Catalyst Design for Tailor-Made Polyolefins: Proceedings of the International Symposium on Catalyst Design for Tailor-Made Polyolefins, Kanazawa, Ja (Studies in Surface Science and Catalysis)

Studies in Surface Science and Catalysis 89 CATALYST DESIGN FOR TAILOR-MADE POLYOLEFINS Proceedings of the International Symposium on Catalyst Design for Tailor-made Polyolefins, Kanazawa, March 10-12, 1994

This Page Intentionally Left Blank

Studies in Surface Science and Catalysis Advisory Editors : B . Delmon and J. T. Yates Vol .89

CATALYST DESIGN FOR TAILOR- MADE POLY0LEFlNS PROCEEDINGS OF THE INTERNATIONAL SYMPOSIUM ON CATALYST D ESIG N FOR TAI LOR-M A DE POLY0LEFINS , KANAZAWA, MARCH 10-12, 1994 Edited by Kazuo Soga

Japan Advanced Institute of Science and Technology, Hokuriku

Minoru Terano

Japan Advanced Institute of Science and Technology, Hokuriku

KODANSHA Tokyo

1994

ELSEVIER Amsterdam - London - New York - Tokyo

Copublished by KODANSHA LTD., Tokyo and ELSEVIER SCIENCE B.V., Amsterdam exclusive sales rights Japan KODANSHA LTD. 12-21, Otowa 2-chome, Bunkyo-ku, Tokyo 112, Japan

for the rest of the world ELSEVIER SCIENCE B.V. 25 Sara Burgerhartstraat, P.O. Box 211, 1000 AE Amsterdam, The Netherlands

ISBN 0-444-98656- 1 ISBN 4-06-2071 86-X (Japan)

Copyright @ 1994 by Kodansha Ltd.

All rights reserved No part of this book may be reproduced in any form, by photostat, microfilm, retrieval system, or any other means, without the written permission of Kodansha Ltd.(except in the case of brief quotation for criticism or review) PRINTED IN JAPAN

List of Contributors Numbers in parentheses refer to the pages on which contributors’ paper begin.

Abe, M. (327) Osaka Research Laboratory, Mitsui Toatsu Chemicals, Inc., Takaishi, Osaka 592, Japan Akino, Y. (1 19) Research Laboratory of Resources Utilization, Tokyo Institute of Technology, Midori-ku, Yokohama 227, Japan Akiyama, M. (327) Central Research Institute, Mitsui Toatsu Chemicals, Inc., Sakae-ku, Yokohama 247, Japan Albizzati, E. (139) Himont “G. Natta” Research Center, P. le G. Donegani, 12 44100 Ferrara, Italy Altomare, A. (257) Department of Chemistry and Industrial Chemistry, University of Pisa, via Risorgimento 35, 56126 Pisa, Italy Arndt, M. (179) University of Hamburg, Edmund-Siemers-Allee 1, Germany Arribas, G. (257) Escuela de Quimica, Facultad de Ciencias, Universidad Central de Venezuela, Caracas, Venezuela Asanuma, T. (327) Osaka Research Laboratory, Mitsui Toatsu Chemicals, Inc., Takaishi, Osaka 592, Japan Bacon, D. W. (81) Chemical Engineering Department, Queen’s University, Kingston, Ontario, K7L 3N6, Canada Beck, S. (193) Fakultat fiir Chemie, Universitat Konstanz, D-78434 Konstanz, Germany

vi

List of Contributors

Berry, I. G . (55) Department of Chemistry, UMIST, Manchester M60 IQD, U.K. Bohm, L. L. (351) Hoechst AG, 65926 Frankfurt(M), Germany Brintzinger, H. (193) Fakultat f i r Chemie, Universitat Konstanz, D-78434 Konstanz, Germany Bujadoux, K. (249) E.C.P. EniChem Polymeres France, Centre de recherche, 62670 Mazingarbe, France Burfield, D. R. (91) Chemistry Department, University of Malaya, 59100 Kuala Lumpur, Malaysia Busico, V. (21) Dipartimento di Chimica-Universita’ di Napoli via Mezzocannone, 4-1-80 134 Napoli, Italy Chu, K. J. (299) Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, 373- 1 Kusong-dong, Y usong-gu, Taejon 305-701, Korea Ciardelli, F. (257) Department of Chemistry and Industrial Chemistry, University of Pisa, via Risorgimento 35, 56126 Pisa, Italy Cipullo, R. (21) Dipartimento di Chimica-Universita’ di Napoli via Mezzocannone, 4-1-80 134 Napoli, Italy Conti, G. (257) Department of Chemistry and Industrial Chemistry, University of Pisa, via Risorgimento 35, 56126 Pisa, Italy Corradini, P. (21) Dipartimento di Chimica-Universita’ di Napoli via Mezzocannone, 4-1-80134 Napili, Italy Dall’ Occo, T. (139) Himont “G. Natta” Research Center, P. le G. Donegani, 12 44100 Ferrara, Italy Dupuy, J. (109) CNRS-Laboratoire de Chimie et Procedes de Polymerisation LCPP BP 69390 Vernaison, France Dyachkovskii, F. S. (201) Institute of Chemical Physics Russian of Scienses, Chernogolovka, 142432, Moscow Region, Russia

List of Contributors vii

Eisch, J. J. (221) Department of Chemistry, State University of New York at Binghamton, Binghamton, NY 13902-6000, U.S.A. Enderle, H. F. (351) Hoechst AG, 65926 Frankfurt(M), Germany Ewen, J. A. (405) Catalyst Research Corporation, 1823 Barleton Way, Houston, TX 77058, U.S.A. Fleissner, M. (351) Hoechst AG, 65926 Frankfurt(M), Germany Galimberti, M. (139) Himont “G. Natta” Research Center, P. le G. Donegani, 12 44100 Ferrara, Italy Gan, S. N. (91) Chemistry Department, University of Malaya, 59 100 Kuala Lumpur, Malaysia Guyot, A. (43) CNRS-LCPP, BP 24-69390 Vernaison, France Han, T. K. (163) Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, Y usong-gu, Daejon 305-701, Korea Hosoda, S. (365) Sumitomo Chemical Co., Chiba Research Laboratory, Sodegaura, Chiba 299-02, Japan Hsu, J. C. (81) Chemical Engineering Departoment, Queen’s University, Kingston, Ontario, K7L 3N6, Canada Hungenberg, K. D. (373) BASF AG, Plastics Laboratory, D-67056 Ludwigshafen, Germany Ihara, E. (237) Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higasi-Hiroshima, Hirosima 724, Japan Ihm, S. K. (299) Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, 373- I Kusong-dong, Yusong-gu, Taejon 305-701, Korea Imai, M. (1 71) Tonen Chemical Corporation, Tonen Corporate R&D Laboratory, Iruma-gun, Saitama 356, Japan

viii

List of Contributors

Inoue, N. (327) Central Research Institute, Mitsui Toatsu Chemicals, Inc., Sakae-ku, Yokohama 247, Japan Ishihara, N. (339) Central Research Laboratories, IDEMITSU KOSAN Co., Sodegaura, Chiba 299-02, Japan Jeong, Y. T. (153) Department of R&D, Korea Petrochemical Industrial Co., Ulsan 680-1 10, Korea Jordan, R. F. (271) Department of Chemistry, University of Iowa, Iowa 52242, U.S.A. Journaud, C. (43) CNRS-LCPP, BP 24-69390 Vernaison, France Kakugo, M. (129) Chiba Research Laboratory, Sumitomo Chemical Co., Sodegaura, Chiba 299-02, Japan Kaminsky, W. (1 79) University of Hamburg, Edmund-Siemers-Allee 1, Germany Kanazawa, S. (471) Tonen Chemical Corporation, Tonen Corporate R&D Laboratory, Iruma-gun, Saitama 356, Japan Kang, K. K. (153) Department of R&D, Korea Petrochemical Industrial Co., Ulsan 680- 1 10, Korea Kao, S. C. (389) Union Carbide Corporation, P.O. Box 670, Bound Brook, NJ 08805, U.S.A. Karol, F. J. (389) Union Carbide Corporation, P.O. Box 670, Bound Brook, NJ 08805, U.S.A. Kashiwa, N. (381) Polymers Laboratories, Mitsui Petrochemical Industries Ltd., Waki, Y amaguchi 740, Japan Keii, T. (1) Japan Advanced Institute of Science and Technology, Hokuriku, Nomi-gun, Ishikawa 923-12, Japan Kerth, J. (373) BASF AG, Plastics Laboratory, D-67056 Ludwigshafen, Germany Kimura, S. (327) Osaka Research Laboratory, Mitsui Toatsu Chemicals, Inc., Takaishi, Osaka 592, Japan

List of Contributors ix

KO, Y. S. (163) Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, Y usong-gu, Daejon, 305-701, Korea Kohno, M. (327) Central Research Institute, Mitsui Toatsu Chemicals, Inc., Sakae-ku, Yokohama 247, Japan. Kojima, K. (365) Sumitomo Chemical Co., Chiba Research Laboratory, Sodegaura, Chiba 299-02, Japan Kuramoto, M. (339) Polymer Research Laboratory, IDEMITSU Petrochemical Co., Ichihara, Chiba 299-01, Japan Lancaster, G. M. (285) Polyolefins and Elastomers Research and Development Laboratories, The Dow Chemical Company, 2301 Brazosport Boulevard, B-1607, Freeport, T X 77541, U.S.A. Langhauser, F. (373) BASF AG, Plastics Laboratory, D-67056 Ludwigshafen, Germany Langlotz, J. (315) Institute of Inorganic Chemistry, Martin-Luther-University Halle-Wittenberg, Geusaer StraBe, D-062 17 Merseburg, Germany Leclerc, M. (193) Fakultat fur Chemie, Universitat Konstanz, D-78434 Konstanz, Germany Lee, D. H. ( 1 53) Department of Polymer Science, Kyungpook National University, Taegu 702-701, Korea Loi, P. S. T. (91) Chemistry Department, University of Malaya, 59 100 Kuala Lumpur, Malaysia Masi, P. (73)(257) EniChem, via Maritano 26, 20097 S.Donato Milanese, Italy Masson, P. (109) CNRS-Laboratoire de Chimie et P r o d d b de Polym&isation LCPP BP 69390 Vernaison, France Menconi, F. (73)(257) EniChem, via Maritano 26, 20097 S.Donato Milanese, Italy Morimoto, M. (237) Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higashi-Hiroshima, Hiroshima 724, Japan

x

List of Contributors

Morini, G. (139) Himont “G. Natta” Research Center, P. le G. Donegani, 12 44100 Ferrara, Italy Mortreux, A. (249) Laboratoire de Catalyse hCiCrog6ne et homogene, URA C N R S 402, USTL, ENSCL, BP 108, 59652 Villeneuve d’Ascq Cedex, France Muller, P. (373) BASF AG, Plastics Laboratory, D-67056 Ludwigshafen, Germany Murata, M. (171) Tonen Chemical Corporation, Tonen Corporate R&D Laboratory, Iruma-gun, Saitama 356, Japan Nakano, A. (171) Tonen Chemical Corporation, Tonen Corporate R&D Laboratory, Iruma-gun, Saitama 356, Japan NG, S. C. (91) Chemistry Department, University of Malaya, 59100 Kuala Lumpur, Malaysia Nodono, M. (237) Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higashi-Hiroshima, Hiroshima 724, Japan Olonde, X. (249) E.C.P. EniChem PolymZres France, Centre de recherche, 62670 Mazingarbe, France Patin, M. (109) CNRS-Laboratoire de Chimie et ProcCd6 d e Polym&-isation L C P P BP 69390 Vernaison, France Pellecchia, C. (209) Dipartimento di Fisica, Universitii di Salerno, 1-8408 1 Baronissi(SA), Italy Pelletier, J. F. (249) Laboratoire de Catalyse hEt6roghe et h o m o g h e , URA C N R S 402, USTL, ENSCL, BP 108, 59652 Villeneuve d’Ascq Cedex, France Petit, F. (249) Laboratoire de Catalyse hEtCrog&ne et h o m o g h e , URA C N R S 402, USTL, ENSCL, BP 108, 59652 Villeneuve d’Ascq Cedex, France Pombrik, S. I. (221) Department of Chemistry, State University of New Y ork at Binghamton, Binghamton, N Y 13902-6000, U.S.A. Robert, P. (109) CNRS-Laboratoire de Chimie et Procgdb d e Polymerisation L C P P BP 69390 Vernaison, France

List of Contributors xi

Rbll, W. (193) Fakultat f i r Chemie, Universitat Konstanz, D-78434 Konstanz, Germany Shariati, A. (81) Chemical Engineering Department, Queen’s University, Kingston, Ontario, K7L 3N6, Canada Shigematsu, Y. (365) Sumitomo Chemical Co., Chiba Research Laboratoy, Sodegaura, Chiba 299-02, Japan Shiomura, T. (327) Central Research Institute, Mitsui Toatsu Chemicals, Inc., Sakae-ku, Yokohama 247, Japan Shiono, T. ( 1 19) Research Laboratory of Resources Utilization, Tokyo Institute of Technology, Midori-ku, Yokohama 227, Japan Soga, K . ( 1 19)(307) Japan Advanced Institute of Science and Technology, Hokuriku, Nomi-gun, lshikawa 923-12, Japan Solli, K. A. (35) Borealis AS, N-3960 Stathelle, Norway Spitz, R. (43)(109) CNRS-Laboratoire de Chimie et Procedb de PolymCrisation LCPP BP 69390 Vernaison, France Stehling, U. (193) Fakultat fur Chemie, Universitat Konstanz, D-78434 Konstanz, Germany Stevens, J. C. (277) Polyolefins and Elastomers Research and Development Laboratories, The Dow Chemical Company, 2301 Brazosport Boulevard, B-1607, Freeport, TX 77541, U.S.A. Sugimoto, R. (327) Osaka Research Laboratory, Mitsui Toatsu Chemicals, Inc., Takaishi, Osaka 592, Japan Sun, L. (81) Chemical Engineering Department, Queen’s University, Kingston, Ontario K7L 3N6, Canada Swogger, K. W. (285) Polyolefins and Elastomers Research and Development Laboratories, The Dow Chemical Company, 230 1 Brazosport Boulevard, B- 1607, Freeport, TX 77541, U.S.A.

xii

List of Contributors

Tait, P. J. T. ( 5 5 ) Department of Chemistry, UMIST, Manchester M60 IQD, U.K. Taube, R. (315) Institute of Inorganic Chemistry, Martin-Luther-University Halle-Wittenberg, Geusaer StraBe, D-062 17 Merseburg, Germany Terano, M. (101) School of Materials Science, Japan Advanced Institute of Science and Technology, Hokuriku, Nomi-gun, Ishikawa 923- 12, Japan Tjaden, E. B. (271) Department of Chemistry, University of Iowa, Iowa 52242, U.S.A. Uemura, A. (365) Sumitomo Chemical Co., Chiba Research Laboratory, Sodegaura, Chiba 299-02, Japan Vindstad, B. K. (35) Statoil R&D, N-7004 Trondheim, Norway Wache, S. (315) Institute of Inorganic Chemistry, Martin-Luther-University Halle-Wittenberg, Geusaer StraBe, D-062 17 Merseburg, Germany Wester, T. S. (35) Norwegian Institute of Technology, Department of Inorganic Chemistry, N-7034 Trondheim, Norway Woo, S . I . (163) Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, Y usong-gu, Daejon 305-701, Korea Yamamoto, I . (365) Sumitomo Chemical Co., Chiba Research Laboratory, Sodegaura, Chiba 299-02, Japan Yamashita, M. (237) Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higashi-Hiroshima, Hiroshima 724, Japan Yasuda, H. (237) Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higashi-Hiroshima, Hiroshima 724, Japan Yim, J. H. (299) Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, 373- 1 Kusong-dong, Y usong-gu, Taejon 305-701, Korea

List of Contributors xiii

Yokote, Y. (327) Central Research Institute, Mitsui Toatsu Chemicals, Inc., Sakae-ku, Yokohama 247, Japan Yoshioka, S. (237) Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higashi-Hiroshima, Hiroshima 724, Japan Ystenes, M . (35) Norwegian Institute of Technology, Department of Inorganic Chemistry, N-7034 Trondheim, Norway Zambelli, A . (209) Dipartimento di Fisica, Universitg di Salerno, 1-8408 1 Baronissi(SA), Italy

This Page Intentionally Left Blank

Contents

List of Contributors ................................................................................................ Preface ...............................................................................................................

1.

2.

3.

4.

5.

6.

7.

V

xix

Articulation of Kinetics of Quasi-Living Stages to those of Slurry Polymerization and an Unified Explanation-Propene Polymerization with MgCl,/EB/TiCl,-Al(C,H,),-(T. Keii)

1

Active Sites and Mechanisms of Stereospecificity in Heterogeneous Ziegler-Natta Catalysts (P. Corradini, V. Busico and R. Cipullo)

21

Dependence of Transient Comonomer Kinetics on Catalyst Design by Magnesium Chloride Supported Polymerization of Ethene and Propene (K.A. Solli, B.K. Vindstad, T.S. Wester and M. Ystenes)

35

A New Mechanism for Hydrogen Activation in Propene Polymerization Catalysts (A. Guyot, R. Spitz and C. Journaud)

43

Rate Enhancement Effects in the Prepolymerization and Copolymerization of Ethylene and a-Olefins (P.J.T. Tait and I. G. Berry)

55

Characterization of Active Sites in Ti/Hf/MgC12 Catalysts by Chiral Reagents (F. Masi and F. Menconi)

73

A New Polymer-Supported Catalysts for Olefin Polymerization (L. Sun, A. Shariati, J.C. Hsu and D.W. Bacon)

81

xvi

Contents

Active Center Determination in Ziegler-Natta Polymerization : an Innovative Dual-Labeling Approach (S.N. Gan, P.S.T: Loi, S.C. N G and D.R. Burfield)

91

Recent Tendency of Research Targets for Industrial Polypropylene Catalysts (M. Terano)

101

10. The Control of Molecular Weight Distributions in Ziegler-Natta Catalysis (R. Spitz, M. Patin, P. Robert, P. Masson and J. DuPuY)

I09

8.

9.

1 1. Synthesis and Application of Terminally Magnesium Bromide-

Functionalized Isotactic Poly (Propene) (T. Shiono, Y. Akino and K. Soga)

119

12. Wide Range Control of Microtacticity in Propylene Polymerization with Heterogeneous Catalyst Systems (M. Kakugo)

129

13. New Heterogeneous Catalysts for Polyolefins (E. Albizzati, T. Dall’Occo, M. Galimberti and G. Morini)

139

14. Change of Internal Donor for Mg(OEt),-Supported TiCI, Catalyst (D.H. Lee, Y.T. Jeong and K.K. Kang)

153

15. Temperature Programmed Decomposition of MgCI,/TH F/ TiC1, Bimetallic Complex Catalyst and its Effect on the Homoand Copolymerization of Ethylene (Y.S. KO, T.K. Han and S.I. Woo)

I63

16. Characterization of Mg/Ti Type Catalysts Prepared from Different Mg Components (M. Murata, A. Nakano, S. Kanazawa and M. Imai)

171

17. Mechanism of the First Steps of the Isotactic Polymerization with Metallocene Catalysts (W. Kaminsky and M. Arndt)

179

18. Reaction Mechanisms in Metallocene-Catalyzed Olefin Polymerization (H. Brintzinger, S. Beck, M. Leclerc, U. Stehling and W.

Roll)

193

19. Role of Ions in Coordination Polymerization of Olefins (F.S. Dyachkovskii)

20 I

20. Copolymerization of Hydrocarbon Monomers in the Presence of CpTiCI, - M A 0 : Some Information on the Reaction Mechanism from Kinetic Data and Model Compounds (A. Zambelli and C. Pellecchia)

209

Contents xvii

21. The Role of Ion-Pair Equilibria on the Activity and Stereoregularity of Soluble Metallocene Ziegler-Natta Catalysts (J.J. Eisch and S.I. Pombrik)

22 1

22. High Molecular Weight Monodisperse Polymers Synthesized by Rare Earth Metal Complexes (H. Yasuda, E. Ihara, S. Yoshioka, M. Nodono, M. Morimoto and M. Yamashita)

237

23. Lanthanocene Based Catalysts for Olefin Polymerization : Scope and Present Limitations (J.F. Pelletier, A. Mortreux, F. Petit, X. Olonde and K. Bujadoux)

249

24. Effect of Ligand and Inorganic Support on Polymerization Performances of Ti and Zr Catalyst (F. Ciardelli, A. Altomare, G. Arribas, G. Conti, F. Masi and F. Menconi)

257

25. Design of Non-Metallocene Single-Site Olefin Polymerization Catalysts (E.B. Tjaden and R.F. Jordan)

27 1

26. InsiteTM Catalyst Structure/Activity Relationships for Olefin Polymerization (J.C. Stevens)

277

27. Novel Molecular Structure Opens Up New Applications for Insite@ Based Polymers (K.W. Swogger and G.M. Lancaster)

285

28. Molecular Weight Distribution Control with Supported Metallocene Catalysts (S.K. Ihm, K.J. Chu and J.H. Yim)

299

29. Highly Isospecific Heterogeneous Metallocene Catalysts Acivated by Ordinary Alkylaluminums (K. Soga)

307

30. Mol Mass Regulation in the Ally1 Nickel Complex Catalyzed 1, 4-cis Polymerization of Butadiene (R. Taube, S. Wache and J. Langlotz)

315

3 1. Syndiotactic Polypropylene (T. Shiomura, M. Kohno, N. Inoue, Y. Yokote, M. Akiyama, T. Asanuma, R. Sugimoto, S. Kimura and M. Abe)

327

32. Syntheses and Properties of Syndiotactic Polystyrene (N. Ishihara, and M. Kuramoto)

339

33. The Industrial Synthesis of Bimodal Polyethylene Grades with Improved Properties (L.L. Biihm, H.F. Enderle and M. Fleissner)

35 1

xviii

Contents

34. Structure and Properties of Ethylene/ a-Olefin Copolymers Polymerized with Homogeneous and Heterogeneous Catalysts (S. Hosoda, A. Uemura, Y. Shigematsu, I. Yamamoto and K. Kojima)

365

35. Progress in Gas Phase Polymerization of Propylene with Supported TiCI, and Metallocene Catalysts (K.D. Hungenberg, J. Kerth, F. Langhauser and P. Miiller)

373

36. Feature of Metallocene-Catalyzed Polyolefins (N. Kashiwa)

38 1

37. Ligand Effects at Transition Metal Centers for Olefin Polymerization (F.J. Karol and S.C. Kao)

389

38. Propylene Polymerizations with Metallocene/Teal/Trityl Tetrakis (Pentafluorophenyl) Aluminate Mixtures (J.A. Ewen)

405

The International Symposium on Catalyst Design for Tailor-made Polyolefins was held at the Ishikawa High-tech Conference Center in Kanazawa, March 10-12, 1994 in memory of the establishment of the Japan Advanced Institute of Science and Technology (JAIST, Hokuriku) through the efforts of President Dr. Tominaga Keii. The symposium had over 200 attendants including 90 foreign scientists from 13 nations. At this meeting various trends in the following were noted. HETEROGENEOUS CATALYSTS Polymerization kinetics and mechanism Unsolved problems Catalyst preparation METALLOCENE CATALYSTS Polymerization mechanism Modeling and modification Applications NEW TRENDS IN T H E POLYOLEFIN INDUSTRY This volume is a collection of 22 invited and 16 contributed papers, which were subjected to scientific review. Unfortunately, the 36 poster papers presented at the symposium have not been included because of limited space. We believe that these proceedings are an excellent guide to the recent developments in both heterogeneous Ziegler-Natta and homogeneous Kaminsky-Sinn catalysts. Large grants from JAIST and the Ministry of Culture and Education, Japan are deeply appreciated. This symposium could not have been held without such invaluable financial support. The editors thank the authors for the superior quality of their presentations as well as for contributing to this volume. Thanks are also extended to Mr. Ippei Ohta of Kodansha Scientific for his invaluable assistance in the editing of this volume. June 1994

Kazuo Soga Minoru Terano

This Page Intentionally Left Blank

I

1. Articulation of Kinetics of Quasi-Living Stages to those of Slurry Polymerization and an Unified Explanation-Propene Polymerization with MgCl,/EB/TiCl,-Al(C,H,),-

Tominaga Keii Japan Advanced Institute of Science and Technolagy, Hokuriku. Tatsunokuchi, Ishikawa 923-12, Japan

INTRODUCTION

For this twenty years the author has been possessed by some curious kinetic behavior of the propene polymerization with a MgC12/EB/TiC1-A1(C2H5)3 catalyst. As described in the previous papers(ll2), the observed kinetic behavior of the polymerization in

a slurry system were so complex that the

usual kinetic analysis could not be applied. Since we supposed that the kinetic behavior suffered from the rapid rate decay of the polymerization, we attempted to develop some

new method that applicable for observing kinetic behavior of

the polymerization free from any rate decay or those, at least, with a negligibly small rate decay.

A stopped flow reactor developed with Terano et

a1.(3) was useful to carry out the polymerization for a short time, such as 0.03-1s at room temperature, where the polymerizations were those of quasipolymerization with constant rates. stages of the polymerizations, we

From the results of the quasi-living could determine precise values of rate

constants of propagation and transfer reaction as well as concentrations of polymerization centers, as communicated before(3)' ( 4 ) t ( 5 ) . observation gives

In addition, the

us some important key to understand our kinetic data of the

slurry polymerization together with those reported by Giannini(6) and Hsu et a1.(7), on the basis of an unified kinetic model, the description of which is the purpose of

this reporting.

First, the problems remain unsolved in the slurry polymerization are summarized. Then, they are discussed in the light of of new data obtained by the stopped flow method. The whole data

articulated are reviewed ahd explained

2 T. Keii unifiedly on the basis of an unified theory of non-uniform active centers.

OBSERVED

KINETICS

OF

SLURRY

POLYMERIZATION;

UWSOLMD

PROBLEMS

The observed kinetic results of the slurry polymerizations carried out under the conditions: temperature (1-65OC), monomer concentration [MI ( 0 . 2 2 0.48mol/dm3), triethy

aluminum

concentration

[A]

(1-100 mmol/dm3)

and

polymerization time t (5s-3h), are summarized in comparison with the results and Hsu et al. ( 7 ) as follows.

reported by Giannini(6) Rat.

docay

, the rate decay is a third

As described in the report

order with respect to rate itself at the beginning stages of the polymerization (0-lh) and order.

it slows down to a second order and then approaches to a first

In the range of

temperature, 1-6S0C,

a

second

approximately could be applied for the time course during

order

decay

0.5-3h at 4loC, as

Re-examining the experimental data, we found that the constant kd is a function of [A] as that

kd'kd'

fAlz61.4 e-17'6 kJ'RTIAl

The rate equation reported by

%,

1[MI = el/(i

S ' ,

0.005<[A]<0.050 mol/dm3

(3)

Hsu et al.(7) is

+ e3e2t)l/&j

(4)

where the three parameters, Bs, are all independent of polymerization time t. Then, the order of rate decay is 1+83, experimental value of 83 is about 1.0-1.5.

that is 2 . 0 - 2 . 5

because of the

The value of 8 3 , however, is 0-0.4

at low values of [A], [A]/[Ti]< 10, where the corresponding decay order is 1-

', though it can also

1.4. The value of 82 is approximately proportional to [A]

be approximated to [A] at the region of low concentrations, [A]/[Ti] < 50

.

I . Kinetics of Propene Polymerization and Unified Explanation 3

The results reported by Giannini(6) may be represented by

which corresponds to a decay of order, 1.5-2. The rate-decay reported by may be represented by that of order, 1.5-2.

Kashiwa et al. ( * )

Woo et al. (’)

confirmed that a first order decay occurs after a rate maximum in the ethene polymerization with MgC12-supported catalyst. Such rate-decay of first order is popular in the polymerizations with TiC13 catalysts. ( l o )However, it should be noted that the usual method to determine the order n , i.e. linear plotting of (l/Rp,t)n against t , is erroneous but for the case of suitable precise data given. Dependence

polymerization

of

problem unsolved.

Rate profile

rate

on

at a fixed t

triethyl

alunimum

[A] is a

was roughly expressed by

a

pseudo-Langmuir-Hinshelwoodform,

where K is an increasing function of t, because of that the concentration [A],,

at which the rate maximum locates

(

Eq.(6)), was shifted to lower value of [A] as t increased.

[A],

observed

= 1/K

from

In addition, the

value of exponent of the denominator of Eq.(6) increased from 2 to a higher value as t increased. by

that

%

The experimental data of Hsu et al. ( 7 ) may be expressed

is proportional to [AI2

and 81 to log(b[A]/[Ti]) or

and

then their rate equation, Eq. ( 2 ) , can be regarded as more useful than ours , Eq.(1)

.

Abnormal polymerization

dependence centers

of

C*

polymerization

rate

on

concentration

has been found firstly by Giannini(6) who applied

CO*-tagging method to the slurry polymerization during rate decay at represented the abnormality of the result is proportional to (C*,)Y

of

65OC. He

by a figure, which shows that R

as well as by a table, in which values of

[MIC*t are decreased with t increases. The data may be represented by

Pat

%, t/

4

T.Keii

~ =const. , ~ ( c * ~)Y

R

y = 1.7-2.2

,

(7)

We obtained similar results(li2) with the polymerization rate at various time, in gaseous system, Tait('l),

by the use of CO-poisoning method developed by

Caunt and

as

RP = const. (c,*-[coI)~,

y - 2 .

(8)

The value of y obtained for the gaseous polymerization with TiC1J-A1(C2H5)2C1 catalyst, where the rate is free from any decay, was -1.5

.

The result obtained

with a slurry polymerization with a similar catalyst by Caunt and Tait(l') seems to be that of y = 2 - 3 .

the values of y

Although

are

different with

authors and polymerization systems, the conclusion that the dependencv of RP on C* is not first order but hiaher one or that RD/rMlC* is not constant but increases with C* is reliable

.

It should be noted that this fact

demands

the same dependency of the transfer reaction rate, Rtr, on C* because of the

-

established experimental fact that values of Mn polymerizations of rate decay. This predicts that of R

on C* and that of transfer rate R t

basis

.

- on

remain

unchanged during

both the abnormal deDendencv

C* must be exr>lained on the same

These issues are important subjects in this study.

OBSERVED

KINETICS

OF

QUASI-LIVINQ

POLYMERIZATION.

We studied the polymerization in its quasi-living stages in the range of time, 0.1-10s,

by means of

a new stopped flow method

experimental conditions used are temperatures , m01/dm3,

and

[MI, 200-1480 mmol/dm3.

10-60°C,

( 3 ) 8 ( 4 ) , ( 5 ) . The

[A], 0.14-230

The important results interested here

are summarized as follows. The

initial

Rp,o = kp

rate

of

polymerization

%,o

could be represented

by

'*o, [A]

This form could be confirmed from the result that Mn/t was independent of [A] while Y/t was a function of [A], where Y was yield at t. the dependency of C*,

on [A] can be represented by

As

'shown in Fig.l.,

1. Kinetics of Propene Polymerization and Unified Explanation

-

m

0

0.02-

[A]/mmol/dm-3 I

50

Fig.1. Dependence of C*,,

100

150

250

200

/ [Ti] on [A]. 2OoC, [MI=72Omol/dm3 ,

[Ti]= 2 . 3mol/dm3, Curve ( 1)-Eq. (11):Curve(2)-Eq. ( 10)

The constants in the above are Ka=106dm3/rnol, and

a=O .0107, [Alo=10-4mol/dm3,Co=O . 0 8 8 and

[A]' = [A]

-[Ale.

Here,

[A], is corresponding to the

concentration at which polymerization rate Rp is zero. The existence of [Alo seems a feature of the polymerizations with MgC12-supported catalysts, as appeared also in the experimental results of Hsu et a1.(7) who forced to put a constant value such as 5 in [A],/[Ti] even the value shifted approximately to 1 at t=O.

In our slurry polymerization the experimental values of [A], increased

slightly from 0.1 to 0.3, in

as

and then

[A],/[Ti]

, as t increased

from 6min. to 3h('),

5

6 T. Keii

[A], was supposed as amount of alikyl aluminum dissipated for alkylation of catalyst. The

remained

time-invariant

unchanged

in

MWD

the

--

The observed values of polydispersity, Mw/Mn, course

of

the

quasi-living

stages

of

the

polymerization, which has been confirmed by Murata(12) in the polymerization with Si02-supported catalyst. Furthermore, the author pointed out the same results

in the kinetic

polymerization A1(C2H5)2C1

.

study of

and Morinaga

(I4)

Kashiwa(13)

with styrene

and butene-1

with propene polymerization with TiC13-

This is unexpected phenomena because it is well established fact

that polydispersity changes from 1 to 2 in homogeneous Ziegler polymerization. In

the heterogeneous polymerizations, although its values observed values

larger than 2 ,

are

it has been predicted by the existing theories as that i t

changes from 1 to higher value or from a large va1.ue to its double. Therefore, this phenomena is curious one.

A

PHENOMENOLOGICAL

Abnormal

ARTICULATION

Dependency

of

Rp

OF

EXPERIMENTAL

RESULTS.

on C* and Rate-decay.

Taking

dimensions of kinetic parameters into account, the dependency of RP,t

On

the C*t’

Eq.(7), can be represented by

Rp,t = kp[M with

(12)

C*o (f(t))Y

f(t) = C*t / C*,.

(13)

For the case of CO-poisoning we may have similarly

The

expression,Eq.(l4), is

contradicted with Eq.(12) because

its form at

[CO] = 0 , Rp,t = ~ P [ M ] C * ~

at

[CO] = 0

Therefore, Eq.(14) should be revised as

(16)

1. Kinetics of Propene Polymerization and Unified Explanation 7

The form becomes the same form as Eq.(9) at t=O and

[CO]=O. However, to

guarantee the complete agreement between them , we must introduce furthermore a dimensionless function of [A], f([Al), such

+ K,[A]

I

)

,

that

C*,/C0

= f ([A]) = Ka[AI'/(l

as shown by Eq. (11). We can have a phenomenological form from the

articulation of

the experimental results summarized above, because f([CO]) can

be involved in f (t), as Rp,t,[A] = kp[MICof([Al)f(t)y

From the above, it can be shown that of y=2 are compatible with

(18)

Eqs.(l) and (8) in the case

a third order deactivation of

f(t) = C*t/C*o = (1 + 2kdt)-'l2. In the case of y=3 a seventh order deactivation of

C*t

(19)

C*t

must

be

supposed.

These correlations, however, are only apparent because such a higher order deactivation reaction of polymeruzatuin centers is not practical. Rate

Equation. As can be seen from Fig.2, a comparison between a

Langmuir-Hinshelwood equation, Eq.(6), (lines-) and the experimental data at are severe at low concentrations of [A] as well

various times, discrepancies as at t>>2h. 0

0

A

Fig.2.Dependence of [Al/Rp,t,

on [A].(Ref.(1) 4loC), lines are from Eq. ( 6 )

8 T. Keii

However,

Eq.(18) in the case of y=2 , together with

gives the following rate equations,Eqs.(20) and discrepancies in Fig.2

Eqs.(l9),(11) and (3),

(21), which

reveals the

as i lustrated in Fig.3.

Rp,t,[ A

Since the order of rate decay at t>2h

approaches first order, we may take it

as 1.5, for which

is compatible and gives

an alternative form, Eq.(23). The agreement with the

experiment is promoted for t>2h, as illustrated n Fig.4.

Rp,t = ~ p ~ ~ l ~ ~ l ~ 1~ + ~kd'[Alt ~ ~ l / ~ A l o ~ / ~

H,ere, it

may

be

worthwhile

to

recognize

.

Eqs. (20),(21) and (23), pseudo Lan-shelwoo

.

that

these

rate

equations,

d forms. are those of a new

(Rideal) tvrJe of rate eauation for surface react' ion. a i c h has not been nized vet.

I . Kinetics of Propene Polymerization and Unified Explanation 9

Rate-determining

The rate-determining

Ste9

step of

this

polymerization is, of course, a monomer insertion into a growing chain.

The

insertion, however, can be either the insertion of a monomer from solution or that from a coordinated state. The two mechanisms can not be distinguished by usual kinetic experiments.

The reason is revealed here by means of the

statistical mechanical theory of heterogeneous reaction (lo). The rate of the insertion of monomer can be represented alternatively, as below. R

P

= (Insertion rate of a solute m

k.[M])(number

of

into a growing chain,

growing chains with a vacant site. C*@(O)

= k,8(0) [M]C*

or

Rp

)

(24)

(Insertion of a monomer coordinated on the site into a growing chain, k',e(M))(number

of growing chains,

C*

)

=k ',Q(M)c*.

(25)

The ratio of the two probabilities, a site occupied by a monomer, Q(M),

and

the site is vacant, 8 ( 0 ) , can be given by statistical mechanics as

where qM is the partition function of a coordinated monomer, solute monomer and Km is adsorption

(

coordination

)

QM,L that of a

constant, respectively.

From the relation between the two probabilities, we can have the relation between the rate constants, k, and km', k, = k,' K,,, As

can easily be seen from Fig.5. that shows the case

of the insertion step,

the transition state TS is in equilibrium with its two initial states.

Rate of

the determining step can generally be represented by insertion reaction rate = (kT/h) [TS] The equilibrium, " TS$(growing

(28)

chain with vacant site) + (solute monomer)', is

represented, denoting the partition functions of TS

10 T. Keii

Transition S

Initial State

Final State

I

6

T-

C3Ha

T .

Init. S

EP

Coordinat. Final 5.

Fig.5. Reaction Scheme and Energy Relation of Insertion Step.

and growing chain by

p * and q,

* , by

and the equilibrium, " T S S (growing chain) + (coordinated monomer)', by

Then, the rate of polymerization in this case can be expressed by

two

different forms in accordance of the use of one from the above two forms for [TS] in

E q . (28).

(25), are only

Such two resultant forms, which are those of E q s . (24) and

two different styles of expression of the insertion rate,

Eq.(28). If the value of K,[M]

is larger than 1, the site is almost occupied by

monomer, the rate of polymerization is zero order with respect to [MI and the observed activation energy is E +Qm. in Fig.1. In the case that Km[M] << 1, the

P

site is almost vacant,

the rate is first order to [MI and the observed

activation energy is E P'

These two situations segm intuitively to be evidences

1. Kinetics of Propene Polymerization and Unified Explanation

I1

for the insertion of coordinated monomer but they are only evidence of Eq.(28) with the two equilibria, Eqs.(29) and (30). Therefore, the most important is that the the transition state TS involves a site which is available to monomer so guarantees a considerable depression of Ep that the

coordination, which

surface reaction occurs in spite of its small frequency factor. Recently, Ystenes(15) proposed a

"

Trigger Mechanism", in which the

transition state of the insertion was assumed as that involved two monomer; one coordinated and the other from solution. This mechanism should lead that the reaction order with respect to [MI changes from 2 to 1 as %[MI a small value(<>l). order dependence of R by u s ( 1 6 ) . where

R

P

increases from

As an evidence, he takes a higher

on [MI than first order, such as a second order reported

P

In our case, however, the order observed at the initial stages, increases, of the propene polymerization

with TiC13/A1(C2H5)3 catalyst changed from 2 to 1 when the polymerization is started on the addition of monomers after the addition of alkyl aluminum. When the polymerization is started on the addition of alkyl aluminum after the introduction of monomers the order is 1 during the initial stages.

So, we

supposed the participation of a monomer in the formation of a polymerization center.

That is, the new mechanism predicts that the order should be 2 to 1

whereas ours that 2 to 0 as [MI increases. supported

by

many

authors.

For

polymerizations with VC14/A1(C2H5)2Y

(

The latter prediction has been

example,

the

homogeneous

propene

Y = C2H5, C1 and Br) at -78OC and [MI=

8.4-12.9 mol/dm3 (16) can be represented by a usual Langmuir-Hinshelwood type (I7) , Rp = CoksY,[MIKa[Al /(I

+

%[MI

+

Ka[A1)2

(31)

The propene polymerization with V(acac)3/A1(C2H5)2C1 under the condition that -78 and -65OC and [MI = 1.0-8.3 mol/dm3 shows also a lower order as represented bY Rp = Coks(K,,,[Ml/(l

where [A]' is

(

+ K,[Ml))(Ka[Al'/(l + K,[AI'))

[A] - [A],)

as described in Eq.(11) (18).

(32)

The rates of these

two polymerizations were constant in the course of polymerizations, 0.5-7h and

12 T. Keii

1-15h.

So, the

"Trigger Model" seems to be difficult for explaining these

experimental results. Hetero-Site

Yodel and

Homo-Site

we can divided the ongoing kinetic models of into two kind, as follows.

From these above discussion

Yodel

the coordination polymerization

As mentioned above,

all kind models should be

based on the two stes, a vacant site on which a monomer can coordinate and a site on which a growing chain is. sites with

those of

polymerizations

with

We can distinct kinetic models of hetero-

homo-stes. Typical MgC12

example of

supported

the

catalyst

former

or

is

that

the with

V(acac)3/A1(C2H5)2C1, where the vacant site is available only for monomer but not for the coordination of alkyl aluminum.

For the latter the polymerizations

with VC14/Al(C2H5)2Y(Y=C1,Br,Et)(I8) or TiC13/ A1(C2H5)2H(19), the two sites of which are equivalent for coordination of both species , is classified. This distinction of kinetic models is only based upon the experimental rate equation.

Almost of the experimental rate equations reported

for the

polymerizations with TiC13/A1(C2H5)2Y(Y=Et,C1)catalysts are Rideal type('')

= kp[MICoKa

which corresponds to the hetero-site model. This form, howeve, can approximated

also

be

by a Langmuir-Hinshelwood type, Eq.(34), that belongs to homo-

site model, as shown by the author.(20)

AN

UNIFIED

THEORY

OF

NON-UNIFORM

ACTIVE

CENTERS.

Here remain the interesting but not easily understandable experimental results, which concern with "time". The author propose explanation of them on a single basis of

here an unified

a non-uniform active center theory

which is articulated his theory " Intrinsic Fluctuation"(4) with the theory of surface heterogeneity proposed by the author. ( 2 ) Intrinsic

Fluctuation of

Rate

conmtants

on Uniform

Surface

:

I . Kinetics of Propene Polymerization and Unified Explanation

I3

Values of rate constants of surface reactions even on a homogeneous surface such as a crystal plane of single crystal are not constant but distributed around the average values. That is, the idea is that all rate constants are not constants but have a width around their average values. This

is due to that

. l. e . is not onlv one the transition state TS of a surface reaction. in u w but manv because manv number of hetero-site wairs belona to one arowina chain. The same situation can be supposed for TS in homogeneous gaseous reactions of second order, for example, but all structures of such TS are completely continuous and then they are reduced into a structure of TS of the lowest free energy. In the case of surface reaction, reacting species are localized and then many TS of distinguished structures , each of them is in the lowest free energy among its infinite number of deformed structures, are possible, though only the few, TS with the nearest neighbor sites( vacant sites of the growing chain) and with

the second neighbor sites

may take the representative role,

in practice. This conception has been proposed by the author with a name

'I

Intrinsic

fluctuation".( 4 ) The next sample may be useful for understanding the conception. Suppose that the most stable TS of the growing chain is that with a monomer on its nearest neighbor site of distance ro and those with a monomer on a site of distance r from the growing chain. If both the changes of potential energy of Ti-C and monomer-site can be expressed by the same parabola form, the activation energy potential

energy

the potential energy at the point of intersection of two

(

curves) is Ep,ro=(ro/2)2 for

the

most

stable

TS

and

Ep,r=(r/2)2 for the letters. The difference in the energy corresponds to the ratio of rate constant, as kp,./kp, ro=exp( E p , ro-Ep,r)/RT. Number of the latter with distance r, on the other hand, increases as 2rdr/ro2. These relation gives that the probability of TS may be expressed, in terms of its $-value, 'dkp.

Then,

the

averaged

value

of

kp

is

given

by

by 6kp(

kp, ro-

kp,rmax)/log(kp,ro/kp,rmax), where kp,rmax is the minimum value of kp. This is only an artificial example. In general, to determine some rational

form for

such probability would be difficult. However, here we can satisfactorily apply the conception for understanding the problem, as follows. We denote a distribution

of values of a rate constant k around its mean

14

T. Keii

value

k

by

T(k/k)dk/k. This distribution function should be assumed as

applicable for kp, ktr and kd of a growing chain, with the relation, providing that pandpare constants.

(35)

and then

k = kd = k p / p = ktr/pp .

(36)

It should be noted that the distribution, T(k/k), and its integration are not over various growing chains but over various k-values of a growing chain. growing chains

are all equivalent on a uniform surface and then the averaged

polymerization rate, R p ,

where C

*

The

for example, can be expressed by

is the total number of the growing chains.

Dependencies

of

%

on

C*

and t.

analyze first the experimental result,Eq. ( 7 ) ,

Basing upon the above theory,

we

curious dependency of R p on C*,

as below. It is assumed that the deactivation of a arowina chain occurs bv a first order reaction. Denoting the probability that a growing chain is living at t by f(kdt), the total number of living chains on the surface can be given

= C*,

C*t'

Using a

f(kdt)

(38)

first order decay, f(kdt) = exp(-kdt), the average

deactivation of

C *t

or Rd,

ds

rate of

can be given by

J',

R d , t = C*o

the

kde-kdtT(k/k dk/k =-dC *t /dt = C*,( -df ( k d t ) /dt )

The averaged rates of polymerization and transfer at t , R p , obtained, remembering the relations,Eqs.(35)and ( 3 6 ) , as

(39)

and R t r ,

can be

15

I . Kinetics of Propene Polymerization and Unified Explanation

= ~[M]R~,

and

(40)

(41)

Rtr,t = w R d , t

These two expressions guarantee Mn-value independent of time during

the

polymerization of rate decay, as that

We assume here that a general form of deactivation of order y for f(kdt) (l+kd(y-l)t)-'/Y-l = f (kdt)

.

(43)

is that of order 2 - ( 1 / y ) .

The corresponding rate-decay of R p , t

The values of y

in the experiments of many workers,Eqs.(7) and ( 8 ) , are ca.2. For this situation we may propose that the functional form of T(k/k) is exp(-k/k), which results in

y=2.

This functional form has been assumed for explaining the

- -

time-invariant nature of polydispersity, %/% , in quasi-living stages. ( 4 ) Time-invariant MWD

during

quasi-living polymerization

We

start from the basic equations of polymerization of a constant rate, denoting number of growing chains and dead polymers

with n in monomer size by N*, and

Nn.

dN*,/dt

7

dN*l/dt

dNn/dt

kp[M1N*,-1 -

(

kp[M] + ktr)N*,

-

(

kp[M]

= ktr N*n

=

ktrN*n

t

ktr )N*1

n = 1.2,.. .

16

T. Keii

The solution of this set of differential equations has been given by Cabrerizo and Guzman(21), as

Z(N*, + N,)

= C*,

(

1 + ktr t

)

Z (N*, + N,)

= C*,

(

1 + ktr t

)

Z n( N*, + N,)

Zn2(N*, + N,)

=C*o(

1 + ktr t + kp[Ml t

= C*02((k

P

)

[Ml/ktr)2(ktrt- 1 + exp(-ktrt)))

From this solutions we have

%/Mn

= Z n2(N*, t N,

)

1

+ Nn)/( In(N*, + Nn))2

(48)

the value of Mw/Mn, in the region or k t >> 1, changes

The latter gives that from

c ( N*,

P

to 2 as t increases from

accepted conclusion for

t << 1/ ktr to t >> l/ktr , which is widely

homogeneous polymerizations.

Here, the author shows that the time-invariant MWD can be explained by means of the averaging procedure with the use of T(k/k) in the case of y = 2 . Averaged value with the use of

Z ( N*,,

+

+

Zn( N*,, h2(N*n

Nn

+

which give that

Nn Nn

= C*,

)

1 +

= C*o ( 1

) )

(

=

xw/K

C*,

= 2

T(k/k) = exp(-k/k) are

ktrt

)

+ ktrt + kP [MI t

(2(kp[Mlt)2/( 1 +

)

(49)

ktrt))

independent of t. This is an explanation of the

time-invariant nature of the polydispersity during quasi-living stages. The observed

time-invariant nature of

the shape o f

GPC-curve of prod,Jced

polymers,i.e. W(1ogM) against logM, during quasi-living stages can also be explained on the same basis, as that

I . Kinetics of Propene Polymerization and Unified Explanation

17

This result shows that GPC-curve shifts with time by log(t/l+ ktrt) but its shape remain unchanged during the quasi-living stages. Thus, the time invariant property of MWD can be explained on the basis of "Intrinsic Fluctuation" theory. However, the theory can not explain any i.e.

broadening of MWD,

-Mw/En

though its values in this propene catalyst

are

small

3-5. ( 2 )

> 2 , in the heterogeneous polymerizations,

polymerization The

broadening

with of

MWD

MgC12-supported in

heterogeneous

polymerizations, however, has been interested by many workers. As well-known, there are the three

rival theories; diffusion control ( 2 2 ) , chain length

control(23) and non-uniform surface(2) . The mathematical procedures involved in respective theories have been so devised that they lead some averaged values of the ratio of kp[Ml and ktr.

So far i t

concerns stationary MWD, only

established criterion to judge the theories is the effect of hydrogen on MWD. It has been proposed by Roe(24) who proved that hydrogen does not affect on MWD only in the case of non-uniform surface. The present author prefers the view of non-uniform surface, mainly on the experimental confirmation of

no effect of

hydrogen on MWDs in many heterogeneous Ziegler-Natta polymerization systems. The behavior of MWD during quasi-living stages is a new one of criteria for judging the rival theories. It will be shown that the time-invariant broad MWD during quasi-living stages is a decisive evidence of the non-uniform surface theory articulated with the above "Intrinsic Fluctuation" theory.

Non-uniform

Surface: The broadening of MWD can be explained on the

basis of a surface heterogeneity that the catalyst surface is not of a single crystal

but

of

polycrystal.

Assuming

that

a

component

surface

i

is

characterized by k p , (i) and its number of active sites is denoted by C*,, (i), and both ktr

and kd are common for all surfaces, the averages of R Q , t and

C * t , Eq.(40) and Eq.(38), over the whole surface

can be represented by

18

T. Keii

(52)

In the case of quasi-living stages, where f (kdt) i

Eqs.(49) and ( 5 0 ) . remain unchanged excepting that kp

by and . P

the above

1 ,

resul

and kp2is substituted

Then, the averaged polydispersity, <M,/%>=2
P

2>/2,

gives a larger values than 2 in accordance with the width of kp-values over the whole catalyst surface.

The details of this procedure has been described in

the previous paper. ( 2 ) Here, some related problems is discussed. proportional to C*y

in the case of

As it was pointed out,

Rp is

changed with t or CO-poisoning, as shown

C*

by Eqs.(8) and ( 9 ) , whereas it is proportional to C * when C* changed with [A], Eq.(lO). Both effects of t and [CO] on C* were represented by

(C*/C*,)y

and

then they are a l s o effective for i R P,t>, as shown by Eq.(51). The effect of [A], however, may be supposed as different with the both effects of t and [CO]. Since

the effect of [ A ] does not involve any rate constant which is subject of

the averaging with is independent of

T(k/k) but involve an equilibrium constant such as I<,

which

"Intrinsic Fluctuation", the effect, f( [A]), might

be

supposed as a result of alkylation of Ti on the surface. If the total number of potential active surface Ti is Co, Eqs.(lO) and (11) may be understood as that

of usual rate equation with constant kp independent of Co. Finally, an problem

in

application

general

of

the

heterogeneous

known problem, the discrepancy

present

kinetics

between

theory

for

an

unsolved

is proposed. It is a well-

initial kinetics and time-course

kinetics. A typical example is that the initial rate of polymerization in batch system, -(d[M]/dt),,

is a first order with respect to [MIo whereas the time-

course rate -d[M]t/dt is a second order with respect to problem, we may propose an explanation that

For the

assuming a first order reaction

on a uniform surface without decay and averaging [MIt ,

([Mlt=[M]oexp(-k C*,t)) P , and the rate at t , (=k C * o [ M l t = k p C * o [ M l o e x p ( - k p C * , t ) ) , with T(k/k)dk/k, we P

1. Kinetics of Propene Polymerization and Unified Explanation

19

have

-d[Mlt/dt = kpC*,[MlO/(1

+ k p C * o t ~ 2 = k p C * o [ M ] o ( [ M l t / [ M2] .~ )

This is an apparent form of second oder.

(54)

For understanding this, Lipman and

Norrish(26) proposed a sophiscate method that during the time course

[MIt

chnges to [MIo and obseve the rate at t which is smaller than the true initial rate at [MIo and t=O. From the rates, thus observed repeatedly, with [MIo at various time, they attributed the discrepancies from the true inital rate to the deactivation of C*. Their method, however, is meaningless because Eq. ( 5 4 ) gives -d[Mlt/dt = kpC*o[Mlo/(l + kpC*ot~2=kp[MltC*o/(1 + kpC*ot)

The change of [ M I t

into [MIo results in the rate

-d[Mlt/dt = kpC*,[M],/(1

Their conclusion that

+ kpC*,t)

C*t= C * , / ( l

+

(56)

(57)

kct)

and they confirmed that the experimantal value of k, kpC*,,

(55)

is

close to that of

which is natural fiom Eq.(56) and very evidence of their mistake. Such

high speed deactivation of

C*

is not real but only an apparent one.

REFERENCES

1. T.ICeii, E. Suzuki, M.Tamura. M.Murata and Y.Doi, Makromol .Chem. 18 3 ,

.

2 2 8 5 - 2 3 Y . Doi , E .Suzuki , M .Tamura , M Murata,I<. Soga , ibid . 1 8 5,1537 - 1 5 5 7 ( 198 4 )

3 . T.ICeii, M.Terano, I<.l
4 . T.I<eii,"Catalytic Olefin Polymerization"(Proc.1ntern.Symp. on

Recent

Developments in Olefin Polymerization Catalysis,l989) ed.T.Keii and

K.Soga, pl-10(1990), T.l<eii," Olefin Polymerization Catalysts" (Proc.

.

20 T. Keii

1st Korean-Japan Workshop on Olefin Polymerization Catalysts,Korea,l988) ed.D.Lee and K.Soga, p3-19(1990).

5 . T.Keii, "Transition Metals and Organometallics as Catalysts for Olefin Polymerization" ed. W. Kaminsky and H. Sinn(1988), p3-12. 6 . U.Giannini, Makromol. Chem. Suppl.5, 210-229(1981).

7. M.F.Cunningham, J.A.Dusseault, C.Dumas and C.C.Hsu,

I'

Transition

Metal Catalyzed Polymerizations Ziegler-Natta and Metathesis Polymerizations" ed. R.P.Quirk p137-150(1988). 8.

N.Kashiwa, N.I
9. I.Kim, J.Kim and S.Woo,"Olefin Polymerization Catalyst( Proc.lst KoreanJapan

Workshop,Korea,l981)p107-142.

lO.T.Keii, "Kinetics of Ziegler-Natta Polymerization", Chapman Hall (1972) ll.P.J.T.Tait, "Preparation and Properties of Stereoregular Polymers" ed, R.W.Lenz and F.Ciardelli,(1978),p85-112 12.M.Murata, A.Nakano, H.Furuhashi and M.Imai,"Catalytic Olefin Polymerization" ed. T.Keii and K.Soga,(1990),p165-176 13.N.Kashiwa and J.Yoshitake, Polymer Bulletin, 11,485(1984! 14.A.Morinaga, Tokyo Inst. Tech. MS Thesis(1976) 15.M.Ystenes, J.Cata1. 129, 383-401(1991) 16 .T.Keii, K.Soga and N.Saeki, J. Poly. Sci . , C , 16,1507 (1967)

17.Y.Doi. M.Takada and T.Keii, Bull. Chem.

SOC.

Jpn.,152,1802(1979)

18.Y.Doi, S.Ueki and T.Keii, M a c r o m o l e c u l e s , l 2 , 8 1 4 ( 1 9 7 9 )

19.M.Tamura, J. Res. Inst. Catalysis, Hokkaido Univ.,28,No.3 (Prof.J.Horiuti, Memorial Band) 22.Cf.S.Floyd,G.E.Mannand W.H.Ray,"Catalytic Polymer-izdtionof olefins" ed. T.Keii and K.Soga,(1986),p339-367 23.M.Gordon and R-J.Roe, Polymer,2,41(1961) 24.R-J . Roe, Polymer,2,60( 1961 ) 25.L.Reich and A.Schindler," Polymerization by Organometallic Compounds", Intersci. Pub.(1966), p318. 26.R.D.A.Lipman and R.G.Norrish, Proc. Roy. Soc.(London) A275, 310 (1963)

21

2. Active Sites and Mechanisms of Stereospecificity in Heterogeneous Ziegler-Natta Catalysts

Paolo corradini, Vincenzo Busico and Roberta Cipullo Universita' di N a p l i Dipartimento di Chimica via Mezzocannone, 4 1-80134 N a p l i (Italy)

-

-

Introduction Forty years after the discovery, the heterogeneous Ziegler-Natta catalysts are still a hard playground for kinetists and modelists. A molecular description of the kinetic behaviour of such catalysts in 1-alkene polymerization requires indeed to take explicitly into account the presence of several types of active sites and the extent to which, for each type of active sites, structural aspects (e.g. primary or secondary monomer insertion into a growing chain with a primary or secondary last-inserted monomeric unit) affect the rate constants of chain propagation and transfer. Forty years, however, did not certainly pass in vain. The formidable advances, on the one hand, in the microstructural characterization of the polymerization products (mainly by 13CNMR) and, on the other, in the techniques of molecular mechanics, as applied to the models of active centers, allowed a evaluation of substantial progress in our understanding of the catalytic processes and in the rational development of better catalysts. What follows is an up-to-date presentation of the main results obtained by us (1) in the experimental study of the regio- and stereospecificity of propene polymerization in the presence of the MgC12-supported llhigh-yieldllcatalysts, that still monopolize the industrial production of isotactic polypropene (2).

22

P. Corradini. V. Busico and R. Cipullo

The active sites classification

in

Mqcl2-supported

catalysts:

a

Propene polymers produced with MgC12-supported catalysts are usually mixtures of macromolecules, which can be separated into fractions with a stereoregularity varying from almost ideally isotactic to prevailingly syndiotactic (314). It is common practice to take as flisotacticvl the polymer fraction insoluble in boiling heptane (C7), and as "index of isotacticityll (1.1.) of a given polypropene sample the percentage by weight of this fraction (the 1.1. may amount from less than 50% for simple MgC12/TiC14-A1R3 systems to more than 95% for catalysts modified with suitable Lewis bases) (3,4). This practice is legitimated by the observation that, in typical cases, the C7-insoluble fraction has a content of m diads well over 95% (4). On the other hand, 13CNMR data indicate that the C7-soluble fraction contains polymer molecules with (a) stereosequences rich in m diads, though with lower stereoregularity than in the C7-insoluble "isotactict1 fraction (content of m diads between 70% and 90%, indicatively), as well as (minor amounts of) (b) stereo sequences rich in r diads (content of r diads typically between 80 and 90%) (4). We propose to call flisotactoidll the stereosequences of type (a); "syndiotactoidll , those of type (b) It seems plausible to assume that the active sites producing the different types of stereosequences are located at transition metal atoms which reside at the edges of MgC12 platelets and may have more than one coordination environment. In the next two sections, we will give results of the microstructural characterization of propene polymers and of

.

oligomers resulting from the polyinsertion process at early reaction stages. These new results are also in favour of the existence of various types of organometallic species acting as catalytic sites; according to the distribution of

2. Active Sites and Mechanisms of Stereospecificity 23

configurations in the stereosequences produced, such sites will be denoted as isotactic, isotactoid and syndiotactoid. We will then discuss preliminary 13CNMR evidence of the formation of stereoblock macromolecules, that we consider to be associated with interconversion phenomena between different types of active sites. In the last section, models of active sites suitable for site interconversion will be presented.

Regiospecificity of propene polyinsertion It has been shown (5,6) that propene polyinsertion at all three classes of active Centers in MgC12-supported catalysts proceeds in the 1-2 (primary) mode (Scheme 1,a). The 13CNMR spectra of isotactic polypropene samples do not show resonances arising from head-to-head or tail-totail monomer enchainments (7). This implies that the concentration of regioinverted units is well below 1 mol%. It has long been suspected, however, that occasional (2-l)-(secondary-)inserted units can slow-down chain propagation, due to the unfavourable steric contacts required for the formation of a head-to-head enchainment (Scheme l,b and c). This has been confirmed experimentally by the I3CNMR analysis of the polymer end-groups formed via H2-induced chain transfer in the presence of highly isospecific catalysts (Scheme l,d and e) ( 8 , 9 ) .

I Ti -CHz- CH - Po

(313 CH3

I I I+c3b - (311-CH - a-

t

Ti

(312-

(313

I - CH - P.

24 P. Corradini, V. Busico and R. Cipullo

In the limit of zero H2 pressure, 10-30% of these transfer events take place at a 2-1 last-inserted unit (Scheme 1,e) (8). This fraction, which may appear at a surprisingly high, is the result of a (much) first sight higher reactivity to the active metal of the small H2 molecule, compared with that of propene (Scheme l,c), at the crowded Ti-CH(CH3)-CH2-Pn moiety (Pn = polymeryl). The latter, therefore, can be viewed as an active site in a Vlormant" state. With increasing H2 pressure (>lo bar), the molecular weight of the products is decreased at a point that their molecular structure can be investigated by means of GC-MS techniques (9,lO). From the precise evaluation of the ratio between iso-butyl and n-butyl end-groups (Scheme l,d and e) in the llhydrooligomerstl as a function of H2 and propene concentration, it is possible to extrapolate, with an appropriate kinetic treatment of the data ( 9 ) , the ratios of and ksp/kps governing the kinetic constants kPP/kPs formation of the various possible constitutional sequences (Scheme 1). The concentration of regioirregular monomer placements is given by the reciprocal of the former ratio. The * fraction, x (d) of active sites in the lldormantll state, in the limit of negligible chain transfer, can be calculated as

-

-

*

In table 1, the values of k /k ksp/kps and x (d) for PP PS' two typical high-yield catalyst systems are compared with corresponding ones for a typical homogeneous isospecific metallocene-based catalyst .(lo). Table 1

-

Regiospecificity of propene polyinsertion for various stereospecific catalyst systems at 60'C Catalyst system (I) MgCl2/BEHP/TiC1,-PES/AlEt3 (11) MgC12/TiC14-A1Et3 (111) rac-(EBI)ZrC12/MA0

1. 1*103

1. o*102 1. o*102

4.8 1.0

0.2

0.1

0.9

0.5

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

BEHP = bis(2-ethylhexy1)phthalate; PES = phenyl-triethoxysilane; EBI = ethylene-bis(1-indenyl); MA0 = methylalumoxane

2. Active Sites and Mechanisms of Stereospecificity 25

In the case of system I, allowing the synthesis of polypropene samples with an I.I.>97% (ll), the data are to be referred prevailingly to the isotactic sites. 2-1 propene misinsertions take place on average once every l o 3 insertion steps at these sites; this notwithstanding, 20% (indicatively) of them are in the dormant state. The interpretation of the results for system I1 is less straightforward. Polypropene samples prepared with this catalyst system, indeed, have an 1.1. of only 30-40% (4). The C7-soluble polymer fraction, that we assume to be formed prevailingly at the isotactoid and syndiotactoid sites, is characterized by a much lower average molecular weight than the C7-insoluble one (4,12). Kinetic determinations (13) indicate that this is .mainly due to a lower average rate of chain propagation at such sites relative to the isotactic ones ( 1 3 ) , and imply that the number of isotactoid and syndiotactoid sites is much higher than that of the isotactic sites. Under hydrooligomerization conditions, therefore, it can be expected that the low-molecular-weight, volatile products which can be characterized by GC-MS are predominantly produced at the isotactoid and syndiotactoid sites in system 11; the kinetic data in Table 1, therefore, should be referred principally to the set of such sites. These are definitely less regiospecific than the isotactic sites (one 2-1 propene misinsertion on average every lo2 insertion steps), and a higher fraction of them ( 5 0 % , indicatively) are in the dormant state (this may explain, at least in part, the lower apparent propagation rate). The above results also suggest that the activating effect of H2 due to the wakening of dormant sites should be higher for system I1 than for system I, as actually reported (8,141.

26

P. Corradini. V. Busico and R. Cipullo

The origin of the stereospecificity The statistical analysis of the distribution of the stereoirregularities within prevailingly stereoregular poly(1-alkene) chains has proved to be a powerful tool for understanding the origin of the stereospecificity (3,15). A distribution of configurations consistent with the enantiomorphic-sites statistics (16) is indicative of a steric control arising from the chirality of the active sites; a bernoullian distribution of the steric diads (17), in turn, results when the stereospecificity is dictated by the asymmetry of the last-inserted monomer unit. The typical stereodefects in isotactic and syndiotactic polypropene chains produced in these two limiting cases are shown in Scheme 2. In C7-insoluble fractions of. polypropene samples prepared with MgC12-supported catalysts, the stereoirregularities are predominantly of type a in Scheme 2 ( 3 , 4 ) . This proves that the isotacticity is a result of the intrinsic chirality of the isotactic sites. The l3CNMB microstructural analysis of C7-soluble fractions is usually less clearcut, due to the contemporary presence of isotactoid and syndiotactoid sequences with high concentrations of stereodefects (4). However, the recent disclosure (18) of catalyst systems producing C7-soluble polypropene fractions with roughly equal amounts of isotactoid and syndiotactoid sequences of relatively high stereoregularity allowed to prove unquestionably, simply on inspection of the I3CNMR spectra (Fig. l), that at least in such samples the stereodefects are mainly of types a and d in Scheme 2.

-

-

_LLLLtLLL... m m m r r m m m

-

.... (b).+ m m m r m m m m

Scheme 2 Typical stereodefects in isotactic (a),(b) and syndiotactic (c), (d) polypropene chains (adapted Fischer notation). Cases (a), (c) : enantiomorphic-sites control; cases (b),(d): chain-end control.

2. Active Sites and Mechanisms of Stereospecificity 27

This leads to the conclusion that the isotactoid sites have a chirality leading to an enantiomorphic control of chain propagation (as it occurs, to an even higher extent, for the isotactic ones), wheareas the formation of chain-end syndiotactoid sequences is the result of a stereocontrol.

13CNMR evidence of active WgCl2-supported catalysts

site

interconversion

in

The 13CNMR distribution of the steric pentads in C7-soluble polypropene fractions produced with MgC12-supported catalysts can be reasonably reproduced within a ggtwo-sitegg statistical model, assuming such fractions as physical mixtures of enantiomorphic-sitescontrolled isotactoid sequences and of chain-end-controlled syndiotactoid sequences (4,15,19). Some mismatch between experimental and best-fitting calculated data, however, is often observed in the region of the E-centred pentads, particularly when analyzing amorphous (diethylether-soluble) fractions. Doi (19) proposed for the first time that this may result from the presence of steric pentads arising from chemical junctions between isotactoid and syndiotactoid sequences; as a matter of fact, the match improves significantly when this possibility is taken into account in the calculations (Fig. 2) (20).

With the availability of I3CNMR spectrometers operating at very high magnetic fields, it is now possible to expand the microstructural determination on propene polymers at the heptad level at least. The 150 MHz I3CNMR spectra of typical diethylethersoluble polypropenes (Fig. 3 ) confirm that the two types of stereosequences are bonded (at least in part); this is indicated, in particular, by the presence of well resolved and intense resonances attributable to the rrrrmm heptad (20). It should be noted, indeed, that this heptad is not expected to arise from stereochemical errors in the type; isotactoid sequences (mainly of the ..mmmrrmmm..

28

P. Corradini, V . Busico and R. Cipullo L

8

c b b

a

Fig. 1 - 13Ch'MR spectrum (methyl region) of a C7-soluble polypropene fraction obtained with a "highyield" catalyst system including 2,b-lutidine as external donor. Chemical shift scale in ppm downfield of TMS.

I

i

-

Fig. 2 Experimental distribution of the steric pentads f o r a typical ethersoluble polypropene fraction, and bestfitting distributions calculated in the framework of a simple "two-site" model and of a Coleman-Fox type "two-site" model taking into account the possibility of active site fnterconversion.

Fig. 3 - 150 MHz 13CNMR spectrum (methyl region) of a typical ether-soluble polypropene fraction (20).

Chemical shift scale in ppm downfield of TMS.

11.3

21.5

2l.O PL"

20.5

20.0

2. Active Sites and Mechanisms of Stereospecificity 29

Scheme 2,a) nor in the syndiotactoid ones (mostly ..rrrmrrr..; Scheme 2,d), but rather from chemical junctions between the two (..mmmmmrrrrr..). From a straightforward calculation of the concentration of junctions, it is found that both types of stereosequences are rather short (around 10 monomeric units, on average). This seems to be the main reason for the lack of crystallinity in the diethylether-soluble fractions. More surprisingly, a careful I3CNMR characterization allowed us to point out (21) the presence of syndiotactoid sequences, though in very low amount, even in typical such C7-insoluble polypropene fractions (Fig. 4); in 2 samples, long (>lo monomeric units) isotactic stereoblocks appear to be spaced by much shorter syndiotactoid sequences. The disturbance to the crystallization which should arise from the forced proximity of the two different types of stereosequences is concomitant with the formation of appreciable amounts of Y-form crystallinity (Fig.5), quite unusual in polypropene samples of very high stereoregularity and molecular weight and said to be favoured with respect to the a-form by a'limited length of the crystallizable stretches (22). The amount of syndiotactoid sequences was found to decrease when the catalyst system includes an l1externall1 Lewis base, particularly when this is an alkoxysilane (21). The overall picture seems to indicate that the active metal centers at the edges of MgC12 platelets (including the isotactic ones) may change reversibly their environment and the resulting steric control on the polymerization in times which may be shorter (or even much shorter) than those of growth of each single polymer molecule. Mechanisms for explaining the formation of stereoblock structures in the case of the homogeneous anionic polymerization of methyl-methacrylate were proposed already in 1963 by Coleman and Fox (23). A similar occurrence was demonstrated by Zambelli and coworkers for the polymerization of propene in the presence of homogeneous V-based catalysts at low temperatures (24).

-

-

30

P. Corradini. V . Busico and R. Cipullo

Fig. 4 - 13CNMR spectrum (methyl region) of a typical C7-insoluble ("isotactic) polypropene fraction. Resonances arising from (small amounts o f ) syndiotactoid sequences are clearly detected. Chemical shift scale in ppm downfield from TMS. 13C satellite bands are starred.

2. Active Sites and Mechanisms of Stereospecificity 31

Possible models of active sites in Hgcl2-supported catalysts llIdealtl structural models, suitable for the explanation of the various kinds of stereoregularities which may arise in the heterogeneous polymerization of propene, were presented in 1982 by Doi (19). Structural models for the explanation of the isotactic propagation, which at variance with those cited above do not involve necessarily the coordination of Al-alkyls to the active transition metal atom, were proposed by us in refs. 25-27 for TiC13 and extended to MgC12-supported catalysts in refs. 28-30 (for related models of active sites for the new homogeneous metallocene catalysts, see e.g. ref. 31). In our view, the stereospecificity arises from non-bonded interactions between the atoms of the growing chain and those of the entering olefin, which may be conditioned by asymmetries in the neighborhood. In this framework, the isotactic sites (Fig. 6 , A ) would derive from dinuclear Ti2C18 species epitactically placed on the (100) edges of MgC12 layers (28-30). This belief has been enforced by the recent discovery of highly efficient bidentate I1internall1 Lewis bases with the two donating atoms at a distance suitable for the chelation

-

-

of the Mg atoms on the (110) edges of MgC12 layers (30,32), thought to prevent the coordination of TiC14 to such edges on which the less stereoregular polymer would be formed (fig. 7,A-B). The common enantiomorphic-sites-type microstructure of the isotactic and isotactoid sequences suggests that the two types of active sites at which they are formed are similar. Possible precursors of isotactoid sites could be chiral Ti complexes in sterically less conditioning locations, as in the hypothetical mononuclear titanium species on a (110) edge of a MgC12 platelet sketched in Fig. 7,A (29,30). The chain-end origin of the syndiotacticity, on the other hand, could indicate a more open environment of the active metal in the syndiotactoid sites, e.g. as in the hypothetical mononuclear titanium species epitactically coordinated to (100) edges in Fig. 6,B or to (110) edges in Fig. 7,B (29,30,33).

32

P. Corradini. V . Busico and K. Cipullo

The transformation of isotactic Sites h t o syndiotactoid sites (as those of F i g . 6,A-B) could correspond to the reversible dissociation of dinuclear Ti adducts on the (100) edges of MgC12 platelets (28,291: that of isotactoid sites into syndiotactoid sites (as those Of Fig. 7,A-3) could involve slight movements and rearrangements of ligands at a single Ti atom residing on the (110) edge of a MgClz platelet. In line with the results presented in the previous sections, the surface concentration of isotactoid and syndiotactoid sites could be reduced by competitive covering of the MgC12 surfaces with suitable Lewis base molecules ( 28,29).

-

Fig. 6 Models o f a c t i v e titanium complexes on (100) edges of M & l Z p l a t e l e t s , hypothesized for: (A) i s o t a c t i c PrOpagatiOn; ( B ) eyndiotactoid propagation.

-

Fig. 7 Models of a c t i v e titanium complexes on (110) edges of MgClZ p l a t e l e t s , hypothesized for: (A) i s o t a c t o i d propagation; ( B ) syndio t a c t o i d propagation.

Acknowledgements The authors wish to thank Dr. L. CaVallO and Profs. G. Guerra, A.L. Segre, M. Vacate110 and A. Zambelli for useful discussions. Financial assistance from the Italian Ministry for the University and from the Italian National Research council (CNR, Progetto FinaliZZatO Chimica Fine) is acknowledged.

2. Active Sites and Mechanisms of Stereospecificity

33

References (1) For previous publications in the field, see refs. 4,6,9

below (2) a. P.C. Barbe', G. Cecchin, L. Noristi, Adv. Polym. Sci. 81, 1 (1987); b. P.C. Barbe' L. Noristi, G. Baruzzi, Makromol. Chem. 193, 229 (1992) (3) P. Corradini, V. Busico, G. Guerra, llMonoalkene Polymerization: Stereospecificity", in: llComprehensive Polymer Sciencegg, Pergamon Press, Oxford 1988, vol. 4, pp. 29-50 (4) V. Busico, P. Corradini, L. De Martino, F. Graziano, A . Iadicicco, Makromol. Chem. 192, 49 (1991) (5) a. A. Zambelli, M.C. Sacchi, P. Locatelli, G. Zannoni, Macromolecules 15, 211 (1982); b. A. Zambelli, P. Ammendola, Progr. Polym. Sci. 16, 203 (1991) (6) V. Busico, P. Corradini, L. De Martino, Makromol. Chem., Rapid Commun. 11, 49 (1990) (7) T. Hayashi, Y. Inoue, R. Chujo, T. Asakura, Macromolecules 21, 2675 (1988) (8) J.C. Chadwick, A. Miedema, 0. Sudmeijer, Macromol. Chem. Phys. 195, 167 (1994) (9) V. Busico, R. Cipullo, P. Corradini, Makromol. Chem. 194, 1079 (1993) (10) V. Busico, P. Corradini, R. Cipullo, Makromol. Chem.,

Rapid Commun. 14, 97 (1993) (11) A. Proto, L. Oliva, C. Pellecchia, A.J. Sivak, L.A. Cullo, Macromolecules 23, 2904 (1990) (12) T. Keii, Y. Doi, E. Suzuki, M. Tamura, M. Murata, K. Soga, Makromol. Chem. 185, 1537 (1984) (13) a. T. Keii, M. Terano, K. Kimura, K. Ishii, in "Transition Metals and Organometallics as Catalysts for Olefin Po1ymerization1l,W. Kaminsky and H. sinn (Eds.), Springer-Verlag, Heidelberg 1987, pp. 3-12; b. M. Terano, T. Kataoka, T. Keii, J. Polym. Sci., Part A 28, 2035 (1990) (14) G. Guastalla, U. Giannini, Makromol. Chem., Rapid

Commun. 4, 519 (1983) (15) For a recent review, see e.g.: M.W. van der Burg, J.C. Chadwick, 0. Sudmeijer, H.J.A.F. Tulleken, Makromol.

34 P. Corradini. V . Uusico and R . Cipullo

Chem. Theory Simul. 2, 399 (1993) (16) R.A. Shelden, T. Fueno, T. Tsunetsugu, J. Furukawa, J. polym. sci., Part A 3, 23 (1965) (17) F.A. Bovey, G.V.D. Tiers, J. Polym. Sci. 44, 173 (1960) (18) R.C. Job, Int. Pat. Appl. WO 90/12816 (Filing Date: 19 April 1990), Applicant: Shell Oil Co. US (19) Y. Doi, Makromol. Chem. Rapid Commun. 3, 635 (1982) (20) V. Busico, P. Corradini, R. De Biasio, L. Landriani, A.L. Segre, submitted for publication (21) V. Busico, P. Corradini, R. Cipullo, R. De Biasio, manuscript in preparation (22) S. Bruckner, S.V. Meille, V. Petraccone, B. Pirozzi, Progr. Polym. Sci.?16, 361 (1991) (23) B.D. Coleman, T.G. FOX, J. Chem. Phys. 38, 1065 (1963) (24) P. Ammendola, X. Shijing, A . Grassi, A. Zambelli, Gazz. Chim. Ital. 118, 769 (1988) (25) P. Corradini, V. Barone, R. FUSCO, G. Guerra, Eur. Polym. J. 15, 133 (1979) (26) P. Corradini, G. Guerra, R. FUSCO, V. Barone, Eur. Polym. J. 16, 835 (1980) (27) P. Corradini, V. Barone, R. Fusco, G. Guerra, J. Catal. 77, 32 (1982) (28) P. Corradini, V. Barone, R. FUSCO, G. Guerra, Gazz.

Chim. Ital. 113, 601 (1983) (29) V. Busico, P. Corradini, in: "Transition Metal Catalyzed Polymerizations", Ed. R.P. Quirk, Cambridge University Press (New York), 1988, pp. 551-562 (30) E. Albizzati, Chim. Ind. (Milan) 75, 107 (1993) (31) P. Corradini, V. Busico, L. Cavallo, G. Guerra, M. Vacatello, V. Venditto, J. Mol. Catal. 74, 433 (1992) and refs. therein ( 3 2 ) E. Iiskola, A. Pelkonen, H. Kakkonen, J. Pursiainen, T. Pakkanen, Makromol. Chem., Rapid Commun. 14, 133 (1993 (33) A structural model for syndiotactic propene polyinsertion in the presence of Tic13-based catalysts has been proposed in: M. Kakugo, T. Miyatake, Y. Naito K. Mizunuma, Makromol. Chem. 190, 505 (1989)

35

3. Dependence of Transient Comonomer Kinetics on Catalyst Design by Magnesium Chloride Supported Polymerization of Ethene and Propene

K.-A. Sollia, B.K. Vindstadb, T.S. WesterC and M. Ystenes' Borealis AS, N-3960 Stathelle, Norway Statoil R&D, N-7004 Trondheim, Norway Norwegian Institute of Technology, Department of Inorganic Chemistry, N-7034 Trondheim, Norway

a

ABSTRACT Titanium based Ziegler-Natta catalysts supported on ballmilled or precipitated magnesium chlorides have been synthesized and evaluated by heptane slurry polymerization. The instantaneous activity change by addition of propene comonomer to ethene polymerization and vice versa has been analyzed by continuously monitoring monomer consumption. The results show that both the support preparation method and the chemical composition of the support greatly influence comonomer reactivity and kinetics. New details about the transient effect of comonomer addition have been revealed. INTRODUCTION There has been scarce literature articles regarding comonomer 'activation effect' and other kinetic peculiarities in a-olefin polymerization after Spitz et al.') reported their results. We have found that a closer investigation of the transient kinetics give new information. EXPERIMENTAL The polymerization reactors used in this work is shown in Figure 1. Both reactors make use of a hollow shaft stirrer rotating at cu 2400 rpm, which eliminates gadliquid mass transport resistance. The steel/glass reactor on the right perform somewhat better on temperature control due to larger heat conductance, but the kinetics observed in the two reactors is the same. Comonomers are added through the septum inlet using an inerted syringe. The polymerization reactor is connected to ethene/propene supply ('A' - 'D'), a vacuum line ('H'), a thermostated water bath ('N'), and a personal computer ('L') as shown in Figure 2. As the monomer is consumed, it is supplied into the reactor through a mass flow meter ('F'). A total pressure of 1 atm is ensured by passing excess monomer through the glycerol filled bubble flask ('E'). At constant reactor pressure and reactor

36

K.A. Solli. B.K. Vindstad. T.S. Wester and M. Ystenes

Figure 1.

Polymerization reactors.

Figure 2.

Polymerization reactor connections.

temperature (that is, constant monomer concentration), the monomer consumption equals the polymer production (the catalyst activity). A thermocouple ('J') with ice/water cold junction ('0')is used for temperature monitoring. An overpressure security valve ('G')is installed for safety reasons. A detailed description is given by Solli2). Two types of catalyst systems has been investigated in this work. Catalyst A is

3. Dependence of Transient Kinetics on Catalyst Design

37

made by ballmilling anhydrous magnesium dichloride with ethyl benzoate internal donor, and thereafter impregnating with excess titanium tetrachloride at 80 "C. Details are given by Nirisen3). Catalyst B is made by slowly precipitating a magnesium dichloride alcoholate in titanium tetrachloride with diisobutyl phthalate internal donor present. The procedures given by Hu and Chien4) were followed. Polymerizations were performed by introducing heptane, triethyl aluminum cocatalyst, ethyl benzoate external donor (if propene polymerization) and monomer, in that order. After equilibration at 50 "C, the catalyst was added. During an ongoing polymerization the appropriate amount of comonomer was injected through the septum using an inerted syringe. Flow data was collected two times per second. RESULTS Figure 3 shows the ratehime curve from four ethene polymerizations using catalyst A with propene injected. The cu 15 vol% propene addition gave a 14 % instantaneous drop in ethene consumption, followed by a cu 50 % increase after 25 min. Addition of 1-hexene gave similar results (not shown). Figure 4 shows the ratehime curve from three ethene polymerizations using catalyst B with propene injected. The cu 15 vol% propene addition gave a 14 % instantaneous drop in ethene consumption, followed by an cu 25 % increase after 50 min.

Catalyst A wlpropene injection

..._ .. .. .. .. .. .. .. . . . . . ..._.

a 4

c

I

: ,, : ,

I

_.-' -

E

-

L?2 0

0

I

20

.

-

40

?

60

- _ _ _ _--- . -C.

- - - - -.*,

--

,, 0 ;..'_ /

.................... ;j

d -

,__-

, .

-

80

b a -

-

1 0 0 120 1 4 0

Time (min)

Figure 3. Propene addition to ethene polymerization using catalyst A. a - no injection, b - 6 vol% injection, c - 15 vol% injection, d - 2 times 15 vol% injection with 6 min separation.

38

K . A . Solli. B.K. Vindstad. T.S. Wester and M . Ystenes

Catalyst B wlpropene injection m

0,lO iij v

0

20

60

40

80

100 120 140

Time (min)

Propene addition to ethene polymerization using catalyst B. a - no injection, Figure 4. b - 15 vol% injection, c - 30 vol% injection.

s 0

,

,

Added 0.5 vol% e ethylene 1

1

4

,

d 4.0 ~ 0 1 %ethylene 1

I

Time [min]

Ethene addition to propene polymerization using catalyst A. 0.5 vol% and Figure 5. 4 vol% added.

The effect of ethene addition to propene polymerizations using catalyst A is shown in Figure 5 and 6. No instantaneous drop in propene consumption was observed (except for the immediate shock from the injection). A cu 4 vol% ethene addition gave a 50 %

3. Dependence of Transient Kinetics on Catalyst Design

39

:: -.E

10

-

9 -

t

8 -

u,

7 -

3

6 -

m

5 -

-5 -

I

I

0

1

1

20

I

I

63

40

Time [rnin]

Ethene addition to propene polymerization using catalyst A. Repeated Figure 6. 4 vol% additions.

increase in propene consumption during 2 min, thereafter the flow curve was normalized during 10-15 min. Figure 6 shows the effect of repeated additions of comonomer.

0

Figure 7.

10

2G

30 T h e [mln]

40

60

60

Ethene addition to propene polymerization using catalyst B. 6 vol% added.

The effect of adding ethene to a propene polymerization using catalyst B is similar to the effect with catalyst A. A cu 6 vol% ethene addition gave a 60 % increase in propene consumption during 2 min, followed by normalization during 10-15 min.

40

K . A . Solli, B.K. Vindstad, T.S. Wester arid M . Ystenes

0

Figure 8.

10

20

30 Time [rnlnl

40

50

60

Addition of 1,3-butadiene and 1,4-pentadiene to ethene polymerizations.

1 e mmoiea trlethylelunilnlum

1 0 mmolen lrlelhylelumlnlum

1

2 mmoiea

2 mmolea 1.3-butedlene

0

10

20

30

40

50

Tlrne [mlnl

Figure 9.

Addition of 1,3-butadiene and 1,bpentadiene to propene polymerizations.

Addition of diolefins have a substantial deactivating effect, both on ethene and propene polymerizations. Figure 8 shows the effect of adding 1,3-butadiene or 1,4-pentadiene to an ongoing ethene polymerization. The unconjugated 1,Spentadiene shows the largest effect, but it should be noted that even large amounts added do not destroy more than cu 80 % of the activity. The effect on an ongoing propene polymerization is shown in Figure 9. Here, the deactivation of the catalyst system is more or less complete even with moderate amounts of diolefin added. The hindered olefins cis or tram 2-butene have also been added to ethene and propene polymerizations. Addition of tram-2-butene gives almost no effect, while an

3. Dependence of Transient Kinetics on Catalyst Design

41

immediate and permanent deactivation effect of cis-2-butene was observed. The effect was minor, but significant (not shown). DISCUSSION In the flow rate curves from ethene polymerizations with propene added, the immediate shock from the addition is clearly distinguished from the following rapid rate reduction, as exemplified in Figure 10. Presumably, the rate reduction is only due to the dilution of ethene in gas phase, and thereof lower concentration in the slurry. But that implies that propene consumption is slow, and negligible just after the addition. Also, the rate reduction is a bit too slow to be explained by rapid concentration changes. Since also the effect of 2-butene isomers and diolefins is unequal, a possible effect of reduced monomer concentration may be eliminated. A larger deactivation effect of cis-2-butene as compared to trans-2-butene has been predicted by Ystenes’) based on the trigger mechanism6).

40

I

I

I

I

I

44

48

52

56

60

Time (min)

Figure 10.

Detail of flow curve from an ethene polymerization with propene added.

The strong effect of both 1,3-butadiene and 1,4-pentadiene suggests a bidentate complexation of diolefins (i.e. both double bonds), especially because 1,Qpentadiene, the most flexible molecule, deactivates strongest. The effect of the injection on the monomer flow rate curve may be split into four distinct steps, clearly seen for the larger additions of comonomer. First comes the immediate shock of the injection resulting from equilibrating of the reactor pressure through the flow meter and the bubble flask. For large and fast injections a negative flow

42

K . A . Solli, B.K. Vindatad, T.S. Wester and M. Ystenes

is observed (only a second or two). Secondly comes a temporary increase in the monomer consumption rate lasting maximum 1 minute. This feature is also seen when catalyst is present, and is attributed to an increase in the solubility of the monomer. The third step (which is not observed for ethene additions to propene polymerizations) is a permanent rate reduction developed within 2-3 minutes. This effect is attributed to an interaction of the added component with the active sites on the catalyst. The fourth step is slow activation (for ethene addition a normalization). This step seems hard to explain, taking into account that replacement of the gas phase just after the propene addition does a influence the rate increase7). It might be an effect of physical changes in the polymer/catalyst particles, for example increased fracturing (increasing the number of active sites), and changed monomer solubility close to the active centre. The rate increase observed for ethene addition to propene polymerization, results in an additional amount of polymer produced which is several times the amount of ethene added. The difference in observations using the two catalyst systems is probably due to both chemical and physical differences. The mechanisms for these differences cannot be explained without a deeper characterization of the catalyst systems. But the important observation is that differences does exist, and may be controlled by an appropriate catalyst design. REFERENCES R. Spitz, L. Duranel, P. Masson, M. F. Damcades-Llauro and A. Guyot, in 1. "Transition Metal Catalyzed Polymerizations - Ziegler-Natta and Metathesis Polymerizations", R. P. Quirk (ed.), Cambridge University Press, Cambridge, 1988, 719. K.-A. Solli, "Kinetic and Spectroscopic Studies Related to the Influence of Donors 2. on Heterogeneous Ziegler-Natta Catalysts", Dr. ing. Thesis, No. 45, Universitetet i Trondheim, Norges Tekniske Hogskole, 1992. 0. Nirisen, "Polymerization of Propylene on Magnesium Chloride Supported 3. Ziegler-Natta Catalysts", Dr.Ing. Thesis, No. 44 Institutt for uorganisk kjemi, Norges Tekniske Hngskole - Universitetet i Trondheim, 1985. Y. Hu and J. C. W. Chien, Journal of polymer science. Part A. Polymer Chemis4. try, 2003 (1988). M. Ystenes, A. Guyot and R. Spitz (ed.), Die Makromolekulare Chemie, Macro5. molecular Symposia, 66 (1993) 71. M. Ystenes, Journal of Catalysis, 129, 383 (1991). 6. B. K. Vindstad, K.-A. Solli and M. Ystenes, Die Makromolekulare Chemie, Rapid 7. Communications, 13,471 (1992).

43

4. A New Mechanism for Hydrogen Activation in Propene Polymerization Catalysts

A. GUYOT, R. SPITZ and C. JOURNAUD CNRS - LCPP, BP 24 - 69390 VERNAISON, France

ABSTRACT

The two main mechanism suggested to explain the activation effect of hydrogen in propene polymerization are the 2-1 insertion mechanism and the allylic mechanism. The analysis by GC-AED (atomic emission detection) of the gas phase of a polymerization carried out in the presence of deuterium, do support the allylic mechanism. It is suggested that the most probable event after a 2-1 insertion is transfer to monomer after monomer coordination ; that gives an allylic dormant structure which may be reactivated by hydrogen or alkylaluminium. The allylic structure may also be formed after a regular 1-2 insertion.

INTRODUCTION

The activation effect of hydrogen in the polymerization of propene and more over higher aolefins has been observed long time ago even with the first generations Ziegler Natta catalysts '). In the case of modern heterogeneous MgC12 supported catalyst, the first observation was published by Guastalla and Gianini2' who pointed out that the activation effect was similar for both the isospecific ans aspecific sites. The data concerning the "ester generation" of high milage catalysts, where both the internal and external Lewis base were aromatic esters, are not very clear. Soga and Shiono3' reported first some decrease of activity for both atactic and isotactic sites, while Chien and Kuo4' indicate a large increase of the number of active site, ever for ethylene polymerization. Later or, Chien and Nozaki reported successive activation and deactivation in the presence of hydrogen5'. Our data has shown transcient activation by HZfollowed by a very rapid deactivation6'. Finally, Albizatti et a17)have reported that this family of catalyst is more sensitive to deactivation than the simpler catalysts without Lewis base. They explain the deactivation effect by a reaction of the ester with Ti-

44

A . Guyot, R. Spitz and C. Journaud

H bond to produce Ti-OR bonds which cannot be easily exchanged with alkylaluminium. According to the same authors, this deactivation reaction does not take place easily with silanes. Indeed the activation effect is much more clear with “silane”generation of catalysts*’. Conflicting data have been published about the number of active sites : this number is increased upon addition of H2 according to Parsons et a19’ from 14C0 incorporation measurements, while neither the number nor the rate constant are changed, according to Yoshikiyo et all0’, using the stop-flow method. An old explanation”’ reactualized by the previous authors”’, was based on the reduction by Hz of the

p

elimination reaction, which cause inhibiting coordination to the site of the chain end double bond. Recently two more convincing explanations have been proposed, both based on reactivation of dormant sites. The initial proposal of Tsutsui12’followed also by Busico et aIi3-l4)and by Chadwick’” suggest that a dormant site is formed after a regioirregular 2-1 insertion. Our proposal’6’ was that a dormant x allylic site results from a transfer reaction to the monomer. The present paper is mainly a discussion of these two proposals, which actually can be reconcilied.

