摘要:
Cationic Group 4 metallocene complexes and their role in polymerisationcatalysis: the chemistry of well defined Ziegler catalysts*Manfred BochmannSchool of Chemical Sciences, University of East Anglia, Norwich NR4 7TJ, UKCationic alkyl complexes of Group 4 metallocenes of the type [MCp,R]+ (M = Ti, Zr or Hf, Cp = C,H,)have been recognised as the catalytically active species in metallocene-based olefin polymerisation catalysts.These highly electrophilic 14-electron species possess a very complex chemistry in which the formation oftemporarily dormant stabilised adducts plays a dominant role. Cationic metal alkyls of this kind are foundto be extremely active polymerisation catalysts, with high stereoselectivities and the potential to producenumerous previously inaccessible polymeric materials.A detailed understanding of the chemistry of thesespecies promises to lead to a new generation of well defined polymerisation catalysts. Metallocene-basedcatalysts already play an increasing role in major industrial polymerisation processes.Since Ziegler's discovery of the polymerisation of ethylene withTiCl,-AlClEt, catalysts just over 40 years ago,' shortlyfollowed by Natta's discovery of the stereoselective polymeris-ation of propene,, the polymerisation of a-olefins has developedinto a giant industry. World production of polyolefins in 1995is estimated to be 53.6 million tons, twice the figure of 1983,and is expected to grow by 50% over the next 10 years.3 Thisindustry is based on heterogeneous catalysts which, after fourdecades of development, have become highly selective and effi-cient. Classical Ziegler catalysts are heterogeneous, with thepolymerisation taking place on dislocations and edges of TiCl,crystals (cf: structures 26a and 26b below).Consequently, thereare many different types of active sites, and the resultingpolymer has a typically broad molecular weight distribution.,Alternatives are chromium-based catalysts, e.g. Cr on silica(Phillips catalyst) which are comparatively less active andpolymerise ethylene to give high-molecular-weight rigidpolymethylene-type polymers; they are inactive towardspropene. Heterogeneous titanium-based Ziegler catalysts, onthe other hand, are more versatile; they are able to polymerisepropene with a very high degree of stereoselectivity and catalysethe copolymerisation of ethylene with higher alk-1 -enes such ashex-1 -ene, important for the production of flexible, non-brittlepolymers for use as films and in packaging.Although these aresuccessful processes which give rise to a wide range of polymerproducts, the diversity of active sites in heterogeneous catalystsleads to an uneven degree of comonomer incorporation, with ahigh incorporation rate in short chains, and little incorporationin the high-molecular-weight fraction, a drawback wheremetallocenes offer particular promise.By the 1980s it appeared that the development ofheterogeneous catalysts was maturing rapidly, with few majoradvances expected. At this juncture several developments in thechemistry of metallocenes came together to produce a furtherquantum leap in polymerisation catalysis.Shortly after the synthesis of the first Group 4 metallocenesby Wilkinson et al., in 1953 the use of these complexes aspolymerisation catalysts was tested.Mixtures of [TiCl,Cp,]and AlClEt, were found to polymerise ethylene withcomparatively moderate activity; propene was not polymer-i ~ e d . ~ . ' These catalysts were however prone to reduction toinactive titanium(rr1) species and could not compete withthe highly active and stereoselective heterogeneous catalysts.* Non-SI units employed: cal = 4.184 J; bar = lo5 Pa.Consequently they found no industrial application, althoughthese soluble systems were of value for mechanistic studies.An important further development began with theunexpected observation that although Ziegler catalysts arevery sensitive to hydrolysis, traces of water actually increasedthe rate of polymerisation in titanocene catalysts, and theformation of aluminoxanes by partial hydrolysis of thealuminium alkyl components was suggested. This phenom-enon was investigated in detail by Sinn and co-workers whofound that the normally inactive system [ZrCp,Me,]-AlMe,becomes highly active upon the addition of water." The same'effect is achieved when AIMe, is partially hydrolysed tomethylaluminoxane (MAO) before the transition-metal com-plex is added.Methylaluminoxane is a poorly characterisedpolymeric glassy substance, usually of molecular weight 900-1200 and the approximate composition [{ MeAlO},], whichconsists of linear, cyclic and cross-linked compounds, probablycontaining predominantly four-co-ordinate A1 centres andsome OAlMe, end groups.The compound dissolves readilyin hydrocarbons such as toluene where, due to the facile ligandexchange in aluminium complexes, it establishes complexsolution equilibria. Samples are usually rich in methyl groups,and even after drying in uucuo contain typically 3 4 % freeA1,Me6;l2 in commercial MA0 samples as much as 3040%trimethylaluminium may be present. Repeated pumping leadsto loss of Al,Me, and the increasing formation of an A1,0,network which no longer dissolves in toluene. The A12Me6content in MA0 solutions guards against such heterogenisation.As Sinn and co-workers" found, MA0 is a much moreeffective activator for metallocene dihalides than aluminiumalkyl halides and even allows the polymerisation of propenewith [ZrCl,Cp,] (although these simple Cp complexes lackthe stereoselectivity of the heterogeneous catalysts).In order toachieve high activity, MA0 has to be employed in a large excessover the zirconocene component, usually at A1 : Zr ratios of103-104:1, so that in such catalysts the cost of MA0 by faroutstrips that of the zirconium complex. In spite of thisdrawback MA0 has become the most widely used activator formetallocene-based catalyst systems, including large-scaleindustrial processes.Mechanistic AspectsAlthough metallocene-based catalysts are, as soluble systems,in principle more amenable to mechanistic study thanheterogeneous catalysts, it proved for a long time ratherJ. Chem.Sac., Dalton Trans., 1996, Pages 255-270 25difficult to underpin the various mechanistic ideas with firmexperimental evidence. In 1959 Breslow and Newburg 'suggested for their [TiCl,Cp,]-AlClEt, system that the trans-ition metal is first alkylated and then forms a halide-bridgedbinuclear complex which is capable of reaction with ethylene6Scheme 1). Some polarisation of the molecule was envisagedwith a partial positive charge on titanium and negative chargeon aluminium. ' This proposal anticipated several features ofthe well known Cossee-Arlman polymerisation mechanism l4for heterogeneous catalysts, such as the Lewis acidity of themetal centre and the mutual cis orientation of the metal-alkylbond and the incoming monomer ligand.There was sub-sequently some debate about the nature of the active speciesand it was widely assumed that a probably halide-bridgedtitanium-aluminium mixed-metal species was involved. Variousplausible intermediates were suggested, for example asformulated by Henrici-Olive and Olive (Scheme 2).Kinetic studies on [TiCl,Cp,]-AlClR, polymerisationsystems provided indications of an 'intermittent' chain-growthmechanism. l 6 A mechanistic scheme was proposed in which thewell known adduct between the titanocene complex and thealuminium alkyl is converted into an active species C* which isable to insert ethylene molecules. After each insertion step theactive complex carrying a polymer chain, C*-P, can becometemporarily dormant as a stabilised complex which isScheme 1Scheme 2cp, CP" +I cp,*2Scheme 3 C* = active [TiC12Cp2]-AICIR2 polymerisation species,P = polymer chainspectroscopically detectable.Renewed activation then allowsfurther chain growth (Scheme 3). These studies did not, ofcourse, provide information about the nature of C*; they didhowever indicate that any metal-olefin n complex involvedcould only be present in minute concentrations. The studies alsoprovided an estimate of the relative rates of ethylene insertioninto Ti-R bonds; Ti-Me reacts with C2H4 120 times slowerthan Ti-Et and 96 times slower than Ti-Pr, in line with theobservation that in this and other systems rapid chain growthis already observed when substantial amounts of potentiallyreactive Ti-Me species are still present in the system.'The chain growth sequence outlined in Scheme 3 iscompatible with the assumption of