"Crowned" Fe4S4 Clusters as Electrochemical Metal Ion Sensors

Article abstract of DOI:10.1002/(SICI)1099-0682(200002)2000:2<253::AID-EJIC253>3.0.CO;2-M

A series of Fe4S4 cluster compounds I, III, and V, in which the cuboidal cluster core is appended with four crown ether thiolate ligands, and II and IV, bearing thiolate ligands without crown ether parts, has been synthesized and characterized. The spectroscopic and electrochemical properties of these compounds are determined by the electronic nature of the thiolate ligands. Only in the case of III, where a very short α-thioacetyl linker was used to connect the crown ether ligands to the cluster core, was a restricted conformational freedom of the ligand observed. A detailed electrochemical study of the influence of alkali and earth alkali metal ions (Li⁺, Na⁺, K⁺, Mg²⁺, and Ba²⁺) on the reversible 2^-/3^- reduction of the cluster compounds was performed. In the case of the crown ether appended clusters I,III, and V, the addition of these metal ions resulted in an anodic shift, i.e. in positive direction, of the reduction potential (modulation effect) and to larger current responses (promotion effect). The magnitude of the modulation effects is determined by the binding affinity of the metal ions in the crown ether ligands, and by the distance between bound metal ions and the redox active cluster core. Variation of the linker between the cluster core and the metal ion binding site resulted in cluster compounds with almost inverse selectivities for e.g. K⁺ and Ba²⁺ in the case of I and III. For the large effects found for compound I a lariat binding mode is proposed.

Full text of DOI:10.1002/(SICI)1099-0682(200002)2000:2<253::AID-EJIC253>3.0.CO;2-M

FULL PAPER
“Crowned” Fe4S4 Clusters as Electrochemical Metal Ion Sensors
Robertus J. M. Klein Gebbink,[a] Stephen I. Klink,[a] Martinus C. Feiters,*[a]
and Roeland J. M. Nolte*[a]
Keywords: Iron-sulfur clusters / Crown ethers / Electrochemical devices / Host-guest chemistry
A series of Fe4S4 cluster compounds, I, III, and V, in which performed. In the case of the crown ether appended clusters
the cuboidal cluster core is appended with four crown ether
I, III, and V, the addition of these metal ions resulted in an
thiolate ligands, and II and IV, bearing thiolate ligands
anodic shift, i.e. in positive direction, of the reduction
without crown ether parts, has been synthesized and potential (modulation effect) and to larger current responses
characterized. The spectroscopic and electrochemical (promotion effect). The magnitude of the modulation effects
properties of these compounds are determined by the
electronic nature of the thiolate ligands. Only in the case of
III, where a very short α-thioacetyl linker was used to
connect the crown ether ligands to the cluster core, was a
restricted conformational freedom of the ligand observed. A
detailed electrochemical study of the influence of alkali and
is determined by the binding affinity of the metal ions in the
crown ether ligands, and by the distance between bound
metal ions and the redox active cluster core. Variation of the
linker between the cluster core and the metal ion binding
site resulted in cluster compounds with almost inverse
selectivities for e.g. K+ and Ba2+ in the case of I and III. For
earth alkali metal ions (Li+, Na+, K+, Mg2+, and Ba2+) on the the large effects found for compound I a lariat binding mode
reversible 2–/3– reduction of the cluster compounds was
is proposed.
closely to a transition metal ion. Upon binding of a metal
ion the redox potential of the transition metal complex is
Introduction
Iron sulfur clusters play various roles as prosthetic changed. This can be caused either by through-space elec-
groups in proteins. Their main function is to catalyze elec- trostatic interactions or communication via various bond
tron transfer reactions, e.g. to another redox active pros- linkages between the binding site and the redox-active
thetic group within a protein, or from one protein to an- center. In the latter case the connector between the binding
other.[1] In a biological environment, tetranuclear Fe4S4 site and the redox-active site should be conducting in order
clusters generally exist in oxidation states that are desig- to achieve the desired effect, whereas in the former case this
nated as 3Ϫ/2Ϫ/1Ϫ when the charges on the cysteinyl ligands is not required. In either case, the transition metal ion
are considered together with the core charge.[2] Synthetic should be easily polarized to undergo a change in its redox
thiolate clusters of this type show a somewhat richer redox potential. Two examples of such polarizable systems are fer-
chemistry; they can exist in four distinct oxidation states, rocene and [Ru(bipy)3]2ϩ, which have both been frequently
4Ϫ/3Ϫ/2Ϫ and 1Ϫ.[3] Early studies on biological and syn- used in electrochemical devices.[5]
thetic systems have shown that the potentials at which these
states interconvert are strongly influenced by the environ-
ment of the cluster. Factors such as hydrogen bonding to
the cluster, electron-donating or -accepting properties of the
thiolate ligands, hydrophobicity of the ligands, and solvent
polarity are all important in determining the actual values
of the redox potentials.[4] In general, it can be said that the
(partial) electron delocalization in the clusters is the reason
why these systems are easily influenced by polarizing forces.
One of the challenges in the field of electrochemical sen-
sory devices is to develop molecular systems that selectively
recognize metal ions in general, and alkali and alkaline
earth metal ions in particular.[5] Many of these systems are
based on transition metal complexes, because they have a
rich and well-understood electrochemistry. In these systems
a binding site for the above-mentioned metal ions is located
Fe4S4 thiolate clusters can also be considered as suitable
building blocks for this purpose, as (i) they have a rich and
well-defined electrochemistry, (ii) their redox potentials are
highly sensitive to polarizing effects, (iii) they can be rela-
tively easily functionalized, and (iv) four ligands can be
used to influence one redox center. For these reasons, we
decided to study thiolate clusters as the redox-active site for
the recognition of alkali and alkaline earth metal ions. As
mentioned earlier, these clusters bear a 2Ϫ or 3Ϫ overall
negative charge in their most stable forms, whereas most
other systems used for this purpose are neutral or positively
charged. These clusters are therefore expected to lack repul-
sive Coulomb forces that disturb the interaction with metal
ions; instead, one can take advantage of attractive Coulomb
forces to favour it.
In this paper, the synthesis, characterization, and proper-
ties as electrochemical devices of a small series of crown
ether substituted clusters will be presented. A simple aza
crown ether compound, 1-aza-4,7,10,13,16-pentaoxacy-
clooctadecane (monoaza-18-crown-6), was used as the me-
tal ion binding unit because of its ease of functionalization,
[a]
Department of Organic Chemistry, NSR Center,
University of Nijmegen,
Toernooiveld, NL-6525 ED Nijmegen, The Netherlands
Fax: (internat.) ϩ 31-24/365-2676
E-mail: mcf@sci.kun.nl
Eur. J. Inorg. Chem. 2000, 253Ϫ264
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
1434Ϫ1948/00/0202Ϫ253 $ 17.50ϩ.50/0
253

R. J. M. Klein Gebbink, S. I. Klink, M. C. Feiters, R. J. M. Nolte
FULL PAPER
and its well-documented binding properties towards metal
In the synthesis of ligand L2H and L3H, the spacer unit
ions.[6] The binding selectivity of this crown ether is mainly was added to the respective secondary amines, di-n-bu-
determined by its size compatibility with a metal ion guest, tylamine and monoaza-18-crown-6, prior to introduction of
and increases in the order Liϩ ഠ Mg2ϩ < Naϩ < Kϩ
ഠ
the protected thiol functionality, as is shown in Scheme 2.
The respective secondary amines were acylated at the nitro-
Ba2ϩ
.
First of all, three cluster compounds I, III, and V were gen position by treatment with freshly distilled α-chloroace-
synthesized with different crown ether thiol ligands. The ac- tyl chloride, followed by substitution of the halide by a
tual binding unit is not changed in this small series, but the thioacetyl group as discussed above. The desired thiols were
spacer units and the nature of the thiol groups are. In ad- then isolated after hydrolytic deprotection of the thioacetate
dition, two cluster compounds II and IV were synthesized by treatment with 3% HCl in methanol. Overall yields were
and to some extent studied, as these cluster compounds 16 and 52% for L2H and L3H, respectively. Because these
lack the crown ether moieties and can therefore be con- compounds lack an internal base, they are less sensitive to
sidered as model compounds.
oxidative dimerization than L1H.
Scheme 2. Synthesis of ligands L2H and L3H: (i) di-n-butylamine,
K2CO3, DMF; (ii) monoaza-18-crown-6, K2CO3, toluene; (iii)
K2CO3, CH3C(O)SH, DMF; (iv) 3% HCl/MeOH
Results and Discussion
Ligand Synthesis
A different strategy was used for the synthesis of L4H
and L5H (Scheme 3). In the first step, 2-thiosalicylic acid
was protected as its disulfide, which was then converted into
the corresponding diacid chloride by treatment with thionyl
chloride. Acylation of secondary amines with this diacid
chloride proceeded smoothly in the presence of triethyl-
amine as a base. The final step, cleavage of the disulfide
bond to yield two identical thiols, was performed in a water/
dioxane mixture with triphenylphosphane as the nucleoph-
ile, under acidic conditions.[9] Pure samples of L4H and
L5H could be obtained by column chromatography. The
overall yields were 48 and 52% for L4H and L5H, respec-
tively.
The synthesis of ligand L1H is depicted in Scheme 1.
Starting from α,αЈ-dibromo-m-xylene, one of the bromide
substituents was replaced by a thioacetyl group using a pro-
cedure described by Kellogg and co-workers[7] in which
freshly distilled thioacetic acid was treated with K2CO3 to
obtain the thioacetate nucleophile. After purification, the
desired monobromide 1 was coupled to monoaza-18-crown-
6 in DMF in the presence of base. The final step involved
the hydrolysis of the thioacetyl function to the desired thiol.
This reaction proceeded smoothly under acidic conditions
in methanol (3% HCl) and under an inert gas.[8] Due to the
presence of an internal base, i.e. the tertiary amine group,
L1H is prone to dimerize under aerobic conditions. This
compound was, therefore, kept strictly under N2 during
workup and storage. The overall yield of L1H was 41%.
Synthesis and Characterization of Substituted Fe4S4
Clusters
Cluster compounds IϪV were synthesized by ligand ex-
change reactions (Scheme 4),[10] and were isolated as di-
anionic species with either tetraalkylammonium or tetra-
phenylphosphonium counter ions. In one approach, these
exchange reactions are based on an acid-base equilibrium,
in which a more acidic thiol replaces a more basic thiolate
ligand from an Fe4S4(SR)4 cluster.[11] By making use of vol-
atile thiols as the leaving group, e.g. tertiary butyl thiol,
the equilibrium is easily driven to one desired product by
Scheme 1. Synthesis of ligand L1H: (i) K2CO3, CH3C(O)SH, DMF;
(ii) monoaza-18-crown-6, K2CO3, DMF; (iii) 3% HCl/MeOH
254
Eur. J. Inorg. Chem. 2000, 253Ϫ264

