Mononuclear, dinuclear, and pentanuclear [{N,S(thiolate)}iron(II)] complexes: Nuclearity control, incorporation of hydroxide bridging ligands, and magnetic behavior

Article abstract of DOI:10.1002/chem.200500156

The mixed N₃S(thiolate) ligand 1-[bis(2-(pyridin-2-yl)ethyl} amino]-2-methylpropane-2-thiol (Py₂SH) was used in the synthesis of four iron(II) complexes: [(Py₂S)FeCl] (1), [(Py₂S)FeBr] (2), [(Py₂S)₄Fe₅^II(μ-OH) ₂]-(BF₄)₄ (3). and [(Py₂S) ₂Fe₂^II(μ-OH)]BF₄ (4). The X-ray structures of 1 and 2 revealed monomeric iron(II)-alkylthiolate complexes with distorted trigonal-bi-pyramidal geometries. The paramagnetic ¹H NMR spectra of 1 and 2 display resonances from δ = -25 ppm to + 100 ppm, consistent with a high-spin iron(II) ion (S=2). Spectral assignments were made on the basis of chemical shift information and T₁ measurements and show the monomeric structures are intact in solution. To provide entry into hydroxide-containing complexes, a novel synthetic method was developed involving strict aprotic conditions and limiting amounts of H₂O. Reaction of Py₂SH with NaH and Fe-(BF₄)₂·6H ₂O under aprotic conditions led to the isolation of the pentanuclear, μ-OH complex 3, which has a novel dimer-of-dimers type structure connected by a central iron atom. Conductivity data on 3 show this structure is retained in CH₂Cl₂. Rational modification of the ligand-to-metal ratio allows control over the nuclearity of the product, yielding the dinuclear complex 4. The X-ray structure of 4 reveals an unprecedented face-sharing, biooctahedral complex with an [S₂O] bridging arrangement. The magnetic properties of 3 and 4 in the range 1.9-300 K were successfully modeled. Dinuclear 4 is antiferromagnetically coupled [J = -18.8(2) cm^-1]. Pentanuclear 3 exhibits ferrimagnetic behavior, with a high-spin ground state of S_T=6, and was best modeled with three different exchange parameters [J= -15.3(2), J′ = -24.7(3), and J″ = -5.36(7) cm^-1]. DFT calculations provided good support for the interpretation of the magnetic properties.

Full text of DOI:10.1002/chem.200500156

DOI: 10.1002/chem.200500156
Mononuclear, Dinuclear, and Pentanuclear [{N,S(thiolate)}Iron(ii)]
Complexes: Nuclearity Control, Incorporation of Hydroxide Bridging
Ligands, and Magnetic Behavior
Divya Krishnamurthy,[a] Amy N. Sarjeant,[a] David P. Goldberg,*[a] Andrea Caneschi,[b]
Federico Totti,[b] Lev N. Zakharov,[c] and Arnold L. Rheingold[c]
Abstract: The mixed N3S(thiolate)
ligand 1-[bis{2-(pyridin-2-yl)ethyl}ami-
no]-2-methylpropane-2-thiol (Py2SH)
was used in the synthesis of four
iron(ii) complexes: [(Py2S)FeCl] (1),
entry into hydroxide-containing com-
plexes, a novel synthetic method was
developed involving strict aprotic con-
ditions and limiting amounts of H2O.
Reaction of Py2SH with NaH and Fe-
(BF4)2·6H2O under aprotic conditions
led to the isolation of the pentanuclear,
m-OH complex 3, which has a novel
dimer-of-dimers type structure connect-
ed by a central iron atom. Conductivity
data on 3 show this structure is re-
tained in CH2Cl2. Rational modifica-
tion of the ligand-to-metal ratio allows
control over the nuclearity of the prod-
uct, yielding the dinuclear complex 4.
The X-ray structure of 4 reveals an un-
precedented face-sharing, biooctahe-
dral complex with an [S2O] bridging ar-
rangement. The magnetic properties of
3 and 4 in the range 1.9–300 K were
successfully modeled. Dinuclear 4 is
antiferromagnetically coupled [J=
ꢀ18.8(2) cmꢀ1]. Pentanuclear 3 exhibits
ferrimagnetic behavior, with a high-
spin ground state of ST =6, and was
best modeled with three different ex-
change parameters [J=ꢀ15.3(2), J’=
ꢀ24.7(3), and J’’=ꢀ5.36(7) cmꢀ1]. DFT
calculations provided good support for
the interpretation of the magnetic
properties.
II
[(Py2S)FeBr] (2), [(Py2S)4Fe5 (m-OH)2]-
II
(BF4)4 (3), and [(Py2S)2Fe2 (m-OH)]BF4
(4). The X-ray structures of 1 and 2 re-
vealed monomeric iron(ii)–alkylthiolate
complexes with distorted trigonal-bi-
pyramidal geometries. The paramag-
1
netic H NMR spectra of 1 and 2 dis-
play resonances from d=ꢀ25 ppm to
+100 ppm, consistent with a high-spin
iron(ii) ion (S=2). Spectral assign-
ments were made on the basis of chem-
ical shift information and T1 measure-
ments and show the monomeric struc-
tures are intact in solution. To provide
Keywords: cluster compounds · en-
zyme models · magnetic properties ·
S ligands · tripodal ligands
Introduction
In recent years there has been significant interest in the syn-
thesis and study of transition-metal thiolate complexes.
Zinc,[1–7] nickel,[8–12] and iron[13–20] thiolate complexes have
been made as models for mononuclear and polynuclear met-
[a] D. Krishnamurthy, Dr. A. N. Sarjeant, Prof. D. P. Goldberg
Department of Chemistry, Johns Hopkins University
3400 N. Charles Street, Baltimore, MD 21218(USA)
Fax : (+1)410-516-8420
ꢀ
alloenzyme active sites with RS M coordination. The al-
kylthiolate functionality is important because of its rele-
vance to cysteinate ligation, but it is also quite challenging
to control from a synthetic perspective. The challenges that
arise include the avoidance of oxidative pathways that gen-
erate disulfide or sulfur-oxygenate species, as well as the
control over the nuclearity of the final metal products. This
latter issue originates from the propensity of alkylthiolate li-
gands to form indiscriminate bridges between metal ions.
Our research group has an ongoing effort in the synthesis
and reactivity of alkylthiolate-containing metal complexes.
E-mail: dpg@jhu.edu
[b] Prof. A. Caneschi, Dr. F. Totti
Dipartimento di Chimica e UdR INSTM di Firenze
Polo Scientifico, Via della Lastruccia 3
50019 Sesto Fiorentino (Italy)
[c] Dr. L. N. Zakharov, Prof. A. L. Rheingold
Department of Chemistry and Biochemistry
University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093 (USA)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
7328
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

FULL PAPER
We are in part motivated by an interest in building models
for metalloenzyme active sites, such as those found in the
iron-containing enzymes peptide deformylase, superoxide
reductase, and nitrile hydratase. Herein we describe the syn-
thesis of four new alkylthiolate–iron(ii) complexes with the
tetradentate ligand 1-[bis{2-(pyridin-2-yl)ethyl}amino]-2-
methylpropane-2-thiol (Py2SH), which bears two pyridine
donors, and one amine and one alkylthiolate donor. This
ligand was previously synthesized for use in biomimetic
copper chemistry.[21,22] Given the relatively easy synthesis of
Py2SH, we were surprised to find that only the copper
chemistry of this ligand had been investigated. Thus we
were motivated to develop its iron chemistry and determine
its geometric and nuclearity preferences for the iron(ii) ion.
Reaction of this ligand with
Results and Discussion
The synthesis of Py2SH followed the method reported by
Ohta and co-workers.[21] The commercially available starting
material 2-vinylpyridine was converted to bis(2-pyridylethyl-
amine) by treatment with NH4Cl,[31] which was then reacted
with neat 1,1-dimethylthiirane at 708C to afford the ligand
Py2SH. It was noticed during the purification of bis(2-pyri-
dylethylamine) that a large amount of primary amine, 2-
pyridylethylamine, had also formed in the NH4Cl reaction.
To improve the overall yield of Py2SH, the 2-pyridylethyla-
mine was recycled and treated with 2-vinylpyridine under
acidic conditions to give an additional amount of bis(2-
pyridylethylamine).[32]
iron(ii) halides gives the mono-
meric high-spin FeII complexes
[(Py2S)FeX] (X=Cl, Br). In
the case of iron, mononuclear
alkylthiolate complexes are
even less common than for
other metals.[14,16,20,23,24] These
complexes exhibit five-coordi-
nate structures in which the
Py2Sꢀ ligand coordinates in a
tetradentate fashion, although the amine donor forms an un-
usually weak interaction with the metal center. Paramagnet-
ic NMR data are presented to characterize the solution-
state structures of these complexes.
Synthesis of [(Py2S)FeCl] (1) and [(Py2S)FeBr] (2): Com-
plexes 1 and 2 were synthesized by generation in situ of the
deprotonated thiolate ligand, Py2Sꢀ with NaOH, followed
by reaction with the appropriate ferrous halide starting ma-
terial (Scheme 1). Complex formation was accompanied
The synthesis of (Py2S)FeII–hydroxide complexes was also
targeted because of our interest in modeling the active site
of peptide deformylase, which contains a coordinated OHꢀ
/H2O ligand at an RS–FeII center. Traditional syntheses of
(L)n M/OH complexes involve self-assembly[25–28] or substitu-
tion[29,30] reactions with excess OHꢀ/H2O in protic solvents.
Here we describe a new strategy involving aprotic condi-
tions (THF) and a nearly stoichiometric amount of water to
give the hydroxide-containing complexes [(Py2S)4Fe5(m-
OH)2](BF4)4 (3) and [(Py2S)2Fe2(m-OH)]BF4 (4). Both of
these complexes display unprecedented structural motifs for
iron(ii). The latter is a dinuclear, face-sharing bioctahedral
complex with an unusual set of single-atom bridging groups
comprising two thiolate S atoms and one hydroxide oxygen
atom. The former complex, which is synthesized under
almost identical conditions to those used for 4, exhibits a
pentanuclear, dimer-of-dimers type structure with a central
Scheme 1.
with the appearance of a bright yellow color. Vapor diffu-
sion of diethyl ether into the filtered reaction mixtures
yielded crystals of 1 and 2 suitable for X-ray diffraction.
Strict anaerobic conditions must be maintained at all times
for these complexes both in solution and in the solid state to
prevent oxidation.
II
iron(ii) ion linking Fe2 (m-OH) units. A surprisingly simple
synthetic method was discovered that allows control over
the formation of either the dimeric or pentameric species.
Given the high-spin nature of the iron(ii) ions in 3 and 4,
their magnetic properties were worthy of investigation.
Magnetic susceptibility studies have revealed antiferromag-
netic coupling for the dimer 4, and ferrimagnetic coupling
for the pentamer 3. Density functional calculations are also
presented in support of the Heisenberg spin Hamiltonian
model for complex 3 to rationalize the ferrimagnetic ST =6
ground state.
X-ray structures of 1 and 2: The molecular structures of
[(Py2S)FeCl] (1) and [(Py2S)FeBr] (2) are shown in Figure 1
and Figure 2, respectively. Selected bond lengths and angles
are given in Table 1. The structures of 1 and 2 show that
Py2Sꢀ is capable of giving discrete, monomeric RS–FeII com-
plexes, with the ligand coordinated in the expected tetraden-
tate fashion. Both complexes exhibit distorted trigonal-bi-
II
ꢀ
ꢀ
pyramidal geometries. The Fe X (X=Br, Cl) and Fe Npy
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7329

