A PART I
SPECTROSCOPIC STUDIES OF
FOUR-AND FIVE-COORDINATE

TRANSITION METAL COMPLEXES WITH

TERTIARY PHOSPHORUS LIGANDS

PART II
REACTIONS OF SOME CYANOPHOSPHINES
WITH DIBORANE AND BORON TRIFLUORIDE _

Thesis for the‘Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
EDWARD JOSEPH LUKOSIUS
1 9 7 2

4-4.; ‘AL

. a LIBRHRY

Michigan State
University

 

'7'

This is to certify that the

thesis entitled
Part I. Spectroscopic Studies of Four- and Five-Coordinate
Transition Metal Complexes with Tertiary Phosphorus
Ligands
Part II. Reactions of Some Cyanophosphines with Diborane and
Boron Trifluoride

presented by

Edward Joseph Lukosius

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Chem; 5&1!

Major professor

 

Date July 28. 1972

0-7639

 
       

 
 

.uéhnay

NOAH & SUNS’
800K BINDERY INC.
LIBRARY amozns
:.. :I’:::::-‘::T. Inn-m

ABSTRACT

PART I
SPECTROSCOPIC STUDIES OF FOUR-AND FIVE-COORDINATE

TRANSITION METAL COMPLEXES WITH TERTIARY
PHOSPHORUS LIGANDS

BY

Edward Joseph Lukosius

Infrared, electronic, and proton nuclear magnetic
resonance data are presented and discussed for the following
complexes: NiL3(CN)3, NiL2(CN)3 (L = (CH3)3P, (CH3)2POCH3.
cnap(ocna),, P(OCH3)3), [NiL5](C104)3 (L = (CH3)21>ocn3.
CHSP(OCH3)2, P(OCH3)3). [C0L51C104 and [RhL5]B(C3H5)4 (L =
P(0CH3)3), and {Ni[(CH3)3P]4](BF4)2. The five-coordinate
complexes are believed to have a trigonal bipyramidal struc-
ture, both in the solid state and in solution, on the basis
of their infrared and electronic spectra. Equilibria of
the following type are found in solutions of the five-co—

ordinate complexes:

 

NiL3(CN)2 _3_ NiL3(CN)2 + L

<

 

+
[ML51n > [ML4]n+ + L

<

 

Isosbestic points have been observed in the ultraviolet-
visible Spectra of these solutions. Electronic factors

which favor the formation of five-coordinate complexes over

Edward Joseph Lukosius
four—coordinate complexes have been discussed on the basis
of the electronic spectra.

The temperature dependence of the proton magnetic
resonance spectra of the above equilibria have been investi-
gated, and are discussed in terms of ligand exchange. The
proton nmr spectra are similar for all complexes; a doublet
at fast exchange rates, sharp singlet at intermediate rates,
and multiplet at slow exchange rates. The doublets at fast
ligand exchange rates are simply 31P-1H coupling, however,
the coupling constants and chemical shifts are an average of
complexed and free ligand in solution. Decoupling of phos—
phorus and hydrogen nuclei at intermediate exchange rates
where singlets are Obtained is believed due to a spin-ex-
change mechanism for which strong 31P-31P coupling is
necessary. Multiplets at slow exchange rates are due to
31P-31P and 31P-1H coupling through the metal. JPCH is
found to be of opposite sign in complexed and free ligand,
while JPOCH has the same sign. Qualitative rates of ligand
exchange have been determined and follow the order: (CH3)3P >
P(OCH3)3 > CHaP(OCH3)3 > (CH3)2POCH3. In the metal complexes
of the typeML5n+ where L = P(OCH3)3, the rate of ligand

exchange follows the order: Co+ > Rh+ > Ni2+.

Edward Joseph Lukosius
PART II
REACTIONS OF SOME CYANOPHOSPHINES WITH DIBORANE
AND BORON TRIFLUORIDE
The following adducts have been prepared by the reaction

of RgPCN (R - CH3. cnao. C3H5O. cana, N(CH3)2) with diborane
and boron trifluoride; [R2PCN]BH3 (R = all of the groups
listed above), [R3PCN13F3 (R = CH3, CH3O, canal and
[RaPCNlBF3(Bfia) (R = CH3. CH30). Nuclear magnetic reso-

nance (118, 31P, 19F, 1H) and infrared data (v are pre-

CN)
sented and discussed. 0n the basis of these data, it is
concluded that the borane molecule is coordinated to the
phosphorus atom and the BF3 molecule is bound to the nitro-
gen atom of the cyanide group in all of the above adducts.
The borane hyperconjugation model of the B-P bond has

been shown to be inconsistent with the nmr data. However,

a correlation between JPB and phosphine basicity has been
observed in the borane adducts. Infrared data have suggested
that the CN bond in the free cyanophosphines is inter-
mediate between a triple and double bond. Experimental and
spectroscopic results have indicated that the cyanide group
decreases the basicity of the phosphorus atom to which it

is bound in these RgPCN ligands,as compared to the corre-

sponding PR3 ligands.

PART I

SPECTROSCOPIC STUDIES OF FOUR-AND FIVEoCOORDINATE
TRANSITION METAL COMPLEXES WITH TERTIARY
PHOSPHORUS LIGANDS

PART II

REACTIONS OF SOME CYANOPHOSPHINES WITH DIBORANE
AND BORON TRIFLUORIDE

BY

Edward Joseph Lukosius

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1972

To Susan

ACKNOWLEDGMENT

It is with sincere appreciation that I acknowledge the
encouragement and direction of Dr. Kenneth J. Coskran during
this investigation.

I am deeply grateful to my wife, Susan, for her under-
standing and inspiration. I wish, also, to thank our
parents, Mr. and-Mrs. Ignatius F. Lukosius of Boston,
Massachusetts and.Mr. and Mrs. John W. Cotter of Dedham,

Massachusetts for their assistance, advice and encouragement.

iii

TABLE OF CONTENTS

INTRODUCTION . . . . .
EXPERIMENTAL . . . . .
RESULTS AND DISCUSSION

Infrared Spectra

Absorption Spectra
Proton Nmr Spectra

INTRODUCTION . . . . .
EXPERIMENTAL . . . . .

Materials . . . .
Preparation

Preparation
Reaction of
Nmr Spectra
Infrared Spectra

of [R
RP(CN3

RESULTS . O O O O O 0
DISCUSSION . . . . . .

[RaPCN]BH3 Adducts

mammal?3 and [RgPCN1BF3(BH3) Adducts

CONCLUSION . . . . . .

BIBLIOGRAPHY . . . . .

PART I

2 With BF3

iv

of [R3PCNlBH3 Adducts
Preparation of [R2PCN1BF3 Adducts
PCN]BF:(BH3) Adducts

Page

11
19
19
41

102
107

109

Table

II.
III.
IV.

V.

VIII.

IX.

XI-

XII-

XIII.

LIST OF TABLES

PART I

Five-coordinate Ni(II) complexes with mono-
dentate ligands . . . . . . . . . . . . . . .

Infrared data, cyanide stretching frequencies

Absorption spectral data for Ni(CN)3 complexes

K for NiL3(CN)2 complexes at 25°. . . . .

inst

Absorption spectral data for Ni(ClO4)2 and
Ni(BF4) complexes . . . . . . . . . . . . . .

Temperatures of singlet maxima and magnetic
susceptibility data . . . . . . . . . . . . .

PART II
Experimental data for 3H3 and BF3 adduct
formation . . . . . . . . . . . . . . . . . .
Nmr and ir data for [RaPCN]BH3 adducts . . .

Nmr and ir data for [RzPCNlBFa and
[RgPCN]BF3(BH3) adducts . . . . . . . . . . .

Values of JPCH' JPOCH' and JPNCH in

R2 PCN and [R2PCN18H3 O o o c o o o o o o o o

PhOSphorus chemical shift data in
R2 PCN and [R2 PCN] BHS o o o o o o o o o o o o

61H(R) in RaPCN and [RzPCN]BH3 . . . . . . .

Change in 631P in RzPCN and [RaPCN]BH8 upon
BF3 adduct formation . . . . . . . . . . . .

Page

20

33

34

64

80
86

87

95

99
101

105

LIST OF FIGURES

PART I
Figure Page

1. (I) Trigonal bipyramidal coordination Sphere
surrounding Ni(II) in [Ni(rsp)c1]c1o4 (TSP =
tris(gamethylthiophenyl)phosphine). (II)
Tetragonal pyramidal arran ement of arsenic
atoms around Ni(II) in Ni diars)(triars)](ClO4)2
(diars = g-phenylenebis dimethylarsine): triars
='bis(g-dimethylarsinophenyl)methyl arsine) . 2

2. Effect of distortion on d orbital energy
levels in Dsh symmetry . . . . . . . . . . . 5

3. Splitting of d orbital energy levels in d8
complexes of C4v symmetry . . . . . . . . . 6

4. Pr0posed dimeric structure of NiL2(CN)3
(L I P(OCH8)3)018 o o o o o o o o o o o o o o 21

5. Ultraviolet-visible spectrum of: (a)
Ni[P(0CH3)3]3(CN), in cnzclz: (b). solution
in (a) with excess ligand . . . . . . . . . . 25

6. Absorption spectrum of (a) Ni[(CH3)3P]3(CN) in
CH3c1a; (b) NiI$C33;3P]2ECN;2 in CH3C12: (c
solid state Ni[ CH3 3P]3 CN

7. Effect of temperature change on the ultraviolet-
visible spectrum of Ni[P(OCH3)3]3(CN)2 in
CHaCJ-z o o o o o o o o o o o o o o o o o O o o 31

2 o o o o o o o o 27

8. Isosbestic points obtained for {Ni[P(OCH3)3]5}-
(Cl-0‘ )2 in CHgCla o o o o o o o o o o o o o o 37

9. 1H nmr spectrum of Ni[(cnagp]2(cn)2 in CH3C12
at various temperatures . . .-. . . . . . . . 44

10. Portion of ultraviolet-visible spectrum of
Ni[(m3)3P]3(CN)3 in CH3C12 o o o o o a o o o 47

11. 1H nmr spectrum of Ni[(CH3)3P]3(CN)2 in
Cflgcla at various temperatures . . . . . . . . 52

vi

LIST or FIGURES (Cont.)
Figure Page

12. 1H nmr Spectrum of NiL3(CN)3 (L - (CH3)3POC33.
CH3P(OCH3)3) in CH3C13 at various temperatures 54

13. 1H nmr spectrum of:(a) Ni[P(OCH3)3]3(CN) in

CH C13: (b) solution in (a) with added P OCH3)3:
(c solution in (b) when cooled . . . .~. . . 57

14. 1H nmr Spectrum of Ni[(CH3) P]3(CN)2 upon
successive addition of (C3333P.. . . . . . . . 60

15. 1H nmr spectrum of Ni[CH3P(0CH3)2yJ(CN)3 in
CH3613 upon addition of CH3P(OCH3 2 . . . . . 63

16. 1H nmr spectrum of {Rh[P(0CH3)3]5}B(C5H5)4 in
C33C13 at various temperatures . ... . . . . . 69

PART II

17. Structure of [M(CO)4L]2 com lexes (M ' Cr or Mo:
R - CH3; 0C3H5. C5115. N(CH3§2)75 9 o o o o o o 75

18. Nmr spectra of [(CH3)3PCN]BH : (a) Time-averaged
13 spectrum at 60.0 MHz. (b3 1H spectrum at 60.0

MHz. (c) 113 spectrum at 60.0 MHz. (d) 31F
spectrum at 24.3 MHz . . . . . . . . . . . . . 85

19. 113 nmr spectrum of [(cnao) PCN]BF3(BH3) at 32.1
MHz: (a) Spectrum at -25 ; (b) Spectrum at -50° 90

20. The ligand cone angle for tertiary phosphorus
compounds35. . . . . . . . . . . . . . . . . . 97

21. Possible structures of [RCNlBF351 . . . . . . 102

vii

INTRODUCTION

Four-coordinate complexes of Ni(II) have been known
for some time.1 In recent years, a large number of five-
coordinate Ni(II) complexes with polydentate ligands have
been prepared}!3 The ability of these complexes to attain
pentacoordination, as well as their resulting structures,
are determined primarily by the restrictive geometry of
these polydentate ligands."6 This is exemplified in
Figure 1, where the structures of two such complexes (1 and
II), as determined by X-ray studies,"v8 are pictured. The
restrictive ligand geometry in Ni[(TSP)Cl]ClO4(I) forces
the phosphorus and sulfur atoms into four positions of a
trigonal bipyramid around Ni(II), allowing the fifth un-
hindered position to be filled by a chloride ion.9 In
[Ni(diars)(triars)](C104)z(II), the triarsine ligand geometry
forces an arsenic atom into the fifth coordination site of a
tetragonal pyramid.10 For monodentate ligands this type of
ligand restrictive geometry is not possible, and indeed,
few five—coordinate Ni(II) complexes with monodentate ligands

have been reported.

P IA:
N 5 (4h! / As
[-—
s/ I A5/ '\A3/
X
I' 'II

Figure 1. (I) Trigonal bipyramidal coordination Sphere
surrounding Ni(II) in [Ni(TSP)Cl]ClO4 (TSP =
tris (g-methylthiophenyl)phosphine).

(II) Tetragonal pyramidal arrangement of arsenic
atoms around Ni(II) in [Ni(diars)(triars)](C104)2
(diars - g-phenylenebis(dimethylarsine); triars
= bis(g—dimethylarsinophenyl)methyl arsine).
This latter group consists of two main types; NiL3X2

(L = tertiary phosphorus ligand; X = CN, Cl, Br, I), and

[NiL51X2 (L - tertiary phosphite; x - c104, BF4). In 1960

the first five—coordinate Ni(II) complex with monodentate

ligands was prepared,11 Ni[P(C2H5)2C5H5]3(CEOCGH5)2 , which

dissociated reversibly in solution to free ligand and the

corresponding four-coordinate complex (Equation 1). Five—

 

NiIP(C2H5)2CSHSIS(CECCSHS)2 <

 

1
P(C2H5)2C6H5 + NiIPICzfls)2CsHalz(CECC6H5)2 ( )
coordinate complexes that have been isolated since that
time are listed in Table I.
The stereochemistry adopted by these complexes is not
determined by ligand steric restrictions, as in complexes
with polydentate ligands. However, with regard to ligand-

ligand repulsion, it has been shown that the trigonal

 

 

 

Table I. Five-coordinate Ni(II) complexes with mono-
dentate ligands.

Compound Reference
NiL3x3

L x
HP(C5H5 )2 Clo Br. I [12]
PC14H13 Cl, Br, I [13]
RPC33H3 (R = CH3. C3H3) C1, Br. I [14]
c6H5P(CH3)2 I. CN [15]
(CH3)3P Br. I [16]
P C4H9) CN [17]
P 0CH3 CN [18]
P 0C3H3§3 CN [19]
P 0C3H 3 CN [18]
P OCH )3CCH3 CN [18]
C3H3P 0CH3) CN [19]
C3H3P 0C3H )3 CN [19]
C3H3P C3H ): CN [20]
C5F5P CH3 2 CN [15]
(C6H5)2PCH8 CN [15]
[NiL31x3

L x
P OCHs 3 C104 [21]
P OCH, 3CCH3 C104 [21]
P 0CH3 3CCH3CH3 C104 [21]
C(CH3) 4CH3 C10, [21]

POCH) 3(CH3) )3 PF, [21]
Ni(CN)33 [221,[23]
Nisr3[(CH3)3P]2 [24]

 

4

bipyramid is the most stable structure, being favored over
the tetragonal pyramid.25 Indeed, X—ray structure studies
of the complexes Ni[C5H5P(OC2H5)2]3(CN)3,25
Ni[C3H3P(CH3)3]3(CN)3,27 Ni(HP(C3H3)3]3I3.28 and
{Ni[P(OCH)3(CH2)3]5)(Clo‘)229 have demonstrated a near
trigonal bipyramidal arrangement of ligands about the Ni(II)
atom, with Erggg_apical cyanides in the appropriate com-
plexes. Only one square pyramidal arrangement of monoden-
tate ligands about the Ni(II) atom has been determined by
X-ray structure studies. In [Cr(NH3CH2CH2NH2)3]Ni(CN)5'
1.5H30,3° the Ni(CN)5-3 ions exist in both a square pyram-
idal. and distorted trigonal bipyramidal structure.

