An Odyssey from Stoichiometric Carbotitanation of Alkynes to Zirconium- catalysed Enantioselective Carboalumination of Alkenes Ei-ichi Negishi and Denis Y. Kondakov Department of Chemistry, Purdue University, West La fayette, Indiana 47907, USA 1 Introduction Until recently, the use of organometals in organic synthesis had been dominated by polar reactions of organometals, such as organolithiums and Grignard reagents, with polar electrophiles, such as alkyl halides, ketones and other carbonyl compounds, as well as nitriles. Although carbon-carbon bond formation via organometallic reactions of nonpolar compounds, such as oligom- erization and polymerization of alkenes, alkynes and dienes, has been known for several decades, most of the early examples were limited to the synthesis of highly symmetrical molecules, such as benzene, cyclododecatriene and polyethylene.* As such, these reac- tions and procedures were not readily applicable to the synthesis of complex organic molecules of low symmetry.Nonetheless, addition of carbon-metal bonds to alkenes and alkynes, termed carbometullation,3 may, in principle, be achieved in a controlled manner so that it would be applicable to the synthesis of unsym-metrical molecules. Carbometallations may proceed by various mechanisms, but those proceeding via pericyclic reactions are of particular interest to us, because they can be facile and highly stereoselective. For the crucial step of such processes, a four- centred syn addition mechanism represented by Scheme 1 may be proposed.If this mechanism indeed operates, the crucial structural requirement for organometallic reagents is the presence or ready LUMO C-M HOMO CaMa 0 00 00 HOMO LUMO Scheme 1 Ei-ichi Negishi received the bachelor S degree from the University of Tokyo in 1958. While he was a research chemist at Teijin, Ltd, Japan, he came to the University of Pennsylvania as a Fulbright Scholar in 1960 and received his PhD degree in 1963. He joined Professor H. C. Brown's research group at Purdue University as a postdoctoral associate in 1966 and became his assistant in 1968. In 1972 he moved to Syracuse University as Assistant Professor and was promoted to Associate Professor in 1976. He returned to Purdue University as Professor in 1979.He is the author of about 250 scientijc publica- tions. His recent work has centred on the use of transition-metal complexes as Catalytic reagents in organic synthe-sis. Some transition metal-catalysed reactions developed by him and his students in-clude Pd-or Ni-catalysed cross-coupling, Pd-catalysed cyclic carbopalladation reac-tions, and Zr-or Ti-catalysed carbometallation reactions. availability of a low-lying metal empty orbital. Since one can write essentially the same mechanism for hydrometallation by merely replacing C with H, we reasoned that those metals which readily participate in hydrometallation, such as B, A1 and Zr, should also participate in carbometallation but that the activation energy for carbometallation would be generally higher than that for the corre- sponding hydrometallation reaction due to the greater steric require- ments of C groups relative to H and more highly directionalized sptz-hybridized C orbitals as compared with the non-directional s orbital of H.We further reasoned that one way of promoting such carbometallation processes might be to resort to dynamic polariza- tiorz between two Lewis acids (or electrophilesj which makes one of them more acidic (or electrophilic), while making the other more basic (or nucleophilicj. As generally accepted, this might indeed be the mode of activation in a wide variety of Lewis acid-catalysed reactions, such as the Ziegler-Natta polymerization4 and the Friedel-Crafts rea~tion.~ As discussed by us some 15 years interactions between two metal-containing Lewis acids 'MIL and *M2L can lead to (i) 'ate' complexation, (ii)dynamic polarization and (iii) transmetallation among others (Scheme 2j, and some of these processes can serve as crucial steps in catalytic cycles.With these simplistic notions in mind, we embarked on a long-range investigation of developing carbometallation reactions of B, Al, Zr and other metals promoted or catalysed by Lewis acids, such as those containing B, Al, Ti, Zr and other metals. 