Cyanide and chloride exchange on homoleptic gold(III) square-planar complexes: Variable pressure kinetic investigation by heteronuclear NMR

Article abstract of DOI:10.1016/S0020-1693(01)00818-0

Kinetic studies of X^- exchange on [AuX₄]^- square-planar complexes (where X=Cl^- and CN^-) were performed at acidic pH in the case of chloride system and as a function of pH for the cyanide one. Chloride NMR study (330-365 K) gives a second-order rate law on [AuCl₄]^- with the kinetic parameters: (k₂^Au,Cl)²⁹⁸=0.56±0.03 s^-1mol^-1kg; ΔH₂^?Au,Cl=65.1±1 kJmol^-1; ΔS₂^?Au,Cl=-31.3±3 Jmol^-1K^-1 and ΔV₂^?Au,Cl=-14±2 cm³mol^-1. The variable pressure data clearly indicate the operation of an I_a or A mechanism for this exchange pathway. The proton exchange on HCN was determined by ¹³C NMR as a function of pH and the rate constant of the three reaction pathways involving H₂O, OH^- and CN^- were determined: k₀^HCN,H=113±17 s^-1, k₁^HCN,H=(2.9±0.7)×10⁹ s^-1mol^-1kg and k₂^HCN,H=(0.6±0.2)×10⁶ s^-1mol^-1kg at 298.1 K. The rate law of the cyanide exchange on [Au(CN)₄]^- was found to be second order with the following kinetic parameters: (k₂^Au,CN)²⁹⁸=6240±85 s^-1mol^-1kg, ΔH₂^?Au,CN=40.0±0.8 kJmol^-1, ΔS₂^?Au,CN=-37.8±3 Jmol^-1K^-1 and ΔV₂^?Au,CN=+2±1 cm³mol^-1. The rate constant observed varies about nine orders of magnitude depending on the pH and HCN does not act as a nucleophile. The observed rate constant of X^- exchange on [AuX₄]^- are two or three orders of magnitude faster than the Pt(II) analogue.

Full text of DOI:10.1016/S0020-1693(01)00818-0

www.elsevier.com/locate/ica
Inorganica Chimica Acta 331 (2002) 257–269
Cyanide and chloride exchange on homoleptic gold(III)
square-planar complexes: variable pressure kinetic investigation by
heteronuclear NMRꢀ
Florence J. Monlien, Lothar Helm, Amira Abou-Hamdan, Andre´ E. Merbach *
Institut de Chimie Mole´culaire et Biologique, Ecole Polytechnique Fe´de´rale, EPFL-BCH, CH-1015 Lausanne, Switzerland
Received 3 September 2001; accepted 15 November 2001
Dedicated to Professor A.G. Skyes
Abstract
Kinetic studies of X− exchange on [AuX4]− square-planar complexes (where X=Cl− and CN−) were performed at acidic pH
in the case of chloride system and as a function of pH for the cyanide one. Chloride NMR study (330–365 K) gives a
second-order rate law on [AuCl4]− with the kinetic parameters: (k2Au,Cl 298=0.5690.03 s−1 mol−1 kg; DH‡2 Au,Cl=65.191
)
kJ mol−1; DS2‡ Au,Cl= −31.393 J mol−1 K−1 and DV‡2 Au,Cl= −1492 cm3 mol−1. The variable pressure data clearly indicate
the operation of an Ia or A mechanism for this exchange pathway. The proton exchange on HCN was determined by 13C NMR
as a function of pH and the rate constant of the three reaction pathways involving H2O, OH− and CN− were determined:
k0HCN,H=113917 s−1, k1HCN,H=(2.990.7)×109 s−1 mol−1 kg and kH2 CN,H=(0.690.2)×106 s−1 mol−1 kg at 298.1 K. The
rate law of the cyanide exchange on [Au(CN)4]− was found to be second order with the following kinetic parameters:
(k2 ) =
Au,CN 298=6240985 s−1 mol−1 kg, DH‡2 Au,CN=40.090.8 kJ mol−1, DS‡2 Au,CN= −37.893 J mol−1 K−1 and DV‡2 Au,CN
+291 cm3 mol−1. The rate constant observed varies about nine orders of magnitude depending on the pH and HCN does not
act as a nucleophile. The observed rate constant of X− exchange on [AuX4]− are two or three orders of magnitude faster than
the Pt(II) analogue. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: High-pressure NMR; Kinetic; Gold(III) complexes; Ligand exchange; Square planar
complications from redox reactions. Soft nucleophiles
are reducing agents and Au(III) is reduced to Au(I).
1. Introduction
However, chloride and cyanide complexes of Au(III)
The chemistry of gold(III) knows recent develop-
are thermodynamically stable in aqueous solutions and
ments with the use of its compounds in the treatment of
also labile. These two complexes are therefore ideal
rheumatism [2] (anti-arthritis Au(I) drugs may be acti-
candidates for a fundamental study involving tempera-
vated in vivo to Au(III) metabolites [3]) and Au(III)
ture, pressure and pH variations.
complexes hold promise as possible anti-tumour agents
The reactions of displacement of ligand in square-
[4]. It is therefore of interest to investigate fundamental
planar transition metal complexes have been described
kinetic properties of the Au(III) complexes in solution.
in books and reviews published in the last years [5]. The
Au(III) is a low spin d8 centre which forms square-pla-
rate law for the general reaction (Eq. (1)) can be
nar complexes, but the high oxidising power of the
expressed by the two term equation (Eq. (2)):
Au(III) complexes limits the number of nucleophiles for
which ligand substitution can be investigated without
k
2
[MX4]n− +4*X− ? [M*X4]n− +4X−
(1)
(2)
kobs=k1+k2 [ligand]
ꢀ Part 97 of the series High-pressure NMR Kinetics, part 96 is Ref.
[1].
The second-order rate constant k2 refers to the bimolec-
ular attack of the ligand on the substrate and the
pseudo-first order rate constant k1 to the attack of the
* Corresponding author. Tel.: +41-21-693 9871; fax: +41-21-693
9875.
E-mail address: andre.merbach@epfl.ch (A.E. Merbach).
0020-1693/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(01)00818-0

