The synthesis of triangulo-trimetal complexes containing both iron(II) and vanadium(II)

Article abstract of DOI:10.1016/S0020-1693(01)00450-9

Reactions between a variety of vanadium(II) and iron(II) starting materials in tetrahydrofuran (thf) solution afforded di-, tri-, and tetra-nuclear complexes containing thf or N,N,N′,N′-tetramethylethylenediamine (tmen) as ligands. Most products were ionic and contained the two metals in different structural units, cation or anion. Trinuclear clusters formulated as [V_3-x-Fe_xCl₅(tmen)₃][BPh ₄](x = 0-3) were obtained and characterised by C, H, N and metal analyses, ¹H NMR, M?ssbauer, and positive ion FAB and electrospray mass spectroscopies, and magnetic moment measurements in solution. We characterised fully by low-temperature single-crystal X-ray diffraction analysis the mixed-metal triangulo-complex salt [V₂Fe(μ-Cl)₃(μ₃-Cl)₂(tmen) ₃][BPh₄].

Full text of DOI:10.1016/S0020-1693(01)00450-9

www.elsevier.nl/locate/ica
Inorganica Chimica Acta 319 (2001) 147–158
The synthesis of triangulo-trimetal complexes containing both
iron(II) and vanadium(II)
David J. Evans b, Peter B. Hitchcock a, G. Jeffery Leigh a,*, Brian K. Nicholson c,
Antonio C. Niedwieski d, Fa´bio S. Nunes d, Ja´ısa F. Soares d
a School of Chemistry, Physics and En6ironmental Science, Uni6ersity of Sussex, Falmer, Brighton BN1 9QJ, UK
b Department of Biological Chemistry, John Innes Centre, Norwich Research Park, Colney, Norwich NR4 7UH, UK
c Chemistry Department, Uni6ersity of Waikato, Hamilton, New Zealand
d Departamento de Qu´ımica, Uni6ersidade Federal do Parana´, Centro Polite´cnico, 81531-990 Curitiba PR, Brazil
Received 9 February 2001; accepted 2 April 2001
Abstract
Reactions between a variety of vanadium(II) and iron(II) starting materials in tetrahydrofuran (thf) solution afforded di-, tri-,
and tetra-nuclear complexes containing thf or N,N,N%,N%-tetramethylethylenediamine (tmen) as ligands. Most products were ionic
and contained the two metals in different structural units, cation or anion. Trinuclear clusters formulated as [V3−x
-
FexCl5(tmen)3][BPh4](x=0–3) were obtained and characterised by C, H, N and metal analyses, 1H NMR, Mo¨ssbauer, and
positive ion FAB and electrospray mass spectroscopies, and magnetic moment measurements in solution. We characterised fully
by low-temperature single-crystal X-ray diffraction analysis the mixed-metal triangulo-complex salt [V2Fe(m-Cl)3(m3-
Cl)2(tmen)3][BPh4]. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Crystal structures; Iron(II) complexes; Vanadium(II) complexes; Triangulo-trimetal complexes
1. Introduction
One complex we intended to use as starting material
was [VCl2(tmen)2] (tmen=N,N,N%,N%-tetramethyl-
ethane-1,2-diamine) [4], but we found [5]that it changes
spontaneously but reversibly under very mild condi-
Iron and either molybdenum or vanadium are
present in the putative active sites of the two most
common nitrogenases [1,2]. In addition, it is evident
that polymetal species are probably involved in the
reduction of N2 by some chemical systems [2]. Despite
what is known about nitrogenase active centres and
what is surmised about these chemical nitrogen-fixing
systems, there are no examples of synthetic pre-formed
metal clusters or assemblies that react with dinitrogen.
This persuaded us that it would be worthwhile to try to
prepare simple mixed-metal cluster complexes, related
to some we had already characterized [3] but containing
both iron and vanadium, that might, under appropriate
conditions, react with dinitrogen.
tions to
a trinuclear complex, [V3(m-Cl)3(m3-Cl)2-
(tmen)3]+ [5]. It seems reasonable to assume that the
reaction of a mononuclear compound with a dinuclear
material might be involved in its formation. Recent
synthetic and mechanistic studies of this reaction sup-
port our hypothesis [6a]. A related reaction route might
therefore allow us to prepare trinuclear mixed iron–
vanadium clusters. We have already shown that the
reaction of iron(II) chloride with tmen forms intercon-
vertible complexes trans-[FeCl2(tmen)2], [{FeCl-
(tmen)}2(m-Cl)2], and [Fe3(m-Cl)3(m3-Cl)2(tmen)3]+ [6].
We now describe efforts to synthesise mixed trinuclear
species [M3(m-Cl)3(m3-Cl)2(tmen)3]+ where M may be
both iron or vanadium. To achieve this we used a
selection of mono- and di-nuclear species of iron and
vanadium as starting materials.
* Corresponding author. Tel.: +44-1273-606755; fax: +44-1273-
67849.
E-mail address: g.j.leigh@sussex.ac.uk (G.J. Leigh).
0020-1693/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(01)00450-9

148
D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
2. Results and discussion
[VCl3(thf)3] is not reduced by SnCl2 in tetrahydrofuran
(thf), though FeCl3 is [7]. Our results imply that the
opposite process, reduction of SnIV by VII, is sponta-
neous. The order of magnitudes of decreasing reducing
power of the various redox couples in thf appears to be
VII/VIII\SnII/SnIV\FeII/FeIII. We were not able to
obtain [Fe2(m-Cl)3(thf)6][BPh4] from the [SnCl5(thf)]−
salt by metathesis using Na[BPh4] and so we were not
able to carry out reactions that were not complicated
by the oxidising effect of SnIV on VII.
2.1. Reactions in6ol6ing [Fe2(v-Cl)3(thf)6][SnCl5(thf)]
This material [7] contains a {M2(m-X)3}+ (X=Cl or
Br) core, frequently encountered in first-row transition
metal ion chemistry [8]. Its reactions with the VII spe-
cies are fast, giving a remarkable succession of colours
for 15–20 min. It transpires that the products formed
are determined by the redox chemistry of the SnIV/VII
pair.
We formulate the bright green product
2 as
[VI2ICl3(thf)6][SnCl5(thf)], consistent with the microanal-
ysis (see Section 4), though this contains a potentially
reactive VII/SnIV couple. Nevertheless, the analysis fa-
vours this formulation rather than [VIIICl2(thf)4]-
[SnIVCl5(thf)], which has also been described. Surpris-
ingly, this solid has been stated to be bright-green, an
unexpected colour for a VIII–thf complex, though it
was never characterised structurally [7].
Reactions of the diiron salt in thf at room tempera-
ture with [VCl2(tmen)2] [4] (proportion 2Fe:1V) or with
[V2Cl3(thf)6][BPh4] [9] (1Fe:1V) produce air-sensitive
magenta solutions. This colour is characteristic of vana-
dium(III)-thf adducts [10]. When concentrated under
vacuum and kept at −20°C for 4–6 days, both solu-
tions yielded air-sensitive red crystals and a white pow-
der. Fractional recrystallisation over a period of 3
months gave the following products: paramagnetic pink
or purple (1) and bright green (2) well-formed crystals,
a colourless diamagnetic crystalline solid (3), and a
white powder (4). The products were all isolated in very
small amounts and frequently as mixtures that were
hand-separated.
The microanalysis of 3 is consistent with the formula-
1
tion ‘Sn(BPh4)2’, but the H NMR spectrum in CD2Cl2
(Fig. 1) also reveals a pair of resonances assignable to
thf protons. The proportion [BPh4]−: thf is 1.00:0.75, as
judged by spectrum integration. This thf appears to be
lattice solvent which is easily lost upon drying, when
the crystals become opaque and white. The 13C NMR
spectrum, obtained from the same diluted dichloro-
methane solution used for the 1H NMR analysis,
showed only traces of thf at l ca. 25 and 67,
Materials (1) had analyses consistent with the empir-
ical formula [VCl3(thf)3], though their colour ranged
from pink to purple. Probably they are formed in the
oxidation of VII by the SnIV. It has been noted that
Fig. 1. 1H NMR spectrum of ‘Sn(BPh4)2·xthf’ (3) in CD2Cl2 (l-scale, 270 MHz). The spectrum was recorded from −40 to 120 ppm but no signals
were found outside the range 0“10 ppm.

