Antiferromagnetic spin-coupling between MnII and amminium radical cation ligands: Models for coordination polymer magnets

Article abstract of DOI:10.1016/j.poly.2007.09.020

One- and two-electron oxidation of the manganese(II) complex [L₂Mn(hfac)₂] (L = 4″,4?-di-tert-butyl-2′,2″,2?trimethoxy-{4-(4′-diphenylaminophenyl)pyridine}) were studied by ultra violet/visible/near infra red spectroscopy, cyclic voltammetry and magnetometry. A one-electron oxidation converts the triarylamine ligand to its radical cation and gives a complex in which the antiferromagnetic coupling between the spin on the ligand and that on the metal J/k_b is -1.5 K. In a dilute frozen matrix and at low temperature this behaves as an S = 2 system. A two-electron oxidation gives [L₂Mn(hfac)₂]^{2{radical dot}+} which at low enough temperatures behaves as an S = 3/2 system but the spin-coupling between the metal and the ligand is weaker (J/k_b = -0.3 K). The weakness of these spin-couplings mean that Mn^II/amminium radical cation complexes are not promising systems on which to base coordination polymer magnets. The equivalent copper(II) complex [L₂Cu(hfac)₂] was also investigated but this decomposes when an attempt is made to oxidise the ligand to its amminium radical cation.

Full text of DOI:10.1016/j.poly.2007.09.020

Available online at www.sciencedirect.com
Polyhedron 27 (2008) 383–392
www.elsevier.com/locate/poly
Antiferromagnetic spin-coupling between MnII and amminium
radical cation ligands: Models for coordination polymer magnets
*
Richard J. Bushby , Colin Kilner, Norman Taylor, Rhidian A. Williams
Self-Organising Molecular Systems (SOMS) Centre and School of Chemistry, University of Leeds, Leeds LS2 9JT, United Kingdom
Received 15 August 2007; accepted 27 September 2007
Available online 5 November 2007
Abstract
One- and two-electron oxidation of the manganese(II) complex [L2Mn(hfac)2] (L = 400,4000-di-tert-butyl-20,200,2000trimethoxy-
{4-(40-diphenylaminophenyl)pyridine}) were studied by ultra violet/visible/near infra red spectroscopy, cyclic voltammetry and magne-
tometry. A one-electron oxidation converts the triarylamine ligand to its radical cation and gives a complex in which the antiferromag-
netic coupling between the spin on the ligand and that on the metal J/kb is ꢀ1.5 K. In a dilute frozen matrix and at low temperature this
behaves as an S = 2 system. A two-electron oxidation gives [L2Mn(hfac)2]2Å+ which at low enough temperatures behaves as an S = 3/2
system but the spin-coupling between the metal and the ligand is weaker (J/kb = ꢀ0.3 K). The weakness of these spin-couplings mean
that MnII/amminium radical cation complexes are not promising systems on which to base coordination polymer magnets. The equiv-
alent copper(II) complex [L2Cu(hfac)2] was also investigated but this decomposes when an attempt is made to oxidise the ligand to its
amminium radical cation.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Keywords: Manganese(II); Copper(II); Molecular magnet; Crystal structures
1. Introduction
have been investigated [6,7]. This paper explores the possi-
bility of using amminium radical cation ligands. We have
used the monopyridyl functionalised triarylamine ligand 3
to make the model ‘monomers’ 4 and 5 (Scheme 1) and then
tried to oxidise the ligands to the amminium radical cation
level. The sign of the spin coupling between the ligand rad-
ical cation and the metal ion centre in the resulting 4nÅ+ and
5nÅ+ complexes is expected to depend on whether the spin
orbitals overlap or are orthogonal [7]. In the MnII com-
plexes, the non-zero overlap between spin orbitals should
result in an antiferromagnetic interaction but, in the case
of the copper complex the spin orbitals are orthogonal
and the interaction should be ferromagnetic [8].
One of the most successful strategies for producing
molecular magnets is that originally pioneered by Gatteschi
and co-workers [1,2] which relies on the creation of coordi-
nation polymers in which spin-bearing ligands and spin-
bearing transition metal centres alternate with each other.
At low enough temperatures, such systems can show bulk
magnetic properties regardless of whether the coupling
between metal and ligand is ferromagnetic or antiferromag-
netic [2]. Systems of this type include a vanadium/TCNE
complex which has a Curie temperature which is above
room temperature but in most such systems there is weak
or very weak spin-coupling between the ligand and the
metal and, as a result, Curie temperatures are also low [3].
Most other examples exploit nitroxide [4] or nitronyl nitrox-
ide [2,5] as the spin-bearing ligand although several others
2. Results
2.1. Synthesis and X-ray crystal structures
*
As shown in Scheme 1, the ligand 3 was prepared by a
Corresponding author. Tel.: +44 113 3436509; fax: +44 113 3436452.
