DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 2437 Structural direction by the dominant metal Michael A. Beswick, Marta E. G. Mosquera and Dominic S. Wright * Chemistry Department, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1EW Assessment of the structures of a number of heterometallic alkali metal/p block metal complexes suggests that the p block metals have a dominant role in dictating the structures of these species, since generally greater covalency of the metal–ligand interactions leads to more geometrically rigid metal co-ordination. 1 Introduction We have recently reported the development of a general synthetic strategy for the preparation of a variety of heterometallic alkali metal/p block metal cage complexes involving the stepwise metallation reactions of primary amines and phosphines (REH2; E = N or P) with organoalkali metal reagents, producing [(REH)M]n (M = alkali metal), followed by their deprotonation by various p block metal reagents 1–3 such as Sb(NMe2)3 4 (Scheme 1).Our principal aims in these Scheme 1 R E H H R E M H R E M Y [R¢M] [Y] M= alkali metal, Y= p block source studies have been the preparation of a broad range of species containing a variety of mixed-metal stoichiometries and the investigation of the dependence of the cage structures and mixed-metal stoichiometries on the reaction system employed. A large number of these species have now been synthesized and structurally characterised, containing a broad spectrum of Group 13, 14 and 15/alkali metal compositions, and we are now in an appropriate position to assess the central factors governing the structures of these species.This assessment is of primary importance to us since the imido and phosphinidene p block metal anion fragments of these cages are novel ligand systems and can be used as robust synthons in the targeted design of heterobimetallic complexes.5 The purposes of this perspective are (i) to highlight the fact that it is the p block metal–ligand frameworks (not the alkali metal–ligand frameworks) which largely control the structures of heterometallic p block metal/alkali metal cage compounds of this type, (ii) to clarify the issues concerned in thermodynamic and kinetic control of p block metal/alkali metal cage compounds in general and (iii) to address the issue of whether ‘ring-stacking’ and ‘ring-laddering’ models 6 are appropriate in the prediction or interpretation of structural influences in systems in which metal valence and covalency provide the fundamental structural influences.Dr. Michael A. Beswick was born in Warrington, Cheshire in 1964. He obtained his first degree at Hatfield Polytechnic in 1990 and his Ph.D. at Cambridge University in 1992, under the supervision of Professor the Lord Lewis. After a period of research at the University of Murcia in Spain, with Professor J. Vicente, he returned to Cambridge where he is now a postdoctoral researcher in inorganic chemistry. His principal interests concern cluster and cage compounds of transition and main group metals.Dr. Marta E. G. Mosquera was born in Astorga (León, Spain) in 1969. She obtained her first degree in 1992 at the University of Oviedo. Her Ph.D. research, under the supervision of Professor V. Riera and Dr. J. Ruiz at the University of Oviedo, concerned the chemistry of ruthenium–diphosphinomethanide complexes.She joined the Wright group in Cambridge as a postdoctoral researcher in 1997. Her research focuses on the chemistry of Groups 14 and 15. Dr. Michael A. Beswick Dr. Marta E. G. Mosquera Dr. Dominic S. Wright Dr. Dominic S. Wright was born in Gosport, Hampshire in 1964. He obtained his first degree at Strathclyde University in 1986 and his Ph.D. at Cambridge in 1989, under the supervision of Dr. R. Snaith. After a college research fellowship with Gonville and Caius College, Cambridge, he was appointed to his current position as a lecturer in inorganic chemistry at Cambridge University.He was the recipient of the 1993 RSC Meldola Medal. Current research interests include synthetic and structural studies of p block metal metallocene compounds, metallacyclic p block metal– ligand