GENERAL INTRODUCTIONFrom Ferrocene to UranoceneBY G. A. GAMLENImperial Chemical Industries Ltd., Petrochemical and Polymer Laboratory,P.O. Box 11, The Heath, Runcorn, CheshireReceived 28 th April, 1969It is characteristic of Faraday Society Discussions that they are held at opportunemoments in the development of the subject. This one is no exception becausemetallo-organic chemistry has reached that particular stage where considerablesimplification is possible through the replacement of facts by concepts. Thus wemay contemplate a return to the position in which chemistry can be regarded asa unified whole with metallo-organic chemistry occupying its place as the bridgebetween organic and inorganic chemistry. This will only be realized, however, ifour theory of bonding is sufficiently good to account for the situations found at theextremes of organic and inorganic chemistry and cast in a sufficiently simple formfor it to be readily comprehensible and profitable to use.There is no doubt that molecular orbital theory can be refined to include allinteractions between electrons and nuclei but unfortunately such calculations requirethe services of a large‘ computer and cannot be considered as handy for the averagechemist.The necessity of this second requirement for success was recognized someyears ago by a number of theoretical chemists and especially by 0rgel.l For thisreason, ligand field theory was devised in which the considerable simplification thatwas made depended upon some severe approximations regarding the contributionof the ligand to the bond.This theory has been of the greatest help in increasing our understanding of thebonding in complex compounds, particularly of those properties which are associatedmainly with the metal ion.With increased understanding, however, there came alsoa fuller realization of the limitations imposed by the approximations used especiallyas regards the state of the ligand once the bond to the metal had been formed. Avaluable task for the theoretical chemists is therefore to look for closer approximationsto molecular orbital theory without sacrificing the utility of ligand field theory. Asexamples of this approach we have the simplified model for octahedral metal carbonylcompounds proposed by Anderson and Brown2 which nevertheless gives a goodrelationship between force constants and ligand characteristics, and Hillier and Mrs.Canadine’s paper using the SCCEI method.The results in this case can now betested by new experimental techniques such as photo-electron and n.m.r. spectro-scopy, which will considerably increase their value.The present-day preoccupation with the role of the ligand springs in part fromthe unusual organic reactions which can be catalyzed by metallo-organic compoundsand which are of increasing industrial interest. To deduce the mechanism of thesecatalytic reactions requires a much deeper understanding of the nature of the metal-ligand bond than we have at present and is hampered by the inadequate and out-mode8 FROM FERROCENE TO URANOCENEformalism with which we try to convey our ideas.For example, the arrow used torepresent the donation of n-electrons from an olefin to a metal,R R\ /CCI1 -,MR/ \Ralthough describing how the bond is formed gives no indication of the nature of thelinks (especially in the olefin) once it has formed.It is clear that there is room for improvement of various terminological aspects ;a number have already been suggested and of these Cotton's hapto nomenclatureparticularly deserves mention as it would remove, inter alia, the confusion surroundingthe symbol 71. This is used indiscriminately to describe complexes in which thebonding is at least partly by non-linear orbital overlap as well as complexes in whichthe ligand donates n-electrons to the metal in a c-bond.Perhaps it is time for aresponsible body such as this Society to suggest to the International Union of Pureand Applied Chemistry that a commission should be set up for a general reform ofmetallo-organic nomenclature-r should it be organo-metallic ?In addition to the many other reasons why this meeting is opportune there is,for example, the realization that this is a huge area of chemistry which potentiallynumbers as many compounds in its fold as does organic chemistry. It also comes ata