electron deficient cationicintermediates, [Cp,Ti-R]+ (2), as the active species C*(Scheme 4). Such a cationic 14-electron complex could beformed by dissociation of the adduct 1 and would possess aco-ordination site suitable for binding an olefin molecule cisto the Ti-alkyl bond, ready for facile alkyl migration and chaingrowth (alkene 'insertion'). After each step the anion, here[AlCl,R,] -, would be able to occupy this co-ordination sitewith formation of a halide-bridged bimetallic complex, repre-senting the temporarily dormant spectroscopically observablespecies C-P, of Scheme 3.The possibility of the participation of a cationic active species[TiCp,Me]+ had been suggested by Shilov and co-workers asearly as 1961.l 7 There was, however, no direct evidence for theexistence of such a species. Dyachkovskii attempted the electro-chemical generation of [TiCp,Me] + in dichloromethane andindeed found that ethylene polymerisation took place only inthe cathode chamber." However, the idea of ionic intermedi-ates did not find widespread acceptance as an explanation for areaction which typically took place in very non-polar solvents.Interest was revived by the discovery of Eisch et al. in 1985that Ph-CS-SiMe, reacted with [TiCl,Cp,] in the presenceof AlC1,Me to give the cationic titanium vinyl complex 3(Scheme 5) formally the product of alkyne insertion into apostulated '[TiCp,Me] + ' intermediate.In the following year we reported the synthesis of the firstcationic titanocene methyl complexes, [TiCp,Me(L)] +X-(L = NH3, NCR, pyridine, etc.; X = PF, or BPh,).,'Although the presence of strongly bound donor ligands Lprevented any reactions with ethylene, the complexes readilyunderwent insertion reactions with nitriles.20,21 At around thesame time Jordan et al.reported analogous cationic zirconiumcomplexes [ZrCp,Me(L)]BPh,, including the labile tetrahydro-furan (thf) complex which partially dissociates in dichloro-methane and polymerises ethylene; the first evidence thatwell defined cationic metallocene complexes were capableof polymerising olefins in the absence of aluminium alkylactivators.22 The related benzyl complex [ZrCp,(CH,Ph)-(thf)]BPh, was similarly active for the polymerisation ofethylene but not p r ~ p e n e , ~ ~ .~ ~ as were the titanocene-ethercomplexes [TiCp,Me(L)]BPh, (L = thf, Et20 or PhOMe).25256 J . Chem. SOC., Dalton Trans., 1996, Pages 255-271 [TiC12Cpd + AICI2Me - [ TiCp2Me+ AIC14-?PhCSSiMe3 I3Scheme 54% do4a 4bAlthough the activities of these donor-stabilised cationiccomplexes were at best modest, they did play a useful role asmodels for the active species. The details of the interaction ofethylene with [MCp,Me]+ has been the subject of numeroustheoretical studies at various levels of sophistication. Theresults tend to be quite sensitive to the methods used and differin the importance attached to features such as the presenceor absence of agostic interactions, the relative stability of theethylene Tc-complex and the height of the activation barriers.However, all agree on the general features such as the necessityof generating a 14-electron [MCp2R] + species.Such a complexhas two low-lying unoccupied orbitals, d, (4a) and d, (4b).,"The d, orbital acts as the acceptor for the incoming ethylenemolecule. The trajectory of the approach of the co-ordinatedolefin towards the alkyl chain on the way to forming the newC-C bond is shown in Scheme 6, based on the ab initiocalculations by Jolly and Maryni~k.,~ The interatomicdistances for the optimised transition-state geometry shown inScheme 6 correspond quite closely to the results obtained byZiegler and co-workers for the analogous zirconium reaction[ZrCp,Me] + + C2H4 using density function methods.28The details of the chain growth mechanism and the structureof the transition state have attracted considerable attention, notleast because of the implication for the stereocontrol in thepolymerisation of alk-1 -enes.In 1983 Brookhart et al. suggestedthat the insertion step may be facilitated by an a-agosticinteraction of one C-H bond of the alkyl ligand with the Lewis-acidic metal centre.29 This is only possible if the alkyl ligandadopts the conformation shown for the CH, ligand in 5(Scheme 6), rather than the conformation resulting from arotation by 60".Jolly and Marynick suggested for [Ti-Cp,Me(C,H,)] + that structure 5 may be primarily adoptedin order to avoid unfavourable interactions between CH, andthe Cp ligands, rather than because of a strong energy gainby a Ti. HC i n t e r a ~ t i o n . ~ ~ By contrast, recent ab initiocalculations by Ahlrichs and co-workers suggest strong a-agostic bonding of one methyl hydrogen in [TiCp,Me] +, to thepoint of an almost 'carbenoid' character of the methyl ligand.These authors do not find any significant role for an ethylenepre-complex on the potential energy surface, suggesting that theinsertion proceeds without definable intermediates or indeedwithout a noticeable activation barrier.30 Others 2 7 3 3 1 havecalculated activation barriers of 7-10 kcal mol-', in goodagreement with experimentally determined values,32 and findcomparatively high ethylene binding energies of 20-30 kcalmolThere is evidence from both Extended Hiickel 3 3 and ab initiocalculations 34 that the alkene complex [ZrCp2Me(C2H4)] +does not initially possess a-agostic interactions but that thesewhich increase in the order Ti < Hf < Zr.315Scheme 6develop as the CH, ligand tilts away from the Zr-C axistowards the C2H4 ligand, with CH, and C,H, hydrogens ineclipsed conformations. This interaction stabilises the transitionstate, with a very low activation barrier (if any).The whole in-sertion sequence is quite exothermic, AG < - 30 kcal mol-' . 3 1The participation of an a-agostic interaction during the chaingrowth process has been elegantly demonstrated by Brintzingerand co-workers35,36 for the hydrodimerisation of E- and Z-BuCH=CHD (Scheme 7). Since Zr HC is favoured overZ r .. . DC the presence of agostic bonding in the rate-determining step would lead to an unequal distribution oferythro and threo products from a non-chiral catalyst, andindeed a stereokinetic isotope effect of 1.3 : 1 (erythro-threo)has been found.35 Scandium catalysts show a similar isotopeeffect.37The primary product of the insertion, the propyl complex 6(Scheme 6), is also stabilised by an agostic interaction, this timewith a y-CH bond. It has been suggested that there is asubsequent rearrangement to a more stable P-agostic structurewhich may represent a 'resting stage' in the catalytic process 36although there may be a substantial energy barrier for such ay-$ rearrangement.31Surprising results concerning the speed of the insertionprocess were recently obtained with ab initio moleculardynamics simulations on [H,Si(C,H,),ZrMe] + + C2H4.34The whole ethylene insertion process seems to take place ona 70-170 fs time-scale.During this process the H2Si(C5H4),ligand framework proved to be remarkably flexible, with(C5H4)-Si-(C5H4) angles changing from ca. 75 to 98 and backto 75". No doubt these figures will be revised; they underlinehowever the extraordinarily high catalytic activities one mightexpect from these catalysts in the absence of other mitigatinginfluences.A consequence of the insertion sequence of Scheme 6 is thealternation of the site for monomer co-ordination after eachinsertion step, a facet that becomes crucial for the stereocontrolof alk- 1 -ene polymerisations (Scheme 8).Chain termination occurs typically via P-hydride elimination,a process that is about three orders of magnitude slower thanthe insertion step.For propene oligomerisations with stericallyhindered [Zr(C,Me,),R] + catalysts termination by P-methyltransfer has also been observed. 