Advanced Routes for Polymer Toughening

Polymer Science Library 10 ADVANCED ROUTES FOR POLYMER TOUGHENING Polymer Science Library Edited by A.D. Jenkins Uni...

Author: E. Martuscelli | P. Musto | G. Ragosta

12 downloads 943 Views 23MB 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

Polymer Science Library 10

ADVANCED ROUTES FOR POLYMER TOUGHENING

Polymer Science Library Edited by A.D. Jenkins University of Sussex, The School of Molecular Sciences, Falmer, Brighton BN1 9QJ, England 1. K. Murakami and K. Ono, Chemorheology of Polymers 2. M. Bohdaneck~ and J. KovdL Viscosity of Polymer Solutions 3. J. Wypych, Polyvinyl Chloride Degradation 4. J. Wypych, Polyvinyl Chloride Stabilization 5. P. Kratochvfl, Classical Light Scattering from Polymer Solutions 6. J. Bartoh and E. Borsig, Complexes in Free-Radical Polymerization 7. Yu.S. Lipatov, Colloid Chemistry of Polymers 8. F.J. Baltd-Calleja and C.G. Vonk, X-Ray Scattering of Synthetic Polymers 9. K. Kamide, Thermodynamics of Polymer Solutions

Polymer Science Library 10

ADVANCED ROUTES FOR POLYMER TOUGHENING

Edited by

E. Martuscelli P. Musto G. Ragosta National Research Council of Baly, Institute of Research and Technology of Plastic Materials, Arco Felice (Naples), Italy

1996 ELSEVIER Amsterdam

- Lausanne

- New

York - Oxford

- Shannon

- Tokyo

ELSEVIER SCIENCE B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam, The Netherlands

ISBN: 0-444-81960-6 (vol. 10) ISBN: 0-444-41832-6 (Series) 9 1996 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. Special regulations for readers in the USA. This publication has been registered with the Copyright Clearance Center Inc.(CCC), 222 Rosewood Drive Danvers, MA 01923. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the U.S.A. All other copyright questions, including photocopying outside of the USA, should be referred to the copyfight owner, Elsevier Science B.V., unless otherwise specified. No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. This book is printed on acid-free paper. Printed in The Netherlands

CONTENTS Preface ...................................................................................................... vii

Introduction ................................................................................................

1

E. Martuscelli, P. Musto, G. Ragosta

P A R T I: T O U G H E N E D

THERMOSETS

Chapter 1. Epoxy Resins .......................................................................... 11

E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi

Chapter 2. Unsaturated Polyester Resins ................................................. 61

E. Martuscelli, P. Musto, G. Ragosta

Chapter 3. Thermosetting Polyimides .................................................... 121

E. Martuscelli, P. Musto, G. Ragosta

P A R T II: T O U G H E N E D

THERMOPLASTICS

Chapter 4. Nucleation Processes in Toughened Plastics ......................... 157

A. Galesh, Z. Barwzak, E. Martuscelli Chapter 5. Isotactic Polypropylene Based B l e n d s .................................. 243

L. D'Orazio, C. Mancarella, E. Martuscelli, G. Sticotti

Chapter 6. Polyamide 6 / E t h y l e n e - c o - V i n y l a c e t a t e B l e n d s ...................... 2 8 9 a Model System of Thermoplastic/Elastomer Pairs

L. D'Orazio, C. Mancarella, E. Martuscelli

vi

Chapter 7. Blends Polyamide 6/Functionalized Rubber . . . . . . . . . . . . . . . . . . . . . . . . .

335

R. Greco, M. Malinconico, E. Martuscelli, G. Ragosta

Chapter 8. PMMA/Rubber Blends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

439

P. Laurienzo, M. Malinconico, E. Martuscelli, G. Ragosta, M. G. Volpe

Chapter 9. Polycarbonate Toughening by ABS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

469

R. Greco

Chapter 10. Rubber Modification of Biodegradable Polymers .............. 527 M. Avella, B. Immirzi, M. Malinconico, E. Martuscelli, M. G. Volpe, M. Canetti, P. Sadocco, A. Seres

vii

PREFACE The problem of polymer toughening elnerged more than twenty years ago and since then has experienced a steady increase in interest aald in the number of researchers all over the world dedicating their efforts to its solution. Although many aspects remain to be clarified and many questions still wait for a definitive answer, nowadays toughened plastics hold a consistent and well defined position ha the polymer market and successfully withstand the competition of advanced engineering plastics whose R&D costs are often prohibitive. hi such a complex and rapidly evolving scenario, we were invited by Professor A. D. Jenkins, Editor in chief of the Polymer Science Library edited by Elsevier Science, to produce a book on this topic. At a first glance we were hltrigued by this opportunity but soon realized how extensive and comprehensive the existing literature is, especially when considering a number of highly appreciated and widely diffused books (see, for instance, the references cited in the introduction). Therefore we decided not to attempt a comprehensive coverage of all the aspects of the subject matter, but rather to concentrate the attention on a limited number of systems which may be regarded as typologies of toughened plastics. From an extensive treatment of these systems we felt it possible to develop general concepts whose importance we learnt to appreciate more mad more as our research activity in the field has proceeded. These concepts, which can never be overestimated in the design mid formulation of tough polylner blends, include the role of the interface in lnultieomponent systems, the chemical reactivity of the blend components, the mode mad state of dispersion of the second phase, the crystalli~fity mad crystallization conditions, the glass transition temperature etc. The present book therefore mainly reports on the long term activity of the Institute of Research and Teclmology of Plastic Materials in the area, which started in the early seventies with the chemical modification of ethylene-propylene copolymers to be used as tougheners of Polyamide-6. The hlteresthlg results obtained prompted us to investigate further such a system both from a fundamental

viii and from an applicative point of view, and from this initial and highly rewarding experience we gained an approach which we have never abandoned along the years, i.e. to make fundamental research on systems of high technological interest. In this respect an essential part of our activity has always been represented by our continuous contact with research teams operating in the industrial world who gave us numerous inputs aald faced us with the problems of the demanding market of innovative materials. We hope that the present contribution can give to the reader the feeling of how intellectually sthnulating aald challenging the field of polymer toughening can be, so as to encourage him to spend his research efforts in such an area. Finally we wish to express our deep gratitude to all the authors who have contributed to this volume and to the many others behind them, each of one can be found in the extended literature cited. A special thank goes to Mr. Giuseppe Narciso, who gave us an invaluable help in the hard task of editing the manuscript.

Ezio Martuscelli PeUegrino Musto Giuseppe Ragosta

National Research Council of Italy; Institute of Research and Teclmology of Plastic Materials Arco Felice (Naples), Italy.

INTRODUCTION

E. Martuscelli, P. Musto, G. Ragosta

National Research Council of Italy, Institute of Research and Technology of Plastic Materials, 80072 Arco Felice (Na), ITALY

Toughened polymers represent a large area of scientific and technological concern. In fact, with the gradual penetration of plastics in areas traditionally dominated by metals and ceramics, new polymeric materials, both thermoplastics and thermoset resins, have been developed for increasingly demanding applications. These materials are able to provide the right combination of lightness and mechanical performance over a wide range of temperatures. In mm~y instmaces a good balance between stiffness and tougheness is required but most of the as sinthesized polymers which exlfibit adequate rigidity are characterized by brittleness and low resistance to crack propagation. Therefore the increase of the intrinsic tougheness of otherwise brittle materials has stimulated a huge mnoum of research efforts both in the academic mid in the industrial world [ 1, 2, 3]. Two different approaches have emerged, both with their own potential and weakness. One is to sinthesize new omo- or copolymers, based on novel monomers as in the case of polycarbonates, polysulphones and polyether-ketones. The second approach consists in modifying existing polymers through the addition of a second polymeric component, a route generally referred to as "blending". Such a method presents the distinctive advantage of being, in general, more economically attractive, since the development of new sinthetic methods is a long and costly process [4].

An excellent example of the blending approach is provided by the rubber toughening in which a small amount of rubber, typically between 5 and 20 % by weight, is incorporated as a disperse phase into a rigid plastic matrix. The resulting blend is characterized by a considerably higher fracture tougheness than the parent polymer; there is an inevitable reduction in the modulus and tensile strength but these losses are far outweighed by the improvement in fracture tougheness. This approach has proven to be very successful and a wide variety of plastics toughened in this way are now commercially available. Among the best known examples are high impact polystyrene (HIPS), and polyvinylchloride (PVC); other plastics which have been toughened using this teclmology include polymethylmethacrylate (PMMA), polypropylene (PP), polycarbonate (PC), Nylons and, most recently, thermosetting resins such as epoxies, polyimides and unsaturated polyesters [5, 6]. In the case of rubber toughened thermoplastics an essential condition to achieve satisfactory results is that the rubbery phase must be finely and homogeneously dispersed within the matrix; furthermore the rubbery particles must be adequately bonded to the matrix. To this end the rubber must posses a solubility parameter sufficiently different from that of the matrix polymer to ensure a fine second phase dispersion but close enough to promote adequate adhesion of the particles to the matrix. Such stringent requirements strongly limit the choice of possible rubber tougheners for a given polymer matrix. To overcome this limitation the concept of compatibilizing agents in the form of block or graft copolymers has been developed and succesfully applied to a wide number of actual cases [7, 8]. Essentially, a suitable

block or graft copolymer whose segments are

chemically equivalent to the blend components is added. Atter the blending process such a copolymer locates preferentially at the interphase thus promoting a better dispersion of the second component and an improved adhesion among the phases. The behaviour of small amounts of compatibilizer in an immiscible blend has been described as that of a classic emulsifying agent, similar to the soap molecules at an oil-water interface [9]. The success of the use of block or graft copolymers (and in some instances also random copolymers) as compatibilizers accounts for many of

the large number of commercially available blends, e.g. HIPS and ABS. The renewed interest towards the reactive melt processing, i.e. reactive extrusion (REX) and reactive injection moulding (RIM), is also due to the recent achievements in the development of copolymer compatibilizers. Moreover, when the thermoplastic matrix is able to crystallize, further factors must be taken into account such as the structure and size of spherulites and lamellar crystals, the spherulite grow rate and the nucleation process, which are all affected by the presence of the dispersed sott phase [ 10-13]. The toughening of thermosetting resins poses different, yet equally challenging issues. Generally, a critical step towards the preparation of a toughened them~osetting blend is to start from a single-phase homogeneous reactants mixture prior to the curing process. For this reason the rubbers usually employed are low molecular weight liquids and are miscible enough to dissolve in the resin prepolymer. However the elastomeric phase separates out during the curing process, giving rise to fine and homogeneous dispersion of the second component in the resin matrix. The phase separation process is of paramount importance in these systems and depends on both kinetic and thermodynamic factors. The understanding of these factors and the ability to control them to obtain the desired morphology of the materials represent one of the main goals of the research efforts in this area. As for thermoplastics, also in the case of thermosetting blends, a certain degree of chemical interaction between the resin and the rubber modifier is required to improve the interfacial adhesion and hence to achieve an effective toughening. Such reactions generally involve the end-groups of the rubber modifier whose functionality is adjusted according to the chemical nature of the matrix resin. An example of an extensively investigated system of this type is provided by the blend composed by a bifunctional epoxy resin and a carboxyl-terminated acrylonitrile-butadiene copolymer (CTBN). This system is particularly versatile since the solubility of the CTBN modifier in the uncured epoxy resin can be adjusted by varying the acrylonitrile content in the copolymer. However the CTBN precipitates out from solution during curing and, at the same time, a chemical

reaction occurs among the carboxyl end-groups of the rubber and the oxirane rings of the matrix. The rubber toughening approach of thermosets fails when high temperature requirements becomes more and more critical. In fact the reduced themml stability of unsaturated, low molecular weight rubbers, renders such blends unsuitable for high temperature applications. Moreover the toughening efficiency of elastomeric modifiers gradually decreases as the cross-linking density of the matrix increases. This is because, as the thermoset resin is cross-linked more tightly, its capability to be plastically deformed is strongly reduced and the dispersed rubber particles are no longer able to induce energy dissipation processes in the matrix. Both of these issues are particularly relevant in the case of the latest, high teclmology termosets like tetrafunctional epoxies and bismaleimide resins which find application in the aerospace and in fl~e electronic industries. For these materials a novel approach has recently emerged, which consists in the fommlation of blends with tough, ductile and thermally stable engineering thermoplastics. Examples of thennoplastics successfully employed in such blend systems are provided by polycarbonates, polyetheresulphones, polyarilsulphones and polyefllerimides. Concurrently with the development of novel tougheners and of more sofisticated technologies to produce multicomponent polymer blends with balanced end-properties, a large amount of efforts has been spent to elucidate the mechanisms of fracture in these complex systems. This in an attempt to be able to control the very many factors which play a role in the fracture behaviour of toughened plastics. Alter more than twenty years of extensive research in this area we may say that we are still far from a complete understanding of the whole phenomenon but very significant advances have been achieved especially in the case of blends based on rubber modifiers. In fact it is now well established that rubber particles with low moduli act as stress concentrators in both thennoplastic and themloset resins, enhancing shear yielding and/or crazing in the matrix, dependhlg on its molecular architecture. In particular in the case of themlosets the

crazing mechanism does not operate, while one important process is the initiation and growth of multiple localized shear yield deformations in the matrix. In addition a cavitation process occurring either in the rubber particles or at the particle-matrix interface often plays a key role. Once formed, these voids grow and so dissipate energy; at the same time they lower the stress required to initiate shear yielding in the matrix thus promoting more extensive plastic deformation [ 14]. hi the present book the theme of polymer toughening is covered in its very diverse aspects. Each chapter deals with a particular material or class of materials and is aimed at giving an overview of the problems encountered and of the proposed solutions. In particular the first part is devoted to thermosetting resins with three chapters dealing with epoxy resins, unsaturated polyesters and bismaleimide resins. The toughening of highly cross-linked epoxies by blending with engineering thennoplastics is covered in Chapter 1. Two different approaches are presented, depending on the nature of the thermoplastic modifier. With bisphenol-A polycarbonate a reactive blending process was developed, by which polycarbonate chains were chemically h~corporated within the epoxy network. A non-reactive blending process was carried out when a polyetherimide was used as toughened. In both the cases satisfactory results were achieved. The second chapter describes the toughening of unsaturated polyester resins by reactive liquid rubbers. Commercially available elastomers were chemically modified in order to enhance their compatibility and reactivity towards the polyester resin. These modified rubbers were molecularly characterized and their chemical

interactions

with the

matrix

during

curing

were

investigated

spectroscopically. The morphology and the fracture properties were found to be strongly dependent on the type of rubber modifier and on the chemical nature of its end-groups. The third chapter on thermosets deals with the toughening of polyimides. For this class of materials the problem has been approached either by adding a themloplastic second component (a polyetherimide), or a reactive liquid rubber. Better results were achieved with the polyetherimide, despite a number of

processing problems encountered along the way. This confirms the reduced efficiency of rubber modifiers in the toughening of very highly cross-linked themaosets. The second part of the book, concerned with thermoplastics, is opened by a chapter on the nucleation processes occurring in toughened semicrystalline plastics. This is a subject of great theoretical and practical interest, and a comprehensive account is given of the most recent advances in the interpretation of such a complex phenomenon. The role played by file molecular and microstructural parameters of the components on the toughening of blends of isotactic polypropylene (iPP) and ethylene-propylene (EPR) copolymers is treated in detail in Chapter 5. Here is shown that the desired fmal properties can be imparted by an appropriate choice of molecular mass and molecular mass distribution, constitution and tacticity of the blend components. A further critical factor is represented by the crystallization conditions through which it is possible to optimize the mode and state of dispersion of the rubbery phase in the semicrystalline matrix. Chapters 6 and 7 both deal with blends based on polyamide-6 (PA6). The fonner treats a blend system whose minor components are ethylene-vinylacetate (EVA) copolymers. The latter chapter covers PA6 blends in which the modifier is a functionalized EPR copolymer. These systems have received a great deal of attention in our Institute and, historically, represent the starting point of our continuing involvement in the field of polylner toughening. Particularly interesting results were obtained with an EPR elastomer functionalized by hlsertion of succinic alfllydride groups along its backbone. Such a modified rubber was used either as interfacial agent in PA6/EPR blends or directly as toughener for the PA6 matrix. Two different routes were followed to prepare the blends: by melt mixing the components in a Brabender-like apparatus or concurrently with the hydrolytic polymerization of e-caprolactam. The results obtained by these two approaches are critically reviewed. In chapter 8 the problelns involved with the toughening of amorphous thermoplastic polymers such as polymethylmethacrylate (PMMA) mid polystyrene

(PS) are discussed. In particular a novel method for the PMMA toughening is presented, which appears simpler to be implemented for large scale applications then those already established. Such a method consists in the dissolution of the rubber modifier (EVA copolymers) in file acrylic monomer which is subsequently radically polymerized. Satisfactory results in terms of tougheness were obtained even with limited amounts of the second component in the blend. Chapter 9 deals with the toughening of polycarbonate (PC) by ABS copolymers. The chapter starts with a critical review of vast amount of literature data, stressing the points which deserve further investigation. Subsequently the results obtained by the author's research group on aspects such as processability, miscibility and thermal behaviour of these blend are presented and discussed. Finally the last chapter is concerned with a class of biomaterials like poly (13hydroxybutyrate) (PHB) and his copolymers produced by bacterial synthesis. These polymers are receiving increasing attention due to their biocompatibility and biodegradability and the improvement of their mechanical properties is becoming a problem of considerable technological interest. Recent results obtained using methods based on bulk or suspension pol3qnerization of the modifier are discussed. References

1.

C.B. Bucknall, "Toughened Plastics", Appl. Sci. Pub., London, 1977.

2.

A.J. Kinloch, R. J. Young, "Fracture Behaviour of Polymers", Appl. Sci. Pub., London, 1983.

3.

J. M. Margolis, "Advanced Thermoset Composites", Van Nostrand Reinhold Co., New York, N. Y., 1986.

4.

M . J . Folkes, P. S. Hope, "Polymer Blends and Alloys", Blackie Academic & Professional, London, 1993.

5.

C. K. Riew, A. J. Kinloch, "Toughened Plastics I: Science and Engineering" Advances in Chemistry Series, 233, ACS, Washington, D.C., 1993.

~

C. K. Riew, "Rubber Toughened Plastics", Advances in Chemistry Series, 222, ACS, Washington, DC, 1989.

.

E. Martuscelli, R. Palumbo, M. Kryszewski, "Polymer Blends", Vol. I, Plenum Press, New York, 1980.

.

A. Galeski, E. Martuscelli, M. Kryszewski, "Polymer Blends", Vol. II, Plenum Press, New York, 1984.

.

D.R. Paul, "Polymer Blends ", Vol. 2, D.R. Paul and S. Newman Eds., Academic Press, London, 1978.

10. E. Martuscelli, "Rubber Modification of Polymers: Phase Structure, Crystallization, Processing and Properties" in "Thermoplastic Elastomers from Rubber-Plastic Blends", S. K. De, A. K. Bhowmick Eds., Ellis Horwood, New York, 1990. 11. E. Martuscelli, "Structure and Properties of Polypropylene-Elastomer Blends" hi "Polypropylene: Structure, Blends and Composites", J. Karger-Kocsis Ed. Chapman and Hall, London, 1995. 12. Z. Bartzak, E. Martuscelli, A. Galeski, "Primary Spherulite Nucleation in Polypropylene-based Blends and Copolymers" in "Polypropylene: Structure, Blends and Composites", J. Karger-Kocsis Ed. Chapman and Hall, London, 1995. 13. E.

Martuscelli, "Relationships Between Morphology, Structure,

Composition and Properties in Isotactic Polypropylene Based Blends" in "Polymer Blends and Mixtures", D. J. Walsh, J. S. Higgins, A. Macolmachie, NATO ASI Series, 1984. 14. A. C. Roulin-Moloney, "Fractography and Failure Mechanisms of Polymers and Composites", Elsevier Appl. Sci., London, 1988.

PART 1 TOUGHENED T H E R M O S E T S

This Page Intentionally Left Blank

11 CHAPTER 1

EPOXY RESINS E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi National Research Council of Italy, Institute of Research and Teclmology of Plastic Materials, 80072 Arco Felice (Na) ITALY.

1. Introduction

Recently epoxy resins with high crosslinking density have received considerable attention owing to their high temperature performances [ 1-3]. These high technology materials find applications in the aerospace industry as well as in other structural applications where the stiffness, the creep resistance for extended periods of time mad the thermal stability are essential [4]. The chemical formulas of the commercially available multifunctional epoxies are reported below:

/ok

A /~/c.=--CH---CH=

CH2--OH--CH2~~N \

TGAP:TriglycidylepoxJdebasedon aminophenol

/o,,,

A

CH2"--CH--CH2\ ~ /~---~ /CH2~CH~CH 2 N---~( ) )---C H2----'~~ ) )'----N C~2?H--CH / ~ / \'~--J/ \CH2__CH__CH2

V

o

TGDDM:TetraglycidilepoxJdebasedon DDM

12

O

/ 0\

O

/\

O-CH2-CH-CH2

O-CH2-CH-CI- ~

/\

O-CH2-CH-CH =

0

Novolac resins: Average value for n: DEN 438 = 1.6 DEN 439 = 1.8

Those resins are crosslinkcd with a number of different hardeners, the more common being reported below:

H2N---( (

) )-----.S.~

(~

))----NH 2

DDS: 4, 4' diarnino diphenyl sulfone

H 2 ~ C H 2 ~ N H 2 DDM: 4, 4' diarnino diphenyl methane

MNA: methylnadic anhydride

When those resins are fully cured, the resulting molecular network exhibits very high values of glass transition temperatures, Tg,(>250~

compared to the

conventional bifunctional epoxies. The stiffness of these materials is maintained up to temperatures approaching the Tg.However, as for other thermosets, the fracture

13 tougheness is extremely low, limiting the application of these materials to situations where the stress is relatively low and preferably static. The well developed approach of using a reactive liquid rubber as a toughening agent, which has meet with considerable success in the case of bifimctional epoxies [5-7], does not provide significant improvements for highly crosslinked resins. The reason why this is so has been dearly demonstrated by Yee and Pearson [8,9]. They tested a series of resins made from epoxies with different starting molecular weight. Their results, reported in Fig. 1, show that, in the absence of rubber, an increase in the molecular weight of the starting resin, which corresponds to a decrease of the cross linking density, has a very limited effect on the fracture toughness. On the contrary, when a rubber is added, the resins with higher cross-link density exhibit a very small improvement in toughness, which increases substantially by decreasing the cross-link density.

---

7

E #

6-

13.. w

I

I

I

9 Rubber modified o unmodified

5-

(/)

In a)

r

J~ o~ =3 0

I-(D

4-

3-

2-

IL_

::3

(J

1 -

(o

IL

u.

0 0

I

I

I

1000

2000

3000

Epoxy

Equivalent

4000

Weight

Figure 1. Fracture tougheness values for a series of DGEBA-DDS epoxies. (After Yee and Pearson).

14 From the point of view of the deformation mechanism this effect can bc cxplaine,d considering that, as the resin is cross-linked more tightly, its capability to bc plastically dcformexi is strongly rexiuce~ and the dispcrsexl robber particles arc no longer able to induce energy dissipation mechanisms such as shear yielding and/or cavitation [10-14]. Therefore the higher is the cross-linking density of the matrix, the lower its toughcnability. Recently a novel approach to toughening of epoxies has emerged, which has the potential of solving the above discussed problems. It consists in the formulation of blends with tough, ductile and thermally stable cngin~ring thermoplastics [15-21]. This altcmativc approach has the additional advantage that other desirable properties of the matrix, such as modulus, yield stress and glass transition temperature, arc not adversely affcctexi by the addition of the modifier.

80

I

i

3.5

I

o Yield Stress m

9 Modulus

m

70

t~

- 3.o

5

E t_

o

9 60-

r,,, .,,.,.==

t,,ol

v

"0

- 2.5

l

50-

4O 0

i

I

10

20

Rubber Content (phr}

Figure 2. Mechanical properties of a CTBN-modified epoxy.

2.0

30

,.,. Q "10

15 On the contrary Fig. 2 demonstrates a considerable decrease of these properties when a reactive liquid rubber is employed as second component [22]. The thermoplastic component can be either incorporated within the thermosetting network by means of a suitable reactive blending process or can be simply mixed with the matrix, giving rise, in most cases, to a completely phase separated system after the curing process. The thermoplastic modifiers cited in the literature are reported below:

c.3

I~

PC: Poly(bisphenoI-Acarbonate)

r,

o

!

] ~n

PEI:a thermoplasticPoly(etherimide)

PSU:Poly(sulfone)

PESU:Poly(ethersulfone)

16 The PSU, PESU based systems have been investigated by several authors [2334]. Some of these systems were found to phase separate after the curing process to give modifier's particles of about 0.5 pm in diameter. Unfortunately the resulting improvement in fracture toughness was modest. These results prompted to investigate other types of thermoplastics, and in the present chapter the systems DEN 438/PC and TGDDM/PEI are discussed in detail.

2. DEN 435/PC system 2.1 Molecular characterization and curing kinetics. A critical step toward the preparation of a successful thermosetting blend is to start from a single-phase homogeneous reactant mixture prior to the curing process. In our case this was achieved by a reactive blending process in which the PC was dissolved at high temperature (220~

for extended periods of time (typically 3 h) in

the uncured resin. At the end of this procedure, a clear, homogeneous solution was obtained for the whole composition range investigated. Then the temperature of the mixture was allowed to decrease to 80~ and the hardener MNA and the accelerator BDMA were added. The mixture was then poured into a glass mold, and the curing process was carried out at 120~ for 20h. A final posteuring step was performed for 5 h at 200~

After this protocol, a visually transparent sheet was recovered. The

various investigated blend compositions, together with their codes are reported in Table 1. The DEN 438/PC mixtures, prior to the addition of the hardener, were examined by differential scanning calorimetry (DSC) and by Fourier Transform Infrared Spectroscopy (FTIR) to insure the miscibility of the system. DSC measurements (Fig. 3) showed that all the mixtures investigated exhibit single glass transition temperatures, Tg, intermediate between those of the pure components, which, as expected, increase with increasing the PC content in the mixture.

17 Table 1.

Codes and Compositions of the Epoxy~PC Blends Code

Epoxy

MNA

Epoxy +

(%)

(%)

B0

49.5

50.5

100

B4

47.5

48.5

96

4

B8

45.0

47.0

92

8

B 10

44.5

45.5

90

10

B 12

43.5

44.5

88

12

B15

41.5

43.5

85

15

B20

39.6

40.4

80

20

MNA

PC

(%)

PC

_66134

9zla

-20

I

l

i

~,

20

60

100

140

T oC

I 180

Figure 3. DSC traces of uncured epoxy resin and of uncured epoxy~PC mixtures. Compositions as indicated.

18 In Fig. 4 the Tg values are reported as a function of composition. This behavior indicates that the high-temperature dissolution process produces a single-phase, homogeneous system over the entire composition range investigated.

310

I

I

I

300 -

lit

I---

290

280 0

I

I

I

0.1

0.2

0.3

0.4

PC Weight Fraction

Figure 4. The glass transiaon temperature, T~ as a fimction of composition for the Epoxy~PC mixtures investigated. Fig. 5 gives the FTIR transmission spectrum in the range 4000-400 cm -1 of a film of the uncured resin cast from CH2CI 2. Characteristic absorptions due to hydroxyl groups are observed in the 37003200 cm -1 range. The C-H stretching region (3200 - 2880 cm -1) is complex and highly overlapped while in the lower frequency range (1600-400 cm -1 ) a number of better resolved peaks are detected. For some of them a well-defined baseline can be identified, thus affording, after proper assignment, their use for analytical purposes.

19 The complexity of the resin's molecular structure makes complete assignment based on normal-coordinate analysis an extremely difficult task. Thus the peaks' assignments rely heavily on the work of Antoon [35] as well as on those of Bellamy [36] and of Colthup et al. [37].

.4-

_j 0-~

,

4000

'~

aooo

20'00 Wavenumbers

x6oo

( e r a - - 1)

Figure 5. FHR transmission spectrum o f uncured epoxy resin at 4000-400 cm -1. The spectrum was obtained on a film cast from CH2CI~

1.5

0 4000

30'00

2obo

Wavenumbers (era-- I)

~o'oo

Figure 6. bTIR transmission spectrum o f polycarbonate at 4000-400 cm -I. The spectrum was obtained on a thin film cast from CH2CI2.

20 In Fg. 6

is reported the transmission FTIR spectrum of PC: the carbonyl

stretclfing vibration produces a strong, well-resolved peak at 1775 cm-l, while the O-C-O stretching mode gives rise to a very intense and complex multiplet with maxima at 1227, 1193, and 1163 cm-1. In the spectrum of a 66/34 epoxy/PC mixture the C=O peak of PC is shifted by 3 cm-l with respect to the peak position detected in pure amorphous PC. This effect is evidenced in Fig. 7A where the spectra of the mixture and of pure PC are compared in the 1850-1700 cm-l range. Fig. 7B shows the result of spectral subtraction performed on the mixture spectrum using pure PC as the reference: the derivative-type feature characteristic of band shift in the reference is evident.

A / \,\

77J~-/~

\Ix.~_.~

,,Q

18bo Waven.umbene

175o (cm-- I )

~

"

tsoo Wavenumbera

t~5o (era-- I )

Figure 7. A) transmission spectra of an 80/20 epoxy~PC mixture o f pure PC m the 2000-1700 cm "1 range; B) subtraction spectrum in the 20001700 cm -I range obtained from the 80/20 epoxy/PC mixture using pure PC as reference.

Previous studies on the vibrational behaviour of PC in solution [38] demonstrated analogous shifts of the C=O peak in chlorinated solvents like CH2CI2 and CHCI3; in particular the shift was increased by increasing the protonating

21 properties of the solvent (-3.0 cm -I for CH2CI 2 and -4.5 cm-I for CHCI3 in a 1% w/w solution). The above effects were interpreted in terms of hydrogen bonding interactions between carbonyl oxygen of the polymer and hydrogen of the solvent. An analogous type of molecular interaction can be assumed to account for the effect observed in the epoxy/PC mixture. In particular, in this case hydrogen bonding may involve the PC carbonyls as proton acceptors and the methylene or methine groups bonded to the electron withdrawing oxygens, as proton donors. The fact that in the mixture spectrum we observe a single, highly symmetrical Vc_o absorption (see Fig. 7A) indicates that most of the PC carbonyls are involved in the above interactions. This in turn implies that the PC is molecularly dispersed in the epoxy matrix, thus confirlning the DSC results which indicated that the epoxy/PC mixture is a single-phase, miscible system. A deeper analysis of the uncured epoxy/PC mixture has been performed in order to investigate whether chemical interactions among the components occur in addition to the previously discussed physical interactions during the dissolution process.For this purpose the PC component was selectively extracted from the system: the mixture was first completely dissolved in a common solvent (CH2C12) and then the PC component was precipitated from solution using acetone, which is a solvent for the epoxy resin and a non solvent for PC. An epoxy/PC mixture of the same composition, but obtained by direct dissolution of the components in CH2CI2 was used as a reference system. The results of the quantitative extraction are summarized in Table 2. It is noted that, while in the reference system the PC is almost completely recovered (ca. 95%), in the mixture obtained by high-temperature dissolution of PC in the resin, only about 50% of PC can be extracted. Also of interest are the GPC data of the various recovered fractions reported in the last three columns of Table IV. It is noted that the PC fraction recovered from the reference system displays a slight increase in both M w and M~ while the molecular weight distribution is narrower with respect to the starting PC. This is likely due to the fact that, during acetone extraction, the lower molecular weight fractions of PC do not precipitate; in fact we recovered only 95% of the total extractable fraction.

22 Table

2.

Results of Quanatative Extraction on Epoxy~PCMixtures and GPC data m

Epoxy

PC

extracted

M . 10-4

M . 10-4

M , / M~

(%)

(%)

Pc (%)

0

100

-

1.2

3.1

2.6

73.5

26.5 b

95

2.4

3.9

1.6

73.5

26.5 c

47.1

1.3

1.9

1.5

aon the total amount of extractable PC. bmixture prepared by direct dissolution in CH2CI2. cmixture prepared by high-temperature dissolution. However the most relevant observation is that the PC fraction recovered from the epoxy/PC mixture obtained at high temperature, shows a marked decrease in the molecular weight moments, while the polydispersity remains scarcely affected. In particular both and M w are about half of the values detected in the reference system. This indicates that, during the high temperature dissolution process, the PC undergoes chain-scission reactions. Moreover the observation that M~ and M~ decrease by about half may be interpreted by assuming a random chain-scission process with no preferential sites along the PC backbone. These processes also account for the lower amount of extractable PC in the case of the epoxy/PC mixture. In Fig. 8 is reported the FTIR transmission spectrum of the PC fraction recovered from the epoxy/PC mixture. The measurement was performed on a thin film cast from CH2CI 2. Comparison with the pure PC spectrum (see insets) evidences the occurrence of a broad absorption centred at about 3528 cm -1, together with a lowintensity peak at 916 cm-1. The first characteristically broad contribution can be readily ascribed to stretching vibrations of self-associated hydroxyl groups. The band at 916 cm"l is attributed to a ring mode characteristic of epoxy groups. Further information on the VoH absorption can be gamed by eliminating the interference of hydrogen-bonding interactions, which, as is well known, produce extensive band broadening and make the spectral region poorly resolved [36,37].

23

0.58

~

0.40

' "

.Q

'

I

9 A I

'

4000

B

2500

i

O

.Q 0.23

0.05 !

i

3622.2

i'

i

1 ....

i"

1

2866.7 211 I. I 1355.6 Wavenumber ( c m -I )

600.00

Figure 8. FTIR transmission spectrum of the PC fraction recovered from the 80/20 epoxy~PC mixture. The spectrum was measured on a thin film cast from CH2CI~ The insets compare this spectrum with that of the pure PC in two different frequency ranges.

The spectra of dilute solutions of this PC fraction and of pure PC in a low polarity solvent (CH:CI2) are reported in Fig. 9, traces A and B, respectively. A sharp peak at 3583 cm-1, not present in the starting polymer, is evident in the PC recovered fractions; this band is partially overlapped with a PC absorption (a Vc_o overtone) at 3527 cm-1. The presence of a well defined singlet with no evidence of structuration indicates the formation of a unique type of hydroxyl groups on the PC chain during the high temperature dissolution process.

24

.2-

3583

(1)

9 0 DrJ

\ aiz7

.1-

36'00

3 5'00

Wavenumbers

34-'00

(cm-- 1 )

Figure 9. Spectra of PC fracaon (curve ,4) and of pure PC (curve B) in the frequency range 3750-3250 cm "1. The spectra were collected on dilute CH2CI2 soluaons. The formation on the PC fraction of the end-group structures of the type

/0\

/ 0\

OH ,

~

e~f-__~

PC chain --1.6

could account for the spectral features observed in the OH stretching region as well as for the presence of the epoxy ring mode at 916 c m -1. The formation of such end-group structures would imply the presence of hydroxyl-terminated PC chains, formed by chain scission of PC during the high-temperature treatment. The chain scission process of PC can occur either by hydrolysis of carbonate groups (scheme

25 A) or by reaction of carbonate groups with alcoholic functionalities present as impurities in the epoxy resin (scheme B).

SCHEME A

O chain scission, decarboxylation

----.---~ ~

H

+I ~ ~ V v ~ ' L II

+ CO2

II

SCHEME B

ci-I~ T" "J~"~

OH OH I

I

o II

~

CH3 T

O /\ .O-CH2-CH-CH2

/ O\ O-CH2-CH-CH2

H3 -

1.6

-1.6

l

e aO,ox

on

26 /\o C1"!2

/',o

OH ,

~

e~~/F-~

I~-~CH2

+ CO2

-

1.6

In either case, the OH-terminated PC chains (II) can further react with epoxy groups forming epoxy terminated PC chains O /\

O-C~-CH-C~ |+

0 I\

o I\

-CH-CH2

O-C~-CH-C~

H2 -

1.6

To obtain further support for the proposed reaction scheme we have performed quantitative FTIR analysis of the groups involved in the reaction (PC carbonyls and epoxy groups). In particular the epoxy functionalities were determined on chloroform solutions of the 64/36 epoxy/PC mixture using the epoxy ring mode at 916 cmq as an analytical band. To eliminate the interference of the solvent absorptions in the region of interest, spectral subtraction of the solvent was performed. It was found that, at~er the high temperature dissolution process, the content of epoxy groups in the mixture decreased by 7.0 mol % with respect to the starting value. It seems noteworthy that when the pure resin was subjected to the same themml treatment, no reduction in the epoxy group content was detected. An analogous approach was used to detenmne the carbonyl group content in the mixture. In this case the solvent used was CH2CI 2. After the thermal treatment, we found a carbonyl group reduction of 7.8 mol %. Both the epoxy and carbonyl group reductions are consistent with the proposed reaction schemes. From the above analysis it was found that for 100 g of reactive mixture 1.5 x 10.2 mol of carbonyl groups was consumed. On the other hand for the same amount of reactive mixture (100 g), the decrease of epoxy groups was found to be 3.2 x 10.2 mol.

27 The fact that the number of epoxy groups consumed is about twice the number of carbonyl groups seems to indicate that PC chain-seission occurs preferentially through the reaction scheme A. In any ease the formation of structures of type I is extremely important in the subsequent curing processes; in fact the epoxy functionalities at the end of the PC chains will take part in the crosslinking reactions, thus incorporating PC backbones within the epoxy network. After the high-temperature dissolution of PC in the epoxy matrix, the hardener MNA and the accelerator BDMA, were added in the desired proportions. A clear, homogeneous mixture was obtained.

.j o

1.. o m

0 - ~ 4000

-,,

SO'O0 - -

20"00

W a v e n u , u ~ b e r s (err'-- 1)

,600

Figure 10. FTIR transmission spectrum of epoxy~ardener mixture at room temperature prior to curing.

The curing process was carried out at 100~ in an environmental chamber directly mounted in the FTIR spectrometer to monitor the progress of the reaction in real time. The pure epoxy resin was subjected to the same thermal treatment as the blend, prior to the curing process.

28 In Fig. 10 is reported the FTIR transmission spectrum of the epoxy/MNA/ BDMA mixture at the beginning of the curing reaction. The MNA gives rise to several absorptions in the C-H stretching region which further complicate such a spectral range. A well resolved doublet at 1857 and 1781 cm-l appears, due to the symmetric and asynmaetric stretching vibrations of the anhydride carbonyls. Additional strong peaks due to MNA are detected at 1228 cm-l (Vc.o) and at 1083, 943, 929, 916, 899 cm-l (anhydride ring modes). Finally, at 798 crn-l a =CH out-ofplane deformation is found. [35-37]. During the curing process we observed the gradual decrease of the MNA absorptions and the concurrent intensity increase of the peaks arising from ester functionalities (see Fig. 11).

40'00

30'00

20'00

10'00

WAVENUMBERS (era-- 1)

Figure 11. FTIR transmission spectra o f the epoxy~hardener mixture at various curing times.

29 These peaks are located at 2963 cm -1 (v~,,,), 2863 cm -I (E.a,,), 1740

c m "1

(Vc_o, ester), 1454 cm -l (8c~,) , 1398 cm -1 (wcx,) 1267 cm -1 (Vc_o + Vc_c) and 1178, 1155, 1127 cm -l (vr

[36,37].

To follow the conversion of the anhydride groups we monitored the intensity decrease of the well-resolved absorption at 1857 cm "1. From the spectral data it is readily possible to calculate the fractional conversion, ct, of the anhydride groups. The tz versus time curves for the pure epoxy resin (curve A) and for the B 15 blend (curve B) are reported in Fig. 12.

12

9

/

o a

0.6

'

A

I

.....

I

~o -6o"oo o - o - o o o

-iy /

.~a~ 6" o

p

s 0

0

0

-

0.8

O.O'r 0

'

J .....

100

I

200

. . . .

800

time (mirO

Figure 12. Anhydride conversion versus time for pure epoxy (curve .4) and for the B15 blend (curve B). The continuous lines represent the zero order fit; the dotted lines represent the first order fit. Both curves exhibit analogous behaviour with an initiallineartrend followed by a plateau region. Moreover, for pure epoxy a higher initialslope as well as a higher value of the finalconversion arc observed in comparison with the blend.

30 Generally for thermosetting systems the overall reaction rate is a function of the temperature, concentration of reactants, reaction mechanism, and the local microviscosity [39]"

dtz

dt

= A e x p ( - E A/ R T ) f ( a ) f (

1"1,)

(1)

where A is the kinetic Arrhenius factor, E A is the activation energy, R is the molar gas constant, T is the temperature (K), f(ot) is a function of the reaction mechanism and the extent of conversion, and f(rlL) is a function of the local viscosity. Under isothermal conditions the reaction proceeds normally until the molecular weight increases to the extent that the glass transition approaches the cure temperature; f(rlL) becomes important only when the material is about to vitrify. In the absence of diffusion control, the general kinetic equation describing the process is:

do:

dt

where K=A exp

9= A e x p ( - E A/ R T ) f ( a ) f ( r l L )

(2)

(-EA/RT).The simplest expression for ritz) is f ( a ) = ( 1 - a)"

(3)

where n is the order of reaction. Integrating eq. (2) for n=0, n= 1 and n=2, we obtain, respectively:

a = Kt

(4)

- I n ( l - a) = Kt

(5)

6t

1-or

= Kt

(6)

31 Equation (4) fits the experimental data up to a conversion of 0.55 for the pure epoxy and up to 0.60 for the epoxy/PC blend (see Fig. 12). The analysis of the kinetic data according to eq. 5 is shown in Fig. 13. A good linear correlation between-log(l-c0 and t is observed for both the systems investigated. The slopes of the straight lines in Fig. 13 give the first order kinetic constants (K = 0.032 min-1 for the pure epoxy; K = 0.011 for the epoxy/PC blend) from which it is possible to calculate the relative conversion-time curves.

I

I

100

150

"

I v

0

0

0

60

200

time (rain)

Figure 13. -LogO-a) versus time for pure epoxy (curve A) and for the B15 blend (curve B). These curves are shown as dotted lines in Fig. 12. R is apparent that the first order kinetic expression fits the experimental data up to about full conversion for the pure epoxy system and up to 0.8 conversion for the epoxy/PC blend. The analysis

32 performed according to eq. (6) (second-order kinetics) did not yield the expected correlation. From the results of the kinetic analysis some conclusions can be drawn: 1. The data seem to indicate that the presence of PC in the system does not alter the overall reaction mechanism of the curing process. 2. Lower values of the calculated rate constants for both the consumption of anhydride groups and formation of ester groups were found in the epoxy/PC blend in comparison with the pure epoxy system.This marked effect could be ascribed to an increase in the bulk viscosity of the system due to the presence of the dissolved PC.Furthermore the PC may participate to the cross-linking process through its endgroup structures, thus perturbing the kinetics of the epoxy matrix as well as its overall molecular structure. 3. The final conversion of anhydride groups in the epoxy/PC blend is considerably lower than that obtained in the pure epoxy system. Moreover in the latter case a first-order rate equation describes the process over the whole conversion range, thus indicating that the curing reaction never becomes diffusion controlled. Conversely, for the blend, an increasing departure of the theoretical curve fxom the experimental data is observed starting from tx = 0.80. This in turn indicates that at tiffs conversion value the Tg of the system has reached the reaction temperature and the process becomes diffusion controlled. This effect could be ascribed to an increase in the Tg of the blend at any conversion with respect to the Tg of the epoxy matrix because of the presence of the PC component. Such a Tg increase as been demonstrated at zero conversion by DSC measurements.

2.2 Mechanical and Fracture Analysis Dynamic mechanical data for the unmodified resin, compression-molded polycarbonate, and a blend containing 20 % w/w of PC, are compared in Fig. 14. All the materials exhibit a primary tan5 relaxation peak corresponding to the glass transition flg). The Tg of the neat epoxy resin occurs at 170~ while the tan6 peak of PC is detected at 150~

The tan5 relaxation peak of the thermoplastic

component is considerably sharper then that of the epoxy resin. The peak temperature

33

in the blend coincides with that observed hi the neat resin; the peak shape remains highly symmetrical but is considerably broader. Although the Tg'S of the two blend components are quite close, the above observation seem to indicate that no phase separation has occurred during the curing process. In fact, in a phase separated system a less symmetrical band shape of the tan5 relaxation peak would be observed, due to the presence of a second-component peak at lower temperature. Similar results were obtained on the other compositions investigated. SkO

11~

0.6

eC

tO[

~ll E' ( P*I

T.n+

U

11.6

LO 8.2

i

Z.O

t0:"

A

7.0

?.8

74

1

7.0

lOO

so

log ['~

150

P, ~

zoo

1 . . o - o,. -" ,

zso

tZ

9o

to

0.4

lOO

1so

zoo

z1so

Tiin +~

c

i

7.O

!

1,

SO

Figure 14. Dynamic-mechanical spectra at 10 Hz for (A) neat resin; (B) pure polycarbonate; and (C) B20 blend. The fracture behavior of both the unmodified and the PC modified epoxy resins has been examined at low and high strain-rates. At low strain rate (2.5 x 10.3 s -l) two basic types of load-displacement curves are recorded, which correspond to two

34 different types of crack-growth behavior. The curve shown in Fig. 15A is observed for the plain epoxy resin and for blends containing less than 10% w/w of PC (blends B4 and B8). ''

1

I

30 load (N)

j

20

I0

displacement (mm) 0

0.2

I

I

0.4

0.6

Figure 15. Load versus displacement curves for plain epoxy resin (curve A) and BIO blend (curve B).

The load rises linearly with strain up to a maximum value where the crack propagates instantaneously causing a rapid drop in the load. The corresponding fracture surfaces exhibit little evidence of plastic deformation, as will be discussed in detail later. The load-displacement curve shown in Fig. 15B is representative of blends containing 10% w/w or more of PC (B 10, B 12, B 15, B20). Here the crack propagates intermittently in a stick-slip fashion. The load increases linearly up to a critical value and than the crack propagates in a stable manner until the stored elastic energy in the sample decreases to such an extent that crack arrest is allowed. Upon reloading the sample the process of crack-growth is iterated up to the complete failure of the sample. Examination of the corresponding fracture surfaces shows clear evidence of ductility.

35 The critical stress intensity factor K c, is determined from the load-displacement curves according to the equation:

K = crY~f~

(7)

where a is the nominal stress at the onset of crack propagation, a is the initial crack length, and Y is a calibration factor depending on the specimen geometry. For three-point bending specimens, Y is given by Brown and Srawely [40].

I

1.6

1,2

P)

I

I

-

E

Z

0,8

-

0,4

-

w

u

0

I

0

5

"

I

'

10

Blend Composition

I

15

20

PC)

Figure 16. Critical stress intensity factor, Kr at low strain rate as a function o f blend composition.

The values of K c are reported in Fig. 16 as a function of the mount of PC in the blend. It is noted that there is a significant increase of K c with increasing PC content;

36 an essentially linear correlation is found between K c and the amount of the modifier in the blend. The toughening effect of PC is more dramatically evidenced in Fig. 17 where G e data are reported. The G c values were calculated from the values of K c and of the elastic flexural modulus E according to the Irvin relation [41] for linear elastic fracture mechanics:

G =~

(8)

E

In terms of G c, the addition of 20% w/w of PC raises the toughness of the epoxy matrix by a factor of about 7.

600

500

-

400

-

E 300

-

U

200 -

/

J

100 -

0 0

I

I

I

5

10

15

Blend Composition (g) Figure 17. Critical strain energy release rate, G c at low strain rate, as a function o f blend composition.

20

37 In attempting to toughen a brittle polymer, the main goal is to increase the toughness without significantly compromising other important properties such as the elastic modulus. The E values reported in Fig. 18 as a ftmetion of blend composition clearly demonstrate that the improved toughness is achieved without sacrificing the stit~ess of the epoxy matrix. In fact, over the composition range investigated, the modulus shows a gradual but very limited decrease from 3.0 GPa in the neat resin to 2.8 GPa in the B20 blend.

4.5 4~

3.5

13.. r

..

.

I

1

......

I

.........

I

-

-

3.0 -~

w

uJ

2.5

-

2.0

-

1.5

-

1.0

i

0

5

. . . . . . . . . . .

I'

"

10

Blend Composition

I'

15

20

PC)

Figure 18. Elastic flexural modulus, E, as a function of blend composition. Fracture measurements were also carried out under impact conditions in order to evaluate the toughness of these materials under rapid loading. The Kc and G e values are reported as a function of blend composition in Figs. 19 and 20, respectively. In

38 this case the K e values were obtained as previously, using eq. 7, while the G e values were estimated by energy measurements according to the following equation:

U

(9)

G = BW@

where U is the fracture energy corrected for the kinetic energy contribution, B and W are the thickness and the width of the specimen respectively and ~ is a calibration factor which depends on the length of the notch and the size of the sample. Values of 9 were taken from Plati and Williams [42]. Apart from a decrease in the absolute values of Ke and G c the general behavior of the impact toughness parameters is analogous to that observed in the low-speed tests.

1.5

I

I

I

5

10

15

1.0-

0.5

0

20

Blend Composition (% PC) Figure ] 9. Critical stress intensity factor, K c, under impact conditions as a function of blend composition.

39 350

300

-

250

-

200

-

150

-

100

-

(M

E

--.j w

o

(.t

500i~ 0

/

I

I

I

I

5

10

15

20

Blend Composition (g)

Figure 20. Critical strain energy release rate, G• under impact conditions as a function o f blend composition.

In particular Kr and Gr increase by factors of about 2 and 5, respectively, compared with the values for the neat resin. A decrease in toughness on increasing the loading rate is a general phenomenon related to the reduced capacity for viscoelastic and plastic deformation which polymeric materials exhibit when the strain rate is enhaneexi. However, in our case it is interesting to note that, even if the strain rate is increased by about 5 orders of magnitude, the observed decrease in the toughness parameters is relatively modest (e.g., 40 % for the B20 composition). The considerable increase in toughness found in the PC-modified epoxy resins can be ascribed to an improved capability of localized plastic deformation of the epoxy/PC network. This enhanced capacity of deformation is due to the particular kind of molecular structure of the network developed during the curing process.

40 As previously mentioned during the high-temperature dissolution process, the PC component interacts chemically with the uncured epoxy resin with formation of epoxy end group. Such groups located at the ends of PC chains can take part in the curing reaction thus incorporating PC chains into the epoxy network. The resulting molecular structure can be sketched as:

( dissolution 220uC

~ 5hr=

curing postcuring

_

The PC molecular segments are longer and more flexible than the segments of a simple epoxy network and can be more easily deformed under loading. The fracture toughness data have been interpreted in terms of the morphological analysis of the fracture surfaces obtained on samples tested at low and high strain rates. It has been found in earlier investigations [43-46] that several characteristic features can be observed on fracture surfaces of epoxy resins, especially when stickslip propagation takes place. These features may fall into three main categories: an initiation region followed by crack arrest lines, a region of slow crack growth, and an area of rapid crack growth which covers the remaining surface of the sample. The scanning electron micrographs of the fracture surfaces of samples tested at low strain rate (Fig. 21) illustrate the above features. In particular, we note that, prior to the crack arrest line AB, all the samples exhibit a smooth and relatively featureless surface which can be associated with fast crack propagation,. Beyond the AB line, the pure epoxy (Fig. 21A) as well as the

41 blends B5 and BS, for which no stick-slip behavior was observed (micrographs not shown), display only the presence of fine markings extending from a restricted area.

Figure 21. SEM micrographs offracture surfaces obtained at low strain rate." A) pure epoxy resin; B) BIO blend; C) B15 blend; D) B20 blend. At high magnification (Fig. 22) these markings appear as wave crests and arise from adjacent sections of the crack front following paths at slightly different levels. On the other hand, when the crack propagation occurs by a stick-slip process, after the crack arrest line a well defined slow-growth region (area ABCD in Figs. 2 I B, 21C and 21D) is observed. The size of this area increases markedly with increasing PC content in the blend. A closer examination of this region at higher magnification

42 (see Fig. 22) reveals the occurrence of V-shaped features that result from events occurring during the arrest and re initiation of crack propagation.

Figure 22. High magnification SEM micrographs offracture surfaces obtained at low strain rate: .4)pure epoxy resin; B) B20 blend. Following the slow growth region, there is a transition to an unstable initiation region where the crack accelerates, giving rise to the formation of fine longitudinal lines, approximately parallel to the crack direction. Note that the crack direction is from left to right in the mierographs. The above observation account for the fracture toughness results and, at the same time, give a clear picture of the events occurring in a stick-slip process: after crack arrest a localized plastic zone develops at the crack tip upon loading, and the crack becomes blunted. This blunted crack grows slowly until sufficient strain energy is stored, through continued loading, to force the crack to propagate rapidly through file undeformed material.

43 Thus the slow growth region observed in figure 2 l b-d, defines the size of the plastic zone at the crack tip. Such a plastic zone is most likely generated via a shearyielding rather than a crazing mechanism. Indeed several authors [7, 9, 47] have concluded that crazing does not usually occur in epoxy materials. Therefore, it might be assumed that localized yielding at the crack tip with consequent notch blunting is the major source of energy dissipation during fracture in both the unmodified and the PC-modified epoxy resins. This mechanism is far more active in the B 10, B 15, and B20 blends because of the rather extensive presence of PC chains within the epoxy network. In fact as shown in Figs. 2 l b-d, an increase in size of the plastic zone is observed on increasing the PC content in the blends. Fig. 23 shows SEM mierographs of some of the investigated samples fractured under impact conditions.

Figure 23. SEM micrographs offracture surfaces obtained under impact conditions: A) pure epoxy resin; B) BIO blend; C) B20 blend.

44 Examination of these surfaces at low magnification did not reveal features characteristic of a stick-slip process because, at high strain rate, this type of crack propagation mechanism is completely suppressed. It is clear that, even at very high magnification (2500 x), no evidence of a dispersed second phase could be detected in the blends, thus confirming the dynamic-mechanical results which, for a B20 composition, indicated the occurrence of a single-phase, homogeneous system. Attempts to selectively etch the fracture surfaces with a solvent for PC (CH2CI 2) failed to reveal any feature distinctive of PC removal. The fracture data were used to calculate the degree of crack tip blunting occurring when these materials are fractured at low strain rate. In order to estimate such a parameter the stress distribution around a blunt crack has to be taken into account. It can be shown [48] that, for a crack under an applied stress of cro, the stress normal to the axis of the crack at a short distance r ahead to the crack tip is given by:

or= Cro~-~r (1 +l + p /p2/rr) "

(lO)

where p is the crack tip radius and a is the crack length. Assuming that the fracture occurs when a critical stress r is reached at a distance r = c, eq.10 becomes [48,49]:

c r ,~~ = (l+p~/2c) "~ cr~/2av (1 + p, / c)

(11)

The term cr 2 ~ ~ can be considered as the critical stress intensity factor Kle for a "sharp" crack, and cr~/no as the stress intensity factor K~ for a blunt crack. Hence eq. 11 may be rewritten as:

K~ = ( l + p , / 2 c ) ~'~ Kk (l+p/c)

(12)

45 This equation relates Kr to the radius of a blunt crack Pc; its validity may be checked by measuring the variation of K~ with pC. Direct evaluation of Pc is very difficult, especially for thermosetting systems. However, as reported by Kinloch and Williams [48], this problem may be circumvented assuming Pc to be equivalent to the crack opening displacement 8c:

(13)

where ay,t is the tensile yield stress and Sy is the yield strain. The above equation has been used to calculate 8~ or p~, and hence the degree of crack blunting. The resulting p~ values are reported in Fig. 23 as a function of blend composition. 15

I

I

,,

I

10 E U

o

I

0

5

I

10

Blend Composition

I

15

20

(~ PC)

Figure 24. Crack-np blunting, Pc, as a function of blend composition.

46 It is noted that the degree of crack tip blunting increases sharply starting from a PC content of 10% wt/wt. This behaviour further confirms that blunting takes place only when stick-slip propagation occurs. In terms of yield behavior, the appearance of such a type of crack-growth mechanism might be viewed as a consequence of the reduction of the yield stress, whiell enlmaces the ability of the material to plastically deform in the vicinity of the crack tip with a consequent increase in the fracture toughness parameters.

3. TGDDM/PEI system For this particular blend system the thermoplastic component was dissolved in a common solvent. In particular the PEI was dissolved in CH2CI2 and mixed with the epoxy resin at room temperature. After complete dissolution of the TGDDM, the solvent was distilled off at 60~ vacuum at 100~

The last traces of CH2CI2 were eliminated under

obtaining a clear, viscous solution. The DDS hardener was added

at 120~ under vigorous mechanical stimng up to complete dissolution. The mixture was poured in an open steel mould and degassed in a vacuum oven at 100~ for 5 h. Finally, the blend was cured for 16 h at 120~

2 h at 150~

2 h at 180~ and

postcured for 4 h at 200~ For the blend containing 30 phr of PEI, due to the very high viscosity of the resulting mixture all the components were dissolved in CH2C12; solvent removal and degassing were performed simultaneously in the vacuum oven at 100~

Blend

compositions and codes are reported in Table 3. DSC measurements were performed on several of these mixtures of varying compositions and are reported in Fig. 25. As in the case of the epoxy/PC system, a single glass transition temperature (Tg) is observed in all the cases, which increases by increasing the PEI content in the mixture. The Tg values are shown in Fig. 26 as a function of composition. The above results indicate that, in the composition range investigated, the PEI is molecularly dispersed in the TGDDM forming a single-phase homogeneous system. It

47 is worth noting that, when the DDS hardener is dissolved in these mixtures, they still remain homogeneous and visually transparent. Table 3.

Codes and composiaons of the investigated TGDDM/PEI mixtures. Code

TGDDM

DDS

PEI

PEI

A0

76.9

23.1

-

-

A5

74.1

22.2

3.7

5.0

AI0

71.5

21.4

7.1

10.0

A15

69.0

20.7

10.3

15.0

A20

67.0

20.0

13.3

20.0

A30

62.6

18.7

18.7

30.0

(%)

(%)

,,,

(%)

oT X Iii

-50

-25

0

25

Temperature

!

I

I

5O

75

100

('C)

Figure 25. DSC thermograms in the temperature range between-50~ and lO0~ for,4) AO; B) AIO; C) A20 and D) A30 blends.

48 I

320

310

I

I

-

300 -

290

-

280

-

270

-

260

'

I--

0

I

.

10

20

I' 30

40

Blend Composition [g PEI) Figure 26. Glass transition temperatures, Tg of the uncured TGDDM/PEI mixtures as a function of composition. The residual heat of curing, AI-Ir normalized to the epoxy resin content, is reported in Fig. 27A as a function of blend composition. We observe a linear increase of AI~ with increasing the PEI content in the blend. After the postcuring step. the neat resin shows no residual M-Ir. Conversely, for all the blends, AHr is considerably lower than that observed in the cured samples but still well detectable (Fig. 27B). Moreover a linear trend, approximately parallel to fl~at observed for the cured samples, is found between AI~ and P EI content. These results can be interpreted considering that, as it will be shown later, the PEI phase separates during the early stages of the curing process. It is likely fllat this PEI phase incorporates a small amount of unrcacted TGDDM which can act as plasticizer [50,51].

49 I

200

150

Ik,,,

!

I

-

100

"r'

50-

0

0

I

I

I

5

10

15

20

Blend Composition (% PEI) Figure 27. Residual heat of curing, AHnas a function of blend composition A) cured samples; B) postcured samples. The postcuring step allows the complete cure of the TGDDM in the epoxy phase as evidenced by the absence of residual AHr in the neat resins. In all the blends, upon postcuring, AI-Ir decreases by an approximately constant value (100 J/g) which corresponds to the AHr value of fl~e cured neat resin. This indicates that, also in file blends, the TGDDM phase is fully cured, but the residual TGDDM dissolved hi the PEI phase is not affected by a postcuring process at 200~ and can only be cured in the DSC experiments where higher temperatures are reached. Dynamic-mechanical spectra of the neat resin, PEI and a blend containing 20 phr of PEI are reported in Fig. 28. In the explored temperature range the pure epoxy resin gives rise to a single tan5 peak at 265~ corresponding to the glass transition. The tan6 peak of PEI is centred at 220~

50 0.6

EPOXY

0.4

I-- 0.2 0

2

PEI

co 1.5

:: 1.0

t~

I-- 0.5 0

0.6

EPOXY

t,o 0.4 I=

I--- 0.2 0 r

r

i

1

SO

100

150

200

Temperature

"

i

250

. . . . . . . "1

300

('C)

Figure 28. Relaxation peaks (tan6) in the temperature range between 90~ and 300~

A) AO; B) PEI; C) A20 blend.

The blend shows two distinct relaxation peaks centreA at 210~ and at 267~ respectively corresponding to the Tg's of file two blend components. Such a result clearly indicates that phase separation has occurred during the curing process. The dynamic-mechanical spectra of the other investigated compositions show the same features and the values of the transition temperatures remain unaffected by blend composition. It is noted that in all the blends the PEI Tg is lower than that of the pure PEI by about 10~ In agreement with the previously discussed DSC data, this effect could be ascribed to the plasticizing effect of the uncured TGDDM present in the PEI

51 phase. The dynamic-mechanical data just discussed are consistent with results previously reported by Bucknall and Gilbert [50]. The fracture behaviour of the neat epoxy resin and of the PEI-modified resins have been examined at low (1 mm/min) and high (1 m/see) deformation rate. The critical stress intensity factor is reported in Fig. 29 as a function of the PEI content in the blend.

2.5

I

I

I

I

I

I

5

10

15

2.0

E

1.5

Z 2~

"~

1.0

u

0.5

0

0

20

Composition (wt ~ PEI}

Figure 29. The critical stress intensity factor, Kr for the TGDDM/PEI system as a function of blend composiaon; (A) high speed tests, (0) low speed tests. For both the low and high speed tests, a significant increase of Kr is observed with increasing the amount of PEI in the blend. In particular a linear correlation between K c and PEI content is found to hold for both the sets of experiments. As expected, the Kr values at high speed are consistently lower than those at low speed,

52 due to the reduced capacity of plastic deformation of the system when the strain rate rises. The toughening effect of PEI is further evidenced in Fig. 30, where G e, is reported as a function of blend composition. As for K~, a substantial increase of G c is found by increasing the PEI content in the blend in both the low and high speed tests. In particular, we found a tenfold increase of Gr at high speed and a 13-fold increase at low speed for a blend containing 19 wt. % of PEI. This enhancement in toughness is of the same order of magnitude as that obtained with bifunctional epoxies toughened by reactive liquid robbers [6,7].

1.2

A

tM

E

I

I

I

0.8

0.4

0

5

10

Composition (wt

15

20

PEI}

Figure 30. The critical strain energy release rate, Gc~for the TGDDM/PEI system as a function of blend composition; (A) high speed tests; (0) low speed tests.

53 In the latter ease, however, a significant reduction of the elastic modulus, E, is generally observed. For the system under investigation, owing to the inherent rigidity of PEI, the elastic modulus shows only a very limited decrease by increasing the PEI content in the blend (Fig. 31).

I

I

13.

I

t ~ . - . . - o

0

LLI

1

0

I

I

I

5

10

15

Composition (wt

20

PEI)

Figure 31. The flexural elastic modulus, E, for the TGDDM/PEI system as a function of blend composiaon. The toughening effect of PEI can be ascribed to the capacity of the PEI phase to the plastically deformed under loading. The fracture surfaces of different epoxy/PEI blends broken at low speed were examined by scanning electron microscopy. To obtain further details, these surfaces were etched with a solvent of the PEI component (CHCI3) prior to examination. The

54 micrographs of both the unetched and etched surfaces are reported in Figs. 32 and 33 for comparison. At low PEI content (10 phr, Fig.32A, 32B) the thermoplastic component segregates into spherical domains with diameters of about 2-3 Ixrn, uniformly distributed within the matrix. The etching completely removes the PEI phase, leaving smooth cavities where the domains were located.

Figure 32. SEM micrographs of the fracture surfaces of samples broken at low deformation rate; ,4) unetchedAlO blend; B) etchedAlO blend; C) unetched A20 blend; 19) etched A20 blend. This result indicates that the PEI has not been crosslinked during the curing process and hence it retains its capability to be plastically deformed under loading.

55

Figure 33. SEM micrographs of the fracture surfaces of samples broken at low deformation rate; A) unetchedA30 blend; B) etched A30 blend at low magnification; C) etched A30 blend at high magnification.

At higher PEI content (20 phr, Fig. 32C, D) a much more complex morphology is observed. In addition to the above described spherical domains of PEI, very large and irregularly shaped regions are observed. Upon etching these regions evidence an inner structure, consisting of clusters of domains which are not dissolved by CHCI 3. These features indicate that in flae above regions a phase inversion has occurred, leading to a structure in which epoxy domains are coated by a continuous skin of PEI. In the areas where phase inversion does not occur, the PEI domains are completely

56 removed by etching as in the case of the blend with lower PEI content. In the blend containing 30 phr of PEI, the fracture surface shows again a single type of morphology, consisting of a continuous and a dispersed phase (Figs. 33A, B, C). The etching process evidences a "pomegranate structure" (Figs. 33B, 33C) clearly indicating that in this case the PEI component forms the continuous phase which is completely removed by the solvent. The epoxy resin is segregated into grain-like domains which are uniformly coated with a thermoplastic skin. Clearly, at this composition the phase inversion is complete. The morphological analysis just presented allows to interpret the fracture behaviour of this blend system. When the PEI is the dispersed phase, the toughening effect can be ascribed to a plastic deformation of the PEI domains. The fracture occurs by brittle failure of the epoxy matrix with the PEI domains bridging the crack and delaying its propagation; in these conditions most of the energy is dissipated by the modifier. Furthermore, the plastic deformation of the PEI domains suggests sufficient interfacial strength among the phases. The etching experiments demonstrated that this interfacial strength is unlikely to be due to chemical interactions. We must assume physical interactions like Van-derWaals forces to be responsible for this effect. When PEI forms the continuous phase, the failure occurs within the thermoplastic skin covering the epoxy domains which remain completely undeformed

(see Figs. 33). In particular, Figs. 33B and 33C

clearly indicate that in this case the crack propagates around, rather than through the epoxy domains. Therefore, yielding of the thermoplastic continuous phase is the main toughening mechanism. At the intermediate compositions (see Figs. 32C, 32D) both the types of failure processes are active. In addition to the above discussed toughening mechanisms, other deformation processes have been proposed to explain the improvement in tougheness for thermoplastic modified epoxies such as crack-pinning [52-54], crack-path deflection [55], microcracking [56-60] etc. From a fundamental analysis of these mechanisms it emerges that four parameters are important to maximize the toughening effect:

57 i. size of the particles ii. strength of the particles iii. adhesion between the particle and the matrix iv. distribution of the particles in the matrix Unfortunately the current routes used to prepare thermoplastic modified epoxies cannot independently control these parameters; for example the use of reactive oligomers increases adhesion but also reduces the particle size. Therefore there is a need to develop new experimental techniques to analyze from a fundamental point of view the influence of each of the above factors on the toughening effect. A deeper understanding of the interplay among the parameters which control the toughening will eventually lead to the development of new preparative approaches for the formulation of such kind of innovative materials. As an example of a promising method for producing thermoplastic-modified epoxies it may be cited the emulsion polymerization technique to produce struea~ed core-shell particles. The use of such core-shell particles should allow independent control of particle size, ductility and adhesion. In fact the structured core-shell particles are preformed, thus their size can be independently controlled with surfaetants. The ductility of the particles can be controlled by a suitable selection of the core polymer. The shell polymer can accommodate reactive groups which would allow independent control of adhesion. References

1.

H. Lee, K. Neville, "Handbook of Epoxy Resins", Mc Graw-Hill, New York, 1967.

2.

C.K. Riew, A. J. Kinloch, "Toughened Plastics r', Advances in Chemistry Series, 223, Washington D.C., 1993.

3.

J.N. Sultan, F. J. Me Gerry, "Microstructural Characteristics of Toughened Thermoset Polymers", Cambridge, MA, 1969.

58 4.

J.M. Margolis, "Advanced Thennoset Composites", Van Nostrand Reinhold Co. Inc., New York, 1986.

5.

C.K. Riew, E. H. Rowe, A. R. Siebert, Adv. Chem. Ser. 154 (1976) 326.

6.

A.F. Yee, R. A. Pearson, J. Mater. Sci. 21, (1986) 2475.

7.

A.J. Kinloch, S. I. Shaw, D. A. Tod, D. L. Hunston, Polymer 24 (1983) 1341.

8.

R.A. Pearson, A. F. Yee, J. Mat. Sci. 24 (1989) 2571.

9.

A.F. Yee, g. A. Pearson, J. Mat. Sci. 21 (1986) 2462.

10. A. J. Kinloch, R. J. Young, "Fracture of Polymers", Appl. Publishers, London, 1983, p. 421. 11. A. J. Kinloeh, "Structural Adhesives: Developments in Resins and Primers", Applied Science, London, 1986. 12. A.J. Kinloch, S. J. Shaw, D. L. Hunston, Polymer 24 (1983) 1355. 13. C.B. Bucknall, T. Yoshii, Br. Polym. J. 10 (1978) 53. 14. C.B. Bucknall, "Toughened Plastics", Applied Science, London, 1977. 15. E.J. Kubel, Adv. Mater. Process. 8 (1989) 23. 16. C.B. Bucknall, I. K. Partridge, Polymer 24 (1983) 639. 17. M. Abbate, E. Martuscelli, P. Musto, G. Ragosta, G. Scarmzi, J. Polym. Sci., Part B: Polymer Phys. 32 (1994) 395. 18. V. Di Liello, E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, J. Polym. Sci., Part B 32 (1994) 409. 19. E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, Die Ang. Makromol. Chem. 204 (1993) 153. 20. V. Di Liello, E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, Die Ang. Makromol. Chem. 213 (1993) 93. 21. E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, Die Ang. Makromol. Chem. 217 (1994) 159. 22. R. A. Pearson, "Toughened Plastics I", C. K. Riew and A. J. Kinloch Eds., Advances in Chemistry Series 223, Washington DC, p. 405 (1993). 23. C.B. Bucknall, I. K. Partridge, Polymer, 24 (1983) 639. 24. R.S. Raghava, Natl. SAMPE Symp., 28 (1983) 367. 25. J.L. Hendrick, I. Yilgor, J. E. McGrath, Polym. Bull. 13 (1985) 201.

59 26. C.B. Bucknall, I. K. Partridge, Polym. Eng. Sci., 26 (1986) 54. 27. J. A. Cecere, J. E. Mc Grath, Polym. Prep. (Am. Chem. Soc. Div. Polym. Chem.) 27 (1), (1986) 299. 28. H. Jabloner, B. J. Swetlin, S. G. Chu, U.S. Patent, 4, 656, 207 (1987). 29. S.G. Chu, B. J. Swetlin, H. Jabloner, U.S. Patent 4, 656, 208 (1987). 30. S.C. Kim, H. R. Brown, J. Mater. Sci., 22 (1987) 2589. 31. R.S. Raghova, J. Polym. Sci., Polym. Phys. Ed. 26 (1988) 65. 32. Z. Fu, Y. Sun, Polym. Prep. (Am. Chem. Soc. Div. Polym. Chem.). 33. R.A. Pearson, A. F. Yee, Polym. Mat. Sci. Eng., 63 (1990) 311. 34. J. Kim, R. Robertson, Polym. Mat. Sci. Eng., 63 (1990) 301. 35. M.K. Antoon, Ph. D thesis, Case Western Reserve University, Cleveland OH. 36. L. J. Bellamy, "The Infrared Spectra of Complex Molecules, Vols. I and II, Chapman and Hall, London, 1980. 37. N.B. Colthup, L. H. Daly, S. E. Wiberley, "Introduction to Infrared and Raman Spectroscopy", Academy Press, San Diego, 1990. 38. D.F. VameU, J. P. Rut, M. M. Coleman, Macromolecules, 14 (1981) 1350. 39. G.B. Enns, G. K. Gillham, J. Appl. Polym. Sci., 28 (1983) 2567. 40. W.F. Brown, J. Srawley, ASTM STP 410 (American Society for Testing and Materials), Philadelphia, 1966, p. 13. 41. G.R. Irvin, in Encyclopedia of Physics, Springer, Berlin, 1958. 42. E. Plati, J.G. Williams, Polym. Eng. Sci., 15 (1975) 470. 43. S. Yami, R. J. Young, J. Mat. Sci., 14 (1979) 1609. 44. D.C. Phillips, J. M. Scott, M. Jones, J. Mat. Sci., 13 (1978) 311. 45. B.W. Cherry, K. W. Thomson, J. Mat. Sci., 16 (1981) 1925. 46. J.A. Schroeder, J. Mat. Sci., 23 (1988) 3073. 47. R. J. Young, "Develompent in Polymer Fracture-l", E. H. Andrews Ed., Applied Science, London, 1979, p. 183. 48. A.J. Kinloch, J. G. Williams, J. Mat. Sci., 15 (1980) 987. 49. S. Yamini, R. J. Young, J. Mat. Sci., 15 (1980) 1823. 50. C.B. Bucknall, A. H. Gilbert, Polymer, 30 (1989) 213. 51. D.J. Hourston, J. M. Lane, Polymer, 33 (1992) 1379.

60 52. F.F. Lange, Philos. Mag., 22 (1970) 983. 53. L.R.F. Rose, Mech. Mater., 8 (1987) 11. 54. F.F. Lange, K. C. Radford, J. Mat. Sci., 6 (1971) 1197. 55. K.T. Faber, A. G. Evans, Acta Metall., 31 (1983) 565. 56. A.G. Evans, S. Williams, P. W. R. Beaumont, J. Mat. Sci., 20 (1985) 3668. 57. P.G. Chambides, R. M. MacMeeking, Mech. Mat., 6 (1987) 71. 58. A.G. Evans, K. T. Faber, J. Am. Ceram. Sot., 67 (1984) 255. 59. J.W. Hutchinson, Acta Metall., 35 (1987) 1605. 60. M. Ortiz, Appl. Mech., 54 (1987) 54.

61 CHAPTER 2

UNSATURATED POLYESTER RESINS E. Martuseelli, P. Musto, G. Ragosta National Research Council of Italy, Institute of Research and Technology of Plastic Materials, 80072 Arco Felice (Na) ITALY.

I. Introduction

Unsaturated polyester (UP) resins represent one of the most important matrices for composite applications [1,2]. They are particularly useful in sheet molding compounds (SMC) and bulk molding compounds (BMC) for manufacturing automotive parts [3-5]. Like other thermosets, UP resins are blended with several additives to enhance their properties. For example the polymerization shrinkage of the material during curing may cause moulding problems such as poor surface quality, warpage and difficult dimension control. Most of these problems are eliminated using low-profile additives like poly-methylmethacrylate and poly-vinylacetate [6-9]. Moreover the UP resins are limited by their brittleness, especially when good impact behavior is required. To overcome this limitation blending with liquid rubbers has been widely used [10-15], often with limited success. The critical point of this approach is the limited solubility of the rubbery component in the unreacted resin: thus the reduced toughening effect maybe due, in part, to separation of the rubber resin mixture before curing. The inhomogeneity of the starting system results in a relatively coarse rubber-particle distribution in the cured resin. The desired results can also be obtained if the reactivity of the rubbery component toward the thermosetting matrix is enhanced so as to promote a finer and more homogeneous dispersion and better interfacial adhesion [ 16-19].

62 In flae present chapter we describe the chemical modification of a series of commercially available liquid rubbers aimed at improving their reactivity towards the UP matrix. The molecular characterization of such reactive rubbers via FTIR spectroscopy is discussed in detail as well as the morphological and mechanical analysis of the resulting blends.

2. The system UP/Isocyanate terminated Polybutadiene A first commercial rubber employed to enhance the toughness of the UP resin was a hydroxyl-terminated polybutadiene (HTPB) having a molecular weight, M , of 3,000 and a functionality of 2. To enhance the reactivity of the HTPB towards the UP resin the end groups of HTPB were transformed into isocyanate groups by reaction with toluene diisocyanate (TDI), according to the following scheme: OCN.

HOWW~OH + 2 HTPB

CH3

~O'~,.-NCO

OCN

,CH3

CI'I3 NCO

(~>--NH-COOVWV~ COO-NH~Q-~ ITPB

TDI

In Fig. 1 are reported the FTIR spectra of HTPB (Fig. 1A) and of ITPB (Fig. 1B) in the range 4000-600 cm-l; both spectra were obtained on thin films cast from CH2C12 . The spectrum of HTPB displays an ill-resolved absorption in the range 37003120 cmq. In particular a broad peak approximately centred at 3360 cm-1 can be distinguished, with a shoulder at about 3600 cmq. These spectral features can be associated with the O-H terminal groups" the shoulder at 3600 cm-1 is attributed to the stretching vibration of non-interacting,

"free" OH groups,

while fl1r

characteristically broad absorption at 3360 cmq is assigned to the stretching modes of OH groups which self-interact through hydrogen bonding. The spectrum of ITPB displays marked differences with respect to the spectrum of the starting material. Besides the very intense VNCo peak at 2275 cm-1 in the O-H, N-H stretching region (3700-3120 cm-1) the broad yon absorption of HTPB is

63 transformed into two partially resolved peaks at 3422 and 3342 cm-]. This doublet is characteristic of the vm~ vibrations of urethane linkages whereby both the position and breadth of the peaks are determined by strong intermoleeular hydrogen-bonding interactions [20]. Furthermore new peaks at 1743 cm-I (with shoulders at 1723 and 1712 cm'l), at 1620, 1595, 1530, 1277, and 1207 cm-l are detected in the ITPB spectrum. It is to be noted that only the doublet at 1620-1595 em-l can be associated with absorptions of TDI (aromatic ring modes). All the others are indicative of the formation of new chemical bonds.

2.40"

|

t..-

_4 ii

9

oo eo e. =e

9 ~

Ii

i

1.60-

..Q t_ 0 ..Q

< o.8o-

0.00 -

4000

i

i

3200

i

'i

2400

i

i

1600

Wavenumber (cmFigure 1. bTIR transmission spectra of HTPB (solid curve) and ITPB (dotted curve) in the frequency range 4000-600 cm-L Spectra from films cast from CH2CI2.

64 In particular the band at 1743

c m -1 c a n

be associated with the carbonyl

stretching mode of urethane groups. However the carbonyl absorption is complex and the two low-frequency components could arise either from intermolecular interactions of the hydrogen bonding type or from the formation of different carbonyl species. We will discuss these spectral features in more detail later. The broad, asynmaetric absorption at 1530 em-1, as well as the doublet at 1217-1207 cm-l, are both characteristic of substituted amides (amide bands II and III) [20-21]. To obtain further details about the chemical linkages formed by the reaction between TDI and HTPB, solution FTIR spectra in Cc14 were measured. This was done in order to eliminate, by dilution in a non polar solvent, the effect of intermolecular interactions wlfieh generally cause a significant broadening in the OH, NH stretching region, as well as band splitting in the carbonyl range. In Fig.2, curve A, is shown the FTIR spectrum of a dilute solution (7.21 mg/ml) of HTPB in CC14 in the frequency range 3850-3260 cm-1.

.04 3.437 3640

r~

a .02 0 m

aa'oo

ae'oo Wavenu.lmbers

a4'oo (era-- I )

Figure 2. FTIR spectra of HTPB (curve .4) and ITPB (curve B) in the frequency range 3850-3260 cm j. Spectra obtained on dilute solutions

in CCI 4.

65 As a consequence of the dissociation of the hydrogen bonding interactions, the broad OH absorption centred at 3360 cm-1 in the solid state spectrum is resolved into a single, sharp peak at 3640 em-1 in the solution spectrum. The ITPB solution spectnma, Fig. 2, Curve B, displays an equally sharp and well resolved singlet at 3437 em-l; such a frequency is typical of the v~ vibration of unassociated -NH-COO- groups. The total absence of any residual OH absorption at 3640 em-1 in the ITPB solution spectrum, clearly indicates that all the hydroxyl terminal groups of the rubber have been transformed into isocyanate groups, according to the reaction scheme previously reported. This finding is confirmed by Fig. 3 where the HTPB and ITPB solution spectra (curve A and B, respectiveiy) are compared in the frequency range 2000-1570 cmq.

1642 .4"

1745

o G)

o m

"~ .2

1882

~~

I 1621

A ls'oo Wavenu.tnbers

~7"oo

(cm-- I )

te'oo

Figure 3. FTIR spectra as in Figure 2 but in the frequency range 2000-1570 cm-l.

It is clear that a single component at 1745 cm-1 is present in the ITPB spectrum, which arises from the Vco vibration of the urethane carbonyls.

66 Thus the complex shape observed in the carbonyl region of the ITPB solid-state spectrum (Fig. 1B) is not due to the formation of more than one carbonyl species but to intermoleeular hydrogen bonding interactions of the type: i

oc.

0.3 o

O C N ~ CH3I ~ H! ,, !

When ITPB is mixed with the UP resin, the following reaction may occur: CH3, y---(NCO

W'WWCOO-NH(\-.-..-/ O ~ §OH ITPB UP

WW~CO0_NH~O}

UP

The kinetics of this process was monitored by FTIP~ following the intensity decrease with time of the (NCO absorption at 2275 cm-1. The isothermal measurement was carried out at 80~ in a temperature chamber directly mounted in the spectrometer. In Fig. 4 the spectra collected at different reaction times for a blend containing 10 wt.-% of ITPB are reported. From the above spectral data, the percent conversion of the NCO groups as a function of time can be calculated and is reportexl in Fig. 5. It is apparent that the reaction proceeds rapidly up to a conversion of about 50 %. At longer times the rate decreases gradually, and the conversion can be consider~ complete aiter about l0 h. In these conditions a triblock copolymer of type A-B-A is formed, where A and B represent an UP chain and a polybutadiene chain, respectively.

67 t (min)

33

d

Illl~~~/

d

I 2680.

0

2585.

I 8

I

2511.

1

2,438.

t 7

2362.

I 2

2287.

t 8

I

2213.

3

213R.

I cl

2064.

.... i 4

1990.

0

Figure 4. bTIR spectra o f a mixture containing 10 wt % o f lTPB in the frequency range o f 2660-1990 cm -1; spectra were collected at 80~ and at varying reacaon times as indicated 100

I

80

,

~~

S

6o

!

40 i 2O 0

0

t

100

I

200

I

i

I

800

400

600

600

Figure 5. Conversion o f NCO groups as a function o f time for the isothermal measurements at 80~

68 On the basis of the above results we may assume that, in the early stages of the process, all the NCO groups present on the surface of the rubbery domains readily react; under such conditions the rate-limiting step is the reaction itself, which is relatively fast owing to the high reactivity of the NCO and OH groups involved. Once the NCO groups at the interface have been consumed, for the reaction to proceed further it is necessary that the remaining NCO end groups migrate from the bulk to the surface of the rubbery domahls. Under such conditions, file diffusion of the NCO end groups towards the interface becomes the rate-limiting step, and the process is correspondingly slowed down. We note that under our experimental conditions the first regime lasts 35 min while the second is more than 15 times longer. However, it is essential to reach complete conversion of all the isoeyanate groups in order to obtain the desired morphology and hence flae desired toughness of the final material. The fracture parameters Kr and G e of the various investigated blend compositions (see table 1) are reported in Figs. 6 and 7, respectively.

Table 1. Codes and compositions of the investigated mixtures. Code B0

UP

Styrene

HTPB

ITPB

(wt%)

(wt%)

(wt%)

(wt%)

-

-

70

30

A10

63

27

10

-

BI0

63

27

-

10

B20

56

24

-

20

The plain resin exhibits very low values of Kr and G~, reflecting the poor crack resistance of a higlfly crosslinked material. Similar results are observed for the blend containing HTPB as the rubbery plmse. In contrast, the blends in which the hydroxyl end groups of the rubber were previously transformed into isocyanate groups, display a marked improvement in fracture toughness. For these blends, in fact, the values of

69 K c and G e are from twice to five times larger than those of the neat resin. However such an improvement is found to be dependent on two parameters: 1) The time period during which the PE/ITPB mixture was allowed to react at 80~ prior to the curing process. 2) The amount of ITPB rubber in the blend. 1.2 1.0

N

0.8

E z

0.6 M,

~.

O.4 0.2

UP

UP/HTPB UP/ITPB UP/ITPB UP/ITPB 90/10 90/10 90/10" 80/20*

Figure 6. K c values for the various investigated blend compositions; (l[]) high speed tests (1 m/sec) ; (11) low speed tests (1 mm/min). *Blend premixed for 600 mm at 80~ In particular, for the B 10 blend, both K~ and G~ increase when the reaction time of the PE/ITPB mixture is increased from 30 to 600 min. As discussed in file previous section, such an effect can be ascribed to the fact flint the reaction between PE and ITPB is complete only after 600 rain, while after 30 min the conversion of isocyanatc groups does not exceed 50 %. Therefore in the fomlcr case the triblock copolymer PE-ITPB-PE is formed to a larger extent, thus producing increased adhesion at the rubber matrix intcrphasr

70

0.6 0.5

N

E

0.4 0.3

u

(..9

0.2 0.1

0

~

~

UP

UP/HTPB UP/ITPB UP/ITPB UP/ITPB 90/10 90/10 90/10" 80/20*

Figure 7. Gc valuesfor the various investigated blend compositions; (El) high speed tests (1 m/sec) ; al) low speed tests (1 mm/min). *Blend premixed for 600 mm at 80~ Using the same preparation procedure, an increase of ITPB content in the blend from 10 to 20 % w/w produces a decrease in the touglmess parameters. This effect may be related to the relatively coarse rubber particle distribution which, as will be shown in the next section, is produced in the blend with the higher rubber content. The fracture results were interpreted in terms of a fractographic analysis performed by SEM. When the unmodified polyester resin is fractured the crack propagation occurs in a brittle, unstable mariner. The scanning electron micrograph of the fracture surface shows, at high magnification, the presence of free lines mainly emanating from the crack initiation region (see Fig. 8). These lines extend approximately along the crack propagation direction and are associated with the point of arrest of the crack front. The absence of any fractographic feature distinctive of plastic flow is consistent with the observed lack of touglmess. The surface of the rubber-modified resins is characterized by a morphology in which the rubber is segregated into spherically shaped domains, homogeneously dispersed in the PE

71

matrix (Figs. 9-12). In particular the micrograph of the AI0 blend (Fig. 9) shows a large number of rubber particles with diameters ranging from 10 to 30 ~tm.

Figure 8. SEM micrograph of the surface of a BO specimen after impactfailure

Figure 9. SEM micrograph of the surface of a sample of the AI Oblend after impact failure.

72 Furthermore most of the rubbery domains appear as circular cavities in the matrLx; such an observation can be explained by assuming that, during the fracture process, the particles break in two pieces which remain bonded to the original sites. After failure, the half-particles retract, due to the vanishing of the triaxial tensile stress and the surfaces of the particles lie below the fracture plane [22]. Moreover the PE matrix around the cavities appears rather flat, indicating its very limited plastic deformation. When HTPB is replaced with ITPB, strong variations in the overall fracture morphology are observed. The micrographs of the B 10 blends (Figs. 10a and 10b) evidence a marked reduction of the particle size (diameters 3-5 ~tm) with respect to the corresponding PE/H

B blend.

These domains are likely to originate from the aggregation of the central rubber blocks of the PE-ITPB-PE copolymers formed prior to the curing reaction. The PE blocks, on the other hand, are likely to be dispersed within the PE/styrene matrix and to take part in the subsequent curing process. In such a way, the condition is realized to obtain strong bonds between the rubbery domains and fl~e PE matrix. It is also evident, especially from Figs. 11 and 12 that some of the particles show well defined holes at the center.This phenomenon, already observed in epoxy-rubber systems, is known as cavitation. The fractured surfaces also display clear evidence of extensive shear yielding of the PE matrix around the rubber particles. This phenomenon is far more pronounced in the ease of the B 10 blend where the PE/ITPB mixture was allowed to react for 600 rain, prior to cure (Fig. 10a). The micrograph of the B20 blend shows fractographie features analogous to those observed in the B10 blends (see Fig. 12), apart from a reduction of plastic shear deformation of the matrix and an increase in particle dimensions (5-20 btm). From the above fractographic analysis it emerges that shear yielding of the matrix is file primary source of energy dissipation during fracture in rubber-modified PE resins. In fact, the stress field associated with the rubber particles in the neighbourhood of a loaded crack leads to the initiation of two basic deformation mechanisms which can strongly interact with each oilier [23-28].

73

Figure 10. SEM micrographs of the surface of a sample of the BIO blend: a) UP/1TPB mixture reactedfor 30 rain at 80~ before curing; b) UP/ITPB mixture reactedfor 600 rain at 80~ before curing.

Figure 11. SEM micrograph of the sample of Fig. lOb at higher magnification.

74

Figure 12. SEM micrograph of the surface of a sample of the B20 blend The UP/1TPB mixture was allowed to react for 600 min at 80~ before curing.

One is the formation and growth of localized shear-yield deformations caused by the fact that during loading the particles produce stress concentration at their equators, acting as sites for the initiation of shear deformation in the matrix. The other mechanism is the cavitation of file rubber particles whereby the loading generates a triaxial state of stress at the crack tip which causes failure and void formation either in the particle or at the particle matrix interphase. Once formed, these cavities lower the stress required for shear yielding, promoting further shear deformation in the matrix. The mechanisms outlined are scarcely effective in the case of the A10 blend, mainly because of the poor adhesion at the interphase. The lack of adhesion, in turn, reflects the absence of chemical hlteraction between PE and HTPB. On the contrary, both the mechanisms operate in the PE/ITPB systems. Here the reaction between PE and ITPB generates triblock copolymers which, acting as emulsifiers, reduce the particle size and firmly bond the two phases together when phase separation occurs.In particular when all the ITPB rubber has reacted with PE the interracial adhesion is ma~,dmized, as can be seen from the micrograph of Fig. 15b, where extensive shear

75 yielding of the matrix is apparent. This effect, together with the achievement of an optimum particle size, results in the higher values of fracture toughness observed in this system.

3. The system UP / maleimido terminated rubber

As a second example of the use of a reactive liquid rubber to toughen polyester matrices, we will describe a system in which the modifier was a maleimido-terminated butadiene-acrylonitrile copolymer (ITBN). This toughening agent was obtained by chemically modifying a conmaercially available amino-terminated rubber (ATBN) having a molecular weight M~ = 3600 and an acrylonitrile content of 18% wt/wt. It is to be noted that the -NH 2 groups on ATBN were obtained by reacting a carboxyl terminated butadiene-acrylonitrile copolymer with methylpentamethylene diamine. Tlms the chemical structure of the ATBN temfinal groups is the following:

ANV'C--N H--C H2--'CH--C H2"--CH2~NH2

The reaction scheme for the preparation of the maleimido-terminated copolymer (ITBN) is the following: ~0 H2N~NV~NH2 + 2 ATBN

O

=

0 0 II II I( C'HN WVW~NH'C~I1 ~,,,c~OH OH.c..,..~ II MATBN II O O I Ac,20, NaAc

0

0

. .o5 ITBN

+2 H20

76 The reaction is carried out in chloroform as solvent, and proceeds in two steps. First a maleamic acid is formed by reacting the -NH2 end-groups with a stoichiometric amount of maleic anhydride at room temperature. Then this intermediate is cyclized by adding to the solution acetic anhydride and sodium acetate. In Fig.13 are compared the FTIR transmission spectra of ATBN (Fig.13A), MATBN (Fig. 13B) and ITBN (Fig. 13C). In the discussion below we will confine the attention on the spectral regions where the amide and carbonyl groups display their characteristic absorptions (3600-3000 cm"l for the NH vibrations and 18001550 cm-1 for the amide modes).

1.5A

._.._____._.__.._

r

g]

.5

30'00

2000 Wavenumbers

1000

(om-l)

Figure 13. FTIR transmission spectra of A) ATBN; B) MATBN; C) ITBN. The ATBN spectrum displays a broad, irregularly shaped absorption centred at 3300 cm-1 which can be readily attributed to the NH modes of the amide linkages in

77 the terminal groups. This peak displays a higher frequency shoulder at approximately 3385 cm-1. The low frequency component is due to hydrogen bonded N-H groups, while the high frequency shoulder is due to free N-H. Typical amide absorptions are detected also in the carbonyl frequency range. In particular two peaks centred at about 1650 and 1545 cm-1 (amide mode I and II) are observed.

1670

1645 [

1545

A.

I

~

s

-

o 5 1545

, ~oo

,8'oo

1~oo

Wa'lenumbers

J ioo

, ~oo

(cm-l)

Figure 14. FTIR transmission spectra o f A) ATBN; B) MATBN; C) ITBN in the frequency range 1900 - 1500 cm -1.

Unlike the essentially isolated NH stetclfing, these two modes are more complex vibrations. In particular the amide I mode is predominantly a C=O stretching, but has significant contributions from a C-N stretching and a C-C-N deformation. Tiffs mode is conformationally sensitive mad has been widely used to detect the different conformations in polypeptides and proteins [29]. In the ATBN spectrum the amide I

78 band clearly displays two unresolved components at 1672 cm -1 and at 1642 cm -l. The two components are assigned to hydrogen bonded carbonyls (at 1642 cm -l) and to carbonyls not involved in specific intermolecular interactions (at 1672 cm-l). Thus in fl~e system under investigation flae following type of hydrogen bonding interactions take place:

I~

CH3

qWV'C--N~CI-12--C H--CH2--CH2--NH 2 I

vVV~C--Ir'-'-C I.-12--CH--C H2--C I.-12--NH2 H

Hydrogen bonding involving carbonyls as acceptors and -NH 2 groups as donors cannot be entirely ruled out. Assuming that the absorptivities of interacting and noninteracting carbonyls do not differ substantially, as it is generally the case, in the ATBN copolymer at room temperature, the population of interacting C=O groups prevails over the population of non interacting groups. Well defined differences are observed in the MATBN spectrum with respect to the spectrum of the parent copolymer. In particular, in the NH stetching region the peak intensity increases substantially; the absorption is now centred at 3277 cm -1 and remains rather broad. This effect is due to the contribution of the newly formed NH groups of the maleamic acid moieties, which further complicate the NH region. The carbonyl region of the MATBN spectrum differs substantially too from that of the parent copolymer (see Fig. 14): a new, lfighly symmetrical peak centred at 1719 cm -1 is observed, which is assigned to the Vc_o vibration of the carboxyl groups in the maleamic acid moiety. A further absorption is found at 1635 cm -1 while the peak originally centred at 1545 cm -l in the ATBN spectrum, strongly increases in intensity. The latter doublet is due to amide band I mad II and the intensity increase arises from the newly fonned amide groups in the maleamic acid units. However the shape of the 1635 cm -1 band is different from that observed in the ATBN spectrum. Its position

79 correspond to that of hydrogen bonded amide carbonyls, although the frequency is 7 cm-l lower than that of the same mode in the ATBN spectrum; furthemlore it is very sharp (FWHH = 15 cm-1) and rather symmetrical. These observations indicate that the molecular interactions realized by the maleamic acid moieties of AMTBN are different and stronger than those occurring among the amide linkages present in the terminal groups of ATBN. Finally, the FTIR spectrum of ITBN (Figs. 13C and 14C) clearly shows the occurrence of the cyclization process. The VNHabsorption decreases as a consequence of the consumption of the NH groups of the maleimide. The Vc_o peak centred at 1719 cm-1 in the MATBN spectrum is shifted at 1709 cm-1 and a satellite band of the

main Vc__o absorption appears at 1775 cna-l. This doublet is characteristic of imides and is due to the symmetric and asymmetric stretching vibration of the imide carbonyls respectively. As expected both the amide band I and II decrease in intensity with the latter shifting at 1545 cm-1, the same value found in the ATBN spectrum. It has been found that the cyclization reaction also proceexts at temperatures above 100~

Since the cyclization of the maleamic acid end groups produces well

detectable features in the infrared spectrum, FTIR spectroscopy represents a particularly attractive technique to follow the kinetics of such a process. In Fig. 15 is reported a conversion, ~, versus time curve collected a 125~ The values were calculated from the spectral data in Ref. 29. It is observed that, under such conditions complete cyclization is approached in about 100 minutes. The spectrum of the thermally cyclized ITBN is very close to that of the product obtained by the solution reaction. However, while the latter remains soluble in all the solvents of the rubber, the former is completely insoluble. Evidently a cross-linkhlk reaction occurs at temperatures of 120~

or above, which is not detected

spectroscopically. Such a process is likely to involve the bismaleimide unsaturations which can be themmlly activated owing to their very high reactivity. These groups give very weak signals in the IR spectrum m~d this accounts for the fact that no detectable differences are observed between the spectra of soluble and hasoluble products. A further observation is that, unlike the ATBN copolymer, the

80 ITBN prepared by the solution method becomes insoluble when treated at 125~ This confirms that the cross-linking involves the maleimide groups and suggests that, during the thermal ring-closure process cyclization and cross-linking occur simultaneously.