Suzuki coupling reaction between the monoboronic acid
E-mail address: r.j.bushby@leeds.ac.uk (R.J. Bushby).
0277-5387/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2007.09.020

384
R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
O
N
OMe
N
OMe
OMe
N
Br
B
O
N
(i-iv)
(v)
MeO
OMe
MeO
OMe
MeO
OMe
1
2
3
F3C
N
CF3
OMe
OMe
O
O
MeO
(vi)
MeO
N
M
N
N
OMe
O
O
MeO
F3C
CF3
4, M = MnII
5, M = CuII
i
+
˚
Scheme 1. Synthesis of the complexes: Reagents: (i) n-BuLi/THF/ꢀ78 ꢁC; (ii) B(O Pr)3/THF/ꢀ78 ꢁC; (iii) H /H2O/ꢀ78 ꢁC; (iv) pinacol/4 A molecular
sieve/D, 46% (overall); (v) 4-bromopyridine hydrochloride/Pd(PPh3)4/Ba(OH)2/DME/Ar atmosphere/D; and (vi) [Cu(hfac)2] or [Mn(hfac)2]/heptane/
ether/DCM.
2 (from the bromocompound 1 [9]) and 4-bromopyridine.
The required 2:1 complexes 4 and 5 were obtained by mix-
ing solutions of the ligand with [Mn(hfac)2] or [Cu(hfac)2].
Recrystalisation of the complexes 4 and 5 from THF/
hexane gave orange and green crystals, respectively. Single
crystal X-ray structures were obtained for all three com-
pounds 3–5, the derived molecular structures are shown
in Fig. 1 and crystal structure refinement data are given
in Section 4 and Table 1. As expected, ligand 3 shows a
propeller-like twist of the aryl residues around an almost
planar central nitrogen. Like other triarylamines, the dihe-
dral angles range between 53ꢁ and 74ꢁ. The torsional angle
between the pyridine and benzene rings of 36ꢁ is also sim-
ilar to that in related compounds [10]. As expected, the
geometry of the ligand is only slightly changed on forming
the complexes 4 and 5 and complexation occurs through
the pyridyl rather than arylamine nitrogen. The Mn com-
plex 4 has dihedral angles between 56ꢁ and 68ꢁ with a tor-
sional angle of 29ꢁ and in the Cu complex 5 the dihedral
angles range between 56ꢁ and 69ꢁ with a torsional angle
of 31ꢁ.
radical cation 3Å+ is chemically stable on the CV timescale.
Both of the complexes 4 and 5 show oxidation potentials
and peak widths that are essentially the same as those of
the uncomplexed ligand 3.1 This is the first indication that
oxidation is occurring at the ligand rather than at the metal
centre: something that is confirmed by the spectroscopy
and magnetometry studies described below. This is not sur-
prising since oxidation of manganese(II) to manganese(III)
in similar complexes is known to require much higher
potentials [12]. If each of the ligands in turn is being oxi-
dised to the radical cation level and the complex remains
intact (for example, 4, [Ar3NMnIIAr3N] ! 4Å+, [Ar3NÅ+-
MnIIAr3N] ! 42Å+, [Ar3NÅ+MnIIAr3NÅ+]) one could argue
that there should be shifts in ligand oxidation potentials
arising from coulombic repulsion between the charge on
the ligand and that on the metal ion (as in 4Å+) and between
the charges on the two ligands (as in 42Å+). However, the
charged centres are well separated from each other and in
dichloromethane these systems are strongly ion-paired so
that the charge–charge repulsion is screened out by the
counter-ion. Based on data reported by McGill and co-
workers [13] for related ion-paired dication diradical spe-
cies in dichloromethane and given the separation between
the charged centres (which is known from crystal structure
data) the shifts in the oxidation potentials due to such cou-
lombic repulsions would be <25 mV. Such small differences
would be difficult to observe and so the similarity of the CV
data for complexed and uncomplexed ligands does not
indicate whether or not the complexes remain intact.
2.2. Cyclic voltammetry
A cyclic voltammogram for the ligand 3 is shown in
Fig. 2 [11]. There is a clean one-electron redox process at
0.71 V (versus silver/silver chloride) with a width at half
height of about 0.1 V. Both forward and reverse scans
match in amplitude and there are no additional peaks of
the type associated with benzidine formation, even after
many cycles [10]. This indicates that a fast reversible elec-
trochemical process is occurring and that the amminium
1
See supplementary material.

R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
385
Fig. 1. X-ray derived structures of the ligand 3 (top), manganese complex 4 (middle) and copper complex 5 (bottom). For the complexes the Mn and Cu
are on inversion centres. Ellipsoid probabilities are 50%.