D. P. Goldberg et al.
2.345(1) Š for [(TPA)FeII(SC6H2-2,4,6-Me3)](ClO4),[34] an
iron(ii) thiolate complex of the ligand TPA, and somewhat
shorter than the salen complex (Et4N)[Fe(5-NO2salen)(S-p-
II
tol)][35] with Fe SAr=2.386(2) Š. However, the series of
ꢀ
recently syntheszied complexes [L8py2FeII(SR)]+ (R=aryl,
ꢀ
alkyl) exhibit Fe SR bond lengths significantly shorter
(ranging from 2.259(2)–2.323(3) Š) than in 1 and 2, with the
short end occupied by the alkylthiolate (R=cyclohexyl)
complex.[16]
The tertiary amine to iron(ii) distances of 2.3875(13) Š
and 2.3521(16) Š in 1 and 2, respectively, are unusually long
and deserve some comment. These distances are significant-
ly longer than corresponding distances in the six-coordinate,
high-spin
(TPA)–FeII
complexes
[(TPA)Fe(BF)-
(MeOH)]ClO4,[36] [(TPA)Fe(CH3OH)2](BPh4)2,[37] [(TPA)Fe-
Figure 1. ORTEP diagram of [(Py2S)FeCl] showing the 50% probability
thermal ellipsoids. Hydrogen atoms have been omitted for clarity.
(O3SCF3)2],[37] and [(TPA)Fe(Cl)2],[33] which exhibit Fe
Namine distances of 2.175(6)–2.220(7) Š. A more analogous
II
ꢀ
comparison to 1 and 2 would come from a series of five-co-
ꢀ
ordinate TPA iron(ii) complexes, however there is only one
example of a five-coordinate iron(ii) species in which TPA
binds in the typical tripodal mode, [(TPA)Fe(SC6H2-2,4,6-
II
Me3)]ClO4,[34] and this complex has a normal Fe Namine
ꢀ
bond length of 2.250(3) Š. The unusual eight-coordinate
complex [(TPA)2Fe](BPh4)2 is the only TPA complex to ex-
ꢀ
hibit an Fe Namine distance (2.389(3) Š) as long as that of
1,[37] but this elongated distance can be attributed to the
crowded nature of the high-coordinate center.
Perhaps an even better comparison than TPA–FeII com-
plexes for 1 and 2 are complexes of the ethylene-spaced
TEPA (TEPA=tris(2-(2-pyridyl)ethyl)amine), but to our
knowledge there are no examples of structurally character-
ized TEPA–Fe complexes. Out of the eight examples of
transition metal–TEPA complexes in the Cambridge Struc-
tural Database,[38] none of them exhibit an unusually long
Figure 2. ORTEP diagram of [(Py2S)FeBr] showing the 50% probability
thermal ellipsoids. Hydrogen atoms have been omitted for clarity.
M Namine bond.[39] Thus the cause of the significantly weak-
ꢀ
II
ꢀ
ened Namine Fe interaction in 1 and 2 has not been conclu-
Table 1. Selected bond lengths [Š] and angles [8] for 1 and 2.
sively determined. It is possible that the strong donating
ability of RSꢀ combined with the presence of another nega-
tively charged ligand (Clꢀ or Brꢀ) reduces the positive
charge on the iron center and thereby weakens the Namine in-
teraction, while the bonding between the py groups and FeII
remains normal because of the p-acceptor ability of pyri-
dine.
Complex
1 (X=Cl)
2 (X=Br)
ꢀ
Fe N(1)
2.1479(14)
2.3875(13)
2.1720(14)
2.3274(5)
2.4299(5)
105.36(5)
133.35(4)
120.20(4)
89.29(5)
87.47(5)
96.67(4)
90.59(4)
93.458(17)
174.03(4)
2.1639(17)
2.3521(16)
2.1420(17)
2.3212(5)
2.6166(3)
104.59(6)
133.86(5)
121.03(5)
88.74(6)
90.06(6)
90.22(5)
97.98(4)
90.019(15)
171.90(4)
ꢀ
Fe N(2)
ꢀ
Fe N(3)
ꢀ
Fe S(1)
ꢀ
Fe X(1)
N(3)-Fe-N(1)
N(3)-Fe-S(1)
N(1)-Fe-S(1)
N(3)-Fe-N(2)
N(1)-Fe-N(2)
N(3)-Fe-X(1)
N(1)-Fe-X(1)
S(1)-Fe-X(1)
N(2)-Fe-X(1)
The angular parameters given in Table 1 show that both 1
and 2 are best described as distorted trigonal bipyramidal,
in which the Namine and halide ligands lie on the axial posi-
tions (Namine-Fe-X: 174.03(4)8 (1); 171.88(4)8 (2)) and the
two py and thiolate donors form the equatorial plane. Al-
II
ꢀ
though the long Fe Namine distance in 1 and 2 suggests a
weakened bonding interaction, the angles around the iron
center clearly rule out a four-coordinate geometry. Angular
data for five-coordinate complexes can be analyzed by cal-
culation of a t index,[40] with the limiting values of t being
0.0 for square-pyramidal and 1.0 for trigonal-bipyramidal
geometry. The t indices for 1 and 2 are 0.68and 0.63, re-
spectively, showing that both compounds are closer to trigo-
nal-bipyramidal than square-pyramidal.
bond lengths are in the normal range for high-spin iron(ii)
II
complexes,[33] with the average Fe Npy distance of 2.160 Š
ꢀ
for 1 and 2.152 Š for 2. There are only a few examples of
monomeric RS–FeII complexes for comparison. The Fe SR
II
ꢀ
ꢀ
distances for 1 and 2 are close to the Fe SAr distance of
7330
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

[{N,S(thiolate)}Iron(ii)] Complexes
FULL PAPER
1
1H NMR spectroscopy: The H NMR spectra for 1 and 2 are
in 1 appear at 32.5 ppm as a separate peak. An assignable
peak for the remaining methylene protons, H-b, was not
found for either 1 or 2. The g protons in both 1 and 2 are
shifted significantly upfield, appearing at d=ꢀ8.2 and
ꢀ15.6 ppm, respectively. They also have the longest T1
values, consistent with literature values for H-g, although
the T1 measurement for H-g in 2 is an approximation be-
cause this peak lies on top of one of the gem-dimethyl
peaks, H-d (the upfield shoulder at d=ꢀ16 ppm). Distinct
T1 values for H-g and H-d in 2 could not be obtained by the
inversion-recovery method because of the overlap of these
resonances, but the measured T1 (17.7 ms) at d=ꢀ15.6 ppm
is close to an average of the T1ꢁs for the analogous H-g
(39.2 ms) and H-d (0.92 ms) peaks for 1. The remaining up-
field-shifted peaks for 1 and 2 are assigned to the gem-di-
methyl protons. For 1, the gem-dimethyl groups appear at
d=ꢀ19.0 and ꢀ24.5 ppm. For 2, one of the gem-dimethyl
groups gives rise to the shoulder at d=ꢀ16 ppm as men-
tioned above, while the other gem-dimethyl peak is a dis-
tinct, broad peak at d=ꢀ23.0 ppm.
shown in Figure 3 and the data are summarized in Table 2.
Both complexes exhibit paramagnetically shifted peaks be-
tween 100 to ꢀ25 ppm, which are typical for high-spin
In solution, 1 and 2 may be expected to contain a mirror
ꢀ
plane containing the Fe S vector and bisecting the Npy-Fe-
Npy angle, and the single set of pyridine resonances (H-a, H-
b1,b2, H-g) is consistent with this symmetry. However, the
inequivalence of the gem-dimethyl groups suggests that this
symmetry is broken by the -CH2C(CH3)2Sꢀ arm. In the solid
state (Figures 1 and 2), the gem-dimethyl groups are clearly
inequivalent due to the conformation of the thiolate arm,
and thus there may be conformational restriction of the
-CH2C(CH3)2Sꢀ group in solution, causing the separate
peaks for H-d in the NMR spectra.[41]
Figure 3. 1H NMR spectra of a) 1 and b) 2 in CD2Cl2 at 25.08C. The
peaks in the diamagnetic region (d=0–10 ppm) are due to trace solvent
and trace amounts of free ligand.
Table 2. 1H NMR chemical shifts and T1 values for 1 and 2 in CD2Cl2 at
258C.
1
2
The NMR assignments are based partly on comparison
with the literature values for five-coordinate[33,34] and six-co-
ordinate[33,36] monomeric TPA- and substituted TPA–FeII
complexes. All of these complexes exhibit pyridine chemical
shifts and 1/T1 values in the order H-a > H-b1,b2 > H-g,
and the assignments for 1 and 2 were based on this prece-
dent. The chemical shifts for H-a in TPA–FeII species typi-
cally fall between d=+110 to +140 ppm, with the upfield
end of this range made up by the five-coordinate [(TPA)Fe-
(SAr)]+ complexes reported by Zang and Que.[34] The rela-
tively upfield-shifted H-a peaks for five-coordinate 1 and 2
match the trend for these species. The H-b1,b2 chemical
shifts and T1 values for 1 and 2 also fall in the same range as
H-b1,b2 for other TPA–FeII species, which are found be-
tween d=23–65 ppm and exhibit T1ꢁs of several millisec-
onds. The peak for H-g in 1 and 2 appears farther upfield
than in most TPA complexes, with the exception being the
[(TPA)Fe(SAr)]+ complexes (Hg ~ꢀ10 ppm), which are
also the closest structural analogues to 1 and 2. There is no
direct precedent to compare the signals for the gem-dimeth-
yl protons (H-d). The significant upfield shift of the H-d res-
onances may be explained by a spin delocalization mecha-
nism, as opposed to a Fermi contact shift mechanism.[42]
Proton
d [ppm]
+90.9
T1 [ms]
d [ppm]
T1 [ms]
a
1.8
13.4, 25.0
39.2
3.7
–
0.92, 1.87
+95.3
+41.5, +32.5
ꢀ15.6
1.7
14.8, 10.1[a]
17.7[a]
b1,b2
g
+42.9, +33.9
ꢀ8.2
a,c
b
d
+32.5
+32.5
10.1[a]
not assigned
ꢀ19.0, ꢀ24.5
not assigned
ꢀ15.6, ꢀ23
–
17.7,[a] 0.98
[a] An “average” T1 value for overlapping peaks (see text for discussion).
iron(ii) complexes. Peak assignments have been made based
on chemical shifts, T1 measurements, and comparison with
the literature. The spectra shown in Figure 3 are highly re-
producible from batch to batch of 1 and 2, and provide a
good fingerprint for these complexes in solution. The far-
thest downfield shifted peaks at d=90.9 and 95.3 ppm for 1
and 2 are assigned to the a-protons, and have the shortest
T1 values among the pyridine protons, as expected based on
proximity to the paramagnetic center. The peaks at d=42.9
and 41.5 ppm for 1 and 2, respectively, are assigned to one
of the H-b1,b2 protons. The other set of H-b1,b2 protons for 1
are assigned to the peak at d=33.9 ppm. Both H-b1,b2 pro-
tons in 1 exhibit T1 values of similar magnitude (d=13.4
and 25.0 ms). In the case of the bromide complex, the other
set of the H-b1,b2 protons is assigned to the relatively broad
resonance at d=32.5 ppm, which overlaps with the methyl-
ene signals from H-a and H-c. In contrast, the H-a,c protons
Incorporation of OHꢀ: synthesis of [(Py2S)4FeII (m-OH)2]-
5
(BF4)4 (3): Complex 3 was synthesized by the deprotonation
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7331

