Dynamic and Static Stereochemistry in Dirnolybdenum and Ditungsten Compounds Containing a Central (M=M)“+Unit BY MALCOLM H. CHISHOLM Department of Chemistry Indiana University Bloomington Indiana 47405 U.S.A. Received 4th September 1979 The chemistry of molybdenum and tungsten in oxidation state +3 is now dominated by compounds containing a central (M=M)6+ core with metal-to-metal distances in the range 2.2-2.3 A. The molecular orbital configuration of the triple bond is dn4arising from the mutual interaction of metal atomic d,z (0)and d,, d,,(n) orbitals. In compounds containing the central (MEM)~+ core the metal atoms may be bonded to three four five or six ligand atoms. Examples of each are given and for a given coordination number the preferred geometry is discussed.Low temperature n.m.r. studies support the view that the structures found in the solid state are also present in solution. Variable temperature n.m.r. studies reveal a wealth of knowledge concerning the dynamic behaviour of these molecules in solution. For example (i) M2(NR2)6 and Mz(NR2)4Yz molecules (Y = halide alkyl or alkoxy group) are molecular propellers; (ii) rotational barriers about metal-to-metal triple bonds in MzY4Xz(M=M) compounds are comparable to those in related Si2Y4X2 and P2Y4 com-pounds; the latter contain a central element-element 0-bond of length 2.2-2.3 A; (iii) molecules of the type L(R0)3M=M(OR)3L’ contain OR groups which are cis and trails to the neutral ligands L and L’. Site exchange between cis and ti’aans groups occurs on the n.m.r.time-scale by a threshold mechanism which does not involve exchange of OR groups between the two metal atoms. This is compared to the fluxional properties associated with MX5 compounds (D3,,+D4J. Finally the absence of any complex containing a central M4I2+ tetrahedral arrangement is noted. The closest approach to such a compound is seen in the tetranuclear Complexes Mo~(,x-F)~(OBU~)~ and Mo~(,LL- F)3(p-NMe2)(OB~t)8 which contain a bisphenoid of molybdenum atoms having two short Mo-Mo distances (2.26 A) and four long Mo-Mo distances (3.75 A) corresponding to localized triple and non-bonding interactions respectively. “ There are literally thousands of chromium(Ir1) complexes which with very few exceptions are all hexaco~rdinate.’’~ This is not surprising in view of the fact that ligand field stabilization favours an octahedral geometry for a d3 ion.Since ligand field stabilization energies increase sizeably in going from the first to the second row and again from the second to the third row within a triad of transition metals one might have anticipated an abundance of molybdenum(Ir1) and tungsten(II1) hexaco- ordinate complexes. Rather interestingly the reverse trend is observed there is but a handful of well authenticated mononuclear molybdenum(Ir1) complexes and to this author’s knowledge not one mononuclear complex is known for tungsten(Ir1). Does this mean that the argument based on ligand field stabilization is fallacious? Certainly not. If we consider the d6 configuration for example we find an abun- dance of six-coordinate octahedral 2nd and 3rd row transition metal complexes e.g.IP PtIV complexes. The main difference between the d3 and d6 systems rests with the former being paramagnetic and the latter diamagnetic and since within any triad of metals for given oxidation state Z+,the effective nuclear charge exerted on the valence electrons decreases down the series then the d3 orbitals in an octahedral en- vironment t& become more diffuse and available for metal-metal bonding. This is M. H. CHISHOLM well demonstrated in the structures of the M2C1;-ions which share a common con- facial bioctahedral geometry D3h,and where the M-M distances change from 3.12 A (M = Cr) to 2.67 (M = Mo) to 2.41 (M = W).2'3 Indeed W2C1$- is diamagnetic and Pauling4 introduced the canonical structures W=W and WgW as part of a resonance hybrid description for the anion.However the types of compounds described in this accowt are of a simpler nature in that there are no atoms directly bridging the two metals. SYNTHESES Though this account does not dwell on syntheses or reactivity patterns of the com- pounds to be described it is worth noting that the basic M& compounds [M = Mo or W X = R(P-elimination stabilized alkyl) or NMe,] are derived from metathetic reactions involving either MoCl, MoCl, wC16 or WC14. Though the details of these reactions are now known we have presented a strong case for the fact that the M2X6 compounds are not formed by the coupling of two reactive mononuclear species.6 The most synthetically useful compounds are the dimethylamido Compounds M,(NMe,),,7y8 from which literally scores of dinuclear compounds con- taining the central (MEM)~ have been prepared.Some of these reactions are sum- + marized in scheme 1. where R =Me CHzCMe3 M -Mo R = Bur PrF or Mo2( OZCNMez)4 (m) + 1-alkene f alkane R = Et i-Pr n-Bu MO~(OR)~L~;= an amine L SOLID-STATE STRUCTURES M2X6COMPOUNDS (x = R NR2AND OR) All these compounds have a central staggered ethane-like M2C6 M2N6or M206 group having virtual D3d~ymmetry.