The use of trimethylsilyl iodide as a synthon in coordination chemistry

Article abstract of DOI:10.1016/S0020-1693(01)00801-5

Trimethylsilyl iodide is shown to be an efficient metathetical reagent for preparing transition-metal iodides from the corresponding chlorides, though often complications can cause problems. These include reduction of the starting metal chloride when its oxidation state is high, due to the reaction of iodide, and even oxidation of low-oxidation-state compounds, presumably by incipient silyl cations. Finally, some very inert chlorides, such as of iridium(III), react too slowly with the iodide under the experimental conditions, and simple reaction with solvent becomes predominant.

Full text of DOI:10.1016/S0020-1693(01)00801-5

www.elsevier.com/locate/ica
Inorganica Chimica Acta 330 (2002) 197–212
The use of trimethylsilyl iodide as a synthon in coordination
chemistry
G. Jeffery Leigh a,*, J. Roger Sanders b, Peter B. Hitchcock a, Jaisa Soares Fernandes c,
Maria Togrou a
a School of Chemistry, Physics, and En6ironmental Science, The Uni6ersity of Sussex, Brighton BN1 9QJ, UK
b Biological Chemistry Laboratory, John Innes Centre, Norwich Research Park, Colney, Norwich NR4 7UH, UK
c Departamento de Qu´ımica, Uni6ersidade Federal do Parana´, Centro Polite´cnico, Jardim das Americas, 81531-971 Curitiba, PR, Brazil
Received 15 October 2001; accepted 22 November 2001
Dedicated to the memory of Luigi Venanzi, a good friend, a teacher, and a gentleman
Abstract
Trimethylsilyl iodide is shown to be an efficient metathetical reagent for preparing transition-metal iodides from the
corresponding chlorides, though often complications can cause problems. These include reduction of the starting metal chloride
when its oxidation state is high, due to the reaction of iodide, and even oxidation of low-oxidation-state compounds, presumably
by incipient silyl cations. Finally, some very inert chlorides, such as of iridium(III), react too slowly with the iodide under the
experimental conditions, and simple reaction with solvent becomes predominant. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Crystal structures; Silyl complexes; Transition metal complexes; Iodide complexes
1. Introduction
from [VO(salen)] and trimethylsilyl iodide (H2salen=
N,N%-ethylenebis(salicylideneimine)).
We have recently been interested in using metal
iodides to prepare complexes containing hydronitrogen
residues that may be implicated in nitrogen fixation
reactions [1], because, iodide should be a good leaving
group. To prepare such iodides, we employed
trimethylsilyl iodide as a reagent. This should form
metal iodide compounds by metathetical reactions
driven by the greater volatility of trimethylsilyl chloride
as compared with trimethylsilyl iodide. In fact, al-
though metathesis often occurs [2], other reactions,
especially redox reactions, tend to complicate matters,
and the ultimate product is not easy to predict. An
alternative metathetical route might involve trimethylsi-
lyl iodide and a metal oxide, because, the strength of
the SiꢀO bonds in the generated hexamethyldisiloxane
might act as effective driver of the reaction. In fact, we
have already described [3] the production of [VI2(salen)]
Normally metathetical reactions involving tri-
methylsilyl iodide might be expected to proceed via
four-centre transition states. For example, 5–6 mol of
Me3SiI per atom of vanadium in thf at 0 °C break the
dinuclear core of [V2IICl3(thf)6][AlCl2Et2] [4] to give
[VI2(thf)4] [3], although it is remarkable that the same
reactants in thf at reflux temperature produce
[VI2ICl3(thf)6]I. These are not redox reactions, but
Me3SiI reacts smoothly with [VCl3(NCMe)3] [5] in ace-
tonitrile to reduce VIII smoothly to VII, giving
[VII(NCMe)6]I4 [4] In this reaction, the I2 produced by
oxidation of the iodide is combined into the anionic
part of the salt. The crystal contains a linear arrange-
ment of four iodine atoms, which we formulated as
I42−. This observation, led us to consider the wider
possibility of preparing low-oxidation-state compounds
using the mild reducing power of iodide, but also to try
and evaluate the general behaviour of Me3SiI as a
metathetical and reducing synthon. In this paper, we
present some data on a general survey of such reac-
tions, and we show that reactions with transition-metal
* Corresponding author.
E-mail address: g.j.leigh@sussex.ac.uk (G.J. Leigh).
0020-1693/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(01)00801-5

198
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
Table 1
Summary of results from reactions of trimethylsilyl iodide
Periodic
group
Simple metathesis or replacement
Metathesis and/or reduction
Metathesis and oxidation
4
5
6
[TiCl3(MeCN)3]“[TiI3(MeCN)3];
ZrCl4“[ZrI3(MeCN)4][ZrI5(MeCN)]
[VO(salen)]“[VI2(salen)] [2];
[V2Cl3(thf)6][AlCl2Et2]“[VI2(thf)4] [3]
CrCl2“[CrI2(NCMe)4];
[TiCl4(MeCN)2]“[TiI2(MeCN)4](I3);
TiIV“TiIII
[VCl3(MeCN)3]“[V(MeCN)6](I4) [3];
VIII“VII
[MoCl4(NCMe)2]“[MoCl4(NCMe)2](I3);
CrCl2“[CrI2(NCMe)4][CrI4(NCMe)2];
CrII“CrIII
[CrCl3(thf)3]“[CrI3(NCMe)3]
MoIV“MoIII
;
[WCl4(NCMe)2]“[WCl4(NCMe)2](I3);
WIV“WIII
8
9
FeI2“[Fe(NCMe)6][FeI4] [14,17];
FeI2“[Fe(thf)6][FeI3(thf)]2 also [16];
[RuCl2(NCMe)4]“[RuI(NCMe)5]I
CoCl2“[Co(NCMe)6][CoI4]
RuCl3“[Ru(NCMe)6]I4; RuII RuII
FeI2“[Fe(thf)6][FeI4]1.2[FeI3(thf)]0.8
;
FeII“FeII+FeIII
K2[IrCl6]“[IrCl3(NCMe)3]; IrIV“IrIII
compounds can indeed form iodides, but that the reac-
tions are not always simple. In some cases, the iodide
seems to act as a reducing agent and the ultimate
metal-containing product is not necessarily what might
have been required. In addition, we have also observed
unexpected oxidations, which may be a consequence of
the formation of incipient silyl radicals produced by
homolysis of the iodide.
crystals. Microanalyses (C, H and N) were compatible
with the formulations [TiIIII3(NCMe)3] or [TiIII2-
(NCMe)4](I4). X-ray structure analysis at 173 K showed
it to be the former, with some residual diethyl ether in
the crystal, as [TiIIII3(NCMe)3]·0.5Et2O (1) (Fig. 1).
Crystallographic data are presented in Table 14 and
selected molecular dimensions are shown in Table 3.
The configuration is the expected meridional octahedral
one, with distortions provoked by the bulk of the
iodide ions.
This complex is one of the first molecular iodides of
titanium with N-donors to be structurally characterised,
apart from [TiI2(py)4]I3 [7]. A related product is
[TiI2(NCMe)4]I3, prepared from TiI4 and dry MeCN
[8]. and it is isostructural with [MoI2(NCMe)4]I3 (see
below).
2. Results
A summary of our results is presented in Table 1,
and the properties of the corresponding products are
shown in Table 2.
2.1. Reactions of compounds of periodic Group 4
Table 2
Effective magnetic moments (veff) and IR data of metal iodide
complexes
The first reaction described represents an attempt to
produce a complex of titanium(II) by a simple reduc-
tion with iodide, as exemplified by our earlier success
[3] with vanadium(II) complex, [V(NCMe)6]I4 from a
reaction of [VCl3(NCMe)3].
Complex
Magnetic moments w(CꢁN) (cm−1
at 293 K (mB) a
)
[TiI3(NCMe3]·1/2Et2O (1)
[TiI2(NCMe)4]I3 (2)
[ZrI3(NCMe)4]
When an excess of Me3SiI was added to a hot, royal
blue solution of [TiCl3(NCMe)3] [6] in MeCN, an im-
mediate change to deep brownish–red was observed.
Diethyl ether was then carefully added to the deep
brownish–red solution, finally affording a large amount
of a deep brown hexagonal plates or long prisms. Both
kinds of material were of the same compound, as
shown by microanalysis or 1H NMR spectroscopy.
Some diethyl ether was present. The paramagnetic
(magnetic moment in solution 1.70–1.75 mB, 238–313
K), moisture-sensitive crystals both appeared solvated,
and apparently lost solvent upon standing under dini-
1.77
diamagnetic
2277(s), 2305(s)
2274(s), 2302(s)
[ZrI5(NCMe)] (3)
[CrI2(NCMe)4] (4)
[CrI2(MeCN)4]
4.62
not determined
2290(s), 2320(m)
[CrI4(MeCN)2] (5)
[CrI3(NCMe)3] (6)
[MoI2(NCMe)4]I3 (7)
[WI2(NCMe)4]I3 (8)
[RuI(NCMe)5]I (9)
[Ru(NCMe)6]I4 (10)
[Co(NCMe)6][CoI4] (11)
3.83
2286(s), 2319(s)
3.82
2273(s), 2319(m)
2290(w)
2291(w)
2281(s), 2311(s)
2349
diamagnetic
diamagnetic
4.87
[IrCl3(NCMe)3]·MeCN (12) Diamagnetic
1
trogen or in vacuo. H NMR analysis also suggested
the presence of a small amount of diethyl ether in the
a In solution in MeCN-d3.