time of rapid expansion in the number of physical techniques available to study thenature of the bond and when the existing techniques are being so improved that theyallow much more stringent tests to be applied to theoretical explanations.Some of these techniques unlock physical properties which were hitherto virtuallyinaccessible ; for example, the electron density in different orbitals around the nucleusof certain elements can be deduced from Mossbauer effect spectra. These have beenused by Greatrex and Greenwood to make structural deductions about a largenumber of substituted iron carbonyls and by Bancroft, Mays and Baxter tocorrelate, among other things, the partial centre shift values of various ligands totheir bonding characteristics.This is a development which is sure to be followedclosely by all chemists interested in bonding.Another powerful technique of great potential which is quite new to the organo-metallic field is photo-electron spectroscopy.' This reproduces directly the electroniclevel structure from the innermost shells to the atomic surface.It is a characteristicfeature of this branch of spectroscopy that chemical structure effects cause shifts ofthe inner levels and thus provide information on chemical bonding. The spectrashown below (fig. 1) illustrate this for platinum and potassium chloroplatinate whileEvans, Green, Green, Orchard and Turner and Braterman and Walker havediscussed the bonding in pentacarbonyl manganese derivatives.It is not possible in a short introduction to discuss all these exciting new techniquesindividually but perhaps one typical example may be quoted : the existence of a cis-effect has been suspected for many years in square planar palladium and platinumcomplexes.Fryer and Smith lo have recently demonstrated its reality in trans-phosphine palladium complexes by estimating the charge on the chlorine atoms fromthe nuclear quadrupole resonance spectra. It is clearly significant that the order ofincreasing cis-effect of the ligands is the reverse of their trans-effectG . A . GAMLEN 9These advances in physical techniques are particularly welcome at the presenttime because they provide an insight into quite complex electronic interactions ofligands which may arise either through the central metal atom or by a field effect.A theoretical treatment, to be satisfactory, therefore has to consider the molecule asa whole ; yet metallo-organic molecules have becomesteadily more and more complex as time goes byand provided bonding problems of increasing com-is one of many possible examples to illustrate thiscomplexity and has been chosen because in additionto well-recognized forms of bonding it also has ametal atom [Fe(B)] which is acting as both a donorand accept0r.lThe polynuclear " cluster " compounds, whichmay show multiple bonding of a high order,12 alsoexemplify this requirement to treat the molecule asa whole and raise the question of whether metalatoms joined together in a cyclic system bear asimilar relationship to their individual parts as, forexample, benzene does to six individual carbonatoms.The compound Rh,(C0)16 which wasoriginally discovered in 1943 and whose true naturewas not revealed for twenty years when its structurewas determined by Corey, Dahl, and Beck l3 illus-trates a number of such points, (fig.3). As wellas being the first example of a hexanuclear metalcarbonyl it was also the first polynuclear metalcarbonyl recognized to contain a carbonyl groupbridging three metal atoms. Significantly, it is alsoan exception to the inert gas rule in that the entiremolecule contains two electrons in excess of the'' xenon '' configuration for each rhodium atomplexity. The compound Fe2(C0)6C6H~02, (fig. 2)KINETIC ENERGY - I I 1CV 80 75 70RINnlNC. FNFRGYFIG. 1 .-Electron spectra from metal-lic Pt and KzPtCls showing shifts ofthe platinum NVI and NVII levels.and it has been suggested that this electron pair is strongly delocalized with low-lyingexcited states.I OH c9FIG.2.-Structure of Fe2(C0)6(C6H802)10 FROM FERROCENE TO URANOCENESimilar bonding problems arise in the unusual compounds containing a singlecarbon atom as ligand such as Fe,(CO),,C, in which the carbon atom is equidistantfrom the five iron atoms and therefore penta~o-ordinated,~~ (fig. 4), and the rutheniumarene complexes such as R U ~ C ( C O ) ~ , ( C H ~ ) ~ C ~ H ~ in which the carbon atom occupiesa6FIG. 3.