D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
149
Table 1
57Fe Mo¨ssbauer parameters for complexes 5–7 and for [Fe3Cl5(tmen)3][BPh4] (recorded at 77 K, referenced against iron foil at 298 K)
a
a,b
Compound
l
(mm s−1
)
q.s. a (mm s−1
)
Y
(mm s−1
)
% Total area/constraint
applied a,c
[V3Cl5(tmen)3]2[FeCl4] (5)
[Fe3Cl5(tmen)3][BPh4] (Ref. [6])
1.03 (1)
1.14 (1)
2.57 (1)
2.00 (1)
0.28 (1)
0.22 (1)
100
100
[M3Cl5(tmen)3][BPh4] (6, sample 1) 1.12 (1), 1.13 (1), 1.13 (1) 2.58 (7), 2.20 (3), 1.86 (4) 0.17 (4), 0.15 (4), 0.15 (2) 18, 49, 33
[M3Cl5(tmen)3][BPh4] (6, sample 2) 1.12 (1), 1.14 (1), 1.13 (1) 2.43 (1), 2.12 (1), 1.83 (1) 0.13 (1), 0.14 (1), 0.13 (1) 17/1, 54/3, 34/2
[V2FeCl5(tmen)3][BPh4] (6, sample 3) 1.106(3)
2.258(7)
2.55 (1)
0.190(6)
0.17 (1)
100
100
[FeCl2(dppe)] (Ref. [18] and this
work)
0.74 (1)
0.89 (1)
[V3Cl5(tmen)3]2[FeCl4(dppe)] (7)
2.32 (1)
0.27 (1)
100
a Numbers in parenthesis correspond to the experimental error in the last significant figure(s).
b Half-width at half maxima.
c See text.
together with very intense phenyl carbon resonances in
of [FeCl4]2− in a number of different salts [13] at 77–80
K (i.s. 1.00–1.06 mm s−1; q.s. 2.62–3.07 mm s−1). The
low solubility of the product in organic solvents made
its purification difficult.
the anion (l 125–140). No clear evidence was found in
1
either H or 13C NMR spectra for p-complex formation
between parallel phenyl rings of two [BPh4]− anions
and the metal(II) cation [11]. Compound 4 was iden-
tified as the well known tetranuclear compex
[Fe4Cl8(thf)6], formed frequently from thf solutions of
iron(II) chloride [12].
Six moles of VII complex should interact with one
mole of FeII complex to form (5), but the atom propor-
tions used in the reaction were approximately 1:1.
Consistent with the presence of an excess of iron(II) in
the reaction mixture, [{FeCl(tmen)}2(m-Cl)2] [6b] was
obtained as a by-product. No attempt was made to
isolate any [FeCl2(tmen)2] [6b] or to determine the
yields of the Fe–tmen complexes, but the formation of
[{FeCl(tmen)}2(m-Cl)2] implies that tmen has been lost
from the vanadium starting material, as expected if the
trinuclear vanadium cation is generated [5]. Again, no
bimetallic trinuclear complex appeared to have been
formed.
The following balanced equation can be written for
this reaction.
2[V2Cl3(thf)6][BPh4]+2[Fe2Cl3(thf)6][SnCl5(thf)]
“2[VCl3(thf)3]+[V2Cl3(thf)6][SnCl5(thf)]+‘Sn(BPh4)2’
+[Fe4Cl8(thf)6]
Apparently no hoped-for bimetallic Fe/V species was
generated in this system.
To test whether [Fe4Cl8(thf)6] can generate mononu-
clear species capable of reacting with dinuclear
[V2Cl3(thf)6][BPh4], the two complexes were allowed to
react in refluxing thf (1Fe:1V atom ratio) in the pres-
ence of tmen. Upon the addition of the amine, the
boiling solution changed from light to deep green.
Filtration, concentration under vacuum and cooling to
−20°C for 5 days did not afford any solid, suggesting
that [MCl2(tmen)2] (M=VII or FeII) and [{FeCl-
(tmen)}2(m-Cl)2], which normally crystallise under these
conditions, were probably not formed. Long deep-green
needles (6) (sample 1) were obtained after slow diffu-
sion of hexane into the thf solution at room tempera-
ture. The product contained Fe and V, in the atomic
proportion 1.5:1 (V:Fe), but C, H, N analysis was
consistent with the rather less precise formulation
[V3−xFexCl5(tmen)3][BPh4], x=0–3 (see Section 4).
The analytical data and our previous results [5,6] with
single-metal systems are consistent with the conclusion
that trinuclear cations had been formed, but possibly in
a mixture containing all combinations of the two metals
within the defined stoichiometry.
2.2. Reactions in6ol6ing [Fe4Cl8(thf)6]
Because of the very long Fe···Fe distances in the
crystals of this compound, it may dissociate in solution
[12b], producing reactive dinuclear intermediates capa-
ble of reacting with mononuclear vanadium materials.
When a thf suspension of this complex reacted with
[VCl2(tmen)2] (atomic ratio 1.1 Fe: 1V), a powder (5)
resembling [V3Cl5(tmen)3]I5 in colour (turquoise), habit,
and solubility was formed. A plausible formulation,
consistent with the microanalysis, is [V3Cl5(tmen)3]2-
[FeCl4]. The {FeIIClx(thf)y} units generated from
[Fe4Cl8(thf)6] in solution might have accepted chloride
from [VCl2(tmen)2], triggering the formation of the
trinuclear vanadium core [5].
The large line widths in the solid state 57Fe Mo¨ss-
bauer spectra of 5 (Table 1) suggested that the sample
was contaminated with iron-containing impurities, but
the isomer shift and quadrupole splitting are typical of
high-spin iron(II) and consistent with reported values

