Redox-niediation of electron-electron spin-spin exchange interactions, |J|, in paramagnetic trinuclear molybdenum complexes: An example of a 'J switch'

Article abstract of DOI:10.1039/a908323e

A series of trinuclear molybdenum nitrosyl complexes has been prepared using the dinucleating ligands 4-(4-hydroxyphenyl)pyridine (HL¹), 1-(4-pyridyl)-2-(4-hydroxyphenyl)-ethene (HL²) and -ethane (HL³), and the trinucleating ligands 3,5-bis(4-hydroxyphenyl)pyridine (H₂L⁴) and 2,6-bis(4-ethenylpyridyl)-4-hydroxytoluene (HL⁵). These complexes are of the type [Mo{OC₆H₄EpyMoCl}₂] (Mo = Mo(NO)Tp^Me,Me, Tp^Me,Me = tris(3,5-dimethylpyrazolyl)borate; E = nothing, CH=CH and CH₂CH₂; py = C₅H₄N or C₅H₃N; from HL¹, HL² and HL³), [{ClMo(OC₆H₄)}₂pyMoCl] (from H₂L⁴), and [ClMo{OC₆H₃Me[CH=CHpyMoCl]₂}] (from HL⁵). The species [Mo{OC₆H₄EpyMoCl}₂] contains one 16 valence electron (ve) ([Mo{OC₆H₄-}₂]) and two 17 ve centres ([(-py)MoCl]), [{ClMo(OC₆H₄)}₂pyMoCl] has two 16 ([ClMo{OC₆H₄-}]) and one 17 ([(-py)MoCl]) ve centres and [ClMo{OC₆H₃Me[CH=CHpyMoCl]₂}] one 16 and two 17 ve centres. Reduction of these species by cobaltocene in tetrahydrofuran/dichloromethane mixtures affords complexes having three 17 ve centres with one unpaired electron per metal centre. The interaction between these impaired electrons in solution is determined by the relationship between |J|, the electron spin-spin exchange interaction, and A_Mo, the molybdenum hyperfine coupling constant, which was detected by EPR spectroscopy. In [Mo{OC₆H₄EpyMoCl}₂], the interaction was dependent on ligand conformation, |J| ≈ A_Mo when E = nothing, |J| ? A_Mo when E = CH=CH and |J| ? A_Mo when E = CH₂CH₂. Reduction of [Mo{OC₆H₄EpyMoCl}₂] to [Mo{OC₆H₄EpyMoCl}₂]^- resulted in exchange between all three spins irrespective of ligand conformation, and the EPR spectra of [{ClMo(OC₆H₄)}₂pyMoCl]^2- and [ClMo{OC₆H₃Me[CH=CHpyMoCl]₂}]^- were similar to that of [Mo{OC₆H₄EpyMoCl}₂]^-. Oxidation reconstitutes the original EPR spectra of [Mo{OC₆H₄EpyMoCl}₂], [{ClMo(OC₆H₄)}₂pyMoCl] and [ClMo{OC₆H₃Me-[CH=CHpyMoCl]₂}]. This behaviour is consistent with full three centre interaction being 'switched on' when the 17:16:17 or 16:17:16 ve systems are reduced to a 17:17:17 ve system, and 'switched off' on reoxidation. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a908323e

View Article Online / Journal Homepage / Table of Contents for this issue
Redox-mediation of electron–electron spin–spin exchange
interactions, |J|, in paramagnetic trinuclear molybdenum
complexes: an example of a ‘J switch’
Elefteria Psillakis, Peter K. A. Shonfield, Abdel-Aziz Jouaiti, John P. Maher,
Jon A. McCleverty* and Michael D. Ward*
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK BS8 1TS.
E-mail: jon.mccleverty@bristol.ac.uk; mike.ward@bristol.ac.uk
Received 18th October 1999, Accepted 7th December 1999
A series of trinuclear molybdenum nitrosyl complexes has been prepared using the dinucleating ligands 4-(4-
hydroxyphenyl)pyridine (HL1), 1-(4-pyridyl)-2-(4-hydroxyphenyl)-ethene (HL2) and -ethane (HL3), and the
trinucleating ligands 3,5-bis(4-hydroxyphenyl)pyridine (H2L4) and 2,6-bis(4-ethenylpyridyl)-4-hydroxytoluene
(HL5). These complexes are of the type [Mo{OC6H4EpyMoCl}2] (Mo = Mo(NO)TpMe,Me, TpMe,Me = tris(3,5-
dimethylpyrazolyl)borate; E = nothing, CH᎐CH and CH CH ; py = C H N or C H N; from HL1, HL2 and
᎐
2
2
5
4
5
3
HL3), [{ClMo(OC H )} pyMoCl] (from H L4), and [ClMo{OC H Me[CH᎐CHpyMoCl] }] (from HL5). The
᎐
6
4
2
2
6
3
2
species [Mo{OC6H4EpyMoCl}2] contains one 16 valence electron (ve) ([Mo{OC6H4–}2]) and two 17 ve centres
([(–py)MoCl]), [{ClMo(OC6H4)}2pyMoCl] has two 16 ([ClMo{OC6H4–}]) and one 17 ([(–py)MoCl]) ve centres
and [ClMo{OC H Me[CH᎐CHpyMoCl] }] one 16 and two 17 ve centres. Reduction of these species by cobaltocene
᎐
6
3
2
in tetrahydrofuran/dichloromethane mixtures affords complexes having three 17 ve centres with one unpaired
electron per metal centre. The interaction between these unpaired electrons in solution is determined by the
relationship between |J|, the electron spin–spin exchange interaction, and AMo, the molybdenum hyperfine
coupling constant, which was detected by EPR spectroscopy. In [Mo{OC6H4EpyMoCl}2], the interaction was
dependent on ligand conformation, |J| ≈ AMo when E = nothing, |J| ӷ AMo when E = CH᎐CH and |J| Ӷ AMo when
᎐
E = CH2CH2. Reduction of [Mo{OC6H4EpyMoCl}2] to [Mo{OC6H4EpyMoCl}2]Ϫ resulted in exchange between all
three spins irrespective of ligand conformation, and the EPR spectra of [{ClMo(OC6H4)}2pyMoCl]2Ϫ and
[ClMo{OC H Me[CH᎐CHpyMoCl] }]Ϫ were similar to that of [Mo{OC H EpyMoCl} ]Ϫ. Oxidation reconstitutes
᎐
6
3
2
6
4
2
the original EPR spectra of [Mo{OC6H4EpyMoCl}2], [{ClMo(OC6H4)}2pyMoCl] and [ClMo{OC6H3Me-
[CH᎐CHpyMoCl] }]. This behaviour is consistent with full three centre interaction being ‘switched on’ when the
᎐
2
17:16:17 or 16:17:16 ve systems are reduced to a 17:17:17 ve system, and ‘switched off ’ on reoxidation.
The construction of stable multi-centre high-spin organic mole-
cules which exhibit ferromagnetic behaviour is a major object-
ive of the molecular magnetic community.1 The synthesis and
stabilisation of such polyradicals remains, however, a difficult
synthetic challenge and while interesting magnetic behaviour
has been detected,2 it is as yet unclear how such species can
be incorporated in magnetic materials. However, molecular
magnets may also be derived from oligonuclear transition
metal complexes containing at least two, and preferably more,
paramagnetic transition metal centres.3 Such oligomeric
coordination and/or organometallic compounds are likely to
have several advantages over organic polyradicals, such as the
possibility of higher spin density, chemical stability, and redox
activity for switching purposes. It is this last possibility which
has attracted our attention.
