Fluorenyl complexes of zirconium and hafnium as catalysts for olefin polymerization Helmut G. Alta* and Edmond Samuelb aLaboratorium für Anorganische Chemie Universität Bayreuth D-95440 Bayreuth Germany bEcole Nationale Superieure de Chimie de Paris (UMR 7576- CNRS) 11 rue P. et M. Curie 75231 Paris Cedex 05 France In the continuously expanding panorama of Group 4 metallocene complexes and the extensive research devoted to their use as olefin polymerization catalysts fluorenyl complexes of zirconium and hafnium play a very special role. This singularity can be explained in terms of a particularly versatile bonding potential of the fluorenyl ligand causing a facile change from h5 ? h3 ? h1 (ring slippage) and thus providing new coordination sites in catalytic cycles.The substitution of fluorenyl ligands in unbridged or bridged (ansa) metallocene complexes allows the variation of steric requirements and symmetry of these complexes. Both parameters have a strong influence on the catalytic activity and stereospecifity as soon as these catalyst precursors are activated with cocatalysts like methylalumoxane (MAO). Synthetic methods X-ray structures and catalyst activities are discussed for several compounds together with updated literature. 1 Introduction and historical perspective Presently the field of metallocene complexes is one of the most attractive research areas in organometallic chemistry undoubtedly because of the versatile character of this family of compounds in a wide variety of reactions.Among these and by far the one of the highest importance is polymerization catalysis. In this respect the reactivity which they exhibit far supersedes any other known reagents and for this reason they are now well established as ‘the compounds of this decade’1 which ‘revolutionized olefin polymerization’.2 The polymers they produce have properties well beyond those obtained by any previously known technology. It should be pointed out that although the term ‘metallocene’was coined in the early 1950s for symmetrical sandwich Helmut G. Alt was born in 1944 and raised in South Bavaria. After high school he joined the army for two years and then he started his chemistry education at the Technische Universit�at M�unchen in 1966. After his PhD (1973) with M.Herberhold he spent 19 months as a postdoctoral fellow with M. D. Rausch at the University of Massachusetts USA. In 1975 he returned to Munich and conducted his preparative Habilitationsarbeit. In 1978 he moved to the newly founded Universit�at Bayreuth and finished his Habilitation in 1980. There he has a permanent position now as an extraordinary professor. His research interests are acetylene complexes and olefin polymerization reactions with metallocene complexes as catalysts. Helmut G. Alt Cl Zr Cl Since the 1950s Cp compounds with ferrocene as the classical example the explosion since 1980 of interest in Group 4 cyclopentadienyl compounds and their derivatives has led them to be covered by this designation in spite of their pseudo-tetrahedral structure and completely different chemical properties compared to the former.Their catalytic properties have become so well established that the term ‘metallocene catalysts’ immediately suggests complexes of Group 4 metals bound to a h5-C5H5 ring or one of their numerous fused ring derivatives. 2TiCl2 (Cp = cyclopentadienyl) has been known to be an active catalyst for ethylene polymerization,3,4 but the most important breakthrough in this field came with the synthesis of bridged chiral bis(indenyl) complexes of zirconium and hafnium and the discovery of their unusually high catalytic activity in combination with methylalumoxanes as co-catalysts. 5,6 Polymerization of propylene was found to be isospecific and exploration in the area of fluorenyls led to the discovery of ansa-cyclopentadienyl fluorenyl complexes as being equally efficient catalysts for the syndiospecific polymerization of propylene.7 Cl M Cl M Cl Cl M Cl Cl M = Ti Zr Hf Edmond Samuel is a Directeur de Recherche emeritus at the Centre National de la Recherche Scientifique (CNRS).He graduated from Paris University (the Sorbonne) where he obtained his PhD on Group 4 cyclopentadienyl and indenyl compounds an area which continued to be his main field of research. In 1970/1971 he spent a year as a postdoctoral fellow with Professor Marvin D. Rausch at the University of Massachusetts (Amherst). His recent collaboration with Professor John Harrod of McGill University (Montreal) led to the discovery of the catalytic properties of Group 4 metallocenes in the dehydrocoupling of silanes.Edmond Samuel 323 Chemical Society Reviews 1998 volume 27 The unusually high activity of these catalysts made it obvious that the fluorenyl ligand must have a strong influence. However a survey of the literature shows that thousands of cyclopentadienyl complexes have been described not quite as many indenyl derivatives and only very few fluorenyl complexes. What is the underlying reason? 