Comprehensive Organometallic Chemistry || Bonding of Unsaturated Organic Molecules to Transition...

19.1 INTRODUCTION 2

19.2 CARBONYL COMPOUNDS 2/ 9.2. / Bonding in Metal Carbonyl Complexes 219.2.2 UV Photoelectron Spectral Studies of Metal Carbonyls 519.2.3 Force Constants in Metal Carbonyl Complexes 919.2.4 Geometries of Metal Carbonyls and Related Species 12

19.2.4.1 Even-electron low-spin complexes 1219.2.4.2 Odd-electron and high-spin complexes 16

19.2.5 Bonding Capabilities of Metal Carbonyl Fragments 16/ 9.2.6 Molecular Orbital Analysis of the Frontier Molecular Orbitals of Metal Carbonyl Fragments 18

19.2.6.1 C2v M (CO)4 fragment 1819.2.6.2 C3v M(CO)3 fragment 2019.2.6.3 C2v M(C0)2 fragment 23

19.2.7 Reactions of Metal Carbonyl Complexes 2419.2.7.1 c\s-Labilization in metal carbonyl complexes 24/ 9.2.7.2 Methyl migration in metal carbonyl complexes 26

19.3 SANDWICH, TRIPLE DECKER SANDWICH, BENT SANDWICH AND HALFSANDWICH COMPOUNDS 28

19.3.1 Sandwich Compounds of the Transition Metals 2819.3.2 Cyclooctatetraene Sandwich Compounds 3119.3.3 Mixed Sandwich Compounds 3219.3.4 Sandwich and fSlippedf Sandwich Compounds derived from Carboranes 3219.3.5 Triple Decker Sandwich Compounds 3519.3.6 Bent Sandwich Compounds 3719.3.7 Half Sandwich Compounds 4019.3.8 Aromaticity in Sandwich Compounds 46

19.4 ALKENE AND POLYENE METAL COMPLEXES 47/ 9.4.1 The Dewar-Chatt-Duncanson Bonding Model 4719.4.2 Asymmetrically Substituted Alkene Complexes 5219.4.3 Conformational Preferences of Metal Alkene Complexes 5419.4.4 Conformational Preferences of Metal Alkyne Complexes 59/ 9.4.5 Metal Polyene and Polyenyl Complexes 6019.4.6 Reactions of Coordinated Alkenes and Polyenes 67

/ 9.4.6.1 Nucleophilic addition reactions of coordinated alkenes 68/ 9.4.6.2 Peri cyclic reactions of coordinated alkenes and polyenes 7019.4.6.3 Alkene insertion and hydrogenation reactions 1119.4.6.4 Ziegler-Natta catalysis and metallacyclobutanes 72

19.5 OXIDATIVE ADDITION AND REDUCTIVE ELIMINATION REACTIONS 76

19.6 ALKYL, CARBENE AND CARBYNE COMPLEXES 7719.6.1 Alky I Complexes 1119.6.2 Metal Carbene (Alkylidene) Complexes 1119.6.3 Metal Carbyne (Alkylidyne) Complexes 80

REFERENCES 80

1

19

Bonding of Unsaturated Organic Moleculesto Transition MetalsD. M. P. MINGOSUniversity of Oxford

2 Bonding of Unsaturated Organic Molecules to Transition Metals

19.1 INTRODUCTION

The rapid development of organometallic chemistry following the discovery of ferrocene inthe early 1950s resulted in the isolation of a wide variety of compounds with unusual and at timesbewildering bonding modes between transition metals and organic molecules or fragments. Thecomplexity of the structures revealed by an ever increasing number of crystal structure determi-nations led organometallic chemists to adopt a Baconian approach to the scientific developmentof the subject,1 whereby they were content to classify the new compounds in an empirical fashionand relied on only the most rudimentary of bonding concepts, such as the 18-electron rule andPauling's electroneutrality principle and synergic bonding model, as the basis of a conceptualframework.2

In 1956 Longuet-Higgins and Orgel3 demonstrated the potential of the molecular orbital methodas a predictive tool in organometallic chemistry by suggesting the existence of cyclobutadienecomplexes two years before they were actually isolated by Criegee and Hubel.4>5 At that timehigh speed computers were not generally available and this type of molecular orbital analysisdepended heavily on symmetry arguments. Consequently, some erroneous conclusions were alsoarrived at by this method of analysis.6 During the last ten years the general availability of highspeed computers has led to a renaissance in theoretical organometallic chemistry, and molecularorbital calculations have been reported at various levels of sophistication. Interestingly, the ap-proach which has had most impact on the experimental chemist during this period has not beenthe very sophisticated ab initio calculations, but the approximate semi-empirical molecular orbitalcalculations based on the extended Hiickel method.7 In particular, the fragment approach to thebonding in organometallic compounds developed so widely by Hoffmann and his coworkers8 duringthe last five years has provided a conceptual framework for the experimental chemist which enableshim to develop approximate bonding models without having to resort to extensive computing.This chapter illustrates the application of this methodology, which utilises perturbation theoryand symmetry arguments extensively,10 to a wide range of structural and reactivity problems inorganometallic chemistry. However, it is recognised that this approach is an approximate oneand therefore wherever possible results from more sophisticated computational techniques havebeen used to check the validity of the conclusions.

19.2. CARBONYL COMPOUNDS

19.2.1 Bonding in Metal Carbonyl Complexes

An understanding of the interactions between carbon monoxide and transition metals is essentialto an understanding of the structures and properties of carbonyl complexes and their substitutedderivatives. Molecular orbital analyses of the bonding in transition metal complexes have generallybeen based on two approaches. The approach which is most widely used and simplest to appreciateis based on elementary symmetry and perturbation theory ideas. The second approach which hasresulted from the widespread availability of high speed computers is based on the rigorous andquantitative derivation of molecular orbital energies and related observable physical quantities.The high symmetries ofcdrbonyi complexes such as Cr(COJ6, Ni(COJ4 and Fe(CO)s togetherwith the wealth of photoelectron, vibrational and NMR data which has been accumulated forthese complexes have made them an ideal testing ground for the development of sophisticatedmolecular orbital (and indeed also valence bond) calculations.1118 Unfortunately, unlike thesituation which pertains to X-ray crystallographic structural analyses, there is no independentlybased reliability index to judge the reliability of a particular calculation. The tendency of authorsto emphasise the merits of their particular mode of calculation makes it difficult for the less spe-cialised reader to make an independent judgement of their validity.

In this section the basic features of the perturbation model of carbonyl bonding will be outlined,and then the more sophisticated calculations will be used to judge its overall reliability. In thiscontext, it is instructive to examine in a little detail the molecular orbitals of the isolated carbonmonoxide molecule. Figure 1 illustrates the molecular orbitals of CO.19 The HOMO is 5 a andthe LUMOs are the degenerate set 2TT*. The greater electronegativity of oxygen compared tocarbon results in a greater localization of the molecular orbitals 4a and ITT on the oxygen atomand 5a and 2TT* on the carbon atom. Also indicated in Figure 1 are the approximate energies ofthe metal J-orbitals of the first row transition elements. The precise energies of the metal ^-orbitalsin a particular complex will depend on the charge on the metal atom, but since the metal charge

Bonding of Unsaturated Organic Molecules to Transition Metals

is likely to lie between 0 and +1, a band can be plotted to indicate the possible d-orbital energyrange.20 A recognition of the relative energies of 4c and 5 a and their electronic distribution clearlyindicates that the metal-carbon c-bonding interactions will be dominated by donation of electrondensity from 5 a to a suitable empty orbital on the metal as illustrated in (1). The situation is lessclear cut for the CO 7r-levels since although the overlap between the 2TT* level and the relevantmetal ^/-orbital, as indicated in (2), is superior to that between \w and the metal d-orbital (see(3)), the latter is not negligible. In earlier accounts the contribution of (3) was largely ignoredand the M—CO bonding was described in terms of (1) and (2) only.21 The important elementsof metal carbonyl 7r-bonding are better represented by the interaction diagram of Figure 2. Thisis a common three orbital interaction diagram and indeed if the interactions between 1 TT and 2TT*with 3dw were identical then the resultant molecular orbitals would approximate to bonding,non-bonding and antibonding, and the non-bonding level would be noded at the central carbonatom as indicated in (4).22 In point of fact the Sd^-lir* interaction predominates and the highestoccupied level in a metal carbonyl complex is better represented by the mixing illustrated in (5).The predominance of the 3f/7r-2x* interaction also suggests that there will be a net transfer ofelectron density from the metal to the carbonyl ligand. Therefore, the synergic bonding modelas first proposed by Pauling23 for metal carbonyls whereby the carbonyl metal a-donation issupplemented by back donation to the 2TT* (CO) level is essentially corroborated. However, inmolecular orbital terms it is also necessary to introduce the suppplementary mixing between 1 wand 2TT* to complete the picture. The essential validity of the three-centre bonding model describedabove is illustrated by electron contour plots derived from SCF-Xa-MSW calculations onCr(CO)6 by Johnson and Klemperer19 and illustrated in Figure 3. The It2g level is primarily CO

(1) (2) (3)

Ti

In 10

Ti

Ni5a

5a

1n

In

4a

|

+1

Ni

Figure 1 Electron density plots of the frontier molecular orbitals of carbon monoxide, adapted from J. B. Johnsonand W. G. Klemperer, J. Am. Chem. Soc, 1977,99,7132. Also shown in the figure are the energies of these mo-lecular orbitals relative to those of the first row transition elements in their zero and singly charged states

3


(4)

3d,

Figure 2 A schematic illustration of the three-centre orbital interaction which occurs as a result of the overlapof the carbon monoxide 1 w and 2w* molecular orbitals with the ^/-orbital of a transition metal with 7r-pseu-dosymmetry, 3dw

17T in character, with a small bonding contribution from 3dx. The 2t2g level is primarily a 3dv-orbital perturbed by a bonding interaction with 2x* (CO) and an antibonding interaction withIT (CO); and 3t2g is primarily 2TT* (CO) in character with an antibonding interaction with 3d-K.

In thiocarbonyl and thionitrosyl complexes the lower electronegativity of sulphur comparedto oxygen and the smaller l7r-27T* energy separation results in an equalization of the 3d7r-2w*and 3dT-1 w interactions and consequently a closer similarity to the bonding, non-bonding, anti-bonding situation described above and illustrated in (4). Lichtenberger and coworkers havecompared CO, CS and NO and NS as ligands using Fenske-Hall molecular orbital calculationsand the reader is referred to refs. 24-27 for a full discussion of the implications of this three-centremodel to the interpretation of UV photoelectron, IR and electron impact studies on complexesof these ligands.

There has been some controversy in the literature recently concerning the relative importanceof the <r-donation and 7r-back donation components of the synergic bonding model. Part of thiscontroversy originates from an absence of unambiguous techniques for allocating in a precisefashion the charge associated with metal and carbonyl within the complex and the precise degreeof build-up of electron density in the internuclear regions.19'28"30 Johnson and Klemperer19 haveestimated on the basis of Xa-SCF-MSW calculations that in Cr(CO)6 the ^-donation representsapproximately 80% of the bonding contribution compared with 20% for 7r-back donation. These

4

3dw + 2TT* \ir

(5)


o«o ©•0

5eg

» •«••

o

It 2g 2t 2g * . »

estimates were based on charge density comparisons in (CO)6 and Cr(CO)6 and must be treatedwith some caution in view of changes in atomic sphere sizes in the independent calculations onthe (CO)6 and Cr(CO)6 entities. Although the strength of bonding represented by these compo-nents differs by a large amount, Johnson and Klemperer19 have suggested that the weaker x-in-teraction is able to charge compensate for the strong c-interaction since Cr d^ to CO 2TT* bondingleads to far more charge transfer than CO 5 a to Cr da bonding of an equivalent magnitude.Sherwood and Hall's29 Fenske-Hall parameter free molecular orbital calculation on Cr(CO)6suggests that the bonding contributions are more equally balanced — 55% for <r-donation and45% for x-back donation. They suggest that their results are in general agreement with the lowtemperature X-ray and neutron diffraction study of Rees and Mitschler.31 Electron populationanalyses showed 1.52 electrons in the 5<r-orbitals and 0.51 in the 27r*-orbitals of each carbonylwhile the experimental values are 1.65 and 0.38, respectively. The gross atomic charge on Cr was0.306 compared to the experimental value of 0.15 ± 0.12.3'

Bursten, Freier and Fenske's30 results using the projected Xa (PXa) technique also indicatethat the estimates of Johnson and Klemperer undervalue the importance of the back donationcomponent. The PXa technique has permitted the authors to avoid the ambiguities inherent inthe interpretation of Xa-SW charge distributions by not relying on the contour plot techniqueutilized by Johnson and Klemperer.19 Their calculations suggest that 1.42 electrons are donatedand 1.25 electrons back donated in Cr(CO)6- Therefore, in summary there appears to be a con-sensus of opinion that metal-carbonyl bonding leads to electroneutrality as originally proposedby Pauling23 although the relative strengths of c-bonding and 7r-bonding components remaina matter of some debate. A delineation of these relative contributions is of course experimentallyvery demanding if not impossible.

19.2.2 UV Photoelectron Spectral Studies of Metal Carbonyls

The UV photoelectron32'33 spectra of metal carbonyls and related species have been extensivelyinvestigated during the last few years. There is general agreement concerning the importantspectral features and especially those associated with ionizations from molecular orbitals localizedpredominantly on the metal. However, there has been a considerable debate concerning the as-signment of the lower lying carbonyl based orbitals. These compounds have proved to be an im-portant testing ground for the more sophisticated theoretical calculations1 ]~17 and for evaluating

5

Figure 3 Electron density contour plots of the 5eg, It2g, 2t2g and 3t2g molecular orbitals in Cr(CO)6 derived fromSCF-X a-MSW molecular orbital calculations (reproduced with permission from J. B. Johnson and W. G.Klemperer, J. Am. Chem. Soc, 1977, 99, 7132). The It2g, 2t2g and 3t2g molecular orbitals can be related to thethree molecular orbitals illustrated schematically in Figure 2. The 5eg molecular orbital corresponds to the lowestunoccupied set of molecular orbitals in Cr(CO)6


the validity of Koopmans' approximation, which equates the negatives of the SCF computed ei-genvalues €/*CFof the molecular ground state with ionization energies of the various molecularorbitals in the following simple fashion:34

IE = - e ? C F

While this theorem saves a great deal of computer time and is a conceptionally satisfying con-nection between molecular orbital theory and photoelectron spectra, it has severe limitations.In particular this 'frozen' orbital approach neglects electron relaxation effects associated withthe formation of the positive ion in the ionization process and effects arising from electron cor-relation. Studies on organometallic compounds have demonstrated that in general Koopmans'theorem is not a good approximation for such systems since the relaxation effects associated withphotoionization from molecular orbitals with a high degree of metal character are quite differentfrom those molecular orbitals with a high proportion of ligand character.

Therefore for organometallic compounds it has proved necessary to perform Hartree-Fockself consistent field (SCF) calculations on the molecular ground state and the relevant positivelycharged states and subtract the differences in energy in order to reproduce the observed ionizationenergies. Such calculations are naturally very time consuming and at times a more limited typeof correction, ASCF, is performed whereby only the wavefunctions of the valence orbitals in theionized states are recomputed. Such calculations have indicated that the different relaxationenergies associated with metal based molecular orbitals arise from a considerable charge migrationtowards the metal atom on going from the ground to the positively charged state.35'36

Many of these difficulties may be circumvented by using the SCF-Xa scattered wave techniques.The Xa calculation allows the direct calculation of ionization energies by the 'transition state'method, whereby the difference in two energy states is computed by means of a single calculationin a transition state that involves electron occupation numbers halfway between the initial andfinal values.37 In general the results of Xa-transition state ionization energies are in betteragreement with the observed ionization energies than those calculated by the more time consumingASCF method. However, the Xa results are dependent on the type of potential employed and theprecise nature of the calculation.3840 Figure 4 compares the results of some Xa calculations onNi(CO)4 and emphasises the important point that the computed eigenvalues are very sensitiveto the precise mode of calculation.

Figure 5 illustrates the He-I photoelectron spectra of Cr(CO)6, Fe(CO)5 and Ni(CO)4.4145

The multiplicities and relative intensities of the ionizations in the 8-10 eV region are assignedunanimously to ionizations from molecular orbitals localized predominantly on the metal andcan be interpreted simply in terms of the ligand field splitting diagrams shown below thespectra.

There is some controversy surrounding the ionizations in the spectral region beyond 13 eV,which consists of many very closely spaced ionizations. In the hexacarbonyls the narrow bandat —13.3 eV has been attributed to a molecular orbital derived largely from the 5 a carbonyl lonepair orbital. The overlapping bands between 14-16 eV represent probably a mixture of the re-maining 5a ionizations and carbonyl ITT ionizations. For a detailed discussion of possible as-signments of these ionizations and the utility of He-II radiation in such studies, two excellent recentreviews are recommended to the reader.3536 Semi-empirical and accurate molecular orbitalcalculations on these molecules are given in refs. 11-14 and 46-60.

The corresponding trifluorophosphine complexes have also been extensively investigated.6166

For each of the PF3 complexes the average ionization energies (weighted where appropriate fordegeneracies) from the predominantly metal MOs are larger for the PF3 complexes than thecorresponding carbonyls, thus suggesting a higher positive charge on the metals in the PF3 com-plexes. Whether the greater charge arises solely from 7r-bonding effects or results from an elec-trostatic perturbation introduced at the metal site by virtue of the proximity of the somewhatnegatively charged fluorines remains to be fully resolved.67

Substituted metal carbonyls. The UV photoelectron spectra of substituted metal carbonylcomplexes ML(CO)s (L = H, Me, Cl, etc.; M = Mn or Re) have been very widely studied,6772

and the theoretical assignments of the relevant ionizations have followed an amusing fluxionalprocess which is described in some detail in reference 35. The problem has now been largely re-solved with the assistance of complementary studies using He-II radiation and the observationof spin-orbit coupling effects for the heavier elements. The generally agreed level ordering forthe HOMOs in MnH(CO)5 and Mn(Me)(CO)s is shown in (5a).

Some of the confusion has arisen because the ionizations associated with these levels occur at


-0.5

£0 - 1 . 0c

SCF-X.-SW

Figure 4 Comparison of the results of Xa calculations on Ni(CO)4 (reproduced with permission from B. I. Kim,H. Adachi and S. Imoto, J. Electron Spectrosc. Related Phenom., 1977, 11, 349)

very similar energies to those of the M—L a-bonding molecular orbital of ai symmetry. For ex-ample, in Mn(Me)(CO)5 the 2Ai ion state has an energy of 9.49 eV compared with 8.65 and 9.12eV for the 2E and 2B2 states arising from ionizations associated with the e and b2 molecular orbitalsillustrated above. For MnH(CO)5 and Mn(Me)(CO)5 the observed differences between the 2Eand 2E&2 ion states are 0.40 and 0.47 eV, respectively, and the larger splitting in the methyl com-

b 2 -

LM(CO)5

(5a)

7


8 10 12 14 16Ionization potential(eV)

18 20

8 10

8 10

12 14 16Ionization potential(eV)

18 20

18 20

Figure 5 He-I UV photoelectron spectra for (a) Cr(CO)6, (b) Fe(CO)5 and (c) Ni(CO)4 (adapted from A. HCowley, Prog. Inorg. Chem., 1979, 26, 46) and their respective ligand field splitting diagrams

pound has been attributed to hyperconjugative effects. The photoelectron spectra of the morecomplex halocompounds MnX(CO)5 (X = Cl, Br or I)70'73 have been similarly assigned.

The UV photoelectron spectra of the substituted carbonyls ML(CO)5 (L = neutral ligand;

8 1U 1 / 14 10 15 ZV

Ionization potential(eV)

• eg a i ^ ^ = t2

e'

Cr(CO)6 Fe(CO)5 Ni(CO)4

d*-Oh d*-D3h dl0-Td

lAlg—- lT2g ! A i ' — • 1 E ' a n d 1 E " ! A , — - lT2 and lE

ion state ion states ion states

8

Bonding of Unsaturated Organic Molecules to Transition Metals 9

M = Cr, Mo or W) have also been investigated and the 2E-2B2 energy separation has been usedto suggest the following order of 7r-acceptor ability:74'75 amine ~ pyridine < sulphide < phos-phine.

19.2.3 Force Constants in Metal Carbonyl Complexes

The vibrational spectra of metal carbonyl complexes, and in particular their IR spectra, havemade an important contribution to their structural characterization and provided a wealth of datafor theoretical interpretation. The rigorous vibrational analysis of metal carbonyl compoundsis complex and has only been completed for the more symmetrical compounds. The metal-carbonand carbon-oxygen force constants arising from such analyses are summarised in Table 1 forM(CO)6 (M = Cr, Mo or W) and Ni(CO)4.76-79 The reduction in C—O force constant oncomplexation is consistent with the 7r-bonding model described in an earlier section, and indicatesa resultant C—O bond order of approximately 2.7.80

Table 1 Force Constants for some Binary MetalCarbonyls1

Cr(CO)6

Mo(CO)6W(CO)6

Ni(CO)4CO

k(CO)(mdyne A"1)

17.217.317.217.919.8

k(M—C)(mdyne A""1)

2.081.962.362.02

1. L. H. Jones, Ace. Chem. Res., 1976, 9, 128.

The CO force constants in less symmetrical metal carbonyl complexes are generally derivedfrom simpler and more approximate methods, the most commonly used being the Cotton-Krai-hanzel method,81 whereby the CO force constants are derived solely from the observed COstretching frequencies. This method clearly has its deficiencies since anharmonicity effects areignored and the interaction constants between cis- and /ra«s-carbonyl ligands are constrainedto bear a simple geometric relationship. In the CO factored force field approximation the latterconstraint is not necessary because additional data available from isotopically substituted de-rivatives are utilised.82 The force constants derived by these methods, although not accurate inan absolute sense, probably accurately reflect the trends in a series of related complexes. Fromthese results it is apparent that the force constants in metal carbonyl complexes are controlledlargely by the extent of back donation permitted to the carbonyls as influenced by the metal oxi-dation state and the electronic characteristics of the other ligands coordinated to the metal.Qualitative discussions of such effects on CO force constants are to be found in several elementarytext books.21

Fenske and his coworkers8386 have completed an extensive molecular orbital analysis of metalcarbonyl, isocyanide and cyanide complexes based on the Fenske-Hall model which indicatedthat in such complexes the ligand orbital populations remained essentially constant except forthe ligand donor orbital and lowest unoccupied orbitals (5a and 2TT* in the case of CO). Thereforein such complexes the electronic distribution can be described by the following: CO (l7r)4(5o-)a-(2TTx)

b(2TTy)b', RNC (l2i\)a(^e\x)

b{}Q\y)b. Consequently the force constants in these complexesare given by the following simple equations (k in mdyn A"1).

k = 37.44 — 8.88a — \6A5b for neutral carbonyl complexesk = 39.32 — 9.78a — 18.926 for cationic carbonyl complexesk = 36.42 - 9.93a - 11.766 for M(CO)sLk = 33.00 — 1.66a — 11.986 for neutral isocyanide complexesk = 30.06 — 5.54a — 13.716 for cationic isocyanide complexes

These equations reproduce the experimentally determined data exceedingly well and linear regression analyses indicated a statistical significance at the 99% level. Brown et a/.87 have demonstrated similar relationships based on semi-empirical MO calculations.

1.0

0.8

0.6

0.4

0.2

0

-0.2

-0.4

-0.6 -

-0.8

*«** $

Increasingn withdrawal

Increasing o withdrawal

10 Bonding ofUnsaturated Organic Molecules to Transition Metals

Based on the assumption that 7r-bonding effects alone are important, Graham88 has proposedthe following relationship for changes in the force constants (of trans- and c/s-carbonyls) inLM(CO)5 relative to MeMn(CO)5:

&k trans = A(7L + 2ATTL

Akcis = A(7L + AxL

A<TL and ATTL represent the a- and ^-contributions to the force constants and may be calculatedif &axiai and A;basal are known. The result of this analysis is presented in a diagrammatic form inFigure 6. The results are in general accord with chemical intuition since Cl for example is classifiedas a poor a-donor and x-donor and SnMe3 as a good c-donor and vr-acceptor.

i

c

B

i_

PL,

-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4a Parameter (mdyne A ~')

Figure 6 Plot of OL and TIX parameters for substituted metal carbonyl complexes Mn(CO)5X derived from forceconstant data (reproduced with permission from W. A. G. Graham, Inorg. Chem., 1968, 7, 315)

Yarbrough and Hall67 have made a similar analysis for the complexes ML(CO)5 (M = Cr,Mo or W; L = PEt3, PMe3, P(NMe2)3, P(OEt)3, P(OMe)3 and PF3). Such an analysis suggeststhat CO and PF3 have almost identical donor/acceptor characteristics. The remaining phosphaneligands have similar characteristics and are superior (7-donors and weaker 7r-acceptors than PF3.U V photoelectron spectral studies on these compounds suggest that PF3 is a better c-donor andweaker vr-acceptor than CO and have established the relative donor and acceptor abilities ofP(NR2)3, P(OR)3 and PR3 shown in Figure 7.

More remarkable is the simple equation proposed by Timney89 relating with amazing accuracythe kco Cotton-Kraihanzel force constant in binary metal carbonyls M(CO)rt to the force constantin the monocarbonyl M(CO) with the same number of d electrons kci, the number of carbonylligands no at an angle 6 to the carbonyl under consideration and a ligand effect constant, be'-

kco = kd + Y,ndb()

Thus for example in a square-pyramidal M(CO)s fragment:

/cco(axial) = kd + 4bcis

/cco(basal) = kd + 3bcis + btrans


0

— 1

Q.

O- 2

03I

- 3

CO

PF

Estimated limits ofexperimental uncertainty

1 1

P(0R)

PR

13 12 11a, (ff-Donor)—

10

Figure 7 Comparison of the relative o and TT characteristics of CO, PF3, P(NR2)3, P(OR)3 and PR3 derivedfrom photoelectric spectral studies (reproduced with permission from L. W. Yarbrough and M. B. Hall, Inorg.Chem., 1978,17,2269)

Typical values of kd are given in Table 2 and the be values for CO are given in Table 3. Burdetthas proposed an explanation for this amazing additive phenomenon in terms of an angular overlapanalysis90 of the 7r-bonding in octahedral metal fragments. The fact that bCiS « btransl^ suggests

Table 2

First transition seriesSecond transition seriesThird transition series

Force Constants kda'b of Metal Monocarbonyl Fragments1

k5

1373

1353

k6

138713811381

ki

1444

1445

ks kg

1498 155415061498

1810

1613

a d is the number of valence electrons associated with the metal. b In N m"""1,1. J. A. Timney, Inorg. Chem., 1979, 18, 2502.

Table 3 Timney's Ligand Effect Constants1^

Ligand

COPF3

PC13PCl2PhPClPh2

PMe3PPh3MeCNP(OPh)3

NOCSN2CpClBrIHMe

b9o(Y)

33.533.230.614

-27.7-21- 1 4

1.3425614.0

1431341127571

126.1141.6109.3825529.8293094

23216052.0

15910610110412992

37.344.935.313.0

-11.0-38.7-31.7-21.9

-0.330.065

5.799

145141125

71

C/QC)\ 1 1

25.516.021

-61- 5 2

45

bno(Y)b

51.444.6

22

a I n N m ~ ' . b In trigonal bipyramid.1. J. A. Timney, Inorg. Chem., 1979,18, 2502.

that trans ligands lead to less overall stabilization of the system than cis ligands, a conclusionwhich is consistent with the observed geometries of metal carbonyl fragments. The Timney


equation can be generalized to substituted carbonyl complexes of the type [M(CO),iYw']<? asfollows:

*co = kd+ ZnebeiCO) + IXA(Y) + qW

and the appropriate values of be(Y) are given in Table 3. The constant k' extends the analysisto charged metal carbonyl complexes.

The magnitude and sign of the vibrational displacement interaction coordinates derived forCr(CO)6 by Jones7679 in a rigorous fashion have also been the subject of theoretical analysesbased on the Xa

19 and Fenske-Hall method.29 The former suggests that the interaction constantsare inconsistent with substantial ^-contribution to the total Cr—C bond order, and the latterfavours moderately strong Tr-bonding.

19.2 .4 Geometries of Metal Carbonyls and Related Species

19.2.4A Even-electron low-spin complexes

The application of low temperature matrix isolation techniques particularly by Turner andOzin and their coworkers91"101 has led to the isolation and structural characterization using IRand Raman techniques of a wide range of coordinatively unsaturated metal carbonyl species, e.g.Cr(CO)5, Fe(CO)4 and Mo(CO)3. These studies have also provided an important stimulus fortheoretical calculations designed to rationalize the geometries and ground state electronicstructures of these kinetically labile species. In particular, extended Hiickel molecular orbitalcalculations by Hoffmann and Elian102 and Burdett103'104 based on the Walsh diagram meth-odology have been successfully used to account for the geometries of many of these molecules.Their conclusions together with those derived by Pensak and McKinney105 using bond lengthand bond angle energy minimization techniques (possible within the framework of the modifiedextend Hiickel method) are summarised in Table 4. Some of these conclusions have also beenverified using ab initio techniques.106'107 A detailed discussion of the electronic arguments usedto interpret the slopes of the energy levels in these Walsh diagrams will not be reproduced herebecause of space limitations. The reader is referred, therefore, to the relevant papers for excellentdiscussions of these aspects.102"107 The intention of this section is to summarise the importantconclusions derived from these studies and to demonstrate that the ground state geometries ofthe coordinatively unsaturated fragments bear important relationships to those of the parent18-electron carbonyl complexes. In presenting the results in such a fashion it is hoped that thegeneral conclusions of these analyses will be more easily remembered and utilized.

In general the geometries of main group covalent compounds can be most simply accountedfor in terms of the Valence Shell Electron Pair Repulsion approach,108 although equivalent andat times more accurate conclusions can be derived from molecular orbital109"113 or crystal fieldcalculations.114'] 15 The well documented isoelectronic basis of the VSEPR approach is representedin a matrix form in Figure 8(a), which clearly illustrates the relationship between the observedgeometry and the number of valence electron pairs. The figure also emphasises the importantstereochemical role played by the lone pair electrons. In general for transition metal complexesit is not possible to construct a corresponding matrix because electrostatic interactions frequentlyplay as important a role as covalent interactions in determining the equilibrium geometry, e.g.the octahedral coordination geometry is particularly widespread for the first row transition metalsbecause of favourable metal to ligand size ratios.116 Nevertheless the structures of those transitionmetal complexes, which have a wide range of low symmetry geometries, must arise from directionalcovalent bond effects. The binary metal carbonyl compounds are good examples of this type oftransition metal compound and it is possible to represent their structures in a matrix form anal-ogous to Figure 8(a), albeit in a less extensive fashion (see Figure 8(b)). The diagonal elementsrepresent the stable 18-electron molecules Cr(CO)6 (octahedral), Fe(CO)5 (trigonal bipyramidal)and Ni(CO)4 (tetrahedral) which conform to the VSEPR geometries if the d electrons are ignoredin the electron pair count. There are no examples of stable seven coordinate 18-electron carbonylcomplexes but the related ions [Mo(CN)7p~ and [Mo(CNBut)7]2+ are known. The former hasa pentagonal bipyramidal (PB) structure in agreement with VSEPR considerations and the lattera capped trigonal prismatic structure (CTP).117118 These are two of the three almost equienergeticstructures available to transition metal seven-coordinate compounds (see ref. 119 for a detailedanalysis of the potential energy surfaces in these molecules).

