaction 5, lead to the formation of low molecular weight aliphatic cyclic carbonates and to a depletion of carbonate and alkylene groups in the blend, if their removal is favoured (54,55,78]. It has been demonstrated by neutron scattering that exchange reactions occur rapidly in P E T at the high temperature required for melt-mixing [79,80]. Similar exchange reactions are expected to occur for other types of polyesters, as well as for PC. Their effects on the chemical structure are limited t o molecular weight redistribution, which could become important when the extent of Reactions 1, 2, and 4 increases. In fact, when polyesters and P C are immiscible, Reactions 1, 2, and 4 occur initially at the interface, where domains containing mainly block copolymers are formed. In these domains, block copolymers can react with each other more easily than with P C or polyester homopolymers, so that the average segment lengths in block copolymers would decrease quite rapidly. Reaction 6 between polyesters and P C homopolymers with the corresponding polyester and P C segments in block copolymer molecules may contribute to a faster redistribution of molecular weight, of the number of segments per molecule of block copolymer, and of the segment lengths in the overall blend. When Reactions 6 occur intramolecularly, cyclic oligomeric esters or carbonates can be formed [21-23,81,82]; however, due to the reversibility of these reactions and the relatively low volatility (high molecular weight) of these products, cyclic oligomers reach an equilibrium concentration which is usually low and this reaction can be considered of minor importance for the control of the final chemical structure. Of course, the result of melt-mixing in terms of chemical modification depends on the relative rates of the above reactions (and of the side reactions described below) which in turn depend on the relative reactivity and concentration of the various reactive groups. In this respect, the miscibility of the starting polymers may influence significantly the reaction rates [60,71-731. It has been demonstrated that most of the above reactions occur only in the presence of a suitable catalyst; no exchange reactions took place when

F. Pilati, M. Fiorini, C. Berti

90

the residual catalyst was removed [48] or when an appropriate inhibitor was added (28,56,70,71]. Therefore, the resulting chemical structure will be strongly dependent on the type and concentration of catalysts and/or inhibitors. Even though some mechanisms have been proposed for metalbased catalysis of exchange reactions [23], the effects of the catalysts on the relative rates of the various reactions is still an open question. Achieving knowledge of the catalytic activity toward the different reactions would be very important for the control of the resulting chemical structure, but it is also a formidable task because of the number of reactions that can take place simultaneously, ignorance of the location and local concentration of the catalyst, the possible change in the catalytic activity during the process, and the increasing miscibility with the progressive formation of block copolymers. For these reasons the effect of catalysts on the reactive blending has usually been studied using an empirical approach, i e . , by examining the overall effects of various catalysts. 2.2. Side reactions

As already mentioned, because of the high temperature required for meltblending of PC with polyesters, the effects of side reactions, such as thermal degradation of polyesters and carbon dioxide release, may become considerable. They are represented schematically below for PET/PC systems: 7) Thermal degradation of polyesters:

8) Carbon dioxide release from aromatic-aliphatic carbonate groups:

0

!@-8-OCH2C&O-C-O@

0

II

+ 0

9) Carbon dioxide release from PC terminal groups:

Effects of Catalysts in Reactive Blending

@-O-k-OH

91

+

The thermal degradation of polyesters (Reaction 7), and particularly of poly(alky1ene terephthalate)s, has been extensively studied [23,83-851. The mechanism involves a six-membered cyclic intermediate with a @-hydrogen elimination; the main results of this reaction are a decrease in the molecular weight and the formation of carboxyl and vinyl ester terminal groups. The reaction rate depends on the chemical nature of the polyester and can be increased in the presence of residual catalyst [83]. This reaction is usually characterised by a high activation energy (about 190kJ/mol) and as a consequence its rate is strongly influenced by temperature. While thermal degradation of polyesters takes place at an appreciable rate in the temperature range usually employed for PC/polyester meltmixing (270-30ODC), polycarbonates are more thermally stable under the same conditions and the effects of thermal chain scission during reactive blending can usually be neglected [86]. Another side reaction that can affect the final chemical structure is carbon dioxide release from aliphatic-aromatic carbonate groups (Reaction 8), leading to the formation of ether groups in the backbone chains. It can occur after the formation of aliphatic-aromatic carbonate groups by exchange reactions (joining points between polyester and PC segments), and plays a significant role only at high degrees of exchange. Finally, carbon dioxide release can occur according to Reaction 9, quickly after the formation of carbonic acid groups from Reaction 2 [28,31]. The main effect in this case is a depletion of carbonate groups in the bulk. The rates of Reactions 8 and 9 are probably unaffected by the catalysts commonly present in the reaction medium. 2.3. Catalyst inhibitors When exchange reactions are not desired, the residual catalysts must be deactivated by using suitable additives (catalyst quenchers) [87-941. They have been widely applied not only in industry in order to prevent undesirable exchange reactions [87,88], but also in fundamental research to avoid the effects of exchange reactions on the results of the characterisation studies; exchange reactions could occur during the characterisation test itself, e.g., during DSC experiments [28,56,70,71] or rheological measurements. The catalyst inhibitors commonly used are organic phosphites and, in particular, di-n-octadecylphosphite (DNOP) and triphenylphosphite [28]. For commercial blends, such as PC/PBT products, phosphorous acid is preferred since it is less expensive and can be more easily added during blend preparation by extrusion [53,88].

92

F. Pilati, M. Fiorini, C. Berti

Although the effectiveness of these inhibitors has been clearly demonstrated, the mechanism of their interaction with the catalyst is not completely understood. In general, it has been assumed that phosphites interact with titanates by exchanging one or more ligands and chelating the metal atom in some way and that the complex formed is no longer available to exchange for alkoxides, thus preventing further catalytic action by the metal. The hydrolysis of phosphites to different phosphonates in polyester and polycarbonate blends has been studied by solid-state 31PNMR [89-941; in blends prepared under anhydrous conditions, the exchange reactions were still taking place, whereas in blends prepared under less rigorous conditions or when the polymers were deliberately exposed to humidity before blending, transesterification was completely suppressed. The optimum molar ratio with respect to the catalyst is not clearly defined. Kollodge and Porter I721 reported that an inhibitor-to-metal catalyst molar ratio of 5:l was the lowest one leading to complete inhibition. Molar ratios ranging from 1 to 5 have been reported to be necessary to achieve complete deactivation of the catalyst [60,70,72](see also Chapter 6).

3. Evolution of the chemical structure during melt-mixing Reactions 1-9 described above can occur simultaneously during meltblending and their relative rates determine the chemical structure of the resulting products. The most important effects are expected from exchange reactions which lead to the formation of block copolymers. As the extent of these reactions increases, an increase of the amount of block copolymer and a progressive decrease in the average length of the blocks occur; in principle, a random copolymer would be obtained for very long reaction times. The decrease in the average length of the polyester segments can occur only by Reactions 4 and 7; since the rate of Reaction 7 is known for the information about the most common poly(alky1ene terephtha1ate)s [83-851, rate of Reaction 4 can be obtained from a knowledge of the rate of decrease of the PET segment length (even though a quantitative kinetic treatment of the data is difficult due to the generally unknown effects of miscibility and morphology changes during melt-mixing). Knowledge of the rate of change of the average length of the P C segments would also be of great importance to get information about the relative rates of Reactions 1, 2, and 4;unfortunately, although it is possible to measure the average length of polyester segments (see below), the average length of PC segments is still unmeasurable (see also Chapter 8). Even though the main effects on the chemical structure and on properties are expected from exchange reactions, some features of the final chemical structure can originate from side reactions. These can be used to prepare polymers with molecular structures which cannot be easily obtained

93

Effects of Catalysts in Reactive Blending

by polymerisation of monomers. It is obvious that the possibility of changing the relative rates of the various reactions would allow the control of the final chemical structure, and hence the preparation of a large number of different polymeric materials from the same starting polymers. Therefore, the appropriate choice of catalysts that are active in different reactions is a potentially attractive route to the preparation of various polymeric materials from the same starting polymers. Before discussing in more detail the effect of catalysts on the chemical structure of the products that can be obtained by melt-mixing of polyesters and PC, it is interesting to comment on the methods available for chemical characterisation. Most of these methods are described in detail in other chapters of this book, so the discussion here is limited to the specific case of polyesters and PC.

3.1. Approaches to the investigation of the resulting chemical structure

As long as the chemical structure changes during blending, a number of changes are induced in the IR and NMR spectra, thermal properties, miscibility, molecular weight and MWD, etc. Therefore, several techniques can be used to detect these chemical modifications. There is no single technique capable of providing detailed information about the structural changes and most techniques give only qualitative information. Their proper choice should be based on the type of information desired and on the extent of the exchange reactions; for an extensive chemical characterisation, the use of several techniques is usually preferable. A brief critical discussion of the most common techniques is reported below. 3.1.1. Spectroscopic methods The most obvious first approach to detecting changes in the molecular structure of the polymer chains is by spectroscopic methods, and in particular NMR and IR spectroscopy. The different wavenumbers of the carbony1 stretching bands of polyesters and PC (- 1720 and 1774cm-', respectively) have been used frequently to obtain information about exThe formation of aliphaticchange reactions [28,30-32,42,46,48,49,54,65]. aromatic carbonate ( m 1764cm-') and aromatic-aromatic ester groups (- 1740cm-') as a result of exchange reactions, would in principle make it possible to follow the progressive decrease of the initial ester and carbonate groups and the increase of the reaction products. Indeed, a progressive change of the spectrum in the carbonyl stretching region (1700-1850cm-'), from two distinct peaks to a single broad peak has been observed for PC/PBT [28,30-32,53,67] and for PC/PET [42,46,48,49,54,65]systems (see Figure 2). However, the relatively low sensitivity of IR spectroscopy, the overlapping of the relatively wide carbonyl bands and the effects of side

-

F. Pilati, M. Fiorini, C. Berti

94

aJ

8

Y

3

B

P

-

c

fr ,? I

1800

1600

1800

1800

1600

Wavenumber in cm-'

1600

Wavenumber in cm-'

Wavenumber in cm-'

1800

1600

Wavenumber in cm-1

1800

1600

Wavenumber in cm-'

Figure 2. FT-IR of: (a) PC;(b) PET;samples from melt-mixing of PET and PC in a Brabender Plasticorder at 270°C for 30min (c); at 270°C for 60min (d); at 270°C for 9Omin plus 30min at 275°C under vacuum (e)

reactions (such as COa release) make the quantitative interpretation of the different contributions almost impossible. As a consequence, IR spectroscopy can be used just as a qualitative technique to demonstrate the occurrence of exchange reactions;what is more, this technique is applicable

Effects of Catalysts in Reactive BIending

95

only when the extent of these reactions is sufficiently high. In some cases, IR spectroscopy can also be used to detect the formation of by-products, such as cyclic ethylene carbonate (a sharp peak at 1811 cm-') in PET/PC systems (Figure 2) [54,55]. NMR provides richer information [28,30-32,40,41,45,48,49,71-731: both lH and 13CNMR have been used to follow the progress of the reaction from a simple blend of homopolymers, via block copolymers with shorter and shorter segments, to random copolymers. The change in the chemical shift due to the replacement of aliphatic moieties by aromatic ones (bisphenol A) on the terephthalate units is sufficiently large to allow the estimation of the relative amount of aliphatic-aliphatic, aliphatic-aromatic and aromaticaromatic terephthalate units, and the degree of blockiness can be derived directly on the crude product [28-32,40,41,71-731. NMR is also very useful for obtaining information about the extent of side reactions [40,41,45].This technique is limited by its low sensitivity to the very first formation of block copolymers, when the concentration of the joining links between PC and polyester segments is very low (but high enough to modify the properties of the blends [71-731).

3.1.2. Selective solubility Evidence of the occurrence of exchange reactions is the change of solubility. Since most poly(alky1ene terephtha1ate)s are far less soluble in common solvents than PC, it is usually possible to find a solvent dissolving P C selectively from PC/polyester blends. As a consequence of exchange reactions, the fraction soluble in the PC solvent usually changes in amount and composition as the reactions proceed; a typical curve of solubility us. reaction time is shown in Figure3. It can be obtained by extraction of a sample with the P C solvent or, better, by dissolving the sample in a common solvent and reprecipitating the polyester-rich fraction in a typical PC solvent. The solubility of the block copolymers formed as a consequence of exchange reactions is dependent on the length and on the relative weight of the blocks in a given macromolecule. For both PBT/PC [28,30] and PET/PC [35,74,77] systems, it was found that the fraction miscible in dichloromethane decreased at first and then increased with the shortening of the blocks, following the progress of exchange reactions. The formation of block copolymers with segments of PC insoluble in dichloromethane is the dominant effect in the first stage of reaction, while the increasing solubility that accompanies the decrease of the average length of the polyester blocks explains the progressive increase of the soluble fraction for longer reaction times. The combination of selective solubility with spectroscopic methods (e.g., 'H NMR analysis of the soluble and insoluble fractions) is highly recommended because it helps to detect block copolymer formation even in that particular hypothetical case when the number of polyester segments soh-

96

F. Pilati, M. Fiorini, C. Berti

20

0

,

~

,

~

.

[

,

~

Figure 3. Selective solubility of products prepared using PET with Ti as residual catalyst: -0- with DNOP added; 4without DNOP (PET/PC 50:50 by wt, Brabender mixer, 275°C and 30 rpm)

ble in the PC solvent compensates exactly for the number of PC segments insoluble in the same medium. Because the difference in soluble and insoluble fractions is a complex function of chemical structure, composition and molecular weight, this method provides only qualitative evidence for the occurrence of exchange reactions, but it is quite sensitive to formation of the very first block copolymer.

,

~

,

Effects of Catalysts in Reactive Blending

97

A typical procedure for the selective solubility tests on PET/PC systems is as follows [70]:to a sample of about 1 g (accurately weighed) taken from each blend, 25 ml of dichloromethane and 1ml of trifluoroacetic acid are added and magnetic stirring is started. As soon as complete dissolution is achieved, precipitation is started by dropwise addition of 75 ml of tetrahydrofuran. When precipitation is completed, the insoluble fraction is separated by filtration, washed with a mixture of dichloromethanetetrahydrofuran (25/75, v/v), and finally with tetrahydrofuran. The soluble fraction is recovered by evaporation of the solvent under reduced pressure after filtration of the insoluble fraction. Soluble and insoluble fractions are finally dried at 100°C under vacuum and weighed. This method was tested on simple blends of P E T and PC obtained by mixing in the Brabender apparatus in the presence of DNOP (0.5 wt%), as catalyst inhibitor of ester exchange reactions, and quantitative separation of PET and P C homopolymers is achieved (the difference between the initial and recovered weights was usually f 5 % ) , even when they were thoroughly mixed [70] (see also Chapter 8). 3.1.3. Selective degradation

Selective degradation based on chemical attacks, destroying completely one of the polymers (and the corresponding segments in block copolymers) in a blend, and leaving unchanged the other polymer (and the corresponding segments in block copolymers), is very helpful and desirable for the characterisation of block and graft copolymers in general. In the particular case of polyester and P C blends, use is made of the different reactivities of ester and carbonate groups toward nucleophiles to depolymerise selectively the PC segments and characterise the remaining PET segments. Nucleophiles, such as piperidine [35] or ammonia, were found t o react much more rapidly with carbonate groups than with ester groups. Under appropriate reaction conditions, this method makes it possible to recover and characterise the segments of polyesters present, in terms of average molecular weight and MWD, in the crude product obtained after reactive blending. In fact it was found that for PET/PC and P B T /P C systems, piperidine destroys the PC segments quantitatively, according to Scheme 1. Bisphenol A and the urea and urethane derivatives, formed as a result of this reaction, are soluble in various solvents and can be easily separated from the residual insoluble PET (or PBT), which is recovered quantitatively for molecular weight and MWD measurements. This method is particularly valuable for the characterisation of the very first stage of exchange reactions, when the quite limited number of joining points between P C and polyester segments is not detectable by NMR spectroscopy. As the exchange reaction between ester and carbonate groups in polyester/PC reactive blending is accompanied by a large decrease in

98

Q

8

F. Pilati, M. Fiorini, C. Berti

0-c

I

0

6

O=b

I

8 0

I

CHz Clz

+

o=cI I

0

0

with R = H , -C-

0I

Q

o=c

Scheme 1

the average molecular weight of the polyester segments, molecular weight measurements of the residual polyester (by intrinsic viscosity, size exclusion chromatography (SEC),etc.) provide information about the rate of Reaction 4. The method is also very helpful for correlating properties with the average length of polyester segments.

Effects of Catalysts in Reactive Blending

99

While this method is probably the most effective for the detection of the initial formation of block copolymers by ester-carbonate exchange reaction, it may become inappropriate at high degrees of exchange, when the solubility of shorter and shorter polyester segments limits or makes difficult their quantitative recovery. A method of complete destruction of polyester segments, leaving unchanged PC segments, would be highly desirable for a more extensive chemical characterisation, but, to our knowledge, such a technique has not been reported so far. A typical procedure of selective degradation of P C blocks in PET/PC systems is as follows [70]: to a sample of about 1 g (accurately weighed), 25 ml of dichloromethane and 1ml of piperidine are added under magnetic stirring. The reaction of carbonate groups with the aliphatic amine is monitored by following the disappearance of the polycarbonate band in the IR spectra of samples taken from the solution at various times. Once degradation is complete, 100ml of methanol is added dropwise, to ensure the complete recovery of even the shortest PET segments. The PET residue is then filtered off, washed with methanol, dried under vacuum at 100°C overnight and weighed. The effectiveness of the degradation procedure was tested on samples obtained by melt-mixing P E T and P C in the presence of DNOP [70]. For the weight fraction recovered, the 'H NMR spectrum and the intrinsic viscosity of the residue confirm that piperidine attacks P C selectively. In fact, the amount of the residue is identical to the initial amount of PET, there is virtually no bisphenol A left in the polymer residue (as shown in I R and NMR spectra), and only a slight decrease of the intrinsic viscosity of P E T is observed (e.g., the intrinsic viscosity of P E T decreased from 0.82 to 0.78dl/g) (see also Chapter 8).

3.1.4. DSC technique

Thermal properties, such as Tg and melting temperature (T,), are frequently used to characterise heterogeneous systems when their respective values for the two polymers differ substantially. This is usually the case of polyester/PC blends, and therefore it was a logical consequence to apply the well-known DSC technique to the study of reactive blending of polyesters and P C [38,42,45-47,50,58,62,64,65,67,68]. A summary of the basic principles of DSC for the characterisation of reactive blending has been reported recently [56]. Changes in T', T,, crystallisation temperature (T,), and enthalpies of melting (H,) and crystallisation (H,) have been extensively used as evidence of exchange reactions. However, these parameters depend on both thermodynamic and kinetic factors other than molecular structure, and can provide only qualitative information about the occurrence of reactions. The method has a particularly low sensitivity in the first stage of reaction, when a limited fraction of the interfacial volume is involved in exchange reactions, and also at very high degrees of exchange, when the system appears homogeneous and crystallinity disappears completely (see also Chapter 8).

100

F. Pilati, M. Fiorini, C. Berti

3.1.5. Other techniques Even though the above techniques are the most widely used, several other methods have been employed to obtain information about the occurrence of chemical reactions in PC/poly(alkylene terephthalate) blends. The formation of block copolymers and the consequent interaction of the different blocks with the corresponding PC and polyester domains in the molten state lead to changes in the rheological properties, recorded as a variation in the torque during blending. Usually the torque increases at first, reaches a maximum and then decreases [70]. A thin-layer chromatography-infrared method has also been developed to identify the products of transesterification [49]. Chemical methods, such as alcoholysis combined with some analytical technique (e.g., chromatography or mass spectroscopy), can be used to obtain information on the details of the chemical structure, in particular at a high extent of exchange [54]. Light microscopy, scanning and transmission electron microscopy, etc., could also provide evidence of changes in morphology induced by exchange reactions [47,66-69,76]. Dynamic mechanical analysis can be used to get information about thermal transitions and mechanical properties [42,58,62,67,76]. Mass spectrometry has also been used in the study of exchange reactions in polymeric systems [37,51,57]. This technique allows the thermal treatment of the polymer blend directly in the ion source of the instrument, so that the evolving products are ionised and continuously detected by repetitive mass scans almost simultaneously to their formation. The detection of products containing the repeat units of the two polymers forming the blend under investigation has been presented as strong direct evidence of the occurrence of exchange reactions. 3.2. Eflects of catalysts in the wactive blending of PC/polyester systems The catalyst used for the bulk polymerisation of polyesters is usually still active toward most of the reactions described above. Initially, the catalyst is present in the polyester domains and is absent in the PC domains; only the portion of the catalyst present at the interface could affect exchange reactions. This fact suggests an alternative approach where the catalyst is added on purpose to polyesters and PC before blending, so that most of it is present at the interface. Good miscibility of the catalyst in the polymers is required; however, the fast and preferential miscibility of a given catalyst in one of the polymers may limit the expected catalytic effects. The role of different catalysts (residual or deliberately added) on the chemical structure resulting from melt-blending of polyesters (mainly PET) and PC are discussed in the following sections.

101

Effects of Catalysts in Reactive Blending

3.2.1. Eflects on the chemical structure of residual catalysts i n polyesters

A residual catalyst is always present in commercial polyesters prepared

by bulk polymerisation, except when it is extracted by dissolutionreprecipitation methods 1481. Unfortunately, in most papers dealing with polyester-polycarbonate reactive blending, the polyesters used are commercial samples, where nature and concentration of the catalyst are unknown and where a catalyst inhibitor was probably added. For this reason, it is not easy to compare the results reported by different authors, and the following discussion is mostly based on our work on PET/PC systems, where the PET samples were prepared with a known amount of catalyst; a list of the catalysts used is reported in Table 1. In contrast to PET, residual metal catalysts in PC are present only when it is prepared by high-temperature polymerisation; since we used commercial PC prepared by interfacial polymerisation, it is assumed to be free of residual catalyst. Table 1. Catalysts used for reactive blending of PC and PET

+ +

Ca(ac)z Sbz03 Mn(ac)n Sbz03 Ti(0Bu)d Ti(0Bu)r Ca(ac)z Ti(0Bu)r Mn(ac)z Ti(0Bu) llMn(C0)s

+ +

+ SbzO3 + SbzO3 I

Ph=

Tb(~)3 Tb(acac)3 .diPy

-00

diPy =

Er (ac)3 E~(NO~)~.BIZC~

(ac) =acetate; (acac)= acetylacetonate;

OBu = n-butoxide; B12C4 = benzo-12-crown-4

102

F. Pilati, M. Fiorini, C. Berti

Catalytic activity in the formation of block copolymers. The first evidence of different behaviour by the various catalysts contained in PET used for reactive blending with PC was observed by comparing the signals of the torque recorded during melt-miuing in a Brabender apparatus. These traces, recorded under the same conditions (amount of polymer blend loaded into the mixer, temperature, rpm, catalyst molar concentration) [70], are shown in Figure4. For most catalysts, the torque increases in the initial period of reaction, passes through a maximum, then decreases until a constant or almost constant level is reached. In contrast, when DNOP was added as catalyst inhibitor, there was no significant change in the torque (Figure 4, curve (a)). It is clear that the torque profiles and the different positions of the maxima reflect different chemical histories and indicate that the reaction path strongly depends on the type of catalyst used. More information about changes in the chemical structure was obtained from NMR spectroscopy. Some typical 'H NMR spectra of the crude products taken from the Brabender after 20min of mixing at 270°C are shown in Figure5. While for most catalysts the spectra were identical to that of a simple blend of unreacted PET and PC (spectrum (b)), spectrum (a) obtained with the sample containing Ti as residual catalyst differs significantly from that of the simple blend. The effect of residual Ti(0Bu)l in PET/PC system was described in [40,41,54,55]:Ti(OBu)4 was found to be a very efficient catalyst for estercarbonate exchange reactions, as visualised by strong 'H NMR signals of aromatic-aliphatic and aromatic-aromatic ester groups, and by changes in the aliphatic moieties of the polyester (4.7-4.8ppm) and in the aromatic (7.1-7.4ppm) and aliphatic (1.6-1.8ppm) signals of bisphenol A. The presence of aromatic-aliphatic ether moieties (6.8-7.2 and 4.5-4.7ppm) and of cyclic ethylene carbonate is also evident in Figure 5a [54]. For the other catalysts tested, the 'H NMR spectra were very similar to that obtained from a simple mixture of unreacted PET and PC (Figure 5b) and are therefore consistent both with a simple blend of PET and PC homopolymers and with PET-PC block copolymers with a very limited number of block junctions. The selective solubility method (based on the dissolution of the samples into dichloromethane/trifluoroacetic acid mixture, 25/1 (v/v)), followed by reprecipitation with tetrahydrofuran [70], combined with 'H NMR analysis of soluble and insoluble fractions can provide more information about the formation of PET-PC block copolymers than 'H NMR alone. In fact, the presence of PET in the soluble fraction and of PC in the insoluble one proves the formation of block copolymers. If no reaction had occurred between PC and PET, they would have been found entirely in the soluble and insoluble fraction, respectively [70]. It seems reasonable to assume that the content of PET in the soluble fraction and of PC in the insoluble one is higher when the extent of ex-

Effects of Catalysts in Reactive Blending

103

Figure 4. Torque signal measured in the Brabender mixer during reactive blending of PET/PC blends (50:50 by wt) at 27OoC and 30 rpm. Residual catalyst in PET: (a) Ca/Sb with DNOP added; (b) Tb(acac)j.diPy; (c) Er(N03)3.B12C4; (d) Ca/Sb; (e) Ce(ac)s; (f) EuL3; (g) SmL3; (h) Ti(0Bu)r [70]

F. Pilati, M. Fiorini, C. Berti

104

I

Y

'

4

0

PPm

Figure 5. 'H NMR spectra (200 MHz) after 20 min of mixing at 270°C in the presence of (a) Ti(0Bu)a and (b) unreacted PET and PC. ALI other catalysts gave spectra almost identical with spectrum (b) [70]

change reactions increases because the initial unreacted blend of PET and PC is progressively transformed into block copolymers (with shorter and shorter segments of PET and PC). Therefore, for the same reaction time, the amount of PET in the soluble fraction and of PC in the insoluble one are expected to be higher at a higher catalytic activity. Based on these arguments, one can obtain some information about the catalytic activity by comparing the compositions of soluble and insoluble fractions after 20 min of melt-mixing at 270°C. These data (Table 2) indicate that various catalysts lead to different products and suggest the following order of catalytic activity: Ti(OBu)4 >> SmL3 > EuL3 > Ca Sb > Ce(ac)s Er(N03)3.B1&4 > Tb(acacf3.diPy 0. By extending the selective solubility approach to different reaction times, the data reported in Figures6, 7 and 8 are obtained. As seen in Figure 6, the soluble fraction first decreases, reaches a minimum and then increases to complete solubility. It is clear that a single data point could be misleading because the same soluble fraction can be obtained for two different times. The entire curve of soluble fraction us. reaction time can better discriminate between different catalyst behaviour and the initial slope of

-

+

N

105

Effects of Catalysts in Reactive Blending

Table 2. Characterisation data of samples obtained by melt-mixing of PET and P C (50/50 by wt) for 20min at 270°C [70] Intrinsic P E T in PC in PET in P C in Intrinsic viscosity Soluble soluble insoluble soluble soluble viscosity of PET fractiona fractionb fractionb fraction' fractiond of P E P segments' (wt%) (wt%) (wt%) (wt%) (wt%) (dl/g) (dl/g)

Catalyst code

Ca/Sb 35.5 29.0 Ce(ac)3 Er(N03)3.B&* 43.4 EuL3 28.3 SmL3 19.9 Tb(acac)J.diPy 50.7 Ti(0Bu)r 100

5.8 7.3 4.1 11.0 26.7 -

37.3 31.7

15.6 33.7 44.1 -

50

4.8 3.6 3.6 6.2 10.6

41.6 17.5 17.5 48.9 70.7 -

-

100

-

0.82 0.79 0.74 0.80 0.69 0.91

0.13 0.47 0.42 0.30 0.22 0.63 0.09

aIn THF/CH&lz/TFA (75/25/1, by vol.) bFrom 'H NMR spectra 'With respect to the entire amount of P E T "With respect t o the entire amount of PC e,fIn phenol/TCE (40/60) at 25OC, ebefore reactive blending, 'after selective degradation

these curves is probably a better indicator to order and quantify the catalytic activity toward exchange reactions. However, for very low extents B

R

1004 100-

-1

I

?580

20-

0

I

I

10

I

20

I

30

I

1

I

I

40

50

Reaction time (min)

'

I

60

Figure 6. Change in the percentage of the soluble fraction as a function of time for different catalysts: 4- Ti; 4Sm; -0- Ce; -A- Er; Eu; -0- Ca/Sb [70]

-+-

F. Pilati, M. Fiorini, C. Berti

106

of exchange or when the change in solubility is very fast, it may become difficult to discriminate between catalysts. Alternative methods based on lH NMR analysis of the soluble and insoluble fractions can be used. Curves of the percentage of PET and PC in the soluble and insoluble fractions, respectively, are shown in Figures 7 and 8. In these cases, too, the initial slopes of the curves can be used for a comparison of the catalytic activity. The above approaches provide evidence of the formation of block copolymers, but do not elucidate the mechanism of the process. Exchange reactions 1, 2 and 4 can be responsible for the formation of block copolymers; since these reactions lead to different changes in the molecular weight of PET and PC segments, the measurement of the rate of these changes would be of utmost importance for a knowledge of the mechanism involved. Unfortunately, only the measurement of the rate of change of the molecular weight of the PET segments was possible after selective degradation of the PC segments. The time dependences of the intrinsic viscosity of the PET chains recovered after selective degradation are shown in Figure 9. For most catalysts, there is a strong decrease of the average length of the PET chains with reaction time. As phenol-ester exchange (Reaction 3) can be reasonably ruled out, the decrease of the molecular weight of PET can be due to thermal chain scission (Reaction 7) and to ester-carbonate exchange (Reaction 4). The thermal chain scission of PET has been studied extensively [83,85] and its rate, calculated at 270°C and reported in

Figure 7. Change in the composition of the soluble fraction as a function of time for different catalysts: Ti; 4-Sm; -0- Ce; -A- Er; Eu; -0- Ca/Sb [70]

+

--+-

107

Effects of Catalysts in Reactive Blending

Reaction time (min) Figure 8. Change in the composition of the insoluble fraction as a function of time for different catalysts: 4- Ti; -EE- Sm; -0- Ce; -A- Er; Eu; -0- Ca/Sb

-+-

1 .o 0.8

+

Y

0.6

Y

0.4

0.2

l

0

'

l

10

'

l

20

'

l

30

'

l

40

Reaction time (min)

'

l

50

'

l

60

Figure 9. Change in the intrinsic viscosity (relative to the starting value) for PET blocks after selective degradation: 4- Ti; -EE- Sm; -0- Ce; -A- Er; -V- Tb; -4- Eu; -0- Ca/Sb. -x- Thermal degradation

F. Pilati, M. Fiorini, C.Berti

108

Figure9, is by far lower than that observed in the same figure for reactive blending in the presence of residual catalysts. Except for Ca/Sb, the order of the catalytic activity toward Reaction 4 is the same as that established with other techniques. The reason for the anomalous behaviour of the Ca/Sb system is unclear and it may be tentatively ascribed to the effects of several contributions: a relatively stronger effect of this catalytic system on the degradation reaction of PET [83] and a slower redistribution of molecular weight through Reactions 6 may be assumed to be responsible for the faster reduction of the average length of the PET segments. These data further suggest that for most catalysts, the major contribution to the formation of block copolymers stems from the ester-carbonate exchange (Reaction 4), even though a contribution from hydroxyl terminal groups of PET (Reaction 1) is expected according to the recent results of Porter

PI

*

The effect of the different catalysts is also reflected on the MWD of both the crude products withdrawn from the Brabender and the residual PET after selective degradation of PC. Some typical SEC curves are shown in Figures 10-12 [70]for samples obtained from melt-mixing in the presence of Ti(OBu)4, Ca/Sb and EuL3, respectively.

10

15

v. (ml)

20

25

Figure 10. SEC curves of the samples obtained by melt mixing of PET and PC for 20 min at 270°C in the presence of Ti(0Bu)r: (c) crude sample after mixing; (d) residual PET after selective degradation. Dashed and dotted lines are for starting (a) PC and (b) PET [70]

109

Effects of Catalysts in Reactive Blending

The SEC curves of crude products after 20 min of mixing are similar to those expected from the combination of the starting PET and PC; the peak shape is similar and, in particular, there is no evidence of a second peak or of broadening at either large or small elution volumes (V,),as would be expected by a substantial contribution of Reactions 1 and 2, occurring at random on the P C chains. On the other hand, after selective degradation, the SEC peaks of recovered PET are shifted to larger V, (lower molecular weights) with respect to initial PET, as expected if ester-carbonate exchange (Reaction 4) took place. The retention volumes and the peak shapes strongly differ from catalyst to catalyst and from those of the initial PET; in particular, we observed that the shift to higher V, values is stronger for Ti(0Bu)r (Figure 10) than that for EuL3 (Figure ll),and that the peak for EuL3 is broader than that for Ti(0Bu)a. While for both Ti and Eu catalysts the SEC peaks after selective degradation are unimodal, for the sample with Ca/Sb the peak is broader and trimodal (Figure 12). These data confirm once more that Ti derivatives are much more efficient than the other catalysts (the shift at larger retention volumes is stronger) and indicate that the process can be very complex (as in the presence of Ca/Sb). Ester-ester exchange (Reactions 6), which favours the redistribu-

7B v

>

3

I

I

I

I

10

15

20

25

Ve (ml) Figure 11. SEC curves of the samples obtained by melt-mixing of PET and PC for 20 min at 27OoC in the presence of EuL3: (c) crude sample after mixing; (d) residual PET after selective degradation. Dashed and dotted lines are for starting (a) PC and (b) PET [70]

F. Pilati, M. Fiorini, C. Berti

110

I 10

I 15

K

I 20

I 26

(ml)

Figure 12. SEC curves of the samples obtained by melt-mixing of PET and PC for 20 min at 270°C in the presence of Ca/Sb: (c) crude sample after mixing; (d) residual PET after selective degradation. Dashed and dotted lines are for starting (a) PC and (b) PET [70]

tion of the molecular weight of PET segments, and other factors such as the migration rate of the catalyst from PET domains to PC domains, the increased miscibility after block copolymer formation, and rheological effects, can be considered responsible for the different peak shapes observed in SEC after selective degradation. For the Ti catalyst, the fast ester-ester exchange reaction can explain the sharp peak observed for recovered PET. On the other hand, slower ester-ester exchange reactions can explain the broadening of the peak for Eu; PET at the interface undergoes exchange reactions with PC at a faster rate than that of Reactions 6 with the core of the PET domains. As already mentioned, several simultaneous effects can be assumed responsible for the complex peak shape in the case of Ca/Sb catalyst.

Effects of long reaction times. The effects of different catalysts on the chemical structure are still more evident at long reaction times. In principle, the exchange reactions would continue until the formation of random copolymers. However, for very active catalysts, such as residues of Ti(OBu)4, side reactions can affect the final chemical structure. l H NMR spectra of samples from a 50/50 blend of commercial PC and of PET

111

Effects of Catalysts in Reactive Blending

30 min at 270’C

TMS

90 min at 27OoC, then 100 min at 29OoC under vacuum -?

Figure 13. ‘H NMR spectra (200 MHz) of products obtained by reactive blending of PC and PET for different times (PC/PET 50:50 by wt, Ti(0Bu)r catalyst, Brabender mixer, 30 rpm)

112

F. Pilati, M. Fiorini, C. Berti

containing residues of Ti(OBu)r, taken at various times of mixing, are shown in Figure 13. Most of the characteristics of these spectra have already been discussed in the previous section. Here it is interesting to note the results at long reaction times with application of vacuum. 'H NMR and Fourier transform (FT) 1R spectra for long reaction times in the presence of residual Ti catalyst in PET are consistent with the occurrence of exchange reactions and with a progressive elimination of carbon dioxide and cyclic ethylene carbonate (EC) from the reacting system. All the ethylidene moieties originating from PET could, in principle, be transformed into EC and therefore a pure polyarylate, poly(bispheno1 A terephthalate), could be obtained by reactive blending of a 50/50 (by mol) blend of PET and PC with the complete removal of EC. Actually, a large percentage of EC was removed (about 20% of the theoretical amount), but its further removal proved impossible. The peaks at 6.9 and 4.3ppm in the 'H NMR spectra, attributable to aromatic and aliphatic protons, respectively, adjacent to the oxygen of an ether group, can explain the above observation. These peaks stem from decarboxylation of aliphatic-aromatic carbonate groups (Reaction 8), previously formed by exchange reactions. They can undergo several competing reactions (further exchange, elimination of EC and release of COz) and, of course, the resulting chemical structure is defined by the straight and reverse relative rates (whereas Reactions 4 and 5 are reversible, the release of carbon dioxide, Reaction 8, is irreversible), and by the rate of removal of EC. From the intensity ratio of the various peaks in the 'H NMR spectrum of a sample subjected to long reactive blending, the chemical structure can be completely elucidated. It consists of units originating from terephthalate (T), bisphenol A (BPA), and aliphatic ethylidene moieties (Eg). 0

0

T

BPA

These units are joined together either by ester (T with Eg and BPA) or ether (Eg with BPA) groups. Information about the progressive change of the chemical composition is provided in Table 3. It should be noted that the molar ratio of BPA and T units does not change significantly even for long reaction times when vacuum is applied. In contrast, due to elimination of EC, the aliphatic units Eg decrease, and the Eg/T molar ratio, initially equal to 1, reaches a constant value of 0.84 for long times of reaction. From 'H NMR spectra, it was also possible to obtain information about the fractions of ester and ether bonds between the various units. The aromatic-aromatic ester groups increase during reaction and, at the end

113

Effects of Catalysts in Reactive Blending

of the process, it was possible to estimate (from the intensity of the peaks in the 8.1-8.4ppm region [28,29,40,41,95])that about 40% of the ester linkages of the T units are with Eg units, and the remaining 60% with BPA units, a situation close to that expected for a random distribution of Eg and BPA units. At the same time, an analysis of the 6.8-7.4ppm region, typical of the aromatic protons of BPA, confirmed that about 40% of the BPA units are linked to Eg units by an ether bond (see Table 3). From the 'H NMR spectra, it appears also that there is no significant amount of residual carbonate groups, and that -COO-CH2CH2 -0- CHzCHz -0OCunits (characterised by signals at 4.5 and 3.85 ppm), which could originate from decarboxylation of dialiphatic carbonate groups (-CH2CH2-O-CO-O-CH2CHz-), are Table 3. Operating conditions and characterisation of the polymeric materials resulting from melt-mixing of PET and PC (1/1 by mol of monomeric units) (541 Sample Sample history

Composition (molar ratioy Eg/T BPA/T

ib1 + 2[c1

[a1+ [bl + [CI 0.22

[qIb (dl/g)

1.03 0.47 Reacted' at 270°C for 30 min in 0.97 a closed system Reacted' at 270°C for 60 min in 0.91 0.99 0.32 0.43 a closed system 0.36 0.37 1.02 Reacted' at 270°C for 90 min in 0.84 a closed system 1.02 0.38 Reacted' at 270°C for 90 min in 0.56 a closed system, plus 30min at 275°C under vacuum Reactedd at 283°C for 45 min in 0.42 a closed system 0.48 Reactedd at 283°C for 45 min in a closed system, plus 35min at 275°C under vacuum Reactedd at 283°C for 45 min in 0.53 a closed system, plus 35min at 275" C and 40 min at 290°C under vacuum Reactedd at 270°C for 90min 0.84 1.05 0.39 0.56 in a closed system, plus 35min at 275°C. 40min at 290"C, and 65 min at-310°C under vacuum a n o m 'H NMR spectra; Eg, T, BPA, and units (a)-(c) are described in the text phenol/l1l,2,2-tetrachloroethane(60/40 by wt) 'In a Brabender Plasticorder PL2000 a tank reactor, with a paddle agitator

114

F. Pilati, M. Fiorini, C. Berti

negligible [55]. It is also interesting to note that the two signals at 4.7 and 4.3ppm (methylene groups adjacent to ester and ether groups, respectively) have exactly the same intensity after a long reaction time. This means that about 50% of the Eg units are linked with T units by an ester bond, and the remaining are joined to BPA units by an ether bond. Since among the degradation products obtained from a complete methanolysis of the final polymer [54], significant traces of molecules deriving from BPA-O-Eg-0BPA moieties were not found, it was concluded that ethylidene moieties are almost entirely present in the polymer chains included between terephthalate and bisphenol A units, to which they are bonded by an ester and an ether group, respectively, as shown below.

The chemical structure of the resulting copolymer can therefore be described by the units (a)-(c), randomly joined to terephthalate units (d).

As is apparent from the above discussion, the final structure of the random copolymer obtained for long reaction times strongly depends on the relative rates of Reactions 5 and 8, and reactive blending can, in principle, be used to obtain random copolymers with different chemical structures by controlling the extent of these reactions. In particular, it was demonstrated [54,55] that Reaction 5 is reversible and that the elimination of cyclic carbonates depends on both the presence of a suitable catalyst and the nature of the aliphatic glycol [78]. Pure bisphenol A polycarbonate was obtained from poly(ethy1ene-aZt-bisphenolA carbonate) in the presence of tin-based

Effects of Catalysts in Reactive Blending

115

catalysts and of Ti(OBu)4, whereas a negligible amoilnt of EC could be removed and a poly(ether-carbonate) was obtained in the absence of catalyst. This means also that the release of COz from aromatic-aliphatic carbonate groups (Reaction 8) can occur in the absence of catalysts and that catalysts can be used to change the relative rates of Reactions 5 and 8. Elimination of aliphatic cyclic carbonates (1,3-propylenecarbonate) was also observed during the reactive blending of poly(trimethy1eneterephthalate) with PC, in the presence of either Ti(OBu)4 or SnBusO [96,97]. On the other hand, when PBT was used in reactive blending with PC, no elimination of cyclic butylene carbonate was observed, even in the presence of Ti(OBu)4 catalyst [96,97].Somewhat different results were obtained with lanthanide residual catalysts in PET. For all the catalysts tested (Table l), it was never possible to obtain a chemical structure close to that expected for random copolymers. The average length of the PET segments tends to a constant value for Sm residual catalyst, which follows Ti catalyst in the catalytic activity order. Also, the formation of ether linkages is very limited and does not increase significantly even for very long reaction times, and elimination of EC has not been observed. 3.2.2. Reactive blending by a deliberate addition of catalysts

As shown above, residual catalysts in PET were found to have a strong effect on the rate of the various reactions that can occur during melt-mixing of PET and PC. Calcium acetate/antimony trioxide, a common catalytic system for PET, revealed a low catalytic activity toward exchange reactions. On the other hand, residues of Ti(OBu)4 activated not only exchange reactions, but also side reactions with the formation of ether moieties, cyclic ethylene carbonate and discolouration. Lanthanide compounds appeared to possess intermediate activities toward exchange reactions, while the contribution of side reactions was found to be rather low, as compared to that of Ti catalyst. It should also be stressed that lanthanide compounds have shown a wide range of catalytic activities, depending on the metal employed. Since the exchange reactions between PET and PC occur at the interface, only a small fraction of the overall amount of catalyst present in PET is believed to be initially active in such a system. Therefore, an alternative approach to accelerate exchange reactions and make them compatible with the low residence times typical of extruders resides in the addition of the catalyst to PET and PC just before blending. In fact, by this procedure, the catalyst is dispersed on the pellet surface, at least initially, and is expected to work more efficiently. In addition, this approach is economically advantageous, since commercial PET, which is less expensive, can be used. Experiments on PET/PC melt-blending were therefore performed in the presence of various metal derivatives added just before blending. Fur-

116

F. Pilati, M. Fiorini, C. Berti

thermore, because the efficiency of a given catalyst is expected to depend also on the miscibility of the catalyst in PET and PC domains and on the renewal of the interface, experiments were carried out using the same metal catalyst with different ligands and different mixing equipments (paddle agitator in a batch reactor [35],Brabender Plasticorder equipped with two counter-rotating double-winged rotors [74,77],and twin-screw co-rotating extruder 175,761). The catalysts used in the experiments carried out in a Brabender apparatus are listed in Table 4; they were dispersed mechanically onto polymer pellets at a concentration of O.O66mol% with respect to PET monomeric units, corresponding to that commonly used in the synthesis of PET. As described above, a first qualitative comparison of the catalytic activities of the lanthanide residues in PET and the fresh catalyst added to commercial polymers can be made on the basis of the torque values recorded during reactive blending, and of the data for the solubility of the resulting products in dichloromethane. The time for the achievement of a maximum in the torque profile can Table 4. Reactive blending of PET/PC (1/1 by wt) in Brabender at 275"C, 48 rpm, in the presence of various catalysts (0.066mol%with respect to PET units) added before blending (741 Catalyst

Soluble fraction Time to the Slope and torque max 10 min 20 min direction of the TC(OC) Tm(OC) (min) (%) (%) solubility curvea (DSC)b (DSC)b

Without catalyst Sm(ac)3 Sm(phthalate)s Sm(acac)3 Sm(o-formyl-

phenolate)^

La2 (oxalate)~ La2 (tartrate)J La2 (phthalimide)~ La(acac)s Al(%=)3 Ca(acac)2 Sr(acac)a Ba(acac)z Zn(acac)z Zn(ac)a

16 12 16 2.5

48 47 49 40

42 38 44 80

J- Slow down J- Slow down J- Slow down t Speed U P

3.8 16 14 14 4 6.5 9 8 6 8 7

35 49 49 47 21 15 40 37 14 39 38

100 43 42 37 74 53 27 22 56 25 26

t Speed UP

aSee Figure 14 bFor the samples obtained by blending for 10 min CNot observed

J- Slow down J- Slow down J- Slow down t Speed up t UP J- Down J- Down t UP J- Down J, Down

144 147 146

248 241 239

-

-C

-

-

-

C

C

143 145 159 C

193 170 168 189 176 157

-C

249 238 242 C

230 236 237 231 236 238

Effects of Catalysts in Reactive Blending

117

be taken for comparison; it was recorded at 16-18min of blending for the simple blend of commercial PET and PC, and at 8 min for the blending of commercial PC with PET synthesised in the presence of Sm(acac)3. The addition of this last catalyst, in the form of a powder dispersed on pellets of commercial PET and PC just before melt-mixing, reduced the time t o the maximum in the torque to 2.5min. The times to maximum for other freshly added catalysts are reported in Table 4. Selective solubility data confirmed the torque results. As seen in Figure 14, Sm(acac)s, freshly added to commercial PET and PC, showed a solubility profile very similar to that of residual Ti in PET, the most active residual catalyst toward exchange reactions. One can conclude that Sm(acac)S, freshly added to commercial polymers, and residues of Ti(0Bu)s in PET 100

80

ke

B

.r(

8 60

c

a

0)

2 0

cn

40

20

Figure 14.Solubility tests on PET/PC blends obtained by reactive blending in a Brabender at 275OC: neat mixture of commercial PET and PC (0);blends of P C with PET synthesized with Sm(acac)s, 0.066mol% with respect to PET units ( 0 ) ; blends of commercial PET and PC obtained with Sm(acac)s added before blending, 0.066mol% with respect to PET units (0);blends of PC with PET synthesized with Ti(OBu)r, 0.066 mol% with respect to PET units (A) [74]

F. Pilati, M. Fiorini, C. Berti

118

show similar rates with respect to exchange reactions, but they lead to the formation of copolymers with quite different chemical structures. In fact, selective degradation of PC moieties in the samples blended for 10 min revealed that in the case of Ti catalyst the intrinsic viscosity value of PET decreased from 0.91 to 0.09dl/g, while addition of Sm(acac)s to commercial PET and P C leads to a decrease in intrinsic viscosity from 0.65 to 0.17 dl/g. Based on the same criteria, the catalysts employed can be ordered as follows (see Table 4): Sm(acac)3 = Sm(c-formyl pheno1ate)a > La(acac)s > Al(acac)s = Ba(acac)z > Zn(acac)z = Zn(ac)z > Sr(acac)z = Ca(acac)z > La(phthalimid)s > Laz(tartrate)s > Laz(oxa1ate)s = Sm(phtha1ate)a = Sm(ac)s = residues of catalysts in commercial PET. In addition to the type of metal, the type of ligand plays also a very important role; carboxylic acid salts of lanthanides appeared to possess very low catalytic Table 5. Properties of PET/PC (50/50 by wt) blends and copolymers obtained by one-step extrusion in a twin-screw extruder in the presence of various catalysts (0.033 or 0.066 mol% with respect to PET) [75] ~~~

~

1771 of 1771 of Solubility resulting PET Tg(I)a T,(II)a T," T : in CHzClz blend" segmentsd Catalyst (dig-') (dl g-'1 ("C) ("C) ("C) ("C) (%) (mol% (wt%)) 0.60 0.60 143 144 249 48 Without catalyst 86 Sm(acac)s 183 232 0.55 0.28 26 109 0.033 (0.045) Sm(acac)3 0.59 0.22 -e 194 232 23 109 0.066 (0.090) La(acac)a 0.58 0.30 24 111 0.033 (0.044) -e 189 236 Pb (acac)2 -e -e -e 0.59 0.17 23 111 0.033 (0.039) Ca(acac)2 0.60 0.63 49 87 143 143 250 0.033 (0.024) Ca(acac)z 0.60 0.43 44 140 170 243 88 0.066 (0.048) Zn(acac)z 0.61 0.46 48 146 145 250 88 0.033 (0.026) Zn(acac)z 0.066 (0.052) 89.5 137.5 166 243 45 0.60 0.43 *Results obtained from dynamic-mechanicalthermal analysis bResults obtained from DSC "Intrinsic viscosity of the final product (tetrachloroethane/phenol,60/40 by wt, 25OC) dIntrinsic viscosity of PET segments after selective degradation of PC moieties in the final product (tetrachloroethane/phenol,60/40 by wt, 25OC) eNot observed

Effects of Catalysts in Reactive Blending

119

activity, probably due to both their solubility and chemical reactivity. It is interesting to note that the same metal catalyst with different ligands showed nearly the same activity when added during polymerisation. As is apparent from the above discussion, the chemical structure of the products that results from reactive blending of PET and P C it can be controlled to a certain extent by choosing a suitable catalyst, the way it is introduced into the blend (as a residue in PET or added on purpose before blending), and the reaction time. However, when reactive blending is carried out in commercial extruders, the residence time is usually short and it can change only within a narrow range. In this case, the reaction time is no longer a parameter for the control of the chemical structure and some other variables, namely the mixing efficiency and the concentration of the catalyst added, have been tested. Various metal catalysts in the form of acetylacetonate salts were tested in a P E T / P C one-step extrusion process, carried out in a twin-screw extruder a t 270-280°C and 10&150rpm, with a residence time of about 1min; they were initially dispersed on the PET pellets at 0.024-O.09 wt% with respect to the overall blend. As seen in Table 5 , from solubility tests and selective degradation it was found that Pb, Sm, and La were the most active catalysts, while Ca(acac)z was quite ineffective. DSC data also supported this conclusion [76]. Sm(acac)S, one of the most effective catalysts when added to commercial P C and PET, was also used in experiments at different concentrations. It was initially dispersed on PET or P C pellets at concentrations from 50

."0

-

0

m 20

0

I

0.02

I

0.04

I

0.06

Catalyst, %

I

0.08

C 1

Figure 15. Influence of the catalyst concentration on the solubility of PET/PC blends and copolymers in dichloromethane [76]

120

F.Pilati, M.Fiorini, C. Berti

0.015 to 0.09 wt% with respect to the overall PET/PC blend, which was then extruded in a twin-screw co-rotating extruder at a residence time of about 1min [75]. The results (Figure 15) suggest that the fraction soluble in dichloromethane first decreases (formation of block copolymers), then remains almost constant with a further increase of catalyst concentration. The value of the solubility at high concentration is typical of that found for the minimum in the solubility us. time curves. It probably reflects effects of limited catalyst d i h i o n from the surface to the inner part of the polymer domains, saturation of the surface, etc. 4. Conclusions

It is now well established that various reactions can occur during blending of polyesters and PC at high temperature, in the molten state. The main initial effect of these reactions is the formation of polyester-PC block copolymers at the interface of polyester and PC phases. These copolymers improve the local miscibility, which in turn favours the further contact of polyester and PC. The amount of block copolymers gradually increases, while the average length of the polyester and PC segments decreases. Side reactions, such as the formation of cyclic carbonates by intramolecular exchange and carbon dioxide release from aromatic-aliphatic carbonate moieties, can take place in parallel with the formation of block copolymers. The chemical structure resulting from reactive blending depends on the relative rates of these reactions which in turn depend on the type of catalyst. It can be present in polyesters as a residue from the bulk polymerisation, or it can be deliberately added to the polymers in an appropriate amount just before blending. A more efficient mixing apparatus can improve the renewal of the interfaces and can be used to accelerate the process. Titanium alkoxides are the most efficient catalysts for exchange reactions and in the case of PET/PC systems the high catalytic activity can be exploited to eliminate EC from the blend, shifting the composition of the resulting polymeric material toward a more aromatic character. The fast exchange in the presence of this catalyst can make the control of the chemical structure difficult when a low extent of exchange (high molecular weight of the blocks) is desired. In this case, or when a low amount of side reaction is desired, other catalysts, such as some lanthanide derivatives, can be more convenient. The addition of an effective catalyst inhibitor can hinder exchange reactions and allow a simple blend of unreacted polymers to be obtained. In conclusion, the proper choice of reaction conditions (polymer ratio, type and concentration of catalyst, reaction time, type of mixing apparatus, catalyst inhibitors, etc.) makes it possible to control the evolution of the chemical structure and to prepare a large number of polymeric materials with a wide range of properties from the same pair of initial polymers. It

Effects of Catalysts in Reactive Blending

121

is possible to obtain products ranging from a simple blend of t h e starting polymers t o polymeric materials with morphologies and properties that gradually change from those of t h e unreacted blend to those typical of a random copolymer of the same composition, with the additional possibility of increasing the aromaticity by withdrawing cyclic aliphatic carbonates from the blend during melt-mixing.

References 1. D. R. Paul, S. Newman, “Polymer Blends”, Academic Press, New York 1978, vols. 1 and 2 2. 0. Olabisi, L. M. Robeson, M. T. Shaw, “Polymer-Polymer Miscibility”, Academic Press, New York 1979 3. D. R. Paul, J. W. Barlow, H. Keskkhula, “Polymer Blends”, in: Encyclopedia of Polymer Science and Engineering, Wiley, New York, 1985, vol. 12, p. 399 4. D. J. Walsh, “Polymer Blends”, in: Comprehensive Polymer Science, Pergamon, Oxford 1989, vol. 2, ch. 5 5. W. J. MacKnight, F. Karasz, “Polymer Blends”, in: Comprehensive Polymer Science, Pergamon, Oxford 1989, vol. 7, ch. 4 6. L. A. Utracki, “Polymer Alloys and Blends, Thermodynamics and Rheology”, Hanser, Munich 1990 7. D. W. Fox, R. B. Allen, “Compatibility”, in: Encyclopedia of Polymer Science and Engineering, Wiley, New York 1985, vol. 3, p. 758 8. D. Heickens, N. Hoen, W. Barentsen, P. Piet, H. Ladan, J. Polym. Sci., Polym. Symp. Ed. 62,309 (1978) 9. R. Fayt, R. Jerome, P. Teyssis, Makromol. Chem. 187,837 (1986) 10. J. Noolandi, Ber. Bunsenges. Phys. Chem. 89, 1147 (1985) 11. A. Vilgis, J. Noolandi, Makromol. Chem., Macromol. Symp. 16, 225 (1988) 12. R. Fayt, R. Jerome, P. Teyssis, J. PoZym. Sci., Polym. Lett. Ed. 24, 25 (1986) 13. P. Teyssis, R. Fayt, R. Jerome, Makromol. Chem., Macromol. Symp. 16,41 (1988) 14. S. B. Brown, C. M. Orlando, “Reactive Extrusion”, in: Encyclopedia of Polymer Science and Engineering, Wiley, New York 1988, vol. 14, p. 169 15. “Reactive E z t m i o n ” , edited by M. Xanthos, Hanser, Munich 1992 16. M. Lambla, “Reactive Processing of Thermoplastic PolymerJ”, in: Comprehensive Polymer Science, Pergamon, Oxford 1992, 1st Suppl., ch. 21 17. R. J. Kumpf, J. S. Wiggins, H. Pielartzik, T k n d s Polym. Sci. 3, 132 (1995) 18. D. Freitag, U. Grigo, P. R. Muller, W. Nouvertnci, “Polycarbonated’, in: Encyclopedia of Polymer Science and Engineering, John Wiley & Sons, New York 1988, vol. 11, p. 648 19. D. C. Clagett, S. F. Shafer, “Polycarbonated’, in: Comprehensive Polymer Science, Pergamon, Oxford 1989, vol. 5, p. 345 20. C. A. Pryde, M. Y. Hellman, J . Appl. Polym. Sci. 25, 2573 (1980) 21. I. Goodman, “Polyesters”, in: Encyclopedia of Polymer Science and Engineering, John Wiley & Sons, New York 1988, vol. 12, p. 1 22. D. B. G. Jaquiss, W. F. H. Borman, R. W. Campbell, “Polyesters, t h e m o plastic?’, in: Kirk-Othmer Encyclopedia of Chemical Technology,John Wiley & Sons, New York 1982, vol. 18, p. 549

122

F. Pilati, M. Fiorini, C. Berti

23. F. Pilati, “Polyesters”,in: Comprehensive Polymer Science, Pergamon, Oxford 1989, vol. 5, p. 275 24. T. Kaneko, M. Takamatsu, JP 115146 (1974); M. Matsukane, C. Azo, JP 54160 (1973); J. Yamamoto, M. Ueno, K. Umeda, J P 80162 (1974); A. D. Wambach, Ger. Pat. 2310742 (1973); Y. Nakamura, R. Hasegawa, H. Kubota, Ger. Pat. 2343609 (1974); J. P. Mercier, J. Devaux, P. Godard, Brit. Pat. 1569296 (1977) 25. T. R. Nassar, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 23,85 (1979) 26. S. R. Murff, D. R. Paul, J. W. Barlow, J. Appl. Polym. Sci. 29,3231 (1984) 27. D. C. Wahrmund, D. R. Paul, J. W. Barlow, J. Appl. Polym. Sci. 22,2155 (1978) 28. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22,229 (1982) 29. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1875 (1982) 30. J. Devaux, P. Godard, J. P. Mercier, R. Touillaux, J. M. Dereppe, J. Polym. Sci., Polym. Phys. Ed. 20,1881 (1982) 31. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1895 (1982) 32. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 33. W. E. Preston, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 29,2251 (1984) 34. A. W. Birley, X. Y . Chen, Brit. Polym. J. 16,77 (1984) 35. F. Pilati, E. Marianucci, C. Berti, J. Appl. Polym. Sci. 30,1267 (1985) 36. X. Y. Chen, A. W. Birley, Brit. Polym. J. 17, 347 (1985) 37. G. Montaudo, C. Puglisi, F. Samperi, Polym. Deg. Stab. 31,291 (1991) 38. B. D. Hanrahan, S. R. Angeli, J. Runt, Polym. Bull. 14,399 (1985) 39. B. D. Hanrahan, S. R. Angeli, J. Runt, Polym. Bull. 15, 455 (1986) 40. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24, 3301 (1986) 41. P. Godard, J . M. Dekoninck, V. Devlesaver, J. Devaux, 3. Polym. Sci., Polym. Chem. Ed. 24,3315 (1986) 42. Z. H. Huang, L. H. Wang, Maleromol. Chem., Rapid Commun. 7,255 (1986) 43. D. Delimoy, C. Bailly, J. Devaux, R. Legras,Polym. Eng. Sci. 28,104 (1988) 44. M. Gaztelumendi, 1. Mondragon, J. Nazabal, Makromol. Chem., Macromol. Symp. 20/21,269 (1988) 45. P. M. Henrichs, J. Tribone, D. J. Massa, J. M. Hewitt, Macromolecules 21, 1282 (1988) 46. T. Suzuki, H. Tanaka, T. Nishi, Polymer 30, 1287 (1989) 47. W. N. Kim, C. M. Burns, J. Polym. Sci., Polym. Phys. Ed. 28,1409 (1990) 48. Li-Hui Wang, Z. Huang, T. Hong, R. S. Porter, J. Macromol. Sci-Phys. B29, 155 (1990) 49. Li-Hui Wang, M. Lu, X. Yang, R. S. Porter, J. Macmmol. Sci-Phys. B29, 171 (1990) 50. Y. V. Konyukhova, N. P. Bessowva, S. I. Belousov, V. I. Fel’dman, Y.K. Godovskii, Polymer Science 33,2262 (1991) 51. G. Montaudo, C. Puglisi, F. Samperi, J. Polym. Sci., Polym. Chem. Ed. 31, 13 (1993)

Effects of Catalysts in Reactive Blending

123

52. R. Legras, D. Delimoy, J. Devaux, Proc. Workshop “Interfaces New Materials”, Elsevier, Amsterdam 1991, p. 120 53. R. P. Joyce, A. P. Berzinis, Compalloy’91, Fourth Int. Congress on Compatibilizers and Reactive Polymer Alloying, New Orleans 1991, p. 67 54. C. Berti, V. Bonora, F. Pilati, M. Fiorini, Makromol. Chem. 193,1665 (1992) 55. C. Berti, V. Bonora, F. Pilati, M. Fiorini, Makromol. Chem. 193,1679 (1992) 56. R. S. Porter, L. H. Wang, Polymer33, 2019 (1992) 57. G. Montaudo, C. Puglisi, in: Comprehensive Polymer Science, Pergamon, Oxford 1989, 1st Supplement, p. 227 58. D. G. Hamilton, R. R. Gallucci, J. Appl. Polym. Sci. 48, 2249 (1993) 59. W. Zheng, Z. Wan, Z. Qi, F. Wang, Polymer 34, 4982 (1993) 60. J. S. Kollodge, R. S. Porter, Polymer 34, 4990 (1993) 61. M. Okamoto, T. Inoue, Polymer 35, 257 (1994) 62. N. Wings, G. Trafara, Angew. Makromol. Chem. 217, 91 (1994) 63. H. Yang, W. Yetter, Polymer 35, 2417 (1994) 64. W. Zheng, Z. Qi, F. Wang, Polym. Int. 34, 307 (1994) 65. W. Zheng, Z. Wan, Z. Qi, F. Wang, Polym. Int. 34, 301 (1994) 66. Zhi-li Liao, Feng-Chin Chang, J . Appl. Polym. Sci. 52, 1115 (1994) 67. A. N. Wilkinson, D. Cole, S. B. Tattum, Polym. BdE. 35, 751 (1995) 68. R. Zhang, X. Luo, D. Ma, J. Appl. Polym. Sci. 55, 455 (1995) 69. D. Delimoy, B. Goffaux, J. Devaux, R. Legras, Polymer 36, 3255 (1995) 70. M. Fiorini, C. Berti, V. Ignatov, M. Toselli, F. Pilati, J. Appl. Polym. Sci. 55, 1157 (1995) 71. J. S. Kollodge, R. S. Porter, Macromolecules 28, 4106 (1995) 72. J. S. Kollodge, R. S. Porter, Macromolecules 28, 4097 (1995) 73. J. S. Kollodge, R. S. Porter, Macromolecules 28, 4089 (1995) 74. V. Ignatov, C. Carraro, V. Tartari, R. Pippa, F. Pilati, C. Berti, M. Toselli, M. Fiorini, Polymer 37, 5883 (1996) 75. V. Ignatov, C. Carraro, V. Tartari, R. Pippa, M. Scapin, F. Pilati, C. Berti, M. Toselli, M. Fiorini, Polymer 38, 201 (1997) 76. V. Ignatov, C. Carraro, V. Tartari, R. Pippa, M. Scapin, F. Pilati, C. Berti, M. Toselli, M. Fiorini, Polymer 3 8 , 195 (1997) 77. M. Fiorini, F. Pilati, C. Berti, M. Toselli, V. Ignatov, Polymer 38, 413 (1997) 78. C. Berti, V. Bonora, M. Fiorini, F. Pilati, B. Sweileh, Macromol. Chem. Phys. 198, 321 (1997) 79. K. P. McAlea, J. M. Schultz, K . H. Gardner, G. D. WignaU, Polymer 27, 1581 (1986) 80. J. Kugler, J. W. Gilmer, D. Wiswe, H. G. Zachmann, K. Hann, E. W. Fischer, Macromolecules 20, 1116 (1987) 81. G. Montaudo, P. Maravigna, “Formation of Cyclic Oligomers”, in: Comprehensive Polymer Science, Pergamon, Oxford 1989, vol. 5, ch. 5 82. U. W. Suter, “Ring-Chain Equilibria”, in: Comprehensive Polymer Science, Pergamon, Oxford 1989, vol. 5, ch. 6 83. H. Zimmermann, in: Developments in Polymer Degradation, edited by N. Grassie, Applied Science, London 1984, vol. 5 84. V. Passalacqua, F. Pilati, V. Zamboni, B. Fortunato, P. Manaresi, Polymer 17, 1044 (1976) 85. J. Devaux, P. Godard, J. P. Mercier, Makromol. Chem. 179, 2201 (1978)

124

F. Pilati, M. Fiorini, C. Berti

86. K.B.Abbas, Polymer 21, 936 (1980) 87. D.J. Brunelle, US 4452932 (1984) 88. J. J. Verhoeven, EP 02641432 (1990) 89. M.F. Cheung, K. R. Carduner, A. Golovoy, H. van Oene, J. Ind. Eng. Chem. Res. 28,476 (1989) 90. A. Golovoy, M.F. Cheung, K. R. Carduner, M. J. Rokosz, Polym. Eng. Sci. 29,1226 (1989) 91. S. Chang, M.F. Sheu, M. H. Chang, J. Polym. Sci., Polym. Chem. Ed. 20, 2053 (1982) 92. A. Golovoy, M. F. Cheung, K. R. Carduner, M. J. Rokosz, Polym. Bull. 21, 327 (1989) 93. K. R.Carduner, R. 0. Carter 111, M. F. Cheung, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40,963 (1990) 94. M. F. Cheung, K. R. Carduner, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40,977 (1990) 95. E.V. Guinlock, R. A. Wolfe, J. C. Rosenfeld, J. Appl. Polym. Sci. 20, 949 (1976) 96. F.Pilati, C.Berti, M. Fiorini, M. Toselli, V. N. Ignatov, V. Tartari, C. CarTaro, R. Pippa, A. Moro, POC’94, 6th International Symposium on Polymer Supported Reactions in Organic Chemistry, Venezia, Italy, 19-23 June 1994, p. 140 97. C. Berti, E. Marianucci, B. Sweileh,F. Pilati, M. Fiorini, 4th AIM Conference on Advanced Topics in Polymer Science, Gargnano (Bs),Italy, 2-7 June 1996, p. 40

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 3

Model Studies of Transreactions in Condensation Polymers

J. Devaux

1. Introduction

Condensation polymers, as defined by Carothers [l],are those polymers obtained by a coupling reaction between polyfunctional monomers, involving the elimination of a small molecule, e.g., water. Later on [2], condensation polymers were defined as polymers with repeat units joined together by functional units of one kind or another, such as ester, amide, urethane, sulfide, and ether linkages. Other classifications, such as Flory’s [3], emphasise the differences in polymerisation mechanisms for the distinction between polymers. A main characteristic of the condensation polymers lies in their polymerisation reaction, which progresses stepwise, through successive equilibrated steps. End-groups of condensation polymers are usually reactive. As thermodynamic equilibria are always dynamic states, condensation polymers in thermodynamic equilibrium are “living” species, continuously exchanging end-groups and chain linkages. As a consequence, in “pure” polycondensates this leads to equilibrium molecular weight distributions. The situation becomes more complex when two, or more, condensation polymers are mixed together, forming homogeneous binary phases. In this case, the exchange of end-groups and chain linkages can lead to distributions of molecular weights but also of molecular compositions. Thermodynamic and kinetic parameters play their role in defining the nature of

126

J. Devaux

the products (copolymers) and the progress of the reaction (transcondensation). Transcondensation reactions can be observed between polyesters, polyamides, polyurethanes, and also polyethers. Mixed transcondensations are also conceivable and, for instance, polyester-polyamide, polycarbonatepolyamide transcondensations were also reported. Reports about exchange reactions, involving various types of condensation polymers, can be found in Chapter 1 of this book or elsewhere. For the sake of consistency, this chapter will mainly refer to transcondensations between polyesters and/or polycarbonate. The use of high molecular weight polycondensates in the study of transreactions has theoretically one major advantage concerning the sensitivity of the analytical detection of low levels of reaction. Indeed, if two macromolecules of different natures, having each a degree of polymerisation of 100, are exchanging only one chain linkage, i e . , undergoing 1%reaction on a molar scale - the final product is a “100% pure” block copolymer exhibiting completely different physicochemical properties (e.g., solubility), rendering its identification obvious. Moreover, even if the number of “mixed” polymer linkages is too low for its determination by any spectroscopic technique, each new link “bears”,on the average, a rather long “tail” of a chemically different nature, enhancing the sensitivity of the detection. Separations by size exclusion chromatography can therefore be based on the specific absorption of each constituent. However, drawbacks inherent to the use of polymers in chemical reactions are numerous, e.g., high viscosities, limited solubilities, reduced numbers of available analytical techniques. Therefore, model studies are often undertaken in order to simplify the characterisation and the quantification of the reaction products. Model studies may aim at the description of the microstructure of the copolymers obtained, i.e., from a blocky to a random distribution of structural units along the chain. Other model studies concern the kinetic analysis of exchange reactions, based on the quantification of new chain linkages. Finally, model studies are also undertaken in order to identify these new links and, in this way, to elucidate the transreaction mechanism. In this latter case, reactions can be followed between low molecular weight species chemically similar to the structure of the polymer chain. Again, the use of one macromolecule and one low molecular weight species has a major advantage, as the transreaction leads to a drastic decrease of the molecular weight of the polymer, which is very accurately measurable. The following sections deal with a model study of the microstructure of copolycondensates arising from transreaction, then with a kinetic model applicable to the same kind of reactions. The third section concerns model transreactions either between one polymer and one low molecular weight species, or between two low molecular weight species. These model reactions are considered in order to identify the reaction mechanisms and also

Model Studies of Transreactions in Condensation Polymers

127

to propose a possible mechanism for transreaction catalysis and inhibition. All these studies are illustrated using the poly(buty1ene terephthalate) (PBT)/bisphenol A polycarbonate (PC) system, which represents a general case of transreaction between polyesters involving four different structural units. Other transesterification reactions from the literature as well as from the author's data are also considered. 2. Theoretical

2.1. Microstructure of copolycondensates from transreactions

2.1.1. Copolycondensates from three structural units [4] Copolycondensates may originate from transreaction between two polycondensates having one structural unit in common. In this case, the copolycondensate will be made of three different structural units. Its general structure can be written:

A1B and A2B are two different repeat units, denoted here as A1 and A2 structural units since B represents the common structural unit; z and y are the respective sequence average lengths and n is a repetition factor such as the degree of polymerisation is n(z y). Later in this chapter, [Ail will mean the molar concentration in structural unit A,. As a general rule, only copolycondensates resulting from a transreaction between two linear polycondensates of high molecular weights will be considered, allowing chain ends to be neglected in statistical calculations. The molar fraction of a structural unit is given by:

+

FA( = t r[Ail Ail' i , j = 172

(2)

a= 1

The molar fractions of repeat units (dyads) can also be defined:

The probability of an A, structural unit being followed by an Aj structural unit is:

i,j=l

128

J. Devaux

From these definitions it can be deduced that:

F A , A= ~ FA( x P A ~, A i, ~j = 1 , 2

(5)

This means that the molar fraction of dyads (&Aj) equals the molar fraction of Ai multiplied by the probability of an Aj being followed by an

Aj.

It is clear that the distinction between [AjAj] and [AjAi] is only formal, both concentrations being equal. For simplicity, a unique notation is used:

The values of FA( can be calculated from the initial concentrations (see Eq. (2)), the fractions of dyads being experimental values. They can be obtained, as shown later, using nuclear magnetic resonance (NMR) or infrared (IR) data. Yamadera and Murano [4] introduced the concept of “degree of randomness”, B , with the following definition:

B = P A ~iA P~A ~ Ai~#, j

(7)

Using Eqs. ( 5 ) and (6), one obtains:

Equation (8) allows the direct calculation of B from experimental data. The value of the degree of randomness, B, characterises the copolycondensate microstructure (see Table 1). Table 1. Copolymer microstructure as a function of the degree of randomness B

B

Copolymer type

1

Random

#j

2

Alternating

Long sequences of identical PA~A, = PA~A, = 0 i, j = 1,2 i # j structural units

O

Block

Case Random distribution of structural units Alternating distribution of structural units

Probability P A ~= A~ FA^

i,j = 1,2

P A ~= AP~ A ~ = A 1~ a, j = 1,2 i

129

Model Studies of Transreactions in Condensation Polymers

The average sequence length can also be determined from the general formula of the copolycondensate (Eq. (1)) and the number of dyads, nb, can be easily deduced from the same formula:

nb(A1A1) = n ( z - 1) nb(AzA2) = n(y - 1) nb(A1Az) = 2n Taking into account that

the average sequence lengths can be calculated:

x=-

1

y=-

1

(14)

PAA

The following relation deserves attention: 1

1

X

Y

B=-+-

2.1.2. Copolycondensates from four structural units [5]

An exchange reaction, taking place between two linear polycondensates having no structural unit in common, will generate a copolycondensate with four different structural units. In general, six parameters are needed to describe this kind of copolycondensate, which can be written as follows:

-[ (A1-B 1)z-( A2--B

1 )v]m-[

(A1-b)

z-(

A2--B

2)Urln-

(16)

A1, A2, B1 and B2 represent the four different structural units; x, y, z and w are the sequence average lengths; m and n are the average lengths of sequences having in common the same B1 or B2 unit.

Again, in this section, only copolycondensates resulting from a transreaction between two linear polycondensates (AlBl), and ( A Z B ~of) ~high molecular weights are considered, allowing chain ends to be neglected in statistical calculations. The following concentration relationship results from these considerations:

130

J. Devaux

and also:

my = n z

(18)

Equation (18) merely states the equality of the numbers of AzBl and dyads. The molar fractions of structural units are defined again by concentration ratios (see Eq. (2)): AlB2

and

C PjI

j=1

It follows from Eq. (17) that:

FA, = FB, (i = 1,2) Spectroscopic measurements, 'H and 13C nuclear magnetic resonance (NMR), make it possible to quantify dyad and triad fractions. Thus, two modes of calculation of the copolycondensate sequence distribution are developed here, based on dyad and triad determinations. Dyad analysis. The fraction of dyads AiBj is defined by: FAiBj =

[AiBj]

t

--[AiBj]

[ ~ i ~ j 2 l i ~ i l i,j=l i= 1

Since

2

2

i=l

j=1

C [Ail = C [Bj] and [AiBj] = [BjAi]:

If the copolycondensate chain could be inspected by proceeding from one end to the other, the probability of finding an A, structural unit followed by a Bj one would be:

j=1

Model Studies of Transreactions in Condensation Polymers

131

From Eqs. (17) and (18) it follows that:

PA,B~ = P B , A ~( i , j = 1,Z) whereas, except for an equimolar system,

P A ~#BP~B ~ A(i , #j) Rom Eqs. (19), (22) and (24), it can be deduced that:

FA,B~ = FA^ P A , B ~(i, j = 1 , 2 ) The degree of randomness B (see Eq. (7)) can be defined as:

Equations (27) and (28) lead to:

The microstructure of the copolycondensate originating from four different structural units can be deduced from the B values, as shown in Table 1. A random copolycondensate ( B = 1) means that the structural unit distribution obeys Bernoulli statistics, but respects the strict A,B alternantion. Values of B < 1 indicate that the structural units tend to cluster in homogeneous sequences AlBl and AzBz (the limiting case of B = 0 corresponds to a copolycondensate with infinitely long ones). Values of B > 1 denote a tendency of AlBz and AzBl to predominate over AlBl and AzBz sequences while the limiting case of B = 2 corresponds to the formation of (infinitely) long sequences A1Bz A1 Bz A1BzAlBz and AzB 1A2B 1AzB 1AzB 1 by the transreaction.

Triad analysis. Let us consider, for instance, a Bj structural unit adjacent to an Ai, on the one side, and to an Ak on the other side. The fraction of AiB .A triads can be defined and the use of lower-case f indicates that 3 .k the fraction of triads AiBjAk is restricted to triads having a Bj central unit:

i,k=l

132

J. Devaux

By inspecting again the structural units along the chain, from one end to the other, one can define the probability of finding an & unit following an AiBj sequence as:

From Eqs. (24), (30), and (31), it can be deduced that: fAiBjAk = FAi

PAiBj

BjAk

FB~

and also that:

k= I The degree of randomness around Bj can be defined by the expression:

Again, the relationship between the copolycondensate microstructure and the values of BB, can be found in Tablel. A definition of BA(can be deduced in a similar way. Average sequence lengths. These can be deduced from the general formula of the copolycondensate (Eq. (16)). For instance, 2, the average length of the sequences A1B1, equal to the ratio of the average number of AlBl sequences and the average number of AlBlA2 sequences, may be writtten as:

Similarly:

Model Studies of Transreactions in Condensation Polymers

133

Assuming that the number of macromolecules remains constant during the transcondensation reaction, the following equations can be written:

If only the dyad fraction can be experimentally determined, the calculation of the average sequence lengths is possible provided that, in all triads, the substitution on the left of a central unit dies not influence the substitution on the right. In this case: (41)

PA;BjAk = &jAk

F’rom Eqs. (25) and (36)-(38), the sequence lengths become:

z=-

1

1 --

PAzBa

w=-- 1 %A1

hzAa

1

-

PAaBi

Fkom these equations, a value of B can be recalculated: 1 1

B=-+x w

(44)

(45)

(46)

Also, a correlation between the degrees of randomness B and B B can ~ be found. The introduction of Eq. (32) into Eq. (34) yields, for i # k:

Two cases are possible:

B B= ~ (PA~B~/FB (i ~ =)j B ) BBj = (pAkBj/FBj)B

( j = k)

(48) (49)

2.2. Kinetics of tmnsmaction

2.2.1. lhnscondensation involving two polymers [6] The kinetics of equilibrated transcondensation can be modelled using the overall reaction mechanism:

NAl-BlN PBT

+ N A ~PC- B ~ Nk’k+ N A ~ - Bcopolyester ~N + NAl-BzN

(50)

134

J. Devaux

For simplicity, the initial molar fractions of polycondensates 1 and 2 will be designated by a and b, respectively: a = (FA~B~)o = FA^ = F B ~ b = (FA~B~)o = FA^ = F B ~

(51) (52)

The molar fraction of AiB2 or AzBl dyads at time t will be given the value

X:

Assuming, its did Murano and Yamadera [7], a second-order reversible reaction dX

= kp(a - X)(b - X ) - k;X2 dt

(54)

for transreaction involving two chain linkages of similar nature (for instance, two ester groups), the copolycondensate can be assumed to be random at equilibrium, and

X e = ab k p = kk

(55) (56)

By this assumption, Eq. (54) can be written as:

A transcondensation ratio r = X/a can be defined, leading, by integration of Eq. (57), to a very simple kinetic expression: In [ b b r ] = k2t Equation (58) allows the determination of the transreaction rate constant k2 by measuring the transcondensation ratio. The latter can be obtained, as shown hereafter, by NMR or IR techniques. 2.2.2. lbnscondensation involving a polymer and a low molecular weight model compound [8]

As illustrated below, reactions between a polycondensate and a low molecular weight model compound are often used in order to study reaction mechanisms. This procedure makes it possible to follow closely the kinetics of these reactions as the transreaction leads to a drastic decrease of the polymer molecular weight, which is very accurately measurable. It can be

Model Studies of Transreactions in Condensation Polymers

135

assumed, as a main hypothesis, that upon reaction with a small molecule, the polymer chains will break randomly. In this case, the Casassa theory of statistical degradation [9] can be used as a basis of the kinetic analysis. If the reaction is a second-order one (first-order with respect to each reagent), the general kinetic expression is given by:

-dp dt

- k2p"41 -

(59)

where p represents the number of polymer reactive groups per unit volume and [A] is the concentration of low molecular weight species. This secondorder Eq. (59) applies strictly to polymers which are not subject to any thermal degradation in the temperature range of the transreaction. Also, the definition of p deserves to be explained in detail. For instance, in the case of P C exemplified below, owing to the well known instability of acid carbonate end-groups, the carbonate links are considered as a whole, and p is assumed to be equal to the number of carbonate groups per unit volume. However, in PBT, p corresponds to the number of ester groups per unit volume, ie., twice the number of repeat units. In the absence of thermal degradation, Eq. (59) can be integrated:

which can be rewritten in a more concise form:

with

In Eq. (62), p , represents the upper limiting value ofp at infinite molecular weight; no and n are the average degrees of polymerisation, at times zero and t respectively; [&I is the initial concentration of the low molecular weight species and EO is the initial number of chain ends. If the concentration of the low molecular weight species is assumed to remain constant over the duration of the reaction, a very simple equation, similar t o the integrated Casassa expression, is obtained: l/n with :

-

l/no = kLt

(63)

136

J. Devaw

When a statistical thermal degradation of the polymer is to be taken into account, Eq. (59) does not hold any more and the reaction rate is given by:

where k represents the kinetic constant of statistical degradation. Equation (65) clearly shows that, in this case, the reaction does not obey any definite order with respect to [A].However, if [A]can be taken as constant, Eq. (65) reduces to an apparently first-order expression which, after integration, gives an equation similar to the Casassa expression (63).

3. Application t o the PC/PBT system 3.1. Microstructural study PC and PBT are polyesters which are partly miscible, and thus susceptible to transcondensation reactions in a homogeneous phase. Their structural formulae are as follows:

PBT

L

CH3

Jn

Results on the methods of characterisation of the exchange reaction between PC and PBT have been reported previously using rather lowperformance spectroscopic methods [lo]. Spectra of higher resolution [ll] are obtained by Fourier transform (FT) 'H NMR, as shown for instance in Figurel, representing the 'H NMR signals of terephthalic protons in a PC/PBT system (60/40 by wt), exhibiting a transcondensation ratio of 0.56. The asssignment of peaks in Figure 1 is reported in Table2 and can be used for any transcondensation level. By integration of the corresponding peaks, the molar fractions of triads (and of dyads, using 13C FT-NMR) can be calculated. Following the theory reported above, a statistical analysis of the microstructure can be undertaken. Selected results [lo] me reported in Tables 3-6 using a PC/PBT (50/50 by wt) system (PC is Lexan 135 from General Electric and PBT is TENITE 6PRO from Eastman Kodak). Transcondensation of PC/PBT systems leads also to significant modifications in the IR spectra. These modifications can be correlated with NMR

137

Model Studies of Transreactions in Condensation Polymers

f

a

1

I

8.75

8.25

PPm [

7.75

Figure 1. 'H NMR (250MHz) signals of aromatic protons for a 60/40 PC/PBT system; transreaction ratio 0.56

Table 2. 'H NMR (250 MHz) chemical shifts of terephthalic protons as a function of the nature of the substituents in a 60/40 PC/PBT system; transreaction ratio 0.56 ~

Nature of substituents

Ri

Proton

Ra

Aliphatic

Aliphatic

Hl-H4

Aliphatic

Aromatic

HI -Hz H3 -H4

Aromatic

Aromatic

Hl-H4

Chemical shift 6 (PP4

8.12 f 0.005 (a)

8.16 f 0.01 8.19 f 0.01 8.28 f 0.01 8.31 f 0.01

(b) (c) (d) (e)

8.36 f 0.02 (f)

138

J. Devaux

Table 3. Fractions of triads centered on B1,and corresponding degrees of randdomness BB,,from 'H NMR results Reaction time (Din) ~ A ~ B~ ~A A~ ~B~ ~A A~ ~B ~ A ~ Theoretical values for a blend Experimental values for a copolyester Theoretical values for a statistical copolyester

30

60 100 200 ~

~

3

1.00 0.71 0.49 0.37 0.29

0.00 0.26 0.43 0.47 0.49

0.29

0.50

~

~

0.00 0.02 0.09 0.15 0.23

0.00 0.53 0.86 0.95 0.99

0.22

1.00

Temperature of reaction 253OC; 50/50 by wt =46.4/53.6 molar ratio

A1: butylene, -(CH2)4-;

A2: bisphenol A,

CH3 B1: terephthalate, -0-C

43

II

0

C-O-;

B2: carbonate, -0-C-0-

II

0 I,

0

Table 4. Evolution of the average length of butylene (x)and bisphenol A terephthalate (y) sequences as a function of time at 253'C Reaction time (min) Theoretical values for a blend Experimental values for a copolyester Theoretical values for a statistical copolyester

Calculation based M,

N

36 500, PC

30 60 100 200

X

?/

166 6.46 3.28 2.57 2.18

1.42 1.64 1.94

2.16

1.87

-

1.15

on triads. PC/PBT 50/50 by wt; mixture of homopolyesters: PBT

zn 15000. N

quantification, making it possible to follow transcondensation in the solid state. Figure 2 shows the IR spectrum of the same sample as in Figure 1, emphasising the new peaks, with respective assignments given in Table 7. The IR assignments given in Table 7 allow calculation of the transcondensation ratio. Indeed, the absorbance A of the 1070cm-' band (aromatic ester) provides a measure of F A ~ = BX ~ while the absorbance A0 of

139

Model Studies of Transreactions in Condensation Polymers

Table 5. Fractions of dyads and correspondingdegrees of randomness B from 13C NMR results; PC/PBT ratio, A1, A2, B1, B2 as in Table 3 Reaction BA ~ ~F BA ~~B FA ~ ~BB~ time (min) F A ~ F 0.00 0.464 0.00 Theoretical values for a blend 0.536 0.00 0.35 0.36 0.09 0.05 Experimental values 31 0.49 0.19 0.24 0.76 for a copolyester 75 0.35 0.19 100 0.32 0.21 0.21 0.20 0.84 200 0.28 0.24 0.26 0.20 0.97 Theoretical values for a statistical copolyester 0.29 0.25 0.25 0.22 1.00 Table 6. Evolution of the average length of sequences as a function of time at 253OC. The calculation is based on dyads, for a PC/PBT system as described in Table 4; z, y, z , w , m and n are defined in Eq. (16)

Theoretical values for a blend Experimental values for a copolyester Theoretical values for a statistical copolyester

Reaction time (min)

z

31 75 100 200

166 5.96 2.82 2.55 2.23

z -

1.17 1.53 1.68 1.85

w

- 5 9 1.33 5.80 1.93 2.58 2.32 2.11 2.32 1.93

2.16 1.87 2.16

m

n

-

-

23 8 38 13 39 13 41 14

1.87 41 15

the 727cm-' band (terephthalic aromatic ring) is a measure of F B ~ = a. From these two absorbances, the transcondensation ratio T = X / a can be derived. Figure 3 shows the linear correlation between transcondensation ratios measured by 'H NMR and IR techniques [ll].A linear regression gives:

3.2. Kinetic study

Using either the 'H NMR or the IR technique, the transcondensation ratio r = X / a can be determined. Figure 4 shows the results obtained using the P C and PBT samples [S] described in Tables 3-6. Kinetic equation (58) appears to be closely followed and the rate constant k ~ given , in Table 8, varies according to Eq. (67) where the activation

J. Devaux

1800

1600

1400

1200

1000

800

600

Figure 2. IR spectrum of a 60/40 PC/PBT system; transreaction ratio 0.56 1.0-

0.8

-

0.6

-

n

X

d

a.

