Arene cluster compounds~~ ~~Brian F. G. Johnson,*'" Paul J. Dysonb and Caroline M. Martin"a University Chemical Laboratories, Lensfield Road, Cambridge CB2 I E W, UK ' Depurtmmt of Chemistry, Imperial College of Science, Technology and Medicine, South Kensington,London S W7 2A Y. UKThe synthesis, isolation and characterisation of a wide arid diverse range of arene clusters based on nuclearitiesof three to eight metal atoms are described. Their use as models for the chemisorption of benzene on themacromolecular surface are considered and matters relevant to the chemistry on the surface such as bondingtype, mobility and reactivity discussed. Certain of these compounds readily undergo photoisomerisation toproduce benzyne derivatives by cleavage of two C-H arene bonds.This process may be tuned by varying thenature of the other ligands present. In the solid many of the arene clusters exhibit strong arene-areneinteractions which are 'graphitic-like' leading to the formation of chains, ribbons and snakes. Use of bridgingbifunctional arenes such as [2.2]paracyclophane and PhCH=CHCH=CHCH=CHPh leads to useful monomericprecursors to arene cluster polymers.Our interest in this area was originally stimulated' by thechance discovery of the arenehexaruthenium derivatives[Ru,C(CO),,(C,H, -,,Me,,)] (n = 0-3) and the subsequentsynthesis of the bis(benzene) derivative [Ru,C(CO), ,(q6-C,H,)(p4-qL : q 2 : q2-C,H,)].2 The observation that in thislatter interesting compound the benzene molecules adopted notonly the familiar q6 terminal bonding mode found in e.g.[Cr-( q6-C,H,,)J. but also the then highly unusual p3 face-cappingmode, led us to investigate the chemistry of arene clusters inmore detail.Over the recent past we have developed a systematic anddiverse chemistry of these materials, not only based on Ru,Cunits, but also for a variety of clusters containing from three toeight nictal atonis. For us, basically four patterns of interesthave emerged, u i z , (i) the cluster-surface analogy, (ii) thephotochemistry of arene clusters and C-H bond-cleavagereactions, (iii) graphitic-like interactions in the solid stateleading t o chains, snakes and ribbons and (iv) the generation ofmonomer and dimer precursors to conducting polymers. Thesethemes are not independent; they rely heavily on one another.Nevertheless, each provides an excellent research area in its ownright and undoubtedly other related areas of interest willemerge.The Cluster-Surface AnalogyThe idea behind these investigations emerged naturally fromour studies of [Ru,C(CO), l(qh-C6H6)(p3-q2 : q2 : q2-C6H6)].Our discovery of this compound corresponded with reports bySomer-jai and others of the interaction of benzene with theRh( 1 1 1 ) surface both in the presence of differing amounts ofCO and in its absence.It seemed apparent to us that ;i clearrelationship existed between our benzene-carbonyl cluster andthe benzene -carbonyl-surface interaction (Fig. 1). The feelingthat our systems, which were so easily studied by the usualvibrational (IR and Raman) spectroscopic techniques, NMRspectroscopy and diffraction methods (X-ray and neutron),might enable a clearer understanding of the surface chemistryled us to contemplate the synthesis of simple model compounds.In the limiting case, the single-crystal metallic surface may beregarded as being composed of approximately close-packedplanar arrays of spheres extending infinitely in two dimensions.A consequence of this long-range periodicity is that the localatomic and electronic structure will be influenced by moredistant atoms in both the surface and the bulk of the metal.\ 1.81(15)A1.33( 15) ASurface(4Ru1.