Vibrational and solid state (CP/MAS) 31P NMR spectroscopic studies of bis(trimethylphosphine)gold(I) halides

Article abstract of DOI:10.1016/S0022-2860(99)00204-5

The 1:2 AuX/PMe₃ (X = Cl,Br,I) complexes have been prepared and studied by infrared, Raman and solid-state cross polarisation magic angle spinning (CP MAS) ³¹P NMR spectroscopy. The new complex bis(trimethylphosphine)iodogold(I) [AuI(PMe₃)₂] was prepared by the reaction of [AuI(PMe₃)] with PMe₃ in dimethylformamide solution. A new method for the preparation of the corresponding bromide complex [AuBr(PMe₃)₂], from [NBu₄][AuBr₂] and PMe₃, is described. The IR, Raman and CP MAS ³¹P NMR spectra of the X = Cl, Br compounds are essentially identical, whereas those for X = I show differences that are suggestive of relatively minor differences in the crystal structure relative to the X = Cl, Br compounds. The latter compounds exist as the ionic species [Au(PMe₃)₂]⁺X^-, and the X = I complex is also essentially ionic, but with a possible weak Au···I interaction; this conclusion is supported by the ¹⁹⁷Au Mossbauer spectrum of this compound. The solid-state CP MAS ³¹P NMR spectra of the chloride and bromide complexes consist of doublets due to the presence of ¹J(¹⁹⁷Au-³¹P) coupling. This study is the first in which such coupling is observed for the P-Au-P coordination environment. The ¹J(¹⁹⁷Au-³¹P) coupling constants are estimated from the doublet splittings.

Full text of DOI:10.1016/S0022-2860(99)00204-5

Journal of Molecular Structure 516 (2000) 263–272
www.elsevier.nl/locate/molstruc
31
Vibrational and solid state (CP/MAS) P NMR spectroscopic
studies of bis(trimethylphosphine)gold(I) halides
a
a,
b
*
E.N. de Silva , G.A. Bowmaker , P.C. Healy
aDepartment of Chemistry, University of Auckland, Private Bag 92019, Auckland, New Zealand
bSchool of Science, Griffith University, Nathan, Queensland 4111, Australia
Received 4 March 1999; received in revised form 23 April 1999; accepted 23 April 1999
Abstract
The 1:2 AuX/PMe3 (X ˆ Cl,Br,I) complexes have been prepared and studied by infrared, Raman and solid-state cross
polarisation magic angle spinning (CP MAS) 31P NMR spectroscopy. The new complex bis(trimethylphosphine)iodogold(I)
[AuI(PMe3)2] was prepared by the reaction of [AuI(PMe3)] with PMe3 in dimethylformamide solution. A new method for the
preparation of the corresponding bromide complex [AuBr(PMe3)2], from [NBu4][AuBr2] and PMe3, is described. The IR,
Raman and CP MAS 31P NMR spectra of the X ˆ Cl, Br compounds are essentially identical, whereas those for X ˆ I
show differences that are suggestive of relatively minor differences in the crystal structure relative to the X ˆ Cl, Br
compounds. The latter compounds exist as the ionic species [Au(PMe3)2]ϩXϪ, and the X ˆ I complex is also essentially
197
…
¨
ionic, but with a possible weak Au I interaction; this conclusion is supported by the Au Mossbauer spectrum of this
compound. The solid-state CP MAS 31P NMR spectra of the chloride and bromide complexes consist of doublets due to the
presence of J(197Au–31P) coupling. This study is the first in which such coupling is observed for the P–Au–P coordination
1
environment. The 1J(197Au–31P) coupling constants are estimated from the doublet splittings. ᭧ 2000 Elsevier Science B.V. All
rights reserved.
Keywords: Gold; Trimethylphosphine; Solid-state NMR; Infrared; Raman
…
1. Introduction
Au Au interaction is intermolecular rather than
intramolecular [4–7,10,11], and these include the
1:1 gold(I) halide/trimethylphosphine adducts
[AuX(PMe3)] [4,10,11]. Resulting from the relative
structural simplicity of the PMe3 molecule, these
compounds are of fundamental importance in the
study of physical properties of gold(I) phosphine
complexes. For example, they are of particular impor-
tance in the understanding of the vibrational spectra of
such complexes, and an analysis of the infrared and
Raman spectra of [AuX(PMe3)] was reported earlier
[12]. Recently, a detailed study on the low frequency
vibrations of these compounds has been reported, and
Although gold(I) complexes with phosphine
ligands are well known [1–3], recently there has
been an increased interest in this type of compound
because many of them show unexpected interactions
between the closed-shell d10 metal centres [4–11].
There are a number of complexes wherein the
* Corresponding author. Tel: ϩ 64-9-373-7599; fax: ϩ 64-9-
373-7422.
E-mail address: ga.bowmaker@auckland.ac.nz
(G.A. Bowmaker).
0022-2860/00/$ - see front matter ᭧ 2000 Elsevier Science B.V. All rights reserved.
