Alternative synthesis, density functional calculations and proton reactivity study of a trinuclear [NiFe] hydrogenase model compound

Article abstract of DOI:10.1002/ejic.200601077

The trinuclear complex [('S₂')(Ni(PMe₃)} ₂Fe(CO)('S₂')₂] (1) {'S₂' ^2- = 1,2-benzenedithiolate(2-)} has been reported as a structural as well as functional model for [NiFe] hydrogenases since it contains key structural features of the [NiFe] hydrogenase active site, and is oxidized by protons to give [1]⁺ and H₂. Complex 1 formed as an unexpected product from the reaction of [Fe(CO)₂('S ₃')]₂ {'S₃'^2- = bis(2- mercaptophenyl) sulfide(2-)} and [Ni(PMe₃)₂('S ₂')]. Both the iron and nickel centers of 1 are chelated by 'S ₂' donors, but not with any 'S₃' ligand. In order to understand this reaction, the new precursor [Fe(CO)₂( ^siS₃)]₂ (2) (^siS₃ ^2- = bis(2-mercapto-3-trimethylsilylphenyl) sulfide(2-)} was synthesized. Compound 2 readily loses its CO ligand to give [Fe(CO)( ^siS₃)]₂ (3), which consists of two 16 valence electron fragments and could be isolated in the solid state. Conversion of 3 to 2 is feasible with gentle bubbling of CO gas for about 2 min. Treatment of 2 with [Ni(PMe₃)₂('S₂')] resulted exclusively in the formation of complex 1, which confirms that all three 'S₂' ligands in 1 originate from [Ni(PMe₃)₂('S₂')]. Therefore, an alternative synthesis of 1, which does not involve any 'S ₃' ligand, has been developed. Density functional theory (DFT) calculations suggest that the oxidation states of the metal centers are Fe ^II and Ni^II and do not change upon oxidation of 1 to [1]⁺. The unpaired electron in [1]⁺ is located to a large extent on the nickel atoms and the adjacent thiolate donor functions. The charge, however, is distributed over the whole cluster, main parts residing on the 'S₂' ligands. Preliminary constant potential coulometric measurements indicate that 1 mediates the reduction of protons to dihydrogen at a mild potential (-0.48 V vs. NHE). Based upon these experimental and theoretical results, plausible mechanisms for this reduction are briefly discussed. Wiley-VCH Verlag GmbH & Co. KGaA, 2007.

Full text of DOI:10.1002/ejic.200601077

FULL PAPER
DOI: 10.1002/ejic.200601077
Alternative Synthesis, Density Functional Calculations and Proton Reactivity
Study of a Trinuclear [NiFe] Hydrogenase Model Compound
Frank Lauderbach,[a] Raju Prakash,*[a][‡] Andreas W. Götz,*[b][‡‡] Marcela Munoz,[b]
Frank W. Heinemann,[a] Ulrich Nickel,[b] Bernd A. Hess,[c][†] and Dieter Sellmann[a][†]
Keywords: Constant potential coulometry / Density functional calculations / Iron / Nickel / NiFe hydrogenases / S ligands
The trinuclear complex [(‘S2’){Ni(PMe3)}2Fe(CO)(‘S2’)2] (1)
{‘S2’2– = 1,2-benzenedithiolate(2–)} has been reported as a
structural as well as functional model for [NiFe] hydro-
genases since it contains key structural features of the [NiFe]
hydrogenase active site, and is oxidized by protons to give
[1]+ and H2. Complex 1 formed as an unexpected product
of complex 1, which confirms that all three ‘S2’ ligands in
1 originate from [Ni(PMe3)2(‘S2’)]. Therefore, an alternative
synthesis of 1, which does not involve any ‘S3’ ligand, has
been developed. Density functional theory (DFT) calcula-
tions suggest that the oxidation states of the metal centers
are FeII and NiII and do not change upon oxidation of 1 to
from the reaction of [Fe(CO)2(‘S3’)]2 {‘S3’2– = bis(2-mercap- [1]+. The unpaired electron in [1]+ is located to a large extent
tophenyl) sulfide(2–)} and [Ni(PMe3)2(‘S2’)]. Both the iron and
nickel centers of 1 are chelated by ‘S2’ donors, but not with
any ‘S3’ ligand. In order to understand this reaction, the new
on the nickel atoms and the adjacent thiolate donor func-
tions. The charge, however, is distributed over the whole
cluster, main parts residing on the ‘S2’ ligands. Preliminary
constant potential coulometric measurements indicate that 1
mediates the reduction of protons to dihydrogen at a mild
potential (–0.48 V vs. NHE). Based upon these experimental
and theoretical results, plausible mechanisms for this re-
duction are briefly discussed.
precursor [Fe(CO)2(siS3)]2 (2) {siS3 = bis(2-mercapto-3-tri-
2–
methylsilylphenyl) sulfide(2–)} was synthesized. Compound
2 readily loses its CO ligand to give [Fe(CO)(siS3)]2 (3), which
consists of two 16 valence electron fragments and could be
isolated in the solid state. Conversion of 3 to 2 is feasible with
gentle bubbling of CO gas for about 2 min. Treatment of 2
with [Ni(PMe3)2(‘S2’)] resulted exclusively in the formation
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
[Fe]-only[4] hydrogenases. [NiFe] hydrogenases represent the
major class among hydrogenases, consisting of a dinuclear
nickel–iron active site (Figure 1).
Introduction
Hydrogenases comprise a fundamental group of bacte-
rial enzymes that catalyze the reversible interconversion of
protons and molecular hydrogen. Thus, they play a key role
in molecular bioenergetics.[1] In recent years, interest in hy-
drogenases has intensified due to their environmental and
industrial potential in the context of the development of a
pollution-free and sustainable hydrogen economy.[2] Hydro-
genases can be divided into two main groups based on the Figure 1. Schematic structures of the active sites of [NiFe] hydro-
genases. Left: oxidized form from D. gigas.[3a,3b] Right: Reduced
form from D. vulgaris Miyazaki.[3d]
type of metal containing active sites, namely, [NiFe][3] and
[a] Institut für Anorganische Chemie der Universität Erlangen-
Nürnberg,
Numerous spectroscopic,[5] structural,[3,6] and theoreti-
cal[7] studies on various enzymes have provided further in-
sight into the structure and function of the active sites of
[NiFe] hydrogenases. Despite these efforts, the exact nature
of the catalytic cycle and the exact oxidation state of nickel
during the catalytic cycle remain uncertain. This has
prompted the synthesis of a number of low-molecular
weight di- and oligonuclear heterometallic complex ana-
logues to the active site of the [NiFe] enzymes.[8,9] Studies
on such compounds can lead to a better understanding of
the properties of the active site and can shed light on the
catalytic reaction that takes place in the enzyme. The ad-
vantage of such models is that they are low-molecular
Egerlandstrasse 1, 91058 Erlangen, Germany
Fax: +49-9131-852-1165
E-mail: raju.prakash@chemie.uni-erlangen.de
[b] Institut für Physikalische und Theoretische Chemie der Uni-
versität Erlangen-Nürnberg,
Egerlandstrasse 3, 91058 Erlangen, Germany
[c] Lehrstuhl für Theoretische Chemie der Universität Bonn,
Wegelerstrasse 12, 53115 Bonn, Germany
[‡] Present address: Institut für Organische Chemie der Universität
Erlangen-Nürnberg, Henkestrasse 42, 91054 Erlangen, Ger-
many.