38 Energetically, P-CH,elimination is preferred,39 while P-H elimination is kineticallymore facile. The resulting [Cp2M-H] + fragment rapidlyinitiates the growth of a new polymer chain.Synthesis and Chemistry of Electron-deficient[ Cp2M-R] + CationsThe mechanistic discussion in the previous section is based onthe assumption that 14-electron species [MCp2R] + , howevertransient, are formed in catalytically active systems, ind/ependentof the nature of the 'activator' (e.g.AlCIEt, or MAO). Theconcentration of [MCp,R]+ in such mixtures is of course notknown, and conventional metallocene-aluminium alkyl mix-J. Chem. SOC., Dalton Trans., 1996, Pages 255-270 251 H2 I H2etc.Scheme 88Scheme 9tures are far too complex to identify reactive intermediatesor to determine their lifetimes, resting and active stages, ordecomposition pathways. A direct synthesis of 'base free'[MCp2R] +X- would not only confirm mechanistic assump-tions and allow the underlying chemistry of such complexes tobe studied, but might also lead to the preparation of essentially'neat' catalytically active sites and therefore provide catalystsof extremely high activity.Early studies had shown that even donor-stabilised metalalkyl cations are very potent electrophiles which do not tolerateconventional 'non-co-ordinating' anions such as BF4- andPF6- but form isolable salts with BPh,- .20921*22 These studieshave also established the protolysis of metallocene dialkyls withammonium salts as a general route to cationic metallocenederivatives [equation (l)].209251 H2 1 H2In 1989 Turner and c o - w ~ r k e r s ~ ~ reported the formationof the zwitterionic complex 7 from the reaction of [Zr-Cp*,Me,] (Cp* = C,Me,) and WHBu3][B(C6H4Et-3),] intoluene.Similarly, the reaction between [ZrCp*,Me,] and theacidic nido-carborane C2B,H,, gave 8 in which the anion isco-ordinated via a hydride bridge (Scheme 9). Both complexespolymerise ethylene rapidly under mild condition^.^' Complex7 evidently arose by electrophilic attack of [ZrCp*,Me] +on the anion. The first cationic 'base free' titanium and zir-conium complexes [MCp',Me][BPh,] were obtained from[MCp',Me,] and wHMe,Ph][BPh,] in dichloromethane at-40 "C in essentially quantitative yield (M = Ti or Zr, Cp' =CSH, or q '-inden~l).~' Under these conditions the compoundsare presumably present as dichloromethane solvates. Theinjection of ca. 5 equivalents of ethylene to solutions of[TiCp',Me]+ (Cp' = indenyl) resulted in the formation ofpolyethylene, while most of the Ti-Me complex remainedunreacted, a behaviour reminiscent of the kinetic results byFink and co-workers on [TiC1,Cp2]-A1ClMe, systems.Although [TiCp',Me] + proved to be thermally labile inchlorinated solvents, the complex polymerised not onlyethylene but also propene, in contrast for example to the[TiCl,Cp,]-AlClEt, catalysts.An analogous reaction allowedthe isolation of piCp*,Me]BPh,;42 like similar pentamethyl-cyclopentadienyl titanium complexes [TiCp*Me(L)] + (L = thfor tetrahydrothiophene) 43944 this compound is catalyticallyinactive, presumably for steric reasons.These studies showed that although equation (1) indicates theformation of a single species, the in situ generation of cationiccatalysts leads in fact to a complex equilibrium of variousadducts of the highly electron-deficient [MCp',R] + cation.The amine liberated during the protolysis step, the counteranionand the solvent are capable of binding to [MCpf2R]+.Underpolymerisation conditions this means that all these adducts arein competition with the only productive species, the olefincomplex (Scheme 10). Although the notion of 'single site'catalysts for metallocenes 4 5 9 4 6 is an apt description as far as thetransition state is concerned, it is evident that there is a wholerange of resting states the concentration of which must beminimised in order to obtain highly productive catalysts. Onlyif the relative importance of these equilibrium complexes andthe underlying chemistry are known can one begin to estimatethe proportion of species that are productive at any one time,and hence evaluate actual catalyst activities.The most important contribution in these equilibria is the co-ordination of the c~unteranion.~' For the tetraphenylborateanion various types of interaction with electron-deficient metalcentres have been found; examples are complexes 9-1 1 .48-50258 J.Chem. Soc., Dalton Trans., 1996, Pages 255-27Scheme 1015CH2Ph11 10Fluoro-substituted tetraarylborates are less basic, although[B(C,H,F-4),]- has been shown to co-ordinate via a single Fatom, as in 12.’lThe role of carboranes as comparatively non-co-ordinatinganions has been mentioned above (Scheme 9).,’ There has beensome success with [ZrCp,Me]+ salts of anions of the type[M(C,B,H,,),]- (M = Fe, Co or Ni); these catalysts possessconsiderable activity for the polymerisation of ethylene andfor ethylene-butene copolymerisation.5 2 On the other hand,carboranyl anions can act as polydentate ligands as in 13 to givespecies of very low catalytic activity. 53cA substantial reduction in cation-anion interaction wasachieved with Turner’s introduction of perfluorotetraphenyl-borate [B(C,F,),] - as the c~unteranion.’~ This borate isconsiderably less basic and less prone to phenyl transferreactions than [BPh,] - and has produced cationic polymeris-ation catalysts of exceptionally high activity.Cationic complexes can be generated from Group 4 metalalkyls either by protolysis using salts of very weakly basicamines such as [NHMe,Ph][B(C,F,),] or, if amine as theby-product is to be avoided, by alkyl-anion transfer with[CPh,][B(C,F,),] (Scheme 1 1, pathway A).55,56 However,although [B(C6F5),]- comes close to the ideal of a ‘non-co-ordinating anion’, at least in dilute solutions, even herethere are detectable metal-fluoride interactions in many cases.For example, the crystal structure of the thorium complex[ThCp*,Me][B(C,F,),] (14) shows close Th F contacts of2.675 and 2.757 A.The ethylene polymerisation activity of 14 isabout 3500 times higher than that of the [BPh,]- ana10gue.’~In the zirconium hydride complex 15 a similar weak co-ordination to two F atoms of [HB(C,F,),]- was found, which14Ph$-R17Scheme 11was surprisingly favoured over anion co-ordination via the B-Hmoiety.’*As an alternative to the synthesis of cationic complexesMarks and co-workers introduced the reaction of metallocenedimethyls with B(C,F,), to give methyl-bridged zwitterioniccomplexes of type 16 (Scheme 11, pathway B).’, This routeoffers the advantage that the product is stabilised by methyl co-ordination and is less polar and significantly more soluble intoluene.Several such complexes have been characterised byX-ray diffraction. 59*60 The Zr-CH,-B bridge is compar-atively weak, with rather long Zr-C distances which dependon the steric requirements of the Cp ligands. Two of theJ, Chem. Soc., Dalton Trans., 1996, Pages 255-270 251 [MCp’Med 18a, Cp’ = C5Me5i8b, Cp’ = C5H3(SiMe&IHtwie19Scheme 12Hmethyl hydrogens are agostically bonded to zirconium.It isthought that under catalytic conditions 16 partly dissociatesto [ZrCp‘,Me] + [BMe(C,F,),] -, and indeed complexes ofstructure 16 are good ethylene polymerisation catalysts.As is generally the case, the activity of 16 depends on thebulkiness of the Cp’ ligands. If Cp‘ = C,H, or C,H,Me,-1,2,the activities are comparable to those of the analogous[ZrCl,Cp,]-MA0 catalysts. Hydrogenolysis of 16 leads to thecationic hydrides [ZrHCp‘,] +; these react with propene to givelow-molecular-weight polymers (Cp’ = C,Me, or C,H,Me,),while the bulkier [ZrH(C,H,Bu‘,-1 ,3),] + selectively dimerisespropene to 2-methylpentenes. 5 9 bIn contrast to the methyl complexes the reaction ofzirconocene dibenzyls with B(C6F5), gives only ionic products,[ZrCp,(CH,Ph)]+[PhCH,B(C,F,),]- (17) (Scheme 11, path-way C).60 In this case the electron deficiency of the metal cationis partly alleviated by the q2-co-ordination of the benzyl ligand,indicated spectroscopically by a high-field shift of ca.30 ppm ofthe ’ NMR resonance of the @so-carbon of the phenyl ring,as well as characteristic changes in the C-H and H-H couplingconstants of the benzylic CH, group. The NMR spectrum of[ZrCp”,(CH,Ph)] + [B(C6F5),] - in toluene at 60 “C confirmedthe stability of the q2-benzyl bonding mode at highertemperatures, i.e. under polymerisation conditions [Cp” =C,H,(SiMe,),-1,3]. Several cationic benzyl complexes havebeen prepared; those with stereoselective ligands, such asrac-[Zr(C,H,)Cp‘,(CH,Ph)] + (Cp‘ = indenyl), proved to beexcellent propene polymerisation catalysts, with similar stereo-selectivities but higher productivities than the correspondingrac-[ZrCl , (C, H,)Cp‘,]-MAO catalysts.’ ,6Not only arene substituents in benzyl ligands are able todisplace a weakly co-ordinating anion from the co-ordinationsphere. The reaction of the half-sandwich compounds [M-Cp‘Me,] [M = Zr or Hf; Cp‘ = C,Me, or C,H,(SiMe,),]with B(C,F,), gives the toluene complexes 18, whereas thezwitterionic complex 19 is only formed if M = Ti (Scheme12).63 Although it is debatable based on the structuralparameters whether the toluene ligand in 18 is q5 or q6 bound,it is best understood as a 16-electron analogue of [MCp,Me,]and is comparatively ~ n r e a c t i v e .