Synthesis and coordination chemistry of a potential precursor to a triarylamminium radical cation ditopic ligand

Article abstract of DOI:10.1016/j.poly.2012.06.059

The multi-step synthesis of a new ditopic ligand (7) is reported, which is a potential precursor to a triarylamminium radical cation. The preparation and full characterization of three new bimetallic first row transition metal [M = Mn (8), Ni (9), Cu (10)] coordination complexes containing this ligand are described, including detailed electrochemical and magnetic studies. Of note, a ground triplet state is observed in the bimetallic Cu complex. Chemical oxidation of these precursor complexes generated persistent complexes containing the dication of the ligand, which is confirmed through UV-Vis and EPR spectroscopies. The variable temperature magnetic properties of the dicationic complexes are described and included in the discussion are results from density functional theory calculations of bimetallic radical cation complexes.

Full text of DOI:10.1016/j.poly.2012.06.059

Polyhedron 52 (2013) 1118–1125
Contents lists available at SciVerse ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Synthesis and coordination chemistry of a potential precursor
to a triarylamminium radical cation ditopic ligand
Kseniya Revunova a, Serge I. Gorelsky b, Martin T. Lemaire a,
⇑
a Department of Chemistry, Brock University, 500 Glenridge Avenue, St. Catharines, Ontario, Canada L2S 3A1
b Center for Catalysis Research and Innovation, Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 3 July 2012
The multi-step synthesis of a new ditopic ligand (7) is reported, which is a potential precursor to a triary-
lamminium radical cation. The preparation and full characterization of three new bimetallic first row
transition metal [M = Mn (8), Ni (9), Cu (10)] coordination complexes containing this ligand are described,
including detailed electrochemical and magnetic studies. Of note, a ground triplet state is observed in the
bimetallic Cu complex. Chemical oxidation of these precursor complexes generated persistent complexes
containing the dication of the ligand, which is confirmed through UV–Vis and EPR spectroscopies. The
variable temperature magnetic properties of the dicationic complexes are described and included in
the discussion are results from density functional theory calculations of bimetallic radical cation
complexes.
We dedicate this manuscript to the
centennial anniversary of Alfred Werner’s
Nobel Prize in Chemistry.
Keywords:
Triarylamine ligands
Coordination chemistry
Bimetallic complexes
Magnetic properties
Electrochemistry
DFT calculations
Ó 2012 Published by Elsevier Ltd.
1. Introduction
radical anions [21]. In fact, Long and Evans’ recent report [22] of sin-
gle molecule magnet behavior in Dy dimers containing N23ꢀ radical
Spin control in molecular systems has been a topic of enormous
interest in the areas of physical organic and coordination chemis-
try for over 25 years [1–5]. Finding ways to promote ferromagnetic
interactions between electrons to align their spin properties to
produce structures with a high spin ground state is a very active
area of research in modern inorganic chemistry. The applied bene-
fits that arise from isolating high spin molecules are well known
and include new single molecule magnets [6–8], magnetic refriger-
ants [9,10] and paramagnetic contrast agents [11], to name a
few.
The metal-radical approach is one synthetic strategy that has
been applied toward the preparation of coordination complexes
with large S values in their ground states [12,13]. This tactic requires
the use of stable free radicals as ligands and takes advantage of a
strong direct magnetic exchange coupling that is generated by coor-
dination of spin density rich donor atoms to paramagnetic transition
metal ions. As another advantage, the sign of the metal-radical mag-
netic exchange coupling can often be predicted in metal-radical
complexes on the basis of the Goodenough–Kanamori orbital sym-
metry rules [14,15]. Stable radicals that have been used as ligands
include nitroxides [16], nitronyl and imino nitroxides [17], verdaz-
yls [18], thiazyls [19], and semiquinone [20] or other N-centered
anion bridging ligands underscores the importance of strong
intramolecular exchange coupling in efforts to increase blocking
temperatures in molecular nanomagnets.