[‡‡]Present address: Department of Theoretical Chemistry, Faculty
of Sciences, Vrije Universiteit Amsterdam, De Boelelaan 1083,
1081 HV Amsterdam, The Netherlands;
E-mail: agoetz@few.vu.nl
[†] Deceased.
Eur. J. Inorg. Chem. 2007, 3385–3393
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3385

R. Prakash, A. W. Götz et al.
FULL PAPER
weight compounds and therefore are often more easily iso- complex should result in a new trinuclear NiFe complex
lated and examined by various physicochemical methods containing at least one derivatized ‘S2’ ligand. Conversely if
than the macromolecular enzyme itself. In addition, elec- the reaction proceeded via a ligand-exchange pathway, the
tronic and steric modifications can be generated by ad- use of a derivatized ‘S3’ ligand in the dinuclear iron precur-
justing the ligand environment.
sor complex should result in exclusively the trinuclear com-
Recently, two [NiFe] model complexes [Fe(CO)2(CN)2(µ- plex 1. Herein we have chosen the twofold silylated deriva-
2–
SCH2CH2CH2S)Ni(S2CNR2)] [R = Et and -(CH2)5-] were tive of the ‘S3’ ligand, siS3–H2 {siS3 = bis(2-mercapto-3-
synthesized which up to now resemble the structural fea- trimethylsilylphenyl) sulfide(2–)}, for the synthesis of the
tures of hydrogenases most closely.[10] However, in practice, derivative of the [Fe(CO)2(‘S3’)]2 precursor. Since the SiMe3
none of the reported sulfur-bridged [NiFe] model com- group is a good NMR probe, it enables the detection of the
plexes shows any hydrogenase-like reactivity.
reaction pathway of the formation of 1.
We have previously reported on the sulfur-bridged trinu-
clear complex [(‘S2’){Ni(PMe3)}2Fe(CO)(‘S2’)2] (1) {‘S2’2–
=
1,2-benzenedithiolate(2–)}, which stoichiometrically re-
duces protons to dihydrogen and in turn gets oxidized to
[(‘S2’){Ni(PMe3)}2Fe(CO)(‘S2’)2]+ ([1]+).[11] The reversible
conversion of 1/[1]+ was also achieved with ferrocenium salt
and NBu4BH4, respectively. Moreover, the structural pa-
rameters (bond lengths and angles) of 1 and [1]+ are nearly
identical. This structural rigidity during the redox reaction
indicates a low reorganization barrier and therefore favors
a rapid electron transfer.[12] This result is insofar highly
interesting as many oxidoreductases undergo redox reac-
tions which are accompanied only by marginal structural
changes. These intriguing facts inspired us to investigate the
properties of 1 in greater detail. Herein, we describe the
Scheme 1. Synthesis of complex 1.
The silylated precursor complex [Fe(CO)2(siS3)]2 (2) was
exact formation pathway and a novel straightforward syn- prepared similarly to the synthesis of the unsubstituted par-
thetic route of 1. Density functional Theory (DFT) calcula- ent compound as reported in the literature,[13] except that
tions reveal important insights into the oxidation reaction the reaction was performed under CO atmosphere. Ad-
of 1 to [1]+. In addition, we report preliminary coulometric dition of the ligand siS3–H2 to [Fe(bda)(CO)3] (bda =
studies which employed 1 as an electrocatalyst for the re- benzylideneacetone) in THF led to a dark green solution
duction of protons to dihydrogen at a low reduction poten- and gas evolution. Keeping this solution at –20 °C for two
tial.
weeks, brown crystals suitable for X-ray structure analysis
were obtained. The molecular structure of 2 is shown in
Figure 2 and selected bond lengths and bond angles are
given in Table 1. The structural parameters of the silylated
centrosymmetric compound 2 differ only marginally from
the parent compound [Fe(CO)2(‘S3’)]2.[13] Complex 2 con-
Results and Discussion
Synthesis
It has already been reported that the trinuclear NiFe sists of two distorted octahedral fragments that are linked
complex 1 formed by the reaction of [Fe(CO)2(‘S3’)]2 {‘S3’2– by an edge. Within these fragments, the thiolate donors are
= bis(2-mercaptophenyl) sulfide(2–)} with [Ni(PMe3)2(‘S2’)] cis to each other. One CO is coordinated trans to a bridging
according to Scheme 1.[11] In this reaction compound 1 was thiolate S atom, whereas the other CO is in trans position
the only isolated product with 18% yield (it is not possible to the thioether S donor. The distances around Fe, includ-
to give the exact stoichiometry for Scheme 1, however, the ing Fe–CO bonds, are in the range usually found for dia-
given numbers denote only the ratio of the initial reactants magnetic octahedral thiolate-thioether complexes.[12] The
used). Both the iron as well as the nickel coordination cen- Fe–thiolate bond lengths are longer than the Fe–thioether
ters of 1 are chelated by ‘S2’ donors, but surprisingly not bond lengths. The Fe–Fe distance of 341.51(6) pm indicates
with any ‘S3’ ligand. Therefore, it is interesting to know how that there is no metal–metal interaction.
1 can be formed from the reactants in which the Fe center
Compound 2 is diamagnetic and sparingly soluble in
is not bearing any ‘S3’ ligand. We can think about two pos- most common solvents. However, it is better soluble than
sible pathways for the formation of 1: in the first route, an the unsubstituted derivative. It is stable in CO saturated sol-
initial reductive cleavage of the ‘S3’ ligand at the iron center vents for about two weeks. On prolonged standing, it grad-
to an ‘S2’ ligand and subsequent reaction of this intermedi- ually decomposes to form a stable five-coordinate dinuclear
ate species with [Ni(PMe3)2(‘S2’)] to form 1; in the second species [Fe(CO)(siS3)]2 (3).
route, a ligand exchange type of reaction in which all ‘S2’
The molecular structure of 3 is depicted in Figure 3 and
ligands originate from the reactant [Ni(PMe3)2(‘S2’)]. If 1 selected bond lengths and angles are given in Table 1. The
would be formed via the reductive cleavage pathway, the use molecule can be described as two edge-linked distorted tri-
of a derivatized ‘S3’ ligand in the dinuclear iron precursor gonal bipyramids. It contains two 16 valence electron frag-
3386
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 3385–3393

Trinuclear [NiFe] Hydrogenase Model Compound
FULL PAPER
When a solution of 3 in THF is either stirred under an
atmosphere of CO for about 30 min or by bubbling of CO
gas for about 2 min, compound 2 is formed again. The IR
spectroscopic monitoring of the reaction confirmed that the
ν(CO) band of 3 at 1975 cm–1 vanished and two bands at
2032 cm–1 and 1993 cm–1 appeared which correspond to the
ν(CO) bands of 2.