“Crowned” Fe4S4 Clusters as Electrochemical Metal Ion Sen
FULL PAPER
properties of related cluster compounds[12][13] and con-
firmed the integrity of the cubane core in these newly syn-
thesized compounds (see Experimental Section). Further
characterization was performed by NMR spectroscopy and
electrochemical techniques (vide infra). NMR showed that
all isolated cluster compounds contained residual solvent.
In spite of the above workup procedures, and after several
crystallization attempts and washings of the isolated prod-
ucts with various nonsolvents, it was not possible to remove
these solvent traces. Elemental analyses of the isolated clus-
ter compounds were therefore not performed.
NMR spectra of Fe4S4 cluster compounds in general are
characterized by large chemical shifts of nuclei close to the
cluster core and by broadened signals, due to the thermal
accessibility of paramagnetic excited states at ambient tem-
perature.[14] These two features can be observed in Figure
1, which shows the 1H-NMR spectrum of cluster III in
[D7]DMF solution. In Table 1, the isotropic shifts are listed
for the protons closest to the cluster core in clusters IϪV.
These shifts are defined as the difference in chemical shift
for a given nucleus in the free ligand and in the cluster com-
pound. For clusters IϪIII the shifts all have a negative sign,
whereas for clusters IV and V some have a positive sign.
The latter observation is a characteristic feature of aryl thi-
olate-substituted clusters and is correlated to the sign of
spin density at the respective carbon atom in the aromatic
ring. For aliphatic ligands, the sign has been shown to be
non-alternative; its magnitude depends strongly on the dis-
tance between the metal center and the hydrogen nu-
cleus.[14]
Scheme 3. Synthesis of ligands L4H and L5H: (i) H2O2, KOH, wa-
ter; (ii) SOCl2; (iii) di-n-butylamine, TEA, CH2Cl2; (iv) monoaza-
18-crown-6, TEA, CH2Cl2; (v) PPh3, dioxane/water/HCl
applying a “dynamic” vacuum. The other approach makes
use of an Fe4S4 cluster that bears good leaving groups such
as ClϪ ligands. The thiol used in these reactions is first de-
protonated in order to make it a better nucleophile and to
prevent the formation of HCl. Both methods were applied
in the syntheses of clusters IϪV since each of them has
its advantages. The first method proceeds under very mild
conditions without the need of any additive, whereas the
second method is much faster but suffers from the fact that
a base has to be added which may lead to counter ion
scrambling.
Scheme 4. Preparation of Fe4S4 cluster compounds by ligand ex-
change reactions
Prior to the synthesis of the cluster compounds, test reac-
tions were performed in situ in an electrochemical cell, and
followed by differential pulse voltammetry (DPV). With
this procedure, the quality of the freshly prepared thiols was
tested, and an indication of the electrochemical character-
Figure 1. 1H-NMR spectrum of cluster compound III in [D7]DMF
(the asterisks denote unassigned peaks and residual solvent peaks)
These general features were also found in temperature-
istics of the final cluster compounds was obtained. On a dependent NMR studies on clusters III and IV, in the tem-
preparative scale, the desired cluster compounds IϪV were perature range 298Ϫ350 K. The experiments showed that
isolated as black oils after filtration of the crude reaction the isotropic shifts for these cluster compounds changed in
mixtures, followed by repeated precipitations from solution, magnitude upon increase of the temperature, due to a
and in vacuo removal of residual solvent.
higher occupation of the first excited state, whereas their
Initial spectroscopic characterization of the isolated clus- sign did not change. The temperature dependencies of the
ter compounds was performed by means of UV/Vis spec- isotropic shifts are listed in Table 1, and were analyzed as
troscopy. Their UV/Vis characteristics closely matched the linear relationships. Similar linear dependencies have been
Eur. J. Inorg. Chem. 2000, 253Ϫ264
255

R. J. M. Klein Gebbink, S. I. Klink, M. C. Feiters, R. J. M. Nolte
FULL PAPER
Table 1. Isotropic shifts of ligand protons in cluster compounds fact that the thiolate ligands in this compound have rela-
IϪV, together with their temperature dependence (temperature
tively strong electron-accepting properties. The halfwave
range: 300Ϫ340 K; solvent: [D7]DMF or [D6]DMSO, see Experi-
potentials of these cluster compounds are therefore solely
determined by the electronic properties of the thiolate li-
gands. In contrast, the E1/2 value of cluster III was found
to be 70 mV more positive than that of cluster II. This is
probably caused by the crown ethers, which are in close
proximity of the cluster core due to the short acetyl spacers.
The exact nature of the influence of the crown ethers on
the reduction potential is not yet understood. NMR experi-
ments showed that the ligands in III have restricted confor-
mations (vide supra).
For all cluster compounds, the peak separations (∆Ep) in
the cyclic voltammograms and the peak widths at half peak
height (W1/2) in the differential pulse voltammograms were
larger than the theoretical values of 59 mV and 90.4 mV,
respectively. These deviations indicate that the electron-
transfer rate between the electrode and the cluster com-
pound is slow on the experimental time scale. A reason for
the electrochemically quasi-reversible redox processes, in
the case of the crown ether-functionalized compounds,
could be the shielding of the cluster core from the electrode.
Direct contact between the electrode and the redox-active
species is desirable, as diminished contact delays the veloc-
ity of electron-transfer between the redox active moiety and
the electrode, resulting in large peak separations. In ad-
dition, the ratio of the peak currents for the backward and
the forward processes (ib/if) in CV did not approach the
value of 1 in most cases.
To obtain insight into the factors that govern the electro-
chemical properties of the above-mentioned cluster com-
pounds, a more elaborate study was undertaken for com-
pounds IV and V. First of all, scan rate dependent CV
experiments were performed. The E1/2 values of both clus-
ters were found to be independent of the scan rate, whereas
∆Ep increased considerably with increasing scan rate. A lin-
ear relationship was found between the cathodic peak cur-
rent (ip,c) and the square root of the scan rate (υ1/2), suggest-
ing that the mass transport of redox-active species toward
the Pt electrode was diffusion-controlled. In addition, it was
found that at higher scan rates the 2Ϫ/3Ϫ reduction was
chemically more reversible. The 2Ϫ/3Ϫ reductions can
therefore be described as chemically quasi-reversible, which
implies that the reduction of the cluster core is followed
by a chemical reaction that is fast enough to influence the
voltammetric analysis (EC mechanism). This can either be
due to the coordinating properties of the solvent, which
would lead to diminished stability of the generated 3Ϫ clus-
ter, or to ligand reactivity.
mental Section)
Com-
pound
Ligand
δ [ppm]
Cluster
δ [ppm]
∆Hiso
[ppm]
∆Hiso/T
[Hz/K]
I
CH2S
CH2S
CH2S
3.62
3.36
3.76
15.08
13.83
15.35
16.58
3.73
Ϫ11.46
Ϫ10.47
Ϫ11.59
Ϫ12.82
3.52
n.d.[a]
n.d.[a]
Ϫ3.32
Ϫ3.32
0.43
II
III
IV
pArH
oArH
mArH
mArH
pArH
oArH
mArH
mArH
7.25
Љ
Љ
Љ
5.32
1.93
0.31
7.97
Ϫ0.72
Ϫ1.03
Ϫ
1.96
Ϫ0.71
Ϫ1.14
Ϫ0.20
Ϫ0.27
Ϫ
8.28
V
7.25
n.o.[b]
5.29
Љ
Љ
Љ
n.d.[a]
n.d.[a]
n.d.[a]
7.96
8.39
[a]
[b]
n.d.: not determined. Ϫ n.o.: not observed.
observed for other cluster compounds in the same tempera-
ture range.[14][15]
An interesting feature from Figure 1 is the fact that the
CH2ϪS hydrogen signals in cluster III appeared as two sin-
glets with approximately equal intensity. Apparently, the li-
gands in III have a restricted rotational freedom, which re-
sults in nonequivalent methylene hydrogen atoms due to the
angular dependence of their isotropic shifts. Similar obser-
vations have been made for the cysteinyl methylene hydro-
genatoms in iron-sulfur proteins and in oligopeptide model
compounds.[16][17] For cluster II, only one signal was ob-
served for the CH2ϪS protons, suggesting that in III the
crown ethers play a determining role in restricting the con-
formation of the ligands.
The electrochemical properties of IϪV were studied by
means of cyclic voltammetry (CV) and differential pulse
voltammetry (DPV) in DMF. This solvent was chosen be-
cause both R4N and Ph4P salts of iron-sulfur clusters are
soluble in it. Numerical data of the study are presented in
Table 2.
Table 2. Electrochemical data for the 2Ϫ/3Ϫ reduction of cluster
compounds IϪV in DMF[a]
CV
ib/if
DPV
E1/2 [V][b]
∆Ep
Ep [V][b]
W1/2
I
Ϫ1.74
Ϫ1.65
Ϫ1.58
Ϫ1.52
Ϫ1.50
0.68
0.86
0.80
0.56
0.49
85
Ϫ1.74
Ϫ1.66
Ϫ1.58
Ϫ1.50
Ϫ1.48
160
100
150
140
284
II
III
IV
V
221
213
170
270
[a]
For conditions see Experimental Section. Pt working electrode.
[b] Potentials are given relative to the Fc/Fcϩ redox couple in the
For cluster compound IV, the solvent and the type of
electrode in the electrochemical experiments were varied. In
solvents with relatively high ε, the E1/2 values of IV were
Ϫ
same solvent.
The halfwave potentials (E1/2) for the 2Ϫ/3Ϫ reduction similar to those of [Fe4S4(SPh)4]2Ϫ, i.e. Ϫ1.52 V and Ϫ1.40
of clusters I, IV, and V compare well with values for simple, V in DMF and acetonitrile, respectively. In dichlorometh-
related thiolate clusters. For [Fe4S4(SCH2Ph)4]2Ϫ and ane, the E1/2 for IV was found to be 100 mV more negative
[Fe4S4(SPh)4]2Ϫ the E1/2 values have been reported to be than for [Fe4S4(SPh)4]2Ϫ. This might be the result of ion
Ϫ1.74 and Ϫ1.52 V, respectively.[18] The corresponding pair formation or (diminished) interactions of the cluster
E1/2 for cluster II has an intermediate value; this reflects the with the solvent. The electrochemical reversibility was im-
256
Eur. J. Inorg. Chem. 2000, 253Ϫ264