I

0.90 0.75

-

0.60

-

0.45

-

0.30

-

0.15

-

0.00

I

.,

I

I

I

I

-

T

i

'J

0

15

30

~

45

j

i

60

75

"l

90

time {mini

Figure 15. Conversion versus time curve for the maleamic acid cyclization o f AMTBN rubber. Data collected at 125~

Blends of UP with ATBN and ITBN rubbers were prepared, whose codes and compositions are reported in Table 2. These materials were analyzed in detail with respect to their morphology and mechanical fracture properties.Ultra thin sections of UP/ATBN and UP/ITBN blends stained with osmium tetroxide were examined by transmission electron microscopy (TEM) in order to obtain information on the internal structure of the rubber particles.

81

Figure 16. TEMmicrographs of a stained ultrathin secaon of the C10 blend Table 2.

Codes and compositions of the investigated blends. Code

UP

ATBN

ITBN

(wt %)

(wt %)

(wt %)

CO

100

C4

96

4

C8

92

8

C10

90

10

C15

85

15

D4

96

-

4

D8

92

-

8

D10

90

-

10

D15

85

-

15

82

Figure 17. TEM micrographs o f a stained ultrathin section o f the DIO blend

Both ATBN and ITBN contain unsaturated sites along their backbone and are therefore dark stained in the presence of osmium tetroxide [30,31], giving good contrast with the saturated UP matrix. This is clearly shown in Figs 16 and 17 where TEM micrographs of UP/ATBN and UP/ITBN blends containing 10% by weight of rubber are reported. For both file types of modifier, file rubbery phase shows a spherical shape, indicating that no defomlation of the rubbery domains during sectioning has occurred. However, most of the ATBN particles appear detached from the UP matrix, while all the ITBN domains remain well bonded to the matrix. This is an indication that better adhesion between the two phases is attained in the UP/ITBN system. The stained ATBN particles (see Fig. 16) show a very coarse internal structure in contrast with the relatively free and homogeneous texture exhibited by the ITBN particles. Moreover both ATBN and ITBN particles do not seem to contain inclusions of UP reshl, indicating that the dispersed phase is made solely of the rubber component. The size of the rubber domains is the same for both the blend systems investigated

83 and ranges from 5 to 10 ~tm. To obtain further details on the nature and distribution of these rubber particles, fracture surfaces obtained by breaking the blend samples in liquid N 2 were examined by scanning electron microscopy (SEM). Prior to the SEM observation, the fracture surfaces were etched with a solvent of the rubber component (CH2C12). Figs. 18A and 18B show the micrographs of the UP resin modified with 10 % of ATBN and ITBN respectively.

Figure 18. SEM micrographs of the fracture surfaces of samples broken in liquid N 2 and etched in CHCI3. A) C10 blend; B) DIO blend Comparing file micrographs appears that the etching treatment removes completely most of the ATBN particles (see Fig. 18A) leaving cavities unifomaly distributed wiflfin the matrix, h~ contrast, the fracture surface of the UP/ITBN blend remains almost completely unaffected by the etclfing treatment (see Fig.18B), thus

84 indicating that the ITBN domains, after the curing process, are no longer soluble hi CHCI3, likely because a cross-linking reaction of the rubber has occurred. This reaction increases the intrinsic stiffiaess of the rubbery domains and this effect could account also for the finer texture observed in TEM analysis, which is likely due to a more efficient microsectioning of the UP/ITBN system with respect to the UP/ATBN blend. The SEM analysis confirms that the two blend systems have a similar size distribution of the rubber particles. A change in solubility of the two rubber modifiers after the curing process was confirmed by selective extraction of the blend components: the blends were first freely grounded and than left in CHCI3 for ? h. The solvent removes the soluble fraction of the rubber as well as the residual uncrosslinked LIP fraction. The quantitative results of such an analysis are reported in Table 3. Table 3.

Results of the selective extraction of the blend components Code

fraction extracted (%)

total amount of UP extracted (%)

total amount amount of of rubber rubber extracted extracted on the total (%) rubber present in the blend (%)

CO

1.55

1.55

C10

12.45

3.30

9.15

91.6

D10

4.47

2.90

1.57

15.7

D15

4.36

2.90

1.45

9.7

It is noted that in pure UP the amount of unreacted UP molecules is negligible (1.6 %). In the 90/10 UP/ATBN blend the total soluble fraction amounts to 12.45 %. Solution infrared analysis showed that, of this fraction 9.15% is ATBN while the

85 remaining 3.3% is unreaeted UP. Thus we were able to extract most of the rubber initially present in the blend (91.6 %). On the contrary, in the 90/10 UP/ITBN blend, the total recovered fraction drastically decreases (4.47 %) due to the fact that only 15.7 % of the initial ITBN content can be extracted. The unreacted UP fraction does not change noticeably (2.90 %). Finally for the 85/15 UP/ITBN blend we found 9.7 % of extractable ITBN and 2.9 % of UP. These results clearly indicate that, unlike ATBN, the ITBN copolymer undergoes chemical reactions during the curing of the UP matrix due to the presence of the reactive maleimide end groups. These reactions make the products insoluble in CHCI 3. In a previous paper [32] we found that the maleimide double bonds easily undergo addition reactions through a radical mechanism in UP/Bismaleimide blends. In UP/ITBN mixtures such a process is further favoured by the presence of a radical initiator in the system. Once formed, the ITBN macroradicals may interact between themselves forming a cross-linked network or may interact with the UP/styrene system being cured. In the latter case some ITBN molecules would be firmly bonded to the UP/styrene network, thereby improving the interracial strength among the phases. This improvement, in turn, will play an important role in determining the low and the large strain properties of the UP/ITBN blend system. In the discussion which follows we will show evidences indicating that both the ITBN reactions mentioned above do occur simultaneously in the UP/ITBN blend system. The Young modulus, E, of the various rubber modified UP resins was determined h~ three-point bending mode at ambient temperature and at a cross-head speed of 1 mm/min using the equation:

L3p E= ~ 4dWB 3

(1)

where d is the displacement, P the load at the displacement d, L is the span and W and B are the widfll and the thickness of the specimen respectively.

86 The E versus composition curves for both UP/ITBN (curve A) and UP/ATBN (curve B) blends are shown in Fig. 19.

2o5

I

1

I

I

I

I

5

10

15

-

2.0I..U 1,,5

-

0

Blend Composition

20

rubber)

Figure 19. Elastic modulus, E, as a function of blend composition, tt) UP/ 1TBN system; B) UP/ATBN system. As expected, for all the materials tested, the modulus decreases gradually with increasing the amount of rubber. In particular a linear correlation is found between modulus and composition. It call be noted, however, that the ITBN based blends display higher values of E with respect to the ATBN blends. This apparently surprising result can be explained by assuming that the ITBN rubber undergoes a crosslinking reaction during the curing of the UP matrix. Such a process enhances the ITBN modulus and therefore the soffe~ting effect of the ITBN rubber is less pronounced thin1 that of the uncross-linked ATBN rubber.

87 To predict large-strain properties, the yielding behaviour of all the investigated blends was examined under uniaxial compression mode. Typical stress-strain curves obtained at the same loading rate and temperature as for the Young modulus, are

12oj

shown in Figs. 20 and 21 for UP/ATBN and UP/ITBN blends respectively. l

:,

l

I

A

100 o'y A

13_

B

80-

C 60-

r

Di

/ /

40-

/ I

20-

0

[

-

0

EY

4

[

8

"'

I

12

"[

16

'"

20

Strain (%) Figure 20. Compressive stress-strain curves at ambient temperature and at a cross-head speed o f l ram~rainfor the UP/ATBNsystem. A) CO; B) C4; C)

c8; D) C10; E) C15. It can be observed that, when loadeA in compression, all the samples yield and flow, contrary to what appears in tension, where they exhibit a completely brittle behaviour. The compressive yield stress, ao,y, evaluated as in Fig. 20 is plotted as a function of blend composition in Fig. 22. For both UP/ITBN (curve A) and UP/ATBN (curve B) flae yield stress decreases with increasing the rubber content. This is a direct consequence of the lower shear modulus of the rubbery phase compared to flTethermosetting matrix, Milch prevent the rubber domains to support a significant amount of fl~e applied stress.

88 120

I

1

I

100 -1

A

!

C

13.. t,~

o

O0 i

0

I

0

'

4

!

!

I

8

12

16

Strain -(~;)

20

Figure 21. Compressive stress-strain curves at ambient temperature and at a cross-head speed of 1 ram~rainfor the UP/1TBN system. A) CO; B) D4; C) D8 D) D1 O; E) D]5. 100

[

I

I

I

8O

60-

o

b" 40

20

0

I

I

i

5

10

15

20

Blend Composition (~ rubber) Figure 22. Compressive yield stress, %.y as a fimction of blend composition. 14) UP/ITBN system; 17) UP/ATBN system.

89 From Fig. 22 it is also evident that, in the composition range investigated, the UP/ITBN system exhibits higher Cc,y values with respect to the UP/ATBN blend. This behaviour is generally observed whenever an improved adhesion is realised between the matrix and the dispersed phase. In fact, in such a case, the rubbery domains are able to act as load bearing components due to the mechanical continuity of the structure. Thus larger stresses are needed to yield the material. The better adhesion realised in the UP/ITBN system may arise from the chemical interaction between the maleimide functionalities of ITBN and the unsaturations of the UP/styrene system during the curing process.

I

1.0

I

,

I

A 0,8

-

N

E Z

3; w

0.6-

u

0.4-

0 . 2 ....

0

I

I

I

5

10

15

Blend composition

20

rubber}

Figure 23. Critical stress intensi~ factor, Kr at low strain rate (1 mm/mit 0 as a function of blend composition. A) UP/ITBN system; 13) UP/A TBN system.

90 On samples cured at 70 ~ and postcured at 100~ fracture mechanics tests at low and high deformation rate were performed to evaluate flae intrinsic touglmess [23,24]. The critical stress intensity factor, Kr determined from flae load-displacement curves is reported in Fig. 23 as a function of the rubber content in the blend. It can be seen that, while in the case of the UP/ATBN system (curve B) the Kr increase is relatively modest, for the UP/ITBN system (curve A) a substantial enhancement of toughness is found. In particular, Kr increases sharply by addition of small amounts of ITBN rubber and than levels off at a rubber content of 10-15%.

!

450

I

o 350

-

250

-

!

A

Owl

E u

(.9 150 -

50 0

I

I

I

5

10

15

20

Blend composition (%; rubber}

Figure 24. Critical strain energy release rate, G c at low strain rate (lmm/min) as a function o f blend composition. ,4) UP/ITBN system; 13) UP/ATBN system.

91 The enhanced touglmess showed by this blend system is further evidenced ill Fig. 24 where the fracture data are expressed by parameter G~ (critical strain energy release rate). The values of G c were calculated from the values of Kr and of the elastic modulus E, according to the equation:

~ =-E-(1- v~)

(2)

where v is the Poisson's ratio, taken as 0.35. In terms of Gr the addition of 10+."15% of modifier rises the toughness of the UP matrix by a factor of about 5 when the rubber is ITBN and of about 2 in the case of ATBN. 1.0

0~

I

!

I

i

I

I

5

10

15

-

0.6w

0 . 4 -t

0.2 0

20

Blend composition (~ rubber] Figure 25. Critical stress intensiO~factor, Kc, under impact conditions (1 m/sec) as a fimction of blend composition. A) UP/ITBN system; B) UP/A TBN system.

92 Fracture measurements were also carried out under impact conditions (1 m/sec) in order to evaluate the toughness of these materials under rapid loading. The corresponding K c and G c values are reported in Figs. 25 and 26 respectively. The behaviour of the impact toughness parameters in analogous to that observed in the low speed tests, apart from a decrease in the absolute values of K c and G c. The whole of the fracture data clearly evidence the higher toughening efficiency of the ITBN rubber with respect to ATBN. As previously mentioned both the type of rubbers may act as sites for the initiation and growth of localised shear deformation in the matrix [34]. This mechanism, which is believed to be the main source of energy dissipation, occurs to a greater extent in the UP/ITBN blend mainly because of the stronger adhesion across the particle-matrix interface, which prevents premature particle debonding.

350 .

I

I

I

A A N

250 E

"3 w r

O

150 -

50 0

I

I""

I

5

10

15

Blend

composition

(~

20

rubber)

Figure 26. Critical strain energy release rate, G c, under impact conditions (l m/sec) as a function of blend composition. ,4) UP/ITBN system; B) UP/ATBN system.

93 Further details on the deformation mechanism occurring in these blend systems were obtained by a fractografic analysis of the surfaces of samples fractures at low deformation rate. In Fig. 27 are shown the SEM micrographs of an UP/ATBN and an UP/ITBN containing 10 % by weight of the modifier. The micrographs have been taken near the notch tip, in the region of crack-growth.

Figure 27. SEM micrographs of the fracture surfaces obtained at low strain rate. A) C10 blend; B) DIO blend.

The micrograph of the UP/ATBN blend (Fig. 27A) reveals that the majority of the rubber particles have been fractured approximately along their equatorial plane. The average diameter of these particles appears to be unchanged, within experimental uncertainty, compared to the diameter of the undeformed particles as determmed by TEM. This seems to indicate that the fracture of the rubber domains occurs without

94 particle cavitation. However, some unbroken particles clearly debonded from the matrix are also observed. The above findings, together with the relative smoothness of the fracture surface around the rubber particles indicate that no significant plastic deformation has occurred ahead the crack-tip. The other UP/ATBN blend compositions investigated (micrographs not reported) display analogous fractografic features. When ATBN is replaced by ITBN (Fig. 27B) we again observe equatorial rupture of the rubber particles, but there is no evidence for particle debonding. Moreover the fracture surface displays clear evidence of extensive plasticity of the UP matrix. The fractografic analysis is in agreement with the touglmess results and gives conclusive evidence that the primary source of energy dissipation is the formation of a plastic zone ahead the crack-tip. In fact the stress field associated with the rubber particles promote the formation of multiple but

localised shear

deformation bands in the matrix. Although file two types of rubber investigated display similar particle size distribution, this mechanism is far more effective in the ITBN modified resin. This may be related, as already pointed out, to a better adhesion at the particle/matrix interface which is realised in this system through a chemical interaction between the rubber and the UP resin. It is to be noted, however, that a further contribution to energy absorption in the UP/ITBN system, arise from the cross- linked structure of the rubber domains which require a higher energy to be broken. We have already stressed that localised yielding in the vicinity of the crock-tip is the main mechanism controlling the degree of plastic deformation and the increased toughness of the investigated rubber modified UP resins. In the light of this observation it is of interest to examine models which combhae yield and fracture data in order to obtaha a quantitative estimation of the deformation process occurring at the crock-tip. It has been shown [23-35] that, for relatively brittle polymers the extent and shape of the localised plastic deformation ahead of a crack tip can be successfully described using the models developed by Dugdale [36] and Irwin [37].

95 The Dugdale model assumes that the yielding of the material at the crack tip renders the crack longer by an amount equal to the length of the plastic zone R. The value of R is related to K c and to the tensile yield stress gy by:

R=z__..Kr

(3)

8

260

210

~ 9

160

n.-

110

I

i

-

o

o

I

A

8

6o

10 0

I

i

I

5

10

15

20

Blend composition (g rubber}

Figure 28. The length of the plastic zone, R, as a function of blend composition A) UP/ITBN system. B) UP/ATBN system. Similarly Irwin advanced the hypothesis that the shape of the plastic zone ahead the crack-tip can be considered approximately circular. From geometrical considerations he concluded that the radius, rp, of this zone, can be estimated, for the plain strain case, by

l( /2

' :Gt J

(4)

96 Comparison of eq. 3 and 4 reveals that the Dugdale model predicts a larger extent of plasticity than the Irwin circular model. However, both the equations predict that the size of the plastic zone is proportional to (K~ / cry)2. The R and rp values calculated according to eq. 3 and 4 respectively are reported as a function of blend composition in Figs. 28 and 29 respectively. The tensile yield stress values were obtained from the uniaxial compression tests, assuming the ratio trt,y / cr~,y to be equal to 0,75 [38-40]. Both the rp and R parameters increase linearly with increasing the rubber content in the blend but the values relative to the UP/ITBN blend are higher than those of the UP/ATBN blend. These results further confirm the higher efficiency of the ITBN rubber in toughening the UP resin. 40

30

E ,~

I

I

-

I

A

20

t_

10

I

I

I

I

0

5

10

15

20

Composition (~ rubber)

Figure 29. The radius of the plastic zone, rp,as a function of blend composition A) UP/ITBN system; B) UP/ATBN system. 4. The system UP/Polybutadiene modified with maleic anhydride. As a last case study we will consider a system ha which the liquid rubber is a Poly isobutylene (PIB) whose chemical fonrmla is reported below:

97 CH3

. I

R R'

.

I I

,

CH3--CH2---CH2-~--CH2 n--~H2--C=C-- R CH3 PIB

where R, R', R"= H, Cl-t3

The technical note provided by the producer states that PIB has a double bond per clmin in the end-group, as determined by NMR spectroscopy. This double bond can be either di- or tri-substituted, while tetra-substituted unsaturations were excluded by UV analysis; moreover the disubstituted double bonds can be of the vinyl or vinylidene type. Succinic anhydride groups (SA) were attached onto the PIB chains through a reaction with maleic anhydride, initiated by benzoyl peroxide [41]. Blends containing both the unmodified PIB and the modified rubber (PIBSA) were prepared using a procedure close to that described for the previous systems. The compositions and codes of the various investigated blends are reported in Table 4.

Table 4.

Codes and compositions of the UP~rubber blends

UP

PIB

PIB G.D. a

(%)

(%)

(%)

UP

100

-

-

E0

90

10

El.5

90

10

1.5

E2.5

90

10

2.5

E2.5 b 90 aG.D.: grafting degree t'premixed 12 h before curing

10

2.5

Code

In Fig.30 are compared the FTIR spectra in the frequency range 4000-450 cm-I of the plain rubber (curve A) mid of a PIBSA sample having a grafting degree of 1.5 % wt/wt (curve B).

98 0.38 0.35

13~6.19 13~9.15 1470.37 ~

0.30

A

9

0,25

0,151 2000

,42

1800

1600

1400

1200

__222 IlXX)

800

CM-1

ot

1366,14 B

0.

I

A 0.20[ o.l/

IT.~ 1860.31 ~

923.33

1716.43

1600

1400

1200

1000

800

600

4t50

CM-1

Figure 30. FTIR spectra in the frequency range 4000-450 cm-1 of plain PIB (c~erveA) and of PIBSA with a g. d ofl.5 % wt (curve B). The peaks arising from the vibrations of the grafted succmic anhydride groups are found in the curve B of Fig. 30 at 1860 cm -l, 1783 cm d, 1716 cm-l and 701 cm -l In particular the doublet at 1860-1783 cm d is due to the in phase and out of phase vibrations of the anhydride carbonyls, while the 1716 cm -1 component arises from the Vc__o vibration of the of the succinic acid. Its presence indicates that a very limited number oi anhydride underwent hydrolysis. The carbonyl region is free from interfering PIB absorptions and the peak at 1782 cm -l has been used to evaluate the amount of SA bound on the PIB backbone. The quantitative determination has been performed on solutions of PIBSA in CH2C12 by means of the calibration curve reported in Fig.31. This curve was constructed using succinic anhydride solutions in CH2CI 2 as standards. The two PIBSA smnples prepared according to the procedure reported in Ref. 41 yielded a grafthlg degree of 1,5 and 2,5 % wt/wt respectively.

99 i

1.2

I

i

I

i

I

0

0.8

-

0.6

-

0.4

-

0.2

-

110 6

/I

o,,oo

,,oo

,,;0

I

i

I

=

I

fr

1.0

i ~

i

L

[ ,,~o

,,~,o

/

0 o"t d3

,,=I:

0

'

)

I

0.1

'

I

0.2

'

I

0.3

Concentration

'

I

0.4

'

I

0.5

'

0.6

(rag/roLl

Figure 31. Calibration curve for the quantitative determination of grafting degree. The inset displays the peaks usedfor the construcaon of the curve. There are several studies dealing with the grafting of maleic anhydride and other reactive monomers onto saturated and unsaturated rubbers, initiated by organic peroxides. In the case of saturated rubbers it has been shown that the grafting occurs preferentially onto the methylene groups and in particular onto the longer and more regular-CH 2- sequences [42,43]. For unsaturated elastomers such as polyisoprene or polybutadiene [44] the MA addition occurs in the vicinity of the double bond, and preferentially in the a position with respect to the unsaturation. In our case we have used FTIR spectroscopy to establish the preferential site of addition along the PIB chain. It is well known that vinyl groups give rise to a medium intensity absorption in the range 900-890 cm-1 due to an out-of-plane deformation mode of the group. This peak goes completely tmdetected in the spectrum of a thin

100 PIB film (see Fig. 30A) due to the very low concentration of the vinyl end-groups into the PIB sample. However, increasing the sample thickness from few microns up to 0.5 rmn, a well resolved absorption is detected in the expected frequency range (see Fig. 32A).

891

.5

96o

8~o

Wavenumbers

(era-- 1 )

86o

Figure 32. FI"IR spectra in the frequency range 900-780 cm "1 o f plain PIB (curve A), fully iodinated PIB (curve B) and PIBSA having a G.D. o f 2.5 (curve C). The spectra were collected on films 0.5 mm thick.

To confirm the assignment of such a peak to a wagging mode of the vinyl groups, halogenation of the PIB end-groups was performed; the spectral profile of file fully iodinated PIB, reported in Fig. 32B in the frequency range 900-780 cm-~, shows the complete disappearance of file peak at 891 cm-1, thus confirming the assignment. The spectrum of file PIBSA samples having a grafting degree of 2.5 is

101 reported in Fig. 32C. Also in this case the component at 891 cm-1 is absent which indicates that all the vinyl end groups have reacted with maleic anhydride. On this basis the following reaction mechanism may be assumed: I -----~

21-

R !

I"

+

=

W~CI-12--C=CH 2

9

R I

WW~CH--C=CI-12

R

R

I

vW~CH=C--C.H2

o olo

VWWCI..I=C--CH 2

Structure HI may ether interact with a PIB molecule to yield PIBSA and structure I by a chain propagation step, or may terminate by coupling or disproportionation with another radical species. A further possibility also cited in the literature [45] but in our opinion less likely to occur, involves as a first step the direct addition of the primary radical I- onto the double bond: R I"

4-

R

I

1

WY~CI-12--C---~CH2

Structure IV will then add an MA molecule to form: R I

-

102 which in turn will evolve in the same way as structure III. It is noted that the total concentration of double bonds in PIB is 0.405 mmol/g and that the maximum concentration of grafted maleic anhydride, corresponding to a grafting degree of 2.5 % wt/wt, is 0.265 mmol/g. Attempts to increase the grafting degree by varying the reaction conditions (temperature, peroxide and MA concentration) were unsuccessful. Thus only about 65% of the total amount of terminal double bonds of PIB are reactive towards MA addition. The remaining 35% of unreactive double bonds may be assumed to be of the trisubstituted type: CH 3

I

WVV~H=C--CH 3

and their lower reactivity could be ascribed to steric effects. When PIBSA is mixed with UP the following esterification reaction may occur: R ~CH=C--CH

O 2 + OH--~A#

PIB

UP

~

R O I II ~V~-,H=C--CH2--T-C--OH PIB II O

UP

which would yield a diblock copolymer of the type PIB/UP. The process has been investigated by FTIR spectroscopy: the PIBSA/UP reaction mixture was hold at 80°C trader vigorous mechanical stirring. At various time intervals different aliquots were withdrawn and analyzed spectroscopically as solution in CH2C12. Using the appropriate concentration range the succinic anhydride doublet is clearly discerned as a low intensity peak centred at 1860.cm-1 and as a well defined shoulder at 1783 cm -1 of the Vc=o mode of the polyester (see Fig. 33). Fig.33 also shows that the intensity of

the Vas,C=o peak of the grafted anhydride decreases

gradually by increasing the reaction time.

103 1.7

=

l,,

t

I

1783 1.3

m

u .r

!.__

0 r

0.9

B~~

0.6

~ ~

0.2

~ 1873

1851

I 1829

'

~ 1807

/ I

1785

1763

Wavenumbers (cm-1) Figure 33. b-TIR solution spectra in the frequency range 1873-1763 cm"s of the plain UP resin (curve A) and of a E2.5 blend (curves B, C,D). Traces B, C and D refer to premixing times of O, 215 and 800 min respectively. To obtain more quantitative information the spectral data of Fig.33 were analyzed by difference spectroscopy. A pseudo-base line was identified in the frequency range 1873-1763 cm -] as the spectral profile of the neat UP resin (Fig. 33, curve A). Subtraction of such a base line from the spectra of Fig. 33 allowed to eliminate the interference of the Vc_o absorption of the UP carbonyls so as to obtain a fully resolved profile of the anhydride doublet (see Fig. 34).

104

.4-

G r cO

.2-

or_

1900

ld50

Wavenumbers

1800 (cm--1)

1750

Figure 34. Subtraction analysis in the frequency range 1920-1750 cm ~ o f the solutwn spectra of the E2.5 blend Curves A, B, C, D and E refer to premixing times of O, 90, 215, 335 and 800 min respectively.

The per cent anhydride conversion was then evaluated from the intensity of the Vc_o peak at 1783 cmq as:

a(t) =

lOO{[Sa]o-[SA1} [Sa]o

[SA

a(t) = 100{1--~--~-jA' C0l

where [SA] is file concentration of succinic anhydride groups in the reactive mixture, C is file concentration of the reactive mixture in the CH2CI2 analytical solution, and the subscripts 0 and t refer to reaction times 0 and t respectively.

105 The results of this analysis are reported in Fig.35.

I

20

I

....

I

!

I

150 0 0

E

10-

0

0 0

I

I

I

I

I

I

200

400

600

800

1000

1200

1400

Time (mini Figure 35. Conversion o f anhydride groups on PIBSA as a function o f the reaction time in a E2.5 blend Premixing performed at 80~

It is immediately apparent that the maximum attainable conversion is about 15 %; the reaction proceeds at an almost constant rate of 0.034 min "1 up to 280 mm where a conversion of 10 % is reached; at longer times the process slows down considerably and the maximum conversion is attained after about 1000 min. The very slow reaction rate compared with that of low molecular weight analogs (i.e. propanol and maleic anhydride [46]) might be ascribed to the fact that the system under investigation is etherogeneous. Under such conditions the rate limiting step is the diffusion of the SA end-groups towards the surface of the rubbery domains. Moreover the limited yield is likely due to fact that two different hydroxyl

106 functionalities are present as UP end-groups, namely primary and secondary OH groups deriving from the polycondensation of the starting monomers iso-propylene glycol and MA. It is known from the literature [45] that secondary alcohols do not react with MA in the absence of suitable catalytic systems; thus, in our case, only a limited number of

OH end groups, i.e. the primary ones, are able to interact

chemically with MA. The diblock copolymer formed upon reaction of the reactive end groups of PIB and of UP is likely to act as an emulsifying agent in the blend, thus improving the dispersion of the rubbery component in the matrix before cure and enhancing the interfacial adhesion after the cross-linking process. As will be shown in the next paragraphs, interesting results in terms of toughness and morphology may be achieved even with limited yields of such a copolymer. The Young modulus, E, of the neat UP and of the various blend compositions investigated, determined in the flexural mode at ambient temperature and at a crosshead speed of I mm/min, are reported in Table 5. Table 5.

Mechanical and morphological parameters of the various investigated blends.

Code

E

ffy,c

R

rp

~

~a

(N/mm 2)

(MPa)

(I.tm)

(Ixm)

(~m)

UP

2300

85.0

19.6

2.6

-

-

E0

2000

74.0

32.8

4.4

75.0

36.0

El.5

2050

73.5

55.0

7.4

14.0

7.6

E2.5

2050

73.2

73.3

9.9

8.0

4.7

E2.5 b

2000

72.1

88.4

12.0

8.0

3.5

astandard deviation bpremixed 12 h before curing

As expected, for all the materials tested, a decrease of modulus, compared with that of the plain resin, is observed.

107 Moreover such an effect is independent of the type of rubber used as well as of its grafting degree. Typical compressive stress-strain curves obtained at the same loading rate and temperature as for the Young's modulus, are shown in Fig.36.

I

120

I

I

I

100 80-

Q_

3; 1/1 I/)

60-

(D L_

u')

4020-

0

I

I

I

I

4

8

12

16

20

Strain (g)

Figure 36. Compressive stress-strain curves at ambient temperature and at I mm/min. (a)pure UP resin;(b) EO blend; (c) E1.5 blend; (d) E2.5 blend; (e) E2.5" blend *Blendpremixed 12 h before curing. The values of the compressive yield stress,

13c,y ,

evaluated according to the

following equation: P (l-s) o'~,y- Ao

(5)

108 where P is the load, A o is the initial cross-sectional area of the specimen and 6 is the strain of the specimen, are reported in Table 5. The same behaviour as that found for file Young's modulus is observed. The decrease of Oc,y upon rubber addition is a direct consequence of the lower shear modulus of the rubbery phase compared with that of the UP matrix, which prevents the rubbery domains to support a significant amount of the applied stress. The fracture behaviour of both the unmodified and the rubber modified resins has been examined at low and high strain-rate.

1.0 0.8

E

0.6

z U

0.4

0.2

UP

EO

E 1.5

E2.5

E2.5"

Figure 37. Critical stress intensityfactor Kcfor the various investigated blend compositions; (H) low speed tests; (I) high speed tests. The UP resin and the blend containing the unmodified PIB as rubbery phase exhibit low values of K c in both test conditions (see Fig. 37). In contrast, the blends containing PIBSA display a marked improvement in fracture toughness. An analogous trend is observed when the fracture data are expressed in term of the parameter, G~, (see Fig. 38).

109 350 300

250 E

200

u

150

0

100 50

UP

EO

E1.5

E2.5

E2.5 ~

Figure 38. Criacal strain energy release rate Gcfor the various investigated blend compositions; (0) low speed tests; (11) high speed tests. The observed improvement of K~ and G c is found to be dependent on two factors: 1. The grafting degree of PIBSA. 2. The time period during which the UP/PIBSA mixture was allowed to react prior to the curing process (premixing step). In particular we observed an increase in the K c and G c parameters with increasing the grafting degree; moreover, for the same graRmg degree of 2.5 by weight, a further enhancement of toughness is achieved by increasing the premixing time from 60 min to 12 h. The effect of grafting degree on the toughness parameters may be ascribed to the formation of a UP/PIB block copolymer which, acting as an emulsifier, reduces the particle size of the rubbery domains and firmly bonds the two phases together once the curing reaction has occurred; the higher the grafting degree of the rubber, the higher the concentration of the emulsifier in the blend. Analogously, the effect of the

110 premixing time can be accounted for by considering that the conversion of MA and hence the amount of the diblock copolymer formed, increases going from 60 min to 12 h (see Fig. 35). The Dugdale and the Irwin models were used to evaluate the mount of plastic deformation ahead to the crack-tip. The R and rp values, calculated according to eq. 3 and eq. 4, are reported in Table 5. Both the R and rp parameters increase by increasing the g ~ g

degree of the PIB rubber and, for the same grafting degree, by

increasing the premixing time. These results further confirm that the efficiency of PIB as toughening agent depends on the mount of the PIB/UP block eopolymer formed during processing.

Figure 39. SEM micrographs of the fracture surfaces of samples broken at low deformation rate: (4) EO blend; 03) E1.5 blend.

111 To confirm the fracture results a morphological analysis by SEM was performed on the UP based blends fractured at low deformation rates (see Figs. 39 and 40). The UP/PlB blend (Fig. 39A) shows few large cavities with a diameter ranging from 80 to 100 ~

corresponding to the positions of the rubber domains pulled off during

fracture. Morcovor the UP matrix around the cavities appears rather fiat indicating its very limited plastic deformation.

Figure 40. SEM micrographs of the fracture surfaces of samples broken at low deformaaon rate: (C) E2.5 blend; (D) E2.5" blend. For the UP/PIBSA blends (Fig. 39B, and Fig. 40), a substantial reduction of file particle size is observed, together with a more homogeneous distribution of the particle sizes. The fracture surfaces of these samples also display evidences of shear deformation of the UP matrix around the rubber particles.

112

Figure 41. SEM micrographs of the fracture surfaces of samples broken at low deformation rate and subsequently etched with CHCI~ vapours; (,4) EO blend; (B) E1.5 blend To put the above observations on a more quantitative basis, we performed a complete image analysis of the fracture surfaces [47, 48]. To obtain meaningful results by this approach it is necessary to achieve good contrast among the phases. To this end an etching treatment of the samples was performed by CHCI3 vapours so as to completely remove the rubber particles from the fracture surfaces. The analysis was carried out on a large number of images (more than 10); examples of the etched surfaces used are reported in Figs. 41 and 42. The frequency versus particle size histogram of the E0 blend (see Fig. 43A) shows a very broad distribution of sizes with an average diameter D of 75 ~tm and a standard deviation, a, of 36.0. Conversely, when PIBSA is used (see Fig. 43B and Fig. 44), the average particle size decreases of about one order of magnitude and the

113 m

distribution becomes substantially narrower. The D and c values of the various investigated blends are collectively reported in Table 5. The above effects are found to depend on the grafting degree, G.D. of PIBSA: in fact, in going from a G.D. of 1.5 to a G.D. of 2.5 D decreases from 14.0 to 8.0 and a goes from 7.6 to 4.7. It is worth to note that the blends having the same g.d. of the rubber, but prepared with different premixing times, exhibit the same values of D , but a goes from 4.7 to 3.5 by increasing the premixing time.

Figure 42. SEM micrographs of the fracture surfaces of samples broken at low deformation rate and subsequently etched with CHCI3 vapours: (C) E2.5 blend; (1)) E2.5" blend

114 0.25 -

=b, U t,-0 13

o"

0 L_ tt ID

>

4-.

0.20

0.15

0.10

o

nr

0.05

0

I

I

10

20

30

40

50

60

70

80

i

90 100 110 120 130 140

Particle Size {pm) 0.3

0 C;

-

0.2

0" 0 t___

U_

i

d) ,==.,

0.1

4)

I

0

~1

3

6

9

~

I,.I

12 15 18 21 24 27 30

Particle Size (pm) Figure 43. Particle size distribution: (,4) EO blend; (B) El.5 blend

115 0.3

(J r (1) ::3 O"

-

0.2

t._ It (I) =.w .,i

0.1

13E

i

4

6

8

10

=

12

i

=

14

L

16

Particle Size (pm) 0.3

-

O r (I) O"
0.2

it

>, (U

0.1

IT"

- - • i

z

6

8

I

10

~

I

,

12

I

14

~

I

16

Particle Size (pm)

Figure 44. Particle size distribution (C) E2.5 blend; (D) E2.5" blend Thus the net effect of this processing step is simply to induce a more homogeneous distribution of the rubber particles within the matrix, while the average diameter remains unaffected. In turn the narrowing of the particle size distribution produces a limited effect on the toughening properties of the blend. The whole of the experimental results described herein indicate that, in order to achieve a significant toughening effect, it is not necessary that the total amount of rubber introduced in the blend is reactive toward the thermosetting matrix.

116 Even with conversion of the reactive functionalities as low as those obtained after one hour of premixing (see Fig 35), a material with interesting toughening properties can be obtained. This in turn implies that fl~e amount of mterfacial agent needed to produce a morphology suitable to achieve the desired properties is rather low. This same behaviour has been found for thermoplastic polymers such as polyamides [49] and could be considered as a general rule in the toughening of brittle polymeric matrices. Thus an alternative approach to achieve the desired results, wlfich would avoid the very long and impractical premixing step, would be to synflaesize separately the compatibilizing copolymer and to add small amounts of it to the unmodified liquid rubber prior to blending. Preliminary results along this research line have recently been obtained by the authors and they appear to be highly promising.

References

1.

J. M. Margolis "Advanced Thermoset Composites", Van Nostrand Reinhold Co., New York (1986).

2.

R. Edelman, P. E. Mc Mahon, "A New DAP-Polyester Resin for Carbon Fibers", Composite Teclmology Review, 1, N.2 (1979) 7.

3.

E. Kubel, Adv. Mater. Process, ... (1989) 17.

4.

R. Bums, "Polyester Molding Compound", Marcel Dekker, New York (1982).

5.

S. Newman, D. Fesko, Polym. Comp. 5 (1), (1984) 88.

6.

E. Melby, J. Castro, "Comprehensive Polymer Science", Vol. 7, S. L. Aggarwal Ed., Pergamon Press, Oxford (1991).

7.

C.P. Hsu, M. Kinkelaar, L. Lee, J. Polym. Eng. Sci., 31 (1991) 1450.

8.

L. Kiale, Y. S. Yang, L. J. Lee, AICHE Symp. Serv., 84 (1988) 1450.

9.

L. Suspene, D. Fourquirier, Y. S. Yang, Polymer, 32 (1991) 1593.

10. G.A. Grosbie, M. G. Phillips, J. Mat. Sci., 20 (1985) 182. 11. G.A. Grosbie, M. G. Phillips, J. Mat. Sci., 20 (1985) 563.

117 12. E. H. Rowe, 34th Ann. Teclm. Conf. SPI Reinf. Plastics/Compos. Inst. 23B (1979). 13. L. Suspene, J. F. Gerard, J. P. Pascault, Polym. Eng. Sci., 30 (1990) 1585. 14. P. Barlet, J. P. Pascault, H. Santerean, J. Appl. Polym. Sci., 30 (1985) 2955. 15. D. Verch+re, H. Santerean, J. P. Pascault, S. M. Moschier, C. C. Riccardi, R. J. J. Williams, Polymer, 30 (1989) 107. 16. M. Malinconico, E. Martuscelli, G. Ragosta, M. G. Volpe, hat. J. Polym. Mat., 11 (1987) 317. 17. M. Malinconico, E. Martuscelli, M. G. Volpe, Int. J. Polym. Mat., 11 (1987) 295. 18. E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, E. Bertotti, J. Polym. Sci., Part B, Polym. Phys. Ed., 31 (1993) 619. 19. E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, Die Ang. Makromol. Chem., 217 (1990) 159. 20. L. J. Bellamy, "The hlfrared Spectra of Complex Molecules" Vols. I and II, Chapman and Hall, London (1980). 21. N. B. Colthup, L. H. Daly, S. E. Wiberley, "Introduction to Infrared and Raman Spectroscopy", Academic Press, San Diego (1990). 22. E. Butta, G. Levita, A. Marchetti, A. Lazzeri, Polym. Eng. Sci. 26 (1986) 88. 23. A. J. Kinloch, R. J. Young, "Fracture Behaviour of Polymers", Applied Science Publishers, London (1983). 24. A. F. Yee, R. A. Pearson, J. Mat. Sci., 21 (1988) 2462. 25. A. J. Kinloch, S. J. Shaw, D. L. Hunston, Polymer, 24 (1983) 1355. 26. A. J. Kinloch, S. J. Shaw, D. A. Todd, D. L. Hunston, Polymer, 24 (1983) 1341. 27. R. A. Pearson, A. F. Yee, J. Mat. Sci., 21 (1986) 2475. 28. C. K. Riew, A. J. Kinloch, "Toughened Plastics I", American Chem. Soc., Washington, D.C. ( 1993).

118 29. M. Abbate, E. Martuscelli, P. Musto, G. Ragosta, M. Leonardi, "A Novel Reactive Liquid Rubber with Maleimide End-groups for the Toughening of Unsaturated Polyester Resins", in print on J. Appl. Polym. Sci. 30. K. Kato, J. Electron. Micros., 14 (1965) 20. 31. K. Kato, Polym. Eng. Sci., 7 (1967) 38. 32. E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, "A Polymer Network of Unsaturated Polyester and Bismaleimide Resins", in print on Polymer. 33. J. C. Williams, "Fracture Mechanics of Polymers", John Wiley and Sons, New York (1984). 34. C. K. Riew, "Rubber-Toughened Plastics", American Chemical Society, Washington, D.C. (1989). 35. A. J. Kinloch, J. G. Williams, J. Mat. Sci., 15 (1980) 987. 36. D. S. Dugdale, J. Mech. Phys. Solids, 8 (1960) 100. 37. G. R. Invin, Appl. Mater. Res., 3 (1964) 65. 38. J. N. Sultan, F. J. McGarry, Polym. Eng. Sci., 13 (1973) 29. 39. A. S. Wronski, M. Pick, J. Mat. Sci., 12 (1972) 28. 40. R. D. Adams, J. Copperdale, N. A. Peppiat, "Adhesion-2", K. K. Allen Ed. Applied Science, London (1978). 41. M. Abbate, E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, J. Appl. Polym. Sci., 58 (1995) 1825. 42. R. Greco, G. Maglio, P. Musto, G. Scarinzi, J. Appl. Polym. Sci., 37 (1989) 777. 43. R. Greco, G. Maglio, P. Musto, F. Riva, J. Appl. Polym. Sci., 37 (1989) 789. 44. B. C. Trivedi, B. M. Culbertson, "Maleic Anhydride", Plenum Press, New York (1982) p. 466-4 72. 45. J. Sheng, X. L. Lu, K. D. Yao, J. Macromol. Sci., A27(2) (1990) 167. 46. B. C. Trivedi, B. M. Culbertson, "Maleic Anhydride", Plenum Press, New York (1982) p. 78-80.

119 47. S. L. Chan in "Fractography and Failure Mechanics of Polymers and Composites", A. C. Roulin-Moloney Ed., Elsevier Appl. Sci., London (1989), Chap. 4, p. 145. 48. J. Serra, "Image Analysis and Mathematical Morphology", Academic Press, New York (1982). 49. R. Greco, M. Malinconico, E. Martuscelli, G. Ragosta, G. Scarinzi, Polymer 29 (1988) 1418.

This Page Intentionally Left Blank

121 CHAPTER 3

THERMOSETTING

POLYIMIDES

E. Martuscelli, P.Musto, G.Ragosta National Research Council of Italy, Institute of Research and Technology of Plastic Materials, 80072 Arco Felice (Na), ITALY

1. Introduction

Polyimides are aromatic-heterocyclic polymeric resins which cure via crosslinking reactions or linear chain-extension reactions to give high temperature resistant materials [1-4]. These resins are able to mantain their excellent properties at temperatures far exceeding those of highly cross-linked epoxies. In fact, while the maximum use temperature of epoxies is about 200~ at temperatures up to 370~

polyimides can be used

Such an outstanding temperature stability is due to

the aromatic-heterocyclic structure of the polymer backbone:

C•)N_R,

C

C

0

0

il

II

n

where R and R' can be varied. This molecular structure, besides its very high thermal and thermo-oxidative stability, also provides high glass transition temperatures.

122

There are three important classes of polyimide resins: condensation polyimides, polymerization of monomeric reactants polyimides (PMR) and bismaleimides (BMI). The condensation polyimides are generally prepared by reaction of an aromatic tetracarboxyl-dianhydride,

such

as

benzophenone

tetracarboxylic

dianhydride, and an aromatic diamine, such as methylene dianiline: O

O II

O + H2N---~CtI2~NH2

0

BTDA (I)

O II OH--C,

O II ..C,

i-,

O

0

~-~

MDA(II)

O II ..C--OH

O H

H

Polyamic Acid

m

O

II

O

II

O

II J

l

]1 0

II 0

Polyimide

-

The properties of the final product can be controlled by an appropriate choice of structure I and II in the reaction scheme above. At the present time, however,

123 the number of commercially available monomers from which to choose is rather limited. The PMR polyimides have been largely developed at the NASA laboratories [5]. These resins are prepared from monomers which are soluble in low boiling solvents, such as methanol or ethanol, resulting in a significant improvement in composite processing over the condensation polyimides [6]. Finally the bismaleimide resins are fomaed by the reaction of an aromatic or aliphatic diamine with maleic anhydride:

O

O

O

1 O

O

o O

O

R cml be varied to obtain tailored characteristics. The cure occurs through two possible mechanisms, depending on the resin composition. If the diamines are used, there are two steps in the process" the first is a Michael addition reaction of the diamine across the double bond:

O

O

O

O

The second step is the free radical polymerization of the residual double bonds, which is responsible for the formation of the tri-dimensional network. In

124 the case of BMI without dialuines, the cure proceeds via the free radical reaction only. In general polyimides, like all the thermosetting materials, suffer a major drawback, namely their britlleness, which is attributed to the aromatic nature and to the high crosslink density of the network. This problem is particularly relevant for the bismaleimide resins which exhibit the highest crosslink density. Several approaches have been divised to overcome this limitation, among which the most promising are: addition of reactive elastomers [7], Michael addition chain reaction [8,9], copolymerization with aUyl temminated copol3aners, modification with themloplastics [ 10]; in the present chapter some of the results obtained by the first and the last approach will be presented and discussed in detail.

2. BMI toughened by thermoplastics The

bismaleimide resin was

4,4'-bismaleimido-dipheyl methane

and

diaminodiphenyl sulfone (DDS) was used as hardener. The thennoplastic modifier was an amorphous polyetherimide (PEI) commercially avilable as Ultem 1000 from Genaral Electric, having M,,= 12,000 and

Mw=30,000.

The chemical structures of the above blend components are reported below:

II 0

BMI

II 0

PEI: a thermoplastic Poly (ether imide)

!>._

125 The PEI was dissolved in CHCI 3 and BMI and DDS were added to the solution under vigorous mechanical stirring. The solvent was removed under vacuum and the mixture was placed in an open steel mold and cured at 160~ for 7 h. Postcuring was performed for 2 h at 180~ and 4 h at 200~

The codes and

composition of the investigated blends are reported in Tab. 1.

Table 1. Codes and compositions of the investigated BML'PEI mixtures.

Code

BMI

DDS

PEI

PEI

(wt %)

(wt % )

(wt %)

(p.h.r.) a

A0

76.9

23.1

-

A10

71.5

21.4

7.1

10.0

A20

66.7

20.0

13.3

20.0

A30

62.6

18.7

18.7

30.0

.

.

.

.

aparts for 100 parts of BMI.

The fracture behavior of the net BMI resin and of the PEI modified resins was examined under impact conditions using a Charpy instrumented pendulum. The data were analyzed according to the linear elastic fracture mechanics approach. The parameters Ko (critical stress intensity factor) and G~ (critical strain energy release rate) were calculated by means of the equations: =

rrqS

l)

where cy is the nominal stress at the onset of the crack propagation, a is the initial crack length and Y is a calibration factor depending on the specimen geometry. U G =~~BWr

2)

126 where U is the fracture energy corrected for the kinetic energy contribution, B and W are the thickness and the width of the specimen respectively, and 9 is a calibration factor which depends on the length of the notch and size of the sample. The values of 9 were taken from Plati and Williams [11]. Figs. 1 and 2 show the values of K c and G c as a function of blend composition. Both the parameters increase significantly with increasing the amount of PEI in the blend; in particular G c increases three times with respect to the reference BMI matrix. Moreover no reduction of the elastic modulus of the material was observed for all the investigated compositions.

t

0.8

...........

I .........

.

-

f

N

E Z

1.

0.6

-

z; (

,,r 0.4

0.2

. . . . . . . . . . .

0

I .

5

.

.

.

.

.

[ . . . . . .

10

I

i,

15

Composition (PEI Wt Figure 1. The critical stress intensity factor, K c, determined under impact conditionsjbr the BMI/PtH system as ajbnction of blend composition.

i

2O

27 250

. . . . . .

I

. . . . .

I

. . . .

1 .....

200 -

E

150

-

""3 u

100

50

0

r

0

'

"

i

5

. . . . . . . . . .

i

10

.

.

.

.

.

i

15

20

Composition (PEI W t %) Figure 2. The critical strain energy re~ease rate, G c, determined under impact conditions for the BMI/PEI system as a function o f blend composition.

The morphological analysis performed on the fractured surfaces clearly indicates that the system is completely phase-separated [12]. However the f'mal morphology strongly depends on blend composition. At low PEI content (see Fig. 3) small PEI domains (about 2 l.tm), uniformly distributed within the BMI matrix cohexist with very large domains whose size ranges from 50 ~ n to 100 ktm. Upon etching these surfaces with boiling vapours of CH2C12, the small domains disappear while the larger ones are not removed by the solvent but become more porous. It is likely that in these regions a phase inversion has occurred. By increasing the PEI content in the blend the larger domains tend to

128 increase in size and to coalesce, and a very limited number of smaller domains of pure PEI is observed (see Fig. 4).

Figure 3. SEM micrographs of the fracture surfaces of samples broken under impact conditions Jbr A) unetched A 10 blend; B) etched A 10 blend.

129

Figure 4. SEM micrographs of thefracture surfaces of samples broken under impact conditionsfor A) unetchedA20 blend; B) etchedA20 blend. Finally, at a PEI concentration of 30 phr, the morphology of the whole fracture surface closely resembles that of the large domains observed at lower PEI content (see Fig. 5).

130

Figure 5. SEM micrographs of thefracture surfaces of samples broken under impact conditionsfor A) unetchedA30 blend; B) etchedA30 blend. The above morphological analysis gives an insight into the fracture behaviour of this blend system. In the regions where PEI is the dispersed phase, the fracture occurs by brittle failure of the BMI matrix with the PEI domains bridging the crack and delaying its

131 propagation. Conversely, in the regions where PEI forms the continuous phase the failure may occur within the thermoplastic phase and the crack propagates around rather than through the BMI domains. Therefore, yielding of the thermoplastic continuous phase is the main toughening mechanism. These two deformation mechanisms operate simultaneously in the BMI/PEI system; their contribution to the over-aU fracture tougheness depends strongly upon the blend composition.

2. BMI thoughened by reactive liquid rubbers It is generally accepted that the addition of a reactive liquid rubber can improve substantially the tougheness of thermosets like epoxy resins. Only a few papers on the modification of BMI resins with rubbers have been reported so far, whereas rubber-modified epoxies have received considerable attention in the literature [ 13, 14]. St. Clair and St. Clair have reported an increase in the fracture tougheness of nadic-terminated polyimides by addition of amine-terminated silicone rubbers [15], whereas Varma et al. [16] have reported an increase in shear strength of BMI resins by the addition of amine-terminate~i rubbers. Shaw and Kinloch observed an improvement in the fracture energy of BMI resins by modification with carboxylterminated butadiene-acrylonitile (CTBN) rubbers [17]. Unlike with epoxy resins, CTBN rubbers are not compatible with BMIs even at high temperatures (110~ However, it has been claimed that during curing between 170~ and 210~ the CTBN rubber reacts via the backbone double bonds with the BMI resin, although no experimental evidence has been provided to conf'mn such a hypothesis. In any event, at the end of the curing process a microphase separated structure similar to CTBN-modified epoxies, is observed. The main disadvantage of this approach is the marked decrease in the high temperature mechanical properties. The elastic modulus at 250~ shows only 0.21 GPa for the 50/50 BMI/CTBN rubber system, which also indicates a significant reduction of the glass transition temperature. As a result of the oxidative sensitivity of the CTBN rubber, the oxidative stability of

132 the BMI/CTBN system is low. Nevertheless, the fracture tougheness enhancement makes this system interesting for adhesive applications. Recently new bismaleimides containing ether linkages were prepared and characterized [18]. Mixtures of two of these products form an eutectic mixture and become easy to process owing to their low melting point and long pot life. Takeda and Kakiuehi [18] studied the modification of these mixture systems with CTBN rubbers. In particular they investigated the mechanical and thermal properties as a function of the amount of and the composition (aerylonitrile content) of the added rubber. The chemical formulas of the investigated BMI resins are reported below: o

o

6

6

where R is:

4,4'-bismaleimidodiphenylmethane(BDM)

2,2-Bis [4-(4-maleimidophenoxy)phenyl]propane (BPPP)

Bis [4-(3-maleimidophenoxy)phenyl]sulfone(3,3'BPPS)

133 In Fig. 6 are reported the results of flexural tests of a resin modified with CTBN containing 17 % of AN. The base resin was a 50/50 wt/wt mixture of BDM and 3,3'-BPPS. The flexural strength increases with increasing the amount of CTBN up to 50 phr and then levels off. However a marked reduction of the flexural modulus is observed by increasing the CTBN concentration.

12

.

.

.

.

I. . . .

l

i

,,,

I

500 "3'1 --,--.,,

E E

(11) x

c,.

,

- 400"

-

0

0.

- 300

,m, I,,.., 4.,.I

m

- 200

I,.,

x

=_.

r or)

""

3 3

I.I_ !

I

o

20

I

40

CTBN

'

'!

. . . . . .

60

I

. . . .

80

Concentration

I

100

100

'

120

(phr)

Figure 6. Flexural strength and flexural modulus as a function of CTBN concentration. After Takeda and KakiuchL

The fracture energy (Gc) was found to increase with the rubber content while decreased by enhancing the AN content in the rubber (see Figs. 7 and 8 respectively). The AN content had a strong effect on the solubility of the rubber in the uncured resin and, as a consequence, on the final morphology of the cured

134 material. In the absence or at low AN coments (up to ~ 20 % wt/wt) the rubber is insoluble in the resin and the blend shows a microphase separation. At higher AN values, the CTBN becomes soluble and no phase separation occurs upon curing. In these conditions the rubber modifier is far less effective in inducing energy dissipation mechanisms [18]. The significant improvement in tougheness observed for phase separated systems, was interpreted as a consequence of a good interfacial adhesion among the phases. In turn such an effect was ascribed to a chemical interaction occurring between CTBN and the resin matrix.

350

t

!

~,

I

....

I

.........

I

300 250 0r

E

(5

200

-

150 -

7

100 50-

'

0

'

'

i ...................

20

! . . . . . .

40

I

60

. . . . . .

I'

80

'

!

100

120

CTBN Concentration (phr)

Figure 7. Fracture Energy (Go) as a function of rubber concentration. After Takeda and Kakiuchi.

135 The authors [18] determined the concentration of carboxyl end-groups of CTBN prior and after a pemixing step, and they found that such a concentration hardly changed. Therefore, in analogy to what happens for polyisoprene radically crosslinked in the presence of catalytic amounts of bismaleimide, they suggested that the crosslinked network contains linkages between the allylic carbon of the rubber and the maleimide double bonds [19].

1 6 0

'

140

-

,,

,

I

,

I

,

.

l

.,,I

I

.....

E 120 -

0 100

80

.... , 0

'

5

I

10

.....

I

. . . . . .

15

I

.

.

.

20

.

.

I

25

"

'~

30

AN Content (wt ~1 Figure 8. Fracture Energy (G~ as a function of AN content in the rubber. After Takeda and KakiuchL

The authors of the present chapter recently reported [20] on the modification of a bismaleimide resin commercially available as Kerimid FE 70026 (RhonePoulenc) which is a mixture af BMI and tolylene-BMI in the weight ratio 60/40.

136

0

Tolylene-BMl

This resin is partially oligomerized by a diamine to improve its processability. In fact it is a glassy solid at ambient temperature and becomes fluid at about 50~

no free primary amine groups are present in the formulation. The

exact molecular structure of this product is proprietary and this limits to some extent the spectroscopic characterization. Nevertheless Kerimid 70026 has been used in place of other, better characterized BMI monomers, owing to the fact the the rubber modifier can be readly dissolved in such a matrix. The toughening agent was a maleimido-terminated butadiene-acrylonitrile copolymer (ITBN) whose preparation and spectroscopic characterization is reported in the second chapter of this book. Initially a kinetic analysis as a function of temperature was performed on the pure polyimide as well as on a typical blend composition (85/15 wt/wt) in order to investigate the effect of the rubber modifier on the curing kinetics of the resin and to identify possible chemical interactions among the blend components. Real-time FTIR spectroscopy measurements were carried out in the transmission mode, placing a thin film (1 - 5 ~m) of the product in a temperature chamber mounted in the spectrometer. In Fig. 9 are reported the the spectra of the neat Kerimid collected at 160~ in the frequency range 3250 - 2700 cm -1. Two peaks characteristic of the maleimide double bonds are located at 3167 cm -1 and 3100 cm -1. The former has been attributed to the first overtone of the C=C stretching vibration, while the latter is due to the C-H stretching mode of the bismaleimide unsaturation [21-23]. Both the peaks are found to decrease with reaction time, thus confirming the gradual disappearence of maleimidic double bonds. However the 3167 cm -1 peak

137 is too low in intensity to afford reliable quantitative evaluation. The 3100 cm -1 peak has a reasonable intensity but suffers from extensive overlapping with the peak at 3070 (aromatic vC_H). The spectral data were thus analyzed by means of subtraction spectroscopy, whereby the spectrum at time zero is subtracted from those collected at longer times:

/'/']

lD

o

C O JE~

29.31

217

.1 -

3100

0 It) _13

30,37

.05

2'oo

30'00

Wavenumbers

(cm--1)

28'00

Figure 9. Real time spectroscopic monitoring of the curing reaction of the neat Kerimid resin in the frequency range 3300- 2700 ctrr'. Spectra collected at 160~

4 = 4-SF.Ao

138 where the subscripts s, t and 0 denote the absorbance of the subtraction spectrum, of the spectrum collected at time t and of the specmun collected at time zero, respectively. The subtraction factor, SF, allows to compensate for differences in thickness between the spectra collected at times zero and t. Its value is obtained by reducing to the baseline an internal thickness peak, i.e. a peak which is invariant with the reactants' conversion (in the present case the aromatic absorption at 1514 cm-1). It has been found that the thickness variation during the process is very limited, so that SF is always close to unity.

.01

-

_

tO

i

i_ 0

--.01

-

I .......

3 0o 1 .

3200

.

.

.

.

.

.

.

.

.

2!72

2965 .

.

.

.

3000 i

Wovenumbers

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

Ccm-1)

.

.

.

.

.

2800 i

Figure 10. Spectral subtraction analysis of the data reported in Fig. 9.

139 The main advantage in using spectral subtraction in the analysis of kinetic data rely on the fact that peaks not affected during the process (the interfering absorption at 3070 cm -1 in the present case) are compensated for and hence completely removed in the subtraction specmma. Conversely positive absorbances from the zero base line generally reflect molecular structures that are formed during the process, while negative absorbances reflect structures that are lost. The subtraction spectra reported in Fig. 10 in the frequency range 3200 2700 cm -l clearly show the gradual development of a completely resolved negative peak of BMI at 3100 cm -], for which a linear and consistent baseline can be identified. The absorbance values so evaluated can be directly used to determine the BMI conversion as a function of time.

0.6 !I.....

i

i

I

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

!

. . . . . . . .

1

_

180~ 0.4

,f

o ~ o , ~ M o-~:~ o

1 0 C0

_..;

6o'c

o

0.2

f

0

~

I

100

. . . . . .

I

200

I. . . . .

300

I

400

. . . . . . . .

I

500

600

Time (min) Figure 11. The conversion o f bismaleimide double bonds as a function o f the reaction #me.for the neat Kerimid resin. Temperatures as indicated.

140 In Fig. 11 are reported the conversion, (x, versus time curves for the neat Kerimid resin at three different temperatures (160~

170~

180~

All the

curves show an initial linear trend whose slope increases by increasing the temperature. At longer times the reaction rates decrease substantially and the curves approach a plateau region. The fmal conversion increases with increasing the temperature. It is noted that at 180~ the final conversion values do not excede 50 %; this is due in part to the rigidity of the resin, whose Tg increases rapidly with conversion. Thus at o~ values close to 0.5 the glass transition reaches the reaction temperature and the curing process is frozen in. In order to achieve a more complete cure it is necessary to process the resin at temperatures exceeding 200oC.

0.8 .!...................... i ................

0.6

I

~

I ..............

180"0

0

~. . . . . . . .

170"C 0

O_ 0

0.4

1

50 ~ 0

0

0.2

~ f O~r0

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

i

200

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

l ....

400

......... i ......................................[' 800 600

Time (min) Figure 12. The conversion o f bL~maleimide double bonds as a function o f the reaction timeJbr the 85/15 Kerimid/ITBN blend. Reaction temperatures as indicated.

141 When the ITBN component is premixed with the Kerimid matrix, noticeable effects are detected in the curing kinetics. The ot-t curves relative to a blend containing 15 % wt/wt of ITBN at 150~

160~

170~ and 180~ are reported

in Fig. 12.

0.6

. . . . .

I

,

.I

. . . . . . . . . ~. . . . . .

t,

,

I

,

,

160~ A

0.2

tl 0

T

0

I

1

100

200

i

300

!

"

:

400

!

'

500

600

Time (min)

0.6

j

........

i 170oC

t

,

t,,

,

i

.

,

i

0.4

-

0.2

-

0 0

100

200

Time (min)

300

400

142 0.8 I| . . . . . .

I .....

l

....

I ....

,

..

J

..

180'C

0.6 l

B

0.4-

0.2-

0

. . . . . . . . 0

,.

50

. . . . . .

i

....

100

,

. . . . .

150

,

200

.......

i~

250

300

Time (min)

Figure 13. Comparison of the bismaleimide conversion profiles relative to the neat Kerimid resin (curves A) and to the Kerimid/ITBN 85,/15 blend (curves B). Reaction temperatures as indicated.

To highlight the differences the kinetic curves relative to neat Kerimid resin (curve A) and to the blend (curve B) at 160~

170~ and 180~ are compared in

Figs. 13. It is readly apparent that both at 160~ and at 170~ the curing process in the blend is considerably retarded; in particular the initial linear trend occurring in the neat resin is almost completely suppressed in the blend. Moreover, for the blend, owing to the slower reaction rate, a plateau region was not reached in the investigated time range although it appears that the ot-t curves of the blend tend to the same final conversion values of the neat resin. However the above retardation effect is found to depend strongly on the reaction temperature: it decreases at 170~ while at 180~ the situation is completely reversed. Here both the reaction rate and the fmal conversion are higher in the blend than in the neat resin. (see Fig.

13).

143 The retardation effect observed at 160~ and 170~ might be accounted for by a viscosity increase occurring when the ITBN rubber is dissolved in the neat resin. This in turn causes a reduction in molecular mobility of the maleimide functionalities involved in the crosslinking process. As the temperature increases such an effect is reduced because of the decreasing viscosity of the blend system and, more importantly, because of the higher rate of production of primary radicalic species which renders the system less sensitive to molecular mobility effects. The observation that at 180~ an opposite effect is found on both the reactio~ rate and the fmal conversion is not easily accounted for. One possible explanation could be the occurrence of finther reaction steps which are inactive at lower temperatures. These processes involve chemical interactions between the rubber unsaturations and the bismaleimide double bonds and have the net effect of complicating the overall reaction mechanism and of accelerating the chain addition process through which the BMI network is built up. Spectroscopic evidence of the occurrence of such reactions at 180~ is discussed below. The problem of the possible chemical interactions between the ITBN rubber and the thermosetting matrix was investigated spectroscopically. In principle these interactions may occur either through the maleimide-end groups of ITBN or by reaction of the double bonds along the ITBN backbone. The first kind of interaction is not amenable to spectroscopic analysis since the rubber end-groups yield the same signals as the maleimide groups of the matrix, and the contribution of the two species cannot be separated. However, since alifatic bismaleimides are more reactive than their aromatic counterparts, it is very likely that the BMI end-groups of the rubber partecipate to the cross-linking process through which the kerimid network is build up. With respect to the unsaturations present along the rubber backbone, it is well known that, in general, they may have three different configurations:

144 ~9 C H 2 \

/H /C--C\ H CH2-~

---CH2\ /CH2---/C=C,, H

H

trans-l,4

cis-l,4

- - ~ C H 2 ~ H ....... CH 11 CH2 1,2-vinyl These configurations yield three distict characteristic frequencies due to the out-of-plane deformation of the =C-H bond at 730, 970 and 915 cm -1, respectively. The ITBN transmission spectrum reported in Fig. 14 shows the presence of two well resolved absorptions at 968 cm -1 and at 915 cm -1 while the characteristically broad peak at 730 cm -1, distinctive of cis-l,4 configurations is absent. Thus in the ITBN copolymer only trans-l,4 and 1,2 vinyl configurations are present, with the former being largely predominant. The knowledge of the ratio gtrans/l~1,2v [24] allows one to evaluate quantitatively the relative population of the above configurations by using the relationship:

Ctrans C],2v

~ ~

Arrans A1,2v

In the present case it is found that 82.3% of monomeric units are present in the trans-l,4 configuration and 17.7% in the 1,2-vinyl configuration. By using spectral subtraction spectroscopy it has been possible to isolate the specmnn of ITBN in the region of interest from that of the blend. This was accomplished by digitally subtracting the contribution of the matrix from the blend spectrum.

145 J

g68

1

g15

.5 j ~ o'o0

3 0 0 0.

. . . . . . .

2 0 i0 0 .

.

.

.

96o

.

.

1'

0~'0 0

Wavenumbers (cm- 1) Figure 14. The ITBN transmission spectrum in the frequency range 4000 450 cm-'. The inset evidences the frequency range where the characteristic group frequencies of the double bonds occur.

The situation is represented in Fig. 15, where curve A refers to the initial blend specman taken at 160~

while curve B represents the result of the spectral

subtraction; curve C is the ITBN specman at ambient temperature and is reported for comparison. It is noted that traces B and C are almost coincident, apart from a sligth broadening of the peaks in the former spectnnn which is due to a temperature effect. The spectral data of Fig. 15 demonstrate the reliability of such an approach but, in order to be able to monitor the fate of the minor component in the blend an appropriate criterion is needed to choose the reference spectrum to subtract at times other than zero. This is because both the blend and the reference spectra change quite substantially with time due to the crosslinking process, and the rate of change of the two spectra are quite different.

146

"1 968

I .,3 915

(.1 to

-s o

~t .2

10'00 950 900 Wavenumbers (cm--1)

Figure 15. ,4) The B15 blend spectrum in the frequency range 1050- 850 cm'; B) The subtraction spectrum Blend- Neat Resin; C) The ITBN spectrum. Thus, in performing the spectral subtraction analysis we chose to use a matrix speetrtun and a blend spectrum having coincident values of conversion; the situation is schematically represented in Fig. 13" to obtain the difference specmma representative of ITBN at time t 1, the matrix spectrum corresponding to the point r I was subtracted from the blend spectrum corresponding to the point s I. In this way "clean" results were obtained over the whole time range investigated. It is noted however that this approach cannot be used for the kinetic curves taken at 180~

for times higher than 50 min. This is because at longer times the

conversion values of the blend excede the maximum conversion of the neat resin

147 and a reliable reference spectrum is no longer available. In these cases the spectral components at 968 and 915 cm" were separated by a curve fitting algorithm; the details of the calculations are reported in Ref. 23. In Fig. 16 are reported the conversions of the trans, 1-4 unsaturations and those of the 1-2, vinyl unsaturations of the ITBN rubber as a function of the reaction time for the process carried out at 160~

and 170~

(Fig. 16 A) and at

180~ (Fig. 16 B). First it is noted that both at 160~ and at 170~ the conversion of the two unsaturated species is close to zero, which indicates that in this temperature range the double bonds along the rubber are not involved in any kind of chemical interaction.

,I

0.8

0,6

-

0,4

-

0.2

L

I

I

I

I

!

I

m

m -

[]

12 0.0

-~

-0.2 0

"H

~

U

I

1

I

50

100

150

,

I 200

,,

, 250

time (min)

9

J 300

I

!

350

400

450

148 0.8

0,6

I

,

I

-

0.4

-0.2

L

I

.//

o~

~

9o

0.2

0.0

,

Oo~176 ~ 4

I 0

~

o

l

.......... 50

, 100

, 150

200

time (rain)

t~gure 16. The conversion o f the trans, 1-4 and o f the vinyl 1-2 unsaturations o f lTBN as a function o f the reaction time. m,["l= 160~ 1700(7..; 0 , 0

-- 180~

A, A =

The open symbols rejkr to the 1-4 trans unsatura-

tions, the solid symbols to the 1,2-vinyl.

A differem picture is found at 180~

Here, aider an induction period lasting

approximately 50 min. the conversion of the above species starts to increase substantially with an approximately linear trend and reaches values of about 60 % by the end of the process. No evidence of a plateau region is found in the investigated time interval: the process through which the rubber unsaturations are consumed would have continued fimher, perhaps to completition, at longer times. Another relevant observation is that the data points relative to both the types of

149 unsaturations can be accomodated on a single curve, which indicates a similar reactivity of both the trans-l,4 and the 1,2-vinyl unsaturations. At this point it is worth to compare the kinetic profile relative to the BMI double bonds (Fig. 13) with those of the ITBN unsaturations (Fig. 16 B). It is found that the rubber double bonds start to react at about 50 min when the BMI conversion is already high (45 %). At this point the BMI reaction rate has already slowed down considerably, while in the region were it was at its maximum the ITBN unsaturations were inactive. Furthermore in the time range where the reaction rate of the rubber double bonds is steady, the BMI reaction rate gradually reduces to zero; towards the end of the process the BMI conversion remains constant while that of the rubber unsaturations continues to increase linearly. In summary it seems that the above processes are not directly correlated, as if they were following two different and independent reacction pathways. A crosslinking process of the ITBN rubber within phase separated rubber domains initiated by catalitic amounts of BMI and/or by the maleimide end-groups of the rubber would explain this effect. In this instance the induction period would represent the time necessary to reach a critical concentration of radicalic species in the rubbery domains. This kind of mechanism has been demonstrated for polyisoprene radically cross-linked in the presence of small amounts of bismaleimide. On the other hand this mechanism would not account for the acceleration of the curing process of the matrix observed in the blend at 180~ and for the increase of the final BMI conversion compared to that in the neat Kerimid resin. In fact chemical processes confined into the rubbe13, domains of a phase separated system would hardly affect the over-all curing mechanism of the BMI continuous phase. Furthermore a morphological analysis of the fracture surfaces of blends of different composition carried out by scanning electron microscopy [20] did not reveal any evidence of a dispersed second phase, even at very high magnification. Thus the electron microscopy analysis indicated the occurrence of a single phase, homogeneous system upon the curing and postcuring processes.

150 The experimental data just discussed cannot be considered conclusive and further investigations on the molecular structure realized upon curing in such a complex network, possibly employing other spectroscopic techniques, are in order to fully account for the experimental observations. However, at temperatures of 180~ and above, extensive chemical interactions between BMI and ITBN can be anticipated which involve both the maleimide end-groups and the backbone unsaturations of the robber. To test the ability of the ITBN rubber in improving the tougheness of the Kerimid matrix, a series of blend compositions were prepared by first dissolving the ITBN rubber into the resin at 120~ for 30 min, and degassing under vacuum for additional 30 min. At the end of this step a clear, visually homogeneous mixture was obtained. The mixture was then poured in a glass mold, cured at 180~ for 5 h and postcured at 220~ for 2 h. The codes and compositions of the investigated blends are reported in Tab. 2

Table 2. Codes and compositions of the investigated blends" Code

BMI

ITBN

(wt%)

(wt %),,,

K0

100

-

K4

96

4

K8

92

8

K10

90

10

KI5

85

15

The fracture properties of such blends were investigated at low (1 mm/min) and high (1 m/sec) rate of deformation. The parameters K e and G~ were calculated using Eqs 1 and 2.

151 In Fig. 17 the K c values are reported as a function of blend composition; for both the testing conditions K e increases linearly with increasing the rubber content up to a maximum of about two times for a 85/15 blend composition. Similar results are found when the fracture tougheness is expressed through the parameter G e (see Fig. 18). In this case, for the low speed tests, an increase of about three times is achieved with respect to the value of the neat resin for the 85/15 blend composition. As shown in Fig. 19 a modest reduction of the elastic modulus is brought about by the addition of the ITBN component.

|

0.8

!

I

. . . . .

-

A ;

0.6-

w

0.4

0.2

0

I

I

4

8

Composition

'

I

.

12 (ITBN Wt

.

.

.

.

16

~)

Figure 17. The critical stress intensity factor, K c, for the KerimidJTBN blend system as a function of composition at high (curve B) and low (curve A) deformation rate.

152 180

I

,

, I

,I

150 -

E

A

12o -

.,,...

d

9o

60

3O

I

0

I

4

I

8

12

16

Composition (ITBN Wt ~)

Figure 18. The critical strain energy re~ease rate, Gc, for the Kerimid / ITBN blend system as a function of composition at high (curve B) and low (curve A) dejbrmation rate.

I

I

I

3.5

13.

(.~

3.0

LU

2.5

.

0

.

.

.

I

.

.

.

5

.

I

.

10

.

.

I

15

20

Composition (ITBN Wt ~1

Figure 19. The flexural elastic modulus, E, for the Kerimid,TTBN blend system as a junction of composition.

153 The observed enhancement in tougheness is somewhat lower than that obtained by using an engineering thermoplastic like polyetherimide as second component. This illustrates the lower efficiency of the rubbery phase with respect to thermoplastics in toughening densely crosslinked thermosetting materials. In this case the enhancement in the tougheness parameters may be ascribed to the incorporation of flexible rubber chains within the BMI network which render the network more flexible and easier to deform under loading. References

1.

V. Crivello, J. Polym. Sci., Polym. Chem. Ed., 11, 1185 (1973).

2.

I.K. Varma, Sangita, D. S. Varma, J. Polym. Sci., Polym. Chem. Ed., 22, 1419(1984).

3.

J.E. White, M. D. Scaia, Polymer, 25, 850 (1984).

4.

C.E. Browing, "Advanced Thermoset Composites", J. M. Margolis Ed., Van Nostrand Reinhold Co., New York (1986).

5.

T.T. Serafini et al., US Patent 3, 745, 149 (1973).

6.

T.T. Serafini "Status Review of PRM Polyimides", ACS organic Coatings and Plastic Chemistry, 40, 469 (1979).

7.

A.J. Kinloch, S. J. Shaw, Amer. Chem. Soc. Polym. Mater. Sci. Eng., 49, 307 (1983).

.

9.

M. Bergain, A. Combet, P. Grosjean, Brit. Pat. Spec. 1, 190, 718 (1973). H. D. Stenzenberger, US Pat. 4, 303, 779 (1981).

10. H. D. Stenzenberger, W. Romer, M. Herzog, P. Konig, 33rd Int. SAMPE Symp., 33, 1546 (1988). 1 I. E. Plati, J. G. Williams, Polym. Eng. Sci., 1_55,470 (1975). 12. V. Di Liello, E. Martuscelli, P. Musto, G. Ragosta, G. Scarinzi, Die Ang. Makromol. Chem., 223, 93 (1993). 13. J.N. Sultan, F. J. Mc Garry, Polym. Eng. Sci., 13, 27 (1973).

154 14. L. T. Manzione, J. K. Gillham, C. A. Mc Pherson, J. Appl. Polym. Sci., 2__66, 889(1981). 15. A.K. St. Clair and T.L. St. Clair, Int. Adhesion Adhesives, 1, 249 (1981). 16. I.K. Varma, G.M. Fohlen, J.A. Parker, D.S. Varma, in Polyimides, K.L. Mittal, Ed., Plenum, New York, 1984, vol. 1, pp. 683-694. 17. A. J. Kinloch, S. J. Shaw, Int. Adhesion Adhesives, 5_, 123 (1985). 18. Shinji Takeda, Hiroshi Kakinchi, J. Appl. Polym. Sci., 35, 1351 (1988). 19. P. Kavacic, R. W. Hein, J. Am. Chem. Soc. 81, 1190 (1959). 20. M. Abbate, E. Martuscelli, P. Musto, G. Ragosta, "Toughening of a Bismaleimide Resin by a Maleimido-terminated Liquid Rubber", submitted to J. Appl. Polym. Sci. 21. D.O. Hummel, K. U. Heinen, H. Stenzenberger, H. Siesler, J. Appl. Polym. Sci., ~

2015 (1974).

22. C. Di Giulio, M. Gautier, B. Jasse, J. Appl. Polym. Sci., 291771 (1984). 23. S. F. Parker, S. M. Mason, K. P. J. Williams, SpectTochimica Acta, 46A, 121 (1990). 24. Silas, Yales, Thornton, Anal. Chem., 31,529 (1959).

PART II TOUGHENED THERMOPLASTICS

This Page Intentionally Left Blank

157 CHAPTER 4

NUCLEATION PROCESSES IN TOUGHENED PLASTICS A.Galeski l, Z.Bartczak ~, E.Martuscelli 2 ICentre of Molecular and Macromolecular Studies, Polish Academy of Sciences, 90-362 Lodz, Poland

2Istituto di Ricerca e Tecnologia delle Materie Plastiche, CNR, Via Toiano 6, 80072 Arco Felice, Italy

I. Introduction

The properties of a polymer can be extensively modified by a physical mixing with another polymer. The properties of a resulting polymer blend depend on the composition and processing and also on the physical state of each component at the temperature of application of the blend. Over the years most theoretical and experimental investigations of properties of polymer blends have concerned the systems containing amorphous components (e.g.[ 1]), although blends with crystallizable components are receiving increasing attention (e.g.[2]). Below melting points of components the blends with crystallizable polymers constitute heterogeneous systems because the components separate from each other during crystallization. However, the miscibility of the remaining amorphous phases of the components may be still possible. The incompatibility induced by crystallization of components affects the mechanical properties of the blends,

158 causing frequently their deterioration. On the other hand the mechanical properties of blends are determined also by the properties of crystalline phases, including the overall crystallinity, crystalline morphology and the sizes of crystallites and their aggregates such as spherulites. The average spherulite size in blends is a very important factor influencing their mechanical properties e.g. the yield stress and ultimate strength of the material [3]. The size of spherulites is controlled mainly by the process of primary nucleation. The primary nucleation behavior depends on both the material properties and thermal treatment i.e. thermal conditions for processing and crystallization. In the case of blends the primary nucleation depends also strongly on the blend composition. Primary nucleation behavior in blends has been recently widely studied. Below a brief survey of habits and nature of nucleation in polymers is given and the results of investigations of prhnary nucleation of spherulites in polymer blends containing crystallizing components are summarized.

2. Primary nucleation in polymers Crystals in polymers are grown from nuclei rather than formed uniformly over the entire volume of the material. In the primary nucleation phenomenon in polymers three paths for nucleation can be distinguished: (i) homogeneous nucleation which takes place if no preformed nuclei or foreign surfaces are present, (ii) heterogeneous nucleation- the nuclei are formed on foreign surfaces which often reduce the nucleus size needed for stable crystal growth, and thus enhance the nucleation process and (iii) specific only for polymers the self-seeding - the nucleation caused by small polymer crystals which survived melting or dissolution of the polymer sample [4]. The classical concept of crystal nucleation based on the assumption that fluctuations in the undercooled phase can overcome energy bmxier at the surface

159 of the crystal was first developed by Gibbs and later by Kossel and Volmer (see general surveys of nucleation by Zettlemoyer [5] and by Price for macromolecules [6]). The rate of nucleation I* has been derived by Turnbull and Fisher [7] to be

I*=(NkT/h)exp[-(AG*+AG~)/kT]

(1)

where N is related to the number of crystallizable elements, AG~ is the energy of formation of a nucleus of critical size and AG~ is the activation energy for chain transport. The formula derivation is based on the above assumptions using the absolute rate theory. Generally, in polymers as the temperature is lowered from the melting temperature a rapid decrease in AG* and a slow increase in AGn occur causing I* to increase. As the temperature is lowered even further, the decrease in AG* becomes moderate but the increase in AG~ more significant which result in a decrease in I*. Therefore, a maximum in I* exists which is related to the ease with which crystallizable elements can cross the phase boundary. The theories of polymer crystallization are still considerably controversial. The behavior of polymer melts on the molecular level at melting temperature and below it is not yet fully understood and creates problems with interpreting crystallization on higher levels. The basic problem lies in quantification of processes which are only qualitatively understood. For this reason one will fmd in the literature a variety of expressions for a given parameter for polymer nucleation and crystallization. Although the expressions may only differ slightly one has a choice in selecting an expression to use for his own data. The calculations may result in 10-20 % or greater departure from the published data. Since the smallest value of AG* is related to the size and shape of the nucleus in such a way that it has the minimum surface free energy therefore the critical dimensions of the nucleus can be calculated for the anticipated geometry of the

160 nucleus. The expressions for critical sizes of nucleus can be obtained by zeroing the first derivative of AG* (the sum of changes in bulk and surface energies of the nucleus) with respect to the dimensions of the nucleus. Similarly several various expressions can be found for the free energy of fusion Agf. Together with various geometries of the nucleus it gives rise to a range of equations which relate important parameters. Hoffman [4,8-10] presented extensive diverse work in this area. For this reason the expressions derived or used by Hoffinan will be quoted in this review. For example for the case of homogeneous primary nucleation and for rectangular shape of the nucleus the free enthalpy of the formation of nucleus of critical size can be described by the following expression (e.g. see Ref.[4])

AG*=(32(yrr~)/(Agr)2=[32~zco(Tm)2]/[(Ahf)2ta(AT)2]

(2)

where ~ and r are the side and end surface free energies of the crystal,

Agf is the

free enthalpy of fusion of the crystal of the chosen geometry, T ~

is the equili-

brium melting temperature, AT= TO-T, and f=2T/(T ~ +T). The term AGn is usually approximated by the WLF equation for the viscous flow:

AGn/kT=U*/[R(T-Too)]

(3)

Based upon these types of calculations, one will find that the typical homogeneous nucleus dimensions are about 103 to 105 A 3 while a typical polymer chain voltune is about 105 to 107 ,~3. Thus, only a small portion of the polymer chain is involved in forming a nucleus. One of the two types of nuclei - fringed micelle - is thought to be a bundle of polymer chains with long sections remaining uncrystallized. There are restrictions to the fringed micelle formation. As it was shown by Flory [11] the strain generated at the crystalline-amorphous interface by polymer molecules which cross the interphase boundary must limit the nucleus and crystal

161 dimensions. A single molecule, therefore, must fold in order to reach the proper dimensions for the formation of a nucleus and fitrther growth. Thus, the crystal continues to grow in chain folded fashion with constant lamellar thickness. Chain folded nuclei are more probable than fringed micelle nuclei in all cases where the segmental mobility of macromolecular chain is high. Another path of primary nucleation is the heterogeneous nucleation studied extensively by Binsbergen [12-17]. The experimental part of his works concern mostly the heterogeneous nucleation in isotactic polypropylene. Based on the vast number of his own and other authors experimental observations Binsbergen derived a theory of heterogeneous nucleation of crystallization in polymers [17]. The formation of a nucleus on a foreign surface involves a creation of a new interface, similarly as in the case of homogeneous nucleation. However, the preexisting foreign surface greatly reduces the free enthalpy of the formation of a critical nucleus, AG*. This lowers the critical size of the nucleus and results in the formation of heterogeneous nuclei at lower undercooling. Again, assuming rectangular shape of a nucleus lying fiat on a foreign surface one can obtain the expression for the free enthalpy of the formation of a nucleus of the critical size:

AG*=( 16Accyce)/(Agf)2=[ 16AacrCre(T~ )2]/(AhffAT)2

(4)

where Aa is the specific interracial free energy difference for the interface: nucleus-foreign surface. Similarly as for the homogeneous nucleation AG* in the equation (4) is proportional to 1/(AT)2. However, for very active foreign surfaces characterized by low value of A~ the critical thickness of a nucleus approaches the molecular thickness, and the formula (4) transforms to a form

AG*=(4bo~gr Tm)/(AhffAT - Ac Tm/bo)

(5)

162 where bo is the molecular thickness. As At~ goes to zero the formula (5) approaches the enthalpy barrier characteristic for secondary nucleation mechanism which is proportional to 1/AT. The kinetic nucleation theory with chain folding provides now the best general tool for understanding the primary nucleation and the growth of polymer crystals at isothermal conditions from unstrained melt [4]. The reptation concept proposed originally by de Gennes [18] was also adapted for the description of chain motion and transport in the melt [8,19,20]. The reptation theory leads to more accurate expressions for the preexponential factor in I~ and for the activation energy for chain transport, AGn, in eq.(1). It also predicts the dependence of the crystallization rate on molecular weight in different regimes of crystallization [9,10]. For example the appropriate expression for the secondary nucleation process predicted by the kinetic nucleation theory with reptation is as follows [8]:

I=(Nol3gpi)/(aon0exp[-4booa~ T~

(6)

where No is the number of reacting species at the growth front, 13g=(~n)(kT/h)exp[Qo - RT], n is the number of macromolecule segments in the melt, ns is the number of stems of width ao, h is the Planck constant, and QD is the activation energy for reptation. K is a constant usually of the order of unity as determined from the experiments. The most spectacular prediction of the reptation concept concerns the mean time of reeling out from the melt an entire macromolecule with one end attached to the nucleus onto the growing from and pulled by crystallization forces [9]:

t= 1.9xlO9n2 [s]

(7)

163 (n is the number of chain units, other required parameters taken for polyethylene) which is very short (order of 10Zs). For comparison, the time for establishing intermolecular entanglements in a polymer melt is approximately four orders of magnitude longer [21] (intermolecular entanglements were removed by dissolution in a solvent followed by quick evaporation of a solvent and careful drying ). The time for restoring the intermolecular entanglements is of order of tens of minutes for polyethylene melt. These two results point out that macromolecules in the melt are relatively immobile in contrast to a crystallizing macromolecule pulled at one end by attractive forces of crystallization. It is also clear in the view of those results that the nucleation of new crystalline layers on an existing crystal is the controlling factor of the crystal growth. Direct evidence for further nucleation step beyond primary nucleation was brought by Wunderlich and Cormier [22] from observation of crystallization of linear polyethylene melt seeded with extended chain crystals. The observable crystal growth is a result of two processes, the first being the nucleation of initiating stems on the surface of the crystal, and the second being the coverage of the surface by new stems beginning at the initial stem. It must be emphasized here that the crystal grows macroscopically in the direction normal to its surface while on the molecular level the elementary growth mechanism is the growth along the crystal surface.

3. General remarks on primary nucleation.

From the expressions for AG* for homogeneous (eq.(2)) and for heterogeneous (eqs.(4) and (5)) nucleations the constant nucleation rate in isothermal conditions is expected. However, in polymer samples there are usually heterogeneous seeds with a broad specmma of Act values resulting in various nucleation rates I*. Also a limited number of those seeds present in samples leads to differentiated

164 exhaustion of particular fraction of nuclei. Moreover, the self-seeding gives rise to an almost instantaneous nucleation. All those attributes of nucleation events cause that the real nucleation process in polymers is a complicated function of time, not just a temperature: I'=I*(T(t),t). The theories of homogeneous, heterogeneous and self-seeded nucleations describe the mechanisms and show the tendency but hardly predict the real habit of nucleation in a given polymers. Hence, the experimental methods of determination of nucleation are of particular importance. The knowledge of nucleation data is often essential for controlling physical properties of polymer; mechanical properties depend to a great extend on the spherulite average size, size distribution and the size of so-called "weak spots" defects of spherulitic structure including cavities and frozen stresses which resulted from volume contraction during crystallization [23,24], all determined by the primary nucleation process. For some applications it is sufficient to determine only the total number of nuclei activated during the crystallization. The simplest way of obtaining this value is from the average spherulite size for samples which are filled with spherulites. The average spherulite size can be obtained from the first moment of a size distribution or of a distribution of chord intercepts with spherulite boundaries [25]. Other average spherulite sizes can be obtained from higher moments of spherulite size distribution. The higher moments of spherulite size dislaibution can be determined on the basis of direct characterization of spherulite patterns (under polarized light microscope, under scanning or transmission electron microscope of thin films or sections of bulk samples - second or third moment of the spherulite size distribution), on the basis of the small angle light scattering (fourth or fifth moment) (e.g.[26,27], also [28]) or of the light depolarization technique (second or third moment) [29,30]. However, if the time dependence of activation of nuclei is required other methods must be used. The data on time distribution of primary nucleation are usually

165 obtained by direct microscopic observation of a crystallizing sample. The odds of this method are the necessity of using thin samples and the condition of crystallization allowing for counting the spherulite centers. Those limitations can be overcome by applying a method of reconstructing the sequence of nucleation events from shapes of spherulite boundaries in already crystallized films [31] and in thin sections for bulk samples [32]. The time lag between the nucleation of two neighboring spherulites can be found from the curvature of their common boundary and this procedure repeated for a chain of neighboring spherulites delivers the data on the time distribution of the activation of nuclei. The time distribution of activation of nuclei should be expressed in number of activated nuclei per volume unit of untransformed fraction of the sample. Calibration of the time axis in that method is made by measurements of the spherulite growth rate. Usually, a given brand of polymer is characterized by an average number of primary nuclei per volume unit at certain crystallization conditions. The average spherulite size in bulk is determined by the number of nuclei per volume unit. For thin films, however, the spherulites as seen in plane are larger. For the thickness of a sample below the average spherulite size in bulk, the thinner the sample the larger the spherulites. The apparent increase in spherulite sizes in thin films results from the constant average number of nuclei per volume unit. The factor complicating this simple relation is the nucleating ability of sample surfaces. The Avrami type of analysis is often erroneously applied for obtaining the nucleation data from differential scanning calorimetry isothermal crystallization experiments and from dilatometry. The reason for this is that the conversion of melt to spherulites is assumed to follow pure sporadic or pure instantaneous modes, ~e .only modes described correctly by. the Awami equation"

ct(t)= 1-exp(-Kt" )

(8)

166 where c~(t) is the degree of the conversion of melt to spherulites. However, the general form of the equation for conversion kinetics of melt to spherulites for isothermal experiments, which was first developed by Avrami is as follows:

a(t)- 1-ex(-pG2

a(t) - 1 - ex

fo I ( t ) ( t - t ) 2 dt I

- -~ pG 3

(t)(t- 0 3 dt

for films

(9a)

for bulk samples

(9b)

where G is the spherulite growth rate constant at a given temperature of crystallization and I(t) is the rate of nucleation. It is evident that eqs.(9a) and (9b) that for I*=IoS(t) (5(0 being the delta Dirac function; this represents the instantaneous mode of nucleation) one obtains in the 'Avrami' exponent t2 and t3 *

c

for two and three dimensional cases, respectively, while for I =Iot, (c being an integer number, sporadic nucleation for c=0) the 'Avrami' exponent is proportional to tr and tr for the two and three dimensional cases, respectively. Therefore the plot of ln[-ln(c~(t))] vs. ln(t) as required for the Avrami type of analysis can lead to a straight line only in two limiting cases: instantaneous and sporadic nucleation modes, both very seldom in a plain form in crystallization of polymers. The proper way of data analysis is by solving the integral equation which follows from eqs.(9a) and (9b)

~2 I(t)(t- t) u dt = - 1 / ( p G u)ln[1- a(t)]

for films

(10a)

~~ I ( t ) ( t - t ) 3 d t = - 3 / ( 4 p G 3)ln[1 - a(t)]

for bulk samples

(10b)

167 The simplest way of solving it is by Laplace transformation, or equivalent, by third (for films) or forth (for bulk) order differentiation against time, t: I(t) = -1/(2pG 2)d3 [ln(1- a(t))]/dt 3

for fihaas

(1 la)

I(t) = -1/(8pG 3)d4 [ln(1 - a(t))]/dt

for bulk samples

(11b)

4

In the literature there are many examples of the treatment of the problem of nonisothermal solidification ( for a broad review see the paper by Wasiak [33]). Most of them are lacking in a firm theoretical background. However, the probabilistic approach to the description of spherulite patterns [34,35] provides now a convenient tool for the description of the conversion of melt to spherulites. In the case of nonisothermal crystallization the conversion of melt to spherulites is described by [34,35 see also 36]:

a(t) = 1-exp{-pf~ I(t)If~ G(s)ds] 2 dt}

for films

(12a)

a(t) = 1-ex~-(4p/3)~~ I(t)I~~ G(s)dsl 3 dt}

for bulk samples

(12b)

Since the growth rate is unambiguously determined in all three regimes of

crystallization by secondary nucleation process and completion rate of the nucleated layer it could be precisely meast~ed in isothermal experiments in thin films as a function of temperature and could be then easily transformed to the function of time provided that the temperature change is monitored during

168 nonisothermal solidification of a polymer. The solution of the eqs.(12a) and (12b) is also by Laplace transformation, or equivalent, by differentiation [36]:

d { 1/G(t)-~d I 1/G(t)-d-d~-{ln[1-a(t)]}l I for films I(t)= -1/(2P)-d-~

{

dE d{

(13a)

dI,n ,a 0 jlltforb mp,es ,3b

I(t)=-l/(8p) d 1/G(t)-d-~ 1/G(t)-d-~ 1/G(t)~

In all above equations the conversion degree of melt to spherulites must not be mistaken for the crystallinity degree. The difficulty in all nonisothermal experiments is that the fractions of the material crystallized at different temperatures differ in the degree of crystaUinity. It is usually considered that nuclei are spread randomly over the sample, except for nuclei formed on outer surfaces in three dimensional samples, on surfaces allowing for transcrystallinity and on surfaces of the second dispersed component e.g. short fibers. However, that is not necessarily true. If the nucleation events occur not only at the very beginning of crystallization but also later during crystallization process then the volume occupied by already advanced spherulites are excluded from further nucleation. The close vicinity of an arbitrarily chosen nucleus is then poorer in other nuclei than more distant regions. Although the nucleation itself is a spatially random process it is limited to the uncrystallized portion of the sample. Such an exclusion always produces a kind of a distance correlation, if the nucleation process is prolonged in time. The spatial correlation of spherulite centers was first observed by Misra, Prud'homme and Stein [37] while respective mathematical formulas for the description of the distance correlation of nuclei for model modes of primary nucleation were derived in Ref.[35] and [36] for isothermal and nonisothermal cases.

169 4. Primary nucleation in polypropylene.

Isotactic polypropylene is now very frequently used as a base for many blends. The world production of polypropylene exceeded in 1986 8 million ton and about 60% of this production went into blends with other polymers. These numbers show an increasing tendency (e.g.[38]). Therefore, the nucleation habits of isotactic polypropylene, being very important example of crystallizing component of polymer blends, are revealed below. Polypropylene has been shown to exhibit several crystalline modifications [39-42] in addition to the most common ot monoclinic structure reported first by Natta et al.[40]. The [3-phase crystallizes from primary nuclei in the hexagonal fashion, although, the [3-phase can also be initiated along a growing front of the a-phase in the temperature gradient. The T triclinic phase was first discovered in low molecular fractions of isotactic polypropylene solidified by slow cooling [41], in commercial polypropylene crystallized under high pressure [42] and in samples of propylene-ethylene copolymer at a low ethylene content which exhibited a peculiarity of a complete crystallization in the T-form [43,44]. In the literature there are no reports on primary nucleation of the T-form spherulites, however, the 7-form is reported to exhibit spherulitic lamellar structure and undergoes the y-or transition on annealing at 147~ and 1 atm [45].

4.1 Nucleation of a form of isotactic polypropylene.

First consistent microscopic data on primary nucleation of isotactic polypropylene were obtained by von Falkai and Smart [46] and also by von Falkai [47]. Their data for samples melt annealed at 180~

prior to crystallization and then

crystallized isothermally in the temperature range from 122 to 145 ~ presented in Table I.

are

170 Table h Basic crystallization parameters o f isotactic polypropylene. ~

Crystallization

Nucleation

Growth rate

temperature [~

density [ 106 cm3]

122.0

2.36

18.0

125.0

1.56

12.0

127.5

1.02

7.0

130.0

0.85

4.3

132.5

0.73

2.6

135.0

0.65

1.6

138.0

0.58

0.86

140.0

0.53

0.59

145.0

0.47

0.27

[taroJmin]

~Data from von Falkai and Smart [46] for isotactic polypropylene of Mw=51 200, heptane extracted, melt annealed before crystallization at 180~ for 15 min

The nucleation in these experiments in polypropylene was found to be heterogeneous and instantaneous with calculated Avrami exponent closely matching 3. (There is somewhat unclear point about the three-dimensionality of their samples used for microscopic observation of crystallization). Since then many authors published nucleation and crystallization data for isotactic polypropylene (e.g.[29,30,48-52], see also [53]). The change from instantaneous to sporadic character of primary nucleation is reported in the literature when the samples were heated up to 200~ and above

171 [48,51]. Under such condition the polarized light microscopy examination revealed an initially constant nucleation rate, however, decreasing for longer crystallization time. The Avrami type of fit to integral exponent was not satisfactory in this case. The course of primary nucleation in isotactic polypropylene down to 70~ was first demonstrated by Burns and Tumbull [54] and by Koutsky, Walton and Baer [55] employing the droplet technique developed originally by Vonnegut for tin and water droplets [56]. One can recognize four distinct regions in the nucleation of isotactic polypropylene melt: (i) immediately below the DSC determined melting point (165-167~

there is a

gap where the crystal nucleation and growth hardly takes place. Neither the present heterogeneities nor introduced nucleating agent can accelerate the nucleation. (ii)Most of published nucleation data concern the region of temperature below ~50~

but above 115~

where regular spherulite are nucleated (see

e.g.[57,58]) (although Binsbergen and deLange [59] observed negatively birefringent sheaves of crystalline lamellae crystallized isothermally at as high as 160~ which is well above Regime II - Regime I transition temperature estimated for isotactic polypropylene at 155~

[58]). This region is the

extended region of activity of heterogeneous nuclei. The number of heterogeneous nuclei is limited during the crystallization. (iii) Some of the heterogeneous nuclei become active at even lower temperature which follows from their smaller size or lower perfection. These nuclei are also limited in nmnber. (iv) Finally at approx.80-85~

and below there is the region of homogeneous

nucleation. The number of nuclei in this region increases rapidly with the decrease of the temperature.