In analogy to the synthesis of 1, complex 2 was treated
with [Ni(PMe3)2(‘S2’)] in THF at 60 °C for about 4 h. Dur-
ing the course of the reaction, the green suspension turned
into a dark-red solution. The IR spectrum of the solution
showed a ν(CO) band at 1916 cm–1. Treatment of the solu-
tion with methanol resulted in a black-grey microcrystalline
1
material. The H and 13C NMR spectra of the isolated ma-
terial (see Experimental Section) showed signals identical
to those obtained for 1. It did not show any signals for
SiMe3 groups, indicating the unique formation of 1. An X-
ray structure determination of a single crystal obtained
from this material further corroborated this result. Thus,
the formation of complex 1 via the ligand exchange type of
reaction has unambiguously been proven.
Figure 2. Molecular structure of 2 (50% thermal ellipsoids, H
atoms and solvent molecules omitted).
Table 1. Selected bond lengths [pm] and angles [°] of 2 and 3.
Based upon this encouraging result we devised an alter-
native way of synthesizing 1 without using the ‘S3’ ligand
(Scheme 2). Treatment of FeCl2·4H2O with two equivalents
of deprotonated ‘S2’2– in THF led to a red solution of
[Fe(‘S2’)2]2– which was subsequently treated with PMe3 and
Ni(BF4)2·6H2O. After addition of one more equivalent of
‘S2’2– a slurry formed. Upon diffusion of CO gas into the
slurry, a dark red solution formed gradually. The IR spec-
trum of the solution exhibited a ν(CO) band at 1916 cm–1.
Further work-up of the mixture afforded complex 1 in an
isolable yield of 18%. It is noteworthy that a strict order of
addition of the reactants is necessary for the formation of
1 in this one-pot synthesis.
2
3
Fe1–Fe1A[a]
Fe1–S1
Fe1–S2
341.51(6)
230.55(6)
227.91(6)
229.64(6)
233.65(6)
179.5(2)
178.1(2)
114.3(3)
94.97(2)
85.49(2)
87.92(2)
271.95(6)
220.5(1)
226.7(2)
221.6(2)
229.0(1)
178.0(4)
–
114.1(4)
74.29(3)
88.06(4)
89.32(4)
Fe1–S3
Fe1–S3A[b]
Fe1–C1
Fe1–C2
C1–O1
Fe1–S3–Fe1A[a]
S1–Fe1–S2
S2–Fe1–S3
[a] Fe1A in 2 = Fe2 in 3. [b] S3A in 2 = S6 in 3.
ments that are linked by thiolate bridges. The CO ligands
and thioether donors are in axial positions with the two
thiolate donors and iron atoms lying in the base plane. Due
to the loss of a CO group the Fe–Fe distance in 3 [271.95(6)
pm] decreases by 70 pm compared to that of 2 [341.51(6)
pm].
Scheme 2. Alternative synthesis of complex 1.
DFT Calculations
Complex 1 readily reacts with protons to give dihydrogen
and [1]+. When the solution of [1]+ is treated with hydride,
the reduced complex 1 is formed again. Both complex 1
and its oxidation product [1]+ were characterized by single-
crystal X-ray structure analysis.[11] In order to gain insight
Figure 3. Molecular structure of 3 (50% thermal ellipsoids, H
atoms and solvent molecules omitted).
Eur. J. Inorg. Chem. 2007, 3385–3393
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3387

R. Prakash, A. W. Götz et al.
FULL PAPER
Table 2. Comparison of selected bond lengths [pm] and angles [°] obtained experimentally and calculated theoretically for 1 and [1]+
along with the structural changes ∆ upon oxidation. The average of experimental values which are symmetry-equivalent in the calculations
is given.
[1]
Calcd.
[1]+
Calcd.
∆
Exp.
Exp.
Calcd.
Exp.
Fe1–S1/S2
Fe1–C1
223.6
173.6
260.5
241.5
241.1
221.2
234.4
90.26
84.76
155.96
55.23
62.39
221.9(1)
173.8(5)
259.6(1)
240.5(1)
237.0(1)
220.3(2)
231.4(1)
90.00(4)
84.01(5)
154.24(5)
55.18(3)
62.41(3)
222.6
175.4
259.7
246.0
237.6
224.1
232.7
90.66
84.17
155.51
56.54
61.73
220.1(2)
174.4(8)
258.3(2)
243.1(2)
232.6(2)
221.3(2)
229.2(2)
91.37(8)
83.33(8)
154.17(9)
56.14(3)
61.93(4)
–1.0
+1.8
–0.8
+4.5
–3.5
+2.9
–1.7
+0.40
–0.59
–0.45
+1.31
–0.66
–1.8
+ 0.6
–1.3
+2.6
–4.4
+1.0
–2.2
+1.37
–0.68
–0.07
+0.96
–0.48
Fe1–Ni1/Ni2
Ni1–Ni2
Ni1/Ni2–S2/S1
Ni1/Ni2–P1
Ni1/Ni2–S5
S1–Fe1–S2
S1–Fe1–S1A
S1–Fe1–S2A
Ni2–Fe1–Ni1
Ni2/Ni1–Ni1/Ni2–Fe1
into this redox process and to determine the details of the structures, while the calculations find a coplanar arrange-
electronic structure, DFT calculations were carried out on ment. However, it was found that the normal modes de-
cluster 1 and its cation [1]+. The symmetry has been con- scribing this movement are very low in both redox states of
strained to C2v for both the neutral and the oxidized cluster the cluster (around 50 cm–1). Thus, such a distortion should
in all calculations. Therefore, one of the PMe3 groups had be facile if a small force is exerted on the cluster due to the
to be rotated around the P–Ni bond as compared to the X- crystal environment. No significant change of the structural
ray structures. Test calculations without symmetry con- parameters upon oxidation of the cluster 1 is observed both
straints showed that the structures with C2v symmetry are for the experimental and the calculated structures. However,
indeed lower in energy, thus legitimating this procedure.
the trend of the changes of the structural parameters upon
As illustrated in Table 2, the theoretical structure deter- oxidation, e.g. slight elongation of the Ni1–Ni2 distance
mination (geometry optimization) yielded bond lengths and and shortening of the Ni1/Ni2–S2/S1 distance, is in general
bond angles in very good agreement with the X-ray struc- well reproduced by the calculations.
ture determination. For clarity, the molecular structure of
1 is depicted in Figure 4.