The potential energy surface for five-coordinate complexes is also very soft120 and althoughFe(CO)5 has a ground state structure based on the trigonal bipyramid it is highly fluxional and

Table 4 Calculated Geometries of Low Spin Binary Transition Metal Carbonyls using Semi-empiricalMolecular Orbital Methods too

;s

Co

«

I53

ICo

3

o

d-Orbital Coordinationpopulation number / ' II2 III3 Examples4

d6 6 M(CO)6 Oh5 C4v (d = 93.5°) C4v(93°) 93° W(CO)4(CS)

Cr(CO)5C4,(^ = 91.4°)4 C2v (90°, 170°) C2v (95°, 165°) Mo(CO)4 C2v (107°, 174°)3 C3v (d = 60°) C3o (0 = 57°) Mo(CO)3 C3v (0 = 65°)

ds 5 D3h D3h Fe(CO)54 DAh Dld(\50°) C2v (Co(CO)4)+ Ni(CN)J~ Z)4/,

Mn(CO)4, Fe(CO)43 C2v (a = 175°, C2v (a = 172°, [Rh(PPh3)3]+ {a = 159°, 6 = 90°)

0 = 90°) 6 = 90°)d10 4 rrf Td Td Ni(CO)4

3 Z)3A C3v(6 = &2°) D3h Ni(CO)3D3hd4 1 D5h Mo(CN)4

7-D5hMoCCNBu1)^ C2v (CTP)

6 D3d Distorted octahedral

Angles defined as follows: -, „

1. M. Elian and R. Hoffmann, Inorg. Chem., 1975,14, 1058.2. J. K. Burdett, J. Chem. Soc, Faraday Trans. 2, 1974, 70, 1599.3. D. Pensak and R. J. McKinney, Inorg. Chem., 1979,18, 3407.4. The examples are derived from J. K. Burdett, Coord. Chem. Rev., 1978, 24, 1.

Number of ligands Number of ligands

Number ofelectronpairs

7

PentagonalbipyramidorCapped trigo-nal prism

6

Distortedoctahedrale.g. Ti(CO)<

Cr(CO)*

5

Cr(CO)5

Ni(CN)52"

>hFe(CO)5

4

Cr(CO)4

XNi(CN)4

2"

Ni(CO)4

3

j

Cr(CO)3

Rh(PR3)3+

d*

d*

d10

(a) . (b)Figure 8 On the left hand side of the figure the shapes of binary main group molecules as derived from the ValenceShell Electron Pair Theory are shown. The disposition of lone pairs is also indicated. On the right hand side ofthe Figure the shapes of transition metal carbonyl and related molecules are presented. Particularly noteworthyis the relationship between the d electron count and the geometries of the metal carbonyl fragments

too

S'

to

iIr̂

Co< ^O

Co

oa

c ^ *


thought to rearrange through a low energy square pyramidal transition state. Indeed the relatedand isoelectronic ion [Ni(CN)5p~ exists in the crystalline state in two forms, a square pyramidaland a trigonal-bipyramidal which is somewhat distorted towards square pyramidal.121 Therefore,in Figure 8(b) the alternative structures for seven- and five-coordination have been presented.

The VSEPR predictions for metal carbonyl fragments with fewer than 18 valence electronsare only accurate for molecules with a completed d shell, i.e. d]0 complexes. For example,Ni(CO)3122 derived by carbonyl loss from the tetrahedral d10 Ni(CO)4 is trigonal planar. Thestructural change is therefore entirely analogous to that relating H3B—CO (tetrahedral) andH3B (trigonal planar).

For metal carbonyl fragments with an incomplete d shell the VSEPR predictions are not ful-filled. However, as the matrix in Figure 8(b) indicates there is a horizontal relationship betweenthe metal d electron configuration and the metal carbonyl geometry. For example, Cr(CO)s,Cr(CO)4 and Cr(CO)3 all have geometries which can be related by sequential ligand loss tothat of the 18-electron d6 octahedral carbonyl Cr(CO)6- The observed geometries correspondto the retention of the maximum number of cis ligands. Although these fragments show slightdistortions away from the idealized octahedrally based fragment geometries, it is noteworthythat the distortions occur in the opposite sense to those which occur in main group compoundswith lone pairs, i.e. the carbonyls bend somewhat towards the vacant octahedral sites. In a localizedorbital sense these distortions represent movements of the ligands towards the empty octahedralmetal hybrid orbitals.

Low spin ds complexes show a similar relationship based on the square pyramidal geometry,i.e. Ni(CN)2~, 18-electron square pyramidal; Ni(CN)4~, 16-electron square-planar; and ML3",14-electron T-shaped (L = PR3).123 In this series it is the unique apical ligand which is removedfirst to generate Ni(CN)4~. For structures based on the seven-coordinate d4 molecules[Mo(CN)v]5~ and [Mo(CNR)7p+ experimental data is lacking. It has been reported that the16-electron Ti(CO)6 molecule is a distorted octahedron but the precise nature of the distortionhas not been unambiguously determined.124

This matrix representation emphasises the important point that whereas for main groupcompounds it is the lone pairs which are stereochemically active, in transition metal carbonylsit is the number of holes in the d-orbital manifold which exerts a stereochemical influence.Therefore, in contrast to main group elements the basic bonding and geometric features of theparent 18-electron metal carbonyl complexes are retained on carbonyl loss. This point not onlyhas geometric relevance but also important consequences for the bonding capabilities of thesefragments and for understanding the widespread importance of the 18-electron rule in metalcarbonyl chemistry. These aspects are discussed in some detail in a subsequent section.

The conclusions derived above are completely consistent with the following rules proposed byBurdett on the basis of the angular overlap model.l25 (1) Neglect any electrons in the metal orbitalsinvolved in metal-ligand ?r-bonding. The angular geometry of the complex is determined primarilyby the occupation numbers of the metal-ligand antibonding cr-orbitals. (2) If the electron occu-pation numbers of degenerate a* orbitals are symmetrical (for example (t2*)6 in a tetrahedralcomplex and (eg*)4 in an octahedral complex) then the predicted VSEPR geometry is observed.(3) If a hole exists in the highest energy molecular orbital, e.g. a 22220 electronic configuration,then the structure is that of maximum overlap with the lobes of the metal dx2-yi orbital. (4) Iftwo holes exist symmetrically in the two highest energy orbitals, e.g. 22200, then the structureis based on an octahedron containing the maximum number of cis ligands.

The electronic basis of these rules is quite easy to appreciate because it focuses attention onthe number of holes in the ^-orbital manifold within the complex. In the angular overlap meth-odology, destabilization of the LUMO implies a corresponding stabilization of the matching andoccupied bonding molecular orbital. Therefore the achievement of maximum overlap betweendx2-y2 and the ligand orbital set (in a ds low-spin complex) results in a maximum destabilizationof the LUMO and the adoption of the most stable metal-ligand geometry. In a d6 complex themaximum destabilization of two unoccupied antibonding molecular orbitals derived from inter-actions between the metal ^/-orbitals and the ligand orbitals is clearly achieved for geometriesbased on the octahedron, where maximum overlap between dx2-yi, dz2 and the ligand eg set isattained.

Burdett's rule (1) has to be modified somewhat in the light of recent structural and theoreticalresults on six-coordinate low-spin d4 complexes.126 On a model based on c-interactions alone,such complexes should have octahedral geometries. Kubacek and Hoffmann126 have demonstratedthat the presence of strong 7r-donor or acceptor ligands in d4 complexes of the type ML^L'^L"^results in stereochemical activity for the hole in the octahedral t2g set. They have rationalized


the type of distortions shown in Figure 9 for six-eoordinate d4 molecules on the basis of theachievement of maximum destabilization of the LUMO within the t2g set.126

Figure 9 Schematic illustrations of the geometries of Mo(CO)2(OBut)2(py)2 (M. H. Chisholm, J. C. Hoffmannand R. L. Kelly, / . Am. Chem. Soc.y 1979,101,7615) and Mo(CO)2(S2CNPri2)2 (J. L.Templeton and B. C.Ward,J. Am. Chem. Soc, 1980,102, 3288). Both of these six-coordinate dA complexes show significant distortions awayfrom octahedral, in the former towards bicapped tetrahedral and in the latter towards trigonal prismatic

Before completing this section it is important to add the following note of caution. The geometricarguments derived above for metal carbonyl fragments are based on the assumption of a strongcovalent interaction between the metal and ligand orbitals. The contracted nature of the metalnd valence orbitals compared to the corresponding metal (n + 1 )s and (n + \)p orbitals suggeststhat the metal d-orbital-ligand interactions will only predominate when the ligands either havelow electronegativities or are able to form short covalent bonds with some multiple bond characterto the metal, e.g. CO, NO, etc. For more electronegative ligands, e.g. F~, NCS~, etc., the metaldf-orbital-ligand interactions will be of less importance and the stereochemical role of holes inthe d subshell will not be as influential. In such complexes, the geometries will be determinedlargely by electrostatic considerations and the octahedral and tetrahedral geometries will pre-dominate. Similarly, the presence of bulky substituents on ligands is likely to result in deviationsfrom the predicted geometries shown in Figure 8(b) leading to structures which minimise theligand-ligand repulsions, e.g. M{N(SiMe3)2ta (6).

(MeaSDzN^J^M—N(SiMe3)2

( M e 3 S i ) 2 N ^

(6)

19.2,4.2 Odd-electron molecules and high-spin complexes

Fewer odd electron carbonyl fragments have been subjected to matrix isolation studies andconsequently the amount of structural information is more limited. The potential energy surfacesfor such fragments are often very soft and at times more than one structure for a particular mol-ecule have been reported. For example for the d9 complex Co(CO)4, a D2d geometry intermediatebetween those described above for Ni(CO)4 (dw) and Ni(CN)4~ (ds) has been reported in additionto a C$v (6 « 100°) geometry.98127 For Mn(CO)5 a C\v structure similar to that described abovefor Mo(CO)5 has been identified but the 6 angle is larger (95° cf. 91°).128

The molecular orbital calculations on odd-electron species have not been as unanimous in theirpredictions as those described above for even-electron complexes. Since such molecules are notcentral to the bonding arguments presented in later sections the reader is referred to refs. 102-105for detailed discussions of the computed equilibrium geometries.

Magneto circular dichroism studies on Fe(CO)4 have established that this molecule has ahigh-spin paramagnetic ground state.129 For such a configuration Burdett has computed a Cioequilibrium geometry (6 — 135°, 0 = 110°)103 close to the experimentally estimated values of0 = 145° and 0 = 120°.

19.2.5 Bonding Capabilities of Metal Carbonyl Fragments

In the previous section dealing with the geometries of metal carbonyl fragments it was arguedthat for a given d electron configuration the metal carbonyl fragments retained the essentialgeometric features of the parent 18-electron carbonyl complex. For example, the d6 M(CO)s,M(CO)4 and M(CO)3 fragments each have geometries which can be related to that of the parentoctahedral M(CO)6 molecule as shown in Figure 8(b). In a localized orbital sense this suggeststhat the d2sp3 hybridization about the metal centre is retained by the metal in each of theseM(CO)rt fragments, leading to the formation of (6 — n) empty and low energy hybrid orbitals

The metal carbonyl fragments show no tendency to form complexes which exceed the 18-electron configuration since there are no suitable empty orbitals available for accepting the ad-ditional electron pairs. And indeed there are no well documented examples of carbonyls or sub-stituted carbonyls which do exceed the 18-electron configuration.

The formation of carbonyl complexes with fewer than 18 electrons is more finely balanced inan energetic sense, since the electrons which are lost in the formation of 17- and 16-electroncompounds represent depopulation of levels which are involved solely, or primarily, in metal-carbonyl back bonding. This energy loss may be compensated for by additional intra- and inter-molecular electrostatic bonding, particularly if some of the carbonyl ligands are replaced by ligandswith inferior 7r-acceptor qualities such as tertiary phosphines. For example, the 18-electron cis-Mo(CO)2(depe)2 complexes may be readilv oxidized electrochemically to the corresponding17-electron trans-[Mo(CO)2(depo)2]+ ions.132"134 The geometric change associated with theoxidation process assists the stabilization of the 17-electron species since the HOMO illustratedin (8) is not involved in metal-carbonyl back donation, in the trans complex.135

(8)

The arguments which have been outlined above for metal carbonyl fragments will not betransferable to other transition metal-containing fragments ML£+, where L is a ligand with little7r-acceptor ability and M is a non-zero oxidation state metal, since the positive charge on the centralatom will increase the d-s-p promotion energies and effectively prevent the type of hybridizationdescribed above for carbonyls. Consequently, the geometries of species such as [M(H2O)6]W+,[MX4]n~, etc., are primarily decided by crystal field and ligand interaction terms and there isnot the same propensity for achieving the 18-electron configuration.

For odd-electron metal carbonyl fragments such as Mn(CO)s and Co(CO)4, the odd electronresides in an appropriately hybridized d,s,p orbital and therefore is sufficiently stereoactive toresult in the dimerization process illustrated in (9) for the formation of Mn2(CO)io from two


which point towards the missing vertices of the octahedron (see (7)). These orbitals are thereforeenergetically and spatially ideally set up to interact with the donor orbitals of Lewis bases to re-generate 18-electron complexes of the type M(CO)nL6_n. It follows that there is a strong drivingforce for the maintenance of the 18-electron configuration for carbonyl and substituted carbonylcomplexes based on the d4, d6, ds and dw fragments illustrated in Figure 8(b). Indeed the drivingforce is so strong that the empty orbital of M(CO)5 (M = Cr, Mo or W) fragments, generatedby UV photolysis of M(CO)6 in Ar, Kr, CH4 and SF6 matrices, appears to function as an acceptororbital and interacts in a weakly bonding fashion with the 'inert' matrix atoms and moleculesdespite their high ionization potentials. This weak interaction has been identified by IR and UVstudies by Turner et al.130 and has been the subject of a theoretical treatment by Veillard.131

>< ^ >| >!

M(CO)6 M(CO)5 M(CO)4 M(CO)3

>b * $(7)


Mn(CO)5 radicals. Therefore, it is possible to draw some analogies between organometallic radicalssuch as Mn(CO)5 and Co(CO)4 and organic or main group radicals such as Me* and S\Ry.However, it must be emphasised that the metal-metal strengths in Mn2(CO)io (~65 kJ mol"1)and Co2(CO)8 (~80 kJ mol"1)136 are considerably smaller than those of the C—C bond in ethane(386 kJ mol"1) or the Si—Si bond in Si2R6 (~340 kJ mol"1).136 As a result the driving force forradical dimerization is much smaller for metal carbonyl fragments than for their organic analogues.Also, the homolytic cleavage of such metal-metal bonds using thermolysis or photolysis is muchmore facile and the reader is referred to refs. 137-140 for a detailed discussion of the chemicalconsequences of such metal-metal bond dissociation processes.

Mn(CO)5 a Mn(CO);

(9)

19.2.6 Molecular Orbital Analysis of the Frontier Molecular Orbitals of Metal CarbonylFragments

Although the localized hybrid orbital schemes described above are ideal for relating the elec-tronic and geometric features of metal carbonyls to those of their substituted derivatives, it is alsoimportant to understand the frontier molecular orbital characteristics of such fragments in orderto appreciate fully the conformational preferences of M(CO)n(r;-polyene) complexes and thebonding in polyhedral complexes containing metal carbonyl fragments. This fragment approachto organometallic complexes has been pioneered and extensively developed by Hoffmann andhis coworkers.8'142 The initial bonding between the metal carbonyl fragment and the polyenefollows from a consideration of the bonding capabilities of their frontier molecular orbitals andin particular their highest occupied (HOMO), and their lowest unoccupied (LUMO) molecularorbitals. The important interactions between the fragments are easily evaluated using symmetryand second order perturbation theory arguments.9'10 The following section summarizes the im-portant bonding characteristics of the most commonly occurring metal carbonyl fragments.

19.2.6.1 C2v M(CO)4 fragment

For such a fragment the two hybrid orbitals shown in (10) which point towards the missingvertices of the parent octahedron are equivalent to a symmetry adapted set of molecular orbitalsof ai and 02 symmetry, (11) and (12), respectively. Contour plots of these orbitals derived fromextended Hiickel calculations are illustrated in Figure 10.102 Of course in addition to these frontierorbitals there are three lower lying molecular orbitals of a2, bi and ai symmetry which correspondclosely to the t2g set of the parent octahedron. The complete set of molecular orbitals for anM(CO)4 fragment is illustrated on the left hand side of Figure 11. For a d6 metal both the higherlying ai and b2 frontier orbitals are empty and therefore can interact with the donor orbitals ofaj and b2 symmetry of a suitable ligand. A pair of cis c-bonding ligands L fulfils these requirementsand the interaction diagram for the formation of an M(CO)4L2 complex is illustrated in Figure

In V(CO)6 the situation is somewhat different since the odd electron resides within the t2g7r-bonding set of the metal octahedron. X-ray diffraction studies at low temperatures141 of thismolecule show only small distortions from the idealized octahedral geometry and therefore thehole in the t2g subset is not sufficiently stereoactive to cause a major distortion away from thegeometry which maximizes the metal-carbonyl back bonding. In addition an extrapolation ofthe metal-metal bond strengths in Co2(CO)g and Mn2(CO)io suggests a V—V bond strengthof ~50 kJ mol"1 for V2(CO)i2 which is presumably insufficient to compensate for the necessaryreorganization of the V(CO)6 fragments towards geometries based on seven-vertex polyhedra(see ref. 126 for an MO analysis of the possible distortion modes available to V(CO)6).


b X

- 2

JFigure 10 Electron density contour plots of the frontier molecular orbitals of an M(CO)4 fragment (reproducedwith permission from M. Elian and R. Hoffmann, Irtorg. Chem., 1975,14, 1058.)

Figure 11 The energies and symmetry characteristics of the frontier molecular orbitals of a d6 M(CO)4 fragmentwith C2v symmetry are shown on the left hand side of the figure. The figure also illustrates the way in which thesefrontier molecular orbitals interact with the orbitals of two additional cis ligands to generate the complexM(CO)4X2

19

11. For a ds metal the additional electron pair must occupy the b2 molecular orbital and conse-quently the C2V Fe(CO)4 fragment is characterized by an acceptor orbital of aj symmetry anda filled donor orbital of b2 symmetry. This combination has important consequences for under-standing the conformational preferences of Fe(CO)4(alkene) complexes and this aspect will bediscussed in some detail below in Section 19.4.3. The nodal characteristics and electron populationsof the frontier orbitals of the Fe(CO)4 fragment bear an important similarity to those of singletcarbene CH2 as shown in Figure 12. This similarity has been described as isolobal and serves tounderline the structural relationship between cyclopropane and triangulo metal cluster chemistryshown in (13).143'144

(11)

(12)

1

20 Bonding of' Umaturated Organic Molecules to Transition Metals

* H HFe(CO)4

Figure 12 An illustration of the isolobal nature of the frontier molecular orbitals of Fe(CO)4 and singlet carbene(CH2). In addition these fragments are pseudoisoelectronic and consequently can replace each other in the moleculesshown in (13)

CH2 CH2 Fe(CO)4 Os(CO)4

H2c/—-^CH2 H2C^-—Ve(CO)4 H2C^—-^Fe(CO)4 (OC)4Os^-—\)s(CO)4

(13)


The three frontier molecular orbitals of an M(CO)3 fragment may be derived by taking sym-metry adapted linear combinations of three octahedral hybrid orbitals. Alternatively they maybe derived from the molecular orbitals of the parent octahedral carbonyl M(CO)6 as illustratedin Figure 13. In this Figure the octahedral molecular orbitals have been defined with respect tothe three-fold axis, since the removal of three fac carbonyls generates the M(CO)3 fragment withthree-fold symmetry. In this coordinate system the t2g set consists of z2, — (2/3)1/2(x2 — y2) —(\/3)l/2yz and (l/3)l/2xy — (1/3)1/2JCZ. The mixing of yz and xz character serves to providemaximal bonding to carbonyl TT*. The corresponding eg set is (l/3)1/2(x2 — y2) + (2/3)'/2^zand (l/3)1/2x^ + (2/3)l/2xz giving maximal antibonding with the carbonyl c-bonding orbitals.

AxCr\

L

y vCr

2al

2e

t2g

lela l

Figure 13 Derivation of the molecular orbitals of a conical M(CO)3 fragment from the molecular orbitals ofthe parent octahedral molecule M(CO)6 (reproduced with permission from T. A. Albright, Trans. Am. Crystallogr.Assoc, 1980,16,35)


Removal of the three fac ligands causes the energy of t2g to rise slightly, because of the loss of^-interaction with the carbonyls. The eg set drops greatly in energy and in addition mixing of pxand py character into these orbitals hybridizes them away from the carbonyls as illustrated in(14). One other orbital (2ai) appears at low energy for the C$v M(CO)3 fragment in Figure 13,and it is descended from the aig* orbital in the parent octahedron. It is rehybridized as shownin (15) due to mixing in of pz character.142'144'145 Contour plots of the frontier orbitals of the C^vM(C0)3 fragment are illustrated in Figure 14.142'145 The tilting of the le and 2e molecular orbitalshas important consequences for the conformational preferences of M(CO)3(r]-polyene) complexes,and this aspect will be discussed in some detail in Section 19.4.5.

(14)

(15)


2e.

1

\ I / ^

le.

Figure 14 Electron density contour plots of the le and 2e molecular orbitals of the M(C0)3 conical fragment.Schematic illustrations of these molecular orbitals are drawn on the right hand side of Figure 13

The 2a i and 2e frontier molecular orbitals bear the isolobal relationship to the frontier molecularorbitals of a BH or CH fragment illustrated in Figure 15.144 Indeed the pseudo isoelectronic re-lationships BH = Fe(CO)3 = Co(77-C5H5) and CH = Co(CO)3 = Ni(77-C5H5) are an importantcomponent of the Polyhedral Skeletal Electron Pair Approach developed by Wade and Mingosand widely used to rationalize the structures of polyhedral metallocarboranes, alkyne complexesand metal cluster compounds.146"150 Figures 16 and 17 give some indication of the way in whichthese isolobal relationships serve as an important substitutional principle for unifying the structuresof these very divergent chemical compounds. (See also Chapter 1 in Volume 1 for a detailed dis-cussion of the Polyhedral Skeletal Electron Pair Theory.)

/B

oc H

22

Figure 15 A schematic illustration of the isolobal relationship between Fe(CO)3 and B—H fragments

Aco)3 /^ M A(CO)3Ir^-->(CO)3 (CO)3Cq^-^Co(CO)3 (OC)3Cc^+HCR (CO)3Co yCR

Ir(CO)3 Co(CO)3 Co(CO)3 CR

Figure 16 An illustration of the substitutional principle derived from the isolobal relationship between Co(CO)3,Ir(CO)3 and CR for a series of tetrahedral molecules


Seven skeletalelectron pairsOctahedron Examples


The frontier molecular orbitals of an M(CO)2 fragment can be derived by the removal of twocis ligands from a square planar M(CO)4 molecule. The resultant frontier molecular orbitalsare illustrated in Figure 18.142'145 For M(CO)2 there are four low lying levels of aj, bi, a2 andai symmetry consistent with the strong preference for ds configurations for square planargeometries, and two higher lying levels of b2 and a i symmetry which strongly resemble the frontiermolecular orbitals of the Civ M(C0)4 fragment discussed above. Therefore, it is not surprisingthat d10 ML2 fragments also exhibit a substantial degree of carbene like character. This propertyhas been exploited extensively, particularly by Stone and his coworkers151"155 for the relatedzerovalent Pt(PR.3)2 complexes. However, it should be emphasised that this analogy does not takeinto account the bonding contribution of the lower lying b\ (dyz) metal orbital, which is responsiblefor the distortions observed in a large number of dxo ML2(polyene) and ML^carborane) com-plexes (see ref. 150 for example). These distortions will be discussed together with the confor-mational aspects in more detail in Section 19.3.7.

hy(s-z)

•• z

X

y

hy(xz-x)

!Ni

x2-y2

Figure 18 A schematic illustration of the frontier molecular orbitals of an angular Ni(CO)2 fragment

closo |XwXI B6H62" Co4(CO),oC2Et2 Os5(CO)5P(OMe) H2Ru6(CO)18

Square pyramid

nido V\ BsH<> (B4H8)Fe(CO)3 (C4H4)Fe(CO)3 Os3(CO)io(C2Ph2) Fe5(CO)15C

m-Di vacantoctahedron

arachno \T/ B*Hi<> Fe4(CO)l3H^

Figure 17 An illustration of the application of the isolobal relationship for a series of closo-, nido- andarac/mo-octahedral polyhedral molecules


19.2.7 Reactions of Metal Carbonyl Complexes

Thermal substitution reactions of transition metal carbonyl complexes may occur by one ormore of several competing pathways:156162 (a) dissociation of CO or other departing ligand inthe rate determining step; (b) an associative process involving entering ligand and metal carbonyl;(c) migration of a ligand, particularly of hydride or alkyl to an adjacent CO, thus exposing a vacantcoordination site at the metal centre; and (d) a radical chain pathway in which the metal centreis labilized towards substitution by a radical abstraction reaction at the metal centre, e.g. ab-straction of H from HRe(CO)s.

a Based on kinetic measurements of Fe(CO)4-(PPh3).1. R. J. McKinney and D. A. Pensak, Inorg. Chem.,

1979,18,3413.

Two questions of major importance regarding dissociation of a ligand from a six-coordinatemetal carbonyl complex arise: (a) what are the influences of other ligands in the complex on therate of dissociation and (b) what is the stereochemical relationship between the departing ligandand those that remain? Kinetic evidence on the rate of CO dissociation in substituted carbonylcomplexes suggests that ligands with N, O or halide donor atoms have a labilizing influence,whereas ligands with P, As, Sb, S, Mn, Au, etc. donor atoms generally have a non-labilizing effect.Mechanistic studies, particularly by Brown and his coworkers,164171 have demonstrated thatthe former class of ligand labilizes the cis carbonyls and with the limited kinetic data availablethey have grouped the following ligands in order of their cis-labilizing effect: CO, Au(PPli3),H-, SnPh3, GePh3, M(CO)« < P(OPh)3 < PPh3 < I" < MeSOj, NC5H5 < MeCO" < Br" <NCO~, Cl~, NOJ. The relative labilization of cis CO groups towards dissociation spans a largerange and the difference in rate constants in a given metal series is of the order of 108. Such effectsare large and compare in magnitude with the range of trans effects seen in substitution reactionsof square planar complexes.

Lichtenberger and Brown172 have examined the c/s-labilizing effect from a theoretical pointof view using the Fenske-Hall molecular orbital model and investigated the energy requirementsfor CO dissociation from Mn(CO)J and Mn(CO)5X (X = Br or H). Consideration of the inter-action between individual CO groups and the remainder of the molecule in the ground statestructures of these species does not provide a rationale for the relative rates of CO dissociation.Although the 7r-overlap populations for the cis and trans carbonyls in Mn(CO)sBr reproducethe trend in labilities, the c-overlap populations follow the reverse order (largely because of thesubstantial interaction between the carbonyl 5a and the metal virtual 4s and 4p levels). The^-interaction with the metal d levels is, however, consistent with the observed kinetics. When the

(kJmor1) (kJmol"1)

Ni(CO)4 — Ni(CO)3 + CO 92.3 86.1Fe(CO)5 — Fe(CO)4 + CO 184.5 174.3a

Cr(CO)6 -> Cr(CO)5 + CO 241.9 159.4

19.2.7. / cis- Labiliza tion in metal carbonyl complexes

Ligand dissociation is the most commonly observed process and the departing ligand is mostoften CO. McKinney and Pensak163 have estimated the dissociation energies for the loss of onemolecule of CO in a variety of carbonyls using modified extended Huckel theory. The calculatedAHdissoo which were obtained by subtracting the total energies of the parent carbonyl and thegenerated fragment in its optimised geometry, are reproduced in Table 5. The agreement betweenobserved and calculated values is good for Ni(CO)4 and Fe(CO)s, but overestimated by some80 kJ mol"1 for Cr(CO)6- The reason for this error is unknown, but the authors have noted thatthe modified extended Huckel method disproportionately favours high symmetry situations withcarbonyls trans to each other relative to lower symmetry situations.

Table 5 Computed Bond Dissociation Energies forM(CO)W — M(CO)n-i + CO and the

Corresponding Enthalpies of Activation A//* derivedExperimentally1


a- and x-overlap populations are combined the cis and trans carbonyls of Mn(CO)5Br appearquite similar and the total overlap population suggests that they might be more inert than thecarbonyls in Mn(CO)6 in direct contradiction with the experimental observations.

An examination of the potential energy surface for the Mn(C0)4X fragment, which wouldreasonably represent the transition state for the subtitution process, has indicated a distinct energyminimum for the Civ fragment illustrated in (16) (see ref. 173 for confirmation of this geometryby matrix isolation studies).

For the homoleptic carbonyl cation Mn(CO)5" an energy minimum is achieved for the C$v ge-ometry illustrated in (17). This difference may be traced to a difference in bonding interactionsbetween Mn(CO)4 and either Br~ or CO. The d6 C2V Mn(CO)4 fragment (as described in Section19.2.6.1) possesses two low lying and empty metal levels. The first of ai symmetry is suitable ford-bond formation with Br~ and the second of b2 symmetry is capable of ^-interaction with theBr 4/7-orbitals as in (18). These interactions are particularly interesting in comparison with thecorresponding interactions in the parent Mn(CO)5Br molecule, where the Br 4p 7r-like orbitalsenter into four-electron destabilizing interactions with the filled metal orbitals derived from themetal t2g set. The simplest statement of this observation is that 7r-donor ligands such as Br~ areable to stabilize the 16-electron Mn(CO)4Br species by effective donation of an additional electronpair from their p^-orbitals. Lichtenberger and Brown also concluded that the coordinativelyunsaturated Mn(CO)4" fragment is stabilized when X is a good a-donor ligand. Interestingly,the attainment of the intermediate (16) is not equally favourable for the paths involving cis andtrans carbonyl dissociation shown in (19), since in the latter case the trans CO dissociation involvesa thermally disallowed process involving a crossing of HOMO and LUMO levels.

XBrH

e

698°

100°

085°90°

(16) (17)

(18)

—CO + CO

X

m-dissociation

+ CO

X

/ra/w-dissociation

(19)

ai ((T-interaction) bj (7r-interaction)


McKinney and Pensak163 have confirmed these conclusions using the modified extended Huckelmethod and calculated the following lability order for Mn(CO)5X complexes: MeCO > SH >Cl > Br > I > Me > (PH3)+ > SnH3 > GeH3 > CO+ > H, which is consistent with the experi-mentally determined series given above, although the calculated activation energies are consistently40 kJ mol"1 too high. They have also calculated that the unrelaxed fragment (20a) is some 60kJ mol~l more stable than (20b) in agreement with the observed regioselectivity for the dissociationprocess.

0C-7T—CO

(20a) (20b)

19,2.7.2 Methyl migration in metal carbonyl complexes

Berke and Hoffmann] 74 have studied the methyl migration reaction shown in equation (1) usingthe extended Huckel method.160'161 Important yet representative points along the reaction pathmay be represented by structures (21a-e). Structure (21a) represents a distortion of the structureof the parent MeMn(CO)s molecule with the crucial carbonyl bent up to 20° to 'anticipate' themigrating methyl group. Structure (21c) was identified as the transition state and represents acalculated activation energy of ~80 kJ mol"1, compared to the experimentally determined valueof 48-60 kJ mol"1. Figure 19(a) illustrates the interactions between the carbonyl 5a- and 2TT*-orbitals and the methyl anion lone pair orbital o"Me- m t n e absence of the metal atom. The dominantfeature of this diagram is a four-electron destabilizing interaction between o"Me- and 5 c, whichis only partially relieved by mixing in of 2TT* (CO). This allows the antibonding lobes to hybridizeaway from each other as shown in (22) but the net effect remains antibonding and the processhas a high activation energy. The metal Mn(CO)| fragment is able to reduce the activation energyby interacting in a two-electron stabilizing fashion simultaneously with (TQO + ^Me- and ceo ~~o"Me- using its 2a' (02) and 3a' (ai) orbitals as shown in Figure 19b. Clearly these interactionsare closely related to those involved for the transition state in the as-labilized substitution processin Mn(CO)5Br, and discussed in the previous section. In the course of the migration process alow lying vacant orbital is produced which is the acceptor orbital of a five-coordinate d6 fragmentand which has mainly dzi character. Interaction of this orbital with the lone pair electrons of eithera solvent molecule or the incoming ligand L is important in defining the thermodynamic viabilityof the reaction.

z

X

(a) (b) (c) (d) (e)

(21)

(22)

oRMn(CO)5 + L — • R—C—Mn(CO)4L (1)


aCO"aMe 2a' Me

la'

la' ac0+aMe

OC OCMe Me O

\C Me

(a) (b)Figure 19 The important interactions between the carbonyl 5u and 2TT* molecular orbitals and the lone pairof the methyl anion in the transition state for a methyl migration reaction are illustrated in (a). The stabilizationafforded to the O—C—Me" moiety by the Mn(CO)| fragment is illustrated in (b). This figure was adapted fromfigures in H. Berke and R. Hoffmann, J. Am. Chem. Soc, 1978,100, 7224

Based on this theoretical model for the transition state Berke and Hoffmann174 proposed thefollowing rationalizations and predictions.