systems, and heterometallic complexes containing novel metalbased ligand arrangements.2438 J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 Fig. 1 Structures of [(Me2N)2Sb3(NCy)4Li] 1, [{Sb2(NCy)4}2Li4] 2 and [{Sb(NCy)3}2Li6]?2Me2NH 3 (top) and imido anion units (bottom) 2 Structural Influences in p Block Metal/Alkali Metal Cages Clearly the outcome of the reactions producing any mixed p block metal/alkali metal cage complex will be highly dependent on the thermodynamic balance between the p block metal– ligand and alkali metal–ligand bond energies.In the case of imido SbIII/Li complexes, all the evidence illustrates that Sb]N bonding dominates the reactions involved.This is witnessed particularly by the formation of alternative imido anions in which the Sb–N frameworks of the antimony precursors are conserved. This point is illustrated in Fig. 1 which depicts the alternative imidoantimony(III) anions formed by the reactions of dimethylamidoantimony(III) reagents with various primary amido or phosphido lithium precursors [Li(REH); E = N or P] and the structures of the heterobimetallic cages resulting. For example, in the reaction of Li1[Sb(HNR)4]2 with Sb(NMe2)3 (1 : 2 equivalents) the (10e) antimony centre of the [Sb(HNR)4]2 anion is preserved in the resulting [(Me2N)2Sb3(NR)4]2 monoanion (e.g.in Li[Sb3(NMe2)2(NR)4] 13b) and in the reaction of the dimer [{Sb(NMe2)(m-NCy)}2] with [(CyNH)Li]n (1:2 equivalents) the Sb2N2 unit is maintained in [{Sb2(NCy)4}2Li4] 2, containing the dianion [Sb2(NCy)4]22.3a The dominance of the antimony(III) frameworks over the outcome of these reactions is also emphasised by the formation of similar complexes of the type [{Sb(ER)3}2Li6],3b, f,g containing [Sb(ER)3]32 trianions (e.g.[{Sb(NCy)3}2Li6]?2Me2NH 3,3b Fig. 1), from a broad range of Li(REH) with Sb(NMe2)3 (3 : 1 equivalents), and by the fact that even where the same organic groups are present the outcome of the reactions is dictated by the reaction sequence [e.g., as in the case of 1, 2 and 3, all of which contain CyN groups (Cy = C6H11)]. Studies of the reactions of heterobimetallic SbIII/Li complexes with metal salts stress the fact that it is the imido antimony(III) monoanion, dianion and trianion units of the cages which are the robust chemical entities.5 Most dramatically, all six of the Li1 cations of the N6Li6 core of [{Sb- (NCy)3}2Li6]?2Me2NH 3 are substituted in the transmetallation reaction with [PbCp2] (Cp = C5H5), giving the heterobimetallic complex [{Sb(NCy)3}2Pb3] 4 (Fig. 2).5a In the reaction of 3 with [KOBut], cleavage of the N6Li6 core occurs (the weaker N]Li bonds being replaced by stronger O]Li interactions), giving the trimetallic cage [{Sb(NCy)3}2Li6]?3KOBut 5 (Fig. 3).5c It should be noted that the lability of the Li]N frameworks is in fact a general feature of all of the imido SbIII/alkali metal cages shown in Fig. 1,5 in which the various imido antimony(III) anions are readily transferred (intact) to a range of main group and transition metal ions, e.g. the reactions of [{Sb2(NCy)4}2- Li4] 2 with Group 11 salts give [{Sb2(NCy)4}2M4] (M = Cu 6 or Ag 7).5b Further evidence of the integrity of the [Sb(ER)3]32 trianions of the [{Sb(ER)3}2Li6] (E = N or P) cages and of the lability of Fig. 2 Conversion of compound 3 into the lead complex [{Sb- (NCy)3}2Pb3] 4J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 2439 their E6Li6 cores comes from variable-temperature 7Li NMR studies of the Me2NH-solvated phosphinidene system [{Sb- (PCy)3}2Li6]?6Me2NH 8 in which a binomial septet is observed even at 290 8C (Fig. 4).3e Semiempirical MO calculations suggest that the six Li1 cations are involved in a dynamic ‘carousel’ process by which they are coupled to the six equivalent P centres of the two intact [Sb(PCy)3]32 trianions of the core. Fig. 3 Cage expansion of compound 3 into the trimetallic [{Sb- (NCy)3}2Li6]?3KOBut 5 Fig. 4 Structure of the cage [{Sb(PCy)3}2Li6]?6Me2NH 8 (top), the 7Li NMR spectrum of 8 (298–208 K) and the associated ‘carousel’ process involved (bottom) The dominance of Sb]N bonding over Li]N bonding in these systems allows direct control