~ ~ Achieving a metallocene-type structure by solvent complexation is evidently preferableto anion co-ordination, a good illustration that the solvent inelectrophilic systems is not necessarily relegated to a spectatorrole.The q2-benzyl ligand in 17 blocks the co-ordination sitenecessary for alkene binding during polymerisation. Neverthe-less, complexes of type 17 are highly active; it is thought thatthey act as a reservoir for the active species, and that the firstinsertion is preceded by an q2-q1 rearrangement.61 Such areversible stabilisation in a way that does not poison thecatalyst (unlike donor ligands) but allows it to participate inequilibria leading to the active species promises to be quite ageneral and useful facet of the chemistry of cationic metallocenealkyls; it helps to prevent the excessively high initial activitythat would be expected if only ‘pure’ active species werepresent, and it may reduce decomposition and deactivationreactions. Neutral metal alkyls fulfil such a stabilising role.The reaction of [ZrCp,Me,] with [CPh,][B(C,F,),] proceedsin stages.The first product formed at low temperature is amethyl-bridged dinuclear species, [(ZrCp,Me),(p-Me)] + (20)(Scheme 13). The stability of this species depends on the Cpligands. Whereas the C,H, complex reacts with further CPh, +in CD,Cl, even at -40 “C, the analogous Me,SiCp’, (Cp’ =indenyl) complex is stable in this solvent up to nearly 20 0C.65The formation of such dinuclear species has also been observedfor cationic titanium 5 6 and thorium 57 complexes.In thesystem [ZrCp,Me,]-B(C,F,), the co-ordination of excess[ZrCp,Me,] to [ZrCp,Me] + competes with the co-ordinationof [BMe(C,F,),]- to give an equilibrium between 16 and 20.66Trimethylaluminium forms even more stable adducts with[MCp’,Me] + , leading to the heterodinuclear complexes[Cp’,M(p-Me),AlMe,]+ (21, M = Zr or Hf).65 The formationof complexes 21 is essentially quantitative; typical examples areshown in Scheme 13 (21a-21d). The complexes are isolable asmicrocrystalline solids or oils. The methyl coupling constantsindicate that the CH, bridge in 20 is most probablyapproximately linear, with a trigonal bipyramidal carbon atom,while the Zr-CH,-A1 angle is acute, as in Al,Me,.The facile formation of 20 and particularly 21 points toanother way, apart from anion co-ordination, in which anactive centre carrying a growing polymer chain can betransformed into a temporarily dormant species, in line withthe ‘intermittent growth’ model (Scheme 3).Ethyl complexes,as models for species with longer polymer chains, formsimilar but less stable dinuclear compounds, as shown in thecase of the hafnium complexes [(HfCpzEt),(p-Et)]+ and[Cp2Hf(p-Et),A1Et2]+ (21e). Evidently the tendency fordinuclear species to dissociate increases with increasing alkylchain length.67It is evident from the mechanistic considerations that neither20 nor 21 possess the vacant co-ordination sites required formonomer binding and chain growth.Nevertheless, the factthat these complexes are extremely active catalysts can beaccommodated by postulating a dissociation equilibrium whichgenerates a common catalytically active species [ZrCp,Me] + ,from both 20 and 21. If so, the productivity of the catalystshould decrease as the concentration of the neutral metal alkylincreases, i.e. as the equilibrium is shifted towards the dinuclearcompounds (Scheme 14). This is indeed observed; for example,the propene polymerisation activity of the catalyst system YLIC-[Zr(Me,SiCp’,)Me,]-[CPh3][B(C6FS),] (Cp’ = indenyl) de-creases with increasing Zr : Ph3C+ ratio, and if AlMe, is added,with increasing A1 : Zr ratio. The effect is quite pronounced at20 “C but still noticeable at polymerisation temperatures of60 “C when dissociation is favoured.The highest productivitiesare obtained at an A1:Zr ratio of 1 : 1 (Fig. 1). The catalystproductivity increases further when AlMe, is replaced byAlEt,, in line with the more facile dissociation of AlEt,a d d u ~ t s . ~ ’ , ~ ’The formation of 21 provides a mechanism for the transferof alkyl chains from the transition metal to aluminium. As aconsequence, high concentrations of aluminium alkyls notonly suppress catalytic activity but also reduce the polymermolecular weight. Under the conditions given in Fig. I , the M ,260 J. Chem. SOC., Dalton Trans., 1996, Pages 2.55-27these sterically very shielded anions are less co-ordinating than[B(C,F,),] - and give higher polymer molecular weights, thereis NMR evidence for the formation of tight ion pairs and anionco-ordination via fluorine contacts.,*There has been considerable speculation in the past aboutthe interaction of MA0 with metallocenes and its mode ofactivation.Equilibria such as 22a-22b have been suggested,and complex 23 was isolated which can be seen as a modelfor the co-ordination of [ZrCp,Me]+ to an A1-0-A1 bondof MA0.69*70 The interaction of solid, essentially tri-methylaluminium-free MA0 with [ZrCp,(' ,CH,),] providesevidence for the formation of a cation-like zirconium methylspecies in MAO-activated catalysts, with a methyl chemicalshift close to that of 16 or [ZrCp,Me]+ in solution.71 Thedata do not, unfortunately, permit the distinction between'unassociated', methyl-bound or oxygen-bound [ZrCp,Me] +.As discussed in the introduction, MA0 exists in solution as amixture of equilibrium species including significant amounts ofAl,Me,, and it seems most likely that under these conditionsheterobinuclear AlMe, adducts of type 21 are the primaryproducts of MA0 a~tivation.,~Whereas the existence of cationic 14-electron metallocenealkyls [MCp,R]+ and their reactivity is by now fairly welldocumented, there is considerably less information about theolefin n: complex that precedes the alkyl migration step (cJ:Scheme 4).In 1974 Tebbe and co-workers described theniobium(m) complex [NbCp,Et(C,H,)] which possesses someof the requirements for a structural model for the n complexbut, as a d2 compound, exhibits extensive back-bonding andshort Nb-C bonds which would not be present in a dozirconium(1v) complex.7 2 Horton's dienyl complex 2473 andthe alkoxy complexes 2S7, give a useful indication of thezirconium-alkene bond distances involved and hence thestrength of this interaction.-1 [MCp2Me2] + CPh,'I 20CPhc2 [MCp2Me] @ 1M21 a, Zr21 b, Hf21 clr21 d24Me21 eScheme 13value of polypropene is more than halved as the Al-Zr ratio isincreased from 1 : 1 to 100: 1 . 6 5Commonly used metallocene catalysts require of courserather high concentrations of aluminium alkyls in order toattain acceptable productivity levels. The aluminium compo-nent plays several important roles: as alkylating agent, as Lewisacid, as a provider of counteranions, as a scavenger and asstabiliser of the active species.As the discussion above shows,appropriate separation of these functions leads to well definedcatalyst systems which require no or very low concentratioris ofaluminium alkyls.The discussion above illustrates that 14-electron cationiccomplexes such as the 'active species' [ZrCp,Me] + are ratherelusive and best prepared in situ. The closest approach so far to'non-stabilised' [ZrCp,R] + was reported recently by Marksand co-workers who showed that the highly hindered anions[B(C,F,SiR,),]- (R, = Pr', or Me2Bu') form isolable saltswith [ZrCp',Me]+ (Cp' = C5H5 or C,H,Me,-l,2). Although25Stereoselective Propene PolymerisationAs Natta discovered, heterogeneous polymerisation catalystspolymerise propene with a very high degree of stereoselectivity.It is thought that polymerisation takes place at the edge ofa TiCI, crystal.The titanium centres are octahedrally co-ordinated; in the active species one of these co-ordination sitesis occupied by the polymer chain, a second by the monomer.Stereoselectivity results because a neighbouring C1 atom (C1* instructure 26) restricts the rotation of the polymer chain. Themonomer then binds to Ti such that steric interaction with thechiral P-carbon of the alkyl chain is minimised (si co-ordinationin 26a) whereas co-ordination of the other n-face (26b) isdisfavoured. The result is highly isotactic polypropene; theregular structure is reflected in a high melting point (165 "C)and high stiffness and hardness.Propene can be polymerised with varying degrees of stereo-control.Atactic polypropene (aPP) has a random orientationof methyl groups along the chain, lacks crystallinity and isusually an oil or wax. In isotactic polypropene (iPP) all methylgroups show the same orientation so that the polymer strandscan align themselves to give crystalline domains. In hemi-J, Chem. SOC., Dalton Trans., 1996, Pages 255-270 26Scheme 14I25- .