(a) Lewis acids enhance the rate of the methyl migration reaction since the electrophilic in-teraction with the carbonyl oxygen atom will lower the energy of 5 a and thereby decrease thefour-electron destabilizing interaction with (T^e- and in a second order way allow the increasedmixing of the latter with 2TT* (CO). Recent experimental work by Shriver and coworkers hasconfirmed this suggestion.175

(b) Electronegative substituents on the alkyl group raise the activation energy since they lowerthe energy of <7Me- and enlarge the four-electron destabilizing interaction with 5<r.

(c) The migration reaction is enhanced by good 7r-acceptor ligands such as NO at the metalcentre since these ligands are capable of accepting the additional metal electron density whichis released by the loss of a carbonyl acceptor orbital arising from the transformation of a carbonylligand into an acyl ligand during the methyl migration reaction.

(d) The methyl migration reaction (equation 2) for the related nitrosyl complex is less fa-vourable than the corresponding carbonyl reaction because the greater 7r-acceptor ability of NOresults in a greater loss in metal-ligand vr-bonding on formation of the alkyl-nitro complex.

(e) The activation energy for the process shown in equation (3) resulting in the formation ofan ethyl complex from an alkyl carbene complex is computed to be very low and has been attributedto the greater stability of the carbene pz acceptor orbital compared to the carbonyl 2TT* level.

Some of these conclusions have been corroborated by other researchers using CNDO and abinitio molecular orbital techniques.176178 Possible intermediates in the related hydroformylationreaction have also been investigated theoretically.179

27

Me . . L^ Me ^

O N - ^ + L — ^ N - 7 ^ - (2)

Me LMe. ^

H 2 C ^ ^ - + L —> CHz-^^— (3)


19.3 SANDWICH, TRIPLE DECKER SANDWICH, BENT SANDWICH AND HALFSANDWICH COMPOUNDS

19.3.1 Sandwich Compounds of the Transition Metals

There have been numerous semi-empirical,180"195 ab initiol96~l9S and Xa14'199 calculations

reported for sandwich compounds, and in particular ferrocene. Consequently, the essential bondingfeatures are now well understood. Figure 20 illustrates the important bonding interactions betweenthe vr-orbitals of the carbocyclic ligands and the metal valence orbitals in Fe^-CsHs^ andCr(ry-C6H6)2- In each case it is the metal ajg (4s), Q\u (4px, 4py), a2u (4pz) and eig (3dxz, 3dyz)orbitals which interact most strongly with the ligand orbitals of matching symmetry derived bytaking in-phase and out-of-phase linear combinations of the ligand 7r-molecular orbitals. Themetal e2g (3dx2-y2, 3dxy) and ajg (3dz2) orbitals remain essentially non-bonding, the former be-cause the d-type overlap with the ligand e2g set is not very large (particularly for ferrocene), andthe latter because the ligand 7r-orbitals of ajg symmetry point towards the nodal cone of the metal3dzi orbital as illustrated in (23). For larger carbocyclic ligands the metal-ligand e2g overlapbecomes more significant, and the energies of these orbitals are more favourably matched.Therefore, the e2g orbital becomes more bonding.

4p•MMWMK

a2inelu

elg'elu

e2g>e2u

elg'elu

a lg'a2u alg'a2u

M

(23)

£>5d Ferrocene Dibenzenechromium D6h

Symmetry (a) (b) Symmetry

Figure 20 Schematic illustrations of the important orbital interactions in (a) FeCrj-CsHsh and (b) 0(77-C6H6)2

Bonding ofUnsaturated Organic Molecules to Transition Metals 29

Table 6 Ground State Electronic Configurations and First lonizationPotentials of some Metallocenes M(r/-C5Rs)2 derived from UV

Photoelectron Spectral Studies1

Compound

V(r?-C5H5)2Cr(irC5H5)2Mn(r?-C5H5)2Mn(r/-C5Me5)2Fe(r?-C5H5)2Fe(r?-C5Me5)2Ru(r?-C5H5)2Os(T?-C5H4Me)2

Co(r?-C5H5)2Co(r?-C5Me5)2Ni(r?-C5H5)2

Electroncount

1516171718181818191920

Molecularground state

4 A l g (e 2 g ) 2 (a l g ) '3E2g(e2g)3(a lg)1

6Alg(e2g)3(alg)'(ey2

2E2g (e2g)3(a lg)2

'Aig (aig)2(e2g)4

•AigCaig)2^)4

^^(ajg)^)4

•A.gCezg^Ca.g)2

2E !g(a lg)2(e2g)4(efg)'2E lg(a lg)2(e2g)4(efg)'3A,g (alg)2(e2g)4(etg)2

Firstionizationpotential

(eV)

6.785.716.915.336.885.888.516.935.564.716.51

1. J. C. Green, Struct. Bonding (Berlin), 1981, 43, 37.

Table 7 Ground State Electronic Configurations and First Ionization Potentialsfor some Bis(arene) and Mixed Sandwich Compounds derived from UV

Photoelectron Spectral Studies1

Compound

Ti(if-C6H6)2V(r?-C6H3Me3)2Cr(7?-C6H6)2

Cr(r?-C6Me6)2M O ( T ? - C 6 H 6 ) 2

W ( T 7 - C 6 H 6 ) 2

Mn(r?-C5H5)(r7-C6H6)

Electroncount

16171818181818

Molecularground state

»A l g(e2g)4

2A l g(e2 g)4(a l gy1A l g(e 2 g) 4(a l g) 2

•A^CejgWa^)2

l A l g (e 2 g ) 4 (a l g ) 2

1A,g(e2g)4(aIg)2

'A.gCe^na,,)2

First ionizationpotential (eV)

5.505.335.455.015.525.406.36

1. J. C. Green, Struct. Bonding (Berlin), 1981, 43, 37.

The non-bonding nature of the metal aig and e2g orbitals in FeCrj-CsHs^ and Cr(r7-C6H6)2provides a facile explanation for the observation that it is possible to synthesise a wide range oftransition metal metallocenes with 15,16 and 17 valence electrons, examples of which can be foundin Tables 6 and 7. The existence of symmetrically bonded metallocenes with 19 and 20 valenceelectrons, e.g. Co(?7-C5H5)2 and NiCrj-CsHs^, is less easy to understand in the context of themolecular orbital diagrams shown in Figure 20, since they require electron population of theantibonding eig (3dxz, 2>dyz)* orbitals. This interesting anomaly will be discussed in more detailin Section 19.3.4.

The relative ordering of the highest occupied ajg and e2g levels in ferrocene and di (benzene) -chromium has been the subject of some controversy.35'36 Early semi-empirical molecular orbitalcalculations and the first ab initio SCF-LCAO calculations suggested the level ordering atg <2̂g < eig*, although the energy level sequence e2g < ajg < eig* was inferred from the optical ab-

sorption spectrum of ferrocene195 and supported by Xa calculations.199 UV photoelectron studiesusing both He-I and He-II radiation suggest within the Koopmans' Theorem approximation thatthe first two bands in the spectrum are assignable to 2E2g (6.88 eV) and 2Aig (7.23 eV) ion states,in agreement with the level ordering a)g < e2g < eig*.32'200~202 The relative intensities of the firsttwo bands reflect the different degeneracies of the e2g and aig orbitals, and the results are consistentwith ESR studies which suggest a 2E2g ground state for the [FeCiy-CsHs^]"1" ion.203204

Table 6 summarises the ground state electronic configurations and first ionization potentialsfor some metallocenes of the first row transition metals. It is interesting to note that althoughMn(r;-C5Hs)2 has a high spin 6Aig [(e2g)

2(aig)1(eig*)2] configuration, the corresponding pen-

tamethylated derivative Mn^-CsMes^ has a low spin 2E2g [(e2g)3(aig)

2] configuration. Electron


diffraction studies205 have demonstrated that this spin change is associated with a substantial(0.3 A) shortening of the metal-carbon bond length.

The molecular orbital interaction diagram for Cr(i7-C6H6)2 illustrated in Figure 20 is verysimilar to that of Fe^-CsHs^, but a slightly larger metal e2g (3dx2-y2, 3dxy) ligand e2g interactionresults in a stabilization of these orbitals relative to ajg {3dzi). I8i,i83,i86 Photoelectron spectralstudies206'207 have confirmed that the first ionization process in Cr(r;-C6H6)2 generates the 2Ajgion state. This ground state has also been inferred for the [Cr(?7-C6H6)2]+ ion on the basis of ESRstudies.186'208'209

A plot of ionization energies for the isoelectronic series of metallocenes Cr(77-C6H6)2,Mn(77-C5H5)(T7-C6H6), and Fe^-CsHs^ shown in Figure 21 effectively illustrates the cross overfrom (e2g)

4 (aig)2 to (aig)2 (e2g)4 configurations. The 3d valence state ionization energies of the

central metal atoms in these complexes are also plotted in the figure, and the fact that the 2Ajgionization energies closely parallel those of the isolated metal atoms is consistent with the non-bonding nature of the ajg (3dz2) orbital.35

7

8

Free atom

Table 7 summarizes the ground state electronic configurations and first ionization potentialsfor some bis(arene) and mixed sandwich compounds.3536 The very low ionization potentials ofthe bis(arene) and pentamethylcyclopentadienyl compounds210 are particularly noteworthy andGreen21' has argued that the low ionization potentials of these compounds have an importantbearing on their ability to activate small inert molecules, such as N2, and C—H bonds.

In recent reviews Cowley35 and Green36 have drawn attention to the following trends in ion-ization energies of transition metal sandwich compounds. (1) The appropriately weighted ion-ization potentials for related first, second and third row transition metal sandwich compoundsincrease in the order 3rd row « 2nd row > 1st row. A similar trend has been noted for metal car-bonyls and rationalized in terms of the ionization potentials of the isolated metal atoms. (2) The2E2g-2Ajg energy separation decreases with the atomic weight of the central metal atom for themetallocenes M(?7-C5H5)2, and this effect has been attributed to the increased covalent bondingin compounds of the second and third row transition elements. (3) If the average ionization energiesfor the ionizations from the ajg and e2g molecular orbitals are calculated by weighting each stateaccording to its relative theoretical cross section then it is found to increase with atomic numberacross the transition series from vanadocene to nickelocene. This of course reflects the effect ofincreased nuclear charge on the central metal atom. (4) Methyl substitution for hydrogen in ametallocene results in a decrease in I.P. of approximately 0.2 eV and for subsequent substitutionsthe effects are additive. Particularly noteworthy in this context is the very low first ionizationpotential of Co^-CsMes^ (4.71 eV), which can be compared with a first ionization potentialof 5.14 eV for gaseous sodium atoms.

1 1 I

(77-C6H6)2Cr C7-C6H6)(77-C5H5)Mn (*/-C5H5)2Fe

Figure 21 A plot of the energies of the 2Aig and 2E2g ion states for the isoelectronic series of metallocenesCr(i7-C6H6)2, Mn(r;-C5H5)(77-C6H6) and Fe^-CsHs^ derived from UV photoelectron spectral studies (reproducedwith permission from S. Evans, J. C. Green and S. E. Jackson, J. Chem. Soc, Faraday Trans. 2, 1972, 68, 249)


a

7s6d

~-alg 5/

(cot)4 - U(cot) U

19.3.2 Cyclooctatetraene Sandwich Compounds

The organometallic chemistry of the lanthanides and actinides has attracted considerable in-terest in recent years.212'216 Although the organo derivatives of these elements are invariablythermally very stable, the considerable degree of ionic character associated with the metal-carbonbonds in these compounds makes them extremely air and moisture sensitive. The isolation of ur-anocene, U^-CsHg^, by Streitweiser and his coworkers represented a particularly significantadvance in this area of chemistry.217'218 For the aromatic cyclooctatetraene dianion C8Hg~, the10 x-electrons occupy the following molecular orbitals: (aig)2 (eju)4 (e2g)

4, and consequentlyin the sandwich compound \J(rf-C%H%)2 the filled ligand 7r-orbitals transform as aig, a2U, eiu, eig,e2g and e2U. In addition to those interactions described above for af-block metallocenes there isthe possibility of bonding interactions between the metal/-orbitals, which transform as a2u, e!u,e2u and e3U in the D^h point group, and the symmetry matching ligand 7r-orbitals.219 Streitweiseroriginally proposed that the interactions between the metal and ligand orbitals of e2U symmetrywere particularly important in stabilizing this class of molecule. A recent MS-Xa calculation,220

the results of which are schematically illustrated in Figure 22, have demonstrated that the metal6d-orbitals also make a significant contribution to bonding and have the effect of depressing thee2g level below that of e2U. The He-I and He-II photoelectron studies on Th(r]-Cslls)2 (Z0) andLK^-CsHg^221"223 which are summarised in Table 8 are consistent with the level ordering indi-cated by the Xa calculations. The additional low ionization energy band observed in the spectrumof U(77-C8Hg)2 when compared with that of Th^-CgHg^ has been assigned to ionizations of thef1 configuration.

Figure 22 A schematic illustration of the molecular orbitals of U(?]-C8H8)2 (adapted from N. Rosch and A.Streitweiser, Jr., J. Organomet. Chem., 1978,145, 195)

Table 8 Ionization Energies21 for Cyclooctatetraene ActinideComplexes MC^-CgHs)!1

M f" e2u e2g e !u, eig

Th 6.79 7.91 9.90, 10.14, 10.65U 6.20 6.90 7.85 9.95, 10.28, 10.56

a In eV.1. J. C. Green, Struct. Bonding (Berlin), 1981, 43, 37.

19.3.3 Mixed Sandwich Compounds

Clack and Warren225"23' have made an extensive investigation of the bonding in the 18-electronmixed sandwich compounds (r;-C5H5)ML (L = 77-C3H3,77-C4H4,17-C5H5,77-C6H6 and r]-CiYl-i)and concluded that the energy of the carbocyclic ligand 7r-orbitals of ej symmetry plays an im-portant role in influencing the charge distributions in the two rings and the relative strengths ofthe metal-carbon bonds. These results based on IND-SCF-MO calculations confirm the essentialfeatures of an earlier and important bonding study by Fischer.183

(24)

5e

A

50

B

50

C

0

5e

4e

4e 1

6a

77

•®xC)50

17

59

16

50

45

29

0

26

Figure 23 The frontier molecular orbitals of the mWo-icosahedral B| 1H11 fragment (24). For reasons of clarityonly the orbital contributions of the atoms contained in the open pentagonal face of the ligand are shown. Theseatoms are designated (A), the five boron atoms in the closed pentagonal face are designated (B) and the cappingatom (C). The relative contributions of these atoms to the molecular orbitals are indicated at the right hand sideof the figure


UV photoelectron spectral studies of tris- and tetrakis-cyclopentadienyl, M(r]-CsHs)3,M ^ - C s H s ^ and M(?7-C5H5)3*THF complexes of the lanthanides and actinides have also beencompleted224 and the reader is referred to two recent reviews35'36 for a detailed discussion of theelectronic structures of these compounds.

19.3.4 Sandwich and 'Slipped' Sandwich Compounds derived from Carboranes

The original synthesis of metallocarboranes was affected by deprotonation of mafo-carboraneanions according to equation (4)232-234 ancj resuJted from a very perceptive analogy between thefrontier molecular orbitals of C2B9H?7 and those of cyclopentadienyl C5H5 by Hawthorne. Ex-tended Huckel molecular orbital calculations on the Bj jHn~ fragment (24), which is isoelectronicand isostructural with C2B9H?]", are reproduced in Figure 23. These have confirmed that the 6ai,5ei and 5e2 molecular orbitals are localized predominantly on the open pentagonal face of theborane anion and bear a strong similarity to the TT-MOS of cyclopentadienyl.235'236 However, in

C2B9H1 2- + NaH —* C 2 B 9 H n 2- .?'"2,Fe(II)> ^Fe(C2B9H12)2- (4)x JL * V - ^

Bonding of Vnsaturated Organic Molecules to Transition Metals 33

12+

a

(25)

Figure 25 illustrates the Walsh diagram for the simple lateral 'slip' distortion shown in (25)derived from extended Hiickel calculations on the hypothetical [Cu(Bj jHi 1)2]w~ anion. For an18-electron d6 complex, the highest occupied level is 5eju and results from the bonding combinationof the metal x, y orbitals and the antisymmetric combination of the ligand 5ej set. This contrastswith the situation in metallocenes (see Section 19.3.1) where the highest occupied levels are derivedprimarily from the metal ajg (dzi) and e2g (dx2-y2, dxy) orbitals, and is a reflection of the lower

contrast to the cyclopentadienyl anion the p^-orbitals of the 6ai, 5ei and 5e2 MOs of BnH27 arenot orthogonal to the open pentagonal face and are rehybridized in such a way that they tilt in-wards. This rehybridization effect arises primarily because the hydrogen atoms attached to theboron atoms, which define the open pentagonal face, do not lie within that plane but above it.

For Bi iHfr there are four additional MOs 4ei and 4e2 which lie between 5&\ and 6ai in energyterms. These orbitals have their maximum amplitudes within the open pentagonal face of the nidopolyhedron and overlap less effectively with the frontier orbitals of the metal. The role of the lowerlying 4ei set is particularly small since it is localized predominantly on the closed pentagonal faceand the capping atom rather than the open face.

The structures of the metallocarborane sandwich compounds synthesized by Hawthorne andhis coworkers are illustrated in Figure 24. The structures of these compounds depend markedlyon the number of ^-electrons present. Those complexes with d5, d6 and d1 electronic configurations,e.g. Fe(III), Co(III) and Ni(III), exhibit a symmetrical sandwich structure (Figure 24a) in whichthe metal atom is nearly equidistant from five carbon and boron atoms from each cage, and thestructure can be viewed as two icosahedra linked through the metal atom. Complexes which havethe carbon atoms in adjacent positions adopt the transoid configuration illustrated in the figurewhereby the carbon atoms in the two icosahedra are as far away from each other as possible. Acisoid arrangement, however, has been found in the d6, Ni(IV) complex Ni(r;-C2B9H] 1)2 and.is illustrated in Figure 24b.233'234

Figure 24 The structures of some sandwich compounds derived from carboranes. The transoid and cisoidstructures found in d5, d6 and d1 complexes are shown in (a) and (b). The slipped structure characteristic of d*and d9 complexes is shown in (c)

In the metallocarborane sandwich compounds which have more than seven ^/-electrons, e.g.d* Ni(II), a distortion from the symmetrical sandwich geometry is observed. Instead of beingequally bonded to three boron atoms and two carbon atoms the metal occupies a position closerto the boron atoms than to the carbon atoms (see Figure 24c for example). In its simplest formthis distortion may be represented by a lateral displacement A of the dicarbollide ligands as il-lustrated in (25), leading to pseudo-7r-allylic bonding at the metal.237 However, it has been notedthat this 'slip' distortion is also accompanied by a slight folding of the C2B3 face away from themetal atom in such a way that the C2/1 symmetry of the molecule is maintained.238 Severalqualitative bonding models were originally proposed to account for these distortions.239242


electronegativity of boron compared with carbon. The lowest unoccupied level in [Cu(Bi jHi 1)2] n~is 5eig (xz,yz)* which is the antibonding combination of the metal xz.yz orbitals with the sym-metric component of the ligand 5ej set. A similar degenerate set is the lowest unoccupied levelin Fe^-CsHs^ (see Figure 20).

The 'slip' distortion is unfavourable for 18-electron complexes because the reduced metal-ligandoverlap integrals which arise from the distortion reduce the bonding nature of the au and bucomponents of the 5eju set shown in Figure 25, and the lower lying 5eig (xz.yz) bonding orbitals.However, when the 18-electron configuration is exceeded the additional electrons populate theantibonding 5eig (xz, yz)* set, whose components (and in particular the ag component shownin Figure 25) are stabilized substantially by the 'slip' distortion. The stabilization of the ag com-ponent is a result not only of a diminution of metal-ligand overlap, but also of a mixing in of metals-orbital character which is symmetry forbidden in the undistorted structure with D^d sym-metry.243

D A(A) C2h

Figure 25 The Walsh diagram for the lateral slip distortion in the sandwich compound [Cu(BnHn)2]/l • Theelectron occupations shown correspond to the 18-electron situation

The extended Hiickel calculations have indicated that single electron occupation of the agcomponent of 5eig (xz, yz)* in a 19-electron d1 complex provides an insufficient driving forcefor the 'slip' distortion and only in 20-22-electron complexes does the stabilization associatedwith the ae and be levels exceed the destabilization associated with the filled bonding levels.243

This is of course in complete agreement with the X-ray analyses on dicarbollide sandwich com-pounds described above and illustrated in Figure 24.237'239


The calculated atomic charges for the unslipped and slipped [Cu(Bj 1H1 i)2p~ complex illus-trated in (26) indicate a build up of electron density at those boron atoms which lie farthest fromthe metal atom, which provides a ready explanation for the observation that the more electro-negative carbon substituents in [Cu(C2B9Hj 1)2]" occupy these positions in the 'slipped' structure(see Figure 24c).

-0.11

0.24 / ^ N -0.08

Cu -0.29

A =0.4 A

(26) (27) (28)

19.3.5 Triple Decker Sandwich Compounds

The possibility that transition metals might form triple decker sandwich compounds was in-dicated by early mass spectrometric results on Ni^-CsHs^ and Fe^-CsHs^,252 but stable de-rivatives have only been isolated and characterized in recent years. Werner and Salzer253 reportedthe synthesis of the 34 valence electron cation [l^^-CsHsh]"1" in 1972, and subsequentlyGrimes,254'255 Herberich256'257 and Siebert258'259 have synthesized a wide range of triple deckersandwich compounds with 30 valence electrons based on carborane, azacarborane and thiacar-borane ligands, e.g. C4BH J, C2B2SH4-, C2BNH5 and C3B2H5~. Some of the complexes derivedfrom these ligands are illustrated in (29). These complexes may also be viewed as metallocarbo-ranes based on the pentagonal bipyramid and their electronic structures rationalized within theframework of the Polyhedral Skeletal Electron Pair Theory as described in Section 19.2.6above.

Co FeB—B/ \

Co Fe

(29)

By analogy with these metallocarboranes it might seem reasonable to suppose that the 19- and20-electron metallocenes Co^-CsHs^ and N^ry-CsHs^ would also show an inclination to undergo'slip' distortions. Indeed the longer metal-ring distances and smaller metal-ligand bond disso-ciation energies in these compounds compared to Fe(r7-C5H5)2 provide convincing evidence thatthe additional electrons in the cobalt and nickel complexes occupy the antibonding 5eig (xz,yz)*orbitals.244"246 The smaller degree of covalent character in the 5eig (xz,yz)* set of the metallo-cenes (arising from the higher electronegativity of carbon compared with boron), however, providesa less effective driving force for such a distortion. It is significant in this context that the 20 electroncomplex W(CO)2(ry-C5H5)2, where the degree of covalent character in the metal-cyclopentadienylbond is greater, shows a 'slip' distortion resulting in the 18-electron allylic structure shown in(27).247 j ^ e structure of the Ru(?7-C6Me6)(7?4-C6Me6) complex (28) can similarly be accountedfor in terms of the 18-electron rule.248"251 The calculations on [Cu(Bi iHj 1)2] n~ also suggest thatthe introduction of electron withdrawing substituents in 1,2-positions of the cyclopentadienylligand should facilitate the 'slip' distortion in nickelocene and related derivatives.


Hoffmann and Lauher260 have successfully accounted for the electronic structures of thesetwo classes of triple decker sandwich compounds using a fragment molecular orbital analysis.The [Ni(?7-C5H5)] fragment has frontier molecular orbitals which closely resemble those describedin Section 19.2.6 for Fe(CO)3. Figure 26 shows that [Ni(77-C5H5)] has three filled low lyingmolecular orbitals which are essentially non-bonding and have predominantly z2 (aj), and x2

— y2, xy (e2) character. The higher lying three orbitals comprise an ei orbital set, which areoutpointing xz, x and yz, y hybrid orbitals, and an orbital of ai symmetry, which is an outpointing5, z hybrid. These orbitals are additionally somewhat antibonding with respect to the ring.

hya I

e ixz yz

z2

ai

xy x2-y2

Figure 26 A schematic illustration of the frontier molecular orbitals of the Ni^-CsHs) fragment

Figure 27 illustrates schematically the interactions between the frontier orbitals of the twoNi(77-05115)2" fragments at the correct internuclear distance for a triple decker sandwich compoundand the central C5H5" ligand. The left hand side of Figure 27 shows how the important frontierorbitals of the two Ni^-CsHs)"1" fragments interact in the absence of the central ring. The eiorbitals split only slightly since they have mainly d character and there is little net overlap. Theoverlap of the aj orbitals is more significant because of the higher proportion of metal s and pcharacter and in- and out-of-phase combinations give rise to bonding a / and antibonding a2r/

orbitals. The right hand side of Figure 27 shows the x-orbitals of the central cyclopentadienylligand. The a2r/ and ei" orbitals interact strongly with the corresponding orbitals of the Ni(t7-C5H5)"1" fragments. In addition there are significant interactions between the (7-orbitals of thecentral ligand of a / and e / symmetry (not shown in the figure for reasons for clarity) and themetal a / and e / orbitals leading to a destabilization of the latter.

In the 30 valence electron triple decker sandwich compounds, the bonding ei" molecular orbitalwhich is delocalized over all three rings and the metal atoms is occupied and the highest occupiedlevels are those which originated from the non-bonding aj (z2) and e2 (x2 — y2, xy) orbitals ofthe isolated Ni^-CsHs)"1" fragments, i.e. the electronic configuration is (e/04 (a/)2 (e2r)4 (e2r/)4

(a2r/)2- However, there is also a low lying level of e / symmetry which is somewhat antibondingwith respect to the outer rings and essentially non-bonding with respect to the central ring. Thisorbital is occupied by four electrons in the 34 valence electron [Ni2(r;-C5H5)3]+ ion.

Arising from this theoretical analysis Moraczewski and Geiger261 have shown that the 36-electron cyclooctatetraene complex (30) undergoes a reversible two electron oxidation to give[Co2(?7-C5H5)2(?7-C8H8)]2+ which presumably has the 34-electron triple decker sandwichstructure illustrated in (31). The instability of the complex (31) towards nucleophilic attack hashindered its complete structural characterization unfortunately.


a

a

a,"+eam i

+e2 '+a,'a

C5H5

Figure 27 A fragment molecular orbital analysis of the bonding in the triple decker sandwich compound[Ni2(T?-C5H5)3]+ (adapted from J. W. Lauher, M. Elian, R. H. Summerville and R. Hoffmann, J. Am. Chem.Soc, 1976,98,3219)

-2e.

2+

(30)

\ /

(31)

19.3.6 Bent Sandwich Complexes

In bent bis(cyclopentadienyl) complexes the rings are not parallel, i.e. the angle between thenormals to the planes defined by the cyclopentadienyl rings is less than 180°. Some of the manyexamples of bent sandwich compounds are illustrated in (32).262 Crystallographic studies haveestablished that the inter-ring angle (6) can vary from 148° in MOCTJ-CSHS^I^ to 126° inZr^-CsHs^b-263"265 From the examples illustrated above it is clear that these compounds donot rigidly conform to. the 18-electron rule, and the existence of structurally related complexeswith 16, 17 and 18 electrons, e.g. Mo(77-CsH5)2X2 (X = halide or pseudohalide), has provideda unique opportunity for testing the theoretical model developed for these compounds using ac-curate X-ray crystallographic structural determinations,266269 UV photoelectron spectral270

and electron paramagnetic resonance studies. These experimental studies have substantiated themolecular orbital picture developed for these complexes by M. L. H. Green, J. C. Green andProut,266 Peterson, Lichtenberger, Fenske and Dahl269 and Lauher and Hoffmann.271 See refs.272-274 for earlier accounts of the bonding in these complexes.

37


HM

H.

HB

H

HM—H

The way in which the d-type molecular orbitals of the M(r7-C5H5)2 fragment change as afunction of the bending angle 6 is illustrated in Figure 28.271 For complexes with d°-d]01 config-urations there is no driving force for the sandwich complex to take up a bent configuration, andthe additional ligands which coordinate to the metal provide an essential stabilizing interactionfor the formation of bent sandwich compounds. Typically the metals in such complexes have fouror fewer valence electrons, and therefore, it is the three low lying levels illustrated in Figure 28that play an important role in the formation of coordinate bonds to the additional ligands L.Contour plots of the relevant b2, lai and 2ai molecular orbitals derived from extended Huckelcalculations are illustrated in Figure 29.271 The b2 molecular orbital has primarily dyz characterand the two ai orbitals contain contributions from the metal s, pz, dx2-yi and dzi orbitals. The1 a i molecular orbital is directed primarily along the y axis (see Figure 29 for a definition of theaxis system) and may be approximately described as a dyi orbital with a small admixture of dxi-z2.Consequently this orbital remains essentially non-bonding or at the most weakly bonding withrespect to ligands L located in the zy plane of the molecule. The 2a i orbital is hybridized awayfrom the C5H5 ligands and its directional characteristics lead to very good overlap with the donororbitals of the additional ligands L. The strong interactions between 2a i and the symmetric ligandcombination, and b2 and the antisymmetric ligand combination are important features of thecorrelation diagram for M(T7-C 5 H 5 ) 2 L 2 complexes shown in Figure 30. In an 18-electron d2

complex the highest occupied molecular orbital is lai, which retains its non-bonding nature ifthe ligands are located along the nodes of this orbital. Since these nodal lines make an angle ofapproximately 78°, an L—M—L bond angle of approximately 78° is anticipated for thesecomplexes. For 16-electron d° complexes, maximal metal-ligand bonding is achieved if the liganddonor orbitals overlap effectively with both 2a i and lai, thereby raising the energy of the empty

e2g>e2-11

la

120'

38

d° 18-electron d° 16-electron dl 17-electron d4 18-electrondl 17-electron d2 18-electrond2 18-electron

(32)

M e—* M

Figure 28 Molecular orbitals of M (r7-C5H5)2 as a function of the bending angle 6 (reproduced with permissionfrom J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc, 1976, 98, 1729)

Bonding of Vnsaturated Organic Molecules to Transition Metals 39

1

IFigure 29 Electron density contour maps of the three frontier orbitals of the M^-CsHs^ angular fragmentThe plots represent a section through the yz plane (reproduced with permission from J. W. Lauher and R. Hoff-mann, 7. Am. Chem. Soc, 1976, 98, 1729)

a, + b2

Figure 30 Interaction diagram for M(rj'CsHs)2^2 formed by the combination of M(r;-C5H5)2 (d°) and two(T-donating two-electron ligands L? (reproduced with permission from J. W. Lauher and R. Hoffmann, J. Am.Chem.Soc, 1976,98, 1729)

la] orbital. This is achieved by an opening out of the L—M—L angle from 78° to approximately110°. For a dx complex a compromise between these two effects leads to a computed intermediateangle of 85°.


The structural data for M(rj-Csli5)2X2 complexes have been summarised by Prout and co-workers. They find that the ranges of L—M—L angles experimentally determined for dn com-plexes are 94-97° (d°), 85-88° (d]) and 76-82° (d2). These values are in complete agreementwith the molecular orbital calculations described above.263266

Very elegant ESR studies of the dx complexes V^-CsHs^Ss and V(?7-C5H4CH3)2Cl2 byPeterson and Dahl267"269 have confirmed that the odd electron does indeed reside in an a 1 orbitalwhich is localized predominantly on the metal and has a high proportion of dyi character. Fur-thermore, UV photoelectron studies by Green, Jackson and Higginson and Dahl, Fenske and theircoworkers have supported the general features of the bonding model developed above.269'270 Inaddition, these studies have shown clearly the presence of two non-bonding molecular orbitalswith ionization potentials between 6 and 7 eV for M^-CsHs^X compounds, one non-bondingmolecular orbital for M(r;-C5H5)2X2 and none for M^-CsHs^Hs, corresponding to the differentdegrees of utilization of the laj, 2a 1 and b2 molecular orbitals in the bent sandwich compounds.An important consequence of this bonding model for the M(r;-C5H5)H3 complexes is that thehydride ligands all lie in the yz plane since this arrangement maximizes the overlap between themetal laj, 2a 1 and b2 orbitals and the hydrogen ls-orbitals.275

19.3.7 Half Sandwich Compounds

Half sandwich compounds of the general type M(r/-cyclic polyene)Lw represent a major classof transition metal organo derivatives. When L is a good 7r-acid ligand, e.g. CO or NO, thesecomplexes adhere rigorously to the 18-electron rule and consequently their stoichiometries canbe readily predicted on this basis. In (33) a typical series of complexes of this type is illustrated.The existence of an extensive series of isoelectronic complexes of this kind has proved to be in-valuable for comparative spectroscopic and theoretical studies (see refs 276-297 for exam-ples).

