PART I ,
SUBSTITUTED GROUP VIB METAL CARBDNYL
COMPLEXES 0F DIMETHYL METHYLPHDSPHONITE ,
A AND BIS (DIMETHYLAMINO) MET HYLPHDSPHINE

PARTII »=_
THE SYNTHESIS AND SPECTROSCOPIC PROPERTIES '
OF SOME CYANOPHOSPHINES AND THEIR 3 _
CHROMIUM AND MOLYBDENUM CARBONYL COMPLEXES

 

 

' ‘ Thesis for the Degree of'PhD
MICHIGAN STATE UNIVERSITY
CHRISTOPHER E JONES
1971

 

.‘ H53“;

 

 

This is to certify that the

thesis entitled
ParT I. Subsfitu¢ed Group VlB MeTal Carbonyl Complexes of

DimeThyl MeThylPhosphonITe and Bis(Dime+hylamino)Methyl-
phosphine. Parf II. The SynThesls and SpecTroscopic
PrOperfles of some Cyanophosphines and Their Chromium and
Molybdenum Carbonyl Complexes.

presented by

ChrisTOpher E. Jones

has been accepted towards fulfillment
of the requirements for

Ph .D . degree in Chemi s‘rry

 

 

KMW

Major professor

Date February I9, I97!

 

0-7639

 

 

 

 

ABSTRACT

PART I

SUBSTITUTED GROUP VIB METAL CARBONYL COMPLEXES
OF DIMETHYL METHYLPHOSPHONITE AND
BIS(DIMETHYLAMINO)METHYLPHOSPHINE

BY

Christopher E. Jones

The syntheses of substituted metal carbonyls of the
type [M(co)6_XLX] where M = Cr, Mo, and W and where X =
1, 2, or 3 for L = CH3P(OCH§)2, and where X = 1 or 2 for
L = CH3P(N(CH3)2)2 are described. The infrared data in
the CO stretching region and CO stretching force constants
are discussed and an order is proposed for the v-acceptor
strength of the ligands: P(OCH3)3 ; CH3P(OCH3)2 > (CH3)3P
> CH3P(N(CH3)2)2 > P(N(CH3)2)3. The proton nmr data is
also discussed for these complexes and tentative assignment

is made for the sign of 2J coupling of phosphorus to the

PH (
methyl protons). In the di- and trisubstituted complexes,
phosphorus—phosphorus coupling is observed in the proton

nmr Spectra with the absolute value oftzJpp being larger in
the trans isomer than in the gi§_isomer for a given com-I
pound and larger in M0 and W compounds than in analogous Cr
compounds. These Spectral data are used to help interpret

the nature of the metal-phosphorus bond and to determine

the stereochemistry of the complexes which are prepared.

 

 

 

Christopher E. Jones
PART II
THE SYNTHESIS AND SPECTROSCOPIC PROPERTIES OF
SOME CYANOPHOSPHINES AND THEIR
CHROMIUM AND MOLYBDENUM CARBONYL COMPLEXES

A rapid and easy method of preparation is presented

for the routine synthesis of compounds of the type RxP(CN)3-x

(where x = O, 1, or 2 and R = Ph-, CH50-, C2H50-, (CH5)2N-,
or CH3-). The infrared, proton nmr, phosphorus—31 nmr,

and mass spectral data are presented and discussed. From
this series of cyanophosphines, the reactions of the com—
pounds thPCN, PhP(CN)2, (CH3)2PCN, (C2H5O)2PCN, and
[(CH3)2N]2PCN with M(CO)4C7H3 (M = or or Mo and C7H§ is
norbornadiene) have been investigated. From these reactions,
yellow crystalline complexes of the form [M(CO)4L]2, where

L is a bridging cyano-phosphine ligand, have been isolated
Their formulation is based on analytical data and their

,fiE —-CN\
\\NC-fi:’

of the infrared and proton nmr Spectral data. These bridged

structure is proposed as (CO)4M M(CO)4 on the basis
complexes react further with another molar equivalent of
ligand (L) or a different ligand (L') to yield complexes
of the type M(CO)4L2 and M(CO)4LL'. The products from these
reactions further support the proposed structure for the
bridged Species. The proton nmr, infrared, and some of the
mass Spectral data of these bridged complexes are also pre-

sented and discussed.

 

 

 

PART I
SUBSTITUTED GROUP VIB METAL CARBONYL COMPLEXES
OF DIMETHYL METHYLPHOSPHONITE AND

BIS(DIMETHYLAMINO)METHYLPHOSPHINE

PART II
THE SYNTHESIS AND SPECTROSCOPIC PROPERTIES OF
SOME CYANOPHOSPHINES AND THEIR

CHROMIUM AND MOLYBDENUM CARBONYL COMPLEXES

By .1
- I I"!

Christopher EDTJones

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

1971

 

 

.._. __ ..,-_4'

 

 

To Gail

ii

 

II“

 

 

 

ACKNOWLEDGMENT

I would like to extend my appreciation to Dr. Kenneth

J. Coskran, for his interest, patience, and encouragement

during this investigation.

I am deeply grateful to my wife, Gail, for her inspira-
tion and unrelenting patience. I wish, also, to thank my
parents, Mr. and Mrs. Theodore E. Jones of Traverse City,
Michigan for their assistance, guidance, and encouragement

during my entire educational career.

iii

TABLE OF CONTENTS

PART I
Page

INTRODUCTION 0 C O O O O 0 C O O O O O O O O O O O O O 1
EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . 12 .
RESULTS MD DISCUSSION . O O O O O C O O O C O O O O O 21

Infrared Spectra . . . . . . . . . . . . . . . . 23
Proton Nmr . . . . . . . . . . . . . . . . . . . 34
Phosphorus-31 Nmr . . . . . . . . . . . . . . . . 45

Stereochemistry . . . . . . . . . . . . . . . . . 50

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 55

EXPERIMENTAL . . O O O C C O O O O O O O . O O O O O O 58 i ‘

Materials . . . . . . . . . . . . . . . . . . . . 59

Preparation of Cyanophosphine Compounds . . . . . 60

ms .. _.._. _,. ‘
1

Reactions of P(CN)3 with Various Metals . . . . . 63

Preparation of Metal Complexes of Cyanophosphine
Ligands Other than P(CN)3 . . . . . . . . . 64

iv

TABLE OF CONTENTS (Continued).

RESULTS AND DISCUSSION . . . . . . . . . . . . .

Cyanophosphine Compounds . . . . . . . . . .
Mass Spectra of Cyanophosphines . . . . . .
Nmr Spectra of the Ligands . . . . . . . . .
Metal Complexes of CyanophOSphines . . . . .

Infrared Spectral Data of Metal Complexes of
Some Cyanophosphine Ligands . . . . . .

Proton Nmr of Metal Complexes . . . . . . .
Mass Spectra of Complexes . . . . . . . . .

Structure of Bridged Complexes . . . . . . .
CONCLUS ION O O O O O O O O O O O O O O O O O O O
SUGGESTIONS FOR FUTURE WORK . . . . . . . . . .

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . .

Page
68

68
68
71
74

75
82
88
89

96
99

101

10.

 

LIST OF TABLES

PART I

CO stretching frequencies and force constants
Secular equations for CO stretching modes .
Proton nmr data . . . . . . . . . . . . . .

31P nmr spectral data . . . . . . . . . . .

PART II

Experimental data for cyanophosphines . . .

Experimental and analytical data for metal
complexes . . . . . . . . . . . . . . .

Mass spectral data for cyanophosphines . . .

Infrared and nmr Spectral data of cyanophos-
phines C O O O O O O O O O O O O 0 O O O

Cyanide and carbonyl infrared frequencies for
metal complexes . . . . . . . . . . . .

1H nmr of complexes . . . . . . . . . . . .

vi

Page

24
32
35
49

62

65

7O

72

76

83

FIGURE

1.

6.

LIST OF FIGURES

PART I
Page

o—bonding through donation of carbon lone
pair to an empty metal d-orbital . . . . . 5

7—bonding through donation of metal d-electrons
to CO antibonding orbitals . . . . . . . . 6

Examples of the CO stretching region of the
infrared spectra for mono— and disubstituted
complexes. A is Mo(CO)5 (CH3P(OCH3)2 ); B is
trans-Mo (co) [CH3P(N( CH3)2)2]2; g'is cfs-
MOZCO) 4[éH3P?OCH3)] )12 2 o o o o o o o o o o 27

Examples of the CO stretching region of the
infrared Spectra for fac- and mer- -trisubsti-
tuted complexes. D is faCeMoE5672:

[CH3P EOCH3g2 2]3 and E is mersw CO)
[CH3P OCH3 213 . . . . . . . . . . . . . . 29

Examples of 1H nmr spectra of metal complexes
which illustrate spectra with and without
the presence of phosphorus-phosphorus coup-
ling. In both spectra L = CH3P(N(CH3)2)2
The more intense downfield peaks are for
the (CH3)2N- group while the less intense
upfield peaks are from the CH3- resonances.
The tran84W(CO)4L2 is a good example of an
anA'X'n type Spectrum . . . . . . . . . . 38

Shapes of the proton nmr resonances of some tri-
substituted complexes of CH3P(OCH3g2 (L).
A is m§£7W§COg3L3; B is fgng(CO 3L3; and
C is ngfCr CO 3L2. The methyl resonance is
at higher field in all cases. Those resonances
in the mggfcomplexes labeled with the letter
t are assigned to the two ligands trans to
each other, and give rise to a near "trip—
let". Those resonances labeled d refer to
the "doublet" which results from the ligand
cis to the other two ligands (see text). In
C, wassignment of the central resonance in the
"triplets“ is not obvious, and is therefore,
not labeled . . . . . . . . . . . . . . . 4O

 

 

 

 

LIST OF FIGURES (Continued).

FIGURE Page

7. 31p nmr spectra of CH2P(OCH3)2, CH2P(N(CH3)2)2,
and three of the metal complexes of these
ligands . . . . . . . . . . . . . . . . . .46

PART II

8. Shapes of the infrared bands in the CO region
§_represents the general shape of the SEE”
disubstituted complexes and B the bridged
complexes . . . . . . . . .'T . . . . . . 78

9. 1H nmr Spectra of the mixed ligand complexes.
In both the Cr and the Mo complexes, L =
[(CH3)2N]2PCN and L' = (CH3O)3P. (See
Table 10 for chemical shifts and coupling
constants.) . . . . . . . . . . . . . . . 86

10. Two possible isomers for the bridged dimeric
complex. Dotted lines indicate expected
cleavage by a phosphorus ligand . . . . . 90

11. Examples of proton nmr spectra used to
determine structure of bridged complex . . 94

viii

 

 

 

 

INTRODUCTION

In recent years, metal carbonyl chemistry has been so
thoroughly investigated that to attempt any sort of compre—
hensive review, even in the specific area of group VIB
carbonyl chemistry, which is of particular interest here,
would be somewhat futile. There have been several reviews
in the general area of metal carbonyl chemistry. An excel—
lent review of transition metal complexes of phosphines,
arsines, and stibines has been published by Booth (1); re—
views of cyclopentadienyl metal carbonyls have been pub-
lished by Wilkinson and Cotton (2), Fischer and Fritz (3,4),
Pauson (5), and Zeiss (6); ultravioletminduced reactions of
metal carbonyls have been discussed by Strohmeier (7); both
Abel (8) and Hieber (9) have published very good reviews of
metal carbonyls;and a very extensive review of the group
VIB metal carbonyl derivatives has been presented by Dobson,
Stolz, and Sheline (10). Therefore, except where such in—
formation is useful in elucidating the work presented here,
repetition of these reviews will be avoided.

A great deal of the work on tertiary phosphorus and
amine complexes of the transition metal subgroup VIB (Cr,
Mo, W) has been done in the last decade. Emphasis has been
placed both on the synthesis of such complexes, and on the

1

 

2
nature of the metal ligand bonding. Metal ligand bonding
has been studied by means of infrared, nmr, and other spectral
techniques. Infrared Spectroscopy has by far received the
most widespread attention in studying metal carbonyl com-
plexes. The number and shape of the infrawred bands in the
carbon-oxygen stretching region can be used to determine
the degree of substitution and the geometry of the metal
complexes. Shifts in the CO stretching frequency upon
substitution can be used to study metal-ligand bonding, but
a more rewarding approach involves the use of CO stretching
force constants. In 1962 Cotton and Kraihanzel(11) published
a rather straightuforward set of secular equations for cal-
culating these force constants for six-coordinate metal
carbonyls of octahedral symmetry. These equations were used
to begin a comparative study of metal-carbonyl, metal-phos—
phorus, and metal-nitrogen bonding in complexes of the type
M(CO)6_XLX, where M = Cr, M0, or W; X = 0,1,2, or 3; and
L = tertiary phosphorus or amine ligands (11,12,13). Al.
though the carbonyl stretching frequencies had been used
previously with some success to discuss the nature of metal
ligand bonding, the use of force constants greatly improved
the interpretations of the infrared data.

The useful application of nmr techniques to transition
metal complex chemistry has been limited somewhat to recent
years. Before 1963, proton nmr spectroscopy was used only
sparingly in this area. Reports of nmr spectra of metal

complexes generally included only changes in chemical shifts

 

3
and coupling constants of the ligands upon complexation.
In 1963 Jenkins and Shaw (23) observed the phenomenon of
“virtual coupling" of phOSphorus atoms in the proton nmr
spectra of trans-PdIz[P(CH3)2(C6H5)]2 and related complexes.
They used this virtual Coupling to determine the structures
of these palladium complexes. The proton nmr spectra of
some of the complexes which Jenkins and Shaw investigated
displayed a methyl resonance consisting of three lines
with the central line varying in intensity from complex to
complex. A doublet, due to the splitting of the protons
by phosphorus,would have been the expected resonance, how—
ever, the resonances of each complexed ligand are further
coupled to the phosphorus atom of the ligand gg§_or EEEEE
to it in the complex. The resulting spectrum may appear
as a 1:2:1 triplet as is found for cases of large phos-
phorus~phosphorus coupling,amd as was observed for the E£22§_
complexes above. If the phosphorus-phosphorus coupling is
weak, the intensity of the central peak may be so small
that the spectrum appears as a doublet. The gig complexes
of palladium were found to give only doublets in the methyl
region of the proton nmr. Therefore, Jenkins and Shaw pro-
posed that the geometries of disubstituted palladium com-
plexes could be ascertained from the magnitude of the phos-
phorus-phosphorus coupling (i.e. the intensity of the third
peak in the methyl resonances of the proton nmr). Later
in 1963 King (37) reported similar spectra for metal car-

bonyl complexes of tris(dimethylamino)phosphine, and made

4

some rough comparisons of the strength of virtual coupling
in going from one metal to another. In 1965 Verkade gt gi.
(48) reported more metal—phosphorus complexes which demon—
strated this phenomenon, and presented some crude estimates
of 31P—31‘P coupling strengths. They also observed coupling
in some complexes of gi§_geometry and made further compari—
sons between metals. Verkade's work was followed by a theo—
retical interpretation by Harris (16,17) of the virtual
coupling of phosphorus atoms in the proton nmr spectra of
these systems. Complexes which have two or more phosphorus
ligands bound to a metal are classified as XnAA'XA type
systems, where X and A represent the protons and phos—
phorus, respectively of one ligand bound to a metal and X'
and A' represent the protons and phosphorus, respectively,
of a second ligand bound to the same metal. The subsequent
development and application of this theory by Verkade (18,
19,20,21) and Shaw (22,27) greatly improved the interpreta—
tion of proton nmr for transition metal complexes. Because
of their work, proton nmr became a useful tool in the
determination of structures of transition metal complexes
and in the study of metal—phosphorus bonding.

Phosphorus—31 nmr has also been applied to metal com—
plex chemistry in recent years. In 1961, Meriwether and
Leto (41) reported 31P chemical shifts of metal complexes.
In 1966 the 31P chemical shifts for several phosphine sub—

stituted tungsten carbonyl complexes were reported by Grim

5
et al. (42), and for molybdenum carbonyl complexes by Lenzi
and Poilblanc (43). Bertrand §£_§l. (19) also reported 31P
data for several metal carbonyl complexes, and other authors
(44,45,49) lent interpretation and data to this area of nmr
spectroscopy.

Relatively little use has been made of ultraviolet and
visible spectra in studies of the group VIB carbonyl com—
plexes. In 1969 Abel gt_§l. (28) studied the ultraviolet
and visible spectra of the isoelectronic series V(CO);,
Cr(CO)6, W(CO)6, and Re(CO);, and discussed the results in
terms of metal—ligand W—bonding.