D. P. Goldberg et al.
of Py2SH in dry, freshly distil-
led THF with sodium hydride,
followed by reaction with a
stoichiometric amount of Fe-
(BF4)2·6H2O under inert at-
mosphere (Scheme 2). Recrys-
tallization from CH2Cl2/pen-
tane gave X-ray quality crys-
tals, and a determination of
the X-ray structure of 3 re-
vealed a pentanuclear complex
with two dimeric, (Py2S)2Fe2(m-
OH) units organized around a
central FeII(m-SR)4 core (vide
infra).
Typical methods of synthe-
ꢀ
sizing (L)nM OH species in-
clude substitution reactions,
such as that between [HB-
(pztBu,Me)3ZnI] and OHꢀ to
give [HB(pztBu,Me)3ZnOH],[28]
and
self-assembly
reac-
tions[25–27,43,44] in protic (e.g.
MeOH) solvent. All of these
strategies failed for the Py2Sꢀ
Scheme 2.
system. For example, reaction of Py2SH with Fe(ClO4)2 or
Fe(BF4)2 in the presence of NaOH, KOH or Me4NOH in
MeOH did not lead to any tractable products. Attempts also
were made to substitute the halide ligand of 1 and 2 with
OHꢀ, but the only product was a dark green precipitate
which was insoluble in most common organic solvents (e.g.
CH3CN, CH2Cl2 and MeOH) and could not be characterized
further. These failures prompted us to attempt conditions in
which the amount of H2O/OHꢀ could be better controlled.
Aprotic conditions (THF) combined with a limiting amount
of H2O from the Fe(BF4)2·6H2O starting material led to the
successful synthesis of the hydroxide species 3, and the OHꢀ
ligand in 3 occupies the open site of the (Py2S)FeII unit, as
designed. These results suggest that the use of aprotic sol-
vent combined with nearly stoichiometric amounts of H2O
Figure 4. ORTEP diagram of the cation of [(Py2S)4Fe5(m-OH)2](BF4)4 (3)
showing the 30% probability thermal ellipsoids. Hydrogen atoms and sol-
vent molecules have been omitted for clarity.
can be a good approach for the preparation of other
II
ꢀ
(L)nM OH compounds.
Molecular structure of [(Py2S)4Fe5(OH)2](BF4)4 (3): The
molecular structure of the cation in 3 is shown in Figure 4,
and selected bond lengths and angles are given in Table 3.
The hydroxide bridge is located trans to the tertiary amine
group for both Py2Sꢀ ligands, and the Namine-Fe-OH vector
makes up the axial direction of a distorted trigonal-bipyra-
midal geometry for both Fe2 and Fe3.
II
II
ꢀ
ꢀ
This molecule can be viewed as two Fe (m-OH) Fe com-
plexes joined together by a central iron(ii) ion (Fe1). The
Fe1 ion is located on a crystallographically imposed twofold
ꢀ
The average Fe Npyridine bond length of 2.11 Š in 3 is
ꢀ
shorter than the Fe Npyridine distances of the five-coordinate
axis, and thus there is only one unique dimeric FeII (m-OH)
complexes 1 and 2. Similarly, d(Fe Namine)av =2.29(1) Š in 3
ꢀ
2
unit. Each iron center in the dimeric unit is five-coordinate
is significantly smaller than the corresponding distance in 1
or 2. Thus the unusually weak interaction between the Namine
and FeII seen in 1 and 2 is not observed in 3. The RSꢀ and
OHꢀ ligands in 3 both share their negative charge with two
iron(ii) ions, whereas in 1 and 2 the thiolate and Xꢀ (X=Cl,
Br) donors provide a full negative charge to a single iron(ii)
center, and therefore the resultant higher positive charge on
and bridged by a single hydroxide ligand. The bridging atom
II
ꢀ
is assigned to a hydroxide group based on Fe O bond
lengths, charge balance considerations, and magnetic sus-
ceptibility data (vide infra). The chelating thiolate arms
form four bridges to the central iron(ii) ion, resulting in a
distorted tetrahedral Fe(SR)4 coordination sphere for Fe1.
7332
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

[{N,S(thiolate)}Iron(ii)] Complexes
FULL PAPER
Table 3. Selected bond lengths [Š] and angles [8] for 3·2CH2Cl2·C5H12.
pyridine and thiolate donors on each iron, and these devia-
tions are reflected in the t values, which are 0.63 for Fe2
and 0.58for Fe3. Considerable distortion from tetrahedral
geometry is also observed for Fe1, with angles ranging from
121.36(6)8 to 94.18(3)8.
ꢀ
Fe1 S1
2.3134(11)
2.3182(10)
2.0140(18)
2.091(3)
2.296(3)
2.118(3)
2.3220(11)
2.0211(18)
2.126(3)
2.295(3)
2.106(3)
2.3375(11)
3.905
116.79(7)
116.13(4)
94.18(3)
121.36(6)
92.21(10)
86.48(10)
116.79(7)
116.13(4)
94.18(3)
O1-Fe2-N1
O1-Fe2-N3
N1-Fe2-N3
N3-Fe2-N2
O1-Fe2-N2
N1-Fe2-N2
O1-Fe2-S1
N1-Fe2-S1
N3-Fe2-S1
N2-Fe2-S1
O1-Fe3-N6
O1-Fe3-N4
O1-Fe3-N5
N6-Fe3-N4
N6-Fe3-N5
N4-Fe3-N5
O1-Fe3-S2
N6-Fe3-S2
N4-Fe3-S2
N5-Fe3-S2
Fe2-O1-Fe3
Fe1-S1-Fe2
Fe1-S2-Fe3
92.21(10)
86.48(10)
114.10(12)
89.89(11)
175.27(10)
92.04(12)
97.71(6)
108.48(9)
137.02(10)
82.90(8)
92.30(10)
86.75(10)
174.86(10)
110.43(12)
92.51(11)
89.87(12)
96.20(6)
109.24(9)
140.06(9)
83.92(8)
ꢀ
Fe1 S2
ꢀ
Fe2 O1
ꢀ
Fe2 N1
ꢀ
Fe2 N2
ꢀ
Fe2 N3
ꢀ
Fe2 S1
Solution-state structure of [(Py2S)4Fe5(OH)2](BF4)4 (3):
Conductivity: The X-ray structure of 3 shows a [(Py2S)4Fe5-
ꢀ
Fe3 O1
ꢀ
Fe3 N4
ꢀ
(m-OH)2]4+ ion with four BF4 counterions, and therefore it
ꢀ
Fe3 N5
should behave as a 1:4 electrolyte in solution if the solid-
state structure remains intact. Electrical conductivity meas-
urements of 3 were carried out in CH2Cl2 to characterize
the nature of the complex in solution.[51–53] Conductivity
measurements are often done in more polar solvents such as
H2O, CH3CN and CH3OH, especially with 1:4 electro-
lytes,[54,55] but these solvents were not used because they
could potentially coordinate to the iron(ii) ions in 3 and dis-
rupt the structure of the Fe5 cluster. Since 3 was crystallized
from CH2Cl2, it was the solvent of choice for conductivity
measurements. Figure S1 in the Supporting Information
shows the Onsager plots of compound 3, and of the refer-
ence compound [nBu4N]PF6, which was used as a standard
1:1 electrolyte. The slope for 3 (slope=2913) is close to four
times that of the reference, [nBu4N]PF6 (slope=627), sug-
gesting that 3 behaves as a 1:4 electrolyte in CH2Cl2, and
that the [(Py2S)4Fe5(m-OH)2]4+ unit remains intact in solu-
tion. These results are also consistent with the 1H NMR
spectrum for 3 (see Experimental Section), which shows a
complex pattern of paramagnetically shifted resonances,
quite distinct from the relatively symmetric NMR spectra
for 1 and 2.
ꢀ
Fe3 N6
ꢀ
Fe3 S2
ꢀ
Fe2 Fe3
S1-Fe1-S1
S1-Fe1-S2
S1-Fe1-S2
S2-Fe1-S2
O1-Fe2-N1
O1-Fe2-N3
S1-Fe1-S1
S1-Fe1-S2
S1-Fe1-S2
S2-Fe1-S2
150.8(1)
137.74(5)
135.39(5)
121.36(6)
ꢀ
Fe2 and Fe3 in 3 may be the cause of the shorter N Fe
bond lengths. The same shortening is not observed in the
ꢀ
Fe SR distances for 3, falling between 2.3134(11)–
2.3375(11) Š, but the bridging mode of the thiolate donors
ꢀ
is expected to lengthen the Fe SR distances.
II
II
ꢀ
ꢀ
The Fe (m-OH) bonds in 3 are on the short end of Fe
(m-OH) distances found in the literature, although these
values are mainly for six-coordinate FeII complexes.[45,46] To
our knowledge, there are only two complexes with five-coor-
dinate iron(ii) ions bridged by
a hydroxide ligand:
II
[FeII(HB(3,5-iPr2pz)3)]2(OH)2,
a
bis(m-hydroxo) complex,
Control of nuclearity: Synthesis of [(Py2S)2Fe2 (m-OH)]-
and [FeII(HB(3,5-iPr2pz)3)]2(OH)(OBz), a m-hydroxo, m-car-
(BF4) ( 4:) Inspection of the structure of the pentanuclear
boxylato complex.[30] These complexes have Fe (m-OH) dis-
complex 3 suggested that removal of the central FeII ion
ꢀ
II
tances of 1.964–2.04 Š, and complex 3 fits well within this
should lead to the isolation of the dinuclear (Py2S)2Fe2 (m-
ꢀ
range. In contrast, the Fe Fe distance of 3.905 Š in 3 is
OH) unit. Given that the overall metal/ligand ratio in 3 is
5:4, we speculated that limiting the FeII to substoichiometric
quantities might lead to the synthesis of the dinuclear
much longer than that in either [FeII(HB(3,5-iPr2-
[FeII(HB(3,5-
ꢀ
pz)3)]2(OH)2
(Fe Fe
3.179(5) Š)
or
II
ꢀ
iPr2pz)3)]2(OH)(OBz) (Fe Fe 3.681(3) Š), and can be corre-
lated with the obtuse Fe-OH-Fe angle 150.8(1)8. Such a
wide angle is found in (porphyrinoid)M-(m-OH)-M(porphyr-
inoid) structures,[47,48] which have only one bridging ligand
connecting the metal centers, as opposed to 3, which has the
S1-Fe1-S2 unit as a second bridge between Fe2 and Fe3.
[(Py2S)2Fe2 (m-OH)] complex. When the reaction was run
with a Py2SH/NaH/Fe(BF4)2·6H2O ratio of 1/1/0.75 in THF,
a yellow precipitate formed immediately as in the case of 3.
However, vapor diffusion of pentane into a solution of the
crude material in CH2Cl2 led to the crystallization of a dif-
ferent product. Analysis of this compound by X-ray diffrac-
tion revealed complex 4, which has the intended hydroxide-
bridged, dinuclear structure, although the thiolate ligands
that previously were bound to the central FeII ion in 3 are
now bridged to the opposite iron(ii) ion in the dimer
(Scheme 2). Crystals of 3 and 4 are easily distinguishable by
eye due to differences in their morphology (needles versus
blocks).
ꢀ
The long Fe S distances likely allows for the unusually
obtuse Fe2-(m-OH)-Fe3 angle. The related [FeII(HB(3,5-
iPr2pz)3)]2(OH)(OBz) has a tighter, more rigid second
bridge in the form of OBz, and consequently has a smaller
FeII-(m-OH)-FeII angle of 137.0(3)8, in agreement with the
m-hydroxo, m-carboxylato motif.[49,50]
The angles around Fe2 and Fe3 reveal distorted trigonal-
bipyramidal geometries for both metal ions. The angle be-
tween the Namine donor and the hydroxide bridge for each
iron is 175.278 and 174.868, consistent with the axial direc-
tion of a trigonal bipyramid. However, significant deviations
from 1208 occurs in the equatorial plane formed by the two
Molecular structure of [(Py2S)2Fe2(m-OH)](BF4) ( 4:) An
ORTEP diagram of the dinuclear complex 4 is shown in
Figure 5, and selected bond lengths and angles are given in
Table 4. The structure is best described as a face-sharing bio-
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7333

