SOME STUDIES ON CYCLOPENTADIFNYL AND

CARBONYL COMPLEXES OF TRANSITION METALS.

A Thesis submitted by

JON ARMISTICE McCLEVERTY, B.Sc.

for the Degree of

Doctor of Philosophy of the

University of London.

Royal College of Science. May 1963. This Thesis I dedicate to

DIANNE.

Her love and understanding is a

continual inspiration. kit

CHAPTER II

ic-CYCLOPENTADIENYL METAL HYDRIDES

THE DI-n-CYCLOPENTADIENYL HYDRIDES OF TANTALUM MOLYBDENUM. AND TUNGSTEN.

The hydride complexes are prepared in high yields by the inter-

action of the anhydrous metal chlorides (Noel" 1C16, or TaC15) with a

solution of sodium cyclopentadienide in tetrahydrofuran containing an

excess of sodium borohydride. Without addition of borohydride to the

sodium cyclopentadienide solution only low yields of the hydrides can be

obtained since the hydriue ion required must come from the solvent or from

traces of excess of cyclopentadiene or cyclopentene present in the

reaction mixture.

The tantalum hydride, (n-C5H5)2TaH3, is a white crystalline

compound which is sensitive to air; the molybdenum and tungsten hydrides,

(n-05H5)2MH2, form yellow crystals, the former being more sensitive to air

than the latter, which can be handled briefly in air. The tantalum

compound is sparingly soluble in light petroleum but it is moderately

soluble in benzene; the other hydrides are soluble in light petroleum as

well as in other common solvents. Solutions of all three compounds

decompose rapidly wren exposed to air. The compounds are soluble in

halogenated solvents and in carbon disulphide but react with them quite

rapidly, the order of reactivity being Ta Mo ACKNOWLEDGMENTS

I would like to express my deep gratitude to

Professor G. Wilkinson for his constant encouragement, patience, and guidance during his supervision of this work.

I am very grateful to Dr. D. F. Evans, Dr, R. Mason, and particularly Ir. L. Pratt, for their invaluable assistance.

I am indebted to my colleagues and friends in the lab. for their continual help and advice, both academic and otherwise. I would also like to thank my mother for helping to ensure that I got as far as this.

My thanks are due to Miss C. M. Ross for her care in typing this thesis.

The award of a scholarship from the Carnegie Trust for the Universities of Scotland, for the period 1960-3963, is gratefully acknowledged. ABSTRACT

The di-I-cyclopentadienyl hydrides of molybdenum, tungsten and

tantalum have been prepared. High-resolution proton magnetic resonance and infrared spectra of these compounds and of the trihydride cations of molybdenum and tungsten have been measured. For the trihydride species of tungsten and tantalum the patterns of the high-field lines show that

there are two equivalent hydrogen atoms in an A2B grouping. This

observation is discussed with reference to the base character and to the

molecular configuration of the di-n-cyclopentadienyl metal hydrides.

High-resolution proton magnetic resonance spectra of hydrido, methyl and ethyl derivatives of transition-metal carbonyl and

%-cyclopentadienyl carbonyl derivatives have been measured, and the data compared with those for non-transition-metal compounds. Infrared

spectra of some hydrides and corresponding deuterides are given. An improved preparation of Re2(CO)to is given and the new complexes

7c-05H5Ru(C0)2X (X = H, D, I, CH3 and C2H5) are described.

The preparation of some propionyl derivatives of the type

II-05H5M(CO)nC0C2H5 (M = Mb, W, Fe, Ru; n = 2, 3) is described. A

novel alkyl transfer reaction in 7c-05H5MO(Q0)3C2H51 where the ethyl group

migrates from the metal to the cyclopentadienyl ring, is also described.

Reaction of n-05H5Rh(C0)2 with CF3I, C2F5I and C3F7I affords

the compounds n-05H5Rh(CO)RfI (Rf = perfluoroalkyl radical). Their

high-resolution fluorine magnetic resonance spectra are discussed.

TABLE OF CONTENTS

Chapter Page

I Introduction 1

II The Cyclopentadienyl Hydrides 12

III Transition—metal Alkyls and Related Derivatives 32

IV Rhodium Perfluoroalkyl Compounds 50

V Experimental 61

References 79 CHAPTERI

INTRODUCTION

Metal hydrides can be classified broadly into four main groups.

These are:

(a) salt-like hydrides of which LiH and CaH, are examples;

(b) covalent transition-metal hydrides such as HMn(C0)5,

(7C-0 5H5 )2WH2, and (Et3P)2PtHC1; (c) complex hydrides like LiA1H4 and NaBH4; and

(d) other hydrides. These include the polymeric BeH2 and MgH2,

CuH, the transition-metal binary hydrides such as PrH3 and UH3 and the non-stoichiometric compounds like TiE1 .7 and the "Weichfelder Hydrides".

The hydrides H2Fe(C0)3 and HCo(C0)41'2 were discovered in 1931

but studies in this field have greatly intensified in the last four years.

Now a large number of transition metals, complexed with a variety of ligands, have been shown to form covalent molecular hydrides and Table 1 displays a selection of these.

At first it seemed that strong-field ligands such as '3P, CO,

CN, or 7-05H5 were necessary to stabiliso the metal-hydrogen bond but recently the existence of hydridic species in systems containing ethylenediamine, ammonia, pyridine or dimethylglyoxime has been 6486 demonstrated. Unusual compounds of rhenium (and technetium), the "rhenohydrides" or "rhenides", have been prepared and their 5051 hydridic nature claimed. They may be analogous to the complex 2.

hydrides of group (c) but they are difficult to prepare, purify, and characterise.

PREPARATION

Transition-metal hydrides can be prepared in a variety of ways. These include:

(1) acidification of sodium metal carbonylates, e.g.

H3PO4 Re2(C0)10 --> NaRe(C0)5 HRe(C0)5 ,. T.H.F. HC1 t-05H5W(C0)3Na n-05H5W(C0)3H, H2O H3PO4 NaV(C0)5(V3P) mr(co)5(140,

(2) sodium borohydride or lithium aluminium hydride reduction of metal halides, e.g. NaBH4 n-05H5Ru(C0)2I > n-05H5Ru(C0)2H, T.H.F. NaBH4. (03P)21An(NO)2Br > (4v3 P)2Mn(NO)2H, T.H.F. LiA1H4 (0'31))3IrliC12 (03P)3IrH3, Et20 NaBH4 [Rh(en)2C12] + [Rh(en)2H2]+9 Et0H

(3) by treatment of phosphine or arsine complexes with hypophosphorous acid, e.g. H(H2P02) (Et20As)3RhC13 3 (Et 20AS )3 RhEC12

3

(4)by treatment of phosphine or arsine complexes with hydrazine hydrate in aqueous solution or aqueous ammonia, e.g.

trans-(Et3P)2PtC1, N2H trans-(Et3P)2PtHC1, - H 20

(5) by treatment of phosphine or arsine complexes with KOH in

ethanol or a polyalooholy e.g.

Bt 20 „ KOH (NH4)20sC16 5 (Et2OP)30sC13 (Et20)30s(CO)HC1, • EtoH KOH (0'3P)3RuC13 (V3P)3Ru(CO)HC1, MeOCH2OH

(6) by direct synthesis from hydrogen and either the pure, finely

divided metal or metal sulphides, e.g.

250 atm. Co + 4C0 2H2 HCo(CO), 180°

(7) by sodium borohydride or sodium amalgam reduction of complex metal

cyanides in aqueous or ethanolic solution, e.g.

2+ Na/Hg Co + CN > (HCo(CN)03- 1 H2O

(8) by protonation in very strong acids, e.g. H2SO4 Fe(C0)5 HFe(C0)5+, BF3,H20 I-05H5W(C0)3H it-05H5W(C0)3H24-1 CF3 COOH HPF6 + (,t-05115)2MoH2 (1-05115)2MoH3 9 H2O HPF6 %-05H5FeMn(C0)7 %-05H5FeMn(C0)7e. H2O 4.

TABLE 1

Some Transition-metal Hydrides

M.p. i J, Compound Colour 7- H liEl References oc ; c. p.s. cm

-05H5)2TiBH4 violet para 1942 55 Ti" 7t-05H5Cr(C0)3H yellow 10 15.46 19,15,18, 87. H2Cr(c0)5 white 9,21,22, 42,53,70. C6H6Cr(C0)3114- yellow soln. 13.55 77

7t-05H5Mo(C0)3H pale yellow 54 15.52 1790 14,19,18, 87. (g-05H5)2MoH2 pale yellow 183-185 18.76 1847 32,69 (n-05R5)2MoH3+ white 16.08 1915 52,69,87 [Mo(C0)6(OH)3H3] yellow 40,41 [n-05H5NO(C0)3]2H+ red brown 30.99 77 soln. ic-05H5W(C0)3H pale yellow 69 17.33 37.7 1845 14,18,19, 87. 7c-05H5W(C0)3H2+ yellow soln. 11.93 77 (n-05H5),WH2 yellow 163-165 22.28 73.2 1896 32,69 (7c-05H5)2WH3+ white 16.08 47.8 1943 52,69,87 16.44 [(I-C15115)jiMo(C0)6De red brown 32.88 77 soln. HMn(00)5 c'less -25 17.50 1728 8,28,29, 44,45. HMn(C0)4(P) yellow 138 1790 82 HMn(N0)2(031))2 yellow 153-154 26.8 32.0 83 H[Mn(C0)4]2P(2 yellow 154-155 84 U2Mn2(C0)9 dark red para 71 06R9Mh(C0)3H orange yellow 88 Mp.. Colour J MH Y References Compound °C c.p.s. cm-1

HTc(C0)5 c'less soln. 79. K2TcH8 white 51.

HRe(C0)5 c'less soln. ca. 100 15.66 1820 7,37,38, 87,78. K2ReH8 white 20.05 1846 36,50,51. (it—05H5)2ReH lemon yellow 161-162 22.80 2030 12,26. (7t—05H5)2ReH2+ white 22.90 2055 12,26,87. (933P)3R0H3 red 155 dia. 2000 65. (0'313)4Re113 yellow 147 dia. 2050 66.

HaFe(C0)4 yellow —70 21.10 11,17,33. HFe3(C0)11 cherry red 24.90 11,17,33. H2Fe3(C0)12 cherry red '24.90 11,17,33. n—05H5Fe(C0)21-1 yellow liq. —10 21.91 1835 32,87. (n—C5H5)2FeH pale green 11.2 46,31. HFe(C0)54. green soln. 18.15 77. + liFe(C0)403P yellow soln. 17.75 35.31 77. HFe(C0)3(0'313)2+ yellow soln. 18.11 77. En—05H5FeMn(C0)7W orange 38.0 1760 77. [C21-14(Et2P)2]2FeHC1 red 155-156 43.60 1840 48,59. [o—C6H4(Et2P)212FeHC1 red 221 40.50 1870 48,59. [o—C6H4(Et2P)2]2FeH2 c'less 22.9 1726 48,59.

6.

M.p. 1- J Compound Colour g I'lli Roforences oC H c.p.s. cm-

7(-05H5Ru(C0)2H yellow liq. -10 20.92 1853 87. (n-05H5),Rue yellow soln. 16.2 46. [n-05H5Ru(C0)2]2H+ red brown 28.58 77. soln.

(Et20)3Ru(CO)HC1 c'less 102-104 16.9 1901? 49. [C2114(Et2P)2]2RuHC1 c'less 175-176 31.6 1938 30,57. [C2H4(Et2P)2]2Ru(MOH c'less 1884 30. [C2114(Et2P)2]2Ru(Et)H pale yellow 1873 30. [o-C6114(Et2P)2]2RuH2 yellow 277-2781 1617 30.

(7c-05H5)20se not 46. obs. (c13P)30s(CO)HC1 c'less 171 2100 58. [C2H4(Et2P)020sHC1 c'less 171-172 35.8 2039 30,57. [C2H4.(Et2P)2]2.0s(SCN)H c'less 200 2009 57. [C2114.(Et2P)2]20sH2 lemon yellow 150-155 1720 57.

HCo(C0)4 yellow -26 20.05 1934 3,4,5,10, 16,17,13, 68,81. HCO(ON) 3- yellow soln. 26.46 27,43. HCo(C0)3(03P) yellow 67,80.

HRh(CN)53- c'less soln. 20.3 26.0 43. trans4Rh(en)2112)2+ brown 31.0 31+1 2100 64,86. cis-[Rh(en)2H2]2+ brown 31.0 31+1 2093 86. cis-Dh(trien)HC1)2+ brown 28.5 27.0 2081 86 1. Compound Colour Mop. H 3MH //.11., References C c.p.s. cm—J-

Rh(NH3)5H2+ brown 2079 86. trans-[Rh(dmg)2HC1] green 26.4 86. trans-Elih(py)4HC1]+ brown 28.6 86. (V2MeAs)3RhHC12 yellow 172-175 2077 56.

Hir(co)4 6. HIr(CN)53- c'less soln. 23.7 47. (0302Ir(CO)HC12 white 315-320 2245 63. (03A8)2Ir(co)Hc12 pale yellow 249-252 2200 63. (913P)3IrHC12 yellow 256 2230 60. (03As)3IrHC12 yellow 240 2170 60. (03Sb)3IrHC12 yellow orange 201 2100 60. (fi3P)3IrH2C1 white 250 2215, 60. 2110 (Et3P)3ITHC12 yellow isomer 81-84 22.4 2114 49. white isomer 99-100 31.5 2213 49. (91303IrH3 'less isomer 173 2130, 61,62. 1750 e'less isomer 152-153 2075 61,62.

H2Ni(C0)6 dark red 73. (131.3p)2xill01 diamag. 32.

(Et3P)2PdHC1 Oless 2035 24. 8.

JNE VH Compound Colour M.p. 1- References 00 H c.p.s. cm-1

EPt(CN)53— c'less soln. 22.15 1800 47. (Et3P)2PtHC1 c'less 80-81 26.9 1276 / 2183 76. (14.5)r (Et3P)2PtH(NO2) c'less 95-98 29.7 /276/ 2150 76. (16.5) (Et3P)2PtH(NO3) c'less 47-49 33.8 1276 / 2242 76. (15.5)r (Et3P)2PtH2C12 c'less 82-100 2254, 76. 2265

HIT(C0)6 yellow 74,85. HV(C0)5(0'313) yellow 85.