Of course, the electronic interactions between metals in such
oligonuclear species are critically dependent on the type of
connection between the constituent metal ions,3 and the control
and manipulation of these interactions is a major objective of
contemporary inorganic and related materials chemistry. If the
metal orbitals are close enough to overlap directly, the nature of
the magnetic interaction (ferro- or antiferro-magnetic) depends
on their relative symmetry, as formalised by the Goodenough–
Kanamori rules,4 and this has been exploited in the preparation
of complexes with predictable magnetic properties.4,5 However
if the relevant orbitals are too far apart to overlap directly then
there may be no electron–electron spin–spin exchange (hence-
forth referred to as exchange (J) interactions), but if such
exchange does occur, then coupling may only occur via the
participation of the bridging ligand orbitals.
The mediating effect of bridging ligands on the exchange
coupling between metals in oligonuclear complexes has received
relatively little systematic attention, in contrast to the extensive
investigations of the effects of structure and topology on the
properties of organic polyradicals.1,6 The identification of suit-
able paramagnetic transition metal components which can be
easily linked within a rigid predominantly carbon-based archi-
tectural framework and which may couple magnetically is now
an important goal in the design and assembly of ‘molecular
magnets’. At the heart of this type of work is the control of
the sign and magnitude of the exchange interaction, J, which
depends substantially on the nature of the pathway linking the
interacting spins.
The magnetic behaviour of organic polyradicals and some
oligonuclear transition metal complexes is also amenable to
study by EPR spectroscopy,6–8 and it is now well-established
that an EPR spectrum for any compound depends on the mag-
nitude of J compared to A, the hyperfine coupling constant.9
When |J| Ӷ A, exchange is negligibly small but when |J| ӷ A it
is detectable by EPR spectroscopy even though it may be too
small to detect by conventional susceptibility measurements. In
neither condition is it possible to easily extract the sign or
magnitude of J. However, when |J| ≈ A more complex second-
order EPR spectra result, and a value for J can be computed by
spectral simulation.10
The EPR spectra of Mo() nitrosyl complexes of the
type described in this paper are characteristic, as we have
described elsewhere,8,11 the electron–nuclei coupling constant
AMo being ca. 5.0 mT for single or isolated paramagnetic mono-
molybdenum nitrosyl centres.12–14 However, in a dinuclear
J. Chem. Soc., Dalton Trans., 2000, 241–249
This journal is © The Royal Society of Chemistry 2000
241

View Article Online
species with one unpaired spin on each Mo centre, the observed
hyperfine splitting is ca. 2.5 mT, i.e. AMo/2, provided |J| ӷ
A.12–14 Such a complex set of multiplets is observed in the
(b) constructing a bridge incorporating three binding sites cap-
able of binding one 17 ve and two reducible 16 ve metal centres,
EPR spectrum of [ClMo{py(CH᎐CH) py}MoCl] (Mo =
᎐
4
Mo(NO)TpMe,Me
,
TpMe,Me = tris(3,5-dimethylpyrazolyl)borate,
py = C5H4N or C5H3N)13 and is characteristic of the situation
9,15
where J ӷ AMo
.
It has long been known that such exchange
coupled spectra are shown by nitroxide diradicals.9 In tri-
nuclear species such as the centrosymmetric [1,3,5-{MoCl-
᎐
(pyC᎐C)} C H ] the observed hyperfine coupling is 1.6 mT
᎐
3
6
3
(AMo/3),11 and similar behaviour is observed in nitroxide
or
triradicals.16
(c) making a similar type of ligand binding two 17 ve and one
reducible 16 ve metal groups.
The asymmetric dinuclear complex [ClMo(pyCH᎐CHC H -
᎐
6
4
OMoCl)], which contains a diamagnetic 16- (phenoxide) and
a paramagnetic 17-valence electron (ve) (pyridine) metal
centre,14 has a ‘singletϩsextet’ multiplet EPR spectrum,† with
AMo = 5.0 mT, consistent with the expected valence-trapped
behaviour, no hyperfine coupling being observed between the
unpaired electron and the adjacent diamagnetic Mo centre.
However, reduction to the isovalent monoanion [ClMo-
(pyCH᎐CHC H OMoCl)]Ϫ (two 17 ve centres) caused the
᎐
6
4
EPR spectrum to change to a ‘singlet ϩ sextet ϩ undecet’
multiplet,† with an observed hyperfine splitting of 2.5 mT
(AMo/2): a situation consistent with |J| ӷ AMo and both elec-
trons coupling equally to both nuclei despite the asymmetry of
the complex. This monoanion is isoelectronic with [ClMo(4,4Ј-
Route (a) could be realised using ligands HL1–HL3, Fig. 1, by
preparing a 16 ve bis(phenolato) complex, e.g. [Mo(OC6H4py)2],
and then attaching 17 ve molybdenum nitrosyl groups to the
pyridyl residues giving [Mo{OC6H4pyMoCl}2]. Ligands H2L4
could afford entry to route (b) and HL5 to route (c). Some
preliminary results have been reported,18 and this paper
describes in full the outcome of our search for switchable three-
centre paramagnets.
pyCH᎐CHpy)MoCl] which has an observed hyperfine splitting
᎐
of 2.5 mT, with J = Ϫ18 cmϪ1, determined from susceptibility
measurements.17 The exchange interaction between the two
molybdenum centres in [ClMo(pyCH᎐CHC H OMoCl)] is
᎐
6
4
‘switched on’ by reversible reduction of the molybdenum
phenolate terminus from a 16 ve to a 17 ve configuration.
Re-oxidation caused the EPR spectrum to revert to its original
17/16 ve state prior to reduction (AMo = 5.0 mT).
It was this ‘switching’ effect which has prompted our search
for similar behaviour in more complex oligonuclear species. For
example, could we make a trinuclear complex containing one
diamagnetic and two paramagnetic centres which, on reduc-
tion, would be converted to a system where full exchange
occurred between all three paramagnetic centres? If this
behaviour was electrochemically reversible, the system would
constitute a ‘J switch’.
Synthetically, this objective could be achieved by three routes,
Fig. 1 Ligands.
all of which exploit the reversible redox couple {Mo(NO)}3ϩ
/
{Mo(NO)}2ϩ
:
Results and discussion
(a) by attaching two 17 ve metal centres via a bridging ligand to
a central reducible 16 ve metal centre, forming an acyclic or
chain-like molecule;
The ligands we have used in developing our synthetic strategies
were prepared by relatively simple routes. We have described
RL1 and RL2 (R = H or Me) before14 and hydrogenation of
MeL2 gave MeL3. Trifunctional Me2L4 was obtained via a
coupling reaction involving 3,5-dichloropyridine and two
equivalents of the Grignard reagent of 4-bromoanisole. The
dipyridine MeL5 was prepared using the Heck reaction, coup-
ling 2,6-dibromo-4-methoxytoluene to two equivalents of
4-vinylpyridine. Demethylation of MeL3, MeL4 and MeL5
using molten pyridinium chloride gave good yields of HL3,
H2L4 and HL5. The ligands all contain two different types of
reaction centre: at least one phenolic OH group and at least one
pyridine N atom (Fig. 1).
The desired complexes were prepared by standard procedures
which we have described in detail before,11,19 and were charac-
terised initially by their elemental analyses and FAB mass
spectra. We used three 4-pyridylphenols HL1, HL2 and HL3 to
assemble trinuclear species via route (a), as shown in Scheme 1.