150
D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
2.4. Physical studies of the triangulo-trimetal species
2.3. Reactions in6ol6ing [{FeCl(tmen)}2(v-Cl)2]
2.4.1. Mass spectroscopy
When solutions of this complex and of [VCl2(tmen)2]
in thf (Fe/V atomic ratio 1:1) were mixed at room
temperature, no reaction was observed even after stir-
ring for 20–24 h. However, the addition of a solution
of Na[BPh4] in thf (Fe/V metal:tetraphenylborate 3:1)
caused an immediate change from blue to grass-green.
A very fine suspension of a white powder (probably
NaCl) was filtered off, leaving a clear green solution.
Layering with a small amount of hexane afforded large
deep green prisms which were again formulated as
[V3−xFexCl5(tmen)3][BPh4] (x=0–3) (6, sample 2),
based on C, H, N and metal analyses.
This preparation was perfectly reproducible. Slightly
different crystallisation conditions (solvent proportions
and solution concentration) produced distinct crystal
shapes (thick plates, needles or long prisms), but the
habit of all crystals in a given batch was normally the
same. Sometimes the crystals contained thf. The
batches all had similar compositions, as shown by the
data reported in Section 4. Crystals of 6 suitable for a
good X-ray analysis (sample 3) were finally isolated
after a new preparation from [{FeCl(tmen)}2(m-Cl)2],
[VCl2(tmen)2] and Na[BPh4] in thf (Fe:V atomic ratio
1.6:1; Fe+V:[BPh4]− 3:1) at room temperature. The
crystals again contained both Fe and V. Our problem
was to demonstrate whether these materials 6 contained
distinct cocrystallised tri-iron and trivanadium species,
or whether they were the sought-after mixed-metal tri-
nuclear species.
Electrospray mass spectrometry (ESMS) is a mild
technique ideally suited for characterising ions in solu-
tion. When a solution of 6, sample 2, in 1,2-C2H4Cl2
was injected, a clean spectrum was obtained which
showed no significant signals other than a complicated
envelope of peaks between m/z 675 and 695, Fig. 2.
These arise from the isotopomers associated with all of
the species [V3−xFexCl5(tmen)3]+, x=0–3, showing
that the sample contains mixed-metal cores, and is not
simply a mixture of the V3 and Fe3 materials. ESMS
cannot be used as a quantitative technique under these
conditions, and the relative intensities of peaks in the
pattern were not completely reproducible. The signal
strengths also consistently varied with time during a
single injection, with the V-rich species being more
dominant in the early part of the acquisition and the
Fe-rich ones later, for reasons that are not apparent.
However, it is clear that all mixed species are present in
6, sample 2, in solution.
When the experiment was repeated using MeCN as
solvent and mobile phase, spectra were less clean with
many peaks at lower masses suggesting fragmentation
of the M3 aggregates, but the mixed V/Fe parent ions
were still detectable. If the mass spectrometer was not
thoroughly flushed, then peaks assignable to [M3Cl5−n
-
(OH)n(tmen)3]+ or to [M3Cl5−n(OEt)n(tmen)3]+ were
found, indicating reactivity with H2O or EtOH, respec-
tively, lingering in the tubing. The V-rich species domi-
nated this type of process, suggesting either that they
undergo Cl−/OR− exchange more readily, or that the
OR− adducts are less stable for the Fe-rich aggregates.
In fact, related triangulo-alkoxotrivanadium species
have been structurally characterized [5b], but none of
their iron analogues is known.
In the FAB mass spectra of 6, samples 1 and 2, the
‘molecular’ ion for [V3Cl5(tmen)3]+ (m/z 678) was al-
ways present. We also always observed a peak at m/z
553, which we have accounted for previously [6b] by a
reaction between the ‘all-iron’ trinuclear cation and the
alcohol matrix (3-nitrobenzylalcohol) used for sample
preparation. This fragment, possibly {(Htmen)2-
[BPh4]}+, was also observed when the all-iron cation
was submitted to FAB-MS analysis, but never arose
from the all-vanadium cation. The m/z 678 and 553
fragments were also observed when a ground single
prismatic crystal of 6, sample 2, was submitted for
MS-FAB analysis, and this supports the hypothesis of
co-crystallisation. No evidence was found in the FAB+
spectra for mixed-metal cations with the {V2Fe} and
{VFe2} cores.
Fig. 2. The parent cation region of the ESMS of 6. The peaks
correspond to a mixture of all species [V3−xFexCl5(tmen)3]+, x=0–
3. Analysis is complicated by overlapping isotope patterns, but the
peaks at m/z 678, 683, 688 and 693 are characteristic of the V3, V2Fe,
VFe2 and Fe3 combinations, respectively.
The FAB-MS is a much harsher technique and the
ablation process may be inducing selective reaction of
the Fe-containing species with the matrix (c.f. the re-

D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
151
Fig. 3. Plot of the reciprocal of the atomic magnetic susceptibility (−1) and effective magnetic moment (veff) versus absolute temperature (T) for
A
[M3Cl5(tmen)3][BPh4] (M=VII and FeII) (6) sample 2. Diamagnetic correction: 1.985×10−4 emu atom−1 MII. The observed changes were
reversible with the increase or decrease in temperature.
sults discussed in the previous paragraphs). Another
possibility is that composition varies with phase, since
ESMS samples are from solution whereas FAB-MS
samples are from the solid, but the ESMS results are
more consistent with the Mo¨ssbauer and X-ray crystal-
lography results which both involve the solid state.
exactly the same composition. They are all likely to
contain more or less mixed-metal cations.
For our preferred curve fitting, (i) the relative intensi-
ties of the signals in the non-constrained spectrum
(sample 1) are compatible with the areas in the con-
strained spectrum of sample 2; (ii) the baseline devia-
tions from zero are small and random in all cases; and
2
2.4.2. Mo¨ssbauer spectroscopy
(iii) the final  value, which reflects the overall quality
The Mo¨ssbauer parameters obtained from spectra of
samples 1, 2 and 3 of 6 at 77 K are shown in Table 1.
The Mo¨ssbauer spectrum of of 6, samples 1 and 2,
(Table 1) were convincingly deconvoluted as three
quadrupole doublets, which we assign to {Fe3}, {Fe2V}
and {FeV2} aggregates. The isomer shifts are all very
similar to the value obtained for [Fe3Cl5(tmen)3][BPh4],
indicating high-spin iron(II) nuclei in trimetallic cations
[6b]. The areas of the three Mo¨ssbauer doublets quoted
in Table 1 are in the proportions 1:3:2, with no con-
straints applied to the analysis. Because the probable
components of the products differ only in the V/Fe
ratio, these areas reflect the number of iron nuclei in
the different complexes. The complexes [Fe3Cl5-
(tmen)3][BPh4] and [V3Cl5(tmen)3][BPh4] are isostruc-
tural and have remarkably similar dimensions [5,6], and
should cocrystallise with few packing problems. Similar
trinuclear cations such as [V3Cl3(OMe)2(tmen)3]+ and
[V3Cl4(OMe)(tmen)3]+ also cocrystallise [5].
Complex 6, sample 3, has similar Mo¨ssbauer parame-
ters to those of samples 1 and 2. However, only a single
species seemed to be present, giving rise to a single
quadrupole doublet. Thus sample 3 was the most likely
to contain a single mixed-metal species. Microanalysis
implied the presence of the trinuclear cation and te-
traphenylborate, but also of protonated tmen (see be-
low). Clearly the three samples are similar but not of
of the fitting model, was equally low for both the
constrained and the non-constrained models. The fits
are only consistent with the hypothesis that 6, samples
1, 2, and 3, are crystallographic mixtures of FeII-con-
taining trinuclear cations differing only in the relative
amounts of iron they contain, and that in sample 3 one
such species, not the tri-iron species, is clearly
predominant.
2.4.3. NMR spectroscopy
The product 6, sample 2, and the homonuclear clus-
ters [V3Cl5(tmen)3][BPh4] and [Fe3Cl5(tmen)3][BPh4] are
not diamagnetic and therefore they do not produce
well-resolved 1H NMR spectra. However, each spec-
trum is characteristic and can be used to distinguish the
complexes in solution. When crystals of the homonu-
1
clear iron and vanadium clusters were mixed, the H
NMR spectrum of their solution in [2H8]-thf was re-
markably similar to that obtained from 6. This also
supports the hypothesis that 6 is a crystalline mixture of
trinuclear species.
2.4.4. Magnetic properties
The magnetic behaviour of the product 6, sample 2
was analysed in [2H8]-thf solution by the Evans method
[14] in the temperature range +50 to −100°C (323–

