TRINETHYLSILYLMETHYL COMPLEXES

OF THE TRANSITIONAL METALS

A thesis submitted

by

WALTER MOWAT, B.Sc.,

for the degree of

Doctor of Philosophy

of the

University of London

Royal College of Science, Imperial College of Science and Technology,

London, S.W.7. October 1972. 2

ACKTIOWLEDGEMENTS

I would like to express my gratitude to Professor G. Wilkinson

F.R.S. for his guidance and encouragement during the supervision of this work and to the Science Research Council for support.

I would like to thank my friends and colleagues in the depart- ment for their valuable help and advice, in particular Dr. Guido

Yagupski, Dr. Matteo Giongo, Dr. Nick Hill, Tony Shortland and

Barbara Wozniak. Also Ray Shadwick, John Clay, Sue Johnson and the late Harold Smith for technical assistance.

Finally I would like to thank Brenda for her companionship over the last three years. 3

CONTENTS

Abstract. 4 Introduction. 5 Chapter I.

Trimethylsilylmethyl complexes. 13

Chapter II.

Neopentyl complexes. 29

Chapter III.

Neopentyl and related alkyls of chromium(IV). 36 Conclusion. 53 Experimental. 55

Abbreviations. 66

References. 67

• 4

ABSTRACT

The factors influencing the stability of the transition metal-carbon bond are discussed. The importance of kinetic stability is emphasised, with particular reference to the 0-elimination/hydride transfer reaction.

Thermally stable binary trimethylsilylmethyl (R) complexes of vanadium, VR4 and VOR3, niobium and tantalum, (11-R)2M2R4, and molybdenum and tungsten, M 2R6, have been prepared by the interaction of the respective metal halide with the Grignard or lithium reagent derived from chloromethyltrimethylsilane, and have been characterised by their spectroscopic properties.

Similar neopentyl (R') complexes have been isolated for titanium and zirconium, MR'S , tantalum, R13 TaC12, and molybdenum,

Mo 2RI 6•

Binary chromium(IV) alkyls, CrR4 (R = neopentyl, neophyl, tritylmethyl and methyl) have been prepared and characterised in the same way. 5

INTRODUCTION

The first stable, isolable compound with a transition metal

to carbon o-bond was the tetrameric trimethylplatinum iodide

[(CH3)3PtI]4, preDa'red by Pope and Peachy in 1907.1 Prior to this,

there had been many attempts to establish carbon linkages with 2 transition metals, e.g. Ti, Zr, Fe and the ready availability of

Grignard and lithium reagents as alkylating agents increased the

scope of potential alkyls but the results were disappointing. The

reason for this lack of experimental success was attributed3 to the

"impossibility of the existence of this type of compound". This 2 idea became firmly established and Cotton states that transition

metal alkyls and aryls are "very much less stable and accessible

than those of non transition elements". Bearing all this in mind one

can realise why there was little useful activity in this field during

the first half of the century. \ During the early 1950's, a few series

of c-bonded alkyl and aryl compounds in which other ligands were

present began to emerge. These were compounds of the type 4 C5H5)2Ti(C6H5 )2, h5- C5H5Fe(C0)2 CH3,5 CH3Mn(C0)56 etc. When

these compounds were shown not only to be readily isolable but in most cases to have remarkably high stability, it was assumed that the phosphines, carbonyls or other ligands present were stabilizing the complexes and that the absence of such ligands would imply loss of stability of the metal-carbon bond. Parshall and Mrowca sum up the feeling of the time "in contrast to simple alkyls some metal

complexes bearing other ligands in addition to alkyl or aryl groups are strikingly stable".

The first report of an isolable binary alkyl is an unsub- 2a stantiated report of trimethylrhenium, but the first proven alkyl 8 was tetramethyltitanium, prepared by Berthold and Groh in 1963. 6

TiMe!, is an unstable, yellow crystalline compound decomposing at

temperatures greater than -70° and this behaviour seemed to conform

with the prevailing theories of alkyl instability. Other binary

alkyls are the detonatively explosive dimethylmanganese,9 bis(trity1)- 10 11 nickel and a few other titanium alkyls, Ti(CH2Ph)4, Ti(Ph)412 12 and Ti(Ph)2. Alkyl complexes which may be classified alongside 13 these include alkyl metal halides e.g. MenTC14-n and Me3 MC12

(M = Nb and Ta),14 ionic salts, e.g. Li41/Ph6,15 Li2NbPh7,16 17 18 Li3CrMe6 and Li2 WPh6 and etherates, e.g.- R3Cr(THF)3

(R = Me,Ph etc)19 and TiPh2(Et20).20

The early conceptions of intrinsic thermal instability of

the transition metal-carbon bond and subsequent "Tr-stabilisation"

by donor ligands can be shown to have little foundation as follows.

The available thermodynamic data on transition metal-carbon bond

energies (Table 1) are hardly extensive but certainly seem to be

TABLE 1. Bond Energies of Transition Metal-Carbon a-bonds.

_1 Bond Compound Energy in KJmol Ref.

Pt - C6H5 (Et3P)2Pt(C6H5 )2 250 21

Pt - CH3 (h57C5N5 )Pt(cH3)3 164 22

Ti - CH3 (h57C5H5)Ti(CH3)2 250 23

Ti - C6H5 (h5-05H5)2 Ti(C6H5)2 350 23

Ti - C2H5 Ti(C2H5 )3C1 130 24

insufficient to support the view that the bond is weaker than those

' between carbon and non-transition metals, or between transition metals

and other first row elements e.g. oxygen and nitrogen in the alkoxides

and dialkylamides respectively, where the isolable compounds are 7

25 numerous. There is additional evidence in the form of force constant data derived from vibrational spectra. The carbon-fluorine force constants of CF3I and CF3Mn(C0)5 have been compared26 and the inference was that the CF3-Mn bond is strong. Similar comparisons between CH3TiC13 and the tin and silicon analogues27 show the titanium value to be of the same order of magnitude as the others.

Finally, the value of 2.28 mdyne A found for the metal carbon bond in TiMe4 is only 20% lower than a Si-Pb group tetramethyl of the same mass.28

The theory of "donor ligand stabilization" of the transition metal-carbon bond can be dismissed when one takes into consideration the homogeneous hydroformylation catalysts RhCl(PPh3)3 and

RhH(C0)(Pa3)3. 29The activity of these complexes is in fact completely dependent on the lability of the metal-carbon bond, and thus we can see that the presence of donor ligands provides no guarantee of stability.

It should also be noted that some substitution-inert octahedral 30 metal complexes have particularly stable transition metal-carbon 2+ 2+ bonds, e.g. Cr(H20)5 Me and Ph(NH 3)5C2H5 . Other substitution inert complexes where coordination sites are completely occupied include 31 adducts of alkyl complexes, e.g. TiMe4L2 and TiMeC13L2 are considerably more stable than the parent alkyls, and chelate complexes, where chelation not only blocks the coordination site but also contributes to the resistance of the bond to homolytic fission e.g. tris[w-dimethylarsino-o-tolyl]chromium(III).52 One can also compare coordinative saturation in binary alkyls. WMe633 is more 8 stable and less reactive than TiMe4 or CrMe4,34,35 and Cr(1-nor)4 is completely air and moisture stable, but Cr(4-cam)3 decomposes instantaneously in either.36 It can be seen that "stabilizing" 8

ligands in compounds such as (h5- CO5)Fe(C0)2R etc. are occupying

coordination sites, thus these compounds may also be considered

substitution inert.

It is clear from the above arguments that we can consider two

distinct types of stability, namely thermodynamic and kinetic. At

this stage it is worthwhile discussing the differences between them

as current theories37-39 propose that it is largely on kinetic grounds

that transition metal alkyls are stable or unstable. Cross and

Braterman39 have attempted to go into this in considerable depth, but

it is the view of the author that due to the lack of physical data

available for the alkyls, particularly the more recent ones, their

conclusions cannot be considered any more significant than those of

Wilkinson.37,38

Thermodynamic stability is concerned with the strength of the

metal-carbon bond relative to that of the alkyl decomposition products,

i.e. M-C vs C-C. Figure 1 shows a thermodynamically unstable alkyl,

and it can be seen that for the complex, represented as M-R, the

energy of its final decomposition products is lower, resulting in a

positive free energy of decomposition. Thus if the activation energy

I

Fd

1

ING

STOLE IDEMPIPOSKitir4 PROOkeT5 Figure 1

•Ea for the transition state can be overcome, it is far more likely

that the reactive intermediates will break up further to the final

decomposition products rather than recombine to give the original 9 complex. As the end products are unreactive (alkanes, alkenes, metal etc.) the reaction is irreversible. Thus to make the compound stable, one must consider ways of increasing the activation energy, the quantity which determines the kinetic stability. The way to do this is to block the decomposition pathways.

There are several mechanistic paths available for decomposition.

Zeiss and co-workers44°-42 have shown that the principal organic decomposition products of the metal alkyl are the respective alkane,

1-alkene and 2-alkene. The alkenes arise mainly through the 13 elimination reaction presumably via a 4-membered transition state

(equn 1), but Zeiss says that some a elimination occurs.

M-CH2CH2R M-H + CH2 = CHR Equn. 1 H---:CH2

The alkane is formed by subsequent interaction of the metal hydride with an alkyl group. The presence of the hydride was demonstrated by using 0-deuterated alkyls and adding deuterium to a scavenger olefin present before the decomposition began. It is agreed41'43 that radical or ionic mechanism (see equn 2 and 3) form a relatively minor part of the decomposition, but the presence of small amounts of cyclic or coupled alkanes shows that they do occur (equn 2).

M-R M.+ R. M+ R-R Equn 2

M-R 41' M++ R ' products Equn 3

The order of alkyl stabilities is normal>secondary>tertiary and Zeiss tentatively suggests that this is due to inductive effects of the alkyl group on the electron density about the metal centre, but it is equally probable that it is the number of available s-hydrogen atoms.

The kinetic arguments for the assumed homolytic fission of alkyl 10

groups in square planar Ni, Pd or Pt complexes has been considered 44 fully by Chatt and Shaw. Other proven decomposition routes include that for RCrC12,(TBF)n,41 which is totally homolytic, with the initial product being a dichlorochromium(II) species, and that of (h5 - C5H5)2TiPh2, which has an internal hydride transfer with elimination of a molecule of benzene.45

The hydride transfer/olefin elimination reaction is the main decomposition route, and thus if it can be blocked, it would be expected that the alkyl would enjoy greater stability. Proof of this has been provided by the studies of Tamura and Kochi.46

Several series of binary transition metal alkyls were prepared and their decompositions showed the following stability order for the alkyl groups:- methylft.dbenzyl..oneopentyl > >n-propyl.fton-butyl>ethyl>t-butyl>i-propyl.

The most stable are clearly those which cannot readily eliminate alkenes. To block the elimination reaction, the ligand must be chosen such that it satisfies at least one of the following conditions:

i) no available hydrogens on the n-atom.

ii) the 0-atom is one which does not form multiple bonds with carbon.

iii) the ligand geometry is such that an olefin cannot be formed.

It may be noted that almost all of the transition metal binary alkyls prepared since the beginning of this work are elimination stabilized.

These are presented in Table 2.

The most impressive contribution has been the series of 36 1-norbornyl and related alkyls prepared by Tennent and Bower.

In this case, the fused alicyclic ring system is inert to 0- elimination, as in iii) above. 11

TABLE 2. Recent binary transition metal alkyls.