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
199
Fig. 1. ORTEP representation of the structure of [TiIIII3(NCMe)3]·0.5Et2O (1).
Clearly, although Me3SiI does not reduce TiIII to TiII,
one might then expect it to reduce TiIV to TiIII and
possibly also ZrIV to ZrIII. Both of these were investi-
gated. The reaction of [TiCl4(NCMe)2] with the silyl
iodide gives [TiI2(NCMe)4](I3) (2) in over 70% yield.
This compound has been reported [8] previously,
though the X-ray structure determination was incom-
plete. Consequently a complete analysis is reported here
(Fig. 2, and Tables 4 and 14). The presence of the I3−
ion is evidence that the iodide is again the reducing
agent. The reaction of ZrCl4 turned out to be more
interesting, giving the diamagnetic [ZrI3(NCMe)4][ZrI5-
(NCMe)] [3], also in over 70% yield. The structure was
determined by X-ray analysis, and confirms that the
ZrIV is present in two different environments. There is
no reduction. The product has been obtained previ-
ously by direct reaction of ZrI4 with MeCN, and struc-
turally characterised [9].
gen bridges [10]. The reaction of CrI3 in MeCN with
zinc amalgam is reported to give blue CrI2(NCMe)2 and
this compound is stated to lose MeCN very easily to
yield anhydrous CrI2 [11]. Our complex forms blue–
green crystals that also showed a tendency to lose
MeCN. The magnetic moment, 4.90 mB, is close to the
expected high-spin spin-only value.
A minor byproduct of this preparation consisted of
brown crystals. X-ray analysis showed that they con-
sisted of a salt [CrI2(MeCN)4][CrI4(MeCN)2] (5). How-
Table 3
,
Bond lengths (A) and angles (°) for [TiI3(CNMe)3]·1/2Et2O (1)
Bond lengths
I(1)ꢀTi(1)
I(2)ꢀTi(1)
I(3)ꢀTi(1)
Ti(1)ꢀN(1)
Ti(1)ꢀN(2)
Ti(1)ꢀN(3)
N(1)ꢀC(1)
N(2)ꢀC(3)
N(3)ꢀC(5)
2.739(2)
2.725(2)
2.720(2)
2.163(9)
2.127(9)
2.167(9)
1.12(2)
2.2. Reactions of compounds of periodic Group 5
1.131(13)
1.130(13)
Some reactions of vanadium compounds have al-
ready been described above, and we did no further
work. VIII can be reduced to VII, but VII compounds
undergo simple metatheses.
Bond angles
N(1)ꢀTi(1)ꢀI(1)
I(2)ꢀTi(1)ꢀI(1)
I(3)ꢀTi(1)ꢀI(1)
N(2)ꢀTi(1)ꢀN(1)
N(2)ꢀTi(1)ꢀN(3)
N(1)ꢀTi(1)ꢀN(3)
N(2)ꢀTi(1)ꢀI(3)
N(1)ꢀTi(1)ꢀI(3)
N(3)ꢀTi(1)ꢀI(3)
N(2)ꢀTi(1)ꢀI(2)
N(1)ꢀTi(1)ꢀI(2)
N(3)ꢀTi(1)ꢀI(2)
N(2)ꢀTi(1)ꢀI(1)
N(3)ꢀTi(1)ꢀI(1)
I(3)ꢀTi(1)ꢀI(2)
176.6(3)
93.30(7)
96.97(7)
86.2(3)
175.1(4)
89.2(4)
88.2(3)
86.2(3)
89.7(3)
91.0(3)
83.5(3)
90.3(3)
94.9(3)
89.8(3)
169.72(8)
2.3. Reactions of compounds of periodic Group 6
The reaction of CrCl2 with trimethylsilyl iodide in
MeCN at reflux yields the diiodide [CrI2(NCMe)4] (4)
rapidly and in high yield. The structure is the conven-
tional trans-octahedral, (Fig. 3 and Tables 5 and 14),
but it is, nevertheless surprising. The corresponding
chloro- and bromo-adducts with MeCN have the stoi-
chiometry CrX2(NCMe)2, [10,11] and the bromide has a
polymeric structure in the solid state with double-halo-

200
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
ever, the crystals were of poor quality and the analysis
was taken only as far as R=15%. The detailed dimen-
sions will not be presented, but the overall structure is
quite clear and is presented in Fig. 4. Both anion and
cation are trans-octahedral, which is not surprising, but
the only sensible formulation would suggest that both
ions contain CrIII. The chloro-homologue of the anion
has been structurally characterised. This compound
might have arisen from the action of adventitious oxy-
gen or even water. However, we do not believe this to
be the case. The only other possibility is that some
stage in the reaction route is more homolytic than
heterolytic, and that trimethylsilyl radicals cause the
oxidation. This will be discussed further below.
Fig. 2. ORTEP representation of the structure of [TiI2(NCMe)4](I3) (2).
When an excess of Me3SiI was added to a hot, purple
solution of [CrCl3(thf)3] in MeCN, an immediate
change to deep brownish–red was observed. Diethyl
ether was then carefully added, finally affording a large
amount of deep red–brown crystals. The X-ray struc-
tural determination (Fig. 5 and Tables 6 and 14)
showed the product to be [CrI3(NCMe)3]·1/2Et2O (6),
generated by simple halide exchange.
It is remarkable that solvated CrI3 exists in two
different forms: as the molecular [CrI3(NCMe)3] (6) and
as the 1:1 electrolyte [CrI2(NCMe)4][CrI4(NCMe)2] (5).
Apparently, both complexes are quite stable and there
is no evidence of interconversion between the two
forms.
Table 4
,
Bond lengths (A) and angles (°) for [TiI2(CNMe)4]·MeCN (2)
TiꢀN(1)
TiꢀN(2)
TiꢀI(1)
I(2)ꢀI(3)
N(1)ꢀC(2)
N(2)ꢀC(4)
C(1)ꢀC(20)
C(3)ꢀC(4)
2.142(4)
2.145(4)
2.6898(4)
2.9185(5)
1.136(7)
1.128(7)
1.457(7)
1.464(7)
N(1)%ꢀTiꢀN(1)
N(1)%ꢀTiꢀN(2)
N(1)ꢀTiꢀN(2)
N(2)%ꢀTiꢀN(2)
N(1)ꢀTiꢀI(1)’
N(2)ꢀTiꢀI(1)
N(1)ꢀTiꢀI(1)
N(2)ꢀTiꢀI(1)
I(3)ꢀI(2)ꢀI(3)%%
C(2)ꢀN(1)ꢀTi
C(4)ꢀN(2)ꢀTi
180.0
91.49(17)
88.51(17)
180.00(19)
89.80(12)
90.24(12)
90.20(12)
89.76(12)
180.0
177.0(4)
177.8(4)
Symmetry transformations used to generate equivalent atoms: %−x,
−y, −z; %%−x+1, −y, −z+1.
In complex 6, the Et2O solvate is disordered and
gives an infinite zigzag line of electron density along the
c-axis. This was modelled in the refinement by includ-
ing five independent peaks as isotropic C atoms. The
geometry of the molecule (Fig. 8) is octahedral with a
meridional configuration. This is common to other six-
coordinate compounds of metals of Groups 4, 5 or 6 in
the oxidation state III. The CrꢀI distances in 6 (mean
,
value 2.6700(16) A) are similar to those found in other
CrIII iodide complexes and significantly smaller than
,
those found in 4 (3.0816(4) A). Evidently, the CrꢀI
bond lengths decrease with increasing oxidation state.
Hathaway and Holah [12] reported the synthesis of a
CrIII iodide complex by oxidation of chromium metal
with iodine in acetonitrile. This complex was formu-
lated as [CrI2(NCMe)4]I based on magnetic moment
measurements and IR and microanalytical data. Com-
plex 6 is isostructural with [TiI3(NCMe)3]·1/2Et2O and
[MoI3(NCEt)3] [13].
The reaction of Me3SiI with [MoCl4(NCMe)2] [14] in
acetonitrile under reflux gives trans-[MoI2(NCMe)4](I3)
(7) in substantial yield. The product has veff=3.83 mB,
suggesting three unpaired electrons and thus more con-
sistent with MoIII than MoIV. The crystal structure of
this material (Fig. 6 and Tables 7 and 15) presents
several interesting features, apart from confirming the
reduction of the molybdenum. The crystal contains
isolated undistorted, octahedral [MoI2(NCMe)4]+
Fig. 3. ORTEP representation of the structure of [CrI2(NCMe)4] (4).
Table 5
,
Bond lengths (A) and angles (°) for [CrI2(CNMe)4]·(4)
Bond lengths
CrꢀN(2)
CrꢀN(1)
CrꢀI
2.057(5)
2.101(4)
3.0816(4)
N(2)ꢀCrꢀN(1)’
N(2)ꢀCrꢀN(1)
N(2)ꢀCrꢀI’
92.01(18)
87.99(18)
89.08(14)
Bond angles
N(1)ꢀC(2)
N(2)ꢀC(4)
1.126(6)
1.131(6)
N(1)ꢀCrꢀI’
N(2)ꢀCrꢀI
N(1)ꢀCrꢀI
92.47(13)
90.92(14)
87.53(13)