152
D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
2.5. The X-ray diffraction analysis of
173 K) (Fig. 3). In order to simplify the analysis of the
data, it was assumed that the sample used was an
equimolar mixture of [Fe3Cl5(tmen)3][BPh4] and
[V3Cl5(tmen)3][BPh4].
[M3Cl5(tmen)3][BPh4] (M=FeII and/or VII) (6)
A single crystal analysis of 6, sample 2, carried out at
room temperature, revealed a highly disordered struc-
ture. Although the presence of triangulo-trinuclear
cations was evident the structure could not be refined
satisfactorily and, of course, there was no possibility of
distinguishing between iron and vanadium in the three
equivalent sites.
The tris(vanadium(II)) complex is known to be anti-
ferromagnetic [5] and the tris(iron(II)) analogue ferro-
magnetic [6b]. In the all-vanadium complex, the
effective magnetic moment per metal ion at 20–25°C is
lower than the ‘spin-only’ value of 3.87vB and decreases
with decrease of temperature [5]. For the trinuclear iron
compound, the effective magnetic moment is compat-
ible with the ‘spin-only’ value of 4.90vB per metal atom
at 20–25°C and increases with the decrease of the
temperature. If we assume that no intermolecular inter-
action occurs in solution, and that the magnetic field
generated by the complexes in the mixture has the same
nature and intensity as those produced in the pure
samples, an ‘average molecule’ of [M3Cl5(tmen)3][BPh4]
would then be expected to show a magnetic moment
intermediate between the two extremes. This seems to
be the behaviour shown in Fig. 3.
The measured veff per metal atom is 3.95vB at 22°C,
slightly higher than the ‘spin-only’ value for three un-
paired electrons per metal ion (as expected for an
‘average’ species) and it increases slowly as the temper-
ature was lowered. Neither the very low veff characteris-
tic of the ‘all-vanadium’ trinuclear complexes [5] nor
the high values given by the {Fe3Cl5}+ core [6b] were
obtained, nor a temperature-dependence characteristic
of either. The experimental evidence is consistent with
the hypothesis of a mixture and possibly of heterometal
clusters.
Single crystal X-ray diffraction analysis at 173 K of
6, sample 3, confirmed that the unit cell contained
co-crystallised [V2FeCl5(tmen)3]+ and [Htmen···thf]+ as
tetraphenylborate salts (Fig. 4). This formulation agrees
well with the microanalytical data (see Section 4). The
presence of two VII and one FeII centres disordered
equally over the metal positions in the cation was
strongly indicated by the final occupancy factors calcu-
lated for these sites during structure refinement. The
best molecular model to agree with the experimental
structure factors was clearly consistent with each metal
site having an occupancy factor of 2/3 for vanadium
and 1/3 for iron. No alternative model, such as the ones
provided by the {V3}, {VFe2} or {Fe3} combinations,
was as satisfactory. This result, together with the over-
all quality of the X-ray data and final refinement
parameters presented in Table 2 points to a confident
structural characterisation of the mixed-metal species
[V2FeCl5(tmen)3]+.
The cation contains the normal {M3Cl3} planar six-
membered ring found in other trinuclear complexes
[5,6,15,16], capped by two chlorine atoms so that each
metal atom has approximately octahedral coordination,
completed by the nitrogen atoms of a chelating tmen.
Fig. 4. ORTEP plot of the unit cell contents in [V2FeCl5(tmen)3][Htmen···thf][BPh4]2 (complex 6, sample 3). The metal sites in the triangulo-complex
are labelled V(1), V(2) and V(3), corresponding to the major site occupancy, but they actually contain two VII and one FeII centres disordered
equally over the three positions (see text). Occupancies are 2/3V+1/3Fe.

D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
153
[V3Cl5(tmen)3]+ and [Fe3Cl5(tmen)3]+ [5,6]. The aver-
age bond lengths and angles in 6 are not very different
from the dimensions of the two homonuclear cations.
Table 2
Crystal data and structure refinement for [V2FeCl5(tmen)3]-
[Htmen···thf][BPh4]2
Empirical formula
Formula weight
C76H113B2Cl5FeN8OV2
1527.34
2.6. Other preparati6e studies
Temperature of data collection (K)
173(2)
,
Wavelength (Mo Ka) (A)
Crystal system
Space group
0.71073
monoclinic
P21/n (no. 14)
2.6.1. Reactions in6ol6ing [FeCl2(dppe)]
An X-ray crystallographic distinction between iron
and vanadium in a trinuclear cluster might be easier
were they bound to different ligands. This approach has
been used by others to help distinguish the different
metal sites in the crystal structure of the trinuclear
[CpMn(FeCp%)2(m-CO)2(m-NO)] (Cp=C5H5 and Cp%=
C5H4CH3) [17].
Unit cell dimensions
,
a (A)
18.0273(4)
13.8745(2)
32.5746(5)
94.277(5)
8124.9(2)
4
,
b (A)
,
c (A)
i (°)
3
,
V (A )
Z
Absorption coefficient (mm−1
)
0.61
The complex [FeCl2(dppe)] (dppe=Ph2PCH2CH2-
PPh2) [18] was described as tetrahedral [19] by analogy
with the structurally characterised [FeCl2(dippe)] [20]
and [Fe(CH2Ph)2(dippe)] [21] (dippe=Pri2PCH2CH2-
PPri2). We have now shown that [FeCl2(dppe)] can be
obtained pure and in high yield (\80%) from a fast
reaction between FeCl2 and dppe in boiling thf (see
Section 4).
q Range for data collection (°)
Reflections collected
Independent reflections
Reflections with I\2|(I)
Completeness to q=21.95° (%)
Final R indices [I\2|(I)]
3.70–21.95
39516
9837 [Rint=0.062]
8127
99.4
R1=0.068,
wR2=0.147
R1=0.084,
wR2=0.153
R indices (all data)
When white [FeCl2(dppe)] was allowed to react with
an equimolar amount of the light blue [VCl2(tmen)2] in
thf, a familiar change to turquoise was observed in the
reaction mixture, and a turquoise powder (7) was iso-
lated. This colour is characteristic of [V3Cl5(tmen)3]+
[5]. The product was insoluble in ethers and only
slightly soluble in dichloromethane, and could not be
recrystallised. All this, and the microanalysis, suggested
Table 3
,
Selected bond lengths (A) and bond angles (°) for [V2FeCl5(tmen)3]-
[Htmen···thf][BPh4]2 (6, sample 3)
Bonding and non-bonding distances about the metal centres
V(1)···V(2)
V(2)···V(3)
V(1)–N(1)
V(1)–Cl(3)
V(1)–Cl(4)
V(2)–N(3)
V(2)–Cl(1)
V(2)–Cl(4)
V(3)–N(5)
V(3)–Cl(3)
V(3)–Cl(4)
3.1771(14)
3.1586(14)
2.212(5)
2.4772(19)
2.553(3)
2.190(5)
2.4938(19)
2.536(2)
V(1)···V(3)
3.1578(13)
V(1)–N(2)
V(1)–Cl(1)
V(1)–Cl(5)
V(2)–N(4)
V(2)–Cl(2)
V(2)–Cl(5)
V(3)–N(6)
V(3)–Cl(2)
V(3)–Cl(5)
2.224(5)
2.4816(18)
2.5621(19)
2.200(5)
2.4942(19)
2.5760(19)
2.210(5)
Fig. 5. View of the trinuclear cation in [V2FeCl5(tmen)3]-
[Htmen···thf][BPh4]2 (product 6, sample 3), showing the triangulo
arrangement of the metal sites and the alternative capping groups.
Occupancies are: 0.91 Cl(4) and 0.09 OH.
2.199(5)
2.500(2)
2.525(2)
2.5016(19)
2.5459(18)
Angles about the metal centres
V(1)···V(2)···V(3) 59.79(3)
V(1)···V(3)···V(2)
Cl(3)–V(1)–Cl(1)
N(2)–V(1)–Cl(1)
N(2)–V(1)–Cl(3)
Cl(2)–V(2)–Cl(1)
N(4)–V(2)–Cl(1)
N(4)–V(2)–Cl(2)
Cl(2)–V(3)–Cl(3)
N(6)–V(3)–Cl(2)
N(6)–V(3)–Cl(3)
60.40(3)
161.20(7)
95.40(14)
98.24(14)
160.7(17)
97.82(14)
95.88(14)
161.33(7)
95.39(16)
98.16(16)
Chlorine atoms Cl(4) and Cl(5) are removed by
N(1)–V(1)–N(2)
N(1)–V(1)–Cl(1)
N(1)–V(1)–Cl(3)
N(3)–V(2)–N(4)
N(3)–V(2)–Cl(1)
83.74(19)
97.24(14)
97.06(14)
82.93(18)
94.62(15)
,
−1.762(3) and 1.795(2) A, respectively, from the least-
squares plane defined by the metal atoms. There is
some replacement (9%) of Cl(4) by an OH (Fig. 5). In
the [Htmen···thf]+ moiety, the protonated amine is
hydrogen-bonded to the oxygen atom of thf (N(7)···O
N(3)–V(2)–Cl(2) 100.50(15)
N(5)–V(3)–N(6)
N(5)–V(3)–Cl(2)
N(5)–V(3)–Cl(3)
83.23(19)
95.95(15)
98.34(15)
,
distance 2.678 A). Selected bond lengths and angles are
shown in Table 3. Table 4 compares the principal mean
dimensions with the corresponding values for
Standard deviations in parentheses.