“Crowned” Fe4S4 Clusters as Electrochemical Metal Ion Sen
FULL PAPER
proved in acetonitrile. The best results were obtained, how- with I, but only 20 mV with IV. The crown ether parts in
ever, when a pyrolytic graphite-edge (PGE) electrode was the ligands thus appear to be essential to give large changes
used. With this electrode, also in DMF, a rather fast elec- in the electrochemical properties of the Fe4S4 core upon
tron transfer was observed, together with a somewhat addition of metal ions. A more detailed investigation of the
higher chemical reversibility. Compared to Pt electrodes effect of the crown ether ligands on the electrochemical be-
PGE electrodes are chemically less inert, i.e. due to surface haviour toward alkali and alkaline earth metal ions was
oxidation, interactions between electrode and redox-active performed with compounds I, IV, and V, and is described
species are more plausible.
below.
In conclusion, it appears that the relatively large organic
ligands influence the electrochemical behaviour of cluster
compounds IϪV to some extent. Due to shielding of the
negatively charged cluster core by the large organic ligands,
which apparently slows down the electron transfer from
electrode to cluster core, the observed electrochemical pro-
cesses were not always ideal. By varying the circumstances
under which the compounds were studied, however, it was
shown that these clusters inherently undergo reversible re-
dox transitions.
Table 3. Changes in Ep for the reduction of compounds I, III, IV,
and V upon addition of 10 equivalents of metal ion[a]
I
III
IV
V
Liϩ
90 (120)
(0)
0 (0)
0 (10)
Naϩ
Kϩ
30 (n.d.)[b]
40[c] (n.d.)[b]
190[d] (290)[e]
190[d] (210)
Ϫ
0 (10)
0 (10)
30 (90)
20 (50)
50 (60)
50 (60)
70 (110)
30 (50)
Ϫ
(130)
(120)
Mg2ϩ
Ba2ϩ
[a]
In parentheses the Ep change when a large excess (100 equiva-
[b]
[c]
lents) of metal ion is added. Ϫ
n.d. ϭ not determined. Ϫ
[d]
Reached after 2 equivalents had been added. Ϫ Reached after 4
Electrochemical Recognition of Alkali and Alkaline
Earth Metal Ions
[e]
equivalents had been added. Ϫ Small additional peak observed.
The effects of alkali and alkaline earth metal ions on the
E1/2 for the 2Ϫ/3Ϫ transition of cluster compounds I, III,
IV, and V, were investigated by performing DPV experi-
ments in the presence of varying amounts of perchlorate
Promotion and Modulation Effects
First of all, it is worthwhile to explain the difference be-
salts of lithium, sodium, potassium, magnesium, and bar- tween promotion and modulation effects on the electro-
ium. The use of the DPV technique has the advantage that chemical properties of redox active species. In general, pro-
very small amounts of material can be used. This technique motion effects involve an improvement of both current re-
also has a higher resolution compared to CV. Figure 2 sponse and electrochemical reversibility of a redox process,
shows the changes in the differential pulse voltammograms and in addition often involve an increase of peak currents
of cluster compound I and of V upon the addition of Ba2ϩ after repetitive CV scans. Modulation effects involve
ions. The peak potentials for the reduction of both com- changes in the potential of a redox process. The electro-
pounds shift to more positive values, i.e. the clusters are chemistry of large biomolecules often suffers from very
more easily reduced to the 3Ϫ form. Compound I exhibited poor current responses due to slow electron transfer from
so-called multi-wave behaviour, whereas V showed a con- the redox active entity, embedded in the protein core, to the
tinuous shift of its Ep (vide infra). Similar titrations were electrode.[19] Addition of a promoter increases the electron-
performed with the cluster compounds III and IV.
transfer rate either by improving the protein-electrode inter-
In all cases, a change in Ep occurs in an anodic, i.e. in a action or by mediating electron transfer. Generally, pro-
positive direction (Table 3). The magnitude of this change moters are polar or charged compounds that bear coordin-
depends on the cluster compounds as well as on the metal ating groups. Sometimes the use of a promoter also results
ions. For example, Ba2ϩ ions give a shift in Ep of 190 mV in an altered E1/2. In that case this compound behaves as a
Figure 2. Changes in Ep upon addition of Ba2ϩ ions for the reduction of (a) compound I (prepared in situ) and (b) compound V, followed
by DPV; numbers correspond to the amount of metal ion added (mol. equiv.)
Eur. J. Inorg. Chem. 2000, 253Ϫ264
257

R. J. M. Klein Gebbink, S. I. Klink, M. C. Feiters, R. J. M. Nolte
FULL PAPER
Figure 3. (a) DPV titrations of compound IV with Ba2ϩ ions, and compound V with Naϩ and Mg2ϩ ions, numbers correspond to the
amount of metal ion added; (b) corresponding titration curves (including those for compound IV and Naϩ) in which Ep, ip, and W1/2
are plotted as a function of the amount of metal ion added
modulator as well. Promoter/modulator effects similar to current response and in smaller W1/2 values (vide supra).
those observed for proteins have also been observed for The titration curves in Figure 3b indicate that changes in
semi-encapsulated, synthetic cluster compounds.[20][21]
these parameters are directly related to the amount of metal
Overlays of subsequent differential pulse voltammograms ion that is added to V. These promotion effects were already
in the titration of cluster compounds IV with Ba2ϩ ions, observed upon addition of less than four equivalents of me-
and V with Naϩ and Mg2ϩ ions are shown in Figure 3a. tal ions per cluster. Binding of more than two monovalent
Figure 3b shows the titration curves in which ip, and W1/2 metal ions by V converts this compound from an overall
(and also Ep, see below) are plotted as a function of the negatively to an overall positively charged species, resulting
amount of metal ion added per cluster. The curves derived in a more favourable interaction with the negatively charged
from the titration of IV with Naϩ ions are also shown. For electrode. This observation is in sharp contrast to the large
compound IV, a small decrease in current response (ip) was excess of Ba(ClO4)2 which is required to promote the elec-
observed upon the addition of monovalent alkali metal trochemistry of metalloproteins and the earlier mentioned
ions. The addition of divalent metal ions clearly resulted in semi-encapsulated cluster compounds.[19Ϫ21] This strongly
an increase in ip as well as in the electrochemical reversi- suggests that specific binding of metal ions in the crown
bility of the redox process (decreased W1/2 value). The ip ether rings causes the observed effects. In addition, com-
and W1/2 values of V improved considerably upon the ad- pound IV, having no specific metal ion binding sites, was
dition of monovalent as well as divalent metal ions, except only affected at metal ion concentrations approaching the
in the case of Liϩ for which only minor improvements were millimolar range.
observed (not shown).
The promoter effects observed for cluster compound I
The fact that divalent metal ions have an influence on the were in line with those obtained for V, albeit less pro-
electrochemical properties of IV indicates that the overall nounced (not shown). Apparently, the larger flexibility of
negative charge of IV to some extent hampers a clean redox the crown ether ligands in I as compared to V results in a
process. Divalent metal ions are assumed to form a layer of better electrochemical behaviour even without the addition
positive charges on a negatively charged electrode, thereby of metal ions.
facilitating the reduction of negatively charged species,
Changes in the peak potentials of compounds IV and V
while monovalent metal ions do not give this effect.[22] This vary from 0 for Liϩ to 70 mV for Mg2ϩ after the addition
phenomenon can be explained by the larger polarizing of 10 equivalents of these ions per cluster compound (Table
strength of the divalent alkaline earth metal ions as com- 3). In general, the changes of the Ep of IV are very small;
pared to the monovalent alkali metal ions.
a maximum modulation of 20Ϫ30 mV for divalent metal
The addition of metal ions to cluster compound V, with ions is observed. This small change may be the result of a
the exception of Liϩ, resulted in an improvement of the weak coordination of the metal ions to the amide function
258
Eur. J. Inorg. Chem. 2000, 253Ϫ264