\,48(0:2) A1.39(2) AClusterFig.1 Comparison of the bonding geometries for benzene on ( a ) theRh( 1 1 1) metal surface and ( h ) the hexaruthenium cluster,[Ru6C(C0)1 l(~6-C6H6)(~,-C6H6)1Hence, adsorption of CO at one site will influence thecharacteristics of sites adjacent to (or even removed from) it.Recent advances in the dynamic theory of low-energy electrondiffraction (LEED) by ordered absorbate overlayers has ledto the structural characterisations of several metal surface-J. Chem.SOC., Dalton Trans., 1996, Pages 2395-2402 239benzene complexes by Somerjai, van Hove and c o - ~ o r k e r s . ~There are at least five possible sites that chemisorbed benzenemay adopt on a close-packed atomically flat metal surface (Fig.to motion in the bulk. In this paper some of our efforts made inresponse to these challenges are reported.2).These are the ‘on-top’ site which has six-fold symmetry, the‘hollow’ site with three-fold local symmetry and the ‘bridge’ sitewith two-fold symmetry. In each overlayer benzene ischemisorbed intact and lies parallel to the close-packed metalsurface. Adsorption at a three-fold hexagonal close-packed(h.c.p.)-type site is clearly related to the p, bonding modeboth cases the C6 ring is expanded showing in-plane Kekuledistortions: C-C bond distances alternate between 1.3 1 ( 1 5 ) and1.83( 15) A, the short bonds lying above single atoms while thelong C-C bonds form bridges linking pairs of metal atoms.Two-fold bridge-site occupancy has been substantiated on aPt( 11 1) structure with a distortion in molecular symmetry toC21,. For us, the challenge to produce model cluster systemswhich correspond to these observations was raised.The metal skeletons of high-nuclearity carbonyl clustercompounds are frequently structurally comparable to fragmentsof the bulk metallic lattice, e.g.[Rhl3(CO),,H,l2-,[ O S ~ ~ C ( C O ) ~ , ] ~ - and [Rh,,(C0)2,]4- may be recognised asfragments of h.c.p., cubic close packed (c.c.P.) and body-centred cubic (b.c.c.) structures re~pectively.~ The series ofosmium clusters based on four, ten, twenty and thirty-five metaland [ O S ~ ~ ( C O ) ~ ~ ] ~ -, follow the C.C.P. growth sequenceprecisely. Significantly, in the dianion [OS~,(CO),,]~ -, eachface of the cluster corresponds directly to the metal(ll1)surface. Smaller carbonyl clusters are typically deltahedra andmany of their metal-core configurations may be regarded asmicroscopic fragments of common close-packed lattices.However, the role of the ligand sphere in stabilising these bulk-like geometries should be appreciated, as should the case inwhich they undergo geometrical transformation as the overallelectron count is modified, e.g.monocapped trigonal bipyramidin [os6(co)1s] (84e) to octahedral in [Os,(CO), ,i2- (86e).Structural comparisons between the chemisorbed state andmetal clusters are best made when there is a correspondence inthe metal, ligands, ligand coverage of the cluster or surface andin the crystallography. Detailed structural analysis of the gas-metal interface is currently only feasible for metal crystalshaving well defined flat surfaces.For chemisorption on theselow-Miller-index surface planes the metal-ligand and -metalconnectivity is respectively smaller and larger than for clusters,although on higher planes the metal co-ordination numberdecreases at irregular features such as step and kink sites.Thus, our challenges as molecular chemists may be describedas: ( a ) the production of systems containing arenes in bondingmodes related to those supposed to exist on the surface; (b) theinvestigation of the change in electronic character and bondingmode of the arene as the number and type