Figure 3. Linear relationship between transcondensation ratios measured by IR (A/Ao) and by 'H NMR ( X / a ) .The linear regression corresponds to (66)

141

Model Studies of Transreactions in Condensation Polymers

Table 7. Characteristic infrared absorptions of the ester and carbonate groups in PC/PBT systems as obtained after transcondensation

IR absorption Groups

(cm-')

Aliphatic ester: -(O>C""

'

1720a

0- C H p

Aromatic ester:

1740' 1070b Aromatic carbonate:

0

1780a

- e - " O # w

Mixed aliphatic-atomatic carbonate: 0

1770'

"C = 0 (stretching) bpara-Disubstituted phenyl coupled to the oxygen atom

50

100

160

200

Figure 4. Transcondensation kinetics for a PC/PBT system (50/50 by wt): (e) 224.5, (A) 234, (W) 243.5, ( 0 ) 253'C

142

J. Devaux Table 8. Rate constant

k2

as a function of temperature

Reaction temperature ("C)

k2

x lo3'

(min-1)

224.5

3.97 4.0gb 7.66 6.08b 13.43 13.71b 21.16

234 243.5 253

aPC/PBT 50/50 by wt = 46.4/53.6 molar ratio bDuplicate values Table 9. Apparent rate constant Weight fraction (%) PBT

PC

20 40 50 60 80

80 60 50 40 20

k2

as a function of the PC/PBT ratio at 243.5"C x lo3 a (min-' )

k2

2.14 7.14 13.6 20.6 34.3

x lo3 (min-1)

k2

-

0.95 3.13 4.44 7.40

aTi concentration 0.021% (by wt) bTi concentration 0.012% (by wt)

energy is given in cal/mol. kz(min-') = 2.29 x 10l1exp(-31320/RT)

(67)

Theoretically, the influence of the polycondensate molar ratio should be fully taken into account in kinetic equation (58) through constants a and b. However, from the experimental results reported in Table 9, k2 appears clearly to be influenced by the PC/PBT ratio. In Table 9, a rapid increase in k2 with PBT concentration is noticeable. This was confirmed to be due to the titanium catalyst content in the PBT used. When by an appropriate washing procedure this T i concentration is reduced from 0.021% to 0.012% by wt, a decrease in the k2 rate constant is observed (Table 9). A logarithmic plot of the apparent rate constant as a function of Ti content yields a straight line (Figure 5 ) which can be best described by the following equation: exp(-31320/RT) kz(rnin-l) = 3.22 x 1025[Ti]2.5

(68)

143

Model Studies of Transreactions in Condensation Polymers

In k2 -4

-5

-6

0

I

-13.5

I

-13.0

In [Ti]

Figure 5. Transcondensation rate constant kz at 243.5"C as a function of Ti concentration in PBT. Open and full circles relate to the lower and the higher Ti contents, respectively (see Table 9)

From the kinetic results reported above, it clearly appears that the PCP B T transcondensation is catalysed by transesterification catalyst residues. This is often the case and necessitates a search for catalyst inhibitors. However, several mechanisms can be invoked, not only for the transcondensation itself, but also for its catalysis and its inhibition. The results of a study of these mechanisms for the PC/PBT system, by means of model compounds, is reported in the next section. 3.3. Mechanism of the PC/PBT transcondensation

3.3.1. Model compound study of the overall mechanism [8] The IR and NMR investigations and the kinetic studies reported above clearly show that the PC-PBT transcondensation gives rise to copolycondensates with four structural units. As summarised in Table 10, this transreaction can result either from alcoholysis of a macromolecular species (PC or PBT) by a hydroxyl-terminated polycondensate (PBT-OH or PC-OH,

144

J. Devaux

Reactions I and I' in Table lo), or from an acidolysis involving a carboxylterminated PBT macromolecule (Reaction I1 in Table 10). The reaction can also proceed via a direct ester interchange between a PC and a PBT macromolecule (Reaction I11 in Table 10). As already mentioned, a hypothetical reaction involving an acid carbonate end-group (- 0 -CO -OH) should not be taken into account, owing to the known instability of this type of carbonate group [12]. The direct condensation reaction, involving a -COOH and an -OH PBT end-group, is not strict0 sensu a transcondensation reaction. It is considered in the final discussion, in the light of the author's recent results of a PBT coupling reaction with poly(ethy1ene terephthalate) (PET) [13,14,15,16]. In this section, the most likely mechanism of the PC-PBT transcondensation will be investigated by the study of model reactions between a polymer (PC or PBT) and a small molecule. In particular, the alcoholysis of PC by a linear alcohol, n-hexadecanol, is used to simulate Reaction I and the alcoholysis of PBT by Chydroxybiphenyl is taken as a model of Reaction 1'. To determine the possibility of PC acidolysis by -COOH endgroups (Reaction II), the reaction between PC and ptert-butylbenzoic acid was investigated. Finally, the reaction between PC and butylene dibenzoate is taken as a model of the direct ester-ester interchange reaction (Reaction I11 in Table 10). As already reported [8], kinetic studies were carried out in benzophenone solutions under nitrogen. In a first set of experiments, the alcoholysis of PC by n-hexadecanol (Reaction I in Table 10) was investigated. This reaction was found to be acccurately described by the kinetic model of Eq. (60). A plot of the rate constant as a function of reciprocal absolute temperature leads to an activation energy of 28.6 kcal/mol: k2 = 1.73 x

(69)

10l2exp(-28.600/RT) cm3mol-'min-l

In a second set of experiments, PC acidolysis by pte&butylbenzoic acid was studied (Reaction I1 in Table 10). This reaction was found to be fast Table 10. Possible mechanisms of the PC-PBT transcondensation; AI, A2, BI, B2 as explained in Table 3

+ +

+PBT-fOH +PC+ +copolyester+ +PC+OH I' +POOH + P B v +copolyester +PBT+OH Transesterification I1 +PBT-fCOOH+ +PC+ +copolyester +PC+OH + COa by alcoholysis

Transesterification by alcoholysis

Direct esterification

I

+ +

I11 SPBTJ + +PC+ +copolyester+ copolyester

Ad32 dyad+ phenol end-group AzBi dyad alcohol end-group AaBl dyad unstable -0-CO-OH end-group AiBz AaBi dyads

+

+

+

145

Model Studies of Transreactions in Condensation Polymers

in the temperature range of interest, as shown by the Arrhenius expression for its kinetic constant: k2

= 2.37 x lo1' exp(-23.400/RT) cm3mol-'min-'

(70)

The possibility of PBT alcoholysis by 4-hydroxybiphenyl (Reaction I' in Table 10) was investigated in a third series of experiments. The overall reaction appeared as a first-order one [8], and the apparent rate constant was found to be equal to the kinetic constant of PBT thermal degradation [17]:

k + Icz[A] = k

(71)

By verifying Eq. (71) for any [A], it follows that k2 0, ie., the alcoholysis of PBT by the P C phenolic end-groups (Reaction I' in Table 10) does not occur at a measurable rate in the temperature range investigated. In a fourth set of experiments, the PC-butylene dibenzoate reaction was studied (Reaction I11 in Table 10). For this reaction, two different mechanisms have to be considered: direct ester-ester interchange and acidolysis of PC by the benzoic acid produced by the pyrolysis of butylene dibenzoate. Assuming a rate constant equal to that of PBT pyrolysis for the butylene dibenzoate pyrolysis and using for the acidolysis step the rate constant determined above for the acidolysis of PC by an acid of the benzoic group (Reaction I1 in Table lo), a steady-state value of the benzoic acid concentration can be calculated for the two-step indirect mechanism. The result (8.9x10-6mol.cm-3 at 250°C) seems too high to be experimentally justified. Thus, the most likely mechanism for the PC-butylene dibenzoate reaction is a direct ester-ester interchange. An Arrhenius plot of the second-order rate constant leads to the determination of a high activation energy of 59.5 kcal mol-': N

k2

= 1.6 x 102'exp(-59500/RT)

(72)

It is worth noting here that all the results of the study on model reactions between a polymer (PC or PBT) and a small model molecule were qualitatively confirmed later by means of more classical reactions involving only small molecules [ll]. From this series of model reactions, conclusions can be drawn on the general mechanism of PC-PBT transreaction. To this purpose, the initial rates of the various possible processes (Table 10) were evaluated for a PC/PBT system (50/50 by wt) at 250°C on the basis of the rate constants reported above (Eqs. (69), (70), and (72)). These values are given in Table 11,with reference to the reaction schemes of Table 10, and are compared to the apparent initial reaction rate of PC-PBT transcondensation calculated from Eqs. (54) and (67). To carry out these calculations, equal numbers of -OH and -COOH end-groups (1.37 x mo1.g-') were assumed for PBT.

J. Devaux

146

Table 11. Calculated initial rates of model reactions and of PC-PBT transcondensation at 250"C Reactions

Initial rate at 250°C (lo* mol g-l min-' )

Scheme I I PC alcoholysis I' PBT alcoholysis

5.1 Not measurable

Scheme 11 PBT pyrolysis PC acidolysis

9.1 10.5

Scheme 111 Direct ester-ester interchange

31.1

PC-PBT transcondensation

1970

Results of the calculations, as reported in Table 11, clearly show that PC-PBT transcondensation is far faster than any simple model scheme suggests. A catalysed mechanism is thus the only logical explanation: it was demonstrated that Ti residues in PBT do act as a catalyst (Eq. (68)). It was demonstrated that the PBT alcoholysis by phenolic species does not even occur in the presence of Ti catalyst [6,11,18], Thus, this rules out the chain mechanism represented by Scheme I in Tables 10 and 11. As Scheme I1 does not induce the microstructure actually observed for the copolycondensate (no equivalence between the amounts of AlB2 and A2B1 dyads), the most likely mechanism for PC-PBT transcondensation is a reversible direct ester-ester interchange catalysed by Ti residues present in the commercial PBT used. At equilibrium, a random copolycondensate with four different structural units, obeying Bernoulli statistics, is obtained. As shown in the theoretical treatment, the rate constants of the direct (kz) and of the reverse (k!J reactions are identical. This result is understandable, taking into account the similar bonding energies of broken and re-formed ester bonds. It should therefore be concluded that, at least at a first a p proximation, the reaction is driven only by entropy changes. 3.3.2. Model compound study of the catalysis mechanism (1 1)

In the majority of the reactions described above, the PBT used contained catalytic amounts of titanium, most probably in the form of tetraalkyl orthotitanate. In order to understand better the catalytic action of this titanium compound on PC-PBT transcondensation, model compound studies were undertaken, involving only low molecular weight species. Dimethyl terephthalate (DMTP) and diphenyl carbonate (DPC) were selected as model compounds for PBT and PC, respectively:

147

Model Studies of Transreactions in Condensation Polymers 0

0

CHJ-O-(!!~~-O-CHI

DMPT

0

DPC

e O - ( ! ! - O a

Two tetraalkyl orthotitanates were investigated, tetra(2-ethylhexyl) orthotitanate (TEHT) and tetrabutyl orthotitanate (TBT): C2H5

I

TEHT

Ti-(O-CH2-CH-(CH2)3-CH3)4 Ti-(O-(CH2)3--H3)4

TBT

In a first series of model reactions, DMTP reacted separately in the melt at 245°C with TEHT and TBT. The reaction products were then separated by high-performance liquid chromatography (HPLC) and analysed. Several concentrations and reaction times were used. A reaction of exchange of substituents between the terephthalate ester and the titanate was found to occur in each case, according to a mechanism already proposed by Mehrota [19], where the first step of exchange can be written as:

(73) 0

C&-O--Ti--(O--R)3

II / \ II

+R - O - C ~ C-O--CHs

R is an alkyl group, 2-ethylhexyl or butyl for TEHT or TBT, respectively. If the reaction duration is long enough, and if the initial concentrations are appropriate, this kind of exchange can lead to tetramethyl orthotitanate and dibutyl or di(2-ethylhexyl) terephthalate, respectively, in the cases of TBT and TEHT. The formation of such new terephthalates was confirmed by HPLC. However, the exchange reaction steps are each equilibrated, so the formation of new terephthalate species has to follow Bernoulli statistics, what was actually observed. Therefore, it was concluded from the results of this set of experiments that an exchange of substituents between terephthalate and orthotitanate esters actually occurs under the experimental conditions studied. Moreover,

J. Devaux

148

this exchange leads to a statistical distribution of substituents between terepht halates and orthot it anates. In a second set of model reactions, DPC reacted in the melt either with TEHT or with TBT, under the same experimental conditions. Again new peaks were detected by HPLC, and were assigned at first to the ester exchange species 0

Ti -(0-R)4

+ D O - cII - o a

-

0

(74)

with the possibility of complete substitution for sufficiently long reaction times. The mixed aliphatic-aromatic carbonates are known, however, to undergo an ester pyrolysis at high temperature: 0

0 0 - C - 0 II- R

-

O-H+COJ+RI (75)

Reaction (75) leads to the formation of phenol, COz and an alkene. In HPLC analyses, phenol was actually observed in significant amounts. No mixed aliphatic-aromatic carbonate was found, and this can be explained by Reaction (75). However, regardless of the nature of the titanate used, the same peak was found to occur in the HPLC chromatograms, which cannot be asssigned to tetraphenyl orthotitanate, which is insoluble under the experimental conditions. As a consequence, another reaction mechanism was sought. This mechanism is similar to that proposed by Mehrota [19] and by Fkadet and Markcha1 [20] for the reaction of tetracoordinated titanates with acids. Thus, the first step of exchange between DPC and a tetraalkyl orthotitanate could be: 0

With an excess of DPC, the reaction could progress upon completion of three exchanges on the titanate:

149

Model Studies of Transreactions in Condensation Polymers

0 30 0 - C - 0 Q II - t

/

Ti-(O--R)4

o \

-

(77)

However, the completion of all possible exchanges is improbable, as intermolecular exchanges are known to produce a more stable structure:

The latter reaction leads to a mixed aliphatic-aromatic carbonate undergoing pyrolysis under the experimental conditions, as shown by Reaction (75). Thus, the final products of the DPC-titanate exchange are phenol, COz, a mixed aliphatic-aromatic ether and a titanium compound (11) with a structure that is independent of the nature of the titanate. This is in agreement with the HPLC results. From Reaction (78), it appears that the intermediate product (I) of Reaction (77) is eliminated, and consequently the equilibrium is shifted to the right. Thus the stoichiometry of the reactions has to be 3 mol of DPC per mol of titanate, which is actually confirmed by the experiment. In a third set of model reactions, DPC reacted with DMTP in the melt (245°C) at a stoichiometric ratio in the presence of different amounts of TEHT, and for varying durations between 1 and 180 minutes. In the absence of titanium catalyst, no reaction was observed. This result seems to be in contradiction to the exchange previously observed between PC and butylene dibenzoate (Reaction I11 in Table 10 and Eq. (72)). First, the effect of a catalyst present in the reaction medium of PC and butylene dibenzoate cannot be totally ruled out, but it seems unlikely owing to the purification procedures and the reproducibility of the results. Also, the very high energy of activation (Eq. (72)) should lead to a very low level of reaction at 245"C, which could be observable on polymer reaction with a low molecular weight molecule, through the enhancement of sensitivity due

J. Devaux

150

to the long tail effect, but hardly measurable on classical reactions between two low molecular weight species. Another possible explanation requires reconsideration of the two-step scheme including the pyrolysis of the butylene dibenzoate followed by fast acidolysis of the PC by the benzoic acid produced (see discussion of Reaction 111in Table 10). The argument of a steady benzoic acid concentration can be overcome by an acidolysis step of zero-order, as proposed by Halek [21].Assuming a two-step mechanism for the transcondensation, the calculations lead, in this case, to an activation energy which is very close to that obtained for PBT pyrolysis (rate-determining step) [17]. Anyway, the slowness of the uncatalysed transcondensation reaction, should it actually occur, fully agrees with the common experience [22-241. It can therefore be concluded that the presence of a catalyst is necessary for a transcondensation reaction to take place at a significant level. When TEHT is added to the stoichiometric mixture of DMTP and DPC, an exchange reaction is easily detected by HPLC. The assignment of the peaks can be made on the basis of the overall mechanism:

[I

0

It

(79)

TEHT

0 0 - C -0-CH3+ As the distribution of exchange products has to be Bernoullian, dimethyl carbonate (DMC) and diphenyl terephthalate (DPTP) are to be found in the reaction products, as well as the mixed intermediates. 0

CHs-0-(3-0-

II

CH3

( D W

(DPTP) All the aromatic compounds were observed during HPLC analysis, except DMC, which could not be observed by the UV detector. Some of the products were found to be due to exchanges with TEHT. A limited kinetic analysis confirmed that the exchange reaction was a first-order one not only with respect to DMTP and DPC, but also with

Model Studies of Transreactions in Condensation Polymers

151

respect to TEHT. Moreover, the order of the reaction was shown to be the same, the direct and reverse kinetic constants being identical. All these model reactions fully confirm the previous results on polymer reactions. The first order with respect to the catalyst is more in agreement with the results of Pilati et al. [22,25,26]than the 2.5-order (Eq.(68)). This latter value was probably obtained due to the unknown structure and residual activity of the titanium catalyst used for the manufacture of the commercial PBT. A first order with respect to the titanate indicates, again following Pilati et al., that the reactional intermediate is probably a coordination complex of this titanate. Thus the overall mechanism of catalysis of the transcondensation between DMTP and DPC model compound could be written as follows: 0

0

0

Ti -(0-R)4 + ~ O - C - IIO ~ + C H 3 - O - C " 0 ! - O - c & DMTP

DPC

TI R-O\

0-R I ,O-R

R-O\

0-R

I ,O-R

J. Devaux

152

R-O\

0-R I 10-R

0 “&oQ (11)

Ti-(O-R)rtCH3-O-C 0

Model Studies of Transreactions in Condensation Polymers

153

3.3.3. Model compound study of the mechanism of inhibition by phosphite esters [ l l ] In previous publications, the author showed that several compounds, particularly those of phosphorus were very active in inhibiting the PC-PBT transcondensation [27,28].In order to understand the mechanism of this inhibition, some of these inhibitors were added to a reaction medium (245°C) containing equimolar amounts of DPC and DMTP, and 1% (by wt) of TEHT. The phosphorus compounds selected for this set of experiments were di-n-octadecyl phosphite (DNOP), diphenyl phosphite (DPP), triphenyl phosphite (TPP) and triphenyl phosphate (TPPa). It was previously shown that at least 2 mol of phosphite per mol of catalyst were necessary to fully inhibit the PC-PBT transcondensation [29]. In order to avoid problems of insufficiency of inhibitor, a two-fold excess (4 mol per mol) was used. Reaction products were analysed by HPLC. Results of these model reactions clearly show that no inhibition occurs when TPP or TPPa is added to the DPC-DMTP reaction medium. The reaction products are in fact identical to those obtained in the absence of inhibitor. On the contrary, DPP and DNOP effectively inhibit transcondensation. Quantitative analysis shows that only 0.4% and less than 5% of terephthalates are phenyl-substituted in the cases of DNOP and DPP, respectively. No changes are observed in the equilibria, nor in the kinetics, in the case of addition of TTPa or TPP. This last result is rather surprising owing to the strong inhibiting activity of TPP in the case of PC-PBT transcondensation. An explanation for this observation has to be found in the mechanism of inhibition. IR and UV-VIS analyses of several mixtures of TEHT with phosphorus compounds were undertaken [ll],showing that TEHT and both DNOP and DPP formed complexes, as evidenced by a broad IR absorption band between 1000 cm-' and 1050 cm-' and by a W-VIS absorption band extending from the UV region to 500nm. No broad IR band was observed for the TPP/TEHT and TPPa/TEHT systems, while the UV-VIS spectra showed no band extending beyond 400-450nm. Literature results also suggest that titanates form complexes with diphosphites [30,31] and that the repective products are yellow in colour. From the results on the first-order TEHT catalysis and on the TEHTdiphosphite complex formation, it seems reasonable to assume that the DPC-DMTP transcondensation is catalysed by a coordination complex of titanate with the esters, and that inhibition results from the formation of a more stable complex between the same titanate and a diphosphite. A

J. Devaux

154

structure of this latter coordination complex is proposed: R-0

R-2-

\

0-R

/

Tj - 9 - R

::

I

H 00 \& \J R’- 0-P P-0-R’ R’-0

/

\

0 -R‘

This suggested structure of the coordination complex explains most of the results obtained. Indeed, the molar ratio of at least 2/1 between the titanate and the phosphite results in hexacoordination of the titanium, inhibiting any further complexation of any ester compound. Also, the need for a diphosphite is explained, as diphosphites are known to tautomerize into diphosphonates (Equilibrium (81)), providing the proton for the hydrogen bridge. 0

H-O-P-(Oa)

2

.-”-(OD)

2

(81)

TPPa and TPP do not provide the formation of such a hydrogenbridged complex. The inhibition of the PC-PBT transcondensation by TPP should result from a chemical reaction of TPP with either -OH or -COOH end-groups of PBT, leading to the in situ formation of DPP, which was extensively studied later [13-161. It should be noted that the complexation of the titanate by the diphosphite does not inhibit the exchange reaction between TEHT and DPC or DMTP. Although the esterification reaction between -OH and -COOH end-groups is not directly concerned, its possible catalysis could be affected [321. Finally, it is worth mentioning that the above model reactions were confirmed by the PC-PBT transcondensation [ll].Moreover, attention was focused on a latency period observed when polymers (PC and PBT), rather than model compounds (DPC and DMTP), underwent transcondensation. This was clearly shown to be due to a miscibility phenomenon: PC and PBT are only partially miscible at the start of the reaction. Thus, during an initial (latency) period, the PC-PBT transcondensation rate is strongly reduced, in contrast to the model reaction between DPC and DMTP taking place in a miscible blend of low molecular weight reactants. Also, during this latency period, TPP added to a PC/PBT blend is transformed into DPP, leading to effective inhibition of the transcondensation. Thanks to

Model Studies of Transreactions in Condensation Polymers

155

the calculations of the reaction kinetics, it was possible to demonstrate that, at a processing temperature of 260°C, this inhibition lowers the level of transcondensation products below ca. mol (see also Chapter 6). 4. General discussion and conclusion

As mentioned in the Introduction, transreactions in condensation polymers are very common, due to the “living” nature of polycondensates in the melt. Therefore, the use of PC-PBT to illustrate the theory, as well as the exemplification by the same polymer pair of the use of model reactions for a profound study of the reaction steps, is certainly a limiting choice. Due to space limitations, it is not possible to present here an exhaustive literature survey of all transcondensation examples published, even in the recent literature. Only specific examples will follow, selected to illustrate some concepts developed in this chapter (see also Chapter 1). PC-PBT transcondensation studies by IR and NMR are now numerous; the same system was recently explored by FT-Raman spectroscopy [33], but only indirect evidence of transcondensation was obtained, in the form of a change in the C=O band shape due to the change in crystallinity of the PBT content. This is not really surprising, as carbonyl stretching does not usually give rise to strong bands in Raman spectroscopy. As major changes due to transcondensation are in the region of C=O stretches, it is concluded [33] that the IR technique is more suited to follow this type of t ransreactions. The characteristics and behaviour of the PC/PET system are very close to those of the the PC/PBT one [34,35]. However, latency time was observed to be considerably longer in the former case, since PET is less miscible with PC than PBT. Also, a loss of carbonate groups from a consecutive reaction was observed to occur and this was assigned to a decarboxylation of aliphatic-aromatic carbonates, with a corresponding loss of bisphenol A to explain the fairly stable molecular weight. A more satisfactory explanation was published later by Berti et al. [36], who reported a loss of cyclic ethylene carbonate. They also gave evidence of the formation of aromaticaliphatic ethers, which was only suggested earlier [34]. The transcondensation between PC and polyarylate (PAr) was also studied by means of model compounds [37]. By kinetic methods, it was shown that an exchange reaction takes place even in the absence of catalyst. The exchange reaction parameters were calculated despite some P C hydrolysis, which was unavoidable under the experimental conditions. Extensive studies of PC/PET and also of PET/PAr systems were undertaken by Fakirov et al. [38]. In these studies, the authors showed that transcondensation was occurring, leading to eventual randomisation, but that, due to crystallisation, annealing at a temperature well below the melting point results in a rearrangement of the random copolycondensates to

156

J. Devaux

blocky ones. While confirming the driving force for randomisation was an entropy increase, they explained the reordering as driven by the large enthalpy difference due to crystallisation (crystallisation-induced reaction). In the same study, the influence of transcondensation catalyst on the rate of both reactions was emphasized. In a subsequent study, Denchev et al. [39]showed that when this transreaction is carried out in a miscible system, PBT/PAr, the sequential reordering can be observed even at an unusually low temperature (140°C). The PBT/PAr miscible system was also studied by Fernandez-Berridi et al. [40], using a theoretical approach very similar to that presented in this chapter. Sequence lengths as well as degrees of randomness were determined. A kinetic study also showed the applicability of the kinetic equation (58) to this system and an Arrhenius plot led to an activation energy of about 40 kcaljmol. Moreover, the absence of a latency period is in agreement with the complete miscibility of the system. A low order of reaction with respect of titanate content was observed, however, with no clear explanation, except that some titanate catalyst may have been inactive but was taken into account. An interesting study of transreaction between PET and a liquidcrystalline polyester (LCP) was undertaken by Hong et al. [41]. It was shown that the addition of LCP to PET caused an improvement of the crystallisation rate of PET but the effect of a transreaction leading t o the formation of block copolycondensate does not seem very clear, except for improving the dispersion of LCP in the PET matrix. PET/PBT systems were used in the study of transreactions [15]because these polyesters are miscible in the melt. 13CFT-NMR [42] made it possible to determine sequence length distributions but some degradation was also observed. A renewing interest was focused on transcondensation studies with the commercial availability of poly(ethy1ene 2,6-naphthalate) (PEN). Guo [43] checked the possibility of transreactions of PEN with PC, PET, PBT, poly(ethy1ene iso-terephthalate) (PEI) and in systems containing a liquid crystalline polyester - polyhydroxybenzoate (PHB). Transcondensation was shown to occur in all the systems studied, except for those containing the LCP. Other studies also demonstrate the occurrence of transcondensation between PET and PEN, improving their miscibility [44]. Annealing temperature was found to influence transcondensation, but the blend composition appeared to have no effect. Also, transcondensation in PC/PEN systems was shown to occur by means of solubility tests, DSC, and FTIR and NMR spectroscopies [45], resulting in improved miscibility. Finally, a recent paper by Fortunato et at. [46]deserves to be mentioned, as it deals with the inhibiting effect of phosphorus compounds on model transcondensation reactions catalysed by TBT. The authors assume the complexation of the titanate by the phosphorus compounds, but they emphasise the effect of -OH groups on the inhibitor. Moreover, they did not

Model Studies of Transreactions in Condensation Polymers

157

notice any inhibiting activity of DPP, which contradicts the result exposed extensively in this chapter. This point, among many others, emphasises the need for continuous research in the field of modelling of transcondensation and related reactions.

References 1. W. H. Carothers, J. Am. Chem. SOC.51,2548 (1929) 2. G. Odian,Principles of Polymerization, 2nd Ed., Wiley Interscience, New York 1981 3. P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca, New York 1953 4. R. Yamadera, M. Murano, J . Polym. Sci.,5A-1,2259 (1967) 5. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci.,Polym. Phys. Ed. 20, 1875 (1982) 6. J. Devaux, P. Godard, J. P. Mercier J . Polym. Sci.,Polym. Phys. Ed. 20, 1901 (1982) 7 . M. Murano, R. Yamadera, J. Polym. Sci. 2,8 (1971) 8. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1895 (1982) 9. E. F. Casassa, J. Polym. Sci. 4,405 (1949) 10. J. Devaux, P. Godard, J. P. Mercier, R. Touillaux, J. M. Dereppe, J. Polym. Sci., Polym. Phys. Ed. 20, 1881 (1982) 11. D. Delimoy, LLM&mgesde polycarbonate de bispheizol A et de polybutyldne tir6phtalate”, PhD thesis, Coll. Hypotheses 7, Academia Ed. Louvain-laNeuve 1988 12. P. D. Ritchie, J. Chem. SOC.1054 (1935) 13. B. Jacques, R. Legras, J. Devaux, E. Nield, Makromol. Chem., Macromol. Symp. 75,231 (1993) 14. B. Jacques, J. Devaux, R. Legras, E. Nield, J. Polym. Sci., Part A , Polym. Chem. 34,1189 (1996) 15. B. Jacques, J. Devaux, R. Legras, E. Nield, Polymer 37, 1189 (1996) 16. B. Jacques, J. Devaux, R. Legras, E. Nield, Polymer 37,4085 (1996) 17. J. Devaux, P. Godard, J. P. Mercier, Makromol. Chem. 179, 2201 (1978) 18. L. Buxbaum, Ger. Pat. 2455025 (1974) 19. R. C. Mehrota, J. Am. Chem. SOC.76,2266 (1954) 20. A. Fradet, E. Marhchal, Adv. Polym. Sci. 43,51 (1982) 21. G. W. Halek, J. Polym. Sci., Polym. Symp. 74,83 (1986) 22. F. Pilati, P. Manaresi, B. Fortunato, A. Munari, V. Passalacqua, Polymer 22,799 (1981) 23. V. D. Keshav, M. R. Hemant, J. Appl. Polym. Sci. 30,205 (1985) 24. P. J. Flory, J. Am. Chem. SOC.6 2 , 2261 (1940) 25. F. Pilati, P. Manaresi, B. Fortunato, A. Munari, P. Monari, Polymer 24, 1479 (1983) 26. F. Pilati, A. Munari, P. Manaresi, Polym. Commun. 25,187 (1984) 27. J. Devaux, P. Godard, J. P. Mercier, “New Polymer Compositions and Their Preparation”, Brit. Pat. 1569296 (1976); Ger. Pat. 27107295 (1977); BE 175656 (1977) ; FR 7706229 (1977); NL 77002524 (1977); JP 26151/77 (1977); US 775360 (1977)

158

J. Devaux

28. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22, 229 (1982) 29. J. Devaux, “Cin6tique et m6canisme de la tmnsesterificatzon entre le poly(butyl6ne te‘re‘phtalate) et Ee poly(carbonate de bisphe’nol A)”, PhD thesis, UCL 1979 30. N. V. Koridze, V. M. Andreyeva, A. A. Tager, A. L. Suvorov, Y. A. Khustaleva, L. I. Widman, Potym. Sci. USSR27, 310 (1985) 31. S. J. Monte, G. Sugerman, FR 2339645 (1977) 32. B. Jacques, J. Devaux, R. Legras, E. Nield, Macromolecules 29, 3129 (1996) 33. M. V. Fellow-Jarman, P. J. Hendra, M. J. J. Hetem, Spectrochim. Acta, Part A 51, 2107 (1995) 34. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Part A, Polym. Chem. 24, 3301 (1986) 35. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Part A, Polym. Chem. 24, 3315 (1986) 36. C. Berti, V. Bonora, F. Pilati, Makmmol. Chem. 193, 1665 (1992) 37. J. Devaux, P. Devaux, P. Godard, Makromol. Chem. 186, 1227 (1985) 38. S. Fakirov, M. Sarkissova, Z. Denchev, Macromol. Chem. Phys. 197, 2837 (1996) 39. Z. Denchev, M. Sarkissova, S. Fakirov, F. Yilmaz, Macromol. Chem. Phys. 197, 2869 (1996) 40. M. Fernandez-Berridi, J. J. Iruin, I. Maiza, Polymer 36, 1357 (1995) 41. S. M. Hong, S. S. Hwang, Y. Seo, I. J. Chung, K. U. Kim, Polym. Eng. Sci. 37, 648 (1997) 42. S. C. E. Backson, A. M. Kenwright, R. W. Richards, Polymer 36,1991 (1995) 43. M. Guo, Polym. Prep. ACS37, 227 (1996) 44. D. W. Ihm,S. Y . Park,C. G. Chang, Y. S. Kim,H. K. Lee, J. Polym. Sci., Part A, Polym. Chem. 34, 2841 (1996) 45. M. J. Fernandez-Berridi, J. J. Iruin, I. Maiza, Macromol. Chem., Rapid Comrnun. 16,483 (1995) 46. B. Fortunato, A. Munari, P. Manaresi, P. Monari, Polymer 35, 4006 (1994)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 4

Copolymer Composition: a Key to the Mechanisms of Exchange in Reactive Polymer Blending

G . Montaudo, C . Puglisi, F. Samperi

1. Introduction

Widespread interest in the development of polymer blends with useful properties has resulted in numerous studies on the exchange reactions occurring in the melt-mixing of polymer systems containing reactive functional groups (ester, carbonate, amide, etc.) [l-401.Blends of condensation polymers such as polyesters, polyamides, polycarbonates, and in general all polymers bearing reactive functional groups inside the chain or at chain ends, may yield reactions of chemical exchange, when mixed in the molten state. These exchange processes are most important, since they lead to the formation of copolymers that may change the blends’ properties drastically. They may be induced by the presence of catalyst residues used in the polymerisation, or may be caused by reactive terminal groups (OH, COOH, NHz), originally present in the polymers or generated in situ by thermal and/or hydrolytic degradation reactions. Variable amounts of random, block, or graft copolymers are thus formed, often acting as partial compatibilisers of the initial blend, and affecting the mechanical properties and the thermal stability of the resulting material. Figure 1 is a schematic representation of the exchange and thermal degradation reactions which may occur in blends of two homopolymers

160

G. Montaudo, C. Puglisi, F. Samperi

500-

2 400n

v

A

5

E& 300-

Y

Ei

G

I

Thermal decomposition

I

100-

Thermally activated ezchange

b

Copolymer AB

...................................................... Molten state

A 200-

+

+

Activated exchange Copolymer AB (End-group or catalyrt

j

....................................................... Solid state A+B

Reaction time Figure 1. Schematic representation of the exchange reactions which can occur in polymer blends containingreactive functional groups as a function of temperature

A and B containing reactive functional groups, as a function of temperature. When the blend is processed in the temperature range 200-3OO0C, an AB copolymer is formed by end-group reaction or catalyst-activated exchange [140],whereas at temperatures above 3OO0C, a thermally activated exchange (ie., in the absence of a catalyst) and a simultaneous thermal decomposition process may occur [33-361. The studies concerning the temperature range above 300°C fall beyond the scope of this chapter. A detailed study of the mechanisms and of the kinetics of exchange and degradation reactions occurring in the molten state is necessary to obtain polymeric mixtures with known and reproducible properties, to avoid detrimental reactions and to control the final properties of the blend obtained. Often, however, the end-groups of the homopolymers used as blend components have not been taken into account, and some of the basic chemical reactions leading to the exchange have been ignored, so that during meltmixing the reactions take place in the complete absence of criteria making possible their chemical control. In this chapter, an attempt is made to provide a general model enabling the description of the possible reactions occurring during melt-mixing of reactive polymers, and to obtain explicit rules allowing their control. Modelling the behaviour of reactive polymer blends is possible by monitoring

161

Exchange Mechanisms in Reactive Polymer Blending

the composition of the copolymers formed in the melt-mixing process. However, the creation of such a model for reactive polymer systems requires the collection of appropriate and specific information. To this purpose, we have analysed various cases (listed below) encountered in the molten-state reactions of condensation polymer blends, and have sought in each case to define their distinctive features. 2. Exchange mechanisms of reactive polymers in the melt

The exchange reactions occurring during melt-mixing of polymer blends containing polyesters, polyamides, and polycarbonates at temperatures below 300°C may be activated by the presence of reactive chain ends or by the presence of appropriate catalysts [l-401.In the following, the exchange processes which may occur in these systems (Figure2) are discussed, attention being focused on the experimental parameters that can be used to discriminate among the various mechanisms. A general question, when dealing with exchange reactions, is the detailed mechanism by which they take place: by direct exchange between inner functional groups (ie., located inside the polymer chains) as shown in Figure2A and B, or by attack of outer functional groups (ie., reactive inner-inner

-

outer-inner

c

.

0

0 Intramolecular

c '

Reactive inner groups (-CO-0-; -0-CO-0-;

Reactive end groups (-OH;-COOH; -NH2)

-CO-NH-)

Figure 2. Schematic representation of the exchange reaction mechanisms which can occur in polymers containing reactive functional groups

162

G . Montaudo, C. Puglisi, F. Samperi

chain ends) on inner groups (Figure2C,D). Our analysis is based on the distinction between inner groups exchange (called here inner-inner) and reactive chain ends exchange (called here outer-inner). The inner-inner intramolecular exchange process, occurring between functional groups of the same homopolymer chain (FigurezA), produces a low molecular weight (MW)cyclic homopolymer and a shorter linear homopolymer chain. When an intermolecular inner-inner exchange process occurs between two homopolymers of different structure (Figure 2B), two linear copolymer chains are formed. The copolymer chains thus generated can, in turn, undergo exchange according to reactions A or B. When a sufficient number of exchanges have occurred, both linear and cyclic random copolymer chains would be present in the mixture [37,39].When the exchange occurs by an intramolecular outer-inner process (reaction C), a cyclic homopolymer and a shorter linear homopolymer chain are produced, similar to reaction A. The intermolecular outer-inner reaction D does not produce two copolymer chains (as in reaction B), because only the polymer chain bearing active chain ends (outer groups) can attack the inner group of the other polymer, yielding a copolymer molecule plus a shorter linear homopolymer chain. In this case, the homopolymer reacting through its inner groups undergoes a random scission process, and its MW should vary according to the random scission law [36]. The thermodynamic approach of Flory [41] for the interchange reactions in polyesters predicts that the molecular weight distribution (MWD) changes with reaction time until it reaches the most probable one. Flory’s result was derived under the assumption that exchange takes place by direct interaction of two inner ester groups [41], but the result is also valid when the driving mechanism is the attack of active chain ends on ester groups. This implies that most polymer properties display a very weak dependence on the exchange mechanism, and one has to select the most sensitive to mechanistic changes. In order to distinguish between inner-inner and outer-inner processes, one may look at parameters such as: (i) composition of the copolymer formed in the exchange reaction; (ii) dependence of the extent of reaction on the concentration of active chain ends; (iii) dependence of the reaction rate on the MW of the homopolymer reacting through its active chain ends. In the ester exchange reactions, the formation of copolymer molecules has been ascertained to occur as a result of two processes: initial formation of a block copolymer from the homopolymers and subsequent rearrangement to a random distribution of units along the copolymer chain [37,39]. In order to establish the specific exchange mechanism involved in the reaction between two different homopolymers A and B, ie., inner-inner or outer-inner, five cases have been assumed and are summarised in Table 1. If in an equimolar mixture of homopolymers A and B (Table 1, Case l), the exchange reaction takes place by direct interaction of two inner groups (assuming also that the reactivity of these groups is nearly the same), the

Inner-Innerait' Case 4

a

Reaction goes to comple- Reaction goes to comple- Reaction stops as soon as tion tion reactive chain ends are consumed Reaction rate dependent Reaction rate dependent Reaction rate dependent on MW of A and B on MW of A on MW of A and B

Initial feed: equimolar mixtures of homopolymers A and B; Equal reactivity of inner groups in homopolymers A and B is assumed

Reaction rate independent of MW of A and B

Reaction goes to completion

Reaction rate dependent on MW of A

Reaction goes t o completion

Copolymer composition is Copolymer composition is Copolymer composition is Copolymer composition is independent of extent of dependent on extent of independent of extent of dependent on extent of reaction reaction reaction reaction

Copolymer AB contains an excess of A chains, in the early stages of reaction; it decreases in the later stages

Copolymer composition is independent of extent of reaction

Copolymer AB formed in Copolymer AB contains early stages contains an an excess of A units excess of A chains; in later stages AB has composition 50/50 (by mol)

Only polymer A has reactive chain ends Unreactive chain ends of type B are generated and do not participate to the reaction Reactive chain ends of type A are consumed and regenerated by thermal cleavage of polymer A during the reaction

Polycarbonate and NH2terminated polyamides

Case 5

Copolymer AB 50/50 (by mol) is formed

____

Case 3

OH-terminated polyesters Capped polyesters and Polycarbonate and OHOH-terminated polyesters or COOH-terminated polyesters Polymers A and B have Only polymer A has reac- Only polymer A has reacreactive chain ends tive chain ends tive chain ends Reactive end groups are Reactive chain ends of Unreactive chain ends of type B are generated and type B are generated and regenerated participate t o the reaction do not participate t o the reaction Reactive chain ends of Reactive chain ends of type A are regenerated type A are consumed durduring the reaction ing the reaction

Case 2

Outer-Innera

Catalyst not needed

Copolymer AB 50/50 (by mol) is formed

Polymers A and B have reactive inner groups

Capped polycarbonate and capped polyesters

Case 1

Catalyst needed below 3OOOC

0

F

Q,

G 9

7

d

6c.

ti'

v1

K

E

5

D

6. 5

164

G. Montaudo, C. Puglisi, F. Samperi

concentration of functional groups of species A and B is equal, and each species has the same probability of reacting. Then, the composition of the AB copolymer formed is predicted to be close to 1:l (feed ratio), regardless of the extent of reaction, and the reaction would proceed to the complete transformation of the two homopolymers. Now, assume that the exchange reaction takes place between active chain ends (for instance, alcoholic groups in polyesters). If both homopolymers possess these reactive end-groups (Table 1, Case2), in an equimolar mixture, the predicted result is the same as above and the molar ratio of components in the copolymer formed should be 1:l. Therefore, in this case, the copolymer composition is irrelevant to the establishment of the mechanism of exchange. A difference can be found in the rates of the two reactions: in the inner-inner process the reaction rate is independent of the MW of both homopolymers, whereas in the outer-inner process it essentially depends on the number of end-groups present. However, if only one of the homopolymers (A) possesses alcoholic reactive chain ends (Table 1, Case3), only the latter would be able to attack the inner ester groups of polyester B, thus cutting (statistically) the polyester B chains into two pieces. Provided that the homopolymer A has a relatively high molecular weight, the AB copolymer initially formed would contain an excess of A units. Taking into account that each exchange reaction generates an alcoholic end-group of B type (not existing initially), which would then participate in the reaction, A and B end-groups would both be present in the reaction mixture. Therefore, the composition of the copolymer formed in the later stages of the reaction is expected to equilibrate and to come close to 1:l (feed ratio), ie., with the progress of the exchange reaction, this system tends to behave as in Case 2, where both types of end-groups are present. Case 3 allows a discrimination between the two mechanisms of copolyester formation. In Case 4 (Table l), only one of the homopolymers possesses alcoholic reactive chain ends (type A), capable of attacking the inner ester groups of polymer B. However, unlike Case3, the chain ends of type B generated in the exchange are assumed to be unreactive, and the reaction should stop when the reactive chain ends initially present are consumed. In this case, the AB copolymer formed would contain an excess of A units, regardless of the extent of exchange. In the final Case 5 (Table l),the situation is the same as in Case4 and the reactive chain ends of type A are consumed as above. However, in Case5 the reaction does not stop because the chain ends of type A are continuously regenerated by a thermal (hydrolytic) cleavage of homopolymer A during the reaction. In this case, the copolymer AB initially formed contains an excess of A chains, which decreases in the later stages of the reaction. The copolymer composition is therefore dependent on the extent of exchange, and the reaction may go to cornpietion. The approach outlined above, based on the determination of the copoly-

Exchange Mechanisms in Reactive Polymer Blending

165

mer composition as a function of the extent of exchange, appears applicable as a general mechanistic criterion for exchange reactions and for the prediction of the copolymer compositions attainable in specific systems. We have investigated the dependence of copolymer composition on the extent of reaction in the systems described in Tablel, and in the following sections we shall compare the predictions with the results observed in the cases of polyester/polyester [37-391 , polyester/polycarbonate [40] and nylon G/polycarbonate [36] blends.

3. Exchange reactions occurring by inner-inner mechanism (Case 1)

3.1. Capped PBT/PC blends To investigate the dependence of copolymer composition on the specific exchange mechanisms, poly(buty1ene terephthalate) (PBT) samples bearing no end-groups were synthesised and reacted in the melt with a high MW polycarbonate (PC) sample for different times [40]. The reactions that can occur in this system are summarised in Scheme la-c. The exchange reaction between high MW PBT and PC, ie., containing non-detectable amounts of reactive end-groups, occurs even at 240°C in the presence of an esterification catalyst, by an inner-inner mechanism (Scheme la) [4-13,401. To explore the dependence of the copolymer composition on the inner-inner exchange mechanism, we have reacted an equimolar mixture of high MW PBT and PC at 240°C in the presence of Ti(OBu)4, in a Brabender mixer for different times [40]. PC/PBT blend samples taken at different time intervals were selectively extracted first with tetrahydrofuran (THF) and then with tetrachloroethane (TCE), and characterised by nuclear magnetic resonance (NMR) spectroscopy. THF is able to extract selectively the unreacted PC, whereas PC/PBT copolymers are soluble in TCE at room temperature. The residue of the extraction is composed of unreacted PBT. Initially, a block copolymer is formed, which randomises as the reaction time increases. The material balance of the reaction, reported in Table2, shows that the weight of copolymer formed is 25% after 2min and that the reaction goes to completion in about 10 minutes (Figure3a). The composition of the copolymer, indicated as the molar ratio of terephthalate to bisphenol A units (obtained by 'H NMR), was monitored as a function of the reaction time and was found to be 50:50 (Figure3b), ie., equal to the feed ratio of the two homopolymers, regardless of the reaction time, its predicted (Table 1) [40]. To verify whether the occurrence of the inner-inner mechanism of exchange is dependent on the presence of the catalyst, we reacted PC and PBT, with the reactive chain ends capped, in the absence of catalysts [40]. The uncatalysed exchange reaction between PC and a PBT sample with

166

G. Montaudo, C. Puglisi, F. Samperi a) direct ester-carbonate exchange (inner-inner)

-0-co-0-

I T +

+ b) alcoholysis (outer-inner)

-0-y-0-

&

*O-CO-O-(CHZ),-O-OC

+ c) acidolysis (outer-inner)

+

-COOH

O H-

+CO1

Scheme 1. Exchange reactions which can occur in the melt-mixing of PBT/PC blends

E.

WC~I

Catalysed PC/PBT (24OOC) Uncatalysed capped PBT/PC (29OOC) Uncatalysed PC/PBT-OH (240' C) Reaction Weight % of Copol. Degree of Reaction Weight % of Copol. Reaction PC/PBT3000 PC/PBT6000 PC/PBT18000 time extracted fractionsa comp.b randomtime extracted fractionsa comp.b time rn nesC (min) 'Copol.JPCJ PBT (min) W C ~ JC C ~ CCe W C ~ JC C ~ i;' (min) Copo1.J PC J PBT

168

G. Montaudo, C. Puglisi, F. Samperi

I

0

~

20

l

40

~

l

60

~

80

l

100

~

120

l

1 D

~

I

Reaction time (min)

0

20

40

60

80

100

120

140

Reaction time (min) Figure 3. Weight percent (a) and composition (b) of the copolymer produced during m e l t - d i g of PC/PBT blends at 240°C in the presence of catalyst (innerinner, Case 1)

the chain ends capped by benzoate groups does not take place at 240"C,but it was found to occur at temperatures above 270°C;the extent of exchange increases with the temperature (Figure 4a) [40]. The time dependencies of the weight and composition of the copolymer produced when an equimolar blend of capped PBT/PC was reacted at 290°C are reported in Table2. It is seen that 35%of the copolymer weight is

'

l

'

169

Exchange Mechanisms in Reactive Polymer Blending

--I t:

a

40

-8 30 8 b O 20-

Y

8

Y

-

d 100-

$ i

'

i

'

i

'

l

'

i

'

i

'

l

Temperature ('C) 100

b

80.* .Y

i60- U

b

a 40-

3 0

.

u8 200

0

10

20

30

40

Reaction time (min)

50

60

'

Figure 4. Capped PBT/PC blends melt-mixed in the absence of catalyst: (a) extent of exchange as a function of temperature after 60min of reaction; (b) composition of the copolymer formed at 290°C us. reaction time (inner-inner, Case 1) formed after 5 min and that the reaction goes to completion within 10min. The molar ratio of terephthalate t o bisphenol A units of the extracted copolymer was found to be 50:50 (Figure4b), ie., equal t o the feed ratio of the two homopolymers, and therefore indicating the occurrence of an innerinner exchange mechanism. This result is in agreement with the predictions (Table 1) and also with the results obtained with the catalysed reaction.

170

G. Montaudo, C. Puglisi, F. Samperi

3.2. Capped PET/PC blends As in the case of the PBT/PC blend discussed above, to correlate the composition of the copolymers with the exchange mechanisms, we have reacted a 50:50 (by wt) blend of high MW P C and an end-capped poly(ethy1ene terephthalate) (PET) sample at 27OoC, in the presence of Ti(0Bu)r [40]. The catalysed exchange reaction of high M w PET and PC is likely to occur by an inner-inner exchange mechanism analogous to that for the PBT/PC system, generating a four-component copolymer (Scheme 2, structures A,B,C,D). However, at this temperature (270°C) the aliphatic carbonate units formed are thermally unstable and undergo thermal decomposition by the evolution of COz and ethylene carbonate (ETC). The evolution of COz causes the formation of ether groups along the copolymer chain, thus introducing a new component (Scheme 2E) into the copolymer

C

C

Ethylene Carbonate

Scheme 2. Exchange reactions occurring in the catalysed melt-mixing of PC/PET blends

Mixing time (min)

I

Weight % of extracted fractionsa Copol. PC I P E T Copol. cornpb Et/TC

T/Bd

TG % Weight losse

I

NMR % Weight loss' C02 ETC

I

I

NMR Molar amount of copolymer structural unitsg MA M B 1 M c I M D ME

8' 0

172

G . Montaudo, C. Puglisi, F. Samperi

sequence [15-18,34,40]. The evolution of C02 and butylene carbonate was not observed during melt-mixing of the PBT/PC system because reaction was at a temperature (240°C) at which the butylenelbisphenol A carbonate moiety is still thermally stable, whereas the evolution takes place at temperatures above 300°C [35]. The material balance of the PET/PC exchange reaction, reported in Table 3, shows that after 2 min the copolymer is essentially a four-component one (Scheme 2A,B,C,D) that becomes later a five-component copolymer (from 4 min up to 40 min). After this time, the A and D carbonate sequences are totally consumed by the evolution of C02 and ethylene carbonate, and a threecomponent copolymer (B,C,E) is produced (Figure 5). The weight of copolymer formed after 2min of mixing is 70% and the reaction goes to completion after 4min (Figure6a). The weight loss of ETC and C02, calculated by NMR, is in agreement with the weight loss data obtained by isothermal thermogravimetric analysis at 270°C of a PET/PC sample in the presence of catalyst (Table 3). The composition of the copolymer obtained in the catalysed reactive blending of PET and PC is represented as a function of reaction time in Figure6b. It can be noted that the molar amount of the terephthalate units in the extracted copolymers is 57%, i e . , identical to the feed composition, and it remains constant during the reaction time, as predicted (Tablel) [401.

60-

.41

50-

3 40-. !’330--

c

c1

20:

6& 10-. 0~

,

0

.

l

.

l

10 20

,

l

,

,

,

l

.

,

.

l

.

l

.

i

.

i

.

,

.

l

30 40 60 60 70 80 90 100 110 120

Reaction time (rnin)

Figure 5. Amounts (mol %) of structural units in the copolymer as a function of the reaction time of a PC/PET blend melt-mixed at 270°C in the presence of catalyst: 0 A; 0 B; A C; * D; V E (structural units as indicated in Scheme 2)

r-

Exchange Mechanisms in Reactive Polymer Blending

$,loo; u 80 8 V 60

c bl 0 .

h

0

-

173 a

40-

3 $? 20o , , , , , . , , , , , , , , , , 80

30

b

0

10

20

30

40

50

Reaction time (min)

60

70

E

Figure 6. Weight percent (a) and composition (b) of the copolymer (mol % of ET units) produced in the melt-mking of PC/PET blend (Case 1) at 270°C in the presence of catalyst

4. Exchange reactions occurring by outer-inner mechanisms

4.1. PET/PTX blends (Case 2)

As mentioned in Section 2, among the examples listed in Tablel, Case2 relates to systems where the exchange reaction is brought about by reactive chain ends (for instance, alcoholic groups in polyesters). If both homopolymers possess these reactive end-groups (Tablel, Case2) in an equimolar mixture, the same result as above is predicted and the molar ratio of components in the copolymer formed should be 1:l. The composition of a copolyester formed during melt-mixing of equimo-

174

G. Montaudo, C. Puglisi, F. Samperi

lar amounts of PET and poly(ethy1ene truxillate) (PTX) was reported to be in agreement with the above prediction [38].The respective experiments yield the same result as in Case 1, and therefore the copolymer composition is irrelevant, in this case, to establishing mechanism of exchange. However, a difference between Case 1 and Case 2 might be found by monitoring the kinetics of the two reactions: in the inner-inner process the reaction rate is independent of the M W of both homopolymers, whereas in the outer-inner process (Table 1) it depends substantially on the number of end-groups present. At present no kinetic data are available for such a system. 4.2. PET/PEA blends (Case 3)

A kinetic study on the composition and sequence of the copolyesters formed in the exchange reactions which occur during melt-mixing of poly(ethy1ene adipate) (PEA) and PET was reported [37]. It is interesting to note that even if melt-mixing occurs in a closed system (where the sample composition is constant), the composition of the resulting copolymer may deviate from the value given by the molar ratio of the two homopolymers present (Table 1, Case 3). A high MW PET sample (practically free of reactive end-groups) was melted with an equimolar amount of uncapped low MW PEA, without catalyst; the exchange kinetics, at 290°C, was followed by analysing the reaction mixture by 'H NMR and fast atom bombardment mass spectrometry (FAB-MS) [37]. The copolyester initially formed in the exchange reaction (Scheme 3) proved to be a block copolymer containing an excess of EA units, as expected for a reaction proceeding by the attack of PEA end-groups on the inner functional groups of PET chains (Table 1, Case3). In the later stages of exchange (as expected), the copolyester sequence was observed to approach a Bernoullian distribution and an equimolar composition [37]. Two different methodologies [42] can be used for the analysis of mass spectra of polymer mixtures undergoing exchange processes: the first approach (data partitioning) considers the exchange reaction as a whole, the second (data best-fitting) focuses exclusively on the modifications of the reaction products, i e . , the copolymer formed [37].In the data partitioning approach, the experimental MS peak intensities are partitioned into various groups, each group having a different size (dimers, trimers, tetramers, pentamers, etc.). Taking a linear combination of the experimental MS intensities belonging to each of these groups, it is possible to obtain a first characterisation of the sample. Each group of peaks yields an independent estimate of the molar fraction of comonomer A in the sample (SA)and in the copolymer (CA) [42]. On the other hand, thermodynamics requires that the number-average sequence lengths ( n A ) and ( n g ) decrease steadily with the progress of the

0

II

0

II

CH2)4-C -O-(CH2)2-O-C 0

II

I

0

0

_Q_0-fc(-H22)0-*

0

II 0

II

1

Further reaction

COPOLYMER

0

0

0

~ ~ - O - ( C H ~ ) Z - O - A M

f -0-( CHZ)~-OH

+

CH2)2-O-C

Alcoholysis (outer-inner)

-( CH2)4-C-O-(

0

II

I

4

Scheme 3. Exchange reactions occurring in the melt-mixing of PEA/PET blends

*C-(

0

f-0-(CH2)2-0-c

*fO

+

G. Montaudo, C. Puglisi, F. Samperi

176

exchange reaction. In fact, copolymer chain configurations exhibiting a large number of consecutive A or B units (and therefore representing highly ordered configurations) have a lower entropy than chain configurations in which A and B units are found randomly along the chain, As stated above, in order to monitor the rearrangements occurring in the structure of the copolymer, the best-fitting method can be used. In this approach, the peak intensities corresponding to homopolymers are completely neglected, since the only peaks directly related to the microstucture of the copolymer are the copolymer peaks [42,43]. The statistical analysis of copolymers makes use of the Markoffian model in order to characterise the sequence of copolymer samples. Assuming a theoretical distribution, and then fitting the calculated abundances with the experimental peak intensities, the copolymer composition can be determined [42-441. According to chain statistics [42-441 , the sequence distribution of a two-component copolymer is completely defined when the four elements (Pxy)of the probability matrix are determined. The latter are linked by the following normalisation conditions:

PAA+PAB= 1

PBB + & A

=1

(1)

Each Markoffian theoretical distribution has an associated composition and an associated number-average length of like monomers [45], given by: CA

= PBA/(PAB i-& A )

( n ~=) ~ / P A B ( n ~=)I/&A

(2) (3)

The FAB-MS spectrum of the powdered mixture of the two polyesters, PEA and PET, before melting (sample MO) is shown in Figure 7a; the complete absence of MS peaks corresponding to species containing ET units is due to the fact that the PET sample was completely free of low M W oligomers (a comparison with the spectrum of a pure PEA sample showed no differences). The FAB-MS spectra shown in Figure 7b,c were obtained from samples melt-mixed for 20min (M20) and 270min (M270), respectively. No linear oligomers were detected: in fact, the labelled peaks in these spectra belong to a single mass series due to the protonated molecular ions of cyclic esters. The MS peak intensities of each of the eight MS spectra of PET/PEA samples, reacted for 10 to 270 min (Table4), were grouped according to the data-partitioning approach to estimate the copolymer composition and the extent of exchange [37].The results indicate an excess of EA units at the early stages of the reaction, which levels to almost equal amounts of EA and ET in the later stages. The extent of exchange (EE) increases with reaction time. In the last three samples the EE is well above 90%, and this implies that the exchange reaction is practically over, ie., the samples represent pure copolymers [37].

Exchange Mechanisms in Reactive Polymer Blending

177

(a)

I%

-

A,E.

517

Figure 7. Positive FAB mass spectra of PET/PEA blend reacted at 290°C for: (a)0 min; (b) 20 min; ( c )270 min However, from the data-partitioning approach it is not possible to determine whether the copolymer number-average block lengths, ( n ~ )(ring) , , decrease steadily, as predicted on thermodynamic grounds. In order to extract the copolymer sequences from the mass spectrometric data, the in-

G. Montaudo, C. Puglisi, F.Samperi

178

Table 4. Results of the analysis on FAB-MS spectra of the PEA/PET melt-mixed samples using the best-fitting approach [37] MO MI0 M20 M30 M45 M60 M90 M120 M270 a

1.00/0.00/0.00/1.00 0.84/0.16/0.37/0.63 0.85/0.15/0.32/0.68 0.82/0.18/0.34/0.66 0.68/0.32/0.48/0.52 0.65/0.35/0.50/0.50 0.59/0.41/0.55/0.45 0.50/0.50/0.66/0.34 0.50/0.50/0.53/0.47

100/0 70130 68/32 65/35 60140 58/42 57/43 56/44 56/44

53 47 52 80 85 96 94 97

6.3 6.5 5.4 3.1 2.8 2.4 2.0 2.5

2.7 3.1 3.0 2.1 2.0 1.8 1.5 1.9

20.0 7.4 5.4 0.4 0.1 0.2 0.8 1.0

First-order MarkoRian P-matrix resulting from best fit

(PEAEAIPEAETIPETEAIPETET)

Calculated copolymer composition (EA/ET molar ratio) using Eq. (2) Calculated extent of exchange Calculated number-average lengths of EA and ET units in the copolymer by Eq. (3) Calculated agreement factor

tensities of the peaks corresponding to homo-oligomers were excluded from the calculation and the reduced set of data was analysed by the best-fitting approach. The MAC04 computer program [42,43) was employed to simulate the experimental data and to characterise the sequence distribution of the eight copolymers contained in the melt-mixed samples. In Table4 (column 2), we report the eight P-matrices (each made of four matrix elements) which yielded the lowest error level (agreement factor, AF) in the simulation process. The remaining columns in Table4 report the attributes

0

50

100

150

200

250

300

Reaction time (min)

Figure 8. Number average lengths (TLEA)and 0 ( n m )us. reaction time of the copolymer formed in the PET/PEA blend reacted at 290°C (Case 3)

179

Exchange Mechanisms in Reactive Polymer Blending

of each P-matrix: the associated copolymer composition, the associated extent of exchange, and the associated number-average block length in the copolymer. The curves in Figure8 show the variation of the copolymer numberaverage lengths ( n ~ )(,n g ) with the reaction time, and indicate that the predicted initial formation of block copolymers actually occurs, as well as that at longer reaction times a random distribution is approached [37]. Figure 9 illustrates the behaviour of the copolymer composition as a function of time: the amount of EA in the copolymer was found to decrease gradually from nearly 100% down to 50%, in excellent agreement with the predictions of the model (Tablel). These results are based on MS spectra taken from the crude melt-reacted samples. In these samples, the composition of low MW co-oligomers (which was detected by FAB-MS) may differ from the composition of the high MW copolymer. Therefore, the computations are based on the assumption that this difference is negligible. The validity of this approximation can be checked by analysing the aminolysed copolymer samples. In fact, during the aminolysis, the size of the high MW macromolecules is reduced and the latter are transformed into low oligomers [46]. Therefore, for these partially degraded samples, the differences between the “true” composition and the composition detected by MS becomes negligible. Copolymer compositions derived from the aminolysed samples (Figure9) display some variations with respect to the crude samples; nevertheless the time-dependence of the copolymer composition is the same [37]. These results indicate that the attempt to derive the overall copolymer composition from the analysis of low MW species is justified, since they actually reflect the overall properties of the sample. In conclusion,

-5

s

50

;

.

. . . ,

50

.

.

..>.

100

,. ”

0

.,. . .. , . . .. , 150

200

Reaction time (min)

.=.-

250

3( 3

Figure 9. Copolymer composition (mol % of the EA units), formed in the meltmixing of the PET/PEA blend at 29OoC,as a function of reaction time, calculated from the positive FAB-MS spectra of (0)crude and ( 0 ) aminolysed samples (Case 3)

180

G.Montaudo, C.Puglisi, F. Samperi

the copolymer composition was found to vary with the reaction time (Figure 9), and this is considered t o be evidence in favour of an active chain-ends mechanism (Table 1, Case 3). Monte Carlo calculations were also performed (391 to simulate the exchange reactions occurring in Casesl-3, and the results confirm the predictions of the present model (Table 1). The method showed a remarkable predictive power in the case of exchange reactions, and the salient features of these Monte Carlo calculations are summarised in the Appendix to this chapter. 4.3. PBT/PC blends (Case4)

The exchange reaction of high M W PC with a PBT sample containing hydroxyl end-groups takes place according to an outer-inner mechanism (Figure2D, Scheme 2B), and it does not require the addition of catalysts. The exchange proceeds by the attack of the hydroxyl end-groups on the PC chains, resulting in block copolymers of PC and PBT joined through butyl carbonate units, plus low MW PC with phenol end-groups. The latter are unreactive at the reaction temperature of 240°C and therefore the exchange stops as soon as the reactive hydroxyl end-groups of PBT are consumed (Table 1) [40]. To investigate the dependence of the concentration of active chain ends on the copolymer composition, three samples of PBT bearing OH endgroups with increasing molecular weight (3000, 6000, 18000) were reacted with PC at 240°C for different times. The data relative to the material balance of the reactions of PC with the three samples of OH-terminated PBT, obtained by selective extraction and by l H NMR analysis, are reported in Table2. It can be noted that the amount and the composition of the copolymers obtained are independent of time (see also Figure 10a,b). The weight of copolymer formed in the reaction increases with the MW of the PBT used (Figure lOa), whereas the copolymer composition shows an excess of PBT units with respect to the feed molar ratio (Figurelob), as predicted for Case4, Table 1[40]. The remarkable difference from the analogous case of the PET/PEA ester-ester exchange (Case3) [37] is that in the ester-carbonate exchange, the reactive hydroxyl groups of PBT are consumed in the reaction and are not regenerated (because phenoxy groups are unreactive: Scheme 2B, Table 1). Instead, in the PET/PEA system the reactive hydroxyl groups are regenerated and the composition of the copolymer shows an excess of PEA units at the onset of reaction, which then equilibrates and comes close to the feed molar ratio of the homopolymers, in the later stages of the reaction [37,39].

181

Exchange Mechanisms in Reactive Polymer Blending

l90

o

0

20

o

40

60

l

80

100

121

Reaction time (min) Figure 10. Weight percent (a) and composition (b) of the copolymer produced in the melt-mixing at 240°C of PC/PBT-OH blend in the absence of catalyst (Case4). PBT-OH M W : W 3000; 0 6000;0 18000

G. Montaudo, C. Puglisi, F. Samperi

182 4.4. PET/PC blends (Case4)

As in the case of PBT/PC blend (Scheme lb,c), the uncatalysed exchange reaction of high MW PC with a PET sample containing carboxyl endgroups occurs by an outer-inner mechanism (Figure 2D) [40]. Melt-mixing of PET/PC was performed at 270°C and was monitored by 13C NMR [40]. Peaks corresponding to carbonyl and to quaternary carbons of terephthalate carboxyl end-groups disappear after 10min and are substituted by the peaks corresponding to carbonyl and to quaternary carbon of aromatic terephthalate units, which are generated during the exchange reaction. These units constitute the links between the PET and PC blocks 100

a

40

0

20

10

30

40

50

70

60

Reaction time (min)

80

I

b

d

.g 70Y

8 608

’a

CI CI

8V 0

-

5040-

30 301 0

I

10

,m

I

20

-

,

]I

30

,

1

40

*

I

50

,

Reaction time (min)

II

60

-,

1I

70

t

Figure 11. Weight percent (a) and composition (b) of the copolymer produced in the melt-mixing of PC/PET-COOH blend (Case4) at 270°C in the absence of catalyst

Exchange Mechanisms in Reactive Polymer Blending

183

in the copolymer. The phenol terminal groups of PC are unreactive and the exchange reaction stops as soon as the carboxyl reactive groups are consumed [40]. As in the case of PBT/PC blend, the weight of PET/PC copolymer formed remains constant throughout the reaction time (about 75%, Figure lla). The composition of the copolymer formed is also constant (Figurellb) and shows an excess of PET units (60% against a molar feed ratio of 50%). This result is in agreement to the predictions made (Table 1, Case4), and also with the results obtained with PBT containing hydroxyl groups [40]. 4.5. N y l o n 6 / P C blends (Case 5 )

Several authors [27-321 have studied the processing parameters, calorimetry, morphology and rheological properties of the nylon 6/PC blends, and have also indicated that some chemical reactions occur when the blends are processed at 240°C. However, scarce structural information was available on the materials produced in the thermal treatment of these blends, and on the extent of the exchange reactions occurring, since these studies did not establish the structure of the products formed and did not provide the material balance of the process. This fact generated the impression that the exchange reactions occurring in the processing are marginal, with little copolymer being formed and with the occurrence of undefined crosslinking reactions [27-321. One interesting aspect that pertains to the nylon 6/PC system is that

+

Scheme 4. Exchange reaction which occurs in the PC/nylon 6 blend

184

G . Montaudo, C. Puglisi, F. Samperi

CH3

+

Scheme 5. High-temperature hydrolysis of nylon 6 and successive reactions of amino end-groups with the inner carbonate groups of PC the two polymers are only partially miscible. However, it has been reported [29-32) that the formation of a copolymer in the processing of the two homopolymers acts as a compatibiliser, thus improving t he mechanical properties of the blends. When a 50/50 (by wt) nylon 6/PC blend was heated for about 60min at 25OoC, a single glass transition temperature (Tg) was found, indicating that a completely homogeneous product was obtained ~71. We have followed the kinetics of the melt-mixing process occurring at 240°C under nitrogen flow, performing the selective extraction and the structural analysis of the materials produced [36]. The reaction of equimolar nylon 6/PC blends could be considered within the overall scheme as an exchange reaction occurring by the attack of reactive amino terminals on the inner carbonate groups, and all the predictions consequent to the reactions summarised in Scheme 4 were verified experimentally [36]. In this example (Tablel, Case5), the same situation as in Case4 is reproduced initially; the reactive chain ends of type A are consumed, producing a nylon 6-PC copolymer and low M W PC chains bearing unreactive phenol end-groups. However, in the nylon 6/PC system, the reaction

185

Exchange Mechanisms in Reactive Polymer Blending

cl L

0

n

‘

8 lo-

w ~

o

~

~

,

~

l

~

l

~

3

l

~

Figure 12. Mol % of nylon 6 in the copolymer, produced in the melt-mixing at 240°C of the equimolar (in repeat units) PC/nylon 6 blends, ws. reaction time (Case5). Nylon 6 M W : ( 0 ) 50000;(0)8000 does not stop because the chain ends of type A (-NH2) are regenerated continuously by a thermal (hydrolytic) cleavage of homopolymer A (nylon 6 , Scheme 5). The carboxyl end-groups of nylon 6 , also generated in the hydrolysis (Scheme 5), do not react with the carbonate groups at this temperature (240°C) [36]. The copolymer AB initially formed contains an excess of A chains, but this excess decreases in the later stages of the reaction (Figurel2). The copolymer composition is therefore dependent on the extent of exchange, and the reaction moves toward completion. Therefore, the exchange reaction yields sizable amounts of copolymer, and under nitrogen flow no crosslinked copolymer is formed. The NMR analysis yielded, in addition t o the copolymer composition, evidence for the presence of urethane units interconnecting the nylon 6 and PC blocks [25]. The proportion of urethane units increases with the processing time, indicating a reduction of the block size as a function of the extent of exchange [36].Our study established the material balance of the reaction products, and provided the detailed structural information necessary to define the chemistry taking place in the melt-mixing process (Table 5). The exchange reaction in Scheme 4 concerns the attack of amino terminal groups on the inner carbonate groups, and according t o Table 1, Case 5, the exchange rate was found to be dependent on the M W of nylon 6 . The strong influence of the amino terminal groups on the reaction kinetics can be deduced from data in Table 5 and Figure 12. In the low MW nylon 6 blend, there is a rapid initial decrease (within about 20 min) in the PC

l

.

l

I

I

I

98 91 86 81 80 74 63 96 82 80 80

82 87 80 78 78

I

32' 37' 45i 47j 51j 54.d 59.d 90 84 77 67

-f

-f

0.8 1.8 2.4 4.0 -f

-f

0.5 1.5

I

0.9 3.0 3.0 2.6

Nylon 6 M w = 7900

0.8 2.0 3.0 3.5 7.0

Nylon 6 MW = 60000 - -f -f -f -

67 60 54 50 48 46 40.5 37j 4d 48i 5d

5 0 5 0 43 39 36 31 19

2 2 4 6 9

-

9 13 16 20 32

46 41 37 35

34 33 29

36

13 18 26 31

49 44 39

Material balance (rnol) in melt-mixed blend samples PCa I NyGa I PCb 1 Nysb I PCc

1 5 7 9 9 11 11

23 800 12 400 10 700 9 900 6 350 5 100 4 800

0.6 4 6

98 91 85 82 80 75 71

a4

7

4 6 7

91 78 73

37 31 24 18

8.5

PC,MWh

Table 5. Composition and material balance of the products formed in the melt-mixing at 24OoC of 1:l (by mol) PC/nylon 6 blends as a function of reaction time

I

32 0 0 37 37 8 11 36 36 15 34.d 20 26.9 33 63 55 52 49

16 500 10 350 7 600 2 500

ComDosition of the extracted copolymer (mol %) Processing Extracted fractions Copolymer Copolymer time (wt %) Nylon6 Urethane compositiong (wt %) (min) Cqe w1a W2b w3c w1+ w2 Cld c2e c3a 10 20 30 40 50 60 75 2@ 27j 24 29 9 16 24 27

10 20 30 50

+

. +

a Trifluoroethanol (TFE)-soluble fraction; insoluble fraction; CHCl3-soluble fraction; determined by l H NMR in the trifluoroacetylated TFEsoluble fraction; determined by 'H NMR in the trifluoroacetylated insoluble fraction; not measurable; g calculated as: (W1 (71 Wz.C ~ ) / ( W I " 2 ) ; determined from the end-groups obtained by 'H NMR analysis; mixture of unreacted nylon 6 and of nylon 6/PC copolymer contained in the T F E soluble fraction; j nylon 6/PC copolymer only. Unconverted nylon 6 is absent, as indicated by copolymer composition and material balance data

187

Exchange Mechanisms in Reactive Polymer Blending

0.06-

\n 4

0.04-

. 0.020.00

‘

0 I

I

20

1

40

Reaction time (min)

I

60

Figure 13. Degree of polymerisation (DP) of low MW PC chains formed in the melt-mixing of the high MW PC/nylon 6 blend at 240°C us. reaction time

present (Table5), due to the attack of the abundant amino terminals on the PC chains. In the high MW nylon 6 blend, instead, the amount of PC starts to decrease at a later stage (Figurel2), since, due to the low concentration of amino terminal groups in the high MW nylon 6, the system has to wait for the formation of the appropriate concentration of amino groups through the high temperature hydrolysis of nylon 6 (Scheme 5). Data in Figure 12 clearly indicate that within the first 20 minutes of reaction, the nylon 6/PC blend behaves like the PBT/PC and PET/PC systems discussed above (Case 4), where the exchange reaction proceeds until the reactive chain ends are consumed and then stops. However, at longer reaction times, when the hydrolysis of the nylon 6 amide groups produces enough amino terminals, both low and high MW nylon 6 blends show the same PC incorporation levels (Figure 12), indicating that the production of amino terminals by hydrolysis is the rate-determining step in the overall

Figure 14. Weight percent of copolymer produced in the melt-mixing at 240°C of the high M W PC/nylon 6 blend as a function of reaction time

188

G . Montaudo, C. Puglisi, F. Samperi

Figure 15. SEM micrographs of a PC/nylon 6 blend (25/75 by wt) melt-mixed at 240°C for 5 min: (a)uncompatibilised; (b) added with 2% of AB PC/nylon 6 block copolymer

kinetics. A remarkable kinetic effect on the composition of the copolymers produced in the two cases is also observable. The incorporation of PC into nylon 6 to form the nylon 6-PC copolymer, shows noticeably different amounts of PC incorporated as a function of time (Figure 12). As expected, the rate is much higher for the low MW nylon 6 blend. Another consequence of the attack of amino terminals on PC (Scheme 4) is that the MW of PC in the blend decreases with the processing time (Figure 13), as imposed by the random scission law (Eq. (4)):

1/DP = 1/DPi + K t

(4)

Finally, data in Figure 14 indicate that the total amount of copolymer formed in the melt-mixing process is quite sizable even at short reaction

Exchange Mechanisms in Reactive Polymer Blending

189

times. Starting from 50/50 molar blends, after melt processing for 75 min, the blend contained 30 mol of PC and 70 mol of nylon 6-PC copolymer [361. As a corollary to this study, the compatibilisation of the nylon 6/PC system was attempted [47]. Nylon 6 and PC are immiscible and form biphasic blends. When the homopolymers are immiscible, their block or graft copolymers can act as compatibilising agents, since they migrate at the blend interface, thus decreasing the surface energy and inducing compatibilisation in the system. We synthesised several ABA and AB nylon 6PC block copolymers, and tested their compatibilising effect on the nylon 6/PC blends. The morphology of the nylon 6/PC blend added with 2% by wt of block copolymers was analysed by scanning electron microscopy, and it showed the presence of only one phase, in contrast to the same blend without addition, which is clearly biphasic (Figure 15) [47]. 5. Conclusions

The exchange reactions occurring in the melt-mixing of polymer blends constituted of polymers bearing reactive functional groups may involve different chemical pathways, i. e., inner-inner and outer-inner mechanisms, as summarised in Table 1. Inner-inner reactions are often catalysed by transition metal complexes, enhancing the formation of the four-centre transition state determining the exchange process. Outer-inner reactions involve the nucleophylic attack of electron-rich end-groups on inner functional groups, and provide a variety of behaviours, as discussed above. Outer-outer exchange reactions usually involve copolymerisation processes and therefore have not been discussed here. The presence of amino, hydroxyl or carboxyl end-groups in the meltmixing process may cause “inner-outer” reactions which influence the yield and composition of the copolymers obtained. Therefore, the content of endgroups should be controlled in order to avoid composition fluctuations in these processes. Inner-inner and inner-outer mechanisms operating in the exchange reactions occurring in melt-mixing processes of a variety of systems have been investigated, making use of appropriate polymers, end-capped or containing active chain end-groups. Monitoring the composition of the copolymer formed in each case proved to be a diagnostic tool for establishing the mechanism of reaction. The diagnostic approach used here makes it possible to control the composition and yield of the copolymer to be produced, and it is generally applicable to other systems where exchange reactions occur.

G. Montaudo, C . Puglisi, F. Samperi

190

Appendix Monte Carlo modelling of exchange reactions The numerical procedure used [39] to simulate ester interchange reactions consists in generating an “artificial” sample, containing lo4-lo5 molecules and in using a random-number generator to pick up at random from the artificial sample the macromolecules which will undergo exchange. In this way, the interchange process is reduced to a series of exchange events. The variable controlling the process is the number of exchange events per initial molecule (S) [1,48-521. In order to simulate the initial stages of the reaction, S is set to a low value (e.g., S = 0.2 or S = 0.5), whereas the final stages of the reaction correspond to higher S values (e.g., S = 3 or S = 4). As the number of exchange events per initial molecule increases, the abundance of macromolecular species in the artificial sample varies. This variation reproduces only approximately the corresponding variation in the actual sample, due to statistical fluctuations. The underlying assumption of the Monte Carlo approach is that the size of the fluctuations decreases and that the approximation becomes better and better as the size of the artificial sample increases. The MOSES computer program is structured in three sections: the vector-filling, the vector-updating and the vector-screening section. The vector-filling section has three subsections. The first subsection accepts the input data, namely: the molar fraction of A in the sample (sA), the number of exchanges per initial molecule (S),the total number of units in the sample ( L ) ,Yo(A) and Yo(B) (the initial number-average degrees of polymerisation), the polydispersity index (D) and the exchange mechanism option (INNER or TERMINAL). The second subsection fills the contenents of the vk vector in order to simulate the molecular size distribution at time t = 0. The third subsection fills the contenents of the wk vector in order to simulate the initial sequence distribution. The wk vector is filled with two strings of the type AAAAAA and BBBBBB. The length of the two strings is equal to SA x L and SB x L. The vector-updating section performs different tasks, depending on the type of exchange reaction to be simulated. When the exchange mechanism option is INNER and the exchange process involves two molecules (see Figure 2B), the program chooses randomly two oligomers from the artificial sample and selects the points &I and Q2 at which the exchange takes place (Q1belongs to the first oligomer and Qz to the second one). In a first step, the program computes the distances, 211, 212, 221 and 222 of Q1 and Q 2 from the chain ends (see Figure2B). In a second step, the program generates two new oligomers in which the monomeric units alternate in a different manner along the chain with respect to the old oligomers. The new oligomers differ from the old ones in length also: the old lengths are z l l 221 and 212 222 whereas the new lengths are 211 212 and 221+ 222. In a third step, the program replaces

+

+

+

Exchange Mechanisms in Reactive Polymer Blending

191

the old oligomers with the new ones. When the exchange involves a single molecule (see Figure 2A), the program chooses randomly an oligomer and selects the points Q1 at which the exchange takes place. Then it selects another point, Q2, belonging to the same oligomer and it forms a cycle consisting of ncycmonomeric units, where ncyc = (Q2 - QI). The program chooses Q2 so that the probability of formation of an oligomer having ncycunits, G(ncyc), follows the simple power law predicted by the Jacobson-Stockmayer theory [53-551: G(ncyc)= 4.242641(n,,,)-5/2

(5)

When the exchange mechanism option is TERMINAL, the program simulates exchange reactions involving end-group attack. The computer code differs from the previously described one in the way the point Q1 (at which exchange takes place) is selected. This point is no longer chosen randomly along the oligomer’s backbone. Instead, it is set at the oligomer’s head. When the exchange involves a single molecule, the dimensions (n,,,) of the cyclic molecule formed depend exclusively on the point Q2. In this case, the process is kinetically controlled [53-551 and G(n,,,) (the probability of formation of an oligomer having ncycunits) is given by:

G (~z,~,)= 2.828423(n,,,) -3/2 The vector-screening section uses the contenents of the vectors v k and wk to compute the number of times, NA,B,,an oligomer identified by the formula A,B, appears in the simulated sample. The knowledge of NA,B, makes it possible to compute the molar fraction of A in the copolymer (cA), the molar fraction of homo-oligomers (eHOM), the weight fraction of homopolymer A (WHA), the weight fraction of homopolymer B (WHB) and the weight fraction of co-oligomers ( W C O ) . The composition of the copolymer formed during melt-mixing displays a strong dependence on the exchange mechanisms. In fact, when the driving mechanism is the attack of active chain ends on ester groups (Figure 2C,D), the copolymer composition (CA) deviates from the composition of the sample (sA), whereas when the exchange proceeds by direct exchange between inner ester groups (Figure 2A,B), no deviation occurs. Thus, the curve which reports the evolution of the copolymer composition at different reaction times can be used to discriminate between the two exchange mechanisms.