(7t—05H5)2TaH3 white 187-189 11.63 1735 69,73. 13.02

—H high—field line seen but position not measured 9.

PROPERTIES

In general, the hydrides containing phosphine or arsine ligands are stable thermally and towards aerial oxidation. The least stable hydrides are those containing only carbonyl or cyanide ligands. As the majority of these complexes are diamagnetic, the presence of a hydridic Proton can be deMonstrated using proton magnetic resonance techniques.

All known diamagnetid transition-metal hydrides show characteristic resonance lines at fields greater than the tetramethylsilane resonance

(e.g. T10.,50). Frequently spin-spin coupling due to ligands (e.g. "POO or metals (e.g. 183w) of spin i gives rise to multiplet structure on these high field lines; metal-hydrogen coupling constants are regarded as evidence for a direct metal hydrogen bond. Metal-hydrogen stretching frequencies are observed in the infra-red in the region 1700-2400 cm.-1, the abSorption being intense in the case of the phosphine hydrides but weak in carbonyl and cyclopentadienyl-carbonyl complexes.

The hydrides show.a variety of acidic and basic properties.

HCo(C0) 28 and HV(C0)585 are strong acids but HV(C0)5(03P), 85 28 RMn(C0)5 and H2Fe(C0) 28 are weaker. 7E-05H5W(C0)3H exhibits both acidic and basic properties. It can be protonated in strong acid media , +77 to give w-05H5W(C0)3H2 but also acts as a very weak acid in aqueous 118 media. (n-05H5)2ReH, and (n-05H5)2Mo/WH2 have basic properties, the rhenium species having a pKa value close to that of ammonia.26'69

Most of the phosphine and arsine hydrides are considered to be neutral 76 although (Et3P)2PtHC1 can absorb a molecule of HC1 forming (Et3P)2PC112C12.

10.

However, this complex does not appear to dissociate and may be regarded as an octahedral Pt Wcompound. Iridium complexes undergo a similar reaction, e.g. (21 313)2Ir(CO)C1 + HCl ---4(913P)2Ir(CO)HC12.119

Most hydrides react readily with halogenated hydrocarbons forming the corresponding halides.

CC14 76 ((f3P)2PtHC1 (913P)2PtC12 + CHC13 3 CH I 35 (n-05H5)2WH2 (%-c5115)2w12 + M14_

Both cobalt and iron carbonyl hydrides act as reducing agents, e.g. 13 acetone being converted to isopropanol. It has been suggested that

HCo(C0)4 is the catalyst in the hydroformylation reaction.

A number of transient hydrides have been suggested as inter- mediates in homogeneous reactions with molecular hydrogen in aqueous solution. Kinetic studies on the system cuprous acetate and quinoline support this view.120,121

STRUCTURES

Only a few compounds have been studied crystallographically and it seems that the hydrogen atom occupies a coordination position around the metal. (Et3P)2PtHBr has been examined123 and shown to have the expected square-planar configuration, with the hydrogen atom (not detected) trans to the bromine atom. (V3P)30s(CO)HBr is octahedral but 122 the hydrogen atom was not found; the three phosphine ligands and the hydrogen atom are in the same plane, with the Br and CO above and below

the plane. 11.

(7g-05H5)21.1oH2 is a wedge-shaped molecule where the hydrogen atoms lie in a plane parallel to and between the planes of the cyclopentadienyl 124 rings. The metal hydrogen distance is estimated at ca. 1.2 X. It is apparent that this hydride has a trigonal-pyramidal configuration where a sterically active lone-electron-pair occupies one of the coordination positions. 112Fe(COL. was studied39 using broad-line proton magnetic resonance techniques and the workers concluded that the hydrogen atoms were 81 about 1.1 A distant from the iron. Recently Lipscomb and co-workers have calculated the theoretical chemical shift of the proton in HCo(C0)4 using approximate covalent radii for the H and Co atoms, and have found striking agreement with the experimental value. They suggest that the high diamagnetic shielding of the proton is due to the 4s and 42 orbitals and not to the close proximity of the proton to the Co atom and its 3d electrons. 12.

CHAPTER II

%-CYCLOPENTADIENYL METAL HYDRIDES

THE DI-N -CYCLOPENTADIENYL HYDRIDES OF TANTALUM, MOLYBDENUM, AND TUNGSTEN.

The hydride complexes are prepared in high yields by the inter- action of the anhydrous metal chlorides (MoC15, WC16, or TaC15) with a solution of sodium cyclopentadienide in tetrahydrofuran containing an excess of sodium borohydride. Without addition of borohydride to the sodium cyclopentadienide solution only low yields of the hydrides can be obtained since the hydriue ion required must come from the solvent or from traces of excess of cyclopentadiene or cyclopentene present in the reaction mixture.

The tantalum hydride, (v-05H5)2TaH3, is a white crystalline compound which is sensitive to air; the molybdenum and tungsten hydrides,

(n-05H5)2MH2, form yellow crystals, the former being more sensitive to air than the latter, which can be handled briefly in air. The tantalum compound is sparingly soluble in light petroleum but it is moderately soluble in benzene; the other hydrides are soluble in light petroleum as well as in other common solvents. Solutions of all three compounds decompose rapidly wren exposed to air. The compounds are soluble in halogenated solvents and in carbon disulphide but react with them quite rapidly, the order of reactivity being Ta > Mo >W. 13.

Like (7c-05H5)2Rell, the molybdenum and tungsten compounds behave 52 as bases and are soluble in dilute aqueous acids and in trifluoroacetic + acid, to give the ions (n-05H5)2MH3 with gaseous hydrogen chloride or bromide the white deliquescent halide salts are formed. The tantalum hydride does not appear to behave as a base, even in trifluoroacetic acid, and is decomposed by aqueous acids.

Chemical Properties. As was mentioned, all of the hydrides react with halogenated hydrocarbon solvents. Dichloromethane reacts much more slowly than carbon tetrachloride while chloroform appears to react very slowly; the proton magnetic resonance spectrum of (n-05H5)2WH2 in chloroform showed the characteristic high-field hydride line which remained unchanged after 5 hours. 18 It had previously been found that other hydrides (e.g. n-05H5Mo(C0)311) reacted easily with carbon tetrachloride, N-bromosuccinimid and methyl iodide to form the corresponding halides (i.e. n-05H5Mo(C0)3C1, tc-05H5Mo(CO)3Br, and n-05H5Mo(C0)3I). It seemed probable that the reaction between halogenated hydrocarbons and the di- and tri-hydrides would give similar compounds.

Carbon tetrachloride reacts quite rapidly in the cold and very rapidly when heated with the hydrides to form light green, insoluble solids whose empirical formula is (n-05H5)2MCln .131

(n-05115)2MBri + CC14 (n-05H5)2MCln + chlorinated hydrocarbons.

With methyl iodide, on refluxing under nitrogen for several hours, the hydrides form dark-green iodides (7c-05H5)2MI25 These are sparingly soluble in acetone, from which they are recrystallised. 14.

Attempts to prepare the methyls, (71-05145)2M(0113)n, by reacting the halides with methyl lithium in tetrahydrofuran, diethyl ether or benzene failed, owing probably to the insolubility of the halides.

Reaction of the hydrides with diazomethane failed to give the methyls, polymethylene being the major product.

Spectroscopic Studies. The most significant spectral informa- tion bearing on the structure of these hydrides comes from high-resolution nuclear magnetic resonance studies. The data are presented in Table 2, and the high-field region for (7-05H5)2TaH3 and (n-05H5)2WH3 is shown in

Fig. 1. The spectra of all of the species are similar to those of + (n-05H5)2ReH and (n-05H5)2ReH2 and also to those of the ferrocenonium ion, (7c-05H5)2FeH 1 and its ruthenium and osmium analogues which were 46 reported to be formed when the neutral molecules (n-05H5)21/1 are dissolved in boron trifluoride hydrate. In all these cases there are two main groups of resonance lines, one at low and one at high fields relative to tetramethylsilane. The low-field line of relative intensity 10 is due to the ton equivalent protons of the n-05H5 groups. The bands from molybdenum and tungsten dihydride show fine structure due to electron- coupled spin-spin interaction with the hydrogen atoms bound to the metal; this splitting is of the same order, ca. 1 c.p.s., as that in

(n-05115)2ReH and (at-05H5)2FeH. The high-field lines occur in the region characteristic for hydrogen atoms bound to transition-metal atoms and the areas, relative to the low—field lines, taken as 10, confirm the stoicheiometries of the various species. 15.

TABLE 2

4igh -resolution Proton Magnetic Resonance Spectra (at 56.45 Mc./sec.)

Hydride line n-C5H5 line Splitting (c.p.s.) Slittinq (c.p.s.) Compound l by other by n-05H5 1-. by hydride hydride protons protons

(n-05H5)2TaH3 a 11.63 d (B)% 9.6 13.02 (A2) 5.24

(n-05H5)2MoH2 a 18.76 - Not resolved 5.64 0.96 Line width (triplet) 1.6 c.p.s. (n-05H5)2MoH71 + b 16.08 - Not resolved 4.38 Line width 1.4 c.p.s. (n-05H5)2WH2 a 22.28e - 0.75 5.76 0.75 "septuplet") (triplet) (n-05H5),WHs+ c 16.08 (B) d Not resolved Not resolved 8.5 16.44 (A2) Line widths 4.39 Line width <2.1 c.p.s. 1.4 c.p.s.

ab In benzene. In trifluoroacetic acid. In concentrated hydrochloric acid, reference Me3C.OHy for which 1- = 8.69. Calculated by using data tabulated by Corio (Chem. Rev., 1960, 60, 363). e The hydride lines of the tungsten compounds show weak satellite doublets from those molecules containing the isotope 183W with spin I- (abundance 14.3%). The tungsten- proton splitting is 73.2 c.p.s. for the dihydride and 47.8 c.p.s- for the A2 protons in the trihydride cation. 16.

For the dihydride species of rhenium, molybdenum, and tungsten

there is only one high-field proton resonance line, as in (7c-05H5)2ReH

and the ferrocenonium ion. For the rhenium and tungsten compounds, the fine structure on these lines is due to coupling with the 7t-05H5 ring

protons although all the expected components (eleven) cannot be resolved.

It thus appears that the two hydrogen atoms in (7t-05H5)2MH2 are in equi- valent environments. This is not true, however, of the three hydrogen atoms in (7t-05H5)2TaH3 and (n-05H5)2WH3 since in these cases the high-field line patterns (Fig. 1) are characteristic of A2B groupings89 with two equivalent protons and one non-equivalent one. The cation

(7t-05H5)2MoH3 in water or in trifluoroacctic acid shows only a single line of relative intensity 3, which suggests that in this case there is a slow exchange with the protons of the solvent; this exchange also occurs in benzene-trifluoroacetic acid mixtures. A solution of (7(-05H5)2WH2 in

D20-DC1 mixtures shows no high-field line, again indicating a slow exchange with the solvent which is complete within a few minutes. In neither the molybdenum or tungsten case is the exchange rate fast enough to make the solvent and hydridic proton lines coalesce.

The infrared spectra of the neutral hydrides have been measured -1 -1 down to 750 cm. and measurements below 750 cm. are given elsewhere.90

The metal-to-hydrogen bond stretching frequencies occur in the region -1 around 2000 cm. .; the band in (7(-05H5)2TaH3 is exceptionally intense, compared with other metal-hydrogen stretching frequencies. Even under the highest available resolution on a grating instrument only one symmetrical band is found for the M-H absorption in molybdenum and tungsten dihydrides, but for (71-05H5)2TaH3 this band has weak shoulders.

18.

FIG. 1

High-resolution nuclear magnetic resonance spectrum in the high-field region of

(a) (n-05115)2WH3+ and (b)(n-05H5)2TaH3. 19.

Discussion. The High-resolution nuclear magnetic resonance

spectra of (n-05H5)2ReH(D) and (n-05H5)2ReH2(D2)+ have shown that the

structure with the hydrogen atom(s) lying between the metal atom and the

n-05H5 rings which had been proposed92 for these compounds (and which has recently been suggested46 as a possibility for the ferrocenonium ion) is unlikely. The main unsettled question in this work was whether the

metal-to-ring axes are linear, as in ferrocene, or are angular. This problem has also been considered by others91 in the light of dipole moment and far-infrared data on the rhenium, molybdenum, and tungsten species.

The observations of the non-equivalence of the hydrogen atoms bound to the metal in the tantalum and tungsten trihydrides, and the ability of the neutral di-n-cyclopentadienyls or hydrides of Mol W, Re, Fe, Rul and Os to take up a proton, suggest a structural interpretation of the properties

of this class of compounds. The experimental results suggest that in the complexes

(n-05H5)2MEx (x = 1, 2, or 3) there are, in addition to the hybrid orbitals involved in metal-to-ring bonding, three bonding orbitals, two

of which are equivalent. These three orbitals can be ()Coupled by

hydrogen atoms exclusively, as in (n-05H5)2TaH3, or by hydrogen atoms and

lone pairs of electrons - thus (n-05H5)2ReH will have two lone pairs. In view of the steric effects of lone pairs in the compounds of transition and non-transition elements which have lone pairs present on the central atom192 it would be remarkable if the lone pairs in the present series of

compounds did not have the same steric consequences and occupy bond 20. positions. Hence the (n-05H5)2MR)t. species could be considered as

5-coordinate, probably with a configuration approximately trigonal bipyramidal, in which case the non-equivalence of the three hydrogen atoms in the trihydrides implies that the metal-to-ring axes are non- linear. In ferrocene, ruthenocene, and, presumably, osmocene, the metal-to-ring axes are certainly linear. While sterically active lone pairs are therefore presumably absent, there must nevertheless be primarily non-bonding (versus the rings) electrons on the metal atom. The addition of a proton to such a linear di-%-cyclopentadienyl compound would be expected to distort the rings as in the above cases. The greatly reduced basicity (protonation of ferrocene occurs in BF3,H20,46 but we find no high-field line in trifluoroacetic or concentrated sulphuric acid), compared with, say, (n-C5H5)2ReH, is consistent with involvement of the non-bonding electrons in subsidiary metal-to-ring bonding. So far no evidence in any case for the uptake of more than one proton has been found, but it is only to be expected on electrostatic grounds that the second base constant would be very small. Even in

BF3,H20 the ratio of the low and high field lines for the rhenium compound is 10:2. In extending these ideas, it seems likely that all di-7c-cyclopentadienyl metal compounds of the general type (n-05H5)2tan will have non-linear metal-to-ring axes. Non-linearity has indeed been confirmed for (1-05H5)2TiC12A1(C2H02. 93It also follows that some compounds such as (n-05H5)2TiX2 should have a vacant orbital. 21.