Electronic Spectra of these five-coordinate complexes
support the assignment of a trigonal bipyramidal structure
in solid and solution. For a low spin, d8, trigonal bi-
pyramidal complex of D3h symmetry, two electronic transi—
tions are expected as the result of metal d orbital Split-
ting.31 v.[(e">4(e')4 -> <e">4<e-)3(a.)11 and v.[<e")4<e'>4
-> (e")3(e')4(a1)1] (Figure 2). Transition v1 is dipole
allowed while v2 is forbidden.32 The electronic spectra
of a number of the complexes in Table I exhibit these transi-
tions; an intense band at low energies (v1) and a weak band
at higher energies (v3).14v1903° However, in a number of
these complexes the transition v1 has been observed to
split into two distinct bands, in both the solid and solu-
tion spectra.15'18'19 In the solid state this splitting

has been attributed to a ground state distortion of the

5
ligand field levels of these complexes because of lattice
packing interactions and the effects of ligand conformations
or because of a second order JahndTeller effect (see Figure
2).33 Thus the symmetry of the complex is reduced and v1

will appear as two bands. The splitting of v1 is expected

to be larger than that of v3 since the dx3_y3 and dxy
orbitals are more sensitive to distortion than the dxz
and dyz orbitals.33

 

 

 

 

 

 

 

 

 

a] d.2" ..
x), 93 1 ”I II 42
. .d.2.,2- -°
. '°dxy-- .
O” a , I '. dxz. . O .
03.1 (or C3V) C2v(or Cs)

Figure 2. Effect of distortion on d orbital energy levels
in D symmetry.
ah

The solution electronic spectra of these complexes
exhibit the same splitting of VI; however, at low tempera-
tures, the two absorptions which constitute v1 are Observed
to coalesce and at ~v77°K, a single, very symmetrical band
is observed. Venanzi and coworkers 33 have proposed a

dynamic JahndTeller distortion of excited states to explain

6
the splitting of VI at high temperatures. These authors
have also preposed that a ground state distortion, in which
different degrees of distortion could exist at different
temperatures, would account for the splitting of VI. A
tetragonal pyramidal structure does not account for the
Observed spectra, Since three bands should be observed at
all temperatures34 (Figure 3). Thus, five-coordinate Ni(II)
complexes with monodentate ligands are believed to have a
trigonal bipyramidal structure in solid and solution as a

result of studies of their absorption Spectra.

 

bl 7} de2.y2

 

 

 

 

 

 

0] — dzz
b2 *‘* dxy
e L‘fi gdxztdyz
Gav

Figure 3. Splitting of d orbital energy levels in d8
complexes of C4v symmetry.
Infrared studies support this assignment of a trigonal
bipyramidal structure. One CN stretching frequency appears

in both the solid and solution (excess phOSphine is added

7
to prevent dissociation) spectra for NiL3(CN)3 complexes,
indicating a Eggggfsubstituted complex.14:19 Taking into
account x-ray and electronic studies mentioned above, this
is most easily assigned to cyanides in apical positions of
a trigonal bipyramid.

Few studies have appeared on the nuclear magnetic
resonance spectra of these complexes. Alyea andMeek15
have reported evidence for the exchange of phosphines on
Ni[C3H5P(CH3)3]3(CN)3 in solution. Upon addition of free
ligand, the methyl doublet resonance shifts upfield toward
that of free ligand, indicating a time-averaging of added
ligand and complexed ligand. In addition, Coskran, et al.13,
reported the same result upon addition of P(OCH3)3 to a
solution which contained Ni[P(OCH3)3]3(CN)2.

Most of the five—coordinate NiL3x3 complexes that have
been isolated have been reported to be reversibly unstable
toward dissociation in solution, according to the following

equilibriumu"20 (Equation 2):

 

NiL3x3 > NiL3x3 + L (2)

<

(Ni(CN)5-3 also dissociates to Ni(CN)4-2).22o23 Electronic
spectra of these complexes exhibit isosbestic points, which
indicate the presence of an equilibrium in solution.12v15v1°v
18'30'23 Indeed, a number of the four-coordinate NiL2X2
complexes have been isolated.”"20 Thermodynamic studies
have been carried out and equilibrium constants reported

for some complexes involved in equilibrium 2.15:35 No

8
studies on the stabilities of the [NiL5]X2 complexes have
been reported.

The relative stability of five-coordinate complexes
toward their corresponding four-coordinate complexes has
been related to both electronic and steric effects.15.19
Gray and coworkers 19 have Shown that good o-donor ligands,
as well as good n-acceptor ligands, should favor five-co-
ordination for Ni(II) complexes with phosphine ligands.

These authors have shown that the relative stability of five—
coordinate complexes of C5H5P(0R)2 and P(OR)3 with nickel
cyanide is determined by a balance of o and w electronic'
effects. They have also suggested that P(C3H5)3 and
(CGH5)2POR do not form five-coordinate complexes with nickel
cyanide, but rather only four—coordinate complexes, because
the bulkiness of these ligands precludes the attainment of
pentacoordination about the Ni(II) atom. Simple oebonding
considerations for these good o-donors would predict the
formation of five—coordinate complexes.19 Alyea and Meek15
report results that suggest the inter-ligand interaction of
(CBH5)3PC3H5 may prevent the formation of a five-coordinate
Complex with nickel cyanide. These authors also found the
tendency to stabilize a five-coordinate complex is less for

a fluorophenylphosphine than for the corresponding, less
bulky, unfluorinated phosphine. However, these latter re-
sults could be due to electronic effects, since the fluoro-
phenylphosphine would not be expected to be as good a o-donor

as the corresponding unfluorinated phosphine. Coskran,

9
gt 213,18 have suggested that steric effects have been overo
emphasized since they have isolated Ni[P(OC6H5)3]3(CN)2
which is stable with respect to dissociation.
With these results, along with the already vast number

of established four—coordinate Ni(II) complexes}

it became
of interest to consider more closely those effects, steric
and electronic, which cause one coordination number over
another. In this study an attempt was made to minimize the
steric factors by utilizing "small"36 monodentate ligands
with the intentions of relieving the ligand-ligand repul-
sions and the restrictive geometry of polydentate ligands
which may sharply influence the coordination number of the
Ni(II) ion. It has been possible to relate the observed
coordination number and geometry of the complex to electronic
effects, i;gfi,the nature of the metal-ligand bond. To ac-
complish this the series of ligands, (CH3)3P, (CH3)2POCH3.
CH3P(0CH3)2, and P(OCH3)3 was chosen for study with
Ni(ClO4)2'6H30, Ni(BF4)2°6H20, and Ni(CN)3. (Those com-
plexes of P(OCH3)3 pertinent to this study have been pre-
pared previously,13v21 and the complexes of (CH3)3P were
reported while this research was in progress.37)

The following five-coordinate complexes have been iso-
lated and investigated in this study: NiL3(CN)2 (L = all of
the ligands in the above series), and [NiL5](ClO4)2 (L =
P(OCH3)21, CH3P(OCH3)2, (CH3)2POCH3). A number of four-
coordinate complexes have also been isolated: gingiL2(CN)2,

(L = (CH3)3PoCH3. CH3P(OCH3)2). trans-Ni[(CH3)3P]3(CN)3.37

and Ni[(CH3)3P]4(BF4)2 .37

10
Relative stabilities of the five—coordinate complexes
toward their corresponding four—coordinate complexes were
investigated by absorption spectral studies. The following

types of equilibria were found to be present in solution

(Equations 3 and 4):

 

NiL3(CN)2 < > NiL3(CN)2 + L (3)
. +2 . +2
NlL5 ;——> NiL4 + L (4)

These equilibria are processes of intermolecular ligand ex-
change, and proton nuclear magnetic resonance temperature
studies were carried out on these systems to determine
relative ligand exchange rates. In this latter study,
similar equilibria (Equation 4) involving the complexes
[CoL31Clo321 and [RhL5]B(C5H5)438 (L = P(0CH3)3) were also
investigated in order to study the effect of the metal on
ligand exchange rates. Both methoxy and methyl ligand
resonances in the proton nmr are simple doublets and thus
facilitated this study. Changes in the proton nmr spectra
were induced by temperature changes and were related to

these exchange processes (Equations 3 and 4).

EXPERIMENTAL

Materials.- Ni(CN)2-4H20 was prepared according to
a method previously given39 and this blue solid was de-
hydrated to buff colored Ni(CN)2 by heating at 1500 under
vacuum for approximately 5 hr. The salts, Ni(ClO4)2-6H20
and Ni(BF4)2-6H30 were obtained from G. Fredrick Smith
Chemical Co., Columbus, Ohio and Alpha Inorganics, Beverly,
Massachusetts, respectively. The tertiary phosphorus
compound, CH3PC12, was received from the Department of the
Army, Edgewood Arsenal, Edgewood, Maryland as a gift and it
was used without further purification. The dehydrating
agent, 2,2-dimethoxypropane was obtained from Eastman Kodak
Co., Rochester, N.Y. Trimethylphosphine was purchased from
Strem Chemicals Im., Danvers, Massachusetts, and converted to
the silver iodide complex‘0 which was stored at about -20°
in a stOppered flask. Trimethylphosphite was purchased
from Aldrich Chemical Company, Inc., Milwaukee, Wisconsin.
The ligand, CH3P(OCH3)2, was prepared according to the method
of Maierfl1 however, CH3PC12 was used as the starting mater-
ial. The observed boiling point (bsoo 60-610) for
CH3P(OCH3)2 agreed well with the literature.41 The ligand,
(CH3)3POCH3, was prepared by heating a 1:1 molar mixture of

(CH3)2PN(CH3)242 and CH3OH at 65° for one hr.and then

11

12
distilling the product at b730 58-590. The reactant,
(CH3)2PN(CH3)3, was prepared by two methods: a reaction
sequence outlined by Burg and Slota,“2 which gave very
poor yields, and by heating (CH3)2PC143 with two moles of
(CH3)3NH in Na-dried diethyl ether at -78° under a nitro-
gen atmosphere. The (CH3)2NH was slowly added through an
addition funnel with a pressure equalizing side arm. The
reaction flask was also equipped with a very efficient
stirrer and a dry ice condenser. After complete addition
of (CH3)3NH, the reaction mixture was allowed to slowly
come to room temperature while magnetically stirred. The
amine hydrochloride was filtered off under suction, and the
product was distilled at 95-1000/760 mm. Reaction of
(CH3)2PC1 or CH3PC12 with CH3OH directly in the presence of
base gave very poor yields of the phosphinite and phos-
phonite, respectively. The 1H nmr spectrum of CH3P(OCH3)2

consists of two doubkfis at 1.13 ppm (CH3, JP = 8.5 Hz)

CH

and at 3.52 ppm (OCH3, JPOCH = 11.1 Hz), and Similarly,

the Spectrum of (CH3)3POCH3 consists of two doublets at

1.19 ppm (CH3, = 6.0 Hz) and at 3.32 ppm (OCH3, J

JPCH POCH

= 13.5 Hz).
The complexes [NiL5](ClO4)221, [C0L51Clo421, and
[RhL5]B[C5H5]438 (L = P(OCH3)3) were prepared according to

previously described methods.

NilCHaP(OCHs )2]3(CN)20- T0 2.2 g (20 [1111101) Of Ni(CN)2

 

suSpended in 50 ml of acetone and cooled in ice water was

13

added 5.4 g (50 mmol) of CH3P(OCH3)2 all at once. The solu-
tion was stirred for approximately 10 hr, filtered under
nitrogen, and concentrated under vacuum to give a dark red
solution. Red-orange crystals formed when the solution was
cooled to -78°. These crystals were filtered, rapidly
washed with -780 acetone, and dried in a stream of dry nitro-
gen.

Aggl, Calcd for C11H27N205P3Ni: C, 30.35: H, 6.22;
P, 21.39; N, 6.44. Found: C, 30.21; H, 6.11; P, 21.48;
N, 6.63.

Ni[CH§P(OCH§)Z]2(CN)2 .- This yellow complex would

 

typically appear as an impurity in the above five-coordinate
complex. Treatment of Ni(CH3P(OCH3)2]3(CN)3 with anhydrous
ethyl ether in a manner analogous to that reported earlier18
produces a yellow powder which is insoluble in acetone and
diethyl ether. Also, if the above mentioned five-coordinate
complex is maintained under a dynamic vacuum for an hour or
more, conversion to Ni[CH3P(OCH3)2]2(CN)2 is apparent by
formation of the yellow solid.

Aggl, Calcd for C3H13N304P2Ni: C, 29.35; H, 5.52;
P, 18.95; N, 8.57. Found: 0, 28.95; H, 5.17; P, 18.95;

N. 8.50.

Ni((CH3)3PoCH3]3(CN)3.- To 1.1 g (10 mmol) of Ni(CN)3

 

suspended in 50 ml of acetone and cooled in ice water was
added 2.8 g (30 mmol) of (CH3)2POCH3 all at once under a

nitrogen atmosphere and the mixture was stirred magnetically.

14
The solution was allowed to stir for several hr, and then
it was filtered, concentrated under vacuum, and cooled to
give red-orange crystals. Attempts at recrystallization in
the presence of excess ligand yielded mostly the four-coor-
dinate complex and repeated attempts at obtaining a satis—
factory elemental analysis were unsuccessful. This complex
could, however, be identified by its infrared spectrum in
the solid state and further characterized by its absorption

spectrum, also in the solid state.

Ni[(CH3)3POCH3]3(CN)2.- As mentioned above, this

 

yellow complex consistently appeared as an impurity in the
five-coordinate complex and could be isolated from such solu-
tions. Also, by allowing Ni(CN); to react with (CH3)2POCH3
in a molar ratio of 1:2 or less and in a manner similar to
the above, this complex could be prepared easily.

A32}; Calcd for C8H18N202P2Ni: C, 32.61; H, 6.13;
P, 21.08; N, 9.53. Found: C, 33.04; H, 6.11; P, 19.78;

N, 9.69.

Ni[(CH3)3P]3(CN)2.- A 50 ml acetone suspension of

 

0.44 g (4.0 mmol) of Ni(CN)3 was evacuated on a high vacuum
line and cooled with liquid nitrogen. The ligand, P(CH3)3.
0.91 g (12 mmol), liberated from the A91 complex by heating,
was condensed onto this frozen suspension and the mixture
was then allowed to warm to room temperature. Stirring
(magnetic) was initiated as soon as possible and a dark red

solution resulted. After the solution had stirred for about

15

1/4 hr, it was removed from the vacuum line, filtered, con-
centrated, and cooled to -780 to give red crystals. Alter-
natively, the acetone solution may be evaporated to dryness
and the resultant red-orange powder recrystallized from g:
hexane or diethyl ether.

A321, Calcd for C11H37N3P3Ni: C, 38.95; H, 7.97;
P, 27.41: N, 8.26. Found: C, 38.53: H, 8.02: P, 26.59;
N, 8.12.

Ni[(CH§)3P]3(CN)3.- Prolonged evacuation of
Ni[(CH3)3P]3(CN)3 would produce this yellow solid which

could be recrystallized from an acetone-diethyl ether mix-
ture (10:1).

5&2}, Calcd for C8H13N2P3Ni: C, 36.50; H, 6.85;
P, 23.60; N, 10.65. Found: C, 36.39: H, 6.88; P, 23.41;
N, 10.43.

[Ni[CH3P(OCH3)a]§)(ClO!)2.- In an acetone-2,2-di-

methoxypropane“ solution (10 ml of each) was dissolved

0.73 g (2.0 mmol) of Ni(ClO4)3‘6H20. This solution was
stirred at room temperature fcr1/2 hr and then 1.25 ml

(> 10 mmol) of CH3P(OCH3)3 was added all at once to give

an orange precipitate and a similarly colored solution.

The solution was stirred for 1/4 hr, filtered, and the solid
was washed well with diethyl ether and dried under vacuum.
Recrystallization can be accomplished from a very concen-

trated acetone solution.

16
5331, Calcd for C15H45C12013P5Ni: C, 22.55: H, 5.64;
P, 19.42; Cl, 8.90. Found: C, 22.42: H, 5.63; P, 19.06;
Cl, 8.94.

[Ni[(CH3)3POCH3]5](C104)2.- This compound was pre-

 

pared in a manner identical to that listed above for
{Ni [CH3P (OCH3 )3] 3 I (ClO4I3 .

Anal. Calcd for C15H45C13013P3Ni: C, 25.08: H, 6.27:
P, 21.55: Cl, 9.88. Found: C, 25.27: H, 6.19; P, 21.47;
Cl, 10.08.