Ti(nor)4 (36) V(nor)'4 (36) Cr(nor)4 (36) Mn(nor)4 (36 Ti(C6 F5 )4 (47) V(benzy1)4* (52) Cr(1-cam)4 (36) .Ti(CH2SiMe3)4* (48) Cr(4-cam)3 (36) Fe(nor)4 (3 Ti(C6H5)3* (49) Cr(t-buty1)4 (35) Ti(CH3)3* (49) Cr(benzyl)4* (34) Co(nor)4 Cr(CH2SiMe3)4 (37) Zr(nor)4 (36) Nb(C6F5)4 (53) Mo2(benzyl)6 (54) Ni [(Bio C2H102 ]22 ( 55 Zr(benzy1)4 (50) Zr(CH2SiMe3)4* (48)

Hf(nor)4 (36) W2(benzyl)6 (54) Hf(benzY1)4 (51) W(CH3)6 (33) W(C6F5)5 (53)

in solution only

nor = 1-norbornyl

1-cam = 2,5,3-trimethylbicyclo[2.2.1]hept-l-y1

4—cam = 2,2,3-trimethylbicyclo[2.2.1]hept-1-y1

There are several preparative routes for metal alkyl complexes. The only generally applicable method for binary alkyls is

the interaction of the binary metal halide with a strong alkylating

agent i.e. Grignard or lithium reagent (equn 4).

MCln nRMgX -4 MRn nMgXCl Equn1+

As will be seen later, the reaction is seldom as simple as equn1+

implies. Less powerful alkylating agents (ZnR2, AiR3, HgR 2 etc.)

lead to partial substitution to give alkyl metal halides. The more

important other methods are as follows:- 12

i) Transition metal complex anions + alkyl halide.

e.g. Mn(C0)5- + RX RMn(C0)5 + X

ii) Oxidative addition. Particularly common method. MeI e.g. Pt(PPh 3)3M19 --)MePtI(PPh 3)2 Me2PtI2(PPh3)2 (57)

Ni(cop), Ph3ccp h3,(ph 3c)2Ni 2COD (58)

iii) Ring expansion. This may be considered as a special

case of oxidative addition, and is well known for

fluoroalkenes.

C e.g. Ni(PPh3)4 ------)(PPh3)2Ni-CF2CP2CF2CF22 (59)

iv) Insertion reactions. Insertion of olefin into

transition metal hydrides is the basis for

homogeneous hydrogenation, hydroformylation etc.

The many other methods are generally restricted to a few complexes. 17;

CHAPTER I

TRIMETHYLSILYLMETHYL COMPLEXES.

The first group of compounds prepared used trimethylsilyl- methyl (R) as ligand. This was for preparative convenience, and also because the ligand satisfies the first two requirements above for blocking the (3-elimination reaction, i.e. it has no 13-hydrogens and silicon cannot form a double bond with carbon. Trimethylsilyl- methyl complexes with "stabilizing" ligands have been reported and appear to be more stable than the corresponding methyls, 48 neopentyls etc. Lappert et a1. have worked on the binary titanium(IV) and zirconium(IV) systems but have been unable to isolate halogen- free crystalline compounds. Similar work was carried out by the author and co-workers with the same results. In view of the 61 relative ease of preparation of the neopentyl analogues, it iis hard to understand why these compounds are not isolable, but the main difficulty Would seem to arise from trace amounts of (Me3SiCH2)2 , the coupling product from the Grignard or lithium reaction. This • compound is a relatively involatile oil and inhibits crystallisation.

The low-melting chromium and vanadium alkyls were rather difficult to crystallise from the initial petroleum solution, but the hardest compounds to crystallise were the dimeric niobium and tantalum alkyls, which, in some cases, required several weeks at -45° to initiate. The crude crystalline material, once filtered and washed, was readily recrystallised from organic solvents. It was possible to eliminate the coupled product by subliming the lithium reagent but this was only carried out for some catalytic runs using the g.l.c.

The compounds were prepared by conventional procedures by the interaction of the metal halide with the lithium or Grignard 14 reagent. The isolable compounds and some of their properties are listed in Table I.1.

Tetrakis(trimethylsilylmethyl)vanadium(IV) and oxotris(trimethylsilyl- methyl)vanadium(V)

From the interaction of the lithium or Grignard reagents With

VC14, the compound V(CH 2SiMe3)4 is obtained. This is similar to

Cr(CH2SiMe3)4 being very volatile, sublimable in vacuum at room 37 temperature, and very soluble in petroleum, benzene etc. It is very air-sensitive, inflaming in air, although under nitrogen at -30° it is stable indefinitely. Also, unlike the chromium compound, it is decomposed by aqueous solutions, alcohols and chlorinated solvents.

However, it does not react under mild conditions with CO, CS2, 502,. alkenes, acetylenes, tertiary phosphines but does react with NO, tetra- cyanoethylene, hexafluorobut-2-yne, dry ammonia and alkyl amines.

The intense visible spectrum of V(CH2SiMe3)4 is consistent with a non-centro symmetric structure and the spectrum compares with those found for similar compounds,62 i.e., V(0Bu)4, 10,900, 13,900 cm and V(NR2)4 , 13,300, 17,600 cm-1 , but with a very strong field, as is the case for the chromium alkyl.

On passing a petroleum solution of V(CH2SiMe3)4 through a cellulose column, the colour changes to pale yellow and from this solution needles of oxotris(trimethylsilylmethyl)vanadium(V) are obtained. The compound can be made directly from V0C13 and a deficiency of the Grignard reagent; if excess Grignard is used, reduction to V(CH2SiMe3)4 occurs. The oxo compound is relatively air stable, sublimable in vacuum at 70°, but is somewhat light sensitive. It is unaffected by water and alcohols and is readily soluble i.n petroleum, benzene and ether. The compound does not react under mild conditions with CO, tertiary phosphines or primary amines. 15

TABLE 1.1 Trimethylsilylmethyl (R) Compounds of Transition Metals.

.. _ . -.... Electronic spectrum Colour form m.p.°C Compound (cm-1 , c

VR4 dark green needles 143 15,625, 23,600 (310) and weak broad band in far i.r. VOR3 lemon yellow needles 75 Rising absorption in u.v. weak shoulder at 26,600.

(ti,R 1)2 N12R4 a) red prisms 152 (dee) Rising absorption in u.v.

(µ-R1)2Ta2R4 a) orange prisms 170 (dec) Rising absorption in u.v.

M02R6 yellow plates 99 Rising absorption. in u.v.

W2R6 orange-brown plates 110 ditto with weak band at 21,300.

a) 11-R1 C-SiMe3 16

The n.m.r. spectrum (Table 1.2) has a sharp methyl resonance but the methylene peak is ca 150 cycles wide evidently due to coupling with 51 V(spin 7/2) which should produce eight lines, although these are not resolved, possibly due to interaction with the nuclear quadrupole or to exchange broadening resulting from intramolecular fluxional behaviour.

The i.r. spectra of all the trimethylsilylmethyl complexes are dominated by ligand vibrations, and these have been assigned.37

The metal-carbon stretches are those of most significance to this _1 work and for VR4, the two bands at 504 and 430 cm have been interpreted37 as those required for a tetrahedron distorted towards a C4v square pyramidal structure, with the unpaired electron occupying the apical position. For VOR3, the two bands at 522 and 468 cm are in accord with a regular C v symmetry. The V = 0 stretch.occurs at

988 cm-1

Vanadium (IV) is a d' system and VR4 is paramagnetic with one -3 unpaired electron. Solutions of this compound less than 10 molar in petroleum show a typical vanadium(IV) spectrum of eight lines, with a g-factor of 1.968 + 0.001 andlwperfine coupling of 50 ± 1 x 10cm.

On cooling, the lines broaden until at 100-120°K a rigid matrix spectrum appears [Figure I.1]. We may measure g et = 1.974 ± 0.001 _4 _1 and A ll =. 111 ± 1 x 10 cm , but it is impossible to obtain the remaining parameters gy and Al. Further cooling leads to a new spectrum, which is rather poorly resolved [Figure 1.2].

The magnetic moment has been measured by n.m.r. line shifts in benzene solution at 308°K and was 1.55 BM, while in the solid a value of 1.30 BM was found using a Gouy-Rankine balance. Such low 63 values are not unusualy for vanadium(IV) compounds and presumably indicate some form of inter-molecular spin coupling. 17

TABLE 1.2 Nuclear magnetic resonance spectra of trimethyl-

silylmethyl (R), Compounds (T values in deuterobenzene)

4101Wna=Gm.2,a17.=WX!Jha..,2r=_,C,,.

CH 3 CH 2 Remarks

Centre 7.5b CH3:CH2 = 9:2

8.50 I p,-CH3 : CH3 : CH2 = 9:18: 4

9.00 p—CH 3: CH 3: CH2 = 9:18:4

7.85 CH3 :CH2 = 9: 2

8.10 CH3:CH 2 = 9:2

J(1 83 W-CH2 ) = 10Hz 4

Figure 1.1. E.s.r. spectrum of VR4. (105 K)

Figure 1.2. •E.s.r. spectrum of VIZ. (95°K)

1,0mT. 19

Bis-p-(trimethylsilylmethylidyne)tetrakis(trimethylsilylmethyl)- diniobium(V) and ditantalum(V)

The interaction of niobium or tantalum pentachloride with the

Grignard reagent gives distinct colour changes on addition of successive equivalents of ligand. The final stage in the reaction is a dimeric species with a carbene bridge, the crystal structure of which has 64 been determined for the biobium complex. The compounds have a quasi-aromatic ring of which the carbene bridges form part (Figure Sitle3 1.3).

R.

They are best prepared by addition of metal halide in ether to excess Grignard reagent, and after extraction into petroleum the dark red (niobium) or orange (tantalum) crude products may be readily recrystallised. Both compounds have a phase change at 72° when the crystals crack, but are unaffected thereafter up to their melting points of 152° and 170° respectively, when they both decompose.

Both the biobium and tantalum compounds are air and moisture sensitive, particularly the latter. Exposure to air rapidly causes formation of a white coating of the hydrated oxide, and occasionally they inflame spontaneously. They are readily soluble in ether, petroleum, and benzene, but decompose slowly in chlorinated solvents and rapidly in protic solvents. They do not chromatograph on alumina, silica or cellulose and fail to sublime in vacuo.

Other than towards oxygen and moisture, the compounds are 20 surprisingly inert. Dry oxygen reacts with a petroleum solution of either to yield yellow, petroleum-soluble oils containing a mixture of species. Dry chlorine and hydrogen chloride gas both react immediately with either compound and the reaction of chlorine with the niobium compound was followed by n.m.r. The reaction was not particularly clean, but the major organic products were chloromethyl- trimethylsilane and trimethylchlorosilane, with several other minor chloroalkylsilane derivatives. Reactants having no affect on either compound include nitric oxide, carbon monoxide, tertiary alkyl and aryl phosphines, amines and alkenes.

The mechanism of the formation of the bridging carbene is unknown. It is assumed that a penta-alkyl complex cannot exist for steric reasons, and since the trisalkylmetaldichloride is known (see below) it must be at the final stages of chlorine elimination that reduction of an a-carbon atom occurs. The presence of the p-silicon atom certainly contributes to the stability of the carbene, as the analogous carbon compound, i.e. the neopentyl, cannot be prepared.