THE 2-1 INSERTION THEORY Using and homogeneous Zirconocene-methylalurnoxane (MAO) isospecific catalyst, Tsutui et all” did observed a rate enhancement of propene polymerization, by a factor of 3 almost obtained with a ratio H2/C3 of 0.05 in the feeding stream as compared with an experiment without hydrogen.They compared the I3C NMR spectra of the polymer prepared in the absence of H2 or with a flow ratio H2/C3 of 1 . The first spectrum show a few 2-1 internal insertions which have been reduced by 25 % for a HdC3 flow ratio of 0.1 and totally supressed when the ratio was increased to 1. The analysis of the chain end were in line with a transfer reaction to monomer via a p elimination

reaction, and negligible transfer to M A 0 in the absence of H2, while, in the presence of H2, the transfer to H2 takes place preferentially after a 2-1 insertion. It was estimated that 80 % of hydrogen ratio was 1. No data were

transfer took place after such a regioirregular insertion when the given for chain ends for lower H2/C3 ratio.

Later on, the same group” have studied the polymerization of butene 1 with the same catalytic system. They observed a stronger rate enhancement by a factor of 60, again obtained with a low H2/C4 ratio of 0.1. In the absence of Hz, each polymer chain did contain 0.2 to 0.3 regioirregular insertion while the chain end indicate a monomer transfer (with

p

elimination) taking place

preferentially after a 2-1 insertion. No vinylidene end group was observed, which should be expected

4. Mechanism for Hydrogen Activation in Propene Polymerization

45

if the monomer transfer took place after a normal 1-2 insertion. The internal irregular placement where supressed in the presence of H2 (H2/C4= 0.1) and the hydrogen transfer was shown to take place exclusively after a 2-1 placement (each chain does contain 1 butyl and 1 pentyl chain end).

The hydrooligomerization studies of Busico et all3) where carried out using propene polymerization only with MgC12 supported stereospecific heterogeneous

catalysts. The

hydrooligomers where studied by GC-MS and 13C NMR analysis. Two families of products were separated, one containing only 1-2 insertions and the second being terminated after a 2-1 insertion. The data were interpreted assuming 4 possible regio addition (primary-primary kpp, primarysecondary kps ...) and 2 possible termination reactions by H2 (kpH and kSH after primary or secondary insertion respectively). The authors assume no noticeable difference between kpH and kSH and, on this basis, they came to the conclusion that kpp/kps = 1.1 lo3 and kSp/kpS = 5 . Then, after a 2-1 insertion a certain time is necessary before a new regular 1-2 insertion will take place ; in the meantime, if a termination of the chain by hydrogen transfer occurs, then the dormant chain may become active again. The authors estimate that in between 10 and 30 % of the sites are actually dormant. Applying the same analysis to the data of Tsutsui, the same authorsl4) concluded that in the case of the zirconocene homogeneous system, the percent of dormant site was in between 80 and 95 %, the ratios kpp/kpg = lo2and kSp/kpS = 0.1

More recently, Chadwick et all5)have studied the chain ends of polypropylene prepared from MgCI2 supported heterogeneous catalysts with various external Lewis bases. They observed up to 20 % of butyl chain ends, indicative of 2-1 insertion followed by hydrogen transfer. The rate enhancement effect of hydrogen was shown to be dependent on the nature of the Lewis base, and some correlation between the rate enhancement effect and the amount of n-butyl chain end has been put in evidence. In addition, it was shown that the correlation was valid both for the very isotactic xylene insoluble fraction and for the more atactic xylene soluble fraction. The rate enhancement was explained again as the regeneration of active species after hydrogen transfer following a 2-1 insertion.

THE X- ALLYL THEORY Our proposal16' was based on a mechanism previouysly discussed by Marks et al") who studied ethylene polymerization with neodymium based catalysts. These catalysts did produce

46

A . Guyot, R . Spitz and c‘. J o u r n a u d

without any unsaturation, and failed to polymerize any a-olefins, due to the formation of x-ally1 species with saturation of the chain end of the polymer. According to that mechanism, the transfer reaction to the monomer does not take place through j3 elimination, but a hydrogen atom is extracted from the monomer to saturate the polymer chain detached from the active site, the latter remaining blocked by a x-ally1 species.

Scheme 1

monomer coordination

Transition state

reductive elimination

The mechanism is also similar to the o-bond mechanism suggested by Siedlelg’, and in line with the theoritical conclusion of T. Ziegler’’’ according to which a C-H scission mechanism of the monomer leaving an insaturated ligand to the active sites and a saturated polymer chain end is much more feasible than the j3 elimination reaction generally agreed up to now.

Our conclusion were based upon the failure to observe more insaturation in the polymer when the polymerization was carried out at higher temperature, and to the fact that the activation effect of hydrogen was much higher at 90°C than at 70°C for the same heterogeneous catalyst in propene polymerization. According to our proposal, H2 allows the reactivation of the dormant x-allylic site by an oxydative addition.

Scheme 2

4. Mechanism for Hydrogen Activation in Propene Polymerization

47

It should result a o alkyl bond or a monomer molecule and an hydride species as an intermediate step.

It can be also observed that the x ally1 structure of a dormant site allows an easy explanation of the effect of alkylaluminium concentration on the steady level of activity of the catalyst, and also of the activation effect of ethylene in propene polymerization'6).

DEUTERATION EXPERIMENTS

In one of the intermediate steps of scheme 2 , a monomer molecule coordinated to the site is formed. It is not obvious that this monomer should be inserted. It may be exchanged with the medium as well, leaving Ti-H site. In such case a part of that monomer will be firther in equilibrium with the gas phase. Then, if hydrogen is replaced by deuterium,one deuterium atom should be included in that monomer molecule.

A corresponding experiment has been carried out using 1 bar of propene and 4 bars of

deuterium. The gas phase has been analysed by GC coupled with AED (atomic emission detection), allowing to separated the deuterated compounds. A reference experiment under hydrogen has also ben carried out. The chromatogram include 4 peaks assigned respectively to

H2 @2),

ethane,

propane and propene. The enrichment in deuterated propene can be obtained upon substraction of the two chromatograms. The result is shown in figure 1. Enrichment is observed in ethane, propane and propene after the major D2 peak. The two first enrichments might be caused by a very slight hydrogenation activity of the catalyst. The rather small effect on the propene is a direct proof of the existence of an allylic mechanism. However it is difficult to conclude about its importance because nothing is known concerning the competition between the exchange of coordinated monomer and its hrther insertion in scheme 2.

More direct evidence should be obtained from NMR analysis, similar to those performed by Tsutsui or by Chadwick at very high resolution. Our available 250 Mhz apparatus is not sensitive enough owing to the rather high molecular weight polymer produced by the heterogeneous catalyst. Experiments will be carried out in a near hture for analysis of the deuterated polymers.

48 A . Guyot. K. Spitr and C. Journaud 40

z

m

a i m

30

28

1

3.8

3.2

3.4

3.6

3.0

4.0

4.2

4.4

4.6

4.0

5.0

T i m (nin.)

Fig.1 Atomic emission detection spectrum of the gas phase above a propene polymerization : differential spectrum between experiments carried out with hydrogen or with deuterium

DISCUSSION OF THE TWO MECHANISMS

The experiments described by the teams of T s ~ t s u i ’ ~ -B’ ~~’s i c o ’ and ~ ~ ~Chadwick’*’ ~’ clearly demonstrate that, whatever is the catalytic system, after a regioirregular insertion, and in the presence of hydrogen, a transfer reaction to hydrogen takes place preferentially, as compared with a new insertion. The correlation shown by Chadwick between the amount of n-butyl chain ends (caused by the transfer reaction after a 2-1 insertion) and the rate enhancement by hydrogen, is a strong argument to support the link between the two phenomenons. However there are some difficulties. The experiments reported by Tsutsui indicate that most of the activation effect is obtained with a rather small percent ofH2 in the flow (5 % in the case of propene12’*10 % in the case of buteneI7’) while a much higher amount of H2 is necessary to supress new insertions after a 2-1 regioirregular insertion ; so the site is not filly dormant after such a 2-1 insertion. On the other hand,

4. Mechanism for Hydrogen Activation in Propene Polymerization

49

Busico et aIz1’stated that the reactivation of the site after a 2-1 insertion is not enough to hlly explain the activation effect of hydrogen. Another difficult point is the fact that the correlation pointed out by C h a d ~ i c k ’ is ~ ’valid for both the xylene soluble and xylene insoluble fractions of the polymers, which means that there is no special correlation between regioselectivity and stereoselectivity. Further, Chadwick has compared experiments with 2 different concentrations of hydrogen, and did not report data in the absence of H2.

On the other hand, the dormant character of allylic structures in propene polymerization is quite clear. It is known that butadiene is coordinated with polymerization catalyst as an allylic structure ; butadiene is indeed a strong retarder of the propene polymerization22’ ; it can be incorporated as a short sequence at the end of a polypropene chain, so forming a dibloc copolymer ; but fbrther reinsertion of propene units after an oligo-butadiene bloc is not possible. On the other hand, in ethylene-butadiene copolymerization, the butadiene does not strongly retard the polymerization and is inserted in the polymer chain as multiblocs ; the reason is that ethylene is able to insert in an allylic structure of the sitez3’. The activation effect of ethylene in propene polymerization is well known24325’ and we have established that it was competitive and not additive to the hydrogen activation effect*’. This observation prompted us to think that the dormant species reactivated by Hz or by ethylene were the same, i-e an allylic structure formed after a transfer reaction to the monomer (scheme 1). A hrther argument in favor of this transfer mechanism was the fact that, even when the polymerization temperature was as high as 95OC in the absence of H2, there are practically no more insaturation in the polymer, as deducted from IR spectroscopy ; then the

p

elimination is not the dominating mechanism of chain termination. In addition, it has been observed

in our laboratory that the molecular weight is not sensitive to the monomer concentration, which means that the expected effect on the propagation reaction is about compensated by a transfer reaction involving the monomer.

So we think that saturated chain ends can be produced even in the absence of hydrogen if the transfer to monomer takes place as shown in scheme 1. It may be observed that if such a transfer takes place after a 2-1 insertion, a n-butyl saturated chain end will be produced as well as if the transfer is from H2, while it should be an isobutyl chain end from the monomer transfer after a regular 1-2 insertion.

50

A . Guyot, R. Spit7 and C. Journaud

Then it i. clear that the presence of n-butyl chain ends is not a signature of a hydrogen transfer after a 2-1 insertion ; it is actually specific for a 2-1 insertion, but there are at least two possibilities to explain their formation ; there is actually a third possibility after a transfer to alkylaluminium and hydrolysis of the polymer-alkylaluminium. The signature of the hydrogen reactivation is actually the n-propyl chain end, assuming a 1-2 insertion in the metal-hydrogen bond of the metal hydride ; however it can also be objected that the same chain end is formed after a

p

elimination reaction.

The main features of our proposal are summarized in Table 1

Table 1

Last insertion

Chain ends after transfer reaction

Transfer agent

New end

Chain termination Before

After

Reactivation by H2 1-2

2- 1

Monomer

Isobutyl

H2

Isobutyl

Monomer

n-butyl

H2

n-butyl

x-ally1

n-propyl n-propyl

x-ally1

n-propyl n-propyl

It remains to decide what will be the preferential mechanism after a 2-1 insertion. A j3 elimination mechanism on a CH2 group is expected to be more difficult than after a 1-2 insertion where a tertiary hydrogen atom is extracted. A a-CH3 elimination mechanism similar to the P-CH3 elimination mechanism recently proposed and discussed by Pesconi et a126’for metallocene catalysts is also hard to believe in the case of heterogeneous catalysts. The insertion of a new monomer unit is a difficult process, although not impossible because, as reported by Chadwick’” a small amount of regioirregular insertion can be observed in the xylene-soluble fractions. But such an insertion involves, before it occurs, the coordination of the monomers ; this coordination may occur also, not being the more difficult and rate determining step of the whole process of insertion after the 2-1 last insertion. If so, monomer transfer after monomer coordination should be the more probable event,

4. Mechanism for Hydrogen Activation in Propene Polymerization

51

insertion. If so, monomer transfer after monomer coordination should be the more probable event, leading to a x-ally1 structure which is, for sure, a dormant state of the site. On the other hand, is it some reason for the hydrogen coordination to be easier after the 2-1 insertion than after a regular insertion ?. If it was the case, the activation effect would be more directly connected to the amount of 2-1 insertion that actually it is. In any case, after one insertion step, there is a competition between

HZand the monomer to occupy the vacant coordination. The coordination of hydrogen is normally followed by hydrogen transfer, hydride formation and reactivation by monomer insertion in between the Ti-H bond. The coordination of monomer may be followed by a transfer reaction or by insertion. The first case gives a dormant x-ally1 structure, and is dominant after a 2-1 insertion, but occurs also after a regular 1-2 addition. The hydrogen is able to reinitiate the x-ally1 structure, whatewer is its origin.

CONCLUSIONS

Although the 2-1 insertion might be responsible for a rather large proportion of chain termination in propene stereospecific polymerization, it may be not the main explanation of the activation effect of hydrogen. The most probable event after such an insertion should be a transfer to monomer leading to a very dormant x-ally1 structure. The reactivation of the active site, then should be done by alkylaluminium, hydrogen or ethylene, when one or both of the two last compounds are present.

AKNOWLEDGEMENTS

The authors are indepted to ELF-ATOCHEM who have supported this work

REFERENCES

1.

P.R. Srinivasan, R. Shasirakant and S. Sivaram in "Hydrogen effects in catalysis" Z. Paal and

.G. Menon, M. Dekker 1988 p 723-745 and references herein

52

A . Guyot, K. Spitz and C . Journaud

2.

G. Guastalla and U. Gianini, Makrom. Chem. Rapid. Comm. 4 5 19 (1983)

3.

K. Soga and Shiono, Polymer Bull. 8 261 (1982)

4.

J.C.W Chien and C. Ikuo, J. Polym. Sci. Polym. Chem. Ed. 24 2707 (1986)

5.

J.C.W. Chien ant T. Nozaki, J. Polym. Sci. Polym. Chem. Ed. 29 505 (1991)

6

R. Spitz, P. Masson, C. Bobichon and A. Guyot, Makromol. Chem. 189 1043 (1988)

7.

E. Albizatti, M. Gaumberti, U. Gianini and G. Morini, Makromol. Chem. Symp. 48/49 223 1991)

8.

R. Spitz, C. Bobichon and A. Guyot, Makrom. Chem. 190 717 (1989)

9

I.W. Parsons and T. Acturki, Polymer Comm. 30 72 (1989)

10.

K. Imaoka, S. Ikai, M. Tamura, M. Yoshikiyo and T. Yano, J. Mol. Catalysis s;! 37 (1993)

11.

E.M.T. Pijpers and B.C. Roest, Eur. Polym. J. 8 1162 (1972)

12

T. Tsutui, N. Kashiwa and A. Mizuno, Makromol. Chem. Rapid. Comm. 11565 (1990)

13

V. Busico, R. Cipullo and P. Corradini, Makromol. Chem. Rapid. Comm. 13 15 (1992)

14

P. Corradini, V. Busico and R. Cipullo, Makromol. Chem. Rapid. Comm.

15.

J.C. Chadwick, A. Miederna and 0. Sudmeijer, Makromol. Chem. 195 167 (1994)

16.

A. Guyot, R. Spitz, J.P. Dassaud and C. Gomez, J. Mol. Catalysis

17.

M. Kioka, A. Mizuno, T. Tsutsui and N. Kashiwa in "Catalysis for Polymer Synthesis" E.J.

Vanderberg and T. Cheng A.C.S. Symp. Serie 496 72 (1992)

21 (1992)

29 (1 993)

4. Mechanism for Hydrogen Activation in Propene Polymerization

18.

53

C. Jeske, H. Lauke, H. Mauermann, P.N. Swepston, H. Schumann and T.J. Marks, J. Am.

Chem. SOC.107 8091 (1985)

19.

A.R. Siedle, W.M. Lamanna, R.A. Newmark, J. Stevens, D.E. Richardson and M. Ryan,

Makromol. Chem. Macromol. Symp. 66 215 (1993)

20

T.K. Woo, L. Fan and T. Ziegler, Makromol. Symp. (in press) Lecture presented at the Int.

Symp. "40years Ziegler Catalysts" in Freiburg Sept. 1993

2 1.

V. Busico, R. Cipullo and Corradini, Makromol. Chem. 194,1079 (1 993)

22.

J. P. Dassaud, A. Guyot and R. Spitz, Makromol. Chem. 194263 1 (1993)

23.

P. Robert and R. Spitz, Makromol. Chem. Macromol. Symp. f& 261 (1993)

24.

R. Spitz, L. Duranel, P. Masson, M.F. Llauro and A. Guyot, in "Transition metals catalyzed

polymerizations : Ziegler Natta and Metathesis Polymerization" R.P. Quirck ed. Cambridge (1988) p 719

25.

S. Lin, Q. Wa and L. Sun in "Studies in Surface Science and Catalysis

Catalytic Olefin

Polymerization" T. Keii and K. Soga ed. Elsevier (1990) p 245

26.

L. Resconi, F. Pietemontesi, G. Franciscono, L. Abis and T. Fiorani, J. Am. Chem. SOC.114

1025 (1992)

This Page Intentionally Left Blank

55

5. Rata Enhancement Effects in the Prepolymerization and Copolymerization of Ethylene and a-Olefins

P.J.T. TAIT and I.G. BERRY Department of Chemistry, UMIST, Manchester M60 lQD, UK. ABSTRACT Some recent results for the copolymerization of ethylene and a-olefins together with data relating to prepolymerization using propylene, butene-1 ,hexene-1 and 4-methylpentene-1 prior to ethylene homopolymerization are presented. Experimental data using the catalyst systems 6-TiC13. 0.33AIC13 and MgC12/EB/TiC14 are reported and include kinetic data, active centre determinations and morphological studies. These results are discussed in the context of current models involving expanding multigrain microreactors.

INTRODUCTION Rate enhancement effects occur when ethylene and a-olefins are copolymerized by many Ziegler-Natta and Ziegler-Natta related catalysts (1-lo), the presence of an a-olefin leading to a significant increase in the rate of consumption of ethylene.. Whilst this phenomenon has been known for some time, relatively few publications have appeared in the scientific literature. An earlier publication from this group (1) summarized the effects associated with the copolymerization of ethylene and octene-1, and ethylene and 4methylpentene-1 , using heterogeneous catalysts such as 6-TiC1,- 0.33A1CI3, and also supported catalysts such as MgC12/ethyl benzoate/TiCI4 and MgC12/dibutyl phthalate/TiCI,. It was established that the observed rate enhancement effects were associated with the presence of increased numbers of active centres. The present study is concerned also with copolymerization effects, but in addition investigates effects associated with a-olefin prepolymerization preceding either ethylene homopolymerization or ethylene a-olefin copolymerization. a-Olefins such as propylene, butene- 1, hexene- 1 and 4-methylpentene- 1 have been employed as prepolymerization monomers along with catalysts of the type 6-TiC13. 0.33AIC1, - AIEt, and MgC12/ethyl benzoate/TiCI4 - AIEt,.

56

P.J.T. Tait and I.G. berry

EXPERIMENTAL Reagents.

Pentamethylheptane (EC180, Erdolchemie) was supplied by ICI plc

(Chemicals and Polymers) and was stored for 24 h over activated molecular sieves (type 4A), then purged with dried nitrogen prior to use. Ethylene (99.9%) and butene-1 (99.4%) gases were supplied by British Oxygen Co Ltd. Propylene gas (99.97%) was supplied by ICI plc (Chemicals and Polymers). All monomer gases were purified by passage through columns of preactivated molecular sieves (types 4A and 13X). Nitrogen gas ('White spot' grade, 99.99%) was supplied by British Oxygen Co Ltd, and was purified by passage through phosphorous pentoxide, potassium hydroxide and columns of preactivated molecular sieves (type 4A). Catalvsts.

6-TiC13. 0.33A1CI3 (Stauffer AA 1.1) was supplied by ICI plc

(Chemicals and Polymers Division). A magnesium supported MgCl,/ethyl benzoate/TiCI4 catalyst was prepared by ball-milling dried MgCI, with SOCI, under nitrogen (MgCI,: SOCI, molar ratio = 20:l). The ball-mill was rotated for 1 h at 20 "C and then ethyl benzoate (BDH, Analar grade, vacuum distilled from CaH,) added (MgCI, : EB molar ratio = 7.5: l), and rotation continued for 90 h. The product was then reacted with TiCI, (MgCI, : TiCI, molar ratio = 1:2) at 80 "C 2 h with stirring, then filtered at 80 "C and washed with nheptane until the filtrate was free of Ti compounds. The solid was vacuum dried at 40 "C for 70 h and made up into a slurry in EC180. The catalyst contained 1.9 wt % Ti. Polvmerization Procedure.

Polymerizations were performed at 1 atm pressure in

an all-glass reactor consisting of a double-walled vessel of approximately 500 cm3 capacity. Water from a thermostatically controlled bath was circulated through the walls of the vessel to maintain a constant temperature of 60 "C. The stirrer was operated at 800 - lo00 rpm. Gaseous monomer was admitted into the reactor by means of a solenoid valve so as to maintain a constant pressure within the reactor. An electrical pulse system was employed to supply gaseous monomer on demand at a fixed pressure, and to simultaneously monitor the rate of reaction so that the instantaneous rate of consumption of gaseous monomer could be determined. The reactor vessel could be evacuated by means of a rotary oil pump or purged with dry, oxygen-free nitrogen. The order of addition of components in Ziegler-Natta polymerization is very important, especially in Ziegler-Natta copolymerization reactions. The following orders of

5 . Rate Enhancement Effects in Olefin Polymerizations

57

addition were observed with a 4 min interval between injection of reaction components:

For homopolymerization:

diluent: monomer: cocatalyst: catalyst

For copolymerization:

diluent: monomer: comonomer: cocatalyst: catalyst

For prepolymerization:

diluent: monomer 1: cocatalyst: catalyst: monomer 2 after removal of monomer 1

It is important to note that the reactor could be evacuated and purged with dry oxygen-free nitrogen so as to remove any traces of the monomer used in the prepolymerization step. The concentration of comonomer or prepolymerization liquid monomer in the reactor was determined by gas chromatography using a Pye Unicam GCD Chromatograph with a carbowax capillary column. The comonomer contents of the copolymers produced were determined by means of infra-red spectroscopy using a Perkin Elmer 298 spectrophotometer. Active Centre Determination.

Radiolabelled Carbon Monoxide ( 14CO) was

supplied by ICI plc (Chemicals & Polymers) in a 1 .Odm3 cylinder under a total pressure of 3 atm and with an activity of 1.5 mCi.

Active centre concentrations were determined using

procedures which have been described previously (1 1,12).

58

P.J.T. Tait and I.G. Berry

RESULTS KINETIC STUDIES CoDolvmerization Figure 1 shows rate-time profiles for the copolymerization of ethylene and 4methylpentene-1 when using different concentrations of 4-methylpentene-l . Comparative results for ethylene homopolymerization are also shown.

r F; f"E

9 . w

I60

120

80

d 40

0 40

80

120

Time / min

Figure 1. .Plot of rate of consumption of ethylene versus time for the copolymerization of ethylene and 4-methylpentene-1 at 60 "C and 1 atm ethylene pressure. [4-MP-1]: 0 = 63 mmol drn-,; 0 = 379 mmol dm-3; = 631 mmol dm-3: ethylene hompolymerization = 0 . Catalyst system 6-TiCI3~0.33AlCl3- AIEt,: [Ti] = 1.2 mmol dm-3: [Ti] : [All = 1:4. It is evident that the presence of an a-olefin leads to a significant enhancement in the rate of ethylene consumption and that there is a particular concentration of a-olefin which will produce optimum results. The actual value of ethylene consumption in copolymerization was found to depend both on the concentrations of a-olefin and aluminium alkyl. Rate enhancement factors, i.e., ratios of optimum copolymerization rates to corresponding rates in homopolymerization, of 4.7, 17.7 and 15.8 respectively were obtained for the data shown in Figure 1. The influence of aluminium alkyl on copolymerization rates should also be

noted. The variations of the rate of ethylene consumption with variation in both 4methylpentene-1 and aluminium alkyl concentrations are shown in Figure 2.

5 . Rate Enhancement Effects in Olefin Polymerizations 59

[All / mmol W3

Figure 2. A three-dimensional surface plot showing variation in average rates of ethylene consumption with aluminium triethyl and 4-methylpentene-1 concentrations. Catalyst system 6-TiCl,~0.33A1Cl3- AlEt,: [Ti] = 1.2 mmol dm": temperature = 60 "C: ethylene pressure

=

1 atm.

As comonomer concentration is increased at a constant cocatalyst concentration the average rate of ethylene polymerization increases. However as the aluminium triethyl cocatalyst concentration is increased at a constant comonomer concentration the average rate of ethylene polymerization decreases slightly and the rate-time profiles become flatter and

reach lower maximum rates. The level of short chain branching (SCB) was found to increase with increasing concentration of 4-methylpentene-1 and to decrease with increasing concentration of aluminium triethyl. Representative values are listed in Table 1.

60

P.J.T. Tait and I.G. Berry

[AIEt,] I

%(aver) I

SCB level

mmol dm-,

g PE(mmo1 Ti h)-'

i-Bu I lOOOC

4.8

72.4

3.0

Catalyst system 6-TiC1,~0.33A1C13-AIEt,: [Ti] = 1.2 mmol dm": [4-MP-1] = 126 mmol dm": temperature = 60 "C: ethylene pressure = 1 atm. Preuolvmerization Ethylene and 4-Methvluentene- 1 CoDolvmerizationand 4-Methvl~entene-1 Preuolvmerization Figure 3 shows rate-time profiles for ethylene homopolymerization after ethylene-4methylpentene-1 copolymerization and also after 4-methylpentene-1 prepolymerization. Comparative results for ethylene homopolymerization are also included. It is apparent that once rate activation has been achieved it continues at the same level even in the absence of an a-olefin. Whilst copolymerization quickly produces very significant rate activation, that achieved by prepolymerization with 4-methylpentene-l , whilst significant, is still low. 160

-

120

-

80

-

40I w

0

1

40

n

u

- -

I.

a

I

1

1

80

120

Catalyst Age / min

Figure 3. Rate-time profiles showing influence of prepolymerization on subsequent ethylene homopolymerizat ion.

5. Rate Enhancement Effects in Olefin Polymerizations 61

Ethylene homopolymerization after ethylene - 4-MP-1 copolymerization = 0 : ethylene homopolymerization after 4-MP-1 prepolymerization =

: ethylene homopolymerization

=o. Catalyst system 6-TiCl3.0.33A1Et, : [Ti]

=

1.2 mmol dm" : [All = 9.6 mmol dm-3:

[4-MP-1] = 379 mmol dm-3: temperature = 60 "C: ethylene pressure = 1 atm.

4-Methvl~entene-1 PreDolvmerization The effect of 4-methylpentene-1 prepolymerization on subsequent ethylene homopolymerization is shown in greater detail in Figure 4. 4-methylpentene-1 prepolymerization was carried out for 10 rnin and 60 rnin respectively before all traces of 4methylpentene- 1 were removed from the reactor and ethylene admitted. The rate-time profiles are quite stable and show a slight decay type character. It is apparent that rate activation occurs in both cases, being more pronounced for the 60 rnin prepolymerization, but in both cases the rate activation is not to the same extent as in the previous case.

-I

J=

.-

I-

d

0

E E Y

W

a ul \

a

c

UI

0

20

40

80

60

100

120

Catalyst age / m i n Figure 4. Rate-time profiles showing the influence of 4-methylpentene-1 prepolymerization on subsequent ethylene homopolymerization. Ethylene homopolymerization after 4-MP-1 prepolymerization for 10 rnin = =

A

: ethylene homopolymerization =

d

and 60 rnin

o

Catalyst system 6-TiC13.0.33A1C1,- AlEt,: [Ti] = 1.2 mmol dm-3: [All = 9.6 mmol dm-3: [4-MP-1]

=

379 mmol dm',: temperature

=

60 "C : ethylene pressure = 1 atm.

62 P.J.T.Tait and I.G. Berry

Propvlene PreDolvmerization The effect of varying lengths of propylene prepolymerization between 1 and 30 min on subsequent ethylene homopolymerization is shown in Figure 5 . The prepolymerization

rate-time profiles are as before plotted from the time ethylene was admitted to the reactor. It is apparent that significant rate enhancement results which after about 20 min is equal to that produced in the copolymerization of ethylene and 4-methylpentene- 1.

0-09

Figure 5 .

n

v

"

u

"

u

u

u

"

f

-

Rate-time profiles showing influence of propylene prepolymerization on

subsequent ethylene homopolymerization. Ethylene homopolymerization after propylene prepolymerization for 1 min 5 min = A ; 10 rnin = 0 and 22 min =

:

= A

;

ethylene-4-MP-1-copolymerization= 0 :

ethylene homopolymerization = 0 Catalyst system 6-TiCI3.0.33A1C1,- AIEt,: [Ti] = 1.2 mmol dm-3: [All = 9.6 mmol d ~ n ' ~ : 60 "C : propylene pressure

temperature

=

[4-MP-1]

379 mmol drn',.

=

=

1 atm: ethylene pressure

=

1 atm:

Butene- 1 Prepolvmerization and Hexene-1 PreDolvmerization. In order to assess the generality of the earlier discoveries prepolymerization experiments were carried out using butene- 1 and hexene-1 as prepolymerization monomers. Representative rate-time profiles are shown in Figures 6 and 7.

5. Rate Enhancement Effects in Olefin Polymerizations

Timelfrom

additionof monomer

63

2 1 1 rnin

Figure 6. Rate-time profiles showing influence of butene-1 prepolymerization on subsequent ethylene homopolymerization. Ethylene homopolymerization after butene- 1 prepolymerization for 1 min = A ; 3 rnin = A ; 5 rnin = 0 ; 10 min = : ethylene-4-MP-1 copolymerization = 0 : ethylene homopolymerization = 0 . Catalyst system 6-TiC1,.0.33A1C13- AlEt,: [Ti] = 1.2 mmol dm": [All = 9.6 mmol dm-,: temperature = 60 "C : butene-1 pressure = 1 atm: ethylene pressure = 1 atm.

-

7

160.

z

.+ -

B

-

w

a

lZ0:#

," 80

CK

40. n

0

"

n

n

9

n

fi

,

Figure 7. Rate-time profiles showing influence of hexene-1 prepolymerization on subsequent ethylene homopolymerization. Ethylene homopolymerization after hexene-1 prepolymerization for 5 min = A and 22 min = 0 . Ethylene-hexene-1 copolymerization = 0 : ethylene homopolymerization = 0 . Catalyst system 6-TiCl3.0.33A1C1,- AlEt,: [Ti] = 1.2 mmol dm-3: [All = 9.6 mmol dm-3: [hexene-1] = 384 mmol dm": temperature = 60 "C : ethylene pressure = 1 atm.

64

P.J.T. Tait and 1.G. Berry

As before, prepolymerization rate-time profiles are plotted from the time at which ethylene was admitted to the reactor system. Successful removal of hexene-1 after prepolymerization was confirmed by GLC analyses. It is evident that prepolymerization with butene-1 and hexene-1 gives rise to significant rate enhancement for subsequent ethylene homopolymerization. Rate enhancement following prepolymerization with hexene- 1 is considered significant since the hexene-1 prepolymer would be expected to be soluble in EC180 at 60 "C. Effect of Prouvlene Preuolvmerization on Ethvlene-4-Methyl~entene-1Couolvmerization Kinetics The effect of propylene prepolymerization

on ethylene4methylpentene- 1

copolymerization was investigated using two catalyst systems, viz., S-TiCI,.O. 33AlC1, AlEt, and MgCI,/EB/TiCl, - Al(n-Oct)3. Figure 8 shows a representative rate-time profile for ethylene-4-methylpentene-1 copolymerization following propylene prepolymerization for 10 min. Comparison of this ratetime profile with that for normal copolymerization shows that very similar rates of enhancement are obtained after some 20 min copolymerization. However the propylene prepolymerized system shows an activated rate value almost from the start of the copolymerization. Comparative plots for ethylene homopolymerization and ethylene homopolymerization after propylene prepolymerization for 10 min are also shown. Comparative results using a supported MgCl,/EB/'I'iCl,

-

Al(n-Oct), catalyst are

shown in Figure 9. All profiles are plotted from the time of catalyst injection. In this case all the relevant copolymerization and homopolymerization rate-time profiles are of a decay type from the beginning of observation. The normal ethylene homopolymerization rate is substantially

increased

by

copolymerization

with

4-methylpentene-l .

Propylene

prepolymerization for about 5 min produces enhanced rates of subsequent ethylene homopolymerization. Indeed the rate-time profile for ethylene homopolymerization following propylene prepolymerization for 5 min is more or less identical to that for ethylene-4methylpentene- 1 copolymerization.

5 . Rate Enhancement Effects in Olefin Polymerizations

-

"

u

26

46

u

v

v

G

ad

66

Time[ f r o m a d d l l i o n o f

G

s

106

monomer 2 )

=

65

~

126

I rnin

Figure 8. Rate-time profiles showing the influence of propylene prepolymerization on subsequent ethylene-4-methylpentene-1copolymerization. Ethylene-4-MP- 1 copolymerization after propylene prepolymerization for 10 min = 0 : ethylene-4-MP- 1 copolymerization = 0 : ethylene homopolymerization after propylene prepolymerization for 10 min = W : ethylene homopolymerization = 0 . Catalyst system 6-TiC13.0.33A1C13- AlEt,: [Ti] = 1.2 mmol drn-,: [All = 9.6 mmol d m 3 : temperature = 60 "C: prop lene pressure = 1 atm: ethylene pressure = 1 a m : [4-MP-1] = 379 mmol dm'

Y

4000-

2000-

0

3 Catalyst Age / min

Figure 9. Rate-time profiles showing the influence of propylene prepolymerization on subsequent ethylene homopolymerization. Ethylene homopolymerization after propylene prepolymerization for 1 min = A ; 3 min = A and 5 min = 0 : ethylene-4-MP-1 copolymerization = 0 : ethylene homopolymerization = 0 . Catalyst system MgCl,/EB/TiCl, - Al(n-Oct),: [Ti] = 0.08 mmol dm": [All = 8.8 mmol dm3: temperature = 60 "C: propylene pressure = 1 atm: ethylene pressure = 1 a m . [4-MP-1] = 379 mmol dm".

66

P.J.T. Tait and I.G. Berry

ACTIVE CENTRE DETERMINATION Active centre determinations were carried out using a ''CO-radiolabelled technique

(11,12). The relevant contact time, t,, was determined to be 20 min ( l l ) , representing the completion of I4CO insertion into active centres and the absence of further incorporation of 14C0 into polymer chains via side reactions. Hence, Ci, the amount of 14C0 incorporated into the polymer at tc = 20 min, can be related to C*, the number of active centres by the equation:

ci The propagation rate coefficient,

c*

$, was calculated using

% where

=

(1)

the equation:

= $ C * [MI

(2)

is the rate of polymerization and [MI the concentration of monomer.

CoDolvmerization Studies Figures 10 and 1 1 show values of C* and kp respectively as functions of polymerization time

for ethylene homopolymerization and ethylene-4-methylpentene-1 copolymerization when

" fi

1

n

n

66

126

Time/ m i n Figure 10. Active centre concentrations during ethylene homopolymerization (

o

) and

ethylene-4-methylpentene-1 copolymerization ( 0 ). Catalyst system 6-TiCI3.0.33A1Cl3- AIEt3: [Ti] = 1.2 mmol dm-3: [All = 9.6 mmol dm-3:

[4-MP-1] = 379 mmol dm-3: ethylene pressure = 1 atm: temperature = 60 "C: t, = 20 min.

5 . Rate Enhancement Effects in Olefin Polymerizations

67

using the 6-TiC1,~0.33A1C13- AIEt, catalyst system. For this catalyst system about 0.3 -

0.4% of the titanium is active in ethylene homopolymerization whilst the much higher figure of 4-6% is active in ethylene-4-methylpentene-1 copolymerization during a 2 h

polymerization. These results are in excellent agreement with those published earlier by Tait, Downs and Akinbami (1). Figure 11 shows that during ethylene-4-methylpentene-1copolymerization the value

$ is slightly lower than it is during ethylene homopolymerization. A similar finding was reported by Akinbami (13,14). This slight reduction in the $ value may arise from the slow

of

insertion of 4-methylpentene- 1 molecules into the growing polymer chain.

i

o o- e 7w d

0

E

m

600\

a

0

Y

400.

0

0

200-

0

Figure 11. Propagation rate coefficient values during ethylene homopolymerization ( 0 ) and

ethylene-4-methylpentene-1copolymerization ( 0 ). Experimental conditions are as for Figure 10. PreDolvmerization Studies Active centre concentrations were determined at various polymerization times during ethylene homopolymerization times during ethylene homopolymerization following a 10 min propylene prepolymerization. Values of C* are shown in Figure 12 whilst corresponding values of

$ are shown in Figure 13. For comparison data points

for ethylene

homopolymerization performed without a prepolymerization step are included.

68

P.J.T. Tait and I.G. Berry

"

-

A

66

126

-

A

1

C a t a l y s t a g e / min Figure 12. Active centre concentrations during ethylene homopolymerization with ( 0 ) and without (

o ) propylene prepolymerization.

Catalyst system &TiC1,.0.33AlC13- AlEt,: [Ti] = 1.2 mmol dm-3: [All = 9.6 mmol dm-3: propylene pressure = 1 atm: ethylene pressure = 1 a m : temperature = 60 "C: t, = 20 min.

400. 20001

0

sd

120 Catalyst a g e / min

Figure 13. Propagation rate coefficient values during ethylene homopolymerization with ( 0 ) and without ( 0 ) propylene prepolymerization.

Experimental conditions are as for Figure 12.

5 . Rate Enhancement Effects in Olefin Polymerizations 69

These results clearly show that following a 10 min propylene prepolymerization the concentration of active centres during the subsequent ethylene homopolymerization reaction is substantially higher than that during a conventional ethylene homopolymerization. Indeed the values approach those of ethylene-4-methylpentene-1 copolymerization for short polymerization times. DISCUSSION It is now clearly established that the presence of an a-olefin leads to enhancement in ethylene consumption and that this phenomenon takes place with a variety of heterogeneous Ziegler-Natta catalysts, including supported catalyst systems. The results presented here confirm earlier results (1) that this rate enhancement arises from an increase in the concentration of active centres. The results which are presented here, however, illustrate for the first time that the enhancement in ethylene consumption that follows a-olefin prepolymerization, or a-olefin copolymerization, is also due to an increase in the concentration of active centres. It should be remembered that the present work establishes that the a-olefii need not remain present during the polymerisation for rate activation to continue. This indicates that a permanent change in the nature of the catalyst is produced early in the polymerization as a result of the presence of the a-olefin. Furthermore, the varying length of prepolymerization time required by different prepolymerization monomers to produce significant ethylene rate enhancement illustrates that the change which occurs in the catalyst depends on the type of a-olefin which is used, the less active a-olefins requiring longer prepolymerization times. The phenomena of increased numbers of propagating centres in the copolymerization of ethylene and a-olefins with the heterogeneous Ziegler-Natta catalysts have been reviewed previously (1). The following effects may be considered important: (i)

Physical disintegration of the catalyst matrix particles thus exposing new potential centres.

(ii)

Formation of new active centres by reactions involving a-olefins.

(iii)

Diffusion phenomenon associated with monomer and aluminium alkyl; also increased rates of diffusion of ethylene through more amorphous regions of polymer.

(iv)

Activation of dormant or potential active centres by complexation or other reactions with a-olefin molecules - "switching on" effects.

70

P.J.T. Tait and I.G. Berry

It is important to recognise that more than one effect may be responsible for rate enhancement resulting from either copolymerization or prepolymerization and involving aolefins, and also that different effects may operate in different catalyst systems. In addition it is necessary to distinguish between rate enhancement arising from an increase in C* and

that arising from higher

$ values.

In general causes may be helpfully divided into two classes depending on the magnitude of the rate enhancement which is generated. Thus primary rate enhancement, having rate enhancement factors of 3 to 10, or even much higher as in some of the data presented in this paper, e.g., 17.7, may be associated with factors (i) and (ii) above, whilst

secondary rate enhancement, having rate enhancement factors of 2-3, may be associated with factors (iv) and (v) above; such factors may explain rate enhancement in ethylene-a-olefin copolymerization shown by some metallocene catalyst systems (15). For the catalyst systems studied in the present investigation, viz., 6-TiC1,~0.33AIC1, AIEt, and MgCI,/EB/TiCI, - Al(n-Oct),, it is believed that the weight of evidence favours physical disintegration of the catalyst particles as being a major cause of the increased numbers of active centres. Visual observation of particle growth during prepolymerization with propylene shows very small particles, as distinct from the larger particles observed during ethylene homopolymerization. Additionally, particle morphology and particle size distribution are very different for ethylene homopolymerization and ethylene-4-methylpentene- 1 copolymerization. During ethylene homopolymerization gross agglomeration of particles occurs producing large polymer particles, the majority being over 500 microns in size for a 2 h polymerization at 60 "C and 1 atm pressure. On the other hand whilst some agglomeration does occur during the early stages of ethylene4methylpentene- 1 copolymerization, breakdown of the agglomerates occurs and the final polymer morphology is much finer (16,17). For prepolymerization with propylene there are no visual signs of agglomeration, and a product having a fine morphology is produced. Figure 14 shows a schematic representation of processes whereby catalyst fragmentation may occur. A multigrain model for the original catalyst particle is adopted (18). Polymerization of ethylene at active centres located within the body of the original catalyst particle, exposed as a result of copolymerization or prepolymerization, explains the rate enhancement. The growing polymer catalyst particles are seen as expanding microreactors each having its integral identity.

5. Raie Enhancement Effects i n Olefin Polymerizations

1

71

Ethylene Homopolymerization

/ \

Prrpolymerization

Copolymerization

Ethylene

Homopolymerization

Figure 14. Schematic representation of catalyst particle fragmentation and polymer growth.

In keeping with this model a minimum amount of polymer formed during the prepolymerization step is necessary to fragment the catalyst, and the slower rate of polymerization of monomers such as 4methylpentene- 1 explains the need for longer prepolymerization times for such monomers to produce significant rate enhancement. In normal ethylene homopolymerization the initial rate of production of polyethylene molecules leads to the formation of a layer of high density polyethylene covering the catalyst particle surface which constrains polymerization to the external surface, and explains the 'low' rate of homopolymerization. In this respect, it is the rate of polymerization of ethylene which is

anomalous,

REFERENCES 1.

P.J.T. Tait, G.W. Downs and A.A. Akinbami in "Transition Metal Catalyzed

Polymerizations", Ed. R.P. Quirk, Cambridge, 1988, p.834. 2.

P.J.T. Tait in "Olefin Polymerizations", Ed. W. Kaminsky and H Sinn, Springer-

72

P.J.T. Tait and I.G. Berry

Verlag, Berlin, 1988, p.309. 3.

G. Fink and E. Kinkelin in "Transition Metal Catalyzed Polymerization", Ed. R.P. Quirk, Cambridge, 1988, p. 161.

4.

D.C. Calabro and F.Y. Lo in "Transition Metal Catalyzed Polymerizations", Ed. R.P. Quirk, Cambridge, 1988, p.729.

5.

P. Pino, P. Cioni, J. Wei, B. Rotzinger and S . Arizzi in "Transition Metal Catalyzed Polymerizations", Ed. R.P. Quirk, Cambridge, 1988, p.1.

6.

R. Spitz, L. Duranel, P. Masson, M.F. Darricades-Llauro and A. Guyot in "Transition Metal Catalyzed Polymerizations", Ed. R.P. Quirk, Cambridge, 1988, p.719.

7.

A. Munoz-Escalona, C. Alarion, A. Albornoz, A. Fuentes and J.A. Sequera in "Olefin Polymerization", Ed. W. Kaminsky and H. Sinn, Springer-Verlag, Berlin, 1988, p.417.

8.

A. Munoz-Escalona, H. Garcia and A. Albornoz, J. Appl. Sci., 34, 977 (1987).

9.

W. Kaminsky, and H. Hahnsen in "Advances in Polyolefins", Ed. R.B. Seymour and

T. Cheng, Plenum Press, New York and London, 1987, p.361. 10.

T. Tsutsui and N. Kashiwa, Polymer Communications, 29, 180 (1988).

11.

P. J.T. Tait in "Transition Metal Catalyzed Polymerizations, Alkenes and Dienes", Part A, Ed. R.P. Quirk, Harwood Academic Pub., New York, 1983, p.115.

12.

P.J.T. Tait, B.L. Booth, M.O. Jejelowo, Makromol. Chem., Rapid Commun. 9, 393 (1988).

13.

A.A. Akinbami, PhD Thesis, Manchester, 1985.

14.

P.J.T. Tait and A.A. Akinbami, to be published.

15.

P.J.T. Tait and A.A. Abozeid, to be published.

16.

I.G.Berry, PhD Thesis, Manchester, 1990.

17.

P. J.T. Tait and I.G. Berry, to be published.

18.

W.H. Ray, G.E. Mann and S . Floyd in "Catalytic Polymerization of Olefins", Ed. T. Keii and K. Soga, Kodansha, Tokyo, 1986, p.330.

73

6. Characterization of Active Sites in Ti/Hf/MgCl, Catalysts by Chiral Reagents

F. MAS1 AND F. MENCONI EniChem, via Maritano 26, 20097 S.Donato Milanese, Italy A. ALTOMARE, F. CIARDELLI, M. MICHELOTTI, R. SOLAR0 Department of Chemistry and Industrial Chemistry, University of Pisa, via Risorgimento 35,56126 Pisa, Italy

ABSTRACT Previous studies concerning with ethylene and a-olefins homo- and copolymerization in the presence of MgC12-supported Ti/Hf catalysts have shown that the kinetics and polymer properties can be affected by the addition of protic compounds. In order to gain more information about these effects and related modifications of active sites, an investigation was carried out by using optically active carboxylic acids and chiral monomers. The influence of the structure of optically active protic compounds on the preferential polymerization of one monomer enantiomer (stereoelective polymerization) has been related to selective molecular interactions of the different catalytic species present in the catalyst. The obtained results are discussed with reference to the role of different transition metals. INTRODUCTION Previous work dealing with the preparation of ethylene homo-1 and copolymers2 in the presence of MgC12 catalysts derived from two transition metals, namely titanium and hafnium, indicated that a broader family of active sites was present as compared with analogous systems derived from only one transition metal. In particular it was demonstrated, by comparison with model catalysts prepared by supporting

74

F. Masi and F. Menconi

on MgC12 either hafnium or titanium, that the species derived from the former had lower productivity but produced macromolecules having larger molecular weight.3 Moreover, during homopolymerization of a-olefins, in particular propene and 4methyl-1 -pentene, Hf-containing sites appeared to give a much better isotactic specific stereocontrol, good values being obtained also without added Lewis bases.4 The different chemical nature of active sites in these bimetallic catalysts was furtherly demonstrated by addition of organic protic compounds resulting in the narrowing of molecular weight distribution as a consequence of selective deactivation of sites producing shorter chains.5 With the aim of obtaining a better insight about the nature of the interaction between active sites and protic compounds, the present paper reports on a preliminary study with chiral (optically active) carboxylic acids used during catalyst preparation and/or activation. Stereochemical information was derived by using the modified catalysts in the polymerization of chiral (racemic) a-olefins having the asymmetric carbon in the a-(3,7-dimethyl-l-octene) or P-position (4-methyl- I -hexene) to the double bond. EXPERIMENTAL Materials. Commercial (R)(S)-4-methyl- 1-hexene and (S)-3,7-dimethyl1-octene (Fluka) having [a]::+13.3 (optical purity 81.4%) were distilled over Na/K alloy and stored under dry argon. (R)(S>-3,7-dimethyl-l-octene was prepared by a procedure already reported.6 (S)-2-Phenylbutanoic acid having [a1265+7 1.3 (optical purity 90.7%) was obtained by crystallization of the salt of the racemic acid (Fluka) with (S)-a-phenylethylamine.7 Commercial (Aldrich) (R)-3,7-dimethyl-6-octenoic acid (citronellic acid) having [a]LO +8.54 (optical purity > 99%) and Al(i-Bu)3 was used as received. Catalvst preParation. The Hf/Ti catalyst PAR15 was prepared as previously described.* CA catalysts were obtained by treatment with chlorinated aluminum alkyls of Ti and Mg carboxylates in n-hexane solution. The starting carboxylic acids were (S)-2-phenylbutanoic acid (CA/PB4), (R)-citronellic acid (CA/C4) and C 1 1carboxylic acid (CA/V2), respectively.* The presence of the carboxylate ligand in the final catalyst was determined by FT-IR (Fig. 1). A solution of 30 mmol of olefin in 30 ml Polvmerization experiments. anhydrous n-heptane was placed in a glass vial under argon atmosphere, then 3.0

6. Characterization of Active Site i n Ti/Hf/MgCI, Catalysts 75

mmol Al(i-Bu)3, pre-reacted or not with the required amount of the chiral acid, and the catalyst (0,l mmol of transition metal) were added in the order and the mixture was kept under stirring at 20 “C. After a suitable time the suspension was poured into 10 ml methanol and the coagulated polymer was dissolved in chloroform, precipitated in methanol and dried under vacuum. The n-heptanelmethanol mixture was distilled and the distillate, after washing with water and drying over Na2S04, was reacted with a slight excess of bromine. The dibromide was distilled under vacuum and reacted with 1.3 g powdered Zn in 5 ml methanol. The olefin/methanol azeotrope was distilled, washed with chilled water, dried over Na2S04 and finally distilled over Na/K alloy. The recovered olefin was characterized by glc analysis and optical rotation measurement.9 Viscosity measurements were performed in Polvmer characterization. tetrahydronaphthalene at 120°C by an Ubbehlode dilution viscometer. Optical rotatory measurements were carried out on polymer solution in cyclohexane (c = 0.5 g/dl) by a Perkin Elmer Mod. 141 spectropolarimeter. NMR spectra were recorded at 120 “C on polymer solutions in 1,2,4-trichlorobenzene/hexadeuterobenzene9: 1 mixtures by a Varian Gemini 200 spectrometer. IR spectra of polymers were measured on cast films by a Perkin-Elmer Mod. 1600 spectrophotometer. IR spectra of CA/C4 and CA/ PB4 catalysts and of Ti and Mg carboxylates were recorded by a Mattson 7000 FT-IR spectrometer equipped with MCT detector. The CA/C4 and CA/PB4 spectra were recorded on the powdered catalysts by using a Spectratech diffuse reflectance accessory. Ti and Mg carboxylate spectra were recorded on n-heptane solutions. RESULTS In order to evaluate if and how the active sites were influenced by the carboxylate ligand, optically active carboxylic acids were used for catalyst preparation. In particular (+) (R)-3,7-dimethyl-6-octenoicacid (C4) and (+) (S)-2-phenylbutanoic acid (PB) were used. The catalyst prepared by coprecipitation of Ti-carboxylates with analogous Mg-derivatives (Ti/COOH molar ratios: starting 0.25, in the catalyst 1.O; Ti/Mg molar ratio 1) was activated with Al(i-C4H9)3. In some cases the chiral carboxylic acid was also used as an “external additive”, being added to the trialkylaluminum before its interaction with the catalyst. FT-IR analysis indicated the presence of carboxylate in the catalyst, mainly bound to titanium as a bidentate ligand (Fig. l).lo A molar ratio carboxylic acidmi = 1 was detected by chemical analysis.

76

F. Mnsi and F. Menconi

2000

1800

1600

1400

1200

1( 10

Wavenumber Figure 1. Diffuse reflectance FT-IR spectrum of CA/C4 catalyst (a). For comparison FT-IR spectra of Ti-carboxylate (b) and Mg-carboxylate (c) in n-heptane solution are also reported. Racemic a-olefins having the asymmetric carbon in the a- (3,7-dimethyl-loctene, DMO) or P-position (4-methyl-l-hexene, 4MH) to the double bond were used. The chiral effect of the carboxylate on the catalytic centers was evaluated by examining chiroptical properties of unpolymerized monomer (polymerization experiments were interrupted at conversions in the range 5-50%),of the whole polymer and of polymer fractions.9 The catalyst prepared from the optically active acid induces the formation of optically active polymers due to the preferential (stereoelective) polymerization of the R antipode of both 4MH and DMO (Table 1). PB gives better results than C4 thus suggesting the importance of the distance of the chiral carbon from the carboxylic group. This would be clearly effective if either the acid is attached (or complexed) to the active sites or promotes the formation of inherently chiral sites during catalyst

6. Characterization of Active Site in Ti/Hf/MgCl, Catalysts

77

Table 1. Polymerization of (R)(S)4-methyl-l-hexene and of (R)(S)-3,7dimethyl-1-ctene in the presence of catalytic systems obtained by reacting CA/C4 and CA/PB4 with Al(i-Bu)3.a Olefin Catalyst. Molar ratio Temp. Duration C0nv.b [ct]ks [c(]2,5 ("C) (h) (%) rec.rn0n.c polymeIcl Olef./Ti C T l e 4MH CA/C4 3600 20 800 8.0 n.d. n.d. CT2 4MH CA/C4 300 20 8 27.7 -0.004 +0.35 CT3 f DMO CA/C4 300 20 18 6.9 +12.431 n.d. CT4 DMO CA/C4 200 20 24 5.3 4 . 0 0 7 n.d. PB 1 4MH CAPB4 200 20 7 35.6 -0.015 -1.10 PB2 g 4MH CA/PB4 200 20 7 27.7 +0.026 -0.81 PB7 4MH CA/PB4 200 0 6 7.8 -0.008 -3.19 PB8 4MH CA/PB4 200 -30 8 0.6 -0.001 - 6.81 PB3 DMO CA/PB4 200 20 7 6.6 +0.023 -1.66 PB4 g DMO CA/PB4 200 20 7 7.8 +0.034 -1.98 a h n-heptane, molar ratio Al/Ti = 30 if not otherwise stated. bEvaluated as (polymer weight/olefin weight)*l00. CNeat. dIn cyclohexane (c = 0.5 g/dl). eMolar ratio AlfIi = 167. fRun carried out in the presence of (S)DMO having [a]: = +13.27 and optical purity = 8 1.4%. gRun carried out in the presence of a small amount of ethylene. Run

coprecipitation. As in previous cases,ll the effect is larger for DMO than for 4MH, and increases with decreasing the temperature from 20 to -30°C (Table 1). Fractionation with boiling solvents of the polymers obtained at 20°C by polymerizing 4MH with CA/C4 and CA/PB4 catalysts indicates that the stereoelectivity is connected to the stereospecific centers giving high molecular weight macromolecules (Table 2). The lower optical rotation, in absolute value, of the cyclohexane extracted fraction with respect to the diethyl ether extracted fraction is connected to the higher stereoselectivity of more isotactic specific active centers.11 Even in the case of PB, which gives the largest asymmetric effect, the polymer optical rotation is much lower than that obtained by isotactic specific polymerization of a pure olefin antipode. Addition of PB to Al(i-C4H9)3 in molar ratio acid/Al = 0.1 before the interaction with the catalyst not containing the optically active acid (run CA2), gave a very modest effect if any. On the contrary in the case of the bimetallic Hf/Ti catalyst the addition of the external optically active acid PB also in molar ratio PB/AlO.l gave

78

F. Masi and F. Menconi

Characterization of the fraction of poly(4MH) samples obtained in the Table 2. presence of CA/C4 and CA/PB4 catalytic systems. Fraction extracted with acetone ethyl acetate diethylether cyclohexane

Run CT2 %

D99,/D964

14 29 27 30

-

0.77 0.80

Run PBl

Yb

%

n.d. +0.05 35,000 +0.31 +0.49 133,000 +0.38 372,000

7 28 23 42

[a125 a

D997/Dg64 [a125 a

-

0.86 0.89

Mvb

-

+0.04 n.d. -1.23 32,000 -1.54 158,000 -0.96 455,000

aIn cyclohexane (c=0.5 g/dl). bEvaluated from viscosity values in tetraline at 120 "C. poly(4MH) with positive optical rotation. As previously observed for ethylene POlymerization,5 also in this case an increase of Evwas observed. These preliminary data confirm a different interaction of the reaction product of A1R3 and RCOOH with the active sites of Hf/Ti catalyst with respect to Ti-only catalysts. The contemporary use of chiral acids as both internal and external additives for Ti catalyst gave a moderate increase of stereoelectivity (Runs PB 1, PB5 and PB6) without variation of Ev(Table 3). Moreover, when the external acid is racemic Table 3. Polymerization of (R)(S)-4-methyl-l-hexene in the presence of catalytic systems obtained by reacting CAN2, PAR15 and CAPB4 with the Al(i-Bu)3/ (S)-2-phenylbutanoic acid adduct.a Run Catalyst

Molar ratios Duration C0nv.b rec.monom. polymer [ u p [ a l p M"e (S)PB/Al 4MH/Ti (h) CA2 C A N 2 0.1 1000 1.0 22.3 +0.004 +0.03 258,000 PA1 PAR15 0 200 5.5 24.8 320 ,000 PA2 PAR15 0.1 200 5.5 22.3 +0.011 +0.52 350,000 PB1 CA/PB4 0 200 7.0 36.6 -0.015 -1.10 237,000 PB5 CA/PB4 0.1 200 5.0 24.0 -0.013 -1.52 232,000 PB6 CA/PB4 0.2 200 5.0 27.1 -0.015 -1.80 230,000 0.1 200 5.0 22.5 -0.011 -1.29 231,000 PB9 f CA/PB4 ah n-heptane at 20 "C. bEvaluated as (polymer weight/olefin weight)-100. CNeat. dIn cyclohexane (c = 0.5 g/dl). bEvaluated from viscosity values in tetraline at 120 "C. fRun carried out in the presence of (R)(S)PB.

6 . Characterization of Active Site in Ti/Hf/MgCl, Catalysts 79

Table 4. Stereochemical relationship and chiral efficiency in the polymerization of (R)(S)-4-methyl-l-hexenewith (S)PB modified cata1ysts.a Run

Catalyst

(S)PB added to

PA2 CA2 PB1 PB7

PAR15 CAN2 CA/PB4 CA/PB4

AIR3

PB8

CAPB4

AlR3 CA/PB4 CA/PB4 (OOC) CA/PB4 (-30°C)

Recovered 4MH Polymer R,(R) d EF e P,b Prev. chiral. PC , Prev. chiral. R,(S) 0.36 0.13 0.49 0.26

R R

0.03

S S

1.25 0.45 0.89 3.07

S S R R

0.97 0.99 1.02 1.06

0.42 0.15 0.89 3.07

S

4.97

R

1.10

4.97

S 1.36 R 1.03 0.34 CA/PB4 0.43 & AlR3 PB9 CA/PB4 CA/PB4 0.36 S 1.24 R 1.02 1.24 a In n-heptane at 20°C. boptical purity of unreacted 4MH. C Optical purity of polymerized 4MH. d Evaluated as (100 - Pp)/(100 + Pp). e Chiral efficiency, evaluated as EF = Pp*[PB]/[Ti]. PB5

CA/PB4

(run PB9) an appreciable stereoelectivity is still observed thus excluding an exchange between external and internal species with massive transfer of RCOO- residues from A1 to active sites and back. Even if in the case of 4MH the asymmetric effect is only partial, it appears however rather interesting to make some quantitative considerations on the relative polymerization rate of the two monomer antipodes and the efficiency (EF) (Table 4). When the optically active acid is used during catalyst preparation, stereoelectivity [Rp(R)/Rp(S)]and in particular chiral efficiency (EF) are larger than when the same acid is simply added to AlR3. However, the ratio Rp(R)/Rp(S)remains, even at low temperature (runs PB7 and PB8), well below the value expected for the attainment of a single configuration catalyst (about 1.80).11-13 Therefore these data can be thus better explained by assuming a chiral perturbation of the stereospecific sites by the optically active carboxylate ligands. The higher efficiency of the internal acid also indicates that when the carboxylic acid is added to AlR3 no stereoelectivity is obtained with titanium, suggesting a different type of interaction by the aluminum

80

F. Masi and F. Menconi

carboxylate and excluding the exchange of ligands. Finally the observed differences between Ti-catalyst and Hfni-catalyst confirm the different capacity of Hf and Ti containing sites to interact with the reaction product between AIR3 and carboxylic acids.5 ACKNOWLEDGEMENTS Partial financial support by MURST-Rome (60%)is kindly acknowledged. Mr. A.Moalli and M t A.Vignati (EniChem) are gratefully acknowledged for their technical assistance. REFERENCES 1. EMasi, S.Malquori, L.Barazzoni, C.Ferrero, A.Moalli, EMenconi, Rhvernizzi, N.ZandonB, A.Altomare, ECiardelli, Makromol. Chem. Suppl. 15, 147 (1989). 2. A. Altomare, R.Solaro, ECiardelli, L.Barazzoni, EMenconi, S.Malquori, F.Masi, Polymer Commun., 33, 1768 (1992). 3. EMasi, S-Malquori, L.Barazzoni, R.Invernizzi, A.Altomare, ECiardelli, J . Mol. Cut., 56, 143 (1989). 4. A.Altomare, R.Solaro, ECiardelli, L.Barazzoni, EMenconi, S.Malquori, EMasi, J . Organornet. Chem., 417,29 (1991). 5. C.Ferrero, F.Masi, EMenconi, L.Barazzoni, A. Moalli, A.A1tomare, R.Solaro , ECiardelli, Polymer, 34, 3514 (1993). 6. P.Pino, ECiardelli, G.P.Lorenzi, Makromol. Chem., 70, 182 (1963). 7. A.Fredga, Ark. Kemi, 7,241 (1954). 8. EMasi, L.Barazzoni, EMenconi, R.Invernizzi, S.Masini, P.Ferrero, A.Moalli, (EniChem Polimeri), Eur. Pat. Appl. Publ. No. 523,785 (1993). 9. ECiardelli, CCarlini, A.Altomare, EMenconi, J.C.W.Chien, in “Transitionmetal catalyzed polymerizations”, R.P. Quirk, Ed., Cambridge University Press, Cambridge, p. 25 (1988). 10. G.B.Deacon, R.J.Phillips. Coordination Chemistry Reviews, 33,227 (1980) 1 1 . P.Pino, A.Oschwald, ECiardelli, C.Carlini, E.Chiellini, in “Coordination Polymerization”, J.C.W.Chien, Ed., Academic Press, p. 25 (1975). 12. J.C. W.Chien, J.C. Wu, C.I.Kuo,J. Polym. Sci., Polym. Chem. Ed., 20,2445 ( 1982). 13. J.Vizzini, ECiardelli, J.C.W.Chien, Macromolecules, 25, 108 (1992).

81

7. A New Polymer-Supported Catalysts for Olefin Polymerization

L. Sun, A . Shariati, J.C. Hsu and D.W. Bacon Chemical Engineering Department, Queen’s University, Kingston, Ontario, K7L 3N6 Canada ABSTRACT A novel Ti-based catalyst supported

on poly(ethy1ene co-

acrylic acid) has been developed for olefin polymerization. The catalyst was tested for its performance with ethylene homo- and co-polymerizations with 1-hexene. The catalyst showed very high activity, comparable to those used in commercial processes, and no appreciable deactiviation was observed in two-hour polymerization. Polyethylene of high melting point, high degree of crystallinity, and molecular wieght were obtained. These properties can be regulated by employing hydrogen and comonomer. INTRODUCTION Since the development of highly active supported ZieglerNatta (2-N) catalysts, a great deal of attention has been focussed on the improvement of the effectiveness of catalyst supports’’Z’. Until very recently, the most successful support materials used in Z-N catalysts have been either magnesium chloride or silica gel, which show very high activities, as documented in many reference^^,^) . These supports are effective, but also have drawbacks. Therefore, the search for more effective catalyst supports is still of both academic and industrial interest. Recently, the use of polymer as support has appeared in both patents5’and open literature6,”.However, the productivities of those reported polymer-supported catalysts for batch polymerization are rather low. The limitations are obvious for many of these catalysts due to their low activity and/or complications in catalyst preparation. In this paper we present a novel catalyst system of high activity prepared conveniently using a commercially available polymer as the support.

82

L. Sun. A . Shariati. J.C. Hsu and D.W. Bacon

Experimental Catalyst Preparation and Characterization The polymer support used was poly(ethy1ene-co-acrylic acid)

obtained from Polyscience Inc. It was dissolved in hot toluene and reprecipitated using a large quantity of methanol. After filtration and being dried, the polymer was ball millled for 120 hours with dry heptane to produce a finely dispersed suspension. 20 mL of di-n-butyl magnesium-triethyl aluminum complex in heptane from Texas Alkyls ( 0 . 5 1 8 mmol Mg/mL of heptane with Mg/A1=8.51) were reacted with 5 0 mL of the suspension at 60°C with adequate agitation for one hour. The reaction mixture was filtered and the solids were washed with dry heptane. The resulting modified polymer support was reacted with 30 mL of TiC1, at 60°C for one hour with vigorous stirring. It was filtered out and washed thoroughly with heptane to produce the pol-ymer-supportedcatalyst. The surface area and chemical composition of the catalsyt were obtained with following results: m2.g-l 1 . 5 7 wt%

Specific surface area by BET method Ti content by the colorimetric method

7.9

Mg content by atomic absorption C 1 content by neutron activation

1.01 wt%

2.4 wt%

Figure 1 shows the IR spectra of the polymer (curve 1) and the catalyst (curve 2 ) . The support shows a strong absorption band at 1 7 0 2 cm-' which can be attributed to C=O stretching. After reacting with Mg and Ti compounds the 1 7 0 2 cm-' band disappeared completely and a series of smaller new absorption bands appeared at 1676, 1 6 3 0 , 1 5 6 7 and 1523 cm", which may be attributed to the complexs of functional groups with titanium and/or magnesium compounds. The evidence is, therefore, clear that the titanium catalyst is coordinated on the polymer support via functional groups. Figure 2 shows the particle size distribution of the catalyst.

7 . Polymer-supported Catalyst for Olefin Polymerization

I

I

1676

100,

I

83

)w---l

I

I

i yd O L 00 1800

1600

1400

rli IIlc

polymer cupport I l l and

0 2

0 3

04

Particle Size, mm

Wave Numbers (ern.') I.ig I IR 5pcclra

01

1200

llle calalyri

(2)

Fig.2 Particle size distribution ot catalyst

Ethylene Homo- and Co-polymerization with I-Hexene To evaluate catalyst performance, slurry polymerizations of ethylene were carried out in a 500 mL stainless steel autoclave equipped with a MagnesDrive agitator. The reactor was operated in a semi-batch fashion under a constant pressure of 2 bar. The polymerization rate was measured based on monomer feed rate by a mass flowmeter. The signals from the flowmeter were fed to a personal computer where the data were processed to generate ploymerization kinetic profiles. Figure 3 shows the effects of catalyst particle size on ethylene homopolymerization, which indicate that the particle size affects the polymerization only at the initial stage of polymerization (first 20 min). The data illustrate a substantial increase in maximum rate when the particle size decreases from 420 pm to 1 2 0 pm. The convergence of rates beyond 20 min implies possible particle fragmentation. The relationship between monomer partial pressure and polymerization rate is nearly linear, as shown in Figure 4. There is a tendency, however, that at higher pressures catalyst deactivation becomes increasely important. The individual effects of cocatalyst, temperature, Al/Ti ratio, hydrogen partial pressure, and comonomer concentration were reported in a previous publication''. Briefly, the most effective

84 L. Sun. A . Shariati, J.C. Hsu and D.W. Bacon

so

I

I

250 micron

Fig.3 Effects of catalyst particle size on ethylene polymerization rate. Polymerization conditions: 70"C, AI/Ti = 161, P,/P, = 8.2, particle size 250 pm.

Time, min Fig.4 Effects of ethylene partial pressure on polymerizatlon. Polymerization conditions are the same as given in Figure 3.

7. Polymer-supported Catalyst for Olefin Polymerization

85

cocatalyst was found to be trihexyl aluminum. The optimum temperature for catalyst acivity is at about 80°C. At that temperature the catalyst was stable during a two-hour polymerization. Outside the optimum temperature, not only the overall rate suffered, but also showed significant catalyst deactivation. The best activity seems to lie in the range of Al/Ti ratio of 50-100. In the presence of hydrogen, the rate of polymerization was enhanced slightly up to approximately PH,/P,,,,=0.O5; beyond that ratio, the catalyst activity was depressed gradually. In copolymerization of ethylene with 1-hexene, the polymer yield, as shown in Figure 5, increased ten-fold as 1-hexene concentration increased to 0.2 mol/L. The increase in rate has been a well established phenomenon.

0.0

0 1

0.2

0.3

0 4

[I-Hexene]. mollL pig.5 Dependence of catalyst yield on I-hexene concentration. [El =O. I I 2 molil.. catalyst 10-15 mg; AliTi = 100: T = 60°C; polymerization time = I S m1n

To study the individual and joint effects of these variables on polymerization, polymerization experiments were further carried out according to a central composite design for five factors. The five factors studied were temperature, alkyl concentration, Ti concentration, P H Z / P C Z Hand 4, comonomer concentration. The assigned

86

L. Sun. A . Shariati. J.C. Hsu and D . W . Bacon

values for these five factors are given in Table 1. The results of statistical analysis were summarized in the following equation: Catalyst Activity = 41.5-2.06[Ti]- 7 8 . 8 ( P H 2 / P c 2 , , 2+2.34 ) [C,H,,l (1) -0.025(TI [C6H,,1-0.020[All [C6Hl,l Interestingly, the individual effects of temperature and alkyl Table 1.

The assigned values for the five factors in the central composition design for the study of copolymerizsation.

No.

Factor

-2

-1

0

1

2

1

Temperature, "C

50

60

70

80

90

0.8

1.6

2.4

3.2

4.0

0.0132

0.0265

0.0398

0.0531

0.0663

2

A 1 Conc.,

3

Ti

mmol/L

Conc., rnmol/L

4

p,,, / p - l i t $

0

0.05

O.L0

0.15

0.20

5

Cornonorner, r n o l / L

0

0.048

0.C96

0.144

0.192

aluminum are not significant, but through a joint effect with comonomer concentration. Polymer Products Characterization

Homo- and co-polymers of ethylene with 1-hexene were synthesized under different polymerization conditions. Some properties of the homopolymer under different hydrogen partial pressures are given in Table 2. Apparently the crystallinity is not affected by the hydrogen pressure, but the MP decrease slightly with hydrogen. Without hydrogen the melt index could not be measured due to high molecular weight (more than one million). The effects of T, Al/Ti, and comonomer on copolymer composition are given in Table 3. The comonomer composition remained fairly constant with temperature, but comonmer content increased with Al/Ti ratio and also with comonmer partial pressure, as one would expect. The comonomer sequence distribution, determined from NMR

7. Polymer-supported Catalyst for Olefin Polymerization

87

analysis, is given in Table 4. Virtually no HHH sequences are detected. The sequences of HHE, EHH and HEH were also found to be relatively moinor. The melting point and crystallinity of the copolymer, determined by DSC, have a large dependence upon the Table 2

Some properties of polyethylene synthesized with polymer-supported catalyst

H2

MPd

Crystal.

Melt Index

Dens i ty

( mmHg 1

(OC)

(5)

(g/10 min)

(g/cm’)

. .

0

141.6

74.3

not det

81

138.7

68.5

not det

160

136.8

71.6

0 . 4 3 ~

0.952

227

136.9

75.9

O.055b; 0 . 6 4 ~

not det

B u l k density (g/cm31

not det.

0.33

not det.

0.32 0.31

.

not det.

a: measured by DSC: b: measured by ASTM method 01238 at condition E (190/2.16).c: measured by ASTM method 01238 at condition N (190/10). Polymerization conditions: Catalyst 10-15 mg; Al/Ti=100: time=2 h.

Table 3. Effects of temperature, Al/Ti, and comonomer concentration on copolymer composition.

heptane 250 mL:

8R

L. Sun. A . Shariati. J.C. Hsu and D.W. Bacon

content of comonomer units incorporated into the polymer as shown in Table 5 . As the 1-hexene content increases from 0 to 5 . 5 mol%, the melting point decreases from 1 4 1 . 7 " to 124.0°C, and crystallinity drops from 7 2 . 2 % to 3 1 . 7 % . An SEM photograph of the copolymer shows a more rubber-like morphology with numerous colddrawn threads, demonstrating a more amorphous structure comparing to that of the homopolymer. Table

4.

Comonomer sequence distribution in ethylene-1-hexene copolymers

0 . 11 2 0 !I?

0 064 0 16

98.7 97.2

0 II?

0 32

94 5

Table

5.

Exp. no.

I3 2 8 5.5

0.0

0.2 05

2 3 4 5 6

97 4 94.6 89.5

0.0 0.0 0.0

0.0 0.4 1.0

I 3 2.4 4 5

0.0

2 5

00

5 2

(1.9

82

96 2 O? 1) X5 3

DSC analysis of ethylene-1-hexene copolymers [C&I

a

(mol/L) 1

2.6 5.2 10.0

0.0 0.029 0.064 0.096 0.16 0.32

Melting point

Crystallinity

(OC)

(%)

141.7 136.0 130.9 129.8 126.7 124.0

72.2 52.5 50.6 47.0 43.7 31.7

a) Comonomer concentration in polymerization runs; [C,H,] = 0 . 1 1 2 mol/L in all runs.

7. Polymer-supported Catalyst for Olefin Polymerization

89

Conclusion

A new catalyst supported on poly(ethy1ene co-acrylic acid) copolymer has been developed, which is capable of producing polyethylene with high molecular weight and high crystallinity, and of copolymerizing ethylene and 1-hexene and higher a-olefins. The activity is comparable to commercial catalysts, and under optimum polymerization conditions, no appreciable deactivation was observed during two-hour polymerization. References 1. U. Giannini, A. Mayr, P. Longi, E. Susa, V. Davoli, D.

Deluca, A. Leccese, and A. Pricca, Ger. Offen 2,125,107 (1971). 2. E. Kinkelin, G. Fink, B. Bogdavonic, Makromol. C h e m . , R a p i d Commun. 7, 85 (1986). 3. C . Dumas and C .C. Hsu, JMS-Rev. Macromol. Chem. P h y s . , 2 4 ( 3 ) , 355 (1984). 4. B.E. Wagner, G.L. Goeke, F.J. Karol, and K.F. George, Eur. Pat. Appl. 0,055,605 (1981). 5. R.P. Nielson, US Paten 4,477,639 (1984); T. Sasaki, T. Ebara, H. Kora, K. Kawai, M . Yamasaki, and S . Kawamata, US Patent 5,051,484 (1991);A.B. Furtek and B.Z. Gunesin, US Patent 5,118,648 (1992); B.E. Hoff, US Patent 4,268,418 (1981); S.A. Bedell, W.M. Coleman, and W.R. Howell, Jr., US Patent 4,623,707 (1986). 6. A.M. Bochkin, A.D. Pomogailo, and F.S. Dyachlovskii, R e a c t i v e P o l y m e r s , 9 , 99 (1988). 7. V.A. Kabanov, I.A. Litvinov, T.V. Budantseva, V.I. Smetanyuk, D o k l . A k a d . N a u k . S S S R , 2 6 2 ( 5 ) , 1169 (1982). 8. L. Sun, C . C . Hsu, and D.W. Bacon, J. Polym. S c i . , P a r t A : P o l y m e r Chemistry. In press

This Page Intentionally Left Blank

91

8. Active Center Determination in Ziegler-Natta an Innovative Dual-Labeling Approach

Polymerization:

SENG-NEON GAN, PATRICK S.T.Loi, SWEE-CHENG NG AND D.R.BURFIELD Chemistry Department, University of Malaya, 59100 Kuala Luopur, Malaysia. ABSTRACT Widely used radiotracer nethods for active center determination, employing 14C0 tags and tritiated alcohol quench, generally provide conflicting results, thus casting doubt on the validity of both methods. This paper seeks to harmonize the apparent discrepancies by comparative studies on the polymerization of ethylene, propylene and 4-methylpent-l-eneI catalyzed by TiC13.AA/A1R3, employing an innovative duel-labeling approach. I t is concluded that quantitative 14C0 tagging of active site requires short contact times and high 14C0 concentrations to minimize interfering side reactions. The tritiated alcohol quench, after the tagging reaction, could provide information on the extent of transfer reaction. Under optimized conditions, tagging and quenching procedures can provide comparable and meaningful data. INTEODUCTION An important aspect of a Ziegler-Natta catalyst is the determination of the number of active centere. Over the past 30 years, a variety of nethods have been proposed. These have been comprehensively re~iewedl-~).The most widely used include: kinetic , inhibition, quenching, and radiotagging methods. Application of these and other methods have led to a wide range of values of active centers, and it is not clear whether the discrepancies is due t o the variation in catalyst activity or the deficiency of the methods. Two particularly sensitive methods which have been widely studied are: (i) quenching methods where the polymerization systen is

92

S.N.Gan, P.S.T. Loi, S.C. NG and

terminated with an excess of

D.R.Rurlield tritiated alcohol, leading to the

destruction of the active site and concomitant labeling of the growing polymer chain, viz: -Ti-P t ROT 4 -Ti-OR t T-P (T=tritium) (1) and, (ii) radiotagging methods where the growing chain interacts with a catalyst poison such as 14C0 whereupon polymerization is halted, and the polymer is tagged through interactions such as:

*co

-Ti-P

0

" + -Ti-* C-P

HX 4

0

-Ti-X t

1' H* C-P

(2)

Both methods have some drawbacks. The quenching approach leads to labeling of non-active polymer chains4) which have undergone transfer reaction with the aluminum alkyl. -Ti-P t -Al-Et 4-Ti-Et t -A1-P I t has to be corrected for kinetic isotope effect ( K I E ) 5 ) .

(3) With 14C0

tagging there is some controversy over the efficiency of the process since reactivation of the catalyst has been demonstrated6). The present study has been designed to examine the compatibility of the two methods through an innovative dual labeling procedure. The active chains are first tagged with 14C-labeled carbon monoxide and then quenched with tritium labeled methanol to provide labeled polymer chains containing both 14C and T ( 3H ) isotopes. The radioactivities of the two isotopes may be assayed individually by liquid scintillation technique because of the differences in the Denergy spectrum. EX PER1MENTAL Materials: Tritiated methanol was synthesized by the tritium exchange between tritiated water and methanol. It was then dried with activated 3A molecular sieves (10% w/v). l4CO (specific activity 4.44 X 10"

dpm/mol) was purchased from New England Nuclear.

Propylene and 4-methylpent-1-ene were purchased from Matheson Chemical Company, U.S.A. Ethylene was purchased from Malaysian Oxygen Company. TiC13.AA was a gift from Stauffer, U.S.A.

Aluminum alkyls

were purchased from Schering, Germany. Solvents and gases were purified by standard procedures and dried over freshly activated molecular sieves.

8. Active Center Determination in Ziegler-Natta Polymerization 93

Polymerization and Labeling Procedure: Polymerieation was carried out in a 2-necked flask fitted with a suba seal cap. Typical concentrations used were [TiC13.AA] = 25 mM, [A1R3] = 95 mM in toluene, and monomer pressure of 1 atmosphere, in the case of propylene o r ethylene. In the polymerization of 4-methylpent-l-eneI the liquid monomer was syringed into the flask to make up a 2 M solution at the beginning of the reaction. The reactor was thermostatted at 30'C

and stirred magnetically. Polymerization time

was generally fixed at 10 minutes before carrying out the labeling procedure. In the quenching experiments, tritiated methanol was added at the specified times. The amount was calculated based on a 2-fold excess of total aluminum-alkyl bond present. The polymer was precipitated into excess acidified methanol, washed, and dried. In the radiotagging experiments, the reactor was first evacuated to remove unreacted monomer. 14C0 equivalent to 20% (mol/mol) of Ti present was then added through the suba seal by syringe and allowed to react for 15 minutes. The polymer was isolated, purified and dried. In dual-labeling experiments, the radiotagging procedure was first carried out. 15 minutes after the addition of 14C0, the unreacted CO was evacuated, and tritiated methanol introduced, and was allowed to react for a further 20 minutes before the polymer was isolated. Radioassay: Polypropylene and poly-4-methylpent-1-ene samples (100mg) were assayed after dissolution in 10 cm3 of scintillant solution (4g/dm3 PPO, 0.4 g/dm3 POPOP in toluene) at oil bath temperatures. Polyethylene samples were suspended in a NE221 Gel Scintillant (Nuclear Enterprises, Scotland). Activities were meaeured with a LKB Wallace Rackbeta Liquid Scintillation Counter (Model 1217) with counting efficiencies determined by an External Standard Ratio (ESR) method after calibration with a series of quenched standards. The accuracy of the ESR calibration was cross-checked

against

certified 3H and 14C-n-Hexadecane internal Standards and the values from the two methods agreed within a 2% range for assay of individual isotopes.

94 S.N.Gan.P.S.T. Loi, S.C. NG and D.K.Burfield

In house synthesized 'H-polypropylene

( 1 , 5 5 2 0 . 0 3 X l o 6 dpm/g)

and 14C-polypropylene (3.32k0.02 X l o 5 dpm/g) were used as secondary standards for the calibrations of the gel scintillant system. For dual-label assay, samples were counted in two channels with window settings of 8

- 90 and 80 - 165 respectively, together with 3H

and 14C standards. Results were calculated using an Interpolation Spline method. Cross-checks with mixtures of previously assayed singly-labeled polymer samples showed generally close agreement for tritium ( 2%) but poorer agreement for I4C estimation (2 4 % ) .

RESULTS AND DISCUSSION General Observations Although radiotracer methods are very sensitive for the assay of low concentrations of active species, the experiments are essentially blind. Certain assumptions need to be made about the nature of the incorporation before any estimation of active centers can be calculated. Model Studies Evidence of CO insertion (Equation 2) have been provided by a number of earlier studies of low molecular weight polymeric systems. Doi et a1 7, carried out an elegant study on propylene polymerization by the V ( A C ~ C ) ~ / A ~ E catalyst ~ ~ C ~ system at -78°C so that at low temperatures, chain transfer reactions are negligible. IR studies showed that the addition of CO led to quantitative incorporation as a polymer end-group. This clearly demonstrates the feasibility of assay of transition metal-polymer bonds by interaction with CO. Moving from a low temperature homogeneous system to a conventional heterogeneous titanium catalyst, Mejelik et al*)

were able to identify polymeric carbonyl groups after tagging with CO. Another significant model study is that carried out by Kakugo et a19). They used GC-MS to identify low molecular weight products formed during the interaction of CO with a TiC13/A1Et2C1 catalyst in the presence and absence of propylene in n-heptane at 60'C.

In the

presence of monomer, after quenching the reaction mixture with water, a series of ketonic products with the general formula: C3H7-CO-

8. Active Center Determination in Ziegler-Natta Polymerization

95

-(C3H6)n-Et where, O
-

Et

t

nC3H6

____)

Ti - ( c 3 H ~ )- ~Et

(4)

These results provide strong evidence of the CO insertion at Ti-C bond. Furthermore, only one propylene unit appears to insert after the blocking of the site by CO and no multiple insertions of CO were observed. Interestingly, kinetic studies showed that the above reactions were continuing at a significant, although declining rate, even after 5 hours. This is most easily understood by a chain transfer reaction

leading to regeneration of the initial Ti-C bond, e.g.: Ti-CgH6-CO-(C3H6)nEt

t

AIEtZCl

t A1Et(C3H6-CO-(C3H6),Et)Cl (7) +Ti-Et In the absence of monomer the sole reaction product was propanal.

Most recently Shiono e t )'1I.

have used a stopped-flow

polymerization method for the study of the early stages of ethylene polymerization using a supported titanium catalyst. At very short reaction time, transfer reactions are negligible and the polymer chains attached to the titanium growth centers are of low molecular weights. Consequently, their end groups may be detected by spectroscopic analysis. Thus treatment of the reaction system with I3CO allows the interaction of the tag with the titanium-polymer bond to be monitored. Both IR and NMR studies showed that CO is incorporated into the macromolecule. Detailed NMR analysis revealed the tag to be incorporated as a ketonic group and appears to indicate that reactions analogous to (6) are occurring, i.e., after tagging, in the presence of excess ethylene, a further unit of monomer is inserted and that this provides a polymeric ethyl ketone on quenching. In addition, there was a build up of low molecular weight carbonyl species soluble in xylene/ethanol mixtures when the polymerization was treated with small amounts of CO for long reaction

96

S.N.Gan, P.S.T. Loi, S.C. NG and D.R.Burfield

periods (20h). These most likely result from chain transfer with aluminium alkyl (cf. reaction 7 ) and a repeat o f the tagging cycle, The overall evidence of these model studies suggests two things. Firstly, carbon monoxide specifically inserts into transition metalpolymer bonds and thus potentially the method can be used for counting numbers of growth sites. Secondly, low molecular weight tagged products may also be formed, through transfer reactions. The UDtake of Carbon Monoxide The uptake of labeled CO can be readily monitored as a function o f time. The results from our work show that there is an initial rapid uptake of CO within the first ten minutes followed by a further continuing slower incorporation (Figure 1).

3.0

2.0

sow

Contact tim /(ntn.)

1 .o

/(x

105)

I

2.30

I2

2.85

22

3.14

32

1.22

0 0

20

10

'4co

contact

>O

40

tI?.?/(rnl")

F i g u r e 1 . T h e ch an g e i n p o l y m e r r a d i o a c t i v i t y w i t h l 4 C O

contact t i m e . SDPM per m i n u t e .

specific disintegration

These findings are essentially the same as numerous earlier studies8s11-16) and must now be regarded as characteristic of the interaction of CO with these systems. The scheme that most adequately fits the published data is similar to the earlier conclusions of Bukatov et a l l ' , namely that

8. Active Center Determination in Ziegler-Natta Polymerization 97

the initia

fast reaction corresponds to site labeling and that the

subsequent slow incorporation is due to side reactions such as chain transfer w th co-catalyst (equation 7 ) followed by reinitiation of chain growth and subsequent retagging. This is consistent with model studies’) , would explain the enhancement in the presence of monomer15) and with increased metal alkyl con~entration’~), and would be expected to give rise to the observed low molecular weight impurities11

.

Quenching Experiments In the quenching experiments, excess CH30T was added after the initial polymerization. All the metal-polymer bonds become tagged with tritium. Ti-PI A1-P’ t CH30T 4 Ti-OCH3, Al-OCH3 + T-P t T-P’ (8) Dual-labeling ExDeriments In the dual-labeling experiments, during the initial polymerization, exchange reactions with AIR3 would have generated some A1-P. The monomer gas was cut off (in the case of propylene and ethylene) and the radiotagging procedure was carried out. 15 minutes after the addition of 14C0, all the active sites would have presumably been tagged to form Ti-l4CO--P. Unpon subsequent quenching with CH30T, the resulting polymer sample would contain radioactivities due to both the incorporated 14C and T iaotopes. Comparison of Single Labeling and Dual Labeling ExDeriments The results obtained from single labeling and dual labeling experiments between polymerizations carried out under similar conditions are compared (Table 1). As expected, the amounts of incorporated 14C0 tags were similar in both the dual-labeling and single labeling experiments since insertion could occur only at the Ti-P bond. On the other hand the amounts of incorporated tritium in dual labeling experiments were much lower than those of single labeling experiments. These are most readily explained as follows.

98

S.N.Gan. P.S.T. Loi. S.C. NG and D.R. Burfield

In the single labeling experiments, all the metal-polymer bonds, MPB, become tagged with tritium (reaction 8). While in the case of dual labeling, only the base metal polymer bond would become tagged with tritium as represented in the following scheme. A1-P

t

Ti-14CO-P

CHQOT

+ Al-OCH3 + Ti-H t

t CH30T

*

Table 1 : Comparison of C

(9)

T-P

t

(active center concentration) and

** C (non-active metal-polymer

bond concentration) ~~~~

No.

Labeling Monomer mode type

(10)

CH30-14CO-P

T

= c**

14c

= C

*

~

[ MPB I /mmol/mol Ti

The sums of the 14C0 tags and the incorporated tritium from the dual labeling experiments agreed quite well with the total metal polymer bonds determined by single labeling quenching experiments. Use of Dual Labeling Techniaue to Demonstrate Transfer Reactions The transfer reactions, such as ( 3 ) and (7), can be demonstrated by dual labeling technique (Table 2) in the same runs, instead of

duplicate runs with single labeling technique. While Ti-P concentration remained low, the total A1-P bonds increased with reaction time as the polymer chains were constantly being transferred from the titanium sites to the aluminum sites.

8. Active Center Determination in Ziegler-Natta Polymerization 99

Polymerization t ime/min.

14C0 tag = [Ti-PI

T incorporated [Al-P]

[ MPB I

nmol/mol Ti

_-______________-____------------------------------------------60 120

0.33 0.41

180

0.44

0.94 2.08 2.20

1.27 2.49 2.64

[TiC13.AA]= 25 m M ; [A1Et2C1]= 95 m M ; monomer=propylene; 30'C CONCLUSION It is concluded that quantitative 14C0 tagging of active site requires short contact times and high 14C0 concentrations to minimize interfering side reactions. Under optimum conditions, I4CO and CH30T dual labeling approach can lead to a mixture of single radioactive isotope labeled polymer chains, with 14C tagging the active metalpolymer bonds, while the incorporated tritium corresponds to the amount of base metal-polymer bonds. REFERENCES 1. D.H.Ballard, "Coordination Polymerization", J.C.W. Chien, Ed., Academic Press, London, 1975, p.223. 2. J.Mejzlik, L.Lesna, and J.Kratochvila, Adv.Polym. Sci. 81, 84 (1987) A.D.Caunt, S.Davies, P.J.T.Tait, in "Transition Metal Catalyzed Polymerizations, Ziegler-Natta and Metathesis Polymerizations", R.P.Quirk,Ed.,Cambridge University Press,Canbridge, 1988, p.105 4. D.R.Burfield, P.J.T.Tait, Polymer l3, 315 (1972) 5. C.F.Feldman, E.Perry, J.Polywer Sci., 46, 217 (1960) 6. J.Mejzlik, M.Lesna, Makromol. Chem. 178, 261 (1977) 7. Y.Doi, M.Murata, K.Soga, Makromol.Chem.Rap.Comm.5,811(1984) 8. J.Mejzlik, M.Lesna, J.MaJor, Makromol. Chem. 184, 1975 (1983) 9. M.Kakugo, H.Sadatoshi, K.Wakamatsu, H.Yoshioka, Polymer Preprints (Japan) 28 650 (1979) 3.

100 S.N.Gan, P.S.T. Loi. S.C. NG and D.R. Burfield

10.

T.Shiono, M.Ohgizawa, K.Soga, Makromol. Chem.

11.

G.D.Bukatov,

V.A.Zakharov,

194, 2075

Yu.I.Ermakov, Makromol. Chem.

(1993)

179,

2097 (1978)

12.

J.C.W.Chien

and C.I.Kuo, J. Polym. Sci., Polym. Chem, Ed.

a,

731 (1985) 13.

V.Warzelhan, T.T.Burger, D.T.Stain,

Makromol. Chem.

183, 489

( 1982)

14.

P.J.T.Tait,

in "Transition-metal Catalyzed Polymerization-

Alkenes and Dienes", Part 1, R.P.Quirk,

Ed., Harwood Academic,

New York,1983, p. 115ff. 15. 16.

A.D.Caunt, Brit. Polym. J., 22 (1981) P.J.T.Tait,

B.L.Booth,

Cornmun. 9 393 (1988)

M.O.Jejelowo,

Makromol. Chern. Rapid.

101

9. Recent Tendency of Research Targets for Industrial Polypropylene Catalysts

M. TERANO

School of Materials Science, Japan Advanced Institute of Science and Technology, Asahidai 15, Tatsunokuchi, Ishikawa 923- 12, Japan

ABSTRACT Recent tendency of research targets for industrial propylene polymerization catalysts was investigated with academical papers and industrial patents. Research for Ti-based Ziegler-Natta catalysts is still very active in the industrial and the academical fields. However, tendency of the research targets was found to have changed gradually from 1990 to 1993, that is, from basic catalyst features to advanced polymer properties.