D. P. Goldberg et al.
effects. All but one of the Fe-S distances are in the range
2.44–2.51 Š and are similar to other bridging thiolate distan-
ces in related FeII (m-SR)2 complexes.[59] The Fe2 S4 dis-
tance is significantly longer at 2.67(1) Š, and although the
ꢀ
2
reason for this long distance is not clear, it may be a conse-
quence of steric strain.
ꢀ
An Fe Fe distance of 2.929(1) Š is observed for 4, and is
shorter than other high-spin FeII face-sharing bioctahedra,
2
such as [(tacn)FeII(m-Cl)3FeII(tacn)]+ (Fe Fe 3.014(2) Š;
ꢀ
tacn= 1,4,7-triazacyclononane) prepared by Wieghardt
et al.[60] or [(thf)FeII(m-Cl)3FeII(thf)]+ (Fe Fe 3.086(2) Š)
ꢀ
prepared by Sobota et al.[61] However, it is significantly
longer than in related metal–metal bonded species such as
FeII (S2C3H6)(CNMe)7]+ (Fe Fe 2.634(2) Š).[62] The relative-
ꢀ
2
ꢀ
ly long Fe Fe distance, combined with the paramagnetic be-
havior of 2 (vide infra), clearly rules out the presence of a
metal–metal bond in 4.
Figure 5. ORTEP diagram of the cation of [(Py2S)2Fe2(m-OH)](BF4) (4)
showing the 30% probability thermal ellipsoids. Hydrogen atoms and sol-
vent molecules have been omitted for clarity.
It is tempting to suggest complex 4 is formed by the
fusion of two distinct monomeric species based on the asym-
metry observed in the X-ray structure. One species could be
a terminal hydroxide complex containing Fe2, in which the
OH ligand is coordinated trans to the tertiary amine as
found for the other Xꢀ ligands (X=Cl, Br, OH) in 1–3,
while the other species could be a solvento complex of Fe1
with THF coordinated in the same site. Upon bringing the
two monomers together, the weakly bound THF on Fe1 is
displaced by the thiolate donor S(2), and the terminal hy-
droxide ligand makes a weak bridging interaction with Fe1.
The third bridge is formed between the remaining thiolate
ligand S4 and Fe2.
The synthesis and structural characterization of 3 and 4
provides an example of how subtle changes in a reaction
system can influence the nuclearity and arrangement of
transition metal complexes. Synthetic control over nuclearity
is usually difficult to achieve, yet it is important in order to
build molecules of defined metal ion content, spatial ar-
rangement, and physical properties. It is especially impor-
tant for paramagnetic metal ions such as high-spin iron(ii),
because of their application in the field of molecular mag-
netism, for which the rational synthesis of polynuclear com-
plexes with new topologies and interesting magnetic proper-
ties is a central theme.[63–65] The magnetic properties of both
3 and 4 are presented below.
Table 4. Selected bond lengths [Š] and angles [8] for 4.
ꢀ
Fe1 O1
2.1613(16)
2.5074(9)
1.9842(16)
2.290(2)
2.188(2)
2.256(2)
2.4406(9)
2.6708(9)
2.256(2)
2.155(2)
2.249(2)
2.5078(8)
2.9288
89.80(6)
166.53(7)
90.82(7)
89.26(8)
78.97(7)
87.58(8)
92.29(8)
83.54(5)
109.829(6)
82.69(6)
O1-Fe1-S3
N1-Fe1-S4
N2-Fe1-S3
N3-Fe1-S3
N1-Fe1-S3
S4-Fe1-S3
O1-Fe2-N5
O1-Fe2-N6
N6-Fe2-N5
N4-Fe2-N5
O1-Fe2-N4
N6-Fe2-N4
O1-Fe2-S3
N5-Fe2-S3
N6-Fe2-S3
N4-Fe2-S3
O1-Fe2-S4
N5-Fe2-S4
N6-Fe2-S4
N4-Fe2-S4
S4-Fe2-S3
Fe1-S3-Fe2
Fe1-S4-Fe2
83.84(5)
161.73(6)
97.38(6)
171.78(6)
92.81(6)
90.48(3)
101.72(8)
168.73 (7)
89.33 (8)
86.89(8)
90.54 (8)
92.34 (8)
87.52 (5)
169.68 (6)
81.66(6)
88.46(6)
81.15(5)
100.43(7)
94.74(6)
169.86(6)
85.39(3)
71.46(2)
69.65(2)
ꢀ
Fe1 S3
ꢀ
Fe2 O1
ꢀ
Fe1 N1
ꢀ
Fe1 N2
ꢀ
Fe1 N3
ꢀ
Fe1 S4
ꢀ
Fe2 S4
ꢀ
Fe2 N4
ꢀ
Fe2 N5
ꢀ
Fe2 N6
ꢀ
Fe2 S3
ꢀ
Fe1 Fe2
Fe2-O1-Fe1
O1-Fe1-N2
O1-Fe1-N3
N2-Fe1-N3
O1-Fe1-N1
N2-Fe1-N1
N1-Fe1-N3
O1-Fe1-S4
N2-Fe1-S4
N3-Fe1-S4
octahedral complex, in which the two iron centers are
joined by two bridging thiolate groups and one bridging hy-
droxide ligand. Face-sharing biooctahedral complexes of the
type L3M(m-X)3ML3 form a large and well-studied group of
molecules.[56–58] However, to our knowledge there are no ex-
amples of complexes in which two iron atoms share one O
Magnetic properties of [(Py2S)2Fe2(m-OH)]BF4 (4): The
magnetic susceptibility data for 4 are presented in Figure 6
as a cT versus T plot. At room temperature, the product of
cT is 5.85 emuKmolꢀ1, which is approaching the spin-only
value for two independent high-spin FeII (d6) S=2 ions. As
the temperature is decreased, there is a steady decrease in
cT down to the final value of 0.043 emuKmolꢀ1 at 1.9 K.
This behavior is consistent with intramolecular antiferro-
magnetic coupling between the two high-spin FeII ions, re-
sulting in a non-magnetic ground state.
and two S bridging atoms. The Fe2(RS)2(O) core also reveals
II
ꢀ
a striking asymmetry in the Fe OH bond lengths. The
ꢀ
Fe1 OH distance is 2.161(2) Š, and is one of the longest
II
ꢀ
ꢀ
Fe OH bonds reported, while the Fe2 OH distance of
1.984(2) Š is one of the shortest.[45] The long Fe OH bond
II
ꢀ
ꢀ
in 4 is trans to the short pyridine bond Fe1 N2 (2.188(2) Š),
and the short Fe OH interaction is trans to the longer ter-
tiary amine bond Fe2 N6 (2.249(2) Š). Thus the asymmetry
in the Fe (m-OH)-Fe unit may be a consequence of trans
II
ꢀ
The experimental data in Figure 6 was fit to the model
shown in Equation (1), which was derived from the spin
Hamiltonian for two interacting S=2 ions (H=ꢀ2 JS1S2).
ꢀ
II
II
ꢀ
7334
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