The n.m.r. spectrum of each compound shows only three peaks,

(see Table 1.2) two Si-Me 3 peaks and one methylene. There is no n.m.r. evidence at all for a proton on the bridging carbon, or a metal hydride, and the i.r. spectra and the X-ray structural data support this. The compounds have nearly identical i.r. spectra which are consistent with the observed structure. The usual.Me3SiCH2- vibrations_1 occur, the metal-carbon stretches are at 492, 463, 438 and 400 cm two for terminal M-C and two for bridging, and the bands at 374 and _1 310 cm are assigned as ring deformations.

The crystal structure of the niobium compound showed that the crystals are triclinic, with space group P1. The dimer has 4

Figure 1.4. Th„-.) Molecular structure of (11.-R)2n2R.,.. 22

symmetry 1- as shown in Figure 1.4. The bond lengths and angles are

given in Table 1.3. The metal-carbon bond lengths are similar to those 65 determined for LijCr2Me E7.4C4H EO. The ring is very close to being

TABLE 1.3 Bond lengths and angles for (µ-R)2Nb2R4. 0 Interatomic distances (A) and angles (0) are the two independent values unless otherwise stated.

Nb-C 1.995(9) Nb....Nb' 2.897(2) Nb'-C 1.954(9) C....C' 2.684(13) Nb-CH2 (mean of 4) 2.160(9) Si-C (mean of 24) 1.860(13)

C-Nb-C' 85.6(4) Nb-C-Si 119.8(6) Nb-C-Nb' 94.4(4) Nbi—c—Si 142.4(5)

planar, and the quL.si-aromatic structure allows us to consider the

niobium as having oxidation state V, which explains the diamagnetism

of the compound without having•to invoke metal-metal interation.

On addition of successive equivalents of Grignard to the

metal halide MC15 , it was hoped that complexes of the form MRnC15_n

would be formed. However, it was Shown that petroleum extraction at

•each "end-point" yielded a mixture of compounds. All were moisture

sensitive but compared to the binary alkyls, relatively stable

towards oxygen. The only compound isolable so far appears to be

TaR 3C12.

Hexakis(trimethylsilylmethyl)dimolybdenum and -ditungsten

From the interaction of the Grignard reagent with MoC15

or WC16 in ether the compounds M2(CH2SiMe3)6 are obtained,

(Table 1.1). These compounds are isomorphous and isostructural by

X-ray diffraction. The n.m.r. spectra (Table 1.2) are also virtually

identical and the physical properties of the compounds are also 23

similar. They sublime in vacuum at ca 120° and are stable in air over

short periods of time although on prolonged exposure they darken and

the end oxidation product appears to be the blue molybdenum and

tungsten oxides.

Both compounds are soluble in petroleum, benzene and ether,

the molybdenum compound giving yellow and the tungsten compound

intense orange-brown solutions; in absence of air these solutions are

stable for at least a week, but are rapidly oxidised in air.

The i.r. spectra are almost identical and as noted later are

consistent with the observed structure.

The compounds are more reactive than the other alkyls. Water,

dilute acids and bases and ethanol do not react with the molybdenum

compound but the tungsten compound reacts rapidly. Concentrated

mineral acids produce tetramethylsilane._ On heating the molybdenum

compound in aqueous 12N HC1, in absence of air, it dissolves to give

a pale yellow-green solution from which the yellow-green solid

Rb3Mo2C18 is obtained on addition of rubidium chloride. On refluxing

the alkyl in glacial acetic acid molybdenum(II) acetate is formed

quantitatively. It is most probable that the Mo-Mo bond remains intact in both reactions.

The reaction with nitric oxide is rapid at room temperature,

but no simple product could be isolated. Carbon monoxide reacts with

both compounds at room temperature and pressure but several products are formed sequentially and the systems have not yet been studied in

detail.

The i.r. spectra of the two compounds are almost identical.

The X-ray structural data54 indicate that the molecules have 2,

symmetry and there are three bands in the range for metal-carbon

stretches as expected, e.g. for 'molybdenum 'AO, 484 and 534 cm . •

'\ C

Figur I.5. The molecular structure of 1,1o2,R(3. 2C)

The crystal structure of the molybdenum compound has been 54 determined and shows the crystals to be monoclinic, with space group Pc. (Figure 1.5) The bond lengths and angles are given in

Table 1.4- The'mOlYbdenum-molybdenum bond distance is only slightly

TABLE 1.4 Bond lengths and angles for Mo2R6.

No. of independent Average values

Mo-Mo 4 2.167A Mo-CH2 24 2.131 Mo-Mo-CH2 24 100.6° Mo-CH2--Si 24 121.1°

67 longer than that found in tetraacetatodimolybdenum (2.12A). Each molybdenum atom has a distorted tetrahedral coordination and the configuration of the methylene groups in the dimer is staggered.

This staggering arises from inter-ligand repulsion,- there being no electronic factor favouring the eclipsed form as there is in 69 Re2C182 . 68The Mo-Mo bond may be considered to be multiple, i.e. a six-electron triple bond. The metal-carbon bonds are comparable with those of the niobium complex discussed earlier, and

Li4[Cr2 Me 8].4C4H80.65

Non-isolable trimethylsilylmethyl complexes

In many cases, the interaction of the lithium or Grignard reagent with metal halides yielded petroleum solutions of metal alkyls, but the products were not isolable (cf. TiR4 and ZrR4 as mentioned earlier). A brief resume of some of these reactions will now be given, but no details will appear in the experimental. 26

TABLE 1.5 Infra-red spectra of the trimethylsilylmethyl compounds

VR4 VOR 3 ( P-R)2 Nb2R4 (1.4-R) 2Ta 2R 4. , Mo2il ,o W2R 6 ., .. ,.,-- 13Maial..36.1.3.11116.1.191MV, -mml 686 vs 681 s . 690 s ' 690 s 691 s 695 vs 669 sh 680 vs 681 vs . 669 sh 669 sh 621 m 619 m. 616 m 619 m 612 m 610 w 615 m '610 m

504 s 522 s 495 m 492 m 534 w 560 sh 486 m 470 m 465 m 484 s 550 m 430 m 430 m 438 m - 440 vw 490 s 399 m 400 m 455 sh

373 m 374 s 310 m . 310.m

275 m 269 m 268 m 281 m 275 m 242 m 240 14 250 w 253 m 249 m 232 m 230 w 230 m 241 sh 235 m 27

Manganese: The addition of two equivalents of RMgC1 to a solution of MnC12/LiC1 in THE produced an almost colourless, involatile, petroleum-soluble oil. This was very air-sensitive, giving brown petroleum-insoluble products. The colourless oil was probably MnR2.

Rhenium: The interaction of up to five equivalents of RMgC1 with ReC15 gave a brown, involatile, petroleum-soluble, air- sensitive oil.

Iron: Although the conditions were varied considerably to try and obtain a clean reaction, the best reaction produced a very sensitive, black, amorphous, petroleum-soluble compound, possibly FeR2. The metal analysis (30.1% Fe) was rather high. (FeR2 requires 24.5, FeR3, 17.6%). The attempts are summarised in Table 1.6.

TABLE 1.6 Reaction of FeC13 with RLi or RMgCl.

Reagent Solvent Temp°C Petroleum Solution Comments colour

Grignard Ether any f black E.s.r. showed Fe(II)

Lithium Ether -60 t black-brown CO gave terminal carbonyl and acyl. Fe(II) again.

• Lithium Petroleum -40 orange Oily solid, invola- tile, decomposes -60 orange at 140°C. Fe(III) by e.s.r. Negative -70 dark orange test for chloride.

Nickel: NiBr2 reacted with two equivalents of RLi in petroleum to yield a brown, paramagnetic, air sensitive solution. The e.s.r. spectrum was inconclusive. The solution reacted with carbon monoxide to give terminal carbons, then acyls. The solution was found to catalyse the isomerisation and hydrogenation of olefins. 28

Cobalt: CoC12 reacted with two equivalents'of Rid in petroleum, with or without phosphines present, to yield an intense blue-green solution, which was oxidised by traces of air to a red-brown colour. Neither solution yielded an isolable compound. The blue solution catalysed the hydrogenation of olefins.

Uranium: The reaction of UC14 in THF with four equivalents of RMgC1 yielded a yellow, moderately petroleum-soluble, sticky solid which could not be purified free from halogen. 29

CHAPThi, II

NEOPENTYL COMPLEXES.

Having used the trimethylsilylmethyl ligand with some success, the next ligand to be considered was naturally the carbon analogue, neopentyl OR I). Previous studies of alkyls with stabilizing n- 60 bonding ligands had shown that neopentyl complexes were signifi- cantly less stable thermally than the trimethylsilylmethyl analogues e.g. cis(Me3SiCH2)2Pd(PEt3)2 is a stable compound but its neopentyl analogue, prepared only in solution, decomposes rapidly. However, it was also noted that the melting points of the neopentyl complexes were higher. and so it was hoped that binary neopentyl complexes would crystallise more easily than their trimethylsilylmethyl analogues. The expectations were fully realised in the case of the titanium and zirconium tetra-alkyls. As noted in Chapter I, the trimethylsilylmethyl complexes were not isolated as crystalline solids but were only partially characterised in solution. The neopentyl analogues crystallise particularly easily out of a very impure reaction mixture and can be purified by recrystallisation. One of the reasons why the initial crystallisation is relatively easy is that the coupled product from the neopentyl Grignard reaction

(Me3CCH2CH2CMe3) is more volatile than the silyl compound, and can be removed from the reaction mixture by gentle heating in vacuo before the petroleum extraction. Thus it is not present to inhibit crystallisation.

The compounds were prepared analogously to the trimethyl- silylmethyl compounds, but because of the lower thermal stability, some analogues were impossible to prepare. They are listed in

Table 4.1. 30

TABLE II.1 Neopentyl (R') compounds of Transition Metals

Compound Colour/form m.p.°C Electronic spectrum

+SIVEM112.121.2S TiR'4 yellow prisms 99 Rising absorption in u.v. ZrR'4 off-white prisms 103 HfR'4 white sticky solid TaR'3 C12 pale yellow needles 115 I I Mo2R'6 yellow plates 135

Tetrakis(neopentyl)titanium(IV), zirconium(IV) and hafnium(IV)

The interaction of neopentyl lithium with titanium, zirconium

or hafnium tetrachloride yields dark, almost black petroleum solutions

from which, after careful recrystallisations, can be obtained the

tetra-alkyls. The compounds are air and moisture sensitive and are

volatile and may be sublimed in vacuo at ca. 70°. They are thermally

unstable at room temperature over long periods of time (> one week)

their stabilities being in the order Ti

are soluble in petroleum, benzene and ether, and are unaffected over

short periods by chlorinated solvents. Their sensitivity to moisture

precludes chromatography.

The n.m.r. spectra (Table 11.2) of the compounds shows only

the expected methyl and methylene resonances for TiR'4 and ZrR'4 .

The spectrum of HfR'4 showed it to be rather impure. The compounds are moisture sensitive to such an extent that a peak at 8.86 T for neopentane always appears unless the sample is prepared very carefully in vacuo. The i.r. spectra of the neopentyls show essen- tially the expected ligand vibrations. For the metal carbon _1 stretches, TiR'4 and ZrR'4 both have a medium band at ca. 500 cm TABLE 11.2 N.m.r. spectra of the neopentyl compounds

Compound CH3 CH2 Remarks

TiRI4 8.62 7.64 CH 3'' CH 2 - 9:2 ZrR' 4 8.62 8.34 11 HfR14 8.62 (obscured) impure TaR'3 C12 8.65 7.23 ft /4021216 8.64 7.09 ft

W2R'6 8.64 7.12 impure

with a weak shoulder at 480 cm . This would be in accord with a

slightly distorted tetrahedral structure. Crystal structures have 70 71 been determined for Ti(benzy1)4 and Zr(benzy1)4 and both show that

the molecule deviates from a regular tetrahedron but it is possible

that this is due to interaction, either steric or electronic, with

the phenyl ring.