154
D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
Table 4
Principal mean dimensions in trinuclear cations
Mean dimensions
Complex
[V3Cl5(tmen)3]+ (Ref. [5])
[Fe3Cl5(tmen)3]+ (Ref. [6b])
[V2FeCl5(tmen)3]+ (6, sample 3)
,
M–Clequatorial (A)
2.500(4)
2.519(6)
3.142(7)
2.214(2)
162.1(2)
77.8(1)
2.490(8)
2.566(12)
3.235(8)
2.194(5)
158.9(2)
81.0(4)
2.491(9)
2.550(17)
3.164(9)
2.206(11)
161.1(3)
78.8(4)
,
M–Clcapping (A)
,
M···M (A)
,
M–N (A)
Clequatorial–M–Clequatorial (°)
M–Clequatorial–M °)
(
(
)
Clcapping–M–Clcapping
°
87.9(1)
82.8(3)
86.5(3)
83.9(5)
88.4(4)
83.3(3)
(
N–M–N °)
that a cluster similar to [V3Cl5(tmen)3]+ [5] had been
formed. The FAB-MS spectrum (m/z 678) confirmed
the presence of [V3Cl5(tmen)3]+ in the solid. The Mo¨ss-
bauer spectrum (Table 1) was consistent with the pres-
ence of a high-spin iron(II) species as in [FeCl2(dppe)]
[18,19], but the larger isomer shift implies a higher
coordination number. The impure product 7 did not
appear to contain the required mixed-metal cluster.
Finally, the reaction between [V2Cl3(thf)6][AlCl2Et2]9
and [FeCl2(dppe)] (VII:FeII 1:2) in the presence of tmen
and Na[BPh4] (V/Fe metal:amine:[BPh4]− proportion
3:3:1) was expected to yield a product such as
[{Fe(dppe)}2{V(tmen)}Cl5][BPh4]. However, the amine
displaced the phosphine from the iron, and pure dppe
was isolated from the reaction mixture. The metal-con-
taining green semi-crystalline product, isolated in small
amount, produced a FAB-MS spectrum very similar to
that of 6, implying the formation of {Fe3}- and {V3}-
triangulo complexes, and similar mixed-metal cores.
even more useful because the many complexes ‘VCl2L2’
may really be trinuclear [5,6].
These complexes were investigated principally to
model the active-site cluster of vanadium nitrogenases,
though they have so far exhibited no dinitrogen chem-
istry. ESMS data support the presence of the mixed-
metal complexes in 6 and the crystallography is best
interpreted in such terms. However, these preparative
pathways cannot be used reliably to obtain pure hetero-
bimetallic products. They are clearly mixed Fe/V spe-
cies. To the best of our knowledge, the materials
described here are the only examples of clusters con-
taining both iron and vanadium in the same structural
unit, apart from some vanadium–iron–sulfur clusters.
However, vanadium–oxygen clusters have been re-
ported to show a singular and unexpected dinitrogen
chemistry [24]. The preparation of iron–vanadium clus-
ters with similar dinitrogen chemistry remains one of
our principal objectives.
3. Conclusions
4. Experimental
The difficulties of complete characterisation of the
mixed-metal complexes suggest that a more reliable
entry into the chemistry of these heterobimetallic spe-
cies needs to be found. However, our data leave little
doubt that mixed-metal trinuclear clusters triangulo-
[M3Cl5(L-k2)3]+ (M=both Fe and V) have been pre-
pared, and the material with two vanadiums and one
iron has been well characterised. Tmen seems to be a
good ligand for vanadium, although it is not electron-
rich and it frequently gives rise to crystallographic
disorder at room temperature. In spite of that, its role
as a leaving molecule during the assembly of the tris-VII
and tris-FeII cations is clearly important [5,6]. Alterna-
tive neutral ligands have not yet been identified. As for
V-starting materials other than [VCl2(tmen)2] and
[V2Cl3(thf)6]+, we have yet to explore possibilities such
as [V(MeOH)6]Cl2 [22] and other VII-complexes with
alcohols. Complexes such as ‘VCl2(EtOH)2’ [23] may be
All operations were carried out under an inert atmo-
sphere in a dinitrogen-filled drybox (Faircrest Engineer-
ing, Croydon) or with standard Schlenk techniques.
Solvents were dried by standard procedures [25] and
distilled under N2 prior to use.
The commercial products diethylaluminium ethoxide
(25% solution in toluene), 1,2-bis(diphenylphos-
phino)ethane (dppe), iron(II) chloride, iron(III) chlo-
ride, sodium tetraphenylborate, tin(II) chloride and
vanadium(III) chloride (Aldrich) were used without
further purification. N,N,N%,N%-Tetramethylethylenedi-
amine was refluxed over molten sodium for ca. 1 h and
then distilled under N2. [VCl3(thf)3] [10a], trans-
[VCl2(tmen)2] [4], [Fe2Cl3(thf)6][SnCl5(thf)] [7] and
[V2Cl3(thf)6][BPh4] [9] were prepared by published
methods. [Fe4Cl8(thf)6] [12] was obtained as a by-
product during the preparation of [Fe2Cl3(thf)6]-
[SnCl5(thf)], as described below. [{FeCl(tmen)}2((-Cl)2]