[{N,S(thiolate)}Iron(ii)] Complexes
FULL PAPER
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢀ
ꢁ
ꢁ
ꢀ
4 should be more covalent than the Fe
J
kT
3J
kT
6J
kT
10J
kT
ꢀ
ꢁ
2exp
þ 10exp
þ 28exp
þ 60exp
Ng2b2
3k
(m-OH) interactions. The antiferromag-
netic coupling in the mixed-valent dimer
[(Fe2(OH)3(tmtacn)]+ (tmtacn=trime-
thyltriazacyclononane) is weaker than in
the mixed-valent dimers of the type
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
cT ¼
ð1Þ
J
3J
kT
6J
kT
10J
kT
1 þ 3exp
þ 5exp
þ 7exp
þ 9exp
kT
[Fe2S2]+, and this phenomenon has been explained in part
by the greater covalency of the S2ꢀ bridges.[66]
All symbols have the usual meaning. A least-squares fit of
the data in Figure 6 using Equation (1) gives J=
ꢀ18.8(2) cmꢀ1, g=2.187(5), and R2 =0.99938. As seen in
Figure 6, the fit is quite good and it is not necessary to in-
Magnetic properties of [(Py2S)4Fe5(OH)2](BF4)4 (3): The
magnetic susceptibility for 3 was recorded over a tempera-
ture range of 2–300 K. The temperature dependence of the
cT product for complex 3 is shown in Figure 7. The high-
temperature value of cT is below the value for five noninter-
acting iron(ii) ions (15.0 emuKmolꢀ1 for g=2.00), and a
minimum in cT is seen near 250 K, which are indications of
a ferrimagnetic species. A monotonic increase in cT is ob-
served upon decreasing the temperature, leading to a maxi-
mum of 21.24 emuKmolꢀ1 at 19 K, which is close to the
value expected for a system with ST =6 (22.5 emuKmolꢀ1).
Below 19 K, cT rapidly decreases to 5.5 emuKmolꢀ1 at 2 K.
This decrease at low temperatures may be due to zero-field
splitting contributions, intercluster magnetic interactions, or
saturation effects. The gross behavior of the curve suggests
an antiferromagnetic arrangement of iron(ii) ions with a fer-
rimagnetic ground state of ST =6.
Figure 6. Plot of cT versus T for complex 4. The solid line is the best fit
obtained from the analytical expression described in the text.
To confirm this qualitative interpretation, a least-squares
fit of the data in Figure 7 was carried out, providing values
for the different exchange parameters. As a first approxima-
tion, the degeneracy and spin-orbit contributions arising
clude a term accounting for paramagnetic impurities. Al-
though there are few examples of high-spin iron(ii) face-
sharing, biooctahedral complexes, the magnetic properties
of the complexes from Wieghardt et al. ([(tacn)FeII(m-
Cl)3FeII(tacn)]+) and Sobota et al. ([(thf)FeII(m-Cl)3FeII-
(thf)]+) have been studied in detail.[60,61] The [(thf)FeII(m-
Cl)3FeII(thf)]+ complex was found to be weakly antiferro-
magnetically coupled, with J=ꢀ4.6 cmꢀ1. In contrast, the
[(tacn)FeII(m-Cl)3FeII(tacn)]+ complex was determined to be
weakly ferromagnetically coupled, with J=+11.6 cmꢀ1. Fer-
romagnetic coupling is unusual in these types of dimers, and
this molecule appears to be an anomaly in terms of its mag-
netic behavior. It should be noted that best fits for both of
these complexes included contributions from zero field split-
ting (1.0<jDj<5.5 cmꢀ1).[60] It was not necessary to include
zero-field splitting (zfs) terms for 4 in order to obtain an ex-
cellent fit (Figure 6), and given the larger J value for 4 it
seems reasonable to ignore the small zfs terms.
5
from the D ground state of a high-spin ferrous ion were ne-
glected, and a single g value for all of the iron ions was em-
ployed. Taking into account the twofold symmetry of the
cluster, three different exchange pathways can be individu-
ated: two different pathways are mediated by a sulphur
bridge (J and J’) and one pathway is mediated by a hydroxo
bridge (J’’). The spin Hamiltonian shown in Equation (2)
was used, with negative J indicating antiferromagnetism
and positive J indicating ferromagnetism. Least-squares fits
The antiferromagnetic coupling in 4 is quite similar to
that in the FeII (m-SR)2 complex [(Lc)2FeII (NO3)2] (HLc =
2
2
2-[[2-(2-pyridyl)ethyl]amino]ethanethiol)), which has J=
ꢀ20.04(4) cmꢀ1, g=2.137(2).[59] The only superexchange
pathways for the latter complex come from the bridging al-
kylthiolate groups, and the similarity of the coupling in this
complex and that found for 4 suggest that the alkylthiolate
bridges in 4 provide the dominant exchange pathways,
making the hydroxide bridge superfluous with regards to
mediating the antiferromagnetic coupling. This analysis is
Figure 7. Plot of cT versus T for complex 3. The solid line is the best fit
ꢀ
obtained from a full-matrix diagonalization.
reasonable in light of the fact that the Fe (m-SR) bonding in
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7335

D. P. Goldberg et al.
of the data were obtained by using a full-matrix diagonaliza-
tion of the model given in Equation (2).[67]
H ¼ ꢀ2JðS1S2 þ S1S2AÞꢀ2J0ðS1S3 þ S1S3AÞꢀ2J00ðS2S3 þ S2AS3A
ð2Þ
Þ
Fits of the data in Figure 7 invariably gave negative J
values for all three exchange interactions, indicating antifer-
romagnetic interactions between all of the iron(ii) ions.
Stronger antiferromagnetic exchange was found for the two
sulfur-mediated pathways as compared to the hydroxo-
bridged pathway. Different fits of the data between 19 K
and 300 K of almost equivalent quality could be obtained by
changing the starting values of the parameters, resulting in
significantly different final J, J’, J’’ values. Nearly equivalent
fits were found with different J/J’ ratios, ranging from fit 1:
J/J’=1, g=2.10(2), J=J’=ꢀ20(1) cmꢀ1, J’’=ꢀ6.7(3) cmꢀ1
with R=2.05”10ꢀ4, to fit 2: J/J’=1.6, g=2.078(2), J=
ꢀ15.3(2) cmꢀ1, J’=ꢀ24.7(3) cmꢀ1, J’’=ꢀ5.36(7) cmꢀ1 with
2
obsꢀccalcd
R=1.77”10ꢀ4 (R=ꢀðc
Þ ). For all initial J/J’ ratios the
2
cobs
value of J’’ was consistently determined to be about
ꢀ6 cmꢀ1.
Figure 8. The model geometry of 3 used for density functional calcula-
Thus this fitting procedure did not allow us to distinguish
between a model in which the two sulfur bridges mediate
nearly equivalent superexchange interactions versus a model
in which they show quite different coupling efficiencies. The
metrical parameters for each sulfur bridge show some small
tions.
2
ˆ
Table 5. Determinants, SCF energies, and expectation values of hS i, ex-
pected and computed (in parenthesis), for compound 3. Only determi-
nants with positive Ms were reported.
ꢀ
ꢀ
ꢀ
differences, with Fe1 S1 2.313(1), Fe1 S2 2.318(1), Fe2 S1
2
ꢀ
ꢀ
ꢀ
ˆ
2.322(1), Fe3 S2 2.337(1), Fe1 Fe2 4.324, Fe1 Fe3 4.307 Š
and Fe1-S1-Fe2 137.7 and Fe1-S2-Fe3 135.58. These small
differences allow for the possibility that these pathways
could exhibit substantially different exchange coupling.
To gain further insight into this problem, quantum me-
chanical calculations were performed based on the density
functional theory (DFT) in the B3LYP-broken symme-
try[68,69] framework. A Gaussian type Ahlrichs TVZ basis set
was used for the iron atom while Ahlrichs VDZ basis sets
were used for the S, N, O, and H atoms. The NWChem 4.6
package[70,71] was used. The crystallographically imposed
twofold symmetry of 3 allowed the simplified model shown
in Figure 8to be used for all of the calculations, in which
symmetry-equivalent atoms were omitted. The gem-dimeth-
yl groups of the two unique Py2Sꢀ ligands were also omitted
in order to simplify the calculations. Hydrogen atoms were
used in place of the two symmetry-related iron atoms to sat-
isfy the sulfur valences. The following spin Hamiltonian was
used [Eq. (3)]:
State/Ms
hS i
SCF energy [a.u.]
HS:
j222i; Ms =6
40.00 (40.09)
10.00 (10.06)
10.00 (10.05)
10.00 (10.04)
ꢀ7030.093412
ꢀ7030.096229
ꢀ7030.096162
ꢀ7030.096733
¯
BS1:
BS2:
BS3:
j222i; Ms =2
¯
j222i; Ms =2
¯
j222i; Ms =2
ꢀ22.3 cmꢀ1. All of the computed J values are antiferromag-
netic, in agreement with the experimental findings. The cal-
culated J and J’ values are in good agreement with the
values obtained from the experimental fit 2, which gave in-
equivalent exchange interactions for the two sulfur bridges.
However, the calculated value of J’’ is significantly larger
than that found through fitting the experimental data. De-
spite this discrepancy, the computed J values suggest that
the best fit of the experimental data includes inequivalent J
and J’ parameters.
In light of the DFT results, a spin-Hamiltonian analysis of
the magnetic structure was done using the set of parameters
from fit 2. An ST =6, j446> (where jSASA’ST >, with SA =
S2 +S3 and SA’ =S2’ + S3’) was found to be the ground state.
The first excited state is 28.6 cmꢀ1 above the ground state.
The picture of the magnetic structure of 3 that emerges
from these calculations is shown in Figure 9. The picture
shows that Fe(2) and Fe(3) (and their symmetry related
ions) are ferromagnetically polarized by the stronger J and
J’ exchange pathways, in spite of the presence of intrinsic
antiferromagnetism in the magnetic exchange parameter J’’.
H ¼ ꢀ2JðS1S2Þꢀ2J0ðS1S3ÞꢀJ00ðS2S3Þ
ð3Þ
We computed three broken symmetry determinants[72] in
addition to the high spin case (see Table 5) to obtain the
three exchange parameters J, J’, and J’’. This calculation
leads to a determinate system of three equations with three
variables, and the solution of such a system gives the follow-
ing values: J=ꢀ15.4 cmꢀ1, J’=ꢀ23.2 cmꢀ1, and J’’=
7336
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

[{N,S(thiolate)}Iron(ii)] Complexes
FULL PAPER
The variation of eigenvalues as a function of 1’=Jav/J’’
where Jav =(J + J’)/2, with fixed J/J’=1, was also calculated
(Figure 10) to check how much the discrepancy between J
and J’ influences the magnetic structure. In this case, a high
spin ground state (ST =6) is retained from 1’=4 down to
1’=2. Accidentally degenerate situations are observed for 1’
<3; the system, therefore, can be described in this case as
experiencing a truly “spin-frustrated” situation in which the
competing exchange interactions lead to
a degenerate
ground state.[74] For 1’=2, an accidentally degenerate
ground state is predicted where different spin states with
ST =4, 5, 6 have the same energies. For 1’=1, all of the mag-
netic interactions have the same magnitude, and again an
accidentally degenerate ground state is calculated, this time
with spin states ST =0, 1, 2. Only for 1’=1/4 is a relatively
separated ground state (ST =2) predicted.
From the previous analyses we can conclude that for most
combinations of J, J’, and J’’, J and J’ must be much higher
than J’’ to have a ground state with ST =6. However, for the
case in which J=J’, an isolated ST =6 ground state can be
obtained with a smaller ratio (1’ < 1). The experimental
and theoretical results for complex 3 provide strong evi-
dence that J and J’ are large enough compared to J’’ to give
a well-separated ST =6 ground state, and therefore the
system does not exhibit spin frustration in the strict sense of
degenerate ground states, although there are still competing
magnetic interactions created by the topology of the mole-
cule.
Figure 9. Spin topology for 3 derived from fitting the experimental data.
The ST =6 ground state is clearly a result of competing ex-
change interactions.
To shed more light on the magnetic structure of complex
3, we calculated the dependence of the spin frustration as a
function of the exchange parameters. The eigenvalues for
various ground states as a function of the ratio 1=((J + J’)/
2)/J’’, with J/J’ fixed to the experimental value of 0.62 (see
fit 2), are shown in Figure 10. At 1=3.7, the competing ex-
UV/Vis spectroscopy of 1–4: The absorption spectra for
compounds 1–4 are shown in Figure 11. All four complexes
exhibit one main peak in the near-UV region that is not
present in the spectrum of the free ligand (Py2SH in CH2Cl2
exhibits a weak band at lmax (e) 339 nm (41mꢀ1 cmꢀ1), and
intense absorption below 300 nm). The two mononuclear
Figure 10. Plot of the variation of the energy of the ground state and the
first excited state as a function of [(J + J’)/2]/J’’, with the J/J’ ratio fixed
at 0.62 (solid line). The variation of the energy is also reported for 1’=
[(Jav’)/2]/J’’, with Jav =J + J’ and ratio fixed at 1.0 (dotted line).
change pathways result in a reversal of the intrinsic antipar-
allel character of J’’ and give rise to a well-separated ground
state (ST =6) with parallel spin alignment of the outer iron
ions, as shown in Figure 9. As 1 varies from the experimen-
tal ratio of 3.7 to 1, the system tends to have a more frus-
trated magnetic structure and an associated decrease of ST
values. On moving to 1 values of 2 to 1/3, quasi-degenerate
ground states are found in which the uniform parallel align-
ment of the outer spins is no longer observed, indicating sig-
nificant spin frustration. The ST =0 state becomes the
ground state for 1=2. For 1=1, as expected for three com-
peting antiferromagnetic exchange pathways with similar
magnitude,[73] ST =0 is only 1.1 cmꢀ1 above the ST =2 ground
state.
Figure 11. Optical absorption spectra of 1 (solid line), 2 (dashed line), 3
(dotted line) and 4 (broken line) in CH2Cl2.
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7337