of metal atomschanges; (c) the monitoring of related changes in bonding as afunction of ligand type and ( d ) the exploration of the migratorypatterns that exist on the cluster surface and their relationshipadopted in [RU6C(CO)1 1(?16-C6H6)(j.13-q2 : q2 : q2-C6H6)].Inatoms, [Os,H,(CO)i21, [ ~ ~ ~ o ~ ( ~ ~ ) ~ ~ ] ~ - ~ [os~o(co)~o12-The model compounds [ M3(CO),(p3-$ : qz : q2-C6H,)] (M = Ruor Os)]The simplest model of the [l 1 11 surface is the metal triangulo-arrangement found in the simplest clusters [Ru3(CO) 12] and[OS,(CO),~]. In this work we took advantage of this fact andexplored routes to benzene-carbonyl models. We were luckythat our first attempted syntheses proceeded moderately easilyand in favourable yields. The best routes to the rutheniumand osmium compounds [M3(CO)9(p3-q2 : q 2 : q2-c6H6)] are,however, different. For ruthenium, treatment of [Ru3(CO), 2]in dichloromethane with Me3N0 in the presence of an excess ofcyclohexa- 1,3-diene leads to the required product in a single-stage process.The corresponding osmium derivative may beprepared in a related fashion (Scheme l), but is best obtainedfrom the very reactive unsaturated cluster [OS~(~.~-H)~(CO)(Scheme 2 ) . 2 The molecular structures of the two compoundsare, as expected, very similar.2’6 In each case the central metaltriangle supports the p,-bonded benzene and nine CO ligands.Owing to the greater precision of the ruthenium structure, theC-H vectors are seen to bend away from the C, plane and Ru,unit, in agreement with the Somerjai model. In addition, in bothcases, the C, ring shows the expected Kekule distortion. Clearlythese compounds are good model systems. Detailed vibrationalanalysis of both these cluster and surface-adsorbed systems iscurrently under investigation.Many other ‘model’ systems have now been prepared anddetailed synthetic routes to clusters based on M,, M,, M6, M,and M, units have been devised.The Ru6C system provides agood example of the synthetic routes employed and also theversatility of the benzene ligand, which may bond to produce avariety of geometric forms (Scheme 3 ) . , One of our primaryobjectives in this aspect of the work has been to increase thenumber of benzene or related molecules supported by thecluster unit, leading ultimately to a system containing onlybenzene on the surface.In our synthetic programme essentially the same approachhas been employed for all systems under investigation.Theinitial cluster is activated by oxidative removal of CO as CO, byreaction with Me3N0, followed by addition of cyclohexa- 1,3-diene which usually bonds first through a single double bondand then through the 1,3-diene unit. By this means complexescontaining six-membered rings bonded to central cluster unitshave been isolated and characterised.The bonding mode adopted by the cyclohexadiene unit inthese systems is not always the same.’ For the Os, system wehave good evidence to suggest that two forms of the compound[Os,(CO),,(C6H,)] exist (Scheme 1). One, for which themolecular structure has been established by single-crystal X-rayanalysis, with the diene bonded to one osmium atom. Thiscompound is reluctant to undergo conversion either to thedienyl derivative by single C-H bond cleavage, or to thebenzene derivative via the cleavage of two C-H bonds.Incontrast, the second isomer, which we suppose contains thediene bridge bonding to two osmium atoms, readily undergoesconversion directly into [Os3(CO),(p3-q2 : q2 : V2-c6H6)]. Thisobservation is significant, indicating that C-H bond cleavage,leading to a face-bridging arene, is assisted by multipleinteractions with the cluster surface in this case. However, thiscannot be the whole story, since in our studies of the Os,systems (Scheme 4) we have been able to isolate andcharacterise the complete range of intermediate compoundsleading to the final benzene derivative., It is important to note,however, that here the hexadiene molecule bonds to only oneosmium atom and benzene apparently adopts only the q6terminal bonding mode.Similar behaviour is noted with theCO, systems where again only the q6 mode for benzene isc3v c 6 V czvFig. 