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
201
cations and the usual linear (I3)− anion. The molybde-
Table 6
,
Bond lengths (A) and angles (°) for [CrI3(NCMe)3]·1/2Et2O (6)
,
num–iodine bond lengths are 2.730(1) A, which is not
unusual for compounds of this type, and the MeCN
ligands are of normal dimensions and almost ideally
Bond lengths
CrꢀN(1)
CrꢀN(3)
CrꢀN(2)
CrꢀI(1)
CrꢀI(2))
CrꢀI(3)
N(1)ꢀC(1)
N(2)ꢀC(3)
N(3)ꢀC(5)
2.045(7)
2.047(7)
2.054(8)
2.6550(15)
2.6628(16)
2.6922(16)
1.130(11)
1.135(11)
1.114(11)
,
linear. The I(l)ꢀI(2) separation (2.916 A) is normal for
an isolated (I3)− ion. The packing diagram shows that
it may be that we are dealing with a doubly bent (I5)3−
ion, coordinated at both ends. This is not sure. Al-
though iodine–iodine interactions are indeed very com-
mon, they are often not strong enough to justify
regarding them as formal two-electron bonds. In the
present case, the normal separations between the three
central atoms, and the long separations (although less
than the sum of the van der Waals radii) between the
Bond angles
I(1)ꢀCrꢀI(2)
I(1)ꢀCrꢀI(3)
N(1)ꢀCrꢀN(3)
N(1)ꢀCrꢀN(2)
N(3)ꢀCrꢀN(2)
N(1)ꢀCrꢀI(1)
N(3)ꢀCrꢀI(1)
N(2)ꢀCrꢀI(1)
N(1)ꢀCrꢀI(2)
N(3)ꢀCrꢀI(2)
N(2)ꢀCrꢀI(2)
N(1)ꢀCrꢀI(3)
N(3)ꢀCrꢀI(3)
N(2)ꢀCrꢀI(3)
92.37(5)
94.32(5)
176.6(3)
89.5(3)
87.3(3)
89.9(2)
93.4(2)
177.5(2)
90.7(2)
90.1(2)
85.2(2)
89.9(2)
88.9(2)
88.1(2)
,
two external pairs of iodine atoms (3.967(1) A) are not
consistent with the presence of the as yet unreported
(I5)3− ion.
The tungsten analogue, [WCl4(NCMe)2] [15] behaves
like the molybdenum, giving [WI2(NCMe)4](I3) (8). The
structure (Fig. 7 and Tables 8 and 15) is also similar,
and we exclude the possibility of (I5)3− on the same
grounds. In summary, trimethylsilyl iodide undergoes
simple metathesis with both CrII and CrIII and there is
no sign of reduction. There is evidence for oxidation of
CrII and CrIII. Both MoIV and WIV are reduced cleanly
to give iodides in the III oxidation state, as one might
now expect.
2.4. Reactions of compounds of periodic Group 8
The iron group presents at least one unusual aspect
of iodide chemistry in that older text books often state
that, although iron(II) iodide exists, iron(III) iodide
does not. This can easily be rationalised by invoking
the reducing power of iodide. However, the complex
ion [FeIIII4]− has been characterised structurally [16], so
this cannot be the whole explanation. The ability of
trimethylsilyl iodide to act as a simple iodide donor
might be a possibility, and so this aspect was also
investigated.
Fig. 4. ORTEP representation of the structure of [CrI2(NCMe)4]-
[CrI4(NCMe)2] (5).
Iron(II) iodide reacts with Me3SiI in MeCN to gener-
ate the known [17] [Fe(NCMe)6][FeI4], as judged by C,
H, N analysis and the IR and Moessbauer spectra. The
last was consistent with the presence of two high-spin
iron(II) species, one of them [FeI4]2−. However, a
similar reaction, but in thf (the only time we used this
solvent throughout the series of experiments), yielded a
crystalline material containing insufficient thf to be
formulated as [Fe(thf)6][FeI4]. The analysis for C, H, N,
and I seemed to fit best to a stoichiometric formula
FeI2(thf)2.5, which we were unable to rationalise. The
X-ray crystal structure analysis proved it to contain
iron in three different environments, with a structure
Fig. 5. ORTEP representation of the structure of [CrI3(NCMe)3]·
1/2Et2O (6).
best represented by the formula [Fe(thf)6][FeI4]1.2
-

202
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
Fig. 6. ORTEP representation of the structure of [MoI2(NCMe)4](I3) (7).
similar to that found in other RuII iodide complexes, in
[FeI3(thf)]0.8. The crystal structure showed considerable
disorder of the two monoanions, so the details will not
be discussed further. However, a similar structure,
[Fe(thf)6][FeI3(thf)]2 has been described [18]. We have
been able to repeat this last preparation and have
confirmed its structure by X-ray analysis. The remark-
able aspect of our first thf-containing product is that it
contains the iron(III) material, [FeIIII4]−, as well as the
two iron(II) species [Fe(thf)6]2− and [FeI3(thf)]+. This
conclusion is based upon the analytical and structural
data described above, and upon Moessbauer spec-
troscopy, but the yield was always small and not repro-
ducible. Neverthless, both iron(II) and chromium(II)
show that, under the appropriate circumstances, they
can be oxidised by Me3SiI to oxidation state III.
,
the range 2.69–2.73 A [21]. The average RuꢀN bond
,
length is 2.019(10) A, which is comparable to the
corresponding values of other RuII acetonitrile com-
plexes [19b,22]. The acetonitrile ligands are almost lin-
ear and it is conceivable that any non-linearity arises
from weak interactions between the I− moieties and
bound acetonitrile molecules in the cations
Table 7
,
Bond lengths (A) and angles (°) for [MoI2(NCMe)4]I3 (7)
Bond lengths
MoꢀN
MoꢀI(1)
NꢀC(2)
I(1)···I(3)c5
2.117(3)
2.730(1)
1.137(5)
3.967(1)
We prepared trans-[RuCl2(NCMe)4] [19] by reducing
a solution of RuCl3·xH2O in acetonitrile with Zn at
room temperature. The disadvantage of this method is
the possible contamination of the product with
ZnCl2·2MeCN. This ruthenium complex has been pre-
pared previously by a variety of methods, including
catalysed H2-reduction of RuCl3·3H2O in acetonitrile
and the photolytic displacement of 1,2-dibutylbenzene
from [{RuCl2(h6-o-Bu2nC6H4)}2] in acetonitrile. The re-
action of trans-[RuCl2(NCMe)4] and Me3SiI in acetoni-
trile under reflux gave orange crystals of diamagnetic
[RuI(NCMe)5]I·MeCN (9).The finding that [Ru-
I(NCMe)5]I is formed preferentially, rather than
[RuI2(NCMe)4] (coordinated acetonitrile instead of io-
dide), is consistent with the good leaving properties of
Bond angles
I(1)···I(3)c5ꢀI(2)c
NꢀMoꢀNc1
NꢀMoꢀNc3
NꢀMoꢀI(1)
153.37(2)
91.2(2)
88.8(2)
90.02(10)
89.98(10)
179.61(2)
Nc1ꢀMoꢀI(1)
MoꢀI(1)ꢀI(3)c5
Symmetry operations used to generate equivalent atoms, c1−x, y,
−z; c2x, −y, −z; c3x, −y, z; c4−x, −y+1, −z; c5−0.5+
x, −0.5+y, −1+z.
1
the iodide. The H NMR spectrum of 9 in MeCN-d3
displays a singlet at l=2.50 ppm due to the methyl
protons of the coordinated acetonitrile groups. This
compound is not the first iodoruthenium(II) complex,
among which is the venerable [Ru(NCMe)2(CO)2I2]
[20].
In the X-ray structure of complex 9 (Fig. 8 and
Tables 9 and 15), the octahedral [RuI(NCMe)5]+ cation
,
has an RuꢀI separation of 2.717(1) A, which is very
Fig. 7. ORTEP representation of the structure of [WI2(NCMe)4](I3) (8).

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
203
Table 8
lar facile reduction of [VCl3(NCMe)3] by an excess of
Me3SiI. It comprises isolated [Ru(NCMe)6]2+ cations
and the I42− anion. The octahedral geometry around
the ruthenium and the bond lengths to Ru are unexcep-
tional. The four-iodine unit is almost ideally linear (the
IꢀIꢀI angle is 179.18°) with a central bond length of
,
Bond lengths (A) and angles (°) for [WI2(NCMe)4]I3 (8)
Bond lengths
WꢀN
WꢀI(1)
I(2)ꢀI(3)
NꢀC(1)
C(1)ꢀNꢀW
2.084(5)
2.7297(6)
2.9204(7)
1.154(9)
177.1(5)
,
2.8402(10) A and two relatively long peripheral bonds
−
,
of 3.326 A. This is apparently best regarded as I2+2I .
Bond angles
N%ꢀWꢀN
N%%ꢀWꢀN
N%%ꢀWꢀI(1)
NꢀWꢀI(1)
NꢀC(1)ꢀC(2)
88.5(3)
91.5(3)
90.26(14)
89.74(14)
178.2(8)
Table 9
,
Bond lengths (A) and angles (°) for [RuI(NCMe)5]I·MeCN (9)
Bond lengths
RuꢀN(5)
RuꢀN(4)
RuꢀN(1)
RuꢀN(2)
RuꢀN(3)
RuꢀI(1)
N(1)ꢀC(1)
N(2)ꢀC(3)
N(3)ꢀC(5)
N(4)ꢀC(7)
N(5)ꢀC(9)
N(6)ꢀC(11)
N(7)ꢀC(13)
2.003(10)
2.007(10)
2.013(10)
2.024(10)
2.049(11)
2.717(1)
1.138(14)
1.148(14)
1.112(14)
1.145(13)
1.150(13)
1.18(3)
Symmetry transformations used to generate equivalent atoms: %x, −y,
z; %%−x, y, −z.
1.21(3)
Bond angles
N(5)ꢀRuꢀN(4)
N(5)ꢀRuꢀN(1)
N(4)ꢀRuꢀN(1)
N(5)ꢀRuꢀN(2)
N(4)ꢀRuꢀN(2)
N(1)ꢀRuꢀN(2)
N(5)ꢀRuꢀN(3)
N(4)ꢀRuꢀN(3)
N(1)ꢀRuꢀN(3)
N(2)ꢀRuꢀN(3)
N(5)ꢀRuꢀI(1)
N(4)ꢀRuꢀI(1)
N(1)ꢀRuꢀI(1)
N(2)ꢀRuꢀI(1)
N(3)ꢀRuꢀI(1)
90.0(3)
88.1(3)
176.9(4)
178.3(3)
89.3(3)
92.6(4)
86.7(3)
89.1(3)
88.4(3)
94.8(3)
91.2(2)
91.0(2)
91.4(3)
87.2(2)
178.0(2)
Fig. 8. ORTEP representation of the structure of [RuI(NCMe)5]I·
MeCN (9).
,
(I···HꢀCH2CN). A significantly short distance (3.484 A
,
vs. 3.7–4.0 A for normal C···C contacts) is found
between methyl carbon atoms of bound acetonitrile
ligands in neighbouring molecules, and this is probably
due to packing pressures in the crowded unit cell (Fig.
9).
The reaction of anhydrous RuCl3 with an excess of
Me3SiI in acetonitrile under reflux yielded brick-red
crystals of the diamagnetic ruthenium(II) complex
[Ru(NCMe)6](I4) (10). Of course, this is not an iodo
complex. The I2 produced at the same time is incorpo-
rated into the anionic part of the complex to form a
polyiodide. Reduction of RuIII to RuII is facile (E° for
red
RuIII–RuII= +0.24 V), so this is not unexpected. Even
[RuCl2(dmso)4] is obtained from
a
solution of
RuCl3·xH2O in dmso under N2 in the absence of any
other reductant [23]. A weak band in the IR spectrum
of 10 at 2291 cm−1 was assigned to w(CꢁN) and in the
1H NMR spectrum in MeCN-d3 solution the methyl
protons of the coordinated acetonitrile groups give rise
to a singlet at l=2.54 ppm.
Compound 10 (Fig. 9 and Tables 10 and 16) is
isostructural with [V(NCMe)6](I4) generated by a simi-
Fig. 9. ORTEP representation of the structure of [Ru(NCMe)6]I4 (10).