-Structure of Rh6(CO)16.the centre of the octahedron of the metal atoms, (fig. 5). This compound, which wasdescribed by Johnson, Johnston and Lewis l5 and whose structure was determinedby Mason and Robinson,16 illustrates again the unexpected novelty of many of themetallo-organic compounds discovered in the past few years.Organo-metallic compounds are fascinating not only in their structural " many-sidedness " and complexity but also in the instability arising therefrom.Just astheory over the past ten years has had to be refined to provide a more dynamicelectronic description of the metal-carbon bond, so it now has to cope with a muchmore dynamic situation at the atomic level if we are to understand the behaviourof fluxional molecules l7 and n-ally1 complexes. This then is the background againstwhich we can trace the development of theory.(who also conceived theuseful rule-of-thumb " inert gas rule '') and then note that valence bond theory l9We may start from the lone pair concept of SidgwicG. A . GAMLEN 11represented a very real advance because it drew attention to the importance of d-electrons in the bonding of transition metal complexes ; by concentrating on bondingelectrons it satisfactorily explained the stereochemical problems then facing chemists.Its shortcomings were demonstrated, however, when improved methods of magneticThe Coordinate Link (Sidgwick 1927)Crystal F i e l d Theory(Bethe 1929, Van VleckValence Bond Theory(Pauling, 1935)Bonding e l e c t r o n s ,stereochemistry. Non-bonding e l e c t r o n s ,s p e c t r a , magnetism.Molecular Orbital Theorya-complexes of 2 /\ block n- complexes(Chatt and Shaw, \ 1959) (Dswar, Chatt, 1951)\/’\ ,\ , I\ /Metal Flho<oalkyls(Cotton, 1964)\FIG.612 FROM FERROCENE TO URANOCENEsusceptibility determination became available.Further, it was quite unable tocontribute an explanation for electronic spectra, although both the magnetism andelectronic spectra could be interpreted in terms of crystal field theory as developedby Van Vleck,20 Schlapp and Penny 21 and Jordahl,22 and in which the emphasiswas placed on the non-bonding electrons. These essentially electrostatic approaches,however, did not deal with the electron delocalization known to occur, as was shownby the work of Owen and Stevens 23 on the e.s.r. spectrum of ammonium hexachloro-iridate(1V). However, as early as 1935, Van Vleck 24 had pointed out that, if partlycovalent bonding was assumed, both m.0. theory and crystal field theory predictedthe same result for the splitting of metal d-levels into two subshells.It was thereforepossible to reconcile these approaches and various quantitative treatments havebeen proposed starting with the first general treatment of octahedral complexes byTanabe and Sugano 2 5 in 1954. Since then the molecular orbital theory contributionsby Cotton,26 Gray, Ballhausen 27 and Jarrgensen 28 have been particularly important.Some of the main steps in the historical development of the theoretical side areillustrated in fig. 6.While theoretical organo-metallic thinking has evolved in this way, an immensegrowth of synthetic techniques and materials has occurred. A major feature on theexperimental side has been the increase in the number of different kinds of moleculeable to act as ligand, including such unrelated entities as carbenes, carbollides andeven nitrogen itself.To reduce this richness and variety of ligand to manageable proportions it hasbeen necessary to set up a system for classifying ligands according to type.Onesuch scheme is shown in the following table.29TYPES OF LIGANDLII I(1) with one'or morefree electron pairs(2) no free electron pairs butwith n-bonding electrons,e.g., ethylene, cyclo-pentadienyl ion.I(c) with additional 71-I I(a) no vacant orbitals toreceive electrons from orbitals that can receivethe metal, e.g., HzO, n-electrons from the metalNH3 e.g., PR3, CN- orbitals, e.g.,OH-, C1-.The hydrocarbyl ligards found in Class 2 may then be sub-classified again accord-(b) vacant or vacatableelectrons that can bedonated to vacant metaling to the number of electrons they donate to the metal.TABLE 1.