/ C o \OC CO

(33)

When L is not such a good 7r-acceptor ligand, e.g. NH3, PR3, SMe2, etc., then the 18-electronrule has less predictive value and some examples of M(r/-cyclic polyene)L2 complexes with 16and 17 valence electrons have been reported in addition to those with 18 valence electrons (see(34) for some typical examples). Besides showing a disregard for the 18-electron rule, suchcomplexes frequently have distorted geometries. These distortions are at times of a minor natureas in (34d) where the benzene ring is no longer planar but takes up the boat conformation illus-trated in (35),298 or major in the case of Cu(PR3)(i7-C5H5) (36a),299 which takes up the 14-electronmonohapto geometry in solution rather than the 18-electron pentahapto geometry (36b).These distortions have been the subject of several theoretical analyses,145'300303 the results ofwhich will be briefly summarized below.

BPh

F5C6 C6F5 F5C6 C6F5

16-electron 17-electron 18-electron 18-electron

(a) (b) (c) (d)

(34)


BPh

R3P PR3 (a) (b)

(35) (36)

(c)

(a) (b)

(37)

(38)

In the previous section an analysis of the frontier molecular orbitals of the M^-CsHs) fragment(see Figure 26) indicated that its bonding capabilities are decided primarily by a set of threeout-pointing hybrid orbitals of aj and e} symmetry. These orbitals lie at higher energies than thenon-bonding ai {dzi) and e2 (dx

2-y2> dxy) orbitals by virtue of the antibonding nature of the dxz,dyz interaction with the ligand ei set, which is only partially mitigated by the rehybridizationresulting from p-orbital mixing (see (37)). Similar arguments also apply to the higher lying ajhybrid orbital. Other M(ri-Cn¥in) fragments have a similar set of out-pointing ei and ai molecularorbitals capable of interacting with the orbitals of the additional ligands L.

Figure 31 illustrates the interactions between these frontier molecular orbitals and those ofa simple a-donating ligand to generate the molecular orbitals of the 18-electron complexesM(T?-C 5 H 5 )L , M(r/-C5H5)L2 and M(77-C5H5)L3. In the M(77-C5H5)L complex (Figure 31a),the ligand <r-donor level interacts exclusively with the ai hybrid frontier orbital of M^-CsHs),and leaves the Q\ hybrid set unaffected. If L is additionally a 7r-acceptor ligand with low lying7T* orbitals of ei symmetry then the ei hybrid set can be substantially stabilized (see (38)). Anindirect consequence of this interaction is a reduction in metal-ring antibonding character whichresults from the more extensive delocalization shown in (38). This situation exists in symmetricalpentahapto complexes such as Ni^-CsHsXNO) (33) and Cu^-CsHsHCO). When L is not agood 7r-acceptor ligand then alternative means of reducing the antibonding nature of the C\M^-CsHs) orbital are sought. A 'slip' distortion similar to that described in detail in Section19.3.4 for metallocarborane sandwich compounds provides an effective way of reducing the an-tibonding nature of the ei set (see Figure 25). In the limit such a 'slip' distortion can result in themonohapto geometry illustrated in (36a) or the dihapto geometry illustrated in (36c). ForCu^'-CsHsXPR^) the former is preferred in solution and the low activation energy fluxionalbehaviour reported for this compound probably proceeds through a dihapto transition state closelyrelated to (36c).299'304


a,(hy)

x2-y2,xy

Cu(7-C5H5) L

(a)

ai(hy)

z2

x y2,xy aib,

Co(rj-C5H5) Co(rj-C5H5)L2

(b)

L

a,(hy)

x2 -

In M(?7-C5H5)L2 complexes the ligand donor set transforms as ai and b2 and consequentlythe antisymmetric ligand combination interacts only with one component of the M^-CsHs) c\set leaving the other component unaffected (see Figure 31b). The latter can be stabilized by the7r*-levels of the ligand L, if L = CO, NO or a similar 7r-acceptor ligand, in the manner illustratedin (39). In the absence of such ligands, a 'slip' distortion would be anticipated for M^-CsHs)!^complexes. However the magnitude of the distortion is likely to be less than that discussed abovefor M^-CsHsJL since only one of the metal-ring antibonding ei components is involved. Thesoft nature of the potential energy surface for these distortions has been illustrated by a detailedstructural study of the isoelectronic metallocarboranes (40), (41) and (42). For these complexesWelch et al.305~301 have shown that the direction of the 'slip' distortion and its magnitude maybe influenced by the substitution pattern of the carbon atoms in the open pentagonal face of theligand. In (40) a pseudotrihapto allylic geometry is observed and (41) has a pseudotetrahaptogeometry. Wallbridge et a/.308"309 have also shown that the extent of the 'slip' distortion in such

Mn(/7-C5H5) Mn(77-C5H5)L3 L3

(c)

Figure 31 Schematic interaction diagrams which illustrate the bonding in the 18-electron half sandwich com-pounds Cu(77-C5H5)L (a), Co(r/-C5H5)L2 (b) and Mn(T7-C5H5)L3 (c)


complexes depends to a large extent on the 7r-acceptor qualities of the ligands coordinated to themetal and the metal d-p promotion energy (see ref. 310 for the effect of cage size).

(39)

N

(40) (41) (42)

In cyclopentadienyl complexes the distortions take on a less dramatic form with the metalshowing little or no displacement from the central position. However, the very different bondinginteractions between the ligand ei set and the metal dxz and dyz orbitals do result in a significantvariation in the carbon-carbon bond lengths within the cyclopentadienyl ring. In Co^-CsMes)-(CO)2 an 'allyl-ene' localization of the carbon-carbon bond lengths is observed (see Figure 32a),31 *whereas for Rh^-CsClsHdiene) a 'diene' localization bond pattern has been noted (see Figure32b).312 This interesting difference results from the fact the nodal characteristics of the HOMOin M^-CsHs)!^ complexes depend on their conformations. In Co(?7-C5Me5)(CO)2 the M(CO)2plane lies perpendicular to the mirror plane passing through the unique carbon atom of the cy-clopentadienyl ring (43) and consequently the HOMO, illustrated in Figure 32(c), has nodal

1.410(4)

1.446(4)1.392(6)

1.433(7)

1.399(6)

1.441(8)

(a) (b)

(c) (d)

Figure 32 Illustration of the influence of conformation on the C—C bond lengths in M(T;-C5R5)L2 18-electroncomplexes. The observed conformations and bond lengths in Co(r7-C5Me5)(CO)2 and Rh^-CsClsHdiene) areillustrated in (a) and (b) respectively and related to the nodal characteristics of cyclopentadienyl 7r-orbitals inthe highest occupied molecular orbital (c) and (d)


characteristics which result in the observed 'allyl-ene' bond localization pattern. For Rh(r;-CsClsHdiene) the alternative conformation (44) is observed in the solid state and consequentlythe HOMO in this complex (see Figure 32d) reflects the nodal characteristics of the comple-mentary component of the cyclopentadienyl ei set. The nodal characteristics of this ei componentgive rise to the observed 'diene' localization bond length pattern. The antibonding nature of themetal-ligand interaction in the HOMOs illustrated in Figures 32(c) and 32(d) also results ina slight folding of the cyclopentadienyl313"315 ligand plane, an effect which has also been notedfor the complex Rh(r7-C6H5BPh3)(P(OMe)3)2 (35), where the dihedral angles defining the boatdistortion are 7.0° and 4.9°.298 In this substituted benzene complex the HOMO has the nodalcharacteristics shown in (45) and the boat distortion reduces the metal-C-1 and -C-4 antibondinginteractions.300 A similar boat distortion has been observed in the 18-electron complex Ni(?j-PhMe)(C6F5)2 (34c).316 The related 17-electron complex Co(r7-PhMe)(C6F5)2 (34b)317 whichhas only a single electron occupying the orbital illustrated in (45) has a planar ring and shortermetal-arene carbon bond lengths, attesting to the fact that (45) is indeed antibonding betweenmetal and the arene ligand.

(43) (44) (45)

From the interaction diagram shown in Figure 31 (b) it is clear that the alternative symmetricconformations for M(r/-CsH5)L2 complexes (43) and (44) are equienergetic and consequentlya low rotational barrier is anticipated, and indeed observed, for M(r?-CnHrt)(CO)2 complexes.Computed rotational barriers for these complexes range from 0.8-0.04 kJ mol"1.300 Removalof the degeneracy of the M(ti-CnHn) Q\ set, for example, by the introduction of heteroatoms intothe polyene, can generate substantial (40-80 kJ mol"1) rotational barriers. Detailed discussionsof such effects are to be found in refs. 235, 300 and 318.

The discussion as presented above has attributed the distortions in M(r;-CwHn)L2 18-electroncomplexes exclusively to the nodal characteristics and antibonding nature of the HOMO. It followsthat the corresponding 16-electron complexes should adopt more symmetrical structures. Whilstthis is generally true, a caveat has to be added regarding 16-electron cyclopropenyl complexesof the type [M(r?-C3R3)(PR3)2]+ (M = Ni, Pd or Pt) which are known to adopt highly distortedstructures (see Figure 33 for examples319'320). For the cyclopropenyl cation, the ej set of 7r-mo-lecular orbitals shown in (46) and (47) shows a high degree of localization and consequently theiroverlap integrals with a metal dxz orbital, for example, will be very sensitive to the location ofthe metal atom above the ligand plane. For example, the change from ry3 to rj2 coordination modeshown in (48) can dramatically improve the dxz ligand overlap. Consequently, the observed dis-

2.05(1) / ', X 1.91(1) 2.09(1)/ ; V^.2.48(l)' ' ' X / '

/ ; x / ; j .

[Ni(7?-C3Ph3)(PPh3)2]+ [Pt(^C3Ph3)(PPh3)2]+

(a) (b)

Figure 33 Illustration of the influence of conformation on the M—C and C—C bond lengths (A) in 16-electronmetal cyclopropenyl complexes


tortions in [M(r;-C3R3)(PR3)2]+ complexes (Figure 33) can be interpreted readily in terms ofthe nodal characteristics of this important bonding component. Alternatively, they can be un-derstood in terms of the localized representations shown in (49) and (50).300 Ref. 300 also givesan interesting discussion of the fluxional behaviour of such complexes and suggests that the per-mutation of phosphine environments observed in the NMR spectra does not proceed through asymmetrical transition state with the metal located above the centre of the triangle.

(46) (47)

(48)

(49) (50)

In the 'piano stool' MXTJ-CSHS)!^ polyene complexes, the ligand donor set transforms as aiand e and consequently interacts strongly with all three of the frontier orbitals of M^-CsHs)(see Figure 31c). Consequently, the most favourable two-electron metal-ligand stabilizing in-teractions are achieved when the M^-CsHs) ei and ai orbitals are empty, i.e. when M = d6 metalion. For such complexes the ?r-bonding abilities of the ligand are not important for stabilizingthe complex, and therefore complexes of this type with cr-donor ligands, e.g. [Rh^-CsHs)-(OCMe2)3p+ are reasonably common.321 No distortions are anticipated for this class of compoundsince both dxz and dyz are involved equally in metal-ligand bonding. Distortions can, however,be induced by an asymmetric ligand set, e.g. in complexes of the type M^-CsHs)!^!/, particularlyif the bonding capabilities of the ligands L and L' are very different, e.g. if L = PR3 and 1/ = H,or L = PR3 and V = CR3.322323

If the total electron count in a M(r]-CnHn)L3 complex exceeds 18 then the additional electronshave to be accommodated in the antibonding components of the dxz, dyz ligand ei molecular or-bitals, and the antibonding nature of the resulting interactions may be relieved by a 'slip' distortion.The structures resulting from these 'slip' distortions can be predicted accurately from the 18-electron rule and some examples are illustrated in (51). For the 18-electron structures shown in(51) there are the alternative 18-electron charged or zwitterionic structures illustrated in (52),which may also be derived from 'slip' distortions and presumably have energies similar to thecorresponding structures shown in (51).324 With these alternative structures available to thecompounds shown in (51) it is not altogether surprising that many of these compounds are fluxionalwith the metal fragment generally undergoing rapid 1,2-haptotropic shifts around the polyeneligand.325-330


CO

(51)

CO

(52)

19.3.8 Aromaticity in Sandwich Compounds

Soon after the discovery of ferrocene it was recognised that one of its characteristic propertieswas its ability to undergo electrophilic substitution reactions with reagents which did not causeits oxidation to the ferrocenium cation.331 Therefore, the chemistry of ferrocene is certainly thatof an aromatic molecule. However, it is difficult to extend the other physically based criteria foraromaticity used in organic chemistry, e.g. thermodynamic, diamagnetic anisotropies and l HNMR data, to define the extent of aromaticity in this and related organometallic molecules.332

Based on the simple chemical reactivity criterion it is interesting that Mn(r/-CsH5)(CO)3 likeferrocene may be classified as aromatic,333 but the corresponding benzene complexes are muchless susceptible to electrophilic attack. Cr(77-C6H6)2 does not show any reactivity with Friedel-Crafts reagents,334 and Cr(r?-C6H6)(CO)3 reacts only with difficulty.335 Certain cyclobutadienecomplexes, e.g. Fe(i7-C4H4)(CO)3 and Coirj-C^^irj-CsHs) also show 'aromatic properties'336

and for the former compound these properties have been widely utilised in the synthesis of highlystrained organic molecules.

The important chemical differences between these molecules clearly require theoretical in-terpretation, although the absence of rigorous quantum mechanical criteria for aromaticity makesit almost impossible to define precisely the electronic origins responsible for these differences.Indeed, as a result of an extensive investigation of the theoretical and experimental data availablefor a wide range of aromatic organic molecules, Lewis and Peters were forced to adopt the con-servative definition of aromaticity as 'having an electron organisation like that of benzene'.332

Fenske and Bursten337 have argued that the reduced aromaticity of Cr(?7-C6H6)2 and Cr(r/-C6H6)(CO)3 compared with benzene arises because of the synergic bonding interactions betweenthe benzene eig and e2g 7r-orbitals and the metal eig (dxz, dyz) and e2g {dxi-yi, dxy) orbitals. Thesereduce the extent of cyclic delocalization in the coordinated benzene molecule. In contrast, thearomatic properties of Fe^-CsHs^ and MnCrj-CsHsXCO^ are thought to arise because themetal back donation from the dxi-yi and dxy orbitals to the cyclopentadienyl e2 7r-orbitals is lessextensive because they are more antibonding, and consequently the metal provides an effectiveway of stabilizing the aromatic C5H J entity.

The aromatic properties of cyclobutadiene complexes, e.g. Fe(??-C4H4)(CO)3, are particularlydramatic by virtue of the antiaromatic properties of the free cyclobutadiene molecule. The conceptof 'metallo-aromaticity' has been introduced337 to account for the remarkable properties of thistype of compound. Metallo-aromaticity was introduced as a heuristic device rather than a rigorousdefinition to indicate that in such a molecule the cyclobutadiene /vorbitals and metal dxz anddyz orbitals have the correct topology and electronic requirements to generate a three dimensionalelectronic system with effective electron delocalization. The three dimensional network arisingfrom the overlap of the carbon /?-orbitals and the metal dxz, dyz set shown in (53) correspondsto a six-vertex graph with two phase dislocations indicated by ~ in (54) and gives rise to threehighly delocalized and stable bonding molecular orbitals338"340 which in Fe(CO)3(r;-C4H4) are


occupied by the four electrons contributed by the cyclobutadiene and the two electrons contributedby the Fe(CO)3 fragment. The electron delocalization is particularly effective in this moleculeowing to (a) the good match of cyclobutadiene and iron orbital energies, (b) the good overlap (S= 0.18) between metal and ligand orbitals and (c) the geometry of the carbonyl ligands whichallows a good separation of metal-cyclobutadiene and metal-carbonyl interactions.337 Whenthese requirements are not met, for example in Cr(r;-C4H4)(CO)4, then there is a correspondingdecrease in those chemical properties associated with aromaticity.

(53) (54)

The topological aspects of this model have been utilised, in conjunction with the concepts ofthe perturbation molecular orbital method, to account for the bond lengths and chemical propertiesof a wide range of cyclobutadiene and diene complexes of Fe(CO)3.338~340 Refs. 341-344 dem-onstrate the extension of this methodology to the problem of localization energies in diene-metalcomplexes.

19.4 ALKENE AND POLYENE METAL COMPLEXES

19.4.1 The Dewar-Chatt-Duncanson Bonding Model

Dewar's345 and Chatt and Duncanson's346 analysis of the bonding in metal alkene complexesowed much to the synergic bonding model proposed for carbonyls by Pauling and described insome detail in Section 19.2 above, but it differed in the important sense that it used the perturbationmolecular orbital formalism rather than the valence bond method to describe the interactionsbetween the frontier orbitals of the alkene and the metal.345 Consequently this mode of analysisled in a transparent fashion to the important conclusion that the alkene must be bonded in a TT-fashion (55) to the metal and accounted for the IR characteristics of the coordinated alkenemolecule in Zeise's salt and related molecules.346 The forward donation component of the bond(see Figure 34) was attributed to the overlap of the filled alkene x-bonding molecular orbital witha totally symmetric orbital on the metal and the back donation component to overlap of filledorbitals on the metal, which are antisymmetric with respect to the mirror plane perpendicularto the metal alkene plane, and the empty alkene TT* orbital. Clearly both these interactions aremaximised for the 7r-bonding (dihapto) geometry illustrated in (55), and result in electron pop-ulation of the alkene 7r*-orbital and depopulation of the alkene 7r-orbital. These changes in electrondistribution are reflected in a lengthening of the C—C bond in the coordinated alkene and alowering of the C—C stretching frequency.347351

V

M

(55)


Figure 34 Schematic illustration of the foward and back donation components of the Dewar-Chatt-Duncansonbonding model for metal-alkene complexes

The relative importance of the two bonding components in the Dewar-Chatt-Duncansonbonding model can be qualitatively evaluated on the basis of the formal oxidation state of the metal,the electron donating characteristics of the ligands coordinated to the metal and the substituentson the alkene. Consequently this synergic bonding model has a flexibility which has proved in-valuable for rationalizing the physical and chemical properties of a wide range of alkene com-plexes.350 Furthermore, the perturbation molecular orbital formalism developed by Dewar hasbeen extended to the more complex bonding situations presented by polyene and polyenyl metalcomplexes. These developments will be described in more detail below.

The reliability of the Dewar-Chatt-Duncanson bonding model has been confirmed by a recentvery accurate SCF-Xa-SW calculation on Zeise's salt (56).352 An interesting feature of thisanalysis is the degree to which the ethylene molecule is retained as a subunit within the complex,thereby confirming the appropriateness of using a perturbation method of analysis. The forwarddonation component of the Dewar-Chatt-Duncanson bonding model can be identified in thecontour electron density plots shown in Figure 35 which illustrate the donation of electron densityfrom the alkene x-bonding molecular orbital to the platinum J-orbitals with the correct symmetrytransformation properties. Figure 36 illustrates an electron density plot for the back donationcomponent of the Dewar-Chatt-Duncanson model. It has been estimated from these calculationsthat 75% of the total binding energy in Zeise's salt originates from the forward donation componentand 25% from the back donation component, in broad agreement with the spectroscopic data whichhave been accumulated for this compound. In addition, accurate neutron diffraction studies onZeise's salt have demonstrated that the C—C bond length is only increased by approximately0.02 A on coordination (see Table 9).353>354

Cl VCl—Pt-

Cl

(56)

Table 9 Structural Parameters for some Symmetrical Alkenes andtheir Complexes

Compound

C2H4C2(CN)4C2F4K[PtCl3(C2H4)]Ni(C2H4)(PPh3)2Pt(C2H4)(PPh3)2

Pt(C2(CN)4)(PPh3)2Pt(C2Cl4)(PPh3)2Fe(C2H4)(CO)4Fe{(HO2C)HC}2(CO)4Rh(77-C5Me5)(C2H4)(PPh3)Ta(i?-C5Me5)(CHCMe3)-

(ij-C2H4)(PMe3)

C—C (A)

1.337(2)1.34(2)1.31(2)1.354(2)1.43(1)1.43(1)1.49(5)1.62(3)1.46(6)1.40(4)1.408(16)1.477(4)

M—C (A)

2.139(10)1.99(1)2.11(1)2.11(3)2.04(3)

2.03(3)2.093(10)2.228(3)2.285(3)


(a)

1

(b)

1

(c)

Figure 35 Electron density contour plots of the c-like bonding orbitals in Zeise's salt derived from Xa calculations.The contour values increase in absolute magnitude with increasing absolute values of the contour labels. The signof the labels gives the sign of the orbital lobes. The x-orbital of ethylene mixes with both the metal dzi anddx2-y2 orbitals to a significant extent (reproduced with permission from N. Rosch, R. P. Messmer and K. H.Johnson, J. Am. Chem. Soc, 1974, 96, 3855)

In lower valent transition metal complexes the degree of back donation from metal to alkeneis more extensive and is reflected in larger changes in the geometrical parameters associated withthe coordinated alkene.355"360 In particular the alkene C—C bond length is increased by as muchas 0.2 A in platinum(O) complexes, and Table 9 summarizes some typical bond lengths in thistype of complex. A similar bond lengthening has been noted for alkene complexes of the earliertransition metals which have a higher formal oxidation state but have ancillary ligands with good(T-donating characteristics, e.g. Ta(r;-C5H5)(C2H4)(CHCMe3)(PMe3) in Table 9, which hasa C—C bond length of 1.477(4) A.359

The effect of substituents on the length of the C—C bond in coordinated alkenes remainssomewhat ambiguous. A structural determination of Rh(77-C5H5)(C2H4)(C2F4) has demonstrateda significant bond length difference for the C—C bonds, i.e. 1.358(9) A for C2H4 and 1.405(7)


Figure 36 Contour plot of the x-like bonding orbital of Zeise's anion derived from Xa calculations (reproducedwith permission from N. Rosch, R. P. Messmer and K. H. Johnson, J. Am. Chem. Soc, 1974, 96, 3855)

A for C2F4.358 But this difference is less apparent when the structures of related ethylene andsubstituted ethylene complexes of platinum(O) are compared (see Table 9).355

The alkene substituents bend back away from the metal on coordination to give a quasi-tetrahedral coordination geometry to the alkenic carbon atoms. It has been noted in a recent reviewthat the bending back is least for hydrogen and is greatest for halogen substituents. This trendparallels that for the metal-alkenic carbon bond lengths where it is found that the shortestmetal-carbon bond lengths are associated with the alkenes bearing halogen substituents.355

For those complexes bearing electronegative substituents on the alkene and containing metalsof low oxidation states, the gross geometrical changes associated with coordination representinteractions which go beyond those which can be reliably treated using perturbation theory anda more appropriate starting point for the discussion of the bonding in these complexes is the me-tallocyclopropane structure illustrated in (57).361'362

Support for the view that the Dewar-Chatt-Duncanson perturbation theory approach is in-appropriate in such complexes is provided by equilibrium studies made by Tolman.363 Accordingto the Dewar-Chatt-Duncanson model the stabilization energy dE resulting from metal-alkenebonding may be estimated from the perturbation theory expression364 dE = A/Air + B/Air*where A and B involve geometric and overlap terms and ATT and ATT* represent the HOMO-LUMO energy separations implicit in the synergic bonding model, and which can be estimatedfrom ionization potential and electronic spectral data.365'366 Generally the energy of the alkene7r*-level is more sensitive to substituent effects than the 7r-level (see Table 10 for example), andtherefore bE is generally dominated by the term B/ATT* in the above equation. Tolman has plottedthe equilibrium constant for equation (5) against the energy of the LUMO of the alkene, and showna reasonable correlation between these parameters (see Figure 37). He noted that only conjugatedalkenes, e.g. styrene and /raws-stilbene, and fluorinated alkenes Ci^n^A-n showed major de-partures from the linear correlation expected from this simple analysis.

50

c ^

(57)

+ 10

uo

+ 5

CM

SI

o0

- 1 - 3 - 4 - 5(eV)

Inherent in the localized bonding representation illustrated in (57) are the two localized M—Ccr-bonds shown in (58). When these bonds are combined as symmetry-adapted linear combinations(as shown in (58)), then the resultant molecular orbitals can be identified with the forward andback donation components of the Dewar-Chatt-Duncanson bonding model. Therefore, the lo-calised bonding description represents a limiting form of the synergic bonding model where theinteraction between metal and ligand orbitals is particularly strong.351'367"369 In such situations,perturbation theory gives a qualitatively satisfactory model. Quantitatively it is deficient sinceit does not take into account in a satisfactory fashion the changes in alkene orbital energies whichresult from the gross geometric changes which accompany coordination. As noted above thesechanges are particularly striking for alkenes bearing halogen substituents coordinated to lowoxidation state metals. Sakaki et 0/.370373 have concluded on the basis of CNDO calculationson the complexes Pt(PH3)2(XY) (XY = C2H2, C2H4, CS2 and CO2) that the bending back ofsubstituents raises the energy of the ligand 7r-orbital and lowers the energy of TT* and thereforesignificantly strengthens the metal-ligand synergic interactions. Fluoroalkenes do not haveparticularly low lying 7r*-orbitals (see Table 10), but the energies of these levels are very sensitiveto the geometry of the alkene and are stabilised substantially by the distortions which accompanycoordination to a metal.372"377 This could perhaps account for the anomalous structural andthermodynamic data associated with complexes of these ligands. There have been suggestions,

M

(58)

Figure 37 Correlation of log K (25 °C) with the energy of the lowest unoccupied orbital of the free alkene. TheC2H4_WFW alkenes are designated by crosses and C2H4_wPh« by squares (reproduced with permission from C.A. Tolman, J. Am. Chem. Soc, 1974, 96, 2780)

Bonding of' Umaturated Organic Molecules to Transition Metals 51

is

NiL3 + alkene ^=^ Ni(alkene)L2 + L (5)


however, that in complexes of fluoroalkenes electron donation from the metal to ligand a* levelsmay give rise to an important bonding contribution.378

a Adiabatic if available. b Separation of energy levels fromUV spectra.1. C. A. Tolman, J. Am. Chem. Soc, 1974, 96, 2780.

Norman has reported the results of SCF-Xa-SW molecular orbital calculations on the com-plexes Pt(O2)(PH3)2 and Pt(C2H4)(PH3)2

379 which have confirmed that the metal to ligandcharge transfer is much more extensive in these complexes than in Zeise's salt, and shown thathis results are consistent with the X-ray photoelectron spectral studies which have been reportedfor these and related complexes.380"383 Ab initio calculations on Ni(C2H4)X2 (X = F~, H2O,PH3, etc.), where the C—C bond length was optimised in terms of the calculated total energy,have demonstrated that the calculated equilibrium bond length for all complexes is longer thanthat for free ethylene. The calculated C—C distance in the nickel(O) complexes was found to belonger than that in the corresponding nickel(II) complexes in broad agreement with the structuralresults summarised in Table 9. These calculations also provide a striking example of the validityof Pauling's electroneutrality principle, since the calculated charge on the nickel atom in thenickel(O) complex Ni(C2H4)(NH3)2 of +0.58 is only 0.25 smaller than that in the correspondingnickel(II) complex Ni(C2H4)(NH2)2, supporting the view that the alkene is behaving as a netLewis base in the higher oxidation state complex and a net Lewis acid in the lower valent com-plex.384

19.4.2 Asymmetrically Substituted Alkene Complexes

In unsymmetrically substituted alkene complexes the metal-carbon bond lengths often differsignificantly.385"394 The theoretical rationalisation of the asymmetric nature of the bonding insuch complexes can be readily appeciated in terms of the Chatt-Dewar-Duncanson perturbationtheory model described above. The relevant structural data for square planar alkene complexesof platinum(II) presented in Table 11 suggest that the metal-carbon bond length to the alkeniccarbon atom carrying the substituent (C-2 in (59)) is invariably longer than that to the unsub-stituted carbon atom (C-l in (59)), an effect which appears to be independent of whether thesubstituent is a net 7r-donor (D) or vr-acceptor (A). Frequently, this asymmetry in bond lengthsis accompanied by a slip distortion of the alkene moiety relative to the metal atom in such a way

Ionizationpotential* TT —> TT* b

Alkene (eV) (eV) -Ew* (eV)

Fumaronitrile 11.15 5.64 5.51Acrylonitrile 10.91 6.43 4.48Methyl vinyl ketone 10.10 5.95 4.15Styrene 8.47 4.33 4.141-Hexene 9.46 7.00 2.46Vinyl acetate 9.19 6.1 3.1/Aww-Stilbene 7.95 3.71 4.24CH2=CHO(CH2)3Me 8.93 6.43 2.50trans-2-Hzxene 9.13 6.84 2.29m-2-Hexene 9.13 6.98 2.152-Methyl-1-pentene 9.12 6.58 2.54MeCH=CMe2 8.67 6.98 1.69C2H4 10.52 7.28 3.24C2F4 10.52 8.88 1.64CH 2 =CHF 10.58 7.44 3.14CH2=CHC1 10.00 6.4 3.6Propylene 9.73 7.10 2.631-Butene 9.58 7.10 2.48CHF=CF 2 10.53 7.61 2.92trans-CHF=CHF 10.38 7.28 3.10

Table 10 Ionization Potentials and ic -+ 7r* TransitionEnergies for Alkenes1


that the centre of the C—C bond lies below the plane defined by the metal atom and the otherligands coordinated to the metal (see (60)). X-ray results on zero oxidation state metal-alkenecomplexes of the type Pt(alkene)(PRs)2 and Fe(alkene)(CO)4, where the alkene is acting as anet Lewis acid, surprisingly show a similar trend in metal-carbon bond lengths (see Table11).

C-l

M:

C-2

(59) (60)

Table 11 Pt—C Bond Lengths in Alkene Complexes

a In the alkene the first carbon is written as C(l), the second as C(2). b R =CH2CH2CH2NH3

f. The two entries refer to two crystalline modifications, one or-ange, the other yellow. c In the complexed ligand the first atom (C) as written isnumbered 1, the second (C, N or O) is labelled 2.