~Two views of the Mo,(NMe,) molecule are shown in fig. 1. The M-NC planes are aligned with the Mo-Mo-N planes thus maintaining Dfdsymmetry and giving rise to six proximal N-methyl groups those lying over the metal-metal bond and six distal N-methyl groups.There is also a large class of M2X,(NR2)4compounds e.,q.,X = halide alkyl or alkoxide group.5* These may be viewed as 1 ,Zdisubstituted ethane-li ke molecules. The halides all crystallize in the anti-rotamer and maintain this in hydrocarbon solutions.lO?ll Fig. 2 shows the * The compounds where X = OR(R = But Pr' Et and Me) have only recently been made and behave in solution like the corresponding alkyls X = CH,R.9 196 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (MEM)6C lb) 2C(13) N(121 lC(12) wbZC(12) FIG.1 .-Two ORTEP views of the Mo,(NM~,)~ molecule (a) almost perpendicular to the Mo-Mo axis; (b)almost along the Mo-Mo axis. In this and in all other structural figures thermal ellipsoids are drawn at the 50% probability level.molecular structure of W2C12(NEt2)4 lo again the N-ethyl groups are arranged in proximal and distal sets. The solid state structures of W2Me2(NEt2)412 and Mo2Me2- (NME2),13also show the anti-rotamer. M2X6L2 COMPOUNDS Mo~(OS~M~~)~(HNM~~)~'~ lS are two examples of nitrogen and W2(OPri)6(py)2 donor adducts to M2X6-type molecules. The geometry about each metal is essentially square planar and the two M03N units are joined in such a manner that they are partially staggered with respect to each other. A view down the W-W bond of the W206N2 skeleton of the W2(OPri),(py) molecule is shown in fig. 3. In the Mo~(OBU~),(O~COB~')~ molecule there are a pair of cis-02COBut ligands which bridge the MozMo bond.' This imposes a virtually eclipsed geometry on the Mo2040& skeleton.In the W2(NMe2)4(PhN3Ph)2 molecule the diphenyltriazenido group is bidentate but does not bridge the WzW bond." An ORTEP view of the molecule is shown in fig. 4 note that the molecule has a C2axis of symmetry. M. H. CHISHOLM FIG.2.-An ORTEP view of the WzC12(NEt2)4 molecule. Note the central W,C12N skeleton has virtual Czusymmetry. N1 02 FIG.3.-The Wz06N2 skeleton of the W2(OPri)6(py)2 molecule viewed down the metal-metal bond. 198 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (MZEM)6t FIG.4.-An ORTEP view of the W2(NMe2)4(PhN3Ph)2 molecule emphasizing the C2 axis of symmetry. Irrespective of the nature of the groups present we have found that when four atoms are coordinated to each metal in an (M=M)'j+ compound the four atoms lie at least roughly at the corners of a square plane.Typically the M-M distance is ~0.02 8 longer in these compounds than in the simple M2X6 compounds.* What about five atoms coordinated to each metal? Where will the fifth ligand position be? Well we only have one structurally characterized example so it is certainly premature to claim a general trend. The central skeleton of the W,(CH,) (02CNEt2)4 molecule is shown in fig. 5. There is a pair of bridging O,CNEt ligands which imposes an eclipsed geometry with respect to each end of the molecule.ls There is also a pair of bidentate O,CNEt ligands and the carbon atom of the methyl group makes up a pentagonal coordination for each tungsten.Although one cannot generalize from a single case the structure of W2(02CNMe,)6l9 is so closely related to that of W2(CH3)2(02CNEt2)4 that it is at least tempting to say that a pattern is beginning to emerge. The central W,(O,C) skeleton is shown in fig. 6. The relationship between the structure of W2(CH3)2(02CNEt2)4 and W2 (O,CNMe,) is most striking the methyl carbon is replaced by an oxygen atom of an axially aligned O,CNMe group. The other oxygen forms a weak/long bond in the axial position. 