-A CLASSIFICATION OF ORGANIC GROUPS WHICH ACT AS LIGANDS TO TRANSITIONMETALS 30no.of electrons name of class examples of organic group1 Yl akyl or aryl groups2 alkene ethylene3 enyl n-all y 14 diene cyclobutadiene, butadiene5 dienyl n-cyclopentadienyl6 triene benzene7 trienyl n-cycloheptatrienyG . A . GAMLEN 13This table runs up as far as seven electrons donated but could now be extended toten for the di-anion of cyclo-octatetraene which is the ligand in the newly discovered~ r a n o c e n e , ~ ~ a sandwich-type compound like ferrocene.r 1Of the various ligand categories in this classification I propose to dwell briefly onlyon the first and second because they form a rough but useful basis for understandingthe multi-electron systems involving delocalization over a number of carbon atoms.Thus in very simple terms and using valence bond language, one can consider then-ally1 group as a resonance hybrid formed from a combination of a one- and two-electron donor system.Similarly, the cyclopentadienyl ligand is, in these terms acombination of an ally1 with an olefin donor.ONE ELECTRON DONORSStable alkyls of the p-block elements have been known for over 100 years and wereinvestigated so extensively by Frankland in the latter part of the 19th century thathe came to be known as the “ father ” of organometallic chemistry. They aretheoretically derived from the hydrocarbon by replacing an atom of hydrogen by itsequivalent of metal.For the most part, the bonding in these compounds fitted inwith current organic theory and presented no problems ; there is, however, the notableexception from the theoretical standpoint of the “ electron deficient compounds ”which can be exemplified by the dimer of aluminium t r i m e t h ~ l . ~ ~ This has beenshown by an X-ray study to have a bridged structure like aluminium(fig. 7). Each bridging methyl is considered to be held symmetrically to the twow,[CH4FIG. 7.4tructure of A12(CH&. Methyl distances are C-C distances (H atoms have been omittedfor clarity).aluminium atoms by only two electrons which are accommodated in three-centreorbitals formed by the overlap of the A1 sp3 and C sp3 orbitals.It is thereforeappropriate that among the papers to be presented at this meeting is one on bondin14 FROM FERROCENE TO URANOCENEin this and related compounds by Levison and perk in^,^^ which suggests the presenceof a rather stronger metal-metal bond than has hitherto been postulated.It was thought for a long time that the d-block elements were unable to formstable alkyls or aryls ; the discovery that ligands of high ligand-field strength such asFro. 8.the cyclopentadienyl anion, carbon monoxide and tertiary phosphines, stabilize bothtransition metal hydrides and alkyls marked a major step out of this confined thinkingand opened up new possibilities.A notable example of an aryl complex is the ruthenium phosphine complex,Ru(naphthalene)(P-P)2, where P-P = (CH3)2PCH2CH2P(CH3)2 discovered byChatt and D a ~ i d s o n .~ ~ Examination by infra-red and n.m.r. has shown that thecompound in solution is an equilibrium mixture of the arene complex and the 0-arylhydrido complex. An X-ray structure determination which has just been finishedby Kilbourn and co-w~rkers,~~ confirms that in the crystal the compound exists asa ruthenium aryl, (fig. 3).It is obviously important to explain how this stability comes about : Chatt andShaw 37 have advanced the following qualitative suggestion. Any fission of the metal-carbon bond is likely to be irreversible because all organo-metallic compounds arethermodynamically unstable. Such a homolytic fission will occur either by thepromotion of an electron into the anti-bonding orbital of the M-C bond or thepromotion of one electron from the M-C bonding orbital.It then follows that theenergy barrier to dissociation can be related to the energy difference between eitherthe M-C bonding orbital and the lower unfilled orbital or the M-C anti-bondingorbital and the highest filIed orbital, whichever is less. A similar explanation hasbeen put forward to explain in part the exceptionally stable ortho-substituted arylcomplexes.Among other stabilizing systems for these one-electron donor complexes arechelating nitrogen donors, such as dimethylglyoxime and especially the corrin ring ;the stable cobalt-carbon c-bond has the distinction of being the only one known tooccur in Nature, and Green, Smith and Tasker 30 discuss the requirements for stabilityof such a bond in terms of the energies of the 4 2 and d , , orbitals.The propertiesof this type of bond are also discussed by Hill, Pratt and Williams 39 who find thatsuch strong trans and cis-influences are present that the co-ordination number ofthe complex changes with change in donor strength of the ligand; more especiallyG . A . GAMLEN 15they find the alkyl ligand to be a very strong donor. Finally in this section thecuriously high thermal stability of the fluoro alkyl complexes deserves mention.TABLE~.