These qualitative views of the nature of the hydride species

have received mathematical support within the framework of molecular-

orbital theory. Ballhausen and Dahl94 have shown that the metal-to-

%-05H5-ring bonding, as made evident by orbital overlaps, is very little

reduced even if the metal-to-ring axes are at angles down to 1350; even

further distortion can doubtless occur with a sacrifice in metal to ring

bonding energy. Further, with angular bonds, it follows immediately that

there will remain on the metal atom three strongly space-directed orbitals

(cf. Fig. 2): q) 0 (an a hybrid), y, and q157. (spdx2y2, dz2 hybrids),

two of which are equivalent (*_y, 4/ y), of precisely the character

required. Twelve electrons are used in the bonding between the cyclo-

pentadienyl ring and the metal; this electronic structure is common to

all sandwich compounds. The surplus electrons are distributed between

the space-directed orbitals. The number of electrons in excess of the

twelve in each case are:

6 in (n-c5H5)2Fe 4 in (n-05H5)2W

5 in (7c-05H5)2Re 3 in (n-05115)2Ta.

When ferrocene is protonated distortion to the angular form may occur

with the hydrogen atom probably located inkii/ o. In (n-05H5)2ReH the + proton could be similarly located, but in (7t-05H5)2ReH2 the proton

magnetic resonance spectra show equivalence of the protons, as is the

case for the molybdenum and tungsten dihydrides, so that presumably + In (7c-05H5)217113 9 (n-05H5)2MoH3 and -y andyi are now occupied. (7c-05H5)2TaH3, all of the orbitals will be occupied, and the expected z

Ay Iy 23.

FIG. 2

Diagram showing expected disposition of

7t—G5H5 rings and non—bonding orbitals

in (7—05H5)5EHn compounds • 25.

FIG. 3

Molecular structure of (n—05H5)2Molie. 26.

non-equivalence of the hydrogen atom in 0 is clearly demonstrated by the appearance of an AB2 group in the proton resonance spectra. The dipole-moment90 and infrared data are consistent with non-linearity of the metal-to-ring axes, but provide little conclusive evidence either for or against thorn. Thus it might have been expected

(cf. ref. 26) that, if the H-M-H bonds in the dihydride species are at an angle, both the symmetric and the asymmetric stretching modes would have been evident in the infrared region. This is not so. However, it is not necessarily surprising, since with the very heavy central metal atoms, together with the fact that if the H-M-H angle is about 900 as the molecular-orbital picture suggests, the coupling could be very small. A similar case occurs for the dialkyl- and aryl-tin dihydrides, where the H-Sn-H group is certainly angular, in which only one Sn-H stretching frequency can be observed.95 It is also possible, in principle, to decide on structures with parallel or twisted rings in

(it-05H5)2M compounds by infrared studies, as discussed, for examplepfor 9697 dicyclopentadienyl-tin and -lead. For the species (n-05H5)2ReH, + (n-05H5)2Mo(W)H2, and (%-esH5)2Mo(W)H3 there are indeed differences in the number of bands observed" in the 400-600 cm.-1 region associated with metal-to-ring stretching frequencies. However, only three or four of the predicted 7-9 bands appear. Further, the measurements were made in mulls where crystal lattice effects provide unknown and often large variables, so that such small differences among weak bands cannot be taken as providing substantial evidence for or against an angular 96 configuration. (A similar conclusion was reached for the tin and 27. lead compounds, even from solution spectra where the spectra were independent of the nature of the solvent.)

Crystallographic Studies. While the spectroscopic evidence suggests that the cyclopentadienyl rings are not parallel to each other, there is no way of determining chemically whether this is correct. The only means by which the structure could be confirmed was by an X-ray crystallographic examination. A study of dihydridodi-n-cyclopentadienyl- molybdenum was carried out subsequent to the work previously described in this thesis .124

The pale yellow crystals, grown by slow sublimation in vacuo, are monoclinic, with a = 14.30 A, b = 5.90 A, c = 10.41 A, - 104.00; the space group is C2/c or Cc. The structure has been determined by Patterson, Fourier and least-square methods the bond lengths shown corresponding to the usual discrepancy index R = 0.090 for the complete two-dimensional data. The mean molybdenum-carbon bond length is 2.29 ± 0.04 X. The mean carbon-carbon bond lengths and bond angles are

1.46 X and 108° with e.s.d's of 0.09 X and 3°, respectively. The molecule is wedge-like, the angle between the eclipsed cyclopentadienyl rings being 25 t 3° (see Fig. 3). A particularly significant feature of the analysis is the location, by difference Fourier methods, of the hydrogen atoms. The mean molybdenum-hydrogen bond length is 1.2 ± 0.2 X, the bond angle

H-Mo-H being 90 ± 10°. This result is in striking agreement with the only previously known metal-hydrogen bond distance of ca. 1.1 I in 28 .

H2Fe(C0)4, as determined by broad-line proton resonance methods.39

The overall geometry of dihydridodi-n-cyclopentadienylmolybdenum is also

consistent with the electronic model described by Ballhausen and Dahl.94

INFRARED STUDIES

Transition-metal-hydrogen stretching frequencies have been

discovered112 in the infrared region 2400-1700 cm-1; bending modes are

considered to be between 600 and 900 cm-1. In the series of hydrides

containing chelating phosphines or arsines and electronegative ligands

such as Ol y NO 2 y and CN 9 assignments of the metal-hydrogen stretching

frequency were relatively easy since the absorptions are fairly

intense.30/48Y5759 This is not so in the carbonyl hydrides. The 11,16 compounds HCo(C0)4 and DCo(C0)4 were extensively studied10, and a -1 very weak band at 703 cm was tentatively attributed to the Co-H

stretching frequency. However, the assignation of the M-H stretching

frequency in HMn(C0)52904945 at 1783 cm-1 indicated that the very 17 intense carbonyl bands might obscure that frequency in HCo(C0)4. -1 Subsequently an absorption at 1934 cm was accepted to be the Co-H - 66 stretching frequency.

The spectra of the Group VI tricarbonyl-n-cyclopentadienyl- 18 metal complexes have been published but no assignment of the metal-

hydrogen stretching modes was attempted. A number of weak absorptions were observed close to the very intense carbonyl bands and these were

also noted'in the spectra of the corresponding alkyls and ,-C5H5Mn(C0)3. 29.

FIG. 4

Infrared Spectra of Tungsten Hydrides

(in the region 2100-1200 cm-1)

A. 7t-05H5W(C0)3H in CS2 solution. B. (11-C H )2MH3 in hexachlorobutadiene mull.

(the dotted lines represent the corresponding deuterides) 114) 0002 008I 0091 00t71 0021 co

e 31.

One of these bands was considerably more intense in the hydride than in the deuteride spectrum and was assigned to the metal-hydrogen stretching frequency A band in the deuteride spectrum in the region 1385-1000 cm 1, absent from the spectrum of the hydride, was assigned to the metal- deuterium stretching frequency. The ratio s)) mH/N710 was 1.395 for the -1 molybdenum and tungsten complexes; the weak bands at 1790 and 1845 cm 18 previously reported without assignment, are the Mo-H and W-H stretching modes, respectively. Solutions for study were made up in a nitrogen- filled polythene bag using carefully degassed carbon disulphide.

The hydrides and deuterides of iron and ruthenium, it-05H5M(C0)2H, are unstable thermally and in air and were handled in a vacuum wherever possible. Once formed, the compounds were rapidly distilled in vacuo into a tube containing degassed carbon disulphide.

The solutions for study were then quickly transferred under a stream of nitrogen to the infrared cells. Despite this technique, the bridging carbonyl band of the binuclear species, [n-05H5M(C0)2)2, appeared in -1 -1 every spectrum recorded (e.g. at 1772 cm for Fe and 1812 cm for Ru).

The M-H stretching frequencies in the cationic hydrides, + (7C-CsH02141Ix , were found by comparing the mull spectra in the range

2300-1700 cm-1. Weak bands at 2055, 1915, and 1943 cm1 , in spectra of the rhenium, molybdenum and tungsten hydrides, respectively (absent in the deuterides), are assigned to the metal-hydrogen stretching mode.

The expected M-D stretches were not found. These should occur

(calculated from.VmHfimp, ca. 1.4) very close to or under the more -1 intense C-H deformation frequencies in the region 1500-1200 cm . 32.

CHAPTER III

TRANSITION METAL ALKYLS AND RELATED COMPOUNDS

INTRODUCTION

Chemical shifts are the result of differences in the extent to which the nuclei are screened from the applied magnetic field by the electrons around them. The total field H, at a nucleus is given by

H !applied(1 - C7)2 where 0 is the nuclear magnetic screening constant. 125 126 McConnell and Pople have considered the screening constant, at any nucleus, to be the sum of three terms:-

6'd. is the diamagnetic contribution to the screening constant from the electron cloud distribution at the nucleus concerned. It depends on induced circulations of electrons around the field direction, which set up a small field at the nucleus in the opposite direction to Happlied. revresents the reduction in the shielding due to restriction of the P - induced electron circulations by the electrostatic fields duo to neigh- bouring bonded atoms in the molecule. It is called the "paramagnetic term" because it is calculated in terms of a "mixing-in", by the applied magnetic field, of low-lying excited electronic states with the ground state of the molecule. For protons, Cr is generally quite small 33.

because there are no low—lying p orbitals associated with the hydrogen atom.

Cro p the "long—range shielding term", is due to the magnetic fields set up by the circulations of electrons on atoms in other parts of the molecule.

To give a net effect, which is not averaged to zero by the Brownian motions of the molecules, the neighbouring atoms must be magnetically anisotropic. This term is important for proton resonances because the diamagnetic shielding by one electron is small.

127 128 Cavanaugh and DaileY , following ShoolerY, studied the high—resolution proton magnetic resonance spectra of CH3X and C2H5X compounds. They found a linear relationship between the internal

6i, — r ) and the chemical shift, of an ethyl group (S.i. = CH3 CH2 electronegativity, , of the group X:— o.684 bi 1.78 (1) Using values of E calculated from (1) they derived an empirical relationship between the "methyl shift", m9 which is the separation, from methane, in cycles/sec., of the resonance of the methyl group at

60 Mc./sec.:-

m 100 € — 151 (2) 0 They found that both m and i varied linearly with electronegativity.

On this basis, they concluded that any influence on the screening due to the magnetic anisotropy of the C—X bond could be disregarded. They interpreted the difference, D, between the shifts of the methyl protons in

CH3X and the 0( protons in the corresponding C2H5X compounds as being due to a "C— C bond shift" arising in the latter compounds and acting equally 34 •

on protons at both ends of the ethyl group. The physical meaning of the

"shift" is not clear although it was suggested that it arose in the magnetic anisotropy of the C-C bond. A similar, though smaller, effect has been observed in the spectra of some hydrocarbons.129 By taking D in a number of compounds as a measure of the "C-C bond shift", subtraction of this value from the f; proton resonance in the ethyl left a residual

shift of 41 cycles/sec. (0.7 p.p.m. at 60 Mc./sec.) while the corres- pending shift for ethane itself is 38 cycles/sec. (0.63 p.p.m. at

60 Mc./sec.). This implies that when corrections for the extra shift due to the C-C bond are applied, all the i6 protons have a similar electronic environment. Support for this idea was found when the analogous series of isopropyl derivatives were studied. It was also suggested that the difference D might have been expected to arise primarily from the difference in electronogativity between the H atom and the CH3 group.

Spiesecke and Schneider130 suggested that in addition to the inductive effect of the group X, the effect of its magnetic anisotropy must be considered. When the magnetic anisotropy contribution is allowed for, a reasonably good correlation with the electronegativity of

X is possible. It appears that equations (1) and (2) hold only for those groups in which no large "neighbour anisotropy" effect is expected98,100,101,102 and it is expected that the use of equations (1) and (2) would give calculated methyl shifts higher than those which are 130 observed in the CH3X compounds when X is a highly anisotropic group.

The apparent validity of (1) in compounds in which X should have a large 35.

"neighbour anisotropy" effect, e.g. C2H5I, may arise because the con—

tributions to the shielding at the O( and ig protons will be in the same 130 direction in the applied field and of comparable magnitude so that the

difference between their respective shifts is cancelled. The net effect

is a move down—field of both the Ol and e, proton resonances.

TRANSITION METAL ALKYLS

The chemical shifts of protons bonded directly to transition

metals in groups such as Mn(C0)5— and 7t—05H5Fe(C0)2—, etc. (Table 1),

when compared with the shifts of hydrogen atoms bonded to main—group

elements of low electronegativity (e.g., Sn; SnH4,t = 6.11)295 show that

they are highly diamagnetically shielded by the transition metal. If the

shielding of the transition—metal hydrides can be regarded as the result

of a large "neighbour anisotropy" effect of the transition metal atom, a

large contribution from this might be expected to be present in the alkyl

derivatives. The proton resonance spectra of these compounds (Tables 2

and 3) show clearly that: (a) there is no correlation between e

calculated from equation (1) and the position of the methyl resonances;

(b) the subtraction of the difference between the shifts of the cK—protons

in CH3X and C2H5X compounds from the (3—shift of the ethyl protons does

not leave a residual p —shift of 41 cycles/sec.; and (c) the methyl

shifts calculated by using equation (2) [except for Re(C0)5CH3] are

0.25— 0.85 p.p.m. lower than the observed values. Hence the proton 36.

resonance spectra of CH3X and C2H5X compounds of the transition elements show that, as with non-transition-metal alkyls, the screening of the alkyl protons cannot be described adequately on the basis of a sum of the con- tributions of the inductive and the paramagnetic effects. Some evidence that the lack of correlation depends on the close proximity of X to the alkyl group is shown by the introduction of a carbonyl group between the transition metal and the ethyl group since the spectra of the resulting propionyl derivatives are all very similar to those of a normal

C2 H5--C- it group. Further, calculations of electronegativities utilising 0 O in C2H5X compounds or the position of methyl shifts in CH3X compounds are not valid and at best may provide, in a very closely related series of compounds, only a qualitative indication of relative electronegativities

(see Table 4). It should also be noted that X is not just an isolated transition-metal atom but a transition metal carrying groups such as ic--.C5H5 and CO which can also produce long-range screening contributions at the alkyl-proton positions.