'Thus it appears that the central (M=M)6+ unit upon expanding the coordination number of each metal from 3 to 4 to 5 goes from trigonal to square planar to penta- gonal planar and only reluctantly will accept a sixth ligand atom in the axial position (axial with respect to the M-M bond).REMARKS ON BONDING In all the compounds a simple analysis of the symmetry types of orbitals required to form M-M and M-L bonds and a consideration of the symmetry properties of * See ref. (56) for a tabular listing of M-M distances containing the central (MEM)~+group. M. H. CHISHOLM FIG.5.-The central W2C2(02C)4 molecule emphasizing that the skeleton of the W2(Me)2(02CNEt2)4 molecule has virtual Czvsymmetry. 09 FIG.6.-The central wz(Ozc)6 skeleton of the W2(02CNMe2)6 molecule emphasizing that the molecule has virtual Czvsymmetry. the metal valence shell orbitals leads to a satisfactory formulation of electronic structure. We may assume that the M=M bond is formed primarily by overlap of metal dZ2 orbitals to give the component and metal dxzand dyzorbitals to give the rc components.This is in accord with the assumption originally made and subse- quently supported by SCF Xcc calculation^^^ for the quadruple bond in Re,Cl;- and Mo,Cli-. Furtheriiiore the detailed electronic structure of Mo,X compounds (X = 200 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (M=M)6' R NMe and OH) was the subject of a recent SCF Xcc SW calculation and here the calculated and observed p.e. spectra were in good agreement.20 Then in M2X6 molecules the metal may use sp2or sd2hybrids to form the three tri- gonal bonds. When X = NR2 or OR ligand to metal n-bonding may also occur to two of the metal orbitals not used in a-bonding.The maximum M-N bond order is therefore 13 in M2(NMe,) compounds and each metal attains a valence shell of 16 electrons. 798 In M2X4L2molecules the four planar bonds may use s px,py d,z_,z-hybrids. In W,(Me),(O,CNEt,) the five quasi-planar bonds may use tungsten s px,py,dx2-y2 and dxyatomic orbitals and in W2(0,CNMe2)6 the additional use of the tungsten pz orbital may be employed to form the weak axial W-0 bond (2.67 A). Such a qualitative picture may be viewed as satisfactory to the extent that it readily accounts for the observed diamagnetic nature of the compounds and the short nature of the Mo-Mo distances which are only ~0.1 8 longer than those found in compounds containing MoZMo bonds. * Furthermore all the compounds are yel- low or orange resulting from a tailing into the visible of higher energy (u.v.) charge transfer bands.Lastly it should be noted that a triple bond consisting of a (T component and two equivalent n-components has cylindrical symmetry and imposes no restriction upon geometry [cf. Re,Cli-where the 6 component of the M-M bond imposes an eclipsed geometry of the two ReCI; units]. The observed geometries for all of the afore- mentioned compounds appear to be totally dominated by the steric requirements of the ligands. All the M2X6 and M2X6L2 compounds adopt staggered geometries because steric repulsive interactions dominate. Only in Mo,(OBu'),(02COBut)2 W2(Me)2(02CNEt2)4 and W2(02CNMe2)6 which contain bridging OCO groups are the geometries eclipsed. DYNAMICAL SOLUTION BEHAVIOUR Since all the compounds are diamagnetic their dynamical solution behaviour is readily investigated by variable temperature n.m.r.spectroscopy. The dialkylamido compounds reveal the expected but rarely before observed diamagnetic anisotropy associated with a triple bond. Variable temperature n.m.r. studies reveal that these molecules are " cheerleader " molecules they whirl as they twirl.? Detailed descriptions concerning the rotations that occur around the M-N bonds and the M=M bonds have been presented elsewhere as has the assignment of proximal and distal resonance^.^^ It is sufficient here to exemplify the phenomenon. Fig. 7 shows the high temperature and low temperature limiting 'H n.m.r. spec- trum of W2C12(NEt2) in [2H,]tol~iene. At high temperatures >130 "C proximal + distal ethyl exchange is rapid on the n.m.r.time-scale while at low temperatures <-16 "C,proximal and distal resonances are frozen out. Three further points are noteworthy. (1) The high-temperature limiting spectrum corresponds to an ABX3 spectrum and the low-temperature limiting spectrum to two ABX3 spectra. Evi-dently the mechanism of proximal $ distal exchange does not remove the diastereo- topic nature of the methylene protons. (2) The spectra correspond to the presence of only the anti rotamer in solution. This is the rotamer found in the solid state; * See ref. (56) for a recent tabular listing of M-M distances in compounds containing a central MEM bond. 7 At Indiana University they cheer " Go Big Red ".M. H. CHISHOLM 201 ~~ 6.O 5.0 4.0 3.0 2.0 1.o 0.0 8 / p.p.m. FIG.7.