-THE THERMAL STABILITY OF SOME FLUOROCARBON COMPLEXES COMPARED W HTHEIR HYDROCARBON ANALOGUESCZFSMn(CO)S stable up to 150" C2HSMn(C0)4 dec. 25".(C6F5)2Zr(n-CSH5)2 dec.218 (c 6H 5 ) 2Zr(n'C5H5)2 not knowntran~-(C~F~)~Ni[P(c~'H~)~]~ not dec. at 21 3" ~ ~ ~ ~ S - ( C ~ H ~ ) ~ N ~ [ P ( C ~ & ) ~ ] ~ dec. 125"There has been little detailed structural work on these classes of compoundsso far. However, initial measurements on bond length imply that the metal-carbonbonds are very similar in the hydrogen and fluorine complexes. Cotton has notedthat the carbon-fluorine stretching frequency appears to be lower than expected inthe metal complexes and has interpreted this in terms of a n-bonding mechanism.It is suggested the fluorine atoms would lower the energy of the a*-orbital of themetal-alkyl system, enough to permit effective back donation of d-electrons from themetal to carbon.40TWO-ELECTRON DONORSIn turning to the olefin complexes which are formed in such profusion by thed-block elements, we remember that Zeise's so indispensable to all reviewersof organo-metallic chemistry, was discovered only a few years after the platinumammines but fell outside existing theories and so remained anomalous and ignored.The description of the metal-alkene bond as a n-complex involving p-bondingwas established by Dewar 42 for silver-olefin complexes and Chatt and Duncanson 43for platinum-olefin complexes leading to what may now be termed the '' conventional "picture of this bond, the main features of which are shown in fig.9.The usefulness of this explanation is apparent from the fact that it has givensatisfactory service over the past 15 years and is only now being seriously re-considered.As was mentioned earlier, the problem today concerns the nature of the olefin onceit has complexed to the metal.In particular, the formation of the metal-olefin bondFIG. 9.4rbitals used in the combination of ethylene with platinum(II).destroys the n-symmetry of the olefin orbitals and removes the 0-71 inequivalence;mixing of the O- and n-components originally present in the olefin then allows somechange of hybridization from sp2 and sp3 so that the carbon atoms take on somealkyl-type character which is reflected in the bond length and angles of the boun16 FROM FERROCENE TO URANOCENEolefin. However, these changes are not great and it has required the recent improve-ments in accuracy and speed which automatic diffractometry brings to structuredetermination by X-rays before some of these rather small changes in a series ofrelated complexes could be measured with sufficient precision.An investigation ofC'3FIG. 10.-Structure of Zeise's salt.this calibre is described by Ibers44 and co-workers for simple alkene complexes.As they have pointed out, it is remarkable that so few results have been recorded forthis important class of compounds, and for this reason I make no apology in showingyou some unpublished results of Owston and co-workers 45 on a new structuraldetermination of Zeise's salt, (fig. lo).*B BFIG. I1 .-Crystal structure of Zeise's salt, K[C2H4PtCI3] . H20. 