of the nature of the imidoantimony(III) anions formed and clearly these units will have a direct bearing on the ultimate structures adopted by the heterobimetallic SbIII/alkali metal cages.Although it cannot be taken for granted that the more covalent p block metal– nitrogen bonds will necessarily be stronger than the ionic alkali metal–nitrogen interactions in heterometallic p block metal/ alkali metal cages in general, what is certain is that even where there is a closer match in bond energies than appears to be present in the antimony(III) systems the more covalent p block metal–ligand bonding will always exert the greatest influence over the structures adopted, since such bonding imparts directionality in the surrounding ligand framework.These more rigid geometric requirements will dominate the ionic and largely non-directional alkali metal–ligand interactions and the alkali metals will have generally a minor role in dictating the structure. One consequence of the greater structural influence of the p block metal framework is that the alkali metal–nitrogen frameworks of heterobimetallic imido complexes should be modified at the expense of the more rigid bonding demands of the p block metal.As the diVerence in the metal–ligand bond energies increases greater distortion of the alkali metal–ligand cores is anticipated.Some evidence for this is seen in the structure of the sodium complex [{Sb(NCy)4}2Na4] 9 (Fig. 5).7 In contrast to the lithium analogue (2 in Fig. 1) which has a tetrahedral arrangement of the four Li1 cations at its centre,3a the four Na1 cations of 9 are distorted into a square-planar arrangement. This occurs as a consequence of the strain induced by the complexation of the larger Na1 cations by the [Sb2(NCy)4]22 dianions, the weaker and more flexible Na]N core being modified as a result.It is noticeable here that distortion of the imidoantimony(III) dianions only occurs at the m-N centres and that the pyramidal geometries of the antimony(III) centres are very similar in 2 and 9. There is emerging evidence that the p block metal– heteroatom frameworks have a dominant role over the structures of heterometallic cage complexes in general. In the reactions of primary amidolithiums [Li(RNH)] with Sn(NMe2)2 only the imido tin(II) cubanes [Sn(NR)]4 are isolated.8 For this reason alternative stepwise metallation procedures have to be employed in order to build the desired mixed SnII/alkali metal arrangements.Using the reactions of Li(REH) (E = N or P) with [Sn(NR)]4 9 (employing the cubane as the base) we showed that heterobimetallic systems are accessible. The fragmentation Fig. 5 Structure of [{Sb(NCy)4}2Na4] 92440 J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 of the imido tin(II) cubane arrangement of the precursor in these reactions exhibit a dependency on the acidity of the primary amido or phosphido lithium.Thus, all four of the ButN groups of [Sn(NBut)]4 9c are replaced in the reaction with Li(CyPH) (4 : 6 monomer equivalents), giving the metallacyclic cage complex [{Sn2(m-PCy)}2(m-PCy)2(Li?thf)4] 10 (Fig. 6), whereas the reaction of the less acidic Li(C10H7NH) with [Sn(NBut)]4 gives [Li(thf)4][Sn3(NBut)(NC10H7)3] 11 (Fig. 7), in which only three of the ButN groups are eliminated.2a The pattern of reactivity observed for [Sn(NBut)]4 and the structures of the products formed strongly suggest that the products are templated by the comparatively thermodynamically robust imido tin(II) cubane.Similar reactivity is observed for the imidoaluminium cubane [MeAl(NMes)]4 (Mes = 2,4,6-Me3- C6H2),10 the reaction with Li(CyPH) giving [{AlMe(m-PCy)}2- (m-PCy)2(Li?thf)4] 12 (the aluminium analogue of 10).1 Although the direct reactions of Sn(NMe2)2 with primary amido lithiums has been unsuccessful in the preparation of heterobimetallic complexes, we have recently shown that the polynuclear dimethylamido reagent [{Sn(NMes)2}- {Sn(m-NMe2)}2] 138 is more well behaved in this respect.The reaction of the latter with Li(2-MeOC6H4NH) gives the ‘pseudo-ladder’ [{Sn(MesNH)(m-NC6H4OMe-2)}2(Li?2thf)2] 14 (composed of a central [Sn(MesNH)(m-NC6H4OMe-2)}2]22 