The calculation of the harmonic forcefield results in a
frequency of 1955 cm–1 (exp. 1910 cm–1 in KBr/1916 cm–1
in CH2Cl2) for the ν(CO) band in 1, and 1998 cm–1 (exp.
1952 cm–1 in KBr/1976 cm–1 in CH2Cl2) in [1]+. The experi-
mentally observed blue-shift (42 cm–1 in KBr/60 cm–1 in
CH2Cl2) upon oxidation is thus confirmed. The Kohn–
Sham molecular orbitals were used for analyzing the elec-
tronic structure of our complexes. A stability analysis[14] of
the restricted Hartree–Fock wavefunction of 1 showed that
it is stable with respect to singlet and triplet instabilities.
The S2 value of 0.787 (S2 = 0.75 for a doublet without spin
contamination) in the unrestricted Hartree–Fock wave
function of [1]+ does not indicate substantial multi-refer-
ence character for both compounds which would render
such an analysis questionable. The results of several types
of population analyses are summarized in Table 3. Note
that the assignment of a charge to an atom within a polya-
tomic molecule cannot be carried out unambiguously. Dif-
ferent kinds of population analyses therefore will give dif-
ferent answers. It should be mentioned that the Mulliken
Figure 4. Molecular structure of 1 (50% thermal ellipsoids, H
atoms and solvent molecules omitted, dashed lines represent the
metal–metal distances).[11]
Differences in experimental bond lengths and bond population analysis[15] in particular is strongly dependent
angles, which are symmetry-equivalent in the DFT calcula- on the basis set and doubtful for an extended basis set, as
tions, can be attributed to packing effects in the crystals. far as absolute values of charges are concerned. Therefore,
Therefore the data from the calculations were compared to the common features of a Mulliken analysis, a natural pop-
the average value of these experimental bond lengths and ulation analysis (NPA),[16] and of a charge analysis based
angles. The plane defined by the ‘S2’ ligand coordinating to on generalized atomic polar tensors[17] are extracted. The
the nickel atoms (denoted as ‘S2Ni’ in the following) and the latter is dependent on the basis set only implicitly by the
axis of the CO ligand are slightly tilted in the experimental way of the general quantum-mechanical description.
3388
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 3385–3393

Trinuclear [NiFe] Hydrogenase Model Compound
FULL PAPER
Table 3. Partial charges and d orbital occupations as calculated by fined to a more restricted region of the molecule. A small
various types of analysis of the density obtained in the DFT calcu-
negative spin polarization arises at the iron atom, but the
lations. ‘S2Fe’ denotes the ‘S2’ ligands coordinating to the iron
main part of the spin polarization is located on the
atom, ‘S2Ni’ the ‘S2’ ligand coordinating to the nickel atoms.
Ni2‘S2Ni’ fragment of the cluster. About two thirds of the
1
[1]+
magnetic moment are located on the nickel d orbitals and
about one third on the ‘S2Ni’ ligand. Approximately two
thirds of the spin polarization on the ‘S2Ni’ ligand are lo-
cated on the thiolate donor functions. This data can be very
nicely visualized by an analysis of the molecular orbitals.
Upon oxidation an electron is removed from the HOMO of
1 which has essentially the same form as the SOMO in
[1]+ and is mainly formed from d orbitals on the nickel
atoms and p orbitals on the adjacent sulfur atoms (Fig-
ure 5). The localization of the unpaired electron at these
atoms with a small contribution at the iron atom can be
clearly seen from a plot of the spin density of [1]+ (Fig-
ure 6). These findings are consistent with EPR measure-
ments, which have been reported in ref.[11]
α+β
α-β
Mulliken population analysis
Fe
Ni
S1/S2
‘S2
S5
‘S2
PMe3
CO
–0.545
0.010
–0.015
0.008
–0.009
–0.223
0.302
0.129
6.925
8.705
–0.569
–0.030
0.053
0.276
0.027
0.026
0.425
0.201
6.925
8.662
–0.116
0.315
0.040
0.100
0.109
0.301
–0.010
0.003
–0.104
0.329
Fe
’
Ni
’
d(Fe)
d(Ni)
Natural Population Analysis
Fe
Ni
S1/S2
‘S2
S5
‘S2
PMe3
CO
–0.129
0.562
–0.012
–0.407
–0.115
–0.727
0.254
–0.135
0.593
0.046
–0.168
–0.080
–0.489
0.339
0.096
7.621
8.909
–0.099
0.305
0.042
0.104
0.113
0.302
–0.009
–0.004
–0.096
0.317
Fe
’
Ni
’
0.038
7.626
8.964
d(Fe)
d(Ni)
Charges from generalized atomic polar tensors
Fe
Ni
–0.679
–0.145
–0.076
–0.171
–0.019
–0.041
0.443
–0.672
–0.021
0.008
0.082
–0.015
0.048
0.486
0.539
–
–
–
–
–
–
–
–
S1/S2
Fe
‘S2
S5
‘S2
’
Ni
’
PMe3
CO
0.435
Figure 5. Plot of the highest occupied molecular orbital (HOMO)
of 1 (H atoms omitted).
The oxidation states of iron and nickel (represented by
their d occupations) do not change upon oxidation of 1 to
[1]+. They are compatible with the formulation of FeII and
NiII. This is consistent with the observation that only very
small structural changes take place when 1 is oxidized to its
cation. For example, the iron atom lies 4.95(1) pm above
the plane defined by the atoms S1, S2, S1A and S2A in the
crystal structure of 1. Its distance from the plane defined
by the atoms S1, S2, S3 and S4 in the crystal structure of
[1]+ is with 4.98(2) pm almost identical, indicating that the
iron center is not affected by the oxidation. The computed
values for these distances in 1 and [1]+ coincide with the
experimental data. No significant differences of the partial
charges on any particular atoms are observed upon oxi-
dation of the cluster. This means that the positive charge in
[1]+ cannot be assigned to individual atoms but is distrib-
uted over the whole cluster. What at first seems surprising
Figure 6. Plot of the spin density in [1]+ (H atoms omitted).