2 Stoichiometric Carbotitanation of Alkynes vs. Formation of Tebbe Reagent and its Reaction with Alkynes With the development of regio- and stereo-selective methods for converting alkynes into tri- and tetra-substituted alkenyl derivatives as one of the major goals, the reaction of alkynes with Al-Ti reagents was considered.Treatment of terminal alkynes with organoalanes was known to give mainly alkynylalanes via terminal H abstraction, and the same reaction of internal alkynes was known to require rather drastic conditions leading to the formation of oligomeric products.6In sharp contrast, the reaction of diphenylacetylene with 2 equiv. of Me,AI and 1 equiv. of Cp,TiCl,, where Cp is v5-cyclopentadienyl, was complete within 12 h at 20-22 "C to give, Denis Y. Kondakov was born in St. Petersburg, Russia, in 1965. He received his MS (I 987) and PhD (1991) degrees from St. Petersburg University, Russiu, where he worked with Professor Alexey Dneprovskii.He was then awarded a JSPS postdoctoral fellowship and joined the group of Professor Tamotsu Takahashi at IMS, Okazaki, Japan. He is currently a postdoctoral research associate with Professor Ei-ichi Negishi at Purdue University. His research interests are mainly in the development of new and syn- thetically attractive reactions mediated by organotratzsition metal compounds. 417 CHEMICAL SOCIETY REVIEWS, 1996 ate permanentpolanzation complexation dynamicpolarizationA. ?M -1' 11 transmetatlation 0 'L, 0 'M-2L0 0 + 1L-28 Scheme 2 upon hydrolysis, >98% (2)-a-methylstilbene in 84% yield, while PhCECPh iodinolysis gave (E)-a-iodostilbene in 75% yield Although the detailed structure of the organometallic product was not established 2 Me3AI CpzTiClp in the original study, it has recently been identified as an alkenylti- tanium complex lX(Scheme 3) As might be expected on this basis, 1 the reaction is only stoichiometric in Ti, and our brief attempts to develop its catalytic version have not so far been successful PhHPh In a concurrent and independent study, Tebbe9 reported that the Me TiCp2CI.AIMe,Cl3., 1 (n =3or2) reaction of Me,AI and Cp,TiCI, in the same 2 1 ratio would produce a bimetallic complex 2, known as the Tebbe reagent, the formation of which must involve a C-H activation as elucidated by Grubbs'O (Scheme 4) Interestingly, the reaction of 2 with PhCrCPh was shown a year later to give a titanacyclobutene 3 presumably via carbotitanation of methylenetitanocene (4) with PhCrCPh Thus, the same reagent combination, I e Me,AI, Cp,TiCI, and PhCSCPh, in the same molar ratio but mixed in different sequences and time intervals has led to two discrete processes These early results were already pointing to the intriguingly multi-faceted nature of Scheme 3 carbometallation reactions of early transition metal-Al reagents 3 Zirconium-ca tal ysed Carboalu mination ofcp2Tic12 Alkynes 2 Me3AI 3.1 Methylalumination of Alkynes Although interesting, the carbotitanation reaction of alkynes7 turned I out to be of limited synthetic scope, besides being only stoichiomet- ric in Ti In a situation of this nature, it is often profitable to screen other metals of the same triad and of the neighbouring groups Indeed, we discovered that the use of Cp,ZrCI, in place of Cp,TiCI, 2 led to a similar but catalytic reaction of much greater synthetic value shown in Scheme 5 30 This reaction is, in principle, competitive with Normant's carbocupration,2 but the two reactions have turned out to be synthetically rather complementary to each other Specifically, the Zr-catalysed carboalumination reaction can readily lbase handle the cases of carbometallation with methyl? l2 allylI4 and benzyl l4 groups These groups do not appear to be readily accom- modated by carbocupration The Zr-catal ysed carboal umination reaction IS relatively unaffected by proximal heteroatoms such as halogens, 0,S,15nand SI,'~~permitting the synthesis of trisubstituted PhCECPh alkenes containing two heterofunctional groups which can be used to synthesize a wide variety of terpenoids and carotenoids On the Ph Ph other hand, proximal heteroatoms significantly affect the regio- and stereo-chemistry of carbocupration As of 1994,the syntheses of Ecp2 over 40 simple and complex natural products have made use of the 3 Zr-catalysed alkyne carboalumination reaction No further discus- sion of the synthetic aspects of the reaction is permitted here, and our previous reviewsJh l6 should be consulted for further details Clarification of the mechanism of the Zr-catalysed methyl- alumination has proved to be very challenging We initially envi- sioned that the reaction might involve (1) methylation of Cp,ZrCI, with Me,AI to produce MeZrCp,CI and Me,AICI, (11) Ph)=S<Ph methylzirconation of alkynes to give alkenylzirconium deriva- ICHp H(or D)CH2 H(or 0) tives, which most likely is promoted by an A1 reagent, and (111) their reverse transmetallation with Me,AlCl to yield the observed Scheme 4 alkenyldirnethylalanes with regeneration of Cp,ZrCl, (Scheme 6) CARBOTITANATION AND CARBOALUMINATION OF ALKENES-E NEGISHl AND D Y KONKAKOV H30*or X = Me and/or CI Scheme 5 RCZCR Scheme 6 Me,,AICl3, + RClCR Scheme 7 RCECR R;I '?4ix2 x = ci andlor R' 1 +/ R1=Me, Et and higher alkylCp2Zr-CI 6 Scheme 8 This was supported by observation of a reversible Me-CI exchange between Me,AI and Cp,ZrCI, by NMR spectroscopyi7 and the stoichiometric reaction of RCrCAIMe, with preformed MeZrCp,CI among others (Scheme 5) l8 However, our sub-sequent study has indicated that it