X-Ray crystallographic and extended X-ray absorption fine structure studies of gold(I) complexes containing weak intermolecular interactions

Article abstract of DOI:10.1039/a707650i

The crystal and molecular structures of [Au₂{S₂CN(C₂H₄OMe)₂} ₂], [Au(PPh₃)(SCH₂CO₂H)], [Au(PPh₃)(SCMe₂CO₂H)] and [Au(PPh₃)(SCH₂CO₂Me)] have been determined. All compounds contain an approximately linear primary co-ordination sphere of ligands about the gold atom, but they differ markedly in their type of intermolecular interaction. In [Au₂{S₂CN(C₂H₄OMe)₂} ₂] the supramolecular array is dominated by an almost linear, polymeric backbone of gold atoms with alternate gold-gold contacts of 2.7902(6) (intramolecular) and 3.1572(7) A (intermolecular), whereas in [Au(PPh₃)(SCH₂CO₂H)] dimers associated through long gold-sulfur contacts of 3.131(2) A are held together in a polymeric chain by hydrogen bonding between neighbouring carboxylic acid residues. The increased steric bulk of the thiolate ligand in [Au(PPh₃)(SCMe₂CO₂H)] caused by the methyl groups on the α-carbon atom precludes association of the gold centres, but hydrogen bonding between carboxylates as in the SCH₂CO₂H compound causes dimerisation. Compound [Au(PPh₃)(SCH₂CO₂Me)] exists as a monomer with no evidence of weak intermolecular interactions. Analysis of ambient-temperature EXAFS (extended X-ray absorption fine structure) measurements on solid samples of the first two compounds yield gold-gold separations of 2.775(2) and 3.271(6) A and 4.188(15) A, respectively. Gold-sulfur separations of 2.290(1) and 3.532(8) and of 2.329(4) and 3.124(19) A, respectively, are also in good agreement with X-ray data.

Full text of DOI:10.1039/a707650i

View Article Online / Journal Homepage / Table of Contents for this issue
X-Ray crystallographic and extended X-ray absorption fine structure
studies of gold(I) complexes containing weak intermolecular
interactions
Peter Bishop,*,a Patsy Marsh,a Alan K. Brisdon,b Brian J. Brisdon*,c and Mary F. Mahonc
a Cookson Matthey Technology Centre, Sandy Lane, Yarnton, Oxford, UK OX5 1PF
b Department of Chemistry, UMIST, PO Box 88, Manchester, UK M60 1QD
c School of Chemistry, University of Bath, Bath, UK BA2 7AY
The crystal and molecular structures of [Au2{S2CN(C2H4OMe)2}2], [Au(PPh3)(SCH2CO2H)], [Au(PPh3)(SCMe2-
CO2H)] and [Au(PPh3)(SCH2CO2Me)] have been determined. All compounds contain an approximately linear
primary co-ordination sphere of ligands about the gold atom, but they differ markedly in their type of
intermolecular interaction. In [Au2{S2CN(C2H4OMe)2}2] the supramolecular array is dominated by an almost
linear, polymeric backbone of gold atoms with alternate gold–gold contacts of 2.7902(6) (intramolecular) and
3.1572(7) Å (intermolecular), whereas in [Au(PPh3)(SCH2CO2H)] dimers associated through long gold–sulfur
contacts of 3.131(2) Å are held together in a polymeric chain by hydrogen bonding between neighbouring
carboxylic acid residues. The increased steric bulk of the thiolate ligand in [Au(PPh3)(SCMe2CO2H)] caused by
the methyl groups on the α-carbon atom precludes association of the gold centres, but hydrogen bonding between
carboxylates as in the SCH2CO2H compound causes dimerisation. Compound [Au(PPh3)(SCH2CO2Me)] exists
as a monomer with no evidence of weak intermolecular interactions. Analysis of ambient-temperature EXAFS
(extended X-ray absorption fine structure) measurements on solid samples of the first two compounds yield
gold–gold separations of 2.775(2) and 3.271(6) Å and 4.188(15) Å, respectively. Gold–sulfur separations
of 2.290(1) and 3.532(8) and of 2.329(4) and 3.124(19) Å, respectively, are also in good agreement with
X-ray data.
Recent advances in gold chemistry have highlighted the flexible
metal–metal bonding modes exhibited by this element.1 Particu-
lar attention has been paid to compounds which contain weak
intermolecular metal–metal interactions with gold–gold separ-
ations typically in the range 2.5–3.5 Å, which are intermediate
between those found in the Au2 molecule at 2.47 Å, and twice
the gold van der Waals radius of 3.60 Å.2 These attractive gold–
gold interactions have been estimated to be as high as 30 kJ
molϪ1 in some compounds,3 and they are considered to be
significant in the stabilisation of hypervalent main group spe-
cies such as [C{Au(PPh3)}6]2ϩ,4 and of considerable utility in
the self assembly and crystal engineering of gold-containing
materials.5
Database analyses have revealed that gold–gold interactions
are frequently observed in formally two-covalent gold() com-
pounds containing auxiliary P- and/or S-donor ligands.6 Gold–
phosphine complexes are commonly used as reagents in
organometallic chemistry,7 and their potential as drugs for the
treatment of rheumatoid arthritis has been explored in some
detail.8 Gold thiolates have also received attention as arthritis
inhibitors and in chemotherapy,9 and have been shown recently
to possess important photophysical properties.10 Therefore fac-
tors affecting the inter- and intra-molecular arrangements in
such compounds are of considerable interest, and although
theoretical studies of weak gold–gold bonding interactions are
well developed,11 they are insufficiently advanced to be used for
predicting detailed structural information.
ous and insoluble, they are not amenable to X-ray single-crystal
analysis. In this paper we explore the use of EXAFS as an
alternative technique for this type of structural analysis, and
show that, contrary to previous expectations,12 weak gold–gold
interactions involving internuclear distances greater than 3 Å
can be determined from EXAFS signals measured at room
temperature.
