Competitive bulk liquid membrane transport and solvent extraction of some transition and post-transition metal ions using acylthiourea ligands as ionophores

Article abstract of DOI:10.1039/b603802f

Competitive transport experiments involving metal ions from an aqueous source phase through a chloroform membrane into an aqueous receiving phase have been carried out using a series of acylthiourea ligands as the ionophore present in the organic phase. The source phase contained equimolar concentrations of cobalt(ii), nickel(ii), copper(ii), zinc(ii), silver(i), cadmium(ii) and lead(ii) with the source and receiving phases being buffered at a number of different pHs. Transport selectivity was observed for silver(i) in all but one case. Selectivity for silver(i) was also observed in a series of parallel solvent extraction experiments carried out under similar source and membrane phase conditions. An X-ray diffraction study of the silver(i) complex of the N,N-dibutyl-N′-benzoylthiourea from this series is reported. There are two crystallographically independent molecules in the asymmetric unit and each molecule consists of four silver(i) ligand complex units, hence giving rise to Z = 8. In each molecule, both chelate and monodentate co-ordination modes of complexation to silver(i) are evident. Hence, each silver is co-ordinated to a sulfur and oxygen atom from one ligand and to a shared (bridging) sulfur from another ligand. Each silver atom is thus co-ordinated to three donor atoms. the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006.

Full text of DOI:10.1039/b603802f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Competitive bulk liquid membrane transport and solvent extraction of
some transition and post-transition metal ions using acylthiourea ligands
as ionophores
Michael M. Habtu,a Susan A. Bourne,b Klaus R. Kocha and Robert C. Luckay*a
Received (in Durham, UK) 14th March 2006, Accepted 22nd June 2006
First published as an Advance Article on the web 13th July 2006
DOI: 10.1039/b603802f
Competitive transport experiments involving metal ions from an aqueous source phase through a
chloroform membrane into an aqueous receiving phase have been carried out using a series of
acylthiourea ligands as the ionophore present in the organic phase. The source phase contained
equimolar concentrations of cobalt(II), nickel(II), copper(II), zinc(II), silver(I), cadmium(II) and
lead(II) with the source and receiving phases being buffered at a number of different pHs.
Transport selectivity was observed for silver(I) in all but one case. Selectivity for silver(I) was also
observed in a series of parallel solvent extraction experiments carried out under similar source
and membrane phase conditions. An X-ray diffraction study of the silver(I) complex of the
N,N-dibutyl-N0-benzoylthiourea from this series is reported. There are two crystallographically
independent molecules in the asymmetric unit and each molecule consists of four silver(I) ligand
complex units, hence giving rise to Z = 8. In each molecule, both chelate and monodentate co-
ordination modes of complexation to silver(I) are evident. Hence, each silver is co-ordinated to a
sulfur and oxygen atom from one ligand and to a shared (bridging) sulfur from another ligand.
Each silver atom is thus co-ordinated to three donor atoms.
applications. A number of studies have looked at the factors
which influence the transport efficiency11–13 and one study has
Introduction
We recently published a preliminary report illustrating the
shown that the transport-limiting step differs from one system
to the next.14
highly selective Ag(I) bulk liquid membrane transport using
N,N-diethyl-N0-camphanylthiourea as well as the unusual
crystal structure formed by this complex.1 The N,N-dialkyl-
N0-aroylthioureas have long been known since their first
synthesis by Neucki in 1873.2 The co-ordination chemistry
To date, there have not been studies using acylthiourea
ligands for potential metal ion transport. We report herein the
first comprehensive study of the use of these ligands for metal
ion transport using bulk liquid membranes, with particular
and potential applications (solvent extraction) of mainly N,N-
dialkyl-N0-aroylthioureas with transition metals was first ex-
reference to the results which were obtained with the ligand
N,N-diethyl-N0-camphanylthiourea.1 This ligand showed re-
plored by the groups of Hoyer and Beyer3 and later Konig,
¨
markably high and selective transport rates for Ag(I) in the
Schuster et al.4 explored the similar chemistry of these ligands
presence of several other transition metal ions. We also carried
with the platinum group metals. We have exploited these
out comparative competitive liquid–liquid extraction experi-
ligands for the analytical and process chemistry of mainly
ments using the above ligands with the seven metal ions. The
the platinum group metals.5,6 The favourable physicochemical
crystal structure of the silver(I) complex with HL1 is also
properties of selected N,N-dialkyl-N0-acylthioureas have re-
described.
sulted in the convenient HPLC determination of platinum
group metals.7
In this study we used the three phase arrangement and the
same cell design as that of Lindoy et al.,15 which involves two
There have been many reports of the transport of transition
buffered aqueous phases (source and receiving phases) sepa-
and post-transition metal ions through bulk liquid membranes
rated by an immiscible organic phase (chloroform) incorpor-
using synthetic ionophores. These included the use of mixed-
ating the ionophore. We have buffered the source and
donor acyclic ligands,8 oxygen–nitrogen donor macrocycles9
receiving phases at a number of pHs to examine the effect
and thioether donor macrocycles.10 The objective of these
which this has on the transport of metal ions.
studies has been to obtain better reagents for the separation
of metal ions in mining, medical and a number of analytical
Experimental
a Department of Chemistry, University of Stellenbosch, Private Bag
X1, Matieland, 7602 Stellenbosch, South Africa. E-mail:
Materials
RCLuckay@sun.ac.za; Fax: þ27 21 808 3849; Tel: þ27 21 808
All reagents were of analytical grade and used without further
3333
b Department of Chemistry, University of Cape Town, Private Bag,
Rondebosch, 7701, South Africa
purification. All ligands were synthesised according to the
method of Douglass and Dains16 and characterised by means
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1155–1162 | 1155

View Article Online
of elemental analysis, IR and NMR, as described pre-
viously.5,17 All aqueous solutions were prepared using de-
ionised water. Chloroform used for the membrane phase
phase to the receiving phase over this period was determined
by atomic absorption spectroscopy (AAS) and confirmed by
inductively coupled plasma-optical emission spectroscopy
(ICP-OES). The experimental results obtained from both
techniques did not differ by more than 4%.
was presaturated with water by shaking
a two phase
water–chloroform mixture, then removing the aqueous phase.