PII: S0022-2860(99)00204-5

264
E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
the n (AuP), n (AuX) and d (PAuX) modes were
assigned in the range of 220–200, 315–160 and
100–75 cmϪ1, respectively [10].
further investigate the dependence of splitting on the
nature of the phosphine ligand.
Gold(I)halide/triphenylphosphine (PPh3) com-
plexes of 1:2 and 1:3 stoichiometry have been
known for some time. Despite the strong tendency
for gold(I) to display linear two-coordination in its
complexes, these compounds [AuX(PPh3)2] (X ˆ
Cl,Br,I) and [AuCl(PPh3)3] show trigonal planar and
tetrahedral coordination, respectively, with the
halogen atom included in the coordination sphere
[13,14]. By contrast, AuX/PMe3 complexes of the
above type are apparently limited to 1:2 stoichiometry
[4,15], and the crystal structure of a chloroform
solvate of the X ˆ Cl complex shows that it is the
ionic compound [Au(PMe3)2]Cl, containing the two-
coordinate [Au(PMe3)2]ϩ complex and separate ClϪ
ions [4]. It was proposed that the X ˆ Br complex
has the same structure as the chloride [15], but this has
not yet been verified, and the X ˆ I complex has not
yet been reported. We report here the preparation of
this complex, an IR and Raman study of a complete
2. Experimental
2.1. Materials
Trimethylphosphine was prepared by a literature
method [25]. Chloroauric acid HAuCl4 (Univar) was
used without further purification. Chloro(trimethyl-
phosphine)gold(I), [AuCl(PMe3)], was prepared
from chloroauric acid and PMe3 by the method
given in Ref. [10]. Iodo(trimethylphosphine)gold(I)
[AuI(PMe3)], was prepared from [NBu4][AuI2] and
PMe3 by following the method outlined in Ref. [10].
Tetra-n-butylammonium diiodoaurate(I), [NBu4][AuI2],
was prepared from [NBu4]I, I2 and Au metal by a
method described in Ref. [24]. Tetra-n-butylammo-
nium dibromo- and dichloroaurate(I), [NBu4][AuX2]
(X ˆ Br,Cl), were prepared according to previously
described methods [26,27]. All syntheses were carried
out in a dry nitrogen atmosphere.
series of unsolvated 1:2 AuX/PMe3 (X ˆ Cl,Br,I)
197
¨
Au Mossbauer spectrum of
complexes, and the
the iodide complex.
2.2. Synthesis of bis(trimethylphosphine)gold(I)
chloride, [Au(PMe3)2]Cl
These compounds are also of interest in that they
provide further examples of a nuclear spin system in
which a spin I ˆ 1/2 nucleus (31P) is coupled to a spin
I Ͼ 1/2 nucleus (197Au, I ˆ 3/2) [16–21]. It has been
shown earlier that the CP MAS 31P NMR spectra
of [AuX(PMe3)], [AuX(PPh3)] and [HgX(PPh3)]ϩ
exhibit doublets as a result of spin–spin coupling to
the I ˆ 3/2 197Au and 201Hg nuclei [10,22,23]. Obser-
vation of scalar coupling in these cases is unexpected,
as it should be averaged to zero by the rapid spin–
lattice relaxation associated with these strongly
quadrupole coupled metal nuclei. All of these cases,
wherein coupling has been observed to date, involve
two-coordinate complexes containing one phosphine
and one halide ligand; hence complexes prepared in
the present study provide an opportunity to examine
whether this splitting also exists in complexes with
two phosphine ligands attached to the metal. It has
been found that the observation of this coupling also
depends on the nature of the ligand, as splitting was
not observed in the case of [AuX(tmpp)] and
[Au(tmpp)2]X [tmpp ˆ tris(2,4,6-trimethoxyphenyl)-
phosphine] [24]. Yet another aim of this study was to
[AuCl(PMe3)] (0.21 g, 0.68 mmol) and acetone
(6 ml) were placed in a 10 ml Schlenk tube. PMe3
(0.07 ml, 0.68 mmol) was added to this mixture with
stirring, whereupon heat was evolved and a white
precipitate formed. After about 3 hours, n-hexane
(3 ml) at 40–50ЊC was added. The Schlenk tube was
cooled in an ice bath and the solution stirred for
30 min. The resulting white powder was collected
by vacuum filtration and stored under N2. Yield
0.1540 g (59% based on Au). m.p. 229–230ЊC
(164–167ЊC [15]). Anal. Found: C, 18.59; H, 4.72.
Calculated for C6H18AuClP2: C, 18.75; H, 4.72%.
2.3. Synthesis of bis(trimethylphosphine)gold(I)
bromide, [Au(PMe3)2]Br
[Bu 4N][AuBr2] (1.717 g, 2.86 mmol) was dissolved
in dimethylformamide (6.5 ml) in a 25 ml Schlenk
tube. To this solution PMe3 (0.32 ml, 3.1 mmol) was
added dropwise with stirring. After a few minutes the
clear solution became cloudy and a fine white preci-
pitate formed. The product was collected by vacuum

E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
265
Fig. 1. STA trace for [Au(PMe3)2]Cl. The endotherms at 156.2 and 231.4 are assigned to loss of 1 mol PMe3 and the melting point of the 1:1
complex, respectively.
filtration, washed with a minimum of chilled
dimethylformamide, dried and stored under N2.