“Crowned” Fe4S4 Clusters as Electrochemical Metal Ion Sen
FULL PAPER
in the ligand, or may be caused by changes in the electrode Naϩ < Kϩ ഠ Ba2ϩ, which is based on size complementar-
surface as pointed out above. At very high metal ion con- ity. From the point of view of charge stabilization, the di-
centration, larger effects were observed with Mg2ϩ and valent metal ions are expected to bind more strongly than
Ba2ϩ, suggesting that either of the two factors became more the monovalent metal ions.[6] The binding constants for the
important in the modulation process.
ligands of V were anticipated to be lower than those of the
For compound V, the modulation effects are more pro- ligands of I. The acyl functions on the nitrogen centers in
nounced than for IV. The titration curves in Figure 3b show the aza crown ether rings of V lead to a host molecule
that the largest effect is achieved after the addition of sev- which is a less efficient metal ion binder. This modification
eral equivalents of Naϩ. For Ba2ϩ, the major effect is could explain the generally smaller changes in Ep found for
achieved after the addition of approximately 20 equivalents V. However, these rules of thumb do not explain the trends
(not shown). After the addition of an excess of modulator, in responses of compounds I and V.
minor additional changes in Ep were observed. These ad-
ditional changes closely match the effects found for com- Ep of compound I on the charge density of the metal ion
pound IV at low modulator concentrations. that causes the modulation. Metal ions bearing a high
Figure 5 shows the dependence of the modulation of the
In Figure 2a, an overlay of sequential voltammograms charge density give rise to a larger change in the reduction
for the titration of compound I with Ba2ϩ is depicted (vide potential than metal ions bearing a low charge density. In
supra). Figure 4 presents the changes in Ep for this com- other words, the higher the polarizing strength of the metal
pound upon the addition of Ba2ϩ, Naϩ, and Kϩ ions. Due ion, the larger the modulation of the redox properties of
to the multi-wave behaviour observed in the Ba2ϩ titration the polarizable Fe4S4 core in I. It is likely therefore that this
and problems with the deconvolution, the Ba2ϩ curve in modulation is governed by electrostatic forces. In this way,
Figure 4 deviates from an ideal titration curve. Nevertheless metal ions with similar binding constants or ionic radii such
it can be concluded that the effect of Ba2ϩ ions is much as Kϩ and Ba2ϩ, but different polarizing effects, bring
larger than the effects of Kϩ and Naϩ ions. The addition about different modulations. Similar observations have been
of very small amounts of both Ba2ϩ and Kϩ results in a made by Hall et al. for their ferrocene based cryptates.[23]
steep rise in Ep, whereas for Naϩ this is not observed. In
accordance with the observations for compound V, it is
likely that the binding of ions in the crown ether ligands of
the cluster contributes most to the observed changes in Ep.
The fact that multi-wave behaviour is observed with Ba2ϩ
,
as well as with Liϩ and Mg2ϩ ions, points to a strong bind-
ing of these metal ions by compound I.
Figure 5. Dependence of the metal ion induced change in Ep,red of
compound I on the charge density of the metal ion
The multi-wave behaviour that is observed for I in combi-
nation with Ba2ϩ, Liϩ, and Mg2ϩ ions, and the magnitude
of the modulations, suggests that the binding of these metal
ions by I is strong. Since the connection between the bind-
ing site and redox site in I is not conducting, the metal ions
should be bound in very close proximity to the redox active
center in order to be able to affect the redox potential to the
observed extent. It is proposed therefore, that compound I
binds metal ions in a lariat-ether fashion,[24] with the oxy-
Figure 4. Changes in Ep upon titration of compound I with Naϩ,
Kϩ, and Ba2ϩ ions
Rationalization of the Modulator Effects
The different changes in Ep upon the addition of stoi- gen and nitrogen atoms of the crown ether ring and the
chiometric amounts of metal ions to cluster compounds I negatively charged thiolate sulfur atom, which coordinates
and V probably arise from the fact that the affinities of the to the cluster, as donor atoms (see Figure 6). In this way
crown ether ligands of the host compounds for the metal the interaction between the binding site and the redox site
ions are different. As was mentioned in the introduction, would be through-bond instead of through-space and the
the binding of metal ions in the monoaza-18-crown-6 li- overall negative charge of the cluster compounds would
gands of I is expected to follow the trend Liϩ ഠ Mg2ϩ
<
contribute to its interaction with metal ions. Similar lariat
Eur. J. Inorg. Chem. 2000, 253Ϫ264
259

R. J. M. Klein Gebbink, S. I. Klink, M. C. Feiters, R. J. M. Nolte
FULL PAPER
complex
free
Ep
Ϫ Ep ϭ Ϫ RT/nF [ln Kred/Kox]
(1)
interactions have been proposed before by Gokel et al. in
nitrobenzene pivoted crown ethers.[25]
In words, the ratio of the binding constants of a guest in
the reduced and oxidized host is given by the change in Ep
brought about by the binding of the guest. In this model,
the modulations that are observed for the cluster com-
pounds reflect the enhanced binding of metal ions by these
compounds upon reduction of the cluster core. Comparison
of the modulations brought about by the different metal
ions will then involve more than just the comparison of
binding constants of the oxidized host. Equation 1 has been
shown by Kaifer to be valid for systems with Kox > 104. For
systems with 1 < Kox < 104, which is expected to be the
range for our compounds, the increase in binding constant
is underestimated by this Equation.[26] The results of appli-
cation of the above-mentioned analysis to the effects found
for I and V are presented in Table 4.
Figure 6. Proposed mode of metal ion binding by compound I
Due to conformational restrictions, the host-guest inter-
action in V is not expected to be similar to that of I. As
in cluster compound I, there is no conducting connection
between the redox center and the binding site in this com-
pound. An effective impact on the redox potential of the
cluster core should, therefore, arise via through-space inter-
actions. The more rigid ligand connection in V compared
to I is likely to result in a less favourable approach of the
metal ion to the redox active center, which could explain
that, in general, smaller modulation effects and no multi-
wave behaviour were observed for compound V. Combining
the values from Table 3 with the curves in Figure 3b, and
taking into account the metal ion charge densities, one can
conclude that the binding constants in V roughly follow the
trend Liϩ < Ba2ϩ < Mg2ϩ < Naϩ < Kϩ. In particular,
acylation of the aza crown ether rings seems to influence
the binding of metal ions with a high charge density in a
negative sense. The modulation effects observed for V can
then be explained by combining the known trend in binding
constants with the polarizing strength of the metal ions.
Another way of dealing with the observed modulation
effects is presented in Scheme 5, which is taken from earlier
papers by Kaifer et al.[26] Starting from a redox-active host
compound (an Fe4S4 cluster in this case) and an uncom-
plexed guest, there are two ways to come to a reduced host-
guest complex. In one way, the guest is bound by the host
compound in the first step, resulting in a facilitated re-
duction of the redox active center. In the second way, the
redox-active center is first reduced, which in turn results in
an enhanced binding of the guest.
Table 4. Ratio of Kred and Kox for binding of different metal ions
in compounds I and V[a]
I
V
Liϩ
33 (13)
3 (1)
Ϫ
Naϩ
Kϩ
7 (3)
7 (3)
5 (2)
2 (1)
5 (2)
Mg2ϩ
Ba2ϩ
1600 (650)
1600 (650)
[a]
Determined after the addition of 4 equivalents of metal ion.
Estimated errors are given in parenthesis.
The titration curves presented in Figure 4 indicate that
more than 2 and possibly 4 metal ions are bound to the
crown ether cluster compounds. From our experiments it
could not be determined whether the four binding sites be-
haved independently from each other, in terms of equal
binding constants for every site, or that these sites influ-
enced each other, e.g. by diminishing the binding of every
next metal ion. It should be noted that applying Equation
1 might lead to overestimated values in the latter case. The
numbers in Table 4 should therefore not be treated as abso-
lute values but merely as representing trends in metal ion
affinities of the cluster compounds.
Conclusion
A number of new thiolate-functionalized iron-sulfur cub-
ane clusters IϪV were prepared and characterized. In I, III,
and V, the cluster cores are appended with binding sites for
metal ions, viz. crown ether ligands. Electrochemical studies
show that the large crown ether ligands hamper, to some
extent, the electrochemical analysis of the cluster com-
pounds. Only in the case of cluster compound III does the
presence of the relatively large ligands have a pronounced
effect on the E1/2 value. Here, the ligands are believed to
have a restricted conformational freedom, resulting in a
more effective shielding of the cluster core.
Scheme 5. Pathways for metal ion complexation in an Fe4S4 cluster
compound; Ec0 and Ef0 are the midpoint potentials for the complex
and the free ligand, respectively
For the respective reduction potentials and binding con-
stants Equation 1 can be derived.[26]
Examination of the promotion and modulation effects of
metal ions on cluster compounds I, IV, and V reveals that
260
Eur. J. Inorg. Chem. 2000, 253Ϫ264