A significant amount of the work involving chromium,
molybdenum, and tungsten carbonyl complexes of phosphorus
and nitrogen donor ligands, has been concerned with the
nature of metal-ligand bonding. The accepted metal—carbon
bonding scheme in metal carbonyls is a "synergic" one in
which bonding is accomplished through o-donation of a lone
pair of electrons from the carbon atom to an empty d orbi—

tal of the metal (Figure 1) in conjunction with a back

QMG + .cao: ——> MW£50

Figure 1.— o—bonding through donation of carbon lone
pair to an empty metal d—orbital (39).
donation of electron density from filled metal d-orbitals
to vacant p—v* antibonding orbitals on the carbonyl group

(Figure 2)o

“Q _Q Mflmma
CM.C7¢: 0CD MQW OE)

Figure 2.— w-bonding through donation of metal d- elec-
trons to CO antibonding orbitals (39).

Bonding between a metal and a coordinately bound
ligand can take place, in general, in one of three ways.
The bonding can invoke a o-bond (ligand to metal donation)
reinforced by a d—v to p-w* metal to ligand w-bond as
stated for the carbonyls. A second method of bonding in-
volves back donation of metal d-v electron density to
empty ligand d-v orbitals in addition to the coordinate
o-bond, and is the bonding system generally accepted for
phosphorus-metal coordinate bonding. The third method is
that believed to exist in nitrogen-metal coordinate bonding
which uses only a o-bond and little or no w-bonding.

In octahedral (or very nearly octahedral) metal com—
plexes, the metal o-orbitals and the metal v-orbitals are
essentially separate with little or no mixing, and therefore,
can be treated independently ("symmetry factoring") (11).
When this idea is applied to metal carbonyl systems such
as Cr(CO)6, Mo(CO)6, and W(CO)6 and their derivatives, it
is possible to discuss metal to ligand w-bonding in terms
of carbonyl stretching frequencies. Since metal d—w to

carbonyl p-W* donation involves an antibonding orbital on

 

 

 

7
the CO group, the carbon oxygen bond should be weakened,
and should cause a lower energy infrared stretching fre~
quency to be observed. This involvement of CO antibonding
orbitals is borne out by a comparison of free CO, which has

1 (10), with group VIB

a stretching frequency of 2133 cm—
metal carbonyls which have stretching frequencies at ap-
proximately 2000 cm.1 (10). This assumption that the
shift to a lower stretching frequency is a result of
donation of metal electron density into the CO antibonding
orbitals is also reflected in the isoelectronic series,
Mn(C0):, Cr(CO)6, and v(co){ which exhibit co stretching
frequencies of 2096, 2000, and 1859 cm_} respectively (15).
These observations substantiate that the stretching fre-
quency is dependent on the available metal electron

density (reflected by the net charge on the metal).

Since all of the carbonyl groups bonded to a metal are

essentially in competition with each other for metal w-elec-
tron density, if a CO group is replaced by a ligand of weaker
w-accepting ability, the remaining C0 groups should withdraw
more electron density into their p-v* antibonding orbitals.
This withdrawal in-turn should weaken the CO bond and pro-
duce a lower energy infrared stretching frequency. Such a
shift to lower energy is in fact observed when carbonyls

are replaced by phosphorus and nitrogen ligands, unless the
ligand is PF3, which appears to be comparable to or better

than CO in v—accepting ability (13,38). The idea that there

 

 

 

8
is an increase in the donation of metal d-v electron density
to the remaining carbonyl groups in these substituted metal
carbonyls is also supported by the available information
on metal—carbon bond strengths and bond lengths in group
VIB carbonyls. Adams (29) and others (30,31) have provided
some information on metal carbon stretChing frequencies,
v(M—C), and metal-carbon-oxygen deformation frequencies,
6(MCO), for metal carbonyls and their derivatives. Verkade
et al0 (32,33) investigated similar infrared data for a
variety of substituted metal carbonyls and made correlations
to the W—bonding ability of the various ligands as compared
to a carbonyl groupo As poorer V—accepting ligands replaced
carbonyl groups, the metal—carbon stretching frequency in—
creased. Thus, as the CO bond gets weaker, the M-C bond
gets stronger as a result of the increased metal—carbonyl
v—bonding. Problems with this hypothesis arise from the
difficulty in assignment of these low energy infrared ab—
sorptions as well as from the difficulty which arises from
the coupling of v(M—C) and 6(M-C-O) (10). An indication
as to the validity of this interpretation of the changes in
the metal-carbon stretching frequencies could be obtained
from crystal structure data, since increases in metal—car—
bon multiple bonding should be accompanied by a shortening
of the M-C bond. Unfortunately there are few structural
data available on these systems. Crystal structures have

been reported for Cr(CO)6 (34) and for Cr(CO)3(PH3)3 (35),

 

9

and the average metal—carbon bond distance indeed does
shorten from 1.909 :t 0.003% for Cr(co)6 to 1.84 i 0.012
for Cr(CO)2(PH3)3. “Even more conclusive are the structural
determinations by Grim SE_§i' (46) of Cr(CO)5[P(C6H5)3] and
Cr(CO)5[P(OC6H5)3]. The Cr-C distances in both of these
complexes are shorter than the corresponding distances in
Cr(CO)6. In addition, if changes in the strength of metal—
carbonyl v—bonding are the cause of changes observed in the
infrared spectra, one would predict that Cr—C distances
which are trans to the phosphorus ligand should be shorter
than Cr—C distances for CO’S gig to phosphorus ligands and
trans to other CO's. Also C~O distances for CO's Eggns to
a phosphorus ligand should be longer than those gig to a
phOSphorus ligand. And lastly, since infrared data predict
that P(OC6H5)3 is a better v—acceptor than P(C6H5)3, the
Cr—P distances and the C«O distances should be shorter, and
all of the Cr-C distances should be longer for the P(OC6H5)3
complex than for the P(C6H5)3 complex. Everyone of these
predictions is found to hold. It is of further interest to
note that a comparison of the bond lengths of these two
complexes with those of Cr(CO)3(PH3)3 also fits the pre—
diction by infrared data that the order of F—accepting abil—
ity for these ligands is P(OC6H5)3 > P(C6H5)3 > PH3.

Some authors (30, 36, 47) have argued that the obser-
able changes in the spectra of group VIB carbonyl complexes

upon substitution of a carbonyl group with a ligand of

 

10

weaker w-accepting ability can and should be attributed to
o—bonding effects. Brown and Darensbourg (36) presented an
interesting argument to this effect based on infrared in-
tensities, frequency measurements, and dipole moment deri—
vatives. However, both the far infrared data and the
slight bit of structural data which are available are in
direct contradiction to their suggestion.

Several substituted metal carbonyl complexes of the
type [M(CO)6_XLX], where M = Cr, M0, or W, x = 1, 2, or 3,
and L = a tertiary phOSphorus ligand (in particular,
P(OCH3)3 (14), P(CH3)3 (14,33), and P(N(CH3)2)3 (37)) have
been investigated previously using infrared and proton nmr
techniques. However, no study has been made on substituted
metal carbonyls with ligands of the type CH3PY2 and (CH3)2PY,
where Y = —OCH3 or -N(CH3)2. In this work, group VIB
hexacarbonyl complexes of dimethyl methylphosphonite,
CH3P(OCH3)2, and b£§(dimethylamino)methylphosphine,
CH3P(N(CH3)2)2, are reported. Although the proton nmr
spectra of metal complexes of P(OCH3)3 (20) and P(N(CH3)2)3
(37) have been reported, the invariance of the chemical
shifts and coupling constants has prevented interpretation
in terms of phosphorus metal bonding. Because of the closer
proximity of the methyl protons to the phosphorus atom in
the ligands used in this work, meaningful changes in the
nmr spectra can be observed and correlated with infrared

data to obtain information about the metal phosphorus bond.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11
Cotton and Kraihanzel (11) type force constant calcula—
tions have been combined with the proton nmr results to
help explain the stereochemistry and bonding in various

substituted group VIB metal carbonyls.

 

 

 

 

 

 

EXPERIMENTAL SECTION

The infrared spectra were obtained on a Perkin-Elmer
Model 225 grating spectrophotometer. Sodium chloride
optics were used andlin all cases, hexane was used as a
solvent. Proton nmr spectra were taken in chloroform
solutions on a Varian Associates Model A-56/60-D instru-
ment with tetramethylsilane as an internal standard. The
phOSphorus nmr spectra were obtained on a Varian associates
Model DA—60 spectrometer at 24.29 MHz and are referenced
to 85% ortho-phosphoric acid as an external standard.
Elemental analyses were performed by Galbraith Laboratories,
Inc., Knoxville, Tenn. The gfisomers of pentane, hexane,
and octane were used where these solvents are involved.

* Melting points were taken in glass capillaries

using a Hoover melting point apparatus.

Materials.— The molybdenum and tungsten hexacarbonyls

 

were purchased from Pressure Chemical Co., and chromium
hexacarbonyl was purchased from Strem Chemical Co. (All hexa—
carbonyls were used without purification.) The ligands,
dimethyl methylphosphonite (CH3P(OCH3)2) and §$§(dimethyl—
amino)methylphosphine (CH3P(N(CH2)2)2). were prepared ac-

cording to previously described methods (51). Purification

12

 

 

I III. || IJ

 

 

 

 

13

was achieved by vacuum distillation at the reported tempera—
tures. Molybdenum norbornadiene tetracarbonyl (52),
molybdenum mesitylene tricarbonyl (53), tungsten mesitylene
tricarbonyl (53,54), and chromium cycloheptatriene tri-
carbonyl (52) were also prepared as described elsewhere.
Acetonitrile pentacarbonyl tungsten and gi§-§£§(acetonitrile)—
tetracarbonyl tungsten were prepared by a modification of
previously described methods (55). These complexes were
prepared by irradiation with ultraviolet light (Hanovia

Lamp 654—A10) in a quartz tube for 2 and 5 hours for the
mono— and disubstituted species, respectively. They were

identified by their infrared spectra (55); no further puri—

fication was performed.

(Qimgthyl methylphosphonite)pentacarbonylchromium,
[(CH3)(CH3O)2PCr(CO)5].— A magnetically stirred mixture
of 2.1 g (9.5 mmole) of chromium hexacarbonyl and 1.0 g
(9.2 mmole) of CH3P(OCH3)2 in 50 ml of octane was caused
to reflux for 12 hr under nitrogen. The solution was then
allowed to cool to room temperature and concentrated under
vacuum to about 2 ml. Approximately 5 ml of pentane was
added, and fractional crystallization of this solution in

a dry ice—isopropyl alcohol bath gave first, Cr(CO)6 and
((CH3)(CH3O)2P)2Cr(CO)4 (both identified by their infrared
Spectra» and then the white, crystalline (CH3)(CH30)2PCr(CO)5.

The solvent was decanted, and when the solid (mp 6°) was

warmed to room temperature a yellow liquid product

 

 

 

 

14

remained. Repeated fractional crystallizations gave the
pure product, and any remaining solvent was removed under
vacuum. Anal. Calcd. for C8H9PCrO7: C, 32.00; H, 3.00;

P, 10.03. Found: C, 32.20; H, 3.15; P, 10.57.

(Bis(dimethylamino)methylphosphine)pentacarbonyl-
chromium, [(CH3)((CH3)2N)2PCr(CO)5].— A magnetically stir—
red mixture of 3.0 g (13 mmole) of chromium hexacarbonyl
and 1.8 g (13 mmole) of CH3P(N(CH3)2)2 in 50 ml of octane
was caused to reflux for 15 hr under nitrogen. The result-
ing solution was concentrated under vacuum to about 2 ml
of a brown liquid. Five milliliters of pentane was added,
and the solution, when cooled to —78°, yielded pale yellow
crystals of (CH3)((CH3)2N)2PCr(CO)5,mp 74-760. The pentane
was then decanted, and any remaining solvent was removed
under vacuum. Vacuum sublimation at about 1000 and 1 torr

,produced pale yellow crystals of the complex. A231; Calcd.
for C10H15N2PCIO5: C, 36.80; H, 4.60; P, 9.50. Found:

C, 36.61; H, 4.73; P, 9.61.

gingis(dimethyl methylphosphonite)tetracarbonyl-
chromium, [((CH2)(CH30)2P)2Cr(CO)4].— A magnetically stir~
red mixture of 5.0 g (9.2 mmole) of chromium hexacarbonyl
and 2.1 g (20 mmole) of CH3P(OCH3)2 in 50 ml of octane was
caused to reflux for 24 hr under nitrogen. The solution
was concentrated to about 10 ml under vacuum, and cooled in

a -780 bath, which caused precipitation of white, needlelike

 

 

 

 

|| I IIIJ

 

 

 

 

15

crystals of cis-((CH3)(CH3O)2P)2Cr(CO)4/mp 75—760. Re—
crystallization was achieved in pentane. Anal. Calcd. for

C10H18P2CIOS: C, 31.60; H, 4.747 P, 16.32. Found: C,

31.70; H, 4.87; P, 16.60. The trans isomer of
((CH3)(CH3O)2P)2Cr(CO)4 was observed (infrared spectra)

after several recrystallizations. It was not, however, pro—

duced directly from the reaction of the carbonyl and ligand.

trans-Bis(bis(dimethylamino)methylphOSphine)tetra—

carbonylchromium, [((CH3)((CH3)2N)2P)2Cr(C0)4].- A mag—

netically stirred mixture of 5.0 g (9.1 mmole) of chromium
hexacarbonyl and 2.5 g (19 mmole) of CH3P(N(CH3)2)2 in 50
ml of octane was caused to reflux for 28 hr under nitrogen

to give a brown colored solution. After it had cooled to

room temperature, this solution was filtered. The filtrate,

when further cooled in a -780 bath, yielded yellow crystals.

The octane was decanted, and recrystallization from pentane

 

gave yellow, needlelike crystals of trans—((CH3)((CH3)2N)2P)2—
Cr(CO)4, mp 88-900. Anal. Calcd. for C14H30N4P2Cr04: C,
38.89; H, 6.95; P, 14.35. Found: C, 38.82; H, 6.89; P,

14.44.

Tris—(dimethyl methylphosphonite)tricarbonylchromium,

[((CH3)(CH30)2P)3Cr(CO)3].— A mixture of 0.3 g (1.0 mmole)
of chromium cycloheptatriene tricarbonyl and 0.4 g (3.7
mmole) of CH3P(OCH3)2 in 50 ml of methylcyclohexane was

caused to reflux for 1 hr during which the initially dark

 

 

 

 

16

red solution gradually turned to light yellow. This solu—

tion was then cooled to room temperature and filtered. The

solvent was removed under vacuum and about 5 ml of octane
was added to the oily residue. When this solution was

cooled in a —780 bath, the oily residue formed again. After

the solvent had been decanted, the residue was again dis-

for 2 hr before the

 

solved in octane and cooled at —200
very pale yellow mixture of Egg— and Egg-(CH3)(CH3O)2P)3—
Cr(CO)3, mp 152—1552 was produced. The product was re-
crystallized from pentane. Aggl. Calcd. for C12H27P2Cr09:

C, 31.31; H, 5.88; P, 20.22. Found: c, 31.15;‘H, 5.88;

P, 20.09.

(Dimethyl methylphosphonite)pentacarbonylmplybdgngm,
ing3)(CH3O)2[Mo(CO)5].— A magnetically stirred mixture of
4.1 (15 mmole) of molybdenum hexacarbonyl and 1.6 g (15
mmole) of CH3P(OCH3)2 in 50 ml of methylcyclohexane was
caused to reflux for 4 hr under nitrogen. The clear to
yellow liquid (CH3)(CH3O)2PMO(CO)5 was isolated in a manner
completely analogous to that used for the isolation of
(CH3)(CH3O)2PCr(CO)5. m. Calcd. for C8H9PMOO7: c,
27.91; H, 2.62; P, 9.01. Found: C, 27.99; H, 2.88; P,

9.26.