Yield 0.8256 g (67% based on Au). m.p. 190–195ЊC
(186ЊC [15]). Anal. Found: C, 17.00; H, 4.40. Calcu-
lated for C6H18AuBrP2: C, 16.80; H, 4.20%.
purged bench with a KBr beam splitter, and uncooled
DTGS (deuterotriglycine sulphate) detector. Far-
infrared spectra were recorded at 4 cmϪ1 resolution
at room temperature as pressed Polythene discs on a
Digilab FTS-60 Fourier-transform spectrometer
employing an FTS-60V vacuum optical bench with
a 6.25 mm Mylar film beam splitter (75–500 cmϪ1
)
2.4. Synthesis of bis(trimethylphosphine)gold(I)
iodide, [Au(PMe3)2]I
or a 12.5 mm Mylar film beam splitter (40–
250 cmϪ1), a mercury lamp source and a pyroelectric
TGS (triglycine sulphate) detector. Data acquisition
and processing was performed using Win-IR software
by Bio-Rad. Raman spectra were recorded at 4.5 cmϪ1
resolution using a Jobin-Yvon U1000 spectrometer
equipped with a Spectrolab Laserspec III Pre-mono-
chromator and a cooled photomultiplier (RCA
C31034A) detector. The samples were presented in
the sample bottles, or on a steel plate with the sample
mounted, and excited with ϳ20 mW of 514.5 nm
[AuI(PMe3)] (0.1786 g, 0.45 mmol) was dissolved
in deoxygenated dimethylformamide (3 ml) in a 25 ml
Schlenk tube with heating. To this solution PMe3
(0.05 ml, 0.45 mmol) was added, dropwise with stir-
ring. The solution was reduced to ϳ 2 ml in volume,
and cooled in an ice bath. The resulting white preci-
pitate was filtered off and washed with a minimum of
dimethylformamide. Yield 0.142 g (67% based on
Au). m.p. 143–147ЊC. Anal. Found: C, 15.04; H,
3.81. Calculated for C6H18AuIP2: C, 15.14; H, 3.81%.
radiation from
a Spectra-Physics model 2016
argon–ion laser. Data acquisition and processing
was performed using Yvon Prism software. Solid
state CP MAS 31P NMR spectra were recorded at
ambient temperature on a Varian Unity-400 spectro-
meter at 161.93 MHz. Single contact times of 2 ms
were used with a proton pulse width of 6.5 ms, a
2.5. Spectroscopy
Mid-infrared spectra were recorded at 4 cmϪ1 reso-
lution at room temperature as CsI discs and as Nujol
mulls between CsI plates on a Digilab FTS-60
Fourier-transform spectrometer employing a nitrogen

266
E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
Table 1
Wavenumbers (cmϪ1) and assignments of the low-frequency bands in the IR and Raman spectra for [Au(PMe3)2]ϩXϪ
[Au(PMe3)2]Cl
IR
[Au(PMe3)2]Br
IR
[Au(PMe3)2]I
IR
Assignment
Raman
Raman
Raman
370
371
354
278
270
250
215
203
188
160
369
354
278
268
251
214
206
187
161
d(PC3)s ϩ n(AuP)s
d(PC3)s ϩ n(AuP)a
d(PC3)a
d(PC3)a
d(PC3)a
n(AuP)s ϩ d(PC3)s
r(PC3)
r(PC3)
r(PC3)
r(PC3)
353
280
273
250
350
279
271
250
347
267
272
202
188
164
199
175
152
206
189
163
150
110
85
206
187
169
154
106
82
110
107
99
71
45
98
d(PAuP)
d(PAuP)
n(AuI)
Lattice
63
50
62
47
79
48
42
proton decoupling field of 60 kHz and a recycle delay
time of 45 s. The samples were packed in Kel-F
inserts within silicon nitride rotors and spun at a
speed of 5 kHz at the magic angle. A total of 32
free induction decays were collected and transformed
with an experimental line broadening of 10 Hz. All
31P chemical shifts were externally referenced to 85%
H3PO4 via solid triphenylphosphine (d ˆ Ϫ9:9 ppm†.
the gravimetric plot in this region showed a weight
decrease of 19.2%, compared with a 20% decrease
expected for the loss of 1 mol of PMe3. The sharp
endotherm at 231.4ЊC corresponds to the m.p. of the
1:1 complex (230–231ЊC [12]). Above 230ЊC the
gravimetric plot indicates the beginning of loss of
the second mole of ligand. Apparently the melting
point of 164–167ЊC [15] corresponds to the decom-
position of [AuCl(PMe3)2] to [AuCl(PMe3)] and
PMe3.
2.6. Thermal analysis
Simultaneous thermal analysis (STA) involving
differential thermal analysis (DTA) and thermogravi-
metric analysis (TGA) were carried out on a Polymer
Laboratories STA 1500. Calibration was done using
an indium metal standard. A 10–15 mg sample was
used in an aluminium crucible under a nitrogen atmo-
sphere at a heating rate of 10ЊC/min from 22 to 260ЊC.