The NO stretching frequencies of the monometallic 1–3 (Fig. 2)
and trimetallic 4–6 were typical and consistent with our formu-
lations.11,14,20 Specifically, νNO in the mononuclear species
occurred at ca. 1654 cmϪ1, somewhat lower than the same
centre in the trinuclear 4–6 because of the electronegativity of
the {(–py)MoCl} group, and much lower than that in mono-
† The isotopes 92Mo, 94Mo, 96Mo, 98Mo and 100Mo have I = 0, with
relative abundance ca. 75%; 95Mo, relative abundance 15.9%, nuclear
magnetic moment Ϫ0.914 µN and 97Mo, relative abundance 9.6%,
nuclear magnetic moment Ϫ0.934 µN have I = 5/2. The EPR spectra of
isolated paramagnetic monomolybdenum nitrosyl centres appear as a
singlet (75% of total signal intensity) overlapped by a sextet (25% of
total intensity). In dinuclear species with one unpaired spin on each Mo
centre, the spectrum appears as a superimposed singlet (I = 0, I = 0; 56%
probability), sextet (I = 0, I = 5/2; 38%) and 1:2:3:4:5:6:5:4:3:2:1
undecet (I = 5/2, I = 5/2; 6%). In trinuclear species with one unpaired
electron per Mo, the EPR spectra appear as a superimposition of a
singlet (I = 0, I = 0, I = 0; 42% probability), sextet (I = 0, I = 0, I = 5/2;
42%), undecet (I = 0, I = 5/2, I = 5/2; 14%) and 16-fold multiplet,
(I = 5/2, I = 5/2, I = 5/2, 2%) with relative intensities. 1:3:6:10:15:21:
25:27:27:25:21:15:10:6:3:1.
242
J. Chem. Soc., Dalton Trans., 2000, 241–249

View Article Online
Fig. 2 Mono-, di- and tri-nuclear molybdenum complexes: Mo = Mo(NO)TpMe,Me
.
phenoxide species such as [ClMo(pyCH᎐CHC H OMoCl)]
᎐
6
4
(ca. 1685 cmϪ1) because two phenoxide groups are more
effective electron donors than one. Reduction of 5 to [5]Ϫ in
tetrahydrofuran/dichloromethane by addition of one mole
equivalent of cobaltocene caused the colour of the solution
to change from red to pale brown and the NO stretching
frequency of the {Mo(OAr)2} group to decrease by 48 cmϪ1
,
bringing νNO into the range characteristic of a 17 ve molyb-
denum() nitrosyl centre.8
Reaction of [MoCl2] with H2L4 using a 2:1 mole ratio gave
dinuclear 7 and trinuclear 8 via route (b). We were also able to
quaternise the uncoordinated pyridine group in 7 using MeI,
giving cationic [9]ϩ, isolated as its PF6Ϫ salt. The NO stretching
frequencies are again consistent with our formulations, but we
did not detect a significant shift to higher frequencies in the
conversion of 7 to [9]ϩ, perhaps because the molybdenum
nitrosyl fragments are too far away from the site of the positive
charge. We were also able to make trimolybdenum complexes
directly using HL5 (route (c)). A mixture of dinuclear 10 and
trinuclear 11 was prepared by the reaction of the ligand with a
fourfold excess of [MoCl2] in the presence of NEt3. The FAB
MS spectra and NO stretching frequencies of these compounds
were as expected and selected data are listed in the Experi-
mental section.
Electrochemical behaviour
The cyclic voltammograms (CVs) in dichloromethane obtained
from most of the complexes described in this paper exhibited
quasi-reversible behaviour at best (potential data are given in
the Experimental section). While the CV curves were sym-
metric, with equal cathodic and anodic peak currents, the peak-
to-peak separations in the CVs were larger than the theoretical
ideal, in some cases being up to 230 mV. However, from several
spectroelectrochemical studies on complexes of this type we
Scheme 1
J. Chem. Soc., Dalton Trans., 2000, 241–249
243

View Article Online
Table 1 EPR spectral data obtained from paramagnetic molybdenum and tungsten complexes, subdivided by ligand type according to bridging
ligands
b
c
b
c
b
c
Complexa
giso
AMo
Complexa
giso
AMo
Complexa
giso
AMo
17 ve phenolato
centres
17 ve pyridyl
centres
17:17 ve mixed
centres
Ligands L1–L3
[1]Ϫ
[2]Ϫ
[3]Ϫ
d
1.963
1.964
1.963
5.1
5.1
5.1
4
5
6
1.978
1.977
1.977
[4]Ϫ
[5]Ϫ
[6]Ϫ
1.974
1.976
1.973
1.6
1.7
1.7
2.4
4.9
Ligand L4
[7]2Ϫ
1.969
1.969
2.6
2.6
8
1.979
4.9
[8]2Ϫ
1.971
1.7
[9]Ϫ
Ligand L5
10
11
1.978
1.978
4.8
2.4
[10]Ϫ
[11]Ϫ
1.973
1.975
2.4
1.8
a Anionic complexes generated by addition of cobaltocene to the precursor complex in CH2Cl2/THF (1:1 v/v). b In CH2Cl2 except for anionic species,
at 293 K. c Metal-hyperfine coupling (95Mo, 97Mo; I = 5/2), in mT. d Second order spectrum.
Because the potential for the [Co(C5H5)2]ϩ/[Co(C5H5)2]
couple is Ϫ1.34 V with respect to the ferrocenium/ferrocene
couple in dichloromethane, we were able to effect an electron
transfer to all those complexes having a reduction potential
equal to or less negative than Ϫ1.30 V, this process being
associated exclusively with a molybdenum mono- or bis-
phenolato fragment. Use of cobaltocene as a reducing agent in
this system is particularly convenient because of its solubility in
solvents compatible with the di- and tri-nuclear species being
investigated, and because its oxidation product, [Co(C5H5)2]ϩ,
is relatively kinetically stable.‡
Fig. 3 Cyclic voltammogram of [Mo{OC6H4pyMoCl}2], 4, showing,
from left to right, the 17/18 ve reductions of the two pyridinyl-
EPR spectral behaviour
molybdenum centres, the 16/17 ve reduction of the single phenolato-
molybdenum fragment, and the 17/16 ve oxidation of the two
pyridinyl-molybdenum centres.
The EPR spectrum of the reduced mononuclear bis(phenolato)
complexes [1]Ϫ (Fig. 4a), [2]Ϫ and [3]Ϫ in THF are typical of
their class, giso = 1.963 and AMo = 5.1 mT (Table 1).8 The tri-
metallic complex 4 is unusual, however, in that it exhibits a
“second-order” spectrum (Fig. 4b; giso = 1.978)§ indicating that
|J| ≈ AMo (see later). In contrast, 5 (Fig. 4c) has a normal
‘singlet ϩ sextet ϩ undecet’ spectrum, the hyperfine splitting
being 2.4 mT, indicating that |J| ӷ AMo, while 6 (Fig. 4d) has a
‘singlet ϩ sextet’ spectrum, AMo = 4.9 mT, consistent with two
non-interacting spins, i.e. |J| Ӷ AMo. These variations in the
relationship between J and AMo are presumably a function of
the bridging ligand where, in 4, the two rings are relatively free
to rotate in solution leading to some loss of delocalisation and
therefore also of electron–electron spin–spin exchange across
the bridge. The inter-ring torsion angle in such ligands in the
solid state would be expected to fall in the range 30–40Њ. In 5 the
bridging ligand is constrained to be essentially planar because
of the ethene link which facilitates a high degree of delocalis-
ation, but in 6 both the insulating effect and the rotational
flexibility of the CH2CH2 group will substantially reduce
delocalisation and hence the effectiveness of exchange. So in 6
the electron spins on each metal centre are effectively isolated.
Exchange coupling between pairs of 17 ve molybdenum
nitrosyl fragments via another diamagnetic metal atom centre
is not unprecedented. We have obtained second order EPR
spectra from ruthenium() porphyrin adducts, [trans-{ClMo}-
have confirmed that the majority of the couples {Mo(NO)}3ϩ
/
{Mo(NO)}2ϩ and {Mo(NO)}2ϩ/{Mo(NO)}ϩ were chemically
reversible.13 Our electrochemical investigations were sup-
plemented by square-wave voltammetry (SWV). We estimated
whether each electrode process was a one- or two-electron
transfer either from the relative intensities of the CV responses
or from the relative areas under each peak in the appropriate
SWVs. We know from previous studies of dinuclear and
trinuclear species that, when there is no significant Coulombic
interaction between the redox centres, the reductions or oxid-
ations will appear as a synchronous two- or three-electron
process.11,14,21 Theoretically, the electron transfer must occur
statistically in separate steps but the effect of this is not
normally resolved by conventional CV or SWV techniques,
and we did not attempt to analyse our electrochemical data
with this in mind. A typical CV of a trinuclear species, 4, in
dichloromethane is shown in Fig. 3.