258
F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
solvent. The k1, if observed, will depend on the nature
of the solvent. The absence of steric hindrance above
and below the plane of the complex produces a charac-
teristic kinetic behaviour with a five-coordinate interme-
diate or transition state.
The first chloride exchange study on tetrachloroau-
rate(III) was performed by Rich and Taube [6] under
conditions where hydrolysis and oxydo-reduction does
not occur. The rate constants were determined using
isotopic labelling technique with radioactive chloride
(36Cl). 35Cl/37Cl NMR can also be used to extract kinetic
information [7]. In this case, the property used to
measure the Cl− exchange rate is the free chlorine spin
relaxation rate. We have extended the temperature do-
main studied and for the first time used variable pressure
chloride 35Cl NMR to obtain the reaction activation
volume.
Early cyanide exchange studies [8] with certain heavy
metal cyanides complex were monitored with counting
techniques using 14C-labelled radiocyanide. However,
cyanide exchange can be better and easier followed using
13C NMR. A preliminary cyanide exchange study on
[Au(CN)4]− was performed by Pesek and Mason [9].
This study was limited to high pH where cyanide is
entirely deprotonated. Our first goal was to perform a
variable pH study to check if HCN could act as a
nucleophile like CN−. Further, by using variable tem-
perature and variable pressure 13C NMR, we intend to
give an answer to the nature of the transition state for
the cyanide exchange on this homoleptic complex.
processor (Titrino 716 from Metrohm), for the addition
of titrant and millivolts readings. A combined glass/ref-
erence electrode with a symmetrical electrode chain from
Radiometer Analytical S.A. (pHC2406L) and Metrohm
713 pH meter were calibrated from titration of HCl with
NaOH under the same conditions (T=298.1 K and
I=0.6 mol kg−1) thus the pH values are defined as
−log[H+]. If the solutions pH values differed by more
than 0.1 unit between the beginning and the end of the
NMR measurement, the data were not accepted.
2.3. NMR measurements
All NMR spectra were recorded in a 9.4 T magnetic
field with Bruker ARX-400 and DPX-400 spectrometers
at 100.63 MHz (13C), 32.633 MHz (37Cl) and 39.206
MHz (35Cl). In 13C NMR experiments, methanol was
used as internal standard for chemical shift (49.3 ppm)
and for B0-homogeneity. In the case of chloride NMR,
1
the H NMR water signal served as reference of the
B0-homogeneity. All aqueous solutions contained 2%
D2O as an internal lock substance. The temperature was
controlled by a Bruker B-VT 3000 unit and measured
before and after spectral accumulation by substituting
the sample with a Pt-100 resistance (accuracy: 0.5 K)
[11].
High-pressure high-resolution liquid NMR spectra
were monitored with a home built probe described
previously [12]. Tetrachloroethylene was used as pres-
surisation liquid. The temperature was stabilised to
90.2 K by pumping thermostated liquid through the
high-pressure vessel.
2. Experimental
Slow exchange reactions were followed using labelling
techniques by mixing two pre-thermostated solutions
using fast injection techniques [13] or manual injection.
The temperature of the fast injection unit was stabilised
by thermostated ethanol circulation at 298.1 K. The
mixing time was measured to be less than 0.5 s in our
conditions. About 50 spectra (1–5 scans each) were
recorded for each experiment and a supplementary delay
between successive spectra was added for slow reactions.
2.1. Chemicals and solutions
K[Au(CN)4] was prepared as described in literature
[10]. K[AuCl4] (Aldrich, p.a.), KCl (Merck, p.a.),
KCF3SO3 (Fluka, purum), K13CN (Cambridge isotope,
99% enriched) were of the highest quality available and
were used without further purification and dissolved in
doubly deionised and distilled water. Solutions contain-
ing ꢀ0.1 mol kg−1 of [Au(L)4]− and varying concen-
tration of KL (L=CN− or Cl−) were freshly prepared
before experiments. The ionic strengths of [AuCl4]− and
[Au(CN)4]− solutions were adjusted to 0.6 mol kg−1 by
adding KCF3SO3 and KNO3, respectively. [AuCl4]−
solutions were protected from light to avoid decomposi-
tion. [Au(CN)4]−/CN− solutions were studied from pH
11.5 (initial pH the solution) to pH 0.6 by adding
concentrated HNO3.
2.4. Programs
The analysis of data using appropriate equations was
accomplished with the non-linear least-squares-fitting
program SCIENTIST [14]. The reported errors correspond
to one standard deviation.
Line widths of NMR signals were obtained by fitting
Lorentzian functions to the experimental spectra using
NMRICMA 2.7 program [15] for MATLAB [16]. Complete
line shape analysis based on the Kubo–Sack formalism,
using modified Bloch equations was also performed
using NMRICMA 2.7 program to extract rate constants
from experimental spectra.
2.2. pH measurements
The pH values were determined by potentiometric
technique. The instrumentation used consists of a titro-

F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
259
Table 1
HCN,H
obs
Experimental 13C NMR line widths of the HCN/CN− signal and rate constants k
of the proton exchange on HCN at 298.1 K as a function
of pH a
(s−1
)
pH (90.1)
Dw1/2 exp (Hz)
k
(s−1
)
HCN,H b
obs
HCN,H b
obs
pH (90.1)
Dw1/2 exp (Hz)
k
2.3
2.4
2.9
3.9
4.6
5.3
6.5
6.8
7.5
21.3
20.4
19.1
19.8
20.7
27.4
80.2
1.14×102
1.20×102
0.95×102
1.18×102
1.34×102
1.38×102
4.39×102
9.43×102
2.80×103
7.9
8.4
8.7
9.1
9.4
6.16×103
3.04×104
5.20×104
1.95×105
3.34×105
6.67×105
1.07×106
1.16×106
1.83×106
148
120
119
87.8
50.6
38.2
13.7
9.7
10.1
10.3
10.7
a CCN=0.410 mol kg−1, I=0.594 mol kg−1 (KNO3).
b Extracted from the experimental spectra by the Kubo–Sack formalism.
3. Data treatment and results
Under our experimental conditions (T=298.1 K, I=0.6
mol kg−1 fixed with KNO3), a pKHa CN value of 9.590.2
was obtained.
3.1. Cyanide solutions
The proton exchange rate law was already reported in
different supporting electrolyte and ionic strength by
Bednar and Jenks [17] and Banyai et al. [18]. Three
different reaction pathways were proposed (Eqs. (4)–
(6)).
In our study of CN− exchange on [Au(CN)4]−, we had
to vary the pH from basic to acidic. Over this large range,
the free ligand present in solutions changes from CN−
at high pH to HCN at low pH. Therefore we have first
determined the dissociation constant KHa CN and the
proton exchange rate constant on HCN.
HCN,H
k
0
HCN+H2O ? CN−+H3O+
(4)
(5)
(6)
HCN,H
HCN,H
k
−0
k
1
HCN+OH− ? CN−+H2O
1
HCN,H
k
3.1.1. H exchange on HCN
−1
HCN,H
k
2
13C NMR spectra were performed from pH 10.7 to 2.3
HCN+CN− ? HCN+CN−
at 298.1 K with a 0.41 mol kg−1 KCN (I=0.6 mol kg−1
)
An overall proton exchange rate constant on HCN,
solution (Table 1). At low pH, HCN is the major species
in the solution. The 13C NMR spectra show a doublet
with a coupling constant 1JCH=272 Hz. The HCN
chemical shift and line widths are 112.4 ppm and 20 Hz,
respectively. In increasing the pH, the doublet broadens
and coalesces around pH 7–8. This phenomenon is due
to the decreasing lifetime of the proton on HCN. At pH
8, the NMR sharp coalesced signal broadens and to move
strongly towards the CN− resonance frequency at 165.1
ppm. At pH\10, the CN− form is predominant and the
line width of the signal stays around 4–6 Hz.
HCN,H
obs
k
, can be obtained by taking into an account the
three pathways (Eq. (7)).
kHobCsN,H= −d[HCN]/[HCN]dt
=k0HCN,H+kH1 CN,H[OH−]+kH2 CN,H[CN−]
=kH0 CN,H+kH1 CN,H10−(pK +pH)
w
KHa CN
KHa CN+[H+]
+kH2 CN,HCCN
(7)
The inverse of the mean lifetime of 13C in the HCN site,
krHCN (1/~HCN=kHr CN), was extracted from experimental
spectra by line shape analysis using Kubo–Sack formal-
ism and the following 3×3 exchange matrix D1 (Eq. (8),
see Appendix A for more details). PHCN and PCN are the
populations of HCN and CN−. The resonance frequen-
cies of HaCN, HbCN and CN− are fixed to 11441, 11169
and 16638 Hz, respectively.
Because the proton exchange is fast on the NMR time
scale, the dissociation constant KHa CN could be deter-
mined by fitting the variation of the chemical shift, lobs
,
of the coalesced signal as a function of pH (Eq. (3)).
KHa CN
[H+]+KHa CN
[H+]
[H+]+KHa CN
lobs
=
lCN
+
lHCN
(3)
sites
HaCN
−kHCN
HbCN
CN−
(8)
0
kHr CN
kHr CN
(PHCN/2PCN)kHr CN (PHCN/2PCN)kHr CN −(PHCN/PCN)kHr CN
Ã
Ã
Ã
r
Ã
D1
=
Ã0
−kHr CN
Ã
Ã
Ã
Ã
Ã