{Ni[(CH3)3P]4}(BF4)3.- The air over a 25 ml solution

 

of absolute ethanol and 0.34 g (1.0 mmol) of Ni(BF4)3-6H30
was evacuated on a high vacuum line and the solution cooled
with liquid nitrogen. Trimethylphosphine (0.30 g, 4.0
mmol) was added to this solution in a manner as indicated
above. A pale red solid formed as this mixture was allowed
to warm to room temperature and stirred magnetically. The
solution was removed from the vacuum line, filtered, and the
solid washed well with diethyl ether and dried under vacuum.
The complex dissolves in ethanol only very slowly to give a
yellow solution, in acetonitrile to give a purple solution
which eventually becomes colorless upon standing, and in
acetone to give a red solution which rapidly decomposes to
give a nearly colorless solution.

523;, Calcd for C13H3533P8P4Ni: C, 26.80; H, 6.70;
P, 23.10. Found: C, 26.65; H, 6.56; P, 22.98.

17

Physical Measurements.- Elemental analyses were per-
formed by Galbraith Laboratories, Inc., Knoxville, Tennessee.
Infrared spectra were obtained in Nujol mulls by use of
either a Beckman Model 12 grating Spectrophotometer or a
Perkin-Elmer Model 225 grating Spectrophotometer. Reflec-
tance spectra were obtained by use of a Bausch and Lomb
Spectronic 600 with BaSO, as the diluent and reference.
Spectra in the solid state were also obtained by use of the
Cary Model 14 spectrophotometer and a Nujol mull of the com-
pound which was “painted” onto filter paper.‘:5 A reference
cell with Nujol only was needed. These "filter paper cells“
were simply taped to the wall of the cell compartment
nearest the detector; however, little difference in the
spectra were noted when the Opposite wall was used. Spectra
were also recorded by use of Nujol mulls mounted between
quartz windows. In the visible region, almost identical
Spectra were obtained by the two methods, however, in the
ultraviolet, the resolution was very poor when quartz win—
dows were used. Solution ultraviolet-visible spectra were
recorded in CH3C12 by means of Unicam Model SP.800B and Cary
Model 14 SpectrOphotometers. The CH2C13 was previously
dried by refluxing over CaHg. Samples were prepared under
nitrogen in ground glass stOppered cells in order to mini—
mize concentration changes while the Spectra were recorded
at various temperatures. A thermostated methanoldwater bath,
reproducible to i 0.10°C, was attached to the Spectropho-

tometer; thus, the samples could be recorded at different

18
temperatures. Proton nmr spectra were recorded by use of a
Varian Model A56/60D analytical Spectrometer, equipped with
a Varian Model V-6040 variable temperature controller. Low
temperatures were determined by measuring the chemical Shift
difference for methanol, while ethylene glycol was used for
high temperatures. All of the Spectra were recorded in
CH3C12 solutions except the high temperature Spectrum of
{Rh[P(OCH3)3LdB(C3H5)4 which was recorded in CH3CN: both
solvents were used after being dried over CaH2. The high
temperature Spectra were taken in thickdwalled nmr tubes
which were sealed under an atmosphere of N3. Chemical Shifts_
were measured by using the solvents as references (CH3C12,
6 = 5.30 ppm: CH3CN, 6 I 2.00).46 Magnetic susceptibilities

‘were determined by Evans' nmr method47 in CH3C13 solutions.

RESULTS AND DISCUSSION

Five-coordinate complexes of the type NiL3(CN)2(L =
(CH3)3P,37 (CH3)3PoCH3, CH3P(OCH3)3, P(oCH3)3)18 are
produced when Ni(CN)3 is allowed to react with an excess
of the above ligands in acetone solution. The corresponding
NiL3(CN)3 complexes are prepared by subjecting the five-
coordinate complexes to high vacuum at room temperature, or
by allowing Ni(CN)3 to react with these ligands in a 1:2
molar ratio. Complexes of the type [NiL5](C104)3 (L =
(CH3)3POCH3, CH3P(OCH3)2, P(OCH3)3)21 are isolated from
acetone-2,2—dimethoxypropane solutions of Ni(ClO4)2-6H20 and
excess ligand. When L = (CH3)3P no five-coordinate complex
could be isolated. However, the complex, Ni[(CH3)3P]4(BF4)2.
was prepared when Ni(BF4)2'6H30 was allowed to react with
(CH3)3P in ethanol solutions, similar in manner to the prep-

aration Of Nil (CH3 )3P] 4 (C104 )2 .37
Infrared Spectra

Complexes of Ni(CN)3 which were prepared in this study
permit easy assignment of the cyanide stretching frequency.
The mull infrared Spectrum of each NiL3(CN)2 (L = (CH3)3P,
(CH3)3P0CH3, CH3P(oCH3)3, P(0CH3)3) complex consists of one
Sharp absorption (Table II), which is consistent with a
trigonal bipyramidal structure with the cyanide groups in

19

20

 

 

 

 

 

 

 

 

 

 

Table II. Infrared data, cyanide stretching frequencies.
L vCN(cm_1) Structural Assignment
NiL3(CN)3
(CH3)3P 2097 L CN
(CH3)3P0CH3 2100
h“
CH3P(oCH3 )2 2104
CN
P(OCH3)3 2125
N1L.(CN)2 L {N
(CH3 )3P 2106 b N j
NC 3
(CH3)3POCH3 2129, 2113 L JN
CH3P(0CH3)2 2122, 2103 Ni
L CN
a
P(OCH3)3 2149, 2144, (Figure 4)
2127, 2122

 

aSee reference 18.

21

E£2n§_apical positions. Infrared Spectral data in the cyanide
region for the corresponding NiL3(CN)3 (L = all ligands in

the above series) complexes is also presented in Table 11.
When L =(CH3)2POCH3 or CH3P(OCH3)3, two cyanide stretching
frequencies are observed, and when L = (CH3)3P a Single
cyanide stretching frequency is observed. These data are
interpreted in terms of a Sigrsquareoplanar geometry for

the complexes of (CH3)3POCH3 and CH3P(OCH3)3, and a Egaggf
Square—planar arrangement for the complex of (CH3)3P. When

L = P(OCH3 )3’ four cyanide stretching frequencies are observed.
Coskran, g£_§£.13 have postulated that a dimer is present

in the solid state, which contains both bridging and terminal

cyanides (Figure 4).

(CN)
NC CN

(C N)

Figure 4. Proposed dimeric structure of
Nil.3(CN)3 (L = P(0CH3)3).18

Absorption Spectra

Ni(CN); Complexes.- The band maxima in the absorption

 

Spectra for both the four- and five-coordinate Ni(CN)2 com-

;pleXes are listed in Table III. These Spectra are recorded

22

.wooasonm u Anmv

 

Asmvnms.omm
Asmvmmc.omm
Armvcmc.msm
Asmvoms.nsm
Asmvmfls.mmm
Armvmec.mmm
Armvomc.mnm

Aruvcue.mmm

Anmvoem.flsmyocm.oom
Asmvnmc.cmm.firmvccm.oeu
Acmvcsc.Armvmmm.mmm
Acnvcms.msm.sem
Armvmes.flsnvmmm.ceu
Asmvmes.mmm.sem
Asmvcue.xsmvnsm.ccu
Acmvcms.mmm.eem

msmtmem

Accounvmmm
Asmvcsvxfirmvmmmtncm
Acuvose.flsmvomm
Asmvcms.xsmvmmm.ncm
Acmvcms.xrmvmmm.mmm
Asmvcmc.firnvmsm.ocu
Accounvmcm.ocm

«Azove_mmAcmovaaz
«AzovaOmAmmuv_Hz
«Azova_cmoom«Ammovahz
«AzovmmemoomaflmmOVHAz
«Azova_aflmmoovmcmo_az
«Azovm_ahmmoovmsmo_az
«Azovamafimmoovm_az

«Azovmflmxmmoovm_az

 

dwcmmfla
mmooxm .uHOnmUV
Adev mCOADQHomnfl

Issuance
Aifiv nCOADQHOmA<

Acumum owaomv
Aiev mcowumuomnm

ocsomaoo

 

 

.moxoamaoo «Azuvwz Mom sumo Hmnuoomm cowumuomné

.HHH magma

23

in both solid and CH2C13 solution media for each complex,
along with the effect of adding excess ligand.

For each five-coordinate complex in solution, there is
a very intense absorption in the region 290-300 mu. When
excess ligand is added to these solutions (and also to the
reference cell), the absorption is not present (Table III).
This Spectroscopic feature is demonstrated in Figure 5 for
CH3C13 solutions of Ni[P(OCH3)3]3(CN)3, which are typical
of the five-coordinate complexes. The presence of this ab-
sorption (290-300 mu) in the solution spectra of the four-
coordinate complexes, and the absence of this absorption in
the solid Spectra and solution Spectra with excess ligand of
the five-coordinate complexes, suggests that the band is due
to the four-coordinate complexes (Figure 6). The rather
constant position of this band in all the Spectra of the
four-coordinate complexes suggests that it is a Ni —> n*CN
charge-transfer band.19 These experimental results suggest
that an equilibrium of the following type is present in

solutions of the five-coordinate complexes:

 

> NiL2(CN)2 + L (3)

NiL3(CN)2 <__—.
This equilibrium is also consistent with the experimental
Observation that four-coordinate complexes can be easily
isolated from solutions of the five-coordinate complexes.
The intense absorption in the visible region, with its
low energy shoulder (325-455 mu), that appear in both the

solid and solution spectra of the five—coordinate complexes,

is assigned to the (e“)4(e')4 -—> (e")4(e')3(a1)1

 

24

.ocmmfla mmmuxm £Dfl3 Amv CH coausaom Anv

.aHORmo an «Azovmmmfimmoovaaez

Amv «mo Esuuommm manwmfl>lu0H0H>mHuHD

.m onsmam

25

onv

obs

.m musmflm

38:35.20;
0mm nmm com

nun

od

3

8

.33
cauoqaosqv

E!

 

 

26

wumum oflaom 0V .«HONTO Gd «Azovu

.mHONmo as «Azovmmmm mmoVHRZ

.«mzovaOnmmmUvgaz
anlsmev Hz Ant
Amv «m0 Esuuoomm cofiumuomnm

.o ouomfim

III}: 1

27

 

.m musmfim

355923.03
0mm

 

aauoqlosqv

28

transition in complexes of D3h symmetry3 (Figure 2). Daw—
son, g£_gl,,33 have discussed the splitting of this band
(p. 5), and their argument implies that this band Should
become more symmetrical at lower temperatures (~a77°K). A
dipole forbidden transition, (e")4(e')4 -+ (e")3(e')4(a1)1,
v2, of low intensity is expected at shorter wavelength than
v1 (Figure 2), but overlapping of intense ultraviolet ab-
sorptions and the above allowed absorptions prohibits accur-
ate assignment of its position. These assignments are in
agreement with Similar assignments which have been made for
other five-coordinate Ni(II) complexes of tertiary phosphorus,
ligands.3

The band assignments that are proposed above assume
that these five-coordinate complexes have trigonal bipyra-

midal structures (D in the solid state and in solution.

3h)
However, it may be argued that these Spectra are those of a
square pyramidal complex (C4v)' Since three spin-allowed
transitions are predicted for these complexes34 (Figure 3).
The proof of this argument would rest in the low temperature
absorption Spectra, as these three bands in C4v symmetry
are independent of temperature, and the two low energy

bands (325-455 mu) would persist at low temperature (~a77°K),
contrary to the predictions for these bands in D3h sym—
metry. At this time, the geometry of these five-coordinate
complexes cannot be unequivocally assigned, but the band

assignments have been made on the basis of the trigonal bi-

pyramidal structure, because it is predominant in Observed

29

complexes of Ni(II) with monodentate ligands“’"‘29 and it is
predicted by ligand-ligand repulsion theories.25

When the absorption spectrum of each five-coordinate
complex in solution is recorded over a narrow temperature
range (0-30°C), isosbestic points are observed. This sup-
ports the proposed equilibrium (Equation 3), since these
isosbestic points indicate an equilibrium.between two ab-
sorbing Species. Figure 7 illustrates the isosbestic point
obtained for Ni[P(0CH3)3]3(CN)2 and is typical of these
five-coordinate complexes.l It can also be noted (Figure 7)
that the low energy bands due to the five-coordinate com—
plexes (325-455 mu) increase in intensity as the temperature
is allowed to decrease, while the bands assigned to four-
coordinate complex (290-300 mu) decrease in intensity. This
latter observation indicates the proposed equilibrium
(Equation 3) shifts toward five-coordinate complex upon a
decrease in temperature.

Estimates of the equilibrium constants, or instability
), for the system

Kinst
NiL3(CN)2 ———> NiL3(CN)3 + L (3)

<

constants (Kinst

have been obtained for these five-coordinate complexes.
{rhese calculations were facilitated by the spread in the
loand maxima positions for the four- and five-coordinate
«complexes. The extinction coefficient (8) for the intense
.absorption in the visible region (the high energy band of
‘VI was used ) for each five-coordinate complex was deter-

rnined from the apparent absorption maxima observed when

 

30

aaOamo ca «AZOVSHSAmmoovm_az
mo Ednuommm oaflawa>uu0a0fl>muuas may no Omsmno musumnomaou mo Doommm

.b oudmflm

31

com

ems

oov

.5 madman
3813:3303

 

 

 

cmn can con man
. . Qo
a
-vo
rw.° UV!
m
Q.
I m D
. m m
.m .u.—
s I
00
ea _
.0;
.oo .

 

Q
N

32
extremely large excesses of ligand were added to solutions
of the five-coordinate complexes of known concentration.
These values of s were then used to calculate the concen-
tration of five-coordinate complex present at equilibrium
with no added ligand and the difference between this concen—
tration and the initial concentration was taken as the four-
coordinate complex concentration and the dissociated ligand
concentration, also at equilibrium. The calculated equi—
librium constants (or instability constants) are recorded in
Table IV along with the calculated e's. The uncertainty in
these values is such that (CH3)3POCH3 and CH3P(OCH3)2 may be
interchanged but it is definite that these constants follow
the order (CH3)3P > P(oCH3)3 > CH3P(OCH3)3 ~ (CH3)3P0CH3.

Surprisingly, there is no regular order in K.

inst paral-

lelling the systematic change in the nature of these phos-
phorus derivatives. This particular order, however, is
observed elsewhere. The band maxima (Table III) for the
allowed d-d transitions (see later) are in approximately the
same order indicating that the P(CH3)3 complex has the
'weakest ligand field interaction. The large extinction
coefficients that are obtained for these five-coordinate
complexes are similar to previous observations,3o15 and dem—
onstrates the strong mixing of ligand and metal orbitals in

these complexes.

33

 

 

 

 

Table IV. Kinst for NiL3(CN)2 complexes at 25°.
4
L Kinst X 10 -5 -1
(M) (mole cm )

(CH3)3P 2.5 2390
(CH3)3P0CH3 0.4 2480
CH3P(ocH3)2 0.5 2506
P(0CH3)3 1.7 2756
Ni(ClO!)2 and Ni(BF$)£_ComplexeS.- The band maxima

 

in the solid state and solution spectra of the five-coordinate
[NiL3](Clo,)3 (L = (CH3)3P0CH3, CH3P(0CH3)3, P(0CH3)321)
complexes and the fOur-coordinate Ni((CH3)3P]4(BF4)2 com-
plex are presented in Table V. The ultraviolet-visible Spec-
tra of the five-coordinate complexes in the solid state and in
solution to which ligand has been added (see later) are simi-
lar to those reported for {Ni[P(oc::H)3(CH3)3]3}(Clo,)3,21:48
*which is trigonal bipyramidal, according to a three-dimen-
sional X-ray crystal structure analysis.29 Assignment of

the bands in the five-coordinate complexes is accomplished
xvith the assumption of a similar structure, and are similar
-to the assignments in the NiL3(CN)3 complexes. The low
(energy absorption bands in the spectra of these complexes
(400-560 mu), assigned to the (e")‘(e')4 -9 (e")4(e')3(a1)1,
‘v1 transition (D3h symmetry) (Figure 2), are Split in the
solid state and solution with added ligand33 (p. 5). The

Imigh energy band (300—320 mu) is assigned to the dipole

34

.uooasorm u Anne

 

Isucomc.cemiasuvccm
Ismccom.ccc.xsnccam
Arncmmm.nse.xsmcocm

Armcocv.mac.xraeocm

oos.AsmCoov.omu cos.ocm
AccounCmNe.Aancmm freeman.osv.1smcomm
Apaches.c~s.xsnvofim Aomomncmme.AsuCoum

Igneous.sov.Araooom Igneous.eoe.xrmcmcm

al.macm.naaiasoc_azc
«AeoaovmsnamooaaxcmocLazw
«AcoHoCHanaxamoocmmmo_azw

«AeoHoCMSHSAamooCSLAzL

 

Aocmowa
mmmoxm .«HOamUv
Ajay meowumHOmnE

Auavumuv Aoumum paaomv
Aiev msoaumnomnd Anew mcoaumnomnd

ocsomaoo

 

 

.moxoamsoo «Avmmvaz osm «AvoHovaz How sumo Hmuuoomm documuomnd .> wanna

35

forbidden, (e")4(e')4 -—> (e")3(e')4(a1)1, v2, transition.