Recent studies of the proton magnetic resonance spectra of ethyl groups bonded to nuclei of spin t have been made, the analysis of the

A3B2X spectra being done by using the effective internal chemical shifts, 98 S' and (S ", as two sets of A3B2 spectra or by spin decoupling.99 The spin-coupling constants of the X nuclei with methyl (JAx) and methylene groups (JBx) are of opposite sign. The spectrum of n-05H5W(C0)3CH3 showed that for the methyl group is 4.1 cycles/sec. J183'.-11

37. TABLE 2 Proton Resonance Spectra of the Transition Metal Hydrides, Alkyls and Propionyls at 56.46 Mc./sec, (SiMe4 used as internal reference in all cases.) Compound 1 8 J Solvent C5H5 M-H CH2 sec.) (c./sec.) j/6 Mn(CO)5H 17.5e . Mn(C0)5CH5 CC14 10.11 Mn(C0)5C2H5 C6H6 8.96 8.55 -22.9 8.02 -0.35 Mn(C0)5C0.C2H5 CH2C12 7.04 9.08 +82.5 7.22 +0.09 Re(C0)5H C4H80 15.66 Re(C0)5CH5 CC14 10.23 Re(C0)5C2H5 C6H6 9.00 8.23 -43.45 7.81 -0.18 Re(C0)5C0C2H5 C6H6 7.53 9.12 +86.40 7.26 +0.08 %-05H5Fe(C0)2H C6H12 5.26 21.91 it-05H5Fe(C0)2CH5/ CC14 5.30 9.89 IT-05H5Fe(C0)2C2H571* C6H6 5.40 8.45 8.75 +16.96 7.63 +0.45 7t-05H5Ru(C0)2H C6H12 4.84 20.92 7C-05H51112(C0)2CH3 CC14 4.76 9.71

TC-05H51111.(00)2C2H5 C6H6 5.31 8.67 9.03 +21.35 7.69 +0.36 n-05H5Ur(C0)5H71 C6H12 5.22 15.46 It-05H5Cr(C0)57571 0014 5.24 9.32 %-cslismo(co)3e c5H12 4.70 15.52 %-c5li5mo(003c113 71 0014 4.73 9.66

11-05H5MO(C0)3C H574 col,. 4.74 8.31 8.56 +14.20 7.61 +0.54 %-c5H5w(co)3H712 c6H12 4.65 17.33 71-c5H5w(c0)3011371 cc14 4.62 9.60 %-c5H5w(co)3c2H # 014 4.65 8.39 8.54 +8.62 7.54 +0.88 %-c5H5w(co)30211:47 cH2c12 4.63 8.34 8.52 +9.55 7.54 +0.79

ir-c5115(c5li5cH3 )co 5.44 9.87 (%-c51-15)2Ti(o3 )27/ 4.03 1o.16 (C6H5CH3 )Mn(C0)572/ 9.6o

Cotton, Do and Wilkinson, J., 1959, 833. 71 Earlier measurements at 40 Mcisec., 1 not made with SiMe4 as internal reference. No analysis of C2H5 spectrum reported. 7171 D. Jones, unpublished work. 38.

FIG. 5.

Proton Magnetic Spectrum of

7c—05H5Ru(C0)2C2H5.

Theoretical spectrum (J/5 = 0.35) shown below.

Magnetic field increasing from left to right; units in S

l II 1 I -1-0 -0.5 0 0.5 1-0 1.5 a•o 40.

FIG. 6.

Proton Magnetic Spectrum of

Re(C0)502115

Theoretical spectrum (J/ = 0.18) shown below.

Magnetic field increasing from left to right; units in 6 .

42.

TABLE 3

Observed and Calculated Chemical Shifts of the Transition-metal Ethyl and Methyl Complexes

(The data are referred in cycles/sec. to the line position of CH4 at 60 Me./ sec.; positive values indicate protons less shielded than those in CH

Methyl Compound d-CH3 0-CH3 1 CH3-X Residual Calc. shift R-X CH3-X in in £ ' -0t-CH2 (2-shift methyl/ifference d. (R = C2H5) ethyl ethyl in ethyl in ethyl shift obs.-calc.

En(C0)5R -16.6 +52.4 +77.0 1.500 69.0 +8.0 -1.0 +15.6

Re(C0)5R -23.8 +50.0 +96.2 1.254 73.8 +23.6 -25.6 -1.8

n-05H5Fe(C0)2R -3.4 +83.0 +65.0 1.985 86.4 -21.4 +47.6 +50.9

%-c5H5Ru(c0)2R +7.5 +69.4 +48.2 2.020 62.3 -14.1 +51.6 +44.0

n-05H5Mo(C0)3R +10.4 +91.4 +76.4 1.951 81.0 -4.6 +44.1 +33.7

IT-05115w(c0)5R +14.0 +86.6 +77.6 1.881 72.6 +5.0 +37.1 +23.1

Calc. from S values in Table 1 by using equation (1). / Cale. from E values by using equation (2).

TABLE 4

Electrons•ativity Data Calculated for the Grou• IV .Alkyls

1-- .' For the , Observed t shift Calc. shift of mi.e4, Element €-- ethyls i'‘CCH 3e CI (p.p.m.) i 3 1 of CH3X vs. CH CH3X vs. CH i --. ! I- I C 2.6 +0.438 9.11 1 2.08 +43.44 +57 +14 Si 1.9 -0.4201'' 10.00 1 1.49 -9.96 -3 +7 Ge 2.0 -0.3077171 9.88 1.57 -2.76 • +6 +8 Sn 1.93 -0.390 , 9.96 1.51 -7.56 0 +8 Pb 2.451 0 9.27 11.78 +33.44 +27 -7 loo • Data from Allred and Rochow. Cale. by using equation (1) In cycles/sec. from CH4 at 60 Mc./sec.; positive values indicate protons less shielded than CH4 (IC 9.83). Calc. by using equation (2). 101 Al 98 • Data from Slomp. Data from Narasimhan and Rogers.

TABLE 5

183W-proton Coupling Constants

Compound J (cycles/sec.)

(7E-05H5)2WH2 H 73.2 (7t-05H5)2WH34- 47.8 for A2 protons of A2B group

It-05H5W(00)311 37.7 I-05H5W(00)30113 4.1 with CH3 protons 7t-05H5W(00)202H5 ca. 4.5 with CH3 protons of -C2H5 group [ 7[-05H5W(00)3 2E6+ / 38.6 [(7t-0 5H5)2MoW(C06 Hr 38.0

• Data from Green, McCleverty, Pratt and Wilkinson, J. Chem., Soc., 1,961, 4854. 7' Data from Davison, McFarlane, Pratt and Wilkinson, J. Chem. Soc., 1962, 3652. 44-

Thus it is likely that in the corresponding ethyl, the values of coupling constants would be J < J < 5 cycles/sec. The main multiplet of the BX AX ethyl spectrum was analysed as an A3B2 system. The analysis of the relatively weak A3B2X spectrum (X = 1831N, abundance 14.28%) would be difficult because the effective internal chemical shifts, St and 6 would not differ greatly from 6. of the main multiplet. At high radio—frequency power, a satellite line was observed at 1.6 cycles /sec. on the high—field side of the principal A6 line which did not correspond to any of the main multiplet lines. This line can be considered as part of the doublet produced by the 183W splitting of the principal line of the methyl group, A6 (the most intense line in observed and calculated 103 A3B2 spectra). The corresponding line on the low—field side of A6 was partially obscured by the stronger lines of the main multiplet.

Thus the 183W coupling to the protons in the p -CH, group is ca. 4.2 183 cycles/sec. The magnitudes of other W—proton spin—spin coupling constants (Table 5) are considerably lower than proton spin coupling constants found in compounds of other heavy elements of spin * such as

111Sn, 119Sn, 3011)b, etc. (e.g., 117/1 19Sn bonded directly to H,

J, = ca. 2000 cycles/sec.;95 183W bonded directly to HI = on-n ca. 30-80 cycles/sec.).

Note. The screening constant, 07, refers to the screening of a bare nucleus and normally cannot be measured directly. It is related to the H — 1 by the relationship E = 6- - 6 = y chemical shift, g Hr H where 67 and (Sr are the screening constants of the given nucleus of the sample and reference, and H and Hr their respective resonant fields. 45

TRANSITION-METAL PROPIONYL COMPLEXES

During studies of some transition-metal ethyl complexes of the type 7s-0 87 5H5M(CO)nCH5' it was found necessary to prepare analogous compounds in which a carbonyl group was inserted between the transition metal and the ethyl radical. Proton magnetic resonance studies showed that such propionyl complexes have "normal" ethyl groups ('C ca. 7-8, CH 2 CH3 ca. 9) in contrast to the case where the ethyl group is bound directly to the metal, e.g. 7t-05H5Mo(C0)3C2H5 (7 CH2 8.34, 17 CH3 8.52). Of the four propionyls prepared, 7c-05H5M(C0)2C0C2H5 (M = Fe, Ru) and 7t-05H5M(C0)3C0C2H5 (M = MO, W), the molybdenum species proved to be the most unstable and difficult to purify. The preparation of the iron, ruthenium and tungsten derivatives was carried out by treating the sodium salts, Tc-05H5M(CO)nNa, in dry tetrahydrofuran with propionyl chloride.104

Although this method was used in an attempt to prepare the molybdenum propionyl, only traces were found (using n.m.r. techniques) and the complex 113 was eventually prepared by reaction with CO under pressure.

The product of the reaction between 7c-05H5Yo(C0)3C2H5 and CO at

100 atm. and at 80° or 1000 was shown to contain two products both having

"normal" ethyl groups (see Table 6). It seemed possible that at temperatures above the melting point of the ethyl (77.5 - 78.50) the C2H5 group could migrate from the transition metal to the cyclopentadienyl ring giving a binuclear molybdenum complex {(It-05H4C2H5)Mo(C0)31 2.

That the ethyl group does indeed transfer above the melting point (in the absence of CO) was demonstrated by heating pure 7t-05H5Mo(C0)3C2H5 for 46.

2 hr. at 100° in a sealed evacuated tube. On cooling, red crystals were formed which were chromatographed on alumina columns, three bands being separated. Although the first and second bands were eluted from the chromatography columns together and could not be cleanly separated by further chromatography, the main product isolated from them was

(7c-05H4C2H5)Mo(C0)3/2. The second unidentified specirs may have been a mono-ethyl (7c-05H4C2H5)Mo(C0)6Mo(7t-05H5)1. The third species was the unsubstituted complex [I-05H5Mo(C0)3}2.

Evidence for the formulation of the first species is as follows.

Proton magnetic resonance studies show that it has an ethyl group similar to that in diethyl ketone (see Table 6) and the infrared spectrum shows the expected aliphatic C-H stretching frequencies in the range 2800-3000 -1 cm • Only two CO stretching frequencies are observed which correspond 114 to those observed in similar ring-substituted complexes. The ultra- violet absorption spectrum in benzene shows A max = 510 mg, = 1415 and X max =.390 mg, 6 = 12,680, again comparable with other similar complexes.114 18 It had already been noted that transition-metal alkyls of the type under discussion decompose to red solid„ when heated or left to stand in daylight, although the products were not investigated at the time.

A number of other similar complex carbonyl methyl and ethyl compounds have now been exposed to ultraviolet radiation but only the above molybdenum ethyl complex showed any tendency to form the {(Tc,C5H4C2H5)Mo(C0)312. 47

All the other complexes, particularly the tungsten ones, gave unsubstituted

binuclear species.

The decomposition of the 5-05H5Mo(C0)3C2H5 is accompanied by the

formation of condensable gaseous hydrocarbons (98%) and a non-condensable

gas (2%). The condensable hydrocarbons were identified by gas-liquid

chromatography as ethylene and/or ethane, n-butane, and twol'C i fractions

(ratio ca. 1:50; b.p. 36 ± 2° and 42 ± 2°). The overall ratios were

ca. 1:8:1:800; the "C5" species may have been diethyl ether and either a

branched pentene, penta-2,3-diene, cyclopentadiene or cyclopentene.

The non-condensable gas contained ca. 99% methane and ca. 1% hydrogen.

The infrared spectrum of the gas showed no peaks in the region 2000-2200

-1 . CM yiind icating that CO was not released during the decomposition.

Discussion. A radical mechanism may be postulated in which

homolysis of the ethyl-molybdenum bond occurs easily in the melt. This gives an ethyl radical which can either dimerise (forming n-C4H10) or

attack the n-05H5 ring forming (71-05H4C2H5)Mo(C0)3. and H.; the larger radical then dimerises. It might be expected that if this mechanism were correct, the main products of the decomposition would be n-butane and

the unsubstituted binuclear molybdenum complex. n-C4Hi0 is observed in

the reaction gases, as well as lower and higher-boiling hydrocarbons;

the formation of these in a radical decomposition is not surprising.

Further, the ethyl complex was never obtained in yields greater than 10%

(based on the starting material), the unsubstituted species being the main

binuclear complex isolated. Also, no substitution on the ring was expected or found when the decomposition was carried out in light petroleum (b.p. 80-100°). 48

TABLE 6

Compound I Carbonyl stretching frequencies T-c1) 'CH tCg 3 n-05115 FC(C0)2C0C 2115 5.56a 7.28 9.10 2026s, 1960s, 1840vw, 17591m, 1657sd

IT—05H5Ru(C0)2C0C2H5 4.69b 7.80 8.99 2032s, 1960s, 1850vw, 1710vw, 1630md

1—05H5Mo(C0)3C0C2H5 5.41a 7.82 8.98 2016s, 1930s, 1708w, 1675sa n—05H5W(C0)3C0C2H5 4.40c 7.12 9.18 2014s, 1932s, 1740w, 1640md

{(1-05H4005)Mo(00)312 1 4.68° 7.08 , 8.81 2015m, 1960s, 1910s

a in benzene. in dichloromethane. in carbon tetrachloride. ketonic carbonyl stretching frequency. 49 •

OTHER METHYL DERIVATIVES

The proton magnetic resonance spectra of the compounds 146 it-05H5(C5H5CH3)Co and (C6H6CH3 )Mh(C0)\3 show that the methyl protons are unusually highly shielded. It had been suggested147 that in the series of compounds I-05H5(C3H5R)Co (R = H, CH39 CC13 , CF3, C6H5), the group R and the carbon atom to which it is attached should lie below the plane formed by the other four carbon atoms of the C5-ring, thus allowing interaction between the transition metal and the group R.