References 1. A. M. Kotliar, J. Macrom. Sci. Rev. 16,367 (1981) 2. M. Lambla, in: Comprehensive Polymer Science, edited by S. Aggarwal, S. Russo, Pergamon, Oxford 1992,First. Suppl., p. 619

192

G. Montaudo, C. Puglisi, F. Samperi

3. H. G. Ramjit, R. D. Sedgwick, J. Macromol. Sci.-Chem. A10, 815 (1976) 4. J. Devaux, P. Godard, J. P. Mercier, Polym Eng. Sci. 22,229 (1982) 5. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1875 (1982) 6. J. Devaux, P. Godard, J. P. Mercier, R. Touillaux, J. M. Dereppe, J. Polym. Sci., Polym. Phys. Ed. 20,1881 (1982) 7. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1895 (1982) 8. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 9. R. S. Porter, L. H. Wang, Polymer 33, 2019 (1992) 10. N. Wings, G. Trafara, Angezu. Makmmol. Chem. 217,91 (1994) 11. A. N. Wilkinson, D. Cole, S. B. Tattum, Polym. Bull. 35, 751 (1995) 12. A. N. Wilkinson, S. B. Tattum, A. J. Ryan, Polymer 38, 1923 (1997) 13. I. Hopfe, G. Pompe, K. J. Eichhorn, Polymer 38, 2321 (1997) 14. S. R. Murff , J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 29,3231 (1984) 15. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux , J. Polym. Sci., Polym. Chem. Ed. 24,3301 (1986) 16. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24,3315 (1986) 17. C. Berti, V. Bonora, F. Pilati, Makromol. Chem. 193, 1665 (1992) 18. M. Fiorini, C. Berti, V. N. Ignatov, M. Toselli, F. Pilati, J. Appl. Polym. Sci. 55, 1157 (1995) 19. S. J. Kollodge, R. S . Porter, Macromolecules 28, 4089, 4097, 4106 (1995) 20. M. Shuster, M. Narkis, A. Siegman, J. Appl. Polym. Sci. 52,1383 (1995) 21. R. Zhang, X. Luo, D. Ma, J . Appl. Polym. Sci. 55, 455 (1995) 22. J. I. Eguiazabal, M. J. Fernandez-Berridi, J. J. Iruin, I. Maiza, J. Appl. Polym. Sci. 59, 329 (1996) 23. D. W. Ihm, S. Y . Park, C. G. Chang, Y . S. Kim, H. K. Lee, J. Polym. Sci., Polym. Chem. Ed. 34,2481 (1996) 24. A. C. M. van Bennekom, D. T. Pluimers, J. Bussink, R. G. Gaymans, Polymer, 38, 3017 (1997) 25. K. H. Wei, H. J. Chong, Macromolecules 30, 1587 (1997) 26. B. Jacques, J. Devaux, R. Legras, E. Nield, J. Polym. Sci.,Polym. Chem. Ed. 34, 1189 (1996) 27. J. I. Eguiazabal, J. Nazabal, Macromol. Chem., Macromol. Symp. 20121, 255 (1988) 28. M.Cortazar, J. I. Eguiazabal, J. J. Irvin, Brit. Polym. J. 21,395 (1989) 29. E. Gattiglia, F. P. La Mantia, A. Turturro, A. Valenza, Polym. Bull. 21,47 (1989) 30. E. Gattiglia, A. Turturro, E. Pedemonte, J. Appl. Polym. Sci. 38,1807 (1989) 31. E. Gattiglia, A. Turturro, E. Pedemonte, G. Dondero, J. Appl. Polym. Sci. 41,1411 (1990) 32. E. Gattiglia, F. P. La Mantia, A. Turturro, A . Valenza, J. Appl. Polym. Sci. 46,1887 (1992) 33. B. Plage, H.-R. Schulten, J. Anal. Appl. Pyrolysis 15,197 (1989) 34. G. Montaudo, C. Puglisi, F. Samperi, Polym. Degrad. Stab. 31,291 (1991) 35. G. Montaudo, C. Puglisi, F. Samperi, J. Polym. Sci., Polym. Chem. Ed. 31, 13 (1993)

Exchange Mechanisms in Reactive Polymer Blending

193

36. G. Montaudo, C. Puglisi, F. Samperi, J. Polym. Sci., Polym. Chem. Ed. 32, 15 (1994) 37. G. Montaudo, M. S. Montaudo, E. Scamporrino, D. Vitalini, Macromolecules 25,5099 (1992) 38. G. Montaudo, M. S. Montaudo, E. Scamporrino, D. Vitalini, Makromol. Chem. 194,993 (1993) 39. M. S . Montaudo, Macromolecules 26,2451 (1993) 40. G. Montaudo, C. Puglisi, F. Samperi, Macromolecules 31,650 (1998) 62,2255 (1940) 41. P. J. Flory, J. Am. Chem. SOC. 42. M. S. Montaudo, A. Ballistreri, G. Montaudo, Macromolecules 25, 5051 (1991) 43. M. S. Montaudo, G. Montaudo, Macromolecules 25,4264 (1992) 44. J. C. Randall, “Polymer Sequence Determination”, Academic Press, New York 1977 45. H. K. Fkennsdorff, Macromolecules 4, 369 (1971) 46. G. Montaudo, E. Scamporrino, D. Vitalini, in: Applied Polymer Analysis and Characterization, edited by J. Mitchell Jr., Hanser, Munich 1992, vol. 11, p. 79 47. G. Montaudo, C. Puglisi, F. Samperi, J. Polym. Sci., Polym. Chem. Ed. 34, 1283 (1996) 48. M. Guaita, 0. Chiantore, Polym. Degrad. Stab. 10,212 (1984) 49. M. Guaita, 0. Chiantore, Polym. Degrad. Stab. 11, 167 (1985) 50. M. Guaita, 0. Chiantore, L. Costa, Polym. Degmd. Stab. 12,226 (1986) 51. M. Guaita, 0. Chiantore, L. Costa, Macromolecules 24,2196 (1991) 52. J. Malac, J. Polym. Sci. C33,223 (1971) 53. H. Jacobson, W. H. Stockmayer, J . Chem. Phys. 18, 1600 (1950) 54. J. A. Semleyen, Adv. Polym. Sci. 21,41 (1976) 55. Cyclic Polymers, edited by J. A. Semleyen, Elsevier Applied Science Publishers, London 1986

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 5

Interchain Transesterification Reactions in Copolyesters

J. Economy, L. A. Schneggenburger, D. F’rich

1. Introduction and background

The role of interchain transesterification reactions (ITR) in copolyesters can hardly be underestimated. It has been known for many years that poly(ethy1ene terephthalate) (PET) is most effectively produced by preparing an oligomer of PET capped at either end with glycol units. These glycol units are then distilled off during polymerisation to form a higher molecular weight polymer via ITR. Another example is provided by the reaction of PET with acetoxybenzoic acid (HBA) to produce a 60/40 PHBA/PET copolyester with liquidcrystalline (LC) character. In this case, one would expect that the HBA units will react with PET to produce copolymer chains with the same composition as the molar ratio. Unfortunately, this reaction does not proceed exactly as predicted, yielding a rather non-uniform 60/40 copolymer with significant variations in composition and in sequence distribution. In the reaction of PHBA with biphenol terephthalate (BPT) to produce the Xydar (PHBA/BPT) family of polymers, a different problem is observed, namely the formation of blocky units during solid state polymerisation which greatly complicate melt processing of this type of polymer. In Section 2 of this chapter, the syntheses of the well-known LC copolyesters are described and explanations are provided for some of the complications that may arise during polymerisation. In Section 3, the potential to solve some of these problems is then considered by the use of

J. Economy, L. A. Schneggenburger,D. Rich

196

high-temperature ITR reactions. Discussionsof several key processes, which effectively illustrate rapid randomisation of copolyesters at elevated temperatures, are included. Section 4 deals with what has been the subject of considerable debate, namely, the mechanism by which the melting point of the copolyesters increases by 50-100°C upon annealing just below their crystal-nematic transition temperature, Ten. For the first time, conclusive evidence is provided based on 13C NMR, demonstrating that this process is due to sequence ordering of the random copolyesters. In Section 5, some of the practical consequences of ITR are considered, such as the development of a new family of thermosetting copolyesters and the formation of adhesive bonds through ITR using LC polyesters or thermosetting polyesters. 2. Synthesis and microstructure

The nature of the microstructure in aromatic copolyesters has been a subject of considerable debate over the past 15 years. The three best LCP-Vectra copolyesters (Hoechst Celanese)

HBA

'IJ

HNA

Xydar terpolymers (Amoco)

t

X-7G(Kodak, Tennesse Eastman) 0 -CHz-CH2-

0-C

PET

0

HBA

Figure 1. Chemical structures of Vectra (PHBA/HNA by Hoechst Celanese), Xydar (PHBA/BPT by Amoco), and PHBA/PET (by Tennessee Eastman)

197

Interchain Transesterification Reactions in Copolyesters

known aromatic copolyesters, PHBA/2,6-hydroxynaphthoicacid (HNA), PHBA/BPT, and PHBA/PET, indicated in Figure 1, have been reported to be random or blocky [l-71. The problems associated with these reports arise from the difficulties in characterising these relatively intractable copolymers. These polymers are insoluble or soluble only in small concentrations

42/58

4

I

Insoluble

I

62/38 I

Chemical shift (ppm) Figure 2. 13C NMR spectra of PHBA/PET fractions with differing solubilities in the 60/40 PHBA/PET copolyester. Numbers indicate the molar ratio of PHBA/PET calculated from NMR

198

J. Economy, L. A. Schneggenburger,D. Rich

in very aggressive solvents. In the initial report on the PHBA/PET system in 1976, the microstructure was described M random [8]. The possibility of compositional variations was not raised. In the late 1980s, a detailed NMR study was carried out showing that the 60/40 PHBA/PET actually consists of two distinct compositions: 44/56 and 62/38 PHBA/PET (Figure 2) [6]. Furthermore, the soluble PET-rich fraction ( w 20%) was shown to be blocky while the insoluble HBA-rich fraction (w 80%) was more random (Figure 3). To understand the microstructure that is present in these systems, it is important to examine the role of the synthetic route to these copolyesters. The PHBA/PET copolyester system presents an unusual degree of complexity. During the synthesis of PHBA/PET, acetoxybenzoic acid should react in the melt with PET via ester-interchange reactions to produce a random copolyester with a 60/40 ratio. What actually occurs is that during further reaction the ratio of PHBA in the copolymer increases above a critical range, resulting in phase separation into a nematic and an isotropic melt, each with distinctly different compositions. Since the two phases are 0

PHBA-PHBA - - - 157.5 (ppm) PHBA-PET - - - 156.8 (ppm)

4-060

I

50

-

40

-

30

-

20 30

I

40

I

I

50

PHBA content (mol %)

I

I

60

I

70

Figure 3. PHBA dyad sequences for PHBA/PET copolyesters: (0) experimental values; ( 0 ) calculated values for totally random copolyesters

199

Interchain Transesterification Reactions in Copolyesters

il I

x

= 0.73

x

= 0.48

x

= 0.47

NB d

b 8

h

... . . . ...,... . . . ... ... . . ... ,.. ... , . . 170

166

162

PPM

158

,

154

. . i

150

Figure 4. 13C NMR spectra in the carboxyl region of the copolyesters HBA/HNA in PFP at 80°C (z is the fraction of HBA units in the copolyester): (a)synthesised from 13C-labelled HBA; (b) the commercially available 73/27 HBA/HNA Vectra, Hoescht Celanese; (c) synthesised from 48/52 HBA/HNA mixture; (d) synthesised from 13C-labelled HNA

200

J. Economy, L. A. Schneggenburger,D. Frich

relatively immiscible, there is little chance of the occurrence of ITR between them, leading to a more uniform composition. Attempts to homogenise the phase separated compositions at 280-300°C under shear did not result in any measurable changes in the composition. Heating at higher temperatures was unsuccessful due to the onset of degradation reactions. The PHBA/HNA system is more suitable for study due to its compositional homogeneity and solubility in pentafluorophenol (PFP), allowing the use of NMR techniques to characterise dyad sequences. This copolyester is prepared by melt polymerisation of the two monomers pacetoxybenzoic acid and 2,6acetoxynaphthoic acid and is far more stable at elevated temperatures than the PHBA/PET. In Figure 4b and 4c, the I3C NMR spectrum of the carboxyl carbon region of the PHBA/HNA copolyesters of the 73/27 and 48/52 systems is shown [9].Also shown in Figure 4a and 4d are the spectra of 13C enriched HBA and HNA containing copolymers, allowing unique identification of the dyad sequences. With this technique, it was possible to determine the reactivity ratios of the two monomers by analysing the 50/50 copolymer after polymerisation to a molecular weight of 2000 [lo]. Examination of the copolymer by 13C NMR showed the same ratio of monomers as in the starting mixture. Furthermore, analysis of the dyad sequences indicated a distribution of the four possible dyads identical to what one might predict for a random copolymer. From these results one may conclude that the reactivities of the two monomers are about the same. However, one could argue that interchain transesterification reactions could occur in PHBA/HNA more rapidly than polymerisation, resulting in the random structure. A unique experiment was carried out, which distinguished between polymerisation and interchain transesterification reactions at 245°C [lo]. As shown in Figure 5, pacetoxybenzoic acid (ABA) with a 13C labelled carbonyl carbon (B*) was reacted with an ABA/HNA dimer (BN). In the absence of ITR, one should observe only B*B and B*B* dyads, but B*N dyads are observed at a concentration of 14%, indicating that a small amount of ITR occurs during polymerisation. Blackwell et al. have examined the microstructure of PHBA/HNA using X-ray diffraction techniques [ll].From their study, which depends on modelling of the diffraction pattern, they concluded that the sequence distribution is random. Windle has suggested that this random sequence could be described as a nonperiodic layer (NPL) crystal [12-141. In this model, the same random sequence in neighbouring chains forms ordered, matching sequences by axial translations between adjacent chains. Alternatively, Blackwell and Biswas have proposed a model suggesting the existence of random sequences as one-dimensional paracrystalline lattices (PCLs) along the chain axis [15,16]. The main difference between these two models is that the PCL model requires crystallites to form as close packing of random chains, without identical matching sequences between adjacent chains. Statistical interpretations have shown that the probability of finding matching monomer units decreases rapidly as one considers an increasing number

>

Interchain Transesterification Reactions in Copolyesters B'BN B'B' I

B'

+ BN

201

245OC

+ (B-B'N)

(B'-BN)

170 min

-'COOH

I

I

170

I

I

166

I

I

162

I

PPM

I

158

I

I

154

I

I

150

I

Figure 5. Evidence of ITR from 13C NMR of oligomers (n = 5) in PFP using 99% 13C-enrichedcarbonyl in the HBA unit

of chains; therefore, Blackwell's interpretation is favoured at present [17]. In the Xydar system (PHBA/BPT), there has also been some confusion as to the interpretation of the microstructure. This copolymer was initially prepared by a combination of melt and solid-state polymerisation reactions. Thus, a mixture of three monomers, namely, ABA, terephthalic acid (TA), and acetylated biphenol (ABP), was heated to 300°C in Therminol-66 solvent. Since TA is insoluble in the melt until 300"C, low molecular weight chains of ABA and ABP form. Once the reaction temperature reaches 275"C, TA begins to react, but the melting point of the mixture rises too rapidly and hampers further stirring. The material is removed, ground, and polymerisation is then continued in the solid state to yield a polymer with a melting point of 408°C. An ABA/ABP/TA oligomer which is reacted at 290°C is still soluble and I3C NMR results indicate that this oligomer has a blocky structure [18]. Thus, it appears that the as-prepared copolymer after polymerisation in the solid state is indeed blocky, but Blackwell has reported that melt-spun fibres show a random sequence, based on simulations of the X-ray diffraction pattern [19]. In subsequent discussions, the nature of this discrepancy will be clarified.

3. Randomisation processes There has been considerable debate over the nature of the sequence distribution of these polymers, as well as the possible changes in the microstruc-

202

J. Economy, L. A. Schneggenburger, D. Rich

ture on heating them in the nematic melt. Some authors have argued that these systems are stable in the melt [l],while others claim that they undergo ordering [4]; finally, we have suggested that the dominant process is randomisation through ITR [7]. Kugler et al. studied interchain transesterification reactions between deuterated and non-deuterated PET chains [20]. The rate of transesterification was determined with small-angle neutron scattering techniques (SANS). They observed that decreases in the block size (molecular weight of the deuterated and protonated segments) occurred with no change in the viscosity of the polymer. Thus, ITR was deduced to be responsible for the randomising effect on the molecular structure. Assuming independence of molecular weight, they calculated the rate constant for ITR in PET at 280°C to be 0.215s per molecule. This corresponds to approximately 10 transesterification reactions occurring per molecule per minute. Similar observations were made by Arrighi et al. in SANS studies of 4,4'-dihydroxya,a'-dimethylbenzalazine copolyesters [21].

1/1 mixture

1/1 HBA/HNA Random copolyester 100

I

200

I

I

300

Temperature, O

400 C

Figure 6. DSC of the PHBA and PHNA mixture, compression moulded mixture, and a 1/1 PHBA/HNA random copolyester

Interchain TransesterificatiooReactions in Copolyesters

203

The potential for very rapid randomising processes in the copolyesters has been demonstrated by heating a 50150 mixture of the two homopolymers of PHBA and PHNA at 450°C at a pressure of only several hundred pounds per square inch (psi) [7]. A relatively low-viscosity melt was

450' C compression moulded

6 +

I

I

I

b

'

I

I

I

I

170

PHBA/PHNA

I

I

I

I

166

I

I

%

1

162

I

I

I

I

158

I

I

I

I

154

I

I

I

I

I

1

150

PPm

Figure 7. Comparison of 13C NMR spectra of a compression moulded mixture of PHBA and PHNA (a)with that of random PHBA/HNA copolymer (b)

204

--J. Economy, L. A. Schneggenburger, D. Rich

formed which extruded from the cracks in the mould within a few seconds. DSC measurements of this new material indicated a T,, of 26OOC which is consistent with a 50/50 random copolymer (Figure 6). Benzoicnaphthoic and naphthoic-benzoic dyad sequences could be detected easily from the 13C NMR spectrum of the compression moulded polymer shown in Figure 7a. Comparison of this 13C NMR spectrum and that of the random PHBA/HNA copolymer (Figure 7a,b) indicates that the structures are identical. Evidently, ITR reactions at this elevated temperature occur at rates in excess of 100 ester interchange reactions per chain per second [71. Randomisation in the nematic melt also occurs in the Xydar copolyesters. In Figure 8, DSC analysis indicates that heating commercial PHBA/BPT (2/1) to 440°C results in a material with a much lower T,, of 370°C. Presumably, the blocky polymer prepared by solid-state polymerisation, with a T,, of 409OC, on heating to 44OOC undergoes rapid randomisation, yielding a more uniform random copolymer. This new material may very well provide a system which is much more easily processed than the currently available product. The occurrence of ITR in polyesters would appear to have a number of useful implications. These include the ability to alter polymer average molecular weight and molecular weight distribution, produce both blocky and random copolymer structures by blending high molecular weight polymers [22,23],and effectively perform localised post-processing steps in exN

A

-84 4 4

4

1 300

(4

409OC

-

-

(b)

350

400

Temperature, OC

450

5 I0

Figure 8. DSC of 2/1 PHBA/BPT Xydar LCP: (a) commercial; (b) randomised at 44OoC for 15 min and annealed at 270°C for 5 h

Interchain TransesterificationReactions in Copolyesters

205

isting polymer structures [24,25]. Another intriguing implication resides in the use of ITR to recycle thermosetting polyester resins [26]. The process depends on acetolysis of the crosslinked polyester network, leading to multifunctional oligomers. These oligomers can react again upon heating to give the identical crosslinked copolyester system.

4. Sequence ordering Another unusual process that occurs upon annealing the copolyesters near their T,, is the dramatic increase in transition temperature by 40-75°C [27-291. As expected, annealing well below the melting point leads to a significant increase in transition enthalpy with little change in the melting temperature. Diffusion processes dominate under these conditions and it is well accepted that ordering occurs by a physical process [5,27-301. 13C NMR analyses of 50/50 HBA/HNA [7] and 73/27 HBA/HNA [31] copolyesters annealed well below Tcndo not indicate any change in dyad sequence distribution. On the other hand, annealing the copolyesters near Tcnnot only increases their melting temperatures, but is accompanied by a change in crystal structure (hexagonal to orthorhombic) [28] and an increase in density [32]. Annealing HBA/HNA copolyesters near T,, dramatically increases the transition temperature throughout the entire compositional range, as illustrated in Figure 9 [33]. Yoon et al. noted that the melting transition of the annealed copolymer is best described as a smectic B (molecules arranged in a hexagonally close-packed array within the layers, ie., there is a long-range hexagonal order in the layers) [34]. In addition there is an abrupt decrease in solubility [33]. These results are indicative of sequence ordering; however, this interpretation cannot be considered conclusive. The mechanism by which this process occurs has led to several quite different interpretations. Winter proposes that ordering is due to a physical process which involves the melting of crystallites when annealing near T,, [27]. Melting is followed by nucleation and crystallisation of longer ordered sequences already present within the random copolyester. Fkom a simple statistical analysis, we have estimated that the probability of pre-existing long sequences of ordered units within a two-dimensional array is less than 1% 133,351. Hence, this proposal can be ruled out since the long ordered sequences necessary to explain the relatively large transition enthalpies observed for nucleation and growth of crystals are not available. Geil suggests that an increase in crystal size and perfection via a physical process is occurring and that Blackwell’s PCL model [16,17] is the “appropriate description of the crystalline state of these copolymers” [36]. Contrary experimental evidence includes results showing that there is no difference between the DSC traces of samples slowly cooled (l”C/min) from

206

J. Economy, L. A. Schneggenburger,D. Rich

200 4 0

20

40

60

Composition (% HBA)

80

1

Figure 9. Transition temperature as a function of composition of HBA/HNA copolyesters the annealing temperature and those of quenched ones [33].One would expect that crystalline domains with increased size and perfection formed from a physical process would be sensitive to heating conditions, so that slow cooling would result in a larger endotherm. Yoon has argued that the increase in melting point during annealing near T,, is a result of molecular weight increase because of the long annealing times [37].However, a recent report by Windle shows that when the molecular weight of a 75/25 HBA/HNA copolyester increases from 8600 g/mol to 30000 g/mol, the increase in melting point is only 12°C (from 283°C to 295OC) [38].Hence, it is unrealistic to attribute changes of 40°C or more to an increase in molecular weight. We propose an ordering mechanism that is chemical in nature and depends on ITR reactions within the existing crystallites present in the copolyester. Thus, near T,, the end-groups have sufficient mobility within the crystallites to facilitate ordering of the microstructure through ITR. These reactions can only occur in the crystallites, since in the coexisting nematic melt no templates are present to promote ordering [7].The driving force for chemical ordering reactions arises from the improved packing and correspondingly higher density associated with a more ordered structure. ITR allow an improved ordering of sequences, but not necessarily complete ordering, since mobility within the crystallites should drop significantly as the transition temperature begins to increase. Obtaining definitive evidence on the nature of the ordering processes has been, until recently, an extremely difficult problem. The use of X-

207

Interchain Transesterification Reactions in Copolyesters

ray diffraction techniques has not proven effective, especially with the HBA/HNA system, due to the relatively low degree of crystallinity in this family of copolyesters. From the transition enthalpies ( A H )of the annealed random copolyesters examined, the degree of crystallinity is estimated to range from 10 to 20%. Use of 13C NMR techniques with these copolyesters is not easy because of their low solubility. However, by preparing a low molecular weight 50/50 HBA/HNA copolyester, MW in the region of 5000 g/mol, it was found that the solubility increased dramatically [35]. We prepared the low molecular weight copolyester with HBA units containing a 13C carbonyl carbon. Using 13C NMR, one could then follow the sequence ordering as the concentration of dyad sequences changes during annealing [39]. DSC results after each annealing cycle are shown in Table 1. 13C NMR of samples annealed at 190°C for 24 and 48 h (to build crystallinity) Table 1. DSC results of annealed samples Sample

(4 (b)

(4 (d) (e)

a

As-prepared 190°C/24 h (a)+23OoC/24 h 190"C/48 h (c)+23OoC/48 h (c)+233"C/48 h

Tcn("C)

AH (J/g)

Crystallinitya (%)

241 242 264 244 268 276

2.3 7.3 4.0 5.7 3.6 2.4

5 16 9 12 8 5

Based on fully crystalline homopolymer of HBA

(AH= 46 J/g

[46])

revealed no changes in the benzoic-naphthoic (BN) and benzoic-benzoic (BB) dyad ratios. However, the BN/BB ratio does change upon annealing near TC,.Table 2 lists the changes which occur upon ordering. The Table 2. I3C NMR results Sample As-prepared 19OoC/24 h 19OoC/48 h

+ 230°C/24 h + 23OoC/48 h

fBN

fBB

Change %

0.515 0.524 0.526

0.485 0.476 0.474

1.7 2.1

-

fractions of BN ( ~ B N )and BB dyads ( ~ B B )initially are 0.515 and 0.485, respectively. Annealing near the T,, for 24h results in a 1.7% increase in BN dyads. This indicates that the sequences are changing from random to alternating during annealing. Since the differences in dyad sequences are measured in the crystalline regions only (- 10% in this sample), the 1.7% change is significant. Increasing the annealing duration from 24 to 48h further changes the ordering from random to alternating (2.1% increase in BN dyads). This proves unambiguously that interchain transesterification reactions occur, leading to sequence ordering. It should be noted that

J. Economy, L. A. Schneggenburger, D.Rich

208

the measurements are reasonably accurate since the experimental error is within the range of f0.2%. It is important to note that these kinds of ordering processes are common to many copolyesters. For example, these ordering reactions were found to occur in HBA/BPT copolyesters. Annealing a commercial 2:l HBA/BPT copolyester (Ten = 409°C) for 5 h at 400°C increased the transition to 456°C (Figure 10) [35].Subsequent heating to 460°C for 10h

250

300

350

400

450

Temperature, "C'

500

Figure 10. DSC of 2/1 PHBA/BPT Xydar LCP: (a)commercial; (b) annealed 5 h at 400°C;(c)heated to 460°C for lOmin, followed by annealing 5 h at 270°C; and (d) annealed 20h at 349°C

Interchain TransesterificationReactions in Copolyesters

209

followed by annealing at 270°C for 5 h yielded a randomised structure with a Tcnof 359°C. This structure could be re-ordered by annealing just below the new Ten. Ordering processes are also observed in a non-liquid-crystalline copolyester derived from sebacic acid, biphenol, and hydroquinone [35], ils well as in a thermoplastic elastomer based on copolyester units [40](see also Chapter 8). 5. Adhesive bonds in polyesters formed by ITR

5.1. Liquid-crystalline copolyesters

We have reported on a completely new approach for developing an adhesive bond between films of LCPs via ITR [25]. LCP coatings on an aluminium substrate, when brought together under heat and moderate pressure, develop an adhesive bond with reasonable lap shear strengths. Typically, such bonds never fail adhesively (ie., at the polymer-metal interface), but always cohesively (within the polymer). It is important to note that these adhesive bonds can be formed above or below the melting point of the LCP,

I

)d( ) 200

250

300

Temperature, OC

350

400

Figure 11. Memory effect after heating 24/76 HBA/HNA copolyester above the Ten: (a)commercial; (b)annealed 3 min at 360'C; ( c ) annealed 15 min at 360°C; (d)annealed 30 min at 360'C

210

J. Economy, L. A. Schneggenburger, D. Rich

which suggests that the primary mechanism for developing the adhesion is through ITR. The possibility of interpreting the adhesion process in terms of chain diffusion and chain entanglement is small, considering the rod-like nature of the chains and the relatively short contact time (see also Chapter 11). We have observed that briefly heating the 24/76 HBA/HNA at 360°C (35°C above its Ten) and then quenching results in complete retention of crystalline order (Figure 11). Only after heating for 15min does the crystallinity disappear after quenching (Figure 11c). This experiment suggests that diffusion processes in the melt are very slow and that there is a strong memory which persists well into the nematic state. It is not clear whether the primary mechanism acting to eliminate crystallinity is physical diffusion, chemical diffusion via ITR, or possibly both. It is safe to argue that both diffusion and ITR are operative over the 15-min time span at 35°C above the TCn(see also Chapter 11). The lap shear strength of 73/27 HBA/HNA bonded to aluminium is shown as a function of processing temperature in Figure 12. The excellent adhesion of the 73/27 HBA/HNA to aluminium may be partly due to the ability of the nematic melt to orient at the polymer-metal interface on cooling from the melt in such a way as to minimise the mismatch in coefficient of thermal expansion (CTE). Presumably, while cooling from the

2ooo

r----proceased at 280'C at 3OO0C processed at 320°C processed at 340'C

-t- proceased

0

100

200

300

Testing temperature ('C) Figure 12. Effect of the processing temperature on the lap shear strength of 73/27 HBA/HNA bonded to aluminium for specimens tested at different temperatures

211

Interchain TransesterificationReactions in Copolyesters

melt, the LCP organises at the interface to minimise the stress associated with contraction of the substrate and thus approximates to the CTE of the substrate. This process could continue well below Ten. In addition, the bonding between aluminium and the polymer would be enhanced through the dipolar interactions of the carboxylate ester with the oxide coating on the aluminium substrate. It is noteworthy that these LCP coatings provide excellent protection of metal substrates against acids, moisture, abrasion, and wear [41]. On the other hand, they have several limitations. The LCP must be coated as a high-viscosity melt at temperatures above 300°C onto the metal substrate. The strength of HBA/HNA at high temperatures is limited to 13&150"C due to the low degree of crystallinity [24]. The anisotropic nature of the LCP also appears to provide a limit to the lap shear strength of one-half to two-thirds of the values observed with epoxies. This should be expected since the transverse mechanical properties of highly anisotropic films are low because of the tendency of these rod-like chains to orient preferentially with reduced bonding in the perpendicular direction. 5.2.

ITR in thermosetting polyesters

We have recently reported the discovery of a new family of crosslinkable aromatic copolyesters [42]. The precursor oligomer is made up of a twocomponent system, where one oligomer has acetoxy end-groups and the other oligomer has carboxylic acid end-groups. The mixture of the two oligomers melts at 10O-22O0C, depending on oligomer molecular weight and composition. The melting temperature is well below the cure temperature, which is in the range 240°C to 320°C. One can design systems with relatively low viscosities by selecting the appropriate monomer feed ratio and molecular weight, thus eliminating the need for solvents. In certain cases, liquid-crystalline character can be observed in the oligomeric melts, using a polarising microscope. This feature is definitely related to the presence of longer rod-like units between branching sites. Interestingly, the optical birefringence persists during curing, suggesting the presence of local order in the crosslinked structure. In Table 3, some of the novel properties of these polymers are listed. Table 3. Properties of crosslinked aromatic copolyesters Glass transition temperature (OC) Weight loss to 400°C in Nz (%) Weight loss to 400°C in air (%) Char yield at 600°C (%) Density (g/cm3) Moisture uptake (wt%) Dielectric constant (25OC, 1 MHz)

225 0 0 40

1.35 0.3 4.7

212

J. Economy, L. A. Schneggenburger, D. Rich

-

Figure 13. DMTA scan of crosslinked aromatic copolyester resin: shear modulus 1.2GPa at 25OC, T.225OC

The outstanding feature of this new thermoset family, compared to epoxies and phenolics, is its excellent high-temperature stability. The use of these types of polymers as thin electrically insulating foamed films with very low dielectric constants is discussed elsewhere [43]. The high-temperature retention of mechanical properties for the cured resin is shown in Figure 13. The dynamic mechanical thermal analysis (DMTA) scan shows almost 80% retention of flexural storage modulus (E') to over 16OOC. The glass

Consolidation

d; ; uAl

T Shaping I

wsembIy of

Figure 14. Process for cost-effective composite fabrication

213

Interchain TransesterificationReactions in Copolyesters

transition temperature is 225”C, as indicated by the peak in the flexural curve. In all cases, the formation of adhesive bonds has loss modulus (El’) taken place in the solid state, after surfaces have been coated and cured. The use of oligomeric mixtures to coat yarns or fabrics continuously, followed by curing of the coating, provides a technique for greatly simplified solid-state processing of composites. Thus, as shown in Figure 14, a carbon fibre fabric or yarn can be impregnated with the melt to the appropriate concentration. Curing can then be carried out in a continuous process and the acetic acid by-product collected separately. Consolidation of the composite laminas is carried out by heating under pressure using ITR. It is noteworthy that one can produce good composite structures with minimal porosity by processing in the more traditional manner without first eliminating the acetic acid. Presumably, the 4% of acetic acid present as end-groups could easily escape from within the composite during moulding. Some consolidation of porosity may occur as a result of continued ITR under pressure. Considerable effort has been made to improve the lap shear strength of the adhesive bond formed between the crosslinked aromatic copolyesters and metal substrates. Typically, the lap shear strength averages about 1900 psi (Figure 15), with failure always at the polymer-metal interface [44]. A series of experiments were designed to determine whether bonding at the polymer-metal interface could be improved. These included the use of milled glass fibre reinforcement , a specially designed liquid-crystalline copolyester thermoset, different metal substrates, such as titanium, and various substrate surface pretreatments. The most dramatic increase in I

12 14

10 -

1500

8-

1000

6-

3

Ti t

1

1

4-

k a er

3

z.

h

500

W

2-

01

0

I

50

I

100

I

150

I

200

10

250

Figure 15. Lap shear strengths of the crosslinkable aromatic copolyester bonded to aluminium, as a function of temperature

214

J. Economy, L. A. Schneggenburger,D. Rich

0

I

I

I

Bondline thickness (pm) Figure 16.Lap shear strength of the crosslinkable aromatic copolyester bonded to aluminium, as a function of bondline thickness

bond strength was achieved by reducing the bondline thickness from 350pm to 100 pm (Figure 16) [45].The failure mode was still adhesive, so the likelihood exists of obtaining even higher lap shear strengths with improved surface treatment. However, since the lap shear strengths are within the range of what we would expect for high-performance adhesives, no further work is planned. 6. Mechanism of adhesive bond formation

The mechanism by which the adhesive bond forms in both LCPs and thermosetting polyesters has been investigated using secondary ion mass spectroscopy (SIMS) and neutron reflection techniques [45]. These complementary analytical techniques were used to examine the entire range of physical and chemical diffusion distances possible in the aromatic copolyesters. Long range effects could be examined by SIMS while neutron reflection would be able to detect localised interfacial diffusion. The contrast across the interface for the aromatic copolyesters was achieved by observing the interpenetration between thin films of deuterated and non-deuterated copolyesters. A schematic of the results is shown in Figure 17. From the SIMS examination of the adhesion/welding of the polymer-polymer interface between LCP films, it appears as though the bilayered sample completely homogenises within the first hour of annealing at 280°C. Neutron reflectivity studies confirm these results. The possible mechanisms for interpenetra-

-

215

Interchain TransesterificationReactions in Copolyesters

73/27 HBA/HNA Lcp

m y

Deuterated layer Non-deuterated layer

N

Cured aromatic copolyester thermoset

lOOpsi

Complete homogenisation

< 300 A interdiffusion (b)

(a)

Figure 17. Interdiffusion of chains of (a) 73/27 d-HBA/d-HNA and 73/27 HBA/HNA and (b)d-crosslinkable aromatic copolyester and crosslinkable aromatic copolyester upon annealing at 280°C

tion consist of physical diffusion of LCP molecules and chemical diffusion through the occurrence of rapid ITR. Because of the limited diffusivity of the LCP molecules at the interface, homogenisation of the films through ITR would appear to be the primary mechanism for interpenetration, but conclusions on the actual mechanism of LCP interpenetration that is operative cannot be drawn solely from the data presented (see also Chapter 11). The degree of interpenetration was measured for thin cured aromatic copolyester thermosets by the same technique as described above for the LCP bilayered films. No observable changes in SIMS depth profile were observed with annealing after 10 h at 280°C. Since the depth resolution of SIMS is 500 A, any changes in the polymer- olymer interface are predicted to occur over length scales of less than 500 . Neutron reflection data suggest that the actual amount of interpenetration is less than 300A after 11 h at 280°C. Interpretation of the SIMS and neutron reflection data for the welding of two cured aromatic copolyester thermoset films is much less ambiguous than that discussed above for LCP thin-film welding. Since the individual films were cured into infinite molecular weight networks prior to joining and annealing, physical diffusion of individual polymer chains or fragments across the interface would not be possible. Thus, the only mechanism available for adhesion across the polymer-polymer interface would be chemical interdiffusion through rapid high-temperature ITR (see also Chapter 11).

x

216

J. Economy, L. A. Schneggenburger, D. Rich

References

Liq. Cryst. Nonlin. Opt. 157, 535 (1988) 2. G. D. Butzbach, J. H. Wendorf, H. J. Zimmerman, Polymer 27, 337 (1986) 3. E. Joseph, G. Wilkes, D. Baird, Polymer 26, 689 (1985) 4. R. W. Lenz, J. Jin, K.A. Feichtinger, Polymer24, 327 (1983) 5. Y. G. Lin, H. H. Winter, Macromolecules 21, 2439 (1988) 6. L. Quach, E. Hornbogen, W. Volksen, J. Economy, J. Polym. Sci., Polym. Chem. Ed. 27, 775 (1989) 7. A. Muhlebach, J. Economy, R.D. Johnson, T. Karis, J. Lyerla, Macromolecules 23, 1803 (1990) 8. W. J. Jackson, H. F. Kuhfuss, J. Polym. Sci., Polym. Chem. Ed. 14, 2043 (1976) 9. A. Muhlebach, R.D. Johnson, J. Lyerla, J. Economy, Macromolecules 21, 3115 (1988) 10. J. Economy, R.D.Johnson, J.R. Lyerla, A. Muhlebach, in: Synthesis and Microstructure of Aromatic Copolyesters, edited by R. A. Weiss, C. K. Ober, American Chemical Society, Washington DC 1990,vol. 435,p. 129 11. J. Blackwell, A. Biswas, Macromolecules 18, 2126 (1985) 12. R. Golombok, S. Hanna, A.H. Windle, Mol. Cryst. Lip. Cryst. 155, 281 (1988) 13. S. Hanna, A.H. Windle, Polymer 29, 207 (1988) 14. A. H. Windle, C. Viney, R. Golombok, A.M. Donald, G. R. Mitchell, Faraday Discuss. Chem. SOC.79,55 (1985) 15. J. Blackwell, A. Biswas, Macromolecules 20, 2997 (1987) 16. A. Biswas, J. Blackwell, Macromolecules 21,3146 (1988) 17. A. Biswas, J. Polym Sci., Polym. Phys. Ed. 30, 1375 (1992) 18. R.D. Johnson, J. Economy, J. Lyerla, A. Muhlebach, APS, Polymer Section, St. Louis, MO 1989 19. J. Blackwell, H. M. Cheng, A. Biswas, Macromolecules 21, 39 (1988) 20. J. Kugler, J. W. Gilmer, D. W. Wiswe, H.G. Zachmann, K. Hahn, E. W. Fischer, Macromolecules 20, 1116 (1987) 21. V. Amghi, J. S. Higgins, R. A. Weiss, A. L. Ciecioglu, Macromolecules 25, 5297 (1992) 22. M. E. Stewart, A. J. Cox, D. M. Naylor, Polymer 34,4060 (1993) 23. E. R. George, R. S. Porter, Macromolecules 19, 97 (1986) 24. J. Economy, A. G. Andreopoulos, J. Adhesion 4 0 , 115 (1993) 25. J. Economy, T.Gogeva, V. Habbu, J. Adhesion 37, 215 (1992) 26. J. Economy, A.G. Andreopoulos, Polym. Adv. Technol. 7,561 (1996) 27. Y.G.Lin, H. H. Winter, Macromolecules 24, 2877 (1991) 28. A. Kaito, M. Kyotani, K. Nakayatama, Macromolecules 23, 1035 (1990) 29. S.Z. D.Cheng, Macromolecules 21, 2475 (1988) 30. G. R. Mitchell, A. H. Windle, Colloid Polym. Sci. 263, 230 (1985) 31. C. W.Potter, Master’s thesis, University of Illinois at Urbana-Champaign, 1994 32. D.J. Wilson, A.H. Windle, H.G. Zachmann, International Conference on Advanced Polymer Materials, Dresden 1993 1. M.T. DeMeuse, M. Jaffe, Mol. Cryst.

Interchain Transesterification Reactions in Copolyesters

217

33. J. M. Kachidza, Master's thesis, University of Illinois at Urbana-Champaign, 1991 34. D.Y. Yoon, N. Masciocchi, L.E. Depero, C. Viney, W. Parrish, Macromolecules 23, 1793 (1990) 35. C. W. Potter, J. C. Lim, G. Serpe, J. Economy, Progr. Pacific Polym. Sci. 3, 271 (1994) 36. J. Liu, F. Rybnikar, P. H. Geil, J. Macromol. Sci-Phys. B35, 375 (1996) 37. D.Y. Yoon, Y. Ando, 0.0.Park, T. E. Karis, D. Dawson, T. Huang, J. Am. Chem. SOC., Polym. Prepr. 37,81 (1996) 38. A. Romo-Uribe, A.H. Windle, J . Am. Chem. SOC.,Polym. Prepr. 37, 83 (1996) 39. L. A. Schneggenburger, P. Osenar, J. Economy, Macromolecules, axcepted March 1997 40. J. Economy, C. Fischer, Polym. Adv. Tech. 5, 295 (1994) 41. D. Frich, J. Economy, Proc. ACS (Div. PMSE) 69, 438 (1993) 42. J. Economy, D. Frich, K. Goranov, J. C. Lim, Proc. ACS (Div. PMSE) 70, 398 (1994) 43. L. A. Schneggenburger, J. Economy, ICEMCM Proc. 65 (1995) 44. D. Frich, J. Economy, Proc. ACS (Div. PMSE) 74, 341 (1996) 45. D. Frich, PhD thesis, University of Illinois at Urbana-Champaign, 1996 46. A. Muhlebach, J. Lyerla, J. Economy, Macromolecules 22, 3741 (1989)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 6

Inhibition of Transreactions in Condensation Polymers N. R.James, S. S. Mahajan, S. Sivaram

1. Introduction 1.1. Polymer blends

There has been tremendous commercial and scientific interest in polymer blends and alloys over the last several years [I]. The terms “blends” and “alloys” have become almost indistinguishable in industry parlance, possibly because of the convenience of semantically equating the two concepts. By strict definition there is a difference, although the basic goal of each is identical, which is to combine two or more polymers and thus make available the best properties of each polymer in a single material with an optimised cost and performance. The difference is that a blend is essentially a mechanical mixture, whereas an alloy is produced by achieving mixing of the polymers at a molecular level in a chemical reactor or extruder. 1.1.1. Compatibility i n polymer blends

Few polymers form truly miscible blends characterised by a single glass transition temperature (T’) and homogeneity on the 5-10nm scale [2]. Miscibility in these systems is attributed to the presence of specific interactions, such as acid-base or ion-dipole, hydrogen bonding and transition metal complexation [3-6]. The majority of the blends are immiscible and possess a phase separated morphology. The physical properties of such blends are limited by the large domain size, poor interfacial adhesion, and the

N. R. James, S. S. Mahajan, S. Sivaram

220

tendency to form unstable morphologies. Heterogeneous blends of technological importance are termed “compatible” [7,8]. The differences between miscible, immiscible and compatibilised blends is illustrated in Figure 1. Miscible blends are thermodynamically stable, molecular-levelmixtures. Immiscible blends are separated into macroscopic phases with minimal interfacial adhesion and unstable morphologies. Compatibilised blends are also macro-phase separated. However, the presence of interfacial agents or chemical bonds stabilises the morphology and increases interfacial adhesion. In such blends, satisfactory physical and mechanical properties are related to the presence of a finely dispersed phase and resistance to gross phase segregation. The presence of a compatibilising agent permits the blending of otherwise incompatible polymers to yield blends or alloys with unique properties, generally not attainable from the individual components. Compatibility is promoted through block and graft copolymers possessing segments with chemical structures or solubility parameters which are similar to those of the polymers being blended. The copolymer may be preformed and added :a) Domain size<0.5nm (b) Domain size>100nx

@ (u) miscible

(b) immiscible

>

:c) Immiscible blend domain size >miscible blend

@

(c) compatibilised

Figure 1. Differences between miscible, immiscible, and compatibilised blends [13]

221

Inhibition of Transreactions in Condensation Polymers

to the mixture of polymers or formed in situ by reaction between co-reactive functional groups on the polymers. The compatibilising agent acts as a polymeric surfactant, which lowers surface tension and promotes interfacial adhesion between the dispersed and matrix polymer phases. Compatibilisation of polymer blends through reactions during compounding is becoming increasingly important and is known as reactive compatibilisation [9-131.

1.2. Reactive compatibilisation Reactive compounding in a twin screw extruder usually involves highly reactive functional groups, which are unstable under processing conditions. The inherent functionality present in end-groups of the polymer pairs as well as polymer-polymer cross reactions are often exploited. Some examples of common compatibilising reactions between functional blend constituents are shown in Figure 2. Compatibility is promoted by the presence of functional groups which undergo chemical reaction. A segment of a compatibilising agent containing reactive functionality (X) compatibilises a polymer having a different structure as well as a different solubility parameter but having a co-reactive functionality (Y). A-B--X

I

+ Y--EI

+A-B-X-Y-E-

I

I

The resultant graft or block copolymer, prepared in situ from a functionalised polymer containing pendant or terminal functionality is a compatibilising agent for unfunctionalised A and E polymers.

The compatibilising agent functions as an organic surfactant and is thought to concentrate at the interface of two incompatible polymer phases and act as an emulsifiers, reducing interfacial tension. It promotes adhesion also through interpenetration and entanglements (see also Chapter 10).

1.3. Transreaction during melt-blending During the last few years, research efforts have been directed towards the development of blends from existing polymers by moving the chemistry from reaction vessels into processing equipment. This is often referred to as reactive processing. The reactions occurring during melt-blending can modify the chemical structure and improve compatibility between polymers which are otherwise immiscible [20].This approach is particularly attractive in the case of polycondensates, which usually contain reactive functional

N. R. James, S. S. Mahajan, S. Sivaram

222

09 0

Imidization [12~4,151

R-NHz

t

0

+ R - D

0

~

0

II

R-OH+ H(R~~)o-c-R' Esterification [12,14]

0

0

~~

0

II + R-0-C-R

0

II

II

+ R'-O-C-R"

R-C-OH

0

+ R-0-C-R

I A II R

Oxamline ring opening [12,14,16]

II

0

N 0 +R'-C-OH U O H II

I

+R -C -N-CH~

-CH~-O-C

0

II

-R

Cross-reaction [12,14,17-191

Ar-0-Ad

+ Ar"-O-Ar"'

Figure 2. Examples of common compatibilised reactions between functional blend constituents [13] groups along the backbone and can undergo exchange reactions under suitable reaction conditions. &active blending can be applied to many polycondensates, opening the possibility for the preparation of a number of new materials by the direct combination of two or more polymers in processing machines such as extruders. This processing technology has many desirable features. It is complementary to polymerisation processes, has a high flexibility and versatility, requires lower capital investment, and is

Inhibition of Transreactions in Condensation Polymers

223

Table 1. Types of chemical reactions performed by reactive extrusion Type

Description

Preparation of high molecular weight polymer from monomer or low molecular weight prepolymer, or from mixture of monomers or monomer and prepolymer Formation of grafted polymer or copolymer from reaction Graft reaction of polymer and monomer Interchain copolymer Reaction of two or more polymers to form random, graft, formation or block copolymer through either ionic or covalent bonds Coupling/crosslinking Reaction of polymer with polyfunctional coupling or branching agent to build molecular weight by chain exreactions tension or branching, or reaction of polymer with condensing agent to build molecular weight by chain extension, or reaction of polymer with crosslinking agent to build melt viscosity by crosslinking Controlled degrada- Controlled molecular weight degradation of high molecular weight polymer (controlled rheology), or controlled tion degradation to monomer Functionalisation/ Introduction of functional groups into polymer backbone, functional-group end-group, or side chain, or modification of existing functional groups modification

Bulk polymerisation

more environmentally friendly. The types of chemical reactions which are performed by reactive extrusion may be divided into different categories as described in Table 1 (see also Chapter 7).

1.3.1. Interchain copolymer formation Interchain copolymer formation by reactive extrusion is particularly useful for compatibilisation of immiscible polymer blends. Interchain copolymer formation may be defined as reaction of two (or more) polymers t o form copolymer. Most of the condensation polymers have nucleophilic endgroups, such as carboxylic acid, amine, and hydroxyl groups. These nucleophilic end-groups can form covalent bonds with suitable electrophilic functionality leading t o copolymer formation when a suitable electrophilic group is attached t o a second polymer. The kinds of electrophilic functionalities that can form covalent bonds across a phase interface within the time constraints of an extruder are cyclic anhydride, epoxide, oxazoline, isocyanate, and carbodiimide. Kotliar [21] has reviewed the interchange reactions involving condensation polymers such as polyesters and polyamides.

N. R. James, S. S. Mahajan, S. Sivaram

224

The various types of interchange reactions in polyesters are

-0-coIntermolecular alcoholysis

+

- OH

-0-coIntermolecular acidolysis

+

-

A

-OH

+

CO

I

0

-

-COOH+

-

-0

A

-0

I

-co

-COOH

-0-coTransesterification

+

-co-0

A

+

I

-co

co I

0

Interchange reactions in condensation polymerisation of polyesters and polyamides have been studied extensively [22-261. It is also known that such reactions occur during melt processing of these polymers and form copolymers which change from block to random structures with time. Very little information is available in the open and patent literature about interchange reactions in polyamide/polyamide blends. Flory [22]suggests that aminolysis occurs with the possibility of amide exchange in polyamide melts. These reactions are -NH-CO-

Intermolecular acidolysis

+

-

A

Intermolecular aminolysis

+

- NHa -NH-CO-

Intermolecular amidolysis

+

-CO-NH-

I

+ -COOH

-co

-COOH -NH-CO-

-NH

.A

-NH2

+

CO-

I

NH

A

-NH

I

-CO

+

-

CO-

I

NH-

Blends of nylon 6 with nylon 6,6 are commercially important since they can give certain desirable properties, such as low-temperature toughness, increased flexibility, etc. Recent patents describe the preparation and properties of these blends [27,28]. According to these patents nylon 6 and nylon

225

Inhibition of Transreactions in Condensation Polymers

6,6 homopolymers undergo facile interchange reactions in the presence of certain catalysts to yield random copolymers during the processing. Verma et al. investigated [29] the melt blending of nylon 6 with nylon 6,6 without the addition of any external catalyst and found that little reaction occurs, as deduced from two distinct melting peaks of the homopolymers in the blend. Prevention of exchange reactions and of fast exchange reactions are both quite important. However, not much work has been done in this field. Hence there is a need to study the reaction kinetics thoroughly to understand the mechanism of transamidation reactions. The chemical processes for interchain copolymer formation in extruder reactors are summarised in Table 2. The types of copolymers formed are illustrated through the reaction of two polymers, namely, AAAAA and BBBBB. Interchain copolymer formation through chain cleavage followed by recombination is not a very useful process, as resulting copolymers are obtained as mixtures having broad molecular weight distributions. In the early Table 2. Chemical processes for interchain copolymer formation in extruder reactors Process

Type of copolymer obtained

Block and random AAAAABBBBBB AABBBBBAAAAA + AABBAAABBB, etc. Block A AAAABBBBB

1.Chain cleavage/recombination

2.End-group of 1st polymer end-group of 2nd polymer 3.End-group of 1st polymer pendant functionality of 2nd polymer

+

+ +

4.Covalent crosslinking:

pendant functionality of 1st polymer pendant functionality of 2nd polymer or Main chain of 1st polymer main chain of 2nd polymer 5.Ionic bond formation

+

Graft

A BBBBB A A-BBBBB A

+

Graft (crosslinked network)

Graft (crosslinked network) Usually graft (usually crosslinked)

A B A-B A B A-B A B

226

N.R.James, S. S. Mahajan, S. Sivaram

stages of such a process, block copolymers may predominate over random copolymer; however, such block copolymers have molecular weights less than the sum of the homopolymers and they may not be effective compatibilisers compared with the block copolymers formed by end-group/endgroup reaction. Block or graft copolymers are formed in extruder reactors through reaction between functionalised end-groups of two different polymers. Because the probability of end-groups reacting within the extruder time scale is low, highly reactive functionality is necessary on the two substrate polymers. Interchain copolymer formation by either covalent or ionic crosslinking is also possible in extruder reactors. In covalent crosslinking, pendant functionality of one polymer reacts with pendant functionalities of a second polymer to form a covalent bond. In ionic crosslinking, acidic groups, such as carboxylic, sulfonic, or phosphoric acid are present on both polymers and are at least partially neutralised by a metal cation, usually divalent, which may form a bridge between the two polymers being extruded. In some cases, polymers containing masked ionomers, chemical groups which form ionic species during the extrusion process, have been used to form copolymers during reactive extrusion. Use of masked ionomers may result in lower energy during extrusion processing since ionic self-associationneed not be overcome before interchain copolymer formation can occur. 2. Control of transesterification in polyester blends

2.1. Introduction

A fascinating feature of blends in the polyester family is the transreaction which is also known as transesterification. Ester exchange reactions have been reported to occur during extrusion of polyester blends [30-331. As transesterification proceeds, blends convert first to block copolymers and finally to random copolymers. The resultant initial blocks and eventual random copolymers are expected to exhibit enhanced mutual miscibility over the original unreacted components. Transesterified copolymers may exhibit some advantages such as higher tensile strength in polycarbonate/polyarylate blends [34], but these blends display poor impact properties [35]. The extent of exchange reaction is a key variable with regard to the phase behaviour of a given blend. There is an increasing interest in the understanding of the exchange reactions that take place in polymer blends between different functional groups involved in the mixtures of some polycondensation polymers. Of particular interest have been the industrially important blends based on polyester/polyester or polyester/polycarbonate. The chemical structure, and therefore the properties, of the resulting polymeric materials, are controlled by the relative rate and extent of several reactions which occur during the melt-blending. Transesterification makes it possible to obtain

Inhibition of Transreactions in Condensation Polymers

227

copolymers with different levels of randomness and composition by changing the weight fraction of each mixed polymer and reaction conditions such as temperature and residence time. Since ester exchange reactions affect mechanical properties, it is essential to control the extent of exchange reactions during melt-blending in order to achieve the desired properties. Devaux et al. found that residual polymerisation catalyst caused ester exchange reactions. Organophosphorus compounds, such as phosphites [36,37],phosphonates [38], and phosphates [39],have been used to inhibit ester exchange reaction in the molten state. Cheung et al. have also proposed the inhibition of ester exchange reactions by organophosphates in a polyarylate/polycarbonate/poly(ethylene terephthalate) (PAr/PC/PET) ternary blend [40]. Control of processing conditions to minimise the extent of exchange reactions in the polycarbonate/polyarylate blends has been discussed by Mondragon and Nazabal[34]. The activity of a catalyst, either present as a residue from polymer synthesis or purposely added before blending, may play a role in controlling the chemical structure of the resultant product. 2.2. Inhibitors f o r transreaction in polyester and polycarbonate blends

2.2.1. Organophosphite catalysts Molten polyesters can undergo transesterification during processing, resulting in chemical structures ranging from diblock copolymers with high molecular weight blocks, to multiblock copolymers with shorter and shorter segments, and eventually to random copolymers. Ester exchange reactions in polyester blends may lead to materials with poor mechanical properties. To inhibit or suppress transesterification, the use of organophosphites has been suggested in the patent and technical literature. The exact mechanism for the stabilising effect of the phosphite is unknown. Devaux et al. have found that organophosphite was useful to inhibit ester exchange reactions by forming an octahedral complex with residual catalysts which catalyse the transesterification reaction. They observed that phosphites, such as di-n-octadecyl phosphite, diphenyl phosphite and triphenyl phosphite were very effective in the inactivation of titanate catalyst residing in poly(butylene terephthalate) (PBT) after polymerisation. Triphenyl phosphite appeared to be the most effective inhibitor after an exchange of its phenol ligand by hydrolysis. Hydrolysis of triphenyl phosphite leads to diphenyl phosphite which upon further hydrolysis generates phosphoric aicd. This acid can tautomerise to phosphonic acid. According to Verhoeven [41], phosphonic acid is a possible titanate inhibitor. Cheung et al. [40] have reported the results of inhibiting the ester exchange reaction by organophosphites in a polyarylate/polycarbonate/poly(ethylene terephthalate) (PAr/PC/PET) ternary blend. According to them, the phosphites

N.R.James, S. S. Mahajan, S. Sivaram

228

appear to be transformed into phosphonates, which are effective inhibitors for the exchange reactions. Using solid-state 31P NMR spectroscopy it was shown that for bis(2,Cdi-t-butylpheny1)pentaerythritoldiphosphite (Ultranox 624), a conversion of the phosphite group to phosphonate, via hydrolysis at high processing temperature, is a prerequisite for an effective inhibition of transesterification. This conversion occurs readily during melt compounding if the polymers are not completely dry. However, if rigorous drying is employed and phosphite conversion does not occur, then transesterification is not arrested. The results indicate that the organophosphite may have reacted with the moisture from the environment during extrusion through side-chain hydrolysis, been converted into the tricovalent phosphoric acid (H-0-P) and later tautomerised into the more stable pentacovalent form (H-P=O) as follows:

+ 2H20

1

H O - P 0< o x o 0 >P -OH

+ 2&OH

11

-0a

The carboxyl end-groups in PET can react with the diphosphite (Ultranox 624)and form phosphonic acid, which is subsequently transformed into a phosphonate, as follows: -l&O-P
>P

+HOOCtPET

+ PET+OOC&OH

Inhibition of Transreactions in Condensation Polymers

229

It was also found that the conversion of the diphosphite in Ultranox 624 occurs not only during the processing but also at room temperature and in tightly closed steel containers, albeit slowly. Thus, in a period of about one yearc of storage in steel containers, sufficient conversion of the previously ineffective inhibitor took place to bring about a substantial improvement in the effectiveness of Ultranox 624 in suppressing transesterification reactions in blends of PET and PBT with polyarylate and polycarbonate. Differential scanning calorimetry (DSC) was employed to evaluate the effectiveness of the inhibitor. When an organophosphite is used to suppress the ester exchange reactions, it has been found that whenever the phosphite is converted into a phosphonate, there is little change in both the melting point and the heat of fusion of PET during prolonged exposure to high temperature in the blends. The addition of an organophosphate to the PAr/PC/PET ternary blend effectively retards ester exchange reactions when the extrusion temperature is up to 280°C. Beyond 280"C, the phosphite stabilised blend shows signs of instability. The criterion for stability is the retention of the cold crystallisation exotherm and the heat of fusion of PET phase. The addition of both organophosphite and carbodiimide produces a blend which is stable with regard to the melting point and the heat of fusion of PET up to an extrusion temperature of 300°C. Incorporation of hindered phenol as a third stabiliser allows an extrusion temperature of 325°C. This temperature range is needed for blends which require higher processing temperatures and longer residence times in the extruder. The ternary blend of PAr/PC/PET shows consistency at temperatures ranging from 280°C to 325°C (see also Chapter 3). 2.2.2. Non-organophosphite catalysts

Sulphuric acid [42] and its salts have been mentioned as inhibitors for PET/PBT/PC blends. Suppression of transesterification in PC/PBT blends has also been achieved by the addition of about 1 wt% of nylon 6,6 or polyacrylamide [43]. In the case of bisphenol A polycarbonate (PC) and PET, the activity of a catalyst, either present as a residue from polymer synthesis, or purposely added before blending, may play an important role in controlling the chemical structure of the find product. Recently Fiorini et al. [44,45]have reported that catalysts based on lanthanide compounds possess a wide range of catalytic activity toward different reactions taking place during PET/PC reactive blending. The properties of the resultant product prepared by reactive blending of PET with PC can depend strongly on the reactions taking place during melt-blending. The selective activity of the catalyst toward exchange and degradation reactions and the solubility of the catalyst within the different phases may be key factors for the process. Hence, the choice of the appropriate catalyst becomes very important as it will allow materials with the desired properties to be prepared.

N. R. James, S. S. Mahajan, S. Sivaram

230

2.2.3. Processing conditions In the case of immiscible mixtures, blending depends not only on the miscibility level of components, but also on the processing conditions. When mixing of components takes place in the molten state, the control of interchange reactions provides an opportunity to obtain alloys which are more homogeneous as compared to the corresponding physical mixture. Transesterification in polyester blends depends strongly on their initial compatibility and the blending conditions, such as temperature, residence time in the molten state, as well as percentage of each component. During transesterification reactions a change in the melt viscosity is expected due to newly formed structures. This viscosity variation produces a corresponding change in the torque required to turn the Brabender mixer. Mondragon and Nazabal [34] have suggested a method of controlling interchange reactions of PC/PAr during processing in Brabender Plasticorder by plotting the torque variation against residence time. They also suggested an exchange reaction mechanism which analyses the viscosity variation as a function of time and blend composition at various temperatures. During the transesterification of PC/PAr, the torque required to turn the Brabender vs. residence time was recorded. This torque is an indicator of melt viscosity. Torque variations of different PAr compositions us. residence time at various temperatures are illustrated in Figure 3. For all compositions the torque drops initially and becomes stable after some time when the two blend components are homogeneously mixed (Figure 3). The viscosity increases progressively with increase in residence time until it reaches a maximum value, which is attributed to exchange reactions between the two polymers. The torque increase is small because these transesterification reactions do not produce crosslinking. A decrease in viscosity at maximum torque shows that the degradation effect is stronger with re-

40%

30 -

15

15 10%

25

50

Time (min)

75

15

30

Time (mi.)

45

- 0

10

20

Time (mi.)

30

Figure 3. Torque variations of PAr compositionsus. residence time at: (a) 25OoC; (b) 27OoC; (c) 29OoC [34]

23 1

Inhibition of Transreactions in Condensation Polymers I

0

20

40

% PAr

60

80

100

Figure 4. Torque data of blends before transesterification (empty symbols) and after transesterification (filled symbols) 'vs. mixture composition at the temperatures used: (1) 290°C; (2) 270°C; (3) 250°C [34] spect t o the exchange reactions. At 250°C (Figure 3a), the torque shows maximum at longer residence time. The torque increases with composition and the major variation occurs at about 25% PAr. Similar results are observed at higher processing temperatures, ie., 270°C and 290°C (Figure 3b,c). At higher temperatures maximum torque appears at shorter times. At these temperatures also major variation occurs at 25% PAr. At lower PAr content the maximum takes a longer time to appear and the torque change appears greater. This could be because of the higher viscosity of PAr, which necessitates the use of higher temperatures during processing when the PAr content increases. As a consequence, interchange reactions are faster when the blend has a high PAr content. In Figure 4 it can be seen that at higher PAr contents the torque of the blend is closer to the torque of pure polymers. This observation could suggest that miscibility exists in PAr-rich compositions. However, the torque behaviour suggests that these polymers are immiscible or partially miscible and the first transesterification step develops in the PAr-rich phase. In

232

N.R.James, S. S. Mahajan, S. Sivaram

the absence of this transesterification reaction the torque data of the blend would have been below the torque data of pure polymers in all compositions. At low PAr content, the amount of copolymer which can be produced in the first step would be low, and hence it would have less influence on the torque data of the blend. In Figures 3 and 4 it can be seen that the interchange reaction consists of two steps. The first step is fast and occurs in the PAr-rich phase, but it is not observable in the torque-time plot. The second step proceeds between the two phases of the blend and produces slope variation in the torque. As shown in Figure 3 at low PAr contents, the variation between the maximum and the stabilised torque is higher because of the smaller amount of PArrich phase. During blend preparation in the Brabender it was observed that when the torque stabilised, the transparency of mixture varied with the temperature applied. Thus, at 250°C, blends with a PAr content of more than 60% were transparent, whereas at 270°C this limit decreased to 50% PAr, and at 290°C the same composition was observed to be transparent. Blends with a lower PAr content appeared opaque at these temperatures. These observations suggest that the miscibility limit in this blend exists between 50 and 60% PAr. Transparency can be due to copolymer formation in the PAr-rich phase, which could act as a compatibilising agent at high PAr contents. Moreover, at higher melt temperatures transesterification takes place more quickly and for this reason the higher the temperature, the lower the PAr content necessary for transparency to be observed. In all compositions when the maximum in the torquetime curves was attained, transparency was observed in both melt and solid states. This is a clear indication of a structural change in the initially immiscible blends due to transesterification and it shows that the copolymers formed are fully amorphous. The above observation indicates that the exchange reaction between PC and PAr can be controlled by selecting an adequate temperature, thus allowing copolymer formation during a different residence time. More recently, it has been observed that ester-carbonate interchange reactions occur even in the solid state at temperatures below 23OOC [46]. Mixtures of oligomers derived from bisphenol A polycarbonate and polyesters (PET or poly(ary1ester)) were precrystallised and subjected to solid-state polymerisation in the temperature range 180-230°C. During this reaction both chain extension and interchange occurred. High molecular weight copoly(ester-carbonate)s were obtained at the end of solid-state polymerisation. Analysis of polymer structure by a combination of techniques, such as selective dissolution, FTIR, and 'H NMR, provide an evidence for the occurrence of carbonate-ester exchange reaction.

233

Inhibition of Transreactions in Condensation Polymers 3. Methods of analysing transreactions in polymer blends

3.1. IR spectroscopy Several methods have been used for analysing transreactions in polymer blends. One of the most commonly used techniques is infrared spectroscopy (IR), which originates from molecular vibrations that cause changes in the dipole moment and polarisability of molecular chains. These spectra are unique to each molecule and hence reflect the chain structure. Transreactions in polycondensates lead to the formation of new copolymers. If we consider a system of PET and PC, new copolymers such as aromatic ester and aromatic carbonate will be formed due to transreactions. The progress of evolution of new components can be inferred by IR spectroscopy [47-501. Specific vibration absorption bands of carbonyl groups in PET and P C are at 1720 and 1780cm-', respectively. The stretching vibration of the carbonyl group in aromatic ester and aromatic aliphatic carbonate show specific absorption bands [47] at 1070, 1740, and 1770cm-l. When

i.1 A 10 min

20 min

30 min

1840

60 min

Wavenumber (crn-l)

1680

R. 80 min

1840

1680

Figure 5. IR spectra of soluble fractions extracted with CHzClz from a PET/PC blend (1:l by wt) melt-mixed at 27OoC for different periods of time [47]

111

N. R. James, S. S. Mahajan, S. Sivaram

234 for 5 min

1740 1719

1720

60 rnin

30 rnin

15 rnin

1743

1744

t717

u 1 7 i 1800

1601

u 1800

160

Wavenumber

1800 (cm-l)

1600

I 1800

1600

Figure 6 . The spectra recorded for PET/PCblends (1:lby wt) prepared at 300°C for: 5 min; 15min; 30 min; 60min [47] PC/PET blends were melt-mixed for various periods of time and extracted with CHZClz, which is a good solvent for PC, significant changes occurred in the IR spectrum of the soluble fraction (Figure 5). For longer periods of mixing a progressive evolution of the band at 1720cm-l was observed. When the blending temperature was raised to 300°C,new absorption bands were seen at 1070 and 1740cm-', and the band at 1780cm-1 was shifted

--li

U

1800 1700 (cm-')

(3)

1800 1700 (cm-

Figure 7. Evolution of C=O stretching bands in a PC/PBT blend (50/50 by wt) as a function of reaction time at 243.5OC: (1) Omin; (2) 5min; (3) 10min; (4) 15min; (-) soluble fraction; (----) insoluble fraction 1471

Inhibition of Transreactions in Condensation Polymers

235

to 1770cm-l. The intensity of the bands at 1070 and 1740cm-' increased with mixing time (Figure 6). The occurrence of new bands and the band shift indicate formation of aromatic ester and aliphatic-aromatic carbonate by interchange reactions. For a PC/PBT system similar observations were made [48]. The progressive appearance of absorptions at 1780crn-' and 1720cm-' in the soluble fraction indicates transesterification (Figure 7). At longer reaction times new bands appeared at 1740cm-l and 1070cm-' and the band at 1780cm-' shifted to 1770cm-'. As mentioned earlier, the bands at 1070cm-' and 1740cm-l are characteristic of an aromatic ester structure. The band at 1070cm-' can be attributed to a complex vibration of the pamdisubstituted phenyl to the right of the ester structure influenced by the neighbouring -COO group [48]. The absorption at 1770cm-' results from the -C=O stretching of a mixed aliphatic-aromatic carbonate. In the case of ternary blends (PC/PAr/PET) the situation is more complex [40], since the aromatic ester groups, the presence of which is an indication of exchange reactions, exist in the system itself. Thus the changes in the infrared spectrum will be more subtle and some of the exchange products will not be distinguished from the reactants. Figure 8 shows the infrared spectra of the unstabilised blend as extruded, the unstabilised blend after isothermal treatment at 300°C for 30min and the triply stabilised blend (details are given in Table 3). In curve 1 (Figure 9) there is evidence of an increase of the peak at 1070cm-', but the corresponding changes in -C=O stretching region (1800-1700cm-') are not clear. Curve Microns

Wavenumber (cm-l) Figure 8. IR spectra of PAr/PC/PET blend without stabiliser and extruded at 28OOC: (1) before isothermal test; (2) after 30min at 300°C; (3) fully stabilised system (sample F, Table 3) [40]

236

N.R. James, S. S. Mahajan, S. Sivaram Table 3. Results of isothermal DSC tests

Sample

Extrusion

11’

I’

280

X

B

300

X

C

300

X

X

D

300

x

R-pstyb

E

325

X

X

F

325

X

X

(“C)

Time at

AHf‘

300°C

(J/g)

111’

temp.

A

a

Stabilisers

(mi4 0 30 0 30 0 30 0 30 0 10 30 0 30 60

X

12.2 10.2 9.7 6.2 12.2 12.0 14.4 10.0 7.7 7.0 5.3 11.8 13.3 10

Tm (“C) 249 245.7 253.7 247 249.7 249.7 254 251 251.8 251.1 244.4 252 248 246

I Ultranox 624;I1 Stabaxol P-100;I11 Ethanox 330 R-Psty, reactive polystyrene AHf, Enthalpy of fusion

5.0

5.5

6.0

6.5

7.0 7.5 8.0

9.0 10.0 11 12

t

3 4

2000

1800

1600

1400

1200

Wavenumber (cm-I)

1000

800

Figure 9. IR spectra of: (1) sample B (Table 3); (2) sample C (Table 3) after 30 min at 300°C [40]

237

Inhibition of Transreactions in Condensation Polymers

Microns

Wavenumber (cm-l) Figure 10. IR spectra of sample E (Table 3): (1) after 10min at 300°C; (2) after 30 min at 300°C [40]

2 clearly shows a reduction of the bands at 1780 and 1770cm-1 and increase in intensity for the bands at 1070cm-' and 1740cm-' indicating Microns 5.0

2000

5.5

1800

6.0

6.5

1600

7.0 7.5 8.0

1400

9.0 10.0 11

1200

1000

Wavenumber (cm-l) Figure 11. IR spectra of sample F (Table 3): (1) before isothermal test; (2) after 60 min at 300" C [40]

N. R. James, S. S. Mahajan, S. Sivaram

238

1

1765

1730

1695

1765

1730

1695

Wavenumber (cm-')

Figure 12. Evolution of C=O stretching bands after different treatment times: (a) Omin; (b) 5min; (c) 15min; (d) 25min (-); soluble fraction; (---) insoluble fraction [50] formation of aromatic ester and loss of carbonate functionality. Figure 10 shows IR spectra of the blend after the addition of stabilisers I and 11. The comparison of the curves proves the improved stability on addition of the stabiliser 11. When the sample is heated at 300°C for 10 and 30min, the absorption at 1 0 7 0 ~ m -increases ~ markedly (Figure 10). Figure 11 shows the spectrum of a fully stabilised system where, even after treatment at 300°C for 60min, the absorption at 1 0 7 0 ~ m -remains ~ constant; an increase in absorption is expected whenever there is an exchange reaction.

239

Inhibition of Transreactions in Condensation Polymers

Considering the PAr/PET system, -C=O stretching bands of pure PAr and PET appear at 1740 and 1720cm-', respectively [50],when no exchange reaction has taken place. When the 50:50 PAr/PET blend is heated at 300"C, transreactions occur and blocks of PAr appear in the insoluble fraction as well as PET sequences in the soluble fraction. When the reaction proceeds, the bands become similar, eventually leading to similar maximum and shape (Figure 12). According to Berti et al. [51]the overlapping of several carbonyl stretching bands does not allow a clear separation of bands and conclusions based on IR data can be misleading. They postulate elimination of ethylene carbonate (EC) as a main reaction, rather than decarboxylation as was assumed by Godard et al. [32]. This was inferred by the presence of a small peak at 1810cm-' observed when melt-mixing of P C and PET was carried out in a closed system, which disappeared when the reaction was performed in a n open system, under vacuum, or in a stream of nitrogen. N

3.2. NMR spectroscopy

Valuable insights regarding transreactions occurring in polycondensates are provided by 'H, 13C1and 31P nuclear magnetic spectroscopies.

'

3.2.1. H and

l3

C nuclear magnetic resonance spectroscopies

For the PC/PBT system, important information can be obtained from the region between 8 and 8.4ppm [48]. The NMR spectra in this region for P C/ P BT (50/50 by wt) observed after different reaction times at 253°C (Figure 13) show progressive appearance of terephthalic ester units substituted by one or two aromatic groups. Significant changes occur in the region between 7 and 7.5ppm corresponding to aromatic protons of bisphenol A 2)

13)

:41

(5)

-L ---d 83 82 81 8C PPm

Jd

83 82 81 8C PPm

83 82 81 8 PPn

83 82 81 80 PPm

u 83 82 81 8C PPm

Figure 13. NMR peaks of protons of terephthalate units in the PC/PBT system (50/50 by wt) after different reaction times at 253OC, solvent CDCk: (1)30min; (2) 60min; (3) 100min; (4) 200min; (5) bisphenol A polyterephthalate [48]

N. R. James, S. S. Mahajan, S. Sivaram

240

Table 4. Evolution of PBT 13C peaks and relative intensities during the reaction (chemical shifts in ppm relative to TMS)

0

0

c5,6 Reaction Attribution C ~ J O Cz,g C3,8 c4,7 Chemical 23.8 65.7 68.1 68.7 168.3 167.5 132.3 132.9 133.4 128.9 129.6 time shifts (min) (PP4

and the region around 4.2ppm corresponding to a methylene group adjacent to oxygen (-CHz-O) as a result of modifications in the surroundings due to interchange reactions. 13C NMR spectra of PC/PBT mixtures during the reaction show that the relative intensities of the carbon signals corresponding to PBT carbons decrease progressively. Table 5. Evolution of PC 13C peaks and relative intensities during the reaction (chemical shifts in ppm relative to TMS)

Attribution C;JZ C L Cklo Cis C;, Cg C; Reaction Chemical 154.8 155.8 156.3 147.7 147.3 119.1 119.6 127.2 148.9 41.6 28.7 time (min) shifts (PP4 Relative intensities

NM: not measurable

241

Inhibition of Transreactions in Condensation Polymers The labelling of carbons is as shown below. (C) 0 0

- C ~ z - ~ H z - ~ - ~9 ~-cHz ~10 -- ~ - ~ ~ z 1

(C’>

2

3

5

6

CHs

0

0

- o - ! - o ~ + a o - ~ - IIo 1

3

4

CH39 10 7

Tables 4 and 5 show changes in the relative intensities of PBT and P C 13C peaks during reaction. The new c3,8 band appearing in the spectrum is due to the substitution of a butylene group by bisphenol A. The splitting of the C4,7 peak at 132.9 ppm into two symmetrical absorptions at f 0 . 5 ppm from the initial band is also a result of the substitution. Considering C5 and Cs the peak at 129.6ppm indicates asymmetrically substituted terephthalate groups while the peak at 128.9ppm characterises the carbon atoms substituted either by two butylene groups or by two bisphenol A groups. The &,9 peak corresponding to the carbon of the -CH2-O groups of PBT is at 65.7ppm. The two new bands at 68.1 and 68.7ppm probably result from the appearance of aliphatic-aromatic carbonate and aliphatic-aliphatic carbonate units. The presence of three types of carbonate ester units (Ci,12) is clear from peaks at 156.3ppm, 155.8ppm and 154.8ppm corresponding to ester carbonate bearing two butylene groups, one butylene and one bisphenol A group and two bisphenol A groups respectively. The C2,11 and Ci,lo signals of bisphenol A groups are split into two bands due to the simultaneous presence of bisphenol A carbonate and bisphenol A terephthalate units in the copolyester. The C;,*, Cg, and C!, bands remain practically unchanged during the reaction. Similar changes occur in the case of PC/PET systems (Tables 6 and 7) [321. (C) 0 0

0

CHn

0

242

N. R. James, S. S. Mahajan, S. Sivaram

Table 6. Evolution of PET I3C peaks and relative intensities during the reaction (chemical shifts in ppm relative to TMS)

0

0

Attribution c1,2 (33,s c4,7 c5,6 Reaction Chemical 66.7 65.7 65.5 63.7 62.9 165.3 164.2 133.3 133.6 130 129.6 time shifts (min) (PP4

Reaction at 235OC.NM: not measurable

The relative intensities of the “carbon”signals (165.3, 133.6, 129.6ppm C S , ~C4,7, , C S , ~corresponding ) to PET carbons decrease progressively. The new peaks appearing at 164.2, 133.8 and 130.0ppm can be attributed to terephthalate units substituted by an aromatic species. The peak corresponding to methylene carbons ( C I , ~undergoes ) a more complex evolution. The peak at 62.9ppm decreases gradually, finally giving rise to two peaks at 65.7 and 63.7ppm. This indicates progressive substitution of a

-

Cg C$ Fhaction C b Ici.9 Ci.8 Attribution C&11 Chemical 156.3 148.9 148.6 120.8 120.2 114.1 127.8 148.0 143.1 42.4 42.0 30.8 time shifts (min) (PP4

Inhibition of Transreactions in Condensation Polymers

243 b

a

J 9.0

8.5

8.0 ppm

7.5

7.0

9.0

8.5

8.0

PPm

7.5

I 7.0

Figure 14. 'H NMR spectra of (a) PAr/PBT blend before transesterification; (b) soluble fraction of PAr/PBT after transesterification 152) neighbouring terephthalate by a carbonate group, which gives rise to two non-equivalent methylenes. At longer reaction times another set of two peaks appears at 66.7 and 65.7 ppm, indicating some reactions consecutive to the transesterification. A carbon dioxide elimination from the aromaticaliphatic carbonate group with ether formation was suggested. Considering the peaks of PC groups (Table 7), C;,,, and C$,lo peaks occurring in PC at 148.9 and 120.2ppm are initially split, two new peaks appear at 148.6 and 120.8 ppm belonging to new bisphenol A terephthalate groups. A comparison of l H N M R spectra of a PAr/PBT blend before and after transesterification (Figure 14) proves new signals at 8.3 and 8.21 ppm between those of terephthalate protons (a) and (b) (Figure 15) (6, = 8.41, 6b = 8.16) which can be attributed to the aromatic protons of the asymmetrically substituted terephthalate unit, indicating transesterification [52]. The absorptions corresponding to aromatic protons of the symmetrically substituted isophthalic units (d) appear at 9.08, 8.56 and 7.78ppm as a singlet, a doublet, and a triplet, respectively. These units can react with the PBT units giving rise to a type (e) structure. The absorptions at 8.87 (singlet), 8.48 (doublet) and 7.71 (triplet) ppm can be attributed to the aromatic protons of the asymmetric isophthalate units. Thus the occurrence of transreactions which drive the chemical structure toward that of a random copolymer is clearly indicated by the modification of all the regions of the NMR spectrum. From lH N M R there is also evidence of the formation of ethylene carbonate and the occurrence of side reactions [51]. In a PC/PET blend it is observed that the molar ratio

N. R. James, S. S. Mahajan, S. Sivaram

244

R - O O C ~ C O O - R

h-OOC-@OO--

R

b c

COO-Ar

I

h - o o c ~

d

COO- R

I

h - o o c ~

e

Figure 15. Structures referred to in the text (521

of ethylene glycol protons to the terephthalate protons changes from 1 to 0.84 while the molar ratio of bisphenol A to terephthalate units does not change. The observed change in molar ratio is assumed to be due to the elimination of EC from the reaction medium. In fact a small sharp peak at 6 = 4.51, typical of EC, is observed when the samples are prepared by melt-mixing in a closed system, but it disappears in the spectra of samples prepared under reduced pressure [51]. Thus 'H and 13C NMR spectroscopies prove to be useful techniques to follow transreactions taking place in polycondensates during melt-mixing (see also Chapter 1). 3.2.2.

31 P

nuclear magnetic resonance spectroscopy

Phosphorus-31 magic-angle spinning (MAS) NMR is useful in tracking the chemistry of phosphites in polymer blends, since it can determine, through the chemical shift interaction, the identity of phosphorus species present in the solid state and, to high precision, the relative and absolute concentrations of these moieties. The efficiency of the technique can be exemplified by the studies on pentaerythritol) the behaviour of Ultranox 624, bi~(2~4di-t-butylphenyl diphosphite (BTBP), an organophosphite stabiliser in a blend of PC, PET, and PAr prepared at 280°C [53]. Three concentrates of Ultranox 624 in polycarbonate were prepared under different conditions in a twin screw extruder (Table 8). The 31P CPMAS (cross-polarisation magic-angle spin-

245

Inhibition of Transreactions in Condensation Polymers

Table 8. Summary of concentrate preparation in twin screw extruder BTBP concentrate

Polycarbonate (Dow’s caliber)

Drying conditions

Compounding atmosphere

A B C

300-15 (low M W ) 300-15 (low M W ) 300-3 (high M W )

Mild Rigorous Rigorous

Air Nz blanket N2 blanket

ning) spectrum of neat BTBP shows an isotropic chemical shift line for BTBP at 115.5ppm (Figure IS), which was confirmed by the use of side band suppression pulse sequences. The broad peak centered at 7ppm is identified as an amorphous orthophosphite or phosphonate. It is reported that organophosphites can undergo oxidation to an organophosphate es-

-

Figure 16. Phosphorus-31 CPMAS spectra of neat BTBP (26% by wt in a nonphosphorus containing organic binder): (a) 3200 kHz, 100 acquisitions, arrows indicate isotropic chemical shifts; (b) 3500 kHz, 10 acquisitions. Both spectra represent 256 acquisitions at 120.5MHz using a 2 ms contact pulse and 10 s recycle time [53]

246

N. R. James, S. S. Mahajan, S. Sivaram

Figure 17. Phosphorus-31 CPMAS spectra of the PC/BTBP concentrate A (Table 8): (a) as extruded, 3500kHz, 14000 acquisitions; (b) after drying in aircirculating oven, 16 h at 100°C, 3200 kHz; (c) after drying under vacuum without heating, 3200 kHz; (d) clear pellets of concentrate A, 3500 kHz; (e) cloudy pellets of concentrate A, 3500 kHz. All spectra represent 256 acquisitions at 120.5MHz using a 2 ms contact pulse and 10s recycle time [53] ter or hydrolysis t o a n organophosphonate and isotropic chemical shifts of these species should be f 1 0 ppm to that of 85% phosphoric acid at 0 ppm.

Solid-state 31PNMR of concentrates. There is an overall increase in the line width in the spectra of PC/BTBP concentrates relative to that of the neat BTBP which can be due to the loss of crystallinity induced by the compounding process (Figure 17). Comparing the spectrum of the concentrate after heating in a n air oven at 100°C for 16 h and that after

247

Inhibition of Transreactions in Condensation Polymers

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I

250

200

150

100 PPm

50

0

1

-50

Figure 18. Phosphorus-31 NMR spectra of the PC/BTBP concentrate A: (a) CPMAS spectrum of concentrate A dried in an air-circulating oven at lOOOC (as in Figure 17b); (b) TOSS FID of the concentrate dried in an oven at 100°C; (c) CPMAS spectrum of pulverised concentrate A stored in ambient conditions for 30 days; (d) CPMAS spectrum of neat polycarbonate [53] vacuum drying, the former shows higher intensity of the resonance near 7ppm. Following heating under air, a fraction of the pellets turned cloudy; their spectrum showed predominance of two impurity peaks at 5.9 and 8.3ppm, apart from that at 7ppm (Figure 17). If we compare the CPMAS of the concentrate with total suppression of spinning side bands (TOSS; when TOSS is employed following the use of single 90' pulse for the creation of transverse magnetisation, the technique is called TOSS FID), it is seen that the signal for phosphorus (115.5ppm) is not visible, indicating that the species is in an immobile environment (Figure 18). But the peak at 7ppm is seen in TOSS FID, also indicating that the species is in a mobile or relaxed environment (Figure 18). TOSS FID confirms the presence of other impurity peaks. When the unheated pulverised sample was exposed to air for 30 days, a similar change occurred when heated under air. Thus the new products must have formed by either

N. R. James, S. S. Mahajan, S. Sivaram

248

, , I

I , 1 I I I I I l I I I I I I I 1 , 1 1 1

250

200

150

100

50

, I l l ,

0

I I I I I I I I

-50

ppm

Figure 19. Phosphorus31 spectra of blends prepared from PC/BTBP concentrate A (Table 8), 3200 kHz: (a) CPMAS spectrum of the PC/BTBP concentrate after drying under vacuum for 16 h in an air-circulating oven; (b) CPMAS spectrum of a polymer blend prepared using the PC/BTBP concentrate; (c) TOSS spectrum of the same blend, 3600 acquisitions, 2 s recycle delay; (d) TOSS PPS spectrum using a 40 ps delay of the same blend, 12 000 acquisitions 153) oxidation or hydrolysis.

Solid-state 31P NMR of polymer blends. The CPMAS NMR spectrum of the actual PC/BTPB concentrate used to produce the polymer blend (acquired at a high signal/noise (S/N) ratio) shows an impurity peak at 95ppm also (Figure 19). The 31PMAS NMR spectrum of the ternary blend shows complete absence of a peak at 115.5ppm (Figure 19). This was confirmed by TOSS and TOSS PPS (TOSS protonated phosphorus suppression) spectra. The former shows resonances at 95 and 7 ppm as well as a n extra signal at 42ppm which could be an additional product of the chemistry of BTPB. The TOSS PPS spectrum for a delay time of 40ps

Inhibition of Transreactions in Condensation Polymers

r , l l l l l l l l l l l l l l l I I I I I

250

200

150

100

1

1

50

1

1

1

0

1

1

249

1

1

1

1

1

1

1

~

-50

PPm

Figure 20. (a) CPMAS spectrum of PC/BTBP concentrate C (Table 8); (b) the corresponding polymer blend; (c) TOSS spectrum of concentrate B (Table 8); (d) the corresponding polymer blend [53]

shows the absence of a peak at 7ppm and a reduction in the intensity of the peak at 95 ppm. The absence of the peak in the TOSS PPS spectrum indicates a protonated phosphorus species in a relatively rigid environment, unlike the species in the original concentrate. Thus the 7ppm species in the blend cannot be a phosphate, but a phosphonate resulting from the hydrolysis of the phosphite triester. For such a species, which has a proton directly bonded to the phosphorus, we can expect short TOSS PPS delays to produce complete loss of intensity. The results of DSC analyses show that whenever the phosphite is converted into phosphonate, the melting point and the heat of fusion of PET are stabilised and transreactions are suppressed [54]. From the above discussion it is clear that 31PNMR spectroscopy provides information on the actual mechanism by which phosphitic stabilisers suppress transreactions in blends of polycondensates.

250

N. R. James, S. S. Mahajan, S. Sivaram

3.3. Dinerentid scanning calorimetry

DSC has been reported to be an effective qualitative technique to assess the

extent of transreactions occurring in blends of polycondensates. Associated with transesterification, there will be a decrease in AH1 [55].At advanced stages of transesterification, the crystallisabIe moiety will lose its ability to crystallise. If we consider blends whose components are noncrystallisable, such as bisphenol A polycarbonate and polyarylate, the T, values of each component will move towards each other with progressive transreactions, eventually giving rise to a single Tgcorresponding to the copolymer formed due to transesterification [40,47,55]. The copolymer produced by transreaction can enhance the miscibility of a system [47]. Multiple Tgvalues for a polymer blend indicate multiple phases. In the case of a PC/PET blend, the higher T, corresponds to PC while the lower Tg corresponds to PET. Figure 21 shows the change in Tg with PET/PC blend composition. The higher T,, related to PC,decreases at low levels of PET, then increases a little when the concentration of PET exceeds 60 wt %. The lower T,, related to PET, increases with increasing PC content through a flat peak and then decreases. The DSC thermograms (Figure 22) of PET/PC blended at different temperatures for 10min show dual transitions when blending temperatures are below 300°C.A single transition is found for the sample blended at 300°C. Figure 23 shows the change in Tgof the blends for various blending times at 270°C.It can be seen that, on prolonged mixing, the difference between the two transition temperatures decreases and they finally merge into one Tg, indicating the progress of transreaction. 1501

I 1 0

I

20

I

40

I

60

Content of PET (%)

I

80

I

100

Figure 21. Change in T,with PFCT/PC blend composition. Samples were annealed at 18OoC for 60min prior to the second run. Measurements were carried out by DSC at a heating rate of 40°C/min [46]

25 1

Inhibition of Transreactions in Condensation Polymers

In the case of blends with a crystallisable moiety, such as P E T with PAr and with PC, the transesterified polymers are noncrystallisable and exhibit a single Tg between that of the starting polymers [31,56]. If we consider blends of PBT with PAr and with PC, blends without transesterification show a single T,,indicating amorphous miscibility of P B T and PAr, but without melting point depression, implying that PBT crystallises with exclusion of PAr. However, transesterified systems show higher T' values than the corresponding physical blends and exhibit a marked melting point depression and a lower PBT crystallinity [57,58]. The physical blends of PC and PAr are found to exhibit two amorphous phases, a pure P C phase and a PAr-rich mixed phase [35], giving two separate T, values at 149°C and 178"C, respectively. On controlled thermal treatment, transreaction between P C and PAr takes place leading to a new copolymer with a single T,, depending on the original binary composition. The progress of reaction from block copolymers to random I

1

I

I

I

I

I

3OO0C 290°C 280OC 270°C

26OoC

PET 80

120

160

Temperature OC

200

Figure 22. Change in DSC Tg behaviour of PET/PC (1:l by wt) blended at different temperatures for 10min [47]

N. R. James, S. S. Mahajan, S. Sivaram

252

copolymers can be traced by DSC [35].The DSC curve of a PC/PAr blend (50/50) processed in a single screw extruder at 305°C (Figure 24) shows a single Tgcorresponding to PC (a PAr-rich phase may also exist). But the blend extruded at 325°C shows a single Tgat 160"C, indicating complete miscibility (Figure 24). The DSC curves of PC/PAr blend film cast from methylene chloride solution reveal two Tgvalues but both transitions seem to have shifted toward each other, indicating partial miscibility (Figure 25). An interesting observation is that the extrudate at 285"C, which shows two distinct Tgvalues, exhibits a single Tgwhen injection moulded at a melt temperature of 290°C (Figure 26) and the originally opaque sample becomes transparent. The same sample when annealed at 180°C turns opaque. A solution-cast film of moulded sample was also prepared, the DSC thermograms of which show reappearance of Tgat 149°C (PC-rich I

I

I

I

min at

27OoC

30

35 60

90

80

100

120

140

Temperature OC

160

Figure 23. Change in DSC T, behaviour of PET/PC (1:l by wt) with blending time at 27OoC [47]

253

Inhibition of Transreactions in Condensation Polymers

I

100

I

,

,

I

120

1

1

,

1

140

,

,

,

1

,

160

,

,

I

180

,

,

,

I

200

,

,

,

I

OC

Figure 24. DSC curves of a PC/PAr blend (50/50) single screw extruned at: (1) 285OC; (2) 305°C; (3) 325OC [35]

phase) , indicating phase separation. However, a solution-cast film of the blend compounded in a twin screw extruder shows a single T,,confirming transesterification in advanced stages (Figure 25).

183OC 100

120

140

160

180

200

Figure 25. DSC curves of 50/50 PC/PAr blends: (1) resins dissolved in methylene chloride at 23OC and cast into a film;(2) twin screw extrudate dissolved in methylene chloride and cast into a film (351

N. R. James, S. S. Mahajan, S. Sivaram

254

I

100

.

*

.

I

120

"

'

I

'

~

3

140

"

"

160

~

~

180

'

"

200

'

'

.

OC

1

Figure 26. DSC curves of PC/PAr blends (50/50): (a) reproduction of curve (1) of Figure 25; (b) injection-moulded A; (c) sample of B dissolved in methylene chloride and cast into a film [35]

260

I 0

5

1

20 15 Time at 277OC (min) 10

I

25

I

30

Figure 27. Change in melting point of copolyester in 50% blend (without additives) with time in the melt [38]

From the above discussion it is clear that DSC can be used as an efficient technique to follow transreactions. There are many reports of the use of DSC in studies regarding the inhibition of transesterification

255

Inhibition of Transreactions in Condensation Polymers 112' 110

u

I

I

5

10

I

I

-

h

h"

108-

104 0

I

15 I

20 I

Time at 277OC (min)

25 I

30

Figure 28. Change in T, of 50% blend (without additives) with time in the melt 1381

[31,38,40,54,55,59].A blend of polycarbonate and the copolyester based on 1 ,Pcyclohexane dimethanol and a mixture of terephthalic and isophthalic acids is known to undergo transesterification [38]. Figures 27 and 28 show 16.0 12.0

h

M

\

c,

v

B

!3

8.0

Icl

0

3

8

4.0

0 0

5

10

15

Time at 277OC (min)

20

25

30

Figure 29. Effect of various additives on the crystallisability of copolyester from 50% blend held in the melt for various times [38]

256

N. R. James, S. S. Mahajan, S. Sivaram

50

100

150

200

250

OC

Figure 30. DSC scan of PAr/PC/PET ternary blend with concentrate A (Table 9), predried at 100°C for 16 h, after being held isothermally for 30 min at 300°C 1541

the change in T, and Tg with time in the melt for the 50% blend containing only the residual catalyst from the copolyester. In order to inhibit the transesterification, arsenic compounds, or commercially available stabilisers were incorporated into the blends in a subsequent extrusion step. The

50

100

150

200

250

OC

Figure 31. DSC scan of PAr/PC/PET ternary blend with concentrate A (Table 9), predried in vacuum at ambient temperature for 16 h, after being held isothermally for 30 min at 3OO0C [54]

257

Inhibition of Transreactions in Condensation Polymers

Figure 32. DSC scan of PAr/PC/PET ternary blend with concentrate B (Table 9), predried at 100°C for 16 h, after being held isothermally for 30min at 300°C 1541

ability of As203 as well as phosphite stabilisers to preserve the crystallinity is evident from the heat of fusion values obtained by DSC (Figure 29).