204
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
I42− groups. For example, in [{UO2(OH)[CO-
Table 10
,
Bond lengths (A) and angles (°) for [Ru(NCMe)6]I4·(10)
(NH2)2]3}2](I4) there is a linear symmetrical I4 with IꢀI
,
separations of 3.40 (peripheral) and 2.70 (internal) A,
Bond lengths
and hydrogen-bonding to all the iodine atoms [28]. An
example of an I42− unit completely free of covalent
interactions with other atoms is in the compound
Tl6PbI10 [29].
RuꢀN(2)
RuꢀN(1)
RuꢀN(3)
I(2)ꢀI(2)%%
N(1)ꢀC(1)
N(2)ꢀC(3)
N(3)ꢀC(5)
2.026(5)
2.027(5)
2.027(5)
2.8402(10)
1.137(8)
1.137(8)
1.142(8)
Contacts between the peripheral iodine atoms and
methyl carbon atoms (C(4) or C(6)) of neighbouring
[Ru(NCMe)6]2+ units (3.856“3.954 A) are consider-
,
Bond angles
ably shorter than the normal van der Waals’ distances
N(2)%ꢀRuꢀN(1)
N(2)ꢀRuꢀN(1)
N(2)ꢀRuꢀN(1)%
N(2)ꢀRuꢀN(3)%
N(1)ꢀRuꢀN(3)%
N(2)ꢀRuꢀN(3)
N(1)ꢀRuꢀN(3)
C(1)ꢀN(1)ꢀRu
C(3)ꢀN(2)ꢀRu
C(5)ꢀN(3)ꢀRu
88.8(2)
91.2(2)
88.8(2)
88.5(2)
92.9(2)
91.5(2)
87.1(2)
176.7(6)
178.2(5)
176.3(6)
3
,
of 4.15 A for I···C(sp ). This reflects the existence of
intermolecular ligand interactions in the lattice.
2.5. Reactions of compounds of periodic Group 9
When an excess of Me3SiI was added to a hot royal
blue solution of CoCl2 in MeCN, an immediate change
to deep green was observed and addition of diethyl
ether afforded a large amount of deep green plates.
These crystals are extremely moisture-sensitive and they
give a blue solution in thf. The X-ray crystal structure
determination (Fig. 10 and Tables 11 and 16) showed
the product, of empirical formula CoI2·3MeCN, to be
[Co(NCMe)6][CoI4] (11), previously prepared either by
dissolving anhydrous CoI2 in acetonitrile, or by the
reaction between cobalt and iodine in acetonitrile [30].
In comparison, the product of the reaction of CoCl2
with MeCN, formally CoCl2·3MeCN, is actually
[Co(NCMe)6][CoCl3(NCMe)] [31]. The effective mag-
netic moment of compound 11, veff=4.87 mB, is greater
that the spin-only value of 3.88 mB, which is not
uncommon for cobalt(II). Two sharp intense absorp-
tion bands at 2311 and 2281 cm−1 are consistent with
coordinated acetonitrile.
Symmetry transformations used to generate equivalent atoms: %−x,
−y, −z; %%−x+1, −y+1, −z−2.
The structure of complex 11 is not unexpected, and
both ions have been characterised structurally in other
situations. The dimensions found here are consistent
with reported data [31,32]. The asymmetric unit con-
tains two discrete cations, both lying on inversion cen-
tres, and one anion in a general position. There is no
evidence of intermolecular Co···Co interactions, the
Fig. 10. ORTEP representation of the structure of [Co(NCMe)6][CoI4]
(11).
These values are comparable to the bond lengths calcu-
,
shortest Co···Co distance being is 8.679 A, but inter-
lated for an isolated I42− ion (2.807 and 3.991 A,
,
molecular ligand interactions are present in the lattice.
For example, contacts between I(4) and carbon atoms
C(1) of neighbouring molecules are significantly shorter
respectively) [24,27]. The separations in the four-iodine
,
unit of [V(NCMe)6](I4) are 3.166, 2.846 and 3.516 A in
order along the chain. The ion I42− is quite common
and has been recognised in several different structural
situations. Generally there are clear covalent links of
some kind, as in [{Cu(NH3)4I4}n] [24,27], with a
CuꢀIꢀIꢀIꢀIꢀCu system bent almost to 90° at each ter-
,
(3.779 A) than the normal van der Waals’ contacts of
3
,
4.15 A for I···C(sp ).
When K2[IrCl6] was treated with a large excess of
Me3SiI in acetonitrile under reflux, a change to deep
brownish–red was observed and pale yellow cubic crys-
tals were obtained in moderate yield. The X-ray crystal
structure analysis showed the product to be fac-
[IrCl3(NCMe)3]·MeCN (12) (Fig. 11 and Tables 12 and
16)). In this reaction, IrIV was converted to IrIII. The
reduction of the metal was accompanied by the forma-
minal I, and
a
linear iodine arrangement. In
[{Ir(C5Me5)}2I6] [25] and some platinum complexes [26]
the arrangement is rather similar, though the angles of
90° open to more than 100° and may reach 126°.
Coordination, however, is not necessary to stabilise

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
205
Table 12
tion of I2, which gives the deep brownish–red colour in
,
Bond lengths (A) and angles (°) for [IrCl3(NCMe)3]·MeCN·(12)
the solution. The ion [IrIVCl6]2− has been reported to
be readily and quantitatively reduced to [IrIIICl6]3− by
Bond lengths
IrꢀN(3)
IrꢀN(2)
IrꢀN(1)
IrꢀCl(2)
IrꢀCl(1)
IrꢀCl(3)
N(1)ꢀC(1)
N(2)ꢀC(3)
N(3)ꢀC(5)
KI or sodium oxalate (E° for IrIV–IrIII= +0.87 V)
1.994(5)
1.999(5)
2.004(5)
2.3299(14)
2.3391(14)
2.3430(14)
1.129(7)
1.133(8)
1.137(7)
red
Table 11
,
Bond lengths (A) and angles (°) for [Co(NCMe)6][CoI4]·(11)
Bond lengths
I(1)ꢀCo(3)
I(2)ꢀCo(3)
I(3)ꢀCo(3)
I(4)ꢀCo(3)
Co(1)ꢀN(1)
Co(1)ꢀN(3)
Co(1)ꢀN(2)
Co(2)ꢀN(5)
Co(2)ꢀN(6)
Co(2)ꢀN(4)
N(1)ꢀC(2)
N(2)ꢀC(4)
N(3)ꢀC(6)
N(4)ꢀC(8)
N(5)ꢀC(10)
N(6)ꢀC(12)
2.6104(7)
2.6029(7)
2.6028(7)
2.6009(7)
2.098(4)
2.112(4)
2.113(4)
2.101(4)
2.117(4)
2.122(4)
1.128(6)
1.123(6)
1.134(6)
1.125(6)
1.133(6)
1.122(6)
Bond angles
N(3)ꢀIrꢀN(2)
N(3)ꢀIrꢀN(1)
N(2)ꢀIrꢀN(1)
N(3)ꢀIrꢀCl(2)
N(2)ꢀIrꢀCl(2)
N(1)ꢀIrꢀCl(2)
N(3)ꢀIrꢀCl(1)
N(2)ꢀIrꢀCl(1)
N(1)ꢀIrꢀCl(1)
Cl(2)ꢀIrꢀCl(1)
N(3)ꢀIrꢀCl(3)
N(2)ꢀIrꢀCl(3)
N(1)ꢀIrꢀCl(3)
Cl(2)ꢀIrꢀCl(3)
Cl(1)ꢀIrꢀCl(1)
90.09(19)
89.19(19)
90.20(19)
89.42(14)
178.58(15)
88.46(13)
88.83(14)
90.94(15)
177.72(13)
90.38(5)
178.33(13)
89.15(14)
89.33(14)
91.30(5)
Bond angles
92.67(5)
N(5)ꢀCo(2)ꢀN(4)
I(4)ꢀCo(3)ꢀI(3)
I(4)ꢀCo(3)ꢀI(2)
I(3)ꢀCo(3)ꢀI(2)
I(4)ꢀCo(3)ꢀI(1)
I(3)ꢀCo(3)ꢀI(1)
I(2)ꢀCo(3)ꢀI(1)
N(1)ꢀCo(1)ꢀN(3)%
N(1)ꢀCo(1)ꢀN(3)
N(1)ꢀCo(1)ꢀN(2)%
N(3)ꢀCo(1)ꢀN(2)%
N(1)ꢀCo(1)ꢀN(2)
N(3)ꢀCo(1)ꢀN(2)
N(5)ꢀCo(2)ꢀN(6)
N(5)ꢀCo(2)ꢀN(6)%%
N(5)ꢀCo(2)ꢀN(4)%%
N(6)ꢀCo(2)ꢀN(4)%%
N(6)ꢀCo(2)ꢀN(4)
85.79(16)
108.97(3)
111.47(3)
108.59(2)
106.85(2)
110.17(3)
110.77(3)
88.60(16)
91.40(16)
89.22(16)
91.03(16)
90.78(16)
88.97(16)
87.32(16)
92.68(16)
94.20(16)
86.90(16)
93.10(16)
[33] and the ionisation of Me3SiI in acetonitrile solution
has been suggested [34].
The most surprising feature in this reaction is that it
did not produce an iridium iodide complex. This is
consistent with the fact that the formation of iridium
complexes is normally very slow, and a catalyst might
be needed to drive the halogen-exchange reaction.
Other related systems, investigated previously by us and
cited above, where the chloride or oxo ligands are not
replaced by iodides involve the preparations of
[V2Cl3(thf)6]I. and of [V3Cl5(tmen)3]I.
Complex 12 adopts octahedral facial-configuration
as in fac-[IrCl3(PMe2Ph)3] and in fac-[IrCl3(tacn)]
(tacn=1,4,7-triazacyclononane) [35]. The asymmetric
unit cell contains acetonitrile of crystallisation in the
lattice. The geometry bond lengths and bond angles
about the iridium are very similar to values observed in
related compounds [36].
Symmetry transformations used to generate equivalent atoms: %−x,
−y+1, −z; %%−x−1, −y, −z.
3. Discussion
3.1. Metathesis reactions
We have shown that Me3SiI can be used as a meta-
thetical reagent to prepare transition-metal iodides.
However, it would appear that the method is best
applied to compounds of low oxidation state, because,
the iodine is able to take part in redox reactions that
Fig. 11. ORTEP representation of the structure of [IrCl3(NCMe)3]·
MeCN (12).