D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
155
were applied by the use of Pascal constants [27]. No
solid-state magnetochemistry measurements were made.
was prepared from an equimolar mixture of anhydrous
FeCl2 and tmen in refluxing thf [6b].
Microanalyses were performed by Mr Colin McDon-
ald (Nitrogen Fixation Laboratory) or Ms Nicola
Walker (Department of Chemistry, University of Sur-
rey), using Perkin Elmer 2400 and Leeman CE 440
CHN elemental analysers, respectively. Vanadium, iron
and tin analyses were performed by Southern Science
(Sussex Laboratory), by ICP-OES.
4.1. Preparation of [Fe2Cl3(thf)6][SnCl5(thf )] and
[Fe4Cl8(thf )6]
Anhydrous SnCl2 (4.1 g, 21.6 mmol) was added to a
yellow solution of FeCl3 (7.0 g, 43.2 mmol) in 200 cm3
of thf. The mixture was stirred for 4h, filtered through
Celite and allowed to stand at room temperature for 3
days. Colourless crystals of [Fe4Cl8(thf)6] were then
filtered off, washed with thf (60 cm3) and dried for 2 h.
Yield: 3.3 g. The pale yellow filtrate was cooled to
−20°C for 2 d to produce small light yellow needles of
[Fe2Cl3(thf)6][SnCl5(thf)]. These were washed with 50
cm3 of cold thf (−20°C) and dried for 2 h under
vacuum. Yield: 4.6 g. Total yield, based on iron con-
tent: 53%. Anal. Found for [Fe4Cl8(thf)6]: C, 30.6; H,
5.30; N, 0.0. C24H48Cl8Fe4O6 requires: C, 30.7; H, 5.16;
N, 0.0%. Anal. Found for [Fe2Cl3(thf)6][SnCl5(thf)]: C,
32.6; H, 5.65; N, 0.0. C28H56Cl8Fe2O7Sn requires: C,
33.0; H, 5.55; N, 0.0%.
NMR solvents, supplied by Goss Scientific Instru-
ments, were dried over molecular sieve before use and
kept under dinitrogen in Schlenk tubes equipped with
grease-free taps. NMR spectra were obtained in the
appropriated deuterated solvents using JEOL GSX-270
equipment. The operating frequencies for the observa-
1
tion of H and 13C were 270.2 and 67.9 MHz, respec-
tively. Tetramethylsilane was used as reference. IR data
were recorded on a Perkin Elmer 883 instrument, from
Nujol mulls prepared under dinitrogen and spread on
KBr plates.
Electrospray mass spectra were recorded on VG Plat-
form II mass spectrometer. The most suitable mobile
phase was found to be 1,2-C2H4Cl2. The lines of the
instrument were first flushed with dry ethanol, and then
overnight with carefully dried 1,2-C2H4Cl2 which was
stored under Ar. Samples, made up under Ar in the
same solvent, were injected via a Rheodyne valve with
a 10−6 dm−3 sample loop. The nebuliser tip was at
3500 V and 60°C, with dinitrogen used as a drying and
nebulising gas. The skimmer cones were maintained at
20 V to minimise fragmentation. Spectra were assigned
from m/z values and isotope distribution patterns that
were simulated using the ISOTOPE program [26].
FAB mass spectra were recorded by Dr A. Abdul
Sada (School of Chemistry, Physics and Environmental
Sciences, University of Sussex), on a VG Autospec
spectrometer (Fisons Instruments), equipped with a CsI
gun at 25 kV (SIMS technique) or on a Kratos
MS80RF machine with xenon at 8 kV (FAB tech-
nique). In both cases, 3-nitrobenzyl alcohol was used as
the matrix.
4.2. Preparation of [FeCl2(dppe)]
Anhydrous FeCl2 (1.0 g, 7.9 mmol) and dppe (3.3 g,
8.3 mmol) were suspended in thf (70 cm3) and heated
under reflux. After ca. 15 min, the light pink suspension
changed to a clear light yellow solution and after 35
min to a very heavy white suspension. The mixture was
heated a further 2 h and was filtered hot, giving a
greenish–white powder and a light yellow filtrate. The
solid was then washed with thf (40 cm3) and thf/hexane
(1:3, 45 cm3) and dried under vacuum. A second crop
of the product was recovered by filtration from the
mother liquor after 3 days at room temperature. Yield:
3.5 g (83%). Anal. Found: C, 59.3; H, 4.60; N, 0.0.
C26H24Cl2FeP2 requires: C, 59.5; H, 4.60; N, 0.0%.
Mo¨ssbauer parameters are presented in Table 1.
4.3. Reaction between [Fe2Cl3(thf)6][SnCl5(thf)] and
[V2Cl3(thf)6][BPh4]
Mo¨ssbauer data were recorded at 77 K by Mrs. J.
Elaine Barclay, using an ES-Technology MS105 spec-
trometer with a 25 mCi 57Co source in a rhodium
matrix. Spectra were referenced against iron foil at 298
K. Samples were pure solids or mixtures with boron
nitride (ca. 50% w/w), ground to fine powders and then
transferred to the aluminium sample holders in the
glove box. The program used for spectra fitting and
parameter calculation was ATMOSFIT 4, written by Dr
Ian Morrison (University of Essex).
The VII-starting material (1.1 g, 1.1 mmol) was com-
pletely dissolved in thf (20 cm3) and the bright green
solution added slowly to a light yellow suspension of
the FeII-starting material (1.1 g, 1.1 mmol) in thf (30
cm3). The colour changed to lemon yellow, light brown,
reddish brown and finally to magenta in ca. 15 min.
The mixture was stirred for 1.5 h, concentrated to ca.
30 cm3 under vacuum, filtered and stored at −20°C.
Four days later, 0.1 g of a pink crystalline material (1,
sample 1) was filtered off, washed with cold thf (3 cm3)
and dried for 30 min. The filtrate was further concen-
trated to ca. 20 cm3, left at room temperature for 4
more days and filtered to give 0.1 g of a white powder
Magnetic moment and magnetic susceptibility mea-
surements were carried out in solution by variable-tem-
perature NMR spectroscopy using the Evans method
[14]. Corrections for the diamagnetism of the ligands