The study reported herein centres on four gold() compounds,
each containing S-donor ligands which were functionalised
with oxygen-containing substituents in order to increase their
solubility in polar media. It forms part of a systematic invest-
igation into the properties of gold thiolates. Single-crystal
X-ray diffraction methods were used to reveal the differing
structural features present in [Au2{S2CN(C2H4OMe)2}2], [Au-
(PPh3)(SCH2CO2H)], [Au(PPh3)(SCMe2CO2H)] and [Au-
(PPh3)(SCH2CO2Me)], and EXAFS studies were carried out
on the first two compounds in order to compare the inter-
nuclear separations obtained from these differing techniques.
Results and Discussion
Preparative methods for gold(I) dithiocarbamates
The first gold() dialkyldithiocarbamates were prepared by
Akerstrom13 by the careful reduction of gold() by sodium
sulfite in saturated salt solution, followed by reaction with an
aqueous solution of the sodium salt of the ligand. This method,
and variations thereon, has been used to prepare a wide range
of gold() dithiocarbamate complexes.14–16 In the present study
reduction to gold() was achieved using aqueous ethyl 2-
hydroxyethyl sulfide as the reducing agent. The solution of the
gold() intermediate so formed was added to an aqueous solu-
tion of the salt K[S2CN(C2H4OMe)2] to afford solid [Au2{S2C-
N(C2H4OMe)2}2], compound 1, which was purified by extrac-
tion into chloroform. The spectroscopic features of 1 are in
accord with data reported on analogues. It has been noted
Experimental determinations of the internuclear parameters
in gold cluster chemistry have depended almost exclusively
upon single-crystal X-ray diffraction studies, and we are aware
of only one previous investigation in which EXAFS (extended
X-ray absorption fine structure) has been used very specifically
to explore weak gold–gold interactions in the solid state.12 As
many potentially interesting and useful compounds in which
such interactions might be important are polymeric, amorph-
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682
675

View Article Online
Table 1 Selected bond lengths (Å) and angles (Њ) for complex 1
Au(1)᎐S(2)
Au(1)᎐S(4)
Au(1)᎐Au(2)
Au(1)᎐Au(2Ј)
Au(2)᎐S(3)
Au(2)᎐S(1)
2.289(2)
2.292(2)
2.7902(6)
3.1572(7)
2.296(2)
2.300(2)
S(1)᎐C(1)
S(2)᎐C(1)
S(3)᎐C(8)
S(4)᎐C(8)
N(1)᎐C(1)
N(2)᎐C(8)
1.736(9)
1.742(9)
1.732(8)
1.723(9)
1.331(10)
1.323(10)
S(2)᎐Au(1)᎐S(4)
174.58(9)
90.67(6)
88.70(6)
98.44(6)
85.35(6)
Au(1)᎐Au(2)᎐Au(1Љ) 147.35(2)
S(2)᎐Au(1)᎐Au(2)
S(4)᎐Au(1)᎐Au(2)
S(2)᎐Au(1)᎐Au(2Ј)
S(4)᎐Au(1)᎐Au(2Ј)
Au(2)᎐Au(1)᎐Au(2Ј) 141.49(2)
S(3)᎐Au(2)᎐S(1)
S(3)᎐Au(2)᎐Au(1)
S(1)᎐Au(2)᎐Au(1)
S(3)᎐Au(2)᎐Au(1Љ)
S(1)᎐Au(2)᎐Au(1Љ)
C(1)᎐S(1)᎐Au(2)
C(1)᎐S(2)᎐Au(1)
C(8)᎐S(3)᎐Au(2)
C(8)᎐S(4)᎐Au(1)
N(1)᎐C(1)᎐S(1)
N(1)᎐C(1)᎐S(2)
S(1)᎐C(1)᎐S(2)
N(2)᎐C(8)᎐S(4)
N(2)᎐C(8)᎐S(3)
S(4)᎐C(8)᎐S(3)
109.1(3)
110.0(3)
107.6(3)
107.3(3)
116.9(7)
116.5(6)
126.6(5)
117.7(6)
117.3(6)
124.9(5)
173.58(9)
86.90(6)
91.68(6)
86.23(6)
98.26(6)
Fig. 1 Molecular structure of compound 1 showing the atom labelling
scheme adopted. Thermal ellipsoids are drawn at the 30% probability
level
Primed and double primed atom positions are related to corresponding
unprimed atoms by the Ϫx ϩ ¹, y Ϫ ¹, Ϫz ϩ ¹ and Ϫx ϩ ¹, y ϩ ¹,
¯
²
¯
²
¯
²
¯
²
¯
²
Ϫz ϩ ¹ symmetry operations respectively.
¯
²
Pri3C6H2S)Au}6] has been fully characterised by X-ray crystal-
lographic methods.18 The thiolates reported in this study also
exhibited very low solubilities in either water or organic solv-
ents, but both 2 and 4 dissolved sufficiently in D2O containing
1
NaOD for the collection of H NMR data, as noted in the
Experimental section.
The alcohol-soluble gold thiolates [Au(PPh3)(SCH2CO2H)] 3,
[Au(PPh3)(SCMe2CO2H)] 5 and [Au(PPh3)(SCH2CO2Me)] 7
were prepared by heating the gold thiolates 2, 4 or 6 under
reflux with 1 equivalent of triphenylphosphine in ethanol. Pre-
vious workers19–21 have prepared phosphine adducts of gold()
thiolates by an alternative route which involves the reaction of
halogeno(tertiary phosphine)gold() compounds with 1 equiv-
alent of thiol in the presence of base. The method used in this
study affords clean products in high yields and avoids any pos-
sible complications caused by using basic solutions of thiols.