“Crowned” Fe4S4 Clusters as Electrochemical Metal Ion Sen
FULL PAPER
spectra were recorded with a BioRad FTS 25 instrument. Ϫ UV/
Vis spectra were recorded with a PerkinϪElmer Lambda 5 instru-
ment. Ϫ 1H-NMR spectra were recorded with Bruker WF-90,
Bruker AC-100, Bruker WM-200, Bruker AC-300, or Bruker AM-
the largest effects occur in the case of the crown ether func-
tionalized cluster compounds I and V. Addition of stoichio-
metric amounts of metal ions leads to large changes in the
electrochemical characteristics of the cluster reduction, im-
plying that metal ions are bound in the crown ether ligands.
Smaller changes arise from electrode modification, which
facilitates the reduction of the negatively charged cluster
compounds, as was shown for compound IV. Reduction of
the cluster compounds enhances their affinity for alkali and
400 instruments (internal standards: CDCl3 as solvent TMS, δH
ϭ
0.00; [D6]DMSO as solvent, δH ϭ 2.50; [D7]DMF as solvent, δH ϭ
2.94). Ϫ Mass spectra were obtained from a VG 7070E instrument
[FABϩ (3-nitrobenzyl alcohol), EIϩ]. Ϫ Elemental analyses were
performed with a Carlo Erba EA 1180 instrument. Ϫ Melting
points were measured with a ReichertϪJung Hotstage 600 instru-
alkaline earth metal ions. A lariat-binding mode is pro- ment (uncorrected values). Ϫ Electrochemical analysis (CV, DPV)
were performed with an EG & G Princeton Applied Research Mo-
del 273 Electrochemistry System equipped with Model 270 Electro-
chemical Analysis Software 200. Reference electrodes: Ag/AgCl
(0.1  LiCl) in DMF, Ag/AgI (0.02  Bu4NI/0.1  Bu4NPF6) in
CH2Cl2, and Ag/Agϩ (0.1  AgNO3) in CH3CN. Supporting elec-
trolyte: 0.1  tetrabutylammonium hexafluorophosphate (TBAH).
All measurements were performed under an inert gas, using a con-
ventional three-electrode electrochemical cell equipped with either
platinum or PGE working and Pt auxiliary electrodes. PGE elec-
trodes were polished with 0.3 µm alumina, rinsed with water, and
posed for the complexation of metal ions to compound I,
resulting in a more or less through-bond interaction be-
tween the redox site and the crown ether bound metal ion.
In this way, binding of metal ions by this compound is fa-
voured by the negative charge of its redox center. Such a
favourable interaction was ruled out, however, for com-
pound V due to restrictions in the conformation of its li-
gands; probably related to the lack of rotational freedom of
the amide bonds. The modulations observed for compound
V seem to take place via through-space interactions. Over- sonicated in water for 10 min before use. Before every session, the
experimental setup was tested using ferrocene and
[Fe4S4(StBu)4](NBu4)2. Halfwave and peak potentials are given
against the reversible Fc/Fcϩ couple. Scan rates were 100 mV/s
(CV) and 10 mV/s (DPV) unless stated otherwise. Concentrations
of electroactive species were between 0.01 m and 1 m.
all, the characteristic features of the modulations measured
for the novel cluster compounds were shown to be governed
by the interplay of crown ether selectivity, ligand flexibility,
and electrostatic forces. As a result, the modulations ob-
served for compounds I and V are in sharp contrast to each
other. Whereas I shows a strong response towards Ba2ϩ α-Bromo-αЈ-thioacetyl-m-xylene (1): To
a solution of K2CO3
ions and hardly any response towards Kϩ, compound V
displays much smaller effects and an opposite selectivity.
The results presented here show that crown ether-func-
tionalized iron-sulfur clusters can serve as electrochemically
active host compounds, and can in principle be applied as
building blocks for the construction of electrochemical sen-
sory devices which are able to display selective responses
towards charged species
(1.05 g, 7.6 mmol) in 4 mL of DMF was slowly added a solution
of freshly distilled thioacetic acid (1.14 g, 15 mmol) in 2 mL of
DMF. After CO2 formation had ceased, the resulting yellow solu-
tion was diluted with the same amount of DMF. This diluted solu-
tion was added dropwise to a solution of α,αЈ-dibromo-m-xylene
(4.0 g, 15.2 mmol) in 20 mL of DMF, and the resulting mixture was
stirred overnight under the exclusion of light. These manipulations
were all performed under N2. After filtration and removal of the
solvent in vacuo, the resulting residue was dissolved in 100 mL of
diethyl ether, washed with water (3ϫ), dried with MgSO4, filtered,
and concentrated in vacuo. Purification by flash chromatography
(silica, ethyl acetate/hexane ϭ 1:10, v/v) yielded 2.1 g (54%) of 1 as
a yellow oil. Ϫ 1H NMR (90 MHz, CDCl3): δ ϭ 2.36 [s, 3 H,
C(O)CH2], 4.11 (s, 2 H, ArCH2S), 4.45 (s, 2 H, ArCH2Br), 7.29
(m, 4 H, ArH). Ϫ EI-MS; m/z: 260 [M]ϩ. Ϫ C10H11BrOS (259.2):
calcd. C 46.35, H 4.28, S 12.37; found C 46.50, H 4.39, S 11.85.
Experimental Section
General Procedures: All chemicals were commercial products and
were used as received. α-Chloroacetyl chloride was synthesized
using a literature procedure.[27] Di-n-butylamine and thioacetic acid
were distilled under reduced pressure. Diethyl ether was distilled
from sodium metal/benzophenone, hexane was distilled from so-
dium metal, and dichloromethane and chloroform were distilled
from CaCl2. Pyridine and triethylamine were distilled from CaH2.
DMF used for electrochemical experiments was stored over BaO
for at least one day, after which it was distilled under reduced press-
ure. Ethyl acetate used for flash chromatography was distilled from
K2CO3/Na2CO3. Hexane used for the same purpose was distilled
once using a rotary evaporator. All other solvents were used as
received. Manipulations involving thiol compounds were per-
formed under N2 unless stated otherwise. Glassware used in thiol
handling was left overnight in a hypochlorite solution to prevent
stench. Manipulations involving iron-sulfur clusters were per-
formed using standard Schlenk techniques. Ϫ Flash chromatogra-
phy was performed on Silica Gel (60 H) purchased from Merck, or
on neutral alumina purchased from Across. Prior to use, neutral
alumina was activated to activity III, according to the Brockman
scale.[28] Ϫ TLC analyses were performed on silica 60 F254 coated
glass plates from Merck. Thiol compounds were detected by treat-
1-(α-Thioacetyl-m-xylyl)-4,7,10,13,16-pentaoxa-1-azacycloocta-
decane (2): A solution of 1 (390 mg, 1.51 mmol) in 5 mL of DMF
was added dropwise to a solution of 4,7,10,13,16-pentaoxa-1-aza-
cyclooctadecane (400 mg, 1.51 mmol) in 15 mL of DMF which
contained some K2CO3 as a base. The reaction mixture was stirred
overnight at 40°C under N2, after which the solvent was removed
under reduced pressure. The resulting oil was dissolved in 40 mL of
dichloromethane, washed twice with a saturated aqueous NaHCO3
solution, dried with MgSO4, filtered, and concentrated in vacuo.
Purification by flash chromatography (silica, CHCl3/MeOH/
TEA ϭ 97:2:1, v/v) yielded 530 mg (80%) of 2 as a viscous yellow
oil. Ϫ 1H NMR (90 MHz, CDCl3): δ ϭ 2.36 [s, 3 H, C(O)CH3],
2.78 (t, J ϭ 6.6 Hz, 4 H, NCH2CH2O), 3.62 (m, 22 H, OCH2,
ArCH2N), 4.11 (s, 2 H, ArCH2S), 7.24 (m, 4 H, ArH). Ϫ FAB-
MS; m/z: 442 [M ϩ 1]ϩ. Ϫ C22H35NO6S (441.6): C 59.84, H 7.99,
N 3.17, S 7.26; found C 59.63, H 8.05, N 3.22, S 7.36.
1-(α-Mercapto-m-xylyl)-4,7,10,13,16-pentaoxa-1-azacyclooctade-
ment of the TLC plates with a nitroprusside reagent.[29] Ϫ FT-IR cane (L1H): A solution of 2 (200 mg, 0.45 mmol) in 5 mL of 3%
Eur. J. Inorg. Chem. 2000, 253Ϫ264
261