This sophisticated "hydrogen-bonding-like" mechanism could account for both the infrared spectra (in the C-H stretching region) of these compounds and for the high diamagnetic shielding of the methyl group in ic-05H5(C5H5CH3)Co.

Recently, an X-ray crystallographic examination of n-05H5-

(C5H5C6H5)Co has shown148 that the phenyl group and the carbon atom to which it is attached lie above the plane of the other four carbon atoms.

The fluorine magnetic resonance spectrum of 7t-05H5(C5H5CF3 )Co shows that the CF3 resonance (at 15.5 p.p.m. on the high-field side of C6H5CF3 ) occurs in the region characteristic of terminal CF3 groups attached to aliphatic fluorocarbons. Any interaction with the transition metal would cause a marked down-field shift (of up to 90 p.p.m.) of the CF3 resonance

(see Chapter IV).

These-facts seem to indicate that the methyl group in the cobalt derivative (and, by analogy, in the manganese) is well isolated from any influence of the metal although no other convincing explanation of the anomalous shielding has been found. 50•

CHAPTER IV

RHODIUM PERFLUOROALKYL DERIVATIVES

INTRODUCTION

The theory of chemical shifts of fluorine nuclei has been 133 developed largely by Saika and Slichter who, together with Hofmann and

Gutowsky,134 had showed earlier that there is good correlation between the

electronegativities of the atoms to which the fluorine nuclei are attached

and their chemical shifts. It might be expected that an increase of

electron density on the fluorine atoms would lead to increased screening of

the 19F nucleus but this simple explanation is quantitatively inadequate. 133 It was suggested that for the fluorine nuclei, variations in

the local paramagnetic term, 0- p (see Chapter III), were the dominant

would not be expected cause of the chemical shifts; changes in ord and To to contribute substantially. In the calculation of the 6 term the lowest electronic state of the molecule is mixed in with small amounts of

excited electonic states. In C-H bonds the appropriate excitation

energies are usually too large to allow 6- to have a significant effect on

the chemical shift, but with fluorine atoms the large negative chemical shifts which are often observed (e.g. UF6, -995 p.p.m. relative to HF)1135 indicate that the paramagnetic term is of major importance, presumably

because the 2p electrons have excitation energies lower than those in a

C-H bond. 51.

Perfluoroalkyl derivatives of transition metals have been 136 studied by Stone and his co-workerso The 19F magnetic resonance spectral

agt a of a number of these compounds, together with some perfluoroalkyl

iodides and some tin and phosphorus compounds, are shown in Table 7.

The chemical shifts of the fluorine atoms of theig-CF2 and y-CF3

groups are relatively constant but thea-CF2 resonance varies from compound

to compound (CX refers to the carbon atom nearest to the metal). When

the ct-CF2 group is bonded directly to a transition metal or to iodine a

large low-field shift is found. When a carbonyl group is inserted

between the perfluoroalkyl group and the metal, or the group is attached

to elements like Sn and P, no large low-field shift is observed. The

401,-CF2 group of the perfluoroethyls is more shielded than that in the

perfluoropropyl compounds, an effect which has also been observed in

fluorocarbons.137 It is notable that when a perfluoroalkyl group is

attached to a metal also bonded to iodine, the N-CF2 resonance moves to

a lover field compared with that of the analogous derivative containing no iodine.

In (C3F7)2PI the electronegativity of the PI group may be quite high, but the of-CF2 resonance is close to that in C3F7Sn(C4H9)3 in which 150 the electronegativity of the Sn is probably considerably lower. alone This would indicate that electronegativity/cannot account for the

unusual shifts of the 0C-CF2 resonance in all these compounds. It 136 was suggested that the presence of large down-field shifts in the

transition metal derivatives, and their absence in the tin and phosphorus 52

compounds, is a result in the former compounds of partially filled d-electronic shells, which can be induced to mix with gp orbitals of the fluorine atoms giving rise to a substantial value of 6"p. This is unlikely for the tin and phosphorus derivatives but iodine can have lower-lying excited states which may mix with the ground state.

THE RHODIUM PERFLUOROALKYLS

It had been found that it-05H5Co(C0)2 forms perfluoroalkyl 138 derivatives of the type 7c-05H5Co(C0)(R)I (Ft = CF5, C2F5, C5F7).

%-05H5Rh(C0)2139 forms similar compounds when it reacts in a sealed tube 140 with CF5I, C2F5I or C5F7I in benzene at 40°. The products form dark-red air-stable needles whose infrared spectra are similar to those of their cobalt analogues. These are the first reported perfluoroalkyl derivatives of rhodium.

The 19F magnetic resonance spectrum of n-C5H5Rh(CO)(CF3)I shows a doublet (splitting 11.7 c.p.s.) at -72.6 p.p.m. The doublet structure arises from electron-coupled spin-spin interaction of the 19F 1 03 nuclei in the CF3 group with Rh (spin 2, 100$ abundance). The very large down-field shift of this resonance, relative to that of CF5Sn(CH5),1 41 is similar to those shown in Table 7b and is consistent with the para- magnetic shift observed in similarly bonded CF2 groups of the compounds in Table 7a. Multiplet structure observed on the doublet lines (six components, splitting 0.5 c.p.s.)is probably due to coupling between the fluorine nuclei and the protons on the cyclopentadienyl ring. The corresponding splitting in the proton resonance of the cyclopentadienyl 53.

protons (T 4.11 in CH2C12) by the CF3 fluorine nuclei was not seen; iO3 neither was the expected coupling of these protons with Rh although the line width at half-height was 2.0 c.p.s.

In the perfluoroethyl compound, the expected structure would be a quartet and a triplet, with extra splitting due to rhodium coupling.

The spectrum was more complicated than this. An unsymmetrical quartet at —4.4 p.p.m. (the c-CF2 group) and a single line at +20.1 p.p.m. (the

CF3 group) was found. The of-CF2 group resonance is a typical AB pair showing that the two fluorine atoms, FA and FE, in this group are not equivalent and are thus able to couple with each other (JAB = 261 c.p.s.).

Each fluorine resonance shows additional splitting into a doublet and again into a quartet. The notable feature of these additional splittings is that they are different for FA and F . The doublet splitting is B 9.7 c.p.s. for the high-field fluorine atom, FB, and 6.3 c.p.s. for the low-field fluorine atom, FA. The separation between the components of the quartets is 1.7 c.p.s. for FB and 2.3 c.p.s. for FA. The doublet splittings presumably arise from coupling with 103Rh, while the quartet splittings are due to coupling between the 40(-CF2 and the CF3 group (J043).

The CF3 line was expected to be a double doublet due to the different couplings with FA and FE but the splittings are so small (ca. 1-2 c.p.s.) that the components overlap and a single broad line (line width at half-height 5.2 c.p.s.) is found even under very high resolution.

The 19F magnetic resonance spectrum of 7c-05H5Rh(C0)(C3F7)I is exceedingly complex. As e/pocted, the 0L—CF2 group appears at low—field 54.

(-8.6 p.p.m.), the at mid-field (+17.1 p.p.m.) and the P-CF2 at high-field (+52.5 p.p.m.). The )/-CF3 group shows "triplet" structure due to coupling with the two fluorine atoms of the Ot-CF2 group but, since these fluorine atoms are not equivalent, this main splitting of the CF3 group is really a double doublet. (It is known from measurements on simple perfluoro- alkyl compounds that Joiy is considerably greater than J or Jci,8 ).136 AY The two doublet splittings are 11.5 c.p.s. and 10.9 c.p.s. The equivalent splittings should be observed in the cx-CF2 group resonances but owing to the difficulty in resolving the two components of the c(-CF2 multiplot, only the 11.5 c.p.s. coupling could be confirmed. The outer two lines of this double doublet show clearly a further splitting into triplets (splitting 1.2 c.p.s.) presumably by the -CF2 group although 136 this is not often observed. The inner line of the double doublet is a quintet (splitting 1.0 c.p.s.) which can be explained by the overlapping of the innermost triplets of the double doublet. The -CF2 group gives rise to an AB pair (JAB = 284 c.p.s.). There is additional splitting on these lines which is observed most clearly on the two central components. The high-field line is a poorly resolved quartet (splitting ca. 1.5 c.p.s.) and the low-field lino a doublet

(splitting 2.0 c.p.s.). The quartet splittings must be due to the t-CF3 group with broadening due to smaller coupling with the fluorine atoms of the o(-CF2 group. Sfu TABLE 7A

Fluorine Chemical Shifts of some Perfluoroethyl and Perfluorop,ropyl Compounds.a

Jtx 6. Compound -CF2 - F2 ()1 -)CF3 'Tay '")< i in c4.s. in c.p.s.

C2F5Mn(C0)5 5.2 20.4 1.5 C2F5COMn(C0)5 50.9 16.7 0.7 C2F5Fe(C0)4I -4.6 19.9 1.8 (C2F5)2Fe(C0)4 10.4 20.1 2.0 C2F5Sn(C4H9)3 56.8 20.3 1.4 C2F5T 1.6 21.8 4.6 C3F7Mn(C0)5/3 2.0 51.7 15.2 12.4 C3F7Fe(C0)4I -7.3 50.5 15.5 11.4 (c3F7)2Fe(o0)4 5.5 51.7 15.0 11.1 C3F7Sn(C4H03 54.6 59.1 16.7 9.3 (o3F7)2Pic 39.3 56.1 17.6 9.2 3.2 03F71 -3.1 54.6 17.1 9.3 4.6 c2F5Rh(c5115)(c0)Ia -4.4 20.1 see text C3F7Rh(C5H5)(00)Id -8.6 52.5 17.1 see text

TABLE 7B Fluorine Chemical Shifts of some Perfluoromethyl Couounds.a

Compound CF3 resonance CF3 resonance (p.p.m.) Compound (p.p.m.) CF3Co(C0)4e -74.1 cF3c0(c5H5)(c0)i -73.5 cF3mh(co)5e -72.9 cF3Rh(c5H5)(c0)I -74.8 CF3Fe(C0)4.I -78.4 CF3Sn(CH3)3f +42.1

a. In p.p.m., relative to benzotrifluoride, positive values indicating lines on the high-field side of this reference. b. Jpy not observed in perfluoropropyl compounds; Jaig not usually resolvable. c. 31P--19F coupling constants: F = 23.6 c.p.s.; 36.2 c.p.s.; Ft = 9.2 c.p.s. d. Line positions corrected for solvent shifts in dichloro- methane (Evans, Proc. Chem. Soc., 1958, 115). e. ref. 149. f ref. 141. 56.

FIG. 7.

19F Nuclear Magnetic Resonance Spectrum of

I—05H5Rh(C0)(C3F7)I

in CH2C12.

(scale in p.p.m. relative to benzotrifluoride, positive values indicating values on the high—field side of the reference).

a. the well resolved lines in the 04—CF2 multiplet.

b. the ' —CF3 group.

q D

09 OS Ot OE OZ 01 58.

The low—field lines, due to the 0(—CF2 group, are very complex. AB = 240 c.p.s.), where the lines show They constitute an AB pair (j additional structure similar to that in n—05113Rh(C0)(C2F5)I. The high— field group is well resolved into a total of sixteen lines,being a doublet doublet quartet. One of the doublet splittings represents the coupling with rhodium (J—F = 4.3 c.p.s.) and the quartet splitting represents the Rh coupling with the CF3 group (Jyy = 11.5 c.p.s.). The remaining doublet structure (splitting 2.0 c.p.s.) suggests that there is coupling with one of the fluorine atoms of the / —CF2 group. The low—field group of lines could not be resolved although the group, as a whole, was more than

11 c.p.s. wide.

DISCUSSION

The non—equivalence of the fluorine resonances of the 04—CF2 and g—CF2 groups exhibited in the rhodium derivatives indicates that some form of hindered rotation of the Rh—C and C—C bonds may be taking place. This is not surprising in view of the relative size of the groups surrounding the rhodium atom. Rotational isomerism should be possible but the relative simplicity of the spectra suggests that there is a rapid rotation about the Rh—C and C—C bonds giving a spectrum which.is a weighted average of those expected for the rotational isomers at the temperature 126 at which the spectra were measured.

The coupling between the rhodium and fluorine nuclei is detectable 87 only on the O(—CF2 group. In 7g—05115W(C0)3C2H5 and (C3F7)2PI, coupling 59

between the 163W and 31P and 1 H and 19 F respectively extends to the -CH:,

and -CF3 groups. However in 7c-05H5W(C0)3CF2CF2H no 183W-19F coupling is reported.142 This effect may be explained in terms of the magnitude of

the magnetic moments of the relevant nuclei (I) ; 1.13, WI 0.12, Rh : -0.09, H ! 2.80, F 2.63, in multiples of the nuclear magneton).

In the perfluoropropyl derivative the coupling constants between

the fluorine atoms of the 0(-CF2 group and those of the y--CF3 group

ay ), are found to be much larger than the coupling constants between the a-CF2 and ?-CF2 fluorine nuclei or between the (-CF2 and 1/-.c173

fluorine nuclei. Values of coupling constants similar to those shown

the derivatives in Table 7 are found in CF3 CF2CFIC1 where Joey

T 143 J ( cAft and in (CF3)2CHCF2OCH3 where the strongest coupling occurs between the fluorine nuclei of the CF3 group and those of the CF2 group while the tertiary proton, though closer, causes much smaller splittings of the CF3 group.144

The values of J Rh-F decrease in the series of.rhodium derivatives in the order CF3 > C2F5 C3F7. The coupling mechanism:3 in fluorine compounds are not fully understood although the magnitudes

of coupling constants are regarded as a measure of the ability of the

electron cloud associated with two nuclei to couple these nuclei. The electron clouds associated with the rhodium, fluorine and 0c-carbon atoms

(and the bonds between them) could be influenced by the electronegativity

of a group attached to the 04-carbon atom (e.g. -CF3 or -02F5). 60.