might actually involve direct addition of the Me-AI bond to alkynes promoted by a ZrCp, derivativei7 (Scheme 7) Thus, for example, Me,AlCI-Cp,ZrCl, is a reasonable methylaluminating agent,17 even though no Me-GI exchange between the two compounds to produce MeZrCp,CI is detectable by NMR spectroscopy In view of the Curtin-Hammett principle, however, rigorous exclusion of the mechanism shown in Scheme 6 cannot be made on the basis of the currently available data Furthermore, other six-centred mech- anisms, such as those represented by 5 and 6, (Scheme 8) must also be given serious considerations Regardless of the precise mechanisms, however, these Zr-catalysed methylalumination reactions must not involve cy C-H activation 3.2 Alkylaluminationof Alkynes via Simple Addition of Alkyl-Metal Bonds The Zr-catalysed alkylalumination of alkynes with alkylalanes con- taining Et and higher alkyl groups initially proved to be problem- atic However, it has been found that the reaction of alkynes with ---.. 2. D30* R' -D (60 -90%) R = Zi?kylor aryl. R' P Et or n-Pr 1. Et2AICI .. '-wI I 2. D,O+ Et' 'D 91% (2'1) Scheme 9 R,AICl-Cp,ZrCl, reagent systems rather than R,AI-Cp,ZrCl, combinations in chlorinated hydrocarbons, e g (CH,CI),, can provide the desired syn alkylalumination products in good yields,'* although its regioselectivity appears to be significantly lower than that of methylalumination (Scheme 9) Relatively little IS known about the mechanism of these reactions, but several four- and six- centred processes similar to those considered for methylalumina- tion may be considered Here again, there is no indication of C-H activation On the other hand, some related reactions of Et,AI and Prn,AI have turned out to represent a major mechanistic surprise, as discussed later 4 Hydroalumination of Alkenes and Alkynes with Triisobutylalane and Zirconocene Dichloride One significant limitation of the Zr-catalysed carboalumination is that isoalkylalanes, e g Bul,AI, do not undergo carboalumination Instead, Bui,Al-Cp,ZrCI, acts as a hydroaluminating agent Both alkenes and alkyne~l~~~ can be hydroaluminated For the hydro- alumination reaction of alkenes, a mechanism shown in Scheme 10 has been proposed I9 Examination by NMR spectroscopy of some reaction mixtures indicates that the initial hydrozirconation prod- ucts, i e chloroalkylzirconocenes,build up and subsequently decay, supporting the transmetallation-hydrozirconation-reverse trans-metallation mechanism The mechanism involving hydrozircona- tion with BulZrCp,Cl has been further supported by the fact that BulZrCp,C1 generated by the treatment of Cp,ZrCI withBu'MgC1 does hydrozirconate alkenes,' and alkynes 22 However, it is not clear at the present time whether BulZrCp,CI, first undergoes dehydrozirconationto give HZrCp,CI, which then hydrozirconates alkenes and alkynes An alternative possibility that BulZrCp2C1 interacts directly with carbon+arbon T bonds via a six centred Scheme 10 CHEMICAL SOCIETY REVIEWS, I996 Scheme 11 transition state must be considered seriously. Regardless of the precise mechanistic details, both Bui,Al-Cp2ZrCI, and preformed BulZrCp,Cl serve as convenient alternatives to HZrCp2C1.However, some significant differences between HZrCp2C1 and Bul,Al-Cp,ZrCl, have also been observed. For example, the reac- tion of dec-5-yne with Bu’~AI-C~,Z~CI, gives, after deuteriolysis, an essentially 1: 1 mixture of (ZJ-5-deuterio-dec-5-ene and (Z)-5-deuterio-dec-4-ene20 and the corresponding reaction of dec- 1-yne produces a mixture of (0-1-deuterio-dec- lene and 1,l-dideuterio-decane20 (Scheme 11).These results indicate that the synthetic scope and mechanistic details of the hydrozirconation processes involving the use of various ‘HZrCp2CI’ equivalents may vary and must therefore be carefully examined and delineated. 5 Dzhemilev Ethylmagnesiation of Alkenes and its Mechanism Involving Cyclic Carbozirconation via p C-H Activation One of our earlier disappointments was that the reaction of Me,Al-Cp,ZrCI, with alkenes, e.g. oct-1-ene, did not provide the desired methylalumination products. In retrospect, this failure was to be expected. Since the desired products are isoalkylalanes, they can undergo competitive hydroalumination of the starting alkenes discussed in the preceding section, unless the desired methyl- alumination reaction is considerably faster than the competing hydroalumination process. Our recent investigation has established that the reaction of oct- 1-ene with Me2Al-Cp2ZrCI2 indeed gives 2- hexyl-oct-1-ene as the major product along with a smaller amount of 2-methyI-o~t-l-ene,2~ both of which must have been formed via carbometallation-dehydrometallationas depicted in Scheme 12.In view of the results with methylalanes described above, Dzehmilev’s report on the Zr-catalysed ethylmagnesiation of alkene~,~(Scheme 13) came as a surprise to us. We were