R. J. M. Klein Gebbink, S. I. Klink, M. C. Feiters, R. J. M. Nolte
FULL PAPER
HCl in methanol was stirred at ambient temperature for 3 d. At
this point, TLC analysis showed one spot that was positive in a
sodium nitroprusside test. To the reaction mixture were then added
20 mL of dichloromethane and 10 mL of water. The layers were
neutralized using a saturated aqueous NaHCO3 solution, after
which the organic layer was washed once again with water. After
drying with MgSO4, filtration, and removal of the solvent under
1.3 mmol) in 2 mL of toluene was added dropwise. The reaction
mixture was stirred overnight at ambient temperature, after which
the solvent was removed in vacuo. The resulting residue was dis-
solved in 50 mL of dichloromethane; this solution was washed
twice with a saturated aqueous NaHCO3 solution, dried with
MgSO4, and filtered. After removal of the solvent under reduced
1
pressure, 350 mg (90%) of 5 was obtained. Ϫ H NMR (90 MHz,
CDCl3): δ ϭ 3.82 (m, 24 H, OCH2, NCH2), 4.24 (s, 2 H, CH2Cl).
reduced pressure, 170 mg (95%) of L1H was obtained as a yellowish
1
oil. Ϫ H NMR (90 MHz, CDCl3): δ ϭ 1.75 (t, J ϭ 7.3 Hz, 1 H, Ϫ FT-IR (KBr, cmϪ1): ν˜ ϭ 3003, 2872 (br, CH2), 1648 (CϭO),
SH), 2.78 (t, J ϭ 6.0 Hz, 4 H, NCH2CH2O), 3.62 (m, 22 H, OCH2, 1455, 1353 (CH2), 1121 (COC), 712 (CH2). Ϫ EI-MS; m/z: 340
ArCH2N, ArCH2S), 7.24 (m, 4 H, ArH). Ϫ C20H33NO5S (399.5): [M]ϩ. Ϫ C14H26ClNO6 (339.8): calcd. C 49.54, H 7.73, N 4.13;
calcd. C 60.28, H 8.09, N 3.51, S 8.09; found C 60.44, H 8.24, N
3.67, S, 8.18.
found C 49.45, H 7.84, N 4.09.
1-[α-(Thioacetyl)acetyl]-4,7,10,13,16-pentaoxa-1-azacycloocta-
decane (6): To prepare this compound, a similar procedure as de-
scribed for the synthesis of 4 was used. A DMF solution of potas-
sium thioacetate, prepared from K2CO3 (100 mg, 0.72 mmol) and
(α-Chloroacetyl)di-n-butylamine (3): To a solution of di-n-butyl-
amine (1.06 g, 7.8 mmol) in CH2Cl2 (75 mL) which contained
1.1 mL of triethylamine as a base, was added dropwise a solution
of α-chloroacetyl chloride (875 mg, 7.75 mmol) in 25 mL of thioacetic acid (100 mg, 1.3 mmol), was added to a DMF solution
CH2Cl2. The reaction mixture was stirred overnight at ambient
temperature and subsequently washed (3ϫ) with a saturated aque-
ous NaHCO3 solution, and dried with MgSO4. Flash chromatogra-
phy (silica, CHCl3/MeOH ϭ 98.5:1.5, v/v) gave 3 as a colourless
containing 5 (350 mg, 1.03 mmol). After workup using dichloro-
methane and flash chromatography (silica, CHCl3/MeOH ϭ 98:2,
v/v) compound 6 (250 mg, 61%) was obtained as a viscous, yellow
oil. Ϫ 1H NMR (300 MHz, CDCl3): δ ϭ 2.38 (s, 3H, CH3),
oil in 26% yield. Ϫ 1H NMR (90 MHz, CDCl3): δ ϭ 0.82Ϫ1.04 3.50Ϫ3.81 (m, 24 H, OCH2, NCH2), 3.98 (s, 2 H, CH2S). Ϫ 13C
(m, 6 H, CH3), 1.14Ϫ1.70 (m, 8 H, CH2CH2CH3), 3.20Ϫ3.43 (m, NMR (75 MHz, CDCl3): δ ϭ 30.5 (CH3), 32.5 (SCH2), 48.0
4 H, NCH2), 4.07 (s, 2 H, CH2Cl). Ϫ EI-MS; m/z: 205 [M]ϩ. Ϫ
(NCH2), 70.0Ϫ71.5 (m, OCH2), 168.5 (NCϭO), 196.0 (SCϭO). Ϫ
C10H20ClNO (205.7): calcd. C 58.38, H 9.80, N 6.81; found C FT-IR (KBr, cmϪ1): ν˜ ϭ 2867 (br, CH2, CH3), 1691 (SCϭO), 1656,
57.95, H 10.17, N 6.87.
1648 (CϭO), 1468, 1450 (CH2, CH3), 1353 (COCH3), 1120 (COC),
732 (CH2). Ϫ EI-MS; m/z: 379 [M]ϩ. Ϫ No reproducible analysis
could be obtained for this compound.
[α-(Thioacetyl)acetyl]di-n-butylamine (4): Under N2, 299 mg
(2.15 mmol) of K2CO3 was allowed to react with 155 µL of
thioacetic acid in a minimum amount of DMF. The resulting solu-
tion was added dropwise to a solution of 3 (446 mg, 2.16 mmol) in
1-(α-Mercaptoacetyl)-4,7,10,13,16-pentaoxa-1-azacyclooctadecane
(L3H): This compound was prepared using the procedure described
DMF, and this mixture was stirred overnight under the exclusion for compound L2H. Starting from 6 (100 mg, 0.24 mmol), 77 mg
of light. After removal of the solvent under reduced pressure, the
residue was dissolved in CH2Cl2 and washed twice with brine and
(95%) of ligand L3H was isolated. Ϫ 1H NMR (300 MHz, CDCl3):
δ ϭ 2.17 (t, J ϭ 7.52 Hz, 1 H, SH), 3.40Ϫ3.76 (m, 26 H, OCH2,
once with water. The organic layer was dried with MgSO4. Flash NCH2, SCH2). Ϫ 13C NMR (75 MHz, CDCl3): δ ϭ 27.0 (HSCH2),
chromatography (silica, CHCl3/MeOH ϭ 99:1, v/v) yielded 4 as a
yellow oil in 65% yield (345 mg). Ϫ H NMR (90 MHz, CDCl3):
δ ϭ 0.84Ϫ1.05 (m, 6 H, CH2CH2CH3), 1.19Ϫ1.73 (m, 8 H,
47.83 (NCH2), 70.03Ϫ71.78 (m, OCH2), 171.1 (NCϭO). Ϫ FT-IR
(KBr, cmϪ1): ν˜ ϭ 2867 (br, CH2), 2546 (SH), 1654 (CϭO), 1448,
1421 (CH2), 1119 (br, COC). Ϫ EI-MS; m/z: 338 [M]ϩ. Ϫ No repro-
1
CH2CH2CH3), 2.39 [s, 3 H, C(O)CH3], 3.22Ϫ3.43 (m, 4 H, NCH2), ducible analysis could be obtained for this compound.
4.37 (s, 2 H, CH2S). Ϫ FT-IR (KBr, cmϪ1): ν˜ ϭ 2959, 2932, 2874
2,2Ј-Dithiobisbenzoic Acid (7):[30] To a solution of thiosalicylic acid
(CH2, CH3), 1695 (CϭOCH3), 1647 (CϭO), 1428, 1377 (CH2,
(8 g, 51.9 mmol) and NaOH (6.4 g, 160 mmol) in 60 mL of water
CH3), 1355 (CϭOCH3), 733 (CH2). Ϫ EI-MS; m/z: 245 [M]ϩ. Ϫ
was added very slowly, by means of a dropping funnel, 25 mL of a
C12H23NO2S·0.5MeOH (260.4): calcd. C 57.44, H 9.64, N 5.36, S
solution of 15% H2O2 in water. During the addition the reaction
12.26; found C 57.46, H 9.49, N 5.68, S 12.07.
mixture was cooled on ice. After the addition was completed the
mixture was slowly allowed to come to ambient temperature. Fi-
(α-Mercaptoacetyl)di-n-butylamine (L2H): To
a
solution of
4
(130 mg, 0.53 mmol) in 3 mL of CH2Cl2 was added 3 mL of 6% nally, it was stirred overnight at 60°C. The mixture was again
HCl in MeOH. The resulting mixture was stirred overnight. At this
point, TLC analysis revealed a single spot that turned red when
treated with nitroprusside. The reaction mixture was diluted with
CH2Cl2 and neutralized with saturated aqueous NaHCO3, washed
once with water and dried with MgSO4. Removal of the solvent in
cooled on ice, and neutralized with aqueous 2 HCl, which resulted
in the formation of a white precipitate. The product was filtered
off, washed with cold ethanol and cold water, and dried in vacuo.
This procedure yielded 6.3 g (80%) of 7 as a white powder, m.p.
298Ϫ299°C. Ϫ 1H NMR (90 MHz, [D6]DMSO): δ ϭ 7.39 (m, 2
vacuo yielded 108 mg (100%) of L2H as a brownish-yellow oil. Ϫ H, ArH4,4Ј), 7.62 (m, 4 H, ArH5,5Ј, ArH6,6Ј), 8.04 (m, 2H, ArH3,3Ј).
1H NMR (200 MHz, CDCl3): Ϫ FT-IR (KBr, cmϪ1): ν˜ ϭ 3500Ϫ2500 (OH), 1673 (CϭO), 735
0.99Ϫ0.89 (m, H,
CH2CH2CH3), 1.26Ϫ1.37 (m, 4 H, CH2CH2CH3), 1.49Ϫ1.61 (m, (ArH). Ϫ FAB-MS; m/z: 306 [M]ϩ. Ϫ C14H10O4S2·0.25EtOH
δ
ϭ
6
4 H, CH2CH2CH3), 2.17 (t, J ϭ 7.43 Hz, 1 H, SH), 3.20Ϫ3.36 (m,
(317.8): calcd. C 56.20, H 3.74, S 20.69; found C 56.20, H 3.36,
6 H, NCH2CH2, CH2S). Ϫ FT-IR (KBr, cmϪ1): ν˜ ϭ 2958Ϫ2863 S 20.81.
(CH2, CH3), 2554 (SH), 1642 (CϭO), 1454, 1430, 1376 (CH2,
N,N,NЈ,NЈ-Tetra-n-butyl-2,2Ј-dithiobisbenzoylamine (9): A suspen-
CH2). Ϫ EI-MS; m/z: 203 [M]ϩ. Ϫ No reproducible analysis could
be obtained for this compound.
sion of 7 (410 mg, 1.33 mmol) in 10 mL of SOCl2 was heated under
reflux overnight, yielding a clear solution. After removal of the
excess of SOCl2 in vacuo, the resulting residue, 8, was dissolved in
5 mL of dichloromethane. This solution was then added dropwise
to a solution of di-n-butylamine (387 mg, 3.0 mmol) and of triethyl-
amine (220 µL) in 50 mL of dichloromethane. The reaction mixture
1-(α؊Chloroacetyl)-4,7,10,13,16-pentaoxa-1-azacyclooctadecane
(5): To a solution of 4,7,10,13,16-pentaoxa-1-azacyclooctadecane
(300 mg, 1.14 mmol) in 15 mL of toluene, containing some solid
K2CO3 as a base, a solution of α-chloroacetyl chloride (150 mg,
262
Eur. J. Inorg. Chem. 2000, 253Ϫ264