Isosbestic points are obtained in the solution spectra
of {Ni[P(OCH3)3]5}(ClO4)2, when the spectra are recorded
over a small temperature range (Figure 8). Similar isos-
bestic points are obtained at 362 mu and 424 mu for
{NilCH3P(OCH3)2]5](C104)2. (The complex
{Ni[(CH3)3POCH3]5}(C104)3 decomposes rapidly in CH2C12 and
did not allow for the observance of isosbestic points.)
Therefore, an equilibrium Similar to that found in the

NiL3(CN)2 complexes is proposed:

 

___> mun” + L (4)

<

[NiL5] +2

Further evidence for this equilibrium is given by the ob-
servation that the solution Spectra of the five-coordinate
complexes are altered upon addition of ligand. The posi-
tions of the low energy bands shift to lower energies when
ligand is added, approximate the pOSitions of the absorptions
in the solid state Spectra (Table V), and indicate that the
four- and five-coordinate bands overlap to a large extent
in the solution spectra. Indeed, the solution spectrum of
[Ni[(CH3)2POCH3]5}(ClO4)3 more closely resembles the spec-
trum of the four-coordinate (Ni[(CH3)3P]4](BF4)3 complex
and the five-coordinate Species is probably largely dissoci-
ated in solution.

Unlike the corresponding NiL3(CN)3 complexes, the
[JN'in]+2 complexes have not been isolated. Their existence,

'though, is tenable in view of the fact that (CH3)3P does

 

36

.mHOnmo ca «AvoHovmonmAnmoovasz now omcfimuno mucflom oaumonmomH

.w ousmflm

37

.w musmam
as; face—20>)

00m. onv ooh om” mum com

  

tron 0:339:

'21:?

--—'-—_.— ,

”M"?
0

r0.

8....

6A

 

.__I.
0.
o

0!
O

O

aauoqlosqv

38

form an isolable NiL4+2 complex, and this complex will as-
sociate with added (CH3)3P in solution to form a five-co-
ordinate complex. The solution and solid state spectra of
this four-coordinate complex are similar (Table V), and
similar to other square-planar, Ni(II) complexes.3‘ How-
ever, when excess (CH3)3P is added to a solution of
(Ni[(CH3)3P]4](ClO4)2, a dark blue color results, and the
Spectrum of this solution resembles those of the five-co- (
ordinate complexes in this work (Table V), and also that I
of the blue complex,[Ni(TAP)Br]ClO4 (TAP = P(CH3CH2AS(CH3)3)3).49

It is worthwhile to comment further on the associative
nature of Ni[(CH3)3P]42+ in solution. When the tetrafluoro-i
borate salt of this complex is dissolved in CH3CN, a dark
purple solution results. Addition of a large amount of
diethyl ether to this solution causes trace amounts of a
blue solid to be precipitated. This solid exhibits a VCN
at ~v2300 cm_1, indicative of a coordinated CH3CN group,50
which is not surprising, in view of the associative
character of this complex. The proposed five-coordinate
complex present in solution when excess (CH3)3P is added to
[Ni[(CH3)3P]4]2+ has not been isolated, however. The dark
blue solution that forms when excess ligand is added rapidly
fades to pale yellow and addition of diethyl ether to this
solution causes precipitation of a pale yellow solid. More-
over, when diethyl ether is added to the initial blue solu-
tion, the same yellow solid is obtained. Attempts at

recrystallization inevitably produced light green solutions

39
indicative of solvated Ni(II). Elemental analyses of the
compound were inconsistent and offered no information about
it. Nevertheless it is tempting to speculate that it is a
Ni(O) complex of (CH3)3P. The complex, [Ni((CH3)3P]4}, has
been prepared and its color is light yellow}6 The mecha-
nism for its formation in this work might be considered to
be similar to that used for preparation of Ni(O) phosphite
complexes!“52 in which excess ligand causes reduction of
Ni(II) to Ni(0) and the excess ligand is oxidized.

Reference to Table V, shows that in the five-coordinate
Ni(ClO4)3 and Ni(BF4)3 complexes, the order of increasing 3
ligand field strength is (CH3)3P < (CH3)3P0CH3 < CH3P(0CH3)3
< P(0CH3), This order is Obtained from those spectra to
which excess ligand has been added. This order of ligand
and field splitting closely parallels the stability of the
five-coordinate complexes. Thus, the five-coordinate com-
plex with (CH3)3P is unstable and cannot be isolated from
solution, while the isolable P(0CH3)3 complex appears to be
stable in solution and its absOrption spectrum is affected
very little upon the addition of ligand. Therefore, it is
proposed that the above order of ligands reflects the stabil-
ity of the five-coordinate [NiL3]+2 complexes toward the

corresponding [NiL4]+2 complexes.

Electronic Effects in Five-Coordinate Complexes.- In

five-coordinate, d3, complexes with trigonal bipyramidal

40

structures four of the filled metal d orbitals can be used
in w-bonding, if 7- acceptor ligands are available.19 In
square-planar, d3, complexes only three d orbitals can be
used in wébonding, for reasons of symmetry.19 Thus, a good w—
acceptor ligand would favor the formation of a five-coordin-
ate complex over a four-coordinate complex. In the ligands
considered in this study, the w—acceptor ability would be
expected to decrease in the order P(0CH3)3> CH3P(OCH3)2 >
(CH3)3POCH3 > (CH3)3P.53 The more electrOnegative OCH3
groups would have the effect of lowering the energy of the
empty phosphorus 3d orbitals, thus enhancing their #-
acceptor character.5‘ At the same time the increased elec—
tronegativity of OCH3 groups should withdraw electron
density from the phosphorus atom, and produce a contracted
and stable o-donor orbital, which would be less suited for
otbonding to the relatively more expanded Ni 43 and 4p
orbitals.1’ Since ligands which bond to these S and p
orbitals in preference to d orbitals favor the five-co-
ordinate structure, the tendency toward five—coordination
Should decrease in the order (CH3)3P > (CH3)2POCH3 >
CH3P(OCH3)3 > P(0CH3)3, according to oébonding arguments?“19

.The order of ligand V-acceptor ability follows the
order of proposed stability of the [NiL3]+2 complexes rela-
tive to [NiL3]+2 complexes. An opposite trend is predicted

by the above oébonding arguments. This suggests that #-

bonding effects are of primary importance in determining

the relative stability of these five-coordinate complexes.

41
Support for this hypothesis is given by the observation that
attempts to isolate the complex, {Ni[(CH3)3P]5}(BF4)3 (p. 39),
result in the apparent reduction of Ni(II) to Ni(O). This
suggests that the increased metal electron density could
not be sufficiently relieved by these poorer v-acceptor
ligands.5‘

The stability of the NiL3(CN)2 complexes toward the
correSponding NiL2(CN)3 complexes decreases in the order
(CH3)3P0CH3 ~ CH3P(0CH3)3 > P(0CH3)3 > (CH3)3P (Table IV).
This does not reflect the Order predicted by either n-ac-
ceptor or o-donor arguments, but suggests that the stability
of five-coordinate complexes is determined by a more equal
combination of these two effects. This is consistent with
observations that suggest the CN- anion is a good n-ac-
ceptor.20 Thus, the cyanide groups reduce the electron
density on the metal and the ligand n-acceptor ability
becomes less important in these NiL3(CN)2 complexes than in

the [NiL5]+2 complexes.
Proton Nmr Spectra

Three classes of complexes were studied, namely, the
four-coordinate complex, Ni[(CH3)3P]2(CN)2, the five-coord-
inate complexes, NiL3 (CN)2 (L = P(0CH3)3, CH3P(OCH3)2.

(CH3)2POCH3, (CH3)3P), and the five-coordinate complexes

ML5n+ (Mn+ = Ni“, L = P(0CH3)3 and CH3P(0CH3)3; M1M =

21'
Co+ and Rh+ 38, L

P(0CH3)3). The complexes, NiL2(CN)2

42
(L = (CH3)3P0CH3, CH3P(0CH3)3. P(0CH3 )3), {Ni[(CH3)3PoCH3]3)-
(€104)3 and {Ni[(CH3)3P]4}(BF4)3 were not studied here be-
Cause of insolubility and instability in solution. The
Co(I) and Rh(I) complexes were included in this study be-
cause they permit comparisons of the effect of the metal on
ligand exchange rates. The temperature dependent proton
nmr spectra obtained for all of these complexes are similar
over their temperature ranges. For each complex there are
three distinct Spectral features that occur at Specific
temperatures. These spectral features are: doublet (high
temperature Spectrum), sharp singlet (intermediate tempera-
ture Spectrum), and multiplet (low temperature Spectrum).

These classes of complexes will be discussed in turn.

Ni[(CH3)3P]3(CN)3.- The temperature dependent proton

 

nmr spectrum of this complex is Shown in Figure 9. At
probe temperature (~'38°) the Spectrum is a sharp Singlet
at 6 = 1.52 ppm. If the solution is cooled, this singlet
broadens and collapses and finally at —65° a three-line
pattern is observed. Alternatively, if the solution is
heated from the point of the Singlet, this resonance is
observed to broaden and finally a sharp doublet forms at 97°
with a coupling constant of about 10 Hz and the chemical
shift is slightly downfield from the singlet.

The general features of these Spectra are identical to
those obtained for some Eggggfsquare-planar Pd(II)55 and
Rh(I)56 methylphosphine complexes. In these systems the

room temperature proton nmr spectra of the pure complexes

43

Figure 9. 1H nmr spectrum of Ni[(CH3)3P]2(CN)2 in
CH2C12 at various temperatures..

 

 

44

a.»

44°

 

 

 

-56°
1
F If 1| PPM
1.70 1.50 1.30

Figure 9.

 

45
were virtually coupled triplets and upon successive addi-
tions of ligand they collapsed forming singlets, and finally,
doublets formed at high ratios of added ligand to complex.
An intermolecular process of ligand exchange between added
ligand and complex was prOposed to account for these Spec-
tral changes. The similarities between these spectra
and those observed in this work for Ni[(CH3)3P]2(CN)3
are obvious and we Similarly propose that a process of
intermolecular ligand exchange is occurring. However, there
is one important difference in the system studied in this
work and that of the Pd and Rh complexes and that is that
no excess ligand has been added to solutions of the nickel
complex. Therefore, this four~coordinate nickel complex
must be dissociatively unstable with respect to free ligand
and a lower coordinated Species. The absorption spectrum
of a methylene dichloride solution of Ni[(CH3)3P]3(CN)3 pro-
duces an isosbestic point at 345 mu.when the Spectrum is
recorded over a relatively small temperature range. (Fig-
ure 10.) Thus, there must be more than one absorbing
Species in solution and this is in agreement with a dis-
sociative process occurring in solution for which the fol-

lowing equilibrium is proposed:

 

NiL3(CN)3 <___:_ 1~1iL(cn)2 + L, L = (CH3)3P (5)

This equilibrium similarly accounts for the intermolecular
ligand exchange process. The three-coordinate nickel com-

plex in the above equilibrium can be classed as a fourteen

46

. .u «
HO mu
CH «Azovmmmmmmmovafiz mo Esnuoomm wanwmfl>luoaofl>muuas mo coauuom

.OH oudmflh

47

.oH ousowm

AiEv faced 962,

can

nun

 

 

.

0.0

J O...

 

oauoqlosqv

48
electron complex (d8 metal ion and three two-electron donor

ligands). Such systems are not at all common, but there is

no other Obvious recourse in the interpretation of the data.
One might speculate, however, that the three-coordinate
complex dimerizes through cyanide bridging forming
[L(CN)Ni(CN)3Ni(CN)L] . In this complex there would be a
square planar arrangement about each four-coordinate Ni.
Bridging cyanide groups are, of course, well known in some
Ni(II) systems.57

Other equilibria that may be considered to be present
in this solution do not account for the observed 1H nmr and
absorption Spectra. For example, the complex Ni[P(OCH3)3]3-
(CN)2, has been prOposed to dimerize through cyanide bridg-
ing.18 A Similar monomer-dimer equilibrium for
Ni[(CH3)3P]2(CN)2 would not involve ligand exchange and
thus, could not explain the 1H nmr Spectra. Also, it is
known that the four-coordinate complex, Ni[(CH3)3P]3(CN)2,
will associate with ligand to form a five-coOrdinate complex
(see later), however, this equilibrium is not present in
solutions of the pure four-coordinate complex because a
characteristic isosbestic point is observed at 335 mu for
the four-coordinate - five-coordinate equilibrium (p. 29)
and only after additional ligand is added toIsolutions of
the four-coordinate complex is this isosbestic point observed.

Therefore, the three Spectral features which are ob-
served, doublet, singlet, and multiplet (triplet) are at—

tributed to an intermolecular ligand exchange process and

49
they represent the Spectra for processes under fast, inter—
mediate, and slow rates of exchange, respectively. It will
be convenient to delay discussing the doublet and singlet
Spectra until after the results of the five-coordinate com-
plexes are presented; but the triplet,or Slow exchange Spec-
trum.will be discussed now.

Numerous examples of Egaggfsquare-planar d8 metal com-
plexes58 of organOphosphorus ligands have exhibited proton
nmr Spectra Similar to the triplet spectrum observed for
Ni[P(CH3)3]2(CN)2 at -65°. This is an example of an
XnAA'Xn' type spectrum where X and A are the H and P
of one ligand respectively, X' and A' are the H and P
of the second ligand, and n = 9. The separation of two

outer lines of the triplet equals |J ,| and the

AX+JAX
central line, which will vary in intensity from complex to
complex, can be qualitatively used as a measure of JAA.:59'5°
the more intense the central line, the larger qAA" There-
fore, the triplet spectrum for this trans—square-planar
nickel complex with an intense central resonance means that
the two ligands are strongly coupled to each other through
the metal. For a complex exhibiting ligand exchange this
coupling could only occur under the limiting conditions of
very slow ligand exchange. Therefore, the triplet spectrum
is representative of the ligand-ligand interaction under

conditions of Slow exchange for this trans-squareoplanar

arrangement of ligands about the metal atom.

50

NiL3(CN)z Complexes.— The five-coordinate Ni(CN)2
complexes of (CH3)XP(OCH3)3_i (x = 0, l, 2, or 3), which
are trigonal bipyramidal with p£3p§_apical cyanides,18
exhibit very Similar temperature dependent 1H nmr Spectra
to the four-coordinate complex, Ni[(CH3)3P]3(CN)3 (see
Figure 11). For these complexes, hOwever, the low tempera-
ture multiplet is a poorly resolved quartet contrasted to
the triplet observed for the four-coordinate complex. For
the complexes of CH3P(OCH3)2 and (CH3)2POCH3, two resonances
are observed for the two types of ligand protons in each
complex (Figure 12). Therefore, two doublets and two sing-
lets are observed for each complex as the temperature of the
solution is lowered: however, only for the OCH3 protons is
the quartet Splitting observed while for the CH3 protons
only a broad Singlet is recorded. The failure to resolve
a quartet is surprising because the OCH3 proton Splitting is
observed in the same ligand and furthermore, the CH3 proton
splitting is readily observed in the five—coordinate (CH3)3P
complex (see Figure 11).

To account for these temperature dependent Spectra we
again propose a process of intermolecular ligand exchange
which is consistent with the postulate that all of these
complexes are dissociatively unstable according to the pro-

posed equilibrium:

NiL3(CN)2 :3 NiL3(CN)2 + L (3)

51

Figure 11. 1H nmr Spectrum of Ni[(CH3)3P]3(CN)2 in

CH2C12 at various temperatures..

52

42°

—-24°

A _390
N0

L7 1 l 1 l

1.60 1.40 1.20 ppm

Figure 11.

 

53

ANA «moovmmmo .mmoomfimmov

..monsumummEmu msoflnm> um uHUnmo CH

u AV «Azuvmqflz mo Eduuoomm Had we

.Na musmflm

54

«IU

omv-

«Amzuornxu H.