- k.-I - < 20-Q+-Eav15- - 0a Yp 10-s h .r .-1 5-k1\ *1 10 20 40 60 80 100Al-Zr ratioFig. 1 Propene polymerisation productivity of a [Zr(Me,SiCp',)-Me2]-[CPh3][B(C,F5),]-AIR, (Cp' = indenyl) catalyst as a functionof the Al-Zr ratio, R = Me (A) or Et (m); toluene, 1 bar propene,20 "Cisotactic PP the orientation of every second methyl group israndom, while in syndiotactic PP (sPP) the methyl orientationsalternate (Scheme 15).Of these, only iPP and, to a morelimited extent, aPP have found commercial uses at present.45Isotactic polymer is produced when the monomer moleculesin sucessive insertion steps bind to the metal via the sameenantioface [either re, re or si, si, Scheme 15(a)], while co-ordination to the opposite enantioface in every second insertionstep leads to sPP.Metallocene catalysts can achieve this stereocontrol in twoways. Ewen showed in 1984 that [TiCp,Ph,]-MA0 catalystsproduce iPP at low polymerisation temperatures (- 50 "C) withthe stereochemistry of the insertion step being controlled bythe chiral P-carbon of the growing chain (Scheme 16).75 Thisleads to a polymer in which a stereoerror is perpetuated andnot corrected ('chain end control', 27a).By contrast, a ligandframework with C, symmetry, as in 28, leads to a polymerCp2Zi - O T t - 0 1 AlMezMeCp2Zr-O~Al,e-O~AlMe3 0 022b23where a misinsertion is corrected ('enantiomorphic site control',27b).76The stereoselective propene polymerisation has recently beenreviewed in some and only the most pertinent aspectwill be addressed here. As the example of [TiCpPh,]-MA0showed,75 a rigid ligand framework (e.g. 28) is in principle notnecessary for stereocontrol but indispensable for polymerisationat usual temperatures ( > 50 "C).The hindered rotation ofmonosubstituted Cp ligands provides a measure of stereo-control by imposing C, symmetry on the metallocene. Aninteresting variation of stereocontrol with temperature wasfound for [Ti(C,H,Pr'),Ph,]-MA0 which produces predomi-nantly isotactic polypropene at - 50 O C , nearly atacticpolymer at - 10 "C and polymer with an increased262 J. Chem. SOC., Dalton Trans., 1996, Pages 255-27syndiotactic component at +lO°C, as a result of subtlechanges in the preferred olefin co-ordination over thistemperature range.77 Hindered rotation is also the reasonwhy [ZrCI,Cp',] (Cp' = 1 -methylfluorenyl) gives predomi-nantly isotactic polymer.78 The main attraction of thesecomparatively simple Cp derivatives is the ease of theirsynthesis.However, neither stereoselectivities nor polymermolecular weights can rival the results of heterogeneouscatalysts.79A significant breakthrough was made with C, symmetric,fairly rigid ansa-metallocenes of type 28 and 29.80 The ligandframework imposes an insertion stereochemistry as outlined inScheme 15(a) and gives very highly active catalysts, althoughhere, too, both molecular weight and stereoregularity (andhence physical properties) of the polymer left room forimpr~vement.'~ For some years there was an extensive searchfor new improved ligand systems. A judicious choice ofsubstituents on both the C, and the C, ring of the indenylskeleton proved to be the key to success, e.g. 30 and 31.81*82Restricting the co-ordination sphere of the metal in this way didnot only improve the stereoselectivity but resulted in a dramaticincrease in the polymer molecular weight and catalystproductivity as well.For example, 30 (R = I-naphthyl)matches heterogeneous catalysts in polymer molecular weight(ca. 900 000) and tacticity but it is about 40 times more active.81It appears that 30 and 31 represent something like an endpointin ligand development for stereoselective polymerisation atpresent .ansa-Metallocenes with Cp derivatives of very different sizesand C, symmetry, as in 32, impose a polymerisationstereochemistry as in Scheme 15(6) and give syndiotacticp~lypropene.'~ In fact, this ligand system proved to beremarkably versatile; the introduction of a methyl or tert-butyl substituent on the cyclopentadienyl ring changed thestereocontrol of every second insertion step sufficiently to leadto hemiisotactic and isotactic polymer, respectively (Scheme1 7).84 Elastomeric polypropene with isotactic and atacticblocks have been prepared with 33 and 34.85~86Closer inspection of the stereoselectivity of propenepolymerisation with ansa-metallocenes has shed light on themechanism by which the ligand framework controls thepolymer structure.Surprisingly, the insertion of propene into[Zr(C,H,)Cp',Me] + (Cp' = indenyl) proceeds in a non-stereoselective manner, insertion into a Zr-Et bond, however, isScheme 16J . Chem. Soc., Dulton Truns., 1996, Pugcs 255-270 26highly sele~tive.~’ According to molecular mechanics calcu-lations by Corradini and co-workers.88 the ligand frameworkdoes not influence the enantiofacial orientation of the monomerdirectly but controls the conformation of the alkyl ligand, hencethe effect is only noticeable if the alkyl chain carries at least twocarbons.The incoming monomer then co-ordinates preferen-tially with whichever orientation minimises repulsive van derWaals interactions with the alkyl ligand.Occasionally mis-insertions occur, and a propene moleculeproduces a 2,l- rather than the 1,2-insertion product, leading toa secondary alkyl chain end (Scheme 18). Whenever thishappens any subsequent insertion, either 1,2 or 2, 1, becomesdifficult. The situation is resolved by a (comparatively slow) p-H elimination step which can lead to isomerisation and furtherchain growth with a 1,3-insertion defect, or to the start of a newpolymer chain.As a result, up to 90% of the potentially activezirconium species may be temporarily dormant because of 2, I -misin~ertions.~~The relatively long-lived nature of 2,l -insertion productsmay also explain the apparent paradox that sterically highlycongested complexes such as 30 or 31 not only give polymerswith much better stereoregularity and much higher molecularweight than 28, but are also significantly more active. Evidentlythe substituents help to enforce the correct stereochemistryof monomer co-ordination; the 2-methyl substituents of theindenyl ring in particular guard against a monomer orientationthat would lead to a 2,l-insertion product.Comparison with Lanthanide CatalystsCationic Group 4 complexes of the type [MCp,R]+ areisoelectronic with neutral metallocene complexes of the29 30lanthanide metals, [LnCp,R].The chemistry and reactivity ofboth groups of metals should therefore be quite similar, andthe question arises why the comparatively simple lanthanidesystems, which are mostly isolable and free of complicationssuch as counteranions, should not be used in preference toGroup 4 metals.There are indeed many similarities. In the 197Os, the groupsof Lappert, Ballard and co-workers 90 obtained methyl-bridgedlanthanide-aluminium complexes [Cp,Ln(p-Me),AlMe,] (35)(Ln = Sc, Y, Gd, Dy, Ho, Er, Tm or Y b) which are analogousto the cationic binuclear Zr and Hf complexes 21 reportedrecently.65 These lanthanide complexes are thermally remark-ably stable and may be purified by sublimation.A short timelater Watson showed that [LuCp*,Me] polymerised ethyleneat a high rate but gave a simple insertion product with propene,[LuCp,*(CH,CHMe,)], which reacted with hydrogen underhydrogenolysis to yield [LuHCp,*], all steps of relevance toZiegler catalysis. The hydrogenolysis of isolable lanthanidealkyls was employed effectively by Marks and co-workers inorder to convert the inactive alkyls [LnCp,*(CH(SiMe,),)](Ln = La, Nd or Lu) into the hydrides [LnHCp*,] which arehighly active, if short lived, ethylene polymerisation catalyst^.'^Similar neodymium catalysts have been used to polymeriseethylene under very high pressure and temperatures in tubereactors (200 O C , 1200 bar) where the short catalyst lifetimespose no di~advantage.