INTRODUCTION After a discovery of metallocene catalyst1), most of the research efforts seem to shift to this type of catalysts. In the field of polyethylene (PE), metallocene catalyst is already commercially used to produce a new special grade of polymer, but for polypropylene (PP) the catalyst is applied only for pilot scale production of syndiotactic PP. For isotactic PP, it is not too much to say that the present industrial production completely depends on the Ti-based heterogeneous catalysts. Many patents of MgC12-supported catalysts have been proposed continuously from the reading PP manufacturers in the Even in a basic field,

102 M. Terano

academical papers on Ziegler-Natta type catalysts are still actively published by many scientists.5-7) From the situation mentioned above, Ti-based Ziegler-Natta catalysts are regarded as the industrial PP catalyst in this paper.

WORLD POLYOLEFIN CAPACITY World PE and PP forecast capacities reported by Payn8) are shown in Table 1.

Table 1. Total 1995 Forecast Capacity * HDPE LLD/HD LLDPE LDPE North America Western Europe Japan Pacific Rim China Middle East Eastern Europe Latin America Other Regions Grand Total

* 10

3,588 4,207 1,118 2,037 285 262 840 658 408

6,373 1,433 533 2,012 281 260 190 0 365

1,111 510 210 846 430 50 370 160

3,914 5,467 1,369 1,899 717 779 1,738 2,018 53

13,403

11,447

3,687

17,954

0

PP

TOTAL

5,651 6,989 2,902 3,936 5 10 1,400 971 1,727 593

20,637 18,606 6,132 10,730 1,793 3,131 3,789 4,773 1,579

24,679

71,170

mtlyr

He also mentioned in the paper about the current world wide polyolefin capacity commitments by metallocene catalysts from various companies, which were 650,000 mt/y for PE and 295,000-320,000 mt/y for PP. They are only a few percent of existing capacity. World polyolefin industries still have to depend on the Ziegler-Natta type heterogeneous catalysts almost completely. In the near future, the situation will not be so largely different. From such a view point, it is discussed in the following that what type of catalyst property is now trying to develop or improve for industrial PP production.

9. Recent Tendency of Polyolefin Research Targets

103

PAPERS AND PATENTSUSED IN THIS STUDY To know about the recent tendency of research targets for industrial PP catalysts, academical papers and industrial patents were studied independently. Papers published from 1990 to 1993 concerning industrial PP catalysts were selected from 10 journals listed in Table 2.

Table2. List of Journal

Die Makromoleculare Chemie Die Makromoleculare Chemie ,Macromolecular Symposia Die Makromoleculare Chemie ,Rapid Communications European Polymer Journal Journal of the American Chemical Society Journal of Applied Polymer Science Journal of Polymer Science Macromolecules Polymer Polymer Bulletin

Patents were limited to the application for Japan and to the term from January 1, 1990 to December 31, 1993 to be investigated by using a computer patent search system with an international patent class; CO8F4/00 (polymerization catalyst). The amount obtained was found to be 1863. Then, the patents directly related to the subject of this paper were selected manually, then used in the following analysis. The quantity was 455.

104 M. Terano

TENDENCY OF RESEARCH TARGETS FOR INDUSTRIAL CATALYSTS Table 3 shows the amounts of paper and patent for each year.

Table3. Paper and patent Tic13 cat. paper patent

Year ~

Supported cat. Paper patent

~~

1990

1

18

9

89

1991

0

21

15

105

1992

3

11

17

121

1993

4

4

20

86

Numbers of supported catalysts are much larger than those of Tic13 catalysts, which reflects the recent industrial importance of the supported catalyst. Amount of patents of Tic13 catalysts becomes quite small in 1993. Although it is said that the Tic13 catalysts have lost the importance in the polyolefin fields, the research of the catalysts is not stopped but continued both in an academical and an industrial fields. Nevertheless, it is no doubt that the MgC12-supported catalysts are most important as an industrial catalyst. Therefore, industrial patents of the catalysts were analyzed to find the tendency of recent research targets. As shown in Table 4, the patents were divided into 6 targets and others. Zucchini9) and Karol'O) have mentioned about the main properties which are required for an industrial polyolefin catalyst. The targets in Table 4 are similar to the properties written in their papers.

9. Recent Tendency o f Polyolefin Research Targets

Table 4. Number of Patent for Each Target Target 1990 1991

1992

1993

Activity

12

15

25

9

Stereospecificity

16

27

26

11

Morphology

18

16

16

16

Polymer Property

9

10

11

10

Copolymerization

8

9

17

13

121

86

~

Total

89

105

105

~~

Activity and stereospecificity are the most basic requirements but the amounts of patents are decreased in '93. Patents concerning morphology and polymer property are proposed on the constant level from '90 to '93. Numbers of copolymerization and molecular weight (distribution) are increasing. Thus, the tendency of research targets changed gradually from '90 to '93 reflecting the difference of industrial requirements for the catalyst. For instance, recently wider molecular weight distribution of the PP is regarded as one of the most important targets for MgC12-supported catalysts because metallocene catalysts can not produce the polymer with that feature. Following 2 Japanese laid open They employed different patents are the typical examples for the target. approaches but obtained similar wonderful results. Ueki et. a1.I l ) claimed to use special type of Si compounds as an external donor. While, Asanuma et.

106 M. Terano

external donor

Si(0Me)2

$O\Si(OMe)2

d

d

G

O \ t-Bu /Si(oMe)2

0-0,

i-Pr F

o M e ) l

S. Ueki, N. Aoki, K. Imanishi, Japanese Laid Open Patent, 5-331234

preparation method MgC12 : 300g, C S H ~ ( C O Z C ~:H 7.5~ )ml ~ Tic14 :60 ml, C~tTiClz: 7g

-

)

ground,40h

aA

( A : log, toluene : 60 ml ) 114"c, 30min.

washing with heptane solid catalyst component

T. Asanuma, K. Morita, S. Kimura, Japanese Laid Open Patent, 4-222804

9. Recent Tendency of Polyolefin Research Targets

a1.12) used a solid catalyst component prepared with 2 Ti compounds. values of Mwfin obtained were 7.5 and 7.9, respectively.

107

The

CONCLUSION The research targets for industrial PP catalysts have been changing from basic catalyst features such as activity or stereospecificity to advanced polymer properties like wider molecular weight distribution, which reflects the difference of industrial requirements for the catalysts. Academical papers as well as industrial patents of MgC12-supported catalysts are still actively proposed and published by many scientists to satisfy the requirements. Therefore, this type of catalysts may have the industrial life for another several decades by improving their property and polymerization technology.

ACKNOWLEDGEMENT The author is grateful to Mr. K. Ishii, Mr. K. Tashino, Mr. K. Ohnishi and Dr. H. Mori for their contribution to this work.

REFERENCES 1. H. Sim, W. Kaminsky, H. J. Vollmer, and R. Woldt, Angew. Chem., Int. Ed. Engl., l9, 390 (1980). 2. Y. Uehara, and F. Shimizu, Japanese Laid Open Patent, 5-97922 (1993). 3. T. Fujita, Japanese Laid Open Patent, 5-93013 (1993).

4. T. Matsumoto, K. Shinozaki, and M. Kioka, Japanese Laid Open Patent, 5- 170843 (1993). 5. M. Terano, M. Saito, and T. Kataoka, Makromol. Chem., Rapid Commun., 13, 103 (1992). 6. Y. V. Kissin, T. E. Nowlin, and R. I. Mink, Macromolecules, 26, 2151 (1 993). 7. I. Kim, H. K. Choi, T. K. Han, and S . I. Woo, J. Polym. Sci., Part A: Polym. Chem., 30, 2263 (1 992).

108

M. Terano

8. C. F. Payn, Proceeding of Metcon'93, 1993, p.5 I . 9. U. Zucchini, Makromol. Chern., Macromol. Symp., 66,25 (1993). 10. A. Guyot, L. Bohrn, T. Sasaki, U. Zucchini, F. Karol, and I. Hattori, Makromol. Chem., Macromol. Symp., 66, 31 1 (1993). 11. S. Ueki, N. Aoki, and K. Imanishi, Japanese Laid Open Patent, 5-331234 (1993). 12. T. Asanuma, K. Morita, and S. Kimura, Japanese Laid Open Patent, 4-222804 (1992).

I09

10. The Control of Molecular Weight Distributions in Ziegler-Natta Catalysis

R. SPITZ , M. PATIN, P. ROBERT, P. MASSON and J. D U P W CNRS - Laboratoire de Chimie et Procedes de Polymerisation LCPP BP 69390 Vernaison France ABSTRACT According to the fact that very narrow molecular weight distributions (MWD) are easily produced with metallocene catalysts, the actual problems related to MWD are all concerned with wide distributions. Metallocene mixtures are able to produce bimodal distribution, requiring extreme differences in sensitivity to transfer reactions, but the expected properties will not be the same than those of wide peak distributions. The classical heterogenous systems produce sometimes broad to very broad distributions and this result is associated to a cluster structure of the transition metal. This cluster structure often forbids to mix 2 species to get the sum of 2 polymers. Examples combining titanium and vanadium chloride are presented. The effect of heat and monomer diffusion on MWD is generally not important. INTRODUCTION Classical heterogeneous Ziegler-Natta catalysts never produce polyolehs with narrow molecular weight distributions (MWD), that means close to the value expected for a polymerization with associated transfer reactions (Mw/Mn)) 2). On the contrary homogeneous catalysts are well known to often produce polymers with the expected narrow distribution. This property is one of the most important as it defines the ability of the polymer to be processed and also some of the final properties of the polymer. The presence of a medium to broad distribution has favored the development of this family of polymers due to an easier processing but developments in 2 main directions are necessary: very narrow and very broad MWD. From a hdamental point of view it would be interesting to understand what are the main phenomena associated to the control of the molecular weight distribution. The discussion of these phenomena was clearly introduced 10 years ago in a review made by Zucchini and Cecchin I ) . 2 main theories have been discussed and remain for a part open: the diffusion theory and the chemical theory. According to the diffusion theory, the barrier to the transfer of monomer or heat results in distributions of polymerization conditions in a growing polymer particle and thus in distributions of chain lengths. The MWD is then the envelop of these distributions. The more important the diffusion barrier, the broader the expected MWD will be. According to the chemical theory the heterogeneous catalysts have 2 or more different active species producing in the same polymerization conditions 2 or more different mean molecular polymers. The MWD is then associated to the number of families and to the differences between these families.

110 R. Spitz. M. Patin. P. Robert. P. Masaon and J. Dupuy

In fact none of these explanations are really satisfjhg : the MWD generally resist to any attempt to make a direct change in the conditions associated to one or to the other explanation. Considering catalysts based on the same particles size, porosities and particle size distributions one would expect that the MWD will be very different ifthese catalysts have very high or very low activities (change of the activity may occur by slight modifications of the catalyst composition, for instance a moderatelyhigh thermal treatment) but this is not the case: MWD are generally almost independent on the level of activity even on ranges covering 3 or 4 orders of magnitude, as well in gas phase as in suspension polymerization. For this reason, the development of a particular family of supported heterogeneous catalysts for propene polymerization (using diakylphtalates and silanes as Lewis bases) has restricted to a unique particular value the MWD of theses polymers. Broader distributions are only observed with Solvay type non supported systems 2 , which, for this reason, remain present on the market. Very wide MWD are obtained by 2 or more stages of polymerization. As the MWD is controlled by the presence of different families of active centers according to the chemical explanation it should be expected that the addition of different chemical species would led to a broadening of the distribution. But attempts to load different families of active centers on the same camer are not often successll and Ziegler catalysts used to produce polymers processed by extrusion blowing are generally not able to produce these polymers in one polymerization step. Contemporary problems related to the control of the MWD can be summarized as follows: i) evaluation and interpretation of the contribution of difksion effects highlighted by the modem theories, ii) possibility of broad controlled distributions with metallocene catalysts, iii) interpretation of the broad MWD observed with heterogeneous systems based on metal chlorides. EXPERIMENTAL Magnesium-titanium catalysts: i) milling: magnesium chloride is milled (6h) and contacted with titanium tetrachloride in large excess at 80 "C for 1 hour, washed with heptane and dried under vacuum. ii) magnesium chloride (30 g) is milled for 6 hours, then comilled with 1.5 ml of titanium tetrachloride. Magnesium-vanadium chloride catalvsts: magnesium chloride is comilled with S i c 4 ( 5 % by weight) and then contacted at room temperature with V C 4 in solution in heptane. Quantitative adsorption of the vanadium compound occurs within seconds. The catalyst is dried under vacuum and recovered as a powder containing all the vanadium and less than 0.5 % Si. Polymerization: The standard conditions of slurry polymerization are: total pressure 8 bars, variable hydrogen pressure, akylaluminium: triisobutylaluminium (TiBA) or trihexylaluminium : 3 mM, temperature: 80 "C. Duration 1 hour RESULTS AND DISCUSSION Diffusion control of the polvmerization. The idea that olefin polymerization could be controlled by d f i s i o n is an old idea which became more popular with the development of high mileage catalysts. The decreasing shape of many kinetic curves seemed to be related to the increase of the size of the polymer particle (the diffusion length) and suggested that the larger the polymer grows,

10. Control of M M D in Ziegler-Natta Catalysis

III

the thicker the difision barrier is. This idea was demonstrated to be false both on theoretical bases 3 .4 ) and on experimental bases ’). The time decreasing kinetics are related to chemical deactivation of the catalysts and a correct diffusion theory leads to a a s i o n bamer decreasing with polymerization time proportionally to the particle radius (r): the transfer properties increase proportionally to the external surface area (r2) and decrease proportionally to the particle size (r). A complete description of these phenomena are to be found in the papers of Ray and ~ o l l . ~ ’As ~ , a~ consequence ). of it, 2 main behaviors are expected, corresponding to 2 different kinetic shapes in the 2 main polymerization processes (gas phase and slurry). In gas phase, the polymerization rate reaches a maximum before 1 min. polymerization time and then decreases continuously. The diffusion barrier decreases continuously and becomes neghgible after a few minutes. The production of polymer during the time Corresponding to a strong diffusion barrier is small. Ifwe perform an integration of the polymer production during 1 hour polymerization the contribution of the difision barrier to the MWD is negligible. Fig. 1 presents the typical computed behavior of 1 catalyst particle (30 p diameter, 4000 g polymer per g catalyst in 1 hour) during polymerization using the same set of parameter than Ray and CON.@ MWD is calculated using a variation of the MW with temperature which was measured experimentally in gas phase polymerization. According to the theory the logarithm of the temperature vanes almost linearly with the logarithm of the particle size, but the variation of the molecular weight is small and the variation of MWD is negligible. The same holds in a larger range of sizes and activities, limited to the conditions avoiding the melting of the polymer particle, 2

1

0’

P

-B

-I

0

-1 1.25

1.75

2.25

2.75

log(d1

figure 1 variation of Mn and M w M vs. particle diameter d during a gas phase polymerization (activity: 4000 g polymer per g catalyst per hour). deltaT is the Merence in temperature between the particle surface and the polymerization medium in K.

I I 2 R. SpitL. M. Patin, P. Robert, P. Masson and J. Dupuy

7

6

LD

-

~~

70000 g l g l h

5 4

..

3

--

2

-.

20000 g l g l h

.-

0. 0

6000glglh

200

400

600

800

1000

1200

1400

1600

diameter (crrn)

Figure 2. Melt-index (190 "C, 5 Kg) as a function of polymer particle size for different activities in slurry polymerization. The situation in suspension polymerization is rather different. The polymerization rate increases generally during 10 min. or more . The diffusion bamer is enhanced during at least a part of this time and polymer production out of steady-state is not negligible. One could thus expect a sensitivity of polymer properties to the polymerization rate and kinetic shape and also, of course, to the particle size and shape. This can be checked by the study of the molecular weight or viscosity as a function of particle size, simply by grading the polymer particles after polymerization. The trend which are expected are: molecular weight strongly decreasing with increasing particle sue; molecular weight strongly decreasing with increasing catalytic activity. Experimental studies are not really convincing. We have used a family of catalysts with the same particle properties but with activities ranging from the activity computed by Ray and coll. to 20 times larger values that means approaching lo5 g polymedg catalyst / h. The melt-index ( E viscosity-') of the polymer is presented as a function of the particle size on the fig. 2. The melt-index increases with increasing particle size but not with the catalyst activity as expected with a diffusion control. For each of these catalysts the differences in molecular weights remain small and the MWD do not change. The differences between the catalysts may be associated to different diffusion effects during the preparation of the catalysts themselves. In order to interpret this results we have to turn back to theory. We have written a computer model of growth of a polymer particle which is rather simple but leads to the same quantitative values than the model of the literature when the same parameter are used '). On the basis of such a model, activities in the range of 50000 g per g per h are almost impossible to get as the polymerization is them restricted to the shell of the polymer particle, the monomer concentration in the core being almost equal to zero. This supposes that the activity in the shell is almost infinite. This discrepancy between theory and experiment is due to a particular choice of the diffUsion parameter set. The values used in the literature correspond to a reaction which is l l l y controlled by diffusion I. The kinetic shape showing a polymerization rate increasing with reaction time corresponds to the decrease of the diffusion bamer with the particle growth. Assuming that the kinetics are controlled by chemistry, we have to use monomer diffusion parameters in medium and in the particle pores which correspond to the upper range of the possible values. The model is then in better agreement with experiment and diffusion plays a minor role except perhaps for the most active catalysts which

10. Control of M M D in

Ziegler-Natta Catalysis

113

are above the values currently used in industry. In other words: the fact that very high activities are observed implies that normal activities cannot be controlled by diffusion. Fig. 3 shows a simulation of an ethylene polymerization at a rate of 25 000 glgm . The maximum of concentration gradient becomes important for large catalyst particles but the effect on molecular weight is limited and the variation in MWD (not presented) is less than 1%. 2

450 400 350

1.75

300 250

1.5

200

8

150 100 50

i

-*-

t

Mn

0 1 0

5

10

15

20

25

30

1.25

-3 0

C1 35

particla radius (pm)

figure 3. Minimum value of the monomer concentration on the surface of the growing polymer particle and in the centre and minimum value of the Mn during polymerization as a h c t i o n of the catalyst particle size. Catalyst activity 25000 g per g per hour in ethylene slurry polymerization. Ethylene pressure 6 bars, ethylene concentration 0.43 M Another argument can be found in the literature concerning metallocene catalysts. Kaminsky et aL9) present a detailed study of polymerizations using a large M y of different metallocene catalysts in ethylene and propene polymerization. In most cases the polymerization medium is heterogeneous, the polymers being insoluble in the diluent. With activities ranging fiom 1 to 100 in relative scales, the MWD remain close to 2. M s i o n effects in a very heterogeneous medium are thus not able to contribute to large changes of the distribution. This is confirmed by the fact that supported metallocene catalysts, which are heterogeneous, are able to produce narrow MWD polymers lo). Chemical interpretation of the MWD The chemical interpretation of the MWD supposes that the distribution is associated to a distribution of chemical species. Each of these species produces a statistical distribution of chain lengths which correspond to the Schulz-Flory most probable distribution . It is often assumed that the presence of 2 or 3 families of active centers is able to explain all the distributions of MW observed experimentally. This is not the case. Any particular distribution curve can be decomposed in a little number of components but this does not mean at all that the number is really small. According to the fact that the MWD are rather stable versus changes in the polymerization conditions (activity, temperature, transfer agents) and that the distributions are seldom bi or polymodal it seems reasonable to assurne that the distribution of MWD is continuous and thus so is also the distribution of active centers. What could explain a continuous distribution of active centers with a continuous distribution of properties? The active centers are both not very different but always different. This could be due to small changes in the environment on the surface of the catalyst. But, as the catalysts are generally supported on small crystallized particles, this does not explain the fact that a great number of

114

R . Spitr, M . I’atin, P. Robert. P. Masson and J . Dupuy

different species are really present, and that the same order of magnitude can be observed with supported as well as not supported species. We suggest another explanation: the active species belong to clusters of the salt of the transition metal: this clusters correspond to the crystallites with the non-supported systems or cover a part of the surface of the supported species. The active centers are on the edges of these clusters, the inner part being inert toward polymerization. This idea explains different features of the systems: a continuous distribution of properties, related to the shapes and sizes of the cluster. AU the active centers have then a different environment. Of course, this does not exclude the idea that the active specie belong to 2 or more families, having for instance a different number of C1 vacancies as it is mentioned in the literature” ). The hypothesis of a relation between the clustered structure of the active center and the distribution affords a new insight on the properties. Lf a new component is added to a clustered structure the behavior of the system will depend on the interaction between the species. The new specie can be of the same nature (only changing the cluster sizes and distribution) or may lead to new chemical species ifthe component are Werent. The only case avoiding any interaction is when the different components are separated onto the catalyst support, giving rise to the situation described by B o b and coll.12’.We will examine the other situations on examples. The simplest case corresponds to one chemical specie associated to one support, for instance a catalyst obtained by milling magnesium chloride and interacting it in different manner with titanium tetrachloride in the absence of any other chemical treatment. The distribution of the titanium clusters will be supposed to depend on the distribution of the chlorine vacancies on the surface of the support. The catalysts obtained by milling of magnesium chloride followed by adsorption of titanium tetrachloride have a MWD which is different of that got by comilling the two component. During the comilling, transient defects on the carrier surface are able to react with titanium, producing a new clustered state. This leads to MWD which are dependent on the way that the 2 components have been contacted more than to the Ti content. The comilled solids, which contain structures which are not present on the catalyst prepared by contacting, have the broadest MWD. Table 1. Ethylene polymerization using two catalysts with the same component (magnesium chloride, titanium tetrachloride ) with the same activity (2000 g polymer per g catalyst per hour) but a different clustering of Ti. Details are given in the experimental part.

comilline.

I milling+ impregnation

surface area 42 I48

Ti (w YO) 1.8

10.5

MFR= Izl/Is 19.7

I 14.3

Catalvsts with 2 components. Case of the metallocenes: The presence of metallocene catalysts producing polymers with very different molecular weight and constant distribution opens the possibility of the production of polymer mixtures with bimodal distributions. The fig. 4 presents the variation of the polydispersity of a mixture of the same mass of 2 polymers, each of them having the same polydispersity index 2, as a h c t i o n of the ratio of the molecular weights of these 2 polymers. The broadening of the distribution is got when the 2 polymers are in equal proportion. It is to be noticed that the difference in molecular weight between

10. Control of M M D in Ziegler-Natta Catalysis

I15

these polymers must be large in order to reproduce the common polydispersities measured with heterogeneous catalysts, and must reach values close to 20 in order to get very broad distributions. 12

figure 4. Polydispersity index (Mw/Mn) as a h c t i o n of the ratio k of the molecular weights of 2 polymers mixed in equal proportion with polydispersity 2. Pmkhrrs = 0.5 (2+k+l/k) Metallocene systems associating Zr and Hf compounds are able to present such differences and bimodal mixtures of this type have been described ). The difference in molten viscosity between the 2 fractions of such polymers will be more larger than 1000 times which may lead to problems during extrusion. Catalvsts with 2 components. Heterogeneous systems: the simplest way to prepare controlled broad MWD should be the association of two different catalytic species with Merent sensitivities to the transfer. It is rather easy to find such species using different transition metals: compared to titanium chloride, vanadium chloride generally produces low molecular weight polymers although zirconium and hafhium lead to ultra high molecular weights. All these compounds are able to polymerize although in the same conditions, but the differences of sensitivity to transfer, the differences in activity and the temperature effect induce some complications in the choice of the polymerization conditions. Catalysts associating such species are documented in the literature, and positive results concerning MWD are reported which the 2 heavier metals associated to titanium 14). Vanadium could be a very interesting choice as supported vanadium catalysts are known to produce polymers with broad distributions and possibly low molecular weights due to an outstanding response to transfer by hydrogen Is ) but its possible uses are restricted due to the low activity and the unwanted properties of V residues. In a preceding paper16’ we have described the preparation of such MgClz / VCl, supported catalysts and their activation. On the basis of this results it seemed promizing to prepare titanium-vanadium supported catalysts in order to have at the same time a high activity and the control of a wide range of MWD. The MWD of the polymers is very broad (up to Mw/Mn =18). Numerous catalysts associating these 2 metals are described in the literature I’ , I 8 ) . Such catalysts are reported for narrow MWD 19) as well as for broad MWD ’ O ) .

R . Spitz, M. Patin. P. Robert. P. Masson and J. Dupuy

I16

Considering the polydispersities measured with vanadium (Pa=18) and titanium (Pb=7) supported separately on the same type of magnesium chloride carrier it is easy to compute the values of the expected polydispersities of mixtures of polymers prepared with these two catalysts. Mixtures are defmed by the ratio k =Mwa/Mwb, and x being the relative fraction of polymer produced by the V catalyst, the polydispersity of the mixture is: P= (I/x).(x+k (I-x)).(x.k.Pb+(I-x).Pa). Broad to very broad distributions are expected in a large range of compositions. Of course the best results are expected with a major contribution of the V catalyst but polydispersities above 15 are possible even when the titanium catalyst produces the 2/3 of the polymer as it can be observed on fig. 5 . The molecular weights of the 2 polymers must be of course v e v different . Such a condition can be hlfilled with a large difference in sensitivitv vs. transfer bv hvdrogen 30

25

--

w

n

15

// /

/

10

0

0.2

0.4

0.6

0.8

1

proportion of polymer 2

figure 5 Polydispersity of a mixture of a polymer with polydispersity index P=7 with a polymer with polydispersity index P= 18 as a function of the proportion by weight and as a fimction of the ratio k= Mwll Mw2 (polymer 1 is for titanium and 2 for vanadium). Catalysts are prepared by successive contacting of magnesium chloride with excess titanium tetrachloride followed by vanadium tetrachloride. In the second step the catalyst looses a part of the Ti fixed in the first step (table 2). Table 2: catalyst modified by vanadium tetrachloride, catalysts slurry polymerization 1 hour 80 "C Catalyst

T

TVI Tv2 TV3

10. Control of M M D i n Ziegler-Natta Catalysis

I17

The activity decreases but the MWD remains unchanged or becomes slightly narrower. The presence of V-type active centers is revealed by a slight increase of the heptane soluble polymer which is typical for the vanadium catalysts. This is observed only for the sample containing more V and associated to a slight broadening of the MWD. All the catalysts are far beyond what is expected . Similar catalysts can be prepared in an opposite manner. Catalysts are first prepared by supporting VCL, on magnesium chloride, then treated with titanium tetrachloride. In that case the MWD, starting from the value corresponding to the vanadium catalysts becomes narrower and the activity increases to rather high levels (table 3). The molecular weights are increased and the heptane soluble fraction which is 2 % with vanadium alone is reduced to 0.5 %.

Table 3: vanadium-titanium catalysts (2 and 1.8 % vanadium respectively). Catalyst

Ti W%

VI"' VlTl v2 V2T I

---1.0

--1.2

productivity g/g/h 1700 4200 300 3000

15

m= I*,/Is

I 0.7 2.5 0.5

23.7 17.3 22.8 14.2

H2 bars 1.3 3.0 1.5 2.5

These behaviors can be interpreted as follows: the MWD associated to the Ti or V catalysts are related to the cluster state of these compounds. If V is added to a titanium catalyst, V binds to titanium (vanadium does not bind easily to magnesium chloride) and a part of the Ti is extracted by the solvent. This results in a slight change in the cluster state of the titanium and in the presence of a small fraction of active centers producing low molecular weight polymer. Starting with the V catalyst, the cluster state of the vanadium is not too modified by the addition of titanium, but titanium is able to add to the surface of magnesium chloride. The presence of a typical broad distribution associated to the V centers is masked by the large increase in activity due t o the titanium, but remains nevertheless present and explains the rather large distributions observed. CONCLUSION The molecular weight distributions of polyolefins are controlled by the chemical composition of the catalysts. The diffusion of monomer and heat transfer generally produce only minor effects, which become noticeable with large particles andor high activities. The chemical effects are related to the presence o f 2 or more types of active centers, each of them producing a narrow molecular weight distribution. The mixtures of 2 types of active center produces narrow (P= 3-4)or bimodal distributions if the ratio of the molecular weights produced respectively becomes large. Broad MWD requires a large number of different active species which can be explained by the presence of clusters, which allow slight differences between all the active centers o f a catalyst. The formation of these clusters make it difficult to associate different species as the catalysts contain then intermediate mixed species which do not produce the expected polymers. This explains that the preparation of

118

R. SpitL. M. Patin, P. Robert, P. Masson and J . Dupuy

catalysts producing broad molecular weight distributions in a large range of polymerization conditions and in different processes remains a difficult problem. Acknowledgments This work was supported by Elf-Atochem. REFERENCES 1 Zucchini and Cecchin , Adv. Polym. Sci , 3 , 1 0 1 , (1993) 2 A. Bernard and P. Fiasse in (( Catalytic O l e h Polymerization )) T. Keii and K Soga ed Kodansha Tokyo p. 405, (1990) 3 S. Floyd, K. Y.Choi, T. W. Taylor, W. H. Ray, J. Appl. Polym. Sci., 2_L ,223 1, (1986) 4 RA. Hutchinson, C. M. Chen, W. H. Ray, J. Appl. Polym. Sci. 44, 1389, (1992) 5 R Spitz, J-L Lacombe, M. Primet and A. Guyot in (( Transition Metal Catalyzed Polymerizations )) R P . Quirk ed., MMI Symposium Series 4, p 389 Harwood (1983) 6 S. Floyd, G.E. Mann,W. H.Ray (( Catalytic Polymerization of O l e h s )) T. Keii, K.Soga ed. Kodansha-Elsevier Amsterdam p. 339 (1986) 7 R Spitz, J. Terle, J. Dupuy in (( Recents progres en g h i e des procedes )) 27, 19, (1993) 8 V.B. Skomorokhov, V.A. Zakharov and V. A. Kirillov Polymer Science 35, 881, (1993) 9 W Kaminsky, R Engehausen, K. Zoumis, W Spaleck and J. Rohrmann, Makromol. Chem. 193, 1643, (1992) 10 S. Collins, W. Mark Kelly and D. A. Collins, Macromolecules 25, 1780, (1992) 1 1 K Soga,T. Shiono, Polym. Bull., 8 ,261, (1982) 12 L. L. Bohm , J. Berthold, R Franke, W. Strobel and U. Wolfmeier in (( Studies in Surface Science and Catalysis )) 25 , T. Keii, KSoga ed. Elsevier Amsterdam p 29, (1986) 13 A. Ahlers and W. Kaminsky, Makromol. Chem. Rapid Commun. 9,457, (1988) 14 (F. Masi, S. Malquari, F. Menconi, C. Ferrero, A. Moali and R InverniZzi in (( Studies In Surface Science and Catalysis 6 (( Catalytic Olefin Polymerization )) T. Keii, KSoga ed. Elsevier Amsterdam, p355 (1989) 15 F. J. Karol, K. J. Kann and B. W. Wagner in (( O l e h Polymerization ) ) , W. Kaminsky and H. Sinn ed. Springer Berlin p. 149 (1988) 16 R Spitz, V. Pasquet, A.Guyot in ((40 years Ziegler catalysis )) Makromol. Chem, Symp. in press. 17 EP 22,658 and EP 57,589 to BP Chemicals (1981) Chem. Abs. 94: 175 856, Chem. Abs. 98:4 896 18 EP 32 734 to Montedison S.p.A.( 1981), Chem. Abs. %:170104 19 DE 3028479 to Asahi Chemical Industry (1981), Chem. Abs. 93:s 056 20 DE 3635028, (1988), Chem. Abs. 1 0 9 : 5 5 449 and DE 3,242,150 to BASF A. G., (1984), Chem. Abs. M : 9 1 656

I I9

11. Synthesis and Application of Terminally Magnesium Bromide-Functionalized Isotactic Poly (Propene)

Takeshi Shiono, Yoshi-hide Akino and Kazuo Soga* Research Laboratory of Resources Utilization, Tokyo Institute of Technology, 4259 Nagatuta, Midori-ku, Yokohama 227, Japan *Japan Advanced Institute of Science and Technology, Hokuriku 15 Asahidai, Tatsynokuchi, Ishikawa 923-12, Japan ABSTRACT Isotactic poly(propene) (PP) having a terminal vinylidene group was prepared with the ethylenebis(4,5,6,7-tetrahydro-lindeny1)zirconium dichloride-methylaluminoxane catalyst system. The polymer produced was treated with borane-dimethylsulfide in toluene, followed by reacting with pentane-lI5-di(magnesium bromide) to obtain magnesium bromide(MgBr1-terminated isotactic PP. The polymer was brought into contact with iodine to give isotactic PP having an iodine group at the chain end in about a Polymerization of methyl methacrylate was then 85 % yield. conducted using the MgBr-terminated PP as an initiator to synthesize isotactic PP-block-poly(methy1 methacrylate) copolymer. INTRODUCTION Terminally functionalized polymers are useful as precursors of block and graft copolymers. The Zn-, Al- and B-terminated isotactic poly(propene)s (PPs) were synthesized by a chain-transfer reaction with metalalkyls’) or by hydrometalation of the terminal vinylidene group formed via a B -hydrogen abstraction.2) These metal-polymer bonds were utilized for the functionalization of chain ends as well as for the synthesis of block-copolymers. In this paper, the magnesium bromide(MgBr)-terminated isotactic PP was synthesized and applied for the synthesis of the isotactic PP-block-poly(methy1 methacrylate) (PMMA) copolymer.

120 T. Shiono. Y . Akino and K . Soga

EXPERIMENTAL Materials. Propene (Mitsubishi Petrochemical Co.) was purified by passing through columns of CaC12, P2O5 and molecular sieve 3 A . rac-Ethylenebis(tetrahydroindeny1)zirconium dichloride (Et[H4IndI2ZrCl2) was prepared according to the literature. 3 , Methylalminoxane (Tosoh Akzo Chemicals Co. ) and borane-dimethylsulfide complex (Aldrich Chemical Co.) were used without further purification. Research grade benzene and toluene commercially obtained were dried over calcium hydride under refluxing for 24 h and distilled on molecular sieve 4A. Methyl methacrylate (MMA, extra pure grade) was dried over calcium hydride and distilled before use. Iodine was purified by sublimation and used as a 0.5 M solution in toluene. Argon (99.9995%) was used without further purification. Pentane-I,S-di(magnesium bromide) in tetrahydrofuran (0.6 MI was synthesized from magnesium turnings (for Grignard reagent, Wako Chemical Co. ) and 1,5dried over molecular dibromopentane (Aldrich Chemicals C o . , sieves 4A). Preparation of isotactic PP. Propene polymerization was conducted with a 200-mL glass reactor equipped with a magnetic stirrer. Toluene (100 mL) and methylalminoxane (1.8 mmol) were added into the reactor under an argon atmosphere, and then propene was introduced at 30 OC until the solvent was saturated with propene. Polymerization was started by adding 4 mL of toluene solution containing 0.01 mmol of Et[H4IndI2ZrCl2 and terminated by the addition of a dilute. solution of hydrochloric acid in ethanol. The precipitated polymer was filtered, washed with plenty of ethanol and dried under vacuum at 60 O C for 8 h. The polymer produced was extracted with boiling acetone to obtain approximately 90 wt% of boiling-acetone insoluble fraction, which was confirmed to have the structure ( I ) from 13C NMR and 'H NMR as reported previously.2bt4)

c c C C I 1 I I c=c-c-c-(c-c)n-c-c-c-c-c(I) The polymer was characterized as follows; melting point = 114 'C, isotactic triad = 0.82, number average molecular weight(Fln) determined by chain end analysis = 5,300, Rn by GPC =

I I . Synthesis of Terminally Functionalized Iso-PP

121

2,200 (polydispersity = 2.0). Such a discrepancy of An may result from the limitation of GPC in a low molecular weight range. The vinylidene content in the polymer estimated by assuming An as 5,300 was 0.19 mmol/g-polymer. Synthesis of MgBr-terminated PP. The terminal C=C bonds of the polymer was hydroborated by borane-dimethylsulfide complex. In a 50-mL Schlenk tube equipped with a magnetic stirrer, ca. 1 g of PP, 10 mL of toluene and 0.086 mmol of borane-dimethylsulfide complex were added under an argon atmosphere, and the mixture was heated at 70 - 80 OC for 2 h. Then 0.173 mmol of pentane-I ,5di(magnesium bromide) in tetrahydrofuran solution was added, and the mixture was continued stirring at 70 - 80 OC for 2 h to obtain the MgBr-terminated PP6). Synthesis of PP-block-PMMA copolymer. Polymerization of MMA with the MgBr-terminated PP was conducted at -78, 0 and 45 OC. In MMA (9.4 mmol) was case of the polymerization at 0 or 45 ' C , added into the reactor containing ca. 1 g of the MgBr-terminated PP in 10 mL of toluene at the polymerization temperature. Whereas, in case of the polymerization at -78 OC, MMA was added at 0 OC and the mixture was quickly cooled down to -78 OC using a dry ice-ethanol bath. Polymerization was quenched by pouring the polymerization mixture into ethanol containing hydrochloric acid. The precipitate was collected and dried under vacuum at 60 OC for 6 h. The polymer obtained was extracted with boiling acetone, and the acetone-insoluble fraction was supplied for analyses. Analytical procedures. H spectra of samples were recorded on a EX-90 or a JEOL FX-100 spectrometer operated at 89.45 or 99.45 MHz in the pulse Fourier Transform (FT) mode. 13C NMR spectra were recorded on a JEOL GX-500 spectrometer operated at 125.65 MHz in the pulse FT mode. In 'H NMR measurements, the pulse angle was 45 and 100 - 500 scans were accumulated in 9 s of pulse reputation. In I3C NMR measurements, broad band decoupling was used to remove I3C-'H couplings. The pulse angle was 45 and 6000 - 9000 scans were accumulated in 9 s of pulse reputation. The spectra were obtained at 60 or 80 OC in CDC13 or C ~ solution D ~ (2 wt % for I H NMR and 1 5 wt % for 1 3 NMR ~ in a 5mm 0.d. tube), using CHC13 (7.24 ppm for 'H NMR and 77.0 ppm for I3c NMR, respectively) or C6H6 (7.15 ppm for I H NMR and 128.0 ppm for 3~ NMR, respectively) as an internal reference O ,

O ,

.

122 T.Shiono. Y . Akino and K . S o p

The gel permeation chromatograms (GPC) of polymers were recorded on Sensyu SSC-7100 equipped with a Shodex GPC UT-806L column at 1 4 5 OC using o-dichlorobenzene as solvent. The molecular weights of polymers were determined by the universal calibration technique. Differential scanning calorimetry (DSC) measurements were made with a Seiko DSC-220. Polymer samples (ca. 3 mg) were encapsulated in aluminum pans. Samples were pretreated at 200 O C for 5 min, chilled with liquid nitrogen, and scanned at 1 0 OC/min. RESULTS AND DISCUSSION Synthesis of MgBr-terminated PP. We have already reported that chlorine-, bromine and iodine terminated isotactic PPs were obtained in high yields(>80%) by or Al-terminated2c) polyhalogenolysis of the Zn-terminated' mer. However, such halogen-terminated isotactic PPs cannot be f,

used as precursors of Grignard reagents due to poor solubility in polar solvents. Therefore, in this paper, the MgBr-terminated isotactic PP was prepared by transmetalation of the B-terminated PP. It is reported that trialkylborane is quantitatively converted to the corresponding Grignard reagent by the reaction with pentane-1,5-di(magnesium bromide) in benzene or toluene according to the following scheme.6) The formation of a relatively stable bicyclic borate might probably cause to shift the equilibrium to the right. R3B

+

2 BrMg(CH2)5MgBr + 3 RMgBr

+

qn [ ( H C) B (CH )

2

3

w 2 5

]+MgBr-

The MgBr-terminated isotactic PP synthesized according to the procedure described in the experimental section was subjected to iodolysis. The iodolysis was conducted by adding the toluene solution of iodine to the MgBr-terminated PP in toluene at 60 O C until the color of iodine did not disappear. Figure 1 shows the 'H spectra of the polymers before and after iodolysis. In the spectrum of the polymer after iodolysis, the resonance of vinylidene proton completely disappeared and a new resonance assignable to methylene protons connected to iodine appeared at around 2.9 ppm. Using the resonances of main chain protons as an internal

I I . Synthesis of Terminally Functionalized Iso-PP

123

standard, the conversion of vinylidene to iodine was estimated to be approximately 85%.

PPM l ' ' ' ' ' ' ' ' ' / ' ' ' ' ' ' ' ' ' 1 ' ' ' ' ' ' ' ' ' 1 ' ' ' ' i ' ' ' ' I ' ' ' ' ' ' ' ' ' ~

5

4

3

2

1

0

Figure 1 90-MHz 'H NMR spectra of isotactic PP. (a):original, (b):after iodolysis Figure 2 shows the I3C NMR spectra of the polymers after hydrolysis and iodolysis, where several weak resonances can be observed in addition to the strong resonances of main chain carbons. All the resonances except for those of the tetramethylene sequence in main chain (marked by * ) are assignable to the carbons of 2-methylpentyl, 2-methylpropyl and 3-iodo-2-methylpropyl end groups, which correspond to the initiation end, and the hydrometalated ends after subjecting to hydrolysis and iodolysis, respectively. The intensities of the 2-methylpropyl end group drastically decreased by iodolysis. Some of the resonances of the 3-iodo-2-methylpropyl group were split into doublet due to the presence of four diastereomers (two pairs of enanthiomers), which are derived from the highly isotactic structure of chain end and non-enanthioselectivity in the hydrometalation process of vinylidene group as reported previously. 3 , For reference, the iodine-terminated isotactic PP was pre-

124 T. Shiono, Y . Akino and K . Soga

k

I,.

Pl

xl n

i2 il-it

PPH 1

"

"

l

50

Figure 2

"

4:

"

1

"

~

40

'

1

'

'

35

"

1

~

30

~

'

"

'

~

~

~

25

1

20

'

'

~

'

15

125-MHz 1 3 C N M R spectra of isotactic PP.

(a):after hydrolysis, (b):after iodolysis

1

~

~

'

I I . Synthesis of Terminally Functionalized Iso-PP

125

pared from the B-terminated polymer by reacting with sodium iodide and chloramine-T7) in the mixture of toluene, methanol and water. The conversion was, however, found to be very low due to poor solubility of the polymer. MMA polymerization by MgBr-terminated PP. Polymerization of MMA was then conducted at -78, 0 and 4 5 O C using approximately 1 g of the MgBr-terminated polymer as an initiator. The polymers produced were extracted with boiling acetone to remove the homo-PMMA. The results are summarized in Table 1.

f"3 -fCH2-fk

c=o I

OCH 2 h

6 h

12 h

24 h

48 h

Figure 3 100-MHz 'H NMR spectra of isotactic PP-block-PMMA obtained at -78 ' C with different polymerization time (2-48 h). The 'H NMR spectra of the acetone-insoluble polymers obtained at -78 O C were illustrated in Figure 3. In addition to the strong resonances of PP, the resonance of methoxy protons can be observed at around 3 . 5 ppm, the intensity of which gradually in-

Table 1

Results of M M A polymerization by MgBr-terminated isotactic PP

Run no.

Polymn. conditions Temp["C]

Time[h]

Yield[ g ] Whole

-

-

-

1

-78

2

1.28

2

-78

6

3

-78

4

-78

5

Insol.a

Properties of acetone-insoluble fraction mol-MMA/mol-Pb

Tg[ "Cl Tm[ "Cl

H [ J/g] Anxl O - 3 c

0

-1 5 . 7

114

54.3

2.2

0.98

0.02

-15.2

113

54.3

2.4

1.14

0.86

0.04

-12.2

113

53.1

2.8

12

1.18

1.04

0.08

-11.0

113

52.3

3.0

24

1.20

1.03

0.09

-10.7

113

50.3

3.0

-78

48

1.43

1.20

0.12

-10.0

110

49.8

2.9

6

0

1

1.27

1.10

0.06

-1 1 . 2

111

50.2

2.6

original PP

7

0

2

1.47

1 .14

0.08

-10.1

113

48.4

2.5

8

0

6

1.22

0.97

0.10

-10.0

114

52.1

2.9

9

0

12

1.24

1.07

0.10

-10.2

113

51 . 3

2.9

10

45

0.25

1.02

0.93

0.03

-1 3 . 2

114

56.6

-

11

45

1

1.03

0.92

0.03

-13.7

111

56.4

-

aweight fraction of boiling acetone-insoluble fraction. bcomonomer ratio determined by ' H NMR spectra. 'determined

by GPC.

I I . Synthesis of Terminally Functionalized Iso-PP

127

creased with an increase in the polymerization time. The methoxy protons could be observed also in the polymers obtained at 0 and 45 'C. The ratios of monomer units (MMAIpropene) in the copolymers were estimated from the relative intensities of the methoxy and hydrocarbon protons (see Table 1). The results are shown in When the MMA polymerization was conducted at - 7 8 or Table 1. 0 'C, the content of MMA in the copolymer increased with prolonging the polymerization time. The number average molecular weight of the resulting polymer also increased with an increase in the polymerization time to reach a constant value, indicating formation of the block copolymer. The thermal properties of the isotactic PP-block-PMMA copolymer were briefly investigated by DSC, the results of which are The copolymer essentially shows the melting shown in Table 1. point(Tm) and glass transition temperature(Tg) of isotactic PP, although Tg slightly increased with an increase in the PMMA content. In conclusion, the MgBr-terminated isotactic PP was synthesized via hydroboration of the vinylidene-terminated polymer followed by transformation of B-C bonds with pentane-I ,5di(magnesium bromide). The MqBr-terminated polymer was found to initiate the polymerization of MMA at low temperatures to produce the isotactic PP-block-PMMA copolymer. ACKNOWLEDGEMENT The authors thank Dr. Toshiya Uozumi (JAIST Hokuriku, Japan) for the GPC measurements. REFERENCES 1. For example, (a)Y.Doi, K.Soga, M.Murata, and Y.Ono, Makromol. Chem., Rapid Commun., 4, 789 (1983);(b)D.R.Burfield, Polymer 25, 1817 (1984);(c)G.Redina, and E.Albizzati, Eur. Pat. Appl., 350059(1989) (d)T.Shiono, K.Yoshida, and K.Soga, Makromol. Chem., Rapid Commun., 1 1 , 169 (1990) ;(e)T.Shiono, H.Kurosawa, and K.Soga, Makromol. Chem., 193, 2751 (1992); (f)H.Kurosawa, Dr. Thesis, Tokyo Institute of Technology (1994) 2. (a)R.Mulhaupt,T.Duschek, and B.Rieger, Makromol. Chem., Macromol. Symp., 48/49, 317 (199l);(b)T.Shiono, and K.Soga,

128 T. Shiono. Y . Akino and K . Soga

3.

4. 5. 6. 7.

Macromolecules, 25, 3356 (1992);(c)T.ShionoI and K.Soqa, Makromol. Chem., Rapid Commun., 13, 371 (1992);(d)T.ShionoI H.Kurosawa, O.Ishida, and K.Soga, Macromolecules, 26, 2085 (1993) F.R.W.P.Wild, M.Wasiucionek, G.Huttner, and H.H.Brintzinger, J. Orqanometal. Chem., 288, 63 (1985);T.Tsutsuir A.Mizuno, and N.Kashiwa, Polymer, 30, 428 (1989) A.Grassi, A.Zambelli, L.Resconi, E.Albizzati, and R.Mozzocchi, Macromolecules, 2 l , 617 (1988) L.M.Brown, R.A,Brown, H.R.Crissman, M.Opperman, and R.M.Adams, J. Orq. Chem., 36, 2388 (1971) K.Kondo, and S.Murahashi, Chem. Lett., 1237 (1979) G.W.Kabalka, and E.E.Gooch, J. Org. Chem., 4 6 , 2582 (1981)

I29

12. Wide Range Control of Microtacticity in Propylene Polymerization with Heterogeneous Catalyst Systems

Masahiro Kakugo Chiba Research Laboratory, Sumitomo Chemical Co. Ltd., 2- I Kitasode, Sodegaura, Chiba, Japan 299-02

ABSTRACT Heterogeneous highly isospecific Ti-Mg catalyst system was shown by a combination of temperature-programmed column fractionation (TPCF) and NMR analysis of the polymer obtained to give polypropylene (PP) which mainly composed of two types of isotactic polymers differing in microtacticity and molecular weight.

This finding agrees with that found previously in 6-TiC13-

AIEt2CI system, showing that the control of the microtacticities and the molecular weights of both isotactic polymers as predominant components will enable us to design the catalyst giving various molecular structures, i.e., microtacticity and molecular weight distribution.

On the basis of this

concept, new catalyst systems have been developed to improve the properties of PP.

INTRODUCTION It is known that heterogeneous Ziegler-Natta catalysts comprise some types of active centers.

The kind and the nature of the active centers will be memorized in the structure of PP produced. The detailed structural analysis of polymer, therefore, gives valuable information about active centers.

Previously, along this idea we studied the detailed structure of PP prepared with 6-

TiC13-AlEtzCI system by a combination of TPCF and NMR analysis.

As a result, the molecular

structure of PP would be explained basically in terms of the presence of three types of active centers.

On the basis of this result, we proposed a model for the active centers 1-3).

GTiC13-

AIEt2CI system, that is, consists of three types of the active centers (Scheme I ) , i.e., ( I ) the highly isospecific active center having four firmly bound C1 ions, an alkyl group, and a CI vacancy, (2) the

130 M . Kakugo

Scheme 1.

Model of the active centers on &Tic13

9

-B

Structure of active center change

Stereospecificity

highly isospecific

low isospecific

' P1

IP,

Stereoregularity of polypropylene

syndiospecific

Stereoblock

SP

nonstereospecific

AP

(SB)

low isospecific center consisting of at least a loosely bound CI ion, and alkyl group, and n CI vacancy, and (3) the nonstereospecific active center consisting of two firmly bound CI ions, a CI

ion, an alkyl group, and two CI vacancies.

In addition, we considered that the low isospecific

active center will become syndiospecific one by configurational change of the CI ion, the alkyl group, and the vacancy during polymerization.

The presence of these active centers results i n the

forination of polymers with different stereoregularities, i.e., highly isotactic (IPl), low isotnctic (IP2), low isotactic-syndiotactic stereoblock (SB), syndiotactic (SP), and atactic (AP).

We

analyzed similarly PPs prepared with Ti-Mg catalyst system and have found that the kind and the nature of the active centers present in Mg-Ti catalyst system are essentially similar to those in the 6 TiC13-AIEt2CI system4). On the basis of this finding , we have designed Ti-Mg catalyst systems with a variety of the stereospecificities and molecular weight distributions.

EXPERIMENTAL Polymerization

Polymerization was carried out in a 3-L autoclave i n liquid propylene or n-

heptane. Solvent extraction

The sample was completely dissolved in boiling xylene and the solution

was cooled gradually to 20°C.

The precipitated polynier was separated by filtration.

polymer soluble in xylene at 20°C was recovered from the filtrate by evaporation.

The

12. Control of Microtacticity in Propylene Polymerization

Fractionation

I3 I

Temperature-programmed column fractionation (TPCF) method1) and cross

fractionation chromatography (CFC) by using a Yuka Denshi CFC-T-150A instrument were applied to the fractionation of polymer obtained. Analysises

I3C N M R spectrum was obtained on a JEOL EX-270 pulsed Fourier transform

NMR spectrometer in o-dichlorobenzene at 135OC.

of polymers in o-dichlorobenzene.

Sample was prepared as 5-50mg/ml solution

The pulse interval was IOs, the acquisition time was 4.2s. the

pulse angle was 4.5".and the number of transients accumulated was 3000-10000.

was determined from the area of the resonance peaks of the methyl region.

Pentad tacticity

The molecular weight

distribution of the samples was determined by a Waters Associates type 150C GPC instrument in 0dichlorobenzene at 145°C.

The melting temperature of the samples was measured on a Perkin

Elmer Type-2 differential scanning calorimeter (DSC).

The sample was premelted in DSC at

22OOC for 5 min. and cooled from 150 to 40°C at a rate of S0C/min.

Thermogram was recorded

by raising from 40 to 170°C at a rate of 5"C/min.

RESULTS AND DISCUSSION 1. Microtacticity distribution of polypropylene p r e p a r e d with Ti-Mg catalyst

system Table I shows the molecular weights and the isotactic pentad fractions of PPs prepared with the Ti-Mg catalyst system and the &Tic13 catalyst system.

Both polymers possessed similar

molecular weight and the polymer obtained with the Ti-Mg catalyst system was 0.98 in mmmm fraction, higher isotactic than that with the &Tic13 catalyst system, 0.96.

In order to know the difference of the microtacticity in detail, we fractionated these polymers by TPCF technique and compared the microtacticity distributions of the polymers.

Figure 1 shows

the cumulative fractionation curves of the polymers in which the elution temperature corresponds to the microtacticity of polymer, that is, the higher isotactic polymer elutes at the higher temperature. Table 1.

Sample list for fractionation

Catalyst system

Mw

mmmm fraction

&Tic13 catalyst

249000

0.96

Ti-Mg Catalyst

22.5000

0.98

132 M. Kakugo

100

80

60 40

20

- SB -

0

20

//

I

I

I

1

I

I

40

60

80

100

120

140

Elution temperature ( " C ) Figure 1.

Cumulative fractionation curves of polypropylenes prepared with &Tic13 catalyst

system and Ti-Mg catalyst system.

The fractionation data indicate that the Ti-Mg catalyst system formed less proportion of AP, SP and

SB compared with the &-Tic13catalyst system.

Figure 2 indicates the differential fractionation

curves of the isotactic parts of the PPs obtained from the data shown in Figure 1 , which clearly shows that both the Ti-Mg catalyst system and the &Tic13 catalyst system give similarly two types of isotactic polymers (IPI and IP2), but the proportion of IPI to IP2 in the Ti-Mg catalyst system is higher than that in &Tic13 catalyst system. higher than that in the &Tic13 catalyst system.

In addition, the rnicrotacticity of IPI is rcmarkably The Ti-Mg catalyst system gave mainly isotactic

polymers, 75% of IP1 and 20'70 of IP2, and a small amount (5%) of low stereoregular polymers,

i t . , AP, SP, and SB. This means that in the Ti-Mg catalyst. system a control of the proportion and microtacticities of IPI and IP2 with keeping the generation of the low stereoregular polymer low will lead to the development of the catalysts in giving a wide variety of PPs.

12. Control o f Microtacticity in Propylene Polymerization

IP,

&TiCI,

1 '

I:

Ti-Mg

0.90

0.95

1.oo

lsotactic pentad fraction Figure 2.

Microtacticity distribution curves of isotactic parts of polypropylenes prepared with

&Tic13 catalyst system and Ti-Mg catalyst system: (-)

observed curves; (- --) calculated curve

(&Tic13 catalyst system: [mmmm]=0.958,Mn=84000 and [mmmm]=o.975, Mn=3 1500, Ti-Mg catalyst system: [rmiriirn]=0.960,Mn=33600 and [mmmm]=0.988, Mn=l68000).

133

134

M. Kakugo

2. Control of the microtacticity of isotactic part The polymerization results with several Ti-Mg catalyst systems, A-D are shown i n Table 2. Although the proportions of atactic polymer are kept at lower level than 2%, the microtacticities of the whole polymers are different from each other, suggesting that these polymers have different microtacticity distributions of the isotactic parts.

Figure 3 shows the relationship between the

generation of atactic polymer and the microtacticity of the whole polymers prepared with catalyst systems A-D (solid line) and conventional catalyst system (dotted line).

In the case of

conventional catalyst system, the change of microtacticity was accompanied usually by a relatively large change of AP generation.

This means that there is a limitation of structural control by using

conventional catalyst system because of the deterioration of the properties of PP caused by AP. On the contrary, in the present Ti-Mg catalyst systems, the microtacticity could be controlled ranging from 0.92 to 0.99 with maintaining the proportions of AP at a low level. The polymers prepared with catalyst systems A-D were fractionated by CFC technique. Figure 4 shows the differential fractionation curves of the isotactic parts of these polymers.

This

figure indicates that the relative portion of the IPI to the IP;! changes slightly among these polymers. However the microtacticities of IP1 and IP2 change significantly, i.e., the whole polymer with higher microtacticity consists of the higher microtacticities of both IP1 and IP2.

Table 2.

fi)

Results of propylene polynierization with various Ti-Mg catalyst systems") Catalyst

Activity

system

(g-PPIniol-Ti h )

A

4000

I .7

3.6

0.92

B

20000

1.1

3.7

0.96

C

45000

0.7

3.8

0.98

D

52000

0.6

4.2

0.99

E

59000

1 .o

6.4

0.98

APD) MwlMn (~1%)

mnitiim

fraction

Polymerization was carried out in a 3-L autoclave in liquid propylene at 80°C in I h.

0)Fraction soluble in xylene at 20°C.

12. Control of Microtacticity in Propylene Polymerization

I 0

'

Conventional \\

catalyst system

\ \ \ \

0 I 0.90;

Figure 3.

'

'

'

' 5 AP (wt%)

'

'

'

10

Relationship between isotactic pentad fraction of whole polymer and proportion of

atactic polymer.

90 100 110 El u l i o n I emper at u r e Figure 4. A-D.

120

130

"C

Differential fractionation curves of polypropylenes prepared with catalyst systems

135

136

M. Kakugo

3. Control of Molecular Weight Distribution

As shown in Table 2 , the polymer prepared with catalyst system E has the same niicrotacticity and almost same proportion of A P as that prepared with catalyst system C, but its molecular weight distribution is remarkably broader.

Table 3.

The average molecular weight of

the IPl and the IP2 in catalyst systems C and E Catalyst system

Mw

IPI

IP2

C

399000 86000

E

415000 71000

101

102

103

104

Chain length Figure 5.

105

1 o6

(A)

Molecular weight distributions of the IP1 and the IP2 of polypropylenes prepared

with catalyst systems C and E.

12. Control of Microtacticity in Propylene Polymerization

137

In order to understand the difference in molecular weight distribution in more detail, the polymers were fractionated by CFC technique.

The molecular weight distributions of IPI and IP2 were

obtained from the summation of molecular weight distributions of the fractions eluted in the temperature regions of IP1 and IP2, respectively.

The calculated molecular weight distributions

are shown in Figure 5 and the average molecular weight of the 1P1 and the IP2 are described in Table 3.

The relative portion of the IP1 to the IP2 is largely different from each other.

In catalyst

system E giving a broader molecular weight distribution the IP2 is higher relatively than that in catalyst system A .

Additionally, a difference between the average molecular weights of the IPI

and the 1P2 is larger than that in catalyst system E.

These differences enable us understand the

reason of the wide molecular weight distribution of PP obtained with catalyst system E.

4. Properties of polypropylenes prepared with catalyst systems A-E

Figure 6 shows the relationship between the microtacticity of the whole polymer and the flexural modulus.

The flexural modulus increases with increasing the microtacticity of the whole polymer.

The molecular weight distribution of polymer also affects the flexural modulus; the broader is

0.90

0.95

lsotactic pentad fraction Figure 6.

Relationship between isotactic pentad fraction and flexural modulus.

1.oo

138

M. Kakugo

E

D

-r B

*/O’ C

A

0.95 lsotactic pentad fraction

0.90

Figure 7.

1

1 .oo

Relationship between isotactic pentad fraction and melting temperature

Figure 7 shows the relationship between the rnicrotacticity and the melting temperature of polymer, indicating that the melting temperature depends primarily on the isotactic pentad fraction, not on niolecular weight distribution. CONCLUSION

We have developed a new catalyst technology using Ti-Mg catalyst system for controlling both rnicrotacticity and molecular weight distribution of isotactic polypropylene.

The properties o f

isotactic polypropylene can be controlled by these catalyst systems in a wide range. ACKNOWLEDGEMENT

The author wishes to express his gratitude to Suinitomo Chemical Co. Ltd. for permission to publish this work.

The author acknowledges helpful discussions with Messrs. H. Sadatoshi, K.

Mizunuina, S . Kishiro, and T. Ebara. It EITE R EN C ES 1. M. Kakugo, T. Miyatake, Y. Naito, and K. Mizunumn, Mncromolecules, 21, 714 (1988)

2 . M. Kakugo, T. Miyatake. Y. Naito, and K. Mizunuma, Mnkrorrzol. Chenr., 190, 505 (1989) 3. M . Kakugo, T. Miyatake, and K. Mizunuma, Macrorirolecufes,24, 1469 (1991) 4. T. Miyatake. K . Mizunuma, in. Kakugo, Stud. in Sui$ Sci. Catal., 56(Cat(71. Olejin Polyrrr.), IS5 ( I 989)

139

13. New Heterogeneous Catalysts for Polyolefins

E. Albizzati, T. Dall'Occo, M. Galimberti, G. Morini Himont "G. Natta" Research Center, P.le G. Donegani, 12 44100 Ferrara (Italy)

ABSTRACT

The key factor that allowed the outstanding development of polyolefins has been the Ziegler-Natta catalysis that played and is still playing an innovative role in this field. More andmore sophisticated new catalysts have made possible advanced polymerizationtechnologies, improvedproductgrades and a broader range of applications, through a better understanding of the relationships among the catalyst architecture, the micro and macrostructure of the polymer and its properties. Aiming to widen the "property envelope" of polypropylene we have to improve the followig aspects: i) polymer architecture control ii) polymer microstructure control iii) modification of the surface properties i iii) modification of the rheological properties. The guidelines that we have followed, by using heterogeneous MgC1, based catalysts, to overcome these problems are reported. INTRODUCTION

The expansion of polyolefins into the worldwide plastic market in the last thirty years has been exceptional. The polyolefins world market share was around 20% of the total thermoplastics market in the 60's while it is reaching almost 60% in the 90's with an average growth rate of 7-8% per year. The key factor that allowed the realization of this outstanding development has been Ziegler-Natta catalysis that had played and is still playing an innovative role in this field.

140 E. A l h i i i a t i . T. Dall'Occo. M. Galirnberii a n d G. Morini

More and more sophisticated new catalysts have made possible advanced polymerization technologies, improved product grades and a broader range of applications. Besides, the research effort was aimed at modifying some particular properties of polyolefins in order to widen their application fields. We believe that the expansion of the Inpropertyenvelopel' of polyolefins has been made possible through the following steps: The continuous development of high yield MgC1, based catalysts. The polymer architecture control, in terms of morphology and porosity. The polymer structure control in order to: i) obtain high crystallinity (isotactic index of PP higher than 99%); ii) regulate MW and MWD; iii) distribute randomly one or more comonomers; iiii) obtain new polymers, by using metallocene based catalytic systems. Modification of the surface properties (e.g.compatibility with other polymers, fillers and fibres) by introducing in the backbone polar groups or unsaturation suitable for crosslinking and other chemical reactions. Modification of the reological properties by introducing long chain branching by postreatment of the polymer or copolymerization with diolefins. In this paper the guidelines that we have followed to achieve these development steps are reported.

1)

HIGH YIELD M N l z BASED CATALYSTS We have recently discovered [l] new MgC1, based catalysts

f o r PP production containing a new family of internal donor belonging to the class of 1,3-diethers. These catalysts are able to give with high yield highly stereospecific PP in the absence of any external donor (fig.1).

13. New Heterogeneous Catalysts for Polyolefins

141

Fig.1 : General formula of 1J Diethers

widely accepted mechanism of stereoregulation by Lewis bases, first proposed by Corradini [2], is based on the competition of the Lewis base with TiC1, for selective coordination to unsaturated magnesium atoms on the different later?l faces of MgC1,. According to this model, dimeric titanium species, responsible for the synthesis of isotactic polymer, should be present on the (100) face, whereas the Lewis base should saturate the vacancies of tetracoordinate Mg atoms present on (110) face, thus avoiding the placement of TiC1, and the consequent formation on this plane of non stereospecific sites. As a consequence of this model, we believe that one of the most important feature of a bifunctional electron donor in propylene polymerization is the distance between the coordinating atoms, that must be suitable for chelating on tetracoordinate magnesium atoms located on (110) face of MgC1,. The results obtained with the catalyst system containing phthalates and silanes, both bifunctional bases, respectively as internal and external donors strongly support the above model. Through molecular calculation and conformational analysis it was possible to identify some diethers, in particular 1,3diethers, that have this right distance. These compounds, tested as internal donors in propylene polymerization give outstanding results in term of activity and stereospecificity. A

142

E. Albizrati. T. Dall’Occo. M. Cialimberti and G. Moririi

In table 1 are reported the performances of catalysts containing different diethers as internal donors. In general it is possible to observe that when the maximun probability value of the oxygen-oxygen distance, determined according to Conformation Statistical Distribution methodology [ 3 ] , is near to 3 Angstroms the catalyst performance is very good, whereas when the distance value is spread out in a wide range the performance is very poor.

Tab. 1: 1,3-DIETHERSBASED MgCl SUPPORTED CATALYSTS

DONOR

0-0 Distance

MILEAGE

A

KgPP/gCat

CI

2.9 - 4.7

4.0

CI

3.9 - 7.2

30.0

XI

2.7 - 4.0

35.0

74.9

Polymeridion conditions’ 4 I reactor,propylene 1.2 Kg. hyckogen 1.7 NI.

[TEAL] 2 5 rnrnoVI. 70’C. 2 hrs

13. New Heterogeneous Catalysts for Polyolefins

2)

143

POLYMER ARCHITECTURE CONTROL

The polymer architecture control allows t h e synthesis of new materials with improved properties like HETEROPHASIC OLEFIN COPOLYMERS and POLYOLEFIN ALLOYS with non-olefinic polymers which HIMONT has called HIVALLOY TECHNOLOQY:

-

HETEROPHASIC COPOLYMERS

Heterophasic polypropylene copolymers [ 4 ] are tough, high impact materials where polypropylene is the continuous phase and an elastomeric phase (usually an ethylene-propylene rubber) is uniformly dispersed within the matrix. Before the discovery and commercial exploitation of the high yield MgC1, based catalysts, such heterophasic copolymers, with high rubber content, were essentially made by melt blending of the preformed polymers in an extruder. There were, however, significant limitations on the properties of polymers that could be blended in this way. For istance, a too great difference in the melt viscosities of the various components of the blend prevents from the formation of domains of appropriate size of low modulus EPR particles. An optimum size is necessary to deconcentrate stress during impact which, according to classic rubber toughening theory, acts to initiate delocalized energy absorption and reduce catastrophic brittle failure. In figure 2 is reported a cross section of a such heterophasic polypropylene copolymer where EPR rubber particles can clearly be seen uniformly dispersed in the polypropylene matrix. To make the optimum heterophasic polymer structures directly in the polymerization reactor it is important to tailor the catalyst not only for the production of the ideal polymeric chemical structure but also for the location of the various polymeric phases. The rubbery phase must be homogeneously dispersed and its size controlled in order to achieve the best stiffness-impact balance. The catalyst must be capable of producing a homopolymeric phase with high isotacticity and then to copolymerize the desired elastomeric material with a high degree of randomness uniformly dispersed within the matrix.

144

E. AlhizLati, T. 1)all'Occo. M. Calirnherti and G. Morini

FIG.2

Cross section of heterophasic copolymer with an high rubber content (2000 x).

One of the most important concept developed from Himont/Montecatini's 40 years of commitment to research in Ziegler-Natta catalysis is the "Reactor Granule Technology" [5] where a controlled polymerization process takes place in each polymer granule according to diffusion and kinetic phenomena, related to the structure of the selected catalyst. It produces a growing, spherical granule that provides a porous reaction bed within which other monomers can be introduced and polymerized to form a polyolefin alloy. Figure 3 is a photograph of this Reactor Granule illustrat.ing the porous nature of the resin particle during the production of a polypropylene heterophasic alloy. This technology allows high rubber-containing blends and alloys to be made directly in the reactor and is not limited to a two components heterophasic system, indeed a third or even more phases can also be introduced. Figure 4 is a photograph of a reactor made PP/EPR heterophasic copolymer in which polyethylene has been incorporated to improve stress whitening on impact of a molded part.

13. New Heterogeneous Catalysts for Polyolefins

FIG. 3

FIG. 4

Reactor Granule (35 x)

.

Cross section of three-phase copolymer PP/EPR/PE after removal of rubber (20.000 x).

145

146 E. Albirrati, T. Dall’Occo, M . Galiniherti and G. Morini

-

HIVALLOY

The Reactor Granule Technology makes possible an exciting new technical frontier, allowing the incorporation and polymerization of non olefinic monomers in a polyolefin matrix. The porous polyolefin granule gives a very high specific surface area and a very high reactivity substrate suitable for easy reaction with non olefinic monomers at level greater than 50% wt. via free radical graft copolymerization. Himont calls this emerging technology Hivalloy [6] and the combining of non olefinic monomers with an olefinic substrate makes possible a family ofmaterials not commercially achievable previously. These resins are expected to bridge the performance gap between advanced polyolefin resins and engineering plastics, and are therefore, truly “Specialty Polyolefinsll. A first target of the Hivalloy family of products will be those applications currently served by ABS. Possessing both olefinic and non olefinic characteristics, Hivalloy products are designed to combine the most desirable properties of PP, such as processability, chemical resistence and low density, with many of desirable features of engineering resins which cannot be achieved with currently available polyolefins, such as improvements in the material’s stiffness/ impact balance, improved mar and scratch resistance, reduced molding cycle time and improved creep resistance. Because of their olefinic base, Hivalloy polymers readily accept minerals and reinforcing agents, providing added flexibility and control over properties further expanding the PP property envelope into the specialty area. In conclusion the basic requirements for a high yield catalyst suitable for Reactor Granule Technology are as follow: 1) High surface area. 2) High porosity. 3) High enough mechanical strength to withstand mechanical Processing, but low enough to allow the forces developed by the growing polymer to break down the catalyst into the microscopic particles that remain entrapped and dispersed in the expanding polymer particles.

13. New Heterogeneous Catalysts for Polyolefins 147 4)

5) 6)

Homogeneous distribution of active sites. Free access of the monomers to the innermost regions of the catalyst. Maintenance of above characteristics with a range of different monomers.

3)

POLYMER STRUCTURE CONTROL

-

HOMOGENEOUS CATALYSIS

Metallocene based catalysts [7,8,9] constitute a new class of extremely active Ziegler/Natta polymerization catalysts, which are able to produce all known polyolefins and are unique in producing a large series of new polymers such as highly stereoregular syndiotactic polypropylene, syndiotactic polystyrene, syndiotactic poly-4-methyl-l-penteneI perfectly random LLDPE and EPR/EPDM rubbers. In our laboratories we have recently discovered [lo] some new catalytic systems able to produce EPR and EPDM rubbers with a controlled microstructure in terms of random distribution of comonomers and microtacticity of the short propylene sequences. The elastomeric properties are particularly outstanding: the raw copolymer behaves as a cured elastomer (tab.2). We believe that the most relevant drawback related to the use of metallocene based catalyst systems is the difficulty in controlling the morphology of the polymer. For this reason we have focused our research activity towards the supportation of metallocenes, aiming at using this catalyst in gas-phase processes. In table 3 the most promising results [ll] that we have obtained are reported: it is worthwhile to point out that the properties of the polymers obtained employing the supported catalyst are very similar to those of the polymers obtained with the unsupported ones.

148

E. Albizzati. T. Dall'Occo, M. Galimberti and G. Morini

TAB. 2

-

MAIN FEATURES OF NEW EPR RUBBERS FROM METALLOCENES

VERY HIGH POLYMERIZATION ACTIVITY Ashes Content: Zr = 0.5

-

2 ppm, CI = 0.3 - 1.5 ppm, A1 < 400 ppm

NARROW DISTRIBUTION OF MOLECULAR MASSES AND CHEMICAL COMPOSITION MwIMn = 2 - 3 Ethylene content of a fraction = +I- 5 % of the ethylene content of the taw copolymer

VERY LOW CRISTALLINITY ethylene I propylene copolymers amorphous in the composition range: 40 - 75 % as ethylene molar content

NEW PHYSICAL - MECHANICAL BEHAVIOUR: OUTSTANDING ELASTICITY Tension Set (200%.23°C. 1')

< 10 (ethylene I propylene copolymer, 70% by moles of ethylene, 3 as I.V.)

ITH SUPPOIUED METAWx)CENE CATACYSTS POLYMER CATALYST

ACTMTY o(-Olefin

Tm

A M I.V.

Mw/Mn

BUM

NOTE

MNSrrY

LLDPE#

EPR*

#

KglgZrlh

wt.%

"C

J/g

dug

elmi

A-Unsupported A-Supported

800 220

15.6 15.6

95 101

77

0.92 2.5 1.4 3.8

0.20 0.44

Fine Light Powder SpherelGranule

6-Unsupported 6-Supported

460 190

11.6 11.7

99

05

91

79

1.3 1.1

2.9 3.6

0.15 0.44

Fine Light Powder SpherelGranule

C-Unsupported C-Supported

1300 1160

38 44

36 2.5 4.4 amorphous 3.4

3.0 2.0

.___ .___

Little Fouling No Fouling

Propane slurry polymerization process at 50°C. Propene slurry polymerization process at 50°C.

57

13. New Heterogeneous Catalysts for Polyolefins

4)

149

INTRODUCTION OF POLAR GROUPS AND UNSATURATIONS

In our laboratories we have developed a new synthetic route [12] to obtain polyolefins containing polar groups linked to the main backbone; so far this route presents only a scientific interest, however it is very important to prepare model compounds in order to evaluate the surface properties of these new materials in terms of compatibility with other polymers, fillers and fibers. The synthetic procedure consists of a copolymerization of propylene (and/or ethylene) with 4-iodo-lbutene in the presence of a vanadium based catalyst followed by a substitution reaction of iodine with different polar groups or a dehydrohalogenation reaction to obtain a pendant vinyl group (fig.5).

REACI'ION PATHWAY

Y - 0 2 = 100

-

100 100

x - 0

Polymerization

-

(x

+

(y

(x

+ 3 + 2)

y)

[CH~-CH~]-~-[-CH~-CH-(R)]-~-[-CH~-CH-(CHZ-CH~~]-~

X- I

K+ I lB-crown-6

[CH~-CH~]-X-[-CH~-CH-(R)]-~-[-CHZ-CH-(CH=CHZ)]-~

[CH~-CH~]-~-[-CH~-CH-(R)]-~-[-CH~-CH-(CH~-CH~-X)

X =

OR [Ow,OCOR, CH(COOR)2 [CHzCOOH], -N-PHTALIMIDE R =

Fb5

-

ALKYL, ARYL GROUP

ETHYLENE PROPENE COPOLYMERS CONTAININ0 #IuR QROUPS PREPAMTION OF MODEL COMPOUNDS

150 E. Albizzati, T. Dall’Occo, M. Galimberti and G. Morini

-

CROSS-LINKABLE PP

By using MgC1, based high yield catalysts it is possible to carry out the copolymerization between propylene and 1,3 butadiene [ 131 to obtain a random copolymer; the insertion of butadiene can be controlled by either the polymerization conditions or the catalytic system. Thus a copolymer with 5-10% of butadiene can be produced with prevailing 1,2 or 1,4 trans enchainment obtaining, in the first case, a saturated polymer with a pendant vinyl group and in the second an unsaturation in the polymer chain. Such unsaturate polymers are very reactive towards the radical grafting of polar monomers such as methylmethacrylate, styrene, butadiene, acrylonitrile or the crosslinking reaction necessary for a thermoplastic elastomer. 5)

MODIFICATION OF RHEOLOGICAL PROPERTIES

Aiming to improve the melt viscosity of PP it is possible to submit the polymer to treatment under specific conditions in order to introduce long chain branching. Such a modified polymer is useful for applications in which a high melt strenght is required as in foams, melt thermoforming and blow molding processes. Similar results are obtained via copolymerization of propylene and alfa-omega diolefins with MgC1, high yield supported catalysts.

13. New Heterogeneous Catalysts for Polyolefins

15 1

PEFERENCEB

[l] E.Albizzati, P.C.BarbB, L.Noristi, R.Scordamaglia, L.Barino, U.Giannini, G.Morini, U.S.Patent 4,971,937(1990) to Himont andP.C.BarbB,L.Noristi, R.Scordamaglia, L.Barino, E.Albizzati, U.Giannini, G.Morini, U.S.Patent 4,978,648(1990) to Himont and G.Agnes, G.Borsotti, G.Schimperna, E.Barbassa, U.S.Patent 5,095,153(1992) to Himont

.

[2] P.Corradini, V.Barone, R.Fusco and G.Guerra-Gazz.Chim.Ita1. 113,601 (1983) [3] R.Scordamaglia, L.Barino in J.K.Seyde1 Ed.QSAR Strategies in the Design of Bioactive Compounds, VCH, Weinheim, 1984, p.299 and R.Scordamaglia, L.Barino, Poster presented at "40 YEARS ZIEGLER CATALYSTS1', Freiburg, Sept. 1-3, 1993. [4] P.Galli, T.Simonazzi and D.Del Duca Acta Polimerica 39(1988) ,81. [5] P.Galli, J.C.Haylock Proceeding of SPE Meeting, Houston 2,24,1991. [6] P.Galli, A.De Nicola - Proceeding of "Strategies for Engineering Thermoplasticsll, Brussels 6,29,1992. Adv. Organomet. Chem. 18,99(1980). [7] H.Sinn, W.Kaminsky [8] J.A.Ewen J. Am. Chem. SOC. 106,6355(1984). Makromol. Chem. [9] A.Zambelli, C.Pellecchia and L.Oliva Macromol. Symp.48/49,297(1991). [lo] M.Galimberti, L.Resconi, E.Martini, F.Guglielmi and E.Albizzati Ital.Pat.App1. MI92A000666 to Montecatini Tecnologie. [ll] E.Albizzati, T.Dall'Occo, L.Resconi, F.Piemontesi, Ital.Pat.App1. MI93A001467. [12] M.Galimberti, U.Giannini, R.Mazzocchi, E.Albizzati and U.Zucchini, E.P.A.489,284 to Himont Inc. and M.Galimberti, U.Giannini, E.Albizzati, S.Caldari and L.Abis submitted to J.Mol.Cat. [13] G.Cecchin, F.Guglielmi and F.Zerega U.S.Patent 4,602,077 to Himont Inc.

.

-

-

-

-

-

-

This Page Intentionally Left Blank

I53

14. Change of Internal Donor for Mg(OEt),-Supported TiCI, Catalyst

DONG-HO LEE1, YOUNG-TAE JEONG2 and KAP-KU KANG2 1. Department of Polymer Science, Kyungpook National University, Taegu 702-701, Korea 2. Department of R & D, Korea Petrochemical Industrial Co., Ulsan 680-1 10, Korea ABSTRACT During the in-situ preparation of Mg(OEt)2/di-n-butylphthalate(DNBP)/TiC4 catalyst, some amount of DNBP converted into its derivatives such as diethylphthalate(DEP) >> ethyl-nbutylphthalate(EBP). With binary DNBPDEP internal donor, DNBP and DEP contents in catalyst could be controlled, and it was found that catalyst activity depends on relative amounts of DNBP and DEP rather than total diester content. With addition of Ti(OBu)4, DEP formation was suppressed and molecular weight distribution of polypropylene became wider. INTRODUCTION For high-efficient catalysts of propylene polymerization, it was found that MgCI2 is a suitable support and the preparation and characterization of MgC12-supported T i c 4 catalyst were studied intensively. ,2) In addition, magnesium alkoxides were reported as efficient support.3) Recently, Mg(OEt)2-supported Tic14 catalysts were prepared by not only physical milling method4) but also chemical reaction and the chemical composition and During the in-situ polymerization behaviours of those catalysts were studied in detail. preparation of Mg(OEt)2-supported Tic14 catalysts, diisobutylphthalate added as internal donor (ID) changes into its derivatives78 and this conversion of ID depends on preparation condition^.^) In this article, Mg(OEt)2/di-n-butylphthalate(DNBP)/TiCl4 catalyst was prepared by chemical reaction method, and the chemical composition as well as propylene polymerization behaviours of catalyst were studied in detail. Especially the changes of DNBP and effects of IDScomposition on catalyst activity and isospecificitywere examined. EXPERIMENTAL Reagents. Propylene(po1ymerization grade, 99.5% purity, Korea Petrochemical Ind. Co., Korea) was dried by passing through two columns of preactivated molecular sieve 4A. Triethylaluminum(TEA, Tosoh Akzo Corp., Japan), TiClq(Toho Titanium Co., Japan), di-nbutylphthalate(DNBP, Aldrich Chemical Co., U.S.A.) and cyclohexylmethyldimethoxysilane (CMDMS, Shin Etsu Co., Japan) were used without hrther purifications. n-Hexane(Tokyo

* Dedicated to Lee's mother deceased on February 28, 1994.

154

D.H. Lee, Y.T. Jeong and K.K. K a n g

Kasei Co., Japan) and n-decane(Aldrich Chemical Co., U.S.A.) were dried over preactivated molecular sieve 4A for 24 h and contained <5ppm water. Magnesium ethoxide(Mg(OEt)2, lmm dia., Huels AG Co., Germany) was used after vibrational milling under nitrogen atmosphere for 6 h at room temperature. Preparation of catalysts, Catalysts were prepared by chemical reaction method5.6): To 80ml of n-decane were added continuously Sg(69.9 mmol) of Mg(OEt)2, 60 ml of neat TiC14, 10.4 mmol of DNBP, and some amounts of Ti(0Bu)q if necessary. The mixture was allowed to react at 90-1200 C for 2 h with vigorous stirring. After cooling the reaction mixture, the solid product was separated, washed several times with 300 ml of n-hexane to remove Tic14 and dried in vacuo. Analvsis of catalvsts. Amount of titanium and magnesium elements in the prepared catalysts was determined by an atomic absorption spectrophotometer(AA, Perkin-Elmer Zeemad3030). A precisely weighed quantity of catalyst(100-150 mg) was dissolved in 10 ml of 0.1 M sulhric acid and diluted to 100 ml with distilled water. The amount of n-butoxy group in catalyst was determined as n-butanol by gas chromatography(GC, HP-5890).7) A precisely weighed catalyst(50-100 mg) was dissolved in 25 ml of 1M sulfkic acid and diluted to 250 ml with water. The capillary column(HP-FFAP, 0.2 mm id., 5 m 1.) was used, and the temperature of column increased from 350 C(ho1ding time, 5 min) to 1600 C with a heating rate of 50 C/min. For quantitative analysis of DNBP and its derivatives, exact amount(l.0-1.5 g) of catalyst was brought into contact with 50 ml of anhydrous methanol at room temperature and the resulted methanol solution with addition of ethyl benzene(EBz) as internal standard was injected to GC. A methyl silicone capillary column(Ultra-l,O.32 mm id., 25 m 1.) was used, and the temperature increased from 700 C(ho1ding time, 3 min) to 2800 C with 100 C/min. The retention times of EBz, diethylphthalate@EP), ethyl-n-butylphthalate(EBP) and DNBP were 2.99, 13.63, 15.70 and 17.54 min, respectively. Polymerization and characterization of polymers. The procedures of polymerization and characterization of polypropylene(PP) were described in the previous papers.4~8) Propylene pressure was kept at 10 kg/cm2 and polymerization was carried out at 700C for 2hr. CMDMS was used as external donor with mole ratio of [TEA]/[CMDMS] = 20. The molecular weight and its distribution were determined by GPC(Waters- 150-CV, Shodex AT-SOWS column).

RESULTS AND DISCUSSION Reaction of di-n-butvlphthalateOlNBP)and MdOEt12As mentioned in the previous papers,7~*,9)some parts of DIBP added as internal donor(ID) converted into its derivatives during the in-situ preparation of Mg(OEt)2-supported Tic14 catalysts. This chemical change of ID was considered due to the chlorination of ester with Tic14 followed by reaction with Mg(OEt)27>8)and the reaction of ester with Ti(OEt),Cl4-, which is produced from Mg(OEt)2 and TiC14.9)

14. Internal Donor for Mg(OEt),-Supported TICI, Catalyst

155

With the above backgrounds, direct reaction of di-n-butylphthalate(DNBP) and Mg(OEt)2 in the absence of Tic14 was carried out at 90, 100 and 115OC, and reaction products of diester were analyzed with GC as shown in Table 1.

Table 1. Reaction Products of DNBP and Mg(OEt)2 Reaction Temp(OC) Diester(mo1efraction)

-.-

100

90 DNBP EBP

DEP

115

DNBP EBP

DEP .

Reaction

0

Time(min)

10 20 30 40

1.00 0.93 0.79

0.00 0.00 0.07 0.00

60

0.60

0.19 0.02 0.34 0.06

90

0.20

0.50

0.30

1.00 0.93 0.86 0.68 0.38 0.12

0.00 0.07 0.14 0.29 0.47 0.47

DNBP EBP

DEP

~

~

0.00

1.00

0.00

0.74

0.22 0.04

0.35 0.00

0.34 0.34

0.00

0.00 0.00 0.04

0.15 0.41

0.31 0.66

-

-

a); mole ratio of DNBP/Mg(OEt)2 = 0.16

As shown in Table 1, DNBP converted into its derivatives such as ethyl-nbutylphthalate(EBP) and diethylphthalate(DEP), and DNBP portion decreased with increasing reaction temperature and time. In addition, the produced amount of EBP was larger than that of DEP in every conditions. With the above results, it was considered that transesterification reaction between DNBP and Mg(OEt)2 occurs even in the absence of Tic14 and DEP produces via EBP stepwisely. This transesterification of DNBP wth Mg(OEt)2 was sensitive to reaction temperature and time. Mg(OEtl2/DNBP/TiClq - catalvst The chemical composition and polymerization behaviours such as activity and isospecificity

of Mg(OEt)2/DNBP/TiCIq catalysts prepared at different temperature were examined and the results are given in Table 2 and 3. As expected, conversion of DNBP into EBP and DEP was also possible during the in-situ preparation of Mg(OEt)2/DNBP/TiClq catalyst. However, DNBP was still remained even at 12OOC, 2 h while DNBP disappeared completely even at 1150 C, 1 h for the reaction with Mg(0Et)z in the absence of TiClq.(Table 1) In addition, DEP was main product for catalyst preparation while EBP was produced mainly for the reaction of DNBP with Mg(OEt)2 only.

156 D.H. Lee, Y.T. Jeong and K.K. Kang

Table 2. Chemical Composition of Mg(OEt)2/DNBP/TiClq Catalysts Obtained with Reaction for 2 h at Different Temperature ~

_

_

_

Catalysts

Reaction

No.

Temperature(°C)

Ti(wt%)

DNBPa)

EBPa)

DEPa)

Total Diestera)

90

3.1

34 1

28

72

44 1

100

3.0

266

44

135

445

110

2.6

144

302

2.3

126 57

32

120

24

144

225

a): concentration, xi03 nunoYgatalyst b); mole ratio of DNBP/Mg(OEt)2 = 0.16

Although the reason of these different results is not clear at the present time, the mutual interactions between components such as complex formation of various diester(DNBP, EBP, DEP) and TiC14, chlorination of diester and formation of Ti(OEt),&x, etc. could exist. With increasing reaction temperature, DNBP amount as well as total amount of diesters decreased while DEP increased, and the titanium content decreased slightly.

Table 3. Activity and Isospecificity of Reaction for 2 h at Different Temperature Catalysts No.

1 2 3 4

Activity a) 1.1(wt%) without H2 27.9 27.2 27.0 27.8

99.0 98.6 98.5 98.4

Mg(OEt)2/DNBP/TiClq Catalysts Obtained with

Activity a) 1.1(wt%) with H2b) 63.4 64.0 63.0 61.1

98.5 98.4 98.3 98.3

a); catalytic activity, kg-PP/g-Ti.h.atm b); hydrogen pressure, 0.18 k@m2

As shown in Table 3, the catalyst activity and isotactic index(I.1) of polypropylene(PP) had less dependence on diester content in the prepared catalyst.

14. Internal Donor for Mg(OEt),-Supported TiCI, Catalyst

157

For Mg(OEt)2/diisobutylphthalate@IBP)/TiC~catalyst, Chadwick et d.9) reported that catalyst activity decreased and fraction of xylene-soluble PP increased with decreasing total diester content in the catalyst. To check the effects of total amount and composition of diesters on chemical composition and polymerization behaviours of catalysts, various catalysts were prepared by adding different mole ratio of DNBPDEP as binary ID and the experimental results are given in Table 4.

Table 4. Chemical Composition and Polymerization Behaviours of Catalysts Prepared with Binary DNBP/DEP Internal Donor at 9OOC, 2h Catalysts DNBP/DEP DNBPa) EBPa) DEPa) Total Diestera) Ti No. (Mole Ratio) (wt %)

Activityb)

5

1 .OO/O.OO

336

27

44

407

2.9

62.7

98.9

6

0.7W0.25

332

27

141

500

2.9

58.8

98.5

7

0.67/0.33

214

24

195

433

2.7

59.7

98.5

8

0.50/0.50

111

13

305

429

2.4

58.3

98.6

9

0.33/067

111

14

366

491

2.4

56.7

98.4

10 11

0.2W0.75 0.00/1.00

109

11

443 485

563

2.3

55.7

98.3

485

2.3

46.1

97.8

-

1.1 (wt %)

a): concentration, x103 mmot/gzatalyst

b): catalytic acfivky, kg-PP/g-Ti,h.atm,hydrogen pressure, 0.18 kg/cm2 c); mobratio of DNSP/Mg(OEt)2 = 0.16

It was expected that DEP formation from DNBP is suppressed with DEP addition if transesterification of DNBP to DEP was equilibrium reaction. However, as shown in Table 3, DEP content increased steadily while DNBP content decreased continuously with increasing amount of added DEP. EBP formation could be neglected compared to DEP. The total content of ID was in range of 0.40-0.55 mmoYg-catalyst with less correlation to added mole ratio of DNBPDEP. Titanium content of catalyst decreased slightly with increasing DEP amount. The catalyst activity showed strong dependence on mole ratios of DEP or DNBP rather than total amount of ID; i.e. catalyst activity decreased with increasing amount of DEP or decreasing amount of DNBP. The heptane-insolubleportion of PP was almost constant irrespective to ID amount and composition.

158 D.H. Lee, Y.T. Jeong and K.K. K a n g

Those above results are contradictory to those of Chadwick.9) The reason is not clear, but it might be due to differences in experimental conditions such as catalyst preparation procedure, catalyst washing medium and external donor, etc. M~~OE~7fl>NsPrriCl~rri~OBu!4 - catalyst From the above experiments, it has been found that DEP is formed by transesterification of DNBP even with the presence of DEP during the in-situ preparation of catalyst and DEP amount should be diminished for high catalyst activity. To suppress the formation of DEP, various amount of Ti(OBu)4 was added with ID in the procedure of catalyst preparation . The effect of Ti(OBu)4 addition on catalyst composition was examined as shown in Table 5 .

Table 5 . Effects of Ti(OBu)4 Addition on Catalyst Composition Catalyst Ti(OBu)4/ Ti

No.

Mg(OEt)2

DNBPa) EBPa) DEPa) Total

DNBP

DEP

Butoxy

Diestera) Fraction Fraction (wt%)

(wt%)

(mole ratio) 386

28

23

437

2.8

332

57

46

435

2.8

259

71

42

372

2.9

250

92

31

373

2.8

156

128

34

3.0

120

80

14

150

94

16

260

2.4

317

124

13

1.5

309

89

0

12

0.0

3.0

13

0.2

14

0.4

15

0.6

16

1.o

17

1.5

18

2.0

2.7

19

3.O

20

4.0

0.05

0.23

0.76

0.10

0.58

0.70

0.11

0.77

0.67

0.08

0.78

318

0.49

0.11

1.04

214

0,56

0.07

1.01

0.58

0.06

0.88

454

0.70

0.03

0.94

398

0.78

0.00

0.90

0.88

a): concentration, x103 mmol/g-catalyst

b); catalyst preparation ;90 OC, 2h

With addition of Ti(0Bu)q in catalyst preparation, DNBP content as well as DEP content decreased while EBP amount increased with much contribution. In this case, minor component was DEP(mo1e fraction; ca. 0. l), which indicated that Ti(OBu)4 has some contribution in transesterification to suppress DEP formation. Total amount of diester and DNBP mole fiaction decreased with increasing Ti(OBu)4 amount for Ti(OBu)4 < 2.0 mole. The butoxy

14. Internal Donor for Mg(OEt),-Supported TiCI, Catalyst

I59

group came from TiCl,(OBu)q_, and its content in the catalyst increased with amount of Ti(0Bu)q. Titanium content was remained almost constant. The effects of Ti(0Bu)q addition on catalyst activity, isospecificity and molecular weight of PP were examined as shown in Table 6.