(gis(dimethylamino)methylphosphinelpentacarbonyl—

molybdenum, [(CH3)((CH3)2N)2PM0(CO)5].- A magnetically

stirred mixture of 2.5 g (9.3 mmole) of molybdenum hexacar~

 

 

Donyl and 1.2 g (9.2 mmole) of CH3P(N(CH3)2)2 in 50 ml of

 

17
methylcyclohexane was caused to reflux for 4 hr under nitro—
gen. The resulting black solution was cooled to room
temperature, filtered, and concentrated under vacuum to
about 2 ml of a brown liquid. The white, crystalline
(CH3)((CH3)2N)2PMO(CO)5, mp 45-470, was isolated and puri—
fied in a manner completely analogous to that used for the
isolation of (CH3)((CH3)2N)2PCr(CO)5. 523;. Calcd. for
C10H15N2PMO(CO)5: c, 32.41; H, 4.05; P, 8.38. Found:

C, 32.60; H. 4.21; P, 8.19.

gig—Bis(dimethyl methylphosphonite)tetracarbonyl-
molybdenum, [((CH3)(CH3O)2P)2MO(CO)4].— A mixture of 1.0 g
(3.3 mmole) of molybdenum norbornadiene tetracarbonyl and
1.0 g (9.2 mmole) of CH3P(OCH3)2 in 50 ml of hexane was
caused to reflux for 1 hr. After the solution was cooled and
filtered and the hexane removed under vacuum, a yellow
solid remained. Recrystallization from pentane gave white,
needlelike crystals of gig—((Cfla)(CH30)2P)2MO(CO)4I mp
67-680. 233;. Calcd. for C10H18P2M008: C, 28.30; H, 4.25;

P, 14.62. Found: c, 28.43; H, 4.36; P, 14.78.

trans—Bis(bis(dimethylamino)methylphosphine)tetra-
carbonylmolybdenum, [MCHngHng)2P)2MO(CO)4].- A mag—
netically stirred mixture of 0.9 g (3.0 mmole) of molybdenum
norbornadiene tetracarbonyl and 1.3 g (9.0 mmole) of
CH3P(N(CH3)2)2 in 50 ml of hexane was caused to reflux for

1 hr under nitrogen. In a manner identical to that used

 

 

illllJ

 

 

18
for the isolation of gig-((CHa)(CH30)2P)2M0(CO)4, yellow,
needlelike crystals of Egggg—((CH3)((CH3)2N)2P)2M0(CO)4,
mp 88—900, were obtained. éggi. Calcd. for
C14H30N4P2MOO4: c, 35.29; H, 6.30; P, 13.03. Found:

C, 35.26; H, 6.37; P, 13.10.

£32:Tris(dimethyl methylphosphonite)tricarbonyl—
molybdenum, [(CH3)((CH3)2N)2RQW>@C93].— A magnetically
stirred mixture of 0.4 g (1.4 mmole) of molybdenum mesitylene
tricarbonyl and 0.5 g (4.6 mmole) of CH3P(OCH3)2 in 25 ml
of hexane was caused to reflux for 3/4 hr under nitrogen.
When the clear solution was allowed to cool to room temper—
ature, precipitation of white, needlelike crystals of fag~
((CH3)(CH2O)2P)3M0(CO)3, mp 171—173? was observed. éggl.
Calcd. for C12H27P3M009: c, 28.60; H, 5.36; P, 18.50

Found: C, 28.56; H, 5.30; P, 18.33.

(Dimethyl methylphosphonite)pentacarbonyltungsten,

[(CH3)(CH3O)2PW(CO)5].- A magnetically stirred mixture of

 

3.0 g (8.4 mmole) of tungsten hexacarbonyl and 0.8 g (7.4
mmole) of CH2P(OCH2)2 in 50 ml of octane was caused to re~
flux for 72 hr under nitrogen. The pale yellow liquid,
(CH3)(CH30)2PW(CO)5, was isolated in a manner completely
analogous to that used for the isolation of (CH3)(CH3O)2P—
Cr(CO)5. Anal. Calcd. for C8H9PWO7: c, 22.21; H, 2.09;

P: 7.16. Found: C, 22.59; H, 2.14; P, 7.01.

 

 

19
(gis(dimethylamino)methylphOSphine)pentacarbonyl—

tungsten, [(CH3)((CH3)2N)2PW(CO)5].- A magnetically stir-

 

red mixture of 1.7 g of (CH3CN)W(CO)5 4.6 mmole) and 0.6 g
(4.5 mmole) of CH3P(N(CH3)2)2 in 50 ml of hexane was caused
to reflux for 4 hr under nitrogen. The solvent was removed
under vacuum and a yellow solid which contained a mixture
0f W(CO)6l (CH3)((CH3)2N)2PW(CO)5, and EEEEE‘
((CH3)((CH3)2N)2P)2W(CO)4 remained. These complexes were
identified by their infrared spectra. Repeated attempts to

purify the desired monosubstituted product failed.

cis-Bis(dimethyl methylphOSphonite)tetracarbonyl—
tungsten,_j((CH3)(CH30)2P)2W(CO)4].- A magnetically stir—
red mixture of 2.0 g (5.2 mmole) of (CH3CN)2W(CO)4 and 1.5 g
(13 mmole) of CH3P(0CH3)2 in 50 ml of hexane was caused to
reflux for 11 hr under nitrogen. The solvent was removed
under vacuum, and the remaining solid when recrystallized
from pentane yielded white crystals of gig—((CH3)(CH30)2P)2-
w(co)4, mp 49—510. 5193;. Calcd. for C10H18P2W08: C, 23.50.

H. 3.52; P, 12.12; Found: C, 23.71; H, 3.42; P, 12.14.

trans—Bis(bis(dimethylaminomethylphosphine)tetracar—

Egnyltungsten. [((CH3)((CH3)2N)2P)2W(CO)4].- A magnetically

stirred mixture of 3.1 g (7.8 mmole) of (CH3CN)2W(CO)4 and

2.1 g (16 mmole) of CH3P(N(CH3)2)2 in 50 ml of hexane was

caused to reflux for 9 hr under nitrogen. The yellow,

_ 0
cryStalline trans-((CH3)((CH3)2N)2P)2W(CO)4, mp 91 93 , was

 

 

 

20
isolated in a manner completely analogous to that used for
the isolation of _c_:_j_._§_—((CH3)((CH3)2N)2P)2W(C0)4. 5122;. Calcd.
for C14H30N4P2WO4: C, 29.80; H, 5.32; P, 11.00 Found:

C, 29.95; H, 5.19; P, 11.23.

fac-Tris(dimethyl methylphosphonite)tricarbggyl-

 

tungsten, [((CH3)(CH3O)2P)3W(C0)3].— A magnetically stir—

 

red mixture of 0.5 g (1.3 mmole) of tungsten mesitylene
tricarbonyl and 0.5 g (5.6 mmole) of CH3P(OCH3)2 in 50 ml
of hexane was caused to reflux for 12 hr under nitrogen.
This solution was cooled to room temperature, filtered, and
the hexane removed under vacuum. The remaining yellow
solid, when recrystallized from pentane, gave white crystals

of Egg-((CH3)(CH3O)2P)3W(CO)3. (See below.)

mer-Tris(dimethyl methylphosphonite)tricarbonyl-

 

tungsten, [((CH3)(CH3O)2P)3W(CO)3]o‘ E2£f((CH3)(CH30)2P)3‘

 

W(CO)3 was prepared in a manner completely analogous to the
preparation of £39-((CH3)(CH30)2P)3W(CO)3 except that the
reaction was caused to reflux for 6 hr in methylcyclohexane.
When this reaction was stopped after 2 hr, infrared analysis
of the product showed the presence of the Eggfisomer. Since
attempts at purification of either isomer resulted in the
loss of the compounds because of decomposition, elemental
analyses were not obtained, and identification was accom-

plished with infrared and nmr spectral data.

 

 

 

 

 

 

 

 

RESULTS AND DISCUSSION

The monosubstituted compounds, (CH3)(CH30)2PM(C0)5
(M = Cr, Mo, W), were prepared in good yields from reaction
of equimolar mixtures of the hexacarbonyl and ligand. The
ease with which the reactions took place decreased in the
order Mo > Cr > W. These complexes are liquids at room
temperature, as Poilblanc and Bigorgne (14) reported for
similar complexes of P(OCH3)3. The isolation and purifica-
tion of these compounds was most easily achieved by fraction—
ally crystallizing a pentane solution in a dry ice—iso-
propanol slush bath. This procedure removed the unreacted
hexacarbonyl and any disubstituted complex which may have
formed, and left the monosubstituted compound in solution.
Further cooling caused precipitation of white crystals of
the monosubstituted complex which melted on warming to room
temperature to give a pale yellow liquid.

The monosubstituted complexes of CH3P(N(CH3)2)2 were
prepared in a similar manner. However, these compounds ex—
isted as crystalline solids at room temperature and
were easily sublimed. The monosubstituted tungsten com—
plex of CH2P(N(CH2)2)2 has not beenisolated in a pure
state. The infrared spectrum of the reaction product ob-
tained from a 1:1 mixture of W(CO)6 and CH3P(N(CH2)2)2

21

 

 

 

 

 

22

showed a mixture of unreacted hexacarbonyl and both the
mono- and disubstituted compounds. Attempted purification
by fractional crystallization did not free the monosub-’
stituted compound of these impurities. The monosubstituted
compound may be unstable and may disprOportionate to the
disubstituted compound and the hexacarbonyl. It is inter—
esting to note that when tris(dimethylamino)phosphine,
0%N(CH2)2)3), was allowed to react with W(C0)2, only the
disubstituted compound was observed although both mono- and
disubstituted compounds of Cr and M0 were isolated (37).

The disubstituted complexes of chromium for either
CH3P(OCH2)2 or CH3P(N(CH3)2)2 were prepared by refluxing
the hexacarbonyl in octane, whereas, themolybdenum com-
plexes were very easily prepared from molybdenum norborna—
diene tetracarbonyl. Previously, disubstituted tungsten
carbonyl complexes had been prepared either by direct re-
action of the ligand with the hexacarbonyl or with a tung—
sten-diene complex such as tungsten cyclooctadiene tetra—
carbonyl (32) or tungsten norbornadiene tetracarbonyl (33).
These diene intermediates were prepared from (CH3CN)3W(CO)2
according to the methods of King and Fronzaglia (54). 'In
the course of this work it was found that (CHaCN)2W(C0)4
could be prepared easily and in good yields by irradiation
of an acetonitrile solution of W(CO)6 with ultraviolet
light. When this complex was allowed to react directly
with CH3P(OCH3)2 or CH3P(N(CH3)2)2, the disubstituted com-

pounds were produced in almost quantitativeyields.

 

 

 

23

The tri-substituted compounds of CH3P(0CH3)2 were
easily formed in good yields from either the metal mesitylene
tricarbonyl complex or the metal cycloheptatriene tricar—
bonyl complex. The ligand CH3P(N(CH3)2)2 would not form a
trisubstituted complex with any of the metals, regardless
of the manner of preparation. When CH3P(N(CH3)2)2 was
allowed to react with molybdenum mesitylene tricarbonyl,
the disubstituted compound ((CH3X03§)2N)2P)2M0(C0)4 was
obtained. This reaction is similar to that of P(N(CH3)2)3
with molybdenum cycloheptatriene tricarbonyl where again,

only the disubstituted compound was isolated (37).

Infrared Spectra.— Table I lists the infrared data
for the carbonyl region of all the compounds studied in
this work. The monosubstituted compounds exhibited the
1three infrared bands, A)2), Agl), and E1 for C4V symmetry
((Figure 3—A). An additional weak band which is assigned
to the Raman active 31 mode was observed. This Raman
(active band becomes weakly active in the infrared since the
:ligand, which has a symmetry lower than the C4V assumed for
:the complex, lowers the overall symmetry of the complex.
@he disubstituted compounds of CH3P(OCH3)2 were primarily
the Eli isomers as was indicated by the four carbonyl
Stretching frequencies expected for C2V symmetry (Figure
B—C) (11). Indications of the Egggg isomers for Cr and W
Were observed in both the infrared and the nmr spectra, but

no trans isomer was observed for Mo. Both the cis and trans

 

 

24

em.ov

 

 

 

 

 

ho.ha o.owwa . . . .
.A . . . .1. 2.: . 1.9.1.2255..5_
Asa ov ea ma o some Ao.oaoav . . . .Aoo N a a a
w . o 0 V A A Ammovzvmamoho
Awm.ov OO mH O.meH . . . . . . VAOU N N N m S
m 0 0 V A A A movzvanmoLso
AV“. NM 5m mHm mfim
vAOUvNASIwGMHu
mm.o bh.mH OH.mH m.NHmH m.¢mmfi o.m¢mH m.wNON vAOUvNAmAmmUOmemUHB
om.o Hw.ma om.ha o.amoa oAn.oeoHV m.oeoa m.amom «AoovnAnAemoovmemoLoz
om.o Se.ma wo.ma m.moma m.HHoH o.emoa o.mmom . on a a a 2
HM ax HM am am 5. am A V A A moovm moro
AHV Amv vAooquzero
em.o oo.na ah.ma o.wmoa vo.msoav o.oeoa h.ooom nAoovAaAaAamovzvammoiz
Hm.o so.ma oa.ma n.85oa o.mwoa o.Hooa c.8som onoVAaAmmooVEemoLs
om.o ow.na om.ha o.oeoa vo.omoav o.ohoa m.oaom aAoovAaAaAamonvammoLos
om.o Ho.oa mm.ma o.mmoa o.omoa o.sooa m.waom onoVANAmmvovmamoLoz
om.o o8.ha mm.ha o.wmoa cAo.vsoHV h.weoa o.aoom onoVANANAmmovzvmmmoLso
om.o oo.ma ow.ma o.seoa c.8woa n.mooa o.aaom HAOOVAaAamoovmmmoLuo
HM. NM HM m Hm mHvHQ “NvH< mAOUsz
m\mmsmvs
.owPSmumcoo Touom also .Qmoflonwswmnm

 

 

mUCmumflOU TOHOH GEM mOHOGTDTSHH1mSHZDubqpn,DJJTFH}JHQBHi

 

 

a
E .
.Ama .mouv monsoosoo “Nassau sues somanmosoo 909mm to some we 2 no osam>m
.Um>uwwno qum
.psmfiqmflmmm 0>Humuswe

U
.N\TC%UE N.o H.t mum mwsmumsoo wouomo

.AH EU o.N HV E0 CH mum wwflosmswwnm

a h

. wGOHfldHOW wflmxwmd

 

 

 

 

% Asm.ov oa.ma mm.ea o.oowa o.wawa c.3woa «AoovaANAamooVEHEULS
uASm.oV mm.wa mm.ea c.8wma o.HSwH o.maoa erovnAaAhmoovmamoLso
w l
as as as am Aswan Advae «Aoovaqzrhos
mm.o ow.ea o.mwwa o.oooa erovaAaAmmoovmmmoLz
He.o om.ea h.Hme o.oaaa aAoovaAaAmmooVEHEULoz
oe.o om.ea o.as.wa o.oooa onoveAaAamoovmmmoLho
th Hvfi m H¢ MAOUV maHEIUNW
M\moG%oE

E0 . mwaoswsvmsm
.ompsmumsoo monom at Q .

 

26
isomers have been reported for disubstituted complexes of
Cr, Mo, and W hexacarbonyls with the similar ligands P(CH3)3
(14,33) and P(OCH3)3 (14,18). The disubstituted compounds
of CH3P(N(CH3)2)2,which were prepared here, were exclusively
EEEEEI as was shown by their infrared spectra which conformed
to D4h symmetry (Figure 3—B). The same Egggg configuration
was reported for the disubstituted complexes of P(N(CH3)2)3
(37), although in the case of Mo,Verkade, e£_§£. (18) have
recently reported the isolation of the gig isomers of both
P(N(CH3)2)3 and P(NCH3Ph)3. This P(N(CH3)2)3 complex of
Mo, however, readily isomerized to the EEEEE isomer in ben—
zene solution at room temperature. Both figgm and ESE“
isomers were observed for the trisubstituted compounds of
Cr and W, however, only the Egg— isomer was obtained for Mo.
These isomers were identified by their infrared (Figure 4—C
and 4-D) and nmr Spectra.