-(a) High (150 "C) and (b) low (-18 "C) temperature limiting 'H n.m.r. spectra of anti-W2C1,(NEt,) obtained in [2H8]toluene at 100 MHz. see fig. 2. (3) There is a large chemical shift separation between proximal and distal methylene proton resonances z2.5 p.p.m. The separation between proximal and distal methylene carbon resonances is much larger z 30 p.p.m. The variable temperature n.m.r. spectra of M2R2'(NR2)4compounds R' = Me Et i-Pr n-Bu CH,CMe and CH,SiMe, and R = Me and Et are more complex A because both anti and gauche rotamers exist in equilibria in ~o1~tion.~~*~~*~~ gauche M2R2'(NR2)4molecule has C2 symmetry and thus has two types of NR groups two are anti to R' and two are mutually anti.The low temperature limiting 'H n.m.r. spectrum for W,(CH,CMe,),(NMe,) in [2H,]toluene is shown in fig. 8. 202 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (MEM)6t i FIG.&-Low temperature limiting 'H n.m.r. spectrum of a mixture of anti andgauche W2(CH2CMe3)2 (NMe2)4 obtained at -65 "C and 270 MHz. Note the relative concentrations of gauche to anti rotamers are z 10:1 and that the methylene protons of the neopentyl ligand are an AB quartet in which the chemical shift separation of the Ha and Hb protons is very large. Note (1) The gauche-rotamer predominates. (2) The methylene protons of CH2CMe ligands are diastereotopic and form an AB pattern. (3) The chemical shift separation of the Haand Hbprotons is now exceedingly large [cf the N(CH2CH,) spectra shown in fig.71. The latter presumably reflects the fact that in the gauche rotamer the pair of bulky CMe groups impose a preferred conformation in which the methylene protons occupy sites which are quite different with respect to the M-M triple bonds' diamagnetic anisotropy. Since the M2R2(NRJ4 molecules appaiently prefer to crystallize in the anti- rotameric form it has been possible to measure the energy of activation for anti-to- gauche isomerization in these molecules. This is slow on the n.m.r. time-scale and can be followed by monitoring the approach to equilibrium EA falls in the range 20-24 kcal mol-l depending upon specific R and R' combination^.^" A point which now naturally arises is by what mechanism does anti +gauche isomerization occur a simple rotation about the M-M bond or by an intramolecular mechanism in which NR2 groups are transferred from one metal atom to the other by way of the formation of dialkylamido bridges cJ2 metal carbonyl site exchange in cluster metal carbonyls? This question is best answered by the examination of a molecule of the formula M2X5Y.Here the X groups naturally fall into three classes as shown below. y\ If rotation about the MEEMbondis frozen out on the n.m.r. time-scale then one should observe 3 different X signals. If rotation is fast on the n.m.r. time-scale then X(2) and X(3) become equivalent but remain distinct from X(1). Finally if exchange of X groups between the two metal atoms occurs rapidly then all X groups become equivalent.We have been able to synthesize molecules of this form. For example when M. €1. CHISHOLM Mo,(C,H,),(NMe,) is treated with tert-butanol in benzene the fascinating reaction shown below occurs 21* Examination of the IH n.m.r. spectrum of MO~(C~H,)(OBU*)~ at -65 "C and 270 MHz shows two types of OBu' groups in the integral ratio 2:3. This is consistent with the view that rotation about the W-W bond is still rapid on the n.m.r. time- scale and furthermore that alkoxy group exchange between metal atoms is slow. Further support for facile rotation about M=M is seen in the low-temperature 'H n.m.r. spectra of M2Me2(OB~t)4 Here we have not yet been able to freeze out anti + gauche isomerization on the n.m.r. time-scale even using high field spectrometers.We attribute the difference in EA to rotation about the M=M bond in compounds of the form M2R2'(NR2)4 and M2R2'(OBu*)4to the cogging effect of the NR2 groups in the former. The compounds M2(NR2)6 and M2X2(NR2)4 which we refer to as cheerleader molecules correspond stereochemically to 1,I ,2,2-tetra-aryl substituted ethanes and in solution behave as molecular propeller^.