2.3. Ultra violet/visible/near infra red spectroscopic studies
advantage of being dichloromethane soluble so that it can
be used in a quantitative titration. The results of these
titrations are shown in Fig. 3 and the numerical data
are summarised in Section 4. The electronic absorption
These studies were carried out using thianthrenium tet-
rafluoroborate (THBF4) as the oxidant [14]. This has the

386
R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
Table 1
Crystallographic data for compounds 3–5
Compound
3
4
5
Formula
C34H40N2O3
524.68
monoclinic
C2/c
17.2432(3)
17.0393(3)
20.5000(4)
90
97.005(2)
90
5978.2(2)
8
1.166
0.074
150(2)
C94H114F12MnN4O14
1806.83
triclinic
C94H114CuF12N4O14
1815.43
triclinic
Formula weight
Crystal system
Space group
ꢀ
ꢀ
P1
P1
˚
a (A)
11.9670(2)
13.9279(3)
15.5052(4)
83.1310(10)
74.3390(10)
66.5150(10)
2282.05(9)
1
11.8763(2)
13.9797(3)
15.4469(3)
81.8890(10)
73.0810(10)
66.5500(9)
2250.02(8)
1
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
3
˚
V (A )
Z
qcalc (g cmꢀ3
)
1.315
0.233
150(2)
1.340
0.333
150(2)
l (mmꢀ1
T (K)
)
Measured reflections
Independent reflections
Rint
46881
5838
0.0963
0.0642
0.1787
ꢀ0.230, 0.406
43394
8913
0.0643
0.0536
0.1440
ꢀ0.490, 0.566
43337
8847
0.0719
0.0591
0.1672
ꢀ0.628, 1.082
R1 (F)a [I > 2r(I)]
wR2 (F2)b (all ꢀd3ata)
˚
Dqmin,max (e A
)
P
P
a
R1 = [iFoj ꢀ jFci]/ jFoj.
P
P
b
wR2 = [ w(F2o ꢀ F2c)/ wFo4]1/2
.
Fig. 2. (Left) Cyclic voltammogram for 10ꢀ3 M solutions of ligand 3 in DCM with 0.1 M tetrabutylammonium hexafluorophosphate as the supporting
electrolyte. Potential relative to Ag/AgCl. (Right) Corresponding deconvoluted dI1/dE plot. Sweep rate 500 mV sꢀ1
.
spectrum (250–1250 nm) for the neutral amine 3 in dichlo-
romethane shows one main absorption band at 354 nm.
When it is oxidised with one molar equivalent of THBF4
the spectrum is replaced by that of the radical cation 3Å+
showing bands at 467, 620 and 920 nm. Because of the
lowered symmetry the D0–D1 transition is split [10,15].
There are sharp isosbestic points at 297 and 377 nm, indi-
cating a clean oxidation. Relative to the uncomplexed
ligand 3Å+, the spectrum of the complex 4Å+ (Fig. 3, bot-
tom) is slightly shifted with the main absorption band
at 370 nm. On adding another equivalent of THBF4, both
of the ligands are oxidised to the radical cation level. The
resulting diradical dication 42Å+ shows three bands at 469,
619 and 925 nm. These changes parallel those for the
uncomplexed ligand 3 but all of the bands are all slightly
shifted. This confirms that the complex does not dissoci-
ate upon oxidation. There is a clean isosbestic point for
the first one-electron oxidation step although this is less
clear for the second (Fig. 3, insets). Whereas the results
for the oxidation of the ligand 3 and of the manganese
complex 4 are straightforward and easy to interpret the
results obtained for the copper complex 5 are not. As
expected the electronic absorption spectrum for the neu-
tral complex 5 in dichloromethane shows one main
absorption band at 369 nm but on oxidation, using either
THBF4 or NOBF4 as the oxidant, a complex series of

R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
387
tage is that, at higher temperatures, the signal becomes
weak relative to the diamagnetic correction and the errors
can be larger. Magnetic measurements were made on the
neutral complex 4, the monoradical monocation 4Å+ and
diradical dication complex 42Å+. For the neutral Mn com-
plex 4, since we have a dilute frozen matrix, the individual
manganese(II) S = 5/2 centres are expected to behave inde-
pendently. It was found that the Curie law was obeyed and
vT was 4.35 emu K molꢀ1 which is about the value
expected for a dilute paramagnet of S = 5/2. This is con-
firmed by plots of the normalized magnetisation versus
the ratio of magnetic field and temperature (M/Msat versus
H/T) which is consistent with a Brillouin function of S = 5/
2 at both 2 and 5 K.1 Magnetic data for the mono and
diradical complexes were obtained by selectively oxidising
the sample with one or two equivalents of THBF4, respec-
tively. For the monoradical monocation 4Å+ the Curie law
is obeyed. Hence, the plot (1/v versus T) at low field
(0.5 T) is linear as shown in Fig. 4 (top) and intercepts
around 0 K. The product of susceptibility and temperature
versus temperature (vT versus T) gradually increases from
2 K and flattens out around 50 K giving a high temperature
vT value of 4.71 emu K molꢀ1 (Fig. 4, middle). This is close
to the value expected for a system with isolated S = 1/2 and
S = 5/2 spins. However, at low enough temperatures the
spins are antiferromagnetically coupled and the system
behaves as if comprised of isolated S = (5/2 ꢀ 1/2) = 2 spin
centres. This is confirmed by plots of the normalized mag-
netisation versus the ratio of magnetic field and tempera-
ture (M/Msat versus H/T) which fit reasonably closely to
a Brillouin function of S = 2 (Fig. 4, bottom). The strength
of the spin coupling (Jab) was determined by fitting the vT
versus T curve using the Bleaney–Bowers equation [16] for
a system with an antiferromagnetic interaction between
S = 5/2 and S = 1/2 spins [17].