2 The fluorenyl ligand and its various bonding modes Sandwich or half-sandwich compounds with fluorenyl as ligand have a history of their own. Shortly after the discovery of ferrocene some indenyl compounds were successfully prepared and studied but fluorenyls remained unknown for a long time although some have been cited as examples in several patents but have never been fully characterized.Attempts to prepare CpFeFlu and (Flu)2Fe failed presumably due to the involvement of the p-electrons of the central five-membered ring system in the aromaticity of the fused six-membered rings. In 1970 King and Efraty reported the synthesis of FluMn(CO)3 and described it as the first unequivocal evidence of a pentahaptobound fluorenyl ligand to a transition metal,8 but this assertion was based on NMR data and was not corroborated by X-ray evidence. However earlier in 1965 the first Group 4 bis(fluorenyl) complex (C13H9)2ZrCl2 had been prepared by Samuel and Setton9 and was characterized some years later by X-ray structure analysis indicating h5_ and h3_ bonding modes for the two fluorenyl ligands.10 All attempts to prepare the Ti and Hf analogs failed.It can thus be rationalized that the unique properties of fluorenyl complexes derive from the fact that the fluorenyl ligand is not simply a benzo-substituted derivative of its wellknown cyclopentadienyl congener but can be considered rather as a CH2 capped diphenyl. Thus the most characteristic feature of this ligand is the facile slippage of the central metal-bound five-membered ring from h5 ? h3? h1 coordination which is rather unusual among Group 4 cyclopentadienyl compounds. This behaviour may account for the difficulty in isolating fluorenyl compounds. However far from being a handicap this could explain their observed high activity in catalysis.On the other hand the rapid decomposition of fluorenyl complexes in donor solvents impeded the diversification of their chemistry in other areas and this remained stagnant. It appears then that there is a stability sequence in the triad Fig. 1 Various bonding modes of fluorenyl ligands in zirconium complexes Chemical Society Reviews 1998 volume 27 324 cyclopentadienyl > indenyl > fluorenyl Thus bis(indenyl)dichlorides of Ti and Zr could be prepared in fairly good yields and both could be hydrogenated to give the corresponding very stable tetrahydroindenyl compounds. The corresponding alkyl and aryl derivatives were then isolated. In the case of fluorenyls the Ti compound has resisted all attempts at isolation to this day and the hydrogenation of the bis- (fluorenyl)zirconium dichloride gave only decomposition products (mainly bifluorenyl).Only the bis(fluorenyl)dimethylzirconium complex11 (C13H9)2ZrMe2 was synthesized in 1974. The instability of the fluorenyls could be explained at first solely by steric constraints. Only years later was it recognized that ring slippage could also be a factor in their instability the reason being thence of traces of residual tetrahydrofuran used as a solvent in the preparation. Once this solvent was banished and attempts inspired by Brintzinger’s discovery in the field of bis(indenyl) chemistry to tether the two fluorenyl ligands in the ‘ansa’ fashion were successful a whole new field progressively opened up and a great number of new fluorenyl compounds of zirconium were synthesized and structurally studied along with their hitherto unknown hafnium analogs.Inevitably their polymerization properties were explored culminating in the discovery of some astounding results. 