Liquid-liquid extraction of palladium(II) and gold(III) with N-benzoyl-N′,N′-diethylthiourea and the synthesis of a palladium benzoylthiourea complex

Article abstract of DOI:10.1016/S0277-5387(02)00948-8

N-Benzoylthioureas have been reported to form complexes with gold (III) and palladium (II) and other transition metals. In this study, an N-benzoyl-N′,N′-diethylthiourea (3f) ligand was used in the solvent extraction of palladium(II) and gold(III) from aqueous chloride media (0.1 mol l^-1 NaCl). The distribution coefficient was determined as a function of both metal concentration in the aqueous phase and extractant concentration in the organic phase. The experimental distribution data were numerically analysed by LETAGROP-DISTR software in order to obtain the thermodynamic model corresponding to the metal extraction. It is found that pH does not affect the metal extraction process in the 1-2 pH range. Synthesis of the palladium benzoyl thiourea complexes was carried out by mixing quantities of metal and ligand solutions in methanol in a 1:2 ratio stoichiometric. Yields of 74 and 80.9% were obtained for the Pd-3c and Pd-3f complexes. In order to confirm the formation of the palladium complexes, NMR, FTIR and MS analyses were performed. From MS analyses a complex stoichiometry 1:2 (metal:ligand) was confirmed. The formation of crystals of palladium N-benzoyl-N′,N′-diethylthiourea complex (Pd-3f) in the methanolic solution allows the characterisation of the complex structure by XRD. The resulting structure is described and discussed. Bis(1,1,-diheptadecyl-3-benzoyl-thioureate)palladium(II) (Pd-3c) and bis(1,1,-diheptadecyl-3-benzoyl-thioureate)palladium(II) (Pd-3f) were used as ionophores in polymeric membrane electrodes. Their potentiometric responses to different anionic metal chlorocomplexes are evaluated and discussed taking into consideration the results obtained in the liquid-liquid distribution studies. A nernstian response was only obtained for AuCl₄^-(PDL = 8.8 × 10^-8) and PdCl₄^2- (PDL = 1.5 × 10^-4 M) with a selectivity coefficient of K_{AuCl4-, PdCl42-}^pot = -3.4, calculated taking AUCl₄^- as being the primary anion.

Full text of DOI:10.1016/S0277-5387(02)00948-8

Polyhedron 21 (2002) 1429ꢀ1437
/
www.elsevier.com/locate/poly
Liquidꢀ
/
liquid extraction of palladium(II) and gold(III) with N-
benzoyl-N?,N?-diethylthiourea and the synthesis of a palladium
benzoylthiourea complex
Merce` Dom´ınguez a, Enriqueta Antico´ a, Lothar Beyer b, Angel Aguirre c,
Santiago Garc´ıa-Granda c, Victoria Salvado´ a,*
a Dept. de Qu´ımica, Universitat de Girona, Campus Montilivi, 17071 Girona, Spain
b Institut fur Anorganishe Chemie, Universitat Leipzig, Johannisallee 29, D-04103 Leipzig, Germany
¨
¨
c Depto. de Qu´ımica F´ısica y Anal´ıtica, C/Julia´n Claver´ıa 8, 33006 Oviedo, Spain
Received 23 October 2001; accepted 3 December 2001
Abstract
N-Benzoylthioureas have been reported to form complexes with gold (III) and palladium (II) and other transition metals. In this
study, an N-benzoyl-N?,N?-diethylthiourea (3f) ligand was used in the solvent extraction of palladium(II) and gold(III) from
aqueous chloride media (0.1 mol lꢁ1 NaCl). The distribution coefficient was determined as a function of both metal concentration in
the aqueous phase and extractant concentration in the organic phase. The experimental distribution data were numerically analysed
by LETAGROP-DISTR software in order to obtain the thermodynamic model corresponding to the metal extraction. It is found that
pH does not affect the metal extraction process in the 1ꢀ2 pH range. Synthesis of the palladium benzoyl thiourea complexes was
/
carried out by mixing quantities of metal and ligand solutions in methanol in a 1:2 ratio stoichiometric. Yields of 74 and 80.9% were
obtained for the Pd-3c and Pd-3f complexes. In order to confirm the formation of the palladium complexes, NMR, FTIR and MS
analyses were performed. From MS analyses a complex stoichiometry 1:2 (metal:ligand) was confirmed. The formation of crystals of
palladium N-benzoyl-N?,N?-diethylthiourea complex (Pd-3f) in the methanolic solution allows the characterisation of the complex
structure by XRD. The resulting structure is described and discussed. Bis(1,1,-diheptadecyl-3-benzoyl-thioureate)palladium(II) (Pd-
3c) and bis(1,1,-diheptadecyl-3-benzoyl-thioureate)palladium(II) (Pd-3f) were used as ionophores in polymeric membrane
electrodes. Their potentiometric responses to different anionic metal chlorocomplexes are evaluated and discussed taking into
ꢁ
consideration the results obtained in the liquidꢀliquid distribution studies. A nernstian response was only obtained for AuCl4
/
pot
2ꢁ
(PDLꢂ
/
8.8ꢃ
/
10ꢁ8) and PdCl4
(PDLꢂ
/
1.5ꢃ
/
10ꢁ4 M) with a selectivity coefficient of K
/
ꢂ
/
ꢁ3.4, calculated taking
/
AuCl-4; PdCl24ꢁ
ꢁ
AuCl4 as being the primary anion. # 2002 Elsevier Science Ltd. All rights reserved.