Table 6. Effects of Ti(0Bu)q Addition on Catalyst Behaviours for Propylene Polymerization Catalyst

Activiv)

I.I(Wt%)

Mw/Mn

Mw/Mn

with H2b)

10-3 mi H2

10-3 H2 b)

6851113

218153.4

Activip)

I.I(Wt%)

No.

without H2 12

27.7

99.3

63.3

98.8

13

21.4

98.4

61.6

98.2

14

17.9

98.0

59.5

98.4

15

14.8

98.2

53.1

98.3

16

13.4

98.6

44.6

98.7

17

12.3

98.7

36.3

98.5

18

10.0

96.8

32.2

98.4

19

7.1

97.0

20.4

97.3

540178.2

191138.7

20

6.7

96.0

16.3

97.1

587180.0

187137.3

216153.0 6301103 22615 1.5 642199.5

206145.0 209142.5

a): catalytic activity, kg-PP/g-Ti.h,atm

b): hydrogen pressure, 0.18 k&m2

The catalyst activity decreased continuously with increasing Ti(0Bu)q amount while 1.1 was constant for Ti(0Bu)q < 2.0 mole. For Ti(0Bu)q > 2.0 mole, activity decreased drastically and 1.1 also decreased slightly due to the large amount of alkoxy titanium active site which has less activity and less stereoregularity.10) To remove the alkoxy titanium species, the catalysts were retreated with Tic14 and catalyst composition as well as polymerization behaviours were examined as shown in Table 7. By retreatment of the catalysts with TiClq, titanium and butoxy contents as well as amount of DNBP decreased while catalyst activity increased with unchanged 1.1. The weight-average molecular weight(Mw) and number-average molecular weight(Md decreased simutaneously with addition of Ti(0Bu)q as shown in Table 6. The polydispersity index(P1, M@n) was measured for various amount of Ti(OBu)4 and the results were plotted in Fig. 1.

160 D.H.

Lee, Y.T. Jeong and K.K. Kang

Table 7. Effect of TiClq Retreatment

Catalyst Ti(wt%) No. 14 16 18 20

Activitya) no H2

Activitya) DNBPC) EBPC) H2b)

DEPC)

(AY@)

(AY@)

(A)/@)

(A)/@)

(A)/@)

(AY@)

2.812.5 2.812.4 2.712.3 1.511.3

17.9127.7 13.4125.0 11.1127.4 5.718.4

59.5170.6 44.6172.9 34.U40.1 16.3130.0

45140 23115 1861173 5671231

65/68 3551335 1251200 290/200 62/61 I 100150 I

Butoxy (wt%)

- - -

(A)/@) 0.7510.64 1.0410.67 1.0310.36 0.7810.50

(A); without Tic14 retrcatmnt (B); with Tic14 (60ml)retreatment

a); catalytic activity, kg-PP/g-Ti.hatm b): hydrogen pressure, 0.18 kg/cm2 c); concentration, x103 mmougatalyst

4.5 4.0

+

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Ti(OBu)q/Mg(OEt)2 (Mole Ratio) Fig. 1 Change of polydispersity index(Mw/M,J of PP with amount of Ti(0Bu)q in absence(0) and presence(@) of hydrogen

As shown in Fig. 1, PI increased 0.9-1.2 more due to larger contribution of low molecular weight portion with amount of Ti(0Bu)q. However, molecular weight distribution became narrower in the presence of hydrogen as shown in the previous paper.11)

14. Internal Donor for Mg(OEt),-Supported TiCI, Catalyst

161

REERENCES 1) P.C. Barbe, C. Cecchin and L. Noristi, Adv. Polym. Sci., 8,1 (1987) 2) T.Keii and K. Sogqeds.), "CatalyticOlefin Polymerization",Kodmha, Tokyo, 1990 3) US.Patent 4,548,951(1985)(Shell Oil Co.) 4) Y.-T. Jeong and D.-H. Lee,Makromol. Chem., 191,1487 (1990) 5 ) Y.-T. Jeong, D.-H. Lee and K. Soga, Makromol. Chem., Rapid Commun., 12,s (1991) 6 ) Y.-T. Jeong, D.-H. Lee, T. Shiono and K. Soga, Makromol. Chem., 192,1727 (1991) 7 ) D.-H. Lee,Y.-T. Jeong and K. Soga, Znd Eng. Chem. Res., 31,2642 (1 992) 8) D.-H. Lee, Y.-T. Jeong, K. Soga and T. Shiono, J. Appl. P o h . Sci., 47,1449 (1993) 9)J.C. Chadwick, A.Miedema, B.J. Ruisch and 0. Sudmeijer, Makromol. Chem., 193, 1463 (1992) 10)T. Garrof,E.Iiskola and P. Sormunen, in "TrmitionMetals and Organometallics(IS Cutulystsfor Olefin Polymerization",W . Kaminsky and H. Sinn(eds.), p. 200, Springer-Verlag,Berlin, 1988 11) D.-H. Lee and Y.-T. Jeong, Eur. Polym. J., 29,883 (1993)

This Page Intentionally Left Blank

I63

15. Temperature Programmed Decomposition of MgCl,/THF/TiCl, Bimetallic Complex Catalyst and its Effect on the Homo- and Copolymerization of Ethylene

Y. S. KO, T. K. HAN and S. I. WOO Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, Yusong-gu, Daejon, 305-70 1, Korea ABSTRACT A MgC12/THF/TiC14 bimetallic complex catalyst was prepared by reacting magnesium chloride with titanium tetrachloride in tetrahydrohran(TH3). During the temperature programmed decomposition(TPD) of the bimetallic complex, THF and 1,4 dichlorobutane were identified by Mass spectroscopy(MS). TI-F decoordinated from Ti species reacted with adjacent CI, resulting in the formation of 1, 4-dichlorobutane. When the MgC12/THF/TiC14 bimetallic catalyst (Mg/Ti=5.2) was heated below 108 OC, the catalytic activity of polymerization increased, while it decreased above 140 OC. In ethylene-1-hexene copolymerization, the lowest catalytic activity was obtained at the molar ratio of hexene to ethylene in monomer feed(CH/CE), 1.14 or 2.22. The comonomer distribution of copolymer prepared with thermally pretreated catalyst was more homogeneous than that of copolymer prepared without thermal treatment. INTRODUCTION The MgC12/THF/TiC14 bimetallic complex catalyst was reported that it had high activity in ethylene polymerization with aluminum alkyl cocatalyst.’) 2) 3) Sobota have concluded that the three different complexes( [Mg(THF)6][TiCI5(THF)], [(THF)4Mg(p-C1)2TiC14], [Mg2(pC1)3(THF)3][TiCI~(THF)] ) were synthesized from the reaction between TiC14(THF)2 and MgC12(THF)2.4) These complexes can be decomposed by thermal energy and the structures of active sites can be also changed due to the decoordination of weakly coordinated THF from the complexes. This may affect the catalytic activity of bimetallic complex. The coordination site and strength of THF to MgC12 and Tic14 will determine its polymerization behavior in ethylene polymerization. Therefore, the temperature programmed decomposition(TPD) study was performed to obtain some informations on the structures of bimetallic catalyst. Comonomer effects of or-olefin on kinetics in ethylene copolymerization and properties of copolymer were reported by many authors due to their industrial importance.’) 6 ) 7 ) In ethylene copolymerization using MgCl2/THF/TiC14 catalyst it was reported that the addition of 1-hexene decreased the rate of ethylene consumption compared to homopolymerization.*) Thermal pretreatment of bimetallic catalyst can influence the kinetics of copolymerization, comonomer

164 Y.S.

KO,T.K.Han and S.I. W O O

distribution and the properties of copolymer significantly. In the present study, the temperature programmed decomposition of MgC12/THF/TiC14 bimetallic complexes of various Mg/Ti molar ratios was performed. The effect of the thermal pretreatment of the bimetallic complex catalyst prepared under various condition(temperature and time) on the ethylene and ethylene- 1-hexene polymerization was investigated. EXPERIMENTS The MgC12/THF/TiC14 bimetallic complex catalyst was prepared by the precipitation method. The reactivity ratio of monomers in ethylene-1-hexene copolymerization was determined after 30 min of polymerization. The detailed procedures for the preparation of catalyst and the polymerization were provided elsewhere.2)3)The TPD and MS experiments were conducted for the analysis of the evolved gas during heat treatment of the samples. The evolved gases were analyzed by MS. The detailed procedures have been given el~ewhere.~) Polyethylene and ethylene- 1-hexene copolymers were fractionated with boiling heptane for 6 hrs in a Soxhlet extraction apparatus. Copolymer composition was measured by the IR method using the calibration curve based on A138dA1368 absorbance ratio as reported by Nowlin et al.l0) Melting points and heats of fusion of polymers were measured by DSC. Two DSC procedures have been used as follows. Method A is that first scan temperature was raised at 20 Wmin and second scan temperature was raised at 5 W m i n from 50 OC to 150 OC. Crystalhities of polymer were calculated by the equation, x(%) = 100 x AH,/293 where AHf is the heat of fusion measured by DSC. Method B is that polymer samples were melted at 160 OC at inert atmosphere for 3 hrs. Then the sample was successively annealed at 125, 113, 97, 87, 78, 69, 5 5 , and 35 OC for 12 hrs at each step. RESULTS AND DISCUSSION Figure 1-(A) shows the TPD spectra of bimetallic catalysts of various Mg/Ti molar ratios monitored by TCD detector. When Mg/Ti was 5.2,the catalyst was decomposed at 108, 140 and 242 OC and only THF was observed at 80 OC. THF and 1,4-dichlorob~tane,however, were observed at 210 OC by MS. As MgiTi ratio decreased, TPD spectra became similar to that of TiClq(THF)2(Mg/Ti = 0). When MgiTi ratio increased, TPD spectra became similar to that of MgC12(THF)2. Figure 1-(B) shows temperature-programmed mass-spectra of THF(mass to charge ratio; 42) and 1,Cdichlorobutane (mass to charge ratio; 55). 1,4-dichlorobutane was not be produced in the spectrum of MgC12(THF)2 and only a small amount of 1,4-dichlorobutane was detected in the TPD spectra of the bimetallic complex of high Mg/Ti molar ratio. It may be concluded that some of THF coordinated from Ti by thermal treatment reacted with adjacent CI to form 1,4-dichlorobutane. Figure 2 shows the polymerization rate profiles polymerized with the thermally-pretreated bimetallic catalyst. Thermal treatments at 80 OC and 108 O C enhanced the activity in ethylene polymerization. Above 140 OC, however, the polymerization activity was decreased. These results could be explained by the fact that new active sites were formed by decoordination of THF during thermal treatment at 80 and 108 OC. However, titanium active sites were unstable above 140 OC.

15. Temperature Programmed

Decomposition of MgCI,/THF/TiCI,

0

100

200

300

100

200

165

300

T e m p e r m ~ u r a .OC 0

ZOO

100

300

T e m p e r a t u r a , ‘C

(A) (B) Figure 1. TPD spectrum (A) and mass spectra (B) for THF(-) and 1,4-dichlorobutane(---)of MgC12/THF/TiC14 catalysts. (a) Mg/Ti=O, (b) Mg/Ti=l.O, (c) Mg/Ti=2.1, (d) Mflk5.2, (e) Mg/Ti = 16.5, (0 Mg/Ti = w. m

I m e , rnin

Figure 2. Ethylene polymerization rate profiles after thermal treatment of bimetallic catalyst(Mg/Ti = 5.2). Thermal treatment condition: (a) none, (b) 80 OC, Smin, (c) 108 OC, 5 min, (d) 140 OC, 5 f i n ; Polymerization condition: Pethylene = 3 atm; T= 70 OC and [AI]/[Ti] =128. From these results, we can propose the plausible change in the structure of the MgC12/THF/TiC14 bimetallic complex as shown in Figure 3. When Mg/Ti is 5.2 the catalyst is a mixture of [MgC12(pL-Cl)3(THF)6]+[TiCI5(THF)]and MgC12(THF)2, which was reported by Sobota et al.4)

166 Y.S. KO, T.K. Han and S.I. Woo

> >

140

c

HWnn

0 I

TI

/

I I I I

\

CI

CI

CI-C-C-C-C-CI I l l 1

Wlr

* THF

w w nn

IIV)

Figure 3. Plausible change in the structure of bimetallic catalyst (MgITi treatment.

=

5.2) during thermal

Ethylene and 1-hexene were copolymerized at 70 OC for 30 min with the catalyst (Mg/Ti = 5.2) thermally pretreated at various temperatures. The activity in polymerization and properties of polymers are summarized in Table I. The hexedethylene (CH/C,) molar ratio was changed id the range of 1.14 - 5.42. As shown in Figure 4(A), ethylene consumption rate (activity) for TT-0 increased, which was explained by the fact that the physical disintegration of decreased as CH/C~ catalyst particle did not happen rapidly by I-hexene during the polymerization with catalyst of high Mg/Ti ratio and that propagation rate of 1-hexene is smaller than ethylene.*) The different trends in changes of activity at various 1-hexene concentrations were observed with TT-1, 2, 3 and 4 as shown in Figure 4(B) and Table I. In the case of TT-1, the activity decreased in the range of CH/CE molar ratio between 0 and 1.14. The activity increased when C H / C is ~ higher than is higher than 1.14. The similar trend for TT-2 and TT-3 was observed. When the catalyst was heated at 108 OC for 60 min, the ethylene consumption rate in copolymerization increased when C,/C, was above 2.22. These results also demonstrate that the new copolymerization active site was formed after the decoordination of THF by heating. Table I shows that the comonomer content of copolymer polymerized by TT-0 was slightly higher than that of TT-I, 2, 3, 4’s. The crystalhities of the copolymers obtained by TT-1, 2, 3, 4

15. Temperature Programmed

167

Decomposition of MgCI,/THF/TiCI,

were higher than that of the copolymers obtained by TT-0 despite of the similar comonomer content. This can be explained by the fact that comonomer distribution of the copolymer became more homogeneous after heat treatment. Each melting peak in Figure 5 is representative of a distinct family of macromolecules (or blocks) with different short chain branching. l )

,

300

I

h

L c

I

I

I

I

-F .c I

200

M

'r

-2

-E

150

h 0

h

I

0

a

100

I M Y

M

x

2

4

I

L

250

.F.

h,

600 I

v

v

a

50

E0

0

I0

0

20

30

0

I

I

I

I

I

5

10

15

20

25

30

TIME(M1N)

TIME(MIN)

Figure 4. Ethylene consumption rate in copolymerization; (A) catalyst not thermally treated, (B) heated at 108 OC for 60min; Copolymerization condition : P=3atm, T=70 OC,[AI]/[Ti]=128.

=

CJC,

0.00

c,/c,=0.00

7--

3 0.5 W/g

0.5 W/g

30

60

90 120 Temperature( 'C)

150

30

60

I20 0 Temperature( C)

150

(A) (B) Figure 5 . DSC thermograms of ethylene-1-hexene copolymers after annealing; (A) not thermally pretreated; (B) heated at 108 OC for 60min.

168

Y.S. KO, T.K. Han and S.I. Woo

Table I. Effect of 1 -Hexene on the Polymerization of Ethylene Copo1ymerization.a Catalyst

CHICE Tm molar (OC) ratio

xc (%)

R ,b C6 in 3tmin Copolym. (mol %)

in feed

0.00

TT-0

TT- 1

TT-2

TT-3

1.14 2.22 3.57 5.42 0.00 1.14 2.22 3.57 5.42 0.00 1.14 2.22 3.57 5.42 0.00 1.14 2.22 3.57 5.42 0.00 1.14 2.22 3.57 5.42

136.0 125.4 123.4 121.8 122.8 134.2 124.3 123.8 123.8 121.7 134.4 126.5 124.4 122.7 122.6 135.9 126.9 125.0 122.8 123.5 135.3 128.1 124.9 122.5 121.5

52.1 126.9 32.3 104.9 30.3 105.7 28.8 82.6 18.1 31.9 55.2 147.1 39.5 129.2 34.1 138.7 33.8 150.0 24.3 174.0 57.4 142.0 42.2 126.0 36.8 156.7 35.3 162.1 27.7 164.2 58.8 179.4 37.5 131.3 33.2 132.7 32.9 157.9 20.3 178.1 54.1 198.4 45.0 174.9 39.6 103.0 33.5 163.2 29.6 206.7

Reactivity ratio by F-RC

Reactivity ratio by M-~d

0.0

‘1 52.6 ‘2 -0.13

1.8 2.6 3.1 4.0 0.0 1.7 2.4 2.8 4.0 0.0 1.7 2.4 3.3 3.5 0.0 1.7 2.4 2.6 3.9 0.0 1.5 1.9 2.4 3.0

‘1 55.1 ‘2 -0.14

67.1 -0.085

77.0 -0.072

‘1

55.3 r2 -0.13 ‘1 56.3 ‘2 -0.12

64.5 -0.098

79.2 -0.072

‘1 52.9 87.0 ‘2 -0.13 -0.090 a AyTi=l28, T = 70 OC, t = 30min, P = 3atm. Catalyst; TT-0 : no thermal treatment, TT-I: 80 OC, 5 min, TT-2: 80 OC, 60 min, TT-3: 108 OC, 5 min, TT-4: 108 OC, 60 min. Activity = kg-polymer(g-Ti hr)-l Calculated by Finemann-Ross equation. Calculated by Mayo-Lewis equation.

TT-4

The reactivity ratio in Table I can be calculated from the copolymer composition by FinemannRoss and Mayo-Lewis equation. Bohm suggested that ‘1 can be evaluated by simplification for low comonomer content. Iz) In our present study, full equation approaches were used because simplification did not work reliably. The Finemann-Ross ( 1 ) and Mayo-Lewis (2) equations are given as follows. d [ H ] I d [ E ] = [ H ] I [r,E[HI ] I [El + 1 ---_( 1 ) [HI / [ E l+ rz

F ( f - 1 ) - F2 - -r, -r, ---- (2)

f

f

15. Temperature Programmed Decomposition of MgCI,/THF/TiCI,

169

where d[Hl/d[E] ( f ) and [HI@] ( F ) are hexendethylene molar ratio in the copolymer and initial molar ratio in feed, respectively. The two procedures gave somewhat different rl and r2 values and negative values of r2. Floyd explained that a physically meaningless negative value of r2 can be taken as an indication of heterogeneity of polymer.'3) CONCLUSIONS TJ3F and l,4-dichlorobutane were produced during the thermal treatment of the MgC12/THF/TiC14 bimetallic catalyst. Some THF coordinated to Ti was decomposed by thermal treatment and reacted with adjacent C1 of Tick, to form 1,4-dichlorobutane. The decoordination of THF resulted in the formation of new active sites. The thermally pretreated catalyst showed the higher activity than the catalyst not heated in ethylene polymerization. In ethylene-1-hexene copolymerization, the catalytic activity of thermally pretreated catalyst increased at high CH/CE molar ratio due to the new active sites. The copolymer polymerized by thermally pretreated catalyst showed a comonomer distribution more homogeneous than that obtained by the not thermally pretreated. REFERENCES 1. Han,J. D., Kim, I., and Woo, S. I., Polymer(Korea),13,147(1989). and Woo,S. I., "Catalytic Olefin 2. Kim, I., Chung, M. C., Choi, H.K., Kim, J. H., Polymerization", Soga, K. Eds., Kondansha Ltd., Tokyo, 1990, p323. 3. Kim, I. and Woo, S. I., Polym. Bull., =,239(1989). 4. Sobota, P., Utko, J., and Jana, Z., J. Organomet. Chem., U19(1986). 5. Kissin, Y. V. and Beach, D. L., J. polym. Sci. Polym. Chem Ed., =,333(1984) 6. Chien, J. C. W. and No&, T., J. Polym. Sci. Polym. Chem. Ed., 227(1993) 7. Jaber, I. A,, Ray, W. H., J. Appl. Polym. Sci., 1709(1993). 8. Kim, I., Kim, J. H., Choi, H.K., Chung, M. C., and Woo, S. I., J. Appl. polym. Sci., 48, 72 1(1993). 9. Choi, H.K., Chung, D. W., Han,T. K. and Woo, S. I., Macromolecules,-6 2 452(1993). 10. Nowlin, T. E., Kissin, Y. V., Wagner, K. P., J. Polym. Sci. Polym. Chem. Ed., 26, 755(1988). 4337(1992). 11. Addison, E., Ribeiro, M., Deffieux, A., Fontanille, M., Polymer, 12. BBhm, L. L., J. Appl. Polym. Sci., 29,279(1984). 13. Floyd, S., J. Appl. Polym. Sci., 3,2559(1987).

a,

a

This Page Intentionally Left Blank

171

16. Characterization of Mg/Ti Type Catalysts Prepared from Different Mg Components

M. Murata ,A. Nakano ,S. Kanazawa and M. Imai Tonen Chemical Corporation Tonen Corporate R & D Laboratory 1- 3 - 1 Nishitsurugaoka , Ohi - machi , Iruma - gun Saitama 356, Japan

Summary M f l i catalysts for propylene polymerization were prepared from several Mg compounds such as MgC1, , MgClJ2-ethylhexylalcohol solution Mg(OEt), or Mg(0Et)Cl. Catalyst structure regarding Ti atom location in the solid was examined by elemental analyses, X R D and XPS. Ti atoms in the catalysts prepared with MgC1, and MgCld2-ethylhexylalcohol solution were locally concentrated on MgC1, surface. In contrast with this , it was suggested that Ti atoms in the catalysts from non - MgCl, components of Mg(OEt), o r Mg(0Et)Cl were highly dispersed in the particle. It was also confirmed that propylene polymerization behavior such as initiation and rate decay was influenced by the catalyst structure. These results strongly indicate that the properties of active site are changed by the kind of Mg component used for catalyst preparation. Introduction It is well known that Mg/Ti type catalysts have excellent polymerization performances for olefin polymerization”. And many different preparation methods have been reported so far. However, most of the research effort in this area has been focused on the discussions of the roles of Mg component or donors”. On the other hand, there are very few studies aiming the understand of a relationship between catalyst preparation process and polymerization behavior. Terano et. al. reported that the propagation rate constants(kp) were changed by catalyst preparation

172 M. Murata, A. Nakano, S. Kanazawa and M. lmai

procedure3’. Although this result indicates that the properties of active site are controlled by catalyst synthetic procedure , the origin to vary the characteristics of sites are still not clear. In this study , the transformation of catalyst structure , which will be induced by the kind of Mg component used as starting material , will be discussed. Experimental Catalyst preparations 10 catalysts prepared by various Mg compounds and procedures combinations were used in this study : Cat - 1 : 10.5g of MgC1, and 1.4g of TiC1, were placed in a 0.31 stainless steel vibration mill pot with 650g of 12mm# balls and ground for 8h at room temperature. Cat - 2 : cat - 1 preparation procedure was repeated with 10.5g of MgC1, and 0.15g of TiC1,. After milling, solid part was washed with n-hexane. Cat - 3 : 10.5g of MgCl, and 0.02ml of di-ethyl phthalate@EBP) were co-ground by same condition as cat - 1 preparation. Solid part was treated with lOOml of TiCl, of 110°C for 2hrs. The obtained product was further washed with n - hexane. Cat - 4 : In place of DEBP , di-n-buthyl phthalate(DNBP) was used in Cat - 3 preparation. solution was prepared by the Cat - 5 : MgC1~2-ethylhexylalcohol(2-EHA) reaction of log of MgC1, and 48.7ml of 2-EHA in n-dodecane at 130°C for 2h. This solution was introduced dropwise to 25Oml of TiC1, cooled at -20°C. The solution was heated up to 120°C for 4h and treated at the temperature for 2h. The solid product was further washed with n-hexane. Cat - 6 : 2.428 of DNBP was dissolved in MgCld2-EHA solution prepared by the same procedure as the case of Cat - 5 . After the conduction of same TiCl, treatment for Cat - 5 , the solid was treated again with 250ml of TiC1, at 120°C for 2h. The obtained solid was further washed with n-hexane. Cat - 7 : 1.75mol/l of n-ButhylMgCl was prepared in di-n-butyl ether from Mg and n-Butylchloride. HC(OEt), was added to the Grignard solution

16. Characterization of Mg/Ti Type Catalysts

173

(HC(OEt)JMg=lmol/mol) at room temperature and gradually heated up to 80°C and maintained at the temperature for 4h. The precipitated solid was washed with n-hexane. log of the solid was treated with 2.2.2-trichloro ethanol(TCE) (TCE/Mg=0.3ml/ml) to obtain Mg component. After washing with n-hexane and toluene, Mg component was treated with lOOml of TiC1, at 90°C for 2h. The obtained product was further washed with n-hexane. Cat - 8 : log of Mg(OEt), and 0.02ml of DNBP was co-ground under the same condition as the case of Cat - 1 preparation. Twice TiCl, treatments at 120°C for 2h each were conducted and the product was washed with n-hexane. Cat - 9 : Mg component was prepared with n-ButylMgCl and Si(OEt),. The solid obtained( log) was treated with DNBP(0.02mol) and Tic&(100ml) at 120°C for 2h. After remove out the solution, solid part was trearted with 1OOml of TiC1, at 120°C for 2h. The product was further washed with n-hexane. Cat - 10 : Mg component was prepared by the same procedure as the case of Cat- 7. And DNBPniCl, treatment was conducted with C a t - 9 preparation method. Characterizations of Catalysts Crystallographic evaluation of catalyst solid was conducted by X-ray diffraction. (XRD) Ti and Mg contents were determined by normal elemental analyses. Ti/Mg molar ratio of catalyst particle surface was measured by X-ray photoelection spectroscopy.(XPS , Kratos Exam 800) It was observed that peak intensities of Ti and Mg were decreased with increasing of X-Ray exposing time. Thus, the amount of the element originally exsisted on the surface was determined by extraporation of the intensity-time curve to time zero. Propylene Polymerization Slurry polymerizations in n-heptane were conducted with various catalyst prepared at 48°C under atmospheric propylene pressure with 4Omml of AlEt, cocatalyst. Initiation rate and polymerization rate decay were investigated from time-rate profile.

174 M. Murata, A. Nakano, S. Kanazawa a n d M. lmai

Results and Discussions Figure 1 illustrates XRD patterns of 4 different catalysts. Each figure shows a) Cat - 1 ,b) Cat - 6 ,c) Cat - 7 and d) Cat - 10 ,respectively. a) (prepared from MgCl,) and b) (prepared from MgCld2-EHA) indicated the typical MgC1, crystal structure with the peaks at 15" ,30° to 35" and 50' . c) (prepared from Mg(0Et)Cl without donor) and d) (prepared from Mg(0Et)Cl with donor) have new peak at around 10' to 13" which was not observed in a) or b). It could be confirmed that all the catalysts from non-MgCl, (Cat - 7 to Cat - 10) had this new peak although catalysts from MgC1, (Cat - 1 to Cat - 6) did not have. These results clearly indicate that new crystal face of long lattice distance, which does not have in the catalyst from MgCl, , is appeared in the solid from non-MgC1,. In another word ,catalysts from non-MgCl, component probably have different crystallographic structure from so-called MgC1,. It might be Ti- Mg cocrystal like structure since the appearance of the new XRD peak is independent of existence or absence of internal donor. In order to investigate Ti atoms situation in the catalyst particle, Ti/Mg ratios of both whole (average) and surface of particle were examined. In table 1 , Ti/Mg ratio of whole catalyst solid and that of surface determined by XPS are summarized. In Cat - 1 to Cat - 4 which were prepared by milling of MgCl,, it could be seen that Ti/Mg ratio of surface was higher than that of whole , This result suggests that Ti atoms are locally concentrated on the solid. Cat - 5 and Cat - 6 prepared from MgCld2-EHA solution had higher Ti/Mg ratio of whole than that of surface ,indicating that this type of catalyst contained Ti atoms located in the particle where the species would be silent by XPS measurement. Cat - 7 to Cat - 10 from non-MgC1, components also had higher Ti/Mg value of whole than that of surface , showing similar phenomenon to the cases of MgCldZEHA. However, the major difference between MgCld2-EHA and non-MgC1, type catalysts was their crystallographic structure of whole particle as shown by XRD evaluation. Based on these results , catalyst particle structures are proposed as shown in Figure 2. In Milling type catalysts of Cat - 1 to Cat - 4 , T i atoms is deposited on MgC1,. In MgClJ2-EHA type (Cat - 5 and

16. Characterization of Mg/Ti Type Catalysts

175

Cat - 6) , small size of particles having MgC1, milling type structure are agglomerated since the particle should have MgCI, crystal structure and some Ti atoms should be exist in the solid. This structure can also be understood from the process of catalyst preparation , that is , fine MgC1, powder might be formed at the early stage of the reaction of MgCld2-EHA solution with excess TiCl, , where the very rapid reaction of 2-EHA and TiC1, may take place and fine particle of MgC1, may be precipitated as the result. In non-MgC1, type catalysts (Cat - 7 to Cat - l o ) , Ti atoms are dispersed in the particle and formation of "Ti/Mg cocrystal" like structure might be speculated. As shown here , it can be concluded that fundamental structures of catalyst are devided to two category. One is that Ti atoms dispersed on MgC1, and the other is that Ti is dispersed in the particle. These drastic change in the structure might be attribated to the Mg component which is used as the starting material for catalyst preparation. Figure 3 shown the time-rate curves of propylene polymerization with different catalyst combined with AlEt,. Here ,three catalysts prepared from different Mg compounds of MgCl,(Cat - 4) , MgC142-EHA(Cat - 6) and Mg(OEt)Cl(Cat - 10) are compared. And all catalysts contain same internal donor of DNBP. Both Cat - 4 and Cat - 6 showed rapid initiation and decay type profile. It is considered that this similarity is from equivalent catalyst structure discussed previously. In compare with this , Cat - 10 showed relatively slow initiation , suggesting slow initiation of Ti atoms located in the catalyst particle. Bottom three figures in Figure 3 shows the reciprocal rate-time curves. In all cases , linear relationships were observed for rate decay polymerizations. These indicated that 2nd order rate decay was took place. From the slope of the line decay constant (kd) values were determined and plotted against Ti/Mg ratio measured by XPS. Figure 4 shows the results. kd values were increased with increasing surface Ti/Mg ratio. This results clearly indicate that kd is controlled by the concentration of surface Ti atoms. However, the correlation of MgC1, type was different from non-MgC1, type. At the same Ti/Mg ratio of surface, non-MgCl,

176 M. Murata, A. Nakano, S. Kanazawa and M. Imai

catalyst has much lower kd value in compare with MgCl, type one. In addition to this it could be expected that active sites in the particle of non-MgCI, type catalyst which is hard to form from MgC1, might be very stable during the course of polymerization. Conclusions Several catalysts were prepared from MgCI, and non-MgCI, from some organo-Mg compounds. In the catalysts with MgCl, as the Mg source, it could be confirmed that Ti atoms were mainly located on MgC&. In contrast with this, non-MgC1, type catalysts from organo-Mg components had the structure that Ti atoms were dispersed in the particle. Changing of initiation rate and decay constant with different catalysts could be understood by consideration of the differences in catalyst structures.

car-I

ca1.4

cat.-7

cat.-10

I5 20 25 30 35 40 45 50 20 (dcgnc)

Figure 1

15 20 25 30 35 40 45 50 28 ( d e w 4

XRDresults

16. Characterization of Mg/Ti Type Catalysts

Table I

Cornparson of Ti/Mg ratios of whole catalyst and surface Elemental Analysis (whole) Ti ( ~ 1 % ) Mg ( w ~ % ) TWg

Catalyst

XPS (surface) TMg (moYmol)

(moUmd)

cat.-l cat.-2 cat.-3

MgClz MgClZ MgC12

3.2 0.3 2.5

22.3 25.4 16.7

0.072

0.10f0.01

0.006

0.012f0.02 0.22f0.03

0.076

4.3 14.5 0.15 0.16 f 0.02 cat.-4 MgClz ................................................................................................................... cat.-5

MgClZ / 2-EHA 10.0 0.34 0.24 f 0.04 14.8 cat.-6 MgClz I 2-EHA 0.11 *0.02 0.14 ................................................................................................................... 2.3 10.9 cat.-7 Mg(0Et)CI 4.5 19.2 0.18 0.16f0.01 cat.-8 Mg(0Et)Z cat.-9 n-BuMgCI ISi(OEt)4 cat.-10 Mg(0Et)CI

Figure 3

1.7 3.9 1.7

17.3 17.7 17.3

0.050

0.038f0.008

0.11

0.060f 0.01 0.030 f0.002

0.050

h p y l e n e Polymerization Rates with Different CaUbWs

177

178 M. Murata, A. Nakano, S. Kanazawa and M. Imai

50

t

Ti/Mg (surface) (mol/mol)

Figure 4 Correlation between Ti/Mg ratio of Catalyst Surface and 2nd Order Rate Decay Constant

references 1) P. C. Barbe , G . Cecchin and L. Noristi , Adr. Polym. Sci. , 81,l(1987). 2) K.Soga ,T. Shiono and Y. Doi , Makromol. chem. ,189, 1531(1988). 3) M.Terano ,T. Kataoka and T. Keii , Catalytic Olefin Polymerization T. Keii ,K.Soga(Eds.) ,Kodansha Tokyo ,1990 P55. 'I

I'

I79

17. Mechanism of the First Steps of the Isotactic Polymerization with Metallocene Catalysts

W. Kaminsky, M. Arndt University of Hamburg

1. Introduction

Since the beginning of the polymerization of propene to isotactic polymers with Ziegler-Natta catalysts it was an open question, what the stereospecifity controls. Because the prochiral olefins like propene or l-butene have no chirality, Natta proposed that the insertion of the monomers take place in a chiral structure1I2). There are two possibilities for this: First, the monomer forms at leaat after the second insertion step into a titanium hydride or titanium alkyl group a chiral carbon. This chirality can influence the next and the following insertion steps (chain end control)=. Ti

-

CH2

-

y 3 CH 1

-

CH2

-

CH2

- CH3

Second, the active center is chiral independent of the polymer chain (enantiomorphic side control). The enantiomorphic site control could be given at the surface of heterogeneous catalysts, forming si- or re-faces4). The si-enantioface is preferred because of the stereo hinderance of the methyl groups. What is mainly important for the stereo control? A lot of scientists prefer the enantiomorphic site control to be the main background for this; others find the chain end control most important. All experiments to add chiral donors to heterogeneous or supported catalysts to find an excess of an optically active oligomer, were not very successful5)

.

The situation changed when chiral metallocenes together with methylaluminoxane as catalysts were used6). It was clear now that a chiral active center is very important to give isotactic poly-

180 W.Karninsky and M. Arndt

mers. But still there were some experiments which show that also metallocenes (biscyclopentadienyltitaniumdiphenyl)can catalyze parts of isotactic polypropylenes'

.

2. Chain End Control To find out how big is the influence of the chain end control on the stereospecifity of polypropylenes catalyzed with metallocenes, titanocenes and zirconocenes with a chiral alkylligand were synthesized. 2-Methyl-butyl was used because of its chirality center in a j3-position to the transition metal. The bis(cyc1opentadieny1)titanium- or zirconium bis(2-methylbutyl compounds (1 and 2) were prepared by reaction of the biscyclopentadienyl metal dichlorids with lithium-2-methylbutyl.

cp\ CP

CH3 CHZ-CH-CHZ-CH~

/ Ti 7H3 'CHZ-CH-CHZ-CH~

'

or

These compounds were used as catalysts for the propene polymerization*) If the mechanism is right , so that these metallocenes react with MA0 forming a cationic spezies by transferring an alkyl group, one 2-methylbutyl group remains at the transition metal. The polymerization was carried out by different temperatures, a metallocene concentration of mol/l in 150 ml toluene, a propene concentration of 1,7 mol/l and a molar Al/Zr ratio of 5 x lo4' At temperatures below -20 OC only two or three polymer chains were formed by every active centre. Table 1 shows the 13C-NMR measured pentads of the resulting polypropylenes.

.

All polymers are soluble in toluene and show an atactic behavior. This means that the chain end control is not very strong, but

17. Mechanism of the First Steps of lsotactic Polymerization

TABLE 1

181

.

13C-NMR Measured Pentades of Polypropylenes Catalyst: Bis(cyclopentadieny1)zirconium bis(2-methylbutyl)

Pentade

30

OC

7

-20

OC

OC

-60

OC

-35

OC

(Ti)

0.052

0.085

0.106

0.140

0.430

0.124

0.173

0.182

0.202

0.224

0.076

0.085

0.075

0.072

0.030

0.100

0.108

0.102

0.120

0.047

0.245

0.248

0.255

0.242

0.188

0.157

0.139

0.132

0.107

0.040

0.036

0.031

0.022

0.016

0.011

0.113

0.069

0.055

0.041

0.012

0.091

0.060

0.062

0.052

0.012

there are some effects. The chain end control increases with decreasing temperature. It is much stronger by the titanium than by the zirconium compound. The pentads contain isolated r dyads which are characteristic for a chain end control. In this model it must be independent for the pentads, which alkylated metallocene compound is used, if the model of the cationic active center for the metallocene catalysts is right. The picture becomes clearer if the isotacticity, the different sequence lengths are calculated by the method of Randallg) (see Tab. 2).

Isotacticity I = [(mm) + 0,5 Isotactic sequence lenght niso = 1 : [(rr) + Syndiotactic sequence length nr = [(rr) + 0,s msequence length nm [(mm) + 0 , 5

-

(M)] x 100 0,s (M)] (mr)] : [0,5(mr)l ( = ) I : [O15(mr)l

182 W. Kaminsky and M. Arndt

TABLE 2

13C-NMR Measured Isotacticities and Iso, S y n and Racemic Sequence Length of Polypropylenes

30

OC

7

OC

-20

OC

-60

"C

-35

OC

50.38

56.15

57.34

60.28

75.33

"is0

2.02

2.28

2.36

2.53

4.09

"rn

1.50

1.69

1.74

1.88

3.17

1.48

1.32

1.29

1.23

1.13

1.1. /%

"r

(Ti)

It is interesting to note that the titanocene shows polypropylenes with a higher isotacticity at 35 O C than the zirconocene at 60 O C . The reason for it could be the shorter bond length between the carbon atom and titanium in relation to zirconium.

-

-

The result is that even starting with chiral alkyl groups on the metallocenes by very low polymerization temperatures it is impossible to prepare highly isotactic polymers.

3. Enantiomorphic Site Control Highly isotactic polypropylenes can be obtained using chiral zirconocenes and MAO. This indicates the high influence of the enantiomorphic site control. To measure the influence of chiral active centres optically active metallocenes were prepared. Two methods to separate the racemic mixture of ethylene(bistetrahydroindeny1)zirconium dichloride or dimethylsilyl(bistetrahydroindeny1)zirconium dichloride into the enantiomers are given. 1. reaction with S-binaphthol 2. reaction with o-acetyl-R-mandelate Among these reactions, diastereomers are formed which could be obtained in pure forms. In the first case the S-zirconocene forms a

17. Mechanism of the First Steps of Isotactic Polymerization

183

complex with the S-binaphthol which crystallizes while the other diastereomer remains in the solution. The o-acetyl-R-mandelacic reacts only with the S-zirconocenes. This complex could be separated from the remaining compounds (see Fig. 1).

n

n = 6-20

FIG. 1

Structural Formula of (S)-[l,l'-ethylenebis(4,5,6,7-tetrahydro-l-indenyl)]zirconiumbis(O-acetyl-(R)-mandelate) and Methylalumoxane

The oligomerization starts when an olefin undergoes insertion into a transition metal hydrogen or methyl bond formed by methylation with methylalumoxane (Scheme 1). There are mainly 1,2 and very few 2.1 insertions. Subsequent insertions lead to chain growth. Chain termination takes place by j3-hydrogen transfer to the transition metal atom or to a complex bound olefin, resulting in formation of the hydrid or alkyl transition metal compound in addition to the oligomer. The former, in furn, allows new insertion steps to occur. The formed dimers do not contain a chiral carbon atom. Optical activity is first observed in trimers and higher oligomers")

.

184 W. Kaminsky and M. Arndt

-&

R

c-; SCHEME 1

R

P

B

0,08

0947

0,05

09x9

R

Dherization of Propene (P) and 1-Butene (B). The Amount of Different Isomers is Measured by Gaschromato9raPhY

4. Propene Oligomerization

The average molecular weights of the produced oligopropenes can be controlled by adjustment of reaction temperature and monomer concentration. Ethylenebis(tetrahydroindenyl)zirconi~di(O-acetyl(R)-mandelate) was used as transition metal compound. To obtain sufficient amounts of product with a propene feed rate ‘of 2,5 to 20 ml/min the reaction time had to be extended to between 15 and 24 hours. The reaction temperature was varied in the range from 20 to 60 O C . Depending on reaction time yields ranged from 18 to 24 g of oligopropenes. For reaction temperatures of 30 OC and above the products are oily liquids whereas oligomerization at 20 OC yielded waxy products. The reaction temperature has a great effect on the average degree of oligomerization”

.

17. Mechanism of the First Steps of lsotactic Polymerization

185

The products contain oligopropenes of various degrees of oligomerization. By means of gas chromatography branched alkenes from dimers up to nonamers could be detected. Figure 2 shows the capillary gaschromatogram of a mixture of oligomers produced at 50 O C . In the various oligopropene fractions a number of by-products (isomers) are formed next to the main component. Table 3 gives the amounts of olefins differing in degree of oligomerization. The table not only shows a shift of distribution maxima with temperature but also makes it clear that the oligomerization can be conducted in a way that mainly trimers through heptamers are formed. Provided that the number of active centers be independent of temperature the amounts of lower oligomers should grow with increasing temperature. Up to a reaction temperature of 60 OC this is the case. when the oligomerization is carried out at 70 O C , however, the sume of dimers to nonamers decreases indicating a slight decrease in the number of active centers. TABLE 3

Massdistribution of Propene Oligomers at Different Temperatures

T [ O C )

30 40 50 60 70

Gew. -%

Di

Tri

Tet

Pen

Hex

Hep

OCt

Non

EDi-Non

0,l 1,3 214 4,O 315

210

311 619 1115

415

413 814 11,3 12,l 1415

4,l 113 9,4 9,4 1115

4,2 6,9 5,8 6,l 610

4,l 5,2 2,8 2,9 215

26,4 49,5 64,3 85,s 8218

512 819 14,4 1115

18,8 17,l

0,3 1212 17,2 16,2

5. Isomers

The propene oligomers synthesized with the (S)-En(IndHq)2Zr(C10H804)2/MAO catalyst predominantly consist of 1-alkenes as they are formed by lI2-insertion and isomeric by-products. In order to record all isomers of an individual degree of oligomerization, the products were analytically separated by gas chromatography over a 50 m capillary column.

186

W. Kaminsky and M. Arndt

Independent of reaction temperature and monomer concentration, the capillary gas chromatogram of the dimersfeatures one main peak corresponding to 2-methyl-1-pentene and several other peaks of lower intensity. The fraction of propene trimers, by contrast, is made up of a mixture of several isomers (Fig. 2).

Tri

Tri,

,

Tri,

Tri,

A

c

I

I

FIG. 2

d

10 12 t [min] Capillary-Gaschromatographic Separation of Propene Trimers Synthesized by 50 O C

The composition varies with reaction temperature and monomer concentration (Tab. 4). TABLE 4 Distribution of Isomers in the Trimer Fraction of the

Propene Oligomerization as a Function of Temperature and Propene Flow Rate mol/l Catalyst: (S)-En(IndH4)2Zr(CloH804)2 5 x MA0 4,2 x 10- mol a1 units/l Reaction time: 24 h; propene pressure: 2,2 bar T [OC]

Propene [ml/min]

30 40 50 50 50

10 10 215 5 10

50 50 60 70

15

20 10 10

Concentration w t . - % : Tril Tri2 Tri3 99,3 96,s 85,4 87,3 88,2 89,2 91,8 80,2 71,l

-

-

113 519

0,4 519 3,3 1,9 1,3

515

417 417 413 713 10,2

1,2

4,3 6,9

Tri4

-

Tri5

111 1,l 0,9

or7 or9 210 211 212 212 212

1,s

st2

2,3

615

-

-

XTri2-5 017

218 14,9 12,o 917 812 817 18,3 25,9

17. Mechanism of the First Steps of Isotactic Polymerization

187

At 30 O C and 10 ml/min of propene feed 99 % of the trimer fraction consists of 2,4-dimethyl-l-heptene (Tril) which is formed by 1,2insertions. A marked decrease in the relative concentration of this main component is observed at higher oligomerization temperaures as well as lower monomer feed rates. Simultaneously, other isomers that are formed by double bond migration, 2,1-, and 1,3insertion gain significance. Scheme 2 assigns structures to the individual isomers.

Tril

2,4-Dimethyl-l-heptene

Tri2

2,4-Dimethyl-2-heptene

Tri3

2,6-Dimethyl-l-heptene

Tri4

2,4,5-Trimethyl-l-hexene

+

2,4,6-Trimethyl-l-heptene SCHEME 2

Isomers of Trimeric Oligopropene

The trimer Tri2 stems from Tril via double bond migration. This reaction becomes increasingly important at higher temperatures. Tri3 is formed through an initial 2,l-insertion followed by a 1,3insertion which, in turn, results from rearrangement of another 2,l-insertion after the first one. This order of events becomes plausible when one considers that a regular 1,2-enchainment is sterically hindered after a 2,l-insertion thus favoring another 2,l-insertion. It is this steric hindrance between two adjacent methyl groups in a 2,1-1,2-sequence that is responsible for the relatively low concentration of Tri4

188

W.Kaminsky and M. Arndt

which contains an initial 2,l-enchaiment followed by two insertions with lI2-orientation. Finally, Trig is formed as Tril by three consecutive 1,2-insertions. This time, however, the initial propene unit is inserted into a Zr-methyl bond as it is formed in a reaction of the zirconocene with methylaluminoxane as opposed to a Zr-hydrogen bond resulting from the common j3-hydride transfer. The isomers Trill Tri2, and Tri5 were positively identified by NMR- and mass spectrometry. The propene tetramer contains two asymmetric carbon atoms. Therefore the synthesis with chiral metallocenes leads to the formation of diastereomers. The optical activity of the chiral oligopropenes was determined at various wavelengths. Polarimetric measurements were not only conducted with product mixtures from oligomerizations at various monomer concentrations and reaction temperatures but also with individual fractions of dimers, trimers, and tetramers. To this end the product mixtures were fractionated by distillation over a split tube column (Table 5). TABLE 5

Specific Optical Rotation [a]25 of the Trimers, Tetramers and Mixed Oligomers at Different Wavelengths and Different Reaction Temperatures and Propane Flow Rates

40 40 40 40 50 50 60 60 70 70

589 546 436 365 589 365 589

365 589 365

+ 1,7 + + + + +

2,o 3,7 6,5 0,9 3,O + 0,4 + 1,5

+

0,08

+

0114

+ 3,5 + 4,3 + 7,3 + 11,8 + 2,8 + 8,7 + 2,2 + 6,6 + 1,8 + 5,6

+ + + + + + + + + +

3,O 3,5 5,8

9,2 2,6 7,6 2,2 6,8 1,9 5,6

17. Mechanism of the First Steps of lsotactic Polymerization

189

The propen oligomers starting with the trimers are dextro rotatory. As expected, the achiral dimer does not show any optical activity. The trimer, 2,4-demethyl-l-heptenef which was produced catalyst ~ O ~ ) ~ bears / M A O S-confiwith the ( S ) - E ~ ( I ~ ~ H ~ ) ~ Z ~ ( C ~ O H guration, since a specific optical rotation [ a ] of~ -6.1 ~ ~ was determined for the R-enantiomer' )

.

With increasing reaction temperatures the specific optical rotation of all oligomers decays. This proves that the stereoselectivity of the organometallic catalyst decreases at higher temperatures. The optical activity of the tetramers is higher than that of the trimers. This increase is caused not only by the additional chiral carbon but also by an increase in stereoselectivity due to the longer alkyl chain attached to the active center. This difference is particularly significant at elevated temperatures. While the specific optical rotation of the trimer is lowered by a factor of 20 in the temperature interval1 from 40 to 70 OC, it is only reduced by one half for the tetramer. The extent of stereoselectivity in the chiral synthesis can be checked by determining the enantiomeric excess of the optically active alkenes in the products. Since no literature data was available for the optical rotation of the enantiomerically pure alkenes, their optical purity was determined through gaschromatographic resolution of enantiomers by means of an optically active column. Thermostable substituted Cyclodextrines are best suited as asymmetric phases"). The trimer, 2 ,4-dimethyl-l-hepteneI was resolved into its enantiomers by capillary gaschromatography with an octakis-(6-0-methyl-2,3-d-O-pentyl)-y-cyclodextrine phase. At low temperatures (20 "C) the formation of the first chiral center proceeds with a high selectivity of 97,6 % leading to an enantiomeric excess of 95,3 %.At higher temperatures the ee-value decreases to 23,8 % at 50 OC and 2,5 % at 70 OC. As expected, the

190

W. Karninsky and M. Arndt

ee-value of the trimer produced with the racemic catalyst is 0 (Fig. 3 ) .

-----J

i

10 12 20

95,3

FIG. 3

10 30 73,4

Cminl T ["c]

10

10

10

lo

t

40

50

60

51,3

23,8

10,9

70 2,s

ee

[%I

Asymmetric Oligomerization of Propane. Gaschromatographic Separation of 2,4-Dimethyl-2-heptene(Trimer) Using Octakis(6-o-methyl-2,3-di-O-pentyl)cyclodextrine

It is evident that at high oligomerization temperatures the isotacticity is low. At the same temperature it is higher for the tetramers. It could be calculated how high is a hypothetic eevalue from isotactic polypropylene by using the mm triads (see Scheme 3 , next page).

The results are given in Tab. 6. TABLE 6

Comparison of Measured and Calculated ee-Values of Propene Oligomers and Polymers Temperature ( ' C )

Tetramers

Polymers

50

7314 51, 3 23,8

38,l

95,s 92,8 90,o 87,s

70

2f5

510

20 30 40

Trimers 95,3

mmmm (0,972) (0,958) (0,939) (0,918) 68,O (0,802)

17. Mechanism of the First Steps of lsotactic Polymerization

191

m

2

K \ \

.

1

K

ka

A-B ee= A+B

1

ee=(2kr2-1)

SCHEME 3

1

jk,2 =-(ee+l)

2

Calculation of the ee-Value from the Isotacticity of Polypropylenes

By an oligomerization temperature of 50 O C , the trimers show an ee-value of 23,8, the tetramers of 38,1, and the polymers of 8 7 , 5 . This shows the great influence of the growing chain on the stereospecifity. In conclusion, to come to a high isotacticity, a chiral metallocene is needed. The first insertion steps show a low stereospecifity which increases with the growing polymer chain.

192 W. Kaminsky and M. Arndt

6. References

1. G. Natta, Angew. Chem. l2, 393 (1956) 2. G. Natta, P. Pino, G. Mazzanti, R. Lanzo, Chem.Ind. 39, 1032 ( 1957 ) 3. V. Venditto, G. Guerra, P. Corradini, R. FUSCO, Polymer 3 l , 530 (1990)

4. L. Cavallo, G. Guerra, L. Olive, M. Vacatello, P. Corradini, Polym.Commun. 30, 16 (1989) 5. P. Pino, P. Cioni, J. Wei, J.Am.Chem.Soc. 109, 6189 (1987) 6. W. Kaminsky, K. Kiilper, H.H. Brintzinger, F.R.W.P. Wild, Angew.Chem. 97, 507 (1985); Angew.Chem.Int.Ed.Eng1. 24, 507 (1985) 7. J.A. Ewen, J.Am.Chem.Soc. 106, 6355 (1984) 8. 0. Rabe, Dissertation Hamburg 1993 9. J.C. Randall, Polymer Sequence Determination, Academic Press, New York 1977 10. W. Kaminsky, A. Ahlers, 0. Rabe, W. K h i g , in: Organic Synthesis via Organometallics, D. Enders, H.J. Gais, W. Keim (eds.), Vieweg, Braunschweig 1993, p. 151 11. W. Kaminsky, A. Ahlers, N. Mtjller-Lindenhof, Angew.Chemie 101, 1304 (1989); Angew.Chem.Int.Ed.Eng1. 28, 1216 (1989) 12. D.E. Dorman, M. Jantelat, J.D. Roberts, J.Organomet.Chem. 36, 2757 (1971)

193

18. Reaction Mechanisms in Metallocene-Catalyzed Olefin Polymerization

H. BRINTZINGER,

S. BECK, M. LECLERC, U. STEHLING and W. ROLL

Fakultat fur Chemie, Universitat Konstanz, 0-78434 Konstanz, Germany

ABSTRACT

1. Studies by 'H NMR on equilibria between contact ion pairs such as Cp2ZrCH3d+...H3C-B(C6F5)3d-and binuclear alkyl zirconocene cations of the type (Cp,ZrCH,),b

- CH,)

+

lead to the conclusion that these binuclear species must

generally be considered as participants in all homogeneous Ziegler-Natta systems.

2. Different polypropene chain lengths, which are obtained from cis- and trans1D - propene with the catalyst en(thind)2ZrC12/MA0,show that exchange of a-H with

a-D atoms affect the rate of chain growth by a large kinetic isotope effect; this supports the notion that an a-agostic interaction facilitates the olefin insertion step. 3. A strong increase in polymer chain lengths, which is caused by the presence of amethyl groups in ansa-zirconocene catalysts, is shown, by the effects of propene pressure on ,M ,

to be due to the suppression of the otherwise predominant direct

I3 - H-transfer to a coordinated olefin molecule by these a-substituents. INTRODUCTION Open questions with regard to the mechanisms of metallocene-catalyzed olefin polymerizations concern the equilbria which lead to catalyst activation and deactivation, the factors which control the rate and stereoselectivity of the olefin insertion step, and the mechanisms of chain termination. Some recent studies related to these questions are reported here.

EXPERIMENTAL

I . Solutions of B(c~F,),

' 1 and of CP~Z~(CH,)~ in

C,D,

(10 - 40 mM) were

combined in various proportions under extreme exclusion of humidity (flamed glassware, glovebox techniques) and their 'H NMR spectra measured at room temperature on a Bruker AC 250 MHz spectrometer.

194 H . Brintzinger, S. Beck. M. Leclerc, U. Stehling and W. Roll

2. Cis- and trans-a-deuterated propene were prepared by lithiation o f cis- and trans-chlorpropene, respectively, and subsequent cleavage with D20. The samples thus obtained were purified by repeated distillation from dry MAO. Polymerizations were conducted at 5OoC with en(thir~d)~ZrCI,/MAOin toluene ([Zrl =

M, Al:Zr

= 1200:l 1 at 1 bar. The molecular weights of the polymer products were determined

from their 13C NMR spectra, run at 13OOC in CD ,C , ,I

by the ratio of n-propyl end-

group and methyl side-chain signals at 14.3 and 20.0-21.8 ppm, respectively. 3. Polymerizations were conducted with MAO-activated Me2Si(benzind)2ZrC12and Me,Si(2-Me-ben~ind)~ZrCI,([Zrl = 1.25

M, A1:Zr = 15800:1, T,

= 5OoC), at

propene pressures between 1 and 7 bar. The molecular weights of the polymer products were determined by GPC (BASF AG, Dept. ZKP).

RESULTS

1. Binuclear Cations in Metallocene-Based Zlegler-Natta Catalysts. Indications for the occurrence of binuclear cations of the type (Cp2ZrCH3),@CH3)

+

have been

reported in several instances.'-4 In the 'H-NMR spectra of reaction systems containing B(C6F5), and an excess of Cp2Zr(CH3), in C&6, we observe at room temperature t w o distinct species of this kind. Based on the chemical shifts and the relative intensities of each of their signal sets, both o f these species are undoubtedly ion pairs of composition (Cp2ZrCH3)2@-CH3)+ H&-B(C&),-;

since one of them becomes more

prominent on dilution at the expense of the other, we assign the former t o a solventseparated and the latter t o an associated ion pair consisting of a binuclear cation and a methyl borate anion (Figure 1). For the reaction described by equation 1, we determine an equilibrium constant K, and H,C-B(C6F,),-

= 1.O

f 0.2; this indicates that Cp2Zr(CH3),

+.

are equally strong Lewis bases toward the cation CpzZrCH3

Even in the presence o f excess Lewis-acid activator A, binuclear cations could be present in amounts comparable t o the contact-ion pair CP~Z~CH,~+.-H,C-Ab-,

if

excess A is capable of efficiently complexing the anion H3C - A - according t o eq. 2: 2 Cp2ZrCH,d+-H3C -Ab-

+ (Cp,ZrCH,),(p

- CH,)

+

f A-H,C

-A -

(2)

18. Mechanisms of Metallocene-Catalyzed Olefin Polymerization

6+5.4

6+0.3

6-o*1

= 6+5.6

195

6+5.7

1.020.2

6-0.1

0 + H,C-B(C,F,), (6+1.3, separated)

=

0.520.1 mM"

0 H3C-B(C6F5)3

(&+lo, associated)

(c.f. Li+ H$-B(C6F,),-

6+0.85)

Figure 1. Equilibria between contact ion pairs, excess dimethyl zirconocene and alternative binuclear zirconocene cations, with 'H NMR shift values. If binuclear cations do not contribute to chain growth, as indicated by a recent study,

4,

but still allow chain termination to occur, their presence might explain the

shortening of chain fengths associated with elevated zirconocene concentration^.^) 2. The Olefin Insertion Step. In previous studies, we have observed stereokinetic

isotope effects 61 for the hydro-oligomerization of cis- and trans-1D-1-hexene by Cp2ZrC12/MA0and en(thind),ZrCI2/MAO;') based Ziegler-Natta catalysts

these and related studies on scandocene-

support the notion that an agostic interaction of an

u - H atom of the migrating polymer chain with the metal center facilitates the olefin insertion step, as proposed by Rooney, Green and B r o ~ k h a r t . ~ ~ ' ~ )

196 H . Brintzinger, S. Beck, M. Leclerc, U . Stehling and W. Roll

We have now studied the polymerization of cis- and trans- 1D-propene with en(thind)2ZrCI/MA0, and find that the mean chain length obtained with the trans isomer, PN(trans) = 128, is about 2.8 times larger than that obtained with the cis isomer PN(cis) = 45. This indicates that the olefin insertion step is favored by a large isotope effect (k,/k,

= 2.8) when an a-H atom, rather than an a-D atom, is placed in

the agostic bridging position, as it is to be expected from consecutive insertion reactions of trans- and cis- 1D-propene, respectively (Figure 2). These results provide experimental support for recent theoretical studies on the course of the olefin insertion step in cationic metallocene catalysts.l1-l3'

0-HT

\ DHC=CMeR'

cis-1D-propene:

kD PN = 45

kD

0-HT

\ DHC=CMeR'

Figure 2. Reaction schemes for consecutive insertions of trans-1D-propen (top), which place an a-H atom in the agostic bridging position, and of cis-1D-propene, which place an a-D atom in this position.

3. Chain Termination Mechanisms. Previous metallocene-based polymerization catalysts have given much shorter polymer chain lengths than classical heterogeneous catalysts; recently however, polymers with molecular weights of several hundred thousands have become available by use of ansa-zirconocenecatalysts with a-methyl s u b ~ t i t u e n t s . ' ~ ~In' ~studies ) on the effect of propene pressure on the polymer

18. Mechanisms of Metallocene-Catalyzed Olefin Polymerization

197

molecular weights, we find the molecular weight of polypropene obained with (CH3I2Si-bridgedbis(indeny1) and bis(bedndenyl1complexes to dependent very liitle on propene concentration (Table 1). This indicatesthat the dominant chain termination process is R-H transfer to a coordinated olefin molecule, in accord with previous evidence from studies on the end-groupdistribution in ethene-propenecopolymers. Catalyst

L

IIMAO

Benzlnd

1

0.31

29 800

88

IIMAO

Benzlnd

2

0.66

35 100

88

IIMAO

Benzlnd

3

1.02

38 600

88

IIMAO

Benzlnd

7

2.43

39600

90

IIIMAO

2-MeBenzlnd

1

0.31

80500

92

IIIMAO

2-MeBenzlnd

2

0.66

137 100

92

IIIMAO

2-MeBenzlnd

3

1.02

182 200

93

II/MAO

2-MeBenzlnd

5

1.72

247 700

93

plbar

c(C3H6)

M ,

% mmmm

Table 1. Effect of propene pressure on the molecular weight of polypropene obtained with MAO-activatedMe2Si(benzindI2ZrCl2(I,top) and Me2Si(2-Me-benzind)2ZrC12(11, bottom). T, 5OOC; [Zrl 1.25*10-6mol/L; [AIl:[Zrl 15 800. With o-methyl substituted ansa-zirconocenesas catalysts, however, the molecular weight of polypropene shows a strong increase with propene pressure (Table 11, in accord with expectations for a chain termination by R-H transfer to the metal center. From a plot of PN-' versus c(C3H6)-' (Figure 31, we determine that both types of catalysts have almost identical rate constants for R-H transfer to the metal (kTM), whereas the rate constant for R-H transfer to olefin (kTo) is about ten times smaller for the complex with a-methyl substituents. These substituents thus appear to interfere with the transition state for R-H transfer to a coordinated olefin (Figure 41, which appears to be sterically rather demanding, as indicated by a relatively large lateral extension angle of more than 180". 'I

198 H. Brintzinger, S. Beck, M. Leclerc, U. Stehling and W. Roll

2.50

2.00

1.50

a \ 0

z

0

1.00

0.50

0.00

0.50

1.00 1

/

1.50 c(C,H,)

2.00

2.50

3.00

3.50

[Vmoll

Figure 3. Plot of PN-' vs. c ( C ~ H , ) - ~for Me2Si(benzind)2ZrC12(I, top) and

Me2Si(2-Me-benzind)2ZrC12(11, bottom). PN-' = c(C,H,)-~ *(kTM/kp) k,,/k,

+ kTo/kp gives

as the slope and kTo/k, as the abscissa intercept of each graph.

Figure 4. Model of the reaction complex for 13-H transfer to a coordinated olefin; (I-

methyl groups (shaded) interfere with the formation of this reaction complex.

18. Mechanisms of Metallocene-Catalyzed Olefin Polymerization

199

ACKNOWLEDGEMENTS Financial support of this work by the VW Foundation and BMFT is gratefully acknowledged.

REFERENCES 1.

X.Yang, C.L.Stern and T.J.Marks, Angew. Chem., Int. Ed. Engl. JQ4,1406 (19911, Organometallics N, 840 (1991).

2.

G.G.Hlatky and H.W.Turner, quoted in ref. 4.

3.

M.Bochmann and S.J.Lancaster, J. Organomet. Chem.,

a, C1 (1992);

M.Bochmann, results reported a t symposium "40 Years Ziegler-Natta Catalysts", Freiburg (1993).

14,91 (1993).

4.

N.Herfert and G.Fink, Makroml. Chem. Rapid Commun.,

5.

W.Kaminsky, M.Miri, H.Sinn and R.Woldt, Makromol. Chem. Rapid Commun.

4, 41 7 (1983); W.Kaminsky, K.Kulper and SNiedoba, Makromol. Chem. Macromol. Symp.

3,377 (1986); W.Kaminsky, A.Bark

and R-Steiger, J. Mol.

Catal., 74, 109 (1992). 6.

L.Clawson, J.Soto, S.L. Buchwald, M.L.Steigerwald and R.H.Grubbs, J. Am. Chem. SOC.,1pz, 3377 (1985).

7.

H.Krauledat and H.H.Brintzinger, Angew. Chem., Int. Ed. Engl.,

B,1412

(1990); M.Leclerc and H.H. Brintzinger, in preparation. 8.

W.E.Piers and J.E. Bercaw, J. Am. Chem. SOC., 1 1 2 , 9 4 0 6 (1990); D.W.Cotter and J.E.Bercaw, J. Organomet. Chem. 417, C1 (19911.

9.

D.T.Laverty and J.J.Rooney, J. Chem. SOC., Faraday Trans. L9, 869 (1983).

10.

M.Brookhart and M.L.H.Green, J. Organomet. Chem.,

m, 395

M.Brookhart, M.L.H.Green and L.Wong, Prog. Inorg. Chem.,

(1983);

X,1 (1986).

11.

M.Prosenc, C.Janiak and H.H. Brintzinger, Organometallics, 11,4036 (1992).

12.

H.Kawamura-Kuribayashi, N.Koga and K.Morokuma, J. Am. Chem. SOC., 114 8687 (1992).

n,432 (1994).

13.

T.K.Woo, L.Fan and T.Ziegler, Organometalllics

14.

WSpaleck, M.Antberg, J.Rohrmann, A.Winter, B-Bachmann, P.Kiprof, J.Behm and W.A.Herrmann, Angew. Chem. Int. Ed. Engl.,

a,1347

(1992);

J.Rohrmann, V.Dolle, A.Winter and F.Kuber, Eur. Pat. Appl. 921 20287.5 (1991 1.

200

15.

H . Brintzinger, S . Beck, M. Leclerc, U . Stehling and W. Roll

E.Karl, W.RtiII, H.H.Brintzinger, B.Rieger and U-Stehling, Eur. Pat. Appl. 92108888.6 ( 1991 1.

m, 428 (1989).

16.

T.Tsutsui, A.Mizuno and N-Kashiwa, Polymer,

17.

P.Burger, K.Hortmann and H.H.Brintzinger, Makromol. Chem., Macromol. Symp., 66, 127 (1993).

20 I

19. Role of Ions in Coordination Polymerization of Olefins

F. S.Dyachkovskii Institute of Chemical Physics Russian Academy of Sciences, Chemogolovka, 142432, Moscow Region, Russia. ABSTRACT Studying of homogeneous csystem Cp2TiC12AlR2Cl it was shown that ions as Cp2TiRf play important role in the formation of active centers. More over by means of electroddysis the composition of ionic active centers have been determiaed for difFerent catalyhc systems. Using mass-spectrum technique the interaction of Cp2TiCH3' ions with ethylene in gas-phase was shown. Quantum-chemical investigation of Ti-C bond leads us to conclusion that deficiency of electron density on the titanium atom result m the deformation of Ti-C bond energy curve and decreasing of activation energq of insertion reaction of olefin mto Ti-C bond The ionic nature of active centers in Zr-cene catalytic systems will be discussed. INTRODUCTION It is well known that free ions are very reactive to unsaturated compound. In the gasphase ions react at very collision with saturated and unsaturated hydrocarbons. In solution the reactivity of ions is decreased due to sohation. But in the non polar solventa the reactivity of ions could be very lugh and even at low concentration their role might be important. Cationic and anionic polymerization processe% m hychcarbon s o h t are well studied, Depending on stabilization of gegenim and nature of solvent a free ions or ion-@ are formed. But role of ions in coordination polymerization of olefins needs more dew investigation. RESULTS AND DISCUSSION In the last years ions structure of active centers m homogeneous catalytic systems based on Ti and Zr is discussed in the litenrture very much. It is believed that active centers of Zr-cene catalyhc systems with MA0 have an ionic nature. In present paper the mechanism of homogeneous coordination p o l y m d o n , the reactivity of Ti€ bond, kinetic of macromolecule formation dependingof polarity of the catalytic complex will be discussed.

202

F.S. Dyachkovskii

Investigation of kinetic and mechanism of olefin polymerization in the presence of complex Cj~TiC12.AlR2Cl(A) showed, that not complex A itself, but particles in equilibrium with the complex are catalyhc active [l]. The rate of ethylene polymerization in the presence of complex A was very much depend on the small amount of impurities in the s o h t and was proportional to the square root of A concentration. It indicated that complex is dissociate for two particles. It was suggested that there are positive ions containing Ti in equilibrium with the complex A, the former bemg in fact ache m polyme&ation. Cp2TiC12 + AlR2Cl= Cp2TiRCl.AlRC13 Cp2TiRCl.AlRCl3 = Cp2TiR+ t AlRCl3Cp2TiR+ + ethylene = polymerization In agreement with this assumption the rate of the reaction of complex A with olefins was found to be proportional conductivity of the solvent and therefore was strongly depended on the nature of the solvent. For instance, it was considerably higher for benzene than for than for benzene. heptane, and for ch-le It was shown that there had been a CoIIVersion of the Ti-CH3 group into a Ti-C3H7 during the reaction of active complex with ethylene: Cp2TiCH3Cl.AlCH3C12 + C2H4 = Cp271'iC3H$L41CH3C12 The kinetic curves of hies reaction in Werend s o h t s is demonstrated on fig. 1.

t, min

Fig. 1. Kinetic cu~ve8for complex Ti-C3H7 formation and decomposition in heptane (l), benzene (2) and chlorobenzme (3).

19. Role of Ions in Coordination Olefin Polymerization

203

The direct proof of the positive charge of catalytically active ions containing titanium was achieved in a study of electroddysis of solution of the complex A [2]. Dichloroethane was used as solvent After complex Cp2TiCH3ClAlCH3C12 was subjected to electrodialysis, the number of ions containing titanium and passing to the cathode chamber was approximately 0.5 of all the ions moving through the membrane to another sides. Hence the majority of positive ions in the solution of complex A WM certain to contain a titanium atom. Titanium was not detected in the anode chamber after electroddym. In electroddysis of the complex Cp2Ti14CH3ClA114CH3C12 (€3) the titanium atoms were found to enter the cathode chamber together with the 14CH3 labeled g r o q ~and in quantities approximating the quantity of titanium. It may be concluded from the results that the catalytically active complex A dissociates on ions in solution according to the scheme Cp2TiCH3ClAlCH3C12 = Cp2TiCH3+ + AlCH3Cl3The study of ionic composition of the Ti% +Al(CH3)2Cl catalyst system in dichloroethane showed that only positive ions contain the titanium atom m this case. It is connected with the dissociation of complex CH3TiC12.AlCH3C13, on ions. The electroddysis method in the Ti(OR)4+AlR3 catalyst system confirmed the existence of the (ROhTi+, (C~HS)A~(OR~T~+ C ~[(C2H+4l(ORh]2Ti+ H~, and complex ions, depending on the AVTi ratio in the initial cataipt system. To make clear the role of ions in polymerization, the solution of the side chambers was enriched with the monomer. It the electroddysia of complexes A and B ethylene polymerization was observed only in the cathode chamber. The quantity of the polymer increased simultaneously with the increase in Cp2TiCH3+ ion mcentration m the cathode chamber. When complex B contained the14CHg labeled groups, the polyethylene obtained was radioactive. This shows that the polymer chain f d o n proceeds by the inmiion of ethylene molecules into Ti€ bond right in the ion Cp2TiCH3' We started to mvestigate the electroconductivity of Cp2ZrC12.AlR2Cl solubion m CH2Cl2. It was shown that conductivity of the complex solution at least 10 h e s higher than additive conductivity of its components. The total conductivity of complex sohtion L is m a good correlationwith the equation A-A m - k C1'2, where C is the concentrationof Zr complex. The fonnation of h e ions frm the complex CpzTiCH3Cl AlCH3C12 m gas-phase by mass-spectrum technique has been studied. For this purpose complex Cp2TiC12 ALMe2C1 was placed into the chamber of maas-spectrometer. The ions formation was detected. If ethylene

-

204

F.S. Dyachkovskii

Ions Calculated 1

~pt+i~12 fC2H4

2

cp2i'icl

3

CmTi

4

+

+

tC2H4 +C$4

CpTiCl K2H4

5

Cp2kH3 +C2H4

I

247.964 275.994 212.995 241.026 178.026 206.058 147.956 175.987 193.037 221.068

Mass Found

A

247.967

0.003

212.999 240.992 178.032 206.098 147.958 176.014 193.040 221.057

0.004 0.034 0.006 0.04

0.002 0.027 0.003 0.011

--

-

* This results were obtained together with Dr.Ueno in Kyoto University. From the figures we can draw the conclusion that different positive ions are formed including [Cp2TiCH3]+ during evaporation and ionivlhion of complex A. There is a good correlation between calculated mass and observed one. In the presence of ethylene no mass change in ion Cp2TiCl2+ was observed. On contrary, mass of ions as Cp2TiCI+, Cp2Ti+, CpTiC1' and CmTiCH3? was increased by the mass of ethylene molecular. The best correlation between calculated and observed mass was found for reaction of Cp2TiCH3+ ion with ethylene what confirmed the reaction Cp2TiCH3+ + C2H4 -> Cp2TiC+17t. It should be noted that last four ions 2-5 have the free coordination site, but Cp2TiC12' in tetrahedron structure has hot.results showed that ions [Cp2TiRIS type interact with ethylene in gas-phase and h e coordination site is important for that interaction. The discovery of Zr-cene/MAO catalysts in early 1980's generated renewed interest in our proposal that C n = + ions are active species m soluble c a m system. The ion structure of active centers [ C m MRLltX' type was shown due to extensive studies by Eischp], Jordar1(4], Boc-51, Zambelli[6], Marks[7] et.al. A key feature of these type of active centers is the vacant coordination site. The organic group R of t h cation ~ is incorporated into polymer chain. Cations exist m equilibriumof contact and sohrent-separated ion pairs.

19. Role o f Ions in Coordination Olefin Polymerization

205

Studylng of Zr-cene/MAO catalyst Fink[l] showed that with increasing dielectric constant of the solvent mixture the propylene polymerization rate increases linearly, but the stereospesiiity of the catalyst decrease strongly. Hence, the .sterecmpea&y of this catalyst system is connected with the existence of a polarized Zr-cT-AI complex or a tight contact ion pair with a stereo regulaung role of counter ion. So, ionic nature of active Bite can be extensivety modified, allowing tuning of steric, chirality and electronic properties. Unfortunately one important point was not discussed in literaape much. That is the influence of poSitiVe charge on the reactivity of M-carbon bond. That quesfion waa examined in 19-10]. The catalytic activity would obviously be conditioned by the reactivity of metal-carbon bond coupled with the existence of coordination site. The high transition metal carbon bond reactivity is caused by its liability and easy deformation. The behavior of potential M-C bond curve of transition metals can be compared with of main group of metas. Potential curves of Ti-C bond in CH3TiCl3 and AI-C bond in AlCH3C12 calculated with Hartree-Fock method are given in fig.2 [9].

-

-960.

-1340

e.SV

cv

@I

-962.

-1341

-964. -1344

-966-1346

-968-1348

!a

-970,

25

3.0

35

4.0

45

-

,

.

, . , -

5.0

.

-

.

-

.

-

.

.

5

F w 2. h t i a t i o n energy for T i c and AI-C bonds. Comparison of potential curves shorn that a more -cant deformation occurs m the Ti-C bond at the same energy of bond excitation. It is umnected with the appearance of "triplet instability" that occurs well before in transition metal derivatives. The pair of electrons, which forms the M-C bond, is localized in d-, p and s-orbitals at the equih'brium distance. Partial unpairing of electrons between Ti and C atoms occurs at a r e h t k b d hctsase of distance and unpaired electron is futhl localized in d-A0 of titanium. By the following streching of the bond the spin density in these orbitals increase up to the complete transition of the electron in the Ti@) fragment formed. The low dif€usionof d-A0 of Ti in mmpatkn to s- and p A O is ofgreat importance. Bond stretchmg c a d a rapid decrease of d-A0 overlapping with carbon

206

F S. Dyachkovskii

orbitals and thus the appearance of "triplet instability" (the rearrangement of valence state toTi(m) &agment) already at a relatively small degree of bond stretchmg. The main group metal (aluminum, for example) does not change practically its valence configuration on bond stretchmg for quite long distance.So, the transition metal is able to easily rearrange its electron structure which leads to the high reactivity of the M-C bond, compare with the main group with anti metal bond. The energy transition into the lower triplet electron-excitedstate AE,,t, bondmg properties relative to the M-C- bond, was used as the relative parameter for the M-C bond reactivity [9,10]. The small value of AE,t indicates the decrease of energetic barrim of reaction m wordination metal sphere and decrease of total energy of the M-C bond. AE,t was quantitatively calculated and thus the influence of various factors (structure, polarization and charge) on the reactivity of M-C bond considered. Calculated AE,t for different configuration and polarity of catalyst complexes are given in Table 2.

Table 2 Characteristics of T i w ) derivatives. Compound

Transition energy (eV) E , t

2.6 2.7 2.2 2.3 1.8

1.4 1.7 2.5 1.9

2.7 1.9

We can see that appearance a positive charge near the titanium atom sharply increase the ability of the transition metal for rearrangement of electron structure, thereby decreasing Es4 So, the positive chasged ions IL,MR]+type and polar complex L M R .X-] could exhibit lugh activity m polymerizatiOn reaction due to a slower increase of potential energy curve of M-C

'

19. Role of Ions in Coordination Olefin Polymerization

207

bond and to a decrease of the activation barrier in the insertion and analogous reaction proceedmg through cyclic transition state. The other possible way for the appearance of charge, as you can see Erom the table, can be "protonization" of complexes. The calculation show, that interaction of a proton with chlorine complexes of titanium leads, almost in all caw, to a decrease of A Es -t . In the presence of strong proton acceptors in the M-complexes the equilikum quantity of protonated structures may be considerable. In these cases the favourable deformaton of M-C bond energy can be occur. Takmg the role of complex "protonation"into account, the mechanism of activation of titanium complex fixed on MgO, alumoxanes, ahnnosilicates was explain [1I]. The estimation of the energy n&,t of protom& iimw of surface complexes shows that they can be considered as the active centers of catalysts on supports. It should be noted that recently [12] ethylene insertion into the Ti-C bond in positively charged ion CH3TiCL3+ was curry out. These calculation showed that metal in transition state remenge its electron structure very much,actually change its d e n t state. This changes leads to a decrease of activation energy of insertion reaction h11-14 to 4 kcaVmole. So, it is confirm the main idea of our model [9, lo]. The role of ions in the kinetic of macrmolecule formation is essential. The propagatitm rate on the ions and on the complexes should be different and it means that M-polymerbond couldbeinactiveorinactivefom: "p + X * "p.X active form "p + Lm --t"p+l + Km -* nptl+k inactiveform

+x Jr -x

+x -x

"pX

"p+P

-

+xit -x

wl+k+.X

So, there is a time when potymer chain grows ('gr) and when it is "sleeping" (hi). The ratio

@b+bl could be very small, dependmg on the condition and catalyBt system The calculation shows that Mw and MWD are vcry Sensitjve to that ratio [13,14]. Hence, the equilibrium between active and inactive form (ions, ion-pair, neutral complex) should be taken into account at the consideration of kinetic of polynerization processes, calculation of the active centers concentration and rate of propagation. In conclusion we can generalize the role of ions m coordination polymerization as following: 1. Ion gives a fiee coordination site on the metal. 2. Positive charge on the metal makes favourable deformation of M C bond. 3. By the nature of solvent possible to change the stereo regularity of active site. 4. Acceptors and donors of electrons change the reaciivity of M-C bond. 5 . Ions equilibrium allowed to regulate M, and MWD of polymer.

208 F.S. Dyachkovskii

REFERENCES 1. F.S.Dychkovsku, A.K.Shilova, and A.E.Shilov, J. Polym. Sci., Part .C, No 16,pp.23332339 (1967). 2. F.S.Dyachkovsku, E.A.Cingoryan, and O.N.Babkina, International J. of Chem. Kinetics, Vol. 13,603-613 (1981). 3. J.J.Eisch, A.M.Piotrowski, S.K. Brownstein, E.J. Gabe, F.L.L,ee, J. Am. Chem. SOC., 107, 7219, (1985). 4. R.F.Jordan, P.K.Bradley, R.E. Lapointe, and D.F.Taylor, New J. Chem., 14. 505511,(1990). 5. M.Bochmann, L.M. Wilson, R.L.Short, Organmetallics, 6, 2556, (1 987). 6. CPellecchb, A.Grassi and A.Zambelli, J. of Molec. Catal ., 82, 57-65 (1993). 7. G. Jeske, H. Lauke, KMammam, P.N.Sweptson, H.Schumann, T.J.Marks., J. Am. Chem Soc., 107,8091 (1985). 8. G.Fink and N.Herfert, International Symp. on Advances in olefin, cycloolek and dioleh polymerization. Lyon, France, Apd , 1992, p.15. 9. V.E.Lvovsky, E.A.Fushman, and F.S.Dyachkmk& J. Molec. Catal., 10, 43-56 (1981) 10. V.E.LVOV&Y, E.A.Fuehman and F.S.Dyachk~~d&, Zh.FiZ. khim., 56, NO.8, 18641878 (1982). 11. V.E.Lvovsky, A.A.Ba& and S.S.Ivanchev,Symposium on Catalysis, Novosibirsk, USSR, 183, (1982). 12. K.Morokuma, Chtm. rev., 91, 823 (1991). 13. E.A.Grigoryan, F.S.Dyachkmk~i, and A.E.Shilov, Kinetics and Mechanism of PoIyreaction, Budapest, Hungary, 11,239-241 (1969). 14. B.I,.Erusalimclkii, S.Cr.Lybezkiiin "Process of ionic polymerization" Chemishy, Leningrad, 1974, pp 33-34.

209

20. Copolymerization of Hydrocarbon Monomers in the Presence of CpTiC1, - M A 0 : Some Information on the Reaction Mechanism from Kinetic Data and Model Compounds

ADOLFO ZAMF3ELLI and m

Diparthento d i Fisica, (SA),

Z

F

m

U n i v e r s i t d di S a l e r n o ,

I-84081

Baronissi

Italy

ABSTRACT The title half-metallocene catalyst is active in the polymerization of olefins, styrenes, and conjugated dienes. An insight into the polymerization mechanism emerges from kinetic data concerning homo- and copolymerization of some of the above monomers, as well as from the structure of some novel cationic zirconium complexes. INTRODUCTION Monocyclopentadienyl titanium derivatives, such as CpTiX3 and CpTiX2 (Cp = V5-C5H5, X = C1 o r hydrocarbyl), after reaction with methylalumoxane (MAO), afford very efficient and versatile homogeneous catalysts that promote polymerization of ethylene and a-olefins,l) polymerization of styrene and substituted styrenes to highly syndiotactic polymers,*) stereospecific polymerization of 1,3-dienes to either c i s - 1 , 4 or 1 , 2 syndiotactic polymers, depending on the particular monomer and the reaction conditions.3, In this paper we will discuss some kinetic data obtained in our laboratory concerning the homopolymerization and the binary copolymerization of some of the above mentioned monomers in the presence of the catalytic system CpTiC13-MAO. Some unexpected results concerning the relative reactivities of different monomers in homo- and copolymerization will be tentatively explained by taking into account the wide spectrum of possible coordination modes and strengths of both the monomers and the growing chain ends.

210

A. Zambelli and C. Pellecchia

The p e r f o r m a n c e s of T i - b a s e d

homogeneous c a t a l y s t s w i l l be

a l s o compared t o t h o s e of s i m i l a r c a t a l y s t s based o n Z r , a n d d i s c u s s e d c o n s i d e r i n g t h e s t r u c t u r e a n d t h e r e a c t i v i t y of some c a t i o n i c organozirconium complexes s y n t h e s i z e d i n o u r l a b o r a t o r y . RESULTS AND DISCUSSION on a n d c The

o

w in

reported

data

o ofn stvrenes.

~

Table

concerning

1,

the

of s t y r e n e , pm e t h y l s t y r e n e , a n d p - c h l o r o s t y r e n e , show t h a t s u b s t i t u t i o n of t h e aromatic r i n g of t h e monomer w i t h a n e l e c t r o n - r e l e a s i n g CH3 g r o u p homopolymerizations

in

comparable

conditions

t o a n i n c r e a s e of t h e p o l y m e r i z a t i o n r a t e , w h i l e a n electron-withdrawing C 1 s u b s t i t u e n t produces t h e opposite effect. leads

Table 1.

R e l a t i v e r e a c t i v i t i e s i n homopolymerization o f s t y r e n e

and s u b s t i t u t e d styrenesa) Monomer

Time i n h

Yield i n g

Relative reactivities

styrene p-Me - s t y r e n e

0.1 0.1

0.57

1

0.75

1.3

p-C1-styrene

18

0.09

0.001

a)

P o l y m e r i z a t i o n c o n d i t i o n s : t o l u e n e , 10 mL; CpTiClg, 3 p o l ; MAO,

3 mmol; monomer, 35 mmol,;

temperature, 20 "C. Data f r o m Ref. 2c.

The r e a c t i v i t y r a t i o s f o u n d f o r b i n a r y c o p o l y m e r i z a t i o n s of s t y r e n e w i t h p-methylstyrene and w i t h p - c h l o r o s t y r e n e ,

reported i n

Table 2, a n d d e f i n e d a c c o r d i n g t o t h e scheme o f L e w i s a n d Mayo:

C*-M1

...

C*-M1..

.

+

+

Mi M2

kll

k12

C*-(M1)2

...

C*-M2(M1).

..

rl

=

kll k12

(where C*-Mi i s a c a t a l y t i c complex bound t o a g r o w i n g c h a i n e n d i n g w i t h monomer i ) show t h a t t h e r e a c t i v i t y of t h e monomers toward a n y g r o w i n g c h a i n e n d is p - m e t h y l s t y r e n e

> styrene > p-chlorostyrene.

20. Copolymerization with CpTiC1,-MA0 Catalyst 21 I

The same order for the reactivity of these monomers had been found by Natta et a1.4) in the presence of isotactic-specific catalysts, and was interpreted by assuming that the rate determining step is an electrophilic attack of the monomer by an electron-deficient active species. Table 2. Reactivity ratios for binary copolymerizations of styrene with substituted styrenes2=) Comonomer (Mp) p-Me-styrene p-C1-styrene

rl

r2

rl'r2

0.49 20

1.5 0.37

0.74 7.4

It is currently believed that the syndiotactic-specific active species are group 4 metal cationic complexes (see below) and this very fact provides a simple rationale for the observed values of the reactivity ratios. S t y r e n e and c o n i w a L & d The CpTiC13-MAO catalyst also promotes cis-1,4 polymerization

of 1,3-butadiene and isoprene, and 1,2 syndiotactic polymerization of 4-methyl-l,3-pentadiene . 3 ) Porri et a1 . 5 ) have recently found an unusual behaviour in the polymerization of (Z)-1,3-pentadiene with the same catalyst: a prevailingly isotactic cis-1,4-poly(l,3pentadiene) is obtained at room temperature, while a syndiotactic 1,2-polymer is obtained at -28 OC. Moreover, the polymerization rate seems to be higher at low temperature. A comparison of the homopolymerization rates of some conjugated dienes and styrene, in comparable reaction conditions, is displayed in Table 3.6) One can see that the reactivity in homopolymerization increases in the order isoprene << styrene < 1,3-butadiene << 4-methyl-1,3-pentadiene. This is reasonably due to the different nucleophilicities of the monomers in the coordination step, as well as to the different reactivities of both the monomers and the growing chain ends in the insertion step. The most impressive feature, in this respect, is the very low polymerization rate of isoprene and the very large reactivity of 4-methyl-1,3-pentadiene.

212 A . Zambelli and C. Pellecchia

Table 3. Relative reactivities in homopolymerization of styrene and conjugated dienesa) Monomer styrene butadiene isoprene 4 -Me-PDb) a)

time in min

yield in mg

15

190

4 1000 3

96 60 45

relative reactivities 1 1.9 0.005 770

Polymerization conditions: toluene, 13 mL;

CpTiClg, 1 0 pmol;

10 mmol; monomer, 22 mmol,; temperature, 18 'C. b)4-methyl1,3-pentadiene; in this run were used: CpTiClg, 0.025 p o l ; MAO,

MAO,

3.5 mmol; monomer, 1 7 mmol

.

Data from Ref. 66.

The results of binary copolymerizations of styrene, isoprene and butadiene (see Table 4) provide some additional information concerning the polymerization mechanism. 6 , Table 4 . Reactivity ratios for binary copolymerizations of styrene, isoprene, and butadiene6) Monomer (MI)

Comonomer (Mz)

styrene styrene but adiene

but adiene isoprene isoprene

r1

r2

rl'rz

0.14 0.35 4.7

11.5 6.3 0.31

2.3

1.6 1.4

The values of the reactivity ratios, determined according to the scheme of Lewis and Mayo, are in agreement with an almost random copolymerization of these monomers, thus confirming the close relationship between the polymerization mechanisms of styrene and conjugated dienes. However, some data are untrivial: the lower reactivity of isoprene in comparison with styrene in homopolymerization apparently contrasts with the higher reactivity of isoprene in copolymerization with the same monomer, although the presence of an even small amount of isoprene leads to a strong decrease of the yield.6 , In addition, notwithstanding that

20. Copolymerization with CpTiCI,-MA0 Catalyst

213

i s more r e a c t i v e t h a n s t y r e n e i n b o t h homo- a n d t h e y i e l d s of a series of copolymerizations performed with a c o n s t a n t c o n c e n t r a t i o n o f s t y r e n e and i n c r e a s i n g c o n c e n t r a t i o n o f b u t a d i e n e show a minimum f o r [CqHg] > 0 (see butadiene

copolymerization,

Table 5)

.

Table 5. Effect of t h e addition of butadiene c o p o l y m e r i z a t i o n s a t c o n s t a n t c o n c e n t r a t i o n o f styrenes) Styrene

butadiene

productivity

(mol/l)

(mol/l)

(g/mmol T i - h )

3.32

0

117

3.22 3.26

0.10 0.21

14.4

3.21

0.39

17.1

in

15.3

a)Polymerization conditions: toluene, 6.5 mL; C p T I C 1 3 , 5 M o l ; MAO, 4.5 mmol; temperature, 18 "C; time, 90 min. Data from Ref. 6b. Assuming t h a t t h e c o n c e n t r a t i o n of t h e a c t i v e species d e r i v i n g from a f i x e d amount o f CpTiC13 a n d MA0 i s n o t a f f e c t e d by t h e p r e s e n c e o f d i f f e r e n t monomers, one c o u l d c o n c l u d e t h a t t h e low homopolymerization r a t e o f i s o p r e n e i n comparison w i t h s t y r e n e i s o n l y d u e t o t h e lower r e a c t i v i t y o f t h e growing c h a i n s e n d i n g with an isoprene u n i t , while t h e higher r e a c t i v i t y of isoprene i n c o p o l y m e r i z a t i o n c o u l d be d u e t o a f a v o u r a b l e c o m p e t i t i o n w i t h s t y r e n e i n t h e c o o r d i n a t i o n s t e p . The k i n e t i c d a t a c o n c e r n i n g homo- a n d c o p o l y m e r i z a t i o n o f b u t a d i e n e s u g g e s t t h a t t h e growing c h a i n s e n d i n g w i t h a b u t a d i e n e u n i t a r e less r e a c t i v e t h a n t h o s e e n d i n g w i t h a s t y r e n e u n i t , a n d more r e a c t i v e t h a n t h o s e e n d i n g with a isoprene u n i t . I t i s widely accepted7) t h a t c a t a l y t i c polymerization of c o n j u g a t e d d i e n e s i n v o l v e s q3 c o o r d i n a t i o n o f t h e growing c h a i n end t o t h e c a t a l y t i c s i t e s

(see F i g u r e 1). F o r s y n d i o s p e c i f i c p o l y m e r i z a t i o n of s t y r e n e ( w h i c h p r o c e e d s t h r o u g h s e c o n d a r y w e proposed an analogous c o o r d i n a t i o n i n s e r t i o n o f t h e monomer) o f t h e growing c h a i n e n d , i n v o l v i n g t h e a r o m a t i c r i n g . 2 c ) T h i s h y p o t h e s i s i s now s u p p o r t e d by t h e s t r u c t u r e o f some c a t i o n i c

214 A. Zambelli and C. Pellecchia

zirconium benzyl complexes, showing very strong qn coordinations (see below).

CH2 \\

/

CH

I

I

I

CH2

CH2

I

CH2 I I

I I

I I

Figure 1. Schematic representation of the propagating species resulting from the insertion of butadiene (A), isoprene (B), 4methylpentadiene (C), and styrene (D) In A-C the growing chain ends are 113 coordinated, while in D involvement of the aromatic ring can lead t o different q n coordinations (see below). Of q 1 could be course, equilibria o f the type, e. g., q 3

.

-

involved in the polymerization.