D. P. Goldberg et al.
complexes 1 and 2 exhibit spectra that are quite similar,
with lmax (e) 430 nm (667mꢀ1 cmꢀ1) for 1 and 422 nm
(572mꢀ1 cmꢀ1) for 2. However, the pentanuclear complex 3
shows a significantly blue-shifted and more intense peak at
364 nm (3770mꢀ1 cmꢀ1). The dinuclear complex 4 exhibits
one main peak in between the peaks for the mononuclear
complexes and the pentanuclear complex, at 409 nm
(1847mꢀ1 cmꢀ1). Although a thorough spectroscopic and the-
oretical analysis of these transitions has not been carried out
thus far, for all three cases these bands can be tentatively as-
signed to thiolate-to-iron(ii) charge-transfer (LMCT) transi-
tions based on their positions, relative intensities, and com-
parison with RS–FeII complexes in the literature.[16]
thiolate bridges, even though they are almost identical from
a structural perspective.
Experimental Section
General methods: All reactions were carried out under an atmosphere of
N2 or argon using a glovebox or standard Schlenk techniques. All re-
agents were purchased from commercial vendors and used as received
unless noted otherwise. Tetrahydrofuran (THF), diethyl ether, and pen-
tane were distilled under N2/argon from sodium/benzophenone. Methanol
and dichloromethane were distilled over CaH2. All solvents were further
degassed by direct bubbling of argon followed by repeated cycles of
freeze-pump-thaw, before storing in a glove box. Sodium hydride (60%
in mineral oil) was washed with hexanes prior to use.
UV/Vis spectra were obtained with a Hewlett-Packard 8453 diode array
spectrophotometer. Infrared spectra were measured either as KBr pellets
or as a thin film on NaCl plates on a Bruker FT-IR spectrometer. Ele-
mental analyses were performed by Atlantic Microlab Inc., Atlanta, GA.
Conclusions and Summary
The ligand Py2SH, which was used previously only for the
synthesis of copper complexes, has been employed in the
preparation of the mononuclear, dinuclear, and pentanu-
clear iron(ii) complexes 1–4. Ligands that provide mixed ni-
trogen/alkylthiolate donation are not common, and Py2SH is
therefore a useful ligand in this regard with respect to the
iron(ii) ion. The mononuclear complexes 1 and 2 exhibit the
Synthetic procedures: The ligand 1-(bis(2-(pyridin-2-yl)ethyl)amino)-2-
methyl propane-2-thiol (Py2SH) was prepared according to the method
published by Ohta et al.[21] The precursor amine, bis(2-pyridylethyl)-
amine, was synthesized by two different methods. One method involved
the reaction of NH4Cl and 2-vinyl pyridine in MeOH/water at reflux, fol-
lowing the method developed by Brady et al.[31] It was noticed that a sig-
nificant side product of this reaction was the primary amine 2-pyridyl-
ethylamine, which was isolated during the vacuum distillation of the
product. Thus a second method to synthesize Py2SH involved the reac-
tion of the primary amine with 2-vinyl pyridine. To a solution of the pri-
mary amine (7.6 g, 62 mmol) in MeOH was added 2-vinyl pyridine (6.5 g,
62 mmol) and acetic acid (3.6 mL, 62 mmol) and refluxed for 24 h. Stan-
dard aqueous workup followed by column chromatography (95:5 CH2Cl2/
CH3OH, neutral alumina) resulted in bis(2-pyridylethyl)amine (10.5 g) as
an oil.[32] The bis(2-pyridylethyl)amine (7.2 g, 32 mmol) thus obtained
was treated with isobutylene sulfide (3.7 mL, 38mmol), following the
method developed by of Ohta et al.[21] to give Py2SH (5.4 g) as an oil
(55%).
[(Py2S)FeIICl] (1): To a stirring solution of Py2SH (40 mg, 0.126 mmol) in
methanol (2.5 mL) was added a methanolic solution (1.5 mL) of NaOH
(10.1 mg, 0.253 mmol) and this mixture was allowed to stir for 1 h. A
clear solution of the deprotonated thiolate ligand thus obtained was
added to FeCl2·4H2O (25 mg, 0.126 mmol) in methanol (2.5 mL). The re-
sulting yellow solution was allowed to stir for 1.5 h and then filtered
through a bed of celite. Diffusion of diethyl ether into the filtrate afford-
ed yellow crystals of 1 (22 mg, 42%). 1H NMR (500 MHz, CD2Cl2): d=
90.9, 42.9, 33.9, 32.5, ꢀ8.2, ꢀ19.0, ꢀ24.5 ppm; IR (KBr): n=2919, 2853,
1601, 1564, 1445, 1364, 1310, 1268, 1104, 1019, 781, 588 cmꢀ1; UV/Vis
(CH2Cl2): lmax (e): 430 nm (667mꢀ1 cmꢀ1); elemental analysis calcd (%)
for C18H24ClFeN3S (405.7): C 53.28, H 5.96, N 10.35; found: C 52.81, H
6.02, N 10.19.
five-coordinate geometries expected for TPA-type com-
II
ꢀ
plexes, but display unusually long Namine Fe bonds when
compared to other TPA-type species. Their solution state
structures match the solid state as revealed by paramagnetic
NMR and UV/Vis data.
A method for incorporating OHꢀ into the coordination
sphere of iron(ii), which relies upon the use of strict aprotic
conditions and a limiting amount of H2O as the OHꢀ
source, was developed after traditional methods of self-as-
sembly and substitution chemistry failed. This method led to
the synthesis of the structurally unprecedented pentanuclear
and dinuclear iron(ii) complexes 3 and 4. It was shown that
production of 3 or 4 could be controlled by fine-tuning the
ligand-to-metal ratio using almost identical reaction condi-
tions.
Magnetic susceptibility studies of 3 and 4 were carried
out. In the case of 4, the two face-sharing, biooctahedral
iron(ii) ions were found to be antiferromagnetically coupled
to give a diamagnetic ground state. From the size of the ex-
changing coupling (J) and comparisons with the literature it
was surmised that the dominant exchange pathway in 4 is
likely the thiolate bridges. The magnetic behavior for 3 is
significantly more complicated, but was successfully mod-
eled through a spin Hamiltonian method and accompanying
DFT calculations. This complex exhibits ferrimagnetic be-
havior, leading to a high spin ST =6 ground state. Analysis
of the exchange coupling pathways reveals that all of the
iron(ii) ions couple in an antiferromagnetic manner, but the
thiolate bridges provide the dominant interactions and force
the iron(ii) ions connected by a hydroxide bridge to exhibit
parallel spin alignment in the ground state. From the fitting
of the data and DFT calculations, it is also concluded that
the exchange coupling differs significantly for the different
[(Py2S)FeIIBr] (2): To a stirring solution of Py2SH (40 mg, 0.126 mmol) in
methanol (2.5 mL) was added a methanolic solution (1.5 mL) of NaOH
(10.1 mg, 0.253 mmol) and this mixture was allowed to stir for 1 h. A
clear solution of the deprotonated thiolate ligand thus obtained was
added to FeBr2 (27 mg, 0.126 mmol) in methanol (3 mL). The resulting
yellow solution was allowed to stir for 1.5 h and then filtered through a
bed of celite. Diffusion of diethyl ether into the filtrate afforded yellow
crystals of 2 (25 mg, 45%). 1H NMR (500 MHz, CD2Cl2): d=95.3, 41.5,
32.5, ꢀ15.6, ꢀ23.0 ppm; IR (KBr): n=2956, 2916, 2862, 1567, 1484, 1444,
1375, 1354, 1314, 1262, 1163, 1147, 1098, 1077, 1061, 1020, 940, 789, 769,
669, 644, 615, 594, 557 cmꢀ1
; UV/Vis (CH2Cl2): lmax (e): 422 nm
(572mꢀ1 cmꢀ1); elemental analysis calcd (%) for C18H24BrFeN3S (450.2):
C 48.01, H 5.37, N 9.33; found: C 48.24, H, 5.36, N 9.36.
Synthesis of [(Py2S)4FeII5(m-OH)2](BF4)4 (3): To a suspension of NaH
(4.8mg, 0.201 mmol) in THF (2 mL) was added a solution of Py 2SH
(53 mg, 0.168mmol) in THF (2 mL). The suspension went clear and the
evolution of gas was noted. The reaction mixture was stirred for 30 min
7338
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