Fig. 3. (Top) Ultra violet/visible/near infra red spectra for the oxidation
for monopyridyl ligand 3 to the radical cation species 3+. (Bottom) UV
spectra for the oxidation for Mn complex 4 to the diradical dication
species 42+. Oxidation performed in dichloromethane using THBF4 at
room temperature.
vT ¼ Ng2lB2 T=kBT ꢁ ½28 þ 10 expðꢀ6J=kBTÞꢂ=½7 þ 5
ꢃ expðꢀ6J=KBTÞꢂ þ Na
The experimental data best fits a weak intramolecular
antiferromagnetic coupling of J/kB = ꢀ1.5 K. For the dica-
tion 42Å+, the Curie law is obeyed, the plot (1/v versus T) at
low field (0.5 T) is linear as shown in Fig. 5 (top) and inter-
cepts around 0 K. The product of susceptibility and tem-
perature versus temperature (vT versus T) gradually
increases from 2 K and flattens out around 30 K giving a
high temperature vT value of 5.06 emu K molꢀ1 (Fig. 5,
middle). This value is close to that expected for a system
with isolated S = 5/2 and a pair of isolated S = 1/2 spins.
If, at low enough temperatures the spins are antiferromag-
netically coupled we should get magnetically isolated
S = (5/2 ꢀ 1/2 ꢀ 1/2) = 3/2 spin centres. This is confirmed
by plots of the normalized magnetisation versus the ratio of
magnetic field and temperature (M/Msat versus H/T) which
fit reasonably closely to a Brillouin function of S = 3/2 at 2
and 5 K (Fig. 5, bottom). The strength of the intramolecu-
lar coupling (Jab) was determined by fitting the vT versus T
changes occurs which is quite different to those normally
associated with amminium radical cation formation. It
appears that even though the amminium ion/copper com-
plexes are stable on the short time-scale of the cyclic vol-
tammetry experiments they are not stable on the longer
time-scale of these spectroscopic studies.
2.4. Magnetometry
Magnetometry studies were carried out using dilute fro-
zen solutions which were prepared by the mixture of syr-
inge and vacuum line techniques described previously [9].
This method minimises exposure to air during sample prep-
aration and transfer, and requires less material than using
dry powered samples. It also has the advantage that inter-
molecular spin interactions, which can complicate the
interpretation of the data, can be ignored. The disadvan-

388
R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
Fig. 4. (Top) ‘Curie’ law plot of the reciprocal of magnetic suscepti-
bility (v) vs. temperature for a frozen dilute solution of the radical
cation 4Å+. (Middle) Plot of product of susceptibility and temperature
(vT vs. T) fitted to the Bleaney–Bowers equation. (Bottom) Plot of
normalized magnetisation vs. ratio of field and temperature (Brillouin
plot).
Fig. 5. (Top) ‘Curie’ law plot of the reciprocal of magnetic susceptibility
(v) vs. temperature for a frozen dilute solution of the diradical dication
4
2Å+. (Middle) Plot of product of susceptibility and temperature (vT vs. T)
fitted to the Bleaney–Bowers equation. (Bottom) Plot of normalized
magnetisation vs. ratio of field and temperature (Brillouin plot).

R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
389
curve using the Bleaney–Bowers equation for a symmetri-
ferromagnetic
coupling
cal (S = 1/2)–(S = 5/2)–(S = 1/2) system with antiferro-
magnetic coupling between the S = 5/2 and S = 1/2
centres and zero coupling between the two S = 1/2 centres
[18].
radical
radical ion
or carbene
metal ion or
ferromagnetic
coupling unit
radical
radical ion
or carbene
~
~
ꢀ
ꢁ
v ¼ Ng2lB2 T =4kBT ꢁf½84expð5J=kBTÞþ35expðꢀ2J=kBTÞ
þ10expðꢀ7J=kBTÞþ35ꢂ=½4expð5J=kBT Þ
Cl
Cl
Cl
C
Cl
Cl
þ3expðꢀ2J=kBT Þþ2expðꢀ7J=kBTÞþ3ꢂgþNa
CuIL3 O C
O
C
C O
O
6
The experimental data best fits weak intramolecular anti-
ferromagnetic coupling with a value of J/kB = ꢀ0.3 K
(Fig. 5, middle).