A correlation between coupling and the electronegativity of attache& groups has been noted in proton resonance spectra.145 It is also

noticeable that those Ok-CF2 fluorine nuclei which are most strongly

coupled with the other fluorine nuclei are also the least shielded. 61.

CHAPTER V

EXPERIMENTAL

GENERAL

Microanalyses are by the Microanalytical Laboratories,

Imperial College.

Infrared Spectra. Measurements were made using a Perkin—Elmer

Model 21 spectrophotometer, and a Grubb—Parsons grating spectrometer in

the region of the M—H stretching frequencies (for the hydrides

(t—05H02MHx). Sodium chloride and calcium fluoride optics were employed.

High—resolution Nuclear Magnetic Resonance Spectra.

Alkyls and ropionyls. Measurements were made with a Varian Associates

Model 4311 spectrometer at 56.46 Mc./sec. and 22 + 2°. Whenever possible,

the alkyl spectra were measured in carbon tetrachloride solutions but, as

it was found that some of the ethyls reacted rapidly with halogenated

solvents, it was convenient to use benzene since the internal chemical

shifts, i, are relatively insensitive to solvent changes. The ethyl 103 spectra were analysed as A3B2 systems by using Corio's tables. In all

cases, the positions of the spectral lines, in cycles/sec., were converted

into units of 6, and from these J/S was calculated. The observed lines

were then compared with those calculated by interpolation from the Jig values in the tables. 62.

21Ydrides. The neutral hydrides were studied either in cyclohexane or

benzene solutions while the trihydride cationic species were measured in

water and also trifluoroacetic acid. Tetramethylsilane was used as the

internal reference (tertiary butanol in aqueous solutions).

Fluoroalkyls. The complexes were studied in either cyclohexane or

dichloromethane solutions. The fluorine resonance lines are given in

p.p.m. relative to benzotrifluoride (as internal reference), positive

values representing lines on the high-field side of this reference.

In all cases the samples were contained in 5-mm. (outside

diameter) Pyrex spinning tubes. Line positions were determined by

conventional side-band techniques.

Absorption Spectra. The spectra were recorded on a Perkin-Elmer

Spectracord 4000 using degassed solvents and silica cells.

PREPARATIONS AND REACTIONS

Preparations and reactions of compounds known or expected to be

sensitive to air were carried out under oxygen-free nitrogen or in a vacuum.

Di-n-cyclopentadienyltantalum trihydride. Anhydrous, freshly sublimed

tantalum pentachloride (36 g., 0.1 mole) was slowly added to an ice-cooled

solution of sodium cyclopentadienide (0.7-0.8 mole) in tetrahydrofuran

(250 ml.) containing sodium borohydride (10 g.). The mixture was stirred for 4 hr. under reflux. After removal of the solvent in a vacuum 63.

the product was sublimed from the residue (ca. 1200/10 2 mm.) and

purified by resublimation at 1000 as white crystals, m.p. 187-189°

(decomp.) (Found: C, 37.0; H, 4.0. C10H13Ta requires C, 38.2; H, 4.2%). The preparation does not always succeed and the reasons for this failure

are not clear. At the beginning of the work yields of the order of 60% were experienced but after introduction of a fresh batch of tantalum metal

powder yields of 10% or less were commonplace. It may have been that the original metal powder was less pure than the more recently acquired

material, "catalytic impurities" being responsible for the high yields of the trihydride. When yields from the fresh tantalum were too low for dried sublimation to be effective, the crude/reaction mixture was extracted with boiling benzene. The solution was evaporated and off—white crystals of

the hydride were obtained by sublimation (ca. 100°20-2 mm.).

Di—n—cyclopentadienylmolybdenum dihydride. The reaction was carried out as above, but with molybdenum pentachloride (27 g.), sodium cyclopenta— dienide (0.5 mole), and sodium borohydride (10 g.). The product was sublimed from the residue and crystallised from diethyl ether— light petroleum (b.p. 40-60°) at —78°. An alternative method of isolation is 26 that used for the rhenium hydride where the dry residue is treated with

3N—hydrochloric acid. After filtering, the solution is neutralised with 2N—sodium hydroxide, and the liberated hydride extracted with ether.

The product was purified by vacuum—sublimation at 50°, to give yellow crystals (ca. 50% based on molybdenum), m.p. 183-185° (Found: C, 51.0;

H, 5.3. C10Hi2Mo requires C, 52.7; H, 5.3%). 64.

Di-m-cyclopentadienyltungsten dihydride. A similar procedure but with

tungsten hexachloride (40 g.), sodium cyclopentadienide (0.6 mole), and

sodium borohydride (10 g.) led to yellow crystals (ca. 65% based on

tungsten), subliming in a vacuum at 80° and having m.p. 163-1650

(Founds C, 38.2; H, 3.9. C10H12W requires Cy 38.0; H, 3.8%). The

tungsten compound is moderately stable in air and after about 15 min. in

air about half of the material can be recovered.

Di-%-cyclopentadienyl hydride salts of rhenium molybdenum and tungsten.

The hydrides were treated with gaseous hydrogen bromide at room temperature.

The yellow crystals rapidly disintegrated, leaving white, very hygroscopic

solids. In presence of excess of hydrogen bromide, the solids are pink.

Aqueous solutions of the cations are obtained by dissolving the bromides in water or by dissolving the hydrides in aqueous acid. The pure salts were precipitated as hexafluorophosphates. The deuterides were prepared by dissolving the neutral hydrides in ca0.1N-trideutero- phosphoric acid and were precipitated with a solution of ammonium hexa- fluorophosphate in ca. 0.01N-trideuterophosphoric acid; in these conditions complete isotopic exchange takes place and high-field lines are absent from the proton magnetic resonance spectra of these solutions in D20. On addition of base, the neutral hydrides are liberated quantitatively. The dihydrides readily dissolve in pure trifluoroacetic give acid to/solutions containing the trihydride ions; the molybdenum solution is green and the tungsten olive green. 65.

Tricarbonyl-n-cyclopentadienyl Group VI metal hydrides and deuterides.

The chromium, molybdenum, and tungsten compounds, n-05H5M(C0)3H(D), were 18 prepared by the standard methods. The deuterides were obtained by

treating the sodium salt, n-05H5M(C0)3Na (10 g.), in tetrahydrofuran (10 ml.) with deuteroacetic acid (3 ml.; 60% isotopic substitution). n-05H5Mo(C0)3H and n-05H5W(C0)3H could also be prepared by dissolving the sodium salt in air-free water. To this was added HC1; the dense yellow-white precipitate formed was filtered off under nitrogen and either sublimed or recrystallised from diethyl ether.

Dicarbonyl-n-cyclopentadienylhydrido-and deutero-iron. The compound was obtained by treating the carbonyl n-05H5Fe(C0)2C1 (0.5 g.) with sodium borohydride (1 g.) in tetrahydrofuran (10 ml.). After 15 min. the red solution had become orange-yellow and was poured into water (50 ml.), and the hydride was extracted into light petroleum (b.p. 30-40°; 10 ml.). Removal of the solvent at 00/0.1 mm. left a pale yellow liquid which was distilled in a vacuum. Careful acidification of an aqueous solution of n-05H5Fe(C0)2Na and evaporation of the ether extract gave a low yield of the hydride. The corresponding deuteride was obtained by reduction of n-05H5Fe(C0)2I in diethyl ether by lithium aluminium deuteride. After removal of the solvent at 0°/0.1 mm. the liquid was fractionally distilled but complete separation from ether is difficult. 66.

Dicarbonyl-n-cyclopentadienylhydrido-and deutero-ruthenium. The hydride and the deuteride were obtained from the sodium borohydride a:_'. lithium aluminium deuteride reduction of I-05H5Ru(C0)2I. They were worked up in the same way as the analogous iron complexes. The ruthenium hydride is a colourless, volatile liquid with an obnoxious odour and, like its iron analogue, is very sensitive to air. It (and the iron hydride) was characterised by its infrared and proton magnetic resonance spectra.

When pure at 25° it rapidly dimerises to the binuclear carbonyl, but in solution the rate is much lower so that the spectra can be obtained readily.

Pentacarbonylethylmanganese. The ethyl was prepared by treatment of an excess of ethyl iodide (5 ml.) with a solution of 1E1(00)5Na (10 gm.) in tetrahydrofuran (30 ml.). The solution was stirred at room temperature for one hour. The ethyl was recovered from the tetrahydrofuran solution by pouring the reaction mixture into water and extracting the product into light petroleum (b.p. 50-400). It is a yellow, volatile, air-sensitive oil and was purified by fractional distillation in vacuo. The proton resonance spectrum of the pure compound Mn(C0)5C2H5 (sealed tube in vacuo) showed marked line broadening during 10 min. (the sample darkened noticeably in this time) which indicates that decomposition gives a para- magnetic species as well as Mh(C0)5C0C2H5. Solutions of the complex in hydrocarbon solvents are more stable. Pentacarbonyl propionylmanganese can be obtained by treatment of the ethyl with carbon monoxide under 115 pressure or from sodium pentacarbonylmanganate with propionyl chloride 104 in tetrahydrofuran. It was found that the compound Mn(C0)5C2H5 was completely carbonylated at 60-80° and 100 atm. in 10 min. 67.

Pentacarbonylethylrhenium. The compound was prepared and purified as its

manganese analogue. By addition of a solution of sodium pentacarbonyl- rhenate (5 gm.) in tetrahydrofuran (30 ml.) to ethyl iodide (5 ml.) and

pouring the reaction mixture into water and petroleum after an hour's

stirring, a yellow petroleum extract was obtained. Evaporation of this

solution left a yellow-green viscous oil. This oil distilled in vacuo with difficulty. Reaction of Re(C0)5C2H5 with carbon monoxide was not

complete after 1 hr. at 100° and 100 atm. in comparison with Mn(C0)5C2H5.

The proton resonance spectrum of the product after two sublimations showed

the presence of some Re(C0)5C2H5. Pentacarbonyl propionyl rhenium could

also be. prepared from sodium pentacarbonylrhenate and propionyl chloride in tetrahydrofuran.104

Reaction of the carbonyl sodium salts in tetrahydrofuran with

ethyl bromide gave large yields of the corresponding propionyls.

n-05H5Fe(C0)2% and n-05H5Fe(CO)C2H5. These complexes were prepared by 189105 methods reported in the literature . The properties of the ethyl are similar to Re(C0)5C2H5 and n.m.r. samples were prepared by vacuum distillation techniques.

Dicarbonyl-n-cyclopentadienylpropionyliron. Tetracarbonyl-di-I-cyclo-

pentadienyldi-iron (5 g.) was stirred vigorously for 15 min. with an excess of 1% sodium amalgam in tetrahydrofuran (50 ml.). The amalgam was separated and the solution containing the sodium salt was dripped slowly, at room temperature, into a solution of propionyl chloride (8 ml.) 68.

in tetrahydrofuran (50 ml.) under nitrogen. The mixture was stirred for 1 hr. and the solvent then removed in vacuo. The residue was extracted with light petroleum (b.p. 40-60°; 2 x 30 ml.) and the extract filtered.

The petrol was evaporated (400/0.1 mm.) leaving an orange-red oil.

Chromatography of this oil on alumina in a mixture of petrol (b.p. 40-60°) and benzene indicated that no further purification was necessary after extraction from the residue. The compound is air-stable, very slightly volatile and soluble in common organic solvents. (Found: C, 51.5;

H, 4.6; 0, 20.7%. C1oH1003Fe requires Cy 51.3; H9 4.6; 09 20.5%.)

Dicarbonyl-n-cyclopentadienylmethylruthenium. Ruthenium carbonyl iodide, l06 128 (Ru(C0)2I0n, was prepared by the action of CO on RuI3 at 250°. The orange powder (9 g.) was added to a solution of sodium cyclopenta- dienide (2 g.) in tetrahydrofuran (200 ml.). The solution darkened immediately and was refluxed for 3 hr., when it was cooled and an excess of 1% sodium amalgam added. After vigorous stirring, the amalgam was removed and the solution dripped into methyl iodide (5 ml.). A precipitate of NaI was observed to form immediately and the reaction mixture was stirred at 50° fcr 1 hr. The solvent was removed and the methyl was sublimed at 400/0.1 mm. from the residue. The compound is a volatile white solid, m.p. 39_400, soluble in common organic solvents.

(Found: C, 40.6; H, 3.4; 0, 13.7. C8H802Ru requires C, 40.9; H, 3.4; 0, 13.5%.) 69.

Dicarbonyl-n-cyclopentadienylethylruthenium. The complex was prepared in the same way as the methyl. The ethyl is a colourless, volatile oil, m.p. ca. —50, and samples for n.m.r. studies were prepared using vacuum distillation techniques. (Found: C, .43.1; H, 4.4; 0, 12.9.

C9H1002Ru requires C, 42.9; H, 4.0; 0, 12.7%.) Both the ruthenium alkyls are more stable than their iron analogues, although solutions of both in carbon tetrachloride decompose rapidly. In the absence of air, they are stable in hydrocarbon solvents.

Dicarbonyl-n-cyclopentadienylpropionylruthenium. This compound was pre- pared in the same way as its iron analogue. It is a pale yellow, air-stable oil, soluble in common organic solvents. (Found: C, 43.3;

H, 308; 0, 17.4%. O1 0 H 1 003HU requires C, 43.2; H, 3.9; 0, 17.3%.) n-05H5Cr(C0)3CH3. This was prepared from C5H5Cr(C0)3Na and CH3I in 18 tetrahydrofuran. Previous preparations have resulted in very low yields of the methyl and the reason for this was ascribed to the very high volatility of the complex. Consequently it was worked up in the same way as Mn(C0)5C2H5. Attempts to prepare n-05H5Cr(C0)3C2H5 using the procedure to make Mn(C0)5C2H5 failed. 18 n-05H5Mo(W)(C0)3CH3 and n.-05H5Mo(W)(C0)3C2H . Literature methods were employed to prepare these compounds. However, the methyls could be extracted conveniently by pouring the tetrahydrofuran reaction mixture into a large volume of water and filtering out the precipitated alkyls.

They could then be purified by sublimation. The ethyls were best 70 • obtained by evaporating the reaction solvent (in vacuo), extracting the solid residue with boiling light petioleUm (b.p. 40-60°) and cooling the petrol solution to -78° when orangeyellowl needles were formed. These were filtered out and dried, further purification being unnecessary.