Generally, complexes containing the [MoCl(OC6H4–)] group
reduce in a one-electron step with formation potentials in the
range Ϫ0.74 to Ϫ0.91 V vs. the Fcϩ/Fc couple, and the corre-
sponding bis(phenoxides) reduce more cathodically. This is
expected because phenoxides are stronger π-donors than
chloride, as we have observed previously.22 The [MoCl(OC6H4–)]
group does not exhibit oxidation behaviour between 0 and
ϩ1.00 V. The complexes containing the [MoCl(py–)] fragment
undergo a one-electron reduction between Ϫ1.70 and Ϫ2.50
V, a process which is largely determined by the nature of the
pyridyl ligand but is irreversible when close to the medium
decomposition at very negative potentials. This group also
undergoes a one-electron oxidation at ca. ϩ0.06 V which is
largely metal-based and independent of the bridging ligand.11–14
(4,4Ј-bipy)[Ru(TPP)](4,4Ј-bipy){MoCl}]
{TPP = meso-5,10,
‡ Neither [Co(C5H5)2] nor [Co(C5H5)2]ϩ exhibit EPR spectra under the
conditions described here.
§ The “second order” effects described here arise from the fact that
J and AMo are of comparable magnitude and should be independent of
the spectrometer frequency; here X-band. However, there will be an
additional second order effect from the spin nuclei and adjacent separ-
ations are unlikely to be identical at X- and Q-band frequencies.
244
J. Chem. Soc., Dalton Trans., 2000, 241–249

View Article Online
Fig. 4 EPR spectrum of (a) a reduced 17 ve bis(phenolato) complex, [1]Ϫ; (b) 4 (17:16:17 ve), J ≈ AMo (second order spectrum); (c) 5, J ӷ AMo
;
(d) 6, J Ӷ AMo; and (e) [6]Ϫ (17:17:17 ve).
15,20-tetraphenylporphyrinate(2Ϫ)},23 where the two spins are
interacting via a 17 atom pathway which incorporates the Ru
atom (18 ve). In 4 and 5, the central diamagnetic 16 ve Mo
centre does not appear to inhibit in any obvious way the
exchange coupling process. However, when the connecting
metal group is paramagnetic, as in trinuclear oxomolyb-
denum() complexes which contain diphenolato bridges
between the MoO3ϩ centres, the exchange coupling is signifi-
cantly reduced.24
and [9]Ϫ, which contain two {MoCl(OC6H4–)} groups, have
hyperfine splittings of 2.6 mT, while that of trinuclear 8 is
typical of a single 17 ve {MoCl(py–)} fragment (AMo = 4.9 mT).
The dianion [8]2Ϫ contains three 17 ve centres and reveals full
three-centre exchange, with an observed line splitting of 1.7 mT,
just like that in the trinuclear [4]Ϫ, [5]Ϫ and [6]Ϫ.
The spectrum of 10, like that of 8, is typical of an isolated
17 ve mono-molybdenum centre, while those of 11 and [10]Ϫ
are similar to those of two interacting 17 ve molybdenum
paramagnets.
Our earlier descriptions of how the sign and magnitude of
the exchange interaction between two {Mo(NO)}2ϩ (17 ve)
fragments may be controlled by bridging ligand topology and
conformation are based qualitatively on the spin-polarisation
principle as we have described elsewhere.8,17,25 Propagation of
the polarisation across the bridging ligand is obviously greatly
facilitated by π-orbitals, but when these are not available, spin
polarisation must occur through the σ-bonding system and the
effect attenuates very rapidly. Provided that the magnitude of
J is relatively large, this simple approach allows us to predict the
sign of J, and our predictions were confirmed by susceptibility
studies in the solid state.8,17,24,25 The effect of this spin polaris-
ation is implicit in contact-shifted NMR spectra of Ni()
aminotroponiminato complexes, where spin density alternates
along conjugated groups attached to the ligand framework.26
Dealing first with the complexes derived from [L4]Ϫ and [L5]Ϫ,
the exchange interaction behaviour of [7]2Ϫ and [9]Ϫ is normal
On reduction of 4, 5 and 6, the spectra appear as
‘singlet ϩ sextet ϩ undecet ϩ 16-fold’ multiplets, although only
the outer components are visible, on the low-field side of the
spectra. These components are very weak, but their relative
intensities are as expected, with giso ≈ 1.974 and a hyperfine
splitting reduced to 1.7 mT (AMo/3). This means that there is
exchange coupling between the three spins. What is remarkable
is the relationship between the spectra of 6 and [6]Ϫ. In the
former (17:16:17 ve) we could not detect any interaction
between the two spins at the complex termini, but in the latter
(17:17:17 ve) apparent full three-centre exchange occurs (Fig.
4e) with all three electrons coupled to all three Mo nuclei accord-
ing to the EPR spectrum. This behaviour is reversible: reoxid-
ation of [4]Ϫ, [5]Ϫ and [6]Ϫ by iodine causes the EPR spectra to
revert to those observed for unreduced 4, 5 and 6. Thus the
terminal unpaired spins in 6 are electronically isolated but on
reduction of the metal ion between them, they undergo
exchange with each other as well as with the adjacent (central)
spin.
and consistent with that we have observed in most other
The EPR spectra of the fully reduced dinuclear species [7]2Ϫ
reduced dinuclear phenolato complexes where |J| ӷ AMo
.
27
J. Chem. Soc., Dalton Trans., 2000, 241–249
245

View Article Online
cule are clearly sufficient to effectively isolate the two spins, so
|J| Ӷ AMo. Reduction of 6 to [6]Ϫ reduces the pathway between
the individual spins from 23 atoms (end-to-end) to 11 atoms
(end-to-centre), and pairwise exchange coupling is obviously
possible, and is consistent with our observations of similar
coupling in [ClMo(pyCH2CH2py)MoCl] (AMo = 2.5 mT) which
has a 10-atom bridge.13
The presence of a reduced 17 ve metal atom in these
trinuclear species appears to diminish significantly exchange
coupling between the paramagnetic centres, an observation we
have made previously in connection with di- and tri-nuclear
oxomolybdenum() complexes.28
Fig. 5 Magnetic coupling between (a) an equilateral triangle and
(b) an acyclic arrangement of three unpaired spins 1, 2 and 3 (the
representation of the spin orientation is for illustration only and has
no implications for the sign of J).
Conclusions
Our results show quite conclusively that we can create simple
‘J switches’, either by assembling two 17 ve and one 16 ve
molybdenum centre and reversibly reducing the latter or, alter-
natively, connecting one 17 ve metal centre to two 16 ve centres
and again reducing the latter. The switching phenomenon is
demonstrated by the molybdenum hyperfine splitting, which is
2.4–2.6 mT in a exchange-coupled two-spin system and ca. 1.6
mT in an exchange-coupled three-spin system.
The most dramatic result is obtained from 6 whose EPR
spectrum shows that the two spins on the 17 ve pyridino-
molybdenum termini are effectively isolated (|J| < 10 MHz =
0.33 × 10Ϫ3 cmϪ1) but on reduction to [6]Ϫ full three-spin
exchange is ‘switched on’. The effect is reversed on oxidation.
Finally, while we have demonstrated that ‘J switching’ effects
are dependent on the oxidation states of the three metal centres,
we note that their energies are far too small to be detected by
solid-state magnetic susceptibility methods. However, under
special conditions, particularly when |J| ≈ AMo, a good estimate
of |J| can be extracted by EPR spectral simulation. By this
method,10 we obtained a value of ca. 1400 MHz for |J| in 4,
which corresponds to 5.0 mT (ca. 0.05 cmϪ1). It remains to be
seen whether oligonuclear metal systems having higher spin
multiplicities and substantially stronger magnetic and elec-
tronic interactions can be made to function similarly.