«.80

.NH mHsmHm
omd

 

 

‘-“

 

55
The poorly resolved quartet observed at low temperatures
represents a spectrum of the type, AA'AFXhXh‘Xh", where A
and x have similar notations as before. This assignment
is based primarily on the similarities in the overall Spec—
tral changes for these five-coordinate complexes and the
previously mentioned four-coordinate complex. In the latter
case, the triplet Spectrum due to ligand-ligand interaction
through the metal is well understood"9 and:fl:is believed that
under similar limiting conditions of Slow exchange, coupling
through the metal between the three ligands in the equatorial
plane does take place for these trigonal bipyramidal com-
plexes and gives rise to the quartet spectrum.

A simple experiment to demonstrate the intermolecular
nature of this exchange process is summarized in Figure 13.
If P(0CH3)3 is added to the five-coordinate complex,
Ni[P(OCH3)3]3(CN)2, at 54° only one doublet is observed
which is Slightly upfield from the original doublet for the
complex only. If this solution is cooled, the doublet col-
lapses, a sharp singlet forms, and eventually a quartet forms
similar to a solution of pure complex, but in addition, a
doublet due to the added P(0CH3)3 appears at this low tem-
perature. Therefore, under conditions of fast and inter-
mediate exchange rates the observed Spectra are representa-
tive of a weighted average of all Species in solution and
the added P(0CH3)3 has been averaged into the above equilib—
rium. However, under conditions of slow exchange, the

doublet Spectrum for the added P(0CH3)3 is Observed along

56

Figure 13. 1H nmr Spectrum of: (a) Ni[P(OCH3)3]3(CN)3
in CH2C12; (b) solution in (a with .
added P(0CH3)3; (c) solution in (b) when
cooled.- . _

 

57

 

 

 

 

 

 

 

 

 

 

54'
b
-24'
C
-75'
C
k I. : 4. ; a 4 2. 4mm
4.00 3.80 3.60 3.40 3.20

Figure 13.

58
with the quartet for the complex. Furthermore, this process
is reversible as one doublet appears if the above solution
is warmed to 54°.

Further support for the proposal that the doublet Spec—
trum in these complexes represents an average of all species
in solution comes from the experiment which is summarized
in Figure 14. AS successive amounts of (CH3)3P are added
to the complex, Ni[(CH3)3P]3(CN)3, at 420 (the temperature
at which the 1H nmr Spectrum of this complex is a doublet),
the separation of the doublet is observed to decrease, pass
through zero, and form a new doublet with physical constants
very much like that of pure (CH3)3P. Therefore, under these
conditions of fast ligand exchange, coupling between ligands
through the metal has been removed and a doublet results

which is simply J but whose coupling constant and

PCH'
chemical shift is strongly influenced by the metal ion com-
plex. The addition of free ligand heavily weights the

solution in favor of uncomplexed ligand as is indicated by

the changes in chemical shift and J . An interesting

PCH
point to note is that at a mole ratio of added ligand to
complex of approximately 7.6, the coupling passes through
zero which means that free and complexed ligands have cou-
pling constants of opposite Signs. Similar observations
have been observed for other methylphosphine complexes“:82

and it is prdbable that J for the complexed ligand is

PCH
negative since a positive Sign has been proposed63 for free

(CH3)3P.

59

Figure 14. 1H nmr spectrum of Ni((CHa) P]3(CN)3 upon
successive addition of (CH3)3P.-

60

[L]

I’PCHI TNil3(CN)2]

 

 

 

9.0 0.0
4.4 H
2.7 2.1
0.0 7.6
J -
1.7 12.8
J
.2 ; e 1 4r— ¢ 499'"
1.50 1.30 1.10 0.90

Figure 14.

61
This same type of experiment was also performed by suc-
cessive addition of small amounts of the corresponding four-
coordinate complex, Ni[(CH3)3P]2(CN)2. Addition of this
complex heavily weights the equilibrium in favor of the
four-coordinate complex and the 1H nmr spectrum indicates

this fact by a Slow change in chemical shift and J to

PCH'
the values of the four-coordinate complex at room tempera-

ture (0 = 1.52 ppm and J = 0.0 Hz).

PCH
If CH3P(OCH3)3 is added to a solution of
Ni[CH3P[OCH3)2]3(CN)2, the doublet for the ocn3 protons
Shifts upfield but there is essentially no change in the
coupling constant. The CH3 protons similarly Shift upfield:
however, the coupling constant passes through zero as be-

fore (Figure 15). Therefore. J (coupling to the OCH3

POCH
protons) does not change Sign on going from free to com-
plexed forms but JPCH does. The same result is observed
for complexes of P(0CH3)3 and (CH3)3POCH3. Since JPOCH has
been predicted to be positive in the free ligands,a3'64 it
is also positive in these complexes. However, JPCH appears
to be negative in complexed ligands, since it is positive
in the free ligands.63
The sharp singlet which is observed at intermediate

temperatures and which is representative of an intermediate
rate of ligand exchange, indicates that the phosphorus and
hydrogen nuclei have been decoupled. The temperatures at

which the singlets for each complex reach a maximum intensity

are listed in Table VI. The mechanism for this decoupling

62

.nfiemoovmemo mo
coeufloow com: «avamu ca «Azovmmuflnmooamnmuaflz mo Eduuuomm use me

.oH «Homes

63

.mn musmflm

 

 

 

 

om _ o: o: end one... owv
_ _ _ _ J Be. a _ _ H J
o. \\. pp—
m «.2
o& X N;—
oi
8 ‘ .
Qd mp,
vim N; Q;—
. mg —
_n /<\/<\ 90 d
_ 2 IU m: MIUO .IU 44

64

Table VI. Temperatures of singlet maxima and magnetic
susceptibility data

 

 

 

T(°C) Xh Xh corr
Complex CH3' CH3O (x103cgsu)a (x103cgsu)b
Ni[(CH3)3P]3(CN)3 28 22 202
Ni[(CH3)3P]3(CN)3 -24 ~'0 251
Ni[(cna)2pocna]3(cu)2 25 16 24 289
Ni[CH3P(OCH3)2]3(CN)2 - 5 1 13 291
Ni[P(OCH3)3]3(CN)3 -17 3.4 296
{Ni[P(OCH3)3]5}(C104)3 >77 32 531
[Ni[CH3P(0CH3)2]5](C104)2 >72 >72 24 500
{Rh[P(ocns)3]5101o, 57 170 645
{Co[P(OCH3)3]5]C104C ~10d 1549 2016

 

aMagnetic susceptibility at the corresponding singlet maxi-
mum temperature. For complexes with both methyl and
methoxy groups, a temperature intermediate between the two
singlet temperatures was used.

bMagnetic susceptibility corrected for diamagnetism. See
reference 70.

cTrace paramagnetic Co(II) impurities are present in this
complex.

dOnly a very broad singlet is formed for this complex.

65

cannot be due to an averaging of cOupling constants of op-
posite signs because at the singlet temperatures the con-
centration of free ligand is very small and moreover, JPOCH
has the same sign in both free and complexed forms. Para-
magnetic relaxation effectsa5'57 similarly cannot be re-
sponsible for the decoupling since the magnetic suscepti-
bilities of all of these complexes in solution at the tem—
peratures where the singlets are at a maximum intensity have
been measured and are essentially diamagnetic. The magnetic
susceptibilities were measured by Evans' nmr method47 and
the corrected susceptibilities for the four- and five—coord-
inate Ni(CN)2 complexes are recorded in Table VI. Some of
these corrected susceptibilities are large, but they are
within the range of diamagnetic behavior.68

Therefore, by analogy with the Pd(II) and Rh(I) com-
plexes, the decoupling mechanism must be similar to that
proposed by Fackler, 22.21.55r59 In an elegant fashion,
theSe authors showed that the phosphorus-hydrogen decoupling
could arise from Spin exchange of the strongly coupled
phosphorus atoms under conditions of chemical exchange. The
similarities between the Pd(II) and Rh(I) systems and the
complexes studied in this work are Obvious, for indeed,
strong 31P-31P coupling across the metal ion is apparent
from the virtually coupled Spectra and,in addition,chemical
exchange occurrs.

Some more qualitative comments may be made, however,

concerning the decoupling mechanism. In the slow exchange

66

limit the exchange time, Te, is very long and Te >> 1/JPP,.

Hence, the virtually coupled spectra are observed for these

complexes whose features are dependent upon J and

PP" JPH'

JPH" As the temperature is increased and the exchange

time decreases, Te approaches 1/J and the large J

PP'
constant becomes a decoupling mechanism for J

PP'
. The

PH

rapid intermolecular chemical exchange of one ligand has
the effect of averaging the spin states, at the site of
phosphorus atom exchange, that the remaining phosphorus
atoms (momentarily bonded to the metal) "see". This spin
flipping causes the spin lattice relaxation rate of the re-
maining phosphorus atoms to increase which in turn, causes
the PéH coupling to be averaged out. It is apparent then

that a large J value is necessary for this mechanism.

PP'

PP' were very small, Te would exceed l/JPP' at rela-

tively slow exchange rates and the induced relaxation rates

If J

sufficient for decoupling would never be realized. If the
rate of exchange is increased beyond the point where the
singlet appears, eventually Te > 1/JPP, and the "resonant"
ligand does not “see" the exchanging ligand. Hence the de-
coupling mechanism is lost, the spectrum appears as if there
were no P-P coupling or at least very weak coupling, and the
fast exchange spectrum is a Simple doublet of separation JPH
which is a weighted average of complexed and uncomplexed
ligand in solution (yigg_§gp£§).

It is interesting to note that in the complexes of

(CH3)2P0CH3 and CH3P(OCH3)2 the singlet maxima for the methyl

67

and the methoxy protons occur at different temperatures for
each ligand (see Table‘vfl. This difference is understand-
able . . f . . .

in View 0 the above discuSSion Since JPCH and JPOCH
for each ligand have different values. Thus, since the
frequency of P-H coupling is different, the two types of

protons in each ligand will reach their maximum relaxation

rate at different temperatures.

.ML5n+ Complexes.- The proton nmr spectra of these

 

complexes are also temperature dependent and the general
features of the Spectra are Similar to the previous two
classes of complexes discussed. Figure 16 shows the Spectrum
for the rhodium complex, (Rh[P(OCH3)3]5}+ at the temperatures
indicated. Again, a high temperature doublet, an intermedi-
ate temperature singlet, and a low temperature multiplet

are observed; however, in this case the multiplet is a well
resolved six line resonance. The corresponding Ni(II) and
Co(I) complexes of P(0CH3)3 show very similar Spectra: how-
ever, the nickel complex decomposes at approximately 90°,
when the spectrum is a broad singlet: therefore, no high
temperature doublet is observed. In addition, the Co(I)
complex exhibited only a very broad singlet intermediate
between the doublet and sextet. This may be due to the trace
paramagnetic Co(II) impurities which were identified by esr
and which may be present from unreacted starting material.
Also, like the Ni(CN)2 complex of CH3P(OCH3)3, the complex,
{NilCH3P(OCH3)3]5}3+, exhibits two resonances for the two

types of protons, and again only the OCH3 protons exhibit

68

Figure 16. 1H nmr SpeCtrum Of [Rh[P(OCH3)3]5}B(C6H5)4
in CH2C12 at various temperatures.

 

l

 

57°
A2 2°
”:6.

IL I T lr fi‘ P Pm
3.80 3.60 3.40

Figure 16.

70
the low temperature splitting pattern while the CH3 reso-
nance is a broad singlet.

The doublets and singlets in these Spectra have the
same interpretation as before--they correspond to fast and
intermediate exchange rates, respectively. Absorption
spectra studies reported herein on the Ni(II) complexes
demonstrated their instability in solution with reSpect to
free ligand and a lower coordinated species. Similar ex-
periments on the rhodium and cdbalt complexes in this work
and elsewhere38 have shown their dissociative instability
and we have observed isosbestic points in the absorption
spectra as the temperature is varied. For the rhodium com-
plex the isosbestic points occur at 280 mu and 336 mu and
for the cdbalt complex at 341 mu and 402 mu. Furthermore,
the correSponding four-coordinate complexes of rhodium,
[RhL4]+ can be isolated for L = P(0CH3)3 and other similar
ligands.38 In the case of Ni(II) and Co(I), these lower
coordinated Species have not been isolated. Therefore, an

equilibrium of the form,

 

11+ n+ ( 6)

M115 < > ML4 + L

 

is present in solutions of all of these complexes. In the
case where M = Rh, this equilibrium has already been pro-
posed in order to account for the ligand equivalence indi-
cated by the singlet in the 1H nmr spectrum at 38° for

{Rh[P(OCH3)3]5]B(C5H5)4.3° This observation is consistent

with the work reported here since a singlet was observed in

71
proton nmr spectrum at approximately 38° (see Figure 16).

Therefore, under the conditions of the above equilibrium,

an intermolecular process of ligand exchange is occurring

in solution where the singlet recorded at intermediate tem-
peratures represents an intermediate rate of ligand exchange
and phosphorus-hydrogen coupling is averaged out due to

Spin exchange similar to the Ni(CN); complexes.

The symmetrical six-line multiplet which is nearly a
1:5:10:10:5:1 Spectrum represents the strong ligand-ligand
coupling through the metal in these trigonal bipyramidal
complexes. It may be considered as an AA'A"A"'A“"Xan'Xn"
Xh"'xn"" type of Spectrum in the limiting case where A and
X again have notations similar to before. The separation
of the two outer lines in the sextet observed for
{Ni[P(OCH3)3]5}2+ is about 11.8 Hz, a value very similar to
the doublet separation at high temperature, and it repre-

sents the sum J + J + J + J A-X

Ax + JA'X A"X A'“ x A""x‘
coupling through the metal (five bonds) should be very small

or almost negligible, thus the separation of the two outer

lines of the sextet Should be approximately JAX. It is

commonly observed54.58 that J for P(0CH3)3 (10.5 Hz)

POCH .
changes very little upon coordination, hence the value of
11.8 Hz for the separation of the two outer lines is con-
sistent with this type of spin system. Also, the fractional
intensity of the two outer lines is predicted to be (1/2)r-1
(r = number of A nuclei) that of the entire resonance."o

For these complexes r = 5, hence, the intensity of the two

72

outer lines is calculated to be 1/16 that of the sextet-—
close to what is actually calculated from the spectrum, and
what would be expected from the above binomial expansion.

The symmetrical nature of the six-line proton nmr pat-
tern for the ML5n+ complexes is interesting since there are
two chemically different ligands in the trigonal bipyramidal
geometry; the two axial ligands and the three equatorial
ligands. (Other five-coordinate geometries would also re-
sult in at least two chemically different ligands.) The 1H
resonance is symmetrical (even at 100 MHz) and indicative of
only one type of ligand and may suggest that the ligands are
rapidly exchanging sites intramolecularly, whereby all lig-
ands would be equivalent. This process could be similar to
the Berry71 mechanism postulated for pentavalent phOSphorus
compounds and M(PF3)5 complexes (M = Fe, Rh, Os). The
multiplet pattern is broad at -90°: however, this is prob-
ably due to solvent effects. At the present time, low tem-
perature spectra in other solvents have not been investigated.
The 31F spectra were recorded for some of these complexes:
however, the resonances are very broad and poorly resolved

and hence, no information can be obtained from them.

Rates of gigand Exchange.- If it is assumed that
there is present a similar ligand exchange rate where the
NiL3(CN)3 complexes all exhibit the singlet maximum, then
the temperature at this maximum singlet (Table VI) should
provide a measure of the relative exchange rates. For ex-

ample, Ni[(CH3)3P]3(CN)3 exhibits a singlet maximum,or a

73
similar rate of exchange.at a lower temperature than
Ni[(CH3)3POCH3]3(CN)2, therefore, (CH3)3P should exchange
more rapidly at a given temperature. A comparison of all
these “singlet" temperatures gives the following relative
exchange rates: (CH3)3P > P(0CH3)3 > CH3P(OCH3)3 >
(CH3)3POCH3. This ordering of these four ligands has been

observed before in the instability constants (K ) for

inst
the same NiL3(CN)2 complexes (p. 33) where the (CH3)3P

complex has the largest K. The correlation of ligand

inst'

exchange rate and Kin suggests that the equilibrium rate

st
is primarily determined by only one of the reactions, either
the forward or reverse (Equation 3). Also, the order of
ligand field strength as noted by the position of the d-d
transitions, is approximately the same as the above with
again the (CH3)3P complex having the lowest ligand field
strength (p. 33).

In a similar manner a qualitative exchange rate order

can be obtained for the P(0CH3)3 complexes of Co+, Rh+, and

Ni2+ (see Table\n3. The order of exchange rate is: Co+ >
Rh+ > Niz+. This order might be expected since Ni2+ with
the highest nuclear charge should be the least labile. In
turn, Rh+ should exhibit greater overlap of its 4d or-

bitals with the ligand orbitals than does cobalt's 3d

orbitals.