dianion complex associated with two thf-solvated Li1 cations) (Fig. 8).2b The latter appears to result from the deprotonation of the primary amido lithium by the MesN groups of 13 followed by elimination of Sn(NMe2)2 (Scheme 2).It is highly significant that no remnant of any N]Li framework (beyond that of a monomer) is preserved in 14. Indeed, the structure, which can be regarded as a co-complex between two Li(MesNH) mono- Fig. 6 Structure of [{Sn(m-PCy)}2(m-PCy)2](Li?thf)4] 10 Fig. 7 Structure of [(Li(thf)4][Sn3(ButN)(C10H7N)3] 11 mers and two SnNC6H4OMe-2 monomers, illustrates from a thermodynamic standpoint that the primary amido lithium precursor is dissembled by lithium solvation and as a result of the preference for Sn]N bonding.The Z-shaped profile of the ladder core of 14 arises directing from the typical pyramidal geometry of the tin(II) centres (N]Sn]N ca. 908). This geometry contrasts with the far flatter ladder arrangements typical of amidolithium ladders,6,11 such as [{[Li(C4H4N)]2?tmen}2] 12 and stresses the dominance of Sn]N bonding and tin(II) valence over the structure of 14.The nature of the reaction producing compound 14 obviously has a profound eVect on the stoichiometry of the product since one Sn is eliminated, leading to the observed 2 : 2 ratio of Sn: Li. As with the related imido SbIII/alkali metal cages, the ultimate structures of the heterobimetallic SnII/alkali metal cage adopted is thermodynamically controlled by the relative metal–ligand bond energies involved and by the key influence of tin(II) valence.However, comparison of the structures of the complexes formed by the reactions of primary amido and phosphinidene compounds with [Sn(NBut)]4 10 and 11 2a and [{Sn(NMes)2}{Sn(m-NMe2)}2]2b provides good evidence for our emerging belief that the type and stoichiometry of the basic anion system produced can be influenced by the reaction employed. Another important factor dictating the structures of these heterometallic species is Lewis base solvation.This factor is of Fig. 8 Structure of [{Sn(MesNH)(m-NC6H4OMe-2)}2(Li?2thf)2] 14 Scheme 2 (i) 12[Li(2-MeOC6H4NH)], thf, 2Sn(NMe2)2 N Sn N Sn N Sn N R R Me Me Me Me Sn N Sn N Li N Li N R' R' R R thf thf thf thf H H 3 Sn(NMe2)2 + 2 RNH2 (i)J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 2441 course in common with the organo and metalloorganic alkali metal complexes, whose structures and association states are well known to be drastically modified by co-ordination of the metal centres.5,11,13 However, the key diVerence between the latter and mixed p block metal/alkali metal complexes is that the higher solvation energies for the more electropositive alkali metals and the closed-shell configurations of p block metals will lead to the preferential or exclusive solvation of the alkali metal cations. This will diminish the competition for the ligand electron density by the alkali metal cations and will have the result of strengthening the control of the p block metal ligand framework over the cage structure of the complexes. Semiempirical PM3 calculations on models of the heterobimetallic ‘pseudo-ladder’ complex [{Sn(MesNH)(m-NC6H4OMe-2)}2- (Li?thf)2] 142b illustrate that the centralised Sn2N2 open-ladder arrangement is therefore strongly influenced by solvation of Fig. 9 Semiempirical PM3 calculations of models of compound 14; heats of formation (DHf) in kcal mol21, cal = 4.184 J Fig. 10 Structure of [{Bi2(NBut)4}(Li?thf)2] 15 Li1.The Sn2N2 open ladder [(LiNH2){Sn(m-NH)}2(LiNH2)] (akin to 14) is only marginally more favourable than the cubane [{Sn(m-NH)}2{Li(m-NH2)}2] and the Li2N2 open ladder [(SnNH){Li(m-NH2)}2(SnNH)]. However, the eVect of monosolvation of Li1 by H2O in these uncomplexed species gives a marked preference for the Sn2N2 open ladder structure (A?2H2O) over the cubane (B?2H2O) or Li2N2-centred ladder (C?2H2O) (Fig. 9). Notably, there is no thermodynamic preference for further solvation of the tin centres of the Li2N2 open ladder, so that the bisolvated Sn2N2 open ladder [{(H2O)2LiNH2}{Sn(m-NH)}2{(H2O)2LiNH2}] A?H2O (Fig. 9) is preferred by 40.5 kcal mol21 to the preservation of