Small samples for analysis were taken from both the source
and receiving phases of each duplicate run after each experi-
ment and analysed. The average flux rate, J (mol per 24 h), for
each transport experiment was calculated based on the quan-
tity of metal ions transported into the receiving phase in a 24 h
period. The transport results are quoted as the average values
obtained from the duplicate runs carried out in parallel; in all
cases the flux values obtained did not differ by more than 5%.
J values equal to or less than 2.2 ꢁ 10ꢂ8 mol per 24 h were
assumed to be within experimental error of zero and have been
ignored in the analysis of results.
Preparation and crystallization of [Ag(L1)]4
The ligand N,N-dibutyl-N0-benzoylthiourea, HL1 (0.10 g, 0.34
mmol) was dissolved in a 30 ml solvent mixture of aceto-
nitrile–water in a volume ratio of 5 : 1, respectively. Sodium
acetate (0.09 g, 0.684 mmol), dissolved in a minimum volume
of water, was added to the ligand solution and the resulting
reaction mixture was warmed under reflux in an oil bath at
50 1C for 45 min. To this solution was added a solution of
silver nitrate (0.06 g, 0.34 mmol), dissolved in a minimum
volume of water, drop-wise, after which the mixture was
heated to reflux at 60 1C for 1 h. At this stage, the colour of
the reaction mixture was milky which became clear on addi-
tion of additional acetonitrile.
For the ‘competitive’ metal ion extraction experiments, the
reagents used were identical to those employed in the transport
experiments. The aqueous phase was buffered at pH = 5.0
with sodium acetate–acetic acid buffer and contained equal
concentrations (1 ꢁ 10ꢂ2 mol dmꢂ3) of cobalt(II), copper(II),
lead(II), silver(I), cadmium(II), zinc(II) and nickel(II) as their
respective nitrate salts. The chloroform membrane contained
the ligands at concentrations of 1 ꢁ 10ꢂ3 mol dmꢂ3 and 2 ꢁ
10ꢂ3 mol dmꢂ3 in separate experiments and (as was done in
the transport experiments), 4 ꢁ 10ꢂ3 mol dmꢂ3 palmitic acid.
The aqueous source phase (3 ml) and the chloroform mem-
brane (15 ml) were contained in a tightly capped polytop vial,
wrapped with parafilm and aluminium foil to prevent light-
induced reduction of Ag(I). These vials were shaken at 120
cycles per minute for 24 h on a Labcon oscillating shaker at
25 1C. The determination of metal ions was performed by
AAS and confirmed by ICP-OES. The results are quoted as
the average value obtained from duplicate runs. Any apparent
extraction of a metal ion of less than 2% relative, was assumed
to be within the experimental error of zero and hence ignored
in the treatment of the results. In all cases, the values between
any two duplicate runs did not differ by more than 2%.
The reaction mixture was cooled to room temperature, after
which excess water (about 20 ml) was added to the flask, to
precipitate the complex and to dissolve the remaining sodium
acetate. The white precipitate which was obtained after over-
night storage in a refrigerator was collected by filtration and
recrystallised from an acetonitrile–chloroform (1 : 1) mixture.
Colourless crystals were obtained after 4 days of slow eva-
poration of the solvent mixture at room temperature. The
crystals were characterised by melting point determination,
elemental (C, H, N and S) analysis and X-ray structure
determination. Mp: 113–114 1C. (Found: C, 48.44; H, 5.83;
N, 7.03; S, 7.73%; C16H23N2SOAg required C, 48.13; H, 5.81;
N, 7.02; S, 8.03%.)
Membrane transport and solvent extraction
An aqueous source phase containing an equimolar mixture of
the seven metal ions cobalt(II), copper(II), lead(II), silver(I),
cadmium(II), zinc(II) and nickel(II), (10 cm3) and an aqueous
receiving phase (30 cm3) were separated by a water presatu-
rated chloroform membrane phase containing the ligand (50
cm3) and 4 ꢁ 10ꢂ3 mol dmꢂ3 palmitic acid. The concentration
of the metal ions was 1 ꢁ 10ꢂ2 mol dmꢂ3 and that of the
chosen ligand was 1 ꢁ 10ꢂ3 mol dmꢂ3 or 2 ꢁ 10ꢂ3 mol dmꢂ3 in
separate experiments. The cell details are the same as those
used by Lindoy et al.15
Crystallographic data collection and structure determination
A suitable crystal was mounted on a thin glass fibre and coated
in silicone-based oil to prevent decomposition. Data were
collected on a Nonius Kappa CCD diffractometer using
graphite monochromated Mo Ka radiation (l = 0.7107 A)
with a detector to crystal distance of 45 mm. 663 oscillation
frames were recorded, each of width 11 in diameter, followed
by 422 frames of 11 width in o (with k a 0). Crystals were
indexed from the first ten frames using the DENZO package18
and positional data were refined along with diffractometer
constants to give the final cell parameters. Integration and
scaling (DENZO, Scalepack18) resulted in unique data sets
corrected for Lorentz-polarisation effects and for the effects of
crystal decay and absorption by a combination of averaging of
equivalent reflections and an overall volume and scaling
correction. Crystallographic data are recorded in Table 1.