3 Various bridged and unbridged fluorenyl complexes The basic strategy to be applied therefore in the preparation of the whole family of fluorenyl zirconium or hafnium complexes is the use of toluene or diethyl ether as solvents. Fig. 1 shows examples for various bonding modes in fluorenyl zirconium compounds established by X-ray structures. They exhibit characteristic metal–ring bond distances according to the bonding hapticity.The fluorenyl ligand in the compound shown in Fig. 1c is in fact a monohapto (sigma-bonded) fluorenyl substituted in a bis(cyclopentadienyl) structure and it is one of the very rare examples of a h1 bonded fluorenyl ligand. The compound bearing two fluorenyls is also known. (Scheme 1).14 A series of unbridged bis(fluorenyl) zirconium compounds bearing a substituent either on the 9-position of the fivemembered ring or on the 4-position of the six-membered ring could be obtained as racemic or meso mixtures as characterized by their 1H and 13C NMR spectra.15,16 Now since all three 1 + 2 BuLi 2 +Cp2ZrCl2 with or without substituents on the chain,17 or chains containing heteroatoms such as Si18,19 or Sn.20 Strategies for their preparation and typical examples are shown in Scheme 2 and below; some complexes have been characterized by X-ray structural analysis.Zr 2 1 – 2 BuH 2 – 2 LiCl A convenient route for the synthesis of these mixed ansa metallocene complexes is the so-called fulvene method,12 used to prepare the various ligand precursors. A modified reaction allows the synthesis of the CH2-bridged derivative (Scheme 3).21 Scheme 1 Some of these compounds are chiral and are of special importance in olefin polymerization. Thus the C2H4 bridged bonding modes were evidenced by X-ray structures the 13C NMR chemical shift of C(9) in the fluorenyl ligand can be used as an indication to classify the bonding modes thus in the h1 case the chemical shift is d = 68 ppm for the h3 case d = 78 ppm and for the h5 case d is around 100 ppm.The dichlorides are usually sparingly soluble but the dimethyl compounds can be easily obtained as stable soluble compounds which give well resolved NMR spectra allowing one to identify their configuration. However the richest chemistry which could be developed in this area was that of the ansa-compounds with various bridges either linking two fluorenyls or a fluorenyl and a cyclopentadienyl ligand. These bridges can be either hydrocarbon chains 2 Bu nLi Li+ –Bu nH Li+ 2 H2O –2 LiOH H Li+ 2 Bu nLi Li+ 2 H2O –2 LiCl Scheme 2 R1 R2 ZrCl2 R2 R1 MCl4 H –2 LiCl MCl4 –2 LiCl Chemical Society Reviews 1998 volume 27 R1 = H R2 = C6H11 R1 = H R2 = Ph R1 = CH3 R2 = C6H11 Cl M Cl Cl M Cl 325 Zr R2E meso Si M Li+ bis(fluorenyl) complexes (C13H8–C2H4–C13H8)MCl2 (M = Zr Hf ) can be prepared according to this reaction (Scheme 4).22 2 H H ZrCl2 The X-ray structure of this bis(fluorenyl) complex shows that the C2H4 bridge does not bisect the bis(fluorenyl)zirconium dichloride moiety but points out of the symmetry plane.The NMR data suggest a dynamic mobility for this bridge. The Zr– Chemical Society Reviews 1998 volume 27 326 Cl Cl Zr Cl R2E Cl rac E = Si Sn; R = Me Ph Si Cl Cl M Cl Cl Si M = Zr Hf NMe2 – NMe2H Et2O Li+ 4 1 LiAlH 2 H2O 2 Bu nLi ZrCl4 –2 Bu nH –2 LiBr Et2O ZrCl2 Fig. 2 Molecular structure of (C13H8–C2H4–C13H8)ZrCl2 (two different views) Scheme 3 C-distances to the aromatic rings vary considerably between 241.7(5) and 269.8(5) pm and indicate a ‘disturbed’ h5 bonding mode (Fig.2)22 A special h3 dinuclear fluorenyl complex can be synthesized 5H4–CMe2–C13H8)ZrCl2 with Li(BHEt3) by the reduction of (C (Scheme 5).12 2 Bu nLi –2 Bu nH 2 Et2O Li+ H Cl Me Me Zr Li[BHEt3] Cl Me Cl H Me Br Zr Zr –2 LiBr Br Me H Cl Me H 2 Bu nLi ZrCl4 Scheme 5 H –2 Bu nH –2 LiCl Et2O Finally fluorenyl ligands can also be a component of halfsandwich complexes,23–25 as shown in the following example (Scheme 6). Scheme 4 4 Fluorenyl complexes as olefin polymerization catalysts As mentioned in the introduction metallocene complexes with Group 4 metals are now widely recognized as excellent olefin polymerization catalysts.The active catalytic species is considered to be a metallocene methyl cation generated by the Me Cl Si Zr Me Cl +2 BuLi + ZrCl4 N –2 BuH –2 LiCl But +Zr(NEt2)4 Me H Si Me –HNEt2 N Me But H H Si Me N NEt2 But Zr NEt2 Et2N Scheme 6 reaction of the metallocene dichloride precursor with methylalumoxane (MAO) the most commonly used co-catalyst. Other co-catalysts such as borates have also been used (Scheme 7). + Me X Cp¢ Cp¢ + MAO [MAOX]– M