Keywords: N-Benzoylthiourea derivatives; Gold; Palladium; Solvent extraction; Ion selective electrode; Crystal and molecular structure
1. Introduction
anion exchangers, and, in accordance with Pearson’s
theory [1], better performance is obtained with sulfur-
containing compounds in the cases of Au(III) and
Pd(II). Of the extractants containing soft atoms, a novel
group of benzoylthioureas has been reported to form
complexes with Cu, Fe(III), Pb, Ni, Zn, Cd, Co, Ag, Pd,
Hg and Au [2]. In particular, N-benzoyl-N?,N?-dihepta-
decylthiourea (3c) and N-benzoyl-N?,N?-diethylthiourea
(3f) (see Fig. 1) have been successfully used in the
extraction of copper(II) [3] and gold(III) [4] from acidic
media. These reagents have also been implemented in a
solid supported liquid membrane system for the selective
The development of selective extractants for the
separation and concentration of precious metals at trace
levels reflects an increasing need for these metals to be
recovered and analysed. Chelating extractants have been
found to be more selective than solvating reagents and
* Corresponding author. Tel.: ꢄ
150.
E-mail address: victoria.salvado@udg.es (V. Salvado´).
/
34-972-418-271; fax: ꢄ34-972-418-
/
0277-5387/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 7 7 - 5 3 8 7 ( 0 2 ) 0 0 9 4 8 - 8

1430
M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ1437
/
2. Experimental
2.1. Reagents and solutions
N-benzoyl-N?,N?-diheptadecylthiourea (3c) and N-
benzoyl-N?,N?-diethylthiourea (3f) were synthesised by
Professor Beyer et al. [15]. Organic solutions of these
ligands were prepared in CHCl3 (HPLC grade, Panreac,
Spain).
Stock solutions of the metals were prepared from the
solid compounds, Na2PdCl4 (Aldrich, UK) and HAuCl4
(Fluka, Switzerland). Working solutions were prepared
by dilution and appropriate volumes of NaCl and HCl
standardised solutions were added to adjust the ionic
strength (0.1 M) and the Hꢄ concentration to the
desired value.
Pro analysis-grade Na2PdCl4 (Aldrich), methanol
(Fluka) and ethanol (Fluka) were used for the synthesis
of solid palladium complexes Pd-3c and Pd-3f.
High molecular weight poly(vinylchloride) (PVC) and
o-nitrophenyl octyl ether (o-NPOE), used to prepare the
polymeric membranes, were purchased from Fluka.
All other reagents were of analytical grade (Panreac,
Spain) and all standard solutions and buffers were
prepared with deionised (Milli-Q, Millipore, Bedford,
USA) distilled water.
Fig. 1. Chemical structure of the ligands N-benzoyl-N?,N?-dihepta-
decylthiourea (3c) and N-benzoyl-N?,N?-diethylthiourea (3f).
transport of palladium(II) [5] and gold(III) [6]. Recently,
the field of organic extractant applications has broa-
dened since metal complexes of reagents used in solvent
extraction have also been studied in the development of
polymeric membrane electrodes taking advantage of the
specific interactions between the carrier, i.e. the metal
complex and the target ion. This is the case of several
metallomacrocyclic and organometallic compounds, for
example metalloporphyrin [7] and vitamin B12 [8]
2.2. Apparatus
A Labinco rotary mixer was employed to agitate
aqueous and organic solutions.
A Spectra A-300 atomic absorption spectrometer
(Varian, Australia) was employed to analyse the metal
remaining in the aqueous phase.
All pH measurements were carried out using a pH-
meter GLP 22 (Crison, Spain) equipped with a com-
bined electrode model 50-02 (Crison).
IR spectra, which were obtained through the use of
with KBr pellets, were recorded on a Satellite FTIR
equipped with a WINFIRST LITE software package form
Mattson Instruments.
The 1H NMR and 13C NMR spectroscopy was
performed on a Brucker DPX200. Samples were run in
CD3CN (for Pd-3f) or CDCl3 for (Pd-3c).
derivatives, as well as silver(I)ꢀ
and several Mꢀtriisobutylphosphine sulfide complexes
(MꢂAu, Ag or Hg) [10ꢀ12].
/
thiourea complexes [9]
/
/
/
In this study, the thermodynamic characterisation of
the solvent extraction of palladium(II) and gold(III)
from aqueous chloride media by N-benzoyl-N?,N?-
diethylthiourea in CHCl3 is undertaken in order to
determine the stoichiometry of the complexes formed in
the organic solution and their corresponding extraction
constants.