Cl
Cl
+1/2
Cl
-1/2
Cl
+1/2
For a dilute frozen solution of the neutral CuII complex
5 the Curie law is obeyed and the plot intercepts around
0 K indicating that there is negligible intermolecular inter-
actions. vT is 0.39 emu K molꢀ1; a value which is consis-
tent with vT = 0.37 emu K molꢀ1 calculated for a dilute
paramagnet of S = 1/2, indicating that the unpaired d elec-
tron of copper(II) is magnetically isolated. Plots of the nor-
malized magnetisation versus the ratio of magnetic field
and temperature (M/Msat versus H/T) which is consistent
with a Brillouin function of S = 1/2 at 2 and 5 K.1
Cl
Cl
Cl
Cl
Cl
Cl
Cl
C
Cl
Cl
C
7
8
Cl
Cl
+1/2
+1/2
MnII(hfac)2
-5/2
N
C
C
N
+1
+1
3. Discussion
9
C
C
It is very difficult to design a high-spin polymer that is
purely organic [19,20] because both the range of ferro-
magnetic coupling units [21] and the number of stable free
radicals is very limited. On the other hand, coordination
polymers allow us to exploit the d orbitals of stable metal
complexes and because we can use two metals with differ-
ent moments or a metal and ligand with different
moments it is not even necessary to engineer ferromag-
netic coupling. This is a much easier strategy. One way
to view the relationship between the purely organic
high-spin polymers and the coordination polymers is illus-
trated in Fig. 6. As can be seen, in terms of the spin-cou-
pling pathways, either a transition metal centre or an
organic ferromagnetic coupling unit (metaphenylene,
3,40-disubstituted biphenyl or 4,400-disubstituted metater-
phenyl, etc.) [21] can have the same function. It mediates
an overall ferromagnetic coupling between radical, radical
ion or carbene centres. Hence, the copper(I) in 6 [22] has
the same function of feromagnetically coupling perchloro-
trityl centres as the tertrachlorometaphenylene in 7 [24],
the manganese(II) in 8 [25] has the same function of cou-
pling carbene centres as the metaphenylene in 9 [23] and
the manganese(II) in 102 has the same function of cou-
pling triarylamminium radical cation centres as the 4,400-
disubstituted metaterphenyl in 11 [9,13,26]. However, in
all systems, whether purely organic or organometallic,
the strength of the interaction is critical [9,10,20] Unfortu-
nately, couplings are often weak and so low Curie temper-
atures result. In our work aimed at purely organic systems
+1
+1
OMe
N
MeO
MnII(hfac)2
-5/2
N
N
10
N
+1/2
+1/2
MeO
OMe
11
N
N
+1/2
+1/2
Fig. 6. Schematic of the parallels between spin-coupling pathways in high-
spin organic polymers and those in coordination polymer magnets (see
text).
we used amminium ion spin-bearing centres ferromagnet-
ically coupled 1,3 through benzene, 3,40-through biphenyl
or (mostly) 4,400-through metaterphenyl [9,10,13,26] In the
present work, we have explored the coupling of similar
amminium ion spin-bearing centres through MnII. These
were generated by one- and two-electron oxidation of a
neutral triarlyamine/MnII complex 4. The complex
remained stable when the ligands were oxidised to the
amminium ion level but the antiferromagnetic spin-
coupling between amminium ion and MnII centres proved
to be weak in the monoradical monocation 4Å+ and
weaker still in the diradical dication 42Å+. This seems to
indicate a decrease in the overlap of the spin orbitals
when the second electron is removed. A possible explana-
tion is in terms of a space charge effect: that increased
2
This work.

390
R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
coulombic repulsion in the diradical dication leads to a
small extension of the Mn–pyridyl bonds and a decrease
in the d–p overlap. If space charge is indeed the cause
it is clear that using radical ion spin-bearing centres is a
less good strategy than using neutral radicals in designing
coordination polymer magnets.
4.2.3. Copper complex (5)
In the same manner, the complex 5 was obtained as
green crystals by recrystalisation from THF/hexane. Anal.
Calc. for C78H82N4O10F12Cu: C, 61.35; H, 5.41; N, 3.67.
Found: C, 61.20; H, 5.55; N, 3.40%.
4.3. X-ray crystallography
4. Experimental
Crystallographic data for 3, 4 and 5 were collected at
150 K using a Nonius Kappa CCD diffractometer (with
graphite-monochromated Mo Ka radiation; k = 0.71073
4.1. General details
˚
A) fitted with an Oxford Cryosystems nitrogen-cooled
The general procedures adopted in the synthetic work
were the same as those described in previous papers [9,10].
low-temperature device. Data processing for all structures
was carried out using DENZO [27] while an absorption cor-
rection was applied using SORTAV [28]. The structures were
all solved by direct methods and refined on F2 using
full-matrix least-squares methods with SHELXL-97 [29].