^^ When the blades are removed as in M,R,(OBU')~ and M,R(OBu') compounds then rotation about the M=M bond becomes much more facile EA < 7 kcal mol-I. Indeed the rotational barriers appear close'iy related to tetra-alkyl silanes R,HSi-SiHR and tetra-alkyl diphosphines R2P-PR2 which in solution also prefer the gauche conforma-tion.26 This comparison is all the more impressive when one recognizes that the Si-Si a-bond distance is ~2.3 A (Mo-Mo is 2.2 A) and the P-P a-distance is 2.2 A (W=W is 2.3 A).Thus we believe that our work has provided the first experimental demonstration that for a non-linear molecule containing a triple bond composed of one 0and two equivalent 7c components the rotational barrier is limited only by the steric factors associated with the substituents on the two elements which are united by the triple bond. Molecules of the type M2(OR)6L2 contain two types of OR groups on each metal atom namely those which are cis and trans with respect to the ligand L. In all cases which we have examined thus far the low temperature limiting n.m.r. spectra reveal two types of OR groups in the integral ratio 2 :1.Perhaps even more fascinating is our observation that the low temperature limiting 13Cn.m.r. spectrum of the W2(OPri),- (py) molecule shows three methyne carbon signals OCH(CH3), in the integral ratio 1 :1:1 which is what is expected for a W206N2 skeleton that has virtual C sym-metry namely the pyridine ligands are adjacent to each other as shown in fig. 3. Al-though in the crystal there are six distinct oxygen atoms it is easy to see that a slight twisting about the WrW bond brings about a time-averaged molecule with an apparent C axis of symmetry thereby making the oxygen atoms fall into three sets (01 OS) (03 07) and (02 06). It then follows that there should be three sets of me- thyne carbon atoms. At room temperature on the n.m.r.time-scale all M2(0R)& molecules show only one type of OR group. This is consistent with rapid (n.m.r. time-scale) cis + trans isomerization. Once again however one would like to answer the question " How is this achieved?" In order to probe such an intriguing matter one must design a * We note in ref. (22) that when the reaction is carried out using the labelled compound Mo2-(CH2CD3),(NMeJ4 with Bu'OH the ethane that is eliminated is exclusively CH,DCD,. The result- ing ethyl ligand is CZH3D2formed from Bu'OH + CD2 :CH2 and has a statistical distribution of deuterium atoms on the cc and p ethyl carbons. 204 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (MrM)6f molecule of the form M2(0R)6L1L2 in which L and L2are two different donor ligands then each end of the molecule is effectively labelled M(1) and M(2).One must also design a molecule where it is possible to show that L1and L2do not hop between the two metal atoms either by an intra- or inter-molecular mechanism. In this regard ligand dissociation must be ruled out otherwise M(l) and M(2) would become equivalent. We think we have been fortunate enough to obtain such a molecule. Crystallo-graphically we have shown that acetylene^,^^ allenes29 and dialkylaminocyanimides 29 add across the M=M bond in the Cp2M02(C0)430 compound in the manner shown in fig. 9. Now it so happens that Mo2(OR) compounds also react with all of the f ;N\ I Mo' FIG.9.-Schematic representations of the Cp,Mo2(C0)4(un) molecules where A un = RC ECR; B un = allene and C un = Me2NCN emphasizing the coordination of the central Mo,(un) group.above. Unfortunately no crystallographic data are at present available on the ad- ducts. Nevertheless if we make the assumption that in the Mo2(oPri),(NCNMe2) one molybdenum atom receives a lone pair of electrons from the terminal nitrogen atom while the other molybdenum atom receives a pair of electrons from the C=N ;It-bond and furthermore that this causes the NCNC2 unit to become planar with a fairly high energy barrier (n.m.r. time-scale) to rotation about the central C-N bond then we are home and dry. This may seem like too much to assume but the amazing fact is that the low-temperature 'H n.m.r. spectrum shown in fig. 10 is entirely consistent with these assumptions.There are four methyne proton resonances labelled A-D in fig. 10 respectively. The methyl region of the OPri ligands is more M. H. CHISHOLM complex and consists of three well separated doublets marked E-G and three over- lapping sets marked H in fig. 10. Six methyl resonances are indeed expected accord- ing to our assumption since (i) each molybdenum atom is labelled (ii) there are cis and trans OPr' ligands with respect to the NCNMe ligand and (iii) the methyl groups FIG.10.