Hoffmann et al.367 have argued that these asymmetric bonding effects reflect the perturbationsintroduced into the olefin TT and TT* levels by the introduction of 7r-donor and acceptor ligands.In general 7r-acceptors lower the energies of TT and TT* compared to ethylene itself and second orderperturbation effects polarize the ethylene TT and TT* levels in the manner illustrated in (61). Incontrast Tr-donor substituents raise the energies of TT and TT*, and polarize these orbitals in theopposite sense (see (62)). In a square planar platinum(II) complex the effect of lowering TT and7T* by the introduction of a 7r-accepting substituent is to enhance the back donation componentillustrated in (63). Since the 7r* orbital in such an alkene has a larger coefficient at the unsub-stituted carbon atom C-1, the metal-ligand overlap illustrated in (63) will result in a strengtheningof the metal—C-l bond relative to the metal—C-2 bond. A 7r-donor substituent raises the energies

PtLi{alkene) ComplexesPt—C(l) Pt—C{2)

Alkene* (A) (A)

CH2=C(OMe)2 2.086(28) 2.798(30)CH2=CHOH 2.098(10) 2.222(9)CH2=CHOEt 2.128(7) 2.208(7)CH2=CH(OR) 2.12(3) 2.20(3)R—CH=CH(OR) 2.13,2.04(2) 2.32,2.33(2)CH2=CHC6H4NMe2-/> 2.137(17) 2.262( 16)CH2=CHPh 2.188(8) 2.219(9)CH2=CHPh 2.180(12) 2.236(10)CH2=CHEt 2.163(25) 2.173(23)CH2=CHBui 2.17(5) 2.26(5)CH2=CHRb .2.11(1), 2.17(3) 2.14(2), 2.19(3)

ML2(alkene) and ML^alkene) ComplexesM—C{1) M—C{2)

Alkene0 M (A) (M—X) (A)

CH2=CH—CN Ni 2.016(10) 1.911(10)C12C=C(CN)2 Pt 2.00(2) 2.10(2)CH2=N(Me)2 Ni 1.884(5) 1.920(4)(CF3)2C=N—N=C(CF3)2 Pt 2.02(1) 2.112(9)(CF3)2C=O Ni 1.89(2) 1.87(1)CH2=CHC(O)OMe Fe 2.092(2) 2.106(2)CH2=CH—CN Fe 2.10(1) 2.09(1)

c(O)RC H 2 = C V Fe 2.098(5) 2.127(4)

R'

/CH(CO2Me)C H 2 = C 1 Fe 2.092(7) 2.024(5)

vCH(CO2Me)

(61) (62)

(63) (64)

In zero oxidation state metal-alkene complexes, the alkene behaves as a net Lewis acid andthe metal-7r* interaction is particularly influential. The effect of a x-acceptor substituent willmirror that described above and result in a shortening of the metal-carbon bond to the unsub-stituted carbon atom. Replacement of one of the alkenic carbon atoms by a more electronegativeatom, e.g. N or O, has the similar effect of lowering the energies of TT and TT*, and since the TT*molecular orbital is more concentrated on the carbon atom, a shortening of the metal-carbonbond length relative to the metal-nitrogen (or -oxygen) bond length is anticipated and is indeedobserved (see Table 11 for examples).

19.4.3 Conformational Preferences of Metal Alkene Complexes

X-ray crystallographic studies have established that alkenes coordinated to transition metalsdisplay definite conformational preferences in the solid state. For example, in square planar ds

complexes the alkene carbon-carbon bond axis lies perpendicular to the square plane definedby the metal and the ligands (see (56) for example) unless the geometric constraints imposed bythe ligand force the alkene to adopt the alternative conformation (65). In trigonal d]0 complexes(66) and trigonal bipyramidal d% complexes (67), the alkenic C—C bond axes are found to liewithin a few degrees of the idealised trigonal coordination planes.355 Octahedral d6 metal-alkenecomplexes generally take up the conformation illustrated in (68). In bent sandwich compoundsand half sandwich compounds structural studies have also established that the preferred groundstate conformations are those illustrated in (69) and (70).271'395'396 Similar conformationalpreferences have been noted for the following ligands which can behave as effective Lewis acids

M //

(65) (66)

M-//

(67) (68) (69) (70)


of TV and TT* and consequently enhances the forward bonding component of the synergic bondingmodel. The relevant metal-alkene interaction illustrated in (64) will also lead to a strengtheningof the metal—C-l bond relative to metal—C-2 because of the larger coefficient at C-l in the alkene7r-bonding molecular orbital.


when coodinated to transition metals: dioxygen, acetylene, allene, ketones, imines, azo compoundsand sulphur dioxide.

Although the X-ray results summarised above give a clear indication of the more stable groundstate conformation, they give no indication of how stable the observed conformation is relativeto alternative conformations. NMR studies have, however, provided the necessary informationregarding the barriers to rotation for the alkene complexes illustrated in (66)-(70), and the resultsare summarised in Table 12.397~401 A more extensive discussion of these NMR studies is givenin Chapter 20.

Table 12 Summary of Experimentally DeterminedRotational Barriers in Metal-Alkene Complexes

MLrt (alkene)

(71) (72) (73) (74)

The observed rotational barriers in metal-alkene complexes MLn (alkene) have been reliablyreproduced using extended Hiickel molecular orbital calculations.367 These calculations haveshown that the wide range of rotational barriers observed for such complexes reflects a subtleinterplay of electronic and steric factors. The electronic component is easy to appreciate in termsof the Dewar-Chatt-Duncanson bonding model introduced above.

The forward donation component of the metal-alkene bond is not going to be sensitive to theconformation of the alkene with respect to the metal-ligand coordination plane since it involveselectron donation from the alkene 7r-molecular orbital to an empty orbital on the metal whichis totally symmetric (see Figure 34). In contrast the alkene TT* molecular orbital has a nodal planecoincident with the metal-alkene bond axis and will seek to maximise its overlap with one of thetwo metal d-orbitals (dxz or dyz if z is defined as the metal-alkene bond axis) with matchingsymmetry characteristics. Therefore the electronic component of the rotational barrier inmetal-alkene complexes will be very sensitive to the relative energies of the metal dxz and dyz

orbitals in the fragment MLn. With this in mind, it is possible to identify the following generalsituations where the metal electronic configuration and geometric disposition of ligands can in-fluence the relative energies of these two metal orbitals.367'402"404

(7) Coordination of the alkene to an axial ly symmetric MLn fragment. In an axially symmetricmetal carbonyl or phosphine fragment, e.g. M(CO)3 of C?,v symmetry, and M(CO)4 or M(CO)5of CAV symmetry, the dxz and dyz orbitals are degenerate. Therefore for any rotation angle of thealkene relative to the metal-alkene bond axis (see (71) for a definition of the rotation angle) itis possible to take a linear combination of the metal dxz and dyz orbitals which gives an equivalentmetal-alkene TT* overlap integral. Consequently in such complexes, it is impossible for any con-formational preference to be set by the metal to alkene back donation component. Only stericor supplementary bonding interactions, not defined by the Dewar-Chatt-Duncanson bondingmodel, can contribute to the observed rotational barriers in such complexes. Experimentally itis found that the rotational barriers in such complexes are rather small (<40 kJ mol"1), inagreement with the above theoretical analysis.

Barrier forMetal electronic rotation

Complex type configuration (kJ mol~')

ML2(alkene) d10 76-105ML4(alkene) af8 42-63ML3(alkene) d™ 42-105MLs(alkene) d6 29-42

L

L—Pt

1/

y•>- Z

X

(75)

L-Pt—/

\7

y•>• z

X

(76) (77)

Figure 38 illustrates the molecular orbitals derived from extended Hiickel calculations for theds PtClJ fragment. The metal dxz orbital lies at slightly higher energies than dyz, but is less lo-calised on the metal since there is a significant contribution from the chlorine 3/?-orbitals to thismolecular orbital (see (77)). Consequently although the energy difference term in the secondorder perturbation theory expression favours the in-plane conformation (76), the overlap term

Cl

ClCl

Pf

2a,

Cl

Cl—Pt

71

71

Figure 38 An interaction diagram for Zeise's salt K[PtCl3(C2H4)] derived from extended Hiickel calculations.The diagram refers to the observed perpendicular alkene conformation (reproduced with permission froiji T. A.Albright, R. Hoffmann, J. C. Thibeault and D. L. Thorn, J. Am. Chem. Soc, 1979, 101, 3801)


Structural studies have established that for ML5(alkene) complexes the observed conformationin the solid state is that shown in (72) rather than the sterically preferred structure illustratedin (73). Albright et al.361 have suggested that this preference arises from the four-electron des-tabilising interaction between the alkene 7r-orbital and the filled metal d^-orbital (see (74)) whichhas its maximum value for conformation (73). They calculate a rotational barrier of 40 kJ mol"1

for the complex M(CO)s(C2H4) in reasonable agreement with the experimentally determinedvalue.

(2) Coordination of the alkene to an non-axially symmetric fragment, MLn, where the metaldxz and dyz orbitals have similar energies and compositions. For a T-shaped ML3 fragment, whichpossesses Cio symmetry, the metal dxz and dyz orbitals are no longer degenerate and thereforeit is no longer possible to define equivalent linear combinations of these orbitals for any particularrotation angle. In such complexes there will only be a large electronic component to the rotationalbarrier if the metal dxz and dyz orbitals differ significantly in either their energies or orbitalcompositions. The alternative symmetrical conformations of an ML3(alkene) complex are illus-trated in (75) and (76). In (75) the alkene 7r*-orbital interacts exclusively with dyz and in (76)exclusively with dxz.

Bonding of Umaturated Organic Molecules to Transition Metals 57

favours the perpendicular conformation (75). The overlap term predominates because the energydifference between dxz and dyz is only 0.4 eV and leads to a slight preference for conformation(75). The electronic factors favouring this conformation are supplemented by steric effects whicharise due to the close proximity of the alkenic carbons and the cis chlorines in the in-plane con-formation (76) .405 It has been estimated that approximately 70% of the observed conformationalbarrier arises from these steric effects and only 10% from the electronic effects described above.367

Albright et al.367 have suggested various strategies for stabilizing the alternative in-planeconformation (76) in square planar alkene complexes based on the use of mono- and di-enes whichare sterically constrained, e.g. (78) and (79). An X-ray crystallographic study on (79) which hasbeen recently reported has confirmed the presence of in-plane and out-of-plane metal alkenemoieties.406

(78) (79)

(3) Coordination of the alkene to a non-axially symmetric fragment, where the metal dxzand dyz orbitals have very different energies and compositions. In d10 metal-alkene complexesof the type Pt(alkene)(PPh3)2 the planar ground state conformation (66) is observed althoughon steric grounds alone the alternative conformation (80) would have been predicted. The electronicbasis of this conformational preference arises because the metal dxz and dvz orbitals in the angularML2 fragment have very different energies and orbital compositions. The frontier molecular or-bitals of the angular ML2 fragment have been illustrated in Figure 18 and discussed in some detailin Section 19.2.6. In the context of the alkene conformational problem the important point is thatthe dxz orbital (81) (b2 in Figure 18), which lies in the ML2 plane, is considerably higher lyingthan dyz (82) (bi in Figure 18) and therefore is able to function as a more effective donor towardsthe alkene TT* molecular orbital. Furthermore, thep-orbital mixing in (81) which is absent in (82)leads to a more favourable overlap integral with the alkene TT* orbital. Therefore in this instanceboth the energy and the overlap terms reinforce each other to lead to a clear electronic preferencefor the planar (66) rather than the perpendicular (80) conformation. The electronic reasons forthe very different energies and orbital compositions of (81) and (82) have been discussed in somedetail elsewhere and the reader is referred to refs. 407 and 367 for a detailed discussion of therelevant second order perturbation theory arguments. Experimental estimates of the rotationalbarriers in such complexes lie in the range 75-105 kJ mol"1. A model calculation on Ni-(CO)2(C2H4), with the carbonyls separated by an angle of 100°, has resulted in a computed barrierof 88 kJ mol"1 for rigid rotation of the ethylene with respect to the Ni(CO)2 unit.367 It has beenshown that the computed rotational barrier is very sensitive to the angle between the carbonylligands and the degree of non-planarity of the ethylene ligand. Furthermore, the barrier is increasedif the carbonyls are replaced by ligands with inferior 7r-acceptor qualities.367

Hi.

(80) (81) (82)

In Section 19.2.6 it was emphasised that the frontier orbitals of an angular d10 ML2 unit anda Civ Fe(CO)4 fragment bear striking similarities and both fragments are isolobal with a carbene.Therefore it is not surprising that Fe(CO)4(alkene) and related af8 complexes show similar con-formational preferences to those noted above for Pt(alkene)L2. The preferred conformation ofthe d% complexes (67) has the alkene lying in the idealised trigonal plane of the trigonal bipyramid.In an orbital sense this preference can be readily interpreted in terms of the frontier orbitals of


Fe(CO)4 illustrated in Figure 11. In particular the higher energy and d-p hybrid nature of themetal dxz orbital (83) compared to dyz (84) will lead to stronger metal-alkene back donationin the observed conformation (67).

""Hinnff

(83) (84)

For complexes of the type Fe(CO)4(alkene), the rotation of the alkene is accompanied by apseudorotation of the metal-ligand coordination sphere, since it is known that in general for ds

five-coordinate complexes the activation energy for such a process is very low. The combined alkenerotation and pseudorotation process is illustrated in (85). The activation energy for this processis substantially smaller than that for the rigid alkene rotation process since in the transition state(85b) the Fe(CO)4 fragment has C$v symmetry and the dxz and dyz orbitals become degenerate.Therefore the conformational preference which is so pronounced in the initial geometry is lostin the transition state.367

e. a

e a

(a)

e. a

(b)

(85)

(c)

The arguments described above for d10 Pt(alkene)L2 complexes may be readily extended tohomoleptic metal-alkene complexes of the type Pt(alkene)3, where it can be demonstrated thatthe in-plane geometry (86) is preferred to the perpendicular conformation (87). This predictionwas made by Rdsch and Hoffmann408 on the basis of extended Hiickel calculations and sub-stantiated by an X-ray crystallographic analysis of (88) by Stone and his coworkers.409 Theelectronic arguments may also be extended to alkene derivatives of half sandwich (69)395 andbent sandwich compounds (70)271 where the different energies of the metal dxz and dyz orbitalslead to substantial barriers for alkene rotation.

Jvi

(86)

M / \

(87) (88)

In octahedral and square planar complexes containing two ethylenes in trans positions, theD^d conformation shown in (89) is more stable than the alternative eclipsed Z>2/, conformationshown in (90). In the former, the alkenes interact separately with an orthogonal set of metal dxz

and dyz orbitals whereas in the latter they share an interaction with a common metal d-orbital.It can be readily shown from perturbation theory arguments that the former interaction is largeras long as the metal d-orbitals and the alkene 7r*-orbitals lie close in energy.410

II—M—/ II—M—1|

(89) (90)


19.4.4 Conformational Preferences of Metal Alkyne Complexes411"415

The bonding in metal-alkyne complexes can in general be interpreted in terms of the Dewar-Chatt-Duncanson bonding model since the alkyne molecule has a pair of TT and TT* molecularorbitals which lie in the MCC plane (TT|| and TT||* in Figure 39) which are entirely analogous tothose described above for ethylene complexes.351 In addition the alkyne molecule has TT and TT*molecular orbitals perpendicular to the MCC plane (wj_ and TT±* in Figure 39) which also havethe correct symmetry properties to interact with the metal d-orbitals. The interaction involving7T _L * is not energetically significant since it involves a 5 type of overlap, with a small total computedoverlap integral. However, the overlap integral involving the filled wj_ molecular orbital is com-parable in magnitude to that involving TT||*, and cannot be ignored. The additional 7r-donor in-teraction available to acetylene confers upon it unusual properties since it can function simulta-neously as a 7r-acceptor and 7r-donor ligand. In terms of the simple electron counting rules it placesthe acetylene ligand in an ambiguous position because it can behave either as a two- (utilisingonly TT||) or a four- (utilising both ir\\ and ir±) electron ligand. In general acetylene behaves asa two-electron ligand. However, there are some complexes particularly of the earlier transitionmetals where it is apparent that acetylene is behaving as a four-electron ligand. Some examplesof complexes which only conform to the 18-electron rule if acetylene is defined as a four-electronligand are illustrated in (91)-(94).416~426

O

7T

7T_L7T *

_L

OIc o

Mo(CO)(HC=CH)(S2CNR2)2

(91)

MoO(HC=CH)(S2CNR2)2

(92)

Pt—III

(93) (94)

Figure 39 An illustration of the potential interactions between the TT and TT* molecular orbitals of acetylene andmetal orbitals with appropriate symmetry characteristics


W(CO)(RCCR)3 (95) is formally a 20-electron compound if the acetylenes are consideredin a similar manner. King427 has shown that in this highly symmetric compound there is onecombination of the acetylene w molecular orbitals (see (96)) which does not find a symmetry matchwith a metal (^-orbital and therefore remains non-bonding. Therefore, this compound does contain18 electrons involved in metal-ligand bonding molecular orbitals.

O

C

W

(95) (96)

The ability of acetylenes to function both as x-acid and ?r-donor ligands is nicely illustratedby the complementary conformations taken up by the acetylene ligand in the carbonyl and oxocomplexes illustrated in (91) and (92). In each case the acetylene takes up that conformation whichsupplements the strong interaction between the metal and either the carbonyl or the oxo ligand.The relevant orbital interactions illustrating this point are shown in (97) and (98). In (97) theacetylene TT||* molecular orbital is aligned in such a way that it supplements the metal-carbonylback donation interaction, whereas in (98) the metal-oxo interaction is supplemented by x-do-nation from the acetylene molecular orbital.426

(97) (98)

19.4.5 Metal Poly ene and Polyenyl Complexes

The 7T molecular orbitals of some non-cyclic polyenes and polyenyls are illustrated in Figure40, and their energies and nodal characteristics are compared to those of the ethylene molecule.367

As the chain length is increased then the singly noded molecular orbital 17ra, which for ethylenecontributes to the back donation component of the Dewar-Chatt-Duncanson bonding model,is stabilized considerably. In the allyl radical this orbital is essentially non-bonding and singlyoccupied, and therefore only in the allyl cation will this orbital be able to function as a good acceptororbital in the same sense that has been described in detail above for ethylene. For butadiene andthe higher polyenes, 17ra is a filled bonding molecular orbital and is able to function only as a donororbital. It will only give rise to a stabilizing interaction if there is a metal d-orbital which is empty,and has matching symmetry characteristics. Allyl and butadiene possess an additional molecularorbital with a single nodal plane, 2TTS, which can interact effectively with a metal d^-orbital. Thepentadienyl radical has a non-bonding and singly occupied molecular orbital with the same nodalcharacteristics. For the higher polyenes the 2TTS level is bonding and doubly occupied and thereforecan only act as a donor function.367'428"430


a

2n.

2n.In

In,

2ntIn

InIn

In.

1n,

lnt

In

Figure 40 Relative energies and nodal characteristics of the 7r-orbitals of acyclic polyenes

\\w\\x

(99) (100)

One important consequence of the simultaneous 7r-donor and 7r-acceptor interactions in thecase of butadiene complexes is the equalization of the C—C bond lengths for the coordinatedbutadiene ligand.431 In the free butadiene molecule the central C—C bond is approximately 0.09A longer than the outer C—C bonds, reflecting the nodal characteristics of the HOMO in thefree ligand (l7ra in Figure 40). Coordination of butadiene to a metal fragment such as Fe(CO)3results in electron loss from this molecular orbital and electron donation into the 2TTS molecularorbital, which has a nodal plane passing through the outer C—C bonds. This redistribution ofelectron density therefore results in a shortening of the central C—C bond at the expense of the

From this analysis it is clear that allyl , butadiene and pentadienyl+ fragments in a formalsense resemble the acetylene molecule described in the previous section, since they can functionsimultaneously as 7r-donor and -acceptor ligands, and will bond most effectively to metal-con-taining fragments which have an empty dxz orbital and a filled dyz orbital. The Fe(CO)3 fragmentsatisfies these requirements in an ideal fashion. The frontier orbitals of this fragment were de-scribed in some detail in Section 19.2.6 and illustrated in Figure 13. This fragment has a highlying orbital of aj symmetry (99) and a lower lying degenerate set of e orbitals (100), which inthe case of Fe(CO)3 contain an electron pair. Figure 41 illustrates the interactions between thefrontier molecular orbitals of this fragment and the 7r-molecular orbitals of the butadiene moleculeand underlines the effective match of donor-acceptor interactions involving the 17ra and 2TTS ligandmolecular orbitals. Similar interaction diagrams are relevant to the bonding situations in[Fe(77-C5H7)(CO)3]

+and [Fe(i?-C3H5)(CO)3].- 102>429>430


2a

lesleala,

la'

Figure 41 Interaction diagram for the complex Fe(CO)3(r/-C4H6) with the conformation illustrated in (106)

outer C—C bonds.432 A recent extensive structural survey of butadiene metal tricarbonyl com-plexes by Cotton et al.431 has confirmed that on average the central C—C bond is 0.02 A shorterthan the outer C—C bonds.

The ability of the Fe(CO)3 fragment to interact in a particularly strong fashion with a pairof cis conjugated double bonds is illustrated in the complex formed between p-divinylbenzeneand Fe(CO)3 (101).433 Each of the vinyl groups and the adjacent two ring carbon atoms are bondedto the iron atoms in a tetrahapto fashion, resulting in a large disruption in the aromaticity of thecentral aromatic ring and the fixation of one of its double bonds.

Fe(CO)

Fe(CO)3

(101)

In the solid state all simple allyl, butadiene, pentadienyl and hexatriene complexes of the typeM(?7-C/7Hrt+2)(CO)3 adopt the conformation illustrated in (102) with a carbonyl ligand pointingtowards the open edge of the polyene.434"440 However, solution NMR studies have indicated thatthe alternative conformation (103) is thermally accessible and rotational barriers in the range22-65 kJ mol"1 have been reported for such complexes.441444 In section 19.2.6 it was noted thatthe M(CO)3 fragment has a set of out-pointing hybrid orbitals which approximate to those ofthe parent octahedron. Therefore the conformational preference noted above can be qualitativelyrelated in a valence bond localized orbital sense to the attainment of maximum overlap betweenthese octahedral hybrid orbitals and localized orbitals of the polyene as illustrated in (104) and(105) for example.


M(104) (105)

(102) (103)

A more detailed and extensive account of the conformational preferences and rotational barriersin these complexes may be obtained from a molecular orbital analysis.145'428430 The orbital in-teraction diagram for the butadiene complex which is illustrated in Figure 41 is representative.For the conformation shown in (106) the major interactions arise from the overlap of 2es (dxz)with 2TTS (butadiene) and 2ea (dyz) with l7Ta (butadiene). Although this molecule possesses onlyCs. symmetry, pseudosymmetry considerations dictate that the lower lying and filled metal orbital,lea, interacts primarily with 27ra. However, the latter interaction is only of minor significancesince 27ra lies quite high in energy and therefore does not interact strongly with the metal orbital.Most of the rotational barrier associated with the conformational change (106) to (107) resultsfrom energy changes involving the HOMO 4a' in Figure 41. In going from (106) to (107) theoverlap between 2es and 2TTS decreases, whilst the interaction between les and 2TTS increases,stabilizing 2a' and further destabilizing 4a' as shown in (108). For the related pentadienyl complex,Albright et al.361 have computed that the energy rise in the 4a' level is due approximately 30%to loss of overlap between 2es and 2TTS and 70% from increased repulsions of les.

(106) (107)

les

2TTS

(108) (109)

For 16-electron polyene complexes the alternative conformation (103) is preferred, since theremoval of two electrons from the HOMO 4a' converts the four-electron destabilizing interactionbetween les and 2TTS shown in (108) into a two-electron stabilizing interaction. An example ofthe change of conformation from (102) and (103) resulting from the loss of two valence electronsis illustrated in (109) .445'446

This conformational change is also relevant to an understanding of the bonding in the complexesillustrated in Figure 42.4 4 7 4 4 9 Albright et a/.430 have rationalized the bond lengths and confor-mations of these complexes in terms of the cycloheptatriene-norcaradiene equilibrium illustratedin (110). On the cycloheptatriene side of the equilibrium, conformation (102) is preferred sinceit represents an 18-electron triene complex. However, at the norcaradiene side of the equilibrium,the bonding approximates to that of a 16-electron butadiene complex and the alternative con-formation (103) is preferred. This interpretation is also consistent with the C—C bond lengthsobserved in the complexes and shown in Figure 42.


1/2.44

1.65

Figure 42 Structural data for some cycloheptatriene-Cr(CO)3 complexes (reproduced with permission fromT. A. Albright, Trans. Am. Crystallogr. Assoc, 1980,16, 35)

(110)

In 18-electron metal polyene complexes the rotational barriers decrease as the chain lengthdecreases (see (111) and (112) for typical examples of experimentally determined rotationalbarriers). This effect can be related to a diminution of the important interaction between thepolyene 2TTS orbital and the metal 1 es and 2es orbitals as the energy of the 2TTS orbital increases(see Figure 40).367'429

Calc. 44Obs. 48-50

Fe+

46 kJ mol"1

55-57 kJ mol"1

(111)

Calc. 30Obs. 40

(112)

4

kJmol"1

M(r7-polyene)L2 and M(?7-polyene)L4 complexes also show distinct conformational preferencesin the solid state which depend on the total electron count. In 18-electron ML2 complexes thepreferred conformation is that shown in (113), which has one of the ligands L pointing towardsthe open edge of the polyene. The complementary conformation with Cs symmetry (114) is pre-ferred for 16-electron complexes.450454 Similar conformational preferences (see (115) and (116))have been noted for M(7;-polyene)L4 complexes. In a localized orbital sense the conformational


preferences of M(r;-polyene)L2 complexes can be interpreted in terms of the attainment of atetrahedral configuration about the metal (e.g. 117) in the 18-electron complexes, and a squareplanar configuration (118) in 16-electron low-spin complexes. In 18-electron M(r?-polyene)L4complexes the observed ground state conformations can be related in a localized orbital senseto the attainment of octahedral coordination about the metal centre, whereas the bicapped tet-rahedral geometry is preferred in the 16-electron complex. These alternative geometries arerepresented in a localized orbital sense in (119) and (120).

M

(113)

(117)

M

L

(114)

L—M—L

i

L(115)

L

L

(116)

"in,, M

(118) (119) (120)

The conformational preferences and rotational barriers in these complexes can also be inter-preted in terms of the interactions between the frontier orbitals of the isolated ML2 (or ML4)fragment, which were discussed in some detail in Section 19.2.6 and the frontier orbitals of thepolyene, which are illustrated in Figure 40. For a 16 electron vr-allyl complex of the type [Pt(^-C3H5)(PPh3)2]+, l7ra is the LUMO of the allyl cation and the HOMO of the dw ML2 fragmentis the b2 orbital shown in (81). Therefore, the bonding situation is entirely analogous to that de-scribed earlier for Pt(C2H4)(PH3)2. The very strong bonding interaction between l7ra and b2 inthe perpendicular conformation (114) is lost upon rotation to the in-plane structure illustratedin (113), and replaced by a weaker interaction between the lower lying metal bi orbital (82) andl7ra. Therefore there is a strong conformational preference for the perpendicular conformation(114) in 16-electron allyl complexes. Extended Hiickel calculations suggest a rotational barrierof more than 150 kJ mol"1 for the rigid rotation process in [Ni(T/-C3H5)(CO)2]+.367 However,such a calculation is somewhat academic since allyl complexes invariably rearrange through analternative r]3-r)1 process. In an 18-electron M(77-03^)L2 complex the additional electron pairwill cause a reversal in the conformational preference. In such a complex both 1 ?ra and b2 are filled,therefore b2 must interact with the higher lying and empty 2TTS orbital. This is only possible forthe in-plane geometry (113). Similar arguments can be used to rationalize the conformationalpreferences of 18- and 16-electron butadiene ML2 complexes.367

The situation in 16- and 18-electron [M(77-C5H7)(PH3)2]+ complexes is somewhat morecomplicated and therefore the interaction diagrams for both conformations are shown in Figure43.455 For reasons of clarity only those orbitals which are important in setting the conformationalpreferences are shown in this figure. For the in-plane conformation, b2 (81) interacts stronglywith 2TTS (pentadienyl) and bi (82) less strongly with l7ra (pentadienyl). For the perpendicularconformation the interactions between the alternative combinations of metal and ligand orbitalsare less strong, because the energies of the orbitals are less favourably matched. The occupanciesof the orbitals shown in Figure 43 are appropriate for an 18-electron complex. For this electroncount, the b2 orbital is involved in a two-electron two-orbital stabilizing interaction, whereas bjis involved in a four-electron two-orbital repulsive interaction. The dominant interaction is theformer involving the b2 orbital and therefore the in-plane conformation (113) is preferred. Fora 16-electron complex two electrons are removed from the HOMO (2a" for the in-plane confor-mation and 2a' for the perpendicular conformation in Figure 43). Figure 43 suggests that since2a' is more antibonding than 2a" then the perpendicular conformation (114) will be favouredin the 16-electron complex.


l a"la"

*Figure 43 Fragment molecular orbital analysis of the bonding in [Fe(?7-C5H7)(PH3)2]~ in its two alternativesymmetric conformations. For this 18-electron complex the conformation illustrated on the left hand side is themore stable, but the alternative conformation is preferred in related 16-electron complexes

The above-analysis is complicated by the fact that in the 18-electron pentadienyl complexesthe 'slip' distortion shown in (121) which results in a trihapto allylic geometry can reduce thefour-electron destabilizing interaction involving bj (see (122)) without diminishing the bondinginteraction involving b2 (see (123)). Extended Huckel molecular orbital calculations by Nurseand Mingos455 have demonstrated that for the complexes [Pt(r;-C5H7)(PH3)2]+, the allylicstructure is the preferred ground state structure. However, since the in-plane pentahapto geometrylies only 50 kJ mol"1 higher in energy a low energy fluxional process is possible. These resultsare consistent with structural and NMR results reported by Maitlis and his coworkers456 on aseries of isoelectronic palladium hexadienyl complexes. They have identified the existence of thefluxional process shown in (124) which has as its transition state the in-plane rj5 conformation.

(121) (122) (123)

L

L'V

(124)

66


The preference shown by the d]0 ML2 fragment to interact with odd-alternant hydrocarbon cationsin a trihapto fashion which utilises the b2 orbital is demonstrated by the wide range of structuresshown in (125)-(128).457~461 All the molecules shown in (125)-(127) undergo 1,3-sigmatropicrearrangements similar to that shown in (124) (see Chapter 20 for a fuller discussion of theseprocesses).

(125) (126) (127)

EtQ

Pt(PPh3)2+

(128)

19.4.6 Reactions of Coordinated Alkenes and Polyenes

A theoretical model which would account for the types of reaction which coordinated alkenestake part in, and which successfully rationalizes the relative rates of such reactions as a functionof alkene substituent, metal oxidation state and electronic characteristics of the other ligandscoordinated to the metal, would clearly have a major impact on the development of organometallicchemistry. Unfortunately to date it has not proved possible to develop such a general frameworkand the few significant theoretical contributions which have been made have generally resultedfrom a detailed examination of specific and well defined reactions. The scale of the theoreticalproblem can perhaps be gauged from the structural and conformational problems discussed above.The metal coordination sphere in an organometallic complex is far less rigid than that surroundinga functional group in an organic molecule, and as the previous section indicated it is possible fora metal-alkene or -polyene complex to be undergoing fluxional processes which either involverotation of the polyene about the metal, a pseudorotation process about the metal centre or acombination of both. Furthermore, the reactions of 18-electron complexes are also strongly de-pendent on the generation of new coordination sites at the metal centre by ligand loss. The acti-vation energies of such processes will depend markedly not only on the electronic characteristicsof the ligand undergoing dissociation but also on its steric requirements. The importance of suchprocesses has been underlined by Tolman's 16-18-electron analysis of reaction pathways in or-ganometallic reactions,462 and his attempt to define the steric requirements of ligands such astertiary phosphines in terms of their cone angles.463

Following on from the outstanding success of the orbital symmetry rules in organic chemistrythere were attempts to adapt these concepts to organometallic chemistry.464"470 In particularthese early theoretical studies indicated that the ability of metal ions, e.g. Ag+, and square planards complexes to catalyze pericyclic organic reactions which had been designated as orbitallyforbidden by the Woodward-Hoffmann rules471 could be understood in terms of simple orbitalcorrelation diagrams. Subsequent studies showed that such a simple view of the role of the metalatom in such reactions was too facile since the metal-catalyzed reactions did not appear to go bya simple concerted reaction but by a series of reactions involving oxidative addition processes atthe metal centre472 and, in the case of catalysis by silver ions, metal-stabilized carbonium ions.473

Subsequently, theoretical studies on organometallic systems have tended to be more conservativein their aims and have concentrated on developing an understanding of specific stoichiometricreactions. The following sections summarize some of the more important and recent developmentsin this field.


19.4.6A Nucleophilic addition reactions of coordinated alkenes

Early studies, particularly by Chatt and his coworkers,474 on square planar complexes ofplatinum(II) and palladium(II) showed that the coordination of mono- and dienes to these metalsled to an enhancement of their rates of reaction with nucleophiles. Subsequently it was shownthat nucleophilic addition to organometallic cations containing coordinated polyenes was a generalorganometallic reaction.475 These reactions have proved to be particularly useful in organicsynthesis since the organic moiety can subsequently be released from the metal centre by oxidationwith iron(III) or cerium(IV) salts.475

In the most general terms these nucleophilic addition reactions (see equations (6) and (7) forsome typical examples476) may be understood in terms of the Dewar-Chatt-Duncanson bondingmodel. For a cationic complex the forward donation component involving the alkene x-orbital

MeO.