The structure was solved using SHELXS-9719 and developed
via alternating least squares cycles and Fourier difference
synthesis (SHELXL-9719) with the aid of the interface
The membrane phase, the source phase and the receiving
phase were then gently transferred in this respective order into
the cells. The cells were thermostated at 25 1C and stirred at
10 rpm by means of a coupled single geared synchronous
motor. Under these conditions, not only was the stirring
process consistent, but also the interfaces between the organic
membrane and the two aqueous phases remained flat and well
defined and transport allowed to take place, against a back
gradient of protons. The cells were covered with cover slips in
order to prevent evaporation of solvents over the 24 h period
and then entirely covered by aluminium foil in order to
prevent the light-induced reduction of Ag(I) in the source
phase. All transport experiments were terminated after 24 h
and the amount of metal ion transported from the source
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
1156 | New J. Chem., 2006, 30, 1155–1162

View Article Online
Table 1 The crystallographic data and structure refinement for
tetra[(N,N-dibutyl-N0-benzoylthiourea) silver(I)] complex
Empirical formula
Formula weight
T/K
C64H92Ag4N8O4S4
1597.216
193(2)
l/A
Crystal system and space group
Unit cell dimensions
0.710 73
Monoclinic, P21/n
a = 17.805(4) A a = 901
b = 21.759(4) A b = 96.34(3)1
c = 36.438(7) A g = 901.
14 030(5)
V/A3
Z, Calculated density
Absorption coefficient. (m/mmꢂ1
F(000)
8, 1.516 g cmꢂ3
1.269
6560
)
Crystal size
Y range for data collection
Limiting indices
0.12 ꢁ 0.20 ꢁ 0.25 mm
2.85 to 25.001
ꢂ21 r h r 21, ꢂ25 r k r 25,
ꢂ43 r l r 43
Completeness to y = 27.87
Max. and min. transmission
Refinement method
Data, restraints, parameters
Goodness-of-fit on F2
Reflections collected, unique
Final R indices [I > 2s(I)]
R indices (all data)
94.2%
0.8626 and 0.7421
Full-matrix least-squares on F2
23 275, 0, 1334
1.030
41 998, 23 275 [R(int) = 0.0265]
0.0537, 0.0918
R1 = 0.1441, wR2 = 0.1692
0.709 and ꢂ0.537 e Aꢂ3
Largest diff. peak and hole
program X-SEED.20 There are two crystallographically inde-
pendent molecules in the asymmetric unit, giving rise to Z = 8
in P21/n.w All non-hydrogen atoms were modelled anisotropi-
cally, except for the carbon atoms in the butyl chains. Since the
carbon atoms in the butyl chains have very high thermal
motion, they were modelled with isotropic displacement para-
meters, and over two disordered positions in some cases (site
occupancy factors were independently modelled for each dis-
ordered chain). Hydrogen atoms on the phenyl rings were
assigned an isotropic thermal parameter of 1.2 times that of
their parent atom and refined using a riding model. No
hydrogens were placed on the butyl chains. The different views
and sections of the crystal structure are shown in Fig. 2 and 3.
Fig. 1 Acylthiourea ligands investigated as potential ionophores.
Cu(II). It can be seen from Table 2 that the percentage
transport of Ag(I) by HL1 is higher than that of HL2. The
sum (AgTr% þ AgTM%) [see Table 2 for key and abbrevia-
tions], which is a measure of the interaction between the
ligands and Ag(I), is slightly higher for HL1 than HL2. HL1
is more soluble in chloroform compared to HL2, so the higher
percentage transport of Ag(I) by HL1 can be the result of an
increased lipophilicity of the ligand because of its longer alkyl
substituents (butyl groups) compared to those of HL2 where
the substituents are ethyl groups. That HL3 is selective for
Cu(II) compared with Ag(I) can be attributed to the presence of
the two –OH groups attached to the alkyl end of the ligand. It
Results and discussion
In this section, ligands having similar structures are grouped
together for the purpose of the discussion. Structures of
ligands are shown in Fig. 1 and all transport results are shown
in Table 2.
Ligands HL1, HL2 and HL3 all posses a benzoyl group and
an acylthiourea motif in their structures but differ in their
substituent groups.
For HL1 and HL2, a source phase pH of 5.5 and a receiving
phase pH of 1.0 were adopted as optimum conditions, so that
the effect of varying the ligand concentration from 1 ꢁ 10ꢂ3 to
2 ꢁ 10ꢂ3 mol dmꢂ3 on the transport efficiency and selectivity
for Ag(I) could be studied. For HL3, we have studied the effect
of varying the ligand concentration as well as pH of the source
phase.
Under the experimental conditions employed, HL1 and HL2
are selective for Ag(I) whereas HL3 selectively transports
Fig. 2 Partial structure and numbering scheme for [Ag(L1)]4 com-
w CCDC reference number 609098. For crystallographic data in CIF
or other electronic format see DOI: 10.1039/b603802f
plex. Phenyl and butyl groups are omitted for clarity.
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1155–1162 | 1157

View Article Online
simultaneously coordinating with both Ag(I) and Cu(II) ions,
thus transporting Cu(II) into the receiving phase while simul-
taneously extracting Ag(I) into the membrane phase. There-
fore, we speculate that the ligand may be binding to Ag(I) ion
through the sulfur and oxygen donor atoms of the acylthio-
urea moiety, and perhaps to Cu(II) through the oxygen atoms
of the two –OH groups.
Attaching electron-withdrawing groups to the benzoyl moi-
ety and examining these effects on metal ion transport, is
reported for ligands HL4 and HL5. Experiments were carried
out with these two ligands, where the pH of the aqueous
phases and ligand concentration were varied. Unfortunately,
these ligands, however, differ greatly in their solubility in
chloroform. Ligand HL5 is very soluble, whereas ligand HL4
is only slightly soluble. Nevertheless, both ligands are selective
for Ag(I), with only a small percentage of Cu(II) also being
transported. Results indicate that ligand HL5 has a higher
efficiency and better selectivity for Ag(I).
At a ligand concentration of 2 ꢁ 10ꢂ3 mol dmꢂ3, the
optimum pH values of the source and receiving phase, which
resulted in selective and efficient transport of Ag(I) by HL5,
were found to be 5.5 and 1.0, respectively. Decreasing the pH
of the source phase from 5.5 to 5.0 did not bring about a
significant change in Ag(I) transport.