At first this localization of the spin density on the nickel
is the fact that no charge ends up on the metal centers. atoms and the adjacent sulfur atoms seems to be contradic-
However, this is in line with the negligible changes in the d tory to the above described delocalization of the positive
level occupation. On the other hand the partial charges on charge in [1]+. Things become clear, however, if we imagine
all ligands are increased upon oxidation. Thus, main parts the oxidation as a two-step procedure. In the first step an
of the positive charge in [1]+ are residing on the ‘S2’ ligands. electron is removed from the HOMO, which leaves an un-
The picture is very different for the spin polarization. As paired electron in an orbital located mainly at the nickel
opposed to the positive charge, the unpaired electron is con- and adjacent sulfur atoms. At the same time the electronic
Eur. J. Inorg. Chem. 2007, 3385–3393
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3389

R. Prakash, A. W. Götz et al.
FULL PAPER
structure of the complex cation relaxes and a substantial tion of H2 was estimated by gas chromatography. After the
donation of paired electrons from the ‘S2’ ligands takes end of and 2 h of electrolysis 0.09Ϯ0.004 and
1
place to compensate for the lack of electrons on the metal 0.15Ϯ0.004 mmol of H2 was detected, respectively. When
centers. Thus, the ‘S2’ ligands serve as a kind of electron the electrolysis was performed under identical conditions
density reservoir, which prevent the buildup of positive but in the absence of the complex, the number of coulombs
charge on the nickel atoms upon removal of one electron.
did not increase with time and no H2 formation was ob-
served. The IR spectrum of the electrolyzed solution exhib-
ited the CO band at 1917 cm–1, which confirmed that the
CO molecule did not dissociate during electrolysis. These
results suggest that complex 1 is capable of mediating the
reaction. But, the amount of charge transferred is nearly
twice as high as the corresponding amount of H2 produced,
which indicates that some side reaction takes place. Never-
theless, 1 seems to be the first [NiFe] complex that mediates
the reduction of protons to H2 at a very mild reduction
potential. Efforts are needed to optimize the coulometric
experimental conditions.
The exact site of 1 where the proton initially interacts
and is reduced to H2 is not yet been understood. The reac-
tion may either be preceded via an initial formation of a
protonated thiolate species or via a NiIII–H like intermedi-
ate.[18] According to the DFT studies on 1 and [1]+, an ini-
tial formation of a protonated thiolate species (at the nickel
centers) is fairly reasonable, because the HOMO has also a
large amplitude on the thiolate donor atoms adjacent to the
nickel atoms. Furthermore, a considerable part of the spin
density of the oxidized cluster is located on these atoms.
Constant Potential Coulometry
The cyclic voltammogram of 1 (given in ref.[11]) exhibits
three quasi-reversible one-electron waves at –1.071, –0.095,
and 0.844 V relative to NHE, which are assigned to the re-
dox couples 10/–, 10/+, and 1+/2+, respectively.[11] The poten-
tial of the second wave lies within the range which has been
observed for several [NiFe] hydrogenase redox reac-
tions.[3c,5d] In addition, the small difference in the X-ray
structure parameters of 1 and [1]+ suggests that the electron
transfer should be quite fast as a result of a low reorganiza-
tion barrier. This exciting result prompted us to check the
ability of 1 in catalyzing the reduction of H+ to H2 by con-
stant potential coulometry. A solution of 3.2 mmol of
HBF4 in CH2Cl2 was electrolyzed in the presence of
0.1 mmol of 1 by a platinum electrode at –0.48 V relative to
NHE for 2 h. Interestingly, the number of coulombs in-
creases with time (Figure 7). H2 gas evolution was observed
during the electrolysis. Samples of the gas phase were taken
at 0, 1 and 2 h of electrolysis, and the amount of the forma-
Conclusions
The formation of 1 from the reaction between [Fe(CO)2-
(‘S3’)]2 and [Ni(PMe3)2(‘S2’)] was not understood up to
now. Through the synthesis of a silylated precursor [Fe(CO)2-
(siS3)]2 (2), an earlier observation of the formation of 1 from
[Fe(CO)2(‘S3’)]2 and [Ni(PMe3)2(‘S2’)] is now understood to
be a ligand exchange process. Based on this result, an alter-
native route for the synthesis of 1 in a one-pot reaction has
been developed. The reported trinuclear [NiFe] complex 1
readily reduces protons to H2 and in turn undergoes oxi-
dation to form [1]+. DFT calculations suggest that the oxi-
dation states of the metal centers are FeII and NiII and do
not change upon oxidation of 1 to [1]+. The unpaired elec-
tron in [1]+ is located to a large extent on the nickel atoms
and the adjacent thiolate donor functions. The charge, how-
ever, is distributed over the whole cluster and main parts
reside on the ‘S2’ ligands. These results are compatible with
the protonation of a thiolate group as a first step of the
reduction mechanism. Preliminary constant potential cou-
lometric results indicate that the trinuclear model complex
has catalytic activity for proton reduction under mild po-
tential.
Experimental Section
General: Unless noted otherwise, all manipulations were carried
out in absolute solvents under exclusion of air using standard
Schlenk techniques. IR spectra of solutions were recorded in CaF2
Figure 7. Top: plot of number of coulombs vs. electrolysis time in
the presence (a) and in the absence (b) of complex 1. Bottom: plot
of mmols of H2 produced per mmol of 1 vs. electrolysis time.
3390
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 3385–3393

Trinuclear [NiFe] Hydrogenase Model Compound
FULL PAPER
cuvettes with compensation of the solvent bands, and solids were
measured as KBr pellets. NMR spectra were recorded at room tem-
perature and chemical shifts are quoted in δ scale (downfield shifts
positive) with the residual signals of the deuterated solvent used
as an internal reference. Mass spectra were measured in the field
desorption (FD) or fast atom bombardment (FAB) mode. Physical
measurements were carried out with the following instruments: IR:
Perkin–Elmer 983, Perkin–Elmer 1600 FTIR, and Perkin–Elmer
16PC FTIR; NMR: JEOL FT-JNM-GX 270, Lambda LA 400.
MS: JEOL MSTATION 700. Constant Potential Coulometry: Ex-
periments were performed on Autolab Potentiostat/Galvanostat
equipped with a personal computer. A coulometric cell assembly
consisting of a Pt-gauge working electrode, a platinum wire counter
electrode, and a 1  SCE reference electrode was used. The refer-
ence and counter electrodes were separated from the experimental
solution by a glass frit. This closed cell assembly (total cell volume
of 310 mL) provides with a probe for sampling. This was fitted
with a rubber septum and the gas samples were taken before, dur-
ing, and at the end of the electrolysis. The amount of H2 gas pro-
duced during electrolysis was detected by gas chromatography. The
compounds 1,2-Benzenedithiol (‘S2’–H2),[19] [Fe(bda)(CO)3],[20]
PMe3,[21] bis(2-mercapto-3-trimethylsilylphenyl) sulfide(2–) (siS3–
H2),[22] and (NMe4)2[Fe(CO)2(‘S2’)2][13] were prepared according
the literature methods.
ing the course, the suspension was changed to a clear solution,
which was stirred again at 60 °C for 4 h. Half of the solvent was
evaporated, the solution was filtered and MeOH (140 mL) was
added. Black-grey microcrystals formed, were filtered, washed with
MeOH (25 mL) and dried in vacuo. Yield: 140 mg (18%). Charac-
terization: similar as above.