Substituted Group VIB hexacarbonyl complexes of ter—
tiary nitrogen and phosphorus ligands have been the subject
of many previous studies (10). An integral part of most of
these studies has been the interpretation of the carbonyl
stretching frequencies and the stretching force constants
in terms of the metal ligand bonding (12,13). Substitution
Of a CO group by another ligand of weaker W—acceptor capacity
causes the frequencies of the remaining CO groupsto decrease
inasmuch as the weaker H-accepting capacity of the ligands

(relative to CO) increases the available metal d-F electron

 

 

 

 

 

 

27

alto-5'1 II.I\
.NI .-

121.89.89.39... .1 m 13.1895.851.89.215... .. m

. ANAmmUOmemUVmAOUVOE mH fl .mmxmamfioo UmusuflquSmHC Usm IOGOE
How mnuommm UTHMHMQH was mo COHme mcflgoponum oo TSP mo mmamamxm

 

 

I.m wusmflm

 

 

 

28

GOO—

 

 

 

 

E0

OOON

~«-—. ~— -— .— ‘ —. _ ._—.__. _-—.—-

—-——-——o——..——-C——-—

 

 

 

.m Tasman

000"

000—
A

 

 

A
A
A
A
A
A
A
A
A

 

CONN

00w—

A

 

 

 

7:8

OOON

A
A
A
A
A
A
A
A
A
A

 

 

GONG

I

 

 

 

 

 

 

29

.mAnAmmoovmmmoAonovzwumE mfl m Cam

III-ll

. mHmhnmoovmmmoAonoVOZIOCM mfi m .mmxmamsoo owusufluwQSmHHchTE Csw_»omm

How mnuommm Cwumnmsfl 030 MO coammu osflnoumupm 00 may mo mTHmmem

I.v mHsmHm

 

 

 

 

 

.w musmfim
7:5 7:5

002. coon CONN 000—. OOON OONN

30

l . .
.
_
A
I

 

 

 

 

 

L
I:

 

 

 

A
_
_
A
A

_

 

 

31
density for back donation to the CO p—w* antibonding orbitals.
Trends in such experimentally observable shifts may then be
interpretable in terms of relative changes in the H-acceptor
properties of the metal—ligand bond.
The method of Cotton and Kraihanzel (11) was used to
calculate force constants from the frequencies given in

Table I. Secular equations for M(CO)5L, C molecular sym—

4V
metry; cis-M(CO)4L2, sz; trans-M(CO)4L2, D411; fac—M(Co)3L3,
C3v; and mer-M(CO)3L3, sz, were solved (see Table 2). In

these equations k1 is the stretching force constant for a
CO group trans to a ligand, k2 is for the CO force constant
gig to a ligand, and ki is a measure of the interaction
constant between CO groups. All stretching interactions
between CO groups should give rise to a term in the poten—
tial energy expression with a positive coefficient (ki)’
since as the CO bonds stretch the W bonds are weakened
slightly and the energies of the CO p-W* orbitals are
lowered. This energy lowering results in an increase in
metal to ligand H bonding to that CO group and a decreasing
availability of electrons for other CO p-W* orbitals with
the result that those CO bonds are strengthened (11). In
addition, consideration of the metal d-F orbitals shows

that two mutually EEEEE CO groups share two metal orbitals
while CO groups 213 to each other Share only one. Therefore,

interactions should be approximately twice as large between

 

 

.AIEO CH we > 0am
.msoum £3

 

32

s

asAdloH x .

00 MO mmma UTUSUGH 039 MO HMUWWMWUmV
. @H H

u
5 K

 

H
A.AN-NAA5 u A

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

am
A
e u AAA??? was. 9.5.8 .
H A was An~< >ao 8
MN» K! M5 UTHMHMGH AAV4 uonovzlmsmnh
AAAN+AMA5 I A can >80 sun
I UTHMHMCH AH AOUVZImHU
H
A.Mmlnxv5 H K owuwumsH .sm
N I m
#5 I A swam A £4
Aaxm+axvs A m mam o Nquoovsrmsmnh
n cmsmm «
a
A.M:Axvs n K nmamm mm
A an
A Amnaxvs u A o
CTHMHMGH Am
C H “KIAH¥+HMVAA HMSN A HAM/M ad
>
Axum AIAHMN+NAA5 Adv No useroAsumao
smfimm H<
a 98 Ell
A.xNIvas H A UTHMHMSH m
a smfimm Hm
#5 u A A
smsmm fl
8
o u _ AeA_xm+mAAs Axum can Anwm >30 uonoAE
.Msm Kuexs CTMMHMCH AHA
woos, wuumsfimm
mcoflumsvm Hdeowm hufl>fluo< msflsoumsum Hmasowaoz madowaoz

 

 

 

.AHHV mmCOE osflsoumuum 00 How mCOHumswm Hmflsomm .N manna

lIIIIIIIIIIIIIIIIlIIIllIIIII-----------

 

IIIIII:I____________________________i

33
two trans CO groups as between two cis CO groups. Finally,
both k1 and k2 should decrease as CO groups are replaced

by weaker H-acceptors, with k1 (trans to the ligand) de-

 

creasing more than k2, while ki should increase in value
(11). Therefore, if the force constants are calculated,
from the CO stretching frequencies, an estimate of V—ac—
cepting ability of various ligands, relative to ca can be

made.

If the values of the CO stretching frequencies and
force constants calculated here (Table 1) for the metal
complexes in which both CH2P(N(CH3)2)2 and CH2P(OCH3)2 are
ligands, are compared with those reported for the CO
groups in complexes in which either P(OCH3)3 (13), P(CH2)3
(13), or P(N(CH3)2)3 (56) are ligands, the relative order
of H-acceptor strength of these ligands can be determined.
For these ligands the W-acid strengths were found to decrease
in the order: P(0CH'3)3 : CH3P(OCH3)2 > P(CH3)3 > CH2P(N(CH3)2)2 >
P(N(CH3)2)3. The position ofVCH3P(N(CHV3)2)2 and P(N(CH2)2)3

in this order may be expected, since a certain degree of

p—E —> d-W, nitrogen to phosphorus H bonding is anticipated.
Support for this statement comes from structural data avail-
able for the compound F2PN(CH3)2 (57), in which a short

P-N bond length was observed along with a nearly planar
arrangement for the P, N, and C atoms. This trigonal planar

arrangement of nitrogen o—bonding orbitals suggests an sp2

 

 

 

 

34
hybridization about the nitrogen. This hybridization
must arise from the donation of the lone pair of electrons
from the nitrogen p-F orbital to the empty phOSphorus d-W
orbital. This partial filling of the phOSphorus d-W
orbitals causes the ligand to be a poorer acceptor of metal
d-F electron density than a ligand such as P(CH3)3 which
has no internal p-H —> d—F bonding. Further support for
the assumption that there is some nitrogen to phosphorus
F bonding in these compounds comes from some force constant
data of Nixon, gg_gl. (58). For di- and trisubstituted
molybdenum complexes of a variety of fluorophosphines they
determined a W—accepting capacity of ROPF2 > R2NPF2 >
RP(F)N’R2 > (R2N)2PF. In the case of P(OCH3)3 and
CH3P(OCH3)2, although there may be some degree of oxygen
to phosphorus p—H —> d—H bonding, the important factor
must be that the electronegativity of the oxygen creates
a slight lowering of the energy of the phosphorus d-w
orbital and an increase in its v-accepting capacity. In
addition, with only one inductive methyl group on the
oxygen, as Opposed to the two on the nitrogen, the oxygen
to phosphorus v—bonding is expected to be somewhat less

than the nitrogen to phosphorus H—bonding (59).

Proton N.M.R. For each ligand, the proton nmr spec-
trum consists of two doublets arising from the two types

of protons coupled to the phosphorus (see Table 3). The

 

 

 

 

 

 

 

 

 

35
Table 3. Proton nmr dataa.
b 2 c ’b 3 c
Compound CH3 JPH OCH3 JPH
CH3P(OCH3)2 1.10 8.5 3.50 11.0
Cr[(CH3P(OCH3)2](CO)5 1.70 4.0 3.60 11.5
Mo[(CH3P(OCH3)2](CO)5 1.70 3.0 3.55 12.0
W[(CH3P(OCH3)2](CO)5 1.80 4.0 3.55 12.0
CiS—Cr[CH3P(OCH3 )2]2(CO)4 1.70 3.5 3.60 11.5
cis~Mo[CH3P(OCH3)2]2(CO)4 1.65 2.5 3.55 12.0
cis-W[CH3P(0CH3)2]2(co)4 1.85 4.0 3.55 12.0
fac-Cr[CH3P(OCH3)2]3(CO)3 1.70 2.5 3.57 11.0
mer—Cr[CH3P(OCH3)2]3(CO)3 1.65d 2.5 3.50d 11.0
1.70e 2.5 3.578 11.0
fac—MO[CH3P(OCH3)2]3(CO)3 1.65 1.0 3.55 11.0
fac—W[CH3P(OCH3)2]3(CO)3 1.75 3.0 3.50 11.0
mer—W[CH3P(OCH3)2]3(CO)3 1.75d 3.0 3.48d 11.0
1.80e 3.5 3.53e 11.0
CH3P(O)(OCH3)2 1.42 17.0 3.65 11.0
"“_"73
N(CH3)2
CH3P(N(CH3)2)2 1.05 8.0 2.60 9.0
Cr[CH3P(N(CH3)2)2](CO)5 1.65 4.5 2.60 10.0
MO[CH3P(N(CH3)2)2](CO)5 1.65 3.8 2.60 10.5
W[CH3P(N(CH3)2)2](CO)5 1.80 4.5 2.60 11.0
EEEEE‘ 1.60 4.0 2.60 10.0
Cr[CH3P(N(CH3)2)2]2(CO)4
trans—
————- 2.65 3.0 2.60 10.5
MO[CH3P(N(CH3)2)2]2(CO)4
EEEEEf
W[CH3P(N(CH3)2)2]2(CO)4 1.80 4.0 2.60 10.5

 

achloroform solutions.

bChemical shifts are in ppm (i 0.05) and are referenced to
tetramethylsilane. u

CCoupling constants are in Hz (i 0.5) and are given as the
absolute value. See text for discussion of signs.

Apparent "doublet" for ligand cis to the two trans ligands.
e
Apparent "triplet" for the two ligands trans to each other.

 

 

36

more shielded methyl group protons resonate at a higher
field than either the OCH3 or the N(CH3)2 protons, and the
N(CH3)2 protons resonate at a higher field than those of
the OCH3 group. When the ligands are complexed to the
metals, there is a downfield shift for the CH3 resonances
due to partial oxidation (as a result of donation of the
lone pair of electrons) of the phosphorus. However, the
chemical shifts and coupling constants of both the N(CH3)2
and the OCH3 groups are relatively insensitive to complexa-
tion because of their greater distance from the phosphorus
atom. The absolute value of 2JPH (coupling to the methyl
protons) is decreased upon complexation to less than half
of its free ligand value, while the absolute value of 3JPH

(coupling to either the OCH3 or N(CH3)2 protons) remains
‘ unchanged or is increased Slightly.

The nmr spectra of the monosubstituted compounds con-
sist of two doublets, while the Spectra of the disubstituted
and the Egg—trisubstituted compounds consist of two reso-
nances, each of which appears as a "triplet", the center
peak of which shows a wide variation in intensity from
compound to compound. This apparent "triplet" results
from both phosphorus—hydrogen coupling and phosphorus-
phosphorus coupling, and is an example of an XnAA'XA type
spectrum which has been observed in a variety of other
PhOSphorus complexes (18,21,22,37). The subscript n

refers to the number of protons (X) that are coupled to a

 

 

 

37
phosphorus (A). Thus n equals 3, 6, or 12 for the CH3,
OCH3, or N(CH3)2 resonances, respectively, in the ligands
CH3P(OCH3)2 and CH3P(N(CH3)2)2. (Figure 5.) The center
line of the observed "triplet" arises from the near coin—
cidence of a large number of X lines and the spectrum may
appear as a 1:2:1 triplet in the limiting case (16). The
intensity of the central peak is then dependent upon this
coincidence of many other Spectral lines and the value of
2JPP is a maximum when the central peak is most intense
(16,17). By using the qualitative approach outlined here,
results similar to those obtained in previous studies (18,
22,37) can be achieved, namely, that 2JPP for 35223 com—
pounds is larger than that for analogous gig compounds,ex-
cept perhaps for Cr, and that coupling for EEEEE compounds
‘ decreases Mo > W >> Cr, and for gig compounds Cr > Mo > w.
Unfortunately, the two ligands used in this study do not
form complexes of the same geometry, so comparisons of the
effect of the ligand on 2JPP cannot be fully ascertained.
One unique feature of the nmr Spectra of the complexes pre-
pared in this work is that two resonances are observed for
each complex (the CH3 resonance and the OCH3 or N(CH3)2
resonance), thus allowing for the possibility of two in—
dependent determinations of 2JPP.
For the compound mer-((CH3)(CH30)2P)3W(CO)3 in addi-
tion to the two "triplets", both the CH3 and the OCH3 protons

exhibited a "doublet" at approximately 0.05 ppm higher field

 

 

 

 

 

 

 

 

Figure 5.—

38

Examples of 1H nmr spectra of metal complexes
which illustrate spectra with and without the
presence of phosphorus—phosphorus coupling.

In both spectra L = CH3P(N(CH3)2)2. The more
intense downfield peaks are for the (CH3)2N-
group while the less intense upfield peaks are
from the CH3— resonances. The trans-W(CO)4L2
is a good example of an XnAA'Xfi type spectrum.

 

 

 

39

 

. m 0H 5m. ..n .m
To: IA
«430927.39:
1.60.?

 

 

 

 

 

 

 

 

 

 

 

 

 

.m.<<.Z IA.

 

Figure 6.

 

— Shapes of the proton nmr resonances of some tr1-

4O

substituted complexes of CH3P(OCH3)2 (L)- .A 15
mer-W(CO) L3; B is fgg~w(co)3L3; and. C 15.
me;—CI(CO)3L3. The methyl resonance is at higher
fIEld in all cases. Those resonances 1n the EEE‘
complexes labeled with the letter t are as- d
signed to the two ligands trans to each other. an
give rise to a near "triplet". Those resonances
labeled d refer to the "doublet" which results
from the ligand cis to the other two ligandS (see
text). In C, aSSignment of the central reset
nance in the "triplets" is not obvious, and 15,
therefore, not labeled.

 

 

—i—

42
(Figure 1A)o These "doublets" can be attributed to the
ligand which is uniquely Egggg to a C0 group and gig to
the other two ligands. These resonances represent one of
the limiting cases of the XnAA'Xé type spectrum where
2JPP is small for phosphorus atoms gig to one another
while the slightly lower—field "triplets" represent the X
part of an XnAA'XA type spectrum wherein 2Jpp is appreci-
able for two ligands trans to each other. The mgg- Cr
compound exhibited a Spectrum consisting of two sets of
overlapping "doublets" (Figure 1C). In the OCH3 proton
resonance, where the phosphorus-hydrogen coupling is quite
large, some indication of a very weak peak could be observed
near the middle of this “doublet" grouping. However, it
was impossible to assign this resonance unambiguously to
either the Egggg ligands or the gig ligands. Again the
chemical shift difference between the two resonances is
approximately 0.05 ppm. Similar spectra have been observed
by Shaw, et al. (22), for the trisubstituted Cr compound of
PhP(OCH3)2 and the analogous Mo compound of (CH3)2PPh.

The slight change observed in 3J for the P-OCH3

PH
system upon complexation is typical (22). The relatively
large change observed in 2JPH for the CH3 group is of
interest because it aids in the assignment of the Sign of
the P-c—H coupling. Manatt, gg_gi. (60) have suggested

that the geminal coupling, P—C-H, (ZJ in a variety of

PH).

trivalent phosphorus compounds is positive and that 2J

 

PH

 

 

 

 

ll

 

 

 

 

43
becomes more negative as the 5 character of the phos-
phorus bonding orbitals to carbon increases. Thus, 2JPH
in (CH3)3P is +2.66 Hz while 2J in (CH3)4P+ and

PH
(CH3)3P=O are -14.4 Hz and —13.4 Hz, respectively (60).
In the latter two compounds, the hybridization around
phosphorus is nearly pure sp3 and consequently the phos—
phorus—carbon bonds should have more 5 character than do
phosphorus—carbon bonds in compounds in which the phos-
phorus has not been partially oxidized from loss of its
lone pair. In CH3P(OCH3)2 and CH3P(N(CH3)2)2 the values

of 2J 8.5 Hz and 8.0 Hz (Table 3), resPectively,are

PH,
probably positive (Manatt, et al. (60)) while the value of

17.0 Hz for 2JPH in CH3P(O)(0CH3)2 (Table 3) is taken as
negative. Therefore, as the phosphorus is oxidized, the s
character of the P—C bond increases and the phOSphorus-
hydrogen coupling constant becomes more negative. In the
metal carbonyl complexes the phosphorus is partially oxi—
dized because of donation of the lone pair. The result of
this donation should be an increase in the s character

of the P-C bond, and the valueCfirZJ for the complexed

PH
ligand should be more negative than that for the free

ligand. Since the absolute values of 2JPH in the complexes

are less than those in the free ligands, (Table 3) this
coupling in the ligand is very likely positive if the 3

character effect is dominant. Although the 2J couplings

PH

could be either positive or negative in the complexes,

 

 

 

44
some evidence can be cited for suggesting that they are
negative. As the carbonyl groups on a metal are succes—
sively replaced by one, then two, then three phosphorus
ligands (which are more basic than CO groups), there is an
increase in the effective negative charge on the metal.
Therefore, each phosphorus atom is oxidized to a lesser »
extent in a trisubstituted complex than in a disubstituted
complex, and to a lesser extent in a disubstituted complex
than in a monosubstituted one. Consequently, there should
be less 8 character in the P—C bonds and 2JPH should
become more positive with increased substitution. Refer-
ence to Table 3 shows that the regular decrease in 2JPH
from [3| to [1) Hz for the molybdenum complexes of
CH3P(OCH3)2 parallels the degree of substitution from the
mono— to the trisubstituted complex. These values of 2JPH,
therefore, are probably increasing from -3 to —2.5 to —1
for the complexes Mo(CO)5L, Mo(CO)4L2, and Mo(CO)3 re—
spectively (L = CH3P(OCH3)2). Similar trends can be noted

(Table 3) for the Cr and W compounds. It is,of course

 

consistent with this rationale that 2JPH for the trisubsti—
tuted molybdenum complex could have passed through zero,
and is actually +1.