Taking advantage of the palladium complexation
properties of these ligands, the solid complexes bis-
(1,1,-diheptadecyl-3-benzoyl-thioureate)palladium(II)
(Pd-3c) and bis(1,1,-diheptadecyl-3-benzoyl-thiourea-
te)palladium(II) (Pd-3f) were obtained and isolated
Elemental analyses were performed using a CHNS-O
elemental analyser EA-1108 from Fisons.
from a mixture of ethanolꢀmethanol solutions. The
/
metallic compounds were characterised and the struc-
ture of the crystal Pd-3f was redetermined by means of
XRD and compared with previous published results
[13]. Pd-3c and Pd-3f were used as ionophores in an ISE
system and the potentiometric response of the ion
selective membrane to the aqueous chlorocomplexes of
precious metals was studied. The selectivity of the
potentiometric response is related to the distribution
of the metal between the solution and the membrane
phases [14].
ESI MS spectra were obtained with a Finnigan
AQATM (ThermoQuest, Manchester, UK). The tem-
perature of the probe was 170 8C and the voltage of the
source was 30 eV. The mobile phase was a mixture of
acetonitrile:water 70:30 and the flow rate 150 ml minꢁ1
.
X-Ray data were collected on a Nonius CAD4 single-
crystal diffractometer Potentiometric responses were
measured with a digital pH/mV meter (micropH-2002,
Crison) and registered on a Yokogawa (LR 4120,
Japan) strip-chart recorder. In order to test different

M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ
/1437
1431
units simultaneously, we used a custom built three-way
commuter.
137.1 (Ph); 170.5, 171.5 (CSꢀ
/
, COꢀ
/
). IR (cmꢁ1) 2923,
2852 (Cꢀ
/
H st), 1492, 1427.
C24H30N4S2O2Pd (Pd-3f) (577.07) Anal. Calc. for: C,
49.95; H, 5.24; N, 9.71; S, 11.11%. Found: C, 50.01; H,
2.3. Procedure for the liquidꢀ
/
liquid distribution studies
5.13; N, 9.63; S, 10.93%. Yield: 80.9%. M.p.: 128ꢀ
130 8C.
1H NMR d (ppm) 8.30 (m, 4H, Ph), 7.63ꢀ
6H, Ph); 3.80ꢀ3.40 (m, 8H, CH2); 1.3 (m, 12H, CH3).
13C NMR d (ppm) 11.7, 12.3 (CH3); 46.2, 47.3 (CH2);
128.3, 129.4, 131.8, 136.8 (Ph); 170.3, 170.5 (CSꢀ
COꢀ H st), 1517, 1456,
). IR (cmꢁ1) 2981, 2931 (Cꢀ
/
Experiments were carried out by shaking equal
volumes (10 ml) of the aqueous and organic phases at
known concentrations in stoppered glass tubes. After
equilibrium was reached, the liquid phases were sepa-
rated and an aliquot was taken from the aqueous phase
for analysis by atomic absorption spectrometry (AAS).
Equilibrium time was determined in preliminary experi-
ments by measuring the metal content in the aqueous
phase as a function of time until the metal concentra-
tions in the aqueous solutions did not vary. The
equilibrium times needed for the extraction of gold
and palladium were 16 h and 5 h, respectively. The
equilibrium metal concentration in the organic phase
was calculated as the difference between the initial and
remaining concentrations in the aqueous phase. In some
experiments, metal concentration in the organic phase
was determined after stripping in order to verify the
mass balance.
/
7.50 (m,
/
/
,
/
/
1411, 1372. ESꢄ MS m/z (Fragment Ion) 577.4 (Mꢀ
/
[H]ꢄ).
2.5. Procedure for crystal structure determination
Unit cell dimensions were determined from the
angular settings of 25 reflections in the 105
/
u 5158
/
range. The space group was inferred by systematic
absences to be P21/n. Intensities were measured using
the v ꢀ2u scan technique with a scan angle of 1.58 and a
/
variable scan rate of 60 s per reflection. The intensity of
the primary beam was checked throughout the collec-
tion by monitoring three standard reflections every 60
min. Profile analysis was performed on all reflections
[17]. Symmetry-equivalent reflections were averaged and
Lorentz and polarisation corrections were applied. The
structure was solved by direct methods using SHELXS-97
[18]. Isotropic least-square refinement on F2 was
obtained using SHELXS-97 [19]. During the final stages,
all positional parameters and the anisotropic tempera-
ture factors of all the non-H atoms were refined. The H-
atoms were located by Fourier difference synthesis and
refined free or riding with common isotropic thermal
The initial metal concentration in the aqueous phase
was varied from 1ꢃ
total ligand (3f) concentration in the organic phase was
varied within the range 1ꢃ 1ꢃ
10ꢁ5ꢀ 10ꢁ3 mol. lꢁ1. In
/
10ꢁ4 to 1ꢃ
/
10ꢁ3 mol lꢁ1 and the
/
/
/
order to study the pH dependence of the extraction
process, experiments were performed with the aqueous
solution at both pH 1 and 2. All experiments were
performed at least in duplicate at a controlled room
temperature (r.t.) of 2291 8C.