The compounds all seem more reactive than the molybdenum,

tantalum or chromium neopentyls. TiR'4 and ZrR'4 react with carbon

monoxide giving terminal carbonyl products but no acyls at one

atmosphere and room temperature, and nitric oxide and pyridine react

also. There is no significant polymerisation of ethylene or propylene

at up to 5 atmospheres and neither compound will react with hydrogen

under normal conditions.

Vanadium neopentyls

The reaction of VC14 or VOC13 with the neopentyl Grignard or lithium reagent produced no stable compound. Since V(CH2SiMe3)4 is

the least stable thermally of the trimethylsilylmethyl complexes (it

decomposes at room temperature in a few days) it is not surprising that the tetra-neopentyl analogue is unstable. The tetra-benzyl 32

compound, although characterised in solution, is also non-isolable.52

Theinteraction in petrol at -60° of VC14 with four equiva-

lents of neopentyl lithium yielded an intense green solution which

turned brown on warming to room temperature. The e.s.r. spectrum • of the green solution showed two S = 1 species, presumably compounds

of the fOrm V(CH2CMe 3)nC1 4_/11 and a free radical. The corresponding reaction with VOC13 at low temperature

yields a brown solution with a normal anisotropic S = 1 spectrum,

different from either of the species obtained from VC14. On warming

the eight-line mobile solution spectrum is obtained but above 170°K

a new fifteen line spectrum is observed (Figure II.1). This signal

has the correct intensity distribution 1,2,3,41 5,60,80,6,51 40,21 1, for two equivalent vanadium nuclei (I = 72 ) interacting with the electronic spin. The g-factor is 1.989(3) and the hyperfine splitting is 3.06(1)mT. Similar spectra have been observed for dimeric 72 vanadium tartrates.

Tris(neopentyl)dichlorotantalum(V)

As mentioned before, the niobium and tantalum systems are very complex, and no binary neopentyls have been prepared. Addition of NbCl5 or TaC15 to excess Grignard reagent produces a dark brown oil containing many species. However, the reaction of three equiva- lents of Grignard reagent with TaC15 does produce, in a relatively clean reaction, the trisalkyl-diehloro species, R'3 TaC12. This 14 compound is analogous to the known methyl complexes of Nb and Ta.

These are yellow volatile oils but decompose at room temperature over several hours. R'3 TaC12 is remarkably stable to heat and may be sublimed in vacuum with a flame without significant decomposition.

It is moisture sensitive but relatively stable towards oxygen. It 9 osr,ounr-Itusur*g.x.ao====.4....rxrat. ji

Fii ur TI,1. E.s.r. spectrum of a vanadium-loopontyl compound with one unpaired nlCCtrrn lntc!nictinp: w1 (.h two oquivnint vnmullin nu oloi. 34

is soluble in petroleum, ether, benzene and halogenated solvents but decomposes in protic solvents.

The n.m.r. spectrum (Table 11.2) showsonly the expected resonances for methyl and methylene. The i.r. spectrum in the region below 600 cm is complex, as both the metal carbon and metal chlorine stretches occur here. An attempt to make the bromide,

T3 TaBr2, to try to assign the metal halogen stretches, was unsuccessful.

It has not yet been possible to isolate any other complexes in this system.

Hexakis(neopentyl)dimolybdenum(III)

The interaction of neopentyllithium with MoC15 in ether yields the compound Mo2R1 6, in the same way as the trimethylsilyl- methyl analogue. The yield of the reaction is lowered considerably if the Grignard reagent is used. The compound is stable in air over short periods, but quite oxygen sensitive in solution. As with

Mo2R6, the terminal oxidation product is the blue oxide. It is volatile, sublimes in vacuo at 130°, but in poor yield due to thermal decomposition. It is soluble in most organic solvents and is relatively stable to protic solvents and water. Concentrated mineral acids liberate neopentane, with the formation of complex anionic molybdenum halides cf. the trimethylsilylmethyl analogue.

Mo2R'6 reacts rapidly with carbon monoxide and nitric oxide, but apart from this, it is rather inert.

The n.m.r. spectrum is simple, and the i.r. suggests a structure the same as that found for the silyl compound.

The interaction of the lithium reagent with WC16 produces a brown oil, presumably containing W2R'6, but it was not possible to isolate this as a pure compound. 35

TABLE II.3 Infra-red spectra of the neopentyl complexes

T1R'4 ZrRi4 TaR13 Cl2 MO2R 6 639 s 639 w 615 w

540 s 542 s 566 s 540 s 505 s 506 s

...... _ca. 495 sh 448 w 446 w 461

410

380 m 372 m

365 s 363 s 351 s 299 s 299 s 305 s 300 m 258 s 36

CHAPTER III

NEOPENTYL AND RELATED ALKYLS OF CHROMIUM(IV)

Examination of Table 2 in the introduction, i.e. the

recently prepared alkyls, shows that the metal with the largest number

of isolable compounds is chromium, and that they are in most cases

chromium(IV) compounds. In fact, the first compound isolated in the

course of this work was tetrakis(trimethylsilylmethyl)chromium(IV) 34,73 and this has since been investigated more fully. Up to this

point, compounds with the metal in the IV oxidation state had been II considered rather uncommon for chromium. The mixed oxides Mit Cr06, 143IIcro5 and M 2 11Cr0474a (M = Sr and Ba), the oxide Cr027413 and the I I 75a II 75b • I ionic fluorides M CrF5:M 2 CrF6 and M CrF6 (M = Li, K, Rb, II Cs; M Ba, Sr, Ca, Cd, Hg, Ag) are well known. The halides, 75a CrX4 are reported but only CrF4 is isolable while CrC14 and CrBr4

exist in vapours, although CrC14 appears to be obtained at low

temperatures by the action of HC1 on Cr(CH2SiMe3)4.37 The best

characterised compounds are the trimethylsilylmethyl,37 the tertiary • alkoxides Cr(0R)4,76 the numerous dialkylamides, e.g. Cr(NEti)4 ,77

and the above -alkyls. It would seem that carbon is now to take its

place along-side the other first row elements, nitrogen, oxygen and

fluorine, as being readily able to form chromium(IV) compounds. It

is expected that, in due course, chromium(IV) complexes with the

'metal bound to four boron atoms will be synthesised.

Synthesis and chemical properties of chromium(IV) alkyls

Tetra-alkyl chromium compounds, CrR4, (R = neopentyl,

neophyl, tritylmethyl and methyl) have been prepared by the inter-

action of the Grignard or lithium reagent derived from the respective

chloride with the tetrahydrofuranate of chromic chloride, CrC13.3THF,

or, in the case of tetramethylchromium, an exchange reaction between 37 methyllithium and chromium(IV) t-butoxide. Details are given in the experimental section. The compounds and some of their properties are listed in Table III.1

The neopentyl, neophyl and tritylmethyl compounds are purple to red, paramagnetic, crystalline solids and tetramethylchromium is a volatile, thermally unstable, maroon oil. They are all readily soluble in organic solvents and are unaffected by water, alcohols, dilute acids and bases. Tetramethylchromium co-distils in vacuo with petrol at -35°C and tetrakis(neopentyl)chromium sublimes in vacuo at

60°C in good yield, but the other compounds, due to their higher molecular weights, decompose before the sublimation temperature is reached.

TABTN, III.1 Properties of Chromium(IV) tetra-alkyls

Electronic ...1 1 Colour, form m.p.(° C) a (B.M.) Compound spectruM(cm ) eff

Cr(CH2CMe3)4 maroon needles 110 18500, 21100 2.7 (1090) Cr(CH2CMe2Ph)4 purple prisms _ca. 120(dec.) 18200, 20500 2.8 (1380) Cr(CH2CPh3 )4 purple prisms _ca. 130(dec.) 17600, 20200 2.6 (1380) CrMe4 maroon oil ca. -60 20000, 22200 - (ca. 600) Cr(CH2SiMe3 )4 (1. purple needles 40 17100, 19400 2.9 (1060) •

by Evans' n.m.r. method C in benzene a in petroleum (CM) d from ref. 37.

The general chemical behaviour of the compound is similar to that of tetrakis(trimethylsilylmethyl)qhromiuM37 in that they fail to react with primary aliphatic and aromatic amines, alkyl and aryl tertiary phosphines, ethylene and liquid alkenes and carbon monoxide.

Their reactivity towards oxygen and nitric oxide is in the order methyl > neopentyles"trimethylsilylmethyl > neophyl >> tritylmethyl.

The tritylmethyl compound is sufficiently stable towards oxygen to be handled in air during its isolation. The oxygen and nitric oxide reactions of the neopentyl compound have been followed by e.s.r. Strong mineral acids react with the compounds yielding III solutions with Cr and the respective alkane and gaseous HC1 and chlorine both react to give, initially, the unstable red species believed to be chromium tetrachloride,37 and ultimately a mixture II III of Cr and Cr chlorides. Bis(1,3 dimethylenetetramethyldi- siloxane)chromium(IV) has been prepared in poor yield and has been only partially characterised by its e.s.r. and electronic spectrum.

The reaction of the di-Grignard with CrC13.3THF could lead to two products after an aqueous extraction; the chelate, or the tetra- alkyl Cr(CH2SiMe20SiMe 3)4. The rhombic splitting observed in the 78 e.s.r. favours the chelate structure. It was not possible to isolate the pure compound as it readily disproportionates to give a reddish-purple, involatile chromium(III) species.

Tetrakis(neopentyl)chromium(IV) and the tetrakis(neopentyl)chromate

(III) ion.

The interaction, in tetrahydrofuran (THF), of neopentyl- lithium with CrC13.3THF produces a blue-black 'solution stable only in THF, and we formulate this species as Cr(CH2CMe3)4 by its spectroscopic properties and by analogy to the trimethylsilylmethyl analogue.37 Attempts to isolate a crystalline solid by addition of large cations failed. The solution is oxidised immediately by trace amounts of oxygen or benzoyl peroxide to the maroon tetrakis(neopenty1)- chromium(IV) and the latter may be reduced electrolytically or 39

chemically to the blue anion. Coulometric measurements on an ethanolic solution with tetra-n-butylammonium iodide as supporting electrolyte show a reduction wave at a half-wave potutial of ca. -1.65V vs the-saturated calomel electrode.

The interaction of neopentylmagnesium chloride with

CrC13.3THF does not produce the tetra-alkyl in a one-step process as is the case with the trimethylsilylmethyl. Addition of up to 2 equivalents produces an orange species which can be isolated as a petroleum soluble crystalline compound, readily disproportionating to give the tetra-alkyl on standing or more rapidly by the action of water or air. The X-band (v = 9.2 GHz) e.s.r. spectrum of the frozen orange solution shows bands at 203, 520 and 850 mT which may be attributed to a species having S = 1 and D!1'0.5 cm . The bands observed at 202 and 850 mT are due to molecules with their principal axis nearly parallel to the magnetic field and the other due to molecules with this axis perpendicular to the field. The presence of a strong distortion suggests that this is a mixed species

Cr[CH2CMe3],;C14_n and the absence of splitting in the perpendicular band at 520 mT indicates near-axial 'symmetry implying n = 1 or 3.