M Cp Cp X X = Me halogen M = Ti Zr Hf Instead of [MAOX]– other anions can be used such as [B(C6H5)4]– [B(C6F5)3Me]– or [B(C6F5)4]– Scheme 7 Among the various co-polymers that can be synthesized those with ethylene and propylene are largely of the most important industrial interest.Presently their market covers an annual worldwide production of 60 million tons of polyolefins and the metallocene catalysts are just beginning to have their share as newcomers with a continually soaring trend. Wherein resides this fascination in these catalysts? The following reasons may be invoked - They show an activity averaging up to a hundredfold compared to conventional Ziegler–Natta and PHILLIPS catalysts. - They can produce different types of polymers in terms of molecular weights molecular weight distributions and long and short chain branching. - They produce polymers with a very small molecular weight distribution.- They offer access to new materials for new applications and new markets. A milestone in the development of metallocene catalysts was the discovery that the symmetry of the catalyst precursor the metallocene dichloride complex determines the stereospecificity of the polymerization of prochiral olefins such as propylene. The pioneering work of Kaminsky26 and Brintzinger27 showed that bridged bis(indenyl) metal dichlorides of Zr and Hf of C2-symmetry or in their racemic forms are ideal catalyst precursors for obtaining isotactic polypropylene. Tethering a fluorenyl and a cyclopentadienyl ligand via a C1-bridge yields a complex of CS-symmetry and as a consequence the stereospecificity of the polymer changes from isotactic to syndiotactic.In order to understand the behaviour of these catalysts it is important to have in mind the well established bases of homogeneous catalysis which hold the formation of the catalytically active species as the triggering step. Razavi provided evidence in favour of the initial formation of a cation by using the complex [(C5H4–CMe2– C13H8)Zr(PMe3)Me]BF4 as a model.28 The next step in the catalytic cycle is the coordination of the monomeric olefin to the metal to give a metal–olefin complex. In the case of a stereospecific polymerization of a prochiral olefin such as propylene this step is of crucial importance since among the four possibilities for coordination one only fits in order to attain high stereospecificity. In the meantime other model complexes have been prepared that show two important features - The CNC axis of the olefinic ligand lies in the plane that bisects the metallocene fragment.- A prochiral olefin such as propylene is coordinated to the metal in a way that the methyl substituent has the least steric hindrance. As examples Cp2Zr(C2H4)(PMe3),29 [Cp(C5H4–CMe2– 9H7)]Zr(PMe3)30 and (C5H4–CMe2–C15H12)Zr(PMe3)- 4H7)31 can be mentioned. Molecular modelling is in agree- C (C ment with such a prediction.32 The actual polymerization step proceeds via the so-called olefin insertion and formation of the polymer chain. In most cases this step consists of an alkyl migration to the olefinic ligand generating a new coordination site at the vacant position of the alkyl ligand i.e.at the back side of the molecule. For catalytically active molecules with C2-symmetry no change in the configuration of the molecule takes place and for this reason prochiral olefins are polymerized to give isotactic polyolefins. However in the case of CS-symmetry the configuration of the catalyst changes during the inversion steps from re to si and vice versa to produce syndiotactic polyolefins. If the chain migration is blocked with a bulky substituent in the 3-position of the cyclopentadienyl ring of the ansa metallocene complex (C5H4–CMe2–C13H8)ZrCl2 the polymer chain cannot undergo this inversion step and the olefin coordination and subsequent insertion occurs from only one side of the molecule. As a consequence of the constant maintenance of the symmetry only isotactic polypropylene is formed.This is a very elegant method to change the tacticity with such mixed cyclopentadienyl fluorenyl complexes from syndiotactic to isotactic.33 It is not surprising that various substituents on the cyclopentadienyl ring have an influence on olefin polymerization because the substituent will be close enough to the metal to interfere with the basic polymerization reaction steps. What