CENTENARY LECTUW* QuadrupIe Bonds and other Multiple Metal to Metal Bonds By F. A. Cotton DEPARTMENT OF CHEMISTRY, TEXAS A & M UNIVERSITY, COLLEGE STATION, TEXAS 77843, U.S.A. 1 Introduction Almost exactly 120 years elapsed between the first discovery' (1844) of a com-pound which contains a quadruple bond, namely, Cr2(02CCH&(H20)~, and the recognition (1964) that quadruple bonds e~ist.~,~? However, following this improbably long hiatus, the field has enjoyed a decade of explosive efflorescence and now stands poised for further broad growth. The purposes of this review are to trace the developments over the past ten years, to summarize our present knowledge, and to point out the areas where further work is most needed and further progress is most likely.The planned and deliberate investigation of quadruple bonds, and other highly multiple bonds, between metal atoms began in the only way that it could -with the first conscious and explicit recognition of a multiple, specifically quadruple, bond between two transition-metal atoms. This occurred only in 1964 in my laboratory in a way that I shall now briefly describe. Our studies4-6 of rhenium(m) chloride and its derivatives, which showed that the characteristic structural unit in the chloride itself is an Re3 triangular cluster, and that this cluster persist in all complexes obtained under mild conditions, e.g. Re3C1123- and Re3Cl~(P&)3, had been essentially concluded. Dr. Neil Curtis was then a guest in the laboratory and I suggested that he examine the aqueous chemistry of rhenium with a view to finding ways of preparing ReIII compounds by reduction of Reo4- and also with the idea of preparing, if possible, some mixed clusters (an Re~OsC112~- cluster was a possibility we specifically discussed).After a short period of exploratory work, trying different reductants and various *First delivered at a meeting of The Chemical Society on 4th February 1974 at Bristol University.t1 am reminded in this connection of the similar history of transition-metal n-compIexes. The first one was discovered in 1829 (W. C. Zeise, Pogg. Ann., 1831, 21,497), but here also, almost exactly 120 years elapsed until the true nature of such complexes was recognized(M. J. S. Dewar, Bull.SOC.chim. France, 1951, C71; J. Chatt and L. A. Duncanson, J. Chem. Soc., 1953, 2979). E. Peligot, Compt. rend., 1844,19, 609. F. A. Cotton, et al., Science, 1964, 145, 1305. a F. A. Cotton, Inorg. Chem., 1965,4, 334. J. A. Bertrand, F. A. Cotton, and W. A. Dollase, J. Amer. Chem. SOC.,1963, 85, 1349; Inorg. Chem., 1963, 2, 1106. F. A. Cotton and J. T. Mague, Inorg. Chem., 1964,3, 1094. F. A. Cotton and J. T. Mague, Znorg. Chem., 1964,3, 1402. Quadruple Bonds and other Multble Metal to Metal Bonds reaction conditions, Curtis presented me one day with a sample of a royal blue compound which had an analytical composition of CsReCI4. Since the ‘CsReCld which had already been shown to be Cs~Re~Cl12was dark red, we were naturally very intrigued, and immediately began work to discover the true molecular formula and structure of this substance.Simultaneously, Dr. Brian Johnson, another guest in the laboratory, had been checking a report’ in the Russian literature in which it was claimed that ReCle3- species could be isolated by reduction of ReO4- in aqueous hydrochloric acid with hydrogen gas. These claims proved erroneous, all of the alleged M13ReC16 compounds being, in fact, MIzReCh compounds.* However, the Russian authors mentioned that a blue-green product, to which they assigned the formula KzReC14, was also obtained. We also observed this product. In view of its colour, we wondered if this too could have been incorrectly formulated; we wondered if it might be ‘KReC14’ and thus related to Curtis’ ‘CsReCld.It was soon shown that indeed the stoicheiometry had been reported incorrectly and that the substance isolated by hydrogen reduction of Reo4- is KReC14,HzO. Since the substance crystallized very nicely, an X-ray crystallographic study of it was undertaken by Mr. C. B. Harris, then beginning his doctoral research. A thorough check of the Russian literature, which was impeded by delays in getting papers translated, then revealed that several other blue or green com- pounds alleged to be rhenium(@ compounds had also been reported,g-ll among them (NH&ReC14, KHReC14, and (pyH)HReCle. With regard to the last two, the proposed presence of H appeared to lack any direct support, but was based, apparently, on the result of a determination of oxidation state by a method devised by the Noddacks in their pioneer studies of rhenium chemistry.12 According to Tronev and co-workers, the oxidation state was found to be +2.Presumably, then, in order to reconcile this result with the analytical data, it was considered necessary to introduce a hydrogen atom, as H+, into the formula. We repeated this work and found that, in our hands,13 the Noddacks’ procedure led to an oxidation number of 2.9 f0.2. While Harris was proceeding slowly with the crystal structure analysis on KReCl4,HzO, learning crystallography as he went, we received an issue of the Zhurnal strukturnoi Khimie containing an article14 on the structure of ‘(pyH)HReCl4’. S. J. Lippard, who had just completed a course in Russian, was pressed into service, and he speedily produced a translation which revealed that this compound had been found to have ‘the dimeric group [RezCl~]*-.Eight chlorine atoms constitute a square prism .. . the rhenium atom has for its V. G. Tronev and S. M. Bondin, Khim. Redkikh, Elementov, Akad. Nauk S.S.S.R., 1954,1, 40.* F. A. Cotton and B. F. G. Johnson, Inorg. Chem., 1964,3,780. V. G. Tronev and S. M. Bondin, Dokiady Akad. Nauk, S.S.S.R., 1952, 86, 87. lo V. G. Tronev and A. S. Kotel’nikova, Zhur. neorg. Khim., 1958,3, 1008. l1 G. K. Babeshkina and V. G. Tronev, Zhur. neorg. Khim., 1962,7,215. la I. Noddack and W. Noddack, Z. anorg. Chem., 1933,215, 182. 13 F. A. Cotton, N. F. Curtis, B. F. G. Johnson, and W.R. Robinson, Inorg. Chem., 1965,4, 326. l4 B. G. Kumetzov and P. A. Koz’min, Zhur. strukt. Khim., 1963, 4, 55. Cotton neighbours one rhenium atom, at a distance of 2.22 A, and four chlorine atoms at a distance of 2.43 A. It may be surmised that four hydrogen atoms are situated between C1 atoms on centres of symmetry .. . and serve to bond the [Re2ClsI4- groups . ..to each other.’ We viewed this extraordinary structure with misgivings. The state of refinement was poor and uncertain. The presence of isolated protons ‘on centres of sym-metry’ was unaccountable. The Re-Re distance seemed unbelievably short and the prismatic rather than antiprismatic array of chlorine atoms seemed inexplic- able. Finally, it was stated that severe difficulty with twinning had been encountered, and we wondered if the difficulties had in fact all been sorted out.Harris was impressed with the urgency of solving the KReC14,H20 structure, which he did in a very short time.15 MirabiZe dictu, KReC14,HzO was found to contain an RezCls entity essentially identical to that described by the Russian workers. Some apparent discrepancies between their dimensions and ours were later resolved, and the best value of the Re-Re distance is 2.24f0.01 A. The structure is shown in Figure 1. Figure 1 The structure of the RepCl,*-ion. Once assured that the structure was correct, the obvious challenge was to interpret it in terms of bonding and electronic structure. This was soon done and resulted in the proposal293 that a quadruple bond, consisting of one 0-,two x-,and a 6-component, exists between the rhenium atoms.In so doing, the ion was taken to be RezCls2-, that is, a compound of rhenium(nI), and the previously postulated hydrogen ions were dismissed as spurious. There has never, in my opinion, been the slightest reason to believe they exist, but the myth that they are, l6 F. A. Cotton and C. B. Hams, Inorg. Chem., 1965,4,330. Quadruple Bonds and other Multiple Metal to Metal Bonds or at least may be, present seems to persist in the Russian literature,16-18 with formulas such as ReC12,CH3COO(H),Ha0,16J7where the significance of (H) is left entirely to the reader’s imagination, unless the descriptionlo of it as ‘free hydrogen (H)’ is considered enlightening.The original discussion3 of the quadruple bond was Lased on qualitative considerations of orbital symmetries and rough estimates of overlap integrals. The Re-Re axis was defined as the z-axis, and the Re41 bonds were assumed to project upon the x-and y-axes. The d22-y2 orbital on each metal atom was assumed to be employed mainly in Re-Cl bonding and the remaining four d-orbitals on each metal atom were used to form the Re-Re bond. Overlap of the dz2orbitals gives rise to a a-bond; overlap of corresponding pairs of the d,, and dyzorbitals leads to formation of a pair of n-bonds. Finally, overlap of the d,, orbitals gives rise to a b-bond. The extraordinary shortness of the Re-Re distance is explained by the high multiplicity of the bond.The eclipsed configur- ation is a consequence of the b-component of the bond since that is the only component which is angle-dependent. As Figure 2 shows, the 6 overlap is maximal in the eclipsed configuration and goes to zero upon rotation by 45” to the staggered configuration. cp 1I (01 (b) Figure 2 Sketches showing (a) the maximization of 6 overlap in the eclipsed configuration and (b)how it becomes zero for a staggered configuration. lo P.A. Koz’min, M. D. Swazhkaya, and V. G. Kuznetzov,Zhur. strukt. Khim., 1967,8,1107. l7 P. A. Koz’min, M. D. Surazhkaya, and V. G. Kunetzov, Zhur. sirukt. Khim., 1970, 11, 313. P.A. Koz’min, Doklady Akad. Nauk S.S.S.R.,1972, 206, 1384. Cotton 2 Elaboration of Dirhenium Chemistry With the existence and structure of Re2Ch2-, as well as a working hypothesis as to the bonding, established, attention was next turned to the chemistry of this and related species.An immediate question concerned the ability of the Re2 unit to persist as the ligands were changed. The existence of the bromo-analogue was already kn0wn.19 It could be prepared directly,13J9 in the same manner as Re2Cls2- by using HBr in place of HCl, or it could be obtained by treating a salt of Re2Cls2- with an excess of HBr.l3 The ligand exchange was shown to be completely reversible.13 The existence of the Re2Brs2- ion, incorrectly described as RezBra4- or H2Re2Br82-, was established in a crude crystallographic study in which inaccurate Re-Re distances of 2.21 and 2.27 8, for two crystal forms were given.20 An accurate study 21 made later gave a metal-metal distance of 2.228(4) A, in excellent accord with the best value, 2.24 k 0.01 8, in Rezcls2-.The insensitivity of the Re-Re distance to replacement of Cl by the larger Br is in accord with the expected strength of the metal-metal bond. The fist non-trivial chemistry of the RezXs2- species to be di~coveredl~~~~ was their reversible reaction with carboxylic acids. When RCOzH is present in excess the reactions proceed quantitatively according to reaction (1); moreover, the reactions are completely reversible. The R~z(OZCR)~XZ compounds had previously been obtained in small yields by reaction of rhenium(v) compounds with carboxylic acids23 and much earlier, by Russian workers,24 but the latter formulated them incorrectly as rhenium@) compounds and their true identity became evident only after they had been rationally synthesized by reaction (1).Their stru~tures~5)~~ are as shown in Figure 3(a). The tetracarboxylate is a limiting stoicheiometry, and species with fewer carboxy-groups and correspondingly more halide ions can easily be envisioned. The intermediate case, Re2(02CR)zX4, is well do~urnented,~8*~~ and structures of both the trans-type [Figure 3(b)] and the cis-type have been reported. The compound Re2(02CPh)214, shown in Figure 4, provides an example of the former28 while Rez(OzCCH&CI4,H20 affords an example of the latter.29 Factors affecting the relative stabilities of the cis and trans structures are not understood.A considerable number of other ligand-substitution reaction~~~~2~J0~31 have l9 G. K. Babeshkina and B. G. Tronev, Doklady Akad. Nauk S.S.S.R., 1963, 152, 100. zo P. A. Koz'min, V. G. Kuznetzov, and Z. V. Popova, Zhur. strukr. Khim., 1965, 6, 651. 21 F. A. Cotton, B. G. DeBoer, and M. Jeremic, Znorg. Chem., 1970, 9, 2143. 22 F. A. Cotton, C. Oldham, and W. R. Robinson, Inorg. Chem., 1966, 5, 1798. 83 F. Taha and G. Wilkinson, J. Chem. SOC.,1963, 5406. 14 A. S. Kotelnikova and V. G. Tronev, Zhur. neorg. Khim., 1958, 3, 1016. 25 M. J. Bennett, W. K. Bratton, F. A. Cotton, and W. R. Robinson, Inorg. Chem., 1968, 7, 1570. 2E C. Calvo, N. C. Jayadevan, C. J. L. Lock, and R. Restivo, Canad.J. Chem., 1970,48,219. 27 F. A. Cotton, C. Oldham, and R. A. Walton, Inorg. Chem., 1967, 6, 214. 28 K. W. Bratton and F. A. Cotton, Inorg. Chem., 1969, 8, 1299. 29 P. A. Koz'min, M. D. Surazhskaya, and V. G. Kuznetsov, Zhur. slrukt. Khim., 1970, 11, 313. 30 F. A. Cotton, N. F. Curtis, and W. R. Robinson, Inorg. Chem., 1965, 4, 1696. 31 F. A. Cotton, W. R. Robinson, R. A. Walton, and R. Whyman, Inorg. Chem., 1967,6,929. 