172 Except for very thin specimens, it is difficult to reach beyond the upper range of activity of heterogeneous nuclei due to the low thermal diffusivity of polymers, intense nucleation and fast spherulite growth at lower temperature. Also the latent heat of fusion liberated by the rapid crystallization during quenching tends to maintain the temperature during crystallization in the upper range of activity of heterogeneous nuclei. This and the instantaneous character of most of heterogeneous nuclei cause that homogeneous nucleation range is rarely reached and many of polymeric objects in technological applications crystallize only from heterogeneous nuclei. Figure 1 illustrates the nucleation activity in isotactic polypropylene (RAPRA, iPP1, Mw=3.07*105, Mn=l.56*104.. density= 0.906 g/cm 3, melt flow index=3.9 g/10 min). The data were taken from Refs.[60-62] and differentiated to represent the contribution of new nuclei activated by the temperature decrease by 1~ (data for the crystallization temperature of 90, 100 and 110~ were obtained by the authors for the purpose of this review employing the method of crystallization described in Ref.[62]). The smnples in a fonn of thin films (20-30 ~ n ) were crystallized isothermally on a microscopic hot stage for the temperature above 115~ while for the temperature below 115~ the samples were obtained by isothermal crystallization in a specially designed crystallization cell enabling to reach isothermal conditions within the sample volume in less than 0.5s. It is seen that the number of nuclei increases initially as the temperature of crystallization decreases. At the temperature of crystallization of 132~ a change of slope of the AI/AT vs. T curve is seen which is apparently associated with the regime II- regime III transition in crystal growth kinetics (reported to be at 135 - 137~ for other brands of isotactic polypropylene as determined in Ref.[10] on the basis of the data taken from Refs. [12], [46] and [63], see also Ref. [58]). The regime IIregime III transition is in fact the change in the intensity and the habit of secondary nucleation which may be considered as heterogeneous nucleation on the polymer

173 crystal surface; similar though not identical transition (slightly different temperature of the transition) should be expected for nucleation on surfaces of other heterogeneous nuclei. Further decrease of the crystallization temperature below 115~ results in saturation of the AI/AT value. Apparently all heterogeneous nuclei present in the sample are able to show up within the time of crystallization below 115~ At the temperature below 85~ a new intense process of homogeneous nucleation takes place. A rapid increase in number of formed nuclei with the decrease of crystallization temperature is observed (see Fig. 1). Early droplet experiments during isothermal crystallization [54,55] also showed that the nucleation in droplets of isotactic polypropylene is thermally activated and the droplets crystallize sporadically in time. Investigations of nucleation performed during continuous cooling could, however, resolve only a singular large peak of nucleation at one particular undercooling. Annealing of the melt has a great influence on primary nucleation in isotactic polypropylene. However, the knowledge of the behavior of primary nucleation during melt annealing in polypropylene was acquired gradually as the understanding of nucleation processes in polymers became better. First extensive study of the effects of thermal history on crystallization of isotactic polypropylene was conducted by Pae and Sauer [64] and Sauer and Pae [65]. Further studies included direct microscopic observations of the formation of spherulites in polypropylene melt subjected to various thermal treatment. Annealing of polypropylene melt prior to crystallization decreases the active fraction of primary nuclei. The crucial factor is the temperature of melt annealing below or above the equilibrium melting temperature, T ~ . At 190~ a vast number of those thermally sensitive nuclei remains untouched while even short exposure to temperature around 220~ decreases the number of active nuclei by orders of magnitude.

174 107

"

L

~

.

.

.

.

.

.

~

'

[

'

10 6 05

10 4 I.....I

[.~ <

03

z

lO 2 1(] 1

9

-J

80

9

,,A,

,

~ ........

100 TEMPERATURE

Fig. 1

l .........

120

[~

J

,

1

140

Nucleation activity in isotactic polypropylene OPP1 by RAPRA). The number of nuclei are differentiated to represent the contribution o f new nuclei activated by the temperature decrease by I~

(o) Data

obtained in isothermal crystallization on a microscopic hot stage; (e)nucleation data in isothermal crystallization in a specially designed crystallization celL

(There is still some controversy about the proper value of the equilibrium melting temperature for a-phase of polypropylene which follows from various extrapolations of melting data; if one applies higher crystallization temperature range for extrapolation the equilibrium melting temperature was found at: 208~ Fatou [66], 220~ Samuels [67] and 208•176 Monasse and Haudin [58]. Mucha [68] found 212~ independent of the crystallization temperature range from the extrapolation of SAXS data of lamellae thickness.) The nuclei disappearing due to

175 annealing are thermally activated heterogeneous nuclei and so-called "self-nuclei" [69]. Self-nucleation is a general term describing nucleation of a melt or solution by its own crystals grown previously. Self-nucleation in polymers is particularly strong because of large temperature range where crystals do not melt entirely. There are suggestions that most of previous observations of the behavior of primary nucleation concerns the self-seeding. The equilibrium melting temperature must be exceeded on melt annealing in order to remove entirely the self seeded nuclei from the melt. A technique of self nucleation was developed [70] based on the observation that the critical nucleus size decreases with decreasing temperature. After melting the temperature is reduced below the melting point, preferably 1015~ higher than the required crystallization temperature, and kept for a period of time for producing embryos. Then the temperature is lowered to the crystallization temperature at which most embryos reach the critical size and become stable nuclei. Using this technique one can increase the number of nuclei by few orders of magnitude and significantly reduce the crystallization time. The other source of thermally sensitive nuclei are polymer crystal fragments in small cracks and cavities of a foreign surface which survive melting because of an increase of their melting temperature due to stiffening effect. At moderate undercooling they originate the growth of spherulites. The structure of early stages of the growth of spherulites in polypropylene was studied by Bassett et al. [7173]. The nucleus is built from more or less regular lamellae showing a multilayer arrangement. Early objects develop by branching, usually at rather large angles, and splaying apart the dominant lamellae. A sheaf-like center is usually seen for the crystallization temperature above 155~ or a cross-hatched structure when viewed in plan for the crystallization temperature 155~ and below. The change of morphology of spherulite centers at 155~ is apparently connected with regime I/II transition expected around 155~

The multilayer arrangement of the very center

176 of a spherulite is explained by Bassett [72] as the result of a shish-kebab type of nucleation on straightened fragments of macromolecules which can always be found in a polymer melt. The conclusion is based on electron microscopy observations of nucleation in isotactic polystyrene. It means that the core of a nucleus is build of a single folded macromolecule or a bundle of elongated macromolecules. The elongated fragment extends as far as several lamellae thickness i.e. 500 -1000 A. The volume of such nuclei agrees with the estimation of the volume of a homogeneous nucleus based on calculations of the critical nucleus size. However, such elongated shape of the nucleus is inconsistent with the intuitive assumption that the nucleus is limited to a single lamella thickness which was made in most calculations and modeling concerning the primary nucleation.

4.2 Nucleating agents for the a-form of isotactic polypropylene. The crystallization of isotactic polypropylene from melt could be enhanced in the region of the temperature where heterogeneous nucleation is observed by adding some extra heterogeneous nuclei. The interest in such experiments was stimulated by industrial efforts to decrease the size of spherulites for improvement of optical and mechanical properties. By adding fmely subdivided foreign material it was shown that solids, liquids, and even gas bubbles are able to nucleate polypropylene spherulites (for the list of older patents see Ref.[74]). However, Binsbergen [ 14] has found in contrary to some patent claims that most inorganic salts and oxides are inactive in nucleating of polypropylene. Beck and Ledbetter [75], Beck [76] and Binsbergen [12,14] tested a large number of substances for their possible nucleating effect on the crystallization of polypropylene. The list of most active nuclei for isotactic polypropylene contains: sodium tertiary butylbenzoate, monohydroxyl aluminum p-tertiary butyl benzoate, sodium pmethylbenzoate, sodium benzoate, colloidal silver, colloidal gold, hydrazones,

177 aluminum salts of: aromatic and cyclo aliphatic acids, aromatic phosphonic acids, phosphoric acid, phosphorous acid, several salts of Ca +2, Ba +2, Cu +2, Co +2, Ga .3, In+3, Ti +4 and V +4 with the above mentioned acids, and also air bubbles. There are also other reports on nucleation activity of certain seeds: indigo [77], talc [78], certain crystallographic planes of calcite (because of alternating polar-nonpolar rows on some cleavage planes) [79], various acetals [80] and some others. Good nucleating agents are insoluble in the polymer or need to be crystallized before crystallization of polypropylene. The important feature of a good nucleating agent appears to be the existence of alternating rows of polar and nonpolar groups at the seed surface. It should be mentioned here that the best cleavage planes of seed crystals are not necessarily the crystallographic planes exposing alternating polarnonpolar rows (if they do exist). The extensive data on nucleating agent of crystallization of polypropylene do not match completely satisfactory to the theory of heterogeneous nucleation which is based on the surface free energy consideration (compare eqs.4 and 5). Crystalline lattice matching type of epitaxy as the major mechanism is also excluded because of large variety of nucleating seeds. For fiat surface the nucleation activity of seeds with a finite nucleation rate is expected rather than their instantaneous activity as observed experimentally. Self nucleation on the surface, in cracks and steps by residual polypropylene crystals can be also excluded because of the constant number of active seeds independent of melt annealing in the range 175280~

The heterogeneous nucleation on steps of seed crystal surface is able to

explain the observed behavior if a proper distribution of step length is assumed and if A~ in eq.(5) is sufficiently small. The reason for that may be a good accommodation of polypropylene crystals on surfaces of seeds. Binsbergen [13] introduced an accommodation coefficient ~ (0<~<1) accounting for reduction of interfacial free energy by increased epitaxy through lattice matching and other effects. He proposed that the existence of alternating rows of polar and nonpolar

178 groups on the surfaces may be the cause of a large value of the accommodation coefficient. The presence of large quantity of a filler in polypropylene changes drastically the crystallization conditions. Some active fillers like talc, aluminum oxide etc. act like strong nucleating agents while inert fillers like chalk show little nucleating ability. Because of strong nucleating activity of active fillers the crystallization of bulk filled polypropylene occurs at higher temperature during quenching. Usually the presence of a filler increases the thennal diffusivity of the composition. It enables for inert fillers to reach lower crystallization temperature.

4.3 Nucleation of ~-modification of isotactic polypropylene. A more highly negative birefringent spherulites of 13 phase crystallize in a hexagonal unit cell [39] rather than in monoclinic fashion as for ot form [40]. The nucleation of 13 form occurs much more rarely in bulk samples than the predominant ~ form. Padden and Keith [57] observed spherulites of 13 form sporadically during crystallization in the range from 128 to 132~

Since then the

conditions of the formation of ~ form were intensively studied. It was found that the 13 fonn is nucleated preferentially in the presence of shearing forces [81]. Rapid quenching was used to produce the 13 form of polypropylene in larger quantities. In contrast to this Shi and Zhang [82] and also Varga [83] reported an interesting observation that the amount of ot phase could be suppressed by slowing down the cooling rate to below 5~

In this way rather pure 13 form can be

obtained. Lovinger et al. [84] showed that a large amount of 13 phase can be obtained by crystallization in a temperature gradient. Although the primary nucleation of 13-phase spherulites is extremely rare in this case, the [3 phase is easily initiated by the growth transition along growing front of or-phase spherulites.

179 The habits of 13phase formation was studied extensively by Varga [83,85-90]. For example, he has found that certain brands of polypropylene, those having high molecular weight, are more susceptible to 13-crystallization. It was established that the nucleation of 13 form is instantaneous, and that the density of nuclei decrease with increasing temperature of crystallization [90]. Those observations and some others (e.g.[91-95]) showed clearly that the 13 form generally occurs at a level of only few percent unless certain heterogeneous nuclei are present [96] or the crystallization occurred in a temperature gradient [84], or in the presence of sheafing forces [97]. Leugering [98] demonstrated that a certain quinacridone dye known as permanent red E3B is very effective in generating spherulites of 13 form for isotactic polypropylene crystallized below 130~

However, its effectiveness depends on

nucleant concentration, dispersion and the cooling rate. Since then a series of other crystalline substances have been found to nucleate 13 form in isotactic polypropylene:

2-mercaptobenzimidazole,

phenothiazin,

triphenodithiazine,

anthracene and phenanthrene [99], pimelic acid [100]. In literature there are suggestions that the observation of the 13 form during slow crystallization results from the retardation of one of sub-stages of a multistage process leading finally to the ct form [101 ]. The 13-ct transition observed at 145~ [102] is supposed to be the confirmation of that concept.

5. Secondary nucleation in isotactic polypropylene. After completion of a folded layer on the surface of the crystal a new surface nucleus must be created for further growth of the crystal. This is called a secondary nucleation process. The most widely accepted expression for the secondary nucleation rate is given by eq.(6). The f3g factor in eq.(6) is the retardation factor because at a large undercooling polymers become very viscous

180 and the reptation is retarded. The temperature dependence of the secondary nucleation rate at low and moderately high undercooling is determined by AG* which is proportional to T~ The completion rate of the layer nucleated on the surface of the substrate, g, is expressed as a difference of attactunent and detachment rates of folds to the nucleus on the surface of the substrate. Fig.2 illustrates the formation and the growth of the surface nucleus as it spreads in the direction parallel to the surface of the crystal. After the formation of a stable nucleus (in the case illustrated in Fig.2 for embryos with more than four stems i.e. for the negative free energy of formation) the layer will be completed with new stems by the attachment detachment mechanism. The expression for the completion rate derived by Hoffman [4] is as follows:

g=ao Q 13gexp[-2aobo~r

(14)

qt being the fraction of the free energy of fusion for the forward reaction and Q being the factor of order of unity. As it is seen from the eq.(14) the completion rate is not a strongly dependent function of temperature. There is a competition during crystallization between the secondary nucleation and the completion of the layer which is nucleated by the secondary nucleation on the substrate. Three cases can be distinguished: (i). the secondary nucleation process is slow allowing for completion of the nucleated layer before the next event of the secondary nucleation, (ii). the secondary nucleation events occur before the completion of the nucleated layer, and (iii). the secondary nucleation occurs so often that it does not allow for the completion of the nucleated layer with new folds.

181

Z

o

Bl

:E B

o

It. h.

o >. W Z W W W L~

0 0

I

2

3

4

5

6

NUMBER OF STEMS

Fig.2

Free energy o f the formation o f a chain folded surface nucleus by the attachment-detachment mechanism. Reaction A is the attachment o f a new stem while reactions B and B1 are the detachment o f a stem [51].

The above three cases are the reason for the existence of three regimes of crystallization and respective changes in polymer morphology. The transitions to different reghnes is possible because the completion rate, g, depends on temperature but much less than the nucleation rate, I. The dependence between the secondary nucleation rate, I, and the observable growth rate, G, is described by a basic relationship: for Regime I with low nucleation rate allowing a rapid completion of the entire substrate length [ 103]

Gi=bolL

( 15a)

for Regime II with the completion rate of the layers, g, allowing for multiple nucleation on the substrate [104]

182 Gu=bo(ig)1/2

(15b)

for Regime III for which the crystallization occurs mainly by the intense nucleation of new stems on the substrate rather than the completion of layers across the surface of the substrate [105]

Gin=boiL '

(15c)

where L is the length of a growth strip of thickness bo, g is the completion rate of a strip and could be identified in tenns of the nucleation process g-a(A-B), A and B being attachment and detachment rates, respectively, L' is the effective substrate length [105] which corresponds to ~(2-3)ao. In general L' is considerably smaller than L. The data for growth transitions in polypropylene obtained by Clark and Hoffman [10] and by Monasse and Haudin [58] are included in Table II. The secondary nucleation can be easily determined from the measurements of the spherulite growth rate based on the knowledge of the basic crystallographic and thermodynamic parameters characteristic for a given polymer. The necessary data for the description of processes involved in crystallization of isotactic polypropylene are collected in Table II.

Table 11: Basic parameters for the description qf nucleation and c~stallization of isotactic polyprop.ylene.

equilibrium melting point, Tm

208~

enthalpy of fusion, Ahf

209.3+_29.9 J/g [106]-

chain conformation

3~ helix [40]

[58,66]

183 ot form unit cell parameters [40]: unit cell

monoclinic,

space group

C6-C2/c

a

6.65 A

b

20.96 A

c

6.50 A 99.3 ~

the main growth direction

a* of reciprocal lattice [ 107]

ao

5.49 A

bo

6.26 A

Regime HI transition (predicted)

155~ [58,108]

Regime II/III transition

137~ [10]

L in Regime I

~0.11 ~ n [8,105]

L' in Regime HI

~2ao=9.1 A [8]

fold surface energy, ~e

122 erg/cm2 [58]

work of chain folding, q

6.6 kcal/mole [ 10]

lateral surface free energy,

9.2-11.5 erg/cm2 [8,58]

activation energy for reptation, QD

1500 cal/mole [8]

T~oin the expression for reptation activation energy

231.2 K [8]

equilibrium melting point

176~ [ 100]

enthalpy of fusion

177J/cm3 [ 100]

chain confonnation

31 helix

13unit cell parameters: unit cell

hexagonal [39]

space group

D4-P3121 [41]

184 a

12.74A

c

6.35A 120~

the main growth plane

(300)

ao

6.36A

bo

5.51A

regime II/III transition

123-129.5~ [100]

fold surface energy, ~

48.2-55.2 erg/cm2 [ 100]

6. Observations of primary nucleation in polymer blends

Primary nucleation in polymer blends, because of high concentration of the second component, is a special case in which other phenomena, obstructing or enhancing various modes of nucleations, take place. Since the primary nucleation habits of a crystallizing polymer in blends depends on the level of miscibility the two types of blends will be reviewed separately.

6.1. Miscible blends

Microscopic observations of crystallization in blends of polymers which are miscible in a molten state have shown that the average size of spherulites in blends is generally larger than the size of spherulites in a plain polymer crystallized under the same conditions. Such observations, rather qualitative, performed frequently for samples crystallized nonisothermally with poorly controlled cooling rate, were reported for the following blends: isotactic polystyrene with atactic polystyrene (iPS/aPS) [109-110], isotactic polystyrene with polyphenylene oxide (iPS/PPO) [111], polycaprolactone

with

polyvinyl

chloride

(PCL/PVC)

[110,112]

poly(ethylene terephthalate) with poly(buthylene terephthalate) (PET/PBT) [110], polyethylene oxide with polymethylene methacrylate (PEO/PMMA) [113] and

185 poly(vinylidene fluoride)(PVDF)/PMMA [ 114,115]. The increase of the average spherulite size in blends as compared to a plain polymer, observed by small angle light scattering technique (SALS), indicates that the number of primary nuclei of spherulitic crystallization in blends (calculated per volume unit of the sample) is lower than in plain crystallizing component. Such a behavior was interpreted as a result of the decrease of the concentration of crystallizable polymer and/or of lowering of a degree of undercooling resulted from a decrease of the melting temperature of crystal phase in a blend [112]. Quantitative investigations of primary nucleation phenomena under the conditions of isothermal crystallization, including the study of separate modes of the nucleation, were performed for two blends of miscible polymers: polyethylene oxide with polymethyl methacrylate (PEO/PMMA)

[116] and isotactic

polypropylene with atactic polypropylene (iPP/aPP) [117,62]. The samples of the PEO/PMMA blend with various composition, prepared by mixing in solution, were crystallized isothermally in the temperature range 4 I~ - 47~ [116]. Two temperatures of melting and melt-annealing of

samples prior to their

crystallization were used in those investigations: 70~ and 80~ temperature of melting (80~

At the higher

all remaining traces of crystallinity were destroyed

in samples, so during crystallization process the self-seeded nucleation was excluded and only the heterogeneous nuclei were active (homogeneous nucleation for PEO requires larger undercooling and at applied crystallization temperature can be neglected). When the samples were melted and melt-annealed at a lower temperature (i.e.70~

both heterogeneous and self-seeding modes of primary

nucleation were active

during isothermal

crystallization.

Fig.3

presents

dependencies of nucleation density, D, (nucleation density means here the number of primary nuclei calculated per volume unit of a crystallizing polymer in the blend) on the blend composition, determined on the basis of microscopical observations of thin films [ 116].

186 Data presented in Fig.3a (samples melted at 80~ prior to crystallization) show that the heterogeneous primary nucleation of PEO spherulites is strongly depressed when PMMA is present in the blend. The stronger depression of heterogeneous nucleation at lower crystallization temperature suggest that only those heterogeneous nuclei become active in the crystallization of the blend for which activation the lowest energy ban'ier must be overcomed. Fig.3b shows that also the self-seeding nucleation mode behaves similarly to the heterogeneous mode of nucleation. The nucleation density presented in this Figure is the result of the activity of both heterogeneous and self-seeded nucleation modes. Similar results were obtained also in the studies of heterogeneous primary nucleation in iPP/aPP blend [62, 117]. The heterogeneous nucleation density as a function of isothermal crystallization temperature determined from Avrami analysis of DSC data is plotted in Fig.4. It is seen that the changes in the nucleation density in the iPP/aPP blend have very similar character to that reported for the PEO/PMMA blend.

8

::3

z

I 0

_4'

J.

J

9

. .

. J

.

l

9

,

II

....

l

_

10 CONTENT OF PMMA (wt%)

;

.....

!

20

187 -

-

-

~

:,,m

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

t,,_

, , _

,

,

,,,,,,

,

,J

B

6

Q

0

;

0

["ig. 3

,;

_

;

.

; ....

t

;

; .....

;

.... 9 . . . .

10 CONTENT OF PMMA (wt.%)

t

20

Dependence of the number of primary nuclei per PEO unit volume (nucleation density D) on the PMMA content in PEO/PA,IMA blends .[or different temperatures of isothermal crystallization, (V): 41~ (e):43~

(0): 45~

(0):47~

Temperature of melt-annealing:

80~ (Fig. 1a), 70~ (Fig. 1b). After Ref [116].

The primary nucleation of iPP crystallization in iPP/aPP blend was studied also for isothermal crystallizations at very high tmdercooling [62]. In this study the samples of a plain iPP and iPP/aPP blends in the form of thin fihns were

188 crystallized isothermally in a special cell at a temperatures ranging from 60~ to 82~

Such extreme conditions (undercooling 100 - 120~

were applied in order

to investigate the homogeneous mode of primary nucleation.

1.51 -

9. . . . . . . . . . . . . . . .

1.0

Q 0.5

0.0

A

.

0

Fig.4

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

10 20 CONTENT OF aPP (wt.%)

,

30

Dependence of nucleation density on the aPP content in iPP/aPP blends. Isothermal crystallization temperatures: (A) 122~ (0) 127~

((3) 131~

(A) 125~

Temperature of melt-annealing 220~

After

Ref Il l 7l.

After completion of crystallization the average spherulite radii (average based on the 5-th order momentum) were determined in the samples by using the S ALS technique. Fig.5 presents the determined values of the average spherulite radius in the blends as a function of crystallization temperature [62]. It was demonstrated

189 [62] that at higher crystallization temperatures (78~176

the heterogeneous

nucleation is the dominant mode.

^ v

10

a a

W FI

I11 tl. if) <

60

65

70

75

80

CRYSTALLIZATIONTEMPERATURE( ~ )

Fig.5

Dependence of the average spherulite radius <:R5> on the crystallization temperature in the samples of iPP and iPP/aPP blends: (0) iPP, (0) iPP/aPP 9:1, ( 0 ) iPP/aPP 8:2, (4) iPP/aPP 7:3. After Ref [62].

The density of this nucleation is constant and independent of the temperature at the applied very high undercooling because of the saturation effect - all heterogeneities induce the formation of heterogeneous nuclei. The homogeneous nucleation, although already active, is still too rare to influence the size of spherulites. This is responsible for the plateau on the dependence of spherulite radius on the crystallization temperature for the range of higher temperatures. However, with farther decrease of the crystallization temperature the rate of homogeneous nucleation increases very rapidly and finally this mode of primary

190 nucleation determines the average spherulite size in the sample. The increase of the homogeneous nucleation rate with the decrease of crystallization temperature induces the significant decrease of average spherulite radius. The effect is clearly seen in Fig.5. In Ref.[62] the description of the homogeneous nucleation mode in blends was proposed. The rates of homogeneous nucleation in plain iPP and iPP/aPP blends were estimated [62]. The results of these estimations are presented in Fig.6. It is seen that for all crystallization temperatures applied the rate of homogeneous nucleation in iPP/aPP blends decreases with the increase of the concentration of aPP in the blend. The changes of the nucleation rate are much stronger than the changes of concentration of iPP in the blend. This shows that the depression of nucleation in blends of miscible pol~aners is not only the result of dilution but there are also some other factors influencing primary nucleation in iPP/aPP system. The problem was discussed in Ref.[62] and will be presented briefly later in this review. An interesting blend of two miscible polyethylenes: conventional linear (HDPE) and linear low density (LLDPE) was studied by Hu et al. [118]. They have shown that in the blend of HDPE with LLDPE both polyethylenes are miscible not only in the molten state but also are able to co-crystallize. A linear dependence of average spherulite radius on blend composition for the whole range of concentration was found. The change in the nucleation density is most probably the result of a different content of impurities inducing the heterogeneous nucleation in both components of the blend. The concentration of the impurities in the sample of the blend is an average of the concentrations in both components before blending. The average concentration of impurities determines the heterogeneous nucleation density in the blend. Because of a similarity of macromolecules of both polyethylenes the other conditions for nucleation are not affected in a blend compared to plain components.

191 0

-

-

'.

.

.

.

.

.

.

.

.

'

'

.......

'

"'

--

9

5 4

60

65

70

75

80

TEMPERATURE (oC)

Fig. 6 Rates of homogeneous nucleation in iPP and in iPP/aPP blend~, calculated per volume unit of iPP in the blend, plotted against crystallization temperature. After Ref [62].

The poly(aryl ether ketone)(PAEK)/poly(ether imide)(PEI) blends are a typical case of a miscible blend with one crystallizable component [ I 19]. The presence of PEI component significantly decreases the bulk crystallization and also crystal growth rate of poly(aryl ether ether ketone) PEEK or poly(aryl ether ketone ketone) (PEKK), but the equilibrium melting temperature and crystal surface free energy are not affected. It was found that the bulk crystallization rate is decreases because of the significant depression in density of primary nuclei due to the presence of PEI, however the decreases in nucleation and in growth rate do not follow the same concentration dependence. A brief review of the crystallization behavior of semicrystalline miscible blends was prepared in the past by Rostami [120].

192 6.2. Blends

with

phase

separation

(immiscible

or partially miscible

components)

The primary nucleation of spherulites in blends of inuniscible or partially miscible polymers was observed and studied mainly in the blends containing isotactic polypropylene (iPP) as a major, crystallizing component (only in few cases both components were able to crystallize) [121-145]. Only a few papers report the observations of primary nucleation in blends containing other crystallizable polymers: Nylon-6 with ethylene-propylene rubber functionalized by grafting on its backbone maleic anhydrite (Ny-6/EPM-g-MA) [146], isotactic polybutene-1 with low density polyethylene (iPB/LDPE) [147], polyphenylene sulfide with high density polyethylene (PPS/HDPE) [ 148,149] and with polyethylene terephthalate (PPS/PET) [150] and with polytetrafluoroethylene [149], polyoxymethylene (POM) with urethane elastomer [151] and PEEK/liquid crystalline polyesters [ 152], PEEK/polysulfones [ 153]. In the referenced studies crystallizing polymers fonned the continuous phase (matrix) whereas second component was dispersed in a matrix in the form of droplet inclusions. Even at the first glance the primary nucleation behavior in blends with phase separated components seems to be more complicated than in miscible systems. Depending on the blend composition and/or experimental conditions various observations have been reported. Both increase or decrease of the nucleation density in blends compared to plain crystallizing polymers have been observed. Below we summarize published results on primary nucleation. Blend of isotactic polypropylene

with isotactic polybutene-1 (iPP/iPB) was

regarded as a blend of miscible polymers. More detailed investigations [121] have shown however, that this blend is phase separated and its components are only partially miscible. In the iPP/iPB blend both components are crystallizable, but

193 iPP can crystallize first at a higher temperature. Siegmann [124] studied the morphology of the iPP/iPB blend crystallized nonisothermally. He has found that the size of spherulites in blends is generally lower than that in plain iPP as well as in plain iPB crystallized at similar conditions, except the case of 50:50 composition in which he has observed a nonspherulitic morphology. Detailed study of isothermal crystallization of iPP in iPP/iPB blends (temperature of crystallization was high enough to prevent crystallization of iPB) showed however, that iPP crystallizes spherulitically also in the case of 50:50 blend [121]. It was found that the density of heterogeneous nucleation (i.e. number of heterogeneous nuclei per volume unit of iPP in the blend) decreases only slightly with the decrease of concentration of iPP in the blend (see Fig.7). Similarly to miscible blends discussed in the previous section, also in that blend of partially miscible polymers, the nucleation at lower crystallization temperature is depressed stronger than at higher. This suggest that during crystallization in a blend only the most active potential heterogeneous nuclei retain their activity. Completely different nucleation behavior was found in other partially miscible polymer blends viz. blends of iPP with various elastomers: ethylene-propylene rubber

(iPP/EPM)

[122,132-136],

ethylene-propylene

diene

terpolymer

(iPP/EPDM) [134,135,145] and polyisobutylene (iPP/PiB) [133,134] and transpolyoctene (iPP/TOR) [143]. In the above cited investigations, in which the DSC, SALS and microscopic techniques were used, the drastic increase of the number of nuclei in the blends with the increase of the concentration of elastomer in the system was found. As an example the results obtained on the basis of microscopic observations of isothermally crystallized blends of iPP with various grades of EPM [133,134] and EPDM [134] are presented in Fig.8.

194

g

4

&

x

Q & m fD Z

tU CI

A

t__

~

9

A

_J

eo ::)

z

I

9

n

0

9

9

!

9

9

10

9

9

|

J.

9

9

20

a

!

9

30

9

I

40

J..

I

50

CONTENT OF iPB (wt.%)

Fig. 7

Dependence of the nucleation density in the iPP/iPB blends on the blend composition. Crystallization temperatures: (A) 119~ 123~

(e) 125~

(0) 130~

(A)

Temperature of melt-annealing:

220~ After Ref [1211. The changes in the number of nuclei in blends compared to plain iPP were commonly attributed to the activity of the elastomer as a nucleating agent for iPP crystallization. In the detailed studies of primary nucleation behavior in iPP/EPM blend [122] it was found that the nucleation behavior in that blend is more complicated than reported in previous studies [132-136]. The drastic changes in nucleation process in the blend were found not only with the variation of composition but also with the variation of both the mixing time and the thermal history of the

samples

prior to their crystallization. In order to obtain the

samples of the blend the components were mixed in the molten state in a mixing extruder.

195 Two sets of samples with the same composition but with different duration of melt-mixing (samples mixed once and twice in the extruder) were prepared and examined for the course of primary nucleation during isothermal crystallization.

..=.,m

e

3O

20

10

.,JL.

10

20

CONTENT OF ELASTOMER (wt.%)

30

iPP/Dutml EPM

30

II

~

20

10

10

20

CONTENT OF ELASTOMER (wt.%)

30

196

iPP/Buna EPM

30

z

20

10

0

10 20 30 CONTENT OF ELASTOMER (wt.%)

40

'tr

iPP/EPDM 30

z

20 V

10 q

I =

9

9

.

0

i

9

9

9

,,a

10

I

I

i

I

I

a

9

20

9

I

30

CONTENT OF ELASTOMER (wt.%)

Fig.8

Number of spherulites per unit area (N/S) as a function of elastomer content in thin film samples of iPP/elastomer blends with different types

of

elastomer

crystallized

at

different

crystallization

197

temperatures: (11) 119~ 131~

(1~) 135~

(T) 123~

(0) 125~

(0) 127~

(A)

The grades of elastomer: (a) Epcar EPM, (b)

Dutral EPM, (c) Buna EPM, (d) EPDM. After Refs. [133,134] Moreover, for each sample three different temperatures of melting and meltannealing prior to crystallization were applied: 190~

230~ and 250~

These

temperatures were chosen in order to study: (1) the heterogeneous mode of primary nucleation- Ta=230~

the temperature sufficient to eliminate the self-

seeded nucleation in iPP, (2) the self-seeding mode - T,=190~

at this condition

of melt-annealing the self-seeded nuclei are not destroyed, so during crystallization both heterogeneous and self-seeded nucleation modes are active and (3) the influence of overheating and partial degradation on nucleation behavior, T~=250~ The results of investigations are presented in Figs.9 a-f. In these Figures the densities of nucleation (recalculated to the volume of iPP in the blend) are plotted as a function of blend composition for various mixing conditions, melt-annealing and crystallization temperatures. As can be seen, the nucleation density increases with increasing concentration of EPM in the blend for all conditions of sample treatment, except for the blends mixed once and annealed prior to isothermal crystallizations at 190~ (Fig.9a, b). Additional, specially designed experiments [122], showed that such unexpected behavior is caused by a significant depression of self-seeded nucleation mode in blends due to partial miscibility of components and the increase of the number of heterogeneous nuclei (compare to Fig.9c). It was demonstrated that the increase of nucleation density with increasing content of EPM in the blend and with increasing

mixing

time observed for other conditions (see Fig. 9) is not the result of activity of EPM as a nucleating agent for iPP crystallization but is the result of the increase of the number of heterogeneous nuclei due to migration of impurities from EPM phase to iPP during preparation of the blend by melt-mixing [122].

198

12 I

A

10 8

~o

6

X

135~

D

4

0

10 20 CONTENT OF EPM (wt.%)

12

B

30

v

10 v

6 X

a

4 2

0

10 20 CONTENT OF EPM (wt.%)

30

199

C

t

x t3

119o~~~-

4

0

10 20 CONTENT OF EPM (wt.%)

1)

30

V

8

s

2 0~ 0

10 20 CONTENT OF EPM (wt.%)

30

200 .

4

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

_,-,, .....

,

,

__.

_

E

3

119o

2 X

ca

1 . O

',lr-

.

.

-i i

-

.

.

.

.

9

. .....

0

""-J.

~

9

~

:.=..-~-

.=

.

9

10

,,

i

, .......

=_ _.

,

I

20

,~

30

CONTENT OF EPM (wt.%)

4

=,,

F

v

"

v

119~

.

x

.

',

o 1 .

0

.

.

.

.

1

10

,

L.

!

=

'

I

.

.

.

.

.

20

.

,1

30

CONTENT OF EPM (wt.%)

Fig.9

Dependence of the nucleation density in the iPP/EPM (DutraO blends on the blend composition. Crystallization temperatures as indicated Samples mixed once (a),(c),(e), and twice (b),(d),09 in the extruder. Temperature of melt-annealing: (a), (b) : 190~ 250~

(c),(d) : 220~

(e),(]):

After Ref [122].

This phenomenon will be discussed in details in the next section of this review. It was also shown in the reported study that partial degradation caused by overheating of the blend induces the decrease of the nucleation density

201 (heterogeneous) in plain iPP as well as in blends (compare Fig.9c to 9e and 9d to 9f). The primary nucleation in iPP/EPDM blends at concentration of EPDM up to 50 wt.% as a function of the temperature of isothermal crystallization was thoroughly studied by Wenig and Wasiak [140]. It was found that nucleation density at 117~ shows three distinct maxima at 5%, 15% an 35% of EPDM in the blend. At 127~ another maximmn at 22 wt% of EPDM in the blend appears. The changes in nucleation activity the authors ascribe to a different dispersion of EPDM component varying with composition. The decrease of spherulite sizes induced by the increase of nucleation density was reported also for blends of iPP with other elastomers viz. with styrene-butadiene rubber (iPP/SBR) [ 132, 135], with polyisoprene (iPP/PiP) [ 132], and with transpolyoctene (iPP/TOR) [143] as well as in blends of polyoxymethylene (POM) with urethane elastomer [151] and of Nylon-6 with ethylene-propylene rubber functionalized by grafting on its backbone maleic anhydrite (Ny-6/EPM-g-MA) [146]. On the other hand, the study of isothermal crystallization of the blend of iPP with polybutadiene (iPP/PBd) showed a significant decrease of the heterogeneous nucleation density in the blend compared to plain iPP [137]. Crystallization, including the primary nucleation phenomena, proceeding in blends of iPP with low and high density polyethylenes (iPP is neither miscible with LDPE nor with HDPE) was studied by several authors [61,60,124-131,142,144]. It was found for the blend of iPP with LDPE that the number of heterogeneous primary nuclei of iPP crystallization, active during isothermal crystallization performed at a temperature above the temperature range for crystallization of LDPE, decreases with increasing concentration of LDPE in the blend [61,125]. The nucleation densities determined in Ref.[ 125] for the blends of iPP with LDPE are presented in Fig. 10. On the other hand, Teh [126] reported the decrease of spherulite sizes (i.e. the increase of nuclei number) in iPP/LDPE blends compared to plain iPP if the

202 samples were crystallized nonisothermally. Similar to the above described nucleation behavior was reported for iPP/LLDPE blends [144]: the number of primary nuclei of iPP spherulites decreases with increasing content of LLDPE in the blends crystallized at isothermal conditions.

...., , , t . . . . . . . . .

,i . . . . .

L ..... J, i . . . .

~. . . . . . . . . . . . . .

1.0

0.8

r

0.6

V

T

I

• 0.4 121 0.2

0.0

Fig. 10

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

0

, _ .

.

I

,

20

,

I

,

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

I

40

CONTENT OF LDPE (wt.%)

I,

60

Dependence o f nucleation density on the LDPE content in iPP/LDPE blends. The isothermal crystallization temperatures: upper curve: 125.8~

lower curve: 128.8~

After Ref [61].

The results concerning the primary nucleation in iPP/HDPE blends were reported in [60,127-130]. Lovinger and Williams [127] observed a strong decrease of spherulite sizes (increase nuclei number) in samples of iPP/HDPE blend compared to plain iPP as well as plain HDPE if the samples were crystallized under nonisothermal conditions after melting at 200~ reported by Noel and Carley [129].

Similar observations were

203 Studies of primary nucleation behavior in iPP/HDPE blend at the conditions of isothermal crystallization [60] showed the changes in heterogeneous nucleation in the blend compared to plain iPP, whereas the self-seeded nucleation was almost not affected. It was found that at the temperature at which only the crystallization of iPP was possible and HDPE remained still in molten state the density of heterogeneous

nucleation

of

iPP

concentration of HDPE in the blend

spherulites

decreases

with

increasing

similarly to iPP/LDPE blend [61], and

iPP/LLDPE blends [144]. However, at the temperature at which the crystallization of both components can proceed at the same time the density of heterogeneous nucleation of iPP spherulites strongly increases with increasing concentration of HDPE in the blend (in the range of concentrations studied iPP constitutes a continuous matrix).

V

x

2

o

0

Fig. 11

135~

10 20 CONTENT OF HDPE (wt.%)

30

Variation in the nucleation density with blend composition for iPP/HDPE

blend~. Crystallization

temperatures

Temperature of melt-annealing: 220~ After Ref [60].

as

indicated.

204 The above described dependencies of nucleation density on composition and a temperature are shown in Fig. 11. Another pair of immiscible polymers in which the primary nucleation was studied is the blend of iPP with atactic polystyrene (iPP/aPS) [123]. In this investigations the increase of the number of heterogeneous nuclei of iPP crystallization with increasing content of aPS in the blend was found. Such increase was found to be stronger for blends mixed in the molten state during longer period of time (see Fig.12). The density of self-seeding nucleation in iPP/PS blend shows also the increase with the increasing concentration of PS in the blend. A separate group of blends are blends with biopolymers: mostly with cellulose and starch. The amount of research on starch/thermoplastic polymers and lignocellulosic/thermoplastic polymers has increased dramatically due to the effort to manufacture environment-friendly plastics. Starch and cellulose are not true thermoplastic polymers although starch can be melted and made to flow trader pressure and shear. Strong nucleating ability of cellulose fibers was observed by Quillin, Caufield and Koutsky [ 154,155] leading to transcrystallinity.

15

. A

119~

A

23~

, X

0

/........~........______.~..s~'-

125~

5,

,,.~..............4~ 9

0

,,

,,

9

I

10

130~ ,,

,,

1

,.

20

CONTENT OF aPS (wt.%)

.t

,

i

30

205 25

B

11

20

~

15

x

10

a

121 125~

0

10

20

30

CONTENT OF aPS (wt.%)

Fig. 12

Dependence of the primary nuclei number per iPP volume unit in the blend on the blend composition for samples of iPP/PS blend mixed twice (a) and three times (b) in extruder. Temperature of meltannealing prior to crystallization 220~

Crystallization temperatures

as indicated. After Ref [123].

Various treatments of cellulose fiber including alkyl ketene dimer, alkenyl succmic anhydride or stearic acid deactivates the nucleating ability of cellulose in the blends with iPP. Detailed and systematic study of starch based polymer blends is scarce. Most of the investigations were aimed at technological aspects of compatibilization (e.g.[156-159]) and not at studying crystallization and nucleation.

7. Observation of primary nucleation in propylene copolymers

Although propylene copolymers are commodity products on a large scale there is very little experimental data in the literature concerning primary nucleation of

206 copolymers containing propylene crystallized from the melt. Several authors found for isotactic propylene copolymerized with relatively small amounts of ethylene [160,161] and butene-1 [162] that the random incorporation of comonomer units into iPP chains promotes the crystallization of polypropylene in the y-form instead of usual a-form. Marigo et al. [161] found in studies of fractionated samples of propylene-ethylene random copolymer that the increasing content of comonomer led to the increase of the tendency of the sample to crystallization in the y-form. They observed that the slow-cooling crystallization of the copolymer fraction with the highest content of ethylene comonomer (6 wt.%) generated the material with crystallites almost entirely in the y-form. The phenomenon of the capability of the random copolymers to crystallization preferentially in y-form is not yet completely understood. It was suggested [ 161] that the insertion of the comonomer units into regular sequences of the iPP chain is responsible for an increase of the number of defects along the chain, which in turn may promote the crystallization of polypropylene in the y-form. The studies of the kinetic of crystallization of the propylene copolymers demonstrated that crystallization rate of random copolymers of propylene with ethylene [160] and butene-1 [163] is substantially reduced compared to iPP homopolymer. This is probably due to decrease of both primary nucleation density and the growth rate of spherulites. Investigations of crystallization of propylene blocks in ethylene-propylene-diene block copolymers revealed that at nonisothennal conditions the temperature of the DSC peak respective to crystallization of propylene blocks in copolymer did not differ markedly from the temperature of the peak of iPP crystallization in the blend iPP/random EPDM copolymer with the fraction of isotactic sequences of propylene in the blend close to that in block copolymer [145]. This result may suggest that the primary nucleation and spherulite growth rate in block copolymer were not retarded in block copolymer compared to those in the blend, although the

207 temperature of the crystallization peak at nonisothermal conditions is not a precise parameter for evaluation of the nucleation and growth parameters (it depends primarily on the experimental conditions, as cooling rate, conditions for heat transfer etc.) and gives only the very qualitative description of crystallization process. More precise study of crystallization in block copolymers was performed by Drzewinski [137] who determined the densities of primary nucleation of isotactic polypropylene blocks

in propylene-butadiene

block

copolymer

crystallized isothermally. On the basis of DSC data analysis he found that the density of primary nucleation in block copolymer is several times higher than in iPP homopolymer crystallized under similar conditions.

8. Phenomena influencing nucleation in blends

The results surveyed in the previous section indicate great effects of the presence of the other component in blends on primary nucleation of crystallizing components, in the sense of interaction at interfaces, degree of dispersion as well as thermal and rheological history. It was shown that practically all modes of primary nucleation: homogeneous, heterogeneous and self-seeding are influenced in blends. The influence of the second polymeric component in a blend on primary nucleation of the crystallizing component depends on: (i) the chemical structure of the second polymer, (ii) its physico-chemical properties including the miscibility and its ability to crystallize and, of course, (iii) its concentration and degree of dispersion in a blend. In some blends of immiscible or partially miscible polymers the conditions of a mixing process appear also an important factor from the point of view of primary nucleation. In Ref.[62] the influence of miscibility of blend components on homogeneous nucleation was considered. The following equation, being the modification of the

208 Turnbull-Fisher equation for nucleation rate in homopolymers, was proposed for the rate of homogeneous nucleation in blends of miscible polymers:

i i

U, _

b,lexpl 32s2se

j

(16)

where:

Iob= Ioc'; Io is the nucleation constant representative for crystallizable polymer and c' is a factor dependent on the concentration of crystallizable polymer in the blend, r The value of c' parameter is within the limits c
209 transition temperature of the system (first exponential term). The free enthalpy Ag' instead of Agf appears in the last term of the proposed equation. It takes into account the necessity of the local separation of blend components during the formation of a stable nucleus. The blend components are compatible in the molten state, thus Ag~<0. Since Agf>0, Ag'=(Ag0- Agb) < Agf, and therefore the energy barrier for the formation of a critical nucleus in the blend of compatible polymers is greater than in plain polymer. This results in the decrease of nucleation rate in the blend of compatible polymers as compared to a plain polymer. The equation discussed above seems to be a good approximation of primary nucleation in the miscible blend of two polypropylene isomers: iPP and aPP [62]. However, it is probably not satisfactory in the case of a blend of two miscible but chemically different polymers, unless the variation of surface energies of the crystal o and o~ in a blend is considered. These energies are different for crystals growing in a blend from those in plain homopolymer because the melt surrounding the crystal is the homogeneous mixture of two chemically different components. This effect was taken into consideration in the description of the crystallization in blends proposed by Rostami [120]. In the way similar to discussed above one can consider the influence of miscibility of blend components on the heterogeneous mode of nucleation as well as on the self-seeding. However, these nucleation modes

are much less sensitive to the

change in the condition of transport. Thus, the change in the energy barrier of the formation of critical nucleus caused by the decrease in the value of free enthalpy, Ag', is practically the only important factor. The increase of that barrier, resulted from the separation of components in the surrounding of the growing nucleus, causes that some of the heterogeneities are now unable to induce the instantaneous nucleation (although in plain polymer they could induce such nucleation) but a time is required for the formation of a stable nucleus on their surfaces. That time may be longer than the time necessary for the crystallization of the sample which

210 begins on other more active nuclei. In this way such heterogeneities, being potential nuclei, are excluded from the crystallization process due to kinetic reasons. As a result the number of heterogeneous nuclei active in the blend is lower than in plain crystallizing component, and decreases with increasing concentration of a second polymer in the blend. Potential nuclei which in plain polymer are less active (higher energy barrier) should loose their activity in blends. This in turn should induce a larger depression of heterogeneous nucleation in blends at lower crystallization temperatures. Such a behavior was indeed reported for iPP/aPP [62,117] and PEO/PMMA [113,116] miscible blends. Similar trends, although less pronounced, were observed also in iPF/iPB blends [121 ], which components are only partially miscible. In those blends the discussed influence of miscibility on heterogeneous nucleation was probably limited to the volume of interfacial layers between components whereas in the volume of iPP matrix, free of iPB molecules, the heterogeneous nucleation was probably not disturbed. As a result the nucleation density in the sample of the blend as a whole was only slightly lowered with respect to that in plain iPP. Based on similar considerations one can also draw similar conclusions for the selfseeding mode of primary nucleation in the case of full miscibility of blend components as well as in the case of partial miscibility. The experimental evidence of such expected behavior of self-seeding comes from the results of investigations of nucleation in samples of iPP/EPM blend mixed once in the extruder [122], in which the decrease of the number of self-seeding nuclei in the blend for the increase of rubber concentration was in fact observed (see Fig.9a). In the samples of the blend mixed twice this was not observed because of large changes in heterogeneous nucleation related to prolonged mixing, screening the changes in self-seeding [ 122]. The above presented concept of the influence of the second polymer component on nucleation behavior of ciTstallizing polymer takes into consideration only a

211 miscibility of components. Thus, it is unable to describe changes of the primary nucleation behavior in cases of blends of inuniscible polymers. In order to explain the nucleation behavior in iPP/LDPE blend (see Fig.8) Galeski, Bartczak and Pracella formulated in [61] the hypothesis of migration of heterogeneities (particles of impurities and/or additives), present in both components before blending, from one polymer toward the other during the process of mixing in the molten state. The driving force for such migration across the interface would be the difference in interfacial free energies of the impurities with respect to both molten components. The tendency to minimize the interracial energies would favor the migration of heterogeneities toward that blend component in which their interfacial energies are lower. In the case of iPP/LDPE blend the migration occurs from iPP towards LDPE. As a result the iPP phase in blend becomes poorer in heterogeneities constituting the potential heterogeneous nuclei compared to plain iPP, so the density of heterogeneous nucleation of iPP spherulites in the blend is lower than in plain iPP. This was indeed observed experimentally. The hypothesis of migration was continued

in the specially designed

crystallization experiment in which the nucleating agent was introduced at the beginning to the volmne of one of the blend components. Then the samples were melt-mixed in repetitive cycles in a mixing device. After each mixing cycle small amount of a blend was collected and the nucleation behavior in that smnple was investigated.

Fig. 13 illustrates the results of such experiment, performed for

iPP/LDPE 8:2 blend nucleated with sodium benzoate. It is seen that the migration of the particles of nucleating agent from iPP to LDPE causes the d~astic decrease of nucleation density with increasing mixing time in the smrlple in which nucleating agent was initially intToduced to iPP phase. The particles of sodium benzoate do not migrate in opposite direction; the nucleation behavior of the

212 sample in which the nucleating agent was introduced initially to LDPE is close to that observed in a control iPP/LDPE blend prepared with no nucleating agent.

1.5 1.0 0.5

0.0

3

4

5

6

7

8

9

NUMBER OF MIXING PASSES

Fig.13

Plots of nucleation density vs. mixing time./br 8:2 iPP,LDPE blends containing 0.2wt% of sodium benzoate, crystallized isothermally at 132~

(a) iPP*/LDPE, (b) iPPLDPE*, (r

iPP/LDPE. The asterisk

indicates the polymer to which the sodium benzoate was initially introduced. After Ref [61].

The phenomenon of a migration of impurities during mixing process was observed and confirmed by similar to the above described experiments with use of nucleating agents in other polymer blends. The migration in the same direction as in iPP/LDPE blend was found in iPP/HDPE blend [60]. The migration in the opposite direction i.e. towards polypropylene was found in iPP/EPM [122] and iPP/aPS [123] blends. In the last two blends the migration induces the significant increase of the density of heterogeneous nucleation in blends with increasing content of the second component and increasing mixing time.

213 The migration was also suggested as a possible cause of changes of primary nucleation density in blends of iPP with polybutadiene (PBd) [137], iPP with LLDPE [144], POM with urethane elastomer [151] and PVDF/Nylon-6 and PVDF/PBT blends [ 165]. Another factor influencing the heterogeneous nucleation in blends is the ability of a second blend component to crystallization. If the second polymer dispersed in a crystallizable matrix can also crystallize under similar conditions as the matrix polymer then the crystals of a second polymer grown at interfaces can act as nucleating agent for crystallization of the matrix. Such behavior was found in the iPP/HDPE blend [60,130]. HDPE dispersed in the iPP/HDPE blend crystallizes below 125~176

with the rate close to the rate of crystallization of iPP.

Crystalline HDPE inclusions dispersed in iPP matrix induce additional nucleation of a number of iPP spherulites. As a result the nucleation density of iPP spherulites in the iPP/HDPE blends increases with increasing content of HDPE in the blend during crystallization below 127~ [60]. It is interesting to note that because of the phenomenon of migration the HDPE inclusions become richer in impurities, thus its crystallization is enhanced, which in turn induces a large increase of the nucleation density of iPP in that blend because of large number of HDPE crystallized inclusions contacting iPP melt. The influence of crystallization of a dispersed polymer on a matrix crystallization can probably explain a decrease of spherulite sizes in nonisothermally crystallized samples of the blends reported in series of papers: iPP with LDPE [126], iPP with HDPE [127] and iPP with iPB [124]. During fast nonisothermal crystallization a simultaneous crystallization of both components is possible. Already crystallized inclusions of a dispersed polymer accelerate the crystallization of a matrix acting as nucleating agent and induce the formation of additional spherulites. Finally the average spherulite radius in those blends becomes smaller than in plain iPP crystallized under the same conditions. The influence of crystallization of a

214 dispersed component on the rise of crystallization temperature of a matrix, which involve the change in primary nucleation, was also reported for the PVDF/Nylon6 and PVDF/PBT blends [ 165]. The studies of iPP/aPS blends [123] have shown that besides the influence of migration phenomenon on the heterogeneous nucleation mode the primary nucleation behavior in those blends is affected also by another phenomenon. This is the ability of the iPP-aPS interface to induce the formation of additional heterogeneous nuclei. Such nucleation ability of iPP-aPS interface is probably connected with the interfacial energy between these pol)qners which induces the reasonable decrease the energy barrier for the formation of heterogeneous nuclei contacting with that interface [ 123]. Another possible reason of activity of the interface toward primary nucleation of crystallization of the matrix might be the possibility of the partial orientation of matrix macromolecules at the interfacial layers. According to the theory of Ziabicki [166] such an orientation could induce the lowering of the energy barrier for the formation of critical size primary nucleus. In this way the interfaces present in the blend volume may induce the formation of a new fraction of prhnary nuclei. An additional class of blends of pol)qners with separation of components is the group of blends in which the minor dispersed component is able to crystallize while the matrix is amorphous or crystallizes, but in an other temperature range. The crystallization of minor component in the blend was studied by several authors: blends of styrene-butadiene rubber (SBR) with iPP [167], EPDM with iPP [145], atactic PS with PE [168,196], PE with polyoxymethylene (POM) [170],

polyvinylidene fluoride

(PVDF)

with Nylon-6

and

polybutylene

terephthalate (PBT) [165] and iPP with POE [171]. It was found in those investigations that if the crystallizing component is finely dispersed in the matrix its crystallization during cooling is often splitted into several distinct steps so called "fractionated crystallization". The origin of those steps is the primary

215 nucleation in isolated molten droplets by different nucleating species -the mechanism similar to the earlier observed by Turnbull and Fisher [54] and by Baer et al [55] in the so-called "droplet experhnent". The steps of "fractionated crystallization" observed as separate DSC peaks can even be separated by more than 60~ [167]. The occurrence of that type of crystallization depends strongly on the state of dispersion of blend components. If the size of inclusions of crystallizing polymer is large, then the crystallization kinetics do not differ markedly from that observed for bulk material. Splitting into separate steps occurs when the inclusions are sufficiently small, so that their number is comparable to the number of nucleating heterogeneities. In that situation all heterogeneities distributed in droplets get a chance to activate a heterogeneous nucleus. In bulk only most active heterogeneities (i.e. those which induce the strongest decrease of the energy barrier for nucleation) give rise to spherulites. Weaker heterogeneities loose the competition and remain inhibited because of kinetic reasons. It is also believed that for sufficiently fine dispersion the homogeneous nucleation of crystallization in some fraction of the inclusions may be observed [ 167,171 ]. This could be possible when the number of crystallizing inclusions dispersed in the matrix is larger than the number of heterogeneities able to induce the nucleation. The detailed discussion of the phenomenon of "fractionated crystallization" including the estimation of number of nuclei and energetical consideration was given by Frensch and Jungnickel [165]. Since the work of Frensch and Jungnickel the fractionated crystallization was observed and identified in many other blends including polyolefins/polymnide 6 [172], polyamide6/

polyolefins

[173]

and those blends with various

compatibilizers [172-174]. The studies of fractionated crystallization are utilized for the determination of activation of heterogeneous nuclei vs. temperature and the temperature dependence of homogeneous nucleation in blends. Recently this powerful method of fractionation of primary nuclei was employed for the

216 determination ofbutene-1 distribution in linear low density polyethylene (LLDPE) by DSC analysis after thermal ffactionated crystallization [175]. Kowalewski, Ragosta, Martuscelli and Galeski [171] made use of the possibilities created by a large temperature gap between heterogeneous and homogeneous modes of nucleation exposed by fractionated crystallization and designed a reversible thermal recording medium. A foil made of iPP/POE blend when cooled below 5~ is opaque, while heated to 55-60~ turns to translucent. It is possible to write on the foil with hot pen and the pattern remains stable within the temperature range of + 10-+50~

Cooling down below 5~ erases the foil and makes it ready

for another writing. The experimental data concerning crystallization of copolymers, though very limited, demonstrate that the primary nucleation behavior in copolymers differs from that in plain homopolymer as well as in the blends. Such difference can be easily understand - the noncrystallizable copolymer units interrupt or terminate crystal growth in the molecular chain direction because they are chemically bonded to crystallizable units and their position along the chain is fixed. Noncrystallizable copolymer units represent defects with various possible disruptive effects. The consequences of their presence on nucleation and growth of crystals depends strongly on their chemical structure, configuration as well as position and concentration along the chain. The numerous theoretical and experimental studies regarding an effect of copolymerization on crystallization process and properties of resulting crystals was reviewed by Wunderlich [176]. Here we will only discuss briefly the primary nucleation in copolymers containing isotactic sequences of propylene. The key feature of chains of a crystallizable copolymer is the order of comonomer units in a copolymer chain which determines the length of crystallizable sequences and the phase structure of a copolymer melt prior to crystallization process. In the case of random copolymers there is random distribution of the length of iPP

217 sequences. Moreover, random copolymers rather do not exhibit phase separation. On contrary, in block copolymers the blocks are quite uniform in length and a phase separation in the molten copolymer is frequently observed. The random length distribution of iPP sequences in random copolymers causes that only a part of the crystallizable sequences can participate in the formation of primary nuclei; too short sequences are excluded from the nucleation process. This decreases the density of primary nucleation as compared to iPP homopolymer. Another important factor limiting the number of primary nuclei in copolymers is the energy barrier due to rejection of noncrystallizable comonomer units out of a forming nucleus, i.e. local phase segregation. This energy barrier is similar in nature to that discussed previously for homogeneous blends, but probably reaches much higher level than in blends because noncrystallizable elements are chemically bonded with segment(s) forming a nucleus. On the other hand, the rejection of noncrystallizable units at the crystal-melt interface changes the interfacial free energies of the forming crystal. This may increase or decrease the energy barrier for critical nucleus formation, depending on the chemical nature of the comonomer. The presence of the comonomer units in the chain influences also the chain mobility, thus modifies the conditions of the transport of the crystallizing segments across melt-crystal interface (again, enhancement or reduction, depending on the nature of the comonomer). The experimental data obtained for propylene random copolymers [ 160,163] suggest that the net result of the discussed barriers is frequently positive and the primary nucleation in random copolymers is significantly lower than in plain homopolymer or in the blends. Different conditions for nucleation are in block copolymers. The crystallizable blocks have similar length, so that practically each of them may participate in nucleation process. The phase separation in block copolymers usually occurs prior to crystallization from the melt, so that there is no additional energy barrier for phase segregation during formation of a nucleus as it is in the case of random

218 copolymers. However, similarly to random copolymers one can expect a variation (frequently a significant decrease [176]) of the interracial free energies of the crystal in a copolymer melt compared to that growing in a parent homopolymer melt, which in turn causes a variation of the energy barrier for critical nucleus formation. Additionally, the interfaces already present in the molten copolymer can play an important role in nucleation process. Copolymer chains pass across those interfaces. This causes a local orientation of the chains near the interface. Such an orientation effect should strongly enhance the primary nucleation in block copolymers. The above remarks suggest the possibility of both: a decrease as well as an increase of the primary nucleation density in block copolymers as compared to a respective homopolymer depending on the

chemical structure of the

noncrystallizable blocks. A probable enhancement of the nucleation in block copolymers is strongly supported by the experimental data obtained by Drzewinski [137] who demonstrated a remarkable increase of the primary nucleation density in propylene-butadiene block copolymer as compared to iPP homopolymer and even more significant increase if compared to the blend with a similar propylene content. The concept of "molecular composite" has led recently to a new class of polymer blends [177,178]. Increasing attention has been paid to the blends containing liquid crystal polymers, and considerable effort has been directed toward research of the processes of phase separation, physical properties, crystallization and melting [179-184]. Owing to the rigid rod nature of liquid crystal molecules, the crystallizable component exhibits unusual crystallization and melting behavior. The rigid rod molecules influence all, the morphology, crystallization kinetics, crystallinity, nucleation and melting and act also as reinforcing species. While the special characteristics of liquid crystal polymers guarantee the improvement of processing and improve the strength and toughness of the blend, it is the crystalline phase which determines the ultimate properties of these materials. The blends with liquid crystalline components of the following

219 high perfonnance polymers have been intensively investigated: poly(ether sulfone) (PES), poly(phenylene sulfide) (PPS), and poly(ether ether ketone) (PEEK) [185187]. As a liquid crystalline component a series of fully aromatic polyesters has been applied. It was shown that liquid crystalline component acts as a nucleating agent promoting the formation of an additional large number of spherulites [188]. It was also found that blending of PPS with liquid crystalline polyester accelerates the crystallization and affects the crystal morphology of PPS due to the nucleating activity of liquid crystalline component [189]. Similar concept of "molecular composite" governs the research of blends of polyolefins with rigid rod molecule oligomers miscible in the amorphous state and in melt [190-197]. The most known

systems

are

blends

of

iPP

and

HDPE

with

hydrogenated

oligocyclopentadiene (HOCP) [ 191-193, 195]. The results of x-ray studies and DSC analysis showed that in iPP/HOCP blends the crystallization of iPP in the orform is inhibited, iPP forms the smectic phase instead[193]. In the system of HDPE/HOCP it was found [195] that the morphology in microspherulitic, however, there is a small fraction of very large spherulites dispersed in a matrix of finer spherulites. These two morphologies depend on the concentration of HOCP and show different crystallization kinetics and melting temperature. It is suggested that the HDPE/HOCP blend forms a three phase system: crystalline phase of HDPE, amorphous HDPE-rich phase, and amorphous HOCP-rich phase [195]. In the case of a blend of poly(4-methylpentene-1) with HOCP it was observed that the addition of HOCP causes a reduction in the overall crystallization rate [198]. The depression is attributed to the diluent effect of HOCP.

9. Concluding remarks The results reported in this paper concerning the primary nucleation behavior in polymer blends show that under usual crystallization conditions the changes in

220 nucleation density can be attributed mainly to the changes in the heterogeneous nucleation mode and in part to the self-seeding. In the discussed studies it was pointed out that there is a variety of phenomena inducing the changes of heterogeneous nucleation in the blends as compared to a plain crystallizing component. The behavior of heterogeneous nucleation in several blends is summarized in Table III. On the basis of the results reported in the discussed papers the conclusion can be drawn that two groups of factors have deciding influence on the changes of primary nucleation in polymer blends:

0

The properties of the minor polymer in the blend including: miscibility with the crystallizing component, glass transition temperature, ability to crystallization and the temperature range in which the crystallization is possible, surface tension of the melt of this polymer.

0

The method of blend preparation (mixing the components) and the parameters of resulting phase structure of the blend including the sizes of inclusions of second component and total surface area of the interface. In the case of melt-mixing very important parameters of mixing process appear to be the duration and the intensity of mixing (shear rate). The migration of additives and impurities occurring during mixing process influences strongly the nucleation behavior, thus the parameters of the mixing process are especially important. In blends in which dispersed polymer is able to crystallize the size of its inclusions, controlled by the parameters of mixing process, is one of the most important factors determining the crystallization kinetics of the blend.

221

Table 111: The phenomena influencing the heterogeneous primary nucleation in polymer blends a),h)

Influence Migration of the of complete impurities or partial __,misg"~,~. . . . . . . . . .

Blend code

iPS/aPS iPS/PPO PCIdPVC PEOIPMMA PAEK/PEI iPP/aPP iPPtiPB iPP/LDPE iPP/I.DPE (crosslinked) iPP/LLDPE iPPtHDPE iPP/EPM iPP/EPDM

Influence of the interface

109,110 III t09,112 113,1t6 119 62,117 121,124 61,125,t26 142

4, $

iPP/SBR iPP/PiB iPPIPiP iPP/TOR iPP/PBd iPP/PS Ny-6/ EPM-g-MA POM/ureth. elastomer iPB/LDPE PPS/HDPE PPS/PET PPS/LCP Ny-6/PVDF PVDF/PBT PEEK/LCP iPPlcellulo~

a)

Crystallizztion of stwa3nd component

I" !'~

144 60,127-I30 122,132-136 134,135,140 145 132,135 133,134 132 143 137 123 146

1'1"I' 1'.I"1' 1' 1" t I" (?) ? (1) I' (?) ? I"? *, O) t

t

t

1"

t

t

t

151

? (?) t (7) t" (?) ? t t

?

(7) (?) 1" 1" 1' 1'

1''11"

147 148,149 150 198 165 165 188 154,155

data concern the crystallization of the polymer mentioned first in the blend code

b) 1' ,1,

an increase of nucleation density, in blend a decrease of nucleation density in blend (number of symbols is related to

the intensity of the effect; (?) means that the mechanisna is suggested in this review, although not evidenced in the cited reference) ~) found for samples crystallized nonisothermally

222 The primary nucleation in the crystallizable copolymers was studied to less extent than blends and still detailed studies are necessary to describe the nucleation behavior in these systems. It emerges from the existing data that the nucleation in copolymers is influenced by many various effects depending strongly on the copolymer composition and ordering of the units along the chain. It would be possible to control quite precisely the primary nucleation by the modification of the chemical structure of the copolymer. References

1. D.R.Paul and S.Newman (eds.), Polymer Blends, Academic Press, New York, 1978. 2. E.Martuscelli, Influence of Composition, Crystallization Conditions and Melt Phase Structure on Solid Morphology, Kinetics of Crystallization and Thermal Behavior of Binary Polymer/Pol3aner Blends, Polym.Eng.ScJ., 24 (1984) 563-586, also V.M.Nadkarni, J.P.Jog, Crystallization Behavior in Polymer Blends, in: Two-Phase Polymer Systems, Ed.L.A.Utracki, Progress in Polymer Science, Hanser Pub., Munich, Vienna, 1991, pp.213239, and Z.Bartczak,

E.Martuscelli,

A.Galeski, Primary Spherulite

Nucleation in Polypropylene Based Blends and Copolymers, in: Polypropylene: Structure, Properties and Blends, Ed.J.Karger-Kocsis, Chapman & Hall, London, 1995 3. K.Fridrich, Analysis of Crack Propagation in Isotactic Polypropylene with Different Morphology, Progr.ColloJdPolym.Sci., 64 (1978) 103-112. also J.L.Way, J.R.Atkinson, J.Mater.ScJ., 7, 1345 (1972) 4. J.D.Hoffman, G.T.Davies and J.I.Lauritzen Jr., The rate of crystallization of linear polymers with chain folding, in: N.B.Hannay (Ed.), Treatise on Solid State Chemistry, Plenum Press, New York 1976, Vol.III, Ch.7, pp.497-614. 5. A.C.Zettlemoyer (ed.), Nucleation, M.Dekker, New York, (1969).

223 6. F.P.Price, Nucleation in Polymer Crystallization, in: Nucleation, A.C. Zettlemoyer (ed.), M.Dekker, New York, (1969). 7. D.Turnbull and J.C.Fisher, Rate of nucleation in condensed systems,

J. Chem.Phys., 17 (1949) 71-73. 8. J.D.Hoffman, Role of reptation in the rate of crystallization of polyethylene fractions from the melt, Polymer, 23 (1982)656-670. 9. J.D.Hoffman and R.L.Miller, Test of the reptation concept: crystal growth rate as a function of molecular weight in polyethylene crystallized from melt, Macromolecules, 21 (1988) 3038-3051. 10. E.J.Clark and J.D.Hoffinan, Regime III crystallization in polypropylene,

Macromolecules, 17 (1984) 878-885. 11. P.Flory, On the morphology of the crystalline state in polymers,

J.Am.Chem.Soc.,$4 (1962) 2857-2867. 12. F.L.Binsbergen and B.G.M.deLange, Heterogeneous nucleation in the crystallization of polyolefms, II. Kinetics of crystallization of nucleated polypropylene, Polymer, 11 (1970) 309-332. 13. F.L.Binsbergen, A revision of some concepts in nucleation theory, Kolloid

Z.Z.Polym.,237 (1970) 289-297. 14. F.L.Binsbergen,

Heterogeneous nucleation in the crystallization of

polyolef'ms: Part I. Chemical and Physical nature of nucleating agents,

Polymer, 11 (1970)253-267. 15. F.L.Binsbergen,

A reconsideration of polymer crystallization theory,

Kolloid Z.Z. Polym., 238 (1970) 389-395. 16. F.L.Binsbergen, Computer simulation of nucleation and crystal growth,

J.Cryst.Growth, 16 (1972) 249-258. 17. F.L.Binsbergen,

Heterogeneous nucleation in the crystallization of

polyolefins, III. Theory and mechanism, J.Polym.Sci., Polym.Phys.Ed., 11 (1973) 117-135.

224 18. P.G.de Gennes, Reptation of a polymer chain in the presence of fixed obstacles, J.Chem.Phys., 55 (1971) 572-579. 19. J.D.Hoffman, C.M.Gutman and E.A.DiMarzio, On the problem of crystallization

of polymers

from

the

melt

with

chain

folding,

Discuss.Faraday Soc., 68 (1979) 177-197. 20. E.A. Di Marzio, C.M.Gutman and J.D.Hoffman, Is crystallization from the melt controlled by melt viscosity and entanglement effects?., Discuss.

Faraday Soc., 68 (1979) 210-217. 21. G.P.Andrianova, Rheological effects of structure formation in polymer melts, J.Polym.Sci.,Polym.Phys.Ed., 13 (1975) 95-112. 22. B.Wunderlich and C.M.Cormier, Seeding of supercooled polyethylene with extended chain crystals, J. Phys. Chem., 70 (1966) 1844-1849. 23. A. Galeski and E.Piorkowska, Localized volume deficiencies as an effect of spherulite growth. I.The two-dimensional case, J. Polym. Sci., Polym. Phys.

Ed., 21 (1983) 1299-1312. 24. A.Galeski and E.Piorkowska, Localized volume deficiencies as an effect of spherulite growth. II.The three dimensional case, J. Polym. Sci. Polym.

Phys. Ed.., 21 (1983) 1313-1322. 25. G.E.W.Schultze and R.Willers, Chord intercepts in the two-dimensional cell modell, J.Polym.Sci., Polym.Phys.Ed., 25 (1987) 1311-1324. 26. R.S.Stein and M.B.Rhodes, Photographic light scattering by polyethylene films, J.Appl.Phys., 31 (1960) 1873-1884. 27. S.Clough, J.J.van Aartsen and R.S.Stein, Scattering of light by twodimensional spherulites, J.Appl.Phys., 36 (1965) 3072-3085. 28. J.V.Champion, A.Killey and G.H.Meeten, Small-angle polarized light scattering by spherulites, J.Polym.Sci., Polym.Phys.Ed., 23 (1985) 14671476.

225 29. J.H.Magill, Crystallization of isotactic polypropylene using a light depolarizartion technique, Nature, 191 ( 1961) 1092-1093. 30. J.H.Magill, A new technique for following rapid rates of crystallization. II Isotactic polypropylene, Polymer, 3 (1962) 35-42. 31. T.Pakula, A.Galeski, M.Kryszewski and E.Piorkowska, Method of determining the kinetics of spherulite primary nucleation from the tnmcation of spherulites, Polym.Bull., 1 (1979) 275-279. 32. A.Galeski and E.Piorkowska, Method of determining the kinetics of spherulite primary nucleation from the spherulite shapes in bulk samples,

Polym.Bull., 2 (1980) 1-6. 33. A.Wasiak, Kinetics of polymer crystallization in nonisothermal conditions,

Chemtracks, Macromol. Chem., 2 (1991) 211-245. 34. E.Piorkowska and A.Galeski, Growth sites in space and time, J.Phys.Chem., 89 (1985) 4700-4703. 35. E.Piorkowska and A.Galeski, Statistical description of spherulite patterns,

J.Polym.Sci.,Polym.Phys.Ed., 23 (1985) 1723-1748. 36. E. Piorkowska, Nonisothermal crystallization of polymers. I. The background of the mathematical description of spherulitic pattern formation,

J.Phys. Chem., (in press) 37. A.Galeski and E.Piorkowska, Statistical approach to the description of spherulite pattern. Two and three dimensional cases, Colloid Polym.Sci., 261 (1983) 1-8. 38. M.M Dumoulin and L.A.Utracki, Melt rheology and morphology of linear low density polyethylene/propylene blends, in: Two-Hase Polymer Systems, L.A.Utracki (ed.), Hanser Publ., Munich, Vienna, 1991, pp.185212 39. H.D.Keith, F.J.Padden Jr., M.M.Walter and H.W.Wyckoff, Evidence for a second crystal form of polypropylene, J.Appl.Phys., 30 (1959) 1485-1488.

226 40. G.Natta and P.Corradini, The crystal structure of isotactic polypropylene,

Nuovo Cimento, SuppL 15 (1960) 40-51. 41. E.J.Addink and J.Beintema, Polymorphism of crystalline polypropylene,

Polymer, 2 (1961) 185-193. 42. J.L. Kardos, A.W. Christiansen and E. Baer, Structure of pressure-crystallized polypropylene, J.Polym.Sci.,A-2, 4 (199) 777-788. 43. A.Marigo, C.Marega, R.Zanetti, G.Paganetto, E.Canossa, F.Coletta and F.Gottardi, Crystallization of the y-form of isotactic poly(propylene),

Makromol. Chem., 190 (1989) 2805-2813. 44. A. Marigo, C. Marego, R. Zanetti, G. Paganetto, S. Schmidt and H.G. Zachmann, Crystallization of the y-form of isotactic poly(propylene), 2 synchrotron radiation measurements, Makromol.Chem. 191 (1990) 19671971. 45. K.D. Pae, y-ct solid-solid transition of isotactic polypropylene, J. Polym.

Sci.,A-2, 6 (1968) 657-93. 46. B. von Falkai and H.A. Stuart, Zur Kristallisation von isotactischem Polypropylen, KolloidZ.Z.Polym., 162 (1959) 138-140. 47. B. von Falkai, Schmeltz- und Kristallisationserscheinungen bei makromolekularen Substanzen. I. Kristallisationskinetische Untersuchungen an isotactischem Polypropylen, Makromol. Chem.,41 (1960) 86-109. 48. J.Majer, Einfluss der Schmeltzbedingungen auf die Kristallisation von Polypropylen, Kunststoffe, 50 (1960) 565-567. 49. F.J.Limbert and E.Baer, Kinetics of nucleation and growth of spherulites in homopolymers, J.Polym.Sci., Part A, 1 (1963) 3317-3331. 50. H.J.Donald, E.S.Hume and L.W.White, Crystallization of polypropylene measured by differential thermal analysis, J.Polym.Sci., Part C, 6 (1964) 93-99.

227 51. S. Hoshino, E. Meinecke, J. Powers, R.S. Stein and S. Newman, Crystallization kinetics of polypropylene fractions, J.Polym.Sci.,Part A, 3 (1965) 3041-3065. 52. U.Johnsen and G.Spilgies, Keimbildung und Kristallisation von nukleiertem und nicht nukleiertem Polypropylen bei konstanter Abktihlgeschwindigkeit,

Kolloid Z.Z.Polym.,250 (1972) 1174-1181. 53. B.Wunderlich, Macromolecular Physics, Vol.II Crystal Nucleation, Growth, Annealing, Academic Press, New York, 1976 pp.235-239. 54. J.R.Burns and D.Turnbull, Kinetics of crystal nucleation in molten isotactic polypropylene, J.Appl.Phys., 37 (199) 4021-4026. 55.

J.A.Koutsky, A.G.Walton and E.Baer, Nucleation of polymer droplets,

J.App/.Phys.,38 (1967) 1832-1839. 56. B.Vonnegut, Variation with temperature of the nucleation rate of supercooled liquid tin and water droplets, J.Colloid.Sci., 3 (1948) 563-570. 57.

F.J.Keith and H.D.Padden, Spherulitic crystallization in polypropylene,

J.Appl.Phys., 30 (1959) 1479-1484. 58. B.Monasse and J.M.Haudin, Growth transition and morphology change in polypropylene, Colloid Polym.Sci., 263 (1985) 822-831. 59. F.L.Binsbergen and B.G.M.deLange, Morphology of polypropylene crystallized from melt, Polymer, 9 (1996) 23-40. 60. Z.Bartczak, A.Galeski and M.Pracella, Spherulite nucleation in blends of isotactic polypropylene with high density polyethylene, Polymer, 27 (1986) 537-543. 61. A.Galeski, Z.Bartczak and M.Pracella, Spherulite nucleation in polypropylene blends with low density polyethylene, Polymer, 25 (1984) 13231326. 62. Z.Bartczak and A.Galeski, Homogeneous nucleation in polypropylene and its blends by small-angle light scattering, Polymer, 31 (1990) 2027-2038.

228 63. L. Goldfarb, Heterogeneous nucleation of

isotactic polypropylene,

Makromol. Chem., 179 (1978) 2297-2303. 64. K.D. Pae and J.A. Sauer, Effects of thermal history on isotactic polypropylene, J.Appl.Polym.Sci., 12 (1968) 1901-1919. 65. J.A.Sauer and K.D.Pae, Effects of thermal history on polypropylene fractions, J.AppLPolym.Sci., 12 (1968) 1921-1938. 66. J.G.Fatou, Melting temperature and enthalpy of isotactic polypropylene,

Eur.Polym.J., 7 (1971) 1057-1064. 67. R.J.Samuels, Quntitative structural characterization of the melting behavior of isotactic polypropylene, J.Polym.Sci., Phys.Ed., 13 (1975) 1417-1416. 68. M.Mucha, Some observations on the melting behavior of isotactic polypropylene crystals, J. Polym.Sci., Polym.Symp., 96 ( 1981) 79-89. 69. B.Wunderlich, Macromolecular Physics,

Vol.II, Crystal Nucleation,

Growth, Annealing, Academic Press, New York, San Francisco, London, 1976, pp.52-96. 70. G.Vidotto, D.L6vy and A.J.Kovacs, Crystallisation et fusion des polymer6s autoensemenc6s. I.Polybut6ne-1, poly6thyl6ne et polyoxy6thyl6ne de haute masse mol6culaire, Kolloid Z.Z.Polym., 230 (1996) 289-305. 71. D.C.Bassett and R.H.Olley, On the lamellar morphology of isotactic polypropylene spherulites, Polymer, 25 (1984) 935-943. 72. D.C. Bassett, Permeanganic etching and its insights into crystalline polymers, in: Developments in Crystalline Polymers-2, Ed.D.C.Bassett, Elsevier Appl.Sci.Pub., London, 1988, p.67-115. 73. R.H.Olley and D.C.Bassett, On the development of polypropylene spherulites, Polymer, 30 (1989) 399-409. 74. B. Wunderlich, Macromolecular Physics, Vol.II, Crystal Nucleation, Growth, Annealing, Academic Press, New York, 1976 pp.44-52.

229 75. H.N.Beck and H.D.Ledbetter, DTA study of heterogeneous nucleation of crystallization in polypropylene, J.Appl.Polym.Sci., 9 (1965) 2131-2142. 76. H.N. Beck, Heterogeneous nucleating agents for polypropylene crystallization, J.Appl.Polym.Sci., 11 (1967) 673-685. 77. V.A.Kargin, T.T.Sogolova, N.Ya.Rapoport and I.I.Kurbanova, Effect of artificial nucleating agents on polymers structure and properties, J. Polym.

Sci., Part C, 16 (1967) 1609-1617. 78. J.Menczel and J.Varga, Influence of nucleting agents on crystallization of polypropylene. I. Talc as a nucleating agent, J. Thermal Anal., 28 (1983) 161-174. 79. T.Kowalewski and A.Galeski, Influence of chalk and its surface treatment on crystallization of filled polypropylene, J.Appl.Polym.Sci., 32 (1989) 2919-2934. 80. T.L.Smith, D.Masilamani, L.K.Bui, R.Brambilla, Y.Khanna, K.Gabriel, Acetals as Nucleating Agents for Polypropylene, J.AppLPolym.Sci, 52, 591 (1994) 81. H.Dragaum, H.Hubeny and H.Muschik, Shear induced [3-form crystallization in isotactic polypropylene, J.Polym.Sci., Polym.Phys.Ed., 15 (1977) 1779-1799. 82. G.Y.Shi and J.Y.Zhang, Study on 13 form polypropylene, Kexue Tongbao, 27 (1982) 290-294. 83. J.Varga, G.Garzo and A.Ille, Kristallisation, Umkfistallisation und Schmelzen der 13-Modifnkation des Polypropylens, Angew. Macromol.

Chem., 142 (1986) 171-181. 84. A.J.Lovinger, J.O.Chua and C.C.Gryte, Studies on the Gt and 13 forms of isotactic polypropylene by crystallization in a temperature gradient,

J.Polym.Sci., Polym.Phys.Ed., 15 (1977) 641-656.

230 85. J.Varga and F.Toth, Annealing of the [3-modification of polypropylene,

Makromol. Chem., Makromol. Syrup., 5 (1986) 213-223. 86. J.Varga, [3-modification of polypropylene and its two-component systems,

J. Thermal Anal., 35 (1989) 1891-1912. 87. J.Varga and G.Garzo, The properties of polymer blends of the 15modification polypropylene and an elastomer, Angew.Makromol.Chem., 180 (1990) 15-24. 88. J.Varga, Y.Fujiwara and A.Ille, 13c~-bifurcation of growth during the spherulitic crystallization of polypropylene, Periodica Polytechnica, Chem.

Eng., 34 (1990) 255-271. 89. J.Varga and G.Garzo, Isothermal crystallization of the [3-modification of polypropylene, Acta Chim. Hungarica, 128 (1991) 303-317. 90. J.Varga and F.S.Toth, Filled compounds of the [3-modification of polypropylene, Angew.Makromol. Chem., 188 (1991) 11-25. 91. A.Turner Jones, J.M.Aizlewood and D.R.Beckett, Crystalline forms of isotactic polypropylene, Makromol. Chem., 75 (1964) 134-158. 92. Y.Fujiwara, Das Doppelschmeltzverhalten der ~-Phase des isotaktischen Polypropylens, Colloid Polym.Sci., 253 (1975) 273-282. 93. W.Ullmann and J.H.Wendorff, Studies on the monoclinic and hexagonal modifications of isotactic polypropylene, Progress Colloid Polym.Sci., 9 (1979) 25-33. 94. K.H.Moss and B.Tilger, Nukleierung and Polymorphie in isotaktischem polypropylen, Angew.Makromol. Chem., 94 ( 1981) 213-225. 95. G.Y.Shi, B.Huang and J.Y. Zhang, Enthalpy of fusion and equilibrium melting point of the ~-form of polypropylene, Makromol.Chem., Rapid

Commm., ~3(1984) 573-578.

231 96. P.Jacoby, B.H.Bersted, W.J.Kissel and C.E.Smith, Studies on the 13crystalline form of isotactic polypropylene, J.Polym.Sci., B, Polym. Phys., 24 (1986) 461-491. 97. H.J.Leugering and G.Kirsch, Beeinflussung der Kristallstnnktur von isotaktischem

Polypropylen

dutch

Kristallisation

aus

orientierten

Schmelzen, Angew.Makromol.Chem., 33 (1973) 17-23. 98. H.J.Leugering, Einfluss der Kfistallsmlktur und der Oberstruktur auf einige Eigenschaften von Polypropylen, Makromol. Chem., 109 (1967) 204-216. 99. J.Garbarczyk and D.Paukszta, Influence of additives on the structure and properties of polymers. 3. Study of 13-ortransition in isotactic polypropylene at various temperatures and times of heating, Colloid Polym.Sci., 263 ( 1985) 985-990. 100. G.Y.Shi, X.D.Zhang and Z.X.Qiu, Crystallization kinetics of 13-phase poly(propylene), Makromol. Chem., 193 (1992) 583-591. 101. J.Garbarczyk, A study on the mechanism of polymorphic transition 13---~otin isotactic polypropylene, Makromol. Chem., 186 (1985) 2145-2151. 102. P.Forgacs, B.P.Tolochko and M.A.Sheromov, Study of the 13-(xsolid-solid transition of isotactic polypropylene by synchrotron radiation, Polym.Bull., 6 (1981) 127-130. 103. J.I.Lauritzen Jr. and J.D.Hoffman, Extension of the theory of growth of chain-folded polymer crystals to large undercoolings, J.Appl.Phys., 44 (1973) 4340-4352. 104. I.C.Sanchez and E.A.DiMarzio, Dilute solution theory of polymer crystal growth: Fractionation effects, J.Res.Nat.Bur.Std., 76A (1972) 213-223. 105. J.D.Hoffman, Regime III crystallization in melt -crystallized polymers. The variable cluster model of chain folding, Polymer, 24 (1983) 3-23. 106. W.R.Krigbaum and I.Uematsu, Heat and entropy of fusion of isotactic polypropylene, J.Polym.Sci., Part A, 3 (1965) 767-776.

232 107. A.J.Lovinger, Microstructure and unit-cell orientation in ot-polypropylene,

J.Polym.Sci.,Polym.Phys.Ed., 21 (1983) 97-110. 108. J.I.Lauritzen, Effect of a fmite substrate length upon polymer crystal lamellar growth rate, J.Appl.Phys., 44 (1973) 4353-4359. 109. F.P.Warner, W.J.Mcknight, R.S.Stein,A small-angle X-ray scattering study of blends of isotactic and atactic polystyrene, J.Polym.Sci., Polym.Phys., 15,2113(1977) 110. R.S.Stein, F.B.Khambatta, F.P.Warner, T.Russel, A.Escala and E.Balizer, X-Ray and Optical Studies of the Morphology of Polymer Blends,

J.Polym.Sci.Polym.Symp., 63 (1978) 313-328. 111. W.Wenig, F.E.Karasz and W.J.MacKnight, Structure and Properties of the System: Poly(2,6-dimethylphenylene oxide) Isotactic Polystyrene. Small Angle X-Ray Studies, J.Appl.Phys., 46 (1975) 4194-4198. 112. F.B.Khambatta, F.Warner, T.Russel and R.S.Stein, Small Angle X-Ray and Light Scattering Studies of the Morphology of Blends of Poly(~caprolactone) with Poly(vinyl chloride), J.Polym.Sci.Polym.Phys.Ed., 14 (1976) 1391-1424 113. E. Martuscelli and G. Demma, Morphology, Crystallization and Melting Behavior of Poly(ethylene oxide)/Poly(methyl methacrylate) Blends, in: Polymer Blends, vol.1, eds.E.Martuscelli, R.Palumbo, M.Kryszewski, Plenum Press, New York, 1980, pp. 101-121. 114.

Y.Hirata and T.Kotaka,Phase separation and viscoelastic behavior of semicompatible polymer blends: poly(vinylidene fluoride)/poly(methyl mathacrylate) system, Polymer J., 13, 273-281 (1981)

115. B. Hahn, O. Hermann-Schonherr and J. Wen&off, Evidence for crystalamorphous interphase in PVDF/PMMA blends, Polymer, 28, 201-208 (1987)

233 116. Z.Bartczak and E.Martuscelli, Spherulites Primary Nucleation in Binary Miscible Blends of Poly(ethylene oxide) with Poly(methyl methacrylate),

Makromol. Chem., 188 (1987) 445-453. 117. Z. Bartczak, Primary Nucleation and Growth of Spherulites in the Blends of Polypropylene, Ph.D. Thesis, Centre of Molecular and Macromolecular Studies, Polish Academy of Science, Lodz 1988. 118. S.R. Hu, T. Kyu and R.S. Stein, Characterization and Properties of Polyethylene Blends I. Linear Low Density Polyethylene with High Density Polyethylene, J.Polym.Sci.Polym.Phys.Ed., 25 (1987) 71-87. 119. B.S.Hsiao, B.B.Sauer, Glass Transition, Crystallization, and Morphology Relationships in Miscible Poly(Aryl Ether Ketones) and Poly(Ether Imide) Blends, J.Polym.Sci., Polym.Phys., 31, 901-915 (1993) 120. S. Rostami, Crystallization Behavior of Semicrystalline Miscible Blend,

Polymer, 31 (1990) 899-904. 121. Z. Bartczak, A.Galeski and M.Pracella, Spherulite Nucleation in Blends of Isotactic Polypropylene with Isotactic Poly(butene-1), J.Appl.Polym.Sci., 54, 1513-1524 (1994) 122. Z. Bartczak, A. Galeski, E. Martuscelli and H. Janik, Primary Nucleation Behavior in Isotactic Polypropylene/Ethylene-Propylene Random Copolymer Blends, Polymer, 26 (1985) 1843-1848. 123. Z. Bartczak, A. Galeski and N.P. Krasnikova, Primary Nucleation and Spherulite Growth Rate in Isotactic Polypropylene - Polystyrene Blends,

Polymer, 28 (1987) 1627-1634. 124. A. Siegmann, Crystalline/Crystalline Polymer Blends: Some StructureProperty Relationships, J.Appl.Polym.Sci., 24 (1979) 2333-2345. 125. E.Martuscelli, M.Pracella, G.Della Volpe and P.Greco, Morphology, Crystallization and Thermal Behavior of Isotactic Polypropylene/Low Density Polyethylene Blends, Makrom. Chem., 185 (1984) 1041-1061.

234 126. J.W.Teh, Structure and Properties of Polyethylene - Polypropylene Blend,

J.Appl.Polym.Sci., 28 (1983) 605-618. 127. A. Lovinger and M.L. Williams, Tensile Properties and Morphology of Blends of Polyethylene and Polypropylene, J.Appl.Polym.Sci., 25 (1980) 1703-1713. 128. E.Martuscelli, M.Pracella, M.Avella, R.Greco and G.Ragosta, Properties of Polyethylene- Polypropylene Blends: Crystallization Behavior, Makromol.

Chem., 181 (1980) 957-967. 129. O.F. Noel III, J.F.Carley, Morphology of Polyethylene - Polypropylene Blends, Polym.Eng.Sci., 24, (1984) 488-492. 130. J. Grebowicz and T.Pakula, Clajstallization and Melting of Components in Blends of Polyethylene and Polypropylene, in: Polymer Blends, vol. 1, eds. E.Martuscelli, R.Palumbo and M.Kryszewski, Plenum Press, New York, 1980, pp. 87-100. 131. F. Rybnikar, Crystallization and Morphology in Blends of Isotactic Polypropylene and Linear Polyethylene, J.Macromol. ScL-Phys., B 27 (1988) 125-141. 132. J. Karger-Kocsis, A. Kallo, A. Szafner, G. Bodor and Z. Senyei, Morphological Study on the Effect of Elastomeric Impact Modifiers in Polypropylene systems, Polymer, 20 (1979) 37-43. 133. E.Martuscelli, C.Silvestre and G.Abate, Morphology, Crystallization and Melting Behavior of Films of Isotactic Polypropylene Blended with Ethylene-Propylene Copolymers and Polyisobutylene, Polymer, 23 (1982) 229-237. 134. E. Martuscelli, C. Silvestre and L. Bianchi, Properties of Thin Films of Isotactic Polypropylene Blended with Polyisobutylene and EthylenePropylene-Diene Terpolymer Rubbers, Polymer, 24 (1983) 1458-1468.

235 135. B.Z.Jang, D.R.Uhlmann and J.B.Van der Sande, Crystalline Morphology of Polypropylene and Rubber-Modified Polypropylene, J.Appl.Polym.Sci., 29 (1984), 4377-4396. 136. N.K. Kalfaglou, Physical characterization of Isotactic Polypropylene/Ethylene-Propylene Copolymer Blends, Angew.Makromol. Chem., 129 (1985) 103-119. 137. M.A.Drzewinski, The Structure - Properties of a Model Set of Crystalline Block Copolymer-Homopolymer Blends, D.Sc. Thesis, Dept.Chem.Eng. M.I.T., Cambridge, 1986. 138. J. Duvall, C. Sellitti, C. Myers, A. Hilmer, E. Baer, Interfacial Effects Produced by Crystallization of Polypropylene with Polypropylene-g-Maleic Anhydride Compatibilizers, J.Appl.Polym.Sci., 52, 207-216 (1994) 139. J.Duvall, C.Sellitti, C.Myers, A.Hiltner, E.Baer, Effect of Compatibilization on the Prope1~ies of Polypropylene/Polyamide-66 (75/25 Wt/Wt) Blends,

J.Appl.Polym.Sci., 52, 195-206 (1994) 140. W. Wenig, A. Wasiak, Interaction between the Components in Isotactic Polypropylene Blended with EPDM, Colloid Polym.Sci., 271, 824-833 (1993) 141.

J. Varga and G. Garzo, The Properties of Polymer Blends of the 13 Modification of Polypropylene and an Elastomer, Angew. Makromol.

Chem., 180 (1990) 15-33. 142.

I. Chodak, I. Janigova and A. Romanov, Crosslinked Polyethylene/Polypropylene Blends, 1. Formation of Insoluble Parts, Crystallization and Melting, Makromol. Chem., 192 ( 1991) 2791-2800.

143. W.Wenig and H.W.Fiedel, Dispersion of Trans-Polyoctylene in Isotactic Poly(propylene), 2. Crystallization Behavior, Makromol.Chem., 192 (1991) 191-200.

236 144. Y.Long, Z..Stachurski and R.A.Shanks, Crystallization Behavior of Isotactic Polypropylene/Lmear

Low

Density

Polyethylene

Blends, Polymer

lnernational, 26 (1991) 143-146. 145. B. Pukansztcy, F.Tudos, A.Kallo and G.Bodor, Multiple Morphology in Polypropylene/Ethylene-Propylene-Diene Terpolymer Blends, Polymer, 30 (1989) 1399-1406. 146. E. Martuscelli, F. Riva, C. Sellitti and C. Silvestre, Crystallization, Morphology, Structure and Thermal Behavior of Nylon-6/Rubber Blends,

Polymer, 26 (1985) 270-282. 147. K.Kishore and R.Vasanthakumari, Crystallization Behavior of Polyethylene and i-Polybutene-1 Blends, Polymer, 27 (1986) 337-343. 148. V.M.Nadkami and J.P.Jog, Thermal and Crystallization Behavior of Alloys of Polyphenylene Sulfide and High Density Polyethylene, Jr. Appl. Polym.

Sci., 32 (1986) 5817-5828. 149. V.M. Nadkami and S. Radhakrishan, Structural Investigations of Powder Processed Poly(phenylene sulfide) Blends, Polym.Eng.Sci., 27 (1987) 547552 150. V.L.Shingankuli, J.P.Jog and V.M.Nadkami, Thermal and Crystallization Behavior of Engineering Polyblends. I. Glass Reinforced Polyphenylene Sulfide with Polyethylene Terephthalate, J.Appl.Polym.Sci., 36 (1988) 335352. 151. E.A.Flexman Jr., New High-Impact Acetal/Elastomer Has Broader Use Potentials, Modern Plastics, 62 (1985) 72-76. 152. Y.Zhong, J.Xu, H.Zeng, Blends of Poly(Ether Ether Ketone) with Thermotropic Liquid Crystalline Polyester I.Morphology, Crystallization, and Melting Behavior, Polymer J., 24, 99-1007 (1992)

237 153. Z.Zhang, H.Zeng, Morphology, Crystallization Kinetics and Melting Behavior of the Blends of Poly(Ether Ether Ketone) with Poly(Ether Sulfone with Cardo Side Groups), Polymer, 34, 4032-4037 (1993) 154. D.T.Quillin, D.F.Caulfield, J.A.Koutsky, Crystallinity in the Polypropylene/Cellulose System. I. Nucleation and Crystalline Morphology,

J.AppLPolym.Sci., 50, 1187-1194 (1993) 155. D.T.Quillin, M.Yin, J.A.Koutsky, D.F.Caulfield, Crystallinity in the Polypropylene/Cellulose System. II.Crystallization Kinetics, J. AppL

Polym. Sci., 52, 605-615 (1994) 156. G.Lay, J.Rehm, R.F.Stepto, et al. U.S.Pat. 5,095,054 (1992) 157. J.L.Jane,

A.W.Schwachbacher,

S.N.Ramrattan,

J.A.Moore,

U.S.Pat.