3.2. Vibrational spectra
The far-IR and low-frequency Raman spectra of
[AuX(PMe3)2] (X ˆ Cl, I) are shown in Figs 2
and 3 and the wavenumbers and assignments of
the low-frequency bands are given in Table 1. The
(X ˆ Cl, Br) complexes have ionic structures
[Au(PMe3)2]ϩXϪ, as will be shown in the discussion
of their vibrational spectra below. For X ˆ I the
complex also seems to be ionic, but with a possible
3. Results and discussion
…
weak Au I interaction. The spectrum of the bromide
3.1. Thermal analysis
is not shown because it is essentially identical to that
of the chloride. The spectra in the region above
500 cmϪ1 are due to the bending, rocking and
stretching modes of the coordinated PMe3 ligand.
These bands do not vary much in intensity or position
STA was carried out because the melting point of
[Au(PMe3)2]Cl differed considerably from the litera-
ture value [15]. The STA curves are shown in Fig. 1.
Between 92 and 167ЊC there is a broad endotherm and

E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
267
Table 2
Calculated and experimental vibrational wavenumbers (cmϪ1) for [Au(PMe3)2]ϩClϪ
Calculated (D3h symmetry)
Experimental (C2v symmetry)
Experimental mean
749
Mode assignment
730
720
680
367
354
A0
A2
E0
749
Al ϩ Bl
n(PC3)a
00
00
00
679
371
353
280
273
250
215
206
189
163
150
110
85
Al ϩ B2
A1
B1
B2
A1
B1
A1
A1
679
371
353
n(PC3)s
A0
A2
d(PC3)s ϩ n(AuP)s
d(PC3)s ϩ n(AuP)a
d(PC3)a
d(PC3)a
d(PC3)a
n(AuP)s ϩ d(PC3)s
r(PC3)
r(PC3)
r(PC3)
r(PC3)
277
252
211
200
E0
A2
A0
E0
277
250
217
177
B2
84
E0
A1
B2
98
d(PAuP)
d(PAuP)
when compared with the previously published data for
[AgI(PMe3)]4 [28]. The assignments in the low-wave-
number region are made with reference to the recently
studied [AuX(PMe3)] series [10].
with those in[AuCl(PMe3)2]·2CHCl3 [4]. The spectra
of both [Au(PMe3)2]X (X ˆ Cl, Br) are nearly iden-
tical; this is due to both complexes existing as ionic
species, so the calculations are relevant to both
complexes. The calculated frequencies were matched
to the experimental values with lowered C2v
symmetry, a subgroup of D3h, taking into account
the effects of this symmetry lowering, which are (a)
the appearance of inactive fundamentals, [A ! 10(R)
(A ! 1(IR,R) and A ! 200(IR)(B ! 1(IR,R)], and (b)
the splitting of degenerate modes, [E0(IR,R) (A !
1(IR,R) ϩ B ! 2(IR,R)]. The final refined Au–P
force constant is 244 N mϪ1, which is similar in
magnitude to the Tl–C force constant in [Tl(CH3)2]ϩ
(220 N mϪ1) [29].
For each symmetric internal mode of PMe3, e.g.
d(PC3)s (A1 in local C3v symmetry), there are two
modes in the complex, found at different frequencies
due to combinations with n(AuP)s or n(AuP)a. The
d(PC3)s ϩ n(AuP)a mode is found at 353, 354 cmϪ1
in the IR and Raman, respectively, for the chloride
complex and 350 and 352 cmϪ1, respectively, for
the bromide complex. The d(PC3)s ϩ n(AuP)s mode
is only present in the Raman spectrum, found at 371
and 369 cmϪ1 for the chloride and bromide, respec-
tively. The d(PC3)a mode (E in local C3v symmetry)
should show four modes in the complex but in this
case only three are seen. These d(PC3)a bands are
found at 280 (shoulder), 273, 250 (sh) and 278, 270
(sh), 250 cmϪ1 in the IR and Raman, respectively, for
For the complexes [Au(PMe3)2]X (X ˆ Cl, Br)
there are too many bands in the low-wavenumber
region to permit assignment based on idealised D3h
symmetry. This is most likely due to a lowering of
symmetry of the complex. We have assumed C2v
symmetry, as it was found in the crystal structure of
[AuCl(PMe3)2]·2CHCl3 [4]. As most of the bands are
caused by internal vibrations of PMe3 molecules,
assignments have been given in terms of the local
descriptions for the PMe3 molecule, as was done for
all [AuX(PMe3)] complexes, except the external
modes n(AuP) and d(PAuP). Normal-coordinate
analysis has been done to support the vibrational
assignments. The comparison between the calculated
frequencies and the vibrational spectra frequencies for
[Au(PMe3)2]Cl is given in Table 2. The normal-coor-
dinate analysis was carried out on [Au(PC3)2]ϩ
assuming the idealised D3h symmetry, where a
combined atomic mass of 15.04 was used for the
methyl groups. The initial force field was approxi-
mated with reference to previously published results
for [Tl(CH3)2]ϩ, of D3h symmetry [29]. The
[Tl(CH3)2]ϩ study involved scaled Hartree–Fock
force field calculations and normal-mode analysis to
obtain vibrational frequencies. Bond lengths and
angles for [Au(PC3)2]ϩ were estimated by comparison

268
E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
Fig. 2. Far-IR spectra of [Au(PMe3)2]X: (a) X ˆ Cl; (b) X ˆ I.