Hydrogen atoms were placed geometrically and treated
with a riding model. Data are summarised in Table 1.
In 400,4000-di-tert-butyl-20,200,2000 trimethoxy-{4-(40-diphenyl-
aminophenyl)pyridine} (3), the tert-butyl group centred
on C27 is disordered and was modelled over two positions
with relative occupancies of 0.71:0.29. The Mn atom in
the complex 4 lies on a centre of inversion and so the
asymmetric unit contains half the complex molecule
together with two disordered molecules of THF. Each
THF molecule was modeled over two positions with
relative occupancies of 0.73:0.27 in one case and
0.72:0.28 for the other. The Cu atom in the complex 5 lies
on a centre of inversion and so the asymmetric unit
contains half the complex molecule together with two
molecules of THF, one of which is disordered. The disor-
dered molecule was modeled over two positions with
equal occupancies.
4.2. Synthesis
4.2.1. 400,4000-Ditertbutyl-20,200,2000trimethoxy-
{4-(40-diphenylaminophenyl)pyridine} (3)
4-Bromopyridine hydrochloride (580 mg, 3.0 mmol) was
added to a mixture of ethylene glycol 1,2-dimethyl ether
(40 ml) and water (2 ml). Argon was bubbled through the
mixture for 0.5 h. Tetrakis-triphenylphosphine palladium
(0) (32 mg, 28 lmol) was added under a stream of argon.
The boronic ester 2 (1.3 g, 2.5 mmol) [9] was dissolved in
DME (10 ml) and degassed under a stream of argon for
0.5 h. The boronic ester solution was cannulated into the
reaction vessel, barium hydroxide (1.54 g, 10 mmol) was
finally added and the reaction mixture was refluxed for
48 h. The reaction was left to cool and the organic phase
was extracted with dichloromethane (3 · 20 ml) washed
with water (2 · 50 ml) and dried over anhydrous magne-
sium sulfate. Purification by flash chromatography gave
the amine 3 as small light red crystals (1.1 g, 84%). Anal.
Calc. for C34H40N2O3: C, 77.82; H, 7.68; N, 5.33. Found:
C, 77.50; H, 7.90; N, 5.20%. MS (EI+) m/z: 524.1 (M+,
100%).m.p. 167.3 ꢁC. 1H NMR (C6D6, 300 MHz) dH:
8.60 (2H, dd, J 7.2 and 1.55), 7.01 (2H, d, J 8.1), 7.0
(1H, m), 6.89 (2H, m), 6.86 (2H, d, J 2.14), 6.76 (2H, dd,
J 8.2 and 2.2), 3.28 (12H, s, OMe), 3.17 (6H, s, OMe),
1.15 (36H, s, Bu). 13C NMR (300 MHz, C6D6) dC: 154.2,
154.0, 153.8, 151.1, 147.9, 145.2, 140.5, 136.2, 126.1,
125.0, 121.3, 120.4, 118.6, 112.8, 111.8, 56.3, 56.2, 34.9,
32.0.
4.4. Cyclic voltammetry
These studies were carried out using a conventional
three electrode system coupled to an EG&G Model 362
scanning potentiostat and the system was controlled by a
PC unit running the ‘Condecon 320’ cyclic voltammetry
software. The working electrode was a small platinum disc,
ca. 2 mm diameter; the counter electrode was a platinum
coiled wire and the reference electrode, a silver wire
immersed in a saturated lithium chloride–dichloromethane
mixture containing the supporting electrolyte tetrabutyl-
ammonium hexafluoroborate (0.1 M). Prior to use, sol-
vents were thoroughly dried and degassed and the
concentrations for the amines were approximately
1 mg cmꢀ3. The ferrocene/ferrocenium couple (fast elec-
tron transfer) was used as a standard and its oxidation
potential was checked before and after each experiment
for the solvent used. For the voltammograms shown in
Fig. 2 the potential was swept in the anodic direction
(upper trace) and the lower trace represents the reverse
sweep in the cathodic direction.
4.2.2. Manganese complex (4)
A solution of [Mn(hfac)2 Æ 3H2O] (45 mg, 0.068 mmol)
was prepared in n-heptane (15 ml) by the addition of the
minimum amount of dry diethyl ether (ca. 1 ml). To this
solution was added a solution of ligand 3 (100 mg,
0.190 mmol) in dichloromethane (5 ml) under nitrogen.
The resulting solution was left under a flow of nitrogen
for 12 h. The resulting orange precipitate was collected.
The complex 4 was obtained as orange crystals (ca.
90 mg crude) and recrystalised from THF/hexane. Anal.
Calc. for C78H82N4O10F12Mn: C, 61.70; H, 5.44; N, 3.69.