about substituents on various positions of the fluorenyl ligand? Studies have demonstrated that their influence is strong indeed.34 This could be due to the hindered or favoured orientation of the polymer chain formed during polymerization and thus modifying the activity of the corresponding catalyst.Table 1 shows results using various mixed ansa metallocene dichloride complexes activated with MAO and the influence of various substituents in positions 2 and 7 of the fluorenyl ligand. Unexpectedly the nature of the bridge is also crucial to the activity of the catalyst and on the molecular weight and tacticity of the polypropylene formed.35 The reason for this behaviour is not quite clear yet and it is rather intriguing. Electronic reasons are unlikely because there are four p-systems in the catalyst molecule that are able to compensate this effect. Razavi discussed a change of hapticity for the fluorenyl ligand in the case of the di(phenyl) bridged derivative (C5H4–CPh2– C13H8)ZrCl2.36 However the interaction of the catalyst cation and the MAO anion must be considered and could contribute to this behaviour such that b-hydrogen elimination which terminates polymer chain growth is blocked.Another aspect in the polymerization step is the fact that we are dealing with an ion pair during the whole polymerization process. If it is possible to separate the cation from the anion the activity should be enhanced because the free coordination site at the metal is of 327 Chemical Society Reviews 1998 volume 27 Table 1 Influence of various substituents on the fluorenyl ligand in isopropylidene bridged zirconium complexes in the syndiospecific polymerization of propylene 4 5 6 3 7 2 8 1 ZrCl Tm b/ rc (%) Activity (kgPP/ mmolCat. h) °C Complex H 2,7-Me2 4-Me 2 4,5-Me2 2,7-Ph2 2,7-(But) 2,7-Mes2 2,7-(MeO)2 2,7-Cl2 94.6 94.9 n.b.77.4 92.7 93.5 n.b. n.b. 92.9 90.5 93.7 72.0 135.0 131.2 111.5 107.9 132.5 142.0 132.7 96.3 n.b. 131.0 121.1 n.b. 33.5 11.2 35.5 9.3 17.6 54.8 77.1 0.3 20.7 26.8 34.9 10.5 2,7-Br2 3,4-Benzo 4,5-Benzo a catalysis. 2 h/ M Kg mol21 82 80 63 29 65 74 150 20 n.b. 60 37.5 154 Molecular weight (viscosimetry). b Melting point. c Ratio of syndiotactic pentads. easier access for the olefin. Simultaneously an increased Lewis acidity at the metal should favour faster kinetics. Indeed the activity of ansa bis(fluorenyl) complexes can be increased by a factor of eight for ethylene polymerization when methyl substituents are placed at positions 4 and 5.They behave as spacers towards the bulky MAO anion (Fig. 3).37 Fig. 3 Influence of methyl substituents in the 4- and 5-position of the fluorenyl ligand on the polymerization activity Mixed ansa cyclopentadienyl fluorenyl complexes with alkenyl substituents in the bridge show a very special behaviour (Fig. 4).38 They can be activated with MAO and then polymerize ethylene with high activity and without ‘fouling’ the reactor. Obviously the activated catalysts are able to form a copolymer with ethylene that is insoluble in hydrocarbon solvents. Such a ‘self-immobilization step’ of a catalyst combines the advantages of homogeneous and heterogeneous Chemical Society Reviews 1998 volume 27 328 Cl Cl Zr Zr Cl Cl Cl Cl Zr Zr Cl Cl Cl Zr 5 Conclusion Cl Fig.4 Some mixed ansa cyclopentadienyl fluorenyl complexes with alkenyl substituents Unlike the bridged fluorenyl complexes the unbridged counterparts are far less active towards olefin polymerization. Attempts have been made to achieve stereospecific polymerization of propylene with unbridged metallocene complexes by introducing bulky substituents in the 9-position of the fluorenyl ligands (Fig. 5).39 In all cases the activity of the corresponding catalysts was lower than in the case of the bridged analogs. The performance of fluorenyl complexes as catalysts in olefin polymerization makes it obvious that they cannot be compared with the cyclopentadienyl and indenyl analogs.Due to the fact that fluorenyl ligands are the best candidates for ring slippage reactions they can provide additional coordination sites at the metal and thus increase the catalytic activity. However this behaviour can also be a severe drawback because the thermal stability of such complexes is lowered. In order to avoid this problem the fluorenyl ligand should be fixed at the metal complex the same way as in ansa-metallocene complexes. An