Fig. 3. Low-wavenumber Raman spectra of [Au(PMe3)2]X: (a) X ˆ
Cl; (b) X ˆ I.
the chloride complex, and similarly for the bromide
complex 279 (sh), 271, 250 and 278, 268 (sh),
251 cmϪ1. These assignments are in line with a
number of complexes of PMe3 with gold(I), silver(I)
or copper(I) halides [10,30]. The frequency of d(PC3)a
is essentially unchanged in similar complexes
mentioned above because no mixing is possible on
symmetry grounds, whereas the frequency of
d(PC3)s is significantly higher than for the free ligand
(305 cmϪ1) because of the mixing with n(AuP)s,a
modes. Related complexes [AuX(PMe3)] showed a
similar shift for the d(PC3)s mode, a significant level
of mixing with v(AuP) has been shown by normal-
coordinate analysis [10]. In the present study the
calculated d(PC3)s ϩ n(AuP) frequency at 367 cmϪ1
is close to the mean experimental frequency of
364 cmϪ1 for the split d(PC3)s ϩ n(AuP)s,a modes.
Additional weak bands at 381 and 314 cmϪ1 are
d(PC3)s and d(PC3)a modes due to trace amounts of
[AuCl(PMe3)] impurity [10]. In this study n(AuP)s ϩ
d(PC3)a are assigned at 215 and 214 cmϪ1 in the
Raman spectra of the chloride and bromide
complexes, respectively. These assignments are
similar to the [AuX(PMe3)] (X ˆ Cl, Br) complexes
at 220 and 209 cmϪ1, respectively. The complexes
[AuX(PPh3)] and [AuX(PPh3)2], yield n(AuP) in the
range 160–180 cmϪ1 [14,31]. The shift to higher
frequency for PMe3 is expected as PPh3 is replaced
by the lighter PMe3. Another reason is that a more
positive charge in the cationic complex permits a
greater degree of s-donation from the stronger base
PMe3. Four modes are expected for the r(PC3)
rocking as this is an asymmetric internal mode (E in
local C3v symmetry). The r(PC3) modes are assigned
at 206, 189, 163, 150 and 206, 187, 169, 154 cmϪ1 in
the IR for the chloride and bromide complexes,
respectively. In the Raman spectrum the r(PC3)

E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
269
107 cmϪ1 in the Raman for the chloride and bromide
complexes, respectively. As the d(PAuP) mode is
doubly degenerate (E0 under D3h symmetry), splitting
occurs. The [Au(PMe3)2]X species is essentially two-
coordinate as in [AuX(PMe3)]. In the latter bromide
complex two bands in the IR spectrum at 61 and
88 cmϪ1 are assigned to d(PAuBr) [10]. The
d(PAuP) assignments in this study are at higher
frequency compared to d(PAuBr), as Br is replaced
by the lighter PMe3 ligand.
The vibrational spectra of [Au(PMe3)2]I show some
significant differences relative to the chloride and
bromide species (Figs.2 and 3; Table 1), and these
…
are attributed to the presence of a weak Au I inter-
action. The most obvious differences occur in the
region below 150 cmϪ1. Thus, the two strong
d(PAuP) IR bands around 100 cmϪ1 in the chloride
and bromide are not present in the iodide, but a rela-
tively weak band at 99 cmϪ1 is assigned to this mode.
The bands found at 79 and 48 cmϪ1 in the Raman for
the iodide are almost coincident between the three
complexes, except for the iodide complex where a
shoulder is found on the lower frequency band at
42 cmϪ1. The bands at approximately 50 cmϪ1 in
[Au(PMe3)2]X complexes are assigned as lattice
vibrations. These bands are assigned at similar
frequencies in [AuX(PMe3)] [10]. For the iodide
complex the band at 71 cmϪ1 in the IR is assigned
as v(AuI), although this would be mixed with
d(PAuP), which would be lowered in frequency by
Fig. 4. CP MAS 31P NMR spectra of [Au(PMe3)2]X: (a) X ˆ Cl, (b)
X ˆ Br; (c) X ˆ I.