Found: C, 61.50; H, 5.35; N, 3.90%.

R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
391
[3] J.M. Manriquez, G.T. Yee, R.S. McLean, A.J. Epstein, J.S. Miller,
Science 252 (1991) 1415.
4.5. Ultra violet/visible/near infra red spectroscopic studies
[4] (a) K. Inoue, H. Iwamura, J. Chem. Soc., Chem. Commun. (1994)
2273;
UV/Vis spectra in the range of 200–650 nm were
recorded using a Perkin–Elmer Lambda 3000 UV/Vis/
IR Spectrophotometer with a resolution of 0.1 nm in
dichloromethane. Experiments were carried out using a
pair of quartz cells with the corresponding solvent as ref-
erence. The path length of the cells was 10 mm. In a typ-
ical experiment, a dilute solution of the model diamine
(b) K. Inoue, T. Hayamizu, H. Iwamura, D. Hashizume, Y. Ohashi,
J. Am. Chem. Soc. 118 (1996) 1803.
[5] (a) K. Fegy, D. Luneau, E. Belorizky, M. Novac, J.L. Tholence, C.
Paulsen, T. Ohm, P. Rey, Inorg. Chem. 37 (1998) 4524;
(b) K. Fegy, D. Luneau, C. Paulsen, T. Ohm, P. Rey, Angew. Chem.,
Int. Ed. 37 (1998) 1270;
(c) A. Caneschi, D. Gatteschi, N. Lalioti, C. Sangregorio, R. Sessoli,
G. Venturi, A. Vindigni, A. Rettori, M.G. Pini, M.A. Novak, Angew.
Chem., Int. Ed. 40 (2001) 1760;
(d) M. Minguet, D. Luneau, C. Paulson, E. Lhotel, A. Gorski,
J. Waluk, D.B. Amabilino, J. Veciana, Polyhedron 22 (2003)
2349.
(1.6 · 10ꢀ5 mol lꢀ1
)
in dichloromethane was titrated
against a known concentration of THBF4 (1 mol lꢀ1), also
in dichloromethane [14]. Ligand 3, kmax 354 nm, 3.50 eV
(log10 e, 4.08); radical cation 3Å+, kmax 467 nm, 2.66 eV
(log10 e, 4.01); 620, 2.02 (3.56); 920(s), 1.34 (3.75); isos-
bestic points 3/3Å+, 297 nm; 377. Complex 4, kmax
370 nm, 3.31 eV (log10 e, 4.78); diradical dication 4Å+, kmax
354 nm, 3.50 eV (log10 e, 4.46); 469, 2.66 (4.58); 619, 2.00
(3.87); 925, 1.34 (4.01); isosbestic point 4/42Å+, 419 nm.
Complex 5, kmax 369 nm, 3.40 eV (log10 e, 4.56).
[6] (a) N. Koga, Y. Ishimaru, H. Iwamura, Angew. Chem., Int. Ed. 35
(1996) 755;
(b) Y. Sano, M. Tanaka, N. Koga, K. Matsuda, H. Iwamura, P.
Rabu, M. Drillon, J. Am. Chem. Soc. 119 (1997) 8246;
(c) S. Karasawa, H. Kumada, N. Koga, H. Iwamura, J. Am. Chem.
Soc. 123 (2001) 9685.
[7] Y. Takano, Y. Kitagawa, T. Onishi, Y. Yoshioka, K. Yamaguchi, N.
Koga, H. Iwamura, J. Am. Chem. Soc. 124 (2002) 450.
[8] W.T. Borden, E.R. Davidson, J. Am. Chem. Soc. 99 (1977)
4587.
4.6. Magnetometry
[9] R.J. Bushby, N. Taylor, R.A. Williams, J. Mater. Chem. 17 (2007)
955.
[10] R.J. Bushby, C. Kilner, N. Taylor, M.E. Vale, Tetrahedron 63 (2007)
11458.
[11] (a) K.B. Oldham, J. Spanier, J. Electroanal. Chem. 26 (1970)
331;
(b) K.B. Oldham, Anal. Chem. 44 (1972) 196;
(c) M. Grenness, K.B. Oldham, Anal. Chem. 44 (1972) 1121.
[12] J. Britten, N.G.R. Hearns, K.E. Preuss, J.F. Richardson, S. Bin-
Salamon, Inorg. Chem. 46 (2007) 3934.
Magnetic measurements were carried out using a
SQUID magnetometer (MPMS, Quantum Design) at the
University of Sheffield. Samples were prepared as a dilute
frozen matrix using the same apparatus and combination
of vacuum line and syringe techniques as previously
described in detail in Ref. [9].
Acknowledgements
[13] R.J. Bushby, D.R. McGill, K.M. Ng, N. Taylor, J. Chem. Soc.,
Perkin Trans. 2 (1997) 1405.