-Low-temperature limiting 'H n.m.r. spectrum of Mo2(0Pri)6(NCNMez) obtained at -45 "C 220 MHz in [2H,]toluene. The methyne proton resonances are indicated A B C and D and the methyl resonances E F G and H. The signals marked with an asterisk are due to Mo,( OPri)6 (M =M).of the cis-OPr' ligands are diastereotopic. There are also two signals of equal intensity for the N-methyl protons which is expected for a planar C,NCN group with restricted rotation about the central N-C bond. [This latter observation is directly analogous to the low-temperature limiting spectrum observed for Cp2M02(CO) (NCNMe,)].29 The only other resonances seen in the spectrum (fig. 10) are assignable to (i) small amounts of Mo,(OPr') which is present as an impurity and (ii) residual protons in the [2H,]toluene solvent. On raising the temperature the methyne proton resonances B and D start to broaden and then coalesce as do three of the methyl doublets namely the doublets indicated by F G and one from H in fig. 10. At this temperature site exchange of three of the OPr' ligands is fast while the other three are still frozen out on the n.m.r.time-scale. There are still two signals of equal intensity for the N-methyl protons which implies that the Mo,(NCNC,) unit is not fluxional. We believe the most reason- able interpretation of the dynamic behaviour of the molecule at 16 "Cis that alkoxy group exchange is occurring rapidly at one molybdenum atom but not at the other. Furthermore it is reasonable to suppose that the rapid site exchange involves the alkoxy groups which are coordinated to the least sterically crowded molybdenum atom namely the one which receives a nitrogen lone pair. On raising the temperature above 16 "C site exchange between the other set of OPr' ligands sets in and finally (>80 "C) all OPr' ligands become equivalent and the N-methyl resonances collapse to a single resonance.This is consistent with the view that the Mo,NCNC unit becomes fluxional in a manner which equilibrates both molybdenum atoms this was found for Cp,Mo,(CO),(NCNMe,). All these tem- perature dependent processes do not involve free Mo,(OPr-i), which is present in solution and are thus considered intramolecular processes. Indeed a plausible explanation for the above and indeed all the M2(0R),L2 compounds is that the ends of the molecules undergo facile square-based pyramidal + 206 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (M=M)6i-trigonal bipyramidal interconversions of the type well known for mononuclear ML5 complexes.32 The only major difference is that for the dinuclear compounds M=M there seems to be a marked preference for the square-based pyramid and thus the trigonal bipyramidal form is either a relatively unstable intermediate or a transition state for cis + trans isomerization shown below.M M M The W,(02CNMe,) molecule displays a particularly fascinating dynamical solution behaviour." Each O,CNC unit is planar and one can reasonably assume* that EA for rotation about the central C-N bond is zl6 kcal mol-l. Thus the carbon resonances of the methyl groups can be used to monitor the motions of the oxygen atoms. Furthermore by using 13CQ2 in the preparation of the compound (see scheme 1) one can readily monitor the three types of carboxylic carbon atoms.? Three distinct chemical processes can be detected which are in increasing order of energy of activation (1) the exchange of O(12) and O(11) sites.This is tantamount to an intramolecular substitution reaction in which an entering axially aligned ligand 0(12) substitutes one of the ligands in the pentagonal plane O(11); (2) Exchange of terminally bonded carbamate groups i.e. 13C resonances associated with C(5) and C(6) coalesce and (3) finally above room temperature exchange between bridging and terminally bonded carbamate ligands become fast on the n.m.r. time-scale. The dynamic behaviour of W2Me2(OzCNEt2)4 also parallels that of W2(0,CNMe,) in solution but here there are only two types of carbamato ligands :bridging and ter- minal. Below room temperature the solid-state structure (fig.5) is frozen out on the n.m.r. time-scale. However above 50 "C rapid exchange of bridging and terminal ligands occurs.19 In contrast to the carbamato complexes which readily exchange bridging and ter- minal groups the