260
F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
spectra were measured between 273.9 and 320.1 K. At
higher temperature, decomposition [19] occurs and
HCN evaporates from the solution. The HCN chemi-
cal shift does not change in the temperature range
studied. kHr CN determined as a function of tempera-
ture, were least-squares fitted to the Eyring law (Eq.
The proton exchange rate constants observed on
HCN,H
obs
HCN, k
, which are directly given by krHCN
(kHobCsN,H=kHr CN), are shown in Table 1. Eq. (7) was
HCN,H
obs
least-squares fitted to the k
values leading to the
following rate constants: kH0 CN,H=113917 s−1
,
kH1 CN,H=(2.990.7)×109
s
−1 mol−1 kg
and
(9)), leading to activation parameters, (k0HCN,CN 298
94920 s−1, DH0‡=3892 kJ mol−1 and DS‡0= −
8096 J mol−1 K−1
)
=
kH2 CN,H=(0.690.2)×106 s−1 mol−1 kg. The pH vari-
HCN,CN
obs
ation of k
together with the relative contribu-
.
tions of the three pathways are shown in Fig. 1.
The activation parameters for the pathway 0 (Eq.
(4)) could be determined at pH 3.7 where its contribu-
tion to the proton exchange is predominant. 13C NMR
k
h
DS‡ DH‡
ꢀ n
0
0
kH0 CN,H
=
B T exp
−
R
RT
T
298.1
DH‡
R
1
1
T
ꢀ ꢁ ꢂn
0
=(kH0 CN,H 298
)
exp
−
298.1
(9)
3.1.2. Cyanide exchange on [Au(CN)4]−
The tetracyanoaurate(III) is known to be very stable
in aqueous solution. The stability constant log i4 was
estimated to be 85 [20]. A 13C NMR pH variation
study (from pH 11.6 to 2.0) was performed with a 0.1
mol kg−1 [Au(CN)4]− aqueous solution at 298.1 K. A
singlet was observed at 104.2 ppm with a line width
less than 1 Hz. The integral of that singlet was con-
stant during the pH variation indicating that no disso-
ciation or decomposition occurred even if no free
cyanide was present in solution. However, when the
pH value was increased to 13.9, the complex was
reduced completely to Au(I) species after less than 1 h
at ambient temperature. The following 13C NMR sig-
HCN,CN
Fig. 1. Variation of the log k
(ꢁ) and of the relative contribu-
obs
tions, ki, of three proton exchange pathways (kH0 CN,H ("); k1HCN,H
(ꢂ) and kH2 CN,H (ꢃ)) as a function of pH at 298.1 K; CCN=0.410
mol kg−1, I=0.594 mol kg−1 (KNO3).
nals could be assigned: [Au(CN)2]− (singlet at 153.6
1
ppm) [21], OCN− (broad triplet at 129 ppm, JCꢀN
=
10.5 Hz) [22], CO2 (singlet at 124.9 ppm) and free
cyanide. Cyanogen (CN)2 was probably present in the
solution with a broad multiplet at 119.6 ppm (the
carbon was enriched at 50%) disappearing slowly due
to evaporation.
The rate law was determined at pH 11.5 in order to
have the major proportion of cyanide in the CN−
form. 13C NMR signals of the free ligand (165.5 ppm)
and of the gold complex (104.2 ppm) broaden by the
addition of free cyanide (Fig. 2, first basic spectrum).
Preliminary tests showed that varying the ionic
strength from 0.1 to 2 mol kg−1 had no influence on
the line width observed. Therefore, solutions contain-
ing 0.1 mol kg−1 of K[Au(CN)4] were prepared at pH
11.3 and variable amounts of KCN were added suc-
cessively into the solutions to cover a concentration
domain from 0.02 to 1.23 mol kg−1. The cyanide ex-
change induces line broadening on the 13C NMR spec-
Au,CN
obs
tra, therefore, the rate constants k
were extracted
using line shape analysis with the Kubo–Sack formal-
ism. The site of the CN− bound to the metal centre
(10485 Hz) was added to the exchange matrix D1. The
new 4×4-exchange matrix D2 is given in Eq. (10).
Fig. 2. 100.63 MHz 13C NMR of the [Au(CN)4]−/HCN/CN− solu-
tion as a function of pH at 298.1 K; CCN=0.434 mol kg−1; CAu
0.102 mol kg−1; I=0.615 mol kg−1 (KNO3).
=

F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
261
Table 2
Au,CN
obs
Variation of the observed cyanide exchange rate constant k
and
its concentration normalised value k%Au,CN for CCN=0.4 mol kg−1 as
obs
a function of pH at T=298.1 K
Au,CN
obs
pH
(90.1)
k
CCN
log
(s−1
)
(mol kg−1
)
k%Au,CN
obs
Kubo–Sack
formalism
11.4
10.8
10.6
10.3
10.1
9.9
9.6
9.4
9.1
9.0
2605
2566
2623
2600
2079
1863
1444
1177
840
694
452
179
52
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
0.434
3.38
3.37
3.38
3.38
3.28
3.24
3.12
3.04
2.89
2.81
2.62
2.22
1.67
1.12
8.8
8.4
7.8
7.1
Au,CN
obs
Fig. 3. Variation of the rate constant k
of cyanide exchange on
[Au(CN)4]− as a function of free cyanide concentration at T=298.1
K and pH 11.390.2. CAu=0.09090.003 mol kg−1
.
14
Isotopic
exchange
3.7
3.3
1.3
1.0
0.8
0.6
3.90×10−3
1.42×10−3
1.70×10−5
8.31×10−6
1.08×10−5
3.37×10−6
0.251
0.263
0.197
0.188
0.290
0.271
−2.20
−2.67
−4.46
−4.75
−4.83
−5.30
Variable pH study represents another possibility also
to determine the rate law by varying cyanide concentra-
tion. The 13C NMR spectra as a function of pH of the
[Au(CN)4]−/HCN/CN− solution in aqueous media is
represented in Fig. 2. At basic medium, the line width
of the bound 13CN is broad (1/Tb2 =2400 s−1 at pH
pKHa CN=9.5, CAu=0.1190.01 mol kg−1, I=0.790.1 mol kg−1
(KNO3), H2O/D2O=5%.
Site
HaCN
HbCN
CN−
[Au(CN)4]−
−kHr CN
0
kHr CN
kHr CN
0
0
0
−kHr CN
D2=
−(PHCN/PCN)kHr CN
−(PCNb/PCN)kAr u
(10)
(PHCN/2PCN)kHr CN (PHCN/2PCN)kHr CN
−(PCNb/PCN)kAr u
0
0
kAr u
−kAr u
The inverse of the mean lifetime of cyanide in the
[Au(CN)4]− site, krAu, together with kHr CN, the inverse
of the mean lifetime of proton in HCN site, were
determined by fitting to the experimental spectra. kAr u is
defined for the exchange of a particular cyanide bound
to the metal centre Au(III), and PCNb is the corre-
sponding population (PCNb=4CAu/Ctot). The observed
kinetic rate constant of cyanide exchange on Au(III),
11.4) and sharpens when pH is decreased (1/Tb2 =12
s−1 at pH 4.6). The HCN/CN− resonance frequency
observed varies only due to the proton exchange be-
tween HCN and CN−. The pKHa CN did not change
within the experimental error (9.490.2).
Au,CN
obs
Due to the large variation of k
, two experimen-
tal techniques (line shape analysis and isotopic ex-
change) had to be used. Cyanide and proton rate
Au,CN
Au,CN
k
(Eq. (11)) is directly given by kAr u (k
=
obs
kAr u). The k
obs
Au,CN
obs
constants were extracted from spectra when line widths
values determined using the Kubo–
Au,CN
obs
are broad (ca. pH 7.0–12.0). The values of the k
and k
Sack formalism are presented in Fig. 3. Analysis of the
[CN−] dependence showed a clean second-order rate
law, without intercept, leading to kA2 u,CN value of
6240985 s−1 mol−1 kg.
HCN,H
obs
are presented in Table 2. Below pH 7, the
exchange reaction is too slow to be observed by line
broadening. However, at pHB4, the exchange is slow
enough to be determined by isotopic labelling tech-
nique. The exchange reaction was followed by rapidly
mixing 13C enriched ligand (H13CN) to a complex
kAobus,CN= −d[Au(CN)4−]/dt[Au(CN)4]−
=kA2 u,CN[CN−]
(11)