156
D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
(4) which was washed with thf (5×2 cm3) and dried.
The filtrate (plus washings) was layered with hexane (35
cm3) and was allowed to stand at room temperature for
two weeks. When it was filtered, a mixture (less than
0.1 g) of bright green needles (2) and plum-coloured
short prisms (1, sample 2) was isolated. The magenta
filtrate was cooled to −20°C for 3 months; another
mixture (less than 0.1 g) of colourless (3) and deep
purple (1, sample 3) thick plates was then filtered off
and washed with cold thf (10 cm3, 0°C). The colourless
crystals turned white upon drying and the orange
filtrate was discarded. Anal. Found (%) for 1, sample 1:
C, 38.3; H, 6.70; for 1, sample 2: C, 38.4; H: 6.65; for
1, sample 3: C, 38.3; H, 6.75. [VCl3(thf)3] requires: C,
38.6; H, 6.50%. Anal. Found for 2 (%): C, 32.5; H, 5.60.
[V(II)2Cl3(thf)6][SnCl5(thf)] requires: C, 33.3; H, 5.60%.
Anal. Found for 3 (%): C, 76.9; H, 5.35. Sn(BPh4)2
requires: C, 76.1; H, 5.30%. Sn(BPh4)2×thf requires:
C, 75.2; H, 5.80%. Anal. Found for 4 (%): C, 29.1; H,
5.20; N, 0.0; Fe, 23.8; Sn, 0.0; V, 0.5. [Fe4Cl8(thf)6]
requires: C, 30.7; H, 5.15; N. 0.0; Fe, 23.7; Sn, 0.0; V,
0.0%.
0.6 g (30%). Found for 6, sample 1(%): C, 50.0; H, 6.90;
N, 8.15; Fe, 5.95; V: 9.0%; ratio V/Fe 1.5. The calcu-
lated values are presented below with the data for 6,
sample 2. IR (cm−1, Nujol mull): 1770 (w); 1824 (w);
1880 (w); 1958 (w); 1580 (m); n(B-aryl) at 1426 (s);
n(C–N) at 1002 (m) and 1018 (s); d(C–H aromatic,
out-of-plane) at 708 (s), 734 (s) and 748 (s); n(M–N) at
442 (m), 470 (m) and 498 (m).
4.6. Reaction between [{FeCl(tmen)}2(v-Cl)2],
[VCl2(tmen)2] and Na[BPh4]
Solutions of [{FeCl(tmen)}2(m-Cl)2] (1.4 g, 2.9 mmol)
and [VCl2(tmen)2] (1.0 g, 2.9 mmol) in thf (35 cm3) were
mixed at room temperature. No significant change of
colour was observed, even after stirring for ca. 20 h. A
colourless solution of Na[BPh4] (1.0 g, 2.9 mmol) in thf
(10 cm3) was then added to the reaction mixture, which
changed immediately to a dark grass-green fine suspen-
sion. The mixture was stirred for further 20 h and
filtered, giving a green filtrate and a small amount of a
grey residue (discarded). The filtrate was concentrated
under vacuum to ca. 50 cm3 and then layered with
hexane (30 cm3). Slow diffusion of the hexane layer into
the thf solution produced large deep green thick prisms
which were filtered off after 5 days, washed with thf/
hexane 1:1 (80 cm3) and dried. Yield: 2.27 g, 77%,
based on the average molecular weight of an equimolar
mixture of the {V3}, {V2Fe}, {VFe2} and {Fe3} trinu-
clear complexes (6, sample 2). The product was then
recrystallised twice from thf/hexane mixtures (2:1); the
yield of each recrystallisation step was 65–70%. Deep
green needles or long prisms of 6, sample 2, were also
obtained under slightly different crystallisation condi-
tions; no analytical technique unequivocally distin-
guished one type of crystal from the other two. The
product is very soluble in thf and CH2Cl2, and insoluble
in diethyl ether, hexane and toluene and it reacts with
acetone and methanol. Found for a representative sam-
ple of 6, sample 2: C, 49.9; H, 6.95; N, 8.25; Fe, 9.0; V,
5.7%; ratio V/Fe 0.7. The metal contents of five prepa-
rations (each analysis carried out in duplicate or tripli-
cate) were in the range 4–10% w/w (for Fe) and 5–11%
w/w (for V), the ratio V/Fe ranging from ca. 2:1 to ca.
1:2. Anal. Calc. for an equimolar mixture of the te-
traphenylborate salts of [V3Cl5(tmen)3]+, [V2FeCl5-
(tmen)3]+, [VFe2Cl5(tmen)3]+ and [Fe3Cl5(tmen)3]+: C,
50.2; H, 6.85; N, 8.35; Fe, V, 8%.
4.4. Reaction between [Fe4Cl8(thf)6] and [VCl2(tmen)2]
Tetrahydrofuran (160 cm3) was added to the mixture
of [Fe4Cl8(thf)6] (2.4 g, 2.6 mmol) and [VCl2(tmen)2]
(3.3 g, 9.2 mmol), producing a suspension which was
stirred for ca. 20 h at room temperature. A large
amount of a turquoise powder (5) was formed, which
was insoluble in the mother liquor even after heating
for 3 h under reflux. The mixture was filtered hot; the
powder was washed with thf (60 cm3) and dried for 2 h
under vacuum. Yield: 2.0 g, 46% based on the total
metal content. From the filtrate, concentrated to ca. 90
cm3 and kept at −20°C for 5 days, ca. 0.1 g of
[{FeCl(tmen)}2(m-Cl)2] was isolated. Anal. Found for 5
(%): C, 27.4; H, 6.30; N, 10.0. [V3Cl5(tmen)3]2[FeCl4]
(C36H96Cl14FeN12V6) requires: C, 27.8; H, 6.25; N,
10.8%.
4.5. Reaction between [Fe4Cl8(thf)6],
[V2Cl3(thf)6](BPh4) and tmen
A mixture of [V2Cl3(thf)6](BPh4) (1.5 g, 1.5 mmol)
and [Fe4Cl8(thf)6] (0.7 g, 0.75 mmol) in thf (120 cm3)
was stirred for 1 h at room temperature and under
reflux for 2 h when everything dissolved. Then tmen (3
cm3, 2.3 g, 19.9 mmol) was added, producing a deep
green solution which was kept under reflux for further
30 min and then filtered. A small amount of a fine
black powder was discarded. The filtrate was concen-
trated under vacuum to ca. 25 cm3, layered with hexane
(20 cm3) and left at room temperature for 3 days. Shiny
deep green needles (6, sample 1) were filtered off,
washed with thf/hexane (1:2, 100 cm3) and dried. Yield:
FAB-MS results obtained for 6, sample 2 (m/z, rela-
tive intensity): 678 [V3Cl5(tmen)3]+ (17%); 553 (residue
from [Fe3Cl5(tmen)3]+ possibly {(Htmen)2[BPh4]}+,
5%); 439 (8%); 269(7%); 117 {Htmen}+ (100%); 72
{Me2NCH2CH2}+ (93%); 58 {Me2NCH2}+ (88%). IR
(Nujol mull, cm−1): 1958 (w); 1875 (w); 1815 (w); 1765
(w); 1581 (s); n(B-aryl) at 1424 (s); n(C–N) at 1003 (m)
and 1018 (s); d(C–H aromatic, out-of-plane) at 708 (s);