the monosolvated Li2N2 open ladder. The balance between the relative energies of the Bi]N and Li]N bonds and lithium solvation appears to underlie the structure of the heterobimetallic BiIII/Li cubane [{Bi2(NBut)4}(Li?thf)2] 15 (Fig. 10),3c whose arrangement is similar to the monosolvated calculational model B?2H2O (Fig. 9). Presumably, the expected closer energies of the metal–nitrogen bonds leads to greater competition for the nitrogen electron density by Bi and Li. This results in a greater influence of the Li1 cations over the Bi]N framework of the [Bi2(NBut)4]22 dianion and in a lower solvation energy for Li1. The monosolvated cubane is now preferred to the bis-solvated open ladder structure. A recent study also illustrates that the electronegativity of the organic substituents can have a strong influence on the structure of p block metal/alkali metal cages.The reaction of the terminal NMe2 substituents of [{Sb(NMe2)(m-NCy)2}2SbK] 16, containing a [{Sb(NMe2)(m-NCy)2}2Sb]2 monoanion and having a similar structure to the Li1 complex 1 (Fig. 1), with CyNH2 gives [{Sb(CyNH)(m-NCy)2}2SbK] 17 in which the spiro structure of 16 is retained.3g However, the reaction of 16 with ButOH gives K[{Sb(m-NCy)}3(m3-NCy)]?h6-C6H5Me 18 in which the antimony(III) anion has rearranged into a nido- Fig. 11 Conversion of the spiro [{Sb(NMe2)(m-NCy)2}2Sb]2 anion of compound 16 into the nido anion of K[{Sb(m-NCy)}3(m3-NCy)(OBut)? h6-C6H5Me 182442 J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 type structure (Fig. 11).3g The reason for this transformation is that the greater electronegativity of the alkoxide substituents increases the Lewis acidity of the antimony(III) centres, the nido arrangement maximising their co-ordination numbers.This result also reiterates the previous conclusion that the p block metal frameworks provide one of the greatest influences over the structures of such cages. 3 Conclusions and General Remarks On the basis of the cyclic ladder structure of the primary amido lithium complex [(ButNH)Li]8 19,14 it was suggested recently that the structure of one such species [{Sb(NCH2CH2Ph)3}2- Li6?2thf] 20,3b whose Sb2N6Li6 core can be regarded as being constructed from the capping of a hexameric imidolithium stack by two antimony(III) centres, is templated by metallation of a rigid hexameric primary amido lithium precursor [(RNH)Li]6 by Sb(NMe2)3 (Fig. 12). The idea that the structures of the cage complexes [{Sb- (NR)3}2Li6] are ‘directed’ by the common hexameric structure of a cyclic primary amido lithium precursor 6 assumes that the N6Li6 units of these species are thermodynamically robust enough to template the eventual cage arrangement and that this unit is sustained during metallation by Sb(NMe2)3.In eVect, the mechanism and the structure of the product are dominated by a single thermodynamic factor. Although we agree that thermodynamic considerations generally play the greatest role in dictating the outcomes of the variety of reactions discussed above and the structures of the heterometallic complexes formed, our conclusion is that structural direction by the primary amido lithium complexes is unlikely and that the bonding demands of the p block metals dominate the kinetics and thermodynamics involved. In this regard, it is noteworthy that the infinite ladder structure of the primary phosphido lithium complex [(CyPH)Li?thf]• 21 (which has previously been structurally characterised) 15 has no bearing on the structure of [{Sb(PCy)3}2Li6]?6Me2NH 8 formed in its reaction with Sb(NMe2)3 3e and that the reaction of the cyclic ladder [(ButNH)Li]8 itself with Sb(NMe2)3 gives [{Sb(NBut)3}2Li6] 22, containing the expected Sb2N6Li6 core (Fig. 13).3f In complexes of the type [{Sb(ER)3}2Li6] 3b,e,f (E = N or P) the eight electron, pyramidal antimony(III) centres have the greatest influence on the molecular architecture of the core (not the ionically bonded, flexible E6Li6 framework) (Fig. 14). Their overall structures are constrained by the rigid geometry of the [Sb(ER)3]32 trianions and the appearance of a E6Li6 cyclic ladder motif, which simply maximises ionic