 

 

PART II

REACTIONS OF SOME CYANOPHOSPHINES WITH
DIBORANE AND BORON TRIFDUORIDE

74

INTRODUCTION

Recently, a rapid and simple method of preparation of
several cyanophosphine compounds, R3PCN (R = CH3, OCH3,
OC2H5. C3H5, N(CH3)2), was reported.72 The compounds
[(CH3)2N]2PCN73 and (C2H50)3PCN74 had been reported pre-
viously, but their preparation involved rather extensive
reaction times. Cyanophosphines have been found to exhibit
unusual coordinating properties. Reactions of these RaPCN
compounds with M(C0)4C7H8 (M = Cr or Mo and C7H3 = norborna-
diene) produce dimeric complexes of the type [M(CO)4L]2.
These latter complexes have been assigned the following
bridged structure (Figure 17) on the basis of infrared and

proton nmr data.

12
(c0) ’\ I“ :N‘M(CO)
“~°FECF1V' 4

Figure 17. Structure of LM(CO)4L]3 complexes (M = Cr or Mo;
R = CH3, oczns. C6H5. N(CH3)2).75
Nixon and Swain prOposed the same structure for the
complex [Mo(CO)4(CF3)2PCN]2.76 Thus, these bridged complexes

have demonstrated that two potential donor sites are present

on these cyanophosphines.
75

 

76

In order to further investigate this dual basicity,
the reactions of these cyanophosphines (R = CH3. CH30,
C3H50, C3H5. N(CH3)2) with the Lewis acids, borane and
boron trifluoride were investigated. Similar reactions of
RP(CN)2 (R = CH3. C6H5) were also investigated. Borane
and boron trifluoride were of particular interest since
they have shown markedly different affinities for phosphorus
and nitrogen base sites in adduct formation. Toward BF3.
the donor strength is in the order N > P, while toward 8H3,
the order is P > N.77'79 For example, in 1:1 adducts of
(CH3)3NPF3 with BH3 and BF3. the borane group is bound to
the phosphorus atom while BF3 is bound to the nitrogen
atom.8°

Many authors""""81 have explained the increased af-
finity of borane for phosphorus to be the result of overlap
between a BH3 molecular orbital of 0 symmetry with a
vacant 3dw (or 4prr)“"83 orbital on the phosphorus atom.
Thus, there is a dv-pw (or pw—pn) component enhancing the
strength of the PéB oébond. In addition, this drift of
electron density lowers the energy of the molecule, since
it partially neutralizes a charge separation of the type
X3P+ - 8H3.79 For the BF3 group, the high electronegativity
of the fluorine atoms does not allow a similar drift of
electron density.79 This "hyperconjugation" argument has
been used to explain the existence of the compounds OCBH384
and F3PBH3,35 and the nonexistence of OCBF3 and F3PBF3.79:°°

"Back bonding" of this type is similar in kind to that which

 

77
occurs between a phosphorus ligand and a transition metal
atom.1 Indeed, for the [M(CO)4L]3 complexes (Figure 17),
Jones and Coskran75 reported the.M-N bonds are broken in

the reaction
[M.(C0),L]2 + 2L ——> 2M(co),L3 (7)

rather than the M-P bonds. However, the hyperconjugative
model of P-B bonds has not been consistent with other ex-
perimental results. For example, the model would predict
that 3H3 would form a more stable adduct with PF3 than
with PFaH. because of enhanced v-acceptor character of PF3
over PFzH,87 but HFzPBH3 is more stable toward dissociation
than F3PBH3.°°

Rudolph and Parry have proposed another P—B bond models.8
The Lewis acidébase interaction energy (E) can be approxi—

mated by
E = F[p+(Fa/2)1 (8)

where F = Lewis acid field strength, [)2 dipole moment of
the lone pair, and a = polarizability of the lone pair.
Borane bond strength reversal (P > N) has been shown to be
a function of the Lewis acid strength,89 and is consistent
with this model. The greater stability of HFzPBH3 over
F3PBH3 toward dissociation has been attributed to the
larger polarizability, a, of the lone pair on PF2H than

that on PF3.87

 

78

It was of great interest to determine whether the ob-
served basicity toward BF3 (N > P) and BH3 (P > N) would
also be observed in the cyanophosphines of this study. Any
spectroscopic data obtained for such adducts might also
help to discriminate between the two P-B bond hypotheses.

Furthermore, since adducts of the type, RsPBH3 (R =
CH3.9° 01130.91 C2H50,91 03115.92 N(CH3)3°3), have been pre-
pared previously, as well as the R3PBF3 (R = CH3.9‘ C3H595)
adducts, a comparison of the experimental results and spec-
troscopic data for the corresponding [R3PCNlBH3 and
[RaPCN]BF3 adducts that are formed in this study helped to
determine what effect the cyanide group has on the o-donor

Character of the phosphorus atom to which it is bound.

 

EXPERIMENTAL

All work was carried out by standard vacuum line

techniques or under inert atmosphere.

Materials.- Boron trifluoride was obtained from
Matheson Scientific Inc., Chicago, Illinois and purified
by passing it through a -145°C trap. Diborane was prepared
by the action of H3804 on NaBH4.96 The cyanophosphines
RgPCN (R = CH3. c330, C3H5O. cana. N(CH3)2) and RP(CN)3
(R = CH3, C3H5) were prepared according to a method pre-
viously described.72 Methylene chloride was dried over

CaH, before use.

Preparation of [RzPCN]BH3 Adducts.— An excess of

 

BgHe was condensed from a calibrated trap into a flask

which contained the R2PCN ligand (R = CH3, CH3O, C2H50.
C5H5. N(CH3)2). The mixture was stirred and allowed to
warm from -196° to room temperature. When the uptake of
B3H6 was complete («'1 hr), as measured by an attached
manometer, the excess B3H5 was condensed into the calibrated
trap. .A molar ratio of approximately 1:1 was obtained for
BH3:R3PCNO(Tab1e\HIL (For R = C5H5 and CH3, the value of
BHazRgPCN was determined when these RaPCN ligands were dis-
solved in methylene chloride solutions (see BF3 reactions).

79

 

 

80

J Jl-..‘

 

.umwc can conuomomn

. «acnmo ca can coauomo¢w

 

mefinmmv_«fizoooomooa

 

 

no.” 1-- oa.n -1- mo.m
so.“ 1-- 6H.n -1- oH.n muxnnmv_axzovoemu_
oo.o mo.o H¢.n oe.m on.n mxmmmvnmm_zomnxnmoov_
oo.H Ho.o me.n Ho.“ en.n mxnmnvemm_zomaxonmov_
no.o mo.o eo.u on.H nm.u mAemmvnmmHzoouxemUVH
oo.o nun oo.v In: ne.v nemmxzoouxnmeoVH
no.” 1-- ov.n nu- am.n mnnm_zomuxonmov_
oo.o -u- vo.n -1- oH.e nemmflzooexnmov_
nu- oo.fi In- on.m nn.v nnmmmzom«~z«xemov_w
-u- oo.o -1- on.n oo.o nemn_zom«xameov_
1-- mo.o -1- um.H no.m nemm_zomuxoemnov_
-11 Ho.o -1- Hm.H sw.n nemmzzomaxoemoVH
1-- no.o -11 en.n be.» memmflzooexnmovn.
ocwpmnonm oaoe ocasmmonm maoe ms cmxmu mo coxmu ocflsmmonm
.lnwm oaoa .lmmm oaoe ohm mo Hoes omam mo Hose mo HOEE ocoomaoo

L

 

 

.coquEHom uoooom «mm ocm nmm Mom memo Hmucoefluomxm .HE/manme

81

Since these adducts were solids, efficient stirring and
complete reaction were not possible when the reactions were
run neat.) White solids were formed when R = CBHa or CH3,
and these solids were recrystallized from a CH2C12-hexane
mixture. These complexes decomposed in a matter of hours
at room temperature. When R = CH30, N(CH3)2, and C2H5O.
light yellow liquids were obtained which contained small
amounts of decomposition products, and also decomposed in

a matter of hours at room temperature.

 

Preparation of [RzPCNlBFa Adducts.- Boron trifluoride (

 

was condensed into a -196° flask which contained the RaPCN
ligand (R - CH3, c330, cans) dissolved in cnzclz. This
mixture was allowed to warm to -78°, until the BFs uptake
was complete (~45 min; when R' CH3O. the reaction was stop-
ped when the uptake slowed markedly). Excess BF3 was con-
densed through a -145° trap into the -1960 trap, in order
to remove CH3C12. A mole ratio of approximately 1:1 was
Obtained for BF3:R2PCN (Table VII). The adducts formed
clear solutions in CH2C12 and were unstable above -20°

toward loss of BF3.

Preparation of [RzPCN]BF3(BH3) Adducts.- The

 

RzPCN(BH3) adducts (R s CH3, CH30, cans) were formed, in.
situ, and allowed to react with BF3 in the manner described
above. This resulted in a molar ratio of approximately
1:1:1 for BH3:BF3:R2PCN (Table VII). These complexes formed
clear solutions in CH2Clé and were unstable with respect to

loss of BF3 above -20°.

82

Reaction of RP(CN)2 with BF3.- The RP(CN)2 ligands

 

(R = CH3, C6H5) were allowed to react with BF3 in the manner
described above. A molar ratio of approximately 2:1 was
obtained for BF3:RKCNL (Table VII). These compounds formed
insoluble white solids and were also unstable with respect

to loss of BF3 above -20°.

Nmr Spectra.- Room temperature 118 spectra were re—
corded on a modified NMR Specialties Model MP-1000 Spectrom-
eter at 60 MHz. Low temperature 113 spectra were recorded
on a Varian Associates Model HA-100 spectrometer at 32.1 MHz.
Phosphorus-31 spectra were recorded on a Varian Associates
Model DA-60 spectrometer at 24.3 MHz. Proton and 19F spectra
were recorded on a Varian Associates Model A56/60 spectrom-
eter at 60.0 MHz and 56.0 MHz, respectively. A Fabri-Tek
Instrument Co. Model 1082 computer was used to time-average
the proton spectra. The 118, 31P, 19F, and 1H chemical
shifts are relative to external B(OCH3)3, 85% H3P04. CC13F.
and TMS. respectively. Spectra were obtained for neat
samples of [RzPCNlBH3 (R = CH30, cznso, N(CH3)2), and

methylene chloride solutions of all other adducts.

 

Infrared Spectra.- All spectra were recorded on a
Perkin-Elmer Model 225 grating Spectrophotometer in methylene
chloride solutions. For boron trifluoride adducts, a low

temperature solution cell cooled to -78° was used.

 

RESUDTS

The reaction of RaPCN (R = 03,, 0330. canao. cans.
N(CH3)3) with Bane resulted in 1:1 borane adducts. First
order nuclear magnetic resonance spectra (113, 31F, 1H)
were obtained for these adducts; the spectra (Figure 18)
of [(CH3)3PCN]BH3 are typical. Coupling constants and
chemical shifts Obtained from the Spectra of these compounds
are recorded in Table VIII, along with the cyanide stretch-
in the 113

ing frequencies (v The observation of J

CN)° PB
spectra for all of these complexes is indicative of adduct
formation on the phosphorus base site, rather than on the
nitrogen. A second borane molecule was not taken up by
these ligands, even when the reactions involved lower tem-
peratures (-78°) and longer reaction times (24 hrs).

The reaction of RaPCN (R = CBS. onso, can,) with BF3
also resulted in 1:1 adducts. Nmr (118. 19F. 31P, 1H) and
ir data (VCN) obtained from the spectra of these compounds
are reported in Table IX. Typically, the 113. 31F, and 19F
spectra are broad singlets, and therefore, only chemical
shifts are reported. However, in the 1H Spectra splitting
patterns are observed for the R protons (R = CH3, cnao)

and J and J values are reported along with the chemi-

PCH POCH
cal shift data. All nmr spectra were obtained at

83

 

84

Figure 18. Nmr Spectra of [(CH3)2PCN]BH3. (a) Time -
averaged 1H spectrum at 60.0 MHz. (b) 1H
spectrum at 60.0 MHz. (c) 11B spectrum at
60.0 MHz. (d) 31P spectrum at 24.3 MHz.

85

1 1 A I 1 l

"'1 __I_
“JR ) 61H(3H3)

I‘m '15:, E ]

 

 

 

(SI '3

I400Hz '1

 

 

d - a» L- |,
5319

Figure 18.

86

.mcououm mamamnuoa mo mcaamsoo

.mwmmnucwumm ca co>fim mum mooam> ocmmfia combo

.mcouonm Hanuoe mo umazm Hmowfimnoo

 

 

o
.mcououm ocoamnuoa mo amaze Hmoweonon .mem unwound mo umwnm H60H80£Um
Amoaavmmsm as so we n.oH ¢.on on . no.o vs.m «Anmovz
. onn.H
Aoofiuvsofim mo Hoe -1- oo o m on no u on o aHm.¢ oamao
Acmemvnofiu co Hoe we «.mn s.mn see- no.o nm.n oamo
Anaemvsoaa on Hos he u.oH n.nn o oe.o Hs.H emu
oxnsflavooau no: so 1-- -1- o.mn 6H 1 -1- one.» ammo
.Asmv
1 5V :5 :5 Es M wages as: sag Es ace
7 mm mm and m . n n m
20 o o n AEnoch mass mans A mmvmao Amvmao

>

 

 

.muuooom mmmnzumnma How mumo He too HEZ

. HHH> THQMB

87

.momosucoumm ca mum muosoom Anmmvzomam mo mosam>n

.mmmmnucmumm ca mum mosam> osmofia momma

 

 

Sam «.3 no Tao 1mm 82: 37 013.886 AnmmVammEonaAosmo:
scan a. 3 an o. so H. mm a. o3 2.. AA 2.. Sum. H “£3 amazon: #8:
some ...i .i .i- n. on o. «3 & onto? 1% on. s v on. .c. ”an Eon: :16:
3am see .i .1- mom on: «.32 van? ~35“;st aeflzomaxoamo:
moms n. v .... u..- oém he: mxmoionl. 6:158. H «mm Sum: «so:
fusov AEeaooae Ammv Assoc Ede Assay 38v Assoc ounces

20> Amvmomo b Anmmvmuno Anmmvmuuo hose mane Amvmuo

 

 

.mnosoom AnmmvnmmHZOmsmH osm nmm_zomsm_ now some He can nsz .xH manna

88
low temperatures («v-40°), since these adducts are not stable
above -20°. Inasmuch as JPB is not observed in the 118
spectra, it is proposed that the BF: molecule is bound to
the nitrogen atom of the cyanide group.

A second BF3 molecule was not taken up when R 3 CH3 or
C3H5, even with longer reaction times (~t24 hrs). Reaction
of RgPCN (R = C3H50, N(CH3)2), or further reaction of
(CH30)3PCN, with BF3 resulted in the decomposition of the

ligand unit. Boron trifluoride was taken up in greater than E

2:1 amounts (BF3:RaPCN), which is prObably due to the reac- 1,-

 

tion of BF3 with oxygen and nitrogen atoms of the substituents.
At room temperature, BF3 was given off, and the proton nmr
spectra indicated the R2PCN ligands were not present.

Table Ix also contains similar nmr and ir data for the
[R3PCN]BF3(BI-13) adducts (R = CH3, cnao), which are formed
in the reaction of BF3 with [RgPCNlBH3. When R - C3H5, the
adduct that is formed is insoluble and no Spectroscopic
data were obtained. The 31F and 19F spectra consist of
broad singlets, while in the 1H Spectra only methyl and
methoxy resonances are Observed. Boron-11 spectra exhibit
resonances due to the BF3 and 3H3 groups (Figure 19). It is
apparent that the 8H3 group is bound to the phOSphorus atom,
since JPB is Observed in the borane resonance. The nmr
spectra for these latter two complexes were also Obtained at
~v-25°, since these adducts are not stable above -20°, and

“thermal decoupling"97 induced by quadrupolar boron coupling

was observed below this temperature, as evidenced by loss of

89

Figure 19. 113 nmr spectrum of [(CH30)2PCN]BF§ 8H3) at
32.1 MHz. ((a) Spectrum at -25°. b) Spec-
trum at -50- . -

91

JPB in the 118 spectra (Figure 19).98 When R = C3350 and
N(CH3)2, decomposition Occurred which was Similar to that
observed in the corresponding 1:1 BF3 adducts.