If it can be assumed that 2JPH in these complexes is
a reflection of the extent of phosphorus to metal o-dona-
tion, then the methyl coupling constants give an indication

of the strength of the o—bond. As the metal atom is varied

 

 

 

1.

 

 

 

45

for a given substitution and ligand, the 2JPH value con—
sistently becomes more positive from tungsten to chromium
to molybdenum although the change in 2JPH is relatively
small. since this change in 2JPH is small, it could very
well be that the o—interaction is either very nearly the
same or perhaps becoming slightly weaker along the series
W ; Cr § Mo.

It should be pointed out that in an XnAA'X'n spectrum,
the separation of the two most intense peaks (the outer
peaks of the apparent "triplet") is given by ‘JAX + JAX'

(16). If J is considered to be very small or almost

AXu
negl1g1ble compared to JAx ,then JAX = IJAX + JAX

2JPH may be taken directly from the spectrum. The rela—

tively large number of bonds (four) through which AX'

a[ and

coupling takes place makes this assumption plausible for
these complexes. The values reported in Table 3 for both
2JPH and 3JPH have been determined directly from the

spectra in this manner.

Phosphorus-31 Nmr.— Phosphorus—31 nmr spectra were
obtained for three of the metal complexes, and for each of
the two ligands (see Figure 7). The spectrum of the ligand
dimethyl methylphosphonite which consisted of a septet of
quartets at 0 = —182.8 ppm, relative to the external
standard (85% ggggg—phosphoric acid), arises from the split—
ting of the phosphorus nucleus by both the three methyl and

the six methoxy protons. The ligand bis(dimethylamino)—

methylphosphine exhibited an unresolvable broad peak at

 

46

F' o_ 31 ‘
1gure 7 P nmr Spectra of CH3P(OCH3)2, CH3P(N(CH3)2)2'

and three of the metal complexes of these ligands.

 

 

 

 

 

 

47

3
1p N.M.R.

432.899!“

L, : cuaumcfla’a’z

. . ,4 "m m
(WWW/fl meyw pp MWWW1MA

2. A “(CO)3"3
.4968P9m
MW
MAM
MoDO’s" VK/
425.799" \\/\'W

8.7 “MILWWW

 

 

“WWW/MM

Figure 7 .

 

 

EII:________________________________—__—’7

48

6 = —86.2 ppm. The Spectra of the complexes were quite
weak and revealed a simple broad absorption from which
only a chemical shift (Table 4) could be ascertained.
These data are consistent with similar data reported for
other complexes (Table 4) in which the phosphorus atom is
deshielded upon complexation with a metal. Meriwether and
Leto (41) have suggested that the factors which affect 31P
chemical shifts upon complexation to a metal include: ex-
tent of 0 bond formation, extent of d—v —¢ p-V back dona—
tion, rehybridization of phosphorus orbitals, electronega—
tivities of substituents, and steric effects. The use of
31F chemical shift data to describe bond types is very
tenuous because of both the number of factors involved and
the interdependence of these factors. Indeed, the only

‘ thing that can be said about these data is that the phos—

phorus is deshielded because of the donation of its lone

pair to the metal.

It is interesting that although the chemical shift of
CH3P(N(CH3)2)2 falls between the values reported for

P(CH3)3 (50) and P(N(CH3)2)3 (50)(Table 4), and the shift

 

of trans—Cr(CO)4[CH3P(N(CH3)2)2]2 falls between the Shifts

 

reported for trans—Cr(CO)4[P(CH3)3]2 (19) and trans—

 

Cr(CO)4[P(N(CH3)2)3]2 (19), the chemical shift of
CH3P(OCH3)2 does not occur, as expected, between the shifts

of P(CH3)3 (50) and P(0CH3)3 (50). It is in fact signifi—

 

cantly more negative than the chemical shift observed for

 

 

 

 

49

Table 4. 31p nmr Spectral data.

 

 

 

Compound Chemical Shift Reference
(ppm)
13(cr13)3 + 62.0 (i1.0) (50)
CH3P(OCH3)2 -182.8 (:1.0)
1>(oc11t3)3 —141.0 (:0.5 (49)
CH'3P(N(c33)2)2 - 86.2 (:1.o>
p(1q(c:1+1'3).,)3 -121.9 ($0.5) (5o)
EE§7M0(C0)3[CH3P(N(CH3>213 '196-5 (i1.0)
Mo<c0)5ICH3P(N(CH3)2)21 -125-7 (il-O)
EEEEE-Cr(co)4[CH3P(N(CH3)2)2]2 —164.3 (i1.0)
MO(CO)5[P(OCH3)3] -161.2 ($0.5) (19)
min—s-Cr(CO)4[P(N(CH3)2)3]2 -178-2 ($0.5) (19)
)

trans-Cr(CO)4[P(CH3)3]2 — 21.0 (10.5 (19)

 

 

 

 

 

II

 

 

 

50
P(OCH$)3. In addition, the 31F chemical shift of the gig—
trisubstituted molybdenum complex of CH3P(OCH3)2 is shifted
downfield to —196.5 ppm. Although the magnitude of this
change is as expected (19), the value of —196.5 is still
more negative than the chemical shifts for complexes of
P(OCH§)3 (19). Van wazer §t_al. (50) have reported an ex-
tensive study of 31P nmr for many phosphorus compounds and
discussed the factors influencing the chemical shifts.
Among these factors are bond angles, bond hybridization,
electronegativity of substituents, and the amount of v—
character in the phOSphorus bonds. Along the series of com—
pounds (CH3)nP(OCH3)3_n (n = O,1,2,3) there should be a
regular change in all of the above effects and this change
should be reflected in the 31P chemical shifts of the com-
pounds in this series. Since these changes in chemical
shifts should be regular, it is not clear at this time why
CH3P(OCH3)2 has a larger negative value for its chemical

shift than does P(OCH3)3.

Stereochemistry.— The observance of both cis and trans
and £39— and mgr— isomers for the disubstituted and tri—
substituted complexes of CH3P(OCH3)2, respectively, with
the apparent predominance of the gig and fag- isomers, is
consistent with the stereochemistry reported previously for
complexes with similar ligands (13,14). I expected that no
stable trisubstituted compounds would be formed with

CH3P(N(CH3)2)2 since similar observations had been reported

 

 

in

 

 

 

51
for P(N(CH3)2)5 (37). If these nitrogen—containing phos-
phorus ligands are actually strong o—donors (37) the in-
creased charge on the metal would be too large to accommo-
date three such ligands. The suggestion (37) that elec-
tronic rather than steric effects are the primary cause for
the lack of formation of trisubstituted complexes with
CH3P(N(CH3)2)2 and P(N(CH3)2)3 (37) gains support from the
fact that the bulky ligand P(C6H5)3 (14) forms stable_f_e_1_g—
complexes. Moreover, the ligands, FP(N(CH3)2)2,
CH3P(F)(N(CH3)2)I and C6H5P(F)(N(C2H5)2) all form EEE'
complexes with molybdenum (61). The stability of such
complexes may result from an increase in the W—acceptor
strength for these fluorine containing ligands over that of
CH3P(N(CH3)2)2 and P(N(CH3)2)3 because of the electronega—
tive fluorine atom. This increase in w—acidity and cor—
responding decrease in o-basicity reduces the increased
negative charge on the metal that must arise when three co
groups are replaced by three phosphorus ligands.

It is interesting that CH3P(N(CH3)2)2 would not form
an isolable monosubstituted complex with tungsten (see
Experimental Section) even though it would form such com—
plexes of chromium and molybdenum. King (37) observed the
same results when he reacted P(N(CH3)2)3 with tungsten.
This latter observation was not expected since monosubsti—
tuted tungsten complexes with various amine ligands, which

should also be strong o-donors, are well known (40,62).

 

 

 

52
In order to explain this observation, an important role
might be invoked for the metal to phosphorus w-bond in
stabilizing these compounds. Whether or not the o—bond
is indeed the stronger bond, it algae is probably not suf—
ficient in this system to stabilize a metal-phosphorus
bond. Thus the v—contribution to the synergic bonding-
mechanism must be sufficiently large before the phosphorus
ligand will form a M—P bond. This assumption is supported
by the observed reactions of tungsten. For the Group VI
hexacarbonyls, the order of metal-CO w—bonding is W'> Cr
> Mo (11,63). If a single CO group in W(CO)6 is replaced
by either CH3P(N(CH3)2)2 or P(N(CH3)2)3 a poor r—accepting
ligand is placed trans to and in direct competition with
the strong W—acid, CO, and an unfavorable situation is
created. Stable monosubstituted amine complexes of tungsten
do form, however, because the increased basicity of the
ligands over that of CH3P(N(CH3)2)2 or P(N(CH3)2)3 and the
resultant stronger o-interaction must be sufficient for
bond formation. Consistent with this argument are the ob-
servations that the most stable disubstituted metal complexes
of the ligands CH3P(N(CH3)2)2 and P(N(CH3)2)2 (18,37) are

the trans isomers, while amine substituted compounds are of

 

the cis geometry (12). When the ligand on the metal is one
of these phosphorus containing ligands, the trans isomer
allows for the increase in metal-phosphorus W-bonding neces-

sary to stabilize the compounds, especially for the tungsten

 

 

 

53
complex, since the ligand is then removed from direct com—
petition with a CO for the metal v—electron density. On
the other hand, the gig isomer allows for maximum metal—CO

v—bonding in the case of the poor v-accepting amine ligands.

 

M

 

 

 

PART II
THE SYNTHESIS AND SPECTROSCOPIC PROPERTIES OF SOME

CYANOPHOSPHINES AND THEIR CHROMIUM AND MOLYBDENUM
CARBONYL COMPLEXES

54

 

 

Piliflui l

 

 

INTRODUCT I ON

Although the metal coordination chemistry of tertiary
phosphine compounds has been investigated extensively in
recent years (1), there has been relatively little attention
given to cyanophosphine compounds of the type [RXP(CN)3_X]
(x = 0,1, or 2). The preparation and characterization of
of several cyanophosphine compounds have been reported (51,
65,66,68,69), however, these preparations usually involved
rather extensive reaction times and the characterizations
involved little spectral data. In 1963, Noth and Vetter
(67) reported the preparation of a disubstituted nickel car—
bonyl complex of [(CH3)2N]2PCN in which coordination pre—
sumably occurred through the phosphorus atom. More recently,
Kirk and Smith reported that P(CN)3 forms unstable adducts
with AlH3, AlCl3, and AlBr3 (69) and that in solution it
forms an adduct with pyridine (70) which was identified
by 1H nmr Spectroscopy. In this pyridine adduct P(CN)3 was
probably acting as the acceptor with coordination occurring
through the phosphorus atom. In this same work some associ-
ation of the cyanide groups in both P(CN)3 and PhP(CN)2
with N,N-dimethylformamide was prOposed on the basis of 1H

nmr data (70).

55

 

 

 

 

56

Because of the very little amount of work reported on
the coordination chemistry of cyanophosphine compounds, I
decided to prepare several of these phosphines and attempt
to form transition metal complexes with them. In this work,
a rapid and easy method of preparation is presented for the
routine synthesis of compounds of the type RxP(CN)3-x
(where x = 0, 1, or 2 and R = Ph-, CH3O-, C2H5 —, (CH3)2N—,
or CH§-). The infrared, proton nmr, and mass spectral data
of these compounds are also given. Several of these phos-
phines were then used in reactions with chromium and molyb-
denum norbornadiene tetracarbonyls. Emphasis has been
placed on determining the effect of the cyanide group on
the coordinating properties of the phosphorus atom to which
it is bound, and on determining if the nitrogen atom, as
well as the phosphorus atom, is a potential o-donor site.
The following cyanophosphine ligands were used: PhZPCN,
PhP(CN)2, (CH3)2PCN, (C2H5O)2PCN, and [(CH3)2N]2PCN. Metal
carbonyls were selected as reactants for these ligands be-
cause carbonyl substitution is known to occur for both
trivalent phosphorus compounds (1,10) and for nitriles (10,
55,80).

During the course of this work Nixon and Swain (83)
proposed a dimeric structure for the complexes
[M(Co)4P(CF5)ZCN]2 (M = Cr or Mo) with (CF5)2PCN acting as
a bridging ligand. I therefore wished to determine, by

means of chemical and spectroscopic evidence, the structure

 

 

 

 

 

 

 

57

of dimers of this type and if this process could be observed

for other cyanophosphine ligands.

 

 

 

 

 

 

 

 

EXPERIMENTAL

The infrared Spectra were run on a Perkin-Elmer model
225 grating spectrophotometer. CHC13 was used as a solvent
for all metal complexes and for P(CN)3 and CH3P(CN)2.
Hexane was used as an infrared solvent for all of the other
cyanophosphines. Proton nmr Spectra were taken on a Varian
Associates Model A-56/60-D instrument with tetramethyl-
silane as an internal standard. For all of the nmr spectra
CHCl3 was used as a solvent, except for the spectra of
complexes of PhZPCN where CHZClz was used. The phOSphorus-
31 nmr spectra were obtained on a Varian Associates Model

DA-60 spectrometer at 24.29 MHz and are referenced to 85%

 

ggghg-phosphoric acid as an external standard. Mass Spectra
for the cyanophosphine compounds were recorded on an
Hatachi-Perkin Elmer model RMU-6 instrument using the

liquid inlet system with an ion source temperature of 2000
and an ionizing voltage of 70 eV for all samples except
P(CN)3, where the solid inlet was used at a temperature of
50° and an ionizing voltage of 70 eV. ,Mass Spectra of the
metal complexes of the cyanophosphine ligands were per—
formed by the Chemical Physics Research Laboratory of the
Dow Chemical Co., who used the direct probe method with

temperatures ranging from 110—2150. Elemental analyses and

58

 

 

 

 

EII______________________________________i

59
osmometric molecular weight determinations were performed

by Galbraith Laboratories, Inc., Knoxville, Tennessee.

Materials.- The compounds PhPClz, thPCl, (C2H50)5P,
were purchased from Aldrich Chemical Company, Milwaukee,
Wisconsin, and CH3PC12 was obtained as a gift from the
Department of the Army, Edgewood Arsenal, Edgewood, Mary—
land. The compounds, c2H501>c12 (71), (C2H50)2PC1 (71),
(CH3)2PC1 (72), (CH3)2NPC12(66), and [(CH3)2N]2PC1 (66)
were prepared by previously described methods. Molybdenum
norbornadiene tetracarbonyl (52) and chromium norbornadiene
tetracarbonyl (52) which were used for the preparation of
the metal complexes were also prepared by previously de-
scribed methods.

The compounds, (CH30)2PC1 and CH30PC12 were prepared
when appropriate molar ratios of (CH30)3P and PCl3 were
mixed in a manner described as follows. When a 2:1 molar
ratio of (CH30)3P to PC13, for example, was stirred in a
N2 atmosphere at room temperature for several min, an exo—
thermic reaction ensued forming (CH3O)2PCl. This product
was identified in the 1H nmr spectrum by a doublet resonance
which was, as expected, downfield (due to decreased shield—
ing) from the similar doublet for (CH30)3P; stirring for 24
hr at room temperature insured complete reaction. The extent
of reaction was easily monitored by following the (CH3O)3P

doublet in the 1H nmr spectrum. When the opposite molar

 

ratio of (CflgohP and Pclé was mixed at room temperature,

 

 

 

I[IIT__________________________________—I
60

(CH30)2PC1 again formed, however, if the reaction mixture
was warmed slightly («700) for 3 hr, CH30PC12 formed. This
reaction was also followed by 1H nmr spectroscopy; CH3OPC12
exhibited a doublet resonance down field from the resonance
of (CH3O)2PC1. Although some decomposition was evidenced
by formation of an unidentified orange solid and CH3Cl
(singlet in the 1H nmr at 2.95 ppm) the reactions were es-
sentially quantitative. Caution: It was not found possible
to purify these chlorophosphitgs by distillation (either
at reduced or at atmgspheric pressure) because decomposition
was accelerated by the application of heat, and invariably
§g_gxplosion resulted. These compounds, whose purity was
checked by 1H nmr, were used without purification and with
good results. It was later found that these two compounds
could be made easily if the appropriate molar ratio of
methanol was added dropwise to PC13 and allowed to stir for
approximately 6 hr. The nmr spectra of the products from
these preparations indicated that less CH3Cl formed with
this method than with the above method. No attempt was made

to distill the products of these preparations.