262
F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
solution at the same pH and temperature using fast
injection unit. Attention had to be paid to the long
relaxation times of bound CN−. Experimental T1 val-
ues as a function of pH, for [Au(CN)4]− and HCN/
CN− (individual and mixed solutions) were determined
to be in a time domain from 15 to 25 s. With cyanide
present and not purged from O2, the relaxation time T1
of [Au(CN)4]− was shorter compared to Pesek and
Mason results [9]. By monitoring the increase in inten-
sity of the complex signal and the decrease of that of
the free ligand, the kinetics constants were determined
with the modified McKay equation (Eq. (12)) [23,24].
Au,CN
obs
x(t)=xꢁ+(x0−xꢁ) exp(−k
t/(1−xꢁ))
(12)
x(t) represents the mole fraction of the coordinated
13CN as a function of the time and x0 and xꢁ are the
limiting conditions at t=0 and at the exchange equi-
librium, respectively. To each of the ꢀ50 spectra of the
fast injection experiment Lorentzian functions were
fitted in order to determine the integrals as a function
of time. The analysis of the experimental data using Eq.
(12) was done by a non-linear least-squares program
Au,CN
obs
with k
, xꢁ, x0 as adjustable parameters. A typical
example of the 13CN spectra observed as a function of
Au,CN
obs
time is given in Fig. 4. The k
values obtained
using fast injection, are included in Table 2. Between
Au,CN
obs
pH 3.7 and 0.6, the k
value decrease over three
orders of magnitude.
A second-order rate law was least-square fitted to all
Fig. 4. (a) Series of 13C NMR spectra after rapid mixing of two
solution containing H13CN and [Au(CN)4]−. The intensity of the
complex signal (Dw1/2=2 Hz) increased as a function of time and that
of the HCN doublet (Dw1/2=20 Hz) is decreasing; CAu=0.109
mol kg−1, CCN=0.276 mol kg−1, I=0.60 mol kg−1 (KNO3), pH
3.3, D2O=2%, T=(29890.2) K; Dt between two spectra 11.57 s
(the 15 last spectra are not presented). (b) Time evolution of the 13C
population of HCN (ꢂ) and [Au(CN)4]− (ꢄ). Fitted parameters
Au,CN
obs
the k
values as a function of pH and CCN concen-
tration leading to rate constant (k2 )
Au,CN 298=(69509
460) s−1 mol−1 kg which is within experimental error
the same as the one found from [CN−] variation at
high pH. For visualisation, the second-order rate con-
Au,CN
Au,CN
obs
stants, k
, were concentration normalised in k%
obs
to a free analytical cyanide concentration of CCN=0.4
mol kg−1 (Fig. 5; Table 2). The strong decrease of the
experimental rate observed with pH indicates that the
reacting species is CN−: no exchange pathway with
HCN as a nucleophile could be detected.
(Eq. (2)): [Au(CN)4]−
kAobus,CN=(1.41690.005)×10−3 s−1
:
x0=(0.0190.01);
xꢁ=(0.6290.01);
.
Variable temperature study (273.1–327.1 K) was per-
formed at pH 11.5 with a solution containing 0.100
mol kg−1 of K[Au(CN)4] and 0.455 mol kg−1 of KCN
(I=0.650 mol kg−1 KNO3) using line shape analysis.
At higher temperatures (\330 K) a slight yellow col-
oration of the solution appears rapidly indicating cya-
nide destruction and Au(III) reduction. Up to 327 K,
no variation of the integrals were observed and the
system was fully reversible. This leads, by fitting Eq.
(9), to the following activation parameters:
(k2
)
Au,CN 298=63009100 s−1 mol−1 kg, DH2‡=40.09
0.8 kJ mol−1 and DS2‡= −37.893 J mol−1 K−1
.
Three series of variable pressure experiments were
performed from 0.1 to 200 MPa at pH 11.5 (298.1 and
278.1 K) and pH 9.5 (279.6 K) with solutions contain-
ing 0.1090.01 mol kg−1 of K[Au(CN)4] and K13CN.
Au,CN
obs
Fig. 5. Variation of the rate constant k%
concentration:
(normalised by the
of cyanide exchange on
[Au(CN)4]− as a function of pH in water at T=298.1 K.
C )
CN=0.4 mol kg−1