/
2.4. Preparation of compounds Pd-3f and Pd-3c
2
2
parameters. The function minimised was [(SwFo ꢁ
/
Fc )/
2
2
Stoichiometric quantities of metal and ligand in a 1:2
ratio were mixed in order to prepare the Pd-3c and Pd-3f
complexes [16]. N-benzoyl-N?,N? -diheptadecylthiourea
palladium complex was obtained by dropwise addition
of a methanolic solution of 0.333 mmol of Na2PdCl4 to
an ethanolic solution of the ligand (0.665 mmol). The
mixture was then shaken during 10 min and left to
Sw(Fo )]1/2 where wꢂ
/
1/[s(Fo )ꢄ
/
(0.0661P)2ꢄ
/
1.3168P]
2
and with s(Fo ) from counting statistics and Pꢂ
(Fo , o)ꢄ2*Fc )/3. Atomic scattering factors were taken
/(Max
2
2
/
from International tables for X-ray Crystallography
[20]. Plots were made with the EUCLID package [21].
Geometrical calculations were made with PARST [22].
Relevant crystal and refinement data are collected in
Table 1. All calculations were made at the Scientific
Computer Centre of the University of Oviedo and on
the X-ray group computers.
precipitate at ꢁ22 8C. The resulting dark yellow solid
/
was vacuum filtered and washed with methanol. N-
benzoyl-N?,N?-diethylthiourea (3f) palladium complex
was prepared following the same steps as for Pd-3c. In
this case, 0.250 mmol of Na2PdCl4 was added to a 0.500
mmol 3f ligand resulting in the formation of orange
crystals.
2.6. Procedure for the potentiometric measurements
Solvent polymeric membranes were prepared by
mixing the ionophore (2 wt.%), the plasticiser (66
wt.%) and the polymer, PVC (32 wt.%). The mixture
was dissolved in 1 ml THF and cast in a 16 mm i.d. glass
ring [23]. The solvent was allowed to evaporate over-
night and a membrane was obtained from which 16 mm
diameter disks were cut and mounted on a Phillips IS-
561 electrode body (Glassblaserei, W. Moller, Switzer-
C84H150N4S2O2Pd (Pd-3c) (1418.68) Anal. Calc. for:
C, 71.12; H, 10.66; N, 3.95; S, 4.52. Found: C, 70.82; H,
10.29; N, 3.87; S, 4.24%. Yield: 74%. M.p.: 41ꢀ
1H NMR d (ppm) 8.29ꢀ
8.25 (m, 4H, Ph); 7.52ꢀ
(m, 6H, Ph); 3.86ꢀ3.74 (m, 8H, CH2); 1.3 (s, 120H,
CH2); 0.95ꢀ
0.88 (t, 12H, CH3). 13C NMR d (ppm) 14.1
(CH3); 27.0ꢀ31.9, 53.1, 51.9 (CH2); 127.9, 129.7, 131.3,
/
42 8C.
/
/
7.30
/
/
/

1432
M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ1437
/
Table 1
Crystal data and structure refinement for Pd-3f
determined in accordance with IUPAC recommenda-
tions [24ꢀ26].
/
Empirical formula
Crystal system
Space group
C24H30N4S2O2Pd
monoclinic
P21/n
3. Results and discussion
3.1. Liquidꢀliquid extraction
˚
a (A)
10.443 (3)
18.55 (2)
13.503 (2)
104.70 (2)
2530 (3)
4
˚
b (A)
/
˚
c (A)
b (8)
3
V (A )
˚
The metal distribution coefficient (D) between the
organic and aqueous phase was calculated as the ratio of
total metal concentration in both phases:
Z
Formula weight
T (K)
577.04
293 (2)
˚
l (A)
0.71073
1.515
1184
[Metal]org
dcalc (mg cmꢁ3
F(000)
)
Dꢂ
(1)
[Metal]aq
Crystal size (mm)
u Range for data collection (8)
Index ranges
0.33ꢃ0.23ꢃ0.20
The experimental data are presented in Figs. 2 and 3
in the form of log D versus log[3f] for several concen-
trations of the metals and at different initial pH of the
aqueous solutions. The great extractability of both
metals, especially gold(III), with 3f has restricted the
metal and ligand concentration ranges which can be
used as when the metal content is low, errors in the
analysis lead to inaccurate D values. The dependence of
1.91ꢀ25.97
/
05h 512, 05k 522,
ꢁ165l 516
5236/4959 [Rintꢂ0.0239]
96.8%
Reflections collected/unique
Completeness to 2uꢂ25.97
Max/min transmission
0.8363 and 0.7496
Full-matrix least-square on F2
4959/0/298
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
1.045
Final R indices [I ꢀ2s(I)]
R indices (all data)
Extinction coefficient
R1ꢂ0.0352, wR2ꢂ0.1000
R1ꢂ0.0606, wR2ꢂ0.1121
0.0094(6)
the metal extraction processes on pH range 1ꢀ2 was
/
studied varying the proton concentration in the aqueous
phase while keeping the total concentration of chloride
at 0.1 mol lꢁ1. As seen in Fig. 3 there is no dependence
of the gold(III) and palladium(II) extraction by 3f on the
pH.