2surfacePossible adsorption sites for benzene on a close-packed metal2396 J . Chem. SOC., Dalton Trans., 1996, Pages 2395-240(ii)Scheme 1diazabic yclo[ 5.4.01 undec-7-ene (dbu)Synthesis of [OS~(CO)~(JL~-C~H~)] from [Os,(CO), 2]. (i) C,H,-CH,Cl,; (ii) octane, heat; (iii) [Ph,C][BF,] followed by 1,s-1 (ii)1+(iii) -~ ~ ~ 3 ~ ~ ~ ~ 9 ~ P 3 - T \ 2 : ~ 2 : ~ 2 - ~ g H 6 ~ 1 [OS3(P-H)(C0)9(P3-~2:~2:~z-CgHg)l+Scheme 2 Alternative synthesis of [OS~(CO)~(JL~-C~H,)] from [Os3H2(CO),,]. (i) C,H,-octane, heat; (ii) [Ph,C][BF,]; (iii) dbuobserved.' A feature of these Co, and Os, tetrahedral systemsis the relative ease with which the terminally bonded areneundergoes exchange with an unco-ordinated arene.For bothsystems this exchange provides a highly convenient method ofproducing different arene cluster derivatives. However, thisbehaviour is not general and systems based on the Ru,C unitare reluctant to undergo exchange.In most other systems (e.g. Scheme 3) the diene is found tobridge one edge of the cluster polyhedron and on treatmentof these complexes with Me,NO, or on pyrolysis, smoothconversion into the appropriate benzene derivative isobserved.'For the Ru,C series of compounds, benzene (and otherarenes) adopt both p3 and p bonding configurations.* With[Ru,C(CO),,(arene)], apart from the derivative derived fromC2.2lparacyclophane (see below), all arenes adopt the terminalJ.Chem. Soc., Dalton Trans., 1996, Pages 2395-2402 239(ii) +++ +t (ii) I (iii)[R&$C(CO)i I (rl 6-c&6)(P 3-C6H7)1*Scheme 3 Synthesis of benzene derivatives of [Ru,C(CO), ,I. (i) Me,NO-C,H,; ( i i ) Me,NO; (iii) hexane, heat: (iv) [Ph,C][BF,]; (v) dbubonding mode and no evidence for face bridging has beenfound.” This is not the case for the bis(benzene) derivative,[Ru,C(CO), l(C,H6)2], which may be easily prepared from[Ru,C(CO), ,] using Me,NO-l,3-C,H8 in a sequentialmanner. Here several different isomeric forms have beenisolated. Significantly, these forms undergo exchange but thebarriers to isomer interconversion are relatively high.Themechanism by which isomerisation occurs has not beenestablished, but we consider that the most likely route involvesan intermediate in which the C , ring spans a polyhedral edgeand is co-ordinated through two ‘allyl’ fragments within it.Although direct examples for this mode have not been foundfor the simple benzene derivatives, a compound containingbenzene bound in this manner has been observed by others in[Rh,(q-C5H5),(p-q3 : q3-C6H,)] l 2 and we have fully charac-terised a related cyclophane derivative, [Ru,(CO),(p-q : q3-In general, the benzene compounds described so far cannotbe prepared directly from the reaction of benzene with theappropriate cluster. The Co, system provides a rare example,where direct reaction does occur with either [Co,(CO),] or[CO,(CO),~] to produce [Co4(CO),(C6H6)].” Interestingly, wehave noted that in the reaction of [RU,(CO)~,] with Me,NO-1,3-C,H8 the presence of benzene greatly enhances the rate,although we have not established its role., Benzyne, rather thanbenzene, derivatives of Os,, e.g.