W2(NMe2)4(PhN3Ph)2 (see fig. 4) molecule appears relatively rigid in s01ution.l~ The molecule contains a C,axis of symmetry and may be considered as a member of the class of gauche-M,X,(NMe,) molecules but with X = the bi- dentate triazenido group. There are therefore two types of NMe groups and at -45 "C and 220 MHz the IH n.m.r. spectrum clearly shows four N-methyl reso- nances of equal intensity two proximal (downfield) and two distal (upfield). On raising the temperature there is a pair-wise collapse to give ultimately two lines at 100 "C which means that even though rotations about the M-N bonds become fast on the n.m.r.time-scale the C2axis of symmetry is rnaintained.l7 Enantiomerization involving either gauche to gauche or gauche to anti to gauche transformations does not occur rapidly on the n.m.r. time-scale. This could have been caused either by a simple rotation about the W=W bond or by the formation of an intermediate in which the two 1,3-diphenyltriazenido ligands bridged the M-M bond cJ the structure of NO,(QB~~)~(Q,COB~~),.~~ The difference in both the static and dynamic stereo- chemistry of the Mo2(OBut),(O,COBut) and W2(NMe2)4(PhN3Ph)2 molecules is once again determined by the steric demands of the ligands bonded to the central * Simple organic carbamate esters Me2NC(0)OR show a barrier to rotation about the C-N bond of 16 kcal mol-l 33 1-Note the molecule has virtual Czvsymmetry.M. H. CHISHOLM 207 (MEM)~+ unit. The M-NC units in M2X2(NMe2)4 compounds are effectively cogged in such a way that even though rotations about M-N bonds may be fast on the n.m.r. time-scale rotation about the M=M is hindered. This fact can be used to advantage in investigating the mechanisms of substitution reactions at these dinuclear centres. For example the observation that anti-W,Cl,(NEt,) reacts with LiCH,- SiMe (2 equiv.) in benzene to give anti-W,(CH,SiMe,),(NEt,) which then slowly isomerizes to a mixture of anti-and gau~he-W,(CH,SiMe,),(NEt,)~indicates that the alkyl-for-chloride ligand exchange must proceed with retention of stereochemistry at tungsten.34 Mi2+CLUSTERS For some time now we have been trying to establish that the dimerization of two (M=M)6+ containing compounds can lead to Mi2+ cluster compounds containing a central tetrahedral M4 unit.Indeed it seemed that for a given ligand X or combina- tion of X Y ligands there should be an equilibrium of the type shown below.6 As the steric bulk of an alkoxy ligand is reduced polynuclear [Mo(OR),], com- pounds are formed e.g. for R = Et and Me.35 These are diamagnetic which indi- cates the existence of metal-metal bonds but as yet no X-ray structural information is available. The closest approach to a tetrahedral Mi2+cluster was recently found in the reac- tion between Mo,(OBU')~ and PF (2 equiv.) which leads to a black compound of empirical formula Mo(F)(OBu'),.In one preparation of this compound crystals suitable for detailed X-ray work were obtained. The unit cell was found to contain one molecule of Mo4(~-F),(OBut),and two molecules of Mo~(,u-NM~~)(,u-F),(OBU~)~.~~ The latter compound was a total surprise to us and we attribute the presence of the di- methylamido ligand to incomplete alcoholysis in the preparation of the starting material Mo,(OBu') (see scheme 1). ORTEP views of the Mo~(,u-F)~(OB~') and Mo~(,u- NMe2)(~-F),(OBut) molecules are shown in fig. 11 and 12 respectively. In both molecules the Mo~ unit is a bisphenoid having two short Mo-Mo distances 2.26 8 (averaged) and four long Mo-Mo distances 3.75 A (averaged). Evidently a fluoride-for-tert-butoxide reaction induces a Lewis base association reaction by the formation of metal-ligand bridges.While substitution of the small and more electronegative fluoride ligand might well be expected to promote a Lewis base association reaction,35 the choice of bridging ligands which is established namely F NMe > OBu' is surprising to us. It seems as if the localized M=M units (2.26 A) are held apart by the fluoride bridges-though we have no way of knowing at this time whether these molecules are formed under kinetic or thermodynamic control. Finally however it should be noted that the geometries about the Mo2F404 and Mo2NF304 units (M=M) are virtually identical to the local Mo,O,O' skeleton of the Mo,(OB~~)~(O,COB~~), (MEM) molecule.There are two molybdenum atoms held together by a metal-to-metal triple bond (no bridging groups) and each molybdenum atom is coordinated to four ligands which lie roughly in a square plane. 