~,As the early studies showed, a major difference betweenlanthanide and Group 4 metal catalysts is the failure of theformer to polymerise alk- 1 -enes.There are also structuraldifferences. Recent calculations showed that the ground-stategeometry of [ZrCp,Me] + is pyramidal (36) while complexes oftrivalent metals such as scandium and lanthanides are trigonalplanar (37),28 i.e. the geometry of 36 is better adapted for olefinco-ordination. In contrast to the somewhat elusive cation 36many complexes of type [LnCp’,Me] are isolable; they formp-CH3 bridged dimers easily, provided the steric bulk of Cp’allows it.’4This propensity to form stable dimers is one feature thatdistinguishes lanthanide complexes from the Group 4 cations.Although dimeric hydrides [( LnHCp*,),] are excellenthydrogenation catalyst^,'^ it could be shown that a p-H ligandreacts with olefins 108-10’0 times more slowly than a terminalhydride.For example, 38 reacts with alk-I-enes only slowly togive remarkably stable p-alkyl complexes of type 39; there is noMe GI231 33 34Scheme 17264 J . Cltem. Soc., Dulron Truns., 1996, Puges 255-27X8,G+new chain growthZ r w - fast growing chaineliminationp-H I3536I1 IMe : +1 2slowXII I37chain with 1 ,SinsertionScheme 18ILn = Y or Lu/39Scheme 19Me23840further reaction, and even the hydrogenolysis of 39 is ratherslow (Scheme 19).95 The closely related aryloxide complex[ {Y(p-H)Cp*(OC6H3Bu',-2,6)},] shows similar behaviour.Bridging hydrides are evidently preferred over p-aryloxoligands, and although ethylene is polymerised rapidly, propeneinserts only once to give a [Y,(p-H)(p-CH,CH,Me)] complexwhich is stable to P-H e l i m i n a t i ~ n .~ ~ The ability to form stablep-alkyl compounds must inevitably lead to catalyst deactiv-ation. Nevertheless, the dimeric hydride 40 does polymerisealk- 1 -enes from propene to hexene to give polymers of moderatemolecular weight, albeit over a period of several days.97Another distinguishing feature of lanthanide chemistry is thepropensity to undergo C-H activation rather than insertion ('0-bond metathesis').Thus [ScCp*,Me] exchanges methyl groupswith methane, reacts with arenes to give metal aryls, andactivates styrenes to give vinyl complexes. One molecule ofpropene is inserted, a second molecule reacts with o-bondmetathesis (Scheme 20).98*99 None of these reactions have sofar been demonstrated for Group 4 metal alkyls.Catalyst Deactivation ReactionsMetallocene dihalide-MA0 catalysts show a more or lessrapid exponential decay in activity in the initial phase ofpolymerisation before stabilising at a lower level. With cationicmetal alkyl catalysts the initial activities are usually very high,with an even more rapid decay. Catalyst deactivation is clearlyan important consideration for any commercial use of thesesystems.However, there is so far very little information aboutpossible deactivation pathways.In early work Kaminsky et a1.l" succeeded in isolating anumber of ethylene-bridged complexes from the reaction ofJ. Chem. SOC., Dalton Trans., 1996, Pages 255-270 26[ZrCp,Et(Cl)] with AlEt, which were thought to be side-products, e.g. 41 and 42.Fischer and Miilhaupt determined the kinetics of catalystdeactivation for the atactic polymerisation of propene with[ZrCl,Cp,]-MA0 catalysts. O 1 A reversible and an irreversibledeactivation step could be distinguished, both second order in[Zr]. Although it is not possible to derive firm conclusionsabout the nature of the deactivated species from suchmeasurements, it is tempting to think that the reversibledeactivation step consists of the formation of alkyl-bridgeddinuclear zirconium complexes of type 20.Other potential deactivation pathways are the formation ofdinuclear species by a- or P-CH activation.The chemistry ofcationic Group 4 alkyls provides examples of both. Thefuhalene complex 43 reacts with CPh,' or B(C6F5), even at-60 "C with immediate elimination of methane to give aScheme 20,AlEt341 42Me,Me-Zr I Q-e I43relatively inactive p-CH, complex 44,'" and the hafnium ethylcomplex 45 decomposes slowly with elimination of ethane togive 46 (Scheme 21), a compound reminiscent of 42.64 Theneutral dicarbollide complex [(ZrCp*(C2B,Hl l)Me},] alsodecomposes via a-H elimination to a stable p-CH, product,albeit at a much higher temperature than 43 (45 "C, 2 h).'',Cationic complexes containing [B(C6F5),] - and relatedcounteranions may be expected to be prone to deactivation byaryl or fluoride transfer, although there is at present no evidencefor this under catalytic conditions.Such reactions do, however,occur slowly at room temperature and above. For example,the reaction of 47 with B(C,F,), in benzene over 24 h gave48, (Scheme 22) lo4 and similarly 16 (Scheme 11) was foundto slowly form the fluoride complex [(ZrCp',Me),(p-F)]-[MeB(C,F,),]. 59bAlternative Catalysts and Ligand SystemsAs the work described above has documented, cyclopentadienylcomplexes possess almost ideal catalytic properties. Neverthe-less, increasing efforts are being made exploring the potential ofnew ligand environments and catalyst types.The aim is twofold:first, to find a ligand system that can rival or even better theextensively patented Cp compounds, and second, to eliminatethe complications of a counteranion and to develop neutralanalogues of cationic complexes.Chelating cyclopentadienyl-amide ligands as in 49 and 50have proved very successful in a range of polymerisations andco-polymerisations. '05*1 O6 Benzamidinato complexes such as47 catalyse the polymerisation of ethylene but are less activetowards propene.'" A more radical departure from thecyclopentadienyl theme are macrocycles such as 51 and 52.Although the zirconocene dialkyls with these ligands have ageometry closely similar to cyclopentadienyl complexes andcationic alkyl derivatives can be made, the catalytic activityof these complexes are at best modest.lo8 The Schiff-basecomplexes 53, too, are poor polymerisation catalysts. lo9Tridentate amide complexes have also been prepared, e.g.54."' Bidentate aryloxides afford a series of metal halide andalkyl complexes, e.g. 55 and 56. In the presence of MA0 some ofthese complexes are moderately active as ethylene polymeris-ation catalysts. Unlike metallocene catalysts these complexesgive polymers with a rather broad molecular weightj'HH44L45 R=SiMe3 46Scheme 21266 J. Chem. Soc., Dalton Trans., 1996, Pages 255-27distribution (M,/M, = 7-24), and it is not certain whether theidentity of the complexes remains intact under polymerisationconditions.' 'An alternative method to the generation of cationic catalystsfrom metallocene dialkyls has recently been developed by Erkerand co-workers who prepared zwitterionic alkyl complexesby attack of B(C6F5) on zirconium(r1) and hafnium(r1) dienecomplexes (Scheme 23). They catalyse the polymerisation ofethylene and propene. ' ' Marks and co-workers, used a similarstrategy to convert titanium-diene complexes of type 49 (X, =diene) into highly active catalysts for ethylene-oct- 1-enecopolymerisations. ' 'The replacement of a Cp ligand in the cation [MCp,R] + by adianionic ligand (L) generates an electroneutral analogue,[MCp(L)R], a potentially useful strategy in order to eliminatethe counteranion influence and to evaluate the role of thepositive charge on the metal.The first attempts in this directionwere made by Jordan and co-workers who used the [C,-B,H carboranyl dianion. The compound [{ZrCp*(C,B,-H,,)Me),,] is a polymer which shows moderate activity forethylene polymerisation and oligomerises propene. ' O 3 Thethermally sensitive titanium derivative is even less reactive anddimerises ethylene very slowly to .butene.' l 4 Other dianionicPh4748Scheme 22Me3C4953 M =TiorZr\ /ZrX2MepSiNIMe&50R/d CH(SiMe,),54 R=SiMe3ligands, such as trimethylenemethane and borolene [C4H4-BXI2- (X =NPri2), afford the anionic complexes 57 and 58,respectively. With MA0 as an activator 57 polymerisesethylene. In neither case have the neutral alkyl complexes[ZrCp*(L)R] been isolated.' 