206
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
complicate the product. There is also the possibility
that oxidations might occur, but this seems only to be
a significant problem in very low oxidation states that
we have not investigated. A further complication stems
from the fact that iodide is a good leaving group
(otherwise, a weakly coordinated ligand) and so in the
presence of solvents which are themselves good donors
no iodides may be isolated. We have observed this in
several cases, but we have not extended our research in
any detail to solvents other than MeCN.
In principle, we should be able to predict whether
reduction will occur by considering the redox potentials
of appropriate couples. Unfortunately there are not
enough data on the required ligand sets and solvents to
make this a useful exercise, and extrapolating from
aqueous solutions often does not seem very helpful.
Our findings suggest that the reduction potentials of the
ZrIV–ZrIII and TiIII–TiII couples in acetonitrile are
probably significantly lower than those of TiIV–TiIII,
VIII–VII, MoIV–MoIII and WIV–WIII, since TiIV, VIII,
MoIV and WIV were all reduced by Me3SiI (standard
complexes studied in this work. We then subtracted
appropriate values of the metal covalent radii from the
observed metal–iodide distances (Table 13). The aver-
,
age effective iodide radius is 1.47 A, but individual
,
values range from 1.28 to 1.91 A. The values of the
MꢀI distances in [ZrI3(NCMe)4][ZrI5(NCMe)] (2.734–
2.877 A) [9] and in [VI2(salen)] (2.747 and 2.770 A) [2]
illustrate a further point. The sum of vanadium(IV) and
,
,
,
iodine(−I) ionic radii is (0.60+2.16)=2.76 A. The
length of a VꢀI covalent bond should be 1.22+1.33=
,
2.55 A. Consequently, the VꢀI separations in
[VI2(salen)] imply a high degree of ionic character [2].
In fact, most of the calculated iodine radii are longer
,
than the covalent radius (1.33 A), indicating significant
polarisation of the MꢀI bond. The large CrꢀI bond
length observed in [CrI2(NCMe)4] is due to the Jahn–
Teller effect, for a high-spin CrII complex.
3.3. Mechanisms of these reactions
The metathetical reactions described in this work
probably proceed via a four-centre transition state as
seems to often be the case.
reduction potentials, E° , in aqueous solution are
red
−0.37, −0.26 and 0.00 V for TiIII–TiII, VIII–VII and
TiIV–TiIII, respectively) [37]. Furthermore, the standard
reduction potentials for the reactions Ru3+ +e“Ru2+
and [IrCl6]2− +e“[IrCl6]3− are +0.24 and +0.87 V,
respectively [38], which confirm that both RuIII and IrIV
should be readily reduced (Table 13).
3.2. Metal–iodide interactions
In some cases, the iodide can also cause reduction of
the metal, as in the preparation of [WI2(NCMe)4]I3
from [WCl4(NCMe)2]. This is reminiscent of the spon-
taneous formation of [TiI2(NCMe)4]I3 from TiI4 in
MeCN [19].The iodine often becomes incorporated into
Iodine is a very soft atom, and it should adapt quite
dramatically to the circumstances in which it is bound.
We have compared the metal–iodide distances in the
Table 13
,
Comparison of effective iodide radii (A) in transition metal complexes
a
Effective iodide radius b
,
,
Metal–iodide complex
MꢀI distance (A)
Covalent radius of metal (A)
[TiI3(NCMe)3]
2.720(3)
2.725(3)
2.739(2)
2.6898(4)
2.734(3)–2.877(3)
2.747(1)
1.32
1.40
1.40
1.42
1.37
1.28–1.43
1.53
1.56
1.91
1.48
1.49
1.52
1.44
1.43
1.456
1.48
1.44
[TiI2(NCMe)4]I3
[ZrI3(NCMe)4][ZrI5(NCMe)][9]
[VI2(salen)][2]
1.32
1.45
1.22
2.779(1)
[CrI2(NCMe)4]
[CrI3(NCMe)3]
3.0816(4)
2.6550(15)
2.6628(16)
2.6922(16)
2.730(1)
1.17
1.17
[MoI2(NCMe)4]I3
[WI2(NCMe)4]I3
[Fe(thf)6][FeI3(thf)]2[18]
[RuI(NCMe)5]I
1.29
1.30
1.165
1.24
1.16
2.7297(6)
2.6205(14)
2.717(1)
[Co(NCMe)6][CoI4]
2.6042(7) c
a Lange’s Handbook of Chemistry, 13th ed., J.A. Dean (Ed.), McGraw-Hill, New York, p. 3/121.
b Calculated by subtracting the metal radius from the metal–iodide bond distance.
c Mean value.

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
207
Table 14
Details of crystal structure determinations
Compound
[TiI3(NCMe)3]·1/2Et2O [TiI2(MeCN)4]I3 (2)
[CrI2(NCME)4] (4)
[CrI3(MeCN)3]·1/2(Et2O) (6)
(1)
Empirical formula
Formula weight
Temperature (K)
C8H12I3N3O·1/2Ti
588.8
173(2)
C8H12I5N4Ti
846.62
173(2)
C8H12CrI2N4
470.02
173(2)
C6H9CrI3N3·0.5(C2H5O)
592.92
173(2)
,
Wavelength (A)
0.71073
monoclinic
C2/c (Number 15)
0.71073
monoclinic
C2/c (Number 15)
0.71073
orthorhombic
Pbca (Number 61)
0.71073
Monoclinic
C2/c (Number 15)
Crystal system
Space group
Unit cell dimensions
,
a (A)
18.672(4)
11.356(4)
19.699(9)
96.19(3)
3519(2)
8
15.5257(7)
10.6257(4)
13.4749(4)
99.611(3)
2191.77(14)
4
9.2302(3)
14.4060(4)
12.6202(4)k
18.4304(17)
11.3155(14)
16.6173(17)
97.200(6)
3438.2(6)
8
,
b (A)
,
c (A)
i (°)
V (A )
3
,
1678.11(9)
4
1.86
4.34
Z
Dcalc (mg m−3
Absorption
coefficient
)
2.22
5.74
2.57
7.43
2.29
6.04
(mm−1
)
q range for data
collection (°)
2–30
4.67–27.50
4.16–24.94
4.76–25.03
Index ranges
05h526, 05k515,
−235l523
−205h516, −135k510,
−175l517
−105h510, −175k516,
−135l514
−215h521, −135k511,
−195l515
Reflections collected 5244
9345
15459
7097
Independent
reflections
Reflections with
I\2|(I)
5114 [Rint=0.1521]
2809
2478 [Rint=0.0592]
1434 [Rint=0.046]
2940 [Rint=0.039]
2008
1182
2484
Completeness to q
% (°)
98.7 (27.50)
97.6 (27.50)
97.0 (27.50)
Data/restraints/param 5114/0/137
2478/0/85
1434/0/70
2940/9/89
eters
Final R indices
[I\2|(I)]
R indices (all data) R1 =0.154,
wR2=0.248
R1=0.0084,
wR2=0.200
R1=0.040, wR2=0.109
R1=0.050, wR2=0.116
R1=0.042, wR2=0.096
R1=0.052, wR2=0.010
R1=0.051, wR2=0.133
R1=0.062, wR2=0
have not searched for any of these postulated silicon
products.
a polyiodide ion. In other reactions oxidation of the
metal ion was observed. Examples are [CrI2(NCMe)4]-
[CrI4(NCMe)2] in which CrII was converted into CrIII,
and [Fe(thf)6][FeI4]1.2[FeI3(thf)]0.8 in which FeII was
converted into FeII and FeIII. The only feasible oxidant
in such case is the trimethylsilyl moiety. In neither case
was the product always reproducible and the isolated
yields of oxidised product were small, apparently being
a function of the quantities of solvent, crystallisation
temperature, etc.
’
M
M
n+ +Me3Si+“M(n+1)+ +Me3Si
n+ +I−“M(n+1)+ +1/2I2
A further possibility for the reduction involves inter-
mediate formation of the higher-oxidation state iodide
followed by homolysis of an MꢀI bond, generating the
lower-oxidation state product along with an iodine
atom and thence I2.
The mechanisms of oxidation and reduction are un-
clear. Assuming that the SiꢀI bond is significantly po-
larised, Me3Si+ I−, as the bond-length considerations
suggest, then the following scheme is plausible. This
would provide a route through which the iodide could
reduce the metal, or the silicon ‘cation’ [39] could
oxidise it. For the oxidation, the incipient silicon cation
would become a radical, which could dimerise to form
hexamethyldisilane or abstract a proton from the sol-
vent or another species to form trimethylsilane. We
4. Experimental
All operations were carried out under an inert atmo-
sphere in an argon-filled drybox or with standard
Schlenk techniques. Solvents were dried by standard
procedures and distilled under N2 prior to use. The
commercial products chromium dichloride, cobalt
dichloride, potassium hexachloroiridate, ruthenium