additional advantage of the fluorenyl ligands is their steric bulk. Since the catalytic species in olefin polymerization are supposed to be metallocene methyl cations the fluorenyl ligands are well suited to keep bulky counter anions such as tetraphenylborate or MAO at a distance.With spacers on the catalyst the separation of an ion pair can be achieved. As a consequence the activity of such a catalytic species increases. Finally substituents at special positions of the fluorenyl framework can have a drastic influence on the catalyst performance and they allow the ‘fine tuning’ of the polymer properties within a certain range. It is very likely that these substituents can influence the orientation of the polymer chain that is formed during polymerization. All these aspects point to the usefulness of fluorenyl ligands in transition metal complexes especially when catalytic processes are studied. However it should be kept in mind that these ligands have unique properties and they should be used with caution.7 References 1 W.-W. du Mont M. Weidenbruch A. Grochman and M. Bochmann Nachr. Chem. Techn. Lab. 1995 43 115. 2 K. B. Sinclair and R. B. Wilson Chem. Ind. 1994 857. 3 D. S. Breslow and N. R. Newburg J. Am. Chem. Soc. 1957 79 5072. 4 G. Natta P. Pino G. Mazzanti and U. Giannini J. Am. Chem. Soc. 1957 79 2975. 5 H. Sinn and W. Kaminsky Adv. Organomet. Chem. 1980 18 99. 6 F. R. W. P. Wild L. Zsolnai G. Huttner and H.-H. Brintzinger J. 7 J. A. Ewen R. L. Jones A. Razavi and J. D. Ferrara J. Am. Chem. Soc. 8 R. B. King and A. Efraty J. Organomet. Chem. 1970 23 527 and Organomet. Chem. 1982 232 233. 1988 110 6255. references therein. 9 E. Samuel and R. Setton J. Organomet. Chem. 1965 4 156. 10 C.Kowala P. C. Wailes H. Weigold and J. A. Wunderlich J. Chem. Soc. Chem. Commun. 1974 993. 11 E. Samuel H. G. Alt D. C. Hrncir and M. D. Rausch J. Organomet. Chem. 1976 113 331. 12 A. Razavi and J. Ferrara J. Organomet. Chem. 1992 435 299. 13 M. Bochmann S. J. Lancaster M. B. Hurthouse and M. Mazid Organometallics 1993 12 4718. 14 M. A. Schmid H. G. Alt and W. Milius J. Organomet. Chem. 1997 541 3. 15 K. Patsidis and H. G. Alt J. Organomet. Chem. 1995 501 31. 16 M. A. Schmid H. G. Alt and W. Milius J. Organomet. Chem. 1996 525 15. 17 B. Peifer M. B. Welch and H. G. Alt J. Organomet. Chem. 1997 544 115. 18 K. Patsidis H. G. Alt W. Milius and S. J. Palackal J. Organomet. Chem. 1996 509 63. 19 P. Schertl and H. G. Alt J. Organomet.Chem. 1997 545–546 553. 20 K. Patsidis H. G. Alt S. J. Palackal and G. R. Hawley Russ. Chem. Bull. 1996 45 2216. 21 H. G. Alt and R. Zenk J. Organomet. Chem. 1996 526 295. 22 H. G. Alt W. Milius and S. J. Palackal J. Organomet. Chem. 1994 472 113. 23 J. Okuda F. J. Schattenmann S. Wocadlo and W. Massa Organometallics 1995 14 789. 24 H. G. Alt K. Föttinger and W. Milius J. Organomet. Chem. 7916 (in press). 25 B. Rieger J. Organomet. Chem. 1991 420 C17. 26 H. Sinn and W. Kaminsky Adv. Organomet. Chem. 1980 18 99. 27 H.-H. Brintzinger D. Fischer R. Mülhaupt B. Rieger and R. Waymouth Angew. Chem. Int. Ed. Engl. 1995 34 1143 and references therein. 28 A. Razavi and U. Thewalt J. Organomet. Chem. 1993 445 111. 29 H. G. Alt C. E. Denner U. Thewalt and M. D. Rausch J. Organomet. Chem. 1988 356 C83. 30 H. G. Alt J. S. Han and U. Thewalt J. Organomet. Chem. 1993 456 89. 31 H. G. Alt and R. Zenk J. Organomet. Chem. 1996 522 177. 32 L. Cavallo G. Guerra M. Vacatello and P. Corradini Macromolecules 1991 24 1784. 33 R. Razavi L. Peters L. Nafpliotis D. Vereecke K. Den Dauw J. L. Atwood and U. Thewalt Macromol. Symp. 1995 89 345. 34 H. G. Alt R. Zenk and W. Milius J. Organomet. Chem. 1996 514 257; 1996 522 39. 35 H. G. Alt and R. Zenk J. Organomet. Chem. 1996 518 7; 295. 36 A. Razavi and J. L. Atwood J. Organomet. Chem. 1993 459 117. 37 P. Schertl Dissertation Universit�at Bayreuth 1996. 38 B. Peifer W. Milius and H. G. Alt J. Organomet. Chem. 1998 553 205. 39 M. A. Schmid H. G. Alt and W. Milius J. Organomet. Chem. 1995 501 101. Fig. 5 Various examples of unbridged metallocenedichloride complexes 6 Acknowledgements Received 3rd April 1998 Accepted 28th April 1998 We thank all our co-workers who have contributed to these results. We also thank the Deutsche Forschungsgemeinschaft and the PHILLIPS Petroleum Company U.S.A. for financial support. 329 Chemical Society Reviews 1998 volume