Preparation of Cyanophosphine Compounds.— The com—
pounds, P(CN)3 (65), (CH3)2NP(CN)2 (66), [(CH3)2N]2PCN (66),
(C2H50)2PCN (68), C2H5OP(CN)2 (68), and PhP(CN)2 (70) were
prepared previously when the appropriate phosphorus halide

was caused to reflux with AgCN for 12 to 24 hr in solvents

 

 

 

—i—

61
such as CHCl3, CC14, and (C2H5)20. This method of prepara—
tion was modified slightly for preparation of the above
compounds and the compounds of this work. If acetonitrile
was used as the reaction solvent, the compounds were
quickly formed when the reactants were stirred for about 2
hr at rrom temperature under a nitrogen atomosphere. The
compound, CH3P(CN)2 was prepared previously (51) when aceto-
nitrile was used as the solvent, however, reflux temperatures
were employed for five hr.

Two methods were used for isolation of the products
from these reaction mixtures: A. After about 2 hr of stir-
ring, the reaction mixture was filtered under nitrogen to
remove the AgCl (and any unreacted AgCN). The acetonitrile
was then removed from this filtrate by water aspiration,
and the remaining liquid was vacuum distilled (vacuum sub-
limed for solids) (see Table 5) to give the colorless liquid
or white solid product. B. The entire reaction mixture
was vacuum distilled, after about 2 hr of stirring. The
acetonitrile was collected first, followed by the cyano-
phosphine product. The silver salts were left in the reac—
tion vessel. This second method of isolation was particu—
larly effective for those phosphorus compounds which formed
insoluble complexes with the Silver salts in the reaction
mixture and which would have been lost in filtration, but
it could also be used as a general method for the isolation

of the lower boiling liquid products.

 

 

 

 

62

Table 5. Experimental data for cyanophosphines.

 

 

Compound

Method of

 

Isolation (mm)
PhZPCN A 1100(0.1)
PhP(CN)2 A 68—70°(0.1)
(CH50)2PCN B 580(10)
CH50P(CN)2 B 650(6)
(C2H5O)2PCN A 510(5)
02H5OP(CN)2 B 510(5)
[(CH3)2N]2PCN B 480(1)
(CH3)2NP(CN)2 B (SS-690(8)
(CH3)2PCN B 40-450(10)
CH3P(CN)2 A 50-60°(0.1)a
P(CN)3 A 50—600

 

aSublimed.

 

 

 

63

It should be noted that (CH50)2PCN, which was prepared
by method B, was also prepared by method A. 'With method A,
after approximately half of the material had been distilled
the distillation pot exploded rather violently, perhaps due to

some unreacted (CH30)2PC1.’Therefore, method B is recommended.

Reactions of P(CN)3 with Various Metals.- The com—

 

pound P(CN)3 was caused to react with a number of transition
metals under many different conditions, and with a variety of
solvents. In every case, no identifiable complex was iso—
lated. Reactions of this phosphine with nickel compounds were
studied extensively. These compounds included Ni(Clog)2-6H20,
Ni(ClO4)2-6CH3CN, Ni(BF4)2°6DMSO, Ni(CN)2, NiBrz, and N112.
(When used in these reactions the Ni(II) hexahydrates were
first dihydrated with 2,2-dimethoxypr0pane.) The compound
Ni(CO)4 was used in attempts to prepare nickel(0) complexes

of P(CN)3. All of these reactions with nickel were run on a
scale of approximately a mmole of metal, and a ligand to
metal molar ratio of anywhere from 1:1 to 10:1. A wide range
of temperatures was used as well as a wide variety of $01-
vents. The solvents which were employed included acetone,
ether, halocarbons, alcohols, acetonitrile, and benzene as
well as others. A brown amorphous solid, which was not ident—
ifiable by infrared spectra or elemental analysis, invariably
was obtained from reactions with Ni(CO)4, while a yellow

to green powder was isolated from an initially yellow to

orange solution for the reactions of Ni(II) salts with the

 

 

 

64
ligand. The various ligand:metal ratios, solvents, and
reaction temperatures appeared to have little effect on
the outcome of these reactions.

Reactions of P(CN)3 were also attempted with Co(II) and
Cu(I)salts and with Fe(0), Cr(0), Mo(0), and W(0) carbonyls.
In almost all cases (except chromium, molybdenum, and tung—
sten which required heat) the addition of P(CN)3 to a solu-
tion of the metal salt or carbonyl was followed by an im—
mediate reaction, as evidenced by a rapid color change.

In all cases of reactions attempted with this ligand no
product other than unidentifiable non-crystalline powders
could be isolated. Even in the cases where elemental
analyses were obtained for the products, no meaningful
interpretation could be gathered from the results, and de-
spite the many reactions no definite metal complex of

P(CN)3 was ever isolated.

Preparation of Metal Complexes of Cyanophosphine
Ligands Other Than P(CN)3

Bridged Complexes.- All of the bridged complexes were

 

prepared when equimolar mixtures of either chromium or
molybdenum norbornadiene tetracarbonyl and one of the cyano—
phosphine ligands were stirred in nfhexane for several

minutes (see Table 6). The nfhexane was removed by water

 

aSpirathiand the remaining solid was recrystallized from

CH2C12 to give yellow crystalline solids of the form

 

 

 

 

 

65

Table 6. Experimental and analytical data for metal complexes.

 

 

Rxn 7C %H

0 OP
compound Time Calc. Found Calc. Found Calc. Found

[Mo(CO)4L]2

L:
pthcn 5 min 48.69 48.68 2.39 2.50 7.40 7.26
PhP(CN)2a 5 min 39.13 36.62 1.36 1.60 8.43 7.81
(CH3)2PCN 5 min 28.47 28.29 2.03 2.07 10.51 10.39
(C2H5O)2PCN 5 min 30.42 30.31 2.82 2.88 8.73 8.59
[(CH3)2N]2PCN 5 min 30.59 30.29 3.40 3.45 8.78 8.65
[Cr(CO)4L]2

L:
thpcu 5 min 54.50 54.64 2.67 2.75 8.27 8.46
(CH3)2PCN 5 min 33.47 33.27 2.39 2.40 12.35 12.50

(C2H50)2PCNa 12 hr -— —_ __ __
[(CH3)2N]2PCN 12 hr 34.95 34.70 3.88 3.84 10.03 10.22

Mo(CO)4L2

L:
Ph2PCN 2 hr 57.14 57.12 3.17 3.23 9.84 9.65
(CH3)2PCN 1 hr 31.41 31.18 3.14 3.22 16.23 15.92
(C2H50)2PCNa 2 hr —- —- -— -- -— ——
[(CH3)2N]2PCNa 1 hr 33.73 31.18 4.82 4.84 12.45 11.31
Cr(CO)4L2

L:
thPCN 4 hr 61.43 59.81 3.41 3.89 10.58 10.28
(CH3)2PCNa 2 hr —— -— —— -— —— ~-
[(CH3)2N]2PCNa 1 hr 36.99 35.97 5.28 5.69 13.64 10.28
Mo(CO)4LL'

L =

[(cng)2N]2pcm 1 hr 30.19 30.30 4.40 4.52 13.00 12.88
L’ = (CH3O)3P

 

 

 

 

 

 

 

66
Table 6. (Continued)

Rxn % c % H % P
compound Time Calc. Found Calc. Found Calc. Found
Cr(CO)4LL'

L:

[(CH.3)2N]2PCN 1hr 33.26 32.72 4.85 4.77 14.32 14.25
L' = (CH3O)3P

 

aReadily decomposed and, therefore, not isolated in purity
suitable for good analysis, however, infrared and nmr Spectra
gave unequivocal evidence as to the identity of these com-
plexes.

 

 

 

 

 

67
[M(CO)4L]2. A molecular weight measurement of

[Mo(CO)4Ph2PCN]2 produced a value of 910; calculated, 838.

Disubstituted Complexes.— The disubstituted complexes
(both M(CO)4L2 and M(CO)4LL') were prepared when a 1:1 molar
mixture of the appropriate ligand and the bridged metal
complex were caused to reflux in nfhexane for one to four
hr (see Table 6). The nfhexane was then removed by water
aspiratbmn and the remaining solid was recrystallized. The
M(CO)4L2 complexes were recrystallized from nfhexane to
which a minimum of dichloromethane was added to dissolve
the complexes, while the M(CO)4LL' complexes were recrystal-

lized from nfpentane.

 

 

 

RESUIES AND DISCUSSION

Cyanophosphine Comppunds.— In contrast to the pre-
viously reported preparative methods for cyanophosphines,
the preparations in acetonitrile proceeded rapidly at room
temperature. All of the compounds were hygroscopic and
usually became yellow to red in color after brief exposure
to air. This apparent decomposition, however, did not seem
to affect further reactions of these compounds in any ap—
preciable way, and several of these compounds were used as

ligands to form chromium and molybdenum carbonyl complexes.

Mass_Spectrapof‘Cyanophpsphines.- Mass spectra were ob-
tained to confirm the identity of the cyanOphOSphines and,
perhaps, to gain some information about the strengths of the
phosphorus-cyanide bonds. (My interest in the phosphorus-
cyanide bond strengths in these compounds arose from the
somewhat peculiar behavior of P(CN)3. Apparently P(CN)§ has
fairly strong intermolecular nitrogen-phosphorus interactions,
both as a solid (76) and in solution (64). In addition,
some of the work of Kirk and Smith (69), as well as some
of the work I did with p(CN)3, seems to indicate that at
least one cyanide is easily removed from the phosphorus
atom under reaction conditions.) Parent ions were observed

for all of the compounds.

68

 

 

69

These data are presented in Table 7 along with other major
identifying peaks. For P(CN)3, I found that the peak at
m/e 57 which can be represented as PCN+, was the most in—
tense peak, followed by CN+ (86% relative intensity and then
P(CN)a+ (25%). The intensity of the PCN+ ion, perhaps demon—
strates the ease with which two of the cyanides are cleaved
from the phosphorus, and further, indicates that under these
conditions (temperature of 50°, ionizing voltage of 70 eV),
P(CN)2+ (8.3% intensity) is easily cleaved to give PCN+._
This trend was, in general, paralleled in the other compounds,
especially in the dicyano Species where mass spectral peaks
for PCN+ and P(R)CN+ were always found to be stronger than
the peaks for P(CN)2+. In the compound PhP(CN)2, C6H4PCN+
was the most intense peak and was formed when a neutral HCN
molecule split off of the parent compound, as evidenced by the
metastable peak observed at m/e 110.6 (calculated at m/e 110.56).

m* 110.6 C6H5P(CN)2+—9 C6H4PCN+ + HCN

The phenyl derivatives exhibited fairly strong parent
peaks. The ion peak at m/e 211 for PhZPCN+ showed a 96%
relative intensity, and the ion peak m/e for PhP(CN)2+ ex—
hibited a relative intensity of 50%.

The phosphites showed strong peaks for those Species
in which the phosphorus-oxygen bond remained intact. For
example, in the Spectrum of (CH30)2PCN, the most intense
peak was at m/e 93 corresponding to P(OCH3)2+ with

P(OCH§)(OH)+ being next strongest at 56% relative intensity.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

zova

 

mm m.w ooH

w.HH H.oo m.mm H.8e m.Ho ammo
o.su m.oo w.o 6.6 ooe zooaflomuv
o.m 5.8m m.am ooH m.o «AzovdzNAamov
m.m h.mm m.oH o.ee o.mm zomofizvomov:
w «.6m m.m m.o¢ w.m m.m oAzovaomao
o.w o.om m.oH H.we m.me zodaaoomaov
m.m a.mm w.mm ooH o.Hm «AsovoOSSU
m.a o.o o.o ooH a.mH zooefiommov
e.m o.om s.o m.ma m.mfi «Azovosa
H.8 m.v ooe m.m mm zomasa
tcdomaoo

lioAzovd «Azovd mmwmvam 2mm. zqmw mm mam zomom

 

Ammmmunmoumm me mxmmm

maflmmausoUH mo muflmcmusH m>flumfimmv

 

 

.mmcazmmogm0flm%o now must Hmnpommm mum:

.h magma

 

71

In the Spectra of the compounds [(CH3)2N]2PCN and

 

(CH$)2NP(CN)2, some ions were observed in which the P—N

bond had been cleaved, as evidenced by the presence of
immonium ions. However, there was a high abundance of

ions in which the P-N bond remained intact. This observa-
tion can be contrasted to the observation that the dimethyl-
aminodihalophosphines showed a high abundance of fragments,
which contained only phOSphorus and the halogen, but a low
abundance of fragments which contained P and N (75). In‘
the spectrum of (CH3)2NP(CN)2, the ion peak at m/e 101,
represented as the fragment P(CN)[N(CH§)2]+ (relative in~
tensity 100%), was found to be much stronger than the mass
peak at m/e 57 for the PCN+ ion (20%) and very much stronger

than the peak at m/e 83 for P(CN)2+ (4%)-

 

Nmr Spectra of the Ligands.- The proton nmr resonances

 

for these cyanophosphines (see Table 8) were observed to
be shifted downfield (less shielding) from the corresponding
chloride compounds, and were, therefore, easily identified
in the reaction mixtures.

The 31P nmr chemical shifts are Shown in Table 8.
These resonances shifted upfield as cyanide groups replaced
organic groups on the phosphorus atom. This observed trend
in the Chemical Shifts is opposite to the downfield trend
exhibited by the corresponding chloride compounds (50). For

example, the 31p chemical shifts for (CH3O)2PC1. CH3OPC12,

ii

 

 

72’

Table 8. Infrared and nmr Spectral data of cyanophosphines.

 

 

 

 

 

1 31 a
Compounds H JPH P CN~1
(ppm) (HZ) (ppm) (cm )
Ph-SP + 6.0b ---—
PhZPCN 7.20 9.0d + 36.4 2173 (s)
Php(CN)2 7.60c 10.0d + 75.6 2184 (s)
(ca,o)3p 3.43 10.5 —141.ob ———-
(CH3O)2PCN 3.75 10.3 —122.4 2186 (s)
CH30P(CN)2 3.88 9.0 - 24.4 2179 (s)
2030 (w)
(C2H5O)3P 3.85e 7.5g —139.0b —--—
1.19
(C2H50)2PCN 4.08e 7.5g —117.0 2180 (s)
1.30
c25509(cu)2 4.20: 7.09 - 17.6 2178 (s)
1.40 2133 (w)
[(CH3)2N]3P 2.44h 8.9h —123.0b —___
[(CH3)2N]2PCN 2.75 10.0 — 66.1 2168 (s)
(CH3)2NP(CN)2 2.90 14.0 + 5.7 2179 Is
. 2060 w
(CH3)3P 0.90 2.7 + 62.0b ___-
(CH3)2PCN 1.47 4.5 + 62.6 2172 (s)
CH3P(CN)2 1.90 7.5 + 81.41 2191 (s)
, 2095 (m)
P(CN)3 ———— —--— +138.3l 2186 s
_ 2096 m
as = strong, m = medium, w = weak. bSee reference 50. I
cMost intense peak. dApproximate coupling to ortho-protons
eMethylene protons. fMethyl protons.
?Coupling to methylene protons. hSee reference 37.

 

l .
CH3CN solution.

 

 

 

73

and PC13 are -169, -180, and -220 ppm, respectively (50).
Therefore, a cyanide magnetically shields the phOSphorus
atom while a chloride deshields it. This positive shield-
ing effect from the cyanides, as well as the trends in the
1H chemical shifts may be attributed at least partly to the
diamagnetic anisotropic shielding arising from the cyanide
triple bond. An X-ray diffraction study of solid P(CN)3
(76) indicated that the P-CEN linkage was nearly linear
(172° angle). Therefore, if the same linkage is linear in
the organic substituted compounds the phOSphoruS atom should
lie in the positive region of acetylenic shielding and ex-
perience an upfield shift, while the protons of the organic
groups may lie in the negative region of shielding and ex—
perience a downfield shift (73). In addition, it is note-
worthy to point out that the small C-P—C bond angle of 94°
found in P(CN)3 (76) indicates that the lone pair on
phosphorus has a large percentage of 5 character. This
should also contribute to the effective magnetic shielding
of the phOSphorus nucleus and give rise to the upfield
shifts.