31 Quadridple Bonds and other MultQle Metal to Metal Bonds R /% 00LXI/x X/Re -Re-0I x/l\C/* R Re I oc Figure 4 The structure of ReP(04CPh)J,. Cotton been carried out on the ion, leading to products such as Re@CN)s2- and RezCl~L2, where L = tetramethylthiourea or a phosphine. The retention of the RexRe bond is shown by i.r.32 and Raman33 spectra and X-ray crystallo- graphically, as in the case of Re2Cls(PEt3)~,~~the structure of which is shown in Figure 5.The Re~(S04)4~- ion, with bridging sulphate ions, has also been n Figure 5 The molecular structure of RezCls(PEt3)z. prepared.35 It should be noted that there are also many reactions of Rezcls2-, and other quadruply bonded species, in which destruction of the ReRe unit occurs.36 A list of all published structures which contain a quadruple bond between two Re111 atoms is given in Table 1. Oxidation and reduction reactions have not yet been extensively studied, but it appears that they sometimes cause extensive structural change, expecially the oxidations.38 When Cl2 and Br2 were used as oxidants towards Re2Cls2- and Re2Brs2-, respectively, the products were Re2Xg1-J-, which appear to have the 3a C.Oldham and A. P. Ketteringham, J.C.S. Dalton, 1973, 2304. a3 J. San Flippo, jun., and H. J. Sniadoch, Inorg. Chem., 1973, 12, 2326. F.A. Cotton and B. M. Foxman, Inorg. Chem., 1968, 7,2135. 36 F. A. Cotton, B. A. Frenz, and L. W. Shive, Znorg. Chem., in the press. 86 J. A. Jaecker, W. R. Robinson, and R. A. Walton, J.C.S. Chem. Comm., 1974, 306. P. A. Koz’min, G. N. Novitskaya, V. G. Kuznetsov, and A. S. Kotel’nikova, Russ. J. Znorg. Chem., 1971,12, 861. *6 F.Bonati and F. A, Cotton, Inorg. Chem., 1967, 6, 1353, Quadruple Bonds and other Multiple Metal to Metal Bonds Table 1 Re-Re Bond lengths in variousquadruply bonded dirhenium(1rr) species Species Compound Re-Re Distance 8, Ref.Re2Clg2-Cs2Re2ChQH20 2.235( ?) 37 Re2Cb2-K2Re2CI8,2H20 2.241 (7) 15 Re~c18~-(C5HeN)2Re2Cls 2.244(15) 28 Re2Ch(H20)22-Cs2Re2Cl8,2H20 2.213( ?) 37 Re2Brs2-CszRezBrs 2.228(4) 21 RezC14(OAc)2 RezCl4(02CCH3)2,2HzO 2.224(5) 29 Re2(02CC3H7)4 Re2(02CC3H7)4(Re04)2 2.251 (2) 26 Rez(OzCPh)4C12 -2.235(2) 25 Re2(S04)4(H20)22-Na2Rez(S04)4,8H20 2.214(1) 35 Rezh(02CPh)2 -2.198(1) 28 RezCls(PEts)z -2.222(3) 34 well known confacial bioctahedral structure. Reduction of RezCls2- to Re2Cls3- can be carried out electrochemically, but the product decomposes with a rate constant of about 0.4 s-l and is thus not likely to be isolable, at least at room temperat~re.3~The one-electron reduction40 of Re2(02CCH3)4C12 has been observed to give a yellow product, Re2(02CCH3)4+ or Re2(02CCH&CI, which is stable for hours in solution.Its e.s.r. spectrum affords valuable information on its electronic structure, as will be noted later. 3 Extension to Other Metals However remarkable and interesting the quadruply bonded dirhenium compounds might be, this chemistry could scarcely have been regarded as important if it began and ended with the element rhenium. Therefore one of our earliest goals, to which effort was directed immediately after the main facts about the dirhenium compounds were established, was to see whether similar, multiply-bonded pairs of metal atoms did not also occur in the chemistry of other metals in their lower oxidation states.This work was in fact undertaken with considerable optimism since it did not seem likely that a phenomenon which was so prominent in the chemistry of one element could be entirely lacking in that of others. The first developments in extending the chemistry to other elements were completely logical ones. The basic concept of the Periodic Table, namely, that elements with analogous chemical properties occur in the same column, naturally led to early consideration of technetium. Also, it was thought probable that elements from neighbouring columns in oxidation states that would make them isoelectronic with ReIII might well behave similarly. These ideas furnished guidance but no guarantee of success, since the move from the third to the second transition series is in the direction of decreasing stability of metal-to-metal bonds, and changes in oxidation number necessarily involve differences in charge and orbital size which could oppose the formation of multiple metal-to-metal 3s F.A. Cotton and E. Pedersen, Znorg. Chem., in the press. 40 F. A. Cotton and E. Pedersen, J. Amer. Chem. Soc., in the press. Cotton bonds. Thus the position was precisely that expressed in the well-known maxim: Die Theorie leitet, das Experiment entscheidet. With respect to technetium, success came very quickly. The literature already recorded a compound with the empirical formula (NH&T~zC18,2Hz0,~~ and we were able to that it contains TCzcls3- ions essentially isostructural with Re2C1g2-.The difference in charge, implying the presence of one more electron than the eight required for the quadruple bond, was the only surprising feature. It has been shown39 that Tc2Cls3- can be reduced electrochemically to Te2Cls2-, which is diamagnetic and has a lifetime of at least 5 minutes. A more recent claim43 that compounds of the stoicheiometry MI8 [Tc2C18]3,4HzO can be prepared is entirely erroneous; the compounds in question are all M13[Tc2C18],2H20, with MI = NH4, K, and CS.~~ Simultaneously with our discovery of the TCzcls3- ion, it was found by Lawton and Mason45 that molybdenum(I1) acetate, which had been reported earlier by Wilkin~on,~~has a dinuclear structure with four bridging acetate groups, and a very short Mo-Mo distance, reported as 2.11 A but more recently47 found to be 2.093(1)A. The structure is essentially the same as that shown in Figure 3(a) but without the coaxial ligands, X.Since MoII is isoelectronic with ReIII, it was immediately obvious that a quadruple bond exists between the molybdenum atoms. In view of the close analogy of Mo2(0CCH3)4 to Rez(OzCCH3)&12, it was natural to think in terms of a reversible reaction comparable to reaction (1) above, which would lead to the molybdenum analogue of Re2Cls2-, namely Mo2Cls4-. Such a reaction can be carried out, although the conditions of temperature and concentration are critical and the MozCl84- ion is readily hydrolysed. A number of salts of the MozCls4- ion were prepared and structurally characterized48-51 by Mr.J. V. Brencic, a visitor from Jugoslavia, in 1967-1969. Since that time there has been an enormous output of new compound~5~-~~ containing the Moz4+ unit. Those which have been structurally characterized are listed in Table 2, where their Mo-Mo distances are given. Virtually any carboxylic acid will serve in the Moz(OzCR)4 compo~nds.~6@~53 Two of the most stable and easy to handle are the acetate and, particularly, the trifluoroacetate,51,55 which is soluble and volatile. It has been shown that one factor that can increase the chemical stability of the carboxylates is the ability of large R groups to impede access of attacking groups to the coaxial positions.56 Amidine anions 41 J. D. Eakins, G.D. Humphreys, and C. E. Mellish, J. Chem. SOC.,1963, 6012. 