Density Functional Theory (DFT) Calculations: The DFT modules
of the program package Turbomole[23] were used and the BP86
exchange-correlation functional[24] with a triple-zeta valence-polar-
ized Gaussian basis set[25] was employed for all calculations. The
resolution of identity (RI) technique[26] was employed to accelerate
[16]
the calculations. The natural population analysis (NPA)
was
done with Gaussian98[27] employing the Kohn–Sham molecular or-
bitals obtained from the Turbomole calculations. The program
gOpenmol[28] was used for the visualizations (Figure 5 and Fig-
ure 6). All structures were fully optimized and characterized as true
minima on the potential energy hypersurface by means of a vi-
brational analysis. Accurately converged SCF results (a termination
threshold of at least 10–8 a.u. for the total energy) were used to
guarantee satisfactory accuracy in the force constant calculations.
The force constants were obtained as numerical first derivatives of
the analytical energy gradients as implemented in the parallel PVM
code SNF.[29]
Constant Potential Coulometry: In a typical experiment, a coulo-
metric cell (320 mL capacity) containing a solution of 1 (80 mg,
0.103 mmol) in CH2Cl2 (100 mL), HBF4 (54% solution in Et2O,
5 mL, 3.2 mmol) and NBu4PF6 (3.874 g, 10 mmol) was electrolyzed
at –0.15 V vs. 0.1  SCE for 2 h (25 µL gas samples were taken
from the cell at every hour intervals) and the formation of H2 gas
was analyzed by GC (TC detector, retention time: 0.733 s). An
identical experiment without 1 was also performed.
[Fe(CO)2(siS3)]2 (2): All manipulations were carried out under CO
atmosphere. To a solution of [Fe(bda)(CO)3] (716 mg, 2.503 mmol)
in THF (5 mL) a solution of siS3–H2 (974 mg, 2.473 mmol) in THF
(5 mL) was added. A gas evolution was observed and the solution
color changed to dark-green. This solution was kept at 20 °C. The
black green microcrystal that formed after 15 d, were separated
from the mother liquor and dried with a stream of CO. Yield:
250 mg (17.5%); C40H48Fe2O4S6Si4·2THF (1153.46): calcd.
C
X-Ray Structure Determination of [Fe(CO)2(siS3)]2·2THF (2·2THF)
and [Fe(CO)(siS3)]2·3THF (3·3THF): Suitable single crystals were
embedded in protective perfluoro polyether oil. Data were collected
at 100 K on a Bruker–Nonius KappaCCD diffractometer using
Mo-Kα radiation (λ = 71.073 pm), and a graphite monochromator.
A numerical absorption correction was applied.[30] for 2·2THF,
while for 3·3THF a semiempirical absorption correction based on
multiple scans was performed (SADABS).[31] The structures were
solved by direct methods; full-matrix least-squares refinement was
carried out on F2 using SHELXTL NT 6.12.[32] All non-hydrogen
atoms were refined anisotropically. All hydrogen atoms were geo-
metrically positioned; their isotropic displacement parameters were
tied to those of their corresponding carrier atoms by a factor of
1.2 or 1.5. In the case of 3·3THF there is pseudosymmetry which
emulates a Pccn unit cell, but is not fully supported by the diffrac-
tion pattern which is consistent with the correct space group
P21212. The Flack parameter of approx. 0.5 suggests a centrosym-
metric space group, though. A solution of the structure in Pccn is
possible, but the subsequent refinement results in a dramatical in-
crease of all R-values and estimated standard deviations. Therefore,
the presence of an inversion twin is the best explanation and consis-
tent with the experimental data. One of the three THF solvent
molecules present in the asymmetric unit is disordered on two dif-
ferent crystallographic twofold axes. This disordered solvent mole-
cule has been refined using constraints and no hydrogen atoms have
been introduced for it. Selected crystallographic data are summa-
rized in Table 4.
49.98, H 5.59, S 16.68; found C 50.43, H 5.93, S 16.67. IR (KBr):
˜
NMR spectra that were readily analyzed could not be obtained.
ν = 2029, 1988 (CO) cm–1. Due to the lability of 2 in solution,
[(‘S2’){Ni(PMe3)}2Fe(CO)(‘S2’)2] (1) Starting from 2: A THF sus-
pension (30 mL) of a mixture of 2 (630 mg, 0.624 mmol) and
[Ni(PMe3)2(‘S2’)] (436 mg, 1.220 mmol) was stirred at 60 °C for 4 h.
During the course of time, the suspension changed to a clear dark-
red solution. This was filtered, and treated with MeOH (240 mL).
Black-grey microcrystals precipitated, were removed after 24 h,
washed with MeOH (15 mL) and dried in vacuo. Yield: 90 mg
(19%). IR (KBr): ν = 1910 (CO) cm–1. IR (CH Cl ): ν = 1916 (CO)
˜
˜
2
2
cm–1. 1H NMR (269.6 MHz, CD2Cl2): δ = 7.33 (m, 4 H, C6H4),
6.81 (m, 4 H, C6H4), 6.67 (m, 2 H, C6H4), 6.33 (m, 2 H, C6H4),
1.97 (s, 18 H, PC3H9) ppm. 13C{1H} NMR (100.4 MHz, CD2Cl2):
δ = 207.3 (CO), 146.7, 131.0, 130.3, 125.2, 123.4, (C6H4), 18.2
(PMe3) ppm. 31P{1H} NMR (161.7 MHz, CD2Cl2): δ = 214.6
(PMe3) ppm. C25H30FeNi2OP2S6 (744.09): calcd. C 38.79, H 3.91,
S 24.85; found C 38.54, H 3.82, S 24.80.
Alternative Synthesis of 1: To a suspension of FeCl2·4H2O (200 mg,
1.001 mmol) in THF (10 mL) a solution of ‘S2’–H2 (0.23 mL,
2.001 mmol) and BuLi (2.5  in n-hexane, 1.61 mL, 4.025 mmol)
in THF (10 mL) was added. The resulting red solution was treated
with PMe3 (0.21 mL, 2.04 mmol) and then with a solution of
Ni(BF4)2·6H2O (684 mg, 2.001 mmol) in THF (15 mL). To this
mixture again a solution of ‘S2’–H2 (0.12 mL, 1.046 mmol) and
BuLi (2.5  in n-hexane, 0.81 mL, 2.025 mmol) in THF (10 mL)
was added. A slurry was formed, to this, additional THF (75 mL)
was added. Then the Schlenk tube was fitted with a septum. CO
gas (60 mL, 2.68 mmol) was allowed to diffuse slowly into the mix-
ture, and the solution was stirred vigorously at 20 °C for 15 h. Dur-
CCDC-608118 and -608119 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Eur. J. Inorg. Chem. 2007, 3385–3393
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3391

R. Prakash, A. W. Götz et al.
FULL PAPER
Table 4. Selected crystallographic data for 2·2THF and 3·3THF.