A large number of metal-radical complexes have been reported,
but these tend to contain one maybe two metal ions at best. There
are very few reports of metal-radical complexes with more than
two metal ions, and many of these have been produced serendipi-
tously (here we refer to discrete molecules; there are a large num-
ber of reported one-, two- and three-dimensional metal-radical
coordination polymers). We [23] and others [24,25] have been
interested in the use triarylamminium radical cations as ligands
for transition metal ions. Ultimately, we are interested in the syn-
thesis of large dendritic-type polytriarylamino precursor ligands
for the production of polymetallic complexes. Chemical oxidation
of these precursor complexes is anticipated to produce metal-
radical complexes with interesting magnetic properties. Recently
we reported the synthesis of 5-(4,40-dimethoxydiphenylamino)-
2,20-bipyridine (Fig. 1, left) and M(hfac)2 complexes (M = Mn, Ni,
Cu) containing this ligand [23]. Chemical or electrochemical oxida-
tion of these precursor complexes produced metal-triarylammini-
um radical cation complexes and in the case of M = Mn, we
observed ferromagnetic metal-radical exchange coupling. We
could rationalize the magnetic properties of these complexes on
the basis of a
p-spin polarization mechanism for exchange cou-
⇑
Corresponding author. Tel.: +1 905 688 5550x3400; fax: +1 905 682 9020.
E-mail address: mlemaire@brocku.ca (M.T. Lemaire).
pling (supported by density functional theory (DFT) calculations).
0277-5387/$ - see front matter Ó 2012 Published by Elsevier Ltd.
http://dx.doi.org/10.1016/j.poly.2012.06.059

K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
1119
Fig. 1. Ligand reported in [23] (11) and ditopic ligand (7) described in the current work.
In an effort to extend this work to produce bimetallic com-
plexes, herein we describe the synthesis of ditopic ligand 7. Bime-
tallic M(hfac)2 complexes (M = Mn 8, Ni 9, Cu 10) containing 7
were prepared and fully characterized, including variable temper-
the sample transport rod. The magnetization of the samples was
scanned over a 5–300 K temperature range for 8–10 and for the
oxidation products of 8–10. Diamagnetic corrections to the para-
magnetic susceptibilities were accomplished using Pascal’s con-
stants. EPR spectra were recorded as acetonitrile solutions at
room temperature in quartz tubes on a BrukerElexsys E580 pulsed
and CW X-band (9 GHz) spectrometer.
ature magnetometry experiments, which identified
a triplet
ground state for 10. These bimetallic precursors are subject to
chemical oxidation and the nature of the oxidation products are
analyzed with the aid of DFT calculations.
2.4. Computational details
2. Materials and methods
All DFT calculations were performed using the GAUSSIAN 09 pack-
age [26] using the B3LYP hybrid functional [27,28] and the TZVP
basis set [29] for all atoms. Tight SCF convergence criteria were
used for all calculations. For each given spin state, the structures
were optimized. The converged wave functions were tested to con-
firm that they correspond to the ground-state surface. The evalua-
tion of atomic charges and spin densities was performed using the
natural population analysis (NPA) [30]. The generation of initial
guess wavefunctions for the spin states with the anti-ferromag-
netic coupling, the analysis of molecular orbitals in terms of frag-
ment orbital contributions, Mayer bond order calculations
[31,32] were carried out using the AOMIX program [33,34].
2.1. General procedures
All reagents were commercially available and used as received
unless otherwise stated. Deaerated and anhydrous solvents were
obtained from a Puresolve PS MD-4 solvent purification system,
and all air and/or moisture sensitive reactions were carried out
using standard Schlenk techniques. 1H/13C NMR spectra were re-
corded on a Bruker Advance 300 (or 600 as indicated) MHz spec-
trometer with
a 7.05 (or 14.1) T Ultrashield magnet using
deuterated solvents. FT-IR spectra were recorded on a Shimadzu
IRAffinity spectrometer as KBr discs. EI/FAB mass spectra were ob-
tained using a Kratos Concept 1S High Resolution E/B mass spec-
trometer. UV–Vis spectra were recorded in CH3CN solution on a
Shimadzu 3600 UV–Vis–NIR spectrophotometer using quartz cuv-
ette cells. Elemental analyses were carried out by Guelph Chemical
Laboratories LTD, Guelph, ON, Canada.