42 F. A. Cotton and W. K. Bratton, J. Amer. Chem. SOC.,1965, 87,921; Inorg. Chem., 1970, 9, 789. 43 M. I. Glinkina, A. F. Kuzina, and V. I. Spitsyn, Zhur. neorg. Khim., 1973, 18,403. 41 F. A. Cotton and L, W. Shive, to be published. 45 D. Lawton and R. Mason, J. Amer. Chem. SOC.,1965, 87, 921. 46 T. A. Stephenson, E. Bannister, and G. Wilkinson, J. Chem. SOC.,1964, 2538. 47 F. A. Cotton, Z. C. Mester, and T. R. Webb, Acta Cryst., in the press. 48 J. V. Brencic and F. A. Cotton, Inorg. Chem., 1969, 8, 7. 49 J. V. Brencic and F. A. Cotton, Inorg. Chem., 1969, 8, 2698. 50 J. V. Brencic and F. A. Cotton, Inorg. Chem., 1970, 9, 346. 51 J. V. Brencic and F. A. Cotton, Inorg. Chem., 1970, 9, 351.35 Quadruple Bonds and other Multiple Metal to Metal Bonds Table 2 Dimolybdenum species having muItlIple bonds with confirmed structures Mo2 Entity Compound Mo-Mo DistancelA Ref. (a) Bond Order 4 MozCl84- ~4MOzCl8,2Hz0 2.139(4) 48 (enH2)2MozCl8,2HzO 2.134(1) 49 (NH4)5MozCIg,Hz0 2.150(5) 50 Li4Moz(CH3)8,4THF 2.147(3) 72 M02(02CCH3)4 2.093( 1) 47 Moz(OzCCF3)4 2.090(4) 54 Moz(OzCCF3)4,2py 2.129(2) 55 M02(N2CPh3)4 2.090(1) 57 K4M02(S04)4,2H20 2.1 lO(3) 58, 59 Moz(SzCOCzH5)4,2THF 2.125( 1) 60 M0z(C3H5)4 2.1 83(2) 68 K3Moz(S04)4,3.5HzO 2.164(3) 75, 59 Moz(CHzSiMe3)s 2.167(?) 76 Moz(NMez)s 2.21 4(2) 78 In A. B. Brignole and F. A. Cotton, Znorg. Syntheses, VoI. 13, ed., F.A. Cotton, McGraw-Hill, New York, 1972, p. 87. S. Dubicki and R. L. Martin, Austral. J. Chem., 1969, 22, 1571. 64 F. A. Cotton and J. G. Norman, jun.,J. Coordination Chem., 1971,1, 161. 66 F. A. Cotton and J. G. Norman, jun., J. Amer. Chem. SOC.,1972,94,5697. E. Hochberg, P. Walks, and E. H. Abbott, Inorg. Chem., 1974, 13, 1824. 67 F. A. Cotton, T. Inglis, M. Kilner,'and T. R. Webb, Znorg. Chem., in the press. 68 C. L. Angell, F. A. Cotton, B. A. Frenz, and T. R. Webb, J.C.S. Chem. Comm., 1973,399. 6B F. A. Cotton, B. A. Frenz, E. Pedersen and T. R. Webb, Inorg. Chem., in the press. eo L. Richard, P. Karagiannidis, and R. Weiss, Znorg. Chem., 1973, 12, 2179. D. F.Steele and T. A. Stephenson, Znorg. Nuclear Chem. Letters, 1973, 9, 777. 6a E.H. Abbott, F. Schoenewolf, jun., and T. Backstrom,J. Coordination Chem., 1974,3,255. e3 J. San Filippo, jun., Znorg. Chem., 1972, 11,3140. 64 J. San Filippo, jun., H. J. Sniadoch, and R.L. Grayson, Inorg. Chem., 1974, 13, 2121. 66 J. V. Brencic, D. Dobenik, and P. Segedin, Monatsh., 1974, 105, 142. 66 G. Wilke, et al., Angew. Chem. Internat. Edn., 1966, 5, 151. 67 W. Oberkirch, Dissertation, Technische Hochschule, Aachen, 1963. F. A. Cotton and J. R. Pipal, J. Amer. Chem. SOC., 1971, 93, 5441. Imperial Chemical Industries, Ltd., 1967, B.P. 1 091 296. 70 V. I. Skohlikova, et al., Vysokomol. Soedineniya, 1968, 10, B, 590 (Chem. Abs., 1969, 70, 50602). 71 R.P. A. Sneddon and H. H. Zeiss, J. Organometallic Chem., 1971, 28, 259. F. A. Cotton, J.M. Troup, T. R. Webb, D. H. Williamson, and G. Wilkinson, J. Amer. Chem. Soc., 1974.96, 3824. 73 B. Heyn and H. Still, 2. Chem., 1973, 13, 191. 74 A, R. Bowen and H. Taube, Znorg. Chem., 1974,13,2245. 76 F. A. Cotton, B. A. Frenz, and T. R. Webb, J. Amer. Chem. SOC.,1973, 95, 4431. 76 F. Huq, W. Mowat, A. Shortland, A. C. Skapski, and G. Wilkinson, Chem. Comm., 1971, 1079. 77 M. H. Chisholm and W. W. Reichert, J. Amer. Chem. SOC.,1974,96, 1249. 78 M. H. Chisholm, W. W. Reichert, F. A. Cotton, B. A. Frenz, and L. Shive, J.C.S. Chem. Comm., 1974,480, 804. Cotton RlNC(RZ)NRl-, which are isoelectronic with RC02-, can also serve as bridges,57 as can the sulphate ion,58~~~ the ethylxanthate ion,6096l apparently also ethyl acetate,62 the trifluoromo,thanesulphonateion,62 and the thiobenzoate ion.61 There is a large derivative chemistry of Mo2Cls4-, principally involving molecules of the type Mo2C14L4, in which there are at least 16varied types of L gro~p.~~-~~ Some Mo~Br4L4 analogues were also rnade.G4 There are several organo-derivatives of the Moz4f unit.The first to be rep0rted,~6,67 Mo2(C3H5)4, has been structurally characterized;68 it has the structure shown in Figure 6, and has been shown to possess catalytic activity for Figure 6 The structure of Mo,(allyl), and Cr2(allyl)4.Hydrogen atoms are omitted. olefin69 and stereospecific butadiene70 polymerization. It appears, however, that it is the ally1 groups rather than the Mo2 unit which are the source of the cata- lytic function.The analogous Crz(allyl)4 has similar activity.71 The reaction of Mo2(02CCH3)4 with LiCH3 in ether72 produces Li4 [Mo~(CH~)~],~(C~H&O, from which derivatives with other ethers can be obtained. The structure of the THF-containing compound has been detem~ined;~z it contains the eclipsed octamethyldimolybdenum anion shown in Figure 7. The preparation of a substance with the empirical formula LizMo(CsH5)2H2,2THF has been reported.73 It would seem very probable that this compound is similar to the octamethyl one and contains the M0z(CsH5)4H4~- anion. A considerable aqueous chemistry of the Moz4+ ion has already been developed by Bowen and Ta~be.7~ They found that the M02Cl8~- ion in O.lM-HS03CF3 is hydrolysed to Mo2Ch2+(aq). The addition of an excess of K2S04 to a solution of K4Mo2CIs in 0.2M-HS03CF3 causes a pink precipitate of ICQ[Mo~(S04)4],2H~0 to form in >90 % yield.This can be dissolved in 0.01M-HS03CFs and a slight excess of Ba(SOsCF3)2 can be added to precipitate all the sulphate and leave what is presumed to be the Mo2*+(aq) ion in solution. Despite numerous attempts in various laboratories (all unpublished) no one has succeeded in precipitating this ion with a nonco-ordinating anion. The Moz4+(aq) has an electronic spectrum not greatly different from that of Mo2(02CCH3)4. 37 Qtradruple Bonds and other Mirltiple Metal to Metal Bonds c Figure 7 The structure of the MO~(CH~),~-ion. The compound Mo2(en)4CL was prepared by treating K4Mo2CI8 with neat ethylenediamine on the steam bath for 30-60 minutes followed by washing with ethanol and ether.It is the first compound containing a cationic complex of Moz4+. It has an electronic spectrum quite similar to that of M02~+(aq). The [Mo2(en)4l4+ complex undergoes aquation only slowly (ca. 30 minutes) and reacts slowly enough with 02 to permit the solutions to be handled without rigorous exclusion of air. In the process of crystallizing pink K4 [Mo2(S04)41,2HzO slowly so as to obtain crystallographically useful crystals, it was found that lavender crystals also f0rrned.