An q 4 monomer

coordination

has

been

postulated

as

a

preliminary step in the polymerization of butadiene and isoprene.7) This could account for the favourable competition of these monomers with styrene in copolymerization. The particularly large reactivity of 4-methyl-1,3-pentadiene is probably due to the higher nucleophilicity of this monomer and possibly, as suggested by P ~ r r i , ~ ,to ~ )an q2 coordination to the catalytic complexes, followed by rapid insertion. The products of the reactivity ratios for styrene-butadiene, styrene-isoprene, and butadiene-isoprene copolymerizations are somewhat larger than 1. This deviation from the perfectly random comonomer sequence distribution could originate from the fact that the electrophilicity of Ti of the catalytic species (see Figure 1) is affected by the mode of coordination of the last unit of the growing chain end, and, in turn, could influence the coordination of the next monomer unit. In fact, the selectivity in attacking nucleophiles of different strength should increase while the strength of the electrophile decreases; consequently, a perfectly

20. Copolymerization with CpTiC1,-MA0 Catalyst 2 I5

random comonomer distribution (r1.r~= 1) cannot be expected when the last unit of the growing chain strongly influences the electrophilicity of the active species. It is worth mentioning that CpTiC13-MAO promotes a substantially block copolymerization of styrene and ethylene . g ) This finding could be accounted for in the framework of the above mechanism. U D 4- C

The active species involved in the CpTiC13-MA0

catalytic

system, as well as in closely related systems based on Ti and Zr compounds not carrying q5 ligands, such as TiBz4, ZrBz4, Ti(OR)4 (Bz = benzyl),lO) were suggested to be cationic complexes analogous to those involved in metallocene-based catalysts,11) on the basis of several indirect evidences .2c) The ionic structure was denied by Chien et al. on the basis of electrodialysis experiments .12) However, strong support to the former hypothesis came from the finding that catalysts performing similarly to the MAO-based systems can be obtained from monocyclopentadienyl or Cp-free hydrocarbyls, according to one of the following reactions:I3)

R=Me; n = l M

=

Ti, Zr

Cp' R

=

Bz; n

=

=

C5H5

(=

Cp) or C5Me5

(=

Cp*)

0, 1

In particular, Cp*TiR3 with B (CgF5)3 affords an extremely active catalyst for syndiotactic-specific polymerization of styrene, rivalling with C P T I C ~ ~ - M A O . ~ ~ ) Although the formation of [Cp*TiR2]+ cationic complexes has been clearly detected by NMR,15) clean isolation and full characterization of these species have not yet been accomplished, due to concurrent decomposition, possibly leading to Ti(II1) species . 1 6 )

216 A. Zarnbelli and C. Pellecchia

On t h e c o n t r a r y , w e s u c c e e d e d i n i s o l a t i n g several a n a l o g o u s Z r c o m p l e x e s , some o f which h a v e b e e n s t r u c t u r a l l y c h a r a c t e r i z e d . The X-ray c r y s t a l s t r u c t u r e s of [ ~ r ~ z 3 ][ B+ Z B ( C ~ F ~ ) ~(] 1- 1 ~ 1 7

[CpZrBz21t [BZB(C6F5)3]- ( 2 ) ,I8 a n d [Cp*ZrBz2]+ [BzB(C6F5)3]- (3)19) are r e p o r t e d i n F i g u r e s 2 and 3 .

Figure

X-ray

2.

( l e f t ) and

Of

crystal

s t r u c t u r e s of

[ZrBz3]+[BzB ( C g F g ) 3 ] - 1 [CpZrBz~l+[BzB(CgFg)3]-2 ( r i g h t ) . 1 7 t r 8 )

Compounds 1-3 are c a t a l y t i c a l l y a c t i v e i n t h e p o l y m e r i z a t i o n e t h y l e n e a t 2 5 O C a n d 1 atm, w h i l e t h e y are n o t a c t i v e i n s y n d i o s p e c i f i c p o l y m e r i z a t i o n o f s t y r e n e . R e a c t i o n of 1 a n d 3 with propene and h i g h e r a - o l e f i n s i n m i l d c o n d i t i o n s affords of

s i n g l e i n s e r t i o n a d d u c t s , 2 0 ) o n e of which,

[Cp*Zr (CH2CHMeCH2Ph)B z l +

[BZB(C6F5)31- ( 4 ) , h a s b e e n s t r u c t u r a l l y c h a r a c t e r i z e d (see F i g u r e 2 ) .19)

The most i n t e r e s t i n g f e a t u r e i n t h e c o n t e x t o f t h e p r e s e n t d i s c u s s i o n i s t h e v a r i e t y o f c o o r d i n a t i o n modes of t h e b e n z y l ligands,

r e l i e v i n g t h e e l e c t r o n i c u n s a t u r a t i o n of Z r . I t i s w o r t h

recalling that styrene

of a metal-benzyl

i n t h e syndiotactic s p e c i f i c polymerization

the polyinsertion

growing c h a i n e n d . Compounds 1 a n d 2

i s 2,1,

are

leading t o

zwitterionic,

owing

to

the

q6

c o o r d i n a t i o n of t h e b e n z y l l i g a n d of t h e b o r a t e a n i o n . I n 2 t h e t w o Zr-bound b e n z y l s a r e s i m p l y q1 ( Z r r e a c h e s a 1 6 - e l e c t r o n c o n f i g u r a t i o n ) , w h i l e i n 1 o n e o f t h e b e n z y l s i s q2 a n d t h e o t h e r

t w o a r e q1 surprisingly,

(Zr

reaches

a

14-electron

configuration).

3 has a completely d i f f e r e n t s t r u c t u r e ,

Quite

with no

20. Copolymerization with CpTiC1,-MA0 Catalyst

c a t ion-anion

217

b o n d i n g i n t e r a c t i o n : Z r h a s o n e q3-benz y l a n d o n e

unprecedented

q7-benzyl

configuration. Finally,

ligand,

reaching

a

18-electron

i n 4 Z r i s bound t o o n e fll-benzyl,

while

t h e -CH2CHMeCH2Ph g r o u p b e h a v e s a s a c h e l a t i n g f11:q6 l i g a n d ; t h i s unusual back-biting

arene coordination,

due

Z r r e a c h e s a 16-

electron configuration.

F i g u r e 3 . X-ray c r y s t a l s t r u c t u r e s of [Cp*ZrBz21t 3 ( l e f t ) and o f

[Cp*Zr (CH2CHMeCH2Ph)B z I t 4 ( r i g h t ), d e r i v i n g from propene s i n g l e i n s e r t i o n i n t o t h e former. The [BzB(CgF5)3]- anions, which are not i n t h e coordination sphere of Z r , are CONCLUSIONS The complexes d e s c r i b e d above show t h e s t r o n g t e n d e n c y o f t h e coordinatively unsaturated Z r cations t o r e l i e v e t h e i r electrond e f i c i e n c y a l s o t h r o u g h u n u s u a l bonding i n t e r a c t i o n s . A s i m i l a r behaviour

i s r e a s o n a b l y e x p e c t e d f o r T i a n d it i s

a l s o e x p e c t e d t h a t i n s o l u t i o n d i f f e r e n t complexes a r e o b t a i n e d d e p e n d i n g on t h e c o m p e t i t i o n f o r c o o r d i n a t i o n o f t h e p a r t i c u l a r ligands,

i . e . t h e c o u n t e r i o n , t h e s o l v e n t , t h e monomer and, when

t h e complexes a r e c a t a l y t i c a l l y a c t i v e , t h e growing c h a i n e n d .

seems l i k e l y t h a t , i n t h e p r e s e n c e o f s u c h d i f f e r e n t monomers a s s t y r e n e , c o n j u g a t e d d i e n e s o r o l e f i n s , a c t i v e species of d i f f e r e n t s t r u c t u r e might b e formed, and t h a t t h e y i n t e r c o n v e r t t o e a c h o t h e r d u r i n g c o p o l y m e r i z a t i o n , a l s o as a c o n s e q u e n c e o f t h e i n s e r t i o n of d i f f e r e n t monomers. The p o s s i b i l i t y o f o b t a i n i n g s y n d i o t a c t i c p o l y s t y r e n e from many d i f f e r e n t c a t a l y t i c c o m b i n a t i o n s (CpTiC13-MA0, CpTiC12-MAOI Cp*TiR3-B (C6F5) 3, C p * T i R 3 - [ H N R 3 1 + [B (C6F5) 4 ] - , TiBz4-MAO, T i ( O R ) 4It

218

A. Zambelli and C. Pellecchia

MAO, Ti (acac)3-MAO (acac = acetylacetonate) , ZrBz4-MA0, Zr(acac)2C12-MAO) suggests that in any case the syndiotactic specificity arises from steric interactions between the last unit of the growing chain end and the monomer to be incorporated next, in the presence of a rather crowded and stereorigid ligand environment. This hypothesis is also supported by the experimentally observed polymer microstructure, which is in agreement with the first order Markovian statistical model of syndiospecific propagation.lo=) The reasons of the higher activity of the catalysts based on Ti compounds in comparison with catalysts based on Zr compounds in promoting styrene polymerization are not clearly established. The possibility of obtaining active catalysts using Ti (111) compounds could suggest that the active species are formed after reduction of the group 4 metal, which is easier for Ti(1V) than for Zr(1V). Another difference between the two metals is the softer Lewisacidity of Zr in comparison with Ti, which is confirmed also by the formation of strong Zr-arene Ic-bonding interactions in the above mentioned cationic complexes: on this basis, one could hypothesize different reactivities of the metal-polystyryl bonds for Zr and Ti. REFERENCES AND NOTES 1. K.Soga, J.R.Park, and T.Shiono, P o l y m . Commun., 2, 310 (1991). 2. (a) N. Ishihara, M.Kuramoto, and M.Uoi, Macromolecules, 2, 3356 (1988); (b) A.Zambelli, C.Pellecchia, and L,Oliva, ibid., 22, 2129 (1989); (c) A.Zambelli, C.Pellecchia, L.Oliva, P.Longo, and A.Grassi, Makromol. Chem., 104 (1991).

u,

3. (a) L.Oliva, P .Longo, A.Grassi, P .Ammendola, and C .Pellecchia, Makromol. Chem., Rapid Commun., JJ, 519 (1990); (b) G.Ricci, S.Italia, A.Giarrusso, and L.Porri, J. Organomet. Chem.,

m,

67 (1993).

4. G.Natta, and F.Danusso, Chim. Ind. (Milan),

a,445

(1958).

5. L.Porri, private communication: G.Ricci, S.Italia, and L.Porri, Macromolecules, in press. 6. (a) C.Pellecchia, A.Proto, and A.Zarnbelli, Macromolecules, E, 4450 (1992); (b) A.Zambelli, A.Proto, P.Longo, and P.Oliva, Makromol Chem., submitted

.

.

20. Copolymerization with CpTiC1,-MA0 Catalyst 219

7. See e.g. : L.Porri, A.Giarrusso, and G.Ricci, M a k r o m o l . Chem., Macromol. Symp., 48/49, 239 (1991). 8. C.Pellecchia, P.Longo, A.Grassi, P.Ammendola, and A.Zambelli, M a k r o m o l . Chem., R a p i d Commun., B, 277 (1987). 9. P.Longo, A.Grassi, and L.Oliva, Makromol. Chem.,

u, 2387

(1990). 10. (a) A.Grassi, C.Pellecchia, P.Longo, and A.Zambelli, G a z z . Chi m. Ital., U , 65 (1987); (b) A.Zambelli, P.Longo,

C.Pellecchia, and A.Grassi, M a c r o m o l e c u l e s , a,2035 (1987); (c) A.Zambelli, P.Ammendola, and A.Proto, i b i d . , 2,2126 (1989). 11. For recent reviews of the matter, see, e . 9.: (a) A.Zambelli,

C.Pellecchia, and L.Oliva, M a k r o m o l . Chem., M a c r o m o l . S y m p . , 48/49, 297 (1991) ; (b) R.F.Jordan, Adv. O r g a n o m e t . Chem., z, 325 (1991). 12. J.C.W. Chien, Z.Salajka, and S.Dong, M a c r o m o l e c u l e s ,

25, 3199

(1992). 13. (a) R.E.Campbel1, Eur. P a t . Appl. EP 421,659 to Dow Chemical

Co. (1991); (b) C.Pellecchia, A.Proto, P.Longo, and A. Zambelli, M a k r o m o l . Chem., R a p i d Commun., U , 663 (1991); (c) idem, i b i d . , 277 (1992).

u,

14. C.Pellecchia, P .Longo, A.Proto, and A.Zambelli, M a k r o m o l . Chem., R a p i d Commun., 265 (1992).

u,

15. D.J.Gillis, M.-J.Tudoret, and M.C.Baird, J. Am. Chem. SO C. 2543 (1993). 16. NMR monitoring of the reaction between Cp*TiBzj and B(C6F5)3

m,

at low temperatures showed the formation of [Cp*TiBz2]+ [BZB(C6F5)3]-, which evolved over a short period of time. ESR experiments at room temperature showed the rapid formation of Ti(II1) species. Unpublished results from our laboratory. 17. C.Pellecchia, A.Grassi, and A.Immirzi, J. Am. Chem. S O C . , m, 1160 (1993). 18. C .Pellecchia, A. Immirzi, A.Grassi, and A. Zambelli, O r g a n o m e t a l l i c s , U , 4473 (1993). 19. Unpublished results from our laboratory. 20. (a) C.Pellecchia, A.Grassi, and A.Zambelli, J. Chem. SO C. , Chem. Commun., 947 (1993); (b) C.Pellecchia, A.Grassi, and A.Zambelli, O r g a n o m e t a l l i c s , 298 (1994).

u,

This Page Intentionally Left Blank

22 I

21. The Role of Ion-Pair Equilibria on the Activity and Stereoregularity of Soluble Metallocene Ziegler-Natta Catalysts

JOHN J. EISCH and SONYA I. POMBRIK Department of Chemistry, State University of New York at Binghamton, Binghamton, New York 13902-6000 U.S.A. STEFAN GifRTZGEN, RAINER RIEGER and WOLFRAM UZICK Witco GmbH, D-59192 Bergkamen, Germany, a subsidiary of Witco Corporation, 520 Madison Avenue, New York, New York 10022, U.S.A. ABSTRACT By means of temperature-dependentmultinuclear NMR studies, electrical conductivity measurements and polymerization activity assessments, the nature of the resulting complexes formed between metallocene alkyls or metallocene dihalides and Group 13 Lewis acids was investigated comprehensively. The influences of solvent polarity, concentration, temperature and the strength and proportion of the Lewis acid upon the structure of the metallocene-Lewis acid complexes furnish cogent evidence for the existence of an equilibrium between contact ion pairs (CIP)and solvent-separated ion-pairs (SSIP). Three systems were considered in this study: 1) Cp2TiQ + AlC13; 2) bisb-butylcyclopentadieny1)zirconiumdimethyl + (C6F~)3B;and 3) bisb-butylcyclopentadieny1)zirconium dichloride + MAO. More polar media and higher dilution have been shown to favor the solvent-separatedion-pair isomer over the contact ion-pair isomer. In pj-basic solvents, such as arenes, evidence indicates that a 1:l:l complex of the metallocene, the Lewis acid and the arene is formed reversibly (solvated cation-anion pair, SCAP). Polymerization activities toward ethylene or propylene, as measured in these studies or reported in the literature, support the conclusion that the solvent-separated ion pairs are the most active catalyst sites but are less stereoselective in syndiotactic polymerization. In a related study, the MqAl-content of MA0 could be reduced to zero by toluene-evaporationsat 25OC. The MqAl-free MAO, upon admixture with (CD,),AI, rapidly displayed an 'H N M R signal revealing the regeneration of "free" Me,Al. This observation is consistent with the rapid methyl group exchange between the terminal Me2A1 groups of MA0 chains, MqAl-(-OAIMe-)-,Me, and the added (CD3),A1 and supports the suggestion that these terminal Me2Al groups are the most reactive structural unit of MAO,

222

J.J. Eisch and S.I. Pornbrik

either toward external Me3A1or toward metallocene dialkyls or dihalides. In this view such terminal M q A l groups are responsible for the heightened cocatalytic activity of MA0 in Ziegler-Natta catalysts. This view conforms with the heuristic Steiger-Kaminsky hypothesis. INTRODUCTION Ziegler-Natta catalysts based on transition metal salts and aluminum akyl cocatalysts have been valued as effective polymerization catalysts for unsaturated hydrocarbons, such as olefins and 1,3-dienes. for over 40 years'). However, the active sites of these possibly multi-site catalysts have never been completely identified because of the complexity of the surface reactions occurring in these heterogeneous systems2). After the initial discoveries of Ziegle?) and Natta4) a second breakthrough in polyolefin catalysis was initiated by Ewen5), by Kaminski and Sinn6) and by Brintzinger'), who were successful in polymerizing olefins stereospecifically by means of bridged cyclopentadienyl transition metal complexes, or stereorigid metallocenes, that had been activated by partially hydrolyzed aluminum alkyls known as aluminoxanes, (-RAl-O-)n. In the intervening decade these stereorigid homogeneous catalysts have attracted much interest in industrial and academic laboratories because of their high activity and their potential for producing polymers with novel and highly specific properties*). Over and beyond their commercial importance, these catalyst systems confront researchers with the question of their molecular mode of action. In attempting to answer this question, we and others have found that such homogeneous catalyst systems lighten the task somewhat by permitting kinetic and spectral measurements to be made in solution. Our report is intended to contribute to an understanding of this molecular mode of action and thereby to supplement the independent studies of various groups, including our own, who have marshaled evidence that ion pairs are involved9) and that some type of ionic intermediate is clearly either the active catalyst site or an essential precursor of such a site.lO) In order to probe the ionic character of such a catalytic site, we chose to examine three combinations of a metallocene and a Group 13 Lewis acid 1) titanocene dichloride (4) and AlCl, (5)9b); 2) bis(Il-butylcyclopentadienyl)zirconium dimethyl (la) and tris(pentafluoropheny1)borane (2); and 3) bis(pbutylcyclopentadieny1)zirconiumdichloride (1b) and poly(methy1aluminoxane) (MAO, 3). The techniques employed were multinuclear NMR spectroscopy and measurements of electrical conductivity and polymerization activity. The last two systems represent active catalysts for the polymerization of ethylene or propylene and hence their molecular modes of action have great practical relevance for Ziegler-Natta industrial polymerization technology. The first system, CpZTiCl2 AlCl,, is not an olefin polymerization catalyst -, but it is closely related to the Natta-Breslow homogeneous catalysts, Cp2TiC12-R,A1,,. Accordingly, the formation of ionic intermediates can be studied in the (4 + 5) system without the disturbing influence of transalkylation (equation 1):

21. Role of Ion-pair on Activity and Stereoregularity

223

In addition, some studies were carried out to learn how "free"MqAl is bonded in MA0 and how readily methyl groups undergo exchange between MA0 and "free" Me3Al. Such information is important in defining the cocatalytic action of MA0 in Ziegler-Natta catalysis. RESULTS The Svstern, TitanoceneDichloride (4) and Aluminum Chloride (51. To obtain a better understanding of how titanocene dihalides and aluminum halides interact and yet avoid group-exchange equilibria (e.g., equation l), the system, cp2TiC12-AlC13was examined by variable-temperature, multinuclear NMR spectroscopy and with variations in concentration and solvent polarity. Presented in Figure 1 are the 'H N M R spectra of 1:l mixtures of Cp2TiC12 and AlCl, in methylene chloride at two concentrationsand at five temperatures between +27 and -30 OC. In both cases, the singlet near 6.80-6.81 pprn at 27 OC resolves into two singlets at 6.90 and 6.77 pprn at -30 OC. The resolution into these two signals upon cooling and the coalescenceinto the one signal upon warming are fully reversible. It is furthermore noteworthy that, of the two, the high-field signal becomes dominant upon dilution (Figure la). Similar observations were made when the I3C N M R spectra of 1: 1

I

mol% Metallocen

Electric conductivity of the system la/2 at 0 OC in toluene solution. Figure 1. Absolute concentration of la: 0 - 149 mm0v1.

224 J.J. Eisch and S.1. Pombrik

mixtures of CpZTiCl2 and AlC13 in CH2C12at two concentrations were examined. The singlet near 123 ppm at 25 OC resolved into two singlets, in a reversible fashion, near 123.3 and 122.6 ppm at -30 O C . Again the high field singlet at -30 OC had significantly higher intensity in more dilute solution. These results demonstrate the existence of two different types of cyclopentadienyltitium compounds that are in ready equilibrium with each other. The variable temperature nAl N M R spectrum of the Cp2TiCl2-A1Cl3 system at 0.1 M in CH2C!12displayed a broad, dominant peak at 104.5 ppm at 25 OC (wlR= 356 Hz) with a minor shoulder peak at 99 ppm. Cooling the sample successively to -20°, -40°, and then -70 OC caused this peak to broaden and then to narrow selectively on its high-field side ( W I R - 563,666, and 300 Hz,respectively). At the same time a new, unresolved peak of comparable area began to emerge in the 98-100 ppm region (win = 356 Hz). Limited solubility of the components and high viscosity at -70 OC prevented better peak resolution. Nevertheless, these "Al N M R results support the presence of at least two different aluminum-containingcomponents, each of which is tetracoordinate. The effect of solvent polarity on the ratio of the two cyclopentadienyltitanium componentspresent in the Cp2TiC12-AIC13system was proved by recording the 'H spectrum in CHzC12containing 0,20, or 33% of CC1, (v/v). As is evident from Figure 2, the singlet

I Figure 2.

1

I,

Y

'H-NMR spectrum of metallocene l a in toluene-d8 at room temperature.

observed at 25 OC (column a) shifts from 6.75 to 6.82 ppm as the solvent contains more CCI, and thereby becomes less polar. In keeping with this, the ratio of the two peaks observed at -30 OC (row b) changes to favor the component at lower field (peak near 6.9 ppm). Since this component is favored by a medium of lower polarity, it must be less polar than the component absorbing near 6.7 ppm. An analogous change in the ratio of these two components was exhibited when the CH,CI, solution was diluted with toluene; integration

21. Role o f Ion-pair on Activity and

Stereoregularity 225

of the two peaks at -30 OC permits the determination of equilibrium constants: high field peak = 120 in pure CH2Q and 0.37 in a 5:l (v/v) mixture of low-field peak CH2C12-toluene. Keq Of

Metallocenes

Lewis Acids

The System, Bish-butvlcvclouentadienyl) Dimethyl (la) and Trisbentafluoropheny1)borane (2). Even at temperatures below -90O C immediate reaction takes place between l a and 2 in toluene solution upon addition of an excess metallocene la to 2, resulting in a two-phase system. A yellow color spontaneously occurs at the phase-contact boundary. Upon warming to ambient temperature the increasing methane evolution indicates the onset of extensive decomposition. Upon cooling below the the freezing point of toluene-d8 (-1 10 "C), a red color develops, which disappears upon warming to ambient temperature and reappears on recooling. This readily reversible thermochromic effect could also be observed with other metallocenes, as for example 1,l'ethylene-bis(indeny1)zirconium dimethyl. Guided by earlier studies of Natta-Breslow catalysts"), we performed electrical conductivity measurements on mixtures of la and 2 in toluene solution in order to obtain evidence for the presence of ions. The results are give in Figure 3. A maximum total conductance of 20.7 pS was measured at 165 mole % of la and was shown to be reproducible. Such increasing conductance is clear evidence for ion-pair formation. Upon increasing the excess of l a to more than 165 mole % we observed a separation of phases. According to the electric conductance figures obtained after phase separation (see Table I), the ionic products are obviously concentrated in the lower layer.

226 J.J. Eisch and S.I. Pombrik

su

h

293 K 258 K

I

248 K 238 K

208 K I

'

6

:

4

'

6'2

'

6'0

'

5'8

'

5'6

'

5'4

5'2

5'0

'

4'8

Expansion of cp proton absorption area in temperature dependent 'H-NMR Figure 3. spectra of equimolax mixtures of la and 2 in tolueneas (a) - (0. The top line shows the signals of pure l a for comparison.

Table 1. Electric conductivity data of separated layers at 0 OC. Maximum concentration of la: 320 mmolefi. Electric Conductivity [ pS ] Excess l a [mole %]

Upper layer Lower layer

186 2.2

110

196 1.1 240

207

0.7 270

Multinuclear N M R Measurements. In the 'H-NMR spectrum of la the signals at 5.70 to 5.90 ppm can be assigned to the Cp-ring protons and the signals at 2.60, 1.50 to 1.70 and 1.10 ppm to the protons of the pbutyl group. The singlet at 0.02 ppm is caused by the protons of the two equivalent methyl groups attached to zirconium. The temperature-dependence of the 'H NMR spectrum of an equirnolar mixture of l a and 2 (cf. Figure 4) shows that the dimethylzirconium grouping no longer is evident while two new signals appear between 0 and 1 ppm, in addition to a signal at 0.4 pprn representing a small amount of methane.

21. Role of Ion-pair on Activity and Stereoregularity

227

zr.0 = 21 upper layer

B.0 = 1:l

a 0 = 2 : 1 lowerlayer

20=12

I\

a 0 = 0:1

Temperature dependent 'H-NMR spectra of equimolar mixtures of l a and 2 Figure 4: in toluene-d8. The top line shows the signals of pure la for comparison. The new signal at 0.5 ppm can be assigned to one methyl group attached to a zirconium cation;the new signal at 0.3 ppm would accord with one methyl group attached to boron. Significant line-broadening of all signals can be observed upon decreasing the temperature. The Cp-ring-proton signals are resolved to form additional multiplets, indicating new components with non-equivalent Cp-rings. The observed appearance of new Cp-proton absorptions upon cooling was found to be completely reversible and is in good agreement with our previous NMR investigations on the system, Cp2TiC12 and Alc13'~). By means of solvent polarity variation the high-field Cp proton signal, @f. infra) which appears upon cooling. can be assigned to that arising from solvent-separated ion pairs. This is also supported by our results illustrated in Figure 4. Furthermore, ion-pair formation is supported by the "B-NMR spectra at different molar ratios of l a and 2. The absorption of covalent boron in 2 appears at 60 ppm. Upon reaction with l a a new signal appears at -14 ppm, which can be assigned to an anionic boron component. The System, Bis(n-butvlcvcloDentadienvlhirconiumdichloride i l b ) and Poly(methy1aluminoxane) (3) Electrical Conductivity Measurements. In contrast with the reaction of l a and 2, no phase separation could be observed upon addition of metallocene l b to a poly(methy1aluminoxane) (3) solution in toluene. However, color changes from colorless to yellow could also be observed. Electric conductivity studies were performed at O°C and at this temperature we found only very slight methane evolution and were able to achieve a good reproducibility in our measurements. A steady, almost linear increase in total conductance was observed by increasing the amount of lb. These results give evidence for

228 J.J. Eisch and S.I. Pombrik

the formation of ion pairs upon reaction of metallocene l b with methylaluminoxane 3 in toluene solution. The electric conductance obtained was of the same magnitude as that observed by the addition of metallocene la to 2: after 140 mole % of l b had been added to 3, the resulting conductance was 17 pS. Mulfinuclear N M R Measurements. The temperature-dependent 'H N M R spectra obtained from a mixture of l b and 3 gave results comparable with those obtained from equimolar mixture of la and 2 M. Figure 4). Besides the signals arising from unreacted l b and 3 (broad high field signal), smaller signals of about 20% comparable intensity could be observed (partly overlapping), These smaller signals show the same shift dependency by variation of the temperature as it was found in the la 2-system. An additional new signal was found at 0.5 ppm indicative of a methyl group. This shift is nearly the same as the shift observed for the cationic Zr-CH3 group in the l a 2-system. hemration of Polv(methv1aluminoxane)Not Containing "Free" Trimethylaluminurn. A toluene solution of poly(methy1aluminoxane (MAO) in toluene (10% by weight from Schering GmbH) displays a broad unresolved 'H NMR signal between +0.50 and -0.75 ppm. This broad signal is surmounted at -0.30 ppm by a sharp singlet arising from so-called "free" trimethylaluminum (TMA) even though the TMA is undoubtedly complexed with oxygen in a OA12-unit. Although the TMA can be diminished by heating MAO, the MA0 undergoes irreversible structural alterations by such thermal treatment. In order to remove the "free" TMA without the risk of such structural change, we have attempted to remove the TMA from MA0 by co-evaporation with toluene at 25OC. By evaporating the original toluene in vacuo and by replenishing the toluene and re-evaporating three more times, the original TMA-signal at -0.30 ppm (estimated at 30% of the methyl groups) had completely disappeared. When such "TMA-free" MA0 was redissolved in toluene, treated with (CD3)3Al (98% minimum deuteriation) and the 'H NMR spectrum remeasured, a strong, sharp signal assignable to "free" TMA was clearly evident. From this facile generation of "free" TMA, it can be concluded that certain MeAl units in the MAO, probably the terminal Me2Al unit, can readily exchange with TMA. Catalvst Productivity Numbers for the Homogeneous Catalysis of Ethylene Polymerization. Measuring and interpreting the kinetics of ethylene polymerization, even under initially homogeneous conditions, are difficult to perform in a reproducible and reliable manner. In order to obtain some convenient, useful measure of the catalyst activity of the Cp2TiMeC1-Me,AlC13-, system, therefore, batch polymerizations were conducted under standardized experimental conditions with individual variations in solvent, temperature, concentration, ratio of the Ti:Al components, and the nature of the Lewis acid aluminum cocatalyst. In order to produce an easily measurable amount of polyethylene, the polymerization runs were conducted for a standard 25-min period, during which time the initially homogeneous catalyst system turned heterogeneous. Such heterogeneity was shown not to be a controlling factor in altering the ranking of productivity numbers as the solvent

-

-

21. Role of Ion-pair o n Activity and Stereoregularity 229

polarity or catalyst concentration was changed by the following experiments. Polymerization runs of short duration (100 s), during which time the catalyst system remained homogeneous (little or no precipitation), gave essentially the same results: increasing the polarity of the solvent and decreasing the catalyst concentration increased the PN values. Table II.

Variations in the Productivity Number (PN) for the Homogeneous, Catalytic Polymerization of Ethylene by Cp2Ti(Me)C1

Me,A1Cl3-, with Reaction

Conditions * ’

run no.

solvent

temp

concn,

ratio

OC

mmolL’

Ti:A

Lewis acid

PN

1

CH,CI,

25

1.o

1:1

MeAICI,

140

2

CH,CI,

25

1.o

1:2

MeAICI,

163

3

CH,CI,

25

1.o

1:8

MeAICI,

140

4

CHZCI2

25

0.26

1:8

MeAICI,

440

5

CH,CI,

25

1.o

1:16

MeAICI,

94

6

CH,CI,

25

1.o

1:8

Me3AI

20

7

CH2CI2

25

1.o

1:8

Me2AICI

137

8

CH2CI2

25

1.o

1:8

AICI,

112

9

toluene

25

1.o

1:8

MeAICI,

49

10

toluene

25

0.26

1:8

MeAICI,

197

11

mesitylene

25

1.o

1:8

MeAICI,

11

12

toluene

70

1.o

1:8

MeAICI,

75

13

toluene

50

1.o

1:8

MeAICI,

69

14

toluene

25

1.o

1:8

MeAICI,

49

15

toluene

0

1.o

1:8

MeAICI,

10

16

toluene

-10

1.o

1:8

MeAICI,

c1

a

Productivity number (PN) is defined as grams of polyethylene per gram of Cp,TiMeCI per atm of monomer per hour and was found in six repeated runs to be reproducible to within f5%.

On the basis of six runs under identical conditions, average productivity numbers were determined as grams of polyethylene per gram of methyltitanocene chloride per atmosphere of ethylene per hour, which values were reproducible to within 3 3 % (Table 11). The catalyst system in solution before polymerization was generally golden colored,

230 J.J. Eisch and S.I. Pornbrik

consistent with the persistence of titanocene(IV). Only with Me3A1as a cocatalyst (run 6) did a deep blue color develop after 5 min of admixing; this indicated alkylative reduction to titanocene0. Catalyst Productivitv Numbers as a Function of the Concentration of the Polv(methvla1uminoxane)(MAO). In contrast with the increase in PN upon dilution of the catalyst concentration, as noted with the CpzTi(Me)C1-MeAlCl2 system, diluting the concentration of the Cp2TiC12-MA0 catalyst had no significant effect on the PN for the polymerization of ethylene. Thus, the following PN (conc. mmoVL) were observed: 212 (OSO), 241 (1.0) and 212 (2.0). DISCUSSION Interaction of Titanocene Dichloride (4) and Aluminum Chloride (5). From the foregoing variable-temperature,multinuclear NMR data, it is evident that the interaction of A1C13 (5) with Cp2Ti(Me)C1(6a), with Cp2Ti(CH2-SiMe3)C1 (6b). and even with Cp2TiC1, leads to the generation of some positive charge at titanium and the formation of the partly or wholly free AlCL-ion. Whether these effects find their explanation in either a contact-ion-pairstructure (6) or a solvent-separatedion pair (7) or both remains to be decided (equation 2).

66a R = M e 6b R = CH,SiMe, 6c R = C I

As to the positive charge on titanium, several data supports such a view: (1) the CH, group in 6a and the CH2 group in 6b are shifted downfield in the 'H spectrum by 1-2 ppm over their positions in the starting titanocenes, analogous to the downfield shifts seen in protons adjacent to carbenium ion centers: (2) the Cp protons in 6a 6c also occur 0.2-0.3 ppm downfield from their positions in the starting titanocenes; (3) in the known crystal structure of Cp2Ti(C1) A1MeC12, the bridging chloro ligand has a Ti-Cl separation of 0.23A. greater than that of the unbridged Ti-Cl bond. Either the presence of 6 or 7 in solution or a rapid equilibration between them at 25 OC would account for these observations. That at least a significantproportion of such 1:l complex with AlC13 exists, below -2OoC, as 7 is indicated by the presence of a relatively sharp, dominant "Al peak at 103 ppm. This peak corresponds exactly to the value reported for the free A1C14- ion. Upon warming, this signal broadens toward the low-field side and yields a maximum at 25OC between 104 and 105 ppm. From the 27Alspectra of similar complexes, this broadening is consistent with the equilibration depicted in equation 2. For example, the crystal structure

-

21. Role of Ion-pair on Activity and Stereoregularity 231

Ph(MeC=C(SiMq)Ti+Cppz- A1C14- shows the presence of individual ions, and displays its 27Alpeak as a sharp singlet at 103.6 ppm (CH2C12). Since the equilibrium of equation 2 was most likely operative in interconverting the bridged structures, 6, into their solvent separated counterparts, 7, we chose to study this equilibrium for the case where R = Cl. In this situation, no competing reaction could occur whereby any akyl group on titanium could be transferred to aluminum. Cooling in either CH2C12or toluene to below -20 "Cgave two well-resolved signals for the Cp-ring protons or carbons. Neither signal occurred where the signals of any uncomplexed Cp2TiC12would have occurred, if the low-temperature NMR spectra arose from the "freezing out" of the extensive dissociation of 17c into its components. The only reasonable interpretation, then, of the low-temperature N M R results is that the equilibration between 6 (contact IP) and 7 (solvent-separated IP) (equation 2) has been "frozen out". That such systems contain ions is evident from the pioneering work of Dyachkovskii and co-workers on the electrical conductivity, electrodialysis, and polymerization of ethylene observed in halocarbons.")* 12) It remained to be learned whether such spectra at -30 OC responded to changes in concentration and in solvent polarity, as would be expected of CIP-SSIP equilibria. In accordance with Ostwald's dilution law, the two-particle, solvent-separatedion pairs (SSIP) should be favored at lower concentrations over one-particle, contact ion pairs (CIP), and that is what is observed. Furthermore, SSIP should be preferred over CIP as the medium becomes more polar, and again this is observed. From these observations, the high-field signal in the 'H and 13CN M R spectra at -30 OC can be assigned to the solvent-separated ion pair and the low-field signal to the contact ion pair. Productivity Numbers of the Cp,Ti(Me)C1-Me,AIClq, Svstem for the Polymerization of Ethylene as a Function of Experimental Conditions. The observed changes in PN (Table II) with variation in the experimental conditions for polymerization (more polar solvent, lower concentration, relative insensitivity to temperature increase above a threshold temperature, and relative insensitivity to the Lewis acid strength of Me,AlC13-,,) can best be reconciled with the conclusion that the solvent-separated ion pair Cp2Ti+Me11 A1Me,C14-, is the most active polymerization initiator in polar, nondonor solvents such as CH2C12 If the contact ion pair Cp2(Me)Ti C1 Me,AlC13-, were the active site, then the PN should have shown an increase at higher concentrations and in polar solvents. The relative insensitivity of PN to the Me,AlC13, species used as the Lewis acid with Cp2TiMeC1is consistent with the interpretation that even the weakest acid, Me2AlC1, used in a 1:l ratio (run 1). is sufficient to produce the SSIP (equation 3) in significant

Cp2TiMeCI

+

MefiICI

A 7

Cp2Ti+ Me AIMe2C12-

(3)

proportions. Used in a 1:2 ratio (run 2), MeAlC12 could shift the equilibrium of equation 3 to the right by the law of mass action. On the other hand, the deleterious effect on PN of a

232 J.J. Eisch and S.I. Pombrik

-

Ph(MeC=C(SMe3)Ti+Ch A l Q - shows the presence of individual ions, and displays its 27Alpeak as a sharp singlet at 103.6 ppm (CHZC12). Since the equilibrium of equation 2 was most likely operative in interconverting the bridged structures, 6, into their solvent separated counterparts, 7, we chose to study this equilibrium for the case where R = C1. In this situation, no competing reaction could occur whereby any alkyl group on titanium could be transferred to aluminum. Cooling in either CH2C12or toluene to below -20OC gave two well-resolved signals for the Cp-ring protons or carbons. Neither signal occurred where the signals of any uncomplexed Cp2TiC12would have occurred, if the low-temperatureNh4R spectra arose from the "freezing out" of the extensive dissociation of 17c into its components. The only reasonable interpretation, then, of the low-temperatureNMR results is that the equilibrationbetween 6 (contact IP) and 7 (solvent-separatedIP) (equation 2) has been "frozen out". That such systems contain ions is evident from the pioneering work of Dyachkovskii and co-workers on the electrical conductivity, electrodialysis, and polymerization of ethylene observed in halocarbons."). 12) It remained to be learned whether such spectra at -30 OC responded to changes in concentration and in solvent polarity, as would be expected of CIP-SSIP equilibria. In accordance with Ostwald's dilution law, the two-particle, solvent-separatedion pairs (SSIP)should be favored at lower concentrations over one-particle,contact ion pairs (CIP), and that is what is observed. Furthermore, SSIP should be preferred over CIP as the medium becomes more polar, and again this is observed. From these observations,the high-field signal in the 'H and 13CNMR spectra at -30 OC can be assigned to the solvent-separatedion pair and the low-field signal to the contact ion pair. Productivitv Numbers of the Cu7Ti(Me)Cl-Me,A1C1?-m Svstem for the Polvmerization of Ethvlene as a Function of Experimental Conditions. The observed changes in PN (Table II) with variation in the experimental conditions for polymerization (more polar solvent, lower concentration,relative insensitivity to temperature increase above a threshold temperature, and relative insensitivity to the Lewis acid strength of Me,A1Cl3.,) can best be reconciled with the conclusion that the solvent-separatedion pair Cp2TitMe 11 A1Me,Cl4-, is the most active polymerization initiator in polar, nondonor solvents such as CH2ClP If the contact ion pair Cp2(Me)Ti C1 MenA1C13, were the active site, then polar solvents. the PN should have shown an increase at higher concentrationsand in The relative insensitivity of PN to the Me,AlC13-, species used as the Lewis acid with Cp2TiMeClis consistent with the interpretation that even the weakest acid, MqAlCl, used in a 1:1 ratio (run 1). is sufficient to produce the SSIP (equation 3) in significant

Cp,TiMeCI

+

MefiICI

7

Cp2Ti+Me AIMe2C12-

(3)

proportions. Used in a 1:2 ratio (run 2), MeA1C12 could shift the equilibrium of equation 3 to the right by the law of mass action. On the other hand, the deleterious effect on PN of a

21. Role of lon-pair on Activity and Stereoregularity 233

large ratio of Lewis acid to Cp2TiMeCl (1:l -,16:1, PN of 140 494, runs 1 and 5 ) can be attributed to a transmethylation that destroys the crucial Ti-Me bond of Cp2TiMeCl. In a similar vein, the negative effect of MqAl (run 6) is readily ascribed to the destruction of Cp2TiMeC1by reduction to T i m . From the lower PN displayed by Cp2TiMeC1-MeAlQ2 as one passes from CH2C12 (140. run 3) to toluene (49,run 9) to mesitylene (11, run 11). it is clear that the solvated cation-anion pair 8 is the dominant ion pair present as an q'-complex in aromatics (equation 4) and that such a SCAP must be the least active polymerization site. The relative

reactivity of these ion pairs in polymerization can therefore be ordered as SSIP >> CIP > SCAP. Nature of the Poly(methv1aluminoxanae) (MA01Used in These Studies. The observation that MAO, which has been separated from any "free" TMA, readily exchanges methyl groups with externally added (CD3)3k1suggests that there are Me-A1 units that can undergo bridging and thus facile exchange with external TMA. We propose that the chain terminus units of MAO, namely the dimethylalumino groups, are the most likely sites for such exchange. Furthermore, such MqA1 units should also be the most reactive methylating sites for transition metal halides. Finally, the chloro- or dichloroalumino sites resulting from such methylation should be both sterically and electronically the strong Lewis acid sites that can foster metallocenium ion formation (equation 5). This proposal is based upon the

1CP2MCI2

Me2AI-MAo)

Cp2MMe2

C'2A1-MA0)

Cp2M+-Me

+

' cI..cI/'

Me,?Af '\Af CI

Me

8

L

(5)

4 % (0-AlMe j-

Me

findings and hypotheses of Steiger and Kaminsky. The fact that MA0 with Cp2TiC12yields a high PN (>200) that is relatively insensitive to dilution of the catalyst supports the conclusion that MA0 initially produces such a large proportion of solvent-separatedion pairs that dilution no longer promotes the dissociation of any significant proportion of contact ion pairs. The Polymerization Catalyst Systems. Bis(n-butylcycloDentadienylhirconium Dimethyl (la) with =)?B (2) and Bis(n-butylcvclomntadieny1hirconium Dichloride (lb) with MA0 (3). The main problem in observing ionic species under conditions typical for industrial polymerization processes is the low polarity of the reactants and the solvent. With weakly coordinating solvents like toluene and the complex anions possibly fonned from the typical methylaluminoxanecocatalyst, the stabilization opportunities for metallocenium

234 J.J. Eisch and S.I. Pombrik

cations are very limited. This, of come, limits the shelf-life of such ion complexes significantly so that we were not able to isolate any complexes from toluene up to now, whereas the solvatecomplexes with THF or acetonitrile are reported as isolated crystals and have been characterized by means of x-ray crystallography?) Unfortunately, these solvated complexes are inactive in polymerization. The present studies of the polymerization catalyst systems, l a with 2 and l b with 3, nicely complement the foregoing investigations of the Cp2TiCl2-A1Cl3 system. Although the temperature-dependentmultinuclear studies of the latter system have furnished the first evidence for the existence of equilibria among contact (CIP),solvent-separated (SSIP) and solvated-cation(SCAP) ion pairs, it remained to be established that such ion pairs were actually present into industriallyimportant Ziegler-Natta catalysts. Our present results with the l a 2 and l b 3 systems accomplish exactly this desired goal. The significant increase in electric conductivity upon reaction of the metallocenes l a or l b with either boron activator 2 or methylaluminoxane 3, respectively, clearly supports the formation of solvent-separatedion pairs and the existence of an equilibrium with contact ion-pairs. The 'H-NMR- and "B-NMR results provide further evidence for ion-pair formation in the l a 2 system. The molecular steps of activation by methylaluminoxaneare believed to be more complex. Some ideas concerning this mechanism of activation by MA0 have already been published and are considered here. The role of MA0 in these catalyst system will surely be the subject of further research. However, it can be stated that ion-pair formation can indeed be observed in olefin polymerization catalyst systems under conditions close to those employed in commercial-scalepolymerization. Imulications of Ion-Pair Formation for StereoregularPolvmerization. Although there is little doubt that solvent-separatedion pairs are the most active catalyst sites for the polymerization of ethylene and of propylene by soluble metallocene-Lewisacid combinations,the further tantalizing question remains: what is the role of such ion pairs in the stereoregulationof a-olefin polymerization? A partial answer to this question may be contained in a most recent, pertinent study by Herfert and Fink.13) These researchers investigated the syndioselectivepolymerization of propylene with the isopropylidene(fluoreny1)cyclopentadienyliirconium dichloride-MA0 catalyst system in CH2C12-toluenemixtures and found that as the polarity of the solvent increased, the rate of polymerization increased but the syndiotacticindex decreased. This suggests that solvent-separatedion pairs are more active sites but they are less syndioselective than contact ion pairs. Additional research on the influence of ion pairs on both syndiotactic and isotactic polymerization indices is needed, in order to settle this important issue.

-

-

-

REFERENCES 1. J. Boor, Jr.: "Ziegler Natta Catalysts and Polymerizations", Academic Press, New York, 1979, p. 670.

21. Role of Ion-pair on Activity and Stereoregularity 235

2. 3. 4. 5.

6. 7. 8. 9a. b. C.

d. 10a.

b. C.

d.

e.

f.

J.C.W. Chien, "Coordination Polymerization", Academic Press, New York (1975). K. Ziegler, Angew. Chem., 76,545(1964). G. Natta, Makromol. Chem.,l6,213 (1955). J.A. Ewen, J. Am. Chem. Soc.,106.6355(1984). W. Kaminsky, CLB Chemie f. Labor und Betrieb, 38,398,(1987). H.H. Brintzingner, et al., J. Organomet. Chem., 2&63 (1985). Cat. SOC.of Japan, 33,536,(1991). J.J. Eisch, A.H. Piotrowski, S.K. Brownstein, E.J. Gabe, F.J. Lee, J. Am. Chem. SOC.,

107,7219(1985). J.J. Eisch, S.I.Pombrik, G.X. Zheng, Organometallics,l2,3856 (1993). P.G. Gassmann, M.P. Callstrom, J. Am. Chem. Soc.,lO9,7875(1987). R.P. Jordan, L.S. Bajgur, R. Willett, B. Scott, J. Am. Chem. S O C . , ~7410 , (1986). M. Bochmann, S.J. Lancaster, Organometallics, 2,633 (1993). J.C.W. Chien, A.D. Rausch, W.M. Tsai, Appl. Organomet. Chem., z (1993). G.G. Hlatky, H.W. Turner,R.R. Eckmann, J. Am. Chem. S O C . , ~2728 , (1989) T.J. Marks,X. Yang, C.L. Stem, J. Am. Chem. S O C . , ~3623 , (1991). R. Jordan, D.J. Crowther, M.C. Baenzinger, A. Verme, Organometallics, 2.2574 (1990). e.g. Y.W. Aelyunas, N.C. Beanzinger, P.K. Bradley, R.F. Jordan, Organometallics,

12.

13, 148 (1994). F.S.Dyachkovskii, in 'Toordination Polymerization" (Ed. J.C.W. Chien). Academic Press, New York, 1975,p. 199. The electric conductivity of a pure toluene solution of l a under these conditions was

13.

determined to be d o 0 5 pS. N. Herfert and G. Fink, Makromol. Chem., 193,773(1992).

11.

This Page Intentionally Left Blank

237

22. High Molecular Weight Monodisperse Polymers Synthesized by Rare Earth Metal Complexes

H. YASUDA", E. IHARA, S . YOSHIOKA, M. NODONO, M. MORIMOTO, and M. YAMASHITA Department of Applied Chemistry, Faculty of Engineering, Hiroshima University, Higashi-Hiroshima 724, Japan

ABSTRACT High molecular weight polyethylene ( M n > 300,000, M,/Mn = 1.6) was obtained by using bulky Sm(I1) or Sm(II1) species such as Me2Si(2-SiMe34-tBu-CgH2)2Sm(THF) and Me2Si[2(3),4-(SiMe3)2-CgH212SmCH(SiMe3)2 as initiator. These complexes also conduct the polymerization of 1-pentene and 1hexene to give isotactic polymers of M n > 20,000 with Mw/Mn = 1.6. By contrast the polar monomers such as methyl methacrylate and alkyl acrylates also proceeded the living polymerization by using Ln(C5Me5)2R (R = H, Me) to give high molecular weight polymers ( M n > 450,000) with extremely low polydispersity (Mw/Mn < 1.05). INTRODUCTION Classical Ziegler-Natta catalyst positions as the 1st generation catalyst for polymerizations of ethylene or propylene. The 2nd genaration catalyst lies in MgC12 and/or donor supported Ziegler-Natta catalyst, which exhibits 100 times higher catalytic activity as compared with the 1st generation catalyst. These two are heterogeneous systems composed of two or three metal components. More recently, Kaminsky and Ewen catalyst systems, Z1€p2Cl/(AlMe-0-)~ and Me2C(flurorenyl)CpHfC12/(A1Me-O-)nYwere found as the 3rd generation catalyst for polymerization of ethylene or propylene. These systems are homogeneous, but the clarification of the polymerization mechanism has failed because of the complex structure of (AlMe-O-)n. Homogeneous organolanthanide complexes such as LaH-(C5Me5)2 belongs to the 4th generation catalyst exhibiting high initiating property for the polymerizations. We have synthesized new series of bulky organo-rare earth(I1) and -(III) complexes as excellent catalyst for polymerization of ethylene and 1olefins. On the other hand, we have first found that organolanthanides such as LnR(C5Meg)2 (Ln = Y, Yb, Sm, Lu; R = H, Me) exhibits excellent catalytic activity for the polymerization of polar monomers such as alkyl methacrylates and alkyl acrylates. These initiators conducted the living polymerization of these monomers

238 H. Yasuda, E. Ihara, S. Yoshioka. M. Nodono, M. Morimoto and M.Yamashita

to afford high molecular weight polymers (Mn > 450,000) with extremely narrow molecular weight distribution (Mw/Mn ~ 1 . 0 5 ) .By taking advantage of the living polymereization properties for both methyl methacrylate(MMA) and alkyl acrylates such as butyl acrylate(BuA), a MMA/BuA/MMA tri-block copolymer-ization was realized by the addition of these monomers in this order. Various lactones could be polymerized in a living fashion.

EXPERIMENTAL All the operations were performed under argon General Consideration. by using standard Schlenk techniques. Ethylene was used without further purification. 1-Pentene and 1-hexene (Aldrich Chem) were dried over CaH2 and distilled before use. Methyl methacrylate and alkyl acrylates were dried over CaH2 for more than 4 days and the distillate was further dried over activated molecular sieves 3A for 5 days. 1H and 13C NMR spectra were recorded on a JEOL GX-500 or a JEOL GX-270 spectrometer. Gel permeation chromatographic analyses were run on a Tosoh Model SC-8010 using columns TSK gel G1000, (32500, G4000 and (37000 in chloroform at 25°C for poly(MMA) and poly(alky1 acrylate) and run on a Waters 150C with Shodex column (AT8OM) using 1,2,4-trichlorobenzene as eluent at 130°C for polyethylene. The molecular weight of these polymers were determined with the calibration curve obtained using standard polystyrene. Preparation of Bulky Sm(ZZ) Complex. To a solution of CgHgNa (420 ml of a 2.38 M solution, 1.0 mol) was added dropwise 2-bromo-2-methylpropane (181 ml, 1.5 mol). After being stirred for 24 h, the reaction mixture was poured into cold water. The organic product was then extracted with hexane and was distilled under reduced pressure (54°C 35mmHg). Resulting tBuCgH5 (49.6 g, 380 mmol) was added dropwise to a suspension of NaH (14.4 g, 380 mmol) in 250 ml of THF at 0°C. Stirring was continued further for 18 h at room temperature. After the removal of solvent, 300 ml of hexane was added to the residue. To the hexane suspension was added dimethyldichlorosilane (23.1 ml, 190 mmol) and the mixture was stirred for 18 h. After quenching the product with cold water, the resulted product was extracted with hexane and distilled to give Me2Si(3-tBu-CgH3)2 (165"C, 0.1 mmHg). Subsequently n-BuLi (97.2 mmol, 60.8 ml of 1.56 M solution in hexane) was added to a THF solution (200 ml) of Me2Si(3-tBu-CgH4)2. After being stirred for 6 h, trimethylchlorosilane (14.7 ml, 116.6 mmol) was added to the solution and stiriring was continued for 12 h. The mixture was quenched with aq. sodium carbonate and the organic layer was distilled to give Me2Si(2-SiMe3-4-tBuC5H3) (190 - 200"C, 0.1 mmHg). Yield, 26%. To a solution of Me2Si(2-SiMe3-4tBu-CgH3)2 (3.01 g, 6.77 mmol) in 60 ml THF was added 8.2 ml of 1.66 M nBuLi/hexane solution(l3.5 mmol) at 0°C. In order to exchange Li metal with K metal, a THF solution (20 ml) of tBuOK(1.52 g, 13.6 mmol) was added to a THF

22. High M.W. Monodisperse Polymers 239

solution of Me2Si(2-SiMe3-4-tBu-C5H2)2Li2. The mixture was refluxed for 12 h to afford potassium derivative. Resulted THF solution of Me2Si(2-SiMe3-4-tBuCgH2)2-K2 (6.77 mmol) was added to a THF solution (80 ml) of SmI2 (6.8 mmol). The mixture was refluxed for 12 h and THF soluble part was evaporated to dryness. The product was recrystallized from THF/hexane to give Me2Si(2-SiMe3-4-tBuC5H2)2Sm(THF)2. To a solution of CgHgNa (1.2 Preparation of BuZky Sm(ZZZ) Complex. mol) in 300 ml of THF was dropwise added SiMe3C1 (1.2 mol, 152.3 ml) at 0°C. The reaction was carried out for 2 days and then the product was quenched with water. Distillation of haxane soluble part (48 mmHg/59"C) gave 52 ml of Me3SiC5H5 in 45% yield (52 ml). Then to the suspension of NaH (0.47 mol, 18.8 g) in 170 ml of THF was added 52 ml of Me3SiCgHg. After stirring the mixture overnight, the resuting solution was centrifuged to remove the excess NaH. To the mother liquor, Me3SiC1 (0.47 mol, 60 ml) was added at 0°C and the mixture was stirred there for 1 day. After quenching with water, desired (Me3Si)zC5H4 (0.2 mol, 50 ml) was obtained from the hexane soluble fraction (66"C/6 mmHg). The colorless (Me3Si)2C5H4 (48 ml) was then added to NaH (18.8 g, 0.47 mol) in THF (130 ml) at 0°C and stirred there overnight. Me2SiC12(0.1 mol, 14 ml) was added dropwise to the resulting solution and stirred at 60°C for 60 h. Thus Me2Si[(SiMe3)2CgH2]2 was obtained in 52% yield by distillation using Kugelrohr. The resulted ligand (1 ml) was lithiated with BuLi (3.8 mmol, 2.4 ml in hexane) by stirring overnight in THF (10 ml). Then the resulted lithium salt was added to the anhydrous SmC13 (1.6 mmol, 0.42 g) suspended in THF (15 ml) and the mixture was refluxed for 8 h. After evaporation of the mixture to dryness, ether (30 ml) was added to the residue. The ether soluble part was evaporated to dryness and hexane was added. From hexane soluble part, Me2Si[2,4-(Me3Si)2C5H2]2SmC12Li(THF)2 (racemo) was obtained as orange crystals. From hexane [2,3-(Me3Si)2CgH2]SmC12Li(THF)2 (C1 insoluble part, Me2Si[2,4-(Me3Si)2-C5H2] symmetry) was obtained as yellow crystals. Respective products (0.23 g, 0.29 mmol) were treated with LiCH(SiMe3)2 (0.29 mmol) for 24 h and the corresponding alkylsamarium (111) were obtained in 20-30% yield. Other rare earth metal compounds(II1) are prepared in essentially the same method as described previously. Polymerizations of Ethylene and Alkyl Acrylates. Ethylene was bubbled at 1 atmospheric pressure at 25°C to the toluene solution included the catalyst for a fixed time and the resulted polyethylene was quenched with water and then dried. Alkyl methacrylates and alkyl acrylates were polymerized with the catalyst for a fixed time by the addition of alkyl(Meth)acrylsates to the catalyst solution in toluene.

240 H. Yasuda, E. Ihara, S. Yoshioka, M. Nodono, M. Morimoto and M.Yamashita

RESULTS AND DISCUSSION I . Polymerization of Ethylene with Organolanthanide(ZZ) and - ( Z l Z ) Species. 1-21 Ethylene polymerization using the conventional rare earth metal(I1) complexes such as Sm(C5Meg)2(THF)2 proceeds to afford the polymer of Mn = 20,000 with relatively low polydispersity. However, when the molecular weight of resulted polyethylene exceeds Mn = 25,000, the polydispersity increased significantly to Mw/Mn > 2.2. The present polymerization should occur in the following equation by changing the Sm(I1) species to Sm(II1).

Table. 1 The Mode of Polymerization with Sm(C5Me5)2(THF)2 Cocn, mM 1.o 1.o 1.o

Time/min Activity(Kg/ mol h) 10-3 Mn

0.5 1 3

5 10 430 410

1.80 2.28 2.46

Mw/Mn 1.25 1.32 2.28

In order to improve this disadvantage, we have designed new type of Sm(I1) complexes with bulky ligand. The preparation was carried out in the following scheme. The resulted Sm(I1) (racemo) complex was active for polymerization of ethylene and provided the polymer of Mn = 330,000 with low polydispersity (Mw/Mn = 1.6). In the similar manner, meso-type Sm(I1) was synthesized by the following scheme. However, meso-type complex produced the high molecular weight polyethylene with rather broad polydipersity.

22. High M.W. Monodisperse Polymers 241

On the other hand, rare earth metal (111) species such as La(C5Me5)2H is known to exhibit good activity for polymerization of ethylene and conducted the polymerization to give Mn = 670,000 with rather low polydispersity (Mw/Mn = 1.8), while Ln-(C5Me5)2CH(SiMe3)2 is completely inert towards ethylene. The Ln(CgMe5)2H is however thermally unstable and was not isolated as yet. Therefore, we have synthesized thermally stable organolanthanides in the following scheme. Hexane Insoluble

'SiMej

Me3SI Li' ,,@MeQ/JiMe3

THF

0 Me3Si

Me

Me,SI

+

L1'

SrnC1,Me&,,

L

Hexane soluble

I@SIMe3

9

0

D

Mezs&;i>Ll: Meas1

V I The resulted alkylorganolanthanide(II1) complex was revealed by single X-ray analysis to be C1-symmetry (one set of SiMe3 groups locate in meso-like position). The precursor of racemi-type complex was also analyzed by X-ray studies as illustrated below. The Cp'-Sm-Cp' dihedral angle is 117", about 15" smaller than that of the non-bridged Sm(CgMeg)2Me. ,.S iMe3 S ' S iMe3 I \ .SiMe3

n CH2=CH2 c

242 H. Yasuda, E. Ihara, S. Yoshioka, M. Nodono, M. Morirnoto and M.Yarnashita

Fig. 1, Molecular structure of Me2Si-

Fig. 2, Molecular structure of Me2Si-

[2(3),4-(SiMe3)2-C5H2]2SmCH(SiMe3)2 [2,4-(SiMe3)2-CgH2]2SmCH(SiMe3)2 (C1 symmetry)

(racemo)

The results of ethylene polymerizations by using these complexes are summarized in Table 2. Table 2. Characterization of Polyethylene Obtained by Bulky Sm(1I) and Sm(II1) Complexes. Initiator

10-3 Mn MwIMn

Activity (kg/mol h)

Me2Si(2-SiMe34-tBu-C~H2)2Sm(THF)2 (racemo) 356 1.60 139 Me2Si[2(3),4-(SiMe3)2-C5H2]2LaCH(SiMe3)2 498 1.88 80 Me2Si[2(3),4-(SiMe3)2-CgH2]2SmCH(SiMe3)2 (C1) 41 3 2.19 33 Me2Si[2(3),4-(SiMe3)2-C~H2]2YCH(SiMe3)2 (C1) 331 1.65 188 Me2Si[2,4-(SiMe3)2-C5H2]2SmCH(SiMe3)2 (racemo) no polymerization Me2Si[2,4-(SiMe3)2-C5H2]2YCH(SiMe3)2 (racemo) no polymerization 590 1.81 [(C5Me5WaHI 2 2 . Polymerizations of 1 -0lefins with Sm(II) and Ln(IIZ) Complexes. Me2Si(CgHq)(N-tBu)YH and Me2Si(2-SiMe3-4-tBu-C5H2)2YH are known to exhibt high catalytic activity for polymerization of 1-olefins such as propylene, 1pentene and 1-hexene. We have found that bulky Sm(I1) such as MezSi(2-SiMe3-4tBu-CgH2)2Sm(THF)2 (racemo) exhibits good intiating property for polymer-

22. High M.W. Monodisperse Polymers 243

ization of 1-pentene and 1-hexene. Especially noteworthy is the formation of highly isotactic poly( 1-olefin) by this catalyst. Isotacticity exceeds over 97%. In Figure 1, 13C NMR spectrum of the resulted poly(1-pentene) is shown. The C3 signal appeared as singlet peak to indicate the formation of isotactic polymer. On the other hand, Me2Si[2,4-(SiMe3)2-C5H2][3,4-(SiMe3)2-CgH2]YCH(SiMe3)2 (C1 symmetry) also exhibits good catalytic activity. However, this initiator provides atactic poly( 1-pentene) or poly( 1-hexene) (Table 3). Table 3. Characterization of Poly(1-olefin) Prepared by Bulky Sm(I1) and Y(II1) Complexes. Monomer 1-pentene

1-hexene

Initiator

1 0 - 3 ~Mw/Mn ~

Me2Si(2-SiMe3-4-tBu-CgH2)2Sm(THF)2 13 Me2Si[2(3),4-(SiMe3)2-CgH2]2YCH(SiMe3)2 16 Me2Si(2-SiMe3-4-tBu-C5H2)2YH 20 Me2Si(2-SiMe3-4-tBu-CgH2)2Sm(THF)2 19 Me2Si[2(3),4-(SiMe3)2-C5H2]2YCH(SiMe3)2 64 Me2Si(2-SiMe3-4-tBu-CgH2)2YH 24

1.63 1.42 1.99 1.58 1.20 1.75

3. Living Polymerization of Methyl Methacrylate3-5 1 Polymerization of methyl methacrylate with organolanthanide(II1) complexes were performed with SmH(C5Me5)2, LnMe(CgMeg)2(THF) (Ln = Y, Yb, Sm) or Ln(CgMe5)2Me2AlMe2 (Ln = Sm, Yb, Lu). As a typical example of polymerization, SmH(C5Meg)2 initiated polymerization is summarized in Table 4.

(6 38.0.mmmm)

1 I

,

m PPM

40

30

20

10

Fig. 3 13C NMR spectrum of poly(1-pentene)

244

H. Yasuda, E. Ihara, S . Yoshioka, M. Nodono, M. Morimoto and M.Yamashita

These reactions involve the following marked features. 1) The polydispersity of resulting polymers are unusually low and reach Mw/Mn=l.O3. 2) Polymerization proceeds very rapidly and is complete in a short period with high initiator efficiency. 3) The polymerizations gave high molecular weight poly(MMA) especially when the concentration of the complex was lower than 0.2 mol%. 4)The polymerization proceeds at wide range of reaction temperature starting from +40 to -95°C. 5) Highly syndiotactic polymerizations occur at lower temperature. Table 4. Characterization of Poly(MMA) Synthesized by SmH(CsMe5)2 ~~

~

Polymerization MMA/initiator Temp. ("C) charged, mol/mol

40 25 0 0 -78 -95

500 500 1500 3000 500 lo00

10-3 Mn

55 57 215 563 82 187

Mw/Mn

1.03 1.02 1.03 1.04 1.04 1.04

IT

conversion %,(reacn. h)

77.3 79.9 82.6 82.3 93.1 95.3

99(1) 99(1) 93m 98(3) 97(17) 82(60)

rr; syndiotacticity These results indicate that the present polymerizations proceed in a living fashion. In fact, the Mn of polymers increased linearly in proportion to the conversion irrespective of the initiator concentration, while M w/Mn remains intact during the polymerization. Consequently, we can readily estimate that the present polymerization occurs in a living fashion. To get further insight into the initiation mechanism, we have demonstrated the stoichiometric reaction at 0°C between SmH(CjMe5)2 and MMA. As a result the 1:2 adduct was obtained as single crystals (mp 132°C). This adduct is active for polymerization of MMA and produced the polymer of Mn = 110,000 (Mw/Mn = 1.03) when 100 equivalents of MMA was added. Deuterolysis of the adduct gave DCMe(C02Me)CH2C(Me)2CO2Me to indicate the formation of Sm-enolete or Sma-carbon bond. To verify the exact structure of SmH(MMA)2(CsMe5)2, single crystal X-ray analysis was performed. The adduct has an eight-membeed ring structure. The enolate group bears a cis configuration and binds with another MMA molecule which coordinates to the metal with its ester group.

22. High M.W. Monodisperse Polymers 245

n

v Fig. 4 X-ray analysis of SmH(MMA)2(C5Me5)2 These results indicate that, in the initiation step, the hydride should attack the CH2 group of MMA to generate a transient Sm-O-C(OCH3)=C(CH3)2 species, and then the incoming MMA molecule may participate in a 1P-addition to afford the eight memberted ring intermediate. Then in the propagation step another MMA molecule may attack the growing end, liberating the coordinated ester group. Syndiotactic polymerization should occur by repeating these reactions, where the coordination site changes alternatively. 4. Living Polymerization of Alkyl Acrylates Living polymerizations of methyl-, ethyl-, and butyl acrylate have not been achieved since their acidic a-H easily takes place nucleophilic addition reaction. However, living polymerizations proceed by the unique function of organolanthanide complexes such as SmMe(CgMe5)2(THF) and YMe(C5Me5)2(THF). The results are shown in Table 5. In these cases, living polymerization gave atactic polymers.

Table 5. Polymerization of Alkyl Acrylate with SmMe(CgMeg)2(THF) Monomer

10-3 Mn

Mw/Mn ~

Methyl acrylate Ethyl acrylate Butyl acrylate

55 65 88

Conversion/% ~~~

1.04 1.04 1.04

99 94 99

246 H. Yasuda, E. Ihara, S . Yoshioka, M. Nodono, M. Morimoto and M.Yamashita

The rate of polymerization of alkyl acrylates increases in the order Bu > Et > Me, by sharp contrast to the order Me > Et > Bu observed in the cases of alkyl methacrylates. Table 6. Properties of Tri-block Copolymers System

Tensil Tensile Elongation Compression % % modulas(MPa) strength(MPa)

Poly(MMA/BuA/MMA) (2551 :24) (8 :72: 20) Poly(MMA/EtA/EtMA) (26:48: 26) Poly(MMA)

46 0.8 119 610

22 0.7

81 163

100 58

22 80

276 21

62 100

BuA , n-butyl acrylate; EtA, ethyl acrylate; EtMA, ethyl methacrylate

-, ~ C - C H ~ ~ C - C H ~ ~ ~ C - C H ~ ~ ~ MMA H+

poly(MMA-BuA-MMA)

Me

H

Me

C02Me I C02Me

fl

,2

C02Me

/, t

i

\

poly(MMA)

I 1

i

j

!

20.0

30.0

Time (min)

Fig. 5 GPC profilesof mono-, di- and triblock copolymers

22. High M.W. Monodisperse Polymers 247

As a result, poly(MMA/BuA/MMA) tri-block copolymer in the ratio of 8:72:20 and poly(MMA/EtA/EtMA) in the ratio of 26:48:26 showed the big elongation and small compression to indicate that these polymers exhibt rubber-like elastic property. The GPC profile of the mono-, di and tri-block polymers are shown below.

5. Living Polymerization of Lactones. As an extention of the presen work, we have examined the polymerization of lactones such as P-propiolactone (PL), E-valerolactone (VL) and E-caprolactone (CL) and foud that these systems proceeds the living polymerization. The resuls are shown in Table 7.

The addition of one equivalent lactone to YOMe(C5H5)2 resulted in the formation of 1:l adduct, which gave upon hydrolysis the original lactone, while lactone polymerization starts by the addition of two equivalent of lactone and one equivalent capric acid was obtained by hydrolysis. This result clearly indicates that 0-acyl bond cleavage occur in the propagation step. The M n increases in proportion to the conversion, while molecular weight distribution remains intact.

5

1.00 I

0

I

25

50

75

100

Conversion (O/o)

Figure 6. M n and M / M n vs. conversion for polymerization of caprolactone with YOMe(CgRg)2

248 H. Yasuda. E. Ihara, S . Yoshioka, M. Nodono, M. Morimoto and M.Yamashita

Table 7. Living Polymerization of Lactones. Initiator [YOMe(C5H5)212 SmMe(CgMeg)2(THF) [YOMe(C5H5)212 SmMe(CgMeg)(THF)

Monomer PL VL CL CL CL

10-3 Mn

60 75 130 39 33

M w/Mn

1.13 1.07 1.06 1.04 1.03

Conversion/% 78 80 95 52 87

References 1) H. Yasuda and H. Tamai, frog. folym. Sci., 18, 1097 (1993). 2) H. Yasuda and E. Ihara, J. Synrh. Org. Chem. Jpn., 51, 931 (1993). 3) H. Yasuda, H. Yamamoto, K. Yokota, S . Miyake, and A. Nakamura, J. Am. Chem. SOC.,114,4908 (1992). 4) H. Yasuda, TH. Yamamoto, Y. Takemoto, M. Yamashita, K. Yokota, A. Nakamura, S. Miyake, Y. Kai, and N. Kanehisa, Macromolecules, 26, 7134 (1993). 5 ) H. Yasuda, H. Yamamoto, Y. Takemoto, M. Yamashita, K. Yokota, S . Miyake, and A. Nakamura, Makromol. Chem. Macromol. Symps. 67, 187 (1993). 6) H. Yasuda, M. Furo, H. Yamamoto, A. Nakamura, S . Miyake, and N. Kibino, Macromolecules, 25,5115 (1992).

249

23. Lanthanocene Based Catalysts for Olefin Polymerization : Scope and Present Limitations

J. F. PELLETIER, A. MORTREUX, F. PETIT Laboratoire de Catalyse hCi4rogPne el homoghe, URA CNRS 402, USTL, ENSCL. BP 108.59652 Villeneuve d h c q Cedex (France)

X. OLONDE AND K. BUJADOUX E.C.P. EniChem Polymeres France, Cenire de recherche. 62670 Mazingark (France)

ABSTRACT The ethylene polymerization has been studied over pentamethylcyclopentadienyl based neodynium and samarium catalysts under various conditions,ranging from low temperature - low pressure (1 atm - 20°C) to those used in a high temperature - high pressure pilot plant (180°C - 1200 b). The catalyst remains stable, but attempts at copolymerization with 1-butene have failed, even in the presence of an ylide as modifier. A comparison with the more conventional Cp2ZrCI2 - MA0 catalyst shows that these lanthanocene catalysts, although more reactive, are not able to copolymerize ethylene with a-olefins under industrial conditions.

INTRODUCTION Olefin insertion into metal carbon bonds and p hydrogen elimination are fundamental reactions occuring in Ziegler-Natta catalysis. The characterization of the active sites is however complicated by their multicomponent composition. Several years ago, tremendous advances in lanthanides and group 3 element chemistry provided well defined alkyl metal complexes which served as excellent models for mechanistic studies. Watson [ 11, Marks [2], Bercaw [3] and Teuben [4] synthesized hydrides and alkyls lanthanocenes, scandocene and yttrocene, which are highly active in polymerization but are also extremely sensitive towards impurities like oxygen and moisture. In general, to be used in industrial processes, Ziegler-Natta catalysts contain a slight excess of cocatalyst (alkyl reagent) as scavenger. We have recently reported the possibility to produce polymerization catalysts by direct alkylation between Cp*2NdC12Li(OEt2)2 1 and common alkylating reagents [ 5 ] . Their behaviour at high temperature was also examined. In this paper, we wish to report some data obtained on other lanthanocene based catalysts and discuss about their behaviour in copolymerizations with olefins.

250 J.F. Pelletier, A . Mortreux, F. Petit, X. Olonde and K . Bujadoux

EXPERIMENTAL All reactions were done with dry solvents under nitrogen. The complexes Cp2*LnC12Li(OEt)2 [6], Cp2*NdCH(SiMe3)2 [2], [Cp2YC1]2 [7] and the ylide [81 were synthesized as previously described in the litterature. AtmosPheric DRSSUE te StS. A double envelope 1 liter flask is dried at 12OOC and purged three times with nitrogen. The flask is then decontaminated from moisture by 500 ml of a 10-2M butylethylmagnesium (BEM) solution in Isopar L (high boiling point saturated hydrocarbons fraction) for 1 hour at 80°C. After evacuation, 500 ml of dry Isopar L are introduced, and saturated with ethylene at 80°C. The catalyst, previously obtained from reaction of the precursor with BEM for 1 hour at ambient temperature in toluene, is then added and the ethylene consumption followed with flowmeters. At the end of the reaction, 10 ml of ethanol are introduced. The polymer is precipitated with a large quantity of ethanol or isopropanol, filtered on a sintered glass, washed with n-heptane and dried in an oven at 80°C for 48 hours.

High temperature-low pressure tests The autoclave (1 liter capacity) is monitored with an external heating allowing to reach 250°C. A constant ethylene pressure of 6 bar is applied and the ethylene flowrate can be varied from 100 to 3000 l/h. A mechanic stirrer rotating at speeds up to 1500 rpm is used. Rotameters allow an accurate measurement of the ethylene consumption vs. time (1 min reaction). Before the reaction, the autoclave is decontaminated with an Isopar L solution of BEM at 16OOC for 45 min. After evacuation, 600 ml of dry Isopar are introduced and saturated with ethylene. The catalyst (10 ml of a 10-2M solution) is injected rapidly via a sas with a nitrogen overpressure of 8 bar. High temperature - high pressure te StS This apparatus has been initially developed by Cdf Chimie [9] and consists of a pilot plant where the operating conditions are very close to the industrial ones. The reaction conditions can be adjusted from 160 to 280°C with a pressure range 600-2000 bar under dynamic conditions. At these high pressures, the reaction medium is homogeneous under supercritical conditions. The temperature of this adiabatic reactor is adjusted and regulated by the catalyst solution flowrate. The average residence time is 40 sec, which allows conversions from 10 to 20%and needs a recycling of the unreacted monomers.

23. Lanthanocene Based Catalysts for Olefin Polymerization 25 I

RESULTS AND DISCUSSION Generalization of ethylene homoDolynerization on lanthanocene based catalvsts At it can be seen in table 1, the catalytic systems consisting of 1 or 2 and BuMgEt

polymerize ethylene with high activity at 80°C. The kinetic and the molecular weights depend on the Mg/Nd ratio and on the polymerization temperature. The initial activity decreases with an increase of Mg/Nd ratio, an induction period being even observed for a Mg/Nd ratio of 20. At low temperature (0°C) the catalytic system is only slightly active and requires a low magnesium content : a Mg/Nd ratio of 10 is sufficient to prevent any polymerization. The polydispersities are much broader than those obtained at higher temperature and the GPC curves show a bimodal distribution. Nevertheless, at 80°C a narrow distribution is controlled by the fast PH elimination chain transfer. Table 1. Temperature effect on ethylene polymerizationa

Catalysts

-

-

1530

7200

4.1

660

2880

1220

2.5

Polymerization Mg/Ln Yield temperature ("C) ratio dmmo1.h.b

Cp*2SmC12Li(OEt2)22

Ob

10

Cp82SmC12Li(OEt2)22

Ob

I

Mw

-

Mw&

0

0.34

Cp*2NdC12Li(OEt2)2 1

80

Cp*2NdC12Li(OEt2)2 1

80

20

230

1800

2920

1.6

Cp*2SmC12Li(OEt2)2 2

80

10

660

1830

2200

1.2

2.5

a Catalysts and BuMgEt were mixed at 20°C for 1 h. Polymerizaiion conditions: P C ~ =H1 bar; ~ [Ln] = 0.4 mmol/l; solvent = Isopar L (500 rn11.b solvent = lolucne (100 m ~ )

As previously shown in our first paper [ 5 ] the amount of cocatalyst is a crucial factor determining the activity of the catalyst. This economical and convenient method for the preparation of the polymerization catalysts has been extended to other lanthanides and yttrium. In table 2 are reported the results obtained at high temperature with several complexes activated by butylethylmagnesium. The yield for the yttrium catalyst is lower than those obtained with neodymium and samarium, a result which can be correlated with the ionic radius [ 2 ] .Only the ytterbium catalyst gave very low activities. This may be related to its reduction into YbII species.

252 J.F. Pelletier, A. Mortreux, F. Petit, X. Olonde and K. Bujadoux

At 160°C. in contrast to the results obtained at low temperature, the effect of the M@d ratio is much less pronounced. All polymers have about the same molecular weight - (k loo0 and Mw/Mn = 1.5). A Mg/Nd ratio of 50 is needed to observe a drop of the initial rate constant (kp). This effect could be explained by the fact that the excess of BuMgEt could react with the fourteen electron active species, via the formation of 3 centers, 2 electrons bridge, leading to a bulky adduct in which the orbital required for olefin complexation is occupied. This interaction must be broken by thermal activation in order to give back the active species as depicted in the following equilibrium (eqn 1)

-

Table 2. Ethylene polymerization : comparison between several catalytic systemsa

M a n ratio

Yield g/mmo~.mn.mo~ 1-1 of C2H4

kP mol/l.s

Cam1ysts

Polymerization temperature ("C)

Cp*2NdC12Li(OEt2)2 1

160

3

1600

1 100

Cp*2SmC12Li(OEt2)22

160

3

1600

1100

Cp*2YC12Li(OEt2)24

160

3

1100

1070

Cp*2YbC12Li(OEt2)25

I60

3

0

0

Cp*2YbC12Li(OEt2)2 5

I20

3

190

150

Cpf2NdCI2Li(OEt2)2 1

160

50

920

700

Cp*2NdCH(SiMe3)2 3

160

0

1160

620

Cp*2NdCH(SiMe3)2 3

160

3

1870

1880

aThe alkylation by BuMgEt was carried out at 20°C for 1 h; polymerization conditions: P ( c ~ H ~=)6 bar; [Ln] = 0,2 mmol/l; solvent = Isopar L (600ml).

23. Lanthanocene Based Catalysts for Olefin Polymerization 253

The bulky, well defined complex Cp*2NdCH(SiMe3)2 3, which does not polymerize ethylene at low temperature and atmospheric pressure [2], is a good catalyst at high temperature (160OC) but has a particular behaviour: the initial rate constant with 3 is about the same as with the catalytic system based on 1 and 50 equivalents of BuMgEt. Indeed the kinetic polymerization curve (fig. 1) shows that the ethylene consumption increases to reach a maximum and decreases, indicating that the first ethylene insertion step is very slow. AS already shown by Bercaw et al. [3], this is probably due to the bulkiness of the alkyl ligand -CH(SiMe3)2 or to the non bonding interaction between the lanthanide and SiMe3 [2]. 4 -

catalysts kp 1 + 3eqBEM 1100

+ 50 eq BEM

I

4 3 without BEM

5

10

15

20

25

30

35

40

45

50

55

700 620

60

Time (s)

Figure 1. Polymerization kinetics at 160OC - 6 bar for catalysts 1 and 3. It can be noticed (table 2) that the highest initial rate constant is achieved when complex 3 is mixed with 3 equivalents of BuMgEt. To explain this and also the fact that, at low temperature (8O0C), the molecular weight decreases at high M o d ratio,as it has already been shown in our previous report [ 5 ] , an alkyl transfer reaction occuring between the neodymium and the magnesium compounds can again be involved (eqn 2). Cp*2Nd CH(SiMe3)?+ MgR2

20°C ------=

"Cp*ZNdR" + RMgCH(SiMe&

(2)

3 The reactivity of 3 towards alkylmagnesium and alkylaluminium has been followed by lH NMR in C6D6. The spectrum obtained with BuMgEt is very complex, but the signal corresponding to the -CH(SiMe3)2 group is shifted from -16 ppm to -0.3 ppm. A similar behaviour is observed with (AIMe3)2 (eqn 3). The reaction appears to be much faster and the complex Cp*zNd(pMe),AIMe2 6 (characterized by microanalytical data) cnstallizes from the solution.

254

J.F. Pelletier, A. Mortreux, F. Petit, X. OIonde and K. Bujadoux

Me.

20°C

c ~ * ~ N d c H ( S i M+e (AlMe,), ~)~

/ \AlMe2 + Me2A1CH(SiMe3)2 Cp*2Nd \ /

(3)

Me

6

However, complex 6 does not polymerize ethylene even at low temperature and under ethylene pressure (70 bar) in contrast with the homologous yttrium complex 141.

mod ifier W

Effect of an

t

. .

s at coDolvmemion with I - b u m

The use of these systems for ethylene- 1-butene copolymerization has already shown that no copolymerization occured in a pilot plant under industrial conditions (1200 bar - 20OOC) [ 5 ] .With the aim to find a catalytic system suitable for copolymerization, attempts have been made to modify the lanthanocene complex by a ligand exchange with the ylide Ph3P=CH-C(O)Ph. Table 3 compares the results observed without and in the presence of 1 equivalent of this ligand at 160OC. Table 3. Comparison of activities of binary systems Cp2*LnC12Li(OEt2)2 + BEM and ternary systems with the ylide Ph3P=CH-C(O)Ph for ethylene homopo1ymerization.a

Ln

kP mol. 1-1 s-1

Nd Sm Ndb Smb a Polymerization conditions : PC,H,

Yield g/mmol.min.mol.l-]of C2H4

1100 1100 3 380 4 900

1500 1600 4 150 5 200 = 6 b [Ln] = 0,1 mM ; M g L n = 3 ; To= 160°C ; 1 min reaction. b

Addition of 1 eq of the ylide for 1 hour at 20°C bcfore the addition of Bu Mg Et.

Due to this enhancement of activity, this new catalytic system has been also tested at high temperature and pressure for ethylene- 1 -butene copolymerization. The results have been compared with the yttrocene and ziconocene catalysts (Cp2YC1)2 and Cp2ZrC12 where only cyclopentadienyle groups are present (table 4).

23. Lanthanocene Based Catalysts for Olefin Polymerization

255

Table 4. High temperature - high pressure ethylene- 1-butene copolymerization testsa. Catalyst

cocatalyst

yield

(ratio)

kdrnrnol

-Mn

Mnmn

Density

vinyF vinyiidenec intemaF

1.8 0.961 0.5 22.6 C~*NdC12Li(OEt2)2~ BEM (3) 23 700 2.3 >0.960 0.4 (CP2YC1)2b BEM (3) 14 7 100 cp2zrc12 M A 0 (100) 8 16 800 3.2 0.9395 0.37 apolymerization conditions: Pressure = 1200 bar; temperature = 180°C; but-I-ene 40% ; b lcq PPh3=CHC(O)Phadded before alkylation. double bonds per loo0 carbons.

-

0

0.05

0.03

0.78

0.05

0.17

Although the neodymium catalyst is more reactive, no copolymerization was observed. The less sterically crowded (Cp2YC1)2 isoelectronic system did not give any copolymerization as well. The infrared polymer analysis shows the presence of vinylic double bonds which are produced by the classical PH elimination process after ethylene insertion. However, vinylidene (CH2=CHR) and internal (-CH=CH-) double bonds are present for the yttrium and zirconium systems, indicating that the comonomer insertion step is possible in these two cases. As compared with zirconium, the density of the polymer produced with yttrium is much higher and the molecular weight lower : I-butene acts as an efficient transfer reagent by PH elimination after primary and secondary insertions as described in scheme 1. Surprisingly, the yttrium catalyst induces more secondary than primary insertions which are usually found with Ziegler Natta type catalyst.

[Nd+Fl

CII3-CH2-CtFCH-CH2-@

INSERTION RUTTKANSFER

Primary insertion

CHrCHl CH.$ Ctl*

Scheme 1. Chain transfer reactions in ethylenebut-1-ene copolymerization

256

J.F. Pelletier, A . M o r t r e u x , F. Petit, X. O l o n d e a n d K. B u j a d o u x

CONCLUSION At least in high temperature-high pressure conditions, these bis Cp* lanthanocene and bis Cp yttrocene based catalysts are not suitable for ethylene - a olefin copolymerization, although their productivity for ethylene homopolymerization is tremendously high considering the yield obtained per overall (catalyst + cocatalyst). To achieve this goal, further work should be done in this area, perhaps via ligand modification(s), to provide catalytic systems which could substitute the zirconocene catalysts in a useful way, that is to reduce the amount of cocatalyst generally required in this fascinating chemistry.

References 1.

P.L. Watson and G.W. Parshall. Acc. Chem. Res., 18. 5 1 (1985).

2.

G. Jeske. H. Lauke, H. Mauermann, P.N. Swcpston, H. Schumann and T.J. Marks, J . Am. Chem.

SOC.,107, 8091 (1985). 3

B.J. Burger, M.E. Thompson, W.P. Cotter and J.E. Bcrcaw, J . Am. Chem. SOC.,112, 1566 (1990).

4

K.H. Den Haan, Y. Wiclstra, J.J.W. Eshuis and J.H. Tcubcn, J . Orgnnomcr. Chcm., 323, 181 (1987).

5.

X . Olonde. K. Bujadoux, A. Mortreux and F. Petit, J . Mol. Caial., 8 2 , 7 5 (1993).

6.

T. don Tilley and R.A. Andersen. Inorg. Chem., 20,3267 (1981)

7.

W.J. Ewans. J.H. Meadows, A.L. Wayda, W.E. Hynter and J.L. Atwood, J. Am. Chem. SOC.. 104,2008 (1982)

8.

F. Ramirez and J. Dershowitz, J. Org. Chem., 22.43 (1957)

9.

J.P. Machon, "Transition Mctal Catalyzed polymerizations Zicglcr-Natta and Metathesis Polymerization", Cambridge University Prcss, R.P. Quirk Ed., Cambridge, 1988. p. 344.

257

24. Effect of Ligand and Inorganic Support on Polymerization Performances of Ti and Zr Catalyst

F. CIARDELLI, A. ALTOMARE, G. ARRIBAS*, G. CONTI Department of Chemistry and Industrial Chemistry, University of Pisa, via Risorgimento 35, 56126 Pisa, Italy. *Escuela de Quimica, Facultad de Ciencias, Universidad Central de Venezuela, Caracas, Venezuela F. MAS1 AND F. MENCONI EniChem, 20097 S.Donato Milanese, Italy

ABSTRACT In the present work the activity was discussed of various soluble complexes of Ti and Zr with phenolate, carboxylate, unsubstituted and variously substituted cyclopentadienyl ligands in ethylene and 1-olefins homo- and copolymerization after activation with aluminum alkyls or MAO. Some of these complexes were also used for preparing catalysts supported on inorganic materials such as silica or zeolites after modification of surface functionality. The discussion of the results takes into accounts steric and electronic effects of ligands and support which allow to modulate catalyst performances INTRODUCTION The chemical mechanism by which metallocene complexes of IV group transition metals can catalyze with great efficiency monoalkenes polymerization is now generally accepted as based on the formation of stable cationic complexes.1 In these last the transition metal bears in addition to a o-alkyl group, a positive charge

258

F. Ciardelli, A . Altomare. G . Arribas. G . Conti, F. Masi a n d F. Menconi

having transferred one electron to a ligand which becomes part of complex anion assisted by a cocatalyst such as alumoxane2-4 or another more conventional Lewis acid.5 Both electron transfer and reactivity of the cation are dependent on the transition metal environment provided by ligands and cocatalysts. Even if many outstanding contributions appeared already in the chemical literature putting lights on these aspects, several points remain still to be clarified. These last refer in particular to the role of aluminum alkyls and alumoxane as well as to the possibility of heterogeneizing the above systems without detracting their properties. In this broad context one objective of this work is to contribute additional evidence on the role of the ligands in combination with cocatalyst in determining productivity of ethylene polymerization. On the other side these aspects are also used in order to develop supported cataly~t6~7 with comparable or improved characteristics in respect to their homogeneous precursors to be used in slurry and gas-phase processes. EXPERIMENTAL Materials. All reactions were carried out under argon atmosphere. Solvents were dried over calcium chloride and then freshly distilled under argon from sodium-potassium benzophenone ketyl. Trimethylaluminum, diethylaluminuni chloride, triethylaluminum, triisobutylaluminum and M A 0 (4.5 M in toluene) (Witco), titanium alkoxides, titanocene dichloride, zirconocene dichloride (Aldrich) were used as received. Titanocene and zirconocene dimethyl derivatives were synthesized by literature methods.8 [Pyrocatecholate(2)]titanium(IV) dichloride, cyclopentadienylpyrocatecholatetitanium(1V) chloride, bisnonanoatetitanium(1V) dichloride, pentamethylcyclopentadienyltitanium(1V) trichloride were prepared by methods described earlier.9 1H and 13C-NMR spectra were recorded by using a Varian Gemini 200 spectrometer. Bis~entamethvlcvclopentadienvlzirconium(IV) dichloride. 3 g (12.35 mmol) of ZrC14 freshly sublimed and 4.3 g of Cp*K were added at -80°C in 100 ml of dry toluene. This mixture was warmed to 2 5 T , then refluxed for three days. Solvent was removed under reduced pressure and the solid yellow-green residue was taken up in 200 ml of chloroform. Petroleum ether (150 ml, 90-100°C) was added and the solvent slowly removed by rotary evaporation. The residual solution

(50 ml) was cooled, and the product was filtered off and washed with cold

24. Effect of Ligand and Support on Polymerization Performances 259

petroleum ether, yield 4 g ( 75%) of pale yellow crystals. Anal. calcd. for C20H30C12Zr: C, 55.56; H, 6.94; C1, 16.40. Found: CS5.20; H, 6.83; C1, 16.95. 1HNMR (CDC13): 6 = 1.98 ppm (s); 13C-NMR(CDC13):6 = 11.95 and 123.52 ppm. Bis-tetraDhenvlcvcloDentadienvlzirconium ( I W dichloride. See ref. 10 Silica treatment. Croxfield-type silica (surface area 300 m2/g; [OH] = 2-10-3 moles/g) was heated for six hours at 300°C under vacuum (0.05 mm). 0.8 ml of a 1.6 M solution of LiMe in diethyl ether were added SiO7-LiMe. dropwise at -78°C to a suspension of 3.04 g silica in freshly distilled THF over a period of 1 hour. The temperature was allowed to rise at 20°C and then the silica was washed with THF, the solvent was evacuated and the solid was finally dried under vacuum for three hours. 45.7 ml of a solution 0.15 M of MgC12 in THF were added SiO7-MgCI7. dropwise under magnetic stirring at room temperature to 30.6 g of silica in 60 ml of dry THF. The solvent was evacuated and the solid was dried under vacuum (0.05 mm) at 75°C. Method 1 /N2TiC12-Si02-MgC12): 2 Preparation of supported catalvsts. ml of a solution 0.26 M in n-heptane of bisnonaoatetitanium(1V) dichloride (N2TiC12) were added to a suspension of 5.2 g of Si02-MgC12 at the room temperature under magnetic stirring. The reaction mixture was stirred over a period 1 hour; then the solid product was washed with n-heptane. Method 2 (Cp2TiC12-Si02-LiMe): 3.1 ml of a 0.098 M THF solution of Cp2TiC12 were added to a suspension of 4.0 g Si02-LiMe in dry THF at room temperature under magnetic stirring. The solid catalyst was washed with THF and then with methanol. The solvents were removed under vacuum (0.05 mm). Method 3 (Cp2ZrC12-HY**AlMe3): 0.92 ml of a 0.05 M solution of Cp2ZrC12 were added to a slurry of 2.5 g HY**-AlMe3 in toluene under magnetic stirring at the room temperature. The reaction mixture was stirred for three days and then washed with toluene. A HY type zeolite with Si/AI ratio of 7.25 was heated Zeolite treatment. at 300°C under vacuum (0.05 mm) over a period of six hours. 100 ml of a 38 70 Exhaustive dealumination of the zeolite (HY **). solution of acetylacetone in methanol were added to 12.4 g of HY zeolite. The suspension was refluxed under magnetic stirring for 12 hours, then the solid product was filtered . This treatment was repeated three times, then the solid was calcinated at 600°C under a stream of dry air for six hours.