Fig. 3 The structure of [Ag(L1)]4 complex showing one of the two
molecules in the asymmetric unit.
is interesting and it shows that apparently small changes in
ligand structure have a significant effect on relative selectivity.
For HL3, the sum (AgTr% þ AgTM%) is greater than (CuTr%
þ
CuTM%). Only Cu(II) is transported into the receiving phase
When comparing ligand HL2 with ligands HL4 and HL5,
since they all have diethyl chains on nitrogen, ligand HL5
shows the highest transport of Ag(I) and the values for ligands
HL2 and HL4 are comparable.
by ligand HL3. This suggests that although about 95% of
Ag(I) is extracted from the source phase into the membrane
phase, it is not stripped into the receiving phase at pH = 1.0
(see Table 2). On the other hand, although Cu(II) was not
transported into the membrane phase as efficiently as Ag(I),
about 13% of the total Cu(II) originally present in the source
phase was transported into the receiving phase. Therefore, it
appears that the formation constant of HL3 must be substan-
tially higher for Ag(I) than Cu(II), not allowing the release of
Ag(I) from the membrane phase into the receiving phase.
Hence, HL3 is a better transport ionophore for Cu(II) com-
pared to Ag(I).
The effect of attaching electron-releasing substituents to
N,N-dialkyl-N0-benzoylthioureas on transport efficiency and
selectivity was studied using ligands HL6 and HL7. As can be
seen from their structures, both ligands possess methoxy
substituents at positions C2, C4 and C6. HL6 and HL7 are
both soluble in chloroform which makes these ligands suitable
candidates for transport studies.
The effect of varying the pH of the source phase from 5.0 to
5.5 and 6.0 was studied for these two ligands which resulted in
no significant change in the percentage transport of Ag(I) by
either ligand. Under the experimental conditions employed,
As can be seen from the table, the sum (CuTM% þ AgTM%)
exceeds 100%. This indicates that the ligand HL3 must be
Table 2 J values for the competitive metal ion transport studies involving HL1–HL10. The experimental conditions were: pH of the source phase
= 5.5, pH of receiving phase = 1.0 and concentration of ligand = 0.002 mol dmꢂ3
Metal ion transport, J/mol hꢂ1 ꢁ 10ꢂ7
HL1
HL2
HL3
HL4
HL5
HL6
HL7
H2L8
H2L9
HL10 a
Ligand number
Co(II)
Ni(II)
Cu(II)
Zn(II)
Ag(I)
Cd(II)
Pb(II)b
AgTr%
—
—
2.2
—
18
—
—
42
43
5.4
—
7.8
—
—
—
—
11
—
—
26
57
—
2.5
Z - N
—
—
5.4
—
—
—
—
—
95
13
16
Z = 0
—
—
1.0
—
9.8
—
—
23
—
—
1.8
—
20
—
—
48
36
4.4
—
11
—
—
2.7
—
13
—
—
31
38
6.6
5.8
4.7
—
—
1.7
—
12
—
—
29
35
4.2
2.3
6.9
—
—
—
—
0.6
—
—
1.5
95
—
—
—
—
—
—
—
—
—
—
57
—
5.0
None
—
—
—
—
34
—
—
81
12
12
7.2
6.8
AgTM
%
41
CuTr%
2.4
4.2
9.6
CuTM
%
Z(Ag(I)/Cu(II))c
Z - N
a
Ligand HL10 results from ref. 1 for comparison. b AgTr% = Percentage of Ag(I) transported into the receiving phase. AgTM% = Percentage of
Ag(I) transported into the membrane phase. CuTr% = Percentage of Cu(II) transported into the receiving phase. CuTM% = Percentage of Cu(II)
AgTr%
CuTr%
c
transported into the membrane phase. ZðAgðIÞ=CuðIIÞÞ ¼
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
1158 | New J. Chem., 2006, 30, 1155–1162

View Article Online
both HL6 and HL7 showed clear selectivity for Ag(I), with
ligand HL6 being only slightly more efficient than ligand HL7.
However, as can be seen from Table 2, the selectivity factor for
Ag(I) compared to Cu(II) is higher for ligand HL7 than ligand
HL6. Comparison of ligand HL6 with HL1 shows that ligand
HL1 transports Ag(I) better than ligand HL6 (both have
dibutyl chains). Ag(I) transport by HL7 results are comparable
to those of ligand HL2 (where both have diethyl chains).
The effect of varying the concentration of HL6 and HL7 on
transport efficiency and selectivity was also studied. The
selectivity factor, Z(Ag(I)/Cu(II)), decreases with an increase
in the concentration of the ligands from 1 ꢁ 10ꢂ3 to 2 ꢁ 10ꢂ3
mol dmꢂ3. A further increase in ligand HL6 concentration
from 2 ꢁ 10ꢂ3 mol dmꢂ3 to 4 ꢁ 10ꢂ3 mol dmꢂ3 also resulted in
a drastic decrease in Z(Ag(I)/Cu(II)). From the results ob-
tained, it is evident that the transport of Ag(I) by ligands
HL6 and HL7 is optimum when the ligand concentration is 2 ꢁ
10ꢂ3 mol dmꢂ3 and the pH value of the source phase and
receiving phase are set at 5.5 and 1.0, respectively.
likely that the ionophore is transporting the metal ion as a
deprotonated species and then becomes protonated again at
the membrane phase–receiving phase interface.
Competitive metal ion liquid–liquid extraction studies
In order to understand the membrane transport results better,
a comparative study of liquid–liquid extraction of Ag(I) from a
mixture of the metal ions: Co(II), Ni(II), Cu(II), Zn(II), Cd(II)
and Pb(II) with the above ligands was conducted.
All results are summarized in Fig. 4. Ligands are grouped in
exactly the same way as they were for the discussion of the
transport studies.