This compound was too sensitive to obtain a reliable analysis. On addition of further Grignard reagent, the colour becomes initially blue, but reverts to the orange colour until an "end-point" is reached at ca. 3 equivalents, after which the colour remains an intense, deep blue. This blue solution is stable only in ether, and addition of air, water or petroleum causes immediate disproportionation or oxidation to the maroon tetra-alkyl. Attempts to crystallise the blue compound failed. The e.s.r. spectrum of this solution, which showed a broad line at g ca. 2 with a sharp line at its centre, is ambiguous but probably indicates octahedral chromium(III). 40

The reaction of Cr(CH2CMe3)4 with nitric oxide yields a solution in which three species are observed in the room temperature e.s.r. spectrum. There is a free radical with hyperfine structure attributable to one nitrogen atom and four equivalent protons, which is probably the dineopentyl nitroxide radical, (Me 3CCH 2 )2 NO% The spin Hamiltonian parameters are: g = 2.0062(3), aN = 1.30(2) mT and a4 - 0.75(2) mT. The other two compounds contain chromium bonded to nitrogen and the stronger of these has g = 1.9730(5) and al4 =

0.59(2) mT with a chromium hyperfine structure that was not suffi- ciently well resolved for the magnitude of the coupling to be established. The other spectrum is about one third as intense and overlaps with the other chromium-nitrogen compound preventing accurate measurements. The g-factor is 1.968 and the nitrogen hyperfine splitting 0.6 mT.

The reaction with oxygen is rapid, giving brown solutions containing a mixture of products. The e.s.r. spectra of solutions exposed to a trace of oxygen reveal two transient S = 1 species analogous to the species A and D observed with Cr(CH2SiMe3)437 but with substantially smaller proton hyperfine couplings.

Tetrakis(neophyl)chromium(IV)

The interaction of neophylmagnesium chloride with CrC13.3THF produces tetrakis(neophyl)chromium(IV) as maroon crystals after oxidation or disproportionation of a deep blue solution in an analogous way to the neopentyl reaction. The lithium reagent was not prepared but the existence of the Cr(CH2CMe2Ph)4- ion was established by coulometric measurements on the tetra-alkyl. A solution in acetone with tetra-n-butylammonium iodide as supporting electrolyte shows a one electron reduction wave at a half-wave potential of ca. -1.97V vs the saturated calomel electrode. 41

Tetrakis(tritylmethyl)chromium(IV)

The interaction of tritylmethyllithium with CrC13.3THF

at -78° produces the anion Cr(CH2CPh3)4- which is readily oxidised

by air to the air-stable, purple tetrakis(tritylmethyl)chromium(IV).

The reaction must be carried out at -78° because above this

temperature there is appreciable rearrangement of the lithium reagent

to give Ph-CH2-CPh2Li. 79The latter does react to form a purple

species, but on warming to room temperature, decomposition occurs

rapidly, presumably due to interaction of the p hydrogen atoms.

Coulometric measurements of a solution of the chromium(IV) alkyl in

acetone show a one electron reduction wave a half-wave potential

of ca. -1.99V vs the saturated calomel electrode.

Tetramethylchromium(IV) t The exchange reaction between Cr(0Bu )4 and a large excess

of freshly prepared methyllithium in petrol yields an unstable maroon

solution of tetramethylchromium(IV). This product is not obtained if

a coordinating solvent, e.g., diethyl ether or THF, is used, and

probably for this reason the reaction of Cr(OBut)4 with methylmagnesium 80 chloride does not yield tetramethylchromium. This exchange reaction

has been also recently reported independently35 and used to prepare

Cr(t-buty1)4 while Me4Cr and other alkyls were characterised in

solution. Addition of ligands such as tertiary amines and phosphines,

or carbon monoxide to the CrMe4-MeLi solution produces orange solutions

which appear to contain chromium(III), probably. either CrMe3 L 3 81 17 (cf. CrMe 3.3THF) or a lithium methylchromate.

If the tetramethylchromium and the petroleum are co-distilled

out of the reaction mixture in vacuo and collected at -78° the

resulting solution is considerably less stable than when the excess lithium reagent is present. It decomposes slowly at -60° and rapidly 42

at room temperature. In this purer state, the compound seems more reactive towards amines and phosphines than the other chromium(IV) compounds but no adducts could be isolated.

A petroleum-free sample (see experimental section) was prepared and a mass spectrum obtained, but this showed no chromium- containing species, hardly surprising due to the very low thermal stability of the compound.

Infra-red Spectra

The i.r. spectra of the isolable compounds is dominated by the ligand vibrations, very few of which change significantly in the alkyls. Most of these vibrations are well-known aliphatic and aromatic bands and have not been included in Table 111.2. This lists only bands from 700 to 250 cm . For a near-tetrahedral structure two i.r. active bands are expected for Cr-C between 500 _1 and 400 cm as observed.

Electronic Spectra

The groundstate for a tetrahedral d 2 molecule is 3 A2 2) and, considering only d-electron transitions, there are three excited triplets, 3 T2(et2), 3 T1(et2) and 3 T1(t2). The electric dipole operator transforms as T2 so that the 3 A 2 3T1 transitions are allowed but 3 A 2 3T2 is forbidden. In the strong field limit

3 A 3 T1(t2) is also forbidden as it is a two electron transition, but it is usually found that the interelectronic electrostatic interaction mixes the two 3 T 1 states and therefore two 3 A 3 T 1 absorption bands are expected.

If the tetrahedron suffers a tetragonal distortion, reducing the symmetry to D 2 d , the degeneracy of the triplets is partially lifted. The T2 states are split into an orbital singlet and a doublet, B2 + E, and T1 is similarly split into A2 + E. The ground 43

TABLE 111.2 Infra-red Spectra of Chromium(IV) tetra-alkyls a

Cr(CH 2CMe Cr(CH2CMe2a)4 Cr(CH2CPh3)4

700 S 700 S

693 s

672 640 633 sharp, w 625 620 601 m 610 590 m 584 w 573 w 563 w 550 w

555 s 550 sh 528 w 534 m 513 w 500 m

508 m

481 m 485 m

450 w 424 m 410 w (br)

400 vw 400 vw

378 m 322 vw 293 m 280

a from 700-250 cm in nujol or vaseline mulls. 44

state becomes 3B1 and electric dipole transitions are allowed to d 3 E.

All the chromium(IV) alkyls have an absorption band at ca. 500 nm and then rapidly rising absorption towards shorter wave- length. The oscillator strength of the 500 nm band is ca. 0.02 which is remarkably large for formally d-d transitions, but is not unreasonably so for a highly covalent non-centrosymmetric system.82

The band is assymmetric and may be reconstructed as two components with intensities in the ratio 2:1 with the stronger at the shorter wavelength. 34 This spectrum has been assigned to a system of slightly elongated tetrahedral symmetry (D 2d), with the two strong bands at ca. 500 nm being assigned to the transitions 3 BI -' 3 E(Ti ) and

3 B I -+ 3 A 2 (T 1 ).

Electron spin resonance spectra

Chromium(IV) is a d2 system and the compounds are all paramagnetic with two unpaired electrons. The magnetic moments

(Table III.1) are all slightly lower than the values predicted from the g values. This could be due to some interaction between molecules in the solid phase. All the chromium(IV) alkyls have e.s.r. spectra, and the best characterised of these is that of 37 Cr(CH2SiMe3)4. It consists essentially of a single broad line with g = 1.99310.005 in petroleum or toluene at room temperature, and on cooling this broadens and splits into two lines, while a new line appears at ca. 1600 mT (Figure III.1). This spectrum is typical 83 of a triplet (S = 1) state with axial symmetry in a rigid matrix.

The other alkyls have spectra which show the same general features but the splitting and structure on the main g = ca. 2 signal varies depending on the alkyl group present. The line-width of the mid- 4

Figure III.1. E.s.r. spectrum of Cr(CH9SiMe3)4 in petroleum. Midfield signal showing presence of two conformers. 40

field signal of CrR 4 (where R trimethylsilylmethyl or neopentyl) passes through a minimum at 140°K and both it and the fine structure splitting increase on further cooling. At low temperatures an additional line appears at g = ca. 2. Careful control of temperature and concentration with Cr(CH2SiMe3)4 have shown that this is due to some magnetically concentrated phase which separates out from the solution at low temperatures, presumably because of.reduced solubility. A solution in petroleum, in which CrR 4 is extremely _3 soluble at room temperature, that is not more than 10 molar, does not show the sharp line. On the other hand a solution in benzene, in which CrR4 is much less soluble at room temperature, shows only _4 a broad line with gra2 unless the concentration is ca. 10 molar or less in which case a normal triplet state spectrum is observed in addition to the g = 2 line.

It is possible to obtain supersaturated frozen solutions by plunging tubes containing the solution into liquid nitrogen.

These samples have weaker g = 2 signals than others of the same concentration that were frozen slowly. If such a supersaturated solution in petroleum is warmed to 135°K or more the g = 2 signal grows in intensity until it is the same as that of the slowly frozen solution. This process takes about an hour at 135°K and is faster at higher temperatures. Solutions in benzene behave similarly but temperatures of 200°K or more are required before this equilibration occurs at a perceptible rate.

The g = 2 signal is not identical with the spectrum of pure solid CrR4, which is also a single line with gr^'2 but with a different line width, so one must presume that the separated phase in the frozen solution samples incorporates some of the solvent.

These pheomena are displayed in a more extreme manner by 47

Cr[CH2CPh 3]4. In all solvents tried except benzene a single broad line (150 mT) centred on g = 2 was always observed. Figure 111.2.

In benzene there is a similar broad line, but, provided the solution was frozen rapidly, a number of additional sharp lines are detected.

These lines are not accurately reproducible or isotropic so they cannot be a true frozen solution spectrum.. Nevertheless some features are almost reproducible; such as the splitting of the outermost pair of lines allowing a rough estimate of D. The lines are very narrow (ca. 2.5 mT) compared to the other alkyls of which

Cr[CH2CMe3]4 has the narrowest lines with 6.0 mT width at 145°K.

The behaviour of CrECH2CPh31, is probably due to its low solubility relative to the other alkyls in all solvents except benzene in which its solubility is about the same as the others.

The anomalous line broadening of CrECH2SiMe31, at low temperatures is similarly interpreted as being due to a mixture of two species, both of which remain in solution. At temperatures below

125°K the z-component of the spectrum splits into two. The resolution is clearest at 115°K since below 110°K one of the z-lines is hidden beneath the stronger x,y line and the overall spectrum appears to be broadened rather than split. It is possible that the two spectra are due to two conformers of Cr[CH2SiMe3]4 which may interchange rapidly at temperatures above 125°K. A similar resolution into two conformers at low temperatures has already been postulated for MoCNMe2h, which has a low temperature splitting in the electronic spectrum.84 Two conformers of

CrECH2SiMe3J 4 may be visualised related to each other by rotation of the chromium-carbon bond in exactly the same way as those considered for Mo[NMe2]4.

The broadening of the main peak of the electronic spectrum

48

140TriT

Fig7.ure 111.2. E.s.r. spectrum of Cr(C1-1„CP7aj4 in benzene. (Mid-field signal only.)