Sublimation frequently caused extensive decomposition.

Tricarbonyl-I-cyclopentadienylpropienylmolybdenum. 7c-05H5Mo(C0)3C2H5

(5 g.) was treated with CO (100 at.) for 3 hr. at room temperature. The red oil which was formed was extracted with light petroleum (b.p. 40-600;

2 x 60 ml.) and the solution filtered and evaporated (400/001 mm.). The residual red oil was shown, by its proton magnetic resonance spectrum in benzene and by its infrared spectrum in carbon disulphide, to contain

7-05H5Mo(C0)3C0C2H5. Further purification by chromatography or by re-extraction into petrol was very difficult because the compound is very sensitive to oxygen and is thermally unstable, decomposing to give [7;-05H5Mo(CO)3]2 and other unidentified products. (Found: C, 42.81

H, 3.4; 0, 21.3%. 0,1111004mo requires C, 43.4; H, 3.7; 0, 20.7%.)

Tricarbonyl-%-volopentadienylpropionyltungsten. This complex was prepared in the same way as the iron analogue from 7-05H5W(C0)3Na (5 g.) and C2H5C0C1 (8 ml.) in tetrahydrofuran (100 ml.). The petrol extract was evaporated in vacuo leaving a waxy orange residue. Purification by chromatography on alumina in a petrol-benzene mixture (40:60) afforded) on removal of the solvent (600/0.1 mm.), a waxy orange solid, m.p. ca. 50.

The complex is air-stable and soluble in common organic solvents. 71.

(Found: C9 33.6; H, 2.6; 0, 16.4%. 011H1004w requires C, 33.8; H, 2.6; 0, 16.4%.)

Hexacarbonyldi-n-ethylcycl2pentadienyldi-molybdenum. Tricarbonyl-n- cyclopentadienylethylmolybdenum was heated at 100° in a sealed evacuated

Carius tube for 3 hr. A red oil was formed which crystallised on cooling.

The tube was opened and the contents washed with light petroleum (b.p.

30-400). The remaining red crystals were dissolved in benzene and chromatographed under nitrogen in a petrol-benzene mixture (40:60).

Three bands separated, the first two being eluted together. The solvent was evaporated leaving a dark-red solid. Further chromatography of this solid in petrol-benzene mixtures gave no separation. The red solid was identified as [n-05H4C2H5Mo(C0)3]2, m.p. 210-211°, soluble in chloroform, carbon disulphide, carbon tetrachloride and benzene. (Found: 0, 43.8;

H, 3.1; 0, 17.8%. C20H1806Mo2 requires C, 44.0; H, 3.3; 0, 17.6%.)

Gas Analysis. The gas from thermal decomposition of

IT-05H5Mo(C0)3C2H5 (ca. 2 g.) was collected and analysed chromatographically on three columns. The first sample was injected onto a silica gel column (column length 7 ft., packed with 28-60 mesh Davison silica gel kept at constant temperature by an acetone vapour jacket). A katharometer detector was used with oxygen-free nitrogen as the carrier gas. Hydrogen and methane were detected; CO, CO2, acetylene, ethylene and ethane were absent. The second sample was injected onto a silicone oil column

(column length 7 ft., the oil supported on 40-60 May and Baker Embacel 72. maintained at constant temperature by an acetone vapour jacket).

Again, a katharometer detector was employed, oxygen-free nitrogen acting as the carrier gas. Methane, ethylene and/Or ethane, and n-butane were identified; all other C2, C3, C4, C5 and C6 hydrocarbons were absent. In order to detect the presence of low-boiling oxygenated hydrocarbons, the third sample was injected onto a 20% polyethyleneglycol column (column length 15 ft., with polyethyleneglycol 400 supported on 40-60 alcoholic potassium hydroxide treated Embacel, maintained at constant temperature by an acetone vapour jacket). A thermocouple detector was used with hydrogen (75%) and nitrogen (25%) as carrier gas. Traces of methane and C2 hydrocarbons were observed as well as n-butane. Two fractions were observed in the Cs region, the smaller being probably diethyl ether impurity and the larger either a branched pentene, penta-213-diene, cyclopentadiene or cyclopentene. The boiling points of the two Cs species were 36 t 2° and 42 t 2° respectively.

Dicarbonyl-ic-oyclopentadienyliodoruthenium. The preparation was 108 analogous to that of n-05115Fe(C0)k 2I. The addition of iodine in carbon tetrachloride to tetracarbonyldi-n-cyclopentadienyldiruthenium1C9 gave the iodide as orange crystals, m.p. 103-105°, which sublimed at 1000/0.1 mm.

(Found: C, 23.7; H, 1.4; I, 36.3%. C7H5102Ru requires C, 24.0; H, 14; 36.3%0.) 73.

111 Decacarbonyl dirhenium. Previous procedures for making this carbonyl

Re2(C0)10 have not involved halides of rhenium as these give only the halogen derivatives, Re(C0)5X (x = halogen). The reductive carbonylation 110 procedure has been adopted to prepare Re2(CO)10 from rhenium trichloride or pentachloride. The halides were treated with a slight excess over the stoicheiometric amount of sodium in dry tetrahydrofuran and with carbon

monoxide at 130°/250-280 atm. in an autoclave (e.g. ReC15, 9.0 g.;

Na, 5.0 g.; tetrahydrofuran, 150 ml.). After 8 hr. the pressure was released and the tetrahydrofuran solution was filtered, acidified with ca. 2N-sulphuric acid, and extracted with diethyl ether. The ether extract was washed repeatedly with dilute sulphuric acid and finally with water and then dried (CaSO4). The solvent was removed (250/0.1 mm.) and

the residue acidified with dilute H2SO4 and steam distilled. The rhenium carbonyl was separated from the distillate by filtration and

purified by sublimation. it-Cyclopentadienyl(methylcyclopentadieno)cobalt. The compound was 117 prepared by a method reported in the literature. It was also formed by treatment of a suspension of cobalticinium iodide (2 g.) with an ethereal solution of methyl lithium [from lithium (0.14 g.) and methyl iodide (1 mi.)]. The solution was stirred vigorously and became orange-red within ten minutes. The ether was evaporated in a stream of nitrogen and the residue extracted with light petroleum (b.p. 40-60°).

The orange-red solution was filtered through kieselguhr and evaporated 74. carefully in vacuo (250/0.1 mm.). The residue was distilled (300/0.1 mm.) onto an acetone-dry-ice-cooled. probe giving the pure complex, identical with that prepared by the other method.

Di-w-cyclopentadienyldimethyltitanium. An ethereal solution of methyl lithium f: from lithium (1.4 g.) and methyl iodide (8 ml.)] was added to a stirred suspension of di-n-cyclopentadienyldichlorotitanium (10 g.) in ether (200 ml.). The yellow-green solution which formed immediately was stirred until it had become orange (3 hr.). The ether was removed in vacuo and the residue extracted with light petroleum (b.p. 40-600).

After filtration, the orange solution was evaporated slowly (25°/0.1 mm.) until orange crystals began to form. The complex was stored at this stage as the dried crystals decomposed spontaneously, even in vacuo.

The yield was ca. 60%, much higher than that obtained when the compound 18 was prepared using methyl magnesium bromide. For proton magnetic resonance studies, the dried crystals were quickly redissolved in carbon disulphide.

Monocarbonyl-n-cyclopentadienyltrifluoromethyl-iodorhodium. Dicarbonyl- n-cyclopentadienylrhodium (0.5 g.) was dissolved in benzene (5 ml.), sealed in an evacuated Carius tube with trifluoroiodomethane (3 ml.) and maintained at 30-40° for 30 hr. Large red crystals were formed. The tube was opened (CO pressure), and the volatile materials removed

(300/0.1 mm.). The red residue was extracted with dichloromethane and the complex precipitated as dark orange-red needles by addition of light petroleum (b.p. 30-40°). 7c-05H5Rh(C0)(CF3)I, m.p. 168-169°, may be 75.

purified by sublimation (1000/0.1 mm.) onto an acetone- dry-ice cooled probe, and is soluble in chloroform, dichloromethane, benzene, ether, and light petroleum (b.p. 30-100°),/acetone but sparingly soluble in carbon tetrachloride and carbon disulphide. (Found: C, 22 .1; H, 1.46;

F, 14.6; I, 32.9%. C7H50113 11a requires 0, 21.4; H, 1.28; F, 14.63

I, 32.4%.)

Monocarbonyl—n-cyclopentadienylpentafluoroethyliodorho dium. The complex was prepared in the same way as its trifluoromethyl analogue, using 140 pentafluoroethyl iodide prepared by the method of Hammldine. The complex, m.p. 145-147°, was purified by extracting the residue from the Carius tube in a soxhlet apparatus with light petroleum (b.p. 40-60°).

The compound is air-stable, soluble in chloroform, dichloromethane, benzene, ether, light petroleum (b.p. 50-40°), sparingly soluble in carbon tetrachloride, carbon disulphide, light petroleum (b.p. 40-60°) and insoluble in light petroleum (b.p. 60-120°). (Found: C 21.9; H, 1.33;

F, 21.4; I, 28.6%. C811.50F5IRh requires C, 21.7; H, 1.13; F, 21.5; I, 28.7%.)

Monocarbonyl-I-cycloyentadienylheptafluoroa2pyliodorhodium. The compound was prepared using heptafluoropropyl iodide and purified in the same way as the pentafluoroethyl complex, to which it is very similar. It forms red crystals, m.p. 118-119°. (Found: C, 17.1; H, 1.62; F, 27.4;

I, 26.2%. C91150F71Rh requires CI 16.5; H, 1.20; F, 27.0; I, 25.0.)

76.

INFRARED SPECTRA

The molybdenum and tungsten dihydrides were measured in CS2

solutions and in Nujol mulls; mull spectra only were obtained for the tantalum trihydride.

(7C-05H5)2Molla: 3060w, 1847 f 2 (Mo-H str., ms), 1415m, 1369m, 1267m, 1121ms, 1064ms, 1055ms, 869w, 813m, 763s.

(n-05H5)2WH2: 3050w, 189 6 + 2 (W-H str., ms), 1410m, 1257m, 1088m, 998m, 885w, 799m, 769m.

(7-05H0 TaH,: 3090w, 1735 + 2 (Ta-H str., vs), 1425m, 1358w, 1265w, 1117w, 1072w, 1020s, 1005s, 905m, 870s, 858m, 848m, 825s, 795s, 720w.

71-05H5Fe(C0)2H: 3113w, 2014vs, 1960vs, 1930sh, 1900w, 1835w (Fe-H str.), 1772w, 1762w, 1670w, 1430m, 1363m, 1330vw, 1289w, 1258w, 1246vw, 1220w, 1168w, 1113m, 1058m, 1013s, 1000s, 936vw, 917w, 830m, 806vw, 782w, 747s, 689s. The band at 1772 cm.-1 is due to the bridging CO of the binuclear species which is slowly formed.

n-05H5Fe(C0)2D: 3110w, 2014vs, 1960vs, 1932sh, 1900vw, 1835vw, 1770w, 1762w, 1670w, 1428m, 1363m, 1330m (Fe-D str?), 1290w, 1258w, 1240, 1219w, 1168w, 1113m, 1060m, 1010s, 1002s, 917w, 830m, 782w, 747s, 689s.

n-05H5Ru(C0)2R: 3100w, 2025vs, 1966-vs, 19388h, 1853w (Ru-H str.), 1812w, 1730w, 1435m, 1353m, 1257m, 1199w, 1090m, 10618h, 1010m, 1001sh, 904vw,

816sh, 805s, 729m, 688m. 77.

7c-C,H,Ru(C0)2D: 3098w, 2025vs, 1966vp, 1935sh, 1852vw, 1810w, 1730w, 1435m, 1353m, 1325vw (Ru-D str?), 1257m, 1199w, 1090m, 1061sh, 1008m,

1000shy 819sh, 806s, 729m, 685m.

7c-05H5Mo(C0)3H: ▪ 2027vs, 2020,41, 1940vs, 1904m, 1790 (Mo-H str.), 1730w, 1640w ... 1350w, 1275w

1-05H5Mo(C0)3D: • 2027v s, 2020sh, 1939vs, 1906m, 1847vw, 1790vw, 1731w, 1635w ... 1353w, 1285m (Mo-D str.), 1262w ...

7t-05H5W(00)3H: 2026vs, 2016sh, 1935vs, 1900sh, 1845m (w-H str.), 1735w, 1635w ... 1352w, 1320w, 1253vw

IT-05H5W(C0)3D: 2026vs, 2016sh, 1935vs, 1900sh, 1843vw, 1735w, 1655w ... 1352w, 1322m (w--D str.), 1253vw

n-05H5Fe(C0)2CH3 : 2016vs, 2005sh, 1925vs, 1845w, 1734w, 1655w ...

%-05H5Ru(00)2CH3 : 3095w, 2960w, 2905my 2028vsy 1960vs, 1933sh, 1845vw, 1805vwy 1740vw, 1425m, 1354m, 1196m, 1105sh, 1062m, 1002s, 977s, 810s,

756m, 680w.

7c-05H5Ru(C0)202H5: •3092w, 3053m, 2910s, 2029vsy 1960vs, 19308h, 1807vwy

1756vw, 1428m, 1371w, 1342m, 1304w, 1294w, 1276w, 1256sy 1242sh, 1222my

1187sh, 1169m, 1160sh, 1128m, 1100s, 1065s, 1022my 1001s, 938w, 906s,

895shy 805w, 745sy 725s, 702m.

7:-0511515.0(00)3 C113: ...2023vs, 1934vs, 1904sh, 1828w, 1730w, 1642w ...

1-05H5Mo(C0)3C2H ...2022vsy 2005sh, 1932vs, 1900sh, 1817w, 1727wy 1643w... 78.

%-05H5W(C0)3 03 : 2019vs, 1924vs, 1843w, 1737w, 1653w ...

n-05115W(00)302H51 2016vs, 1928vs, 1841w, 1762w, 1667w ...

it-05H5Ru(C0)2I: 3083w, 2055vs, 2007vs, 1976sh, 1840vw, 1420m, 1374vw, 1345w, 1262m, 1090m, 1058m, 1001m, 920w, 823s, 756w, 666w.

lt-O5H5Rh(00)CF3I: 2098vs, 2052w, 1982w, 1865vvw, 1787vw, 1672vw, 1351w, 1073vs, 1040sh, 989sh, 846sh, 832m, 818m.