modes are found at 203, 188, 160 and 206, 187,
161 cmϪ1. The extent of splitting in the local E
modes for PMe3 which occurs as a consequence of
the ligand bonding to the Au atom would be expected
to increase with decreasing frequency of the mode, in
agreement with the fact that the splitting of r(PC3) is
greater than d(PC3)a. The d(PAuP) bands are found at
110, 85, 106 and 82 cmϪ1 in the IR and 110 and
…
the Au I interaction. For the same reason, n(AuP)
should be lower in the iodide, in agreement with an
observed reduction of ca. 15 cmϪ1 (Table 1). Further
…
support for the hypothesis of a weak Au I bond
can be obtained from recent structural data for the
corresponding tricyclohexylphosphine complexes
[AuX(Pcy3)2] (X ˆ Cl,Br,I). The X ˆ Cl complex is
Table 3
The CP MAS 31P NMR chemical shifts and coupling constants J/Hz (1J(197Au–31P) or 1J(201Hg–31P)) for [Au(PMe3)2]ϩXϪ and related species
Compound
X ˆ Cl
X ˆ Br
X ˆ I
Ref.
d
1J(MP)
d
1J(MP)
d
1J(MP)
[Au(PMe3)2]X
[AuX(PMe3)]
[AuX(PPh3)]
7.6
Ϫ 9.6
30
344
648
520
7.5
Ϫ 6.9
32
340
–
500
2727
1.9
Ϫ 4.4
36
–
553
410
This work
[10]
[22,23]
[23]
[HgX(PPh3)]NO3
30.5
3091
32.2
33.0
2364

270
E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
ionic, whereas the X ˆ I complex shows a weak
93 cmϪ1 [32].
such as LAuX, for which the 197Au nuclear quadru-
pole coupling constant is large [10,23]. The quadru-
pole coupling constant is also expected to be large in
the linear two-coordinate [Au(PMe3)2]X species, and
the predicted doublet splitting is observed in the X ˆ
Cl,Br compounds. The coupling constants
1J(197Au–31P) calculated from the observed spacing
of the doublets are 344 and 340 Hz ( ^ 10 Hz), and
these are compared with the values for related
…
˚
Au I interaction (Au–I ˆ 2.896 A) with n(AuI) ˆ
3.3. CP MAS 31P NMR spectra
The solid state CP MAS 31P NMR spectra of
[Au(PMe3)2]X (X ˆ Cl,Br,I) are given in Fig. 4.
The parameters obtained from the spectra are listed
in Table 3. The spectra of the chloride and the
bromide compounds consist of doublets, this being
the first reported case of doublets in a P–Au–P envir-
onment. The doublet separations are independent of
X, being 567 and 561 Hz ( ^ 20 Hz) for X ˆ Cl,Br,
respectively, consistent with a similar ionic structure
for both complexes. Doublets were also found for
[AuX(PMe3)] (X ˆ Cl,I) with splittings of 1070 and
913 Hz, respectively [10], which are larger than those
in the present study. The large nuclear quadrupole
moment of 197Au (100% natural abundance, nuclear
spin I ˆ 3/2) and the asymmetric coordination envir-
onment in these complexes would be expected to give
rise to rapid quadrupolar relaxation, so that any
1J(197Au–31P) scalar coupling would be averaged to
zero, and the signal would appear as a narrow singlet.
The MAS NMR spectra of spin I ˆ 1/2 nuclei coupled
to a quadrupolar nucleus have been discussed by
various authors [16–21]. For the case where the quad-
rupolar nucleus has spin I ˆ 3/2, the signal due to the
spin 1/2 nucleus splits into a quartet as expected, but
the relative positions of the lines in the quartet depend
on the nuclear quadrupole coupling constant x of the
spin 3/2 nucleus and on the dipolar coupling constant
D between the two nuclei. For the case where the
scalar coupling constant J is much greater than D, a
symmetrical four-line pattern is observed when x is
small [16]. This situation has been observed for the
compound [Au(dppey)2]Cl [dppey ˆ cis-1,2-bis(di-
phenylphosphino)ethylene] wherein a near-tetrahe-
dral coordination geometry about the Au atom is
assumed to result in a 197Au quadrupole coupling
constant close to zero [33]. For the opposite case
where x is very large, a mixing of the spin states of
the spin 3/2 nucleus occurs, and the quartet pattern
becomes two superimposed doublets, with a spacing
of about 1.65 times the 1J(197Au–31P) coupling
constant [10,16,21]. To date this situation has only
been observed in linear mixed ligand complexes
1
complexes in Table 3. The J(197Au–31P) values for
[AuX(PMe3)] (X ˆ Cl,I) with one phosphine ligand
are 648 and 553 Hz, respectively [10], and are consid-
erably higher than the values found in the present
study, which is in agreement with the well-established
1
trend that J(M–P) decreases with an increase in the
number of phosphine ligands bound to the metal atom
1
M [34]. Lower J(197Au–31P) values were also found
for [Au(dppey)2]ϩ at 200 Hz [33]. The decrease in
1J(197Au–31P) with increasing phosphine coordination
number is based on the assumption that the spin–spin
coupling is due to the Fermi contact interaction. Thus
the coupling constant should be proportional to 6s
electron density at the Au nucleus, which in turn is
proportional to the 6s character of the Au–P bonds.