further intrigued by the fact that neither methylmagnesium nor higher alkylmagnesium derivative^^^ participated satisfactorily in this reaction, but we had little intention to pursue these puzzles, as our main interest in the organozirconium area had already been shifted to a seemingly unrelated topic of the chemistry of low oxidation state ‘ZrCp,’ derivatives.Following the initial and promising discovery that enynes undergo ‘ZrCp,’-promoted bicyclization leading to the formation of monocyclic and bicyclic organic compounds26 (Scheme 14), we embarked on a systematic investigation on (i) the P-H abstraction reaction of dialkylzirc~nocene~~to produce alkene zirconocenes or zirconacyclopropanes,28(ii) their ring expansion reactions with alkenes and alkynes via carbo~irconation,2~and (iii) various sub- sequent reactions of five-membered zirconacycle~~~ (Scheme 15). A number of other workers have also contiibuted to this area, but the scope of this review does not permit a detailed presentation of their significant contributions.For further details of these processes, recent reviews by us3’ and others32 as well as pertinent references therein should be consulted. In one specific example, we have found that treatment of Cp,ZrCI, with 2 equiv. of EtMgBr gives Et,ZrCp, which smoothly decomposed at or above 0 “C to produce (ethyl- ene)zirconocene, which reacts with alk- 1-enes to give ‘pair’-select- ively and regioselectively 3-alkyl-substituted zirconacyclopentanes in nearly quantitative yield.33 When 3, rather than 2, equiv. of EtMgBr was accidentally employed, however, a totally different set of products consisting of a 2-ethylalkylmagesium derivative and I -100% (1:l) IM.ML,], stoichiometric process I minor I catalyticcycleI I major ML, = Zr and/or Al group Scheme 12 Scheme 13 (ethylene)zirconocene, which could be trapped as its PMe, complex, was obtained.33 Clearly, the third equivalent of EtMgBr reacted with the zirconacyclopentane derivative, and the course of the reaction has been clarified as shown in Scheme 16.33Perhaps more significantly, however, it one day dawned on us that a series of the three discrete stoichiornetrric reactions, one of which was a totally serendipitous discovery, would add up to Dzhemilev’s Zr-catalysed ethylmagnesiation of alkenes (Scheme 17).More sugges- tive and less detailed concurrent and subsequent studies by others34 CARBOTITANATION AND CARBOALUMINATION OF ALKENES-E. NEGISHI AND D.Y. KONKAKOV 42 1 n =lor2 Z = C, Si, Ge, or Sn group 0 Scheme 14 I Z Z L \ Z Scheme 15 1- Et Scheme 16 Scheme 17 have also revealed similar, but not necessarily the same, cyclic mechanisms. It is striking and instructive that the reaction which might have initially appeared to involve a straightforward addition of an ethyl-metal bond to an alkene actually involves such an intri-cate cyclization-ring opening process. Furthermore, these intri- guing findings have sent us a clear and burdensome warning that, for any carbometallation reactions involving Zr or perhaps even other related transition metals, such as Ti and Hf, cyclic mech- anisms via C-H activation must be considered along with the more straightforward addition processes.6 Zirconium-catalysed Cyclic Carboalumination of Alkynes via Bimetallic C-H Activation As indicated in Section 3, the reaction of alkynes with tri-alkylalanes-Cp,ZrCI, reagent systems in chlorinated hydro-carbons, e.g. (CH,CI),, proved to be rather complex, even though the corresponding reaction of R,AICI-Cp,ZrCI, was much cleaner and synthetically useful. Thus, for example, the reaction of dec-5- yne with Et,AI (3 equiv.) and 10 mol% of Cp,ZrCI, in (CH,CI), at 23 “C for 3 days produced, after deuteriolysis, 7 as the major product along with a couple of alkyne dimers (8 and 9) and a hydroalumination product 10 formed as minor byproducts8 (Scheme 18).Incorporation of two deuterium atoms in 7-9 was a strong indication that some cyclic carbometallation processes via C-H activation must have taken place. The fact that the Me group in the Et moiety of 7 was only 50% deuteriated indicated that it must have been formed via partially cyclic and partially acyclic processes. This reaction was reported as early as 19782J2335but these intricate details had remained unnoticed until recently. In contrast with the reactions with methylalanes, Et,AICI, and Pr,AICI , proceeding readily in chlorinated hydrocarbons (vide supra), that with Et,AI proceeded faster and more cleanly in non-polar solvents, e.g. hexanes, producing nearly exclusively a cyclic carboalumination product represented by 11, which gave 7 and 12 upon deuteriolysis and iodinolysis, respectively (Scheme 19).In the light of the mechanism of the Dzhemilev ethylmagnesiation discussed earlier, we initially assumed that this reaction too must proceed via Et,ZrCp, formed by double transmetallation reaction of Et,AI and Cp,ZrCI, and (ethy1ene)zirconocene. The latter is known to undergo a ‘pair’-selective ring expansion reaction with alkynes to give the corresponding zirconacyclopentenes306 which CHEMICAL SOCIETY REVIEWS, 1996 LD (50% D) 3 Et3AI 10 (5%) nBuCZCBu-n 0.1 cp~zrcl2 7 (67%) (02) -23 'C,3 d + n-Bu nPr nau 8u-n 8 (2%) 9 (13%) Scheme 18 IFBU Bu-nDCI P D (>98% D) D(>98% D)3EtSAI 0.1 CMrClz 7 (92%)n-BuCZCBu-n ,(hexanes I 23 'C, 6 h THF 12 (54%) Scheme 19 R R yZcP2 Scheme 20 R = n-Bu Scheme 21 may possibly undergo a more or less thermoneutral double trans- metallation to give 11 and Et,ZrCp, (Scheme 20).However, we ?;-"became doubtful about this mechanism, when we failed to detect 1. Et3Al(1 equiv)Cp2ZrC12 (1 equiv) -mBu even a trace of Et,ZrCp,. Our doubt became a reality, when addi- -tion of 13 (20 mol%) to a 1:3 mixture of-dec-5-yne and Et,AI CH3 failed to induce the expected catalytic and cyclic carbometallation p;-"reaction. In fact, no reaction was observed. Consequently, the n-BuCf CBu-n 1. Eta Al(3 equiv)Cp2ZrCla (1 equiv) n-Bu intermediacy of 13 and hence the entire mechanism shown in 2. DCI, D20Scheme 20 must be ruled out.CH2D For clarification of the mechanism of this reaction, a series of n-Bu, ,Bu-ndetailed earlier studies of the reaction of Et,AI with Cp,ZrCI, by Sin@-3' and Kamin~ky,~ in the 1960s and 1970s involving NMR 3Et3AI + Cp$C12 1. 23 'C, 24 h t and X-ray analyses proved to be very informative. These workers 2. nBuCECBu-n CHD2 have found that the reaction of Cp,ZrCl, with Et,AI in a 1 : 1 ratio in 3. DCI, D20 C6D6 rapidly produces a mixture of 14a and 14b, which is relatively Scheme 22stable in the absence of an excess of Et,AI . With an excess of Et,AI, however, a C-H activation process takes place to give 15 which is subsequently converted to a more stable product 16 via another evant to the catalytic process. To further probe the course of this C-H activation process. The X-ray structure of 16 has also been reaction, 15 was prepared cleanly in 83% yield by the treatment of obtained.We propose that 15 and 16 are formed via 17 and 18, EtZrCp,CI, preformed by hydrozirconation of ethylene with respectively, by a novel bimetallic p C-H activation followed by HZrCp,Cl, with 1 equiv. of Et,AI. Since there is no Et,AICI, in this ring expansion, as shown in Scheme 21. With these Zr-A1 bimetal- case, which must interfere with the reaction of EtZrCp,CI with lic species in mind, we carried out the reaction of dec-5-yne with Et,AI, it is considerably faster and cleaner than the reaction of Et,AI and Cp,ZrCI, in benzene under three different sets of condi-Cp,ZrCI, with an excess of Et,Al. The reaction of 15 prepared from tions as indicated in Scheme 22 and obtained the mono-, di-, and tri-EtZrCp,CI and Et,AI, with dec-5-yne was, as expected, very fast deuterio derivatives of (Z)-5-ethyl-dec-5-ene. Both reaction and complete in 10 min at 23 "C to regenerate EtZrCp,CI and the conditions and the formation of the dideuteriated product 7 clearly expected alane product 11 (Scheme 23).Interestingly, little or no indicate that the reaction of 15 with alkynes is the one which is rel- interaction between the two compounds was detectable by NMR CARBOTITANATION AND CARBOALUMINATION OF ALKENES-E NEGISHI AND D Y KONKAKOV n-CpZZr.+ AI(CH2CH3)2CI’ R 23 ‘C, 10 min 15 CI Et w 19 11 Scheme 23 2 Et3AI 17 Scheme 24 spectroscopy Assuming that the initial carbometallation product of the reaction of dec-5-yne with 15 is 19 (or its regioisomer), a cat- alytic cycle shown in Scheme 24 may be proposed for the Zr-catal- ysed reaction of dec-5-yne with Et,AI It consists of (1) carbometallationof dec-5-yne with 15 to give 19 or its regioisomer, (I/) Et-alkenyl exchange between Zr and A1 species and dissocia- tion to give 11 and EtZrCp,CI, (111) complexation of the latter with Et,Al to form 17 detectable by NMR spectroscopy, and (IV) its bimetallic C-H activation reaction to regenerate 15 It is worth men- tioning here that carbozirconation of alkenes with CI-bridged Zr-A1 cyclic bimetallic reagents has been recently reported 38 A novel and critically significant notion reflected in the catalytic cycle shown in Scheme 24 is a bimetallic pC-H activation process of 17 and 18, requiring (I) p C-H containing alkylzirconocene moiety, ([I)one C1 atom for tying Zr and A1 through a C1 bridge, and (1“) a trialkylalane, e g Et,AI, rather than a di- or mono-alkylalane The first requirement needs no further comment The requirement of one CI atom was originally indicated by the failure to use 13 as a catalyst To further substantiate this conclusion, 1 equiv of Et,AICI was added to 13 As expected, the reaction produced EtZrCp,CI and 11,and this mixture indeed catalysed the cyclic car- boalumination of dec-5-yne with Et,AI In this reaction, Et,AICI serves as d source of C1 Evidently, one C1 atom per one Zr atom is needed for the Zr compound to act as a catalyst