interactions between the chemically robust anion units, has no necessary mechanistic significance.In a general sense, the nature of the p block metal anion units formed in an individual reaction system obviously has a primary role in dictating the ultimate structure of the heterometallic cage formed with alkali metal cations. The structure adopted relies mainly on the interplay between the p block metal–ligand and alkali metal–ligand bond energies, on the diVerent (ionic and covalent) bonding requirements of the alkali and p block metals and on the presence of Lewis base solvation.Assessment of these structures shows the geometric rigidity of the p block metal frameworks, which results from greater covalency and the requirements of metal valence, is a predominant structural factor. Since the alkali metal–ligand Fig. 12 Proposed structural direction of compound 20 by the supposed cyclic ladder [(PhCH2CH2NH)Li]6 precursor NH Li HN Li NH Li NH Li HN Li NH Li N Li N Li N Li N Li N Li N Li Sb Sb 2 Sb(NMe2)3 –6 Me2NH frameworks present in these complexes are almost entirely ionic in nature, it is clear that the alkali metal cations will exert a lesser influence.The dominance of the p block metal frameworks is reinforced by preferential Lewis base solvation of the alkali metal cations. Bearing in mind the complicated factors governing the formation and structures of these heterometallic cage arrangements, no one global theory is appropriate in their general rationalisation or prediction.In this context, although the ‘ring-stacking’ and ‘ring-laddering’ models have been of immense value in the prediction of the structures of a large range of metalloorganic and organoalkali metal complexes,6 it should be noted that all these systems are dominated by ionic interactions. In mixed p block/alkali metal arrangements, in which there is a large disparity between the character and bonding demands of the metals, it should not be expected that related or similar structural arrangements will be adopted because (i) the p block metal valence generally places large demands on structure, (ii) the geometries of the p block metal centres will not be modified significantly by non-directional alkali metal–ligand interactions and (iii) the alkali metal–ligand frameworks are generally the weakest and the most readily solvated and are therefore subject to breakdown and distortion.A key point here is that the formulation of the ‘ring-stacking’ and ‘ring-laddering’ models deliberately ignores whether the alkali metal–ligand bonds are ionic or covalent. Indeed, in the analysis of ring, stack or ladder options consideration of ionicity or covalency is largely irrelevant since the structures and the detailed variations in metal–ligand bond lengths within them can be rationalised equivalently either by assuming overlap between the metal and ligand orbitals or that the most favourable ionic interactions occur by the alignment of the ligand lone pairs towards the Li1 cations.7 It is precisely this point, however, which will lead to the breakdown of these models as predictive tools in heterometallic p block/alkali metal systems where the directionality and rigidity of the p block metal–ligand bonding outweights the propensity for the alignment of the ligand lone pairs towards the alkali metal cations.Thus, it is not just the relative strengths of the alkali metal–ligand and p block metal–ligand bonds which aVect the outcome of the conflict between the two metals for ligand electron density and the resulting dominance of either metal over the structure, but unfavourable distortion and electronic rearrangement of the p block metal also provide potent contributions to the thermodynamic balance.Fig. 13 Reaction of [(ButNH)Li]8 with Sb(NMe2)3 and the formation of [{Sb(NBut)3}2Li6] 22 Fig. 14 An ‘ionic’ formulation (a) in which the Sb31 ions reside in the N6Li6 ‘bed’ 6 and a ‘covalent’ formulation (b) in which the [Sb(NR)3]32 anions are associated by six Li1 cations N Li N Li N Li Li N Li N Li N Sb Sb N Li N Li N Li Li N Li N Li N .... Sb Sb (b) (a)J. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 2443 4 Acknowledgements We gratefully acknowledge the EPSRC (postdoctoral research grant for M. A. B., 1994–1997), the Leverhulme Trust (postdoctoral research grant for M. A. B., 1997–2000), the Royal Society (D.S. W.), the NuYeld Foundation (D. S. W.) and the Spanish Government (M. E. G. M.) for financial support. 5 References 1 R. E. Allan, M. A. Beswick, P. R. Raithby, A. Steiner and D. S. Wright, J. Chem. Soc., Dalton. Trans., 1996, 4135. 2 (a) R. E. Allan, M. A. Beswick, N. L. Cromhout, M. A. Paver, P. R. Raithby, A. Steiner and D. S. Wright, Chem. Commun., 1996, 1501; (b) R. E. Allan, M. A. Beswick, N. Feeder, M. Kranz, M. E. L. G. Mosquera, P. R. Raithby and D.S. Wright, Inorg. Chem., 1998, 37, 2602. 3 (a) R. A. Alton, D. Barr, A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, J. Chem. Soc., Chem. Commun., 1994, 1481; (b) A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, Angew. Chem., Int. Ed. Engl., 1994, 106, 1334; Angew. Chem., Int. Ed. Engl., 1994, 33, 1277; (c) D. Barr, M. A. Beswick, A. J. Edwards, J. R. Galsworthy, M. A. Paver, M.-A. Rennie, C. A.Russell, P. R. Raithby, K. L. Verhorevoort and D. S. Wright, Inorg. Chim. Acta, 1996, 248, 9; (d ) M. A. Paver, C. A. Russell and D. S. Wright, Angew. Chem., 1995, 107, 1077; Angew. Chem., Int. Ed. Engl., 1995, 34, 1545; (e) M. A. Beswick, J. M. Goodman, C. N. Harmer, A. D. Hopkins, M. A. Paver, P. R. Raithby, A. E. H. Wheatley and D. S. Wright, Chem. Commun., 1997, 1879; ( f ) M. A. Beswick, N. Choi, C. N. Harmer, A. D. Hopkins, M. McPartlin, P. R. Raithby, A. Steiner, M.Tombul and D. S. Wright, Inorg. Chem., 1998, 37, 2177; ( g) A. Bashall, M. A. Beswick, C. N. Harmer, A. D. Hopkins, M. McPartlin and D. S. Wright, Chem. Commun., 1998, 261. 4 A. Kiennemann, G. Levy, F. Schue and C. Tanielian, J. Organomet. Chem., 1972, 143, 35; F. Ando, T. Hayashi, K. Ohashi and J. Kotetsu, J. Nucl. Chem., 1991, 15, 2011; K. Moedritzer, Inorg. Chem., 1964, 3, 609; W. Clegg, N. A. Compton, R. J. Errington, G. A. Fisher, M. E. Green, D. C. R. Hockless and N.C. Norman, Inorg. Chem., 1991, 30, 4680; M. M. Olmstead and P. P. Power, Inorg. Chem., 1984, 23, 413. 5 (a) M. A. Beswick, C. N. Harmer, M. A. Paver, P. R. Raithby, A. Steiner and D. S. Wright, Inorg. Chem., 1997, 36, 1740; (b) D. Barr, A. J. Edwards, S. Pullen, M. A. Paver, P. R. Raithby, M.-A. Rennie, C. A. Russell and D. S. Wright, Angew. Chem., 1994, 106, 1960; Angew. Chem., Int. Ed. Engl., 1994, 33, 1875; (c) A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R.Raithby and D. S. Wright, Angew. Chem., 1995, 107, 1088; Angew. Chem., Int. Ed. Engl., 1995, 34, 1012. 6 K. Gregory, P. von R. Schleyer and R. Snaith, Adv. Inorg. Chem., 1991, 37, 47; R. E. Mulvey, Chem. Rev., 1991, 20, 167. 7 M. A. Beswick, N. Choi, C. N. Harmer, A. Hopkins, M. A. Paver, P. R. Raithby, M. McPartlin and D. S. Wright, J. Chem. Soc., Dalton. Trans., 1998, 517. 8 R. E. Allen, G. R. Coggan, P. R. Raithby, A. E. H. Wheatley and D. S. Wright, Inorg. Chem., 1997, 36, 5202. 9 (a) R. E. Allan, M. A. Beswick, M. A. Paver, P. R. Raithby, M.-A. Rennie and D. S. Wright, J. Chem. Soc., Dalton. Trans., 1995, 1994; (b) M. Veith and G. Schlemmer, Chem. Ber., 1982, 115, 2141; (c) M. Veith and O. Recktenwald, Z. Naturforsch., Teil B, 1983, 38, 1054; (d ) H. Chen, R. A. Bartlett, H. V. R. Dias, M. M. Olmstead and P. P. Power, Inorg. Chem., 1991, 30, 3390. 10 P. P. Power and K. M. Waggoner, J. Am. Chem. Soc., 1991, 113, 3385. 11 M. A. Beswick and D. S. Wright, Comprehensive Organometallic Chemistry, 1995, vol. 1, ch. 1. 12 D. R. Armstrong, D. Barr, W. Clegg, S. M. Hodgson, R. E. Mulvey, D. Reed, R. Snaith and D. S. Wright, J. Am. Chem. Soc., 1989, 111, 4719. 13 W. N. Setzer and P. v. R. Schleyer, Adv. Organomet. Chem., 1985, 24, 353; C. Schade and P. v. R. Schleyer, Adv. Organomet. Chem., 1987, 27, 169. 14 N. D. R. Barnett, W. Clegg, L. Horsburgh, D. M. Linsay, Q.-Y. Liu, F. M. McKenzie, R. E. Mulvey and P. G. Williard, Chem. Commun., 1996, 2321. 