~~*75X-Ray crystallography showed that they have the composition K3Mo2(S04)4,3.5H20 and contain the [Moz(S04)4]3- ion, which has a structure very similar to that of [Mo2(S04)4]4- (Figure 8).According to the accepted description of the quadruple bond, the electron lost from the 4-ion to give the 3 -ion should come from the &bonding orbital, thereby giving the latter species a 2Bzg ground state. There is considerable evidence to support this idea and thus, indirectly, to support the general picture of the quadruple bond. Thus, as Table 2 shows, loss of this weakly bonding &electron causes a lengthening of the Mo-Mo bond, by about 0.05 A. The two ions are related by an electrode potential of +0.22 V. The 3-ion is paramagnetic, with p = 1.65 BM, gll = 1.891, and 38 Cotton Figure 8 The structure of the MO~(s04)4~-entity which occurs with n = 4 in K4[Mo2(S0J4],2H20and n = 3 in K,[MO~(SO~)~],~.SH~O.There are oxygen atoms in the coaxial positions.gl= 1.903. The e.s.r. spectra show that the spin Hamiltonian has axial symmetry and that the unpaired electron is evenly distributed over two magnetically equivalent molybdenum atoms. Qualitative arguments show40 that the quadruple- bond formulation in quantitative form (See Section 4) would require that both g values be <2, as observed. Before leaving molybdenum, it should be mentioned that not only does this element seem, on present knowledge, to be the most prolific former of quadruple bonds, but it also has a marked tendency to form triple bonds. The two best- characterized examples are Mo2 [CH2Si(CH3)3]676 and Mo2 [N(CH3)2]6.77*78 The structure of the latter is shown in Figure 9, and the metal-metal distances in the two compounds are listed in Table 2.A similar compound,7Q Moz [OC(CH3)3]6, presumably has the same sort of structure. The substance [(q5-C5Hs)Mo(C0)2]2 7B M. H. Chisholm, personal communication. Quadruple Bonds and other Multiple Metal to Metal Bonds C Figure 9 The molecular structure of Mo2[N(CH,),],. The view is almost down the molecular three-fold axis. has been briefly described;8u it presumably has a triple Mo-Mo bond by analogy with { [q5-C5(CH3)5]Cr(C0)2>2.s1 Tungsten.-The great readiness of molybdenum to form M-M bonds of orders 3 to 4 naturally led to the hope that tungsten would also have a rich chemistry of similar compounds. This has not proved to be the case so far. The reaction of W(CO)s with glacial acetic acid does not appear to give Wz(OzCCH3)4, but instead some sort of trinuclear species,s21*3 possibly W3(02CCH&O.With a number of other acids, e.g. C&,C02H, CsF5C02H, C~H~COZH, and C3F7CO2H, compounds of the correct stoicheiometry have been obtained,s2 but none of them have formed crystals, and none are therefore substantiated structurally. The triply-bonded tungsten compound76 W2 [CH2Si(CH3)3 16 has been prepared and reported to form crystals isomorphous with those of the molybdenum compound. R. C. Job and M. D. Curtis, Inorg. Chem., 1973, 12,2510. J. Potenza, P. Giordano, D. Mastropaolo, A. Efraty, and R. B. King, J.C.S. Chem. Comm., 1972, 1333. Es F. A. Cotton and M. Jeremic, Synth. Znorg. Metal-Org.Chem., 1971, 1, 265. 83 T. A. Stephenson and D. Whittaker, Inorg. Nuclear Chem. Letters, 1969, 5, 569. Cotton A triply-bonded tungsten compound, W2[N(CH3)2]6, can be made but has not yet been fully purified (separated from W [N(CH3)2]6} or structurally characterized. If the inaccessibility and intractability of compounds with quadruple W-W bonds results from relative weakness of those bonds, which is not necessarily the case, we must naturally try to understand why this bond should be significantly weaker than those in the Mo24+ and Re26+ cases. The answer is not yet obvious to this writer. A decrease in stability from M024+ to W24+ might be attributed to the presence of the 4f14 shells in the latter impeding the close approach necessary for the &overlap to be effective.However, this factor would not be expected to diminish greatly on moving to Re26+; at least I do not think so. Further efforts towards the preparation of tungsten compounds would seem to be pertinent. Chromium.-Though in many other cases multiple M-M bonds found in com- pounds of the second- and third-rows have no parallel in the first-row elements, such is not the case here; there are numerous compounds containing the Crz4+ entity. Those structurally characterized are listed in Table 3. Table 3 Compounds containing multiple bonds bet ween chromium atoms which have confirmed structures Cr2 Entity Compound Cr-Cr DistancelA Ref. (a) Bond Order 4 Crz(OzCCH3)4,2H20 Cr2(02CCH3)4,2H20 2.3855(5) 84 [Cr2(CO3)414-Mg2Crz(C03)4,6H20 2.22( ?) 86 D~(C~HS)~ L~~C~~(C~HS)~,~THF 9014-1.9735) [c~~(cIG)~14-hCrz(CH3)8,4THF 1.980(5) 89 Crz(ally1)4 Cr2(allyl)4 1.97( ?) 87, 88 (b)Bond Order 3 ( [q5-Cs(CH3)5 ICr(C0)z }2 2.276(2) 81 It is fitting to begin by mentioning Cr2(03CCH3)4,2H20, first reported1 in 1844 but only in 1970 accurately described as to structure and bonding.84 The Cr-Cr distance is 2.39 A and the compound is isostructural and isoelectronic with Mo2(02CCH3)4, except for the presence of the coaxial water molecules.It is therefore reasonable to propose that it contains a quadruple bond. This would require that the compound be diamagnetic, and it probably is. Reported suscepti- bilities for this and a large number of other Cr2(02CR)4L2 compounds85 are 84 F.A. Cotton, B. G. DeBoer, M.D. LaPrade, J. R. Pipal, and D. A. Ucko, J. Amer. Chern. SOC.,1970, 92, 2926; Acra Crysr., 1971, B27, 1664. S.Herzog and W. Kalies, 2.anorg. Chem., 1964,329,83;1966,351,237and other references cited therein. 86 R. Ouakes, Y. Maouche, M.-C. Perucaud, and P. Herpin, Compt. rend., 1973,276,C, 281. 87 G. Albrecht and D. Stock, 2. Chern., 1967,7, 321. T.Aoki, A. Furusaki, Y. Tomiie, K. Ono, and K. Tanaka, Bull. Chem. SOL.Japan, 1969, 42, 545. J. Krausse, G.Marx, and G. Schodl, J, Organometallic Chern., 1970, 21, 159. J. Krausse and G. Schodl, J. Organometallic Chern., 1971, 27, 59. 