J. Chem. SOC., Faraday Trans. 1, 1984,80, 129-134 The Role of [2 + 21 Cycloaddition-type Reactions in Catalysis Activation of H-H, C-H and C=C Bonds by Metal Complexes BY JAMES G. HAMILTON AND JOHN J. ROONEY* Department of Chemistry, The Queen’s University, Belfast BT9 5AG, Northern Ireland Received 25th April, 1983 It is argued that direct [2 + 21 cycloadditions and cycloreversions involving Mt=C and C=C species and quasi-metallacyclobutanes are the key steps in olefin metathesis. This theoretical concept is then extended in a novel fashion to describe the mechanisms of many other important reactions in homogeneous and heterogeneous catalysis, and especially the activation of C=C, H-H and C-H bonds by metal-alkyl complexes where the metal ion is electron deficient and the alkyl ligand assumes ‘carbenoid ’ character.Detailed investigations of the stericl and kinetic2 aspects of the ring-opening polymerization of norbornene and its various derivatives using olefin metathesis catalysts, including EtAlCl, have convinced us2y that the mechanism involves a direct [2 + 21 cycloaddition and cycloreversion, as shown in scheme 1 . C C c - c c=c [Mtl C [M tl- - -C [Mtl-C [ M t l - c C [Mt]=C ‘ I - + Scheme 1. Since this theory requires that a metallacyclobutane per se be only present as a transition-state species between reactant and product quasi-metallacyclobutane inter- mediates, it is in sharp contrast to the generally accepted idea of independent coordination of olefin and metallacyclobutane formation. If these steps are so distinct an initial orthogonal orientation of the Mt=C and C=C x orbitals is required, and this is implicit in various theories5 that contend that puckering of the metallacycle is the source of various stereospecific features of the overall process (scheme Scheme 2.Scheme 1, in contrast to scheme 2, has also recently received strong support from theoretical calculations,6 and indeed is clearly the only one possible when transients such as Al=C3 and P=C7 mediate metathetic reactions of olefins. That an Al=C species seems to form from EtAlC1, in norbornene and participates in formal 129130 [2 + 21-CYCLOADDITION REACTIONS IN CATALYSIS [2 + 21-type reactions indicates that the n orbitals, HOMO and LUMO, are essentially located on the A1 and C centres, respectively, such that the canonical representation of the mechanism in scheme 3 is a good approximation. c=c + [All 1 C [All-- C’ c ’ c c - c I + I / = / I - [All- C [All- ‘C Scheme 3.When the overlap integral for the p* atomic orbitals is small there is an inherent theoretical problem still unresolved uis-&uis the degree of concertedness and the application of Woodward-Hoffmann theory as opposed to a stepwise mechanism (cf scheme 3), and this has recently been discussed* for [2 + 21 cycloreversions. Indeed Hoffmann and coworkers6 have recently considered this problem in respect of the olefin metathesis mechanism and conclude that canonical forms depict better the quasi-metallacyclobutane predicted by the interaction of the Mt=C and C=C species. A case in point, which well illustrates this difficulty, is that the highly strained olefin adamantene,g which is essentially a 1,2-diradical, reacts with simple alkenes to form cyclobutane derivatives. The main criterion therefore for such [2 + 21-type reactivity of a general A-B species of this type is associated with a weak n component in the bonding in the unit and not its polarity, cf.Al+-C-, P+-C- and C’-C’, although polarity is obviously important in influencing energetic and steric aspects4? of the reaction. Such A-B species are also expected to be widespread in transition-metal chemistry where the n bond is now a @-p* or 8-8 rather than a pk-pn component, but the basic theory for [2+2]-type reactions still seems to apply. Our main purpose now is to argue that A-B species of the above type may not only participate directly in olefin insertion reactions (cf.schemes 1 and 3) but also in H-H and C-H activation by a similar [2 + 21 mechanism. Further examples of such reactive units are Mt-Mt pairs as found on metal surfaces and in metal cluster compounds, even dinuclear compounds, cation-anion pairs, e.g. the Al-0 redox sites in the surface of y-Al,O,, and perhaps even ‘carbenoid’ metal-alkyl units in electron-deficient complexes, as well as metallacarbenes. In all cases the n component in the A-B bond can be well represented by zwitterionic and diradical canonical forms and we will now use these to describe reactions where a [2 + 21-type mechanism seems to apply for each of these examples. \ [ M t l + H R\C/,H I - H, [Mtl ‘IM 11 ri“‘ R \ / I1 H-[Mtl R’C H~ RLC I = .