Any excess beyond this ratio would become inhibitory Presumably, Et,AICI or EtAICI, compete for Zr with Et,AI and produces stable double C1-bridged bimetallic species, such as 14,which probably have to be converted to 17 for the formation of 15 This provides a plausible explanation for the third requirement indicated above 7 Zirconium-catalysed Enantioselective Carboalumination of Alkenes 7.1 Enantioselective Methylalumination Catalytic enantioselective carbon-carbon bond formation involving simple alkenes without heteroatom functional groups represents a highly desirable but formidable synthetic challenge One of the ulti-mate goals in our study in this area has been to achieve enantiose- lective carbometallation of alkenes under the influence of chiral zir- conocene or titanocene derivatives, but our earlier attempts were all unsuccessful As mentioned in Section 5, the reaction of alkenes with Me,Al-Cp,ZrCI, failed to provide the methylalumination products (Scheme 12) due to competitive hydroalumination (Scheme 10) For successful observation of Zr-catalysed methyl- alumination, all but the initial methylmetallation in Scheme 12 must be effectively blocked This requires a carbometallation process which is faster than competing hydrometallation and carbometalla- tion processes Although this appeared to us to be very wishful, we were encouraged by the known asymmetric, if non-enantioselec-tive, Kaminsky-type alkene polymerization reacti~n,”~ which must proceed via a series of carbometallation processes favoured over potentially competitive hydrometallation processes Provided that such a favourable carbometallation process could be devised, the next key question was if it could be highly enantioselective We rea- soned that such a process for at least methylmetallation would have to involve either four-centred direct carbozirconation similar to that shown in Scheme 6 or six-centred processes similar to those repre- sented by 5 and 6,preferably the former, for effective alkene face selection Ironically, all these structural and mechanistic apprehensions were swept away by the surprisingly favourable observation of conversion of oct- 1-ene into an 88% yield of (2R)-2-methyl-octan- 1-01in 72% ee by the reaction shown in Scheme 25 23 How does this reaction avoid chiral product-depleting hydrometallation’ One pos- sible explanation is that the hydrometallation process may be asso- ciative as indicated by 20, which would be increasingly hindered as the steric requirements of the ligands increase This point is cur- rently under investigation Although not certain, the relatively high 1.MqAI (1 equiv.) Ct2ZrCp2* (21) (8 mol %) (CH&I)z, 22 ‘C, 12 h -RYOH R4 2.02 Scheme 25 CHEMICAL SOCIETY REVIEWS, 1996 20 21 22 Scheme 26 23 65%, 33%ee 63%, 92% 88 Scheme 27 % ee figure appears to be consistent with either a four-centred direct carbozirconation process similar to Scheme 6 or a six-centred process similar to that represented by 5.Judging from the observed absolute stereochemistry of the product, Erker 's chiral zirconium complex containing 1 -neomenthylindene(21)40must select the re face of monosubstituted alkene, provided that the carbometallation step involves a syn addition, as indicated in 22 for the four-centred version.The Zr-catalysed enantioselective methylalumination reaction of monosubstituted alkenes appears to be reasonably general with respect to the carbon groups in alkenes. Both chemical and optical yields are in reasonable ranges, although further improvements are clearly desirable. The reaction also appears to be compatible with certain heteroatom functional groups, such as alcohols and amines. The experimental results obtained with 21 are summarized in Table 1.7.2 EnantioselectiveAlkylalumination Having developed the Zr-catalysed enantioselective methyl- alumination of alkenes, we turned our attention to the development of similar alkylalumination reactions. However, the initial outlook was rather bleak. Our attempts to develop an enantioselective pro- cedure based on Dzhemilev's ethylmagne~iation~~ were very dis- appointing leading only to very low % ee figures. In this connection, however, it is noteworthy that favourable results have been obtained by Hoveyda41 through the use of special classes of alkenes, i.e. cyclic ally1 ethers. Dzhemile~~~~~-~ also reported recently the Zr- catalysed reaction of monosubstituted alkenes with Et,AI pro-ducing aluminacyclopentanes (23).Unfortunately, however, the reaction of dec-1-ene in the presence of 21 led only to the product of 33% ee (Scheme 27).In an attempt to observe an acyclic alkylalumination process, Et,AICI in conjunction with 21 was used, but the results were very disappointing. Recalling significant solvent effects observed in the carboalumination reaction of alkynes (Section 6),the reaction of dec-1-ene with Et,Al (1 equiv.) and a catalytic