F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
263
is undergoing slow chemical exchange between two
magnetically distinct environments, the transverse re-
laxation rate is given by Eq. (16) [28] where 1/Tf2Q is the
quadrupolar relaxation rate of the free halide in ab-
sence of exchange and ~f, the mean lifetime of the
chloride in the free environment.
1/Tf2=1/Tf2Q+1/~f
(16)
This lifetime, in the free state, is related to the lifetime
in the bound one, ~b, and its inverse, kb, by Eq. (17). Pb
and Pf represent the population of bound and free
ligand.
1/~f=(Pb/Pf)(1/tb)=(Pb/Pf)kb
(17)
Fig. 6. Variation of the rate constant kA2 u,CN of cyanide exchange on
[Au(CN)4]− as a function of pressure. (ꢁ): T=298.1 K, pH 11.5,
The ratio of the quadrupolar relaxation rates of
different chlorine isotopes is proportional to the
squared ratio of their nuclear electric quadrupole mo-
ments according to Eq. (18) [7].
C
CN=0.457 mol kg−1; (ꢅ): T=278.1 K, pH 11.5, CCN=0.390
mol kg−1; (ꢆ): T=298.1 K, pH 9.5, CCN=0.443 mol kg−1
.
(1/T )/(1/T )=[Q35/Q37]2=1.610
(18)
35
2Q
37
2Q
The pressure effect on the cyanide exchange on
gold(III) is represented in Fig. 6. The pressure depen-
dence of the rate constant (Eq. (13)) leads to an average
activation volume of DV‡2= +291 cm3 mol−1. The
activation volumes were assumed to be pressure
independent.
In an exchanging system, at a given temperature, the
line widths (Dw1/2=1/(pT2f ), Eq. (16)) for two different
isotopes will only differ in the quadrupolar relaxation
term 1/Tf2Q. Note that we neglect the kinetic isotopic
effect on the exchange rate constant that Hall et al. [29]
have estimated to a few percent for chloride.
DV‡
RT
ꢁ ꢂ
kA2 u,X=(k2Au,X)0 exp
−
2 P
(13)
A set of seven aqueous solutions was prepared with
0.10 mol kg−1 [AuCl4]− complex concentration and
with free Cl− concentration varying from 0.04 to 0.5
mol kg−1. The pH values of the solutions were between
2.7 and 3.7, without adding acid, which allowed to
neglect hydrolysis and deprotonation. 35Cl and 37Cl
relaxation rates, 1/Tf2, were recorded as a function of
temperature between 273.1 and 364.9 K using all solu-
tions. The observed 1/Tf2 values were found to be
independent, within experimental error, of the Cl−
concentration at all temperatures. Therefore, an aver-
age value was calculated for 1/Tf2 at each temperature
(Fig. 7). This finding allowed to simplify Eq. (19)
(combination of Eqs. (2), (16) and (17) and Pb/Pf=
4[AuCl−4 ]/[Cl−]) by neglecting the term including the
rate constant k1 (Eq. (20)).
3.2. Chloride exchange on [AuCl4]−
The kinetics study of Cl− exchange on the [AuCl4]−
complex (Eq. (1), with X−=Cl−) can be complicated
by the aquation of the complex (Eq. (14)) and deproto-
nation of the aquo complex (Eq. (15)).
[AuCl4]−+H2O?[AuCl3(H2O)]+Cl−
[AuCl3(H2O)]? [AuCl3(OH)]−+H+
(14)
K
a
(15)
The stability constant log i4 of [AuCl4]− was estimated
to 26 relative to the hypothetical [Au(H2O)4]3+ [25] and
a pKa value of 0.6 was determined by Bekker and Robb
[26] for [AuCl3(H2O)]. According to the calculations
made by Peck et al. [27] for [AuCl4]− in 1 M NaCl and
using stability constants from various authors, the hy-
drolysis is negligible below pH 6.0.
1/T2f =1/T2f Q+4[AuCl−4 ](kA1 u,Cl/[Cl−]+k2Au,Cl
1/T2f =1/T2f Q+4k2Au,Cl[AuCl4−]
)
(19)
(20)
For chlorine nuclei in complexes of diamagnetic
metal ions, the quadrupolar relaxation rates are large
Even if the temperature domain, where the exchange
contribution is predominant, is relatively large, the
extrapolated rate constant at 298.1 K cannot be deter-
mined without important experimental error. Two dif-
ferent studies of chloride exchange on [AuCl4]− were
performed using 36Cl isotopic exchange [6] and NMR
techniques [30]. The rate constant values obtained by
Sharps et al. are in a good agreement with our values at
(from 104 to 107
s
−1) due to a strong electric field
gradient [7]. Consequently, the signal of bound chlorine
is too broad to be observed and only signals of the free
halide can be measured. The study of the line width of
pairs of isotopes 35Cl/37Cl allows the separation of
quadrupolar relaxation from broadening due to chemi-
cal exchange. On systems where a quadrupolar nucleus

264
F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
Fig. 8. Pressure dependence of the observed 35Cl relaxation rate,
1/Tf2, of a 0.102 mol kg−1 [AuCl4]− and 0.527 mol kg−1 free Cl−
solution; (ꢂ) 277.5 K; (ꢄ) 340.8 K; (ꢃ) 351.5 K.
Taube are consistent with our experimental values.
Therefore, we have added the rate constant k2 deter-
mined at 273.1 and 283.1 K using isotopic exchange
(same analysis was reported [31]). Both values were
expressed in s−1 and divided by a factor of 4 according
to Taube, for comparison with our data. Eq. (9) to-
gether with Eqs. (18) and (20) and were fitted to the
averaged experimental relaxation rates and to both
Taube values, leading to the relaxation parameters (1/
298
Tf2Q,35
)
(or (1/Tf2Q,37 298), EfQ and to the activation
)
parameters DH‡2 and DS2‡ (or k2298) (Table 3; Fig. 7).
Fig. 7. Temperature dependence of the averaged observed relaxation
rates normalised by the complex concentration, 1/(Tf2×4[Au(CN)4−])
for 35Cl (ꢂ) and 37Cl (ꢄ) in aqueous solution (average values from
seven solutions with free Cl− concentration ranging from 0.05 to 0.4
mol kg−1) and temperature dependence of ln kA2 uCl from Ref. [2]
(ꢃ). The dotted lines represent exchange (– · –) and quadrupolar
relaxation contributions (----).
Ef
R
1
T
1
298
ꢀ ꢁ ꢂn
Q
1/Tf2Q=(1/Tf2Q
)
298 exp
−
(21)
The activation volume related to the quadrupolar
relaxation process, DVQ‡ , and the activation volumes
DV‡2 for Cl− chemical exchange were determined from
the pressure dependence of 1/Tf2 at three different tem-
peratures using Eqs. (20), (16) and (22).
high temperature, but an important deviation1 is ob-
served at 298.1 K with their extrapolated (k2Au,Cl 298
.
)
DV‡
RT
ꢁ ꢂ
1/Tf2Q=(1/T2f Q)0 exp −
Q P
(22)
However, the data obtained at low temperature by
35Cl high-pressure NMR experiments were performed
at 277.5 K (99.9% of quadrupolar relaxation contribu-
tion), 340.8 K (22.9% of exchange contribution) and
1 According to Fig. 4 of Ref. [30], the kex should be e2.03 and
probably not 102.03
.
Table 3
Kinetic and quadrupolar parameters for Cl− (0.04–0.5 mol kg−1) exchange on [AuCl4]− (0.1 mol kg−1) (I=0.6 mol kg−1 KCF3SO3)
Quadrupolar parameters
Kinetic parameters
298
(1/Tf2Q
)
EQf
DV‡Q
(kA2 u,Cl 298
)
DH‡2
DS‡2
DV‡2
(s−1
)
(kJ mol−1
)
(cm3 mol−1
)
(s−1 mol−1 kg)
(kJ mol−1
)
(J mol−1 K−1
)
(cm3 mol−1
)
35Cl
37Cl
4091
38.190.8 a
2591
26.190.6 a
9.491
−0.390.2
0.5690.03
65.191
−31.393
−1492
8.890.4 a
−0.290.5 a
a For comparison: values obtained with a 0.6 mol kg−1 KCl solution.