1
Compounds 2, 4 and 6 were characterised by IR, H and 31P
NMR spectroscopies and elemental analyses. The 31P chemical
shift data are similar to those reported for other P᎐Au᎐S con-
taining analogues,20,21 suggesting that the phosphorus environ-
ment is only slightly dependent upon the nature of the sulfur
ligand attached to gold.
Crystals of compounds 3, 5 and 7 suitable for structure
determinations were grown from hot ethanol, hot PriOH and
ethanol respectively.
Fig. 2 Crystal packing diagram of compound 1 viewed down the c
axis of the unit cell and showing the polymeric backbone of gold atoms
in the direction of the b axis
Crystal structures
The crystal structure of compound 1 consists of two gold()
atoms in a slightly puckered eight-membered ring composed of
two gold, four sulfur and two carbon atoms (Fig. 1). Each gold
atom is co-ordinated to two sulfur atoms [Au(1)᎐S(2) 2.289(2),
Au(1)᎐S(4) 2.292(2), Au(2)᎐S(3) 2.296(2), Au(2)᎐S(1) 2.300(2)
Å], one from each bridging ligand, so achieving almost linear
co-ordination with S(2)᎐Au(1)᎐S(4) 174.58(9)Њ and S(1)᎐Au(2)᎐
S(3) 173.58(9)Њ (Table 1). In addition to a short intramolecular
gold–gold separation of 2.7902(6) Å, molecules are linked by
longer intermolecular gold–gold interactions [3.1572(7) Å], so
generating a supermolecular array dominated by a polymeric
backbone of gold atoms (Fig. 2). Both [Au2(S2CNPri2)2] and
[Au2(S2CNBu2)2] have been shown to exhibit similar structural
features,22,23 but with a much longer gold–gold intermolecular
interaction of 3.40 Å in the former complex.22
previously that the infrared-active ν(C᎐N) mode is sensitive
to both the chain length and the steric bulk of the alkyl
substituent.17 For this compound a value of 1494 cmϪ1 for this
mode is very similar to that noted for gold() dialkyldithio-
carbamates with short unbranched alkyl substituents R = Et
(1496) or Pr (1490 cmϪ1).15 Crystals of 1 suitable for a single-
crystal X-ray determination were grown from dichloro-
methane–methanol.
Gold(I) thiolates and their triphenylphosphine adducts
The homoleptic gold thiolates [Au(SCH2CO2H)] 2, [Au(SC-
Me2CO2H)] 4 and [Au(SCH2CO2Me)] 6 were prepared via
reduction of gold() as for compound 1, followed by addition
of the thiol. The compounds were formed as yellow (2 and 6)
and white (4) powders in nearly quantitative yields. Compound
2 has been cited previously9 but was not fully characterised.
The polymeric nature of most homoleptic gold() thiolates
results in poor solubility, and as a result only [{2,4,6-
The structure of compound 3 can also be described as a
supermolecular array displaying weak inter- and intra-
molecular associations, but neither involves gold–gold inter-
actions. Instead two alternative structural features are evident
676
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682

View Article Online
Fig. 5 Molecular structure of compound 5 showing the atom labelling
scheme adopted. Thermal ellipsoids as in Fig. 1
Fig. 3 Molecular structure of compound 3 showing the atom labelling
scheme used, and the interactions between neighbouring molecules via
gold–sulfur contacts around an inversion centre. Thermal ellipsoids as
in Fig. 1
Fig. 4 Crystal packing diagram of compound 3 viewed down the b axis
showing the two forms of intermolecular association along the a axis.
Hydrogen atoms, which were included in calculated positions except
for the carboxylic acid proton H(1), have been omitted for clarity. The
atom H(1) was located and refined at a fixed distance of 0.9 Å from
O(1)
Table 2 Selected bond lengths (Å) and angles (Њ) for compound 3
Au(1)᎐P(1)
Au(1)᎐S(1)
2.254(2)
2.308(2)
Au(1) ؒ ؒ ؒ S(1Ј)
S(1)᎐C(19)
3.131(2)
1.815(10)
Fig. 6 Crystal packing diagram of compound 5 showing dimer form-
ation. The carboxylic acid proton was not located
P(1)᎐Au(1)᎐S(1)
S(1)᎐Au(1)᎐S(1Ј)
178.85(8)
86.55(8)
Au(1)᎐S(1)᎐Au(1Ј)
C(19)᎐S(1)᎐Au(1)
93.45(8)
105.1(3)
Primed atom positions are related to corresponding unprimed atoms by
the Ϫx, Ϫy, Ϫz symmetry operation.
Table 3 Selected bond lengths (Å) and angles (Њ) for compound 5
Au(1)᎐P(1)
Au(1)᎐S(1)
2.2582(14)
2.2896(14)
S(1)᎐C(19)
1.848(6)
(Figs. 3 and 4, Table 2). First, neighbouring pairs of molecules
in which each gold atom is co-ordinated approximately linearly
[178.85(8)Њ] by one sulfur atom [2.308(2) Å] and one phos-
phorus atom [2.254(2) Å] interact with each other in a ‘head to
tail’ arrangement via sulfur–gold contacts around an inversion
centre [Au(1) ؒ ؒ ؒ S(1Ј) 3.131(2) Å, S(1)᎐Au(1)᎐S(1Ј) 86.55(8),
Au(1)᎐S(1)᎐Au(1Ј) 93.45(8)Њ]. Secondly, the carboxylate groups
of neighbouring molecules are hydrogen bonded [H(1)᎐O(2)
1.41(3) Å, O(1)᎐H(1)᎐O(2) 162.7Њ]. These two categories of
alternating intermolecular contacts produce a one-dimensional
polymeric backbone which dominates the supramolecular array
(Fig. 4).