For the two-coordinate species the bonds involve sp
hybrids, for which the Au 6s character is 0.5. The
1
decrease in J(197Au–31P) with increasing number of
phosphine ligands may be explained by “Bent’s rule”
which states that the atomic s character concentrates
in orbitals directed toward electropositive substituents
[35]. This implies that the PMe3 tends to bond with the
Au 6s orbital, whereas the halogens X tend to bond
using the Au 6p orbitals. Bent’s rule was derived from
a systematic comparison of the physical properties of
molecules and a correlation with qualitative models
such as valence bond structures and bond hybridisa-
tion [35]. Independent experimental evidence for this
197
¨
has been obtained in Au Mossbauer spectroscopic
studies of [AuClL] and [AuL(PPh3)]ϩ (L
ˆ
SMe2,PPh3,PEt3 etc.) [36]. A tentative rationalization
of the present experimental results is that in the 1:1
complexes the Au 6s orbital is involved in bonding
with only one PMe3 molecule, so the Au 6s character
of the bond is high. Whereas in the 1:2 complexes the
6s orbital must be “shared” between two PMe3 mole-
cules, so the Au 6s participation in each of the Au–P
bonds is reduced, resulting in a smaller J(197Au–31P)
1
value.

E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
271
¨
Fig. 5. Gold-197 Mossbauer spectrum of [Au(PMe3)2]I.
¨
The spectrum of [Au(PMe3)2]I shows only a broad
singlet, shifted upfield by 6 ppm towards the free
ligand value. The difference between the iodide and
the other two halides in the NMR is consistent with
the different vibrational spectra. In the [Au(tmpp)2]X
complexes all the halides exist as purely ionic species,
and as a result d was the same for all the halides [24].
The doublet-splitting phenomenon is only seen in
linear or nearly linear complexes with certain combi-
nations of ligands. For the 1:2 iodide a long range
[37]. In the latter case it was concluded that the Moss-
bauer parameters are indicative of coordination of the
iodide ligand to the gold atom [37]. In particular, a
decrease in the quadrupole splitting is expected in
going from two-coordinate [AuL2]ϩ to three-coordi-
nate [AuXL2] [37], and the fact that the quadrupole
splitting in [AuI(PMe3)2] is significantly less than
those observed for [AuL2]ϩ [PF 6]Ϫ (e.g. 10.18 and
9.43 mm sϪ1 for L ˆ PEt3 and PPh3, respectively [37])
supports the view that the iodide is coordinated to the
gold atom in this complex.
…
Au I bond would probably result in the complex
showing a significant distortion from linearity. It is
likely that the iodide complex lies between two- and
three-coordination around the metal centre. This gives
further evidence for the different structure of the
iodide species.
4. Conclusion
The new bis(trimethylphosphine) gold(I) iodide
complex was synthesised. The far-IR and Raman
spectra show significant differences between
[Au(PMe3)2]X (X ˆ Cl,Br) and X ˆ I. The
solid state 31P NMR of the chloride and bromide
consists of doublets due to 1J(197Au–31P)
coupling, which is not averaged to zero by relaxa-
tion effects. The absence of this splitting for the
iodide adds support to the proposed different
structure of this complex, which involves weak
coordination between the iodide ligand and the
gold centre.
¨
3.4. Mossbauer spectroscopy
197
¨
The
Au Mossbauer spectrum of the iodide
complex was obtained in order to gain further insight
into the coordination environment in this complex.
The spectrum is shown in Fig. 5, and yields the
following parameters: isomer shift d ˆ 3.38 mm sϪ1
,
quadrupole splitting DE ˆ 8.71 mm sϪ1. These are
similar to the parameters d ˆ 4.38 mm sϪ1, DE ˆ
8.79 mm sϪ1 obtained previously for the corre-
sponding triethylphosphine complex [AuI(PEt3)2]

272
E.N. de Silva et al. / Journal of Molecular Structure 516 (2000) 263–272
[14] P.G. Jones, G.M. Sheldrick, J.A. Muir, M.M. Muir, L.B.
Acknowledgements
Pulgar, J. Chem. Soc.: Dalton Trans. (1982) 2123.
[15] H. Schmidbaur, R. Franke, Chem. Ber. 105 (1972) 2985.
[16] E.M. Menger, W.S. Veeman, J. Magn. Reson. 46 (1982) 257.
[17] A.C. Olivieri, J. Magn. Reson. 81 (1989) 201.
[18] A.C. Olivieri, J. Am. Chem. Soc. 114 (1992) 5758.
[19] R.K. Harris, A.C. Olivieri, Prog. Nucl. Magn. Reson. Spec-
trosc. 24 (1992) 435.
[20] A.C. Olivieri, J. Magn. Reson. A 101 (1993) 313.
[21] A.C. Olivieri, Solid State Nucl. Magn. Reson. 1 (1992) 345.
[22] P.F. Barron, L.M. Engelhardt, P.C. Healy, J. Oddy, A.H.
White, Aust. J. Chem. 40 (1987) 1545.
We acknowledge support of this work by grants
from the New Zealand University Grants Committee,
the University of Auckland Research Committee and
the Australian Research Council. We thank Dr Al
Neilson for providing the trimethylphosphine which
was used in this study, Mr. Robbie Buckley for the
synthesis of [NBu4][AuBr2], Dr. Peta Harvey for
recording the solid-state 31P NMR spectra, and Prof.