“Crowned” Fe4S4 Clusters as Electrochemical Metal Ion Sen
FULL PAPER
was stirred overnight at ambient temperature, after which it was
washed (3ϫ) with a saturated aqueous NaHCO3 solution. The or-
ganic layer was dried with MgSO4, filtered, and concentrated. After
purification by flash chromatography (alumina, CH2Cl2/EtOH ϭ
99:1, v/v) compound 9 (650 mg, 82%) was obtained as a yellowish,
viscous oil. Ϫ 1H NMR (90 MHz, CDCl3): δ ϭ 0.90 (m, 12 H,
CH2CH2CH3), 1.50 (m, 16 H, CH2CH2CH3), 3.10 (t, J ϭ 8.2 Hz,
4 H, NCH2CH2), 3.52 (t, J ϭ 7.4 Hz, 4 H, NCH2CH2), 7.24 (m, 6
H, ArH4,4Ј, ArH5,5Ј, ArH6,6Ј), 7.67 (m, 2 H, ArH3,3Ј). Ϫ FAB-MS;
m/z: 529 [M ϩ H]ϩ. Ϫ No reproducible analysis could be obtained
for this compound.
and (Bu4N)2[Fe4S4(StBu)4][32] were synthesized using published
methods.
(Bu4N)2[Fe4S4(L1)4] (I): To a stirred solution of (PPh4)2[Fe4S4Cl4]
(38.2 mg, 0.033 mmol) in 15 mL of DMF was added a solution of
L1H (65 mg, 0.163 mmol, 5 equiv.) in 5 mL DMF. After this ad-
dition was completed, 2 mL of 0.081  [Bu4N]OH in methanol was
added dropwise to the reaction mixture. After stirring for several
hours, the reaction mixture was filtered and then concentrated in
vacuo, followed by the addition of Et2O to induce precipitation of
the desired product. Storage at Ϫ20°C overnight resulted in the
formation of the tetrabutylammonium salt I, which was isolated as
a black oil after decanting the supernatant and drying in vacuo
(yield not determined). Ϫ 1H NMR (90 MHz, [D6]DMSO): δ ϭ
0.81Ϫ1.10 [m, br., 24 H, CH3 (NBu4)], 1.19Ϫ1.81 [m, br., 32 H,
CH2CH2CH3 (NBu4)], 2.68 (NCH2, the structure of this peak is
obscured by the solvent peak), 3.02Ϫ3.36 [m, 16 H, NCH2 (NBu4)],
3.36Ϫ3.75 (m, 104 H, OCH2, NCH2, NCH2Ar), 7.20 [m, br., 4 H,
ArH], 7.49Ϫ7.90 (m, 12 H, ArH), 15.08 [ds, br, 8 H, SCH2]. Ϫ UV/
Vis (DMF): λmax (nm) ϭ 300 (sh), 420.
N,N-Di-n-butyl-2-mercaptobenzoylamine (L4H): A solution of 9
(528 mg, 1.0 mmol) and triphenylphosphane (263 mg, 1.0 mmol) in
a mixture of aqueous 37% HCl (2.5 mL), water (7.5 mL), and diox-
ane (40 mL) was stirred overnight at 40°C. The organic solvent was
removed under reduced pressure and 100 mL of dichloromethane
was added to the residue. The organic layer was washed (3ϫ) with
a saturated aqueous NaHCO3 solution, dried with MgSO4, filtered,
and concentrated. Purification by means of flash chromatography
(alumina, CH2Cl2/EtOH ϭ 99:1, v/v) yielded 420 mg (80%) of L4H
(Bu4N)2[Fe4S4(L2)4] (II): To a solution of the thiol ligand L2H
(55 mg, 0.27 mmol) in 5 mL of DMF was added dropwise a solu-
tion/suspension of (Bu4N)2[Fe4S4(StBu)4] (54.6 mg, 0.056 mmol) in
10 mL of DMF. The resulting black solution was stirred for several
hours under dynamic vacuum, and overnight under N2. After fil-
tration, the volume was reduced in vacuo to about 1 mL and a
large amount of Et2O was added. The Schlenk vessel was placed
at Ϫ20°C overnight, which resulted in the precipitation of a black
powder. The product was finally isolated by decantation of the
supernatant, washing with Et2O, and drying under vacuum. This
1
as a viscous oil. Ϫ H NMR (90 MHz, CDCl3): δ ϭ 0.90 (m, 6 H,
CH2CH2CH3), 1.50 (m, 8 H, CH2CH2CH3), 3.10 (t, J ϭ 8.0 Hz, 2
H, NCH2CH2), 3.52 (t, J ϭ 8.0 Hz, 2 H, NCH2CH2), 4.74 (s, 1 H,
SH), 7.25 (m, 4 H, ArH). Ϫ FT-IR (KBr, cmϪ1): ν˜ ϭ 3055 (Ar),
2958, 2932, 2872 (CH2, CH3), 2519 (SH), 1628 (CϭO), 1462, 1425,
1377 (CH2, CH3), 773, 747 (ArH). Ϫ EI-MS; m/z: 265 [M]ϩ. Ϫ
C15H23NOS·0.5EtOH (288.4): calcd. C 66.62, H 9.09, N 4.86, S
11.11; found C 66.71, H 8.64, N 5.31, S 11.01.
N,NЈ-(2,2Ј-Dithiobenzoyl)bis(4,7,10,13,16-pentaoxa-1-azacyclo-
octadecane) (10): To prepare this compound, a similar procedure
as described for 9 was used. Starting from 4,7,10,13,16-pentaoxa-
1-azacyclooctadecane (395 mg, 1.5 mmol), 478 mg (80%) of 10 was
obtained as a yellowish, viscous oil after flash chromatography (al-
umina, CH2Cl2/EtOH ϭ 95:5, v/v). Ϫ 1H NMR (90 MHz, CDCl3):
δ ϭ 3.71 (m, 24 H, OCH2, NCH2), 7.39 (m, 2 H, ArH4,4Ј), 7.62
(m, 4 H, ArH5,5Ј, ArH6,6Ј), 8.04 (m, 2 H, ArH3,3Ј). Ϫ FAB-MS;
m/z: 835 [M ϩ K]ϩ, 819 [M ϩ Na]ϩ, 797 [M ϩ H]ϩ. Ϫ
C38H56N2O12S2·0.5EtOH (831.2): calcd. C 56.81, H 7.31, N 3.44,
S 7.88; found C 57.14, H 7.04, N 3.71, S 7.55.
1
procedure gave II as a black oil in 91% yield (84 mg). Ϫ H NMR
(200 MHz, [D6]DMSO): δ ϭ 0.87Ϫ0.97 (m, 48 H, CH3), 1.29Ϫ1.56
(m, 64 H, CH2CH3, NCH2CH2), 3.18Ϫ3.33 (m, 32 H, NCH2),
13.83 [s, br, 8 H, SCH2]. Ϫ UV/Vis (DMF): λmax (nm) ϭ 301, 415,
620 (sh).
(Bu4N)2[Fe4S4(L3)4] (III): This compound was prepared using the
procedure described for compound II. Starting from L3H (98.4 mg,
0.29 mmol,
5
equiv.) and (Bu4N)2[Fe4S4(StBu)4] (69.6 mg,
0.058 mmol), a black oil was obtained. The yield was not deter-
1
mined. Ϫ H NMR (200 MHz, [D7]DMF): δ ϭ 1.05 [s, br, 24 H,
CH3 (NBu4)], 1.49 [s, br, 16 H, CH2CH3 (NBu4)], 1.85 [s, br, 16 H,
NCH2CH2 (NBu4)], 3.40Ϫ4.13 [m, 112 H, OCH2, NCH2, NCH2
(NBu4)], 15.35, 16.58 [d, br, 8 H, SCH2]. Ϫ UV/Vis (DMF): λmax
(nm) ϭ 300, 410, 660 (sh).
2-Mercaptobenzoyl-4,7,10,13,16-pentaoxa-1-azacyclooctadecane
(L5H): This compound was prepared in a similar way as ligand
L4H. Starting from 10 (200 mg, 0.25 mmol) and triphenylphos-
phane (66 mg, 0.25 mmol), compound L5H was obtained in 75%
yield (150 mg) after flash chromatography (alumina, CH2Cl2/
EtOH ϭ 95:5, v/v). Ϫ 1H NMR (90 MHz, CDCl3): δ ϭ 3.50Ϫ3.63
(m, 24 H, OCH2, NCH2), 3.86 (s, 1 H, SH), 7.25 (m, 4 H, ArH).
Ϫ EI-MS; m/z: 399 [M]ϩ. Ϫ No reproducible analysis could be
obtained for this compound.
(Bu4N)2[Fe4S4(L4)4] (IV): This compound was prepared using the
procedure described for compound II, using L4H (18.5 mg,
0.7 mmol, 5.8 equiv.) and (Bu4N)2[Fe4S4(StBu)4] (14.3 mg,
0.12 mmol). THF was used as the nonsolvent in the isolation of
1
the black oil, IV (yield was not determined). Inspection of the H-
NMR spectrum showed that this compound was contaminated
Synthesis of Cluster Compounds: In situ synthesis of cluster com-
pounds IϪV for initial electrochemical characterization was per-
formed by the stepwise addition of 4Ϫ6 equivalents of thiol ligand
and 4Ϫ6 equivalents of base to the appropriate amount of
(PPh4)2[Fe4S4Cl4] in DMF. A stock solution of 0.081  of [Bu4-
N]OH in methanol was used as the base. Differential pulse voltam-
1
with some free ligand L4H. Ϫ H NMR (200 MHz, [D6]DMSO):
δ ϭ 0.64Ϫ1.57 (m, 112 H, CH2CH2CH3), 2.95Ϫ3.45 [m, 32 H,
NCH2, NCH2 (NBu4)], 3.73 [s, br, 4 H, pArH], 5.32 [s, br., 4 H,
oArH], 7.97 [s, br., 4 H, mArH], 8.28 [s, br, 4 H, mArH]. Ϫ UV/
Vis (DMF): λmax (nm) ϭ 307 (sh), 458.
mograms were recorded after the addition of each equivalent ligand (Bu4N)2[Fe4S4(L5)4] (V): This compound was prepared using the
and base, and after equilibration. Typically, 1.8 mg (1.54 µmol) of
(PPh4)4[Fe4S4Cl4] was used in these experiments. Quantitative syn-
thesis of the cluster compounds for detailed characterization was
performed by using either the chloro cluster reaction starting from
(PPh4)2[Fe4S4Cl4], or the ligand exchange reaction starting from
(Bu4N)2[Fe4S4(StBu)4]. The starting materials, (PPh4)2[Fe4S4Cl4][31]
procedure described for compound II. Amounts used were: 32 mg
(0.08 mmol, 6.3 equiv.) of L5H and 15.18 mg (0.013 mmol) of
(Bu4N)2[Fe4S4(StBu)4]. The yield of the black oil, V, was not deter-
mined. Ϫ H NMR (90 MHz, [D6]DMSO): δ ϭ 0.79Ϫ1.09 [m, 24
H, CH3 (NBu4)], 1.09Ϫ1.78 [m, 32 H, CH2CH2CH3 (NBu4)],
3.01Ϫ3.94 [m, 114 H, NCH2, OCH2, pArH, NCH2 (NBu4)], 5.29
1
Eur. J. Inorg. Chem. 2000, 253Ϫ264
263