P(1)᎐Au(1)᎐S(1)
179.34(5)
C(19)᎐S(1)᎐Au(1)
104.8(2)
ligand will prevent the form of association noted in compound 3
above. To investigate this, the crystal structure of [Au(PPh3)-
(SCMe2CO2H)] 5 was determined with the result illustrated in
Figs. 5 and 6. Unlike the previous two structures, the gold atom
is strictly two-co-ordinate [P(1)᎐Au(1)᎐S(1) 179.34(5)Њ], with
gold–sulfur and –phosphorus bond lengths of 2.2896(14) and
2.2582(14) Å respectively (Table 3). Dimers are formed from
hydrogen bonding between pairs of carboxylate residues.
Although the acid proton could not be located with certainty,
Molecular models indicate that on increasing the bulk of the
thiolate ligand the large steric bulk of the triphenylphosphine
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682
677

View Article Online
Fig. 7 Molecular structure of compound 7 showing the atom labelling
scheme adopted. Thermal ellipsoids as in Fig. 1
Table 4 Selected bond lengths (Å) and angles (Њ) for compound 7
Fig. 8 (a) Background-subtracted gold L(III)-edge EXAFS spectra for
Au(1)᎐P(1)
Au(1)᎐S(1)
2.265(2)
2.298(2)
S(1)᎐C(19)
1.848(12)
compound 1 (——, experimental × k3; –––, curved-wave theory ×
k3), and (b) corresponding Fourier-transform magnitudes in arbitrary
units (——, experimental; –––, theoretical)
P(1)᎐Au(1)᎐S(1)
178.96(8)
C(19)᎐S(1)᎐Au(1)
101.3(3)
the O(1) ؒ ؒ ؒ O(2) separation at 2.605(6) Å in 5 is very similar to
that found in compound 3 and in [Au(CNBu)(SC6H4CO2H)].24
The lack of associative interactions between gold centres has no
appreciable effect upon either the gold–phosphorus separations
or the Au᎐S᎐C angle, but it does appear to modify the gold–
sulfur bond length. Thus in 5 the latter is significantly shorter
than the corresponding distance noted in 3 [2.308(2) Å].
Although compound 7 is the methyl ester of 3 it shows no
tendency to associate intermolecularly. The Au(1)᎐P(1) and
Au(1)᎐S(1) separation at 2.265(2) and 2.298(2) Å respectively
are very similar to those in 3, and the molecule shows a very
similar P᎐Au᎐S arrangement (Fig. 7, Table 4). The most not-
able structural differences between the two complexes occur
within the thiolate ligands, but these are small. In 7 the
Au(1)᎐S(1)᎐C(19) angle at 101.3(3) is approximately 4Њ less than
the corresponding angle in 3, and the S(1)᎐C(19) separation at
1.848(12) Å may be slightly longer than in the free acid ana-
logue. Nevertheless these parameters are well within the range
of values observed for other complexes of this type,21–23 and
there seems no obvious steric reason why 7 does not dimerise. It
has been suggested recently21 that, in the absence of steric hin-
drance or solvent molecules, complexes LAuX (with L a neutral
donor ligand and X an anionic group) are generally found as
aggregates with close gold–gold contacts regardless of the
nature of L and X. In the structures of 3, 5 and 7 discussed
above we have noted an extremely subtle interplay of weak
interactions involving gold–gold, gold–sulfur, hydrogen-
bonding, and van der Waals forces, which provide some inter-
esting exceptions to this generalisation.
Fig. 9 (a) Background-subtracted gold L(III)-edge EXAFS spectra for
compound 3. Details as in Fig. 8
giving rise to an atom separation of over 3 Å, have been less
frequently detected at room temperature by EXAFS due to
thermal disorder which dampens the signal amplitude.12 We are
aware of only two early reports27 in gold chemistry where
attempts were made to extract this type of structural inform-
ation from room-temperature EXAFS data, and for the two
compounds studied there were no comparative X-ray data
available for corroboration of the EXAFS results. In this study
gold L(III) EXAFS spectra were recorded at room temperature
EXAFS Studies
The EXAFS measurements were carried out on compounds 1
and 3, both of which exhibit short- and long-range heavy-atom
interactions. This is a powerful technique for analysing the local
environment about a specific atom, and has been used very
successfully to detect relatively strong interactions at room
temperature.25,26 However weaker interactions, especially those
678
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682

View Article Online
Table 5 Comparative data for EXAFS and X-ray crystallographic analyses of compounds 1 and 3
Compound
Shell
Distancea/Å
2σ2 b/ÅϪ2
X-Ray distancec/Å
Rd (%)
1
Au᎐S
2.290(1)
2.775(2)
3.271(6)
3.532(8)
2.329(4)
2.232(4)
3.124(19)
4.188(15)
0.005(1)
0.008(2)
0.015(3)
0.008(2)
0.001(1)
0.001(1)
0.039(5)
0.012(3)
2.294
2.790
3.157
3.569
2.308
2.254
3.131
4.001
17.9
Au᎐Au
Au ؒ ؒ ؒ Au
Au ؒ ؒ ؒ S
Au᎐S
Au᎐P
Au ؒ ؒ ؒ S
Au ؒ ؒ ؒ Au
3
18.9
a Standard deviation in parentheses; systematic errors in bond distances arising from data collection and analysis procedures are ca. 0.02 Å.
b Debye–Waller factor. c Averaged values. d R = [(χtheory Ϫ χexptl)k3dk/χexptlk3dk] × 100.
in transmission mode out to k = 15 ÅϪ1, where k = the photo-
electron wavevector, on samples prepared as finely ground
solids mulled with boron nitride. The results were analysed
using standard procedures.28,29 The contributions due to long-
distance heavy-atom interactions dominate the EXAFS spectra
at >10 k and therefore several data sets were collected and aver-
aged for each compound, and the averaged data multiplied by
k3 to compensate for the fall-off in intensity at higher k values.