1-05H5Rh(C0)C2F512 2098s, 2048vw, 1299m, 1272w, 1263sh, 1191s, 1070s,

1063sh, 1023w, 1017sh, 998vw, 990vw, 908s, 831w, 818m, 732m, 657w,bro6A.

yilL4/1?(CO)C3F7I: 2985w, 2099s, 2046vw, 1408w, 1322s, 1242shy 1230s, 1197sy 1172sh, 1163m„ 1091s, 1061s1iy 1053m, 1022sh, 1013w, 990vw, 9327wy

846shy 833my 819s, 800s, 723s, 662w. 79

REFERENCES

1. Hieber and Leutert, Naturwiss., 1931, 12, 360. 2. Hieber and Leutert, Z. Anorg., 1932, 204, 145; Ber., 1931, 64B, 2832. 3. Coleman and Blanchard, J. Amer. Chem. Soc., 1936, 58, 2160. 4. Hieber, Angew. Chem., 1936, 49, 463. 5. Hieber, Schulten and Marin, Z. Anorg., 1939, 240, 261. 6. Hieber and Lagally, Z. Anorg.) 1940, 1142, 321. 7. Anderson, Quart. Reviews, 1947, 1, 331- 8. Hurd) Sentell and Norton, J. Amer. Chem. Soc., 1949, 71, 1899. 9. Rhomberg and Owen, J. Amer. Chem. Soc., 1951, 73, 5904. 10. Sternberg, Wender, Friedel and Orchin, J. Amer. Chem. Soc., 1953, 75, 2717. 11. Friedel, Wender, Schufler, and Sternberg, J. Amer. Chem. Soc., 1953, j, 3951. 12. Birmingham and Wilkinson, J. Amer. Chem. Soc., 1955, 77, 3421. 13. Jorgensen, Acta Chem. Scand., 1955, 2_, 605. 14. Fischer, Hafner and Stahl, Z. Anorg., 1955, 282, 47. 15. Fischer and Hafnery Z. Nat., 1955, 10b, 140. 16. &igen, Magee and Gallup, J. Amer. Chem. Soc., 1956, 78, 4185. 17. Cotton and Wilkinson, Chem. and Ind., 1956, 1305. 18. Piper and Wilkinson, J. Inorg. Nuclear Chem., 1956, 3, 104. 19. Hieber and Werner, Ber., 1957, 22, 278. 20. Hieber and Brendel, Z. Anorg., 1957, 282, 324. 21. Behrens and Klek, Z. Anorg., 1957, 22Z, 151. 22. Behrens and Weber, Z. Anorg., 1957, al, 122. 80.

23. Hieber and Werner, Ber., 1957, 90, 286. 24. Chatty Duncanson and Shaw, Chem. and Ind., 1958, 859. 25. Rosenblum and Santery J. Amer. Chem. Soc., 1959, 81, 5517. 26. Green, Pratt and Wilkinson, J. Chem. Soc., 1958, 3916. 27. Griffith, Pratt and Wilkinson, Nature, 1958, 182, 466• 28. Hieber and Wagner, Z. Nat., 1958, 13b, 339. 29. Wilson, Z. Nat., 1958, 12h, 349. 30. Chatt and Haytery Proc. Chem. Soc., 1959, 153. 31. Rosenblum and Santer, J. Amer. Chem. Soc., 1959, 81, 5517. 32. Green, Street and Wilkinson, Z. Nat., 1959, 14b, 738. 33. Davison and Wilkinson, unpublished work.

34. Hieber and Legally, Z. Nat., 1959, 141), 132. 35. McCleverty and Wilkinson, unpublished work. 36. Colton, Dalziely Griffith and Wilkinson, Nature, 1959, 112, 328. 37. Hieber and Braun, Z. Nat., 1959, at, 132. 38. Down and Wilkinson, unpublished work.

39. Bishop, Down, Emtagey Richards and Wilkinson, J. Chem. Soc., 1959, 2484.-

40. Hieber, Englert and Rieger, Z. Anorg., 1959, 300, 295. 41. Hieber and Englerty Z. Anorg., 1959, 222, 311. 42. Behrens and Haag, Z. Nat., 1959, 14b, 600. 43. Griffith and Wilkinson, J. Chem. Soc., 1959, 2757. 44. Cotton, Down and Wilkinson, J. Chem. Soc., 1959, 833. 45. Edgell, Asatoy Wilson and Angell, J. Amer. Chem. Soc., 1959, 81, 2022. 46. Curpheyy Santer, Rosenblum and Richards, J. Amer. Chem. Soc., 1960, 82, 5249. 81.

47. Griffith and Wilkinson, unpublished work. 48. Chatty Hart and Hayter, Nature, 1960, 55. 49. Chatt and Shaw, Chem. and Ind., 1960, 931. 50. Ginsberg, Miller, Cavanaugh and Dailey, Nature, 1960, 112, 528. 51. Floss and Grosse, J. Inorg. Nuclear Chem., 1960, 16, 36, 44. 52. Fischer and Hristidu, Z. Nat., 1960, 111, 135. 53. Behrens and Haag, Ber., 1961, 21, 312, 320. 54. Hieber, Kroder and Zahn, Z. Nat., 1960, 15b, 325. 55. Noth and Hartwimmer, Ber., 1960, 21, 2238. 56. Lewis, Nyholm and Reddy, Chem. and Ind., 1960, 1386. 57. Chatt and Hayter, J. Chem. Soc., 1960, 2605. 58. Vaska and Diluzio, J. Amer. Chem. Soc., 1961 83, 1262. 59. Chatt and Hayter, J. Chem. Soc., 1961, 5507. 60. Vaska, J. Amer. Chem. Soc., 1961, 83, 756. 61. Hayter, J. Amer. Chem. Soc., 1961, 83, 1259. 62. Malatesta, Angoletta, Araneo, and Canziani, Angew. Chem., 19615 73, 273. 63. Vaska and Diluzio, J. Amer. Chem. Soc., 1961, 83, 2784. 64. Wilkinson, Proc. Chem. Soc., 1961, 72. 65. Freni and Valenti, Gazz., 1961, 21, 1357. 66. Malatesta, Frani and Valenti, Angew. Chem., 1961, ,3„.3,, 273. 67. Hieber and Lindner, Z. Nat., 1961, 16b, 137. 68. Edgell and Summitt, J. Amer. Chem. Soc., 1961, 83, 1772. 69. Green, McCleverty, Pratt and Wilkinson, J. Chem. Soo., 1961, 4654. 70. Podall, Prestridge and Shapiro, J. Amer. Chem. 'Soc., 1961, 83, Z05. 82.

71. Hieber, Beck and Zeitler, Angew. Chem., 1961, f3, 364. 72. Behrens and Lokofer, Ber., 1961, 21, 1391. 73. McCleverty and Wilkinson, Chem. and Ind., 1961, 288.

74. Werner and Podall, Chem. and Ind., 1961, 144. 75. Werner and Podall, Z. Nat., 1961, 16b, 477. 76. Chatt and Shaw, J. Chem. Soc., 1962, 5075. 77. Davison, McFarlane, Pratt and Wilkinson, J. Chem. Soc., 1962, 3653. 78. Beck, Hieber and Braun, Z. Anorg. Chem., 1961, 308, 23. 79. Hileman, Huggins and Kaesz, Inorg. Chem., 1962, 1, 933. 80. Hieber and Lindner, Ber., 1961, 24,,273. 81. Stevens, Kern and Lipscomb, J. Chem. Phys., 1962, 32, 279. 82. Hieber, Faulhaber and Theubert, Z. Anorg., 1962, 11,1, 125. 83. Hieber and Tengler, Z. Anorg., 1962, 318 136. 84. Green and Moelwyn—Hughes, Z. Nat., 1962, El, 783. 85. Hieber, Winter and Schubert, Ber., 1962, 21, 3070. 86. Gillard and Wilkinson, J. Chem. Soc., 1963, in press. 87. Davison, McCleverty and Wilkinson, J. Chem. Soc., 1963, 1133. 88. Winkhaus, Z. Anorg.) 1963, 3220 404. 89. Jackman, "Nuclear Magnetic Resonance Spectroscopy", Pergamon Press) London, 1959. 90. Fritz, Hristidu, Hummel and Schneider, Z. Nat.) 1959, 15b, 3604. 91. Liehr, Naturwiss., 1957, 44, 61. 92. See Gillespie and Nyholm, Quart. Rev., 1957, 1, 339; "Progress in Stereochemistry", Butterworths, London, 1958, Vol. II; Nyholm, op cit., Vol. I, 1954. 83.

Natta, Corradini and Bassi, J. !Tiler. Chem. Soc., 1958, 80, 755.

94. Ballhausen and Dahl, Acta Chem. Scand., 1961, 15, 1333. 95. Potter and Wilkinson, unpublished work. 96. Dave, Evans and Wilkinson, J. Chem. Soc., 1959, 3684.

97. Fritz and Fischer, J. Chem. Soc., 1961, 547. 98. Narasimhan and Rogers, J. Chem. Phys., 1961, 21., 1049. 99. Evans and Maher, Proc. Chem. Soc., 1961, 208. 100. Allred and Rochow, J. Inorg. Nuclear Chem., 1958, 5, 269. 101. Slomp, J. Amer. Chem. Soc., 1962, 84, 673- 102. Narasimhan and Rogers, J. Amer. Chem. Soc., 1960, 82, 5983. 103. Corio, Chem. Rev., 1960, 60, 363. 104. Closson, Kozikowski and Coffield, J. Org. Chem., 1957, 22, 598. 105. Piper and Wilkinson, Naturwiss., 1956, 43, 57. 106. SInchukerev, Kolbin and Ryabov, Z. Neorg. Khim., 1961, 6, 517. 107. Piper and Wilkinson, J. Inorg. Nuclear Chem., 1956, 2, 38. 108. Piper and Wilkinson, J. Inorg. Nuclear Chem., 1956, 3, 38. 109. Fischer and Vogler, Z. Nat., 1962, 17b, 421. 110. Podall, J. Chem. Educ., 1961, 38, 187, and references therein. 111. Hieber and Fuchs, Z. Anorg. Chem., 1941, 248, 256. 112. Green, Angew. Chem., 1960, 11, 719, and references therein. 113. Closson, Coffield and Kozikowski, International Conference on Coordination Chemistry, London, 1959, Special Publication No. 13, 126. 114. Abel, Singh and Wilkinson, J. Chem. Soc., 1960, 1321. 115. Pitcher, Buckingham and Stone, J. Chem. Phys., 1962, 362 124- 84.

116. King, Stafford, Treichel and Stone, J. Amer. Chem. Soc., 1961, 83, 3604. 117. Green, Pratt and Wilkinson, J. Chem. Soc., 1959, 3753. 118. Lewis, Science Progress, 1961, Ai, 67. 119. Vaska, International Conference on Coordination Chemistry, Stockholm, 1962, Proceedings Abstract, pp. 266-267. 120. Halpern, Quart. Rev., 1956, 10, 463 and references therein. 121. Winfield, Aust. J. Sci. Res., 1951, A4, 384; Boston, King and Winfield, Advances in Catalysis, 1957, 2, 312 (Academic Press, New York, 1957). 122. Orioli and Vaska, Proc. Chem. Soc., 19622 333. 123. Owston, Partridge and Rowe, Acta Cryst., 1960; 13, 246. 124. Bennett, Gerloch, McCleverty and Wilkinson, Proc. Chem. Soc., 1962, 357. 125. McConnell, J. Chem. Phys., 1957, 27, 226. 126. Pople, Schneider and Bernstein, High Resolution Nuclear Magnetic Resonance (McGraw-Hill Book Company, Inc., New York, 1959), pp. 175-183. 127. Cavanaugh and Dailey, J. Chem. Phys., 1961, 34, 1099. 128. Schoolery, J. Chem. Phys., 19532 21, 1899; Dailey and Schoolery„ J. Amer. Chem. Soc., 1955, 77, 3977. 129. Bothner-By and Naar-Colin, Ann. N.Y. Acad. Sci., 1958, 12, 833; Narasimhan and Rogers, J. Chem. Phys., 1959, 31, 1302; Tillieu and Guy, J. Chem. Phys., 1956, 24, 1117.

130. Spiesecke and Schneider, J. Chem. Phys., 1961, 35, 722. 131. McCleverty and Wilkinson, unpublished work. 132. Manchot and Konig, Ber., 1924, 57, II, 2130. 133. Saika and Slichter, J. Chem. Phys., 1954, 22, 26. 134. Gutowsky and Hoffman, J. Chem. Phys., 1951, 19, 1259. 85.

135. Schoolery, Varian Tech. Inform. Bull., 1955, 1, 3; Schoolery and Weaver, Ann. Rev. Phys. Chem., 1955, 6, 433.

136. Pitcher, Buckingham and Stone, J. Chem. Phys., 1962, 36, 124.

137. Tiers, J. Amer. Chem. Soc., 1956, 78, 2914; Gutowsky, Nuclear Magnetic Resonance. Physical Methods, Part IV of Technique of Organic Chemistry, edited by A. Weissberger (Interscience Publishers, New York, 1960), Vol. I, 3rd ed., p. 2663.

138. King Stafford, Treichel and Stone, J. Amer. Chem. Soc., 1961, 83, 3593• 139. Fischer and Bittler, Z. Nat., 1961, 16b, 835.

140. Haszeldine, J. Chem. Soc., 1952, 4259- 141. Chambers, Clark and Willis, Canad. J. Chem., 1961, 22, 131. 142. Treichel, Morris and Stone, J. Chem. Soc., 1963, 720.

143. Crapo and Sederholm, J. Chem. Phys., 1960, 33, 1583.

144. Muller, Lauterbur and Svatos, J. Amer. Chem. Soc., 1957, 12, 1807. 145. Glick and Bothner—By, J. Chem. Phys., 1956, gl, 362. 146. Jones and Wilkinson, unpublished work.

147. Churchill and Mason, Proc. Chem. Soc., 1963, 112.

148. McClellan, J. Amer. Chem. Soc., 1961, 83, 1598. 149. Cotton and Wilkinson, Advanced Inorganic Chemistry, Interscience Publishers, 1962.