F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
265
Table 4
Rate constants (s−1 M−1 a) of proton exchange on HCN in aqueous medium at 298 K for reaction pathways i
b
i
Catalyst
H2O
pKa
log ki
log k−i
Conditions
Reference
0
1.6
10.6 c
10.7 c
11.6 c
293.1 K; I=1.0 M (KCl)
[17]
[18]
this work
1.690.1
2.0590.07
298.1 K; I=1.0 M (NaClO4)
298.1 K, I=0.6 mol kg−1 (KNO3)
2
1
CN−
OH−
9.0
9.09
9.5
B5
293.1 K; I=1.0 M (KCl)
298.1 K; I=1.0 M (NaClO4)
298.1 K, I=0.6 mol kg−1 (KNO3)
I=0.01–0.05 M
[17]
[18]
this work
[37]
6.590.1
5.890.2
9.6 d
15.74
13.80
13.75 f
9.990.2
9.590.1
9.590.2 a
3.2b
4.8 e
5.3 c
295.1 K; I=1.0 M (KCl)
[17]
[18]
this work
298.1 K; I=1.0 M (NaClO4)
298.1 K, I=0.6 mol kg−1 (KNO3)
a Except ki of this work expressed in s−1 mol−1 kg and k0 in s−1
b Experimental value.
.
c Calculated value.
d Ultrasonic absorption value.
e Recalculated value: log k−HC1 N,H=log kH1 CN,H−pKw+pKaHCN=9.5−13.8+9.09=4.8.
f Ref. [38].
351.5 K (40.5% of exchange contribution) allowing to
separate both activation volumes DV‡Q and DV2‡ (Fig.
8). Ideally, a higher temperature should have been used
to better define DV‡2. However, this was not feasible due
to irreversible decomposition of [AuCl4]− above 370.0
K. From the eight adjustable parameters (two activa-
tion volumes, three zero-pressure rate constants and
three quadrupolar relaxation rates at the different tem-
peratures) two values were fixed as obtained from the
dent. The CN− catalysed pathway 2 (Eq. (6)) con-
tributes between pH 5 and 10 with a maximum of 50%
to the overall rate. The OH− catalysed pathway 1 (Eq.
(5)) contributes to ꢀ40% between pH 7 and 9 and
becomes predominant above pH 9.5. If the general
observation is similar to that of Glaser and co-workers
[18], we found different values for the individual rate
constant (Table 4) mainly due to a difference of a factor
2 in the exchange matrices.
variable temperature study (k2277.5 and (1/T2f Q
)
351.5) due
With
log k−HC0N,H−log k0HCN,H
k−HC0N,H=4×1011 −1 s−1 indicating a diffusion-con-
a
pKHa CN=9.5, we calculate (pKHa CN
reverse rate constant
=
to their minor individual contribution to 1/Tf2Q. The
activation values resulting from the fitting procedure
are given in Table 3.
)
a
M
trolled reaction [36]. Our reverse rate constant is about
one order of magnitude faster than those from litera-
ture (Table 4) which can be partially attributed to our
higher pKHa CN value. The activation parameters DH0‡
(Table 5) is close to the thermodynamic enthalpy
4. Discussion
4.1. Proton exchange on HCN
DH°=44.6 kJ mol−1 (0.5 M, NaCl) determined by
0
Solis et al. [34] for HCN dissociation (Eq. (3)) as
excepted for a reverse diffusion-controlled reaction.
Values of pKHa CN have been determined in a number
of studies under a large range of experimental condi-
tions. The formation of HCN occurs below pH 11 and
depends upon the ionic strength, the nature of the
supporting electrolyte and the temperature of the solu-
tion. The general trend observed is the following: the
higher the ionic strength, the higher the pKHa CN value.
The data published are included in the range between
8.78 (318.1 K, extrapolated to I=0 mol kg−1) and
10.01 (298.1 K, I=5 mol kg−1) [32–35]. Our value of
9.5, measured at I=0.6 mol kg−1, is in good agree-
ment with the observed trend.
Our experimental value for the rate constant k1HCN,H
,
obtained for the OH− catalysed pathway, is consistent
with literature data obtained by NMR techniques and
ultrasonic absorption measurements [37]. The value of
Table 5
Activation parameters of the proton exchange on HCN catalysed by
H2O at acidic pH in water (pathway i=0)
I (M)
pKHa CN DH‡0
(kJ mol−1
DS‡0
Reference
)
(J mol−1 K−1
)
1.0 M
9.09
4594
3892
−64912
−8096
[18]
The proton exchange on HCN proceeds via three
(NaClO4)
pathways (Eqs. (4)–(6)) involving H2O, OH− and CN−,
0.8 mol kg−1 9.5
(KCN)
this work
respectively (Fig. 1). At low pH (B4) the predominant
HCN,H
obs
pathway 0 is H2O catalysed and k
is pH indepen-

266
F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
HCN,H
the reverse rate constant k
was calculated (pKw−
more dilute solutions (0.008–0.07 M for free chloride
and 0.01 M for the [AuCl4]−) at 273.1 and 283.1 K.
They could therefore also observe a kA1 u,Cl pathway
beside the kA2 u,Cl pathway. With the NMR technique,
needing a much higher Cl− concentration, it was not
possible to detect the kA1 u,Cl pathway and a pure sec-
ond-order rate law was observed under our experimen-
tal conditions (kA1 u,Clꢂk2Au,Cl[Cl−]). The combined
analysis of our high-temperature data and of Taube’s
low temperature data allows to define the activation
parameters well, DS‡ and DH‡, for the k2Au,Cl pathway
(Table 6).
−1
HCN,H
−1
pKaHCN=log k1HCN,H−log k
pKw determined for our ionic medium (pKw=13.75)
[38] and this value is in good agreement with the Glaser
recalculated value.
) using pKHa CN and
The value of the rate constant kH2 CN,H, of the proton
exchange catalysed directly by CN−, is similar to the
ones already published. This pathway is mainly active
at pH values just below the pKHa CN around pH 7 where
the sum of the catalysis by H3O+ and OH− is at its
minimum.
4.2. X− exchange on [AuX4]−
The cyanide exchange is 105 times faster than the
chloride one on the homoleptic gold(III) complexes.
The effectiveness of the nucleophile attack is not only
due to its basicity but also to its polarisability [40,41].
From the infrared and Raman spectra of aqueous
[Au(CN)4]− solution, Jones and Smith [45] have de-
duced high CN and AuC force constants indicative of a
strong metal–ligand s-bonding and a weak metal–lig-
and p-back-bonding. Cyanide will increase the effective
positive charge on the metal centre as compared to
chloride, which has no metal–ligand p-back-bonding,
and thus facilitates bond formation with the incoming
nucleophile. The difference in reactivity between the
cyanide and the chloride exchange is reflected in a large
difference in the DH‡ terms whereas the DS‡ terms are
very similar (Table 6).
The general mechanistic picture for substitution reac-
tions on square-planar complexes is well established
and involves a concurrent bimolecular attack by the
solvent and the ligand on the metal centre (rate=(k1+
k2[ligand]×[complex]) [39]. The k1 term is the rate
constant for the solvolytic pathway, the k2 term corre-
sponds to the direct attack by the incoming ligand. In
an early study of Cattalini and Tobe [40] of ligand
substitution on gold(III) complexes, it was concluded
that rates of ligand substitution in gold(III) complexes
are very sensitive on the nature of both the incoming
and the leaving group [25]. Furthermore it was found
that the bond formation and bond breaking are nearly
synchronous processes. The substitution on Au(III) in-
volves a pentacoordinated transition state rather than a
pentacoordinated intermediate. This proposal is com-
patible with Baddley and Basolo suggestions [41]. Cat-
talini [42,43] underlined also that the solvent plays at
least two different roles in the substitution reaction. On
one hand, it can behave as a nucleophile (pathway k1).
On the other hand, factors such as solvation changes of
the reacting species and solvent–structure variations of
the transition state must be also taken into account (for
both pathways). The symmetrical ligand exchange on
square-planar complexes usually follow a second-order
kinetic with negative entropy and volume of activation
leading to the assignment of an associative mode of
activation [44]. The solvolytic pathway k1, is usually
negligible.
Even if the entropy values are similar for both chlo-
ride and cyanide substitutions, we observed significantly
different activation volumes. In the case of CN−, the
DV‡2 Au,CN= +291 cm3 mol−1 is less negative com-
pared to DV‡2 Au,Cl= −1492 cm3 mol−1, and even
slightly positive. The value observed for the chloride
exchange corresponds to an associative interchange Ia
or to a limiting associative mechanism A. However,
there is no straightforward explanation for the cyanide
exchange. Small DV‡ values for ligand exchange on
homoleptic square-planar complexes have already been
reported with the assignment of an associative activa-
tion mode [44]: −2.2 cm3 mol−1 for [Pd(H2O)4]2+
,
−0.2 cm3 mol−1 for [Pd(DMF)4]2+, −3.1 cm3 mol−1
for [Pd(MeNC)4]2+, −4.6 cm3 mol−1 for [Pt(H2O)4]2+
and −3.7 cm3 mol−1 for [Pt(MeNC)4]2+. The interpre-
tation of these volumes in molecular terms is not trivial,
as these solvates undergo large changes in coordination
geometry on going from ground state to transition
state. Tetrasolvate species may have their two axial sites
occupied by loosely bound solvent molecules. The
simultaneous loss of one of the axially bound solvent
molecules with the formation of a new bond with the
second solvent molecule in going to a five-coordinate
transition state or intermediate may mean that through
compensation, the overall volume change is small or
negligible. The difference between the chloride and the
cyanide can be explained by specific solvation changes
The cyanide exchange on [Au(CN)4]− follows a
purely second-order rate law. If a kA1 u,CN first-order
term would exist, it should be less than 10−5 s−1 with
regard to the order of magnitude of the observed rate at
low pH, where the cyanide ligand is still active even if
its concentration is less than 10−9 mol kg−1. The ab-
sence of the kA1 u,CN term allows to conclude that HCN
does not react as a nucleophile and the absence of a
solvolytic pathway.
We found by chlorine NMR that the rate law for the
chloride exchange on [AuCl4]− is second order in the
temperature range of 330.1–365.1 K. Rich and Taube
[6] worked using radioactive isotopic technique with