260

F. Ciardelli, A. Altomare, G . Arribas, G. Conti. F. Mas; and F. Menconi

HY **-AIMeJA

1 ml of 2.0 M solution of AlMe3 in toluene was added to a

toluene slurry of 1.5g of dealuminated zeolite and the reaction mixture was kept under magnetic stirring for three days at room temperature; the solid was washed with toluene until AlMe3 elimination. HY**-Nia-AlMeT. 2.5 ml of a 0.04 N solution of nickel (11) nitrate were added under magnetic stirring to a slurry of 2.142 g HY** in water. The solid product was filtered, heated at 300°C for 20 hours and then suspended in 30 ml of freshly distilled toluene. 1 ml of 2.0 M solution of AlMe3 was added at room temperature to the resulting suspension and the reaction mixture was stirred for 4 hours and then washed with dry toluene. HY **-MeSiC1. 20 ml of distilled Me3SiC1 were added to 3 g of HY** and the mixture was refluxed overnight. The excess Me3SiC1 was eliminated under vacuum (0.05 mm). Polvmerization experiments. Ethylene polymerization experiments were carried out by introducing the cocatalyst solution in toluene and the catalytic slurry in the same solvent into the reaction vessel under argon atmosphere. After 10 minutes ageing ethylene was introduced and its partial pressure kept at 1 bar during polymerization time. RESULTS AND DISCUSSION Soluble catalvsts. A possible correlation between ligands of the original transition metal complex and cocatalyst was initially investigated by examining the productivity per g atom of transition metal (SA) or per g atom of transition metal and g atom aluminum @A*), obtained for different Ti or Zr complexes with aluminum alkyls or MAO, respectively.9 The maximum values obtained and the corresponding conditions are reported in Table 1. In the case of titanium, when Cp is not present AlEt3 and M A 0 give SA of the same order of magnitude, but higher Al/Ti ratios are necessary for the latter. In any case the A l n i ratio for optimum productivity is much lower than for CpzTiX2 which however provides much larger activity per Ti atom (SA). Even if the presence of Cp ligandl 1 substantially activate the complex versus MAO, SA* remains usually lower than with AIEt3. Cp*TiC13 has an unusual behaviour and is always more active with AIMe3 rather than MAO. This is substantially in agreement with an excessive Ti reduction with the massive M A 0 necessary.

24. Effect of Ligand and Support on Polymerization Performances

261

Table 1 Polymerization of ethylene by soluble titanium complexes activated with different aluminum derivative@ RUIl

H1

H2 H3 H4 H5 H6 H7 H8 H9 H10 H11 H12 H13 H14 H15 HI6 HI7 H18

Catalyst Ti(n-OBu)4 g Ti(n-OBu)4 g Ti(2-EH)4 l1 Ti(2-EH)4 l1 PcTiC12 PcTiC12 CpPcTiCl j CpPcTiCl J N2TiC12 N2TiC12 Cp*TiC13 I Cp*TiC13 I Cp2TiCl2 Cp2TiC12 cp2zrc12 cp2zrc12 Cp*2ZrC12 O (Cp$4)2ZrC12 P

Cocatalyst AIEt2CI MA0 AIEt2CI MA0 AIEt3 MA0 AlEt3 MA0 AlEt3 MA0 AIMe3 MA0 AIEt2C1 MA0 AlEt3 MA0 MA0 MA0

[AlI/[Mtl 6.0 200.0 5.O 30.0 1.5 100.0 9.0 100.0 4.0 100.0 8.0 200.0 2.5 3000.0 4.0 1500.0 1500.0 3000.0

S.A.b

S.A.*C

0.1 0.7 0.9 0.8 5.7 7.2 0.4 33.4 5.8 16.7 7.2 1.1 3.7 300.0 -

0.01 3 0.004 0.180 0.027 3.800 0.072 0.040 0.330 1.450 0.167 0.900 0.005 1.480 0.100

372.5d 185.6e 80.4f

0.248 0.124 0.010

-

a Optimum productivity at Pethylene = 1 bar, T = 25 "C. b Kg PE/g atom Mt-h. Kg PE/g atom Mt-g

atom Al-h. Mn*10-' = 0.35, Mw-10-5= 0.78; e q = 7.5 in decaline at 135OC; fMII.10-5 = 0.04,

Mw. lo-' = 0.17. g Tetrakis-(n-butanolate)titanium(IV).hTetrakis(2-ethylhexanolate)titaniurn(IV). [Pyrocatecholate(2)] titanium (IV) dichloride. 1 Cyclopentadienylpyrocatecholatetitanium(IV) chloride. Bisnonanoatetitanium(1V) dichloride. 1 Pentamethylcyclopentadienyltitanium(1V) trichloride. Biscyclopentadienyltitanium(1V) dichloride. '1 Biscyclopentadienylzirconium(1V) dichloride. 0 Bispentamethylcyclopentadienylzirconium (IV) dichloride. P Bistetraphenylcyclopentadienylzirconium (IV) dichloride.

The characterization of the catalytic systems and their kinetic behaviour as well as polymer features suggest that the polymerization mechanism is substantially the same for all examined transition metal complexes in the presence of either aluminum alkyls or alumoxane. The formation of active species should consist of

262

F. C'iardelli, A . Altomare. G. Arribas. G . Conti, F. Masi a n d F. Menconi

alkylation of titanium by replacement of chlorine atoms with alkyl groups and consequent formation of unsaturated active species.*2 When Cp ligands are not present, AIR3 and M A 0 give more or less the same productivity, and a lower amount of the former is requested. Such behaviour can be tentatively interpreted by considering that AIR3 is the actual cocatalyst and therefore a consistent amount of M A 0 is necessary to attain the optimal AlR3/Ti ratio. When the Cp ligand is present,l3) M A 0 becomes more effective in giving high productivity at rather high MAO/Ti ratios, thus suggesting that, at least at that concentration level, a different activation mechanism can be operative. The different polymerization rates and activities observed when changing catalyst can arise from different propagation rate constants (k,) and/or active sites concentration ([C*]). These differences are a result of the effect of ligands on the reaction between titanium complexes and cocatalyst, which determines [C* I and steric and electronic effects on the Ti-C bond, which in turn affect the k, value. In the case of MAO, stabilization and activation of cationic species is probably also effective. Considering the higher interest and potentiality of metallocene complexes we examined only Zr-derivatives of this type. Cyclopentadienyl and pentamethylcyclopentadienyl Zr derivatives did not show activity to high molecular weight polyethylene with aluminum alkyls as cocatalyst (Table 1). In the presence of M A 0 the use of substituted Cp-ligands (Table l), providing different electronic and steric properties, has a substantial effect on both productivity and molecular weight. Thus while with bispentamethylcyclopentadienylzirconium dichloride high molecular weight polyethylene was obtained with good productivity, even if SA is about 50% less than with Cp2ZrC12, bistetraphenylcyclopentadienylzircon~unidichlorideg gave only C6-C30 oligomers with interesting productivity. The formation of short chains cannot be merely attributed to the rather large amount of M A 0 used, as concentration of this last has not a remarkable effect on molecular weight at the polymerization temperature used.14 Pre 1im ina ry experiments with supported cat a1y st s Supported catalvsts. were performed by using the metal complexes described in the previous section and silica, which had been treated as described in the experimental. Table 2 reports the best productivity obtained and the related conditions adopted.

24. Effect of Ligand and Support on Polymerization Performances

263

Bisnonanoatetitanium dichloride (N2TiC12) on silica (run S 1) or silica/MgC12 (run S2) shows more or less the same productivity as in solution (run H9) after activation with aluminum alkyls. Table 2. Polymerization of ethylene with various titanium and zirconium catalysts supported on silicaa Run

SI s2 s3 s4 s5 S6 s7 S8 s9 s10 s11

Catalyst N2TiC12/Si02 N2TiCIdSi02/MgC12 N2TiC12/Si02/MgC12 N2TiC12/Si02/MgC12 Ti(2-EH)4/Si02 PcTiCIdS i 0 2 Cp2ZrC12/Si02 Cp2TiCIdSi02 Cp2TiC12/SiOfliMe Cp2TiC12/Si02/LiMe Cp2TiC12/Si02/LiMe

Cocatalyst AIEt2CI AIEt2CI MA0 MA0 AIEt2CI AIEt2C1 AIEt3 AIEt2C1 AIEt2CI MA0 MA0

[All/[Mt]

30 35 I50 760 30 24 10 25 30 I00 1000

S.A.b 3.4 6.7 6.9 24.2 2.4 5.5

S.A.*C

0.113 0.191 0.005 0.03 1 0.079 0.229 -

15.0d 33.0e 13.4 41.3

0.145 1.100 0.134 0.04 1

aSilica Croxfield : 300 in*& ; [OH] 2.10-3 moles/g. bKg PE/g atom Mt-h. CKg PE/g atom Mt-g atom Al-h. dMv-lO-5 = 5.92. eMv-10-5 = 16.2.

The same complex on Si02/MgC12 needs a larger amount of M A 0 (runs S3 and S4) to display the same productivity as in the solution (run H10). In the case of Cp2TiC12 some important differences can be observed thus suggesting that the Cp ligand still plays a certain role. With AIEt2CI as cocatalyst, this last complex when supported on silica (run S8) or LiMe pretreated silica (run S9) results more productive than in the solution. However with MA0 the Si02LiMe supported species show (runs S10 and S1 1) comparable activity as with AlEt2CI and much lower than in solution (run H14). It seems therefore reasonable that during supportation some of the Cp ligands are removed15 despite the treatment with LiMe should have converted silanol groups into -SiOLi groups. Cp2ZrCl2 on silica does not give polyethylene with AIEt3, similarly to what happens in solution (runs S7 and H15, respectively). Soga et al. did however report that good yield can be obtained with Cp2ZrC12 supported on Si02 pretreated with C12SiMe2 and activated with trialkylaluminum.16

264

F. Ciardelli, A . Altomare, G . Arribas, G . Conti, F. Masi and F. Menconi

On the basis of these results showing the possibility of analyzing different effects in supported metallocene complexes as well as the importance of the functionalization of the support surface induced us to use HY-zeolite as a support. This crystalline and better defined material appeared more promising in the attempt to produce supported metallocene catalysts displaying similar performances as in solution. When directly supported on merely thermally treated HY -zeolite, Cp2ZrC12 displays (Table 3) rather modest activity with aluminum trialkyls (runs Z1 and Z3), which is substantially improved by using MA0 (run Z8), even if remaining below the value obtained in solution under similar conditions (Table 1, run H16). As this result could be in some way connected to a modification of Zr-complex due to reaction with the silanol functionalities, the zeolite was treated with AlMe3.15 The catalyst prepared with this last support provided improved activity (one order of magnitude) with AlMe3 and excellent activity with M A 0 giving productivity comparable to that expected for the analogous complex in solution (run Z15). Also good activity was obtained in the ethylene/propene copolymerization, with about 20% mol of a-olefin in the copolymer (run 216). In an analogous experiment in toluene solution CpzZrC12 with MA0 (AI/Zr = 1500) gave SA = 2,200 kg/mol Zrohebar with 20% mol propene in the final product. These results suggested the possibility of achieving supportation of metallocene species on zeolite supports. In order to investigate these aspects in better detail Cp2ZrMe2 and more thoroughly purified HY-zeolite were used. Indeed, even the dealuminated zeolite contains both Si-OH groups and extraframework aluminum. The first ones can be simply removed by treatment with trimethylaluminum which occurs with methane evolution and conversion of all -OH groups into Si-O-A1 oxane species.17

lSi-OH 0

+ Al(CH3)3-

-1Si-O-Al(CH3)2 0

+

CH4

The effectiveness of this treatment is clearly shown by the fact that productivity increases of one order of magnitude for the zeolite supported Cp2ZrC12 when activated with trialkylaluminum cocatalyst (Table 3, runs 21-23). Extraframework aluminum can be removed by exhaustive extraction at 50°C with a solution of acetylacetone in ethanol.18 After this treatment the 27Al-NMR (MAS) spectrum shows (Fig lb) a single resonance at 57.45 ppm of the tetrahedral

24. Effect of Ligand and Support on Polymerization Performances

265

Al,19 whereas resonances at 0 and 30-50 ppm of the extraframework A1 (Fig. la) are completely lacking.20 However silanol groups are still present and treatment with CpzZrMe:! is accompanied by CH4 evolution associated with the Zr-carbon bond cleavage. After extraction with acetylacetone, the zeolite was then pretreated with 2M solution of AlMe3 in toluene and successively washed with dry toluene until the test for aluminum was negative; the resulting support shows in the 27AlNMR (MAS) spectrum three resonances at 60.4, 33.1 and 2.2 ppm (Fig. 2a), suggesting that two different species of aluminum, associated with absorbed AlMe3 and the reaction product of AlMe3 with silanol groups, are now present. The addition of CpzZrMez to the modified zeolite does not produce any CH4 evolution, and the 27AI-NMR (MAS) spectrum shows the same resonances as before the zirconocene addition, the relative intensities resulting only moderately changed (Fig. 2b). Moreover, in the 29Si-NMR (MAS) spectrum only the resonance of the Si(OA1) species at -107 ppm can be observed,*1 indicating that no change occurred for Si after addition of the zirconocene to the zeolite pretreated with AlMe3. All these indications suggest that no chemical reaction occurred during complexation and the zirconocene in the support has maintained its original structure. Polymerization experiments carried out with this modified zeolite (HY**) are reported in table 4.

Figure 1. 27Al-NMR (MAS) spectra Figure 2. 27AI-NMR (MAS) spectra of a) untreated and b) acetylacetone of HY zeolite treated with a) AlMe3 and extracted HY zeolite (see text). b) AIMe3 + Cp2ZrMe2 (see text).

266

F. Ciardelli. A . Altomare. G. Arribas. G. Conti, F. Masi and F. Menconi

Table 3. Polymerization of ethylene with zeolite supported Cp2ZrCl2 activated with different aluminum alkyl derivatives.a Run

Cocatalyst

[AI]/[Mt]

Surface treatment

S.A.b

z1

10 50 12

-

22 23

AIMe3 AIMe3 AIEt3

HY -AIMe3

3 37 3

0.28 0.73 0.28

28 Z 15d-e Z I 6d7f

MA0 MA0 MA0

1500 1500 1500

HY -Al Me3 HY-AlMe3

195 2800 2280

0.13 1.87 1.52

-

S.A.*C

Table 4. Polymerization of ethylene with zeolite supported Cp2ZrMe2 activated with different aluminum a k y l derivatives.a Run

Cocatalyst

24 z5 Z6

AIMe3 AIMe3 AlMe3

z9 z10 z11 212

MA0 MA0 MA0 MA0

[AI]/[Mt]

Surface treatmentb

S.A.C

100 100 100

HY**-AIMe3 HY**-Me3SiC1 HY **-Ni+2-AIMe3

30

0.30

-

-

-

-

1500 3000 1500 1500

HY**-AIMe3 HY**-AIMe3 HY **-Me3SiC1 HY**-Ni+Z-AIMe3

181

0.12 0.20 0.10 0.2s

59 1 150 382

S.A.*d

aPetI1ylene = 1 bar, T = 25 "C. bHY** = exhaustive dealurnination with acetylacetone/etlianol. CKg PE/g atom Zr-h. dKg PE/g atom Zr-g atom A1.h.

When supported on HY **,Cp2ZrMe2 with MA0 as cocatalyst gave comparable activity as in solution and needed comparable amounts of M A 0 (run ZlO), as expected by a fixation on the zeolite surface and inside the channels22 of unmodified species.13 Replacement of AlMe3 with Me3SiC1 for HY ** pretreatment gave a support without free silanol groups and extraframework aluminum. With this last support no activity was detected for Cp2ZrMe2 in the presence of AlMe3

24. Effect of Ligand and Support on Polymerization Performances

267

cocatalyst (run ZS), whereas with M A 0 analogous productivity was achieved as with the HY**-AIMe3 support (run Z1 l), indicating that addition of alumoxane is necessary for producing active species. Interchange with N i ( N 0 3 ) 2 brought to the fixation of Ni++ species on the zeolite surface.23

+

Ni(NO,),

-

1 -Si I\

0 0 Ni

I f

+

2HN0,

2

Again Cp2ZrMe2 on this modified support has no activity with AIMe3 (run Z 6 ) , while excellent activity was observed with M A 0 (run 212). Preliminary experiments with Cp2TiMe2 supported on HY** and activated with M A 0 (Al/Ti = 2,000) gave polyethylene with SA = 600 kg/mol Tiehebar. Kinetic analysis indicated that some interesting differences existed between soluble and zeolite-supported species in the case of Cp2ZrMe2 on HY**-AIMe3. Variation of polymerization rate with time indicated that the soluble system is initially more active than the supported one, but shows a typical decay profile of polymerization rate (Rp) vs. time.24 By contrast R, of the zeolite supported system remains almost constant in the first 60 minutes and already after 20 minutes is higher than for the soluble catalyst (Figure 3). Then the comparable productivity of the zeolite supported catalysts with respect to the corresponding systems in solution, despite the lower initial activity, is a consequence of the better stability of the active sites in the former systems. These preliminary results show that the use of properly treated zeolite supports allows to obtain heterogeneized zirconocene species showing appreciable activity with constant R, thus indicating that entrapment in the zeolite channels prevents deactivation reactions and allows to modulate catalytic activity by molecular modification of the support without preventing activation by MAO. While the diameter of channels of the zeolite HY used in this study is high enough to accommodate the cyclopentadienylzirconium complexes, they may exert a certain

268

F. Ciardelli. A . Altomare. G . Arribas, G. Conti. F. Masi and F. Menconi

shape selection towards different molecular species present in the M A 0 mixture. This last effect could be responsible for the substantial stability of active sites number as indicated by the time dependence of the polymerization rate.

.

WI

**..

c1

2.0 -

'b..,

1.5-

o

Cp2ZrMe2-HY**-AlMe3/MA0

-0.1

0

10

20

30

40

50

60

Time (min) Figure 3

Variation with time of ethylene polymerization rate (Rp) in the

presence of homogeneous and supported catalysts (at 25OC, PetIlylene = 1 bar).

ACKNOWLEDGEMENT Partial support by MURST-Rome (60%) is gratefully acknowledged. G.C. thanks SNS-EN1 for PhD fellowship REFERENCES 1 M.Bochmann, S.J.Lancaster, Oryanometulfics, 12, 633 (1993) 2 C.Sishts, R.M.Hathorn, T.B.Marks, J.Am.Chem.Soc.,114, 11 12 (1992) 3 J.A.Ewen, H.J.Elder, Makromol.Ckem.,MacromolSymp., 66, 17.9 (1993) 4 D.J.Crowther, R.F.Jordan, Makromol.Chem.,MacromoI.Symp.,66, 121 (1 993) 5 J.C.W.Chien, W.Song, M.D.Rausch, Macr,omolecules, 26, 3239 ( 1993) 6 K.Soga, M.Kaminaka, Makromol.Ckeni. , Rapid Commun., 13, 221 (1992) 7 W.Kaminsky, F.Renner, Makromol.Chem. ,Rapid Commuii., 14,239 (1993)

24. Effect of Ligand and Support on Polymerization Performances

8 9 10 11 12 13 14

15 16 17 18 19 20 21 22 23 24

269

E.Samue1, M.D.Rausch, J.Am.Chem.Soc., 95,6263 (1973) G.Conti, G.Arribas, A.Altomare, F.Ciardelli, J.Mo1.Cat. (1994) in press. G.Conti et al, in preparation K.Soga, J.R.Park, T.Shiono, Polymer Commun., 32,310 (1991) P.Pino, U.Giannini, L.Porri, in “Encyclopedia of Polymer Science and Engineering”, ~01.8,Wiley Interscience, New York, 1984, p.148 J.C.W.Chien, D.He, J.Polym.Sci.,Part A , 29, 1603 (1191) L.Resc0ni.F. P i e mo n t e s i , G .Francis c o n o , L .A b i s , T.Fiorani, J.Am.Chem.Soc., 114, 1025 (1992) S.Collins, W.M.Kelly, D.A.Holden, Macromolecules ,25, 1780 (1992) K.Soga, H.J.Kim, T.Shiono, Makromol.Cltem., Rapid Commun., 14, 765 (1 993) G.A.Nesterov, V.A.Zakharov, G.Fink, W.Fenz1, J.Mol.Catal., 69, 129 (1991) F.Ciardelli, A.Altomare, G.Conti, G.Arribas, B.Mendez, A.Ismaye1, Makroniol .Cliem. ,Macromol. Symp., in press G.Engelhardt, D.Miche1, “High resolution Solid-state NMR of silicates and zeolites”, Wiley, New York, 1987 J. Klinowski, Ciieni. Rev., 91, 1459 (1991) J. Klinowski, Inorg. Chem., 22,63 (1983) G.A.Ozin, C.Gil, Ciiem. Rev., 89, 1749 (1989) R.I.Soltanov, E.A.Paukshtis, E.N.Yurchenko, B.A.Dadashev, S.E.Mamedov, B.A.Gasymov, Kinet. Karal., 25 (3), 618 (1984) D.Fisher, S.Jungling, R.Miilhaupt, Makroniol. Chem., Macromol. Symp., 66, 191 (1993)

This Page Intentionally Left Blank

27 I

25. Design of Non-Metallocene Single-Site Olefin Polymerization Catalysts

Erik B. Tjaden and Richard F. Jordan.

Department of Chemistry, University of Iowa, Iowa City, Iowa, USA 52242

Absffact: Cationic group 4 metal alkyl complexes containing teuaaza macrocycle or tetradentate Schiff base ligands, e.g. (Meg-taa)Zr(R)+,(Meq-taen)Zr(R)+,and (FSacen)Zr(R)+,are prepared by protonolysis of suitable neuual diakyl precursors. These complexes display electrophilic behavior, but are less active for ethylene polymerization than CpzZr(R)+cations. INTRODUCTION Cationic group 4 metallocene akyl complexes Cp2M(R)+ (M= Zr,Hf) have been extensively exploited as olefin polymerization catalysts. Fundamental organometallic studies of these and related group 3 and f-element metallocene systems have provided a working rationale for the high activity and selectivity of these systems.' The key steric and electronic properties of Cp2M(R)+ species which are important for catalytic activity are: (i) the do metal elecaon configuration, (ii) the highly unsaturated metal center, and (iii) the availability of vacant coordination sites cis to alkyl ligand. A current challenge is to exploit the general insights gained from studies of CpzM(R)+ systems to develop new classes of single site catalysts with improved and/or complementary properties. Our approach to this problem is to design new types of cationic early transition metal alkyl complexes which are structurally and electronically similar to Cp2M(R)+ species, but which are based on non-metalloceneancillary ligands. RESULTS AND DISCUSSION In a previous study, we investigated the synthesis and chemistry of (N4-macrocycle)M(R)+ complexes (1,2; M = Zr,Hf) incorporating Meg-taa or Meq-taen ligands in place of Cp2 iigands.2 The pockets of these macrocycle ligands are too small to accommodate the large ZrIV and H e V ions, so the metal lies above the N4 plane and additional ligands/substrates are forced to coordinate cis to the alkyl ligand (3). Cationic (N4-macrocycle)M(R)+ species coordinate a variety of ligandshbstrates, exhibit non-classical metal-akyl bonding modes (i.e., agostic interactions) characteristic of electrophilic metal systems, and undergo C-H activation and akyne insertion

277

E.B. Tjaden and R.F. Jordan

reactions. These species also polymerize ethylene in the absence of cocatalysts, but activities are far lower than for Cp2M(R)+catalysts.

The most likely reason for the lower activities of (N4-macrocycle)M(R)+catalysts 1 and 2 is that strong electron donation from the macrocycle ligand results in reduced metal electrophilicity and less effective olefin coordination and activation. To circumvent this problem we are exploring systems containing tetradentate N2022- Schiff base ligands which are expected to be weaker donors due to the higher elecmnegativity of oxygen vs. nitrogen. Floriani has prepared a series of (acen)MXZ halide complexes, some of which adopt cis-MX2 structures.3 However, it has proved rather difficult to convert these precursors to alkyl derivatives. We have developed more direct routes to (acen)MRz and (acen)M(R)+complexes and investigated the reactivity of these systems. The tetradentate ligands Fg-aCen (4a) and &-acen (4b) are readily prepared via condensation reactions (eq l).4 Neutral dialkyl complexes (%-acen)ZrR'Z (5a-5c) are obtained directly via alkane elimination reactions of ZrR'4 compounds and (&-acen)H2 (eq 2). A single crystal X-ray analysis of (Fg-acen)Zr(CH~CMe-& (5a. Figure 1) revealed a mgonal prismatic geometry. The C-Zr-C angle (1300) is larger than in (Nq-macrocycle)MR2or Cp2MR2 complexes (85 - 950).

(R6-acen)y

4a, R I F 4b,R=H

25. Design o f Non-Metallocene Single-Site Catalysts 273

Protonolysis5 of 5a with the ammonium reagent [HNMe2Ph][B(C6F5)41 yields (Fgacen)Zr(CHzCMe3)(NMe2Ph)+ (6a) as the NMe2Ph adduct (eq 3). The fact that 6a retains coordinated amine indicates that this species is more electrophilic than (N4-macrocycle)M(R)+or Cp2M(R)+ species, which generally do not coordinate NMezPh.

5a

A single crystal X-ray analysis (Figure 1) established that 6a is structurally similar to Sa, although the Fg-acen ligand adopts a more planar conformation and the NMezPh-Zr-CH2CMe3 angle is large ( 1700).6

FZS

\

@Cl3

Figure 1. Structures of (Fg-acen)Zr(CH~CMe3)2(5a) and (Fg-acen)Zr(CH~CMe3)(NMezPh)+ (6a). The B(C&5)4- anion of 6a is not shown. Complex 6a undergoes ligand exchange reactions (PMe3, RCN) and inserts polar substrates (CO. ketones), but exhibits only low ethylene polymerization activity. We initially hypothesized that this results from tight amine binding which inhibits coordination of the olefin. In the presence of 2 equiv Al(iBu)g, added to scavenge the amine, in siru-generated 6a is a moderately active catalyst

274

E.B. Tjaden and R.F. Jordan

(eq 4). Under the same conditions, the non-fluorinated catalyst derived from 5c is much less active (eq 4). These results prompted us to explore the synthesis of fluorinated buse-free cations via protonolysis reactions using bulkier ammonium reagents. The reaction of 5a with [HNMePh2][B(C6F5)4] affords the base-free cation (F6acen)Zr(CH2CMe3)+(78) which can be isolated as an analytically pure solid (eq 3). The bulky, weakly basic amine NMePh2 does not coordinate to Zr. Surprisingly 7a is a poor ethylene polymerization catalyst (eq 5). This indicates that the role of Al(iBu)3in eq 4 is more complex than originally thought. Complex 7a can be activated for polymerization with I equiv Al(iBu)3 (eq 5); however, additional Al(iBu)3 does not increase activity. Thus the active species in eq 4 and 5 is formed by reaction of 7a and 1 equiv Al(iBu)3. 1) [ H N M ~ z P ~ I [ B ( G F ~ ) ~ I /R

( R6-acen)Zr,

R'

2)2 equiv AI~BU), 3) 3 atm ethylene

*

toluenekhlorobenzene 30 min, 50 OC

-.

-

...

(4)

-

5a, R = F; R' = CHpCMe 14,000 (g)(mol)-'(atrn)-'(hour)-' 5b, R = F; R' CHzPh 18,000 800 Sc, R = H; R' = CHzCMe3

iz -

No Activity

0 (F6-X811)Zr-C (C6Fd4' 78

H,CMe,

-.

1 AI(bu),

benzene, 45 OC

._

10,000 (g)(mol)-'(atm)-'(hour)"

At present, the mechanism by which Al(iBu)3 activates 7a for ethylene insertiordpolymerization is unknown. The low reactivity of 7a in the absence of AlR3 may result from non-optimum orientation of the coordinated olefin and the neopentyl ligand in the putative (F6acen)Zr(CH2CMe3)(ethylene)+ intermediate. If this intermediate is structurally similar to 6a, with the coordinated olefin replacing the amine, the olefin-Zr-alkyl angle would be too large (ca. 1700) for facile migratory insertion. The Al(iBu)g cocatalyst may bind the alkoxide oxygens of 7a, forcing the neopentyl ligand and the vacant coordination site into a more cis-like arrangement. It is also possible that 7a is inherently unreactive due to the steric bulk of the neopentyl ligand, which is expected to disfavor olefin insertion. The reaction of 7a with Al(iBu)3 may generate a Zr-H

25. Design of Non-Metallocene Single-Site Catalysts 275

species which is more reactive. Control experiments with Cp2Zr(R)+catalysts indicate that the role of Al(iBu)3 in eq 4 and 5 is not simply to scavenge impurities from the reactor. Experiments designed to elucidate the role of A1 cocatalysts in this system are in progress. SUMMARY

Cationic, group 4 metal (N4-macrocycle)M(R)+ and (Rg-acen)M(R)+ complexes can be prepared using synthetic routes developed for Cp2M(R)+ species. These non-metallocene systems exhibit electrophilic behavior, but are less active than Cp2M(R)+ species for olefin polymerization. Efforts to modify the ancillary ligands to increase polymerization activities are in progress. REFERENCES

2

ti

Jordan, R. F. Adv. Organomet. Chem. 1991,32, 325. Uhrhammer, R.; Black, D. G.; Gardner, T. G.; Olsen. I. D.; Jordan, R. F. J . Am. Chem. SOC. 1993,115, 8493. (a) Corazza, F.; Solari, E.; Floriani, C.; Chiesi-Villa, A.; Guastini. C. J . Chem. Soc., Dalton Trans. 1990, 1335. (b) Solari, E.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C. J . Chem. SOC., Dalton Trans. 1992, 367. Liu, H. Y.; Scharbert, B.; Holm, R. H. J . Am. Chem. SOC. 1991,113, 9529. Hlatky, G.; Turner, H. W.; Eckman, R. R. J . Am. Chem. SOC. 1989.11 I , 2728. The X-ray structure of 6a was determined by Prof. Jeff Petersen at West Virginia University.

This Page Intentionally Left Blank

277

26. InsiteTM Catalysts Polymerization

Structure/Activity

Relationships

for Olefin

-

Polyolefins and Elastomers Research and Development Laboratories, The Dow Chemical Company, 2301 Brazosport Boulevard, Freeport, TX 77541 ABSTRACT

The Dow Chemical Company has developed a new family of polyolefins using Constrained Geometry Catalyst Technology (CGCT). The technology is being commercialized under the tradename INSITEm. These INSITE"' technology polymers are characterized by a narrow molecular weight and comonomer distribution. In general, a narrow molecular weight and comonomer distribution would be expected to improve physical properties at the expense of processability. The unique molecular structure of INSITEm technology polymers delivers increased physical properties without sacrificing processability. The improved processability is believed to be the result of significant amounts of long-chain branching. The structure/activity relationships of the family of Constrained Geometry catalysts which give rise to these unique polymers will be discussed. INTRODUCTION There have been numerous interesting developments in the polyolefins industry in recent years. One of the most exciting areas has centered around the development of homogeneous single-site catalysts. These single-sitecatalysts produce ethylene alpha-olefin copolymers with properties that are different when compared with traditional LLDPE and ULDPE polymers. The Dow Chemical Company has developed a new family of polyolefins using Constrained Geometry Catalyst Technology (CGCI'). The technology is being commercialized under the tradename INSITEm. INSITEm technology polymers are characterized by a narrow molecular weight and comonomer distribution. In general, a narrow molecular weight and comonomer distribution would be expected to improve physical properties at the expense of processability. The unique molecular structure of INSITE"' technology polymers delivers increased physical properties without sacrificing processability. The improved processability is believed to be the result of significant amounts of long chain branching. This paper will address the structure / activity relationships of the family of Constrained Geometry catalysts which gives rise to these unique polymers. EXPERIMENTAL The Constrained-Geometry complex syntheses and olefin polymerization conditions were as previously described'. Cyclic voltammetry was conducted in an argon filled drybox in a standard H cell comprising two electrode wells separated by a fine glass frit, platinum working and counter electrodes, and a silver reference electrode. The solvent was 1,2-difluorobenzenecontaining tetra-nbutylammonium tetrakis-pentafluorophenylborate supporting electrolyte.

278

J.C. Stevens

SINGLE-SITE CATALYSTS Metallocene catalysts based on bis-cyclopentadienyl complexes activated with M A 0 have been known for some time. The Kaminsky catalyst is a bis-cyclopentadienyl zirconium catalyst. This complex, when activated with MAO, can produce single-site olefin polymers with high efficiency at low temperatures. Unfortunately, in a high temperature low pressure solution process this catalyst system produces low molecular weight polymers. Additionally, a large amount of the expensive aluminoxane cocatalyst is required for optimum efficiency. C G n CATALYSTS

We have recently discovered a family of new Constrained Geometry Catalysts that allows Dow to produce unique polyolefin polymers in a low pressure solution process. The key catalyst features are shown in Structure 1. The catalysts are monocyclopentadienyl Group 4 complexes with a covalently attached donor ligand. The donor ligand stabilizes the metal electronically, while the short bridging group pulls the donor ligand away from its "normal" position. This has the effect of sterically opening up one side of the complex.

R

I

M = Ti, Zr, Hf Structure 1. General structure of Constrained Geometry Catalyst.

It is possible to synthesize a large number of derivatives of this basic structure and study the structure activity relationships in a rational manner. This paper will address the effect of changing the substituents on the cyclopentadienyl ring, the bridging (R2) group. the coordinating group, and the R3 group.

In general, the open nature of the catalytic site in the Constrained Geometry catalysts does not allow for much steric control of the polymerization reaction, and homopoly a-olefins are generally atactic to slightly syndiotactic. The degree of tacticity obtained under commercially useful conditions is so low that the catalysts should be considered to be atactic. The open nature of the active site allows the copolymerization of a wide variety of olefins with ethylene. Normal a-olefins such as propylene, butene, hexene, and octene are readily incorporated in large amounts. In addition, non-traditional olefins such as styrene can be incorporated. Styrene / ethylene copolymers containing significant amounts of styrene and having a high molecular weight have not been available in the past, as conventional polyolefin catalysts will not copolymerize ethylene with styrene to any appreciable extent. The catalyst activity, when activated with between 50 and 1,OOO equivalents of MA0 is excellent. Catalyst efficiencies between 150,000 and 750,000 g of polymer per gram of metal are obtained, depending on the reactor temperature, specific catalyst, MA0 level and other process variables.

26. INSITETMCatalyst Structure/Activity Relationships

279

(a) Figure 1. X-Ray Crystal smcture of [(tetramethyl-q5-cyclopentadienyl)(N-tbutylamido)dimethylsilyl] titanium dichloride. (a) Front view; (b) side view. The X-ray crystal structure of a titanium tetramethylcyclopentadienyl Constrained Geometry complex bridged with a single dimethylsilaneis shown in Figure 1. In general, the bond distances are unremarkable and are consistent with other known titanium (IV) cyclopentadienyl complexes. The titanium is unsymmemcally bound to the Cp ring, due to the covalent attachment of the amide through the bridging group. Several interesting features can be seen from the side view of this molecule, shown in Figure lb. In this view, it can be seen that the silicon bridging atom has been pulled out of the plane of the Cp ring and that there is a considerable amount of strain in the pseudo-4-membered ring formed by the Cp-Si-N-Ti. The silicon atom of the bridge is pulled out of the plane of the Cp ring by 0.87 A. In addition, the nitrogen atom of the amide ligand has been pulled down from a "normal" position due to the covalent attachment to the Cp ring. The angle f m e d by the Cp centroid, the metal. and the amido nitrogen is 107.6". Comparable Cp-M-N angles for non-constrained complexes are in the range of 115 - 120 O.2 The crystal structure of the analogous zirconium derivative is shown in Figure 2. Again, the bond distances are normal for a zirconium (IV)complex, and are 0.10 - 0.15 A longer than the titanium derivative, consistent with the larger covalent radius of zirconium. The side view of the same molecule (Figure 2b) a ain shows that a large amount of ring strain is evident. The silicon atom of the bridge is pulled 0.84 out of the plane of the cyclopentadienering, and the Cp-ZrN angle is 102.0 '. These titanium and zirconium complexes have a sterically open active center as a result of covalently attaching the amide ligand to the Cp ring. Selected bond distances for these and several other constrained geometry complexes are shown in Table 1.

x

M Ti

Zr Ti Ti

R2 -(SiMe2)-(SiMez)-(SiMe2)2-(CH2)2-

1.909 2.056 1.913 1.909

2.256 2.397 2.277 2.309

2.329 2.455 2.318 2.345

2.329 2.463 2.406 2.345

2.436 2.539 2.429 2.391

2.436 2.540 2.494 2.391

2.262 2.405 2.283 2.282

2.262 2.414 2.291 2.282

Table 1. Selected crystallographic bond distances (in A) for constrained geomeuy complexes.

280 J.C. Stevens

Figure 2. X-Ray Crystal structure of [(tetramethyl-q5-cyclopentadienyl)(N-tbutylamido)dimethylsilyl] zirconium dichloride. (a) Front view; (b) side view. Another constrained geometry catalyst with a slightly longer ethylene bridging p u p is shown in Figure 3. For this complex, the Ti - N bond length is identical to the silane-bridged species. However, the longer bridge allows the titanium to occupy a position more nearly centered over the Cp ring. The side view of this molecule (Figure 3a) shows that the longer bridge is less strained than the short single silyl bridge. The Cp-Ti-N angle is still quite acute in this catalyst, at 107.9 O. The bridge length can be increased further by the use of a disilyl group, as shown in Figure 3b. For this complex, the Cp-Ti-N angle is 120 O , and the active site is much less open than the shorterbridged complexes.

(a)

(b)

Figure 3. X-Ray Crystal structures of (a) [(tetramethyl-q5-cyclopentadienyl)(N-t-

butylamido)ethanediyl]titanium dichloride; and (b) [(tetramethyl-q~-cyclopentadienyl)(N-tbutylamido)teaamethyIdisilyl] titanium dichloride.

26. lNSITETMCatalyst Structure/Activity

Relationships 28 I

Polymerization Results The Constrained Geometry catalysts are effective LLDPE catalysts in high temperature solution polymerizations. Table 2 shows the results for the titanium complex bridged with a single dimethylsilane group. As can be seen, the catalyst is effective at temperatures as high as 160 OC. giving a useful melt index product even in the presence of hydrogen as a molecular weight control. The data shows that as the temperature of the polymerization is increased, the melt index of the polymer increases. At the same time, the density of the polymer increases, indicating that less octene is incorporated at higher temperatures. The second set of runs at 140 O C show that hydrogen is an effective molecular weight control with this catalyst.

Temperature 9J 110

130 140 160

octene

AH2

mL

kPa

150 150 150 150

345 345 345 345

AI:Ti

250 250 500 500

yield R

Mw

density

I2

98 147 128 90

161,000 136,000 66,500 53.000

0.9140 0.9197 0.9174 0.9317

0.15 0.15 3.34 10.66

182 0.9063 1.72 0 500 7.91 690 500 157 57,500 0.9075 All runs with 20 pmoles of [(C5Me4)SiMe2N(t-Bu)]TiC12,2OOO,d solvent, 450 psig ethylene, 10 140

140

300 300

minutes reaction time. Table 2. Ethylene / Octene Copolymer Production using Constrained Geometry Catalysts Table 3 shows that ultra-low density elastomers can easily be produced with constrained geometry catalysts. Fractional melt index elastomeric ethylenebtene resins with densities between 0.87 and 0.85g/mL can be obtained with high efficiencies. The open nature of the active site allows efficient actene incorporation at relatively low actene concentrations. The 0.855 density polymer is over 53 weight percent octene, as determined by l3C NMR. The last two runs runs show that the catalyst responds well to hydrogen, allowing excellent control of melt index over a wide range.

Temperature

[octene]

AH2

80 100 100 100

1.59 1.59 1.85 1.85

172 172 0 345

yield

density

I2

67.0 0.8672 co.10 70.0 0.8700 co.10 110.0 0.8570 0.66 131.9 0.8552 7.48 All runs with 10 p o l e s of [(C5Me4)SiMe2N(t-Bu)]TiCl2,1200 mL solvent. 450 psig ethylene, 10 minutes reaction time. Table 3. Ethylene / Octene Elastomer Production using Constrained Geometry Catalysts

282 J.C. Stevens

The effect of substitution on the cyclopentadiene ring is shown in Table 4. As the groups on the Cp ring are modified to make the ligand more electron withdrawing, there is a marked decrease in efficiency and melt index, as well as an increase in density. The electron density at the metal center for these complexes can be examined by looking at the electrochemical reduction potential for the complexes, or the Jc-H for the dimethyl derivatives. This data indicates that high efficiencies and high comonomer incorporation is correlated with increased electron density at the metal center.

Efficiency npolylgTi C5(CH3)4 150.000 C5H4 59,000 indenyl 31,000

Cp

density dmL 0.8850 0.9070 0.9179

I2

JC-H~

Hz 10.1 2.92 0.92

118.5 119.5 120.4

~ ~ / r b &/p

V -1.49

-1.28

L Y:

.. ,,a

,di

.II

\cl

All runs with 10 pmoles of catalyst. MAO, looomL solvent, 200 mL octene, 450 psig ethylene, 130 "C, 10 minutes run time. a) C-H coupling constant for respective Ti methyl complexes. b) Ti 314 couple, vs. SCE in 1,2-difluorobenzene. Table 4. Catalysis using various Cp derivatives The same effect is observed as the substituent on the nitrogen atom is varied. Table 5 shows that substitutingprogressively more electron-withdrawinggroups on the amido nitrogen leads to decreased efficiency, comonomer incorporation, and melt index.

R

t-Bu Cd5 4-F-CgHq

Efficiency R poly / n Ti 150,000 27,000 15,000

density dmL 0.8850 0.9087 0.9400

I2 10.1 6.37 2.90

GSi+

8, .q 4i

/

R

All runs with 10 pmoles of catalyst, MAO,lOOOmL solvent, 200 mL octene, 450 psig ethylene, 130 OC,10 minutes run time. Table 5. Catalysis using various Amido derivatives The nature of the bridge has a large effect on the activity of the constrained geometry catalyst. Table 6 shows the effect of substituting different bridging groups. As the bridge length is increased from a single dimethylsilyl to a disilyl bridge, the efficiency decreases five fold, while the amount of comonomer incorporated into the polymer decreases, a shown by the increase in the polymer density. The titanium catalysts with an acute Q-M-N angle have the highest efficiency and greatest amount of octene incorporation. This effect can be explained by the more crowded nature of the active site with the longer bridges, as was shown with the crystal structure shown earlier. The crystal structure of the intermediate length ethylene bridged complex showed it to be slightly more crowded than the single silyl bridged complex, and the density of the product reflects this fact. The all-hydrocarbon ethylene

26. lNSITETMCatalyst Structure/Activity Relationships

283

bridge imparts a favorable combination of steric and electronic factors, as shown by the high efficiency and low melt index.

oa

R2 (Si(CH3)2h Si(CH3h

Efficiency

density

g poly / g Ti

dmL 0.9441

23,000 150,000 560,000

120.5 107.6 107.9

(cH2)2

0.8850

0.9190

I2

6.14 10.13 0.21

a) Cp centroid - Ti - N angle. All runs with 10 pmoles of catalyst, MAO, looomL solvent, 200 mL octene, 450 psig ethylene, 130

"C. 10 minutes run time. Table 6. Catalysis using various bridged derivatives It can be seen in Table 7 that the substitution of zirconium for titanium has a dramatic effect. In each case, the zirconium analog has a lower efficiency, and gives a higher density and melt index product.

CP

R2

M Ti

C5(CH3)4 CS(CH314

si(CH3)2 si(CH3)2

indenyl indenyl

Si(CH3)2 Si(CH3)2

zr

CS(CH3)4 CS(CH314

(CH2h

zr

zr Ti Ti

I2

Efficiency

density

g poly / g Ti

dmL 0.8850 0.9571 0.9179

10.13 >250 0.92

3 1,000 30,500

0.9427

9.34

560,000 33.000

0.9190 0.9610

150,000 110,000

R21Cq

jpq

)I,

0.21 >250

All runs with 10 pmoles of catalyst, MAO, looomL solvent, 200 mL octene, 450 psig ethylene, 130 OC, 10 minutes run time. Table 7. Catalysis Results Comparing Ti and Zr Constrained Geometry Complexes. CONCLUSIONS Constrained Geometry catalysts allow the production of a unique family of olefinic polymers. The proper selection of the metal, bridging group, and other substituents allows the control of product properties in a high temperature process. With the proper selection of catalyst variables. products ranging from high molecular weight elastomers to high density polyethylene can be produced. ACKNOWLEDGMENTS The author would like to acknowledge David R. Wilson, Peter N. Nickias, and Robert Mussel1 of The Dow Chemical Company. In particular, the author would like to acknowledge the work of Phil R. Rudolf of The Dow Chemical Company for the solution of the X-ray crystal structures.

284 J.C. Stevens

REFERENCES 1. James C. Stevens, Francis J. Timmers. David R. Wilson, Gregory F. Schmidt. Peter N. Nickias, Robert K. Rosen, George W. Knight, Shih-yaw Lai. European Patent Application 416.815 A2. Aug. 30, 1990.

2. N. W. Alcock, G. E. Toogood, M. G. H. Wallbridge, Acta. Cryst. (1984). C40.598-600.

285

27. Novel Molecular Structure Opens Up New Applications for Insite@ Based Polymers

Director of Research and Development and G. M. Lancaster, Manager of

,Development

Polyolefins & Elastomers R&D, The Dow Chemical Company 2301 Brazospotl Blvd, 8-1607, Freeport, Texas USA 77541

Abstract The Dow Chemical Company has developed a new class of olefin copolymers utiliiing a single site catalyst with constrained geometry and trademarked INSITEQP Technology. The technology can deliver a wide range of new and innovative polymers ranging from polyolefin plastomers (POP, <20 wt.% octene comonomer) and polyolefin elastomers (POE, >20 wt.% octene comonomer) to high

performance polyolefins and polyethylenes. This paper will focus on two major advantages of this new technology which have created new polymer design capabiiities/rules for application and product development: 1) INSITEQPTechnology provides tor independent control of processability, and 2) INSITEQDTechnology provides an unprecedented control of molecular architecture. The use of these new design Capabilities will also be explored in three distinctively different development programs. The use of these new polymer design capabilities coupled with processlproductmodeling capabilities greatly reduces the application and product development time. Int roductlon The Dow Chemical Company has developed a new class of olefin copolymers based on a single site catalyst with constrained geometry and trademarked INSITEQPTechno1ogy.l The technology can deliver new and innovative polymers ranging from polyoiefin plastomers (POP, <20 wt.% octene comonomer) and polyoiefin elastomers (POE, >20 wt.% octene comonomer) to high

performancepolyolefins and polyethylenes. These materials have superior physical properties due to their narrow molecular weight, and comonomer distributions. In June 1993, Dow announced the startup of a 125 million pound plant to produce POP and POE polymers utilking INSITEQPTechnology. Dow's INSITE@ Technology polymers (ITPs) offer improved physical and mechanical propertieswhen used in elastomer, plastomer and polyethyleneapplications as well as enhanced mall processabiiity. The improved physical and mechanical properties result from the narrow molecular weight and short chain branching distributions of the new polymers compared to conventional

286

K . W . Swogger and G.M. Lancaster

polymers. The enhanced melt processability results from insertion of long chain branches in the polymer b a ~ k b o n e ,the ~ * subject ~ of a recently allowed patent.4 In a major effort to rapidly commercializethese new polymers, a significant R&D program has been on-going linking customer application performance requirements to product design, material science, structure property relationship, and process capability. Dow can establish this direct linkage because of the differences of ITP polymers from conventional polyoletins. An example of the material science - product design linkage is being presented separately by Steve Chum, et.

This paper will

focus on two major advantages of this new technology which have created new polymer design capabilities/rules for application and product development: 1) INSITE@Technology polymers (ITPs) have independent control of processability, and 2) INSITE@Technology provides an unprecedented control of molecular architecture. Aspects of these two concepts and the resulting design capabilitieslrules will be discussed and three application developments shown as examples. Coupling the new design rules with process/product modeling capabilities greatly reduces the development cycle time for successful applications. New Rules for Polymer Deslgn The ability of INSITE@ Technology’s single site catalyst with constrained geometry to polymerize higher levels of alpha olefins with ethylene as well as novel alpha olefins offers new and decidedly different processing capabilities and physicaVmechanica1properties to the industry. POPS and POE’s provide unique properties across a wide range of densities and melt indices. The control of processability independent of MWD is a major design rule change. Because they do not have LCB, conventional copolymers made from Ziegler-Natta catalysts and metalloceneor single site catalysts (SSC) homogeneous copolymers require the MWD (via process changes or via blending) to be broadened or the incorporation of a processing aid in order to improve processability. ITPs are unique in their ability to have enhanced processability without broadening of the MWD and sacrificing performance prope~iies.~~’ The presence and control of long chain branching (LCB) in INSITE@Technology polymers offers good processability in addition to the narrow molecular weight distribution. In LLDPE no long chain branching (LCB) exists and 110112 is used as a measure of the processability or flowability of a polymer and can indicate the polydispersity or molecular weight distribution (Mw/Mn) of the LLDPE polymers (Figure 1). In LDPE where high levels of LCB exist, melt tension is used to characterize the melt elasticity of the polymer (Figure 2). The conventional rheological parameters do not adequately describe the unique relationship between the ITP structure, processing, and performance properties because of the presence of LCB (Figure 3 and Figure 4). In addition to melt index and density that can be used with ITP to characterize the flow and

27. INSITE@ Based Polymers 287 physical properties of the polymer, a new parameter has been proposed to better and more completely describe the LCB effects of ITP. This additional parameter is called the Dow Rheology Index or DR18 The Dow Rheological Index (DRI) is a processing performance index which characterizes the long chain branching effect of ITP's (Figure 5) independent of melt Index. DRI is defined as the extent that the rheology of ITP deviates from the rheology of the conventional homogeneous polyolefins that do not have LCB. DRI is defined by the following normalizedequation: DRI = (3.65E6 TO =

Characteristic Relaxation Time

qo

=

T O ~ O -1)

/lo;

Zero Shear Viscoslty

The parametersTO and qo are determined by a nonlinear regression of the experimental data numerically filied to the generalized Cross equation. The index ranges from 0 (for all SSC polymers which do not have LCB) to 30. Combined with MI, the DRI can be used to determine many existing measures of processability. DRI can be used to

calculate the flowability of ITP polymers (i.e., the ease of pumping the polymer through an extruder or the speed which an injection mold can be filled). Figure 6 shows the relationship between the high shear viscosity and DRI. The DRI can also be used to determine the melt elasticity of ITP polymers (i.e., the melt tension, bubble stability, Neck in, Draw resonance, hot green strength, etc.). Figure 7

shows the relationshipof melt tension and DRI. The type of LCBs contained in ITP is different from LDPE in three distinct ways (Figure 8): 1) ITP contains longer and fewer LCB's than high pressure LDPE, 2) ITPs can be designed with a controlled level of LCB along the polymer backbone5 and 3) unlike LCB in LDPE, ITPs LCBs are essentially linear and unbranched. The mechanical and physical properties of ITPs (POP and POE) with various DRls (level of LCB) are shown in Table 1. The relatively b w number and high chain length does not impact the physical or mechanical properties of ITPs but does increase the processability as noted via the increasing DRI. The design rules traditionally used to design a new product for a Ziegler-Natta catalyzed LLDPE are shown in Table 2. The new design rules for ITPs are shown in Table 3. Note that with conventional Ziegler Natta catalyzed LLDPE many of the parameters used to control or design the polymer are coupled and are affected strongly by the production conditions and catalysts. This forces the polymer designer, scientist or englneer to make tradeoffs or compromises. By contrast, INSITEQP Technology breaks these existing rules. Polymer design parameters (density, MW, MWD-2. and

288

K . W . Swogger and G.M. Lancaster

LCB) using INSITE@Technology can be controlled independently. This change enables the polymer designer, scientist, or engineer to design polymers without the current compromises.

As discussed in earlier papers, Dow's INSITEB Technology and process allows the control of molecular architecture to a new level. Dow can design polymers to meet Performance requirements because of our kinetic understanding of reactor operation and the knowledge of structure/property relationships of these new polymers.8 Combining this control of kinetics and structure with the long chain branching gives Dow unique design capabilities.2 By using models we can design a molecule based on structurelproperty relationships, determine plant conditions to make the polymer and know what the polymer is after it is made. Appllcatlon & Product Development This type of control allows us to do product design quite differently than in the past. We are asking our customers for Performance requirements rather than product characteristics (Table 4) for both existing and new applications. This is especially true for the POEs, where ethylene-octene copolymers have not been available until now, and for POP'S, where the new design rules have created new performance opportunities. Polymer specifications for existing applications must be thoroughly reviewed. Usually the specifications are based on the old rules of polymer design and the compromises of performance and processability are often built in. Thus, the polymer designer must go back to performance requirements in order to fully exploit the new relationships that are being established. The INSITE@Technology application and development program links (via predictive modeling and scientific understanding) the Performance requirements of the customers' application with the material science, the processing / fabrication science relationships and the polymer microstructure I property relationships in order to design the polymer. Then utilizing the kinetic process / product model the process conditions to make the designed polymer are defined and the product is produced. In order to fully understandthis process let's examine three applications. Sealant Application (Blown Fllm) The performance requirements for a sealant application are shown in Table 5. This application requires a Nylon barrier layer and will be produced on blown coextrusion equipment. In addition to the performance requirements there are a few additional polymer design rules that should be examined. The linear, ethylene octene structure of POPS brings toughness for extra product protection and /or gauge reduction, plus thermal stability and compatibillty with other polyolefins for co-extrusion. The narrow MWD and comonomer distributions produce lower extractables, excellent optics, and lower heat sealing temperature8 to provide packages with improved organoleptic pedormance, enhanced package appeal, and faster packaging speeds. The LCB content enhances the

27. INSITE@ Based Polymers

289

processability (i.e., improves both the melt elasticity and pumping efficiency) of the resin and eliminates melt fracture (Figure 9) from occurring at high line film extrusion line production rates.7 Utilizing the new polymer design rules (MW, comonomer content, LCB, MW=2 and comonomer type) of ITPs, a polymer can be designed which has significantly improved extrusion, sealing performance, optics, and toughness. Typical sealant materials considered for this application are shown in Table 6. Compared to the competitive materials, the plastomers exhibit improved heat seal strength over a wide temperature range (Figure 10). Plastomer 1 sealed at temperatures similar to the EVA and ionomer, while Plastomer 2 sealed at temperatures 5 to 10°C below the competitive polymers. Ultimate seal strengths were a minimum of 30% greater for the plastorners compared to the competitive polymers. Initiation temperatures for the EVA and Plastomer 1 were similar (approximately 95°C). The lowest initiationtemperature was noted for Plastomer 2 at approximately82"C, slightly lower than that of the ionomer. Ultimate hot tack strengths of the plastomers were 60% higher than the ionomer. The EVA copolymer exhibited poor ultimate hot tack strength. (Figure 11). In summary, plastomers offer outstanding heat seal and hot tack strengths at low sealing temperatures, resulting in faster packaging line speeds and reduced leaker rates. When this is combined with the excellent processability, toughness, puncture resistance and optics, these polymers change the rules for applications requiring high performancesealant materials. Face Mask (Injectlon Moldlng) The performance requirements for the face mask application are shown in Table 7. The main performance requirements for this application is processability and the elimination of plasticizers or processing aids. The injection molding pressures of three ITP polymers and one homogeneous polymer are shown in Table 8. The trial shows the importance of DRI since broad changes in density and MI complicates the use of I10112 flow relationships for polymer characterization. The excellent optics and physical properties of ITP allow for injection molded parts to have optical and physical properties as good as or better than 1-PVC molded parts since no plasticizer is used. Figure 12 shows the effect density or % comonomer has on the stress strain properties of a molded plaque. The excellent processability of ITP's allow parts to be Injection molded on existing molds with excellent optics and flexability and thereby the rules for material selection for injection molded articles have been changed. Wlre & Cable (Banbury Compounded & Cable Extruslon) The key performance requirements for a flexible wire and cable insulation are shown in Table 9. In addition to the performance requirements the application must meet or exceed the current

290 K . W . Swogger and G.M. Lancaster specification (UL1281 Class 45). One of the key performance requirements is the polymers ability to accept high levels of filler and modifiers. A typical formulation is shown in Table 10. The POE family of polymers (>20 wl 'YO octene) have excellent filler and modifier acceptance. Utilizing the new polymer design rules (MW, comonomer content, LCB and comonomer type) and the unique features (the improved physical properties resulting from the narrow MWD) of ITPs, moderate molecular weight iTPs can provide the cross link efficiency, and mechanical properties comparable to very high MW EPDM rubbers (Table 11). From the data one can see that even though the starting Mooney viscosity is much lower than the EPDM's, the ITP based formulations exhibit excellent crosslinkability and filler and oil acceptance. In fact, the ability to have a lower MW and thus, a lower viscosity allows a high level of filler and oil loading to be achieved during the compounding stage. This greatly enhances the compounding efficiency. The LCB of the ITP achieves the required melt strength and results in the elimination of melt fracture during wire line processing. In addition, the narrow MWD and octene comonomer allow a high level of oil to be incorporated without any observation of "bleeding" or phase separation that is the general rule for medium MW EPDM rubbers. The meeting and exceeding of the performance requirements by these polymers combined with its excellent compounding efficiency change the rules for materials used for flexible wire and cable insulation.

Summary Because of Dow's ability to control molecular architecture, once performance requirements are understood it is relatively easy and fast for Dow to either deliver the polymer to the customer or tell him that it can not be done. This has a tremendous benefit to both our customers and Dow. Our customers see quicker response, less trial and error, less cost, very consistent products, and more ability to get the performance they require. Dow sees less cost and resources utilized, quicker development, more satisfied customers, and a wider range of application markets. Focused market development continues in North America and Europe, in conjunction with a wide range of customers under secrecy agreements. The target markets for POP'S and POE's include packaging, automotive, wire and cable, medical, and consumer and industrial goods. The INSITE@Technology application and product development process has been designed to take advantage of the new set of polymer design rules which are being built around the following: 1) ITP's have independent control of processability, and 2) INSITEQPTechnology provides an unprecedented control of molecular architecture. Wide scale commercial product availability into specific markets will be announced during 1993. 63 INSITE is a Trademark of The Dow Chemical Company

Paper Presented at SPO '93 Conference, USA. Used with permission of Schotland Business Research, Inc.

27. INSITEQ Based Polymers 291

Bibliography Trademark Announcement, December 1992. Swogger, K.W., and C.I. Kao, Proceedingsof SPE PolyolefinsVII International Conferecence, Feb. 1993, ANTEC 1993. Swogger, K.W., 'The Material Properties of Polymers Made from ConstrainedGeometry Catalyst", Proceedings of the Sec. Int'l Bus. Forum on Specialty Polyolefins, SPO '92, p. 155165, Sept., 1992

Story, B.A., and G.W. Knight, "The New Family of Polyolefinsfrom INSITE@Technology", Proceedings of Metcon '93 Worldwide Metallocene Conference, p. 112-123, May, 1993 Chum, P.S., Third InternationalBusiness Forum on Specialty Polyolefins, SPO Conference, Sept. 1993. Mergenhagen. L.K., and N.F. Whiteman, "Polyolefin Plastomers'As Sealants In Packaging

-

Applications", 1993 TAPPI Polymers, Laminations and Coatings Conference, Sept. 1993. Edmondson, M.S., and S. E. Pirtle, "CGCT: New Rules for Ethylene Alpha-Olefin Interpolymers

- Processing-Structure- Property Relationshipsin Blown Films', SPE Antec '93 Conference Proceedings Technical Papers Volume XXXIX, p. 63 - 65, May 1993 Lai, S, and G. W. Knight, "Dow Constrained Geometry Catalyst Technology (CGCT): New Rules

-

for Ethylene Alpha-Olefins Interpolymers Controlled Rheology Polyolefins", SPE Antec '93 Conference Proceedings Technical Papers Volume XXXIX, p. 1188-1192, May 1993 Knight, G.W., and S. Lai, "Constrained Geometry Catalyst Technology: New Rules for Ethylene

--

Alpha-Olefin lnterpolymers Unique Structure and Property Relationships" Proceedingsof the SPE Polyolefin Vlll International Conference, p.226-241, Feb. 1993

292 K . W . Swogger and G.M. Lancaster

PROPERTY Toughness Modulus Fbwability (Il~dlp) Melt Strength Melt Index

MAJOR PARAMETERS MW,MWD, Density Density (0.912 to 0.960)

MINOR PARAMETERS

Mw

Mw,MWD Mw MW of HI3 Frartinn

MANY OF THESE PARAMETERS ARE COUPLED 8 AFFECTED STRONGLY BY CATALYSTS

PROPERTY Toughness Modulus

MAJOR PARAMETERS Density (SCB) Average Density (0.86510 0.958) LCB, SCB LCB

Fbwability (Ildlp) Mek Strength Mek Index

w

MINOR PARAMETERS

Mw w

w LCB

DENSITY, LCB AND MW CAN BE CONTROLLED INDEPENDENTLY

E 4

-

TYPlCAl -R F F P FOR RESINS IN THF SO'S Customers Will Need To Know Performance Requirements Such As:

*

Stiffness Bubble Stability Impactfroughness Processability Abuse Resistance Dimensional Stability Taste and Odor

* *

-

Sealability Optics Printability Handling (Conversion) Tear Resistance Weatherability FDA

27. INSITE" Based Polymers 293

LE 5

-

PFRFORMANCF REQUlREMENTS FOR SEALANT APPLICATION

Fabrication Process: Breakthrough: Applicable Regulatons:

Blown Film Coextrusion High Hot T a d and Wide Heat Seal Range FDA Direct Food Contact

4: POLYMER TYPE . POP1 POP 2 ULDPE EVA

COMONOMER 9.0 wt % C8 12.0 wt % C8 9.0 wt % C8 9.0 wt % VA 1

Innnrnar

POP ULDPE EVA

7 Fabrication Process: Breakthrough: Applicable Regulatons:

- PFRF-

/Ns hn\

--

MELT INDEX (dglmln) 1.0MI 1.0MI 1.0 MI 2.0 MI 1

MI

N I

Polyolefin Plastomer Ultra Low Density Polyethylene Ethylene Vinyl Acetate

R F Q J J J T S FOR FACF-

Injection Molding (Sprue) Improve Processabiliy, Elimination of Platicirer. Non-PVC 51OK FDA ApprovaWAverage 9-12 Months

v Msterlal M Density DRI I10/12 Ini. Pressure

DRI 4.5 3.1 NA NA

PERFORMANCF INJFCTIOY

Sample 1

Sample 2

Sample 3

Sample 4

10.87 0.872 0.53

10.26 0.903

10.00 0.880 0.00

7.1

10.82 0.887 0.35 7.8

7.1

5.8

1017

in73

1143

1 A35

0.30

294 K.W. Swogger and G.M. Lancaster

Fabrication Process:

Banbury Internal Mixing / Wire Line Extrusion With Steam Continuous Vulcanization INSITE@Technology Polymers Can Provide Cross Link Efficiency, Wire Line Extrusion And Mechanical Properties Comparable To High Mw EPDM UL 1581 Class 45 (90 and l05OC EPR)

Breakthrough: Applicable Specifications: Feature

Teat Method

I Wire Smoothness 8

I Lower Cvcle Times 8 Lower I Pass SDark Test on Wire I

I

I

Unmet Nerdllmprovement

Mln. Requlrement

? Drop Temperature o Scorch. Cures Within CV Tube > 700 psi > 250%

Wire Line Extension Tensile Elongation Heat Age T&E Retention At 121°C/10d 50% of Original Value and 135'ffd

I

f Higher

ASTM D-638 ASTM 0-638

Lower A 0 Levels Required

ASTM 0-573 8 0-638

Untreated Clay Paraffinic Oil Peroxide Coagent Antioxidant

2.5 1.6

Vinvl Silana

Base Polymer Liinimum Torque Maximum Torque T90 Mooney @ 250°F (ML) Minimum Torque Maximum Torque Delta 3 Crosslinked 4OO0F Tensile Strength (psi) Tensile Q 100% Strain (psi) Elongation Shore A Hardness

I Sag At Strainer Extruder

I Good Melt Strength

INSITE@ Technology

Vlatelon@ 7000 EDPM

Royalone@ 539 EPDM

4 12 6.7

11.5 6.3

14.5 35 6.4

16.5 27 6

47.5 60 6.5

58 73 5

1113 752 218 73

1197

1139 706 214 67

37

837

170 74

27. INSITE" Based Polymers 295

FIGURE 1 FLOW CHARACTERISTICS MWDvs 110/12 nwan 11

10 8

6

Heterogeneous

7

ITP

6

5

Homogemr

4

"":"":'":"":"'':"":"":"":'"'I

1 4

6

6

7

D

0

l

0

1

1

1

2

1

3

iron.

FIGURE 2 FLOW CHARACTERISTICS MWD vs Melt Tens MwMn Hetorogeneoll. ITP

1.6$ 1

. . . .

: 1.6

.

.

.

.

:

. 2

.

.

.

:

-

. . 25

.

:

.

.

.

.

4

3.1

3

Molt Tonrlon (Orom)

FIGURE 3 110A2 RATIO vs LCB 0.85 1.15 Melt Index AND 0.87 0.935 Density ITP

-

-

14.

I, 13-

-=

11-

I

I I10-

:i

7

n

# - .

*

I

- . - . - . - . - ' -

296

K.W. Swogger and G.M. Lancasier

FIGURE 4 MELT TENSION RATIO vs LCB 0.85 1.15 Meit Index AND 0.87 0.935Denslv ITP

-

-

Moll Tonsla. am.

FIGURE 5 DRI ve PREDICTED LCWlooOOC 0.5-30MI& .870 0.920 Density ITP POLYMERS

-

-

DRI 20

10 : 5 2 -

I

0

1

0.5

1.5

2

PRED. LCB/1OOOOC

35,000

-

30.000

-

25.000

-

20,000

-

16.000

10,000 6,000

-

1 0.5 MI

L

-<

27. INSITE@ Based Polymers 297

FIGURE 7 DRI vs MELT TENSION r

Melt Tension, grams

l t * 5MI

n

/

1

12

. I

2

-0

0.5 MI

1

4

6

8

0

14

16

DRI

FIGURE 8 COMPARISON OF BRANCHING LDPE vs ITP

lDPE

IIe

MWD

BROAD

NARROW

BRANCH LENGTH

200 300'

-

1300 1600''

3-7

0.3 0.8

-

# OF CARBONS

NUMBER LCB I POLYMER CHAIN

-

LITERATURE ESTIMATION *PREDlCTED FROM KINETIC MODEL

FIGURE 0 MELT FRACTURE ITP vs HETEROGENEOUS RESIN APPARENT SHEAR RATE 1 I SEC 10000

--

DOWLEXO M

"HI1

-

A

ITP 1 No k l t F r u l u n

11((11111~~~~~'

111111111~~1~~

,,,,,,,

,(1111)1

YUT

rtuciunc

,,111("(

U J W A U MELT

nuciunc

10

1.OEI

I

APPARENTSHEARSTRESS DmEsicu 2

q.0 E 7

298

K.W. Swogger and G.M. Lancaster

FIGURE 10 HEAT SEAL STRENGTH OF COEXTRUDED FILMS H U T SEAL STRENGTH (LWIN)

10

lm

100

110

120

S E M BAR TEMPERANRE (DEG C)

FIGURE 11 HOT TACK STRENGTH OF COEXTRUDED FILMS HOT TACK STRENQTH(WIN)

14

pop2

I*=-=@

--c

12 10

Pop,

P-1

-c

8

EVA

6

laom

4

9 -

2 0 80

m

100

110

SEAL EAR TEMPERATURE (DEQ C )

120

299

28. Molecular Weight Distribution Control with Supported Metallocene

Catalysts

SON-KI IHM,KYUNG-JUN CHU and JIN-HEONG YIM Department of Chemical Engineering, Korea Advanced Institute of Science and Technology, 373- 1 Kusong-dong, Yusong-gu, Taejon 305-701, Korea

ABSTRACT Several kinds of supported Cp2MeC12 (Me = Ti, Zr, Hf)catalysts were prepared with three different SiO2-supports; i.e., SiO2/MAO, Si02/Al(C2H5)3 and Si02/(C2H5)MgCl. The Si02-supported metallocene catalysts show fairly good activity for ethylene polymerization. The molecular weight distribution (MWD) of polyethylene obtained over Si02-supported Cp2ZrC12 and Cp2HfC12 catalysts was narrow, and the shape of MWD was unimodal. In case of SiO2/MAO and Si02/Al(C2H5)3, supported Cp2TiCl2 catalyst resulted in polyethylenes with bimodal MWD. In case of Si02/C2H5MgCl, however, the supported Cp2TiCI2 catalyst gave polyethylene with unimodal MWD.

INTRODUCTION Since the discovery of highly active metallocene catalyst system for olefin polymerization1), several reports have been made on the efforts to overcome the impediment to the commercialization of homogeneous catalyst (requiring large amount of methyl aluminoxane (MAO)). One approach is preparation of cationic, do metallocene by strong Lewis acid instead of Another approach is immobilization of metallocene compounds on silica or other supports4d). Chien et al.4) reported that Et(Ind)2ZrC12 immobilized to Si02 pretreated with small amount of MA0 showed high catalytic activity for olefin polymerization. Soga et al.536) reported that MgCl2-, Al2O3-, and Si02-supported Et[IndH&ZrCl2 could conduct propylene polymerization in a fairly good yield with commercial cocatalysts; i.e., Al(CH3)3, AI(C2Hg)3 or Al(i-C4Hg)3. During immobilization of the metallocene, the silica support was pretreated with small amount of MA0 to enhance the catalytic activity. In this work, metallocene compounds ( Cp2TiC12, Cp2ZrC12 or Cp2Hflc12) were immobilized to modified SiO2-supports pretreated with small amount of MAO, Al(C2H5)3 and (CzHs)MgCl, respectively. Ethylene polymerization was carried out ~

~

0

~

3

~

)

.

300 S.K. Ihm. K.J. C h u and J.H.Yirn

over those catalysts. The catalytic activity and molecular weight distribution of polyethylenes were compared.

EXPERIMENTAL Materials. Ethylene and nitrogen were purified by removing traces of residual oxygen and moisture with columns packed with oxygen scavenger (Fischer RIDOX) and molecular sieve 5A (Aldrich) respectively. Toluene (J.T.Baker) used in the polymerization and catalyst preparation was dried by refluxing through a distillation column over sodium metal under dry nitrogen. Cp2MeCI2(Me = Ti, Zr, Hf) (Aldrich), Al(CH3)3 (Aldrich), AI(C2H5)3 (Aldrich), Al(i-CqHg)3 (Aldrich), (C2H5)MgCI (Aldrich), and MA0 (Tosoh Akzo Co.,Ltd.) were used as received without fbrther purification. Catalvst preparation. Silica (Davison 952) was dehydrated at 600OC for 10 hrs under nitrogen atmosphere. Hydroxyl content of dehydrated silica measured by TGA (thermal gravimetric analyzer)(Dupont 990) was found to be 1 mmoVg silica. Three different compounds (MAO, Al(C2H5)3 and (C2H5)MgCI) were used to make different supports (SiO2/MAO, Si02/Al(C2H5)3 and Si02/(C2H5)MgCI respectively). 15 g of dehydrated Si02 was reacted with 22 mmol of each of those compounds at 500C for 1 hr. The solid part was washed with plenty of toluene and dried with nitrogen purging. 3 g of supports were reacted with 150 ml of 7.7 x M Cp2MeCI2 (Me = Ti, Zr, in 100 ml toluene at 4OoC for 30 mins, washed with plenty of toluene, and dried. The contents of metals in catalysts were measured by ICP (inductively coupled plasma) ( a - 3 5 10) spectrophotometer. Polvmerization and polymer characterization. Ethylene polymerization was carried out in a glass reactor (500 ml) with 300 ml of toluene. Catalyst or solution of Cp2MeCI2 was added to the reactor and ethylene was introduced. After ethylene saturation, cocatalyst was added and polymerization was started. The polymerization rate was estimated from the monomer consumption rate, which was measured by mass flow meter (Type 825 of Datametrics) connected to 16-bit PC via an AD/DA converter (ADL 1 100 of Analog Design). The polymerization was stopped by adding acidic methanol. The precipitated polymer was washed with methanol and dried at 60OC in vacuum. IR spectra of the supports were recorded under nitrogen atmosphere over the range of 4000-2500 cm-I using FT-IR @omem MB-102) with a difise reflectance accessary. The molecular weight distribution was measured by gel permeation chromatography (Waters 150CV) at 145OC with 1,3,5trichlorobenzene as solvent.

28. MMD Control with Supported Metallocene Catalysts 301

RESULTS Table 1 shows the results of ethylene polymerization. The catalytic activity of titanocene with MA0 showed the lowest activity among different metallocene catalysts. The reason was attributed to bimolecular deactivation of titanocene with aluminum compound7). The Cp2MeC12 (Me = Ti, Zr, Hf) compounds with MA0 showed higher catalytic activity than the supported counterparts. SiO2/MAOsupported Cp2ZrC12 catalyst with MA0 cocatalyst showed the highest activity among supported metallocenes. Al(CH3)3, Al(C2Hg)3 and Al(i-CqHg)3 were also used to compare with MA0 as cocatalyst. Al(i-CqHg)3 seemed to give higher activity than Al(CH3)3 or A I ( C ~ H S ) ~ ~ ) . Table 1, Figure 1 and Figure 2 show the molecular weight distribution of polyethylenes obtained from different metallocenes. Homogeneous catalyst based on titanocene, zirconocene or hafnocene with MA0 cocatalyst yielded polyethylenes which showed low polydispersities. Cp2TiC12 with MA0 showed the highest molecular weight, and Cp2ZrC12 with MA0 showed the lowest molecular weight among the different metallocene catalysts. It is well known that polyolefins having a bimodal molecular weight distribution can be obtained by polymerizing ethylene in the presence of a catalyst system comprising two or more different transition metal compounds (each having different propagation and termination rate constants for ethylene polymerization) and MAO*,9). In our homogeneous metallocene catalysts, bimodal molecular weight distribution was also obtained by mixing two metallocenes (titanocene and zirconocene). Among the single metallocene catalysts supported on Si02/MAO, only Cp2TiC12 showed bimodal molecular weight distribution of polyethylene. It is speculated that the interaction of titanocene and aluminum species in MA0 of Si02MAO could affect the molecular weight distribution being bimodal. Accordingly, the MWD of polyethylenes obtained over titanocene catalysts supported on Si02/AI(C2H5)3 and Si02/(C2Hs)MgCI were compared. It was noted that AI(C2Hs)3 contains aluminum species but (C2Hs)MgCI does not. It was found that bimodal molecular weight distribution was obtained over Cp2TiC12 /Si02/AI(C2Hs)3 catalyst but not over other supported metallocene catalysts. Figure 3 shows the IR spectra of Si02, dehydrated Si02, SiO2/MAO, MAO, Si02/AI(C2Hs)3 and Si02/(C2Hs)MgCI respectively. There are many types of surface groups on the silica surface; single hydroxyl (I), hydrogen bound hydroxyls (11), paired hydroxyls (III), and adsorbed water (IV)10, 11). After dehydration at 6OOOC under nitrogen atmosphere, only the single hydroxyl group remained and the others disappeared. After the reaction with MAO, Al(C2H5)3, or (C2Hs)MgCI, the single hydroxyl group disappeared.

Table 1. Activity and molecular weight distribution of polyethyleneprepared with different metallocene catalysts RUn

catalysts)

Cocatalystb)

Yield

Activity

Mn

Mw/Mn

k)

(Kg PWmol Me.atm.hr)

ShapeofMWD

No. 1.

4.5

682

(x 10-4) 5.2

2.64

UIlhlodal

2.

3.1

479

5.6

2.55

UUiXllodal

2.57

UUiXllodal

3.

2.8

424

7.3

4.

3.4

5 15

4.9

2.6

Unimodal

5.

0.9

136

5.2

3.1

6.

0.8

121

5.0

11.8

UUilllodal Bimodal

7.

0.7

106

6.1

2.3

Unimodal UUilllodal

8.

0.4

61

5.5

3.7

9.

1.4

212

10.3

2.5

UUilllOdd

10.

1.o

152

10.8

2.9

Unimodal

11.

0.8

121

5.4

9.2

Bimodal

12.

4.4

667

4.0

10.0

Bimodal

13.

0.9

136

3.9

10.3

Bim&

14.

1.o

152

6.1

8.0

Bimodal

(Continued)

RUn

Catalyst

cocatalyst

Yield

k)

No.

Activity (Kg PWmol Me.atm.hr)

Mn (x

Mw/Mn

ShapeofMWD

lo4)

15.

Cp2ZrC12/SiO2/A1(C2H5)3

MA0

0.43

67

5.4

3.9

Unimodal

16. 17.

Cp2HfC12/SiO2/AI(C2H5)3 Cp2TiC12/SiO2/AI(C2H5)3

MA0 MA0

0.35 0.13

54 20

8.5

Unimodal

5.2

2.2 5.6

18.

33

3.2

4.0

Unimodal

19.

CP~Z~C~~/S~O~/(C~H MA0 ~ ) M ~ C I 0.21 0.12 MA0 Cp2HfC12/Si02/(C2Hg)MgCI

19

4.3

4.9

Unimodal

20.

Cp7TiC17 /SiO?/(C?H5)MgCI

12

8.9

4.0

Unimodal

MA0

0.08

Polymerization condition : pressure = 10 psig, temperature = 60 OC, polymerizationtime = 30 mins a) 7.7 x 10-6 mol of Cp2MeC12 (Me Zr, Hf,Ti) was used. b) 0.025 mol of cocatalyst was used. c)&,e) Cp2TiC12Kp2ZrC12mol ratio was 9.

Bimodal

304 S.K. Ihm. K.J. Chu and J.H.Yim

2

3

4 5 6 7 log (Molecular Weight)

0

Figure 1. Molecular weight distribution of polyethylene prepared with different metallocene catalysts : (a) Cp2MeCI2 + MAO, (b) Cp2MeC12 /SiO2/MAO + MA0 and (c) Cp2MeC12/Si02/ M A 0 + Al(i-CqHg)3 : (-) ;Me=Zr, (- - -) ;Me=Ti and (- - ) ; Me=Zr+Ti

-

2

3 4 5 6 7 log (Molecular Weight)

0

Figure 2. Molecular weight distribution of polyethylene prepared with Si02/AI(C2H5)3 and Si02/(C2H5)MgCI supported metallocene catalysts : (a) Cp2MeC12/Si02/AI(C2H5)3,(b) Cp2MeC12/Si02 /(C2H5)MgCI : (-) ; Me=Zr, (- -) ;Me=Ti and (- . - ) ; Me=Hf

-

28. MMD Control with Supported Metallocene Catalysts 305

dehydrated SiO,

4000

I

I

I

I

I

3750

3500

3250

3000

2750

2500

Wave Number (cm-')

Figure 3. IR spectra of different supports H./H

Figure 4. A plausible model of active site for different SiO2-supported metallocene catalysts

306 S.K. Ihm. K.J. Chu and J.H.Yim

From the infrared spectra of SiO2-supports and molecular weight distribution of polyethylenes, the active site nature of supported rnetallocene can be proposed as Figure 4. SUMMARY The molecular weight distribution of polyethylene could be controlled by properly combining the nature of metallocene and the modification method of silica support. REFERENCES 1. H.Sinn, W.Kaminsky, Adv. Organomet. Chem B, 99 (1980). 2. R.F. Jordan, C. S. Bujgur, R. Willett, and B. Scott, J. Am. Chem. SOC.108, 7410 (1986).

3. P. Longo, L. Oliva, A. Grassi, and C. Pellecchia, Makromol. Chem. 190, 2357 (1989). 4. J.C.W. Chien, D. He, J. Polym. Sci., Part A: Polym. Chem 29, 1603 (1991).

5 . M. Kaminaka, K. Soga, Makromol. Chem., Rapid Commun. 12,367 (1991). 6. K. Soga, M. Kaminaka, Makromol. Chem., Rapid Commun. 13,221 (1992).

7. J.C.W. Chien, B.P. Wang, J. Polym. Sci., Part A: Polym. Chem, 26, 3089 (1988). 8. J.A. Ewen, "Catalytic Polymerization of Olefin", Elsevier Press ,1986, p 271. 9.A. Ahlers, W. Kaminsky, Makromol. Chem., Rapid Commun. 9, 457 (1988). 10. T.E. Nowlin, R.I.Mink, F.Y. Lo and T. Kumar, J. Polym. Sci., Part A : Polym. Chem., 29, 1167 (1991). 11. Y.I. Yermakov, B.N. Kuznetsov, and V.A. Zakharov," Catalysis by Supported Complexes", Elsevier Press ,1981, p 59.

307

29. Highly Isospecific Heterogeneous Metallocene Catalysts Activated by Ordinary Alkylaluminums

KAZUO SOGA Japan Advanced Institute of Science and Technology, Hokuriku, 15 Asahidai, Tatsunokuchi, Ishikawa Pref. 923-12, Japan ABSTRACT Tetrachlorosilane or 1,1,2,2-tetrabromoethane was reacted with the surface hydroxyl . groups of silica gel and the resulting chemically modified silica gel was brought into contact with lithium salt of indene. The catalyst precursors thus prepared were then reacted with zirconium tetrachloride to obtain the immobilized heterogeneous metallocene catalysts. Polymerization of propene was conducted with them using either methylalumoxane or common trialkylaluminums as cocatalyst. Since the catalysts may contain a mixture of meso and racemic isomers which give atactic and isotactic polypropene, the polymer produced was fractionated by extracting with boiling heptane. It was found that the catalysts can be easily activated by ordinary trialkylaluminums to give highly isotactic polypropene with the melting point as high as over 162 “C. Similar catalysts were also prepared using fluorene as ligand, which gave not syndiotactic but highly isotactic polypropene.

INTRODUCTION Poly(a-olefins) of any structure (isotactic, hemiisotactic, syndiotactic and atactic) can be obtained with metallocene catalysts simply by tailoring the stereorigid catalyst precursor, basically according to the local symmetry. Much effort has been recently paid to modify the metallocene catalysts for partial use. It was demonstrated that interconnection of a pair of ligands with single-bridge causes a marked increase in the isotacticity as well as molecular weight of polypropene’.’’. Immobilization of metallocene on the solid surface is also in progress. We have already reported3’ that metallocene catalysts supported on Al,O,, MgCl,, MAO-treated SiO, etc. are easily activated by ordinary alkylaluminums. Whereas, Kaminsky et al.4) obtained highly isotactic polypropene with high molecular weights using a single-phase catalyst composed of Et(Ind),ZrCI,, MA0 and SO,.

308 K . Soga

More recently, we have developed a new type of highly isospecific SiO, supported metallocene catalysts which can be activated by ordinary alkylaluminums. EXPERIMENTAL Materials: Propene and toluene of research grade purity commercially obtained from Takachiho Chem. Co. were further purified according to the usual procedures. SiO, (Fuji Davison Co. # 952) was calcined at 200, 400 or 900 "C for 6 h under a reduced pressure. (CH,),Si(Ind),ZrCI,, Et[IndH,],ZrCI,, iPr(Flu)(Cp)ZrCI, were prepared according to the literat~re~"'~'. Methylalumoxane (MAO) and alkylaluminums were donated from Tosoh Akzo Co. The other chemicals of research grade purity were commercially obtained and used without further purification. Preparation of Supported Catalysts: The synthetic procedure of the CI,Zr(Ind),Si-SiO, catalyst is described below as an example. A solution of SiO, (2.5 mmol) in toluene was dropwise added to a suspension of 4.3 g of SiO, in 70 cm3 of toluene, followed by refluxing for 48 h under agitating with a magnetic stirrer. Modified SiO, was separated by filtration and washed with a large quantity of THF. Then, to the SiO, in 30cm' of THF was dropwise added a suspension of lithium salt of indene (5 mmol) in THF at 0 "C under nitrogen atmosphere. The mixture was heated up to room temperature and kept standing for 12 h with a vigorous stirring. The solid product was separated and washed with a large amount of THF to obtained the catalyst precursor. The catalyst precursor was brought into contact with a solution of Li(n-C,H,) (5.5 mmol) in n-hexane, followed by reacting with

X I , 2THF (2.5 mmol in THF) at room temperature for 12 h. The resulting solid product was separated, washed with a large quantity of THF and diethylether and finally evaporated to dryness under vacuum to obtain the CI,Zr(Ind),Si-SiO, (I) catalyst. The procedure is schematically shown in Figure 1.

lndene

CtF'

/ 7 -7 7 /"

L1-lndene

A

77

7

s102

s102

I

lndene

\ /

sio2

Eapplw~tPi8uarallOn

Filtering 6 Washlng

''

77

SIO*

SICI,

\

Lblndene Indene-LI Indene Indene \ / \Sl' SI / \ ZrC14.2THF 2Ll(n-C4H ) THF,r.f.

( 5 mmol ) Lithium salt of lndene

s102

(5 5 mmol)

tWashlng 6 Drylng

THF, Stirring at 0%

2Un-C4Hd Sbrring at r t for 12h

Washlng

Stirring at r t

(Zr content 5 9 104mol(rr)/g-Si02) by ICP

Figure 1

Scheme and procedure

Of

catalyst (I)

preparation

'Or

29. Highly lsospecific Heterogeneous Metallocene Catalysts

309

The CI,Zr(Ind),Et-SiO, (11) catalyst was prepared using tetrabromoethane in place of SiCI, (Figure 2). Whereas, the CI,Zr(Flu),Si-SiO, (111) catalyst was prepared according to the procedure shown in Figure 1 using fluorene i n place of indene.

(11)

2Ll( n-C4H9)

"\ /c' 7r

/-\

Activation

-

Li

lndene lndene M A 0 or TlBA

Toluene

)-(

u sio,

Figure 2

ZrCi4 2THF

THF,r.l.

*

IndeFe lndene Li

Y P

sio2

Scheme of catalyst (11) preparation.

Polymerization and Analytical Procedures : Polymerization of propene was conducted at 40°C in a 100cm3 stainless steel reactor equipped with a magnetic stirrer using toluene as solvent. Polymerization was terminated by adding acidic methanol and the polymer obtained was adequately washed with methanol, followed by enacting with boiling 1,2,4-tnchlorobenzene to remove the catalyst ash. Polymerization of propene was also carried out using the corresponding homogeneous catalysts for reference. The contents of Zr in the catalysts were analyzed by an ICP-OES-spectrometer (Jobin Yvon, JY-70-PLUS). The molecular mass distributions (MMD) of polymers were measured at 145 "C by gel-permeation chromatography (GPC, Waters 15OC) using o-dichlorobenzene as solvent. The melting points (Tm) of polymers were measured on samples which had been previously melted at a heating rate of 10 "C/min. The microstructure of polymers was mainly determined by I3C NMR. The spectrum was recorded at 120°C with a JEOL GX-270 spectrometer operating at 67.8 MHz. Polymers were dissolved in 1,2,4-trichlorobenzene/benzene-d6 (9/1 by vol) up to 10 wt-%. RESULTS AND DISCUSSION The catalysts (I) and (11) are considered to contain both aspecific (meso) and isospecific (racemic) active sites. The polymers obtained were, therefore, fractionated by extracting with boiling heptane. The results of propene polymerization over the catalysts(1) and (11) together with some analytical data of polymers are shown in Tables 1 and 2, where the results obtained with the corresponding homogeneous catalysts are also indicated for reference.

310 K . Soga

Table I

Results of propene polymerization with the C1,Zr(1nd),Si-SiO2 catalyst".

Catalyst

Activity in Mw Cocatalyst Amount of Yield cocatalyst (in g) kg(PP)/mol(Zr) (lO*gmol'l) (in mmol)

(CH,),Si(lnd),ZrCl,

MA0

5

0.37

74

3.0

MA0 MA0

1

0.32

5.4

34.0

3

0.33

5.6

Tm

1.1. iso.pentads (in %) mmmm(%)

(in "C)

-

142.1

32

83.3

153.0 159.2

68

94.3

156.1 162.3

67

156.7 163.0

68

MA0

10

0.36

6.1

-

AI(i-C,,HJ,

1

0.22

3.7

72.0

153.9 158.0

80

AI(i-C&),

3

0.25

4.3

-

158.6 162.2

76

Cl&(Ind),Si-SiO,

98.0

a) SiO, was calcined at 4WoC

Table 2

Results of propene polymerization with the CI,Zr(Ind),Et-SiO, catalyst"

Catalyst

Et[IndH,],ZrCI,

ClJr(Ind),Et-SiO,

Cocatalyst

Amount of cocatdyst (in mmol)

Yield Mw (in g) (lo4gmol")

MA0

3

2.07

MA0 AKi-C.,&), Al(n-C,HJ, AKC,Y), WCHJ,

1

0.47

1

0.40

1

0.3

Tm in"^)

I 1 1.0

1.1.

iso.pentads

(in o/o)

mmmm(%)

-

71.0

149.9 160.0

46

93.0

158.2

55

91.1

1

0.18 0.21

157.7 153.2 161.2

67 68

I

0.1 1

156.8 162.1

61

48.5

-

-

a) SiO, was calcined at 400 "C

It is instantaneously obvious from Tables 1, 2 and Figure 3 that the isotactic fraction (1.1.) as well as molecular weight, [mmmm] pentad and melting point of isotactic PP drastically increase by using the present immobilized metallocene catalysts. Besides, the supported catalysts can be activated by ordinary alkylaluminums. Among the trialkylaluminums used in the present study, Al(i-C4HJ3 showed the highest activity.

29. Highly Isospecific Heterogeneous Metallocene Catalysts 3 1 1

-

I

a) 1 mmol of Al(i-C4Hg)3

b) 3 rnmol of Al(i-C4Hg)3

-----.c) 1 mmol of MA0 d) 3 mmol of MA0

/

e) 10 mrnol of MA0

\ 140 150 1GO 170 180 190 T e m p e r a t u r e ("C)

Figure 3

DSC charts of isotactic PP (boiling heptane insoluble fraction) obtained by changing the amount of cocatalyst.

To check the microstructure of isotactic PP in more detail, some of the boiling heptane insoluble polymers were analyzed by 1 3C NMR, which did not display any peak assignable to the irregular propene units resulted from 1,3-insertion. The disappearance of such irregular units in addition to the very high [mmmm] value might cause a marked increase in Tm.

312

K. Soga

Most of the supported metallocene catalysts reported so far have been devised to immobilize metallocenes on the solid surface utilizing the ionic interactions between the C1-ligands of metallocenes and the surface active sites. Whereas, in the present catalysts, zirconocene may be fixed on SiO, more rigidly as schematically illustrated in Figures 1 and 2. However, the present catalysts are supposed to possess two kinds of active species which differ in mobility as shown in Figure4.

ct f1

ctZr , 'i Indene lndene /

,z\

fZ:

lndene Inde ....

lndene

\ / i"""'" Si

\ /

.Si.

SiOz

More rigidly immobilized style

SiOz Less lmmoblllzed s t y l e (B)

(A)

Figure 4

\

Plausible structures of the CI,Zr(Ind),Si-SiO, catalyst.

In fact, most of the isotactic PP obtained here display two melting points. It is supposed, therefore, that the more rigidly fixed species (A) is responsible for the production of higher isotactic PP. To confirm it, three kinds of the CI,Zr(Ind),Si-SiO, catalysts were prepared using the SiO, calcined at 200, 400 and 900 "C, and polymerization of propene was conducted at 40 "C over them. Since the concentration of surface hydroxyl groups increases with a decrease in the calcinating temperature of SO,, it is expected that the fraction of (A) becomes predominant with decreasing the calcinating temperature. The results of propene polymerization (Table 3) are in good agreement with this consideration.

Table 3

Results of propene polymerization with the ClTr(Ind),SiSiO, catalyst using the SiO, calcined at different temperatures.