5,115,000 (1992) 158. J.L.Willet, U.S.Pat. 5,087,650 (1992) 159. U.Vaidya, M.Bhattacharya, Properties of Blends of Starch and Synthetic Polymers Containing Anhydrite Groups, J.AppLPolym.Sci., 52, 617-628 (1994) 160. M.Avella, E.Martuscelli and G.Della Volpe, Composition-Properties Relationships in Propene-Ethene Random Copolymers Obtained with HighYield Ziegler-Natta Supported Catalysts, Makrom.Chem., 187 (1986) 19271943. 161. A. Marigo, C.Morega, R.Zanetti, G.Paganetto, E.Canossa, F.Coletta and F.Gottardi, Crystallization of the y-form of Isotactic Poly(propylene),

MakromoL Chem., 190 (1989) 2805-2813. 162. G.P.Guidetti, P.Busi and I.Giulianelli, Structure-Properties Relationships in Some Random Copolymers of Propylene, Eur.Polym.J., 19 (1983) 757759.

238 163. L.Crispino, E.Martuscelli and M.Pracella, Influence of Composition on the Melt Crystallization of Isotactic Random Propylene/1-Butene Copolymers,

Makromol.Chem., 181 (1980) 1747-1755. 164. D.J.Blundell and A.Keller, The Concentration Dependence of the Linear growth Rate of Polyethylene Crystals from Solution, J.Polym.Sci. part B, 6 (1968) 433-440. 165. H.Frensch and B.J.Jungnickel,

Some Novel Crystallization Kinetic

Pecularies in Finely Dispersing Polymer Blends, Coll.Polym.Sci., 267 (1989) 16-27. 166. A.Ziabicki,., Generalized Theory of Nucleation Kinetics. III. Nucleation in Dilute Systems and/or in Systems with Limited Number of Effective Single Elements, Jr.Chem.Phys., 9 (1977) 1638-1643. 167. A.Ghijsel, N.Groesbeek and C.W.Yip, Multiple Crystallization Behavior of Polypropylene/Thermoplastic Rubber Blends and its Use in Assesing Blend Morphology, Polymer, 23 (1982) 1913-1916. 168. A. Aref-Azar, J.N.Hay, B.J.Marsden and N.Walker, Crystallization Characteristics of Polymer Blends. I. Polyethylene and Polystyrene,

J.Polym.Sci. Polym. Phys.Ed., 18 (1980) 637-643. 169. M.Baitoul, H.Saint-Guirons, P.Xans and P.Monge, Etude par analyse thermique differentiell de melanges diphasiques de polyethylene bass densite et de polystyrene, Eur.Polym.J., 17 (1981) 1281-1288. 170. N.Klemmer and B.J.Jungnickel, Anomalie in Mischungen

Ober eine kristallisationskinetische

aus Polyethylen und Polyoxymethylen,

Coll.Polym.Sci., 262 (1984) 381-386. 171. T.Kowalewski, G.Ragosta, E.Martuscelli and A.Galeski, Crystallization of Poly(ethylene oxide) in Isotactic Polypropylene-Poly(ethylene Blends, J.Appl.Polym.Sci., submitted for publication.

oxide)

239 172. T.Tang, B.Huang, Fractionated Crystallization in Polyolefms-Nylon 6 Blends, J.AppLPolym.Sci., 53, 355-360 (1994) 173. O.T. Ikkala, R.M.Holsti-Miettinen, J.Seppala, Effects of Compatibilization on Fractionated Crystallization of PA6/PP Blends, J.Appl.Polym.Sci.,49, 1165-1174(1993) 174. M. Pracella, S. de Petris, M.Psarski, A.Galeski, Structure-properties relationships in blends of Nylon-6 with polyolefins functionalized with acrylic acid, Polymer, (in press) 175. G. Balbontin, I.Camurati, T.Dallocco, A.Finotti, R.Franzese, G.Vecellio, Determination of 1-Butene Distribution in LLDPE by DSC Analysis After Thermal Fractionated Crystallization, Angew.Makromol.Chem., 219, 139160 (1994) 176. B.Wunderlich, Macromolecular Physics, vol.II, Crystal Nucleation, Growth, Annealing, Academic Press, New York, 1976, pp.259-271. 177. G.Kiss, ,Polym.Eng.Sci., 27, 410 (1987) 178. M.Takayanagi, T.Ogata, M.Morikawa and T.Kai, J.Macromol.Sci., Phys. BI7, 591 (1980) 179. C.J. Ong, F.P.Price, Blends of poly(e-caprolactone) with poly(vinyl chloride). II. Crystallization kinetics, 3Polym.Sci.,Polym.Symp., 63, 59-75 (1978) 180. R.S.Stein, F.E.Khambatta, F.P.Warner, T. Russell, A. Escala, E. Balizer, Xray and optical studies of the morphology of polymer blends, J.Polym.Sci.,

Polym.Symp., 63, 313 -328 (1978) 181. F.G.Joseph, G.L.Wilkes, D.G.Baird, Polym.ScLTechnol., 28, 197 (1984) 182. M. Pracella, E. Chiellini, D. Dainelli, Blends of poly(tetrametylene terephthalate) with a liquid-crystalline polyester, poly(decamethylene 4,4'terephthaloyldioxy-dibenzoate), 2. Crystallization kinetics and melting behaviour, Makromol.Chem., 190, 175-184 (1989)

240 183. S.K.Bhattacharya, A.Tendolkar, A.Misra,,Mol.Cryst.Liq.Cryst., 153, 501 (1987) 184. G.D.Butzbach, J.H.Wendorff, H.Zimmermann, Polymer, 27, 1337 (1986) 185. H.Zeng, C.Young, in: Interfaces in Polymer Ceramic and Metal Matrix Composites, H.Ishida Ed., Elsevier, Amsterdam (1988) 186. K.Mai, H.Zeng, Dynamic mechanical behaviour and phase separation of mixtures of polysulfone with bis[4-(4-ethynylphenoxy)phenyl]-sulfone or poly(phenylene sulfone), Makromol.Chem., 187, 1787-1799 (1986) 187. Z. Zhang, H. Zeng, "c-MRS International '90 Proceedings', Vol.3, 1990, p.55 188. J.Xu, W.Xian, H.Zeng, Polyethersulphone induced crystallization with a liquid crystal polymer, Polym.Commun., 32, 336 (1991) 189. Y.Zhong, J.Xu, H.Zeng, Blends of Poly(Ether Ether Ketone) with Thermotropic Liquid Crystalline Polyester I.The Morphology, Crystallization and Melting Behavior, Polymer J., 24, 999-1007 (1992) 190. E.Martuscelli, C.Silvestre, M.Canetti, C.De Lalla, A.Bonfatti, A.Seves, Crystallization and Thermal Behaviour of Blends of Isotactic Polypropylene and Hydrogenated Oligo(Cyclopentadiene), Makromol.Chem., 190, 2615 (1989) 191. V. Di Liello, E. Martuscelli, G. Ragosta, P. Buzio, Mechanical and High Speed Fracture Properties of Polypropylene and Hydrogenated Poly(Cyclopentadiene) Blends, J.Mater.Sci., 24, 3235 (1989) 192. A.Cecere, R.Greco, A.Taglialatela, Isotactic Polypropylene/Hydrogenated Oligo(Cyclopentadiene) Blends: 1.Thermal and Dynamic-Mechanical Behaviour of Annealed Samples, Polymer, 33, 1411 (1992) 193. S.Cimmino, E.Di Pace, F.E.Karasz, E.Martuscelli, C.Silvestre, Isotactic polypropylene- hydrogenated oligo(cyclopentadiene) blends - phase

241 diagram and dynamic-mechanical behaviour of extruded isotropic films,

Polymer, 34, 972-976 (1993) 194. B. Marcandalli, G. Testa, A. Seves, E. Martuscelli, Oxygen Permeation through Fihns of Polypropylene/Hydrogenated Oligo(Cyclopentadiene) Blends, Polymer, 32, 3376-3380 (1991) 195.

S. Cimmino, E. Di Pace, E. Martuscelli, L.C. Mendes, C. Silvestre, High Density Polyethylene/Hydrogenated Oligo(cyclopentadiene) Blends: Phase STructure and Thermal

and Mechanical Properties, J.Polym.Sci.,

Polym.Phys., 32 2025-2035 (1994) 196. S.Cimmino, P.Guarrata, E.Martuscelli, C.Silvestre, P.P.Buzio, Morphology, Phase Structure and

Thermal

Behaviour

of Films

of Isotactic

Polypropylene/Hydrogenated Oligo(cyclopentadiene) Blends: 1.Extruded Isotropic Films, Polymer, 32, 3299-3304 (1991) 197. S. Cimmino, E. Di Pace, E. Martuscelli, M. Monaco, C. Silvestre, Crystallization and Thermal Behaviour of Poly(4-Methylpentene-1)/Hydrogenated Oligo(cyclopentadiene) Blends, 4th European Symposium on Polymer Blends, Capri, Italy, 1993, p. 150 198. L.I. Minkova, M.Paci, M.Pracella and P.Magagnini, Crystallization behaviour of polyphenylene sulfide in blends with a liquid crystalline polymer, Polym.Eng.Sci., 32, 57 (1992)

This Page Intentionally Left Blank

243 CHAPTER

ISOTACTIC

5

POLYPROPYLENE

BASED

BLENDS

L. D'Orazio, C. Mancarella, E. Martuscelli, G. Sticotti

National Research Council of Italy, Institute of Research and Technology of Plastic Materials, 80072 - Arco Felice - Naples, ITALY.

1.

Introduction

It is well known that the rubber modification of polypropylene ( PP ) accomplished by melt mixing the preformed polymer with suitable elastomers turns out to be very useful in realizing materials with improved final use properties. For the production of high-impact polypropylene, as well as, for the production of polyolefinic thermoplastic elastomers ( TPO ), ethylene-propylene copolymers ( EPR ) and ethylene-propylene-diene terpolymers ( EPDM ) were found to be the most suitable additives. In the last decade many papers have been reported in the scientific literature especially concerning with the effects of blending and crystallization conditions and with the structure-properties relationships of PP/elastomer pairs [ 1 - 26 ]. Generally the PP and the elastomer are immiscible above and below the melting temperature, as in these blends, in the melt the minor rubbery component segregates in its own domains. When the PP is allowed to crystallize then the crystalline phase is nucleated and grows from such a heterogeneous melt. The process of crystallization, especially if it is very fast ( high undercooling ), freezes, in a first approximation, the melt morphology of the amorphous phase. At the end of crystallization, the material

244 is characterized by the presence of spherulites, ( larger or smaller according to the density of nucleation and undercooling ), that have occluded in intra- or inter-spherulitic regions the elastomeric domains. Moreover the rubbery component may interfere with both the primary and secondary nucleation processes and with the kinetic and thermodynamic factors of the PP crystallization inducing deep modification in its final phase structure, i.e. in the texture, size and size distribution of the spherulites, in the inner structure of the spherulites ( lamellar and inter-lamellar thickness ), in the physical structure of the inter-spherulitic boundary regions and amorphous inter-lamellar regions and in the nature of the molecular interconnections between these structural elements ( number and type of tie molecules ). The properties of a such material will be, therefore, the result of a complex combination of several factors, not easily rationalized, related to the mode and state of dispersion of the rubbery component, the crystalline structure and texture, the structure of the interface and to the blending, processing and crystallization conditions. To sum up, despite the need for a rigorous understanding of the dominant structural factors determining the properties of PP/elastomer pairs, none of the cited papers contains a vertically integrated investigation on the influence of the molecular characteristics of both the blend components ( constitution, configuration, molecular masses and molecular mass distribution ) on the development of the phase structure in the melt and in the condensed state, which mainly the final properties of PP/elastomer materials depend from. Such a type of study was carried out by D'Orazio et al., that investigated, for a given isotactic polypropylene/EPR composition, the influence of the molecular characteristics of EPR random copolymers and of the processing and crystallization conditions on the properties of isotactic polypropylene/EPR

245 blends [ 27 - 30 ]. Actually our work on isotactic polypropylene/EPR pairs is in progress to assess the influence of the chain constitution and microstructure of EPR copolymers synthetized by means of catalysts having very high stereospecific activity. In comparison with the EPR copolymers investigated so far and obtained by means of the so called "traditional" catalysts, such elastomers

show

in fact comparatively less intra- and

inter-molecular

homogeneity ( wider distribution of composition and longer sequences of structural units ), absence of stereo- and regio-irregularities and wider molecular mass distribution. In this chapter a review of the undertaken work specifically regarding the influence of the molecular structure of the rubbery component on the impact behaviour of isotactic polypropylene/EPR blends is reported. It will be demonstrated that desired toughening can be imparted to such materials selecting the isotactic polypropylene ( iPP ) and EPR components according to their molecular masses, molecular mass distribution, constitution and tacticity and choosing crystallization conditions able to optimise the mode and state of dispersion of the rubbery phase and the crystalline texture.

2.

Influence of the molecular structure of the E P R c o m p o n e n t

2.1

Influence of the molecular mass

The influence of the molecular mass of samples of EPR random copolymers on the melt rheology, phase structure and properties of iPP/EPR blends has been investigated by D'Orazio et al. [ 27 - 30 ]. The pioneering study was carried out on blends containing two samples of EPR copolymer having practically the

246 same propylene content ( C 3 ) and molecular mass distribution ( MMD ), but differing in their molecular mass, as measured by the Mooney viscosity value, with values of 67 and 45 [ 27 ]. Such EPR samples will be referred to as EPR1 and EPR2 respectively. The molecular characteristics of the blend components are reported in Table 1 together with the glass transition temperature ( Tg ) and the observed melting temperature ( T' m ).

Table 1

Molecular characteristics, glass transition ( Tg ) and observed melting temperature ( Y'm ) of plain isotactic polypropylene ( iPP ) and ethylene-propylene copolymers ( EPR ).

Mooney

Materials

iPP

Melt index viscosity (g/10 min) ML(I+4)

C3

content (%wt/wt)

1.7

Mw/Mn

Tg (o C )

T'm (o C )

7.4

- 10

166

EPR1

67*

28

3.5

- 40

45

EPR2

45*

27

3.5

-41

48

* Measured at 100~ The iPP and EPR copolymers were mixed in a Werner mixer at 230 ~ with a blending time of 3 minutes. Blends with composition iPP/EPR 80/20 ( wt/wt ) were prepared. After blending the materials were injection molded at 230 ~ with a mold temperature of 60 ~ The oscillatory shearing flow properties, namely the complex viscosity 1"1" ( defined by q* = rl' - i 1"1", where q' is the dynamic viscosity or the real part of

247 the viscosity and I"1" is the imaginary part of the viscosity ), the storage modulus G' ( defined by G' = co rl", where co is the frequency of the oscillation in radians per second ) and the loss modulus G" ( defined by G" = co rl' ) of the homopolymers and blends were determined at 200 ~

by means of a Rheometrics

Mechanical Spectrometer in the plate-plate mode with a constant strain of 10% and an angular frequency ranging between 0.1 and 100 rad/sec [ 27 ]. The mode and state of dispersion of the EPR components were analyzed by Scanning Electron Microscopy ( SEM ) only in the core of the injection molded bars in order to eliminate the probable effect of the mold walls on concentration and shape of the EPR particles [ 27 ]. Figure 1 shows the dependence of the logarithm of the modulus value of the complex viscosity (11"1"1) upon the logarithm of the investigated frequencies for iPP/EPR1 and iPP/EPR2 blends. For the sake of comparison in each plot the [1"1*1 logarithm of the single components are also reported. As shown, in the whole range of explored frequency such iPP/EPR

blends

exhibit a decrease in viscosity value with increasing frequency; i.e. such iPP/EPR are pseudoplastic melts. To be noted, moreover, that mixing results in a decrease in viscosity below the mean value of the plain components and that such a decrease is larger at low frequency. This effect is designated as a "negative deviation" [ 31 ] from the following logarithm rule of mixtures that applies at constant temperature and shear rate [31,32]:

log rl = qbl log rll + ~)2 log ]'12

(1)

where r I is the viscosity of the mixture, 1"11 and 1"12 are the viscosities of the two components measured at the same temperature, and ~1 and q~2 are their volume

248 fractions. The finding that such iPP/EPR blends are negative deviation blends ( NDB ) agrees with results obtained by Danesi et al. [ 6 ] using a capillary rheometer. 6.0

5.5 ~

'

~

13~

_~ ~.o; -

O

-.=9 4.5

9

4.0 3.5 -1.0

-0.5

0.0

0.5

1.0

1.5

2.0

Log co O iPP

O EPR 1

& iPP/EPR1 exp.

X iPP/EPR 1 theor.

6.0

5.5

--

t31 o

-J

5.0

4.5

4.0

3.5 -1.0

-0.5

0.0

0.5

1.0

1.5

2.0

Log co O iPP

Fig. 1

El EPR2

A iPP/EPR2 exp.

X iPP/EPR2 theor.

Logarithm of the modulus value of the complex viscosity ([q*l) as function of logarithm of the frequency ( 03 )for iPP and EPR single components and for iPP/EPR blends.

249 Taking into account that in oscillatory measurements on polymer melts the frequency ( co ) becomes analogous to shear rate ( ~/ ) [ 33, 34 ] and assuming an approximate equivalence of q* and apparent viscosity ( 1]a ) [ 34 - 39 ], the zero-shear viscosity qo of both single components and blends was calculated by using the following modified Cross-Bueche equation [ 40 ]:

no

= l+(tx~/)m

( 2 )

1]a

where TIo is the zero-shear viscosity, ~ is a parameter that according to Cross should correspond to the characteristic relaxation time related to molecular weight for the linear polymer solution and m gives a measure of the shearthinning of the melt, i.e. a measure of the decrease in viscosity with increasing rate of shear. According to Iwakura et al. [ 41 ] for polymer melts c~ is related to the size of the apparent flow unit; the reciprocal of ~ corresponds to the shear rate at which

l]a=

viscosity rio and

rio/ 2. From the lines

1/

l]a v e r s u s

~' m the zero-shear

a values are easily obtained from the reciprocal of the

intercept and from the slope respectively. The m, 11o and c~ values of the single components and blends are reported in Table 2; for the blends the zero-shear viscosity values calculated assuming the additivity logarithm rule ( rio' ) are also reported. As shown in such a table, the

11o blend values

exhibit a very large

negative deviation from the logarithm additivity rule; to be pointed out that the extent of such a deviation increases with increasing EPR average molecular mass. The m parameter assumes comparatively lower value ( 4/7 ) for the blend

250 containing the EPR2 copolymer indicating a less severe shear-thinning in nonNewtonian region. Moreover as far as c~ parameter is concerned, it can be deduced that in the blend system the transition from Newtonian to pseudoplastic flow is able to start at a frequency higher than that of the plain iPP and the shift towards higher frequency tends to increase with increasing EPR average molecular mass.

Table 2

Application of Cross-Bueche equation: values of m, ot and qo for plain iPP and EPR copolymers and for iPP/EPR blends," the zero-shear viscosity values calculated assuming log additivity ( qo ') for iPP/EPR blends are also reported.

SAMPLE

m

ot ( sec )

qo ( P )

qo' ( P )

iPP

2/3

0.751

130707

EPR1

4/7

0.712

582270

EPR2

6/11

0.366

199729

iPP/EPR1

2/3

0.212

72850

176198

iPP/EPR2

4/7

0.528

85286

142262

The analysis of the EPR mode and state of dispersion, carried out by SEM in the core of transverse microtomed surfaces exposed to boiling xylene vapours to remove EPR, shows that the minor component segregates in almost spherical shaped domains with size strongly dependent on the molecular mass of the EPR

251 copolymer.

A finer dispersion

is achieved

in blends containing

EPR2

( see Fig. 2 ).

Fig. 2

Scanning electron micrographs of smoothed and etched surfaces of iPP/EPR blends.

In such a blend the copolymer domains have a diameter ranging between 0.1 and 0.5 p m , whereas in the blends containing EPR1 the range of particle size is broader ( 0.1 p in + 1.0 la m ). The values of the number- average particle size ( D~ ) found for EPR2 and EPR1 are 0.30 lam and 0.40 ~m respectively. Therefore the dispersion coarseness of such EPR copolymers increases with increasing their melt viscosity, i.e. with increasing phase viscosity ratio defined

252 as la - vii / q2

where 1"11is the viscosity of the dispersed phase and r12 that

of the matrix ( see Table 3 ). An opposite trend was observed by D'Orazio et al. in polyamide6/ethyleneco-vinylacetate ( PA6/EVA ) blends; the results of this study are shown in detail in the next chapter.

Table 3

Phase viscosity ratio ( 111 /112 ), range of particle size ( D ) and number-average particle size ( Dn ) for iPP/EPR blends.

SAMPLE

1"11/1"12

D ( gm )

On

iPP/EPR1

4.5

0.1 + 1.0

0.40

iPP/EPR2

1.5

0.1 + 0.5

0.30

( la m )

The type of dependence of the size of the dispersed particles upon the phase viscosity ratio observed in iPP/EPR and PA6/EVA blends agrees qualitatively with the predictions of the Taylor-Tomotika theory [ 42 - 44 ]. According to this theory, the instability coefficient ( q ) of a cylindrical thread suspended in a viscous liquid is given by the following expression"

- T---J----(1-x2)F(x,g)= ~/ f2(x,g) q - 2rlo R 21"1oR

( 3 )

where ~, is the interfacial surface tension, x = 2nR/~,, R is the diameter of the thread, )~ is the varicosity of the thread and g is the phase viscosity ratio.

253 It was found by Tomotika that, for a given value of It, the maximum instability occurred at a certain definite value of ~,, indicating that drops of definite size would be formed and that ~, changed with it. According to equation ( 3 ) the R versus log ~t function should show a minimum in the vicinity of it= 1 [ 44 ]. The notched Charpy impact strength values for the plain iPP and the iPP/EPR1 and iPP/EPR2 are reported in Fig. 3 as a function of the test temperatures. As well known the plain iPP shows very poor impact properties in the whole range of explored temperatures and, for a test temperature below the glass transition temperature (

Tg ) of the EPR copolymers (< 40 ~

), no improvement

in the iPP impact strength is obtained irrespective of the EPR molecular mass.

1200

O"J c

r E

-

800 ~

E

>,..~

,.c

400

,,c 0 Z

01#= -60

-40

-20

T - - . 0 - - iPP

Fig. 3

.~.

0

20

40

(oc)

iPP/EPR1

,_,,* iPP/EPR2

Notched Charpy impact strength as a function of temperature for plain iPP and iPP/EPR blends.

254 For a test temperature higher than EPR Tg and close to iPP Tg ( -10 + 0 ~

),

the enhancement in iPP impact behaviour strongly depends on the molecular mass of the dispersed phase ( see Fig. 3 ). As a matter of fact much better properties are shown by the EPR1 containing blend. At the temperature of 0 ~ the impact value shown by such a blend is about 14 times as high as that shown by the plain iPP, whereas iPP/EPR2 blend exhibits an impact value just three times as high as that exhibited by the plain iPP. The very different behaviour of EPR1 and EPR2 has been related to their different average particle size ( see Table 3 ). EPR particles ranging in size between 0.1 and 1.0 ~ m with an average diameter of 0.40 g m result more effective for iPP toughening than particles ranging between 0.1 and 0.5 g m. The possibility that in iPP/EPR1 blend a more effective inter-particle distance between two nearest-neighbour dispersed domains is achieved should be likely also considered. In order to assess the effects of ~ when iPP and EPR components have melt viscosity values close to each other, i.e. in the vicinity of the minimum expected by the Taylor-Tomotika theory, the same sample of iPP was blended with two other samples of EPR copolymers having, for constant C3 content ( ~43% wt/wt ), suitable molecular masses [ 29 ]. Such EPR samples will be referred to as EPR3 and EPR4. The average molecular masses and the molecular mass distribution ( MMD ) of the plain iPP and such EPR copolymers, determined by means of Gel Permeation Chromatography ( GPC ), are reported in Table 4. The blending and the samples preparation procedures were kept the same employed in the previous work; blends with composition iPP/EPR 80/20 were analyzed [ 29 ]. The oscillatory shearing flow properties of the single components and blends were determined at 200 ~

by means of a Rheometrics Mechanical

255 Spectrometer in the plate-plate mode with a constant strain of 10% and an angular frequency ranging between 0.1 and 100 rad/sec.

Table 4

Number-average molecular mass ( Mn ), weight-average molecular mass ( Mw ), z-average molecular mass ( Mz ) and molecular mass distribution ( M w / Mn ) for plain iPP and EPR copolymers.

SAMPLE

Mn. 10 3

Mw. 10 3

Mz. 10 3

Mw / Mn

iPP

65

484

2782

7.4

EPR3

40

110

300

2.8

EPR4

70

180

500

2.8

As shown in Fig. 4 and as expected, iPP/EPR3 and iPP/EPR4 melts are pseudoplastic in the whole range of explored frequencies. It is, at the same time, very surprising to find that mixing iPP and such EPR copolymers results in a decrease in blend viscosity according to equation ( 1 ). This finding, in disagreement with the results previously obtained, indicates that in such iPP/EPR blends there is no mutual influence of the single components despite their melt immiscibility and heterogeneity. The m, c~ and

rlo values for plain iPP and EPR copolymers and for

iPP/EPR blends calculated by applying Cross-Bueche equation are reported in Table 5 together with rio' values of the blends. To be pointed out that these iPP/EPR blends, as far as 1"1ois concerned, show a positive deviation from equation ( 1 ) contrary to what previously shown by

256 blends based on the same iPP sample, but containing different EPR samples ( EPR1 and EPR2 ); moreover the extent of such a positive deviation increases on increasing the EPR average molecular mass (see tables 4 and 5).

4.4

iPP/EPR3

4.0 7--

3.6

O1 o ..i

3.2

o Experimental A Theoretical

2.8 2.4

9

i

-3

!

I

-2

!

I

-1

!

I

0

!

1

I

|

I

2

3

Log ( o~ )

4.4

9

4

4.0 A

--

3.6

ol o _1

3.2

,1r

O Experimental ATheoretical

2.8 2.4

9

-3

!

-2

9

!

9

!

-1 Log

Fig. 4

9

0

!

1

.

!

2

9

i

3

(~ )

Logarithm of the modulus value of the complex viscosity ([11"1) as function of logarithm of the frequency ( 03 )for iPP/EPR blends.

257

Table 5

Application of Cross-Bueche equation: values of m, ~ and rio for plain iPP and EPR copolymers and for iPP/EPR blends," the zero-shear viscosity values calculated assuming log additivity ( ~o ') for iPP/EPR blends are also reported.

SAMPLE

m

c~ ( sec )

qo ( P )

rio' ( P )

iPP

2/3

0.722

164100

EPR3

4/7

0.267

218390

EPR4

4/7

0.281

271000

iPP/EPR3

4/7

1.110

188000

174053

iPP/EPR4

4/7

1.050

221141

181967

Therefore it has been supposed that in absence of shear these iPP/EPR systems made of single components having melt viscosity values close to each other may be

described

in terms

of a continuos

two-phase

model,

where

the

macromolecules of one phase are physically entrapped into the macromolecules of the other phase. As shown in Table 5, the m parameter assumes for both the blends the value of 4/7 suggesting that the blend system undergoes a shear thinning less severe than that of the plain iPP in non-Newtonian region. The blend c~ values are comparable to each other; such values are longer than that shown by the plain iPP, indicating that the transition from Newtonian to

258 pseudoplastic flow starts at a frequency lower than that of the plain iPP; an opposite result was obtained for EPR1 and EPR2 containing blends [ 27 ]. Taking into account that EPR3 and EPR4 copolymers have a viscosity lower than that EPR1 and EPR2 copolymers, but closer enough to iPP viscosity, the contradiction among the results indicates that very different rheological behaviour can be exhibited by iPP/EPR blends depending on the molecular masses of the EPR component, that in additional affect the phase viscosity ratio. The analysis by SEM of the mode and state of dispersion of such EPR copolymers realized in injection molded blend samples shows that a layered structure transversal to the mold filling direction, according to the schematic model reported in Fig. 5, is developed.

9

o~

O

9

9 o.e

9 o

O

v

9

" 9

0

9

9

Oo O

1

9 9

9

M.ED. I

Fig. 5

Schematic model of layered structure generated in injection molded samples of iPP/EPR blends.

259 As shown moving from the border toward the core of the samples three different layers are found: 1.

a skin surface (S) where no dispersed domains of EPR can be observed; the thickness of such a layer ranges between 15 and 20 g m.

2.

an intermediate transition layer (I) where the concentration of the EPR domains increases going toward the core of the samples with gradient characteristic; the thickness of such a layer ranges between 20 and 25 ~tm.

3.

a core (C) showing an EPR droplet-like morphology; the range of the diameters of the EPR particles are reported in Table 6 together with the number-average of such diameters Dn and the ~ value of the blends.

Table 6

Phase viscosity ratio ( 111 /112 ), range of particle size ( D ) and number-average particle size ( Dn )for iPP/EPR blends.

SAMPLE

TI1 /1"12

D (/.t m )

On ( g m )

iPP/EPR3

1.33

0.1 + 0.4

0.25

iPP/EPR4

1.65

0.1 + 0.6

0.35

m

The trend of

D n

values versus log g confirms that

D n

decreases with

decreasing log g in agreement with the expectation on the basis of the TaylorTomotika theory. The notched Izod impact strength values for the plain iPP and the iPP/EPR3 and iPP/EPR4 blends are reported in Fig. 6 as a function of test temperatures.

260 As expected and in good agreement with the results previously showed concerning iPP/EPR1 and iPP/EPR2 blends, for test temperature below - 40 ~ no toughening is imparted to blend material. The fractographic analysis of the broken surfaces of both plain iPP and such iPP/EPR blends shows that all samples break in a brittle fashion; the fracture surfaces, in fact, show a crack path where practically no matrix yielding takes place. Moreover for test temperatures ranging between - 40 and

20 ~

the fracture surfaces of the plain iPP show no stress-whitening phenomenon, whereas the fracture surfaces of the blend samples, except the skin layer made of plain iPP, are more or less whitened. Particularly, the surfaces of the blend samples containing the EPR3 copolymer are completely whitened, whereas those of the blend samples containing the EPR4 copolymer shows only slight whitening localized in the core of the sample.

800 -

A

E

e-

600

400

l... r

E

200

9

|

-60

-40

|

-20

9

0

20

40

T ("C) O' iPP

Fig. 6

IEIiPP/EPR3

/k iPP/EPR4

Notched Izod impact strength as a function of temperature for plain iPP and iPP/EPR blends.

261 Taking into account that for test temperatures below iPP

Tg the

blend material

shows no improved strength, the observed stress-whitening phenomenon is mainly to be associated with cavitation during the test and/or a possible orientation contribution. For test temperatures ranging between - 10 and 0 ~ higher impact strength values are shown by blend samples containing EPR4 copolymer ( see Fig. 6 ). The fractographic analysis of the surfaces of the blend samples broken at 0 ~

reveals that the size of the fracture induction area, where the material

breaks in ductile fashion, depends on the molecular structure of the EPR copolymer. As shown in Fig. 7, considerable larger induction area is observed in the blend sample containing EPR4 copolymer.

Fig. 7

SEM micrographs of fracture surfaces broken at 0 ~ iPP/EPR3 and ( b ) iPP/EPR4 blends.

of ( a )

Moreover the surfaces fractured at 0 ~ of the blend samples containing EPR3 copolymer are completely whitened, whereas the fracture surfaces of the blend samples containing EPR4 copolymer exhibit whitening localized only at the fracture induction area.

262 To sum up the best impact properties are shown by the blend samples ( iPP/EPR4 ) that undergo stress-whitening with lowest intensity ( supporting that a cavitation process, rather than multicraze formation, is to be associated with the observed whitening ) and largest induction area. The very different behaviour of the EPR3 and EPR4 copolymers as impact modifiers has been related to their different average particle size ( see Table 6 ). The EPR domains ranging in size between 0.1 and 0.6 ~tm are more effective for iPP toughening and/or for achieving a more effective inter-particle distance according to Wu [ 45 ]. The Dn value able to optimize the impact strength of the iPP is comparable to that found while studying iPP/EPR1 and iPP/EPR2 blend systems. For test temperatures close to room temperature, the impact strength shown by the blend samples containing EPR4 copolymer holds higher than that shown by blend samples containing EPR3 copolymer. To be pointed out that with increasing test temperature, the intensity of the stress-whitening phenomenon and the volume of the material involved increase strongly with no dependence on EPR molecular structure, indicating that multicraze formation may occur. Therefore for test temperature close to room temperature, the fracture mechanism active in such materials likely results in a combination of shear yielding and multicraze formation. The results of our studies on iPP/EPR systems showed that the molecular mass of the EPR copolymer determines the phase structure of the blend, both in the melt and after crystallization process. Moreover a general correlation has been established between the mode and state of dispersion of EPR and the melt rheological parameters, that is between the value of the number-average particle size (Dn) and the particle size range ( D ) of the rubbery phase, as determined by SEM in crystallized samples, and the melt phase viscosity ratio (~t). This

263 correlation has been made under the approximation that the iPP crystallization process freezes the melt morphology of the EPR amorphous phase. A general correlation has been established also between the impact behaviour of iPP/EPR

blends and the mode and state of dispersion of EPR

copolymer, that is between the value of the impact strength for test temperature higher than EPR

Tg and the

value of the number-average particle size ( Dn ) of

EPR phase. Taking into account that Dn value shown by EPR copolymers increases with increasing log g ( see Tables 3 and 6 ) according to the prediction on the basis of the Taylor-Tomotika theory, required toughening can be imparted to iPP/EPR blend materials by optimizing the melt phase viscosity ratio; i.e. for a given iPP sample by optimizing EPR molecular mass.

2.2

Influence of the m o l e c u l a r mass distribution

The influence of the molecular mass distribution of the EPR phase has been studied in iPP/EPR blends prepared with two copolymers ( EPR5 and EPR6 ) having practically the same C3 content ( 43% wt/wt ) and weight-average molecular mass ( Mw ), but differing in their molecular mass distribution ( M w / M n - 4.0 and 13.3 respectively) [ 29 ]; the molecular characteristics of the blend components, determined by GPC, are reported in Table 7. The iPP and EPR copolymers were mixed in a Banbury mixer at 200 ~ with a blending time of 5 minutes. Blends with composition iPP/EPR 80/20 ( wt/wt ) were prepared. After blending the materials were injection molded at 200 ~ C with a mold temperature of 60 ~

264 m

Table 7

Number-average molecular mass ( Mn ), weight-average molecular mass ( Mw ), z-average molecular mass ( Mz ) and molecular mass distribution ( M w / Mn ) f o r plain iPP and EPR copolymers.

SAMPLE

Mn. 10 3

M--w.10 3

Mz. 10 3

iPP

65

484

2782

7.4

EPR5

50

200

800

4.0

EPR6

15

200

1000

13.3

Mw

/ Mn

The dependence of the logarithm of the modulus value of the complex viscosity

< I*1

uoon the logarithm of the investigated frequencies shown by

such iPP/EPR blend systems confirms, as expected, that iPP/EPR5 and iPP/EPR6 are pseudoplastic melts; to be noted, furthermore, that the values of log Irl*l agree with those predicted by Equation ( 1 )(see Fig. 8)indicating, irrespective of the EPR molecular mass distribution, no mutual influence between the two components. Analogous results have been obtained for iPP/EPR blends containing EPR copolymers characterized by the same MMD with value of 2.8, but different molecular masses ( EPR3 and EPR4 samples ) ( see Figs. 4 and 8 ). The rheological parameters of both single components and blends have been derived as in the previous paragraph by using the Cross-Bueche equation; the results are summarized in Table 8.

265 4.4-

A

9

5

3.6

i -ir i

O1 O .J

3.2

O Experimental A Theoretical

2.8 2.4

I

-3

9

-2

g

9

-1

g

9

0

e

9

1

I

2

Log(~ )

4.4-

9

6

4.0 A m

m

E"

3.6

v

O1 O --I

3.2

O Experimental

A Theoretical

2.8 2.4

w

-3

e

-2

w

e

-1

w

g

0

9

!

1

w

!

2

9

g

3

Log(~ )

Fig. 8

Logarithm of the modulus value of the complex viscosity (I~]*1) as a function of logarithm of the frequency ( co )for iPP/EPR blends.

266

Table 8

Application of Cross-Bueche equation." values of m, ~ and rio for plain iPP and EPR copolymers and for iPP/EPR blends," the zero-shear viscosity values calculated assuming log additivity ( qo ') for iPP/EPR blends are also reported.

SAMPLE

m

a ( sec )

qo ( P )

qo' ( P )

iPP

2/3

0.722

164100

EPR5

4/7

1.735

331016

EPR6

4/7

1.761

307500

iPP/EPR5

4/7

1.143

287356

189622

iPP/EPR6

4/7

1.152

263300

186765

It is worth noting that such blends, as far as qo values are concerned, show a positive deviation from equation ( 1 ); moreover the extent of such a deviation decreases with increasing the MMD of the copolymer; i.e. with decreasing its number-average molecular mass ( see Table 7 ). Positive deviation from equation ( 1 ) of qo values have been also shown by blends containing EPR copolymers having the same MMD with value of 2.8, but different molecular masses ( EPR3 and EPR4 samples ) ( see Table 5 ). The SEM analysis of the mode and state of dispersion of EPR5 and EPR6 copolymers, generated in injection molded blend samples shows that a layered structure according to the schematic model reported in Fig. 5, is developed;

267 analogous layered structure has been found in injection molded samples of blends containing EPR3 and EPR4 copolymers. To be pointed out that in the core of the sample with reference to Fig. 5 iPP/EPR6 blend exhibits a more coarse dispersion of the EPR spherical shaped domains with a broader particle size distribution. As shown in Table 9 D range and Dn value measured for EPR6 copolymer are noticeably higher than that measured for EPR5 copolymer and expected according to melt phase viscosity ratio value, as predicted by the Taylor-Tomotika theory.

Table 9

Phase viscosity ratio ( 1"11/112 ), range of particle size ( D ) and number-average particle size ( Dn )for iPP/EPR blends.

SAMPLE

111/1]o

D ( btm )

On ( lu.m )

iPP/EPR5

2.02

0.1 + 0.6

0.35

iPP/EPR6

1.87

0.4 + 1.2

0.80

In additional the range of the EPR particle size affects the impact properties of such materials. As shown in Fig. 9 the blend containing the copolymer with narrower Mw/Mn ratio ( EPR5 ) shows, for test temperatures above the Tg of the EPR copolymer, better impact properties. From this study it has been shown that the molecular mass distribution of EPR copolymer is a relevant structural factor of the elastomeric phase in determining its mode and state of dispersion in iPP matrix both in the melt and in the condensed state.

268 800

A

"E

600

,.C ,4,-I

=

400

L ,4-,I

0

"

E

200

0 -60

-40

-20

0

20

40

T (~ O iPP

I-I iPP/EPR6

/k iPP/EPR5

Fig. 9 Notched Izod impact strength as a function of temperature for plain iPP and iPP/EPR blends. Recalling the direct correlation established in the previous paragraph between the value of the impact strength of iPP/EPR blends and the value of the EPR number-average particle size ( D n ), the molecular mass distribution of EPR copolymer turns out to be a relevant factor also in controlling toughening of iPP/EPR blend materials.

269 2.3 Influence of the propylene content

The influence of the propylene content ( C3 ) along the chain of random EPR copolymers on the melt rheology, phase structure, crystallization and impact properties of iPP/EPR blends has been also investigated [ 27, 29 ]. The sample of iPP, whose average molecular masses and MMD are reported in Table 4, has been blended with samples of EPR copolymers, having for almost comparable molecular masses and molecular mass distribution, increasing C 3 content ( wt/wt ). The molecular characteristics of such EPR

samples are reported in Table 10.

Table 10 Molecular characteristics of EPR copolymers.

( wt/wt )

Mooney Viscosity ML (1+4)

EPR1

28

67

EPR5

43

EPR7

43

85

EPR8

58

55

SAMPLE

C3%

Mn. 10 3

Mw. 10 3

Mz. 10 3

Mw/Mn

3.5

50

200

800

4.0

5.0

73

259

860

3.5

The values of the blend zero-shear viscosity, m and (z parameters, derived as previously by using the Cross-Bueche equation, the melt phase viscosity ratio

270 and the range of the EPR particle size, as measured by SEM in injection molded samples, are reported in Table 11 together with EPR C3 content. It is worth noting that, for the investigated C3 content at least, the melt rheological parameters of the blends and the EPR mode and state of dispersion, show no systematic dependence upon such a structural factor of the elastomers. With increasing EPR C3 content from 28% to 43% ( wt/wt ) ( EPR5 and EPR7 copolymers ), opposite rheological behaviour by iPP/EPR blends are shown depending on EPR molecular mass and molecular mass distribution, which also affect the melt phase viscosity ratio [ 27 ]. To be recalled, in connection with, that for the same C3 content equal to 43% ( wt/wt ) along copolymer chain iPP/EPR5 blend exhibits positive deviation of rio values from equation ( 1 ) ( see Table 8 ); whereas negative deviation of % values from equation ( 1 ) is shown by iPP/EPR7 blend [ 27 ].

Table 11

Values of zero-shear viscosity ( rlo ), m, ~, ~t and range of EPR particle size ( D ) f o r iPP/EPR blends," the EPR propylene content ( C3 ) are also reported.

SAMPLE

%(P)

m

c~ (sec)

g

D (lam)

C3 %

iPP/EPR1

72850

2/3

0.212

4.5

0.1+1.0

28

iPP/EPR5

287356

4/7

1.143

2.0

0.1+0.6

43

iPP/EPR7

65163

2/3

0.198

7.5

0.1+1.5

43

iPP/EPR8

171615

4/7

1.089

-_-2.0

0.1+0.4

58

271 With increasing EPR

C 3 content

further ( 58 % wt/wt ) ( EPR8 copolymer ),

iPP/EPR8 blend shows a dependence of logarithm of modulus value of complex viscosity (Irl*l)upon logarithm of frequency (co)

quite comparable to that

shown in Fig. 8 by iPP/EPR5 blend [ 29 ]; the values of log

[n*lbeing in

agreement with those predicted by equation ( 1 ) indicating no mutual influence between the blend components. Mnreover as sho,-i~: Table 11 the numberaverage particle size and particle size range of EPR8 phase result to be determined by the melt phase viscosity ratio, as predicted by the TaylorTomotika theory. Impact strength values shown by such iPP/EPR pairs [ 27, 29 ] confirm, in agreement with results obtained studying iPP based blends containing different EPR copolymers, that better impact properties for test temperature higher than EPR Tg are shown by blends containing EPR phase dispersed in spherical shaped domains ranging in average size between 0.35 gm and 0.40 gm. From this study it has been shown that the mode and state of dispersion of EPR phase with high C3 content and then the impact behaviour of such iPP/EPR blend materials can be controlled by optimizing melt phase viscosity ratio. As a matter of fact the increase of propylene content from 28% up to 58% ( wt/wt ) along the chain of random EPR copolymers does not cause iPP/EPR systems to become more interconnected, as no evidences of higher affinity between propylenic sequences of copolymer and iPP matrix macromolecules have been found both in the melt and in the condensed state. Therefore it has been concluded that the propylene content of EPR random copolymers, at least up to 58% wt/wt, is no a relevant structural factor in determining phase structure development in iPP/EPR blend systems, which toughening of such materials depend from.

272 3.

Influence of the crystallization conditions

The influence of the crystallization conditions on the structure of phases and inter-phase developed after complete crystallization of the iPP has been studied under controlled crystallization conditions by means of Optical Microscopy ( OM ) and Small Angle X-Ray Scattering ( SAXS ) in film samples isothermally crystallized at relatively low undercooling

( A T ) in a range of

crystallization temperature ( T~ ) ( 123 + 145 ~ ) by D'Orazio et al. [ 28 ]. In the same paper the combined effect of undercooling and EPR molecular structure and composition on the kinetic and thermodynamic parameters related to the crystallization process of the iPP phase has been also investigated. Furthermore the thermal behaviour and the super-reticular structure of the iPP phase developed in the isothermally crystallized blend samples have been compared with that developed in the injection molded blend samples [ 29 ]. The obtained results are summarized as follows: 9 The radial grow rate of the iPP spherulites is independent of the presence of EPR phase. 9 The apparent melting temperature tends to decrease only when the preliminary crystallization occurred at a low value of AT ( T~ = 145 ~

)

( see Table 12 ). 9 The equilibrium melting temperature ( Tm ) of the blends, determined from equation [ 46 ]:

T' m = Tm (1 - 2cye / AH. L c )

(

4

)

where T' m is the apparent melting temperature, Tm the equilibrium melting temperature, AH ( 47 ) the melting enthalpy of fusion of 100% crystalline

273 iPP and L c the lamellar crystal thickness, is characterized by a slight depression unaffected by composition and molecular characteristics of the EPR copolymer ( see Table 13 ). Taking into account that iPP and EPR are immiscible in the melt such observed depression has been ascribed to kinetic and morphological effects rather than to a thermodynamic effect. The crystallinity of the iPP phase, for a given Tc, depends on the EPR content. As a matter of fact the crystallinity index of the iPP phase ( X c ) ( see Table 14 ) decreases with increasing the EPR content irrespective of its molecular characteristics, indicating that the presence of the rubbery phase influence the crystallization process of the iPP. The desmeared SAXS profiles of all the blend samples, isothermally and non-isothermally crystallized, exhibit well defined maxima; typical Lorentz corrected desmeared patterns are shown in Figs. 10 and 11.

Table 12 Observed melting temperatures ( T'm ) f o r plain iPP and iPP crystallized isothermally from its blends with EPR copolymers as a function of crystallization temperature ( Yc ). T'm (~

SAMPLE Tc=123~

Tc=128~

iPP

162

163

iPP/EPR 1 90/10

161

iPP/EPR1 80/20

Tc-133~

Tc=138~

Tc=145~

166

169

176

162

166

169

173

161

162

165

168

174

iPP/EPR2 90/10

162

162

165

168

173

iPP/EPR2 80/20

161

163

165

168

174

274

Table 13 Equilibrium melting temperature ( Tm ) and surface free energy of folding ( cye ) of plain iPP and iPP crystallized from its blend with EPR copolymers.

Tm(~

SAMPLE

(3"e ( erg/cm 2 )

iPP

195

133

iPP/EPR1 90/10

193

100

iPP/EPR1 80/20

189

75

iPP/EPR2 90/10

190

105

iPP/EPR2 80/20

191

85

Table 14 Crystallinity index of iPP phase, ( Xc ) as a function of crystallization temperature ( Tc ). Xe (iPP)% SAMPLE Tc=123~

Tc=128~

Tc=133~

Tc=138~

Tc=145~

iPP

43

44

43

45

46

iPP/EPR1 90/10

34

35

35

36

38

iPP/EPR1 80/20

26

27

28

29

30

iPP/EPR2 90/10

35

36

36

37

38

iPP/EPR2 80/20

28

29

29

29

30

275

16-

I

14

-

12

-

lo-

(a...)

8

-

6--

4

-

2

-

/ I

I

I

0.002

I

I

0.004

I 0.006

s(nm) ra

Fig. 10

iPP

+

iPP/EPR ( 9 0 / 1 0 )

0

iPP/EPR ( 80/20 )

Typical desmeared SAXS patterns of isothermally crystallized iPP and of iPP isothermally crystallized from its blends with EPR copolymer.

The distance between two adjacent crystalline lamellae of the iPP, i. e. the long spacing ( L ) has been derived from the maxima position by applying the Bragg equation. Assuming for the iPP spherulite fibrillae a two phase model, consisting of alternating parallel crystalline lamellae and amorphous layers, from the L values the crystalline lamella thickness ( Lc ) has been calculated by using the following relation:

L~

-

Xc(iPp) "L ( Pc / P a )(1 - X~(~pp) ) + Xc(iPp

(

5

)

276 where Xc(iPp) is the DSC crystallinity index of the iPP phase and 9c and Pa

are the densities of the crystalline and amorphous iPP phase respectively. Subtracting the obtained Lc values from the L values, the average thickness of the amorphous inter-lamellar layer ( La ) has been obtained. In this calculation, in agreement with what found by SEM analysis, the EPR domains have been assumed to be located in inter-fibrillar regions. 16000 -

iPP/EPR

12000 A

t~ r

.m

8000 s_

,.,..=

iPP

4000

'

0

!

5

9

!

10

9

i

15

9

i

20

S 10 3 ( A ~ )

Fig. 11

Typical desmeared SAXS patterns of injection molded samples of plain iPP and iPP/EPR blends.

The L, Lc and L a values of injection molded samples of plain iPP and iPP/EPR blends are reported in Table 15; in Table 16 the Lc and L a values of isothermally crystallized samples of plain iPP and iPP/EPR blends as a function of crystallization temperature ( Tc ) are reported. As shown in such tables, the thickness of the crystalline lamella of the iPP isothermally and non-isothermally crystallized from its blends decreases,

277 whereas the thickness of the amorphous inter-layer increases. Thus when iPP crystallizes in presence of EPR copolymers, the phase structure developed in the blends is characterized by lamellar thickness and inter-lamellar amorphous layer, respectively, lower and higher than that shown by plain iPP.

Table 15 Long period ( L ), lamella thickness ( Lc ) and inter-lamellar amorphous thickness ( La ) for injection molded samples of plain iPP and iPP/EPR blends. 0

0

0

SAMPLE

L(A)

Lc(A)

La(A)

iPP

161

56

105

iPP/EPR1

163

47

116

iPP/EPR2

168

45

123

iPP/EPR3

174

45

129

iPP/EPR4

166

45

121

iPP/EPR5

172

48

124

Moreover for a given Tc, both Lc and

La

values depend on the blend

composition; in fact with increasing EPR content in the blend the iPP lamellar thickness decreases, whereas the inter-lamellar amorphous layer increases. In order to explain these results it has been assumed that EPR molecules with low molecular masses, because of their higher mobility, diffuse into the iPP inter-lamellar amorphous layer forming domains more or less interconnected with the amorphous iPP phase, thus increasing its thickness and hindering the iPP crystal growth.

278

Table 16 Lamella thickness ( Lc ) and inter-lamellar amorphous thickness ( La ) for isothermally crystallized samples of plain iPP and iPP/EPR blends. Tc = 123 ~ SAMPLE

O

T c = 133 ~ O

O

Tc = 138 ~ O

O

Tc = 145 ~ O

O

O

Lc(A) La(A) Lc(A) La(A) Lc(A) La(A) Lc(A) La(g)

iPP

81

108

90

120

111

135

134

157

iPP/EPR1 90/10

68

131

77

142

87

156

106

174

iPP/EPR 1 80/20

47

135

60

156

66

162

81

190

iPP/EPR2 90/10

66

123

81

144

87

154

105

171

iPP/EPR2 80/20

52

132

63

153

69

168

92

216

9 The surface free energy of folding cye of the iPP lamellar crystals determined according to equation ( 4 ) are strongly depressed by the presence of the EPR phase almost irrespective of its molecular characteristics; this effect also being composition dependent ( see Table 13 ). The finding that the iPP crystals grown in presence of EPR phase have a less regular surface has been accounted for assuming that the iPP crystal surface is perturbed by the diffusion of EPR molecules with low molecular masses into the iPP inter-lamellar amorphous layer. 9 The average dimension of the iPP spherulites decreases with increasing undercooling; furthermore for a given T~ their neatness and regularity decreases with increasing the EPR content ( see Fig. 12 ).

279

Fig. 12

Optical micrographs taken at crossed polarizers of plain iPP and iPP/EPR (30/70) blends isothermally crystallized at 124 ~ 134 ~ and 139 ~

Large amorphous inter-spherulitic contact regions at very low AT ( T~ = 145 ~ ) are developed whose dimensions tend to increase with increasing EPR content. Moreover the growth of this amorphous inter-spherulitic regions is more pronounced for blends containing EPR2 copolymer ( see Fig. 13 and

14 ). This observation has been accounted for the smaller

dimensions of EPR2 particles dispersed in the melt ( see Table 3 ), which may be more easily ejected by the crystallization front.

280

Fig. 13

Optical micrographs taken at crossed and parallel polarizers of iPP/EPR1 ( 90/10 ) blends isothermally crystallized at 124 ~ 133 ~ and 145 ~

From all the above results it has been concluded that the influence of Tc on toughening of iPP/EPR blends is to be related to the following. At low Tc the iPP crystallizes at a very high rate and the chains are not able to disentangle. The material will exhibit a microspherulitic texture and crystallites with relatively thin lamellae with a noticeable number of tie molecules, i.e. molecules linking the crystallites among them ( this is analogous to the

281 entangled network originally existing in the melt ); the EPR particles are suddenly trapped by the fast growing iPP spherulites. Therefore homogeneous, interconnected and ductile material is produced.

Fig. 14

Optical micrographs taken at crossed and parallel polarizers of iPP/EPR2 ( 9 0 / 1 0 ) blends isothermally crystallized at 123 ~ 133 ~ and 145 ~

On the other hand, with reducing undercooling the crystallization rate decreases, the macromolecules in the melt are able to disentangle from each other and to migrate towards the crystalline substrates; few tie molecules, able to carry the load, bridge the relatively thicker formed lamellae. Moreover at low AT

282 molecules with low tendency to crystallize can be rejected into the inter-lamellar and inter-spherulitic regions. Also the EPR particles can be partly rejected during the crystallization by the advancing front of the growing spherulites. A brittle behaviour will be, in this case, exhibited by iPP/EPR materials.

4.

Concluding remarks

It has been demonstrated that, for a given isotactic polypropylene/ethylenepropylene random copolymer pair, molecular mass and molecular mass distribution of both components are the dominant structural factors together with crystallization conditions in determining the phase structure in the melt and in the solid state, which mainly toughening of iPP/EPR systems depends from. Assuming that the iPP crystallization process especially if it is very fast ( high undercooling ) freezes the melt morphology of the EPR amorphous phase, it has been shown that, for a given blend composition and mixing procedure, a general correlation is established between mode and state of dispersion of EPR domains, in the melt and in the condensed state, and melt rheological parameters of the components. Such a correlation is between EPR dispersion degree ( range of EPR particle size and EPR number-average particle size ), as determined in crystallized samples, and melt phase viscosity ratio. The type of dependence of the size of EPR dispersed particles upon the melt phase viscosity ratio is found to agree qualitatively with the prediction of the Taylor-Tomotika theory; both range of EPR particle size and EPR number-average particle size increase with increasing melt phase viscosity ratio. It has been moreover shown that a general correlation is established also between toughening of iPP/EPR blends and dispersion degree of EPR

283 copolymers. Such a correlation is between the impact strength values of iPP/EPR blends and the number-average particle size of EPR copolymers (Dn). For test temperatures higher than EPR Tg, the value of D n able to optimize the toughening of such materials is fixed at 0.35 + 0.40 lam. Taking into account that D n value increases with increasing logarithm of melt phase viscosity ratio ( l a ) , showing a minimum in the vicinity of la = 1, the toughening of iPP/EPR blends can be directly correlated with the melt phase viscosity ratio; i.e., for a given iPP sample, with molecular mass and molecular mass distribution of EPR copolymer. In additional it has been shown that, for a given crystallization process, the toughening of iPP/EPR blends will be partly dependent also upon crystalline texture, as at the end of crystallization process iPP/EPR material is characterized by the presence of iPP spherulites ( larger or smaller according to nucleation density and undercooling ), that have occluded mainly in intra-spherulitic regions EPR domains. It has been demonstrated that EPR copolymers modify the inner structure of iPP spherulite ( crystalline lamella thickness and amorphous inter-layer thickness ), and physical structure of inter-spherulitic boundary regions and amorphous inter-lamellar regions. Therefore desired toughening can be imparted to iPP/EPR blended materials by suitably selecting iPP and random EPR components according to their molecular mass and molecular mass distribution and by choosing crystallization conditions able to optimize the overall phase structure.

5.

(~

Abbreviations

and symbols used

Parameter related to the apparent flow unit in a melt

284 C3

Propylene content of EPR phase

D

Range of particle size

Dn

Number-average particle size

DSC

Differential Scanning Calorimetry

AH

Melting enthalpy

AT

Undercooling

EPDM

Ethylene-propylene-diene terpolymer

EPR

Ethylene-co-propylene

EVA

Ethylene-co-vinylacetate Viscosity

q

Dynamic viscosity ( real part of the viscosity ) Imaginary part of the viscosity

q*

Complex viscosity

qO

Zero-shear viscosity

'~O

Zero-shear viscosity calculated assuming the log additivity rule

']a

Apparent viscosity

vii

Viscosity of dispersed phase

r12

Viscosity of matrix Volume fraction Interfacial surface tension Shear rate

G'

Storage modulus

G~

Loss modulus

GPC

Gel Permeation Chromatography

iPP

Isotactic polypropylene Long spacing between lamellae

285 Varicosity La

Amorphous inter-lamellar thickness

Lc

Crystalline lamella thickness

m

Parameter related to shear-thinning of a melt

m

M M

Number-average molecular mass

n

Weight-average molecular mass

W

n

M

Z-average molecular mass

Z n

Mw/M n

Molecular mass distribution

M.F.D.

Mold Filling Direction

MMD

Molecular Mass Distribution Phase viscosity ratio

NDB

Negative Deviation Blends

CO

Frequency of the oscillation

OM

Optical Microscopy

PA6

Polyamide 6

PP

Polypropylene

q

Instability coefficient

R

Diameter of the thread

Pa

Amorphous density

Pc

Crystalline density 2 sin O/L 0 the angle between the atomic plain and both the incident and reflected beams L the wavelength of the X-ray

~e

Surface free energy of folding

SAXS

Small Angle X-ray Scattering

SEM

Scanning Electron Microscopy

286 Tc

Crystallization temperature

Zg

Glass transition temperature

Tm

Equilibrium melting temperature

T' m

Apparent melting temperature

TPO

Polyolefinic thermoplastic elastomer 2~R/~ Crystallinity index

Xc

References

1. Onogi, S., Asada, T. and Tanaka, A. J. Polym. Sci. (A-2) 1969, 7, 171 2. Kryszewski, M., Galeski, A., Pakula, T., Grebowicz, J. and Milezarek, P. J. Appl. Polym. Sci. 1971, 15, 1139 3. Speri, W. M. and Patrick, G. R. Polym. Eng. Sci. 1975, 15, 668 4. Thamm, R. C. Rubber Chem. Technol. 1977, 50, 24 5. Laus, Th. Makromol. Chem. 1977, 60/61, 87 6. Danesi, S. and Porter, R. S. Polymer 1978, 19, 668 7. Plochocki, A. P. in "Polymer Blends" (Eds. D. R. Paul and S. Newman), Academic Press, New York, 1978, Ch. 21, p. 319 8. Kresge, E. N. in "Polymer Blends" (Eds. D. R. Paul and S. Newman), Academic Press, New York, 1978, Ch. 20, p. 293 9. Karger-Koksis, J., Kallb, A. Szafner, A., Bodor, G. and Sengei, Zs. Polymer 1979, 20, 3 7 10. Ho, W. K. and Salovery, R. Polym. Eng. Sci. 1981, 21, 839 11. Karger-Kocsis, J., Kallb, A. and Kuleznev, V. N. Acta Polym. 1981, 32, 578

12. Kojima, M. J. Macromol. Sci,-Phys. (B) 1981, 19, 523

287 13. Stehling, F. C., Huff, T., Speed, C. S. and Wissler, G. Appl. Polym. Sci. 1981,26,2693 14. Karger-Kocsis, J., Kiss, L. and Kuleznev, V. N. Acta Polym. 1982, 33, 14

15. Martuscelli, E., Silvestre, C. and Abate, G. Polymer 1982, 23, 229 16. D'Orazio, L., Greco, R., Mancarella, C., Martuscelli, E. Polym. Eng. Sci. 1982, 22, 536 17. Karger-Kocsis, J. and Kuleznev, V. N. Polymer 1982, 23, 669

18. Dao, K. C. J. Appl. Polym. Sci. 1982, 27, 4799 19. D'Orazio, L., Greco, R., Martuscelli, E. and Ragosta, G. Polym. Eng. Sci. 1983, 23, 489 20. Martuscelli, E. Polym. Eng. Sci. 1984, 24, 563 21. Karger-Kocsis, J., Kall6, A. and Kuleznev, V. N. Polymer 1984, 25, 279 22. Karger-Kocsis, J., Kiss, C. and Kuleznev, V. N. Polym. Commun. 1984, 25, 122 23. Yang, D., Zhang, B., Y., Fang, Z., Sur, G. and Feng, Z. Polym. Eng. Sci. 1984, 24, 612 24. Dao, K. C. Polymer 1984, 25, 1527 25. Karger-Kocsis, J. and Csikai, F. Polym. Eng. Sci. 1987, 27, 241 26. Martuscelli, E. (Rubber modification of polymers: phase structure, crystallization, processing and properties)

in "Thermoplastic Elastomers

from Rubber-Plastic Blends" (Eds. S. K. De, Anil K. Bhowmick), Ellis Horwood, N. Y. 1990 27. D'Orazio, L., Mancarella, C., Martuscelli, E. Polymer 1991, 32, 7, 1186 28. D'Orazio, L., Mancarella, C., Martuscelli, E., Sticotti, G. J. Mat. Sci 1991, 26, 4033 29. D'Orazio, L., Mancarella, C., Martuscelli, E., Sticotti, G. Polymer 1993,34, 17,3671

288 30. D'Orazio, L., Mancarella, C., Martuscelli, E., Sticotti, G. J. Appl. Polym. Sci. 1994, 53, 387 31. Utraki, C.A. and Kamal, M.R. Polym. Eng. Sci. 1982, 22, 96 32. Lee, B. L. and White, I. L. Trans. Soc. Rheol. 1975, 19, 481 33. Nielsen, L E. "Polymer Rheology", Marcel Dekker, New York, 1977 34. Ferry, J. D. "Viscoelastic Properties of Polymers", 2nd Edn, Wiley, New York, 1970 35. Nielsen, L. E. "Mechanical Properties of Polymers and Composities", Marcel Dekker, New York, 1974, Vol. 1 36. Cox, W. P. and Merz, E. H. J. Polym. Sci. 1958, 28, 619 37. Onogi, S. Masuda, T. and Ibaragi, T. Kolloid Z. 1968, 222, 110 38. Onogi, S., Fujii, T., Kato, H. and Ogihara, S. J. Phys. Chem. 1964, 68, 1598 39. Verser, D. W. and Maxwell, B. Polym. Eng. Sci. 1970, 10, 122 40. Cross, M. M. J. Appl. Polym. Sci. 1969, 13, 765 41. Iwakura, K. and Fujimura, T. J. Appl. Polym. Sci. 1975, 19, 1427 42. Taylor, G. I. Proc. R. Soc. Lond. (A) 1934, 146, 501 43. Tomotika, S. Proc. R. Soc. Lond. (A) 1935, 150, 322 44. Tomotika, S. Proc. R. Soc. Lond. (A) 1936, 153, 308 45. Wu, S. Polymer 1985, 26, 1855 46. Hoffman, J. D. SPE Trans. 1964, 4, 315 47. Brandrup, S. and Immergut, E. M. "Polymer Handbook" Interscience, New York 1975, 5, 24

289 CHAPTER 6

POLYAMIDE6/ETHYLENE BLENDS:

A

- co - VINYLACETATE

MODEL-SYSTEM

THERMOPLASTIC/ELASTOMER

OF PAIRS

L. D'Orazio, C. Mancarella, E. Martuscelli

National Research Council of Italy, Institute of Research and Technology

of

Plastic Materials, 80072 Arco Felice -Naples, ITALY.

I.

Introduction

Several ethylene-vinylacetate ( EVA ) copolymers differing in molecular mass and, for a given molecular mass, in vinylacetate ( VAc ) content have been melt mixed

with polyamide6 ( PA6 ) in order to improve its physical-

mechanical behaviour. The aim has been to create a model-system of thermoplastic/elastomer

immiscible

blends

to

rationalize

the

complex

combination of several factors, which the final use properties of such materials depend from. In absence of chemical reactions between blend components the success in obtaining materials with desired properties depends mainly on the control of the factors determining the phase structure in the melt and in the solid state [ 1 - 3 ]. The physical-mechanical behaviour of thermoplastic/elastomer immiscible blends therefore depends on the following factors:

290 1.

the molecular characteristics of both blend components ( composition, constitution, configuration, molecular mass, molecular mass distribution;

2.

the blending, processing and crystallization conditions. Relevant results of the model study carried out on the influence of these

factors in controlling the toughening of PA6/EVA systems in this chapter are reviewed.

2.

Influence of the molecular structure of the EVA component

2.1 Influence of the molecular mass

The influence of the molecular mass of EVA copolymers on the melt rheology, phase structure and properties of PA6/EVA blends has been investigated by D'Orazio et al. [ 4 - 8 ]. The pioneering study was carried out on samples of PA6/EVA blends based on a sample of PA6 having a number-average molecular mass ( M n ) of 2.3.10 4 and containing EVA copolymers having, for constant vinylacetate ( VAc ) content, different molecular masses as measured by their Melt Flow Index ( MFI ) values of 28.0 and 168 g/10 minutes [ 5 ]. Such EVA samples will be referred to as EVAHM and EVALM respectively. In the same paper blends of PA6 with a sample of ethylene-vinylacetate-acrylic acid terpolymer having VAc content and MFI value quite comparable to that of EVAHM copolymer were also investigated in order to weigh the effect of carboxylic pendant groups on the interfacial adhesion between dispersed phase and PA6 matrix. The molecular characteristics of the plain PA6, EVA copolymers and terpolymer are reported in

291 Table 1 together with their glass transition temperature ( Tg ) and observed melting temperature ( T' m ).

Table 1

Molecular characteristics, glass transition ( Tg ) and observed melting temperatures ( T' m ) of plain PA6, EVA copolymers and ethylene-vinylacetate-acrylic acid terpolymer.

SAMPLE

% VAc

MFI

Tg

T' m

% acrylic acid

( wt/wt )

(g/10 min)

( oC )

( oC )

(wt/wt)

48

219

PA6

EVAHM

28.8

28.0

-27

59

EVALM

28.8

168

-21

56

Terpolymer

26.0

175

-22

64

0.5 + 1.0

All the investigated blends were obtained by melt mixing the components in a single screw extruder operating at 50 r.p.m, and at the temperature of 260 ~ Before blending the PA6 was dried in stove at 90 ~ for 24 hours. The examined blend compositions are listed in Table 2. The extrudate materials were injection molded at the temperature of 220 ~ with a processing cycle of 40 sec; the temperature of the mold was 40 ~

All

the obtained samples were stored in a double polyethylene sealed envelope in order to avoid water absorption.

292

Table 2

Blend composition investigated

COMPONENT

BLEND COMPOSITION ( % wt/wt )

PA6

100

95

90

80

0

COPOLYMER

0

5

10

20

100

The analysis by SEM of the EVA mode and state of dispersion developed in injection molded samples reveals that the minor component segregates in spherical shaped domains with anisotropic distribution. At the outer skin of the samples no EVA domains are observed; on moving from the skin towards the core of the samples the concentration of the EVA particles increases with gradient characteristic. The presence of an outer skin consisting of a plain PA6 layer has been explained by preferential wetting of the mold wall with the PA6 as proposed by Southern and Ballman [ 9 ]. The EVA particle size is found to depend on copolymer molecular mass strongly; moreover, according to Han, no evidence of chemical reaction between PA6 and EVA copolymer is observed [ 10, 11 ]. A finer dispersion is achieved in blends containing EVAHM; in such a blend the copolymer domains have a number-average particle size of 1.5 ~tm ( see Fig. 1 ) irrespective of blend composition. On the other hand in the blends containing EVALM the number-average particle size increases with increasing the copolymer content. The values found are 2.0, 3.0 and 5.0 ~tm for blends containing in order 5 %, 10 % and 20 % of EVALM copolymer. Therefore the dispersion coarseness of such EVA copolymers decreases with increasing their molecular mass, i.e. with increasing phase viscosity ratio defined as p = ril/rl: where rl~ is the viscosity of the dispersed phase and 1"12that of the matrix.

293

Fig. 1

Scanning electron micrographs of fracture surfaces of PA6/EVAHM blend ( a ) and PA6/EVALM blends ( b ).

In the case of PA6/ethylene-vinylacetate-acrylic acid blends the spherical shaped domains of the terpolymer have a number-average diameter of 1.5 gm irrespective of blend composition. Furthermore, contrary to what observed in PA6/EVA blends, such domains are embedded into the PA6 with evidence of adhesion ( see Fig. 2 ). Taking into account that the terpolymer has almost the same VAc content and MFI of EVALM copolymer ( see Table 1 ), it has been concluded that the presence of carboxylic pendant groups along the EVA copolymer promotes

294 interfacial interactions, inducing both finer dispersion degree of the minor component and adhesion between the phases.

Fig. 2

Scanning electron micrograph of fracture surfaces of PA6/ethylenevinylacetate-acrylic acid blends.

The Izod impact strength values for room temperature tests of the plain PA6 and of the investigated blends are reported in Fig. 3 as a function of blend composition. As shown, better impact properties are exhibited by blends which lower size of EVA dispersed particles have been measured for ( PA6/EVAHM ); moreover the extent of the observed improvement is larger with increasing the copolymer content. The fractographic analysis of the Izod fracture surfaces of all the investigated blend samples shows that the mode and fracture

mechanism at

room temperature are determined by a combination of shear yielding and multiple crazes formation mechanisms. All the broken surfaces exhibit, in fact, a fracture induction zone covering an area localized in the middle of the notch front with signs of plastically deformed material indicating that a diffuse shear yielding takes place throughout. The remainder of the surface is characterized by

295 the presence of steps, which arise from adjacent sections of the fracture front following paths at different levels.

10 A

~

9

O

E O

~

8

v ,,C

~

7

o

6

C O

E

"~ o

5

N

0

m

m

m

u

m

5

10

15

20

25

% Copolymer ( ~dwt )

9 PA6 ---B-- PA6/EVA(LM) - - ~ - - PA6/EVA(HM) --O-- PA6/Terpolymer

Fig. 3

Izod impact strength of PA6/EVA and PA6/ethylene-vinylacetateacrylic acid blends as a function of composition.

The area of the fracture induction region, and consequently the material volume undergone to plastic deformation, in blend samples is larger than that of plain PA6; moreover such regions exhibit a stress-whitening phenomenon indicating multiple crazes formation during the fracture. To be recalled that no stresswhitening is shown by broken samples of plain PA6. It is worth noting that for a given blend composition larger induction areas and higher volume of stress-whitened material are shown by blends which higher impact strength values are measured for ( PA6/EVAHM ). Therefore the very different behaviour of EVAHM and EVALM copolymers, in affecting amount and interactions between shear yielding and multiple crazes

296 formation fracture mechanisms, has been related to their different average particle size. EVA particles ranging in size between 1.0 ~tm and 2.0 ~tm result more effective for PA6 toughening than EVA particles ranging in size between 2.0 pm and 5.0 ~tm. Taking into account that the EVA dispersion becomes finer with increasing melt phase viscosity ratio, the factor determining the different degree of dispersion of such EVA copolymers, and in additional the toughening of such materials, turns out to be their different molecular masses. It has been interesting to observe that for a given blend composition the improvement in impact strength values exhibited by the blends containing the ethylene-vinylacetate-acrylic acid terpolymer is lower than that exhibited by the PA6/EVAHM blends ( see Fig. 3 ); even tough a better interfacial adhesion is observed ( see Fig. 2 ). Considering that the average particle diameter achieved in both the blends is the same ( 1.5 gm ), this result indicates, in agreement with the results obtained by Wu [ 12 ], that in absence of chemical reactions between the blend components the interfacial adhesion is no an essential factor to realize toughened thermoplastic/elastomer materials. The influence of the molecular mass of EVA phase on the melt rheological behaviour

of PA6/EVA blends has been investigated by means of capillary

rheometer in a range of temperatures ( 230, 240, 250 and 260 ~ ) [ 6 ]. All the blend samples, based on a sample of PA6 with M n= 18000 and containing 10 % ( wt/wt ) of EVA copolymer, were forced through a capillary of radius 0.5 mm and length 30 mm under an applies pressure ranging between 5 9107 dyne/cm 2 and 80. 107 dyne/cm2; the shear rates at the walls ( 3~) were measured at a given temperature.

297 The molecular characteristics, the glass transition ( Tg ) and the observed melting temperature (T'm) of the plain PA6 and EVA copolymers are reported in Table 3.

Table 3

Molecular characteristics, glass transition ( Tg ) and observed melting temperature (T'm) of EVA copolymers together with Tg and Y'm of plain PA6.

SAMPLE

MFI

% VAc

Tg

T'm

( g/10 min )

( wt/wt )

( oC )

( oC )

48

226

PA6 EVA 1

300 + 500

20

- 19

80

EVA2

3.0 + 4.0

20

- 15

88

EVA3

300 + 500

30

- 19

60

EVA5

3.0+4.0

30

-15

63

The dependence of the logarithm of the apparent viscosity ( r l a ) on the logarithm of the shear rate ( ~ ) for all the investigated PA6/EVA blends shows that the apparent viscosities values decrease as the rate of shear increases in the whole range of explored temperatures and shear rate; i.e. PA6/EVA melts are pseudoplastic. The shear dependence of the viscosity has been analyzed by using a modified Cross-Bueche equation [ 13 ]:

~1o = l + ( c ~ y ) q a

m

( 1 )

298 where ~!o is the zero-shear viscosity, a is a parameter that according to Cross should correspond to the characteristic relaxation time related to molecular weight for the linear polymer solution and m gives a measure of the shearthinning of the melt, i.e. a measure of the decrease in viscosity with increasing rate of shear. For polymers with a distribution of molecular masses Cross claims that m is ( M n / M w ) 1/5 [ 14 ]. According to Iwakura et al. [ 15 ] for polymer melts ~ is related to the size of the apparent flow unit; the reciprocal of c~ corresponds to the shear rate at which rla = rio/2. From the lines 1/rla versus 11q

the zero-shear viscosity 1"1o and

cz values are easily obtained from the

reciprocal of the intercept and from the slope respectively. Plots of 1/qa versus 7 m

for the plain PA6 are linear for m = 8/9 ( usually

linearity is obtained with a lower m value ( 2/3 ) ( 16 ) indicating a severe shearthinning in the non-Newtonian region; whereas the PA6/EVA blends show linearity for m = 2/3. The 1"1oand c~ values of the plain PA6 and its blends with EVA copolymers are reported in Tables 4 and 5 as a function of temperature. As shown in such tables,

for all the investigated samples both 1"1oand c~ values decrease with

increasing the temperature. To be noted, moreover, that for a given temperature qo values lower and higher than that of the plain PA6 are respectively shown by blends containing EVA phase with lower ( EVA1 and EVA3 ) and higher ( EVA2 and EVA5 ) molecular mass. As far as ~ parameter is concerned, for a given temperature c~ blend values increase with increasing EVA molecular mass. Such rheological results, showing that in absence of shear the molecular mass of the copolymer determines the flow mechanism of the PA6 in presence of EVA

299 phase in the form o f m o r e or less d e f o r m a b l e droplets, have been interpreted as follows.

Table 4

Application of Cross-Bueche equation." values of zero-shear viscosity ( rio ) and c~ parameter for plain PA6 and PA6/EVA blends.

T (~ C )

PA6

PA6/EVA1

PA6/EVA2

rio ( P a . s)

ot ( s )

11o ( P a . s)

ct ( s )

qo ( P a . s)

ot ( s )

230

347.2

6 . 4 6 . 1 0 -4

212.8

1 . 1 9 . 1 0 -3

618.5

4 . 9 9 . 10 -3

240

247.4

4.77 910 -4

140.9

8.20 910 -4

401.8

2.91 910 -3

250

182.6

3.36 910 -4

105.3

6.64 910 -4

265.3

1.75 910 -3

260

159.0

3.07 910 -4

78.1

6.09 910 -4

164.2

9.23 910 -4

Table 5

Application of Cross-Bueche equation." values of zero-shear viscosity ( rio ) and ot parameter for plain PA6 and PA6/EVA blends.

T (~ C )

PA6 qo ( P a . s)

PA6/EVA3 a

( s)

PA6/EVA5

1"1o( P a . s)

c~ ( s )

11o ( P a . s)

c~ ( s )

230

347.2

6.46 910 -4

266.7

1.78 910 -3

729.3

6.61 910 -3

240

247.4

4.77 910 -4

177.0

1.10 910 -3

441.6

3.39 910 -3

250

182.6

3 . 3 6 . 1 0 -4

137.4

9 . 5 0 . 1 0 -4

339.2

2 . 6 2 . 10 -3

260

159.0

3.07 910 -4

104.5

8.41 910 -4

221.6

1.42 910 -3

300 Whereas EVA copolymers with low molecular mass reduce PA6 entanglement capability likely by increasing average-end-to-end distance between PA6 macromo|ecules and/or by increasing PA6-EVA interface and slippage, with increasing EVA molecular mass PA6/EVA systems may be described as a continuous two-phase model with macromolecules of one phase entrapped into macromolecules of the other phase. For PA6 and PA6/EVA blends plots of log 1"1oversus the reciprocal of the temperature are straight lines; thus the temperature dependence of the viscosity has been accounted for by an exponential relation of the type [ 17 ]:

(2)

rio - A exp(AE* / RT)

where A is a constant and AE* is the activation energy for the viscous flow.

Table 6

Activation energy ( AE* )values according to equation ( 2 ) of plain PA6 and PA6/EVA blends together with Melt Flow Index ( MFI ) of EVA copolymers.

EVA MFI (g/10 min)

SAMPLE

AE*. 103 ( J/mol )

PA6

26.02

PA6/EVA1

32.34

300 + 500

PA6/EVA2

42.22

3.0 + 4.0

PA6/EVA3

30.10

300 + 500

PA6/EVA5

36.90

3.0 + 4.0

301 As shown in Table 6 the blend AE* values are higher than that of the plain PA6 and increase with increasing EVA molecular mass indicating a growth of the volume of the flow element. The analysis by SEM of the EVA mode and state of dispersion developed in the extrudate blend samples shows an anisotropic distribution of the dispersed phase ( see Fig 4 ); the size of the EVA domains increases on moving from the skin towards the core of the filaments with gradient characteristic, in agreement with results obtained for injection molded samples of different PA6/EVA blends.

Fig. 4

Scanning electron micrograph of transversal smoothed surface exposed to boiling xylene vapours of PA6/EVAHM extruded blend.

Spherical shaped domains by EVA phase in the core of the filaments are formed ( see Fig 5 ); such domains are more or less elongated along the flow direction by higher shear-stress occurred in the outer regions of the filaments. The measured ranges of EVA particle size are summarized in Table 7 together with the melt rheological parameters of the blends and EVA Melt Flow Index.

302

Fig. 5

Scanning electron micrographs of cryogenical fracture surfaces of extruded samples of PA 6/EVA1 and PA 6/EVA2 blends.

Table 7

Zero-shear viscosity ( rio ) and characteristic relaxation time ( ot ) values at extrusion temperature ( 240 ~ C), activation energy for viscous flow ( AE* ) values and range of particle size ( D ) of PA6/EVA blends; the EVA Melt Flow Index ( MFI ) is also reported.

SAMPLE

EVAMFI rlo(Pa-s) ( g / 1 0 rain )

c~ ( s )

AE*.103 ( J/mol )

D (pm)

PA6/EVA1

300 + 500

140.9

8.20. 10 -4

32.34

2.0 + 14.0

PA6/EVA2

3.0 + 4.0

401.8

2.91 . 10 .3

42.22

2.0 + 7.0

PA6/EVA3

300 + 500

177.0

1.10 910 .3

30.10

2.0 + 14.0

PA6/EVA5

3.0 + 4.0

441.6

3.39 910 -3

36.90

0.8 + 4.0

As shown, in such a table and by Figs. 5 and 6, the dispersion coarseness of EVA copolymers is confirmed to increase with decreasing their melt viscosity, i.e. with decreasing phase viscosity ratio, in agreement with the results

303 previously shown obtained studying injection molded samples of different PA6/EVA blends [ 5 ]. An opposite trend was observed by D'Orazio et al. in isotactic polypropylene/ethylene-propylene copolymers ( iPP/EPR ) blends; the results of this study have been shown in detail in the previous chapter.

Fig. 6

Scanning electron micrographs of cryogenical fracture surfaces of extrudate samples of PA 6/EVA3 and PA 6/EVA5 blends.

The type of dependence of the size of the dispersed particles upon the phase viscosity ratio observed in the case of PA6/EVA and iPP/EPR blends agrees qualitatively with the prediction of the Taylor-Tomotika theory [ 18 - 20 ]. According to this theory, the instability coefficient ( q ) of a cylindrical thread suspended in a viscous liquid is given by the following expression:

q - ~-----~--(1-x2)F(x,g) - "/ ff2(x,g) 211oR 2qoR

( 3 )

304 where ~/ is the interfacial surface tension, x = 2~R/~,, R is the diameter of the thread, ~ is the varicosity of the thread and ~t is the phase viscosity ratio. It was found by Tomotika that, for a given value of ~t, the maximum instability occurred at a certain definite value of ~, indicating that drops of definite size would be formed and that ~ changed with ~t. According to this theory plot of average particle diameter ( D n ) versus

log ~t

should show a minimum in the vicinity of ~t- 1 ( see Fig. 7 ) [ 18, 21 ]. Referring to Fig. 7 the different trends in the plot of D n versus log ~t observed for PA6/EVA and iPP/EPR blends are accounted for by assuming that the data points of PA6/EVA blends lie on the left-hand branch of the curve, the opposite occurring for iPP/EPR blends.

PA6/EVA

iPP/EPR

r

E3

!~=1

Log

Fig. 7 Average diameters ( D n ) of dispersed particles as function of the logarithm of phase viscosity ratio ( ~t ). Trend as predicted by TaylorTomotika [ 18, 21 ].

305 2.2 Influence of the vinylacetate content

The influence of the vinylacetate content ( VAc ) of EVA phase has been studied in PA6/EVA blends based on same PA6 sample used in the study concerning with the influence of EVA molecular mass, but containing three copolymers ( EVA4, EVA6 and EVA7 ) having, for the same MFI value, increasing VAc content [ 5 - 8 ]. The molecular characteristics, glass transition temperature ( Tg ) and observed melting temperature ( T'm ) of such copolymers are reported in Table 8. Table 8

Molecular characteristics, glass transition temperature ( Tg ) a n d observed melting temperature ( T'm ) of EVA copolymers together with Tg and T' m of plain PA 6.

SAMPLE

MFI

% VAc

Tg

T'm

( g/10 min )

( wt/wt )

( ~C )

( oC )

48

226

PA6 EVA4

30 + 40

30

-17

60

EVA6

30 + 40

35

-16

53

EVA7

30 + 40

40

- 15

55

The blends, all containing 10 % (wt/wt) EVA, were obtained by extruding the two components in a double screw extruder operating at 80 r.p.m, and at a temperature of 240 ~

The extrudate materials were injection molded at a

temperature of 230 ~ with a processing cycle of 30 sec; the temperature of the mold was 40 ~

306 The rheological parameters of such blends have been derived as in the previous paragraph by using the modified Cross-Bueche equation

[ 13 ]; the

results are summarized in Table. 9. As shown in such a table, for a given temperature the blend rio values increase with increasing the VAc content; the extent of such an increase being larger at lower temperatures. The activation energy for the viscous flow ( AE* ) values, derived according to the exponential relation ( 2 ), are 27.65. 103 J/mol for PA6/EVA4 blend, 31.67.

103 J/mol for PA6/EVA6 blend and 47.66. 103 J/mol for PA6/EVA7

blend.

Table 9

T (~ )

Application of Cross-Bueche equation." values of zero-shear viscosity ( qo ) and ct parameterforplain PA6 andPA6/EVA blends.

PA6/EVA4

PA6/EVA6

PA6/EVA7

qo (Pa. s)

c~ ( s )

qo (Pa. s)

ot ( s )

11o (Pa. s)

c~ ( s )

230

534.1

4.23. 10 .3

626.3

6.08. 10 -3

1043.3

1.27. 10 -2

240

433.2

3.83 910 -3

395.6

3.30 910 -3

549.3

5.77 910 -3

250

290.2

2.29 910 -3

338.0

3.32 910 -3

386.9

3.79 910 -3

260

225.3

2.08 910 -3

229.9

1.96 910 -3

239.6

1.80 910 -3

It is worth noting that the volume of the flow element in PA6/EVA melts increases with increasing VAc content along EVA chain. Such a result has been accounted for by assuming that with increasing the VAc content some kind of association between the two types of macromolecules occurs. The VAc content along EVA chain affects in additional the thermal behaviour of EVA component in PA6/EVA blends. The D.S.C. thermograms of

307 PA6/EVA injection molded blends show endothermic peaks due to the melting of the ethylene blocks of EVA copolymers; the corresponding temperatures increase with increasing copolymer VAc content, as the following

T' m

values are

measured: 75 ~ for EVA4 component, 80 ~ for EVA6 component and 85 ~ for EVA7 component. By comparing such T' m values with those shown by plain EVA copolymers ( see Table 8 ) it emerges that the T' m values of EVA phase in the blend overcome those of plain EVA copolymers noticeably. Such results have been related to a probable dissolution into PA6 of EVA macromolecules with higher number of VAc groups and presumably lower molecular mass; owing to such a dissolution EVA phase with longer ethylenic sequences and then higher T' mvalue is found to remain. The calorimetric measurements performed showed, moreover, that plain PA6 and PA6 crystallized in presence of EVA phase exhibit multiple fusion peaks probably due to recrystallization phenomena in the course of heating and polymorphic transition [ 7,8 ]. It is well known that the PA6 may crystallize from the melt in two different crystalline forms [ 22 ], namely ~ and 7, and that the nucleation, growth and relative amount of such forms depend on the crystallization conditions. For the crystallization conditions imposed by the injection molding process it was found by means of WAXS that 32 % of plain PA6 crystallizes in 7 form. The WAXS patterns ( Cu Kct, Ni-filtered radiation), were collected by a RigakuDenki MicroLaue camera ( sample-film distance = 4 cm ) and analysed by a microdensitometer. From the densitometer traces the ~/ form index ( Iv ) was calculated for plain PA6 and PA6 crystallized in presence of EVA phase by applying the following Kyotami [ 23 ] equation:

308 H 2

I,y = H1 + H2 + H3

( 4 )

where H1 and H3 are the heights of the (200) and (002) + (202) crystalline reflections of ~ form and H2 is the height of the (001) crystalline reflection of PA6 7 form, in 2 0 region between 20 ~ and 25 ~ It is worth noting that I v values exhibited by PA6 crystallized in presence of EVA copolymers are higher than that exhibited by plain PA6 and increase with increasing EVA VAc content ( see Fig. 8 ). From such results one infer that the capability of EVA copolymer to nucleate PA6 7 crystalline form increases with increasing the VAc content along its chain.

PA6/EVA7

40 PA6/EVA5

38 PA6/EVA6 PA6/EVA2

36

~

34

32

30 0

10

20

30

40

EVA VAc % ( wt/wt )

Fig. 8

Index of PA6 7 form ( Iv ) versus EVA VAc content for plain PA6 and PA 6/E VA blends.

309 The analysis by SEM of the mode and state of EVA dispersion, performed on smoothed and subsequently etched transversal and longitudinal surfaces of PA6/EVA injection molded samples, reveals a layered structure according to the schematic model reported in Fig. 9. To be recalled that an anisotropic distribution of EVA phase has already been shown by extruded and injection molded samples of blends containing different EVA copolymers [ 5, 6 ]. As shown in Figs. 9 and 10 moving on from the skin toward the core of the sample the following four layers are observed: 1. an outer skin layer presumably containing no copolymer; 2. a skin layer where the EVA segregates into ellipsoidal and/or cylindrical shaped domains tangentially

to the flow direction according to the flow

pattern proposed by Tadmor [ 24 ] on the observation of Rose [ 25]; 3. an intermediate layer where the ellipsoidal and/or cylindrical shaped domains tend to assume a more or less spherical shape;

tb

d/5~

q- t)

(--)

(7)

-,

O

(-)

C

..

(-) () (-)

o

f--

f-

~-)

O

(~

(3

-

() () (~ ( ) O o I

d--~

o (D

MFD 4

~-) d)

M, t

X

Fig. 9

Schematic model of layered structure of injection molded samples of PA6/EVA blends transversally to Melt Flow Direction ( MFD ), sample thickness (1 mm ); outer skin of plain PA6 ( M ) , skin layer (S), intermediate layer ( I ) and core (C).

310 4. a core showing EVA droplet-like morphology to be ascribed presumably to relaxation and/or breaking up of the previously formed domains.

Fig. 10

Scanning electron micrographs of smoothed surfaces exposed to boiling xylene vapours and of cryogenic fracture surfaces of PA6/EVA4 and PA6/EVA 7 blends with reference to Fig. 9.

The layered distribution generated in PA6/EVA blends by injection molding process with reference to Fig. 9, together with the stereological parameters of EVA domains, are summarized in Table 10. As shown in such a table the thickness of the layer free of EVA domains exhibited by PA6/EVA4 blend is about 2.5 times as high as that shown by PA6/EVA7 blend; such a finding has been related to the different viscosity of the two melts at the injection temperature ( 230 ~ ) ( see Table 9 ).

311

Table 10 Layered distribution in injection molded samples of PA6/EVA blends with reference to Fig. 9." layer code, layer thickness, shape and size of EVA domains. Layered distribution with reference to Fig. 9

M

S

PA6/EVA4

PA6/EVA7

layer thickness %

2.4

0.8

layer thickness %

10

10

ellipsoidal

ellipsoidal

minor axis = 0.2 + 0.4

minor axis - 0.6 + 1.2

major axis = 0.6 + 1.6

major axis -- 1.4 + 6.0

9

5

shape of EVA domains

spherical

ellipsoidal and spherical

size of EVA domains (~tm)

0.4 + 1.0

minor axis - 0.6 + 1.0 major axis = 2.4 + 4.4 spherical - 0.2 + 0.8

29

34

shape of EVA domains

spherical

spherical

size of EVA domains (~tm)

0.6 + 2.6

0.8 + 2.4

shape of EVA domains

size of EVA domains (gin)

layer thickness % I

layer thickness % C

To be noted that in both S and I layers finer dispersion and higher resistance to deformation along the Melt Flow Direction are shown by EVA copolymer with comparatively lower VAc content ( EVA4 ), whereas in the core of the samples EVA particle size is almost independent of EVA VAc content ( see Table 10 and

312 Fig. 10 ) and is to be expected according to melt phase viscosity ratio (see

Fig. 7). The tensile elastic behaviour of PA6/EVA blends strongly depends upon

EVA VAc content as shown by Fig. 11 reporting the trend of the elastic modulus ( E ) values of plain PA6 and PA6/EVA blends versus EVA VAc content [ 7 ].

PA6/EVA4 ~ P A 6 / E V A 5

] - -

~ 3

6/EVA2

PA6/EVA6

PA6/EVAI

9

0

,

10

.

,

20

,

|

30

9

,

40

,

|

5O

EVA VAc % ( wt/wt )

Fig. 11

Elastic modulus ( E ) versus VAc content of EVA copolymers for PA 6/EVA blends.

It has been surprising to observe that an increase of VAc content from 20 % to 30 % (wt/wt) results in a modulus increased up to values quite comparable and even higher than that shown by plain PA6. On further increase of VAc content ( 35 % and 40 % wt/wt ) no further E increase is observed ( see Fig. 11 ). The analysis by SEM of the mode and state of EVA dispersion developed in blend samples strained up to the cold-drawing region ( point ( a ) on the curve of Fig. 12 ) shows that the original layered structure reported in Fig. 9 becomes modified according to the schematic model reported in Fig. 13.

313 7oo

(b)

600

/

500

(a)

400

J

/

..~ 300 200 100 9

0

Fig. 12

|

9

1

|

|

2

3

9

|

4

Stress-strain curve for plain PA 6.

j

J MFD c..._~.Q

C

,

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

~__~ .....

~

.......

0

+sL, M

Fig. 13

o

j X

Schematic model of layered structure of injection molded samples oJ PA6/EVA blends strained to cold-drawing," sample thickness ( O.8 ram). Key as for Fig. 9.

314 As shown in Fig. 13 moving from the border towards the core of samples three layers are found instead of four: 1. a skin surface where no EVA domains are observed; 2. an outer layer where EVA domains are deformed along the draw direction assuming mainly ellipsoidal shape; 3. a core where EVA domains assume ellipsoidal shape. The layered distribution developed in the PA6/EVA samples strained just beyond the yield point with reference to Fig. 13 are summarized in Table 11.

Table 11 Layered distribution in injection molded samples of PA6/EVA blends strained to cold-drawing with reference to Fig. 13." layer code, layer thickness, shape and size of EVA domains. Layered distribution with reference to Fig. 13

M

S

PA6/EVA4

PA6/EVA7

layer thickness %

2.0

1.3

layer thickness %

10

11.7

ellipsoidal

fibre

minor axis = 0.4 + 0.8

breadth = 0.2 + 0.6

major axis = 0.8 + 2.4

length = 4.0 + 16

38

37

ellipsoidal

ellipsoidal

minor axis = 0.4 + 1.6

minor axis = 0.4 + 1.6

major axis = 1.6 + 5.0

major axis = 1.2 + 4.4

shape of EVA domains

size of EVA domains (~tm)

layer thickness % C

shape of EVA domains

size of EVA domains (~tm)

315 As shown by the morphological results summarized in such a table, copolymer VAc content results an important parameter in determining the morphological variations induced by straining in the mode and state of dispersion of EVA copolymers developed in S layer. The general trend is that the deformation undergone by EVA domains along draw-direction increases with increasing the copolymer VAc content, as ellipsoidal shaped domains and fibres are respectively shown by EVA4 and EVA7 copolymers. In the core of the samples the size of EVA domains holds almost independent of the EVA VAc content (see Table 11 ). By further straining the dumb-bell shaped specimens to break ( point ( b ) on PA6 stress-strain curve reported in Fig. 12 ), the layered structure developed in the sample strained to cold-drawing becomes modified according to the model schematically shown in Fig. 14.

r ....

j

J MFD

<;; L~> M~

Fig. 14

....

- -

-.....

JzV x

Schematic model of layered structure of injection molded samples of PA6/EVA blends strained to break," sample thickness ( O.6 mm ). Key as for Fig. 9.

As shown, only two layers are found: 1. an outer skin free of EVA domains whose thickness remains unaffected by strain ( see Tables 11 and 12 );

316 2. a core where EVA phase forms ellipsoidal shaped domains; the size of such domains resulting in the break-up and relaxation of EVA domains generated in the samples strained to cold-drawing. The layered distribution generated in injection molded samples strained to break with reference to Fig. 14 is summarized in Table 12. As shown in such a table, irrespective of copolymer VAc content, comparable ranges of minor and major axis by the EVA ellipsoidal shaped domains are shown.

Table 12 Layered distribution in injection molded samples of PA6/EVA blends strained to break with reference to Fig. 13: layer code, layer thickness, shape and size of EVA domains. Layered distribution with reference to Fig. 13 M

C

PA6/EVA4

PA6/EVA7

layer thickness %

2.0

1.1

layer thickness %

48

49

ellipsoidal

ellipsoidal

minor axis = 0.6 + 1.4

minor axis = 0.6 + 1.2

major axis = 1.6 + 6.0

major axis = 2.0 + 5.0

shape of EVA domains

size of EVA domains (pm)

Both PA6 and PA6 crystallized in presence of EVA phase exhibit, after the yielding point, a cold-drawing region preceded by formation of necking and fibre rupture [ 8 ]. Plain PA6 samples show no stress whitening at either cold-drawing or fibre rupture, whereas PA6/EVA samples show at cold-drawing in their central part a

317 more or less stress-whitened zone, fibres becoming completely stress-whitened at break. In particular at the cold-drawing the intensity of the stress-whitening phenomenon increases with increasing the EVA VAc content [ 8 ]. It is worth recalling, in connection with, that with increasing EVA VAc content the dispersion coarseness, shown by EVA copolymers at the cold-drawing, also increases ( see Table 11 ). Taking into account that the stress-whitening phenomenon is to be associated with multiple crazes formation and/or cavitation during tests the above findings, in agreement with those obtained while studying yielding behaviour of PA6/EVA blends [ 7 ], confirm that the capability of EVA domains to initiate multiple crazes formation and/or to be mechanically equivalent to a void, considering the very weak adhesion at interface [ 4 - 8 ], decreases with increasing copolymer dispersion coarseness. The values of the stress at break ( OrB ) and elongation at break ( eB ) of plain PA6 and PA6/EVA blends are reported in Table 13. To be noted that both OrB and eB values shown by PA6/EVA blends are lower than that exhibited by plain PA6 showing that the presence of EVA phase reduces the capability of such materials to be plastically deformed. Nevertheless a general correlation has been established between the final tensile properties of PA6/EVA blends and dispersion coarseness of EVA copolymers, that is between OrB and eB values and number-average of the major axis of EVA ellipsoidal shaped

domains

( D ), as measured by SEM in PA6/EVA fibres strained to break; both OrB and eB values in fact increase with increasing D values ( see Figs 15 and 16 ). Such results have been related to the following: smaller EVA domains are in size, higher are the stress concentrations induced by external load with subsequent lower overall stress and higher is the hindrance generated by EVA

318 phase to PA6 cold-drawing and then the instability in flow, which causes premature rupture of the blend materials.