[14] B. Boduszek, H.J. Shine, J. Org. Chem. 53 (1988) 5142.
[15] S. Amthor, B. Noller, C. Lambert, Chem. Phys. 316 (2005) 141.
[16] B. Bleaney, K.D. Bowers, Proc. Roy. Soc. Lond. Ser. A 214 (1952)
451.
We thank EPSRC for funding this work and Dr. H.
Blythe (Physics, University of Sheffield) for assistance with
the SQUID magnetometry.
[17] H. Kumada, A. Sakane, N. Koga, H. Iwamura, J. Chem. Soc.,
Dalton Trans. (2000) 911.
[18] M.H. Dickman, L.C. Porter, R.J. Doedens, Inorg. Chem. 86 (1986)
2595.
Appendix A. Supplementary material
CCDC 634674, 186580 and 636674 contain the supple-
mentary crystallographic data for ligand 3, Mn complex
4 and Cu complex 5. These data can be obtained free of
charge via http://www.ccdc.cam.ac.uk/conts/retrieving.
html, or from the Cambridge Crystallographic Data Cen-
tre, 12 Union Road, Cambridge CB2 1EZ, UK; fax:
(+44) 1223-336-033; or e-mail: deposit@ccdc.cam.ac.uk.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.poly.2007.
09.020.
[19] (a) K. Itoh, in: D. Gatteschi, O. Kahn, J.S. Miller, F. Palacio (Eds.),
Magnetic Molecular Materials, Kluwer, Dordrecht, 1991;
(b) M.M. Murray, P. Kaszynski, D.A. Kaisaki, W.H. Chang, D.A.
Dougherty, J. Am. Chem. Soc. 116 (1994) 8152;
(c) R.J. Bushby, D.R. McGill, K.M. Ng, in: O. Kahn (Ed.),
Magnetism a Supermolecular Function, Kluwer, Dordrecht, 1996;
(d) P.M. Lathi, in: P.M. Lathi (Ed.), Magnetic Properties of Organic
Materials, Marcel Dekker, New York, 1999;
(e) A. Rajca, J. Wongsriratanakul, S. Rajca, Science 294 (2001)
1503;
(f) N.A. Zaidi, S.R. Giblin, I. Terry, A.P. Monkman, Polymer 45
(2004) 5683.
[20] (a) A.H. Oka, T. Tamura, Y. Miura, Y. Teki, J. Mater. Chem. 11
(2001) 1364;
References
(b) A. Rajca, Chem. Rev. 94 (1994) 871.
[21] D.A. Dougherty, Acc. Chem. Res. 24 (1991) 88.
[22] D. Maspoch, D. Ruiz-Molina, K. Wurst, J. Vidal-Gancedo, J.
Veciana, Dalton Trans. (2004) 1073.
[1] (a) A. Caneschi, D. Gatteschi, P. Rey, R. Sessoli, Inorg. Chem. 27
(1988) 1756;
[23] (a) K. Itoh, in: D. Gatteschi, O. Kahn, J.S. Miller, F. Palacio (Eds.),
Magnetic Molecular Materials, Kluwer, Dordrecht, 1991;
(b) K. Matsuda, N. Nakamura, K. Takahashi, K. Inoue, N. Koga, H.
Iwamura, ACS Sym. Ser. 644 (1996) 142.
(b) A. Caneschi, D. Gatteschi, J.P. Renard, P. Rey, R. Sessoli, Inorg.
Chem. 28 (1989) 1976, 2940 and 3314.
[2] A. Caneschi, D. Gatteschi, R. Sessoli, P. Rey, Acc. Chem. Res. 22
(1982) 392.

392
R.J. Bushby et al. / Polyhedron 27 (2008) 383–392
[24] J. Veciana, C. Rovira, M.I. Crespo, O. Armet, V.M. Domingo, F.
Palacio, J. Am. Chem. Soc. 113 (1991) 2552.
[25] S. Karasawa, H. Kumada, N. Koga, H. Iwamura, J. Am. Chem. Soc.
123 (2001) 9685.
[26] (a) R.J. Bushby, D. Gooding, M. Thornton-Pett, M.E. Vale, Mol.
Cryst. Liq. Cryst. 334 (1999) 167;
(c) R.J. Bushby, in: J.S. Miller, M. Drillon (Eds.), Magnetism:
Molecules to Materials II, Molecule-Based Materials, Wiley, New
York, 2001.
[27] Z. Otwinowski, W. Minor, Methods Enzymol. 276 (1997) 307.
[28] R.H. Blessing, Acta Crystallogr., Sect. A 51 (1995) 33.
[29] G.M. Sheldrick, SHELXL-97: Program for the Refinement of
Crystal Structures, University of Go¨ttingen, Go¨ttingen, Germany,
1997.
(b) R.J. Bushby, D. Gooding, M.E. Vale, Philos. Trans. Roy. Soc. A
357 (1999) 2939;