No reaction occurred when RP(CN)2 (R = C3H5. CH3) was
allowed to mix with 3236 at low temperatures (-78°) and
long times (~o24 hrs). The 2:1 adducts formed from the re-
action of BF3 with RP(CN)3 (R = c311,, 033) were insoluble
and unstable above -20°. No Spectroscopic data were Obtained

for these compounds.

 

DISCUSSION

For convenience, the 1:1 borane adducts will be dis-

cussed separately from the other adducts of this study.

 

[RaPCNlBHgyAdducts

31P - 113 Coupling Constant, JPB.- In a discussion

of the relative merits of the two P-B bonding models

described previously, J is of particular importance,

PB
since it is a property of the bond in question. The sign

of JP has been indicated to be positive33t99 in borane

B
adducts. Therefore, the adducts in TableVIII are listed in

order of increasing JPB’

Rudolph and Shultz99 have reported an empirical cor—

relation between the magnitude of JPB and dative bond

strength for a series of smoothly varying phOSphine ligands

such as F3 —anP —> 8H3. Me3_anP -—> 8H3, (MezN)3_nFnP —>

BH3, and F3XP -9 8H3 (x = F, Cl, Br). These authors have

also reported that J values of similar magnitude in borane

PB
adducts of phosphines which differ markedly may not reflect

dative bond strength, but quite different JPB values reflect

correct relative bond strengths. Therefore, the JPB values

in this study (Table VIII) indicate the following decreasing
order of P-B bond strength in [RaPCNlBH3: R = N(CH3)2 ~'

C2H50 ~ CHaO > CH3 > C3115. 2
9

93

The borane hyperconjugation model is not consistent
with this order of P-B bond strength. If the P-B bond is
stabilized by "back bonding“, bond strength should depend
on the ligand's w-acceptor ability.83 In studies of metal
carbonyl complexes and their carbonyl stretching force
constants, Jones and Coskran5‘ found the following order of
decreasing w-acceptor ability: P(0CH3) >P(CH3)3 > P[N(CH3)3]3.
If this same relative order is asSumed to be present in the
corresponding cyanophosphine ligands, it is not reflected in
the above P-B bond strength order, since {[(CH3)3N]PCN}BH3

has a larger J than [(CH3)2PCN]BH3.

PB
Cowley and Damasco83 have found an empirical correla-
tion between JPB (P-B bond strength) and the basicity of the
phOSphine. This is consistent with the model of Rudolph and
Parry (Equation 8), since the lone pair polarizability. o.
is related to the phosphorus basicity.87 The JPB values in
this study reflect the same order of basicity as proposed
by Cowley and Damasco83 for R3P (R a N(CH3)3, CH30, CH3) and
RPH, (R = CH3, C3H5) ligands, and thus, they are also con-
sistent with equation 8. It should be noted that these
authors83 report the following order of decreasing JPB
(ligand basicity): P(0CH3)3 ~'[(CH3)2N]3P > (CH3)3P. How-
ever, on the basis of base displacement reactions, Reetz91
has reported the following basicities: (CH3)3P ~'
[(CH3)3N13P. Foester and Cohn1°° have reported that base

strengths toward borane decrease in the order CH3PF2 >

(CH3)2NPF2 > CH30PF3, based on displacement reactions,

94

although J decreases in the order (CH3)2NPF2 > CH30PF3 >

PB
CHaPFa. The measurement of large basicities for these
methylphosphines toward borane may be due to entropy
effects“1 in the base displacement reactions. For example,
dissociation constants of (CH3)3NB(CH3)3 and (CH3)3PB(CH3)3
suggest the phosphine is the stronger base, while enthalpies
of formation of these adducts indicate the N-B bond is
stronger than the P-B bond.101 Therefore, the order of
basicities as proposed by Cowley and Damasco,83 and followed
in the corresponding cyanophosphines of this study, may in-
deed be correct, and thus be reflected in J .

PB
The relatively high base strength (and high JP in the

B
borane adducts) of (CH30)3PCN, (C3H50)2PCN, and [(CH3)2N]2PCN
is probably due to (srd) wébonding between the filled 2p
orbitals of oxygen and nitrogen and the empty 3d orbitals
of phosphorus. The vébonding increases the electron density
on the phosphorus atom, and thus, increases the basicity.“2
This type of (p-d) wébonding has been shown to be signifi-
cant in structural studies of HzNPF31°3 and CH30PF2.1°‘
The nitrogen atom of HzNPFz is in a planar environment,”3
and the [(POC) of 123° in CH30PF2 suggests sp2 hybridiza-

tion in oxygen orbitals.1°‘

31 _1 - .
P H (Substituent) Coupling Constants, JPng JPOCH'

_ l - 63,63 63.64
JPNCH' The coupling constants JPCH' JPOCH’ and

JPNCH64 in the free cyanophosphines are assumed to be posi-

tive. There is an increase in the absolute value of all

these couplings in their borane adducts (Table X).

95

 

 

 

 

Table x. Values of J , JP , and JPNCH 1n R2PCN and

[RaPCN1B33 .

JPCH' JPOCH or IJPCH I ' IJPOCHI J Sign: Of

R .m .. W .. semi .. 130:: ”SE
[RaPCN13H3 ' PNCH
[RzPCN]BH3

CH3 4.5 10.2 _10 .2 E1
CH30 10.3 12 .2 +12 .2 '
C2350 7-5 9.0 + 9.0 1..-
N(CH3 )2 10.0 10_5 ”0.5

 

These results can be interpreted in terms of the in-
creased R-P-R and NC-PeR bond angles in the borane adducts
as compared to the free ligands, which results in an in-
crease in the s character of the phosphorus bonding orbi-
tals ("isovalent hybridization hypothesis").1°5 Muller and
Pritchard 105 have observed an increase in JCH as the amount
of S character in the CH bond increases. Cowley and
Damasco83 have proposed a similar relationship between JBH

and the amount of 2s character in the B-H bond. Manatt.

et al.,°3 have prOposed that J becomes more negative as

PCH
the amount of 3 character in the phosphorus bonding

orbitals increases in a variety of trivalent phosphorus com-
pounds. For example, JPCH

in (CH3)3P is +2.66 Hz, where
1 ' .
the Z.(CPC) of 98.6° 07 suggests a large amount of p orbi-

tal contribution to the bonding orbitals, while in the

96

nearly pure Sp3 hybridized (CH3)4P+ and (CH3)3P=O. JPCH

is -14.4 Hz and —13.4 Hz, respectively.63 Therefore, for

[(CH3)3PCN]BH§ J must be negative (-10.2 Hz) in order to

PCH
become more negative upon adduct formation. This change in

the Sign of JPCH for methylphosphines upon complexation has

been observed previously55:61 (p. 58). However, for JPOCH

and JPNCH similar Sign changes are not observed54 (p. 61),

and only small changes in the magnitude of JPOCH and JPNCH

are observed upon complexation.54'°‘ Therefore, these JPOCH

and JPNCH values for the borane adducts are assigned posi-

tive values. Thus, upon adduct formation, JPCH decreases

while JPOCH and JPNCH increase.

118 - 1H Coupling Constant, JBH.- The sign of JB has

H
been assumed positive.108 Valuesgf.JBH Obtained in this
study are not consistent with a simple isovalent hybridiza-
tion hypothesis.“5 Electronegativities of the substituent
groups decrease in the order CH30 o'C3H50 > N(CH3)3 > CBH5 >
CH3,1°9 and the 5 character in the B-H bond should decrease
in the same order.“5 From Table VIII, it can be seen that
JBH of [(CH3)3PCNJBH3 is not consistent with this order.
However, since the above hypothesis depends on an increase
in the borane bonding angles ([.(HBH)),1°5 the bulkiness of

the phosphine may have an effect on the order of J Tol-

BH'
man36 has measured the “cone angles" (Figure 20) of a num-
ber of tertiary phOSphorus compounds, and the method used

would predict the following order of increasing cone angle

97
for the RaPCN ligands: R = CH30 ~'C2H5O «'cas > cana. (The
value of2282 is used as the average value of a Ni-P bond.
However, the relative order of ligand cone angles should not

change, if an average P-B bond distance is chosen instead.)

 

Figure 20. The ligand cone angle for tertiary
phosphorus compounds.36

Due to the two methyl groups in the N(CH3)2 group, the
cone angle in the corresponding cyanophosphine would be
expected to be larger than the CH30. C3H50, and CH3 ligand
angles, but smaller than the ligand angle with the bulky
C5H5 group. Therefore, the ability of the borane group to
increase H-B-H bond angles, or increase 5 character in
the B-H bond1°5 (JBH), should decrease in the following
ligand order due to ligand steric requirements: R 8 CH3 c:
C3H5O ~'CH3O > N(CH3)2 > CoH5. This order is consistent

with the J values obtained for these borane adducts (Table

BH
VIII).

The above argument considers that an increase in H-B-H
bond angles is caused by o-bond electronegativity effects,
although the magnitude of this increase is due to ligand
steric requirements. However, the borane hyperconjugation

model also allows for increased H-B-H bond angles, with

98
increased ligand v-acceptor ability.83 Although the order
of JBH (Table VIII) does not strictly mirror the ligand 7-
acceptor ability, since (C5H5)2PCN is a better w-acceptor
than (CH3)2PCN,1°°o11° hyperconjugation can still be con-
sistent with the above model. It assumes a negligible
effect on the sterically determined, relative J order due

BH
to electron density drift in the B-H bonds.

31F - 1H Coupling Constant, J The observed

PBH' '
values of JPBH (Table VIII) follow-the order of substituent
electronegativity, 1:9,, OCH3 > N(CH3)2 > CH3."9 The iso-
valent hybridization hypothesi31°5 would predict an increase
in the 8 character of both the P-B and B-H bonds as
substituent electronegativity increases. If the PéB-H

bonding situation is analogous to P-C-H, then J H should

PB
decrease with increasing 8 character in the phosphorus

o o as
bonding orbitals, as does JPCH' In order for JPBH of these
cyanophosphine adducts to decrease with increasing electro-

negativity, J must be assigned a negative Sign. However,

PBH
this result is ambiguous, since it neglects the contribution
of the ligand and borane adduct bond angles33'105 to the

amount of 3 character in the P-B-H bonds.

31P Chemical Shift, 031P.- Table XI shows the change
in 031P of these cyanophOSphines upon adduct formation.

If only the molecular ground state electron distribution
is considered, a downfield shift of 031P might be expected

because of the o-donor character of these trivalent

99

 

 

3+ so? an? zoo «A Onmov
8+ co 1 ST zomaxoamaov
3+ on u no .. 2232265:
on: 3 .. on + zomaxomoov
«on o no + zomaxnmov
.Aemmv . .
Amcfinmmonnvmwno I Auoooom flammv . Aemmv .
TQMHOQ “H00 N mfiflflq AUUHpHvUM QCMHOvafiQO .HATCfl—hmuogmvmflflQ ”Cwfignmmogm

 

 

.nmmmzumnma ocm zomnm cw mumo umwnm HmowEOso monogamocm .HN magma

100
phOSphorus compounds.111 A small downfield shift may be
attributed to the w-acceptor character of such ligands.111
There are two other contributions, however, to the chemical
shift; a paramagnetic term, derived from electronic asym-
metry around the nucleus,87 and the contribution from the
other atoms in the molecule.112 This latter contribution
is usually small and does not appreciably affect the chemi-
cal shift.112 Phosphines with substituents that contain
lone pairs next to the phosphorus atom have a large para-
magnetic contribution to the 31P chemical shift.33o113 The
paramagentic contribution is reflected by large negative
values of 031P for these ligands, and an upfield shift upon '
adduct formation.33:113 The 31P chemical Shifts Observed in
this study are consistent with these theories (Table
IX), and are similar to the chemical shifts obtained in

fluorOphosphine-borane83 adducts and phosphoryl compounds.113

In Chemical Shifts, 01H(BH3) and 01H(R).- Table XII

 

shows the change in 01H of the substituents upon adduct
formation. In general, 01H(R) shifts downfield, and indi-
cates a primary dependence of 01H(R) on the molecular ground
state electron distribution. The protons of the substi-
tuents are slightly deshielded by a drift of electron density
toward the borane unit.

Borane proton chemical shifts generally move upfield as
J increases (Table VIII). This is consistent with phos-

PB

phine basicity arguments, since the more basic phOSphine

101
will shield the borane protons to a greater extent and move

61H(BH3) upfield.

Table XII. 01H(R) in RzPCN and [RZPCN]BH3.

 

 

 

 

R 61H(R2PCN)1 51H([R2PCN]BH3)
4.08a 4.21a
°2H5° 1.30b 1.33b
can5 7.20C 7.63°
N(CH3)2 2.75 2.74
aMethylene protons. bMethyl protons.

CChemical shift of largest peak.

Cyanide Stretching Frequency, vCN°— The cyanide stretch-
ing frequency of these cyanophosphines increases upon adduct
formation (Table VIII). An increase in this frequency has
been observed in compounds of the type RCN, when boron com-
poundsn-fi'118 or transition metals5°o113I117 coordinate to
the nitrogen atom. It has been suggested,115o117 that the
CN bond in the free nitriles is intermediate between a
double and triple bond, due to a mesomerism of the following
type: R-CEN(III) <—-9 R-C:N-(IV). Once the nitrogen atom
becomes coordinated, such a phenomenon is not possible be—

cause of the different geometries and hybrid orbitals

involved (Figure 21). An X-ray structural study has shown

102
structure V is present in CH3CN-BF3.118 Therefore, the CN
bond is a triple bond in the coordinated complexes and is

indicated by an increase in v .

CN
+ O.
R-CEN—8F3 R-C=N\ -
5P sp2 BF3
Y EL

Figure 21. Possible structures of [RCN-lBF3.51

If in the cyanophOSphines of this study, a mesomerism
+ _
of the form RaPCEN(VII) <-> RaPc=N (VIII) can be drawn,
then adduct formation on the phosphorus atom should cause
a drift of electron density away from the cyanide group,
which makes structure VIII less favorable, and an increase

in VCN should be observed.

[RzPCN]BF3 and [RzPCNlBF3(BH3) Adducts

 

31P - 11B Coupling Constant, JPB'- The 11B nuclear

magnetic resonance spectra of [R2CN]BF3(BH3) (R = CH3, CH30)
indicate that the borane unit is bound to the phosphorus

atom, since JPB is observed in the borane resonance (Figure
19). Thus, the BF3 molecule is bound to the other base site,

the nitrogen atom of the cyanide group. In these complexes.

JPB is slightly smaller in magnitude than that observed for

the corresponding [RzPCNlBH3 adducts (Table VIII). The

smaller values are consistent with the correlation of JPB

103
with phosphine basicity, since the electronegative BF3 group
should reduce the basicity of the phOSphorus atom and de—
crease JPB' The borane hyperconjugation model is not con-
sistent with this observation, since the BF3 group would

be expected to increase the w-acceptor ability of the phos-

phorus atom, and lead to an increase in JPB'

31 _1 - - _
P H (Substituent) Coupling Constant, JEQE and JPOCH'

The 31P - 1H coupling constants in [RgPCN]BF3(BH3) (R = CH3.
CHSO) adducts (Table IX) do not change from those observed
in the RzPCN(BH3) adducts (Table VIII). A BF3 molecule on
the nitrogen atom of the cyanide group should not cause a
large change in the S character of the phosphorus bonding

orbitals and J and JP should not change substantially.

PCH OCH

Likewise, in the [RgPCNlBFs (R = CH3. CH30) adducts JPCH

and JPOCH do not change from those observed in the free
ligands (Table X), which again indicates the BF3 molecule

is bound to the nitrogen atom.

31F Chemical Shift, 03:3.- The phOSphorus chemical

 

shift in the [RZPCNlBFa (R = CH3, CH30, cans) adducts moves
slightly downfield in all cases. If the BF3 molecule was
bound to the phosphorus atom, large changes in 031P might be
eXpected, similar to those obtained in the [RzPCN]BH3 adducts.
The consistently small downfield shift is indicative of a
small deshielding of the phOSphorus nucleus111 caused by

the electronegative BF3 molecule on the nitrogen atom.

Further evidence for a N-BF3 bond is given by the similar

104
effect of BF3 on 031P of the free ligand and borane adduct
(Table XIII). Since the BFg molecule is bound to nitrogen
in theIRzPCNiBFg(BH§) adducts it is probably bound in a

similar manner in [RgPCNlBFa adducts.