Table 13 Stress at break ( CYB ) and elongation at break ( eB ) for plain PA6 and PA 6/E VA blends.

SAMPLE

cyB" 102 ( Kg/cm 2 )

eB

PA6

5.8 + 0.6

3.8 + 0.4

PA6/EVA1

3.4 + 0.4

2.4 + 0.4

PA6/EVA2

3.9 + 0.4

3.1 + 0.3

PA6/EVA3

4.5 + 0.3

2.9 + 0.4

PA6/EVA4

4.0 + 0.8

1.8 + 0.7

PA6/EVA5

4.4 + 0.6

2.4 + 0.4

PA6/EVA6

2.8 + 0.1

1.0 + 0.4

PA6/EVA7

3.0 + 0.4

1.3 + 0.6

Taking into account that D values of EVA phase in PA6/EVA samples strained to break have been found to decrease with increasing EVA VAc content [ 7, 8 ], both cyB and ~B values of PA6/EVA blends have been related to such a structural factor of copolymers. Plots of cyB and ~B values of PA6/EVA blends versus EVA VAc content show in fact that both stress and elongation at break of such materials decrease with increasing VAc content along EVA chain ( see Figs. 17 and 18 ). From this study it has been shown that also the VAc content along the chain of EVA copolymer, as well as its molecular mass, is an important structural factor in determining the overall phase structure of PA6/EVA blend systems, which toughening of such materials depends from.

319 600

500

PA6/EVA5 PA6/EVA4

A

e,i

E 400 0

n

PA6/EVA3

/~.~

9 J

9 PA6/EVA2 9 PA6/EVA1

,- 3 0 0 9

PA6/EVA7

PA6/EVA6 200

100

n

2

9

4

|

|

6

8

o

D (pm)

Fig. 15

Stress at break ( orb ) of PA6/EVA blends as a function of numberaverage of major axis of ellipsoidal shaped EVA domains ( D ).

PA6/EVA2 PA6/EVA5 ~

=

PA6/EVA4 e ~,,~,,,,,,,~ j O P A

2

O PA6/EVA3 6/EVA1

9 PA6/EVA7 9PA6/EVA6

9

2

n

9

4

n

6

9

9

8

o

D (pm)

Fig. 16

Strain at break ( gB ) of PA6/EVA blends as a function of numberaverage of major axis of EVA ellipsoidal shaped domains ( D ).

320 700

PA6

600 A

E 500

~

PA6/EVA3 -9 PA6/EVA5

m 400

PA6/EVA2 9

~/EVA4

PA6/EVAI 9

300 200

9

0

m

9

10

l PA6/EVA6 9

u

9

20

m

9

30

PA6/EVA7

ul

9

40

u

50

EVA VAc % ( wt/wt )

Fig. 17

Stress at break ( ~B ) of plain PA6 and PA6/EVA blends as a function of EVA VAc content.

PA6 PA6/EVA2 9 6 ~

9 PA6/EVA3

PA

9 PA6/EVA5

9 PA6/EVA7 PA6/EVA6

9

0

,

10

,

,

20

,

|

30

9

n

40

9

i

50

EVA VAe % ( wt/wt )

Fig. 18

Deformation at break ( ~B ) of plain PA6 and PA6/EVA blends as a function of EVA VAc content.

321 2.3 Influence of the processing conditions

The influence of post-blending processing conditions on the melt rheology, phase structure and impact properties of blends obtained by melt mixing a sample of PA6 with a sample of EVA copolymer having 28.8 (wt/wt) of vinylacetate was investigated by D'Orazio et al. [ 4 ]. The PA6 used throughout the work was SNIAMID ASN 27/S produced by SNIA with a number-average molecular mass ( M n ) of 2.3 9104", the EVA copolymer was the same sample already referred to as EVAHM. The blends were obtained by mixing the components in a single screw extruder ( L/D = 25 mm; = 30 mm ) at 50 r.p.m, and at the temperature of 260 ~

the compositions

investigated are reported in Table 14.

Table 14 Blend composition investigated.

BLEND COMPOSITION ( % wt/wt )

COMPONENT

PA6

100

90

70

40

0

EVAHM

0

10

30

60

100

The extrudate materials were transformed according to two different processing: capillary extrusion and injection molding. The capillary extrusion was performed at very low shear rate by means of a capillary rheometer. The samples were extruded without any capillary directly from the extrusion chamber through a die having a diameter of 5.5 mm at 220 ~

240 ~

and 260 ~

Three different

residence times ( t r ) ( 3, 6, 10 rain ) were used. Injection molded samples were obtained at same molding temperatures as those used for extrusion with the

322 following t r inside the cylinder of the press: 0, 3, 6 min. It was impossible to extend t r up to 10 min because of EVA degradation. The rheological characterization of the single components and of the blends was carried out by means of a constant force capillary rheometer. The samples were forced through a capillary of radius 0.5 mm and length 30 mm under an applied pressure of 3 9106 dyne/cm 2 . The values of apparent viscosity ( Via ) for the plain PA6 were: 1100 P at 220 ~ 260 ~ The

850 P at 240 ~

and 600 P at

for EVAHM copolymer they were 260 P at 170 ~ and 175 P at 180 ~ Via values of the PA6/EVAHM blends as a function of temperature and

residence time are summarized in Table 15.

Table 15 Apparent viscosity values ( 1]a ) as a function of extrusion temperature and residence time ( t r )for PA6/EVAHM blends.

T (~

tr ( m i n )

na ( P )

PA6/EVAHM 90/10

PA6/EVAHM 70/30

PA6/EVAHM 40/60

220

3 6 10

317.7 361.4 333.3

97.1 94.9 95.8

42.9 49.2 48.4

240

3 6 10

333.3 285.7 331.5

99.0 96.7 97.1

55.6 42.2 46.1

260

3 6 10

285.7 256.4 230.7

96.7 101.0 98.4

46.1 50.0 37.5

323 As shown in such a table for a given shear stress and temperature, the apparent viscosity of the blends is lower than that of the plain PA6 and decreases with increasing EVA content with no systematic dependence on t r. For a given composition, the blend rla values are slightly influenced by the temperature at least in the range explored; this finding has been accounted for by degradation of EVA copolymer and/or by modification of the mode and state of dispersion of the minor component. The analysis by SEM of the phase morphology developed in extrudate blends shows that, for a given blend composition, the extrusion temperature and/or the residence time strongly affect the mode and state of dispersion of the minor component. The morphological results for PA6/EVAHM blends containing 10 % and 30 % (wt/wt) of copolymer are summarized in Table 16.

Table 16 Shape and size of EVAHM domains in PA6/EVAHM blends as a function of extrusion temperature and residence time.

T ( ~ ) t r ( min )

shape

size ( ~tm )

PA6/EVAI-IM (90/10)

shape

size ( ~tm )

PA6/EVAHM (70/30)

3

spherical

1.0 + 2.0

spherical

1.0 + 2.0

10

ellipsoidal

minor axis = 1.5 major axis = 2.5

cylindrical

1.3 + 2.6

3

spherical

1.0 + 4.0

spherical

2.0 + 5.0

10

spherical

1.0 + 4.0

spherical

2.0 + 5.0

220

260

As shown in such a table at lower extrusion temperature (220 ~

EVA

domains assume spherical shape for shorter t r (3 min); with increasing t r such

324 domains are deformed along the flow direction assuming ellipsoidal or cylindrical shape. With enhancing extrusion temperature up to 260 ~ irrespective of tr, EVA copolymer segregates in spherical shaped domains; the size of such domains being larger than that shown at lower extrusion temperature. The EVA selective dissolution with boiling xylene vapours showed that at composition PA6/EVA 40/60 (wt/wt) EVA copolymer represents the continuos phase including polyamide as dispersed phase [ 4 ]. As shown by Fig. 19, PA6 is elongated along the extrusion direction giving rise to irregular rods in region closer to the filament border, while assuming spherical shape ( number-average diameter = 5.0 gm ) going toward sample centre. Inside PA6 rods small EVA particles ( number-average diameter lower than 1 gm ) are occluded.

Fig. 19

Scanning electron micrographs of cryogenic fracture surfaces of PA6/EVAI4M ( 40/60 ) blend extruded at 220 ~ with tr of 3 min." a) core, b) border of the filament.

For lower extrusion temperature by increasing tr PA6 rods are observed even in regions closer to the core of the filament. With increasing processing temperature, irrespective of tr, PA6 cylindrical and spherical shaped domains are respectively observed in region next to edge and in the core of the filament. The size of such domains are considerably larger ( diameter ranging between

325 10 gm and 30 gm ) than that shown by PA6 at 220 ~

On the surface of PA6

domains spherulites are clearly observable.

Fig. 20

Scanning electron micrographs of cryogenic fracture surfaces of PA6/EVAHM ( 40/60 ) blend extruded at 260 ~ with tr of 3 rain: a) core, b) border of the filament.

The finding that in blends with EVA matrix dispersed PA6 phase segregates in rod-like structure has been related to the high value of melt viscosity ratio. The values of Izod impact strength of plain PA6 and investigated blends for room temperature tests are reported as a function of composition in Fig. 21. For a given value of tr, the PA6 impact values slightly decrease with increasing the injection temperature; on the other hand, for lower processing temperatures the PA6 impact values slightly increase with increasing tr. Such a behaviour has been related to different crystallization conditions ( density of nucleation and undercooling ) affecting the PA6 phase structure (spherulites size, lamellar and inter-lamellar thickness, structure of inter-spherulitic boundary regions, number and type of tie molecules). The fracture surfaces of plain PA6 broken at room temperature show a distinct induction region starting from the middle of the notch and covering a limited area where PA6 undergoes plastic deformation; the remainder of the

326 sample exhibits a rough topography typical of a fast fracture. No stresswhitening phenomenon is observed. Such results indicate that at room temperature the impact energy is mainly related to yielding process in the fracture induction region.

10 ~E E ~a 8

tr = 0 sec

( 6 [

tr = 180 sec

J~

~a L r "c,a

tr = 360 sec

E N

4

.

9

10

.

9

20

.

30

40

% EVAHM (wt/wt)

10

b

E r

J~

tr = 0 sec L

tr = 180 sec ~L

E

t, = 360 sec

N

4

9

0

|

10

9

|

9

20

|

30

9

|

40

% EVAnM (wt/wt)

Fig. 21

Izod impact strength values of plain PA6 and PA6/EVAHM blends as function of EVA content for different residence times( tr ); a) samples molded at 220 ~ b) samples molded at 260 ~

327 By adding to PA6 10 % (wt/wt) EVA copolymer an improvement in Izod impact is achieved ( see Fig. 21 ). For a given injection temperature, such values tend to increase with decreasing residence time; whereas no appreciable effects of the injection temperature, irrespective of t~, are found. The fracture surfaces of 90/10 PA6/EVA blends show a stress-whitened induction region localized around the notch, where PA6 undergoes plastic deformation; the remainder of the sample is involved in rapid crack propagation. The EVA minor component segregates in spherical shaped domains with a number-average diameter ranging between 1.0 gm and 2.0 gm ( see Fig. 22 ).

Fig. 22

Scanning electron micrographs of Izod impact fracture surfaces of PA6/EVAI4M (90/10) blend," a) fracture induction region, b) fracture propagation region.

Such findings indicate that in the induction region the mode and mechanism of fracture of such

PA6 material result in combined effect of multiple crazes

propagation and shear yielding,

in agreement with results obtained while

studying different PA6/EVA systems [ 5 ]. With increasing EVA content ( 30% wt/wt ), the impact values of PA6/EVA blends decrease. It is worth noting that the amount of such a decrease depends on processing conditions. As shown in Fig. 21 the impact values of the

328 blends decrease with increasing both injection temperature and residence time in the cylinder of the press. The fractographic analysis performed shows that in such blends there is a limited induction region starting from the middle of the notch with slight stresswhitening phenomenon. Moreover the range of the EVA particle size is considerably wider ( 1.0 ~tm + 5.0 lam ) than that shown by blends containing 10 % EVA copolymer ( 1.0 ~tm + 2.0 pm ) ( see Figs. 22 and 23 ).

Fig. 23 Scanning electron micrograph of Izod fracture surface (propagation region ) of PA6/EVAI4M( 70/30 ) blend.

The impact behaviour of PA6/EVA 70/30 blends has been related to EVA dispersion coarseness; particle size ranging between 1.0 gm + 5.0 gm are less effective than particles ranging between 1.0 gm + 2.0 lam in promoting formation and propagation of multiple crazes. The observation that the impact values of such materials decrease with increasing injection temperature and residence time have been related to degradation process of the blend components and/or changes in the PA6 phase structure [ 4 ].

329

3. Concluding remarks

Polyamide6/ethylene-co-vinylacetate blends have been studied in detail as a model-system of thermoplastic/elastomer pairs in absence of chemical reactions between the blend components. It has been demonstrated that the chemical composition and constitution and molecular masses of both the blend components,

together with the processing conditions, are the dominant

structural factors in determining the overall phase structure and consequently toughening of such materials. For a given blend composition a general correlation has been established between the melt rheological parameters of the blends and the mode and state of dispersion of the minor component in the melt, that is between the values of the number-average particle size and the particle size range of the rubbery phase, as determined by SEM in crystallized samples, and the melt phase viscosity ratio. This correlation was made under the approximation that the morphology of the amorphous phase in the melt is frozen by the crystallization process, especially if it is very fast ( high undercooling ). In particular the dispersion coarseness of EVA phase in PA6 matrix decreases with increasing phase viscosity ratio, i.e. with increasing copolymer molecular mass. The type of dependence of the size of the elastomeric particles upon the phase viscosity ratio agrees qualitatively with the prediction of the TaylorTomotika theory. It is worth underlining that analogous correlation and type of dependence of the size of the elastomeric particles upon the phase viscosity ratio has been determined studying very different thermoplastic/elastomer immiscible blends made of iPP and EPR random copolymers ( see chapter 1 ). General correlations have been established also between toughening of PA6/EVA materials and dispersion degree of EVA copolymers. In particular the

330 impact strength values of PA6/EVA blends have been correlated to EVA particle size; it has been shown in fact that in thermoplastic/elastomer pairs the interfacial adhesion is no an essential factor to improve such properties. For room temperature tests EVA particles with comparatively lower size ( 1.0+ 2.0 lam ) result more effective for achieving toughened PA6 based materials. Taking into account that the value of the number-average EVA particles size (Dn) decreases with increasing la logarithm, the impact behaviour of PA6/EVA blends has been related to melt phase viscosity ratio; i.e., for a given PA6 sample, to molecular masses of EVA copolymers. Moreover the uniaxial tensile final properties of PA6/EVA blends have been correlated to number-average size of major axis of EVA ellipsoidal shaped domains (D) generated in PA6/EVA samples strained to break. For room temperature tests both CYBand ~B values of PA6/EVA blends increase linearly with D values. Considering that D values have been found to decrease with increasing VAc content of EVA phase [ 7, 8 ], both CYBand

~B values of

PA6/EVA blends have been related to such copolymer structural factor. Finally it has been demonstrated that the toughening of PA6/EVA blends is partly dependent also upon processing and consequently crystallization conditions, that influence both mode and state of dispersion of minor rubbery component and matrix phase structure ( spherulites size, lamellar and interlamellar thickness, structure of inter-spherulitic boundary regions, number and type of tie molecules ). To sum up required toughening can be conferred to thermoplastic/elastomer blend materials, such as PA6/EVA pairs, by suitably selecting the thermoplastic polymer and elastomer according to their chemical composition and constitution and molecular mass and by choosing post-blending processing conditions able to

331 optimize both the mode and state of dispersion of minor component and crystalline texture.

5. Abbreviations

and symbols used

Parameter related to the apparent flow unit in a melt D

Range of particle size

D

Number-average size of major axis of ellipsoidal shaped domains

Dn

Number-average size of spherical shaped particles

DSC

Differential Scanning Calorimetry

AE*

Activation energy for the viscous flow

AH

Melting enthalpy

AT

Undercooling Elastic modulus

EPR

Ethylene-co-propylene

EVA

Ethylene-co-vinylacetate Strain Elongation at break Viscosity

qO

Zero-shear viscosity

qa

Apparent viscosity

ql

Viscosity of dispersed phase

q2

Viscosity of matrix

7

Interfacial surface tension

332 Shear rate H1

Height of the (200) crystalline reflection of PA6 ~ form

H2

Height of the (001) crystalline reflection of PA6 7 form

H3

Height of the (002) + (202) crystalline reflections of PA6 form

iPP

Isotactic polypropylene

Iy

Index of PA6 y form

Z

Varicosity

m

Parameter related to shear-thinning of a melt

m

Mn

Number-average molecular mass

n

MW

Weight-average molecular mass

M.F.D.

Melt Flow Direction

M.F.I.

Melt Flow Index Phase viscosity ratio

PA6

Polyamide6

q

Instability coefficient

R

Diameter of the thread

r.p.m.

Revolution per minute

SEM

Scanning Electron Microscopy Stress

~B

Stress at break Temperature

Tc

Crystallization temperature

Yg

Glass transition temperature

T

Apparent melting temperature

~m

t r

Residence time

VAc

Vinylacetate

333 27tR/)~ X

Crystallinity index

c

WAXS

Wide Angle X-Ray Scattering

References

Martuscelli, E., Palumbo, R. and Kryszewski, m. (eds) (1980) Polymer Blends: Processing, Morphology and Properties, Vol. I, Plenum Press, New York. ,

Martuscelli, E., Palumbo, R. and Kryszewski, m. (eds) (1984) Polymer Blends" Processing, Morphology and Properties, Vol. II, Plenum Press, New York.

,

,

.

,

D'Orazio, L., Mancarella, C., Martuscelli, E., J. Mat. Sci. 23 (1988) 161 D'Orazio, L., Mancarella, C., Martuscelli, E., J. Mat. Sci. 21 (1986) 989 D'Orazio, L., Mancarella, C., Martuscelli, E., J. Mat. Sci. 22 (1987) 429 Addonizio, M.L., D'Orazio, L., Mancarella, C., Martuscelli, E., J. Mat. Sci. 24 (1989) 2939

,

o

Addonizio, M.L., D'Orazio, L., Martuscelli, E., Polymer 32 (1991) 109 Addonizio, M.L., D'Orazio, L., Martuscelli, E., J. Mat. Sci. 28 (1993) 3793

,

Southern, J. H. and Ballman, R.L., J. Appl. Polym. Sci. 24 (1979) 693

10.

Han, C.D. and Chuang, H.K., J. Appl. Polym. Sci. 30 (1985) 2431

11.

Han, C.D. and Chuang, H.K., J. Appl. Polym. Sci. 30 (1985) 165

12.

Wu, S., Polymer 26 (1985) 1855

13.

Fujimura, T., Iwakura, K. Int. Chem. Eng. 10 (1970) 683

14.

Cross, M.M.J. Appl. Polym. Sci. 13 (1969) 765

15.

Iwakura, K., Fujimura, T. J. Appl. Polym. Sci. 19 (1975) 1427

334 16.

Cross, M.M.J. Colloid Sci. 20 (1965) 417

17.

Van Krevelen, D.W., Properties of polymers, Elsevier Scientific Publishing Company, New York (1976)

18.

Taylor, G.I. Proc. R. Soc. Lond. (A) 146 (1934) 501

19.

Tomotika, S. Proc. R. Soc. Lond (A) 150 (1935) 322

20.

Tomotika, S. Proc. R. Soc. Lond (A) 153 (1936) 308

21.

Karger-Kocsis, J., Kallo, A. and Kuleznev, V.N. Polymer 25 (1984) 279

22.

Holmes, R., Bunn, C.W. and Smith, D.~I. J. Polym. Sci. 17 (1955) 159

23.

Kyotani, M. and Mitsuhashi, S. J. Polym. Sci. (A-2) 10 (1972) 1497

24.

Tadmor, Z. J. Appl. Polym. Sci. 18 (1974) 1753

25.

Rose, W. Nature 191 (1961) 242

335 CHAPTER 7

BLENDS POLYAMIDE 6/FUNCTIONALIZED RUBBER

R. Greco, M. Malinconico, E. Martuscelli, G. Ragosta

National Research Council of Italy, Institute of Research and Technology of Plastic Materials, 80072 ArcoFelice, (Na), ITALY

Introduction

Multicomponent systems consisting of a thermoplastic matrix and a rubbery component are generally prepared according to two different blending routes [1-3] 1. Melt miring of the high molecular weight thermoplastic polymer with an elastomer. 2. Dispersion of the rubbery component during the polymerization of the thermoplastic polymer. The melt mixing method 1 utilizes well established technologies and is therefore generally preferred to obtain rubber modified thermoplastics of improved impact properties As the two components are immiscible, the addition of a third component (compatibilizing agent) is frequently used to increase the adhesion between the elastomeric and the thermoplastic phases and to achieve finer dispersions of the rubbery particles in the thermoplastic matrix [4]. Modified elastomers beating functional groups along the chain are frequently used as precursors of compatibilizing agents because they may form graft copolymers reacting with the thermoplastic polymer during the blending process (Reactive Blending). The functionalized elastomer may also be the only rubber

336 component.

The use of such modified rubbers is particularly attractive for

condensation polymers which have reactive chain ends and are generally obtained by a step polymerization mechanism.

These features greatly enhance the

potential use of method 2. In the following sections, some results achieved in the reactive blending of polyamides (PA) with functionalized ethylene-propylene rubbers (EPM) are reviewed. Several examples dealing with PA6 and PA66/polyolefin systems have been reported in patents [5-9]. The materials are generally melt blended and graft copolymers or networks are formed by means of radical initiators or directly by mechanical degradation, with an appreciable improvement of the impact resistance of the polyamides.

Authors [10] have studied the PA6/isotactic

polypropylene (iPP) system to which an isotactic polypropylene functionalized by insertion of anhydride groups onto flae chain had been added. During the melt mixing the anhydride groups react with PA6 amino end groups to yield an iPP/PA6 grai~ copolymer.

A similar approach has been followed by other

authors on PA6/high density polyethylene blends [11]. In both cases the results were very promising. The previously cited routes 1 and 2 have both been explored for the realization of a EPM-modified PA6 characterized by interesting technological performances, particularly resistance to impacts at temperatures below the ambient temperature, where PA6 is normally deficient due to its glass transition well above room temperature. In the following Sections, moving from the description of solution methods of functionalization of EPM rubbers, the review will address all the aspects of both routes to the achievement of compatibilized systems PA6/EPM.

337 1 Functionalization of amorphous ethylene-propylene copolymer by free radical initiated grafting of unsaturated molecules

Introduction Functionalization of polyolefins can be accomplished by grafting unsaturated molecules beating functional groups through a radical reaction initiated by organic peroxides [12-16]. The grafting reactions were carried out under various experimental conditions and the dependence of the fimctionalization degree on the type of ~-13 disubstituted unsaturated molecules, on the reaction time, on the nature and concentration of the peroxide initiator, was particularly investigated.

1.1 The functionalization reaction The EPM copolymer employed is an ethylene-propylene amorphous copolymer, 60% by moles in C2. In a typical grafting reaction, EPM is dissolved in anhydrous solvent in a flask equipped with a nitrogen inlet and a refrigerator. When the dissolution is complete, the proper amount of the chosen unsaturated molecule is added together with the peroxide, and the reaction is carried out at the selected temperature for a fixed length of time. The reaction is stopped by precipitating the polymer in a non-solvent. Degradation of the polymeric backbone is checked by viscosity measurements. The amount of the grafted anhydride in anhydride modified EPM is determined by titrating the acid groups aRer complete hydrolysis of the anhydride groups. The evaluation of the amount of grafted ester groups in EPM-gdibutylmaleate (EPM-g-DBM) and EPM-g-diethylfumarate (EPM-g-DEF) is performed by analysis of the IR spectra of these copolymers in carbon

338 tetrachloride solutions. A calibration curve is previously obtained from the ratio of absorbance of the band at 1735 c m 1 of the dibutyl and diethyl succinate and of that of the solvent at 1291 c m 1, at different concentrations of ester. In order to insert randomly single functional groups onto polyolefin chains, maleic or fumaric acid derivatives may be considered appropriate unsaturated molecules. In fact, 1,2-disubstituted alkenes have the double bond associated with appreciable steric hindrance and show a low tendency to homopolymerize compared with vinyl monomers. The homogeneous grafting of maleic anhydride (MAH) onto EPM initiated by dibenzoylperoxide (DBPO) follows a reaction pattem that can be schematically represented as follows [17]:

o

O

+R" __RH

~

II

O

(P)

(P')

(P")

O

§ P', S"

P" represents EPM macroradicals generated by hydrogen abstraction from EPM chains promoted by the primary free radicals, R', arising from thermal decomposition of DBPO. New macroradicals P" are formed by the addition of P" to the double bond of MAH. This latter reaction is favoured by the strong electron-attracting properties of the double bond of maleic or fumaric acid and their derivatives, as demonstrated by their high tendency to form altemate copolymers [18].

339 The fate of P" macroradicals has not yet been fully clarified.

Successive

addition of unsaturated molecules to P" macroradicals may be considered unlike on the basis of their steric hindrance and, consequently, no formation of a real graft copolymer should occur

Recombination reactions of P" with primary

radicals R" or disproportionation with P" or P" macroradicals have been proposed

~, H---C~ ~0 ~ ~<~--~'H--(' //0~ H---C~/~O H---C%~ 0 (a)

Hr---C%0

Y

'

"

(b)

+

)--~.'H~C~

~c}

340 as termination steps for the above grafting reaction onto iPP or LDPE, especially when it is carried out in the molten state [19-21]. A careful inspection of the IR spectra of MAH-modified EPM does not reveal absorption bands attributable to (a), (b), (c) groups which should be formed in disproportionation or recombination reactions.

In our opinion, P~ macroradicals preferentially

abstract hydrogen atoms from other carbon atoms of the same or of a different EPM chain or from solvent molecules thus generating monosubstituted succinic anhydride groups. 1.2 Effect of the reactivity of the double bond and of the nature of the initiator on the funcfionalizafion degree

In Fig.1 the variation of the fiuactionalization degree of EPM with the increase of DBPO concentration is reported for MAH and DBM.

The

occurrence of a graRing reaction is checked by the appearance in the IR spectrum of the polymeric reaction product of strong carbonyl stretching absorption at 1775

cm 1

and 1840 cm ~ in the case of MAH and at 1735 cm 1 in

the case of DBM. Quantitative evaluations of the inserted functional groups are performed according to IR methods for the ester groups and to potentiometric titration methods for the anhydride group [17]. As expected, both curves of Fig. 1 initially show a marked increase in the functionalization degree, which most likely corresponds to increasing numbers of free radicals per macromolecule. At higher initiator concentrations, the grafting degree eventually reaches a constant value. This trend should be ascribed to a combination of effects deriving from: (a) partial saturation of reactive sites on EPM; (b) decreasing efficiency of the initiator due to the recombination reactions among primary radicals R" and (c) increasing probability of secondary reactions of P" macroradicals.

Point (c) will be discussed later in detail.

The highest

functionalization degree attainable in the case of DBM is about 1/6 of that found in the case of MAH.

341

60<W "1oo)

t=o ~o

40 o/O

20-/ Ec~

9

()

[]

'

160

n

'

260

mmol of DBPO per 100 g of EPM

Fig. 1. Dependence of the grafting degree of DBM (El) and MAH (o) on the DBPO concentration. Reaction conditions" T=139~ xylene=100 mL; t =180 min; EPM=5.0 g; MAH=5.9 g; DMB =13.6 g The behaviour of DEF, not reported in Fig.l, is very similar to that of DBM. The higher reactivity of MAH may be related to an enhanced activation of the double bond towards the addition of P" macroradicals due to a stronger electron-attracting property of the anhydride group with respect to the ester group. The above results prompted us to test a different peroxide initiator, dicumylperoxide (DCPO), in the grafting reaction in order to enhance the graRed amount of DBM and DEF onto EPM. In fact, stability and polarizability of free radicals R ~ produced by peroxide decomposition, may affect the hydrogen abstraction reaction from the polyolefin. The choice of DCPO is suggested by the literature data on MAH - modified polypropylene by means of several radical initiators [16]. We have, therefore, investigated the effect of the nature of the initiator on the initial rate of the grafting reaction, Ri, in the well studied case of MAH grafting onto polyolefins. In Fig. 2 the amount of reacted MAH against the reaction time in xylene at 139~ reported.

by using DBPO and DCPO as initiators, is

342 -r

< 5 "0

~4 o

~ 3 0

%2

I4D---O

x

"" 1 0

E

200 300 400

reaction time, min

Fig. 2. Rate of MAH grafting onto EPM initiated by DCPO (N) and DBPO (e). Reaction conditions: T=139~ xylene=100 mL; EPM=5.0 g; MAH=5.9 g; DBPO and DCPO concentration=2xl0 2 molL 1 The R~ values are obtained from the slope of the respective curves in the initial period of the reaction (Ri = 2.7 x 10-5 mol L 1 s~ and 1.2 x 10-5 mol

L "1 s "1

for DBPO and DCPO, respectively). On the other hand, in organic solvents the decomposition rate constant Ka of DCPO at 139~

is much lower than that of

DBPO (ca.3.2 x 10~ s~ and ca. 2.6 x 10.2 s~, respectively [22]).

We have

therefore estimated the efficiency of the different primary radicals R" in the grafting reaction of EPM by MAH from the values of R~/R4 ratios, where R4 = Ka c is the rate of peroxide decomposition in our experimental conditions. The R~/Ra values are found to be 1.8 and 5.0 x 10-2 for DCPO and DBPO, respectively, thus indicating a remarkably higher efficiency of DCPO as initiator. In Fig. 3 the functionalization degree of EPM by DBM in the presence of DBPO and DCPO is compared at various peroxide concentrations. Different reaction times are used in the case of DCPO and DBPO in order to reach the final grafting degree in both cases.

The behaviour of the two

peroxides isqualitatively similar but, as expected, higher grafting degrees are found by using DCPO for fixed values of peroxide concentration.

|

30

343

4

nn

a"10 20" mn

~

-G------13"-"

10 1

E,--

I

I

0

I

I

100

200

mmol of peroxyde per 100 g of EPM Fig. 3. Influence of the initiator concentration (o DCPO, r-1 DBPO) on the grafting degree of DBM. Reaction conditions" T=139~ xylene=100 mL; t=180 min (DBPO) and 360 min (DCPO); DBM=13.6 g; EPM=5.0 g These results led us to compare the reactivity of different disubstituted alkenes towards EPM by using DCPO as initiator. In Fig. 4 the number of millimoles of inserted functional groups per 100 grams of EPM is reported for MAH, DBM, DEF, and itaconic anhydride (IAH). "ID

100

t._

I=

, p . / 1 if'"

t~O c

"o ~

60

/ I,

i O~ -

-

o

Oll~

20

E'5

EE

0

100

200

300

400

reaction time, min Fig. 4. Dependence of the grafting degree on the reaction time for different unsaturated molecules. Reaction conditions: T=139~ xylene=100 mlL; EPM=5.0 g; DCPO=0.50 g; MAH (ri)=5.9 g; DBM ( i ) and DEF (O)=13.6 g; IAH (0)=6.7g

344 As previously found when DBPO is used (see Fig. 1), in this case MAH once again shows higher reactivity than that of other investigated alkenes. The double bond configuration does not play an important role as shown by similar reactivities exhibited by DEF and DBM.

It is worth noting the higher

functionalization degree obtained in the case of IAH. In Fig. 5 the dependence of the functionalization degree on DCPO concentration is reported for IAH, DBM, and DEF.

40

j

t_

::}

"-"6 t~

u}~}) ~0

E

30

9 s

S

S

S

9 ~P

,,o"

S

S

S

~176

o~

E'~ EE

/

10

#

I

s#

~176 9

!

0

!

|

2;0

50 100 150 mmol of DCPO per 100 g of EPM

Fig. 5. Influence of the DCPO concentration on the graRmg degree for different unsaturated molecules. Reaction conditions T=139~ xylene=100 mL; t=360 min; EPM=5.0 g; IAH (O)=6.7 g; D B M ( i ) and DEF(FI)=13.6 g Once again, higher functionalization degrees are found for IAH with respect to those obtained for DEF and DBM.

This result is worth a more detailed

analysis as the double bond of IAH is expected to be less activated towards the addition to P~ macroradicals than the double bond of DEF or DBM. We stm,~est that the observed behaviour should be ascribed to a higher homopolymerization tendency of IAH (1,1-disubstituted alkene) [23] with respect to that of DEF and DBM (1,2-disubstituted alkenes).

345 Therefore, not only single IAH moieties but also short poly(itaconic anhydride) chains may be attached to EPM backbones as a result of further reactions between P.' macroradicals and IAH units and/or of coupling reactions between short growing poly(itaconic anhydride) radical chains and P- or P" macroradicals.

In fact, from the reaction mixture of EPM, IAH, DCPO in

xylene, significant amounts of IAH oligomers could be isolated and characterized as monoethylester derivatives by IR and 1H-NMRspectroscopy.

O--C I

CH 2 I

O--C~0

The IR spectra of the oligomers are characterized by the disappearance of the 1640-cm 1 absorption (=CH2 group) of the monomer and by the presence of strong carbonyl absorption at 1705 cm ~ and 1735 cm"~. The absence of the double bond is confirmed by their ~H-NMR spectra which do not show olefmic proton signals in the range of 5-76. No such oligomers are detected in the reaction mixture when DBM and DEF are used as unsaturated molecules.

Analogous results are also obtained by

carrying out attempted radical homopolymerization of IAH and DBM in xylene at 13 9~ using DCPO in the absence of the polymeric substrate.

1.3 Degradation of modified EPM Closely related to the grafting of unsaturated molecules onto the EPM backbone is a decrease of the reduced viscosity of the polyolef~c substrate ~l,od, observed in all cases, due to secondary reactions of P" and P-' macroradicals. This general behaviour may be qualitatively related to a decrease of the molecular weight during the grafting reactions and must be ascribed to chain scission of the polyolefin [12,24] according to the following reaction pattems"

346

,,,,,~CH--C H2-C H~C H2-C H2,,,,,,~ I CH 3 |

,,,,,,,~CH=CH2 + CH3~CI--I~CH2~CH2,~ or !

,,,,~C H 2 ~ C - - C H 2 ~ C H2,,,,,~ I CH 3 ,""~CH 2 ~ C ~--------CH 2 I CH 3

r

+ (~H2

Other secondary reactions of disproportionation or recombination may also occur.

These reactions do not change (disproportionation) or even increase (recombination) molecular weight of EPM. In our experimental conditions the chain scission appears to be preferred and this is in agreement with the monornolecular mechanism of the reaction while the other secondary reactions proceed through a birnolecular mechanism and could be more probable if the reaction is carried out in molten state. We have investigated the influence of the peroxide and of the nature of the solvent on the EPM degradation. Fig. 6 shows the variation of flood of modified EPM with the reaction time, using DCPO as source for free radicals and xylene or chlorobenzene as solvents.

In both cases the EPM degradation is

characterized by an initial fast rate which subsequently decreases. A comparison between the pattem of the two curves seems to indicate that a larger extent of degradation occurs in chlorobenzene.

347

O)

%%

-~ 2 -6

__

G

o

~

9

M

26o

D__

460

reaction time, min

Fig. 6. Dependence of reduced viscosity of DBM-modified EPM on the reaction time in chlorobenzene (El) and xylene(o). Reaction conditions: T=139~ (121) or 132~ (o); solvent=100 mL; EPM=5.0 g; DCPO=0.50 g; DBM=13.6 g This fact can be interpreted on the basis of the previous observation that, when xylene is used, reaction of primary radicals with the solvent molecules can lead to the formation of radical solvent species, S ~ and, consequently, to a lower number of P" macroradicals produced per time unit. It cannot be excluded, however, that P" macroradicals also give rise to chain scission reactions, possibly through a mechanism revolving hydrogen abstraction from the chain. An indirect support to this hypothesis is brought up from the observation that the EPM degradation promoted by free radicals m the absence of DBM is lower than that found in the presence of DBM.

The

dependence of rl,~a of modified EPM on the increasing amounts of DCPO and DBPO used is shown in Fig. 7. The increasing concentration of the initiators is accompanied by a decrease of rl,oa that is more remarkable for DCPO under the same reaction conditions.

348

3

1

'

2b0

mmol of peroxyde per 100 g of EPM

Fig. 7. Influence of the peroxide concentrations ( D DCPO, 9 DBPO) on the reduced viscosity of DBM-modified EPM. Reaction conditions: T=139~ xylene=100 mL; DBM=13.6 g; EPM=5.0 g The different behaviour of DCPO and DBPO has to be related to the higher efficiency of the former to generate P" macroradicals and, consequently, to favour chain scission reactions.

2 Reactive blending of polyamide 6 and functionalized EPM rubbers by melt mixing

Introduction In the following section we report our studies on PA6/EPM blends as obtained by melt mixing processing. EPM molecules bearing succinic anhydride groups (EPM-g-SA) are used as precursor of a "compatibilizer", which forms during the melt mixing. EPM and EPM-g-SA are melt mixed to PA6 to obtain binary PA6/EPM or PA6/EPM-g-SA and ternary PA6/EPM/EPM-g-SA blends. When EPM-g-SA is

349 employed, the formation of an (EPM-g-SA)-g-PA6 graft copolymer is assumed, since no solvent is able to separate blend components completely.

On the

contrary, complete separation is obtained as solvent extraction of PA6/EPM blends.

0 II --CH--C

\

CH2-C

0

/

+ H2 N,~,,,,,,,~C O N H,~,,,,,,,,,~

r

II

O 0 II -CH--C

\

/

CH2--iCI O

However, indirect evidence of formation of a graft copolymer could be provided on the basis of a model reaction between EPM-g-SA and aliphatic amines leading to amidic or imidic linkages according to the reaction temperature [25].

Only the impact properties of the binary PA6/EPM-g-SA initially

containing 20 or 30% of EPM-g-SA were found to be very satisfactory, whereas the temary blends showed lower impact performances.

2.1 Influence of processing on phase structure, mechanical and impact properties Our efforts have been devoted to study the influence of melt mixing procedure and of composition on the blend morphology and hence on its impact resistance. In fact, two mixing techniques have been used:

350 (a) one-step mixing, in which all the components PA6, EPM and EPM-g-SA are simultaneously introduced in the static melt mixer (Brabender-like apparatus): (b) two-step mixing, in which EPM and EPM-g-SA are premixed in the same apparatus before the final mixing with PA6. The two mixing procedures are conceptually different [26]. In the former, in fact, the mixing of the components and the grafting reaction start simultaneously and the two processes can strongly interfere along the whole process. In the latter the mixing of the rubbers EPM and EPM-g-SA is separately accomplished and the reactants are only successively brought together. In this way the gra~mg reaction can occur in a more regular and uniform fashion. Only in the case when the time of reaction is much longer than the mixing time are the two procedures equivalent in practice. Also binary PA6/EPM-g-SA blends are studied to provide a meaningful comparison with the morphology and properties of the ternary ones. The polyamide-6 (PA6) used in this work, has a number-average molecular weight (Mn) of 2.3x104. The ethylene-propylene random copolymer (EPM) has a weight-average molecular mass (Mw) of 1.80x105, an ethylene content (C2) of 60 mol % and a glass transition temperature (Tg) of-60~ The EPM-g-SA with a graft content of maleic anhydride of 2.7 in weight is prepared following the same procedure as previously described [25].

2.1.1 Blends preparation and samples conditioning prior testing

One-step-mixing. All the binary and temary blends are prepared in a Brabender-like apparatus by simultaneous melt mixing of all the components at a temperature of 260~ with a mixing time of 20 min and at a roller rotation speed of 32 r.p.m.

351

Two-step mixing. This procedure is used only for temary blends. First EPM and EPM-g-SA are premixed in the same Brabender at 130~ for 10 min at 32 r.p.m. Secondly the rubber mixture obtained is melt mixed with PA6 under the same operating conditions as listed in the above paragraph. The blend C* (80/15/5) obtained in this way is processed again in the Brab~ader-like apparatus in conditions (265~

and 64 r.p.m.) different from

those m~ationed in the above paragraph. This severe second treatment is made to check, at least in one case, the morphological stability of these systems. All the initial bl~ad compositions are reported in Table 1. The materials obtained by both mixing procedures are compression moulded at a temperature of 260~

In this way it is possible to obtain 1 mm to 3 mm

thick sheets, from which specimens for mechanical tensile measurements and specim~as to perform Izod impact tests are obtained respectively by means of a hollow punch and of a milling machine. iii ii ii

i

Bl~ad code PA6/EPM/EPM-~;-SA 100/0/0

PA6 (wt %) 100

iii

EPM (wt %) 0

EPM-g-SA (wt %) 0

Code A

8O/lO/lO

8o

lo

lO

B

80/15/5 80/20/0 80/0/20 70/0/30 80/10/10 80/15/5 80/18/2

80 80 80 70 80 80 80

15 20 0 0 10 15 18

5 0 20 30 10 5 2

C D E F B* C* G*

Table 1. Initial bl~ad composition (the asterisks mark the temary blends obtained by the two-step mixing procedure) The polyamide properties are very strongly dependent on their water content, since water acts as a plasticizer, lowering the glass transition temperature. [271.

Therefore all the specim~as are water conditioned before testing

352 The compression moulded specimens are microtomed and metallized by polaron sputtering. The microtomed surfaces are then observed by a scanning electron microscope at different magnifications.

2.1.2 Mechanical tensile properties and morphology Stress-strain curves for binary and temary blends are reported in Fig. 8, all referred to that of pure PA6, which shows the typical behaviour of a semicrystallme polymer, with a yield point, a cold-drawing region and fibre rupture (curve A). a

/

500

~

400

D

/

54)0

A

400 E

300

o

O)

300

200

200

100

100

E

31o

C

tt

B

015

110 E

.It,

3.0

Fig. 8. Stress-strain curves for (a) binary and (b) temary blends referred to PA6 homopolymer behaviour. Code and compositions as indicated in Tab. 1 The modulus E, the stress O'y and the elongation ~y a t yield, the strength

CYb

and the elongation eb at break and the blend code are all reported as a function of composition in Table 2. Curve D (Fig. 8a) relative to the PA6/EPM (80/20) binary blend, shows a modulus E of 4.5 Kg cm 2, lower than that of PA6 (6.0 Kg cm2). This decrease

353 in modulus is due to the rubber contribution, whose modulus is about three decades smaller than that of the PA6 matrix. The blends can be placed in order of decreasing modulus: the two-step mixing temary blends (Fig. 8b) containing 10% and 5% of EPM-g-SA (curves B* and C* respectively) both exhibit an E value of about 3.4 Kg cm2; the corresponding temary blends obtained by one-step mixing (curves B and C respectively) show a 2.9 Kg cm 2 value; the E value of the remaining temary mixture, containing only 2% of EPM-g-SA, is slightly lower (2.6 Kg cm2); and finally the binary blends P A 6 / E P M -g-SA (curves E and F), containing 20% and 30% of the functionalized rubber, show values of 2.5 and 2.1 Kg cm 2 respectively.

Curve code A B C B* C* D E F G*

Composition PA6/EPM/ EPM-~-SA 100/0/0 80/10/10 80/15/5 80/10/10 80/15/5 80/20/0 80/0/20 70/0/30 80/18/2

E x 10 3 ~y X 10 .2 (kg cm "2) (kg cm "2)

E;y

6 . 0 _ + 0 . 5 4.1_+0.2 0.3_+0.02 2.9_+0.5 2.0_+0.2 3.5_+0.4 3.3_+0.2 4.5_+0.2 2.5_+0.2 2.1_+0.1 2.6_+0.2 -

O b x 10 .2 (kg cm 2)

5.6_+0.4 2.5_-+0.2 2.9_+0.1 2.2_+0.2 2.7_+0.3 2.8_+0.1 1.9_+0.2 1.9_+0.3 2.3_-/-0.2

eb

2.50_+0.3 0.9_+0.2 0.9_+0.1 0.6_--~.1 1.4_+0.2 0.5_+0.1 0.3_+0.1 0.7_-20.1 0.7_+0.2

Table 2 Moduli, stress and elongation at yield and at breakage of binary and temary blends as a function of initial blend composition The blends containing various amount of EPM-g- SA and 20% of total robber all have very close E values (ranging from 2.5 to 3.4 Kg cm -2) which are lower than that of the PA6/EPM (80/20).

This finding can be tentatively

attributed to the effect of the graft copolymer EPM-g-SA-g-PA6 on the existence of interfacial zones between the PA6 matrix and the rubbery dispersed particles,

354 in which there is probably an increase in free volume.

This effect can also

explain the decrease in modulus of such alloys with respect to the PA6/EPM blend containing the same amount of total rubber (20 wt%) but no EPM-g-SA, and therefore no interracial zones. The slight differences between blends B, C, B*,C* and G* can be ascribed to different morphological features (subsequently discussed). The lowest E value (2.1 Kg cm 2) that of blend F, can be attributed to its higher rubber content (30%) with respect to that of the previously considered blends (20%). A parameter that seems to be more interesting as a source of structural information is the elongation at break ~b, as this expresses the capability of the material to be plastically deformed, which in tum depends on the overall structure and morphology of the system. In Figs. 8a and 8b, it is possible to note that for all blends there is more or less sharp reduction of eb with respect to the PA6 performance (eb = 2.5). PA6 is a semicrystalline polymer undergoing a classical spherulitic-fibrous morphological transformation bymeans of diffuse or localised cold drawing (according to temperature and rate of deformation used during the testing) and therefore behaves very plastically.

The addition of rubber and functionalized

rubber can strongly modify the system. Two main reasons can be invoked to explain the reduction in eb for the blends: (1) 'hindrance' of the free EPM rubber particles to cold drawing of the matrix, leading to an unstable flow which causes premature rupture of the specimen, (2) a 'network' effect all over the system induced by the presence of EPMg-PA6, which, acting as an interfacial agent between dispersed and continuous phases, renders the system more or less interconnected and therefore less capable of flowing.

These two causes can work either separately or simultaneously

depending on the initial blend composition as well as on the mixing procedure.

355 In the case of the PA6/EPM (80/20) blend only the first effect will be effective and the very low eb value (-~0.5) is due to the presence of very large particles with a poor adhesion to the matrix, as shown by SEM observations (see below). Also the lowering of the strength ~b with respect to PA6 can be due to the presence of such particles. In fact the extemal load will reduce around the particles very high stress concentrations with subsequent local yielding at an overall stress lower than in the case of PA6 homopolymer. This hypothesis is strongly confirmed by stress-whitening effect observed on the blend specimens, indicating a diffuse craze formation. However, since there is no adhesion of the rubber particles to the matrix, only craze initiation will occur whereas craze termination will not.

Therefore the crazes, developing perpendicular to the

elongation direction, will easily coalesce into macroscopic cracks, leading to premature failure of the specimen soon aRer the beginning of cold drawing. If one compares the one-step mixing temary blends with that of blend D, the following observations can be made : (a) eb is larger (0.9); (b) stress-strain curves show a smoother change in slope, indicating that at equal overall stresses the material yields more easily; (c) a more intense stress-whitening effect is observed, indicating more effective craze formation; (d) SEM micrographs show particles of smaller size, and furthermore some of the rubbery domains seem to show a certain adhesion to the matrix; however, their size distribution is still very large (see Figs. 9, 10, 15) As the amount of total rubber is the same as for blend D, the increase in ductility and adhesion can only be attributed to substitution of part of the EPM by EPM-g-SA in the blends.

356

Fig. 9. SEM micrographs of microtomed surface of temary blend B (80/10/10), (a) 640 x; (b) 5000x

Fig. 10. SEM micrographs of a microtomed surface of temary blend C (80/15/5): (a) 640 x; (b) 5000x ......................

Fig. 11. SEM micrographs of a microtomed surface of temary blend B* (80/10/10): (a) 640 x; (b) 5000x

357 In fact, assuming the formation of an EPM-g-SA-g-PA6 graft copolymer during mixing this will act as an emulsifier. In other words, it will be segregated at the interface between some EPM particles and the PA6 matrix, plasticizing such a region.

In this way the local stress at yield is lowered around such

particles and this leads to an increase in the efficiency of the termination of crazes, radiating from each particle, in front of the neighbourmg ones.

In

conclusion, the lowering of ab with respect to PA6 is due mainly to the 'hindrance' effect and the increase with regard to blend D to an augmented efficiency of multicraze phenomenon due to some partial (even though not homogeneous) action of the graft copolymer as an interfacial agent. Therefore no appreciable overall 'network' effect seems to be effective in this case. Considering now the two-step mixing blend B*, compared with the one-step blend B of the same composition, one can observe: (a) ~b is lower (0.6); (b) the shape of the stress-strain curve is very similar; (c) no stress-whitening effect is observed in this case. (d) SEM micrographs (Fig. 11) show a very homogeneous morphology with a very fine texture (particle sizes 0.1 ~tm). From the above considerations it appears that the EPM-g-SA-g-PA6 is very finely dispersed into the matrix, fully exploiting its role as an interracial agent. Therefore the very small particles adhere strongly to the matrix, making the system well interconnected. This will decrease the microscopic extensibility of the material. It is interesting at this point to understand why a change in the mixing procedure has led to such a different result. When all three components (EPM, EPM-g-SA and PA6) are simultaneously mixed, the graining reaction starts before the two rubbers can be mixed together. Therefore the reaction does not proceed homogeneously leading to a chaotic morphology in which particles free

358 or linked to the matrix by EPM-g-SA-g-PA6 will coexist. The result is the very large distribution size as observed in Fig. 9.

In contrast, when EPM and

EPM-g-SA have been premixed, the reaction with PA6 will occur regularly on the rubber surface, continuously renewed by the tearing due to shear forces. The results will be the completion of the reaction and a continuous size reduction of the particles during mixing. The effect of decreasing the EPM-g-SA content but still using the two-step mixing can lead to some different features.

In fact,

comparing blend C* with blend B*, one can note that: (a) eb is much larger (1.4); (b) the shape of the stress-strain curves are very similar; (c) almost no stress-withening e ~ : ~ is present in this case; (d) SEM micrographs (Fig. 12) show very fine morphology but with a particle size distribution (the mean particle size being larger than for B* blend).

Fig. 12. SEM micrographs of nficrotomed surface of temary blend C* (80/15/5): (a) 640 x;(b) 5000x The observation suggests that a good exploitation of EPM-g-SA-g-PA6 as an interfacial agent has been achieved, but, differently from the previous case, the "network" effect is less effective than for blend B* as the copolymer content

359 is lower.

Moreover, a higher ext~asibility is obtained without any craze

formation.

The only possible explanation for such features seems to be a

predominant shear yielding mechanism, favoured by a certain lower degree of interconnection and by a different particle and interparticle size distribution [28]. This topology allows for a larger cold drawing of the matrix up to rupture than in the case of blend B*, whose structure is much tighter. By further decreasing the copolymer content in blend G*, still obtained by a two-step mixing (2% of EPM-g-SA), the behaviour comes back closer to onestep mixing blends B and C. In fact eb is equal to 0.7 and a stress-withening effect is present. These two results indicate that the EPM particles re-acquire their individuality and hence a multicraze mechanism is acting again. Finally going to the binary blend E which contains a very high amount of EPM-g-SA and no EPM, eb reduces to 0.3 since the system becomes even more strongly interconnected than in the case of blend B*. Also the morphological analysis (Fig. 13) shows similar overall features to blend B*, indicating that the two blends behave very similarly, with some minor differences. The results seem to suggest that only part of EPM-g-SA reacts to form the graR copolymer with PA6 whereas the remainder plays the same role as the EPM particles in temary blends. Of course it is difficult quantitatively to assess how much EPM-g-SA has reacted and if the mechanisms are really identical, as one would infer from the corresponding morphologies, which seem to be qualitatively similar (see Figs. 11 and 13). Finally for binary sample F containing 30% of the functionalized EPM the observed modulus decrease and slight eb increase (0.7) with respect to blend E are both very likely due two to its higher rubber content, yielding a coarser gram with respect to the previous case (Fig. 14).

360

Hg. 13. SEM micrographs of microtomed surface of binary blend E (80/0/20)" (a) 640x; (b) 5000x

Hg. 14. SEM micrographs of microtomed surface of binary blend F (70/0/30): (a)640x; (b) 5000x

361

Fig. 15. SEM micrographs of microtomed surface of binary blend D (80/20/0): (a) 640 x; (b) 5000x 2.1.3 Impact properties and morphology The energy necessary to break semi-beam cantilever specimens is detected by a fracture pendulum at a given temperature. The test temperature is changed by means of a home-made liquid-nitrogen apparatus.

Therefore, curves of the

resilience as a function of the test temperature are obtained for all the blends.

D ~

G)

n,'

Rv

// ,9 m

==

==,

~ - ~

A

B

i! I I

Temperature Fig. 16. Typical resilience versus temperature curve for a semicrystallme polymer

362 Mainly two variables can be defined by such curves: (a) the temperature Ti of the beginning of the brittle-ductile transition (see Fig. 16), obtained from the intersection of the linear extrapolation of the resilience value of R at very low temperature (AB) and the straight line relative to the uptumed portion of the curve (CD); (b) the R value of the plateau at low temperatures below Ti. The resilience R obtained by Izod tests as a function of temperature is reported in Fig. 17, for PA6 homopolymer and all binary and temary blends.

70,

C*/i 50/

'E -

,/

/

, / ' -

G*

.."

i

I

!f

J /:B //; 9 I

30_

10-50

/," .,~ '/ /,"

,.,. /,,,/,, e

/'" /..,.-.-":---f..,Z--'" -30

-10

T(oc)

10

30

Fig. 17. Izod resilience R as a function of test temperature for PA6, binary and ternary blend, whose composition, Rv and T~ values are indicated in table: Code A B C B* C* D E F G*

Composition: PA6/EPM/EPM-~-SA 100/0/0 80/10/10 80/15/5 80/10/10 80/15/5 80/20/0 80/0/20 7O/0/3O 80/18/2

Rv (KJ m 2)

T~(~

2.0 6.5 4.0 6.4 50 2.5 4.5 6.5 6.4

0 10 16 -45 -35 10 -37 -47 9

363 The curve symbols, the composition, R value at - 45~

Rv, and the

beginning of the brittle-ductile transition Ti are also listed.

The PA6

homopolymer exhibits a very low Rv value (2 kJ m2) and a transition

temperature Ti of 0~

The one-step mixing blends B and C show a certain

improvement in Rv but a worsening with respect to Ti (10~

and 16~

respectively). This behaviour is in agreement with SEM micrographs (Fig. 9 and 10) of microtomed surfaces at low (640 x) and high magnification (5000x). There are indeed very large and irregularly shaped domains (10-50 ~tm) with a broad size distribution and a poor adhesion to the matrix. For the blends of the same composition obtained by two-step mixing, the impact properties change very strongly as evidenced by curves B* and C*. This is probably due to the anhydride groups of EPM-g-SA, which are very finely dispersed in the EPM rubber. Therefore, the grafting reaction with (EPM-g-SA)-g-PA6 formation can occur more regularly, giving rise to a very homogeneous texture. Such an effect on the morphology is demonstrated by SEM micrographs (Fig. 11 and 12). In fact the grain is very fine in both cases. Furthermore, that corresponding to the blend B*, having a relatively higher modified rubber content, is finer than that of the blend C*.

Such a finding has already been

discussed and confirmed by the mechanical tensile properties.

An analogous

improvement in the impact properties had previously been obtained [18], confirmed in this work, only using a much higher amount of EPM-g-SA as shown by curves E (blend with 20 wt% of such component) and F (with 30% of the same modified rubber). As it is possible to observe from Fig. 17, blends E and C* and F and B* show a very similar trend, respectively. Furthermore, it is to be underlined that F contains 30% of total rubber whereas B* only 20%. This can lead to the same conclusion previously reported in this paper, i.e. that for these binary blends only a certain amount of EPM-g-SA would react, whereas the remainder would act only as a rubber so~ phase. This seems to be confirmed also by SEM micrographs (Fig. 13 and 14) where it is possible to observe that

364 the grain of blend E is coarser than the corresponding temary C* blend and that of F is even coarser than that of B*. The particular efficiency of the two-step mixing procedure is also demonstrated by the comparison of blend D impact properties containing 20% of unmodified EPM with those relative to the G* blend having the same percentage of total rubber but 2% EPM-g-SA. As it is possible to see, with such a small amount of modified rubber the resilience-temperature curve shows a large improvement.

This effect can be well understood if one compares the

corresponding morphological features. In fact, as shown from SEM micrographs (Fig. 15 and 18) when modified rubber is not present, large particles with no adhesion to the PA6 matrix can be observed (Fig. 15). By adding only 2% of the functionalized rubber, a good homogeneity is already achieved (even though the texture is coarser than in the case of greater amounts of EPM-g-SA (Fig;. 11 and 12).

Hg. 18. SEM micrographs of microtomed surface of temary blend G* (80/18/2): (a)640x; (b) 5000x

365

2.1.4 Concluding remarks The results obtained ill the case of the PA6/EPM-g-SA binary blend indicate that considerable amounts (20-30%) of modified rubber must be used to obtain a toughelled polyamide at low temperatures.

Such large quantifies imply high

costs since the modified rubber is obtainable by a long and complex solution process involving large solvent quantities [25]. Therefore an interesting result has been achieved by working on ternary blends in which a great proportion of EPM-g-SA (50-75%) is replaced by pure EPM, which is a commercial low-cost rubber. The goal has been achieved by using a two-step procedure in which the mixing of the rubbers has been separated by the successive grafting reaction in bulk with PA6 [26].

This gave file possibility of realizing very fine and

homogeneous morphologies suitable to ~ahance strongly the resistance of these blends.

Both features improve more and more with increasing EPM-g-SA

content (cfr. Figs. 10, 11 and 17) but already with a 5% blend C* (80/15/5) the polyamide exhibits a behaviour comparable to that of the binary blend with a 20% content ble~ad (80/20/0). This achievements is important not only from a practical point of view but also from a scie~atific one. The two-step procedure in fact can be applied conceptually to all the cases when mixing proceeds in parallel with a fast chemical reaction.

Fig. 19. SEM micrographs of nficrotomed surface of temary blend C* (80/15/5) processed a second time in the Rheocord at 260~ and 64 r.p.m.: (a) 640 x; (b) 5000x

366 A further preliminary result of the graft copolymer EPM-g-PA6 on the morphological stabilization of these systems aRer successive processing has been shown in the case of blend C* (cf. Fig. 12 and 19). This is very important for complex materials which must be shaped successively (i.e. in rejection-moulded items).

2.2 Influence of grafting degree of rubber on phase structure, mechanical and impact properties

In all the studies reported so far, a ftmctionalized rubber containing 2.9 wt% of grafted anhydride was used throughout.

It is very important to know the

influence of the degree of grafting of the functionalized EPM-g-SA rubber on the mode and the state of dispersion of the minor component, the impact behaviour and the tensile mechanical response of some binary PA6/EPM-g-SA and the temary PA6/EPM/EPM-g-SA blends (prepared by the double-step mixing procedure). EPM-g-SA with graR contents of maleic anhydride of 0.6, 2.4 and 4.5 wt% were then prepared and tested.

Blend code PA6/EPM/ EPM-~;-SA 100/0/0 80/20/0

PA6 (wt %)

EPM (wt %)

EPM-g-SA (wt %)

Grafting degree DG (wt %)

100 80

20

-

-

8o/o/2o

8o

-

2o

o.6

80/0/20 80/0/20 80/15/5 80/15/5 80/15/5

80 80 80 80 80

15 15 15

20 20 5 5 5

2.4 4.5 0.6 2.4 4.5

Table 3. Initial blend composition

367

2.2.1 Blend preparation EPM-g-SA with graft contents of maleic anhydride of 0.6, 2.4 and 4.5 wt% were prepared following the procedure already described [17,25] Binary PA6/EPM and PA6/EPM-g-SA blends were prepared in a Brabender-like apparatus by simultaneous melt mixing of the components at a temperature of 260~

with a mixing time of 20 min and a roller speed of 32

r.p.m. For the preparation of temary PA6/EPM/EPM-g-SA blends, the EPM and EPM-g-SA were first premixed in the same Rheocord apparatus at 130~ for 10min and at 32 r.p.m. The rubber mixture was then melt mixed with PA6 under the same operating conditions as for the binary blends. The initial composition of all binary and temary blends investigated are reported in Table 3 [29]. The final blends as obtained from the mixer were compression moulded in a heated press at a temperature of 260~

to obtain sheets 1 mm and 3 mm thick.

The former were used to get dumbbell-shaped specimens for mechanical tensile tests. The latter were used to make parallelepiped-shaped samples to perform Charpy impact tests on notched specimens.

2.2.2 Mode and state of dispersion of rubbery components Microtomed surfaces of binary and temary blends were exposed for 30 mm to boiling xylene vapour and subsequently examined by a scanning electron microscope, after coating with gold-palladium alloy. It was observed that the xylene selectively dissolved the rubbery phase, leaving the PA6 undissolved. SEM micrographs of microtomed surfaces, taken after etching, for the binary and temary blends are shown in Fig. 20, 21 and 22. In the case of PA6/EPM (80/20) binary blend (Fig. 20), the rubber segregates from the PA6 matrix in spherically shaped domains regularly distributed throughout the whole sample.

368

Fig.20. SEM micrograph of microtomed surface of binary PA6/EPM blend after etching (640 x) The dimensions of such domains are very large, the diameter ranging from about 10 ~m to about 30~tm. The walls of the cavities left seem to be very smooth, indicating no adhesion between matrix and rubber. The degree of grafting of EPM-g-SA plays an important role in determining the mode and state of dispersion of the rubbery components in both binary PA6/EPM-g-SA and temary PA6/EPM/EPM-g-SA blends.

Fig. 21. SEM micrographs of microtomed surfaces of binary blends PA6/EPM-g-SA at increasing degree of grafting (DG)after etching: a) DG=0.6% (2500x); b) DG=2.4% (2500x); c) DG=4.5% (2500x)

369 From the micrographs of Fig. 21 and 22 it emerges that: (i) The dimensions of the etched domains and the relative distribution are much lower than in the case of blends containing unmodified EPM (compare Fig. 20 with Fig. 21 and 22).

Fig. 22. SEM micrographs of microtomed surfaces of ternary blends PA6/EPM/EPM-g-SA at increasing degree of grafting (DG) after etching: a) DG=0.6% (2500x); b) DG=2.4% (2500x); c) DG=4.5% (2500x) (ii) In both PA6/EPM-g-SA systems, the dimensions of the etched domains drastically decrease with increase of the degree of grafting (DG) of the EPM-gSA rubber.

For PA6/EPM-g-SA with DG =0.6% the diameter of the holes

ranges from l~tm to 5 ~tm while for DG=4.5% the average diameter is less than 1]am (compare Fig. 2 l a and 21 c ). k can be observed that for the same value of DG the average dimensions of the etched domains seem to be larger in the case of 80/15/5 ternary blends (compare Fig. 21 and Fig. 22). (iii) It can be noted that the amount of material etched, in the case of the PA6/EPM-g-SA binary blend containing rubber with DG=4.5 % seems to be smaller than in the other cases. Such an observation may be accounted for by the fact that in such a system a larger concentration of unextractable (EPM-g-SA)-gPA6 graft copolymer is formed. In fact the PA6 branches of the (EPM-g-SA)-gPA6 graft copolymer are firmly into the PA6 matrix. dispersed

phase

is

less

and

less

extractable

Therefore the rubber

the

higher

the

DG.

370

2.2.3 Mechanical tensile properties Typical nominal stress vs. strain curves for PA6 homopolymer and for all the binary blends investigated are shown in Fig.23. The corresponding moduli E and the ultimate properties such as the stress Orb and the elongation at break eb are summarized in Table 4. As it is possible to see, the performance of PA6 (curve A) is strongly modified by rubber addition.

N

E o 6.OA

'o "-

X

4.0-

C

u~

E

2.0-

I

0

I

1.0

I

I

Strain

2.0

I

I

3.0

Fig. 23. Stress-strain curves for P A6 homopolymer and for binary blends at increasing grafting degree (DG) (A) PA6; (B) DG=0%;(C) DG-0.6%; (D) DG=2.4%; (E) DG=4.5% Composition PA6/EPM/ EPM-~-SA 100/0/0 80/20/0 80/0/20 80/0/20 80/0/20

DG

E x 10"3 (kg cm -2)

OrbX 10 "2 (kg cm 2)

0.6 2.4 4.5

6.9_+0.9 3.7_+0.5 4.0_+0.4 3.3_+0.2 3.6_+0.3

5 . 2 _ + 0 . 4 2.2_+0.4 2 . 2 _ + 0 . 2 0.3_+0.1 2.5_+0.1 0.75_+0.2 2 . 4 _ + 0 . 2 1.2_+0.3 3 . 2 _ + 0 . 3 1.6_+0.4

eb

Table 4. Mechanical tensile properties of PA6 homopolymer and binary blends

371 The overall effect is a decrease in the values of all the above mentioned parameters (see Table 4). The lowering of the modulus, however, seems to be scarcely influenced by the nature of the rubber (EPM or EPM-g-SA), whereas more sensitive is the elongation at break. In fact,

eb

changes drastically depending on the type of

rubber and the degree of grafting (DG) of the EPM-g-SA used. The very low eb value observed for the PA6/EPM blend (curve B) can be attributed essentially to the presence of very large EPM particles with a poor adhesion to the matrix, as shown by SEM micrographs (see Fig. 20). These domains act as gross material defects, causing premature rupture of the specimen soon after the beginning of yield. Only a few whitened bands of microvoids are visible along the specimens, indicating an ineffective mechanism of craze termination since there is no adhesion of the robber particles to the matrix. As EPM is substituted with EPM-g-SA (curves C, D, and E), the ability of the material to be plastically deformed is raised. In this case an intense stresswhitening phenomenon is observed on the blend specimens due to diffuse craze formation.

This effect increases more and more with increasing degree of

grafting of the EPM-g-SA. This finding is probably related to the presence of an (EPM-g-SA)-g-PA6 graft copolymer formed during mixing. This will act as an interfacial agent between the dispersed (unreacted EPM-g-SA) and continuous phases. As a matter of fact, a greatest homogeneity with a finer dispersion of the rubber component (see SEM micrographs of Fig. 2 l) is achieved with respect to the PA6/EPM blend. This overall moqghology facilitates the formation of crazes as well as their termination, so that the material is able to sustain larger plastic deformation prior to fracture. Passing to the temary PA6/EPM/EPM-g-SA blends, the relative tensile properties are reported in Table 5.

All the blends show, within the limit of

experimental errors, the same eb values with varying degree of grafting, in contrast with their different morphologies (Fig. 22).

This disagreement is

372 probably due to the presence of some defects created in the pressure moulding of the material during specime~a preparation. Such features overcome the intrinsic morphological defects of the material constituted by the rubbery dispersed particles and cause a premature rupture of the specimens, therefore masking the real structure of the ternary blends.

2.2.4 Impact properties The resilience R, obtained by a Charpy test, as a function of temperature is reported for binary and ternary bl~ads in Fig. 24 and 25 respectively. The PA6 homopolymer is take~a as the refercaace material in order to evaluate the improvement in impact properties of the bletads.

5o4 40 "E 3 0 e: 2 0 10 ~_

_

_

O F

"

T

-50

-30

_..._...----'/ -

,

-10

T(~

..

,

10

!

30

Fig. 24. Charpy resilience (R) as a function of test temperature for PA6 homopolymer and for binary blends at increasing DG: (e) PA6; (+) DG=0%; (m) DG=0.6%; (O)DG=2.4%; (D) DG----4.5% As it is possible to see, the pure polyamide exhibits very brittle behaviour with very low R values, which remaill unchanged over the whole investigated temperature range. A similar trend is observed for the PA6/EPM blend

373 (Fig. 24). Such behaviour is in agreement with the morphology of this blend in which large domains with a broad size distribution and no adhesion to the matrix are present.

40 r

'E 30 v

a:: 20 10 I

-50

I

-30

I

-10

T(~

I

10

I

30

Fig.25. Charpy resilience (R) as a function of test temperature for PA6 homopolymer and for temary blends at increasing DG: (O) PA6; ( ' ) DG=0.6%; (rn)DG=2.4%; (e) DG=4.5%

Composition PA6/EPM/EPM-g- S A

DG

E x 103 (Kg cm 2)

(~b X 10 .2 (K~; crn 2 )

~b

80/15/5 80/15/5 80/15/5

0.6 2.4 4.5

3.6_+0.1 3.4_+0.1 4.0_+0.5

2.5_+0.1 2.5_+0.1 2.6_+0.1

0.7_+0.15 0.8_+0.1 0.75_+0.2

Table 5. Mechanical tensile properties of temary blends

374 For blends of the same composition but containing EPM-g-SA as rubbery phase and for ternary blends as well, a large enhancement of impact properties is observed with respect to pure PA6. From the trend of the curves, it emerges that R increases with increasing degree of grafting (DG) of the EPM-g-SA and that the binary blends show higher R values compared to those of the corresponding ternary blends.

Furthermore, all these blends show a marked variation in R

values over the temperature range in which the behaviour of the material changed from a brittle to a ductile mode of failure.

The location of this transition

temperature is a function of DG and blend composition. A shift towards lower temperature is observed with increasing DG value of EPM-g-SA and for the same DG passing from ternary to binary blends.

All the above-mentioned

features suggest that for these blends a good exploitation of (EPM-g-SA)-g-PA6 graft copolymer as an interracial agent has been obtained. The extent of this effect for the blends, as demonstrated by the morphological analysis, is larger the higher DG.

In fact, for a given blend composition, better homogeneity with

smaller domain dimensions of the rubbery phase is achieved in the direction of an increase in the DG values. On the other hand, the finding that the investigated ternary blends exhibit lower impact properties than the corresponding binary ones can be related to a lower amount of (EPM-g-SA)-g-PA6 as interfacial agent, essentially due to the smaller initial percentage of EPM-g-SA. In fact, we have shown how by increasing the EPM-g-SA content in the rubbery phase it was possible to get ternary blends with better impact properties than those of the corresponding binary blends. This finding suggests that, for the latter only, part of the EPM-g-SA can react to form graft copolymer with PA6, whereas the remainder plays a role similar to pure EPM in temary blends.

2.2.5 Concluding remarks The possibility of improving the impact properties of polyamides by the addition

of

ftmctionalized

EPM

has been confirmed for both binary

375 PA6/EPM-g-SA and temary PA6/EPM/EPM-g-SA initial blends. The formation of a graft (EPM-g-SA)-g-PA6 copolymer and its effectiveness for toughening mechanism, has received support over a range of degrees of graRing from 0.6 up to 4.5 wt%. In the investigated range the morphology and impact properties are better the higher the DG for both the binary and temary blends analysed. From the data of the present section (restricted to one binary composition (80/20) and one temary (80/15/5) one it could be inferred that the binary and temary blends behave quite similarly with respect to impact mechanisms.

In

other words, the functionalized rubber added to PA6 (binary system) is only partially converted, during melt mixing, to form on the interface the (EPM-g-SA)-g-PA6 copolymer responsible for PA6 toughening. The remainder, contained in the middle of the rubbery particles, behaves like pure EPM.

3 Reactive blending of polyamide 6 and functionalized EP rubbers concurrently with matt~ polycondensation In the next sections, few experiences carried out in the field of blending during polymerization of matrix polymer will be described. The attention is still concentrated on the system PA6/EPM. If such approach proves to be successful on a system like this, characterized by a complete lack of interracial adhesion of the final polymers, and a complete immiscibility of the intermediates (caprolactam and EPM polymers), it is conceivable that the results can find even more successful application in those systems where the starting chemical physical situation is more favourable.

3.1 Influence of polymerization conditions and blend composition. The rubber component used is the already described amorphous ethylene propylene copolymer (EPM) having 60 percent by mole of C2. A modified EPM

376 was prepared, as described previously [17,25] bearing, along its backbone, 3 percent by weight of succinic acid groups (EPM-g-SA). The blends were prepared according to two slightly different procedures and coded as "S" and "B" types [30]. "S" blends contain only 10 percent whereas "B" ones contain 20 percent by weight of initial total rubber (EPM-g-SA) [7] The subscript of the complete codes reported in all the tables and figures stands for EPM-g-SA weight percentage.

3.1.1 Polymerization conditions "S" codes blends

In a glass vial, equipped with a side ann, EPM and EPM-g-SA are dissolved in anhydrous xylene.

After dissolution, caprolactam (CL) and

aminocaproic acid (ACA) (2% of total lactam weight) are added without occurrence of a clear phase separation. By raising the temperature to 200 ~ C, the xylene is slowly distilled through the side arm in a graduated recovery flask. The last traces of xylene are removed at 260~ and 600 torr. More ACA (2% by weight) is then added and the polymerization is carried out for 4 hrs at 260 to 270~ under a slow stream of nitrogen and vigorous stirring. The crude reaction mixture is finely ground, extracted in a Soxhlet apparatus with boiling methanol and dried under high vacuum at 130~ for 24 hrs. The polyamides, separated by the rubbery phase by solvent extraction, are characterized by viscosity measurements. "B" coded blends

The synthetic procedure for a "B" coded blend is substantially similar to that previously described for the S blend, but no xylene is used to "homogenize" the "components", and therefore the distillation step was unnecessary and the ACA is added all at once together with the CL. The yield of MeOH extractable products and Mn are essentially the same as above.

377 Blend code P So $2 $10 B0 B2 B5

CL

ACA

EPM

95.5 85.5 85.5 85.5 76.0 76.0 76.0

5.0 2.0 (+2.5)* 2.0 (+2.5)* 2.0 (+2.5)* 4.0 4.0 4.0

0 10 8 0 20 18 15

EPM-g-SA Preparative method 0 bulk 0 solution 2 solution 10 solution 0 bulk 2 bulk 5 bulk

Table 6. Preparation of rubber-toughened Polyamide 6 by bulk and solution

methods Figures in bracket refers to the additional amount of ACA added alter xylene distillation (see text) The hydrolytic synthesis of polyamide (PA6) has been reviewed by Sebenda [31], and may be summarized by the following equilibria:

N--H

+

H20

~

"

O II

HO--C

CL

N,H2

/

ACA ring opening

"w"NH2 +

O=C

N~H CH2/5

--~

I

~NHC

O II

addition ~w'NH2 + ~,,w'COOH

~

condensation

"-- ,,,w~NHCO,,,,,,,~

~

H2

378 Instead of using water, we added ACA directly as initiatior. It is likely that in the early stage of polymerization EPM-g-SA may be at least partially involved in the following side equilibrium:

H--COOH

+ H2N-~CH2s~--COOH

=

H2--COOH EPM-g-SA

0 H --C I-i~C--N--~C H2s~-COOH CHz-COOH

ACA

I

As matter of fact, IR analysis of the rubber component recovered from the reaction mixture in the early stage of the polymerization shows the presence of both amide and acid-C=0 stretching absorption at 1705, 1645 and 1535 cm -~ in agreement with this structure I. Subsequently intermediate I will be probably involved in a polycondensation equilibrium with PA6 growing chains giving rise to the formation of a (EPM-g-SA)-g-PA6 gratt copolymer. Appropriate

modifications

of

the

standard

conditions

used

to

homopolymerize PA6 were tested in the preparation of blends of PA6 and EPM and/or EPM-g-SA in order to ensure both a high polymerization degree and good dispersion of the rubber component. In Table 6 are summarized blends codes, compositions and preparative methods of investigated binary and ternary blends. For sake comparison we have also prepared a PA6 homopolymer sample under the same experimental conditions. Because EPM and EPM-g-SA are both immiscible with CL, in the earlier experiments concerning the preparation of blends with 10 percent by weight of total rubber, the "homogenisation" of the reaction mixture was accomplished by

379 first dissolving EPM and/or EPM-g-SA in xylene at 139~ and by subsequently adding CL and ACA ("S" coded series). The temperature was left to rise to 260 to 270~ while xylene distilled off. At this stage, when CL polymerization is not yet started, we have found that the presence of EPM-g-SA has a significant effect on the homogenisation degree of the mixture. In fact, in the binary CL/EPM-g-SA mixture no phase separation was apparent; in the temary CL/EPM/EPM-g-SA, a fine and stable suspension of robber in CL was observed, while in the binary CL/EPM mixture a coarse phase separation was found to occur. These fmdmgs are likely to be related to some emulsifying effect of compound I (see Eq. 4). Small amounts of water, deriving from stage 1, were found to distil together with xylene.

This and the consumption of ACA via equilibrium 4 causes a

lowering of initiatior concentration which was balanced by adding more ACA. The CL polymerization was subsequently carried out at 260 to 270~ for 4 hrs under N2 atmosphere and vigorous mechanical stirring. Absence of oxygen was necessary in order to avoid yellowing of the polymerization mixture. Visual observation of the molten mixture at the end of the polymerization indicated that the dispersion of the rubber domains into the blend increases gradually with the amount of the added modified EPM. As results on analogous blends prepared by melt mixing showed that higher contents of total rubber were necessary to significantly improve the impact properties, we investigated the preparation of a new series of binary and temary blends containing 20 percent rubber by a hydrolytic process. Furthermore, in order to bring experimental conditions closer to those used in the industrial process of the hydrolytic polymerization of CL, we have also investigated the preparation of the above blends by a bulk process ("B" coded blends) Accordingly, all the starting materials were mixed together, in absence of xylene, and CL polymerization was performed as described for the "S" series (no further addition of ACA is needed during this process).

380 In the case of bulk blend preparation, the efficiency of stirring is more important than in the "solution" process in order to achieve an acceptable dispersion of components. k is interesting to note that under our experimental conditions and using a mechanical stirrer in a cylindrical glass vial, blends with a total rubber content of 20 percent in which the ratio EPM-g-SA/total rubber exceeds 0.25 were too viscous to be efficiently stirred [30]. Therefore only ternary blends containing 2 and 5 percent modified rubber could be prepared: the yields of the crude polymerization products are almost quantitative.

A methanol soluble fraction

was found, as generally occurs in PA6 samples prepared by ring opening of CL [32]. This fraction is mainly formed by equilibrium monomer and lower cyclic oligomers. In Table 7 some characterization data of the blends are reported. Formic acid extractions of the methanol insoluble products may be used to obtain pure PA6 from the blends as PA6 is soluble while EPM separates as a supematant solid phase. This technique was found suitable only for So, $2, Bo and B2 blends, where little or no (EPM-g-SA)-g-PA6 grait copolymers is present. In B5 the larger amount of (EPM-g-SA)-g-PA6 graft copolymer, acting as emulsifier, causes such a delay of the phase separation phenomenon that clear solutions cannot be obtained even after months.

On the other hand, a clear

solution is obtained in the case of $10 blend indicating that all the PA6/EPM graft copolymer is solubilized in the formic acid or is at least very finely dispersed. These results, which may be regarded as Molau tests [33] dearly show the emulsifying effect of the (EPM-g-SA)-g-PA6 copolymer. The molecular mass of the polyamide samples recovered from So, $2, Bo and B2 blends are satisfactory and practically identical to that of a pure PA6 sample prepared in a testexperiment (see Table 6). The material obtained by synthesis and purification through extraction is compression molded in a common heated press at a temperature of 260~

into

sheets of two different thickness for mechanical tensile tests and for Izod

381 Impact tests. Both types of samples are conditioned in water at 90~ in order to obtain in any case the same amount of absorbed water (about 3 percent by weight) [25].

3.1.3 Mechanical tensile tests

The experimental results for PA6 homopolymer and for all the binary and ternary blends investigated are reported in Fig. 26. The behaviour goes from a very ductile fracture of pure PA6 to a less ductile one depending on the blend type and composition. The Young modulus (E) of blends is lower than that of plain PA6 and in all cases decreases with an increase of the overall rubber content.

This behaviour is probably due to the decrease of the overall

cristallinity content of the material.

600J. G (kg/cm 2)

P

.~e,,/ S~ S2

300 S~o

100 0

Bo

E (%)

B5

40

120

200

300

Fig. 26. Stress-strain curves for PA6, binary and temary blends The ultimate properties and particularly the elongation at break (lSb) seem to be more interesting as source of structural information. Such a parameter shows very high values (300 percent of total deformation) for PA6 (curve P), whose specimens appear to be still transparent after stret&hing. The eb of So and $2

382 samples decrease to about 180 percent and their specimens show a stress whitening phenomenon probably due to diffuse craze formation around the rubber particles during elongation. Such an effect is higher in the case of $10 blend, where the (EPM-g-SA)-g-PA6 graft copolymer cannot act as interracial agent because no EPM rubber is present.

Blend code P So $2 $10 B0 B2 B5

MeOH soluble, percent 20.3 15.2 18.2 15.2 22.0 24.7 21.0

Composition (PA6/EPM/EPM-g-SA) Initial Final 100/0/0 100/0/0 90/10/0 88/12/0 90/8/2 87.5/10/2.5 90/0/10 88/0/12 80/20/0 74.4/25.6/0 80/18/2 74.4/23/2.6 80/15/5 74.7/19/6.3

M~*

22.100 18.400 21.600 21.000 19.500

II

Table 7. Characterization of the prepared binary and temary blends (*Number average molecular weight of PA6 recovered from the blend, determined according to ref. 36) The elongation at break of the Sl0 blend is lowered to about 65 percent This low eb value is due to the influence of the high (EPM-g-SA)-g-PA6 content, which renders the system more strongly interconnected and therefore less able to undergo the coldrawing process (see Fig. 29c). For the B blends, containing 20 percent total rubber, the elongation at break values are 30 percent for Bs, 50 percent for B0, and 110 percent for B2. Moreover the B2 specimens exhibit a slight stress-whitening effect due to a multiple crazing mechanism acting in the material during its elongation.

The

craze formation produces an apparent volume increase of the blend and hence generates eb values higher than those corresponding to B0 and B5 samples, for which no stress whitening effect is observed.

383 The lower

E:b

values of B0 and B5 depend on different reasons.

For B0

blends the presence of very large EPM domains as observed in SEM micrographs (see Fig. 30), may hamper the cold-drawing and cause the premature fracture of the specimen. The coarse dispersion of rubber particles can be related to the weak stirring power of the system during and at the end of polymerization.

For B5 blend the situation may be even worse because, as

previously mentioned, the melt viscosity of blends increases with increasing EPM-g-SA content and therefore in this case large rubbery domains are also present (Fig. 32).

Moreover at least part of the functionalized rubber is

converted to (EPM-g-SA)-g-PA6 which acting as an inteffacial agent between matrix and dispersed phase, renders the system locally more interconnected and hence less capable of cold-flow. Numerical values of E, eb and of the ultimate tensile strength (Oh) are reported in Table 8 as a function of the total rubber amount in the blends. From these data it emerges that blends have values of Oh lower than that of PA6. This decrease seems to be larger in the case of blends with higher EPM-g-SA content (S10 and Bs).

3.1.4 Morphology of microtomed blend surfaces Samples of binary and temary blends were faced in an ultramicrotome at room temperature; subsequently the smooth surfaces were exposed to boiling xylene vapours or immersed in boiling xylene before being prepared for SEM examination. The rubber phase is selectively dissolved by xylene whereas the PA6 matrix remained unaffected. Fig. 27a shows SEM micrograph of the untreated smooth surface of the So sample. As can be seen, the EPM segregates from the PA6 matrix in spherically shaped domains regularly distributed throughout the whole surface. Exposure to boiling xylene vapours produces holes corresponding to the dissolution of the

384 EPM inclusions and evidences the wide size distribution of such domains, ranging from 5 ~tm to about 20 ~m (Fig. 27b).

Fig. 27. SEM micrographs of smoothed surfaces of S0 sample: a) before etching (640x); b) after etching (640x) e Code

P So $10 $2 B2 B5 B0

E x 10-3

~b

~b

(Kg/cm 2)

(Kg/cm 2)

(percent)

610+20 360+20 270+10 350+20 220+20 180+30 190+10

310+10 180+_20 67+15 200+20 110+30 30_+10 50+_5

3.9_+0.1 3.7_+0.1 3.2_+0.1 3.2_+0.1 2.4_+0.2 2.6_+0.1 2.7_+0.1

Table 8. Tensile moduli, strength and elongation at break as a function of composition. The complete dissolution of EPM seems to suggest that there is no adhesion at the interface with the PA6 matrix. This conclusion is strongly supported by the fractography analysis as reported elsewhere in this section. In the case of $2

385 blend, protrusions emerging from the dispersed domains are deafly visible on the smooth surface (see Fig. 28a). The etching of the surface with boiling xylene vapours is able to dissolve only part of the materials contained in the dispersed domains, leaving the protruding structures practically unchanged and isolated from the rest of the material (see. Fig. 28b and 28c). Moreover, such structures are connected to the underlying matrix.

Such observations strongly indicate that in $2 blends the

dispersed domains are of a multiphase type, i.e., they contain a rubbery phase that is permeated by irregular polyamide structures protruding from the matrix and strongly bonded to it.

Fig. 28. SEM micrographs of smoothed surfaces of $2 sample: a) before etching (640x); b) after etching (640x); c) after etching (2500x)

386 The presence of a certain amount of (EPM-g-SA)-g-PA6 graft copolymer may contribute, together with the shear stresses induced by the stirring system, to the formation of such a complicated overall morphology. Smooth surfaces of the $10 binary blend show the presence of dispersed domains randomly distributed throughout the whole sample. (see Fig. 29a). It is to be emphasized that the relative area occupied by such domains seems to be lower than the rel,fi, . . . . . .

~..... :he initially added EPM-g-SA.

As show in _~ig. _~ : '~ ~~,o,,-g :'~:~ xylene vapours are able to dissolve completely only the central region of such domains. Fibrils connecting the PA6 matrix with the undissolved adjacent regions of the dispersed domains are dearly evidenced by the etching technique (see Fig. 29c).

c Fig. 29. SEM micrographs of smoothed surfaces Sl0 sample: a) before etching (640x); b) after etching (640x); c) after etching (2500x)

387 The above observations may be accounted for by assuming that during the process of blend preparation only part of the EPM-g-SA modified rubber is involved in the formation of (EPM-g-SA)-g-PA6 graft copolymer and that such copolymer, because of a combined effect of the mterfacial forces and of the mechanical stirring, at the end of the polymerization segregates, partially at least, at the interface between the PA6 and EPM-g-SA rubber, thus acting as a real interracial agent. Consequently the dispersed domains observed in the $10 blend are multicostituent in nature and daaracterized also by the fact that the concentration of the (EPM-g-SA)-g-PA6 constituent is not uniform throughout the domains, but is larger at the boundary. In the case of blends with overall rubber content of 20 percent after the exposure of the smooth surfaces to boiling xylene vapours, the rubber phase is not completely removed, and it was necessary to immerse the samples directly in boiling xylene.

Fig. 30. SEM micrograph of smoothed surface of B0 sample before etching (640x) SEM studies

performedon B2 and B5 specimens treated with boiling xylene

reveal well defined but irregular rubbery domains dispersed in the PA6 matrix with a wide size distribution (see Figs. 31 and 32 ). It is to be pointed out that for the ternary blends no protruding structures are observed.

388

Fig. 31. SEM micrographs of smoothed surfaces of B2 sample: a) before etching (320x); b) after etching (320x) The boiling xylene in the case of B blends seems to be able to completely dissolve the rubbery domains, as shown by the smooth walls of the remaining cavities, contrary to what was observed for S type blends containing EPM-g-SA.

Fig. 32. SEM micrographs of smoothed surfaces of B5 sample: a) before etching (320x); b) after etching (320x)

389 These morphological characteristics may suggest that the (EPM-g-SA)-gPA6 formed during the blend preparation process in bulk is quantitatively insufficient to act efficiently as an interracial agent and therefore the dispersed phase exhibits relatively large particles with poor adhesion to the matrix. The possibility that the (EPM-g-SA)-g-PA6 copolymer is finely dispersed in the PA6 matrix with no connection to the rubbery domains must be also taken into consideration.

This phenomenon of "dissolution" could produce drastic

modifications in some intrinsic matrix properties such as PA6 spherulite size and thus in the crack propagation mechanism.

3.1.5 Fractographic analysis With the view of elucidating the influence of composition and test temperature on the mode and mechanism of fracture of "S" and "B" blends, a morphological analysis of fracture surfaces by means of SEM has been performed. The fracture surfaces of PA6 homopolymer broken at -20~

(Fig. 33a),

show no induction zone but only a fast crack propagation zone throughout the specimen. At-2.5~

a limited induction zone begins to appear (see Fig. 33b)

indicating that a transition from brittle to ductile behaviour is occurring.

Fig. 33. SEM micrographs of fracture surfaces of PA6 sample at different test temperature: a)-20~ (40x); b)-2.5~ (40x)

390 The So blend at-20~ (Fig. 34a) shows a fracture mechanism that is very similar to that of PA6 with a fast crack propagation zone extended to the whole sample but with a rougher surface indicating a slightly higher energy dissipation than in the case of PA6. Spherically shaped domains of EPM copolymer with a diameter ranging from 5 ~tm to 20 ~tm randomly distributed throughout the whole surface are observed (see Fig. 34b).

Fig. 34. SEM micrographs of S0 fracture surfaces at different test temperature: a)-20~ (40x; b)-20~ (1250x); c) 11~ (40x); d) 11~ (1250x) No adhesion seems to exist between the dispersed particles and the PA6 matrix as evidenced by the very smooth walls of the cavities.

The fracture

process and, in turn, the morphology of the fractured surfaces of such a binary blend are drastically changed by increasing the test temperature.

391 SEM micrographs of broken surfaces of samples tested at 1 I~

show a

distinct induction zone starting from the notch and covering a discrete area where the sample undergoes a plastic shear yielding mechanism (Figs. 34c and 34d). The remainder of the sample exhibits - as in the previous case - a sudden crack propagation. The area of the induction region increases with increasing test temperature. At -20~

the $2 ternary blend also fractures in a brittle manner.

The

fracture surface of such a mixture shows a macroscopic roughness (see Fig. 35a) From a careful inspection of high-magnification SEM micrographs, there seems to be evidence of rubbery domains very well embedded into the PA6 matrix as stmc~gested by the presence of regions with a very smooth surface structure (Fig. 35b). With increasing test temperature a distinct reduction region of plastic deformation by a shear yielding mechanism appear (Fig. 35c). In this region irregularly shaped rubbery domains connected to the PA6 matrix by means of fibrils are observed (Fig. 35d). The above findings are further evidence that the formed (EPM-g-SA)-g-PA6 copolymer may act as an "interracial agent" promoting the adhesion between the PA6 matrix and the rubbery dispersed phase.

l~ig. 35. SEM rmcrographs of $2 fracture surfaces at different test temperature: a)-20~ (40x); b)-20~ (1250x); c)-10~ induction region (640x); d) +5~ (2500x)

392 Fig. 36 shows SEM micrographs of fracture surfaces of Sl0 binary blend. At -20~ the macromorphology of the broken surface is typical of a fast fracture exhibiting a rough stepped topography (Fig. 36a). It is important to note that, because of the high rate of crack propagation, no distinction between the phase can be observed at this temperature.. These morphological result are consistent with the glassy nature both of the matrix and copolymer at the temperature and rate of fracture.

Hg. 36. SEM micrographs of Sl0 fracture surfaces at different test temperatures: a)-20~ (40x); b)+9.5~ (40x); c)+9.5~ induction region (640x); d) propagation region (5000x)

393 With increasing temperature the samples undergo considerable plastic deformation and the fracture surface shows an reduction region involving a larger area than in the case of the So blend (Fig. 36b). In this region rubbery domains strongly adherent to the matrix are observed. Such observations may have a more general significance as they indicate that the morphology of a multiphase blend appears to be resolved into its structural elements, on Izod broken surfaces, only when the rate of crack propagation is low; in tum this depends upon testing temperature and type of fracture mechanism. It is interesting nevertheless that even in case of resolved morphology the dements may appear strongly deformed because of the fracture process and different morphological features may be seen in different regions of the broken samples surfaces according to the local rate of crack propagation. It may be concluded that the impact strength and mechanical properties of a multiphase blends should be more correctly correlated with morphological parameters that result from microtomed surfaces or thin sections etched by using suitable solvents, where the structural dements that emerge undergo only small deformation. The morphological features observed on broken surfaces and fracture mechanism exhibited by the B0 binary blends are quite similar to those described by So mixture.

Hg. 37. SEM micrograph of B0 fracture surface at -20~ (640x)

394 However in the case of the B0 blends the average diameter of EPM domains is larger. This finding is partly due to the msut~cient stirring of the mixture during the synthesis (compare Fig. 37 with Fig. 34b). Fig. 38 a shows SEM micrographs of B2 fracture surfaces. At the lowest temperature (T = -20~

the blend exhibits an irregular multicrack propagation

process. It is possible to note, in fact, radial crack distribution around the largest rubbery domains.

This can involve a certain amount of fracture energy

dissipation even at such a temperature. Spherically shaped domains of the rubbery phase, random dispersed throughout the whole surface, with a wide size distribution can be observed (Fig. 38b).

An increase of test temperature drastically changes the fracture

mechanism. In fact, the whole sample undergoes a plastic deformation indicating a transition to a more ductile fracture behaviour (Fig. 38c) consistent with a shear yielding fracture mechanism.

c Fig. 38. SEM micrographs of B2 fracture surfaces at different test temperatures: a)-20~ (40x); b)-20~ (640x); c)+5~ (40x); d)+5~ (640x)

395 The fracture surface of the B5 sample, broken at -20~

exhibits a rough

stepped topography typical of a fast fracture (see Fig. 39a). Because of the high rate of crack propagation, at this test temperature no distinction between the phases can be detected. A transition from a brittle to a ductile impact behaviour by increasing temperature is produced. SEM micrographs of broken surfaces shows a clear limited reduction region plastically deformed (Fig. 39b) with the material volume revolved increasing with increasing test temperature. It is important to note the presence, in the slow crack propagation region, of well-defined rubbery domains, randomly distributed, showing a certain degree of adhesion to the PA6 matrix (Fig. 39c).

Fig. 39. SEM micrographs of B5 fracture surfaces at different test temperatures: a)-20~ (40x); b)+3.5~ (40x); c)+3.5~ induction region (640x)

396 3.1.6. Impact tests

The Izod impact strength values (R) for all the S and B coded blends are reported as a function of the testing temperature in Figs. 40a and 40b respectively.

25 t

R (K Jim 2 ) Composition c"/. wtw~ ~A6 EPM EPM-.q-S,e 100 0 0 90 10 0

,CODE

20

P So

$2 90 $10 90

15

8 0

R (K Jim 2 )

25

CODE

20

S10 10

P

Bo

B=

Composition I_*/.wtw.~ PA6/EPM ~PM-g-S/~

/

o1=olo L /

1001

01

0

I

/

/

/

// /

I /

80118 1 2

I / /

//

B2

10

I0

0 -25

i

-15

Tj,

T (~ i

-5.0

i

5.0

a

P !

15

-25

-15

T (~ I

-5.0

I

5.0

I

15

b

Fig. 40. Impact resistance (R) for PA6, binary and temary blends at different testing temperatures: a)"S" samples; b) "B" samples The PA6 homopolymer (Curve P) exhibits very low R values at temperatures below -10~

In such regions the binary and the temary S blends

show (see Fig. 40) only a slight increase of R. This is certainly due to the total rubber content (10 per cent in weight), which is insufficient to reinforce suitably the PA6 at those temperatures.

Such behaviour is in agreement with the

fractographic analysis that shows fracture mode characterized by a fast crack propagation involving the whole sample (see Figs. 34a, 35a and 36a). Moreover, for S blends the brittle-ductile transition temperature seems to be scarcely dependent on the rubber content as well as on the nature of the added rubber.

397 For all the range of temperatures explored, the S10 blend has R values higher than those of the So and $2 blends. This behaviour may be related to the fact that, as shown by the morphological investigation, in the case of $10 blends the dispersed domains are strongly linked to the PA6 matrix (see Fig. 29c). At temperatures higher than that of transition, such blends show a larger reduction region with respect to the other two S blends (see Figs. 34c, 35c and 36c) where it is possible to dissipate a large amount of fracture energy by plastic deformation. The impact properties of the B0 blend are characterized at lower temperatures by R values only slightly higher than those of the So blend and pure PA6 and by a brittle-ductile transition shifted towards a temperature that is higher than that of any other investigated blend. Such behaviour may be related to the large dimensions of the rubbery dispersed domains and to the very poor adhesion with the PA6 matrix. As shown in Fig. 40b, the BE and the B5 coded blends show, at temperatures below the transition, R values about 3 and 2 times higher than that of pure PA6, respectively. At higher temperatures beyond the transition, the R value of B2 is also larger than that of Bs. It is somewhat surprising that the B2 blend, containing only 2 percent of the functionalized rubber, shows better impact properties than the B~ blend. The finding that blend B2 containing only 2 percent modified EPM-g-SA rubber, exhibits better impact properties than any other blend investigated cannot be accounted for by a particularly high degree of adhesion between matrix and dispersed domains, as the latter are completely removed by boiling xylene (see Fig. 31).