However, due to poor signal-to-noise ratio the data sets were
truncated at k = 14 ÅϪ1 rather than smoothing or Fourier filter-
ing the data.
The Fourier transform of the background-subtracted
EXAFS data for compound 1 (Fig. 8) is dominated by a single
broad peak at ca. 2.3 Å with additional weaker features at
around 2.8, 3.3 and 4.5 Å. The Fourier transform for 3 is simi-
larly dominated by a peak at 2.3 Å, but it has just one other
distinct feature at around 4.2 Å (Fig. 9).
Analysis of the data for compound 1 was undertaken initially
using a single-shell model consisting of two ligated sulfur
atoms. Iteration of this model resulted in a Au᎐S separation of
2.291 Å and a R factor of just under 30%. This value could only
be improved significantly by including a second shell corres-
ponding to another gold atom at a distance of ca. 2.8 Å. Iter-
ation of this model resulted in a Au᎐Au distance of 2.78 Å and
a R factor of 20%, so defining the primary co-ordination sphere
of the gold atoms. Addition of shells corresponding to weak
or non-bonded contacts between heavy atoms (Au ؒ ؒ ؒ Au,
Au ؒ ؒ ؒ S) afforded further improvement to R resulting in a final
value of 17.8%. Attempts were made to model Au ؒ ؒ ؒ Au shells
as Au ؒ ؒ ؒ S shells. This resulted in significantly worse fits for 1
and negative Debye–Waller factors for 3. On this basis we con-
clude that the back scattering arising from the heavy gold atoms
and the other much lighter atoms in the molecules can be dis-
tinguished. The addition of each successive shell was tested for
statistical significance using the procedure of Joyner et al.30
Only shells which passed at the 1% significance level are
included in the final model. In all cases the addition of non-
bonded shells resulted in a decrease in the R factor whilst the
distance of the bonded shells remained essentially invariant.
Two procedures were used to check on the shell co-ordination
numbers used in the initial modelling. In the first the occupancy
of the shells was included in the final iteration process, and in
the second the occupation number for each shell was mapped
against its Debye–Waller factor. Neither of these produced a
change in any co-ordination number. The final set of param-
eters obtained for compound 1 is listed in Table 5 which also
contains the corresponding X-ray data.
not surprising that there is some difficulty in distinguishing
clearly between these two shells in view of the similar distances
and electronic configuration of the back-scattering atoms. A
decrease in the R factor was only obtained when more distant
shells corresponding to non-bonded atoms were included in the
modelling. There is no signal at around 2.8 Å corresponding to
a direct Au᎐Au interaction as occurs in 1. This is in accord with
the X-ray data which showed a completely different structural
arrangement in the two compounds. The next most obvious
shell in 1 is located at around 4.2 Å and including this as
a shell of gold atoms [corresponding to a non-bonded
Au(1) ؒ ؒ ؒ Au(1Ј) separation], followed by iteration, resulted in a
decrease in the R factor. Inclusion of shells of non-bonded
Au᎐S and Au᎐C atoms at ca. 3.1 and 4.2 Å respectively in the
final modelling of the data resulted in a final R factor of 18.9%.
All shells in the final model passed the significance test at the
1% level. The data in Table 5 refer to all gold–heavy atom sep-
arations in the range 0–4 Å obtained for 1 and 3 by both
methods of structural analysis. The close correspondence of the
two sets of data bears testimony to the potential of EXAFS as a
structural tool for the evaluation of close and distant contacts
between heavy atoms corresponding to both strong and weak
interactions, especially for compounds whose structures cannot
be evaluated using single-crystal X-ray methods.
Experimental
All solvents were dried and purified using standard pro-
cedures.31 Sodium tetrachloroaurate dihydrate, sulfanylacetic
acid, ethyl 2-hydroxyethyl sulfide, bis(2-methoxyethyl)amine,
carbon disulfide, potassium hydroxide and triphenylphosphine
were reagent grade and used as commercially supplied. The
thiol HSCMe2CO2H was synthesized following literature pro-
cedures.32 Infrared spectra were recorded on a Perkin-Elmer P-
E 1720X FT-IR spectrometer using KBr discs, proton, carbon
and phosphorus NMR spectra on a JEOL GSX 270 machine
with shifts reported as δ values (ppm) relative to an internal
SiMe4 standard. Elemental analyses were performed by the
Department of Pure and Applied Chemistry, University of
Strathclyde.
Syntheses
K[S2CN(C2H4OMe)2]. Potassium hydroxide (4.20 g, 0.076
mol) was stirred with (MeOCH2)2NH (6.25 g, 0.05 mol) in acet-
one (200 cm3) whilst cooled in ice. Carbon disulfide (5 cm3,
excess) in acetone (20 cm3) was added and the reaction flask
stoppered. Stirring for 3.5 h resulted in an orange solution and
some unchanged potassium hydroxide which was filtered off.
Addition of diethyl ether (500 cm3) to the filtrate afforded a
pale cream solid. This was collected and recrystallised from hot
propan-2-ol to give a pale cream microcrystalline product.
Yield 8.03 g, 65% (Found: C, 34.1; H, 5.9; N, 5.65.
C7H14KNO2S2 requires C, 34.0; H, 5.7; N, 5.7%). ν max/cmϪ1 (in
KBr) 1495s (CN), 983 (CS). δH(D2O) 3.38 (s, 6 H, OMe), 3.79
(t, 4 H) and 4.28 (t, 4 H).