Under the conditions employed, consistent to the transport
results, HL1 and HL2 showed selective extraction for Ag(I)
compared with the other six metal ions. At a concentration of
2 ꢁ 10ꢂ3 mol dmꢂ3, ligand HL1 extracted about 91% of Ag(I)
and 4.5% of Cu(II). When the ligand concentration was
decreased by half, the percentage extraction of Ag(I) by ligand
HL1 was decreased as expected (only about 52% Ag(I) was
extracted). Although the decrease in the concentration of
ligand HL1 from 2 ꢁ 10ꢂ3 mol dmꢂ3 to 1 ꢁ 10ꢂ3 mol dmꢂ3
resulted in a decrease in the overall extraction efficiency for
Ag(I), the selectivity for Ag(I) over Cu(II) was enhanced. A
similar trend and percentage of Ag(I) extraction was observed
with ligand HL2.
Ligands H2L8 and H2L9 are both soluble in chloroform.
These are both mono-alkyl substituted ligands and are
grouped together and it was found that under the experimental
conditions employed, these ligands are not good at transport-
ing any of the seven metal ions. This also mirrors their poor
performance in the solvent extraction of the platinum group
metals.4
The crystal structure data of ligand HL1 and ligand HL2
with Ag(I) shows a 1 : 1 mole ratio stoichiometry.32 Therefore,
if it is assumed that this structure mirrors conditions in
solution, 2 ꢁ 10ꢂ3 mol dmꢂ3 (3 ꢁ 10ꢂ5 mol) of the ligands
(HL1 or HL2) is theoretically required to extract the total
amount of Ag(I) (1 ꢁ 10ꢂ2 mol dmꢂ3, 3 ꢁ 10ꢂ5 mol)
completely in the aqueous phase.
In the case of HL3, Ag(I) and Cu(II) were co-extracted by
92% and 41%, respectively; extraction results are consistent
with the transport results obtained, which showed that HL3
transported relatively more Ag(I) than Cu(II) into the mem-
brane phase (see Table 2). A decrease in the concentration of
HL3 from 2 ꢁ 10ꢂ3 mol dmꢂ3 to 1 ꢁ 10ꢂ3 mol dmꢂ3 did not
Under the experimental conditions reported in Table 2, 95%
of the Ag(I) ion initially present in the source phase was
extracted into the membrane phase using H2L8 but only about
1.5% of Ag(I) was stripped into the receiving phase. This
indicates that the formation constant of Ag(I) with H2L8 is
probably too high which thus inhibits the release of the
complexed Ag(I) from the membrane phase into the receiving
phase. Therefore, this ligand can be a good extracting reagent
but not a good transport ionophore for Ag(I). Ligand H2L9
did not transport any of the seven metal ions in the source
phase, but there is a substantial quantity of Ag(I) in the
membrane, this being once again most likely due to a large
formation constant with Ag(I).
Throughout the transport experiments, we have incorpo-
rated 4 ꢁ 10ꢂ3 mol dmꢂ3 palmitic acid in the membrane phase
and shown that at this concentration it does not transport any
of the metal ions in the source phase. Palmitic acid is widely
used in transport experiments involving a variety of ligands as
ionophores.9,21 Since the metal ions that we studied are in the
form of nitrate salts, NO3ꢂ may be transported as counter ions
along with the divalent metal ions into the receiving phase in
order to maintain electro-neutrality through ion pairing or
adduct formation. Therefore, to avoid the uptake of nitrate
ions, it becomes important to incorporate palmitic acid into
the chloroform membrane. An additional benefit of adding
lipophilicity in the form of the long-chain acid is to inhibit any
‘‘bleeding’’ of ‘‘partially’’ hydrophilic species (such as the
protonated or deprotonated ionophore and/or its correspond-
ing charged metal complex) from the organic membrane phase
into either of the aqueous phases.21–23
Fig. 4 Comparison of the competitive extraction studies involving all
ligands. The experimental conditions were: pH of the aqueous phase
= 5.0 and concentration of ligand = 2 ꢁ 10ꢂ3 mol dmꢂ3. 1 represents
HL1; 2 represents HL2, etc.
The fact that there is a large amount of metal ion in the
receiving phase compared to the amount of ionophore used
and because we have used the pH gradient method, it is most
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1155–1162 | 1159

View Article Online
significantly affect the percentage extraction of Ag(I) but
decreased the percentage extraction of Cu(II) from 41% to
31%. In order to understand this result, further studies of
ligand HL3 with Ag(I) and Cu(II) in solution are required.
At a ligand concentration of 2 ꢁ 10ꢂ3 mol dmꢂ3, the
extraction efficiency and selectivity of ligands HL1–HL3 for
Ag(I) follows the order: HL3 > HL2 D HL1 for efficiency and
HL2 > HL1 > HL3 for selectivity. Therefore, based on the
above results, we can say that varying the alkyl end of a
benzoylthiourea ligand from butyl to ethyl enhances selectivity
for Ag(I) extraction. However, introducing more oxygen
donor atoms by adding hydroxyl-groups to the alkyl end of
HL2 to give HL3, resulted in the extraction of Cu(II) together
with Ag(I).
At a ligand concentration of 2 ꢁ 10ꢂ3 mol dmꢂ3, HL4
extracted about 94% of Ag(I) and 3.6% of Cu(II). Similarly,
HL5 extracted about 92% of Ag(I) and 5.1% of Cu(II). These
results are consistent with the transport results obtained
showing that both ligands were selective for Ag(I) and Cu(II),
with major preference for Ag(I).
Comparison of the extraction efficiency of the mono-alkyl
substituted acylthioureas with the dialkyl-substituted analo-
gues shows that the former ligands are more efficient and
selective for the extraction of Ag(I).
The co-ordination chemistry toward the platinum group
metals reviewed by Koch6 on these ligands indicated that the
oxygen atom of the carbonyl group in mono-alkyl substituted
acylthioureas is not available for complex formation with
metal ions due to intra-molecular hydrogen bonding. There-
fore, based on this argument we can expect ligands H2L8 and
H2L9 to co-ordinate with Ag(I) through only the sulfur donor
atom. If this is the case, then we would expect the ligands to
extract Ag(I) in a 1 : 1 ligand to metal mole ratio. In other
words, 2 ꢁ 10ꢂ3 mol dmꢂ3 (3 ꢁ 10ꢂ5 mol) of the ligand would
be theoretically required to extract the total amount of Ag(I)
(1 ꢁ 10ꢂ2 mol dmꢂ3, 3 ꢁ 10ꢂ5 mol) in the aqueous phase.