0, Pt ; 0, C1; 0, 0; 0, C ; 0, K.The two features of this particular determination are the carbon-carbon distancewhich at 1-37 A is barely significantly different from the distance in the free ligand.The second feature is that although one would expect the platinum-chlorine distancetrans- to the ethylene to be longer than either of the two cis-platinum-chlorine distances*A more recent and closer analysis has led to some changes in the values of the bond lengthsshown in the original diagram.The C-C bond distance is now 1.37 A; Pt-Cl (cis) 1.30 and 1.31while Pt-Cl (trans) is 1.33 A. These changes do not affect the general conclusions of the lectureG. A . GAMLEN 17it is not in fact very greatly different. The reason for this is clear when one examinesthe crystal structure of the complex, (fig. 11) : it is seen that the interaction betweenadjacent atoms in this remarkable array is such as to seriously affect any changes dueto the trans-effect.This emphasizes how careful one has to be in interpreting datafrom the crystal, because of the magnitude of the crystal forces which can distort theeffect one is attempting to observe.One of the interesting features of the bonding model proposed by Chatt andDuncanson was that the olefin would be locked perpendicular to the plane of thecomplex by virtue of the orbital hybridization used to accomplish back donation intothe anti-bonding orbitals of the olefin. The recent application of n.m.r. by Cramer 46to rhodium complexes and by Lewis 47 et al. to platinum complexes has establishedthat in some cases the olefin rotates about the olefin-metal bond (fig. 12) even thoughthe " frozen out " configuration has the olefin perpendicular to the plane of the system.However, the ethylene ligands do not readily exchange with free ethylene and thisrotation can be accommodated in terms of a Chatt-Dewar model.Clearly, the geometry of unsaturated ligands when co-ordinated to transitionmetals is a major point of discussion for this meeting as is the proposition ofMcQueeney, Mason and Tow1 48 that the ligand charge distribution when complexedis virtually identical with that possessed by an isolated ligand molecule in its firstexcited state.The archetypal molecule for this idea is the remarkable compoundPt(CS2)(PPh3)2 described previously by Baird, Mason, Ray and Wilkinson 49 butanother example is furnished by the recent structure determination by Hewitt et aLS0of bis-allenerhodium acetylacetonate which shows that the central C=C=C systemis no longer linear, the angle being reduced to about 148".There is also a significantdifference in the central C=C=C bond lengths, the longer bond being the one thatis formally co-ordinated.It has already been noted that our understanding of the nature of the metal ligandbond has been greatly helped by the development of a whole armoury of physicaltechniques and the papers at this meeting illustrate very comprehensively how thewide range of physical determinations yield accurate data which the " non-specialists "can apply to a whole series of related compounds and purposes.Amongst thesepurposes some mention must be made of the increasing use by the practical chemistof symmetry arguments in, for example, vibrational spectroscopy as well as in simpleU H 1 1C Me A Me C,\A C' \FIG. 12.4lefin orientations in the pIatinum (II) olefin complex, Pt (acac) C1 (cis-but-2-ene).molecular orbital calculations. The power of symmetry arguments is demonstratedby the correctness of the prediction that cyclobutadiene would be stabilized throughcomplex formation, and more recently their use to predict the existence of uranocene.No review of this kind is complete without a glance into the future. It is appro-priate to consider first the central metal atom, and particularly the consequences o18 FROM FERROCENE TO URANOCENEthe very recently discovered uranocene with its unique feature that the n-molecularorbitals of the ring share electrons with thef-atomic orbitals of the uranium.Itspreparation could pave the way for a whole new area of chemistry involving theactinide, lanthanide and rare earths-much as the discovery of ferrocene 18 yearsago set in motion the search for organo-metallic sandwich compounds based onthe d-block elements.As regards the ligand, knowing the state of the ligand once it has bonded isfundamental to predictive metal-complex catalysis in organic processes. Clearly,a great deal of further thinking and experimentation is required here, to take us outof the age of Enlightened Empiricism.Some mention must also be made of the rationalization that becomes possiblenow that the nature of the products of a concerted reaction involving a metal ioncan be predicted by the application of the extended Woodward-Hoffman rules asdescribed by Pettit 51 and his co-workers. This understanding should result in organicchemistry of a new elegance and sophistication-but it too will be only achievedif the nature