Initial modelling for the data on compound 3 was under-
taken based on a two-shell model containing one S and one P
atom. Both of these contributions are contained within the sin-
gle broad envelope seen in the Fourier transform. Modelling of
the data in terms of these two shells alone yielded an R factor
of 22% with Au᎐S and Au᎐P separations of 2.33 and 2.23 Å
respectively. On reversing the order of these shells a slight
increase in the value of the R factor resulted, although it is
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682
679

Table 6 Crystallographic data for compounds 1, 3, 5 and 7
Complex
1
3
5
7
Empirical formula
M
C14H28Au2N2O4S4
810.56
C20H18AuO2PS
550.34
C22H22AuO2PS
578.39
C21H20AuO2PS
564.37
Crystal system
Space group
Monoclinic
P21/n
Orthorhombic
Pbca
Monoclinic
P21/n
Monoclinic
P21/n
a/Å
b/Å
c/Å
14.612(2)
10.774(2)
14.678(2)
101.88(2)
2261.3(6)
4
11.575(1)
15.346(2)
21.688(3)
—
3852.4(8)
8
8.732(1)
14.705(2)
8.0370(10)
17.385(3)
98.48(1)
2032.2(5)
4
17.988(4)
13.938(3)
95.28(2)
2180.0(7)
4
1.762
6.93
β/Њ
U/Å3
Z
Dc/g cmϪ3
2.381
13.351
1.898
7.839
1.845
7.433
µ(Mo-Kα)/mmϪ1
F(000)
Crystal size/mm
No. data collected
1520
0.2 × 0.2 × 0.2
3499
2112
0.2 × 0.2 × 0.25
2963
1120
0.15 × 0.15 × 0.2
3640
1088
0.2 × 0.2 × 0.25
3166
Maximum, minimum transmission
Data, restraints, parameters
Goodness of fit
1.000, 0.497
3496, 0, 248
1.101
1.000, 0.477
2952, 1, 230
1.149
1.17, 0.80*
3398, 0, 247
0.776
1.000, 0.285
3166, 0, 237
0.815
R1, wR2 [I > 2σ(I)]
(all data)
0.0295, 0.0641
0.0572, 0.0742
1.129, Ϫ0.748
0.0332, 0.0532
0.0969, 0.0721
0.671, Ϫ0.586
0.0226, 0.0637
0.0418, 0.0714
0.848, Ϫ0.606
0.0287, 0.0925
0.0493, 0.1162
0.747, Ϫ0.505
Maximum, minimum residual electron density/e ÅϪ3
Weighting scheme, wϪ1, where P = (Fo2 ϩ 2Fc2)/3
Extinction coefficient
2
2
2
2
σ2(Fo ) ϩ (0.0362P)2 ϩ 8.2330P
σ2(Fo ) ϩ (0.0060P)2 ϩ 18.8583P
σ2(Fo ) ϩ (0.0699P)2 ϩ 0.0496P
σ2(Fo ) ϩ (0.1173P)2 ϩ 0.1860P
0.001 09(15)
—
0.000 42(15)
0.0009(3)
Details in common: λ(Mo-Kα) 0.709 30 Å; T = 293(2) K; θ range 2–24Њ. * Maximum and minimum absorption corrections quoted.

View Article Online
[Au2{S2CN(C2H4OMe)2}2] 1. Sodium tetrachloroaurate dihy-
drate (5.00 g, 12.6 mmol) was dissolved in water (150 cm3) and
added dropwise to a stirred solution of ethyl 2-hydroxyethyl
sulfide (4.00 g, 37.7 mmol). When the solution became colour-
less it was added to K[S2CN(C2H4OMe)2] (3.40 g, 13.8 mmol)
dissolved in water (50 cm3). Addition of chloroform (50 cm3)
dissolved the resulting sticky yellow precipitate. The chloroform
layer was removed from the aqueous solution and allowed to
stand overnight, which resulted in the deposition of fine lumi-
nescent needles. These were collected, washed with methanol
and dried in vacuo. Yield 3.68 g, 72.4% (Found: C, 20.7; H, 3.5;
N, 3.4. C14H28Au2N2O4S4 requires C, 20.7; H, 3.5; N, 3.5%).
ν max/cmϪ1 1494s (CN), 1009, 984 (CS). δH(CDCl3) 3.36 (s, 6 H,
OMe), 3.75 (t, 4 H) and 4.22 (t, 4 H).
[Au(PPh3)(SCH2CO2Me)] 7. Compound 6 (1.00 g, 3.30
mmol) was heated under reflux with triphenylphosphine
(0.87 g, 3.30 mmol) and ethanol (120 cm3) for 1.5 h. Ethanol
was removed from the cooled solution to afford a viscous
oil which was dissolved in a little diethyl ether. Careful addition
of hexane produced a white microcrystalline product which
was separated, washed with hexane and dried. Recrystallisation
from hot propan-2-ol gave fine white needles of 7. Yield 1.38 g,
74% (Found: C, 44.9; H, 3.80. C21H20AuO2PS requires C, 44.7;
H, 3.55%). ν max/cmϪ1 1732 (CO) and 338 (Au᎐S). δH[(CD3)2CO]
3.48 (s, 2 H, CH2), 3.58 (s, 3 H, Me) and 7.54–7.70 (m, 15 H,
aromatics). δP[(CD3)2CO] 40.0.
X-Ray crystallography
Many of the details of the structure analyses carried out on
compounds 1, 3, 5 and 7 are summarised in Table 6. All meas-
urements were made on a CAD4 automatic four-circle diffract-
ometer, and data were corrected for Lorentz, polarisation and
absorption33 throughout. Data for 1, 5 and 7 were also cor-
rected for extinction. Structures were solved and refined using
SHELX 8634 and SHELXL 93.35 Refinements were based on F2
for all complexes except 5, where F data were used. Full-matrix
least-squares anisotropic refinements were implemented in all
cases. The asymmetric units of 1, 3, 5 and 7 shown in the figures
(unprimed numbers) were produced using ORTEX.36 Hydrogen
atoms were included at calculated positions where relevant
except on disordered carbon C(4) in 1, which shared a site
occupancy of 1 with position C(4A) in the ratio 58:42.
Unfortunately the acidic proton in 5 could not be located with
certainty.