In order to determine the composition and formation
constants of the extracted species, the distribution data
was analysed using LETAGROP-DISTR software [27]. The
best set of equilibrium constants for a given model is
found by a numerical analysis based on the minimisa-
tion of the square sum, U, defined as:
Largest difference peak and hole
ꢁ3
0.651 and ꢁ0.446
˚
(e A
)
land). In this study, between calibrations we stored
electrodes which were either dry or soaked in diluted
10ꢁ4 M Pd(II) chloride solution after having tested
different conditioning solutions (10ꢁ3 or 10ꢁ4 M Pd(II)
chloride) for periods of between 2 h and 24 h without
observing any significant improvement in their slope
and stability.
The membranes were conditioned for between 2 h and
24 h in 10ꢁ3 or 10ꢁ4 M PdCl4 solutions or were kept
dry.
X
2
U ꢂ
(log Dcalc ꢁlog Dexp
)
(2)
2ꢁ
N
An Orion double-junction AgꢀAgCl electrode (90-02-
/
00) was used as the reference electrode. The outer
compartment was filled with the same solution as in
the test solution.
Different volumes of standard analyte solutions were
added to a 10.0 ml aliquot of 0.100 M KClꢀHCl pH 3.0
/
to determine the response of the ISE to precious metal
chlorocomplexes. Differences in potential were mea-
sured with the pH/mV meter and registered on the strip-
chart recorder. All experiments were carried out in a
thermostatic cell at 25.090.1 8C. Calibration plots were
/
obtained by representing the potential of the electrode
versus the logarithm of the concentration of the analyte
in the sample solution. Response time was calculated as
the time elapsed to be within 1 mV of the steady-state
potential. Analytical parameters, i.e. slope of the linear
portion, practical detection limit (PDL), linear range,
operational lifetime and selectivity coefficients were
Fig. 2. Experimental distribution data of gold(III) and palladium(II)
at pH 1. The total ligand concentration in the organic phase was varied
from 3.16ꢃ
/
10ꢁ5 to 1ꢃ10ꢁ3 M.
/

M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ
/1437
1433
related to the equilibria involved in the system i.e.
hydrolysis of the metal, formation of the metal chlor-
ocomplexes. The results of the numerical analysis are
given in Table 2. The goodness-of-fit for any model may
be evaluated from the values of U and s (log D) defined
as s (log D)ꢂ
/
(U/Nꢁ
NK)1/2 where NK is the number of
/
constants to be adjusted.
The best fit to the experimental data is given by
models consisting of one species with a 1:1 stoichiometry
i.e. AuCl3L and PdCl2L. As can be seen in Table 2, the
model formed by the species with stoichiometries 1:1
and 1:2 (metal:ligand) gave the same values of U and s
as the models formed by the 1:1 species. This result
means that the addition of the 1:2 species to the latter
model did not improve the fit of the data and, hence, the
most realistic model is that composed by 1:1 species.
Previous studies on gold(III) solvent extraction from
aqueous chloride media by N-benzoylthiourea deriva-
tives [4] found that the species responsible for gold
extraction by 3f in cumene was AuCl3L2 and its
Fig. 3. Experimental distribution data of gold(III) and palladium(II)
as a function of pH. The total ligand concentration in the organic
phase was varied from 1.0ꢃ
/
10ꢁ4 to 1.0ꢃ10ꢁ3 M.
/
where Dexp represents the experimental values of the
distribution coefficient and Dcalc is the corresponding
value calculated by the program solving the mass
balance equation for the ligand, the metal and Hꢄ,
assuming a proposed set of chemical species and
formation constants. Np is the number of experimental
points.
extraction constant log K2ꢂ
/
5.990.4. In our case, the
/
use of chloroform, which is more polar than cumene,
favours the formation of 1:1 species. Muhl et al. [2]
¨
studied the extraction of palladium and gold (10ꢁ4 M)
with N-benzoyl-N?,N?-dibuthylthiourea (3e) in benzene,
whose concentration was varied from 2.5ꢃ
10ꢁ4 to
/
10ꢁ2 M. From the representation of log D versus
log[3e], they obtained slopes of two for palladium and
approximately one for gold which indicated the forma-
tion of a palladium species of 1:2 (metal:ligand)
stoichiometry and a 1:1 gold: ligand complex. The
In the experimental conditions used in this work (0.1
M NaCl and low pH), palladium and gold are present in
2ꢁ
the aqueous solution as the chlorocomplexes PdCl4
ꢁ
and AuCl4 and the experiments were performed at
such low metal concentration that only mononuclear
species are present. Then, the metal extraction reaction
can by described by using the following equation:
differences in the results of Muhl et al. with respect to
¨
the stoichiometry of the palladium species is attributable
to the higher ligand concentrations at which they have
conducted their experiments.