F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
267
Table 6
Exchange rate constant (k2M,X 298
)
and activation parameters for X exchange on [MX4]n− in water for M=Au(III) and Pt(II) a
Complex
Entering ligand
(kM2 ,X 298
(s−1 M−1
)
DH‡
DS‡
DV‡
Method
Reference
)
(kJ mol−1
)
(J mol−1 K−1
)
(cm3 mol−1
)
[Au(CN)4]−
[AuCl4]−
CN−
CN−
6240985 b
3900 (297.0 K)
40.091
2891
−37.893
−10093
+291c
this work
[21]
NMR
Cl−
0.5690.03b
65.191
−31.393
−1492
NMR and
isotopic
isotopic
UV
this work
H2O
H2O
2.7×10−5 d
1×10−4 d
[6]
[50]
64.9
[Pt(CN)4]2−
[PtCl4]2−
CN−
1191 b
25.190.4
87.9
−14292
−2792
−1792
NMR
[48]
H2O
1.8×10−7 d
isotopic
[47,49]
a The rate constant reported are for the exchange of a particular coordinated X−. The rate constant for an unspecified X− is therefore four
times greater and the corresponding DS‡ value is R ln 4=11.5 greater.
b k2 in s−1 mol−1 kg.
c Averaged value (see Fig. 6).
d k1 in s−1 M−1=k1 (in s−1)/4[H2O], for the exchange of a particular X−
.
of the reactant and the transition state. In particular,
coordinated and free CN− are well known to form
hydrogen bond with water and even protonation can
occur [20]. For [Au(CN)4]− the slightly positive DV‡
value could as well be assigned to a synchronous I
values calculated based on the fitted parameters of the
rate laws and supplementary figures are available from
the authors upon request.
mechanism. This is further supported by the early
observations that the ligand substitution rate is sensi-
tive on the nature of both the incoming and the leaving
group [25].
Acknowledgements
Financial support for this research provided by the
Swiss National Science Foundation is acknowledged.
It is of interest to compare the reactions of Au(III)
complexes with the analogous Pt(II) ones. The Au(III)
complexes are more reactive than the Pt(II) complexes
towards substitution (Table 6) [46,47]. The ligand ex-
change is 600 times faster on [Au(CN)4]− than
[Pt(CN)4]2− [48] and occurs in both cases only through
a CN− nucleophilic attack. The exchange of chloride
on [AuCl4]− and [PtCl4]2− occurs by fundamentally
different processes. Depending on the chloride concen-
tration and on the temperature, the exchange on
[AuCl4]− involves a nucleophilic attack by Cl− or by
Appendix A
The line shape of the 13C NMR spectra of KCN
dissolved in water as a function of pH are affected by
proton exchange on HCN and can be calculated with
the Kubo–Sack formalism which allows for a calcula-
tion of the NMR line shape with chemical exchange
between different sites [51]. Due to the spin–spin
coupling between 1H and 13C a three sites exchange
has to be considered. A site is therein characterised
by a resonance frequency, ꢀi=2pwi, a population Pi
and a transverse relaxation in the absence of exchange,
1/T2i.
H2O [49]. The chloride attack is negligible on [PtCl4]2−
.
For the aquation pathway, Au(III) [50] is approxi-
mately 150 times faster than Pt(II) (Table 6).
In conclusion, the chloride exchange on [AuCl4]− can
occur either through a direct nucleophilic attack of Cl−
or an aquation pathway. The variable pressure experi-
ment (DV‡2 Au,Cl= −1492 cm3 mol−1) clearly indicate
the operation of an Ia or A mechanism. For the cyanide
exchange on [Au(CN)4]−, the observed rate constant
varies about nine orders of magnitude depending on the
pH domain where HCN does not act as nucleophile.
The NMR line width is calculated with Eq. (A1).
V(ꢀ)=C×Re(1K−1P)
(A1)
where 1 is the unity vector, C is a scaling factor and P
is a column vector with fractional populations of the
different sites and the matrix K is given by Eq. (A2).
K=D+R−i(V−ꢀI)
(A2)
5. Supplementary material
where I is a three-dimensional unit matrix and R and V
are the diagonal matrices of resonance frequencies
and transverse relaxation rates given by Eqs. (A3) and
(A4)
Supplementary material including table listing the
reactions conditions, the rate constants obtained, the

268
F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
a
b
with PCN/~CN=PH CN/~HCN+PH CN/~HCN
P =
CN/~CN
−1/T
0
0
0
Ã
Ã
Ã
2HCN
Ã
R=Ã
0
0
−1/T2HCN
Ã
(A3)
P
HCN/~HCN PCNkCr N=PHCNkHr CN
At pH\7 the Kubo–Sack formalism was also used
Ã
Ã
Ã
Ã
0
−1/T2CN
to extract rate constant of cyanide exchange on
[Au(CN)4]−. The exchange matrix accounts for two
exchange processes, the proton exchange on HCN (as
described before) and the cyanide exchange on the
[Au(CN)4]− complex which is introduced as a fourth
site in the matrix D2.
−ꢀHCN+JCH/2
0
0
0
Ã
Ã
Ã
Ã
V=Ã
0
0
−ꢀHCN−JCH/2
Ã
Ã
Ã
Ã
Ã
0
−ꢀCN
We consider that the exchange can occur only via a
CN− ion and HCN itself is inactive. This consideration
was confirmed by the rate law found for the Au(III)
metal centre.
(A4)
The exchange matrix, D1, for the proton exchange on
HCN in water is given by Eq. (A5).
HhCN HiCN
CN−
(A5)
1/~
D1= Ã0
0
1/~HCN
Ã
Ã
Ã
HCN
References
Ã
-1/~HCN 1/~HCN
1/(2~CN
Ã
Ã
Ã
[1] F.A. Dunand, R.S. Dickins, D. Parker, A.E. Merbach, Chem.
Eur. J. 7 (2001) 5160.
Ã
Ã
1/(2~CN
)
)
−1/~CN
[2] (a) F. Zamora, P. Amo-Ochoa, B. Fischer, A. Schimanski, B.
Lippert, Angew. Chem., Int. Ed. Engl. 38 (1999) 2274 (and
references therein);
We distinguish two sites (HaCN and HbCN) depending
on the a/b state of the proton spin) accessible from the
(b) S.J. Lippard, Platinum, Gold and Other Metal Chemothera-
peutic Agents. In: American Chemical Society Symposium Se-
ries, vol. 209, American Chemical Society, Washingon, DC,
1983.
initial CN− site by a single step exchange. The lifetime
HCN
~
and the fractional population PHCN are exactly
the same for both HCN sites and the probability of a
direct ‘ab flip–flop’ exchange being negligible.
D1 can be expressed in terms of kHr CN, the inverse of
the mean lifetime in HCN site.
[3] C.F. Shaw, in: H. Schmidbaur (Ed.), Gold: Progress in Chem-
istry, Biochemistry and Technology, Wiley, New York, 1999.
[4] P.J. Saddler, R.E. Sue, Met. Based Drug 1 (1994) 107.
−kHCN
0
kHr CN
kHr CN
Ã
Ã
Ã
r
Ã
D1=Ã0
−kHr CN
Ã
Ã
Ã
P
HCN/(2PCN)kHr CN
P
HCN/(2PCN)krHCN −PHCN/PCNkHr CN
Ã
Ã