Exchange coupling here occurs via a 13-atom delocalised
bridging ligand. The EPR spectra of the 17:17:17 ve species
[8]2Ϫ and [11]Ϫ are extremely similar to those of [1,3,5-{MoCl-
(pyE)} C H ], where E = CH᎐CH or C᎐C.14 In this last species
᎐
᎐
᎐
3
6
3
we envisage that the couplings |J1|, |J2| and |J3| (Fig. 5a) are all
equal and substantially greater than AMo, the exchange being
propagated via the bridging ligand, each pathway involving 15
atoms. The couplings in [8]2Ϫ and [11]Ϫ may also be regarded as
pairwise interactions, although the path lengths are different.
For [8]2Ϫ there will be two couplings via 8 atoms and one via 13
atoms, while in [11]Ϫ there are two 10-atom and one 15-atom
pathways. The important point is that exchange must involve
primarily the bridging ligand atoms (see below).
In the acyclic species 4, 5 and 6, spin-polarisation must
involve the molybdenum d orbitals. The effect of this is implied
in Fig. 5b, the symmetry of the three molecules requiring that
|J1| = |J2|. We consider that |J3| is negligible if, indeed, there
is any ‘through space’ or solvent-mediated exchange coupling
between the terminal paramagnetic centres. In order to explore
this possibility, and in the absence of suitable crystals for X-ray
studies, we constructed the energy-minimised structures of 4, 5
and 6 using the CAChe suite of programs (see Experimental
section).28 These calculations estimated the distances between
the terminal metal centres to be 30.3 in 4, 36.0 in 5 and 36.7 Å
in 6, probably too far to permit significant through-space or
solvent-mediated exchange coupling in solution. We presume
that the exchange between the three spins in solution is
accounted for by two pairwise interactions, expressed as J1 and
J2 (Fig. 5b). This would be consistent with the solid state
magnetic behaviour of [MoO(TpMe,Me)Cl(OC6H4O){MoO-
(TpMe,Me)}(OC6H4O)MoO(TpMe,Me)Cl], where the global value
of J is derived from consideration of the pairwise interaction
represented by the exchange spin Hamiltonian H = –J(S1S2 ϩ
S1S3) and the interaction between the terminal spins, i.e. the
J(S2S3) term (2 and 3, Fig. 5), J3, is negligible.24 Broadly similar
behaviour has been observed in chain-like organic tri-radicals,
although the EPR spectral information is less informative
because of the low abundance of nuclei with I = 1/2.29
Experimental
General
The starting materials [MoCl2],31 HOC6H4py and HOC6H4-
CH᎐CHpy14 were prepared by standard procedures. Solvents
᎐
for reactions and electrochemistry were carefully pre-dried, and
all reactions were carried out under dinitrogen. The new com-
pounds were usually purified by column or plate chromato-
graphy using silica gel 60 (70–230 mesh) with CH2Cl2 alone
or mixed with either n-hexane or THF as eluent. Chemical
reduction was achieved using freshly prepared cobaltocene in
a mixture of tetrahydrofuran and dichloromethane, and
oxidation was carried out by using dilute solutions of iodine in
the same solvent mixture.
The following instruments were used for routine spectro-
1
Assuming that |J3| is negligible or non-existent, the subtle
relationships between J1, J2 and AMo are determined by the
bridging ligand structure and transmitted through the central
{Mo} group. In 5, |J| must be very small (it is Ϫ6 cmϪ1 from
scopic work: H NMR spectroscopy, JEOL GX-270 or λ-300
spectrometer (in CDCl3 unless otherwise stated, J in Hz);
electron-impact and fast-atom bombardment (FAB) mass
spectrometry, VG-Autospec; EPR spectrometry, Bruker ESP-
300E; UV-VIS spectrophotometry, Perkin-Elmer Lambda 2;
FT-IR spectrometry, Perkin-Elmer 1600.
solid-state susceptibility measurements of [ClMo{py(CH᎐
᎐
CH)4py}MoCl]30) but it is obviously significantly larger than
AMo, the coupling being transmitted through the {Mo} centre
via two relatively planar bridges involving a 23-atom pathway.
The bridge in 4 is shorter (19 atoms) but, as we mentioned
above, there is rotational flexibility about the inter-ring bonds
and so delocalisation could be significantly reduced but not
Electrochemical measurements (cyclic and/or square wave
voltammetry) were made using a PC-controlled EC&G PAR
model 273A potentiostat, with platinum wire working and
counter electrodes, a saturated calomel electrode as reference,
pre-dried CH2Cl2 as solvent and [NBun ][PF6] (ca. 0.1 M) as
4
totally eliminated, leading to a situation where |J| ӷ AMo
.
base electrolyte. Metal complexes for electrochemical examin-
ation were ca. 10Ϫ3 M. Ferrocene was added as internal stand-
ard, and all potentials are quoted relative to the ferrocenium/
ferrocene couple.
Complex 6 contains two saturated CH2CH2 links in the bridges
and although the pathway contains the same number of atoms
as 5, this saturation and the rotational flexibility of the mole-
246
J. Chem. Soc., Dalton Trans., 2000, 241–249

View Article Online
For the modelling of 4, 5 and 6, each complex was drawn in
the CAChe Editor and the energy of the structure was sub-
sequently minimised in the molecular mechanics component of
the suite. The program attempts to minimise the overall energy
of the structure by an iterative process, using Allinger’s stand-
ard MM2 molecular mechanics force field,32 continuing until
the change in energy for a structure refinement is less than
0.004 kJ molϪ1, but we are aware that this final value does not
necessarily represent a global energy minimum, since the local
minimum can be highly dependent on the starting geometry.
3,5-Bis(4-hydroxyphenyl)pyridine (H2L4). This was prepared
by demethylation of MeL4 using pyridinium chloride at 200 ЊC
for 3 h according to a published procedure.9 After the reaction
mixture was cooled to 140 ЊC and water added, the organic
impurities were extracted with CH2Cl2 and the pH of the solu-
tion was adjusted to ca. 6 by adding aqueous NaOH. The
resulting precipitate was filtered off and crystallised from aque-
ous methanol giving the ligand as a white crystalline solid
(84%) (Found: C, 77.3; H, 4.9; N, 5.4. C17H13NO2 requires C,
77.5; H, 5.0; N, 5.3%); EI mass spectrum m/z 263 [Mϩ] (requires
263); 1H NMR δ 7.00 [4H, d, J = 8.8; phenyl H3, H3Ј, H5, H5Ј],
7.75 [4H, d, J = 8.8; phenyl H4, H4Ј, H6, H6Ј], 7.90 [1H, s;
pyridyl H2], 8.92 [2H, s; pyridyl H1, H1Ј].
Ligands
1-(4-Methoxyphenyl)-2-(4-pyridyl)ethane (MeL3). A suspen-
sion of trans-1-(4-methoxyphenyl)-2-(4-pyridyl)ethene (MeL2;
1 g, 4.7 mmol),14 0.1 g of catalyst (Pd on activated carbon) and
ethanol (100 cm3) was stirred under H2 overnight. The solution
was then filtered through Celite to remove the catalyst and any
unreacted starting material. The solvent was then evaporated
in vacuo leaving 1-(4-methoxyphenyl)-2-(4-pyridyl)ethane as a
white powder (0.63 g, 63%) (Found: C, 78.2; H, 7.5; N, 6.6.
C14H15NO requires: C, 78.8; H 7.1; N, 6.6%); EI mass spectrum
m/z 213 [Mϩ] (requires 213); 1H NMR δ 2.88 [4H, br s; H5, H5Ј,
H6, H6Ј], 3.79 [3H, s; OC(H)3], 6.81 [2H, d, J = 8.8; phenyl H3,
H3Ј], 7.04 [2H, d, J = 8.8; phenyl H4, H4Ј], 7.06 [2H, d, J = 6.0;
pyridyl H2, H2Ј], 8.47 [2H, d, J = 6.0; pyridyl H1, H1Ј].