D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
157
735 (s) and 749 (s); n(M–N) at 441 (m); 469 (m) and
497 (m).
Product 6 was sometimes isolated with lattice solvent
and needed to be filtered carefully to prevent the crys-
tals from becoming opaque on washing and drying.
Analytical data for such a solvate. Anal. Found: C,
51.8; H, 7.40; N. 7.75. Calc. for [M3Cl5(tmen)3](BPh4)-
(thf, as an equimolar mixture of the {V3}, {Fe3},
{V2Fe} and {VFe2} solvates: C, 51.3; H, 7.10; N,
7.80%.
Acknowledgements
We acknowledge the financial support from the
Brazilian Research Council (CNPq) and the British
Council to J.S. de S., and from the BBSRC.
References
The preparation of 6, sample 3, was carried out
essentially as described above, from 1.3 g (2.7 mmol)
[1] (a) R.L. Robson, R.R. Eady, T.H. Richardson, R.W. Miller, M.
Hawkins, J.R. Postgate, Nature (London) 322 (1986) 388;
(b) R.R. Eady, G.J. Leigh, J. Chem. Soc., Dalton Trans. (1994)
2739.
of [{FeCl(tmen)}2(m-Cl)2], 1.2
g
(3.4 mmol) of
[VCl2(tmen)2] and 0.95 g (2.8 mmol) of Na[BPh4], all
dissolved in thf (Fe/V proportion 1.6:1; metal/te-
traphenylborate 3:1). The reaction mixture was stirred
at room temperature for 20 h, filtered through Celite
and layered with hexane (thf/hexane proportion 1.5:1).
Green prisms were filtered off after 3 days at room
temperature and were analysed for C, H, N, Fe and V
contents. Anal. Found for 6, sample 3: C, 60.2; H, 7.44;
N, 7.66, total metal content: 10.3% (5.6% V; 4.7% Fe).
[2] (a) N.P. Luneva, S.A. Mironova, A.E. Shilov, M.Yu. Antipin,
Y.T. Struchkov, Angew. Chem., Int. Ed. Engl. 32 (1993) 1178;
(b) T.A. Bazhenova, A.E. Shilov, Coord. Chem. Rev. 144 (1995)
69;
(c) A.E. Shilov, Cattech. 3 (1999) 72.
[3] D.L. Hughes, G.J. Leigh, J.R. Sanders, J.S. de Souza, J. Chem.
Soc., Chem. Commun. (1996) 1985.
[4] J.J.H. Edema, W. Stauthamer, F. van Bolhuis, S. Gambarotta,
W.J.J. Smeets, A.L. Spek, Inorg. Chem. 29 (1990) 1302.
[5] (a) D.L. Hughes, L.F. Larkworthy, G.J. Leigh, C.J. McGarry,
J.R. Sanders, G.W. Smith, J.S. de Souza, J. Chem. Soc., Chem.
Commun. (1994) 2137;
Calc.
for
[V2FeCl5(tmeda)3][tmedaH···thf][BPh4]2,
C76H113B2Cl5FeN8OV2: C, 60.4; H, 7.54; N, 7.41%,
total metal content: 10.4% (6.7% V; 3.7% Fe).
(b) P.B. Hitchcock, D.L. Hughes, L.F. Larkworthy, G.J. Leigh,
C.J. Marmion, J.R. Sanders, G.W. Smith, J.S. de Souza, J.
Chem. Soc., Dalton Trans. (1997) 1127.
[6] (a) A.C. Niedwieski, A.R. Raimondi, D.L. Hughes, G.J. Leigh,
F.S. Nunes, J.F. Soares, in: F.O. Pedrosa, M. Hungria, M.G.
Yates, W.E. Newton (Eds.), Nitrogen Fixation: from Molecules
to Crop Productivity, Kluwer Academic, New York, 2000, p. 62
and unpublished data;
(b) S.C. Davies, D.L. Hughes, G.J. Leigh, J.R. Sanders, J.S. de
Souza, J. Chem. Soc., Dalton Trans. (1997) 1981.
[7] Z. Janas, P. Sobota, T. Lis, J. Chem. Soc., Dalton Trans. (1991)
2429.
[8] P. Sobota, Polyhedron 11 (1992) 715.
[9] J.A.M. Canich, F.A. Cotton, S.A. Duraj, W.J. Roth, Polyhedron
6 (1987) 1433.
[10] (a) L.E. Manzer, Inorg. Synth. 21 (1982) 138;
(b) P. Sobota, J. Ejfler, S. Szafert, K. Szczegot, W. Sawka-Do-
browolska, J. Chem. Soc., Dalton Trans. (1993) 2353.
[11] (a) R.J. Haines, A.L. du Preez, J. Am. Chem. Soc. 93 (1971)
2820;
4.7. Reaction between [FeCl2(dppe)] and [VCl2(tmen)2]
A light blue solution of the V(II)-starting material
(0.5 g, 1.4 mmol) in thf (20 cm3) was mixed with a
white suspension of [FeCl2(dppe)] (0.7 g, 1.4 mmol) in
thf (30 cm3). The mixture changed immediately to a
turquoise suspension of 7, which was stirred for ca. 20
h and then filtered. The solid (0.4 g) was washed with
thf (2 cm3) and hexane (20 cm3) and dried under
vacuum. The filtrate (light purple) was discarded. The
product was insoluble in thf, diethyl ether and hexane,
and only slightly soluble in dichloromethane. Anal.
Found for 7: C, 36.5; H, 6.05; N, 7.95. [V3Cl5(tmen)3]2-
[FeCl4(dppe)] requires: C, 38.1; H, 6.20; N, 8.60%.
FAB-MS (m/z, relative intensity): 882 (9%); 852 (0.5%);
678 [V3Cl5(tmen)3]+ (9%); 479 (7%); 147 (39%); 110
(100%); 72 (26%); 67 (40%).
(b) G.J. Kruger, A.L. du Preez, R.J. Haines, J. Chem. Soc.,
Dalton Trans. (1974) 1302;
(c) R.J. Haines, A.L. du Preez, J. Organomet. Chem. 84 (1975)
357;
(d) J.R. Sanders, J. Chem. Soc., Dalton Trans. (1973) 743.
[12] (a) V.K. Bels’kii, V.M. Ishchenko, G.M. Bulychev, A.N. Prot-
skii, G.L. Soloveichik, O.G. Ellert, Z.M. Seifulina, Yu.V. Rak-
itin, V.M. Novotortsev, Inorg. Chim. Acta 96 (1985) 123;
(b) F.A. Cotton, R.L. Luck, K.-A. Son, Inorg. Chim. Acta 179
(1991) 11.
[13] E. Slade, A. Houlton, J. Silver, D.J. Evans, G.J. Leigh, J. Chem.
Soc., Dalton Trans. (1993) 1217.
[14] (a) D.F. Evans, J. Chem. Soc. (1959) 2003;
(b) D.F. Evans, G.V. Fazakerley, R.F. Phillips, J. Chem. Soc. A
(1971) 1931;
4.8. X-ray diffraction analysis of
[V2FeCl5(tmen)3][Htmen···thf ][BPh4]2 (6, sample 3)
A long, green, air-sensitive prism, ca. 0.25×0.15×
0.10 mm3 was mounted on a glass fibre and cooled on
a Nonius Kappa CCD area detector diffractometer. No
deterioration correction was necessary and no absorp-
tion correction was applied. All the non-hydrogen
atoms were allowed anisotropic thermal parameters.
Refinement was by full-matrix least-square methods on
F2. Programs used: WINGX and SHELXL-97 [28]. For
details, see Table 2.
(c) D.F. Evans, T.A. James, J. Chem. Soc., Dalton Trans. (1979)
723.
[15] (a) J.J.H. Edema, A. Meetsma, S. Gambarotta, J. Chem. Soc.,
Chem. Commun. (1990) 951;

158
D.J. E6ans et al. / Inorganica Chimica Acta 319 (2001) 147–158
(b) J.J.H. Edema, R. Duchateau, S. Gambarotta, C. Bensimon,
Inorg. Chem. 30 (1991) 3585.
[22] F.A. Cotton, S.A. Duraj, L.E. Manzer, W.J. Roth, J. Am.
Chem. Soc. 107 (1985) 3850.
[16] (a) P. Sobota, J. Utko, S. Szafert, Z. Janas and T. Glowiak, J.
Chem. Soc., Dalton Trans. (1996) 3469;
[23] L.F. Larkworthy, M.W. O’Donoghue, Inorg. Chim. Acta 71
(1983) 81.
(b) D.A. Handley, P.B. Hitchcock, G.J. Leigh, Inorg. Chim.
Acta (in press).
[24] N.P. Luneva, L.A. Nikonova, A.E. Shilov, Kinet. Catal. 21
(1980) 1041.
[17] M.E. Barr, A. Bjarnason, L.F. Dahl, Organometallics 13 (1994)
1981.
[18] R.A. Henderson, J. Chem. Soc., Dalton Trans. (1988) 509.
[19] J.E. Barclay, G.J. Leigh, A. Houlton, J. Silver, J. Chem. Soc.,
Dalton Trans. (1988) 2865.
[25] P.D. Perrin, W.L.F. Armarego, Purification of Laboratory
Chemicals, 3rd edn., Pergamon, New York, 1988.
[26] L.J. Arnold, J. Chem. Educ. 69 (1992) 811.
[27] A. Earnshaw, Introduction to Magnetochemistry, Academic
Press, London, 1968, p. 4.
[20] A.R. Hermes, G.S. Girolami, Inorg. Chem. 7 (1988) 1775.
[21] A.R. Hermes, G.S. Girolami, Organometallics 6 (1987) 763.
[28] G.M. Sheldrick, SHELX-97, Program for the refinement of crystal
structures, University of Go¨ttingen, 1997.
.