I

50

,

l

,

,

l

100

l

l

.

l

l

150

l

.

.

l

l

200

.

l

l

l

l

250

l

.

I

OC

Figure 33. DSC scan of PAr/PC/PET ternary blend with concentrate C (Table 9), predried at 100°C for 16h, after being held isothermally for 30min at 300°C 1541

Cheung et al. reported the use of DSC to evaluate the effectiveness of stabilisers to inhibit transesterification in a ternary blend of PAr, PC and

N. R. James, S. S. Mahajan, S. Sivaram

258

I

I

100

50

150

200

250

Figure 34. DSC thermograms of PET, PC, and PAr {40]

PET (50/20/30) [54]. The criterion is based on the retention of the melting point and the heat of fusion of PET [31,40,58]. Figures 30, 31, 32, and 33 show DSC thermograms of ternary blends stabilised with transesterification inhibitors, namely, Ultranox 624 and

I

1.2. I

50

100

150

200

250

Figure 35. DSC thermograms of PAr/PC/PET (50/20/30) blend without stabiliser: (1) second step after extrusion at 280°C; (2) after annealing at 190°C for 1 h; (3) after 10min at 280°C [40]

259

Inhibition of Transreactions in Condensation Polymers Table 9. Summary of concentrate preparation in twin screw extruder Concentrate

Type of polymer

Drying conditions

A

PC PC PET

4 h, 100°C 18 h, 130°C 18 h, 130°C

B C

Compounding atmomhere Air

Nz blanket

N2

blanket

Stabaxol P-100. The stabilisers were incorporated into the blend by first preparing concentrates. The conditions of drying the resins for preparing the concentrates are summarised in Table 9. The stabilisers used were 0.5% by wt of Ultranox 624 and 0.25% by wt of Stabaxol P-100. Figure 30 shows the behaviour of the blend with concentrate A (predried at 100°C for 16h) after being held isothermally at 300°C for 30min. The retention of melting point of PET is clear from the melting endotherm. Figure 33 (blend with concentrate C (predried at 100" C for 16 h)) also shows the crystallinity of the blend. Figures 30 and 31 do not show melting endotherm indicating the occurrence of transesterification. 31P NMR shows that concentrate A (as used to prepare the blend shown in Figure 30) and concentrate C (used to prepare the blend shown in Figure 33) contain phosphonate groups, and that concentrate A (used to prepare blend in Figure 31) and concentrate B (used to prepare blend in Figure 32) do not contain phosphonate species which can effectively complex with the residual catalyst present in PET

Extrusion temp.: 28OoC

50

100

150

200

250

'c

Figure 36. DSC thermograms of PAr/PC/PET without stabiliser: second scan (-); eighth scan (---) [40]

N. R. James, S. S. Mahajan, S. Sivaram

260

Extrusion temp.: 28OoC

100

50

150

200

Figure 37. DSC thermograms of PAr/PC/PET with Ultranox 624: (-) scan; (---) eighth scan (sample A, Table 3) [40]

second

leading to the inhibition of transesterification. The effectiveness of a stabiliser system consisting of phosphites, polycarbodiimide, and hindered phenols which successfully inhibits transesterification at temperatures up to 325°C in a ternary blend of PET, PC, and PAr has been examined by thermal cycling using DSC [40]. Figures 34 and 35 show the DSC thermograms of PET, PC, and PAr and PAr/PC/PET (50/20/30) blend without stabiliser. The thermogram of the blend immediately after extrusion at 280°C clearly shows PET features (a glass transition temperature at 83"C, a cold crystallisation exotherm at 145"C, and a melt-

I

I

50

100

I

I

150

I

I

200

I

I

250

OC

Figure 38. DSC second scan thermograms of PAr/PC/PET with Ultranox 624 extruded at 280°C (-) and at 300°C (---) [40]

26 1

Inhibition of Transreactions in Condensation Polymers

I

Extrusion temp.: 300'C

50

100

150

250 OC

200

Figure 39. DSC thermograms of PAr/PC/PET with Ultranox 624 and Stabaxol P-100: (-) second scan; (---) eighth scan (sample C, Table 3) [40]

ing endotherm at 250°C). The glass transition temperatures of PAr and PC are not seen clearly because of PET'S cold crystallisation exotherm. Curve (2) of Figure 35 shows a shift in glass transition temperatures of PC and PAr (150°C and 180°C instead of 145°C and 190°C) and gives evidence of the occurrence of transesterification to some extent. The third curve, which corresponds to the thermogram after annealing at 280"C, reveals a single Tg at 125°C and proves complete transesterification. When phosphite (Ultranox 624) was incorporated, the blend was fairly stable (extrusion temperature 280"C, annealed at 300°C for 30min; T, decreased by 3°C and AH, by 2 J/g). However, when the extrusion temperature was raised to 3OO0C, a larger decrease in AH, and T, was observed. A combination of Ultranox and Stabaxol P-100 (polycarbodiimide) stabilises the blend at

50

100

150

200

250 'C 8

Figure 40. DSC second scan thermograms of PAr/PC/PET with Ultranox 624 and Staboxol P-100 extruded at 3OO0C (-) and at 325'C (---) [40]

262

N. R. James, S. S. Mahajan, S. Sivaram

Figure 41. DSC thermograms of PAr/PC/PET with Ultranox 624, Stabaxol P100 and Ethanox 330, extruded at 325OC: (-) second scan; (---) eighth scan (sample F, Table 3) [40] the extrusion temperature of 300°C. Upon annealing at 300°C for 30min, a strong decrease in both A H f and Tmwas observed. Addition of Ethanox 330 (hindered phenol) gave a blend which was stable at 300°C for 60min. Table 3 gives the details of the study on stabilisers. Cycling experiments can be used effectively to assess the efficiency of the stabiliser combinations (401. Figure 36 shows the second- and thirdcycle thermograms of the blend without stabilisers (extrusion temperature 280°C). It is clear that the third-cycle thermogram shows a decrease in Tmand A H f due to exchange reactions. Figure 37 shows the thermograms of blends stabilised with Ultranox 624. It can be seen that the secondcycle and eighth-cycle thermograms look similar, which indicates stability of the system. The comparison of the second cycle of the sample extruded at 280°C and that at 300°C shows that at 300°C a marked decrease in the cold crystallisation exotherm and the heat of fusion occurs (Figure 38). When another stabiliser, Stabaxol P-100, was also incorporated, even at the extrusion temperature of 300°C the second and the eighth cycles look the same, suggesting a stable system (Figure 39). Again, when the temperature is raised to 325°C the blend shows a decrease in stability (Figure 40). The incorporation of a third component in the stabiliser system imparts stability to the blend as shown by Figure 41. 3.4. Size-exclusion chromatography

The reaction mechanism and catalytic behaviour of the various catalysts can be understood from the information of the change in average block

Inhibition of Transreactions in Condensation Polymers

263

lengths with mixing time. In a PC/PET block copolymer the change in block lengths has been studied by size-exclusion chromatography after selective degradation of the PC blocks, leaving the PET blocks unaffected [44,45](see also Chapter 8). 4. Conclusions

Reactive processing/blending is a very promising technology for preparing new polymeric materials from existing polymers. There is an increasing interest in the understanding of the exchange reactions during reactive processing. The chemical structures and the properties of the resulting polymeric materials are controlled by the relative rate and extent of several reactions occurring during melt-blending. Several analytical techniques have been used which can detect the changes in the chemical structure after reactive processing. The activity of a catalyst, either present as a residue from polymer synthesis or purposely added before blending, may play an important role in controlling the chemical structure of the final product. Since exchange reactions affect mechanical properties, it is very crucial to control the extent of exchange reactions during melt-blending in order to achieve the desired properties. Organophosphorus compounds, such as phosphites [36,37],phosphonates [38], and phosphates [39], have been used to inhibit ester exchange reactions in the molten state. Lanthanide compounds [44,45]possess a wide range of catalytic activity toward different reactions taking place during PET/PC reactive blending. Hence the choice of the appropriate catalyst is very important to obtain the final product having the desired properties. Reactive blending of condensation polymers has the potential to produce a range of polymer structures from block to multiblock and fully random copolymers. The ability to control structures and sequences in transreactions of condensation polymers offers significant opportunities in creating new properties of well known polymers.

Acknowledgements One of the authors, Ms. Nirmala James, acknowledges the Council of Scientific and Industrial Research (CSIR), New Delhi, India, for the award of a Senior Research Fellowship.

References 1. L. A. Utracki, “Polymer Alloys and Blends: Thermodynamics and Rheologf, Hanser Publishers, New York 1990 2. J. A. Manson, L. H. Sperling, “Polymer Blends and Composites”, Plenum Press, New York 1976 3. C . Beretta, R. A. Weiss, Am. Chem. SOC.PMSE Prep. 56, 556 (1987)

264

N. R. James, S. S. Mahajan, S. Sivaram

4. R. Tannenbaum, M. Rutkowska, A. Eisenberg, J. Am. Chem. SOC.,Polym. Prepr. 27,345 (1986) 5. J. Y. Lee, P. C. Painter, M. M. Coleman, Macromolecules 21,954 (1988) 6. E. M. Pearce, T. K. Kwei, B. Y. Min, Am. Chem. SOC.PMSE Prep. 50, 16 (1984) 7. J. A. Manson, L. H. Sperling, “Polymer Blends and Composite.?, Plenum Press, New York 1976, pp. 87-93 8. M. Xanthos, M. W. Young, J. A. Biesenberger, Polym. Eng. Sci. 30, 355 (1990) 9. K. Kreisher, Plastics Technol. 35,67 (1989) 10. B. Brown, in: Reactive Extrusion: Principles and Practice, edited by M. Xanthos, Hanser Publishers, New York 1992 11. C. Tzoganakis, Adu. Polym. Technol. 9, 321 (1989) 12. M. Xanthos, S. S. Dagli, Polym. Eng. Sci. 31,929 (1991) 13. R. J. Kumpf, J. S. Wiggins, H. Pielartzik, %rids Polym. Sci. 3,132 (1995) 14. N. C. Liu, W. E. Baker, Adu. Polym. Technol. 11, 249 (1992) 15. J. R. Campbell, S. Y. Hobbs, T. J. Shea, V. H. Watkins, Polym. Eng. Sci. 30,1056 (1990) 16. V. J. Triacca, S. Ziaee, J. W. Barlow, H. Keskkhula, D. R. Paul, Polymer 32,1401 (1991) 17. EP 485834 A2 (1992), Mobay Corp., Inv.: R. J. Kumpf, D. I(. Nerger, R. Wehrmann, H. Pielartzik, Chem. Abstr. 117,172302a (1992) 18. EP 353478 A1 (1990)) Dow Chemical Co., Inv.: M. J. Mullins, E. P. Woo, Chem. Abstr. 113,7 9 2 7 0 ~(1990) 19. R. J. Kumpf, R. Archey, W. Kauthold, A. D. Meltzer, H. Pielartizk, J. Am. Chem. SOC.,Polym. Prepr. 34,580 (1993) 20. S. B. Brown, C. M. Orlando. “Reactiue Extrusion”, in: Encyclopedia of Poly-

mer Science and Engineering, 2nd Edition, edited by A. Klingsberg, T. Baldwin, John Wiley and Sons, New York 1988, vol. 14, p. 169 21. A. M. Kotliar, J. Polym. Sci., Macmmol. Rev. 16,367 (1981) 22. P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca

1953 23. P. J. Flory, J. Am. Chem. SOC.,62,1057 (1940) 24. P. J. Flory, J. Am. Chem. SOC.,64, 2205 (1942) 25. R. Yamadera, M. Murano, J. Polym. Sci., 5A-1,2259 (1967) 26. G. della Fortuna, E. Oberrauch, T. Salvatori, E. Sorta, M. Bruzzone, Polymer 18,269 (1977) 27. US 4417032 (1983), Allied Co., Inv.: Y.P Khanna, E. A. Turi, S. M. Aharoni, T.Largman, Chem. Abstr. 100,52525c (1984) 28. US 4861838 (1989), Allied Signal Inc., Inv.: Y. P. Khanna, Chem. Abstr. 112,56977x (1990) 29. A. Verma, B. L. Deopura, A. K. Sengupta, J. Appl. Polym. Sci. 31, 747 (1986) 30. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 31. M.Kimura, G. Salee, R. S. Porter, J. Appl. Polym. Sci. 29,1629 (1984) 32. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Part A 24,3301 (1986)

Inhibition of Transreactions in Condensation Polymers

265

33. K. P. McAlea, J. M. Schultz, K. H. Gardner, G. D. Wignall, Polymer 27, 1581 (1986) 34. I. Mondragon, J. Nazabal, J. Appl. Polym. Sci. 32, 6191 (1986) 35. A. Golovoy, M. F. Cheung, H. van Oene, Polym. Eng. Sci. 27, 1642 (1987) 36. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22, 229 (1982) 37. US 5646208 (1997), Amoco Corp., Inv.: W. W. Cattron, R. J. Schiavone 38. W. A. Smith, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 26,4233 (1981) 39. FR 2567137 (1986), Rhone-Poulenc Specialities Chimiques, Inv.: Y. Bonin, M. Logeat, Chem. Abstr. 105,98561t (1986) 40. M. F. Cheung, A. Golovoy, R. 0. Carter 111, H. van Oene, Znd. Eng. Chem. Res. 28, 476 (1989) 41. Neth. Appl. NL 8602460 (1988), General Electric Co., Inv.: J. J. Verhoeven, Chem. Abstr. 109, 74627a (1988) 42. EP Appl. 295730 A1 (1988), General Electric Co., Inv.: J. J. Verhoeven, W. M. M. Rovers, Chem. Abstr. 110, 174475~(1989) 43. DE 2751969 A 1 (1978), Ciba-Geigy, Inv.: J. Habermeier; Chem. Abstr. 89, 1114029 (1978) 44. M. Fiorini, C. Berti, V. Ignatov, M. Toselli, F. Pilati, J . Appl. Polym. Sci. 55, 1157 (1995) 45. M. Fiorini, F. Pilati, C. Berti, M. Toselli, V. Ignatov, Polymer 38,413 (1997) 46. S. B. Hait, S. Sivaram, Macromol. Chem. Phys., in press 47. L. H. Wang, 2. Huang, T. Hong, R. S. Porter, J. Macromol. Sci-Phys. B29, 155 (1990) 48. J. Devaux, P. Godard, J. P. Mercier, R. Touillaux, J. M. Dereppe, J. Polym. Sci., Polym. Phys. Ed. 20, 1881 (1982) 49. 2. H. Huang, L. H. Wang, Makromol. Chem., Rapid Commun. 7, 255 (1986) 50. J . I. Eguiazabal, G. Ucar, M. Cortazar, J. J. Iruin, Polymer27, 2013 (1986) 51. C. Berti, V. Bonora, F. Pilati, M. Fiorini, Makromol. Chem. 193, 1665 (1992) 52. M. Valero, J. J. Iruin, E. Espinosa, M. J. Fernandez Berridi, Polym. Commun. 31, 127 (1990) 53. K. R. Carduner, R. 0. Carter 111, M. F. Cheung, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40, 963 (1990) 54. M. F. Cheung, K. R. Carduner, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40, 977 (1990) 55. A. Golovoy, M. F. Cheung, K. R. Carduner, M. J. Rokosz, Polym. Eng. Sci. 29, 1226 (1989) 56. J. I. Eguiazabal, M. E. Calahorra, M. M. Cortazar, J. J. Iruin, Polym. Eng. Sci. 24, 608 (1984) 57. R. S . Porter, J. M. Jonza, M. Kimura, C. R. Desper, E. R. George, Polym. Eng. Sci. 29, 55 (1989) 58. M. Kimura, R. S. Porter, G. Salee, J. Polym. Sci., Polym. Phys. Ed. 21, 367 (1983) 59. J. S. Kollodge, R. S. Porter, Polymer 34, 4990 (1993)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 7

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends: Effect on Molecular Structure, Semicrystalline Morphology, and Thermal Properties

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

1. General introduction

It is known that polycondensates, such as polyesters, polycarbonates, and polyamides, are chemically very reactive during melt processing. Polycondensates give rise to transreaction processes when they are brought into the molten state [1,2]. These transreaction processes, also called exchange reactions, are transesterification reactions in the case of polyesters and polycarbonates, and transamidation reactions in the case of polyamides. Transesterification or transamidat ion can occur during melt-mixing of a homopolycondensate or polycondensate blends. Transreaction processes in homopolycondensates give rise to a different molecular weight distribution [3-51, whereas interchange reactions occurring in polycondensate blends generate copolymers, consisting of sequences of the blend components. Using these transreaction processes, it is possible to design copolymers with different levels of randomness and composition during melt-blending. The level of randomness is a function of the blend composition, the mixing temperature, and the residence time in the melt [6].

268

K. L. L. Eersels, A. M. Aerdts, G . Groeninckx

Transreactions in polycondensate blends strongly depend on their initial degree of miscibility. In miscible polycondensate blends, transreactions occur to a large extent because of the intimate contact between the components on a molecular level. By controlling the interchange reactions in these miscible blend systems during the melt-mixing process, it is possible to design tailored block or random copolymers with useful properties. In immiscible polycondensate blends, the control of the interchange reactions occurring at the interface between the phases allows the preparation of blend systems with a finer phase dispersion than in the corresponding physical mixtures; this is due to self-compatibilisation [7,8]. The control of the interchange reactions in molten polycondensate blends may provide a new route to the preparation of copolymers directly during melt processing. Transreaction processes in polyester blends have already been extensively studied, often with respect to self-compatibilisation [9-111. However, very little fundamental information is available in the literature about exchange reactions in melt-mixed polyamide/polyamide blends. The present chapter is mainly devoted to the investigation of the reactive melt-blending process of polyamides, in particular the aliphatic polyamide 46 (PA 46) with the aromatic polyamide 61 (PA 61), and its effect on the molecular structure, the semicrystalline morphology, and the thermal properties of the resulting copolyamides. PA 46 (DSM product StanylB, a condensation product of 1,4diaminobutane and adipic acid), is a semicrystalline aliphatic polyamide; it has high amide content and high chain regularity [12,13].Because of its considerable strength and high melting temperature, a variety of interesting applications are envisaged for this polyamide in the field of engineering plastics and industrial yarns [14,15]. Other advantages are the high heat distortion temperature ( H D T = 160°C), high stiffness at elevated temperatures, low creep and good fatigue resistance, due to the high degree of crystallinity ( w 70%) and the small size of the spherulites [16]. A major disadvantage of PA 46, however, is its high moisture uptake. PA 46 is very well suited for use in the automotive industry, especially for “under the hood” applications, and in the electrical and electronic industries, where it can be used for devices such as switches and printed circuit boards. PA 61 (Bayer product Durethan T40@,a condensation product of 1,6diaminohexane and isophthalic acid), is an amorphous aromatic polyamide characterised by a high glass transition temperature (2’ = 13OoC),transparency and reduced moisture uptake [17]. PA 61 however suffers from a low stiffness above T,. Applications of PA 61 are mainly in parts where transparency is important, such as packaging materials, glass frames, and windows. The first objective of this study is to identify the transamidation reactions occurring as a result of melt-mixing PA 46 and PA 61, and to examine in detail the effect of these transreactions on the chain microstructure of

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

269

the resulting copolyamides. Subsequently, the influence of the molecular structure of the copolyamides formed on their thermal behaviour (crystallisation and melting) is evaluated, as well as the morphology-thermal property relationships. The relationships between the various topics studied are represented schematically in Figure 1. Molecular structure of the resulting copolyamides

T

Reactive melt processing of polyamide blends

+Semicrystalline morphology +

1

Thermal properties of the copolyamides

Figure 1. Schematic representation of the relationships studied in this chapter

Molecular characterisation and semicrystalline morphology studies reveal very important links between the reactive melt processing of polyamide blends and the properties of the resulting copolyamides. However, the molecular characterisation and especially the analysis of the semicrystalline morphology for this type of blend are often omitted in the literature, despite their crucial importance for the understanding of the final properties. The main goal of this chapter is to contribute to a better fundamental understanding of the “reactive melt processing - molecular structure - semicrystalline morphology - thermal propertied’ relationships of polyamide blends. 2. Influence of the processing conditions on the thermal behaviour of PA 46/PA 61 blends

2.1. Introduction

Transreaction processes in polycondensate blends strongly depend on their initial state of miscibility and on the blending conditions. These include blending temperature, intensity and duration of mixing, and the presence of catalyst or inhibitor. The number of interchange reactions is rather limited for immiscible blend systems because of the small reaction volume at the interface between the separate phases. In miscible blends, however, the total number of reactions is much larger because of the intimate contact between the polymer chains. In this chapter, polyamide 46 and polyamide 61 are used to study the transreaction processes during melt-blending. Both polyamides exhibit miscibility over the entire blend composition range. The influence of transreactions occurring in this blend system on the crystallisation and melting behaviour of PA 46 is described. Several blending parameters, such as the extrusion temperature, extrusion time, and screw rotational speed, were varied. The obtained blends were subjected to ther-

270

K. L. L. Eersels, A. M. Aerdts, G.Groeninckx

ma1 analysis, aiming at the establishment of the effect of the melt-blending conditions on their thermal behaviour.

2.1.1. h n s a m i d a t i o n reactions Very little is published in the open literature about transamidation processes in polyamides, or about the mechanisms of these reactions. It is known that three possible reactions can occur between polyamide chains at elevated temperatures: aminolysis, acidolysis and amidolysis [1,7] , and they are similar to the transesterification reactions. However, the functional groups are now amine, carboxyl, and amide groups. The possible transamidation reactions are given in Figure2. . One of the methods of examining the kinetics of amide interchange reactions is by using sealed tubes, where transamidation takes place. The reaction products are subsequently studied by analytical methods, such as gas chromatography, to determine the rate of transreactions. Beste and Houtz investigated the influence of the presence of water and acid on the rate of transamidation using this method [18].They found that transamidation reactions are favoured by water and hydrogen ions and that the rate is proportional to the square root of the carboxyl group concentration. Accordingly, they proposed hydrolysis of the polyamides, followed by crossamidat ion, as a possible mechanism of transamidation. According to Miller, amide exchange was found to involve amide acidolysis and amide aminolysis with no detectable contribution of direct reaction between the amide groups [19]. Several articles on transamidation processes were published by Korshak and Frunze in the early 1960s [20-221. It has been shown that the addition of a salt of the other blend component, as well as heating two homopolyamides in the presence of catalytic amounts of the monomeric salt, accelerates degradation of the macromolecules and facilitates the formation of block copolymers. The information at present available does not appear to be comprehensive enough to demonstrate unambiguously the mechanisms of transamidation. Also, no studies we know of in the open literature that propose

-Ri-CO-NH-Ri-RrCOOH

Acidolysis

-R1-CO-NH-R2-R1-COOH

Figure 2. Possible transreaction processes in polyamides

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

271

catalysts for the transamidation reactions. In this respect, only water and acid are effective. However, some other catalysts, such as phosphite compounds, are claimed in the patent literature [23].

2.1.2. Effects of transreaction processes in blends of polglcondensates The occurrence of transreaction processes in blends of polycondensates often has a pronounced influence on the chain microstructure, and hence on the properties of the end products. As a result of transreactions taking place between polycondensates during the melt-mixing process, block copolymers, and eventually random copolymers are formed. The molecular structure of the block copolymers generated alters continuously, as long as the blend stays in the molten state. In fact, the block copolymers, as formed at the beginning of the melt-mixing process, will finally transform into random copolymers after long mixing times. This evolution of the molecular structure is represented schematically in Figure 3. The extent to which randomisation occurs depends on factors such as the chemical structure of the blend components, the blend composition, the reaction conditions and the presence of catalytic species (see also Chapter 8). In the case of miscible blends, generation of copolymers will occur to a large extent because of the intimate contact between all polymer chains within the miscible phase. However, in immiscible blends, the copolymers will be formed at the interface between the phases. This will enhance selfcompatibilisation because the block copolymer molecules formed in situ

Homopolymers

1

Two-block copolymers

Segmented block copolymers

Figure 3. Schematic representation of the molecular changes resulting from transreaction processes

272

K.L. L.Eersels, A. M.Aerdts, G.Groeninckx

at the interface are composed of sequences of the two immiscible blend components. This will cause a decrease of the interfacial tension between the phases of the blend, and as a result a finer phase dispersion will be obtained. Moreover, since the transreacted chains facilitate compatibilisation, this will also accelerate the reaction of the non-reacted chains, thus enhancing compatibilisation to an even larger extent. The achievement of improved compatibility between the blend components using transreaction processes has already been studied extensively for polyester, polycarbonate [24-261 and polyamide [7,8,27] blends. Hereby, it is often observed that the two separate glass transition temperatures of the immiscible blend components merge systematically into one glass transition temperature with increasing reaction time. This clearly illustrates the impact of transreaction processes, proceeding from complete incompatibility via partial compatibility to complete compatibility (see also Chapter 10). The decreasing particle size as a function of increasing reaction time, and thus as a function of the transreaction level, can be visualised in some cases by scanning electron microscopy (SEM) [9,28]. The transformation of homopolymers into random copolymers via block copolymers can also be illustrated by the change in the solubility of the blend components as a function of the reaction time [29,30]. The reduced solubility observed with melt-mixed blends when dissolving these blends in selective solvents can be ascribed to the occurrence of transreaction processes between the blend components, leading to the formation of less soluble block copolymers. However, the solubility of the blend components increases again after long reaction times, due to the short sequence lengths of highly converted copolymers with better solubility [31,32]. Interchange reactions can also change considerably the crystallisation and melting behaviour of crystallisable blend components. It is known that interchange reactions between the blend components will reduce the ability of crystallisable blend components to crystallise [9,26,29,33].The decreasing ability to crystallise can be ascribed to the shortening of the sequence length of the crystallisable components in the copolymer formed. Other causes could be the formation of graft or crosslinked copolymers [34,35]. This often leads to a lower degree of crystallinity and to a lower crystallisation rate. The occurrence of transreactions can also give rise to a melting point depression of the crystallised component [36,37]. This phenomenon is often used to follow the extent of transreactions in the melt; the continuous depression of the observed melting point is a measure of the extent of the transreaction processes [24,29,35,38]. The melting enthalpy also decreases as a function of the reaction time, indicating a lower crystallinity of the crystallised component in the transreacted blend systems [7,34,39,40]. It is important to mention that only minor changes in the melting temperature and the melting enthalpy are observed when the blends are prepared by solution casting [7,40,41]. This implies that the large changes in crys-

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

2 73

tallisation and melting behaviour are undoubtedly caused by transreaction processes, and only to a minor extent by physical interactions between the blend components. Transreaction processes are also able to increase copolymer segment lengths by chemical rearrangements. Several papers in this field were published by Lenz; it has been found that certain copolymers capable of undergoing reversible reorganisation reactions can be converted from a random to a block structure when one of the repeat units can be incorporated into the existing crystalline phase and the other cannot [4244]. This is achieved by annealing the copolymer at high temperatures below its melting point; it is called crystallisation-induced transreactivity. Units in the amorphous phase are assumed to be more reactive than those in the crystalline phase. The copolymers which have undergone cryst allisation-induced rearrangement exhibit altered crystallinity and melting behaviour. Both melting enthalpy and melting temperature are increased] as a result of the increased chain segment length of the crystallisable component [45,46] (see also Chapter 8).

Fakirov et al. studied the transreaction processes in microfibrillar reinforced composites [47]. Transreactions beween condensation polymers in the melt as well as in the bulk solid state take place at the interfaces] resulting in the formation of a copolymeric interphase. The latter plays the role of a compatibiliser and a synergetic effect in mechanical properties is observed (see also Chapter 8). 2.2. Coprecipitation versus melt-mixing

Melt-extrusion of polymers must always be carried out with the necessary caution, especially when reactive polymers are melt processed. A wide variety of possible reactions during the extrusion can completely alter the final properties of the polymer material, as illustrated in this section, where the thermal properties of polyamide blends obtained by coprecipitation or melt-mixing are compared. Table 1 represents the crystallisation behaviour of pure PA 46 and blends with PA 61 prepared by both coprecipitation and melt extrusion. Pure PA 46 crystallises at 262°C by cooling from the melt at 10"C/min. The PA 46/PA 61 (50/50 by wt) blend prepared by coprecipitation crystallises at 260"C, 2°C lower than pure PA 46. However, the melt-processed polyamide blend crystallises at 243"C, ie., almost 20°C below the crystallisation temperature of pure PA 46. Obviously, the crystallisation rate of PA 46 is affected to a large extent by melt-mixing with PA 61. The decrease of the crystallisation rate of PA 46 in the melt-mixed blends can also be illustrated by quench-cooling experiments from the melt. The solution-prepared blend shows a broad Tg region located between the T, values of the homopolymers, from which miscibility can be concluded (Figure 4). It can also be noticed that the crystallisation rate of PA 46 in

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

274

Table 1. Crystallisation and melting behaviour of PA 46/PA 61 blends; scanning rate 10"C/min Tc ("C) 262

PA 46 PA 46/PA 61 (50150 by wt) solution-prepared PA 46/PA 61 (50/50 by wt) melt-mixed

Tm

("C)

293 290 276

260

243

the pure state and in the precipitated blend is even high enough to allow complete crystallisation during the melt-quenching process. The quenched blend obtained by melt-mixing also reveals a Tgbetween those of the two homopolyamides, indicating miscibility. This is in agreement with the mean field interaction model of Ellis. The expression for &lend for a mixture of PA 46 (Az&-=) and PA 61 (AzCT-z) is given by Eq.(1) 1481. The subscripts z and y refer to the respective mer contribution to the molar volume of the polymer (z= 0.7255, y = 0.4702).

a

b

d

30

60

70

90

110

130

Temperature ("C)

160

170

190

Figure 4. Representative thermograms of the homopolyamides and blends of PA 46/PA 61 (50/50 by wt) after quench-cooling from the melt: (a) PA 46; (b) solution prepared blend; ( c ) melt-mixed blend, first heating scan; (d) melt-mixed blend, second heating scan; (e) PA 61

Reactive Melt, Processing of Aliphatic/Aromatic Polyamide Blends

275

With XAB= 7.984, X A I=~ 1.584 and X B C ~=I 2.288; Xblend is found to be equal to -0.0126. Above T,, however, a cold-crystallisation process takes place. Apparently, the crystallisation rate of PA 46 in the melt-mixed blend has decreased t o such an extent that complete crystallisation during melt quenching is no longer possible. The decreased crystallisation rate of PA 46 in the melt-mixed blend results from the incorporation of non-crystallisable PA 61 between the crystallisable PA 46 sequences due to transamidation reactions. Further crystallisation of PA 46 in this blend can only occur when the crystallisable PA 46 chains gain enough mobility at elevated temperatures, ie., above T, of the blend. The same trend in thermal behaviour is observed in Tablel, where the crystallised blends are being melted after slow cooling from the melt (lO"C/min). Pure PA 46 melts at 293°C (peak melting temperature), the blend obtained from coprecipitation at 290aC, and the melt-mixed blend at 276°C. The pronounced decrease in melting temperature for the latter blend is much larger than one would expect for a purely physical blend; this can be explained by the formation of block copolyamides during the reactive blending process. 2.3. Influence of processing conditions

The extent of the transamidation reactions is dominated by the processing conditions. This is reflected by a change in the thermal behaviour (crystallisation and melting) of the polyamide blends as a function of the varying mixing parameter(s), namely the screw rotational speed, extrusion temperature, and time. Their influence on the thermal behaviour of the blends is examined separately by changing only one parameter while the other two are kept constant.

Influence of the screw rotational speed. The influence of this processing parameter is determined by varying the screw rotational speed of the twin-screw mini-extruder from 50 to 200 rpm, the extrusion temperature (300°C) and time (2min) being kept constant. When the blends processed at different screw rotational speeds are compared to each other, no significant differences can be detected in the thermal behaviour; crystallisation at about 253°C and melting at about 287°C are observed in all cases. It is assumed that once the polyamides have been mixed on a molecular level, the progress of the transamidation processes remains unaffected by more intensive mixing. This can be understood on the basis of the miscibility of the blend system studied; more intensive mixing does not change the reaction volume in the miscible blend where transamidation takes place, and thus the rate of the transreaction processes remains constant. For this reason, the screw rotational speed is set at 100rpm in the experiments concerning the effects of processing temperature and time. In the case of

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

276

Table 2. Influence of the extrusion time on the crystallisation and melting behaviour of PA 46/PA 61 blends (50/50 by wt); scanning rate 10"C/min Extrusion time (min)

Tc ("C)

Tm ("C)

0.5 1 3 5

251 250 247 243

285 283 280 276

immiscible polyamide blend systems, it is expected that the extent of the transreaction processes will be affected by the intensity of mixing, as the surface area between the phases can be changed by varying this processing parameter.

Influence of the extrusion time. Blends are extruded at a constant extrusion temperature (325°C) for various extrusion times (0.5-5 min). Longer extrusion times are avoided to prevent excessive thermal degradation. In contrast to the screw rotational speed, the extrusion time affects the thermal behaviour of the blends obtained. Longer extrusion times result in lower crystallisation temperatures of the crystallisable component (Table 2). The crystallised blends exhibit lower melting temperatures with increasing extrusion time, while the melting enthalpy slightly decreases. Conclusions from these data concerning the volume crystallinity should be drawn with caution, because of contributions from crystal defects [49]. An explanation for the decrease in crystallisation and melting temperature can be found, by taking into account the transamidation reactions. Block copolyamides with rather long sequences of the crystallisable PA 46 are formed at the beginning of transamidation. However, at longer extrusion times the crystallisable PA 46 sequences shorten, and consequently crystallisation becomes more difficult. This process is discussed in detail in Sections 4 and 5. Table 3. Influence of the extrusion temperature on the crystallisationand melting behaviour of PA 46/PA 61 blends (50/50 by wt); scanning rate 10"C/min Extrusion temperature ("C)

' T ("C)

Tm ("C)

295 300 305 315 325

253 250 247 244 243

286 283 280 278 276

2 77

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

Influence of the extrusion temperature. Blends are extruded for a constant extrusion time (5 min) but at varying extrusion temperatures (between 295°C and 325°C). The influence of the extrusion temperature on the thermal behaviour is very similar to that of the extrusion time; Table 3 indicates that higher extrusion temperatures lead to lower crystallisation and melting temperatures. Figure 5 summarises the effect of the melt-blending time on the peak crystallisation temperature of PA 46/PA 61 blends (50/50) at different extrusion temperatures. It is clear that the transamidation processes are not only affected by the extrusion time, but they are also enhanced by higher extrusion temperatures. The variation of these two extrusion parameters alters the extent of the transamidation processes in the same way. It is possible to reach a certain degree of transamidation either by extruding the blend at high temperatures for short times or by extruding at lower temperatures for longer times. It can be concluded that there exists a time-temperature equivalence effect in respect to the occurrence of the transamidation processes. Similar results were found in 265

u 260 e h

E

g

Y

f

255

Y

."8

4 Y

250

6

fa

245

240

0

1

2

3

4

5

6

7

8

Melt residence time (min)

9

10

11

Figure 5. Influence of the extrusion conditions on the peak crystallisation temperature of PA 46/PA 61 blends (50/50 by wt) during slow cooling from the melt (10"C/min). Extrusion temperature: o 295°C; + 30OoC; 0 305°C; 0 315°C; A 325°C

278

K.L. L. Eersels, A. M. Aerdts, G . Groeninckx

the past regarding transesterification reactions in polyester blends. It appears that transesterification is also controlled mainly by the temperature and time of the blending operation (501.

3. Influence of the blend composition on the thermal behaviour of PA 46/PA 61 blends The main objective of the previous section was to demonstrate the influence of the mixing temperature and the time on the melting and crystallisation behaviour of PA 46/PA 61 blends at a constant blend composition. All the experiments were performed on a laboratory scale by means of a minimixer/extruder. In order to determine to what extent the previous results are reproducible, the data described in this section are obtained with blends prepared in a large-scale double-screw extruder. This is a very important step, regarding the applicability of the transamidation reaction concept for the preparation of block copolyamides during melt-extrusion. In contrast to the previous section, the extrusion conditions remain unchanged, but the blend composition is altered. After melt-extrusion, injection-moulded Samples are prepared from the compounded blends. All blend compositions are systematically studied by thermal analysis, and comparison is made with blend compositions prepared by coprecipitation from a common solvent.

3.1. Crystallisation and melting behaviour From the previous section on PA 46/PA 61 blends prepared by means of a mini-extruder, it becomes clear that the thermal behaviour of the crystallisable PA 46 component is influenced by transreactions with the amorphous aromatic PA 61 component during melt-mixing of the blend. The extent of the transreaction processes between PA 46 and PA 61 and, as a consequence, the average crystallisable segment length of PA 46 in the copolymers formed, can be controlled by the extrusion temperature and the extrusion time. Under constant extrusion conditions, it turns out that the initial blend composition strongly influences the crystallisation and melting behaviour of PA 46 in the blends. Figure6 shows the decrease in crystallisation temperature for the various blend compositions resulting from the different modes of preparation. The crystallisation peak temperatures suggest that transreaction processes occur also when the blends are compounded using a large-scale ZSK doublescrew extruder (residence time: 2.5 min). The crystallisation peak temperature strongly depends on the blend composition. The blend with 30 wt% PA 46 crystallises at about 240°C when the blend is cooled from the melt at lO"C/min. This is more than 20°C below the crystallisation peak temperature of pure PA 46. It can also be seen that the blend with only 15wt%

2 79

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends 2 70

260

250

--.

200

0

n

A

0

-

_.",.____.~

,

__

220

210

A P

' ---

W

~

P

A

_ _ I -

P

240

230

_..-...-A

-

-_.-_..-...-_.I_.

a o

10

20

I

I

I

I

I

I

1

30

40

50

60

70

ao

90

100

Figure 6. Influence of the blend composition on the peak crystallisation temperature of PA 46/PA 61 blends during slow cooling from the melt (10"C/min): A solution prepared blends; 0 extruded blends; o injection moulded blends

PA 46 is still able to crystallise around 200OC;however, the observed crystallisation peak becomes very broad, indicating a low crystallisation rate.

Section 2.3 describes the extent to which the crystallisation rate of the block copolyamides formed is reduced by the decreasing length of the crystallisable PA 46 sequences. The average PA 46 sequence length is determined by the number of transreaction processes between the two chemically different polyamides, and depends on the extrusion temperature and duration. However, the number of transamidation reactions between PA 46 and PA 61 depends also on the volume fraction of the two blend components. An excess of PA 61 in the blends would result in fast conversion of homopolyamide 46 into a segmented block copolymer with rather short crystallisable sequences after a short extrusion time; this would lead to an abrupt drop in the crystallisation temperature of PA 46 in the blends. When PA 61 is present at a low concentration, a short extrusion time would affect the crystallisable PA 46 molecules only to a small extent, resulting in a slight decrease in the crystallisation temperature (i. e., crystallisation rate). The decrease in crystallisation temperature of the melt-mixed blends

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

280

0

250

0

I

10

I

20

I

I

1

I

I

70 Composition (wt % PA 46) 30

40

50

60

I

80

I

90

100

Figure 7. Influence of the blend composition on the peak melting temperature of PA 46/PA 61 blends after crystallisation from the melt (l@'C/min): A solution prepared blends; 0 extruded blends; o injection moulded blends is much more pronounced than that of the solution-prepared blends. It should be emphasised that the measurement of the dynamic crystallisation temperature of all blends always implies an additional melt residence time between the melting and the crystallisation of the blends, which cannot be avoided. During this melt residence time (5-7 min at a scanning rate of lO"C/min), transreaction processes would occur, thus lowering the crystallisation temperature of the blends. This is also the case for the blends obtained by coprecipitation, which means that these blends have also undergone some transreaction processes at the onset of crystallisation. The blends obtained by extrusion followed by injection moulding reveal a similar crystallisation behaviour but the crystallisation peak temperatures are slightly below those of the extruded blends. This can be easily understood if an additional melt residence time of 1-1.5 min during the injection moulding process is taken into account; transamidation proceeding during this time interval results in slower crystallisation kinetics of the injection-moulded blends as a consequence of the shorter crystallisable sequences of PA 46. Similar remarks can be made when the melting temperatures of the crystallised blends are compared. The melting temperature of the extruded

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

281

blend with 30wt% ) PA 46 is about 274"C, i.e., almost 20°C below the melting temperature of pure PA 46; a very broad and weak melting region is found around 260°C for the blend with 15wt% PA 46 (Figure7). It can also be seen that the melting temperatures of the injection-moulded samples are again just below those of the extruded blends but far below those of the solution-prepared blends. 4. Molecular characterisation of PA 46/PA 61 blends by means

of lSC NMR

In Sections 2 and 3, the way in which melt-mixing of chemically different polyamides results in the formation of block copolymers is described. It is clear that the block lengths of the copolyamides formed depend on the extrusion temperature, extrusion time, and initial blend composition. The relationship described between the melt-mixing conditions of the polyamide blends and the thermal properties of the resulting copolyamides is strictly qualitative of nature. No direct evidence is given regarding the conversion of homopolyamides into block copolyamides with sequence lengths depending on the melt processing conditions. It is, however, important to have a fundamental insight in the change in the molecular structure of the polyamides due to the occurrence of the transreaction processes, in order to understand the altered final properties. In this section, a quantitative approach using 13C nuclear magnetic resonance (NMR) is proposed to characterise the copolyamides resulting from the transamidation reactions during melt-mixing of PA 46/PA 61 blends. In the study of the relationship between melt processing, molecular structure, and thermal properties, it is very important to know the effect of the melt processing conditions and the blend composition on the molecular structure of the resulting copolyamides, as well as on their resulting crystallisation and melting behaviour. Exact knowledge of the average length of the homologous blocks of the aliphatic polyamide 46 built into the copolyamide would provide valuable information concerning the kinetics and crystallisation ability. The characterisation of ester-interchange reactions in polyester blends as a function of the blend composition, melt temperature, and melt residence time is discussed in the literature. This has been done for poly(hydroxy ether of bisphenol A) (PC) blends by means of Fourier transform infrared spectroscopy (FTIR) [51] and for PC/poly(ethylene terephthalate) (PET) blends, using the NMR technique [52]. Other examples of studies where NMR has been used as a method of characterisation are the interchange reactions in polyarylate (PAr)/poly(butylene terephthalate) (PBT) blends [53,54] and the sequence randomisation of wholly aromatic copolyesters [55]. Regarding polyamide blends, some attempts have been made in the literature to characterise the extent of interchange reac-

282

K . L. L. Eersels, A. M. Aerdts, G . Groeninckx

tions in mixtures of poly(m-xylene adipamide) and polyamide 6 [7] and also in mixtures of nylon-2, y polymers where z indicates the number of carbon atoms separating the nitrogen atoms in the diamine, and y the number of straight-chain carbon atoms in the dibasic acid [8]. However, these analyses by means of 'H NMR are not sufficiently convincing. The determination of the average block length and sequence length distribution of the copolyamides resulting from exchange reactions can be performed by means of high-resolution NMR. NMR has been applied in several ways to polyamides, as described in three different reviews [56581. The first review deals with the application of lH NMR, discussing the assignment of the resonance signals in the spectra of commercially available polyamides [56], and the second one deals with the application of 15N NMR for the characterisation of co- and terpolyamides [57]. As far as our study is concerned, the third review is the most interesting one and deals with lH and 13C NMR studies of polyamides in solution [58]. In the latter review, the possibilities and limitations of the sequence determination in aliphatic, mixed aliphatic-aromatic, and aromatic copolymers are discussed. According to the literature, 'H NMR has been used as a tool to identify polyamides and copolyamides but, on the basis of the investigation by Kricheldorf et d. [59], it is concluded that lH NMR is not a general tool for discriminating between aliphatic copolyamides and blends of homopolyamides. Since the pioneering work of Kricheldorf and co-workers, it is well known that the shift differences for blends of homopolyamides,alternating copolyamides, and random copolyamides occur in the 13C carbonyl region [59-67]. This is possible when the polyamides are dissolved in sulfuric acid or in fluorosulfonic acid. Thus, 13C NMR has been found to be a very powerful tool for the determination of the sequence length distribution in copolyamides. The carbonyl resonance signals of copolyamides are sensitive to changes in the sequence length distributions and can be analysed in terms of dyad sequences [63]. 15N NMR can also be used to determine sequences; however, the advantage of 'H and 13C NMR compared to 15N NMR is that the quantitative measurements are much more accurate [57]. A few reports deal with the quantification of the copolyamide structure in terms of the number-average block length and the degree of randomness. Kricheldorf et a2. have published results on a binary copolyamide obtained by copolymerisation, in terms of the average block length [63-65]. More data were published for copolyesters [68,69],whereas a complete theoretical description of the determination of the number-average block length and degree of randomness for a four-component polycondensate has been given by Devaux et al. [32,70] (see also Chapters 1 and 6). In this section, the 13C NMR spectra of a blend and a copolymer of the semicrystalline aliphatic PA 46 and the amorphous aromatic PA 61 are studied. The copolymer is formed during reaction in the molten state (melt extrusion). On the basis of the 13C NMR experiments, the dyad sequence concen-

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

283

trations in the copolyamide are determined from the 13C carbonyl resonance signals, and from these results the average copolymer composition, the number-average block length, the degree of transamidation, and the degree of randomness are computed. 4.1. Theoretical considerations 4.1.1. Qualitative interpretation of the l S C NMR spectra of PA 46 and PA 61

As already mentioned, 13C N M R is an effective tool for distinguishing between a blend and a copolymer [59]. Chemical shift differences in the carbonyl resonance signals of 1-3 ppm are reported, due to variations in the chain length [63]. More CH2-groups in the monomeric units of the polyamides give rise to a shift to lower field of the carbonyl signal. To distinguish between a 50150 blend of two homopolyamides of PA 46 and PA 61 and a 50/50 copolymer of PA 46 and PA 61, comparison is made of the two I3C N M R spectra of the reaction product of PA 46 and PA 61 obtained after 90min of melt-mixing at 315°C and of a physical mixture of the two homopolyamides, respectively [71]. In the expanded 13C N M R spectra of the carbonyl region, two new peaks are clearly observed in the extruded reaction product. It can be concluded that transamidation reactions have taken place. The four signal peaks of the carbonyl carbon resonances can be accounted for in terms of dyad sequence distributions (Figure 8). The exchange reaction taking place between PA 46 composed of two monomeric units of different nature, adipic acid (ADI) and diaminobutane (DAB), and PA 61 composed of isophthalic acid (IA) and diaminohexane (DAH), results in a four-component copolyamide. According to Devaux et al. [70], this can be described by a general formula: -[(A1B1)2 - (A1B2)y], - [(A2Bl)Z - (AzBz)w],where A1, A2, B1 and B2 represent the monomeric units ADI, IA, DAB, and DAH, respectively. The letter indicates the different chemical nature of the monomer unit (A is a dicarboxylic acid, B is a diamine) and the number indicates the original homopolymer. This means that AlBl is the homologous sequence in PA 46 and A2B2 is the homologous sequence in PA 61. Moreover, 2,y, z, and w represent the average length of the various sequences; rn and n are the mean lengths of blocks having in common the same A1 or A2 unit. The assignment of the four carbonyl carbon resonances in the copolyamide in terms of dyad sequences is given in Figure8. The dyad sequences can be measured quantitatively in the N M R spectra. The knowledge of the relative dyad concentrations together with the use of a statistical model make it possible to calculate the number-average lengths of the homologous blocks (see Section 4.1.2) and the degree of randomness (see also Chapter 3).

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

284

4.1.2. Determination of the percentage of tramamidation

The percentage of transamidation can easily be derived from the carbonyl region in the 13C NMR spectrum of the reaction product after extrusion of I

ADI-DAH

178

176

174

’

li2

’

PPM Figure 8. Expanded 62.5 MHz 13CNMR spectra showing only the carbonyl region of a mixture of PA 46 and PA 61 (50/50 by wt) melt processed at 315°C for different extrusion times as indicated on the right

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

285

the 50/50 blend of the two homopolyamides PA 46 and PA 61 (Figure8). There will be 100% transamidation when all AlBl units are changed into AlBz units and all A2B2 units into AzBl units. A completely random copolymer with a copolymer composition of 50/50 has undergone 50% of transamidation, having an equal amount of A1B1, AlBz, AzBz and AzBl units. Thus the percentage of transamidation (9) can be calculated, based on the fractions of A1B2 and A2B1 dyad concentrations according to the following formula:

Q(%) = ( F A ~+BF~A ~ B x 100 ~)

(2)

where F A ~ is B the ~ fraction of an A,Bj dyad and is defined by

The percentage of transamidation in the melt-mixed PA 46/PA 61 (50/50) blend, with the 13C NMR spectrum shown in Figure8 (90min of extrusion), is 12.16%. For other blend systems, such as PA 6/PA 66 and PA 6/PA 48, the degree of transamidation after 1 h at 295°C is 20% and 45%, respectively [63].

4.1.3. Quantitative determination of the degree of randomness and the average block length The degree of randomness and the number-average block length have been defined for different systems. The analysis of a standard two-component AB copolymer is well explained by Koenig [72]. Other studies concern the structures of three-component [68] and four-component [32,70] polycondensates. Here we deal with a four-component copolyamide system and the same statistical model can be applied, as developed by Devaux et al. for exchange reactions between two linear polycondensates (AlBI), and (A2B2), (p and q are the numbers of units in polymer 1 and 2, respectively) with degrees of polymerisation high enough for the effects of chain ends to be neglected in the calculations [70]. The following concentration relations are valid:

The degree of randomness is represented by

[AiBj] = [All [A21

(i, j = 1, 2)

286

K. L. L. Eersels, A. M. Aerdts, G. Groenindcx

When a copolymer exhibits completely random statistics, then

[AiBj] = [All [A21

therefore x = 1

(7)

Values of x > 1 indicate that the units have a greater tendency to alternate. When x < 1, the units tend to cluster in homogeneous sequences and thus the copolymer exhibits a block character. When x = 2, one deals with a completely alternating copolymer, and in the case when x = 0, the copolymer is completely diblock, or the system is a mixture of two homopolyamides. The general relation between the degree of randomness (x)and the degree of transamidation (!I?) is represented by: = 2FAlFAzX

(8)

When the copolymer is completely random, the percentage of transamidation for a 50/50 copolymer is 50%0,and for 70130 and 85/15 compositions, the percentage of transamidation is 42% and 25.5%, respectively. The number-average sequence length of AlBl groups is defined as

B~ the probability of finding an A, unit followed by where P A ~represents a Bj unit ( P A ~=B[AiBj]/[Ai]). ~ We assume that in all the triads the substitution on the left of a central unit does not influence the substitution on the right (701. In the same way, the number-average lengths of the A1B2, A2B1, and AzB2 sequences, y, z , and w, respectively, are defined as

In a random copolymer with equal molar ratio of A1 and A2 (50/50), x = 1 and the number-average length of AlBl and A2B2 sequences is x = 2 and w = 2, respectively. In this particular example the average number of AlBl units is 104 in the beginning, and after 15% of transamidation the average number of AlBl units next to each other is about 6.7.When the transamidation reaction in a polyamide system is very fast, it is very difficult to achieve long homologous block lengths. On the other hand, when the transamidation process is very slow, short block lengths (tending to

287

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

random structure) are obtained after very long mixing times, and probably other reactions, such as degradation, can be involved. The mean lengths of blocks having in common the same A1 or A2 units, i e . , blocks of the aliphatic or aromatic type (m and n, respectively), can be determined and may provide some additional information. According to Devaux et al. [70] it can be deduced that

where p is the number of AlBl units in polyamide 46 (polymer 1) and q is the number of AzBz units in polyamide 61 (polymer 2). FA^ and FA^ are the molar fractions of polymer 1 and polymer 2 in the copolymer, respectively. 4.2. Crystallisation behaviour of PA 46/PA 61 copolymers,

prepared by melt-blending, as a function of the extrusion temperature, extrusion time, and blend composition

The time dependences, by extrusion at 315"C, of crystallisation peak temperature and crystallisation enthalpy of the PA 46/PA 61 (50/50) blend 260

250 240 -230

u

I 0

0

-8

-

-

0

- 9

- 3 0n2

0

h" 220 210

-

200

-

- 40 -35 D

0

W

190

AHc (J/g)

Tc ('C)

I45

09

W

- 25

0 0

0 1

,

1

.

1

,

1

,

- 20

0 1

,

1

,

15

Figure 9. Influence of the extrusion time on the crystallisation peak temperature T, and the crystallisation enthalpy AH, of PA 46/PA 61 blends (50/50 by wt) during slow cooling from the melt (1O0C/min).Extrusion temperature 315OC

288

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

are plotted in Figure9. The crystallisation enthalpy is given in joules per gram of the blend. It can be seen that both the crystallisation peak temperature and the crystallisation enthalpy are decreasing greatly during the first 60 min of the melt-mixing process. This decrease is less pronounced at longer extrusion times. The influence of the extrusion temperature on the crystallisation peak temperature and the crystallisation enthalpy of the PA 46/PA 61 (50/50) blend is represented in Figure 10; the extrusion time is set at 60 min. It is clear that the crystallisation rate and the crystallisation enthalpy strongly decrease with increasing extrusion temperature in the 295-325°C range. From the above data, it can be concluded that crystallisation of PA 46 in the blends becomes more difficult when the polyamide blends are melt-mixed for long times and at high temperatures, as already discussed in Section 2 [73]. However, the blends composed of 50% PA 46 and 50% PA 61 are still able to crystallise, even after very long melt-mixing times at high temperatures. The ability to crystallise can also be affected by changing the blend composition under constant melt-mixing conditions "741. The crystallisation of PA 46 becomes more difficult at higher PA 61 content in the blends. " 1

Tc ("C) A&

0

225

-

(J/d :32 0

0

- 30

0

-28

- 09

V

h

k 220 -

-26

Y

-

- 24

0

215

0

-

- 22 0

210

0

'

l

'

"

l

'

l

'

l

'

\

0 v 9

l

'

l

'

- 20 18

Figure 10. Influence of the extrusion temperature on the crystallisation peak temperature Tc and the crystallisation enthalpy AHc of PA 46/PA 61 blends (50/50 by wt) during slow cooling from the melt (1O0C/min). Extrusion time 60 min

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

289

Crystallisation is inhibited in blends consisting of 70wt% or more PA 61; this is also confirmed by their transparency at room temperature.

PA 46/PA 61 copolymers, prepared by melt-blending, as a function of the extrusion temperature, eztrusion time, and blend composition

4.3. Molecular structure of

Influence of the extrusion time. The dependence of the 13C NMR spectra of the polyamide blend PA 46/PA 61 (50/50 by wt), melt-processed at 315"C, on the extrusion time has been studied. The expanded 13C NMR spectra of the carbonyl region of the different copolyamides formed are represented in Figure 8, where the assignments of the carbonyl resonance signals of the PA 46/PA 61 copolyamide system are also indicated. It is seen that the peak intensities of the different dyad sequences change, depending on the extrusion time. The peak intensities at 171.71ppm and at 177.62 ppm, ie., the concentrations of dyad sequences AzBl and A1B2, respectively, increase with the extrusion time, suggesting a growing number of transamidation reactions. As a consequence, the crystallisable PA 46 sequences and the amorphous PA 61 sequences shorten

ADI-DAB IA-DAH ADI-DAH IA-DAB

30

57 0 0

20

40

60 ao Extrusion time (min)

100

120

140

Figure 11. Number-average block length of the various sequences of PA 46/PA 61 copolymers as a function of the extrusion time at 315°C.Blend composition PA 46/PA 61: 50/50 by wt

K. L. L. Eersels, A. M. Aerdts, G . Groeninckx

290

with increasing melt-mixing time. In Figure 11 the number-average block length of the various sequences is plotted against the melt-mixing time. It can be seen that the average block length of PA 46, ie., homologous groups consisting of AlBl next to each other, amounts to seven after melt-mixing at 315°C for 120min and is clearly still sufficient to allow crystallisation (see Section 4.2 and Chapter 8 ) . The length of the amorphous PA 61 blocks is decreasing in a similar way. Figure 12 represents the dependence of the degree of transamidation and the degree of randomness on the extrusion time at 315°C. The highest percentage of transamidation obtained by melt-mixing at 315°C for 120min is 14%, which means that 14% of the polyamide bonds of the two blend components PA 46 and PA 61 (AIB1 and A2B2) are converted into new polyamide bonds, forming AlBz and A2B1. This corresponds to a degree of randomness of 0.27. A fully random copolymer would have a degree of transamidation of 50% and the value of x would be 1. An alternating copolyamide would have a degree of transamidation of 100% and x would be 2. It is clear from the above data, and especially from the x-values, that the melt-generated copolymers still have a block char-

Transamidation --t 80

h

11

Oa8

t

- 0.6 8

530-

B

3i

Y

- 1

.r(

- 0.4

1 4 0 1

G

- 1

hdomness

-

Extrusion time (min)

Figure 12. Degree of transamidation and degree of randomness of PA 46/PA 61 copolymers as a function of the extrusion time at 315°C. Blend composition PA 46/PA 61: 50150 by wt

Reaxtive Melt Processing of Aliphatic/Aromatic Polyamide Blends

291 100

ADI-DAB --c

IA-DAH -0-

ADI-DAH

+

IA-DAB

-+

- 80 - 60

0

20

40

60

Blend composition (wt % PA 61)

80

100

Figure 13. Number-average block length of the various sequences as a function of the PA 46/PA 61 blend cornposition. Extrusion temperature 315OC; extrusion time 90 min acter, which explains their ability to crystallise. From these results, it is also obvious that transamidation in the melt does not provide a practical method for preparing high molecular weight random copolyamides by meltblending high molecular weight homopolyamides. A similar conclusion was drawn by Beste and Houtz in their study of amide interchange reactions between sebacamide and N , N'-diacetylhexamethylenediamine [18].

Influence of the extrusion temperature. Melt-mixing of PA 46 and PA 61 for a constant period but at varying extrusion temperatures gives rise t o a different chain microstructure of the resulting copolymers. The influence of extrusion time is similar: the block length of PA 46 in the copolymer formed decreases as extrusion temperature increases. This explains the decreasing crystallisation temperature during a controlled DSC cooling experiment with increasing extrusion temperature, as shown in Figure10; the greater difficulty of crystallisation is also reflected by a lowering of the crystallisation enthalpy. Increasing the melt-mixing temperature raises the degree of transamidation from about 10% at 295°C to about 14.5% when the blends are extruded at 325°C [6];the degree of random-

K . L. L. Eersels, A. M. Aerdts, G . Groeninckv

292

ness is increased from 0.19 to 0.28. The temperaturetime dependence of the degree of transamidation is illustrated by the following example: one could reach a block length of e.g., seven AlBl units of PA 46 in the copolymer, which corresponds to a block length of 102.981, either by melt-mixing PA 46 and PA 61 (50150 by wt) for 120min at 315°C or by melt-mixing the same blend for 60min at 325°C. In both methods, the degree of transamidation will be 14.5% and the degree of randomness 0.27-0.28. However, the copolymers obtained are nowhere near random.

Influence of the blend composition. The sequence length distribution of PA 46 and PA 61 in copolyamides, obtained by reactive extrusion, changes dramatically if the overall starting blend composition is changed. Figure 13 shows the resulting block lengths of PA 46 and PA 61 after meltmixing the blends at 315°C for 90min. The homologous blocks of the PA 46 sequences AlBl decrease from 30.5 for the blend composition PA 46/PA 61 85/15 by wt to 2 for the blend composition PA 46/PA 61 15/85 by wt. The PA 46 block length after melt extrusion of 30% PA 46 and 70% PA 61

-

3 Randomness 'Ihmsamidation

t 8ot

loo

n

E

20

I-

i

-

0

20

40

60

Blend composition (wt % PA 61)

80

100

Figure 14. Degree of transamidation and degree of randomness as a function of the PA 46/PA 61 blend composition. Extrusion temperature 315"C,extrusion time 90min

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

293

amounts to 3.6, corresponding to a block length of 52.9A, which is clearly too small to preserve the crystallisability. This explains the failure to crystallise of the latter blend composition, as already observed in Section 2.2. From a statistical point of view, one would expect that the sequence length of PA 61 would decrease in a similar way when PA 46 is added. However, the block length of PA 61 drops from 36 A2B2 units to 11 when only 15% of PA 46 is added. The further decrease in the sequence length of PA 61 is rather small at higher weight contents of PA 46. The difference in shortening of the block sequences of PA 46 and PA 61 depending on the blend composition can be explained either by a difference in reactivity between PA 46 and PA 61 or by the presence of a catalyst in PA 61. However, no traces of catalyst are found in either of the blend components, not even on the ppm scale. In Figure 14, the degree of randomness and the degree of transamidation are plotted as functions of the blend composition. Here also, it can be noticed that the degree of randomness, which is by definition independent of the blend composition, increases with increasing weight content of PA 61. From a statistical point of view, it could also be expected that the degree of transamidation reaches a maximum at a 50150 PA 46/PA 61 weight ratio. Instead, the maximum shifts to a higher PA 61 content (about 70%). All these observations indicate a higher reactivity of the blends containing a large weight fraction of PA 61, but their explanation requires further investigations. 5 . Characterisation of transamidation reactions in PA 46/PA 61

blends using gradient elution chromatography

Molecular characterisation of block copolymers formed during reactive extrusion is often feasible using NMR or FTIR [32,38,75-771. However, NMR and FTIR have restrictions when applied to the characterisation of reactive blend systems with a low reaction rate. Since melt-extrusion in most cases implies a short melt residence time, only a limited number of reactions would have occurred by the end of the extrusion process. The characterisation of the reaction products by NMR or FTIR becomes even more difficult when the chemical nature of the bonds of the initial polymer and of the reacted products is very similar. This is the case for melt-mixed

-

-NH-(CHZ)~NH-CO-(CH~)~-COPA 46-co-6 I Transamdation 4-

-NH-(CHz)rNH-CO-( CH2)rCO-

+

PA 46

-NH-(CH2 )gNH-CO

PA 61

-NH-(CH2)rNH-CO

PA 46-co-6 I

Figure 15. Polyamide bonds before and after transamidation of PA 46 and PA 61

294

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

PA 46 and PA 61. The new amide bonds of the copolyamides formed as a

result of transamidation differ only slightly from the amide bonds of the homopolyamides. As can be seen in Figure 15, the only difference arising during the transreaction processes is that some butane neighbours of the amide bonds are replaced by a hexane neighbour and vice versa. The technique of gradient elution chromatography (GEC) overcomes the resolution problem of NMR or FTIR with such blend systems. GEC is a high-performance liquid chromatography method, allowing the determination of the chemical composition distribution (CCD) of copolymers [78,79]. A precipitation-redissolution mechanism, combined with a distinct contribution of adsorption, makes it possible to separate and characterise the copolymers. The GEC mechanism is represented schematically in Figure 16. Polymers with different chemical compositions have a precipitation/redissolution transition taking place at different solvent compositions. The precipitation/redissolution mechanism can be achieved by a twosolvent gradient, using a non-solvent A and a strong solvent B. Moreover, the adsorption/desorption transition of the chemically differing polymers could also occur at different solvent compositions when a polar column is used. In this way it is possible to separate polymer blends into the blend components. In this section, GEC is used to characterise PA 46/PA 61 blends, meltmixed for a short period (i.e., less than 10min). The degree of transamidation after the melt-mixing process is lower than 5%. GEC appears to be a suitable characterisation method for the study of the effect of the extrusion conditions (temperature and duration) on the degree of transamidation [80].The possible influence of the molecular weight and the end-group concentration on the transamidation reaction rate is also examined. Melt-

Polymer A Polymer B Composition A/B copolymer (%)

Figure 16. Schematic representation of the GEC mechanism

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

295

extrusion times of 15min or longer yield a degree of conversion which is high enough to allow characterisation by 13C NMR (see Section 4). This illustrates the complementary character of GEC and NMR/FTIR.

5.1. Influence of melt-blending conditions on the degree of transamidation Influence of the extrusion time. Blends composed of 90 wt% PA 46 and 10 wt% PA 61 are first prepared by precipitation from a common solvent. It is stated in Section 2 that PA 46 and PA 61 are fully miscible, regardless of the blending technique (from solution or by melt-mixing). However, the use of solution-prepared blends for the melt-mixing experiments eliminates the melt-blending time required for homogenisation. If this procedure is followed, the experimental extrusion time is the actual reaction time without mixing effects. The reason for using a PA 46/PA 61 90/10 weight ratio is that the low reaction rate at this blend composition allows the better control of the transamidation reactions. Moreover, this blend composition leads to improved detection of the blend components leaving the column, because the aromatic PA 61 is highly absorbent at 230-235nm (conjugated bonds), in contrast to the aliphatic PA 46 which is non-absorbent at 235nm and, as a consequence, is more difficult to detect. Therefore, it is useful to have a high concentration of PA 46 in the blend. Figure 17 represents the GEC chromatograms of the blends taken at 195nm as a function of the extrusion time. A GEC run is also performed with a purely physical blend without transamidation reactions (t = 0). It can be seen that PA 46 and PA 61 can be separated successfully by a difference in retention times of about 30 min. When the blend is melt-extruded, it can be noticed that the peak of PA 61 is systematically disappearing and a broad signal emerges between the retention times of pure PA 46 and pure PA 61. This broad signal can be ascribed to the formation of copolymers composed of PA 46 and PA 61. The copolymers have a retention time which is intermediate between those of PA 46 and PA 61. Figure 18 shows the corresponding chromatograms, taken at 235nm, where only PA 61 can be detected. In this way, it is even more apparent that homopolyamide 61 is converted into copolymer during melt extrusion. At first, the peak broadens on the left-hand side of the signal with increasing extrusion time, but later also shifts to lower retention times since the content of pure PA 61 is exhausted. GEC measurements can also be utilised to perform quantitative calculations on the dependence of the transamidation processes on the extrusion time. For this purpose, a third spectrum is needed which is the difference of the spectra recorded at 195nm and 235 nm. The difference between the two spectra yields a chromatogram where only PA 46 is visualised. It is possible t o correlate area and mass using PA 46 and PA 61 standards be-

296

K. L. L. Eersels, A. M. Aerdts, G . Groeninckv

cause the sensitivity of detection of neither of the blend components alters when the homopolyamides are being transformed into copolymers. This is verified experimentally. By using narrow time intervals for integration of the chromatograms (e.g., between 59 and 60min of retention time) and the correction factors for PA 46 and PA 61 obtained from the standards, the amount of a certain copolymer composition present in the analyzed sample can be calculated. mA1

195 nm

t=O

PA 46/PA 61: 90/10 I

10

I

I

30

I

I

60

I

I

70

I

Retention time (min) Figure 17. GEC chromatograms of PA 46/PA 61 (90/10 by wt) blends as a function of the extrusion time. Extrusion temperature: 315OC, wavelength of detection: 195nm

297

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

Figure 19 represents the mass distribution of PA 46 in the blend during the first 10min of the melt-mixing process. It can be seen that at the beginning of the extrusion process (extrusion time = 0), PA 46 is entirely in the form of homopolyamide (ie., 90% of the blend). However, as extrusion proceeds, PA 61 is built in PA 46 and copolymers are formed. This can be seen by the decreasing mass of homopolyamide 46 and the development of a broad shoulder next to the homopolymer location. This conversion process mAU

A

235 nm

t=O

-24

t = 4'00''

I

10

I

I

30

I

I

so Retention time (min)

PA 46/PA 61: 90110 I

1

70

I

Figure 18. GEC chromatograms of PA 46/PA 61 (90/10by wt) blends as a function of the extrusion time. Extrusion temperature 315OC,wavelength of detection 235 nm

298

K. L. L. Eersels, A. M. Aerdts, G . Groeninckx

is even more visible in Figure20, where the mass distribution of PA 61 in the blend is shown. It has to be mentioned that the mass of neat PA 61 at the beginning of the extrusion process is less than 10% of the blend composition; this can be attributed to a chromatographic peak broadening, for which no correction is applied. The peak broadening is significant only at high PA 61 contents due to the strong dependence of composition on the retention time with the gradient used. The effect could be minimised by applying a shallower gradient, but this in turn would increase the time required for the experiment. Figures17 and 18 show that the signals of the copolyamides are very broad, indicating a wide chemical composition distribution (CCD) of the copolymers. This implies that the copolymers formed during the first

Figure 19. Mass distribution of PA 46 in the PA 46/PA 61 (90/10by wt) blend as a function of the extrusion time. Extrusion temperature 315°C

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

299

10 min of melt-mixing are block copolymers rather than random ones. An almost random copolyamide yields narrower signals, as illustrated in Figure21. The almost random copolymer is obtained by melt-mixing of 30% PA 46 and 70% PA 61 for 90min at 315OC. Influence of the extrusion temperature. In Sections 2 and 4, the effect of the extrusion temperature on the transamidation reactions is described. In order to study the increase of the reaction rate as a function of the extrusion temperature during short melt-mixing times, identical blend compositions are extruded for 2 min at different extrusion temperatures, i e . , at 300°C and at 315OC, respectively. The GEC chromatograms taken at 235 nm are given in Figure 22, and a noticeable difference in the end prod-

Figure 20. Mass distribution of PA 61 in the PA 46/PA 61 (10/90 by wt) blend as a function of the extrusion time. Extrusion temperature 315°C

300

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

10

30

50

Retention time (min)

70

90

Figure 21. GEC chromatograms of neat extruded PA 46, neat extruded PA 61 and almost random PA 46/PA 61 copolyamides (30/70 by wt): (a) PA 46 (200nm); (b) PA 46/PA 61 almost random copolymers (200nm); (c) PA 46/PA 61 almost random copolymers (235nm); (d) PA 61 (235nm)

mAU 235 nm

10

30

50

Retention time (min)

70

Figure 22. GEC chromatograms of PA 46/PA 61 (90/10 by wt) blends as a function of the extrusion temperature. Extrusion time 2 min, detection wavelength 235 nm

301

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

ucts can be deduced. It is seen that the retention time of the copolymers obtained by melt-mixing at 315°C is shorter than that of the copolymers obtained by melt-mixing at 300°C.The observed difference in the copolymers' as a function of the extrusion temperature illustrates once again the high sensitivity of GEC as a molecular characterisation method. 5.2. Influence of end-groups o n the degree of transamidation

By analogy with transesterification reactions, transamidation can occur via different reaction mechanisms, e.g., reactions involving the end-groups (aminolysis and acidolysis) or direct interchange reactions (amidolysis). The dominant mechanism of transreaction in polyester blends composed of PC and PBT or PBT and PAr is reported to be direct transesterification [81,82]. However, the dominant reaction type in polyamide blends is still unknown. One of the possible ways to verify whether transamidation proceeds via end-groups resides in the use of blocking groups, making the end-groups unavailable for transreaction processes. However, as mentioned above, melt processing of polyamides causes degradation during which new functional end-groups are created. Another way to verify the influence of the end-groups on the transamidation reactions is by varying the molecular weight and/or the ratio of amine/carboxyl end-groups in one of the initial polyamides (i. e., before melt-blending), to look for differences in the resulting copolymers. The extrusion time should be kept as short as possible in order to avoid degradation causing changes in the end-group ratios during melt-mixing. In these experiments, the concentration of end-groups and the endgroup ratio of PA 46 are varied. The PA 46 characteristics are given in Table4. The melt-extrusion time is limited to 2min. No significant influence of the end-group concentration of the initial PA 46 on the resulting copolyamides after melt-mixing is detected. From the similarity in the elution behaviour of the obtained copolyamides,it can be concluded that there Table 4. Characteristics of PA 61 and PA 46 batches Polyamide PA 46 (Batch 1) (Compacted powder) PA 46 (Batch 2) (Extruded pellets) PA 46 (Batch 3) (Compacted powder) PA 61 (Extruded pellets)

MUJ

M,

Amine end-group

(g/mol) 23400

(g/mol)

(g/mol)

(meq/g)

79000

0.005

20700

51100

86400

0.018

0.050

24800

69900

120500

0.018

0.057

8800

28900

67000

0.030

0.075

M,

50400

Carboxyl end-group (meq/g) 0.040

302

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

are no indications of a difference in reactivity depending on the end-groups. On the basis of previous observations, a transamidation mechanism via hydrolysis followed by recombination (2. e., direct transamidation) is assumed. In this case, it is important to mention that moisture and acidity catalyse hydrolysis of the polyamides, which accelerates the transamidat ion reactions [18]. 6. Morphological structure of melt-processed PA 46/PA 61 blends 6.1. Semicrystalline morphology of melt-processed PA 46/PA 61 blends The thermal properties of the PA 46/PA 61 blends are related to their semicrystalline morphology [83]. Hence, it is useful to investigate both the crystalline ordering of the crystallised PA 46 component and the amorphous-crystalline supermolecular ordering. Wideangle X-ray scattering (WAXS) is a suitable technique for the investigation of the crystalline ordering which includes the study of the crystal lattice spacings. Based on the difference in electron density between two phases, transmission electron microscopy (TEM) and small-angle X-ray scattering (SAXS) make it possible to gain information on the mutual ordering of the crystalline and amorphous regions in the blends. 6.1.1. Time-resolved wide-angle X-ray d i f i c t i o n It is well known that semicrystalline polyamides can exhibit polymorphism at room temperature; both monoclinic and/or triclinic unit cells can be obtained, reflecting different chain alignments and stackings of hydrogenbonded sheets [84,85]. These structures are often indicated as a-structures. However, heating the polyamides to higher temperatures can result in a change of the crystalline unit cell; the stable a-structure is then transformed into an unstable y-structure which can be described by a pseudohexagonal unit cell [86]. The crystal unit cell changes in polyamides as a function of temperature have already been reported by several authors (87-891. The solid state transition from one type of crystal lattice into another as a function of temperature is known as the Brill transition (Tb).The nature of the Brill transition is considered to be related to a temperature dependent chain packing [86]. At high temperatures, the oscillation of the CH2 groups with respect to the chain axis might include a widening of the spacing between the hydrogen-bonded sheets [87]. Other possible reasons for the change of the unit cell as a function of the temperature are the anisotropy of the thermal expansion [go] or the development of a three-dimensional network of hydrogen bonds between the chains [91,92]. Here, the crystalline chain segments are supposed to perform rotational jumps of 60" around their long

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

4.4

303

-

4.3 4.2 h

$4.1

-

*I 1 a :3

3.8

-

3.7

-

3.6

-

3 . L

-

40

80

-

120

I -160 200

Temperature ("C)

240

280

Figure 23. Temperature dependence of the spacings,d(lm) and d(olo),(llo), of PA 46/PA 61 blends on cooling from 300°C to 50°C. Cooling rate 1O0C/min. Blend composition (PA 46/PA 61): + (85/15);o (70/30);x (30/70);V (15/85) axes, breaking the hydrogen bonds in the initial hydrogen-bonded sheet and creating new ones with the molecules in the neighbouring sheet. This mechanism is known as the 60" flip-flop motion [93]. It is obvious that the melting behaviour of polyamides should be related to the high-temperature pseudohexagonaly-structures that are present just before melting, rather than to the low-temperature a-structures. Similar comments can be made when PA 46 blends are cooled from the melt. Crystallisation will occur in the pseudohexagonaly-form, which will then transform into the more stable a-structure below the Brill temperature. It is interesting to follow the effect of melt-mixing PA 46 with an amorphous aromatic PA 61 on the crystal lattice of PA 46. The copolyamides, as formed during reactive melt-blending, are composed of crystallisable and noncrystallisable sequences. In Figure 23, the lattice spacings of several blend compositions, as obtained by melt processing, are given as a function of temperature. The blends are crystallised from the melt at a cooling rate of 10"C/min. It can be seen that all blend compositions crystallise in the pseudohexagonal form and are transformed into the more stable a-form below the Brill temperature. The Brill transition temperature during cooling decreases significantly (by 15°C) for the blend containfrom the melt (Tb,J ing 70%PA 61. It can also be seen in Figure 23 that the splitting of the d(loo) and d(olo)l(llo)reflections of the same blend is less pronounced. The blend composed of 15% PA 46 and 85%PA 61 does not show a Brill transition and

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

304

remains mainly in the pseudohexagonal form, even at lower temperatures. High concentrations of amorphous PA 61 in the blend clearly hinder the transformation of the PA 46 crystals from a pseudohexagonal form into a monoclinic/triclinic crystal structure, and a pseudohexagonal frozen-in structure is obtained at room temperature. A similar observation is reported by Zimmerman et al., who found that increasing the concentration of PA-2Me6T (based on 2-methylhexamethylenediamineand terephthalic acid) in melt-mixed blends with PA 66 inhibits the normal transition of PA 66 from the hexagonal to the triclinic form [94]. The change in the lattice spacings of the blends, slowly crystallised during heating, is represented in Figure 24. The crystal-into-crystal transformation seems to be reversible. It is important to notice that the crystalline lattice parameters of the pseudohexagonal structures just before melting are independent of the blend composition. Moreover, it can be seen in Figwe25 that the peak width at half-height of the PA 46 crystals at 240°C remains the same in all blend compositions. The independence of the lattice spacings and the peak width of the bIend composition suggests that the PA 61 units, which cannot be incorporated in the PA 46 crystals, do not affect the PA 46 crystal perfection. Nevertheless, the melting temperature of PA 46 in the blends decreases with increasing weight content of PA 61 (Section 3.1). Although no differences in lattice spacings could be observed above 200°C as a function of the blend composition, it is assumed that these

4.3 4.4

100)

3*71

-

3.6

3.511 40

I

80

'

120

I

160

'

I

200

I

I

240

I

I

280

'

Temperature ("C) Figure 24. Temperature dependence of the spacings, d(100)and d(olo),(llo), of PA 46/PA 61 blends on heating from 50°C to 300°C.Heating rate lO"C/min.Blend composition (PA 46/PA 61): + (85/15);o (70/30);x (30/70);V (15/85)

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

305

100/0 85/15

20

18

22

24

26

2 0 (degrees)

Figure 25. WAXS diffraction patterns of different PA 46/PA 61 blend compositions at 240°C

4*41

4.3

:

3.5

:

40

:

~

80

i I

120

l i

l i

160

l i

,

200

,

I

I

240

,

I

280

Temperature ("C) Figure 26. Temperature dependence of the spacings, d(100)and d(olo)l(llo), of PA 46/PA 61 (50/50) blends on cooling from 300°C to 50°C. Cooling rate lO"C/min, extrusion temperature 315"C, melt residence time: 4min (extruded plus injection-moulded); o 5 min; A 120 min

+

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

306

3*71

3.6

3.51'

40

I

80

I

120 I

I

160 I

I

200 I

Temperature ("C)

I

240 I

I

280 I

'

Figure 27. Temperature dependence of the spacings, d(loo) and d(olo)/(llo),of PA 46/PA 61 (50/50) blends on heating from 50°C to 300°C. Heating rate 10°C/min, extrusion temperature 315OC, melt residence time: + 4min (extruded plus injection-moulded);o 5 min, A 120 min crystals differ in their Tb at high temperatures and in their lattice spacings at 5OOC. However, differences in crystal structure of the various blend compositions near T,,, cannot be found at present, and it is doubtful that these differences cause the pronounced melting point depressions. Figure 26 represents the d(loo)and d(llo)/(olo)lattice spacings of a 50/50 blend, obtained after different melt residence times, as a function of temperature during cooling from the melt. The dependence of the sequence lengths of the polyamide copolymers on the melt residence time is studied in detail in Section 4. It is found that a prolonged melt-mixing time results in an increase in the degree of randomness and a decrease in the block copolymer sequence lengths, giving rise to a decrease in the crystallisation temperature of the blends. However, the melt residence time does not seem to determine the lattice parameters of PA 46 just after crystallisation. Only a small increase of the d ( l l ~ ) / ( ~ ~spacings o) can be noticed at lower temperatures with increasing melt residence time. This observation accompanies a decrease in the Brill temperature. The evolution of the lattice spacings during heating and melting of the blends is shown in Figure27. The lattice spacings just before melting are independent of the melt residence time, although the melting temperature decreases to a large extent with increasing melt residence time. Summarising, it can be concluded from the timeresolved WAXS me* surements that the blend composition and the processing conditions affect

307

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

the Brill transition temperature and the lattice spacings at room temperature but do not seem to influence the lattice parameters immediately after crystallisation or just before melting.

6.1.2. Time-resolved small-angle X-ray d i m c t i o n The background-corrected SAXS patterns of the different PA 46/PA 61 blend compositions, slowly cooled from the melt (lO"C/min), are given in Figure28. They reveal broad peaks indicating a large distribution of the long period L . The total scattered intensity at small angles markedly decreases with increasing concentration of PA 61 in the blends, and this can be the result of a combination of several factors. Assuming a twophase system, the scattered intensity is related to the invariant ( q ) 2 ,which is related t o the degree of crystallinity 4 and the difference between the electron density of the crystalline phase (pc) and the amorphous phase (Pa):

(15)

$ ) ( ~ c- p a)'

(q)2 N-

(i) The electron density of the amorphous phase pa of the blend will increase with a rise in the amount of PA 61 in the amorphous phase, since the density of PA 61 (1.18g/cm3) is slightly higher than that of amorphous PA 46 (l.10g/cm3). Furthermore, it appears from the WAXS measurements that pc of PA 46 is unaffected by the blend composition for the blends with 50wt% of PA 61 or less, and even decreases to some extent when

I

I

1

I

I

I

I

I

I

I

L

I

I

I

I

t

I

I

t

0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 0.02 3'( 1 / 4

Figure 28. Background-corrected SAXS curves of PA 46/PA 61 blends, slowly cooled from the melt (lO°C/min)

308

K. L. L. Eersels, A. M. Aerdts, G . Groeninckx

-

280 260

-

240

x

3 220-

4 200-

*[::-:

J

140

120100

-

60 80

4

0

1

40

f

i

1

80

f

i

'

120

f

i

'

160

f

i

f

200

i

Temperature ("C)

1

1

f

240

i

280

f

i

l

~

Figure 29. Variation of the long spacing L versus temperature for PA 46/PA 61 blends on cooling from 3OO0Cto 50°C. Cooling rate lO0C/min, blend composition (PA 46/PA 61): (100/0); (85/15); o (70/30); A (50/50); x (30/70)

+

280 260 240

-

-

0

3 220-

*I:

0 .

4 200-

J

140

-

120-

80

100

-

60 4 0

1

1

1

1

1

f

i

1

1

1

1

1

1

1

1

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

309

the blend contains 70 wt% of PA 61 or more. As a result, (pc - pa)' will decrease, which will affect the scattered intensity. (ii) An increase in the volume fraction of PA 61 in the blends will result in a decrease in the total degree of crystallinity 4. If 4 < 50%, +(l- 4) will become smaller with decreasing blend crystallinity and thus the invariant will also decrease. However, #(1- #) changes only slightly when # is varied between 30 and 70%. (iii) In addition to the previous arguments, the most important reason for the decrease in scattered intensity could be the decrease in stacking order of the lamellae of PA 46 sequences when the weight fraction of PA 61 is increased in the blend. This is illustrated by TEM micrographs in the next section. Figure 29 shows large variations in the mean long period L as a function of the blend composition. The rise in weight content of PA 61 from 15 to 85% increases the observed long period L from about 100 to 230 A due to segregation of PA 61 in the interlamellar amorphous regions. Figure30 represents the variation of the long spacing L veTsus temperature for the PA 46/PA 61 (50/50) blend on slowly cooling from the melt after different melt residence times. It can be seen that L decreases at prolonged melt residence times. A similar decrease of L has also been observed when the extrusion temperature is increased. This can be ascribed to a reduction in the interlamellar amorphous phase thickness and/or the crystalline region thickness.

a

6.1.3. h n s m i s s i o n electron microscopy

A better insight into the semicrystalline blend morphology can often be obtained by combining SAXS measurements with TEM observations. The

Figure 31. Lamellar morphology of PA 46, slowly crystallised from the melt

310

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

Figure 32. Lamellar morphology of PA 46/PA 61 (50/50) blend, slowly crystallised from the melt

Figure 33. Lamellar morphology of PA 46/PA 61 (30/70)blend, slowly crystallised from the melt

lamellar structure of PA 46 is shown in Figure31. It can be seen that the morphology consists of densely packed lamellae. The least stained areas are the crystalline regions while the heavily stained areas represent the amorphous phase. The crystalline lamellae have a thickness of about 30-40A. The addition of 50% by weight of PA 61 results in more disordered sheaflike structures (Figure 32); the crystalline thickness is about 25-35A. The

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

311

blend composed of 30% PA 46 and 70% PA 61 has a crystalline thickness of about 20-30A. Only the crystals with the longest crystalline sequences, which are able to fold, can be observed (Figure 33). The crystalline lamellae are covered by a heavily stained layer. Staining experiments with pure PA 61 reveal that this polyamide, stained under comparable conditions, yields bright micrographs. This leads to the conclusion that the heavily stained layers along the crystalline lamellae represent mainly amorphous PA 46. The region between the crystalline lamellae consists of amorphous PA 61 and amorphous PA 46. The strong decrease in the lamellar stacking order with increasing weight content of PA 61 leads to the decrease in the SAXS intensity of the blends, as mentioned in Section 6.1.2; only the ordered stacks of crystalline lamellae contribute to the SAXS diffraction pattern. 6 . 2 . Relation between the crystalline morphology and the melting behaviour of the PA 46/PA 61 copolymers

The quantitative interpretation of the melting behaviour of the PA 46/PA 61 copolyamides formed during reactive melt extrusion is quite difficult because of the high degree of complexity of the structures formed. In the previous section, it is shown on the basis of the WAXS data that melt-mixing of PA 46 and PA 61 does not seem to have a substantial effect on the perfection of the PA 46 crystals at the onset of melting. Although large melting point depressions are observed as a function of the blend composition and the melt processing conditions, the crystal perfection of the pseudohexagonal structures just before melting remains unaffected, regardless of the blend composition and processing conditions. From SAXS measurements and TEM observations, a variation of the long range stacking order can be noticed with increasing weight content of PA 61. In addition to the loss of stacking order, a small decrease in the crystalline thickness could also be detected from the TEM micrographs. The relationship between the crystalline lamellar thickness and the resulting melting temperature of semicrystalline homopolymers is described well by the Hoffman-Weeks equation [95,96]:

T,,, = T L ( 1 where T: is the thermodynamic equilibrium melting point, I , - the lamellar thickness, 0, - the fold surface free energy, and AH; - the melting enthalpy of a perfect infinite crystal. For PA 46, the lamellar thickness can be calculated from the experimental melting temperature and the following reference values [49]: T; = 350°C, 0, = 75 x J/m2, AH: = 210 J/g (251 J/cm3). A lamellar thickness of 36A is calculated in this way, which is within the range of lamellar thicknesses as observed by TEM. It can be derived from Eq. (16)

312

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

that a small decrease in the crystalline lamellar thickness of PA 46, due to the reactive melt-mixing process with PA 61, can cause large melting point depressions. A second approach consists in the calculation of the melting point depression of the semicrystalline copolymers under thermodynamic equilibrium, using the expression formulated by Flory [97]:

where TAis the equilibrium melting temperature of the random copolymer, T; is the equilibrium melting temperature of the semicrystalline homopolymer, AH, is the heat of melting per mol of crystalline units A at Tk,R is the gas constant, and p,, is the probability of a crystallisable homopolymer unit A being followed by another homopolymer unit A. The probability p,, drops as the transamidation reactions proceed, resulting in a decreasing melting point TA.Assuming a random distribution of crystallisable and non-crystallisable units, the sequence propagation probability p,, is directly related to the concentration of crystallisable units 2,; then Eq. (17) becomes (981:

where xA is the molar fraction of crystallisable units A. The Flory equation is only valid for random copolymers. However, at the beginning of the melt-mixing process, the PA 46/PA 61 copolyamides formed are still block copolymers, ie., far from random. A modified expression for non-random block copolymers composed of crystallisable sequences A and non-crystallisable sequences B is proposed by Baur [99]; it differs from Flory’s equation by an additional term, 2R(1 - pAA)zA.