2,6-Bis(4-ethenylpyridyl)-4-hydroxytoluene (HL5). The con-
version of 2,6-bis(4-ethenylpyridyl)-4-methoxytoluene (1.0 g)
to the corresponding phenol was achieved using molten
pyridinium chloride according to a published procedure.33
A
tan product was obtained (0.81 g) (Found: C, 80.3; H, 5.8;
N, 8.8. C21H18N2O requires C, 80.2; H, 5.8; N, 8.9%); EI mass
spectrum m/z 314 [Mϩ] (requires 314).
Metal complexes
Complexes of 4-(4-hydroxyphenyl)pyridine, 1-(4-pyridyl)-2-
(4-hydroxyphenyl)-ethene and -ethane. [Mo(OC6H4py)2], 1,
3,5-Bis(4-methoxyphenyl)pyridine (Me2L4). To a stirred solu-
tion of 3,5-dichloropyridine (1 g; 6.75 mmol) and [NiBr2-
(PPh3)2] (0.150 g) in THF (50 cm3) was added dropwise a
solution of the Grignard reagent prepared from 4-bromo-
anisole (3.5 g; 18.7 mmol) and Mg turnings (0.50 g; 20.6 mmol)
in dry THF (20 cm3). Stirring was continued overnight and,
following hydrolysis with aqueous NaHCO3, the organic
solvent was evaporated in vacuo. The residue containing Me2L4
was extracted with several portions of dichloromethane, the
fractions combined and the solvent reduced to low bulk in
vacuo. Ethanol was added giving a white precipitate of Me2L4
which was filtered off and washed several times with ethanol
affording the ligand as a white crystalline solid (0.63 g, 86%)
(Found: C, 77.9; H, 5.9; N, 5.1. C19H17NO2 requires C, 78.3; H
5.9; N, 4.8%); EI mass spectrum m/z 291 [Mϩ] (requires 292); 1H
NMR δ 3.87 [6H, s; 2 × OMe], 7.03 [4H, d, J = 8.8; phenyl H3,
H3Ј, H5, H5Ј], 7.57 [4H, d, J = 8.8; phenyl H4, H4Ј, H6, H6Ј],
7.96 [1H, s; pyridyl H2], 8.73 [2H, s; pyridyl H1, H1Ј].
[Mo(OC H CH᎐CHpy) ], 2, and [Mo(OC H CH CH py) ],
᎐
6
4
2
6
4
2
2
2
3. A mixture of the appropriate ligand (HL2, HL3 or HL4; 2.8
mmol), NEt3 (1 cm3) and [MoCl2] (0.593 g; 1.2 mmol) was
refluxed overnight in toluene (50 cm3). The solvent was evapor-
ated in vacuo, the residue redissolved in a minimum amount of
CH2Cl2 and the solution chromatographed on a preparative
TLC plate (20 cm × 20 cm) using CH2Cl2/THF (9:1 v/v) as
eluent. The major brown band was collected each time and
filtered through Celite to remove the silica gel. Each fraction
was then evaporated in vacuo to ca. 5 cm3 and the complexes
were precipitated, as brown microcrystalline powders, by
addition of n-pentane and collected by filtration, yielding
the monometallic complexes 1, 2 and 3, respectively.
[Mo(OC6H4py)2], 1 (Found: C, 59.4; H, 5.5; N, 15.4.
C37H38N9BO3Mo requires C, 59.6; H, 5.6; N, 15.6%): FAB mass
spectrum 764 (requires 764); IR 1654 cmϪ1 (νNO); λmax/nm
(ε/10Ϫ3 MϪ1 cmϪ1) 280 (26), 433 (6.0); Ef = Ϫ1.10 V.
[Mo(OC H CH᎐CHpy) ], 2 (Found: C, 61.4; H, 5.6; N, 14.4.
᎐
6
4
2
C41H42N9BO3Mo requires C, 61.5; H, 5.7; N, 14.7%): FAB mass
spectrum 817 (requires 816); IR 1655 cmϪ1 (νNO); λmax/nm
(ε/10Ϫ3 MϪ1 cmϪ1) 331 (51), 465 (15); Ef = Ϫ1.14 V.
[Mo(OC6H4CH2CH2py)2], 3 (Found: C, 61.1; H, 6.0; N, 14.5.
C41H46N9BO3Mo requires C, 61.3; H, 6.2; N, 14.6%): FAB mass
spectrum 822 (requires 820); IR 1651 cmϪ1 (νNO); λmax/nm
(ε/10Ϫ3 MϪ1 cmϪ1) 274 (sh), 425 (7.5); Ef = Ϫ1.31 V.
2,6-Bis(4-ethenylpyridyl)-4-methoxytoluene (MeL5). A mix-
ture of 2,6-dibromo-4-methoxytoluene (2.2 g, 7.85 mmol),
4-vinylpyridine (2.9 g, 28 mmol), triethylamine (3.8 g, 37.6
mmol), palladium acetate (45 mg, 0.20 mmol) and triphenyl-
phosphine (105 mg, 0.40 mmol) was heated and stirred in a
Schlenk tube under nitrogen at 100 ЊC (ca. 5 d). The solid mass
which had formed was dissolved in CH2Cl2 (100 cm3) and
extracted with water (200 cm3). The organic layer was collected,
dried with magnesium sulfate, filtered and evaporated in vacuo.
The desired compound was crystallised from acetone and dried
in vacuo giving an off-white powder (1.91 g) (Found: C, 80.2;
H, 6.1; N, 8.6. C22H20N2O requires C, 80.5; H, 6.1; N, 8.5%);
EI mass spectrum m/z 328 [Mϩ] (requires 328).
[Mo{OC H pyMoCl} ], 4, [Mo{OC H CH᎐CHpyMoCl} ],
᎐
6
4
2
6
4
2
5, and [Mo{OC6H4CH2CH2pyMoCl}2], 6. A mixture of the
appropriate mononuclear complexes 1, 2 or 3 (0.30 mmol),
NEt3 (1 cm3), and [MoCl2] (0.45 g; 0.9 mmol), was refluxed in
toluene (40 cm3) for 2 d. The solvent was evaporated in vacuo,
the residue dissolved in CH2Cl2 and the solution chromato-
graphed on silica gel. Initial elution with CH2Cl2/hexane
(9:1 v/v) gave unreacted [MoCl2] and the green by-product
[{MoCl}2(µ-O)]. Pure CH2Cl2 was then used to elute first small
amounts of binuclear complexes, which were discarded, and
later the relatively non-polar trimetallic complexes. Recrystal-
lisation from concentrated CH2Cl2 solutions by the addition of
pentane afforded the desired products 4, 5 or 6, respectively, as
brown powders.
1-(4-Hydroxyphenyl)-2-(4-pyridyl)ethane (HL3). This was
prepared in the same way as HL2,14 but using trans-1-(4-
methoxyphenyl)-2-(4-pyridyl)ethane (MeL3) (0.65 g, 3.2 mmol)
as starting material. Removal of the solvent at the final stage of
the reaction yielded the compound as a white powder (0.55 g,
85%) (Found: C, 78.0; H, 6.9; N, 7.4. C13H13NO requires C,
78.4; H, 6.6; N, 7.1%). EI mass spectrum m/z 199 [Mϩ] (requires
199); 1H NMR δ 2.87 [4H, br s; H5, H5Ј, H6, H6Ј], 6.73 [2H, d,
J = 8.6; phenyl H3, H3Ј], 6.94 [2H, d, J = 8.6; phenyl H4, H4Ј],
7.06 [2H, d, J = 6.1; pyridyl H2, H2Ј], 8.45 [2H, d, J = 6.1;
pyridyl H1, H1Ј].