2.5. Synthesis
2.5.1. 5-Bromo-2-pyridinecarboxaldehyde (1)
Following general procedure of Wang et al. [35], a solution of
BuLi (2.5 M in hexanes, 10.0 ml, 25.3 mmol) was slowly added to
a solution of 2,5-dibromopyridine (5.0 g, 21 mmol) in dry toluene
(250 ml) at ꢀ78 °C and the mixture was stirred for 5 h. Next, dry
DMF (2.15 ml, 27.6 mmol) was added. After stirring for 1 h at
ꢀ78 °C, the solution was warmed up to ꢀ10 °C and the reaction
was quenched with saturated NH4Cl aqueous solution. The organic
phase was separated and the aqueous phase was washed twice
with ethyl acetate. All organic fractions were collected, dried over
MgSO4 and the solvent was removed by rotary evaporation. Col-
umn chromatography (silica gel, hexane/ethyl acetate 15:1) pro-
duced pure 1 (1.8 g, 45%). All characterization is consistent with
the literature values.
2.2. Electrochemical measurements
Cyclic voltammetry (CV) experiments were performed with a
Bioanalytical Systems Inc. Epsilon electrochemical workstation.
Compounds 7–11 were dissolved in anhydrous solvent (CH3CN),
filtered, and then deaerated by sparging with N2 gas for 20 min.
Analyte concentrations were approximately 10ꢀ3 M containing
0.1 M supporting electrolyte (Bu4NPF6). A typical three-electrode
set-up (a glassy carbon working electrode, Ag/AgCl reference elec-
trode, and a platinum wire counter electrode) was used. Ferrocene
was used in all cases as an internal standard to calibrate the refer-
ence electrode and was oxidized at a potential of +0.51 V in our set
up. The scan rate for all CV experiments was 100 mV/s or as other-
wise stated in the text.
2.5.2. 5-Bromo-2-(1,3-dioxolan-2-yl)pyridine (2)
Ethylene glycol (2.0 ml, 36 mmol) and p-toluenesulfonic acid
(400 mg, 2.30 mmol) were added to a solution of 5-bromo-2-pyri-
dinealdehyde (4.0 g, 22 mmol) in toluene (30 ml) and reaction
mixture was refluxed with Dean–Stark apparatus for 2 days. After
cooling to room temperature, the mixture was quenched with 10%
Na2CO3 aqueous solution and extracted with ethyl acetate. Organic
fractions were washed with water, brine, and then dried over
MgSO4. Solvent was removed by rotary evaporation. The product
was purified by column chromatography (silica gel, hexane/ethyl
acetate 15:1) to produce 4.2 g (85%) of a pure brown oil. 1H NMR
2.3. Variable temperature magnetic susceptibility measurements and
EPR spectroscopy
Solid state variable temperature magnetic susceptibility mea-
surements were recorded on a superconducting quantum interfer-
ence device (SQUID) magnetometer (Quantum Design MPMS) with
a 5.5 T magnet (temperature range 1.8–400 K) in an external field
of 2000 Oe. Samples of 8–12 were carefully weighed into gelatin
capsules, which were loaded into plastic straws, and attached to
(300 MHz, CDCl3)
d 8.64 (d, 1H, J = 2.3 Hz), 7.83 (dd, 1H,

1120
K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
J = 8.3 Hz, 2.3 Hz), 7.40 (d, 1H, J = 8.3 Hz), 5.78 (s, 1H), 4.07 ppm (m,
4H).
(w), 1762 (w), 1636 (m), 1585 (w), 1569 (w), 1531 (w), 1508 (s),
1437 (w), 1388 (m), 1315 (m), 1294 (m), 1246 (s), 1166 (w),
1135 (m), 1091 (s), 1026 (s), 943 (m), 898 (w), 875 (w), 829 (m),
750 (m), 722 (m), 700 (m), 642 (w), 524 cmꢀ1 (m).