Compound
2·2THF
3·3THF
Formula
C48H64Fe2O6S6Si4
1153.41
C50H72Fe2O5S6Si4
1169.50
orthorhombic
P21212
Mr [gmol–1]
Crystal system
Space group
triclinic
¯
P1
a [pm]
b [pm]
c [pm]
α [°]
971.05(5)
1155.9(1)
1358.4(2)
103.238(7)
99.060(5)
103.119(6)
1.4097(3)
1
1.359
0.865
0.35ϫ0.32ϫ0.14
0.764/0.898
34052
2117.5(3)
27.526(4)
982.7(1)
90
90
90
5.7278(13)
4
1.356
β [°]
γ [°]
V [nm3]
Z
ρcalcd[gcm–3]
µ[mm–1]
0.852
Crystal size [mm3]
Tmin./Tmax.
0.32ϫ0.23ϫ0.08
0.756/0.934
39756
13148
9953
635
438/–420
4.47; 8.60
Measured reflections
Independent reflections
Observed reflections[a]
Refined parameters
Largest diff. peak and hole [enm–3]
6715
5335
304
585/–640
3.58; 8.76
[b]
R1[a]; wR2 (%)
[a] [IϾ2σ(I)]. [b] All data.
Struct. Bond. 1998, 91, 1–29; c) M. Frey, Struct. Bonding
(Berlin) 1998, 90, 97–126; d) R. Cammack, Nature 1999, 397,
214–215; e) M. W. W. Adams, E. I. Stiefel, Science 1998, 282,
1842–1843; f) M. J. Maroney, G. Davidson, C. B. Allan, J. Fig-
lar, Struct. Bonding (Berlin) 1998, 92, 2–65.
a) S. Foerster, M. Stein, M. Brecht, H. Ogata, Y. Higutchi, W.
Lubitz, J. Am. Chem. Soc. 2003, 125, 83–93; b) H.-J. Fan, M. B.
Hall, J. Am. Chem. Soc. 2002, 124, 394–395, and references
cited therein; c) P. E. M. Siegbahn, M. R. A. Blomberg, M.
Wirstam née Pavlov, R. H. Crabtree, J. Biol. Inorg. Chem. 2001,
6, 460–466, and references therein.
a) M. Y. Darensbourg, E. J. Lyon, J. J. Smee, Coord. Chem. Rev.
2000, 553–561; b) A. C. Marr, D. J. E. Spencer, M. Schröder,
Coord. Chem. Rev. 2001, 219, 1055–1074; c) D. J. Evans, C. J.
Pickett, Chem. Soc. Rev. 2003, 32, 268–275; d) E. Bouwman, J.
Reedijk, Coord. Chem. Rev. 2005, 249, 1555–1581.
a) S. Ciurli, S.-B. Yu, R. H. Holm, K. K. P. Srivastava, E.
Munck, J. Am. Chem. Soc. 1990, 112, 8169–8171; b) W. Saak,
S. Pohl, Angew. Chem. 1991, 103, 869–870; Angew. Chem. Int.
Ed. Engl. 1991, 30, 881–883; c) J. Zhou, M. J. Scott, Z. Hu, G.
Peng, E. Munck, R. H. Holm, J. Am. Chem. Soc. 1992, 114,
10843–10854; d) C. Junghams, W. Saak, S. Pohl, J. Chem. Soc.
Chem. Commun. 1994, 2327–2328; e) R. L. Holliday, L. C.
Roof, B. Hargus, D. M. Smith, P. T. Wood, W. T. Pennington,
J. W. Kolis, Inorg. Chem. 1995, 34, 4392–4401; f) C. H. Lai,
J. H. Reibenspies, M. Y. Darensbourg, Angew. Chem. 1996,
108, 2551–2554; Angew. Chem. Int. Ed. Engl. 1996, 35, 2390–
2393; g) F. Osterloh, W. Saak, S. Pohl, J. Am. Chem. Soc. 1997,
119, 5648–5656; h) E. Bouwman, R. K. Henderson, A. L.
Spek, J. Reedijk, Eur. J. Inorg. Chem. 1999, 4, 217–219; i) S. C.
Davies, D. J. Evans, D. L. Hughes, S. Longhurst, J. R. Sanders,
Chem. Commun. 1999, 1935–1936; j) T. Glaser, T. Beissel, E.
Bill, T. Weyhermüller, V. Schünemann, W. Meyer-Klaucke,
A. X. Trautwein, K. Wieghardt, J. Am. Chem. Soc. 1999, 121,
2193–2208; k) G. Steinfeld, B. Kersting, Chem. Commun. 2000,
205–206; l) M. Razavet, S. C. Davis, D. L. Hughes, C. J. Pickett,
Chem. Commun. 2001, 847–848; m) M.-C. Chalbot, A. M.
Mills, A. L. Spek, G. J. Long, E. Bouwman, Eur. J. Inorg.
Chem. 2003, 453–457; n) M. C. Smith, J. E. Barclay, S. P.
Cramer, S. C. Davies, W.-W. Gu, D. L. Hughes, S. Longhurst,
D. J. Evans, J. Chem. Soc. Dalton Trans. 2002, 2641–3647; o) D.
Sellmann, F. Geipel, F. Lauderbach, F. W. Heinemann, Angew.
Acknowledgments
We gratefully acknowledge the financial support provided by the
Deutsche Forschungsgemeinschaft (DFG), Sonderforschungsbe-
reich (SFB-583) and the Fonds der Chemischen Industrie.
[7]
[1] a) P. M. Vignais, B. Billoud, J. Meyer, FEMS Microbiol. Rev.
2001, 25, 455–501; b) R. K. Thauer, A. R. Klein, G. C. Hart-
mann, Chem. Rev. 1996, 96, 3031–3042; c) S. P. J. Albracht, Bi-
ochim. Biophys. Acta 1994, 1188, 167–204; d) T. Lissolo, S. Pul-
vin, D. Thomas, J. Biol. Chem. 1984, 259, 11725–11729; e)
M. W. W. Adams, L. E. Mortenson, J.-S. Chen, Biochim. Bio-
phys. Acta 1980, 594, 105–176.
[8]
[9]
[2] V. Artero, M. Fontecave, Coord. Chem. Rev. 2005, 249, 1518–
1554.
[3] a) A. Volbeda, M. H. Charon, C. Piras, E. C. Hatchikian, M.
Frey, J. C. Fontecilla-Camps, Nature 1995, 373, 580–587; b) A.
Volbeda, E. Garcin, A. L. de Lacey, V. M. Fernandes, E. C.
Hatchikian, M. Frey, J. C. Fontecilla-Camps, J. Am. Chem.