[Mo{OC6H4pyMoCl}2], 4 (Found: C, 47.7; H, 5.2; N, 18.9.
C67H82N23B3Cl2O5Mo3 requires C, 46.9; H, 4.9; N, 19.2%): FAB
mass spectrum 1681 (requires 1681); IR 1670, 1616 cmϪ1 (νNO);
λmax/nm (ε/10Ϫ3 MϪ1 cmϪ1) 286(30), 325(38), 383(19), 428(18);
Ef = Ϫ2.11, Ϫ1.05, ϩ0.03 V.
J. Chem. Soc., Dalton Trans., 2000, 241–249
247

View Article Online
[Mo{OC H CH᎐CHpyMoCl} ], 5 (Found: C, 44.8; H, 4.7;
[MoCl{OC H Me[CH᎐CHpyMoCl]} ], 11 (Found: C, 46.8;
᎐
6 3 2
᎐
6
4
2
N, 16.4. C71H86N23B3Cl2O5Mo3 requires C, 44.7; H, 4.7; N,
16.2%): FAB mass spectrum 1732 (requires 1733); IR 1670,
1611 cmϪ1 (νNO); λmax/nm (ε/10Ϫ3 MϪ1 cmϪ1) 281(27), 369(53),
426(26), 489 (sh); Ef = Ϫ1.99, Ϫ1.06, ϩ0.04 V.
H, 5.5; N, 18.7. C66H83N23B3Cl3O4Mo3 requires C, 46.9; H, 5.5;
N, 19.1%): FAB mass spectrum 1689 (requires 1689); IR 1676,
1609 cmϪ1 (νNO); λmax/nm (ε/10Ϫ3 MϪ1 cmϪ1) 285 (sh), 327(35),
415 (sh), 487(9); Ef = Ϫ1.93, Ϫ0.88, ϩ0.03 V.
[Mo{OC6H4CH2CH2pyMoCl}2], 6 (Found: C, 44.8; H, 5.6;
N, 16.3. C71H90N23B3Cl2O5Mo3 requires C, 44.6; H, 4.6; N,
16.2%): FAB mass spectrum 1737 (requires 1737); IR 1655,
1617 cmϪ1 (νNO); λmax/nm (ε/10Ϫ3 MϪ1 cmϪ1) 276(29), 314 (sh),
417(12), 467 (sh); Ef = Ϫ2.12, Ϫ1.30, ϩ0.05 V.
Acknowledgements
We thank the EPSRC for the provision of a studentship to
P. K. A. S. and the European Commission for support through
contract EC CHRX CT94-0538 (to A.-A. J.). We are very
grateful to Professors Jaume Veciana, Jean-Pierre Launay and
Dante Gatteschi and their colleagues for many stimulating
conversations concerning ‘molecular magnetism’.
Complexes of 3,5-bis(4-hydroxyphenyl)pyridine. [{MoCl-
(OC6H4)}2py}], 7, and [{MoCl(OC6H4)}2pyMoCl], 8. A mix-
ture of H2L5 (0.08 g; 0.29 mmol), [MoCl2] (0.30 g; 0.6 mmol)
and NEt3 (1 cm3) in dry toluene (40 cm3) was refluxed for 8 h.
After cooling and evaporating the solvent in vacuo, the residue
was chromatographed on silica gel. Initial elution with CH2Cl2/
hexane (9:1 v/v) gave unreacted [MoCl2] and the green by-
product [{MoCl}2(µ-O)]. Pure CH2Cl2 was then used to separ-
ate the desired trinuclear complex 8. The solvent polarity was
then increased by adding THF (up to 2.5% v/v) and the first
major purple fraction contained the binuclear complex 7. Each
fraction was concentrated in vacuo to ca. 5 cm3 and complexes
7 and 8 were precipitated by addition of n-pentane.
References
1 D. A. Dougherty, Acc. Chem. Res., 1991, 24, 88; H. Iwamura and
N. Koga, Acc. Chem. Res., 1993, 26, 346; K. Yoshizawa and
R. Hoffman, J. Am. Chem. Soc., 1995, 117, 6921.
2 T. Sugawara and A. Izuoka, Mol. Cryst. Liq. Cryst., 1997, 305A, 41
and references therein.
3 O. Kahn, Molecular Magnetism, VCH, New York, 1993; R. J.
Bushby and J.-L. Paillaud, in Introduction to Molecular Electronics,
eds. M. C. Petty, M. R. Bryce and D. Bloor, Edward Arnold,
London, 1995, p. 72; A. Caneschi, D. Gatteschi and R. Sessoli,
Mol. Cryst. Liq. Cryst., 1996, 278A, 177; F. Paul and C. Lapinte,
Coord. Chem. Rev., 1998, 180, 431.
4 J. B. Goodenough, Phys. Rev., 1955, 100, 564; J. Phys. Chem. Solids,
1958, 6, 287; J. Kanamori, J. Phys. Chem. Solids, 1959, 10, 87; A. P.
Ginsberg, Inorg. Chim. Acta Rev., 1971, 5, 45.
5 O. Kahn, Struct. Bonding (Berlin), 1987, 68, 89; K. Nakatani, P.
Bergerat, E. Codjovi, C. Mathonière, Y. Pei and O. Kahn, Inorg.
Chem., 1991, 30, 3977; S. W. Gordon-Wylie, E. L. Bominaar, T. J.
Collins, J. M. Workman, B. L. Claus, R. E. Patterson, S. A.
Williams, B. J. Conklin, G. T. Yee and S. T. Weintraub, Chem. Eur.
J., 1995, 1, 528; F. Lloret, G. de Munno, M. Julve, J. Cano, R. Ruiz
and A. Caneschi, Angew. Chem., Int. Ed., 1998, 37, 135.
6 H. Iwamura, Pure Appl. Chem., 1993, 65, 57; J. S. Miller and
A. J. Epstein, Angew. Chem., Int. Ed. Engl., 1994, 33, 385; S. Rajca,
Chem. Rev., 1994, 94, 871; S. Rajca and A. Rajca, J. Am. Chem. Soc.,
1995, 117, 9172.
7 T. Weygand, T. Costuas, A. Mari, J. F. Halet and C. Lapinte,
Organometallics, 1998, 17, 5569; N. Le Narvor and C. Lapinte,
C. R. Acad. Sci., Ser. IIc: Chim., 1998, 1, 745.
8 J. A. McCleverty and M. D. Ward, Acc. Chem. Res., 1998, 31, 842.
9 R. Brière, R.-M. Dupeyre, H. Lemaire, C. Morat and A. Rassat,
Bull. Soc. Chim. Fr., 1965, 11, 3290.
[{MoCl(OC6H4)}2py], 7 (Found: C, 42.8; H, 5.1; N. 17.5.
C47H44N15B2Cl2O4Mo2 requires C, 48.5; H, 4.9; N, 17.3%): FAB
mass spectrum 1179 (requires 1179); IR 1683 cmϪ1 (νNO); λmax
nm (ε/10Ϫ3 MϪ1 cmϪ1) 275 (sh), 306(26), 502(16); Ef = Ϫ0.87 V.
/
[{MoCl(OC6H4)}2pyMoCl], 8 (Found: C, 45.3; H, 5.0; N,
18.9. C62H77N22B3Cl3O5Mo3 requires C, 45.5; H, 4.7; N, 18.8%):
FAB mass spectrum 1635 (requires 1637); IR 1685, 1616 cmϪ1
(νNO); λmax/nm (ε/10Ϫ3 MϪ1 cmϪ1) 279(40), 316 (sh), 492(17);
Ef = Ϫ2.20, Ϫ0.83, ϩ0.09 V.