F.J. Monlien et al. / Inorganica Chimica Acta 331 (2002) 257–269
269
[5] (a) F. Basolo, R.G. Pearson, Mechanism of Inorganic Reactions,
second ed., Wiley, New York, 1965;
[26] van Z. Bekker, W. Robb, Inorg. Nucl. Chem. Lett. 8 (1972) 849.
[27] J.A. Peck, D. Tait, B.I. Swanson Jr., G.E. Brown, Geochim.
Cosmochim. Acta 55 (1991) 671.
[28] A.C. McLaughlin, J.S. Leigh Jr., J. Magn. Reson. 9 (1973) 296.
[29] C. Hall, D.W. Kydon, R.E. Richards, R.R. Sharp, Proc. R. Soc.,
A 318 (1970) 119.
(b) R.J. Cross, Adv. Inorg. Chem. 34 (1989) 219;
(c) S.F. Lincoln, A.E. Merbach, in: A.G. Sykes (Ed.), Advances
in Inorganic Chemistry, Academic Press, New York, vol. 421,
1995, p. 1;
(d) M.L. Tobe, J. Burgess, Inorganic Reaction Mechanisms,
Longman, London, 1999.
[30] J.A. Sharps, G.E. Brown Jr., J.F. Stebbins, Geochim. Cos-
mochim. Acta 57 (1993) 721.
[6] R. Rich, H. Taube, J. Phys. Chem. 58 (1954) 1.
[7] B. Lindman, S. Forsen, NMR, Basic Principles and Progress No.
12, Springer, Berlin, 1976, p. 177.
[8] A.W. Adamson, J.P. Welker, M. Volpe, J. Am. Chem. Soc. 73
(1950) 4030.
[9] J.J. Pesek, W.R. Mason, Inorg. Chem. 22 (1983) 2958.
[10] J.M. Smith, L.H. Jones, I.K. Kressin, R.A. Penneman, Inorg.
Chem. 4 (1965) 369.
[11] C. Ammann, P. Meier, A.E. Merbach, J. Magn. Reson. 46
(1982) 319.
[12] (a) U. Frey, A.E. Merbach, High Pressure Res. 2 (1990) 237;
(b) U. Frey, L. Helm, A.E. Merbach, R. Roulet, Advanced
Applications of NMR to Organometallic Chemistry, Wiley, New
York, 1996;
(c) A. Cusanelli, L. Nicula-Dadci, U. Frey, A.E. Merbach,
Inorg. Chem. 36 (1997) 2211.
[13] P. Berhard, L. Helm, A. Ludi, A.E. Merbach, J. Am. Chem. Soc.
107 (1985) 312.
[31] D. Hugi-Cleary, L. Helm, A.E. Merbach, Helv. Chim. Acta 68
(1985) 545.
[32] R.M. Izatt, R.T. Christensen, R.T. Pack, R. Bench, Inorg.
Chem. 1 (1962) 828.
[33] S.J. Broderius, J. Anal. Chem. 53 (1981) 1472.
[34] J.S. Solis, P.M. May, G. Hefter, Aust. J. Chem. 49 (1996) 651.
[35] P. Verhoeven, G. Hefter, P.M. May, Miner. Metall. Process. 7
(1990) 185.
[36] M. Eigen, Angew. Chem., Int. Ed. Engl. 3 (1964) 1.
[37] J. Stuehr, E. Yeager, T. Sachs, F. Hovorka, J. Chem. Phys. 38
(1963) 587.
[38] R.M. Smith, A.E. Martell, Critical Stability Constant 4, Plenum
Press, New York, 1976.
[39] R.G. Wilkins, Kinetics and Mechanism of Reactions of Transi-
tion Metal Complexes, VCH, Weinheim, 1991.
[40] L. Cattalini, L. Tobe, Inorg. Chem. 5 (1966) 1145.
[41] H.W. Baddley, F. Basolo, Inorg. Chem. 3 (1964) 1090.
[42] L. Cattalini, Prog. Inorg. Chem. 13 (1970) 263.
[43] L. Cattalini, V. Ricevuto, A. Orio, M.L. Tobe, Inorg. Chem. 7
(1968) 51 (and references therein).
[44] U. Frey, A.E. Merbach, D.H. Powell, in: J.J. Delpuech (Ed.),
Dynamics of Solutions and Fluid Mixtures by NMR, Wiley,
Chichester, 1995 (Chapter 6).
[45] L.H. Jones, J.M. Smith, J. Chem. Phys. 41 (1964) 2507.
[46] L. Cattalini, A. Orio, M.L. Tobe, J. Am. Chem. Soc. 89 (1967)
3130.
[47] L.F. Grantham, T.S. Elleman, D.S. Martin, J. Am. Chem. Soc.
77 (1955) 2965.
[14] SCIENTIST, version 2.0, Micromath, Inc., Salt Lake City, UT,
1995.
[15] L. Helm, A. Borel, NMRICMA 2.7, University of Lausanne, 2000.
[16] MATLAB® program, version 5.3.1, Mathworks, Inc., 1999.
[17] R.A. Bednar, W.P. Jenks, J. Am. Chem. Soc. 107 (1985) 7117.
[18] I. Banyai, J. Blixt, J. Glaser, I. Toth, Acta Chem. Scand. 46
(1992) 142.
[19] H. Klenk, A. Griffiths, K. Huthmacher, H. Itzel, H. Knorre, C.
Voigt, O. Weiberg, Ullmann’s Encyclopedia, A8, p. 159 and
references therein.
[20] A.G. Sharpe, The Chemistry of Cyano Complexes of the Transi-
tion Metals, Academic Press, New York, 1976, p. 279.
[21] J.J. Pesek, R. Mason, Inorg. Chem. 18 (1979) 924.
[22] M.R. Beck, V. Gaspar, D.P. Goel, Inorg. Chem. Acta 33 (1979)
L147.
[48] F.J. Monlien, A. Abou-Hamdan, L. Helm, A.E. Merbach, High
Pressure Res. 18 (2000) 319.
[49] H.E. Brower, L. Hathaway, K.R. Brower, Inorg. Chem. 11
(1966) 1899.
[50] W. Robb, Inorg. Chem. 6 (1967) 382.
[51] (a) R. Kubo, J. Phys. Soc. Jpn. 9 (1954) 888;
(b) R.A. Sack, Mol. Phys. 1 (1958) 163;
[23] H. McKay, J. Am. Chem. Soc. 65 (1943) 702.
[24] S.F. Lincoln, A.E. Merbach, in: A.G. Sykes (Ed.), Advances in
Inorganic Chemistry, Academic Press, San Diego, 1995.
[25] L.H. Skibsted, Adv. Inorg. Bioinorg. Mech. 4 (1986) 137.
(c) L.W. Reeves, K.N. Shaw, Can. J. Chem. 48 (1970) 3641.