However, inconsistent with our expectation, decreasing the
ligand concentration to 1 ꢁ 10ꢂ3 mol dmꢂ3 had no significant
effect on the percentage extraction of Ag(I) by these ligands.
However, interestingly, reducing the ligand concentration
further to 5 ꢁ 10ꢂ4 mol dmꢂ3 resulted in a decrease of Ag(I)
extraction almost by half. From the above results, it seems
that one mole of the ligand (H2L8 or H2L9) is extracting
almost two moles of Ag(I). In the absence of more evidence,
it is difficult to discuss these results any further.
Comparison of the extraction efficiency of HL4 and HL5 for
Ag(I) at a ligand concentration of 2 ꢁ 10ꢂ3 mol dmꢂ3 shows
the order HL4 > HL5. The extraction efficiency of ligands
HL4 and HL5 can also be compared with that of HL2 resulting
in the order HL4 > HL5 D HL2. At a ligand concentration of
2 ꢁ 10ꢂ3 mol dmꢂ3, the order of extraction selectivity for Ag(I)
compared with Cu(II) is HL2 > HL4 > HL5. The percentage
extraction results for HL4, HL5 and HL2 are very close to one
another. This means that the electron-withdrawing groups
The crystal structure for tetra[(N,N-dibutyl-N0-
benzoylthiourea) silver(I)] complex
ꢂ
(NO2 and Clꢂ) did not result in any discrimination on the
Crystallographic data for the above complex are shown in
Table 1. The crystal structure of the tetra[(N,N-dibutyl-N0-
benzoylthiourea) silver(I)] complex is shown in Fig. 2. Selected
bond lengths and bond angles are listed in Tables 3 and 4,
respectively. The structural analysis for this compound shows
extraction behaviour of the ligands.
On the other hand both ligands HL6 and HL7 have electron-
releasing methoxy substituents attached to the benzene ring.
Both ligands are efficient extraction reagents for Ag(I). Under
the conditions employed, consistent with the transport results,
the extraction results indicate selective extraction of Ag(I) by
both HL6 and HL7. These extraction results are consistent
with a 1 : 1 metal : ligand ratio as the actively extracted
complex. Although a decrease in the ligand concentration
from 2 ꢁ 10ꢂ3 mol dmꢂ3 to 1 ꢁ 10ꢂ3 mol dmꢂ3 resulted in
a decrease in the percentage extraction of Ag(I) by both
ligands, the selectivity for Ag(I) was enhanced as would be
expected.
Table
benzoylthiourea) Ag(I)] complex
3
Selected bond lengths [A] for tetra[(N,N-dibutyl-N0-
Bond
Length/A
Bond
Length/A
Ag1–Ag2
Ag1–S4
Ag1–O41
C41–N41
C41–S4
3.0510(9)
2.4616(17)
2.359(4)
1.313(8)
1.774(6)
1.242(7)
Ag5–Ag6
Ag5–S5
Ag5–O81
C81–N82
C81–S8
3.0806(9)
2.4266(17)
2.342(5)
1.330(9)
1.776(8)
1.231(8)
The extraction value for ligand HL6 is the same as that for
ligand HL1 and the extraction value for ligand HL7 is the same
as that for ligand HL2. The following percentage extraction
pattern is obtained for ligands HL2, HL4, HL5 and HL7
(which all have diethyl chains), HL4 > HL5 > HL2 > HL7,
but all four ligands have roughly the same extraction figure.
The extraction trends show that ligands H2L8 and H2L9 are
excellent reagents for the liquid–liquid extraction of Ag(I).
Under the conditions employed, consistent with the transport
results, the extraction results indicate that these ligands are
very efficient and selective for Ag(I) extraction compared to the
other six metal ions. At a ligand concentration of 2 ꢁ 10ꢂ3 M,
98% of Ag(I) was extracted by H2L8. When the ligand con-
centration was decreased to 1 ꢁ 10ꢂ3 mol dmꢂ3, the extraction
of Ag(I) decreased to 95%. A similar trend and percentages of
Ag(I) extraction were obtained with H2L9.
C42–O41
C82–O81
Table
benzoylthiourea) Ag(I)] complex
4
Selected bond angles [1] for tetra[(N,N-dibutyl-N0-
Bond
Angle/1
Bond
Angle/1
Ag1–Ag4–Ag3
S1–Ag1–S4
79.40(3)
148.63(6)
122.94(13)
77.70(5)
105.1(2)
117.3(4)
129.6(6)
125.0(5)
118.6(6)
125.5(6)
119.4(6)
115.0(6)
Ag5–Ag8–Ag7
S5–Ag5–S8
80.36(3)
146.26(6)
129.75(12)
77.50(5)
105.2(2)
124.0(4)
128.5(6)
125.7(5)
117.0(6)
126.9(6)
118.7(6)
114.1(6)
O11–Ag2–S2
Ag1–S1–Ag2
C11–S1–Ag1
C12–O11–Ag2
C11–N11–C12
N11–C11–S1
N11–C11–N12
O11–C12–N11
O11–C12–C13
N11–C12–C13
O51–Ag6–S6
Ag5–S5–Ag6
C51–S5–Ag5
C52–O51–Ag6
C51–N51–C52
N51–C51–S5
N51–C51–N52
O51–C52–N51
O51–C52–C53
N51–C52–C53
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
1160 | New J. Chem., 2006, 30, 1155–1162

View Article Online
Table 5 Average bond lengths [A] from other structures for comparison
Complex
Cs–S
Co–N
Cs–N
Co–O
Ref.