(I [Mtl [ M t l Scheme 4.J.G . HAMILTON AND J. J. ROONEY 131 Because of the very close relationship of the ring-opening polymerization of cy- cloalkenes to Ziegler-Natta polymerization of alk- 1 -enes it was SuggestedlO that a hy- drido-metallacarbene is generated from a metal-alkyl species and that it is the former and not the latter which propagates the addition polymerization. However, there is now growing evidence that, while this may be correct in some cases,ll a ‘carbenoid’ metal-alkyl with an a-H bridging over to the metal, but not fully transferred to it, can be a ground-state species12 which may undergo olefin insertion,13 as shown in scheme 4. The a-H bridging species can first be considered as a three-membered ring. Complete rupture of the C-H bond results in one extreme, the hydridometallacarbene, but severance of the C-Mt bond gives a zwitterion or diradical.It is thus clear why such canonical representations which illustrate the II. character of the C-Mt bonding, >CCH H \ [M t i R H \ [Mtl’ are an important part of the description of the bridging ‘carbenoid’ metal-alkyl unit. It may even be useful to consider the analogous formation of a metallacyclopentane from two alkenes and a metal ion, e.g. tantalum,14 in the same manner, with emphasis on the three-membered ring representations of the metal-alkene complex, as it engages the second alkene molecule (scheme 5). c= c c- c C-$ c /-, +. ll--Pc C [M tl iMt1 iM t j - L “MtI/ 1 - \ / - - I Scheme 5. A most attractive feature of the above mechanism for Ziegler-Natta polymerization is that it may explain the very high stereoselectivities often observed, since there is evidence15 that H migration towards the metal from the a-methylene group (scheme 4) can be remarkably selective.The widening of the C-C-Mt angle required of the ‘carbenoid’ mode of behaviour of the metal-alkyl is also diametrically opposed to the narrowing of this angle which would be required if the chain transfer step of cis-P-H elimination is to take place, so a strong impetus in many cases towards formation of high polymer is understood. The [2+2] theory also leads naturally, without participation of dz orbitals, even though that may be highly beneficial for transition-metal catalysis, and without prior coordination of olefin, to a good mechanism for the oligomerization of ethylene mediated by Al(alkyl),, where the electron-deficient propagating species may be depicted by the following canonical forms : I R \ c / H H/I ‘*I / \ R R R‘ \ /H H 2- Several ‘carbenoid’ metal-alkyl species also seem to be capable of activating H-H and C-H bonds, as evidenced by a growing number of examples described in the132 [2 + 2]-CYCLOADDITION REACTIONS IN CATALYSIS , recent literature.Thus several do-Zr complexes, e g . (Cp),Zr(alkyl)Cl, react directly with dihydrogen,16 which is also used as a chain-transfer agent in typical Ti-based alk-1-ene polymerization systems.17 In the absence of d electrons H-H cannot be cleaved by the usual oxidative addition mechanism so a direct [2+2]-type attack also seems warranted16 (scheme 6).The molecular-orbital treatments of [2, + 2,] and [2, + 2,J cycloadditions are essentially equivalent. R CH, Scheme 6. Another very interesting example is the recently described metal-alkyl compound, Lu(q5-C5Me,),R, which not only catalyses propylene polymerizationla but also cleaves C-H bonds in tetramethylsilane at 20 OC.19 The most interesting example of all may well be that of alkylcobalamins, where there is a long-standing mystery of how paraffinic C-H bonds in substrates are activated in such a highly selective fashion in the vicinal interchange processes catalysed by B,, coenzyme.2o Since the C-C-Co bond angle is ca. 125” the alkyl group attached to the Co ion may well be ‘carbenoid’ in character and quite reactive in a [2+2]-type reaction with substrate (scheme 7), in the same way that the lutetium-alkyl compound reacts with tetramethylsilane.l9 [C 01 R’ [C 01 Scheme 7. Just as described for Ziegler-Natta polymerization (scheme 3), highly stereospecific behaviour will have its origins in the selectivity with which one of the two hydrogen atoms of the a-methylene group moves towards the metal thereby turning this group into a highly reactive chiral centre. Other homogeneous examples are intramolecular addition of a C-H bond across a Ta=C bond21 and dissociative addition of dihydrogen to a Ta=Ta pair of ions2, in a couple of tantalum compounds. In the heterogeneous field there are many such reactions, e.g. the carbanionic activation of paraffins even at room temperature, for deuterium exchange23 by the redox sites on y-Al,03 (scheme 8).