208
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
trichloride, ruthenium trichloride hydrate, titanium te-
trachloride, trimethylsilyl iodide, vanadium trichloride,
and zirconium tetrachloride were used without further
purification. Starting materials such as [VCl2(tmen)2]
[40] were prepared by literature methods as cited above
in the text.
weighed under argon, dissolved in the appropriate sol-
vent mixture (MeCN-d3+marker) in the glove box.
X-ray crystal data were collected by the 2q−ꢀ scan
method at 173(2) K using an Enraf–Nonius CAD4
diffractometer. During processing, the data were cor-
rected for absorption by semi-empirical „-scan meth-
ods. The structures were solved by direct methods in
SHELXS and refined by full-matrix least-square methods
in SHELXN [42]. All non-hydrogen atoms were refined
anisotropically. Diagrams of the molecular structure of
complexes were drawn with the ORTEP package [43].
Details of the structure determinations are shown in
Tables 14–16.
Microanalyses were carried out at the University of
Surrey using a Leenan CE 440 CHN elemental analyser
or by MEDAC, Brunel Science Centre, Surrey. IR
spectra were recorded on a Perkin–Elmer Spectrum
One model FT IR spectrometer, from Nujol mulls
prepared under dinitrogen and spread on KBr plates.
NMR spectra were obtained in the appropriate deuter-
ated solvents using a Bru¨ker 300 or 500 MHz instru-
ment. (13C, 1H)-HETCOR NMR, 1H{1H}-NOE and
4.1. Preparation of [TiI3(NCMe)3]·1/2Et2O (1)
1
variable temperature H NMR experiments were car-
ried out by Dr. Tony Avent, University of Sussex. Mass
spectra were taken by Dr. Ali Abdul-Sada, at the
University of Sussex, using a Kratos M580RF instru-
ment for FAB spectra (and 3-nitrobenzyl alcohol as a
matrix material), and a Fisons VG Autospec for EI
spectra. Mo¨ssbauer data were recorded at 77 K by
Professor Jack Silver, University of Greenwich and Dr.
D.J. Evans, John Innes Centre, Norwich. Spectra were
referenced against iron foil at 298 K. Determinations of
magnetic moments in solution were carried out in
MeCN-d3 by NMR spectroscopy using the Evans
method [41] with tetramethylsilane, cyclohexane or ben-
zene as the marker molecule. Samples were carefully
[TiCl3(NCMe)3] (3.3 g, 11.9 mmol) was completely
dissolved in 120 ml of acetonitrile (120 cm3) giving a
royal blue solution which was heated to approximately
60 °C for 30 min with a large excess of Me3SiI (8.5
cm−3, 12.0 g, 59.7 mmol). When it changed to deep
brownish–red, the reaction mixture was reduced to
approximately 25 cm3 by distilling off solvent and
layered with 95 ml of diethyl ether (1/4) and then
allowed to stand at room temperature (r.t.) for 3 days.
Deep brown crystalline material was filtered off,
washed with of diethyl ether (30 cm3) and dried under
vacuum. Yield 4.9 g (70%). When exposed to the air for
5–15 min, the crystals show extensive deliquescence,
Table 15
Details of crystal structure determinations
Compound
[MoI2(NCMe)4]I3 (7)
[WI2(NCMe)4]I3 (8)
[RuI(NC)5]I·(MeCN) (9)
Empirical formula
Formula weight
Temperature (K)
C8H12I5MoN4
894.66
173(2)
C8H12I5N4W
982.57
173(2)
C10H15IN5RuI(C2H3N)
601.19
173(2)
,
Wavelength (A)
0.71073
monoclinic
C2/m (Number 12)
0.71073
monoclinic
C2/m (Number 12)
0.71073
orthorhombic
Pnma (Number 62)
Crystal system
Space group
Unit cell dimensions
,
a (A)
15.489(3)
10.567(4)
6.725(2)
99.33(2)
1086.1(6)
2
15.4597(12)
10.5847(6)
6.7508(5)
99.097(5)
1090.8(1)
2
2.99
12.36
16.4318(7)
21.3680(12)
11.8990(7)
,
b (A)
,
c (A)
i (°)
V (A )
3
,
4177.9(4)
8
1.91
3.71
Z
Dcalc (mg m−3
)
2.74
7.70
Absorption coefficient (mm−1
)
q range for data collection (°) 2–35
4.69–24.99
−175h518, −125k512,
−75l58
4.55–22.94
−185h517, −235k522,
−105l513
Index ranges
05h524, 05k517,
−105l510
2555
Reflections collected
5721
12227
Independent reflections
Reflections with I\2|(I)
Completeness to q % (°)
Data/restraints/parameters
Final R indices [I\2|(I)]
R indices (all data)
2492 [Rint=0.218]
2083
1001 [Rint=0.070]
923
98.3 (24.99)
1001/0/49
R1=0.034, wR2=0.085
R1=0.037, wR2=0.086
2800 [Rint=0.075]
2493
94.2 (24.99)
Parameters 202
R1=0.052, wR2=0.124
R1=0.060, wR2=0.128
2492/0/48
R1=0.039, wR2=0.093
R1=0.049, wR2=0.098

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
209
Table 16
Details of crystal structure determinations
Compound
[Ru(NCMe)6]I4 (10)
[Co(NCMe)6][CoI4] (11)
[IrCl3(NCMe)3]·MeCN(12)
Empirical formula
Formula weight
Temperature (K)
C12H18I4N6Ru
854.99
173(2)
C12H18Co2I4N6
871.78
173(2)
C8H12Cl3IrN4
462.77
173(2)
,
Wavelength (A)
0.71073
triclinic
0.71073
monoclinic
P21/n (Number 14)
0.71073
monoclinic
P21/n (Number 14)
Crystal system
Space group
(
P1 (Number 2)
Unit cell dimensions
,
a (A)
8.1495(13)
8.3764(15)
8.9536(15)
89.297(12)
87.099(11)
79.665(11)
600.5(2)
1
12.3494(7)
12.1984(8)
17.3396(9)
9.4628(4)
15.4517(7)
10.1471(3)
,
b (A)
,
c (A)
h (°)
i (°)
k (°)
V (A )
92.112(4)
91.596(2)
3
,
2610.3(3)
4
2.22
1483.1(1)
4
2.07
Z
Dcalc (mg m−3
)
2.36
Absorption coefficient
5.80
6.01
9.52
(mm−1
)
q Range for data collection 4.02–30.07
(°)
3.85–28.00
3.92–25.02
Index ranges
−115h511, −105k511,
−125l510
−155h516, −165k511,
−225l516
−95h511, −175k518,
−105l512
Reflections collected
6411
15187
7699
Independent reflections
Reflections with I\2|(I)
Completeness to q % (°)
Data/restraints/parameters
Final R indices [I\2|(I)]
R indices (all data)
3224 [Rint=0.081]
2609
91.2 (30.07)
3224/0/106
R1=0.061, wR2=0.158
R1=0.074 wR2=0.170
5936 [Rint=0.053]
5256
94.3
5936/0/221
R1=0.039, wR2=0.099
R1=0.045, wR2=0.103
2611 [Rint=0.044]
2634
99.7 (25.02)
2611/0/145
R1=0.027, wR2=0.064
R1=0.032, wR2=0.067
they absorb water and quickly dissolve to give a pink-
ish-purple solution. The was recrystallised from aceto-
nitrile–diethyl ether (1/6) at r.t., to give deep brown flat
needles suitable for X-ray structural analysis. Found:
C, 16.2; H, 2.40; N, 7.45; Ti, 8.25; I, 66.4.
C6H9I3N3Ti·1/2Et2O requires: C, 16.3; H, 2.40; N, 7.15;
Ti, 8.13; I, 64.7%.
4.3. Preparation of [ZrI3(NCMe)4][ZrI5(NCMe)] (3)
To a hot (60 °C) colourless solution of ZrCl4 (1.20 g,
5.15 mmol) in acetonitrile (60 cm3) a large excess of
Me3SiI (4.2 cm3, 5.90 g, 29.5 mmol) was added. Imme-
diately after addition, the colour changed to light or-
ange–red. After approximately 30 min and the mixture
was concentrated to approximately 30 cm3 under vac-
uum. Orange crystals separated upon standing at
−20 °C for 3 days. Yield: 2.61 g, (72%). Found: C,
8.7; H, 1.10; N, 5.2; I, 73.1. C10H15I8N5Zr2 requires: C,
8.6; H, 1.1; N, 5.0; I, 72.4%.
4.2. Preparation of [TiI2(NCMe)4]I3 (2)
A yellow solution of [TiCl4(NCMe)2] (1.12 g, 4.12
mmol) in of acetonitrile (40 cm3), was heated to ap-
proximately 60 °C and allowed to react for 30 min with
a large excess of Me3SiI (3.5 cm3, 4.92 g, 24.6 mmol).
Immediately after addition, the colour changed to deep
brownish–red. The brick-red precipitate was filtered
off, washed with diethyl ether and dried under vacuum.
A second crop of crystalline solid was isolated from the
filtrate after cooling it overnight to −20 °C. Yield:
2.52 g (72%). Found: C, 11.7; H, 1.45; N, 6.8.
C8H12I5N4Ti requires: C, 11.4; H, 1.45; N, 6.6%. The
product was recrystallised from acetonitrile at −20 °C
to give brick-red needles suitable for X-ray structural
analysis.
4.4. Reaction of CrCl2 with Me3SiI
To a warm (60 °C) pale green suspension of CrCl2
(0.92 g, 7.49 mmol) in acetonitrile (60 cm3), an excess of
Me3SiI (3.6 cm3, 5.06 g, 25.3 mmol) was added. Imme-
diately after the addition, the colour changed to dark
greenish-brown and the reaction mixture was then
heated under reflux for 0.5 h. The solution was concen-
trated under vacuum to approximately 20 cm3. Upon
cooling to −20 °C overnight, a dark greenish-brown
crystalline solid precipitated out, which was filtered off
and dried under vacuum Found: C, 10.8; H, 0.98; N,