7. General conclusions Blends of semicrystalline aliphatic PA 46 and amorphous aromatic PA 61 are prepared by coprecipitation from a common solvent and by meltmixing using a mini-extruder. The blends are found to be fully miscible over the entire composition range by both blending methods. However, comparison of the blends reveals substantial differences in crystallisation and melting behaviour, depending on the blending method. This is ascribed to the occurrence of transamidation reactions during melt-miuing of the homopolyamides, resulting in the formation of block copolyamides composed of crystallisable and non-crystallisable sequences. The extent of transamidation is governed by the extrusion temperature and duration. Increasing one of these processing parameters enhances the transamidation

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

313

reactions and shortens the crystallisable segment lengths. This leads to lower crystallisation rates and melting points of the melt-mixed blends. When the melt-blending experiments are performed on a large-scale double screw extruder under the same conditions, it appears that the transreaction processes occur to a similar extent to those in the mini-extruder. These upscaling experiments also reveal that the average sequence length of PA 46 in the copolymer formed is dependent not only on the melt processing conditions, but also on the overall blend composition. Low concentrations of PA 61 in the blends lead to a small decrease in the crystallisation temperature, while a large decrease in the crystallisation temperature is observed at high contents of PA 61. This can be ascribed to the extent of transamidation reactions occurring between the two polyamides, which is dependent on the blend composition. A high content of PA 61 in the blends results in a fast conversion of homopolyamide 46 into a segmented block copolymer with rather short crystallisable PA 46 sequences, decreasing the crystallisation rate greatly. Direct evidence of the transamidation reactions in the PA 46/PA 61 blends is provided by 13C NMR measurements. The carbonyl resonance signals of the PA 46/PA 61 copolymer are assigned to four types of dyad sequences in polyamide blends, that have reacted for long times of meltmixing in the mini-extruder. This makes it possible to distinguish physical blends of PA 46/PA 61 from polyamide copolymers. Moreover, the degree of transamidation and the degree of randomness as well as the number-average block length of the copolyamides can be calculated using the method for a four-component polyamide system. It is shown that the transamidation reactions occur very fast at the beginning of melt-mixing, but slow down at longer extrusion times. High extrusion temperatures speed up the transamidation processes. From 13C NMR measurements on PA 46/PA 61 blends with different compositions, differences in reactivity could be derived. Moreover, it is also observed that an average PA 46 block length of 3-4 repeat units has become too small to be crystallisable. PA 46/PA 61 copolymers, as obtained during short extrusion times, are not suited for I3C NMR characterisation because only a limited number of reactions occur during the short melt residence time. However, copolymer formation at low conversions can be investigated through gradient elution chromatography, which is a technique based upon the solubility of the polymers and their adsorption to the stationary phase. Using GEC, it is concluded that the homopolyamides PA 46 and PA 61 have already been converted into block copolyamides within the first 10 min of melt-mixing, and increasing the extrusion temperature accelerates this process. However, the variation of the PA 46 end-group concentration and end-group ratio at the beginning of extrusion does not appear to influence the elution behaviour of the copolymers formed, and a transamidation mechanism of hydrolysis and recombination is suggested. The semicrystalline morphology of the PA 46/PA 61 copolyamides is

314

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

investigated by time-resolved S A X S and WAXS measurements, using synchrotron radiation, and the melting behaviour of these copolyamides is related to their semicrystalline morphology. From the WAXS data, it can be derived that only an excess of amorphous PA 61 in the blends can affect the PA 46 crystal unit cell by hindering or inhibiting the Brill transition during cooling from the melt. There are, however, no indications from the timeresolved WAXS measurements that the PA 46 crystal unit cell perfection at the onset of melting is influenced by the blend composition or the extrusion conditions. As a consequence, the observed melting point depressions cannot be ascribed to less perfect PA 46 crystals. However, a decreasing lamellar thickness of the crystallised PA 46 sequences with increasing PA 61 content in the blends is observed by TEM. Moreover, thermodynamic calculations reveal that the melting point depression is very likely due to a reduction of the thickness of the crystalline PA 46 lamellae in the blends. The type of PA 61 segregation in the polyamide blends varies depending on the blend composition and the conditions of crystallisation from the melt. As a general conclusion, it can be stated that reactive melt-blending of polyamides opens a new route to the preparation of novel copolyamides with various degrees of randomness and compositions. Control of the interchange reactions in polyamide blends is very important for the preparation of copolyamides with well-defined chain microstructures. It clearly appears that transreactions in polyamide blends always proceed as a function of melt residence time, which is undesirable with respect to product stability during melt processing. Therefore, methods are needed to terminate the transamidation reactions in the melt at a predetermined extent. One of these methods consists in the use of end-blocking products or inhibitors. On the other hand, faster transreactions are desirable in self-compatibilisation applications. In this case, the role of suitable catalysts is very important. Other topics of interest are the influence of transreactions on the molecular weight of the resulting copolymers and on their ultimate mechanical properties.

Acknowledgements The authors are grateful to DSM Research, Geleen, The Netherlands, for financial support of a part of the research described in this chapter, as well as to H. Repin, W. Bruls, S. Eltink, R. Leeuwendal, J. Van Asperen, Sj. Van der Wal, Y . Mengerink, and N. Meijering for valuable discussions, and to M. Koch for his help during the synchrotron experiments at EMBL in Hamburg, Germany. G. Groeninckx is also indebted to the Research Council of the K. U. Leuven, and to the Fund for Scientific Research - Flanders (Belgium) for the financial support offered to the MSC laboratory.

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

315

References 1. A.M. Kotliar, J. Polym. Sci., Macromol. Rev. 16, 367 (1981) 2. G. Montaudo, C. Puglesi, F. Samperi, J. Polym. Sci., Polym. Chem. Ed. 32, 15 (1994) 3. H. Tobita, Y. Ohtani, Polglmer 33, 2194 (1992) 4. J. G. Lertola, J. Polym. Sci., Polym. Chem. Ed. 28, 2793 (1990) 5. A.M. Kotliar, J. Polym. Sci., Polym. Chem. Ed. 11,1157 (1973) 6. K.L.L. Eersels, A.M. Aerdts, G. Groeninckx, Macromolecules 29, 1046 (1996) 7. Y. Takeda, D. R. Paul, Polymer 32, 2771 (1991) 8. Y. Takeda., D. R. Paul, Polymer 33, 3899 (1992) 9. J.M. Martnez, J. Nazdbal, J.I. Eguiazdbal, J. Appl. Polym. Sci. 51, 223 (1994) 10. A. Legros, P. J. Carreau, B. D. Favis, A. Michel, Polymer 35, 758 (1994) 11. R. Y. Zhang, X. L. Luo, D. Z. Ma, J. Appl. Polym. Sci. 55, 455 (1995) 12. Eur. Patent 38094 (1981), DSM, Inv.: R. J. Gaymans, E. H. J. P. Bour, Chem. Abstr. 96, 96636m (1982) 13. Eur. Patent 39524 (1981), DSM, Inv.: R. J. Gaymans, E. H. J. P. Bour, Chem. Abstr. 96, 69637n (1982) 14. R. J. Gaymans, T. E. C. Van Utteren, J. W. A. Van den Berg., J. Schuyer, J. Polym. Sci., Polym. Chem. Ed. 15, 537 (1977) 15. E. Roerdink, J. M. M. Warnier, Polymer 26, 1582 (1985) 16. H. M. J. C. Creemers, Kunststof en Rubber 8, 18 (1991) 17. W. Nielinger, B. Brassat, D. Neuray, Angew. Mukronzol. Chem. 98, 225 (1981) 18. L.F. Beste, R. C. Houtz, J. Polym. Sci. 8 , 395 (1952) 19. I.K. Miller, J. Polym. Sci., Polym. Chem. Ed. 14, 1403 (1976) 20. T. M. Frunze, V.V. Korshak, V.A. Makaxkin, Vysokomol. Soedin. 1, 500 (1959) 21. V. V. Korshak, T. M. Frunze, Lu I-nan, Vysokomol.Soedin. 2, 984 (1960) 22. V. V. Korshak, T. M. Frunze and Lu I-nan, Vysokomol.Soedin. 3,665 (1961) 23. US 4417032 (1983), Allied Co., Inv.: Y. P. Khanna, E. A. Turi, S. N. Aharoni, T. Largmann, Chem. Abstr. 100, 52525c (1984) 24. T. Suzuki, H. Tanaka, T. Nishi, Polymer 30, 1287 (1989) 25. R. S. Porter, L.H. Wang, Polymer 33, 2019 (1992) 26. W. G. Zheng, Z. N. Qi, Polym. Int. 34, 307 (1994) 27. T. S. Ellis,Polymer 33, 1469 (1992) 28. E. Gattiglia, A. Tarturro, F. P. Lamantia, A. Valenza, J. Appl. Polym. Sci. 46, 1887 (1992) 29. W. A. Smith, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 26, 233 (1981) 30. J. I. Eguiazabal, G. Ucar, M. Cortazar, J. I. Iruin, Polymer 27, 2013 (1986) 31. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22, 229 (1982) 32. J. Devaux, P. Godard, J. P. Mercier, R. Toullaux, J. M. Dereppe, J. Polym. Sci.,Polym. Phys. Ed. 20, 1881 (1982) 33. R. S. Barnum, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 27,4065 (1982) 34. J. I. Eguiazabal, M. Cortazar, J. I. Iruin, G. M. Guzman, J. Macromol. Sci.Phys. B27, 19 (1988)

316

K. L. L. Eersels, A. M. Aerdts, G. Groeninckx

35. M. Cortazar, J. I. Eguiazabal, J.I. Iruin, Eur. Polym. J. 30, 901 (1994) 36. J. I. Eguiazabal, M. Cortazar, 3. J. Iruin, J. Appl. Polym. Sci. 42,489 (1991) 37. M. T. Heino, J. V. Seppiila, Acta Polytechnica Scandanavica, Chemical Technology and Metallurgy Series, No. 214, Helsinki 1993 38. W. G. Zheng, Z. N. Qi, Polym. Int. 34, 301 (1994) 39. M. Kimura, R.S. Porter, J. Polym. Sci., Polym. Phys. Ed. 21, 367 (1983) 40. M. Kimura, G. S a l e , R. S. Porter, J. Appl. Polym. Sci. 29, 1629 (1984) 41. J. I. Eguiazabal, M.E. Calahorra, M. M. Cortazar, J. J. Iruin, Polym. Eng. Sci. 24, 608 (1984) 42. R. W. Lenz, S. Go, J. Polym. Sci., Polym. Chem. Ed. 11, 2927 (1973) 43. R. W. L e u , S. Go, J. Polym. Sci., Polym. Chem. Ed. 12, 1 (1974) 44. R. W. Lenz, A. N. Schuler, J. Polym. Sci., Polym. Symp. 63, 343 (1978) 45. R. W. Lenz, J. I. Jin, K. A. Feichtinger, Polymer 24, 327 (1983) 46. G. Chen, R. W. Lenz, Polymer 26, 1307 (1985) 47. S. Fakirov, M. Evstatiev, Adv. Mater. 6, 395 (1994) 48. T. S. Ellis, Macromolecules 24, 3845 (1991) 49. W.H.P. Derks, J. A.H.M. Moonen, F. J. W. Ramaekers, C. J. Kooij, IUPAC Conference, Montreal 1990 50. M. E. Stewart, A. J. Cox, D. M. Naylor, Polymer 34, 4060 (1993) 51. X. Yang, P. C. Painter, M. M. Coleman, Macromolecules 25, 4996 (1992) 52. W. G. Zheng, Z. H. Wang, Z. N. Qi, Polymer 34, 4982 (1993) 53. D. M. Miley, J. Runt, Polymer 33, 4643 (1992) 54. M. Valero, J. J. Iruin, E. Espinosa, M. J. Ferkdez-Berridi, Polym. Commun. 31, 127 (1990) 55. J. I. Jin, J. H. Chang, K. Hatada, K. Ute, M. Hotta, Polymer 33, 1374 (1992) 56. J. Kelm, H. J. Kretzschmar, H. Zimmer, Kumtstofle 71, 514 (1981) 57. H. R. Kricheldorf, PUR Appl. Chem. 54, 467 (1982) 58. G. P. M. van der Velden, J. Beulen, H. de Keijzer, Rec. Z'rav. Chzm. Pays-Bas 110, 516 (1991) 59. H.R. Kricheldorf, E. Leppert, G. Schilling, Makromol. Chem. 175, 1705 (1974) 60. H.R. Kricheldorf, Makromol. Chem. 179, 2133 (1978) 61. H. R. Kricheldorf, E. Leppert, G. Schilling, Makromol. Chem. 176,81 (1975) 62. G. Schilling, H.R. Kricheldorf, Makromol. Chem. 176, 3341 (1975) 63. H. R. Kricheldorf, R. Miilhaupt, Angew. Makromol. Chem. 65, 169 (1977) 64. H. R. Kricheldorf, W.E. Hull, J. Macromol. Sci-Chem. A l l , 2281 (1977) 65. H.R. Kricheldorf, W.E. Hull, J. Polym. Sci., Polym. Chem. Ed. 16, 2253 (1978) 66. H.R. Kricheldorf, K.H. Reith, J. Polym. Sci., Polym. Chem. Ed. 16, 379 (1978) 67. H.R. Kricheldorf, B. Coutin, H. Sekiguchi, J . Polym. Sci., Polym. Chem. Ed. 20, 2353 (1982) 68. R. Yamadera, M. Murano, J. Polym. Sci. 5A-1, 2259 (1967) 69. M. Murano, R. Yamadera, Polym. J. 2, 8 (1971) 70. J. Devaux , P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1875 (1982) 71. A.M. Aerdts, K.L.L. Eersels, G. Groeninckx, Macromolecules 29, 1041 (1996)

Reactive Melt Processing of Aliphatic/Aromatic Polyamide Blends

317

72. J. L. Koenig, “Chemical Microstructure of Polymer Chains”, Wiley, New York 1980 73. K. L. L. Eersels, G. Groeninckx, Polymer 37,983 (1996) 74. K. L. L. Eersels, G. Groeninckx, J. Appl. Polym. Sci. 63,573 (1997) 75. X. Yang, P. C. Painter, M. M. Coleman, Macromolecules 25,4996 (1992) 76. A.Muhlenbach, J. Economy, R. D. Johonson, T. Karis, J. Lyerla, Macromolecules 23, 1803 (1990) 77. M. Valero, J. J. Iruin,E. Espinosa, M. J. Fernandez-Berridi, Polym. Commun. 31,127 (1990) 78. G. Glockner, H. G. Barth, J. Chromatogr. 499,645 (1990) 79. G. Glijckner, J. Appl. Polym. Sci., Appl. Polym. Symp. 51,45 (1992) 80. K.L.L. Eersels, G. Groeninckx, Y.Mengerinck, Sj. van der Wal, Macromolecules 29,6744 (1996) 81. J. Devaux, P. Godard, J.P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 82. M. J. Fernandez-Berridi, J. J. Iruin, I. Maiza, Polymer 36, 1357 (1995) 83. K.L.L. Eersels, G. Groeninckx, M. Koch, H. Reijnaers, Polymer, in press 84. J. P. Parker, P. H. Lindenmeyer, J. Appl. Polym. Sci. 21,821 (1977) 85. R. J. Gaymans, D. K. Doeksen, S. Harkema, “Integration of Fundamental 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99.

Polymer Science and Technology“, edited by L. Kleintjens, P. Lemstra, Elsevier, Amsterdam 1986, vol. I, p. 573 R. Brill, J. Prakt. Chem. 161,49 (1942) T. Itoh, Jpn. J. Appl. Phys. 15,2295 (1976) C. Ramesh, A. Keller, S. J. E. A. Eltink, Polymer 35,2483 (1994) E.D.T. Atkins, J. Macromol. Sci., Pure Appl. Chem. A31, 691 (1994) D.R. Holmes, C. W. B u m , D. J. Smith, J. Polym. Sci. 17,159 (1955) R. Brill, Makromol. Chern. 18,2940 (1956) G. F. Schmidt, H.A. Stuart, 2. Naturforsch., Teal A 13,222 (1958) H. G. OK, A. Peterlin, J. Polym. Sci. 9A-2,1449 (1971) J. Zimmerman, E.M. Pearce, I.K. Miller, J.A. Muzsio, I.G. Epstein, E.A. Hosewood, J. Appl. Polym. Sci. 17, 849 (1973) B. Wunderlich, “Macromolecular Physics”, Academic Press, New York 1980, vol. 3, Crystal Melting J. D. Hoffman, J. J. Weeks, J. Res. Nat. Bur. Stand. (A) 66,13 (1962) P. J. Flory, %ns. Faraday SOC.51,848 (1955) W.H. Jo, D. H. Baik, J. Polym. Sci., Polym Phys. Ed. 27,673 (1989) V.H. Baur, Makromol. Chem. 98,297 (1966)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 8

Sequential Reordering in Condensation Copolymers

S. Fakirov, Z. Denchev

1. Evidence of the occurrence of chemical interactions in blends of condensation polymers During recent decades, polymer blends have become of prime commercial importance [l].Currently, they constitute over 30% of the polymer market. The blends of condensation polymers are of particular interest because of their unique ability to undergo interchange reactions (additional condensation and transreactions, called also exchange reactions or mid-chain reactions) [2]. The latter are reversible and can be performed either in the molten [2] or in the solid state (31 under appropriate conditions. Elevated temperatures between 180°C and 300°C and addition of catalysts are among the basic requirements in this respect. Although the study of transreactions started 50 years ago with the pioneer work of Flory [4], even now some details of the process are not completely understood. There is no common consent in the literature on the relative importance of the three possible mechanisms - alcoholysis, acidolysis or direct ester exchange [5-121. Interchain reactions take place whenever the conditions required (appropriate temperatures and chemical compositions) are available. They occur in homopolycondensates, as can be concluded from the upgrading of molecular weight due to additional condensation; transactions have been convincingly confirmed by studying blends of protonated poly(ethy1ene terephthalate) (PET) with deuterium-labelled ( d 4 ) PET by small-angle neutron scat-

320

S. Fakirov, Z. Denchev

tering. The results show that ester interchange reactions take place rapidly in the melt and also proceed at lower temperatures (about 15°C below the melting of PET), resulting in the formation of a block copolymer consisting of deuterium-labelled and non-labelled PET segments [13,14]. As far as transreactions in typical homopolycondensate blends comprising polyesters are concerned, the first clear evidence of their occurrence was obtained much earlier by Kricheldorf by means of 13C nuclear magnetic resonance (NMR) [15]. In this work, numerous homopolyesters were synthesised and characterised. Also, binary blends of homopolyesters were prepared and characterised after heating for 8 h at 275°C in the presence of a catalyst. The interpretation of the 13C NMR spectra confirms undoubtedly that copolymers are obtained under these conditions (see also Chapter 1). Since in the course of transreactions the physical properties of the system constituents change due to the formation of new components, various techniques have been used for the detection of these changes, infrared (IR) spectroscopy among them. The IR spectra reflect molecular vibrations that cause changes in the dipole moment and polarisability of the molecular chains; they are unique to each molecule and therefore characterise the chain structure, especially the concentration of the constituent groups and the intramolecular forces acting among them. In several of the polyester blend studies, new components, resulting from transreactions, have been detected by IR spectroscopy [16-201 (see also Chapter 6). NMR is an even more powerful tool in the analysis of polymer microstructures, providing information also about the miscibility and the chemical changes in polyester pairs [7,18-241. The kinetics of transreactions has been studied by this technique [7]. The sensitivity of spectroscopic techniques is insufficient to detect the single bond per chain necessary to start transreactions. Nevertheless, the thermodynamics of interaction between blend components can be changed sensitively from the onset of transesterification. Therefore, methods that detect the thermal or thermomechanical behaviour are widely used in such blend studies. Differential scanning calorimetry (DSC), differential thermal analysis (DTA), and dynamic mechanical thermal analysis (DMTA) are used as essential methods in the study of interchange reactions. The factors determining the glass transition temperature, T’,change with the progress of transesterification. Blends are converted first into block copolymers and finally into random copolymers [6,10,11,18,25-311 because all repeating structural units are equally likely to react. The relative populations of homopolymers and block copolymers depend on the mols of bonds interchanged [11,26,31].Therefore, the properties of a blend in which transesterification is possible depend on the rate and degree of completeness of the exchange reaction. This should be more evident in miscible or partially miscible polymer blends. The resultant initial block and eventual random copolymers are ex-

Sequential Reordering in Condensation Copolymers

32 1

pected to exhibit enhanced mutual miscibility, compared to the original non-reacted components. It is also believed that transreacted chains facilitate and accelerate the compatibilisation and chemical interaction of the chains [12]. Transesterification enables copolymers with different composition and randomness to be obtained. The latter is a function of temperature and residence time in the melt, and of the percentage of each homopolymer in the blend. Generally, transesterified copolymers with blocky sequential order are crystallisable in the event that the starting homopolymers are crystallisable, too. Concerning the order in random copolymers and its influence on their crystallisation ability, Hanna et aZ. [32] have summarised the two opposite opinions existing in the literature: (i) In polymer blends comprising at least one crystallisable component, the above-mentioned transreactions lead to the formation of copolymer chains in which two or more different types of chemical units are joined in a random sequence. The authors consider such chains unable to contribute fully to the three-dimensional periodicity of a crystal lattice. As a consequence, one has to predict amorphisation of blends in which transreactions are possible, leading to a high degree of randomness, which was observed in practice by Kimura et al. for PETlpolyarylate (PAr) copolymers (331; (ii) According to the other point of view, thoroughly set forth in [32], copolymers containing units of crystallisable polymers, can show certain crystallisation ability even at a very high degree of randomness. The authors proved this idea by means of computer models of systems of random copolymer chains. It was thereby demonstrated that there is a probability of segregation of certain proportions of similar units exceeding a given length, forming rather perfect, albeit small and isolated, crystals. Also, it is found possible for identical but random sequences to segregate, creating lateral order in a unit with no periodicity in the direction of the chain axis. Bearing in mind that a copolymer with blocky sequential ordering should crystallise if at least one of the starting homopolymers is crystallisable and that a copolymer with random type of sequential ordering is more or less non-crystallisable, the study of the behaviour of the crystalline phase(s) can offer an insight in the occurrence and depth of chemical interactions between the blend components. At the same time, the properties of the amorphous phases will support these conclusions since at the stage of complete randomisation, particularly in equimolar blends of condensation polymers, a single glass transition should be observed instead of the two glass transitions, typical of a mechanical blend or of a block copolymer.

1.1. Evidence derived from the behaviour of the crystalline phases Recently, melt-blended PET and polyamide 6 (PA 6 ) were studied by different techniques, the blend being drawn after extrusion and annealed for various times around the melting temperatures of the blend constituents

322

S. Fakirov, Z. Denchev

[34].Methods, such as DSC and wide-angle X-ray scattering (WAXS), that follow the properties of the crystalline phases, clearly demonstrated the existence of two crystalline phases corresponding to the starting homopolymers as long as no intensive chemical interactions have taken place. At the stage of complete randomisation, however, these crystalline phases disap pear. If the thermal treatment is performed above the melting temperature of one of the blend constituents, but below T, of the other, the latter preserves its crystalline structure, as can be concluded from the figures below. Figure 1 shows curves for the second heating of blends crystallised from the melt of an isotropic blend or after drawing and annealing of the same material at different temperatures [34]. WAXS transmission patterns of PET/PA 6 blend films, zone-drawn and then annealed at various temperatures and durations, are represented in Figure 2. For the unannealed zone-drawn blend (Figure 2a), the concentration of the amorphous halo in the equatorial direction and the breadth of the crystalline reflections suggest a very good orientation of the chains in

60

120

180

240

300

T ("C)

Figure 1. DSC curves of PET/PA 6 blend (ultraquenched, zone-drawn and additionally annealed at different temperatures and for various durations) taken in second heating mode. Annealing temperatures T, and durations t,: (a) unanT =22OoC, t a = 5 h; (d) nealed, undrawn; (b) unannealed, drawn to A=4.2; (c) ' T,=22OoC, t a = 2 5 h (e) Ta=24OoC, ta=5h; (f) Ta=24OoC, t,=25h (341

Sequential Reordering in Condensation Copolymers

323

Figure 2. WAXS transmission patterns of PET/PA 6 blend, zone-drawn and annealed at different temperatures and for various durations: (a) zone-drawn, unannealed; (b) T, = 220°C, t, = 5 h; ( c ) Ta= 240°C, t a = 5 h; (d) T a = 24OoC, ta

= 25 h [34]

the axial direction and the presence of small and imperfect crystallites. The disappearance of the halo and the sharpening of reflections in the equatorial and meridional directions after additional thermal treatment at 220°C for 5 h (Figure 2b) are indications of the improvement of chain orientation and perfection and/or growth of crystallites. In the case of the sample annealed at 240°C for 5 h (Figure 2c), the PA 6 diffraction is in the form of Debye rings, suggesting isotropy of that component. At the same time, the orientation and perfection of the crystallites in the PET fraction remain unchanged. Prolonged annealing (25 h) at the same temperature leads to the almost complete disappearance of the PA 6 Debye rings, as well as to a n improvement of the perfection of the PE T crystallites (Figure 2d). The samples treated at lower temperatures, i e . , below T, of PA 6 or above it but for shorter times (Figure 2c) have a common characteristic feature. The results obtained by DSC and WAXS provide clear evidence of the coexistence of two crystalline phases. They arise from the two blend components, regardless of whether these components are oriented or isotropic (Figure 2a-c). When annealing at 240°C is extended to 25 h, the situation changes drastically - there is no more evidence of the existence

324

S. Fakiiov, Z. Denchev

of PA 6 crystalline phase. The corresponding DSC curve (Figure l f ) and WAXS pattern (Figure 2d) look as if the sample is free of PA 6 component. To explain this effect, one has to remember at this point that the annealing temperature of 240°C is above the Tm of PA 6 and well below the melting of PET. Therefore, under these conditions, PA 6 is completely molten which is not the case for PET. During the 25 h annealing, all the PA 6 reacts with the amorphous part of the PET, resulting in the formation of copolymers. From the DSC and WAXS results (Figures If and 2d, respectively) one can conclude that either the copolymers formed are in random sequential ordering or the existing blocks (if any) have suppressed crystallisability. As evident from the discussion below, it is the former conclusion that is supported by the behaviour of the amorphous phases. 1.2. Evidences derived f r o m the behaviour of the amorphous phases

As already mentioned, conclusions about the existence and number of the amorphous phases are based mostly on the presence and number of glass transition temperatures. DMTA is commonly considered as the most reliable technique for the determination of Tgand for this reason it was applied to the PET/PA 6 blend subjected to drawing and annealing. The results are summarised in Figure 3 [35]. Again, annealing at lower temperatures, or at higher ones (240°C)but for shorter times, leads to the formation of two well-defined and spatially separated amorphous phases, giving rise to two glass transition temperatures TPETand Tg”‘ for the PET and PA 6 amorphous phases, tespectively (kigure 3c,d). Annealing at 240°C for 5 h (Figure 3e) results in a decrease in the PA 6 amorphous fraction. After annealing for 25 h it disappears completely (Figure 3f). This observation indicates that after such a thermal treatment, all of the amorphous PA 6 is involved in a copolymer with the amorphous PET. The behaviour of both crystalline (Figures 1, 2) and amorphous (Figure 3) phases clearly indicates that the PA 6 starting component does not exist any more either in the crystalline or in the amorphous phases. This is due to its involvement in a copolymer with a supposedly random structure as a result of transreactions taking place during prolonged annealing at high temperatures. 1.3. Evidence derived fmm the behaviour of crystalline and amor-

phous phases

1.3.1. IR study and weight control after selective extraction The DSC and WAXS measurements described in Section 1.2 were performed with annealed PET/PA 6 blends without any further treatment. The same samples were also studied by IR spectroscopy after application of selective extraction, accompanied by weight control of the fractions. This

Sequential Reordering in Condensation Copolymers

-150

-100

-50

0

50

100

Temperature ("C)

325

150

:

10

Figure 3. Temperature dependence of the loss modulus E" in the range from -150 to 200°C. Annealing temperatures and durations are as follows: (a) unannealed, undrawn; (b) unannealed, drawn to X = 4.2; (c) T, = 220°C, t, = 5h; (d) T, = 22OoC, t , = 25 h; (e) T,= 240°C, t , = 5 h; (f) T, = 240°C, t, = 25 h [35] approach aims at the quantitative removal of the PA 6 component from the system, using formic acid as a selective solvent. The IR spectra taken after extraction of PET/PA 6 films with different thermal prehistories are given in Figure 4 [36]. It is seen that the peak intensities at 3300cm-' are quite different. Since this is the wavelength characteristic of the amide group, the respective peak intensity can be considered as a measure of the amount of PA 6 fractions in the blends. The weight losses after extraction are given in Table 1. Taking into account that the weight ratio of the components in the blend is 1:1, and that PET is insoluble in formic acid, the weight losses refer solely

326

S. Fakirov, Z. Denchev

Table 1. Weight losses as a result of selective extraction of PA 6 (based on the PA 6 content) of a PET/PA 6 blend (1:lbywt) after various thermal treatments (361

Sample as in Figure 4 Annealing temp. ("c) Annealing time (h) Weight loss (%)

a

b

-

-

98

96

-

C

d

e

f

220 5 91

220 25 80

240 5 62

240 25 22

to PA 6. However, they depend strongly on the annealing temperature for the thermally untreated samples they are about 100% whereas after annealing at 240°C for 25 h they drop to about 20% (Table 1). It can be concluded that practically the entire PA 6 fraction is extracted in thermally untreated samples, in contrast to those annealed at 240°C for 25 h. The rise in the intensity of the IR absorption band at 3300cm-', as well as in the range of 2200-2800~m-~,which is characteristic of the vibrations of the -CH2- groups, follows the same trend since the copolymer is enriched in -CH2- groups, as compared to homo-PET (Figure 4). Thus, it is quite clear that the amount of PA 6 incorporated in a copolymer increases with annealing temperature and duration.

b C

d C

f

t 3 i O O 3600

t

3400 3200 3000

2800 (cm-')

Figure 4. 1R spectra recorded after extraction in formic acid of PET/PA 6 (1:l by wt) films with different thermal prehistories: (a) as quenched; (b) as drawn; (c) Ta = 22OoC, t a = 5h; (d) Ta = 220°C, ta = 25h; (e) Ta = 240°C, t , = 5h; (f) Ta = 24OoC, ta = 25 h [36]

Sequential Reordering in Condensation Copolymers

327

1.3.2. Scanning electron microscopy study after selective extraction Scanning electron microscopy (SEM) studies were carried out with the same PET/PA 6 samples [36]. Bearing in mind the fact that PET and PA 6 sxrongly differ in their orientation state, it was possible to illustrate the occurrence of intensive chemical interactions between the two blend components. After cold drawing of the isotropic blend, both blend constituents are in a well oriented state, as can be concluded from the WAXS patterns shown in Figure 2a,b. The shorter thermal treatment at 240°C followed by cooling results in crystallisation of PA 6 in the isotropic state (Figure 2c). At the same time, the PET component preserves its good orientation. It wax possible at this stage to remove the chemically non-bonded PA 6 by its dissolution in formic acid. As a result, it was possible to observe directly the PET microfibrils. Two typical micrographs are displayed in Figure 5. Considering the blend morphology, it is seen that a highly fibrillised PET structure formed at the stage of orientation by drawing is preserved after annealing (Figure 5a). Fibrils with diameters of about 1pm are predominant. It should be noted that during annealing at T, = 240°C these perfect individual fibrils form highly oriented aggregates (Figure 5b). At firs; glance this fact is surprising. However, one should consider the chemical processes taking place at this high temperature and extended annealing time of 25 h. In accordance with the studies by WAXS, DSC, and DMTA, this observation can be explained by transreactions between PET and PA 6 and formation of copolymers comprising the entire amount of PA 6 and the amorphous part of PET. These copolymers are resistant to the selective solvent. Therefore, the micrographs reveal highly interconnected microfibrils (Filpre 5b) , forming larger morphological entities.

Figure 5. SEM micrographs (magnification 5000~)taken from PET/PA 6 (1:l by wt) films after selective extraction with formic acid. Thermal treatment before extraction: (a) T, = 220°C, t, = 25h; (b) T, = 24OoC, t, = 25h [36]

328

S. Fakirov, Z.Denchev

1.3.3. Light microscopy after selective extraction The occurrence of transreactions in blends of condensation polymers can be visualised by observation of thin slices of the blend at different stages of the thermal treatment, subjected to selective dyeing. The PET/PA 6 blends are suitable for such observations because of the possibility to dye the PA 6 constituent selectively using a wool-type dyestuff [37]. In PET/PA 6 blends, if there is no chemical interaction between the homopolymer components, one should expect well-defined phase boundaries between the particles of the components since these blends are known to be immiscible [l].The situation changes completely with the progress of transtreactions, as seen in Figure 6. Since the dyestuff used is exclusively polyamideoriented, it does not change the colour of homo-PET. After blending and orientation, a rather homogeneous distribution of the two components can be expected and subsequent annealing at 220°C improves the morphology (Figure 6a,b). Isotropisation at 240°C (Figure 6c) results in the formation of larger PA 6 domains (as compared to the two previous cases), although the fibrillised matrix is also visible. Since the isotropisation temperature is high enough for the occurrence of chemical reactions, the appearance of the larger dark areas (Figure

Figure 6. Microphotographs in polarised light of thin slices of PET/PA 6 blend (1:l by wt): (a) PA 6 selectively dyed after drawing; (b) dyed after annealing at 220°C for 5h; (c) dyed after annealing at 240°C for 5h, and (d) dyed after annealing at 240°C for 25 h. Magnification 38Ox [37]

Sequential Reordering in Condensation Copolymers

329

6c) should be attributed to the formation of block copolycondensates, the colouration of which is due to the content of PA 6 units. This suggestion is supported by the result of annealing at the same temperature but for a longer time (25 h, Figure 6d); in this case PA 6 is in the molten state while PET preserves its microfibrillar structure. Due to the reactions described above, a complete transformation of the matrix from a homopolymeric (PA 6 and amorphous PET) into a copolymeric state takes place, resulting in the almost uniform colouration of the entire sample (Figure 6d), except for the PET fibrils that are smaller than the resolution of the instrument used.

1.4. Evidence derived from chromatographic methods

As a consequence of chemical interactions in a blend of condensation poly-

mers, the skeletal structure of macromolecules, the molecular weight (MW) and the molecular weight distribution (MWD) should vary simultaneously to an extent depending on the depth of the chemical changes that take place. Therefore, chromatography techniques sensitive to these changes would also provide evidence of the occurrence of transreactions [12].

n I

PET

PETIPC copolyrmr

PC

I

10 rnin

30 min

60 min

A

Q

Spot Point

-

Distance of migration

Figure 7. Thin-layer chromatograms of PET/PC (50/50 by wt) blends, meltmixed for various times, developed first with dichloromethane, and then with phenol/tetrachloroethane at 270°C [21]

As reported by Wang et al. [21], a combination of thin-layer chromatog-

330

S. Fakirov, Z. Denchev

raphy (TLC) and IR spectroscopy has been used to identify the products of transreactions in a 50/50 (by wt) PET/bisphenol A polycarbonate (PC) melt-mixed blend after different heat treatments. The change in the TLC curves as a function of the reaction time at 270°C (Figure 7) supports the hypothesis of almost complete transformation of the starting homopolymers PET (the left peak) and PC (the right peak) into a copolymer product (the peak in the middle) after a reaction time of 140min (Figure 7e) [21]. More recently, Fiorini et al. [38] studied the effects of various catalysts employed for PET synthesis on the transreactions in PC blends of the same homopolymer using the method of size exclusion chromatography (SEC). The PET/PC blends were melt-mixed for relatively short times (2030min at 270°C) and then the PC blocks were selectively degraded to low molecular weight products. (The procedure of aminolysis was previously reported by Pilati et al. in [5]). In such a way, the carbonate groups of the P C component become involved into low molecular weight urethanes and diols. At the same time, the PET sequences remain unaffected by the degradation process (see Section 5.2). One can see from the SEC curves in Figure 8 that the Ti(0Bu)rcatalysed reactive blending (Figure 8d) significantly decreases the molec-

s

I

- I

I0

I 20

I

15

V,

I

25

,ml.

Figure 8. SEC elution curves of melt-mixed PET and PC (20 min at 270°C) in the presence of Ti(0Bu)r: (c) crude sample after mixing; (d) residual PET after selective degradation. Dashed and dotted lines depict the starting PC (a) and PET (b), respectively. Elution solvent: hexafluoroisopropanol (HFIP)/C&C12/CHCL (5/20/75 by wt) at 0.6ml/min. Columns: PLgel, 500, Id,and 104A, at 25°C with UV detection. V,, elution volume (381

331

Sequential Reordering in Condensation Copolymers

ular weight (ie., the length) of the PET sequences, as compared to that of the starting PET (Figure 8b). This observation was supported by comparison of the intrinsic viscosity values of non-reacted homo-PET ([q] = 0.91 dl/g) and that of the PET sequences separated after the selective degradation of the thermally treated PET/PC blend ([q]= 0.09dl/g) [38]. The SEC curves in Figure 8 also indicate that without the application of selective degradation, this method cannot be used to obtain reliable data for the block lengths of any type. The SEC method, combined with preliminary selective degradation of the PC sequehces, was recently applied to follow the changes of the PET sequence length in different PET/PC copolymers [39]. These results will be discussed in detail in Section 5.3 of this chapter. What should be mentioned here is that the SEC method, combined with selective degradation, 'H and 13C NMR provides strong evidence supporting the occurrence of transreactions in thermally treated PET/PC blends. 2. Melting-induced sequential reordering in condensation copolymers

It is generally accepted that, in polyesters where the chains are terminated by hydroxy or carboxy groups, four types of interchain reactions are possible: (i) additional (post) condensation; (ii) exchange reaction between hydroxy and ester groups (intermolecular alcoholysis); (iii) intermolecular exchange between carboxy and ester groups (acidolysis) and (iv) transesterification (ester-ster exchange reaction) [2,11]. The reversibility of all these reactions is worth noting. In-depth studies by Devaux et al. [40] on poly(buty1ene terephthalate) (PBT)/PC blends show that direct transesterification is the most likely interaction mechanism in these blends (see also Chapter 3). As a result of direct ester-ster interchange reactions, the blend of homopolymers transforms into a block copolymer which subsequently turns into a random one: (A),

+ (B), +

* * *

(A)z- (B)v- (A),

* * *

+ . . .ABBABAABA

* * *

(1)

The block and the random copolymers formed are expected to exhibit enhanced mutual miscibility, compared to the starting non-reacted components [9,11,12,17,20,32,33,4143].Moreover, the transreacted chains probably also facilitate and accelerate the compatibilisation and interaction between non-reacted chains [12]. Therefore, it is important to control the interchange reactions in order to obtain a consistent product [44]. Devaux et al. [40] also demonstrated that the rates of the direct and reverse reaction are identical. This is quite understandable, taking into account that the bonding energies of broken and re-formed bonds are almost the same in both directions. Nevertheless, randomisation proceeds as a practically irreversible process (under given conditions) and eventually a

332

S. Fakirov, Z. Denchev

random copolyester obeying Bernoulli statistics is formed [40]. To explain this effect, one should bear in mind the well known relation between the equilibrium constant, K, and the changes in both reaction heat, AH, and react ion entropy, AS: RTln K = AH - TAS

(2)

In the case of ester-ester interchain reactions at high temperatures, usually above T, AH = 0 and randomisation is driven only by the large entropy increase originating from the transition of the block copolymer into a random one (see Eq. (1)). This results in the final randomisation of the blend and in a drastic change of its properties -crystallisation ability, solubility, transition temperatures, etc. In the attempt to explain some of these effects, we adopted the above-mentioned [32,33] general viewpoint that, with the progress of transreactions in the blend, the random copolymers formed should reveal decreased (if any) crystallisability as compared to one or both of the starting homopolymers. For this reason, the terms “random copolymer” and “randomisation” are used here to denote a copolymer unable to crystallise and the process of obtaining it, respectively. Accordingly, the terms “block copolymer” or “regeneration (restoration) of blocks” denote a copolymer displaying crystallisability, or the corresponding process. It is well known that IR spectroscopy [16-201 and ‘H and 13C NMR [7,18-20,22-241 are mostly used in the analysis of polymer microstructure, allowing a deeper insight into miscibility and chemical changes in polyester pairs. All these techniques, however, have limited application to blends constituted of PET and PBT because these two polymers are insoluble in almost all the solvents commonly employed for spectroscopic examination. This hampers the use of quantitative IR and NMR analyses for such samples [45]. Conflicting reports on the degree of miscibility of PBT and PC, or PET and PC, may be partly attributed to the lack of complete identification and of quantitative measurements of interchange reactions in these systems [12]. This is another reason for deriving conclusions concerning sequential order in condensation copolymers from their ability to cryst allise. The sequential reordering was followed in copolymers prepared by meltblending of the following homopolymers [46,47]:

333

Sequential Reordering in Condensation Copolymers

PAr tere/iso

= 1 :1

The above homopolymers were all thoroughly vacuum-dried at 100°C for 24 h. The PET/PC, PET/PAr, and PBT/PAr binary blends were produced by short melt-mixing and denoted as mechanical mixtures. Various feed ratios were used with some blends, so as to produce copolymers enriched in the crystallisable or non-crystallisable constituent. Generally, no additional amounts of transreaction catalyst were employed, unless otherwise specified. To follow the changes in the thermal characteristics of the blends with the progress of the transreactions, samples of each mechanical mixture were taken and subjected to thermal treatment directly in the DSC apparatus under a dry nitrogen atmosphere. The treatment cycle is shown schematically in Figure 9. Each sample was heated to the highest annealing temperature (280°C for the PET-containing samples and 290°C for the PBT-containing ones) and kept at this temperature for a given time period, followed by cooling to room temperature. A second heating to 280°C was applied when annealing was performed below Tg (dashed lines). The DSC curves were recorded only in a heating mode (Figure 9, the thick solid line). Similar treatment was given to neat PET, PC, PBT, and PAr controls, as well as to the blend with an increased amount of transesterification catalyst. Interchange reactions in blends of condensation polymers depend strongly on their initial compatibility and on the blending conditions [12]. These include temperature, duration of mixing, preparation method, viscosity match and the presence of catalysts [9,21]or inhibitors [48-511. Different conclusions have been suggested in the literature concerning the blend properties, simply because the blends studied have been prepared in different ways [12]. Sequential reordering in blends of immiscible and miscible partners will be discussed separately for the sake of clarity. 2.1. Melting-induced sequential reordering in condensation copolymers obtained from blends of immiscible partners In immiscible homopolymer blends, transreactions are expected to occur only at the interface between the different phases. Therefore, one may ex-

334

S. Fakirov, Z. Denchev

pect that in such samples the process of melting-induced reordering from block to random copolymers will be relatively slow, requiring relatively long annealing times [46]. The DSC heating curves for an equimolar PET/PC blend, thermally treated for different time periods at 280°C, are shown in Figure 10. Neat PC (Figure lOa), which is a typical non-crystallisable polymer, reveals a PC clear glass transition Tg at 149°C with no indication of crystallisation or melting processes. PET homopolymer (Figure lob) shows a well-defined crystallisation peak with a crystallisation temperature (T,) at about 140"C, followed by a melting peak with T, at 258OC. The PET glass transition temperature (TiET) is observed at about 70°C. The behaviour of the PET/PC equimolar blend changes with the progress of thermal treatment. Prolonged annealing at T, = 280°C, a temperature considerably higher than the Tm of PET, strongly hampers its crystallisation ability (Figure 1Oc-i). This effect is better seen in Table 2, where the changes in the degree of crystallinity (wc), T, and Tgare summarised for PET/PC blends of various composition as dependent on the annealing time at 280, 235, and 245°C. The heat treatment cycles shown * does ' : in Figure 9 by dashed lines were applied. The data indicate that T not seem to be sensitive to the annealing conditions, at least in this particular case (Figure 1Od-i, Table 2). At the same time, a well-pronounced depression is observed during annealing at 280°C. To exclude any doubt that the effects observed in Figure 10 are a result of physical consolidation [12] (ie., related to peculiarities in the crystallisation kinetics) rather than of interchange reactions, the cycle of thermal

TF

above T,,,

Annealing time

-

Figure 9. Schematic representation of the heating-annealing-cooling cycles used in the DSC measurementsfor studying the amorphisation,and regeneration of the crystallisability of blends of condensation polymers. DSC curves were recorded in heating mode only, by scanning the sample from room temperature (RT) to the highest annealing temperature (Ta= 280 or 29OoC), as shown by the thick solid line (Tm = melting temperature) [46)

335

Sequential Reordering in Condensation Copolymers

treatment (Figure 9, solid lines) was applied to neat PET and PC. The thermograms obtained are like those in Figure 1Oa for PC and 1Oc for PET. From these curves the values of T,, Tm,and wc are derived and shown in Table 2. From Table 2, one may conclude that the annealing conditions do not change the T,, T, and w, values for neat PET and PC. This observation strongly supports the assumption that the effects observed in Figure 10 are

n

-

475

'd7-

I

50

1

100

I

150

I

200

I

250

Temperature in

45 0

I

300 O

C

Figure 10. DSC curves of the equimolar PET/PC blend, taken in heating mode at 10°C/min, after various annealing times at 280, 235, and 245°C. Sample weight: 3.54 mg. Curves of the starting PET and PC are also given for comparison. Sample weights: 19.0 and 32.0 mg, respectively. The heating-ooling cycle denoted by the dashed line in Figure 9 was performed before obtaining the DSC traces [46]

336

S. Fakirov, Z. Denchev

not caused by crystalliiation kinetics during the thermal treatment. A further support of the last statement can be found in the experiments carried out with PET/PC blends differing in their composition, also presented in Table 2. The blends were subjected to the treatment shown in Figure 9 by the solid lines. These measurements were stimulated by the assumption that the lower the PET content in the blend, the shorter the randomisation time. On the other hand, PET/PC blends of molar ratios higher than 5.7:l.O should always be crystallisable, regardless of the duTable 2. Annealing conditions, melting temperature Tm,glass transition temperature T,,and degree of crystallinity w c for neat PET and PC and their blends with different compositions given as molar ratios; data derived from DSC curves taken in a heating mode, as shown by the thick solid line in Figure 9 [46] Annealing conditions

Transition temperatures Crystallinity

Temperature ("C) Time (min) Tg("C) Tm ("C) PET/PC = 100/0 280 280 235

0 165 260

280 280 280

0 300 600

280 280 235 235 245

0 240 100 300 200

wc (%)

78 78 78

263 259 258

37 35 36

75 80 80

250 239 237

38 27 26

90 92 80 82

250 254 223 223 228

PET/PC = 85/15

PET/PC = 70/30

-

37a

13 17 27

PET/PC = 50/50 280 0 250 39 280 45 78 232 37 280 75 80 228 22 80 228 11 280 105 280 135 80 228 8 280 165 80 225 5 280 265 82 218 5 235 30 82 235 60 235 120 200 21 235 240 196 21 200 26 235 360 245 30 223 31 245 60 223 38 245 180 223 38 After annealing at 2350 C and cooling to RT (Figure 9, the thin solid lane) 235 60 80 220 240 20' 235 260 74 220 247 378

337

Sequential Reordering in Condensation Copolymers Table 2 continued Annealing conditions

Transition temperatures Crystallinity Tg("C) T, ("C) ' w c (%) Temperature ("C) Time (min) PET/PC = 30/70 0

280 280 280 280 280 235 235 235 235 245 245

15 30 60 90 60 160 260 360 100 200

280 280 235

0 165 260

79 144 79 144 79 135 102 102 102 102 98 93 89 89

218 213 212 202

240 240 236 236

-

-

240 240 240 238 228 250 250

18 23 27 27 34a 34

-

-

PET/PC = 0/100

a

150 148 148

3Ba 338 22a 15a

-

wc is calculated on the basis of the two melting endotherms in the DSC curve

ration of the thermal treatment. This is because even at a high degree of randomisation, PET sequences long enough to crystallise will be present in the copolymer. The DSC traces for PET/PC blends of various compositions are shown in Figure 11A-C. In accordance with our expectation, the blend with the lowest PET content (PET/PC=30/70 mol %) requires the shortest time of 90min to obtain a product unable to crystallise (Figure 11A, curve (e)), against 240min for the 70/30 composition (Figure 11C, Table 2). Another peculiarity of the first blend, being the richest in PC, is the possibility of resolving its glass transition (T,Pc), which was impossible with the rest of the blends, containing less PC. Bearing in mind that the blend components are immiscible, the fact that TPETand Ti" get closer to each other (Figure 11A, curves (a)-(d), Table 23 and with the progress of annealing at 280°C finally merge into one Tg (Figure 11A, curve (e), Table 2) can be considered as an additional and independent proof of the occurrence of transreactions resulting in the formation of copolymers with more or less random sequential ordering [46]. The thermal behaviour of the blend richest in P E T (PET/PC=85/15, Figure 1l C) shows that even prolonged (up to 600min) annealing at 28OOC does not result in a complete loss of crystallisability, ie., in this case no amorphisation occurs, in contrast to all the other samples. Therefore, one may expect transformation of the two starting homopolymers into a block copolymer with longer PET blocks. This conclusion is derived from the PET PET changes in Tg . For shorter annealing times (up to 150min), Tg is fairly constant while for t , = 600min it increases slightly (Table 2, 85/15

338

S. Fakirov, Z. Denchev

composition, Figure 11C). In order to put an end to all doubt about crystallisation kinetics effects, an equimolar blend, containing additional amounts of tetrabutyl titanate, catalysing ester interchange reactions, was studied by DSC. A strong effect of the catalysts upon transesterification in PC-polycaprolactone systems has recently been observed by Shuster et al. (521, by Miley et aZ. [53] on PBT/PAr blends, and by Pilati et at. [5] on PET/PC blends. The DSC curve of an equimolar PET/PC sample containing 0.08% tetrabutyl titanate taken before annealing at 280°C is identical to curve (c) in Figure 10. A drastic shortening of the time required to obtain an amorphous (as revealed by DSC) product is registered - 60min against 265min for the same composition in the absence of catalyst (Figure lOi). Another approach to the qualitative estimation of the sequential order in copolymers is to test their solubility. The results from these experiments show a 100% solubility of the non-crystallisable (according to DSC) equimolar PET/PC sample. Bearing in mind the results of Devaux et aZ. [29], whose data concerning a 50/50 wt% PBT/PC blend demonstrated that a totally soluble sample corresponds to a degree of randomness of more than 53%, the solubility test can be considered as another proof of randomisation. Porter et aE. also accept that the copolymer solubility increases with the progress of the randomisation process [12]. In addition to the PET/PC blends, the same experiments on meltinginduced sequential reordering were performed with the PET/PAr blend also known to be immiscible [1,12]. The results of prolonged heating at 280°C followed by annealing at 235°C (see the treatment cycle denoted in Figure 9 by solid lines) are shown in Figure 12. Basically, this blend behaves in the same way as the previous system. In addition to the melting peak of PET, two well-resolved glass transition temperatures, TiETand T r , can be detected. At short annealing times at 280”C, the values of TiETand T,PA’ correspond to those of the neat homopolymers. A longer annealing time at Ta= 28OOC (Figure 12e) results in merging of the two PET-PA1 glass transitions of the homopolymers into one Tg . At the same stage the melting peak of PET disappears completely after a continuous decrease with the rise oft,. The observed complete amorphisation and the appearance of a single T’ET-pArat 110°C (Table 3) can be explained solely by transreactions leading at this high temperature to complete randomisation of the copolymer. This is in accordance with many recent reports (12,441. Previous communications of Porter et aE. [33,45] disclosed that heating at 280°C for over 16h yields a singlephase blend, displaying the same variations in Tg- and T,-values which are explained by randomisation. A peculiarity of the PET/PAr blend, compared to the PET/PC ones, is that much longer time is required for the obtaining of an amorphous (as re-

339

Sequential Reordering in Condensation Copolymers

100

50

150

-

200

300

250

Temperature in OC

,

After annealing time (min) at 245°C

e-

-1

d-

g; 0

a

PET-PC

?

T

<

I

After annealing time (min) at 235OC

300

0 After time (min) annealing

,

at 280°C PET

50

100

50

100

150

150

200

250

300

Temperature in OC

200

250

300

Temperature in OC

Figure 11. DSC curves of PET/PC blends taken in heating mode at 10"C/min, according to the cycle in Figure 9 (solid line). The blend compositions (molar ratios) are: (A) 30/70;(B) 70/30; (C) 85/15 [46]

340

S. Fakirov, Z. Denchev

Table 3. Annealing conditions, glass transition temperature T,, melting temperature Tm,and degree of crystallinity wc for neat PET and PAr and their equimolar blend. The data are derived from DSC curves taken in a heating mode, as shown by the thicker solid line in Figure 9 [46] Annealing conditions

Transition temperatures (“C)

Temp. (“C)

Time (mi.)

280

0

280 280 280 235

0 420 920 360

280

0

TgPET TgPA’

Tm

PET/PAr = 100/0 78

-

75 75

187 190

PET/PAr = 50/50 110 87

PET/PAr = 0/100 192

-

Crystallinity We

(%)

258

37

255 247 240

33 27 33

-

-

vealed by DSC) product - 920 min against 265 min for PET/PC under the same treatment conditions. The reason may be that PAr contains bulkier structural units than PC does, and therefore, due to steric hindrance, its structure is less favourable for chemical reactions. As long as the randomisation of the copolymers studied is performed at a temperature well above T,,,and T, of the homopolymers, an estimation of the weight losses of the homopolymers as well as of their equimolar blend proves to be appropriate, using the same annealing conditions as those in the DSC experiment. Moreover, as repeatedly reported for PET/PC and 360 min at

yTLET 1

0

.

1

50

.

1

100

.

1

150

.

1

.

1

.

l

200 250 300 Temperature in OC

Figure 12. DSC curves of an equimolar PET/PAr blend, taken in heating mode at 10”C/min, after various annealing times at 280°C and 235”C, according to the cycle denoted by solid lines in Figure 11. Sample weight: 7mg [46]

34 1

Sequential Reordering in Condensation Copolymers

Table 4. Weight losses of neat PET and PC and their blend (50/50by mol) under various annealing conditions, as revealed by TG analysis [46] Sample Composition Weight losses during designation linear temperature (wt%) increase (%)

From 50 to 280°C

PET PC PET/PC

100 100 50150

PET

100 100

PC

PET/PC

50/50

1.7 1.2 15.0

From 50 to 235OC 0.0 0.0 2.0

Weight losses during isothermal

annealing (%) At 280°C (t, =165 min) 0.8 0.5 6.0

At 235°C ( t , =260 min) 0.0 0.2 0.2

PBT/PC blends [7,9,12], decomposition under the above conditions may occur with evolution of low molecular weight products such as butylene and ethylene carbonate (EC). In order to demonstrate whether or not thermal degradation is responsible for the observed disappearance and reappearance of crystallinity, neat PET and P C samples and their equimolar blend have been studied by thermogravimetry (TG). Data derived from the T G traces are summarised in Table 4. The results in Table 4 show a rather different thermal stability of the neat homopolymers, as compared to their blend. Neat homopolymers lose less than 2% of their weight only during the first heating up to 280°C. During the following isothermal annealing for 165min at 280°C the weight losses are less than 1wt%. No weight losses are detected after the subsequent cooling and heating to 235"C, keeping the sample at this temperature for 260min. The results of the T G analysis of neat PET are consistent with the DSC measurements where no changes in the T,,,and wc values are found during the entire thermal treatment (Table 2, neat PET). Our observations disagree with those of Murff et al. [54], who report a significant drop in the heat of fusion of neat P E T after annealing for 30 min at 290°C. The equimolar PET/PC blend behaves differently. It loses 15wt% during the first heating up to 280°C and 6 wt% during the following isothermal annealing a t 280°C (Table 4). The observed differences in the behaviour of the neat homopolymers and of their blend can be explained with ester-ester interchain chemical reactions taking place in the blend and, as previously reported, resulting in evolution of volatile components (EC). However, the observed drop (between 2- and 10-fold) of melting enthalpy and crystallisation ability during annealing at 280°C cannot be attributed solely to weight losses in the blend that do not exceed 6 wt%. On the other hand, it should be mentioned that according to the findings of Berti et al. [9], the reported effective evolution

342

S. Fakirov, Z. Denchev

of EC in PET/PC blends, annealed within the 270-290°C temperature range, occurs in the presence of tetrabutyl titanate, which is not the case with our samples. The most reliable information concerning the sequential order in copolymers can be obtained by NMR analysis. Due to the limited solubility of polyesters and particularly of PET and PBT, in some cases the structural information cannot be obtained easily, as mentioned above and stressed by Porter [12]. These difficulties arise when the structure of the copolymer includes long PET or PBT blocks. Only more or less random copolymers are soluble in the typical solvents used for NMR analysis. Bearing in mind these peculiarities of the blend under investigation, an attempt was made to determine the sequential order of the PET-PC copolymer at the stage of randomisation, and of the copolymer with recovered crystallisation ability, where a blocky structure is expected. For the sake of comparison, a physical mixture of neat PET and PC was also analyzed. The resulting lH NMR spectra are displayed in Figure 13. A well-expressed difference between the spectra of the physical mixture and those of the copolymer at the stage of randomisation (Figure 13, spectra (a) and (b)) was detected. In the first spectrum, the four protons of the terephthalic residue give a signal at 8.02ppm. The peak does not reveal a fine structure because these protons are magnetically uniform. The signals in the 6.60-7.00ppm region correspond to the PC aromatic protons. Quite different is the spectrum of the amorphous (as revealed by DSC) blend: two new peaks at 8.30 and 8.20ppm appear and the multiplet in the 6.407.20 ppm region, reflecting the PC protons, is already deformed and has a disturbed fine structure. All this is a clear evidence that chemical reactions have taken place. 2.2. Melting-induced

sequential mordering in condensation copolymers obtained from blends of miscible partners

As repeatedly reported, there is evidence that exchange reactions depend strongly on miscibility [55].For this reason, it is of particular interest to follow them in a typical miscible polyester blend such as PBT/PAr. Kimura et al. [56] have studied PBT/PAr blends, before and after transesterification, by thermal and dynamic mechanical testing. Blends prepared by solution precipitation show a single T,, situated between those of the individual polymers and increasing with PAr content, which indicates amorphous miscibility of PBT and PAr. Melting point depression of PBT crystals is not observed and the melting point of PBT is almost constant up to high PAr contents. Also, PBT crystallises by separation from the amorphous miscible PBT/PAr phase. When blends are held at 250°C for 16h, transesterification occurs, as shown by the higher T,-values, compared to those of the respective physical blends. A considerable melting point depression and lower PBT crystallinity at higher PAr content are also noted. These results

343

Sequential Reordering in Condensation Copolymers

are consistent with a reaction in the polymer mixture, first to block copolymers and finally to random ones. New 'H, 13C NMR and IR evidence has recently been provided to confirm transesterification in these blends [16,57]. In addition, unusual X-ray diffraction results have been obtained with this blend [58]. The X-ray data are shown to be consistent with a nematic order of PBT in the blends, which is a cybotactic nematic structure, having both parallel alignment of chains and correlation of the centres of repeating units. By controlling the exchange reactions (ie., the sequential order) in another miscible blend based on PBT and PC, through an appropriate choice of the processing temperature and time, copolymers are obtained with mechanical properties higher than those of the corresponding physical blends [59]. The DSC heating curves after thermal treatment at 290°C for the 50/50wt% (62/38molar ratio, referred to repeating units) PBT/PAr blend

1

8.4

.

I

8.0

.

1 - 7

7.6

1

7.2

.

I

6.8

'

I

6.4

d in ppm from TMS Figure 13. 30GMHz 'H NMR spectra of an equimolar PET/PC system after

different heat treatments: (a) mechanical mixture of homopolymers; (b) after annealing for 165min at 280°C (randomised sample); ( c ) after annealing the randomised sample (b) for 260min at 235°C (restored blocks). Solvent: CFJCOOH CDCL; TMS standard 146)

+

S. Fakirov, Z. Denchev

344

at 140°C

8

PAr i

0

'

l

50

'

100

l

~

150

l

200

~

250

l

~

l

'

l

300

Temperature in OC

Figure 14. DSC curves of the PBT/PAr (50/50 wt%,or 62/38 molar ratio), taken in heating mode at 10°C/min, after various annealing times at 290°C and 14OoC. Sample weight: 7.62 mg. Curves of the starting PET and PC are given for comparison. Sample weight: 12.0 and 9.Omg, respectively [47]

are shown in Figure 14. A substantial difference in the crystallisation ability of the two blend constituents is observed. While PAr (Figure 14a) reveals only one transition at about 190°C with no indication of crystallisation or melting processes, PBT (Figure 14b) shows well defined crystallisation and melting peaks at about 130°C and 22OoC, respectively. In addition, PBT exhibits also a glass transition Tg at about 65°C. Figure 14c reflects the thermal behaviour of the 50/50wt% PBT/PAr blend prior to its high-temperature annealing. The Tgof PAr is not seen either in heating or in cooling modes (the cooling curves are not displayed) because the cold crystallisation peak of PBT overlaps with Tgof PAr, as observed also by Golovoy et al. for the same system [41]. The DSC curve of the starting homopolymer mixture indicates that during the blending procedure no observable transesterification has taken place in the blend, which is in accordance with previous reports [56]. During subsequent annealing of the blend for various times at a temPBT perature considerably higher than T, , a strong reduction of the crystalli-

345

Sequential Reordering in Condensation Copolymers

sation ability is observed (Figure 14d,e). These changes are summarised in Table 5. A pronounced depression of the PBT melting temperature is observed at intermediate annealing durations (60 and 160min), together with a two-fold decrease of w,. Thermal treatment at 290°C for 260min results in a complete disappearance of the crystallisability of PBT - melting peaks are no longer observed (Figure 14f). The only characteristic feature of the DSC curve is a single Tg at about 65°C. This can be regarded as an Table 5. Annealing conditions, glass transition temperature T,, melting temperature Tm,and degree of crystallinity wc for neat PBT and PAr and their blends. The data are derived from DSC curves taken in a heating mode, as shown by the solid lines in Figure 9 [47] Annealing conditions Temp. ("C) 290 290 290 290 290 140 140 140

Time (min) 0 60 160 260 360 100 300 400

Transition temperatures ("C)

T:', TgPA' PBT/PAr = 100/0 wt% 58 58 58 58 58 58 58 58

-

-

Crystallinity

T,

wc (%)

225 220 218 218 218 216 216 216

37 35 35 35 35 36 37 37

PBT/PAr = 90/10 wt% (94/6 molar ratio) 290 290 290 290 290 290 290 290 290 140 140 290 140 140 140

0 60 160 260 560

58

-

-

177 174 174

220 215 210 200 200

PBT/PAr = 50/50 wt% (62/38 molar ratio) 0 60 160 260 60 100 100 100 300

400

57 75 58 58 75 70 70 70 75 75

142 -

-

-

-

220 200 275 -

160 160 160 160 160

PBT/PAr = 0/100 wt% 290 290 290 290 290 140 140 140

0 60 160 260 360 100 300 400

-

-

-

-

-

192 192 192 192 192 192 192 192

-

-

-

35 36 28 26 26 36 35 16 0 11 24 0 20 20 20

-

-

-

346

S. Fakirov, Z. Denchev

indication that the thermal history described has caused quite advanced transesterification, resulting in a random copolymer, as generally accepted [6,10,18,25-311. Further evidence in favour of sample randomisation caused by transreactions in the melt can be found in the data of the homopolymer controls. Neat PBT and PAr samples were subjected to the same thermal treatment as the 50/50wt% blend. The DSC curves for PBT and PAr are almost the same as curves (c) and (a), respectively, in Figure 14 and remain unchanged during the total cycle of heat treatment. This can be concluded also from the data of the neat blend partners presented in Table 5. The changes that take place in the homopolymers during annealing are negligible compared to those in the PBT/PAr blend, where interchain reactions are possible. Apparently, these transreactions are responsible for the blend amorphisation. In favour of this statement is also the experiment with a blend containing much more PBT (PBT/PAr = 94/6 by mol, referred to repeat units, or 90/10 by wt) than the equimolar one. The shorter treatment of this blend at 29OOC does not lead to substantial changes in the crystallisation behaviour or in the corresponding parameters. However, when the annealing time increases to 560 min, significant changes take place. The new crystallisation parameters are typical of hindered crystallisation due to the involvement of PAr moieties in the PBT chains via transesterification. Another characteristic feature of this blend is that transreactions do not lead to complete amorphisation, as in the 50/50 PBT/PAr sample (Figure 14f). Obviously, the PBT-enriched sample contains sequences of PBT repeating units that are long enough to crystallise. Their average degree of polymerisation n should be above 15, as determined by the homopolymer molar ratio in the blend. Another proof in favour of randomisation are the data from the solubility test. They show a 100% solubilisation of the randomised, i.e., noncrystallisable (as revealed by DSC) 50/50wt% sample. This finding is in good agreement with the results of Devaux et al. [18], as well as with our results on solubility of the PET/PC blend after randomisation [46]. The weight losses of the 50/50 blend and of the neat homopolymers during the annealing cycles, as determined by TG, are summarised in Table 6. Unlike the previously studied PET/PC sample, the present blend is much more stable. This substantially increased resistance to thermal degradation of the PBT/PAr blends is to be related to differences in the chemical composition. It is worth mentioning here that the chemical structure of the partners in these blends does not allow elimination of low molecular weight alkylene carbonates as in the case of the PC-containing polyester blends studied [9]. There are other systems too, in which evolution of low molecular weight product is not to be expected. As already mentioned in Section 2.1, recent detailed studies on PET/PA 6 blends by means of different techniques showed that prolonged annealing above the melting temperature of PA 6

347

Sequential Reordering in Condensation Copolymers

and below the T,,,of PET, leads to the complete involvement of PA 6 in a copolymer with PET. This was proved by selective extraction and IR measurements [36] as well as by selective dyeing [37], while DSC [60,61], WAXS [62], and DMTA [35] studies demonstrated the complete disappearance of PA 6 as a separate phase after such treatment, ie., the polyamide component was entirely included in a PET-PA 6 copolymer. Most probably, the melting-induced sequential reordering should be affected by the specific chemistry of the interchange reactions in each particular case and by the chemical composition of the homopolymer blend constituents.

3. Cryst allisation-induced sequential reordering in condensation copolymers Taking into account the fundamental consideration that transreactions do not determine the type of sequential ordering and are just a means for its realisation, it seems reasonable to expect that if it is possible to create the necessary conditions in the random copolymer, one could govern the direction of the reversible transformation represented by Eq. ( l ) ,shifting it toward formation of blocks. When dealing with condensation copolymers, it might be possible to observe such a situation, when new factors appear in the system, acting more strongly than and in opposition to those causing randomisation to take place. A good example in this respect is crystallisation. It becomes effective when a randomised condensation copolymer is brought to a temperature below the melting of the crystallisable blend component. Quite important results concerning interchain reactions have been obtained in studies of the randomisation of PET with its deuterated analogue [13,14]. The investigations performed [13] have led to the interesting assumption that the course of exchange reactions could favour polymer Table 6. Weight losses of neat PBT and PAr and their blend (50f 50 wt% or 62f 38 molar ratio) under various annealing conditions as revealed by TG analysis [47] Sample Composition Weight losses during Weight losses designation (wt%) linear temperature during isothermal annealing (%) increase (%) From 50 to 290°C At 290°C (t, = 260 min) PET 100 0.50 1.20 PC

PET/PC

PET PC

PET/PC

100 50/50 100

100 50/50

1 .oo 0.45

From 50 to 140°C 0.00 0.00 0.20

0.75 2.25

At 140°C ( t ,

= 400 min) 0.00 0.00 0.20

348

S. Fakirov, Z. Denchev

crystallisation. The same idea was discussed somewhat earlier by Lenz et at. [6346]. Studying crystallisation-induced reactions of copolyesters in a heterophase solid-liquid system, the authors [66] noticed an unusual phenomenon, namely the formation of block structures upon heating random copolymers. The most effective reorganisation of a random copolymer into a block one has been observed with a sample containing 78.8% ethylene terephthalate units and 21.2% ethylene 2-methylsuccinate units. After a 30 h annealing at 220"C, this sample displayed a small but detectable increase of the number-average sequence length of the terephthalate blocks (from 5.3 to 6.8), a two-fold crystallinity increase and a 9°C rise in the melting point. Similar small changes of the same properties have been reported by these authors for liquid crystalline copolymers [66]. However, as emphasised by Muhlenbach et al. [67],such results concerning crystallisationinduced polymerisation should be interpreted with caution. Blend compositions close to 80/20 by mol (referred to as repeat units) in favour of the crystallisable polymer, applied in the above studies [63-66],could have limited the randomisation to sequences of 5-6 repeat units, thus affecting the reliability of the conclusions about crystallisation-induced reordering.

3.1. Evidence of crystallisation-induced reordering derived from the crystalline phase behaviour After the appearance of numerous unambiguous results demonstrating the transition of a block copolymer into a random one through transreactions, we looked for more convincing proofs of the existence of crystallisationinduced sequential reordering in the opposite direction, ie., from a random copolymer to a block copolymer, realised again by means of transreactions. Evidence in favour of the above effect may be found in the restored crystallisability of amorphous and presumably random PET/PC copolymers (Figures 10-12,14). Attention should be paid to the curves obtained after annealing of a non-crystallisable copolymer sample at temperatures below the melting point of the crystallisable component. Let us come back to the DSC traces of the three systems studied. The behaviour of the PET/PC blends during annealing at 235°C ( i e . , below the Tm of PET) deserves special attention. For all PET/PC compositions except 85/15 and the PET/PAr blend, almost complete regeneration of the crystallisation ability is observed (Figure 1Oj-n, Figure 11A, curves (f)-(k), Figure 11B, curves (c)-(e), Figure 12f). Regeneration is revealed by the trend to restoration of the initial w, and, in the majority of cases, of the Tm values, as can be concluded from Tables 2 and 3. The formation of larger blocks of both components is supported also by the behaviour of the amorphous phases. At the stage of complete randomisation only one amorphous phase exists, characterised by one glass transition temperature, PET-PAr T9PET-'0(Figure 11A, curve (e), Table 2) and Tg (Figure 12e, Table 3). Shortly after regeneration of the blocks, there is only one Tgin the PET/PC

Sequential Reordering in Condensation Copolymers

349

DSC curve (Figure 11A, curve (f), Table 2) which practically coincides with that of the amorphous PET/PC sample of the same composition (Figure 11A, curve (e)). Increasing the annealing time at 235°C results in lower PET Tg values, closer to Tg (Figure 11A, curves (g)-(i), Table 2). In the PET-PAr PET/PAr blend, the single Tg for the randomised sample (Figure PET-PAr PET-PAr 12e) decreases toward TiET,i e . , Tg (random) > Tg (block) > TiET.Such a behaviour of Tg is consistent with the assumption of block restoration processes. The same effect can be seen in Figure 11B, curves (b), PET-PC PET-PC of the random copolymer (curve (b)) is higher than Tg (c): Tg (block) (curve (c)). The behaviour of the 50/50 wt% PBT/PAr blend after amorphisation (Figure 14f) is also quite interesting. Annealing this sample at 140°C results in restoration of a melting peak (Figure 14g,h) and of more than half of the initial crystallinity (Table 5). What is more, this amorphisationcrystallisation cycle can be repeated with the same sample, as shown in Figure 14i-1. It should be noted that the second randomisation at 290°C takes half of the time of the first one - 100 against 260min (Figurel4f,i). This has to be attributed to the shorter block length, restored during the crystallisation at 140"C, as compared to the initial one. The subsequent regeneration of the crystallisation ability (Figure 14j-1) is characterised by PBT preservation of the same relatively low T, -value of 160°C and half of the initial degree of crystallinity (Table 5). While the randomisat ion of condensation block copolymers is well documented, well understood and generally accepted [12,44] this is not the case for the opposite direction of the reversible transition block i+ random copolymer. The regeneration of the crystallisation ability (expressed by the values of T,, Tgand enthalpy of melting AH or wc,respectively), as reported first by Lenx et al. [63-66] and observed by Miihlenbach et al. [67], by Economy et al. [68], by Potter et al. [69], and by us [46,47],could be explained at least by two approaches. The first one, discussed by Economy et al. [68] and supported by Porter [12] and by Pakula [70], does not assume the occurrence of chemical processes during crystallisation of the more or less randomised copolymer and therefore it can be called physical approach. In contrast, the chemical approach explains the regeneration of crystallisation ability as a result of interchain reactions during crystallisation, leading to the regeneration of the block structure. This is postulated by Len2 et al., supported by Economy et al. [63-66] as well as by Stamm [71] and accepted in [46,47].These two approaches are described in more detail below. In the physical approach, one assumes that the two homopolycondensates A and B interact under appropriate conditions, forming a copolymer with a very broad block length distribution, the amount of random copolymer being insignificant, as shown schematically in Figure 15. A transition

S.Fakirov, Z. Denchev

350

from Stage I11 to Stage I1 takes place via interchain reactions. During subsequent cooling below T,, crystallisation starts when a selection of crys tallisable blocks, i e . , those with degree of polymerisation n > 5-6, takes place. It is presumed in this approach [70] that the crystallisation process requires long times because of the hampered crystallisation ability of the block copolymer, as compared to the homopolymers. Thus, the experimentally observed regeneration of the crystallisation ability should be a result of a physical process - selection of blocks of an appropriate length that allows crystallisation. This approach excludes randomisation (Stage I, Figure 15); the experimentally observed loss of crystallisation ability (in this particular case after annealing at 280°C) is considered to be due to the relatively rapid cooling of the melt in Stage 11. The chemical approach assumes that during melting of the A/B blend, interchain reactions result in the complete randomisation of the copolymers, i e . , a transition from Stage I11 to Stage I occurs (Figure 15). However, the random copolymer obtained is also characterised by some block length distribution, which is not as broad as in the former case (Stage 11). One can assume that blocks with n = 5-7 have some frequency of appearance, as shown schematically in Figure 16. When the melt of the randomised copolymer (Stage I, Figure 15) is brought to a temperature somewhat below T, i e . , at a temperature enabling the crystallisation of one of the reacting species, crystallisation starts. To this purpose, nuclei are formed from blocks with n > 5. By the crystallisation (or “precipitation”) of blocks, they leave the phase where the reaction takes place (the melt). As a result their concentration in the melt is upset, as shown schematically by the dashed line in Figure 16, and the reaction proceeds to restore this concentration, i.e., the random copolymer is converted into a block one. As more blocks axe generated,

0

1

2

3

4

5

6

7

8

-

50 100 150 200

-Block length in repeat units

Figure 15. Probability distribution W, of the block length for copolymers built up of two condensation homopolymers, A and B, respectively. For Stages I, I1 and 111, see the text [46]

351

Sequential Reordering in Condensation Copolymers

this tendency progresses and more random copolymer is converted into a block copolymer. The existence of blocks with n > 5 (Figure 16) shifts the reversible transition to the “block” direction via transreactions at this particular temperature, so long as conditions suitable for the occurrence of chemical reactions are available. It should be noted that “precipitation” represents formation of crystallites which do not participate in the interchain reactions, or at least do not contribute to the transition from a random to a block copolymer or vice versa, for the following reasons. It is generally accepted that interchain reactions take place only in amorphous polymers or in the amorphous portions of semicrystalline polymers. Even if exchange reactions occur in crystallites, as assumed by Kugler et a!. [13] and demonstrated by Fakirov et al. [60-62], they take place only in crystal defects; these transreactions do not affect the sequential order of the copolymer when they axe in blends since they involve only one of the partners, i.e., either homo-A or homo-B. Since the physical and chemical processes described depend strongly on temperature and time, additional experiments under appropriate temperature conditions with preliminarily randomised amorphous copolymers could shed some light in the attempts to distinguish between the relevance of the physical and the chemical approach. Figure 17 shows the DSC curves of a non-crystallisable (as revealed by DSC) equimolar PET/PC blend recorded in a heating mode after annealing this sample at different temperatures. One can see that annealing of the sample at 280°C for 165min results in almost complete amorphisation and presumably randomisation of the copolymer (Figure 17a). Curve (b) in the same figure is obtained after additional annealing of this sample at 235OC for 260 min and cooling to room temperature, before it was scanned to 280°C. Two well-resolved melting peaks are observed. If the same ran-

0

1

2

3

4

5

6

7

8

9

10

11

Block length in repeat units

Figure 16. Frequency distribution F, of the block length for copolymersfrom two condensation homopolymers, A and B, respectively [46]

352

S. Fakirov, Z. Denchev

1

50

'

1

100

.

l

150

.

l

200

'

I

250

'

I

300

Temperature in OC

Figure 17. DSC curves of an equimolar PET/PC blend, completely randomised at 280"C, taken in heating mode at lO"C/min, after: (a) annealing at 280°C for 165min; (b) annealing at 235°C for 260min, followed by cooling to room temperature; (c) annealing at 235°C for 260min followed by heating to 280°C and immediate cooling to room temperature; (d) sample (c) during the heating from 235°C to 280°C; (e) annealing at 165°C for 300min followed by cooling to PET PET PET room temperature. Tg ,T', ,T", :glass transition temperature of PET and melting temperatures of the less and more perfect PET crystallites, respectively [461

domised sample after the same annealing time at the same T, = 235°C is first heated to 280°C before its immediate cooling to room temperature, the DSC curve taken during the subsequent heating to 280°C looks different (Figure 17c); only one melting peak is observed. As to its shape and temperPET ature position, this peak is quite similar to the lower-melting peak T', in the previous experiment (Figure 17b). Comparison of these two last thermograms leads to the important conclusion that it is the chemical a p proach that has to be preferred in the explanation of crystallisation-induced reordering, leading to restoration of melting in the copolymer sample. This conclusion is supported also by the next two curves (Figure 17d,e). Taking into account the essential assumption in the physical approach that the crystallisation takes long times for selection of crystalline blocks in a copolymer characterised by a very broad block length distribution (Figure 15, Stage II), another DSC measurement was performed. Obviously, all crystallites created during annealing at 235°C completely disappear when the sample is heated from 235°C to 280°C. The only time left for a new crystallisation to take place is during subsequent cooling to room temperature (at lO"C/min). This time is supposed to be too short for a

353

Sequential Reordering in Condensation Copolymers

selection of crystallisable blocks to occur, according to the postulates in the physical approach. This could only mean that, during annealing at 235"C, the number of crystallisable blocks is increased via transreactions, these blocks being able to crystallise even during the short cooling time, as can PET be concluded from the appearance of a well-shaped melting peak TI, (Figure 17c). These crystallites should be rather imperfect and therefore characterised by a low T,. If the crystallites grow isothermally at T, = 235°C for a longer crystallisation time ( t , = 260min), their T, should be PET

considerably higher, which is actually observed (curve (b), TI', ). The result displayed in Figure 17d leads to the same conclusion. It reflects the melting behaviour of crystallites created at 235°C (without cooling the sample from 235°C to room temperature). As expected, the PET melting peak is closer to the higher melting peak TI', of curve (b). In the next experiment, the same randomised sample (Figure 17a) is annealed at 165°C long enough (t, = 300min), in an attempt to invoke crystallisation. The temperature selected is very appropriate for crystallisation of PET [72] and at the same time is not high enough to allow the occurrence of chemical reactions [3,73]. Thus, if crystallisable blocks were available, as assumed in the physical approach (Figure 15, Stage II), under these conditions, so favourable for crystallisation of PET, crystallites should have appeared. The DSC curve of this sample, however, does not show any sign of melting (Figure 17e). This experiment strongly supports the concept of the chemical approach. One has to accept that in the amorphous sample crystallisable PET blocks are not present (Figure 15, Stage I). They appear in a sufficient amount only as a result of crystallisationinduced sequential reordering when the randomised sample is brought to a PET T, below T, but high enough for chemical reactions to take place. With the progress of interchange reactions, leading to randomisation, a decrease in the PET melting temperature is observed in all the blends studied, and finally a complete loss of their crystalline features is reached. This corresponds to a general decrease of the system ordering, eventually attaining complete sample amorphisation. (Tables 2,3,5). At the same time, the opposite tendency to restoration of the system ordering is visualised by the reappearance of melting endotherms on prolonged heating at 235°C (in some cases at 245°C as well), as seen in Figures 10-12 and 14, and Tables 2, 3, and 5. The behaviour of the two initial glass transition temperatures also supports the concept of restoration of the blocks via transreactions; after randomisation, they merge into one Tg and then, upon regeneration of the PET blocky structure, this single Tg shifts to lower values, approaching Tg (Figure 11A, Table 2). Such a tendency can be explained by the formation of larger amorphous areas that are more homogeneous in chemical composition, presuming the existence of longer sequences of the same repeat

354

S. Fakiiov, Z. Denchev

units. The experiments with increased amounts of catalyst have to be interpreted in accordance with the chemical approach, too. The reason for the effects observed can be that the blocks restored via crystallisation-induced sequential reordering are relatively short (5-6 repeat units). This proves that it is the chemical interchain reactions which are responsible for the decrease in crystallinity. It should be stressed again that the lower the PET content, the shorter the time required for amorphisation or block restoration (Table 2). This observation is also in favour of the chemical approach. As mentioned above, in all blends studied (with the exception of the PET/PC 30/70 composition), the glass transition temperature of PC cannot be resolved. This could be due to an insufficient amount of PC and/or to the presence of some crystallinity in the samples. It has been suggested in the literature [74] that in such cases the glass transition temperature of PC may sometimes be hard to detect since crystallinity can restrict the mobility of the amorphous phase and suppress the increase in C, at Tg. This phenomenon, called in some cases the “rigid amorphous phase”, has also been observed by Cheng [75]. In accordance with the finding of Nassar et al. [76]who have shown that PET/PC blends containing 60wt% PET or less reveal a distinct Tgpc,we also have observed a very well-resolved glass transition in the blend with the highest PC content (Figure 11A). The behaviour of during the randomisation and particularly during the block restoration cycles deserves some attention, too. As expected, during randomisation T, and the crystallinity, w,,decrease. While regeneration of blocks results in a stronger tendency to complete restoration of PET crystallinity, the value of T, remains low for shorter annealing times (Table 2). A similar behaviour was reported quite recently by Potter et al. [69]. for a phydroxybenzoic acid/2,6-hydroxynaphthoicacid random copolymer system annealed well below the crystal-to-nematic transition. In the interpretation of the lower T, values after block restoration, one should take into account the following considerations. It is a well established fact that in the case of homopolymers T, increases with the rise of T, or t, (at T, = const). For PET this phenomenon has been thoroughly studied by Fischer et al. [77] and Fakirov et al. [78]. In homopolymers, the lamellar thickness I , depends on the crystallisation conditions alone, while the perfection of crystallites depends also on the chemical and structural perfection of the chains. In many cases the last two factors limit the lamellar thickness and strongly affect the perfection of the lamellae. The best example in this respect are the copolymers with relatively short crystallisable blocks (containing usually between six and ten repeat units). Combined WAXS and SAXS measurements of poly(ether ester) thermoplastic elastomers containing PBT as hard (crystallisable) segments show that the lamellar thickness 1, reaches about 50A, regardless of the treatment conditions. What is more, dealing with the same copolymers, it is

TF

Sequential Reordering in Condensation Copolymers

355

found that the usual rise in the long spacing L with the increase of T, and t, is due to expansion of the amorphous intercrystalline regions [79]rather

than to crystal thickening. Crystallinity (Table 2) shows an increase, and grows during subsequent cooling as well, contributing to the increase in wc.This can be concluded from the two PET melting peaks appearing in some cases (Figure 11A, Table 2, Figure 17b). It is interesting to note that the low melting peaks TI, are slightly below the annealing temperature T, of 235OC, ie., the less perfect population of crystallites appears during the cooling stage. Obviously, they consist of PET sequences of less than the typical 5-7 repeat units and, regardless of their much lower T,, they contribute to the rise in W C

-

The higher melting peak TI‘, of the blend PET/PC = 50/50 by wt in the mechanical blend and T, of the 30/70 blend correspond to (Figure lOc, Table 2). This population of more perfect crystallites arises during annealing at T, = 235°C and presumably comprises PET blocks longer than those forming crystallites that display the TI, values and are in the molten state at T, = 235°C. The behaviour of the second system, PET/PAr = 50/50mol%, in respect to TI,, TI’, and w c (Figure 12, Table 3) is quite similar to that of PET/PC. Since both PET/PC and PET/PAr blends are immiscible, interchain reactions can take place only at the interfaces rather than in the bulk. The observation of Nassar et al. [76] that the amorphous regions in PET/PC blends containing between 10 and 60wt% PET are phaseseparated, with a PET-rich and a PC-rich phase, does not change the situation significantly. It only implies that a small amount of a polymer dissolved in another polymer phase has more favourable opportunities for chemical reactions. As mentioned in Section 2.1, a convincing proof of the occurrence of transreactions, leading to new sequential ordering in the system, is given by the NMR analysis of the copolymer (Figure 13a,b). Unfortunately, the differences between the spectra (b) and (c) (taken at the randomisation stage and after block regeneration, respectively) are not large enough to enable distinction between a random and a blocky copolymer structure. The reasons for these results are related to the peculiar solubility of PET and PBT, as stressed by Porter [12]. In the case of random copolymer, the whole of the sample was subjected to analysis because of its complete solubility, whereas in the “blocky” one only the soluble part was analysed, the latter being more or less random, as observed in practice. Failing to obtain a direct structural confirmation of the proposed block restoration by means of NMR, we looked for another indirect proof. The viewpoint was adopted that random copolymers should have strongly decreased (if any) crystallisability, as compared to the starting crystallisable homopolymer blend constituents [33]; this was also concluded from our

TF

356

S. Fakirov, Z. Denchev

DSC studies. On the contrary, regeneration of blocks in the copolymers should result in restored crystallisation ability with measurable Tm, w,, and Tg.This enables the implementation of the WAXS technique as another appropriate method. Figure 18 shows WAXS curves of neat PBT, neat PAr, and their 50/50 wt% blend at different stages of annealing, i. e., with different degrees of sequential ordering. Curve (a) shows the typical amorphous structure of PAr. Neat PBT is characterised by five strong reflections positioned in the 28 = 15-30" range (Figure 18b). In curve (c) of the PBT/PAr mechanical mixture the same five reflections remain, being however much weaker, obviously due to overlapping with the amorphous halo of the PAr component. The curve of the block copolymer (d) looks the same as curve (c), most probably because the PBT blocks are long enough to crystallise. Taking into account the DSC results, one may conclude that curve (d) reflects the structure of a block copolymer. Curve (e) in Figure 18 is obtained from an annealed PBT/PAr blend

I

10

,

I

20

I

30

1

40 28, degree

Figure 18. WAXS diffraction patterns of PBT (a) and PAr (b) homopolymers and of their 50/50 (by wt) blend after Merent heat treatments: (c) mechanical mixture; (d) 160 min at 29OoC(block copolymer);(e) additional 260 min at 29OoC (randomised sample); (f) after annealing the randomised sample (e) for 160min at 14OoC(restored block copolymer) [47]

Sequential Reordering in Condensation Copolymers

357

not revealing (by DSC) any crystallisation or melting ability. WAXS measurement also indicates an entirely amorphous structure in this case. In our opinion, this is an unambiguous proof of random distribution of the PBT and PAr units along the chain axis which excludes the presence of PBT sequences long enough to crystallise. Of particular importance is curve (f) in Figure 18, taken from a sample with regenerated crystallisability (as revealed by DSC). It looks exactly like the curve of the PBT/PAr block copolymer (d), but is obtained by additional thermal treatment of the random copolymer. For this reason, one can assume that this sample represents a copolymer with restored PBT blocks of a length that enables crystallisation. Finally, the explanation proposed by Kugler et al. (131 of the earlier PET reported extremely high T, is in favour of the statement that the observed effects (Figures 1CL12, and 14) are results of sequential reordering in the copolymers studied and do not originate from peculiarities in the crystallisation kinetics under the conditions applied. These authors report an unusually high T, of PET after extremely long annealing of several months. Kugler et al. [13] explain this observation by crystal thickening via transesterification as concluded from their small-angle neutron scattering experiments in blends of protonated and deuterated PET. Figure 19 illustrates schematically the transition from a random into a block copolymer, suggesting that crystallisation is the driving force for sequential reordering after cooling a randomised blend to a temperature below the T, of the crystallisable component. It is quite evident that sequences of 2-3 repeat units are not long enough to form a lamella of thickness 1,. The lowest values of 1, are about 50-60 A, as mentioned above and demonstrated for PBT copolymers [79].Assuming these values of 1, to be the lower limit and taking into account that the length of the PET repeat unit in the chain-axis direction is about 11 A, it is easy to conclude that crystallisation of PET can occur only if PET blocks of at least 5-6 units are available. This axiomatic requirement can be considered as the basic driving force toward the blocky structure when crystallisation conditions

Random copolymer

Block copolymer

Figure 19. Schematic representation of the transition from random to block copolymer taking place via transreactions under the influence of crystallisation (461

358

S. Fakirov, Z. Denchev

are established. This model is easy to explain by the chemical approach. Concerning Stage I in Figure 15, it was already supposed that along with the predominant sequences of 2 4 units of PET, blocks of 5-7 repeat units can also PET be available. Cooling the randomised copolymer melt to a T, below T, creates conditions for crystallisation of these longer sequences. After keep ing them for some time at this Ta,they form crystallisation nuclei. For the further growth of crystallites, new supplies of longer blocks are necessary. They can be formed only via transreactions, if those previously available (from the randomisation process) are exhausted (Figure 16). The model depicted in Figure 19 demonstrates another important peculiarity of crystallisation-induced sequential reordering. Formation of lamellae of thickness I , not only stimulates sequential reordering in favour of the blocky structure, but also restricts the block length to the value of 1,; so one can speak in this case of "microblock copolymers" with quite uniform block length distribution.

n

T i c = 139OC b TOPc = 157OC 8

0

50

TOPET= 88OC

100 150 200 Temperature in O C

Figure 20. Temperature dependences of the loss modulus El' (in bending mode) for neat homopolymers, their equimolar mixture, and copolymers with different sequential order: (a) neat PET; (b) neat PC; (c) mechanical mixture; (d) block copolymer; (e) randomised sample; (f) restored block copolymer [80]

359

Sequential Reordering in Condensation Copolymers

3.2. Evidence of crystallisation-induced reorderkg derived fi.om the amorphous phase behaviour At this point, the crystallisation-induced reordering phenomenon was considered, mostly as revealed by the properties of the crystalline phases. With the PET/PC system, the DSC method and instrumentation used allowed us to follow the behaviour of the amorphous phases as a function of the crystallinity changes only in the blend compositions with predominant concentrations of the non-crystallisable component (Figure 11A). On the other hand, it is well known that the changes in Tg of polymer blends can characterise their miscibility [1,12]. Therefore, the changes in Tg values with the progress of transreactions in the PET/PC system were followed, using a more sensitive technique (DMTA), in order to get further evidence for the direction of these transreactions in the system under investigation [80]. To this purpose, PET and PC homopolymers were melt-mixed in a Brabender mixer under the conditions indicated in Table 7, to obtain a mechanical mixture as well as block and random copolymers. The restored block copolymer was produced by prolonged annealing of a random copolymer sample. Figure 20 depicts the temperature dependences of loss E" modulus (in bending mode) for PET and PC homopolymers,their equimolar mechanical mixture as well as for three PET/PC copolymers with various block lengths. All data derived from these curves are tabulated in Table 8. One can see that the neat homopolymers are distinguished by very well shaped loss modulus maxima, reflecting their glass transition temperatures at 88 and 157°C for TiETand Ti",respectively (Figure 20a,b, Table 8). The curve of the equimolar mechanical mixture (c) represents a superposition of (a) and (b), the Tg values of the blend components being slightly displaced from those of the neat homopolymers - now the TgpET and T,Pc appear at about 90 and 139"C,respectively, ie., the maxima are getting closer to each other. This trend becomes better expressed in the sample with the expected blocky structure (Figure 20d, Table 8), since the glass Table 7. Conditions of preparation of PET/PC samples by means of melt-blending in a Brabender Plasticorder PL 2000-6 mixer Sample Thermal treatment Temperature ("C) Duration (min)

PET/PC PET-PC PET-PC (mechanical (block (random mixture) copolymer) copolymer) 280 5

280 60

280 100

PET-PCa (restored block copolymer) 235 100

a The copolymer with restored blocks is obtained from the randomised sample under the conditions indicated

360

S.Fakirov, Z.Denchev

transition temperatures of the starting components get even closer, the bimodal character of the curve still being obvious. Bearing in mind the results of our previous investigations on the same (461 and other (46,471annealed blends of condensation polymers, it can be assumed that the PET/PC copolymer produced is distinguished by relatively long blocks of identical repeating units, originating from the starting homopolymers. The two types of blocks form spatially separated amorphous phases, characterised PET by individual glass transition temperatures, Tg and Tpc, respectively. In comparison with both neat homopolymers and the mec\anical mixture, their values are closer to each other (Figure 20a-d, Table 8). Curve (e) in Figure 20 corresponds to the sample with a supposed random sequential ordering. Similarly to the neat homopolymers (Figure 20a,b), in this case one observes only one peak with Tg= 100°C, situated PET relatively close to Tg . The detection of a single glass transition between those of the starting homopolymers is a proof of miscibility [1,12]. It is well known that PET and PC homopolymers are immiscible (see also curve (c) in Figure 20, characterising their mechanical mixture). Therefore, one has to assume that transreactions result in significant changes in the chemical composition of sample (e), preventing the formation of separate phases of either PET or PC. In a copolymer comprising immiscible partners, this effect could be ascribed to a more or less random distribution of PET and PC residues along the chain axis. This conclusion is in good agreement with Table 8. Values of the loss modulus E", storage modulus E', loss factor tand, glass transition temperature T,, and viscosity 71, as obtained from DMTA measurements on neat PET, neat PC, and their mechanical mixture as well as on equimolar copolymers with various sequential orders Sample PET-PC Neat Neat PET/PC PET-PC PET-PC PET PC (mechanical (block co- (randomco- (restoredblock mixture) polymer) polymer) copolymer)

E", MPa (at 25°C) 1.34 1.75 0.47 0.43 0.74 0.63 Tg("C) from E"

0.82

0.54 0.44

100 5.25

96 124 3.89

0.16 1.44 0.10 0.65 0.12 0.43

0.98

0.13 0.58

102 88

peak position E', MPa (at 25OC) Bending tan b peak height Tg("C) (fromtand peak position) T, ("C) (from E') Viscosity Q* (Pas) at 28OoC and frequency0.1rad/s

88 157 15.1 8.32

162 160

30.3 185.0

90 139 8.91

107 132 8.91

101 148 92 144

125 123

108 100

105 124 94 125

20.5

27.0

2686.0

27.3

Sequential Reordering in Condensation Copolymers

361

the behaviour of the crystalline phase of P E T (Figure l0ij). Finally, Figure 20f demonstrates the temperature dependence of the loss modulus E" of the sample with supposedly restored blocky structure as a result of crystallisat ion-induced sequential reordering. This curve has two Tgvalues at 96 and 124"C, although they are not so well resolved as in curves (c) and (d). Nevertheless, the similarity is quite obvious between the three curves, characterising the mechanical blend, the block copolymer, and the sample with restored blocks, respectively. Starting from a more or less homogeneous amorphous phase which is typical of a random copolymer (Figure 20e), crystallisat ion-induced reordering obviously leads t o the formation of two amorphous phases that are spatially separated and large enough to have their individual Tgvalues. These conclusions are also supported by the temperature dependence of the storage modulus El (in bending mode), shown in Figure 21. The values of the storage modulus E', taken a t 25"C, are shown in Table 8. As seen in Figure 2la,b,d,e, neat PET and PC, and the block and random copolymers, reveal only one transition. While the monomodal course of curves (a), (b) and (e) can be expected and is easy to explain by the existence of only one amorphous phase, this is not the case of curve (d) of the block copolymer, where the second transition remains unresolved.