amount of Cp,ZrCI, was carried out in (CH,Cl),. After deuteriolysis, 3-(deuteriomethyl) undecane was obtained in 37% yield along with a cu. 20% yield each of 2-ethyl-dec-1-ene and 1-deuteriodecane. The reaction must have undergone acyclic ethyla- lumination to the extent of 57%, but a competing hydroalumination reaction must have depleted the ethylalumination product to the extent of 20%.Encouraged by these results, we then ran the same reaction in the presence of 21 in place of Cp,ZrCI, and observed, for the first time, a favourable ethylalumination which appears to proceed via non-cyclic ~arbometallation~~ (Scheme 27). The use of CH,CI, or CH,CHCI, at 0 "C or, preferably -25 "C, was optimal. As summarized in Table 2, a variety of monosubstituted alkenes have been converted to ethyl- and higher alkyl-aluminated prod- Table 1 Zirconium-catalysed methylalumination of monosubstituted alkenes' Substrate rlh Product Yield."% ee.% -12 -OH 88 72 12 92 74 12 80 65OV\,H 24 77 70 528 30 85 I2 81 74 HO-0 12c Et2N-OH 79 75 Et2N-0 96d Et2N-OH 68 71 The reactions were run using 8 mol% of 21, I equiv.of Me,AI in 1,2-dichloroethane at 22 "C. Isolated yields. Threefold excess of Me,AI was used Twofold excess of Me,AI was used ucts. The observed chemical yields are lower than those of methyl- alumination by 10-15% presumably due to competitive hydro- metallation, but the % ee figures observed under the optimized conditions were mostly in the 90-95% range. Here again, uniform and predictable re face selection has been observed. 8 Epilogue Our odyssey in the area of carbometallation promoted or catalysed by early transition metals started about 20 years ago with very simple and perhaps naive notions, such as those shown in Schemes CARBOTITANATION AND CARBOALUMINATION OF ALKENES-E NEGISHI AND D Y KONKAKOV Table 2 Zirconium-catalysed alkylalumination of monosubstltuted alkenes‘ Substrate R of R,AI Solvent Temp 1°C Time fh Quenching agent Product Yieldb % %ee 25 4 02 mBuFOH 65 68 25 4 02 it mBuFOH 70 68 25 4 02 Et RBUCOH 72 67 25 6 02 OBu,OH 57 81 0 6 02 Et mBUTOH 63 92 -25 6 02 mBuFOH Et 60 94 25 6 02 Et mB”FOH 70 86 0 24 02 Et m6uFOH 74 93 0 12 02 0 0 10 25 0 24 02 10 0 12 02 62 91 0 The reactions were run using 8 mol% of 21 1 equiv of R,AI unless otherwise stated Threefold excess of R,AI was used Twofold excess of R,AI was used Isolated yields 1 and 2 From the beginning it has encountered a series of unex- pected results and puzzles, and the results obtained by us and others have almost always been much more varied and intricate than ini-tially thought Nonetheless, a number of expected and synthetically useful reactions, such as Zr-catal ysed carboalumlnatlon of alkynes, Zr-catalysed or -promoted hydrometallation of alkenes and alkynes, a wide variety of ‘ZrCp,’-promoted and -catal ysed reactions, and Zr-catalysed enantioselective carboaluminatlon of alkenes, have been discovered and developed As sometimes stated, simple, naive and imperfect theories and notions are clearly better than no the- ories and notions, provided that they are fundamentally sound and employed with discretion After all, Scheme 1 is essentially the same as the better known Dewar-Chatt-Duncanson model, one of the most fundamental theories in organotransition metal chemistry The only difference is that a (T bond is used as a HOMO in place of a filled unbonding orbital The synthetic utility of carbometallation in the synthesis of natural and unnatural complex organic molecules has been well founded over the past few decades and has become undisputable Many additional applications of the reactions and procedures discussed in this review will be forthcoming At the same time, it IS still highly desirable to promote additional methodological and mechanistic investigations so that both acade- mic and industrial synthetic chemists may take full advantage of the carbome tal lation-based methodology Acknowledgements We are deeply indebted to our coworkers, whose namesappearinourpaperscltedherem,especiallyDrs D E Van Horn, T Yoshida,J A Miller,D R Swanson,T Takahashi,C J Rousset, D Choueiry and N Suzuki Our collaboration with Professor Takahashi’s research group at Institute for Molecular Science, Okazaki,Japan,has been particularly fruitful Ourresearch in thisarea has been mainly supported by the National Science Foundation References 1 E Negishi, ‘Organometallics in Organic Synthesis’, Wiley-Intersclence, New York, 1980, pp 532 2 J P Collman, L S Hegedus, J R Norton and R G Finke, ‘Principles and Applications of Organotransition Metal Chemistry,’ University Science Books, Mill Valley, California, 1987 3 For use of the term ‘carbometuflutlon’ see (a)D E Van Horn and E Negtsht,J Am Chem Soc ,1978,100,2252,(6) E Negishi,PureAppl Chem , 1981,53, 2333, (c) J F Normant and A Alexakis, Synrhesis, 1981,841 4 (a) J Boor, ‘Ziegler-Natta Catalysts and Polymerization,’ Academic Press, New York, 1978, (b)H Sinn and W Kaminsky,Adv Organornet Chem ,1980,18,99 5 ‘Friedel-Crafts and Related Reactions,’ ed G A Olah, Intersclence Publishers, New York, 1963 6 (a)T Mole and E J Jeffery, ‘Organoaluminum Compounds,’ Elsevier, 426 Amsterdam, 1972, (b)G Zweifel and J A Miller, Org React, 1984, 32,l 7 D E Van Horn, L F Valente, M J Idacavage and E Negishi, J Organomet Chem ,1978,156,C20 8 Unpublished results obtained in our laboratories 9 F N Tebbe,G W Parshall and G S Reddy, J Am Chem Soc, 1978, 100,3611 10 K C Ott, E J M deBoer and R H Grubbs, Organometallics, 1984,3, 223 1 I F N Tebbe,G W Parshall and D W Ovenal1,J Am Chem SOC ,1979, 101,5074 12 E Negishi, D E Van Horn and T Yoshida, J Am Chem Soc , 1985, 107,6639 13 J F Normant and M Bourgain, Tetrahedron Lett ,1971,2583 14 J A Miller and E Negishi, Tetrahedron Lett ,1984,25,5863 15 (a) C L Rand, D E Van Horn, M W Moore and E Negishi, J Org Chem , 1981, 46, 4093 (b) E Negishi, F T Luo and C L Rand, Tetrahedron Lett ,1982,23,27 16 (a) E Negishi and T Takahashi, Aldrichrm Acta, 1985, 18,31, (b)E Negishi and T Takahashi, Synthesis, 1988, 1 17 E Negishi and T Yoshida, J Am Chem Sac ,1981,103,4985 18 T YoshidaandE Negishi,J Am Chem Soc , 1981,103,1276 19 E Negishi and T Yoshida, Tetrahedron Lett, 1980, 1501 20 E Negishi, D Y Kondakov, D Choueiry, K Kasai and T Takahashi, J Am Chem SOC ,1996,118,9577 21 E Negishi, J A Miller and T Yoshida, Tetrahedron Lett, 1984, 25, 3407 22 D R Swanson, T Nguyen, Y Noda and E Negishi, J Org Chem , 1991,56,2590 23 D Y Kondakov and E Negishi, J Am Chem Soc ,1995,117,10771 24 U M Dzhemilev, 0 S Vostnkova and R M Sultanov, Izv Akud Nauk SSSR, Ser Khim ,1983,213 25 C J Rousset, E Negishi, N Suzuki and T Takahashi, Tetrahedron Lett , 1992,33,1965 26 (a)E Negishi, S J Holmes, J M Tour and J A Miller, J Am Chem SOC ,1985,107,2568,(b)E Negishi,F E CederbaumandT Takahashi, Tetrahedron Lett, 1986, 27, 2829, (c) E Negishi, S J Holmes, J M Tour, J A Miller, F E Cederbaum, D R Swanson and T Takahashi, J Am Chem Soc ,1989,111,3336 27 (a) E Negishi, D R Swanson and T Takahashi, J Chem SOC ,Chem Commun, 1990, 1254, (b) E Negishi, T Nguyen, J P Maye, D Choueiry, N Suzuki and T Takahashi, Chem Lett , 1992,2367 28 (a)T Takahashi, D R Swanson and E Negishi, Chem Lett ,1987,623, (b)T Takahashi, M Murakami, M Kunishige, M Saburi, Y Uchida, K Kozawa, T Uchida, D R Swanson and E Negishi, Chem Lett, 1989, 761,(c)D R SwansonandE Negishi,Organomeralfics,1991,10,825 CHEMICAL SOCIETY REVIEWS, 1996 29 (a) D R Swanson, C J Rousset, E Negishi, T Takahashi, T Seki, M Saburi and Y Uchida,J Org Chem ,1989,54,3521,(b)C J Rousset, D R Swanson, F Lamaty and E Negishi, Tetrahedron Lett ,1989,30, 5105 30 (a) T Takahashi, T Fujimori, T Seki, M Saburi, Y Uchida, C J Rousset and E Negishi, J Chem Soc ,Chem Commun ,1990,182, (6) T Takahashi, M Kageyama, V Denisov, R Hara and E Negishi, Tetrahedron Lett , 1993,34,687, (c)C Coperet, E Negishi, Z Xi and T Takahashi, Tetrahedron Lett ,1994,35,695 31 (a) E Negishi, Chirnica Scripta, 1989, 29, 457, (b) E Negishi, in ‘Comprehensive Organic Synthesis,’ ed L A Paquette, Pergamon, 1991, Vol 5, 1163, (c)E Negishi and T Takahashi, Ace Chem Res , 1994,27,124 32 S L Buchwald and R B Nielsen, Chem Rev, 1988,88,1047 33 T Takahashi, T Seki, Y Nitto, M Saburi, C J Rousset and E Negishi, J Am Chem Soc , 1991,113,6266 34 (a)A H Hoveyda and Z Xu, J Am Chem Soc ,1991,113,5079, (b) K S Knight and R M Waymouth, J Am Chem Soc ,1991,113,6268, (c) D P Lewis, P M Muller, R J Whitby and R V H Jones, Tetrahedron Lett ,1991,32,6797 35 For related studies, see (a) U M Dzhemilev, A G Ibragimov, 0 S Vostrikova and G A Tolstikov, Izv Akad Nauk SSSR, Ser Khim ,1989, 196, (6)U M Dzhemilev, A G Ibragimov, A P Zolotarev and G A Tolstikov, Izv Akad Nauk SSSR, Ser Khim , 1989, 1324, (c)U M Dzhemilev, A G Ibragimov, A P Zolotarev, R R Muslukhov and G A Tolstikov, Izv Akad Nauk SSSR, Ser Khim , 1991,2570, (4U M Dzhemilev, A G Ibragimov and A P Zolotarev, Mendelrev Commun , 1992,135 36 (a)H Sinn and G Opermann,Angew Chem Int Ed Engl ,1966,5,962, (b)H Sinn and E Kolk, J Organoniet Chem , 1966,6,373 37 (a) W Kaminsky and H Sinn, Liebigs Ann Chem ,1975,424, (6)W Kaminsky and H J Vollmer, Liebigs Ann Chem , 1975, 438, (c)W Kaminsky, J Kopf, H Sinn and H J Vollmer, Angew Chem Int Ed Engl ,1976,15,629 38 G Erker, R Noe and D Wingbermuhle, Chem Ber ,1994,127,805 39 (a) H Sinn, W Kaminsky, H J Wollmer and R Woldt, Angew Chem , 1980,92,396, Angew Chem Int Ed Engl , 1980, 19,390, (b)H H Brintzinger, D Fisher, R Mulhaupt, B Rieger and R M Waymouth, Angew Chem ,1995,107,1255,Angew Chem Int Ed Engl ,1995,34, 1143 40 G Erker, M Aulbach, M Knickermeier, D Wingbermuhle, C Kruger, M Nolte and S Werner, J Am Chem Soc ,1993,115,4590 41 J P Morken,M T DidiukandA H Hoveyda,J Am Chem Soc ,1993, 115,6997 42 D Y Kondakov and E Negishi,J Am Chem Sac, 1996,118,1577