[Au(SCH2CO2H)] 2. A solution of sodium tetrachloroaurate
dihydrate (23.9 g, 0.060 mol) in water (150 cm3) was added
dropwise with stirring to ethyl 2-hydroxyethyl sulfide in water
(100 cm3). Stirring was continued until the solution had
decolourised indicating reduction of AuIII to AuI. Sulfanyl-
acetic acid (6.07 g, 0.066 mol) in water (150 cm3) was added
carefully to the reaction vessel. An off-white emulsion was im-
mediately formed but on continued stirring a yellow gelatinous
precipitate resulted. This was filtered off, washed with water,
ethanol and finally diethyl ether, and dried overnight at 50 ЊC
to give a bright yellow powder. Yield 16.8 g, 97% (Found: C,
8.55; H, 1.1. C2H3AuO2S requires C, 8.3; H, 1.05%). ν max/cmϪ1
3446 (br) (OH), 1689s (CO) and 327w (Au᎐S). δH(D2O in the
presence of NaOD) 3.30 (br, CH2).
[Au(PPh3)(SCH2CO2H)] 3. Compound 2 (1.00 g, 3.47
mmol) was heated under reflux in ethanol (100 cm3) with
triphenylphosphine (0.91 g, 3.47 mmol). The colourless solu-
tion so formed was cooled resulting in the formation of a
white crystalline product which was collected, washed with
ethanol and diethyl ether, and dried in vacuo. Yield 1.47 g, 77%
(Found: C, 43.7; H, 3.0. C20H18AuO2PS requires C, 43.6; H,
3.3%). ν max/cmϪ1 3447 (br) (OH) and 1684s (CO). δH(CD3OD)
3.50 (s, 2 H, CH2) and 7.50–7.59 (m, 15 H, aromatics).
δP(CD3OD) 39.3.
CCDC reference number 186/813.
EXAFS Measurements
The EXAFS studies were carried out on finely divided mix-
tures of the compound and boron nitride (Aldrich) in the
approximate ratio 1:10 packed into a 1 mm thick aluminium
spacer and held in place using sticky tape. The gold L(III)-edge
transmission spectra were recorded at the Daresbury Synchro-
tron Radiation Source operating at 2 GeV (ca. 3.2 × 10Ϫ10 J)
with an average operating current of 210 mA on station 9.2
using a Si(220) monochromator offset to 50% of the rocking
curve for harmonic rejection. Three data sets were collected at
room temperature in k space and averaged to improve the
signal-to-noise ratio. No sample decomposition was detected
during the experiment. Calibration of the monochromator
position was carried out using thin gold foil. The EXAFS data
treatment utilised the programs EX27 and EXCURV 92.28 The
background was removed by fitting the pre-edge by a straight
line and subtracting this from the spectrum. The atomic con-
tribution to the oscillatory part of the absorption spectrum
was approximated using a high-order polynomial and the
optimum fit judged by minimising the intensity of chemically
insignificant shells at low r in the Fourier transform. Curve
fitting to the background-subtracted data used single-
scattering curve-wave theory, with phase shifts and back scat-
tering calculated using the default ab initio method within
EXCURV 92.
[Au(SCMe2CO2H)]ؒH2O 4. Reaction of sodium tetrachloro-
aurate (2.00 g, 0.010 mol), ethyl 2-hydroxyethyl sulfide (2.66 g,
0.025 mol) and dimethylsulfanylglycolic acid (1.32 g, 0.011 mol)
using a similar method to that for compound 2 above, resulted
in the formation of the product as a white powder, 2.94 g, 93%
(Found: C, 14.4; H, 3.1; S, 9.6. C4H9AuO3S requires C, 14.4; H,
2.7; S, 9.6%). ν max/cmϪ1 3447 (br) (OH) and 1694s (CO). δH(D2O
in the presence of NaOD) 1.70 (s, Me).
[Au(PPh3)(SCMe2CO2H)] 5. Compound 4 (0.50 g, 1.58
mmol) was heated under reflux in ethanol (100 cm3) with tri-
phenylphosphine (0.41 g, 1.58 mmol) for 1.5 h. The resulting
colourless solution was left to cool and the solvent then evapor-
ated to leave a white oil. Trituration with diethyl ether produced
a white microcrystalline powder which was collected and
washed with more diethyl ether before being dried. Recrystal-
lisation from hot ethanol afforded crystals of suitable quality
for a single-crystal X-ray study. Yield 0.84 g, 92% (Found: C,
45.7; H, 3.8. C22H22AuO2PS requires C, 45.7; H, 3.8%). ν max
/
Acknowledgements
cmϪ1 3447 (br) (OH) and 1676 (CO). δH(CD3OD) 1.64 (s, 6 H,
Me) and 7.50 (m, 15 H, aromatics). δP(CD3OD) 39.6.
We thank the Director of the Daresbury Laboratory for the
provision of facilities.
[Au(SCH2CO2Me)] 6. Reaction of sodium tetrachloroaurate
(6.00 g, 15.1 mmol), ethyl 2-hydroxyethyl sulfide (4.00 g, 37.7
mmol) and HSCH2CO2Me (1.85 g, 16.6 mmol) using a similar
method to that for compound 2 above afforded the product as
a pale yellow powder which was filtered off, washed with solv-
ent and dried. Yield 4.05 g, 89% (Found: C, 12.0; H, 1.50.
C3H5AuO2S requires C, 11.9; H, 1.70%). ν max/cmϪ1 1730 (CO).
References
1 D. M. P. Mingos, J. Chem. Soc., Dalton Trans., 1996, 561;
H. Schmidbaur, Gold Bull., 1990, 23, 11; Chem. Soc. Rev., 1995, 391.
2 A. Bondi, J. Phys. Chem., 1964, 68, 441.
3 H. Schmidbaur, W. Graf and G. Müller, Angew. Chem., Int. Ed.
Engl., 1988, 27, 417.