TiET = 94OC T i c = 125'C TiET-PC= 100°C

TSPET-PC= 123OC

TiET= 92OC T i c = 144'C T i c = 16OoC TiET = 88OC Temperature in OC

Figure 21. Temperature dependence of the storage modulus E' (in bending mode) of neat polymers, their equimolar mixture and copolymers with different sequential orders: (a) neat PET; (b) neat PC; (c) mechanical mixture; (d) block copolymer; (e) randomised sample; (f) restored block copolymer (801

362

S. Fakirov, Z. Denchev

At the same time, the equimolar mechanical mixture (Figure 21c) and the supposed copolymer with restored blocky structure (Figure 21f) show two clear transitions. In both cases the second, sharper, transition originates from the PC-rich amorphous phase, while the first, smoother, transition refers to the PET-rich amorphous phase. The similarity between the shapes of the curves of the neat homopolymers (Figure 21a,b) and that of the random copolymer (Figure 21e) supports the assumption that in these cases one deals with a single homogeneous amorphous phase. This means that in the sample expected to have random sequential ordering (Figure 21e), transreactions under appropriate conditions do actually result in a more or less complete randomisation. The shape of curves (c) and (f), characterised by two transition regions, is in favour of the coexistence of two chemically different and spatially separated amorphous phases, giving rise to two glass transition temperatures. ') Apparently, they belong to a PET-rich amorphous phase (with a lower T and a PC-rich amorphous phase (with a higher T'). Such a behaviour is typical of a mechanical mixture and of a copolymer with blocky structure. This means that the sample illustrated by curve (f)) has to be a block copolymer with blocks restored as a result of crystallisation-induced sequential reordering via transreactions that take place under the new set of treatment conditions (Table 7). The dependences of the viscosity q* at 280°C on the frequency for the homopolymers, their mechanical mixture, and their equimolar copolymers with various sequential orders are shown in Figure 22. One can see that the curves reflecting the viscosity of neat PET (a), of the mechanical mixture (c), and of the blocky copolymers (d) and (f) almost coincide. The curve for neat PC is of the same type, but it is slightly shifted to higher viscosity values. Completely different in position and slope is the curve characterising the random copolymer (e). This result clearly demonstrates the different behaviours of copolymers with random and with blocky sequential order. Discussing the above DMTA results, it is worth noting that one has to be certain that a really random (i.e., monophase) PET/PC copolymer is used as the starting material for the preparation of a sample, characterised by reappearance of two amorphous phases - PET-rich and PC-rich, i.e., a restored block copolymer. For some reasons (e.g., improper test conditions), the resolution of the two amorphous phases might be suppressed even in a PET/PC block copolymer with long homopolymer sequences. If such a material is subjected to prolonged annealing at 235"C, the reappearance of two glass transitions would be the result of physical consolidation, rather than of transreactions that cause crystallisation-induced sequential reordering. In such a case the effect of restoration of two separate phases would be a result of enhanced PET crystallisation at the longer annealing times applied, i e . , it would not be a structural rearrangement within the polymer chains but a morphological rearrangement of existing polymer chains. This problem was discussed thoroughly in the previous section.

363

Sequential Reordering in Condensation Copolymers

Let us summarise briefly the DMTA data that demonstrate the randomisation of the PET/PC blend. The first fact is that T:,' and Tic PET-PC merge into one Tg (Figures 20 and 21, curves (c)-(e), Table 8). What PET-PC is more, the value of the single Tg (determined from the tan6 temperature dependence) is very close to that calculated according to the widely used empirical Fox equation [81] 1/Tg = Wl/Tglf

(3)

W2/Tg2

Here, Wi is the weight fraction of component i and Tgi is the Tg of PET-PC component i. So, Tg (calc) = 111°C; which is very close to the experPET-PC = 108°C (Table 8). imental value of Tg The merging of TiETand Tgpcreflects the drastic decrease of the length of these sequences. This does not simply mean miscibility of existing PET and PC amorphous phases but their disappearance as separate structures due to formation of a PET/PC copolymer built up of consecutively linked PET and PC sequences, comprising 2-5 homopolymer residues. The second and more illustrative observation demonstrating the formation of a random copolymer from the PET/PC blend is the result of viscosity measurements (Figure 22, Table 8). Only the sample supposed to be random reveals a drastic increase in viscosity for the frequency of O.lrad/s - by a factor of 100; for the rest of the samples (Table 7) the viscosity values are almost the same over the entire frequency range.

I

0.1

1 .o Frequency f in rad/s

I

10.0

1

Figure 22. Dependences of viscosity q* on frequency at 280°C of homopolymers, their equimolar mixture, and copolymerswith different sequential order: (a) neat PET; (b) neat PC; (c) mechanical mixture; (d) block copolymer; (e) randomised sample; (f) restored block copolymer [80]

S. Fakirov, Z. Denchev

364

The following facts are noteworthy to prove restoration of the blocky structure of a random sample via transreactions: (i) the PET/PC sample revealing a reappearance of two amorphous phases is obtained by controlled annealing of a monophase random copolymer and (ii) this same sample behaves similarly to the block copolymer or even to the mechanical mixture of the same components. Therefore, based on the above DMTA data, one can conclude that the expected blocky sequential order is actually restored. Additional support for this statement can be found in the reappearance of two separate glass transition temperatures (Figures 20f and 21f, Table 8) after annealing a sample having a single T, (Figures 20e and 21e, Table 8). Summarising the results of the crystallisation-induced sequential reordering, one can conclude that the entire cycle “homopolymers+ block copolymer + random copolymer + block copolymer” seems to be realised with two binary immiscible and a miscible blend of condensation polymers. 4. Miscibility-induced sequential reordering in condensation copolymers obtained from miscible and immiscible partners

Crystallisation-induced sequential reordering is only one of the possible cases in which the sequential order in condensation copolymers is changed. It was worth looking for copolymer systems where the transition from a more or less random copolymer to a blocky one is governed by factors not related to crystallisation.

4.1. Background One can design an equimolar terpolymer in a two-stage preparation process. Let us have three condensation homopolymers (A)n, (B), and (C),; the first two, (A), and (B),, are immiscible, while (B), and (C), are miscible. During the first stage of preparation, randomisation in the A/B blend takes place: (A),

+ (B), +

*

*

ABBBAABBABBAABB* - ., (AB),,,

(4)

After completion of the randomisation, the copolymer (AB), is meltblended with the third homopolymer (C), and further randomisation occurs, resulting in a random terpolymer: (AB),

+ (C), +

. AABCCABBACBCC * *

* * *

, (ABC),,,,,

(5)

This randomisation takes place in the melt via interchain reactions at temperatures above the melting points of all crystallisable components [12291. If the randomised terpolymer (Eq. (5)) is kept for a longer time in the molten state, one can expect changes in the sequential order in favour of the blocky structure: (ABC),

+*

*

.AAAAABBBBBCCCCC. . . ,-. *(A), - (B), - (C);

*.

(6)

Sequential Reordering in Condensation Copolymers

365

This expectation is based on the miscibility of the homopolymers (B)m and (C), . Since miscibility requires interaction between longer polymer chains, a tendency toward the formation of longer B- and C-sequences should exist. If one takes into account that transreactions frequently occur in the melt, the formation of blocks (Eq. (6)) seems to be very probable. In order to check the possibility of restoration of the blocky structure in a random condensation copolymer using the miscibility factor (i.e., to realise miscibility-induced sequential reordering), three homopolycondensates are applied, two of them being miscible. The polymers used are: PBT, polyamide 6,6 (PA 66), PC and PAr, PBT and PAr being miscible in any composition [82]. Similarly to the cases described above [46,47], the conclusions about sequential order are based exclusively on DSC data, assuming that sample amorphisation and observation of a single Tgare results of randomisation, and that regeneration of the crystallisation ability is due to the restoration of blocks. One more reason to focus attention on only the DSC traces is that the literature data on solubility of block and random copolymers are not consistent. Moreover, the application of NMR techniques is impossible in this case because of obvious difficulties arising from the very complex structure of the systems studied and from the peculiar solubility of the polymers to be mixed [121. 4.2. Experimental observations

DSC curves for the PBT/PC/PAr system are shown in Figure 23. The two non-crystallisable homopolymers PC and PAr are solution-blended and the solvent is removed. The blend is then kept for various times at 240"C, a temperature well above the Tg of both components (Figure 23b-f). After each annealing time the sample is cooled down to room temperature and the DSC curve is taken by scanning to 280°C. This cycle is repeated until evidence appears for sample randomisation, i. e., the two separate glass PC transitions Tg and Tg" in curve (c) of the unannealed PC/PAr blend PC-PAr in curve (f) of the same sample after 300min merge into a single Tg at 240°C. It is a common practice to consider such a behaviour of the Tg as an indication of formation of a block copolymer which transforms into a random one when a single Tg is observed due to transreactions (121. This observation is in a good agreement with the findings of Kimura [25] and Porter [83] for PBT/PAr blends. The authors came to the conclusion that the PBT/PAr random copolymers are also amorphous, with no tendency toward crystallisation being observable by thermal analysis. The reaction between P C and PAr, both polymers containing bisphenol A units, has been simulated by Devaux et al. [84], studying the reaction between PC and diphenyl terephthalate or isophthalate, used as model compounds of a poly(ary1carbonate). They came to the interesting conclusion that transreactions occur without any catalyst. In our case, there is

S. Fakirov, Z. Denchev

366

50

100

150

200

250

300

Temperature in OC

Figure 23. DSC curves of the PBT/PC/PAr blend, taken in heating mode at 10°C/min, after various annealing times at 240°C and 28OOC. Sample weight: 45mg. Curves of the starting neat PBT, PC, and PAr are given for comparison [821 no doubt that after annealing for 5 h at 240°C complete randomisation of the equimolar PC/PAr copolymer is achieved. At this stage the sample is removed from the DSC pan and solvent-mixed with PBT in order to obtain an equimolar blend of PBT/PC/PAr in which PC and PAr are involved in a copolymer. Without any thermal treatment, the first heating run reveals a strong melting peak of PBT (Figure 23h) quite similar in its area and temperature position to that of neat PBT (Figure 23g). This indicates that one is dealing with a physical blend of neat PBT and PC/PAr random copolymer. After application of the same thermal treatment cycle as with the former blend of PC/PAr, but at a higher annealing temperature of 280"C, one observes an interesting result: at shorter annealing times (up to 500min) an almost complete amorphisation of the PBT-PC-PAr terpolymer occurs, as can be concluded from the disappearance of the PBT melting peak (Figure 23j). At this stage, a single '2' at about 100°C is also observed. The slight change in the curve at about 160°C could hardly be

367

Sequential Reordering in Condensation Copolymers

assigned t o a glass transition. Rather, this is the start of a very broad and flat melting peak (from 160 to 220"C), arising from PBT crystallites. The curve also indicates a more or less complete amorphisation of the PBTPC-PAr copolymer. Prolonged annealing between 800 and 1500min at the same temperature results in the recovery of the crystallisation ability (Figure 23k-m). It is important to stress here that this effect takes place after a relatively long thermal treatment of the sample in the molten state, Le., above T, of PBT. In other words, the restored crystallisation ability does not originate from crystallisation-induced sequential reordering, as proved for immiscible [46] and miscible [47] blends of the same partners. The sequential reordering in this case occurs in the melt (at 280°C). One has to assume that the respective driving force is related to the tendency of P B T and PAr to mix, since it is known that they are miscible in any composition [12]. The experiments performed with the binary PC/PAr and ternary PBT/PC/PAr blends have been repeated with a similar system comprising PBT, PA 66 and PAr. The only difference is that the experiments were

After annealing

After annealing time (min) at 280°C

PA66

b

do

1bO

1;O

ZbO 2b0 360 Temperature in O C

Figure 24. DSC curves of the PBT/PA 66/PAr blend, taken in heating mode at 10°C/min, after various annealing times at 280, 290 and 200°C. Sample weight: 62 mg. Curves of the starting neat PBT, PA 66, and PAr are given for comparison I821

368

S. Fakirov, Z. Denchev

started with a crystallisable PBT/PA 66 blend. The respective DSC curves of the blends (c)-(h) and (j)-(n) are shown together with those of the neat polymers (a), (b), (i) in Figure 24. As in the previous system, the physical mixture of the two starting components PBT and PA 66 behaves like the homopolymers. The only difference is that in the curves of the neat comPBT ponents one observes, in addition to the glass transitions Tg and TY , PA 66 and the melting peaks T T and T, , the crystallisation peaks T,': and T," ". Thermal treatment for 120min at T, = 280"C, ie., above the melting point of both polymers, does not change the picture qualitatively and ' values is observed (Figure 24d). Signifionly a slight decrease in the T cant changes occur at prolonged annealing. For t, = 370min there is no sign of crystallisation or melting (Figure 24h), which is assumed to be a proof of a complete amorphisation of the PBT-PA 66 copolymer. Using a common solvent, the PAr consitiuent is added at this stage, distinguished by a well-resolved Tg (Figure 24i)). Complete amorphisation of this blend consisting of the randomised PBT-PA 66 copolymer and neat PAr is achieved after annealing for 200 min at 290°C (Figure 24j). This result could be an indication that under these conditions PAr is not only incorporated into the pre-existing copolymer, but the terpolymer obtained is completely randomised. The most interesting results are obtained when annealing at this high temperature is extended up to 600min and particularly to 11OOmin: restoration of crystallisation is observed (Figure 24kJ). The melting peaks are rather broad (centered between 160 and 270°C), because the crystallisation of PBT and PA 66 takes place under very unfavourable conditions - only during the cooling of the melt from 290°C to room temperature. For this reason, an attempt is made to split the single broad melting peak into two peaks by creating better crystallisation conditions. The sample annealed at 290°C for ll00min (Figure 241) was cooled to 200"C, kept at this temperature for 15min and only then cooled further to room temperature. The DSC curves (m) and (n), taken after two such annealing procedures P A 66 at 2OO0C, reveal two very well-resolved melting peaks, T r and T,, . These results lead to the conclusion that after randomisation of the terpolymer comprising PBT, PA 66, and PAr by short annealing at 290"C, further annealing at the same temperature results in regeneration of the blocky structure of the copolymer. Similarly to the previous system (PBT/PC/PAr, Figure 23), the transition of the random copolymer into a blocky one cannot be driven by crystallisation factors, since it takes place at temperatures far above the melting of the crystallisable components. Having in mind that the two starting binary blends (PC/PAr and PBT/PA 66) consist of immiscible constituents and only the third component is miscible with one of the first two, it may be assumed that it is miscibility that accounts for the observed regeneration of the blocky structure. In other words, here one can speak about miscibility-induced sequential reordering

Sequential Reordering in Condensation Copolymers

369

in random condensation copolymers. 4.3. Models and thermodynamic considerations

As already mentioned, the entropy increase is the major driving factor for the transition of a homopolymer mixture first into a block copolymer and finally into a more or less random copolymer via transreactions. The “entropic” direction of these transreactions is well studied and generally accepted [5,7-11,17,20,33,43,44,85,86]. Factors that may act toward restoration of the blocks (ie., toward an entropy decrease) are not so obvious, but we believe they still exist. Numerous phenomena require the presence of longer sequences of repeating units in the chains. To them belong crystallisation [6345], miscibility [12,25], dipole-dipole forces, acidbase attraction [1,87],ion-ion interaction [88],and hydrogen bonding [89]. The best illustration in this respect is the crystallisation process. It can take place only when the chains contain crystallisable blocks, long enough to produce a minimum lamellar thickness of 50-80A. In the majority of cases, this means sequences of 5-8 repeat units. In randomised condensation copolymers of appropriate composition, miscibility could act in the same way as crystallisation, and even longer sequences are expected than in the case of crystallisation, where the lamellar thickness plays a limiting role. For miscibility, such a limitation does not exist, and the longer the blocks, the more effective the mixing. The demand for longer sequences for mixing to occur can be the driving factor for the restoration of blocks in the copolymer system (Eq. (5) or (6)). The results of the present study show that the tendency of the blend components to mix should be stronger than the general tendency toward entropy increase. Comparing the temperature conditions of both crystallisation-induced and miscibility-induced sequential reordering, it is worth mentioning that the restoration of blocks in the former process requires a temperature always lower than that at which the sample randomisation is achieved. For this reason, the reversible transition from block to random copolymer should not be regarded as an equilibrium process. In miscibility-induced sequential reordering, however, the restoration of blocks occurs at the same temperature as does the formation of a random copolymer. Hence, in this particular case, one could consider the above transition (Eq. (5) or (6)) as an equilibrium process. Further comparison of crystallisation-induced and miscibility-induced reordering shows that the transreactions responsible for the realisation of these processes are the same in each direction (both toward random and toward block copolymer) and so are the starting and final products in respect to their type and chemical composition. They differ only in the type of sequential order of the repeat units. Therefore, one should conclude that transreactions should be regarded just as a tool for the realisation of a given sequential order, but that they do not determine it. Other factors, namely

370

S. Fakirov, Z. Denchev

entropy or miscibility, that affect a given sequential order, predetermine the final product in the system. It is important to emphasise again that this consideration is valid only if the “starting” and the “final” products are identical in every respect except for the type of sequential order. All the concepts related to miscibility-induced sequential reordering are visualised in the schematic models displayed in Figures 25 and 26. Figure 25 illustrates the case when the random copolymer obtained from two immiscible condensation homopolymers, (A),,and (B),,,,is blended with the condensation homopolymer (C),. The components (C), and (B),,, are m i s cible. When this blend is annealed at high temperatures in the melt, the (C), component reacts with the copolymer and the initial block terpolymer transforms into a random terpolymer (Figure 25b). When the random terpolymer (ABC), (Figure 25b) is kept for a longer time at the same temperature, due to the tendency of (B),,,and (C), to mix, transreactions in the “opposite” direction take place, resulting in the formation of a blocky structure (see Eq. (5) or ( 6 ) ) . Dephasing should also take place with the appearance of two phases, the first comprising the (A),,,component and the second one the (B),,, and (C), blocks, as shown in Figure 25c. As a result, the random terpolymer (ABC), is converted into block terpolymer (A), - (B), - (C), . . Let us now suppose that one of the components, e.g., (C),, is crystallisable and the other two, (A),, and (B),,, non-crystallisable. On cooling of the molten terpolymer below TZ (Figure 26a), crystallisation should take place (as shown schematically in Figure 26b) if a blocky structure is really available. As a result, one may expect evidence oE (i) the crystalline

--

a

--

b

C

Figure 25. Schematic representation of the miscibility-inducedchanges in the sequential order of a terpolymer, taking place via transreactions during shorter and longer annealing. The condensation homopolymers are: (A),, (a), (B),,,(o), and (C), ( 0 ) ; (A),, and (B),,, are immiscible and non-crystallisable, (C),is crystallisable and miscible with (B),,,[82]

Sequential Reordering in Condensation Copolymers

371

phase of the (C), component (with its Tz);(ii) the amorphous phase of the same component (with its Tgc) that is not completely mixed with the amorphous phase of (B),, which should display its TgB,too and (iii) the (A)n component (with its T:). If this is the actual case, during subsequent heating of the crystallised blocky terpolymer the four transition temperatures Tgc,T:, Tg*, and T: should be observed in the DSC curve, together with some degree of crystallinity of the (C), component - w,".All our experimental results seem to be in favour of the transitions described above and illustrated by the models in Figures 25 and 26. Analysing the system PBT/PC/PAr, we considered the merging of the PC PC-PAr (Figure 23f) as a reliable proof two Tg and TgpA'into a single Tg of randomisation of this binary blend. The next important step is the randomisation of the terpolymer, as illustrated in Figure 25b. This effect takes place after longer annealing at higher temperature (Figure 23j). The third stage is the regeneration of blocks by an even longer thermal treatment at the same high temperature, as shown schematically in Figure 25c. Evidence of the occurrence of this transition can be found in the appearance of two glass transitions and one melting temperature, in accordance with our expectations. One does really observe in Figure 23k the transitions PBT PC TgpBTand T, , as well as Tg . In Figure 26, the corresponding generc c A alised designations are Tg , T,, and Tg . The only missing temperature is Tr', corresponding to T,", most probably because of overlapping with the melting peak of PBT. Bearing in mind the miscibility of PBT and PAr, PBT-PAr one should expect only one Tg instead of two T,PBTand Tg"'.This is not the case because most probably the crystallisation of PBT leads to separation of the amorphous PBT and PAr phases.

;below

a

b

Figure 26. Schematic representation of the crystallisation and dephasing of a block tercopolymer taking place during cooling of the melt. The symbols are the same as in Figure 25. T,and Tmare the glass transition and melting temperatures of the respective components (A),,, (B)m, and (C), [82]

372

S. Fakirov, Z. Denchev

Tz

The second system under investigation PBT/PA 66/PAr, behaves in a rather similar way. After the block regeneration, one should expect here , 66, 66 the following transition temperatures: TiBT, T T , and TYr. At this stage one detects in the DSC curves very slight indications of TJBT (Figure 24k), and of overlapped and T z " (Figure 24k,l). The abPAC sence of Tg can be ascribed again to superposition of its interval on the melting peak; an additional reason could be the lower concentrations of this component. The observed regeneration of the crystallisation ability after prolonged anneahg in the melt cannot be a result of chemical reactions taking place during the cooling to room temperature, i e . , due to a crystallisationinduced sequential reordering. The reason is that if the blocks have been restored only during the cooling process, this should have happened also during cooling after shorter annealing of the melt, i e . , at the stage of randomisation (Figure 23f and Figure 24h). Since this is not the case, one has to accept that the non-isothermal crystallisation during cooling of the melt is possible only when it is preceded by block restoration in the melt. Finally, the apparent lack of consistency in the interpretation of the effects observed deserves some explanation from the thermodynamic point of view. At first glance, thermodynamics should favour either a phaseseparated system (in the melt), consisting of long homopolymer blocks, or a homogeneous melt of a random terpolymer. In the former case, the driving force for intermediate formation of a completely randomised terpolymer would be missing. In the latter case, homopolymer blocks long enough to crystallise should not be restored at all. However, this contradiction is only ostensible, as demonstrated by a more detailed analysis of the matter. The entropy increase of the process of randomisation of a block copolymer A S depends on the entropy of the initial state S1 and final state S2. Obviously, AS' is maximal when S2 corresponds to a completely randomised sample with a statistical sequence distribution, and & to the starting blend of homopolymers or rather to a diblock formed in the first interchain reaction. As shown by theoretical [6,27] and experimental measurements [9&92], a random copolymer chain with a statistical distribution of the repeating units is to be represented as a combination of homodyads. In other words, the transformation of a diblock copolymer into a random one built up, on the average, of homodyads is characterised by a maximum entropy increase AS', the latter being the driving force for this transformation. The entropy change AS" should be much smaller, when the transformation is from copolymer molecules comprising many blocks of 56 homosequences each to the random distribution described above (or vice versa), i e . , AS' << AS'. For this reason, the smaller value of AS" (ie., the trend toward randomisation) could be suppressed by other factors in the system, e.g., miscibility. The trend of the miscible components to mix, which acts in a direction

Tg"

TF

Sequential Reordering in Condensation Copolymers

373

opposite to that of the entropy, also exists at the very beginning of the randomisation process, but this trend is obviously too weak to suppress the entropy effect. It may become effective and govern processes for which the entropy increase is small, for instance, the last stages of randomisation. Only in this way can one explain why complete randomisation via transreactions requires such extremely long times or - the most frequent case - why the randomisation ceases at triads and tetrads, as recently reported [92]. This effect, disclosed earlier for a poly(ethy1ene adipate)poly(hexamethy1ene terephthalate) copolymer [15], has been recently observed with the PET/PA 6 system [93]. Strictly speaking, reliable experimental data, indicating the presence of homodiad sequences, are obtained only with random copolymers synthesised from monomers (e.g., two diacids and two diols), rather than as a result of randomisation of a homopolymer blend or a block copolymer [15]. NMR analysis of the PET/PA 6 copolymers shows that both the random copolymer and that with restored blocks have the same sequence length of about four units. Nevertheless, the random sample is not able to crystallise while the “blocky” one is crystallisable, as revealed by DSC and WAXS [47].This drastic difference in the behaviour of copolymers with virtually the same sequence length can be explained by the different types of sequence length distribution. Unfortunately, NMR analysis can determine just average values of the sequence lengths, which is insufficient for an accurate characterisation of the sequence length distribution. Quite similar is the situation with the crystallisation-induced sequential reordering, where sequences of at least 5-6 repeat units are required for lamellae formation. Bearing in mind the very small value of AS’’, characterising the entropy difference between the random copolymer and the copolymer with restored blocks, it seems acceptable that AS’’ might be compensated by enthalpy changes to realise the “anti-entropic” direction of the process, i. e., crystallisation. In summary, the antagonism between the entropy and other factors during sequential reordering in condensation copolymers can cause effective reversible transformations only if the homopolymer sequences are short enough, ie., if the block copolymer can be considered as a microblock copolymer. 5. Study of the sequential order in condensation copolymers by

means of SEC after selective degradation

Gel permeation chromatography (GPC) or, to use the more modern term, size exclusion chromatography (SEC), is unquestionably the most popular analytical procedure for rapid determination of MWDs in synthetic polymers. This became possible after three fundamental contributions made between 1964 and 1971. The investigations of Moore in the field of porous

374

S.Fakirov, Z. Denchev

polystyrene gels helped the production of stable, hydrophobic column packings [94]. The second major development was the work by Casassa [95-981, who formulated a theoretical model of SEC using statistical mechanical calculations. His very important contribution was the prediction that SEC is an equilibrium, entropy-controlled process of size exclusion that is independent of temperature and governed by the ratio of the radius of gyration of a polymer to the average pore size of the packing [99]. The third theoretical contribution, on which SEC is based, is the concept of universal calibration that was first introduced by Benoit et al. [loo]. In the majority of cases, this concept allows the determination of correct molecular weights even when the polymer standards used for column calibration and the samples to be studied are of different structure, or chemical compositions.

5.1. Basics of the SEC technique SEC separates molecules according to their effective molecular size in solution. The sample solution is introduced onto the column, which is filled with a rigid or gelled porous packing. The sample is carried by the solvent (the mobile phase) through the column. Size separation takes place by repeated exchange of the solute molecules between the bulk solvent of the mobile phase and the stagnant liquid phase within the pores of the packing. The SEC curve is a molecular size distribution curve, but when a concentration-sensitive detector is used, the curve expresses size distribution in terms of weight concentration. By calculation, the row data are converted into a molecular weight distribution curve and the respective molecular weight averages can be derived [loll. SEC has a unique separation mechanism, compared to all other chromatographic methods. In contrast to gas or liquid chromatography, the mobile and stationary phases in SEC are of the same material, namely the solvent being pumped through the column. Therefore, separation does not depend on the different solubilities of the solute in the mobile and stationary phases. The latter is enclosed by pores and a molecule must fit a single pore. If it is too large, it remains in the mobile phase and elutes quickly from the column, ie., it is totally excluded. The passage of a smaller molecule is delayed to the extent to which it can fit into the pores. The stay in the pores is temporary and the molecule eventually elutes from the column [loll. Such a description of the process of size exclusion follows from the finding [95] that the SEC retention mechanism can be interpreted by considering just the equilibrium between free molecules and molecules confined in the pores, a. e., the process should be entropy-controlled and independent of temperature, provided that enthalpic interactions (adsorption or chemical reactions) between polymer and column packing are absent and that the so-called macromolecular crowding effects are not present either. In situ shear degradation of the polymer to be studied must also be avoided. In practice, the undesirable surface interactions are eliminated by selecting a

Sequential Reordering in Condensation Copolymers

375

mobile phase that is more strongly adsorbed than the sample by the column packing. Trace amounts of highly polar materials in the mobile phase can also minimise adsorption. Concentration overloading is kept to a minimum by applying sufficiently low sample concentrations. Sample retention is measured as t R , i. e., the time required for a peak to elute from the column following sample injection. The tR value is sensitive to the experimental conditions (flow rate, specific columns). The retention volume VR,however, accounts for flow rate differences. To calculate VR,the mobile phase volume flow rate F must be known as well as the t~ values, since VR = FtR. Although peak retention reported as VRis not sensitive to flow rate changes, it can still vary with the differences in column size and the instrument dead volume. Bearing in mind all these considerations, the general retention equation for SEC is given by [loll:

+

+

VR= Vo KSECV, B

(7)

where VOis the interstitial liquid volume between the packing particles, KSECrepresents the ratio of the average solute concentration in the pores to that outside the pores. If V, is the internal pore volume, the product KSECV, represents the portion of the retention volume VR,owing to the interaction of the polymer with the pores of the column packing. Obviously, for molecules too large to enter the pores this term is equal to zero. The B term accounts for the adsorption effects and other enthalpic processes between the sample and column packing, which should be minimised so that B x 0. It is beyond the aims and scope of this chapter to consider the SEC equipment, such as columns and column packings, detectors, and pumps, and the details of data handling. Useful information in this respect can be found in [102,103]. Calibration of the system, however, plays an important role in data evaluation because it is critical for the obtaining of correct results and therefore requires special attention. The SEC chromatogram is a graph of the weight fraction of a polymer sample against elution volume or weight of the mobile phase (see Figure 7). Relative sample comparison is valid only for data obtained under the same experimental conditions. Transformation of the elution curves into MWD curves (Figure 8) enables the comparison of SEC data obtained with different instruments or the quantitative treatment of different samples obtained under the same conditions. This is because the MWD of a sample is an intrinsic polymer property, in contrast to the elution curves. In fact, the MWD contains all the important information, which has to be revealed accurately by proper system calibration. The first step in each SEC study is to establish a calibration curve relating the elution volume to the molecular weight of various samples of narrow polydispersity (usually PS). Absolute accuracy is achieved only if the sample to be studied by SEC is polystyrene too, for there is no reason

376

S. Fakirov, Z. Denchev

why calibration established by PS standards should be valid for other polymers. On the other hand, very often it is practically impossible to make a new calibration curve whenever a new polymer is studied. A possible way to overcome this problem is to assume that, regardless of the nature of the polymer studied, PS calibration is valid. This may lead to erroneous molecular weight data since the relationship between the real sample molecular weight and that obtained with PS standards is unknown. Since the PS calibration is applied to samples with a different chemical structure more often than one may expect, in each such case an indication should be given that one is dealing with the so-called “equivalent PS molecular weight”. If the real molecular weight of the sample is needed and there is no way of using narrow MWD standards of the sample material to obtain the calibration curve, one may use the universal calibration of Benoit et al. [loo]. The method transforms the PS peak-position calibration curve into a suitable curve characterising all polymer types. In general, SEC calibration curves for polymers of different types merge into a single plot when the calibration data are given as log[q]M, as illustrated in Figure 27, instead of the usual log M scale, as in the PS peak-position calibration curve. In the universal calibration procedure, one first constructs the PS calibration curve, using a set of narrow MWD PS standards. For each polystyrene MW value on the curve a corresponding value of [q] is determined, e.g., by means of a capillary viscometer. Alternatively, the value of the intrinsic viscosity corresponding to a particular molecular weight can be calculated by using the Mark-Houwink equation, if the K and a val-

Elution volume, THF solvent

Figure 27. Universal calibration plot of Benoit et al., demonstrating that the molecular hydrodynamic volume [7]M governs SEC separation [loo]

377

Sequential Reordering in Condensation Copolymers

ues for polystyrene at a given temperature and solvent are available. Since these parameters are temperature-dependent , they must correspond to the temperature at which the column is operating. To construct the universal calibration plot, the corresponding [q] and M values are multiplied and the logarithm of this product is plotted versus the elution volume for that particular M value. Provided that the Mark-Houwink coefficients a, and K, of the sample are known for the given mobile phase at the appropriate temperature, the molecular weight of the sample M , at each elution volume could be calculated [loll: logM, = [1/(1+ a,)] log(K/K,)

+ [(I + a ) / ( l +a s )lo] g M

(8)

where Q and K are the Mark-Houwink coefficients of the PS standards. Although the universal calibration concept gives the researchers a highly useful approach, some of its limitations should be pointed out. First, one needs to know the accurate Mark-Houwink coefficients of the sample; second, if the sample is chemically heterogeneous (e.g., a copolymer), there would not be a single set of Mark-Houwink coefficients, but variable values. At present, efforts are being made to solve these problems by introducing new detectors, whereby the intrinsic viscosity at each elution volume can be determined directly [103]. 5.2. Selective degradation of PC-containing

condensation copolymers

As described for the first time in [5], the P C blocks are selectively attacked and removed from the system as low molecular weight compounds by reaction of the PET/PC copolymer samples with a CHzClz solution of piperidine at room temperature, according to the scheme given below.

+ H-”J

where R’ and R” can be

/ \

R’OH

+ R”-0-C-N II

0 or

+ R’-0-C-NII

ara R”OH

0

3

and/or -CHz-CHz-

S. Fakirov, Z. Denchev

378 PC DEGRADATION PRODUCTS

I 0

CH3

I1

I11

I :I1 : I11 = 2 : 1 : 1

These low molecular weight compounds are found to be 2-(4hydroxypheny1)-2-(4-piperidinocarbon ylox ypheny l)propane) (I), 2,2-bis (4piperidinocarbonyloxypheny1)propane (11) , and bisphenol A (111) [5]. Once degradation is completed, recovery of even the shortest PET blocks is performed by precipitation of the CH2C12 solution with methanol [38].Thereby, compounds I, I1 and I11 remain in the solution and the PET residue can be filtered off, washed, dried and studied by SEC. Since the method of selective degradation leaves the PET blocks unchanged and does not create additional bonds between them [5,38], it provides information about the average molecular weight of the PET segments in the PET/PC copolymer obtained after melt-mixing. Unfortunately, the above method of selective degradation cannot be used with all polycarbonate types. The attempts to selectively degrade poly(2,2-dimethyltrimethylenecarbonate)(DTC) to low molecular weight products, by applying the above procedure, failed even when prolonged times and elevated temperatures were employed in the process of aminolysis [104]. Obviously, the carbonate group in the DTC polymer is much more stable than that in bisphenol A PC due to the different chemical structure of the substituents.

5.3. Sequence length determination in poly(ethy1ene terephthalate) - bisphenol A polycarbonate ovrndorn copolymers a5 revealed by combined NMR and SEC studies 5.3.1. NMR studies The NMR technique remains the most powerful tool for sequential analysis of condensation copolymers. PET-PC copolymers were prepared by

379

Sequential Reordering in Condensation Copolymers

reactive melt-mixing for 180min at 280°C as described in [39]. For the characterisation of these supposedly random PET/PC copolymers we used the method of Devaux et al. [18] based on the earlier works of Kricheldorf [15], and Yamadera et al. [27]. It was interesting to apply a combination of selective degradation of PC and subsequent analysis of the remaining PET-containing residue by means of NMR in order to compare the results to those derived from the NMR spectra of non-degraded samples. Amorphous (as revealed by DSC) PET-PC copolymers obtained by melt-mixing were studied [39]. These copolymers were completely soluble in chloroform and were isolated by solvent evaporation and/or precipitation with methanol. Use was made of the most characteristic range in the 'H-NMR spectra of the PET-PC copolymers: the signals situated between 8.0 and 8.4 ppm. They belong to the benzene protons of the terephthalic acid residue and are sensitive to the sequence length changes (see also Chapter 3). In Figure 28, the 'H NMR spectra of an amorphous PET-PC copolymer are depicted before (a) and after (b) the selective degradation of the PC units. B1 and Bz represent the acidic residues in the system (terephthalic and carbonate, respectively). The ethylene glycol and bisphenol A moieties are denoted A1 and A2, respectively. The structures of the corresponding fragments are as follows:

- c e f -

II

0

Bi

0

-OCH2CH20-

A1

A2

-c-

II

0

B2

From Figure 28, spectrum (b), it becomes evident that the residual product after P C elimination does not comprise PET sequences only; as one might expect, it would have revealed then only a singlet at a chemical shift 6 = 8.06ppm1 which is not the case. Even though aminolysis affects just the OC(0)O- groups of PC, the ester groups of any type remaining unaffected, the resultant structure after elimination of the PC repeat units is supposed to be rather complex. Both spectra in Figure 28 represent a superposition of the four signals of an asymmetrically substituted (with both alkyl and aryl moieties) terephthalic acid residue Alk-Ar (or A2B1A1) and the two signals of two symmetric terephthalic acid derivatives Alk-Alk (or A1B1A1) at 8.06 ppm and Ar-Ar (or A2B1Az) at 8.28ppm. The former represents a PET unit and the latter the unit of an aromatic terephthalic polyester. The relation-

380

S. Fakirov, Z. Denchev

,

,

ppm 8.4

'

1

8.2

8

8.0

Figure 28. 'H NMR spectra of a non-degraded (a) and degraded (b) PET-PC random copolymer 1391

ship between the different types of terephthalic protons derived from the corresponding peak areas is also worth mentioning. After selective elimination of the PC sequences in all amorphous (as revealed by DSC) PET-PC copolymer samples, the relation of Ar-Ar:Ar-A1k:Alk-Alk = 1:6:9. The approach of Devaux et al. [18] was applied to calculate the coefficients x and y which represent the average degree of polymerisation of PET and bisphenol A terephthalate sequences (see also Eq. (10)). The following relations were used for this purpose:

As seen in Table 9, in all samples where no selective degradation was performed rather unusual x-values were obtained. They were lower than those of a completely random copolymer which would normally indicate a trend toward alternating copolymer. The x-values obtained after selective degradation, however, are higher than those theoretically calculated for fully random copolymers (see the bold values in Table 9). We consider them more reliable than the results obtained without degradation, for the following reasons. The degradation procedure eliminates from the system to be studied by NMR not only the PC residues but also all low molecular weight by-products formed during the melt-mixing and prolonged thermal annealing, so the risk of misinterpreting the results due to peak overlapping

381

Sequential Reordering in Condensation Copolymers

Table 9. Average degree of polymerisation of ethylene (x)and bisphenol A (y) terephthalate segments derived from the 'H NMR spectra in presumed random PET-PC copolymers obtained by melt-mixing for 180min at 280°C [39] No. 1 2 3 4

5 6 7 8

Method of random Annealing at copolymer separation 235OC/24 h Coagulation Coagulation Coagulation Coagulation Solvent evaporation Solvent evaporation Solvent evaporation Solvent evaporation

Selective degradation of PC sequences

No No

Yes Yes No No

Yes Yes

No

Yes No

Yes No

Yes No

Yes

Theoretical values for random copolymers [ 181

x

y

2.67

1.72

1.34 1.23

2.51

1.17

1.62

1.23 1.44 1.15

1.65

1.72 2.66

2.50 2.16

1.39 1.32

1.87

is lower. Furthermore, bearing in mind the entrophic nature of the driving force toward randomisation [30,46,47], it seems very unlikely that in a real system and for finite annealing times the final state of randomisation alternating PET and PC homodyads, not to mention x-values below 2.0 would be reached. The same PET-PC samples were studied by 13C NMR.The results are given in Table 10, where x, y, z,and w correspond to the indices in the following general formula [MI:

and are calculated using the following relations:

Again, B1 and B2 represent the acidic residues in the system (terepht halic and carbonate, respectively) the ethylene glycol and bisphenol A moieties are denoted by A1 and A2. Therefore, [AlBl] and [AzBz] are the concentrations of pure PET and PC units, respectively; likewise, [AlBz] and [A2B1] are the concentrations of the PET-PC transitional structures. One can see in the above tables that, as expected, the method of sample separation (coagulation or solvent evaporation) does not affect significantly the PET sequence lengths. The values for samples with the same thermal prehistory determined by the two NMR methods are also close enough. According to both NMR techniques applied, copolymers with fully random sequence distribution (consecutively linked homodyads of PET and P C) were not obtained. In this respect our results are in favour of the supposition of Kricheldorf [15] and of Backson et al. [92] who pointed out

382

S. Fakirov, Z. Denchev

Table 10. 13C NMR sequential analysis of PET-PC copolymers, differing in thermal prehistory and method of separation [39].Sample numbering as in Table 9 Sample no.

X

1.64 2.80 1.72 2.59 1.70

2.76 1.55 2.60

v

2.55 1.56 2.45 1.63 2.42 1.57 2.82 1.62

x

W

1.38 3.00 1.47 2.81 1.40

3.57 1.33 3.11 1.55 3.40 1.56 2.96 1.59

2.77 1.51

2.70

Theoretical value for a random copolymer [18] 2.16 1.87 2.16 1.87 13C NMR species studied for [ A ~ B I and ] [ A ~ B I ] the : C=O carbon signals of the terephthalic acid residue at 165.6ppm and 164.3ppm, respectively; for [ A ~ B z ]and [AzBi]: the peak of the H-bonded benzene carbons of the bisphenol A residue at 121.1 and 120.83 ppm, respectively. Bold numbers depict the values for selectively degraded samples.

that completely random copolymers could only be prepared in the process of primary polycondensation (starting from monomers, e.g., of two diacyl dichlorides and two diols or diphenols) rather than by transreactions in binary polyester or polyamide blends. Comparison of the NMR data of PET-PC copolymers before and after aminolysis enabled us to elucidate the sequential ordering in the residual PET-containing fragment that remains after PC selective elimination according to the scheme in Section 5.2. We believe that this fragment accurately reflects the PET units sequence in the starting PET-PC copolymer before its selective degradation. Based on the l H NMR data, the following model structure could be suggested for the average residual fragment remaining after the PC selective degradation:

383

Sequential Reordering in Condensation Copolymers

The designations I, I1 and I11 are for the Alk-Ar, Alk-Aik, and Ar-Ar terephthalic protons, respectively. According to this theoretical structure, the relation Ar-Ar:Alk-Ar Alk-Alk = 1:4:6 (if the average PET sequence is considered to comprise 3 PET units). Comparison with that practically established, which is 1:6:9 (or 0.7:4.0:6.0),suggests that the theoretical value corresponds relatively well to the NMR data. At the same time, the real degraded copolymer would probably contain not only the above theoretical sequence. One could also expect various sequences representing different fragments of the theoretical one, containing more asymmetric AlkAr moieties and a little longer PET blocks.

5.3.2. SEC studies As seen from its theoretical structure, the average residual fragment obtained after selective degradation of the PC units should have a molecular weight between 2500 and 2900 depending on the length (two or three units) of the PET sequences. Therefore, a further evidence in favour of the structure proposed might be found by SEC investigations of degraded PET-PC copolymers. Figure 29 and Table 11 summarise all the SEC data of the molecular weights (MW) and molecular weight distributions of the PET-PC copolymers before and after degradation and removal of the PC units. The selective degradation causes a drastic decrease in the molecular weight of the starting samples. In the case of neat PC (Table l l ) , the average weights of the products of aminolysis are close to that of bisphenol A, which is one of

I

I

I

I

I

I

'

384

S. Fakirov, Z. Denchev

Table 11. Molecular weights M,,and Mw and polydispersity coefficient Mw/M,, for PET-PC copolymers with different thermal prehistory before and after PC selective degradation [39] Sample

Thermal treatment Selective Molecular weights from duration of the degradation SEC” and PET/PC blend their distribution

M, Mw

at 280°C (min) PET-PC equimolar PET-PC equimolar PET-PC equimolar PET-PC equimolar Bisphenol A PC homopolymer PC homopolymer

45b 45 265‘ 265

Mw/M?l

No

19690 51332 2.607 5027 8611 1.713 No 10266 28161 2.755 1590 2717 Yes 1.709 303 312 No 1.029 No 18567 29800 1.605 348 361 Yes 1.037 ” Equivalent polystyrene MW, SEC conditions as indicated in Figure 29 Crystallisable (according to DSC) PET-PC copolymer (longer blocks) ‘Non-crystallisable (according to DSC) PET-PC copolymer (shorter blocks)

Yes

the low molecular weight degradation products. This is an additional indication that complete selective degradation of the PC units is carried out. The values of a selectively degraded PET-PC random copolymer (prepared by melt-mixing of an equimolar PET/PC blend for 180min at 280°C) deserve special attention. They strongly support the proposed hypothetical structure of the residual PET-containing fragment. All the degraded PET-PC random copolymers studied by NMR (Tables 9 and 10) reveal and 2600-3100 for A&,, i e . , in similar values of about 1700-1800 for all cases the theoretical MW of 2500-2900 is between M,, and Mw determined by SEC. Since these are equivalent polystyrene M W s , one should not look for a perfect coincidence between the theoretically calculated and experimentally found values. However, the similarity is sufficient to s u p port the structure of the residual PET-containing fragment, as suggested on the basis of NMR data. This fragment contains two sequences of 2-3 PET repeat units separated by long spacers built up of terephthalic and bisphenol A moieties and is therefore non-crystallisable. Judging from Table l l , melt-mixing for 45 min at 280°C apparently gives rise to a similar fragment , comprising 3-fold longer PET sequences. Based on the above results, the combination of selective degradation of PC sequences and subsequent PET block length determination by NMR in a PET-PC copolymer seems to give results that are more logical and more reliable than those obtained by the same technique but without elimination of the PC units. This is especially valid for PET-PC copolymers with a sequence distribution approaching the random one obtained after prolonged annealing of the starting homopolymer blend.

Sequential Reordering in Condensation Copolymers

385

6. Conclusions

The repeatedly reported randomisation of molten block copolycondensates is observed and proved by various techniques sensitive to the changes in the crystalline and/or the amorphous phases. The process is accompanied by a loss of crystallisation ability. This melting-induced sequential reordering is driven mostly by the entropy increase. Restoration of crystallisation ability is observed by annealing of random copolycondensates, and attributed to regeneration of crystallisable blocks. This cystallisation-induced sequential reordering is driven by upsetting the random t)block copolymer equilibrium during the annealing of the random system. Restoration of crystallisation ability is also observed in random terpolymers obtained by melting-induced sequential reordering of ternary homopolymer blends, in which two of the constituents are miscible over the entire concentration range. The effect is established by subjecting these terpolymers to prolonged annealing at the same temperature as copolymer preparation. In this miscibility-inducedsequential reordering the process of block restoration is driven by the miscibility factor. Both randomisation and block regeneration strongly depend on ternperature, transesterification catalyst, and miscibility of blend components. These processes are composition-sensitive and seem to be related to the specific chemistry of the interchange reactions and starting homopolymer structures. The results obtained show that in the transition from block to random copolymer, (i.e., from longer to shorter blocks), the sample crystallisability decreases, attaining eventually a complete (as revealed by DSC) amorphisation. Likewise, regeneration of melting endotherms in the crystallisationinduced reordering process should reflect the restoration of longer PET blocks. Of course, the absence of melting (crystallisation) should not be considered as an indication of complete randomisation. During the crystallisation-induced reordering, one should not exclude entirely the possibility of segregation of units belonging to different copolymer chains, caused by the intensive segmental motion. The non-periodic layer crystallites thus formed might be organised in domains which, along with the newly formed blocks, could contribute to the restoration of the melting endotherms.

Acknowledgement The authors gratefully acknowledge the financial support of the Deutsche Forschungsgemeinschaft (DFG FR 675/21-1).

386

S. Fakirov, Z. Denchev

References 1.

A. Utracki, “Polymer Alloys and Blends”, Hanser Verlag, Miinchen 1989,pp.

1-15 2. J. Flory, “Principles of Polymer Chemistry”, Cornell University Press, Ithaca, New York 1953,pp. 127-155 3. S.Fakirov, “Solid State Reactions in Linear Polywndensated’, in: Solid State Behavior of Linear Polyesters and Polyamideq edited by J.M. Schultz, S. Fakirov, Prentice Hall, Englewood Cliffs, NJ 1990,pp. 1-75 4. J. Flory, J. Am. Chem. SOC.64,2205 (1945) 5. F. Pilati, E.Marianucci, C. Berti, J. Appl. Polym. Sci. 30,1267 (1985) 6. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 7. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24, 3315 (1986) 8. G. Montaudo, M. S. Montaudo, E. Scamporrino, D. Vitalini, Macromolecules 25,5099 (1992) 9. C. Berti, V.Bonora, F. Pilati, M. Fiorini, Makmmol. Chem. 193,1665(1992); 10. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22,229 (1982) 11. A. M. Kotliar, J. Polym. Sci., Macromol. Rev. 16,367 (1981) 12. R. S.Porter, L. H. Wang, Polymer 33,2019 (1992) 13. J. Kugler, L. Gilmer, D. Wiswe, H. Zachmann, K. Hahn,E. W. Fischer, Macromolecules 20, 1116 (1987) 14. K. McAlea, J. Schultz, K. Gardner, S. Wignall, Polymer 27, 1581 (1986) 15. H. Kricheldorf, Makrumol. Chem. 179,2138 (1978) 16. M. Valero, J. J. Iruin, E. Espinosa, M. J. Fernandez-Berridi, Polym. Commun. 31, 127,(1990) 17. L. H. W a g , Z. Huang, T. Hong, R.S. Porter, J. Macromol. Sct.-Phys. B29, 155 (1990) 18. J . Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1881(1982) 19. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24,3301 (1986) 20. J. I. Eguiazabal, G. Ucar, M. Cortazar, J. J. Iruin, Polymer 27,2013 (1986) 21. L. H. Wang, M. Lu, X. Yang, R.S.Porter, J. Macromol. Sci.-Phys. B29, 171 (1990) 22. G. v.d. Vel’den, G. Kolfschoten-Switsamas, A. Veermans, Polym. Commun. 28, 169 (1987) 23. H. Li, A. Wong, MMI Press Symp. Ser. 2,395, (1982) 24. P. Henrichs, J. Tribone, D. Massa, J. Hewitt, Macromolecules 21,1282(1988) 25. M. Kimura, R.S. Porter, Anal. Calorim. 5,25 (1984) 26. A. M. Kotliar, J. Polym. Sci., Polym. Chem. Ed. 11, 1157 (1973) 27. R.Yamadera, M. Murano, J. Polym. Sci., Polym. Chem. Ed. 5,2259 (1967) 28. P. Kresse, Faserforsch. Teztaltech. 11,353 (1960) 29. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1875 (1982) 30. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1895 (1982)

Sequential Reordering in Condensation Copolymers 31. 32. 33. 34. 35.

36. 37. 38.

387

A. M. Kotliar, J. Polym. Sci., Polym. Chem. Ed. 13, 973 (1975) S. Hanna, A. H. Windle, Polymer 29, 207 (1988) M. Kimura, G. Salee, R. S. Porter, J. Appl. Polym. Sci. 29, 1629 (1984) S. Fakirov, M. Evstatiev, J . Schultz, Polymer 34, 4669 (1993) T. Serhatkulu, I. Bahar, B. Erman, S. Fakirov, M. Evstatiev, D. Sapundjieva, Polymer 36, 2371 (1995) M. Evstatiev, N. Nikolov, S. Fakirov, Polymer 37, 4475 (1996) S. Fakirov, M. Evstatiev, Adv. Mater. 6,395 (1994) M. Fiorini, C. Berti, V. Ignatov, M. Tosseli, F. Pilati, J. Appl. Polym. Sci.

55, 1157 (1995) 39. Z. Denchev, A. DuChesne, M. Stamm, S. Fakirov, J. Appl. Polym. Sci. 68, 429 (1998) 40. J . Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1907 (1982) 41. A. Golovoy, M. F. Cheung, K. R. Carduner, M. J . Rokosz, Polym. Eng. Sci. 29, 1226 (1989) 42. R. L. Miller, R. V. Brooks, J . E. Briddell, Polym. Eng. Sci. 30, 59 (1990) 43. J. I. Eguiazabal, M., J. J . Iruin, J. Appl. Polym. Sci. 42, 489 (1991) 44. M. E. Stewart, A. J. Cox, D. M. Naylor, Polymer34, 4060 (1992) 45. J. S. Kollodge, R. S. Porter, Polymer 34, 4990 (1993) 46. S. Fakirov, M. Sarkissova, Z. Denchev, Macromol. Chem. Phys. 197, 2837 (1996) 47. Z. Denchev, S. Fakirov, M. Sarkissova, F. Yilmaz, Macromol. Chem. Phys. 197, 2869 (1996) 48. S . G. Hovenkamp, J. Polym. Sci., Polym. Chem. Ed. 9, 3617 (1971) 49. M. F. Cheung, A. Golovoy, R. 0. Carter, H. van Oene, Ind. Eng. Chem. Res. 28, 476 (1989) 50. K. R. Carduner, R. 0. Carter, M. Cheung, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40, 963 (1990) 51. N. F . Cheung, K. R. Carduner, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40, 977 (1990) 52. M. Shuster, M. Narkis, A. Siegmann, J. Appl. Polym. Sci. 52, 1383 (1994) 53. D. M. Miley, J . Runt, Polymer 33, 4643 (1992) 54. S. R. Murff, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 29, 3231 (1984) 55. R. Koningsveld, K. Kleintjens, J. Polym. Sci., Polym. Symp. Ed. 61,221 (1977) 56. M. Kimura, R. S. Porter, J. Polym. Sci., Polym. Phys. Ed. 21, 367 (1983) 57. E. Espinosa, M. J. Fernandez-Berridi, I. Maiza, M. Valero, Polymer 34, 382 (1993) 58. C. R. Desper, M. Kimura, R. S. Porter, J. Polym. Sci., Polym. Phys. Ed. 22, 1193 (1983) 59. I . Mondragon, J. Nazabal, J. Appl. Polym. Sci. 32,6191 (1986) 60. M. Evstatiev, J. M. Schultz, S . Petrovich, G. Georgiev, S. Fakirov, K. Friedrich, J. Appl. Polym. Sci. 67, 723 (1998) 61. S. Fakirov, M. Evstatiev, S. Petrovich, Macromolecules 26, 5219 (1993) 62. S. Fakirov, M. Evstatiev, “Interfacial Chemical Interactions in Condensation Polymeric Composites”, in: The Interfacial Interactions in Polymeric Composites, edited by G. Akovali, NATO AS1 Ser., vol. 230, 417 (1990)

388 63. 64. 65. 66. 67.

S. Fakirov, Z. Denchev

R. Lenz, S. Go, J. Polym. Sci., Polym. Chem. Ed. 11,2927 (1973) R. Lenz, S. Go, J. Polym. Sci., Polym. Chem. Ed. 12, 1 (1974) G. Chen, R. Lenz, Polymer 26, 1307 (1985) R. Lenz, J. Jin, K. Feichtinger, Polymer 24, 327 (1983) A. Muhlenbach, J. Economy, R. Johnson, T. Karis, J. Luerla, Macromolecules

23, 1803 (1990) 68. 3. Economy, R. Johnson, A. Muhlenbach, J. Lyerla, ACS Symp. Ser. 435 (Liq. Cryst. Polym.), 129 (1990) 69. C. W. Potter, J. C. Lim, G. Serpe, J. Economy, Preprints of the PPP-3 Conference (1993) 70. T. Pakula, private communication 71. M. Stamm, private communication 72. G. Georgiev, D. Simov, Chim. Ind. (Sofia) 52, 16 (1980) (in Bulgarian) 73. F. van Atwerpen, “Kinetics of Crystallization Phenomena of Spherulited’ , in: Poly(ethy1ene terephthalate), G. J. Thieme, Nijmegen, The Netherlands 1971, p. 45 74. J. O’Reilly, F. Karasz, H. Bair, J. Polym. Sci., Polym. Chem. Ed. 6 , 109 (1964) 75. K. Cheng, Macromolecules 20, 2802 (1987) 76. T. R. Nassar, D. R. Paul, J. W. Barlow, J. Appl. Polym. Sci. 23, 85 (1979) 77. E. W. Fischer, S. Fakirov, J. Mater. Sci. 11, 1041 (1976) 78. S. Fakirov, E. W. Fischer, R. HofFmann, G. Schmidt, Polymer18,1121(1977) 79. S. Fakirov, A. Apostolov, C. Fakirov, Int. J. Polym. Mater. 18, 51 (1992) 80. Z. Denchev, M. Sarkissova, H.-J. Radusch, T. Luepke, S. Fakirov, Macromol. Chem. Phys. 199, 215 (1998) 81. T. G. Fox, Bull. Am. Phys. SOC.1,123 (1956) 82. S. Fakirov, M. Sarkissova, Z. Denchev, Macromol. Chem. Phys. 197, 2889 (1996) 83. R. S. Porter, Thermochim. Acta 134, 251 (1988) 84. J. Devaux, P. Devaux, P. Godard, Makromol. Chem. 186, 1227 (1985) 85. W. A. McDonald, A. D. W. McLenaghan, G. McLean, R. W. Richars, S. M. King, Macromolecules 24, 6164 (1991) 86. W. Klee, D. 0. Hummel, Makmmol. Chem., Macromol. Symp. 52,141 (1991) 87. Z. L. Zhou, A. Eisenberg, J. Polym. Sci., Polym. Phys. Ed. 21, 595 (1983) 88. R. Longworth, US 3239176 (1967) 89. F. Kekufuta, H. Shimizu, J. Nakamura, Macromolecules 14, 1178 (1981) 90. M. Murano R.Yamadera, J. Polym. Sci., Polym. Phys. Ed. 2, 8 (1971) 91. K. L. Eersels, A. M. Aerdts, G. Groeninckx, Macromolecules 29, 1046 (1996) 92. S. Backson, A. Kenwright, R. Richards, Polymer 36, 1991 (1995) 93. M. Evstatiev, S. Fakirov, K. Friedrich, Appl. Compos. Mater. 2, 93 (1995) 94. J. C.Moore, J. Polym. Sci. A2,836 (1964) 95. E. F. Casassa, J. Polym. Sci. B5,773 (1967) 96. E. F. Casassa, Y. Tagami, Macromolecules 2, 14 (1969) 97. E. F. Casassa, Sep. Sci. 6, 305 (1971) 98. E. F. Casassa, J. Phys. Chem. 75, 3929 (1971) 99. H. C. Barth, J. Polym. Sci. B34, 1705 (1996) 100. Z. Grubisic, P. Rempp, H. Benoit, J. Polym. Sci., Polym. Lett. Ed. 5 , 753 (1967)

Sequential Reordering in Condensation Copolymers

389

101. W. Scouten, LLSzzeExclusion (Gel Permeation) Chromatography", in: Encyclopedia of Polymer Science and Engineering, edited by H. F. Mark, N. Bikales, Ch. Overberger, G . Menges, J. Wiley, New York 1988, vol. 3 pp. 501-523 102. W. Yau, J. J. Kirkland, D. D. Bly, Modern Size-Exclusion Liquid Chromatography, J. Wiley, New York 1974 103. C . Jackson, H.G. Barth, Handbook of Size-Exclusion Chromatography, edited by C . 3 . Wu, Dekker, New York 1995 104. Z. Denchev, A. Bojkova, A. DuChesne, H. Hocker, H. Keul, S. Fakirov, Macromol. Chem. Phys. 199, 2153 (1998)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 9

X-ray Analysis of Transesterification in Blends of Thermotropic Copolyesters

J. Blackwell, C . M. McCullagh

1. Introduction

A number of copolyesters give rise to non-periodic X-ray diffraction patterns, from which it is possible to obtain information about their sequence, i e . , microstructure. This effect was first observed for the wholly aromatic copolyesters prepared from phydroxybenzoic acid (HBA) and 2-hydroxy6-naphthoic acid (HNA), which form nematic liquid-crystalline melts as a result of their highly extended conformations. X-ray fibre diagrams of these copolymers contain non-periodic layer lines that are reproduced with high accuracy in the simulated diffraction data for models consisting of arrays of parallel chains of completely random sequence [l-41.Modification of the sequence statistics destroys this match, and all but minimal deviations from randomness can be ruled out, so that the X-ray technique provides a simple means to investigate the microstructure. Random sequences for lower molecular weight specimens of copoly(HBA/HNA) have been confirmed by proton nuclear magnetic resonance (lH NMR) analysis of dyad frequencies [ 5 ] , but there have been few such investigations due to the need to prepare deuterated analogues and their relatively low solubility. The X-ray techniques also provide analytical information, in that the d-spacings of the meridional maxima are shifted systematically with the monomer ratio and are specific to the particular composition. Thus one has a simple and

392

J. Blackwell, C. M. McCullagh

unique method of investigating the transesterification process, which will be the subject of this chapter. The occurrence of non-periodic scattering was a striking result when first reported, because the observed data also contained sharp Bragg reflections on the equator and on the layer lines that are characteristic of crystallinity [l].The properties of these materials are those of semicrystalline polymers, and it was natural to assume that there could be a blocky structure in which specific sequences were segregated. However, it was subsequently shown that three-dimensional models constructed of chain segments with non-identical random sequences could account for the observed data, provided there was limited axial register at their centres [6-81. More complex explanations in terms of segregation of identical non-repeating sequences from those in the overall random microstructure have also been proposed [9], although such segregation is not necessary to account for the observed data. Similar aperiodic scattering had been observed earlier for fibres of a number of other wholly aromatic copolyesters [lo], and has also been reported for analogous copolyamides and copolyimides [11,12]. As seen below, the aperiodic effect arises from the structural correlations in an extended chain conformation, in which there are approximately constant but unequal advances for the different monomers. This situation pertains in copoly(HBA/HNA) because of the chemical linkages and the rigidity of the aromatic and ester groups. However, the same kind of non-periodicity is observed for structures in which the chains contain “kinked” moieties, such as the TechnoraB copolyimide (Teijin) [ll]containing 1 , s as well as 1,4-linked phenylenes, and wholly aromatic copolyester carbonates [13],in which the carbonate groups might be expected to lead to non-linearity. Close examination of the data for the X7G copolyester prepared by reaction of HBA and poly(ethy1ene terephthalate) (PET) also reveals the presence of non-periodic meridional reflections [14], showing that the effect is not confined to the more rigid, wholly aromatic systems. This was demonstrated unambiguously for copoly(ethy1ene terephthalate/ethylene 2,6-dinaphthoate), which was found to give non-periodic data that changed systematically across the entire composition range [15]. The present chapter focuses on the copoly(HBA/HNA) system, which has received the most extensive study in view of its commercial application as the VectraB series of resins (Hoechst Celanese). Figure 1shows the X-ray scattering along the chain axis direction for melt spun fibres of six different comonomer ratios. At low HBA contents, four peaks are observed at scattering angles below 28 = 50°, but there are only three peaks at higher HBA contents. The peak positions can be seen to change continuously with monomer ratio: in fact the d-spacing of the first maximum varies from 6.64a to 8.08A as the HBA content decreases from 80% to 20%. This interference corresponds approximately to the average advance per monomer unit along the fibre axis, which is 6.35a for HBA and 8.37a for HNA. The

393

X-ray Analysis of

.-F

B

.I .-e,c Y

30jro 26/75

t

Figure 1. 8/28 X-ray diffractometer scans of the scattering along the chain axis direction of melt-spun fibres of copoly(HBA/HNA)for different monomer ratios: 75/25; 73/27;58/42;40160; 30170;and 25/75.The data are normalised so that the peaks at 28 = 43O have the same intensity measurements are sensitive enough to allow determination of the monomer ratio to within f l %in the intermediate composition range. These data present an ideal way to investigate possible non-randomness in monomer sequence, as it should occur initially in blends of two copoly(HBA/HNA) preparations with different monomer ratios. The data are very sensitive to changes in sequence distribution resulting from transesterification, so that the kinetics can be investigated. Transesterification of melt-blended polyesters in the isotropic state is reviewed in great detail in Chapters 1-3, 7 and 8 of this book, but there has been less attention to blends of liquid-crystalline polyesters. Economy et al. [5] reported that there was rapid transesterification when poly(HBA) and poly(HNA) were mixed as powders and held under pressure at 450OC. These homopolymers are infusible, but can be sintered above a liquidcrystalline transition in the 33&350°C range. Application of heat and pressure to the 50/50 molar ratio mixture of the homopolymers resulted in a

394

J. Blackwell,C. M. McCullagh

copolymer with a solid-nematic transition at approximately 254"C, which matched that of the 50150 random copolymer synthesised directly from the monomers, and pointed to complete transesterification. They also reported a reduction in the rate of transesterification between poly(HBA) and poly(HNA) for higher molecular weight polymers, which they attributed to either the decreased number of end-groups or a higher melt viscosity. However, other studies of a mesomorphic polyester [16] suggested that transesterification is independent of molecular weight in the isotropic melt (see also Chapter 5). The general physical manifestations of transesterification are shifts in the thermal transitions, increased compatibility, and changes in solubility. The changes in chemical structure have been followed by infrared (IR) [17,18] and NMR spectroscopy [5,19,20]. The latter techniques are most readily applied where the component monomers have significantly different chemical structures, and where the initial and final products are soluble. Lenz et al. [21] have described sequence changes in HBA-containing copolymers due to transesterification, and suggested that non-random sequence distributions can result from preference for homopolymer crystallisation. Analogous work with isotropic melt blends of polyesters suggested conversion to random or blocky microstructures, depending on the conditions (22-241. Blocky sequences can result from crystallisation-induced reaction at lower temperatures, whereas randomisation occurs at higher temperatures. Economy et al. [25] suggested that non-randomisation may occur when copoly(HBA/HNA) is annealed just below the solid-nematic transition. However, such non-randomness is not consistent with the available X-ray data, which rules out all but minimal blockiness, and has not been confirmed by other analytical techniques. Transesterification between isotropic polyesters has also been investigated using small angle neutron scattering (SANS) to follow changes in the apparent molecular weight in melt blends of deuterated and non-deuterated PET [15,26,27], from which it was possible to determine activation energies for the interchange reaction. Random scission and recombination of chains resulted in the formation of a block copolymer in which the deuterated blocks became progressively smaller as transesterification proceeded, resulting after about 1 h at 315°C in the formation of a completely random copolymer of deuterated and non-deuterated monomers. The different scattering cross sections of hydrogen and deuterium meant that the deuterated PET segments were effectively dissolved in a non-deuterated PET solvent, and hence their molecular weights could be determined. The extent of transesterification was determined to be a linear function of reaction time: the data followed an Arrhenius relationship, with an activation An activation energy of 152 kJ/mol, and a rate constant of 1.87 x energy was obtained for a more rigid, wholly aromatic copolyester [26], and differences in rate constants for different molecular weight polymers suggested an activated end-group mechanism.

X-ray Analysis of Transesterification

395

The work reviewed below analyses the progress of transesterification in blends of 75/25 and 30/70 copoly(HBA/HNA). Melt blends of the 75/25 and 30/70 copolymer were prepared with an overall 60/40 monomer ratio, and then subjected to sustained heat and pressure. The reaction was followed by monitoring the positions of the diffraction maxima along the chain axis direction for melt-spun fibres, and by thermal analysis. Thereafter, the diffraction effects due to progressive transesterification were simulated by extending the analyses of the scattering by non-periodic chains to include the changes in sequence that result from interchange reaction between chains of different composition. It will be seen that it was possible to reproduce the observed X-ray data, which could then be used to investigate the reaction kinetics. The experimental details are given in the papers by McCullagh et al. [28,29].

2. Scattering by aperiodic polymer chains The data in Figure 1 contain the meridional maxima seen in the fibre diagram. The greatest sensitivity to monomer composition is seen in the first maxima, which occur at d = 6.67, 7.15, and 7.90A for the 75/25, 58/42 and 30/70 copolymers, respectively. Figure 2 shows the simulated diffraction data for isolated chains of random sequence having the same compositions, which were predicted following the methods described in [6]. Briefly, the scattering along the chain axis direction, I ( Z ) ,is calculated as the Fourier transform of the z-axis correlation function of the structure, where Z is the reciprocal space coordinate and z is the distance along the chain axis:

Fjk(z) is the Fourier transform of the cross-convolutionof monomer j with monomer k: u

v

f is the atomic scattering factor and z is the atomic coordinate along the chain axis direction; j,u and k,v designate the uth atom of monomer j and vth atom of monomer k, respectively; pj is the molar fraction of monomer j , and zk is the length of monomer k projected along the chain axis. For a completely random copolymer, each Mjk term is simply pk, the content of monomer k. Non-randomness is simulated by modification of the M j k terms to allow for different combinatorial probabilities. Simple visual inspection shows that there is excellent agreement between the observed and simulated data in Figures 1 and 2, respectively.

J. Blackwell, C. M. McCullagh

396

1

HBA/HNA

73/27

40160

I

0

10

20

30

40

50

28 (degrees)

Figure 2. Simulated X-ray scattering for idmite (slightly sinuous) chains of copoly(HBA/HNA) with the same monomer ratios as in the observed data in Figure 1 The actual d-spacings agree well within the experimental uncertainty, and the small differences in intensity are probably due to packing effects that are not considered in Eq. (1).When one changes the statistics to consider blocky structures, then the peaks shift and split, and the good agreement is lost, so that all but minimal non-randomness can be ruled out [6-81. 3. X-ray analysis of copolyester blends

Curve a in Figure 3 shows the diffractometer scan for a melt blend of the 75/25 and 30170 copolymers, with an overall monomer ratio of 60140; the specimen was in the form of fibres drawn by hand immediately after rapid but thorough melt-blending. The data were recorded as a 8/28 scan along the fibre axis direction [28]. The d-spacings of the observed maxima are given in Tablel, together with those observed for the 75/25, 58/42, and 30/70 copolymers (from Figure 1).The data for the blend contain the maxima for the two pure components: in particular, we recognise the peaks at

397

X-ray Analysis of Transesterification

h

Y

Y

0

LO

20

30

40

50

28 (degrees)

Figure 3. 6/26 X-ray dzractometer scans along the fibre axis direction for: (a) a melt blend of 75/25 and 30/70 copoly(HBA/HNA) with an overall monomer ratio of 60/40, immediately after blending; (b) the same melt blend after compressionmoulding at 315°C for 60min

6.67 and 7.90A as due to the 75/25 and 30170 copolymers, respectively. Similar data are obtained for a physical mixture of fibres of the pure components. Melt-blending has not changed the peak positions, which means that the sequence distribution is as yet unaffected by transesterification. Some small differences do occur between the relative intensities for the initial blend and the physical mixture, which are probably due to poorer Table 1. d-Spacings (A)for the observed scattering maxima along the chain axis direction for fibres of copoly(HBA/HNA) specimens ~~~

~~

58/42

~

30/70

Physical mixturea blendb

Initial blend'

7.90

7.91

3.06

6.67 4.06 3.04

6.70 4.12 3.08

2.07

2.89 2.07

2.88 2.08

75/25

7.90 7.15f0.07

7.12 6.67 4.08

2.96f0.03 2.08f0.01

Moulded

2.87 2.08

2.96 2.08

amixture of fibres of 75/25 and 30/70 copoly(HBA/HNA); overall monomer molar ratio 60/40 binitid melt blend of 75/25 and 30/70 copoly(HBA/HNA); overall monomer molar ratio 60/40 'the same melt blend after compression-moulding at 315°C for 60 min

J. Blackwell, C. M. McCullagh

398

3-dimensional ordering of the chains. Figure 4 shows differential scanning calorimetry (DSC) scans for the physical mixture (curve (a)) and the initial melt blend (curve (b)). For the physical mixture, transitions at 287 and 303°C are observed, which are similar to the values of Tm for the pure components. For the melt blend, a single Tmappears at 227"C, well below that at 248°C seen for the 58/42 copolymer (T, for the 60/40 copolymer would be essentially the same). Thus, the thermal data suggest that melt-blending has resulted in the formation of a compatible blend, ie., a solution, an intimate mixture at the molecular level. The melt-blending also results in a significant (albeit qualitative) decrease in melt viscosity, to the extent that the melt becomes much easier to stir on the hot plate, consistent with the formation of a solution. DeMeuse and Jaffe [30]report a negative deviation from the rule of mixtures for the viscosities of melt blends of 30/70 and 75/25 HBA/HNA. A fundamental conclusion from the analysis of the scattering by aperiodic chains is that the positions of the maxima along the chain direction (ie., their &spacings) depend only on the monomer separations ( z j ) , proportions ( p j ) and the combination statistics ( M j k ) .These parameters are unaffected by mixing of parallel molecules, so it is unimportant whether the molecules are mixed at the molecular level or separated into large domains, as in the physical blend: in either extreme and for all intermediate degrees of mixing, the peak positions will be the same. Thus a specimen made up of random 30/70 and random 75/25 copolymer chains with an overall ratio of 60/40 will exhibit layer lines at d = 7.90 and 6.67& whereas the random

I

175

200

.

,

225

.

,

250

.

,

275

.

Temperature ("C)

,

300

.

325

Figure 4. DSC second heating s c m s of the same specimens for which the X-ray data are shown in Figure 3: (a) the physical mixture of melt spun fibres of 75/25 and 30/70 copoly(HBA/HNA); (b) the initial melt blend; (c) the melt blend after compression-moulding at 315' C for 60 min

X-ray Analysis of Transesterification

399

60/40 copolymer has a single intermediate layer line at d = 7.15 fi. I ( 2 ) in Eq. (1) is calculated as the Fourier transform of the z-axis correlation function, which effectively averages the scattering over all possible sequences. For a mixture of two copolymers, I ( 2 ) is based on the composition weighted sum of the correlation functions for the 75/25 and 30/70 components, and this sum is different from the correlation function for the pure random 60/40 copolymer. As a simple example, compare the dimer probabilities for 50/50 copoly(A/B) with those for a mixture of 75/25 and 25/75 copoly(A/B) in which the overall A/B monomer ratio is 50/50. In the random 50/50 copolymer, the molar fractions of four possible dimers are equal: [AA]= [AB]= [BA]= [BB]= 1/4. However, in the mixture, the ratios are [AA]= [BB]= 5/16 and [AB]= [BA]= 3/16. Similar differences are derived for the dimers, trimers, etc. Thus after Fourier inversion of the correlation functions, the predicted scattering patterns must be different, and for the blend we predict the non-periodic layer lines for the two independent components. So, for example, if we were able to prepare an intimate mixture of Kevlar and DNA (deoxyribonucleic acid) with all the chains arranged parallel to the fibre axis, we would see the layer lines predicted for both molecules independently, rather than some average of the two. This still pertains, even if the two polymers that are mixed are random sequences of the same monomers but with different compositions. Of course, the above discussion deals only with the positions of the layer lines along the chain axis direction. Mixing the chains is likely to have a major effect on the intensity distributions along the layer lines, just as this depends on crystallinity and axial stagger. But the positions of the layer lines along the fibre axis direction are unchanged by such interference effects. A change in the latter positions can only occur if there is a change in the combination statistics, and only transesterification could cause it in the present system. Curve (b) in Figure 3 is the diffractometer scan of the melt-blended specimen that was compression-moulded at 315°C for 60min. It is seen that there is only a single peak in the d = 6-8fi range, at d = 7.15& and the data now resemble those in Figure 1 for the 58/42 copolymer. The d-spacings of the observed maxima are given in Table 1 and are within the experimental error of those of the 58/42 copolymer. The small difference between 58/42 and 60/40 does not affect this conclusion. The 2% difference in monomer ratio would shift the first maximum for the random copolymer from d = 7.18a to d = 7.14& based on predictions following ref. [6]. T, values for the 58/42 and 60/40 copolymers are approximately the same [31].Curve (c) in Figure 4 shows the DSC data for the same compressionmoulded specimen, which has a single T, at 248"C, characteristic of the 60140 random copolymer and indicative of complete randomisation of the two component copolyesters. Figure 5 shows the scale-expanded X-ray diffractometer scans in the 20 = 7-17" region for melt-blended specimens that were compression-

J. Blackwell, C. M. McCullagh

400

a

10

12

14

16

26 (degrees) Figure 5. 8/28 X-ray diffractometer scans for 28 = 7-17' along the chain axis direction for the melt blend of 75/25 and 30/70 copoly(HBA/HNA) after compression-moulding at 315'C for 0, 9, 18, 27, 36, 45, 54, 60, 66, and 72min; overall monomer ratio of 60/40 in all specimens

moulded at 315°C for times ranging between 0 and 72min. The maxima characteristic of the two starting copolymers shift slowly toward each other as the moulding time increases. The component peaks were resolved using a Gaussian-Lorentzian curve fitting program, and their d-spacings are plotted against moulding time in Figure6. We observe a steady change in d-spacings until the two peaks converge after 60min. DSC scans of the same specimens are shown in Figure 7, where a steady increase in T, with moulding time is seen, from 227°C characteristic of the compatible blend of the two initial copolymers, to 248°C for the 60/40 random copolymer. There is an approximately linear change in T, with time, which correlates

40 1

X-ray Analysis of Trmsesterification 1

8.02

8‘ 6.5 4

0

10

20

30

40

Time (min)

50

60

I

70

Figure 6. Plots of d-spacing against moulding time for the first and second intensity maxima for the melt blend of 75/25 and 30/70 copoly(HBA/HNA) (data from Figure 5)

with the changes in d-spacing in Figure6. From the above results, it is clear that melt-blending the 30/70 and 75/25 copolymers results first in the formation of a compatible blend. Meltpressing at 315°C results in slow progressive transesterification, yielding a completely random copolymer of intermediate composition after 60 minutes. Our data indicate that X-ray diffraction can be used relatively easily to follow the transesterification reaction in the HBA/HNA copolymers. The kinetics of the reaction is investigated further in the next section. DeMeuse and Jaffe [30] had interpreted their X-ray data for melt blends of copoly(HBA/HNA) to suggest that interchange does not occur in the nematic state. They reported that the first meridional maxima appeared to merge into a single broad peak, which they suggested to be composed of the two peaks of the original components. This apparent conflict with our results is probably due to incomplete mixing of the initial blend: transesterification leading to changes in sequence distribution of an incompatible or poorly mixed blend would occur preferentially at the edges of the domains, producing copolymers of intermediate compositions in these regions. The first diffraction maxima of the latter copolymers would occur between those of the initial copolymers, resulting in a “filling in” of the region between the two initial peaks. At the early stages of the reaction, one would obtain a single broad peak in the region d = 8.5-6.3 A,which would gradually sharpen and become a narrower peak at d = 7.13 A when the (60/40) copolymer becomes fully randomised and homogeneous.