McO

(129)

(7)

OMe

(130)

is likely to outweigh the back donation component and lead to the development of a partial positivecharge on the alkenic carbon atoms and thereby increase their susceptibility towards nucleophilicattack. The enhanced reactivities of cationic polyene complexes towards nucleophiles may beinterpreted in a similar fashion.

This simple view does not represent the only possible pathway for nucleophilic attack at an alkenecoordinated to a metal centre. A detailed theoretical investigation has been made477 of the reactionsbetween hydride ion and coordinated ethylene in the complexes [Fe(CO)5(C2H4)]2+, [Fe(?y-C5H5)(C2H4)(CO)2]+, [PtCl3(C2H4)]- and Ni(C2H4)(PH3)2, and concluded that in all theseexamples the degree of back donation to the alkene is sufficient to deter direct nucleophilic attackon the alkene in its ground state symmetrical r/2 geometry. They attribute this deactivation ofthe alkene to a rise in energy (compared to the free alkene) of the alkene TT* acceptor orbital oncoordination. The carbon /vorbital coefficients in the alkene TT* acceptor orbital are also reducedas a result of the bonding interaction with the metal ^-orbitals. Therefore in terms of a frontierorbital model for reactivity both the energy term and overlap terms are less favourable fornucleophilic attack at the coordinated alkene compared to the free alkene.478 These argumentsdo, however, neglect the stabilizing effect resulting from the development of positive charges onthe alkenic carbon atoms in a cationic complex. In a self-consistent field approach this would havethe effect of lowering the energy of the acceptor ?r*-orbital and lead to a more favourable inter-action with the hydride donor orbital.

Eisenstein and Hoffmann477 have argued that an important component of the activation energyfor nucleophilic attack at the coordinated alkene arises from a slipping motion of the alkene relativeto the metal-alkene bond axis prior to bond formation between the nucleophile and the alkeniccarbon atom. This proposed slip distortion is entirely analogous to that discussed in Section 19.4.2for unsymmetrically substituted alkenes coordinated to metals and illustrated in (60). In orbitalterms this slipping motion has the effect of lowering the energy of the x* acceptor orbital andincreasing the coefficient of the/?-orbital at the carbon atom which lies farthest from the metalatom, as shown in (131). In a localized orbital sense the slip distortion has the effect of generatingthe metal-stabilized carbonium ion illustrated in (132). This model can account for the regiose-lectivity of some nucleophilic addition reactions of unsymmetrically substituted alkenes. In theabsence of overriding steric effects nucleophilic attack generally occurs at the alkenic carbon atom

(6)


bearing the substituent. If the substituent is a 7r-donor then it is capable of stabilizing the slippedtransition state. This model has also provided a rationalization for the observation that a phosphineprefers to replace a carbonyl in [Mn(CO)5(C2H4)]+, whilst it adds to the double bond in [Fe(r?-C5H5)(C2H4)(CO)2]+. Replacement of three carbonyl ligands by a cyclopentadienyl ring in theabove complexes has the effect of lowering the energy of the acceptor orbital localized predomi-nantly on the alkene, thereby promoting nucleophilic attack at the alkene. This is an interestingconclusion since this difference in reactivity would not have been predicted on the basis ofmetal-induced polarization effects. A discussion of nucleophilic addition reactions based on abinitio molecular orbital calculation is to be found in Ref. 479.

+CM

C

(131) (132) (133)

When more than one type of alkene or polyene is coordinated to a metal in an organometalliccation then a different sort of regioselectivity problem arises. For example, the compound shownin (133) has five possible sites where nucleophilic attack could occur. Yet in practice such reactionshave been shown to proceed with a high degree of selectivity and, for example, in (133) nucleophilicattack occurs exclusively at the position marked 1. Davies, Green and Mingos476'480'481 haveproposed a set of rules for predicting the regioselectivities of these reactions. These rules werederived from a simple perturbation theory model which relates the charge on specific carbon atomsto the hapticity of the coordinated ligand, i.e. the number of carbon atoms of the polyene orpolyenyl coordinated to the metal, the parity of the hapto number, i.e. whether an even or oddnumber of carbon atoms are coordinated to the metal, and finally the degree of cyclic conjugationin the coordinated polyene. These rules in order of decreasing importance are:

1. Nucleophilic attack occurs preferentially at even coordinated polyenes which have no un-paired electrons in their highest occupied molecular orbitals.

2. Nucleophilic addition to open* coordinated polyene is preferred to closed polyenes.3. For even-open polyenes nucleophilic attack at the terminal carbon atom is always preferred;

for odd-open polyenes attack at the terminal carbon atom occurs only if ML^ is a strong electronwithdrawing group.

The quantum mechanical basis of these rules and their applications have been extensivelydiscussed elsewhere and therefore will not be discussed in detail here.476'480 Some illustrativeexamples of the rules are shown in (134)-(137). More detailed calculations, particularly by Clack

Fe—C—O

(135) (136)

<zxH

(137)

* In this context ligands which are cyclically conjugated, e.g. benzene and cyclooctatetraene, are described asclosed, and non-cyclically conjugated polyenes, e.g. butadiene and pentadienyl, are described as open.


and his coworkers, dealing with the regioselectivities of nucleophilic addition reactions are to befound in refs. 482 to 486.

The conformation of the metal-polyene complex can also have a profound influence on theregioselectivity of the nucleophilic addition reaction. Schilling, Hoffmann and Faller487 havemade a detailed analysis of the ligand asymmetry on the regioselectivity of nucleophilic additionto [Mo(77-C5H5)(^C3H5)(NO)(CO)]+. Extended Hiickel molecular orbital calculations forthe alternative conformations of this ion shown in (138) and (139) have demonstrated that thecalculated net charges on the terminal carbon atoms of the allyl group not only differ within theallyl group but undergo a reversal on changing from exo to endo. Simple arguments based onthe supposition that nucleophilic attack will occur preferentially at the carbon atom bearing themore positive charge suggest that nucleophilic attack will occur at C(l) for both isomers. Con-siderations based on the orbital coefficients of the allylic carbon p-orbitals in the LUMO of bothisomers lead to the same conclusion. These theoretical results are in complete agreement withthe experimental findings.

-0.16

-0.07

exo endo

(138) (139)

Albright and Carpenter have shown that in Cr(CO)3(arene)488 complexes the regioselectivityof nucleophilic and electrophilic attack is controlled not only by the substituent on the arene ringbut also by the conformation of the Cr(CO)3 unit. They have demonstrated that the arene carbonswhich are eclipsed with respect to the carbonyls are preferentially attacked by nucleophiles.Electrophilic substitution is preferred at the staggered arene carbon atoms. For donor substitutedCr(CO)3(arene) complexes, the effect of the conformation and the electronic effects of the donorgroup operate in concert and nucleophilic substitution is directed meta to the substituent.488 Thesame relationship holds for acceptor-substituted Cr(CO)3(arene) complexes and substitutionis directed para to the substituent. When the steric requirement of the substituent is large, e.g.Bul, a para rather than meta attack predominates because the conformation of theCr(CO)3(arene) complex is set by steric rather than electronic effects.

19.4.6.2 Pericyclic reactions of coordinated alkenes and polyenes

There have been numerous attempts to use qualitative symmetry-based arguments as a basisfor rationalizing metal-assisted concerted pericyclic organic reactions.464470 Unfortunately,experimental tests of the predictions arising from these theoretical analyses have been rare. Inpart this situation has arisen because of the difficulties associated with finding a system whichis experimentally well defined and simultaneously amenable to theoretical analysis. The dramaticcatalysis of the quadricyclene-norbornadiene conversion by various rhodium(I) complexes489

appeared to be an example par excellence of the way in which a symmetry forbidden reactioncould become allowed by utilising the af-orbitals of the transition metal, until it was shown toproceed in a stepwise non-pericyclic fashion.490 Later studies on the metal-catalysed ring openingof bicyclop .1.0] butane derivatives also showed that the two C—C bonds of the substrate werebreaking in a non-concerted manner.491

One of the few metal-promoted transformations which still appears to be a true pericyclic processis the ring opening of a set of strained cyclobutene iron(tricarbonyl) complexes reported by Pettitand coworkers,492 and illustrated in equation (8). Their suggestion of a reversible loss of carbonmonoxide from the complex followed by an irreversible ring opening is consistent with kineticstudies. The second step appears to be a genuine pericyclic reaction. Unfortunately, it is not possible

/

4 *" n " Fe(CO)

(8)

Fe(CO)4 Fe(CO)3


to determine the influence of an Fe(CO)3 unit on the stereochemistry of a cyclobutene ring openingfrom this study, since the prohibitively large strain energy associated with the incorporation ofa trans double bond in a six-membered ring precludes the isolation of any products which mighthave come from conrotatory ring opening or from the alternative disrotatory mode which bendsthe breaking carbon-carbon c-bonds away from the metal. Pinhas and Carpenter493'494 haveargued that the observed disrotatory process is symmetry allowed because it permits a favourablemixing between the TT* orbital of the C—C bond undergoing ring opening and the metal-cyclo-butene back donation component (see (140)).

(140) (141)

They also have studied the related metal-assisted ring opening of 2-phenylmethylenecyclo-propene illustrated in equation (9) and demonstrated that the reaction proceeds by a disrotatorymode which bends the breaking carbon-carbon cr-bond away from the metal. Although both thepossible disrotatory ring opening pathways are symmetry forbidden, Pinhas and Carpenter493'494

have argued that the observed process is less forbidden since it minimises the four-electron de-stabilizing interaction between the carbon-carbon c-bond and metal-cyclopropene back donationcomponent as shown in (141). Albright461 has arrived at similar conclusions for methylenecy-clopropene-ML2 complexes. However, more theoretical work is required to understand why thealternative conrotatory ring process, which is not symmetry forbidden, does not take place. Ex-tended Hiickel calculations have indicated that the observed disrotatory and the conrotatory modesare almost equienergetic for the iron(tricarbonyl)methylenecyclopropenecomplex.493494

19.4.6.3 Alkene insertion and hydrogemtion reactions

The alkene insertion reaction shown in equation (10) is clearly a key step in many catalyticprocesses, alkene hydrogenation and hydroformylation being the most important, and consequentlylas received some attention from theoreticians.495'496 The scale of the theoretical problem whena real system is studied can perhaps be gauged by Thorn and Hoffmann's495 analysis of the reactionbetween ethylene and PtHCl(PH3)2 which required a detailed analysis of all the intermediatesillustrated in Figure 44. In addition it proved necessary to consider the polytopal rearrangementsavailable to these intermediates.

Several workers have noted that an important prerequisite for the alkene insertion reactionis the coplanar transition state illustrated in (143).497'498 This geometry mitigates the four-electrondestabilizing interaction between the donor orbital of the hydrido ligand and the alkene 7r-orbitalby allowing some mixing in of the alkene TT* orbital which also lies in this plane. A similar sym-metry restriction pertains to the reverse reaction — ^-elimination from a coordinated ethyl li-gand.498 The planar transition state is energetically unfavourable for a five-coordinate hy-dridoalkene ds platinum complex since the strong 7r-donating hydrido ligand shows a preference

R R R

/^^= ^ 2^==~=r —• { (9)Fe(CO)4 Fe(CO)3 /p

Fe(CO)3

— M —H + >< —* —M—CH2—CH3 (10)


Figure 44 The intermediates considered in Thorn and Hoffmann's analysis of the alkene insertion reactionPtHCl(PR3)2 + C2H4 -* Pt(Et)Cl(PR3)2 (reproduced with permission from D. L. Thorn and R. Hoffmann, J.Am. Chem. Soc, 1978,100, 2079)

for the axial site of a trigonal bipyramid and the alkene has a strong conformational preferencefor the in-plane equatorial geometry (67). The electronic reasons for the latter have been discussedin some detail in Section 19.4.3. When these site and conformational preferences are combinedas in (144), then it can be seen that the resultant geometry is unfavourable for the insertion step.For four-coordinate hydridoethylene complexes the trans isomer is thermodynamically morestable than the corresponding cisy but the calculated activation energy for the insertion reactionfrom the trans isomer is much greater than that from the cis isomer,495 largely because the formerrequires a pseudotetrahedral transition state, which is energetically unfavourable for a ds pla-tinum(II) complex. Thorn and Hoffmann495 consider that it is most likely that the cis transitionstate is achieved by dissociation processes involving three-coordinate platinum(II) intermediates.The computed activation energy for this process is smaller than for that involving five-coordinateintermediates. These conclusions are in broad general agreement with the experimental obser-vations.495

H "///

M

(143)

H

(144)

Dedieu499 has reported a similar investigation using ab initio molecular orbital techniques ofthe alkene insertion reaction in octahedral d6 rhodium complexes as models for Wilkinson'scatalyst. He found that the early stages of the reaction are best described as an ethylene insertionprocess, and that some polytopal rearrangements occur simultaneously with relaxation from thetransition state. The whole insertion process was computed to be exothermic with a rather moderateactivation energy. He also concluded that a good vr-donor ligand in the coordination sphere pro-motes the alkene insertion rather than the hydride migration mechanism and lowers the computedactivation energy.

19.4.6.4 Ziegler-Natta catalysis and metallocyclobutanes

The symmetry arguments discussed above are of course also relevant to the related problemof alkene insertion into metal-alkyl bonds, which is a key step in the Cossee-Arlman mechanismfor Ziegler-Natta polymerization of alkenes.500 Consequently this and related processes have


Me

ClCl

T'i—^-C

Cl C

n

Cl/,,,, MQTi

C l ^ l^ l Cl

nMer=C

Cl

ci ci

n

Cl—Ti2

•c

Cl

Figure 45 Schematic illustration of the intermediates proposed by McKinney for Ziegler-Natta catalysis (re-produced with permission from R. J. McKinney, J. Chem. Soc, Chem. Commun., 1980, 490)

Me

(145)

(146)

been studied by several workers using different levels of approximation in their molecular orbitalcalculations.501"506 The problem associated with such calculations is that a variety of assumptionshas to be made concerning the nature of the key active intermediate in these heterogenous titaniumhalide catalysts. McKinney507 has reported molecular orbital calculations on an alternativemechanism for Ziegler-Natta catalysis. This mechanism which is illustrated in Figure 45 involvesthe coupling of two cis coordinated alkenes to give a metallacyclopentane intermediate, followedby reductive elimination.

Coordination of two ethylene molecules to the titanium halide moiety can give rise to the twoalternative conformations illustrated in Figure 45. The preferred orientation of the ethylene ligandsis very dependent on the formal oxidation state of the titanium. For titanium(I V) the perpendicularconformation is favoured, but in the corresponding titanium(II) complex the occupation of a metal^/-orbital which can simultaneously back bond to both alkenes leads to a preference for the planarconformation (145). This molecular orbital, which is illustrated in (146), has an in-phase overlapbetween the adjacent carbon atoms of the two alkenes. This favourable interaction gives rise toa low activation energy for the coupling reaction leading to the metallacyclopentane complexshown in Figure 45. The corresponding reaction for the titanium(IV) complex is not favourablesince it represents, formally at least, a titanium(IV) to titanium(VI) oxidation state change. Thetitanium(IV) methylmetallacyclopentane complex illustrated in Figure 45 can either undergoa cyclobutane elimination reaction, or a methyl migration reaction to form a titanium(II)-«-pentylcomplex. McKinney507 has argued that the latter becomes a favourable process as a result of smalldistortions in the metallacycle which lead to an in-phase overlap between the donor orbital of themethyl group and a donor orbital of one of the a-carbon atoms of the metallacycle. The proposedmechanism can account for the stereoselective polymerization of propylene since reductive couplingof one side of the metallacycle with the polymer chain would place a chiral centre either a or ftto the titanium centre. Green et at.508 have proposed an alternative mechanism for polymerizationwhich involves carbenes generated by a-elimination processes. However, this mechanism hasnot been the subject of such a detailed theoretical analysis.


Metallacycles, as exemplified by the metallacyclopentane complex described above, are ofteneasily formed and decomposed. They have been proposed as key intermediates in numerousmetal-catalyzed reactions and are clearly central to an understanding of the role played by metalcomplexes in catalysis.509'510 Figure 46 summarises some of the relevant organic and organo-metallic reactions derived from this key intermediate. From a theoretical point of view the easiestdecomposition pathways to analyse for metallacyclopentanes are those leading to two moleculesof ethylene (the reverse of the reaction discussed above in the context of Ziegler-Natta catalysis),and that leading to the formation of cyclobutane.51 !'512 A recent study of nickel(II) metallacy-clopentanes Ni(C4H8)LAI (n = 1-3) has led to the following interesting symmetry distinctionbetween these alternative decomposition pathways: If a metallacycle degradation to form cy-clobutane is symmetry-allowed, then the formation of ethylene will usually be symmetry-for-bidden, and vice versa.512

Li or MgBrMgBr

/M 2 +

HorH M° + 2

w

Figure 47 A schematic illustration of the evolution of the orbitals of a (CH2)4 fragment. In the centre the tet-ramethylene dianion is shown and its transformation either into cyclobutane (left hand side) or two moleculesof ethylene (right hand side), with the concomitant loss of two electrons. (Reproduced with permission of R. J.McKinney, D. L. Thorn, R. Hoffmann and A. Stockis)

O + M° *Figure 46 The importance of metallacyclopentanes in organometallic transformations

The electronic reasons for this symmetry distinction can be appreciated by reference to Figure47 which illustrates the evolution of the orbitals of a tetramethylene dianion when it is transformed


either into cyclobutane (left hand side) or two molecules of ethylene (right hand side). Thesetransformations require the simultaneous loss of two electrons to an empty orbital on the metal.In a formal sense this corresponds to a formal decrease of the metal oxidation state by two. Al-though both transformations are formally reductive elimination processes, the formation of cy-clobutane requires that the electron pair be transferred to the metal centre via an orbital of b2symmetry. In contrast the formation of two molecules of ethylene takes place with the electronpair being transferred to the metal centre via an aj orbital. In d% ML2 complexes there is onlyone orbital which is vacant on the metal centre and depending on the orientation of the ML2 unitrelative to the tetramethylene dianion plane it can have ai or b2 symmetry, but not both. Thecomplementary nature of the resultant orbital correlations is illustrated in Figures 48 (a) and 48 (b).The reductive elimination of cyclobutane from a square planar four-coordinate complex is sym-metry allowed (Figure 48a), but the corresponding elimination from the four-coordinate transcomplex shown in Figure 48(b) is symmetry forbidden. For the latter, reductive elimination ofethylene is an allowed process and that of cyclobutane forbidden.512

L•

LNi

L

Ni

L

a

- H - (a)

L

Ni

L

L

NiVL

a

b

a2

a(b)

Figure 48 The symmetry allowed transformation of c/s-Ni(C4H8)L2 —*• NiL2 + C4H8 is shown in (a) and thesymmetry forbidden transformation trans-N[(C4H$) -*- NiL2 + C4H8 is shown in (b). (Reproduced with thepermission of R. J. McKinney, D. L. Thorn, R. Hoffmann and A. Stockis)

Since the direct conversion of the square planar and trans-Ni(C4H$)L2 structures is a forbiddenreaction,5'3 cyclobutane formation is predicted to be the exclusive product of reductive eliminationfrom Ni(C4Hg)L2 complexes. It has been proposed that the trans-Ni(C4Hs)L2 complex is animportant intermediate in the degradation of five-coordinate nickelacyclopentane complexesNi(C4Hg)L3, since it should decompose readily by the symmetry allowed process to give ethylene.These theoretical conclusions are in broad general agreement with the elegant experimental studiesperformed on theNi(C4H8)Lw system by Grubbs and hiscoworkers.513514

A similar symmetry based analysis has been reported for the Fe(CO)3(C2H4)2 system. Thispaper also analyses the regioselectivities of reactions involving asymmetrically substituted al-kenes.511


19.5 OXIDATIVE ADDITION AND REDUCTIVE ELIMINATION REACTIONS

Oxidative addition of single bonded substrates X Y to transition metal complexes as exemplifiedby equation (11) represents one of the most common reactions available to low valent transitionmetals.515 The reaction occurs with a very wide range of substrates and there are many welldocumented examples for dw,ds and d6 complexes. The general classification of such reactionsas oxidative addition reactions masks the complexity of the alternative mechanistic pathwaysfor achieving the final product which has both X and Y coordinated to the metal. From a theo-retical point of view the most attractive transition state is the three-centre symmetrical transitionstate illustrated in (147). However, such a transition state is likely to be important only when XYis a homonuclear diatomic molecule such as H2.516 For heteronuclear diatomics and polar sub-strates in general, alternative linear transition states are important. To complicate matters furtherthere is some evidence that some oxidative addition reactions proceed by radical pathways.517

The symmetry restrictions even for the oxidative addition reaction proceeding through the sym-metrical transition state (147) are minimal, and this process is symmetry allowed for many ofthe alternative polytopal geometries of d10 and ds complexes. The reasons for this can perhapsbe best appreciated by considering the symmetry constraints for the reverse reaction — reductiveelimination. For a pair of cis methyl or hydrido ligands in a transition metal complex with Civsymmetry, the aj and b2 metal ligand combinations illustrated in (148) correlate with the aj (a)and hi (a*) molecular orbitals of the isolated ethane or dihydrogen molecules. For the reductiveelimination to be a symmetry allowed process the additional electron pair occupying b2 (c*) hasto be transferred to the metal centre via an orbital of b2 symmetry.510'518~520 Therefore, the the-oretical nature of the problem is entirely analogous to that discussed in Section 19.4.6.4 for theelimination of cyclobutane from a metallacyclopentane complex. The angular ML2 ds fragmentand the ML4 d6 fragment of C20 symmetry both have as their lowest unoccupied molecular orbitala metal-localized orbital of b2 symmetry (see Figures 11 and 18 for illustrations of the relevantorbitals) capable of accepting the additional electron pair. Therefore for square planar ds andoctahedral d6 complexes containing cis alkyl or hydrido ligands, the reductive elimination processis symmetry allowed. By the principle of microscopic reversibility the reverse reaction — oxidativeaddition involving dxo ML2 or ds ML4 complexes — is also symmetry allowed. However, somereorganisation energies may be involved in attaining these geometries from the more symmetricalground state geometries. Reference 521 gives a detailed analysis of all these factors for the oxidativeaddition of hydrogen to square planar rhodium(I) complexes.

X1

ivr :

(147)

Tatsumi et al. 52° have made a detailed analysis of reductive elimination of ethane and hydrogenfrom organotransition metal complexes of the type MR2(PRa)2, and evaluated the role of thecentral metal atom and the peripheral ligands on the activation energy for the reductive eliminationprocess. The following conclusions emerged from this analysis. (1) In the four-coordinate complexthe activation energy is reduced if the leaving groups are good c-donors. (2) Stronger donor ligandstrans to the leaving groups increase the activation energy for elimination. (3) The activation energyis substantially lower for nickel than for platinum four-coordinate complexes, reflecting the relativeenergies of the antisymmetric M—R bonding orbitals in the nickel and platinum complexes. (4)T-shaped trans-MLR2i arising from dissociation of L from ML2R2, encounters a substantialbarrier to polytopal rearrangement to C15-MLR2, which in turn has an open channel for reductiveelimination of R.2- (5) If the leaving groups are poor c-donors, cis-trans isomerization in thethree-coordinate manifold is calculated to be more favourable than reductive elimination.

ML* + XY —*- MX(Y)Ln (11)

ai b 2

(148)


19.6 ALKYL, CARBENE AND CARBYNE COMPLEXES

19.6.1 Alkyl Complexes

An early view that the inability to isolate stable transition metal alkyl complexes resulted froman inherent weakness of the transition metal-carbon cr-bond was supported by theoretical cal-culations.522 Fortunately this view did not prevail and the experimental work of Chatt523 andWilkinson524 and their coworkers in the 1950s demonstrated that thermally robust metal alkylcomplexes could be isolated if ligands such as CO, 77-C5H5 and PR3 were also coordinated to thetransition metal. Such ligands are important in the isolation of robust alkyl complexes becausethey raise the activation energies for the potential decomposition pathways open to the metal alkylmoiety. Two of the more important decomposition pathways, ^-elimination and reductive elim-ination, have been discussed from a theoretical point of view in Sections 19.4.6.3 and 19.5. The^-elimination process from a metal ethyl complex requires an available empty coordination sitecis to the ethyl group capable of accepting the migrating /3-hydrogen. Since carbonyl and 77-C5H5generally form stable 18-electron complexes which are inert to ligand dissociation, they effectivelyblock the /^-elimination process. In square planar platinum(II) complexes, which have 16-electronconfigurations, there is a vacant coordination site, but the relatively high energy of the acceptororbital on the metal leads to a reasonably large activation energy for /3-elimination. For complexesof the type PtL2R2 the presence of strong cr-donating ligands such as phosphines raises the acti-vation energy for reductive elimination of alkane, as noted in Section 19.5. Therefore, stablecomplexes of the type Pt(PR3)2R2 can be readily isolated.523

There are other decomposition pathways open to metal alkyl complexes, e.g. a-eliminationand bimolecular elimination processes, but these have not been the subject of such detailed the-oretical analyses.

During the last ten years a wide range of homoleptic metal alkyl complexes of the transitionmetals has been synthesized. The contributions of Wilkinson524 and Lappert525 and their coworkersto this field have been particularly noteworthy. The isolation of these complexes has been achievedby using ligands which do not readily undergo /^-elimination processes, e.g. CH2SiMe3, Me, and1-norbornyl. Thermodynamic studies on these compounds have demonstrated that transitionmetal-alkyl bond strengths are in line with expectations based on electronegativity considerationsand are not inherently weak.526

In general the bonding in metal alkyl complexes can be accurately described in a localized orbitalsense involving the overlap between sp3 hybrid orbitals on the carbon atoms and an appropriatehybrid orbital on the metal. UV photoelectron spectral studies on homoleptic metal alkyl complexeshave indicated that in complexes such as WMe6 the major bonding component arises from overlapof the hybrid orbitals on carbon with the metal d- and s-orbitals.527530 There is also somestructural and photoelectron spectral evidence supporting the view that there is a weak hyper-conjugative interaction between the filled metal ^/-orbital and the alkyl o"*-orbitals in metal alkylcomplexes and particularly when the hydrogen atoms are replaced by fluorine.531

19.6.2 Metal Carbene (Alkylidene) Complexes

Singlet carbene has a filled orbital of (7-pseudosymmetry and an empty /?-orbital of 7r-pseu-dosymmetry (see Figure 12 for an illustration of the relevant orbitals) and therefore is capableof entering into synergic bonding interactions with transition metals in an analogous fashion tothat described above for alkenes. The first examples of carbene complexes of the transition metalswere reported by Fischer and his coworkers in 1964.532 These carbene complexes were synthesizedby sequential nucleophilic and electrophilic addition to metal carbonyl complexes and thereforethe substituents on the carbene were invariably 7r-donating groups. The range of such carbenecomplexes has been greatly extended to other metals, particularly by Lappert and his coworkers.533

Structural studies on these substituted carbene complexes have demonstrated that the MCR2unit is planar, and the metal-carbon distances show a considerable bond shortening comparedto related alkyl complexes which is consistent with a substantial degree of multiple bond characterbetween metal and the carbene carbon atom.534 Photoelectron spectral studies on such complexesare also consistent with this interpretation.535'536 Chemically these substituted carbene complexesare electrophilic in character and, for example, readily undergo nucleophilic substitution reactions.

In 1965 Guggenberger and Schrock537 reported the structural characterization of the firstexample of an unsubstituted carbene complex, viz. Ta(?7-C5H5)(CH2)(Me). Subsequently,Schrock and his coworkers538 have synthesized a wide range of 10-, 14-, 16- and 18-electron

78 Bonding of Unsaturated Organic Molecules to Transition Metalsm

tantalum and tungsten alkylidene (CHR) complexes which have a number of interesting structuraland chemical features. For example, the structure illustrated in (149) has alkyl, alkylidene andalkylidyne groups simultaneously coordinated to the metal and provides an internally calibratedcomparison of the relative lengths of these types of bonds. The relevant bond lengths are W—alkyl,2.258 A; W—alkylidene, 1.942 A; W—alkylidyne, 1.785 A. X-ray and neutron diffraction studiesby Williams, Churchill and Stucky539543 on Schrock's alkylidene complexes have demonstratedthat the bonding in such complexes deviates significantly from that which might have been pre-dicted on the basis of a simple synergic model involving forward donation from the carbene donororbital to the metal and back donation from the metal to the empty p-orbital on the carbene. Inthese complexes the alkylidene ligand appears to undergo the rotational distortion illustrated in(150), whereby the ligand pivots about the alkylidene carbon atom. This leads to a reduction inthe M—C—H bond angle (angles as small as 78° have been reported) and a corresponding in-crease in the M—C—R bond angle (bond angles up to 170°). Since the substituents on the carbenecarbon atom are generally rather bulky, e.g. C(Me)3, there is clearly a steric component to thesepivoting distortions. Molecular orbital calculations by Goddard et al.544 on model compoundshave indicated that there might also be an important electronic component to the observed dis-tortion mode. In particular the alkylidene deformation has been traced to an intramolecularelectrophilic interaction between the acceptor on the metal and the carbene lone pair orbital asillustrated in (151). It has been argued that the bulky substituents on the metal and carbene protectthe metal centre from intermolecular interactions and control the extent of carbene pivoting. Asecondary interaction between the metal and carbene weakens the C—H bond of the alkylideneand attracts the a-hydrogen to the metal. However, full transfer of hydride from the carbene tothe metal is a forbidden reaction for 14-electron five-coordinate complexes.

Bu

™ r v R

C CH2Bul H

' " (150)

P P = dmpe

(149)

(151)

Chemically the Schrock alkylidene complexes show significant differences when comparedwith the Fischer carbene complexes. For example, they are nucleophilic rather than electrophilicin character. Goddard et al.544 have attributed this difference to an extremely effective overlapbetween the tantalum d-orbitals and carbene /^-orbital in the back donation component of thealkylidene-metal bond. In the tantalum complexes the d-p interaction is so strong that the metald-carbon 2p antibonding orbital is high lying and effectively delocalized. Consequently it is lessavailable for nucleophilic attack than the corresponding orbital in the Fischer-type carbenecomplexes, where the metal-carbon overlaps are less effective. Goddard et al.544 have also analysedthe theoretical aspects of the ^-elimination reaction characteristic of these compounds whichleads to the formation of carbyne complexes.

The similarity in the bonding capabilities of carbenes and alkenes suggests that the argumentsdeveloped in Section 19.4.3 regarding the conformational preferences of metal alkene complexescan be readily adapted to gain an understanding of the conformational preferences of metal carbenecomplexes. For example, the observed conformation of Ta(r;-C5H5)2(CH2)(Me) illustrated in


(152) is consistent with the interaction between the carbene acceptor orbital and the filled non-bonding lai orbital of the Ta^-CsHs^I^ framework illustrated in Figure 3Q. In this compoundthe experimental barrier to carbene rotation is too high for it to be determined in an NMR ex-periment. The computed rotational barrier is in excess of 100 kJ mol"1.

(152)

The conformational preferences of ethylene and carbene ligands on the same metal take ona special significance when the Herrison-Chauvin mechanism for alkene metathesis is consid-ered.545 In particular only the collinear alkene and carbene conformation illustrated in (153) willlead to metallacyclobutane formation and be productive in metathesis. In a d6 complex the collinearconformation is unfavourable from a 7r-bonding point of view since the carbene and alkene areforced to share the same metal d-orbital for back donation, rather than maximize their interactionswith a pair of orthogonal d-orbitals. Therefore this conformation will only be favoured if thereis a significant direct carbene-alkene bonding interaction, i.e. if the carbene and alkene arecoordinated to a metal ion with a small radius. The collinear arrangement is preferred for d2

complexes which have only one filled ^/-orbital capable of interacting with the empty orbitals ofthe carbene and the alkene.546

Eisenstein and Hoffmann546 have argued that for d2 to d6 complexes neither a metal (car-bene) (alkene) (153) nor a metallacyclobutane (155) is the stable geometry but instead an inter-mediate non-classical structure (154) is preferred. They cite recent structural evidence on tita-nium(II)-substituted cyclobutene complexes547 to support this unusual proposal. They suggestthat the role of the metal in alkene metathesis is primarily to hold the two reactive componentstogether in a geometry favourable for their interaction.