[23] L.-J Baker, G.A. Bowmaker, P.C. Healy, B.W. Skelton, A.H.
White, J. Chem. Soc.: Dalton Trans. (1992) 989.
[24] L.-J Baker, R.C. Bott, G.A. Bowmaker, P.C. Healy, B.W.
Skelton, P. Schwerdtfeger, A.H. White, J. Chem. Soc.: Dalton
Trans. (1995) 1341.
¨
F.E. Wagner for obtaining the Mossbauer spectrum.
References
[25] M.L. Luetkens Jr, A.P. Sattelberger, H.H. Murray, J.D. Basil,
J.P. Fackler Jr, Inorg. Synth. 28 (1990) 305.
[26] P. Braunstein, R.J.H. Clark, J. Chem. Soc.: Dalton Trans.
(1973) 1845.
[27] R.W. Buckley, P.C. Healy, W.A. Loughlin, Aust. J. Chem. 50
(1997) 775.
[28] H.G.M. Edwards, D.W. Farwell, J. Mol. Struct. 197 (1989)
203.
[1] R.J. Puddephatt, in: G. Wilkinson (Ed.), Comprehensive
Coordination Chemistry, 5, Pergamon Press, Oxford, 1987,
p. 861.
[2] R.J. Puddephatt, The Chemistry of Gold, Elsevier,
Amsterdam, 1978.
[3] C.A. McAuliffe, W. Levason, Phosphine, Arsine and Stibine
Complexes of the Transition Elements, Elsevier, Amsterdam,
1979.
[29] P. Schwerdtfeger, G.A. Bowmaker, P.D.W. Boyd, D.C. Ware,
P.J. Brothers, A.J. Nielson, Organometallics 9 (1990) 504.
[30] G.A. Bowmaker, R.D. Hart, B.E. Jones, B.W. Skelton, A.H.
White, J. Chem. Soc.: Dalton Trans. (1995) 3063.
[31] G.A. Bowmaker, Effendy, J.V. Hanna, P.C. Healy, B.W.
Skelton, A.H. White, J. Chem. Soc.: Dalton Trans. (1993)
1387.
[32] G.A. Bowmaker, C.L. Brown, R.D. Hart, P.C. Healy, C.E.F.
Rickard, A.H. White, J. Chem. Soc.: Dalton Trans. (1999)
881.
[33] S.J. Berners-Price, L.A. Colquhoun, P.C. Healy, K.A. Byriel,
J.V. Hanna, J. Chem. Soc.: Dalton Trans. (1992) 3357.
[34] J.G. Verkade, J.A. Mosbo, in: J.G. Verkade, L.D. Quin (Eds.),
Phosphorus-31 NMR Spectroscopy in Stereochemical
Analysis, VCH, Deerfield Beach, FL, 1987, p. 425.
[35] V. Jonas, C. Boehme, G. Frenking, Inorg. Chem. 35 (1996)
2097.
[36] P.G. Jones, A.G. Maddock, M.J. Mays, M.M. Muir, A.F.
Williams, J. Chem. Soc.: Dalton Trans. (1977) 1434.
[37] G.C.H. Jones, P.G. Jones, A.G. Maddock, M.J. Mays, P.A.
Vergnano, A.F. Williams, J. Chem. Soc.: Dalton Trans.
(1977) 1440.
[4] K. Angermaier, E. Zeller, H. Schmidbaur, J. Organomet.
Chem. 472 (1994) 371.
[5] D.B. Dyson, R.V. Parish, C.A. McAuliffe, R.G. Pritchard, R.
Fields, B. Beagley, J. Chem. Soc.: Dalton Trans. (1989) 907.
[6] H. Schmidbaur, G. Weidenhiller, O. Steigelmann, G. Mu¨ller,
Z. Naturforsch. Teil B 45 (1990) 747.
[7] H. Schmidbaur, P. Bissinger, J. Lachmann, O. Steigelmann, Z.
Naturforsch., Teil B 47 (1992) 1711.
[8] H. Schmidbaur, W. Graf, G. Mu¨ller, Angew. Chem. 100
(1988) 437 Angew. Chem., Int. Ed. Engl. 27 (1988) 417.
[9] H. Schmidbaur, A. Stu¨tzer, P. Bissinger, Z. Naturforsch Teil B
47 (1992) 640.
[10] K. Angermaier, G.A. Bowmaker, E.N. de Silva, P.C. Healy,
B.E. Jones, H. Schmidbaur, J. Chem. Soc.: Dalton Trans.
(1996) 3121.
˚
´
[11] S. Ahrland, K. Dreisch, B. Noren, A. Oskarsson, Acta Chem.
Scand. Part A 41 (1987) 173.
[12] D.A. Duddell, P.L. Goggin, R.J. Goodfellow, M.G. Norton,
J.G. Smith, J. Chem. Soc. A (1970) 545.
[13] G.A. Bowmaker, J.C. Dyason, P.C. Healy, L.M. Engelhardt,
C. Pakawatchai, A.H. White, J. Chem. Soc.: Dalton Trans.
(1987) 1089.