J. Blackwell, C. M. McCullagh

402

248

60 min 239

27 min

235

1 234

18 min

231 9 min

227

175

200

225

250

275

Temperature ('C) Figure 7. DSC second heating scans for the specimens used to generate the X-ray data in Figure 5: melt blends of 75/25 and 30/70 copoly(HBA/HNA), compression-moulded for the specified times. The values of T, are shown by each peak

4. Kinetics of transesterification

The formalisations presented above to predict the X-ray scattering for aperiodic chains include treatment of all possible comonomer sequences, and are sufficiently general to allow extension to consider progressive interchange reactions. Our simulation of the scattering patterns (301 utilised atomic coordinates for HBA and HNA based on the structures of low molec-

X-ray Analysis of Transesterification

403

ular weight model compounds [32-351. The HBA coordinates were identical to those used previously IS]. For HNA, we used two sets of coordinates corresponding to the cis and trans conformations of the monomer. Analyses of the mechanical properties of HBA/HNA fibres [36] suggest that cisHNA is the predominant conformation in these copolymers. The cis:trans ratio for HNA was refined to obtain the best agreement between the observed and calculated d-spacings for each comonomer composition, and ratios of 100:0, 81:19 and 68:32 were used for the 75/25, 60140 and 30170 copolymers, respectively. Non-linearity of the extended chains was modelled by incorporating a truncated Gaussian distribution of lengths ( z j ) for each monomer, centered on an average length that was 0.12A less than the distance between the ester oxygens, as has been described previously [6]. 4.1. Random transesterification

The blend is treated as if it is a mixture of two totally different copolyesters: B'/N' and B"/N" designate the monomers of the 30170 and 75/25 copolymers, respectively, which have molar fractions w and 1 - w, respectively [30].Initially we considered homogeneous mixing of the molecules, and random interchange defined by a cross reaction parameter, r , that increases from 0 (no reaction) to 1 (complete randomisation) during the reaction. The treatment models detectable reactions only: those occurring between 75/25 chains, for example, do not affect the sequence statistics. At the start of the reaction (r = 0), the Mjk terms of Eq. (1) form a 4 x 4 matrix:

Here the zeroes indicate there has been no cross reaction. When transesterification is complete (T = l ) , there is equal (composition weighted) probability for a given monomer to be followed by one that originates from the same or the other copolymer. After complete randomisation, the structure is described by:

which leads to I ( 2 ) for the random copolymer of intermediate composition. Over the course of the reaction, the intermediate copolymers can be thought of as block copolymers composed of alternating segments of the two starting copolymers, the lengths of which decrease progressively. The matrix of Mjk

J. Blackwell, C. M. McCullagh

404

4

a

12

28 (degrees)

16

20

Figure 8. Simulated X-ray scattering along the chain axis direction for randomly transesterified blends of 30/70 and 75/25 copoly(HBA/HNA) in the range T = 0-1, in 0.1 increments of T

L = W + r(1 - w) R = L [ ( l - r ~ / L ) / (-l w)] Calculated f(2) data in the region of the first non-periodic maxima, derived using the random transesterification model for different values of r , are shown in Figure8. As T increases from 0 (no reaction), the peaks

X-ray Analysis of Transesterification

405

originally at d = 6.70 and 7.88A shift toward each other at r = 0.1, and merge into a single peak at d = 7.16A when r = 0.30. This peak becomes narrower at r = 0.40, beyond which its position and width are constant. By definition, randomisation is not complete until r = 1, but the breadth of the peaks predicted for random copolymers and their relatively close proximity means that for r = 0.40 one cannot distinguish the transesterification product from the completely random 60/40 copolymer. In order to explore further the sensitivity of the scattering data to r , we repeated the above simulations for transesterification in a hypothetical 60/40 blend of the two homopolymers, poly(HBA) and poly(HNA). The first maxima predicted for the original homopolymers shifted toward one 0.70, and became another and also broadened, until they merged at r indistinguishable from the peak at d = 7.16A for the random copolymer at r 0.85. This difference in behaviour of the homopolymer and copolymer blends is simply due to the fact that the copolymer components are already closer to the intermediate composition. For the blend of the 30/70 and 75/25 copolymers, at r = 0.40, the monomer pairs probabilities are [BB]= 0.375, [BN]= [NB]= 0.225, [NN]= 0.175, which are very similar to those for the random 60/40 copolymer, where [BB]= 0.360, [BN]= [NB]= 0.240, [NN] = 0.160. Consequently, further monitoring of the transesterification reaction is beyond the limits of the experimental error. For the homopolymer blend, a comparable situation is not reached until r = 0.90. For completeness, we considered the possibility that non-random sequence distribution may occur due to unequal rates of interchange at the different monomers. The combinatorial probabilities can be altered by weighting each hfjk term by a kinetic factor for the difference in reaction rates at each monomer. We also considered possible non-randomness due to the existence of domains, ie., incomplete mixing, such that a fraction of one of the polymers would be unavailable for reaction with the other. In all cases, deviations from non-randomness led to divergence from the excellent obtained for perfectly random reaction in a homogeneous blend, and need not be considered further. N

N

4.2. Wnsesterification kinetics

If x represents the number of transesterification events in a polymer with n ester linkages, the probability of reaction, P, at any given site is

P = 1 - (1 - l/n)z

(8)

The number of transesterification events per monomer is u = 5/72, and as n+ca

J. Blackwell, C. M. McCullagh

406

4

I

I

I

I

a

12

16

20

28 (degrees)

Figure 9. Simulated X-ray scattering along the chain axis direction for randomly transesterified blends of 30/70 and 75/25 copoly(HBA/HNA) in the range r = 0.00-0.02, in 0.02 increments of T (These data would fall in the region of the lowest three curves in Figure 8)

If all ester bonds are equally accessible, the first reaction at a given site alters the chain statistics by altering first nearest neighbour probabilities. (Subsequent reactions at the same site have no effect on the overall statistics [30].) For the component of molar fraction w, the probabilities of the same or different pairwise combinations as a result of the reaction with the other copolymer species are

The ratio PsamelPdifferent can also be derived from the starting monomer ratios and the combinatorial probabilities in matrix (5):

407

X-ray Analysis of Transesterification 8.0-

onon

nn

7.5 -

(4 7.0

;

onn

a %n

nononnn

6.5

5

8.0

1

0.0

b' X

0.2

0.6

0.4

0.8

1.0

Y

Figure 10. (a) Predicted d-spacings (from Figure 9) predicted d-spacings (0 and 0 ) vs. v

us. r ; (b)

observed ( x ) and

Hence

and u=ln(+

r w(1 - r)

)

Thus, we can determine u as a function of time since r can be determined during the reaction by matching the observed and calculated d-spacings of the non-periodic scattering maxima. Figure 9 shows the predicted scattering for blends of the 30/70 and 75/25 copolymers for r = 0-0.20,in increments of 0.02. The component peaks were resolved using a Gaussian-Lorentzian peak fitting program, in

J. Blackwell, C. M. McCullagh

408

1

la5: U

I

315% 295OC

1.0-

I

0.0

0

1000

2000

3000 4000 5000 t (sec)

6000

Figure 11. Number of reactions per monomer, u, ws. reaction time, t , at 315'C (0)and 295°C (A)

the same way as was done for the observed data. The d-spacings derived from the peak positions are plotted against r in Figure 10a, which is very similar to the plot of the observed d-spacings against time in Figure 6 , provided that we assign the data recorded after 60 minutes in the melt to r = 0.40. If we assume that the reaction proceeds linearly with time, as has been reported for other polyesters [l0-12], the reaction is complete after 150min, and the reaction time is t = 9 x 103r seconds. This allows us to calculate v, the number of transesterification events per monomer, for each value of T using Eq. (13), and the observed and calculated d-spacings are replotted against v in Figure lob. The monomeric transesterification rate constant, k, is given by v = kt, and v is plotted against t in Figure 11. The upper straight line is for the experimental data in Figure 6 , and leads to k3150c = 2.3 x 10-4s-'. The lower plot is the data obtained for a second series of blend specimens held in the melt for different times at 295°C [l], and yields IC29pc = 0.83 x 10-4s-'. Assuming an Arrhenius relationship, we estimate an activation energy of 142 kJ/mol, in good agreement with the previously reported values of 152kJ/mol for PET [lo] and 142-173kJ/mol for a wholly aromatic copolyester [ll].However, the rate constants are an order of magnitude smaller than the value of k z s o ~ c= 1 . 8 7 ~ measured for PET, which may reflect the nematic structure and higher melt viscosity of copoly(HBA/HNA), i. e., a lower molecular mobility. 5. Conclusions

Wide-angle X-ray diffraction is a simple method for investigation of the transesterification reaction in wholly aromatic copolyesters, and probably therefore in a wide range of polyesters. The non-periodic scattering maxima for the transesterification intermediates in blends of 30/70 and 75/25

X-ray Analysis of Transesterification

409

copoly(HBA/HNA) can be simulated by a model in which there is complete molecular dispersion and random reaction between chains. The activation energy derived for the completely random reaction is very similar to those reported by other authors in neutron diffraction studies of transesterification of non-liquid-crystalline polyesters. However, the rate constants are an order of magnitude lower than for the latter systems, which may reflect the differences in melt viscosity and mobility in the liquid-crystalline and isotropic states.

Acknowledgement The work in this laboratory was supported by NSF MRG No. 91-22227.

References 1. R. A. Chivers, J. Blackwell, G. A. Gutierrez, Polymer 25,435 (1984) 2. J. Blackwell, A. Biswas, G. A. Gutierrez, R. A. Chivers, Faraday Disc., Chem. SOC.79, 73 (1985) 3. J. Blackwell, R. A. Chivers, G. A. Gutierrez, A. Biswas, J. Macromol. Sci.Phys. B24, 39 (1985) 4. A. H. Windle, C. Viney, R. Golombok, A. M. Donald, G. R. Mitchell, Faraday Disc., Chem. SOC.79, 73 (1985) 5. A. Muhlebach, J. Economy, R. D. Johnson, T. Karis, J. Lyerla, Macromolecules 23, 1803 (1990) 6. A. Biswas, J. Blackwell, Macromolecules 21,3146 (1988) 7. A. Biswas, J. Blackwell, Macromolecules 21,3152 (1988) 8. A. Biswas, J. Blackwell, Macromolecules 21,3158 (1988) 9. S. Hanna, A. H. Windle, Polymer 29, 207 (1988) 10. J. Blackwell, G. A. Gutierrez, Polymer 2 3 , 671 (1982) 11. J. Blackwell, R. A. Cageao, A. Biswas, Macromolecules 2 0 , 667 (1987) 12. T.-Z. Wu, S. N. Chvalun, J. Blackwell, S. Z. D. Cheng, Z. Wu, F. W. Harris, Polymer 36,2123 (1995) 13. A. I. Schneider, J. Blackwell, H. Pielartzik, A. Karbarch, Macromolecules 24, 5676 (1991) 14. J. Blackwell, G. Lieser, G. A. Gutierrez, Macromolecules 16,1418 (1983) 15. X. Lu, A. H. Windle, Polymer 37, 2027 (1996) 16. M. H. Li, A. Brulet, P. Keller, C. Strazielle, J. P. Cotton, Macromolecules 26, 119 (1993) 17. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22,229 (1982) 18. L. H. Wang, Z. Huang, T. Hong, R. S. Porter, J. Macromol. Sci-Phys. B29, 155 (1990) 19. A. Muhlebach, R. D. Johnson, J. Lyerla, J. Economy, Macromolecules 21, 3115 (1988) 20. J.-I. Jin, J.-H. Chang, K. Hatada, K. Ute, M. Hotta, PoZymer33,1374 (1992) 21. R. W. Lenz, J. Jin, K. A. Feichtinger, Polymer 24,327 (1983) 22. A. M. Kotliar, J. Polym. Sci., Macromol. Rev. 16,367 (1981) 23. R. S. Porter, L.-H. Wang, Polymer 33,2019 (1992)

410

J. Blackwell, C. M. McCuUagh

24. R. S. Porter, J. M. Jonza, M. Kimura, C. R. Desper, E. R. George, Polym. Eng. Sci. 29, 55 (1989) 25. J. Economy, R. D. Johnson, J. Lyerla, A. Muhlebach, in: Liquid Crystalline Polymers, edited by R. A. Weiss, C. K. Ober, ACS, Washington, DC 1990 26. J. Kugler, J. W. Gilmer, D. Wiswe, H. G. Zachmann,K. Hahn, E. W. Fischer, Macromolecules 20, 116 (1987) 27. W. A. MacDonald, A. D. W. McLenaghan, G. McLean, R. W. Richards, S. M. King, Macromolecules 24, 6164 (1991) 28. C. M. McCullagh, J. Blackwell, A. M. Jamieson, Macromolecules 27, 2996 (1994) 29. C. M. McCullagh, J. Blackwell, A. M. Jamieson, Macromolecules 30, 4837 (1997). 30. M. T. DeMeuse, M. Jaf€e, Mol. Cryst. Liq. Cryst., Nonlzn. Opt. 157, 535 (1988) 31. G. W. Calundann, M. J d e , Proc. Robert A . Welch Conf. Chem. Res., Synth. Polymers, p. 247, 1982 32. J. M. Adams, S. E. Morsi, Acta Crystallogr. B32, 1345 (1976) 33. J. P. Hummel, P. J. Flory, Macromolecules 13, 479 (1980) 34. J. Trotter, Acta Crystallop. 14,101 (1961) 35. H. C. Watson, A. Hargreaves, Acta Crystallop. 11,556 (1958) 36. M. J. Troughton, A. P. Unwin, G. R. Davies, I. M. Ward, Polymer29, 1389 (1988)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 10

Effects of Transreactions on the Compatibility and Miscibility of Blends of Condensation Polymers

M. Xanthos, H. Warth

1. Principles of blend compatibilisation

Few polymers form truly miscible blends characterised by a single glass transition temperature (Tg) and homogeneity on a 5-10nm scale with domain sizes comparable to the dimension of a macromolecular statistical segment. The majority of blends are immiscible, i e . , possess a phase separated morphology. Blends of this type are often preferred over the miscible ones since they combine some of the important characteristics of both blend constituents. Blend composition, viscoelastic properties of the components, and interfacial adhesion are among the parameters known to control the size and morphology of the dispersed phase and its stability to coalescence. Heterogeneous blends of technological importance are termed “compatible” and they constitute the majority of the commercial blends introduced in the past 30 years. In such blends, often known also as alloys, satisfactory physical and mechanical properties are related to the presence of an interphase modified through compatibilisation and the formation of a fine dispersed morphology resistant to phase separation. Polymer compatibility, which may lead ultimately to complete thermodynamic miscibility, may be enhanced by various methods. In addition to cocrystallisation and cocrosslinking, strong interactions such as acidbase or ion-dipole, hydrogen bonding and transition metal complexation

412

M. Xanthos, H. Warth

between suitably functionalised components, have been shown to improve compatibility. More commonly, compatibility is promoted through interfacially active copolymers (e.g., block, graft, random) with segments capable of specific interactions and/or chemical reactions with the blend components. The copolymers may be added separately, or formed in situ through chemical reactions, often catalysed by the addition of low molecular weight (MW) compounds. Reactive and non-reactive compatibilisation of polymer blends has been the subject of numerous reviews [l-71. During blending of a pair of suitably functionalised polymers A and B, interchain block, graft, or random copolymers may be formed at various concentrations through covalent or ionic bonding. According to Brown [3], the following are five general types of chemical processes/reactions by which interchain copolymer formation has been achieved in extruder reactors:

1. Chain cleavage/recombination (main chain of A/main chain of B or main chain of A/end-group of B); block and/or random. 2. End-group of A/end-group of B; block. 3. End-group of A/pendant functionality of B; graft. 4. Covalent cross-linking: pendant functionality of Alpendant functionality of B or main chain of A/main chain of B; graft (crosslinked network). 5. Ionic bond formation between A and B; usually graft (usually crosslinked). The compatibilisers formed in situ have segments that are chemically identical to those in the respective unreacted homopolymers and are thought to be located preferentially at the interface, thus lowering interfacial tension, and also promoting mechanical interlocking through interpenetration and entanglements. The majority of commercial compatibilised blends belong to Categories 3-5 above and are produced through reactions that can easily take place across phase boundaries within extrusion residence times. Examples of such reactions involve the following functionalities: anhydride or carboxyl with amine; epoxy with anhydride or carboxyl; oxazolin with carboxyl; isocyanate with hydroxyl or carboxyl; carbodiimide with carboxyl. Macroradical recombination in polyolefins and interchain ionic salt formation are examples from Categories 4 and 5 , respectively. Interchange reactions that could potentially compatibilise binary blends of condensation polymers (polyesters, polyamides and their combinations) belong to Category 1 above and will be reviewed in this chapter with emphasis on methods of following transreactions. 2. Dansreactions applied t o blend compatibilisation

A list of possible transreactions between polyesters, polyamides, and their combinations that could lead to the formation in situ of interchain copolymers is shown in Tablel. &actions include direct interchange of groups

413

Effects of Transreactions on Compatibility and Miscibility

present in the main chains (ester/ester, amidelamide and esterlamide), and intermolecular reactions of groups in the main chain with terminal end-groups (e.g., hydroxyl and carboxyl in polyesters, amine and carboxyl in polyamides). These reactions may be considered to involve chain cleavage, followed by recombination of free end-groups to give an equilibrium molecular weight distribution of random and/or block copolymers [3]. In the early stages of such a process, block copolymers may predominate over random copolymers, but the exact nature and distribution of different types of copolymers have not been always determined in the literature. The complex and competitive processes involved plus the variety of experimental conditions can result in a range of reaction rates and resultant molecular chain microstructures [B].The resulting copolymers may be obtained as mixtures with lower MW homopolymers usually having broad molecular weight distribution (MWD). Such block copolymers may have molecular weight less than the sum of the two homopolymers and may not be as effective as blend compatibilisers as are block copolymers formed by the end-group/end-group reaction (Category 2 above) [3]. Reviews covering transreactions in polyester and in polycarbonate binary blends have apTable 1. Possible transreactions between polyesters, polyamides and their combinations Reaction type

Reactive group

Co-reactive group

Ester interchange (transesterification) Alcoholysis

Ester

Aminolysis

Ester

Hydroxyl (alcohol, phenol) Amine

Acidolysis

Ester

Carboxyl

Amide-ester exchange Amide-amide exchange (transamidation) Alcoholysis

Amide

Ester

Amide

Amide

Ester

Transreact iona

+ B-COO-B A-COOR + B-OH F! A-COO-B + R-OH A-COOR + B-NHz 3 B-CO-NHB + R-OH A-COOR + B-COOH S A-COO-B + RCOOH A-CO-NHA + B-COO-B 3 A-CO-NHB + B-COO-A A-CO-NHA + B-CO-NHB A-COO-A

F! A-COO-B

Ester

Amide

Aminolysis

Amide

Acidolysis

Amide

F! A-CO-NHB

Hydroxyl (alcohol, A-CO-NHR + B-OH phenol) 3 A-COO-B RNHz Amine A-CO-NHR B-NHz F! A-CO-NHB RNHz Carboxyl A-CO-NHR B-COOH S A-CONHB RCOOH

"A and B indicate the respective polymers

+ + + + +

414

M. Xanthos, H. Warth

peared in the last two decades [&lo]. Transreactions take place more easily at the high temperatures required for melt-blending and strongly depend on initial component miscibility and blending conditions. These include temperature, mixing time, viscosity match, and the presence of residual catalysts from polymerisation (e.g., Ti compounds, Sbz03 in polyesters) and inhibitors. Transreactions can occur in both miscible and immiscible blends. In the former, miscibility is caused directly only by interaction of components and transreactions may play a secondary role. In the latter, the reactions presumably occur at the interfacial regions between the components. Several techniques have been used to follow transreactions and differentiate them in terms of their effects on miscibility from strong interactions. Application of these techniques in polyester systems has been reviewed by Porter and Wang [8]. Table2 lists available techniques applicable to the characterisation of transreacted systems. FT-IR has been used extensively to detect new components produced by reactions. NMR methods seem to be the most accurate means to determine degrees of transreaction and have provided insight about miscibility and the chemical changes occurring within polyester pairs; however, NMR methods may become inaccurate at very small or very high levels of transesterification (see also Chapter 1). DSC, differential thermal analysis (DTA), and dynamic mechanical thermal analysis (DMTA) have also been used as primary methods for the study of interchange reactions, since thermal or thermomechanical behaviour changes from the onset of transreactions. Usually, the Tgof immiscible components is shifted and a new single Tg may appear at high extents of transreaction. For crystallisable components, since compositional sequence in the main chains is changed by the reacTable 2. Characterisation techniques applicable to transreaction Method

Measured DroDertv

'H Nuclear magnetic resonance (NMR) 13C NMR Infrared (IR), Fourier transform (FT) IR spectroscopy Differential scanning calorimetry (DSC) Thermogravimetric analysis (TGA) X-ray fluorescence Size-exclusion chromatography (SEC) Rheometry Fluorescence

Randomness of copolymer (triad) Miscibility Crystallinity

Cross polarisation/Magic-angle spinning 13C NMR

T,, Ts, crystallinity

Content of specific components (%) Residual catalyst Blend components Blend properties Composition as per component wavelength Miscibility

Effects of Transreactions on Compatibility and Miscibility

415

tions, melting and crystallisation temperatures (T, and T,)and fractional crystallinity may be depressed (see also Chapter 5). Crystallinity related changes may also be followed by X-ray diffraction (see also Chapter 9). Since the skeletal structure, MW, and MWD are changed simultaneously by transreactions, thin-layer chromatography and solubility evolution have been used to follow reaction products. Morphological changes attributed to transreactions may also be followed by scanning or transmission electron microscopy (SEMITEM) (see also Chapter 8). It should be noted that heating of polymer melts or solutions at elevated temperatures may promote further transreactions; this means that the randomness of the examined polymer blend/copolymer may be changed during DSC, TGA, or rheology measurements, or even during heating required to completely dissolve semicrystalline polymers in suitable solvents. Compatibility improvement through copolymers produced by transreactions may not be advantageous in all cases. Extensive transreactions, leading to the formation of random copolymers, may result in the loss of unique properties of the blend components, a severely depressed melting temperature compared to the crystalline component, molecular weight degradation, and even COz evolution in some polyester blends [ll].Deterioration in certain properties such as embrittlement in polycarbonate (PC)/polyester blends has been reported by Golovoy et al. [12]; reduced solvent resistance in semicrystalline polyester blends because of reduced crystallinity has been reported by Kimura et al. [13]. Many experimental studies have sought to prevent extensive transesterifkation rather than promote it. Chemical strategies have been devised to control ester-exchange reactions by using various stabilisers such its organophosphite additives, especially in the molten state; see, for example, Cheung et al. [14]and Golovoy et al. [15] for poly(ethy1ene terephthalate) (PET)/polyarylate (PAr), and Cox et al. [16] for PET/poly(ethylene naphthalate) (PEN) (see also Chapters 3 and 6). There exist numerous polyester/polyester, polyester/polyamide and polyamide/polyamide blends that have been commercialised in the past 20 years. Although some of these systems may be inherently miscible due to strong interactions, compatibilisation due to controlled transreactions would undoubtedly play a rolefor their success in the market place. In his review of polymer alloys and blends, Utracki [17] identified several suppliers from USA, Europe, and Japan associated with the production of the following blends: polyamide (PA)/PC - three suppliers; poly(buty1ene terephthalate) (PBT)/PC - eight suppliers; PET/PC - seven suppliers; PBT/PET - three suppliers; PAr/PET or PC - two suppliers. Typical characteristics of these blends include good mechanical properties, high heat distortion temperature, good solvent and chemical resistance, and good processability. Utracki’s data are examples of the status of condensation polymer blends up to 1993. A more up-to-date list would undoubtedly be much longer, given the advances in the development of

416

M. Xanthos, H. Warth

specialty and engineering blends involving, for example, liquid-crystalline polyesters and the current interest in PEN as a complement or substitute for PET in barrier applications.

3. !Ramreactions applied to specific binary blends 3.1. Polyester/polyester blends 3.1.1. Polycarbonate/poly (ethylene terephthalate) Blends of PET and bisphenol A-carbonic acid copolymer (PC) are completely immiscible over a wide range of compositions. During melt processing transesterification occurs and, with an increase in the reaction extent, the compatibility of the blends increases from completely incompatible to partially compatible, then to complete compatibility. In addition, with an increase of reaction extent, the crystalliisation ability and crystallinity of PET decreases [18,19]. If an equimolar blend of PET and PC is subjected to thermal treatment at 280"C, an amorphous random copolymer is formed that no longer reveals melting or crystallinity. Additional annealing of such samples below the melting temperature of PET results in restoration of the crystallisation ability; this is due to crystallisation-induced sequential reordering, from random to block copolymer, by means of transreactions which close the cycle of transformations from two homopolymers via block and random copolymer back to a block copolymer. The latter copolymer shows DMTA peaks of two amorphous phases, clearly separated and with distinct individual Tgvalues [20-223. Yoon et al. [23] studied the competition between phase separation and transreactions and found that: (i) The global structure coarsens at 200°C due to the dominance of phase separation over transesterification and melts at 220°C due to the dominance of transesterification at the domain interface; however, transesterification is slow but still significant,even at 200°C. (ii) An intricate balance of transesterification and phase separation rates controls global and interfacial structures. (iii) Interfacial structures become measureable under certain conditions, and the interfacial region between PC or PET and the copolymers generated by transesterification increases with time. Kinetic studies using IR and NMR spectroscopysuggested that the most likely mechanism is a direct ester-ester interchange reaction which can also be catalysed by additives such as Ti derivatives [24]. Catalysts enhancing transesterification have been ranked by reactivity [25]: Ti(0Bu)r >> Sm(o-formylphenolate)3 > Eu(o-formylphenolate)3 > Ca Sb > CeAc3. Other compounds based on Er and T b show lower catalytic activity (for more details see Chapter 2).

+

Effects of Transreactions on Compatibility and Miscibility

41 7

Transesterification is inhibited by phosphites, imide- and/or oxazolinecontaining compounds, and optionally hindered phenols [26]. Using 31P NMR it was also shown that phosphonates, obtained by the in situ hydrolysis of phosphite stabilisers, inhibit transesterification during melt processing [15,27].

3.1.2. Polycarbonate/poly(butylene terephthalate) The miscibility of PC and PBT is controversially discussed in the literature. In solution-cast blends immiscibility was found [28], while in the melt, partial miscibility has usually been observed [29,30]. It is known that transesterification takes place in the melt and that copolyesters formed by transesterification change the compatibility of PC and PBT [31]. Transesterification is a function of processing conditions [32], and depends on time of thermal treatment [28] and temperature [33], as well as on blend composition, where higher PC concentration results in higher degrees of transesterification [28]. Tattum et al. [29] prepared a series of PC/PBT blends of varying weight fraction (10:90-9O:lO) via melt-blending. Microscopy showed that the blends formed dispersed-continuous or bicontinuous two-phase morphologies, dependent on the weight fraction and rheological behaviour of the homopolymers. Blends containing 10% of either homopolymer showed no multiphase structure and the development of a significant interphase. No evidence was found for the occurrence of direct interchain transesterification, although transesterification via acidolysis occurred at temperatures higher than 26OOC owing to degradation of PBT. Increasing transesterification resulted in a progressive reduction in T, and T,,as well as in the degree of crystallinity, with the development of a mixed-phase glass transition at around 90°C. Transesterification also induced a significant change in blend morphology, from a coarse (5-10 pm) bicontinuous structure, when uncatalysed, to a submicron bicontinuous structure at low degrees of reaction [34]. Devaux et al. [35-371 suggested a direct reversible ester-ester interchange reaction as the most likely mechanism (see also Chapter 3). The transesterification can be catalysed by the titanium residue present in PBT and can be stopped at various levels by some additives capable of complexing the titanium catalyst [34]. Phosphorus compounds are mainly used as transesterification inhibitors, e.g., phosphites [32,38], phosphates [39,40], phosphonates (15,411, and phosphoric acid [42]. Other stabilisers include sulfates or sulfites [43], boric acid [44], amino- or hydroxy-acids [45], e hydroxy aldehydes or ketones [46], and organosilicates [47].

3.1.3. Polycarbonate/polyarylate On controlled thermal treatment of physical blends of uncatalysed bispheno1 A polycarbonate and PAr (a bisphenol A terephthaloyl/isophthaloyl

418

M.Xanthos, H.Warth

copolyester) blends at 250"C, reaction progression from block to random copolymer has been traced by DSC,13CNMR, and SEC;completely transreacted blends were amorphous and showed a single Tgvalue [48-501.Transesterification has also been studied in blends prepared by extrusion in a Brabender batch mixer at 30-40 rpm and 250-300°C [51,52]and in extruders [12].Blends between 15/85 and 85/15 parts polycarbonate/polyarylate were prepared on a twin screw extruder (TSE) at 300rpm and temperature profile 260-300°C. A 50:50 blend was also prepared using a single screw extruder (SSE)at 285°C. The materials were characterised by DSC which showed two Tgvalues for blends compounded on the SSE (or solution-cast) but only one Tgfor blends prepared on the TSE. Blends prepared on the SSE phaseseparated during annealing, while those prepared on the TSE did not. Moulded parts from blends prepared on the TSE showed tensile and impact values below those predicted by the rule of mixtures due to a combination of molecular weight degradation and transreactions during processing. Transesterzcation of PAr with PC occurs during devolatilisation of mixed polymer solutions in an extruder [53].The relative amounts of the two solutions varied between 80:20 and 10:90.The combined solution was concentrated to 45% solids and fed to a devolatilising TSE with 50s r e s idence time at 220-280°C and 20-60s residence time at 280-320°C. For each formulation the product was isolated free of solvent and had a single Tg by DSC. Moulded test parts had notched impact strength higher than test parts of comparable control formulations made by coextruding a mixture of the solid homopolymers at 360°C with vacuum venting

3.1.4. Poly(ethy1ene terephthalate)/poly(butylene terephthalate) PET and PBT are miscible in the amorphous phase. Their blends have a single glass transition which varies with composition. Thermal treatment results in transesterification and upon cooling PET crystallises faster than PBT; blends show multiple crystallisation peaks which are associated with chain entanglement, as found by Yu and Choi [54,55].By contrast, Escala and Stein [56]found by X-ray, DSC, and IR spectroscopy no indication of cocrystallisation, with crystallisation rates affected primarily by the degree of supercooling of each component in the blend and by the influence of blending on the glass transition temperature. Transesterification is temperature dependent. Backson et al. [57]heated PET and PBT homopolymers and their mixtures in the absence of oxygen at 573 K for 30 min and at 476K for 6 h. At 573 K, transesterification was complete and a single random copolymer was obtained. Two block copolyesters were formed when the mixtures were heated at 476 K for 6 h. Kim and Ha [58]studied the various competing different mechanisms and concluded that the direct ester-ester interchange reaction prevails over the alcoholysis reaction.

Effects of 'lkansreactionson Compatibility and Miscibility

419

Jacques et al. [59] also proved the occurrence of ester-interchange reactions during PET/PBT blend processing by 13C NMR measurements. By this method it was shown that titanium alkoxide is an efficient catalyst, and that triphenyl phosphite can be used as a stabiliser. Other transesterification catalysts include organotinanates and organozirconates [60]. Besides triphenyl phosphite [59,61], other stabilisers can be used, which include triphenyl phosphate [62], hydrogen phosphates of Zn or Ca [63], phosphate ester [64] and the reaction product of Sbz03 with talc (Oncor 75RA) [65] (see also Chapter 2). Some authors [15,27] suggest that conversion of the phosphite group to a phosphonate moiety, probably via hydrolysis, is a prerequisite for an effective inhibition of transesterification. This conversion occurs at room temperature over long periods of time or readily during melt compounding if the polymers are not completely dry. However, if rigorous drying is employed and phosphite conversion does not occur, then transesterification is not arrested (see also Chapters 3 and 6). 3.1.5. Poly(ethy1ene terephthalate)/poly(ethylene naphthalate) The properties of PEN expand possible polyester markets to include applications which demand a level of performance beyond the capabilities of current PET packaging. Needs for a clear polyester with properties between those of PET and PEN exist for hot-fill packaging, cosmetic packaging, pharmaceutical packaging, niche film and fibre applications, small size carbonated beverage bottles and returnable, refillable bottles [66]. By controlled transesterification the properties can be custom-designed. A certain degree of crystallinity is needed for injection/stretch blow moulding applications, while it is not needed for extrusion blow moulding or sheet production. The desired degree of transesterification is 25-35%. Typical residual catalysts, such as those based on Ti, Sb, Co, and Mn in concentrations between 20-500 ppm, facilitate transesterification. However, phosphorous stabilisers, e.g., phosphoric acid, and phosphites, may be present at 30-200 ppm and will hinder transesterification. The transesterification process for PET/PEN is very fast, if samples are heated above Tg.Three different mechanisms are believed to take place in the melt: alcoholysis, acidolysis, and direct ester exchange. The last one seems to be the most important according to the majority of authors. The process is divided into two steps: the first step is miscibility, which is diffusion controlled, and the second step is the transesterification, which follows Arrhenius' law, and requires good interfacial contact of the reacting species [67-701. The degree of transesterification is a function of temperature and mixing time; for some authors it is also a function of residual catalyst and inhibitor content as well as naphthalate concentration [11,27,71]. The effects of certain processing and compositional variables on the degree of transreaction are summarised in Table 3. Two classes of transesterification inhibitors are reported in the lit-

420

M. Xanthos, H. Warth

Table 3. Effect of some variables on deeree of transreaction Variable Original New Effect on degree of transreaction Mixing time Temperature Naphthalate content

1.5min

270°C 50 mol%

4.5min 310°C 80 mol%

10-30% 5-28% 18-23%

erature: phosphorus- and vanadium-containing compounds. Similarly, phosphites (dioctadecyl, triphenyl, diphenyl, dibenzyl, decyldiphenyl, bis(dibutylphenyl)pentaerythritol, distearylpentaerythritoldi, etc.) are commonly reported as inhibitors [ll].Some authors claim that only the hydrolysed and aged products of these compounds, the phosphonates, are active, while others claim that there is no influence at all [27,72].Inhibitors are applied at concentrations which correspond to 2 moles per mol of residual catalyst. The chemistry involved is very complex and yet not completely understood (see also Chapters 3 and 6). 3.1.6. Poly (ethylene terephthalate)/polyarylate In solution-mixed components, heating for over 16 hours at 280°C yields single-phase blends due to transesterification, as confirmed by FT-IR analysis of carbonyl streching bands of samples treated for different times [73-751. From thermal data, it was suggested that randomisation of the copolymers initially formed proceeds after 10 min, with completion after 10h at 280°C [13].In a detailed study of transesterification in blends of PAr with PET [76], prepared by extrusion at 265-270°C with a maximum of 3min residence time, and compression moulded at 26&350"C with short and long cycle times, only those blends moulded at higher temperature and longer cycle time showed evidence of transesterification. A 50:50 blend of PAr with PET extruded on an SSE showed significant transesterification only above an extruder set temperature of 320°C to yield a single amorphous phase.

3.1.7. Phenoxy blends Phenoxy-bisphenol A polycarbonate (PC) blends undergo significant interchange reactions at elevated temperatures to form essentially random copolymers of complex architecture. IR spectroscopyhas been used to study this process, because it is sensitive to the transformation of the original aromatic/aromatic carbonate groups of the PC to aromatic/aliphatic and aliphatic/aliphatic carbonates [77]. Transesterification is a function of the blend composition, time, and temperature. Mondragon et al. [78] confirmed the immiscibility of PAr/phenoxy blends by measurements of the softening temperature. The DSC technique was used to show the variation produced in the structure of the blends as a consequence of interchange reactions. The influence of the exchange pro-

Effects of Transreactions on Compatibility and Miscibility

42 1

cess on the mechanical behaviour was shown by the improvement observed in the elastic modulus of the reacted mixtures compared with that of the corresponding unreacted blends.

3.1.8. Liquid-crystalline polymer (LCP) blends

Porter et al. [8,79] discussed the relationship between miscibility and transesterification for a variety of polyester blends, including various polyesters with LCPs. Recent results on blends of LCPs with PC, PBT, and PET are summarised below.

With PC. Lin and Yee [80] studied the thermal behaviour of blends of a liquid-crystalline copoly(ester amide) (Vectra B950) with two isotropic polymers, one amorphous (PC) and one semicrystalline (PET). It was found that the glass transition temperature of PC decreases with increasing Vectra concentration in the blend, which suggested a partial miscibility. The miscibility is enhanced through heat treatment at elevated temperatures, presumably due to a transesterification reaction. Moreover, the presence of the amorphous polycarbonate hinders the crystallisation of the liquidcrystalline polymer in the blends. Wei e t al. [81,82]studied the transesterification mechanisms and rate in blends of PC and random liquid-crystalline polyester copoly(oxybenzoatepterephthalate). It was found that the ester-ester interchange in the two polymers took place within 15min when the blend was annealed at 260°C in vacuum. In the annealed blend, the bisphenol A unit in the polycarbonate reacted first with the terephthalate unit and then with the oxybenzoate unit in the LCP. As the transesterification in the blend continued for about 1h, the liquid-crystalline phase in the blend disappeared. The originally immiscible blend (two distinctive glass transition temperatures) became completely miscible after annealing (single glass transition temperature). With PBT. Kil et al. [83]prepared PBT/LCP blends in order to understand the effects of the transesterification on the properties, crystallisation behaviour , complex viscosity and molecular weight of the reactive blends. Because of hindered crystallisation, the T, of the blend decreased with increasing LCP content. The transesterification reaction had no effect on the crystallisation behaviour and complex viscosity. The reactive blends showed fast crystallisation due to annealing and increased molecular weight due to end-to-end reactions. Lee e t al. [84]showed that in situ compatibilisation of PBT with LCP could be achieved by a transesterification reaction catalysed through the addition of dibutyltin dilaurate. The high MW copolymers acted at the interface and reduced the size of the dispersed LCP phase, and enhanced interfacial adhesion, thus leading to an early increase of the flexural strength. The occurrence of transesterification was also confirmed by Jo et al. [85].

422

M. Xanthos, H. Warth

With PET. Hong et al. [86] studied the in situ compatibilisation of a PET/LCP blend via transesterification reactions in a twin screw extruder. The liquid-crystalline polymer, LC 3000, enhanced the crystallisation rate of PET, by acting as a nucleating agent. The LCP lowered the blend v i s cosity above T,, (crystal-nematic transition temperature), thus acting as a processing aid. The addition of dibutyltin dilaurate as catalyst caused an increase of the viscosity of the blends, reduced the size of the dispersed phase, enhanced adhesion to the matrix, and led to an enhancement of the mechanical properties of the two immiscible phases. The optimum catalyst content was about 500ppm, when the reaction proceeded in 90/10 PET/LCP blend systems. Hwang et al. [87] prepared copolyesters by reactive blending of PET and a thermotropic LCP, poly(phydroxybenzoic acid-co-ethylene terephthalate). Transesterification and chain scission reactions occurred depending upon reaction conditions, and resulted in reorganisation of chain sequence. The crystallisation of PET was delayed and the melting temperature was depressed with increasing LCP content due to exchange reactions [881* Heino et al. I891 showed that several catalysts, such as SbO3, stannous octoate or zinc acetate, enhanced the compatibility of the blends via transesterification reactions between the blend components, PBT (Arnite DO 4-300) with two thermotropic LCPs (Vectra A 950 and Rodrun LC 3000). 3.2. Polyamide/polyamide blends Transamidation of nylon 6 with nylon 6,6 without catalyst has been studied in a an extruder at 215-280°C at different blend ratios [go]. The products were characterised by DSC. Transamidation of nylon 6 with nylon 6,6 at 280°C in a single screw extruder in the presence of 1%phosphite catalyst gave products with a single T, and a lower heat of fusion than either of the homopolymers [91]. Moulded test parts showed improved tensile strength and elongation, and decreased modulus compared to either of the homopolymers due to decreased crystallinity resulting from copolymer formation. Takeda and Paul [92] studied blends of poly(m-xylene adipamide) and nylon 6 prepared by extrusion at 260°C (opaque, two glass transitions) and 29OOC (transparent, single glass transition). Phase homogenisation was the result of interchange reactions, as shown by thermal, dynamic mechanical, and NMR analyses. A single phase developed after as few as five interchange reactions per molecule. Extrusion-blended amorphous aromatic/semicrystalline aliphatic blends, such as nylon 6/nylon 61-co-T (an amorphous aromatic copolyamide based on iso/terephthalic acid and hexamethylenediamine), studied by Xanthos et al. [93],showed a single Tgthat increased almost linearly with increasing aromatic nylon content, as would have been expected from

Effects of Transreactions on Compatibility and Miscibility

423

property additivity rules for miscible systems. Data obtained from DSC measurements on pellets and tan b (loss factor) us. temperature plots on injection moulded specimens indicated good agreement between the two methods. Similarly, melt flow index and mechanical property data essentially obeyed additivity rules. However, T, of the crystalline nylon 6 component decreased with decreasing concentration of nylon 6, a possible effect of interchange reactions.

3.3. Polyamide/polyester blends

3.3.1. Poly(ethy1ene terephthalate)/polyarnide

Pillon et al. [94-961 have formed nylon 6,6-PET block copolymer in situ through an ester-amide interchange reaction in the melt. The process was run in either a Brabender mixing chamber at 265-295°C and 91rpm or in an SSE at 30CL350"C and 10-2lrpm with 2 4 m i n residence time. p Toluenesulfonic acid (TSA; 0.2 wt %) was added to catalyse the reaction. The extent of copolymer formation in the extruded product could be calculated using 400MHz 'H NMR spectroscopy since the chemical shift of the terephthalate aromatic ring protons shifted as ester linkages were replaced by amide linkages. Maximum copolymer level (23%) was obtained by extrusion of equimolar blends with TSA at 330-350°C with 2 min residence time. Blends extruded either without TSA or at less than 310"C, or precipitated from trifluoroacetic acid solution, showed no evidence of the interchange reaction. FT-IR analysis of the melt-prepared blends indicated no shift of the carbonyl group of PET or PA (an indication of absence of hydrogen bonding); by contrast, such a shift was observed in the solution cast blends. These results confirm that the formation (in solution) or lack (in the melt) of hydrogen bonding is independent of the presence of the interchange reaction. Further studies on uncatalysed and melt-catalysed blends indicate an increase in the degree of crystallinity of PET, modification of the catalysed blend morphology as a result of the interchange reaction and embrittlement due to the increase in crystallinity [96,97]. In a series of publications describing morphologies of the so-called "microfibrillar reinforced composites'' consisting initially of PET/nylon 6 homopolymer blends, evidence of transreactions between the blend components was drawn from the application of techniques such as DSC, WAXS, DMTA, and microscopy. In addition to reactions at the interface between PET fibrils and a polyamide matrix produced by post-extrusion processing, solid state transreactions resulted in modification of the matrix to an initially crystallisable block copolymer followed by the formation of a non-crystallisable random copolymer [98-1011.

424

M. Xanthos, H. Warth

3.3.2. Polycarbonate/polyamzide Eguiazabal and Naaabal [lo21 studied the interchange melt reaction between nylon 6 and bisphenol A polycarbonate (PC) in a Brabender mixer at either 240” or 250°C and 30rpm. The interchange reactions were complex including acidolysis, aminolysis, and amidolysis of carbonate bonds. The products were analysed by selective solvent extraction and DSC. The Tgvalues of the two polymers approached each other as mixing time increased to 45min but the relatively small changes indicated a low degree of interchange reaction. In a similar study [103,104],additional criteria for copolymer formation were IR spectroscopy and SEM morphological analysis; the latter showed the presence of discrete PC domains at a high PC content in the blend after 15min mixing and homogeneity after 45 min at low PC percentages as a consequence of increasing copolymer formation. Chemical reactions between the two polymers give rise to low MW compounds which cause decrease in the Tgof PC and reduction of the overall crystallisation rate of nylon 6. The exchange reactions between amino-terminated nylon 6 and the inner carbonate groups occurring by melt-mixing at 240°C were studied by Montaudo et al. [105]. The exchange reaction yields sizable amounts of copolymer after relatively short times of melt-mixing. NMR analysis of the copolymer yielded besides the copolymer composition, evidence of the presence of urethane units interconnecting the nylon 6 and PC blocks. The amount of urethane units increased with reaction time, indicating a reduction of the block size as a function of the extent of exchange. The interchange reactions leading to the progressive formation of copolymers which homogenise a mixture of nylon 6/PC and affect its thermal transitions were studied by calorimetric techniques [lo61 (see also Chapter 4).

References 1. M. Xanthos, Polym. Eng. Sci. 28, 1392 (1988) 2. M. Xanthos, S. S. Dagli, Polym. Eng. Sci. 31,929 (1991) 3. B. Brown, in: Reactive Extrusion: Principles and Practice, edited by M. Xanthos, Carl Hanser Verlag, Munich, New York 1992, ch. 4 4. N. G. Gaylord, J. Macromol. Sci-Chem. A26,1211 (1989) 5. N. C. Liu, W.E. Baker, Adv. Polym. Technol. 11, 249 (1992) 6 . L. A. Utracki, “Polymer Alloys and Blends”, Carl Hanser Verlag, Munich, New York 1989 7. M. Xanthos, S. H. Patel, in: How to Manage Plastics Waste: Technology and Business Opportunities, edited by A. L. Bisio, M. Xanthos, Carl Hanser Verlag, Munich, New York 1994, ch. 13 8. R. S. Porter, L. H. Wang, Polymer 33, 2019 (1992) 9. A. M. Kotliar, J. Polym. Sci., Macromol. Rev. 16, 367 (1981) 10. R. S. Porter, A. Casale, Polym. Eng. Sci. 25, 129 (1985) 11. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22, 229 (1982)

Effects of Transreactions on Compatibility and Miscibility

425

12. A. Golovoy, M. F. Cheung, H. Van Oene, Polym. Eng. Sci. 27, 1642 (1987) 13. M. Kimura, G. Sdee, R. S. Porter, J. Appl. Polym. Sci. 29,1629 (1984) 14. M. F. Cheung, A. Golovoy, R. 0. Carter, H. Van Oene, fnd. Eng. Chem. Res. 28,476 (1989) 15. A. Golovoy, M. F. Cheung, K. R. Carduner, M. J. Rokosz, Polym. Bull. 21, 327 (1989) 16. J. A. Cox, M. E. Stewart, F. A. Shepperd, R. R. Light, WO 92/02584 (1992) 17. L. A. Utracki, Polym. Eng. Sci. 35,2 (1995) 18. B. D. Hanrahan, S. R. Angeli, J. Runt, Polym. Bull. 15,455 (1986) 19. W. G. Zheng, Z. N. Qi, F. S. Wang, Polym. Int. 34,307 (1994) 20. Z. Denchev, M. Sarkissova, S. Fakirov, F. Yilmaz, Macromol. Chem. Phys. 197,2869 (1996) 21. Z. Denchev, M. Sarkissova, H.-J. Radusch, T. Luepke, S. Fakirov, Macromol. Chem. Phys. 199,215 (1998) 22. S. Fakirov, M. Sarkissova, Z. Denchev, Macromol. Chem. Phys. 197,2837 (1996) 23. H. Yoon, Y.Feng, Y.Qiu, C. C. Han, J. Polyrn. Sci., Polym. Phys. Ed. 32, 1485 (1994) 24. P. Godard, J. Dekoninck, V. Pevlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24,3301 (1986) 25. M. Fiorini, C. Berti, V. Ignatov, M. Toselli, F. Pilati, J. Appl. Polym. Sci. 55,1157 (1995) 26. H. Van Oene, M. F. Cheung, A. Golovoy (Ford Motor Co.), EP 321126 A1 890621 27. A. Golovoy, M. F. Cheung, K. R. Carduner, M. J. Rokosz, Polym. Eng. Sci. 29,1226 (1989) 28. G. Pompe, L. J. Haussler, J. Polym. Sci., Polym. Phys. Ed. 35,2161 (1997) 29. S. B. Tattum, D. Cole, A. N. Wilkinson, Plast. Rubber Compos. Process. Appl. 25,477 (1996) 30. 2. Huang, Y.Zhang, Y. Liu, Gongneng Gaofenti Xuebao 9,489 (1996) 31. N. Wings, G. Trafara, Angew. Makromol. Chem. 217,91 (1994) 32. X. Kong, A. Ajji, Adv. Polym. Blends Alloys Technol. 3,73 (1992) 33. A. W. Birley, X. Y . Chen, Br. Polym. J. 16,77 (1984) 34. A. N. Wilkinson, S. B. Tattum, A. J. Ryan, Polymer38, 1923 (1997) 35. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1895 (1982) 36. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1881 (1982) 37. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1901 (1982) 38. K.J. Idel, J. Kirsch, F. Paul, L. Meier, H. G. Gehrke (Bayer A.-G., FRG), DE 3842219 A1 900621. 39. R. R. Gallucci (General Electric Co., USA), US 5354791 A 941011 40. J. Van Helmond (General Electric Co., USA), US 5252653 A 931012; E P 543125 A1 930526. 41. J. R. Decarli, Polim.: Cienc. Tecnol. 3, 16 (1993) 42. M. Pellow-Jarman, M. Hetem, Plast. Rubber Compos. Process. Appl. 23,31 (1995)

426

M. Xanthos, H. Warth

43. W. M. M. Roovers, J. J. Verhoeven (General Electric Co., USA), US 4965308 A 901023 44. J. A. Tyrell (General Electric Co., USA), US 4560722 A 851224 45. R. J. McCready (General Electric Co., USA), US 4452933 A 840605 46. D. J. Brunelle (General Electric Co., USA), US 4452932 A 840605 47. D. G. Hamilton (General Electric Co., USA), EP 601784 A2 940615 48. M. Kimura, R. S. Porter, J . Polym. Sci., Polym. Phys. Ed. 21, 367 (1983) 49. M. Kimura, R. S. Porter, Anal. Calorim. 5 , 25 (1984) 50. J. Devaux, P. Devaux, P. Godard, Macromol. Chem. 186, 1227 (1985) 51. I. Mondragon, J. Nazabal, Polym. Eng. Sci. 25, 178 (1985) 52. I. Mondragon, J. Nazabal, J. Appl. Polym. Sci. 32, 6191 (1986) 53. D. L. Freitag, L. Bottenbruch, M. Schmidt, US 4533702 (1985) 54. Y. Yu, K. J. Choi, Proc. 53rd SPE ANTEC, 1711 (1995) 55. Y. Yu, K. J. Choi, Polym. Eng. Sci. 37,91 (1997) 56. A. Escala, R. S. Stein, Multiphase Polymers, Adv. Chem. Ser. 176,455 (1979) 57. S. C. E. Backson, A. M. Kenwright, R. W. Richards, Polymer36,1991 (1995) 58. J. H. Kim, W. S. Ha, Han’guk Somyu Konghakhoechi 31, 788 (1994) 59. B. Jacques, J. Devaux, R. Legras, E. Nield, J. Polym. Sci., Polym. Chem. Ed. 34, 1189 (1996) 60. G. Sugerman, S. J. Monte, Interfaces Polym., Cemm., Met. Matrix Compos., Proc. Int. Conf. Compos. Interfaces, edited by H. Ishida, Elsevier, New York 1988, p.657 61. B. Jacques, J. Devaux, R. Legras, E. Nield, Polymer 38 5367 (1997) 62. Y. Liu, Y. Tong, L. Wu, Gongneng Gaofenzi Xuebao 8 , 8 (1995) 63. R. Van Der Meer, C. L. Verbraak, J. W. Heuseveldt, J. Van Helmond, H. H. Ketels (General Electric Co., USA), Eur. Pat. Appl. 609538 A1 940810 64. Y. Bonin, M. Logeat (Rhone-Poulenc Specialites Chmiques, Fr.), F’r. Demande 2567137 A1 860110 65. A. A. Deguia (Celanese Corp., USA), Eur. Pat. Appl. 63031 A1 821020 66. C. L. Heisey, D. C. Hoffman, J. A. Zawada, Book of Abstracts, 211th ACS National Meeting, New Orleans, LA 1996, POLY-232 67. M. Guo, Book of Abstmcts, 211th ACS National Meeting, New Orleans, LA 1996, POLY-230 68. M. Guo, J. Am. Chem. Soc., Polym. Prepr. 37, 227 (1996) 69. M. Guo, W. J. Brittain, Book of Abstmcts, 215th ACS National Meeting, Dallas 1998, POLY-107 70. M. Guo, Book of Abstmcts, 213th ACS National Meeting, San Francisco 1997, POLY-548 71. R. Po, E. Occhiello, G. Giannotta, L. Pelosini, L. Abis, Polym. Adw. Technol. 7, 365 (1996) 72. M. E. Stewart, A. J. Cox, D. M. Naylor, Proc. 51st SPE ANTEC, 1222 (1993) 73. R. Porter, Themochim. Acta 134, 251 (1988) 74. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24, 3315 (1986) 75. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24, 3310 (1986) 76. L. M. Robeson, J. Appl. Polym. Sci. 30, 4081 (1985) 77. X. Yang, P. C. Painter, M. M. Coleman, Macromolecules 25, 4996 (1992)

Effects of Transreactions on Compatibility and Miscibility 78. 79. 80. 81. 82. 83.

427

I. Mondragon, P. M. Remiro, J. Nazabal, Eur. Polym. J . 23,125 (1987) R. S. Porter, J. Am. Chem. SOC.,Polym. Prepr. 31,296 (1990) Q. Lin, A. F. Yee, J. Mater. Sci. 32,3961 (1997) K. H. Wei, J. C. Ho, Macromolecules 30,1587 (1997) K. H. Wei, K. F. Su, J. Appl. Polym. Sci. 56, 79 (1995) S. B. Kil, 0. 0. Park, K. H. Yoon, C . I. Hwang, Korea Polym. J. 5, 145

(1997) 84. J. Y. Lee, J. Jang, S. M. Hong, S. S. Hwang, Y. Seo, K. U. Kim, Int. Polym. Process. 12,19 (1997) 85. B. W. Jo, J. H. Chang J. I. Jin, Polym. Eng. Sci. 35,1615 (1995) 86. S. M. Hong, S. S. Hwang, Y. Seo, I. J. Chung, K. U. Kim, Polym. Eng. Sci. 37,646 (1997) 87. C. I. Hwang, S. B. Kil, 0. 0. Park, Polym. Other Adv. Mater., Proc. Int. Conf. Front. Polym. Adv. Mater., edited by P. N. Prasad, J. E. Mark, J. F. Tung, Plenum, New York 1995, p. 93 88. C. Chang, P. Guo, M. M. Denn, Polym. Adv. Technol. 7,168 (1996) 89. M. T. Heino, J. V. Seppala, Acta Polytech. Scand., Chern. Technol. Metall. Ser. 25,214 (1993) 90. N. M. R. Schott, B. Sanderford, Coat. Plast. Prepr., ACS Div. Org. Coat. P l a t Chem. 37,73 (1977) 91. Y .P. Khanna, E. A. Turi, S. M. Aharoni, T. Largman, US 4417032 (1983) 92. Y. Takeda, D. R. Paul, Polymer 32,2771 (1991) 93. M. Xanthos, J. F. Parmer, M. L. La Forest, G. R. Smith, J. Appl. Polym. Sci. 62,1167 (1996) 94. L. Z. Pillon, L. A. Utracki, Polym. Eng. Sci. 24, 130 (1984) 95. L. Z. Pillon, L. A. Utracki, D. W. Pillon, Polym. Eng. Sci. 27,562 (1987) 96. L. Z. Pillon, J. Lara, D. W. Pillon, Polym. Eng. Sci. 27,984 (1987) 97. M. R. Kamal, M. A. Sahto, L. A. Utracki, Polym. Eng. Sci. 21,985 (1981) 98. M. Evstatiev, N. Nicolov, S. Fakirov, Polymer 37,4455 (1996) 99. S. Fakirov, M. Evstatiev, S. Petrovich, Macromolecules 26,5219 (1993) 100. M. Evstatiev, S. Fakirov, Polymer 33,877 (1992) 101. S. Fakirov, M. Evstatiev, Adv. Mater. 6,395 (1994) 102. J. Eguiazabal, J. Nazabal, Makromol. Chem., Macromol. Symp. 20/21,255 (1988) 103. E. Gattiglia, F.P. La Mantia, A. Turturro, A. Vdenza, Polym. Bull. 21,47 (1989) 104. E. Gattiglia, A. Turturro, E. Pedemonte, J. App. Polym. Sci. 38,1807 (1989) 105. G . Montaudo, C. Puglisi, F. Samperi, J. Polym. Sci., Polym. Chem. Ed. 32, 1, 15 (1994) 106. M. Cortazar, J. I. Eguiazabal, J. J. Iruin, Br. Polym. J. 21,395 (1989)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Chapter 11

Effect of Transreactions and Additional Condensation on Structure Formation and Properties of Condensation Polymers

F. J. Balt6 Calleja, S. Fakirov, H. G . Zachmann*

1. Relationship between interchain reactions and structure of

condensation polymers

In his review on compatibility and transesterification in binary polymer blends, Porter [l]stresses that a fascinating feature of the blends of the polyester family is their easy transesterification, which has the following consequences: (i) it can open a new route to enhanced compatibility and preparation of novel copolymers with specific degrees of randomness and compositions; (ii) it can lead to more uniform polymers by minimising molecular weight fluctuations in the melt stream during polymerisation and processing; and (iii) it can provide for chemical healing of laminates of condensation polymers [l-261. 1.1. Effect of interchain reactions on structure formation and properties of condensation polymers

High-performance polymer materials are distinguished by a structure usually involving (i) maximal orientation of macromolecules; (ii) perfect crys'Prof. H. G. Zachmann passed away on 28.4.1996

430

F. J. Balti Calleja, S. Fakirov, H.G.Zachrnann

tallites built up of extended chains; and (iii) a relatively high molecular weight of the polymer. The lower ability of condensation polymers, as compared to polyolefins, to form such a structure is related to their peculiarities - larger cross-section of the macromolecules, more irregular chain structure, lower molecular weights. The desired supermolecular structure is usually achieved by thermal and mechanical (orientation) treatments. During thermal treatment at elevated temperature, together with structural changes (crystallisation and relaxation) condensation polymers could undergo chemical interactions, e.g., additional condensation, exchange reactions, ek., due to the presence of reactive groups in these polymers (-CONH-, -COOH, -NH2, -OH, -COO-); this is in contrast to polyolefins, where only degradation could take place [27]. Additional condensation and trans- (exchange) reactions affect not only the processing conditions and mechanical behaviour of linear polycondensates (this will be discussed further in this chapter), but they can be used to eliminate defects in the structural entities, originating from chain ends, entanglements and chain folds, as shown schematically in Figure 1. It should be added here that, in addition to the molecular weight increase, additional solid-state condensation results in the rise of the number of tie molecules. In contrast to the chemical aspects of solid-state reactions, their role in the creation of the physical structure in condensation polymers is rather underestiwated, as emphasised by Wunderlich [28]. The ideal structure resulting in optimal mechanical parameters should be characterised by chains stretched as much as possible, of the highest molecular weight, perfectly aligned to each other. Solid-state reactions are expected to contribute to the fulfillment of these requirements by the achievement of maximum molecular weight values and creation of a perfect structure by elimination of defects (Figure 1). Starting from the experience available [29-311, it seemed that such a structure can be realised by two-stage cold drawing, each stage followed by high temperature annealing [32]. The structural and chemical changes in oriented polycondensates taking place during these thermal and mechanical treatments are outlined in Figure 2. After the first cold drawing (up to X = 5 ) , the polymers are subjected to annealing in vacuum at a temperature T, close to but below the melting one (T,= 260°C in the case of poly(ethy1ene terephthalate) (PET), Figure 2a). In addition to crystallisation and relaxation, chemical reactions take place during this treatment, resulting in the elimination of defects (Figure 1) and an increase in molecular weight, and hence in improved drawability [27,31,33].These changes occurring during annealing are shown schematically in Figure 2a,b. When an oriented and annealed, partially crystalline, high molecular weight material (Figure 2b) is subjected to a second drawing, conformational changes related to the complete extension (at moderate stresses) of the chains in the amorphous regions are effected along the direction of the external strain. Further drawing could result in the destruction of crys-

Effect of Transreactions on Structure and Properties

431

tallites by chain defolding and alignment in the draw direction. Such a conformational transition has already been proved for PET [34],nylon 6 (PA 6) [35]and poly(buty1ene terephthalate) (PBT) [36].The structure created as a result of such mechanical treatment is shown schematically in Figure 2c. The final morphological structure of highly oriented polycondensates illustrated in Figure 2d is similar to that of polymers obtained from liquidcrystalline mesophases (parallel alignment of the chains dong the fibril axis). It is quite natural to expect that polymers with such a supermolecular structure should be characterised by very good mechanical properties due to the large number of chains bearing the external mechanical strain and to the presence of strong molecular interactions. In view of the practical application of polycondensates with improved mechanical properties, evidence has been provided about the extent to which the above processes actually occurs using wide- and small-angle Xray scattering (WAXS, SAXS) as well as mechanical tests. The results of the mechanical tests of samples with different draw ratios (A = 5 or 20) are presented in Table 1. They are a further experimental

-COOH

YN-

--0 coox -c lo-- R

AC

--*

-

-co

NH-TR,

-NY

N H - S (COAC

-CO~NH-

-COOH

-NY CO-NHIH/OOC-

AC

Figure 1. Schematic representation of elimination of defects through solid-state reactions in linear polycondensates: AC additional condensation; TR transreactions

432

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

proof of the difference in the morphological structures of both materials.

It is seen that the samples with X = 5 annealed at 26OOC for 6 h show an

extremely high deformation ability (E = 320%, Table 1, Sample 3). The latter is an indication that, under these conditions, the chemical reactions illustrated in Figure 2b have taken place with a subsequent rise in the molecular weight. Taking into account that the fivefold drawn samples are subjected to a second drawing at room temperature (after annealing at 26OoC),the total draw ratio amounts to X = 20. It is seen in Table 1 that the tensile strength and elasticity modulus of the samples with X = 20 are almost twice as high as those of the samples with X = 5 (a = 0.51 GPa and E = 13.8GPa against r~ = 0.22GPa and E = 9.3GPa, Table 1, Samples 1).After annealing at 2OO0C, the samples with X = 20 have E = 18.6 GPa and u = 0.6 GPa. These parameters become slightly lower when annealing

l-

T (260'C)

I

I

i

I

b

Figure 2. Models illustrating the physical and chemical changes in oriented partially crystalline PET subjected to additional drawing and annealing with fixed ends: (a) structure of the sample after cold drawing (A = 4.5); (b) structure of the same sample after annealing (Ta= 260OC); (c) structure of the same sample after a second cold drawing (A = 15-20); (d) structure of the same sample after a second annealing (Ta= 250-270°C). AC additional condensation; TR trans (exchange) reactions; the dashed line delimits two separate microfibrils [32]

433

Effect of Transreactions on Structure and Properties

is carried out at 260°C but their values are three times ( E ) and four times (a)higher than those of the samples with X = 5 annealed at the same temperature (Table 1, Samples 3). These substantial differences in the mechanical properties of both materials are mainly due to the relatively larger number of chains under stress bearing the external strain, as well as to the strong molecular interactions in the highly oriented material. This statement is also supported by the fact that the samples with X = 20 show a rather low and constant elongation at break (16-18%) regardless of the annealing temperature, in contrast to the case of X = 5 (38-320%). In this way solid-state interchain reactions contribute to the improvement of the mechanical properties of polyesters and polyamides mainly through the elimination of defects and increase in molecular weight.

1.2. Structure formation in blends of condensation polymers with interchain reactions occurring to various extents Blends of crystallisable polymers are interesting subjects for investigation. Among others, the following questions arise: Are the polymers in the blend miscible or partially miscible? How is the crystallisation behaviour of each component influenced by the presence of the other component? Do any chemical reactions such as transesterification take place between the two components and how do these reactions influence miscibility or depend on miscibility? Extensive studies of the miscibility and crystallisation behaviour have been performed on blends of, among others, poly(scapro1actone) and polystyrene [37,38], of poly(ethy1ene oxide) and poly(methy1 methacrylate) [39], of poly(viny1idene fluoride) and PA6 [40], and of isotactic polypropylene with different rubbers [41]. It was shown that the second component influences the crystallisation kinetics and morphological structures in many different ways depending on composition and miscibility. There also exist some studies of miscibility and transesterification, mainly in blends of polycarbonate (PC) and PBT [9,42], as well as of PC and PET [43-461. Clear evidence was found that transesterification plays an important role in blending such systems (for more details on this topic Table 1. Mechanical tests of drawn PET bristles annealed under constant strain A = 5, annealed for 6 h

Sample no. 1 2 3

T, ("C) Unannealed 200 260

U

(GPa) 0.22 0.26 0.14

X = 20, annealed for 2 h

E

E

9.3 11.5 6.5

61 38 320

(GPa)

(%)

c7

E

&

(GPa)

(GPa)

(%)

0.51 0.60 0.57

13.8 18.6 16.7

17 16 18

434

F. J. BaltB Calleja, S. Fakirov, H. G . Zachmann

see also Chapter 10). Neutron scattering experiments [47,48] have revealed that within a single-component system, such as PET, above the melting point a large amount of transesterification takes place within a few minutes. While there exist many studies on miscibility and crystallisation, the question of how transesterification is related to them is still not completely understood. This problem was investigated using blends of PET and poly(ethylene-2,6-naphthalene dicarboxylate) (PEN) (491. Both materials are able to crystallise. The melting point Tm and glass transition temperature Tgof PEN are 270 and 120°C, respectively, and those of PET are 268 and 70°C, respectively. Initial studies on blends of PET and PEN indicated that these polymers form a single-phase system when annealed for 2min in the melt but it was unclear to what extent transesterification has taken place. For this reason, PET and PEN were mixed by coprecipitation from solution in order to avoid any transesterification and then melt-pressed for different times in order to study the extent of transesterification by means of different methods [49]. Figure 3 represents the temperature dependence of the loss modulus, G”,of blends containing 30 wt % PET obtained in the amorphous state by melt pressing at 280’C for different times t m and quenching in ice72

PET/PEN 30/70

0.2

0.5

2

10 45

Figure 3. Temperature dependence of the loss modulus G” during heating of PET/PEN blends containing 30 w t % PET after different times tm of meltpressing at 280°C [49]

435

Effect of Transreactions on Structure and Properties

water. Two well-separated maxima of G" are observed after melting for 0.2 and 0.5 min. The first maximum appears at 72 and 78"C, respectively. The temperature of 72°C corresponds to the glass transition maximum of neat P E T [50]. However, the maximum of the blend is broader than that of neat PET. The second maximum appears at 103 and 105"C, respectively, being lower by about 15°C than the glass transition maximum of neat PEN [50]. This indicates that, after short melting times such as 0.2 and 0.5min, two different phases coexist, one rich in PET and the other one rich in PEN. With increasing melting times, the two maxima approach each other, resulting in a broad maximum at about 85"C, after a melting time of 2 min, and a sharp maximum at 100°C after melting for 45min. This indicates that a single phase is formed during these longer periods of melting. The maximum at 65°C visible in the two last spectra is the p* maximum of PEN. Similar results are obtained on blends containing 44 and 60 wt % PET. The two peaks observed after melting of less than 2min merge into a single one after longer melting times. As expected, the peak of PEN becomes weaker with increasing PET content. In the case of 70 wt % PET the PEN peak disappears completely. A broad peak at 123°C is caused by crystalli-

PET/PEN 30/70

I

t,, min 0

0.2

0.5

2

10 45 1

50

1

1

100

1

1

150

1

1

200

1

1

250

1

1

300

T,OC

Figure 4. DSC curves of PET/PEN blends containing 30 wt % PET after different times t, of melt pressing at 280°C and quenching in ice-water [49]

436

F. J. Baltd Calleja, S. Fakirov, H.G . Zachmann

sation of PET [49]. Figure 4 shows the DSC curves of the blends containing 30wt % PET after different times of melt-pressing (as in Figure 3). The DSC curve of the as-precipitated powder (t, = 0) is also represented. The samples molten for 0.2 and 0.5min show two exothermic peaks, a small one in the region of 140-150°C, which can be attributed to the crystallisation of PET, and a larger one at approximately 18OoC,attributed to the crystallisation of PEN. After melting for 2 min, the PET peak is no longer apparent. If the melting time is further increased, the other peak also gradually disappears. The as-precipitated powder ( t , = 0) has already crystallised to some extent at the beginning of heating in the DSC. Therefore, the crystallisation peak is relatively small. A melting peak at 266"C, corresponding to the melting of PEN, with a small shoulder at 251"C, can be seen in the cases oft, = 0,0.2 and 0.5 min; it can be attributed to the melting of PEN (266°C) and PET (251°C). With increasing t m of previous melting, the melting peak gradually shifts to lower temperatures, becomes less intense, and finally disappears. These observations suggest a transesterification process between PET and PEN leading to a random copolyester. It is known from previous studies [51] that random copolyesters of PET and PEN of the compositions investigated here crystallise either to a small extent or not at all. Similar results are obtained with other compositions of the same blend. With increasing amount of PET, the crystallisation peak at the lower temperature increases while that at the higher temperature decreases, justifying the attribution of these peaks to the crystallisation of PET and PEN, respectively. No crystallisation and melting are observed fort, = 45 min in the PET/PEN 30/70 and 44/56 samples and for t , = 10 min in the 60/40 and 70/30 samples [49]. Further information on the crystallisation behaviour is obtained by WAXS measurements. Unfortunately, the crystal reflections of PET and PEN lie quite close together. Nevertheless, interesting results were obtained by means of studies of the crystallisation kinetics. Figure 5 shows the change of the WAXS during isothermal crystallisation at 161.5"C of a blend containing 60 wt % PET melt-pressed for 0.5min. One can clearly recognise that the PET reflection grows within the first 3 min while the PEN reflection starts to grow after 8 min. Thus, PET crystallises faster than PEN. The corresponding results obtained from a blend containing 44 wt % PET and crystallised at 204.5"C show that PEN crystallises faster than PET [49]. In this way the rate of crystallisation of blends with different compositions at various temperatures was studied and the half-times of crystallisation of PEN were determined from the increase of the intensity of the reflection at 15.5" with time; for PET, the reflection at 17.3" was used. The data obtained led to some conclusions concerning the rate of isothermal crystallisation. It was found that in the high-temperature range of crystallisation (T, > 150°C for PET and T, > 190°C for PEN) the half-

Effect of Transreactions on Structure and Properties

437

PEN PET t, 45 40 35

30 25 20

15 10 5

Figure 5. Changes of WAXS during isothermal crystallisation of a PET/PEN blend containing 60wt% PET at 161.5"C after rapid heating from the glassy state [49]

time of crystallisation of the blend components is shorter than that of the neat materials. How can this be explained? In the case of phase separation, the half-time of PEN will either remain unaffected by the presence of PET or it will decrease due to a nucleation effect. Therefore, it can be concluded that miscibility does exist and two possible effects may be considered: (i) a melting point depression - however, no melting point depression is observed in the DSC diagrams after melt-pressing for 0.5min; (ii) dilution of PEN by PET, due to which the rate of formation of crystal nuclei is decreased, and this could be the explanation of the observed effect. In the low-temperature range, PEN crystallises in the blend at lower

438

F.J. Balti Calleja, S.Fakirov, H. G . Zachmann

temperatures and PET at higher temperatures than in the pure state. This can again be explained by assuming partial miscibility. As the glass transition temperature of PET (70°C)is lower than that of PEN (120"C), the presence of PET increases the mobility of the PEN molecules while that of PEN decreases the mobility of the PET molecules. Further, the following experimental results clearly demonstrate that transesterification between PET and PEN takes place at 280°C in the blend of these components obtained by coprecipitation from solution: (i) the decrease in the melting point with increasing melting time t , (Figure 4); (ii) the disappearance of the crystallisation and melting peak in the DSC curves during heating with 10"C/min after melt-pressing for 45 min; (iii) after 45min of melt-pressing - even if annealed for a long time at high temperature -no crystallisation occurs when the PET content is 60wt %, and only one component, namely the one which predominates, crystallises at the other compositions. The WAXS, DSC, and DMTA curves of the material obtained by melt-pressing for 45min are the same as those of a copolyester obtained by the usual synthesis [51]. Thus, it can be concluded that transesterification is completed after 45min of melt-pressing. On the other hand, in a mixture of the two components composed of grains with a diameter of 0.5 mm or larger, the rate of transesterification is drastically reduced. Thus, transesterification alone cannot result in a singlephase system when the system is initially phaseseparated on a scale of 0.5 mm. Another point to consider is whether transesterification within short times is possible if the separate phases are smaller than 0.5mm. It is shown [52] in another two-component system consisting of the isotropic ternary copolyester of PEN, PET, and poly(p-hydroxybenxoic acid) (PHB) (35:35:30) and the liquid-crystalline binary copolyester PET-cePHB (30:70) that no transesterification occurs within 45 min, even if blending has been performed by coprecipitation from solution. This may indicate that single phase formation is required for transesterification to take place. Finally, the appearance of a single peak in G" after 2min of melting indicates that a single phase is formed. It is clear that some miscibility already exists in the films melt-pressed for 0.2 and 0.5 min. This is supported by the following results. (i) The rate of crystallisation of PET and PEN in the blend differs from that of the pure phases. (ii) The glass transition peak of PEN in the blend appears at a temperature lower by 15°C than that of neat PEN. (iii) In the blend containing only 30% PEN, the glass transition peak of PEN in the G" curve has completely disappeared. This indicates that PEN is dissolved in PET though the Tgof PET remains unchanged. Thus, a conclusion can be drawn that coprecipitation is accompanied by phase separation, the two phases being highly dispersed. During subsequent melt-pressing a single phase is formed within 2 min by diffusion. It should be stressed that, as is well known, a single glass transition does not necessarily mean that miscibility exists on a molecular scale. A single maximum also appears when the components are separated, forming

Effect of Transreactions on Structure and Properties

439

phases that are smaller than ca. lOA [53]. The question to what extent transesterification has occurred within the first 2 min of melt-pressing is difficult to answer. By dissolution of the sample after melt-pressing and a second coprecipitation, small changes in the DSC diagram are observed, indicating that some transesterification has taken place [49]. The question arises of whether miscibility is enforced by the transesterification process against thermodynamics. In other words, can a single phase be formed within such a short time just as a consequence of transesterification, even if there is no thermodynamic tendency to form a single phase? The results obtained with the liquid-crystalline PET-co-PHB (30:70) and the isotropic PEN-co-PET-co-PHB (35:35:30) seem to indicate that this is impossible. Transesterification and therewith a single phase are not achieved if thermodynamics do not favour at least partial miscibility. If, however, the homopolymeric components of the blend are partially miscible, the formation of copolymers by transesterification may increase the miscibility, thus resulting in a copolyester showing complete miscibility of the components [49] (see also Chapter 10). The microhardness technique allows one to follow not only the changes in the mechanical properties of blends of condensation polymers but also the extent of chemical interactions. Microindentation involving a mechanical deformation on a small scale is one of the simplest methods of determining the microhardness of a material [54,55]. Microhardness, H , is also a technique which offers direct information on microstructural changes in polymers [56-59] and blends [60], being, in the latter case a helpful tool for the assessment of the degree of interpenetration of the blend components. The crystallisation behaviour and the physical ageing of PET and PEN were investigated extensively by means of microhardness studies [6144]. A linear relationship was found empirically between the degree of crystallisation and the microhardness, H , of polymers. Since the time dependence of the relative crystallinity of a polymer can often be described by an Avrami equation [6547], a similar approach was proposed to characterise the microhardness evolution with time of a polymer film [56,62],

where HYa” and Ha,i are the microhardness of the fully crystallised and fully amorphous polymer, respectively, t , is the crystallisation time, and G and n are the modified Avrami parameters for microhardness. As in the Avrami equation, the crystal growth, G is proportional to the number of nuclei per unit volume and the rate of nucleus formation [65]; it is strongly temperature-dependent. The Avrami exponent, n, is characteristic of the nucleation type and the crystal growth geometry, and may show values ranging from below 1 to far above 6. A correspondence list between crystallisation type and Avrami exponent can be found in [66].

440

F. J. Baltd Calleja, S. Fakirov, H.G.Zachmann

This correspondence, however, is not uniquely fixed to any one set of conditions. Additional information on nucleation, morphology, and mechanism is needed to fully interpret the exponent n. Furthermore, the values of the Avrami parameters obtained from microhardness data may differ slightly from those obtained from crystal growth analysis, if the relationship between the two is not perfectly linear. In a recent study of the microhardness measured at room temperature of amorphous PET/PEN blends as a function of both composition and blending time, t,, it was observed that, for a given processing time, the microhardness measured at room temperature first increased with increasing blending time, then reached a maximum value for processing times ranging between 2 and lOmin, and finally decreased for longer melting times [68]. Blends containing different concentrations, x , , , ~and ~ x O v P Eof N the starting materials were obtained by coprecipitation from solution in hexafluoroisopropanol. Thin amorphous films of PET/PEN copolymer were then obtained from the precipitated powder by melt pressing in vacuum at a temperature T, = 280OC during a time t, = 2min followed by quenching in ice-water. Microhardness of the blends was determined at various crystallisation temperatures, T,,as a function of the crystallisation time, t,. The hardness is given by H = K F / d 2 ,where F is the force applied by the indenter, d is the diagonal length of the indentation, and K = 1.854, in SI units, is a geometrical factor. The microhardness of the heated copolymer films was measures as a function of the time t,, until it reached a maximum value, Ifma",at which it stabilised. The rise in microhardness of the copolymers in Figure 6 is due to the crystallisation of the pure PET and PEN, respectively, remaining in the blend, and follows an Avrami-type Eq. (1). For temperatures lower than ', of PEN, 125"C, ie., lower or equal to the glass transition temperature, T only PET crystallises, while for temperatures higher than 145°C both PET and PEN crystallise. The hardening rate increases with temperature, as expected in crystallisation kinetics related phenomena. The xO,PET = 0.9 blend maximum hardness plateau, Ifmax= 118MPa, is quasi-independent of T, between 95 and 125"C, suggesting that in all cases PET has reached its maximum degree of crystallinity. Similarly, the maximum hardness plateau of the x,,,,, = 0.1 blend, Hma"= 160MPa, is also independent of T, between 145 and 175°C. Thus it can be concluded that at the end of each test, both PET and PEN have reached their maximum degree of crystallinity. It was found [69] that maximum hardness is obtained for the blend containing 10% PEN (118MPa) and that the blend containing 40% PEN shows the lowest hardness value (45MPa), even though neat PEN is harder than PET. This effect can be attributed to the transformation of part of the PEN and PET initially present into an amorphous transesterified polyester of lower hardness than PEN [49],as discussed above. In order to understand better the mechanisms involved in the hardening

441

Effect of Transreactions on Structure and Properties

.-

A

150

z 100

9

W

% 50

PET/PEN (10/90) 0

0

50

150

100

200

250

t, (min)

Figure 6. Microhardness, HI as a function of crystallisation time, t,, for different heat treatment temperatures, T,,for a copolymer with xO,pET= 0.1 [68] process of the PET/PEN blends, an analytical method is proposed, relating parameters of the individual components. The following assumptions are made for the development of the model. (i) Microhardness is an additive function of the microhardness values of the individual components present in the blend [56], i

where zi is the concentration of i in the blend with i = PET, PEN, and NT, where N T refers to the transesterified polyester. (ii) A fraction cr of the component of minor nominal concentration, X O , ~ will , react with an equal quantity of the other component to form a volume fraction 2 a x 0 , ~ of transesterified polyester. The actual volume fraction of polyester, x N T, and of PET and PEN, x i , with i = PET, PEN can thus be expressed as a , function of the nominal concentration, X O , ~ as,

(iii) The microhardness of a material is related to its degree of crystallinity through an Avrami-type Eq. (1). (iv) PET and PEN are allowed

F. J. Baltd Calleja, S. Fakirov, H. G . Zachmann

442

to crystallise above their respective Tg (70°C for PET and 120°C for PEN) whereas the copolyester does not crystallise upon heat treatment and, consequently, its microhardness is assumed to be constant throughout the test. The microhardness of the copolymer films can be described by Eq. (2). The volume fraction of the different components is defined by Eq. (3). The fraction, a,of the component of minor concentration xo,,, which transesterifies with an equal amount of the other component, can be determined by introducing Eq. (3) into Eq. (2) and rearranging to yield (4)

a=

+

x0,rnHrn Z O , M H M- H Z O , ~ ( H-IMHm - 2HNT)

with XO,M > XO,, and HM and H, their respective hardness. At the end of each test, the degree of crystallinity of each component is known, and the corresponding value of their microhardness together with the maximum value of the sample hardness can be introduced into Eq. (4) to determine a. The microhardness of PET and PEN is related to their degree of crystallinity and follows an Avrami-type Eq. (1) if the crystallisation temperature, T,, is larger than their respective T,, or else it remains constant. The microhardness of the transesterified copolyester, H,,, is assumed to be constant throughout the heat teatment. The value of the hardness parameters in Eq. (1) and H=,i, can be determined experimentally on neat PET and PEN. The microhardness evolution with time shows that the copolymer hardens following two different regimes, with two distinct slopes, nl and n2. In view of this result, which can be generalised to all the samples tested, Eq. (1) should be written as (5) (nn-ni) with k = 1 , 2 and GI = G2tc1+2

where t c 1 4 is the time at which the change in hardening kinetics takes place. It was found [69] that the fraction of transesterified copolyester depends on blending conditions, T,, and t,, and the concentration, Z O , ~ , of the various components only, and is independent of the crystallisation temperature, T,.It increases with increasing degree of blending from a = 0.010.04 for X0.i = 0.1 to a = 0.93 for Z O , ~ ~ ,= 0.6. For a given blend composition, X O , ~ a, higher degree of transesterification was calculated for the films with excess of PET (ie., X O , ~ ~>, X O , ~ ~ , . , ) . As already pointed out, Andresen et al. [49] suggest that a single phase has to be formed for transesterification to take place; after 2 min at 280"C, dilution of PEN and PET is completed and the blend is formed as a single phase. They conclude that a certain degree of transesterification is expected after processing for 2 min but are unable to quantify it. Their DSC results

Effect of Transreactions on Structure and Properties

443

show that, after processing for 2min, the melting peak of PEN disappears in blends containing 60 or 70 wt % PET, suggesting that most PEN had transesterified (ie., high a),whereas the blends containing 30 wt % PET clearly show the melting peaks of both PET and PEN, suggesting a lower value of a,in good agreement with these results [69]. The crystallinity-related hardening of PET/PEN blends pressed at 280°C for 2 min, investigated at high temperature as a function of composition and crystallisation temperature between 95 and 175"C, leads to the following conclusions. The microhardness first decreases due to relaxation of internal stresses in PET, followed by an abrupt rise caused by the crystallisation of the individual components in the blend. The analytical model derived describes well the experimental microhardness data and allows the degree of transesterification undergone by the blend during manufacture to be assessed. It should be stressed that this value is difficult to reach otherwise. After 2min at 280"C, only a few percent of copolyester is formed in the films with a low concentration of the PET component whereas a high degree of transesterification is reached as the blend composition tends toward 60 wt % PET. A two-stage hardening process is assumed, characterised by Avrami exponents nl = 1.2-1.5 and n2 = 3.2-5.2 for both P E T and PEN, suggesting a branching fibrillar crystal geometry [69]. 1.3. Eflect of polymer structure and morphology on chemical interactions in condensation polymers

The structural reorganisation of crystalline polymers during annealing is hindered by the existence of folds and tie molecules. Their removal considerably enhances the transition from a metastable structure to a more perfect one. Such a possibility is observable in the chemical reaction of backbone chains, with elimination of the strained parts of macromolecules [70]. It is well known that the stressed parts of macrochains in amorphous regions are characterised by higher chemical reactivity [71]; accordingly, they can be broken selectively. After crystalline reorganisation, these bonds can be re-formed. As pointed out by Wunderlich [72], this field of investigation has been neglected, although it offers new prospects in polymer physics and chemistry. The first observations in this direction were those of Winslow [73],who noted that the polyethylene density steadily increased with the progress of oxidative destruction, due to chain cleavage in the amorphous regions, followed by further crystallisation and reorganisation. Similar chemical reactions were used to eliminate the amorphous regions (the so-called etching technique). An excellent review of these investigations can be found in the work of Wunderlich [28]. Data dealing with solid-state degradation reactions and polymer morphology (including linear condensation polymers) are summed up by Wegner [74] but they are beyond the scope of the present book.

444

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

The studies of Miyagi and Wunderlich [75,76] provide a typical example of the effect of morphology on solid-state post-reactions in linear condensation polymers. Melt-crystallised PET and etched oligomer lamellae (hydrolysed to remove chain folds) from the same polymer were annealed in vacuum at temperatures between 200 and 260°C for 3-48 h. The annealed samples were analysed, with determination of molecular weight, X-ray lowangle spacing, density, heat of fusion, and melting point variation, at different heating rates. In every case the crystal lamellar surfaces proved to be chemically reactive. Chain folds and chain ends on the surface were chemically converted into tie molecules between different locations (points) of the same lamella. The increase in molecular weight resulting from solid-state post-condensation during annealing is shown in Table 2. Table 2. Molecular weight increase of previously etched PET as a result of annealing at 250°C under vacuum [75,76] Annealing time (h) Molecular weight

0 2300

3

6

12

24

9700

14200

19600

25200

In addition to PET, there are other polymers suitable for a similar removal of chain folds and elimination of stress points in the tie molecules, as well as for subsequent polycondensation and improvement of the crystalline structure; these are polyamides, other polyesters, polyurethanes and polyureas. Under suitable conditions, transreactions and post-condensation below the melting point may occur [77]. Another illustration of the effect of polymer morphology on solid-state post-condensation reactions is provided in the studies of Lenz et al. [78821. These authors noted that the isomerisation reactions of five- and sixmember acetal rings in copolyesters,at temperatures well below the melting point, were in the opposite direction to that which could be expected from the data for acetal ring equilibration in a homogeneous melt. This apparent non-equilibrium behaviour was rationalised through continuously removing five-member ring units by crystallisation during the reaction, thereby forcing the reaction in the direction controlled by that capacity to crystallise, irrespective of the normal thermodynamic properties of the reaction itself. Similar results were obtained by Lenz et al. [Sl] on the cis-trans photoisomerisation reaction of 1,4-poIybutadiene, carried out below the melting point, with polymer films of high trans-1,4-content. Under appropriate conditions of temperature and polymer composition, the reaction reverts to non-equilibrium behaviour, attributed to the irreversible recrystallisation of repeating units after isomerisation from cis to trans structures. The interpretation of the composition us. time behaviour observed was based on the incorporation of trans units into crystalline regions on the lamellar fold surface.