Calcinating

Cocatalyst

Temp. (in "C)

Amount of

Yield

Tm

1.1.

cocatalyst (in mmoi)

(in g)

(in "C)

(in %)

200

A1(i-C4W3

3

0.26

162.3 158.2

75

400 900

3

0.25

162.2 158.6

76

3

0.24

-

159.5

3

0.14

161.4

-

400

Al(i-C,Y), A1(i-C4W3 MA0 MA0

3

0.33

162.3 156.1

900

MA0

3

0.17

200

-

159.1

31 84 67 27

29. Highly lsospecific Heterogeneous Metallocene Catalysts

313

Polymerization of propene was then carried out at 40°C using the Cl,Zr(Flu),Si-SiO, (111) catalyst. The polymer obtained was fractionated by extracting with boiling heptane. In

Figure 5 are illustrated the I3C NMR spectra of (a) whole polymer, (b) boiling heptane soluble polymer and (c) boiling heptane insoluble polymer obtained with the CI,Zr(Flu),SiSiO, - A1(i-C4H& catalyst system.

,,I\

40

20

30

10

W"

.

3'5.

.

30.

2'5

20

15

10

Figure 5 "CNMR spectra of polypropene obtained with the CI,Zr(Ru),Si-SiO, catalyst : (a) whole polymer (b) boiling heptane soluble polymer (c) boiling heptane insoluble polymer.

, - - - . . -. . 35

. . .

30

25

20

15

ld

314

K . Soga

Surprisingly, the boiling heptane insoluble polymer (c) was found to be highly isotactic. Some additional data on the typical boiling heptane insoluble polymers are shown inTable4. Table 4

Results of propene polymerization with the CI,Zr(Flu),Si-SiO, catalyst”. ~ _ _ _

Cocatalyst

Amount of cocatdyst (in mmol)

Yield (in g)

Mw (lo4grnol.’)

Tm (in “c)

mmmm or rrrr (%)

i-Pr(Flu)(Cp)ZrCI,

MA0

13

3.03

3.9

123.0

r r r r = 77

CI,Zr(Flu),Si-SiO,

MA0 AI(i-C,HJ,

15 15

0.5 1 0.45

33

160.3 163.9

mmrnm = 96

Catalyst

a) SiO, was calcined at 400’C. x =??

However much more information should be necessary to speculate the structure of isospecific sites in the catalyst (Ill). In conclusion, it was found that highly isospecific heterogeneous metallocene catalysts, which are activated by ordinary trialkylaluminums, can be prepared by fixing the ligands on the surface of SiO,. A more detailed study on the improvement of both catalyst activity and isospecific selectivity is now in progress, the results of which will be published elsewhere.

REFERENCES 1. T.Mise, S.Miya, H.Yamazaki, Chemistry Letters, 1853 (1989) 2. W.Speleck, M.Antberg, J.Rohrmann, A.Winter, B.Bachmann, P.Kiprof, J.Behm, W.A.Hemnann, Angew. Chem. Int. Et. Engl., 31, 1347 (1992) 3. K.Soga, M.Kaminaka, Makromol. Chem., 194, 1745 (1993) 4. W.Kaminsky, F.Renner, Makromol. Chem., Rapid Commun., 14, 239 (1993) 5. F.R.W.P.Wild, M.Wasiucionek, G.Huttner, H.H.Brintzinger, J. Organomet. Chem., 63, 288 (1985) 6. J.A.Ewen, M.J.Elder, Makromol. Chem., Macromol. Symp., 48/49, 253 (1991)

315

30. Mol Mass Regulation in the Ally1 Nickel Complex Catalyzed 1, 4-ck Polymerization of Butadiene R. TAUBE, S. WACHE and J. LANGLOTZ Institute of Inorganic Chemistry, Martin-Luther-University Halle-Wittenberg, Geusaer StraBe, D-06217 Merseburg, Germany ABSTRACT For the industrial important 1,4-cis polymerization of butadiene the polymerization degree ii can be regulated in a theoretically founded way using the C12-allyl nickel(I1) complex [Ni(C12Hl,)][B(C6H3(CF3)2)4] 1 as the catalyst. A proper kinetic reaction model has been derived with the insertion reaction of butadiene into the allyl nickel bond and the 0-hydride elimination from the growing polybutadienyl chain as the rate determining reaction steps for the chain propagation and for the transfer reaction of the catalyst to the monomer, respectively, by which a new polymer chain is formed. Accordingly the polymerization degree fi is determined by the ratio of the corresponding rate constants kp/kiirwhich strongly depends on the cation-anion interaction as it is shown for the complex catalysts [Ni(C12Hl,)]03SCF3/10(AlF3 0,s toluene) and [Ni(c12Hl,)]PF6/NEt4PF6 in comparison with the technical nickel catalyst Ni(O2CR),/BF,*0Et2/AlEt3.

-

INTRODUCTION The technical synthesis of 1,4-cis polybutadiene is carried out at present as solution polymerization using titanium-, cobalt-, nickel- and quite recently also neodymium-containing ZIEGLER-NATTA catalysts l). Each of these catalysts has been developed entirely empirically to a high degree of activity and selectivity. In the frame of our comprehensively mechanistic investigations of the allyl nickel complex catalyzed butadiene polymerization we were able to show that the technical nickel catalyst NI(02CR)2/BF3 OEt2/A1Et3, which was developed by the Bridgestone Tire Company under the guidance of J. Furukawa 2, already in the middle of the sixties, in its structure can be described adequately as a polybutadienyl nickel(I1) complex coordinated to a polymeric fluoroaluminate anion 3). With the synthesis and characterization of the cationic CI2-allyl nickel(I1) complex [Ni(C12Hl,)][B(C,H3(CF3)2)4] 1 as a highly ac-

-

316 R. Taube, S . Wache and J . Langlotz

tive catalyst for the 1,4-cis polymerization of butadiene 4) we have proved conclusively, that in accordance with the reaction model derived by us for the ally1 nickel complex catalyzed butadiene polymerization 5 ) the cationic polybutadienyl butadiene nickel (11) complex [RC,H,N~(C~H,)I+ is the real cis-catalyst, cf. reaction scheme in Figure 1

.

Figure 1. Reaction scheme for the catalysis of the 1,4-cis polymerization of butadiene with the CI2-allyl nickel(I1) cation from the complex 1 as precatalyst in a non-coordinating solvent like toluene, benzene, dichloroethane or without solvent in liquid butadiene. Accordingly, the cation [Ni(Cl2HI9)]+, which is present mainly in the thermodynamic more stable syn-form b, reacts during a very short initiation period via the less stable but more reactive anti-form a under insertion of butadiene to the anti-polybutadienyl complex a. In consequence of the very rapid anti-syn isomerization this complex also does exist in equilibrium (cf. K3) with the more stable syn-complex d , which must be regarded as the stable store complex under conditions of polymerization. With butadiene the polybutadiene complexes 0 and f are formed as the real catalysts. By the much higher reactivity of the less stable anti-complex Q formation of cis-units are catalyzed in ac-

30. Mol Mass Regulation in 1.4-cis-Butadiene Polymerization 3 I7

cordance to the so called anti-cis and syn-trans correlation. Since all the equilibria can supposed to be rapid the insertion reaction of butadiene k2c has to be taken as the rate determining step in the catalytic cycle. Thus, the catalytic activity is determined thermodynamically by the concentration of the r14-cis-butadiene complex in the anti-form e and kinetically by its reactivity k2c. Therefore a naturally limit to catalytic activity is given by the coordination of n-bonds from the growing chain to the nickel, which have to be substituted by the coordination of butadiene according to the equilibrium K2. The insertion reaction of butadiene into the allyl nickel(I1) bond takes place under formation of a o-bond between the terminal C-atoms of both components in a correspondingly modified n-coordinated state 7 ) , and the new ~ ~ - b u t e n ygroup l is obtained in the anti-configuration (cf. complex g) in accordance to the principle of strongest interaction and least structure variation, respectively. Furthermore, to avoid a highly unstable transition state for the insertion step by coordinative unsaturation at the nickel, it has to be supposed, that the next n-bond from the growing chain is coordinated to the nickel and supports the insertion step energetically thereby essentially. For sterical reasons the coordination of the next double bond could take place easier in the anti-complex e than in the syn-complex f giving rise to the cis selectivity, for which from the CURTIN-HAMMETT principle 9, the relation Sc,t = k2c (k2tK5)-1 can be derived 6 , lo). Besides catalytic activity and selectivity the molecular weight regulation is the third important aspect in controlling the polymer properties. We succeeded now in completing our reaction model on the allyl nickel complex catalyzed 1,4-cis polymerization of butadiene by elucidating also the mechanism of the molecular weight regulation 11)

-

EXPERIMENTAL Butadiene polymerization. The polymerization of butadiene was investigated in toluene as solvent with the C12-allylnickel(II) complex [Ni(C12H19)] [B(CsH3(CF3)2)4] 1 as the catalyst 4 , under the following variations of the experimental conditions: 5.1 M. Nickel concentration Butadiene concentration [BD],: 1.3 2.7 [Nil: 0.45 M. Conversion C = ([BD], [BDj)/[BDl0: 0.30

-

-

0.55.

-

-

318 R. Taube, S. Wache and J. Langlotz

Reaction temperature T and reaction time t: 25 OC (30, 60 min); o Oc (30, 180, 240 min); 40 OC (3, 5, 7, 10 min); 50 OC (5 min). Besides the turnover number TON = [BDIoC/[Ni]t in mOl BD/(mol Ni ah), as the measure of catalytic activity, the cis-trans selectivity SCit, the conversion C and the polymerization degree n has been determined 3, 4, 6). 0-Bydride elimia8tion. To generate the diene end group in a proper high concentration a short chain polybutadiene with Mn k 500 600 g mo1-l was synthesized by using the less active C12-a1lylnickel(I1) complex [Ni(C12Hlg)][FB(C6F5)3] 2 as the catalyst 12).

-

11.3 mg of the complex 2 were added to a solution of butadiene in

150 ml toluene with [BD], = 0.1 M (Ni : BD = 1 : 1000). The polymerization has been carried out under shaking at room temperature. After 4 hours the reaction was stopped by adding 10 pl concentrated HClaq, and all the solvent was destilled off. From the obtained polybutadiene 0.4 g were solved in 10 ml toluene, 0.15 ml (1 mmol) of the azodicarbonicacid diethylester C2H50C(0)N=N(O)COC2H5 5 h up was added, and the yellow reaction solution was boiled 4 to nearly complete decolouration. Then the solvent was removed, the remaining viscously oil was washed several times with methanol and solved in CDC13 to get a 0.1 M solution of the tetrahydro-1,2diazine derivative, which was identified by its 13C NMR spectrum in comparsion with the spectrum of the same DIELS-ALDER product prepared under identical conditions with penta-l13(E)-diene. 13C NMR (CDC13, 22,l MHZ, 25 OC) 6 156.0, 154.0, 128.6, 124.9, 67.6, 61.7, 61.4, 38.4, 13.8, 13.7; DIELS-ALDER product from penta-l13(E)-diene 6 155.1, 154.6, 128.7, 121.9, 49.7, 61.6, 61.3, 42.0, 13.4, 13.4, 17.8.

-

RESULTS AND DISCUSSION Tha catalytic propertieo of 1. Under standard conditions, M I [BD], = 2 M I T = 25 OC and t = which means [Nil = 2 30 min the polymerization in toluene gives the following results: C k 0.5; TON = 1.2 lo4 mol BD/(mol N i * h); Scit: 93 % cis, i(,, k 7.5 lo4 g mol'l, z 3.2 lo4 g mol'l, and &/xn = 2. For the mechanistic investigations the stability of complex 1, which can be handled on air for a short time without decomposition, and its relatively good solubility is very usefull, but by high activity some limits in the variation of the reaction condi-

xn

30. Mol Mass Regulation in 1,4-cb-Butadiene Polymerization

319

tions were given. Tha r8tm of the C h 8 h prop8g8tion. If in the reaction mechanism of the ally1 nickel complex catalyzed 1,4-cis polymerization of butadiene (Scheme I) all the equilibria are rapid and the insertion reaction of butadiene via the reaction channel k2c is the rate determining step the rate of chain propagation rp can be described by the rate law rp = kp[Ni][BD], where kp is the propagation rate constant. Provided that the whole amount of nickel in the used catalyst 1 is catalytically active and that there is no deactivation during the reaction time t, then with the relation [Nil = [Niltot = const. the rate law takes pseudo first order rp = k[BD] where k = kp[Ni]. Since the concentration of butadiene is given by [BD) = [BD], .-kt the conversion of butadiene C = ([BD],,-[BD] ) / [BD], is described by eq (1): c = 1 e-kt = 1 e-kp[Ni]t (1)

-

-

From eq (1) follows eq (2) f o r the propagation rate constant simple transformation:

kp =

$

by

2,3 1 19[Ni]t 1-C

4

Figure 2. The dependence of the conversion C = ([BD]o-[BD])/[BD]o from the product of the nickel concentration [Nil in mol’l and the reaction time t in seconds for 12 runs of polymerization at 25 OC.

320 R. Taube, S. Wache and J. Langlotz

Figure 2 shows the determined dependence of the conversion of butadiene lg(l(1-C)) from the product of catalyst concentration and reaction time [Ni]t/2.3. From the slope of the regression line kp = 3 1 mol-ls-l is found and the correctness of eq (2) is proved. At 0 OC the propagation constant decreases to $ sa 0.5 1 mo1-ls-l and at 50 OC it increases to kp = 11 1 mol-ls-l. Tho rat. of tho tranafor reaction. Without any transfer reaction the polymerization degree 6 should be given by the so called theoretical chain length Y accordingly to eq (3): V = [BD]oCINi]-l

(3)

-

Since a ratio V / ( E ) = 5 12 was found in dependence of the reaction conditions a transfer reaction of the catalyst to the monomer must take place giving rise to the formation of more than one polymer chain during the reaction time t. The most probable course of the transfer reaction is outlined schematically in Figure 3. After the formation of the 0-C3 polybutadienyl complex A a hydrido-diene complex B can be formed by Bhydride elimination. The hydrido-diene complex B reacts quickly with butadiene under substitution of the polybutadiene from the nickel and formation of a crotylnickel(I1) complex C by butadiene insertion into the hydrido-nickel bond.

[a

f‘ i

+

H’:

i-

iA

B

C

Figure 3. Reaction scheme for the 8-hydride elimination as the rate determining step of the transfer reaction of the catalyst to the monomer. The formation of the 1,3-diene end group could be established unequivocally by trapping the diene in a DIELS-ALDER reaction with

30. Mol Mass Regulation in I +cis-Butadiene

Polymerization

321

.

the strong dieneophil azodicarbonicacid diethylesterl3 (cf Experimental Procedures). The transfer reaction can be described kinetically by eq (4):

If the R-hydride elimination kl assumed to be the rate determining step in the transfer reaction followed by the more rapid substitution and insertion reaction k2, then by application of the BODENSTEIN principle for the hydrido-diene complex B the rate law for the transfer reaction ru = kl[Ni] can be derived. The preposition is that complex B is a very reactive intermediate, whose concentration can be regarded as constant in the stationary state of the reaction course, and whose reversed reaction can be neglected correspondingly to the relation k,l << k2[BD]. With the rate of the transfer reaction also the formation rate of the polymer chains is given, which is defined by rii = d[X]/dt, where [XI means the molar concentration of the polymer chains. After the reaction time t the concentration of polymer chains generated by the transfer reaction is given by [XI = ku[Ni]t with ku = kl. Additionally every catalyst molecule bears one growing chain. Therefore the polymerization degree A is obtained as the ratio of the converted amount of butadiene [BD],C and the total concentration of the formed polymer chains bonded to nickel or being free [Nil + [XI, according to eq (5):

..

-

_

[Nil + [XI

_

~ __

(5)

[Nil (1 + kut)

-

By introducing the theoretical chain length cf. eq (3) and after the proper transformation, eq (6) is obtained for the rate constant of the transfer reaction:

Figure 4 shows the relation between the parameters 3- F i and kist for 4 polybutadiens, which were prepared under standard conditions at 25 OC.

322

(v

R. Taube, S. Wache and J. Langlotz

- ii). 103

8

0

0,2

0,4

0,6

0,8

1

1,2 ii.t.106

1,4

[s]

Figure 4. Relation between the theoretical chain length v = [BDIoC/[Ni] and the polymerization degree 5 according to eq (6) for 4 runs of polymerization under standard conditions at 25 OC. By the slope of the regression line the rate constant k~ = s-l is given. For 0 OC and 50 OC the values of the rate constant are kli sa 5 s ' l , respectis-l and Q z 1,5 6

5,

vely. Thus, % shows a similar dependence of temperature as and the ratio of both rate constants kp/Q can be regarded as independent of temperature to a good approximation under the given reaction conditions. Tho control of thm polymmrin8tion dogroo. Elimination of the reaction time t by combination of eqs (2) and (6) gives eq (7) or eq (8) for the polymerization degree E:

According to eq (8) 'i can be calculated in dependence of the initial butadiene concentration [BDIo, the nickel concentration [Nil = [Niltot, and the conversion C. The validity of eq (7) follows from Figure 5, where the relation between the parameters(\)/'ii)- 1 and 2.3 [lg(l/(l-C))]/[Ni] is shown.

30. Mol Mass Regulation in 1,4-cis-Butadiene Polymerization

323

From the reciprocal slope of the regression line the ratio of the rate constants kp/kiizz 635 1 mol-1 can be obtained. The experimental error amounts about loo 1 mol-l. Y - 1

n

0

2

4

6

a

10

Figure 5. Relation between the modified polymerization degree ( v /E) 1 and the quotient from conversion and the nickel concen-

-

tration 2.3 [lg(l/(l-C))]/[Ni] according to eq (7) for 10 runs of polymerization in the range of temperature between 0 and 50 OC.

Conaluaiona. Thus, it has been established for the first time that in the case of the 1,4-cis polymerization of butadiene, catalyzed by the C12-allylnickel(II) complex 1, the polymerization degree or chain length n can be exactly described by a simple chain propagation-transfer reaction model with the 0-hydride elimination as the rate determining step for the transfer reaction of the catalyst to the butadiene, which starts the generation of a new chain. With the given ratio kp/kU as a specific parameter of the catalyst and in the frame of the usable experimental conditions the polymerization degree could be varied between 5 = 400 till n = 2500 in agreement with the reaction model. Thus, it can be concluded that the whole amount of the catalyst 1 is indeed catalytically active. Furthermore the deactivation of the catalyst, which has been observed at a longer time of reaction 14) ,can be neglected to a very good approximation under the experimental conditions used in this work. As we have found the polymerization degree can be increased till to ii = 3000, that is the range of technical interest, by mass polymerization in liquid butadiene at 0 OC. Under these conditions

324 R. Taube, S . Wache and J. Langlotz

the concentration of butadiene has reached the maximal value of 12 M, but the ratio kp/kU decreases to ca. 280 f 20 1 mol-1. Very probably this is a medium effect. Furthermore in the solution polymerization the ratio of the constants kP /kU shows an interesting dependence on the concentration of the catalyst. If the nickel concentration is increased to M the ratio kp/kU increases also by one order of mag[Nil = nitude in consequence of a corresponding decrease of the transfer constant kU. Under these conditions the polymerization degree n approachs the theoretical chain length V at corresponding short reaction time. In this way the polymerization degree can be increased also to an extent as it is characteristic for the technical nickel catalyst 3 , 15). The decrease of the transfer constant kU with increasing catalyst concentration could indicate some influence of the anion on the rate of the 8-hydride elimination, possibly by an increasing penetration ion pair formation 16). A similar effect has been established by us in the case of the 1,4-cis polymerzation of butadiene with the C12-allylnickel(II) complex [Ni(C12H19)I P F as ~ the catalyst 1 7 ) . If the concentration of the anion is increased by adding the corresponding tetraethylammonium salt NEt4PF6 in a ratio Ni : X = 1 : 100 the ratio kp/kU increases by the factor of 3 cf. the results given in Table 1. Table 1. Mol mass regulation in the 1,4-cis polymerization of butadiene by [Ni(C12H19)]X catalysts

30. Mol Mass Regulation in I ,4-rk-Butadiene Polymerization

325

Obviously the interaction with the anion opens an additional possibility to regulate the molecular weight very sensitively and we investigate this effect further to deepen our understanding of the catalytic structure-reactivity relationship and to design tailormade catalysts for the 1,4-cis polymerization of butadiene of technical interest.

Acknowledgment. We thank Prof. Dr. G. MUller from the Fachhochschule Merseburg for the polymer characterization, and we are indebted to the Deutsche Forschungsgemeinschaft and to the Fonds der Chemischen Industrie for financial support. REFERENCES 1. L. Porri, and A. Giarrusso, in G. C. Eastmond, A. Ledwith, S. RUSSO, and B. Sigwalt (Eds.), Comprehensive Polymer Science, Vol. 4, Part 11, Pergamon Press, Oxford, 1989, p. 53 108 2. J. Furukawa, Pure Appl. Chem. 42, 495 (1975) 3 . R. Taube, J. Langlotz, G. MUller, and J. MUller, Makromol. Chem. 194, 1273 (1993) 4. R. Taube, and S. Wache, J. Organomet. Chem. 428, 431 (1992) 5. R. Taube, U. Schmidt, J.-P. Gehrke, P. Bahme, J. Langlotz, and S. Wache, Makromol. Chem., Macromol. Symp. 66, 245 (1993) 6. R. Taube, and J. Langlotz, Macromol. Chem. 194, 705 (1993) 7. R. Taube, S. Wache, J. Sieler, and R. Kempe, J. Organomet. Chem. 456, 131 (1993) 8. R. Taube, P. Bohme, and J.-P. Gehrke, J. Organomet. Chem. 399, 327 (1990) 9. J. I. Seemann, Chem. Rev. 83, 83 (1983) 10. R. Taube, J.-P. Gehrke, P. Bohme, and K. Scherzer, J. Organomet. Chem. 410, 403 (1991) 11. R. Taube, S. Wache, and H. Kehlen, J. h e r . Chem. SOC. submitted for publication 12. R. Taube, S. Wache, and J. Sieler, J. Organomet. Chem. 459, 335 (1993) 13. W. Adam et al., Angew. Chem. Int. Ed. Engl. 19, 762 (1980) 14. S. Wache, Dissert., Reihe Chemie, Verlag Shaker, Aachen 1993 15. T. Yoshimoto, K. Komatsu, R. Sakata, K. Yamamoto, Y. Takenchi, A. Onishi, and K. Ueda, Makromol. Chem. 61, 139 (1970) 16. G. Boche, Angew. Chem. Int. Ed. Engl. 31, 731 (1992) 17. R. Taube, and J. Langlotz, publication in preparation

-

This Page Intentionally Left Blank

327

31. Syndiotactic Polypropylene

l'l'hiomura, M.Kohno, N.Inoue, Y.Yokote, M.Akiyama, T.Asanuma* R.Sugjmoto*

, S.Kimura *,and M.Abe *

Central Research Institute, Mitsui Toatsu Chemicals, Inc., 1190 Kasama-cho, Sakae-ku, Yokohama 247 "Osaka Research Laboratory, Mitsui Toatsu Chemicals, Inc., 1-6 Takasago, Takaishi, Osaka 592, Japan ABSTRACT The joint research and development program of syndiotactic polypropylene (SPP) by Mitsui TOatSii Chemicals and Fina group have resulted in successful large scale production in Finals' existing plant. Preceding to the plant trial, it was necessary to retrofit the catalyst to the plant requirements, reducing the catalyst cost down to the tolerable level. Seemingly minute modifications in the ligand structure revealed the unexpected effects in polymerization.

As was

pointed by some authors, SPP was dogged by the difficulties in its processing due to its sluggish rate of crystallization. We have analyzed this behavior and found out that, by blending with isotactic polypropylene (IPP), SPP could be processed with the conventional machines under the conventional working conditions, keeping intact the favorable features of SPP. INTRODUCTTON The discovery of the syndiotactic polymerization of propylene with zirconocene catalysts(1) has attracted the attention from several research centers(2). Patent literature is already abundant in relation to the improvements in the catalyst performance (activity, tacticity, molecular weight, molecular weight distribution, fluff morphology, prevention of the reactor fouling, etc.). The notable features of SPP are its transparency, resilience and anti-irradiation property, while the sluggishness in solidification by its melt

328 T.Shiomura

processing have posed serious problems.

Parallelling the improvements

in the catalyst and its adaptation to the mass production plant, the efforts to enhance the marketability of SPP have been continued by both companies. CATALYST The first syndiospecific metallocene catalyst which Ewen, Razavi, et al. at Fina Oil and Chemical Co. reported in 1988(1), gave SPP with rather small molecular weight and showed the limited H 2 response. (Table 1)

We have tried zirconocenes wi.th varying ligand structure.

(.Table 2) Table 1.

Hydrogen response of metallocenes(4)

Yetallocene

PH2 (ka/cn2) 0 0.3 2.3

UeZCtCp.P I u) ZrCl2

Ye2C (Cp. 2.7-d i - t -BuPlu)ZrC I2

-

Table 2.

Tm (C)

450,000 498.000

(dl/R) 0.82 0.78 0. 71

137 136 I37

0.3 1.2 2.3

82.000 516.000 703,000 656.000

0. 81 0.60 0.58 0. 49

134 134 135 136

0 0.3 1.2 2.3

55.000 386.000 369.000 543.000

2.35 1.91 1.67 I. 24

132 130 131 130

0

PhPC(Cp. Plu)ZrC12

v.

I.

Activity (e/E/h) 188,000

m n n . condition

Temperature: 60 C Duration: liquid pool

1 h

I h

Effect of substituents on the fluorene ring(5)

Yetrllocene

YAO

Tp

(me) 2

(me) 420

(C)

(8)

40

PhZC(Cp.2.7-di-t-Bu Plu)ZrCl2

1

170

PhZC(Cp.2.7-di-TYS Plu)ZrCIZ

2 2 2

Ph2C (Cp. P I u) ZrC I2

PhBc(c~. H8PI u)ZrCI2

Polrmn. condition:

Yield

I.V.

fm (C)

(d I /I)

74

Activity (R/B/h) 37,000

4.18

138

40

79

79.000

2.35

147

320 320

40 (20

78 39

39.000 419.000

2.32 2.71

142 147

410

60

202

101,000

0.69

(

)

rrrr (-)

0.892

0.935 0.580

L i q u i d pool (4 toluene. propylene pressure: 3 kg/cmZ-C) Duration: 1 h. Propylenp(1iquid): 0.75 I(except 4 )

Table 1 shows the polymerization results with sever1 metallocenes. Under the similar condition, effect of H2 on the activity and molecular weight of the polymer obtained varied greatly with the seemingly slight modification on the ligand structure.

Table 2 reports the

31. Syntheses and Properties of Syndiotactic Polypropylene

Figure 1. H-NMR spectra of metallocenes

I

Ad-

b) PhZC(Cp. 2.7-di-t-Bu FldZrC12

c) Ph2C(Cp. 2.7-d i -TMS F I u) ZrC12

329

330 T. Shiomura

serious tacticity loss incurred by partial hydrogenation of the fluorene ring. (TMS:trimethyl silyl; H8:octahydro) Concerning the use of the so-called ionizing ionic compounds to replace methyl alumoxane (MAO), we have discovered the use of metallocene halide pretreated with organometallic compounds (typically, AlEt3) in combination with the ionizing ionic compounds (typically, Ph3C.B(C6F5)4)

was more convenient way(61 than the popular, but

tedious way which used the neutral metallocene (typically, metallocene Table 3 shows that in this purpose AlMe3 fell far below

dimethyl(7)).

higher alkyl substituted aluminum compounds and that halogen substituted aluminum alkyl was deleterious.

The catalyst activity of our

system was higher than the standard system rising MAO, although there was some decrease in molecular weight (expressed as intrinsic viscosity, I . V . ) of the product polymer. Table 3.

Use of the ionizing ionic compound as cocatalyst(6) ---Me2C(Cp,Flu)ZrC12

Zirconocene

E cpd. Molar ratio

AIR3

(mR)

2.5 2.5 2.5 2.5 2.5 2.0 2.0 2.0

2.0 2.0 2.0 2.0 2.0 2.0 2.5 2.5

/ A1R3 / Ph3C.B(ChF5)4

(DR)

50 AIMe3 AlEt3 80 AliBu3 140

(me) 11 11 11

AIOct3 250

11 11 4.7 4.7 4.7 4.7

AIEtZCI 80 AlEt3 AlEt3 AIEt3 AlEt3 AlEt3 A1813 AlEt3 AlEt3 AIEt3

MA0 MA0

22

37 65 86

2.1

43 43 43 43 43

4.7 8.4 13 21

340 mg 670

Polmn. condition

Zr: Al:B 1:120:2 1 :120:2 1:120:2 1:120:2 1:120:2 1: 40:l I : 70:l 1:120:1 1:lsO:l 1: 80:0.5 1: 8O:l 1: 80:2 1: 80:3 1: 80:5 1 :1000 1:2000

Yield (e) 2. 3 158 104 153 0 8.5 84 73 50 40 83 149

131 104 118 131

system---

A c t i vi t y (fl/R/h)

500 31.000 22.000

31.000 0 2,100 23.000 18.000 12,000 10.000 23,000

37.000 32.000 26.000

24,000 26,000

1. 16 1.1R

I. 12

142 143 138

I . 20

I . 32 1.14 I. 07 1. 14 1.14 1.14 I . 17 1.12 1.44 1.40

142

148

t o l u e n e 1 I: p r o p y l e n e pressure: 3 kg/cm2; temperature: 20 C: duration: 2 h

: solvent:

The idea that the roll of MA0 is binary: first, to alkylate zirconocene dichloride and secondly, to act as the strong Lewis acid to detract the alkyl group to give zirconocene cation which is the true catalytic species suggested the partial replacement of costly MA0 by AIR3 would work.

In this respect, we found the specific structure

31. Syntheses and Properties of Syndiotactic Polypropylene 331 of the metallocene was the key factor in the effectiveness of such re-

placement(8).

In Table 4 , the parent zirconocene ( Me2C(Cp9Flu)ZrC12

)

worked poorly in the new system, whi.1e the modified zirconocenes adapted themselves more satisfactorily to the new system and made it possible to dramatically reduce the required amount of MA0 to maintain the high catalytic activity.

Table 5 shows AlMe3 was poor in this

capacity compared to higher alkyl analogs. Table

4.

Partial replacement of MA0 by AliBu3 MA0

Zirconocene

AliBu3

YeZC(Cp. Plu)ZrCl2

2.0 1.0

(me) 400 5.3

Ye2C(Cp.2.7-di-t-Bu Plu)ZrCl2

3.0 1.0 1.0 1.0 1.0

640 1.1 5.3 5.3 21

0 72 18 72 72

2.0 1.0

420 5.3

72

(me)

Ph2C(Cp. Plu)ZrCl2

Activity

Tm

(g)

(ke/e/h)

(C)

84 2.0

127 2

143

18

4403 13 56 78 85

497 13 58 78 215

149 151 151 150 150

0.95 1.01 0.91 0.98 0.97

80 60

74 40

37 40

140 148

4.18 4.67

0 72

20 60 17 60 60 80

0

Polrmn. condition: liquid propylene 0.85 I (ezcept for temperature 40 C

Table

5.

I.V. (dl/e) 1.27 1.35

Yield

Duration (min)

(ad

# :

3.0 I )

Effect of A1R3 species

Ye2C(Cp.2.7-di-t-Bu Plu)ZrCI2

(me)

MA0

Yield

AIR3

(me)

(08)

(sf

Activity (ka/e/h) 5.6

Tm (C)

I . v. (dl/e)

148

0.88

151

0.98

1.0

421.3

1.0

5.3

AIYe3

27

1.0

5.3

AlEt3

42

33

33

l4g

0.97

1.0

5.3

AliBu3

73

78

78

150

0.98

1.0

5.3

AlOct3

135

66

66

150

0.93

5.6 4.5

4.5

Polrmn. condition: liquid propylene 0.65 I ; temperature 40 C Zr:YAO:AIR3 molar ratio = 1: 50 : 200 ( t 2r:YAO = 1: 250)

332 T.Shiomura CRYSTALLIZATION B E H A V I O R The s p h e r u l i t e s o b s e r v e d b y t h e p o l a r i z e d m i c r o s c o p e f o r compression-molded

s h e e t s ( 2 mm t ) o f i s o t a c t i c P P ( I P P ) ,

isotactic

random c o p o l y m e r ( C 2 c o n t e n t = 4 . 9 w t % ; h e r e a f t e r d e s i g n a t e d a s MFL) a n d SPP a r e shown i n t h e same s c a l e ( F i g u r e 2 ) . s h o w i n g SPP p r o d u c e d v e r y small s p h e r u l i t e s . I n F i g u r e 3 , l i g h t s c a t t e r i n g p a t t e r n s o f i s o t h e r m a l l y c r y s t a l l i z e d ( a t 50 C ) s h e e t of MFL, SPP a n d SPPIMFL

80120 mixture a r e shown, t h e l a t t e r m i x t u r e g i v i n g r i s e t o e x t r e m e l y small quadri-foil.

Figure

4 shows t h e p r o g r e s s i o n of s p i n o d a l decom-

p o s i t i o n f o r SPPIMFL 50150 s y s t e m a t 240 C .

A f t e r t h e p r e p a r a t i o n of

t h i n s h e e t of h o m o g e n e o u s l y b l e n d e d s p e c i m e n ( s o l u t i o n b l e n d e d a n d t h e s h e e t w a s h e a t e d up t o 7 4 0 C , h e l d a t t h i s temp-

precipitated),

e r a t u r e f o r t h e p r e s c r i b e d time, t h e n q u e n c h e d . The p r o g r e s s of t h e s p i n o d a l d e c o m p o s i t i o n was o b s e r v e d u n d e r p h a s e - c o n t r a s t

microscope.

T h u s , w e c o u l d f o l l o w t h e i n i t i a l f o r m a t i o n of s m a l l , u n i f o r m l y d i s p e r s e d d o m a i n s ( a ) , t h e s u b s e q u e n t g r o w t h o f t h e d o m a i n s ( b ) , and t h e development of t h e i n t e r - c o n n e c t e d domains ( c ) . t h e c o m p o s i t i o n r a t i o 50/50 was i n c o m p a t i b l e

SPP a n d I P P i n

e v e n a t 240 C .

Crystall-

i z a t i o n b e h a v i o r was a l s o examined w i t h DSC. The homogeneously b l e n d e d s p e c i m e n w a s h e a t e d up t o 220 C , h e l d f o r t h e p r e - s e t t o crystallize.

time t h e n c o o l e d

The r a t e of h e a t i n g a n d c o o l i n g was + / -

F i g u r e 5 a ) a n d b ) show t h e r e s u l t s f o r SPP/MFL = 70/30.

20 C l m i n . . In this

c o m p o s i t i o n r a t i o , MFL c r y s t a l l i z e d f i r s t a t ca. 100 C , t h e n d i d SPP a t c a . 70 C .

When t h e s p e c i m e n was s e t t o c o o l i m m e d i a t e l y a f t e r

i t r e a c h e d 220 C , t e m p e r a t u r e of c r y s t a l l i z a t i o n ( T c ) f o r MFL was

9 8 C , a n d 74 C f o r SPP.

On p r o l o n g e d h e a t i n g , Tc f o r MFL i n c r e a s e d ,

w h i l e b r o a d e n e d e x o t h e r m p e a k a n d d e c r e a s e d T c f o r SPP r e s u l t e d i n . SPPIMFL = 8 0 / 2 0 b e h a v e d d i f f e r e n t l y ( F i g u r e 5 c ) a n d d ) ) .

In t h i s

c a s e , T c c o r r e s p o n d i n g t o MFL d i d n o t a p p e a r a n d t h e s i n g l e peak c o r r e s p o n d i n g t o SPPappeared a t 69 C and r e m a r k a b l e c h a n g e i n t h e form of e x o t h e r m p e a k was n o t a p e a r e d a f t e r 1 h o u r k e e p i n g a t 220 C. hase-contrast

The p

m i c r o s c o p e d i d n o t e v i d e n c e t h e o c c u r e n c e of s p i n o d a l

d e c o m p o s i t i o n a t t h i s b l e n d c o m p o s i t i o n e v e n a f t e r 1 h o u r a t 160 C . However, a t l o w e r t e m p p e r a t u r e , t h e s e p a r a t e g r o w t h of I P P a n d SPP c r y s t a l s were o b s e r v e d by XRD. The a d v a n t a g e b r o u g h t a b o u t w i t h SPP/ IPP b l e n d i s t h e a c c e l e r a t e d r a t e of c r y s t a l l i z a t i o n , and t h e e x c e l l e n t t r a n s p a r e n c y of t h e f a b r i c a t e d p r o d u c t s . F o r p r a c t i c a l conveni-ence, t h e e x t r u d a t e f r o m Melt I n d e x e r ( 2 3 0 C ) was d i p p e d i n a b a t h k e p t a t

3 1. Syntheses and Properties of Syndiotactic Polypropylene 333

Figure 2 Spherulites developed i n the compression molded p l a t e n a ) IPP (homo)

(

2m t h i c k )

IPP (random: MFL) c ) SPP b)

Figure 3 L i g h t scattering patterns o f t h e isothermally r r y s t a l l ired sheet a ) MPL b ) SPP C)

SPP/MPL=80/20

334 T. Shiomura

Figure 4 Progression of spinodal decomposition f o r SPPIMP1=50/50 at 240 C alafter 5min., h)af ter 10rnin.. C)after 2Ornin. (Phase-contrast rnicrnscope!

Figure 5 Progression of spinodal decomposition at 220 I: for SPP/MF1=70/30 and 80/20 a) SPP/MF1=70/30: cooled immediately after heated u p to 220 C. 11) ditto: after holding f o r 1 h a) SPP/MF1-80/20: cooled immediately after heated u p t o 220 C. h) ditto: after holding for I h

a)

b)

C>

d) L

b

3 I . Syntheses and Properties of Syndiotactic Polypropylene 335

the preset temperature and time for the onset of turbidness (denoted

T) and time for complete solidification (denoted as S) were measured. Figure 6 shows the nucleation effect of IPP in SPP. as

PROCESSING

AND

APPLICATION

Blending of SPP with IPP has provided a real break-through for the processing and fabrication of SPP, which otherwise was difficult to handle with the conventional processing machines in the conventional operating conditions(9). Thus, the transparent articles with moderate rigidity can now be injection-molded without the trouble of sticking of SPP to the mold cavity.

Film and sheet can be casted with the conventional machine

settings, and even fabrication of blown film of unusual transparency are produced by quenching either by water or by air.

Applied to

thermoforming process, sheet can be turned into transparent tray. More interesting is the success of calendering of SPP (11). since IPP was hitherto notorious by its poor performance in calendering (lo).

Trans-

parent SPP will, we hope, find its way into the field occupied by plasticized PVC. Another feature of SPP j s its resistance to irradiation.

Embrittlement and discoloration can be avoided by using

SPP in place of IPP (12).

In Figures 7 and 8, several items

fabricated from SPP are illustrated.

TABLE 6. Mechanical properties of the blended materials SPH1002 SPGl5O SPC151 SPC152 SPC153 SPC154 SPC155 IPP IPP 100 80 70 60 80 70 60 IPP(homo:MI =8) 20 30 40 tlOO IPP(random;MI =8) 20 30 40 ttl00 174 185 368 255 164 216 230 176 kg/cm2 160 Tensile Yield Str. 566 600 616 618 %00 450 456 X 443 509 Elongation 75 2 234 220 105 50 50 65 DuPont Impact Str.23C $.cm(l/O"D) 103 -1oc <1.5 lzod Impact Str. 23C lip.cm/cm NB NB 10.8 7.5 NB NB NB 3.5 12 -1oc 2.3 2.3 2.2 2.1 2.2 2.2 2.1 1.6 Flexural Modulus I(g/cmZ 5400 7100 8200 9600 6200 7100 7000 16400 7600 Strength 201 272 308 350 242 273 269 501 276 Rockwell Hardness R 74 85 92 96 83 85 85109 85 Vicat Softening T. C 111 115 118 120 111 113 113 153 124 Heat Deflection T. C 82 75 83 91 72 76 80 112 81 Gloss x >loo >lo0 >loo 95 >I00 >lo0 >lo0 88 89 Transmittance x 91 88 75 66 90 82 74 82 84 Haze x 32 13 27 41 5 13 25 88 57

SPP(MI= 10)

Remarks

tlPP(hm:MI=4). t*lPP(random:MI=1.5). NB : do not break

336 T. Shiomura

Figure 6 Timc for crystallization

Figure 7 Calenderfd sheet and cast shee! from SPP

7 1 : .

i ,

1

0

..__I

..

I i i

.

20

TABLE 7

.

s

--4

~

.. ..

..

$0 60 i c (C)

100

80

Figure 8 Injection molded articles from SPP a) stewed caps. b) s y r i n g e barrels. c ) tumblers

Comparison of

injection molding conditions Machine: JSW JlOOE-C5 SPP-A IPP (MI=21) (MI-10) Temperature Cylinder C 190/210 190/210 Nozzle C 210 210 Mold C 30 30 Pressure Injection % Boost X

70 40

50 30

Cycle time Inj. /Boost sec Cooling sec

5 30

3 8

WeiRht

B

7.4

7.3

3 1. Syntheses and Properties of Syndiotactic Polypropylene 337

Literature

1) J.A.Ewen, L.Jones, A.Razavi and J.D.Ferrara, J. Am. Chem. SOC., 110 , 6255 ( 1 9 8 8 ) ; J.A.Ewen, M.Elder, L.Jones, L.Haspeslagh, J.Atwood,S.Bott and K.Robinson, Makromol. Chem. Makromol. Symp. ~

48/49

,

253 (1991)

2 ) (a) Y.Chatani, H.Maruyama, K.Noguchi, T.Asanuma and T.Shiomura, J.

Polym. Sci., Part C, Polym. Letters,B, 393 (19gO);

Y.Chatani.

H.Maruyama, T.Asanuma and T.Shiomura, ibid., Part B, Polym. Phys., 29,

1649 (1991);

T.Asanuma, S.Nakanishi, T.Shiomura and T.Kanaya,

Sen-i Gakkaishi.49, 260 (1993);

T.Asanuma, T.Shiomura, Y.Hirase,

T.Matsuyama, H.Yamaoka, A.Tsuchida, M.Ohoka and M.Yamamoto, Polym. B u l l . , z , 79 ( 1 9 9 2 ) ;

T.Asanuma, Y.Nishimori, M.Ito, N.Uchikawa

and T.Shiomura, ibid.,B, 567 ( 1 9 9 1 ) ;

T.Asanuma, Y.Nishimori,

M.Ito, and T.Shiomura, Makromol. Chem., Rapid Commun.,g,

315

(1993)

(b) E.Shamshoum and D.Rauscher,"MetCon

(lgg3)"',

173- (1993);

' 9 3 (Houston), May 26-28

E.Shamshoum. S.Kim, L.Sun, R.Paiz, M.Goins

and D.Barto1, "SPO '93 (Houston), Sept. 21-23 (1993)", A.Razavi, D.Vereecke, L.Peters, D.V.Hessche,

205 (1993);

K.Den Dauw,

L.Nafphiotis and Y.de Froimont, ibid., 105 (1985); H.N.Cheng and J.A.Ewen, Makromol. Chem.

190,1931

(1989);

J.A.Ewen, M.J.Elder,

R.L.Jones, S.Curtis and H.N.Cheng,"Catalytic Olefin Polymerization", Kodansha(Toky0)-Elsevier,

439

(T.Keii and K.Soga, Eds.),

(1990)

R 30, 319 M.Antberg, V.Dolle, S.Haftka, J.Rohrmann, W.Spaleck,

(c) S.Haftka and K,KOnnecke, J. Macromol. Sci.-Phys., (1991);

A.Winter and H.J.Zimmermann, 48/49,

Makromol. Chem. Makromol. Symp.,

333 ( 1 9 9 1 )

(d) G.Balbontin, D.Dainelli, M.Galimberti and G.Paganetto, Makromol. C h e m . , m , 693 ( 1 9 9 2 ) ; P.Sozzani, M.Galimberti and G.Balbontin, Makromol. Chem., Rapid Commun.,&, 305 (1992); P.Sozzani, R.Simonutti and M.Galimberti, M a c r o m o l e c u l e s , x , 5782 (1993) ( e ) G.R.Hawley, T.G.Hil1,

P.P.Chu,

R.L.Geerts,

S.J.Palacka1

H.G.Alt;' " S P d - ' P 3 (Houston), Sept. 21-23 (1993)'",

91 (1993)

3)

Chemical Week, May 18, 7 ( 1 9 9 3 )

4)

Mitsui Toatsu Chemicals, Inc., Jpn. Appl. No. 04-138.960

and

338 T. Shiomura

5)

Mitsui Toatsu Chemicals, Inc., Jpn. Appl. No. 03-713,419;

05-074,

229; 6)

Mitsui Toatsu Chemicals, Inc., W092/01723

7 ) Fina Technology, Inc., Jpn. Kokai 03-179.005;

03-179,006

8 ) Mitsui Toatsu Chemicals, Inc., Jpn. Kokai 04-02-8,703 cf. Idemitsu Kosan Co. Jpn. Kokai 60-217.209; Mitsui Petrochemical Ind. Jpn. Kokai 6 3 4 9 , 5 0 5 9 ) Mitsui Toatsu Chemicals, Inc., EP 414,202 cf. EP 466,926.; EP 419,677; EP 414,047; EP 428,972; EP 4 5 1 , 7 4 3 cf. Jpn. Appl. No. 05-274,072; 05-274,073; 05-271,694; 05-270.136;

lo)

05-274,074; 05-271.693; 05-266.875; 05-262,429; 05-275.440 P.Prentice, Polymer, 22, 250 ( 1 9 8 1 ) ; F.Altendorfer and

A.Wolfsberger, Kunststoffe, cf.R.D.Leaversuch,

80, 691

(1990)

Modern Plast. Int., Aug. 1 6 ( 1 9 9 1 ) ; J.Ogando,

Plast. Technol., Feb. 110 ( 1 9 9 3 ) 11) Mitsui Toatsu Chemicals, Inc., Jpn. Kokai 05-162.158 1 2 ) Mitsui Toatsii Chemicals, Inc., EP 431,475; Jpn. Kokai 03-250,030

339

32. Syntheses and Properties of Syndiotactic Polystyrene

F.ISHMARA*. AND M. KURAMOTO** * Central Research Laboratories, IDEMITSU KOSAN Co.,Ltd., 1280 Kami-izumi Sodegaura, Chiba 299-02, Japan **Polymer Research Laboratory, IDEMITSU Petrochemical Co., Ltd., Anesaki-Kaigan, Ichihara, Chiba, 299-01, Japan

ABSTRACT Homogeneous titanium compound and methylaluminoxane(MAO) system is an effective catalyst for syndiospecificpolymerization of styrene. A comparison of the stereoregularities of the polypropylene and the polystyrene formed by various metallocene catalysts is studied. (CgH6)2C(rl-CgHq)(rl-C9H6)TiC12 / M A 0 system give homogeneous catalyst, for the polymerization of propylene giving isotactic rich polypropylene and of styrene to give syndiotactic polystyrene. Heterogeneous titanium compound containing halogen makes a mixture of isotactic and syndiotacticcomponents.The amount of syndiotacticpolystyrene obtained is dependent on the molar ratio of A1 to Ti. The result of ESR measurement suggests that Ti 3+ species are important as a highly active site for producing syndiotactic polystyrene (SPS). Syndiotactic polystyrene (SPS) is a new crystalline engineering thermoplastic. With a melting point of 270 "C and its crystalline nature, SPS has high heat resistance, excellent chemical resistance, water/ steam resistance. The rate of crystallization is very fast in comparison with isotactic polystyrene (IPS),thus, SPS can be used in a number of forming operations, including injection molding, extrusion and thermoforming.

INTRODUCTION The control of stereoregulaxity is practically important both in the development of new polymers or tailor-made polymers and in the control of polymer properties. When a vinyl monomer (CH2=CHR) is polymerized, the three types of polymers can be obtained ; Atactic, Isotactic and Syndiotactic. When there is a random arrangement of R groups, the structure's

340 N. lshihara and M. Kurarnoto

called atactic. When all the R groups lie uniformly on the same side, the structure's called isotactic. And finally if the R groups occupy positions alternatively above and below the backbone plane, the structure's called syndiotactic. tensivethe Since research discovery concerning of Ziegler-Nana the stereospecific catalyst, poex-

4 1

?A

lymerization of olefins has been carried out. In most cases, isotactic polymers are obtained and syndiotacticpolymers are rare. However, we have I succeeded in synthesis highly syndiotactic poly146.0 146.0 styrene in 1985 in Central Research Laboratories C b m k d ShUl (ppm) of IDEMITSU KOSAN Co., Ltd.1)-2) The 13Cn, 1 m e 7 a w w ~ l p c h d k p m ~Wc I ~ NMR spectra of three types of polystyrenes are ~~~1~~~~~~~~~~~~~~~~~ .given in Fig. 1. Atactic polystyrene is one of the most common plastics in the world. However, the softening point of this polymer's not so high. So, the use of this plastics at high temperature's restricted. Isotactic polystyrene which was discovered by Natta in 1955 is a polymer with a high melting point, 24OOC (degree centigrade). It should be a plastic with high heat resistance. Therefore, many companies tried to industrialize this polymer. However, the crystallization rate of this type of polystyrene is too slow for practical use. Thus this polystyrene has not been industrialized yet. On the other hand, our syndiotactic polystyrene has a high melting point, 270OC. It is higher than that of isotactic polystyrene. Furthermore the crystallization rate of this polymer is so fast that this polystyrene could be industrialized as a plastic with high heat resistance (Table 1). 1

Table 1 The propertles of three types of polystyrenes

Atactlc PS

lsotactlc PS 1955 G. Natta

Syndlotactlc PS 1985

N. lshlhara l IDEMITSU KOSAN C0.LTD.1

Amorphous Crystallization Rate Tg W)

Tm("C)

100

-

Crystalline Slow 99

240

Crystalline Fast 100 270

32. Syntheses and Properties of Syndiotactic Polystyrene

341

RESULTS AND DISCUSSION bperties of SPS Fig. 1 shows the 1 3 C - M spectra of the expanded phenyl C1 carbon of three types of polystyrenes, isotactic,atactic and syndiotactic. In polystyrene the resonance of methylene and phenyl C1 carbon of polymer reflects the conformations of polymer. In particular phenyl C1 carbon provide the best guide to determine stereoregularity of polystyrene. The spectrum of atactic polystyrene shows five main peaks corresponding to its various configurational sequences. The spectrum of isotactic polystyrene shows a sharp singlet at lower magnetic fields corresponding to mmmm pentad configuration. In contrast, the spectrum of syndiotacticpolymer displays a sharp singlet at higher magnetic fields corresponding to rm pentad configuration. The syndiotacticity was more than 99%. Fig. 2 shows you the IH-NMR spectra of the methine and methylene proton signals of the three types of polystyrenes. It was reported that the methylene proton signal of the atactic polystyrene was only a broad resonance and the two methylene protons in isotactic polystyrene were nonequivalent. In agreement with this observation, the spectrum shown here had eight peaks due to the signals of two nonequivalent methylene protons. However, as shown here, the methylene proton signal of the syndiotactic polymer shows only a hiplet. This suggests that two methylene protons of this polymer are equivalent and that the structure’s pure syndiotactic. A well-defined X-ray diffraction pattern of this polymer is quite different from that of isotactic polystyrene. The identity period measured from the fiber spectrum of this structure is about 5.1 A. The result indicates that the crystalline form of SPS has a trans planar conformation like this. Recently, from further investigation it has been suggested that there existed not only a zig-zag planar structure with annealing, but also a helical structm upon crystallization from dilute solution (Fig. 3).DSC and IR observation indicate that a solid-solidphase transition from the p to a form occurs at 190 “C.a-form of SPS is more stable than the pfonn 3). The rate of crystallization of SPS is several orders of magnitude higher than isotactic polystyrene (Fig. 4). Maximum crystallization rate. occurs near 160 “C.The crystallinity of SPS,as well as its hydmarbon nature yields excellent resistance toward moisture, steam and various chemicals. Typical properties for these products are shown in this Table 2. A wide range of products have been formulated with SPS,including glass reinforced resins and I

342 N . lshihara and M. Kuramoto

M. Kobayashi. T. Nakaoki, N. Ishihara; Macromolecules, 22,4377 (1989)

Flg. 3 Schernatlc representationof molecular structures of a-SPS and &SPS.

0.8 h

-E

T

C

0.6

Y

i! 5

-1

1

0.4

0.2

0 0

100

200

300

CrystalllzatlonTemperature ( "C) Fig. 4 Crystalllzatlon rate wlth temperature for SPS

32. Syntheses and Properties of Syndiotactic Polystyrene 343

Table 2 Summary of Physical Propertiesfor SPS Products

Property

Neat Resin

SPS 1WGlasa

30%Glass

Filled

Fllled

I

PET

30%Glass ! Fllkd

SpecHlc Gravlty

1.01

1.09

1.26

I I

1.55

Tenslle strength ( MPa )

35.3

71.6

118

I

152

Tenslle Elongatlon (%)

20

3.1

Flexural Strength ( MPa )

63.7

115

185

Flexural Modulus ( MPa )

2550

4000

9020

lzod Impact ( KJhn2 )

10.0

8.8

2.5

10.8

DTUL 1.82 MPa ( O C )

95

130

251

0.45 MPa ( OC )

110

262

269

2.6

2.8

2.9

Dlelectric const [ 1MHz ] Dlelectrlc loss tangent [ 1MHz I

I

I I I

: I I

I I I

I

I I I

;

2.5 196 9810 8 245 250 3.5

I

< 0.001

< 0.001

<0.001

, I

0.007

impact modified grades, for specific applications. These SPS products exhibit exceptionalelectrical performance, low specific gravity and toughness and modulus competitive with other high heat crystalline engineering thermoplastics.

. .

Stemspecific ~olvmenzatlonof s

~em

The results of polymerization of styrene using different metal compounds with methylaluminoxane as cocatalyst are given in the Table 3. Titanium haride compounds and even titanium compounds lacking halogen atoms can produce syndiotactic polystyrene. The polymer conversion was found to vary according to the ligand on titanium. Their polymerization activities indicate that titanium metallocene complexes with one cyclopentadienyl ligand yield the heighest activity for SPS.The data indicate that substituents on their cyclopentadienyl ligand which are elecaon donor groups generally yield higher polymerization activities. This result suggests stabilization of active site by electron releasing substituents. The cationic nature of Group N B metal active centers for olefin polymerization has been proposed in recent publications 4). Zr compounds were also found to catalyze syndiotactic polystyrene. In comparison with the titanium compounds, the zirconium compounds show lower activity and lower stereoregularity, which could arise from the less electrophilicand the larger ionic radius of zirconium in comparison with titanium. The activity of the Cp*TiC13 is higher than the unsubstituted cyclopentadienyl complex. SPS molecular weight can be controlled primarily by temperature via P-hydride elimination as

344 N . lshihara and M. Kuramoto Table 3

Polymerization01 styrene using various metal compounds with MA0 ~

Catalyst mmol.

40 40 40 40 30 45 30 30 25 25 25

0.05

0.05 0.05 0.05 0.05 0.05 0.05 0.05

0.05 0.05 0.05 0.01 0.05 0.01 0.05

4.1 2.1 3.8

Syndiotactic Syndiotactlc Syndiotadc Syndiotactlc Syndiotactlc Syndlotactlc SyndlotacNc Syndiotadc Syndiotactic Syndiotactic Syndiotactic Syndlotactic Atactic Atactic Atactic Syndlotactic Syndiotactic Atactic Atactic Atactic Atactic

