5502
Z. Wang et al. / Journal of Organometallic Chemistry 692 (2007) 5501–5507
triisopropylsilyl, which was treated as the alternative
NH
S4Fe4
amino protective groups. Alkylsilylamine as starting mate-
rial can react with paraformaldehyde to form the
bis(hydromethyl) derivative 2a–c (Scheme 1). Treatment
of thiol complex 3 with 2a–c afforded diiron azadithiolates
4a–c. The alkylsilyl groups were simultaneously removed
by the excess CF3COOH existing in the solution to afford
5 in isolated yields of 10% for 1a, 32% for 1b and 36%
for 1c (Scheme 1). But we did not obtain the stable nitro-
gen-protective complex 4a–c, probably due to unstable
Si–N bond in the presence of trace acid. Comparing with
the isolated yield for 1b (32%) and 1c (36%), the yield for
1a (10%) is quite low. Under the same experimental condi-
tion, the unisolated yield of our method is obviously higher
than that of Rauchfuss’ method.
S
S
S
Cys
H2O
CO
Fe
Fe
OC
NC
C
CN
O
Fig. 1. The active site of Fe-only hydrogenase.
2-CH2C4H3O) [12–16], less work has been done to the
construction of this parent core structure, i.e., (l-SCH2)2-
NH[Fe2(CO)6]. Recently, Rauchfuss et al. reported the first
synthesis of this model complex by a three-component con-
densation starting from iron carbonyl sulfides, formalde-
hyde, and amines. However, the yield in this reaction was
low (37%) and no crystal structure was reported [17]. The
synthetic difficulty may probably be attributed to the
unstability of the azadithiolate-bridged complex in which
the central atom of the chelate is a secondary amine. We
decided to assess whether organosilicon reagent could be
applied to such reaction as a N-protection reagent. To this
end, alkylsilyl chloride was chosen and subjected to ami-
nating with NH3 or KNH2, followed by treating with form-
aldehyde, and finally reacted with Fe–S complex to achieve
the condensation. In addition, since the electron-donating
characteristics of tertiary phosphine ligands towards the
Fe atom are similar to those of CNÀ ligands and the com-
plicated protonation on the cyanide nitrogen atom can be
avoided, the PR3-monosubstituted complexes (l-
SCH2)2NH[Fe2(CO)5PR3] (R = Ph (6), Me (7)) were pre-
pared for a study of the influence of phosphine ligands
on the coordination structures and the protonation ability.
In this paper, we would like to report this modified proto-
col for the synthesis of (l-SCH2)2NH[Fe2(CO)6] (5), the
conversion to its PR3-coordinated products as well as their
electrochemical properties.
Monosubstituted complexes (l-SCH2)2NH[Fe2(CO)5L]
(L = PPh3 (6), PMe3 (7)) were readily prepared in moder-
ate to reasonable yields by treating 5 with PR3, according
to literature procedures (see Scheme 2) [18,19]. Considering
the different solubility and coordination ability of the ter-
tiary phosphine, different solvents were used for the indi-
vidual reactions, with hexane for trimethylphosphine and
toluene for triphenylphosphine. While the reaction of 5
with 1 equiv of trimethylphosphine, which has a good
coordinated ability, can be controlled at the monosubsti-
tuted stage, the reaction to form complex 6 required
2 equiv of PPh3 in refluxing toluene.
1
The products 5–7 were characterized by IR, H and 31P
NMR spectra and elemental analyses. The IR spectra of
these three complexes in KBr show three strong CO bands
in the region of 1910–2080 cmÀ1, identical to the carbonyl
stretching pattern for terminal CO bonds. In comparison
with all carbonyl complex 5, the average value of the bands
for monosubstituted complex 6 and 7 are lowered by 36
and 60 cmÀ1, respectively.
2.2. Molecular structures of complexes 5 and 6
2. Results and discussion
Single-crystal X-ray diffraction analysis of 5 and 6
(Fig. 2) shows the usual distorted square-pyramidal geom-
etry around the iron centers. The Fe–Fe distances of the
two complexes are somewhat shorter than those in the
structures of H2-uptake enzyme from DdHase (Desulfovib-
rio desulfuricans) and H2-evolving enzyme from CpHase
2.1. Synthesis and spectroscopic characterization
Rauchfuss and co-workers have synthesized the core
structural molecule (l-SCH2)2NH[Fe2(CO)6] (5) by the
treatment of 3 with a premixed solution of paraformalde-
hyde and (NH4)2CO3 or hexamethylenetetramine in yield
of 37% [17]. However, the isolated yield for the target com-
plex is very low (<20%) when we repeated their work. We
reasoned that the bridge N could be protected to increase
the stabilities. Many amine protective groups such as acyl
chloride and carbamate are usually removed by treatment
with strong base. In contrast, the silyl group can be
removed quite easily by treatment with trace acid or in a
mild, natural conditions, for instance by using TBAF.
Thus, considering the unstability of the Fe2S2 skeleton in
basic conditions, we choose some silyl reagents as the
amino protective groups to achieve the transformation to
complex 5, including triethylsilyl, t-butyldimethylsilyl and
˚
(Clostridium pasteurianum) (ca. 2.6 A) [3,4,10], but still in
good agreement with the structural data of Fe–Fe bonds
˚
(2.49–2.51 A) found in other diiron azadithiolates com-
plexes [12,15,17]. The CO displacement by one molecule
of triphenylphosphine ligand has only a small effect on
˚
the Fe–Fe distance [2.519(8) A in 6] as compared to that
˚
of 5 [2.505(1) A]. Similar to (l-SCH2)2NH[Fe2(CO)4
(CN)2] [17], the N–H of 5 and 6 are all axial as anticipated
[11] and the bridging-N atoms are of sp3-hybridization. In
the solid state, the nonbonding Fe–H distance between the
H atom at the bridging-N atom and the nearest Fe atom of
˚
5 is almost consistent with 6 (3.34 and 3.38 A, respectively),
which is significantly longer than that of dicyanide substi-