[oS3(p-H)2(Co),(c6H,)1,are formed directly on reaction of the activated cluster[OS,(CO)~,(M~CN)~] with C,H6; no evidence has been foundthat [Os,(CO),(C,H,)] is an intermediate in this reaction, butwe have discovered that on irradiation or prolonged heating[Os3(CO)9(C6H6)] is converted into the same benzynederivative (see below). l 4C16H16)] (Fig. 3).13Electronic character and change in bonding mode of areneA number of interesting features have emerged but we remain along way from truly understanding the factors which govern thebonding mode adopted by arenes on the cluster surface.Apartfrom [Ru2(CO),(p-q3 : q3-C1 6H 16)] we have no direct evidencefor the bis(ally1) bonding mode (Fig. 3) although there is limitedevidence to suggest the formation of unstable edge-bridgedintermediates in the p3 to q’ conversion process.The q‘ terminal bonding mode is commonly observed for allcluster nuclearities and seems to be independent of the metal-metal connectivities within the clusters, involving a connec-tivity (M-M) of two in [M,(CO),(R,C2)(C,H6)],16 threeC(C0)12(C,H,)]16 (basal isomer 1 ) and four in [Ru,C-(CO),,(C,H,)] l 7 (apical isomer2) and [Ru,C(CO),~(C,H,)].’Migration of the arene over the cluster surface to producedifferent isomeric forms may be induced chemically l 6 orthermally.I 7 * l 8 At present there appears to be no correlationbetween the number of carbonyl groups also present and thearene bonding type. For systems containing more than onearene moiety, as in [Ru,C(C0),,(C6H6),], we detect a slighttendency for face bonding to be preferred, but this may be moreapparent than real!One exciting observation which has arisen from our work isthat the CH chemical shift in the ‘H NMR spectra of the rangeof q6-bonded benzene derivatives [Ru,C(CO) 12(q6-C6H6)] (6(q6-C6H,)] (6 5.40) undergoes a systematic variation as afunction of the number of metals in the cluster unit.lgAlthough great care should always be exercised in anycorrelation of this type, it is reasonable to assume that the shiftin [Os,H,(CO),,(C6H6)],8 [CO4(CO),(C6H6)] and [RU.j-5.93), [Ru6C(CO),,(~6-C6H6)] (6 5.56) and [Ru,H,(CO),,-2398 J.Chem. SOC., Dalton Trans., 1996, Pages 2395-240+(iii),,F'Scheme 4(iii) heat, hexane; (iv) 3.2 equivalents Me3NO-CH,Cl,-C,H,Synthesis of some benzene derivatives of [Os,H,(CO),,]. (i) 3.2 equivalents Me,NO-CH,Cl,-1 ,3-C6H,; (ii) CH,Cl,-l ,3-C,H8;Fig. 3 Molecular structure of [Ru,(CO),(p-q3 : q3-Cf6HI6)]reflects the change in electron density available for bondingon the cluster surface and, in any event, appears to be a moresensitive probe than e.g. the CO stretching vibration in carbonylclusters.A similar variation is noted for the 'H resonances of the tworings in face-capping [2.2]paracyclophane derivatives (seebelow).20 In this case the degeneracy of the two rings is removedon co-ordination and the chemical shift observed at 6 6.51 forthe parent ligand is seen to separate.As the number of metalsincreases from three to six the separation also increases; for[Ru3(CO),(p3-q2 :q2 :q2-c16H16)] the values are 6 3.76 andthe values are 6 3.43 and 7.44 (A = 4.01). In this case the7.22 (A = 3.46) and for [RU~C(CO),~(~L,-T~~: q2 :q2-C16H16)]bonding between the co-ordinated ring and the cluster affectsits p, interactions with the other ring. This is a particularlysensitive probe of the cluster surface and it would beinteresting to examine the interactions of cyclophane with thebulk surface.Bonding as a function of ligand typeOur investigations in this area have been limited to date.Undoubtedly the presence of CO on the cluster surface affectsthe nature and type of bonding mode adopted by the arene.Inone key experiment we have observed that reaction of thetriangulo-clusters [M3(CO)9(p3-q2 : q2 : q2-C,H6)] (M = Ruor 