210
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
4.6%). A second batch of crystalline solid was obtained
from the filtrate by carefully layering with diethyl ether
(40 cm3). Yield: 2.35 g (67%). Recrystallisation from
acetonitrile layered with diethyl ether (1/2) afforded
greenish-blue crystals which were shown to be
[CrI2(NCMe)4] (4) by a single X-ray crystal structure
analysis. The X-ray structural analysis was carried out
on the crystals without previously isolating and drying
them to avoid loss of solvent. The product is extremely
air-sensitive. It burns in air, giving a pink smoke.
From the same recrystallisation pot, a few dark
brown crystals were also obtained, which were analysed
crystallographically as [CrI2(NCMe)4][CrI4(NCMe)2]
(5).
diethyl ether (40 cm3) and dried under vacuum. Yield:
1.49 g, (66%). Found: C, 10.7; H, 1.40; N, 6.1.
C9H13.5I5N4.5W requires: C, 10.8; H, 1.35; N, 6.3%.
Deep brown crystals suitable for X-ray structural anal-
ysis separated from a diluted acetonitrile solution at
−20 °C.
4.8. Preparation of FeI2·2.5thf
Iron(II) iodide was prepared by heating clean iron
powder, 55% hydriodic acid and methanoluntil no
more dihydrogen was evolved, filtering, and cooling the
filtrate to allow crystallisation of [Fe(MeOH)6]I2. This
was carefully heated in vacuum for 2 h at 100 °C to
leave FeI2. Iron(II) iodide (3.1 g, 10 mmol), thf (50
cm3) and Me3SiI (4.0 cm3, 5.8 g, 29 mmol) were heated
together under reflux, filtered hot, and allowed to cool
to r.t. for 14 h. Black crystals were filtered of and dried
in vacuum. Yield 3.6 g (71% calculated on the empirical
formula). Found: C, 24.1; H, 4.4, Fe 11.4.
C10H20FeI2O2.5 requires: C, 24.5, H, 4.1; Fe 11.3%.
Magnetic moment veff=5.36 mB calculated on iron
content.
4.5. Preparation of [CrI3(NCMe)3]·12Et2O (6)
[CrCl3(thf)3] (1.57 g, 4.19 mmol) was dissolved in
acetonitrile (50 cm3), giving a purple solution, which
was heated to approximately 60 °C with a large excess
of Me3SiI (3.0 cm3, 4.18 g, 20.9 mmol). Immediately
after addition, the colour changed to deep brownish–
red. The reaction mixture was heated under reflux for
approximately 30 min, then reduced to approximately
20 cm3 and then layered with diethyl ether (40 cm3).
After 3 days a dark brown crystalline material was
filtered off, washed with diethyl ether (20 cm3) and
dried under vacuum. Yield: 2.08 g (84%). (Found: C,
16.2; H, 2.40; N, 7.4. C8H14CrI3N3O0·5 requires: C,
16.2; H, 2.4; N, 7.1%). The product was recrystallised
from acetonitrile–diethyl ether (1/2) at r.t., to give deep
red–brown plates suitable for X-ray structural analysis.
4.9. Preparation of [RuCl2(NCMe)4]
To a brownish–red solution of RuCl3·xH2O (2.10 g)
in acetonitrile (80 cm3), an excess of Zn dust (1.20 g,
18.4 mmol) was added. The reaction mixture was al-
lowed to stir for 3 days at r.t. A gradual change from
brown–red to yellow was observed. The mixture was
then filtered to separate ZnCl2 and Zn, which were
discarded. The filtrate was evaporated to dryness to
produce a yellow solid, which was washed with diethyl
ether (30 cm3) and dried under vacuum for 2 h. Yield:
2.60 g, 7.74 mmol.
4.6. Preparation of [MoI2(NCMe)4]I3 (7)
Trimethlysilyl iodide (8.0 cm3, 11.2 g, 56 mmol) was
added to a dark red solution of [MoCl4(NCMe)2] (3.2
g, 10 mmol) in MeCN (100 cm3) and the mixture heated
to reflux. The volume was reduced to approximately 40
cm3 by solvent distillation and the residue stored at
−20 °C. A red powder was filtered of and recrys-
tallised from MeCNas brown needles. Yield 4.0 g
(45%). Found: C, 10.7; H, 1.40; N, 6.10. C8H12I5N4Mo
requires: C, 10.7; H, 1.40; N, 6.30%.
4.10. Preparation of [RuI(NCMe)5]I·MeCN (9)
To a warm (60 °C) yellow solution of [RuCl2-
(NCMe)4] (1.03 g, 3.06 mmol) in acetonitrile (80 cm3),
Me3SiI was added (1.4 cm3, 2.04 g, 10.2 mmol). The
reaction mixture was heated under reflux for approxi-
mately 30 min, when the colour changed from yellow to
orange. The reaction mixture was then filtered and the
residue on the sinter (cream-coloured material) was
discarded. The filtrate was concentrated to approxi-
mately 30 cm3 under vacuum. Orange crystals separated
upon standing at −20 °C for 4 days. Yield: 1.20 g
(65%). Found: C, 21.0; H, 2.8; N, 12.3. C10H15I2N5Ru
requires: C, 21.4; H, 2.7; N, 12.5%. Recrystallisation
from acetonitrile layered with diethyl ether (1/1) at r.t.
afforded orange plates suitable for X-ray structural
analysis.
4.7. Preparation of [WI2(NCMe)4]I3 (8)
To a warm (50 °C) brownish–red suspension of
[WCl4(NCMe)2] (0.94 g, 2.30 mmol) in acetonitrile (120
cm3), a large excess of Me3SiI (2.0 cm3, 2.76 g, 13.8
mmol) was added. The reaction mixture was heated
under reflux for approximately 1 h, after which the
volume of the solution was reduced to approximately
30 cm3 under vacuum. After 3 days at −20 °C dark
brown crystalline material was filtered off, washed with