[{MoCl(OC6H4)}2pyMe][PF6], 9[PF6]. A mixture of the start-
ing complex 7 (0.10 g; 0.085 mmol) and methyl iodide (1 cm3)
was refluxed in CH2Cl2 (20 cm3) for 18 h. After removal of the
solvent in vacuo the solution was chromatographed on a short
(12 cm long) column of alumina (Brockmann activity IV) using
MeCN (99%)/aqueous KPF6 (1%) as eluent. Small amounts of
by-products eluted first before the major product. To the major
brown fraction containing the pure product was added 5 cm3 of
saturated aqueous KPF6. After evaporation to dryness the solid
was extracted from the excess KPF6 by partitioning between
water and CH2Cl2. The product (9ϩ) was precipitated from the
dried (MgSO4), concentrated CH2Cl2 solutions by addition of
pentane. [{MoCl(OC6H4)}2pyMe][PF6], 9[PF6] (Found: C, 43.9;
H, 4.5; N, 15.2. C48H58N15B2Cl2F6O4Mo2P requires C, 4.44; H,
4.7; N, 15.2%): FAB mass spectrum 554 (requires 553); IR 1685
cmϪ1 (νNO); λmax/nm (ε/10Ϫ3 MϪ1 cmϪ1) 316(30), 389(17),
456 (sh); Ef = Ϫ1.55 (reduction of quaternary site), Ϫ0.74 V.
10 E. Psillakis and P. K. A. Shonfield, PhD Theses, University of
Bristol, 1997 and 1999; P. K. A. Shonfield, A. Behrendt, J. C. Jeffery,
J. P. Maher, J. A. McCleverty, E. Psillakis, M. D. Ward and
C. Western, J. Chem. Soc., Dalton Trans., 1999, 4341.
11 A. J. Amoroso, A. M. W. Cargill Thompson, J. P. Maher, J. A.
McCleverty and M. D. Ward, Inorg. Chem., 1995, 34, 4828.
12 A. Das, J. P. Maher, J. A. McCleverty, J. A. Navas and M. D. Ward,
J. Chem. Soc., Dalton Trans., 1993, 681.
13 S. L. W. McWhinnie, J. A. Thomas, T. A. Hamor, C. J. Jones,
J. A. McCleverty, D. Collison, F. E. Mabbs, C. J. Harding, L. J.
Yellowlees and M. G. Hutchings, Inorg. Chem., 1996, 35, 760.
14 A. Das, J. C. Jeffery, J. P. Maher, J. A. McCleverty, E. Schatz,
M. D. Ward and G. Wollerman, Inorg. Chem., 1993, 32, 2145.
15 D. C. Reitz and S. I. Wessman, J. Chem. Phys., 1960, 33, 700.
16 A. Hudson and G. R. Luckhurst, Mol. Phys., 1967, 13, 409.
17 A. M. W. Cargill Thompson, D. Gatteschi, J. A. McCleverty,
J. A. Navas, E. Rentschler and M. D. Ward, Inorg. Chem., 1996,
35, 2701.
18 E. Psillakis J. P. Maher, J. A. McCleverty and M. D. Ward,
Chem. Commun., 1998, 835.
19 J. P. Maher, J. A. McCleverty, M. D. Ward and A. Wlodarczyk,
J. Chem. Soc., Dalton Trans., 1994, 143.
20 J. A. McCleverty, G. Denti, S. J. Reynolds, A. S. Drake, N. El Murr,
A. E. Rae, N. A. Bailey, H. Adams and J. M. A. Smith, J. Chem.
Soc., Dalton Trans., 1983, 81.
21 S. M. Charsley, C. J. Jones, J. A. McCleverty, B. D. Neaves and
S. J. Reynolds, J. Chem. Soc., Dalton Trans., 1988, 301.
22 N. Al-Obaidi, M. Chaudhury, D. Clague, C. J. Jones, J. C. Pearson,
J. A. McCleverty and S. S. Salam, J. Chem. Soc., Dalton Trans.,
1987, 1733.
Complexes of 2,6-bis(4-ethenylpyridyl)-4-hydroxytoluene.
[MoCl{OC H Me[CH᎐CHpyMoCl]} ], 11, and [MoCl{OC -
᎐
6
3
2
6
H Me[CH᎐CHpyMoCl](CH᎐CHpy)}], 10. A mixture of HL7
᎐
᎐
3
(0.25 g, 0.8 mmol), [MoCl2] (1.57 g, 3.2 mmol) and NEt3 (2 cm3)
was stirred and refluxed in toluene (60 cm3) for 24 h. The
solvent was then removed in vacuo and the residue was purified
chromatographically over silica using CH2Cl2 containing THF
(2% v/v). The trinuclear complex 11 separated first, and was
isolated, after solvent removal, as a microcrystalline black
powder (0.59 g, 43%). Dinuclear 10 was eluted after 11 had
been removed, by increasing the THF content (to 5%) of
the CH2Cl2. It was isolated as a microcrystalline black powder
(0.18 g, 18%).
[MoCl{OC H Me[CH᎐CHpyMoCl](CH᎐CHpy)}],
10
᎐
᎐
6
3
(Found: C, 50.3; H, 5.6; N, 18.2. C47H61N16B2Cl2O3Mo2
requires C, 49.7; H, 5.1; N, 18.2%): FAB mass spectrum 1231
(requires 1231); IR 1676, 1607 cmϪ1 (νNO); λmax/nm (ε/10Ϫ3 MϪ1
cmϪ1) 280 (sh), 311(34), 489(7); Ef = Ϫ1.92, Ϫ0.91, ϩ0.03 V.
248
J. Chem. Soc., Dalton Trans., 2000, 241–249

View Article Online
Amoroso, T. N. Branston, A. Das, J. P. Maher, J. A. McCleverty,
23 J. A. McCleverty, J. A. Navas and M. D. Ward, J. Chem. Soc.,
Dalton Trans., 1994, 2415.
M. D. Ward and A. Wlodarczyk, Polyhedron, 1997, 16, 4353.
24 V. Â. Ung, S. Couchman, J. C. Jeffery, J. A. McCleverty, M. D.
Ward, F. Totti and D. Gatteschi, Inorg. Chem., 1999, 38, 365.
25 V. Â. Ung, D. A. Bardwell, J. C. Jeffery, J. P. Maher, J. A.
McCleverty, M. D. Ward and A. Williamson, Inorg. Chem., 1996,
35, 5290; V. Â. Ung, A. M. W. Cargill Thompson, D. A. Bardwell,
D. Gatteschi, J. C. Jeffery, J. A. McCleverty, F. Totti and M. D.
Ward, Inorg. Chem., 1997, 36, 3447.
28 CAChe Scientific, Beaverton, OR, 1994.
29 A. Rajca and S. Utampanya, J. Am. Chem. Soc., 1993, 115, 2396.
30 D. Morusef, DEA Thesis, Université Paul Sabatier, Toulouse,
1996–7.
31 S. J. Reynolds, C. F. Smith, C. J. Jones and J. A. McCleverty, Inorg.
Synth., 1985, 23, 4; A. S. Drane and J. A. McCleverty, Polyhedron,
1983, 2, 53.
26 D. R. Eaton, A. D. Josey, W. D. Phillips and R. E. Benson, Discuss.
Faraday Soc., 1962, 34, 77; D. R. Eaton, W. D. Phillips and
D. J. Caldwell, J. Am. Chem. Soc., 1963, 85, 397; R. H. Holm and
M. J. O’Connor, Prog. Inorg. Chem., 1971, 14, 241.
32 N. L. Allinger, J. Am. Chem. Soc., 1977, 99, 8127.
33 C. Dietrich-Buchecker and J.-P. Sauvage, Tetrahedron, 1990,
46, 503.
27 R. Cook, J. P. Maher, J. A. McCleverty, M. D. Ward and A.
Wlodarczyk, Polyhedron, 1993, 12, 2111; A. Abdul-Rahman, A. A.
Paper a908323e
J. Chem. Soc., Dalton Trans., 2000, 241–249
249