0s) with alkynes R2C2 leads to the displacement of thebenzene to a terminally bonded mode and the concomitantbond formation of the alkyne in the familiar 20, n: mode acrossthe triangular face to form the complex [M3(CO),(R2C,)(q6-C&6)]. l 6 For the series of compounds [Os,(CO),L(q6-C6H6)] (L = C2H4, PR,, etc.) the barrier to rotation of theC6H6 moiety is found to vary but not by very much.21 Atpresent we are studying the variation in the 'H chemical shiftfor a series of compounds in which a CO ligand is replaced by aseries of phosphine ligands with differing basicities.Migratory patterns of the arene ligand over the cluster surfaceAs we have commented previously, a range of different isomericforms have been observed for the series of bis-substitutedcompounds [RU,C(CO)~ l(c6H6)2] (Scheme 5) and themonosubstituted system [Ru,C(CO), 2(C6H6)].In each casealmost every possible isomeric form is observed (except the bis-p3 form!) for [Ru,C(CO), 1(c&)2]. Interconversion is readilymonitored spectroscopically but at no stage have stable formsbeen observed in which the arene straddles the edge.Nevertheless, we believe this to be the most likely intermediate(see above).J.Chem. Soc., Dalton Trans., 1996, Pages 2395-2402 239f&...._ 1 ‘‘Jq6 - bnsalScheme 5q6 - apical cr3Photochemistry of Arene Clusters and C-HCleavage ReactionsWe have observed that on irradiation of solutions of[ O ~ , ( c O ) ~ ( p ~ - q ~ : q2 : q2-C6H6)] in toluene the benzynederivative [Os3H2(CO),(p3-q2 : q1 : q ‘-C6H4)] is produced.Interest in this reaction centres around the activation of thetwo adjacent ring C-H bonds. Although the mechanism bywhich the photochemical reaction occurs has not been fullyestablished, the accepted view is that 0s-0s bond cleavage((J __+ o*) occurs as the initial step. Although related work hasnot been carried out for other triosmium arene compoundscontaining C,H,Me, C,H,Me2-1,3, C6H3Me3- 1,3,5, etc., wehave established that the direct reaction of a wide range ofarenes with [Os,(CO), o(MeCN)2] generates the appropriatebenzyne derivatives.2 2 No evidence for the intermediate form-ation of p3-arene intermediate compounds has been found andit is probable that the reaction takes place in these instancesvia a p-bonded intermediate. Evidence for a reaction of thiskind has been found in our related studies of the reaction ofcyclohexene with [Os,H,(CO) 2] or [Os,H,(CO), o(MeCN),].Here, totally unexpectedly, cleavage of one of the ‘olefin’ C-Hbonds to generate the cyclohexenyl derivative [Os,H,(CO),-(C6H9)], which undergoes a second C-H bond cleavage onheating to form the ‘yne’ derivative [OS4H2(CO)9(C6H,)], inwhich the C6H, moiety spans the tetrahedral Os, face.Relatedstudies of the corresponding ruthenium systems have led tothe isolation of a series of yne-like complexes based on the[RU,(CO)~ ,(C6H8)] unit but, in this case, presumably becauseof the weaker Ru-Ru bonds, the yne unit straddles a butterflyRu, arrangement.23 Again, olefinic C-H bond cleavage hasoccurred in preference to methylene CH, bond cleavage.Graphitic-like Interactions in the Solid StateMany of the compounds reported in this article have been thesubject of detailed analysis by X-ray diffraction studies onsingle crystals. Several points of interest emerge. For several ofthe monosubstituted derivatives, based on Os, and Ru6C unitswith simple arenes, the organic rings are observed to interlockto produce layers of organic substrate and layers of the clusterunits [Fig.4(a) and ( l ~ ) ] . ~ “ In the bis-substituted derivatives ofRu,C the various different isomeric forms give rise to differentcrystallographic packing arrangements. Thus, in trans-(q6-C&)2] the rings from adjacent molecules in the crystal aresuperimposed albeit with some displacement and the inter-ringdistances are compatible with those observed in graphite. Inthese examples, interaction between the rings occurs throughout[RU6C(CO)1 l(q6-C6H3Me3-1,3,5)2] and tranS-[RU6C(CO)11-nnx oWFig. 