Such behaviour is more likely to be related to an optimum size

distribution of dispersed domains produced together with a rather drastic modification of intrinsic matrix properties following a process of "dissolution" of (EPM-g-SA)-g-PA6 copolymer and/or nucleation effects that gives rise to PA6 spherulites with lower dimension and different structure of the interspherulitic

398 boundary regions [34]. Optical observations of thin sections of PA6 and B2 show that in the case of the blend sample dimensions of the PA6 spherulites are indeed smaller than those observed in plato PA6. For the sake of comparison, in Fig. 41, the impact behaviour of B2 and B5 blends is compared with that of a temary PA6/EPM/EPM-g-SA 80/15/5 blend prepared by melt-mixing of components [34]. From the examination of the trend of the curves in Fig. 41, it emerges that the impact properties of B2 and B5 are slightly better than those of blends obtained by melt-mixing. This finding is in agreement with the SEM observations that indicate a rather similar overall morphology of the samples [34]. R (K Jim 2 )

50-

30 -

10

0-

C

.......

-z5

.-

T (~

-g

1'5

Fig. 41. Comparison of impact behaviour of PA6 and B2, B5 blends with a PA6/EPM/EPM-g-SA (80/15/5) temary blend prepared by melt mixing process (Curve C5). Blend C5 was obtained by melt mixing the components in Brabender-like apparatus at 260~ with a mixing time of 20 mm and rotational speed roller of 32 rpm 3.1.7 Concluding remarks

Caprolactam polymerizations via a hydrolytic process may be carried out in the presence of rubbery components such as EPM and/or modified EPM (EPMg-SA). Two different methods of blends preparation were followed:

399 the first (the solution method, "S" coded blends) involved a preliminary dispersion of the rubber in a suitable solvent able to dissolve the caprolactam and the initiatior; the second (bulk method, "B" coded blends) is characterized by the fact that the rubber is directly added to caprolactam and initiatior and dispersed by mechanical stirring before polymerization. It is found that the degree of conversion of caprolactam to PA6 and molecular mass are practically independent of the content of functionalized EPM-g-SA. Tensile strengths and elongations at break of binary and ternary blends are slightly lower than those of pure PA6. A more marked positive effect on the impact properties of PA6 is found when a small amount of EPM-g-SA (2 and 5 percent ) is added to binary PA6/EPM (80/18) or (80/15) blends. Better impact

behaviour

was

obtained

in

the

case

of the

B2 blend

(PA6/EPM/EPM-g-SA 80/18/2). Evidence mainly supported by morphological and structural analysis shows that EPM-g-SA acts as an emulsifier and interfacial stabilizer between the matrix PA6 and the dispersed main component EPM. As a matter of fact, a more fine dispersion of the rubbery component is observed in blend containing a certain amount of functionalized EPM-g-SA rubber. The interracial activity of EPM-g-SA is due to a true chemical reaction with growing PA6 chains through the formation of a graft copolymer of the (EPM-g-SA)-g-PA6 type. Finally it must be pointed out that the main advantage of the method of blend preparation described in the present paper resides on the fact that with a suitable reactor the number of overall operations is reduced with respect to the melt mixing preparation. Nevertheless, the properties of the blends obtained are similar.

400 3.2 Influence of rubber reactivity and functionalization degree

It has been shown how attempts to prepare binary or temary blends PA6/EPR/PER-g-SA containing high amounts of EPM-g-SA (more than 10% w/w) failed because of the high reactivity of the grafted anhydride groups towards-NH2, end groups of growing polyamide chains [30]. As a consequence of the formation of highly grafted EPM chain, the polymerization mixture was too viscous to be efficiently stirred during the hydrolytic process, thus decreasing the degree of dispersion of the rubber phase attainable. On the basis of those previous results and considering that ester groups exhibit a lower reactivity towards amine groups, a more stirrable reaction mixture during the polymerization and, consequently, a finer dispersion of the rubbery particles in the final polyamide matrix could be expected if an EPM modified by gaffing ester groups is used as a functionalized rubbery component. In the present section we report the preparation and characterization of rubber-modified PA6 obtained directly during the hydrolytic polymerization of CL

in the presence of EPM modified by grafting dibutyl succinate groups,

EPM-g-DBS. The aims of this investigation were: (a) to set up the experimental conditions of the CL polymerization in the presence of EPM-g-DBS; (b) to shed some light on the chemistry involved in the blend preparation and (c)

to

find

and

to

analyse

correlations

between

composition,

functionalization degree of the rubber used, mode and state of dispersion of rubbery components as well as impact behaviour of the blends.

3.2.1 Blend preparation and model reactions

The preparation of EPM-g-DBS samples, each with different amounts of grafted DBS, was accomplished in solution according to the procedure in reference 17. The hydrolytic polymerization of e-caprolactam in the presence

401 both of functionalized EPM and of a mixture of EPM and EPM-g-DBS was initiated by e-aminocaproic acid and made under experimental conditions analogous to those used previously (i.e. 260~

4h) under vigorous stirring in a

nitrogen atmosphere [30]. The total amount of rubber in the feeding was kept constant to a value of 20% by weight since our previous investigations already showed that a blend composition of PA6/total rubber close to 80/20 by weight represents a good compromise between good toughening improvement and reasonable values of tensile modulus [30]. To evaluate in a semi-quantitative way the reactivity of ester groups toward primary amine groups and the influence of this reaction on the molecular weight of the PA6 prepared, a set of reactions involving low molecular weight compounds beating ester or amine groups was preliminary investigated. Furthermore, the products isolated from these "model reactions" are easily characterized and provided us with useful information on some aspects of the chemistry involved in the blend preparation [35]. I

I

Code

PA6 DBS1 DBS2 DBS3

CL ACA DBS A CA (molx 0 3) (molxl0 -3) (molxl0 3) DBS 400 397 402 398

18 20 18 38

0 1.4 3.7 8.6

14.3 4.9 4.4

CL Inherent M~ D B S viscosity(b (g mol 1 (dig "1) x 10"3)(c)

284 109 46

1.30 1.14 0.58 0.37

23.0_+0.5''" 19.0_+0.4 7.0_+0.2 3.5_+0.1

II

Table 9. Hydrolytic polymerization of e-caprolactam in the presence of dibutyl succmate (a) (a) Reaction temperature: 260~ reaction time: 4 h, (b) In m-cresol at 25~ c=0.50g dl 1, (c) Calculated by using the equation of reference 36

402

Hydrolytic polymerization of ~-caprolactam in the presence of dibutyl succinate. Polymerization reactions of caprolactam were made in the presence of different amounts of dibutyl succinate. The polyamides 6 obtained were purified according to conventional procedures; their inherent viscosities were measured and approximate M~ values were calculated by using the empirical relation [36]:

1VI.=15 600~inh1"49

The results are summarized in Table 9 and show a decrease of the molecular weight by decreasing the CL/DBS molar ratio. This behaviour could be expected as a consequence of a stoichiometric imbalance ofoCOOH and -NH2 end groups caused by the reaction between ester groups of DBS chains.

and the growing PA6

The occurrence of this reaction was confirmed by the presence of n-

butanol in the volatile products evolved during the polymerization [35]. Molecular weights, high enough for practical uses, may therefore be reached when the CL/DBS molar ratio is kept close to 2.Sx102+3.0x102. For a blend containing 20 wt% of EPM-g-DBS (DBS = 3wt%) it is possible to calculate, from the data of Table 9, that M, values close to 19x103 g mol l may be foreseen for the polyamidic matrix, provided that gaffed ester groups have a reactivity similar to that of free succinate groups and that the presence of the rubber phase does not influence the polymerization rate.

Reaction of EPM-g-DBS with tridecylamine The reactivity of the sur162

groups towards primary amines and the

nature of the reaction products were investigated by reacting pure dibutyl succinate and tridecylamine (TDA). The reaction was effected in the absence of solvent at 260 ~ C for 4 h, with use of an excess of amine. The isolated reaction product shows the well known amide bands at 3300 cm 1 , 1640 ern1

and

403 1545

c m "1

in the i.r. spectrum while the C=0 ester stretching band at 1738

c m "1

is

absent, thus indicating that only tridecyl succindiamide is formed. EPM-g-DBS was

also treated with TDA in similar reaction conditions.

Two different

-NHJ-COOR molar ratios (10/1 and 2/1) were used. Two sets of reactions were accordingly made, with various reaction times in the range 30-360 min. Analysis of the recovered volatile products formed during the reaction showed the presence of butanol as the mare component.

The extent of the reaction was

followed by i.r. analysis made on thin films of the recovered polymer, by which means the gradual disappearance of the band at 1738 cm 4 and the simultaneous growth of a doublet at 1700 cm 4 (s) and 1770 cm 4 (w), attributable to a cyclic imide group, were observed. Plots of R, defined as the absorbance ratio of the 1738 cm 4 and 1700 cm 4 absorption, against time, for both sets of reactions, are reported in Fig. 42. 2.0 1.5 I:Z:: 1.0 0.5

0

260 360 4oo T (min)

Fig. 42. Dependence of the 1738/1770 c m "1 absorbance ratio, R, on time for the reaction of EPR-g-DBS with TDA. Molar ratio TDA/DBS" 9 2/1; 9 10/1 It is shown that a higher excess of amine groups increases the amount of the reacted ester groups at a given time and that after 240 mm R reaches, for both curves, a limiting value that changes only slightly after an additional 120 mm. Compared with previous results on the reactivity of grafted anhydride groups [25, 30, 37], these results indicate that grafted ester groups have a markedly lower reactivity towards amino groups by comparison with those of the

404 anhydride groups. Moreover, probably a fraction of grafted ester groups does not react because of the fact that in an heterogeneous reaction, only a fraction of ester groups is accessible to the -NH2 groups. It is worthwhile remarking that the chemical and spectroscopic evidence indicates the occurrence of the following reaction:

O II

--CH

C--OC4H9

CH 2 ,,

L

I

+ (313H27NH2

C--OC4H 9

II O

-

C4 HgOH

CH--(~ O

/N-CI3H27 0

thus suggesting that the presence of the macromolecular chains preferentially leads to the formation of a cyclic imide, while a diamide was obtained for free DBS, as shown before.

The following scheme reports the proposed reaction

patterns, when free and grafted DBS are used respectively: O

O

II C--OC4H 9

CH2

+ C13H27NH 2

R--CH

C--OC4H 9 II O

II CH2-C

/

- C4H2H

/

R--CH

C--OC4H 9

II O

(I)

+ C13!'i27NH2 ~ - C4 H9 OH

O

\

II C--NH(CH2)12CH3

CH2

N--(CH2)12CH3

R-CH--C II O

0

CH2-- ICI--NH(CH2)12CH3 --CH

R = EPR

C--NH(CH2)12CH3 II 0 R = H

405 Structure (I) shows a hypothetical intermediate that can give rise to a cyclic imide if DBS is graRed onto EPM or to a diamide if free DBS is involved in the reaction.

Polymerization of ~-caprolactam in the presence of EPM rubbers. Binary PA6/EPM-g-DBS and temary PA6/EPM/EPM-g-DBS blends with a general composition PA 6/total rubber close to 80/20 were prepared by using EPM-g-DBS beating 0.6, 2.0, 4.6 and 6.0 wt% of DBS. A binary PA6/EPM blend was also prepared for comparison purposes. Experimental conditions close to those used in the industrial process of hydrolytic polymerization of s caprolactam were used, and the polymerization was made according to the reported procedure [30], with the use of aminocaproic acid (ACA) as initiator. The starting materials, CL, ACA and EPM and/or EPM-g-DBS, were all mixed together in a vial previously degassed and vigorously stirred at 260~ for 4 h. the prepared blends, after grinding, were extracted by methanol to remove unreacted CL and cyclic oligomers. The products recovered from methanol consisted of about 20 wt% of CL + ACA in the feed. In addition to the well known equilibrium occurring in hydrolytic polymerization of CL, the presence of EPM-g-DBS in the reaction mixture can give rise to the equilibrium

t

i

H

CH

O II C--OC4H 9 + C II O

H2N--R--COOH

~

OC4H 9

R = ~(CH2) 5

~NCOOH

"--

2-C 0

or

R=

.... (CH2) 5

(A)

C--NH(CH2)5~n

406 (A) is subsequently involved in a condensation step with growing PA6 chains fonnmg (EPM-g-DBS)-g-PA6 graft copolymer species, their composition and structure depending on the functionalization degree of EPM and on the molecular weight of the graRed polyamide chains. It is worth noting that, unlike the analogous reaction used to prepare binary blends with 20wt% of EPM-g-SA, in the present case the reaction mixtures allow efficient stirring throughout the 4h required for polymerization. This behaviour may be accounted for by the lower reactivity of the succinate groups with respect to the anhydride group, as shown before and consequently by the lower number of PA6 grafted chains in the former case. It has also been observed that during the preparation of the binary PA6/EPM-g-DBS blends a fine dispersion of the rubber component occur within the first 2h of the process, whereas, for ternary blends, more than 3 h are required and, under the same reaction conditions, the higher the EPM-gDBS/EPM ratio, the faster dispersion is reached. i

Code

i

(we-O/'o) PA6 PA6/EPR A5 B5 C5 D5 A10 B10 C10 D10 A15 B15 C15 D15 A20 B20 C20 D20 Part A

i

Feed composition GraRing degree Intrinsic viscosity" CL/EPM/EPM-g-DB S EPM-g-DBS EPM-g-DBS

0.6 2.0 45 6.0 0.6 2.0 45 6.5 08 1.8 48 5.6 0.6 2.0 4.6 56

(dl gl) 1.54

0.90 1.54

m

1.57 1.24

0.96 0.89 1.54

0.96 0.89

(wt-%) 100/0/0 80/20/0 80/15/5 80/15/5 80/15/5 80/15/5 80/10/10 80/10/10 80/10/10 80/10/10 80/5/15 80/5/15 80/5/15 80/5/15 80/0/20 80/0/20 80/0/20 80/0/20

407 EPM-g-DBS/EPM ratios of 3.0, 1.0 and 0.33 were used in the formulation of the temary blends to investigate the dependence of the morphology of the blends either on the composition or on the chemical structure of the different gaffed copolymers (EPM-g-DBS)-g-PA6 formed. All of the blends prepared and investigated are reported in Table 10 together with the code, the composition after methanol extraction and the inherent viscosity of the PA6 recovered by treatment with formic acid.

Code

Methanol soluble Blend composition fraction CL/EPM/EPM-g-DBS (wt-% on CL+ACA) (wt-%) PA6 20.3 100/0/0 PA6/EPR 22.0 74/26/0 A5 15.0 77/17/6 B5 18.7 76/18/6 C5 18.7 76/18/6 D5 20.0 76/18/6 A10 18.7 76/12/12 B10 17.5 76/12/12 C10 25.0 75/12.5/12.5 D10 20.0 76/12/12 A15 16.2 77/6/17 B15 17.5 77/6/17 C15 20.0 76/6/18 D15 20.0 76/6/18 A20 17.5 77/0/23 B20 20.0 76/0/24 C20 17.5 77/0/23 D20 18.7 76.5/0/23.5

Inherent viscosity b of PA6 (dl g-l) 1.3 1.0 1.0 1.1 1.1 1.0 0.9 -

Part B

Table 10 Characterization of PA6 and of the prepared binary PA6/EPM and PA6/EPM-g-DBS blends and temary PA6/EPM/EPM-g-DBS blends in tetrahydronaphtalene at 135~ b. m m-cresol at 25~ c = 0.50 g dl ~

408

3.2.2 Blend analysis. An analysis of the blends prepared was made on B20, C20 and D20 binary blends and B10, C10 and D5 temary blends by selective extraction of the components by two different solvents. By treatment with formic acid, emulsion were obtained in every case, thus deafly revealing the emulsifying effect of (EPM-g-DBS)-g-PA6 formed. A complete phase separation could also not be reached after several weeks, and more or less opaque solutions, according to the content of EPM-g-DBS used, were obtained with a supematant rubbery phase. As an example, we report the procedure followed on the C20 blend and illustrated below. C20 Blend Formic acid

Supernatant rubbery phase

fA

Soluble fraction Solvent evaporation Pure EPR - g - DBS

Opaque solution

~ Solvent evaporation Pure PA6 +(EPR-g-DBS )-g-PA6 rich In PA6

Residue:(EPR.g.DBS)-g-PA6 rich in EPR

A sample of this blend was treated with formic acid. A supematant rubbery phase was separated and the opaque formic acid solution was evaporated and the recovered product characterized by i.r. analysis. It mostly consists of PA6, as indicated by the pattern of the spectrum, almost identical to that of pure PA6. A very weak absorption at 1738 cm1 shows, however, the presence of small amounts of ester groups.

These findings suggest that the formic acid phase

contains mainly pure PA6 and that the suspended free particles most likely consist of (EPM-g-DBS)-g-PA6 copolymers with a composition rich in polyamide, but also beating unreacted ester groups on the EPM chains. A rough evaluation of M~ of the polyamide recovered from formic acid solution, attained

409 by using the previously reported relation between the measured inherent viscosity and M~, led to a value close to 13.000. The isolated supematant rubbery phase represents 80 wt% of the initial rubber in the feed. Successive treatment of the rubber with xylene reveals that it is almost 65% soluble. Infra-red analysis of the soluble part shows that this consists of pure EPM-g-DBS, whereas the residue shows absorption both at 1640 cm 1 and at 1738 cm 1, characteristic of amide and ester groups respectively. Reasonably, this product, insoluble both in xylene and formic acid, consists of (EPM-g-DBS)-g-PA6 graft copolymers rich in EPM. These results indicate that only about 50% of the initial EPM-g-DBS used in the feed have reacted during the polymerization of the caprolactam and that at a given grat'tmg degree a wide range of composition of (EPM-g-DBS)-g-PA6 graft copolymer was obtained.

3.2.3 Mode and state of dispersion of rubbery components. Scanning electron microscopy has proved to be a valuable technique for a careful inspection of the overall morphology of rubber-modified PA 6. Specimen surfaces were smoothed by using an ultramicrotome and analysed either as obtained or after exposure to xylene vapours. The latter technique proves to be very useful for resolving the morphology of blends, especially for specimens with a high degree of interconnection between the different phases.

Surface SEM.

micrograph (before and alter etching) related to binary and ternary blends are reported in Fig. 43, 44 and 45. The removal of the rubbery phase (consisting of EPM and/or unreacted EPM-g-DBS) shows the morphological features, particularly for the B20 blend, where the dispersed phase is intimately interconnected to the matrix.

410

Fig. 43. SEM micrographs of smoothed surfaces of binary PA6/EPM 90/10 blend: a) before etching (640x); b) after etching (640x)

Fig. 44. SEM micrographs of smoothed surfaces of B20 blend: a) before etching (640x); b) after etching (640x)

Fig. 45. SEM micrographs of smoothed surfaces of C5 blend: a) before etching (640x); b) after etching (640x)

411 Micrographs of smoothed surfaces of blends, after etching arranged at increasing weight ratio (M) of functionalized EPM-g-DBS rubber over total rubber content are shown in Fig. 46-49. The figures refer to values for the degree of functionalization of EPM-g-DBS of 0.6, 2, 4.6 and 6, respectively.

r~g. 4 / . Ut~M rmcrograpt~s or-smoothed surtaces of B-coded samples after etching: a)B5; b) B15; c) B10; d) B 20

412 The mode of dispersion of the rubber in the temary blends, as deafly revealed by etching consists mainly of spherical domains whose dimensions range from about 1 ~tm to more than 50 ~tm (see, as an example, blend D10). The SEM investigation, made mainly on etched smoothed surfaces, leads to the following observations: (i) temary PA6/EPM/EPM-g-DBS blends show a quasi-bimodal mode of dispersion of rubbery components; domains with a size distribution around 1 ~m coexist with very large domains (more than 50 ~tm across).

This bimodal

morphology seems to be more accentuated in temary blends containing EPM-gDBS with higher gra~Jng degree (G). Moreover, increasing M for a given G of EPM-g-DBS, the number of domains per unit area with a diameter larger than 10 ~tm diminishes, whereas the number of smaller sized domains increases (see Figures 46 and 49).

Fig. 48. SEM micrographs of smoothed surfaces of C-coded samples after etching: a)C5; b)C15; c)C10; d)C 20 (ii) Binary PA6/EPM-g-DBS blends show a very fine texture, even though sporadic, medium-size domains are observed for A20 blends prepared from the EPM-g-DBS with lowest G (see Fig. 46d, 47d, 48d and 49d).

413

Fig. 49. SEM micrographs of smoothed surfaces of D-coded samples aider etching: a) D5; b) D15; c) D10; d) D 20 It emerges from the above that the mode and state of dispersion of rubbery components, especially for ternary blend, is strongly dependent not only on the composition, but also on the G value EPM-g-DBS used.

The type of

morphology observed in ternary blends is probably related to the fact that at the early stages of CL polymerization several different processes, often in competition with each other, are effective. Among them the most important are: (a) the dispersion of free EPM and EPM-g-DBS m the reaction mixture (CL, oligomers and PA6 molecules); (b) the possible preferential mixing of EPM-g-DBS and EPM; (c) the reaction of EPM-g-DBS with PA 6 oligomers to form 0EPM-gDBS)-g-PA6 graft copolymer molecules; (d) the dispersion of (EPM-g-DBS)-g-PA6 into the rubbery phase; (e) the dispersion of (EPM-g-DBS)-g-PA6 into the PA6 phase. The relative rate and weight of each of these processes will be influenced by composition, viscosity and G of functionalized rubber. It is likely that when the G value is higher, processes (c) and (e) prevail over the other, and consequently

414 only part of EPM will be efficiently made compatible with PA 6, because not all the rubber has the chance to be in contact with both PA 6 and (EPM-g-DBS)-gPA 6 molecules. This would explain why in temary blends containing EPM-gDBS with higher G one observed, together with a fine dispersion of rubbery particles, the presence of very large domains likely to contain EPM and unreacted EPM-g-DBS. The small particles derive from the fraction of EPM and/or EPM-g-DBS that during the mixing had the chance of being emulsified with PA 6 by (EPM-gDBS)-g-PA 6 graft copolymers, while the large particles consist mainly of rubber that had not been made compatible. For PA 6/EPM-g-DBS binary blends, where there is no free EPM to be emulsified, the anomalous behaviour of the A20 blend could be explained by considering the low G value of EPM-g-DBS (0.6 %) and the low reactivity of the ester group. The molecules of (EPM-g-DBS)-g-PA6 graft copolymer formed during the polymerization, having therefore few PA6 grafts per EPM chain, will be unable to emulsify all the unreacted EPM-g-DBS.

3.2.4 Impact properties. The Izod impact strength, R, measured at two different testing temperatures (-25~ and-10~

is shown as a function of the weight ratio M (functionalized

rubber/total rubber), at constant G in Fig. 50 for binary and ternary blends. From the trend of the curves it emerges that generally R increases with the increase of M: binary PA 6/EPM-g-DBS blends (M=l) being characterized by the highest impact-strength values. It must be pointed out that no complete rupture of the specimens was observed for the blend D20 at- 10~ As shown by Fig. 51, the effect of the grafting degree of EPM-g-DBS on R is dependent on the composition of the blends. For ternary blends with the lowest M values (80/15/5 blends; see Fig. 5 la). R seems to be scarcely dependent upon G.

These blends show an impact behaviour only slightly better than

415 that of the corresponding PA6/EPM (80/20) binary blend. For 80/10/10 temary blends a maximum is observed in R for G values of about 2.

A more

complicated trend is shown by the curves R against G for temary blends with higher M.

Together with the maximum, a minimum at hither G values is

observed (see Fig. 5 lc, d). 20

2O

I

~E 15 I0

5

b

a

0

r

. 0.25

0.50

0.75

1.00

25

l

0.25

/

0.50

i

i

25

20

/

20

NE

'~E I0

15

~: I0

5

1.00

0.75

5

0

C

"' 0 i 25

I

0,50

I

0.75

I

1.00

0

I

0.25

l

0.50

i

0.75

i

1.00

Fig. 50. Impact resistance (R) of binary and temary blends at increasing ratio M (functionalized rubber/total rubber) at two different temperature: 9 =-10~ 9 =-25~ graRing degree: a) 0.6%; b) 2.0%; c) 4.6%; d) 6.0% In binary PA6/EPM-g-DBS (80/20) blends, high R values are found at the lowest G value; as a matter of fact a small variation of R is shown when G ranges from 0.6 to 4.5. At -25~ the largest value of R is attained for a value for G of 6. The fact that binary PA6/EPM-g-DBS blends generally exhibit, at least at the temperatures investigated, the highest R values may probably be related to the small size and rather sharp distribution of dimension of particles dispersed in the matrix of such materials (see Fig. 46d, 47d, 48d and 49d). Nevertheless, especially for temary blends, the dependence of impact behaviour on composition

416 and G for rubber cannot be simply interpreted in terms of mode and state of dispersion of rubbery components as it emerges from the morphological analysis.

20

-- 20 t

~EE

NE 15 "3

I0

=

5

I0 5

0

2

4

6

E

0

2

4

6

25 20 15 I0

5|_td | --

,

2

,

,

0

4

| 2

4

i

9 1 6

8

Fig. 51. Impact resistance (R) of binary and temary blends at increasing grafting degree (G) at two different temperatures: 9 = -10~ 9 = -25~ a) M = 0.25; b) 0.5; c) 0.75; d) 1.0 The fact that a fine texture is not the only condition to give a better impact behaviour is demonstrated by the observation that for a grafting degree of 2 the binary PA 6/EPM-g-DBS (80/20) blends have practically the same R value of the temary PA 6/EPM/EPM-g-DBS (80/5/15) blend, which shows a completely different morphology characterized by a quasi-bimodal distribution of dispersed domains (compare with Fig. 47 c, d). At the same time it may be noted that D-type blends with composition 80/10/10 and 80/5/15 seem to have almost the same overall morphology (compare Fig. 49b, c) and are characterized by a rather different impact behaviour (see Fig. 46d).

The dependence of some mechanical parameters

obtained by tensile tests, such as modulus (E) and ultimate strength ~b on M for different grafting degrees is shown in Fig. 52. It may be observed that in the limit of experimental errors, these quantifies are practically almost independent

417 of the grafting degree, though the mode and state of dispersion of particles is different from one blend to another.

~4

' C:::)

co 0

4-

X

X

-

~E 3 0

0 0

~

9

0

9

tm

(..)

~2

v2

o

o.' o

11o

i o 0

o

0.2

m

o"

~m 0

0.go

Fig. 52. Elastic modulus (E) and breaking resistance ((~b) for binary and ternary blends at increasing ratio M. EPM-g-DBS grafting degree: 9 = 0.6%, 9 2.0%, O = 4 6 % , D = 6.0%; .:. = 0%

This result, together with the observed impact behaviour of binary and temary blends, leads to the conclusion that the mechanical response of a rubbermodified PA 6 must be the outcome of the combination of several factors, generally composition dependent. The type of phase structure, as detected by SEM morphological analysis, is one of these factors: others, such as size and distribution of dimensions of spherulites, lamella thickness and structure of intedamellar and interspherulite regions related to the PA 6 matrix, according to the findings of Martuscelli et al [38]. could play an important role in determining the final properties of blend materials.

3.3 I n f l u e n c e o f b l e n d i n g c o n d i t i o n s a n d d e g r e e o f grafting

In the present section a modification of the above methodology of preparation of rubber toughened polyamide 6 is described, aimed to obtain blends in bulk and in the presence of EPM-g-SA. As previously described [30], it was not possible to polymerize CL in the presence of more than 10% of succinic anhydride modified EPM due to the high reactivity of succinic moieties

418 towards NH2 end groups. The mare effect of such reactivity was a viscosity increase of the polymerization medium already at early stages of the polymerization, which prevented a regular development of the blend morphology. On the basis of such experience, and of the experience in the field of melt mixing of preformed polymers, the modification consists in the addition of the second elastomeric phase at selected times during the polycondensation of the matrix polymer.

3.3.1 Rubber addition and kinetics of PA6 polymerization

All the blends prepared and investigated in the present paper are reported in Table 11. In this Table, the codes are as follows: A is pure Polyamide 6 (PA6); B is a binary PA6/EPM blend; C are blends with EPM-g-SA at 0.7% degree of grafting, and D are blends with EPM-g-SA at 1.4% degree of grafting.

The

subscript numbers (0, 40 and 120) indicate the time of polymerization at which the rubber was added (polymerization time PT).

Code A B Co

C4o C12o Do D4o

D120

DG (%) -

PT" (min) -

Extractable in MeOH(%) 13.5

0

0

-

0.7 0.7 0.7 1.4 1.4 1.4

0 40 120 0 40 120

37.5 27.5 17.8 42.3 29.7 22.2

Composition (%) PA6 EPM-g-SA b C

71.4 74.4 76.8 69.7 73.4 75.7

28.6 25.6 23.2 30.3 26.3 24.3

Table 11. Characterization of PA6 and of the prepared binary PA6/EPM and PA6/EPM-g-SA blends (" See text for defmition; b pure PA6, r PA6/EPM 77.5/22.5 binary blend)

The conditions of polymerization are the same as described in the previous section 3.2; the only important difference is that the EPM-g-SA phase is added not at the beginning of polymerization, but at selected times (PT) and under

419 efficient mechanical stirring. The M~ values of the PA6 obtained are reported as a function of the polymerization time, together with the values of the MeOH extractable fractions, in Fig. 53. It can be seen that the largest increase in the molecular weight is obtained during the first 2h of polymerization, when the polycaprolactam reaches an M~ value of about 16000, 80% of the final value (around 20000). In order to study the influence of the degree of polymerization on the morphology and mechanical properties of the final blend, the rubber was added to the reacting mixture after 0, 40 and 120 min (PT).

30 v

60 ~ 20

tll

40

x

~ x 1"

20 0

0

60

120 180 240 Time (min)

300

Fig. 53. Mn of plato PA6 (0) and MeOH extractables (El) as function of the polymerization time The choice of 120 min as the limiting time of rubber addition has been made on the basis of the following two considerations: (1) the viscosity of the reacting medium for times e x ~ g

120 mm

becomes too high to ensure intimate mixing of the rubber inside the polymerizing matrix, at least with our single-blade stirrer.

More efficient mixers (i.e. a

Brabender apparatus) would be necessary as we reach conditions doser to the melt mixing of high-molecular-weight polymers.

420 (2) We know that a reaction occurs between groups grafted onto EPM and growing PA6 chains, which leads to the formation of an imidic linkage [30]. From the stoichiometry of this reaction it happens that, for ~

= 16000 and

EPM-g-SA at 1.4% by weight of anhydride, the amount of residual amino end groups is almost equal to the amount of anhydride groups. For longer PT, the further decrease in the amino end groups will leave a useless excess of anhydride groups in the system. The influence of PT on the MeOH extractables for C and D type blends is shown in Fig. 54. It can be seen that the addition of grafted EPM increases the amount of extractables.

40!I v

1

ffl

~ ~6

30"

L_ X

9

"r" 0

20-

:~ 10-1 0

60

240

PT (rain)

Fig. 54. MeOH extractables as a function of PT: (e)PA6/EPM-g-SA (DG=I 4%) blends; (A) PA6/EPM-g-SA (DG-~ 76%) blends; (.:.) plain PA6 This effect is more pronounced when the EPM-g-SA is added at the beginning of the polymerization reaction and when the degree of graRing of EPM is larger.

Such behaviour can be explained on the basis of the following

consideration: EPM-g-SA may be considered as a monofunctional tennmator of polymerization and it is known in the literature that similar molecules lead to a decrease in the equilibrium conversion and in the molecular weight of the

421 resulting homo-PA6 [39]. Furthermore, once a growing PA6 chain has reacted with the anhydride group of EPM-g-SA, it can continue its growth only via the COOH end group.

Such a group is known to be much less reactive than an

amino group towards addition to cyclic caprolactam, and it can only condensate with the amino end group of a second growing polycaprolactam chain: The effect of this will be a decrease of the M~ of the grafted PA6 chains with respect to the homo P A6. Another effect that must be taken into account is the influence of the rubber on the 'viscosity' of the polymerizing medium. It was observed in fact that: (1) the addition of the rubber always leads to a more viscous final polymeric melt; (2) the increase in viscosity is more marked when the rubber is added at high PT; (3) no differences were found between the two sets of experiments carried out with rubbers at different degrees of graining; (4) increase in the stirrer speed has a negative effect on the efficiency of mixing as the polymeric melt tends to climb the stirrer shaft. On the contrary, we found it useful to decrease the stirrer speed in the last hour of polymerization, at least for blends with PT=2h.

3.3.2 Blend analysis In a previous section

we described a method of characterization of

PA6/EPM-g-SA blends based on selective altemate solvent extractions [35]. A similar protocol (see Fig. 55) is followed for the blends obtained in the present approach. Finely ground samples of blends were treated overnight with formic acid to remove all the homo-PA6. The resulting emulsion was lef~ to stand in a separatory funnel for the time necessary to give a stable phase separation (this usually takes from a few days to several weeks). At this point, the supematant

422 solid phase was collected, washed with formic acid, coagulated with methanol, dried and weighed. Afterwards, this sample was treated with xylene to eliminate and to weight the rubbery fraction that did not react with polyamide-6. What remains, if anything, should likely be practically pure graft copolymer, (EPM-g-SA)-g-PA6.

C40 Blend Formic acid (RT) ~r

Supernatantsolid rubberyphase Xylene / , r Solublefraction ~ [Solvent ~

~evaporation

PureEPR- g - SA

Opaquesolution I

I Solventevaporation Pure~PA6 + (EPR-g-SA)-g-PA6 richinPA6 "~ Residue:

(EPR-g-SA)-g-PA6 rich in EPR

Fig. 55. Steps in the selective extraction of blend C40 with formic acid and

xylgrle. It must be pointed out that the formic acid solution is never transparent, even after weeks, suggesting that some graft copolymer having longer PA6 chains does not give sediment from formic acid. The results of HCOOH and xylene extractions, together with the final blend composition (i.e. after MeOH treatment), are reported in Table 12. From the data it can be concluded that: (1) for the blends coded C, containing the rubber with lower degree of grafting, the residue of HCOOH extraction, that is rubbery phase, is always less than the initial rubber. A similar effect is not found in the blends coded D.

423 (2) For the series coded C, the C120 shows the lowest amount of HCOOH residue. Code

Composition (%) HCOOH Extraction

Xylene extraction"

PA6

Extract

B

Co C40 C 120 Do D40

D120

(%)

(%)

71.4 74.4 76.8 69.7 73.7 75.7

EPM-~-SA Extract b 73 28.6 77 25.6 78 23.2 86 30.3 68 26.3 72 24.3 76

Residue 25 22 22 11 30 27 20

-

60 -

26 22 27

Residue 24 39 45 70 75 68

Table 12 Characterization of the prepared blends by selective solvent extraction (" Xylene extraction is carried out on the residue of HCOOH extraction; b PA6/EPM 77.5/22.5 binary blend) A possible explanation of such experimental evidence is a follows.

The

reaction between anhydride groups and growing PA6 chains will cause the formation of a shell of reacted rubber around a core of unreacted EPM-g-SA. To allow an extensive reaction of the core, such a shell should be 'soft' and the droplet should break down under shear into smaller spheres. When the degree of grafting is too high, the shell formed is quite tough and the stirring is less able to break the emulsion.

In contrast, at a lower degree of grafting, the shell is

breakable and the grafting reaction can continue during the polymerization. From the above, we can conclude that for the blends coded C, we always have the formation of relatively high amounts of grafted (EPM-g-SA)-g-PA6 molecules, which upon HCOOH extraction remain reside the polyamide phase. Such an effect is more evident for the C~20blend, where the length of grafted PA6 chains is similar to that ofhomo-PA6. Further support for this explanation is given by the last columns of Table 12, where we report the xylene extraction of the residue from HCOOH extraction

424 (owing to the small amount of recovered material, it was not always possible to carry out gravimetric analysis). In the case of blends C, we suppose that the preliminary HCOOH extraction removes all the (EPM-g-SA)-g-PA6 molecules very rich in PA6, which constitute the outer shell of the dispersed phase, and leaves as a residue the core of the particles, mainly made up of less reacted rubber with few PA6 chains on it. Thus, the residue of HCOOH extraction of blends C should be very soluble in xylene, as we actually found (see last columns of Table 12). It must be pointed out that the residue of xylene extraction shows, on DSC analysis, the existence of a PA6 crystalline phase. On the contrary, the preliminary HCOOH extraction in the case of blends D should be unable to remove all the (EPM-g-SA)-g-PA6 from the outer shell of the particles, as many of the grafted molecules are highly and tightly entangled with each other and with the core of the domains, owing to the higher degree of gra~ing. Such an outer shell, rich in short PA6 chains, is also insoluble in xylene, as evidenc~ by the large residue after xylene extraction.

Furthermore, the residue after xylene

extraction, in this case, does not show evidence of a crystalline phase, supporting the idea that those chains are very short. The IR spectra corresponding to the blend C40 at different steps of extraction (see Figure 55), together with the spectrum of pure EPM-g-SA, are reported in Fig. 56 and 57 respectively. It is evident that in the HCOOH residue are present the absorptions of EPM-g-SA and PA6 (see 2500-3500 cm 1 region). Furthermore, we can note, in the carbonyl stretching region, the 'amide I' and 'amide II' band (1640-1545 cm -1) together with a group of three bands (1709,1728 and 1778 cml),

whose attribution is somewhat difficult: the

1728 cm 1 band could be due to the esterification of part of the residual succimc anhydride groups following the methanol extraction, the band at 1778 cm -1 being the symmetric stretching of anhydride.

The 1709 cm 1 band could either be

considered as the acidic moiety of the said hemi-ester or, together with the 1778 cm ~ band, it might be the imide synunetric and asymmetric stretching.

425

!'

4600

3()00 1800 1000 Wavenumber (cm-~)

6()0

Fig. 56. I.R. spectra of the extracted fractions of the blend C4o: (A) HCOOH residue; (B) xylene-msoluble fraction of the HCOOH residue; (C) xylene-soluble fraction of the HCOOH residue; (D) pure EPM-g-SA (see Scheme in Fig. 55)

D

C

B

2200

2000

1800

1600

1400

Wavenumber (cm -1)

Fig. 57. Carbonyl stretching region of the i.r. spectra shown in Fig. 56

426 A similar analysis, carded out on the residues of xylene extractions on C4o and D40 blends (see Fig. 55) is reported in Fig. 58. It can be seen that the specXra always show the presence of amidic linkages together with bands corresponding to the polyolefinic backbone.

B

I

4600

I

3000

1800

|

1000

600

Wavenumber (cm-1)

Fig. 58. I.r. spectra of the xylene-msoluble fraction of the HCOOH residue for the blends with PT = 40 mm: (A) C4o blend; (]3) D4o blend Such amidic groups seem to be more evident in the case of the blends coded C, as a result of the presence of few long PA6 molecules instead of the shorter, even more frequent, chains likely to occur inside the D series. Scanning electron micrographs of microtomed and etched surfaces of all investigated blends are shown in Fig. 59-61.

The PA6/EPM blend (Fig. 59)

exhibits the morphology typical of an incompatible system.

The rubber is

segregated in very large spherically shaped domains, whose dimensions range from 20 l~m to more than 100 t~m. Furthermore, walls of the leR cavities are very smooth, indicating no adhesion to the PA6 matrix. As EPM is substituted by EPM-g-SA, a strong modification in the mode and state of the dispersion of the rubber component is achieved.

Fig. 59. SEM micrographs of a microtomed surface of a hydrolytic PA6/EPM (80/20) binary blend after etching

Fig. 60. SEM micrographs of microtomed surfaces of hydrolytic PA6/EPM-g-SA (80/20) binary blend after etching (degree of grafting DCr-~.7% and PT as indicated)

428 The extent of such an effect, as shown by Fig. 60 and 61, is dependent on the degree of grafting (DG) of the EPM-g-SA used and on the stage of the reaction at which the rubber was added to the reacting mixture (PT).

Fig. 61. SEM micrographs of microtomed surfaces of hydrolytic PA6/EPM-g-SA (80/20) binary blends aider etching (degree of gaffing DG = 1.4% and PT as indicated) From an inspection of Fig. 60 and 61 it emerges that: (1) the

dimensions of etched domains relative to the blends containing

EPM-g-SA with DG-~.7% are much lower than those of the blend containing unmodified EPM (compare Fig. 59 and 60). Moreover by increasing PT the average dimensions of such domains decrease, leading to a very fine and homogeneous texture.

In fact, for the blend obtained with the highest PT,

particles less than 1~tm are present (Fig. 60c or 60d).

429 The above observation can be accounted for by assuming that during the process of blend preparation part of the EPM-g-SA is involved in the formation of (EPM-g-SA)-g-PA6 graft copolymer, which can act as an emulsifier and/or interracial agent. The effectiveness of such a copolymer will be the greater the higher the molecular weight of the grafted polyamide chains. Thus an increase in PT leads to blends with a finer dispersion and an improved interracial adhesion because the graft copolymer formed has segments of molecular weight comparable to those of the PA6 homopolymers. (2) A very irregular distribution of rubbery phase is observed in the case of D blends (see Fig. 61). Tiny domains (~1 ~tm) coexist with medium

size

(~ 10 ~tm) and very large ones (50-100 ~m). Some of them are only partially dissolved by the boiling xylene vapour.

This quasi-bimodal morphology,

however, seems to be more evident in Do and D40 blends (Fig. 6 la and 6 lb). In fact, with increasing PT the number of large domains tends to diminish, whereas the number of smaller sized domains increases (Fig. 61c).

Such an effect, as

already mentioned, is due to an improved action of the formed (EPM-g-SA)-gPA6 graft copolymer, whose PA6 length branches becx~me more and more compatible with PA6 matrix. It is interesting at this point to understand why a change in the degree of grafting of the EPM-g-SA produces such a marked effect on the overall morphology of the resulting blends. For this purpose, two aspects must be taken into account: the grafting reaction and the efficiency of mixing during the polymerization process.

When the DG of EPM-g-SA is low, the overall

reactivity towards the growing PA6 chains is relatively low.

Therefore, the

grafting reaction may occur gradually on the rubber surface continuously renewed by the sheafing forces induced by stirring. This allows the achievement of good mixing and dispersion of the rubbery component. The results will be the completion of the reaction and a continuous size reduction of the particles by the emulsifying action of the grafted copolymer, as evidenced in Fig. 60. On the

430 contrary, at higher DG the reactivity of EPM-g-SA increases and therefore the gaffing reaction starts very rapidly, giving rise on the rubber surface to a graff copolymer rich in PA6 chains. The presence of

such highly grafted

(EPM-g-SA)-g-PA6 chains, strongly anchored to the growing PA6 matrix, enhances the viscosity of the mixture. This will decrease more and more the ability of the stirring apparatus to get a satisfactory dispersion. Consequently all the anhydride groups within the gaffed EPM-g-SA particles are lost with respect to the emulsifying action. Thus the efficiency of the graft copolymer, formed under such conditions, will be much lower if compared to that obtained in the previous case, and therefore a coarser morphology with larger and irregular domains is achieved. The Charpy impact strengths R of PA6 homopolymer and of blends C and D as a function of temperature are reported in Fig. 62. As can be seen, R of pure PA6 (curve A) remains constant at very low values, over the whole investigated temperature range.

The corresponding

fracture surface shows the appearance typical of a brittle material. A similar trend with R values slightly lower is observed for the PA6/EPM blend (curve B). Such behaviour is in agreement with the morphology of this blend and no adhesion to the matrix is present. Passing to the PA6/EPM-g-SA blends containing rubber with DG of 0.7% (curves C) a large improvement in the impact performance is observed with respect to PA6.

The extent of this effect is higher the greater the time of

polymerization at which the rubber is added. Furthermore, the curve of R against temperature shows a clear transition from a brittle to a ductile mode of fracture along the test temperatures. The blend with code C~20 shows the lowest brittleductile transition temperature together with the best impact behaviour.

It is

interesting to observe that the zone around the notch tip of specimens broken at temperature around the transition starts to present a phenomenon.

stress-whitening

431 60-

C120

50-

40E v

30C4o 20-

10-

_

" '9. . . . ....................... 9

-50

I

-40

I

-20

I

0

--

i

20

B

I

40

T(~

Fig. 62. Charpy resilience (R) as a function of test temperature for PA6 homopolymer and for binary PA6/EPM-g-SA blends. Curves are as follows" Code A B Co C4o C12o Do D4o D12o

Composition PA6 PA6/EPM (80/20) PA6/EPM-g-SA (80/20) PA6/EPM-g-SA (80/20) PA6/EPM-g-SA (80/20) PA6/EPM-g-SA (80/20) PA6/EPM-g-SA (80/20) PA6/EPM-~-SA (80/20)

GD (%) 0.7 0.7 0.7 1.4 1.4 1.4

PT (min) 0 40 120 0 40 120

This effect, due to multicraze formation, is more pronounced for high values of PT and test temperature. The blends containing EPM-g-SA with DG of 1.4% exhibit, at equal values of PT, lower impact performance than blends C (compare curves C and D in Fig. 62). The impact behaviour observed in the case of blends C and D is certainly related to the mode and state of dispersion of the rubbery component in the final material.

The blend characterized by more fine regular and homogeneous

432 morphology exhibits the best impact behaviour (blend C120). Thus it may be concluded that, when the blends are prepared concurrently with the hydrolytic polymerization of caprolactam, factors such as degree of grafting of the rubber and the time of polymerization at which the rubber is added to the reaction mixture play a decisive role in detemfining the end properties of the modified PA6. The chemical structure of the (EPM-g-SA)-g-PA6 graft copolymer formed during the blending and polymerization process and the viscosity of the reactive mixture, in fact, depend on such factors. When the degree of grafting of EPM-g-SA used is relatively low and especially at higher values of PT, (EPM-g-SA)-g-PA6 copolymer is likely characterized by PA6 segments whose length is comparable to those of plain PA6. Such a copolymer will have a higher capability to act as emulsifier and compatibilizing agent between the PA6 matrix and the unreacted molecules of EPM-g-SA.

4 General conduding remarks Methods to obtain rubber-modified polyamide-6 (PA6) with improved impact behaviour were described. Such methods consist of the fight formulation of binary and/or temary blends containing as rubbery component an ethylenepropylene rubber (EPM) random copolymer and/or a suitable functionalized EPM. The functionalization of EPM was performed by means of homogeneous gratimg reactions of unsaturated molecules such as maleic anhydride, to form EPM-g-succinic anhydride (EPM-g-SA),

and dibutyl

maleate, to

form

EPM-g-dibutyl succinate (EPM-g-DBS). Two routes have been followed to prepare the blends: 1) melt mixing in a Brabender-like apparatus. 2) concurrently with caprolactam polymerization; In the first case the temary alloys were obtained

by two different

procedures: (a) all the components were introduced at one time in the mixer

433 (single-step mode); (b) the two rubbers were premixed separately before final mixing with PA6 (double-step mode). Mechanical tensile tests at room temperature and Izod impact tests at various temperatures on notched specimens have been performed as well as morphological analysis on smooth and/or etched samples of all blends. Very high impact properties have been achieved for binary blends containing EPM functionalized by routes (1) and (2) and for temary PA6/EPM/EPM-g-SA blends by route (lb). The results have been related to the mode and state of dispersion of the rubber, to the degree of grafting of the EPM and to the blending procedures used.

The influence of rubbery components on the structure and

crystallization behaviour of the PA6 matrix has also been investigated. The

features

observed

in

PA6/EPM-g-SA

(EPM-g-DBS)

and

PA6/EPM/EPM-g-SA (EPM-g-DBS) blends were interpreted by assuming that during mixing a graft copolymer, able to act as an interracial and emulsifying agent, is formed between the functionalized EPM rubber and PA6. In the case of blends with EPM-g-SA, this graft copolymer (EPM-g-SA)-g-PA6 is obtained by means of a heat-induced condensation

between the carboxylic groups of

EPM-g-SA molecules and the-NH2 ~ad groups of polyamide molecules.

The

morphology observed in the case of PA6/EPM/EPM-g-SA temary blends is accounted for if it assumed that the graft copolymer partially acts as an 'interracial agent' (improving adhesion between the matrix and the dispersed phase, and decreasing the average dimension of the EPM rubber domains) and partially forms much smaller domains very rich in (EPM-g-SA)-g-PA6 strongly adherent to the matrix. The influence of the degree of grafting (DG) of EPM-g-SA on the morphology and the tensile and impact properties of binary and temary blends prepared by melt mixing has been investigated. Freer and more homogeneous dispersions of the rubbery domains and better impact properties are obtained with increasing degree of grafting of the EPM in the blends. At equal DG values

434 and for the composition used, the binary PA6/EPM-g-SA alloys show a better behaviour than the ternary ones. These results are related to the presence of an (EPM-g-SA)-g-PA6 graft copolymer formed during melt mixing, which acts as an interfacial and emulsifying agent. Attempts to prepare binary or ternary blends by route 2 containing high amounts of EPM-g-SA (more than 10% w/w) failed for two reasons: the high reactivity of the grafted anhydride towards -NH2 of growing PA6 chains, and the relatively high degree of grafting of the EPM-g-SA used (3% by weight of succinic groups). As a consequence, highly grafted EPM chains are formed from the beginning of the polymerization of caprolactam, producing an abrupt increase in the viscosity of the mixture. Such blends are then too viscous to be stirred efficiently. In order to investigate 80/20 PA6(EPM+modified EPM) blends, in which the amount of modified EPM ranged from 0 to 100% of the total rubber content, by route (2), it was necessary to use EPM-g-DBS. In fact, because of the relatively low reactivity of the ester groups towards -NH2 groups of growing PA6 chains, it was possible to prepare blends with higher EPM-g-DBS content, and to study the influence of different degrees of grafting on the morphology and properties of the blends. In the last section we have described a modification of route 2.

Such a

modification is based mainly on the fact that the functionalized rubber (EPM-g-SA) is added not at the beginning of the polymerization reaction of caprolactam but after selected times. This allows us to study the influence of the molecular

mass

of

preformed PA6

chains

on the structure of the

(EPM-g-SA)-g-PA6 graft copolymer, on the mode and state of dispersion of the rubbery component in the final blends and thus on its properties.

Only binary

PA6/EPM-g-SA blends with an 80/20 (w/w) composition were investigated. Moreover, in order to avoid problems arising from too high degrees of grafting of the rubbery phase, which lead to gelation of the polymerizing mixture, we decided to limit our investigation to a low degree of functionalization of

435 EPM-g-SA. So, two functionalized rubbers, differing in degrees of grafting (0.7 and 1.4 wt% of grafted groups), were used in the preparation of the blends. The approach proved to be successful, and the results obtained were comparable to those of melt-mixing of preformed polymers, in terms of toughening of the matrix PA6 polymer.

References

1. E. Martuscelli, R. Palumbo, M. Kryszewski Eds., in "Polymer blends: processing, morphology and properties" Vol. 1, New York, (1980) 2. M. Kryszewski, A Galeski and E. Martuscelli Eds, in "Polymer blends: processing, morphology and properties", Vol. 2, New York, (1984) 3. E. Martuscelli, in "Thermoplastic elastomers for rubber-plastic blends", S.K. De and A.K. Bhowmick Eds, Ellis Howood, New York, p. 28 (1990) 4. G. Maglio, R. Palumbo in Ref. 2, p. 21 5. A.J. Chompff, US Patent 3380948, (1975) 6. J.H. Devis, UK Patent 1403797, (1973) 7. R.J. Zetlin, US Patent 199591, (1962) 8. G. Illing, Kunststoffe, 1, 275, (1968) 9. UK Patent 248229, (1962) 10.F. Ide, A.J. Hasegawa, Appl. Polym. Sci., 18, 963 (1974) 1I.D. Braun, U. Einsenohr, Kunststoffe, 65, 139 (1975) 12.F.P. Baldwin, G. Ver Strate, Rubber Chem. Technol., 45 (3), 834 (1972) 13.G.D. Jones, in "Chemical Reaction of Polymers", E.M. Fettes, Ed., Interscience, New York, p. 247 (1964) 14.D. Braun, U. Eisenlhor, Angew. Makromol. Chem., 55, 43 (1976) 15.F. Severini, Chim. Ind. Milan, 60, 743 (1978) 16.Y. Minoura, M. Ueda, S. Minozuma, M. Oba, J. Appl. Polym. Sci., 13, 1625 (1969)

436 17.G. De Vito, N. Lanzetta, G. Maglio, M. Malmconico, P. Musto, R. Palumbo, J. Polym. Sci. Polym. Chem. Edn., 22, 1335 (1984) 18.T.C. Trivedi, B.M. Culbertson, in "Maleic Anhidride", Plenum, New York, Chap. 10, (1982) 19.N.G. Gaylord, M. Metha, J.Polym. Sci. Polym. Lett. Ed., 20, 481 (1982) 20.F. Ide, K. Kamada, A. Hasegawa, Chem. High Polym. (Japan), 25, 107 (1968) 21.R. Greco, P. Musto, G. Maglio, G. Searinzi, in "Rubber toughened plastics", C.K. Riew Ed., ACS Publ., Washington, Adv. Chem. Series, 222 (17), 359 (1989), 22.J.C. Masson, in "Polymer Handbook", J. Brandrup, E.H. Immergut, Eds., Wiley-Inter-Science, New York, (1975) 23.B.E. Tate, in "Vinyl and diene monomers", E.C. Leonard, Ed., Wiley, New York, Vol. 1, Chap. 4. (1970) 24.E. Borsig, A. Fiedlerow, M. Lazar, J.Macromol. Sci. Chem. A16, 513 (1981) 25.S. Cimmino, L. D'Orazio, R. Greco, G. Maglio, M. Malinconico, C. Mancarella, E. Martuscelli, R. Palumbo, G. Ragosta, Polym. Engin. Sci., 24, 48 (1984) 26.S. Cimmino, F. Coppola, L. D'Orazio, R. Greco, G. Maglio, M. Malinconico, C. Mancarella, E. Martuscelli, G. Ragosta, Polym., 27, 1874 (1986) 27.W.M. Barentsen, D. Jeinken, P. Piet, Polym., 15, 122 (1974) 28.S. Wu, Polym. Eng. Sci., 27, 335 (1987) 29.R. Greco, M. Malinconico, E. Martuscelli, G. Ragosta, G. Scarinzi, Polym., 28, 1185 (1987) 30.S. Cimmino, L. D'Orazio, R. Greco, G. Maglio, M. Malinconico, C. Mancarella, E. Martuscelli, P. Musto, R. Palumbo, G. Ragosta, Polym. Eng. Sci., 25, 193 (1985) 31 .J. Sebenda, J.Macromol. Sci. Chem., A6, 1145 (1972)

437 32.M.I. Kohan, in "Macromolecular Syntheses", J.A. Moore Ed. Collective Vol. 1, p.93 (1977), John Wiley & Sons, New York 33.G. Illing, in "Polymer blends: processing, morphology and properties", E. Martuscelli, R. Palumbo, M. Kryszewski, Eds., Vol. 1, p. 167, Plenum Press. New York, (1980) 34.A. Casale, F. Speroni, A. Filippi, E. MartusceUi, R. Greco, R. Palumbo, G. Maglio, M. Malinconico, N. Lanzetta, G. Ragosta, L. D'Orazio, It. Patent N ~ 21204/84. 35.R. Greco, N. Lanzetta, G. Maglio, M. Malinconico, E. Martuscelli, R. Palumbo, G. Ragosta, G. Scarinzi, Polym., 27, 299 (1986) 36.J.A. Moore, Ed. in "Macromolecular Syntheses", Collective Vol. 1, John Wiley and Son, New York, p.93 (1977) 37.M. Avella, R. Greco, N. Lanzetta, G. Maglio, M. Malinconico, E. Martuscelli, R. Palumbo, G. Ragosta, in "Polymer blends: processing, morphology and properties", M. Kryszewsky, E. Martuscelli, R. Palumbo, Eds., Plenum Press, New York, p. 191 (1980) 38.E. Martuscelli, F. Riva, C. Sellitti, C. Silvestre, Polym., 26, 163 (1985) 39. G. Odian, in "Principles of polymerization", Wiley, New York, (1981)

This Page Intentionally Left Blank

439 CHAPTER 8 P M M A / rubber blends P. Laurienzo, M. Malinconico, E. Martuscelli, G. Ragosta, M.G. Volpe National Research Council of kaly, Institute of Research and Technology of Plastic Materials, 80072 Arco Felice (Na) ITALY.

1. Introduction

The toughening of poly(methylmethacrylate) (PMMA) with rubber particles while keeping the optical transparency characteristic of the tin-toughened matrix is reported in literature and is obtained by a complex multistage process [1,2]. The rubbers are obtained by a suspension process in which a shell of crosslinked rubbery butyl acrylate-styrene copolymer is polymerized around a core of PMMA. The rubber is then added in the melt to a PMMA matrix. The coreshell structure (Fig. 1) is necessary to assure a good stress transfer between the phases and, at the same time, a better matching of the refractive indices of the two phases in order to get a transparent material.

Fig. 1. Schematic representation of the core-shell structure in commercial high-impact PMMA

440 To achieve the desired level of toughening, high amounts of rubber (20-30% by weight) are necessary, with an unavoidable dimmt~on of elastic moduli. PMMA/poly(ethylene-co-vinylacetate) (EVA) blends consist of a finely dispersed rubbery phase in a glassy matrix of PMMA.

Such blends are

characterized by an interpenetrated network 0PN) morphology, with very small rubbery particles. It is interesting to start with an analysis of IPN systems, with particular emphasis on the HIPS and related structures, as the synthetic process here followed for the PMMA/EVA rubber system is somewhat similar to that industrially used for the toughening of Polystyrene. An IPN is defined as a combination of two polymers in network form, at least one of which is synthetized and/or cross-linked in the immediate presence of the other [3-8]. An IPN can be distinguished from simple polymer blend, graft and block in two ways: 1

-

an IPN swells, but does not dissolve in solvents,

2 - creep and flow are suppressed. We can distinguish four basic types of IPN: -

Sequential IPN

-

Simultaneous IPN (SIN)

- Thermoplastic IPN - Gradient IPN In forming a sequential IPN, (see Scheme 1), the synthetic steps are taken in the following order: a) polymer I is synthetized b) polymer I is crosslinked c) monomer II plus crosslinker is swollen in d) monomer 11 is polymerized with crosslinking and e) phase separation between networks I and 11 takes place.

441 1

2

3

4

Scheme 1

a) and b) may be simultaneous or sequential in time, and e) is usually simultaneous with d), but starts after d) has proceeded to the point where the free energy of mixing becomes positive. For SIN's, monomer 11 (or prepolymer 11) is added before step b). Thus to a greater or lesser extent, the two networks are formed simultaneously. Network I chains are stretched and diluted by network 11 in a sequential IPN, but only diluted in a SIN, altering many morphological and physical properties.

Of

course it is required that the two polymerizations be non-interfering reactions, such as by stepwise and chain kinetics. In a gradient IPN, the composition is varied within the sample at a macroscopic level.

This is conveniently carried out by soaking a sheet of

network I in monomer 11 for a limited period of time, and then II is rapidly polymerized before diffusion equilibrium can occur. This is the method followed in chain polymerization. For step polymerization, such as polyurethanes, one of the component monomers is swollen into the bead first and the second monomer added later.

442

1.1. Thermoplastic IPN

In such IPN, some degree of flow potential is maintained in the material, by means of physical bonds. Thermoplastic IPN behave as thermosets at the use temperature, but as thermoplastics at some more elevated temperature. Dual phase continuity or phase inversion is controlled by two factors: the volume fraction of each component and its melt viscosity.

Obviously, equal

volume fractions and equal melt viscosity promote dual phase continuity. In the study shown in Scheme 2, Kraton G thermoplastic elastomer was used as polymer I. Styrene and methacrylic acid were dissolved into the Kraton G and polymerized in situ. Chemically Blended Thermoplastic IPN's Structural Units ethylene - buthylene center block t'%,, styreneend block ~ physical crosslink site, polymer II

~~

Swell in o, V (monomermix II)

I: ~

ib tltb4ock ~~,,r

Swollen r o t e

II:

-

Neutralization

~

?

Na§O.._C_R NaOH 0 ~Ho--c--R ~ ~ c

FN

0

styrene

physical crosslink site, polymer II Detail

IPN Precur:~r

i?,- C--O-N~"

Na+O " - O_

!

Scheme 2

Upon neutralization, with shearing, the poly(styrene-methacrylic acid) copolymer melt viscosity increased and a phase inversion took place. Let us now briefly discuss the factor determining the phase domains size in the sequential IPN's. Because of the very small entropy of mixing and positive

443 heat of mixing, phase separation is the usual case in two component polymer mixtures, rather than the exception. Far from being undesired, as we have seen previously, phase separation frequently yields unexpected synergisms such as toughening. The researches in the field of polymer blends, grafts, blocks and IPN's, suggest that the size of the domains, among several other variables, is important in determimng physical and mechanical behaviour.

Experimentally, phase

domains sizes range from a few hundred angstroms to several microns, with polymer blends having the largest domains and block r

and IPN's the

smallest. It is known that the domains size in block copolymers is principally controlled by the individual block molecular weights and the ratio of their molecular weights. In the case of IPN's the phase domain size of polymer II depends on the crosslink density of polymer I, the volume fraction of each polymer, the interracial tension and the temperature. Sequential IPN's have been widely investigated in order to elucidate the problem of phase continuity. In a revealing experiment on PBA~S system with PS as second component, it was found that: 1

-

above 20% of polymer network 1I, its phase domains structure was

continuous, 2 - throughout the composition range studied, polymer network I was continuous too.

1.2. HIPS

Let us describe in some details the HIPS production and properties [9-13].

444

The large scale production of HIPS started with Dow's discovery of the graft interpolymerization process. In this process a solution of rubber in styrene monomer is prepolymerized with sheafing agitation until approximately 30% styrene conversion.

A phase inversion process happens during this stage.

Subsequent polymerization (so called "finishing cycle") is carried out in bulk without agitation or in suspension.

With the introduction of transmission

electron microscopy in studying polymer blends, it become apparent that a HIPS system consists of a continuous phase of polystyrene and a dispersed phase containing rubber particles with polystyrene occlusion (Fig.2). Used rubbers are styrene-butadiene rubbers (SBR) and low cis- or high cis-l,4-polybtaadiene (LCPB or HCPB).

Fig. 2. HIPS structure by TEM microscopy of blends with HCPB rubber phase. Darker region is polybutadiene phase: (a), TPS-17: RPV = 29.2%; (b), TPS-18: RPV = 28.5%. The results of HIPS have been interpreted on the basis of graft interpolymerization reactions and of phase diagram shape, which leads to the phase inversion process. In Fig. 3 the phase changes in the polymerization of styrene in the presence of different type and Mw of the rubbers are recorded, in terms of motor power

445 input versus styrene conversion.

Phase inversion takes place between the

maximum and the minimum of the curves, which have the same sigmoidal shape.

30.0

28.0

motor power input (watts)

26.0

24.0

22.0

20.0 0-;-'

12

'

20

'

28

% styrene conversion

Fig. 3. Phase changes in the polymerization of styrene in the presence of different rubbers. Rubber content 7%; rpm = 250. (~) TPS-21: LCPB, M~ = 160,000 g mol-' ; (o) TPS-22: HCPB, M~ = 120,000 g mol-'; (O) TPS-23: SB-8, = 320,000 g mol -~. The rubber particles come to existance at the phase inversion point; keeping the same amount of rubber, the agitation starts at approximately the same styrene conversion, while its duration depends on the viscosity of the system. Adding a SBS (styrene-butadiene-styrene) triblock polymer (SB8) of Mn=320,000, inversion takes place within a range of about 5% styrene conversion.

On the

other hand, in an HIPS with low-cis PB rubber, Mn=160,000, phase inversion is fast within the range of 1%. As a first conclusion, we can say that an attempt to improve the impact strenght by increasing the amount of rubber in the initial solution will make phase inversion more difficult and would require efficient and costly stirrers. In

446 high viscosity robber solutions lower polymerization rates could be more helpful towards completion of the phase inversion.

1.3. Mechanical properties of HIPS

In Table I the tensile properties and impact strenght of HIPS polyblends containing LCPB and HCPB are shown.

Table I

Tensile properties and impact strenght of TPS polyblends containing LCPB or HCPB as robber phase.

Type and

Rubber

Young's

Tensile

Impact

amount

phase

modulus

strenght

Elongation

strenght

Polyblend

of rubber

volume (%)

(MN/m2)

(MN/m2)

(%)

(kJ/m2)

TPS-16

7 parts LCPB

28.0

1310

20.3

28.0

7.9

TPS-21

7 parts LCPB

23.0

1520

18.4

20.0

8.6

TPS-17

7 parts HCPB

29.2

1380

19.9

45.0

9.9

TPS-18

7 parts HCPB

28.0

1385

20.3

45.0

10.0

TPS-22

7 parts HCPB

21.5

1725

22.5

13.0

10.0

Styron 457* ?LCPB 1827 18.6 35.0 7.3 *Styron 457 is a high impact grade polystyrene prepared by Dow Chemical Co.

The robber has a modulus of around 20 MNm 2 while glassy PS has a modulus around 3000 MNm 2. It is evident that the addition of the rubber causes a strong decrease in the modulus. The decrease is higher increasing the RPV (Rubber Phase Volume).

447 The tensile strenght increases due to the decrease in the volume of the softer and weaker rubber phase. The elongation decreases with decreasing the rubber phase volume, while the impact strenght is not influenced very much. The major mechanism of energy absorption is deformation of the matrix polymer, although some energy is taken up in associated deformation of the rubber particles. In HIPS, the dominating mechanism of yielding is multiple craze formation which causes a measurable increase in volume. Plots of volume strain against elongation show in fact a slope of 1, which is indicative of pure craze formation. Zero slope are expected for materials deforming exclusively by shear yielding.

1.4. RTPMMA

Let us now discuss about toughemng methods of PMMA [14-16]. It has been approached by either IPN technique or blending method. IPN was obtained in late '70 by Allen and co-workers. The method followed was to dissolve elastomer (PU) precursors in MMA monomer, then the urethane catalyst is added (DBTI). The cross-linking points of the elastomer network were introduced by trifunctional polyols. Gel swollen with vinyl monomer was obtained which, as the vinyl polymerization proceeded, gave rise to phase separation process. It has been found that the physical and mechanical properties of the composites depend on: 1

-

initiator concentration: the impact stmnght stays constant, while the

modulus decreases at high initiator concentration; 2 - molar ratio of isocyanate to hydroxyl groups: shear modulus and impact strenght shows a plateau at molar ratio 1:1, while decreases at lower and higher;

448 3 - time interval between gelation and MMA polymerization.

Impact

strenght stays constant, while shear modulus shows marked increases with time; 4 - PMMA/PU ratio: tensile strenght and shear modulus decrease increasing the PU fraction, while the impact strenght increases (see fig.4). 20% or more of PU was found necessary in order to get good properties.

2.0

E

o E

-

-

1.5

~- 30

E 20 "~

1.0

Z

x

u)

I.-

v

0.5

0

~

~ 0

10

f

t

2

w

4

Polyurethane (% by wt)

,

,

6

0

Fig. 4. Effect of weight fraction of the PU elastomer on the mechanical properties of the composites. (11), Shear modulus; (V]), tensile strength; (.), notched impact strength Blending methods of toughening of PMMA constitute the totality of commercially available RTPMMA. The method consists essentially in the melt mixing of an "ad hoc" polymerized robber and PMMA.

The rubber has a

complex structure, being the product of an emulsion polymerization in which the core is normally PMMA with multiple shells around consisting of a rubbery layers, crosslinked poly(n-butylacrylate-co-styrene), surrounded by a glassy layer of poly(methylmethacrylate-co-ethylacrylate).

449 A

u 120 oe ,

o

le.

o 100

'7

E ~

80

0

~" lU

60

a. E

411

a.

20

,m

L lU ,C

r~n

Q

~

40

60

[3

r

0

-60

-40

-20

0

20

80

Temperature (*C)

Fig. 5. Relationship between razor-notched Charpy impact energy and temperature for PMMA (~) and RTPMMA (.) containing 36 vol% of rubber particles. The reason for such complex structure is to assure, together with the toughening effect, also a proper matching of the refractive indices of matrix and dispersed phase. According to the literature, the toughening effect (see fig. 5) becomes relevant only for rubber consists of as much as 30% by volume with an unavoidable diminution of the elastic modulus of the base PMMA. In a much more simpler reactive blending approach, we have shown that high impact performances together with the retention of large part of the original transparency of base PMMA can be obtained with very unexpensive ethylene-covinylacetate (EVA) rubbers.

In fact, it is possible to polymerize the methyl-

methacrylate (MMA) by a radical-initiated process in the presence of the preformed rubber [17,18]. The EVA rubber is characterized by the fact that it is soluble m the MMA monomer at temperatures >60~

while it is immiscible with

PMMA. As a consequence, the growing of a PMMA phase can, in principle, give rise to a complex morphology with the separation of a minor phase

450 constituted by EVA droplets trapping some PMMA particles.

The final

microstructure that then might develop, schematically represented in Figure 6, can be described as a multicore shell structure, very similar to that industrially realized. We have demonstrated that the addition of as much as 7% of rubber is sufficient to cause a high improvement of the impact properties without a significant loss of tensile modulus, while the optical transparency, typical of methacrylic matrices, is maintained in most of the blends [19-21]

Fig. 6. Schematic representation of the multicore~shell PMMMEVA synthetic blends.

structure in

2. Synthesis of Blends

The blends were obtained by the following simple method: the EVA copolymers were dissolved in MMA monomer in a ratio 7/100 by weight. After adding a 0.2% by weight of benzoyl peroxide as radical source, a first stage of polymerization was effected under efficient stirring. When the viscosity reached a critical level (normally after about 100 min.), a final curing step was effected in

451 a mold kept at 80~ for 12 h. It is worth noting that the MMA monomer (Fluka product) has not been purified from the inhibitor.

Table H

Codes, composition, and rl.,Y of the used EVA copolymers

Code

VA % by weight

(dL/g) LMWl8

18

0.54

LMW28

28

0.54

MMW40

40

0.70

MMW20

20

0.83

HMW20

20

1.04

HMW9

9

1.06

HMW10

10

1.30

HMW36

36

1.38

~)c = 0.25 g/(100mL toluene) at 30~

As can be observed in Table II, the employed EVA copolymers, kindly supplied by Dupont, show a VA content ranging from 9 to 40% by weight with rlmh values ranging from 0.5 to 1.4 dL/g (viscosimetric analysis was effected in toluene by a Ubbehlode viscosimeter). molecular weight

The employed codes

refer to low

(LMW), medium molecular weight ( M W ) ,

and high

molecular weight (HMW) EVA, followed by the VA content (% by weight). The same codes are hereafter used to identify the prepared blends. The plain PMMA used as reference was prepared by a conventional radical process initiated by organic peroxide (0.2 wt %).

452

3. Morphological Analysis

Electron microscopy spectroscopy was used to characterize the final morphology of the blends and to investigate on the phase inversion process and on the fracture mechanisms. A Scanning Electron Microscope (SEM), Philips 501 model, was used. The analysis has been performed on fracturated surfaces, coated with a thin layer of gold/palladium alloy. Some blends were subjected to a smoothing and polishing procedure and observed after exposure to n-heptane vapors (20 min) to remove the EVA phase.

3.1 Phase inversion

To study the morphological development of phases in our blends, we observed smoothed and polished surfaces after n-heptane extraction. The final morphology of a blend having medium-viscosity EVA is reported in Fig. 7. It is evident that the morphology revealed by the extraction of EVA is very complex. Particularly, the PMMA constitutes the matrix while large regions of EVA with subincluded PMMA particles are deafly evident. High viscosity EVA blends show the same final morphology depicted in Fig. 7. The revealed morphology is very similar to that reported for HIPS [4,5]. For HIPS, it also reported that the absence of stirring leads to a "not inverted" morpholgy, in which a matrix of rubber surrounds large spherical polystyrene domains even if the rubber represents only 7% of the total polymeric material. This situation was obtained also by us and is shown in Fig. 8 where the same blend of Fig. 7 was polymerized under stirring for only 90 mm, i.e., before the attainment of the proper viscosity for "phase inversion". Upon etching with n-heptane, the tiny network of rubber that surrounded PMMA particles is removed. It is evident that PMMA, altough present in a proportion larger than

453 90%, nevertheless constitutes the dispersed phase embedded in a continuous network of EVA copolymer.

Fig. 8. SEM micrograph of a smoothed surface of MMW20 blend (90 min of stirring) aRer etching with n-heptane (640x)

454 The above results deafly show that a so-called phase inversion process must occur during the polymerization. The final morphology then strongly depends upon the stirring conditions. For low-viscosity EVA, even after 110 mm of stirring, the morphology of the cured blend, reported in Fig. 9 after smoothing and etching with n-heptane, shows a "not reverted" situation, with large domains of PMMA surrounded by a tiny shell of EVA. This demonstrates that the MW of the rubber also plays an important role in the development of the final morphology.

Fig. 9. SEM micrograph of a smoothed surface of LMWl8 blend after etching with n-heptane (640x)

3.2 Fractographic Analysis The fractographic analysis of PMMA and PMMA blends surfaces after impact testing at room temperature is reported in Figures 10-13. For all the figures, the notch front is on the left-hand side.

455

Fig. 10. SEM micrograph of fracture surface at room temperature of plain PMMA (320x) The fracture surface of PMMA (Fig.10) shows a series of brittle fracture bands or striations, oriented perpendicular to the crack propagation.

The

formation of these bands is because, above a certain crack speed, the craze preceding the crack front undergoes branching. [22]. At stn~ciently high stress levels, these crazes undergo fracture, causing surface roughening (bands), a deceleration of the crack, and a drop in the stress amplitude around the crack tip of very large domains in a matrix of a tiny layer of EVA copolymer. A similar structure, as already mentioned, is consistent with the absence of phase inversion during the radical polymerization of MMA.

Apart from this observation, the

fracture surface does not reveal any fractographic feature that is below that necessary to initiate branching crazes. The fracture then reverts back, the crack spee~ rises again, and branching reoccurs. The repetition of this process gives rise to the banded appearence of the fracture surface.

Figure 11 shows the

fracture morphology of a blend containing low viscosity EVA copolymer. This blend is characterized by a major-phase PMMA dispersed in the shape distinctive

456 of plastic deformation mechanisms (crazes and/or shear bands) from which crack development can take place.

Fig. 11. SEM micrograph of fracture surface at room temperature of LMWl 8 blend (320x) Completely different are the fracture surfaces of blends containing medium and high-viscosity EVA copolymers, as can be seen from the micrographs of Figures 12-13, respectively. The overall morphology of these blends seems to be very similar. In fact, both photos reveal domains (less than 1 micron in size) finely dispersed and well embedded in the matrix.

Moreover, signs of an

extensive plastic deformation in the matrix are also evident.

In view of the

intrinsic morphologies revealed by the heptane extraction (see previous paragraph), we may attribute the small particles of Figs. 12 and 13 to the glassy PMMA particles subinduded in the rubbery domains, these last domains being highly deformed during the fracture. This indicates that a large amount of energy is dissipated, probably in the form of crazes, during the impact process.

The

above considerations account for the very high impact toughness observed in such materials.

457

Fig. 13. SEM micrograph of fracture surface at room temperature of HMW20 blend (320x)

458 The influence of the VA content of EVA copolymers on the impact properties of blends can be explained as follows: for low molecular weight EVA, where, as we have previously reported, the phase reversion process does not occur during the stage of polymerization under stirring, an increase in the VA content is to improve the polarity of EVA copolymers and to create a stronger interracial adhesion between EVA and PMMA with a consequent improvement in teh impact properties. For MMW blends, such effects are less evident. No large differences are observed in the impact properties as function of VA content of EVA rubbers, at least at room temperature.

4. Mechanical properties 4.1. Impact Behavior

The impact properties were analyzed according to the Linear Elastic Fracture Mechanics (LEFM) approach [23].

The procedure used for the

calculation of the critical strain energy release rate (Gc) and the critical stress intensity factor (Kc) is reported elsewhere [24]. Charpy-type specimens (6.0 mm wide and 60 mm long) were cut by a mill and notched with a fresh razor blade. Then, they were fracturated at different temperatures and at an impact speed of 1 m/s by using an instrumented Charpy pendulum. To obtain sheets of 3.00 mm thickness, the blends and the pure PMMA were compression-moulded in a heated press at 200~ and at a pressure of 240 atm. For simplicity, the impact values can be collected into three groups as function of the molecular characteristic of employed EVA, i.e.: (a) Blends containing low-viscosity EVA (rl~ = 0.54 dL/g). (b) Blends containing medium-viscosity EVA (rl~ = 0.7-0.83 dL/g). (c) Blends containing high-viscosity EVA (rl~ = 1.04-1.38 dL/g).

459 In Figures 14-16, the values of Gc (energy release factor) vs. temperature are reported for, respectively, low-, medium-, and high-viscosity EVA.

Gc (KJ/m=)

2.0

O LMW 18 VA 18% ~inh 0.54 9LMW 28 " 28% .... 0.54

1.5

J

1.0

-~-'~"~

~

PMMA

0.5[

-80

-60

-40

-20

0

T(~

Fig. 14. Energy-release factor Gc of plato PMMA and of PMMA/lowviscosity EVA blends as function of the testing temperature

Gc( KJ/rr~ 3.0

*

/

MMW40

9 MMW 20

VA 40% ~ i n h 0.70

......

2.0.

f

1.0~ v

9

-80

-at

A

.

.

v

&

-60

,

v

I

-40

.

.

v

I

-20

.

v

I

0

v

PMMA

I

I

20

40

T(*C)

Fig. 15. Energy-release factor Gc of plato PMMA and of PMMA/medium viscosity EVA blends as function of the testing temperature.

460

G ( K J / m =) 3.0

9 H M W 20 VA 20% ~ i n h zx H M W 36 " 36% . . . . o H M W 10 ,, 10% . . . . 9 HMW 9 9% . . . .

1.04 1.38 1.30 1.06

2.0

PMMAIPVAC 1.0

-80

PMMA

-60

-40

-2

0

20 T(~.)

Fig. 16. Energy-release factor Gc of plato PMMA, of PMMA/highviscosity EVA, and of PMMA/PVAc blends as function of the testing temperature. As can be observed from the impact data, the blends obtained with lowviscosity EVA are characterized by Gc values similar to, or worse than, that of pure PMMA, in all the ranges of investigated temperatures. Going to medium-viscosity EVA (see Fig. 15), the behaviour of blends is definitely more satisfactory.

In fact, the values of Gc for both kinds of

copolymers are well above the values of PMMA for all the investigated temperatures. The major improvement is observed at temperatures higher than 10~

where PMMA still behaves as a brittle material, while the blends undergo

brittle-to-ductile transition. The differences in the Gc values of the two blends are relevant at temperatures below 0~

while their curves tend to overlap at

room temperature. A similar trend is observed for the blends reported in Fig. 16 having EVA copolymers with the highest molecular weights. Another important variable that we have studied is the vinyl acetate (VA) content. In Table 111, the Gc values at 20~ are reported. For blends with low MW EVA, the Gc diminishes, increasing

461 the VA content, but at high MW EVA (i.e., rl~ > 1 dL/g), the Gc goes through a maximum.

Table Ill

Codes of the blends, % of VA content of the rubber phase in each blend, and critical strata of energy release rate (Gc).

Code

VA content (%wt)

Gc (20~

LMWl8

18

1.3

LMW28

28

0.9

MMW40

40

3.4

MMW20

20

3.5

HMW20

20

3.6

HMW9

9

2.3

HMWl0

10

2.5

HMW36

36

3.2

PMMA

-

1.0

PMMA/PVAc

100

1.1

4.2. Tensile properties

As a high elastic modulus is an important characteristic of PMMA, tensile testing was performed in order to investigate about the amount of its forecastable diminution in these new rubber-modified PMMA based blends. We tested the

462 plain PMMA used as reference and the MMW20 blend, as it showed complete phase-inverted

morphology,

excellent impact properties

and

outstanding

trasparency. Tensile testing was performed using an Instron machine. Specimens of 1.6 9 4.5 930 mm were obtained from sheets compression-moulded at 200~ heated press at 240 atm. The samples were tested at 20~

in an

and at a constant

cross-head speed of 10 mm. mm -~ . From the stress-strata plots, the values of Young elastic modulus (E) and of ultimate tensile strength ( ~ ) were obtained (Table IV). It can be observed that the diminution of E is neglectable, as consequence of the low weight content of the rubber.

On the contrary, the ~

value shows an abrupt diminution,

comparable to those observed in classical PMMA-rubber toughened materials. This diminution is a consequence of the high apparent volume ratio between dispersed phase and matrix, due to the core-shell structure which develops in these systems.

Table IV Young's elastic modulus (E) and ultimate tensile strenght ( ~ ) of plain PMMA and MMW20 blend Code

E (Kg. cm -:)

cr~ (Kg. cm -~)

PMMA

1.5.104

440

MMW20

1.3 9104

250

5. Thermal Analysis The thermal properties of EVA rubber and of a PMMA-EVA blend (namely, the MMW20 blend) have been analyzed in the DSC experiments

463 reported in Figs. 17-18, respectively. A Mettler System TA-3000, equipped with a control and programming unit (microprocessor TC 10A), was used. The system was provided with a calorimetric cell DSC-30, which allowed temperature scans from -170 to 600~

The experiments were performed under

a nitrogen flow. Fig. 17a shows the DSC melting trace for EVA MMW20 and Fig. 17b shows the crystallization trace. The EVA rubber has two clear melting peaks (at about 50~ and 90~

and two clear crystallization peaks (at about 65~

and

40~

1"

1"

E

o_ E ,-

.u

,-

-,--1 ~ - ~

-100

T"--r"-'-T

-50

0

~'-"-~'-"--r --'-"-"

50

Temperature

-1-~" " - " - "- I . . . .

100

150

1 ~

200

9~ " - T ' -

250

2so

2oo'-';Co

,~o

Temperature

(~

(a)

~'o

o

.so

-,oo

(~

(b)

Fig. 17. DSC traces of EVA" (a) meltmg~ and (b) crystallization.

Figs. 18a and 18b show DSC traces for the melting and crystallization, respectively, of the MMW20 blend. crystallization peaks between 50~

The blend shows weaker melting and and 90~

Fig.18 also shows a transition at about 110~ transition of the PMMA phase.

and between 70~

and 25~

this corresponds to the glass

464

1"

1'

u E o

i.u

Temperature (~

(a)

Temperature (0C)

(b)

Fig. 18. DSC traces of MMW20 blend during" (a) melting, and (b) crystallization.

6.Graft copolymers formation The high impact properties of HIPS are at least partially interpreted as a result of the formation of graft copolymer species between polybutadiene and the growing chains of polystyrene. In our case, we used radical polymerization to grow PMMA, and EVA copolymers are constituted by long polyethylene sequences that are known to be reactive toward radical copolymerization. To get indirect proof that grafted EVA-g-PMMA species are formed in our system, we prepared and characterized a blend using poly(vinyl acetate) as a second phase. The Gc values at different temperatures reported in Figure 16 and the value of Gc at room temperature reported in Table 111 show that this blend has very poor impact properties, comparable with those of plain PMMA.

465 Also, the fracture surface (see Fig 19) resembles that of unmodified PMMA.

The surface is covered with brittle fracture bands originated from a

cyclic process of formation and breakdown of crazes.

Fig. 19. SEM micrograph of fracture surface at room temperature of PMMA/PVAc blend (160x). We believe that this result indicates that, in the absence of graft copolymerization reaction between PMMA and the dispersed phase, an effective toughening process cannot occur, although the dispersion of the minor phase is very intimate.

This could also explain why the impact properties of HMW

blends seem to go through a maximum as a function of VA content (HMW20>HMW36). In fact, we can suppose that a balance occurs between the increase in interracial compatibility between EVA and PMMA by increasing the VA content (better impact properties) and the decreased ability to form graft

466 copolymer species, due to the diminution of the length of polyethylenic sequences on EVA chains at higher VA content.

7. Optical properties The

PMMA/EVA

blends

which

show

a

complete

phase-inverted

morphology keep a transparency at room temperature very close to that of pure PMMA. This is due to the effect of equalization of the refractive indices of the two components, related to the particular multicore structure. In fact, a blend prepared simply by melt-mixing preformed EVA and PMMA polymers in the same weight percentage (i.e., 7/100 w/w) in a Brabender-like apparatus is completely opaque at any temperature. Moreover, these blends exhibit a very peculiar and interesting optical behavior [19-21].

In fact, altough almost completely transparent at room

temperature, their transparency gradually decreases increasing the temperature until, at T > 70~

they become completely opaque. This temperature-dependent

opacification, which is reversible, has been ascribed to the melting of ethylenic sequencies of the EVA component, which gives rise to a light-scattering phenomenon; taking into account the microstructure of the blend, the description of this kind of scattering can be carried out in the Rayleigh-Gans approximation [251. In conclusion, the optical behavior of these blends appears attractive in view of their possible applications as a temperature-controlled optical device or as an active element for thermally induced optical bistability.

References 1) C.J. Hooley, P.R. Moore, M. Whale, M.J. Williams, Plast. Rubber Process. Appl. k 345 (1985)

467 2) C.B. Bucknall, J.K. Partridge, M.V. Ward, J. Mater. Sci.19, 2064 (1984) 3) L.H. Sperling, "Interpenetrating Polymer Networks and Related Materials", Plenum Press, New York (1981) 4) Yu.S. Lipatov and L.M. Sergeva, "Interpenetrating Polymeric Networks", Naukova Dumka, Kiev (1979). 5) D. Klempner, Angew. Chem. 90, 104 (1978) 6) D.L. Siegfried, D.A. Thomas and L.H. Sperling, J. Appl. Polym. Sci., 26, 177(1981). 7) J.K. Yeo, L.H. Sperling and D.A. Thomas, Polym. Eng. Sci., 21,696, (1981) 8) D. Klempner and D.C. Frisch, eds., "Polymer Alloys 11", Plenum Press, New York (1980). 9) G.E. Molan, J. Polym. Sci., A-3, 4235 (1965) 10)C.G. Bucknall, "Toughening Plastics", Applied Science, London (1977) l l)K. Kato, Polym. Eng. Sci., 7, 38 (1967) 12)G.F. Freeguard, M. Karmarkar, J. Appl. Polym. Sci., 15, 1649 (1971) 13)K. Sardelis, H.J. Michels, G. Allen, J. Appl. Polym. Sci., 28, 3255 (1983) 14)G. Allen, M.J. Bowden, D.J. Blundell, F.G. Hutchinson, G.M. Jeffs, J. Vyroda, Polymer, 14, 597 (1973) 15)N. Shah, J. Mat. Sci., 23, 3623 (1988) 16)M.E. Fowler, H. Keskkule, D.R. Paul, Polymer, 28, 1703 (1987) 17)P. Laurienzo, M. Malinconico, E. Martuscelli, G. Ragosta, M.G. Volpe, Italian Patent N ~ 47946A89 18)P. Laurienzo, M. Malinconico, G. Ragosta, M.G. Volpe, Angew. Makromol. Chem. 170, 137 (1989) 19)G. Carbonara, P. Mormile, G. Abbate, U. Bemini, P. Maddalena, M. Malinconico, in "Physical Concepts of Materials for Novel Optoelectronic Device Applications I: Material Growth and Characterization",M. Razeghi ed., Proc. Soc. Photo-Opt. Instrum. Eng., ~

688 (1990)

468 20)U. Bemini, G. Carbonara, M. Malinconico, P. Mormile, P. Russo, M.G. Volpe, Appl. Opt., 31, 5794 (1992) 21)U. Bemini, P. Russo, M. Malinconico, E. Martuscelli, M.G. Volpe, P. Mormile, J. Mater. Sci., 28, 6399 (1993) 22)F. Coppola, R. Greco and G. Ragosta, J. Mater. Sci., 211775 (1986) 23)A.J. Kinloch and R.J. Young, "Fracture Behavior of Polymers", Applied Science, London, 1983 24)M.J. Doyle, J. Mater. Sci., 18, 687 (1983) 25)G. Abbate, U. Bemini, P. Maddalena, S. de Nicola, P. Mormile and G. Pierattini, Opt. Comm., 70(6), 502 (1989)

469

CHAPTER 9

POLYCARBONATE TOUGHENING BY ABS Roberto Greco

Institute of Research and Technology of Plastic Materials (IRTEMP) of Italian National Research Council (CNR) - Via Toiano, 6 - 80072 - Arco Felice (Napoli) Italy.

1. Introduction

Polycarbonate (PC) and Acrylonitrile-Butadiene-Styrene (ABS) blends are commercial products since many years; they have received a particular attention in patents and technical applications [1-3], such as in automotive industries. The reason of their success on the market is due to their excellent thermal, mechanical and impact performances. The two components offer, in fact, a good compensation of properties, as summarized in table 1, where positive and negative technological aspects of both PC and ABS are schematically illustrated. It is of a particular interest to the purpose of this paper to analyse the blend properties and particularly the impact performances, with respect to their simple and peculiar way of preparation. Their good mechanical and impact behaviours are obtained, in fact, by simple melt-mixings in common equipments, without addition of any specific additive for component compatibilization. This achievement can be considered rather an exception when compared with more complex techniques of toughening

used

for

other

incompatible

systems.

Suitable

mterfacial

compatibilizing agents must be generally added, in fact, to the mare components in order to reach the goal. Furthermore this is achieved by more or less complex

470 procedures, such as, for instance, reactive blendings. Several of these examples are illustrated elsewhere in this book. Therefore, since the behaviour of PC/ABS blends, is not yet clearly understood from a scientific point of view, it is worth to undertake a systematic investigation of these blends. In recent years, in fact, only a limited number of papers have been issued in literature, concerning some of their properties. As a matter of fact, PC/ABS alloys consist of four polymeric species, PC, polystyrene (PS), polyacrylonitrile (PAN) and polybutadiene (PB), the last three in form of copolymers, S-co-AN (SAN) and PAN grafted onto PB, all compounded in complex multiphase systems.

Table I Comparison of technological qualitative behaviours between PC and ABS technopolymers.

Behaviour

Positive

PC

"

ABS

high heat distortion T

economy

mechanical resistance

processability

low T toughness

impact strength

transparency

notch sensitivity

dimensional stability electric properties

Negative

processability notch sensitivity stress cracking chemical resistance

low heat distortion T

471 For a systematic study, necessary to highlight the intimate reasons of the blend performances, it is convenient to briefly describe chemical constitution, processing and properties of the single components. The successive step is that of describing the behaviour of PC/SAN blends before affording the more complex PC/ABS multicomponent systems. Finally the PC/ABS blend properties will be analysed and discussed. The acquired direct knowledge on PC/ABS blends could be useful for developing similar compatibilization techniques to be utilized for other systems.

2. Blend components

2.1 PC PC, consisting of linear thermoplastic polyesters of carbonic acid with aliphatic or aromatic dihydroxy compounds, can be represented by the general structure: O

II 0

C~O

A detailed description of all aliphatic and aromatic polycarbonates, a very large family of condensation polymers [4,5], is, however, well beyond the scope of this paper. Here only the Bisphenol-A Polycarbonate (PC-BPA), the most used polycarbonate for commercial blends with ABS is of interest, whose structure is the following:

L

o_c_o

PC-BPA will be simply indicated as PC in the remainder of this paper. It has a good thermal stability and the dry PC may be kept for hours at 310 ~ in the molten state, thermal degradation starting above 400 ~

472 Its injection moulding grades possess all average molecular weight (MW) generally comprised between 22000 and 32000. Tensile, impact and flexural strength increase with MW up to about 22000, beyond this value their improvement becomes much smaller, whereas the melt viscosity increases more sharply. Therefore a compromise is necessary to obtain a viscosity sufficiently high to get satisfactory mechanical properties but sufficiently low to impart flow characteristics suitable for filling complex moulds. PC is thermally and mechanically stable up to its Tg, lying around 150 ~ (its storage modulus, G', is only slightly dependent on temperature over this temperature interval). PC exhibits in solid state a toughness higher than other amorphous glassy polymers, such as polystyrene and polymethylmethacrylate. This has been attributed to a broad G" maximum, the so-called ~/-relaxation, occurring at about -100 ~ I61. A brittle-to-ductile transition occurs below room temperature. It depends on several variables, such as temperature, MW, loading rate, state of stress, thermal treatments (physical ageing and annealings), additives, external media, sample thickness and notch sharpness. PC, although exhibiting tough behaviour in stressstrain and in unnotched impact tests, is very sensitive to sharp notches and to specimen thickness.

2.2 SAN Styrene-acrylonitrile (SAN) copolymers are low-cost materials of increasing commercial importance, prepared from acrylonitrile I CH 2 = CHCN ] and styrene ICH2 =

CHC6Hs] monomers [7,8]. The monomers mixture is generally polymerised

according to the azeotropic ratio (76 wt % of styrene) in order to better control the copolymer composition.

473 SAN copolymers solve the problem of the bad processability of PAN, being easily moulded and shaped by conventional equipments. They are strong, rigid and transparent materials, with high dimensional stability, craze resistance and strong resistance to liquids, such as water, acid and basic aqueous solutions, detergents and bleaches as well as to solvents, such as oils, gasolines and kerosenes. They exhibit a better stress-cracking resistance than general-purpose PS in several environments, together with a good thermal stability, low creep behaviour, excellent tensile and flexural strength, surface hardness, great rigidity and good resistance to weather agents. Most of these characteristics are improved with increasing the AN amount in the copolymers, showing best behaviours in the range of 20-35% in weight of AN in SAN. Beyond this values a yellowing effect increases too, requiring suitable additives. The combination of the properties above illustrated has made SAN suitable for end-uses in several fields such as buildings, automotives, major and minor appliances in domestic equipments, packaging, home furnishings and others.

2.3 ABS

Acrylonitrile-butadiene-styrene (ABS) copolymers are a large family of thermoplastic

materials,

containing

an

elastomeric

component,

usually

polybutadiene (B) or a polybutadiene-based copolymer, in form of domains well dispersed in a thermoplastic matrix of SAN [9-11]. The SAN is graRed onto the elastomer in order to obtain a suitable dispersion and a reduction of the domain size. ABS can be produced by several techniques: a) mechanical blending of SAN and a SAN-co-B copolymer in common mixing equipments. b) polymerization: there are three commercial processes for ABS manufacturing:

474 b~) emulsion, involving a two step process: 1) production of an elastomeric substrate, made by PB, or Styrene-co-B (SBR) or AN-co-B (NBR) random copolymers; the S or AN amount must be less than 35 %, in order to keep the rubber glass transition temperature (Tg) contribution sufficiently low; the reaction is carried out in a batch reactor and a careful control is made on the particle size (in the range of 0.05-0.5 pan) and on the crosslinking degree, both affecting the gratting efficiency of the second stage of the process; 2) copolymerization of S and AN and simultaneous grafting reaction of SAN with the rubbery substrate formed in the first step. High rubber contents can be obtained by this technique ( up to 50 % ). These materials are often blended with SAN or other ABS materials to vary their initial rubber concentration and particle size distribution. Advantages are low temperatures and pressure used in the process and a wide range of products available. Disadvantages are high energy requirements. b2) suspension polymerization, involving a two-step process: 1) the rubber, dissolved in a styrene-acrylonitrile mixture together with a chain-transfer agent and an initiator, is charged to a prepolymerization reactor; the particle size (generally in the range of 0.5-5 lam) is partially controlled by adjusting MW and stirring intensity; the phase inversion sets the maximum rubber amount which can be incorporated since higher rubber contents increase too much theviscosity and consequently the average particle size; 2) the prepolymer is charged to a suspension reactor along with water, suspending agent, initiator and a chain-transfer agent. Water and energy consumption are lower, but the wastewater is more concentrated than for emulsion process.

b3) bulk polymerization, involving a continuous two-step process. A butadiene rubber is dissolved in a styrene-acrylonitrile mixture together with a transfer agent, an initiator and a diluent, for controlling the viscosity, in a

475 prepolymerization vessel. Discrete rubber particles including SAN and monomer are formed, whose size (0.5-10 ~tm) is controlled by a high shear stirring. The prepolymerized material is continuously charged to a polymerization reactor, where the rubber particles are cross-linked, retaining the shape previously acquired. The rubber percentage is limited up to 15-18 %, since, beyond this concentration, the viscosity becomes too high and the material processing too difficult. Advantages of this method are low energy requirements and low wastewater amounts and, hence, low costs of production. Disadvantages are high cost equipments, minor product flexibility, due strong limitations in processing highly viscous polymer melts, less complete conversion from monomer to polymer. This last effect requires, for most ABS, a devolatilization process, in order to freed them from residual monomer prior to compounding of the final product. All these methods of preparation yield a family of ABS materials with a large flexibility, depending on composition, MW, degree of graRmg, rubber particle size and morphology, allowing the tailoring of properties suitable to meet specific enduses. However, in spite of the different processing utilized, the morphology mainly consists, in all the cases, of a rubbery phase dispersed in a compatible way in a SAN matrix, due to the grafting of SAN onto the B-based rubber. ABS are engineering thermoplastics exhibiting good processability, excellent toughness and sufficient thermal stability. They have found application in many fields, such as appliances, building and construction, business machines, telephone, transportation, automotive industries, recreation, electronics and others.