41 Quadruple Bonds and other Multiple Metal to Metal Bonds always weakly paramagnetic, with apparent magnetic moments of 0.50-0.85 BM.It seems quite likely that this is due to traces of chromium(u1) compounds as impurities, although there is also the possibility that the &bond is so weak that a 66* triplet state is detectably populated. Jt is possible that this uncertainty could be resolved by an e.p.r. study. As will be seen presently, there are other structurally characterized compounds with Cr-Cr quadruple bonds, and one with a triple bond, in which the Cr-Cr distances are much shorter than that in the acetate. This simply illustrates the fact that quadruple bonds can vary in strength and that bond multiplicity is not a direct or single-valued index of bond strength. This situation prevails for bonds of other orders as well.There are, for example, wide variations in the strengths of single bonds, as illustrated by the series C-C (356 kJ mol-I), N-N (160 kJ mol-l), 0-0(146 kJ mol-l), and F-F (158 kJ mol-l). In the case of Cr2(02CCH3)4,2H20 the relatively strong binding of the coaxial ligands (Cr-0, 2.27 8,) is probably related to the relative weakness of the Cr-Cr bond. In Moz(OzCR)4 compounds, where the Mo-Mo bonds are very short, coaxial ligands are either absent or only weakly attached. The exact interplay between these two factors is not clear, but a reciprocal relationship between the strengths of the two bonds does appear to exist. An interesting compound which is similar to the acetate is Mg2Cr2(C03)4,6H20. This contains the [Cr2(C03)4(H2O)2’J4- ion, where the carbonate ions play the same bridging role as do the acetate ions in the acetate and the H2O molecules serve as coaxial ligands.86 In this case the Cr-Cr bond length is 2.22 8, and the Cr-OH2 bond length is 2.32 A, shorter and longer, respectively, than the corresponding bonds in the acetate.We turn now to several compounds in which there are extremely short Cr-Cr bonds. These ‘superbonds’ are the shortest metal-to-metal bonds known. Crz(allyl)4 has a structures7988 analogous to thaF of Moz(allyl)4 (Figure 6), with a Cr-Cr distance of only about 1.97 A. There is also a [Cr2(CH3)8I4- ion with a structuresg essentially identical to that of [Mo2(CH3)sl4- (see Figure 7); the Cr-Cr distance is 1.980(5) A. A closely related species is the [Cr2(C4H8)4I4- ion, in which there are four chelating -(CH&-units, two on each Cr atom.90 The Cr-Cr bond here is the shortest M-M bond presently known, with a length of 1.975(5) A.In addition to the structurally characterized compounds of dichromiumfu) just mentioned there are a number of other reported compounds which seem certain to contain Cr-Cr quadruple bonds. Aside from the numerous carboxylato- bridged species other than the acetate,85,91 there are several organometallic compounds. The compound Cr2 [(CH2)2P(CH3)2 kg2 presumably contains a Cr-Cr quadruple bond and eight Cr-C bonds, but it is not known whether the structure is (a) or (b) of Figure 10, or possibly the variant of (a) in which the two 91 P. Sharrock, T. Thiopanides, and F. Brisse, Canad, J.Chem., 1973, 51, 2963. 92 E. Kurras, U. Rosenthal, H. Mennenya, G. Oehme and G. Engelhardt, 2. Chem., 1974,14, 160. Cotton Figure 10 Two possible structures for Cr, [(CH2)2P(CH,)2]4. sets of chelate rings are eclipsed rather than staggered. Compounds which may contain [Cr2(C5H10)4I4- have been rep~rted.~O The compounds LizCr(U-CsH40)2, 2Ether and several similar onesg3 may well contain a [Cr2(o-C1&0)4]~- ion in which the o-CsH402- ions occupy a set of positions similar to those of bridging carboxy-groups. There is also C~(U-C~H~OM~)~,~~ which could contain dinuclear molecules, though its low solubility perhaps indicates a polymeric structure. It should be noted, to complete the picture regarding chromium(II), that there are also many compounds, including important ones, which do nut have metal- metal bonds according to magnetic or crystallographic data, or both.These include Cr12,95 CrC12,4H20,96 several with the formula M12Cr(S04)2,xH20 where MI is an alkali some M12CrC14 MI~C~BI-~(H~O)~,~~ and Cr [N(SiMe& ]2(THF)2.100 There is no evidence for the existence of a dinuclear aquo-ion (through no effort to detect it is recorded). It has been proposedlO1 that the rate-controlling step in the redox and substitution reactions of 83 F. Hein, R. Weiss, B. Heyn, K. H. Barth, and D. Tille, Monatsber. Deut. Akad. Wiss. Berlin, 1959, 1, 541. s4 F. Hein and D. Tille, 2.anorg. Chem., 1964, 329, 72. ss F. Besrest and S. Jaulmes, Acta Cryst., 1973, B29, 1560.96 H. G. von Schnering and B. H. Brand, 2.anorg. Chem., 1973,402, 159. 87 A. Earnshaw, L. F. Larkworthy, K. C. Patel, and G. Beech, J. Chem. SOC.(A), 1969, 1334. H. J. Seifert and K. Klatyk, 2.anorg. Chem., 1964, 334, 11 3. L. F. Larkworthy and A. Yavari, J.C.S. Chem. Comm., 1973, 632. looD. C. Bradley, M. B. Hursthouse, C. W. Newing, and A. J. Welch, J.C.S. Chem. Comm., 1972, 567. R. D. Cannon and J. S. Lind, J.C.S. Chem. Comm., 1973, 904. 43 Quadruple Bonds and other Multl'ple Metal to Metal Bonds Cr2(02CCH3)4(H20)2 in aqueous media is the dissociation : Cr2(02CCH3)4 2 Cr(02CCH3)2 It thus appears that complexes of the dinuclear cation Crz4+ play a very important role in the chemistry of chromium(@, but is not as dominant as is Moz4+ in the chemistry of molybdenum(I1).The latter, of course, also has an extensive cluster chemistry102 involving Moe clusters with Mo-Mo single bonds, and this has no parallel whatever in the chemistry of chromium(r1). Two compounds containing triple Cr-Cr bonds have been reported. The first was (q5-C5Me5)2Cr2(C0)4, which consists of two (q5-C5Me5)Cr(C0)2 units joined by a Cr-Cr bond which is 2.276(2)A long.1°3 On the reasonable assump- tion that each chromium atom is to have an 18-electron configuration, this should be considered a triple bond. Very recently it has been shown104 that the presum- ably isostructural (q5-C5H5)2Cr2(C0)4 forms readily on thermolysis of the highly strained (q5-C5H5)2Cr2(C0)6.*05 Heteronuclear Quadruple Bonds.-The possibility of forming mixed metal species [MM'(02CR)4], containing two Group VI metals, is very obvious, but little has yet been published.The isolation of CrMo(02CCH3)4 has been re- ported.lO6 The structure has not yet been determined, but if a band in the Raman spectrum (and also in the ix.) at 392 is assigned to Cr-Mo stretching, then the Cr-Mo force constant is only about 2/3 that for Mo-Mo in Mo2(02CCH3)4. McCarley and co-workers107 have also prepared this compound as well as several MoW(02CR)4 compounds and shown that both Moz(OzCR)4 and MoW(02CR)4 species can be oxidized by halogens to such products as Mo2(02CR)4+13- and MoW(02CR)4I. Ruthenium and Iron.