[{N,S(thiolate)}Iron(ii)] Complexes
FULL PAPER
and then treated with a THF solution (2 mL) of Fe(BF4)2·6H2O (56 mg,
0.168mmol), and a yellow precipitate resulted immediately. The reaction
mixture was then stirred for 16 h. The volatiles were removed under
vacuum and the resulting yellow solid was redissolved in dichlorome-
thane and filtered through a bed of celite. Brown crystals of 3 (26 mg,
34%) were obtained by diffusion of pentane into the dichloromethane
solution of the filtrate. 1H NMR (400 MHz, CD2Cl2): d=94.5 (br), 64.0
(br), 60.0, 49.0, 41.9, 38.5, 37.0 (br), 36.0, 24.5, 22.0, 18.0 (br), 10.5 (br),
ꢀ0.5, ꢀ1.5, ꢀ7.0 (br), ꢀ18.0 ppm; IR (thin film): n=3034, 2963, 2926,
2863, 1730, 1668, 1568, 1484, 1444, 1364, 1316, 1158, 1052, 875, 665, 646,
594, 520 cmꢀ1; UV/Vis (CH2Cl2): lmax (e): 364 nm (3770mꢀ1 cmꢀ1); ele-
mental analysis calcd (%) for C72H98Fe5N12S4O2B4F16 (1918.3): C 45.07, H
5.14, N 8.76; found: C 45.15, H 5.25, N 8.40.
(1808–t–908) method. Crystals of 1 and 2 were isolated from mother
liquor and washed repeatedly with Et2O to remove impurities prior to
analysis by NMR. Crystals of 3 were simply isolated from the mother
liquor and used directly for NMR experiments.
Magnetic susceptibility studies: Magnetic susceptibility and magnetiza-
tion measurements on powdered samples were performed on a Cryogen-
ics S600 SQUID magnetometer operating between 0 and 6 T in the range
1.8–300 K. The data were corrected for the diamagnetic contributions
using Pascals’ constants.
Conductivity: Conductivity measurements were carried out in dichloro-
methane with an Accumet AR-20 Conductivity meter at 258C, using an
Accumet conductivity cell (cell constant K=1.0 cmꢀ1). The 1:1 electro-
lyte [nBu4N]PF6 was used as a standard. The data obtained from the
measurements were used to calculate Le, the equivalent conductivity at
different concentrations. A plot of Le as a function of the square root of
concentration (c1/2), followed by extrapolation through a linear least-
squares fit resulted in L0, the conductivity at infinite dilution, as the y in-
tercept. Onsager plots were then constructed by plotting the difference,
Synthesis of [(Py2S)2FeII2(m-OH)]BF4 (4): To
a suspension of NaH
(5.8mg, 0.244 mmol) in THF (2 mL) was added a solution of Py 2SH
(64.3 mg, 0.203 mmol) in THF (2 mL). The suspension went clear and the
evolution of gas was noted. The reaction mixture was stirred for 30 min
and then treated with a THF solution (2 mL) of Fe(BF4)2·6H2O (51.6 mg,
0.152 mmol), and a yellow suspension resulted immediately. The reaction
mixture was then stirred for 16 h. The volatiles were removed under
vacuum and the resulting yellow solid was redissolved in dichlorome-
thane and filtered through a bed of celite. Orange crystals of 4 (78mg,
45%) were obtained by diffusion of pentane into the dichloromethane
solution of the filtrate. 1H NMR (400 MHz, CD2Cl2): d=79.8(br), 41.0,
32.4, ꢀ14.2 ppm (br); IR (thin film): n=3032, 2957, 2917, 2852, 1723,
1568, 1484, 1379, 1360, 1321, 1262, 1158, 1137, 954, 664, 640 cmꢀ1; UV/Vis
(CH2Cl2): lmax (e) 409 nm (1511Mꢀ1 cmꢀ1); elemental analysis calcd (%)
for C36H49Fe2N6S2OBF4 (844.44): C 51.20, H 5.84, N 9.95; found: C 50.59,
H 5.72, N 9.65.
Paramagnetic 1H NMR spectroscopy and T1 measurements: The
1H NMR spectra for the high-spin iron(ii) complexes 1 and 2 were mea-
sured on a Varian Inova FT-NMR instrument at 500 MHz (1H). The
1H NMR spectrum for complex 3 was measured on a Varian Unity FT-
NMR instrument at 400 MHz (1H). Chemical shifts (in ppm) are refer-
enced to the residual solvent peak CHDCl2. Longitudinal relaxation
times (T1) were measured using the inversion–recovery–pulse sequence
L0ꢀLe, as a function of c1/2
.
X-ray crystal structure determinations: Diffraction intensity data were
collected with a Bruker Smart Apex CCD diffractometer at 150 K for 1
and 2 and 173 K for 3. Crystal data, data collection, and refinement pa-
rameters are given in Table 6. The space groups for 1–3 were chosen
based on the systematic absences and intensity statistics. The structures
were solved by the direct methods, completed by subsequent difference
Fourier syntheses, and refined by full matrix least-squares procedures on
F2. SADABS[75] absorption corrections were applied to the data for 1–3.
All non-hydrogen atoms in 1–3 were refined with anisotropic displace-
ꢀ
ment coefficients except the B and F atoms of a disordered BF4 ion in 3,
which were refined with isotropic thermal parameters. The H atoms in 1
and 2 were found on the Fourier maps and refined with isotropic thermal
parameters. The H atoms in 3 were treated as idealized contributions.
The Flack parameters for the non-centrosymmetrical structures 1 and 2
are 0.006(9) and 0.005(5), respectively. All software and sources of scat-
Table 6. Crystallographic data for 1, 2, 3·2CH2Cl2·C5H12, and 4.[a]
1
2
3·2CH2Cl2·C5H12
4
empirical formula
fw
T [K]
C18H24ClFeN3S
405.76
150(2)
C18H24BrFeN3S
450.22
150(2)
C79H114B4Cl4F16Fe5N12O2S4
2160.36
173(2)
C36H49BF4Fe2N6OS2
844.44
100
cryst system
space group
a [Š]
b [Š]
c [Š]
orthorhombic
Pna21
14.9482(11)
8.9368(6)
14.1757(10)
–
orthorhombic
Pna21
14.9037(6)
8.9422(4)
14.4390(6)
–
monoclinic
C2/c
13.0575(16)
24.876(3)
31.214(4)
–
95.349(2)
–
10095(2)
4
triclinic
P1
¯
12.030(2)
12.7114(17)
12.806(2)
87.467(13)
89.840(15)
73.306(13)
1873.7(5)
2
a [8]
b [8]
–
–
–
–
g [8]
V [Š3]
1893.7(2)
4
1924.3(1)
4
Z
color, description
crystal size [mm]
yellow, block
0.25”0.20”0.10
1.423
1.052
SADABS
yellow, block
0.25”0.15”0.10
1.554
2.973
SADABS
(0.880)
yellow-brown, needle
0.48”0.23”0.15
1.421
0.9680.943
SADABS
(0.843)
orange, block
0.18”0.11” 0.031
1.497
1calcd [gcmꢀ3
]
m [mmꢀ1
]
absorption
correction
Numerical
(0.836)
(Tmin/Tmax
)
reflections collected
unique reflections
11020
4311(0.0188)
11156
3814(0.0175)
31819
11523(0.0337)
18888
12201(0.0417)
(Rint
)
obsd reflections
R1 (I>2s(I)))
wR2 (I>2s(I)))
GOF2
41483652 70
0.0215
0.0501
78
0.0191
0.0426
0.992
7544
0.0612
0.1827
1.054
0.0561
0.1402
0.994
1.001
max;min
0.291;ꢀ0.2180.454;
ꢀ0.181
0.95;ꢀ0.94
1.550;ꢀ0.915
[a] For all crystal structure determinations MoKa radiation (l=0.71073 Š) has been used.
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7339