+ R-H - A[-0 Al-0 Scheme 8. The [2+2]-type reactions have long since been recognized as occurring on metal surfaces but are called dissociative adsorption, e.g. of H,, or associative adsorption, e.g. of etmlene, on a contiguous pair of sites. An exciting extension of this type ofJ. G. HAMILTON AND J. J. ROONEY 133 mechanism may well reside in the recent discoveries of homologation reactions of paraffins24 and of 0lefins~~9 26 on various catalysts, one of simultaneously promotes olefin metathesis. This implies that adsorbed methylenes are formed from ethylene etc. on a pair of metal sites according to scheme 9, and then react with alkene by another [2 + 21 cycloaddition which is a key step in the homologation process.c c c-c [Mtl C MI- c Etc - l l + l l --I I Scheme 9. Even acetylenes are now known to fall apart to give metallacarbynes upon addition across a W=W triple bond as has been observed in recently described reactions of a dinuclear complex of tungsten.28 Heteronuclear units such as C=N and C=O behave in like fashion. The insight which this recent branch of catalysis and organometallic reactions gives into the mechanism of Fischer-Tropsch synthesis is extremely imp~rtant.~' Note added in proof: B e n ~ z e ~ ~ has reported that carefully purified EtAlCl, does indeed catalyse the cross-metathesis of norbornene and cis-pent-2-ene. Huu Thoi Ho, K. J. Ivin and J. J. Rooney, J. Mol. Catal., 1982, 15, 245. Huu Thoi Ho, B. S. R. Reddy and J. J. Rooney, J.Chem. SOC., Faraday Trans. I , 1982,78, 3307. K. J. Ivin, J. J. Rooney and C. D. Stewart, J. Chem. SOC., Chem. Commun., 1978, 603. J. G. Hamilton, K. J. Ivin, J. J. Rooney and L. C. Waring, J. Chem. SOC., Chem. Commun., 1983,159. M. Leconte and J-M. Basset, J. Am. Chem. SOC., 1979, 107, 7296. 0. Eisenstein, R. Hoffmann and A. R. Rossi, J. Am. Chem. Soc., 1981, 103, 5582. E. Schaumann and R. Ketcham, Angew. Chem., Int. Ed. Engl., 1982, 21, 225. W. Bums and M. A. McKervey, unpublished results pN. Bums, Ph.D. Thesis (Queens University of Belfast, 1976)l. lo K. J. Ivin, J. J. Rooney, C. D. Stewart, M. L. H. Green and R. Mahtab, J. Chem. SOC., Chem. Commun., 1978, 604. l 1 H. W. Turner and R. R. Schrock, J. Am. Chem. SOC., 1982, 104, 2331. l2 Z. Dawoodi, M. L.H. Green, V. S. B. Mtetwa and K. Prout, J. Chem. SOC., Chem. Comrnun., 1982, 1410. l3 D. T. Laverty and J. J. Rooney, J. Chem. SOC., Faraday Trans. 1, 1983, 79, 869. S. J. McLain, J. Sancho and R. R. Schrock, J. Am. Chem. SOC., 1979, 101, 5457. l5 W-K. Wong, W. Tam and J. A. Gladysz, J. Am. Chem. SOC., 1979, 101, 5440. l6 K. I. Gell, B. Posin, J. Schwartz and G. M. Williams, J. Am. Chem. SOC., 1982, 104, 1846. *' See J. Boor Jr, Ziegler-Nutta Catalysts and Polymerizations (Academic Press, New York, 1979), p. ' U. Khlabunde, N. F. Tebbe, G. W. Parshall and R. L. Harlow, J. Mol. Catal., 1980, 8, 57. 251. P. L. Watson, J. Am. Chem. SOC., 1982, 104, 337. I* P. L. Watson, J. Chem. SOC., Chem. Commun., 1983, 276. 2o See J. M. Pratt, in BZ2, ed. D. Dolphin (Wiley, New York, 1982), vol 1, p. 325. 22 A. P. Sattelberger, R. B. Wilson and J. C. Huffman, Znorg. Chem., 1982, 21, 4179. 23 C. S. John, C. Kemball, E. A. Pearce and A. J. Pearman, J. Chem. Res., 1979, (S) 400, (M) 4830. L. Chamberlain, I. P. Rothwell and J. C. Huffman, J. Am. Chem. Soc., 1982, 104, 7338.134 [2 -k 21-CYCLOADDITION REACTIONS IN CATALYSIS 24 C. O’Donohoe, J. K. A. Clarke and J. J. Rooney, J. Chem. Soc., Farada-v Trans. I , 1980, 76, 345; 25 F. Hughes, B. Besson, P. Gussiere, J. A. Dalmon and J-M. Basset, Nouv. J . Chim., 1981, 5, 207. 26 R. A. Strehlow and E. C. Douglas, J . Chem. Soc., Chem. Commun., 1983, 259. 27 P. P. O’Neill and J. J. Rooney, J . Am. Chem. SOC., 1972,94,4383. 29 L. Bencze, ISOM 5 (Graz, Austria. August 1983). J . Chem. SOC., Chem. Commun., 1979, 649. R. R. Schrock, M. L. Listemann and L. C. Sturgeoff, J. Am. Chem. Soc., 1982, 104, 4291. (PAPER 3/652)