Effect of Transreactions on Structure and Properties

445

Furthermore, Lenz and Go [78] observed that random copolymers of cis- and trunsl,4-cyclohexylenedimethylene terepht halate underwent ester-interchange reorganisation at temperatures just below the melting point. This reorganisation in the solid state is the conversion of a random into a block copolymer, evidenced by the change in copolymer properties. The driving force in this anti-entropic ordering process is believed to be the irreversible expansion of the crystalline regions following replacement of c i s by trans-glycol units. Processes of this type are also instrumental in the direct preparation of block copolymers by a solid-state ester exchange reaction. Relevant and interesting experiments were carried out by Crystal [83,84]on selenium with folded chains in the crystalline regions. The contribution of the amorphous phase in semicrystalline condensation polymers to the increase in the reaction rate was illustrated in [85] on copolyamide (nylon6,6/PA6) and in [86] on PET copolyester (flexible ethylene adipate as comonomer units): these have a higher rate of solidstate postcondensation than the corresponding homopolycondensates. The cause for this is the diminished crystallisability of the copolymers. For the same reason, the synthesis of copolycondensates in the solid state (using crystalline monomers) proceeds at a higher rate than does the synthesis of the corresponding homopolycondensates under identical conditions, as found in [87]. If the staring material is a semicrystalline polycondensate rather than not an amorphous one, there are two possible effects on the reaction rate. First, the crystalline phase, as such, restricts chain mobility and diffusivity, thus reducing the reaction rate. On the other hand, as the end-groups are concentrated in the amorphous regions, a higher reaction rate may be expected, due to the higher local concentration of reagent. A further advantage of the use of a semicrystalline polymer as a starting material is that polymer particles do not agglomerate in the reactor, as do amorphous ones [88]. 2. Chemical interactions on the interfaces and interphases of condensation polymers

2.1. Homochemical healing When two samples of the same amorphous polymer are brought into close contact at a temperature above the glass transition, the interface gradually disappears and the mechanical strength of the polymer-polymer interface increases until, at very long contact times, the full fracture strength of the bulk polymer is reached. At this point, the junction surface has in all respects become indistinguishable from any other surface within the bulk material, ie., the junction has “healed”. The study of polymer healing is important for the understanding, at the level of chain dynamics, of the mechanical strength of the polymer-polymer interface, because it is related to

446

F. J. BaltA Calleja, S. Fakirov, H. G. Zachmann

various polymer engineering problems, such as polymer-polymer welding, preparation of bulk thermoplastics from small resin particles, or the formation of a continuous protective film by fusion of individual resin particles from a polymer dispersion. The earliest systematic studies of polymer healing were those of Voyutskii, who proposed a molecular interpretation of the phenomenon [89]. Physical healing, representing a mutual self-diffusion across the interface, supposes high chain mobility, ie., only the chains from amorphous phases can be involved as long as the process is carried out below the melting point [3,30]. Condensation polymers, being usually crystalline and distinguished by higher chain rigidity, are less inclined to physical healing than polyolefins. However, due to their ability to undergo solid-state reactions, they show a new type of healing - chemical healing - in addition to the physical type [3,22-25,30,90-95]. An important advantage of chemical healing as a welding technique resides in the fact that chemical interaction resulting in joining two pieces in contact is possible, even between chains incorporated in the crystallites. It seems quite reasonable to expect that the reactions of exchange or post-condensation in the solid state can be performed on the interface between two samples of condensation polymers, provided that (i) the contact is adequate and (ii) the corresponding reaction conditions are available. Obviously, the chemical bonds formed will contribute to the cohesion of the two polymer samples. Assuming that the junction of the two polycondensate samples is due to INTERFACE I

TR

I

Figure 7. Schematic view of the chemical healing process in semicrystalline condensation polymers at temperatures just below melting: TR transreactions; AC additional condensation; D diffusion [23]

Effect of Transreactions on Structure and Properties

447

(i) mutual diffusion (D), (ii) transreactions (TR), and (iii) solid-state additional condensation (AC), as represented schematically in Figure 7, the conclusion is that transreactions, mainly involving ester-ester groups, are the most important factor in chemical healing. This conclusion is supported by the results of Devaux et al. [5,96-981, discussed in detail in Chapter 3; they demonstrated that the direct exchange reaction is the dominant one in the melt of a PBT/PC mixture. Lower molecular weight samples would otherwise need lower contact areas, because they comprise many more endgroups and therefore are more liable to undergo additional polycondensation than higher molecular weight samples. Furthermore, mutual diffusion is greater in lower molecular weight samples, which also reduces the contact area. The influence of the healing temperature on the bonding effect was studied and in this way the main contribution of transreactions to the welding process was demonstrated [24]. Healing conditions and results of static mechanical measurements are summarised in Table 3. Table 3. Healing conditions and static mechanical data for welded PET samples (241

Healing temperature ("C) for 24 h duration 140 160 180 200 220 240 250 Stress at break (kg/cm2) _ _ 824123 193 f 26 181 f24 210 f- 32 Interface shear 0 0.4 6.6f 2.8 6.2f3.4 stress (kg/cm2) The first conclusion to be reached is that the fracture behaviours of at the welded samples depends strongly on the healing temperature (Th): high Th values, fracture (rupture) of the strip occurs outside the contact area; at low healing temperatures debonding of the interface occurs. The potent influence in this particular case of temperature on the healing effect suggests that most bonding forces are engendered by the chemical reactions taking place during the healing process, rather than by interpenetration of the chains due to mutual diffusion. The solid-state post-condensation rate decreases exponentially as the temperature drops still further below the melting temperature; this is demonstrated with PA6 I991 and PET [loo], but with the latter, the reaction did not in fact take place at temperatures below 180°C [100,101]. The temperature dependence of the transesterification reaction is quite similar. It was found by Droscher and Schmidt [lo21 that after annealing for 63 h at 2OO0C, 72% of the diethylene glycol units were transferred in PET (after 2 h, they were less than 3%) and about 25% of them were transferred after 14 h at 225OC. Thus, the conclusion to be drawn is that at treatment temperatures below 180°C, the chemical reaction does not contribute to

448

F.J. BdtB Calleja, S. Fakirov, H. G. Zachmann

the healing effect. In contrast, the closer the healing temperature is to the melting point, the more effective are transreactions and post-condensation in the welding process. In summary, it may be concluded that the results obtained demonstrate a new molecular mechanism of the healing phenomenon in semicrystalline linear condensation polymers - healing as a result of chemical reactions between contiguous macromolecules on the interface. It is to be expected that chemical healing is a generic property of all polymers, characterised by their capacity to undergo solid-state interchain reactions. Experiments with other polyesters and polyamides support this conclusion, as shown below. 2.2. Hetemchemical healing and healing with coupling agents

Starting from the fact that exchange reactions are possible not only between polycondensate molecules of identical chemical composition, but also between macromolecules of different chemical composition [5,81-83,9799,104,1051, heterochemical healing can be expected to occur. In order to prove this assumption, experiments with polyamides and polyesters have been carried out [22]. These two types of polycondensates were selected for two reasons: (i) dicarboxylic acids react with glycols as well as with diamines and (ii) these partners in the healing experiments have very close melting temperatures, i e . , healing can be conducted at a temperature equally favourable to either polymer. Experiments were carried out with such healing partners as PA66 and PET, and PA6 and PBT [22], and it was demonstrated that heterochemical healing does actually take place. Surprisingly, the bond in the hetero-cases was stronger than in the homo-cases, where the bond, not the strip, failed. Alternative means, such as coupling agents or catalytic enhancement, not only provide evidence of chemical healing, but also offer an attractive way of achieving chemical bonding in somewhat milder conditions. The application of coupling agents rests on the fact that chemical reactions between condensation polymers and low molecular weight acids, alcohols, or amines are also possible [106]. In the case of a polycarboxylic acid or a polyol, their interaction with different macromolecules results in crosslinking. When such an agent is spread on the contact surface, unaided or in combination with a suitable transreaction catalyst, molecules from either contacting surface are involved in the reactions, thus contributing to the bonding of the two pieces. The following compounds were used as coupling agents: pyromellitic acid, benzophenone tetracarboxylic acid, oxalic acid, malonic acid, pentaerythritol, and pyromellitic acid with Zn(0Ac)Z [106]. These data indicate that solid-state reactions between linear condensation polymers and low molecular weight compounds provide an alternative way of conducting chemical healing involving macromolecules.

Effect of Transreactions on Structure and Properties

449

2.3. Chemical healing in crosslinked polyamides

Chemical crosslinking was conducted prior to healing; it was expressed by the formation of -CH2- bridges between the nitrogen atoms of amide groups through methoxy methylation, as described in [107,108].It was found that samples of PA 66, healed at the same temperature but for different periods of time, showed different stresses at break, which were higher for longer healing times [3]. What could be the reason for the increase in the stress at break with increasing healing temperature and duration? Obviously, it is not due to an increase in the degree of crystallinity during the healing process, since all the samples discussed were crosslinked, and that excluded any further crystallisation. Insofar as chemical crosslinking is effective only in the amorphous regions, the crystalline regions afford some freedom of motion with respect to the defects incorporated in the crystallites. It can be expected that, during healing, these defects would leave the crystallites, moving into the amorphous regions, as demonstrated by Fischer et al. [lo91 for polyethylene and by Fischer and Fakirov [110]and Zachmann et al. [lll]for PET. As a result of this process, the perfection of crystallites improves. The higher the healing temperature and duration, the higher the crystallites' perfection. Such an improvement of the crystallite structure contributes to an increase in the stress at break [30,33].These assumptions seem to be reasonable because healing may be regarded as annealing (thermal treatment in the solid state). More important evidence is that samples with higher initial crystallinity, due to heat treatment prior to healing, have higher initial shear stress values than those with lower crystallinity (under identical healing conditions). Furthermore, crosslinked nylon 11 samples showed larger shear stress values than the non-crosslinked ones (8.6 kg/cm2 as against 3.4 kg/cm2). At first glance, these two interesting observations, higher interface shear stress values of the crosslinked samples as well as those with higher crystallinity than of the non-crosslinked and less crystalline ones, seems to be in disagreement with the basic concepts of solid-state reactions in polymers. It is generally accepted that any restriction of the chain motion, including functional group mobility, leads to a drastic drop in the rates of solid-state interchain reactions. Crystallinity considerably restricts, and crosslinking nearly prevents, any mobility; nevertheless, results from chemical healing experiments indicate increased reactivity in the solid state. Evidently, the considerations of the relationship between chain (or chain segment, or functional group) mobility and reactivity in the solid state are valid for systems with no mechanical stress points in the macromolecules. In cases where the system abounds in entanglements and stressed tie molecules, as a result of crystallisation or relaxation processes caused by thermal or mechanical treatment, it seems that a different concept is required. This concept, emphasised by Wunderlich [28,72],presupposes increased reactivity at stressed

450

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

interchain contact points. In the cases discussed above, the second factor (stressed contact points) probably contributes more to the increase in the rate of solid-state interchain reactions than the first one (chain mobility) [31. Furthermore, the same observations lead to another important conclusion on the dominant role of exchange reactions in the welding process. In fact, an exchange reaction can take place between any two ester-ester or amide-amide groups, or between functional chain end-groups and ester or amide groups, if the respective partners in the transreaction are close enough to each other. The same also applies to additional solid-state polycondensation, i.e., to a reaction solely between functional chain end-groups. Moreover, taking into account that (i) the concentration of the partners for transreaction is much higher than that of terminal groups and (ii) that crystallinity and crosslinking suppress mobility, terminal groups are far less likely to collide. At the same time, crystallisation and crosslinking lead to an increase in the rate of transreactions, mostly of the ester-ester and amide-amide types, due to the increased concentration of stressed contact points. Such circumstances are favourable to the transreactions and the conclusion follows that the healing effect is mainly the result of interchain transreactions, involving ester-ester or amide-amide groups. The established fact that higher treatment temperatures correspond to higher interface shear stress values may also be interpreted as “more numerous and more stressed contact points increase chemical reactivity”. The higher the healing temperature, the higher is the Brownian motion. This means a higher stress at the interchain contact points, and hence higher transreaction rates. The results of the stress effect at contact points on chemical reactivity are in very good agreement with those of Wunderlich [28,70,72]and Porter et al. [71]on polymer stress reactions. These experiments with crosslinked polyamides clearly indicate that, in their case, the healing effect is exclusively due to solid-state reactions taking place at the contact interface rather than to diffusion processes. In conclusion, “purely” chemical healing is observable in crosslinked polycondensates, as “physical” healing is greatly hindered by diffusion restrictions. 2.4. l’hnslwrctions at the phase boundary of semisolid blends of

condensation polymers

In the cases described above, the interchain reactions take place between partners which are in the solid state, i.e., at temperatures below their melting points. The other extreme case is when the blend partners are in the molten state, i.e., above their melting temperatures. This case is described in detail in Chapters 3, 7 and 8. Here a special case is discussed, i.e., when the blend of condensation polymers is thermally treated at a temperature which is inbetween the melting points of the components. Under such conditions, the lower-melting partner is in the molten state, while the

Effect of Transreactions on Structure and Properties

451

higher-melting one remains solid. From the kinetic point of view, these conditions are more favourable for chemical interactions than those of chemical healing, and less favourable compared to the case when both partners are molten. The preparation of the recently developed [112] microfibrillar reinforced composites (MFCs) is a good illustration of such partners and of the extent of interactions on the phase boundary. Unlike the classical macrocomposites (e.g., fibre-reinforced ones) and the molecular composites (with “single” rod-like macromolecules as reinforcing elements), the third type of polymer composites is reinforced by microfibrils. The fibrils are created during MFCs preparation by drawing the polymer blend, leading to the orientation of both components (fibrillisation) followed by melting of the lower-melting component (isotropisation) with preservation of the oriented microfibrillar structure of the higher-melting component. It is important to note that in addition to isotropisation during short (several hours) thermal treatment, chemical reactions (additional condensation and transreactions) between condensation polymers in the melt [lo51 as well as in the bulk solid state [3] take place at the interfaces, resulting in the formation of a copolymeric interphase. The latter plays the role of a compatibiliser, i e . , one is dealing with a self-compatibilisation effect, as long as there is no need to introduce an extra synthesised copolymer of the blend components according to the usual approach [113-1161 (see also Chapter 10). Compatibilisation can be effective only at the initial stages of chemical interactions. During thermal treatment the interphase grows and involves all of the isotropic (molten) component and the amorphous portions of the fibrillised one in block copolymers, thus transforming the homopolymeric matrix into a copolymeric one. In the case of crystallisable homopolymers, the block copolymers also crystallise. As the chemical interactions progress, the block copolymers randomise and convert themselves into statistical copolymers. An important result of the randomisation is the loss of the crystallisation ability of the matrix, which can influence the overall behaviour of MFCs. Evidence of these physical processes occurring during the preparation of MFCs and the various stages of the interfacial chemical interactions was obtained by means of DSC, WAXS, S A X S , IR, and solubility measurements of binary and ternary blends of PET (T, = 265”C), PA6 (T, = 222”C), and PBT (T, = 225°C). This matter is considered in more detail in the first section of Chapter 8. The difference in the size of the reinforcing elements, and particularly the method of preparation of MFCs, in contrast to the macrocomposites, makes MFCs similar to natural composite materials, where the fibrils and the matrix grow simultaneously with interpenetration by formation of chemical bonds. MFCs make it possible to mimic such natural materials due to the occurrence of transreactions between the blend partners.

452

F. J. Balti Calleja, S. Fakirov, H.G.Zachmann

This type of reactions in immiscible polymer blends presumes some competition between phase separation and transesterification. The competition was recently studied by Han et al. [116] as a function of time and temperature, by DSC and small-angle neutron scattering (SANS) in PET/PC blends. It was found that the global structure coarsens at T < 200°C due to the dominance of phase separation over transesterification and melts at T > 220°C due to the dominance of transesterification at the domain interface. However, transesterification is slow but still significant even at T < 220°C. An intricate balance of transesterification and phase separation rates controls global and interfacial structures. The latter become measurable under certain conditions, and the interfacial thickness between P C or PET and the copolymers generated by transesterification increases with time. DSC results are consistent with results obtained by SANS, but the latter is more sensitive than the former and differentiates the structural change at different length scales caused by phase separation and transesterification [116]. 2.5. Chemically released d i f i s i o n via tmnsmactions in condensation polymers

The healing experiments with crosslinked polyamides and the results obtained may be considered as an indication of the existence of chemically released diffusion in the solid state by motion of chain segments caused by inter- and intrachain chemical reactions, mainly of the exchange type. Thus, as in molten and dissolved polycondensates, there is, in solid condensation polymers, a continuous chemical interaction, leading to an equilibrium state. At the same time, this means that even in the solid state, the chemical composition of macromolecules does not remain constant with time, i. e., parts of macromolecules are constantly being exchanged. The experiments with crosslinked polyamides also indicated that chemically released diffusion is the main (and possibly the only) mode of mass transfer in polycondensates. It seems very likely that segment transfer through the bulk is not only due to physical diffusion processes but to chemical ones as well. As long as the temperature is high enough, the segment no longer appertains to its initial macromolecule. Due to transreactions between different macromolecules, macromolecule parts frequently change their immediate neighbours. Furthermore, the chemical composition of such a segment can change, if transreactions occur between chemically different molecules. Thus, in the case of linear condensation polymers, the classical picture of self-diffusion in bulk polymer, or the reptation model, as proposed by Edwards [117] and de Gennes [118], needs to be complemented. In cases where transreactions are possible (transesterification, transamidation, transetherification, etc.), chemically released diffusion of the macromolecules can occur. This aspect further demonstrates that at temperatures close to melting,

Effect of Transreactions on Structure and Properties

453

mass transfer in such polymers takes place, as in the melt, along with chemical reactions. The contribution of the latter to mass transfer greatly depends on the reaction conditions (temperature, pressure, catalyst, medium, etc.). Chemically released diffusion is apparently insensitive to molecular weight. The determining factor seems to be the number of favourable contacts between different chains. This number is expected to increase with increasing temperature, due to chain relaxation. The entanglements present (stressed points) act in the same direction. The higher their concentration, the greater the contribution to the chemically released diffusion. These results on the chemical healing phenomenon shed some light on certain fundamental aspects of the chemistry of condensation polymers. Diffusion (mass transfer) in linear polycondensates at elevated temperatures (close to the melting point) is believed to be achievable mainly by chemical reactions between macromolecules (entanglements, stressed points, favoured contact points, etc.). The mass transfer by such chemically released diffusion means that the chemical composition (even in h e mopolyesters) steadily changes. These conclusions are derived from chemical healing experiments with crosslinked polyamides, i. e., samples with totally inhibited or greatly restricted segment mobility [3] (see also Chapter 5). 3. Effect of interchain reactions in condensation polymers on their

mechanical properties

3.1. Mechanical properties as revealed by tensile experiments

The mechanical properties best characterising polymers as materials depend on the molecular weight in the following way. Tensile strength and tear resistance [119] (Figure 8) along with the elastic modulus E increase with molecular weight [120-1241. Stability to impact and stress cracks is considerably improved when the molecular weight is increased [124].A particularly marked increase in the elongation at break with increasing molecular weight has been observed. PET bristles, for example, after quintuple drawing at room temperature followed by additional solid-state polycondensation, are able to undergo another triple drawing at room temperature, thus achieving a final 15fold elongation [33]. The rupture force capable of destroying the polymer sample increases with increasing molecular weight. The fracture mechanism is also greatly dependent on the value of the molecular weight [125]. At the same time, increasing the molecular weight leads to a decrease in crystallinity and surface hardness [126]; these two parameters are also very temperaturesensitive [126]. The increase in the molecular weight as a result of thermal treatment a t elevated temperatures in vacuum or in an inert gas flow has a remarkable effect on the drawability of condensation polymers, as shown in Figure 9.

454

F. J. B d t 6 Calleja, S. Fakirov, H. G. Zachmann 500

- 100

-- 70

- 80

400 -

-60

- 40 100 -60

I

10 I

18

I

I

14

18

[VI [d/d I

I

24

30

~,.10-~

2E b

20

Figure 8. Dependence of the tensile strength TS and tear resistance TR on the intrinsic viscosity [q] and molecular weight , respectively, for poly(ether esters) with various molecular weights achieved by additional solid-state condensation

cb

160

(%)

i'

Figure 9. Dependence of the relative elongation at break (Q) on the annealing temperature (T,)for 6h in vacuum (with continuous pumping) for drawn bristles of PBT (a) and PET (b) [30]

One can see that the relative elongation at break, q,,of PET and PBT increases with increasing T,. This dependence is particularly well expressed

Effect of Transreactions on Structure and Properties

455

in the case of PET - all samples annealed below 250°C reveal a relative deformation at break of less than 60% while those treated at T, of 255 or 260°C exhibit an extremely high elongation of about 200%. Taking into account that these samples have been subjected to drawing (A = 5.5) before annealing, the total draw ratio achieves the value of 1O:l and 15:1, respectively (311. Similar results are reported in [112] for drawn and annealed PET films. Several authors [126-1391 have studied highly oriented PET fibres or films. For instance, two-stage drawing was performed, the first stage being carried out at 100°C and leading to a draw ratio of 4:1, while the second stage at 225°C resulted in a draw ratio of 6:l [138]. In the present case, PET bristles were highly drawn (A of about 15) at room temperature [31,33] (Figure 9). Slusallek and Zachmann [111,139] reported the same result for drawn P E T films annealed above T, = 240°C. From these results, it can be concluded that the chemical changes during annealing really affect the deformation ability when carried out in vacuum and at high temperature and for a long duration.

3.2. Mechanical properties as revealed by microhardness 3.2.1. In blends with partial chemical intemctions

Hardness determination has been shown to be a promising technique for the microstructural investigation of multicomponent blends and can provide information on the degree of interpenetration of the blend components [SO]. The case of blends of low- and high-density polyethylene (PE) is an example where hardness can be described in terms of an additive system of two independent components [140]. However, in systems such as PE/polypropylene prepared from semidilute solution or PBT/PC blends, a deviation from the additivity law is detected [141,142]. In these cases, the deviation of the hardness from the additive behaviour of the separate components can be related to the changes of crystallinity and to the thickness of the crystals. It is worth mentioning that in blends of condensation polymers, in contrast to polyolefins, the blend partners can interact chemically and form copolymers. For this reason, the observed deviation from the additivity law for such systems can also be related to the chemical changes taking place during blending. In order to distinguish between the contribution of crystallinity factors and chemical changes, one has to study condensation blends in the glassy state. A series of blends of condensation polymers and the respective copolymers formed by transreactions has been studied by applying the microhardness, H , technique [143]. The mechanical properties of random copolymers of PET with PHB and PEN with PHB, forming liquid-crystalline melts when the PHB content exceeds 30% [144], were investigated as a function of composition [143]. In the PEN/PHB copolyesters containing up to 50% PHB, crystals of PEN are formed, whereas in systems containing

456

F. J. BaltA Calleja, S. Fakirov, H.G . Zachmann

80-90% PHB, crystals of PHB are found. It is known that hardness is directly related to the macroscopic mechanical properties of the materials [56,57,145].The mechanical behaviour is interpreted in the light of several microstructural parameters, such as crystal thickness, polymorphic crystal forms, and the fraction of crystalline content [56]. In the case of quenched glassy copolyesters containing PHB, the additivity of microhardness of the separate components is postulated. PHB and its copolymers containing more than 50% PHB cannot, however, be solidified in the fully amorphous form [144]. For the crystallised samples of PET/PHB and PEN/PHB it is shown that the microhardness of the PET and PEN crystals is an increasing function of the crystal thickness [143]. On the other hand, at high concentrations of rigid PHB units, the samples always crystallise and show a hardness increase proportional to the PHB content. In the case of copolyesters of PET/PEN, both copolymer units having flexible chain segments and lacking liquid-crystalline behaviour, the entire range of copolymers can be prepared in the amorphous state. This enables the study of the influence of composition and molecular structure on the microhardness value [51]. Samples of random PET/PEN containing between 10 and 80mol% PEN were synthesised [146] and amorphous films were obtained by melt pressing of these samples as well as of neat PEN, followed by quenching in ice-water. The films were crystallised by annealing the glassy materials at various temperatures. In the samples containing 0-30 mol % PEN, the PET sequences crystallise while the PEN segments remain in the amorphous regions. On the contrary, in the samples containing 80 and 100% PEN, the PEN sequences crystallise in the a-polymorphic form [144] and the PET segments are expelled in the non-crystalline regions. Only the samples containing 50 and 60% PEN, annealed at high temperature, do not crystallise 1511Figure 10 shows the variation of the microhardness of the quenched and that of the crystallised samples, H , as a funcamorphous samples, Ha, tion of the PEN content. It is seen that Ha increases linearly with the increasing concentration of PET units, wPET,according to the prediction of a mechanical parallel model:

where wPENis the total concentration of PEN units within the copolymer. The higher values of H shown in Figure 10 for the crystallised samples are related either to the presence of PET (left) or PEN (right) crystals. When the concentration of PET units prevails, the probability of bundles of PET sequences agglomerating and forming crystallites also increases. On the other hand, at higher concentrations of PEN segments, the probability of these sequences packing in the form of crystalline aggregates is also higher. The increase in crystallinity and crystal thickness consequently gives rise

Effect of Transreactions on Structure and Properties

457

40 350

1

D.

0

20

0

40

0

60

w (% PEN)

80

. ... . . .

1 100

Figure 10. Experimental microhardness values, H , us. PEN content (mol%) for quenched materials; Ha,annealed samples. The crystal hardness values, H e , are derived using Eq.(6) [51] t o higher H values. In addition, the hardness of the crystals, H,, is calculated using the additivity relationship of crystalline and amorphous hardness values [147] for the PET/PEN compositions 100/0,90/10,20/80, and 0/100: (7)

H

=H

+

, ~ L H a ( l - QJ,)

where a~ = l,/L represents the so-called linear crystallinity. Equation (7) is shown t o apply for P E T samples crystallised at various temperatures and times of crystallisation [61]. For the 20/80 and 0/100 PET/PEN compositions, the H, values are almost the same because the PEN crystal thicknesses and crystallinities in both cases are very similar. On the other hand, the H, value for the copolyester with 90/10 PET/PEN composition is slightly higher that that for 100/0 PET/PEN due t o the smaller PET crystal thickness of the latter. In conclusion, unlike PHB copolymers with a high concentration of PHB units, unable to quench into the amorphous state, glassy amorphous materials can be produced from the above flexible copolyesters over the entire range of compositions. In the case of the amorphous PET/PEN copolymers, the microhardness reveals a simple additive behaviour over the entire composition range. Similarly to the results obtained with PET/PHB and PEN/PHB systems, the annealed samples show the lowest microhardness values at concentrations close to 50%. This finding is consistent with the fact that, even after annealing, the samples of these compositions are

458

F. J. Balti Calleja, S. Fakirov, H. G.Zachmann

always amorphous. However, the rise in PET or PEN content results in increased microhardness due to the contribution of developing crystalline regions. The influence of crystal thickness, crystallinity, and the hardness of the crystals can be quantitatively accounted for. Further, attention is focused on the mechanical behaviour as evaluated by microhardness measurements of PET/PEN blends and/or copolymers in the glassy state. By exclusion the occurrence of crystallisation in the blends, it is possible to follow the effect of chemical composition in a pure form on microhardness as well as the extent of transreactions taking place in the blends. For this reason, the microhardness of films of PET/PEN blends prepared by coprecipitation from solution, followed by melt-pressing and quenching, has been determined [68]. The miscibility, transesterification, and crystallisation properties of these films are discussed in Section 1.2 [49]. The PET/PEN compositions chosen are between 10/90 and 90/10. It is interesting to examine the variation of the mechanical properties (microhardness) of these PET/PEN blends as dependent on both composition and melt-pressing time, t,, and to compare the results with those obtained with random copolyesters of PET and PEN [51]. As discussed in Section 1.2, studies on blends of PET and PEN indicate that their thermal behaviour depends strongly on the time of compression moulding, t,, before quenching the fdms in ice-water [49].Let us remember that for times between 0.2 and 0.5min two glass transition values, T,, are observed by means of dynamic mechanical analysis (Figure 3), indicating the presence of two amorphous phases, a PEN-rich phase and another PETrich one. For t m in the range of 2 to 45min, only a single value of T, and, consequently, a single phase are found to exist. Within this time range, transesterification of the two components also takes place, resulting in a copolyester of PET and PEN (491. The PET/PEN blends are completely amorphous after quenching from the melt, as revealed by DSC experiments [49]. Figure 11 illustrates the variations in microhardness with melt-pressing time for various PET/PEN compositions. In all cases, H shows first a rapid initial increase with time, exhibiting a maximum just before t, = lOmin, and then, for longer times, a gradual decrease down to values which can be even lower than the initial ones. In order to explain the variation of H us. t,, it is convenient to analyse the results of hardness as a function of composition. It is seen in Figure 11 that for a melt-pressing time of about 0.2 min, H increases linearly with increasing PEN concentration, according to the prediction of the mechanical parallel model given by Eq. (6). These results indicate that the microhardness of the PET/PEN blends shows similar values to those obtained for PET/PEN amorphous random copolyesters [51]. As t , is increasing up to 10 min one observes a shift of the straight line toward higher values. Finally, for t , = 45min, the lowest microhardness is observed. As already mentioned, at short melt-pressing times of about 0.2-0.5min

459

Effect of Transreactions on Structure and Properties

180

h

h

\

170

4

W

,44156

140 130

30170

60140

P

120 I

o

I

10

1

ao

\

30

40

90/10 50

60

Figure 11. Microhardness, H , us. pressing-time,t,, for different PET/PEN compositions e, 10/90; +, 30/70;A, 44/56; 0,60/40;W, 70/30;0,90/10 [67] two phases are observed, and at times of 2min or more a single phase is found. For t, of about 10-45min, no crystallisation and melting during heating (10"C/min) are observed in the DSC, indicating the formation of an amorphous copolyester of PET and PEN by transesterification during melt-pressing. The initial increase in hardness up to t, = 0.5 min could be attributed to the corresponding shift of Tgtoward higher temperatures. It is known that, in the case of amorphous blends, temperature is the dominant factor in determining the yield behaviour of the glassy material [141]. The further increase in hardness up to t, = 2 min could be ascribed to the transition from a two-phase system into a single amorphous phase, composed of interpenetrating molecules of both polymers. Such a homogeneous system should offer a higher mechanical resistance to yield and to plastic deformation. Finally, H increases further up to the melt-pressing time of about lOmin, when copolyesters of PET and PEN are already formed by transest erification. A question arises at this stage: why does H gradually decrease with increasing t , if the molecular weight and the viscosity remain practically constant? One possible explanation could be that at the beginning of t r a n s esterification the copolyester has a rather block-like character. Only after longer times does it become a random copolymer. Therefore, the above results indicate that the microhardness of the block copolyester is greater than that of the random one. The presence of blocks could lead to microphase separation between PEN and PET blocks. It seems, then, reason-

460

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

able to assume that sequences of blocks with the same chemical composition, packed in parallel, would yield less easily than parallel copolymer sequences of statistical composition. In conclusion, in order to achieve the optimum mechanical properties of these blends, one should apply melt-pressing times in the range of 510 min. Otherwise, the mechanical properties reflected by microhardness can be reduced by 10-15%.

3.2.2. In copolymers from blends of condensation homopolymers The PET/PEN system discussed above has the peculiarity that, depending on the treatment conditions, mainly temperature and duration, one can obtain copolymers with a blocky or random type of sequential order. For this reason, it is of interest to extend this study on the mechanical properties of blends of condensation polymers where chemical interactions cause profound changes in the sequential order, resulting in complete randomisation. The blend partners PET and PC are appropriate in this respect and blends are prepared by means of coreactive mixing (temperature of 275-290°C, vacuum of 0.1-0.2 mm Hg, magnesium hydrohexabutoxy-orthotitanate as a catalyst). After 45min the respective blend is extruded and granulated [148]. Figure 12 shows DSC curves of neat PET and PC and their blends after coreactive blending and film preparation. As expected, the homoPET reveals both a well-defined glass transition and a crystallisation peak at about PET/PC (wt %) 100/0

90/10 m/30 50150

60

1

80

I

100

I

120

T ("1

I

140

I

160

180

Figure 12. DSC curves of homoPET and homoPC and their coreactive blends with various compositions [148]

461

Effect of Transreactions on Structure and Properties

65 and 120"C, respectively; the homoPC shows only a glass transition at about 130°C. The observation of only one glass transition temperature in all blends, which increases with the rise of the initial PC content, is more striking. There is no sign of crystallisation in the range studied (60-18OoC), either. These findings support the conclusion drawn on the basis of WAXS results and the observations reported for the same blends [116] that the samples are in the glassy state. This peculiarity of the blends is very important for the subsequent microhardness measurements performed at room temperature. Table 4 summarises the glass transition temperatures derived from the curves displayed in Figure 12, together with other experimental and calculated data. Table 4. Glass transition temperature T, measured experimentally and calculated according to the equations of Fox (Eq. (8)) and Couchman (Eq. (9)), the difference in the specific heat AC, between liquid and glassy states at T,,the microhardness H , and the depressions in glass transition and microhardness, AT, and AH, respectively, for neat PET and P C as well as for their coreactive blends with various compositions 11481 Composition (PET/PC) wt %

mol %

T, ("C)

AC,

H

AT,

AH

Molar ratio Exp. Calc. Calc. (cal/deg) (MPa) ("C) (MPa) Eqm E d 9 )

100/0

100/0

-

90/10 70/30 50/50 30/70 10/90 0/100

92.6/7.4 75.5/24.5 56.9/43.1 36.1/63.9 12.8/87.2 0/100

12.50/1 3.08/1 1.32/1 1/1.77 1/6.81 -

70 0.078 89 72.5 70.8 0.0805 87 80.0 75.5 0.0725 93 89.7 82.6 0.0759 100 101.9 93.7 0.0658 113 117.9 113.0 0.0658 128 0.041

128 129 131 138 143 149 152

58

39 41

34 28 15 0

24 23 21 14 9 3 0

Figure 13 shows the dependence of microhardness on the composition of the coreactive blends. The H values of the two homopolymers are also given. One observes an almost linear decrease of microhardness with the rise of PET content in the blends, ie., with increasing concentration of the component with lower H value. The solid straight line in the same Figure 13 illustrates the microhardness values calculated according to the mechanical parallel model (Eq. (6)) using the H values of amorphous PET and of PC. The experimentally measured and calculated values for the coreactive blends are in rather good agreement. In the case of crystalline polymers and copolymers, it is known that the microhardness depends primarily on the crystal characteristics, such as size, perfection, chain conformation, etc. [56,57].Dealing with completely amorphous samples, these parameters obviously cannot be used; the only available quantity is the glass transition temperature. For this reason an

F. J. Balt6 Calleja, S. Fakirov, H. G. Zachmann

462

120 I 0

I

I

I

I

20

40

60

80

100

PET (wt %) Figure 13. Dependence of microhardness H on the composition of coreactive PET/PC blends. The values of neat PET and PC are also given. The solid line reflects the H values calculated according to Eq.( 5 ) [148] attempt is made to look for a relationship between H and T,, bearing in mind that they should be sensitivein a similar way to the blend composition and to the occurrence of chemical reactions, leading to the formation of copolymers. In Figure 14, the Tgvalues are plotted for the two PET and PC homopolymers as well as for their coreactive blends. The data are taken from the DSC curves illustrated in Figure 12. In contrast to the microhardness behaviour (Figure 13), the glass transition temperature decreases non-linearly with increasing PET content in the coreactive blends. The Tg value of the blend richest in PET deviates from the general trend, even being slightly higher than the preceding one (Figure 14). This could be related to the possible effect of crystallinity on Tg.Obviously, the cooling conditions in the present case do not exclude crystallisation, as in the case of homo PET. Similarly to the approach to microhardness, the expertimentally evaluated Tgvalues are compared to theoretically derived ones, using two equations frequently applied to polymer blends and copolymers: the Fox equation [149], which is close to the mechanical parallel model serving for the microhardness calculation (Eq. (6)):

where T, is the glass transition temperature of the binary blend, T,, and Tg,are the glass transition temperatures of polymers 1and 2, respectively, while w1 and w2 are their mass fractions.

463

Effect of Transreactions on Structure and Properties

140

60

r

'

0

I

I

20

40

I

60

PET (wt %)

I

80

100

Figure 14. Dependence of the glass transition temperature T, on the composition of coreactive PET/PC blends. The values of homo PET and homo PC are also given. The data are evaluated from the DSC curves displayed in Figure 12. The dashed line reflects the Tgvalues calculated according to Eq.(8), and the solid line those according to Eq. (9) (1481

The equation of Couchman [150]is mostly used for miscible blends:

(9) lnTg = (XiAC,, lnT,,

+ X2ACp, lnT,,)/(X~AC,, + XzAC,,)

where T, is the glass transition temperature of the blend, Xi is the molar fraction of the component i, and ACpi is the difference in the specific heat between the liquid and glassy states at T,. While Eq. (8) supposes a linear change of T, with the change in composition, Eq. (9) predicts a monotonic dependence of Tgupon concentration, and when T,, > T,, and AC,, < AC,, , the glass transition of the blend Tg is a concave function of blend composition. The values of T, for coreactive blends differing in composition, as derived from Eqs. (8) and (9),are given in Table 4 and plotted in Figure 14 together with the experimental data. It can be concluded that

tho

moaqmrerl T- v a l i i o c aro r l n c n r tn

thncn

AOrifrDTJ

frnm

t h o Fny

oniiatinn

r'igure 13 is a plot 01 the relationship between the microharaness aepression A H = HPC-HBL and the glass transition depression AT, = Tpc-TgBL B& for coreactive blends differing in composition, where Hpc and H denote the microhardness of homo PC and of the coreactive blend, respectively, and T",' and TgBL are the glass transition temperatures of homo PC and of the coreactive blend, respectively. A linear correlation between AH and AT, is visualised in Figure 15.

F. J. Bdt&Calleja, S. Fakirov, H. G.Zachmann

464

_ -

0

10

20

30

40

SO

60

70

AT, (“1 Figure 15. Relationship between the microhardness depression AH and the depression in the glass transition temperature mgfor PET/PC blends differing in composition [148] Considering the reports [6,16,151-1541 and taking into account the experimental conditions applied for coreactive blending in this study (275290°C, duration of 45 min, transesterification catalyst), intensive chemical interactions should be assumed. It could also be expected that these reactions would lead not only to the formation of copolymers but would result in a more or less complete randomisation of the sequential order of the repeat units; in this case, the initially two-component blends should be converted into single phases. What is more, due to randomisation, these single phases should be amorphous. This is actually the case, as concluded from the DSC curves displayed in Figure 12 -unlike homoPET, the blends do not crystallise and they exhibit only one glass transition temperature. The conclusion that the initial PET/PC blends are converted via chemical interactions into amorphous singlephase materials has important consequences when the “blend” is characterised with respect to its mechanical properties. Unfortunately, this peculiarity of such “blends” is often d i s regarded or unterestimated, which results in erroneous conclusions, e.g., concerning the miscibility of the starting partners [l]. A second detail concerns the method of expression of the “blend” composition. Since condensation polymers are reacted and converted into copolymers, the latter being uniform with respect to the number of components (as well as to the number of phases, provided that no phase separation via crystallisation or dephasing takes place), it seems more reasonable to express the ratio of the components in mol% rather than in wt %. This re-

Effect of Transreactions on Structure and Properties

465

flects more realistically the system composition and at the same time it is the expression as molar ratio that suggests the character of the sequential order in the chains, assuming complete randomisation. For this reason, Table 4 displays the compositions in wt %, mol% and molar ratios. The fact that the molar ratio reflects the block length upon complete randomisation allows direct conclusions about the crystallisation ability of the copolymers obtained. For instance, in the present case only the blend richest in PET (90/10) is potentially crystallisable. For the rest of the blends, the PET “blocks” are too short to form lamellae with a thickness of 5040A which represents the lowest limit for crystallisation [153]. It should be pointed out that in the case discussed here, the differences between wt % and mol % are not that large because the molecular weights of the two repeat units are quite close to each other, but in other cases these differences could be significant. Bearing in mind the outlined peculiarities of blends of condensation polymers, particularly when they consist of one component and of one phase (this case is an exception rather than a general rule, since block copolymers usually consist of two, three, or more phases) , the application of the additive law for the evaluation of their characteristics does not seem quite justified. The observed good agreement between the measured microhardness values and the calculated ones using Eq. (6) (Figure 13) suggests an important conclusion in this respect. Basically, the application of the additive law to blends assumes the presence of spatially well defined regions of chemically and structurally uniform moieties, which is not the case for the amorphous PET/PC copolymers since a single Tgis observed (Figure 12, Table 4); it follows that the contributions of the two species to the microhardness of the copolymers are transferred via the respective repeat units building up the copolymer molecules. Since in the present case no crystalline phases are observed which can independently contribute to the copolymer microhardness, the H values can be regarded only as arising from and depending on the chemical composition of the copolymer. Possibly, one is dealing with the intrinsic contributions of the two types of repeat units to the microhardness of the copolymeric solid. It can be concluded that the additive law can be applied on the molecular level, i e . , the microhardness of amorphous copolymers, differing in their compositions, obeys the additive law based on the microhardness of the respective amorphous homopolymers, provided that no other factors affect the copolymer microhardness. It is important to note that the additive law is applicable to blends of miscible polymer pairs. This is demonstrated for blends of poly(methy1 methacrylate) and poly(viny1idene fluoride) [140]. The situation with the dependence of the glass transition temperature on the composition of the L‘blends’’is rather similar. The empirical Fox equation [149] (Eq. (8)) is commonly used for the evaluation of Tgof polymer blends. This equation is recommended for mis-

466

F. J. BaltB CaUeja, S. Fakirov, H.G.Zachmann

cible blends -a detail which is often disregarded. Other equations account for other factors affecting Tgin addition to the mass fraction of the components, such as Couchman’s one [150] (Eq. (9)),recommended for random copolymers and miscible polymer blends, giving much better correlation of experimental observations than any other similar expression. This fact can be taken as evidence underscoring the importance of the entropy, rather than the volume, as the main factor controlling the onset of glass transition [155]. This equation was initially claimed to be derived purely from thermodynamic considerations, based on the concept of the continuity of the entropy of mixing at the glass transition temperature [155]. Dealing with blends of condensation polymers which are left to react to the stage of complete involvement of the starting homopolymers into copolymers, new factors affecting Tgappear. The case when the copolymers obtained are non-crystallisable is of particular interest. Such an amorphous system is very close to a blend of miscible polymers with respect to the number of components and phases. At the same time, there is an essential difference - the appearance of a new type of chemical bond. In addition to the effect of the chemical composition and structure of the repeating units on T,, which is the same in both cases, the generation of a new type of chemical link would affect the internal mobility of the groups in a single chain and thus would contribute to the changes in Tg.For this reason, equations accounting for the internal rotation around a single bond seem to be more appropriate for the evaluation of T, of amorphous copolymers obtained from a blend. A good example for such a relation is the expression ~561

T, = 262/mR + CI

where J2 is the cohesive energy density (CED), m is a parameter describing the internal mobility of the groups in a single chain, R is the gas constant and CIis a constant. It is noteworthy that CED provides an integrated measure of the strength of the secondary bonds in a compound. Materials with strong secondary bonds show high CED values. Thus, the above equation accounts for the two most important factors determining the value of T,: (i) the possibility of rotation around single bonds and (ii) of formation of secondary bonds with the surrounding atoms. In addition, it offers a link between the glass transition temperature and microhardness of polymer glasses since the cohesion energy is a basic factor in the determination of microhardness as well, as reported in [157]. By this common dependence of T, and H upon cohesion energy, the almost linear relationship between their depression of Lw and AT,, caused by the increasing amount of PET in the copolymers (Figure 15), can be explained.

Effect of Transreactions on Structure and Properties

467

4. Some practical aspects of the chemical interactions in

condensation polymers

4.1. Copolycondensates resulting f m m solid-state additional

condensation

Obviously, if the basic characteristics of the polycondensation process are taken into account, the prospect of using prehomopolymers to obtain block copolycondensates seems very attractive. Nevertheless, work done in this field is still very limited. The first investigations were those of Boye I1581 and Kibler et al. [159]. On the basis of their obervations, Lenz and Go [78] succeeded in preparing a block copolycondensate of poly (&/trans 1,4cyclohexylenedimethylene terephthalate). Solid-state polycondensation was conducted with crystallised, low molecular weight prepolymers below their melting points. The formation of block copolyesters was confirmed by examining the melting behaviour and solubility of the final products. Both melting point and solubility were found to be similar to those of the unequivocal block copolymers, formed by the coupling reaction of homopolymers [78]. In [160], use is made of solid-state additional condensation to obtain a PET copolyester containing diethylene glycol as a comonomer unit (ranging from 1 to 15 mol %). The type of comonomer sequence in the chain is not ascertained, but it is likely that the copolyester obtained is a random one. This presumprion rests on the fact that the prepolymers are prepared from the three comonomers in a common solvent. 4.2. Copolycondensates resulting

from transreactions i n the melt

An approach differing from the classical one - the synthesis of copolycondensates from an initial comonomer mixture [161] - leads to the formation of copolycondensates by interchange reactions in the melt. Exchange reactions are mainly studied in mixtures of polycondensates of identical chemical compositions, but differing in average molecular weight [lo81 or in polycondensate systems having one monomer in common [78,79,162-1671. In an extended study, Devaux et al. investigated various aspects of exchange reactions between PC and PBT in the molten state [5,96-981. These authors were the first to consider the statistical aspects of an exchange reaction between two polycondensates of differing chemical structure [96]. Such a reaction generates a four-component (Al, A2, B1, and B2) copolycondensate. In a PC-PBT system, A1 and A2 are the butylene and bisphenol A groups, respectively, with B1 and B2 being terephthalate and carbonate units, respectively. Spectral data provide a very coherent picture of the PC-PBT copolyester structure, which results from an exchange of aliphatic ester (A1B1) and aromatic carbonate (A2B2) sequences, simultaneously with formation of bisphenol A terephthalate (A1B2) and butylene carbonate

468

F. J. Balti Calleja, S. Fakirov, H.G . Zachmann

(AzB1) in equimolar quantities. The statistical analysis (degree of randomness and average sequence length) clearly shows that the exchange reaction gradually leads to the formation of a copolyester with a random distribution (statistical copolyester). The appearance of an insoluble fraction at the onset of the exchange reaction is due to the initial formation of a copolyester with a long PBT sequence, insoluble in dichloromethane. With the progress of the reaction, the PBT sequence is shortened and a fully soluble product is gradually formed. The conclusion from the statistical analysis is that a sequence length of less than ca. 7 units is required for the formation of a completely soluble copolyester. More details on this subject can be found in Chapter 3. In a very recent study [ll], Berti et al. report that copolyesters are prepared by reacting PET and PC. 4.3. Compatibilisation by means of interchange reactions

The block, as well as the random, copolymers formed via interchain reactions (additional condensation and/or transreactions) are expected to exhibit enhanced mutual miscibility, compared to the starting nonreacted components [1,8,10,13-201. Moreover, the transreacted chains probably also facilitate and accelerate compatibilisation and interaction between nonreacted chains [l].Therefore, it is important to control the interchange reactions in order to obtain a consistent product [21]. Interchange reactions in blends of condensation polymers depend strongly on their initial compatibility and on the blending conditions [I]. These include temperature, duration of mixing, preparation method, viscosity match, and the presence of catalysts (1511 or inhibitors [168-1721. Different conclusions are drawn in the literature concerning the blend properties, simply because the blends studied are prepared in different ways [l].The effect of transreactions on compatibility in blends of condensation polymers is also discussed in [l]and in Chapter 10. 4.4. Prepamtion of laminates from films of condensation

polymers by means of interchain reactions

A generalisation pertaining to oriented films is that the higher the degree of orientation in one direction, to impart high levels of mechanical properties such as tensile strength, the lower is the level of properties in the transverse direction. In uniaxially oriented films, both tensile strength and elongation at break in the direction transverse to orientation can be so low that during ordinary processing or manipulation the film fibrillates, becoming useless for some applications which can otherwise employ the high tensile strength. Biaxially oriented films with a good balance of properties can also be produced, but in an effort to make biaxially oriented films of very high tensile strengths by segmental drawing, the high strength attained in the first-direction draw is diminished by the second-direction draw. Thus, attempts have been made to fabricate cross-lapped, interfacially bonded

Effect of Transreactions on Structure and Properties

469

laminar structures. Chemical healing can be accelerated by spreading catalysts of transreactions or additional condensation on the contact surfaces [go].

Lamination through chemical healing makes it possible to overcome a peculiarity of liquid-crystalline polymers - their molecular orientation depends strongly on the article thickness. For this reason, the thickness increase leads to a loss of the unique mechanical properties of this class of polymers. Thicker articles can be obtained using thin (15-200pm) highly oriented foils by lamination (parallel or cross-plied) and bonding through chemical healing [91]. Chemical healing is not limited to the bonding of films; other structures with surfaces suitable for the required contact can also be subjected to healing. The method is applicable to strips, bands, billets, tubes, rods, and pipes. The fact that self-bonding occurs without fusion means that the initial orientation is retained. The tendency toward relaxation and disorientation is reduced to a minimum by the fact that the heat treatment is conducted on pressed cross-plied films. 4.5. Upgrading of molecular weight of condensation polymers by

means of additional condensation in the solid state

Production of synthetic polyester and polyamide fibers requires polymers with a molecular weight of about 20000. Such polymers have an optimum viscosity for spinning, as well as mechanical properties typical of textile fibres. At the same time, this viscosity is too low for processing by injection moulding or extrusion. For injection molding, a molecular weight of about 30 000-40 000 is required an even higher one for extrusion - about 60 000. Increased molecular weight not only leads to higher melt viscosity but also improves the mechanical properties. Condensation polymers of higher or ultrahigh molecular weight can be considered as engineering plastics [3]. Linear polycondensates (polyamidesand polyesters) of molecular weight above 30 000 are hard to obtain by melt polycondensation. Therefore, their production is conducted in a two-stage process: (i) melt polycondensation and (ii) additional solid-state polycondensation. The second stage may in fact be carried out with polymers in the form of chips, fibres, bristles, films, fabrics, or finished products (clothes, etc.). In industrial practice, additional solid-state polycondensation may be carried out in three ways: (i) under vacuum, (ii) in a stream of inert gas (stationary bed), and (iii) in a fluidised bed. Each of these techniques has its advantages and drawbacks. In (ii) and (iii), a continuous process can be maintained, which is not possible with the vacuum method. However, the vacuum technique has other important advantages. Along with additional solid-state polycondensation, some other treatments may be conducted such as dyeing, plasticising, adding substances that improve processing or operational properties, etc.

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

470

4.6. Recycling of condensation polymers by means of interchain

reactions

4.6.1. Application of additional solid-state condensation for

mycling purposes

It is well known that in the production of polyamide and polyester fibres, wastes amount to 10% of the total production [173]. In the production and processing of polyester films, wastes are some 30% or more [174]. These wastes are often hydrolysed so as to recover the monomers, or burnt for energy production. Wastes from PA 6 fibre production are mostly used in injection moulding, although, owing to their low molecular weight, they have poor processing and operational properties [175]. These disadvantages can be avoided by carrying out an additional solid-state condensation, resulting in a molecular weight increase. As a first consequence, such a treatment increases the melt viscosity and thus improves the processing properties; this is of particular interest for PET wastes [176]. It is demonstrated with PA6 wastes (secondary PA6) that, after regranulation and heat treatment in vacuum for different periods of time (10 to 50h), the molecular weight grows from 18000 to 40000 [177]. Figure 16 shows the effect of molecular weight on the stress at break, Ub,for secondary PA 6. It is obvious that the molecular weight increase from 19 000 to 30 000 leads to a 55% increase in the stress at break. A further molecular weight increase to 40000 does not affect the tensile strength. It is worth noting that the maximal strength value of the secondary material (660 kg/cm2) is significantly higher than that of the fresh material used for fibre production

18

26

-

34

42

Mn.lO-s

Figure 16. Dependence of the stress at break C b on the molecular weight A?fn achieved by additional solid-state condensation under vacuum at 210°C for various durations from PA 6 processing wastes [loll

471

Effect of Transreactions on Structure and Properties (bb = 550 kg/cm2).

The molecular weight increase of the wastes of linear condensation polymers greatly affects not only their mechanical characteristics but also the natural and artificial ageing behaviour of these materials. This is of particular significance for PA6, because of the drop in its tensile characteristics to one-half their initial values following one-year natural ageing [178]. Table 5 summarises the data on stress at break and Brinnell hardness for secondary PA 6 samples of different molecular weights before and after artificial ageing. It is seen that, after ageing, the stress at break drops by some 20% for all samples. Nevertheless, the Ub values for high molecular weight samples are comparable to those for the initial material that has not been subjected to ageing (cTb = 550kg/cm2), which is used in fibre production. The higher hardness is due to additional crystallisation during the artificial ageing cycles. Table 5. Mechanical characteristics before and after artificial ageing of PA6 prccessing wastes with various molecular weights as a result of additional solid-state condensation [179] Property before and after ageing Stress at break (kg/cm2) Brinnell hardness (kdcm')

Average molecular weight (Mn) 18 000

24 000

29 200

33 400

38 800

Before After Before After Before After Before After Before After 420

290

560

500

640

530

660

570

650

520

15

18

17

21

18

21

18

21

19

21

A most important characteristic of engineering plastics is the shape and dimensional stability during utilisation. Investigations are accordingly conducted with untreated wastes of PA6 and PET, with wastes of increased molecular weight (ca. 30 000), and with glass-reinforced (30%)samples. As anticipated, measurements after artificial ageing show the highest dimension distortions to the typical of low molecular weight PA6. Much more stable are the higher molecular weight samples; the reinforced ones show almost no deviation [175]. Similar behaviour is observed with these same samples with regard to the effect of duration of artificial ageing on elongation up to the inception of neck-building; in this case, PET samples show much less deformability than PA 6 [175). The conclusion from these few examples is that additional solid-state condensation is a useful tool for the recycling of wastes from the production of polyamide and polyester films and fibres. By increasing the molecular weight of these wastes, materials are obtained that have processing, me-

472

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

chanical, and operational properties typical of engineering plastics. Simplicity is an important characteristic of this recycling method. There is no need to use, remove, or regenerate solvents, or to add other reactants. There is no need to melt, stir, and regranulate the polymer. Furthermore, during additional solid-state condensation there is almost no by-product formation, or, if any are formed, they are innocuous (water or glycols) and in minute quantities. It is for this last reason that additional solid-state condensation raises no environmental problems. Therefore, this recycling belongs to the category of wasteless technologies and is very promising. 4.6.2. Application of transreactions in condensation polymers for

recycling purposes

In Section 2.4 of this chapter, the preparation of microfibrillar reinforced composites, MFCs, is described briefly as an illustration of the occurrence of transreactions on the phase boundary of polymer blends. It is emphasised that these reactions not only contribute to the compatibilisation of the blend components, but also determine the characteristics of the composite matrix. This new type of composite exhibits mechanical parameters (Young's modulus and tensile strength) which are higher by 30-50% than the weightaverage values of the components, and comparable to those of glass fibrereinforced composites having the same matrix. In addition to the synergistic effect on mechanical properties, MFCs offer further important advantages: reinforcement of polymer by polymer, no mineral additives, reduction in weight, improved mechanical integrity, easy processing, no need for extra compatibilisers, control of the crystallisation ability of the matrix and of its solubility, complete recycling, and repetition of the processing. Another important advantage of MFCs arises from the fact that wastes of condensation polymers can be used as raw materials for the preparation of MFCs since these wastes preserve their reactivity. If the wastes are characterised by decreased molecular weights due to degradation processes during the use of the articles, upgrading in molecular weight can be undertaken (see Section 2.4). Suitable partners for recycling via MFCs preparation could be, e.g., PET beverage bottles and carpets with PA6 as a major component. 4.7. Improvements of the finished-product properties

Interchain reactions not only contribute to the improvement of the material's mechanical characteristics, but also to the properties of the finished product, ie., resistance to chemicals and stability to heat, to environmental factors, etc. Of practical importance are the operational properties of textile fabrics for technical applications (filters, sacks in galvanoplastics, separators in the production of batteries). For instance, the fabric in this last case consists of 60% glass fibres and 40% polyester fibres, both highly resistant to sulfuric acid. (The test requires the loss of weight to be less than

Effect of Transreactions on Structure and Properties

473

15% [180], after a 10-day immersion in a 5% sodium bichromate solution in sulfuric acid of 1.30g/cm3 density, at 80°C). Pure (100%) commercial P E T fabric, subjected to the same test, shows weight losses of 50% and more [181]. As Table 6 shows, when this same commercial fabric is subjected t o additional solid-state condensation, the solubility of the material drops drastically. The data also indicate that it is possible to regulate solubility by extending the heat treatment at a high temperature. As additional solid-state condensation is conducted at temperatures close to the melting point, after this treatment the polymer achieves maximum crystallinity. This results in a heightened resistance to chemicals, and greater stability of the shape of the finished product at elevated temperatures. Resistance to ageing and to climate factors also improves after such treatment. In this way, one may improve the main operational properties of these materials when used as filters, battery separators, galvanoplastic sacks, etc. The introduction in industry of additional solid-state condensation permits heat treatment of ready-made finished articles at an early stage of their production (fibres, fabrics, or films). This is possible by either the continuous or discontinuous process. Finally, the molecular weight markedly affects the polymer finished product properties. Products prepared from polymers with higher molecular weights are notable for their higher resistance to chemicals, heat, and ageing factors. When used, such finished products maintain more stable shapes and dimensions [1751. Table 6. Weight losses after extreme test conditions [180]and stress at break before and after additional solid-state condensation of commercial PET fabrics 11811

Property

Starting material

After post-condensation

Weight losses (%)

23 33 41 53 50 49 52 54

3.5 6-13' 4-T 6-17' 64 58 61 56

Stress at break (kg/cm2)

a The weight losses values depend on the duration of the additional solid-state condensation

474

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

Acknowledgement S. Fakirov gratefully acknowledges the financial support from the Deutsche Forschungsgemeinschaft (DFG FR 675/21-1).

References 1. R. S. Porter, L. H. Wang, Polymer 33, 2019 (1992) 2. A. M. Kotliar, J. Am. Chem. Soc., Polym. Prepr. 21, 180 (1980) 3. S. Fakirov, “Solid State Reactions an Linear Polywndensates”, in: Solid State

Behavior of Linear Polyesters and Polyamideq edited by J . M. Schultz, S. Fakirov, Prentice Hall,Englewood Cliffs, NJ 1990 4. F. Pilati, E. Marianucci, C. Berti, J. Appl. Polym. Sci. 30, 1267 (1985) 5. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20,

1901 (1982) 6. P. Godard, J. M. Dekoninck, V. Devlesaver, J. Devaux, J. Polym. Sci., Polym. Chem. Ed. 24, 3315 (1986) 7. G. Montaudo, M. S. Montaudo, E. Scamporrino, D. Vitalini, Macromolecules 25, 5099 (1992) 8. C. Berti, V. Bonora, F. Pilati, M. Fiorini, Makromol. Chem. 193,1665 (1992) 9. J. Devaux, P. Godard, J. P. Mercier, Polym. Eng. Sci. 22, 229 (1982) 10. A. M. Kotliar, J. Polym. Sci., Macromol. Rev. 16, 367 (1981) 11. C. Berti, V. Bonora, F. Pilati, M. Fiorini, Makromol. Chem. 193,1679 (1992) 12. S. Hanna, A. H. Windle, Polymer 29, 207 (1988) 13. M. Kimura, G. Salee, R. S. Porter, J. Appl. Polym. Sci. 29, 1629 (1984) 14. A. Golovoy, M. F. Cheung, K. R. Carduner, M. J. Rokosz, Polym. Eng. Sci. 29, 1226 (1989) 15. R. L. Miller, R. V. Brooks, J. E. Briddell, Polym. Eng. Sci. 30,59 (1990) 16. L. H. Wang, Z. Huang, T. Hong, R. S. Porter, J. Macromol. Sci.-Phys. 29, 155 (1990) 17. W. Klee, D. 0. Hummel, Makromol. Chem., Macromol. Symp. 52,141 (1991) 18. J. I. Eguiazabal, G. Ucar, M. Cortazar, J. J. Iruin, Polymer 27, 2013 (1986) 19. J. I. Eguiazabal, M. Cortazar, J. J. Iruin, J. Appl. Polym. Sci. 42,489 (1991) 20. W. A. Smith, J. W. Barlow, D. R. Paul, J. Appl. Polym. Sci. 26,4233 (1981) 21. M. E. Stewart, A. J. Cox, D. M. Naylor, Polymer34, 4060 (1992) 22. S. Fakirov, Polym. Commun. 26, 137 (1985) 23. S. Fakirov, J. Polym. Sci., Polym. Phys. Ed. 22, 2095 (1984) 24. S. Fakirov, Makmmol. Chem. 185, 1607 (1984) 25. S. Fakirov, N. Avramova, J. Polym. Sci., Polym. Phys. Ed. 25, 1331 (1987) 26. N. Avramova, S. Fakirov, Angeur. Makromol. Chem. 140, 63 (1986) 27. S. Fakirov, “Structure and Properties of Polymer#, Sofia Press, Sofia 1985 28. B. Wunderlich, Macromolecular Physics, Academic Press, New York 1973, vol. 1 29. S. Fakirov, “Alternative Appmaches to Highly Oriented Polyesters and

Polyamides With Improved Mechanical Properties”, in: Oriented Polymer Materials, edited by S . Fakirov, Huthig & Wepf Verlag, Zug 1996 30. S. Fakirov, “Polywndensation Polymers and Solid State Reactions”, Nauka i Izkustvo, Sofia 1983 (in Bulgarian)

Effect of Transreactions on Structure and Properties

475

31. S. Fakirov, E. W. Fischer, ZUPAC Znt. Symp. on Macromolecules, Aberdeen 1974, Abstr., p. 85 32. S. Fakirov, M. Evstatiev, Polymer 31,431 (1990) 33. S. Fakirov, D. Stahl, Angeur. Mukromol. Chem. 102,117 (1982) 34. S. Fakirov, Colloid Polym. Sci. 256, 115 (1978) 35. S. Fakirov, I. Seganov, Vysokomol.Soedin. A23, 766 (1981) (in Russian) 36. M. Evstatiev, I. Seganov, S. Fakirov, Acta Polym. 39,191 (1988) 37. H. Tanaka, T. Nishi, Phys. Rev. A39, 783 (1989) 38. Y. Li, M. Stein, B. J. Jungnickel, Colloid Polym. Sci. 269, 772 (1991) 39. G . C. Alfonso, T. P. Russell, Macromolecules 19, 1143 (1986) 40. H. Frensch, B. J. Jungnickel, Colloid Polym. Sci. 267, 16 (1989) 41. E. Martuscelli, Polym. Eng. Sci. 24,563 (1984) 42. D. C. Wahrmund, D. R.Paul, J. W. Barlow, J. Appl. Polym. Sci. 22,2155 (1978) 43. T. R.Nassar, D. R. Paul, J. W. Barlow, J. Appl. Polym. Sci. 23,85 (1979) 44. T. Suzuki, H. Tanaka, T. Nishi, Polymer 30,1287 (1989) 45. T. 0. Ahn, S. Lee,H. M. Jeong, Eur. Polym. J. 25,95 (1989) 46. C. A. Murthy, S. M. Aharoni, Polymer 28,2171 (1987) 47. J. Kugler, J. W. Gilmer, D. Wiswe, H. G. Zachmann, K. Hahn,E. W. Fischer, Macromolecules 20, 1116 (1987) 48. K. P. McAlea, J. M. Schultz, K. H. Gardner, G. D. Wignall, Polymer 27, 1581 (1986) 49. E. Andresen, 13. G. Zachmann, Colloid Polym. Sci. 272,1352 (1994) 50. D. Chen, H. G. Zachmann, Polymer 32,1611 (1991) 51. C. Santa Cruz, F. J. BaltS Calleja, H. G. Zachmann,D. Chen, J. Muter. Sci. 27,2161 (1992) 52. E. Andresen, PhD Thesis, University of Hamburg, 1994 53. L. A. Utracki, ‘‘Polymer Alloys and Blends”, Hanser, Munich 1989, p. 64 54. B. Darlix, B. Monasse, P. Montmotonnet, Polym. Testing 6,107 (1986) 55. Y. Deslandes, E. AlvaRosa, F. Brisse, T. Meneghini, J. Muter. Sci. 26,2769 (1991) 56. F. J. Baltd Calleja, Adv. Polym. Sci. 66,117 (1985) 57. F. J. Baltd Calleja, Zknds in Polym. Sci. 2,419 (1994) 58. F. J. BaltB Calleja, C. Santa Cruz, T. Asano, J. Polym. Sci., Polym. Phys. Ed. 31,557 (1993) 59. C. Santa Cruz, F. J. Baltd Calleja, T. Asano, I. M. Ward, Phil. Mag. A68, 209 (1993) 60. F. J. BaltS Calleja, C. Santa Cruz, C. Sawatari, T. Asano, Macromolecules 23,5352 (1990) 61. C. Santa Cruz, F. J. Baltd Calleja, H. G. Zachmann, N. Stribeck, T. Asano, J. Polym. Sci., Polym. Phys. Ed. 29,819 (1991) 62. F. J. Baltd Calleja, C. Santa Cruz, T. Asano, J. Polym. Sci., Polym. Phys. Ed. 31,557 (1993) 63. D. R. Rueda, A. Varkalis, A. Viksne, F. J. Balt.6 Calleja, H. G. Zachmann, Macromol. Chem. Phys. 195,3869 (1994) 64. D. R. Rueda, A. Varkalis, A. Viksne, F. J. Balt.6 Calleja, H. G. Zachmann, J. Polym. Sci., Polym. Phys. Ed. 33,1653 (1995) 65. U. Eisele, “Zntroduction to Polymer Physics”, Springer-Verlag, Berlin 1990, ch. 7

476

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

66. B. Wunderlich, “MacromolecularPhysics”, Academic Press, New York 1976, vol. 2, ch. VI 67. J. M. Schultz, “Polymer Materials Science“, Prentice Hall, Englewood Cliffs, N J 1974, ch. 9 68. F. J. BaltB Calleja, L. Giri, H. G. Zachmann, J. Muter. Sci. 32,1117 (1997) 69. M. T. Connor, M. C. Garcia, D. R. Rueda, F. J. Baltri Calleja, J. Muter. Sci. 32,000 (1997) 70. B. Wunderlich, “Chemical Reactions Involving the Backbone Chain of Macrumolecular Crystals”, in: Reactions of Polymers, edited by Moore, Riedel, Boston, MA 1973 71. A. Casale, R. S. Porter, “Polymer Stress Reactaod’, Academic Press, New York 1978 (vol. l),1979 (vol. 2) 72. B. Wunderlich, ‘LMacrumolecularPhysics”, Academic Press, New York 1977, vol. 3 73. F. H. Winslow, W. Mateyek, J. Am. Chem. SOC.,Polym. Prepr. 7,540 (1966) 74. G. Wegner, Angew. Makromol. Chem. 58/59, 37 (1977) 75. A. Miyagi, B. Wunderlich, J. Polym. Sci., Polym. Phys. Ed. 10,2073 (1972) 76. A. Miyagi, B. Wunderlich, J. Polym. Sci., Polym. Phys. Ed. 10,2085 (1972) 77. W. Griehl, J. Zarate, Plastverarbeiter 18,891 (1967) 78. R. W. Lenz,S. Go, J. Polym. Sci., Polym. Chem. Ed. 11,2927 (1973) 79. R. W. Lenz, S. Go, J. Polym. Sci., Polym. Chem. Ed. 12, 1 (1974) 80. R. W. Lenz,E. Martin, A. N. Schuler, J. Polym. Sci., Polym. Chem. Ed. 11, 2265 (1973) 81. R. W. Lenz, K. Ohata, J. Funt, J. Polym. Sci., Polym. Chem. Ed. 11, 2273 (1973) 82. R. W. Lenz, R. A. Awl, W. R. Miller, E. H. Pryde, J. Polym. Sci. 8A-1, 429 (1970) 83. R. G. Crystal, J. Polym. Sci. 7A-2, 1356 (1969) 84. R. G. Crystal, J. Polym. Sci. 8A-2, 1755 (1970) 85. J. Zimmerman, J. Polym. Sci. 62,995 (1964) 86. M. I. Simonova, E. M. Aizenshtein, V. V. Shevchenko, Khim. Volokna 1973, No.5, 14 (in Russian) 87. A. I. Koretskaja, G. I. Kudrjavtsev, A. A. Konkin, Vysokomol. Soedin. 6, 434 (1964) (in Russian) 88. US 4161578 (1979), Inv.: J. H. Dale 89. S. Voyutskii, Autoadhesion and Adhesion of High Polymers, WileyInterscience, New York 1963 90. US 4715919 (1987), University of Delaware, Inv.: S. Fakirov, J. M. Schultz, Chem. Abstr. 108,168805~(1988) 91. US 4902369 (1990), University of Delaware, Inv.: N. Avramova, S. Fakirov, Chem. Abstr. 112,218405~(1990) 92. N. Avramova, S. Fakirov, Angew. Makromol. Chem. 179, 1 (1990) 93. US 4895612 (1990), University of Delaware, Inv.: N. Avramova, S. Fakirov, J. M. Schultz, Chem. Abstr. 111,59131g (1989) 94. US 4980237 (1990), University of Delaware, Inv.: N. Avramova, S. Fakirov,

J. M. Schultz

95. N. Avramova, Polymer 34, 1904 (1993) 96. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1875 (1982)

Effect of Transreactions on Structure and Properties

477

97. J. Devaux, P. Godard, R. Touillaux, J. M. Dereppe, J. Polym. Sci., Polym. Phys. Ed. 20, 1881 (1982) 98. J. Devaux, P. Godard, J. P. Mercier, J. Polym. Sci., Polym. Phys. Ed. 20, 1895 (1982) 99. S . Fakirov, N. Avramova, Acta Polym. 33,271 (1982) 100. G. Georgiev, D. Simov, Khim. Ind. (Sofia) 52,16 (1980) (in Bulgarian) 101. S. Fakirov, N. Avramova, Plasticheskie Massi (Moskow)No. 4, 54 (1979) (in Russian) 102. M. Droscher, F. G. Schmidt, Polym. Bull. 4, 261 (1981) 103. J. Devaux, P. Godard, J. P. Mercier, Makromol. Chem. 179,2201 (1978) 104. Ger. 2543631 (1976), Inv.: N. Saiki, Y. Takeuchi, H. Saiki, I. Iwakuni 105. P. J. Flory, “Principles of Polymer Chemistry”, Cornell University Press, Ithaca 1953 106. S. Fakirov, Vysokomol.Soedin. B-27,653 (1985) (in Russian) 107. T. Arakawa, F. Nagatoshi, N. Arai, J. Polym. Sci. 7A-2, 1461 (1969) 108. T. Arakawa, F. Nagatoshi, N. Arai, J. Polym. Scz., Polym. Lett. Ed. 6,513 (1968) 109. E. W. Fischer, H. Goddar, G. F. Schmidt, Makromol. Chem. 118,144 (1968) 110. E. W. Fischer, S. Fakirov, J. Mater. Sci. 11, 1041 (1977) 111. K. Slusallek, H. G. Zachmann, Kolloid 2. 2. Polym. 251,865 (1973) 112. S. Fakirov, M. Evstatiev, Adv. Mater. 6,395 (1994) 113. M. Fischer, in: The Interfacial Interactions in Polymeric Composites, edited by G. Akovali, Kluwer, Dordrecht 1992, p. 415 114. R. Holsti-Miettinen, J. Seppaelae, 0. T. Ikkala, Polym. Eng. Sci. 32, 868 (1992) 115. A. R. Padwa, Polym. Eng. Sci. 32,1703 (1992) 116. H. Yoon, Y. Feng, Y. Qiu, C. C. Han, J. Polym. Sci., Polym. Phys. Ed. 32, 1485 (1994) 117. S. F. Edwards, Proc. Phys. SOC.92,9 (1967) 118. P. G. de Gennes, J. Chem. Phys. 55, 572 (1971) 119. G. K. Hoeschele, Chimia 28,544 (1974) 120. I. M. Ward, “Mechanical Properties of Solid Polymers”, Interscience, New York 1971 121. E. E. Gull,N. N. Kuznetsov, “Structure and Mechanical Properties of Polymers, Khimia, Moscow 1966 (in Russian) 122. H. H. Kausch, ‘LPoZymerh c t u r e ” , Springer, Berlin 1978 123. H. Kaufer, “Arbeiten mit Kunststoflen”, Springer Verlag, Berlin 1978 124. K. Friedrich, S. Fakirov, J. Mater. Sci. 20,2807 (1985) 125. S. Fakirov, M. Evstatiev, Plaste Kautschuk 31,351 (1984) 126. J. H. Dumbleton, T. Murayama, Kolloid 2. 2. Polym. 220,41 (1967) 127. R. J. Samuels, J. Polym. Sci. 10A-2, 781 (1972) 128. J. H. Dumbleton, Polymer 10,539 (1969) 129. W. H. Cobbs, R. L. Burton, J. Polym. Sci. 10,275 (1953) 130. H. G. Kilian, H. Halboth, E. Jenkel, Kolloid 2.2. Polym. 172,166 (1960) 131. H. G. Zachmann, H. A. Stuart, Makromol. Chem. 41,131, 148 (1960) 132. K. Dimov, Chr. Betschev. R. Lasarova, Iv. Alexandrova, Faserforsch. Textiltechn. 21,492 (1970) 133. R. Bonart, Ullmans Encyclopaedze der technischen Chemie, Band IT, Verlag Chemie, Weinheim 1976

478

F. J. Baltd Calleja, S. Fakirov, H. G. Zachmann

134. W.J. Dulmage, A. L. Geddes, J. Polym. Sci. 31,499 (1958) 135. H.A. Stuart, Ann. N.Y. Acad. Sci. 83, 3 (1959) 136. R. Bonart, Kolloid Z. Z. Polym. 199, 136 (1964); 210, 16 (1966); 213, 1 (1966);231,438 (1969) 137. J. L.Koenig, M. J. Hannon, J. Macromol. Sci.-Phys. B1, 119 (1967) 138. D. C. Prevorsek, J. P. Sibilia, J. Macromol. Sci.-Phys. B5, 617 (1971) 139. K. Slusallek, H.G. Zachmann,Angew. Makromol. Chem. 53, 35 (1976) 140. J. Martinez-Salazar, F. J. BalG Calleja, J. Mater. Sci. Lett. 4,324 (1985) 141. F. J. Baltd Calleja, L. Giri, Z. Roslaniec, T. Ezquerra, J. Macromol. Sci.Phys. B36, 335 (1997) 142. F. Ania, J. Martinez-Salazar, F. J. Balta: Calleja, J. Mater. Sci. 24, 2934 (1989) 143. F.J. Baltd Calleja, C. Santa Cruz, D. Chen, H. G. Zachmann, Polymer 32, 2252 (1991) 144. S. Buchner, D.Chen, R. Gehrke, H. G. Zachmann, Mol. Cryst. Liq. Cryst. 155,357 (1988) 145. B. Martin, J. M. Pereiia, J. M. Pastor, J. A. de Saja, J. Mater. Sci. Lett. 5, 1027 (1986) 146. D. Chen, PhD Thesis, University of Hamburg, 1990 147. F.J. BaltB Calleja, H. G. Kilian, Colloid Polym. Sci. 266,29 (1988) 148. F.J.BaltB Calleja, L. Giri, T. A. Ezquerra, S. Fakirov, Z. Roslaniec, J. Macromol. Sci.-Phys. B36, 655 (1997) 149. T. G. Fox, Bull. Am. Phys. SOC.1, 123 (1956) 150. P.R. Couchman, Polym. Eng. Sci. 24, 135 (1984) 151. L. H. Wang, M. Lu, X. Yang, R. S. Porter, J. Macromol. Sci.-Phys. B29, 171 (1990) 152. G.v. d. Velden, G. Kolfschoten-Smitsamans, A. Veermans, Polym. Commun. 28,169 (1987) 153. S . Fakirov, M.Sarkissova, Z. Denchev, Macromol. Chem. Phys. 197, 2837 (1996) 154. W. Zheng, Z. Wang, Z. Qi, Polym. Int. 34, 301 (1994) 155. R. J. Roe, “Glasstransition”, in: Encyclopedia of Polymer Science and Engineering, edited by H. F. Mark, N. M. Bikales, C. G. Overberger, G. Menges, Wiley, New York, vol. 7,p. 531 156. U.W.Gedde, “Polymer Physics”, Chapman and Hall, London 1995,p. 78 157. J. Martinez-Salazar, J. Garcia, F. J. Balta: Calleja, Polym. Commun. 26,57 (1985) 158. C. A. Boye, J. Polym. Sci. 55, 275 (1961) 159. US 3117950 (1964), Inv.: C. J. Kibler et al. 160. M.Droscher, Makromol. Chem. 181,789 (1980) 161. J. M. Smith, C. J. Kibler, B. J. Sublett, J. Polym. Sci. 4A-1, 1851 (1966) 162. P. Kresse, Faserforsch. Textiltechn. 11,353 (1960) 163. T. Kiyotsukuri, H.Kashiwabara, R. Immure, J. Chem. SOC.Japan, Ind. Chem. Sect. 69,1812 (1966) 164. K.Yoda, Makromol. Chem. 136,311 (1970) 165. M. Murano, R. Yamadera, Polym. J . 2, 8 (1971) 166. S.Chans, M. F. Seu, J. Appl. Polym. Sci. 28,3289 (1983) 167. J. Budin, J. Vanicek, Proc. First Eur. Symp. on Thennal Analysis, Heyden 1976,p. 168

Effect of Transreactions on Structure and Properties

479

168. S. G. Hovenkamp, J. Polym. Sci., Polym. Chem. Ed. 9, 3617 (1971) 169. M. F. Cheung, A. Golovoy, R. 0. Carter, H. van Oene, Ind. Eng. Chem. Res. 28, 476 (1989) 170. K. R. Carduner, R. 0. Carter, M. Cheung, A, Golovoy, H. van Oene, J. Appl. Polym. Sci. 40, 963 (1990) 171. M. F. Cheung, K. R. Carduner, A. Golovoy, H. van Oene, J. Appl. Polym. Sci. 40, 977 (1990) 172. US 3752866 (1973), Triber Industries, Inv.: M. L. Doerr, Chem. Abstr. 80, 4773i (1974) 173. Z. A. Rogowin, “Basics of Synthetic Fibers Chemistry and Technologf, vol. 2, Khimia, Moscow 1965 (in Russian) 174. K. Kopetz, “Recycling zlon Kunststoffen” , Chemische Werke Huls 1976 175. G. Ivanov, R. Treneva, S. Fakirov, Plate Kautschuk 29, 93 (1982) 176. S. Fakirov, G . Evstatiev, I. Mladenov, Inf. Bull. Khim. Prom. No.6,25 (1977)

(in Russian)

177. C. Y . Cha, J. Am. Chem. Soc., Polym. Prepr. 6, 84 (1965) 178. K. Ilieva, PhD Thesis, Centre for Plastics Application in Machine Building, Sofia 1980 (in Bulgarian) 179. S. Fakirov, Kunststofle 74, 218 (1984) 180. Bulgarian State Norms (BDS) 13055/75 181. S. Fakirov, I. Svetoslavova, Inf. Bull. Khim. Prom. No. 3, 67 (1979) (in

Russian)

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Author index Aerdts, 57 Al-Hadad, 47 Andresen, 442 Arrhenius, 145, 156, 394, 408, 419 Arrighi, 202 Avrami, 439, 441, 442 Backson, 25, 381, 418 Baur, 312 Bazobal, 24 Bekmirzaeva, 47 Benoit, 374, 376 Bernouli, 19, 131, 146, 147, 150, 174, 332 Berti, 84, 155, 239, 341, 468 Beste, 54, 55, 270, 291 Biswas, 200 Blackwell, 200, 201, 205 Boltzmann, 6, 8, 13 Boye, 467 Bragg, 382 Brill, 302, 303, 306 Brinnell, 471 Brown, 412 Carothers, 125 Casassa, 135, 136, 374 Cheung, 227, 257, 415 Cheng, 354 Choi, 418 Couchman, 466 Cox, 415 Crystal, 445 Debye, 323 de Gennes, 452 DeMeuse, 398, 401 Denchev, 156

Devaux, 19, 57, 227, 282, 283, 285, 287, 331, 338, 346, 365, 379, 380, 417, 447, 467 Droscher, 447 Economy, 349, 394 Eduards, 452 Eguiazabal, 424 Eksner, 37 Ellis, 8, 274 Escala, 418 Fakirov, 55, 71, 155, 273, 351, 354,449 Fernandez-Berridi, 156 Fiorini, 229, 330 Fisher, 354, 449 Flory, 37, 54, 125, 162, 224, 312, 319 Fortunato, 156 Fourier, 14, 16, 281, 395, 399 Fox, 465 Fradet, 148 Fries, 62 Frunze, 53, 270 Gallot, 48 Gauss-Lorentz, 400, 407 Geil, 205 Go, 445, 467 Godard, 84, 239 Golovoy, 340, 415 Guo, 47, 156 Ha, 48, 418 Halek, 150 Han,452 Hanna, 321 Heino, 422 Hoffmann, 311 Hong, 156, 422

482 Houtz, 54,55, 270, 291 Hwang, 422 Izoda, 48 Jacobson, 191 Jacques, 419 J&e, 398,401 Jo, 416,421 Kibler, 467 Kil, 421 Kim, 418 Kimura, 321,342,365, 415 Koenig, 285 Kokkala, 48 Kollodge, 47, 92 Kopinke, 49 Korshak, 37,40,53,270 Kotliar, 55, 225 Kricheldorf, 282,320,379,381 Kugler, 202,351,357 Kumar, 47 Larmor, 6,9, 13, 15 Lee, 421 Lei, 41 Lens, 273,348,349,394,444,445,467 Lewis, 48 Lin, 421 Linko, 47 Ma, 49 MarBchal, 148 Markoff, 176 Matthew, 47 McCullagh, 395 Mehrota, 147, 148 Mercier, 84 Miley, 338 Miller, 55,270 Miyagi, 444 Mondragon, 227,230,420 Montaudo, 46,49, 71,84,85,424 Moore, 374 Muhlenbach, 348,349 Murano, 19, 128, 134 Murff, 341

Author index Nassar, 354,355 Nazabal, 227,230,424 Ogata, 35,55 Ootoshi, 46 Otton, 41 Overhauser, 22,28 Pakula, 349 Paul, 83,84,422 Pilati, 151, 330, 338 Pillon, 423 Planck, 4 Porter, 83,84, 92,338, 342,349,355, 365,414,421,429,450 Potter, 349,354 Shuster, 338 Slusallek, 455 Schmidt, 447 Stamm, 349 Stein, 418 Stockmayer, 191 Takeda, 422 Tattum, 417 Tijama, 45 Utracki, 415 Verhoeven, 227 Verma, 225 Voyutskii, 446 Wang, 83,329, 414 Weeks, 311 Wegner, 443 Wei, 421 Wichterle, 37 Windle, 200, 206 Winslow, 443 Winter, 205 Wunderlich, 430,443,444,449, 450 Xanthos, 422 Yamadera, 17, 19, 128, 134,379 Yee, 421 Yoon, 205,206, 416 Yu, 418 Zachmann, 449,455 Zeeman, 7 Zimmermann, 304

Transreactions in Condensation Polymers Stoyho Fahirov copyright 0 WILEY-VCH Verlag GrnbH. 1999

Subject index

Blending (Mixing) Melt-, 83-85, 90, 92, 115, 269, 271 Crystallisation after, 287 Molecular structure after, 289 Morphology after, 302 Reactive, 79, 80, 115 Blends, polymer Definition, 219 PC/PBT, 84, 93, 95, 97 PC/PET, 83, 86, 93, 95, 97, 101, 102, 113 Reactions in, 85 Acidolysis, 86, 224 Alcoholysis, 86, 224 Amidolysis, 224 Aminolysis, 224 Effect of time, 110 Exchange, 87, 117 Intermolecular, 88 Intramolecular, 87 Phenolysis, 87 Side, 90 Technique for, 101 Catalyst, 89 Activity, 102-105 Of exchange reactions, 90 Residual, 102 Chromatography, Size-exclusion (SEC), 108, 109, 262 Coefficient, Of thermal expansion, 210 Compatibilisation, 223, 317, 411, 450452 By interchain (trans-) reactions, 412, 469 Non-reactive, 412

Of blends, 411 Of polyester/polyester blends, 416 Reactive, 412 Self-, 271, 272 Condensation, Additional, 429 For recycling, 470 Upgrading of MW, 469 Condensation polymers, Definition, 125 Copolymers, Block, 320, 334, 355, 357, 361 Random, 320, 334, 355, 357, 361, 368 Degradation, Selective, 97, 98 Thermal, 90, 91 Differential scanning chromatography (DSC), 250-262, 334, 340 Diffusion, Chemical, 210, 214 Chemically released, 452 Physical, 210, 213 Distribution, Block length, 351, 352, 379 Molecular weight, 125, 329 Equation of, Avrami, 439 Couchmann, 463 Fox, 461, 462, 465 Mark-Howink, 377 Exchange reactions, 159 Catalysed, 167, 170, 171, 175 Compatibilization through, 189 Copolymer composition after, 165 During m e l t - d i g , 161, 174

Subject index Inner-inner, 161-166, Intermolecular, 161 Intramolecular, 161 Mechanisms of, 161, 163 Modelling of, 189 Outer-inner, 161, 163, 166, 172 Thermally activated, 160 Healing, Chemical, 446 In crosslinked polyamides, 449 Heterochemical, 448 Homochemical, 445 Physical, 446 With coupling agents, 448 Inhibitors, 91 Interchain reactions, 429 Copolycondensates by, 467 Effect of structure, 429, 433, 444 On crystallisation, 436 On mechanical properties, 453 In interphases, 445 In PET/PEN blends, 434438 On interfaces, 445 Effect on microhardness, 439-449, 455 Of semisolid blends, 448 Recycling by, 470 Liquid crystalline polymers, 196, 209 Vectra, 196 Xydar, 196, 204 Microscopy, Transition electron (TEM), 309-311 Modulus, Elastic, 453 Loss, 434 Polymer alloys, 219 Polymer compatibility, 219 Polymer miscibility, 219, 220 Quenchers, 91, 92 Randomisation, 201 Of Xydar, 204 Rate of, 204 React ions Acidolysis, 38

Alcoholysis, 34 Between amine and amide groups, 50 Block-restoration, 347, 348 Catalysts, 35, 36, 39 13C NMR in, 22 Ester groups, 32, 33 Esterolysis, 40 For hyperbranched polymers, 61 'H NMR in, 17 Initiators, 43 Interchange (exchange), 1,2,32,331 Randomisation, 331, 332 Recent developments, 41 Sequence analysis, 42, 43 With Si-0 bonds, 57 With urethane and urea groups, 66 Scattering, Neutron, 89 Small-angle neutron (SANS), 202 Small-angle X-ray (SAXS), 307 Time resolved, 307-309 Wide-angle X-ray (WAXS), 302305, 322 X-ray, By aperiodic chains, 395 Non-periodic (aperiodic), 391, 392, 395 Reaction kinetics by, 395, 402 Sequence ordering, 205, 319 Crystallisation-induced, 347 Melting-induced, 331 Of immiscible partners, 333 Of miscible partners, 342 Miscibility-induced, 364 Models of, 369, 370 Solubility, Selective, 95-97, 102, 117, 119 Spectrometry, Fast atom bombardment mass, 174, 177, 180 Spectroscopy, 93 13C NMR, 22, 185, 201, 203, 239, 320, 321, 332, 381 Accuracy, 30 Block length by, 25, 28, 29 Block length distribution by, 30

Subject index 'H NMR, 3, 139, 239, 332, 342, 379 Basics, 3 Chemical shift, 9 Chemical shift correlated spectrum (COSY), 20, 21 Experimental, 12 For interchange reactions, 17 Fourier transformation, 14, 15 High resolution, 16 Relaxation, 7 Saturation, 7 Spin-spin coupling, 10 Infrared (IR), 93-95, 139-152, 155, 233, 324, 332 NMR, 95, 104, 110-114 14N NMR, 31 15N NMR, 30 31PNMR, 244-249 "Si NMR, 30 Transamidation, 270, 271 Block length by, 281-287 Distribution, 281, 282 Characterisation by GEC, 293 Degree of randomness, 285 Effect of end-groups, 301 Influence of blend composition, 278 Of extrusion temperature, 277 Of extrusion time, 276 Of screw rotational speed, 275 Molecular characterisation by, 281 Percentage - determination of, 284,285 Transcondensation, 126 Transesterification, Activation energy, 394 Blocky structure, 396 By SANS, 394 Effect on viscosity, 398 Kinetics by X-ray, 405 Monomer sequence, 391, 393,394,395 Random, by X-ray, 402 Transition, Of Brill, 302, 303 Polymorphic in polyamides, 302

485 Transreactions, 125, 126, 319 Average sequence length, 132 Catalysis of, 145, 227 Catalysts, 143, 145 Consequences in blends, 267 Copolycondensates from, 127 Degree of randomness, 128,129,133 During processing, In LCP blends, 421 In PA/PA blends, 422 In PC/PA blends, 424 In PC/PAr blends, 417 In PC/PBT blends, 417 In PC/PET blends, 416 In PET/PA blends, 423 In PET/PAr blends, 420 In PET/PBT blends, 418 In PET/PEN blends, 419 In phenoxy blends, 420 Techniquesfor characterisation, 414 Dyad analysis, 130 Inhibition of, 153, 225 In PC/PBT, 136 Kinetics of, 133, 139 Mechanism of, 143, 146 Rate constant of, 134, 142 Rate of, 135, 142, 146 Ratio of, 134, 136, 137 Revealed by DMTA, 324, 359 By DSC, 322, 323, 334 By IR, 325 By light microscopy, 328 By SEC, 330 By SEM, 327 By TLC, 329 By WAXS, 322, 323 Theoretical, 127 Thermodynamics, 369 Triad analysis, 131 With low molecular weight compound, 134