R. J. M. Klein Gebbink, S. I. Klink, M. C. Feiters, R. J. M. Nolte
FULL PAPER
[12]
B. V. DePamphilis, B. A. Averill, T. Herskovitz, L. Que, Jr., R.
[s, br, 4 H, oArH], 7.96 [s, br, 4 H, mArH], 8.39 [s, br, 4 H, mArH].
Ϫ UV/Vis (DMF): λmax (nm) ϭ 307 (sh), 452.
H. Holm, J. Am. Chem. Soc. 1974, 96, 4159Ϫ4167.
[13]
´
See for example: S. Ciurli, M. Carrie, J. A. Weigel, M. J. Carney,
T. D. P. Stack, G. C. Papaefthymiou, R. H. Holm, J. Am. Chem.
Soc. 1990, 112, 2654Ϫ2664.
[14]
[15]
[16]
[17]
[18]
R. H. Holm, W. D. Philips, B. A. Averill, J. J. Mayerle, T. Her-
skowitz J. Am. Chem. Soc. 1974, 96, 2109Ϫ2117.
J. G. Reynolds, E. J. Laskowski, R. H. Holm, J. Am. Chem.
Soc. 1978, 100, 5315Ϫ5322.
Acknowledgments
This work was supported by the Dutch Foundation for Chemical
Research (SON) with financial aid from the Dutch Foundation for
Scientific Research (NWO). The authors would like to thank Dr.
J. G. M. van der Linden for fruitful discussions and Mrs. A. M.
Roelofsen for help with electrochemical experiments. In addition,
R. Fokkens of the Institute for Mass Spectrometry at the Univer-
sity of Amsterdam is acknowledged.
See for example: I. Bertini, F. Briganti, C. Luchinat, A. Scoz-
zafava, M. Sola, J. Am. Chem. Soc. 1991, 113, 1237Ϫ1245.
See: A. Nakamura, N. Ueyama, Adv. Inorg. Chem. 1989, 33,
39Ϫ67, and references cited therein.
H. L. Blonk, Ph. D. Dissertation, Univ. of Nijmegen, 1991.
[19] [19a]
A. M. Bond, H. A. O. Hill in Metal Ions in Biological Sys-
tems, vol. 27 (Ed.: H. Siegel, A. Siegel), Marcel Dekker, Inc.,
[19b]
New York, 1991, p. 432Ϫ494. Ϫ
F. A. Armstrong, H. A.
Heering, J. Hirst, Chem. Soc. Rev. 1997, 26, 169Ϫ179.
[20] [20a] C. F. Martens, M. M. G. Bongers, P. J. A. Kenis, R. Czajka,
[1] [1a]
[1b]
R. Cammack, Adv. Inorg. Chem. 1992, 38, 281Ϫ322. Ϫ
M. C. Feiters, J. G. M. van der Linden, R. J. M. Nolte, Chem.
[20b]
H. Beinert, R. H. Holm, E. Münck, Science 1997, 277,
653Ϫ659.
Ber./Recueil 1997, 130, 23Ϫ33. Ϫ
C. F. Martens, H. L.
Blonk, T. Bongers, J. G. M. van der Linden, G. Beurskens, P.
T. Beurskens, J. M. M. Smits, R. J. M. Nolte, J. Chem. Soc.,
Chem. Commun. 1991, 1623Ϫ1625.
[2]
Only recently was evidence provided for an all-ferrous Fe4S4
cluster in biology, viz. in the Fe protein of Azotobacter vinelandii
[21] [21a]
nitrogenase. The structure of this cluster has not been deter-
mined yet. See:
G. P. F. van Strijdonck, J. A. E. H. van Haare, P. J. M.
Hönen, R. C. G. M. van der Schoor, M. C. Feiters, J. G. M.
van der Linden, J. J. Steggerda, R. J. M. Nolte J. Chem. Soc.,
Dalton Trans. 1997, 449Ϫ461. Ϫ [21b] G. P. F. van Strijdonck, J.
A. E. H. van Haare, J. G. M. van der Linden, J. J. Steggerda,
R. J. M. Nolte, Inorg. Chem. 1994, 33, 999Ϫ1000.
F. A. Armstrong, P. A. Cox, H. A. O. Hill, V. J. Lowe, B. N.
Oliver, J. Electroanal. Chem. 1987, 217, 331Ϫ366.
C. D. Hall, N. W. Sharpe, I. P. Danks, Y. P. Sang, J. Chem. Soc.,
Chem. Commun. 1989, 419Ϫ421.
[2a]
H. C. Angove, S. J. Yoo, B. K. Burgess, E.
[2b]
Münck, J. Am. Chem. Soc. 1997, 119, 8730Ϫ8731. Ϫ
K. B.
Musgrave, H. C. Angove, B. K. Burgess, B. Hedman, K. O.
Hodgson, J. Am. Chem. Soc. 1998, 120, 5325Ϫ5326.
R. H. Holm, S. Ciurli, J. A. Weigel, Prog. Inorg. Chem. 1990,
38, 1Ϫ74.
J. M. Berg, R. H. Holm in Iron Sulfur Clusters (Ed.: T. G.
Spiro), J. Wiley & Sons, New York, 1983, pp. 1Ϫ66.
P. D. Beer, Adv. Inorg. Chem. 1992, 39, 79Ϫ157.
[3]
[4]
[5]
[22]
[23]
[24]
[25]
[6] [6a]
R. M. Izatt, K. Pawlak, J. S. Bradshaw, R. L. Bruening,
For a review on lariat ethers, see: G. W. Gokel, Chem. Soc. Rev.
1992, 41Ϫ47.
Chem. Rev. 1991, 91, 1721Ϫ2085. Ϫ [6b] R. M. Izatt, J. S. Brad-
shaw, S. A. Nielsen, J. D. Lamb, J. J. Christensen, D. Sen, Chem.
Rev. 1985, 85, 271Ϫ339.
A. E. Kaifer, D. A. Gustowski, L. Echegoyen, V. J. Gatto, R.
A. Schultz, T. P. Cleary, C. R. Morgan, D. M. Golli, A. M.
Rios, G. W. Gokel, J. Am. Chem. Soc. 1985, 107, 1958Ϫ1965.
S. R. Miller, D. A. Gustowski, Z. Chen, G. W. Gokel, L. Eche-
goyen, A. E. Kaifer, Anal. Chem. 1988, 60, 2021Ϫ2024.
R. Adams, L. H. Ulich, J. Am. Chem. Soc. 1920, 42, 599Ϫ611.
H. Brockman, H. Schodder, Ber. Dtsch. Chem. Ges. 1941,
74B, 73Ϫ78.
[7]
[8]
B. Strijtveen, R. M. Kellogg, J. Org. Chem. 1986, 51,
3664Ϫ3671.
[26]
H. Okuno, K. Uoto, T. Tomohiro, Y. Youinou, J. Chem. Soc.,
Dalton Trans. 1990, 3375Ϫ3381.
[27]
[28]
[9]
L. Overman, S. Smoot, J. D. Overman, Synthesis 1974, 59Ϫ60.
L. Que, Jr., M. A. Bobrik, J. A. Ibers, R. H. Holm, J. Am.
Chem. Soc. 1974, 96, 4168Ϫ4177.
[10]
[29]
[30]
G. Tonnies, J. J. Kolb, Anal. Chem. 1951, 23, 823Ϫ826.
Although this compound is commercially available, its purity
was not sufficient for our purposes.
[11]
Kinetic studies on this type of exchange reaction have been per-
formed by Henderson et al. They showed that in the absence
of acid a thiolate ligand first dissociates from the cluster, leaving
one iron center with a vacant coordination site, followed by
nucleophilic attack at this center of a second thiolate. In the
presence of acid, the dissociation of the first thiolate is facili-
tated by protonation. See: R. A. Henderson, K. E. Oglieve, J.
Chem. Soc., Dalton Trans. 1993, 1467Ϫ1472.
[31]
[32]
G. Christou, C. D. Garner, J. Chem. Soc., Dalton Trans. 1979,
1093Ϫ1094.
G. B. Wong, M. A. Bobrik, R. H. Holm, Inorg. Chem. 1978,
17, 578Ϫ584.
Received March 11, 1999
[I99094]
264
Eur. J. Inorg. Chem. 2000, 253Ϫ264