9.5

99.2 100 1.o

40 10

2.0 5.4 5.7 6.0 0.4 0.4 0.1 1.3 3.2 2.0 1.4 0.7 1.8

20

80.8

40

40 8

0.05 0.2 0.02 0.05 0.02 0.25

~~~

[All Conversion Stereospedndty mmoi. wt%

40 28 16 40

1 1

1 1 2 2 2 2 2 2 2 1 3 1

1 6

5 4 4 4 4

Polymn. cond. St / Toluene (rnlIrnl); I ) , 6) ( I 8 0 /loo), 2) (23 150). 3) (100 I50) 4) (50 / IOO), 5) (40 I go), Polyrnn. Temp., time; 1) 4) 50 "C. 2hn 5) 90°C, 4hn 6) 20 OC. 10 days

-

Table 4

Org. Al in rnoldrn'

PolymerlzaUonof styrene using CpTiCI3 with various organoaluminum compounds Meld in g 0.8

TMA TEA TlBA TW

(0.05) 0.1 (0.05) 02 10.05) 0.1 (0.4).+ HzO(0.4) 17.6 TEA (0.4) + HzO(0.4) 0.8 MA0 (0.4) 14.9 MA0 (0.2) + TMA(0.2) 7.1 M A 0 (0.2) + TEA(02) 0.3 MA0 (0.2) + TiBA(0.2) 15.5 CIAO (0.4) 94.7 Me2AIOAIMe2. (0.6) 0.1 0.1 (Me2Ai0)2AIMe (0.5) 0.3 (Me2A10)flo (0.5) MA0 (Mw < 500 ) (0.4) Trace M A 0 (Mw. 500) (0.4) 14.9

COnv. Stereospeclflclty In W Y O

0.5 0.1 0.1 0.1

Atact. Atact. Atact. Atad.

10.8 0.5 9.2 4.4 0.2 9.5 100 0.1 0.1 0.2

Synd. synd. Synd. Synd. Synd. Synd.

9.2

Atact.

Atact. Atact. Atact. Synd. Synd.

CpnCi, 5.0 x 10 4 mc11.dm.~ Toluene 0.1 d d , Styrene 8.7 mldrnJ 50 "C, 2 hr a Me2AK)U + Me2AlCI b 2Me2AIOLI + MeAICI, c 3Me2AIOU+ ACl,

32. Syntheses and Properties of Syndiotactic Polystyrene

Table 5

Polymerization of styrene using various Ti compounds with MA0 Catalyst mmol

TIC13(AA) nc~,(so~vey)

M(I(OEt)2/EB/TlCI4

TICI,

w

w

4

[AI]I[Ti] Conversion wt%

Stereospecificity

Iso.PS + Iso.PS + Iso.PS + Iso.PS +

0.2

1000

8.2 2.0 1.9 0.9

2.0

0.2 0.2

50 500 1 m

2.9 1.1 1.4

40 5 0.2

10

40 500

7.2 0.4 0.7

Is0.PS Iso.PS +

0.3

( Atactic PS) Synd.PS Synd.PS

100

1.0

0.2 lo00 1.0 20

10 50 500

2 2 0.2

with CpTiClg with various organoaluminum compounds is 107 summarized in the Table 4. When T trialkylaluminum was used as cocatalyst, it was not given SPS. As a cocatalyst, methylaluminoxane is 5 10' 3 good for synthesis of SPS. As can 5

5 -

2.5 0.9

Synd.PS Synd.PS Synd.PS Synd.PS

lso.PS(84)+ Synd.PS(16)

Iso.PS(lP)+ Synd.PS(88) Iso.PS(lO)+ Synd.PS(S0)

Synd.PS Synd.PS

-

\

c p w c13

345

346

N. lshihara and M. Kuramoto

rc) Al / Ti

= 1000

Flg.6 Aromatic C1 carbon spectra of polystyrenes obtained with Mg supported TICI, I Malox catalyst In 1,2,4-trlchlorobenezene at 130 O C with a JNMGX-270 spectrometer (at 67.8 MHz)

0 llonaolld 0 Tlln aolutlon

Al I TI molar ratio

Flg. 7 Effect of A1 I TI molar ratlo on TI concentratlon In the solutlon and on the solld

32. Syntheses and Properties of Syndiotactic Polystyrene 347

obtained (Table 5 ). In general, the amount of syndiotactic polystyme increases with increasing a molar ratio of A1 to Ti. Fig. 6 shows a typical %NMR spectra of the polystyrene obtained by heterogeneous catalyst system. This single peak shows highly isotactic sequence and this single peak shows highly syndiotactic sequence. No peak between these two peaks indicates the absence of APS. There would be two types of polymer arising from two different active sites. In order to clarify this problem, some experiments are carried out. M A 0 was reacted with Ti supported on Mg compounds in toluene for about 30 minutes at mom temperature and then the suspension was filtered to separated the solution from the solid phase. The solid phase was washed with n-heptane. Styrene was polymerized using both the solution part and the solid part. Isotactic polystyrene was obtained from the solid part and syndiotacticpolystyrene was obtained from the solution part. These results indicatethat a part of Ti on the carrier was migrated in solution on reaction with MAO, to give a soluble catalytic species that polymerizes styrene to SPS. The amount of Ti that goes into solution depends on the amount of MAO. Heterogeneous catalyst with halogen gives polymers involved isotactic component. The amount is dependent on the A1 to Ti ratio. On the other hand, homogeneous catalyst without halogen does not give isotactic polymer. Furthennore. the titanium concentration of the solid and the solution were investigated. The results are given in this Fig. 7. The titanium concentration in solution increased with increasing the molar ratio of A1 to Ti. It is suggested that the amount of Ti on the carrier that goes into solution depends on the molar ratio of A1 to Ti The effect of a molar ratio of M A 0 to Ti on the activity in the case of CpTiCl3 and MA0 catalyst system was investigated. The activity increases with increasing of the molar ratio of MA0 to Ti. The amount of Ti 3+ species measured by ESR increases with increasing of the ratio of MA0 to Ti, too. This suggests that M A 0 acts as ducting agent from Ti4+ to Ti3+. Ti 3+ might be a active species for synthesis of SPS. MA0 acts as a weak reductant. Therefore, much amount of M A 0 is needed to increase the activity. The addition of triisobutylaluminum to CpTiCl3/ MA0 catalyst was investigated. As a result, it was found that the addition of TIBA increases the activity. However, excess addition of TIBA decreased the activity. The stereochemistry of styrene insertion into the carbon-metal bond of the catalyst has been investigated6). The regiochemistry of insertion can in principle proceed in two ways, primary or secondary. N-propyl benzene was detected by gas chromatography after quenching by methanol. The presence of n-propylbenzene suggests that insertion of styrene proceeds by a secondary process. Ethylbenzene was detected, too. Ethylbenzene arises from reinitiation via secondary insertion after P-hydrogen elimination. And it is observed that the amount of ethylbenzene is larger than that of n-propylbenzene. This fact suggests polymer chain formation is mainly initiated via metal-hydrogen bonds and terminated by P-hydrogen elimination.(Scheme 1) From the investigation of dueterated polymer, the coupling constant of these copolymers show copolymer (a) has a trans confonnation and copolymer (b) is gauche. Independently of

348 N . lshihara and M. Kuramoto

this work, Prof. Wbelli 7) also observed a trans conformation from copolymer (a). It was concluded that the double-bond opening mechanism in syndiotactic polymerization was cis opening (Scheme 2) 8).

6

Tho addlion mod. of I otyrmr mokcuk to I mtaCCH3 bond

n-CHg

+

HmCH2

Tho tmnsfrr roactlon arklng irom p- hydrogrn rllmlnrtkn

6=cH2 I .

[&hydrogen ellmlnatlon

TI5H-CH3

]

Scheme 1 The mechanism of syndiospecific polymerlzation of styrene

-

Cis openlng copolymer(a)

n

By 'H-NMR analysis of a syndiotactic copolymer of perdeuterated styrene

W copolymer(b)

Scheme 2 The mode of additlon of the growing chain end to the double bond of styrene

ene C w These new ansa metallocenes have been applied to polymerization of styrene. Table 6 shows the relationships between the catalyst and stereoregularity of polymer. These titanocenes produced both atactic polypropylene and syndiotactic polystyrene. These chiral metallocenes produced both isotactic polypropylene and atactic polystyrene. However, this

32. Syntheses and Properties of Syndiotactic Polystyrene 349

titanocene could produce marvelously both the stereoregular polymers which isotactic polypropylene and syndiotactic polystyrene. This is the first example that two different stereoregular polymers can be obtained using the same catalyst. As describe before, in syndiotactic polymerization of styrene the secondary insertion occurred. On the other hand, in isotactic polymerization of propylene the primary insertion occurred. Two different types of insertion could occur in the same catalyst. From these results, The mechanism of syndiospecific polymerization of styrene should be different from that of isospecific polymerization of olefin.

Relationship between the catalysts and tho sterooregularitios d the polypropylene and polystyrene products

Tabb 6

Stereoregularlty Catalysts PP

PS

I

Syndlotactk

I

[mr] = 1.0

-f

a

At8ctk

lSyndlotllctlcl [mrl= 1.0

(Syndbtactk [mm] = 0.67

Abtk

I

[mrrl= 1.0

Atactk

[mm] = 0.19

Atactk [mm] = 0.52

Atactk

Atactk

When various ring-substituted styrene were polymerized using CpTiClg with methylaluminoxane catalyst system, the corresponding syndiotactic polystyrenes were obtained The spectra of each atactic polymer show many main peaks corresponding to their various configurationalsequences. The spectra of each isotactic polymer shows a single sharp peak at lower magnetic field corresponding to the mmmm pentad configuration. In contrast, at a high magnetic field the chemical shifts of the phenyl C1 carbon show in each case a single sharp new peak. These results show that polymers are highly syndiotactic. The melting point of syndiotactic poly(ring-substituted)styrene and the corresponding isotactic polymer. Almost all the syndiotactic polymer produced a high degree of crystallinity,

350 N . Ishihara and M. Kuramoto

and all the Tm values of syndiotactic polymers arc higher than those of isotactic polymers with the corresponding substitute. These results imply that the syndiotactic polymer will be useful in industry. Fig. 8 shows the relation between the Hammett's s value of each substituent and reactivities in polymerization. It is observed that monomer reactivity is enhanced by electron-releasing substituents in an aromatic ring. The results suggest that a cationic titanium would be active site for the stereospecific polymerization of styrene to SPS.

substltuents p'BuSt

pMeSt

substltuents pFSt

mMeSt

p-ClSt

mClSt

t

(T

Fig.8 Relation between the Hammetta value for each substituent and reactivities in

polymerization

REFERENCES 1. N. Ishihara, T. Seimiya, M. Kuramoto, M. Uoi, Macromolecules, 1p.2464 (1986). 2. N. Ishihara, M. Kuramoto, M. Uoi, Macromolecules, L 3356 (1988). 3. M. Kobayashi, T. Nakaoki, N. Ishihara, Macromolecules, 22,4377 (1989). 4. C. Pellecchia, P. Longo, A. Proto, A. Zambelli, Makrornol. Chem., Rapid C o m u n . , 12 265 (1992). 209 (1987). 5. M. Kuramoto, N. Ishihara, M. Uoi, Polymer Preprint Japan, 6.N. Ishihara, D. Phill. Thesis, Oxford (1990). 7. A. Zambelli, P. Long, C. Pellecchia, A. Grassi, Macromolecules, 2035 (1987). 8. N. Mitani, N. Ishihara, T. seimiya, T. Ijitsu. T. Takyu, Polymer Preprint Japan, 2 1152 (1988).

35 1

33. The Industrial Synthesis of Bimodal Polyethylene Grades with Improved Properties

L.L. BOHMt H.-F. ENDERLE and M. FLEIBBNER Hoechst AGt 65926 Frankfurt (MI, Germany

ABSTRACT It is well-known for a long time that polyethylenes with broad, bimodal molecular mass distributions have improved properties in respect to processibility, impact properties, and stress crack resistance. Polyethylenes with very broad bimodal molecular mass distributions can be synthesized using new highmileage catalysts with excellent behaviour in average molecular mass regulation by hydrogen and copolymerization with 1-olefines. The appropriate technology is the cascade process. It is possible to generate reactor blends which form a polymer alloy in the solid state. Such materials have properties so far unknown. New types of engineering polyethylene materials are accessible by progress in catalyst development and polymerization technology in combination with a basic understanding of structure - properties relations of the polymer in the solid state. INTRODUCTION High and medium density polyethylenes are standard polymers with a broad range of applications. Therefore a wide spectrum of different grades must be available to fulfill the requirements. For injection moulding one of the basic processing techniques low molecular mass compounds with narrow molecular mass distributions are used because of their fast relaxation behaviour to avoid frozen-in orientations of the mould, and the higher toughness at low melt viscosity. In the solid state impact resistance depend on the number average of molecular mass, whereas melt viscosity is mainly determined by the weight average of molecular mass 112).

-

352

L.L. Bohm, H.F. Enderle and M. Fleissner

However, processing of the melt becomes more difficult with increasing average molecular mass or decreasing melt flow rate 3 1 4 ) for polymers with narrow molecular mass distributions. To overcome these limitations in processing for higher molecular mass and therefore tougher materials the molecular mass distribution must be considerably broadened for these extrusion or blow moulding grades to produce films, pipes and large containers for example. Broadening of the molecular mass distribution strongly decreases melt viscosity at high shear rates 4 1 5 ), considerably reduces extrusion defects like melt fracture 6 1 7 ) and increases melt strength 8-11). Products with broad molecular mass distributions both unimodal and bimodal and their advantageous rheological behavior 12) are well-known and available since longtime 13)

.

New types of catalysts together with state of the art cascade polymerization technology now offer the opportunity to produce polyethylene grades with very broad bimodal molecular mass distributions 14). The material is produced by forming both the low and the high molecular mass fraction during the polymerization process giving a reactor blend. It becomes now obvious that such products show the best compromise in respect to processibility of the melt and solid state properties like impact strength, rapid crack growth resistance and especially longterm stress crack resistance. The extremely low average molecular mass fraction improves flowability of the melt and forms mainly the crystalline phase in the solid state which increases stiffness and Youngs modulus. The high molecular mass fraction is especially responsible for good mechanical properties. If the long chains are further copolymers of ethylene and l-olefines, the stress crack behavior can strongly be improved 15). With the cascade technology it is now possible to design the molecular mass distribution as well as the comonomer distribution to get a material with an excellent compromise of short term and long term mechanical properties. These new materials can be

33. Industrial Synthesis of Bimodal Polyethylene Grades 353

regarded as polymer alloys engineering products for systems 16,171.

which open new applications as example in gas distribution

CATALYST/COCATALYST-SYST~~ TO produce polyethylene on industrial scale using the cascade

technology it is necessary to have an appropriate catalyst/ cocatalyst-system 14). The essential requirement in this case is to generate a low molecular mass fraction with a viscosity number in the range below 80 cm3/g corresponding to a melt flow rate (MFR 19015) 3) of more than 500 g/10 min, and a high molecular mass fraction with a viscosity number above 500 g/cm3. Appropriate catalyst/cocatalyst-systems are improved high-mileage Ziegler-type catalyst in combination with alkylaluminum as cocatalyst because average molecular mass can be regulated with hydrogen. In this case a high-mileage titanium-magnesium-chloride-catalyst together with triethylaluminum as cocatalyst is used. The melt flow rate value depends on the hydrogen/ethylene ratio either in the diluent or in the gas phase of the slurry reactor 18). Fig. 1 shows this correlation for the slurry polymerization process for homo- as well as copolymerization with 1-butene.

Figure 1. Average molecular mass regulation with hydrogen and catalyst productivity as a function of the hydrogen/ethylene ratio in the gas phase of the slurry reactor

354 L.L. Bohrn, H.F. Enderle and M. Fleissner

These experimental results show the excellent hydrogen response which offers the opportunity to cover the whole range of MFRvalues up to more than 1000 g/min, corresponding to average molecular mass values determined by viscometry between 104 and up to 106 g/mol. It is again well-know that hydrogen strongly inhibits the activity of the catalyst/cocatalyst-system as pointed out on fig. 1. But the catalytic productivity (KA) remains still in the range of 10 kg/mmol titanium (corresponding to 5 ppm Ti in the final polymer) under technical process conditions even for MFR-values in the range of 1000 g/10 min. A further improvement can be seen if l-butene is used. l-olefines increase activity in respect to homopolymerization 1 9 ~ 2 0 ) . So from the point of average molecular mass regulation with hydrogen and catalytic activity this catalyst/cocatalyst-systems fulfills the recommendations to be introduced in an industrial process. To generate products with very broad bimodal molecular mass distributions using the cascade technology the catalyst/ cocatalyst-system must produce a polymer with an unimodal and narrow molecular mass distribution. It is known that single-site catalysts like the metallocenes 21-24) form polymers with the Schulz, Flory most probable distribution and a polydispersity q / M n of 2 z5). On the other hand supported high-mileage Zieglercatalysts are multisite catalysts giving mostly polymers with logarithmic normal distributions and polydispersities q / M n much higher than 2 26). To get products with very broad bimodal this means a molecular mass molecular mass distributions distribution with two distinct maxima for a 50:50 % b.w. blend there must be a certain distance between the average molecular mass of the low and high molecular mass fraction. This ratio Mv.2/Mv,1 (MVs2 the viscosity average of molecular mass of the high molecular mass fraction and Mve1 for the low molecular mass fraction) depends on the polydispersity (q/Mn)o of the basic polymer as given by the catalyst/cocatalyst-system 14). The correlation between the polydispersity of the blend (q/Mn) and the ratio Mv.2/Mv.1 is shown on fig. 2 with (S/Mn)o as a variable parameter.

-

33. Industrial Synthesis of Bimodal Polyethylene Grades 355

M".2/M". 1

0'

Mv.2/Mv.1

0.5

I

I 1

ml

Figure 2. Formation of polymers with bimodal molecular mass distributions using the cascade technology The conclusion from this consideration is that the ratio +.2/MV.1 must increase with increasing polydispersity values of the basic polymer (S/Mn)o. However, in all cases the ratio MV.2/$,1 must be high which can only be achieved using a catalyst/cocatalystsystem with excellent hydrogen response together with a two stage polymerization process either discontinuous or continuous. In this case the polydispersity (Q/Mn)o of the basic polymer is about 5. So by changing the mass fraction of the low molecular mass compound ml together with the ratio M,.2/Mv.1 a wide range of polymers with broad molecular mass distributions are accessible as shown on fig. 2 271~4).These results point out that the ratio Mv.2/Mv.1 must be at least higher than 10 to get products with very broad bimodal molecular mass distributions and two distinct peaks (detected by gpc). According to our experience it is impossible to get a homogeneous polymer blend by mechanical mixing of two compounds with an average molecular mass ratio $,2/MV,1 higher than 10. So the mechanical mixing in kneaders of extruders is not a proper way to get high quality very broad bimodal products with excellent performance. But the cascade technology offers this opportunity because both fractions are finely divided in each other inside the small polymer particles formed during

356

L.L. Bohm, H.F. Enderle and M. Fleissner

polymerization. The key point is the particle forming process 18,26,28).

If the catalyst is designed in the proper way the catalyst particle is desintegrated at the beginning of the polymerization process, and the primary particles are spread evenly over the whole and expanding polymer particle 29). Around each primary particle (size appr. 5 to 15 nm) the polymer is formed. In the course of polymerization each polymer particle passes through reactor 1 to get the low molecular mass fraction and the reactor 2 to get the high molecular mass fraction or vice versa. This process is reproduced inside each polymer particle leaving the cascade which means that around each primary particle layers of both polymer fractions are built up. The different residence time of the particles has to be taken into consideration. Consequently inside each polymer particle the low and high molecular mass fraction are finely and evenly distributed in each other as show schematically on fig. 3 .

polymer particle

c

500 pm

log I

Figure 3 . Polymer particle morphology distribution using the cascade technology

and

molecular

mass

For further improvement of polymer properties ethylene must be copolymerized with 1-olefines like 1-butene or 1-hexene. For Ziegler-catalystlcocatalyst system it is again well-known that the r-parameters differ by orders of magnitude 30-32) which only means

33. Industrial Synthesis of Bimodal Polyethylene Grades 357

that the comonomer must be added in excess. This catalyst/ cocatalyst-system fulfills all requirements in respect to polymer particle morphology, polymerization activity, molecular mass regulation with hydrogen, copolymerization with 1-olefines, and molecular mass distribution. REACTOR BLEND FORMATION AND SOLID STATE PROPERTIES The cascade technology offers the opportunity to form reactor blends composed of two very different fractions regarding average molecular mass and copolymer composition 1 4 ) . Polyethylene is a semicrystalline material in the solid state in which the hard crystalline phase and the soft amorpheous phase are connected by tie molecules to build up a physical network. By designing the reactor blend it is possible to generate the appropriate solid state structure with a high number of tie molecules if the coil diameter of the long chains in the melt (radius of gyration) is at least twice the thickness of the crystalline lamellae plus the thickness of the amorphous region 33-37).

Principle of Figure 4. tie molecules formation in a bimodal polymer alloy

The fraction of tie molecules (number of tie molecules/total of macromolecules) in the solid state can be calculated as described elsewhere 34-37). Failure time of a sharp notched bar in glycol under constant stress 38) and the fraction of tie molecules depend in the same way on average molecular mass (fig. 5).

358 L.L. Bohm, H.F. Enderle and M. Fleissner

0.1

100

Figure 5. Tie molecule fraction and failure time as function of average molecular mass

g ,o

f I

aol 0.001

1

0.1 106

i1

0.m1 107

106

MwWW

These data show that long term stress crack resistance depends on the tie molecule fraction in the solid state, and both properties are proportional to each other. Reactor blending consequently offers the opportunity to build up a material which can be regarded in a common held belief as a polymer alloy in the solid state because opposite properties like stiffness or Youngs modulus, impact values and long term stress crack resistance are improved at the same time. If the comonomer (1-butene, 1-hexene) is mainly incorporated in the long chains density is increased for the bimodal material in comparison to an unimodal copolymer as shown on fig. 6. 0.97

Figure 6. Density of unimodal and bimodal copolymers at the same melt flow rate (MFR 19015 = 0.45 9/10 min)

I

,

I

___,1

I-

\

I... .......

-

. ..

o.w

Although density is higher for the bimodal material the comonomer in the long chains is much more effective in respect to stress

33. Industrial Synthesis of Bimodal Polyethylene Grades 359

crack resistance. In comparison to unimodal homo- or copolymers the failure time 38) is higher by approximately 2 orders of magnitude.

Correlation Figure 7. of failure time and melt flow rate MFR 190/5 of sharp notched bars under stress for unimodal homoand copolymers, and bimodal copolymers

Such excellent values can only be realized if the long chains are copolymers with short chain branches along the main chain. It has been shown that chain diffusion within the crystalline lamellae occurs 39). Under stress, crack propagation inside the semicrystalline material is hindered by fibrillation (crazing) at the crack tip. Tie molecules are stabilizing the fibrils provided they are fixed in the crystalline regions. To reduce rapid lateral diffusion of tie molecules through the crystallites a small number of short chain branches must be present along the main chain. This can be achieved by copolymerization with 1-olefines like 1-butene and 1-hexene. Further requirements are high toughness and resistance against rapid crack propagation for pipes under pressure. The relevant parameter is the critical strain energy release rate GIc 4 0 t 4 1 ) . For unimodal homo- and copolymers (density in the range of 0 . 9 4 0 to 0.960 cm3/g) GIc depends mainly on the number average of molecular mass 2 ) . However, polymers with broad bimodal molecular mass distributions have much higher impact values at the same Mn value, obviously due to a stronger contribution of the high molecular mass fraction.

360 L.L. Bohm, H.F. Enderle and M. Fleissner

Figure 8. Impact values GIc as function of the number average molecular mass Mn for unimodal and bimodal homoand copolymers

13

o-homoPo*lmr

15: u

a-10: 5-

0 103

1o4

105

M,Iglmoll

Such materials also show high resistance versus rapid crack propagation measured at pipes under internal gas pressure. This is again very important to avoid damage of a pressurized pipe over a large length initiated by external violence 16142) or stress cracking in defect butt fusion weld. This property is tested on pressurized pipes using the small scale steady state (S4)test 43). The S4-test reveals the minimum internal pressure pc for the brittle pipe failure mode. pc is the critical pressure below which steady state rapid crack propagation cannot sustained. Fig. 9 give the results of measurements on 110 mm/lO mm pipes of the bimodal pipe resin in comparison to a standard pipe material (measurements made at Gastec M I , The Netherlands). Again this test shows a strong improvement of the bimodal material in respect to standard unimodal resins.

Figure 9. Crack length versus internal pressure (S4-test)

33. Industrial Synthesis of Bimodal Polyethylene Grades 361

CONCLUSIONS New developments of high mileage Ziegler catalyst/cocatalystsystems in combination with advanced cascade technology now offer the opportunity for the production of reactor blends with very broad bimodal molecular mass distributions. The blend consists of a low average molecular mass homopolymer and a high average molecular mass copolymer which are finely distributed in each polymer particle during polymerization. By melting these polymer particles a homogeneous melt is formed which transfers to a polymer alloy in the solid state. The low molecular mass homopolymer fraction mainly builds up the crystalline lamellae, and the high molecular mass copolymer fraction expecially forms the amourpheous phase with a high number of tie molecules connecting the crystalline regions in this physical network. This solid state structure can be regarded as a polymer alloy because such materials combine mechanical properties in a manner hitherto unknown. They have high stiffness or Youngs modulus, high impact values, high resistance versus rapid crack propagation and unusual high values in enviromental stress crack behaviour. They represent a new class of pipe materials (PE 100) which can be used for gas distribution systems under pressure 16)

.

ACKNOWLEDGMENT These investigations have been carried out in the plastics R&D department of Hoechst AG, Frankfurt. We want to express our thanks to our colleagues for their support. REFERENCES l.M. Fleissner, Angew. Makromol. Chem. 94, 197 (1981) 2.M. Fleissner, Angew. Makromol. Chem. 105, 167 (1982) 3.ISO (International Standard) 1133-1981 4.C.D. Han, “Rheology in Polymer Processingvv, Academic Press, New York, 1979 5.H. Watanabe, T. Kotaka, Macromolecules 17, 2316 (1984) 6.G.V. Vinogradov, Rheol. Acta 12, 273 (1973) 7. E. Uhland, Rheol. Acta U, 1 (1979) 8.C.D. Han, T.C. Yu, K. Kim, J. Appl. Polym. Sci. 15, 1149 (1971)

362

L.L. Bohm, H.F. Enderle and M. Fleissner

9.W.W. Graessley, M.J. Struglinski, Macromolecules 19, 1754 (1986) 10.C.D. Han, J.Y. Park, J. Appl. Polym. Sci. 19, 3291 (1975) ll.M. Fleissner, Int. Polym. Proc. 11, 229 (1988) 12.J.P. Montfort, G. Marin, P. Monge, Macromolecules 19, 1979 (1986) 13. A.M. Birks, L.E. Dowd, "Polyethylene Filmt8,SPE Technical Papers 25, 714 (1979) 14.L.L. Bohm, D. Bilda, W. Breuers, H.F. Enderle, R. Lecht, in press 15.R. Hayes, W. Webster, Plast. Inst. Trans. 3 2 , 219 (1964) 16.J.M. Greig in "The 1990's and Beyondll, The Plastics and Rubber Institute 1992, p. S3B/l/l 17.H. Backer, R. Dewitt, Kunststoffe 82, 739 (1992) 18.L. Bohm, Chem. Ing. Techn. 56, 674 (1984) 19.P.J.T. Tait, G.W. Downs, A.A. Akinbami, in "Ttansition Metal Catalyzed Polymerizations", Ed. R.P. Quirk, Cambridge University Press, Cambridge, 1988, p. 834 20.R.A. Hutchinson, W.H. Ray, J. Appl. Polym. Sci. 4 l , 51 (1990) 21.H. Sinn, W. Kaminsky, Adv. Organomet. Chem. 18, 99 (1980) 22.J. Okuda, Nachr. Chem. Techn. 41, 8 (1993) 23.M. Bockmann, Nachr. Chem. Techn. 41, 1220 (1993) 24.R. Mulhaupt, Nachr. Chem. Techn. 41, 1341 (1993) 25.L.L. Bohm, J. Berthold, R. Franke, W. Strobel, U. Wolfmeier, in Studies in IISurface Science and Catalysis 25, Catalytic Polymerization of Olefinst1,Eds. T. Keii, K. Soga, Elsevier, Amsterdam, 1986, p. 29 26.L.L. Bohm, R. Franke, G. Thum, in "Transition Metals and Organometallics as Catalysts for Olefin Polymerization11, Eds. W. Kaminsky, H. Sinn, Springer Verlag, Berlin, 1988, p. 391 27.L.L. Bbhm, Angew. Makromol. Chem. 89, 1 (1980) 28.M. Kakugo, H. Sadatoshi, M. Yokoyama, K. Kojima, see (26), p. 433 29.R.A. Hutchinson, C.M. Chen, W.H. Ray, J. Appl. Polym. Sci. 44, 1389 (1992) 30.Y.V. Kissin, Adv. Polym. Sci. 15, 91 (1974) 31.L.L. Bohm, Makromol. Chem. 182, 3291 (1981)

33. Industrial Synthesis of Bimodal Polyethylene Grades 363

a,

32.L.L. Bohm, J. Appl. Polym. Sci. 279 (1984) 33.L.~. Bohm, H.F. Enderle, M. Fleissner, Adv. Mater. Q , 234 (1992) 34.A. Lustiger, R.L. Markham, Polymer 24, 1647 (1983) 35.N. Brown, I.M. Ward, J. Mater. Sci. 18, 1405 (1983) 36.Y.-L. Huang, N. Brown, J. Polym. Sci. B, Polym. Phys. 29, 129 (1991) 37.A. Lustiger, M. Ishikawa, Proc. 11. Plastic Fuel Gas Pipe Symp., Oct. 1989, Am. Gas Association 38.M. Fleissner, Kunststoffe 77, 45 (1987) 39.K. Schmidt-Rohr, H.W. Spiess, Macromolecules 24, 5288 (1991) 40. J.G. Williams, E. Plati, Polym. Eng. Sci. 15, 470 (1975) 41.J.G. Williams, "Fracture Mechanics of Polymers", Harwood, Chichester, 1984 42.W.F. Miiller, E. Gaube, Kunststoffe 7 0 , 72 (1980) 43.P. Vanspeybroeck, "Plastics Pipes VIII'l, The Plastics and Rubber Institute, 1992

This Page Intentionally Left Blank

365

34. Structure and Properties of Ethylene / a-Olefin Copolymers Polymerized with Homogeneous and Heterogeneous Catalysts

S, HOSODA, A.UEMURA, Y. SHIGEMATSU, I. YAMAMOTO, and K. KOJIMA

Sumitomo Chemical Co. Ltd., Chiba Research Laboratory, Kitasode 2-1, Sodegaura-city, Chiba, 299-02 Japan ABSTRACT Structure and properties of various ethylene / or-olefin copolymers were comparatively investigated. Copolymers polymerized with a heterogeneous Ziegler-Natta catalyst showed a wide intermolecular structural distribution, contrary to a narrow distribution of the copolymers obtained with a homogeneous catalyst system. This structural distribution was found to be correspondingly reflected in the distribution on a crystalline level such as a lamella thickness distribution (LTD). Mechanical properties of the copolymers were strongly influenced by the structural distribution; i.e., the narrower the chemical composition distribution and LTD, the stronger the impact strength. This result could be explained from the viewpoints of tie molecular concentration and its orientational homogeniety with deformation. INTRODUCTION Linear low-density polyethylene (LLDPE) has gained a steadfast position during these 15 years as a third commercial polyethylene through its superior properties. LLDPE is now manufactured by various kinds of polymerization process with various types of heterogeneous Ziegler-Natta catalyst. The authors have been investigating the structure and properties of LLDPEs'), and found that a wide intermolecular distribution of chemical composition (CCD) and a wide lamella thickness distribution (LTD) are characteristic to LLDPE, and basic mechanical properties are strongly influenced by the structural distributions. On the other hand, new grades of ethylene / a-olefin copolymer have been recently produced commercially with metallocene catalyst systems, and these copolymers are to have much more excellent film properties like high transparency, high impact strength, etc than those of LLDPE. Then in this article, structure and properties of ethylene / a-olefin copolymers prepared with solid Ziegler-Natta catalysts and homogeneous catalyst systems were comparatively investigated, and the large-strain mechanical strength is discussed from the viewpoints of structural homogeniety and tie molecules.

366 S. Hosoda, A. Uemura, Y . Shigematsu. I. Yamamoto and K. Kojima

EXPERIMENTAL Commercial LLDPEs and copolymers prepared in our laboratory with solid Ziegler-Natta catalyst and soluble catalyst systems were used as samples in this study. Intermolecular structural distribution was obtained by the combination of temperature rising elution fractionation (TREF)4)and GPC/FT-IR system. Degree of short chain branching (SCB) was determined by FT-IR using methyl symmetric deformation band at 1378 cm-' (normal) or 1384 cm-' (iso). Crystalline structure was investigated by DSC, transmission electron microscope (TEM), X-ray diffraction and smal angle X-ray scattering (SAXS). Two-step staining method using chlorosulfonic acid and uranyl acetate was used for TEM observation. Lamella thickness distribution was obtained from the measurement both of lamella thickness and length for 200 - 400 lamellae from various fields of view in TEM micrograph. Stress-strain curves of press-molded sheets were measured at a drawing rate of 50 m d m i n at 25 "C. Tensile impact strength test was carried out following ASTM 1822 at -70, -30, and 25 "C. Extrusion blown films of various thickness were obtained by Placo K-40R manufactured at 170 - 190 "C. Out put was 20 - 25 kg/hr. Optical and mechanical properties of the blown films were measured following ASTM. RESULTS AND DISCUSSION Molecular Structural Distribution Typical example of a bird's eye view of copolymer (comonomer; butene- 1, MFR=2, density=0.9 19) obtained with a heterogeneous catalyst system is shown in Fig. 1. The figure clearly exhibit that the copolymer is composed of main three peak tops. The highest one is characterized by a very distinctive sharp peak of low SCB and relatively high molecular weight. This is a common component of all kinds of LLDPE produced with various polymerization processes, and some superionties in thermal and mechanical properties are considered to be related to the presence of this component. The middle one is the component having the average degree of SCB and molecular weight, and the lowest one is composed of the polymers of highest degree of SCB and lowest molecular weight. The sectional view of Fig.1 at a definite molecular weight gives CCD composed of multipeaks, each of which has been proved to follow a probabil-

Fig. 1. Structural Distribution of LLDPE (comonomer: C4'-l).

Fig. 2. Structural distribution of C,' / C i - 1 copolymer prepared with a metallocene catalyst system.

34. Structure and Properties of Olefin Copolymers 367

ity distribution5)in a similar manner as a high-pressure low-density polyethylene6) produced under uniform polymerization condition. Contrary to the LLDPE sample above, copolymer prepared with homogeneous catalyst system has characteristic narrow structural distribution composed of a sharp single top as shown in Fig. 2. Some of the typical data of the structural distributions are shown in Table 1. The coefficient of variation for CCD, o/X ,of the copolymer prepared with the homogeneous catalyst system ( Q; standard deviation of the degree of SCB, average degree of SCB) is very small (0.3), whereas CCD of LLDPE prepared with heterogeneous catalyst systems is rather wide and o/5varies from 0.5 through 0.9 depending on the polymerization condition and the catalyst used. Further molecular weight distribution (MWD) of the former is narrow (Mw / Mn = 2) compared to that of LLDPEs (Mw / Mn = 4 - 10) which is largely dependent on the catalyst, too.

x;

Table 1 .

Yolecular Structural Distribution o f Ethylene / Butene-1 Copolymers.

sample catalyst

YFR

density

SCB

g/loiin

g/cm3

I/~OOOC

u/Z

Yw/Yn

A

hetero.

1.0

0.921

19.4

0.92

10.3

B

hetero.

1.0

0.921

16.9

0.71

4.2

C

hetero.

0.8

0.919

15.3

0.50

4.1

D

homo.

1.1

0.911

20.0

0.30

2.0

u , standard deviation o f the degree o f short chain branching (SCB) 51, average degree o f SCB

Crystalline Structure Lamellar morphologies were comparatively observed by TEM for two kinds of ethylene / butene- 1 copolymers. As already reported7’, hot-press molded LLDPE exhibits a spherulite structure, and lamellar crystals of various shapes were observed under high magnification. As

Fig. 3. Lamellar morphologies of (a)VLDPE (comonomer; C4’-l) and (b) the copolymer prepared with a metallocene catalyst system.

368

S. Hosoda, A. Uernura, Y . Shigematsu, I. Yarnarnoto and K . Kojirna

shown in Fig. 3 (a), in case of very low-density polyethylene (VLDPE; MFR=l, d=0.900) prepared with a heterogeneous catalyst system, we can clearly recognize two different types of lamellae. One is thick and growing straight, the other is thin and very short as if it was a branch from the former straight and long lamella. On the contrary, the lamellar morphologies of the copolymer obtained with a rnetallocene catalyst system (MFR=I, d=0.910; Fig. 3 (b)) is very simple, and the lamella thickness distribution is narrower than that of VLDPE which spreads over from 16 to 5 nm. The wide distriution of lamella thickness of VLDPE is considered to be attributable to the broad CCD rather than to MWD, judging from the result of TEM observation for LLDPE fractions’). Consequently, the narrower lamella thickness distribution and simple morphology of the copolymers prepared with the metallocene catalyst would be due to their sharp CCD. The differences of lamella thickness and long period manifest themselves in the same manner as the corresponding differences of DSC endothermic curves and SAXS patterns.

Mechanical Properties Some of the mechanical and optical properties of extrusion blown film are shown in Table 2. Copolymers obtained from the homogeneous catalyst system are superior to that prepared with heterogeneous one in the performancies such as impact strength, transparency, and the amount of extract. In this section discussion is focused on the relationship between the largestrain mechanical strength and structural homogeniety.

Table 2. Properties o f VLDPEs prepared w i t h homogeneous and heterogeneous c a t a l y s t systems. s a w Ie

sc- 1

comonomer ca t a I ys t

butene-1 homo.

YFR (g/lOnin) d e n s i t y (g/cm’) Tensile I n a c t Strength ([g- cn/cn) n-C6 EXTRACT (%) Blown F i I n P r o p e r t i e s ” Haze (%I

Gloss (%I

Dart I n a c t (kg-cm/mm) Tear !trength(kg/cm) YD / TD U.S. ( k /cm’) PD / TD

St r e n t

f:

1.7 0.909

sc-2 hexene-1 homo. 1 .a 0.908

ZN- 1 butene-1 hetero. 1.4 0.903

1132

2930

0.3

0.4

7 80 6.5

3.7 130

3.8 125

7.0 105

2920

>4000

400

46 / 60

83 / 120

550 / 410

580 / 560

36 / 74 370 / 290

1 ) U e l t . Ten .: 170 ‘C, Die: 125#-2.0t, B . U . R . : 1.8, F i l m thicEness: 30 p , Output: 20-25 kg/hr

Stress-strain curves of hot-press molded samples of ethylene / butene- 1 copolymers were measured. Comparing two kinds of samples having the same average molecular weight

34. Structure and Properties of Olefin Copolymers 369

and the same density, the copolymer obtained with a homogeneous catalyst system exhibited the higher ultimate strength and the higher stress increase after necking ( 0,). Tensile impact strength ( oT,) was measured for press-molded samples having the density range of 0.9 17 - 0.919 g/cm3, and the results are shown as a function of molecular weight in Fig. 4. aTrincreased with increasing molecular weight and ethylene/ hexene- 1 copolymers showed higher impact strength than ethylene / butene-l copolymers of the same molecular weight. Further copolymers prepared with homogeneous catalyst sytem exhibited higher impact strength than those with heterogeneous catalyst system. As reported already@, distributions of lamella thickness and crystallinity play a key role governing the impact strength of ethylene / a-olefin copolymers; i.e., the narrower the distribution, the higher the strength. Accordingly the above result is quite reasonable considering that the copolymers prepared with homogeneous catalyst system give narrow crystalline structural distribution as described above.

Fig. 4. Tensile impact strength as a function of molecular weight for C2’/C4’-l and C2’/C6’-lcopolymers prepared with solid

Ziegler-Natta catalyst(Z-N)and homogeneous catalyst (Homo.) systems.

In general tie molecules are regarded significant in the mechanical properties of crystalline polymers under large strain. Distribution of the concentration and the orientation of tie molecules in a polymer of solid state are particularly interesting in case of a crystalline ethylene / aolefin copolymer. Then tie molecular ratio was estimated from the two different points of view. One is the method proposed by Huang and Browng), which is based on a molecular dimension in a molten state. The molecules which has the end-to-end distance in the melt equal to or greater than the distance between adjoining lamellar crystals will probably make a tie molecule. The probability of a given end-to-end distance, r, of a random coil is given by

where a is constant and b2 = 3/2 ?. Here Fis the root-mean-square value of r and given by

370 S. Hosoda, A. Uemura. Y . Shigernatsu, I . Yamamoto and K . Kojima

Here n’ is the number of links. For polyethylene, extension factor in melt, D = 6.8 and link length, 1 = 0.153 nm. The probability of a tie molecule of the mono-disperse polymer is given by Eq.(3), where L is a critical distance for rand was chosen as 2lC + lo (lc, lamella thickness OD

r’ exp ( - b 2 r 2 ) d r p

- 31

(3)

-

I, r2

exp (-b’r’)dr

calculated from melting temperature using Thomson-Gibbs equation; la, thickness of amorphous phase). The probability was then calculated for the cross-fractions of various molecular weight and was exhibited finally as a function of the crystallinity (Fig. S)x). The tie molecule fraction shows a convex type dependence on the crystallinity, and it decreases steeply in the region of high crystallinity ( X , > 70 % ) and low crystallinity ( X, < 20 % ) for molecular weight above 50x lo3. Further, tie molecules become substantial for the molecules having molecualr weight above 20x103 and the crystallinity of 40 -50 %. Probability was additionally calculated for two kinds of model cases of different crystallinity distribution as shown in Fig. 6. One case, (a), is a homogeneous structural distribution having the average crystallinity of SO %. Another,(b), is a wide distribution composed of three components each of which has average crystallinity of 77 %, 50 % and 30 %, respectively (average crystallinity is 50 %, too). The former case would be corresponding to the metallocene catalyzed polyethylene and the latter to a typical LLDPE. As a result, tie molecule fraction of the former was calculated to be 28 % for molecular weight of lOOxld and the fraction is much larger than that of the latter case (18 %).

0.14

0.4

I

0.12

.-5 0 . 1 0

0.3

2 0.08

a 0.2

a

.L 0.06

3

2 0.04

0.1

0.02 0.0 0

20

40 60 80 T O O Crystallinity wt%

Fig. 5. Tie molecule probability as a function of crystallinity. MWXIO.~;(a) 99, (b) 50 (c) 20, (d) 10.

0.00 0

20

40 60 80 100 Crystallinity wt%

Fig. 6. Model of crystallinity distribution for the calculation of tie molecule probability. Average crystallinity is 50 % for both. See text for detail.

34. Structure and Properties of Olefin Copolymers

371

The another approach tried in this study to estimate tie molecule fraction is the measurement of molecular orientation in the amorphous phase of polyethylene. Orientation function of the amorphous phase, fa,,,, can be expressed as followslo),

where-f;a,,rJ is the orientation function of amorphous CH, sequences in trans conformation, y denotes the fraction, and subscripts 1, e and 1 are tie, end and loop, respectively. Jam) can be calculated from the infrared dichroism"*12)using the bands at 2015 and 1894 cm-' as follows,

where pc is the fraction of crystalline CH, sequences in rruns conformation as expressed in the form of Eq.(6)"),

&,,)was calculated for a drawn sample from the infrared dichroism for two kinds of ethylene As shown in Fig. 7, sample prepared with a homogeneous catalyst exhibited higher degree of orientation in the amorphous phase compared to the sample with a heterogeneous catalyst. The fraction of each amorphous component was calculated usingham)at the drawing ratio (A) of 5 and shown in Table 3. The results clearly indicate that the sample of homogeneous structure possesses much more tie molecules at a drawn state than the sample of a wide structural distribution. This can be easily explained from the crystallinity dependence of tie molecule orientaion for ethylene / butene- 1 copolymers"). It is reasonable to consider how many tie molecules of the same degree of / butene-1 copolymer of the same molecular weight and density.

Table 3 . Number Fraction o f LOOP, END and T I E in the Amorphous Phase o f Drawn Polyethylenes ( 1 = 5 ) .

Y.W. x10'

density g/cm'

(C)

110

0.919

(F)

99

(K)

114

sample

Fraction

LOOP

END

TIE

0.60

0.13

0.27

0.919

0.52

0.14

0.34

0.919

0.47

0.06

0.47

samples (C) and ( F ) : prepared with heterogeneous Ziegler-Natta c a t a l y s t . sample ( K ) ; prepared with homogemneous catalyst: EXTENSION RATIO

Fig. 7. Orientation function of amorphous CH sequences in trans conformation for c,*~c,I copolymers.

372 S. Hosoda, A. Uemura, Y. Shigematsu, 1. Yamamoto and K. Kojima

orientation exist in a drawn state, regardless of being intrafibrillar or interfibrillar, strongly influences the mechanical strength of the sample. Considering the remarkable crystallinity dependence, the distributions of both the concentration and the orientation of tie molecules in the samples having homogeneous structure would be narrow because of the narrow distributions of the crystallinity and the lamella thickness. The copolymer of homogeneous structure is thus considered to contain much more actually efJicienr tie molecules in the deformed state than those of wide structural distribution, and eventually exhibit a higher mechanical strength.

References 1)s. Hosoda, "Trends in Polymer Science", Council of Scientific Research Integration (India), in press (1994) 2)C.S.Speed. B.C.Trudel1, A.K.Mehta, F.C.Stehling. Polyolefins, VII, 45 (1991) 3)K.W.Swogger, SPO '92, 153 (1992) 4)M. Aoyagi, Y.Sato, S. Hosoda, A. Uemura, Bull. Inst. Chem. Research. Kyoto Univ., 69, 155 (1991) 5)s. Hosoda, Polymer J., 20.383 (1988) 6)K. Shirayama,T. Okada, S. Kita, J. Polym. Sci. A-3.907 (1965) 7)s. Hosoda, K. Kojima, M. Furuta, Makromol. Chem.. 187. 1501 (1986) 8)s.Hosoda, A. Uemura, Polymer J., 24, 939 (1992) 9)Y-E. Huang, N. Brown, J. Polym. Sci., Polym. Phys. Ed., 28,2007 (1990) 1O)W. Glenz, A. Peterlin, J. Polym. Sci., A-2, 9, 1191 (1971) 11)s. Hosoda, Makromol. Chem., 185,787 (1984) 12)s. Hosoda. M. Furuta, Makromol. Chem., Rapid Commun., 2.577 (1981)

313

35. Progress in Gas Phase Polymerization of Propylene with Supported TiCI, and Metallocene Catalysts

K.D. HUNGENBERG, J. KERTH, F. LANGHAUSER, P. MULLER BASF AG, Plastics Laboratory, DS7056 Ludwigshafen

ABSTRACT Product demands for polypropylene polymers cover a wide range of properties. The application of process, process parameters and catalyst (TiCl, and metallocenes) as tools for tailoring product properties is demonstrated in terms of molecular mass, molecular mass distribution, melting behaviour and mechanical and optical properties of hornopolymers, random copolymers and bi-phasic impact polymers.

INTRODUCTION Polypropylene is one of the fastest growing bulk polymers, and more and more changes from a commodity to an intelligent polymer with specialized applications. This is mainly due to the fact, that its properties can be adjusted over a wide range; from stiff to very tough, opaque to transparent '.'),

melting point

from 120 to 166"C, and so its properties can be tailored to fulfil the special demand of the customer.For the production of such a versatile range of tailor-made products, two preconditions must be fulfilled. One of them is an appropiate process. Fig. 1 shows a typical plant for the Novolen' Gas Phase

-

Process, consisting of a two reactor - cascade, which is primarily designed for the production of impact modified polypropylene, but it is also well suited for the production of simple random or homopolymer with high catalyst yields by increasing catalyst residence time, or for the production of polmers with broadened molecular mass distributions, etc.. Moreover, the flexibility of this process makes it possible, to adjust product properties by variation of some few process parameters.

Flgurel. Scheme of reactor cascade for the stirred bed gas phase process

The other presumption is an appropiate catalyst. Modern MgC12-suppottedTiCI, catalysts 'I enable us to produce a wide variety of products with high catalyst yield. In addition, the new metallocene catalysts 57)

374 K.D. Hungenberg. J. Kerth. F. Langhauser, P. Miiller offer access to new polymeric structures and enlarges the possibilities of tailoring product properties. But to be used in a heterogeneous process, all catalysts have to be supported in a way, that the morphology of the growing polymer grain is preserved during polymerization to ensure its handling in the plant. While for TiCI,-catalysts this problem is solved, either by prepolymeridng or by using inorganic supports like silica4),for the metallocene catalysts there are still various possibilites under investigation. Table 1 shows some of these methods. The support can either be "inert", that is the catalytic system, metallocene and methylalumoxane, is just impregnated and so only physically adsorbed, or one of the catalyst components is chemically fixed on the support. We have tested these various possibilities in batch experiments, and compared some product properties and judged the processability of the resulting polymer with respect to fines, coarse grains, lumps, temperature control and so on. From these batch experiments, methods 111 and IV are further developed and adopted for use in continous plants as in fig. 1.

Tabk 1: Cornpartson of dinerent methoda of supporting Me,Si(2-Me-lnd)JtCIJMAO

M,[kg/mol] Mu/Mn

SUPpofl

m.p. YC]

None

260

2.7

154

PP-powder

123

2.4

145

Processability

AIR3-treated

sio2

I

Chemically fixed MA0

Chemically fixed Metaliocenel2)

220

172

I

I

2.0

3.0

I

l

144

144

I

l

++

I

+

~102

With these preconditions fulfilled, it is now possible to tailor product properties either by process or by catalyst and this paper will show how both strategies can be used to get a variety of different target compounds with speaal emphasis on the now well established high-mileage supported TiCI, catalysts and the new metallocenes.

TAILORING MOLECULAR MASS Beside the above mentioned classes of polypropylene, molecular mass and molecular mass distribution are the most important criteria for classifying PP. So, for industrial purpose, one major interest lies in very simple methods to adjust molecular mass and distribution, without, if possible, changing other polymer properties simultaneously For MgCIJTiCI, catalysts, fig. 2. demonstrates the flexibility of a gas phase process in handling even

35. Progress in Gas Phase Polymerization of Propylene

375

extremely high hydrogen concentrations and so offering access to a very wide range of grades simply by

Z'

Fig. 2: Dependence of M, and M JMn on hydrogen concentntbn and temperatun for MgCljTiCl, Fig. 3: Dependence of M, and tactkny

on tempenture for MgClJrCl,

varying the hydrogen partial pressure in the reactor, or, as an additional parameter, the reaction temperature. The shape of the molecular mass distribution (D=M, I M, ) remains fairly constant over the whole range of molecular masses. Alp0 polymer tacticity and properties related to it are constant over this range@, or increase slightly with temperature (fig.3). So, for the well established Ziegler catalysts, these process parameters are best suited to adjust molecular weight independently. On the other hand, for the new metallocene catalysts, one of the most important problems, which has to be solved before any chance of commercialization, was to get high enough molecular weights under

technical conditions. In this case, tailoring product properties, i. e. molecular weight, by tailor-made catalysts has been very successful (tab. 2).

Table 2. Dependence of mokcular m a r s and melting

point on l i a n d structure of I (acc. to m, at W C

R

R

R

m.p.

M,

1 II

H

H

Me

141

60000

Me

i-Pr

Me

152

460000

4

u

A systematic variation of the substituent pattern of I led to a class of catalysts covering a range of about 1 decade in molecular mass. These drastic effects on molecular weight are attributed to a reduction of Lewis acidity by the alkyl groups in 2- or 4- position, which in turn reduces chain termination by P-H-abstraction.

376

K.D. Hungenberg, J . Kerth, F. Langhauser, P. Miiller

TAILORING MOLECULAR MASS DISTRIBUTION Besides molecular mass, the molecular mass distribution determines processing and application, and so PP-grades may be classified according to there mmd or dispersion index D-MJM,: Standard grades:

014-6

Controlled rheology PP:

D = 2.5-3.5 for injection moulding, fibres films etc.

High melt strength PP:

D > 10 for blow moulding, thermoforming etc.

Scheme I summarizes several possibilities for the production of these products, supposing either metallocene catalysis which usually gives polypropyleneswith D-2, or supported TiCl,catalyst, which gives D-5. Scheme I: Variatin of mokcukr mars dhtribution of polypropylene depending on catalyst and procem used

Narrow distributions can easily 'be obtained either from metallocene catalysts directly, or from PP with Ziegler type catalysts by peroxidic degradation in the extruder. Standard grades are readily accessible from Ziegler catalysts, but for metallocene catalysts the mmd of the single catalyst must be broadened either by blending polymers of different M, cascade operation of two or more reactors, or bi-centered catalysts. This also holds for the production of high melt strength grades with Ziegler catalysts lo).

I

Fig.4:

Change of M, and D with comporltiin of a binary blend of 2 potymen. Mm,,=3MX)O,Mm,,=300000 A,

Fig. 5:

A for Do=5, 0,O

for Do=2of the blend componantr

Broadened rnmd with TCi, in a reactor cascade or with bi-metallic catalYst in a singk reactor

To attain these broad mmd with metallocene catalysts may be difficult, because the polydispersion index

D=MJM, as a measure of breadth of the mmd for a mixture of polymers is D =Xw,*M,, I w Y n j=Do I w Y , ,

XWNn,,

So, the maximal dispersion index is limited by the dispersion index Doof the individual mmd's and their

35. Progress in Gas Phase Polymerization of Propylene 377 molecular weights. Fig. 4 shows the dependence of the resulting M, and D of a binary system of 2 very different polymers on the relative mass fraction of the low molecular weight component for two different values of Do . For Do=2, a typical value for metallocene catalysis, D=5 is the maximum for the combined distribution, whereas with D,=5 (Ziegler catalysts) values of D>10 can easily be obtained. Fig. 5 shows, that such a broadening can either be attained by tailoring the catalyst or the process. Here a bi-metallic TiN - Ziegler catalyst is used in a single reactor. The same distribution can be made with a usual TiCI,catalyst, but in a reactor cascade, where the 2 reactors from fig. 1 are operated at different hydrogen concentrations and temperatures. So there exist several possiblities of tailoring a product; the best way, catalyst or process variation, must be determined in every case individually. TAILORING MELTING BEHAVIOIUR Beside propylene homopolymers, random copolymers with minor amounts of ethylene or butane become more and more important. These polymers are tougher and more flexible than homopolymerswith a much lower melting point down to 12O'C, depending on the amount of comonomer. The decrease in melting point is caused by disturbing of the helical conformation by the incorporated ethylene unit"). The same effect results, when the disturbing ethylene unit is replaced by another group, which does not fit into the polypropylene helix. Here one can take advantage of one feature of the metallocene catalysts. Whereas most of the metallocenes show a very high stereospecifity, which is comparable or higher than that of TiCI,-catalysts, their regiospecifity is usually less than that of TiCI,-catalysts. So, practically all polypropylene types produced with metallocenes show some misinsertions other than the normal 1-2-link, for example 1-3- or 2-1-links. Tabb 3: Comparison of two b w melting polymen

1 properties by process

properties by catalyst

PP with Me,Si(2-Me-Benzind)&C12

E P Copolymer with TiCI, 146

145

M , [kglmol]

190

200

AHrn,[Jhl

91

99

cristallinity,[%I

54

59

85

92

~

mmmmpent.,

[%I

10 c,

2.3 2-1-links

shear modulus, [N/mm?

570

705

vicat soft. temp., I'C1

72

misinsertionllOOOC

hexane extract, 1%1

1.5

~~

a4

I

-

-

-

~~

~

~~

0.2

Table 3 gives a Comparison of two polymers. one produced with Me2Si(2-Me-Benzind)&C~and the other one was produced with TiCI,. The properties of the latter one, melting point, molecular mass and mmd were adjusted to rather the same values as for the metailocene product simply by some process

378

K.D. Hungenberg. J. Kerth, F. Langhauser, P. Muller

parameters, that is copolymerization with ethylene and peroxide degradation of the reactor product.

So it is obviously possible to tailor polymers with the same basic properties (melting point, molecular mass) either by process or by catalyst, but a closer look to table 3 reveals some differences. The metallocene polymer, though having a lower melting point, behaves similar to a Ziegler homoplymer, when regarding properties like vicat softening or stiffness, which are related to cristallinity. Cristallinity is rather high, because contrary to the Ziegler polymer, misinsertions are homogeneously distributed over all chains, so decreasing crystall stability and melting point, but not the overall cristallinity. This homogeneous distribution of misinsertions becomes clear from fig. 6. The metallocene polymer is homogenous in molecular mass, but also in crystall stability, that is fractionation temperature, whereas the Ziegler random copolymer shows a broad range of fractionation temperatures, the different fractions also containing different amounts of comonomer. These differences in homogeneity are the reason for one of the main advantages of the metallocene polymer; it shows a very low value for hexane extract according to FDA regulations.

arbitrary unit

I homopolymer

Fi. 6:

temp. [‘ ‘cl

.

Y

302

‘

[g/mo~]

EIP-nndom copolymer

HetemgeneHy in chemical comporttion and mokcuhr mars from cmrc hsctionatlon by TREF and GPC of EIP- copolymer with WI, and homopolymer with Me~i(Z-Me-Benzind)@I,

TAILORING BI-PHASIC POLYMERS The third class of PPgrades are the block or impact-modifiedpolymers. This is a typical case, where polymer properties are tailored by process. Stiff homopolymer from the first reactor of a cascade of stirredbed gas phase reactors (fig.1) is modified in the second reactor by polymerizing an EIP rubber within the homopolymer grains. The rubber fractions can be increased to more than 50%. Product morphology and properties strongly depend on the rubber fraction (fig. 7).The rubber forms discrete particles until both polymers are present in roughly the same amounts; then a structure like an interpentrating network develops. Low temperature impact strenght is drastically increased with the rubber fraction. Another possibility to adjust properties by process, is the variation of the matrix of such impact polymers. For a new class of impact polymers the homopolymer of the matrix is modified by feeding small amounts of ethylen to the first reactor, and so producing a random copolymer as the matrix of the impact polymer. Depending on the amounts of comonomer, the melting point can be varied over a wide range, and also stiffness and impact resistance can be adjusted (fig. 8). Moreover, the optical properties of the impact polymer can be changed drastically. Usually, bi-phasic polymers are not transparent, because

35. Progress in Gas Phase Polymerization of Propylene 379

Fig. 7: Morphology and properties of impact polymers wfih increasing rubber fraction from 2. reactor

of the large, dispersed rubber particles. By adequate adjustment of process parameters however, such random impact polymers can be made transparent.

Fig. 8: Dependence of stiffness and impact strength on rubber fraction from

2. reactor and melting point of the matrix polymer in 1. reactor: Transparency of 15% rubber product: homopolymer matrix :

8%

EIPcopolymer matrix:

76%

But process is not the only tool to vary impact polypropylene; also the catalyst plays an important role. This can best be demonstrated by looking at the performance of the different generations of catalysts. In table 4, the most important features of impact polymer production with Stauffer-TiCI,, MgCIJTiCI, and a metallocene are compared. Obviously, the potential of TiCI,-catalysts is much less than that of supported TiCI,; maximum rubber content is limited, mechanical properties are much behind, and properties can not be varied independently. With the TiCI,-catalyst all important properties, rubber content, molecular mass of rubber and matrix, and stiffness of matrix can be vaned independently over a wide range simply by some process parameters in either one of the two reactors. For the newest class of catalysts, the metallocenes, a further tool to vary polymer properties is the ligand structure. The results for Me2Si(2-Me-Benzind)?rCl, given in tab.4 are rather new and by far not complete, but it demonstrates that it is now possible to produce impact polymers in continous reactors with metallocenes, and the potential of these new class must now be explored.

380 K.D. Hungenberg, J. Kerth, F. Langhauser, P. Milller Tsbk 4: Performance of 3 d h n n t catalysts In impact polymer pmductbn. Pr6p.rU.a

1

8n for

20% rubber.

Stauffer-Tic14 benzoate co-catalyst -25% lZOd (-2O'C)

6 U/m'

500 N/mm'

<200 rubber fraction and parameters

molecular weight not independent

SUMMARY It was shown that catalyst and process are important took for tailoring polypropylene grades. By adequate

application of these took, a wide variety of polymer structures in terms of molecular mass, molecular mass distribution, comonomer content and distribution, microstructure of propylene links etc. are accessible, leading to a broad range of application properties like melt flow behaviour, stiffness, impact strength, melting point, extractables, optical properties etc. REFERENCES 1. K.D. Hungenberg, J. Kerth, F. Langhauser, B. Maranke, R. Schlund, in: G. Fink, R. MOlhaupt, H. H. Brintzinger (eds.), "40 Years of Ziegler Catalysis", to be published at Springer, Heidelberg, 1994 2. P. Galli, J.C. Haylock, Prog. Polym. Sci.,

Is,443 (1991)

3. E. Seiler, B. G(lller, Kunststoffe, Bp, 1085 (1990) 4. EP 288845, EP 306867 5. EP 519237 6. J.A. Ewen, J. Am. Chem. SOC.,$J@ 6355 (1984)

7. W. Kaminsky, K. KOlper, H. H. Brintzinger, F. Wild., Angew, Chem., Int. Ed.,

a,507 (1985)

8. W. Spaleck, M. Antberg, J. Rohrmann. A. Winter, B. Bachmann, P. Kiprof, J. Behm, W. Hermann Angew. Chern.,m, 1373 (1992) 9. K. D. Hungenberg, M. Kersting, Polym. Mater.' Sci. Eng., 10. K.D. Hungenberg, DECHEMA Monogr.,

a,556 (1992)

m,501 (1992)

11. H.J. Zimmermann. J. Macromol. Sd.-Phys. 12. H. H. Brintzinger, private communication

w, 141 (1993)

38 1

36. Feature of Metallocene-Catalyzed Polyolefins

N. KASHIWA Polymers Laboratories, Mitsui Petrochemical Industries Ltd. Waki, Yamaguchi 740, Japan INTRODUCTION Since the catalyst system of b i s c y c l o p e n t a d i e n y l z i r c o n i u m dichloride (CpeZrC12) and methylalumoxane (MAO) produces polyethylene having narrow molecular weight distribution with high activity was found by Sinn and Kaminsky”, metallocene catalyst system is in the limelight as a new type of catalyst system. Recently, the plans for the commercialization of metallocene catalyzed polyolefins are often announced and some of them has been already on market. For example, EXXON Chemical Co. marketed linear low density polyethylene (LLDPE) made by high pressure ion process, and DOW Chemical Co. also did LLDPE and elastmer made by solution process. Furthermore, syndiotactic polypropylene and syndiotactic polystyrene are now in the stage of market sounding. Looking back the historical current of highly active metallocene catalyst ‘system (Fig. 11, it started from finding the combination of Cp2ZrC12 and MA0 by IS80 1988 1990

Brintzinger et al.?’ found isospecific catalyst such as racethylenebisindenyl zirconium dichloride (rac-Et(Ind)zZrC12).

-

I

I

Sinn and Kaminsky on 1980 as described above. After five years, Kaminsky and

1

C.L

.b *?

&

n’

I

/ / \ d Q

&
J&

+

4 3% t

CoaL

MA0

MA0

-

Stemolpeclndly

It.

ho

:m

I18

Year

1980

1985

1988

1989

MA0

Fig. 1 Hlstory of metallocene catalysts

382

On

N . Kashiwa 1988

Ewen3)

succeeded

to

synthesize

the

syndiotactic

polypropylene using isopropylidene(cyclopentadienyl-9fluoreny1)zirconium dichloride. Further, MA0 free catalyst systemJ) by employing boron compound such as triphenylcarbenium tetrakis (pentafluorophenyllboron, [CPh3][B(CBFa)J], instead of MA0 was found. After that, by modifying the ligand of metallocene etc., the catalyst performance i s further improved and the polymer with unique characteristics appears. Today, by the metallocene catalyst technology, polyolefin industry i s on the transition stage of the breakthrough. In this paper, some aspects for metallocene catalyzed isotactic polypropylene and polyethylene are discussed. Metallocene-based Polypropylene The typical example of the first generation metallocene catalyst with isospecificity is rac-Et(Ind)zZrCl,. The polypropylene produced at 5 0 ' C by this catalyst has low melting point (ca. 130'C by DSC), low isotacticity ([mm]=ca. 85% by "C-NMR) and low molecular weight (below Mw=50,000) in comparison with commercialized polypropylene. Fig. 2 shows the relationship between the melting point and meso triad [mml. points of the

The melting polypropylene

._ 0

150

produced by rac-Et (Ind)2ZrC12 a p 140.were lower than that by heterogeneous Ti catalyst at the same 130I I [mml. A s a plausible reason 100 95 90 a5 for the lower melting point, i t lsotacticity mm(%) is considered that the regioFig. 2 Dependence of melting point on isotactlcity irregular units based on 2 , l ( ) content of rsglolrrsgular lnsertlon and 1.3-insertion of propylene S o , it is very important for monomer exist in polymer chain'". the production of polypropylene with high melting point t o enhance not only isospecificity but regiospecificity of metallocene catalyst. If i t i s possible t o produce the non-regioirregular polypropylene by metallocene catalyst, it i s interested in what degree of improvement in melting point. It seems that the symmetricity, the steric bulkiness, and the position of substituent groups on cyclopentadienyl o r c

36. Feature of Metallocene-Catalyzed Polyolefins

indenyl

ligand

are

key

factors

for

controlling

383

stereo- and

regiospecificity, molecular weight of polymer and activity. Recently, Koga et al.B) reported to be able to estimate the stereo- and regiospecificity of metallocene catalyst by ab initio method. In fact, Spalek et a1.7’ showed that dimethylsilylbis(2-methyl-4-naphtylindenyl)zirconium dichloride with bulky ligand can produce polypropylene with melting point and molecular weight comparable to commercialized polypropylene. There i s flexual modulus as one of the important factors in the physical properties of polypropylene. The flexual modulus of metallocene catalyzed polypropylene is reported to be higher than that of heterogeneous Ti catalyzed one at the same melting point (Fig. 3 ) , it i s interested in whether the advantage of modulus i s keeped or not in metallocene catalyzed polypropylene with the Fig. 3 Flexural modulus vs. melting point high melting point Ref. A . Montagua & J.C. F l o y d , YetCon‘93. P171(1993) like the polypropylene by Spalek.

Metallocene-based Polyethylene In this section, first, LLDPE produced by gas phase polymerization in our pilot plant are introduced. Fig. 4 shows the composition distribution (CD) of metallocene catalyzed LLDPE (ca. 0 . 9 0 g/cm”) measured by G P C - I R . Metallocene catalyzed LLDPE has almost same comonomer content in each fraction independent of molecular weight. This significant narrow CD appears on the physical property of polymer as shown below. The dependence of polymer density on comonomer contents i s different between heterogeneous Ti and metallocene catalyzed LLDPE (Fig. 5 ) . The heterogeneous Ti catalyzed LLDPE needs more comonomers than metallocene catalyzed one in the same

384 N . Kashiwa

Yetallocene catalyzed

Heterogeneous T i

f

V

:L 0

0 1o2

1o4

1o8

10' Mw

F i g . 4 D i s t r i b u t i o n o f s h o r t c h a i n branching(GPC-IR a n l y s i s )

density .

Furthermore,

metallocene LLDPE lowers

catalyzed with the

of density, while of heterogeneous Ti catalyzed LLDPE hardly lowers (Fig. 6). Regarding the physical properties of film, in drop that

0.940.

,

,

'

'

"

"$

0'930-

\ 0)

.Z 0.920

g

-

al

n

0.910

-

0.900

film

,

I

I

'

'

1

I

'

(30 ,u m ) on density.

u

v1 c

v

c

.-cc

100

-

0 0

60 0.88 0 910

0 920

0 930

I

I

I

0.89

0.90

0.91

0.92

0,940

D e n s i t y (g/cn')

F i g . 7 Impact s t r e n g t h vs d e n s i t y

D e n s i t y (g/cn3)

F i g . 6 M e l t i n g p o i n t vs. d e n s i t y

0.93

36. Feature of Metallocene-Catalyzed Polyolefins

The

dart impact

385

strength of

metallocene based LLDPE i s significantly high and that degree i s remarkable at lower density, i.e. about three heterogeneous Ti times of catalyzed LLDPE at the density of around 0.923 g/cm3. The heat seal property and the clarity of metallocene catalyzed LLDPE are also superior to heterogeneous Ti catalyzed Rat LLDPE (Figs. 8, 9 ) . From the above results, a crystal structure o f metallocene catalyzed LLDPE i s characterized as illustrated in Fig. 10. It i s thin lamella thickness and many tie molecules with more uniform compared with heterogeneous T i catalyzed LLDPE. For many applications, the good processability of polymer i.e. high flowability and high melt tension is necessary.

Heterogeneous T i c a t a l y z e d LLDPE

Thick lamel l a (High Tm) Fen t e i m o l e c u l e s ( l o w i m p a c t )

120 ,-. 0

-

v

,

100

n 0

c E

.c

m

.-c

2 a

v)

-

-

MI-1.7

-

d-0.907

-

-

MI-1.0 d-0.904

-

80 -

60 -

40 -

c

-

I

--

-

-

g 200

Business

:-nl Broad CD

Narrow CD

10.5

-

?3

3.5

Metal locsna Catalyzed

mat.rogansous Ti Calalyrid

LLOPES(EXXPOL~")

(Osnsity-0.907)

LLOPE~

(Density-0.904)

F i g . 9 C l a r i t y of

Blown F i l m

R e f . TschnIcal brochure o f EXXON Chemical Company * A New Family Ot Linear Ethylene Poly.ers*

U e t a l l o c e n e c a t a l y z e d LLDPE

Thin lamella (low Tm) Many t e i m o l e c u l e s ( h i g h i m p a c t )

F i g . 10 P l a u s i b l e model o f c r y s t a l s t r u c t u r e s

386 N. Kashiwa

But it i s expected that the narrow molecular weight distribution

(MWD) provides bad processability. Figs. 11 and 12 shows the flowability and the melt tension of various LLDPE at the molten state, respectively. For the #1 type of LLDPE, the dependence of the melt viscosity on shear rate is small and the melt tension is also low as expected. However, we found that the selection of metallocene catalyst lead to improve processability of LLDPE. Namely, the # 3 type has a significantly lower melt viscosity at the processing shear rate region (high shear) and In addition, the very high melt tension like HPLDPE. intermediate type between the #1 and #3 types is also obtained.

I

I

1

lllll.l

1 oo

lo-'

,

,

, ,

I

,

, , ,I

, , ,,,,,

1 o3

1 o2

10' (sec-1)

.i. F i g . 1 1 Melt viscosity v s . shear rate

+

10-1' 10.'

'

' " 1 ' 1 4 1

1 oo

'

'

10'

'

"

,

I

MI (g/lOmin) F i g . 12 Melt t e n s i o n v s . m e l t index

,

J

1o2

36. Feature of Metallocene-Catalyzed Polyolefins 387

Several

yeas

ago,

we

reported

that

ethylene/octene-1

copolymer with a unique physical property could be obtained by the catalyst system of rac-ethylenebisindenylhafnium dichloride and MAOJ’ by solution polymerization. This copolymer showed high flowability in spite of a narrow MWD polymer as shown in Table 1. Herein, 110 and I 2 are the values of the melt index Table 1. Relationship between flowability and molecular weight distribution

__

Catalyst

Mw/Mn

12.7

2.6 2.5

metallocene #1

5.6

2.4

homogeneous V

5.9

2.2

heterogeneous Ti

7.6

6.0

13.1

~~

measured at 1 9 0 ~C under 10 kg-loading and 2 . 1 6 kg-loading, respectively, and the I l o / I r ratio is an index of polymer flowability in molten state. Generally, in previously known copolymer, the Mw/Mn i s directly proportional to the I l u / I l ratio and the polymer with Mw/Mn=2-3 have 1 1 0 / 1 3in the range of 5-7. The polymer produced by Hf catalyst, however, had for example, I1,/IZ=13.1 at Mw/Mn = 2.6. In future, metallocene catalyzed LLDPE with excellent properties such as high impact strength, high clarity, good heat sealability, anti-blocking and good processability will employ in a broad range of application as a new polyethylene. Out look Today in fifteen years later since Kaminsky discovery, metallocene catalyzed polyolefin technologys have reached in the level of commercialization although some issues such as the production cost and the molecular weight of polymer are still remained. Further, metallocene catalyst have possibility to produce the polymer with new physical properties as unexpected from the previously known polymer and will build up a new polyolefin world.

388 N. Kashiwa

References 1)H. Sinn, W. Kaminsky, Adv. Organomet. Chem., B , 99(1980) 2)W. Kaminsky, K. Kulper, H. H. Brintzinger, F. W. P. Wild, Angew. Chem., 97, 507(1985) 315. A. Ewen, R. L. Jones, A. Razavi, J. D. Ferrara, J. Am. Chem. SOC., 110, 6225(1988) 4a)R. F. Jordan et al., J. Am. Chem. SOC.,108,1718(1986) 4b)EP-A-0277004 5a)H. N . Cheng. J. A. Ewen, Makromol. Chem., 190, 1931(1989) 5b)K. Soga, T. Shiono, S . Takemura, W. Kaminsky, Makromol. Chem. Rapid Commun., 8 , 305(1987) 6)H. K. Kuribayashi, N . Koga. K. Morokuma, J. Am. Chem., 114, 8687(1992) 7)W. Spalek presented in "International Symposium 40 Years Ziegler Catalysts" Freiburg, Sept.'93 8)JP-A-276807/1990

389

37. Ligand Effects at Transition Metal Centers for Olefin Polymerization

J. Kard and Sun-Chueh Kao Union Carbide Corporation, P.O. Box 670, Bound Brook, New Jersey 08805, USA

ABSTRACT Ethylene polymerization catalysts based on titanium, vanadium, chromium, and zirconium display dramatic changes when the ligand environment at the active site is altered. Catalysts with a ligand environment involving chloride, alkoxide, electrondonor compounds, alkyl, cyclopentadienyl, or indenyl groups show distinctive polymerization characteristics. These characteristics include polymer molecular weight and molecular weight distribution. Catalyst response parameters such as comonomer incorporation, hydrogen response, and catalytic activity are influenced by ligand environment at the active site. Changes in the level of unsaturation at active metal centers can also have a profound effect on polymerization kinetics and the nature of the final reaction products. Alpha-olefins play a subtle role in altering the structure of the active centers. Copolymerization studies provide a powerful tool for probing the steric and electronic features of olefin polymerization catalysts.

l " In the development of polymerization catalysts much attention has focused on the capabilities and limitations of various catalyst families for ethylene and propylene polymerization. It became apparent from polymerization studies that each catalyst family (Ti, V, Cr, Zr) displayed its own unique personality in the polymerization process.' Continuous attention over the years has aimed at modifying or altering catalyst behavior in order to make the catalyst more flexible in the polymerization process and to allow the catalyst to provide the widest range of product opportunities. Changes in the ligand environment at the active polymerization site represent a profound method for controlling and changing polymerization performance over a wide range (Table I). The importance of ligand modification to control the performance of olefin polymerization catalysts will make the 1990s the "decade of the ligand." While process conditions and reactor types do play a role in regulating polymer properties, the chemical nature of the active site is a major determinant of the product characteristics and the process parameters needed to produce the product. In this paper we illustrate the effect certain ligand modifications have exerted in the polymerization of ethylene using high-activity titanium, vanadium, chromium, and zirconium catalysts. In addition, we attempt to provide a perspective on different techniques of ligand modification with olefin polymerization catalysts.*

390 F. J . Karol and S.C. Kao

Table I. Ligand Effects on Olefin Polymerization Behavior Catalytic Activity/Klnetics Polymer Molecular Weight Polymer Molecular Weight Distribution Comonomer Incorporation and Compositional Distribution Stereoregulation

DlSCUSSlON Ppproaches to l&,and Modification of .-P lo Several approaches to ligand modification of polymerization centers offer significant opportunities to change or alter polymerization behavior (Table 11). In some cases the modification occurs prior to use of the catalyst in the polymerization process. In other cases the change occurs while the polymerization reaction is in progress. Table II. Routes to Ligand Modification of Polymerization Centers Directly Synthesized Complex Reversible/Non-Reversible Lewis Bases Ligand Exchange Reactions a-Olefins as Ligands Unsaturated Metal Centers by Ligand Abstraction Directly synthesized organometallic compounds such as chromocene, (C5H5)2CrI or zirconocene dichloride (C5H5)2ZrC12, lead to fixed cyclopentadienyl ligands at the metal center. The presence of the cyclopentadienyl ligand is manifested in the polymerization features of the catalyst.3-5 Reversible or non-reversible Lewis bases can be added prior to or during the polymerization reaction.6 Typical types of Lewis bases include ethers, esters, ketones, amines, siloxanes, and phosphine oxides. Many electron-donor compounds (E) are believed to interact with both the active site (M) and the organoaluminum cocatalyst,

I -M-

I

I +E

-M*E

I

(11

R A t E -R#E (2) Reactions illustrated by eqs 1 and 2 may be viewed as reversible or non-reversible depending on the specific compounds involved. It appears there exists a rapid ligand exchange between zirconocene dichloride and methylalumoxane via numerous

37. Ligand Effects at Transition Metal Centers 391

combined equilibria. When Lewis acids or bases are added to these catalysts, it is possible to influence these dynamic equilibria.’ Ligand exchange reactions serve to alter the ligand environment at the transition metal center in order to regulate or improve the polymerization performance of the catalyst. Reactions of magnesium compounds with titanium halides have been widely used as a route to boost the catalytic activity of titanium-based catalysts.8~9 Addition of halocarbons to vanadium-based catalysts have provided a means to regulate the chloride environment at the vanadium active site.1110 Kinetic data from ethylene, a-olefin copolymerization studies show that the presence of the a-olefin can lead to an initial rate enhancement.lll12 Data with vanadiumbased catalysts suggest that the a-olefin can act as a coordinating ligand during the polymerization process1113 (eq. 3). CX

0 I

- v -0+ CX

I

P

I

p

- v -0

d

P = growing polymer chain C, = a-olefin

(3)

polymerization Coordinatively unsaturated transition metal-based compounds are believed to be the site for olefin polymerization and oligomerization reactions. Highly unsaturated chromium,l4 vanadiurn,ls and even palladium catalysts15 have been described for olefin polymerization or copolymerization. The extent of unsaturation and geometry of such catalytic complexes play an important role in determining the reactivity and selectivity of such complexes.

ChromiumCatalvsis. Comprehensive studies have evaluated the effect of directly synthesized n-bonded organic iigands attached to chromium on polymerization parameters of several supported chromium-based catalysts.3116l17 In these studies differences in hydrogen response, comonomer incorporation, and polymerization activity with these catalysts suggested that the nature of the active site was different due, at least in part, to changes in ligand environment. In these supported catalysts with n-bonded ligands, it appears that in many cases at least one of these ligands (L) remains coordinated to the active chromium centers during the polymerization processla (eq 4).

392

F. J. Karol and S.C. Kao

- 0 - R + nCH2 = CH2

kP

* $r+

CH2-CH, t n R

(4)

Silica-supported bis(cyclopentadieny1)-, bls(indeny1)-, and bis(fluoreny1)-chromium catalysts show good activity In ethylene polymerization.16 The capability of cyclopentadienyl chromium-based active centers to produce lower molecular weight polyethylenes in the presence of hydrogen was observed to decrease for the following series of ligands: cyclopentadienyl > indenyl > fluorenyl L 9-methylfluorenyl. This decrease in hydrogen response is probably related to the decreased electron density at the chromium center in the ligand series above. The CrOdSiO2 catalyst has very little, or any, response to hydrogen as a chain transfer agent. As a result of the high hydrogen response with the chromocene catalyst, a highly saturated polyethylene is produced. Polyethylenes produced with the CrOslSi02 catalyst usually have one double bond per molecule, indicating a different chain transfer process from the chromocene catalyst.19 Thermal aging of the chromocene catalyst led to removal of the cyclopentadienyl ligand and loss of the high hydrogen response of the catalyst.18 Polyethylenes produced with the chromocene catalyst are considered relatively narrow in molecular weight distribution. Addition to the catalyst of ethers,no ammonia,21 or siloxanes22 prior to the polymerization led to modified catalysts which produced polymers with a more narrow molecular weight distribution. Certain chromium-containing catalysts provide examples of unsaturated metal centers generated by ligand abstraction.14 A homogeneous ethylene polymerization catalyst, chromium(1ll) 2-ethylhexanoate and hydrolyzed triisobutylaluminum (PIBAO) can produce small quantities of 1-hexene as well as polyethylene. Addition of dimethoxyethane to the catalyst solution led to a significant Increase in 1-hexene selectivity to 74%.23 The principal byproduct was polyethylene, although small amounts of butenes and octenes were also produced. The modified catalyst had a rate of 1.2 mol/mol Crosec for 1-hexene generation (eq 5).

37. Ligand Effects at Transition Metal Centers

393

One view of the origin of the unsaturated chromium centers relates to the conversion of chelating carboxylate ligands surrounding the chromium center to oxide ligands.14 The net effect of such an interaction In the presence of PIBAO is to convert the chromium center into three-coordinate structures (eq 6) which satisfy the requirements of the proposed mechanism of 1-hexene formation.

The rate of 1-hexene formation was dependent on the square of ethylene pressure. Addition of dienes at very low levels resulted in the inhibition of l-hexene formation. These observations can be understood if two ethylene molecules are coordinated with the active chromium site in the activated complex involved in the rate-determining step.

alalysk. Bimetallic complexes containing magnesium, titanium, and electron-donor molecules when combined with aluminum alkyls show high catalytic activity in ethylene polymerization.24-27 The ligand exchange reaction between MgC12(THF)2 and TiC14(THF)2 in tetrahydrofuran yields a yellow crystalline salt [ M Q ~ C I ~ ( T H F ) ~ ] + ~ ~ C I ~The ( T Hcrystal F ) ] - . structure of this salt has been defined in our laboratories and by another group.28 The presence of MgC12 in titanium-based catalysts serves to increase the number of active centers for polymerization. The THF acts as a solvent for the reactants and participates in complex formation rendering the complex stable and permitting the exchange reaction to occur. Removal of THF from

394

F. J. Karol and S.C. Koo

the complex by organoaluminum compounds provides a route for introducing coordinative unsaturation at the titanium center. Generation of 1-butene from ethylene with high selectivity is well-known technology.29-30 One titanium catalyst is based on directly synthesized titanium(lV) alkoxides and alkylaluminum cocatalysts (eq 7). C2H4

R,AI + Ti(OR)4

R,AI

+ TiCI4

C2H4

high Selectivity to 1-butene

(7)

PE

Because of the fixed alkoxide and chloride ligand environments at the two titanium centers, it was possible to simultaneously dimerize ethylene to 1-butene and copolymerize (eq 8) the resultant 1-butene with ethylene.31-33 The catalyst systems are compatible with each other and operate under the same reaction conditions of temperature, monomer pressure, and solvent. Recent investigations with certain polypropylene catalysts illustrate how ligand abstraction from titanium-based catalysts can lead to polymerization centers with two vacant sites.34 The catalyst was prepared using TiCI3*3Pyridine/MgCln in the presence of A12(C2H5)3C13. The catalyst combined with (C2H5)aAI selectively gave atactic polypropylene.

a,,

01, C2H5#,0 +E

-

L


C2H5

i!)

,E

(9)

iC !; C The aspecific catalyst could be easily converted into an isospecific one by using an electron donor such as ethylbenzoate (eq 9).35 The reaction kinetics and kinetic order with several high-activity catalysts (Ti, V) provide insight into the extent of unsaturation at the active sites.' 3 Homopolymerization with a TiC13(THF)3/MgC12/Et2AlCl/Et3AI catalyst showed a second-order dependency of activity on ethylene partial pressure. Similar results were observed using a VC13(THF)s-based catalyst for ethylene polymerization. These results using transition metal halide-electron-donor complexes suggest that ligand-abstraction reactions to remove the electron donor can provide metal centers with more than one site of unsaturation. Such a result is not easily obtained when using a TiCl3-based catalyst with no added electron-donor.

37. Ligand Effects at Transition Metal Centers 395

Since the early 1960s certain amido complexes of titanium (IV) have been examined as catalysts with alkylaluminum compounds with no unusual results.37-39 Recently cyclopentadienyl amido complexes, particularly of titanium and zirconium have received worldwide attention.40-42 These complexes in a configuration of constrained geometry can be activated by methylalumoxane or cationic activators to provide highly active ethylene polymerization catalysts. Many of these catalysts show outstanding copolymerization behavior. The constrained geometry provides a means to reach an unsaturated, highly exposed metal center for polymerization. Certain amido systems have been described which do not require a cyclopentadlenyl ligand.43 The concept of constrained geometry deals with the approach of generating an active catalytic center that has an open structure with considerable space for incoming molecules to react with minimal steric hindrance. One Important result from the use of such catalytic complexes is the ability of long chain and polymeric a-olefins to incorporate into the chain and thereby provide a route to long-chain branching in polyethylene molecules.44 m. C The oxidation state of active vanadium catalysts for ethylene polymerization has received attention by different investigators.lllo14514s Some investigatorsl0145 have suggested that V(I11) Is the oxidation state of the catalyst and deactivation occurs by reduction to V(II). Other studiesi146 lend support to the proposal that V(II) represents the important oxidation state in ethylene polymerization. Regardless of the actual oxidation state of the active vanadium center, it is clear that a halide ligand is present at the active center of many vanadium catalysts. The family of organohalides represents the primary activity promoters in vanadium catalysis. With some vanadium catalysts, the presence of these promoters appears to regulate the extent of alkylation or reduction of the polymerization centers. With a supported VC13(THF)3/Si02-based catalyst, addition of organohalides (R'X) in the presence of trialkylaluminum cocatalysts leads to a divalent alkyl vanadium halide (eq 10) in a ligand exchange reaction. VC13(THF)3/Si02 + R3AI + R'X + [RVX] (10) Because of the highly unsaturated nature of this vanadium compound, it is most likely complexed to aluminum compounds in the reaction medium (eq 11). The strength of these complexes with aluminum compounds will no doubt depend upon the Lewis acidity of the vanadium center. The V(III) complexes are probably stronger than comparable V(II) ones.

396

F. J . Karol and S.C. Kao

Addition of reversible Lewis bases such as electron-donor compounds to olefin polymerization catalysts is a widely used method to control the behavior of these catalysts. Among the electron-donor compounds studied were oxygen-containing compounds such as ether-alcohols, amino-alcohols, and ether-esters.47 Other systems included phosphine oxides and phosphates,48 silanes,49 and aluminum alkoxides.50 All these electron-donor compounds were able to modify the vanadium active centers and produce polyethylenes with a more narrow molecular weight distribution. Studies of ethylene copolymerization using a VC13(THF)s/Si02 + Et2AICI + Et3AI catalyst revealed that the structure of the comonomer had a significant effect on the increase in polymerization actlvityl3. This observation was interpreted in terms of a mechanistic model Involving two vacant coordination sites at the active centers (eq 3). Like the homogeneous chromium catalysts description earlier, the unsaturated vanadium centers are believed to be generated by ligand abstraction reactions. Electron donors and comonomers are viewed as Lewis-base ligands which can coordinate at the sites of unsaturation and influence chain initiation, propagation, and termination. Addition of stoichiometric quantities of 1,2-dimethoxybenzene to the vanadium-based catalyst led to almost total deactivation of the catalyst13 (eq 12). Such a stolchiometric reaction lends support for the unsaturated nature of the vanadium center depicted in eq 12.

0

I

(12)

P

OCH3

P = growing polymer chain

ZirconiumCatalvsis. A directly synthesized ligand environment has a major impact on the polymerization features of zirconocene-based catalyst~4~5~51~52. Ligands that contain electron-donating groups were able to increase catalytic activity for ethylene polymerlzation carried out in toluene53154or hexane diluent. Steric effects became important when bulky groups or multiple groups were used. Zirconocene catalysts with higher degrees of structural rigidity normally produced ethylene polymers of higher molecular weights. All zirconocene catalysts, regardless of ligand configuration, produced ethylene polymers with a narrow molecular weight distribution. In our studies, zirconocene-based catalysts displayed a wide range of responses to a-olefins such as l-hexene. The ethylene reactivity ratio varies by a factor of 45 with the metallocenes studied. Other workers have reported a variation of

37. Ligand Effects at Transition Metal Centers 397

nearly 200 In the reactivity ratio for ethylene-propylene copolymerizations using different metallocene catalysts.53 The data in Table Ill show the effect of electron-donatinggroups and the indenyl ligand on polymerization activity and 1-hexene incorporation when simple, unbridged metallocenes are used with methylalumoxane. For comparison purposes, the data shown are given on a relative activity basis. The actual activity of the zirconocene catalyst run at 85°C in n-hexane as a diluent was 39.4 kg PE/mmol Zr* hr* 100 psi C2H4. Studies on the effect of ligand structure with the zirconocene catalysts showed that indenyl and cyclopentadienyl ligands showed similar copolymerization behavior (Table IV, V). The indenyl compound was about 15% less active than the parent zirconocene dichloride system. Single bridging atoms (Si, C) between the cyclopentadienyl ligands improved comonomer incorporation, although polymerization activity in some cases was lower than the parent zirconocene dichloride. Zirconocene catalysts (Table Vl) containing a 02C< or (CH3)2C< bridging group and a mixed cyclopentadienyl/fluorenylligand environment showed outstanding 1-hexene incorporation42. For zirconocene catalysts containing indenyl ligands, the presence of a -CH2CH2- or (CH3)2Si< bridging group led to higher comonomer incorporation with equal or better catalytic activity than the unbridged indenyl catalyst (Table VII). Substitution on the ring system in the 2-position led to lower activity and lower 1-hexene incorporation. The unsaturated metallocenium ion arises by formal transfer of CH3- from zirconium to acidic, three coordinate aluminum sites (eq 13). CH3

I

Cp,Zr(CHJ,

t

CAI0

+-

-

(Cp2ZrCHJ' [(CH&AIO]-

(13)

Incoming monomer must compete with and displace the counterion in a contact ionpair or solvent separated ion-pair.55 Other methods for generating unsaturated cationic metallocenium ions include one-electron oxidation, protonation, cation ligand abstraction, or Lewis acid ligand abstraction from cyclopentadienyl zirconium dimethyl compounds.56 The overall impact of the environment at the active center in these catalysts is impressive. The results indicate that considerable control over the behavior of these catalysts can be realized by selecting the appropriate ligand system.

398

F. J . Karol and S.C. Kao

C

o

p

o

l

y

i

~ as s a Probe for l i w n d Environment.

Copolymerization studies within a family of catalysts can serve as a valuable probe for assessing the nature of the active site for olefin polymerization57158. With the (C5H&Cr/S102 catalyst, the effective ethylene reactivity ratio is 72 for an ethylenepropylene copolymerization.3 This high value was attributed to the presence of a cyclopentadienyl ligand at the active chromlum center. The ethylene reactivity ratio (Ti = 48) for ethylene-propylene copolymerization using a Cp2ZrC12/MAO catalyst59 is also a high number. This result can be attributed to the presence of cyclopentadienyl ligands and to the use of zirconium as the active metal center. In one study the ethylene reactivity ratio increased in the series Ti e Zr < Hf.58 The data in Table Vlll compares the results of the current study with investigations carried out by other workers.53 The relative reactivity ratios for ethylene-propylene copolymerization paralleled, in general, the results for ethylene-I -hexene copolymerizations. Ansa-metallocenes with one or two atoms as bridging groups were more effective for copolymerization than the unbridged metallocenes. Previous work with highly active Ziegler-Natta catalysts for copolymerization revealed that reactivity ratio values were predominately determined by entropic factors.60 It is not known at this time whether the reactivity ratio values for metallocene catalysts are similarly determined. One might expect that entropic factors would not exert as much of an influence with metallocene active sites having a more exposed environment. Recent analysis of steric effects in ansa-metallocene-basedcatalysts has focused on "coordination gap aperture" and "obliquity angles" as parameters for structurereactivity relationships.61 The parameters provide, respectively, a measure of the openness of the metallocene center and a measure of those structural elements which are most likely to control stereoselectivity of ansa-metallocene catalysts.

REFERENCES 1. F. J. Karol, K. J. Cann and B. E. Wagner in Transition Metals and Organometallics as Catalysts for Olefin Polymerlzation, W. Kaminsky and H. Sinn, Eds., Springer-Verlag, New York 1988, pp 149-161. 2. F. J. Karol, S. C. Kao, Paper Presented at 12th Summer School on Coordinatlon Chemistry, Karpacz, Poland, June 6, 1993. 3. F. J. Karol, G. L. Karapinka, C. Wu, A. W. Dow, R. N. Johnson and W. L. Carrick, J. Polym. Sci., A-1, 2621-2637 (1972). 4. H. Sinn, W. Kaminsky, H-J. Volmer and R. Woldt, Angew. Chem. lnt. Ed. Engl., U,390 (1980). 5. H. Sinn and W. Kaminsky, Adv. Organomet. Chem., U, 99 (1980). 6. J. Boor, Ziegler-Nafta Catalysts and Polymerizations, Academic, New York, 1979, pp 213-239.

37. Ligand Effects at Transition Metal Centers 399

7. D. Fischer, S. Jungling and R. Mulhaupt, Makromol. Chem., Macromol. Symp., 191-202 (1993). 8. P. Pino and R. Mulhaupt, Angew. Chem. lnt. Ed. Engl., -19, 857-875 (1980). 9. P. Galli, L. Luciani, G. Cecchin, Angew. Makromol. Chem., 94,63-89 (1981). 10. G. G. Evens, E. M. J. Pijpers and R. H. M. Seevens in Transition Metal Catalyzed Polymerizations, R. P. Quirk, Ed., Cambridge Univ., New York, 1988, pp 782798. 11. P. J. T. Tait, see reference 10, pp 834-860. 12. D. C. Calabro and F. Y. Lo, lbid., pp 729-739. 13. F. J. Karol, S-C. Kao and K. J. Cann, J. Polym. Sci., Polym. Chem. Ed., 254 1-2553 (1993). 14. R. M. Manyik, W. E. Walker and 1.P. Wilson, J. Catal., &Z, 197-209 (1977). 15. E. Drent, J. A. M. van Broekhoven, M. J. Doyle, J. Organomet. Chem., fl, 235251 (1991). 16. F. J. Karol, W. L. Munn, G. L. Goeke, B. E. Wagner and N. J. Maraschln, J. Polym. Sci., Polym. Chem. Ed., M,771-778 (1 978). 17. F. J. Karol and R. N. Johnson, J. Polym. Sci., Polym. Chem. Ed., U,1607-1617 (1975). 18. F. J. Karol and C. Wu, J. Polym. Sci., A - l , U , 1549-1558 (1974). 19. F. J. Karol, G. L. Brown and J. M. Davison, J. Polym. Sci., Polym Chem. Ed., jl, 413-424 (1973). 20. G. L. Brown, W. C. Cummings and I.J. Levine, U.S. Patent 4,115,639 (1978); F. J. Karol and C. Wu, U.S. Patent 4,086,408 (1978). 21. F. J. Karol, US. Patent 3,813,381 (1974). 22. F. J. Karol and C. Wu, U.S. Patent 4,086,409 (1978). 23. J. R. Briggs, J. Chem. SOC.,Chem. Commun.,U, 674-675 (1989). 24. U. Giannini, E. Albizatti, S. Parodi and F. Pirinoli, U. S. Patents 4,124,532 (1978) and 4,174,429 (1979). 25. K. Yamaguchi, N. Kanoh, T. Tanaka, N. Enokido, A. Murakami and S. Yoshida, U.S. Patent 3,989,881 (1976). 26. A. Greco, G. Bertolini and S. Cesca, J. Appl. Polym. Sci., 2045-2061 (1981). 27. F. J. Karol, G. L. Goeke, B. E. Wagner, W. A. Fraser, R. J. Jorgensen and N. Friis, U.S. Patent 4,302,566 (1981). 28. P. Sobota, J. Utko and T. Lis, J. Chem. SOC.,Dalton Trans., 2077-2079 (1984). 29. K. Ziegler and H. Martin, US. Patent 2,943,125 (1960). 30. C. E. H. Bawn and R. Symcox, J. Polym. Sci., 139 (1959). 31. D. L. Beach and Y. V. Kissin, J. Polym. Sci., Polym. Chem. Ed., 22, 3027-3042 (1 984). 32. Y. V. Kissin and D. L. Beach, ibid., 24, 1069-1084 (1986). 33. K. J. Cann, M. W. Chen and F. J. Karol, US. Patent 4,861,846 (1989).

a,

a,

a,

400

F. J . Karol and S.C. Kao

a,

34. T. Shiono, H. Uchino and K. Soga, Polym. Bull., 19-21 (1989). 35. K. Soga and J. R. Park in Catalytic Olefin Polymerization, T. Keii and K. Soga, Eds., Elsevier, New York, 131-138 (1 990). 36. British Patent 969,074 (1964). 37. J. G. Hefner, U.S. Patents 4,892,914 (1990) and 4,956,323 (1990). 38. T. Sasakl, et al., EPA 0349886 (Priority Date 6/28/88) and EPA 0509233 (Priority Date 3/15/91). 39. J. G. Hefner, et al., EPA 0476671 (Priority Date 9/20/90). 40. J. C. Stevens and D. R. Neithamer, EPA 0418044 (Priority Date 9/14/89); also U.S. Patent 5,064,802, (1991) and US. Patent 5,132,380 (1992). 41. J. C. Stevens, et al., EPA 0416815 (Priority Date 8/31/89). 42. J. A. M. Canlch, U.S. Patents 5,026,798 (1991); 5,055,438 (1991) and 5,096,867 (1992). 43. J. A. M. Canich, WO 92/12162 (Priority Date 12/27/90). 44. S. Y. Lai, et al., U.S. Patent 4,272,236 (1993). 45. D. L. Christman, J. Polym. Sci., A-l,j.Q, 471-487 (1972). 46. P. D. Smith, J. L. Martin, J. C. Huffrnan, R. L. Bansemer and K. G.Caulton, lnorg. Chem , 2997-3002 (1985). 47. S. C. Kao, K. J. Cann, F. J. Karol, A. E. Marcinkowsky, M. G. Goode and E. H. Theobald, US. Patent 4,923,938 (1990). 48. S. C. Kao and F.J. Karol, US. Patent 4,886,771 (1989), and 4,988,784 (1991). 49. S. C. Kao and F.J. Karol, U.S. Patent 4,999,327 (1991) 50. S. C. Kao and F.J. Karol, US. Patent 5,030,605 (1991). 51. J. A. Ewen, J. Am. Chem. SOC.,.1_Q6,6355(1984). 52. J. A. Ewen, R. L. Jones, A. Razavi and J. D. Ferrara, J. Am. Chem. SOC.,U, 6255 (1988). 53. A. Zambelli, A. Grassi, M. Galimberti, R. Mazzocchi and F. Piomentesi, Makromol. Chem., Rapid Commun.,j.2, 523-528 (1991). 54. P. J. T. Tait, private communication. 55. A. R. Siedle, W. M. Lamanna and R. A. Newmark, Makromol.Chem., Makromol. Symp., 215-224 (1993). 56. J. A. Ewen and M. J. Eider, EPA 0426638 (Priority Date 10/30/89) and

a,

w,

a,

references therein.

a,

57. G. Natta, F. Danusso and D. Sianesi, Makromol. Chem., 238 (1959). 58. F. J. Karol and W. L. Carrick, J. Am. Chem. SOC.,MI2654-2658 (1961). 59. J. A. Ewen in Catalytic Polymerization of Olefins, T. Keii and K. Soga, Eds., Elsevier, New York 1986, p 284. 60. L. L. Boehm, J. App. Polym. Sci., 29, 279-289 (1984). 61. K. Hortmann and H. H. Brintzinger, New Journal of Chemistry, 51-55 (1992).

a,

37. Ligand Effects at Transition Metal Centers 401

Table 111. Effect of Cyclopentadienyl Ring Substituent On Copolymerlzation Rate and 1-Hexene Incorporation Relative Behavior 1H-CH3

2.08

1.39

-Bu

1.52

0.84

-H

1.oo

1.oo

-(CH3)5

0.50

0.14

lndenyl

0.86

0.97

Table IV. Effect of Metallocene Structure On Polymerization Activity and 1-Hexene Incorporation

d

1-He-

Electron Donating Groups

Cp

t

t

Steric Effects

CP

3.

3.

lndenyl vs Cp

--

A

0

Single Atom (Si,C) Bridging Group

CP

(3.1

t

Single Atom (Si) Bridging Group

Ind

0

t

Single Atom (C) Bridging Group

Cp/Flu

3.

t

402 F. J . Karol and S.C. Kao

Table V. Effect of Bridging Group in Ansa-Metailocene On Copolymerization Rate and 1-Hexene incorporation

Activity

D

None 0 42c\

Relative Behavior 1H --

1.03

1.97

1.00

1-00

0.60

1.42

Table Vi. Effect of Bridging Group in Ansa-Metallocene On Copolymerization Rate and 1-Hexene Incorporation Activitv Cyclopentadienyi

--

Relative Activity 1H ,-

1.oo

1.oo

0.64

6.33

0.62

4.1 7

Cyclopentadienyl/ 0

Fiuorenyi

42c\

Cyclopentadienyl/ Fiuorenyi

(CH3)2C

:

37. Ligand Effects at Transition Metal Centers 403

Table VII. Effect of Bridging Group in Ansa-Metallocene On Copolymerization Rate and 1-Hexene Incorporation D

R

i

Relative Behavior ACtMtV 1 2.33 5.1 0

--

w

-CH2CH2(CH3)2Si1

2-CH3

1.01

4.75

1.oo

1.oo

0.63

3.61

1Single atom (Sl) bridging group.

Table VIII. Comparison of Relative Reactivity Ratios With Metallocene Catalysts vst

Relative Reactivity Ratios ~/pa,42 F/Hb (This m v )

(Me5CP)2ZrC12

5.21

7.14

(MeCp)2ZrC12

1.25

0.72

CppZrClp

1.OOd

1.OOd

(Me2Si)Cp2ZrC12

0.50

0.51

Et(Ind)2ZrClp

0.14

0.20

Me2C(Cp)(Flu)ZrClp

0.027C

0.24

aRun in toluene at 50%; E/P = ethylenelpropylene bRun in n-hexane at 85%; E/H = ethylendl -hexene CRun at 25’C dResults normalized to 1.OO for CppZrClp in both studies does not imply reactivity ratio for U P and U H copolymerizationsare the same

This Page Intentionally Left Blank

405

38. Propylene Polymerizations with Metallocene/Teal/Trityl (Pentafluorophenyl) Aluminate Mixtures

Tetrakis

John A. Ewen Catalyst Research Corporation, 1823 Barleton Way, Houston, Texas 77058 ABSTRACT Propylene polymerizations with dichlorometallocenes as pre-catalysts were investigated with trityl tetrakis (pentafluorophenyl)aluminate (Al-F20) as a synthetic reagent for the active zirconium monoalkyl cations. Triethylaluminium served as an alkylating agent for Zr and as a convenient water scavenger. Isospecific (Et[Indl2Zr(R)+) and syndiospecific (Me2C[CpFlulZr(R)+) cations paired with Al-F20 are significantly more stereospecific than the corresponding systems with boron and methylamluninoxane counter anions under certain polymerization conditions. The Al-F20 containing catalysts are less active than the B-F20 systems. INTRODUCTION It is well known that ligand steric effects influence chiral metallocenium catalyst stereospecificities. In contrast, there is only a limited amount of information on the dependence of polymerization stereospecificities on the structure and composition of counter anions paired with the active species in metallocene polymerizations. Initial investigations of the influence of anions in propylene polymerizations revealed that Me2C[CpFlulZr-R+ is more stereospecific when paired with MAO- than when paired with smaller counter anions such as CH3B(C6F5)- (B-F15) and B(CgF514- (B-F20) under conditions most favoring contact ion pairs.6) 7 ) Propylene polymerizations with Al(C6Fg)g- (Al-F20) paired with Me2C[CpFlulZr-R+ and Et[Indl2Zr-R+are presented and compared to the results with MAO- and the B-F20 systems. 8)

406 J. A. Ewen

The A1-F20 is the first example of a well defined anion providing better stereospecificities than previously obtained with MA0 derived counter anions. EXPERIMENTAL Polymerizations, polymer characterizations, and metallocene syntheses and IH NMR analyses have been previously described.9, The trityl synthetic procedure for tetrakis (pentafluoropheny1)aluminate is outlined in Scheme I.8, 16 mmol of n-butyllithium (1.6 M in hexane) were added slowly to an equivalent of bromopentafluorbenzene in 60 mL of toluene at -78°C and stirred for 2 hrs. 4 mmol of A1Br3 in 15 mL of toluene were added slowly to the cold slurry of precipitated, white LiCgFg. The mixture was warmed to room temperature and 4 mmol of Ph3CC1 in 20 mL of CH2C12 were added to obtain an orange slurry. The filtrate was dried under vacuum, crushed, and washed by stirring with pentane. The somewhat air sensitive yellow solids were further washed for about 12 hours with boiling hexane. 2.45 grams of [Ph3C][Al(C6F5)4-]were obtained as a bright yellow powder. The preparation was done under nitrogen using Schlenk techniques and avoided the intermediary preparation of the explosive Al(CgFg)3. Synthetic data and detailed procedures are in the published patent application. Scheme I

LiAl(B6F5)4 + Ph3CC1

4:l tol/CH2C12 b [Ph3Cl [ A - LiC1; 92%

The NMR and elemental analyses obtained for [Ph3C][A1(CgFg)4] were: lH-NMR (CD2C12): 8.26 t

38. Propylene Polymerizations with Catalysts 407

t(2), 7.66 d(2) (TMS 6 = 0.00 ppm); I9F-NMR (CD2C12): -124 (2), -159 (1) -166 (2) (C6F6 6 = -163.7 ppm); Obsd: 52% C; 1.6% H. Theory: 55% C; 1.6% H RESULTS AND DISCUSSION Polymerization results with Me2C[CpFlulZrC12 in conjunction with MAO, [Ph3C][B(CgF5)4], and [Ph3Cl [Al(CgF5)4] are recorded in Table 1. Table 1. Zr,

Bulk Polymerizations with Me2C[l-Cp-l'-FluIZrC12

Anion

p o l

1.1

MA0

(0.459) B-F20 1.1 (9 p o l ) 11.3 Al-F20 (13 p o l )

TEAL Eff., 10-3-Mv m.pt., rrrr ( m o l ) kg/g-cat.h OC %

0

Pol. Temp. 78

. I

5OoC

133

140

86

2

94

104

131

79

2

20

116

138

84

P o l . Temp. = 6OoC

1.1

MA0

(450 mg) 2.9 Al-F20 (25 p o l )

0

276

122

136

-

0.8

17

-

141

-

The other polymerization conditions were 1.4L propylene with time = 60 minutes. Subsequent to Zr alkylation by TEAL, the trityl cation performs either a carbanion abstraction or a b-hydride abstraction to generate the active sites: Zr-CH2CH3 + Ph3C(+)

-9

Ph3CH + H2C=CH2 + Zr+

(4)

At 50°C polymerization temperature the Al-F20 and MA0 systems have similar stereospecificities; both of which are higher than those obtained with B-F20 (Table 1). At 7O0C,

408 J. A. Ewen

higher than those obtained with B-F20 (Table 1). At 7 0 " C , where increased contact ion pairing is anticipated for organometallic systems,6 e 7 , the B-F20 system is less stereospecific than at 50°C. In contrast, the stereospecificity of the Al-F20 is increased at 70°C relative to 50°C. Indeed, the A1-F20 system surpasses the stereospecificity of the MA0 analog at 70°C. The isospecific system (Table 2 ) showed closely analogous results with an unusually low temperature dependence for stereoregulation with Al-F20. Table 2.

Bulk Polymerizations with Et[l-Indl2ZrC12

Zr

Anion

I

P o l

TEAL , (mmol)

Eff., 10-3-Mv kg/g-cat.h

m.pt., mmrmn OC

%

Pol. Temp. = 6OoC

2.4 1.4 3.1

MA0

(450 mg) B-F20 (8.6 p o l ) A1-F20 (16.8 p o l )

0

97

38

135

85

0.8

90

36

135

85

0.8

18

50

141

86

44

140

86

Pol. Temp. = 50°C

12.0

Al-F20 (60 p o l )

2.0

11

Some undefined steric interactions between the anions and the active species are presumably responsible for these unprecedented differences. It is noteworthy that B-F20 and Al-F20 can be anticipated as having similar ionic radii in spite of significantly differing B and A1 aromatic carbon bond lengths (Fig. 1). l o t 11) The novel differences in stereospecificities with differing anions offers new opportunities to control catalyst stereospecificities by tailoring the structure and composition of the anions. They also provide a qualitative rationale for changes in tacticities and stereospecificities with supported

38. Propylene Polymerizations with Catalysts 409

metallocenes.12) However, this empirical approach offers no simple working hypotheses analogous to the ligand steric effects. It is also noted that since the catalyst efficiencies with Al-F20 are consistently lower than obtained with MA0 and that the chemical stability of 0

8-Car = 1.67 A

0

&Car = 1.87 A

Figure 1. Relative Sizes of [BPh4]-, [AlPh41- and Me2Si [Indl22the counter anions and possible poisoning of the active sites by anion degradation products are also considerations. At this stage of development the logical Cp ligand effects on catalyst design clearly have practical advantages over empirically determined anion effects. REFERENCES 1. J.A. Ewen, L. Haspeslagh, M.J. Elder, J.L. Atwood, H. Zang, H.N. Cheng in Transition Metals and Organometallics as Catalysts for Olefin Polymerization; W. Kaminsky and H. Sinn, Eds.; Springer-Verlag, Berlin, 1987, p . 281 2. S. Miya, T. Mise, H. Yamazaki in Catalytic Olefin Polymerization; T. Keii and K. Soga, Eds.; Elsevier, Amsterdam, Oxford, New York, Tokyo, 1989, p. 483. 3. W. Roll, H.H. Brintzinger, B. Reiger, R. Zolk, Angew. Chem. Int. Ed. Engl., 2 9 , 279 (1990)

410 J. A. Ewen

4. 5.

6.

7. 8.

S. Collins, W.J. Gauthier, D.A. Holden, B.A. Kuntz, N.J. Taylor, D.G. Ward, Organometallics, 10, 2061 (1991) W. Spalek, M. Antberg, J. Rohrmann, A . Winter, B. Bachmann, P. Kiprof, J. Behm, W.J. Herrmann, Angew. Chem. Int. Ed. Engl., 31, 1347 (1992) J.A. Ewen, M.J. Elder, R . L . Jones, L. Haspeslagh, J.L. Atwood, S . G . Bott, K. Robinson, Makromol. Chem., Macromol. Symp. 48/49, 253 (1991) J.A. Ewen, M.J. Elder, Makromol. Chem., Macromol. Symp. 66, 179 (1993)

M.J. Elder and J.A. Ewen, Eur. Pat. Appl., Pub. N o . 0,573,403 ( 8/12/93) J. A. Ewen, M.J. Elder, R.L. Jones, S. Curtis, H . N . 9. Cheng, in Catalytic Olefin Polymerization, T. Keii and K. Soga, Eds. ; Elsevier, Amsterdam, Oxford, New York, Tokyo, 1990, p. 439 10. J.F. Malone and W.S. Mcdonald, Chem. Comm. p.444 (1967) 11. T.L. Blundell and H.M. Powell, Acta Cryst. Sect. B, p . 2 3 0 4 (1971) 12. W. Kaminsky, Polymerization results presented at this symposium.

STUDIES IN SURFACE SCIENCE AND CATALYSIS Advisory Editors: B. Delmon, Universite Catholique de Louvain, Louvain-la-Neuve, Belgium J.T. Yates, Universityof Pittsburgh, Pittsburgh, PA, U.S.A.

Volume 1

Volume 2

Volume 3

Volume 4

Volume 5

Volume 6

Volume 7

Volume 8 Volume 9

Volume 10

Volume 11

Volume 12 Volume 13 Volume 14

Preparation of Catalysts LScientific Bases for the Preparation of Heterogeneous Catalysts. Proceedings of the First International Symposium, Brussels, October 14-17,1975 edited by B. Delmon, P.A. Jacobs and G. Poncelet The Control of the Reactivity of Solids. A Critical Survey of the Factors that Influence the Reactivity of Solids, with Special Emphasis on the Control of the Chemical Processesin Relation to PracticalApplications by V.V. Boldyrev, M. Bulens and B. Delmon Preparation of Catalysts II. Scientific Bases for the Preparation of Heterogeneous Catalysts. Proceedings of the Second International Symposium, Louvain-la-Neuve, September 67,1978 edited by B. Delmon, P. Grange, P. Jacobs and G. Poncelet Growth and Properties of Metal Clusters. Applications to Catalysis and the Photographic Process. Proceedings of the 32nd International Meeting of the Societe de Chimie Physique, Villeurbanne, September 24-28,1979 edited by J. Bourdon Catalysis by Zeolites. Proceedings of an International Symposium, Ecully (Lyon), September9-11,1980 edited by B. Irnelik, C. Naccache, Y. Ben Taarit, J.C. Vedrine, G. Coudurier and H. Praliaud Catalyst Deactivation. Proceedings of an International Symposium, Antwerp, October 13-15,1980 edited by B. Delmon and G.F. Froment New Horizons in Catalysis. Proceedings of the 7th International Congress on Catalysis,Tokyo, June30-July4,1980. PartsAand B edited by T. Seiyama and K. Tanabe Catalysis by Supported Complexes by Yu.1. Yermakov, B.N. Kuznetsov and V.A. Zakharov Physics of Solid Surfaces. Proceedings of a Symposium, Bechyhe, September 29-October 3,1980 edited by M. LazniEka Adsorption at the GasSolid and Liquid-Solid Interface. Proceedings of an International Symposium, Aix-en-Provence, September 21-23,1981 edited by J. Rouquerol and K.S.W. Sing Metal-Support and Metal-Additive Effects in Catalysis. Proceedings of an International Symposium, Ecully (Lyon), September 1616,1982 edited by B. Imelik, C. Naccache, G. Coudurier, H. Praliaud, P. Meriaudeau, P. Gallezot, G.A. Martin and J.C. Vedrine Metal Microstructures in Zeolites. Preparation - Properties -Applications. Proceedings of a Workshop, Bremen, September 22-24,1982 edited by P.A. Jacobs, N.I. Jaeger, P. Ji& and G. Schulz-Ekloff Adsorption on Metal Surfaces. An Integrated Approach edited by J. Benard Vibrations at Surfaces. Proceedings of theThird International Conference, Asilomar, CA, September 1-4,1982 edited by C.R. Brundle and H. Morawitz

Volume 15 Volume 16

Volume 17

Volume 18

Volume 19

Volume 20

Volume 21

Volume 22 Volume 23 Volume 24

Volume 25

Volume 26

Volume 27 Volume 28

Volume 29 Volume 30

Volume 31

Volume 32 Volume 33 Volume 34

Heterogeneous Catalytic Reactions Involving Molecular Oxygen by G.I. Golodets Preparation of Catalysts 111. Scientific Bases for the Preparation of Heterogeneous Catalysts. Proceedings of the Third International Symposium, Louvain-la-Neuve, September 6-9,1982 edited by G. Poncelet, P. Grange and P.A. Jacobs Spillover of Adsorbed Species. Proceedings of an International Symposium, Lyon-Villeurbanne, September 12-16,1983 edited by G.M. Pajonk, S.J. Teichner and J.E. Germain Structure and Reactivity of Modified Zeolites. Proceedings of an International Conference, Prague, July9-13,1984 edited by FA. Jacobs, N.I. Jaeger, P. Jit, V.B. Kazansky and G. Schulz-Ekloff Catalysison the Energy Scene. Proceedings of the 9th Canadian Symposium on Catalysis, Quebec, P.Q., September 30-October 3,1984 edited by S. Kaliaguine and A. Mahay Catalysis by Acids and Bases. Proceedings of an International Symposium, Villeurbanne (Lyon), September 25-27,1984 edited by B. Imelik,C. Naccache, G. Coudurier, Y. Ben Taarit and J.C. Vedrine Adsorption and Catalysis on Oxide Surfaces. Proceedings of a Symposium, Uxbridge, June 2&29,1984 edited by M. Che and G.C. Bond Unsteady Processesin Catalytic Reactors by Yu.Sh. Matros Physicsof Solid Surfaces 1984 edited by J. Koukal Zeolites: Synthesis, Structure, Technology and Application. Proceedings of an International Symposium, Pottorot-Portorose, September 34,1984 edited by B. Driaj, S. HoEevar and S. Pejovnik Catalytic Polymerization of Olefins. Proceedings of the International Symposium on Future Aspects of Olefin Polymerization, Tokyo, July 44,1985 edited by T. Keii and K. Soga Vibrationsat Surfaces 1985. Proceedings of the Fourth International Conference, Bowness-omwindermere, September 15-19,1985 edited by D.A. King, N.V. Richardson and S. Holloway Catalytic Hydrogenation edited by L. Cervenq New Developments in Zeolite Science and Technology. Proceedings of the 7th International Zeolite Conference, Tokyo, August 17-22,1986 edited by Y. Murakami, A. lijima and J.W. Ward Metal Clusters in Catalysis edited by B.C. Gates, L. Guczi and H. Knozinger Catalysis and Automotive Pollution Control. Proceedings of the First International Symposium, Brussels, September 8-1 1,1986 edited by A. Crucq and A. Frennet Preparation of Catalysts IV. Scientific Bases for the Preparation of Heterogeneous Catalysts. Proceedings of the Fourth International Symposium, Louvain-la-Neuve, September 1-4,1986 edited by B. Delmon, P. Grange, P.A. Jacobs and G. Poncelet Thin Metal Films and Gas Chemisorption edited by P. Wissmann Synthesis of High-silica Aluminosilicate Zeolites edited by P.A. Jacobs and J.A. Martens Catalyst Deactivation 1987. Proceedings of the 4th International Symposium, Antwerp, September 29-October 1,1987 edited by B. Delmon and G.F. Froment

Volume 35 Volume 36

Volume 37

Volume 38

Volume 39

Volume 40

Volume 41

Volume 42 Volume 43 Volume 44

Volume 45 Volume 46

Volume 47 Volume 48

Volume 49

Volume 50

Volume 51 Volume 52

Volume 53

Keynotes in Energy-RelatedCatalysis edited by S. Kaliaguine Methane Conversion. Proceedings of a Symposium on the Production of Fuels and Chemicalsfrom Natural Gas, Auckland, April 27-30,1987 edited by D.M. Bibby, C.D. Chang, R.F. Howe and S.Yurchak Innovation in Zeolite Materials Science. Proceedings of an International Symposium, Nieuwpoort, September 13-17,1987 edited by P.J. Grobet, W.J. Mortier, E.F. Vansant and G. Schulz-Ekloff Catalysis 1987. Proceedingsof the 10th North American Meeting of the Catalysis Society, San Diego, CA, May 17-22,1987 edited by J.W. Ward Characterization of Porous Solids. Proceedings of the IUPAC Symposium (COPSI),Bad Soden a. Ts., April 26-29,1987 edited by K.K. Unger, J. Rouquerol, K.S.W. Sing and H. Kral Physicsof Solid Surfaces 1987. Proceedingsof the Fourth Symposium on Surface Physics, Bechyne Castle, September 7-1 1,1987 edited by J. Koukal Heterogeneous Catalysis and Fine Chemicals. Proceedings of an International Symposium, Poitiers, March 1517,1988 edited by M. Guisnet, J. Barrault, C. Bouchoule, D. Duprez, C. Montassier and G. PBrot Laboratory Studies of Heterogeneous Catalytic Processes by E.G. Christoffel, revised and edited by Z. P a d Catalytic Processesunder Unsteady-State Conditions by Yu. Sh. Matros Successful Design of Catalysts. Future Requirements and Development. Proceedingsof the Worldwide Catalysis Seminars, July, 1988, on the Occasion of the 30th Anniversary of the Catalysis Society of Japan edited by T. lnui Transition Metal Oxides. Surface Chemistry and Catalysis by H.H. Kung Zeolites as Catalysts, Sorbents and Detergent Builders. Applications and Innovations. Proceedings of an International Symposium, Wurzburg, September 44,1988 edited by H.G. Karge and J. Weitkamp Photochemistry on Solid Surfaces edited by M. Anpo and 1 .Matsuura Structure and Reactivity of Surfaces. Proceedings of a European Conference, Trieste, September 13-16,1988 edited by C. Morterra, A. Zecchina and G. Costa Zeolites: Facts, Figures, Future. Proceedingsof the 8th International Zeolite Conference, Amsterdam, July 10-14,1989. Parts Aand B edited by P.A. Jacobsand R.A. van Santen Hydrotreating Catalysts. Preparation, Characterization and Performance. Proceedings of the Annual International AlChE Meeting, Washington, DC, November 27-December 2,1988 edited by M.L. Occelli and R.G. Anthony New Solid Acids and Bases. Their Catalytic Properties by K. Tanabe, M. Misono, Y. Ono and H. Hattori Recent Advances in Zeolite Science. Proceedings of the 1989 Meeting of the British Zeolite Association, Cambridge, April 17-19,1989 edited by J. Klinowskyand P.J.Barrie Catalyst in Petroleum Refining 1989. Proceedings of the First International Conference on Catalysts in Petroleum Refining, Kuwait, March 5-8,1989 edited by D.L. Trimm, S.Akashah, M. Absi-Halabi and A. Biahara

Volume 54

Future Opportunities in Catalytic and Separation Technology edited by M. Misono, Y. Moro-oka and S. Kimura Volume 55 New Developments in SelectiveOxidation. Proceedings of an International Symposium, Rimini, Italy, September 18-22,1989 edited by G. Centi and F. Trifiro Volume 56 Olefin Polymerization Catalysts. Proceedings of the International Symposium on Recent Develo'pments in Olefin Polymerization Catalysts, Tokyo, October 23-25,1989 edited by T. Keii and K. Soga Volume 57A Spectroscopic Analysis of Heterogeneous Catalysts. Part A Methods of Surface Analysis edited by J.L.G. Fierro Volume 578 Spectroscopic Analysis of Heterogeneous Catalysts. Part B: Chemisorption of Probe Molecules edited by J.L.G. Fierro Introduction t o Zeolite Science and Practice Volume 58 edited by H. van Bekkurn, E.M. Flanigen and J.C. Jansen Heterogeneous Catalysis and Fine Chemicals II. Proceedings of the 2nd Volume 59 International Symposium, Poitiers, October 24.1990 edited by M. Guisnet, J. Barrault, C. Bouchoule, D. Duprez, G. Perot, R. Maurel and C. Montassier Chemistry of Microporous Crystals. Proceedings of the International Symposium Volume 60 on Chemistryof MicroporousCrystals,Tokyo, June 26-29,1990 edited by T. Inui, S. Narnba and T. Tatsumi Volume 61 Natural GasConversion. Proceedings of the Symposium on Natural Gas Conversion, Oslo, August 12-17,1990 edited by A. Holmen, K.-J. Jens and S. Kolboe Volume 62 Characterization of PorousSolids II. Proceedings of the IUPAC Symposium (COPS II),Alicante, May 6-9,1990 edited by F. Rodriguez-Reinoso, J. Rouquerol, K.S.W. Sing and K.K. Unger Volume 63 Preparation of Catalysts V. Proceedings of the Fifth International Symposium on the Scientific Bases for the Preparation of Heterogeneous Catalysts, Louvain-la-Neuve, September 3-6,1990 edited by G. Poncelet, P.A. Jacobs, P. Grange and B. Delrnon New Trends in CO Activation Volume 64 edited by L. Gucri Volume 65 Catalysis and Adsorption by Zeolites. Proceedings of ZEOCAT 90, Leipzig, August 20-23,1990 edited by G. Ohlmann, H. Pfeifer and R. Fricke Volume 66 Dioxygen Activation and Homogeneous Catalytic Oxidation. Proceedings of the Fourth International Symposium on Dioxygen Activation and Homogeneous Catalytic Oxidation, Balatonfured, September 10-14.1990 edited by L.I. Simandi Volume 67 Structure-Activity and Selectivity Relationships in Heterogeneous Catalysis. Proceedings of the ACS Symposium on Structure-Activity Relationships in Heterogeneous Catalysis, Boston, MA, April 22-27,1990 edited by R.K. Grasselli and A.W. Sleight Volume 68 Catalyst Deactivation 1991. Proceedings of the Fifth International Symposium, Evanston, IL, June24-26,1991 edited by C.H. Bartholomew and J.B. Butt Volume 69 Zeolite Chemistry and Catalysis. Proceedings of an International Symposium, Prague, Czechoslovakia, September 8-13,1991 edited by P.A. Jacobs, N.I. Jaeger, L. Kubelkova and B.Wichterlova

Volume 70 Volume 71 Volume 72

Volume 73 Volume 74 Volume 75 Volume 76 Volume 77

Volume 78

Volume79

Volume 8 0

Volume 81

Volume 82

Volume 83

Volume 8 4

Poisoning and Promotion in Catalysis based o n Surface Science Concepts and Experiments by M. Kiskinova Catalysis and Automotive Pollution Control II.Proceedings of the 2nd International Symposium (CAPoC 21, Brussels, Belgium, September 10-13,1990 edited by A. Crucq New Developments in Selective Oxidation by Heterogeneous Catalysis. Proceedingsof the 3rd European Workshop Meeting on New Developments in Selective Oxidation by Heterogeneous Catalysis, Louvain-la-Neuve, Belgium, April 8-10,1991 edited by P. Ruiz and B. Delmon Progress in Catalysis. Proceedings of the 12th Canadian Symposium on Catalysis, Banff,Alberta,Canada, May25-28,1992 edited by K.J. Smith and E.C. Sanford Angle-Resolved Photoemission. Theory and Current Applications edited by S.D. Kevan New Frontiersin Catalysis, Parts A-C. Proceedings of the 10th International Congress on Catalysis, Budapest, Hungary, 19-24 July, 1992 edited by L. Guczi, F. Solymosi and P.T6tBnyi Fluid Catalytic Cracking: Science and Technology edited by J.S. Magee and M.M. Mitchell, Jr. New Aspects of Spillover Effect in Catalysis. For Development of Highly Active Catalysts. Proceedingsof the Third International Conferenceon Spillover, Kyoto, Japan,August 17-20,1993 edited by T. Inui, K. Fujimoto, T. Uchijima and M. Masai Heterogeneous Catalysis and Fine Chemicals 111, Proceedingsof the 3rd International Symposium, Poiters, April 5 - 8,1993 edited by M. Guisnet, J. Barbier, J. Barrault, C. Bouchoule, D. Duprer, G. PBrot and C. Montassier Catalysis: An Integrated Approach t o Homogeneous, Heterogeneous and Industrial Catalysis edited by J.A. Moulijn, P.W.N.M. van Leeuwen and R.A. van Santen Fundamentals of Adsorption. Proceedings of the 4th International Conference on Fundamentals of Adsorption, Kyoto, Japan, May 17-22, 1992 edited by M. Suzuki Natural Gas Conversion II. Proceedings of the Third Natural Gas Conversion Symposium, Sydney, Australia, July 4-9, 1993 edited by H.E. Cuny-Hyde and R.F. Howe New Developments in Selective Oxidation II. Proceedings of the Second World Congress and Fourth European Workshop Meeting, Benalmbdena, Spain, September 20-24, 1993 edited by V. C o d s Corberen and S. Vic Bellbn Zeolites and M'lcroporous Crystals. Proceedings of the International Symposium on Zeolites and Microporous Crystals, Nagoya, Japan, August 22-25, 1993 edited by T. Hattori and T. Yashima Zeolites and Related Microporous Materials: State of the Art 1994. Proceedings of the 10th International Zeolite Conference, Garmisch-Partenkirchen, Germany, July 17-22, 1994 edited by J.Weitkamp, H.G. Karge, H. Pfeiifer and W. Htblderich

Volume 8 5 Volume 8 6 Volume 87

Volume 8 8

Volume 8 9

Advanced Zeolite Science and Applications edited by J.C. Jansen, M. Sttkker, H.G.Karge and J. Weitkamp Oscillating Heterogeneous Catalytic Systems by M.M. Slin'ko and N.I. Jaeger Characterization of Porous Solids 111. Proceedings of the IUPAC Symposium (COPS 111). Marseille, France, May 9-12, 1993 edited by J. Rouquerol, F. Rodriguez-Reinoso, K.S.W. Sing and K.K. Unger Catalyst Deactivation 1994. Proceedings of the 6th International Symposium, Ostend, Belgium, October 3-5, 1994 edited by B. Delmon and G.F. Froment Catalyst Design for Tailor-made Polyolefins. Proceedings of the International Symposium on Catalyst Design for Tailor-made Polyolefins, Kanazawa, Japan, March 10-12, 1994 edited by K. Soga and M. Terano