The phsophorus-hydrogen coupling constants, also listed
in Table 8, varied with the different organic groups on the

phosphorus. The coupling, J increased in absolute value

PHI
when either -CH3, -N(CH3)2, or —Ph was replaced by cyanide

However, as cyanide groups replaced either -OCH3 or -OC2H5,

JP decreased in absolute magnitude.

H

 

 

 

 

74
Metal Complexes of Cyanophosphines.- If one to one
molar ratios of ligand and metal norbornadiene tetracarbonyl
were allowed to react, dimeric complexes involving bridging

cyanophosphine ligands were obtained.

 

M(CO)4C7H8 + RZPCN > [M(CO)4R2PCN]2

These bridged complexes were easily prepared When molybdenum
norbornadiene tetracarbonyl was allowed to react with the
ligands. The chromium analogs, however, were prepared with
slightly more difficulty. For both metals the formation of
the bridged complexes was observed as the appearance of a
bright yellow precipitate. A facile reaction was observed
for diphenyl- and dimethylcyanophosphines for both metals,
however, the diethylcyanophosphite and bis(dimethylamino)-
cyanOphosphine ligands, although readily reacting with I
molybdenum, reacted substantially Slower with chromium. In
addition, these latter complexes of chromium were not nearly
as stable to decomposition as the other complexes which
were prepared.

When two moles of ligand were added to the metal nor-
bornadienetetracarbonyl complex, the bridged complex would
again precipitate from solution. However, if this solution
was then refluxed for about an hour, the bridged species
gradually went into solution as the monomeric disubstituted
complex [M(CO)4L2] formed. In similar manner, the bridged
SPecies could undergo a reaction with a second mole of a

different phosphorus ligand (L') to give M(CO)4LL'.

 

 

 

75

 

p

 

M(CO)4C7H8 + RZPCN > [M(CO)4R2PCN]2 + RZPCN

[M(CO)4(R2PCN)2]

These disubstituted monomeric species were found to be un-
stable to air possibly because of the hygroscopic nature
of the cyanide group. (The compound, P(CN)3, is reported
to be extremely sensitive to moisture, yet inert to oxygen

at room temperature (74).)

Infrared Spectral Data of Metal Complexes of Some Cyano—
phosphine Ligands.- Table 9 lists the infrared data for the car-
bond and cyanide stretching regions for all of the complexes
isolated and characterized from the reactions of some of
these cyanophosphines with either molybdenum or chromium
carbonyls. Figure 8 Shows the shapes and intensities of
the infrared bands in the CO region of the disubstituted
complexes and the bridged complexes which were prepared.

All of the disubstituted complexes of molybdenum were white
and of the gi§_configuration, while the chromium analogs
were yellow and Erangedisubstituted with the exception of
the chromium complex of (CH3)2PCN which was pig, The C2V
(gig) and D4h (Egans) symmetries of these complexes were
easily determined from the number and intensity of the car—
bonyl stretching frequencies (11) (Table 9, Figure 8). The
Eggag- configuration found in the disubstituted complexes

of chromium may indicate that a sterically hindered arrange—

ment exists in the cis isomers of the bulkier ligands and

 

 

 

 

 

 

 

76

Table 9. Cyanide and carbgnyl infrared frequencies for

metal complexes.

 

 

b

 

Compounds v CEN v c.=_oC
[Mo(c0)4L12

L'=
pthcu 2161 2031,1945,1886
Php(CN)2 2170,2155 2044,1965,1920
(CH3)2PCN 2163 2028,1935,1887
(C2H5O)2PCN 2164 2037,1946,1906
[(CH3)2N]ZPCN 2159 2025,1932,1881
[Cr(CO)4L]2

Li=
pthcn 2158 2023,1930,1885
(CH3)2PCN 2161 2020,1934,1885
(C2H5O)2PCN 2150 2025,1985,1930
[(CH3)2N]2PCN 2155 2018,1927,1886
Mo(CO)4L2

L =
pthcn 2183 2039,1955,1932
(CH3)2PCN 2183 2040,1950,1925
(C2H50)2PCN 2183 2062,1997,1970
[(CH3)2N]2PCN 2176 2037,1945,1924
Cr(CO)4L2

L =
Ph2PCN (trans) 2182 1925
(CH3)2PCN 2182 2031,1945,1918

2173 1917

[(CH3)2N]2PCN (trans)

 

 

 

77

Table 9. (Continued)

 

 

 

Compound V CEN v CEOC
Mo(CO)4LL'

L = [(CH3)2N]2PCN 2175 2035,1960,1920
L' = (CH30)3P

L = [(CH3)2N]2PCN 2175 2040,1965,1930
L' = P(OCH2)3CCH3

Cr(CO)4LL' (trans)
L [(CH3)2N]2PCN 2185 1918
L' = (CH30)3P

 

aCHC13 solutions.

bCN stretching frequencies were all medium to very weak in

intensity.

c . . . . . .
Disubstituted monomeric complexes were of C15 configuration
unless otherwide indicated.

 

 

 

 

 

 

78

.mmxmameoo Ummtflhfl

 

map M Unfilmwxoamfioo touduflpmflsmflelmflo map mo macaw Hmumsom

map mpeommumou m

.c0flwmn 00 may CH meson toumumsfl 03» mo mommsm

I.w onsmflm

79

 

 

 

 

 

om mnsmflm

 

 

 

 

 

 

 

 

80
that a cis—trans rearrangement takes place after initial
attack Of the bridged complex by a ligand (see later).
This steric hindrance is apparently not significant for the
larger molybdenum atom, and the gig configuration is re-
retained. For the cyanophosphine-bridged metal complexes,
the CO stretching vibrations probably conform to a local
C2V carbonyl environment (about each metal) which requires
four CO stretching frequencies (2A1 + B1 + B2). The more
intense B2 band is generally of lower energy than the B1 band
for the gig-disubstituted complexes (13). For the dimers
studied, however, the reverse may be true, since the B1 band
appears to be a lower energy shoulder of the B2 band (Fig-
ure 8). Regardless of the assignment of these bands, how-
ever, the Shapes of the infrared absorptions in the CO
region (Figure 8) were always indicative of the type of
complex present (monomer or dimer).

The cyanide stretching mode (v was helpful in identi—

CN)
fying the resulting complexes, although not nearly as useful
as the carbonyl bands. The CN stretching frequency for the
bridging ligands occurred at lower energy than that

for the free ligand. The observed shifts to lower energy
for the bridging ligands ranged from 9 to 30 cm—l. The
largest shift occurred for the dimeric Cr complex of
(C2H5O)2PCN. Nixon and Swain (83) observed a similar shift

of 40 cm"1 for the Cr and Mo complexes of (CF3)2PCN. This

shift to lower energy for the CN stretch in these bridging

 

 

 

 

 

 

81
cyanophosphines is opposite to the shift observed for a
bridging thiocyanate ion in the complex [Pté(PBr§)éClé(SCN)2]
(77). Bridging SCN groups are reported to have higher
energy CN stretching frequencies than terminal SCN groups.
Furthermore, the CN stretching frequencies for nitrile
(RCN) complexes (55,78) are similarly reported to be of
higher energy than the stretching frequencies of uncomplexed
nitriles. Gerrard, eE_§i, (79) and Stolz, ei_ai. (80) have
discussed the increase in VCN for nitriles upon coordination
in terms of mesomeric forms of the uncoordinated nitrile
which are not possible in the complexed nitrile. In the
R2PCN bridged complexes the lower value of VCN (lower than
the value of the free ligand) may be the result of the
strain on the M—P-C—N angles in these bridged complexes.
If the nitrogen atom is sp hybridized (-CEN:) then a nearly
linear arrangement of the atoms P-C-NeM would result (M =
Cr or Mo). This linear arrangement would place a strain on
the M—P—C angles. (If the P~C-N system is nearly linear,
the M-P-C bond angle would have to be close to 90° for the
bridged complex.) However, if the Sp hybrid orbital on
nitrogen has some additional p character,which in the
limit would be Sp2 hybridized (—CEN:), then a smaller
C-N-M angle would result. This Smaller bond angle would
be more compatible with a bridged structure since some ring
strain would be relieved. The relatively small shift from

free to complexed ligand in v indicates that the CN bond

CN’

 

 

(I!

 

 

 

82
is still a triple bond.' (An sp2 hybridization on nitrogen
would require a CN double bond.) However, the addition of
some slight amount of p character to the sp orbitals
would leave the triple bond intact, yet not only relieve

ring strain but lower v slightly.

CN
The cyanide stretching region for the bridged molybdenum
complex of the ligand PhP(CN)2 exhibited two different cya—
nide stretching frequencies. These frequencies indicated
the presence of two distinct cyanide groups, one coordinated
and one not coordinated. Therefore, despite the two avail-
able nitrogen bonding Sites, the P-C-N bridge is preferred.
For the disubstituted monomeric metal complexes of these

cyanOphOSphines, the CN stretching frequency for the terminal

CN group either mxmnned essentially the same as the free

. L.“
3

ligand or actually increased Slightly in energy from the

value of the free ligand.

Proton Nmr of Metal Complexes.— The proton nmr spectra
of the bridged complexes are not significantly different
from the spectra of the free ligands except for a general
downfield shift in the resonances (Table 10). If the organic
groups attached to the phosphorus were either -Ph or —OC2H5,
the spectra were complicated by the addition of proton—
proton coupling to phosphorus-proton coupling. However,
for complexes of either (CH3)éPCN or [(CH3)2N12PCN

Simpler Spectra were obtained which were

 

 

 

83

Table 10. 1H nmr of complexesa.

 

 

 

Compounds 5c JP-H
[Mo(CO)4L]2
L =
thPCN 7.55m ___-
Php<cu>2d
(CH5)2PCN 1.94d 6.0
(C2H50)2PCN 1.30m (CH3) -___
3.70m CH2 7,0
[(CH3)2N]2PCN 2.82d ‘ 11.6
[Cr(c0)4L12
L =
thPCN 7.50m -___
(CH3)2PCN 1.92d 5.0
(C2H50>2PCN 1.41m (CH3) ____
4.20m CH2 7.0
[(CH3)2N]2PCN 2.79d 11,0
Mo(CO)4L2
L :
PhZPCN 7.40m -———
(CH3)2PCN 1.95”t" 5.7
(€2H50)2PCN 1.45m (CH3) -___
4.20m CH2 7.0
[(CH3)2N]2PCN 2.82“t" 11.6
Cr(CO)4L2
L =
thPCN 7.40m ——-—
(CH5)ZPCN 1.93"t" 6.8
[(CH3)2N]2PCN 2.76"t" 11.8

 

 

 

 

 

 

 

 

 

84

Table 10. (Continued)
C ounds 0C J

omp P—H
Mo(CO)4LL“
L = [(CH3)2N]2[CN 2.71“t" 11.8
L' = (CH3O)3P 3.55"t" 11.2
L = [(CHV3)2N]2PCN 2.75%" 11.6
L' — P(OCH2)CCH3 4.19"t" (CH2) 4.1
Cr(CO)4LL‘
L = [(CH3)2N]2PCN 2.74"t" 11.6
L' = (CH30)3P 3.64"t" 11.2

 

aAll spectra of complexes taken in CHClé, except nmr Spec—
tra of phenyl derivatives which were obtained in CH2C12.

bChemical shifts are in ppm (i 0.05) and are referenced to
tetramethylsilane. Coupling constants are in Hz (10.5).

cm = multiplet, d = doublet, “t" = apparent triplet due to

virtual coupling.

dNot sufficiently soluble for nmr spectrum.

 

 

 

 

85
useful in structural determinations (see below). The di—
substituted complexes [M(CO)4L2] of these latter two ligands
exhibited XnAA'X'n type Spectra (16,23) where X and A are
the protons and phosphorus of one ligand and X“ and A' are
are the protons and phosphorus of a second identical ligand
gig or trans to the first. These spectra consisted of a

doublet of separation (J + JAX"[ (16) plus a central peak

AX
of varying intensity. The magnitude of this central reso—
nance has been related to the magnitude of JPP“ (16,18,48).
The disubstituted molybdenum and chromium complexes which
were gig and 3323: respectively, (with the exception of gig-

r(CO)4[(CH3)2PCN]2) exhibited very weak central peaks, small
31P—31P coupling, consistent with other disubstituted com—
plexes of this geometry (18,21,22,23,37). It has been ob—
served that gig— disubstituted chromium complexes have
very strong phosphorus-phosphorus coupling (18,21,22,82,89)
(stronger than iigng—disubstituted chromium). This coupling
is demonstrated by the intense central peak observed for
the complex cis—Cr(CO)4[(CH3)2PCN]2 (see Figure 11).

The proton nmr spectra of the mixed ligand complexes

[M(CO)4LL'] can be considered to be MmABYn spin Sys—
tems similar to those discussed by Ogilvie, 3341i. (82),
where A and B are the tWO phosphorus atoms and M and Y are
the protons of the organic groups attached to the phosphorus
atoms A and B,respectively. For these systems, if J <<

AB

(VA - VB) (where VA and VB are the chemical shifts of the

 

 

 

 

86

Figure 9.- 1H nmr spectra of the mixed ligand complexes.
In both the Cr and the Mo complexes, L =
[(CH3)2N]2PCN and L' = (CH30)3P. (See Table

10 for chemical Shifts and coupling constants.)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88
protons on ligands A and B), then a doublet is observed
for each set of protons Split by the adjacent phosphorus

atom. If J 2 (VA - v ), then perturbation of the doub—

AB B

let Spectrum occurs (See Figure 9), and if JAB

apparent "triplets" due to virtual coupling are observed

>> (VA — VB),
(see Figure 9). Of the three mixed ligand complexes re—
ported in Table 10 only the Mo compound, where L =
[(CH3)2N]2PCN and L' = (CH30)3P, exhibited apparent "trip—
lets" in the 1H nmr spectrum with very weak central peaks
(Figure 9). In a previous study (82) of mixed ligand
complexes, only examples of the first two cases outlined

AB << (VA _ VB) and JAB z (VA — VB), were observed

The spectrum of the lone Cr mixed ligand complex is best

above, J

described as that of a perturbed doublet, JAB 3 (VA - VB),
(Figure 9), while the spectrum of the remaining Mo complex
with L = [(CH3)2N]2PCN and L = P(OCH2)3CCH3 is essentially

just two doublets; an example of the case where JAB <<

Mass Spectra of Complexes .— Nixon and Swain reported

 

(83) that they observed mass spectra for their dimeric com—
plexes of (CF3)2PCN characterized by fragmentation patterns
showing stepwise loss of all eight carbonyl groups while
the (MPCN)2 ring system remained intact. Of the complexes
reported in this work, mass spectra have been obtained for
[Mo(co)4L]2 where L = PhZPCN, PhP(CN)2, (CH3)2PCN, and

[(CH3)2N]2PCN and for {Cr(CO)4[(CH3)2N]2PCN}2. In these

 

 

 

 

 

89
cases, only the complexes with L = [(CH3)2M2PCN showed
parent ions and displayed fragmentation patterns similar
to those reported by Nixon and Swain (83). For the PhéPCN
and (CH3)2PCN complexes, the predominant fragmentation
patterns originated from the recombination ions of (CO)5MoL+
and Mo(CO)6+. In these cases there may very well be rapid
cleavage of the dimer, as a result of the metal—nitrogen
bonds breaking, leaving a species of the form M(CO)4L+,
followed by combination with CO. The PhP(CN)2 complex,

however, showed only rapid decomposition of the entire com-

plex.

Structure of Bridged Complexes.— The identity of the

 

yellow crystalline solids obtained by the displacement of
norbornadiene from C7H8M(CO)4 with these cyanophosphine
ligands was established by elemental analysis as being
[M(CO)4L]n, and the value of 2 for n, which was expected
because of the two potential bonding Sites on each ligand,
was confirmed by molecular weight measurement in one case
(see Experimental) and by the observation of some parent
ions in the mass Spectra. Therefore, as suggested by Nixon
and Swain for the Cr and Mo complexes of the ligand (CF3)2PCN
(83), a dimeric structure involving two bridging cyanophos—
phine ligands iS assumed.

There are two possible isomers for this type of dimer
(excluding CO bridging) (see Figure 10): one in which a

given metal atom is bound to the phosphorus atom of one of

 

 

 

9O

.psmofla msuosmmosm m. mom moms/moan Uwuoomxo oumofltca mwsfla
twpuon .xmameoo comment compass“ map How. numeOmH manflmmom 959

I.OH mesmflm

 

 

91

 

.oH madman

/\
/. \lTon/ _\

<</ II
\ \ .