208 STEREOCHEMISTRY IN COMPOUNDS CONTAINING (MEd't~f)~' Cl111') c (1 21') C(211') c(111') c ( 221') FIG.11.-The central skeleton of the Mo~(,u-F)~(OBU')~molecule. M. H. CHISHOLM C(211) "'1 d >' c (A11) I' FIG.12.-The central skeleton of the Mo4(p-NMez)(p-F),(0Bu')amolecule. For financial support of this work we thank the Research Corporation the donors of the Petroleum Research Fund administered by the American Chemical Society the National Science Foundation the Office of Naval Research the Marshal H. Wrubel Computing Center and the Tax Payers of the State of Indiana.This author is also grateful for all the talented co-authors referenced in this work and in particular to Prof. F. Albert Cotton who was instrumental in promoting this work uia collabora-tion during this author's term at Princeton University. F. A. Cotton and G. Wilkinson in Advanced Inorganic Chemistry (Interscience Publishers 3rd edn 1972) section 25-C-4 p. 830. F. A. Cotton Rev.Pure Appl. Chem. 1967 17 25 and references cited therein. R. Sailant and R. A. D. Wentworth Inorg. Chem. 1969 8 1226 and references cited therein. L. Pauling The Nature of the Chemical Bond (Cornell University Press 3rd edn 1960) p. 437. For recent reviews of chemistry associated with these compounds see (a)M. H.Chisholm and F. A. Cotton Accounts Chem. Res. 1978 11 356 and (b) M. H. Chisholm Transition Metal Chem. 1978 3 321. M. H. Chisholm M. W. Extine R. L. Kelly W. C. Mills C. A. Murillo L. A. Rankell and W. W. Reichert Inorg. Chem. 1978 17 1673. 210 STEREOCHEMISTRY IN COMPOUNDS CONTAlNlNG (MZdVf)6+ M = Mo M. H. Chisholm F. A. Cotton B. A. Frenz W. W. Reichert L. W. Shive and B. R. Stults J. Amer. Chem. Soc. 1976 98 4469. M = W M. H. Chisholm F. A. Cotton M. Extine and B. R. Stults J. Amer. Chem. Soc. 1976 98,4477. M. H. Chisholm and J. Garman results to be published. lo M. H. Chisholm F. A. Cotton M. W. Extine M. Millar and B. R. Stults J. Amer. Chem. SOC. 1976,98,4486. l1 M. H. Chisholm F. A. Cotton M. W. Extine M. Millar and B. R. Stults Inorg.Chem. 1977 16,320. l2 M. H. Chisholm F. A. Cotton M. W. Extine and B. R. Stults Inorg. Chem. 1976,15,2244. l3 M. H. Chisholm F. A. Cotton M. W. Extine and C. A. Murillo Inorg. Chem. 1978 17,2338. l4 M. H. Chisholm F. A. Cotton M. W. Extine and W. W. Reichert J. Amer. Chem. SOC., 1978 100,153. l5 M. Akiyama M. H. Chisholm F. A. Cotton M. W. Extine D. A. Haitko D. Little and P. E. Fanwick Inorg. Chem. 1979,18,2266. l6 M. H. Chisholm F. A. Cotton M. W. Extine and W. W. Reichert J. Amer. Chem. Soc. 1978 100,1727. l7 M. H.Chisholm J. C. Huffman and R. L. Kelly Inorg. Chem. 1979,18 3554. l8 M. H. Chisholm F. A. Cotton M. W. Extine and B. R. Stults Inorg. Chem. 1977 16,603. l9 F.A. Cotton Accounts @em. Res. 1978 11 225 and references cited therein.2o F. A. Cotton G. G. Stanley B,. J. Kalbacher J. C. Green E. Seddon and M. H. Chisholm Proc. Nat. Acad. Sci. 1977 74 3109. M. H. Chisholm D. A. Haitko and C. A. Murillo J. Ameu. Chem. Soc. 1978,100,6262. 22 M. H. Chisholm and D. A. Haitko J. Amer. Chem. Soc. 1979 101 6784. 23 R. D. Adams and F. A. Cotton in Dynamic Nuclear Magnetic Resonance Spectroscopy ed. L. M. Jackman and F. A. Cotton (Academic Press N.Y. 1975) p. 489. 24 M. H. Chisholm and D. A. Haitko results to be published. 25 K. Mislow Accounts Chem. Res. 1976 9 26. 26 S. G. Baxter D. A. Dougherty J. P. Hummel J. F. Blount and K. Mislow J. Amer. Chem. Soc. 1978 100 7795. 27 W. I. Bailey M. H. Chisholm F. A. Cotton and L. A. Rankel J. Amer. Chem. Soc. 1978,100 5764. 28 W. I. Bailey M.H. Chisholm F. A. Cotton C. A. Murillo and L. A. Rankel J. Amer. Chem. Soc. 1978 100 802. 29 M. H. Chisholm F. A. Cotton M. W. Extine and L. A. Rankel J.Amer. Chenz. Soc. 1978,100 807. 30 R. J. Klinger W. Butler and M. D. Curtis J.Amer. Chem. Soc. 1975,97,3535; 1978,100,5034. 31 M. H. Chisholm and R. L. Kelly Inorg. Chem. 1979,18,2321. 32 A. D. English S. D. Ittel C. A. Tolman P. Meakon and J. P. Jesson J. Amer. Chem. Soc. 1977 99 117 and references cited therein. 33 E. Lustig W. R. Benson and N. Duy J. Org. Chem. 1967 32 851. See also discussion in M. H. Chisholm and M. W. Extine J. Amer. Chem. Soc. 1977,99 782. 34 M. H. Chisholm and M. W. Extine J. Amer. Chem. Soc. 1976,98 6393. 35 M. H. Chisholm F. A. Cotton C. A. Murillo and W. W. Reichert Inorg.Chem. 1977,16,1801. 36 M. H. Chisholm J. C. Huffman and R. L. Kelly J. Amer. Chem. Soc. 1979,101 7100.