5,1 'In both 57 and 58 the lithium countercation co-ordinatesstrongly to the chloride ligands, no doubt as the result of theaccumulation of electron density on the metal due to thepresence of doubly negatively charged IT ligands. Such areduction in the Lewis acidity of the metal centre is unlikelyto be helpful to catalysis. An alternative is the use ofcyclopentadienyl ligands with negatively charged substituents,Cp--Z-. A series of complexes with Z = B(C,F,), has beenprepared according to Scheme 24 in which the counteranion isassociated with the borato substituent and not co-ordinated tohalide. l 7 The dihalides 59 in the presence of MA0 are highlyactive and produce high-rnolecular-weight polyethylene, as docatalysts generated from the dimethyls 60, [CPh,][B(C,F,),]and AlBu',.The reaction of 60 with CPh,' in the presence ofAlMe, affords the highly hydrocarbon-soluble zwitterioniccomplex 61. Systems of this kind offer promise as a route tosingle component polymerisation catalysts which can beemployed without the need for activators.57 5851I t -FScheme 235256 55J. Chem. SOC., Dalton Trans., 1996, Pages 255-270 26..6059X = C6F5R = H or SiMe,R61Scheme 24Fig. 2 Comparison of the properties of isotactic polypropene madewith a heterogeneous catalyst (0) with the properties of (a)metallocene-derived isotactic polypropene (O), and (b) syndiotacticpolypropene ( H 1OutlookThe last decade has seen rapid progress in the understanding ofmetallocene-catalysed olefin polymerisations, and the mechan-istic principles are now widely accepted.The emergence ofmetallocenes from the 'black-box era' of ill-defined catalystbrews was made possible by the rational application offundamental concepts of organometallic reaction pathways andbonding modes, such as agostic interactions. Ligand design andreaction conditions can now be tailored to give a predeterminedstereochemistry and, to a more limited extent, a particularactivity or polymer molecular weight.One of the main advantages of metallocene catalysts is theirversatility. Polymer properties can be varied within wide limitsto provide not just substitutes but alternatives to currentlyavailable polymers made with heterogeneous catalysts.I t is thecombination of various properties that is of interest to thepolymer user, and these can vary widely for structurally closelyrelated polymers made by different routes, as showndiagramatically in Fig. 2.46The discussion above has concentrated primarily on thepolymerisation of ethylene and propene. Metallocenes andrelated complexes are of course applied much more widely, suchas in ethylene-alk- 1 -ene copolymerisations, the (syndiotactic)polymerisation of styrene, the cyclic polymerisation of 1,5-dienes, the polymerisation and block-copolymerisation ofmethacrylates, and in the copolymerisation of ethylene withnorbornene to give highly transparent polymers with interestingspeciality optical and medical applications.Although the control of stereoselectivity in propenepolymerisations has been a major focus of interest in the pastI0 years, not least in industrial research laboratories, thesecomplexes with their quite sophisticated ligand environmentsdo not yet appear to have been introduced in large-scaleprocesses (however, there is optimism 3).Syndiotactic polypro-pene, too, has not yet found its niche. It is in non-stereoselectivecopolymerisations that metallocenes proved their worth mostand found applications most readily, the major advantage ofeven simple CSH, complexes being their ability to uniformlyincorporate higher alk- 1 -enes into a polyethylene chain, inde-pendent of the polymer molecular weight. These catalysts canbe supported on MAO-coated silica (thereby reducing theAl-Zr ratio to a tolerable 200-300: 1) and retrofitted into exist-ing plants.Tens of thousands of tons of metallocene-derivedlinear low-density polyethylene (LLDPE) are already beingproduced in this way. The industrial interest is reflected in thenumber of patent applications for metallocene catalysts whichhas risen worldwide from only two in 1982 to over 600 in1 993.46 The chemistry of catalyst-support interactions in theseheterogenised metallocene catalysts is not yet well understood.No doubt this will be an important focus of interest in years tocome.References1 K. Ziegler, E. Holzkamp, H. Breil and H. Martin, Angew. Chem.,2 G. Natta, Angew. Chem., 1956,68, 393; 1964, 76, 553.3 Nachr.Chem. Tech. Lab., 1995,43,823; see also A. N. Thayer, Chem.Eng. News, 1995, 73, 15.4 P. D. Gavens, M. Bottril. J. W. Kelland and J. McMeeking, inComprehensive Organometallic Chemistry, eds. G. Wil kinson,F. G. A. Stone and E. W. Abel, Pergamon, Oxford, 1982, vol. 3.5 G. Wilkinson, P. L. Pauson, J. M. Birmingham and F. A. Cotton,J. Am. Chem. Soc., 1953,75, 101 1.6 D. S. Breslow, US Par., 537039, 1955; D. S. Breslow andN. R. Newburg, J . Am. Chem. Soc., 1957,79, 5072.7 G. Natta, P. Pino, G. Mazzanti, U. Giannini, E. Mantica andM. Peraldo, Chim. Ind. (Paris), 1957, 39, 19; G. Natta, P. Pino,G. Mazzanti and U. Giannini, J. Am. Chem. Soc., 1957,79, 2975.8 K. H. Reichert and K. R. Meyer, Makromol. Chem., 1973,169, 163.9 W. P. Long and D.S. Breslow, Liebigs Ann. Chem., 1975,463.10 A. A. Andersen, H. G. Cordes, J. Herwig, W. Kaminsky, A. Merck,R. Mottweiler. J . Pein, H. Sinn and H. J. Vollmer, Angew. Chem.,1976,88,689; Angew. Chem., Int. Ed. Engl., 1976, 15,630.1955,67, 541; K. Ziegler, Angew. Chem., 1964,76, 545.p. 475.1 1 H. Sinn and W. Kaminsky, Adv. Organomet. Chem., 1980,18,99.12 L. Resconi and S. Bossi, Macromolecules, 1990,23,4489.13 D. S. Breslow and N. R. Newburg, J. Am. Chem. Sue., 1959,81,81.14 P. Cossee, J. Catal., 1964,3,80; E. J. Arlman, J. Card., 1964,3, 89;E. J. Arlman and P. Cossee, J. Catal., 1964,3, 99.15 G. Natta and G. Mazzanti, Tetrahedron, 1960, 8, 86; H. Sinn andF. Patat, Angew. Chem., 1963,75,805; G. Henrici-Olive and S. Olive,Angew. Chem., 1967,79, 805; Angew.Chem., Int. Ed. Engl., 1967,6.790.16 G. Fink and R. Rottler, Makromol. Chem., 1981, 94, 25; G. Fink,R. Rottler and C. G. Kreiter, Makromol. Chem., 1981,%, 1; G. Fink,W. Fenzl and R. Mynott, Z. Naturforsch., Teil B, 1985, 40, 158;R. Mynott, G. Fink and W. Fenzl, Angew. Makromol. Chem., 1987,154,l.17 A. K. Zefirova and A. E. Shilov, Dokl. Akud. Nauk. SSSR, 1961,136, 599; F. S. Dyachkovskii, A. K. Shilova and A. E. Shilov,J . Polym. Sci., Part C, 1967, 16,2333.18 F. S. Dyachkovskii, in Coordination Polymerisation, ed. J. C. W.Chien, Academic Press, New York, 1975, p. 199.19 J. J. Eisch, A. M. Piotrowski, S. K. Brownstein, E. J. Gabe andF. L. Lee, J . Am. Chem. Soc., 1985,107, 7219.20 M. Bochmann and L. M. Wilson, J .Chem. Soc., Chem. Commun.,1986, 1610; M. Bochmann, L. M. Wilson, M. B. Hursthouse andR. L. Short, Organornetallics, 1987,6, 2556.21 M. Bochmann, L. M. Wilson, M. B. Hursthouse and M. Motevalli,Organometallics, 1988, 7, 1 148.268 J . Chrm. Soc., Dalton Trans., 1996, Pages 255-2722 R. F. Jordan, C. S. Bajgur, R. Willet and B. Scott, J. Am. Chem. Soc.,23 R. F. Jordan, R. E. LaPointe, C. S. Bajgur, S. F. Echols and24 Review: R. F. Jordan, Adv. Organomet. Chem., 1991,32,325.25 R. Taube and L. Krukowka, J. Organomet. Chem., 1988,347, C9.26 J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc., 1976,98, 1729.27 C. A. Jolly and D. S. Marynick, J. Am. Chem. Soc., 1989,111,7968.28 T. K. Woo, L. Fan and T. Ziegler, Organometallics, 1994, 13, 432and 2252.29 M.Brookhart and M. L. H. Green, J. Organornet. Chem., 1983,250,395; M. Brookhart, M. L. H. Green and L. Wong, Prog. Inorg.Chem., 1988,36, 1.30 H. Weiss, M. Ehrig and R. Ahlrichs, J. Am. Chem. Soc., 1994, 116,4919.31 H. Kawamura-Kuribayashi, N. Koga and K. Morokuma, J. Am.Chem. Soc., 1992, 114, 8687; T. Yoshida, N. Koga and K.Morokuma, Organometallics, 1995, 14, 746.32 G. Natta and I. Pasquon, Adv. Catal., 1959, 11, 1; J. C. W. Chien,J. Am. Chem. Soc., 1959, 81, 86; J. P. Machon, R. Hermant andJ. P. Houteaux, J. Polym. Sci., Polym. Symp., 1975,52, 107.33 M. H. Prosenc, C. Janiak and H. H. Brintzinger, Organometallics,1992,11,4036.34 R. J. Meier, G. H. J. van Doremaele, S. Iarlori and F. Buda, J. Am.Chem. Soc., 1994,116,7274.35 H.Krauledat and H. H. Brintzinger, Angew. Chem., 1990, 102,1459; Angew. Chem., Int. Ed. Engl., 1990, 29, 1412; M. Leclerc andH. H. Brintzinger, J. Am. Chem. Soc., 1995, 117, 1651.36 Review, H. H. Brintzinger, D. Fischer, R. Miilhaupt, B. Rieger andR. Waymouth, Angew. Chem., 1995, 107, 1255; Angew. Chem., Int.Ed. Engl., 1995,34, 1143.1986,108,7410.R. Willet, J. Am. Chem. Soc., 1987, 109,4111.37 W. E. Piers and J. Bercaw, J. Am. Chem. Soc., 1990,112,9406.38 J. J. W. Eshuis, Y. Y. Tan, A. Meetsma and J. H. Teuben,Organometallics, 1992, 11, 362; L. Resconi, F. Piemontesi, G.Franciscono, L. Abis and T. Fiorani, J. Am. Chem. Soc., 1992,114,1025.39 G. Sini, S. A. Macgregor, 0. Eisenstein and J. H. Teuben,Organometallics, 1994, 13, 1049.40 G.G. Hlatky, H. W. Turner and R. R. Eckman, J. Am. Chem. Soc.,1989,111,2728.41 M. Bochmann, A. J. Jaggar and J. C. Nicholls, Angew. Chem., 1990,102, 830; Angew. Chem., Int. Ed. Engl., 1990,29, 780.42 M. Bochmann and A. J. Jaggar, J. Organomet. Chem., 1992, 424,c5.43 M. Bochmann, A. J. Jaggar, L. M. Wilson, M. B. Hursthouse andM. Motevalli, Polyhedron, 1989, 8, 1838.44 J. J. W. Eshuis, Y. Y. Tan and J. H. Teuben, J. Mol. Catal., 1990,62,277; J. J. W. Eshuis, Y. Y. Tan, A. Meetsma, J. H. Teuben,J. Renkema and G. G. Evans, Organometallics, 1992, 11, 362,45 A. D. Horton, Trends Polym. Sci. (Cambridge, UK), 1994,2, 158.46 M. Aulbach and