-About ruthenium we know only enough to suggest that there may be significant things still to learn.In 1966 Stephenson and Wilkinsonl08 reported the preparation of a series of compounds with the unusual stoichei- ometry Ru2(02CR)4Cl, and having magnetic susceptibilities suggesting the presence of three unpaired electrons per formula unit. The true nature of these compounds was established X-ray crystallographically~~9 a few years later ;the significant portion of the structure of the butyrato-compound is shown in Figure 11. The very short Ru-Ru distance, 2.281(4)A, implies that a very strong M-M bond is present. An orbital scheme was also proposed to explain the presence of three unpaired electrons within the general framework of the quadruple-bond scheme. lo' F. A. Cotton, Accounts Chem. Res., 1969, 2,240. lo3J.Potenza, P. Giordano, D. Mastrapaolo, A. Efraty, and R. B. King,J.C.S. Chem. Comm., 1972, 1333. lo4 P. Hackett, P. S. O'Neill, and A. R. Manning, J.C.S. Dalton, 1974, 1625. lo6R. D. Adams, D. E. Collins, and F. A. Cotton, J. Amer. Chem. Soc., 1974,96, 749. lo6 C. D. Gamer and R. G. Senior, J.C.S. Chem. Confm., 1974, 580. lo' R. E. McCarley, R. J. Hoxmeier, and V. Katovic, personal communication. lo8T. A. Stephenson and G. Wilkinson, J. Znorg. Nuclear Chem., 1966, 28, 2285. lo@M. J. Bennett, K. G. Caulton, and F. A. Cotton, Znorg. Chem., 1969, 8, 1. Cotton Figure 11 A portion of the structure of RU~(O~CC~H,)~C~.The long, angular CI bridgesbetween Ru,(O&H,)~+ units are to be noted. Very recently, Ruz(OzCC3H7)4CI has been more thoroughly investigated.ll0 The magnetic susceptibility from 60 to 300 K and the e.p.r.spectrum in solution show conclusively that Ru2(02CC3H7)4+ has a quartet ground state and that the unpaired electrons are equally shared by the two metal atoms, thus ruling out any mixed-oxidation-state (RuII, RuIII) formulation. It was also shown that oneelectron reduction to Ru2(02CC3H7)4 occurs quasi-reversibly at potentials in the range 0.00 to -0.34 V, depending on solvent. The product, presumably Ru2(02CC3H7)4, appears to be diamagnetic; a crystalline specimen of this could not be obtained. Stephenson and Wilkinsonlos had also tried, unsuccessfully, to isolate crystalline samples of various compounds, e.g. Ruz(OzCCH&(HzO) Ru(OaCR)z(py)4, and Ru(O2CR)a(py)z, which appear to contain Ru in the oxidation state II.It would seem, however, that under appropriate conditions it should be possible to do so. A compound of apparent composition F. A. Cotton and E. Pedersen, Znorg. Chern., in the press. Quadruple Bonds and other Multiple Metal to Metal Bonds Ruz(02CCH3)4.5(H20)1.5 was also reported, and its structure would be of interest. Although no compounds with quadruple bonds are known, iron ranks a close second to chromium among first-row metals in forming multiple M-M bonds. Two compounds, (1) and (2), containing Fe-Fe triple bonds have been reported. Compound (1)ll1 has an Fe-Fe distance of 2.177(3) A, which is about 0.1 A shorter than the triple bond in the very similar [(q5-Me5C5)Cr(C0)~]z.In (2)112 each set of three P atoms belongs to a HC(CH2PPhz)3 ligand and the accompany- ing anion is BPhd-. The Fe-Fe distance here is 2.34 A. We have here another good example of the fact that bond multiplicity is simply a qualitative measure of the number of electron-pair interactions and not an index of bond strength (or length). In compound (2), there seem likely to be three bond components, cr + 2.n, but they are weaker than in (1). There are three compounds containing Fe-Fe double bonds. The first,l13 (3), has an M-M distance of 2.215A. Compound (4)114 is a homologue of (3) and has nearly the same distance, 2.225(3) A. Compound (5),115 with different bridging groups, has a distinctly longer Fe-Fe distance, 2.326(4) A. It is interesting to note that the latter is scarcely shorter than the length (2.37 A) of the formal single bond in (6),116 though the latter is definitely anomalous among single Fe-Fe bonds, which are generally 2.5-2.8 A long.Rhodium.-Dirhodium tetra-acetate dihydrate is a compound which contains the shortest known84 Rh-Rh bond, 2.3855(5) 8.On the basis of comparisons with many Rh-Rh single bonds, which have lengths in the range 2.68 to 2.94 &I1* it has been suggestedllg that this bond must be a multiple one, probably with an order of 3. Recently two compounds, (7) and (8), with double bonds have been 111 S.-I. Murahashi, et al, J.C.S. Chem. Comm., 1974, 563. P. Dapporto, G. Fallani, S. Midollini, and L. Sacconi,J. Amer. Chem. SOC., 1973,95,2021. 113 K.Nicholas, L. S. Bray, R. E. Davis, and R. Pettit, Chem. Comm., 1971, 608. 114 H.-J. Schmitt and M. L. Ziegler, 2. Naturforsch., 1973, 28b,508. 115 J. Calderon, S. Fontana, E. Frauendorfer, V. W. Day, and S. D. A. Iske, J. Organometallic Chem., 1974, 64, C16. 116 P. E. Baikie and 0. S. Mills, Znorg. Chim. Acta, 1967, 1, 55. 11' 0. S. Mills and J. P. Nice, J. Organometallic Chem., 1967, 10, 337. 11* K. G. Caulton and F. A. Cotton, J. Amer. Chem. SOC.,1971, 93, 1914. Cotton OC\ /4 F0 Fe =Fe OC’ ‘L’ ‘co (3) L = BU‘C=ICBU‘ described.ll9 Since these bonds have lengths of 2.46 and 2.55 A, the previous proposal receives further support. The structure of Rh2(02CCH&(PPh&, which is analogous to that of the tetra-acetate dihydrate, has been determined.120 The phosphines are fairly strongly bonded [Rh-P = 2.479(4)A] and the Rh-Rh bond is longer than in the hydrate, namely 2.449(2) A.(7) X=CO; Y=CPh2 (8) x = Y= CPh, A number of formate complexes, mostly of the type [Rh(02CH)2L]2, have been reported.121 The basis for the dinuclear formula is an unpublished X-ray crystal- lographic study of the compound with L = 0.5H20, which is said to ‘contain [Rh(OzCH)2(H20)12 and [Rh(02CH)2 12 units alternating in an infinite chain,’ but no bond distances are disclosed. The preparation of the aquo-ion Rh24+(aq) has also been described.122 This has been obtained in solution by reduction of Rh(HzO)&l+ with Cr2+; no solid compound has been isolated. It is oxidized slowly by air and reacts rapidly with various ligands, for example with acetate ion to generate what appears to be Rh2(02CCH3)4.4 Electronic Structures It is obvious that for moderately complex molecules containing atoms of very high atomic numbers, e.g. Mo2CIs4- and Rezcls2-, conventional MO calcu-113 H. Ueda, Y. Kai, N. Yauoka, and N. Kasai, 21st Symposium on Organometallic Chemistry, Japan, 1973, Abstract No. 214. laoJ. Halpern and G. Khare, personal communication. lal I. I. Chernyaev, E. V. Shenderetskaya, A. G. Maiorova, and A. A. Karyagina, Russ. J. Inorg. Chem., 1965, 10, 290, and earlier work cited therein. lZ2F. Maspero and H. Taube, J. Amer. Chem. SOC.,1968,90, 7361. Quadruple Bonds and other MuItiple Metal to Metal Bonds lations are barely, if at all, feasible.Therefore, until very recently, theoretical discussion of the electronic structures of quadruply bonded dinuclear species, and others closely related, has been of an essentially qualitative nature, and there was no important development beyond the basic qualitative proposals made by this writer some 9 years ago. Experimental approaches to questions of electronic structure have also been few. Aside from some early attempts to interpret a few electronic ~pectra~~t53J2~ the measurement of bond lengths has been the only experiment a1 met hod consis tent 1y used. In the past few years this picture has begun to change. The SCF scattered-wave Xa (SCF-SW-Xa) method of cal~ulationl~~ appears to afford a practicable avenue of theoretical approach, and experimental studies employing e.p.r.measurements of paramagnetic species generated electrochemically, as well as 0-10-Anti bonding Orbitals Sfg --20 -E 30- 0 < 40-0 2. Cr) 282, - (3 z a w 50- 5E" n- M-M Bonding Orbi to Is u' 60- 70 -282, -Ligand Orbitals and Metal -Ligand Bonding 80 -OrbitalsI I I I-90 Figure 12Aportion of the energy-leveldiagram for Mo2Cla4-according to an SCF-S W-Xor calculation by Norman and Kolari. Levels shown are those with at least 20% metal character. la3F.A. Cotton and C.B. Harris, Znorg. Chem., 1967, 6,924. la*K. H. Johnson, Adv. Quantum. Chem., 1973,7,143. Cotton more sophisticated measurements and interpretations of optical spectra, have begun to appear.The M02Cl8~- ion has been treated by the SCFSW-Xa method by Norman and K0lari.1~5 Their results strikingly confirm all of the essential features of the original proposal3 concerning the quadruple bond. Figure 12, which is adapted from the work of Norman and K~lari,~~~ shows some of their results. It is evident that a little above the metal-ligand and ligand lone-pair orbitals, and below the anti-bonding orbitals, is a set of Mo-Mo bonding orbitals, in the order (r, ?I, 6, Figure 13 The Mo-Mo u (4 A,) bonding orbital, in the xy plane, of MopCls4-accordingto the SCF-S W-Xu calculation of Norman and Kolari. la6 (a) J. G. Norman, jun. and H. J. Kolari, J.C.S. Chem. Cornrn., 1974, 303; (b) J. Amer.Chem. SOC.,in the press. Quadruple Bonds and other Multble Metal to Metal Bonds precisely as expected. The 6*-orbital is not far above the &orbital, also as expected, since this is the weakest component of the quadruple bond. Perhaps the most important result of this calculation is that the 6-,n-, and &orbitals are found to have mainly metal d character (76-93%), so that the original crude approximation3 of describing the quadruple bond in terms of Q , n-, and &overlaps of pure d-orbitals is, in essence, validated. Figures 13-1 5 show contour diagrams of the 0-, n-, and &bonding orbitals. The a-(4Alg)orbital (Figure 13) arises mainly from an overlap of metal 4dz2functions. The outer lobes and equatorial rings of the 4dz2 functions are CT anti-bonding with respect to Figure 14 One of the Mo-Mo n(SEu) bonding wavefunctions of MoZClB4--,from the SCF-SW-Xa calculation of Norman and Kolari.Cotton Mo-Cl interactions, as shown by the nodes between Mo and C1atoms, but there is some appreciable overlap with the C13p orbitals to give a bonding contribution as well. Thisa-orbital has the lowest (76 %) metal dcharacter of those forming the quadruple bond. The n-orbitals, one of which is shown in Figure 14, are obviously the result of 4dn-4dz overlap. Finally, Figure 15 shows one of four equivalent sections through the maximum electron density of the &orbital. This one has the highest metal dcharacter (93 %) and looksjust as one would expect for a 4d6-4d6 overlap.In addition to the basic description of the quadruple bond,3 which the Figure 15 A section through one of the maximalplanes (halfway between the xz and yzplanes) of the Mo-Mo 6(2Bzg)bonding orbital of Mo2Clsa-from the SCF-SW-Xor calculation of Norman and Kolari. Quadruple Bonds and other MultlIple Metal to Metal Bonds SCF-SW-Xa calculations so strikingly confirm, this writer subsequently suggested that there might also be two essentially non-bonding a-orbitals, directed outwards along the M-M axis, and formed mainly from metal s-and pz-orbitals, and that at least one of these would have an energy similar to those of the 6-and 6*-orbitals. The SCF-SW-Xa calculation does not support this sugges- tion. While this is of no importance in describing the ground state of Mo2Cls4-, or other systems in which there are eight or fewer M-M bonding electrons, it is important with regard to the excited states of these molecules and for all species such as Tc2Ch3-, Ruz(OzCR)4+, and Rhz(OzCCH&(Hz0)2, in which there are more than eight electrons.The molecular and electronic structures of these species have previously been rationalized by the writer in ways which employed at least one of these a non-bonding orbitals, but, as Norman and Kolari suggest, it may not be too difficult to rationalize them in other ways. One case, however, that at present seems to offer difficulty is Re2C14(PEt3)4, which has the structure126 shown in Figure 16. This contains two more electrons Figure 16 lXe structure of Re,Cl,(PEt,),, with the ethyl groups omitted for clarity. than Re2Cls2- or Re2Cls(PEt3)2 and is subject to considerable internal crowding that should tend to stretch the Re-Re bond.If the two additional electrons occupy the 6*-orbital, thus nullifying the &bond, as the diagram of Figure 12 would suggest, it is not easy to understand why the Re-Re distance of 2.232(6)A is not significantly longer than those in Re2Cls2- [2.241(7)A] and RezCl6(PEt& [2.222(3)A]. From the spectroscopic side, there is evidence to support the level scheme of F. A. Cotton, B. A. Frenz, J. R. Ebner, and R. A. Walton, J.C.S. Chem. Comm., 1974,4. Cotton Norman and Kolari. They have shown that the observed bands in the visible and U.V.spectra of Mo2Cls4-can be satisfactorily assigned using their diagram; in so doing they assign the lowest observed band to the 6 3 6* transition, whereas, if a non-bonding a-orbital were to lie below the 6*-orbital, this transition would be assigned to 6 -+a. For Re2Ck2- and Re2C16(PEt&, Cowman and Gray127 have offered experimental evidence which favours assigning the lowest observed band to the 6 -+ 6* transition in these species, too. The writer had previously rejected that assignment because the transition is extremely weak although a 6 --f 6* transition is orbitally allowed. Recent work in the writer’s laboratory40 involving an e.p.r. study of species with unpaired electrons that must occupy orbitals above the &orbital also gives results that cast doubt upon the earlier proposal that a a non-bonding orbital is next above the &orbital.This is not the place to pursue the interesting but currently controversial question of the detailed electronic structures of the quadruply bonded and related species. There are many other data bearing on the question. The foregoing discussion merely serves to show that this aspect of the field as well as chemical aspects remain challenging subjects for further research. la’ C. D. Cowman and H. B. Gray, J. Amer. Chem. SOC.,1973,95,8177.