The extraction of gold by N-benzoylthiourea deriva-
tives can take place after the reduction of gold(III) to
gold(I) by the ligand as was found in the characterisa-
MeCln4ꢁ ꢄqHꢄ ꢄrL(org)
U MeClsHqLr(org) ꢄ(4-s)Clꢁ
(3)
By considering ideal behaviour in the organic phase
and assuming the activity coefficients of the aqueous
solution to be constant, the equilibrium constant of the
reaction can be written as:
Table 2
[MeClsHqLr]org[Clꢁ](4ꢁs)
Results of the numerical treatment of experimental data on gold(III)
and palladium(II) extraction by N-benzoyl-N?,N?-diethylthiourea in
chloroform
Kqr ꢂ
(4)
[MeCln4ꢁ][Hꢄ]q[L]or rg
The distribution coefficient, D, for metal extraction
can be expressed by:
Model
Species
log K
U
s (log D)
Model 1 (Au AuCl3L
extraction)
4.0790.10
0.7165
(Npꢂ58)
0.1151
q
r
X X
Dꢂ
Kqr[Hꢄ]q[L]rorg[Clꢁ](4ꢁs)
(5)
AuCl3L2 7.78 max. 7.40
4.0790.06
1
1
Model 2 (Au AuCl3L
0.7165
(Npꢂ58)
0.413
0.1151
0.09
extraction)
Model 3 (Pd
extraction)
Hence, a plot of log D versus log L would give a
preliminary estimation of the stoichiometric coefficients
by slope analysis [28]. The values obtained from this
graphical treatment were considered as the starting
point of the numerical calculations and were introduced
into the input of the program along with the constants
PdCl2
L
3.7290.18
(Npꢂ51)
PdCl2L2 7.69 max. 8.29
PdCl2L 3.7390.20
Model 4 (Pd
extraction)
0.413
(Npꢂ51)
0.09

1434
M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ1437
/
Table 3
Selected interatomic and intermolecular bond distances (A) and angles
e.s.d. Fig. 5 shows the molecular structure of the
complex with the atomic labelling.
˚
(8) for complex Pd-3f
3.2.2. Potentiometric results
Potentiometric responses to PdCl4
accurately determined in 0.10 M KClꢀ
Bond distances
2ꢁ
have been
HCl pH 3.0,
PdꢀO?
PdꢀO
PdꢀS?
PdꢀS
2.015 (3)
2.016 (3)
2.224 (2)
2.231 (1)
/
which was chosen to ensure that more than 95% of the
added analyte was in the PdCl42ꢁ form. The influence of
the membrane composition and working conditions on
the electrode response to this palladium chlorocomplex
was evaluated by systematically changing certain vari-
ables. Both palladium ionophores (Pd-3c and Pd-3f)
were tested and a significantly greater signal stability
was achieved when Pd-3c was used as the ionophore.
The drift in potential obtained with Pd-3f membranes is
explained by the leaching of the ionophore into the test
solution due to the lower lipophilicity of ligand 3f in
comparison to 3c, which has longer alkylic side chains
(see Fig. 1). Subsequent experiments were mainly
focussed on Pd-3c based membrane ISEs. Preliminary
experiments were undertaken in order to evaluate the
optimum composition of the liquid membranes and
their conditioning. No significant improvement in the
slope and stability of the electrodes was made by varying
the time and the composition of the conditioning
solution. The best reproducibility of the Nerstian
response was obtained with 2 wt.% Pd-3c as the
ionophore and o-nitrophenyl octyl ether (o-NPOE) as
the solvent mediator.
Bond angles
O?ꢀPdꢀO
O?ꢀPdꢀS?
OꢀPdꢀS?
O?ꢀPdꢀS
OꢀPdꢀS
86.3 (1)
93.41 (8)
178.9 (1)
176.7 (1)
93.43 (8)
86.79 (4)
S?ꢀPdꢀS
tion by XRD of the crystal obtained by mixing a
ꢁ
AuCl4 solution with a solution of 3f in CHCl3 [29].
The 1:1 (gold:ligand) complex formed has a Au(I)
centre, one ligand molecule in which the sulfur atoms
are oxidised and one chloride atom.
3.2. Spectral properties
The IR spectra of the complexes Pd-3c and Pd-3f
described in this work are assigned in the Section 2.
Comparison of these spectra with those of ligands 3c
and 3f (not shown here) reveals the absence of the bands
corresponding to the Nꢀ
present in the ligands. The resonance signals in the H
NMR and in the 13C NMR spectra can be assigned
/
H and Cꢁ
/
O bonds which were
1
After testing all precious metal chlorocomplexes (Fig.