Rappe and Goddard548 have made a detailed analysis of alkene metathesis, epoxidation andde-epoxidation reactions involving tungsten and chromium oxo complexes using the GeneralisedValence Bond Method. They have argued that the oxo ligands play an important role in alkenemetathesis catalysts because the transformation shown in equation (12) is associated with asubstantial increase in the metal-oxo bond energy. In valence bond terms the metal-bond orderincreases from two to three.

*'*X \AUM '" ^ UM' A ^ ^ LflM , / \

^^ *̂** ^̂̂ ^̂̂ .̂^^ ^^ >^s* ^̂̂̂̂

(153) (154) (155)

O s . +g CH2

^M=CH2 + j f —• ^ M ^ \ H 2 (12)^ CH2


19.6.3 Metal Carbyne (Alkylidyne) Complexes

Several examples of carbyne complexes have now been synthesized particularly by Fischer,Schrock and their coworkers.538'549 The bonding in these complexes has not been the subject ofdetailed theoretical studies although in principle the bonding must be closely related to that re-ported for the related oxo, nitrido and alkylimido complexes.550 Structural studies on carbynecomplexes have shown that in some cases the W—C—R group is not linear, e.g. trans-W(CO)4l(CCPh) has a W—C—Ph bond angle of 162° whereas in other compounds, e.g.trans-W(CO)4(CCMe), the C—C—Me angle is 180°. Extended Hiickel calculations on relatedalkylimido complexes, e.g. Re(NMe)Cl2(PR3)3 have demonstrated a very soft potential energysurface for the bending of the Re—N(Me) moiety suggesting that the M—N—C and M—C—Cbond angles may be affected by crystal packing forces.551

REFERENCES

1. F. A. Cotton and G. Wilkinson, 'Basic Inorganic Chemistry', J. Wiley and Sons, New York, 1976, p v.2. J. W. Richardson, in 'Organometallic Chemistry', ed. H. Zeiss, Reinhold, New York, 1960, p. 12.3. H. C. Longuet-Higgins and L. E. Orgel, J. Chem. Soc, 1956,1969.4. W. Hiibel and E. H. Braye, / . Inorg. Nucl. Chem., 1959,10, 250.5. R. Criegee and G. Schroder, Liebigs Ann. Chem., 1959, 623, 1.6. F. A. Cotton, J. Chem. Soc, 1960, 400.7. R. Hoffmann, / . Chem. Phys., 1963, 39,1397.8. R. Hoffmann, Science, 1981, 211, 995.9. L. Libit and R. Hoffmann, J. Am. Chem. Soc, 1974, 96,1370.

10. F. A. Cotton, 'Chemical Applications of Group Theory', Wiley-Interscience, New York, 2nd ed., 1971.11. I. H. Hillier and V. R. Saunders, Mol. Phys., 1971, 22,1025.12. E. J. Baerends and P. Ros, Mol. Phys., 1975, 30,1735.13. J. Demuynck and A. Veillard, Theor. Chim. Ada, 1973, 28, 241.14. K. H. Johnson, Adv. Quantum Chem., 1973, 7,143.15. K. H. Johnson, Annu. Rev. Phys. Chem., 1975, 26, 39.16. R. F. Fenske, Prog. Inorg. Chem., 1976, 21, 179.17. D. E. Ellis and G. S. Painter, Phys. Rev. (B), 1970, 2, 2887.18. S. P. Walch and W. A. Goddard, III, J. Am. Chem. Soc, 1976, 98, 7908.19. J. B. Johnson and W. G. Klemperer, J. Am. Chem. Soc, 1977,99, 7132.20. C. J. Ballhausan and H. B. Gray, 'Molecular Orbital Theory', W. A. Benjamin, New York, 1964, p. 120.21. F. A. Cotton and G. Wilkinson, 'Comprehensive Inorganic Chemistry', J. Wiley and Sons, New York, 4th

ed., 1980.22. R. Hoffmann, M. M.-L. Chen and D. L. Thorn, Inorg. Chem., 1977,16, 503.23. L. Pauling, 'Nature of the Chemical Bond', Cornell University Press, Ithaca, N.Y., 3rd ed., 1960.24. J. L. Hubbard and D. L. Lichtenberger, Inorg. Chem., 1980,19, 3865.25. D. J. Lichtenberger and R. F. Fenske, Inorg. Chem., 1976,15, 2015.26. J. L. Hubbard and D. L. Lichtenberger, Inorg. Chem., 1980,19, 3865.27. M. A. Andrews, Inorg. Chem., 1977,16, 496.28. K. H. Johnson and U. Wahlgren, Int. J. Quant. Chem., Symp., 1972, 6, 243.29. D. E. Sherwood, Jr., and M. B. Hall, Inorg. Chem., 1980,19,1805.30. B. E. Bursten, D. G. Freier and R. F. Fenske, Inorg. Chem., 1980,19,1810.31. B. Rees and A. Mitschler, J. Am. Chem. Soc, 1976, 98, 7918.32. D. W. Turner, C. Baker, A. D. Baker and C. R. Brundle, 'Molecular Photoelectron Spectroscopy', Wiley

Interscience, London, 1970.33. J. H. D. Eland, 'Photoelectron Spectroscopy', Butterworths, London, 1974.34. T. Koopmans, Physica, 1934, 1, 104.35. A. H. Cowley, Prog. Inorg Chem., 1979, 26, 46.36. J. C. Green, Struct. Bonding {Berlin), 1981, 43, 37.37. J. C. Slater and K. H. Johnson, Phys. Rev. (B), 1972, 5, 844.38. N. Rosch, W. G. Klemperer and K. H. Johnson, Chem. Phys. Lett., 1973, 23,149.39. E. J. Baerends, Ch. Oudshoorn, and A. Oskam, J. Electron. Spectrosc Relat. Phenom., 1975, 6, 259.40. B. E. Bursten and R. F. Fenske, / . Chem. Phys., 1977, 67, 3138.41. J. C. Green, D. I. King and J. H. Eland, Chem. Commun., 1970, 1121.42. D. R. Lloyd and E. W. Schlag, Inorg. Chem., 1969, 8, 2544.43. B. R. Higginson, D. R. Lloyd, P. Burroughs, D. M. Gibson and A. F. Orchard, J. Chem. Soc, Faraday Trans.

2, 1973,69,1659.44. E. J. Baerends, Ch. Oudshoorn and A. Oskam, J. Electron Spectrosc Relat. Phenom., 1975, 6, 259.45. S. Evans, J. C. Green, A. F. Orchard, T. Saito and D. W. Turner, Chem. Phys. Lett., 1969, 4, 361.46. I. H. Hillier, M. F. Guest, B. R. Higginson and D. R. Lloyd, Mol. Phys., 1974, 27, 215.47. I. H. Hillier, J. Chem. Phys., 1970, 52, 1948.48. H. B. Jansen and P. Ros, Theor. Chim. Ada, 1974, 34, 85.49. W. C. Nieuwpoort, Phillips Res. Rep. Suppl., 1965, 6, 1.50. A. F. Schreiner and T. L. Brown, J. Am. Chem. Soc, 1968, 90, 3366.51. A. Serafini, J.-M. Savariault, P. Cassoux and J.-F. La Barre, Theor. Chim. Ada, 1975, 36, 241.52. K. H. Johnson and F. C. Smith, Jr., Phys. Rev.(B), 1972, 5, 831.


53. B.-I. Kim, H. Adachi and S. Imoto, / . Electron Spectrosc. Relat. Phenom., 1977,11, 349.54. I. H. Hillier, M. F. Guest, B. R. Higginson and D. R. Lloyd, Mol. Phys., 1974, 27, 215.55. N. A. Beach and H. B. Gray, J. Am. Chem. Soc, 1968, 90, 5713.56. J. Demuynck, A. Strich and A. Veillard, Nouv. J. Chim., 1977,1, 217.57. A. Serafini, M. Pelisser, J.-M. Savariault, P. Cassoux and J.-F. La Barre, Theor. Chim. Acta, 1975, 39,

229.58. I.H. Hillier and V. R. Saunders, Chem. Commun., 1971, 642.59. I. H. Hillier and V. R. Saunders, Mol. Phys., 1971, 22, 1025.60. H. B. Jansen and P. Ros, Theor. Chim. Ada, 1974, 34, 85.61. J. F. Nixon, J. Chem. Soc, Dalton Trans., 1973, 2226.62. R. A. Head, J. F. Nixon, G. J. Sharp and R. J. Clark, J. Chem. Soc, Dalton Trans., 1975, 2054.63. R. A. Head, J. F. Nixon and R. J. Clark, / . Organomet. Chem., 1977,135, 209.64. R. A. Head, J. F. Nixon, N. P. C. Westwood and R. J. Clark, / . Organomet. Chem., 1978,145, 75.65. J. F. Nixon, R. J. Suffolk, M. J. Taylor, J. G. Norman, D. E. Hoskins and D. J. Gmur, Inorg. Chem., 1980,

19,810.66. J. F. Nixon, R. J. Suffolk, M. J. Taylor, J. C. Green and E. A. Seddon, Inorg. Chim. Ada, 1981, 47,

147.67. L. W. Yarbrough and M. B. Hall, Inorg. Chem., 1978,17, 2269.68. S. Evans, J. C. Green, M. L. H. Green, A. F. Orchard and D. W. Turner, Faraday Discuss. Chem. Soc,

1969,47, 112.69. S. Cradock, E. A. V. Ebsworth and A. Robertson, J. Chem. Soc, Dalton Trans., 1973, 22.70. M. B. Hall, J. Am. Chem. Soc, 1975, 97, 2057.71. M. F. Guest, M. B. Hall and I. H. Hillier, Mol. Phys., 1973, 25, 629.72. D. L. Lichtenberger and R. F. Fenske, Inorg. Chem., 1974,13, 486.73. D. L. Lichtenberger, A. C. Sarapu, and R. F. Fenske, Inorg. Chem., 1973,12, 702.74. M. A. Weiner and M. Lattman, Inorg. Chem., 1978,17,1084.75. B. R. Higginson, D. R. Lloyd, J. A. Connor and I. H. Hillier, J. Chem. Soc, Faraday Trans. 2, 1974, 70,

1418.76. L. H. Jones, R. S. McDowell and M. Goldblatt, Inorg. Chem., 1969, 8, 2349.77. L. H. Jones, J. Mol. Spectrosc, 1970, 36, 398.78. L. H. Jones and B. I. Swanson, Ace Chem. Res., 1976, 9, 128.79. L. H. Jones, Coord. Chem. Rev., 1966,1, 351.80. F. A. Cotton and R. M. Wing, Inorg. Chem., 1965, 4, 314.81. F. A. Cotton and C. S. Kraihanzel, J. Am. Chem. Soc, 1962, 84, 4432.82. D. J. Darensbourg and M. A. Murphy, Inorg. Chem., 1978,17, 884.83. R. F. Fenske, Pure Appl. Chem., 1971, 27, 61.84. K. G. Caulton and R. F. Fenske, Inorg. Chem., 1968, 7,1273.85. R. F. Fenske and R. L. DeKock, Inorg. Chem., 1970, 9, 1053.86. M. B. Hall and R. F. Fenske, Inorg. Chem., 1972,11, 1619.87. D. A. Brown and R. M. Rawlinson, J. Chem. Soc (A), 1969, 1530.88. W. A. G. Graham, Inorg. Chem., 1968, 7, 315.89. J. A. Timney, Inorg. Chem., 1979,18, 2502.90. J. K. Burdett, 'Molecular Shapes', John Wiley & Sons, New York, 1980.91. A. J. Rest and J. J. Turner, Chem. Commun., 1969, 375.92. M. A. Graham, A. J. Rest and J. J. Turner, J. Organomet. Chem., 1970, 24, C54.93. M. A. Graham, M. Poliakoff and J. J. Turner, J. Chem. Soc (A), 1971, 2939.94. M. A. Graham, R. N. Perutz, M. Poliakoff and J. J. Turner, J. Organomet. Chem., 1972, 34, C34.95. P. A. Breeze and J. J. Turner, J. Organomet. Chem., 1972, 44, C7.96. J. K. Burdett, Coord. Chem. Rev., 1978, 27, 1.97. J. K. Burdett, J. M. Grzybowski, R. N. Perutz, M. Poliakoff, J. J. Turner and R. F. Turner, Inorg. Chem.,

1978,17, 147.98. G. A. Ozin and A. Vander Voet, Acct. Chem. Res., 1973, 6, 313.99. E. P. Kundig and G. A. Ozin, J. Am. Chem. Soc, 1974, 96, 3820.

100. G. A. Ozin and A. Vander Voet, Prog. Inorg. Chem., 1975,19, 105.101. G. A. Ozin, Ace Chem. Res., 1977,10, 21.102. M. Elian and R. Hoffmann, Inorg. Chem., 1975,14, 1058.103. J. K. Burdett, J. Chem. Soc, Faraday Trans 2, 1974, 70,1599.104. J. K. Burdett, Adv. Inorg. Chem. Radiochem., 1978, 21, 113.105. D. A. Pensak and R. J. McKinney, Inorg. Chem., 1979,18, 3407.106. J. Demuynck, A. Strich and A. Veillard, Nouv. J. Chim., 1977,1, 217.107. P. J. Hay, J. Amer. Chem. Soc, 1978,100, 2411.108. R. J. Gillespie, 'Molecular Geometry', Van Nostrand-Reinhold, London, 1972.109. R. Hoffmann, J. M. Howell and E. L. Muetterties, / . Am. Chem. Soc, 1972, 94, 3047.110. M. M. L. Chen and R. Hoffmann, J. Am. Chem. Soc, 1976, 98,1647.111. A. D. Walsh, J. Chem. Soc, 1953, 2260, 2266, 2296, 2301 and 2306.112. B. M. Gimarc, 'Molecular Structure and Bonding', Academic Press, New York, 1979.113. M. B. Hall, Inorg. Chem., 1978,17, 2261.114. Y. Takahata, G. W. Schnuelle and R. G. Parr, / . Am. Chem. Soc, 1971, 93, 784.115. G. W. Schnuelle and R. G. Parr, J. Am. Chem. Soc, 1972, 94, 8974.116. L. E. Orgel, 'An Introduction to Transition Metal Chemistry', John Wiley, New York, 1960.117. D. L. Lewis and S. J. Lippard, / . Am. Chem. Soc, 1975, 97, 2697.118. M. G. B. Drew, Prog. Inorg. Chem., 1977, 23, 67.119. R. Hoffmann, B. F. Beier, E. L. Muetterties and A. R. Rossi, Inorg. Chem., 1977,16, 511.120. F. A. Cotton, J. Organomet. Chem., 1975,100, 29.121. K. N. Raymond, P. W. Corfield and J. A. Ibers, Inorg. Chem., 1968, 7, 1362.


122. R. L. DeKock, Inorg. Chem., 1971,10, 1205.123. Y. W. Yared, S. L. Miles, R. Bau and C. A. Reed, J. Am. Chem. Soc, 1977, 99, 7076.124. R. Busby, W. Klotzbucher and G. A. Ozin, Inorg. Chem., 1977,16, 822.125. J. K. Burdett, Inorg. Chem., 1975,14, 375.126. P. Kubacek and R. Hoffmann, / . Am. Chem. Soc, 1981,103,4320.127. O. Crichton, M. Poliakoff, A. J. Rest and J. J. Turner, J. Chem. Soc, Dalton Trans., 1973, 1321.128. H. Hiiber, E. P. Kundig, G. A. Ozin and A. J. Poe, J. Am. Chem. Soc, 1975, 97, 308.129. B. Davies, A. McNeish, M. Poliakoff and J. J. Turner, J. Am. Chem. Soc, 1977, 99, 7573.130. R. N. Perutz and J. J. Turner, / . Am. Chem. Soc, 1975, 97, 4791.131. J. Demuynck, E. Kochanski and A. Veillard, J. Am. Chem. Soc, 1979,101, 3467>132. A. M. Bond, R. Colton and J. M. McCormick, Inorg. Chem., 1977,16,155.133. A. M. Bond, R. Colton and J. J. Jackowski, Inorg. Chem., 1975,14, 274.134. R. Colton and J. M. McCormick, Aust. J. Chem., 1976, 29,1657.135. D. M. P. Mingos, J. Organomet. Chem., 1979,179, C29.136. J. A. Connor, in 'Transition Metal Clusters', ed. B. F. G. Johnson, J. Wiley & Sons, New York, 1980, p.

345.137. R. A. Faltynek and M. S. Wrighton, J. Am. Chem. Soc, 1978,100, 2701.138. H. B. Abrahamson, C. C. Frazier, D. S. Ginley, H. B. Gray, J. Lilienthal, D. R. Tyler and M. S. Wrighton,

Inorg. Chem., 1977,16,1554.139. M. S. Wrighton, Chem. Rev., 1974, 74, 401.140. D. R. Kidd and T. L. Brown, J. Am. Chem. Soc, 1978,100,4095.141. S. Bellard, K. A. Rubinson and G. M. Sheldrick, Ada Crystallogr. 1979, B35, 271.142. R. Hoffmann, T. A. Albright and D. L. Thorn, Pure Appl. Chem., 1978, 50, 1.143. D. M. P. Mingos, Adv. Organomet. Chem., 1977,15, 1.144. M. Elian, M. M.-L. Chen., D. M. P. Mingos and R. Hoffmann, Inorg. Chem., 1976,15,1148.145. T. A. Albright, Trans. Am. Crystallogr. Assoc, 1980,16, 35 and references therein.146. K. Wade, Adv. Inorg. Chem. Radiochem., 1976,18,1.147. K. Wade, in 'Transition Metal Cluster Compounds', ed. B. F. G. Johnson, J. Wiley & Sons, New York, 1980,

p. 193.148. D. M. P. Mingos, Nature (Phys. ScL), 1972, 236, 99.149. R. Mason and D. M. P. Mingos, in 'MTP International Review of Science, Physical Chemistry, Series 2',

Butterworths, London, 1975, vol. 11, p. 121.150. D. M. P. Mingos, Trans. Am. Crystallogr. Assoc, 1980,16, 17.151. F. G. A. Stone, / . Organomet. Chem., 1975,100, 257.152. T. V. Ashworth, M. Berry, J. A. Howard, M. Laguna and F. G. A. Stone, J. Chem. Soc, Chem. Commun.,

1979,45.153. G. K. Barker, M. Green, T. P. Onak, F. G. A. Stone, C. B. Ungermann and A. J. Welch, / . Chem. Soc, Chem.

Commun., 1978, 169.154. N. M. Boag, M. Green, R. M. Mills, G. N. Pain, F. G. A. Stone and P. Woodward, / . Chem. Soc, Chem.

Commun., 1980, 1171.155. J. A. K. Howard, J. C. Jeffrey, M. Laguna, R. Navarro and F. G. A. Stone, J. Chem. Soc, Chem. Commun.,

1979, 1170.156. F. Basolo and R. G. Pearson, 'Mechanisms of Inorganic Reactions', 2nd ed., Wiley, New York, 1967.157. R. J. Angelici, Organomet. Chem. Rev., 1968, 3,173.158. D. A. Brown, Inorg. Chim. Acta Rev., 1967,1, 35.159. H. Werner, Angew. Chem., Int. Ed. EngL, 1968, 7,930.160. K. Noack and F. Calderazzo, / . Organomet. Chem., 1967,10,101.161. A. Wojcicki, Adv. Organomet. Chem., 1973,11, 87.162. B. H. Byers and T. L. Brown, J. Am. Chem. Soc, 1975,97, 947.163. R. J. McKinney and D. A. Pensak, Inorg. Chem., 1979,18, 3413.164. J. D. Atwood and T. L. Brown, J. Am. Chem. Soc, 1976,98, 3160.165. J. D. Atwood and T. L. Brown, J. Am. Chem. Soc, 1976,98, 3155.166. C. L. Hyde and D. J. Darensbourg, Inorg. Chem., 1973,12, 1286.167. D. J. Darensbourg, G. R. Dobson and A. Moradi-Araghi, J. Organomet. Chem., 1976,116, C17.168. D. J. Darensbourg and A. H. Graves, Inorg. Chem., 1979,18,1257.169. J. D. Atwood and T. L. Brown, J. Am. Chem. Soc, 1976,98, 3160.170. M. A. Cohen and T. L. Brown, Inorg. Chem., 1976,15,1417.171. T. L. Brown and P. A. Bellus, Inorg. Chem., 1978,17, 3726.172. D. L. Lichtenberger and T. L. Brown, / . Am. Chem. Soc, 1978,100, 366.173. T. M. McHugh, A. J. Rest and D. J. Taylor, J. Chem. Soc, Dalton Trans., 1980, 1803.174. H. Berke and R. Hoffmann, J. Am. Chem. Soc, 1978,100, 7224.175. S. B. Butts, S. H. Strauss, E. M. Holt, R. E. Stimson, N. W. Alcock and D. F. Shriver, J. Am. Chem. Soc,

1980,102,5093.176. V. Bellagamba, R. Ercoli, A. Gamba and G. B. Suffritti, / . Organomet. Chem., 1980,190, 381.177. D. Saddei, H.-J. Freund and G. Hohlneicher, J. Organomet. Chem., 1980,186, 63.178. J. M. Howell and J. F. Olsen, unpublished results quoted in reference 174.179. N. Fonnesbech, J. Hjartkjaer and H. Johansen, Int. J. Quantum Chem., 1977,12, 95.180. J. H. Ammeter, N. Oswald and R. Bucher, Helv. Chim. Acta, 1975, 58, 671.181. M. F. Rettig and R. S. Drago, J. Am. Chem. Soc, 1969, 91, 3432.182. D. A. Brown, Transition Met. Chem., 1966, 3, 1.183. R. D. Fischer, Theor. Chim. Acta, 1963,1,418.184. I. H. Hillier and R. M. Canadine, Faraday Discuss. Chem. Soc, 1969, 47, 27.185. R. Prins, J. Chem. Phys., 1969, 50,4804.186. R. Prins and F. J. Reinders, Chem. Phys. Lett., 1969, 3, 45.


187. M. F. Rettig, in 'N.M.R. of Paramagnetic Molecules, Principles and Applications', ed. G. N. Lamar, W.de W. Horrocks, Jr. and R. H. Holm, Academic Press, New York, 1973, p. 217.

188. E. M. Shustorovich and M. E. Dyatkina, Dokl. Akad. Nauk SSSR, 1960,133, 141.189. D. R. Armstrong, R. Fortune and P. G. Perkins, J. Organomet. Chem., 1976, 111, 197.190. A. T. Armstrong, D. G. Carroll and S. P. McGlynn, J. Chem. Phys., 1967, 47, 1104.191. A. Botrel, P. Dibout and R. Lissillour, Theor. Chim. Ada, 1975, 37, 37.192. J. P. Dahl and C. J. Ballhausan, K. Dan. Vidensk. Selsk. Mat.-Fys. Medd., 1961, 33, 5.193. R. F. Kirchner, G. H. Loew and U. T. Mueller-Westerhoff, Theor. Chim. Ada, 1970, 41, 1.194. J. H. Schachtschneider, R. Prins and P. Ros, Inorg. Chim. Ada, 1967,1, 462.195. Y. S. Sohn, D. N. Hendrickson and H. B. Gray, J. Am. Chem. Soc, 1971, 93, 3603.196. M. Coutiere, J. Demuynck and A. Veillard, Theor. Chim. Ada, 1972, 27, 281.197. P. S. Bagus, U. I. Walgren and J. Almlof, J. Chem. Phys., 1976, 64, 2324.198. M. F. Guest, I. H. Hillier, B. R. Higginson and D. R. Lloyd, Mol. Phys., 1975, 29, 113.199. N. Rosch and K. H. Johnson, Chem. Phys. Lett., 1974, 24,179.200. S. Evans, M. L. H. Green, B. Jewitt, A. F. Orchard and C. F. Pygall, / . Chem. Soc, Faraday Trans. 2,1972,

68, 1847.201. S. Evans, A. F. Orchard and D. W. Turner, Int. J. Mass. Spedrom. Ion Phys., 1971, 7, 261.202. J. W. Rabalais, L. O. Werme, T. Bergmark, L. Karlsson, M. Hussain and K. Siegbahn, J. Chem. Phys.,

1972,57,1185.203. R. Prins and F. J. Reinders, J. Am. Chem. Soc, 1969, 91,4929.204. R. Prins, Mol. Phys., 1970,19, 603.205. A. Almenningen, S. Samdel and A. Haaland, J. Chem. Soc, Chem. Commun., 1977, 14.206. S. Evans, J. C. Green and S. E. Jackson, J. Chem. Soc, Faraday Trans. 2, 1972, 68, 249.207. S. Evans, J. C. Green, S. E. Jackson and B. R. Higginson, J. Chem. Soc, Dalton Trans., 1974, 68, 304.208. R. D. Feltham, P. Sogo and M. Calvin, / . Chem. Phys., 1957, 26, 1354.209. G. H. Olive and S. Olive, Z. Phys. Chem. (Frankfurt am Main), 1967, 56, 223.210. C. Cauletti, J. C. Green, M. R. Kelly, P. Powell, J. van Tilborg, J. Robbins and J. Smart, J. Eledr. Spedrosc,

1981,19,327.211. M. L. H. Green, Pure Appl. Chem., 1978, 50, 27.212. T. J. Marks, Progr. Inorg. Chem., 1978, 24, 51.213. A. L. Wayda and W. J. Evans, J. Am. Chem. Soc, 1978,100,7119.214. M. F. Lappert, P. I. W. Yarrow, J. L. Atwood, R. Shakir and J. Holton, / . Chem. Soc, Chem. Commun.,

1980,987.215. P. L. Watson, J. Chem. Soc, Chem. Commun., 1980, 652.216. J. Holton, M. F. Lappert, D. G. H. Ballard, R. Pearce, J. L. Atwood and W. E. Hunter, / . Chem. Soc, Dalton

Trans., 1979,45.217. A. Streitweiser, Jr., U. Muller-Westerhoff, G. Sonnichsen, F. Mares, D. G. Morrell, K. O. Hodgson and

C. A. Harmon, / . Am. Chem. Soc, 1973, 95, 8644.218. A. Streitweiser and U. Muller-Westerhoff, J. Am. Chem. Soc, 1968, 90, 7364.219. R. G. Hayes and N. Edelstein, J. Am. Chem. Soc, 1972, 94, 8688.220. N. Rosch and A. Streitweiser, Jr., J. Organomet. Chem., 1978,145, 195.221. J. P. Clark and J. C. Green, / . Chem. Soc, Dalton Trans., 1977, 505.222. J. P. Clark and J. C. Green, J. Organomet. Chem., 1976,112, C14.223. I. Fragala, G. Condorelli, P. Zanella and E. Tondello, / . Organomet. Chem., 1976,122, 357.224. J. C. Green, unpublished results quoted in reference 36.225. D. W. Clack and K. D. Warren, J. Organomet. Chem., 1978,149, 401.226. D. W. Clack and K. D. Warren, / . Organomet. Chem., 1978,161, C55.227. D. W. Clack and K. D. Warren, J. Organomet. Chem., 1978,157, 421.228. D. W. Clack and K. D. Warren, J. Organomet. Chem., 1978,162, 83.229. D. W. Clack and K. D. Warren, Inorg. Chim. Ada, 1978, 27,105.230. D. W. Clack and M. Monshi, / . Organomet. Chem., 1976,116, C41.231. J. D. Zeinstra and W. C. Nieuwpoort, Inorg. Chim. Ada, 1978, 30, 103.232. M. F. Hawthorne, D. C. Young and P. A. Wegner, J. Am. Chem. Soc, 1965, 87, 1818.233. M. F. Hawthorne, / . Organomet. Chem., 1975,100, 97.234. K. P. Callahan and M. F. Hawthorne, Adv. Organomet. Chem., 1976,14, 145.235. D. M. P. Mingos, J. Chem. Soc, Dalton Trans., 1977, 602.236. D. A. Brown, M. O. Fanning and N. J. Fitzpatrick, Inorg. Chem., 1980, 19, 1822.237. R. M. Wing. / . Am. Chem. Soc, 1967, 89, 5599; 1968, 90, 4828.238. H. M. Colquhoun, T. J. Greenough and M. G. H. Wallbridge, J. Chem. Soc, Chem. Commun., 1976,

1019.239. R. M. Wing, J. Am. Chem. Soc, 1970, 92, 1187.240. L. F. Warren and M. F. Hawthorne, J. Am. Chem. Soc, 1968, 90, 4823.241. P. A. Wegner, Inorg. Chem., 1975,14, 212.242. C. Glidewell, J. Organomet. Chem., 1975,102, 339.243. D. M. P. Mingos and M. I. Forsyth, J. Organomet. Chem., 1978,146, C37.244. A. Almenningen, E. Gard, A. Haaland and J. Brunvoll, J. Organomet. Chem., 1976,107, 273.245. J. D. Cox and G. Pilcher, 'Thermochemistry of Organic and Organometallic Compounds', Academic Press,

London, 1970.246. H. A. Skinner, Adv. Organomet. Chem., 1964, 2, 49.247. G. Huttner, H. H. Brintzinger, L. G. Bell, P. Friedrich, V. Bejenke and D. Neugebauer, J. Organomet. Chem.,

1978,145,329.248. G. Huttner and S. Lange, Ada Crystallogr.,Sed. B, 1972, 28, 2049.249. M. Y. Darensbourg and E. L. Muetterties, J. Am. Chem. Soc, 1978,100, 7425.250. P. L. Timms and R. B. King, J. Chem. Soc, Chem. Commun., 1978, 898.


251. A. Lucherini and L. Porri, / . Organomet. Chem., 1978, 155, C45.252. E. Schumacher and R. Taubenest, Helv. Chim. Acta, 1964,47,1525.253. H. Werner and S. Salzer, Angew. Chem., Int. Ed. Engl., 1972,11,930.254. D. C. Beer, V. R. Miller, L. G. Sneddon, R. N. Grimes, M. Mathew and G. J. Palenik, J. Am. Chem. Soc,

1973,95,3046.255. R. Weiss and R. N. Grimes, / . Organomet. Chem., 1976,113, 29.256. G. E. Herberich, J. Hengesbach, U. Kolle and W. Oschmann, Angew. Chem., Int. Ed. Engl., 1977, 16,

42.257. G. Herberich, J. Hengesbach, U. Kolle> G. Huttner and A. Frank, Angew. Chem., Int. Ed. Engl., 1976,15,

433.258. W. Siebert and K. Kinberger, Angew. Chem. Int. Ed. Engl., 1976,15,434.259. W. Siebert and W. Rothermel, Angew. Chem., Int. Ed. Engl., 1977,16, 333.260. J. W. Lauher, M. Elian, R. H. Summerville and R. Hoffmann, / . Am. Chem. Soc, 1976, 98, 3219.261. J. Moraczewski and W. E. Geiger, Jr., J. Am. Chem. Soc, 1978,100, 7429.262. P. C. Wailes, R. S. P. Coutts and H. Weigold, 'Organometallic Chemistry of Titanium, Zirconium and

Hafnium', Academic Press, New York, 1974.263. K. Prout, T. S. Cameron, R. A. Forder, S. R. Critchley, B. Denton and G. V. Rees, Acta Crystallogr., Sect.