Soc. 1996, 118, 12989–12996; c) M. Frey, Struct. Bonding
(Berlin) 1998, 90, 97–126; d) Y. Higuchi, H. Ogata, K. Miki,
N. Yasuoka, T. Yagi, Structure 1999, 7, 549–556.
[4] a) J. W. Peters, W. N. Lanzilotta, B. J. Lemon, L. C. Seefeldt,
Science 1998, 282, 1853–1858; b) Y. Nicolet, C. Piras, P. Leg-
rand, C. E. Hatchikian, J. C. Fontecilla-Camps, Structure 1999,
7, 13–23.
[5] a) Z. J. Gu, J. Dong, C. B. Allan, S. B. Choudhury, R. Franco,
J. J. G. Moura, I. Moura, J. LeGall, A. E. Przybyla, W. Rose-
boom, S. P. J. Albracht, M. J. Axley, R. A. Scott, M. J. Ma-
roney, J. Am. Chem. Soc. 1996, 118, 11155–11165; b) A. L.
de Lacey, E. C. Hatchikian, A. Volbeda, M. Frey, J. C. Fontec-
illa-Camps, V. M. Fernandez, J. Am. Chem. Soc. 1997, 119,
7181–7189; c) Y. Montet, P. Amara, A. Volbeda, X. Vernede,
E. C. Hatchikian, M. J. Field, M. Frey, J. C. Fontecilla-Camps,
Nat. Struct. Biol. 1997, 4, 523–526; d) A. K. Jones, S. E. Lamle,
H. R. Pershad, K. A. Vincent, A. P. J. Albracht, F. A. Am-
strong, J. Am. Chem. Soc. 2003, 125, 8505–8524; e) M. Brecht,
M. van Gastel, T. Buhrke, B. Friedrich, W. Lubitz, J. Am.
Chem. Soc. 2003, 125, 13075–13083.
[6] a) The Bioinorganic Chemistry of Nickel (Ed.: J. R. Lancas-
ter Jr), VCH, Weinheim, 1988; b) J. C. Fontecilla-Camps,
3392
www.eurjic.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2007, 3385–3393

Trinuclear [NiFe] Hydrogenase Model Compound
FULL PAPER
Chem. 2002, 114, 654–656; Angew. Chem. Int. Ed. 2002, 41,
632–634; p) M. C. Smith, J. E. Barclay, S. C. Davies, D. L.
Hughes, D. J. Evans, Dalton Trans. 2003, 4147–4151; q) D.
Sellmann, R. Prakash, F. W. Heinemann, M. Moll, M. Kli-
mowicz, Angew. Chem. 2004, 116, 1913–1916; Angew. Chem.
Int. Ed. 2004, 43, 1877–1880; r) P. A. Stenson, A. Martin-Be-
cerra, C. Wilson, A. J. Blake, J. McMaster, M. Schröder, Chem.
Commun. 2006, 317–319.
[24] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098–3100; b) J. P. Per-
dew, Phys. Rev. B 1986, 33, 8822–8824; c) S. H. Vosko, L. Wilk,
M. Nusair, Can. J. Phys. 1980, 58, 1200–1211.
[25] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100,
5829–5835.
[26] K. Eichkorn, F. Weigend, O. Treutler, R. Ahlrichs, Theor.
Chem. Acc. 1997, 97, 119–124.
[27] Gaussian 98, Revision A.11.3, M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G.
Zakrzewski, J. A. Montgomery Jr, R. E. Stratmann, J. C. Bu-
rant, S. Dapprich, J. Millam, A. D. Daniels, K. N. Kudin,
M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R.
Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.
Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma,
N. Rega, P. Salvador, J. J. Dannenberg, D. K. Malick, A. D.
Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V.
Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox,
T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M.
Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Re-
plogle, J. A. Pople, Gaussian, Inc., Pittsburgh PA, 2002.
[28] a) L. Laaksonen, J. Mol. Graphics Modell. 1992, 10, 33–34; b)
D. L. Bergman, L. Laaksonen, A. Laaksonen, J. Mol. Graphics
Modell. 1997, 15, 301–306.
[10] Z. Li, Y. Ohki, K. Tatsumi, J. Am. Chem. Soc. 2005, 127, 8950–
8951.
[11] D. Sellmann, F. Lauderbach, F. Geipel, F. W. Heinemann, M.
Moll, Angew. Chem. 2004, 116, 3203–3206; Angew. Chem. Int.
Ed. 2004, 43, 3141–3144.
[12] D. Sellmann, J. Sutter, Prog. Inorg. Chem. 2003, 52, 585–681.
[13] D. Sellmann, F. Geipel, F. Lauderbach, F. W. Heinemann,
Chem. Eur. J. 2002, 8, 958–966.
[14] R. Bauernschmitt, R. Ahlrichs, J. Chem. Phys. 1996, 104,
9047–9052.
[15] a) R. S. Mulliken, J. Chem. Phys. 1955, 23, 1833–1840; b) R. S.
Mulliken, J. Chem. Phys. 1955, 23, 1841–1846.
[16] A. E. Reed, R. B. Weinstock, F. Weinhold, J. Chem. Phys. 1985,
83, 735–746.
[17] J. Cioslowski, J. Am. Chem. Soc. 1989, 111, 8333–8336.
[18] V. Grass, D. Lexa, J.-M. Savéant, J. Am. Chem. Soc. 1997, 119,
7526–7532.
[19] J. Degani, R. Fochi, Synthesis 1976, 7, 471–472.
[20] A. J. P. Domingos, J. A. S. Howell, B. F. G. Johnson, J. Lewis,
Inorg. Synth. 1990, 28, 52–55.
[21] W. Wolfsberger, H. Schmidbaur, Synth. React., Inorg. Met.-
Org. Chem. 1974, 4, 149–156.
[22] D. Sellmann, R. Prakash, F. Geipel, F. W. Heinemann, Eur. J.
Inorg. Chem. 2002, 2138–2146.
[23] a) R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kölmel, Chem.
Phys. Lett. 1989, 162, 165–169; b) O. Treutler, R. Ahlrichs, J.
Chem. Phys. 1995, 102, 346–354.
[29] a) C. Kind, M. Reiher, J. Neugebauer, B. A. Hess, SNF, Uni-
versität Erlangen-Nürnberg, 1999–2000; b) J. Neugebauer, M.
Reiher, C. Kind, B. A. Hess, J. Comput. Chem. 2002, 23, 895–
910.
[30] P. Coppens, Crystallographic Computing (Copenhagen). 1970,
225–270.
[31] SADABS 2.06, Bruker AXS, Inc., Madison, WI, 2002.
[32] SHELXTL NT 6.12, Bruker AXS, Inc., Madison, WI, 2002.
Received: November 16, 2006
Published Online: May 29, 2007
Eur. J. Inorg. Chem. 2007, 3385–3393
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
3393