6), a Nernstian response was only obtained for
4ꢁ
AuCl42ꢁand PdCl4 in 0.100 M KCl pH 3.0 medium.
easily (see Section 2). The disappearance of the Nꢀ
/
H
signal in the H NMR spectra and the shift in the 13C
1
RhCl5(H2O)2ꢁ, PtCl6 and PtCl4 did not give any
2ꢁ
2ꢁ
NMR spectra of the signals assigned to the CꢀO and Cꢀ
/
/
response and, therefore, would not have interfered in the
2ꢁ
S bonds, which in the complex appear close together,
shows once more the co-ordination of the palladium
through the O and S atoms of the ligand.
measurement of PtCl4 in a mixed PGM solution. The
ꢁ
calibration parameters for AuCl4 potentiometric re-
sponses were determined by using the optimum working
conditions described above for PtCl42ꢁ. The electrodes,
with membranes containing 2% of Pd-3c ionophore,
The ESꢄ MS spectrum of the Pd-3f complex shows a
peak at m/zꢂ577.4 which corresponds to a 1:2 (me-
/
tal:ligand) stoichiometry. In Fig. 4, the experimental and
calculated isotopic patterns of compound Pd-3f (which
agree with the presence of the Pd atom in the complex)
are depicted showing the complete correlation between
them. The high degree of fragmentation of Pd-3c makes
it difficult to interpret the mass spectrum (not shown).
presented Nernstian slopes of 59.691.9 mV per decade
/
for AuCl4ꢁ. However, the sensitivity and linear dynamic
range are very different for the two chlorocomplexes.
The measured PDL value for AuCl4ꢁ was 8.8ꢃ
/
10ꢁ8
10ꢁ4 M. A selectivity
3.4 was calculated taking
M
2ꢁ
while for PdCl4 it was only 2ꢃ
/
pot
coefficient of K
/
ꢂ
/
ꢁ
/
AuCl-4; PdCl24ꢁ
ꢁ
AuCl4 as being the primary anion. These selectivity
coefficients were calculated by the matched potential
method [30,31]. The greater response of the ISE system
for gold is a result of the higher extraction constant
3.2.1. X-ray structure determination
In order to achieve more accurate solid state struc-
tural data for the complex compound, the crystal
structure of Pd-3f was determined. The structure found
agrees with previously published results [13]. In Tables 1
and 3, the crystallographic data and selected bond
length and angles are presented. That the new determi-
nation is a significant improvement is shown by the R
factor, the hydrogen atom locations and the refinement
of the H bonding description, and the overall decrease in
obtained for gold in the liquidꢀliquid system.
/
4. Conclusions
Solutions of N-benzoyl-N?,N?-diethylthiourea (3f) in
CHCl3 were used in the solvent extraction of gold (III)

M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ
/1437
1435
and palladium (II) from aqueous chloride media (0.1 M
NaCl) giving high extraction rates especially for gold.
By treating equilibrium data with the LETAGROP-DISTR
software we have concluded that gold and palladium
extraction can be explained by the formation of 1:1
(metal:ligand) species with extraction constants of log K
the aqueous solution does not affect the metal extraction
process in the 1ꢀ2 pH range
The solid complexes bis(1,1,-diheptadecyl-3-benzoyl-
thioureate)palladium(II) (Pd-3c) and bis(1,1,-diheptade-
cyl-3-benzoyl-thioureate)palladium(II)Pd-3f were pre-
/
pared and isolated from a mixture of ethanolꢀ
/
methanol solution obtaining yields of 74 and 80.9%,
(AuCl3L)ꢂ
/
4.07 and log K (PdCl2L)ꢂ3.73. The pH of
/
Fig. 4. ESꢄ MS mass spectra of Pd-3f. Comparison of the (b) found and (c) calculated isotopic pattern of compound Pd-3f.

1436
M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ1437
/
5. Supplementary data
Further details of the complex Pd-3f structure deter-
mination are available as a cif (crystallographic infor-
mation file) magnetic file from The Director, CCDC, 12
Union Road, Cambridge CB2 1EZ, UK (fax: ꢄ44-
/
1223-336033; e-mail: deposit@ccdc.cam.ac.uk or www:
http://www.ccdc.cam.ac.uk) on request quoting deposi-
tion number 172658.
Acknowledgements
This study has been part-financed by CYCIT projects
(QUI99-0749-C03-03 and BQU2000-0219). We would
Fig. 5. Molecular structure of bis(1,1,-diheptadecyl-3-benzoylthiour-
eate)palladium(II) with the atom numbering scheme.
like to thank J. Delo´s for carrying out the liquidꢀliquid
/
distribution experiments and Dr. A. Roglans for her
help in the interpretation of MS spectra.
respectively. The characterisation of these complexes
was performed by NMR, FTIR and MS analyses and
their stoichiometry was confirmed to be 1:2 (metal:li-
gand) by MS.
References
[1] R.G. Pearson, J. Am. Chem. Soc. 85 (1963) 3533.
[2] P. Muhl, K. Gloe, F. Dietze, L. Beyer, Z. Chem. 26 (1986) 81.
¨
The structure of the Pd-3f crystal was redetermined by
means of XRD. The R factor, hydrogen atom locations
and the refinement of the H bonding description, and
the decrease in e.s.d. are significant improvements.