F. Kuber, Chem. Unserer Zeit, 1994,28, 197.47 M. Bochmann, Angew. Chem., 1992,104, 1206; Angew. Chem., Int.48 A.D. Horton and J. H. G. Frijns, Angew. Chem., 1991, 103, 1181;49 M. Bochmann, G. Karger and A. J. Jagger, J. Chem. Soc., Chem.50 C. Pellecchia, A. Immirzi, A. Grassi and A. Zambelli, Organo-51 A. D. Horton and A. G. Orpen, Organometallics, 1991,10, 3910.52 G. G. Hlatky, R. R. Eckman and H. W. Turner, Organometallics,1992, 11, 1413; H. W. Turner and G. G. Hlatky, Eur. Pat. Applic.,277003, 1988; Chem. Abstr., 1989,110,5829 1 b.53 (a) D. J. Crowther, N. C. Baenziger and R. F. Jordan, J. Am.Chem. Soc., 1991, 113, 1455; (b) C. Kreuder, R. F. Jordan andH. Zhang, Organometallics, 1995, 14, 2993; (c) D. J. Crowther,S. L. Borkowsky, D. Swenson, T. Y. Meyer and R. F. Jordan,Organometallics, 1993, 12, 2897.54 H. W. Turner, Eur. Pat. Applic., 277004, 1988; Chem.Abstr., 1989,110,58290a; A. G. Massey and A. Park, J. Organomet. Chem., 1964,2,245.55 J. A. Ewen and M. J. Elder, Eur. Pat. Applic., 426637 and 426638,1991; Chem. Abstr., 1991,115,136987cand 136988d; J. C. W. Chien,W. M. Tsai and M. D. Rausch, J. Am. Chem. Soc., 1991,113,8570.56 M. Bochmann and S. J. Lancaster, J. Organomet. Chem., 1992,434,C1.57 X. Yang, C. L. Stern and T. J. Marks, Organometallics, 1991, 10,840.58 X. Yang, C. L. Stern and T. J. Marks, Angew. Chem., 1992, 104,1406; Angew. Chem., Int. Ed. Engl., 1992,31, 1375.59 (a) X. Yang, C. L. Stern and T. J. Marks, J. Am. Chem. Soc., 1991,113, 3623; (b) 1994, 116, 10 015,Ed. Engl., 1992,31, 1181.Angew. Chem., Int. Ed. Engl., 1991,30, 1152.Commun., 1990, 1038.metallics, 1993, 12,4473.60 M.Bochmann, S. J. Lancaster, M. B. Hursthouse and K. M. A.61 M. Bochmann and S. J. Lancaster, Organometallics, 1993, 12, 633.62 M. Bochmann and S. J. Lancaster, Makromol. Chem., RapidCommun., 1993, 14, 807.63 D. J. Gillis, M. J. Tudoret and M. C. Baird, J. Am. Chem. Soc., 1993,115,2543.64 M. Bochmann, 0. B. Robinson, S. J. Lancaster, M. B. Hursthouseand S. J. Coles, Organometallics, 1995, 14, 2456.65 M. Bochmann and S. J. Lancaster, Angew. Chem., 1994,106, 1715;Angew. Chem., Int. Ed. Engl., 1994,33, 1634.66 T. Haselwander, S. Beck and H. H. Brintzinger, in Ziegler Catalysts,eds. R. Miilhaupt, G. Fink and H. H. Brintzinger, Springer, Berlin,67 M. Bochmann and S. J. Lancaster, J. Organomet. Chem., 1995,497,68 L.Lia, X. Yang, A. Ishihara and T. J. Marks, Organometallics, 1995,69 E. Giannetti, G. M. Nicoletti and R. Mazzocchi, J. Polym. Sci., Part70 G. Erker, M. Albrecht, S. Werner and C. Kriiger, Z. Naturforsch,71 C. Sishta, R. M. Hathorn and T. J. Marks, J. Am. Chem. Soc., 1992,72 L. J. Guggenberger, P. Meakin and F. Tebbe, J. Am. Chem. Soc.,73 A. D. Horton and A. G. Orpen, Organometallics, 1992,11,8.74 Z. Wu, R. F. Jordan and J. L. Peterson, J. Am. Chem. Soc., 1995,75 J. A. Ewen, J. Am. Chem. Soc., 1984,106,6355.76 W. Kaminsky, K. Kulper, H. H. Brintzinger and F. R. W. P. Wild,Angew. Chem., 1985,97,507; Angew. Chem., Int. Ed. Engl., 1985,24,507.77 G. Erker and C. Fritze, Angew. Chem., 1992, 104, 204; Angew.Chem., Int. Ed. Engl., 1992,31, 199.78 A.Razavi and J. L. Atwood, J. Am. Chem. Soc., 1993,115,7529.79 G. Erker, R. Nolte, R. Aul, S. Wilker, C. Kriiger and R. Noe, J. Am.Chem. Soc., 1991,113,7574.80 H. Schutenhaus and H. H. Brintzinger, Angew. Chem., 1979, 91,837; Angew. Chem., Int. Ed. Engl., 1979, 18, 777; F. R. W. P. Wild,L. Zsolnai, G. Huttner and H. H. Brintzinger, J. Organomet. Chem.,1982,23,233.81 W. Spalek, F. Kiiber, A. Winter, J. Rohrmann, B. Bachmann,M. Antberg, V. Dolle and E. F. Paulus, Organometallics, 1994, 13,954.82 U. Stehling, J. Diebold, R. Kirsten, W. Roll, H. H. Brintzinger,S. Jiingling, R. Miilhaupt and F. Langhauser, Organometallics,1994, 13, 964.83 J. A. Ewen, R. L. Jones, A. Razavi and J. D. Ferrera, J. Am. Chem.Soc., 1988, 110, 6255.84 J.A. Ewen, M. J. Elder, R. L. Jones, L. Haspeslagh, J. L. Atwood,S. G. Bott and K. Robinson, Makromol, Chem., Macromol. Symp.,199 1,48149,253.85 G. H. Llinas, S. H. Dong, D. T. Mallin, M. D. Rausch, Y. G.Lin, H. H. Winter and J. C. W. Chien, Macromolecules, 1992, 25,1242.Malik, Organometallics, 1994, 13, 2235.1995, p. 181.55.14, 3135.A, 1985,23,2117.Teil B, 1990,45, 1205.114, 1 1 12.1974,96, 5420.117,5867.86 R. M. Waymouth and G. W. Coates, Science, 1995,267,222.87 P. Longo, A. Grassi, C. Pellecchia and A. Zambelli, Macromolecules,1987, 20, 1015; A. Zambelli, C. Pellecchia and L. Oliva, Makromol.Chem., Macromol. Symp., 1991,48149, 297.88 L. Cavallo, G. Guerra, L. Oliva, M. Vacatello and P. Corradini,Polym. Chem., 1989,30, 17; L.Cavallo, G. Guerra, M. Vacatello andP. Corradini, Macromolecules, 1991,24, 1784.89 P. Corradini, V. Busico and R. Cipullo, Makromol. Chem., RapidCommun., 1992, 13, 21; V. Busico, R. Cipullo and P. Corradini,Makromol. Chem., Rapid Commun., 1993,14,97.90 D. G. H. Ballard and R. Pearce, J. Chem. Soc., Chem. Commun.,1975, 621; J. Holton, M. F. Lappert, G. R. Scollary, D. G. H.Ballard, R. Pearce, J. L. Atwood and W. E. Hunter, J. Chem. Soc.,Chem. Commun., 1976, 425; J. Holton, M. F. Lappert, D. G. H.Ballard, R. Pearce, J. L. Atwood and W. E. Hunter, J. Chem. Soc.,Dalton Trans., 1979,45.91 P. L. Watson, J. Am. Chem. Soc., 1982,104, 337; P. L. Watson andT. Herskovitz, ACSSymp. Ser., 1983, 212, 459.92 H. Mauermann, P. N. Swepston and T. J. Marks, Organometallics,1985,4, 200; G. Jeske, H. Lauke, H. Mauermann, P. N. Swepston,H. Schumann and T. J. Marks, J. Am. Chem. Soc., 1985,107,8091.93 X. Olonde, A. Mortreux, F. Petit and K. Bujadoux, J. Mol. Catal.,1993,82, 75.94 T. J. Marks and R. D. Ernst, in Comprehensive OrganometallicChemistry, G. Wikinson, F. G. A. Stone and E. W. Abel, Pergamon,Oxford, 1982, vol. 3, p. 202.J. Chem. SOC., Dalton Trans., 1996, Pages 255-270 2695 D. Stern, M. Sabat and T. J. Marks, J. Am. Chem. SOC., 1990,112,96 C. J. Schaverien, J. Chem. SOC., Chem. Commun., 1992, 11.97 E. B. Coughlin and J. E. Bercaw, J. Am. Chem. SOC., 1992, 114,98 P. L. Watson and G . W. Parshall, Acc. Chem. Res., 1985, 18, 51.99 M. E. Thompson, S. M. Baxter, A. R. Bull, B. J. Burger,M. C . Nolan, B. D. Santarsiero, W. P. Schaefer and J. E. Bercaw,J. Am. Chem. SOC., 1987,109,203.100 W. Kaminsky and H. J. Vollmer, Liebigs. Ann. Chem., 1975, 438;W. Kaminsky, J. Kopf, H. Sinn and H. J. Vollmer, Angew. Chem.,1976,88, 688; Angew. Chem., Int. Ed, Engl., 1976, 15, 629.101 D. Fischer and R. Mulhaupt, J. Organomet. Chem., 1991, 417,c7.102 M. Bochmann, T. Cuenca and D. T. Hardy, J. Organomet. Chem.,1994,484, CIO.103 D. J. Crowther, N. C. Baenziger and R. F. Jordan, J. Am. Chem.SOC., 1991, 113, 1455.104 R. Gomez, M. L. H. Green and J. L. Haggitt, J. Chem. SOC., Chem.Commun., 1994,2607.105 J. A. M. Canich, Eur. Pat. Applic., EP 420436, 1990; J. C. Stevensand D. R. Neithammer, Eur. Pat. Applic., EP 418044, 1990; J. C.Stevens, F. J. Timmers, D. R. Wilson, G. F. Schmidt, P. N. Nicklas,R. K. Rosen, G. W. Knight and S. Y. Lai, Eur. Pat. Applic., EP416815, 1990.106 J. A. M. Canich, US Pat. Applic., 5,026,798 and 5,055,438, 1993;H. W. Turner, G. G. Hlatky and J. A. M. Canich, PCT Int. Appl.,WO 9319103, 1993.107 A. N. Chernega, R. Gomez and M. L. H. Green, J. Chem. SOC.,Chem. Commun., 1993, 1415.9588.7606.108 R. Uhrhammer, D. G. Black, T. G. Gardner, J. D. Olsen andR. F. Jordan, J. Am. Chem. SOC., 1993, 115, 8493; D. G. Black,D. C. Swenson, R. F. Jordan and R. D. Rogers, Organometallics,1995,14,3539.109 E. B. Tjaden, D. C . Swenson, R. F. Jordan and J. L. Peterson,Organometallics, 1995, 14, 371.110 F. G. N. Cloke, P. B. Hitchcock and J. B. Love, J. Chem. SOC.,Dalton Trans., 1995, 25.11 1 A. van der Lindeh, C . J. Schaverien, N. Meijboom, C. Ganter andA. G. Orpen, J. Am. Chem. Soc., 1995,117,3008.112 B. Temme, G. Erker, J. Karl, H. Luftmann, R. Frohlich andS. Kotila, Angew. Chem., 1995, 107, 1867; Angew. Chem., Int. Ed,Engl., 1995,34, 1755.113 D. D. Devore, F. J. Timmers, D. L. Hasha, R. K. Rosen,T. J. Marks, P. A. Deck and C. L. Stern, Organometallics, 1995,14, 3132.114 C. Kreuder, R. F. Jordan and H. Zhang, Organometallics, 1995,14,2993.11 5 G. C. Bazan, G. Rodriguez and B. P. Cleary, J. Am. Chem. SOC.,1994,116,2177.116 R. W. Quan, G. C. Bazan, A. F. Kiely, W. P. Schaefer andJ. E. Bercaw, J. Am. Chem. SOC., 1994,116,4489.1 17 M. Bochmann, S. J. Lancaster and 0. B. Robinson, J. Chem. SOC.,Chem. Commun., 1995,208 1.Received 1 1 th October 1995; Paper 5/06738C270 J. Chem. SOC., Dalton Trans., 1996, Pages 255-27
ISSN:1477-9226
DOI:10.1039/DT9960000255
出版商:RSC
年代:1996
数据来源: RSC