2OmT

TT :Tt)c.trum o: Cr(H,CT,':es), in pF..-troleum. 1+9

of Cr(CH2SiMe3)4 at low temperatures has been similarly interpreted.73, 85

The x,y lines of. Cr[CH2CMe 3]4, behave in the same fashion as those of Cr[CH2SiMe3]4, so we may presume that there is a similar resolution into two species, but in the neopentyl it is never possible to resolve the z-lines of the spectrum. (Figure 111.3).

The spectra obtained by Ward et al. for the neopentyl and other alkyls showed only a sharp band at ca. 0.3 mT and a much weaker one 86 at ca. 0.16 mT. It would appear that the concentrations used in _3 this case ( 3 x 10 M) were such that only the magnetically concentrated signal was observed at mid-field.

A spectrum of two species is also detected in frozen solutions of Cr[CH2CMe2Ph]4. (Figure 111.0. In this case the difference between the two forms is so great that the x,y line and the x',y' line are resolved, but the z' line is always concealed under the x,y one.

E.s.r. spectra were also observed with solutions of CrMe4 and Cr[(CH2SiMe2)20]2 (Figures 111.5 and 6) but were not well enough characterised to permit detailed study.

The Spin Hamiltonian parameters of the alkyls, together with those of Cr(CH2SiMe3)4 for comparison, are collected in

Table 111.3.

50

Fre 111.4. E.s.r. spectrum of Cr(CH201:..e,,Ph)4 in ':.)7;troleum.

40mT

spectrum of CrYe,„ in petroleum.

-n m

( „ 1 :;.;.s.r. spectrum of Crar..42:31e20Sie2CH,)2 in petrol TABLE 111.3 Spin Hamiltonian parameters of frozen solutions of Cr114.

(D/cm 1 ) (E/cm 1 ) (11°K) Zit gL

Me 110 ^1 1.99 a ^)0.005 CH2CMe3 145 1.990(3) 0.011(1) CH 2CMe2Ph 110 - 1.988(5) 0.021(5) CH 2CMe2Ph 110 - 1.992(5) 0.049(5) CH 2CPh3 ° 110 r'-' 2.0 a 0.07 0.01 CH 2SiMe 3 145 1.986(3) 1.990(3) 0.063(5) CH 2SiMe 3 110 1.986(3) 1.989(3) 0.073(5) CH 2 SiM 3 110 1.981(3) 1.989(3) 0.089(5) 12 t( CH 2SiMe2) 0 145 1.994(5) a 0.123(10) 0.017(3) CH 2Ph c 110 r" 2.0 a r'' 0 .2

a Average g-factor

b Provisional values from improperly frozen solutions

A. J. Shortland Ph.D. Thesis (ref. 73)

52

34 The zero-field splittings, D, have been derived from the

following equations:-

{ X1 D = 4 14 E(T 2 )J 14 [3 12 2(Z2 )3

x2.1. • - 4 - X211 w[lE(T2 )] 14[1 P2 (11 2)] J

1 ,) 4 X2 6 ( A + 8B+ 2C)2

- 8X( XII- X.I. ) 1 - 1 A A + 8B + 2C [

where Xli and Xj., the spin-orbit coupling parameters, are derived

from the g factors as follows:-

8ko g = 2.0023 - WE3 B2 ( 2)]

8X1. g jt = 2.0023 ItIC3 E(T2 )J

All other values are obtained from the electronic spectra. These

derivations have been presented much more fully in reference 34. 53

CONCLUSION

The recent emergence of the trimethylsilylmethyl and

related alkyls, the methyls, the benzyls and the norbornyl and

related alkyls is clearly proof that by blocking the 0-elimination/

hydride-transfer reaction, binary alkyls of considerable stability

can be prepared, and in most cases, isolated as pure crystalline

compounds. However, if one also takes into consideration the

chemical reactivity, it is apparent that the least reactive are

those which are most hindered sterically, e.g. (tritylmethyl),,Cr

and the borbornyls. (This is closely related to the coordinative

saturation of the alkyl complexes). Cr(t-butyl),, is an example of

a compound which exists purely because of steric reasons. Molecular

models show that in the space immediately surrounding the metal atom,

we have the following order for steric hindrance for the simple

methyl-substituted alkyls:

t-butyl > neopentyl > trimethylsilylmethyl' and the order of reactivity towards oxygen etc. is the reverse, as expected. The reason why Cr(t-butyl),, does exist is that the four t-butyl ligands are so crowded together at room temperature that it is impossible to attain the transition state required for the 0-elimination reaction. However, at higher temperatures,

(ca. 80°) this reaction does take place, and consequently it is the least thermally stable of the three. The trimethylsilylmethyl is the most stable thermally, but this is possibly due to deloca.lisation of the du electrons via du -4 du bonding which is impossible for neopentyl. The phenyl-substituted alkyls decompose at their melting points, perhaps due to some form of interaction of the phenyl ring with the Cr metal to give arene-type complexes. 54

One of the features which became increasingly apparent

during the progress of the work was the complete lack of uniformity

of conditions required for the preparation of each alkyl. Some

required lithium reagent, others required Grignard reagent, and

there were often unusual oxidations or reductions. The trimethyl-

silylmethyl system had the following changes of oxidation state:

IV VVOG13 -4 V R4 V - - III III (MO C15)2 -4 MO R3-M0 R 3

W C16 W R3-W

The reduction of the ligand to give the carbene bridges in the cases

of the niobium and tantalum compounds was even more surprising.

;There was very little consistency even within a particular series of

alkyls of the same metal. For the chromium system, the trimethyl-

silylmethyl Grignard produced the tetra-alkyl in a one-step process,

while the neopentyl and neophyl went through three stages, and the

methyl and tritylmethyl Grignards failed to give Cr(IV) specieS. I No explanation is offered for this, nor why chromium(IV) should be

the preferential oxidation state for binary alkyls. It is possible

to prepare Cr(III) alkyls by having sufficiently bulky groups,

e.g. (4-cam)3Cr36 and this is analogous to the tris(diisopropyl-

amido)chromium(III).94 Attempts by Kruse35 to prepare Cr(III) alkyls

using Et3CCH2- and other bulky 13 substituted alkyls failed, but it

is to be expected that substitutents on the a-position as well

would increase the chances of isolating further binary chromium(III)

species. 55

EXPERIMENTAL

Microanalyses by Beller, Mttingen, Bernhard, MUlheim and

Imperial College Tlicroanalytical Laboratories. Metal analyses were obtained using a Perkin-Elmer Atomic Absorption instrument Model 303, or standard gravimetric methods. The analytical data are collected in Tables E.1-3...... Spectroscopic instruments. Perkin-Elmer R14 and R12A n.m.r. spectro- meters, Varian E-12 X band e.s.r. spectrometer. Infra-red spectra were recorded on a Perkin-Elmer Model 325 spectrophotometer call- _1 brated with polystyrene over the region 5000-200 cm . Samples were _1 run between KBr plates for the region 4000-450 cm and between _1 polyethylene plates in the region 500-200 cm . Electronic spectra were recorded on a Cary 14 Spectrophotometer as solutions in light petroleum, benzene or tetrahydrofuran. Mass spectra were recorded on a A.E.I. MS 9.

The X-ray structural data were collected on a Siemens four- circle ,diffractometer, using Cu-Ka radiation. The structures were solved by heavy atom, Patterson and Fourier methods. For Mo2R6, the crystals were mounted in a Lindemann tube inside a nitrogen-filled

"dry bag", but the crystals of the niobium compound were too sensitive for this method, and the apparatus shown in Figure E.1 was used instead.

Preparations All preparations and other operations were carried out in oxygen-free nitrogen, argon (for preparation of lithiup reagents) or in vacuum unless otherwise stated.

Where possible, commercial alkyl halides were used as received. Neophyl chloride was prepared by the alkylation of benzene with 3-chloro-2-methylpropane using, sulphuric acid as a A

t O un.

Rubber sl e,eve„.

Lill.J.:111 till, . • , • (/ i / p-t--J—r_..;--f?

Fircure E. 1. Apparatus for mounting air-sen it vu crstals. 57

87 catalyst and tritylmethylchloride was prepared by the reaction of

tritylsodium with dichloromethane and recrystallised before use.S8

Tetrakis(t-butoxy)chromium(IV) was prepared by the method of Krauss 89 and Munster and purified by sublimation. Solvents were dried and

degassed prior to use. Petroleum had b.p. 3040° except for

chromatography when it had b.p. 60-80°. Chromatography was carried

out using acid washed alumina (Spencer Chemicals for Industry,

Type H) and cellulose powder (Whatman CF-11). G.l.c. analyses were

obtained using a Perkin-Elmer F-11 instrument with Kent Chromalog

integrator, flame ionization detector and squalene columns.

Trimethylsilylmethylmagnesium chloride was prepared in 90 diethylether as ca. 1M solution; the yields were over 90%. The 91 lithium reagent was prepared in cyclohexane. When specially

,pure reagent was required the solvent was stripped and the lithium 4 alkyl sublimed (10 mm Hg,, 105°),re-dissolved in an appropriate

solvent and standardised; g.l.c. tests of these solutions showed

only traces of Ce 3SiCH2)2 and Me4Si. Neopentyllithium was prepared

in the same way. The neopentyl92 and neophyl93 Grignard reagents - 1 ' were prepared by literature methods, as was tritylmethyllithium. 9k

The di-Grignard reagent derived froal,3-bis(chloromethyl)tetra-

methyldisiloxane was prepared in a similar way to the trimethyl-

silylmethyl Grignard reagent and a sample was hydrolysed and shown

by n.m.r. to be essentially hexamethyldisiloxane. There were no

Si-CH2-C1 protons observed. 58

Chapter I

TABLE E.1 Analytical data for trimethylsilylmethyl compounds

Compound Required Found H Other C H Other

C16 1144 48.1 11.0 12.8 (V) 46.3 10.6 12.4 (V) C12 H 054 V 43.9 10.1 15.5 (V) 44.3 9.8 14.9 (V) C12 H3 1 NbSi 3 40.9 8.8 23.9 (Si) 39.4 8.4 24.4 (Si) 26.4 (Nb) 26.6 (Nb)

C12 H31 Si3 Ta 30.8 7.2 30.9 7.0

C12 H33 M0Si3 40.3 9.2 41.0 9.3 C12 113 3 Si3 W 32.3 7.4 31.2 6.8

• ■ • . • • • . Tetrakis(trimethylsilylmethyl)vanadium(IV) a) From VOC13. To a solution of trimethylsilylmethylmagnesium chloride (300 mmol) in diethylether (250 ml) was added VOC13 (5 ml,

50 mmol) in diethylether (150 ml) dropwise at room temperature over about 1 hr. The deep green solution was filtered through a frit, the solvent stripped and the residue extracted with petroleum giving a deep green solution. After concentration in vacuum to 200 ml, cooling to -78° and collection at that temperature, the product was recrystallised from petroleum (150 ml ) at -30° giving dark green needles, which were filtered cold, washed with chilled petrol and dried in vacuum (4 g, 20%). b) From VC14. To a solution of trimethylsilylmethyllithium (13 mmol) in petroleum (50 ml) was added VC14 (0.3 ml, 3 mmol) in petroleum (5 ml) at -78°. On stirring and slowly warming to room temperature a green solution was obtained which was filtered, 59

concentrated in vacuum to 20 ml, and cooled to -78°. The crystals

were treated as above (yield, ca. 30%).