4 Schematic representations of the molecular organisation inthe solid, leading to molecular ‘chains’.In the related system cis-[Ru,C(CO), l(q6-C6H6)2] the interactions are similar butbecause of the distribution of the arenes about the Ru6C clusterunit the ‘graphitic-like’ interactions lead to molecular ‘snakes’throughout the solid. For the system [Ru,C(CO), ‘(q6-C6H6)(p3-C6H6)] the q6 rings on adjacent molecules in thesolid interact in a graphitic-like manner as do the adjacent p3-ring systems. This also leads to molecular ‘snakes’ [Fig. 4(b)].These interactions are of fundamental importance, particularlysince they provide a mechanism or route by which charge2400 J. Chem. SOC., Dalton Trans., 1996, Pages 2395-240transfer may occur throughout the lattice in well designatedpathways. Very recently, using ultraviolet laser desorption massspectrometry, we have recorded direct evidence for suchaggregated species in the gas phase directly from neutral benzenemolecular clusters.In an effort to extend the studies of thisphenomenon we are currently investigating the formation ofarene clusters containingmuch larger polycyclic systems. It is ourintention to produce organometallic cluster materials in whichring-ring interactions are maximised. To date we have prepared[Ru,(CO)~(C, ,HI0)] which shows the desired effect.25Monomer and Dimer Precursors to ConductingPolymersFollowing on from these studies (see above) we have synthesiseda whole range of C2.2lparacyclophane cluster derivatives.” AtFig. 6 Molecular structures of [RU,C(CO)~,.(~~-P~(CH~),P~}][n = 1 (a) or 2 (b)]the onset of these studies it was our intention to producemonomers with one cyclophane bonded to the central unit andthen dimers in which the cyclophane moiety is bondedsimultaneously to two separate cluster units and thereby servingas a bridge allowing each separate cluster unit to exchangeelectron density.Ultimately, our goal is to generate polymericchains of such materials which hopefully will serve asconductors or semiconductors. We are also interested inproducing clusters containing two cyclophane units which packin chains, snakes or ribbons as with the simpler arenes, butthrough which charge transfer may be mediated. Examples oftwo C2.2lparacyclophane clusters which have been preparedand characterised are shown in Fig. 5. Several general featuresappear.First, the cyclophane ligand shows a marked tendencyto function as a face cap rather than as an q6 terminal group.However, this is not always the case and our attempts toprepare [Ru6C(CO), l(C16H16)2], the first cluster with twoface-capping rings, have resulted in the q6, p3 complex.Secondly, as commented above, the ‘H NMR spectrum ofthe co-ordinated C, 6H16 ligand shows highly significant ‘Hchemical shift values which appear to be sensitive to the numberof metal atoms within the cluster. Finally, we have observedthat in the compound [Ru~C(CO),,(~~-C,~H,,)(~-C,H,)] it isthe C atoms of the co-ordinated ring that sit directly above themetal atoms of the triangular face and not, as in all other cases,the C=C bond. Significantly, major changes in the configurationof the cyclophane rings are also observed.In related studies we have prepared derivatives of linkedarene systems such as Ph(CH,),Ph ( n = 0-2) 26 and PhCH=CH-J.Chem. Soc., Dalton Trans., 1996, Pages 2395-2402 