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682
681

View Article Online
4 H. Schmidbaur, Pure Appl. Chem., 1993, 65, 691; F. Scherbaum,
A. Grohmann, B. Huber, C. Kruger and H. Schmidbaur, Angew.
Chem., Int. Ed. Engl., 1988, 27, 1544.
5 A. Ulman, Chem. Rev, 1996, 96, 1533 and refs. therein.
6 S. S. Pathaneni and G. R. Desiraju, J. Chem. Soc., Dalton Trans.,
1993, 319.
7 R. J. Puddephatt, Comprehensive Organometallic Chemistry, 1st
edn., eds. G. Wilkinson, F. G. A. Stone and E. W. Abel, Pergamon,
Oxford, 1984, vol. 2, p. 765.
M. E. Kerr, D. M. Ho, R. C. Elder, A. E. Bruce and M. R. M.
Bruce, Inorg. Chem., 1993, 32, 2506.
21 M. Nakamoto, W. Hiller and H. Schmidbaur, Chem. Ber., 1993, 126,
605.
22 R. Hesse and P. Jennische, Acta Chem. Scand., 1972, 26, 3855.
23 P. Jennische, H. Anacker-Eickhoff and A. Wahberg, Acta Chem.
Scand., 1975, 31, 5143.
24 W. Schneider, A. Bauer and H. Schmidbaur, Chem. Ber., 1997, 130,
641.
8 R. V. Parish and S. M. Cottrill, Gold Bull., 1987, 20, 3.
9 A. K. H. Al-Sa’ady, K. Moss, C. A. McAuliffe and R. V. Parish,
J. Chem. Soc., Dalton Trans., 1984, 1609; P. J. Sadler, Adv. Inorg.
Chem., 1991, 36, 1 and refs. therein.
10 V. W.-W. Yam, C.-L. Chan and K.-K. Cheung, J. Chem. Soc., Dalton
Trans., 1996, 4019; S. D. Hanna and J. I. Zink, Inorg. Chem., 1996,
35, 297.
11 Y. Jiang, S. Alvarez and R. Hoffmann, Inorg. Chem., 1985, 24, 749;
M. J. Calhorda and L. F. Veiros, J. Organomet. Chem., 1994, 478, 37;
P. Pyyko, J. Li and N. Runeburg, Chem. Phys. Lett., 1994, 218, 133;
N. Kaltsoyannis, J. Chem. Soc., Dalton Trans., 1997, 1.
12 W. B. Jones, J. Yuan, R. Narayanaswarmy, M. A. Young,
R. C. Elder, A. E. Bruce and M. R. M. Bruce, Inorg. Chem., 1995,
24, 37.
13 S. Akerstrom, Ark. Kemi, 1959, 14, 387.
14 J. Dobrowolski, Z. Badowski and I. Kwiatkovska, Rocz. Chem.,
1976, 50, 53.
15 H. J. Blaauw, R. J. Nivard and G. J. M. van der Kerk, J. Organomet.
Chem., 1964, 2, 236.
16 F. Forghiera, C. Preti, L. Tassi and G. Tosi, Polyhedron, 1988, 7,
1231.
25 N. Carr, J. G. Crossley, A. J. Dent, J. R. Gouge, G. N. Greaves,
P. S. Jarrett and A. G. Orpen, J. Chem. Soc., Chem. Commun., 1990,
1369.
26 H. Bertagnolli and T. S. Ertel, Angew. Chem., Int. Ed. Engl., 1994,
33, 45.
27 M. A. Mazid, M. T. Razi, P. J. Sadler, G. N. Greaves, S. J. Gurmon,
M. H. J. Koch and J. C. Phillips, J. Chem. Soc., Chem. Commun.,
1980, 1261; R. C. Elder, K. Ludwig, J. N. Cooper and M. K.
Eidsness, J. Am. Chem. Soc., 1985, 107, 5024.
28 See, for example, A. K. Brisdon, J. M. Holloway, E. G. Hope,
W. Leveson, J. S. Ogden and A. K. Saad, J. Chem. Soc., Dalton
Trans., 1992, 447.
29 S. J. Guman, N. Binsted, J. W. Campbell and I. Ross, J. Phys. Chem.,
1984, 17, 143; 1986, 19, 1845.
30 R. W. Joyner, K. J. Martin and P. Meehan, J. Phys. Chem., 1987, 20,
4005.
31 D. D. Perrin, D. R. Perrin and W. L. F. Armitage, Purification of
Laboratory Chemicals, 2nd edn., Pergamon, Oxford, 1980.
32 E. Biilman, Liebigs Ann. Chem., 1906, 348, 120.
33 DIFABS, N. Walker and D. Stewart, Acta Crystallogr., Sect. A,
1983, 39, 158.
17 J. B. Miller and J. L. Burmeister, Synth. React. Inorg. Metal-Org.
Chem., 1976, 50, 53; H. J. A. Blaauw, R. J. Nivard and G. J. M. van
der Kerk, J. Organomet. Chem., 1964, 2, 236.
18 L. Schröter and J. Strähle, Chem. Ber., 1991, 124, 2161.
19 G. E. Coates, C. Kowala and J. M. Swan, Aust. J. Chem., 1966, 19,
539.
34 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467.
35 G. M. Sheldrick, SHELXL 93, a Computer Program for Crystal
Structure Refinement, University of Göttingen, 1993.
36 P. McArdle, J. Appl. Crystallogr., 1994, 27, 438.
20 E. Delgado and E. Hernandez, Polyhedron, 1992, 124, 2161;
Received 23rd October 1997; Paper 7/07650I
R. Narayanaswarmy, M. A. Young, E. Parkhurst, M. Oullette,
682
J. Chem. Soc., Dalton Trans., 1998, Pages 675–682