Ag4L4 (L = N,N-diethyl-N0-benzoylthiourea)
AgL3 ꢃ SH (L = N,N-diethyl-N0-benzoylthiourea)
{AgL1}4 (L = N,N-dibutyl-N0-benzoylthiourea)
1.776
1.681
1.768
1.302
1.381
1.313
1.341
1.417
1.361
1.247
1.218
1.238
30
31
This work
an empirical formula of [AgL1]4, forming an interesting
cluster.
C22ꢃ ꢃ ꢃN21ꢃ ꢃ ꢃO21, C32ꢃ ꢃ ꢃN31ꢃ ꢃ ꢃO31 and C42ꢃ ꢃ ꢃN41ꢃ ꢃ ꢃO41,
respectively.
The ligand HL1 coordinates to Ag(I) in both a mono- and
bidentate fashion within the [AgL1]4 cluster. The nearly
squarely orientated silver atoms in this molecule are alter-
nately bonded to each other with sulfur bridges forming an
eight-membered ring of alternating silver and sulfur atoms.
The bridging sulfur atoms lie above and below the plane
through the silver atoms in this configuration. The Ag–Ag
separations lie between 2.9687(9) and 3.0949(9) A and are thus
not significantly longer than in metallic silver (2.889 A). Each
Ag atom is co-ordinated to two S- and one O-atom and thus
has a co-ordination number of 3. The co-ordination geometry
around the silver atom is distorted trigonal as can be seen from
the bond angles around each silver atom (see Table 4). The Ag
atoms are not in the same plane with the coordinated O and S
atoms of the ligands. Ag1 lies 0.0288 A from the plane defined
by S1ꢃ ꢃ ꢃS4ꢃ ꢃ ꢃO41, Ag2 lies 0.0667 A from the plane defined by
S1ꢃ ꢃ ꢃS2ꢃ ꢃ ꢃO11, Ag3 lies 0.0462 A from the plane defined by
S2ꢃ ꢃ ꢃS3ꢃ ꢃ ꢃO21 and Ag4 lies 0.0033 A from the plane defined
by S3ꢃ ꢃ ꢃS4ꢃ ꢃ ꢃO31.
While the [AgL1]4 cluster packs with butyl chains in rela-
tively close proximity to one another, there are no direction-
specific interactions in this structure. The two crystallographi-
cally independent [AgL1]4 clusters differ most in the conforma-
tion of the butyl chains, as might be expected. The determined
bond lengths show that the bonding situation of the ligands in
the current Ag complex shows a widespread bond delocalisa-
tion in the chelate ring.
The bonding mode for the tetra[(N,N-dibutyl-N0-ben-
zoylthiourea) Ag(I)] complex is similar to the one reported
earlier for the Ag(I) complex with HL2 (HL2 = N,N-diethyl-
N0-benzoylthiourea)30 but different from the one related struc-
ture of a Ag(I) complex reported with N,N-diethyl-N0-ben-
zoylthiourea, [Ag(HL)3SH].31 In the latter case, the ligands act
as uncharged monodentate sulfur ligands and form a distorted
tetrahedral Ag(I) complex in which the bond lengths in the
ligands remain largely unchanged upon co-ordination. The
fourth co-ordination site is occupied by a hydrogen sulfide ion
which is presumably formed from the ligand during complex
formation. Similar monodentate sulfur ligand co-ordination to
silver through the sulfur atom had also been reported pre-
viously.32 A comparison of the important average bond
lengths for the molecular structure of tetra[(N,N-dibutyl-N0-
benzoylthiourea) Ag(I)] complex with the previously published
Ag(I) complexes with N,N-diethyl-N0-benzoylthiourea is
presented in Table 5.
The ligands co-ordinate as bidentate monoanions. The
conformation of the molecule with respect to the thio-carbonyl
and carbonyl moieties is definitely twisted as reflected by the
torsion angles. The S–Ag–S angles are between 138.27(6) and
148.63 (6)1. The usual Ag–S bond lengths for acylthiourea
Ag(I) complexes are in the range 2.45–2.54 A.24 In our case,
they vary within the ligand chelate rings from 2.457(2) to
2.510(2) A and are significantly longer than those that do not
form part of the ligand chelates [2.399(2) to 2.4304(9) A]. The
Ag–O separations in the chelate rings vary from 2.312(4) to
2.359(4) A.
Conclusion
Under the conditions employed, all the N,N-dialkyl-N0-ben-
zoylthioureas (HL), with the exception of N,N-di-(2-hydro-
xyethyl)-N0-benzoylthiourea, HL3, were efficient and selective
for Ag(I) transport. The N-alkyl-N0-benzoylthiourea (H2L8
and H2L9) ligands were good at transporting Ag(I) into the
membrane phase, but not into the receiving phase.
The competitive metal ion extraction results show that all
the ligands are very efficient in selectively extracting Ag(I) from
a mixture of the seven metal ions. The mono-alkyl substituted
aroylthioureas (HL8 and HL9) were not good at transporting
Ag(I) or any of the other six metal ions. However, these ligands
showed a profound selectivity and efficiency for the extraction
of Ag(I).
Within the chelate rings, there is only partial delocalisation
of the bond order. Even though all the C–S, C–O and C–N
bond lengths lie between the values for single and double bond
order, the C–S separations are longer [1.752(6) to 1.783(7) A)]
compared to those found in other acylthioureas25–27 and
thioamides.28,29 The C–N distances do not show the usually
observed equality and are all shorter than the average single
C–N bond length of 1.472(5) A.5 They are significantly longer
on the acyl side [1.330(9) to 1.392(19) A] than on the
acylthiourea side [1.296(10) to 1.330 (9) A]. The C–O bond
lengths [1.226(7) to 1.249(8) A)] suggest co-ordination of the
oxygen atom.