15 E. Hey-Hawkins and S. Kurz, Phosphorus, Sulfur, Silicon Relat. Elem., 1994, 90, 281. Received 26th January 1998; Paper 8/00687CJ. Chem. Soc., Dalton Trans., 1998, Pages 2437–2443 2443 4 Acknowledgements We gratefully acknowledge the EPSRC (postdoctoral research grant for M. A. B., 1994–1997), the Leverhulme Trust (postdoctoral research grant for M. A. B., 1997–2000), the Royal Society (D. S. W.), the NuYeld Foundation (D. S. W.) and the Spanish Government (M. E.G. M.) for financial support. 5 References 1 R. E. Allan, M. A. Beswick, P. R. Raithby, A. Steiner and D. S. Wright, J. Chem. Soc., Dalton. Trans., 1996, 4135. 2 (a) R. E. Allan, M. A. Beswick, N. L. Cromhout, M. A. Paver, P. R. Raithby, A. Steiner and D. S. Wright, Chem. Commun., 1996, 1501; (b) R. E. Allan, M. A. Beswick, N. Feeder, M. Kranz, M. E. L. G. Mosquera, P. R. Raithby and D. S. Wright, Inorg. Chem., 1998, 37, 2602. 3 (a) R. A. Alton, D. Barr, A. J. Edwards, M. A. Paver, M.-A.Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, J. Chem. Soc., Chem. Commun., 1994, 1481; (b) A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, Angew. Chem., Int. Ed. Engl., 1994, 106, 1334; Angew. Chem., Int. Ed. Engl., 1994, 33, 1277; (c) D. Barr, M. A. Beswick, A. J. Edwards, J. R. Galsworthy, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby, K. L. Verhorevoort and D. S. Wright, Inorg. Chim. Acta, 1996, 248, 9; (d ) M.A. Paver, C. A. Russell and D. S. Wright, Angew. Chem., 1995, 107, 1077; Angew. Chem., Int. Ed. Engl., 1995, 34, 1545; (e) M. A. Beswick, J. M. Goodman, C. N. Harmer, A. D. Hopkins, M. A. Paver, P. R. Raithby, A. E. H. Wheatley and D. S. Wright, Chem. Commun., 1997, 1879; ( f ) M. A. Beswick, N. Choi, C. N. Harmer, A. D. Hopkins, M. McPartlin, P. R. Raithby, A. Steiner, M. Tombul and D. S. Wright, Inorg. Chem., 1998, 37, 2177; ( g) A. Bashall, M. A. Beswick, C. N. Harmer, A. D.Hopkins, M. McPartlin and D. S. Wright, Chem. Commun., 1998, 261. 4 A. Kiennemann, G. Levy, F. Schue and C. Tanielian, J. Organomet. Chem., 1972, 143, 35; F. Ando, T. Hayashi, K. Ohashi and J. Kotetsu, J. Nucl. Chem., 1991, 15, 2011; K. Moedritzer, Inorg. Chem., 1964, 3, 609; W. Clegg, N. A. Compton, R. J. Errington, G. A. Fisher, M. E. Green, D. C. R. Hockless and N. C. Norman, Inorg. Chem., 1991, 30, 4680; M. M. Olmstead and P. P. Power, Inorg. Chem., 1984, 23, 413. 5 (a) M. A. Beswick, C. N. Harmer, M. A. Paver, P. R. Raithby, A. Steiner and D. S. Wright, Inorg. Chem., 1997, 36, 1740; (b) D. Barr, A. J. Edwards, S. Pullen, M. A. Paver, P. R. Raithby, M.-A. Rennie, C. A. Russell and D. S. Wright, Angew. Chem., 1994, 106, 1960; Angew. Chem., Int. Ed. Engl., 1994, 33, 1875; (c) A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, Angew. Chem., 1995, 107, 1088; Angew. Chem., Int. Ed. Engl., 1995, 34, 1012. 6 K. Gregory, P. von R. Schleyer and R. Snaith, Adv. Inorg. Chem., 1991, 37, 47; R. E. Mulvey, Chem. Rev., 1991, 20, 167. 7 M. A. Beswick, N. Choi, C. N. Harmer, A. Hopkins, M. A. Paver, P. R. Raithby, M. McPartlin and D. S. Wright, J. Chem. Soc., Dalton. Trans., 1998, 517. 8 R. E. Allen, G. R. Coggan, P. R. Raithby, A. E. H. Wheatley and D. S. Wright, Inorg. Chem., 1997, 36, 5202. 9 (a) R. E. Allan, M. A. Beswick, M. A. Paver, P. R. Raithby, M.-A. Rennie and D. S. Wright, J. Chem. Soc., Dalton. Trans., 1995, 1994; (b) M. Veith and G. Schlemmer, Chem. Ber., 1982, 115, 2141; (c) M. Veith and O. Recktenwald, Z. Naturforsch., Teil B, 1983, 38, 1054; (d ) H. Chen, R. A. Bartlett, H. V. R. Dias, M. M. Olmstead and P. P. Power, Inorg. Chem., 1991, 30, 3390. 10 P. P. Power and K. M. Waggoner, J. Am. Chem. Soc., 1991, 113, 3385. 11 M. A. Beswick and D. S. Wright, Comprehensive Organometallic Chemistry, 1995, vol. 1, ch. 1. 12 D. R. Armstrong, D. Barr, W. Clegg, S. M. Hodgson, R. E. Mulvey, D. Reed, R. Snaith and D. S. Wright, J. Am. Chem. Soc., 1989, 111, 4719. 13 W. N. Setzer and P. v. R. Schleyer, Adv. Organomet. Chem., 1985, 24, 353; C. Schade and P. v. R. Schleyer, Adv. Organomet. Chem., 1987, 27, 169. 14 N. D. R. Barnett, W. Clegg, L. Horsburgh, D. M. Linsay, Q.-Y. Liu, F. M. McKenzie, R. E. Mulvey and P. G. Williard, Chem. Commun., 1996, 2321. 15 E. Hey-Hawkins and S. Kurz, Phosphorus, Sulfur, Silicon Relat. Elem., 1994, 90, 281. Received 26th January 1998; Paper 8/00687C