/7mulla/

m¢m20w_

 

 

 

92

the cyanophosphine ligands and to the nitrogen atom of the
other (A), and a less symmetric isomer with both phosphorus
atoms bound to the same metal (B). The products obtained
from the reactions of these bridged complexes with other
ligands, as well as the proton nmr and infrared spectra,
indicate that the more symmetric of the two isomers is the
only isomer present.

When a dimeric complex bridged with ligand L reacts
with L', I would expect the metal-nitrogen bond to be cleaved
to yield two moles of a disubstituted product. If I assume
that the less symmetric structure (B) is present, then the
bridged complex would be expected to cleave asymmetrically
when it reacts with L' to yield two distinct complexes,
M(CO)4L2 and M(CO)4L'2. However, if the more symmetric struc—
ture (A) is present, I would expect symmetric cleavage of
the complex to yield two moles of M(CO)4LL'. The mixed
ligand complex is isolated exclusively from reactions of
[M(CO)4L]2 with L“. Therefore, if no rearrangement occurs
during reaction, isomer B can be ruled out. Additional
Support for isomer A comes from the nmr and infrared Spectra.
A has only one carbonyl environ—

The more symmetric isomer

ment, while structure B would have two distinct carbonyl

environments (one about each metal). In the infrared

Spectra only one carbonyl environment is observed. The

 

nmr Spectra of the cis disubstituted complexes,

M(CO)4L2, where L = (CH3)2PCN or [(CH3)2N]2PCN:

 

 

 

93
clearly Show the presence of virtual coupling. If in the
bridged complex, isomer B were present, with two phosphorus
atoms bound in a gig-configuration to a single metal atom,
one would similarly expect to see some indication of virtual
coupling in the nmr spectra. The spectra of the dimers,
however, are sharp doublets (Figure 11) with no evidence
of virtual coupling, a further indication of the presence

of the more symmetric isomer (A).

 

 

94

.xmamfioo tmqoflufl
mo musuodnum ocHEkumt Op toms muuoomm HES couonm map mo mmamemxm

I.HH onsmflm

 

 

95

1:!Ifi

 

.HH madman
«ad «300.: .255:
<42
and £3092. .853 .8933
i is: a...
and zo..«.z«.n:§ .253...

.515“

 

 

3.2.2 I.

 

CONCLUSION

It was found that cyanophosphines could be easily
prepared from the corresponding phOSphorus chlorides and
silver cyanide if acetonitrile was used as a solvent.
These compounds were then observed to form coordination
compounds with chromium and molybdenum carbonyls. Because
of the potential donor Sites on both phosphorus and nitro-
gen, the formation of the cyanophosphine bridged dimers
was not totally unexpected. The ease with which these
reactions took place was somewhat surprising, however,
since even though nitrile coordination is well known (10,
55,80), replacement of a diene with a nitrile is not com-
mon. In fact nitriles have been used as intermediates for
the formation of diene or triene metal carbonyls (54).
The second step in these reactions, cleavage of the dimer
by the phosphorus of a second mole of ligand was antici-
pated Since phosphines are known to replace nitriles in
group VIB metal carbonyl complexes (89).

Comparison of the infrared Spectra of the molyb—
denum complexes of (CH3)2PCN, PhZPCN, and (C2H50)2PCN
to those of P(CH3)3, PPh3, and P(OC2H5)3, indicates
that the cyanophosphine complexes always exhibit the

higher CO stretching frequencies. This observation

96

 

 

97
indicates that the F—accepting capacity of the phosphorus
atom is increased by the bonding of cyanide group to it.
An indication of the strong effect that these cyanide groups
have on the phosphorus—metal bonding can be obtained by
comparison of the cyano group to a chloride. Halides, as
substituent groups on a phosphorus, are generally thought
to cause a pronounced increase in the w-accepting capacity
of phosphorus ligands (13). The complex Mo(CO)4(PClZOC2H5)2
with two chlorides on the phosphorus, has CO stretching

1

frequencies at 2063, 1977, and 1968 cm’ (13), while

Mo(CO)4[P(CN)(OC2H5)2]2 has co frequencies at 2062, 1997,
and 1970 cm—1. This Shift may indicate that cyanide groups
cause an even larger increase in the ability of a phosphorus
ligand to accept metal d—F electron density than do chlorides.
Figure 10 shows the two possible isomers of the
bridged complexes. The drawings are not intended to indi—
cate a planar molecular structure of the ring system.
Nixon and Swain (83) suggested by comparison to the planar
arrangement of the eight membered ring in the thiocyanate
bridged Pt complex, Pt2(PBr3)2C12(SCN)2, for which a crys—
tal structure is known (87), that their c0mplex might have
a planar ring system. If, as I have suggested in the
infrared section here, the nitrogen (or carbon) atoms have

some added p character to the Sp orbitals, the PCN bonds

might bow out slightly to relieve ring strain. If this

 

 

 

L..___

98
in fact happens, then a planar arrangement of the ring
system should result. If, however, these sp orbitals
on both C and N acquired a significant amount of p
character, the ring would pucker, thereby eliminating

the possibility of planarity.

 

 

 

SUGGESTIONS FOR FUTURE WORK

There are several logical extensions of the chemistry
investigated here. The most obvious work would be to study
further the complexes already prepared. A thorough study
of the carbonyl stretching frequencies and force constants
would help to clarify the nature of the bonding in these
Cyanophosphine complexes. More interesting, however, would
be a crystal structure determination to discover the exact
configuration of the proposed eight membered ring. This
ring structure would give information suggesting the orbital
configurations (hybridizations) about the phosphorus, car—
bon, and nitrogen. Since the P—C-N bond angles in P(CN)3
are about 1720 (76), and in PFZCN approximately 1650' (88),
a remarkable deviation from linearity for this P—C-N group
might be found in these complexes.

A most interesting reaction to investigate might be
that of the Sigfdisubstituted complexes of the type Eig-
M(CO)4(R2PCN)2 with another diene complex, M'(CO)4diene.

This reaction could produce a bridged dimer with two differ—

ent metals (Mo and Cr), or if the same metal were to be used

for both groups, and if there were no rearrangement upon reac-

tion, the less symmetric of the two isomers given in Figure 10

would be produced in which both phosphorus atoms are bound

99

 

 

llll J

100

to the same metal

Two other areas of interest for the extension of this
work would be, first, to attempt the preparation and char—
acterization of other metal complexes to determine if
cyanophosphine bridges will form with other metals, such
as Fe, Ni, Pd, and Pt, and secOndly, to investigate the
reaction of these cyanophosphines, RZPCN and RP(CN)2, with

various boron Lewis acids. This study Should provide in—

formation about the relative basicities of phosphorus versus

the cyanide lone pair of electrons.

 

BI BLIOGRAPHY

 

11.

12.

13.

14.

15.
16.

17.

BIBLIOGRAPHY

G. Booth, Advan. Inorg. Chem. Radiochem., 6, 1 (1964).

G. Wilkinson and F. A. Cotton, Progr. Inorg. Chem., i,
1(1959).

 

E. O. Fischer and H. P. Fritz, Advan. Inorg. chem.
Radiochem., i, 55 (1964).

 

E. O. Fischer and H. P. Fritz, Amgmh Chem., 73, 353
(1961). " -

P. L. Pauson, in "Organometallic Chemistry," (H.
Zeiss, ed.), New York: Reinhold, 1960, p. 346.

H. Zeiss, in "Organometallic Chemistry," (H. Zeiss,
ed.), New York: Reinhold, 1960, p. 380.

W. Strohmeier, Angew. Chem. Intern. Ed. English, 3,
730 (1964). N

E. W. Abel, Quart. Rev. (London), 12/ 133 (1963).

 

W. Hieber, Adv. Organgmetal. Chem., 8” 1 (1970).

 

G. R. Dobson, I. W. Stolz, and R. K. Sheline, Advan.
Inorg. Chem. Radiochem., 8, 1 (1966 .

 

F. A. Cotton and C. S. Kraihanzel, J. Amer. Chem. Soc.,
§é, 4432 (1962). “I”

F. A. Cotton and C. S. Kraihanzel, Inorg. Chem., 2,
533 (1963).

F. A. Cotton, Inorg. Chem., 3, 702 (1964).

R. Poilblanc and M. Bigorgne, Bull. Soc. Chim. France,
1301 (1962).

 

W. Hieber and T. Kruck, Z. Naturforsch., igb, 709 (1961).
R. K. Harris, Can. J. Chem., 42” 2275 (1964).

R. K. Harris, Inorg. Chem., 5, 701 (1966).

101

 

 

 

 

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

 

m m_’1‘t_FH—__ -4. __.__ ,0.

102

F. B. Ogilvie, J. M. Jenkins, and J. G. Verkade, J;
Amer. Chem. Soc., 92/ 1916 (1970).

R. D. Bertrand, F. B. Ogilvie, and J. G. Verkade, g;
Amer. Chem. Soc., 92/ 1908 (1970).

F. B. Ogilvie, R. L. Keiter, G. Wulfsberg, and J. G.
Verkade, Inorg. Chem., 8” 2346 (1969).

R. D. Bertrand, D. A. Allison, and J. G. Verkade, J;
Amer. Chem. Soc., 92, 71 (1970).

J. M. Jenkins, J. R. Moss, and B. L. Shaw, J. Chem.
Soc. (A), 2796 (1969).

J. M. Jenkins and B. L. Shaw, Proc. Chem. Soc., 279
(1963).

J. M. Jenkins and B. L. Shaw, J. Chem. Soc. (A), 770
(1966). ' I

J. M. Jenkins and B. L. Shaw, J. Chem. Soc. (A), 1407
(1966).

J. R. Moss and B. L. Shaw, J. Chem. Soc. (A), 1793
(1966). ““““““‘

P. R. Brookes and B. L. Shaw, J. Chem. Soc. (A), 1709
(1967). '

E. W. Abel, R. A. N. Mclean, and S. P. Tyfield, J;
Molec. Spect., 39, 29 (1969).

D. M. Adams, J. Chem. Soc., 1771 (1964).

M. A. Bennett and R. J. H. Clark, J. Chem. Soc., 5560
(1964).

A. A. Chalmers, J. Lewis, and R. Whyman, J. Chem. Soc.
(A), 1817 (1967).

A. C. Vandenbroucke, Jr., D. G. Hendricker, R. E. Mc

Carley, and J. G. Verkade, Inorg. Chem., Z, 1825 (1968).

J. M. Jenkins and J. G. Verkade, Inorg. chem., 6, 2250
(1967).

A. Whitaker and J. W. Jeffery, Acta. Cryst. 23” 977
(1967).

G. Huttner and S. Schelle, J. Organometal. Chem., i9,
p9 (1969).

 

 

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.
52.
53.

54.

103

D. J. Darensbourg and T. L. Brown, Inorg. Chem., l,
959 (1968).

R. B. King, Inorg. Chem., Ex 936 (1963).

R. J. Clark and P. I. Hoberman, Inorg. Chem., 4” 1771(1965).

F. A. Cotton and G. Wilkinson, "Advanced Inorganic
Chemistry," New York: Interscience Publishers,
1967, p. 728.

R. J. Angelici and Sr. M. D. Malone, Inorg. Chem.,
3, 1731 (1967).

L. S. Meriwether and J. R. Leto, J. Amer. Chem. Soc.,
83” 3192 (1961).

 

S. O. Grim, W. McFarlane, and D. A. Wheatland, Inorg.
Nuc. Chem. flLetters, 2, 49 (1966).

M. Lenzi and R. Poilblanc, Compt. Rend. Series C, 263,
674 (1966).

 

S. I. Shupack and B. Wagner, Chem. Commun., 547 (1966).

G. S. Reddy and R. Schmutzler, Inorg. Chem., 6, 823
(1967). N

H. J. Plasts, J. M. Stewart, and S. O. Grim, J. Amer.
Chem. Soc., 2;, 4326 (1969).

R. E. Dessey and L. Wieczorek, J. Amer. Chem. Soc., 91,
4963 (1969).

J. G. Verkade, R. E. McCarley, D. G. Hendricker, and
R. W. King, Inorg. Chem., 4, 228 (1965).

K. J. Coskran, R. D. Bertrand, and J. G. Verkade, J.
Amer. Chem. Soc., 89” 4535 (1967).

V. Mark, C. H. Dungan, M. M. Crutchfield, and J. R.
VanWazer, Topic in Phosphorus Chem., 5, 227 (1967).

L. Maier, Helv. Chim Acta, 2667 (1963).

R. B. King, Organometal. Syn., 1, 124 (1965).

N

 

R. J. Angelici, J. Chem. Educ., 4g” 119 (1968).

 

R. E. King and A. Fronzaglia, Inorg. Chem., 2, 1837
(196

 

55.

56.

57.

58.

59.

60.

61.
62.
63.

64.

65.
66.
67.

68.

69.

70.

71.

72.

73.

74.

104

B. L. Ross, J. G. Grasselli, W. M. Ritchey, and H. D.
Kaesz, Inorg. Chem., 2, 1023 (1963).

R. L. Keiter and J. G. Verkade, Inorg. Chem., 8, 2115
(196

E. D. Morris, Jr. and C. E. Nordman, Inorg. Chem.,
8, 1673 (1969).

C. G. Barlow, J. F. Nixon, M. Webster, J. chem. Soc.
(A), 2216 (1968).

J. G. Verkade, private communication.

S. L. Manatt, G. L. Juvinall, R. I. Wagner, and D. D.
Elleman, J. Amer. Chem. Soc., 88” 2689 (1966).

R. Schmutzler, J. Chem. Soc., 5630 (1965).
W. Strohmeier and K. Gerlack, Chem. Ber., 94/ 398 (1961).
R. B. King, Inorg. Nucl. Chem. Lett., 5, 905 (1969).

N

P. G. Kirk and T. D. Smith, J. Inorg. Nucl. Chem., 30,
892 (1968). -

P. A. StaatS and H. W. Morgen, Inorg. Syn., 6, 84 (1960).
H. Noth and H. J. Vetter, Chem. Ber., gg, 1109 (1963).
H. Noth and H. J. Vetter, Chem. Ber., 9g, 1479 (1963).

K. A. Petrov, L. G. Gatsenko, and H. A. Neimysheva,
Zh. Obshch. Khim, 29, 1827 (1959).

P. G. Kirk and T. D. Smith, Chem. Commun., 42 (1968).

P. G. Kirk and T. D. Smith, J. Chem. Soc. (A), 2190
(1969).

Van Ebbelhard Fluck and J. R. Vanwazer, Z. Anorg. Allg.
Chem., 307, 113 (1961).

G. W. Parshall, J. Inorg. Nucl. Chem., 12” 372 (1960).

J. R. Dyer, "Applications of Absorption Spectroscopy
of Organic Compounds," Englewood Cliffs, N.J.:
Prentice Hall, Inc., 1965, p. 74—79.

V. J. Goubeau, H. Haiberle, and H. Ulmer, Z. Anorg.
Allg. Chem., 311, 110 (1961).

 

75.

76.
77.

78.

79.

80.

81.

82.

83.

84.

85.

86.
87.
88.

89.

105

 

W. Z. Borer and K. Cohn, Anal. Chim. Acta, 47, 355
1969). “N

K. Emerson and D. Britton, Acta Cryst., i1, 1134 (1964).
J. Chatt and L. A. Duncanson, Nature, 178, 997 (1956).

H. J. Coerver and A. C. Curran, J. Amer. Chem. Soc.,
89, 3522 (1958).

W. Gerrard, M. F. Lappert, H. Pyszora, and J. W. Wallis,
J. Chem. Soc., 2182 (1960).

I. W. Stolz, G. R. Dobson, and R. K. Sheline, Inorg.
Chem., 2, 323 (1963).

T. R. Johnson, R. M. Lynden—Bell, and J. F. Nixon, 1L
Organometal. Chem., 21, p15(1970).

F. Ogilvie, R. J. Clark, and J. G. Verkade, Inorg. Chem.,
8, 1904 (1969).

J. F. Nixon and J. R. Swain, J. Organometal; Chem., EL,
13 (1970).

F. A. Miller, S. G. Frankiss, and D. Sala, Spectrochim.
Acta, 2i, 775 (1965).

F. W. Bennett, H. J. Emeleus, and R. N. Haszeldine,
J. Chem. Soc., 1565 (1953).

K. J. Packer, J. Chem. Soc., 960 (1963).
P. G. Owston and J. M. Rowe, Acta Cryst., 13, 253 (1960).
K. Cohn, private communication.

C. E. Jones and K. J. Coskran, Inorg. Chem., 12, 55
(1971).

 

 

 

 

 

 

 

“0151611113111:will)“