1H(Substituent) Chemical Shift, 01H.- The proton
chemical shifts of the substituent groups in the [RaPCN]BF3
(R = CH3. CH30, CoHa) adducts (Table IX) are moved down-
field from the corresponding free ligand values (Table XII).
These protons are deshielded by the drift of electron density
toward the electronegative BFa group. In the [RaPCNlBF3(BH3)
(R - CH3, CH30) adducts, the borane unit enhances the elec-
tron density drift away from the substituents and 01H(R) is
moved further downfield (Table IX). The effect of BF, ad—
duct formation on 01H of the substituents in both free
ligands and borane adducts is similar (Table IX), which

again indicates the coordination of BF3 to the nitrogen atom.

19F Chemical Shift, 619F.- The fluorine chemical
Shifts of the [RaPCN]BF3(BH3) (R = CH3, CH30) adducts are at
lower field than the corresponding [RzPCN]BF3 adducts (Table
IX), since the shift of electron density toward the borane

unit deshields the fluorine nuclei.

11B Chemical Shifg, 0¥1§,- The boron chemical shifts
of the SP3 units in [RaPCN]BF3(BH3) (R = CH3, cnao) are at
higher field than in the corresponding [RzPCNlBF3 adducts.

The shifts are in accord with the fluorine chemical shifts

105

 

 

o . ossu Ros: nmmmzomaflonmoV.
oH mos- «as- zooaxosmov
ms mu 1 o nmm_zomsxnmov_
as on + so + zoasxsmov

Auosoom «savanna n

uosoom ocmuog n Auoooom common
um maesmmosmvmsno Aposoom mmvmsno so onecmnosmvmsno ossomsoo
n mnno<

 

 

.coquEHom poooom «mm com: nmmnzomnma can zumam ca memo cw omcmsu .HHHN magma

106

that indicate a drift of electron density away from the
fluorine atoms and onto the boron atom, which increases the
shielding of the boron nuclei. Boron chemical shifts in
the borane units follow the same trend, indicating a drift
of electron density from the borane hydrogen atoms, and
increased shielding of the boron nuclei in [RZPCNlBF3(BH3)

as compared to the corresponding [RaPCNlBH3 adducts.

Cyanide Stretching Frequency, VCN.- The cyanide
stretching frequencies of the [RzPCNiBFa adducts (Table IX)
agree with the assignment of an N-B bond. As shown pre-
viously (p. 102), the CN bond in the free cyanophosphines

is intermediate between a triple and double bond. Upon
adduct formation on the nitrogen atom, the CN bond has more
triple bond character (structure VII). Thus, the CN stretch-
ing frequency should increase, and to a greater extent than
in the borane adducts, where the mesomerism is only partially
shifted. The cyanide stretching frequency in the
[RzPCNlBF3(BH3) (R = CH3. c330) adducts is similar to that
observed in the corresponding [RzPCN]BF3 adducts, since the

maximum bond order is attained in both cases.

CONCLUSION

Spectroscopic results have indicated that the 8H3
molecule is bound to the phosphorus atom and the BF3 mole—
cule is bound to the nitrogen atom of the cyanide group in
all the adducts of this study. The formation of 2:1 adducts
(BF3:RP(CN)2) with RP(CN)3 (R = CH3, c635) ligands supports
this asSignment. '

In general, the borane hyperconjugation model has been
shown to be inconsistent with the nmr data obtained for the
borane adducts in this study. A correlation between JPB
(P-B bond strength) and phosphine basicity has been observed.
This is not to say that borane hyperconjugation does not
occur in these adducts, only that this effect does not
determine the observed values in the nmr data.

The JPB values of the [RaPCN]BH3 adducts (Table VIII)
are consistently lower than those found in the borane adducts
of the corresponding PR383I113 compounds, and indicate that
replacement of R by a CN group decreases the basicity of
the phosphorus atom. Experimental results also support this
hypothesis, since borane adducts could not be formed with
RP(CN)2 (R = CH3, cans) although [RaPCNlBH3 adducts are
readily prepared. Furthermore, boron trifluoride does not

coordinate to the phosphorus atom of R2PCN (R = CH3. CH3O).

107

108

although BF: does coordinate to the corresponding PR3 com-

pounds.94I°5

BIBLIOGRAPHY

10.

11.
12.

13.

14.

15.

16.

17.

BIBLIOGRAPHY

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

M, O. Workman G. Dyer, and Devon‘W. Meek, Inorg. Chem,
6, 1543 (1967).

M. J Norgett, J. H. M. Thornle and L. M. Venanzi,
Coord. Chem. ReV., 2, 99 (1967) .

Luigi Sacconi, Pure Appl. Chem" 11, 95 (1968).
C. Furlani, Coord. Chem. ReV., 3, 141 (1968).

L M. Venanzi, Angew. Chem., Internat. Edn. (in English),
3, 453 (1964).

L. P. Haugen and R. Eisenberg, Inorg, Chem., 8,1072
(1969).

B. Bosnich, R. S. Nyholm, P. J. Pauling, and.M. L.
Tobe, J. Amer. Chem. Soc., 90, 4741 (1968).

G. A. .Mair, H. M. Powell, and L. M. Venanzi, Proc.
Chem. Soc., 170 (1961).

L. Sacconi, P. Nannelli, N. Nardi, and V. Campigli,
Inorg. Chem., 4, 943 (1965).

 

J. Chatt and B. L. Shaw, J. Chem. Soc., 1718 (1960).
R. G. Hayter, Inorg. Chem., 2, 932 (1963).

J. W. Collier, F. G. Mann, D. G. Watson, and H. R.
Watson, J. Chem. Soc., 1803 (1964).

D. A. Allen, F. G. Mann, and J. T. Millar, Chem. and
Ind., 2096 (1966).

E. C. Alyea and D. W..Meek, J. Amer. Chem. Soc., 91,
5761 (1969).

Otto Dahl, Acta Chem. Scand.,~23, 2342 (1969).
G. N. Schrauzer and P. Glockner, Chem. Ber., 91, 2451

(1964).
109

18.

19.

20.

21.

22.

23.

24.

25.
26.

27.

28.

29.

30.

31.
32.

33.

34.

35.

36.

110

K. J. Coskran, J. M, Jenkins, and J. G. Verkade, J. Amer.
Chem. Soc., 99, 5437 (1968).

B. B. Chastain, E. A. Rick,-R. L. Pruett, and H. B.
Gray, J. Amer. Chem. Soc., 99, 3994 (1968).

R. Ri o, C. Pecile, and A. Turco, gnorg. Chem., 9,
1636 (1967).

K. J. Coskran, T. J. Huttemann, and J. G. Verkade,
Advances in Chemistry Series, No. 62, American Chemical
Society, Washington, D.C., 1967, p. 590.

K. N. Raymond and F. Basolo, gnorg. Chem., 9, 949 (1966).

J. S. Coleman, H. Petersen, Jr., and R. A. Penneman,
gnorg, Chem., 2, 135 (1965). '

B. B. Chastain, D. W. Meek, E. Billig, J. E. Hix, Jr.,
and H. B. Gray, gnorg. Chem., 1, 2412 (1968).

J. Zemann, Z. Anorg. Allg, Chem., 324, 241 (1963).

J. K. Stalick and J. A. Ibers, gnorg. Chem., 9, 1084
(1969).

J. K. Stalick and J. A. Ibers, Inorg. Chem., 9, 1090
(1969).

J. A. Betrand and D. L. Plymale, Inorg. Chem., 9, 879
(1966).

E. J. Riedel and R. A. Jacobson, Nucl. Sci. Abst., 99,
27936 (1967).

K. N. Raymond, P. W. R. Cornfield, and J. A. Ibers,
gnorg. Chem., Z, 1362 (1968).

L. Sacconi, J. Chem. Soc., 9, 248 (1970).

M. J. Norgett, J. H. M. Thornley, and L. M. Venanzi,
J. Chem. Soc.,.g, 540 (1967).

J. W; Dawson, H. B. Gray, J. E. Hix, Jr., J. R. Preer,
and L. M. Venanzi, J. Amer. Chem._§9c., 99, 2979 (1972).

J. R. Preer and H. B. Gray, J. Amer. Chem. Soc., 99,
7306 (1969).

P. Rigo, G. Guastalla, and A. Turco, ;norg3 Chem., 9,
375 (1969).

C. A. Tolman, J. Amer. Chem.§gc., 99, 2956 (1970).

 

 

37.

38.

39.

40.

41.

42.

43.
44.
45.

46.

47.
48.

49.

50.

51.

52.

53.

54.

55.

111

K. A. Jensen and Otto Dahl, Acta Chem. Scand., 22,
1044 (1968) . 1...... ““

Linda.M. Haines, gnorg, Chem., 12, 1685 (1971).

w; C. Fernelius and J. J. Barbage, gnorg. Syn., 2,
227 (1946). "’

K. A. Jensen, P. H. Nielsen, and C. T. Pedersen, Acta
Chem. Scand., 11, 1115 (1963).

L. Maier, Helv. Chim. Acta, 39, 2667 (1963).

A. B. Burg and P. J. Slota, Jr., J. Amer. Chem. Soc.,
99, 1107 (1958).

G. W. Parshall, J. Inorg. Nucl. Chem., 12, 372 (1960).
K. Starke, J. Inorg. Nucl. Chem., 11, 77 (1959).

R. H. Lee, E. Griswold, and J. Kleinberg, gnorg, Chem.,
g, 1278 (1964).

Varian Associates, “High Resolution NMR Spectra
Catalog," National Press, 1963, p. 1,4.

D. F. Evans, J. Chem. Soc., 2003 (1959).

T. J. Huttemann, Jr., B. M. Foxman, C. R. Sperati, and
J. G. Verkade, Inorg. Chem., 4, 950 (1965).

G. S. Benner, W. E. Hatfield, and D. W..Meek, Inorg.
Chem., Q, 1544 (1964).

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

R. S. vinal and L. T. Reynolds, Tnorg. Chem., §, 1062
(1964).
A. A. Orio, B. B. Chastain, and H. B. Gray, £gprg: Chim.
Acta, Q, 8 (1969).

J. Chatt, G. A. Gamlen, and L. E. Orgeo, J. Chem. Soc.,
1047 (1959).

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

J. B. Fackler, Jr., J. A. Fetchin, J. Mayhew, W. C.
Seidel, T. J. Swift, and M, weeks, J. Amer. Chem. Soc.,
2.1.. 1941 (1969 ) .

56.
57.

58.

59.

60.
61.

62.
63.

64.

65.
66.
67.

68.

69.
70.

71.

72.

73.
74.

75.

112
A. J. Deeming and B. L. Shaw, J. Chem. Soc., 597 (1969).

D. Dows, A. Haim, and W; K. Wilmarth, J. Inorg. Nucl.
Chem., 21, 33 (1961).

P. B. Ogilvie, J. M,. Jenkins, and J. G. Verkade, J.
Amer. Chem. Soc., 92,1916 (1969), and references-
therein.

R. K. Harris, Can. J. Chem., 22, 2275 (1964).

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

A. R. Cullingworth, A. Pidcock, and J. D. Smith,
Chem. Comm., 89 (1966).

W} McFarlane, Chem. Comm., 58 (1967).

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

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

R. Engel, Chem. Comm., 133 (1970).

L. Frankel, J. Mol. Spect., 22, 273 (1969).

R. Engel and A. Jung, J. Chem. Soc., 2, 1761 (1971).
F. A. Cotton and G. Wilkinson, “Advanced Inorcanic
Chemistry,“ second edition, New York: Interscience

Publishers, 1966, p. 634.

J. p. Fackler, Jr., :norga Chem., 9, 2625 (1970).

I. M. Kolthoff and P. J. Elving, ”Treatise on Analytical
Chemistry," Part I, Vol. 4, New York: Interscience
Publishers, 1963, p. 1778.

R. S. Berry, J. Chem. Phys., 32, 933 (1960).

C. E. Jones and K. J. Coskran, Inorg. Chem., 10, 1536
(1971).

H. Noth and H. J. Vetter, Chem. Ber., 96, 1109 (1963).

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

C. E. Jones and K. J. Coskran, gnorg. Chem., 12, 1664
(1971).

113

76. J. F. Nixon and J. R. Swain, J. Orgonometal. Chem.,

77. F. G. A. Stone, Chem. Revs., 66, 101 (1958).

78. T. D. Coyle and F. G. A. Stone, grggress in Boron
Chemistry, 1, 135 (1964).

79. W. A. G. Graham and F. G. A. Stone, J. Inorg. Nucl.
Chem., 6, 164 (1956).

80. M. A. Fleming, R. Wyma. and R. C. Taylor, Abstracts,
Symposium on Molecular Structure and Spectroscopy,
Columbus, Ohio, June 1962.

81. T. D. Coyle, H. D. Kaesz, and F. G. A. Stone, J. Amer.
Chem. Soc., 61, 2989 (1959).

82. L. C. Cusachs and J. R. Linn, J. Chem. Phys., 66, 2919
(1967).

83. A. H. Cowley and M. C. Damasco, J, Amer. Chem. Soc.,
66, 6815 (1971).

84. A. B. Burg and H. J. Schlesinger, g. Amer. Chem. Soc.,
§g, 780 (1937).

85. R. W. Par and T. C. Bissot, J. Amer. Chem. Soc., Z6,
1524 (195r).

86. A. B. Burg, Rec. Chem. Progr., 16, 159 (1954).

87. R. W. Rudolph, Ph.D. Dissertation, University of.Michi-
gan, Ann Arbor, Michigan (1966).

88. R. W. Rudolph and R. W. Parry, J. Amer. Chem. Soc., 66,
1621 (1967).

89. D. E. Young, G. E. McAchran, and S. G. Shore, J. Amer.
Chem. Soc., 66, 4390 (1966).

 

90. F. Hewitt and A. K. Holliday, J. Chem. Soc., 530 (1953).
91. T. Reetz, J. Amer. Chem. Soc., 66, 5039 (1960).

92. R. A. Baldwin and R. M. Washburn, J. Org. Chem., 26,
' 3549 (1961).

93. T. Reetz and B. Katlaksky, J. Amer. Chem. Soc., 66,
5036 (1960).

94. ,M. A. A. B69 and H. C. Clark, Can. J. Chem., 66, 119
(1960).

95.

96.

97.

98.
99.

100.
101.
102.

103.

104.

105.
106.

107.

108.

109.

110.
111.

112.

114.

114

M. A. A. Beg and H. C. Clark, Can. J. Chem., 66, 283
(1962).

‘W. L. Jolly, "Synthetic Inorganic Chemistry," Englewood
Cliffs, New Jersey: Prentice Hall, Inc., 1960, p. 158-
160.

C. H. Bushweller, H. Beall, M. Grace, W. J. Dewkett,
and H. S. Bilofsky, J. Amer. Chem. Soc., 66, 2145 (1971).

H. L. Hodges and R. W. Rudolph, private communication.

R. W. Rudolph and C. W. Schultz, J. Amer. Chem. Soc.,
66, 6821 (1971).

R. Foester and K. Cohn, Inorg. Chem., in press.
H. c. Brown, J. Chem. Soc., 1248 (1956).

W. A. Hart and H. H. Sisler, Inorg. Chem., 3, 617
(1964). ,3

A. H. Brittain, J. E. Smith, P. L. Lee, K. Cohn, and '
R. H. Schwendeman, J. Amer. Chem. Soc., 66, 6772 (1971).

E. G. Codding, C. E. Jones, and R. H. Schwendeman,
private communication.

H. A. Bent, Chem. Revs.,66, 275 (1961).

N. Muller and D. E. Pritchard, J. Chem. Phys., 66,
768 (1959).

L. S. Bartell and L. O. Brockway, J. Chem. Phys., 66,
512 (1960).

E. B. Whipple, T. H. Brown, T. C. Farrar, and T. D.
Coyle, J. Chem. Phys., 66, 1841 (1965).

J. H. Letcher and J. R. Van Wazer, J. Chem. Phys., 66,
815 (1966).

C. A. Tolman, J. Amer. Chem. Soc., 66, 2953 (1970).

L. S. Meriwether and J. R. Amer. Chem. Soc.,

§§, 3192 (1961).

Leto, J.
A. Saeka and C. P. Slichter, J. Chem. Phys., 66, 26
(1954).
K. J. Packer, J. Chem. Soc., 960 (1963).

D. R. Armstrong and P. G. Perkins, Chem. Comm., 377
(1965).

115.

116.

117.

118.

115

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

H. J. Coerver and C. Curran, J. Amer. Chem. Soc., 66,
3522 (1958).

I. W. Stoltz, G. R. DObson, and R. K. Sheline, Inorg.
Chem., 6, 323 (1963).

G. Jugie and J. P. Laurent, Bull. Soc. Chim. Fr.,
838 (1970).

302

59

lfi‘L’sIMW 0314

 

W