3. Blends

3.1 PC/SAN blends A certain number of papers have analysed the properties of blends made by

476 SAN and PC [12-55] . One of the main aims, was to get information on their behaviour as a necessary step for affording the analysis of the more complex multicomponent PC/ABS systems. Keitz et al. found experimentally [12] that a number of properties reached a maximum value, for blend specimens having a AN percentage in SAN, ranging from 25 up to 27 % in weight: 1) lap shear adhesion of compression moulded laminated sheets of PC and SAN copolymers; 2) mechanical tensile modulus and elongation at break; 3) notched impact Izod strength; 4) inward shifts of the Tg of PC and SAN blend components with respect to the homopolymer values, detected by DSC and dynamic-mechanical tests. These findings were interpreted as due to PC and SAN adhesion, induced by a partial miscibility varying with SAN internal composition in the blends. A simple binary interaction model, already utilized for a copolymer, made of monomer 1 and 2, mixed with a pol3qner (3), was proposed for interpreting this behaviour

[12,13]. The overall interaction energy density B, as a function of three component binary interaction parameters, can be written as follows: B = B,3 ,J,', + B23 'b'2 - B,2 ,J,', 'b'2

(1)

where ~'~ and {D'2 a r e the monomer volume fractions in the copolymer. The first two terms, representing interaction parameters between polymer 3 with 1 and 2 monomers are linearly additive, as shown in Fig. 1. The third term, taking into account the intramolecular interaction parameter between the two monomers, is of a quadratic form. In the case of endothermic mixing, B~j in eq. 1 are all positive, then B, as a function of ~'~, tends to exhibit a minimmn ( dZB/d~}'l 2 - 2 B12 > 0 ).

477 Furthermore [ 14] for: B,2

>

(2)

(~mB-13 + -x/B23) 2

B becomes negative and therefore a miscibility window must exist in the composition range between +'~, and ~'~b, as shown in Fig. 1.

B Bls

B2s i

B12) 0 o

Figure 1.

1

Interaction parameter B versus ~b'~, for different B~e values,

greater than O, increasing in the arrow direction (after Keitz et al. [12]).

In our case SAN is the copolymer and PC the polymer: the maximum in the curve of lap-shear stress versus SAN composition could be an evidence of the existence of a miscibility window. It is interesting to note, for later discussion, that the above mentioned range of AN contents (25-27 wt %) is very close to the azeotropic composition of SAN, often used as matrix phase for ABS materials, as already mentioned above.

478 In Fig. 2 the lap-shear stress of adhesion, between PC and SAN sheets, is reported as a function of the AN percentage in SAN. !

I

I

I

I

I

10

20

30

40

50

60

_

T

(Pa)

0

70

AN (~)in SAN

Figure 2. Average lap-shear stress as a function of SAN internal composition (after Keitz et al. [12]).

Mendelson [15] studied the miscibility in blends of SAN copolymer with PC and several other polymeric species. PC and SAN were found to be phase separated but partially miscible in a broad concentration range of SAN (23-70% AN), as evidenced by Tg measurements. This result was in partial agreement with those of Keitz et al. [ 12] and of Locati et al. [ 16]. He assumed that a Tg linear model was applicable to each of the separated glassy phases (made by SAN-rich and PC-rich domains), as in the case of a homogeneous blend of two miscible polymers. He used the additivity model of a modified Gordon-Taylor equation [ 17], where the difference between the thermal expansion coefficients in liquid and solid states is taken as a constant for all polymers, as proposed by Tobolsky [18] :

479 Tglblend -- X

SAN

Tg

SAN

+ X PC Tg

PC

(3)

He calculated from the model that: a) less PC entered the SAN-rich phase than SAN the PC-rich phase; b) the SAN composition had only a small effect on the PC-SAN miscibility; that is, on the proportion of PC-rich and SAN-rich phases in the overall blend as well as on the distribution of PC and SAN between these two phases. This last finding was in contrast with the conclusions of Keitz et al. [12], for whom a maximum of miscibility existed in a restricted SAN composition range, as reported above. This could indicate that the application of the Tg model was not suitable for interpreting the data. TEM observations showed rather diffuse phase boundaries between PC and SAN domains. Both the findings, the conflicting results of the Tg model application as well as the TEM features, suggest a certain degree of interpenetration of PC and SAN domains across the interface. From mechanical tensile tests the amount of material deformed by crazing was found to decrease with increasing the PC content in the blend. Gregory et al. [19] analysed, by means of a torsion pendulum, the dynamicmechanical behaviour of multilayered composites, made of PC and SAN alternating laminates. Two series were made, consisting of 49 and 193 alternating layers of same overall thickness but of varying composition. The outer layers consisted of PC in all cases. A novel damping peak was observed in between those corresponding to the Tg of the two constituents. The presence of this peak was almost independent from a variety of parameters, such as molecular orientation, composition, thermal history, thermal cycling, number and thickness of the layers. The peak disappeared only when the planar structure of the layers was disrupted. Its origin was attributed to a particular temperature dependence of the viscoelastic parameters in the layer composite in the appropriate temperature regimes. The same authors investigated on the deformation behaviour of two series of similar coextruded multilayered composites [20] by macroscopic tensile tests,

480 performed at different strain rates. Optical microscopy was used to correlate the microscopic mechanisms observed in the two phases, with the modes of deformation observed in bulk. Three kinds of modes were observed (see Fig. 3) within a single bulk composition: 1) curve A~ :a brittle fracture at low strains, without yielding; 2) curve A2 : a ductile yielding followed by a rupture during the neck formation; 3) curve B : a ductile yielding with a stable neck formation followed by cold-drawing and rupture at high strains. The final mode of failure depended on the relative thickness of PC and SAN layers, as determined by composition and strain rate.

A2 B

F,

Figure 3. Typical modes of fracture of PCA7AN blends, observed by Gregory et al. [20].

Optical microscopy revealed craze initiations in the SAN layers which induced successively shear bands in the PC layers at the craze tips. This interaction between crazes and shear bands hampered the crack propagation delaying the rupture by a stress delocalization. Thermal, mechanical and impact properties of blends containing PC and two different types of SAN (containing 5.5 and 30% in weight of AN), obtained by

481 injection moulding were analysed by Skochdopole et al. [21]. Parameters, such as strength, modulus, heat distortion temperature under load, showed a linear dependence with blend composition. Others, such as elongation at break and dart impact energy dropped from PC down to SAN level at a percentage of about 40% of SAN. Izod impact toughness followed the same trend but the drop occurred at a much lower SAN content (only 10%). The Tg analysis evidenced only a limited solubility between the components, with the blend, containing SAN with a higher AN content (30%), the more soluble in PC. Berger et al. [22] studied the microdeformation at constant strain rate of casted thin films (0.4 lam) by optical microscopy and TEM. The average tensile strain for void formation, ev, was 0.13 for SAN and 0.23 for PC. Both PC and SAN underwent shear yielding, ev decreased by the PC addition to SAN up to 60% in the blend and the voids were initiated by crazes at the PC-SAN boundaries, ev decreased alter an annealing, performed at about 90~

and the shear yielding

mechanisms were suppressed, favouring crazes formation in SAN and in SAN-rich phases. A model, was proposed to interpret the phenomenon. Kim and Bums [23] examined the miscibility of PC and SAN by determining, on solution casted and extruded samples, Tg and ACp (Cp difference between rubbery and glassy state) of blends,. The Tg of PC decreased with increasing SAN content and viceversa that of SAN was enhanced with augmenting PC amount. ACp of both PC and SAN decreased with increasing amount of the second component. The overall behaviour was attributed to a partial dissolution of each component in the conjugate phase. From the experimental data of Tg and ACp the authors calculated the apparent weight fractions of the two components dissolved in the PC-rich and in SAN-rich phases. This was accomplished by the Couchman relationship [24,25] used to describe the Tg dependence on composition in miscible blends. It appeared that SAN dissolved more in PC-rich phase than PC in SAN-

482 rich phase, confirming the previous finding of Mendelson [15]. The polymerpolymer interaction parameter was calculated by the Flory-Huggings theory [26]. The extrudate swell exhibited a maximum and the viscosity a minimum at a weight blend ratio of 50/50, showing a positive and a negative deviation from a linear additivity rule respectively. A reasonable explanation was given for this double effect: when one of the fluids is the matrix, it flows mostly along the tube wall, dissipating more energy than the particles of the dispersed phase. These, on the other hand, being preferentiaY~,~transported along the centre of the tube, are elongated and store more elastic energy than the matrix. When the two phases are co-continuous a reciprocal lubricating effect and their deformation are maxima, giving rise to a minimum viscosity and a maximum die swell in the middle composition range. Guest and Daily [27] analysed the thermal behaviour of PC/SAN blends. The SAN was also dissolved in a suitable solvent and reprecipitated in order to remove from it all the low molecular weight species, that are generally present in styrenic

Table 2. Glass transition data for PC, SAN copolymers and (70/30) PC~SAN blends (after Guest et al. [27]). i

Sample

i

Tgpc,

ATpc

(of) PC

155

(oc) 5.0

SAN* Blend(70/30)*

152

8.0

SAN** Blend(70/30)**

155 * as received

\

TgSAN, ATsAN

6.0

108.5

6.75

113.5

7.0

114.5

5.5

114.5

5.5

** reprecipitated

483 polymers in amounts, ranging between 1 and 2 wt % [28]. A comparison was made between the Tg and the half-width of the glass transition, ATg, of PC and SAN, between "as received" and "reprecipitated" SAN and the corresponding blend values (see Table 2). As it is possible to note, only the "as received" materials show classical inward Tg shitts and ATg changes of blended PC and SAN with respect to the pure component values, whereas in the "reprecipitated" materials such variations are almost completely suppressed. Hence this phenomenon is clearly due to the migration of low molecular weight species, such as monomers and oligomers, contained in commercial SAN, towards PC domains. Also the viscosi~, ratio of the two components can be influenced by the presence of these plasticizers, determining in turn somewhat the blend morphology. In a successive work the same authors [29] showed the ability of dynamicmechanical spectra in revealing morphological features of PC/SAN blend obtained by different processing conditions, such as compression and injection moulding. The injection moulded samples exhibited, in fact, a Tg (G") peak, broader than that corresponding to compression moulded specimens, having the same composition. In other words, they showed somewhat an apparent greater degree of mixing. This effect was related to the mterfacial regions obtained by processing. A parallel analysis by SEM showed a coarse uniform structure throughout compression moulded samples. By comparison, injection moulded samples exhibited a similar morphology in the centre and a stratified, fine, laminar and oriented structure in the skin regions. This kind of morphology seems to resemble that of multilayer sheets studied by Gregory et al. [19, 20] , indicafng tlaat, also in this case, the layers continuity plays the major role. Compression and injection moulded specimens of PC/SAN blends were analysed by dynamic-mechanical torsional tests by Mclaughin [30] as well. The two kinds of specimens showed different morphological microstructures: a) for

484 compression moulding specimens, SAN spherical particles in a PC matrix at high PC/SAN ratios, or PC spherical particles in a SAN matrix at low PC/SAN ratios throughout the sample, viceversa; the onset of co-continuity occurred at a PC/SAN ratio, ranging between 50/50 and 60/40; b) for injection moulding specimens, spheroidal particles in the core and lamellae in the skin, at very high or very low PC/SAN ratios, and sheets all over the sample in an extended range between 50/50 and 80/20. An intermediate peak between the two Tg became distinguishable at high frequencies in plots of tan ~5versus temperature in the latter case. This finding was related to the co-continuity obtained in the injection moulded specimens, which could be idealised as alternating sheets, similar in structure and behaviour, to the previously mentioned multilayered composites analysed by Gregory et al. [ 19,20]. Huang et al. [31] , somewhat confirmed the previous suggestion [12] of an optimum compatibility between PC and SAN in blends with SAN, containing 25% in weight of AN, by means of theoretical calculations based on the solubility parameters of the two components. Quintens et al. [32,33] analysed the dependence of the viscoelastic properties from the morphology. They annealed, at a temperature higher than the Tg of PC (200~

two injection moulded samples of PC/SAN blends, of different

composition (70/30 and 60/40). A gradual coarsening of the microstructure was obtained with increasing the annealing time. As already described elsewhere [19,20,29,30], the storage modulus plateau increased in value in the temperature field ranging in between the two components Tg. Furthermore the coarsening induced a loss of ductility in tests at high deformations. These findings were attributed to the reduction of the PC/SAN interface, consequent to the induced morphological changes. The influence of SAN composition on tensile stress-strain behaviour and phase morphology of a 60/40 PC/SAN blends, containing SAN copolymers of different AN content (in the range 0-34 wt % of AN), was analysed by Quintens et

485 al. [34,35]. Maximum stress and elongation at fracture exhibited a maximum for blends with a SAN copolymer having 25% AN and the morphological dispersion was finest at this composition. A phase morphology coarsening, consequent to the thermal treatment described above, resulted in a loss of ductility but the maximum was not shifted, confirming previous fmdings [12,31]. Also in this case annealings at high temperatures destroyed SAN layers continuity produced during the injection moulding process. This occurred by the break up and coalescence of the elongated SAN particles, influencing the viscoelastic behaviour in between the Tg of the two components. The rate of the process was determined by the viscosity ratio of the two components at the temperature of annealing. The morphological trend, as a function of annealing time, could be monitored by the decreasing contribution of the SAN phase to the overall viscoelastic behaviour of the blend. Takahashi et al. investigated the rheological behaviour of several polymers (HDPE, PP, PS, Ny 66, PC, SAN) [36] and that of PC/SAN blends [37] at very high shear rates. For homopolymers the generalised flow curve, consisted of two non-Newtonian regions separated by a transition or by a Newtonian zone. In the first non-Newtonian zone the viscosity decrease was attributed to a lowering of entanglement density. In the second one the macromolecular chains underwent to a scission mechanism. Then, in spite of the molecular weight decrease, the chains were still able to reentangle with each other, giving rise to a sudden viscosity increase. Finally the viscosity decreased again at higher strain rates, due to a double effect: a) a new disentanglement process of the previously reentangled chains; b) a further chains snapping mechanism. For PC/SAN blends the trend was rather similar to that of the two components and a cylindrical multilayer model was proposed to fit the data. The effect of the high shear rate extrusion on the component compatibilization of a (70/30) PC/SAN blend was investigated by Takahashi et al. [38], as well. The apparent volume of the SAN particles dispersed in the PC matrix, decreased with

486 increasing shear rate, as shown by SEM observations,. Moreover repeated extrusions decreased this volume more and more and a shit~ of the Tg of PC was detected. This indicated that SAN could be partially dissolved into the PC matrix by means of an intense mechanical sheafing. Alternatively free radicals could be produced by the chains scissions. These could favourite the formation of interfacial agents between PC and SAN, helping the PC and SAN compatibilization. The emulsifying effect of the PC-g-SAN copolymers, formed in this way, would enhance the adhesion between PC and SAN, reduce the SAN domain sizes and create effective interlayer zones. Shah et al. [39] analysed ternary blends, formed by PC, one of two SAN (containing 13 % and 25 % of AN) and one of three aliphatic polyesters. Each of the three polyesters was capable to promote in binary blends a single Tg behaviour when mixed with SAN (25 % AN), SAN (13 % AN) or PC. All three of them were able to render miscible normally immiscible binary PC/SAN (25 % AN) or PC/SAN (13 % AN) blends. Each ternary blend showed, however, SAN-rich and PC-rich regions where immiscibility occurred. SAN (25% AN) was more easily solubilized in PC than SAN (13 % AN) by the best of the three polyesters. This indicates that the AN content plays an important role also in this phenomenon, confirming what previously found for binary PC/SAN blends [12,31,34,35]. Im et al. [40] presented a review on physicomechanical properties, such as tensile, impact and fatigue of PC/SAN microlayer composites. Particular attention was paid to irreversible deformations and damage mechanisms as well as to ductility improvements of the brittle components. Elastic moduli of compression and injection moulded PC/SAN blend specimens were measured at temperatures between the PC and SAN Tg by Arends [41]. In the intermediate blend composition range their values differed by about a factor of three, probably due to different modes of connectivity among the particles

487 of the nmlor component. The importance of phase continuity was evidenced for compression and moulding specimens by modulus data versus SAN composition. The data agreed quite well the behaviour predicted by an empirical relation based on the percolation approach in finite domains, devised from Monte Carlo simulations. The interaction between PC and SAN at a given AN content (about 25% by weight) was analysed again by Callaghan et al. [42] in regards to the molecular characteristics of PC and SAN. It was shown that the ~ o components become entirely miscible in blends when the molecular weight of each component is lower than about three thousand. The interaction energy was calculated by equation of state theories, developed by Flory-Huggms [26,43,44] and Sanchez-Lacombe [4550], allowing predictions of interfacial tensions.

3.5

I

'1

I

I

I

I

I

3.0 2.5

E

::L E~

J

2.0 1.5 1.0 0.5 t

I

I

I

I

I

I

5

10

15

20

25

30

35

40

AN (1~1 in SAN

Figure 4. Average diameter of dispersed SAN particles versus AN content in SAN in PC~SAN blends (after Callaghan et al. [42]).

488 The data are reported in Fig. 4 as particle average diameter, d~, versus AN wt % in SAN:

a minimum is observed at about 25 % of AN, consistent with

morphological observations made by Quintens et al. [34]. The interfacial tension appears to be directly proportional to the average SAN particle size in a PC matrix, provided that the components viscosity and the shear field are kept fixed, as shown also in some other case by Takeda et al. [51]. PC blends with commercial SAN purified from oligomeric species exhibited almost no measurable shifts in the Tg of either phase, as already found by Guest et al. [27], supporting their view of SAN oligomers partitioning between the two phases rather than the hypothesis of the partial miscibility made by several authors [12,15,21,23,31]. A series of SAN with different AN contents were selected in a way that their melt viscosity was maintained nearly constant and blended with a PC of a similar viscosity. In this conditions the morphology should be independent of rheological factors and should reflect only the interfacial tension. Mechanical properties of ternary PC/SAN/MBS (methacrylate-butadienestyrene copolymer) blends were analysed by Cheng et al. [52]. In particular the impact strength of a series of binary PC/SAN and ternary blends, containing SAN with different AN contents, as a function of this parameter, exhibited a maximum at about 25% of AN, as shown in Fig. 5. The blends were obtained by a simultaneous mixing procedure. Considerations of surface energy and component pair miscibility were successfully used to predict the MBS particles location. Most of the MBS impact modifier particles were found to be located at the PC/SAN interface, being trapped by surface forces. The mixing procedure as well as mechanical and impact deformation mechanisms were important, however, in determining this location in some cases.

489 500

- .9 E

I

I

I

I

I

I

400

--3

..c

300

C"

k--

"-'

200

E

100 ( _

PC/SAN

( 0

0

j

J

J

I,

J

5

10

15

2O

25

I

30

35

AN (~)in SAN Figure 5. Impact strength of SAN-based materials as a funcaon of AN percentage m SAN for PC/SAN/MBS and PC/SAN blends, as indicated (after Cheng et aL [52]).

The interfacial tension between PC and a series of SAN copolymers having AN contents from 0 to 40 % in weight were determined by measuring capillary thread ruptures at 200~ by Watkins et al. [53]. The interfacial tension plotted versus AN % exhibited a minimum at about 15 wt % of AN, a value lower of that previously reported [12,31,34,35,39,42,52] for the optimum value of miscibility between PC and SAN (about 25 wt %). When the interfacial tension was at a minimum and PC and SAN viscosities were similar, very thin threads could last for several minutes before rupture. This could be the reason of the formation of filaments and/or of stratified lamellae in the skin of injection moulded samples under the action of both high, shear and extensional, flow fields.

490 Inward Tg shifts of PC and SAN were detected on PC/SAN blends and the Flory-Huggins parameters were calculated by Kolarik et al. [54]. The compositions of the conjugate phases were calculated by the Fox's equation. The apparent solubility of SAN in PC or PC in SAN increased with decreasing PC or SAN volume fraction in blends respectively. These dependencies were tentatively ascribed to the presence of intermixed zones existing at the boundaries between PC and SAN domains. The yield stress followed a classical rule of mixing, indicating that the phase adhesion was sufficiently good to provide effective stress transfer between the phase domains during yielding and cold drawing. Janarthanan et al. [55] confirmed the existence of a maximum in a plot of interfacial fracture toughness versus AN % in SAN for PC/SAN blends at about 24 wt.% of AN in SAN, as previously found by Keitz et al. [12].

20 16

G

12 r

{J/m z)

AN '

I

'

2

I

4

'

I

6

.

I

I

8

Wt ~ of benzonitrile

Figure 6. Fracture toughness of PC~SAN interface, Gc, versus benzonitrile content for two different values of AN % (after Janarthanan et al. [55]).

491 They showed in addition, by using benzonitrile as a model oligomer, that low MW SAN species preferentially migrate toward the PC/SAN interface. This increases the PC/SAN interphase thickness reducing the entanglements between PC and SAN macromolecules. In fig. 6 the fracture toughness of adhesion between PC and SAN is reported against the amount of benzonitrile for two blends containing SAN with a different AN wt %. Both the curves undergo to a monotonic decrease, starting from different initial values depending on the AN % in SAN. All the literature results, above described, can be briefly summarized as follows: a) PC and SAN are completely immiscible in blends, both in melt [23] as well as in solid state [ 12,41 ]. b) Their blends show inward shifts of blend Tg values with respect to homopolymers ones. These Tg shifts have been often interpreted by several authors as due to partial miscibility [12,15,21,23,31,54] of the components. The real cause is due, indeed, to migrations of SAN low M.W. species towards PC domains during the melt-mixing [27,42,55]. c) There is a narrow interval of SAN composition (around 25 wt % of AN in SAN), where a variety of overall properties and particularly, the adhesion between PC and SAN, exhibit maxima [12,31,34,35,39,42,52,53,55]. This behaviour has been explained by a binary interaction model [12,13]. The interfacial tension between PC and SAN shows a minimum in the same AN composition range in SAN [12,42,52,53,55]. d) Low M.W. SAN species tend to migrate from SAN phase towards PC and enrich the PC/SAN interface, diluting the macromolecular concentration in the interzones. The net result is a decrease of the number of molecular entanglements between PC and SAN in the interzones at the PC/SAN boundaries. This lowered interconnectivity decreases, in turn, the adhesion between PC and SAN phases I55].

492 e) Different blend processings (compression and injection moulding), can give diverse morphological features, which can be monitored by dynamic-mechanical tests, as shown by several authors [19,20,29,30,32,33,41]. The storage modulus (G') versus T shows measurable changes (a new peak or a shoulder) in between the Tg of the two components. This suggests that PC and SAN changes of Tg in blends depended not only on the thermodynamics of the interface but also on the final morphology of the systems (changing, for instance, the surface to volume ratio). Therefore any external change, such as thermal treatments, yielding variations in the morphological characteristics of the systems, can be easily detected by this technique.

3.2 PC/ABS blends

The behaviour of a 50/50 commercial blend by nuclear magnetic resonance (NMR) in a range of temperature from 100 ~ up to 500 ~

by SEM and TEM

was analysed by Stefan et al. [56]. The transitions associated with PC, SAN, and PB were monitored. It was found that, at temperatures close to the Tg of PC, the chain motions of the zones surrounding the PC were strongly influenced by the PC itself. Therefore the PC apparent volume (interaction zone) increased of about 50% in comparison with its actual value. At temperatures, close to the Tg of PB, the rubber affected the surrounding regions even at a higher extent, so that the PB apparent volume increased twice its actual volume. The larger interaction of the rubber was attributed to the diverse morphology (small round particles), as well as to the higher flexibility of the PB chains. The improved blend impact performances with respect to pure PC was related to the apparent volume of the rubber phase. These findings suggests the existence if interacting regions between PC and ABS phases. The melt viscosity of PC/ABS blends as a function of composition, measured by Dobrescu et al. [57], is shown in Fig. 7. The viscosity goes through a minimum,

493 both at low (y - 1 sec -1) and high shear rate (y = 103 sec-1). The depth of the minimum is the larger the lower the shear rate. The composition value at which the minimum occurs seems to depend on the shear rate value. The trend of the curve is due the PC/ABS immiscibility in the melt. This effect induces a reciprocal lubricating effect of the two components during the mixing, lowering the overall internal friction of the material.

8000

k,~

"

I

'

I

'

!

'

I

'

I

'

'/

'

I

"'

i

''

I

' ........

CtJ

a.

0~

800

0 o

80

0

10

20

30

40

50

60

70

80

90

100

Wt ~; ABS or PMMA

Figure 7. Viscosity as a function of ABS or PMMA percentage for PC/ABS (full circles) and for PC/PMMA (empty circles) at two different shear rates as indicated (after Dobrescu et aL [57]).

In the same figure the curves relative to PC/PMMA blends are reported by comparison as well. They present no minima but a smooth decreasing trend,

494 particularly at high shear rates. The different behaviour of PC/PMMA blends with respect to PC/ABS ones depends on the PC and PMMA miscibility. Kim and Bums [23] found that the viscosity of PC/ABS blends, measured in a capillary rheometer, went through a minimum at a weight blend ratio of 50/50, as found by Dobrescu et al. [57] (see Fig. 7). In Fig. 8, the extrudate swell ratio is reported as a function of SAN, ABS or Kodar

contents.

The

last

one

is

a

copolyester

formed

from

1.4-

cyclohexanedimethanol and a mixture of terephtalic and isophtalic acids, which is known to be compatible with PC.

1.8 9"" - ' ~ " "" "-=.,.,~j~ P C / ABS 0

1.6

L_

~ ~5

1.4

1.2

1

0

0.2

0.4

0.6

0.8

1

Wt ~ SAN, ABS or Kodar

Figure 8. Die swell ratio as a function of SAN, ABS or Kodar percentage (after Kim et al. [23]).

495 The comparison between the three kinds of blends shows a very pronounced maximum for both PC/SAN and PC/ABS blends, at about 50 % of PC content. The third component exhibit, instead, a very fiat curve. These findings clearly indicates immiscibility in melt between PC and SAN or ABS and compatibility between PC and Kodar. The reasons of this behaviour have been already illustrated above for PC/SAN blends and they do not change in the case of PC/ABS ones. Both the Tg of PC and ABS of PC/ABS blends were measured by DSC. That of PC decreased almost linearly with increasing the ABS content. That of ABS, on the other hand, increased linearly with increasing the PC content (in an analogous manner as described for PC/SAN blends).

150

"i'

I

'

I

'

I

'

140

(}.. ~"

ABS " -9 0

130

i-

120

_.._~_.~ABS

110

......... 100

II~|

0

~

I

10

~

~ I

20

,,

i

30

S ,

i

40

,

A i

50

9

N i

60

,

i

70

,

i

80

9

i

90

100

PC (%}

Figure 9. Comparison between Tg of PC (circles) and of SAN or ABS (squares)" for PC/SAN blends (filled symbols) and for PC/ABS blends (empty symbols) (after Kim et al. [23]).

496 The ATg value of PC was greater than that of SAN at equal composition values, whereas the ATg of SAN and ABS followed a parallel trend, as shown in Fig. 9. This effect was attributed to the polybutadiene chains acting as an additional plasticizer for PC. The polymer-polymer interaction parameter was calculated also for PC/ABS blends by the Flory-Huggins theory, showing very close values to those calculated for PS/SAN blends. Cooney [58] tested the photostability of films of a commercial PC/ABS blend, found to be rather sensitive to UV and visible radiations. Particularly the polybutadiene on the specimen surface was easily oxidised and crosslinked. This effect embrittled the surface, causing superficial cracks on bending, which lowered the material impact strength. Suarez et al. [59] found that, for extruded sheets as well as for injection moulded bars, the modulus and the tensile yield strength were nearly additive whereas the elongation at break showed a minimum as a function of the composition. The notched Izod impact strength was much lower for ABS than for PC. It levelled off at the ABS value up to 50% of PC content and then increased almost linearly up to the PC value. Weber et al. [60] studied the morphology of PC/ABS blends in relation to the PC/SAN composition (keeping constant the amount of the SAN-g-B copolymer) and to the melt processing temperature. The Vicat temperature slightly depended on the melt temperature and increased with increasing the PC content. The morphology and the impact resistance were strongly influenced not only by the blend composition but also by the melt temperature. An enhancement of latter parameter caused, indeed, a demixing of PC and SAN and a worsening of the impact performances. However, when PC was the matrix this effect became negligible.

497 The influence of reprocessing PC/ABS blends on their physical properties was studied by Eguiazabal et al. [61]. Two processing cycles affected only slightly the properties. For a higher cycle number a change of the rubbery phase (cross-linking and oxidation) was observed. Therefore density, MFI, stress and elongation at break showed drastic variations after the first two cycles, whereas the small deformation properties were almost unaltered by reprocessmg. The toughening mechanisms of these blends were analysed by Ishikawa et al. [62] by three point bending of round notched bars. The addition of small ABS amounts (2%) decreased the PC toughness whereas contents from 5% up to 20% yielded larger and larger improvements. A stress whitening effect was attributed to formation of voids due to the fracture of the interphase existing between PC and ABS. Radusch et al. [63], studied dynamic-mechanical and dielectric properties, suggesting the existence of a partial compatibility between PC and ABS in the boundary layers of the order of magnitude of 1 nm. The macroscopic properties were, therefore, determined by the interactions among the phases in this layers. An optimal concentration for toughened behaviour was about 70 % PC. Lee et al. analysed the solid-state morphology of an injection moulded PC/ABS (90/10) blend [64]. A bead-and-string structure was observed in the skin regions of the plaques, whereas in the middle an isotropic ABS phase was dispersed in the PC matrix. In a subsequent paper [65] the analysis was extended to the entire blend composition. Three composition ranges were identified: a) a PCrich blend, where the situation was that described above; b) a mid range (between PC/ABS 70/30 and 60/40), where the previous structure evolved to a coalesced stratified morphology at the edges and to a coarse dispersed ABS phase in the centre, with some regions of co-continuity with PC; c) an ABS-rich blend with dispersed PC domains. Qualitatively the above results were explained as due to the

498 melt flow pattems during the mould filling and to relaxations and coalescence acting during the cooling of the materials prior to the complete solidification. Chun et al. [66] found a synergistic effect in notched Izod impact strength extended in a broad composition range (PC/ABS from 80/20 up to 10/90), wider than in other authors' findings. This was attributed the PC/SAN miscibility, to a suitable ABS composition (low B content) and to an improved mixing device. Triphenyl phosphate and a brominate phosphate were compared as flame retardant additives for PC/ABS blends by Green [67]. The latter was shown to be more efficient as flame retardant and to give higher heat distortion temperatures than the former. Real time small-angle X-ray scattering was performed by Bubeck et al. [68] on a series of rubber-modified thermoplastics in order to investigate the modes of deformation in tensile impact tests in such materials. In particular for PC/ABS blends the predominant mechanisms were shear yielding in the PC and associated rubber particle cavitation in the ABS. The cavitation mode seemed to provide a direct relationship between rubber content and impact toughness. The mechanism did not change whether the tensile impact direction was perpendicular or parallel to that of the injection-moulding. At last crazing, though precursor of the final fracture, occurred after the prevailing non-crazing mechanisms, contributing only for a few percent to the total plastic deformation. Suyun [69] found that the fracture morphology of PC/ABS blends with high notched impact strength was a synergistic combination of the rupture characteristics of PC (smooth surfaces and striation with branches) and ABS (microcavities and parabolic markings). The fracture morphology of blends with low impact strengths was very different from that of the two components. Both fracture morphology and impact strength were dependent on PC and ABS characteristics as well as on blend composition.

499 The viscosity-composition relationship of PC/ABS systems was analysed by Kumar et al. [70] by using the Lecyar model. This model is a simplified version of the more accurate Mc Allister's one. The experimental data showed a strong negative deviation from the linear trend with respect to the homopolymer values, as found by other authors [23,57,58]. A good fit of the data with the model was obtained. An accurate fractographic stud), was performed by Lee et al. [71] on PC/ABS single notched specimens obtained by injection moulding and fractured in tensile mode in an Instron Machine at a relatively low strain rate (48 mm/sec). Stress whitening was observed on the surfaces of ABS and of all the blends but PC, indicating voiding formation during the deformation and fracture. Plane stress flow lines were observed for PC, PC/ABS 90/10 and 70/30, accompanied by a lateral contraction of the overall sample. Also the 50/50 blend exhibited some features of plato stress at the specimen's edges. In the centre, however, a valley on one surface and a corresponding ridge on the conjugate surface evidenced a plain strata region. The resulting mechanism was therefore a sort of mixed fracture mode. No suckingin (overall contraction) was observed on 30/70 and 10/90 PC/ABS blends. A characteristic feature (called herringbone or chevron fracture) was however still visible in the centre of the specimen with narrow shear lips. Also this kind of failure mechanism was considered to be of a ductile nature. The ABS surface showed no sucking and no shear lips at the edge, indicating a macroscopic condition of plain strain. From the above observations a progression in the fracture mode with varying the composition was described. There was, in fact, a gradual change from a shear fracture of PC under plain stress conditions, to a craze failure of ABS, under plain strain conditions. The mare influence of the ABS addition to a PC matrix was the cavitational mechanism of the rubber particles. The plain strain observed in the central region of the specimens in intermediate compositions, was due to the increase of the ABS content, while the shear lips at the edges,

500 characteristic of the plain stress conditions became narrower, completely disappearing for pure ABS. An S-shaped curve was observed between the ductileto-brittle transition temperature

and the

composition.

The

most

ductile

compositions were PC/ABS 70/30 and 60/40 whereas the most brittle were 30/70 and 10/90. In a subsequent paper Lee et al. [72], performed a fractographic and a morphological analysis on an injection moulded specimen of PC/ABS composition 70/30. Fracture occurring perpendicularly to the injection direction yielded a herringbone feature in the surface. The one occurring parallel to it yielded an inverse herringbone feature. The blend toughness decreased in the second case. The herringbone was determined by interactions of the main crack with secondary cracks started along the centre line. The inverse herringbone had the same origin but the secondary cracks were initiated near the edges. Such differences in behaviour between perpendicular and parallel directions were attributed to the PC orientation induced by the shear flow during the processing. Vibration welding was applied by Stokes et al. [73] to ABS blended with other thermoplastics, among which PC. It was found that the weld strength of ABS, coupled with PC, was 92% of the value obtained by ABS coupled with itself. Moreover the fracture in tensile mode, performed in direction perpendicular to the weld surface, let~ on the PC halves several ABS islands. Both the findings appeared to be a further evidence of the good adhesion existing between the two resins. A systematic work on PC/ABS blends was initiated a few years ago by Greco et al., whose preliminary data have been published in several papers [74-77]. In a first work [74] two complementary techniques, suitable to selectively etch one of the two components, were presented. The aim was to get a phase contrast suitable for accurate

scanning electron microscopy

(SEM)

observations,

501 particularly in the composition ranges were one of the component is the matrix and the other is dispersed in it. In this case a cross-check is a convenient method. A unique etching technique would give, in fact, good information only for compositions where the phase to be etched is the dispersed one. In the opposite case, the result would be rather uncertain due to loss of the remaining dispersed particles from the etched substrate (matrix).

Figure 10. SEM micrographs of smoothed surfaces ~f a 70/30 (PC/ABS) blend: a) PC matrix etched by a NaOH aqueous solution; b) ABS dispersed particles etched by an acid aqueous solution (after Greco et al. [74])

502 In figs. 10 morphological features evidenced by two etching techniques are shown for a 70/30 (PC/ABS) blend, first etched and then coated with an Au/Pd alloy fihn, for scanning electron microscopy observations. In Fig. 10a the PC matrix has been completely etched by a PC hydrolysis, induced by a NaOH aqueous solution (30% w/v of NaOH), leaving on the surface ABS dispersed particles. Their surfaces are not completely smooth, as one would expect from immiscible components, but several small holes are present on them, due to complementary PC etched protrusions. This indicates a very extended interfacial contact area existing between PC and ABS phases. In Fig. 10b the complementary etching has le~ unaltered the PC matrix and therefore, the numerous holes, visible on the surface, represent the sites where the

Figure 11. SEM micrographs of smoothed surfaces of a 40/60 (PC/ABS), cut along two orthogonal directions, etched by an acid solution (after Greco et al. [74]).

503 ABS particles were sitting before their etching out from the matrix, by a strong oxidant acid solution. In Fig. 11 SEM micrograph of two sections, perpendicular to each other, of the same 40/60 (PC/ABS) blend are shown: the etching, made by the acid solution, reveals the shape of cylindrical PC domains in the ABS matrix. From a complete set of micrographs, encompassing all the blend composition, it was possible to establish at which composition a phase inversion between PC and ABS occurred. In a second paper [75] a brief literature review on PC/SAN and PC/ABS properties was presented, of which this paper represents a more extended version. 100

.

.

,

.

,

,

7O 2~ 'I--

50 4O 30

I-

20

]--

"~"~"~'..._

"-/

16

o"'~-~-"'-"-.----.~

_

93 2

10

"-64 0

20

40

60

80

100

ABS (%1

Figure 12. Ratio of mixing torque, T, to roller speed of rotation, n, (T/n), as a function of PC % and n values in PC/ABS(M) blends' (after Greco et al. [761).

The ratio of a Brabender-like torque (T) to roller speed (n), as a function of

504 ABS content, decreases with the ABS addition from the PC value at all n (at low values the curves exhibit a slight decrease with a series of intermediate peaks). For high roller speeds (n - 32-64 rpm) the curves, alter an initial sharp decrease, levels off on the final ABS value at a blend composition of about 40 % of ABS.

160

'

I

'

l

'

I

'

I

i

PC

140

1:3

~O o w

I-120 B,,....~

[] rl

ABS

100

0

20

---O.~

' - " ~ - ' - "---fi--- - .~

40

60

[]

80

~---,

100

ABS (%) Figure 13. Tg of PC and ABSfrom different sources: full circles, (Kim et al. [23]); empty squares, (Greco et al. [76]).

Processability, thermal, mechanical and impact behaviour were analysed [76] successively on the same PC/ABS(M) blend utilized in the first paper [74]. Improved PC processability by ABS addition was confirmed, as shown in Fig. 9. Inward Tg shifts, as found by other authors, both for PC/SAN [12,15,16, 21,23,27,42,54] and PC/ABS [23,60] systems, were detected by DSC. The data (squares) are shown in Fig. 13, compared with those of Kim et al. [23] (circles).

505 The PC Tg lie very close; the ABS ones follow a parallel trend. This behaviour can be due to a similar ABS internal concentration of the two types of blends. Stress-strain curves of PC, ABS and PC/ABS blends, relative to tensile tests made at R.T. on unnotched specimens, at a low deformation rate (0.1 mm/min), are shown in Fig. 14. Going from pure PC and adding more and more ABS the yield peaks lower and broaden down to a PC content of about 50 %; below this value the curve shapes resemble those of pure ABS, indicating that a phase inversion occurs, as confirmed by morphological observations as well [74]. The yield peaks and the elongation at break lowering, with decreasing the PC content in the blend, are evidences of a decreased blend ductility. 60

50

0

PC....,

. . . . . .

PC-8O

40 ~_~0-20 ~: "--" IO

30

\

~ AB S

PC-lO

PC-50 PC-40

2O 10 0

=

0

Figure 14.

I

10

i

I

20

,

I

30

=

I

40

=

I

50

Stress-strain curves of PC, ABS and PC/ABS(M) blends as

indicated (after Greco et al. [76]).

The application of the Kemer's model [78-80] to the experimental data of the Young modulus showed a very good adhesion between PC and ABS domains all over the composition range. The experimental points (empty circles) lie, in

506 fact, along the lines of perfect adhesion of the Kemer's model, as shown in Fig. 15,

1.4

/

'

'

'

'

/

L 1.2

'

/

Perfect

/

o

'

'

'

'

/ /

Adhesion

_1

o

1.0 ---~ LU

o Adhesion 0.8

0.6

0.4

0

0.2

0.4 ABS

volume

0.6

0.8

1

fraction

Figure 15. Curves of the Kerner's models for perfect and no adhesion between the phases, compared with experimental points (empty circles) (after Greco et al. [76]).

where the model curves of no (very poor) adhesion are reported as well. A synergistic effect was observed for Charpy impact strength and for maximum impact stress at about 25 wt. % of ABS, as shown in Fig. 14. A series of transitions from plain-strain to plain-stress conditions were observed at the different compositions, as clearly shown in Fig. 16. Starting from pure PC a brittle fracture (B) was observed for PC amounts ranging from 100 up to 80 wt %,

507 followed by a very ductile zone (D) between 80 and 50 % of PC (a strong synergistic effect is represented by two peaks in both E and ~ at about 75% of PC content).

12

,

11

- B

10

-

i

,

i

,

i

D

,

i, !

B

: ! !

,

]

SD

9 8 OL.

7

E

-

b 5

'

V

~

0

~

'

I

1

~l

0

.,--"J~'~----"~

ILl

!

I

20

~

I

t

40

I

60

t

I!

80

100

ABS(M)

l~gure 16. Max. stress (left hand-side axis) and energy to crack initiation (right hand-side axis) versus PC content. Plot impact behaviour zones: D, ductile; SD, semiductile; B, brittle (after Greco et al. [76]).

With a further increase of the ABS content (PC content between 50 and 20%) again the behaviour becomes brittle (B) and finally the ABS semiductilc (SD) fracture is approached. This alternation of impact mechanisms is directly detcctable by a parallel fractographic analysis. In Figs. 17 a few micrographs show fracture surfaces of specimens representative of the four composition zones, exhibiting an alternation of impact behaviour. The first one (a) shows classical features of a PC surface fracture, with

508

Figure 17. SF)~/Imicrographs ~ (a), PC; (b), 75/25 and (c), 50/50, PC/ABS blends ; (d) ABS (after Greco el al. [76]).

509 a relatively slow crack initiation region, where ribs are visible, and a successive zone of very high propagation rate, resulting in a smooth region. In this case plainstrain conditions can be judged from the inspection of the overall specimen shape as well, exhibiting no lateral contraction. Next surface (b), relative to a blend, containing 75 % of PC, corresponds to the g and E peaks (zone D) in Fig. 16: patterns of flow lines and a marked lateral contraction of the specimen are a clear evidence of plain-stress conditions. Next surface (c), of a rough appearance, represents a blend, made of 50 % of PC, where again the breakage occurs under plain-strain conditions. The last one (d) is relative to the semiductile behaviour of pure ABS, where slight flow lines reveal that a certain amount of material deformation occurs all over the surface prior to the crack opening. A comparison of the above illustrated results with those obtained using the same PC blended with a second ABS(B) in the PC/ABS blends was presented in the last paper of the series [77].

Table 2. ,Some characteristics of three ABS used in PC/ABS blend~ [77, 81]. i

TRADE

CODE

AN %

B%

S%

NAME Sinkral

MFI (ASTM D-1238)

B32

B

27

22

51

4

Sinkral M 122

M

23

14

63

18

Sinkral

A

27

11.5

61.5

26

A12

In the present paper processability, thermal and impact behaviour of a third ABS(A) has been added to the comparison. Extended trade name, code, internal composition, mclt flow index (MFI) calculated according ASTM D-1238, of the

510 three commercial ABS, manufactured by Enichem Inc., with the common tradename of Sinkral, are reported in Table 3. The main differences of the used ABS consists in the B content, decreasing from B (22%) to M (14%) to A (11.5 %), as well as in the MFI, increasing in the same order. The two ABS, M and A, are rather similar in composition being slightly different only in the AN content and in the MFI. The values of all three, however, lie very close to the AN range of maximum adhesion between PC and SAN in PC/SAN blends [12,31,34,35,39,52,53, 55].

500

I

'

'

'

I

'

'

'

'I

'

'

'

I

,

w

'

'I

_

_

Q_

c

50 tI--

ABS{A)

BS(B) -

;

{M)

I

I

I

,

I

1

,

I

I

I

i

lO

I

I

I

I

I

,

,

-

i

I

10o

n (s - 1 )

Figure 18. T/n as a function of n for PC/ABS blends containing three ABS of different internal composition, M, B and A, as indicated in the figure and in Table 3 (after Greco et al. [77, 81]).

A more complete rheological characterization is provided in Fig. 18, where

511 the ratio of the Brabender torque per unit volume to the roller rotation speed, T/n, with the dimensions of an apparent viscosity (MPa-s), is reported as a function of n for PC and the three examined ABS. The PC has a less pseudoplastic behaviour than the three ABS, of which the ABS(B), having a higher B content, exhibits a higher internal friction (T/n has the same dimensions of an apparent viscosity, even though the complex pattem of the mixing process is completely different from those of common viscometers) than the other two ABS, during the treatment in a Brabender-like mixer. Moreover its curve crosses the PC one at an n value of 16 rpm. Therefore, at the rotation speed at which the blends were prepared (32 rpm), its "viscosity" is

45 40

=

2

rpm

35

ao ~

I---

-

25 20 -

15 10

-

M i

I

20

i

I

I

40

I

60

i

I

80

n

100

ABS (%)

Figure 19. Ratio of torque, T, to roller speed of rotation, n, (T/n), at n value of 32 rpm, as a function of PC %, for blends containing different ABS; codes as indicated in table 3 (after Greco et al. [77, 81]).

512 lower than that of PC. The other two ABS show very close values with respect to each other but much lower than the PC ones. These rheological features can have great influence in the processability of the blends and, of course on their final morphology and properties. As a reference to real conditions of the blend mixing, in Fig. 19, the curves of the same viscometric parameter, T/n, have been reported as a function of blend composition. As it is possible to see, all three ABS reduce the PC internal friction, greatly improving the PC processability when PC is the matrix of the system. The best effect is provided by the ABS(B), the most viscous among the three ABS probably because of its higher B content. The greater B amount would enhance the reciprocal lubricating effect necessary to lower the internal friction of the system. With increasing the ABS content a phase inversion occurs and, therefore, the properties of the ABS matrix become predominant for all the blends. For PC/ABS(M) and PC/ABS(A) the internal friction is very low at high ABS contents compared with the PC one at 32 rpm (mixing roller speed) and T/n levels off on the ABS value. Only the PC/ABS(B) exhibits a very pronounced minimum, since, when ABS(B) is matrix, T/n must gradually come up to the high ABS value, due to its large internal friction, comparable with that of PC. The thermal behaviour of the three types of blends is reported in Fig. 20, where the Tg of PC and ABS are plotted as a function of the blend composition. The most relevant feature if the higher ATg of PC, between the blend PC/ABS(B) and the PC homopolymer, with increasing the amount of ABS, in comparison with the two other types of blends, M and A. This effect somewhat confirms the marked contribution to Tg shitts provided by the B, previously illustrated in Fig. 7 [23], where the ATg of PC was larger for PC/ABS than for PC/SAN blends (containing no B).

513 A migration of low MW species of B (m addition to those of SAN species [27]) from the ABS domains towards those of PC along the PC/ABS boundaries was invoked by the authors when the SAN component was substituted by the ABS one. Also in our case it seems that the higher the B amount in the ABS, as it is for ABS(B) in regards to ABS(M) and ABS(A), the larger the contribution given by this component to the PC plasticization, in the PC/ABS interzones. 155

,

,

,

,

,

,

;

,

,

,

,

,

,

,"

,

""

145

M

0

135

A

Q

I--

125

A M

115

o

B 105

,

0

i

---L.___.____

9

0

9

!

20

i

,

I

I

40

,

I

60

Ii

,

i

!

80

,

,,

i

100

ABS % Figure 20. PC and ABS Tg versus ABS % for PC/ABS blends, containing different ABS, codes as indicated in table 3 (after Greco et al. [77, 81]).

The impact performance (as energy to crack initiation, E as a function of blend composition) of the three blends, tested in flexural Charpy mode on sharply notched specimens, are compared in Fig. 21. The results are similar for PC/ABS(M) and for PC/ABS(B), with some significant differences: a) the synergistic peak, present in both materials, has a different shape (sharp and tall for the former, broad and low for the latter); b) the

514 composition at which the first brittle-to-ductile transition occurs (see Fig. 14) is lower for ABS(B) than for ABS(M): only 10% of ABS is sufficient to toughen the PC in the first case, whereas in the second case it is necessary add at least 20 % ABS to reach similar results. This seems to be due the higher B content of the former. The third blend PC/ABS(A) shows no synergistic effect in all the blend composition, probably because of a different morphology. 10

8

&-"

6

uJ

4

2

0

0

20

40 ABS

60

80

100

1~}

Figure 21. Impact Charpy strength as a function of ABS content for different ABS, codes as indicated in Table 3 (after Greco et al. [77, 81]).

A more complete and accurate analysis on the effect of the ABS intemal composition will be presented in a forthcoming paper [81]. Other authors have analysed the influgnce of the ABS composition on PC/ABS blends, as well, even though in a non systematic way.

515 Kurauchi et al. [82] analysed two blends, of a commercial PC with two commercial ABS, having a different (AN/B/S) composition: a) 22/29/49 ; b) 20/37/43. Only the blend made with the first ABS (a) showed synergistic effects in tensile stress-strain parameters (energy absorption and elongation at break) and in unnotched impact Charpy strength, exhibiting maxima at high PC contents, as a function of composition. In the other blend, with ABS (b), having a greater rubber amount than (a), the corresponding parameters showed, instead, an almost linear trend at high PC contents and minima at low PC amounts. This effect could be due to the different B amounts of the two ABS(a, 29% and b, 37%). In the second case, in fact, the B could encapsulate most of the SAN, reducing the PC/SAN surface of contact. In addition, as a secondary effect, the AN content in b) is slightly lower than that in a), reducing in the first case the PC affinity with SAN. Morbitzer et al. [83] analysed thermal, dynamic-mechanical, stress-strain and notched impact behaviours of two PC/ABS blends, obtained according two different procedures: I) SAN and a SAN-g-B copolymer were preblended in a 60/40 ratio and then various amounts of PC were added to this blend; II) PC and SAN were preblended in a 50/50 ratio and then different amounts of a SAN-g-S copolymer were added to this blend. In both cases the Tg of PC, SAN and B, clearly detected by thermal and by dynamic-mechanical techniques, changed their values with varying the blend composition. The first two Tg varied as reported by other authors for PC/SAN blends [7,10,16,18,23,37,41 ]. The Tg of B decreased to lower temperatures with increasing the PC amount (decreasing the ABS and consequently the B amount). This effect was attributed to the thermal stresses caused during the blend cooling by the different thermal expansion coefficients of the grafted rubber particles and the surrounding SAN matrix [84]. The stress fields in the matrix, around the rubber particles, tended to overlap in dependence of the rubber content and of the final morphology. Since

516 these stresses were lower than the interfacial forces, the rubber particles underwent to a negative hydrostatic pressure. This induced an increase in free volume and a consequent decrease in Tg. The stress and the elongation at break, measured at room temperature, as well as the impact strength, measured at -20~

exhibited a maximum at a composition

of about 80 % of PC for blends of series I. Chiang and al. [85] analysed blends containing ABS of two diverse composition (AN/B/S): a) (22/19/59) and b) (20/32/48).The Brabender torque exhibited a monotonic decreased starting from the PC value with ABS addition; the PC processability improvement was larger for (b) than for (a). This difference in behaviour was attributed to a diverse morphology existing in the two systems. In other words the outer shell of the ABS particles was supposed to be made by AN in (a) and by B in (b), the latter yielding a worse compatibility with PC. This created a larger lubricating effect and hence a lower torque. Both the samples exhibited a broad minimum in the elongation at break at a different composition: 30% ABS for (b) and 50% ABS for (a). The lzod impact strength showed a pronounced maximum at a PC/ABS ratio of 90/10 for (b) and only a small one for (a). Both the effects were attributed, once again, to their different B content. Herpels and Mascia [86] used two different ABS with 20% and 30% of B in PC/ABS blends. The two kinds of blends were compared at equal rubber level in the final product. Small changes of Tg of the three phases were interpreted in terms of a partial compatibility between AN and PC. No correlation of the rubber blend content was found with fracture, obtained at low and high propagation speed. However a toughness increase was obtained under plain-strain conditions. The B was found to be encapsulated in the SAN matrix. Lombardo et al. [87], analysed over the full range of blend composition PC/ABS blends based on different types of ABS:

517 a) SAN (with no B) containing 25 % AN (SAN 25); b) an emulsion-made ABS, with 50% rubber content and very small uniform particle size (about 0.1 ~tm); c) a mass- made ABS, with 16% rubber content and large rubber particle sizes (0.5-1 ~tm). The sample a) exhibited lower modulus and tensile strength, but excellent performances in standard and sharp notched tests at R.T. as well as at low temperatures close to the Tg of the rubber. These results were attributed to the high rubber content but it was stressed that composition, rubber particle size and distribution could have also been partially responsible of the impact behaviour. The influence of the type of rubber particles was, therefore, separated from that of the rubber content. This was accomplished by making blends of SAN 25 with different ABS, in order to prepare PC/ABS blends with same rubber concentrations (5 %, 10 %, and 15 %). The results of the impact properties showed that small and uniform rubber particles (yielded by emulsion-made ABS) toughen PC/ABS materials at lower rubber concentrations and lower temperatures than large rubber particles (obtained by mass-made ABS). It was stressed, however, that also other factors, such as morphology, MW, grafting degree, agglomeration of particles, etc., can play some role in determining the above described behaviour. Wu et al. [88] studied the influence of MW of PC, in the range of M w going from 1.8x104 up to 3.6x105, on PC/ABS blend properties. They found that an increase of MW improves fracture toughness of blends at low temperature. Higher impact strengths, higher critical strain-energy release rate and lower brittle-ductile transition temperatures were, in fact, observed. However, as one would expect, the M.W. increase yields a too high melt viscosity, which renders difficult the blend processability.

518 Therefore a compromise must be reached in choosing the M.W. of PC, in order to take into account both the effects. With the highest M.W. used (3.6x105), the best compromise between impact properties and processability, relatively to the blend composition, was provided by the PC/ABS (65/35) blend.

4. Concluding remarks PC/SAN blend analysis and results can be partially utilized for PC/ABS blends, particularly with respect to component processability, miscibility and thermal behaviour. The rubber-SAN addition modifies mechanical behaviour and particularly improves, of course, the impact performances of the PC/ABS blends.

4.1 Processability A series of papers report on the lowering of the PC viscosity (or of the internal overall friction measured by the torque in mixing apparatus) by means of ABS addition, up to about 30-40% of ABS, indicating an improved PC processability [23,31,57,58,70,75,85,88]; this effect is detectable in PC/SAN blends as well [23,31] and is substantially due to a complete immiscibility of the two blend components.

4.2 Miscibility and thermal behaviour PC and ABS are completely immiscible both in melt as well as in the solid state. Several authors [23,60,63] have considered the two components partially miscible. This was made on the basis of the inward shifts of Tg detected for PC/ABS blend components (with reference to pure PC and ABS Tg values), as already done for PC/SAN systems. But for the latters this effect was clearly shown to be due to low M.W. SAN species migration towards PC boundary domains and not to partial miscibility of PC and SAN [16,27].

519 It is to be noted that these shit, s result to be larger for PC/ABS than for PC/SAN blends [23], probably because of an additional contribution to PC plasticization by low M.W. species of B. As a further evidence of this effect it was found, in PC/ABS blends containing different ABS, that the higher the B amount in the used ABS, the larger the shifts [77]. Some authors have proposed the existence of interzone layers at the PC/SAN [15,54,55] and PC/ABS [56,62,63,76] boundaries, responsible for the good adhesion between the phases; in these regions, in fact, there is the possibility of entanglement formation between PC and SAN or ABS chains.

4.3 Mechanical properties In general PC exhibits a very ductile behaviour in tensile tests on unnotched specimens. The ABS addition lowers the yield stress value, broadening more and more the relative peaks; reduces the cold-drawn ability of the PC matrix and the consequent elongation at break of the blends [59,76]. After the phase inversion, the blend curves resemble to that of pure ABS. The mechanism of deformation is mainly shear yielding in pure PC and in blends in which PC is the matrix. With increasing the ABS content the mechanism changes smoothly to crazing. The sequence of the microstructure deformation is craze initiation and propagation in the ABS, followed by arrest of crazes in the PC, and finally by extensive matrix and rubber particle deformations [ 15]. In contrast with these conclusions other authors found, by flexural [62] and by tensile tests [68], that voiding around rubber particles and cavitation, coupled with shear yielding in the PC matrix, are the main mechanisms of stress relaxation and toughening. Crazing, although precursor to final fracture, comes later on after noncrazing mechanisms and provides only a small contribution to the overall plastic deformation.

520 4.4 Impact properties Several authors have found synergistic effects in impact performances PC/ABS blends when PC is the matrix [56,61,63,66,69,76,77,82,83,85,86,88] only in same case [59] the curve of impact strength versus composition exhibited a minimum for intermediate values. From impact curves as well as by fractografic analysis [71,72,76,77] it was possible to establish that, with varying blend composition, an alternation of brittleto-ductile mechanisms occurred. Passing, in fact, from PC to blends containing more and more ABS, plain-strain to plain-stress overall transitions were observed. These effects at a microscopic level were likely due to a percolation mechanism [76] similar to the one proposed by Wu for impact behaviour of polyamides toughened by functionalized EPR rubber [89-92].

4.5 Influence of the ABS type on toughening The ABS composition is an important variable with respect to blend properties: a) The S/AN ratio is generally very close to the azeotropic S/AN ratio used for making SAN (about 25 wt % of AN) for most of ABS, whose value gives best adhesion between PC and SAN; therefore it has a relative influence; b) The PB amount is, instead of the utmost importance in determining the impact behaviour of PC/ABS blends. However literature results are somewhat contradictory: sometimes an increase of PB yielded synergistic effects [77,81]; in other cases minor improvements or even a worsening [59,82,85,86] of properties were obtained with such an increase. The reason is that a variety of parameters, other than ABS composition, can influence the impact performance: a) ABS and PC molecular characteristics [82]; b) molecular orientation; c) different processings and processing variables; d) thermal history; e) ABS rubber particle type, size and size distribution [87]; f) test

521 variables, such as temperature, strain rate, specimen thickness, notch radius, etc.; g) interfacial adhesion; h) low M.W. contents of ABS.

Acnowledgements This work was partially supported by II~ Progetto Fmalizzato Chimica Free e Secondaria of Italian CNR. The author wishes to thank D. R. Paul for sharing unpublished manuscripts on PC/SAN and on PC/ABS blends.

5. References 1. D. R. Paul and J. W. Barlow, " Polymer Blends (or Alloys) ", in J. Macromol. Sci.- Rev. Macromol. Chem., C18 (1) (1980), pp. 109-168. 2. D. R. Paul and J. W. Barlow and H. Keskkula, in Encyclopedia of Polym. Sci. and Eng., H. F. Mark, C. G. Overberger and G. Menges Eds., Wiley, New York, 1985, vol. 12, pp. 399-461. 3. L. A. Utracki, "Polymer Alloys and Blends" Thermodynamics and Rheology ", Hanser Publ., Munich, 1989. 4. L. Bottenbruch "Polycarbonates ", in Encyclopedia of Polym. Sci. and Techn., H. F. Mark, N. G. Gaylord and N. M. Bikales Eds., Wiley, New York, 1969, Vol. 10, pp. 710-764 5. D. Freitag, U. Grigo, P. R. Muller and W. Nouvertnr, " Polycarbonates ", in Encyclopedia of Polym. Sci. and Eng., H. F. Mark, C. G. Overberger and G. Menges,Eds., Wiley, New York, 1985, vol.ll, pp 648- 718. 6. N. G. Mc Crum, B. E. Read and G. Williams, "Anelastic and Dielectric Effects in Polymeric Solids", Wiley, New York, 1967, Ch. 13, pp. 534-537. 7. G. P. Ziemba, "Acrylonitrile-Styrene Copolymers", in Encyclopedia of Polym. Sci. and Techn., H. F. Mark, N. G. Gaylord and N. M. Bikales Eds., Wiley, New York, 1969, Vol. 1, pp 426-435.

522 8. F. M. Peng, " SAN Copolymers ", in Encyclopedia of Polym. Sci. and Eng., H. F. Mark, C. G. Overberger and G. Menges, Eds., Wiley, New York, 1985, vol. 1, pp 452-470. 9. C. B. Bucknall, " Toughened Plastics ", Appl. Sci. Publ., Essex, England 1977, pp. 90-107. 10. G. A. Momeau, W. A. Pavelich, L. G. Roettger, " ABS Resins ", in Kirk Othmer Encyclopedia of Chem. Techn., 3rd edition, H. F. Mark, D. F. Othmer, C. G. Overberger and G. T. Seaborg Eds., Wiley, New York , 1978, pp. 442-456. 11. A. Lebovits, "Acrylonitrile-Butadiene-Styrene Copolymers ", in Encyclopedia of Polym. Sci. and Techn., H. F. Mark, N. G. Gaylord and N. M. Bikales Eds., Wiley, New York, 1969, Vol. 1, pp. 436-444. 12. J. D. Keitz, J. W. Barlow and D. R. Paul, J. Appl. Polym. Sci., 29 (1984) 3131. 13. D. R. Paul and J. W. Barlow, Polymer, 25 (1984) 487. 14. I. C. Sanchez, Ann. Revs. Mater. Sci., 13 (1983) 387. 15. R. A. Mendelson, J. Polym. Sci., Polym. Phys. Ed., 23 (1985) 1976. 16. G. Locati and G. Giuliani, "Dynamic Mechanical Analysis of PC/SAN Blends ", in" Rheology ", Proc. VIIIth Int. Congr. Rheol., G. Astarita, G. Marrucci and L. Nicolais, Eds., Plenum Press, New York, 1980, Vol.3, p. 205. 17. M. Gordon and J. A. Taylor, J. Appl. Chem., 2, 493 (1952. 18. A. V. Tobolsky, "Structure and Properties of Polymers", Wiley, N. Y., 1960, p. 85. 19. B. Gregory, A. Hiltner and E. Baer, Polym. Eng. Sci., 27 (1987) 568. 20. B. Gregory, A. Siegmann, J. Im, A. Hiltner and E. Baer, J. Mat. Sci, 22 (1987) 532.

523 21. R. E. Skochdopole, C. R. Finch and J. Marshall, Pol3qn. Eng. Sci., 27 (1987) 627. 22. L. L. Berger and E. J. Kramer, J. Mat. Sci., 22 (1987) 2739. 23. W. N. Kim and C. M. Bums, Polym. Eng. Sci., 28 (1988) 1115. 24. P. R. Couchman, Macromolecules, 11 (1978) 1156. 25. P. R. Couchman and F. E. Karasz, Macromolecules, 11 (1978) 117. 26. P. J. Flory, " Principles of Polymer Chemistry", Comell Univ. Press, Ithaca, New York, 1953. 27. M. J. Guest and J. H. Daily, Eur. Polym. J., 25 (1989) 985. 28. R. F. Boyer, " Styrene Polymers ", in Encyclopedia of Polym. Sci. and Techn., H. F. Mark, N. G. Gaylord and N. M. Bikales Eds., Wiley, New York, 1970, Vol. 13, p. 309. 29. M. J. Guest and J. H. Daily, Eur. Polym. J., 26 (1989) 603. 30. K. W. Mclaughin, Polym. Eng. Sci., 29 (1989) 1560. 31. J. C. Huang and M. S. Wang, Adv. Polym. Tech., 9 (1989) 293. 32. D. Quintens, G. Groeninckx, M. Guest and L. Aerts, Polym. Eng. Sci., 30 (1990) 1474. 33. Ibidem, 30 (1990) 1484. 34. Ibidem, 31 (1991) 1215. 35. Ibidem, 31 (1991)1207. 36. H. Takahashi, T. Matsuoka and T. Kurauchi, J. Appl. Polym. Sci., 30 (1985) 4669. 37. H. Takahashi, T. Matsuoka, T. Ohta, K. Fukumori, T. Kurauchi and O. Kamigaito, J. Appl. Polym. Sci., 36 (1988) 1821. 38. Ibidem, 37 (1989) 1837. 39. V. S. Shah, J. D. Keitz, D. R. Paul and J. W. Barlow, J. Appl. Pol3~. Sci., 32 (1986) 3863.

524 40. J. Im, E. Baer and A. Hiltner, in "High Performance Polka-hers ", E. Baer and A. Moet Eds., Hanser, Munich, 1990, pp. 175-198. 41. C. B. Arens, Polym. Eng. Sci, 32 (1992) 841. 42. T. A. Callagan, K. Takakuwa, D. R. Paul and A. R. Padwa, Polymer, 34 (1993) 3796. 43. P. J. Flory, J. Chem. Phys., 10 (1942) 51. 44. M. L. Huggins, J. Chem. Phys, 9 (1941) 440. 45. I. C. Sanchez and R. H. Lacombe, J. Chem. Phys, 80 (1976) 2352. 46. I. C. Sanchez and R. H. Lacombe, J. Polym. Sci., Polym. Lett., 15 (1977) 71. 47. I. C. Sanchez and R. H. Lacombe, Macromolecules, 11 (1978) 1145. 48. I. C. Sanchez, Ann. Rev. Mater. Sci., 13 (1983) 387. 49. I. C. Sanchez in "Encyclopedia of Physical Science and Technology", R.A. Meyer ed., Academic Press, N.Y., 1987, Vol.11, p. 1. 50. I. C. Sanchez, Macromolecules, 24 (1991) 908. 51. Y. Takeda and D. R. Paul, J. Polym. Sci., Part-B, Polym. Phys., 30 (1992) 1273. 52. T.W. Cheng, H. Keskkula and D. R. Paul, Polymer, 33 (1992) 1606. 53. V.H. Watkins and S. Y. Hobbs, Polymer, 34 (1993) 3955. 54. J. Kolarik, F. Lednicky and M. Pegoraro, Polym. Networks Blends, 3(3) (1993) 147. 55. V. Janarthanan, R. S. Stein and P. D. Garrett, J. Polym. Sci., Part-B, Polym. Phys., 31 (1993) 1995. 56. D. Stefan and H. L. Williams, J. Appl. Polym. Sci., 18 (1974) 1451. 57. V. Dobrescu and V. Cobzaru, J. Polym. Sci., Polym. Symp., 64 (1978) 27. 58. J. D. Cooney, Polym. Eng. Sci., 22 (1982) 492. 59. H. Suarez, J. W. Barlow and D. R. Paul, J. Appl. Polym. Sci., 29 (1984) 3253. 60. G. Weber and J. Schoeps, Die Ang. Makromol. Chemie, 136 (1985) 45.

525 61. J. I. Eguiazabal and J. Nazabal, Polym. Eng. Sci., 30 (1990) 527. 62. M. Ishikawa and I. Chiba, Polymer, 31 (1990) 1232. 63. H. J. Radusch, G. W. Schreyer, H. Kausche, 3rd Dresden Polymer Discussion on" High Performance Materials ", Dresden, 16-19 April, 1991. 64. M. -P. Lee, A. Hiltner and E. Baer, Polymer, 33 (1992) 675. 65. Ibidem, 33 (1992) 685. 66. J. H. Chun, K. S. Maeng and K. S. Suh, J. Mat. Sci., 26 (1991) 5347. 67. J. Green, J. Fire Sci., 9 (1991) 285. 68. R. A. Bubeck, D. J. Buckley, Jr., E. J. Kramer and H. R. Brown, J. Mat. Sci., 26 (1991) 6249. 69. L. Suyun, Gaofenzi Cailiao Kexue Yu Gongcheng, 7(6) (1991) 62. 70. G. Kumar, R. Shyam, N. Sriram, N. R. Neelakantan and N. Subramanian, Polymer, 34 (1993) 3120. 71. M.-P. Lee, A. Hiltner and E. Baer, Polym. Eng. Sci, 32 (1992) 909. 72. M.-P. Lee, A. Hiltner, E. Baer and Che-I Cao, J. Mat. Sci., 28 (1993) 1491. 73. V. K. Stokes and S. Y. Hobbs, Polymer, 34 (1993) 1222 74. L. S. Dong, R. Greco and G. Orsello, Polymer, 34 (1993) 1375. 75. R. Greco and A. Sorrentmo, Adv. Polym. Techn., 13 (1994) 249. 76. R. Greco, M. F. Astarita, L. S. Dong and A. Sorrentino, Adv. Polym. Techn., 13 (1994) 259. 77. R. Greco and L. S. Dong, Proc. of the IVth Europ. Symp. on

"

Polymer

Blends ", Capri May 24-26, 1993, Die Makrom. Symp., R. Greco and E. Martuscelli Eds., 78 (1994) 141. 78. E. H. Kemer, Proc. Phys. Soc., 69B (1956) 808. 79. R. A. Dickie, J. Appl. Polym. Sci., 17 (1973) 45. 80. T. Kunori and P. H. Geil, J. Macromol. Sci. - Phys., B18(1) (1980) 93. 81. R. Greco, " PC/ABS Blends: Influence of ABS Composition on Properties ", in preparation.

526 82. T. Kurauchi and T. Ohta, J. Mat. Sci., 19 (1984) 1699. 83. L. Morbitzer, H. J. Kress, C. Lindner and K. H. Ott, Die Angew. Makromol. Chemie, 132 (1976) 19. 84. L. Morbitzer, D. Kranz, G. Humme and K. H. Ott, J. Appl. Polym., Sci., 20 (1985) 2691. 85. W. Y. Chiang and D. S. Hwung, Polym. Eng. Sci., 27 (1987) 632. 86. J. J. Herpels and L. Mascia, Eur. Polym. J., 26 (1990) 997. 87. B.S. Lombardo, H. Keskkula and D. R. Paul, J. Appl. Pol)a-n. Sci., 54 (1994) 1697. 88. J.-S. Wu, S.-C. Shen and F.-C. Chang, J. Appl. Polym. Sci., 50 (1993) 1379. 89. S. Wu, J. Polym. Sci. - Phys Ed., 21 (1983) 699. 90. S. Wu, Polymer, 26 (1985) 1855. 91. S. Wu, Polym. Eng. Sci., 27 (1987) 335: 92. A. Margolina and S. Wu, Polymer, 29 (1988) 2170.

527 CHA_IVrER I0

R U B B E R MODIFICATION OF BIODEGRADABLE POLYMERS

M. AveUa, B. Immirzi, M. Malinconico, E. Martuscelli, M.G. Volpe

National Research Council of Italy, Institute of Research and Technology of Plastic Materials, 80072 Arco Felice (Na) ITALY

M. Canetti, P.Sadocco, A. Seves*

Stazione Sperimentale Carta, Cellulosa, Fibre Tessili Naturali e Vegetali, 20100 Milano ITALY

1 Introduction

In order to develop materials with improved performances, copolymerization and blending have proved to be both alternative routes to introduction of new polymers.

Blending is the less expensive method.

Furthermore, while

copolymerization is confined to primary producers, polymer blending is basically a variation of compounding or formulating in general and many plastics, rubber and coating converters are well able to effect trials.

The success of such

approach is well documented by the large number of multicomponent polymeric systems introduced [ 1-3)]. As compatibility in polymers is usually an exception, different polymers, when mixed together, give rise to incompatible blends. In order to reduce the interracial tension, most successful proved to be the addition of "compatibilizers" or the induction of chemical or physical-chemical interaction during the blending

528 process [4]. This last case is quoted Reactive Polymer Blending (RPB) [5]. The main idea of RPB is to impart chemical reactivity to the polymers to be mixed, suitable to induce their compatibilization. The requirements for reactive blending in the melt consist of sufficient mixing to establish the desired morphology, presence of functionalities on both polymers and suitable reactivity across the melt phase boundary to form copolymers. Generally this kind of blending is used for polymers that contain functional groups in the backbone chains [6-7], which can undergo exchange reactions, such as amide-amide and amide-ester [8] interchange and trans-esterification under suitable reactive conditions. A different approach is when the modification is carried during the polymerization of one or both components. Interpenetrating Polymer Networks (IPN's) all fall in this last category: by a proper choice of monomers, crosslinkers and polymerization conditions, completely new products have been realized and, in some cases, starting from existing polymers. Reactive blending concurrently with the polymerization reaction is a very convenient method to obtain rubber-modified thermoplastic materials.

The

advantages of this procedure towards the melt-mixing of the components are numerous: - the polymerization of the matrix and the preparation of the blends are made in a single step, saving time and reducing the machining of the materials, which always produces some degradation; - the dispersion of the rubber component into the matrix is finer than that obtained with a melt-mixing process; -

in RPB by melt-mixing the range of the obtainable morphologies is limited,

normally, to spherical inclusion of one polymer into the other, while in the present case, unusual morphologies (core-shell, "salami" structure, etc.) are frequently encountered. Reactive blending is already applied in the case of rubber toughening of several thermoplastics.

It is found that by polymerizing caprolactam in the

529 presence of finely dispersed ethylene-co-propylene rubbers (EPR) bearing functional groups, or ethylene-co-vinylacetate rubbers (EVA), high performance polyamide 6 (PA6)-based multiphase blends with well defined morphologies are obtained [9-13].

Similar results are found for the polymerization of

dimethylterephthalate and 1,4-butanediol in the presence of dispersed reactive rubbers. More entangled morphologies are reported for the methyl methacrylate (MMA) polymerization in the presence of EVA rubbers [14] and in styrene polymerized in the presence of poly(butadiene-co-acrylonitrile) rubbers [15-16]. RPB approaches have been tempted for the toughening of naturally occurring polyesters, and the results are hereafter reported. Poly(13-hydroxybutyrate) (PHB) is a biotechnologically produced polyester that constitutes a carbon reserve in a wide variety of bacteria.

Rex:ently,

copolyesters of 3-hydroxybutyrate (HB) and 3-hydroxyvalerate (HV) have been isolated from Alcaligenes Eutrophus [17-18]:

/~H~

0II ~ #~2Hs

0II

-J~CH ,,,CI-12----C--O/~kCH-- CI-12--C ----0 HB

HV

PHBV are highly crystalline polymers with melting points and glass transition temperatures similar to polypropylene. Due to the characteristic of biodegradability (through non toxic intermediates) and processability, PHBV are being developed and commercialized as ideal candidates for the substitution of non-biodegradable polymeric materials in commodities application [19-20]. Their commercial development on such large scale is severely limited. Until recently, the prohibitive cost, narrow processability window (the difference between degradation and melting temperature), and, above all, the low impact

530 resistance around room temperature and below due to very high crystallinity and relatively high glass transition have prevented larger commercial usage. Independently, several research teams are approaching the problem of extending the potential

of PHBV through

various

blending methods.

Fundamental research has been carried out on solvent cast blending of PHB and poly(ethylene oxide) (PEO) [21]. Interesting correlation were found relating to the crystaUinity of blends at different composition. Blends of PHB and poly(methyl methacrylate) (PMMA) were obtained by melt mixing [22] and preliminary results were reported on their chemical-physical interaction. Other results on the thermodegradative behaviour of PHB-based blends [23] showed an enhancement of thermal degradation temperature by addition of poly(epichlorohydrin). When facing the toughening of PHBV, one has to keep in mind that not only the mechanical performance have to be improved, but also the biodegradability of the matrix must be saved. In this respect, RPB seems to be a realistic path: in fact, the chemical interactions are normally confined to few sites in the interfacial region, leaving unchanged the biodegradability of the matrix.

The only

condition, then, is to properly select a second polymer able to impart all the requested mechanical properties. We have used three different polymers in conjunction with PHBV: - poly(butyl acrylate), a low glass transition temperature (To rubber with a reported biocompatibility; - polycaprolactone, a low Tg semicrystalline polymer with a well recognized biodegradability [24-25]. - atactic poly(epichlorohydrin), elastomeric polymer non biodegradable. This review then deals with three different topics: 1) toughening of PHBV by "in situ" polymerization of butyl acrylate (PBA) following two different blending methods: bulk polymerization and suspension

polymerization;

531 2) toughening of PHBV by reactive melt blending with preformed polycaprolactone (PCL) in the presence of organic peroxide. 3) solution blending of PHB with atactic poly(epichlorohydrm).

System 1" toughening of PHBV by "in situ" polymerization

2.1 Bulk polymerization We have developed a method [26] in which the PHB (or PHBV) powders, as they came out from the bacterial polymerization and subsequent purification process, are thoroughly mixed with proper amount of acrylate monomers (or a mixture of acrylate monomers) and free-radical initiator resulting in a minor elastomeric phase intimately dispersed in the polyester matrix. The typical preparation of a blend is afferthere described: 70 g of polyester powder are intimately mixed with 30 g of acrylic monomer, like butyl acrylate, into which 60 nag of benzoyl peroxide (0.2 wt-% of acrylic monomer) are dissolved. The mixture is gently stirred for 24 h at room temperature. Subsequently, this homogeneous mixture is warmed to 90-100~ (extemal temperature) and left at this temperature under stirring for more 24 h, to allow acrylate polymerization. The residual monomer is extracted by vacuum stripping at 80~ and the recovered amount of the resulting blend is about 95 g. The yield of polymerization is 80-90%. It is worth to note that the inhibitor, always present in reagent grade acrylate, was not removed, in order to allow sufficient time for homogenization of reactants before polymerization. PHB homopolymer (PHBV0) and PHBV having 4 mol-% and 7 mol-% of HV comonomer (PHBV4 and PHBV7 respectively) were tested.

Although

several PHBV/PBA compositions were tested, it was found that by this method 30 wt.-% of poly(butyl acrylate) (PBA) is necessary in order to obtain relevant increase in the impact performance.

532 It was found that, occasionally, in the preparation of PHBV4/PBA blends, the acrylate polymerization occurs in regions macroscopicaUy separated from the polyester. However, at 0 %HV content, the growth of PBA occurs mainly on the PHBV powders which tend to agglomerate into larger grains. Such an effect, although not yet fully understood, could be related to change in the chemicalphysical accessibility of PHBV powder surface to BA monomer. The prepared "bulk" blends are given in Tab. 1. Sample PHBV0/PBA 70/30 PHBV4/PBA 70/30 PHBV7/PBA 70/30

% PHBV4 (g) % PBA ~g) 70 30 70 30 70 30

I I

Table 1. Composition of prepared "bulk" blends 2.2 Suspension polymerization [27-29] The bulk polymerization is scarcely reproducible, due to the lack of control of the mass polymerization of acrylic; in fact, by using an organic peroxide as polymerization initiator, like dibenzoyl peroxide, we observe an increasing of reaction temperature, that can cause the volatilization of monomers. It is mainly for the problem of exothermicity, that industrially the acrylics are polymerized in emulsion. These observation prompted us to attempt the suspension polymerization of butyl acrylate (or methyl methacrylate) in the presence of preformed PHBV powder.

By such modification we aimed to achieve a better control of

polymerization conditions.

We then prepare an aqueous emulsion of PHBV

powder to which the acrylic monomer together with peroxide is added. Under stirring, the acrylic polymerizes through a radical process induced by thermal

533 decomposition of peroxide.

The acrylic chains grow exclusively on PHBV

grains, probably due to an energetic gain. The typical preparation of this type of blend is the following: in a cylindrical reactor, 70 g of PHBV0 powder are dispersed in 400 mL of an aqueous solution containing 2.27 g of Na2HPO4, 0.13 g of NaH2PO4 and 4 g of a polyacrylic acid as emulsifying agent. The reactor is placed in a n oil bath at 110~ (extemal temperature) and 30 g of butyl acrylate, containing 0.3 g of benzoyl peroxide (1 wt-% of acrylic monomer) are added under mechanical stirring at fixed and controlled speed. The reaction proceeds for 24 h. After this time, the resulting blend is filtered, washed with boiling water and dried in vacuum. As aspected, this procedure really allows a better control of the polymerization condition in terms of heat exchange.

The acrylic polymerizes

through a radical process induced by the thermal decomposition of the peroxide. Moreover, probably, as consequence of favourable energetic balance, the acrylic phase grows exclusively onto PHBV0 chains; an important side effect of this is that the final shape of the blend particles is perfectly spherical, as normally achieved in suspension polymerization, and this is an advantage for subsequent processing of raw materials. The pol3nnerization yields are always very high (in the same range of bulk polymerization). In Table 2 we report the prepared "'suspension" blends

Sample PHBV0/PBA 80/20 PHBV0/PBA 70/30

% PHBV0 (~) 80 70

% PBA (~) 20 30

Table Z. Composition of "suspension" blends

It is worth to remark that the PHBV0 used for suspension process is different from PHBV0 of the bulk process. Particularly, they belong to different

534 bacterials straw, and this might have influences on the microstructure, and, hence, on the chemical-physical properties.

2.3 Blend characterization

In Tab 3 and 4 we report the thermal parameters of the bulk blends as obtained in differential scanning calorimetry experiments. Differential thermal analysis was carried out by using a Mettler TC 3000 differential scanning calorimeter. Two series of experiment were performed: in the first one, the samples was heated from room temperature to 200~ of 10~

cooled at the same rate down to -100~

10~

at a rate

and re-heated to 200~

In the second experiment, settled to visualize the glass transition

temperature, the sample was heated from room temperature to 200~ 20~

at

at

quenched down to -100~ and re-heated to 200~ at 20~ It is possible to observe that the transition temperatures of PHBV do not

change upon blending. Particularly Tg values are similar, thus indicating a bulk immiscibility of the two blend components. From the values reported in Tab. 3 [Part A) and Part B)] it is evident that the presence of PBA strongly depresses the enthalpy of fusion of PHBV, as well as their crystallization temperatures (particularly for PHBV0) and the enthalpy of crystallization. i lll

Sample PHBV0 PHBV4 PHBV7 PHBV0/PBA 70/30 PHBV4/PBA 70/30 PHBV7/PBA 70/30

Tm (~ I RUN

~ I * (J/g) I RUN

1740 160.9 156.0 176.3 164.8 160.3

122 97 82 86 73 57 i

Part A

T~(~ cryst

AH* (J/g) cryst

73 63 . . . . 55 60 . . . .

76 39 16 27

535

Tr (~

Sample PHBV0 PHBV4 PHBV7 PHBV0/PBA 70/30 PHBV4/PBA 70/30 PHBV7/PBA 70/30

49.0 57.2 62.7 53.3 57.2 52.1

AH* (J/g)

Tm(~

AH* (J/g)

c~st

11 RUN

11 RUN

12 37 43 43 17 34

174.9 165.6 163.0 175.1 166.0 164.7

110 88 70 70 63 50

Part B Table 3. Melting temperature Tm,, crystallization temperature Tr and enthalpic content All of the PHBV0, PHBV4 and PHBV7 and their "bulk" blends with 30 wt-% of poly(butyl acrylate), Part A) I RUN and Crystallization, Part B) II RUN (* the values are normalized with respect to the PHBV0 content) The reduction in the ability of PHBV to crystallize upon slow cooling from the melt is a clear indication of a reduction in primary nucleation of PHBV's in the presence of PBA. This must be related also to a "reacted interface" between the matrix and the rubber, which causes a reduction in the overall crystallization rate.

Sample PHBV0 PHBV4 PHBV7 PHBV0/PBA 70/30 PHBV4/PBA 70/30 PHBV7/PBA 70/30

T~ (~ 2.5 1.1 -0.1 2.5 1.2 1.2

Tm (~ 173 5 165 6 163 3 175 1 167.3 163 4

Table 4. Glass transition temperature Tg and melting temperature Tm of PHBV0, PHBV4 and PHBV7 and their "bulk" blends with 30 wt-% of poly(butyl acrylate).

536

2.4 Suspension polymerization Thermal analysis has been also used for the characterization of blends obtained by suspension polymerization. Surprisingly, following the same thermal treatment, rather different results are obtained. Particularly intriguing are the results obtained by the slow thermal process. It seems that the presence of 20% of PBA phase increases the enthalpy of melting and crystallization of PHBV0 phase. Similarly, in the experiments designed to obtain information on Tg's, we observe an increase in Tg of PHBV0 in the presence of PBA. As we have no reason to expect such differences from bulk to suspension polymerization process, we rather believe that the results can be at least partly attributed to the different origin of the PHBV0 used for bulk and suspension processes.

Sample PHBV0 PHBV0/PBA 80/20 PHBV0/PBA 7O/3O

Tm (~ I RUN 175.1 178.9 174.5

ii

~ (J/g) I RUN 78.0 107.1

Tr (oC) 86.1 73.4

All (J/g) c~st 64.3 75.4

Tm (~ II RUN 175.1 177.5

AH (J/g) 11 RUN 76.9 93.1

81.1

76.8

52.4

164.0

76.3

i i

i

Table 5. Melting temperature Tm, crystallization temperature Tr and enthalpic content AH of the PHB/PBA suspension blends with 20 and 30 wt-% of poly(butyl acrylate). iii

Samples

Tg PBA Tg erm (oC) , (~

PHBV0 PHBV0/PBA 80/20 PHBV0/PBA 70/30

---41.6 -43

ii

ii

Tr176 cryst

AH(J/g) cWst

Tin(~ 11 RUN

AH(J/g) 11 RUN

1.4 6.4

53 57.6

32.7 36.9

167.0 173.3

80.6 70.5

4.9

57.7

37.9

171.6

74.9

Table 6. Glass transition temperature Tg, melting temperature Tin, crystallization temperature Tr and enthalpic content AH of the PHB/PBA suspension blends with 20 and 30 wt-% of poly(butyl acrylate)

537 It is conceivable that the present PHBV0 contains either impurities or low molecular weight species which are removed upon the treatment with our reaction medium

2.5 Etching procedure Morphological characterization was carried out on the A u ~ d coated surfaces of the dumbbell specimens using Philips SEM 501 electron microscope; the samples of blends were smoothed with a LKB Ultramicrotome by means of a glass knife and etched with b ~ o l SEM

micrographs

of

smoothed

vapour in order to remove the acrylic phase. and

butanol-etched

PHBV0/PBA

PHBV4/PBA molded samples are shown in Fig. 1 and 2.

and

For the sake of

comparison, also PHBV4 was etched with n-butanol, leaving an unmodified surface. The smoothing technique was sometimes by itself sufficient to remove the rubber from its domains, especially when a not freshly made glass knife was used, as result of the shearing of the domain contour (see Fig. 2b). Of course, the use of n-butanol is necessary in order to fully reveal the morphology. It can be seen how the rubber is segregated in very small domains. No large differences were found in rubber domain size distribution between the blends containing PHBV0 and PHBV4 :.

~,

or :

....

. . . " ~ . . . . : .;~:n

,... '.....~.;~,.~=.

: , ' . : ; ' ~ , ~.

"....',,'~,;,.,..:.,.:;~

:~_ e...,

:

~e- . , - ' . , 't "

,.

r ,. ." * o~" * .?'' "

,

,',i

. " - ~', V

~'

~.": ,, ". ..:,~,~:/'... .....

~ '"'"'~.."

'~_'..,:., 9 ~ ;~ ~ ' ~ , ~

" 9 " " " # l l" " . ' . ~ t ~~ ' . ' ~ ~'" ,;~ . " ~ '9i " ""

p;..,,~.;.,.~..~. O.~,?~,~l~.~r~.l

,-',,i.

a b Fig. 1. SEM micrographs of PHBV0/PBA 70/30 blend: a) before etching, b) after etching

538

Fig. 2. SEM micrograph of PHBV4/PBA 70/30 blend: a) before etching b) after etching As it has been observed, by swelling experiments, that the permeability of PHBV to BA is very poor, it is conceivable that the polymerization occurs in the interstices between the aggregated polyester powders (the agglomeration is caused by the purification process following the biotechnological synthesis) (see Fig. 3).

Fig. 3. SEM micrograph of original PHBV4 powder

539 The post-treatments of the blends (mixing in the Rheocord chamber and compression molding) should not substantially alter the original morphology obtained in synthesis, due to the partially crosslinked nature of bulk-polymerized PBA [30].

2.6 Impact properties The impact properties were analyzed according to the Linear Elastic Fracture Mechanism (LEFM) approach [31].

The procedure used for the

calculation of the critical strain energy release rate (Gc) is reported by Coppola et al. [32]. Charpy-type specimens (6.0 mm wide and 60 mm long) were cut and notched with a fresh razor blade.

Then they were fractured at different

temperatures and at an impact speed of 1 m/s by using an instrumented Charpy pendulum. The impact properties of samples of PHBV0 and PHBV4 polymers and of their blends are reported in Fig. 4. It is evident the positive influence exerted by the rubber on the fracture toughness of polyesters. The effect is particularly marked at temperatures close and above room temperature.

In fact, while the PHBV0 and PI-IBV4

homopolymers are still rather brittle at room temperature, their blends are much more ductile. The enhancement of properties is more pronounced for PHBV0, which is known to be very brittle due to high crystallinity. The effect is less pronounced on PHBV4, probably because the addition of valerate comonomer is already effective in the induction of ductility [33]. There is a clear shill of the brittle to ductile transition in blends, compared with homopolymers. Once again, the effect is dramatic for the very brittle and stiff PHBV0.

540 Fracture t o u g h e n e s s v e r s u s T e m p e r a t u r e

,.

G (kJIm 2)

..

a

i-I PHB pure OPHBIPBA 9PHBV 4 mol % HV OPHBVIPBA

m

..

..

I

u

O-

-1 O0

I

I

-80

-60

I

- 40

T

I

I

I

l

-20

0

20

40

(oc)

Fig. 4. Fracture toughness of PHBV/PBA blends as a function of temperature By comparison of the micrographs of the blends by means of SEM, we note that the particle size distribution is quasi-bimodal, (i.e., larger spherical particles of 10 ~tm diameter co-exist with fine particles of less than 1 ~m size). It is known that, in thermoplastics technology, a bimodal distribution of a dispersed rubber is more effective in the induction of plastic deformation [34].

2.7 Graft interpolymerization Swelling experiments have been carried out on several blends in order to investigate the possible formation of cross-links between PHBV and PBA "bulk" blends.

It is, in fact, well known that the addition of peroxides to aliphatic

polyesters, like polycaprolactone, can reduce the formation of macroradicals by extraction of labile hydrogen from the polymeric backbone [5]. It is conceivable that PHBV can undergo similar reaction and so, in the presence of BA monomer

541 that PHBV can undergo similar reaction and so, in the presence of BA monomer and/or growing PBA macroradicals, intergraR processes may occur according to scheme 1.

I

I

0

0

I

H--

C--CH

I

3

I

?H 2

' C

+

R'

CH 3

I

~

?H 2

c =o

RH

+

RH

c--0

PHBV

PHBV

o

o

I

I

H--

+

C

CH 3

I

H2 C=O i

H~

+

R'

~

I

C --CH

I

' CH I C--O

PHBV

3

PHBV

macroradical (I)

I

O

I

O

-~4H9

I , C I

CH3

CH 2

I

c =o PHBV

+

o

I c--o

I

H--C --ell 2

I

C-

CH3

I

CH 2

O--C

I

!

CH 2 - - C H

I

C--O--C4H 9

II O

PHBV

macroradicals (II) The results of the swelling experiments are somewhat intriguing. In fact, by using chloroform, a common solvent for PHBV and PBA, we obtained a

542 supematant phase which accounts for about 10% of the charged "bulk" blends. DSC experiment on such residual phase did not show clear evidences of an elastomeric phase (no Tg was recorded at low temperature).

In our opinion a

possible explanation could be the following: the growth of the macroradical (II) by further addition of BA monomer is severely limited by the heterogeneity of the reacting medium. The fate of such macroradical (IT) is likely to be either the formation of short grafts or the termination on a second PHBV chain (crosslinks). On the contrary, dissolving in chloroform the "suspension" blends, we obtained much higher amount of a "graft copolymer" (about 30 wt-%), in which both transition temperatures of PHBV and PBA were found (see Fig. 5). In this case, as we found that BA chains growth only on PHBV powder, we hypothesize that this increases the possibility to obtain long graft before the PBA chain meets another PHBV chain. We hypothesize that the increased yield of grafting is due to the fact that the PBA phase grows as thin layer inside the PHBV0 powder. The improvement of phase distribution, together with the better control of polymerization kinetic, should allow the growth of longer grafted chains of PBA once they have reacted onto PHBV0 substrate.

Tg (PBA)

Tg (PHBVO)

E 0

I

-100

I

I

0

I

1

100

. . . . . . . I

I

200

Fig. 5. DSC trace of PHBV4/PBA 80/20 "suspension" copolymer

543

3 Biodegradability of PHBV To fully exploit the possible technological development of our materials it is very important to check ton which extent the substitution of 30 wt-% of PHBV with an acrylate phase influences the biodegradation of matrices. It has been reported [35] that a Gram positive bacterium which produced extracellular enzymes that degrade the homopolymer PHB when blended with the non-biodegradable atactic poly(epichlorohydrin) was isolated and tentatively assigned as Aureobacterium Saperdae. We studied the microbial degradation of plato PHBV and PHBV/PBA blends with Aureobacterium

Saperdae.

The PHBV-based blends are

characterized by mechanical and morphological methods as a function of bacterial degradation.

3.1 Enzymatic degradation Aureobacterium Saperdae cultures, where the only carbon source was the polymeric sample, were used to degrade to different extent pure PHBV4 and PHBV4/PBA blends (80/20 and 70/30 weight ratios). The micro-organism was pre-cultured overnight on 0.1% LB broth, about 3.5 mL aliquots of this culture were used to inoculate 500 mL flasks containing 100 mL of mineral medium (mineral medium composition: 1 mg/mL NH4CI, 0.5 mg/mL MgSO4.7H20 and 0.005 mg/mL CaC1.2H20 in 66 nM KH2PO 4 (pH = 6.8). The flasks were added with the polymeric samples and incubated at 30~

under shaking, for 15

days. In addition, control experiments were run to verify chemical hydrolysis of polymeric samples immersed in mineral medium at 30~

after 15 days no weight

loss of the samples was found. Cultures at different polymer degradation percentages were stopped and the samples were used to perform various morphological analysis.

544 The percentage of polymer degradation was determined by measuring the weight loss of the sample during bacterial attack.

Having checked the non-

biodegradability of PBA phase, we normalized the weight loss on the PHBV content, thus obtaining the percent of degradation in blends. Polymer samples were removed from the culture medium at different time lengths, washed several times with distilled water and dried to constant weight under vacuum. 3.2 Bacterial degradation of PHBV4 and PHBV4/PBA blends

As previously reported, Aereobacterium Saperdae is effective in the of PHB homopolymer [35]. Thus samples of plain PHBV4 and PHBV4/PBA blends (80/20 and 70/30 weight ratios) were utilised by Aureobacterium Saperdae as the only carbon source, and several polymeric specimens at different percentage of weight loss were analyzed. Preliminary tests carried out with pure PBA revealed that Aureobacterium Saperdae can not degrade this polymer. The thickness of the polymeric samples was measured before and after the bacterial attack.

0"12 !

~

~

0.1

u~ (9 C .~ r

Be

0.06

E!

0.04

=,-=

J=

P-

0.02 -

0

t

10

ii

20

t

=

I

30 40 50 Weight loss (%)

!

60

"

|

70

Fig. 6. Thickness decreasing due to microbial attack for PHBV4/PBA blends as a function of weight-loss (PHBV4 (m); PHBV4/PBA 70/30 (A); PHBV4/PBA 80/20 (D)

545 In Fig. 6 the thickness decreases are reported for the samples as a function of weight loss. As general consideration, we observed that the degradation proceeded with the same mechanism for the PHBV4 and for the blends, in fact the thickness decreases corresponded to the percentage of weight loss, indicating that the polymer erosion, by the degradative enzymes, procx~ed via surface dissolution. As further check of the exclusive surfacial erosion, test samples were cut m the transverse direction and the inner state was examined: no differences where found while degradation proceeds, for PHBV4 and PHBV4/PBA blends.