240CH=CHCH=CHCH=CHPh. ’’ Again, our objective is simplyto produce cluster systems linked by an organo bridge andultimately to produce organometallic cluster materials. Themolecular structures of the series of compounds [Ru,C-(CO),,(Ph(CH,),Ph)] have been established and a couple areshown in Fig. 6.AcknowledgementsWe thank the EPSRC and NATO for financial support. P. J. D.would like to thank The Royal Society for a UniversityResearch Fellowship.References1 B. F. G. Johnson, R. D. Johnston and J.Lewis, Chem. Commun.,2 M. P. Gomez-Sai, B. F. G. Johnson, J. Lewis, P. R. Raithby and3 G. A. Somerjai, J. Phys. Chem., 1990, 94, 1013.4 E. L. Muetterties, T. N. Rhodin, E. Baud, C. Brucker andH. Pretzer, Chem. Rev., 1979,79, 91 and refs. therein.5 A. J. Blake, P. J. Dyson, B. F. G. Johnson, C. M. Martin,J. G. M. Nairn, E. Parisini and J. Lewis, J. Chern. SOC., DaltonTrans., 1993,98 1.6 B. F. G. Johnson, J. Lewis, M. Martinelli, A. H. Wright, D. Bragaand F. Grepioni, J. Chern. Soc., Chem. Cornmun., 1990, 364;D. Braga, F. Grepioni, B. F. G . Johnson, J. Lewis, C. E. Housecroftand M. Martinelli, Organometallics, 1991, 10, 1260.7 P. J. Dyson, B. F. G. Johnson, J. Lewis, M. Martinelli, D. Braga andF. Grepioni, J . Am. Chem. Soc., 1993,115,9062.8 B.F. G. Johnson, A. J. Blake, C. M. Martin, D. Braga, E. Parisiniand H. Chen, J. Chem. SOC., Dalton Trans., 1994,2167.9 I. V. Khand, G. R. Knox, P. L. Pauson and W. E. Watts, J. Chem.Soc., Perkin Trans. I, 1973,975.10 D. Braga, F. Grepioni, E. Parisini, P. J. Dyson, A. J. Blake andB. F. G. Johnson, J. Chem. SOC., Dalton Trans., 1993,2951.1967, 1057; J. Chem. Soc. A, 1968,2865.A. H. Wright, J. Chem. Soc., Chem. Commun., 1985, 1682.11 P. J. Dyson, B. F. G. Johnson, D. Reed, D. Braga, F. Grepioni and12 J. Miiller, P. E. Gaede and K. Qiao, Angew. Chem., Int. Ed. Engl.,13 A. J. Blake, P. J. Dyson, B. F. G. Johnson and C. M. Martin,14 M. A. Gallop, B. F. G. Johnson, J. Lewis, A. McCamley and15 P. J. Dyson and B. F. G. Johnson, unpublished work.16 D. Braga, F. Grepioni, B. F. G. Johnson, J. Lewis, M. Martinelliand M. A. Gallop, J. Chem. Soc., Chern. Commun., 1990, 53;D. Braga, F. Grepioni, B. F. G. Johnson, E. Parisini, M. Martinelli,M. A. Gallop and J. Lewis, J. Chem. SOC., Dalton Trans., 1992, 807.17 D. Braga, F. Grepioni, P. Sabatino, P. J. Dyson, B. F. G. Johnson,J. Lewis, P. J. Bailey, P. R. Raithby and D. Stalke, J . Chem. Soc.,Dalton Trans., 1993,985.18 D. Braga, F. Grepioni, E. Parisini, P. J. Dyson, B. F. G. Johnson,D. Reed, D. S. Shephard, P. J. Bailey and J. Lewis, J. Organomet.Chem., 1993,462,301.19 B. F. G. Johnson, C. M. Martin, D. Braga, F. Grepioni andE. Parisini, J. Chem. Soc., Chern. Commun., 1994, 1253.20 P. J. Dyson, B. F. G. Johnson, C. M. Martin, A. J. Blake, D. Bragaand E. Parisini, Organometallics, 1994, 13, 21 13.21 M. A. Gallop, M. P. Gomez-Sal, P. R. Raithby and A. H. Wright,J. Am. Chem. SOC., 1992, 114, 2502.22 B. F. G. Johnson, J. G . M. Nairn, D. B. Brown, J. Lewis,M. A. Gallop and D. G. Parker, Chern. Eur. J., 1995, 1,252.23 C. M. Martin and B. F. G. Johnson, unpublished work.24 D. Braga, F. Grepioni, P. J. Dyson and B. F. G. Johnson, J. Cluster25 J. G. M. Nairn and B. F. G. Johnson, unpublished work.26 A. J. Blake, B. F. G. Johnson, D. Reed and D. S. Shephard, J. Chem.27 D. S. Shephard and B. F. G. Johnson, unpublished work.E. Parisini, J. Chem. Soc., Dalton Trans., 1993, 2817.1993,32, 1697.J. Chem. Soc., Chem. Commun., 1994, 1471.R. N. Perutz, J . Chem. Soc., Chew?. Commun., 1988, 1071.Sci., 1992,3, 297 and refs. therein.Soc., Dalton Trans., 1995, 843.Received 18th January 1996; Paper 6/00428H2402 J. Chem. SOC., Dalton Trans., 1996, Pages 2395-240