The C–N bonds between the chelate ring and the amino
group [1.330(9) to 1.392(9) A] show, like in all N,N-dialkyl-N0-
acylthioureas and their complexes, partial double bond char-
acter in the C–N(C4H9)2 group. The phenyl groups are rotated
by 9.29(0.87), 8.98(0.64), 3.69(0.47) and 6.04(0.60)1 with
respect to the planes defined by C12ꢃ ꢃ ꢃN11ꢃ ꢃ ꢃO11,
Finally, the crystal structure of the Ag(I) complex with N,N-
dibutyl-N0-benzoylthiourea ligand, HL1, showed interesting
monodentate and bidentate co-ordination modes. The struc-
ture shows that four silver ions are bound to four molecules of
the ligand.
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1155–1162 | 1161

View Article Online
12 J. D. Lamb, J. J. Christensen, S. R. Izatt, K. Bedke, M. S. Astin
and R. M. Izatt, J. Am. Chem. Soc., 1980, 102, 467.
13 T. M. Fyles, Can. J. Chem., 1987, 65, 884.
14 R. M. Izatt, R. L. Bruening, M. L. Bruening, G. C. LindH and J. J.
Christensen, Anal. Chem., 1989, 61, 1140.
Acknowledgements
We thank the University of Stellenbosch, the NRF (GUN
2053064) and THRIP (project 1562) for generous financial
support of this work. We also thank Helgard Raubenheimer
for useful comments of the manuscript.
15 P. S. K. Chia, L. F. Lindoy, G. W. Walker and G. W. Everett, Pure
Appl. Chem., 1993, 65, 521.
16 I. B. Douglass and F. B. Dains, J. Am. Chem. Soc., 1934, 56, 719.
17 M. C. Matoetoe and K. R. Koch, Magn. Reson. Chem., 1991, 29,
1158.
References
18 Z. Otwinowski and W. Minor, Methods Enzymol., 1996, 276, 307.
19 G. M. Sheldrick, SHELXL-97, Program for refinement of crystal
1 H. Berhe, S. A. Bourne, M. W. Bredenkamp, C. Esterhuysen, M.
M. Habtu, K. R. Koch and R. C. Luckay, Inorg. Chem. Commun.,
2006, 9, 99.
2 E. Neucki, Ber. Dtsch. Chem. Ges., 1873, 6, 598.
3 (a) L. Beyer, E. Hoyer, H. Hartman and J. Liebscher, Z. Chem.,
structures, University of Gottingen, Germany, 1997G. M. Shel-
¨
drick, SHELXS-97, Program for solution of crystal structures,
University of Gottingen, Germany, 1997.
¨
20 L. J. Barbour, J. Supramol. Chem., 2001, 1, 189.
21 M. Akhond and M. Shamsipur, J. Membr. Sci., 1996, 117, 221.
22 A. Safavi and S. Rastegarzadeh, Talanta, 1995, 42, 2039.
23 J. Kim, T. H. Ahn, M. Lee, A. J. Leong, L. F. Lindoy, B. R.
Rumbel, B. W. Skelton, T. Strixner, G. Wei and A. H. White, J.
Chem. Soc., Dalton Trans., 2002, 3994.
1981, 81, 21; (b) P. Muhl, K. Gloe, F. Dietze, E. Hoyer and L.
¨
Beyer, Z. Chem., 1986, 81, 26.
4 (a) K. H. Konig, M. Schuster, B. Steinbrech, G. Schneeweis and R.
¨
Schlodder, Fresenius’ J. Anal. Chem., 1985, 321, 457; (b) P. Vest,
M. Schuster and K. H. Konig, Fresenius’ J. Anal. Chem., 1991, 341,
566.
¨
24 A. G. Orpen, L. Brammer, F. H. Allen, O. Kennard, D. G. Watson
and R. Taylor, J. Chem. Soc., Dalton Trans., 1989, 51.
25 N. R. Kunchur and M. R. Truter, J. Chem. Soc., 1958, 2551.
26 M. M. Elcombe and J. C. Taylor, Acta Crystallogr., Sect. A: Cryst.
Phys., Diffr., Theor. Gen. Crystallogr., 1968, 24, 410.
27 H. W. Dias and M. R. Truter, Acta Crystallogr., 1964, 17, 937.
28 M. R. Truter, J. Chem. Soc., 1960, 997.
29 W. Walter, S. Harter and J. Voss, Acta Crystallogr., Sect. B:
Struct. Crystallogr. Cryst. Chem., 1976, 32, 2876.
30 R. Richter, F. Dietze, S. Schmidt and E. Hoyer, Z. Anorg. Allg.
Chem., 1997, 623, 138.
5 K. R. Koch, C. Sacht, T. Grimmbacher and S. Bourne, S. Afr. J.
Chem., 1995, 48, 71.
6 K. R. Koch, Coord. Chem. Rev., 2001, 473.
7 A. N. Mautjana, J. D. S. Miller, A. Gie, S. A. Bourne and K. R.
Koch, Dalton Trans., 2003, 1952.
8 S. S. Lee, I. Yoon, K. Park, J. H. Jung, L. F. Lindoy, A. Nezhadali
and G. Rounaghi, J. Chem. Soc., Dalton Trans., 2002, 2180.
9 J. Kim, A. J. Leong, L. F. Lindoy, J. Kim, J. Nachbaur, A.
Nezhadali, G. Rounaghi and G. Wei, J. Chem. Soc., Dalton Trans.,
2000, 3453.
10 M. Fainerman-Melnikova, L. F. Lindoy, S.-Y. Liou, J. C. McMur-
trie, N. P. Green, A. Nezhadali, G. Rounaghi and W. N. Setzer,
Aust. J. Chem., 2004, 57, 161.
31 U. Braun, R. Richter and J. Sieler, Z. Anorg. Allg. Chem., 1985,
529, 201.
32 S. N. Banarjee and A. C. Sukthankar, J. Indian Chem. Soc., 1963,
40, 573.
11 J. M. Lehn, A. Moradpour and J.-P. Behr, J. Am. Chem. Soc.,
1975, 97, 2532.
ꢀc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
1162 | New J. Chem., 2006, 30, 1155–1162