D. P. Goldberg et al.
tering factors are contained in the SHELXTL (5.10) program package
(G.Sheldrick, Bruker XRD, Madison, WI).
[20] F. M. Macdonnell, K. Ruhlandtsenge, J. J. Ellison, R. H. Holm, P. P.
Power, Inorg. Chem. 1995, 34, 1815–1822.
[21] T. Ohta, T. Tachiyama, K. Yoshizawa, T. Yamabe, T. Uchida, T. Ki-
tagawa, Inorg. Chem. 2000, 39, 4358–4369.
[22] T. Ohta, T. Tachiyama, K. Yoshizawa, T. Yamabe, Tetrahedron Lett.
2000, 41, 2581–2585.
[23] S. C. Shoner, A. M. Nienstedt, J. J. Ellison, I. Y. Kung, D. Barnhart,
J. A. Kovacs, Inorg. Chem. 1998, 37, 5721–5726.
[24] J. A. Halfen, H. L. Halfen, T. C. Brunold, A. T. Fiedler, J. Am.
Chem. Soc. 2005, 127, 1675.
[25] D. K. Garner, S. B. Fitch, L. H. McAlexander, L. M. Bezold, A. M.
Arif, L. M. Berreau, J. Am. Chem. Soc. 2002, 124, 9970–9971.
[26] B. M. Bridgewater, G. Parkin, Inorg. Chem. Commun. 2001, 4, 126–
129.
Diffraction intensity data were measured with an Oxford Diffraction
Xcalibur3 diffractometer equipped with an Enhance (Mo) X-ray Source
(l=0.71073 Š) operated at 2 kW power (50 kV, 40 mA) at 100 K for 4.
Crystal data, data collection, and refinement parameters are given in
Table 6. The frames were integrated and a face-indexed absorption cor-
rection was applied with the Oxford Diffraction CrysAlisRED software
package. The space group for 4 was chosen on the basis of the systematic
absences and intensity statistics. The structure was solved by direct meth-
ods and refined using the Bruker SHELXTL (v6.1) software package.
Analysis of the data showed no decay. All non-hydrogen atoms in 4 were
refined with anisotropic displacement coefficients; the hydrogen atoms
were treated as idealized contributions.
[27] R. Alsfasser, S. Trofimenko, A. Looney, G. Parkin, H. Vahrenkamp,
Inorg. Chem. 1991, 30, 4098–4100.
[28] A. Looney, R. Han, K. Mcneill, G. Parkin, J. Am. Chem. Soc. 1993,
115, 4690–4697.
CCDC-263220 (1), CCDC-263221 (2), CCDC-263219 (3), and CCDC-
263201 (4) contain the supplementary crystallographic 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.
[29] S. Hikichi, T. Ogihara, K. Fujisawa, N. Kitajima, M. Akita, Y. Moro-
Oka, Inorg. Chem. 1997, 36, 4539–4547.
[30] N. Kitajima, N. Tamura, M. Tanaka, Y. Morooka, Inorg. Chem.
1992, 31, 3342–3343.
[31] L. E. Brady, M. Freifelder, G. R. Stone, J. Org. Chem. 1961, 26,
4757–4758.
[32] E. Gojon, P. Dubourdeaux, N. Gon, J. M. Latour, New J. Chem.
1988, 12, 931–939.
[33] D. Mandon, A. Machkour, S. Goetz, R. Welter, Inorg. Chem. 2002,
41, 5364–5372.
[34] Y. Zang, L. Que, Jr., Inorg. Chem. 1995, 34, 1030–1035.
[35] R. N. Mukherjee, A. J. Abrahamson, G. S. Patterson, T. D. P. Stack,
R. H. Holm, Inorg. Chem. 1988, 27, 2137–2144.
Acknowledgements
We are grateful to the National Institutes of Health (GM62309 to D.P.G.)
for support of this work. D.P.G. also acknowledges the Alfred P. Sloan,
Jr. Foundation. We also thank Dr. Ananya Majumdar for assistance with
NMR measurements. A.C. and F.T. acknowledge grants from Italian
MIUR, FIRB project, and E.C. RTN, QUEMOLNA, FP6–504880.
[36] Y. M. Chiou, L. Que, Jr., J. Am. Chem. Soc. 1995, 117, 3999–4013.
[37] A. Diebold, K. S. Hagen, Inorg. Chem. 1998, 37, 215–223.
[38] F. H. Allen, Acta Crystallogr. Sect. B 2002, 58, 380–388.
[39] See for example: a) X. D. Xu, C. S. Allen, C. L. Chuang, J. W.
Canary Acta Crystallogr. C 1998, 54, 600 and b) E. H. Alilou, A. E.
Hallaoui, E. H. E. Ghadraoui, M. Giorgi, M. Pierrot, M. Reglier,
Acta Crystallogr. Sect. C 1997, 53, 559.
[40] A. W. Addison, T. N. Rao, J. Reedjik, J. van Rijn, G. C. Verschoor, J.
Chem. Soc. Dalton Trans. 1984, 1349–1456.
[41] A. L. Gavrilova, B. Bosnich, Chem. Rev. 2004, 104, 349–383.
[42] R. S. Drago, Physical Methods for Chemists, 2nd ed., Saunders Col-
lege Publishing, 1992, p. 500.
[43] R. Alsfasser, M. Ruf, S. Trofimenko, H. Vahrenkamp, Chem. Ber-
Recl 1993, 126, 703–710.
[44] For the self-assembly of a related calixarene zinc-aqua species, see:
O. Seneque, M. N. Rager, M. Giorgi, O. Reinaud, J. Am. Chem. Soc.
2001, 123, 8442–8443.
[45] V. L. MacMurdo, H. Zheng, L. Que, Jr., Inorg. Chem. 2000, 39,
2254–2255.
[46] R. A. Reynolds, W. O. Yu, W. R. Dunham, D. Coucouvanis, Inorg.
Chem. 1996, 35, 2721–2722.
[1] G. Parkin, Chem. Rev. 2004, 104, 699–767.
[2] V. V. Karambelkar, D. Krishnamurthy, C. L. Stern, L. Zakharov,
A. L. Rheingold, D. P. Goldberg, Chem. Commun. 2002, 2772–2773.
[3] S. C. Chang, V. V. Karambelkar, R. D. Sommer, A. L. Rheingold,
D. P. Goldberg, Inorg. Chem. 2002, 41, 239–248.
[4] S. Chang, R. D. Sommer, A. L. Rheingold, D. P. Goldberg, Chem.
Commun. 2001, 2396–2397.
[5] J. N. Smith, Z. Shirin, C. J. Carrano, J. Am. Chem. Soc. 2003, 125,
868–869.
[6] B. Benkmil, M. Ji, H. Vahrenkamp, Inorg. Chem. 2004, 43, 8212–
8214.
[7] J. Seebacher, M. H. Shu, H. Vahrenkamp, Chem. Commun. 2001,
1026–1027.
[8] J. J. Smee, M. L. Miller, C. A. Grapperhaus, J. H. Reibenspies, M. Y.
Darensbourg, Inorg. Chem. 2001, 40, 3601–3605.
[9] C. A. Grapperhaus, M. Y. Darensbourg, Acc. Chem. Res. 1998, 31,
451–459.
[10] V. E. Kaasjager, E. Bouwman, S. Gorter, J. Reedijk, C. A. Grapper-
haus, J. H. Reibenspies, J. J. Smee, M. Y. Darensbourg, A. Derec-
skei-Kovacs, L. M. Thomson, Inorg. Chem. 2002, 41, 1837–1844.
[11] M. L. Golden, M. V. Rampersad, J. H. Reibenspies, M. Y. Da-
rensbourg, Chem. Commun. 2003, 1824–1825.
[47] D. R. Evans, R. S. Mathur, K. Heerwegh, C. A. Reed, Z. W. Xie,
Angew. Chem. 1997, 109, 1394–1396; Angew. Chem. Int. Ed. Engl.
1997, 36, 1335–1337.
[12] C. A. Grapperhaus, C. S. Mullins, P. M. Kozlowski, M. S. Mashuta,
Inorg. Chem. 2004, 43, 2859–2866.
[48] W. R. Scheidt, B. Cheng, M. K. Safo, F. Cukiernik, J. C. Marchon,
P. G. Debrunner, J. Am. Chem. Soc. 1992, 114, 4420–4421.
[49] S. C. Payne, K. S. Hagen, J. Am. Chem. Soc. 2000, 122, 6399–6410.
[50] J. A. R. Hartman, R. L. Rardin, P. Chaudhuri, K. Pohl, K. Wie-
ghardt, B. Nuber, J. Weiss, G. C. Papaefthymiou, R. B. Frankel, S. J.
Lippard, J. Am. Chem. Soc. 1987, 109, 7387–7396.
[51] W. J. Geary, Coord. Chem. Rev. 1971, 7, 81–122.
[52] J. F. Coetzee, G. P. Cunningham, J. Am. Chem. Soc. 1965, 87, 2529–
2534.
[13] J. A. Kovacs, Chem. Rev. 2004, 104, 825–848.
[14] J. Shearer, R. C. Scarrow, J. A. Kovacs, J. Am. Chem. Soc. 2002, 124,
11709–11717.
[15] S. C. Shoner, D. Barnhart, J. A. Kovacs, Inorg. Chem. 1995, 34,
4517–4518.
[16] J. A. Halfen, H. L. Moore, D. C. Fox, Inorg. Chem. 2002, 41, 3935–
3943.
[17] T. C. Harrop, P. K. Mascharak, Acc. Chem. Res. 2004, 37, 253–260.
[18] J. C. Noveron, M. M. Olmstead, P. K. Mascharak, J. Am. Chem. Soc.
2001, 123, 3247–3259.
[53] R. D. Feltham, R. G. Hayter, J. Chem. Soc. 1964, 4587–4591.
[54] T. Chandra, R. A. Allred, B. J. Kraft, L. M. Berreau, J. M. Zaleski,
Inorg. Chem. 2004, 43, 411–420.
[19] C. A. Grapperhaus, M. Li, A. K. Patra, S. Poturovic, P. M. Kozlow-
ski, M. Z. Zgierski, M. S. Mashuta, Inorg. Chem. 2003, 42, 4382–
4388.
[55] Y. H. Chiu, J. W. Canary, Inorg. Chem. 2003, 42, 5107–5116.
7340
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7328– 7341

[{N,S(thiolate)}Iron(ii)] Complexes
FULL PAPER
[56] B. Kersting, D. Siebert, D. Volkmer, M. J. Kolm, C. Janiak, Inorg.
Chem. 1999, 38, 3871–3882.
[57] F. A. Cotton, D. A. Ucko, Inorg. Chim. Acta 1972, 6, 161–172.
[58] R. H. Summerville, R. Hoffmann, J. Am. Chem. Soc., 1979, 101,
3821–3831.
[59] M. Mikuriya, T. Kotera, F. Adachi, M. Handa, M. Koikawa, H.
Okawa, Bull. Chem. Soc. Jpn. 1995, 68, 574–580.
[60] U. Bossek, D. Nꢂhlen, E. Bill, T. Glaser, C. Krebs, T. Weyhermꢂller,
K. Wieghardt, M. Lengen, A. X. Trautwein, Inorg. Chem. 1997, 36,
2834–2843.
[61] Z. Janas, P. Sobota, T. Lis, J. Chem. Soc. Dalton 1991, 2429–2434.
[62] J. D. Lawrence, T. B. Rauchfuss, S. R. Wilson, Inorg. Chem. 2002, 41,
6193–6195.
[63] L. Toma, R. Lescouezec, J. Vaissermann, F. S. Delgado, C. Ruiz-
Perez, R. Carrasco, J. Cano, F. Lloret, M. Julve, Chem. Eur. J. 2004,
10, 6130–6145.
son, J. Nieplocha, T. V., M. Krishnan, E. Brown, G. Cisneros, G.
Fann, H. Fruchtl, J. Garza, K. Hirao, R. Kendall, J. Nichols, K. Tse-
mekhman, M. Valiev, K. Wolinski, J. Anchell, D. Bernholdt, P. Bor-
owski, T. Clark, D. Clerc, H. Dachsel, M. Deegan, K. Dyall, D.
Elwood, E. Glendening, M. Gutowski, A. Hess, J. Jaffe, B. Johnson,
J. Ju, R. Kobayashi, R. Kutteh, Z. Lin, R. Littlefield, X. Long, B.
Meng, T. Nakajima, S. Niu, M. Rosing, G. Sandrone, M. Stave, H.
Taylor, G. Thomas, J. van Lenthe, A. Wong, Z. Zhang, “NWChem,
A Computational Chemistry Package for Parallel Computers, v. 4.5”
(2003), Pacific Northwest National Laboratory, Richland, Washing-
ton 99352–0999, USA.
[71] R. A. Kendall, E. Apra, D. E. Bernholdt, E. J. Bylaska, M. Dupuis,
G. I. Fann, R. J. Harrison, J. L. Ju, J. A. Nichols, J. Nieplocha, T. P.
Straatsma, T. L. Windus, A. T. Wong, Comput. Phys. Commun. 2000,
128, 260. See references therein for the Ahlrichs TZV/VDZ basis
sets and for the B3LYP functional.
[64] G. Aromi, C. Boldron, P. Gamez, O. Roubeau, H. Kooijman, A. L.
Spek, H. Stoeckli-Evans, J. Ribas, J. Reedijk, Dalton Trans. 2004,
3586.
[72] A. Bencini, F. Totti, Int. J. Quantum Chem. 2005, 101, 819.
[73] E. Libby, J. K. McCusker, E. A. Schmitt, K. Folting, D. N. Hendrick-
son, G. Christou, Inorg. Chem. 1991, 30, 3486–3495.
[65] D. Gatteschi, A. Caneschi, R. Sessoli, A. Cornia, Chem. Soc. Rev.
1996, 25, 101.
[74] O. Kahn,
Molecular Magnetism, VCH Publishers, New York,
1993.
[66] D. R. Gamelin, E. L. Bominaar, M. L. Kirk, K. Wieghardt, E. I. So-
lomon, J. Am. Chem. Soc. 1996, 118, 8085–8097.
[75] G. M. Sheldrick, SADABS (2.01), Bruker/Siemens Area Detector
Absorption Correction Program, Bruker AXS, Madison, Wisconsin,
USA, 1998.
[67] D. Gatteschi, L. Pardi, Gazz. Chim. Ital. 1993, 123, 231.
[68] L. Noodleman, J. G. J. Norman, J. Chem. Phys. 1979, 70, 4903–4906.
[69] L. Noodleman, J. Chem. Phys. 1981, 74, 5737–5743.
[70] T. P. Straatsma, E. Apra’, T. L. Windus, M. Dupuis, E. J. Bylaska, W.
de Jong, S. Hirata, D. M. A. Smith, M. Hackler, L. Pollack, R. Harri-
Received: February 12, 2005
Published online: September 6, 2005
Chem. Eur. J. 2005, 11, 7328– 7341
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7341