Oxotris(trimethylsilylmethyl)vanadium(V)

A green petroleum solution of V(CH2SiMe 3)4 (as obtained

above before crystallisation) was absorbed on a cellulose column

and eluted slowly with petroleum under nitrogen. The green band

initially formed turned yellow as it moved down the column. The

eluate was collected, concentrated and cooled to 0° to give yellow

needles. (ca. 50% based on VC14). _ . . Bis-p,-(trimethylsilylmethylidyne)tetrakis(trimethylsilylmethyl)-

diniobium(V)

To a solution of trimethylsilylmethylmagnesium chloride

;(550 mmol) in diethylether (500 ml), was added NbC15 (27 g, 100 mmol)

in diethylether (250 ml) at room temperature over three hours. The

dark red-brown solution was filtered and the solvent stripped. The

residue was extracted with petroleum and filtered. 'After concen-

tration in vacuum to 50 ml, cooling to -35°C for several hours and

collection at that temperature the compound was recrystallised

from petroleum (25 ml) at -40°C giving red-brown prisms (6 g, 20%).

Bis-g(trimethylsilylmethylidyne)tetrakis(trimethylsilylmethyl)-

ditantalum(V)

As for the niobium compound but using TaC15 (35 g). The

orange-red petroleum extract left an orange oil which was dissolved

in diethylether (150 ml) and allowed to stand at -35°. Light orange

prisms of the compound slowly formed and were collected at this

temperature, washed with chilled ether and dried in vacuum. At • this stage the compound could be readily recrystallised from

petroleum. (7 g, 15%). 60

Hexakis(trimethylsilylmethyl)dimolybdenum(III)

To a solution of trimethylsilylmethylmagnesium chloride

(110 mmol) in diethylether (100 ml) was added MoC15 (5.4 g, 20 mmol)

in diethylether (50 ml) at room temperature over one hour and the

mixture stirred for 5 hours. After filtering and stripping the

solvent, the brown oily residue was extracted with petroleum (30-40°)

giving a dark brown solution. This was concentrated in vacuum to

ca. 20 ml, transferred to an alumina column and eluted with petroleum.

The yellow fraction on slow concentration in vacuum gave crystals

of the compound (1 g, 15%).

HeXakis(trimethylsilylmethyl)ditungsten(III)

To a solution of trimethylsilylmethylmagnesium chloride

(110 mmol) in diethylether (150 ml) was added a slurry of WC16

,(10 g, 25 mmol) in diethylether (300 ml) at room temperature over

one hour with vigorous stirring. The green colour formed initially

gradually turned red-brown. After stirring for two hours the

solution was decanted and the solvent stripped. The brown oily

residue was extracted with petroleum (30-40°), concentrated in

vacuum to ca. 30 ml, transferred to a cellulose column and eluted

with petroleum. Concentration of the red-brown fraction to ca.

30 ml and cooling to -40° gave light brown crystals. These were

filtered at -40° washed with chilled petroleum and dried in

vacuum (2 g, 20%). 61

Chapter II

TABLE E.2 Analytical data for neopentyl compounds

Compound Required Found C H Other C H Other

C20 H44 Ti 72.3 13.2 71.0 12.8

C2o Higi Zr 64.0 11.7 64.3 12.2

C15 H33 Cl2Ta 38.7 7.1 15.3 (C1) 38.7 7.1 15.0 (Cl) 38.9 (Ta) 38.7 (Ta) C13 H33 No 58.3 10.7 58.5 10.6

.. • , • . . • . •'.• • .• • • Tetrakis(neopentyl)titanium(IV)

To a solution of neopentyl lithium (80 mmol) indiethyl-

ether (100 ml) was added a solution of the etherate of TiCly

(20 mmol) in diethylether (100 ml) at room temperature over 30

minutes And the mixture stirred for 15 minutes. The solution was

filtered and the solvent stripped, and the resulting dark yellow oil extracted with petroleum. This solution was concentrated in vacuum to about 40 ml and then further concentrated by blowing off.solvent (to initiate crystallisation) until a substantial amount of dark crystalline material 'was formed. The solution was cooled to -10°, filtered and washed several times with chilled petroleum, yielding clean, yellow crystals of the compound (1 g, 30%).

Tetrakis(neopentyl)zirconium(IV)

As for the titanium tetra-alkyl, using ZrC14 (25 mmol).

The dark petroleum solution yielded an impure solid. Off-white crystals of the compound were obtained from this by recrystal- 62

lisation at -25° (2.0 g, 23%). Tetrakis(neopentyl)hafnium(IV)

As above, using HfCl4 (6.2 g, 20 mmol) in 50 ml of diethyl-

ether. A sticky, white solid was obtained, but could not be further

purified by recrystallisation (ca. 1 g, 11%).

Tris(neopentyl)dichlorotantalum(V)

To a solution of tantalum pentachloride (3.7 g, 10 mmol)

in diethylether (100 ml) was added a solution of neopentylmagnesium

chloride (30 mmol) in diethylether (40 ml). After filtering and

stripping the solvent, the yellow oil was extracted with petroleum,

concentrated to 50 ml and cooled to -30°. The pale yellow crystals

were filtered at this temperature and washed with chilled petroleum

(3 g, 75%). The solid could be readily purified further by subli- mationin vacuo at 200°.

Hexakis(neopentyl)dimolybdenum(III)

To a solution of neopentyllithium (30 mmol) in diethylether

(80 mlYwas added MoC15 (2 g, 6 mmol) in diethylether (50 ml) at room temperature over 15 minutes and the mixture stirred for 30 minutes. After filtering and stripping the solvent, the brown, sticky residue was extracted with petroleum. This solution was concentrated to about 5 ml and, after destroying the excess lithium reagent with a few drops of degassed methanol, was transferred to an alumina column and eluted with petroleum. The yellow fraction, on slow concentration in vacuum, gave crystals of the compound (0.3 g, 15%). The analogous tungsten Compound could only be prepared as an impure oil. 63

Chapter III

TABLE E. Analytical data for chromium(IV) compounds

Required Compound Found C H Cr C H

C20 H44 Cr 71.4 13.9 14.7 72.2 13.0 14.5 C10 H52 Cr 82.2 8.9 8.9 81.3 9.0 9.0

C ED H68Cr 88.9 6.3 4.8 88.6 6.4 5.0

The tetrakis(neopentyl)chromate(III) ion

To the stirred suspension of CrC13.3THF (1.4 g, 4 mmol) in

',THF (50 ml) at -40°, was added dropwise over 15 minutes a solution of neopentyllithium (16 mmol) in THF. An immediate reaction gave a dark blue-black solution which was allowed to warm to room temperature and filtered. The reaction appears to be quantitative.'

The solUtion may be oxidised to Cr(CH2CMe 3)4 by bubbling through it small quantities of air until the colour changes to a deep maroon, but the tetra-alkyl is more conveniently prepared as follows.

Tetrakis(neopentyl)chromium(IV)

To a stirred suspension of CrC13.3THF (1 g, 3 mmol) in diethylether (50 ml) was added dropwise at room temperature over ten minutes a solution of neopentylmagnesium chloride (13 mmol) in diethylether (25 m1)., The deep blue mixture was allowed to stir for one hour until it became maroon. The solutiort was then hydrolysed with saturated ammonium chloride solution (100 ml) and after separation, the ether layer dried (MgSO4), decanted and concentrated to ca. 5 ml. The solution was transferred to an alumina column and eluted with petroleum. The maroon fraction was 64

collected and concentrated to small volume and cooled to -30°.

Crystals of the compound separated and were collected at this temperature (0.3 g, 30%).

Tetrakis(neophyl)chromium(IV)

To a stirred suspension of CrC13.3THF (6 g, 15 mmol) in diethylether (100 ml) was added dropwise over 10 minutes at room temperature a solution of neophylmagnesium chloride (60 mmol) in diethylether (40 ml). The dark blue mixture was allowed to stir for

1 hour, then sufficient air was bubbled through the solution until it turned to a deep burgundy colour. This solution was then hydrolysed with saturated• ammonium chloride solution and'extracted into petroleum, dried, (Na2SO4) concentrated to 15 ml and trans- ferred to an alumina, column. The wine colour absorbed on the column and after sufficient petrol had been passed, down to wash out organic impurities the Column was eluted with toluene, and the coloured fraction collected, concentrated, and petroleum added.

On chilling to -60°C, crystals of the compound separated and were collected at this temperature (1.1 g :30%).

Tetrakis(tritylmethyl)chromium(IV)

To a stirred suspension of CrC13.3THF (1.8 g, 5 mmol) in

THE (100 ml) at -78°, was added as rapidly as possible, a solution of tritylmethyllithium (25 mmol) in THE (50 ml) at -78°. A reaction took place at once and the colour of the lithium reagent disappeared, but by the end of the addition, the bright red colour of excess lithium reagent predominated. The solution was allowed to warm to room temperature and air bubbled through until the colour changed to purple (2 or 3 mins.) The reaction mixture was hydrolysed and extracted into toluene and chromatographed on alumina eluting with 65 toluene. .Fine purple crystals of the compound were obtained by concentration of the toluene solution and addition of petroleum

(1.1 g, 20%). - - ...•.•.,• • •. Tetramethylchromium(IV)

To a suspension of freshly prepared methyllithium (ca.

2.0 g, 100 mmol) in petroleum (30 ml) was added a solution of

Cr(0But)4 (1.0 g, 3 mmol) in petroleum (20 ml). The mixture was allowed to stir at room temperature for 15 minutes and the maroon solution filtered. On distillation in vacuum the volatiles were collected at -78°C yielding a petroleum solution of the compound free from methyllithium.; Yield before distillation is probably quantitative, after distillation ca. 10-20%.

Petroleum-free CrMei, may be prepared by treating solid chromium(IV) t-butoxide with solid methyllithium at 30° in vacuo, collecting the volatile products, and separating the tetramethyl-, chromium by trap-to-trap distillation in vacuum.

Bis(1,3-dimethylenetetramethyldisiloxane)chromium(IV)

To a stirred solution of CrC13.3THF (1.2 g, 3 mmol) in

THE (100 ml) at room temperature were added dropwise over 15 minutes

5 mmol of the di-Grignard reagent in diethylether (25 m1). The reaction mixture was stirred for 15 minutes and hydrolysed with saturated ammonium chloride solution, extracted into petroleum and chromatographed on alumina. On removal of solvent in vacuo at room temperature a purple oil was obtained and characterised as the compound by its e.s.r. and electronic spectra.

66

ABBREVIATIONS

Me methyl

Et ethyl

Pri isopropyl t Bu tertiary butyl

Ph phenyl

nor 1-norbornyl

1-cam 2,31 31 trimethylbicyclo[2,2,1]hept-l-y1

4-cam 2,2,3,trimethylbicyclo[2,2,1]hept-l-y1

i.r. infra-red

u.v. ultra-violet

e.s.r. electron spin resonance

n.m.r. nuclear magentic resonance

g.l.c. gas-liquid chromatography

Temperatures are given in degrees Centrigrade unless otherwise stated.

67

REFERENCES

1. W. J. Pope and S. J. Peachy, Proc. Roy. Soc., 1907, 23, 86.

2. For references see a) F. A. Cotton, Chem. Revs., 1955, 55, 551. b) H. Zeiss, Organometallic Chemistry, Reinhold, 1960.

3. G. A. Razuvaev and V. N. Latyaeva Russ. Chem. Rev. (Uspekhi Khim.), 1965, 34, 251.

4. L. Summers, R. H. Uloth and A. Holmes, J. Amer. Chem. Soc., 1955, 77, 3604. 5. G. Wilkinson and T. S. Piper J. Inorg. and Nucl. Chem., 1956, 3, 104. 6. R. D. Closson, J. Kozikowski and T. H. Coffield, J. Org. Chem., 1957, 22, 598.