of the bonding in metallo-organic compounds is better understood.My best thanks are due to Prof. J.Lewis and my colleagues at the Petrochemical& Polymer Laboratory for their help and advice in preparing this Introductory paper.L. E. OrgeI, Report 10th Soluay Council (Brussels, 1956), p. 311. ’ W. P. Anderson and T. L. Brown, this Discussion.I. H. Hillier and Mrs. R. M. Canadine, this Discussion.F. A.Cotton, J. Amer. Chem. Soc., 1968,90,6230.R. Greatrex and N. N. Greenwood, this Discussion.G. M. Bancroft, M. J. Mays and B. E. Baxter, this Discussion.ESCA: Atomic, Molecular and Solid State Structure Studied by Means of Electron Spectroscopyby K . Siegbahn, C. Nordlina, A. Jahlman, R. Nordberg, K. Haunrin, J. Hedman, G. Johansson,T. Bergman, S-E Karlsson, I. Lindgren, B. Lindberg. Nova Acta Regiae Societatis ScientiarumUpsaliensis, ser. IV. vol. 20, (Uppsala, 1967).S. Evans, J. C. Green, M. L. H. Green, A. F. Orchard and D. W. Turner, this Discussion.P. S. Braterman and A. P. Walker, this Discussion.lo C. W. Fryer and J. A. S. Smith, private communication.l1 A. A. Hock and 0. S. Mills, Proc. Chem. SOC., 1958, 233.J. V. Brencic and F. A.Cotton, Inorg. Chem. 1969, 8,7.l3 E. R. Corey, L. F. Dahl and W. Beck, J. Amer. Chem. Soc., 1963,85,1202.l4 E. H. Braye, L. F. Dahl, W. Hubel and D. L. Wampler, J. Amer. Chem. SOC., 1962,84,4633.l5 B. F. G. Johnson, R. D. Johnston and J. Lewis, Chem. Comm., 1967, 1057.l6 R. Mason and W. R. Robinson, Chem. Comm., 1968,468.l7 F. A. Cotton, this Discussion.l8 N. V. Sidgwick, The Electronic Theory of Valency, (Oxford University Press, London, 1927).l9 L. Pauling, The Nature ofthe Chemical Bond (3rd ed., Cornell University Press, New York,2o J. H. Van Vleck, Theory of Electric and Magnetic Susceptibilities (Oxford University Press,21 R. Schlapp and W. G. Penney, Physic. Reu., 1932,42,666.22 0. M. Jordahl, Physic. Rev., 1934, 45,87.23 J. Owen and K. W.H. Stevens, Nature, 1953,171,836.24 J. H. Van Vleck, J . Chem. Phys., 1935,3,805.25 Y . Tanabe and S. Sugano, J. Phys. SOC. Japan, 1954,9,753,766.26 F. A. Cotton, Chemical Applications of Group Theory (Interscience, New York, 1963).27 C. J. Ballhausen and H. B. Gray, Molecular Orbital Theory (Benjamin, New York, 1964).28 C. K. Jprrgensen, Absorption Spectra and Chemical Bonding in Complexes (Pergamon, Oxford,29 R. S. Nyholm, Report. 10th Soluay Council CBrussels, 1956), p. 230.30 M. L. H. Green, Organometallic Compounds, vol. U. The Transition Elements, ed. G. E. Coates,st A. Streitwieser Jr., V. MiiUer-Westerhoff, J. Amer. Chem. Soc., 1968,90,7364.1960).London, 1932).1962).M. L. H. Green, and K. Wade, (Methuen, London, 1968)G . A . GAMLEN 1932 P. H. Lewis and R. E. Rundle, J. Chem. Phys., 1953,21,986.33 R. G. Vranka and E. L. Amma, J. Amer. Chem. SOC., 1967,89,3121; see also T. Onishi andT. Shimanouchi, Spectrochim. Acta, 1964,20,325.34 K. A. Levison and P. G. Perkins, this Discussion.35 J. Chatt and J. M. Davidson, J. Chem. SOC., 1965, 843.36 S. D. Ibekwe, B. T. Kilbourn, Miss U. A. Raeburn and D. R. Russel, Chem. Comm., in press.37 J. Chatt and B. L. Shaw, J. Chem. SOC., 1959,705 ; 1960, 1718.38 M. Green, J. Smith and P. A. Tasker, this Discussion.39 H. A. 0. Hill, J. M. Pratt and R. J. P. Williams, this Discussion.40 F. G. A. Stone, Endeavour, 1966,25,33 ; M. I. Bruce and F. G. A. Stone, Preparative Inorganic41 W. C. Zeise, Pogg. Ann. 1827,9,632.42 M. J. S. Dewar, Bull. SOC. Chim. France, 1951, 18, C71.43 J. Chatt and L. A. Duncanson, J. Chem. SOC., 1953,2939.44 L. Manolovic-Muir, K. W. Muir and J. A. Ibers, this Discussion.45 Miss M. Black, R. H. B. Mais and P. G. Owston, private communication.46 R. Cramer, J. Amer. Chem. SOC., 1964,86,217 ; 1967,89,4621.47 C. E. Holloway, G. Hulley, B. F. G. Johnson and J. Lewis, J. Chem. SOC. A , 1969,53.48 R. McWeeny, R. Mason and A. D. C. Towl, this Discussion.49 M. C. Baird, G. Hartwell Jr., R. Mason, A. I. E. Rae and G. Wikinson, Chem. Comm., 1967,92, see also M. C. Baird and G. Wilkinson, J. Chem. SOC. A, 1967, 865.50 T. G. Hewitt, K. Anzenhofer and J. J. DeBoer, Chem. Comm., 1969,312, see also T. Kashiwagi,N. Yasuoka, N. Kasai and M. Kukudo, Chem. Comm., 1969, 317. '' R. Pettit, H. Sugahara, J. Wristers and W. Merk, this Discussion.Reactions, W. L. Jolly, ed., (Interscience, New York), 1968, 4, 177