An

analogous decrease in both percentage of weight loss and percentage of film thickness, was also obtained with microbial polyesters films degraded by Aureobacterium Saperdae [35] and by Alcaligenes Faecalis depolymerase [36]. As shown by Fig. 7, plato PHBV degraded of 70% of its initial weight during 15 days of bacterial attack. 100 80 A

0

60

,o 2O

oi

I

0

3

I

I

6

9

12

15

Time (day) I-I PHBV 4 neat

II PHBV80/PBA20

~

PHBV70/PBA30

Fig. 7. Weight loss due to microbial attack for PHBV4/PBA as a function of time (PHBV4 (n); PHBV47PBA 80/20 (m); PI-IBV4/PBA 70/30 (~D)

546 Moreover, while in 80/20 blend, the bacteria consumed almost all the available PHBV4 (70 % of PHBV4 contained in the blend), the degradation stopped at 40% of available PHBV4 in the blend 70/30. The reason of such effect can be related to the morphology of blends (see below).

3.3 Morphological characterization SEM analysis of the surface of pure PHBV4 (see Fig. 8a-8d) samples after bacterial attack evidenced the homogeneous superficial erosion caused by the degradative enzymes, while no changes took place inside the sample. During bacterial degradation of the 80/20 blend, pieces of PBA component released in the culture were macroscopically visible.

As consequence of the

bacterial attack, the PHBV4 present on the surface was eroded and pieces of the dispersed PBA component were released allowing new PHBV4 zones to be accessible to the degradative enzymes

~ - ~ . . . 7

,ji;. ........

..

,,~v.~.

, ~,. ,,~ . *

.

,,,..~, .

b

i ~

:

.,2

Fig. 8. SEM micrographs of PHBV4 samples after microbial attack: a) 0% deg. (80x); b) 20% deg. (640x); c) 35% deg. (640x); d) 70% deg. (640x)

547 The SEM analysis of the degraded blend (see Fig. 9a-9d) confirms this degradation mechanism: while the degradation proceeds, large aggregates of rubber-like domains become evident, surrounded by eroded PHBV4 matrix.

Fig 9. SEM micrographs of PHBV4/PBA 80/20 sample after microbial attack: a) 0% deg. (80x); b) 23% deg (40x); c) 35% deg. (40x); d) 50% deg. (40x) The degradation, through surface erosion and PBA abiotic release, continued up to about 50% of overall weight loss. In 80/20 samples degraded at this extent the PHBV4 was no more accessible to the degradative enzymes. Anyway, such extent is very close to the amount of degradable PI-IBV4 found in the homopolymer sample test. In the case of the 70/30 blend only 20% of weight loss could be reached compared to the 50 theoretically available in 15 days bacterial exposition. No release of the PBA component was macroscopically visible in the bacterial culture and the thickness of the degraded samples did not significantly change aRer degradation.

SEM analysis (see Fig.

10a-10c)

evidenced the

impoverishment of PHBV4 zones. At the surface of the degraded blend there remains a PBA continuous phase containing small holes due to the removed PHBV4 domains (see Fig. 10c).

548

Fig.10. SEM micrographs of PHBV4/PBA 70/30 samples aRer microbial attack: a) 0% deg. (80x); b) 10% deg.(40x); c) 21% des. (40x) Such morphology do not allow a deeper penetration of bacteria, with, as consequence, an inhibition of further degradation.

3.4 Tensile behaviour

Tensile tests were performed at room temperature with an Instron machine at 10 mm/mm cross-head speed on dumbbell specimens of 1 mm thickness. The sample was cut from a compression moulded sheet prepared by heating the powder at 185~ for 5 min without pressure, then applying a pressure of 10 MPa for 2 mm at the same temperature. In Figs. 11 and 12 the trends of Young's modulus (E) and the strength at break (Ob) against the percentage of weight loss for PHBV4 and PHBV4/based blends were shown.

549

3.5 A

t~

a.

(3 = 2.5 10 O E 1.5 r

-

~

, m

.

....

1/1 t~ UJ

[3

D

0.5 i

0

....

!

I

i

I

t

I

I

10

20

30

40

50

60

70

Weight loss (%) 9PHBV 4 neat

r'IPHBV80/PBA20

A PHBV701PBA30

Fig. 11. Young's elastic modulus trends due to microbial attack for PHBV4/PBA blends as a function of weight-loss (PHBV4 (m); PHBV4/PBA 80/20 (n); PHBV4/PBA 70/30 (~) From these figures the following can be deduced: - as previously described [37] the presence of PBA particles in PHBV matrix produced a strong improvement of impact properties in the analyzed temperature range, due to the toughening effect of rubber phase (stress absorbers); however, according to above description, a decreasing of Young's modulus (see Fig. 11) and strength at break (see Fig. 12) was observed for the PHBV4/PBA blends with respect to PHBV4 homopolymers; - the modulus calculated for PHBV4 sample (Fig. 11) seemed not to vary with biodegradation process, while crb showed only a slight decreasing (see Fig. 12); this behaviour can be explained by the surfacial pathway of bacteria action, that did not allow a drastic fall of mechanical properties; -

also the reduction of Ob can be due to surfacial erosion of bacteria that can

act as crack initiator points;

550

25== a.

20-

~=

15--

A

=E L

.Q r

~p=m10~ T--5

0

0

F,

10

20

I I PHBV 4 neat

i i 30 40 Weight loss (%) I"1 PHBV801PBA20

l 50

i....... i 60 70

PHBV701PBA30

Fig. 12. Strength at break trends due to microbial attack for PI-IBV4/PBA blends as a function of weight-loss. (PHBV4 (m); PHBV4/PBA 80/20 (121); PHBV4/PBA 70/30 ( ~ ) - the PHBV4/PBA 80/20 blend, as well as PHBV4 homopolymer, presented an almost constant trend both for Young's modulus and strength at break up to 45% of weight loss; for higher weight loss (>50%), the samples seemed to loose mechanical consistency, as consequence of the increased mechanical detachment of PBA from the test sample; - finally, the PHBV4/PBA 70/30 showed a regular decreasing of mechanical properties both for the modulus and strength at break. This latter behaviour can be explained again with the morphology produced by erm3maatic degradation. In fact, the erosion of PHBV from the blend surface leads to an increase in the rubber content which for this blend is not detached from the surface, at contrary of the previous case. Consequently, the tensile mechanical properties regularly decrease with the increase in the degradation time.

551 System 2: Reactive melt blending of PHBV 4 Introduction Poly-e-caprolactone

(PCL)

and poly(]3-hydroxybutyrate-co-13-hydroxy-

valerate) (PHBV) are gaining interest from producers of polymer based goods, due to their technological properties and their inherent biocompatibility and biodegradability [3 9-44]. Since PCL can be blended with a variety of other polymers to improve their deficient properties [24-25] and PHBV is a highly crystalline polymer which is thermo-processable but brittle, a blend of the two materials could give promising results. linkage,

Since both PCL and PHBV possess the same functionality, the ester one

can,

in

principle,

employ transesterification

to

induce

compatibilization. Unfortunately, the extreme sensitivity to thermal degradation of PHBV and the high temperatures normally required for ester-ester interchange, render such a route to compatibilization impractical. In the present work we approached the problem of reactive blending of PCL and PHBV in a static mixer by means of the addition of peroxides, i.e. dibenzoylperoxide (DBPO) and dicumylperoxide (DCPO). According to the type of peroxide, two different temperatures were employed in blend preparation. For comparative purposes, the same blends had been prepared in the absence of peroxides, and the observed differences in behaviour between mechanical and reactive blends were attributed to the hypothetical formation of graR copolymer species.

Moreover the present paper is concemed with the isolation and

characterization

(thermal,

thermogravimetric

and

spectroscopic)

of the

intergraRed species from the blends prepared in presence of peroxides. Thermal and mechanical properties of all prepared blends were investigated.

552

4.1 Preparation of PCL/PHBV 70/30 with peroxide at high and low temperature Tab. 7 shows a summary of the blends and the conditions of their preparation.

The codes include the initials of the polymers, the blend

composition (weight ratio on a basis of ten) and one or two letters which indicate whether the blends are prepared at high (H) or low (L) temperature, and whether in the absence (M) or in the presence (P) of 0.5 wt.-% of peroxide. PCLPHBV 73 HP was prepared by mixing of 28 g of PCL, 12 g of PI-IBV and 0.2 g of DCPO (0.5 wt.-%) in the reaction chamber of a Haake Rheocord at a mixer roller speed of 32 r.p.m, for 10 mm at 160~

After mixing, the molten

blend was quenched by a stream of compressed air. By the same method, sample PCLPHBV 73 LP was prepared by reacting 28 g of PCL and 12 g of PHBV together with 0.2 g (0.5 wt.-%) of DPBO at 100~ For comparison, 40 g of homopolymers were mixed in the presence of 0.2 g of DCPO or DBPO in the same experinaental condition.

4.2 Preparation of PCL/PHBV 70/30 without peroxide at high and low temperature PCLPHBV 73 HM was prepared by mixing of 28 g of PCL and 12 g of PHBV in the reaction chamber of a Haake Rheocord at a mixer roller speed of 32 r.p.m, for 10 min at 160~

After mixing~ the molten blend was quenched by

a stream of compressed air. By the same method, sample PCLPHBV 73 LM was prepared by reacting 28 g of PCL and 12 g of PHBV at 100~

553 Code PHBV P PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCL P

37 37 37 37 73 73 73 73

LM LP HM HP LM LP HM HP

PCL

PHBV

T

(wt.-%)

(wt.-%)

(oc)

0 30 30 30 30 70 70 70 70 100

100 70 70 70 70 30 30 30 30 0

160 100 100 160 160 100 100 160 160 160

Peroxide DCPO DBPO DCPO DBPO DCPO DCPO

Table 7. Code and composition of the prepared blends 4.3 Extraction with chloroform 10 g of PHBV/PCL 73 P blend were treated with 500 mL of chloroform in a flask equipped with a reflux condenser and heated at 60~ under stirring for 24h. The mixture was then transferred into a separatory funnel and left to equilibrate for a few days.

The formation of two phases was observed: a bottom clear

solution and an upper phase of insoluble in chloroform.

The CHC13-insoluble

phase was recovered, washed repeatedly with CHC13, dried and then analysed. About 280 nag of solids were obtained.

The chloroform solution was

filtered, evaporated and the extract was analyzed too. By the same way 10 g of PI-IBV/PCL 37 P blend were mixed with 500 mL of chloroform to give about 100 nag of an insoluble phase.

The results of the

gravimetrical analysis are collected in Tab. 8.

Code PI-IBV/PCL 73 Res. PHBV/PCL 37 Res.

Amount of CHC13 insoluble 280 mg 100 m~;

PHBV/PCL wt. ratio in the residue 94/6 69/31

Table 8. Gravimetrical analysis of blend residuals after CHC13 extraction

554 Also PHBV P and PCL P were treated with chloroform: while PCL P completely dissolves to give a 2% (w/v) solution, PHBV P is almost insoluble, even after heating, and only swells under the action of the solvent. Pure homopolymers as well as mechanical blends dissolve completely in chloroform which is in fact a good solvent for both of them. 4.4 T h e r m a l analysis

Two series of DSC experiments were performed: in the first, the sample is slowly heated from -100~ to 200~

in the other type of experiment, designed to

study the behaviour of glass transition temperature (Tg), two heating rtms were performed at 20~

with a very fast quenching stage between them. In Tab.9

and 10, respectively, the results obtained from the two series of experiments are reported. For sake of simplicity Tab. 9 has been split in two parts" in Part A, the blends processed at 100~

are reported and compared whit the polymers, while

in Part B, the blends processed at 160~

are collected.

For comparative

purpose, PHBV, PHBVP, PCL and PCLP are included in both parts, but it must be noticed that PHBV and PCL samples are processed at 160~ and PHBVP and PCLP are treated with DCPO.

Code PHBV PI-IBV P PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCL P PCL

Part A

T.~ PCL

(oc)

37 37 73 73

LM LP LM LP

61.0 59.9 60.6 60.1 60.4 64.3

Tn~PHBV

(oc)

159.2 172.1 150.7 162.4 161.3 160.7 -

Xc,PCL

(wt.-%) 35.7 33.7 42.3 38.7 37.5 45.8

Xc,PHBV

(wt.-%) 58.0 58.8 55.0 55.7 45.6 44.3 -

555 Code PHBV PHBV P PCLPHBV 37 HM PCLPHBV 37 HP PCLPHBV73 HM PCLPHBV 73 HP PCL P PCL

Part

Tm.PCL

Tm.PHBV

(oc)

(oc)

-

159.2 172.1 161.7 171.8 164.8 160.6 -

-

60.3 61.1 60.2 60.6 60.4 64.3

Xc,PCL

Xc,PHB

(wt.-%) (wt.-%) 38.3 43.0 43.1 39.6 37.5 45.8

58.0 58.5 49.9 52.4 41.7 42.4 -

B

Table 9. Thermal and structural data of homopolymers and blends. Part A: Blends processed at 100~ Part. B Blends processed at 160~ From Tab. 9 B it is possible to observe an increase of PHBV's

Tm when

treated whit DCPO (PHBVP); such an effect is observable also in PCLPHBV 37 HP, compared to PCLPHBV 37 HM.

On the contrary the

Tm's of blends

obtained by the low temperature method are almost unaffected by the presence of DBPO. The above finding seems to suggest that some structural change occurs in PHBV when treated in the melt with DCPO. As a matter of fact, PHBV is no longer soluble in chlorinated solvents aRer DCPO treatment, indicating the formation of crosslinks. The same effect on the melting point of PHBV is found in PHBV-matrix blends, while it is absent in PCL-matrix blends, probably as consequence of the radical scavenging effect of PCL. The peroxide seems to have no influence on the PCL melting point, as well as on its solubility characteristics. In Figs. 13 and 14 crystallinity content (xo) of PCL and PHBV, respectively, is plotted as a function of the PCL percentage in the blends. A comparison of Fig. 13a and 13b shows that the blends prepared at 100~ without and with DBPO, respectively, exhibit a similar thermal behaviour, i.e. in

556 both cases x~ of PCL and PHBV decreases with decreasing their content in the blends

70

x~.(%)

X, (%)

7O

o}

60

b)

6O 5040

30

;0

'

0

20

i 60

i 80

100

30

I

I , 20

0

I

40

PCL(wt.-%)

I

60

I

100

80

PCL (wt.-%)

a b Fig. 13. Crystallmity content of PCLPHBV blends prepared at 100~ as a function of blend composition: a) without peroxide, b) with peroxide (e) PHBV, (o) PCL

Some interesting differences are found between blends prepared at 160~ without and with DCPO, as shown in Fig. 14a and 14b, respectively.

x,:(~) 70

X (%) 7O

60q

60-

4O

4O-

b)

o)

30

0

'

20

;o-

'

6o

PCL (wt.-%)

'

80

~o

1

~o0

~'---"--'------~o I

20

I

30

/

40

1

50

I

100

PCL (wt ;%)

Fig. 14. Crystallinity content of PCLPHBV blends prepared at 160~ as a function of blend composition: a) without peroxide, b) with peroxide (e) PHBV, (0) PCL

557 In fact, while for the blends prepared without peroxide the x~ values of PCL and PHBV decrease with decreasing their content in the blends (Fig. 14a), in the blends prepared with DCPO, PCL's x~ increases by decreasing its percentage in the blends (Fig. 14b). No significant differences are found between blends prepared with and without DBPO.

The origin of the above differences may be explained by a

possible compatibility of the two polymers in the molten state and the formation of grafted species between PHBV and PCL. In the above hypothesis, the blends are in homogeneous liquid state during processing at 160~

Upon cooling

PHBV crystallization may occur in mechanical blends with total rejection of PCL which is still in a liquid state. PCL then crystallizes in its own domains and is not influenced by the presence of PHBV crystals. Such a process is controlled by the crystallization rate of PHBV, which is known to be a slowly crystallizable polymer [45]. On the contrary, for the blends processed at 160~ with DCPO, the crystallization of PHBV from the homogeneous melt may not lead to total rejection of PCL phase, due to the presence of PHBV/PCL-graffed species. As a consequence, PCL crystallizes aRerwards in the interstice left by PHBV crystallization, which might well exert a positive influence on the overall crystallinity of PCL. The Tg's of the two polymers and of their blends are collected in Tab.10. Detectable variations in the Tg values are observed coveting a range of maximum 6~

At the moment, it is difficult to see a regular trend, as the

investigation is restricted to two compositions.

558 Code

%PCL

Tg.PHBV

(oc)

PHBV PHBV P PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCL P PCL

(oc)

5.1 4.1 3.2 1.8 -0.8 -2.7 0.5 3.5 3.4 - 1.0 -

-

37 37 37 37 73 73 73 73

LM LP HM HP LM LP HM HP

-62.5 -54.1 -59.4 -61.1 -54.0 -59.1 -56.4 -71.4 -58.2 -62.5

Table 10. Glass transition temperatures of polymers and blends

From the analysis of DSC traces in quenching experiments, it is possible to observe an other phenomenon: while PHBV quenched from the melt undergoes only a partial crystallization, as can be observed from the crystallization peak obtained in a second heating run (see Fig. 15a), the same polymer shows a complete crystallization in quenching upon treatment with DCPO (see Fig. 15b).

Temperature in ~

o

x ~

...... -50

.... 0

,

50

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

100 150 20 0 Temperature in ~

i

-50

.

.

.

.

i

0

.

.

.

.

i

50

,

-

-

-

I

.

.

.

.

i

.

.

.

.

i

100 150 200 Temperature in ~

Fig. 15. DSC traces of: a) PHBV (11 RUN), b) PHBV P (II RUN)

559 (B)

I -,oo

-~o

o

~o

~oo

~o

~oo

Temperolure in ~

_1oo-~'o

o

~o

a

~oo

~o

~oo

TemperQture in ~

b

Fig. 16. DSC traces of: a) PCLPHBV 37 P (II RUN), b) PCLPHBV37 P Res. (11 RUN) Blend PCLPHBV 37 HP behaves accordingly.

Once again, a possible

explanation of such results may be that DCPO produces some degree of crosslinking in the molten of PHBV. As a consequence the crystallization rate of the polyester is strongly accelerated. Thermogravimetrical experiments were performed by heating the samples from 40~ to 500~ at a heating rate of 20~

In Fig. 17 the TGA traces of

PCL and of PHBV are reported.

100

200

300

/.13.2 400 500 600 Temperoture in ~

283.8 100

200

300

Fig. 17. Thermogravimetrical traces of: a) PCL, b) PHBV

400 500 600 Temperoture in ~

560 Identical behaviour is recorded from PCL P and PHBV P. In Fig.18 the TGA pattem of the residue to chloroform extraction obtained from the PHBV/PCL 73 P blend is compared with that of the parent blend.

(A)

I/I

(:5

B

Ii

275.7 100

200

277,9

300 400 500 Temperoture in ~

100

200

300

/.00 500 600 Temperoture in ~

a

b

Fig. 18. Thermogravimetrical traces of: a)PCLPHBV 37 P res, b) PCLPHBV 37 P From the figure it seems evident that the chloroform residue is made up prevalently of PHBV with a little content of PCL, since the peak at 275~ is due to the degradation of PHBV and the shoulder at 420~

to the degradation of

PCL. TGA thus confirms DSC data about the formation and composition of copolymeric species.

4.5 FTIR spectroscopy analysis In Fig. 19 the FTIR spectra of PHBV, PCL and of the chloroform residue from the PHBV matrix blend are reported.

The spectra are recorded with a

Nicolet 5DXB spectrophotometer with a resolution of 4 cm-1. The spectrum of the residue is close to that of PHBV, bt~ the presence of a peak at 737 cm ~, due to skeletal motions of (-CH2-)5 in PCL, reveals the small, but detectable, presence of such polymer; in addition, the peak absorbance at 2941 cm 1 slightly increase

561 with respect to the one at 2988 cm 4 in PHBV, due to the contribution at 2941 cm 1 of the stretching of methylene sequences present in PCL. ,! t

/ \ t'

I--- ......

800

--I--- . . . .

750

4-- . . . . . .

700

-+---------

650

~,

600

Wovenumber

----+---'

550

I

500

.....

+

. . . . .

450

I

Z.O0

~n c m -1

Fig. 19. FTIR spectra of: a)PCL, b) PCLPHBV 37 PRes., c) PI-IBV

4.6 Mechanical properties In Tab. 11 the relevant parameters obtained from the tensile tests

of all

blends and the two polymers are reported.

Code PI-IBV PI-IBV P PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCLPHBV PCL P PCL i

E 10-3

37 LM 37 LP 37 HM 37 HP 73 LM 73 LP 73 HM 73 HP

~B

10 -2

(Ks/cm5

(Ksdcm5

20.O 16.0 14.8 12.7 15.7 11.3 4.6 5.3 4.2 5.2 2.3 3.6

2.22 3.28 2.09 1.74 2.11 1.43 1.70 1.29 1.30 1.83 4.20 5.01

Table 11. Mechanical data of the overall blends and of the pure polymers

562 The obtained data show that: -

the elastic modulus (E) decreases for PHBV- based blends with respect to

the modulus of PHBV. for the above blends, as well as for PHBV, lower values of moduli are

-

obtained when peroxides are used; E values of PCL-based blends are greater than that measured for PCL

-

which, in turn, has a higher E value compared to PCLP; -

for the above blends higher values of moduli are observed in the presence

of peroxides; -

the tensile stress at break (OB) is generally lower for blends then for

unreacted polymers, much lower in the case of PCL - based blends; -

the tendency of fibre formation in PCL is generally suppressed in its

blends. Some general explanations of the above findings are presented in the following: o the addition of ductile PCL to stiff PHBV is likely to reduce its elastic modulus.

Moreover, we have previously seen how the presence of peroxide

increases the crystallization rate of PHBV; this, likely, leads to smaller spherulites and, thus, to more ductile matrices.

The same effect may help to

explain the strong reduction of E modulus in PHBVP, compared to PHBV; - the addition of stiff PHBV to ductile PCL, on the other hand, causes a great increase in the elastic mod'~.i.i. .~--. . . ":::~

"" ....

~-.~..... " matnx " blends. The presence

of the peroxide in these blends causes a further increase in the E values. This should be attributed to the presence of grafted species at the interface between the two polymers.

We must also note that the presence of peroxide causes a

decrease in modulus in plain PCL; this might be related to degradation occurring in PCL. In view of this, the increase of moduli in the blends with peroxide seems even more relevant;

563 -

the general reduction of the tensile stress at break

((~B) in blends can be

ascribed to an insufficient interracial strength between the two polymers. This in tum, strongly depends on the amount of gaffed species produced by the radical process. Moreover, in PCL-based blends, cold drawing is suppressed and this leads, in general, to very small values of ~B. Cold drawing is still possible in PCLP, but the degradation, likely induced by the peroxide, causes a decrease in CB of the resultant fibres. The increase in ~B for peroxide-treated PHBV may be attributed to the fast crystallization of PHBVP.

4.7 Fracture analysis

The morphology of specimens broken in tensile tests has been analyzed to understand the mode and the state in which the polymeric components are dispersed and their interaction during the tensile deformation. The appearance of selected blends is shown in Figs. 20 and 21. The broken surface of PCLPHBV 37 HM and PCLPHBV 37 HP are compared in Figs. 20a and 20b, respectively.

Fig. 20. Fracture surfaces of: a) PCLPHBV 37 HM, b) PCLPHBV 37 HP

564 It is evident that, when blended mechanically, PCL disperses in spherical domains that show no interaction with the surrounding matrix, as evidenced by the smooth cavities left over by the removed PCL particles. On the contrary, evidence of plastic deformation of PCL particles is frequently encountered in peroxide-treated blends (Fig. 20b). This explains the reduction of E values found for such blends, if compared to those without peroxide previously mentioned. The appearance of PCL matrix blends is shown in Figs. 21a and 21b, respectively, where the fracture surface of tensile bars of PCLPHBV 73 HM and PCLPHBV 73 HP are pictured.

Fig. 21. Fracture surfaces of: a)PCLPHBBV 73 HM, b)PCLPHBV 73 HP

The plastic deformation of PCL, still possible to some extent in mechanical blends, obstructs the detection of PHBV particles (see Fig. 20a).

On the

contrary, the peroxide treated blend (Fig. 20b) dearly shows the PHBV particles on the broken surface, as the cold drawing of the matrix PCL is completely suppressed. The addition of the dispersed phase to the matrix polymer seems to

565 be quite high and this explains the increase of elastic modulus in peroxide-treated blends (see previous paragraph)

5 Condusions As a concluding remark, we have shown how it is possible to perform reactive blending of two bioaffine materials by the thermal decomposition of peroxides added during processing. As expected, when PHBV, a hard polymer, is present, the drawing process of PCL matrix is deterred, while the modulus of the blend increases. On the other side, PCL imparts enhanced ductility to rigid PHBV matrix. Organic peroxides

added during melt processing reduce chemical

interactions between PHBV and PCL to form a graft copolymer (scheme 2);

R' + , ' , ' ~ O ~ C

H I I

O II

~C

H2--C ~

*~ RH + , " ~ O - - C - - C

I

R'

R'

,''~O--C--

I

9

0

H

II

CH2 ~ C ~

CH3

O--C--

I

~

{~H3 O--C--

~Nww, O

+ ~

I

O

II

CH2 ~C,,*w~

C - - Cl--I2 ~C,,,,,,,,~ I II

CH 3

O

0 II Cl--I2 ~ C ' " ' ~

CH3

O II

H2-C

566

~O--C--

9

i

0 0 li i~ CI-I2 --C~w~, + ~ , , ~ 0 - - (CI-12)4 - - C H - - C ~ w ~ ,

CH~

I H

1

O

CH 3

I

~O--C--

I

II

CH2 - - C ~

~,,,, 0 - - ( C H 2 ) 4 - - C - - C ~ II H

-

-

0

Solvent extraction reveals also the presence of crosslinked PHBV; the copolymer is formed in small quantities, as the interaction in the melt is

substantially restricted to the interracial regions between the two immiscible polymers.

We can thus confirm the hypothesis that the formation of graft

copolymers is responsible for important differences found between PHBV/PCL blends obtained by simple processing and reactive processing method.

567

System 3: solution blending of PHB 6 Introduction In the present section blends between PHBV0 and atactic elastomeric poly(epichlorohydrin) (aPECH) are described, obtained by dissolution of the blend components in a common solvent.

Although no chemical interaction is

recognized for the present system, chemical-physical interactions are effective enough to give a system well entangled. Preliminary mechanical investigation, still in progress, reveals that the system is promising. The present work reports the results concerning the crystallization, thermal behaviour, optical microscopy, phase structure and influence of aPECH on biodegradation of PHBV0/aPECH blends [35, 45-46]. For thermal and phase structure investigation, the binary blends were prepared by using a poly(D(-)3 hydroxybutyrate) (Zeneca, ICI group) with a M w - 166000 and M~/M~ = 2.85, and a poly(epichlorohydrin) (Aldrich), purified by centrifugation and filtration of a

dichloromethane solution,

with

a

Mw = 1078000 and M~M~ = 3.52. The films were obtained by solution casting from dichloromethane and then dried under vacuum at 80~

Blends with weight

ratios of 80/20, 60/40, 40/60, and 20/80 PHBV0/aPECH were prepared. For the biodegradative tests, films of pure PI-IB and PHBV0/aPECH blends, which had a thickness of 40 + 5~m, were prepared by solvent-casting techniques from dichloromethane solution using glass Petri dishes as casting surfaces.

Blend

films with weight ratios of 80/20, 65/35, 60/40, 50/50, and 40/60 were obtained. The solution cast films were aged for three weeks, at room temperature, to reach equilibrium crystallinity prior to analysis [47].

7.1 Morphological and thermal characterization The thermal properties of the PFIBV0 homopolymer and blends were analyzed by using a Perkin-Elmer DSC-4 with a Perldn-Elmer 3600 Data Station (TADS System). The glass transition temperatures (Tgs) were determined by

568 heating the quenched sample from 213 to 463 K, the values were taken at the midpoint of the transition. After quenching plato PHBV0 and PHBV0/aPECH blends are completely amorphous and exhibit a single glass transition temperature (T), whose value depends on composition. The appearance of a single T, suggests the presence of a single homogeneous amorphous phase, i.e. that the two components are miscible. The experimental values show a good agreement with the theoretical Fox values [48]

1 Wpner o W,a,ecn =~ + ~ Tg T, em~ro T,~Ee,,

(1)

where Wprmv0 and T~,rmv0, and W,eECHand T~,eECa are the weight fractions and the glass transition of PI-IBV0 and aPECH, respectively. The dependence of Tg on composition is shown in Fig. 22, where the solid curve was calculated using the Fox equation (Eq. 1).

273

b.?' 263

253 I

I

I

I

0.2

0.4

0.6

0.8

1.0

Mass fraction of PECH

Fig. 22. Glass transition temperature (Tg) versus composition of PHBV0/aPECH blends. Solid curve was calculated using the Fox eq. (1)

569

7.2 Morphology and spherulite growth rate The morphology and the radial growth rate (G) values of PHBV0 spherulites were studied employing an optical polarizing microscope with an automatic hot-stage Mettler model FP 82 controlled by Mettler FP80 Control Processor.

Blend samples were melted at 458 K for 1 rain and then rapidly

cooled to the crystallization temperature (To).

During the isothermal

crystallization process, the radii of the growing spherulites were measured as a function of time by taking photomicrographs at different intervals of time. Thin films of plain PHBV0 and PHBV0/aPECH blends, when observed under the optical polarizing microscope during the isothermal crystallization processes, show birefringent spherulitic structures.

After crystallization the

samples appear to be completely filled with impinged spherulites for all the aPECH concentration studied.

The spherulite dimension, at constant To,

decreases with increasing concentration of non-crystallizable component. The spherulite radius, R, increases linearly with time for plato PHBV0 and PHBV0/aPECH blends, for all Tr investigated. For all sample the isothermal radial growth rate, G = dR/dt, was calculated at different Tr

300 A

-== 9 E E .

200

-

-

i

:L

100 -

0

358

368

378

388

398

Tc (K)

Fig. 23. Spherulite radial growth rate (g) versus crystallization temperature (To) for PHBV0/aPECH blends. PHBV0 (wt-%): (o)100; (n)80; (.)60; (0)40; (m)20

570 As shown in Fig. 23, for a given T~, the addition of aPECH to PHBV0 causes a depression of the G values, thus allowing the control of the isothermicity of the PHBV0 crystallization at lower T c values. The experimental growth rate data were analyzed acr~rdmg to the polymerdiluent theory [49-51]. The equation describing the G values of a crystallizable polymer in a one-phase melt containing a second polymer acting as a diluent assumes the following form: U* In G - In ~2 + R(T~ - r~)

_0

.2T=~ ln~b2 = a = l n G o - AK g AT r, A r f

(2)

where Go is the pre-exponential factor that includes all terms that are taken as effectively independent of temperature. energetic

contribution

of diffusional

The U*/R(T c - Too) contains the processes

of the

amorphous

and

crystallizable material to the growth rate: U* is the sum of the activation energies for the chain motion in the melt of the crystallizable and non-crystallizable molecules and Too (Too = T g - C , where C is a constant) is the temperature below which such motion cease. TOm is the equilibrium melting temperature and 62 is the volume fraction of crystallizable polymer. The term f is a correction factor that takes into account the temperature dependence of the melting enthalpy All and is given empirically by f = 2To/(Tin + To). The term Ks contains the free energy required to form a nucleus of critical size, the heat of fusion and the T~ The slope of the straight lines obtained, for plain PHBV0 and PHBV0/aPECH blends, by plotting a versus 1/T c AT f, gives the Ks values reported in Table 12. Values of U* = 10.5 kJ mo1-1 and C = 51.6 K were chosen to give the best fit least squares lines through the data.

571

_

PHBV0/aPECH

i0 "3Kg (K2)

10 7 ~~ (J an "2)

100/0 80/20 60/40 40/60 20/80

4.3 3.5 2.9 2.1 1.5

38 32 27 20 14

_

_

Table 12. Values of Ks and c~o for plain PHBV0 and PHBV0/aPECH blends In our range of crystallization temperatures, Ks can be expressed as"

K

-- nb~

(3)

AHk where n is a variable changing according to the regime of crystallization [51], b0 is the distance between two adjacent fold planes, k is the Boltzmann constant, AH is the enthalpy of fusion per unit volume, and ~ and oe are the lateral and folding surface free energies.

At the undercooling used in this study, the PHBV0

crystallizes according to regime 1II [44], then, according to the Hoffman theory (50)the n variable was set equal to 4. The values of c~e reported in Table 12, were calculated with b0 = 5.76 A [52] and c~ = 0.25 b0AH [53]. The oevalue of 38 erg -2 cm calculated for plain PHBV0 is in good agreement with the value detemainod by measurements of lamellar thickness [44]. A depression of the % value with the increasing of the fraction of the non-crystallizable component in the blend was observed in other miscible blend systems [51, 54-57].

7.3 Isothermal bulk crystallization kinetics

The weight fraction, Xt, of the material crystallized at time t was calculated using the relation:

572

[[(dH / at )at X t ._

.1o-

(4)

Io (dH / dt)dt where the first integral is the heat generated at time t and the second is the total heat when crystallization is complete.

The isotherms of crystallization of

PHBV0 and PHBV0/aPECH blends, compared at the same T~, evidenced that the overall crystallization rate constant progressively decreases by increasing the amount of aPECH in the blend. In Fig. 24 the half-time of crystallization, t0.5 (defined as the time taken for half of the crystallinity to develop), is plotted against To values for some blend composition A

w

Io o

2000

-

1500

-

1000

-

500 -

-

0

I

353

I

373

I

393

Tc ( K )

Fig. 24. Half-time of crystallization (t0.5) versus crystallization temperature (to) for PHBV0/aPECH blends (symbols as Fig. 23)

573 The overall kinetic rate constant K. was calculated by using the Avrami equation [58]:

Xt - l - e x p ( - K t" )

(5)

where n is a parameter depending on the geometry of the growing crystals and on the nucleation process. PHBV0/aPECH

T c (K)

to. 5 (s)

n

100/0

373 378 383 388 393 396 398 401 403

60 82 140 228 488 665 850 1302 2216

2.0 1.9 2.0 2.2 2.1 2.1 2.0 2.0 2.3

1.35 x 6.98 x 2.32 x 8.37 x 1.71 x 8.97 x 4.82 x 2.11 x 7.28 x

10 -4 10 -5 10 -5 10 -6 10 -6 10 -7 10 -7 10 -7 10 -8

80/20

373 378 381 383 386 388 391 393 363 368 373 376 378 381 383

319 521 605 714 882 1209 1247 2498 280 382 561 796 988 1557 1695

2.2 2.2 2.2 2.2 2.4 2.3 2.3 2.1 2.4 2.5 2.5 2.5 2.5 2.5 2.7

1.81 x 6.05 x 4.34 x 3.00 x 1.87 x 9.26 x 8.64 x 1.84 x 4.59 x 2.10 x 7.93 x 3.28 x 1.90 x 6.03 x 4.86 x

10 -6 10-7 10 -7 10-7 10 -7 10 -8 10 -8 10 -8 10 -7 10 -7 10 -8 10 -8 10 -8 10-9 10-9

348 353 358 363 368 371 373

224 360 532 673 960 1320 1636

2.6 2.9 2.6 3.0 2.8 2.8 2.7

60/40

i

40/60

K n (s -n)

2.13 x 10 -7 5.76 x 10 -8 1.95 x 10 -8 1.02 x 10 -8 3.81 x 10 -9 1.57 x 10 -9 8.69 x 10 -10

Table 13. Values of t0.5, n and K, at various To values for plato PHBV0 and PHBV0/aPECH blends.

574 =.

X !

0-

V

t... "3"

O _.1 378

1

388

393 398 I

2

403 K I

3

Log [t (s)]

I

4

Hg. 25. Log[-ln(1-xt)] versus log t according to the Avrami equation for pure PHBV0 For each To the values of n and K~, reported in Table 13, were determined from the slope and the intercept, respectively, of straight lines obtained by plotting log[-ln (1-Xt) ] versus log t (Fig. 25). The Avrami exponent, n, is non-integral with a value between 2 and 3. Similar anomalous data were also observed in the case of poly(ethylene oxide) blended with poly(ethyl methacrylate) [51].

Contrary to the theoretical

prediction [58], in almost all cases the values of n are non-integral. This fact may be accounted for by mixed growth and/or surface nucleation modes and secondary crystallization, even if this latter process does not seem to occur in our case. Experimental factors such as erroneous determination of the zero time (time when the polymer starts to crystallize from the melt), and of the enthalpy of crystallization of the polymer at a given time can cause n to be non-integral [59]

575

7.4 Melting behaviour The rate of heat evolution during the isothermal crystallization was recorded as a function of time starting on samples melted at 458 K for 1 mm and rapidly cooled to the desired Tr

After crystallization the samples were heated to the -1

melting point at a scanning rate of 10 K rain.

The observed melting

temperature (T'm) was obtained from the maximum of the first endothermic peak. The d.s.c, curves of plain PHBV0 and PHBV0/aPECH blends isothermally crystallized, showed two melting peaks, with the peak appearing at the lower temperature corresponding to the melting of the original crystal of the isothermally crystallized sample. The second endothermic peak is caused by the melting of the reorganized crystal formed during the heating process [60-61]. In fact, as shown in Fig. 26, immediately after the first melting peak, an exothermic peak is registered. The lower Tm depends on the To while the higher T~ is almost constant [60].

A . . . . . . . .

~176

I

393

I

I

413

I

I

433

Temperature (K) Fig. 26. D.s.c curves of (A) pure PHBV0 and (B) 40/60 PHBV0/aPECH blend, isothermally crystallized at 373 K

576 The

T'm values of plain PHBV0 and PHBV0/aPECH blends linearly

increase with T~ for a wide range of undercoolmg. By increasing the fraction of aPECH, a depression of the

T'm values can be observed for every To explored

(Fig. 27).

453

433

,,.-E F....

413 348

378

438

408

Tc (K)

Fig 27. Observed melting temperature (T'm) of PI-IBV0/aPECH blends as a function of crystallization temperature (Tr (symbols as Fig. 2) The experimental data can be fitted by the equation of Hoffman [62]"

m

-

yT

+

1-

m

where 7 is the morphological factor and

(6)

TOm is the equilibrium melting

temperature. In Eq. 6 1/7 assumes values between 0 ( when T'm = TOmfor all To) and 1 (when

T'm = Tr

Therefore, the crystals are most stable at 1/? = 0 and

unstable at 1/y = 1. The extrapolated

TOm value is the lower the higher the

content of aPECH in the blend is (Table 14).

577 PHBVO/aPECH

TOm(K)

100/0

461 + 3

80/2O 6O/4O 40/60 20/80

455 449 441 432

i

+2 +2 + 1 +_2

i

Table 14. Equilibrium melting temperature values for plain PHBV0 and PHBV0/aPECH blends.

The 1/3' values are very similar (about 0.3) for plain PHBV0 and PHBV0/aPECH blends, i.e. independent on composition. It is of interest to note that calorimetric measurements of other polymer crystals also yielded comparable values for 1/?. Such a result indicates that aPECH is able to act as a diluent for PHBV0 and the two polymers are miscible in the melt phase [62-63]. The melting point depression according to the Flory-Huggins theory is related to the polymer-polymer interaction parameter ~ 12 according to the relation [63-64]:

i , ll 1t

- R V 2 T;....b

T;....e + m 2 +

1t 1

m,- ~1 : f l - Z 1 2 q ~

(7)

Subscripts 1 and 2 represent the non-crystallizable and the crystallizable polymer, respectively.

All is the perfect crystal heat of fusion of the

crystallizable polymer, V is

the molar volume of the polymer unit at the

equilibrium melting temperature, rn is the degree of polymerization, To~p and T~

are the equilibrium melting temperatures of the pure crystallizable

component and of the blend, respectively, r is the volume fraction of the components in the blend and R is the universal gas constant. The following parameter values have been used in our calculation: AH = 12.6 kJ mol ~ [65]; Vl=80 cm 3 mol -1 [66]; V 2 = 76 cm 3 mol -~ [66]; m 1 =

578 1742; m2 = 7565. Using the values of TOm and T'm reported in Table 14, the plot shown in Fig. 28 is obtained. The experimental points may be interpolated by a line with an intercept at the origin (13) and a slope (X~2) of-0.068.

The

negative parameter X~2 in the PI-IBV0/aPECH system should suggest that the two components can form a compatible mixture which is thermodynamically stable above the equilibrium melting temperature.

~40 x

2-

i

2

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

(x lo)

i

.

4

.

.

.

.

--'

"-

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

I

--

6

Fig. 28. 13versus ~2 according to Eq. 7 for PHBV0/aPECH blends The fact that the line does not pass through the origin can be ascribed to the morphological effect and/or to a composition dependence of 7~2. The equation derived by Kwei and Frisch [67] for non-infinite molecular weight polymers allows the proportionality constant for the morphological contribution, C, and X~2to be calculated:

579

q~lRTm,p

k, ml J

2m2

/

- - --/~'12Zm; 1

(8)

R

Plotting the left-hand side of Eq. $ versus Ol should yield a straight line where the intercept at the origin is C/R and the slope is

-Zl2T0m,b.

The results are

shown in Fig. 29. v v

.a.

150

-

100

-

.,,...

oE

E I,..-. ! e~

E

g-

.g

50
0

ols r

1.0

Fig. 29. Application of the Kwei-Frisch equation on melting temperature depression for PHBV0/aPECH blends. The left-hand side of Eq. $ is plotted versus r

From this one determines a value of 200 for C and a value of-0.054 for Z

12"

The X12 value depends on composition and its value is similar to that derived from the Flory-Huggins equation. By comparing the two terms on the fight-hand side of the Eq. 8, one finds that C/R is greater than the term X12 T~

r

and thus

it may be concluded that morphological effects strongly influence the melting point depression of PHBV0/aPECH blends (Table 15).

580 PHBV0/aPECH

C/R

100/0 80/20 60/40 40/60 20/80

101 101 101 101 101

Z12 T~

t~1

~12 at TOm

0 -4.50 -9.22 -14.19 -19.43

-0.054 -0.055 -0.056 -0.057 -0.058

Table 15. Data from Kwei-Frisch equation (equation 8). 7.5 Structural characterization

The samples for the study of the phase structure were prepared on the basis of the thermal properties obtained from d.s.c, data.

Plato PHBV0 and

PHBV0/aPECH blends were melted at 573 K for 2 rain, then isothermally crystallized at 373 K. Then all isothermally crystallized samples were annealed for 120 rain at the annealing temperature (T) calculated as T = T' m- 19 K. The times of isothermal crystallization and the T~,w~o~are reported in Table 16. PHBV0/aPECH

t (min)

T a (K)

100/0 80/20 60/40 40/60

30 40 55 227

414 413 409 404

Table 16. Time used for isothermal crystallization at 373 K (t) and annealing temperature (Ta) for plain PHBV0 and PHBV0/aPECH blends.

The isothermally crystallized and the annealed samples, when observed under optical polarizing microscope, appear to be completely filled with impinged spherulites.

For all blends no aPECH-separated domains were

observed neither in the intraspherulitic regions nor in the interspherulitic contact zones. D.s.c. scans conducted for isothermally crystallized plato PHBV0 and PHBV0/aPECH blends, showed that for pure PHBV0 the T g of 277 K was no

581 longer detectable, while for the PHBV0/aPECH blends a single glass transition appeared at a temperature value close to the pure aPECH T g of 253 K. The T g event was more pronounced with increasing aPECH fraction.

7.6 Wide-angle X-ray diffraction The WAXD pattems were recorded on a Siemens diffractometer model D500 (Cu K~ radiation, ~, =1.54 A). The degree of crystallinity was calculated from diffracted intensity data in the range 20 = 11-40 ~ by using the area integration method [68].

The amorphous contribution was calculated from

diffracted intensity data of plain PHBV0 and blends quenched in liquid nitrogen. Lattice imperfections were not considered.

The apparent crystal sizes were

calculated from the line-broadening data collected with a scanning rate of 0.1 2 0 -1

deg min . A nickel standard sample was employed to determine the mstnunental broadening. The data reported in Table 17 show that the crystallinity x~ of the blends decreases with increasing aPECH fraction, while the crystallinity of the PHBV0 component X~,HBV0 is similar for plato PHBV0 and 80/20 and 60/40 PHBV0/aPECH blends, and lower for the 40/60 blend. i

PHBV0/ Isothermally aPECH c~stallized x~a,i-mvQ,, x~ 100/0 0.54 80/20 0.42 0.52 60/40 0.33 0.55 40/60 0.19 0.46

Annealed Xr 0.65 0.49 0.39 0.24

X~ 0.61 0.64 0.61

Isothermally crystallized (020t .(1101 208 152 271 158 289 211 284 234

Annealed (020/ 254 289 352 323

(1 i01 202 228 276 301

i

Table 17. Crystalline weight fractions x~ , crystallinity of PHBV0 component X~HBV0 and apparent crystal sizes Dhkl (A) of plain PHBV0 and PHBV0/aPECH blends.

582 The annealing produces a general enhancement of the crystallinity and the Xr

index reaches similar values for plain PHBV0 and all blends. The apparent crystal sizes Dhkl of PHBV0 in the direction perpendicular to

the (110) and (020) crystallographic planes were calculated using the Scherrer equation [68]: -

-

/80 COS0hM where 130 is the half width in radians of the reflection corrected for the instrumental broadening, and k is the wavelength of the X-ray radiation employed. The shape factor K is set equal to unity, and so the size data have to be considered as relative data [68]. The (002) reflection is disturbed by overlapping of other diffractions, and therefore the crystal dimension along the c axis cannot be calculated. The lateral crystal sizes of PHBV0, reported in Table 17, increase with increasing aPECH fraction in the blends.

This fact may be justified by the

decreased PHBV0 crystallization rate in the presence of aPECH.

In fact, the

choice of a constant temperature (373 K) enhanced the time of crystallization with increasing aPECH fraction (Fig. 14). The annealing treatment increases the above mentioned crystal dimensions and the aPECH presence seems not to interfere with the lateral increment of PHBV0 crystal dimensions.

7.7

Small-angle X-ray scattering.

SAXS data were measured at 25~ using a Huber 701 chamber [69] with a monochromator glass block. Monochromatized CuKa X-rays (X = 1.54 A) were employed. The intensity was counted at 85 angles of measurement in the range 20 = 0.1-3.8 with three different step intervals. To reduce the statistical error of counting, for each sample the mean intensity values were obtained from 16 scans

583 with a time of 21 h for a complete measurement.

The standard deviations

calculated for the intensity values at each counting angle showed a low degree of spread of the intensity data around the average values. The raw data were first corrected for sample adsorption and then the background was subtracted. By application of the indirect transformation method developed by Glatter [70-71],

the

corresponding

three-dimensional propagated

correlation

statistical

error

functions band

together

were

with

calculated

the from

unsmoothed and smeared experimental scattering data. For the desmearmg the geometries of the incident beam profile and of the detector were considered [70,72]. The scattering profiles of plain PHBV0 and some PHBV0/aPECH blends show the presence of a maximum, which is associated with the long period, L, resulting from the presence of macrolattice formed by centres of adjacent lamellae. For all S AXS measurements the abscissa variable, Q, was calculate by:

Q - 4rc

senO

2

(lo)

After desmearing the intensity were Lorentz corrected and the L values were calculated by:

27s L =

Qm where Qm is the abscissa value at the maximum of the plot (Figure 30). For the isothermally crystallized 80/20 and 60/40 PHBV0/aPECH blends, with respect to plain PHBV0 the peak position slightly shifts towards higher Q values and the peak broadens as more aPECH is added to PHBV0. 40/60 PHBV0/aPECH blend the peak does not appear at all.

For the

The annealing

produces an enhancement and a better resolution of the peak for plain PHBV0, while for blends the peak resolution does not improve at all.

584

:5 V A

0 V

0 ....I

-e-O 4-0 --e-O 4-0

!

0

,5 10 10 2 x Q ( A oI)

!

15

Fig. 30. Small angle X-ray scattering profiles of PHBV0/aPECH blends: (O) 100/0; (e) 80/20; ( , ) 60/40; (e) 40/60 In Table 18 the L values are reported for the isothermally crystallized and annealed samples. PHBV0/aPECH 100/0 80/20 60/40

Isotermall~rc~stallized 79 72 72

Annealed 93 93 82

i

Table 18. Long period distances (A) of plain PHBV0 and PHBV0/aPECH blends. A general enhancement of the long-period values can be observed as a consequence of the annealing treatment and for the blend samples the long-period distance remains slightly lower than that of plain PHBV0. The fact that the L values do not increase but are actually slightly lower for blends than for plain PHBV0 support the hypothesis that the aPECH is absent from the interlamellar PHBV0 zones. The presence of aPECH in those regions

585 could only be indicated by a very low PHBV0 lamellar thickening with blending, which is very unlikely given the crystallinity and lateral crystal dimension data obtained by WAXD.

In fact, when isothermally crystallized, the PHBV0

component reached about the same crystallinity value with blending as in the pure state.

Besides, it was observed that the lateral crystal sizes of PHBV0

increased with blending. The absence of the aPECH in the mterlamellar PHBV0 zones was confirmed by TEM measurements (not reported) which showed that the thickness of PHBV0 lamellae and of amorphous interlamellar zones are substantially the same in the plain PHBV0 and in the 60/40 PHBV0/aPECH blend. These observations together with the fact that the optical microscopy has shown the system to be completely volume filled with spherulites, and no segregation of aPECH component was observed in the interspherulitic contact zones, suggests that the non-crystallized component is segregated in the interfibrillar zones, which are larger than the interlamellar regions but smaller than the overall spherulite. To explain the scale of rejection of aPECH in the PHBV0 crystallizing matrix during crystallization from the melt, the Keith and Padden equation can be used. i.e. 8=D/G [73]. This expression places the scale of segregation on a some what quantitative basis, where 8 is the dimensional order of segregation, D is the diffusional coefficient of the non-crystallizing component in the crystallizing matrix, and G is the spherulitic growth rate. The parameter 8 has dimensions of length and represents the distance that the rejected component may move during the time of crystallization. If 8 is comparable with interlamellar distances, then the rejected component may reside between lamellae. As reported in the "Morphology and spherulite grow rate" paragraph, the presence of aPECH decreases the PHBV0 spherulite growth rate G, and the aPECH T g value is lower than that of crystallizing PHBV0. These facts give a relative high diffusion term D in the ~5parameter. The aPECH molecules diffuse away from the front of PHBV0 crystallization at such a rate so as not to remain

586 segregated between lamellae, but in any case the mobility of aPECH is not sufficient to let it move away from the spherulites. Thus, aPECH will reside in the interfibrillar zones. Considering the high aPECH M w value, a monomolecular dispersion in the interfibrillar zones of aPECH molecules with a random-coil conformation can produce scattering. In fact the volume filled by one molecule of aPECH having a weight-average molecular weight of 1078000 and a specific volume of 3

0.735 cm g

-1

has been deduced, and by the assumption of spherical shape a

radius of about 70 A was calculated. The presence of scattering arising from the dispersed aPECH molecules finds a first confirmation in the results obtained from the annealed samples. As demonstrated by WAXD investigations about the crystallinity and crystal dimensions of PHBV0, the annealing treatment produces for plain PHBV0 and blends a rearrangement and perfectioning of the crystalline region. In fact, for plain PHBV0 the profile of the Lorentz plot were better resolved after the annealing treatment (Fig. 30).

However, for the 80/20 and 60/40 blends an

increase in broadness was observed after the annealing treatment. This poorer resolution can be explained by the interference of the scattering arising from aPECH molecules dispersed in the interfibrillar zones, and this interference seems to increase with the annealing treatment. In the case of the 40/60 PHBV0/aPECH blend, for both isothermally crystallized and annealed samples the contribution to the scattering of the aPECH inhomogeneity in the system is such that the application of the Lorentz approach to the experimental scattering profile is compromised. In order to analyse the scattering arising from the presence of aPECH in the interfibrillar zones of blends the Debye-Bueche relation was used. This relation is generally applicable to the scattering intensity from the electron density fluctuation of an inhomogeneous system, it can be written as [74]:

587 [I(Q)]

-112 __

[ g 3<.2 >/3 ]-1/2[1 -[- Q213]

(12)

where (n 2) is the mean-square density fluctuation in the system, 1r is the correlation length of the fluctuation and K s is a proportionality constant. For the scattering from an inhomogeneous system with correlation length 1r [I(Q)]

2

-1/2

is

linear when plotted against Q and 1'2r can be estimated from the ratio of slope to intercept. In Fig. 31 the Debye-Bueche plots for blend samples are given. It can be observed that at small Q2 values the plots are nearly linear and they deviate from linearity at larger Q2. The deviation is caused by the addition of the scattering from crystalline-amorphous region, while the lmearity at smaller Q2 represent scattering due to the inhomogeneity of the system [75].

In fact, for the

isothermally crystallized blends the lmearity continues to larger Q2 as the PHBV0 fraction decreases, giving the indication that in the blends besides the crystalline region there exist some amorphous inhomogeneity.

1

0

I

2

0

|

i!! -.5

.20 1 ",

3

I

.4

15

.3 .10

.2

.05 0

.25

.75 103 x Q2

!

1.25

0

Fig. 31. Debye-Bueche plots of PHBV0/aPECH blends. Isothermally crystallized; (e) 80/20; (~) 60/40; (m) 40/60: Annealed: (O) 40/60. The axes of this last curves are indicated by the arrows

588 In Fig. 31 the curve of the 40/60 PHBV0/aPECH annealed blend is also shown and it can be noted that the linearity continues to a Q2 value higher that for the 40/60 PHBV0/aPECH isothermally crystallized blend. The contribution of aPECH to the scattering increases as a consequence of the annealing treatment. The 1o values calculated for all blends do not significantly vary with the aPECH content or annealing treatment; the average of the obtained 1c is 45+5 A. These results indicate that no aggregation phenomena occur with increasing aPECH content in the blends. For the 40/60 PHBV0/aPECH blends, which showed a greater contribution by aPECH to the scattering a three-dimensional analysis of the scattering profile was conducted with the computation of the three-dimensional distance distribution function p(r) [70,72].

Gaussian curves were obtained, indicating

that the scattering is due to globular particles of aPECH and there was a good correlation between the scattering profiles approximated by the program and the experimental results [70]. In Fig. 32 the three-dimensional distance distribution plot for the 40/60 PHBV0/aPECH isothermally crystallized blend is reported. .8-

Q.

.4-

I

I

100

I

I

200

r(,~) Fig. 32. Three-dimensional distance distribution function p(r) for the isothermally crystallized 40/60 PHBV0/aPECH blend

589 The radius of gyration of the whole particle Rg, was calculated from the p(r) function by [70,72]:

R

~ =

*

03)

2Iop(r)d r

The Rg values obtained were about 70 A for both 40/60 isothermally crystallized and 40/60 annealed PHBV0/aPECH blends.

These values give a

measure of the dispersion of aPECH molecules with globular symmetry in the interfibrillar zones, and with the reference to the aforementioned calculated dimensions of a single aPECH molecule, aPECH appears to be dispersed at molecular level. It must be pointed out that the polydispersity of the system and the difficulties in separating the scattering that arises from the crystalline region make the experimental measure of aPECH dispersion merely an indicator of the order of magnitude of dispersion.

8 Biodegradation study PHBV0 degrading bacteria were obtained from samples of garden soil. The most rapidly growing strain was isolated by plating the cultured bacteria on agar medium containing powder of plain PHBV0 as the only carbon source.

This

Gram+ strain was characterized and tentatively designated as Aureobacterium Saperdae. Well grown A. Saperdae cultures were prepared in mineral medium containing PHBV0 as the only carbon source (mineral medium composition: 1 mg/mL of NH4C1, 0.5 mg/mL of MgSO4.7H20 and 0.005 mg/mL of CaC12.2H20 in 66 mM KH2PO 4 (pH = 6.8)). For bacterial degradation studies about 5 mL aliquots of this culture were used to inoculate 500 mL flasks containing 100 mL of mineral medium, in order to obtain 0.1 value of optical density at 540 nm (O.D.540).

To obtain comparable results, film samples

590 calculated to contain 150 nag of PHBV0 were added to each flask, then the flasks were incubated at 30~ under shaking. For the biodegradability tests, fihns of pure PHBV0 and PHBV0/aPECH blends, which had a thickness of 40+_5 ~m, were prepared by solvent-casting techniques from dichloromethane solution using glass Petri dishes as casting surfaces.

Blend films with weight ratios of 80/20, 65/35, 60/40, 50/50 and

40/60 were obtained. The solution-cast films were aged for three weeks, at room temperature to reach equilibrium crystallinity prior to analysis [47]. The

bacterial

attack

was

conducted

using

plain

PHBV0

and

PHBV0/aPECH blend films obtained from solvent-casting technique that allowed the preparation of relatively large quantities of films with homogeneous thickness.

Preliminary tests carried out with pure aPECH revealed that A.

Saperdae could not use this polymer as the sole carbon source. In Fig. 33 the O.D. 540values of the A. Saperdae cultures are plotted against the incubation time for plain PHBV0 and blends of different PHBV0/aPECH weight ratios. E

~ 1

o

~.. 1.0

v

e-

@ "0

._~

_

0

_ 0.1 Time in d

Fig. 33. Growth curves of A. Saperdae cultures where the only carbon sources was: (m) pure PHBV0; (O) 80/20; (~) 65/35; (n)60/40 and (o) 50/50 PHBV0/aPECH blends. Growth was completely inhibited with the 40/60 PHBV0/aPECH blend

591 The reported O.D.540 data are average values obtained by several culture experiments. It can be observed that the bacterial growth rate decreases in the presence of aPECH and decreases further with increasing aPECH content. Little growth occurred with the 50/50 blend and it was completely inhibited with the 40/60 blend. The decrease in the bacterial growth rate with blending could have been a result of the dispersion between the two polymers, which results in the dilution of the PHBV0 molecules on the film surface. After the stationary phase of bacterial growth was reached, the degraded films were extracted and the percentage of weight loss determined (Tab. 19). PHBV0/aPECH 100/0 80/20 65/35 60/40 50/50 40/60 i

Weig~htloss in % 100 63 37 11 3 0

Table 19. Average percentage of weight loss of blend films at the stationary phase of bacterial growth The percentage of weight loss for blend films decreased with increasing aPECH fraction, while plain PHBV0 was completely degraded.

In addition

control experiments were run to verify chemical hydrolysis of polymeric films immersed in mineral medium at 30~

After 15 d no weight loss of the films was

revealed. Since extracellular PHBV0 depolymerases have been found in bacterial broth of some microorganisms such as some bacteria (Pseudomonas Iernoignei [76], Alcaligenes faecalis TI [77], Comamonas sp. [78], and Pseudomonas pickettii [79]), and a fungus (Penicillum funiculosum [80]), cultures of A. Saperdae were assayed for PHBV0 depolymerase activity.

An aliquot of A.

Saperdae culture grown to the end of the exponential phase on mineral medium

592 containing PHBV0, was centrifuged at 10.000 rpm in order to eliminate the cells. The presence of PHBV0 depolymerase activity in the supematant was assayed at 45~

following the decrease in turbidity of a stable suspension of PHBV0

granules. The assay mixture (2.5 mL) contained supematant aliquots, 125 pmol of Tris-HC1 buffer (pH 8.0) and 2 mg of PHBV0 granules. The reaction was started by the addition of the PHBV0 granule suspension and followed by the decrease in turbidity at 650 nm, and it was calculated that 1 mL of supematant produced the degradation of 80 ~g PHBV0/mm, demonstrating that degradation still occurs in the absence of cells. In order to verify the stability of the aPECH component during the bacterial degradation, the composition of some degraded blends was determined by 1H NMR analysis. 1HNMR analyses was used to determine the composition of the biodegraded

blend samples and was carried out on a Brucker AM-500

spectrometer.

The 1H NMR spectra were recorded in CDC13 solutions of the

blends. The chemical shifts were referred to tetramethylsilane used as mtemal standard.

An average of

32 scans was accumulated, and the quantitative

analysis was carried out by the evaluation of the integrals of the signals. The results were reported in Tab. 20.

PHBV0/aPECH Weight loss PHB~0/aPECH PHBV0/aPECH initial in % b~r H NMR calculated 80/20 75 20/80 20/80 80/20 63 45/55 46/54 65/35 41 40/60 41/59 65/35 37 44/56 45/55 60/40 11 57/43 55/45 Table 20. Composition of degraded blend films at the stationary phase of bacterial growth. Assuming that aPECH was not attacked, new PHBV0/aPECH weight ratios of the degraded blends were calculated referring the % of weight loss to the initial

593 PHBV0 content of the blends. These calculated weight ratios, reported in Tab. 20, are coincident with those obtained by 1H NMR analysis confirming that during the bacterial degradation of the blends the aPECH component is untouched and/or that there is not any abiotic aPECH release. If this was not the case, the 1H NMR weight ratios would be richer in PHBV0 content than the calculated ones. Plain PHBV0 film samples degraded at different % of weight loss were prepared by stopping A. Saperdae cultures at different times, on the degraded samples the number-average molecular weight (Mn) value was determined by GPC analysis.

As shown in Tab. 21 the M n values remained relatively

unchanged during bacterial degradation.

Weight loss m% 0 22 48 61 94

Mn

Mw

Mw/M n

58400 45800 42500 45000 60000

166400 149000 142000 184000 184000

2.85 3.26 3.55 3.43 3.07

Table 21. Average molecular weights of plato PHBV0 degraded to different percentages of weight loss. This result indicates that the degradative enzymes act on the surface layer of the film and polymer erosion proceeds via surface dissolution. Analogous results were obtained by Doi et al. [81] for PHBV0 biodegradation with extracellular PHBV0 depolymerase isolated from Alcaligenes faecalis The GPC analysis of blend films revealed two partially overlapping peaks of the PHBV0 and the aPECH components. No shifts of the peaks were observed after the degradation treatment, showing that probably no important changes of the PHBV0 Mn values took place in the degraded blend samples. further evidence that no hydrolytic degradation took place.

This is a

594 The WAXD crystallinity index (x~) of solution-cast films of pure PHBV0 and blends, before and after bacterial degradation, are reported in Tab. 22, together with the crystallinity index of the PHBV0 component (X~HBV0). PHBV0/aPECH initial 100/0 80/20 65/35 60/40 50/50 40/60

Weight loss in % 100 63 37 11 3 0

Before degradation x~ x~yrmvo 0.55 0.46 0.58 0.36 0.55 0.34 0.56 0.30 0.60 0.23 0.57

Aider degradation x~ xo~,rmvo 0.54 0.28 0.60 0.26 0.57 0.31 0.56 0.29 0.60 n.d. n.d.

Table 22. Crystalline weight fraction x c and crystaUinity of the PHBV0 component x c PHBV0, for plain PHBV0 and PHBV0/aPECH blends, before and after bacterial degradation. The xr values of unattached polymer decreases with increasing the content of uncrystallizable aPECH component, while the x~, vrmv0 values are similar for plain PHBV0 and blends.

After bacterial attack, the x~ values of the 80/20,

65/35 and 60/40 PHBV0/aPECH blends are lower in direct accordance with the new PHBV0/aPECH weight ratios shown in Tab. 20 for these degraded blends They are richer in aPECH content as a consequence of the PHBV0 degradation. In contrast, the xr vnav0 values did not change in the degraded blends. Therefore, it may be concluded from the WAXD data that the weight loss of the blend films is solely due to the loss of the PHBV0 fraction in agreement with the assumption made above. The SEM analysis (not reported) of the film surfaces revealed surface erosion of plain PHBV0 and PHBV0/aPECH blends after bacterial attack. The film thickness was determined for plain PHBV0 and 80/20 blend before attack and after a 60% of weight loss. A corresponding decrease of plain PHBV0 film thickness of approximately 60% was measured indicating a surface erosion caused by the degradative enzymes. The observed decrease in both percentage of

595 weight loss and percentage of film thickness, was also obtained by Doi et al. [81] with PHBV0 films degraded by Alcaligenes faecalis depolymerase. A different result was observed for the 80/20 PHBV0/aPECH blend: aider the attack the film thickness remained unchanged, while large cavities (several ~tm) were observed inside the film. Highly degraded 80/20 film (where only 25% of the PHBV0 initially present in the blend is left) maintains the initial dimensions probably due to the aPECH dispersion in the blend. In fact, the composition analysis of the degraded blends revealed that there were neither degradation nor abiotic release of aPECH during the bacterial attack. The intemal large cavities observed in the degraded blend probably could be zones where PHBV0 was extensively degraded and where aPECH was collapsed onto the remaining structure after bacterial attack.

9 Conclusions

Miscibility in the melt between PHBV0 and aPECH is confirmed by the detection of a single glass transition and by the influence of aPECH on the spherulite growth rate and on the overall crystallization rate. The depression of the equilibrium melting temperature appeared to be strongly influenced by morphological effects. After PHBV0 crystallization from the melt the aPECH molecules are rejected into the interfibrillar zones where they probably assume a random-coil conformation. In spite of the molecular dispersion of aPECI-I, the detection for the crystallized blends of a glass transition temperature value close to the one of aPECH, indicates weak interactions at the segmental level between the two polymers.

The trend of aPECH molecules to assume a random-coil

conformation together with the high molecular weight can hinder the close contact between aPECH and PHBV0 molecules in the continuous interfibrillar zone.

The annealing treatment promotes a general perfec~ionmg and

rearrangement of the sample morphology, enhancing the crystallinity and the

596 crystal dimensions of PHBV0 in the pure state and in the blends, and probably favouring the trend of the aPECH molecules to assume a globular conformation. A. Saperdae is a Gram positive bacterium which extracellulary degrades PHBV0 blended with aPECH.

The biodegradability of blended PHBV0

decreased but was not completely compromised by the presence of aPECH component. Bacterial attack of blends containing relatively low aPECH percentages causes a drastic reduction of the PHBV0 content, while the dimension of the film remained unchanged due to the aPECH dispersion inside the blend. In fact, as a consequence of bacterial attack neither degradation nor abiotic release of aPECH was observed.

References

1. Sperling, "Interpenetrated Polymer Networks and Related Materials", Plenum Press, New York (1981) 2. Yu.S. Lipatov and L.M. Sergeeva, "Interpenetrated Polymer Networks", Naukova Dumka, Kiev (1979) 3. D. Klempner, Angew. Chem, 90, 104 (1978) 4. D. Klempner and D.C. Frisch, eds., "Polymer Alloys II", Plenum, New York (1980) 5. M. Narkis, "Reactive polymer blending by free radical processes", in "European Regional Meeting", PPS, September 1991, Palermo (Italy) 6. F. Pilati,, E. Marianucci, C. Berti, J. App. Polym. Sci., 30, 1267 (1985) 7. L.H. Wang, Z. Huang, T. Hong, R.S. Porter, J. Macromol. Sci., Phys., B29, 155(1990) 8. L.Z. Pillon, L.A. Utracki, in "Ecole d'Application des Hautes Polymeres, 3eme Conference sur la reactivite chimique des polymeres en course de transformation", Strasbourg, France (1984), p. 57

597 9. S. Cimmino, F. Coppola, L. D'Orazio, R. Greco, G. Maglio, M. Malinconico, C. Mancarella, E. Martuscelli, G. Ragosta, Polymer, 27, 1874 (1986) 10.G. Illing, Kunststoffe, 1, 275 (1968) 11.P. Laurienzo, M. Malinconico, E. Martuscelli, M.G. Volpe, Polymer, 30, 835 (1989) 12.F. Pilati, P. Manaresi, B. Fortunato, A. Munari, V. Passalacqua, IUPAC 26th Symposium on Macromolecules, Mainz 1979, Preprint vol. 1, 231 13.B. lmmirzi, M. Malinconico, E. Martuscelli, Polymer, 32, 1649 (1991) 14.P. Laurienzo, M. Malinconico, G. Ragosta, M.G. Volpe, Angew. Makromol. Chem., .!70, 137 (1989) 15.G.F. Freeguard. M. Karmarkar, J. Appl. Polym. Sci., 15, 1649 (1971) 16.G.F. Freeguard. M. Karmarkar, J. Appl. Polym. Sci., 16, 69 (1972) 17.E.A. Dawes, P.J. Senior, Adv. Microb. Phys., 10, 135 (1973) 18.P.A. Holmes, L.F. Wright, S.H. Collins, Eur. Pat. Appl., 54459 (1981) 19.E.R. Howells, Chem.lnd. (London), 7, 508 (1982) 20.P.P. King, J. Chem. Technol. Biotechnol., 32, 2 (1982) 21 .M. Avella, E. Martuscelli, Polymer, 29, 1731 (1988) 22.N. Lotti, G. Ceccorulli, M. Pizzoli, M. Scandola, "Miscele di PHB con polimetilmetacrilati", in "X Convegno Italiano di Scienza e Tecnologia delle Macromolecole", Ferrara, 6 ottobre 1991 23.E. Martuscelli, P. Guarrata, A. Seves, A. Siciliano, "Leg,he polimeriche a base di PHB caratterizzate da elevata resistenza al calore", CNR Italian Patent pending, 506/91 24.G.L. Brode, S.V. Koleske, J. Macromol. Sci., Chem. A6, 1109 (1972) 25.S. Iannace, N. De Luca, L. Nicolais, C. Carfagna, S.J. Huang, J. Appl. Polym. Sci., 41,2691 (1990) 26.IT 20293A (1990), Invs: E. Martuscelli, M. Avella, M. Malinconico, G. Ragosta

598 27.Lovell, T.H. Shah, F. Heatley, "Chain transfer to polymer in emulsion polymerization of n-butil acrylate studied by 13C nmr spectroscopy and GPC", Polym. Comm., 32, 98 (1991) 28.J.A. Moore, "Macromolecular synthesis", Coll. Vol. 1, p.23 (1977), J. Wiley, Eds, N.Y. 29.IT 95A/000051 (1995), Invs: M. Avella, M.E. Errico, B. lmmirzi, M. Malinconico, E. Martuscelli, A. Pudia 30.K.H.S. Bagdasar'yan, "Theory of free radical polymerization", Israel Program for Scientific Translations, Jerusalem, p. 187 (1968) 31.A.J. Kinlock, R.J. Young, Fracture behaviour on Polymers, Applied Science, London, p. 74 (1983) 32.F. Coppola, R. Greco, G. Ragosta, J. Mater. Sci., 21, 1775 (1986) 33.ICI Bulletin MLB 8710, Technical bulletin on MBL Biopol 34.D.L. Hunston, A.J. Kinloch, S.J. Shaw, S.S. Wang, Proc. Am. Chem Soc. Meeting, Kansas City, p. 320 (1982) 35.P. Sadocco, C. Bulli, G. Elegir, A. Seves, E. Martuscelli, Makromol. Chem., 194,2675(1993) 36.u Doi, in "Microbial Polyesters", VHC Publisher Inc., 1990 37.M. Avella, B. Immirzi, M. Malinconico, E. Martuscelli, G. Orsello, A. Pudia, G. Ragosta, Die Angew. Makromol. Chem., 205, 151 (1993) 38.R.D. Fields, F. Rodriguez, R.K. Finn, J.Appl. Polym. Sci., 18, 3571(1974) 39. C.V. Benedict, W.J. Cook, P. Jarrett, J.A. Cameron, S.J. Huang, J.P. Bell, J. Appl. Polym. Sci., 28, 327(1983) 40.C.V. Benedict, W.J. Cook, S.J. Huang, J. Appl. Polym. Sci., 28, 335, (1983) 41.Y. Doi, Y. Kanesawa, M. Kunioka, T. Saito, Macromol., 23, 26 (1990) 42.P.A. Holmers, Phys. Technol., 16, 32 (1985) 43.F.P. Delofield, M. Dondoroff, N.J. Pellezzani, C.J. Lusty, R. Contopoulos, J. Bacteriol., 90, 1455 (1965)

599 44.P.J. Barham, A. Keller, E.L. Otun, P.A. Holmes, J. Mater. Sci., 19, 2781 (1984); P.J. Barham, ibid, 19, 3824 (1984) 45.E. Dubini Paglia, P.L. Beltrame, M. Canetti, A. Seves, B. Marcandalli and E. Martuscelli, Polymer, 34, 996, (1993) 46.P. Sadocco, M. Canetti, A. Seves and E. Martuscelli, Polymer, 34, 3368 (1993) 47.S. Bloembergen, D. A. Holden, G. K. Hammer, T. L. Blum and R. H. Marchessault, Macromolecules, 19, 2865 (1986) 48.T.G. Fox, Bull. Am. Phys. Soc., 2, 123 (1956) 49.J. Boon and J. M. Azcue, J. Polym. Sci. A2, ,6,885 (1968) 50.D. Hoffman, Polymer, 24, 3, (1982) 51.S. Cimmino, E. Martuscelli, C. Silvestre, M. Canetti, C. de Lalla and A. Seves, J. Polym. Sci., Polym. Phys. Edn, 27, 1781, (1989) 52.M. Yokouchi, Y. Chatani, H. Tadokoro, K. Teranishi and H. Tani, Polymer, 14, 267 (1973) 53.J.D. Hoffman, Polymer, 26, 803 (1985) 54.E. Martuscelli and G.B. Demma, "Polymeric Blends: Processing Morphology and Properties", Plenum Press, New York, 1980 55.E. Martuscelli, M. Pracella, Wang Ping Yue, Polymer, 25, 1097 (1984) 56.S. Cimmino, E. Di Pace, E. Martuscelli, C. Silvestre, Makromol. Chem Rapid Commun., 9, 261 (1985) 57.E. Martuscelli, C. Silvestre, C. Gismondi, Makromol. Chem., 186, 2161, (1985) 58.J. Avrami, J. Chem. Phys., _7, 1103, (1939) 59.D. Grenier and E. R. Prud'homme, J. Polym. Sci., Polym. Phys. Edn, 18, 1655(1980) 60.H. Mitano, P.J., 19, 1241, (1987) 61.T. Arakawa, F. Nagatoshi and N. Arai, J. Polym. Sci., Polym. Phys. Edn., 7, 1461, (1969)

600 62.D. Hoffman and J.J. Weeks, J. Res. Natl. Bur. Std., 66, 13, (1962) 63.T. Nishi, T. Wang, Macromolecules, _8, 909, (1975) 64.P.J. Flory "Principles of Polymer Chemistry", Comell University Press, Ithaca 1953 65.P. Greco, E. Martuscelli, Polymer, 30, 1475 (1989) 66.D.W. Van Krevelen, "Properties of Polymer", Elsevier, Amsterdam, 1990 67.T.K. Kwei and H. L. Frisch, Macromolecules,, 11, 1267, (1978) 68.L.E. Alexander, "X-ray Diffraction Method in Polymer Science", WileyInterscience, New York, 1969, p. 137 69.E.Schnabel, R. Hosemann and B. Rode, J. Appl. Phys., 43, 3237 (1972) 70.O.Glatter and O. Kratky, "Small Angle X-ray Scattering", Academic Press, London, 1982, p. 119 71. O. GlaRer, J. Appl. Cryst., 13,577 (1980) 72. O. GlaRer, J. Appl. Cryst., 10, 415 (1977) 73.H.D. Keith and F. J. Padden, J. Appl. Phys., 35(4),1270, (1964) 74.P. Debye and A. M. Bueche, J. Appl. Phys., 20, 518, (1949) 75.S. Nojima, Y. Terashima and T. Ashida, Polymer, 27, 1007, (1986) 76.K. Nakayama, T. Saito, T. Fukui, Y. Shirakura, K. Tomita, Biochim. Biophys. Acta, 827, 63, (1985) 77.T. Tanio, T. Fukui, Y. Shirakura, T. Saito, K. Tomita, T. Kaiho, Eur. J. Biochem., 124, 71, (1982) 78.D. Jendrossek, I. Kajoke, R.B. Habibian, A. Steinbuchel, H.G. Schlegel., J. Environ. Polym. Degrad., 1, 53 (1993) 79.K. Yamada, K. Mukai, Y. Doi, Int. J. Biol. Macromol., 15, 215, (1993) 80.C.L. Brucato, S.S. Wong, Arch. Biochem. Biophys., 290, 497 (1991) 81.Y. Kumagai, Y. Kanesawa and Y. Doi, Makromol. Chem., 193, 53 (1992)

601

Subject index

Accelerator 27

Anhydride conversion 29

Accommodation coefficient 177, 178

Arrhenius factor 30

Acetic anhydride 76

Atactic poly (epichloridrin) 532, 534,

Acrylate monomers 535

547

Acrylonitrile-Butadiene-Styrene

Aureobacterium Saperdae 547, 593

copolymer (ABS) 469, 473

Average particle diameter 304

- composition 509, 510, 512, 513,

Avrami equation 165, 166, 577

514, 515, 516, 517, 520 -

-

emulsion-made 517

Benzoyl peroxide 97, 551

mass-made 517

Biocompatibility 532, 555

- type 520

Biodegradability 530, 532, 547, 555,

Activation energy 30, 300, 302, 306

595, 601

Adhesion 54, 57, 69, 94, 290, 293,

Bismaleimide resins 123, 124, 135

294, 296, 317, 330, 335, 371, 375,

Blend 125, 184, 190, 192, 194, 197,

391,359, 389, 430, 433,476, 478

201, 202, 439, 470, 471, 492, 518,

Alloys 470

538

Amide

- binary 352, 359, 374, 380, 406,

-

- bands 64, 78, 79 -

418, 433

absorption 77

linkage 76

- components 471 -

81, 97, 106, 291, 292, 293, 296,

Amino -

-

323, 329, 495, 496, 497, 498,

groups 75

503, 507, 556, 560, 576

terminated butadiene-acrylonitrile copolymer 75, 76

composition 16, 46, 50, 52, 68,

Blending 243, 244, 254, 291 -

conditions 290, 321,330

602 - mechanical 473

Composites 479, 484

-

reactive 470, 528, 534, 554

Compressive yield stress 88, 107

solution 534, 571

Conversion of alfllydride groups 105

time 246, 263

Copolymer

-

-

Bragg equation 275

- graft 335, 349, 353, 369, 387, 399, 413, 429, 433, 464, 465,

Brittleness 124

466 Calibration

random 350, 432

-

- curve 98

Copolymerization 124, 527

- factor 35, 38

Core

Carbonyl -

-

-

shell particles 57

-

absorption 64

-

stretching 20

Crack, 455

terminated butadiene-acrylonitrile

- arrest line 41

copol3aner (CTBN) 75

- growth 34

Carboxylic groups 293 Cavitation 14, 72, 74, 93, 261,262, 317, 499

structure 439, 450, 466

initiation 507, 508, 513, 519

-

pinning 56

-

propagation 519

-

Chain scission 25, 27

- tip blunting 44

Chemical

Craze 294, 295, 317, 327, 328, 447,

fonnulas of epoxies 11

-

-

interactions 89, 143

Co-continuity 484 Cold drawing 312, 315, 316, 317, 318 Compatibilizing agent 469 Compatible 414, 429

455 initiation 480

-

propagation 519

-

-

tip 480

Crazing 498, 519 -

mechanism of 519

Critical

603

-

-

strain energy release rate (Gr 36,

Defonnation mechanism 72, 73, 488

52, 68, 69, 90, 92, 109, 125, 152

Diblock copolymer 106

stress intensity factor (Kr 35, 51,

Dicumil peroxide 551

68, 69, 80, 91, 108, 125, 151

Dielectric properties 497

Cross-Bueche equation 249, 255,

Differential

264, 269, 297, 298, 306

(DSC) 16, 46, 462, 463, 464, 495,

Cross linking

534

-

calorimetry

Dispersion coarseness 251,292, 302,

agent 440

- density 13,442, 443

317, 328, 329

Crystalline

Distribution

- forms 307

- anysotropic 292, 301,309

- lamella thickness 273, 275, 283

- layered 310, 314, 316

Crystallinity index 273, 276, 559,

-

Crystallization 157, 433

-

- conditions 243, 244, 245, 272,

-

-

of the particles 54, 57

Domain

599

size 442, 443 structure 443

282, 290, 307, 330

Droplet 171, 172, 192, 215

fractionate 214, 215

Dugdale model 106

- process 262, 272, 273, 282, 283 -

scanning

rate 281

Dynamic mechanical -

analysis 32, 50

Curing process 16, 27, 48

- behaviour 483,479, 497

Debye-Bueche relation 586, 587

Elastic modulus 126, 131, 152, 312

Decarboxylation 25

Elastomer 335

DDS

(4,4'

diamino

diphenyl

- at break 317, 318

sulphone) 18 DDM

(4,4'

methane) 12

Elongation

diamino

diphenyl

- at fracture 485 Emulsifying agent 74, 106, 109

604 Engineering thermoplastics 14

Filler 178

Enzymatic degradation 543, 550

Films 185, 186

Epoxy

Flame retardants 498

-

-

equivalent weight 13

Flexural strength 133

resins 11, 12, 13

Flory-Huggins theory 577

Etching 127, 500, 501,502, 503

Flow mechanism 298

Ethylene-propylene copolymers 205,

Fourier

206

Spectroscopy (FTIR) 16, 18, 64, 76,

Ethylene-Propylene

Transfonn

Infrared

136, 560

monomers

(EPM) 336, 340, 346, 365, 380,

Fox equation 568

399, 414

Fractographic

-

g - SA 348, 357, 369, 378, 397,

328, 454

417, 425, 337, 344

Fractography 499, 500, 507

- g - SA- PA6 349, 369, 380, 391, 429 Extrudate swell 482, 494 Extruded sheets 496 Extruder 194, 195

analysis

260,

294,

Fracture 479, 498, 499, 506, 507 -

analysis 32, 563 behaviour 34, 51, 118, 125

-

-

energy 133 fast 325

-

-

double screw 305

-

front 295

-

single screw 321

-

induction 260, 261,294, 295, 326

Extrusion -

capillary 321

-

chamber 321

-

direction 324

mechanics tests 100

-

-

mechanism 262, 296, 327, 352, 390

-

surface 72, 260, 261, 294, 325, 326, 430

Failure -

mode of 480

Fibre rupture 316

-

tougheness 12, 13, 36, 68, 69, 131

605 Glass transition temperature 12, 14,

- behaviour 245, 254, 263, 271,

18, 32, 47, 50, 246, 253, 350, 351,

328, 330, 366, 395, 398, 400,

478, 482, 490, 491, 495, 512, 513,

414, 416, 430, 496, 504, 507

529, 535, 554, 567, 595

- modifiers 262

Grafting

- perfomlance 496, 513, 520

- degree 97, 109, 341, 366, 409,

- properties 253, 262, 267, 269,

417, 424, 434, 435, 517 - of maleic anhydride 338, 420,

-

271,294, 321,539, 549 - strength 253, 259, 261,262, 263,

428, 429, 432

268, 283, 294, 295, 296, 325,

of unsaturated molecules 337,

326, 329, 470, 489, 498, 506,

345, 400

514, 520 -

test 38, 92, 125

Hardener 12, 17

- tougheness 38, 43, 51, 52

Heat

Incompatible blends 426, 438

-

-

distortion temperature 470

Injection moulded 246, 247, 258,

of cure 48, 49

266, 270, 272, 276, 277, 291, 292,

High impact polystyrene (HIPS) 443,

301, 303, 305, 307, 311, 313, 314,

452

315,316

Hoffinan theory 571

Interaction

Hydrogen bond 21, 63, 65, 66, 77,

- energy 476, 487

78 -

Hydroxyl terminated -

parameter 476, 477, 482, 496

Inter

polybutadiene rubber 61 -

- PC24 -

fibrillar regions 276 lamellar amorphous layer 276, 277,278

Image analysis 113 -

Impact

particle distance 254, 262

- penetrating

polymer

(IPN) 440, 449

network

606 -

spherulitic

amorphous

Long spacing

contact

2

7

5

Low profile additives 61

regions 279 Interface 69, 74 Interfacial

Maleic anhydride 76

-

adhesion 74

Maleimide groups 85

agents 469

Maleimido

-

-

-

fracture tougheness 490

-

tension 487, 488, 489

-

-

-

-

Maximum stress at fracture 485

terminated polybutadiene rubber

Mechanical -

-

Kinetic

-

-

terminated rubber 75, 136

groups 62, 68

62

-

temainated butadiene-acrylonitrile copolymer (ITBN) 75

Irwin model 95, 110 Isocyanate

double bonds 136

analysis 30, 32, 136

behaviour 469, 480, 504, 519

-

degradation 336

-

properties 349, 366, 393, 419, 530, 549

constant 31 equation 30

analysis 32

-

resistance 470, 505

Kwe and Frisch equation 578

Melt

Kyotami equation 307

- mixing 243, 291, 321, 335, 348, 367, 375, 398, 418, 432

Linear Elastic Fracture Mechanics

-

rheology 245, 269, 290, 321

(LEFM) 125, 458

-

viscosity 472, 492, 493, 499

Liquid

Melting

-

crystalline 218, 219

-

rubber 13, 52, 62, 75

Logaritlun rule of mixtures 247

-

point 157, 174

- temperature 529, 535, 570, 575, 595

607 Methyl nadic anhydride 12, 27

- mass 244, 245, 249, 250, 253,

Michael reaction 123, 124

254, 255, 258, 262, 263, 264,

Micro

266, 267, 268, 269, 270, 277,

- cracking 56

278, 282, 283, 285, 289, 290,

-

292, 296, 298, 300, 301, 305,

defomlation 481

307, 318, 321,329, 330, 332

- voids 371 Migration 211, 212, 213, 214, 483,

-

mass distribution 244, 245, 254,

491,513, 518

255, 263, 264, 267, 268, 269,

Miscibility

270, 282, 283

- in PC/SAN blends 478, 481,487, 518 -

window 477

- structure 245, 261,262, 272 -

weight 62, 75

Morphological analysis 40, 110, 127,

Mixing

452, 537, 546

- teclmiques 349, 350, 353, 429

Morphology 70, 349, 352, 361, 371,

- torque 503, 510, 511,512, 516

386, 393, 407, 416, 426, 432, 483,

Mode and state of dispersion 244,

484, 496, 497, 498, 502

245, 247, 258, 262, 263, 266, 267,

-

droplet-like 259

270, 271, 282, 292, 301, 315, 323,

- melt 243,263, 282

329, 330

Multicraze formation 262

Modulus 14, 36, 37, 53, 85, 106 -

-

loss 247

Necking 316

storage 247

Negative deviation blends 247, 249,

Mold 291,292, 305

270

Molecular

Notch sensitivity 470

- characteristics 290, 297, 305

Nuclear Magnetic Resonance (NMR)

-

-

composite 218 interactions 21

492, 592 Nucleating agents 176, 177, 197, 212,213

608 Nucleation -

dual continuity 442, 443

-

density 185, 186, 189, 192, 197,

inversion 55, 442, 443, 445, 449,

-

200, 201,204, 244, 283 -

heterogeneous

450, 452, 453, 454, 503

161, 163, 164,

173, 174, 189, 193, 197, 203,

- morphology 323 -

205, 213, 220 - homogeneous 161, 163, 164, 189

e

p

a

r

a

t

i

o

n

50, 376, 379, 408,

421,440, 443 -structure 244, 262, 269, 277,

- primary 158, 163, 164, 168, 170,

282, 287, 289, 290, 318, 321,

173, 184, 185, 186, 188, 192, 195, 202, 204, 205, 207, 217

s

325, 329, 330 -

viscosity ratio 252, 258, 263,

- secondary 179, 180, 182

267, 270, 271, 282, 283, 292,

- self 175, 177, 197, 203

296, 302, 303, 304, 312, 329, 330

Optical -

-

-

bistability 446 density 589 microscopy 480, 481

Photostability 496 Plane -

strain 499, 506, 509, 520

- stress 499, 506, 509, 520

- properties 446

Plastic

- transparency 439, 450

- deformation 56, 295, 325, 327

Orientation 214, 261 Oscillatory shearing flow 246, 254

-

-

shear deformation 72, 94, 111 zone 42, 95, 96

Plasticizer 48, 50 Particle size 72, 251,254, 262, 263,

Poisson ratio 91

268, 270, 271, 282, 283, 292, 301,

Poly

311,328, 329

- acrylonitrile (PAN) 470, 472

-

distribution 267

Phase

- amide 336, 348, 351, 365, 376, 400, 417, 432

609 -

-

b

i

s

p

h

e

n

o

l

-

A

carbonate 15, 16,

-

469, 471

-

butadiene 99, 444, 470

-

- butylacrylate 530, 536 -

-

-

vinylacetate 62, 464, 465, 466

butylacrylate-styrene 439, 448

Polymerization 335, 375, 400, 409,

caprolactam 419, 421

413, 418, 432, 474, 474, 474, 527,

condensation 375, 378, 418

530, 530, 536 -

shrilukage 61

Processability 470, 472, 504, 516,

- ethersulfone 15

517,518

- ethylene, high density 336

-

sulfone 15

Polymer-diluent theory 570

- ether imide 15, 46, 124

-

urethane 441,447, 448

Processing 244, 287, 290, 321, 330,

ethylene-vinylacetate 439

483, 486, 492, 497, 499, 483, 486,

esterresins 61, 75

492 -

13-hydroxybutyrate-co- conditions 321,327, 329, 330

hydroxyvalerate (PHBV) 529 -

-

-

- cycle 291,305

imides 121,122

Pseudoplastic

isobutylene 96

- flow 250

isoprene 99

- methyl methacrylate 61,439 - propylene 336, 341 -

-

-

- melts 247, 264, 297

Radial growth rate 272

styrene 470

Radical polymerization 123

styrene-butadiene 444

Reaction

styrene butadiene-styrene 445 -

-

styrene-co-acrylonitrile

(SAN) -

470, 471, 472, 473, 474, 476, 481, 484, 486, 488, 489, 490, 491,473, 474, 472, 477, 520 -

styrene-methacrylic acid 442

mechanism 101 time 69

Reactive elastomers 124, 131 Refractive index 439, 449, 466 Relaxation time 298, 302 Reptation 162, 180

610 Residence time 321, 322, 323, 326,

Spherulite

327, 328

-

average dimensions 278

Rheological behaviour 485, 510, 511

-

Rubber

- growth rate 166, 569, 585, 595

-

-

particles 71,487, 488, 517

-

in blends 389, 397, 417

phase volume (RPV) 444, 446,

-

radius 188, 199, 219, 222

447

size 165, 190, 193, 201

-

SAXS profiles 273, 275, 276, 584 Scamling electron microscopy (SEM) 41, 53, 70, 83,452, 457, 500 -

analysis 250, 258, 266, 276, 292, 309, 312, 323

Selective dissolution 324 Shear bands 480

-

lips 499

-

-

-

dimensional 470, 472, 473

-

thennal 471

Stick-slip propagation 34 Straining 315 Stress -

at break 317

-

concentration 317

-

cracking 470, 473 strain curves 87, 107, 505

-

-

thi~ming 249, 250, 257, 298

-

yielding 14, 72, 74, 111, 262, 294, 295, 327, 519

Sherrer equation 582

size distribution 244

Stability

rate 247, 249, 296, 298, 321,493 stress 301,323

-

fibrillae 275

-

whitening 260, 261, 262, 295, 317, 325, 328

Structure -

interphase 272

-

layered 258, 266, 267, 309, 312,

Shish kabab 176

313,315

Size of the particles 54, 57 -

rod-like 325

-

super-reticular 272

Slow-growth region 41 Specimen 472, 481, 481 Spectral subtraction 20, 138, 144

Succinic anhydride groups 97

611 Surface - energy 209 -

free energy of folding 274, 278

System -

-

Thermal properties of PC/ABS blends 504, 512, 518

-

of PC/SAN blends 480, 482, 518

-

Thermoplastics 124

incompatible 469

Thermoplastic modifier 15

multiphase 470

Tie molecules 244, 280, 281 Toluene diisocyanate 62

Tan 5 32

Toughening 245,254, 259, 262, 263,

Taylor-Tomotika theory 252, 254,

280, 282, 283, 290, 296, 318, 329,

259, 263, 271,303, 329

330, 375, 401, 435, 439, 440, 443,

Temperature

447, 448, 449, 465,469, 470, 500

- apparent melting 272 - equilibrium melting 272

-

-

agent 13, 136 mechanism 56, 497

- extrusion 302, 322, 323, 324

- polyamide 365,377, 417

- melting 290, 291,297, 305, 332

Transparency 470

- mold 246, 263

Transmission electron microscopy (TEM) 80, 444

- processing 324, 325 Tensile -

Triblock copolymer 66, 69, 72, 74

elastic behaviour 312, 584 Ultimate tensile strength 462

- properties 317, 351, 352, 363, 370, 380, 416, 433 -

test 479

Ternary blends 486, 488 TGAP (triglycidyl epoxide based on

Undercooling 243, 272, 278, 281, 282, 283,325,329 Uniaxial compression 86 Urethane - carbonyls 65

aminophenol) 11 -

TGDDM (tetraglycidil epoxide based

linkages 63

on diamino diphenyl methane) 11, 46 Vicat temperature 496

612 Vinylacetate

content

of

EVA

copolymer 289, 290, 293, 305, 306, 307, 308, 311, 312, 315, 316, 317, 318, 320, 327, 330 Viscosity 247, 249, 251, 252, 254, 255,257, 258, 482, 494, 499 - apparent 249, 297, 322, 323 - complex 246, 247, 264, 271 - dynamic 246 -

-

Mooney 246 ratio 483 zero-shear 249, 269, 298

Viscoelastic properties 484 Volume of flow element 301,306

WAXS patterns 307, 581 Welding 500

Yield 352, 371 -

-

point 314 stress 14,490

Yielding 260 -

behaviour 86

- mechanism 43, 56 - process 326 Young elastic modulus 426

Advanced Routes for Polymer Toughening

Read more

Advanced Polymer Chemistry

Read more

Advanced Polymer Processing Operations

Read more

Advanced Polymer Processing Operations

Read more

Escape Routes

Read more

Nanoassembly for Polymer Electronics

Read more

Advanced Polymer Nanoparticles: Synthesis and Surface Modifications

Read more

Advanced ESR Methods in Polymer Research

Read more

Advanced Polymer Nanoparticles: Synthesis and Surface Modifications

Read more

Polymer Synthesis Polymer-Polymer Complexation

Read more

Routes of Reference

Read more

Routes of Reference

Read more

Routes: Eloge de l’autonomadie

Read more

Polymer Analysis Polymer Theory

Read more

Polymer Analysis Polymer Theory

Read more

Polymer Analysis Polymer Physics

Read more

Polymer Synthesis Polymer Catalysis

Read more

Serbian Wine Routes

Read more

Polymer Membranes for Fuel Cells

Read more

Rheology for Polymer Melt Processing

Read more

Routes to Cellulosic Ethanol

Read more

Mechanics of Transformation Toughening and Related Topics

Read more

Advanced Stories for Reproduction

Read more

Advanced thermodynamics for engineers

Read more

Advanced Mathematics for Applications

Read more

Advanced Thermodynamics for Engineers

Read more

Advanced Thermodynamics for Engineers

Read more

Advanced Selling for Dummies

Read more

Advanced Lingo for Games

Read more

Advanced thermodynamics for engineering

Read more

Recommend Documents

Advanced Routes for Polymer Toughening

Polymer Science Library 10 ADVANCED ROUTES FOR POLYMER TOUGHENING Polymer Science Library Edited by A.D. Jenkins Uni...

Advanced Polymer Chemistry

Advanced Polymer Chemistry Advanced Polymer Chemistry Problem Solving Guide Manas Chanda Indian Institute of Science...

Advanced Polymer Processing Operations

ADVANCED POLYMER PROCESSING OPERATIONS Edited Nicholas by P. Cheremisinoff, Ph.D. NOYES PUBLICATIONS Wesiwood, N...

Advanced Polymer Processing Operations

ADVANCED POLYMER PROCESSING OPERATIONS Edited Nicholas by P. Cheremisinoff, Ph.D. NOYES PUBLICATIONS Wesiwood, N...

Escape Routes

johann christoph arnold escaperoutes for people who feel trapped in life’ life’ss hells 1 This e-book edition of Esca...

Nanoassembly for Polymer Electronics

Encyclopedia of Nanoscience and Nanotechnology www.aspbs.com/enn Nanoassembly for Polymer Electronics Tianhong Cui Uni...

Advanced Polymer Nanoparticles: Synthesis and Surface Modifications

ADVANCED POLYMER NANOPARTICLES Synthesis and Surface Modifications ADVANCED POLYMER NANOPARTICLES Synthesis and Surfa...

Advanced ESR Methods in Polymer Research

0471450677.ffirs.qxd 8/16/2006 11:40 AM Page i ADVANCED ESR METHODS IN POLYMER RESEARCH 0471450677.ffirs.qxd 8/1...

Advanced Polymer Nanoparticles: Synthesis and Surface Modifications

ADVANCED POLYMER NANOPARTICLES Synthesis and Surface Modifications ADVANCED POLYMER NANOPARTICLES Synthesis and Surfa...

Polymer Synthesis Polymer-Polymer Complexation

Synthesis, Characterization and Properties of Octa-Arm Polyisobutylene-Based Star Polymers Sunny Jacob, Joseph P. Kenned...