7 • G. W. Parshall and J. J. Mrowca, Adv. Organometal. Chem., 1968, 7, 157. 8. H. J. Berthold and G. Groh, Z. anorg. Chem., 1963, 319, 230.

9. C. Beerman and K. Clauss, Angew. Chem., 1959, 71,,627. 10.' G. Wilke and H. Schott, Angew, Chem. Int. Ed., 1966, 5, 583. 11. v. Zucchini E. Albizzati and V. Gianini, J. Organometallic Chem., 1971, 26, 357.

12. V. N. Latjaeva, G. A. Razuvaev, A. V. Malisheva and G. A. Kiljakova, J. Organometallic Chem., 1964, 2, 388.

13. e.g. C. Beerman and H. Bestian, Ang. Chem., 1959, 71, 618.

14. G. L. Juvinall, J. Amer. Chem. Soc., 1964, 86, 4202. 15. E. Kurras, Angew. Chem.,1960, 72, 635. 16. B. Sarry and V. Dobrusskin, Angew. Chem., 1962, 74, 509. 17. E. Kurras and J. Otto, J. Organometallic Chem., 1965, 4, 114. 18. B. Sarry, M. Dettke and H. Grossman, Z. anorg. Chem., 1964, 329, 218. 19. H. H. Zeiss and R. P. A. Sneedon, Angew. Chem. Int. Ed., 1967, 6, 435-

68

20. V. N. Latyaeva, L. I. Vyshinskaya, V. B. Shur, L. A. . • Fyodorov and M. E. Volpin, J. Organometallic Chem., 1969, 16, 103. 21. S. J. Ashcroft and C. T. Mortimer, J. Chem. Soc. (A), . • . • . . • 1967, 930.

22. K. W. Egger, J. Organometallic Chem., 1970, 24, 501. 23. V. I. Tel'noi, I. B. Rabinovitch, V. P. Tikhonov, V. N. Latyeova, L. I. Vyshinskaya, G. A. Razuvaev, Doll. Akad. Nauk., (U.S.S.R.), 1967, 174, 1374. [Chem. Abs. 68, 43809s] 24. N. M. Chirkov, Kinetics and Catalysis (Kinet. i Katal.) 1970, 11, 269. 25. See work by D. C. Bradley and various co-workers. 26. F. A. Cotton and R. M. Wing, J. Organometallic Chem., 1967, 9, 511.

27. C. L. Lau, Rec. Tray. Chim., 1965, 84, 429. 28. H. H. Eysel, H. Siebert, G. Groh and H. J. Berthold, Spectrochim. Acta, 1970, 26A,1595.

29. e.g. P. S. Hallman, B. R. McGarvey and G. Wilkinson, J. Chem. Soc. (A) 1968, 3143. 30. a) M. D. Johnson and N. Winterton, J. Chem. Soc. (A), 1970, 507. b) K. Thomas, J. A. Osborne, A. R. Powell and G. Wilkinson, J. Chem. Soc. (A), 1968, 1801.

c) M. Ardon, K. Woolmington and A. Pernick, Inorg. Chem., 1971, 10, 2812. d) C. T. Loo, L-Y. Goh and S. H. Goh, J.C.S. Dalton, 1972, 585- •.e 31. a) R. Tabacchi and A. Jacot-Guillarmod, Chimia, 1970, 24, 271. b) R. J. H. Clark and A. J. McAless, J. Chem. Soc. (A), 1970, 2026. ... - c) G. W. A. Fowler, D. A. Rice and J. D. Wilkins, J. Chem. Soc. (A),, 1971, 1920.

69

32. A. Tzschach and H. Nindel, J. Organometallic Chem., 1970, 24, 159.

33. A. J. Shortland and G. Wilkinson, Chem. Comm., 1972, 318. 34. W. Mowat, A. J. Shortland, N. J. Hill and G. Wilkinson, - - J.C.S. Dalton, in press.

• .-..• . ••• 35. W. Kruse, J. Organometallic Chem., 1972, 42, C39. 36. B. K. Bower and H. G. Tennent, J. Amer. Chem. Soc., 1972, 94, 2512.

37. W. Mowat, A. Shortland, G. Yagupsky, N. J. Hill, M. Yagupsky and G. Wilkinson J.C.S. Dalton, 1972, 533. 38. G. Wilkinson, Pure and Applied Chem., 1972, 30, 627.

39. P. S. Braterman and R. J. Cross, J.C.S. Dalton, 1972, 657. . . , 40. R. P. A. Sneeden and H. H. Zeiss, J. Organometallic Chem., 1970, 22, 713. 41. R. P. A. Sneeden and H. H. Zeiss, J. Organometallic Chem., 1970, 26, 101.

42. R. P. A. Sneeden and H. H. Zeiss, J. Organometallic Chem., 1971, 27, 89.

43. G. M./Whitesides, I. R. Stedronsky, C. P. Casey and J. S. Filippo, Jr. J. Amer. Chem. Soc., 1970, 92, 1426.

44. J. Chatt and B. L. Shaw J. Chem. Soc., 1960, 1718. 45. J. Dvorak, R. J. O'Brien and W. Santo, Chem. Comm., 1970, 411. . _ . 46. a) M. Tamura and J. K. Kochi, J. Organometallic Chem., 1971, 29, 111.

- ••• • . . - , b) M. Tamura and J. K. Kochi, Bull. Chem. Soc., Japan, 1971, 44, 3063.

47. G. A. Razuvaev, V. N. Latyaeva, G. A. Kilyakova, G. Ya. Mal'kova, Dokl. Akad. Nauk., (U.S.S.R.) 1970, 191, 620. .-••••• • [Chem. Abs., 73, 14956K].

48. M. R. Collier, M. F. Lappert and M. M. Truelock, J. Organo- . . .. • . • . • •'. •'.

metallic Chem., 1970, 25, C36.

70

49- W. Schafer and K. H. Thiele, Z. anorg. Chem., 1971, 381, 205. 50. U. Zucchini, E. Albizzati and U. Giannini, J. Organometallic Chem., 1971, 26, 357. - ..... - ..... 51. J. J. Felten and W. P. Anderson, J. Organometallic Chem., 1972, 36, 87. 52. S. D. Ibekwe and J. Myatt, J. Organometallic Chem., 1971, 31, C65.

53. E. Kinsella, V. B. Smith and A. G. Massey, J. Organometallic Chem., 1972, 34, 181.

54. F. Huq, W. Mowat, A. Shortland, A. C. Skapski and G. Wilkinson, Chem. Comm., 1971, 1079.

55. D. A. Owen and M. F. Hawthorne, J. Amer. Chem. Soc., 1971, 93, 873- 56. e.g. for methyl, R. B. King, Organometallic Synetheses, 1965, 1, 147.

57. J. Chatt and B. L. Shaw, J. Chem. Soc., 1959, 705. 58. G. Wilke and H. Schott, Angew. Chem. Int. Ed., 1966, 5, 583.

59. C. S. Cundy, M. Green and F. G. A. Stone, J. Chem. Soc. (A), 1970, 1647. 60. B. Wozniak, D. Ruddick and G. Wilkinson, J. Chem. Soc. (A), 1971, 3116. 61. W. Mowat and G. Wilkinson, J. Organometallic Chem., 1972, 38, C35.

62. E. C. Alyea and D. C. Bradley, J. Chem. Soc. (A), 1969, 2330. 63. J. Selbin, Chem. Rev., 1965, 65, 153. 64. F. Huq, W. Mowat, A. C. Skapski and G. Wilkinson, Chem. Comm., 1971, 1477. - 65. J. Krausse, G. Marx and G. Schl3dl, J. Organometallic Chem., 1970, 21, 159.

66. M. J. Bennett, J. V. Brencic and F. A. Cotton, Inorg. Chem., 1969, 9, 1060.

67. D. Lawton and R. Mason, J. Amer. Chem. Soc., 1965, 87, 921. 71

68. F. A. Cotton and C. B. Harris, Inorg. Chem., 1967, 6, 924.

69. F. A. Cotton, Accounts Chem. Res., 1969, 2, 240 and references quoted. 70. I. W. Bassi, G. Allegra, R. Scordimaglia and G. Chioccola, . ... " . • . .•. J. Amer. Chem. Soc., 1971, 93, 3787.

71. G. R. Davies, J. A. J.• Jarvis, B. T. Kilbourn and A. J. P. Pioli, Chem. Comm., 1971, 677.

72. R. E. Tapscott and R. L. Belford, Inorg. Chem., 1967, 6, 735.

73. A. J. Shortland, Ph.D. Thesis, University of London, 1972. 74. a) K. A. Wilhelmi and 0. Jenssen, Acta Chem. Scand., 1961, 15, 1415. b) B. L. Chamberland, Mat. Res. Bull., 1967, 2, 827.

75. a) H. C. Clark and Y. N. Sadana, Canad. J. Chem., 1964, 42, 50.

b) .G. Siebert and W. Hoppe, Naturwiss., 1971, 58, 95.

76. E. C. Alyea, J. S. Basi, D. C. Bradley and M. H. Chisholm, J. Chem. Soc. (A), 1971, 772.

77. J. S. Basi, D. C. Bradley and.M. H. Chisholm, J: Chem. soc. (A), 1971, 1433. 78. N. J. Hill, Personal Communication.

79. E. Grovenstein, Jr. and L. P. Williams, Jr., J. Amer. Chem. Soc., 1961, 83, 412. 80. T. Tsuda and J. K. Kochi, Bull. Chem. Soc., Japan, 1972, 44, 648...... • .• 81. H. H. Zeiss and R. P. A. Sneedon, Angew. Chem. Int. Ed., 1967, 6, 435. . - 82. J. S. Griffith "The Theory of Transition Metal Ions", Cambridge University Press, London, 1961, p. 292 and 308.

83.' a) J. R. Norris and S. I. Weissman, J. Phys. Chem., 1969, 73, 3119. b) J. H. Freed, G. V. Bruno and C. Polnaszch, J. Chem. Phys., 1971, 55, 5270. 72

84. D. C. Bradley and M. H. Chisholm J. Chem. Soc. (A), 1971, 2741. 85. N. J. Hill, Personal Communication. 86. G. A. Ward, W. Kruse, B. K. Bower and J. C. W. Chiln, J. Organometallic Chem., 1972, A?, C43. 87. W. T. Smith, Jr. and J. T. Sellas, Org. Synth., 1963, Coll. Vol. IV, p. 702.

88. H. E. Zimmerman and F. J. Smentowski, J. Amer. Chem. Soc., 1957, 79, 5455. 89. H. L. Krauss and G. Munster, Z. anorg. Chem., 1967, 352, 24. 90. F. C. Whitmore and L. H. Somme r, J. Amer. Chem. Soc., 1946, 68, 481. 91. H. L. Lewis and T. L. Brown, J. Amer. Chem. Soc., 1970, 92, 4664. 92. H. Zimmer, I. Hechenbleinker, O. A. Homberg and M. Danzik, J. Org. Chem., 1964, 29, 2632.

93. G. M. Whitesides, E. J. Panek and E. R. Stedronsky, J. Amer. Chem. Soc., 1972, 94, 232.

94.. E. C. Alyea, J. S. Basi, D. C. Bradley and M. H. Chisholm, Chem. Comm., 1968, 495.