B, 1974,30,2290.264. L. J. Guggenberger, Inorg. Chem., 1973,12, 294.265. K. M. Melmed, D. Coucouvanis and S. J. Lippard, Inorg. Chem., 1973,12, 232.266. J. C. Green, M. L. H. Green and C. K. Prout, J. Chem. Soc, Chem. Commun., 1972, 421.267. J. L. Peterson and L. F. Dahl, / . Am. Chem. Soc, 1974, 96, 2248.268. J. L. Peterson and L. F. Dahl, / . Am. Chem. Soc, 1975, 97, 6416.269. J. L. Peterson, D. L. Lichtenberger, R. F. Fenske and L. F. Dahl, / . Am. Chem. Soc, 1975,97, 6433.270. J. C. Green, S. E. Jackson and B. Higginson, / . Chem. Soc, Dalton Trans., 1975, 403.271. J. W. Lauher and R. Hoffmann, J. Am. Chem. Soc, 1976, 98,1729.272. H. H. Brintzinger and L. S. Bartell, J. Am. Chem. Soc, 1970, 92,1105.273. C. J. Ballhausan and J. P. Dahl, Acta Chem. Scand., 1961,15, 1333.274. N. W. Alcock, / . Chem. Soc (A), 1967, 2001.275. R. Bau, R. G. Teller, S. W. Kirtley and T. F. Koetzle, Ace Chem. Res., 1979,12,176.276. M. F. Guest, I. H. Hillier, B. R. Higginson and D. R. Lloyd, Mol. Phys., 1975, 29,113.277. I. H. Hillier and V. R. Saunders, Mol. Phys., 1972, 23, 449.278. J. A. Connor, L. M. R. Derrick, M. B. Hall, I. H. Hillier, M. F. Guest, B. R. Higginson and D. R. Lloyd,

Mol. Phys., 1974,28,1193.279. J. A. Connor, L. M. R. Derrick, I. H. Hillier, M. F. Guest and D. R. Lloyd, Mol. Phys., 1976, 31, 23.280. M. B. Hall, I. H. Hillier, J. A. Connor, M. F. Guest and D. R. Lloyd, Mol. Phys., 1975, 30, 839.281. M.-M. Coutiere, J. Demuynck and A. Veillard, Theor. Chim. Acta, 1972, 27, 281.282. D. W. Clack and W. Smith, Inorg Chim. Acta, 1976, 20,93.283. M. M. Rohmer, J. Demuynck and A. Veillard, Theor. Chim. Acta, 1974, 36,93.284. D. A. Brown, N. J. Fitzpatrick and N. J. Mathews, J. Organomet. Chem., 1975,88, C27.285. D. A. Brown, N. J. Fitzpatrick, N. J. King and N. J. Mathews, / . Organomet. Chem., 1975,104, C9.286. N. J. Fitzpatrick and N. J. Mathews, J. Organomet. Chem., 1973, 61, C45.287. D. A. Brown, H. L. Clarke and N. J. Fitzpatrick, J. Organomet. Chem., 1973, 47, Cl 1.288. D. L. Lichtenberger and R. F. Fenske, J. Am. Chem. Soc, 1976,98, 50.289. D. L. Lichtenberger, D. Sellman and R. F. Fenske, / . Organomet. Chem., 1976,117, 253.290. T. H. Whitesides, D. L. Lichtenberger and R. A. Budnik, Inorg. Chem., 1975,14,68.291. D. G. Carrol and S. P. McColynn, Inorg. Chem., 1968, 7,1285.292. P. J. Hauser, R. S. Evans and A. F. Schreiner, Theor. Chim. Acta, 1973, 32, 87.293. P. G. Perkins, I. C. Robertson and J. M. Scott, Theor. Chim. Acta, 1971, 22, 299.294. D. W. Clack, M. Monshi and L. A. P. Kane-Maguire, / . Organomet. Chem., 1976,107, C40.295. D. W. Clack, M. Monshi and L. A. P. Kane-Maguire, J. Organomet. Chem., 1976,120, C25.296. D. A. Brown and R. M. Rawlinson, / . Chem. Soc (A), 1969, 1534.297. G. Schmitt, S. Ozman, B. Hoffman and J. Fleischhauer, J. Organomet. Chem., 1976,114,179.298. M. J. Nolte and G. Gafner, Acta Cry stallogr., Sect. B, 1974, 30, 738.299. G. M. Whitesides and J. S. Fleming, / . Am. Chem. Soc, 1967,89, 2855.300. T. A. Albright, R. Hoffmann, Y.-C. Tse and D. Ottavio, / . Am. Chem. Soc, 1979,101, 3812.301. D. M. P. Mingos, Nature (Phys. Sci.), 1971, 229, 193.302. R. Mason, Chem. Soc. Rev., 1972,1,431.303. N. Trong Anh, M. Elian and R. Hoffmann, J. Am. Chem. Soc, 1978,100,110.304. F. A. Cotton, Ace Chem. Res., 1968,1, 257.305. A. J. Welch, / . Chem. Soc, Dalton Trans., 1975, 1473.306. A. J. Welch, J. Chem. Soc, Dalton Trans., 1975, 2270.307. M. I. Forsyth, A. J. Welch and D. M. P. Mingos, J. Chem. Soc, Dalton, Trans., 1978, 1363.308. H. M. Colquhoun, T. J. Greenhough and M. G. H. Wallbridge, J. Chem. Soc, Dalton Trans., 1979,

619.309. H. M. Colquhoun, T. J. Greenhough and M. G. H. Wallbridge, / . Chem. Soc, Dalton Trans., 1978,

303.310. G. K. Barker, M. Green, F. G. A. Stone and A. J. Welch, J. Chem. Soc, Dalton Trans., 1980, 1186.311. L. R. Byers and L. F. Dahl, Inorg. Chem., 1980,19, 277.312. V. W. Day, K. J. Reimer and A. Shaver, J. Chem. Soc, Chem. Commun., 1975, 403.313. L. F. Dahl and C. H. Wei, Inorg. Chem., 1963, 2, 713.314. M. R. Gerloch and R. Mason, J. Chem. Soc, 1965, 296.315. M. R. Churchill, Inorg. Chem., 1965, 4, 1734.316. K. J. Klabunde and K. Neuenschwander, Inorg. Chem., 1980,19, 1221.


317. L. J. Radonovich, F. J. Koch and T. A. Albright, Inorg. Chem., 1980,19, 3373.318. T. A. Albright and R. Hoffmann, Chem. Ber., 1978, 111, 1578.319. M. D. McClure and D. L. Weaver, / . Organomet. Chem., 1973, 54, C59.320. C. Mealli, S. Midollini, S. Moneti and L. Sacconi, Angew. Chem., Int. Ed. Engl., 1980,19, 931.321. P. M. Maitlis, Ace. Chem. Res., 1978,11, 301.322. D. M. P. Mingos, P. C. Minshall, M. B. Hursthouse, S. D. Willoughby and K. W. A. Malik, J. Organomet.

Chem., 1979,181,169.323. W. Rigby, H. B. Lee, P. M. Bailey, J. A. McCleverty and P. M. Maitlis, J. Chem. Soc.,Dalton Trans., 1979,

387.324. P. Hofmann, Z. Naturforsch, Teil B 1978, 33B, 251.325. F. A. Cotton, A. Davison and J. W. Faller, J. Am. Chem. Soc, 1966,88,4507.326. F. A. Cotton, A. Davison and A. Musco, J. Am. Chem. Soc, 1967,89, 6796.327. F. A. Cotton, 'Dynamic Nuclear Magnetic Resonance Spectroscopy', ed. L. M. Jackman and F. A. Cotton,

Academic Press, New York, 1975, p. 378.328. K. Vrieze, in 'Dynamic Nuclear Magnetic Resonance Spectroscopy', ed. L. M. Jackman and F. A. Cotton,

Academic Press, New York, 1975, p. 441.329. B. E. Mann, J. Chem. Soc, Dalton Trans., 1978, 1761.330. F. A. Cotton, B. G. De Boer and M. D. Laprade, 'Internat. Congr. Pure Appl. Chem.', 23rd Spec. Lect.,

1971,6,1.331. R. B. Woodward, M. Rosenblum and M. C. Whiting, J. Am. Chem. Soc, 1952, 74, 3458.332. D. Lewis and D. Peters, 'Facts and Theories of Aromaticity', Macmillan, London, 1975.333. G. E. Coates, M. L. H. Green and K. Wade, 'Organometallic Compounds', 3rd ed., Methuen, London, vol.

2, 1968.334. W. E. Silverthorn, Adv. Organomet. Chem., 1975,13,47.335. B. Nicholls and M. Whiting, / . Chem. Soc, 1959, 551.336. R. Pettit, J. Organomet. Chem., 1975,100, 205.337. B. E. Bursten and R. F. Fenske, Inorg. Chem., 1979,18,1760.338. D. M. P. Mingos, J. Chem. Soc, Dalton Trans., 1977, 20.339. D. M. P. Mingos, / . Chem. Soc, Dalton Trans., 1977, 26.340. D. M. P. Mingos, J. Chem. Soc, Dalton Trans., 1977, 31.341. W. C. Herndon, J. Am. Chem. Soc, 1980,102,1538.342. J. Aihara, Bull. Chem. Soc Jpn., 1978, 51,1541.343. W. C. Herndon, Isr. J. Chem., 1980, 20, 276.344. B. J. Nicholson, J. Am. Chem. Soc, 1966,88, 5156.345. M. J. S. Dewar, Bull. Soc. Chim. Fr., 1951,18, C71.346. J. Chatt and L. A. Duncanson, / . Chem. Soc, 1953, 2939.347. F. R. Hartley, Chem. Rev., 1968,69,799.348. F. R. Hartley, Angew. Chem., Int. Ed. Engl., 1972,11, 596.349. H. W. Quinn and J. H. Tsai, Adv. Inorg. Chem. Radiochem., 1969,12, 217.350. P. Heimbach and R. Traunmuller, 'Chemie der Metall-Olefin Komplexes', Verlag Chemie, Weinheim,

1970.351. E. O. Greaves, C. J. L. Lock and P. M. Maitlis, Can. J. Chem., 1968, 46, 3879.352. N. Rosch, R. P. Messmer and K. H. Johnson, / . Am. Chem. Soc, 1974, 96, 3855.353. W. C. Hamilton, K. A. Klanderman and R. Spratley, Ada Crystallogr., Sect. A, 1971, 25, 5172.354. J. A. J. Jarvis, B. T. Kilbourn and P. G. Owston, Acta Crystallogr., Sect. B, 1971, 27, 366.355. S. D. Ittel and J. A. Ibers, Adv. Organomet. Chem., 1976,14, 33.356. L. M. Muir, K. W. Muir and J. A. Ibers, Faraday Discuss. Chem. Soc, 1969,47, 84.357. J. M. Baraban and J. A. McGinnety, / . Am. Chem. Soc, 1975,97,4232.358. L. J. Guggenberger and R. Cramer, / . Am. Chem. Soc, 1972,94, 3779.359. A. J. Schultz, R. K. Brown, J. M. Williams and R. R. Schrock, / . Am. Chem. Soc, 1981,103,169.360. K. Jones and C. Kruger, Angew. Chem., Int. Ed. Engl., 1980,19, 520.361. A. D. Walsh, Trans. Faraday Soc, 1949,45,179.362. J. A. McGinnety and J. A. Ibers, Chem. Commun., 1968, 235.363. C. A. Tolman, J. Am. Chem. Soc, 1974, 96, 2780.364. M. J. S. Dewar, 'The Molecular Orbital Theory of Organic Chemistry', McGraw-Hill, New York, 1969.365. C. R. Brundle, M. B. Robin, N. A. Kuebler and H. Basch, J. Am. Chem. Soc, 1972,94,1451.366. G. Belanger and C. Sandorfy, J. Chem. Phys., 1971, 55, 2055.367. T. A. Albright, R. Hoffmann, J. C. Thibeault and D. L. Thorn, J. Am. Chem. Soc, 1979,101, 3801.368. W. H. Baddley, J. Am. Chem. Soc, 1968, 90, 3705.369. J. Chatt, Chimia Inorganica, Acad. Nazi. Lincei, Roma, 1961,113.370. S. Sakaki and H. Kato, Bull. Chem. Soc Jpn., 1973, 46, 2227.371. S. Sakaki, H. Kato, H. Kanai and K. Tarama, Bull. Chem. Soc Jpn., 1975, 48, 813.372. S. Sakaki, H. Kato and T. Kawamura, Bull. Chem. Soc Jpn., 1975, 48,195.373. S. Sakaki, N. Kuduo and A. Ohyoshi, Inorg. Chem., 1911,16, 202.374. A. C. Blizzard and D. P. Santry, / . Am. Chem. Soc, 1968,90, 5749.375. J. H. Nelson, K. S. Wheelock, L. C. Cusachs and H. B. Jonassen, Inorg. Chem., 1972,11,422.376. J. H. Nelson, K. S. Wheelock, L. C. Cusachs and H. B. Jonassen, J. Am. Chem. Soc, 1969, 91, 7005.377. K. S. Wheelock, J. M. Nelson, L. C. Cusachs and H. B. Jonassen, J. Am. Chem. Soc, 1970, 92, 5110.378. J. N. Murrell and C. E. Scollary, J. Chem. Soc, Dalton Trans., 1977, 1034.379. J. G. Norman, Jr., Inorg. Chem., 1977,16, 1328.380. C. D. Cook, K. Y. Wan, U. Gelius, K. Hamrin, G. Johansson, E. Olsson, H. Siegbahn, C. Nordling and K.

Siegbahn, / . Am. Chem. Soc, 1971,93, 1904.381. D. T. Clark, D. B. Adams and D. Biggs, J. Chem. Soc, Chem. Commun., 1971, 602.382. W. M. Riggs, Anal. Chem., 1972, 44, 830.


383. R. Mason, D. M. P. Mingos, G. Rucci and J. A. Connor, / . Chem. Soc, Dalton Trans., 1972, 1729.384. B. Akermark, M. Almemark, J. Almlof, J. E. Backvall, B. Roos and A. Stdgard, J. Am. Chem. Soc, 1977,

99,4617.385. C. Pedone and E. Benedetti, J. Organomet. Chem., 1971, 29,443.386. S. Merlino, R. Lazzaroni and G. Montagnoli, J. Organomet. Chem., 1971, 30, C93.387. A. De Renzi, B. DiBlasio, G. Paiaro, A. Panunzi and C. Pedone, Gazz. Chim. Ital., 1976,106, 765.388. S. C. Nyburg, K. Simpson and W. Wong-Ng, J. Chem. Soc, Dalton Trans., 1976, 1865.389. F. Sartori and L. Leoni, Acta Crystallogr., Sect. B, 1976, B32, 145.390. F. A. Cotton, J. N. Francis, B. A. Frenz and M. Tsutsui, / . Am. Chem. Soc, 1973, 95, 2483.391. A. McAdam, J. N. Francis and J. A. Ibers, / . Organomet. Chem., 1971, 29, 149.392. L. J. Guggenberger, Inorg. Chem., 1973,12,499.393. D. J. Sepelak, C. G. Pierpont, E. K. Barefield, J. T. Budz and C. A. Poffenberger, / . Am. Chem. Soc, 1976,

98,6178.394. R. Countryman and B. R. Penfold, Chem. Commun., 1971, 1598.395. B. E. R. Schilling, R. Hoffmann and D. L. Lichtenberger, J. Am. Chem. Soc, 1979,101, 585.396. P. Hofmann, Angew. Chem., Int. Ed. Engl., 1977,16, 536.397. U. Koemm, C. G. Kreiter and H. Strack, / . Organomet. Chem., 1978,148,179.398. C. D. Cook and K. Y. Wan, Inorg. Chem., 1971,10, 2696.399. J. A. Segal and B. F. G. Johnson, J. Chem. Soc, Dalton Trans., 1975, 677 and 1990.400. S. T. Wilson, N J . Coville, J. R. Shapely and J. A. Osborn, J. Am. Chem. Soc, 1974, 96,4038.401. J. W. Faller, Adv. Organomet. Chem., 1977,16, 211.402. T. Zeigler, Inorg. Chem., 1979,18,1558.403. T. Yoshida, K. Tatsumi and S. Otsuka, Pure Appl. Chem., 1980, 52, 713.404. K. Tatsumi, F. Fueno, A. Nakamura and S. Otsuka, Bull. Chem. Soc. Jpn., 1976, 49, 2164 and 2170.405. C. E. Holloway, G. Hulley, B. F. G. Johnson and J. Lewis, J. Chem. Soc. (A), 1969, 53.406. L. L. Wright, R. M. Wing, M. F. Rettig and G. R. Wiger, J. Am. Chem. Soc, 1980,102, 5949.407. D. M. P. Mingos, in 'Metal Olefin Complexes', ed. M. A. Bennett, Wiley, London, 1981, in press.408. N. Rosch and R. Hoffmann, Inorg. Chem., 1974,13, 2656.409. M. Green, J. A. K. Howard, J. L. Spencer and F. G. A. Stone, J. Chem. Soc, Chem. Commun., 1975,

449.410. T. A. Albright and J. K. Burdett, Inorg. Chem., 1979,18, 2112.411. F. L. Bowden and A. B. P. Lever, Organomet. Chem. Rev. (A), 1968, 3, 227.412. B. W. Davies and N. C. Payne, Inorg. Chem., 1974,13, 1848.413. G. B. Robertson and P. O. Whimp, / . Am. Chem. Soc, 1975, 97, 1051.414. B. W. Davies, R. J. Puddephatt and N. C. Payne, Can. J. Chem., 1972, 50, 2276.415. G. R. Davies, W. Hewertson, R. H. B. Mais, P. G. Owston and C. G. Patel, J. Chem. Soc. (A), 1970,

1873.416. J. L. Davidson, M. Green, D. W. A. Sharp, F. G. A. Stone and A. J. Welch, J. Chem. Soc, Chem. Commun.,

1974,706.417. M. Bottrill and M. Green, / . Am. Chem. Soc, 1977, 99, 5795.418. J. W. Faller and H. H. Murray, / . Organomet. Chem., 1979,172,171.419. J. A. K. Howard, R. F. D. Stansfield and P. Woodward, J. Chem. Soc, Dalton Trans., 1976, 246.420. J. L. Davidson and D. W. A. Sharp, J. Chem. Soc, Dalton Trans., 1975, 2531.421. J. L. Davidson, / . Chem. Soc, Chem. Commun., 1980, 113.422. J. W. McDonald, W. E. Newton, C. T. C. Creedy and J. L. Corbin, J. Organomet. Chem., 1975, 92,

C25.423. J. W. McDonald, J. L. Corbin and W. E. Newton, / . Am. Chem. Soc, 1975, 97,1970.424. P. L. Watson and R. G. Bergman, / . Am. Chem. Soc, 1980,102, 2698.425. J. L. Templeton and B. C. Ward, / . Am. Chem. Soc, 1980,102, 3288; 6568.426. P. S. Braterman, J. L. Davidson and D. W. A. Sharp, J. Chem. Soc, Dalton Trans., 1976, 241.427. R. B. King, Inorg. Chem., 1968, 7, 1044.428. R. Hoffmann and P. Hofmann, J. Am. Chem. Soc, 1976, 98, 598.429. T. A. Albright, P. Hofmann and R. Hoffmann, / . Am. Chem. Soc, 1977, 99, 7546.430. T. A. Albright, P. Hofmann and R. Hoffmann, Chem. Ber., 1978, 111, 1591.431. F. A. Cotton, V. W. Day, B. A. Frenz, K. I. Hardcastle and J. M. Troup, J. Am. Chem. Soc, 1973, 95,

4522.432. M. R. Churchill and R. Mason, Adv. Organomet. Chem., 1967, 5, 93.433. R. E. Davis and R. Pettit, / . Am. Chem. Soc, 1970, 92, 716.434. M. J. Barrow, O. S. Mills, P. L. Pauson and F. Hagul, Acta Crystallogr., Sect. B, 1974, 30, 1635.435. M. R. Churchill and F. R. Scholer, Inorg. Chem., 1969, 8, 1950.436. T. Debaerdemaeker, Angew. Chem., Int. Ed. Engl., 1976,15, 504.437. L. Weber, C. Kruger and Y.-H. Tsay, Chem. Ber., 1978, 111, 1709.438. P. E. Baikie and O. S. Mills, J. Chem. Soc (A), 1968, 2704.439. J. Stegemann and H. J. Lindner, J. Organomet. Chem., 1979,166, 223.440. R. Seip, Acta Chem. Scand., 1972, 26,1966.441. E. L. Muetterties and F. J. Hirsekorn, / . Am. Chem. Soc, 1974, 96, 7920.442. P. A. Dobosh, D. G. Gresham, D. J. Kowalski, C. P. Lillya and E. S. Magyar, Inorg. Chem., 1978, 17,

1775.443. L. Kruczynski and J. Takats, J. Am. Chem. Soc, 1974, 96, 932.444. C. G. Kreiter, M. Lang and H. Strack, Chem. Ber., 1975,108,1502.445. B. Lubke and U. Behrens, J. Organomet. Chem., 1978,149, 327.446. K. Hoffmann and E. Weiss, J. Organomet. Chem., 1977,128, 237.447. R. L. Beddoes, P. F. Lindley and O. S. Mills, Angew. Chem., Int. Ed. Engl., 1970, 82, 293.448. M. J. Barrow and O. S. Mills, / . Chem. Soc (̂ 4), 1971, 1982.


449. P. E. Baikie and O. S. Mills, J. Chem. Soc. (A), 1969, 328.450. E. Benedetti, G. Maglio, R. Palumbo and C. Pedone, J. Organomet. Chem., 1973, 60, 189.451. Y. Kitano, T. Kajimoto, M. Kashiwagi and Y. Kinoshita, J. Organomet. Chem., 1971, 33, 123.452. G. Perego, G. Del Piero and M. Cesari, Cryst. Struct. Commun., 1974, 3, 721.453. P. W. Jolly and G. Wilke, 'The Organic Chemistry of Nickel', Academic Press, New York, 1974, vol. 1, p.

264.454. C. Kruger, personal communication quoted in reference 367.455. C. R. Nurse and D. M. P. Mingos, J. Organomet. Chem., 1980,184, 281.456. B. E. Mann and P. M. Maitlis, J. Chem. Soc, Chem. Commun., 1976, 1058.457. S. H. Taylor, B. E. Mann and P. M. Maitlis, / . Organomet. Chem., 1978,145, 255.458. A. Keasey, P. M. Bailey and P. M. Maitlis, J. Chem. Soc, Chem. Commun., 1978, 1.42.459. Y. Becker and J. K. Stille, / . Am. Chem. Soc, 1978,100, 845.460. A. Sonoda, B. E. Mann and P. M. Maitlis, / . Chem. Soc, Chem. Commun., 1975,108.461. T. A. Albright, J. Organomet. Chem., in press.462. C. A. Tolman, Chem. Soc. Rev., 1972,1, 337.463. C. A. Tolman, Chem. Rev., 1977, 77, 313.464. F. D. Mango and J. H. Schachtschneider, / . Am. Chem. Soc, 1967, 89, 2484.465. M. J. S. Dewar, Angew. Chem., Int. Ed. Engl., 1971,10, 761.466. R. G. Pearson, 'Symmetry Rules for Chemical Reactions', Wiley, New York, 1976.467. K. Fukui and S. Inagaki, J. Am. Chem. Soc, 1975,97,4445.468. R. Pettit, H. Sugahara, J. Wristers and W. Merk, Faraday Discuss. Chem. Soc, 1969, 47, 71.469. W. Th. A. M. van der Lugt, Tetrahedron Lett., 1970, 2281.470. R. Traunmuller, O. E. Polansky, P. Heimbach and G. Wilke, Chem. Phys. Lett., 1969, 3, 300.471. R. B. Woodward and R. Hoffmann, 'Conservation of Orbital Symmetry', Academic Press, New York,

1969.472. L. Cassar and J. Halpern, Chem. Commun., 1970, 1082; 1971, 40.473. L. A. Paquette and J. C. Stowell, J. Am. Chem. Soc, 1970, 92, 2584.474. J. Chatt, L. M. Venanzi and L. M. Vallarino, J. Chem. Soc, 1957, 2496 and 3413.475. J. P. Collman and L. S. Hegedus, 'Principles and Applications of Organotransition Metal Chemistry',

University Science Books, Mill Valley, California, 1980.476. S. G. Davies, M. L. H. Green and D. M. P. Mingos, Tetrahedron, 1978, 34, 3047.477. O. Eisenstein and R. Hoffmann, J. Am. Chem. Soc, 1980,102, 6148.478. I. FJeming, 'Frontier Orbitals and Organic Chemical Reactions', J. Wiley, New York, 1976.479. B. Akermark, M. Almemark, J. Almlof, J. E. Backvall, B. Roos and A. St^gard, J. Am. Chem. Soc, 1977,

99,4617.480. S. G. Davies, M. L. H. Green and D. M. P. Mingos, Nouv. J. Chim., 1977,1,445.481. S. G. Davies, M. L. H. Green and D. M. P. Mingos, in 'Reactions of Coordinated Ligands', ed. P. R. Bra-

terman, Plenum Press, 1982, in press.482. D. W. Clack and L. A. P. Kane-Maguire, / . Organomet. Chem., 1979,174,199.483. D. W. Clack and L. A. P. Kane-Maguire, J. Organomet. Chem., 1978,145, 201.484. D. W. Clack, M. Monshi and L. A. P. Kane-Maguire, / . Organomet. Chem., 1976,107, C40.485. D. W. Clack and K. D. Warren, J. Organomet. Chem., 1978,162, 83.486. D. W. Clack, M. Monshi and L. A. P. Kane-Maguire, / . Organomet. Chem., 1976,120, C25.487. B. E. R. Schilling, R. Hoffmann and J. W. Faller, J. Am. Chem. Soc, 1979,101, 592.488. T. A. Albright and B. A. Carpenter, Inorg. Chem., 1980,19, 3092.489. H. Hogeveen and H. C. Volger, J. Am. Chem. Soc, 1967, 89, 2486.490. K. C. Bishop, III, Chem. Rev., 1976, 76, 461.491. M. Sakai, H. Yamaguchi, H. H. Westberg and S. Masamune, J. Am. Chem. Soc, 1971, 93, 1043.492. W. Sleigeir, R. Case, J. S. McKennis and R. Pettit, / . Am. Chem. Soc, 1974, 96, 287.493. A. R. Pinhas and B. K. Carpenter, / . Chem. Soc, Chem. Commun., 1980, 15.494. A. R. Pinhas, A. G. Samuelson, R. Risemberg, E. V. Arnold, J. Clardy and B. K. Carpenter, J. Am. Chem.

Soc, 1981,103,1668.495. D. L. Thorn and R. Hoffmann, / . Am. Chem. Soc, 1978,100, 2079.496. S. Sakaki, H. Kato, H. Kana and K. Tarama, Bull. Chem. Soc. Jpn., 1975, 48, 813.497. P. S. Braterman and R. J. Cross, Chem. Soc. Rev., 1973, 2, 271.498. R. G. Pearson, Top. Curr. Chem., 1973, 41, 75.499. A. Dedieu, Inorg. Chem., 1980,19, 375.500. P. Cossee, Reel. Trav. Chim. Pays-Bas, 1966, 85, 1151.501. D. R. Armstrong, P. G. Perkins and J. J. P. Stewart, J. Chem. Soc, Dalton Trans., 1972, 1972.502. P. Cossee, P. Ros and J. H. Schachtschneider, in 'Proceedings of the 4th International Congress on Catalysis',

1970,287.503. D. R. Armstrong, R. Fortune and P. G. Perkins, J. Catal., 1976, 42, 435.504. K. S. Wheelock, J. H. Nelson, J. D. Kelly, H. B. Jonassen and L. C. Cusachs, J. Chem. Soc, Dalton Trans.,

1973, 1457.505. S. Sakaki, H. Kato, H. Kanai and K.Tarama, Bull. Chem. Soc. Jpn., 1974, 47, 377.506. J. Ph. Grima, F. Choplin and G. Kaufmann, / . Organomet. Chem., 1977,129, 221.507. R. J. McKinney, J. Chem. Soc, Chem. Commun., 1980, 490.508. K. J. Ivin, J. J. Rooney, C. D. Stewart, M. L. H. Green and R. Mahtab, J. Chem. Soc, Chem. Commun.,

1978,604.509. P. S. Braterman, J. Chem. Soc, Chem. Commun., 1979, 70.510. B. Akermark and A. Ljungqvist, J. Organomet. Chem., 1979,182, 59.511. A. Stockis and R. Hoffmann, J. Am. Chem. Soc, 1980,102, 2952.512. R. J. McKinney, D. L. Thorn, R. Hoffmann and A. Stockis, J. Am. Chem. Soc, 1981,103, 2595.513. R. H. Grubbs, A. Miyashita, M. Liu and P. L. Burk, J. Am. Chem. Soc, 1977, 99, 3863.


514. R. H. Grubbs and A. Miyashita, J. Am. Chem. Soc, 1978,100,1300.515. J. P. Collman and W. R. Roper, Adv. Organomet. Chem., 1968, 7, 53.516. R. G. Pearson, Ace. Chem. Res., 1971, 4,152.517. M. F. Lappert and P. W. Lednor, Adv. Organomet. Chem., 1976,14, 345.518. B. Akermark, H. Johansen, B. Roos and U. Wahlgren, J. Am. Chem. Soc, 1979,101, 5876.519. S. Komiya, T. A. Albright, R. Hoffmann and J. K. Kochi, J. Am. Chem. Soc, 1976, 98, 7255.520. K. Tatsumi, R. Hoffmann, A. Yamamoto and J. K. Stille, Bull. Chem. Soc. Jpn., in press.521. A. Dedieu and A. Strich, Inorg. Chem., 1979,18, 2940.522. H. H. Jaffe and G. O. Doak, / . Chem. Phys., 1953, 21,196.523. J. Chatt and B. L. Shaw, J. Chem. Soc, 1959, 705.524. W. Mowat, A. Shortland, G. Yagupsky, N. J. Hill, M. Yagupsky and G. Wilkinson, J. Chem. Soc, Dalton

Trans., 1972, 533 and references therein.525. P. J. Davidson, M. F. Lappert and R. Pearce, Ace Chem. Res., 1974, 7, 209.526. J. A. Connor, Top. Curr. Chem., 1977, 71, 71.527. S. Evans, J. C. Green and S. E. Jackson, J. Chem. Soc, Faraday Trans. 2, 1973, 69, 191.528. M. F. Lappert, J. B. Pedley and G. Sharp, J. Organomet. Chem., 1974, 66, 271.529. L. Galyer, G. Wilkinson and D. R. Lloyd, / . Chem. Soc, Chem. Commun., 1975, 497.530. J. C. Green, D. R. Lloyd, L. Galyer, K. Mertis and G. Wilkinson, J. Chem. Soc, Dalton Trans., 1978,

1403.531. M. R. Churchill, Perspect. Struct. Chem., 1970, 3, 91.532. E. O. Fischer, Pure Appl. Chem., 1970, 24,407.533. D. J. Cardin, B. Cetinkaya and M. F. Lappert, Chem. Rev., 1972, 72, 545.534. O. S. Mills and A. D. Redhouse, / . Chem. Soc (A), 1969, 1274.535. T. F. Block and R. F. Fenske, J. Am. Chem. Soc, 1977, 99,4321.536. M. C. Bohm, J. Daub, R. Gleiter, P. Hofmann, M. F. Lappert and K. Ofele, Chem. Ber., 1980, 113,

3629.537. L. J. Guggenberger and R. R. Schrock, / . Am. Chem. Soc, 1975, 97, 6578.538. R. R. Schrock, Ace Chem. Res., 1979,12, 98.539. M. R. Churchill, F. J. Hollander and R. R. Schrock, J. Am. Chem. Soc, 1978,100,647.540. A. J. Schultz, J. M. Williams, R. R. Schrock, G. A. Rupprecht and J. D. Fellmann, J. Am. Chem. Soc, 1979,

101, 1593.541. M. R. Churchill and W. J. Youngs,«/. Chem. Soc, Chem. Commun., 1978, 1048.542. M. R. Churchill and W. J. Youngs, Inorg. Chem., 1979,18,1930.543. M. R. Churchill and W. J. Youngs, Inorg. Chem., 1979,18, 2454.544. R. J. Goddard, R. Hoffmann and E. D. Jemmis, / . Am. Chem. Soc, 1980,102, 7667.545. N. Calderon, J. P. Lawrence and E. A. Ofstead, Adv. Organomet. Chem., 1979,17,449.546. O. Eisenstein and R. Hoffmann, / . Am. Chem. Soc, 1981,103,000.547. R. J. McKinney, T. H. Tulip, D. L. Thorn, T. Coolbaugh and F. N. Tebbe, J. Am. Chem. Soc, 1981,103,

5584.548. A. K. Rappe and W. A. Goddard, III, / . Am. Chem. Soc, 1980,102, 5114.549. E. O. Fischer, Adv. Organomet. Chem., 1976,14,1.550. D. L. DuBois and R. Hoffmann, Nouv. J. Chim., 1977,1,479.551. D. M. P. Mingos, unpublished results.

Copyright © 1982 Comprehensive Organometallic Chemistry Elsevier Ltd.

Comprehensive Organometallic Chemistry || Bonding of Unsaturated Organic Molecules to Transition...

Documents

Transcript of Comprehensive Organometallic Chemistry || Bonding of Unsaturated Organic Molecules to Transition...