G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
211
4.11. Preparation of [Ru(NCMe)6]I4 (10)
References
[1] A. Hills, D.L. Hughes, G.J. Leigh, R. Prieto-Alco´n, J. Chem.
Soc., Dalton Trans. (1993) 3609.
[2] D.L. Hughes, U. Kleinkes, G.J. Leigh, M. Maiwald, J.R.
Sanders, C. Sudbrake, J. Weisner, J. Chem. Soc., Dalton Trans.
(1993) 3093.
[3] P.B. Hitchcock, D.L. Hughes, G.J. Leigh, J.R. Sanders, J.S. de
Souza, C.J. McGarry, L.F. Larkworthy, J. Chem. Soc., Dalton
Trans. (1994) 3683.
[4] F.A. Cotton, S.A. Duraj, L.E. Manzer, W.J. Roth, J. Am.
Chem. Soc. 107 (1985) 3850.
[5] L. Vilas Boas, J. Costa Pessoa, in: G. Wilkinson, R.D. Gillard,
J.A. McCleverty (Eds.), Comprehensive Coordination Chem-
istry, vol. 4, Pergamon Press, Oxford, 1987, p. 453.
[6] S.I. Troyanov, A.I. Tursina, Koord. Khim. 10 (1991) 2102.
[7] S.I. Troyanov, G.V. Mazo, V.B. Rybakov, K.V. Rudkina, Ko-
ord. Khim. 16 (1990) 466.
[8] S.I. Troyanov, G.V. Mazo, Zh. Neorg. Khim. 33 (1988) 889.
[9] S.I. Troyanov, G.V. Mazo, M.A. Simonov, Vestn. Mosk. Univ.
Ser. Khim. 27 (1986) 575.
Anhydrous RuCl3 (0.73 g, 3.52 mmol) was suspended
in acetonitrile (80 cm3) and allowed to react with a
large excess of Me3SiI (2.5 cm3, 3.52 g, 17.6 mmol). The
reaction mixture was heated under reflux for approxi-
mately 1 h, when the colour changed to deep brown-
ish–red. Brick-red crystals separated upon cooling to
r.t. A second crop of microcrystalline solid was ob-
tained from the mother liquor after concentrating the
solvent to approximately 30 cm3. Yield: 2.05 g, (68%).
Recrystallisation from acetonitrile layered carefully
with diethyl ether (1/2) at r.t. afforded orange crystals
suitable for X-ray structural analysis. Found: C, 17.1;
H, 2.20; N, 9.4. C12H18I4N6Ru requires: C, 16.9; H,
2.10; N, 9.8%).
[10] D.G. Holah, J.P. Fackler Jr., Inorg. Chem. 5 (1966) 479.
[11] D.M. Halepoto, L.F. Larkworthy, D.C. Povey, B.J. Tucker,
G.W. Smith, J. Chem. Soc., Dalton Trans. (1990) 699.
[12] B.J. Hathaway, D.L. Holah, J. Chem. Soc. (1965) 537.
[13] F.A. Cotton, R. Poli, J. Am. Chem. Soc. 110 (1988) 830.
[14] A.V. Butcher, J. Chatt, J. Chem. Soc., Dalton Trans. (1975)
2639.
4.12. Preparation of [Co(NCMe)6][CoI4] (11)
A royal-blue solution of CoCl2 (0.94 g, 7.23 mmol) in
acetonitrile (50 cm3) at 60 °C was treated with a large
excess of Me3SiI (3.4 cm3, 4.78 g, 23.9 mmol). The
colour of the reaction mixture changed from blue to
green immediately after the addition. The mixture was
then heated under reflux for a further 30 min and then
reduced under vacuum to approximately 20 cm3. Deep
green plates, suitable for X-ray structural analysis were
obtained upon layering with 40 cm3 of diethyl ether at
r.t. Yield: 2.60 g, (82%).
[15] E.A. Allen, B.J. Brisdon, D.A. Edwards, G.W.A. Fowles, R.G.
Williams, J. Chem. Soc. (1963) 4649.
[16] S. Pohl, W. Saak, Z. Anorg. Allg. Chem. 523 (1985) 25.
[17] S. Pohl, W. Saak, Z. Naturforsch. 39b (1984) 1236.
[18] W. Saak, S. Pohl, Z. Anorg. Allg. Chem. 552 (1987) 186.
[19] (a) W.E. Newton, J.E. Searles, Inorg. Chim. Acta 7 (1973) 349;
(b) D.E. Fogg, S.J. Rettig, B.R. James, Can. J. Chem. 73 (1995)
1084.
[20] R.J. Irving, J. Chem. Soc. (1956) 2879.
[21] (a) M.A. Billadeau, W.T. Pennington, J.D. Petersen, Acta Crys-
tallogr. Sect. C 46 (1990) 1105;
4.13. Reaction of K2[IrCl6] with Me3SiI
(b) T. Ju¨stel, J. Bendix, N. Metzler-Nolte, T. Weyhermu¨ller, B.
Nuber, K. Wieghardt, Inorg. Chem. 37 (1998) 35;
(c) L.F. Dahl, D.L. Wampler, Acta Crystallogr. 15 (1962) 946;
(d) H. tom Dieck, W. Kollwitz, I. Kleinwa¨chter, W. Rohde, L.
Stamp, Transit. Met. Chem. 11 (1986) 361;
To a suspension of K2[IrCl6] (0.98 g, 2.03 mmol) in
acetonitrile (90 cm3), a large excess of Me3SiI (3.0 cm3,
4.22 g, 21.1 mmol) was added. The reaction mixture
was heated under reflux for approximately 4 h, and
then filtered through Celite to separate any unreacted
starting material. The deep brown-red filtrate was con-
centrated to approximately 20 cm3 under vacuum and
cooled to −20 °C for 6 days. Pale yellow cubes sepa-
rated out, which were shown crystallographically to be
[IrI3(NCMe)3]·MeCN (12). Yield: 0.42 g (45%).
(e) P. Homanen, M. Haukka, S. Luukkanen, M. Ahlgre´n, T.A.
Pakkanen, Eur. J. Inorg. Chem. (1999) 101.
[22] W. Luginbu¨hl, A. Ludi, A. Raselli, H.-B. Bu¨rgi, Acta Crystal-
logr. Sect. C 45 (1989) 1428.
[23] C.A. Helleren, D.Phil. Thesis, University of Sussex, 1998.
[24] L.J. Sæthre, O. Gropen, J. Sletton, Acta Chem. Scand. Ser. A 42
(1988) 16.
[25] E. Dubler, L. Linowsky, Helv. Chim. Acta 58 (1975) 2604.
[26] A. Millan, P.M. Bailey, P.M. Maitlis, J. Chem. Soc., Dalton
Trans. (1982) 73.
[27] K.D. Buse, H.J. Keller, H. Pritzkow, Inorg. Chem. 16 (1977)
1072.
[28] Y.N. Mikhailov, V.G. Kuznetsov, E.S. Kovaleva, Zh. Strukt.
Khim. 9 (1968) 710.
Acknowledgements
[29] A. Rabenau, H. Schulz, W. Stoeger, Naturwissenschaften 63
(1976) 245.
[30] B.J. Hathaway, D.G. Holah, J. Chem. Soc. (1964) 2400.
[31] F.A. Cotton, L.M. Daniels, G.T. Jordan IV, C.A. Murillo,
Polyhedron 17 (1998) 589.
We gratefully acknowledge financial support from
the British Council and CNPq (Brasil) to J.S.F., from
EPSRC to M.T., from BBSRC, and experimental help
from Dr. D.J. Evans (John Innes Centre) and Professor
J. Silver (University of Greenwich) for Moessbauer
spectroscopy.
[32] (a) M. Veith, B. Go¨dicke, V. Huch, Z. Anorg. Allg. Chem. 579
(1989) 99;

212
G.J. Leigh et al. / Inorganica Chimica Acta 330 (2002) 197–212
[39] (a) J.B. Lambert, J.A. McConnell, W. Schilf, W.J. Schulz, Jr., J.
(b) M.R. Pressprich, R.D. Willett, Acta Crystallogr. Sect. C 47
(1991) 1188.
[33] S. Fukuzumi, J.K. Kochi, Inorg. Chem. 19 (1980) 3022.
[34] G.A. Olah, S.C. Narang, Tetrahedron 38 (1982) 2225.
[35] M. Gajhede, K. Simonsen, L.K. Skov, Acta Chem. Scand. 47
(1993) 271.
Chem. Soc., Chem. Commun. (1988) 455;
(b) J.B. Lambert, L. Kania, W. Schilf, J.A. McConnell,
Organometallics 10 (1991) 2578;
(c) J.B. Lambert, S. Zhang, S.M. Ciro, Organometallics 13
(1994) 2430;
(d) G.A. Olah, A. Husain, B.P. Singh, A.K. Mehrotra, J. Org.
Chem. 48 (1983) 3667.
[36] (a) C.E. Briant, K.A. Rowland, C.T. Webber, D.M.P. Mingos, J.
Chem. Soc., Dalton Trans. (1981) 1515;
[40] (a) J.J.H. Edema, W. Stauthamer, F. van Bolhuis, S. Gam-
barotta, W.J.J. Smeets, A.L. Spek, Inorg. Chem. 29 (1990) 1302;
(b) J.S. de Souza, D.Phil. Thesis, University of Sussex, 1995.
[41] (a) D.F. Evans, J. Chem. Soc. (1959) 2003;
(b) D.F. Evans, G.V. Fazakerley, R.F. Philips, J. Chem. Soc. A
(1971) 1931;
(c) D.F. Evans, T.A. James, J. Chem. Soc., Dalton Trans. (1979)
723.
[42] (a) G.M. Sheldrick, SHELX-76 Program for Crystal Structure
Determination, University of Cambridge, 1976;
(b) R.D. Gillard, S.H. Mitchell, P.A. Williams, R.S. Vagg, J.
Coord. Chem. 13 (1984) 325;
(c) L.R. Gray, D.J. Gulliver, W. Levason, M. Webster, Acta
Crystallogr. Sect. B 38 (1982) 3079;
(d) J.-J. Bonnet, Y. Jeannin, J. Inorg. Nucl. Chem. 35 (1973)
4103;
(e) L.M. Vogler, B. Scott, K.J. Brewer, Inorg. Chem. 32 (1993)
898;
(f) K.R. Dunbar, S.O. Majors, J.-S. Sun, Inorg. Chim. Acta 229
(1995) 373;
(g) N. Kanematsu, M. Ebihara, T. Kawamura, J. Chem. Soc.,
Dalton Trans. (1999) 4413.
(b) G.M. Sheldrick, SHELXN, an extended version of SHELX
,
1977;
(c) G.M. Sheldrick, SHELXS, Program for Crystal Structure
Determination, University of Go¨ttingen, 1986;
(d) G.M. Sheldrick, Acta Crystallogr. Sect. A 46 (1990) 467;
(e) G.M. Sheldrick, SHELX-93, Program for Crystal Structure
Refinement, University of Go¨ttingen, 1993.
[37] (a) P.W. Atkins, Physical Chemistry, fifth ed., Oxford University
Press, Oxford, 1994, p. C21;
(b) M.S. Antelman, The Encyclopedia of Chemical Electrode
Potentials, Plenum Press, New York, 1982.
[38] F.A. Cotton, G. Wilkinson, In: Advanced Inorganic Chemistry,
fifth ed., Wiley-Interscience, New York, 1988, p. 887 and p. 915.
[43] ORTEP, Program for Diagrams, C.K. Johnson, Report ORNL-
3794, Oak Ridge Laboratory, Tennessee, USA, revised in 1971.