The Pd-3c and Pd-3f complexes were used as iono-
phores in the development of polymeric membrane
electrodes. The membranes prepared with 2% of Pd-3c
[3] F.Z. El Aamrani, A. Kumar, J.L. Cortina, A.M. Sastre, Anal.
Chim. Acta 382 (1999) 205.
[4] F.Z. El Aamrani, A. Kumar, L. Beyer, J.L. Cortina, A.M. Sastre,
Solv. Extr. Ion Exch. 16 (6) (1998) 1389.
[5] E. Antico´, A. Masana, M. Hidalgo, V. Salvado´, M. Valiente,
Hydrometallurgy 35 (1994) 343.
[6] F.Z. El Aamrani, A. Kumar, L. Beyer, J.L. Cortina, A.M. Sastre,
Hydrometallurgy 50 (1998) 315.
2ꢁ
gave Nernstian responses to PdCl4 and AuCl4ꢁ, with
[7] D. Gao, J.-Z. Li, R.-Q. Yu, G.-D. Zheng, Anal. Chem. 66 (1994)
2245.
[8] C. Palet, M. Munoz, S. Daunert, L.G. Bachas, M. Valiente, Anal.
˜
an average of ꢁ
1.5ꢃ
10ꢁ4 M for palladium, and a slope of ꢁ
per decade and a PDL of 8.8ꢃ
10ꢁ8 M for gold.
/
30.99
/
0.8 mV per decade and a PDL of
/
/
59.6 mV
/
Chem. 65 (1993) 1533.
Fig. 6. Potentiometric response. Selectivity pattern of noble metal chlorocomplexes.

M. Dom´ınguez et al. / Polyhedron 21 (2002) 1429ꢀ
/1437
1437
[9] F.Z. El Aamrani, A. Sastre, M. Aguilar, L. Beyer, A. Florido,
Anal. Chim. Acta 329 (1996) 247.
[20] International Tables for X-Ray Crystallography, 1974, vol. IV.
Kynoch Press, Birmingham: (Present distributor: Kluwer Aca-
demic, Dordrecht).
[10] A. Florido, L.G. Bachas, M. Valiente, M. Villaescusa, Analyst
119 (1994) 2421.
[21] A.L. Spek, in: D. Sayre (Ed.), Computational Crystallography,
Clarendon, Oxford, 1982, p. 528.
[22] M. Nardelli, Comput. Chem. 7 (1983) 95.
[11] J. Bricker, S. Daunert, L.G. Bachas, M. Valiente, Anal. Chem. 63
(1991) 1586.
[12] S. Daunert, A. Florido, J. Bricker, W. Dunaway, L.G. Bachas, M.
Valiente, Electroanalysis 5 (1993) 839.
[23] G.J. Moody, R.B. Oke, J.D.R. Thomas, Analyst 95 (1970) 910.
[24] G.G. Guibault, R.A. Durst, M.S. Frant, H. Freiser, E.H. Hansen,
T.S. Light, E. Pungor, G. Rechnitz, N.M. Rice, T.J. Rohm, W.
Simon, J.D.R. Thomas, Pure Appl. Chem. 48 (1976) 129.
[25] Y. Umezawa, K. Umezawa, H. Sato, Pure Appl. Chem. 67 (1995)
508.
[13] G. Fitze, L. Beyer, J. Sieler, R. Richter, J. Kaiser, E. Hoyer, Z.
Anorg. Allg. Chem. 433 (1977) 237.
[14] E. Bakker, P. Buhlmann, E. Pretsch, Chem. Rev. 97 (1997) 3083.
¨
[15] H. Hartmann, I. Reuther, J. Prakt. Chem. 315 (1973) 144.
[16] J. Hartung, G. Weber, L. Beyer, R. Szargan, Z. Anorg. Allg.
Chem. 523 (1985) 153.
[26] R.P. Buck, V.V. Cosofret, Pure Appl. Chem. 65 (1993) 1849.
[27] D.H. Liem, Acta Chem. Scand. 25 (1971) 1521.
[28] M. Aguilar, in: S. Alegret (Ed.), Developments in Solvent
Extraction, Ellis Horwood, Chichester, 1988.
[17] (a) M.S. Lehman, F.K. Larsen, Acta Crystallogr. Sect. A 30
(1974) 580;
(b) D.F. Grant, E.J. Gabe, J. Appl. Crystallogr. 11 (1978) 114.
[18] G.M. Sheldrick, SHELXS-97. Program for the Solution of Crystal
Structures, University Gottingen, 1997.
[29] W. Bensch, M. Schuster, Z. Anorg. Allg. Chem. 611 (1992) 99.
[30] V.P.Y. Gadzepko, G.D. Christian, Anal. Chim. Acta 164 (1984)
279.
¨
[19] G.M. Sheldrick, SHELXS-97. Program for the Refinement of
Crystal Structures, University Gottingen, 1997.
[31] A.S. Attiyat, A.M. Kagry, M.A. Badawy, H.R. Hanna, Y.A.
Ibrahim, G.D. Christian, Electroanalysis 2 (1990) 119.
¨