2.5.3. 5-Iodo-2-(1,3-dioxolan-2-yl)pyridine (3)
5-Bromo-2-(1,3-dioxolan-2-yl)pyridine (0.90 g, 3.9 mmol) and
N,N0-dimethylethylenediamine (34 mg, 0.39 mmol) were added to
a solution of copper (I) iodide (37 mg, 0.19 mmol) and sodium io-
dide (1.17 g, 7.8 mmol) in 10 ml 1,4-dioxane (prior to the reaction
inorganic salts were recrystallized and dried in vacuo). The reaction
mixture was refluxed for 3 days. After cooling to room temperature
the reaction was diluted with 25% NH4OH aqueous solution and
extracted with CHCl3. The organic fractions were washed with
brine and dried over MgSO4. Column chromatography (silica gel,
hexane/ethyl acetate 15:1), which gave yellow-white shiny crys-
tals (865 mg, 80%). 1H NMR (300 MHz, CDCl3): d 8.81 (d, 1H,
J = 1.9 Hz), 8.04 (dd, 1H, J = 7.9, 1.9 Hz), 7.32 (d, 1H, J = 8.3 Hz),
5.79 (s, 1H), 4.10 ppm (m, 4H). 13C NMR (75.5 MHz, CDCl3): d
155.4, 145.1, 122.4, 103.1, 65.6 ppm. HRMS, Calc. for C8H8NO2I:
276.9599. Found: 276.95949. FT-IR (KBr): 3449 (w), 2965 (w),
2896 (m), 2875 (m), 2750 (w), 1840 (w), 1635 (w), 1572 (m),
1552 (w), 1468 (w), 1381 (s), 1364 (s), 1206 (m), 1129 (m), 1024
(m), 1004 (s), 975 (s), 943 (s), 862 (m), 825 (m), 753 (w), 717
(w), 668 (m), 624 cmꢀ1 (m).
2.5.6. 5,50-(4-Methoxyphenylazanediyl)dipicolinaldehyde (6)
N-(6-(1,3-Dioxolan-2-yl)pyridin-3-yl)-6-(1,3-dioxolan-2-yl)-N-
(4-methoxyphenyl)pyridin-3-amine (0.10 g, 0.24 mmol) was
stirred with 5% HCl (3 ml) in THF (5 ml) for 2 days. The reaction
mixture was made basic with NaOH pellets and extracted into
ethyl acetate. The organic layer was dried over MgSO4 and the sol-
vent was removed by rotary evaporation. A pure, yellow product
was obtained by column chromatography (silica gel, hexane/
EtOAc, 4:1), 68 mg (85%). 1H NMR (300 MHz, CDCl3): d 9.96 (s,
2H), 8.49 (d, 2H, J = 2.3 Hz), 7.86 (d, 2H, J = 8.7 Hz), 7.47 (dd, 2H,
J = 8.7, 2.3 Hz), 7.12 (m, 2H), 6.96 (m, 2H), 3.84 ppm (s, 3H). 13C
NMR (75.5 MHz, CDCl3): d 191.7, 147.4, 146, 143.4, 136.2, 128.6,
128.2, 122.7, 116, 55.6 ppm. HRMS, Calc. for
C19H15N3O3:
333.11134. Found: 333.11055. FT-IR (KBr): 3429 (w), 3045 (w),
3008 (w), 2929 (w), 2812 (w), 2709 (w), 2507 (w), 2359 (w),
2034 (w), 1702 (s), 1555 (s), 1480 (s), 1387 (w), 1319 (s), 1287
(s), 1247 (s), 1211 (s), 1030 (s), 819 (s), 736 (w), 669 (w), 617
(w), 547 cmꢀ1 (w).
2.5.4. 6-(1,3-Dioxolan-2-yl)-N-(4-methoxyphenyl)pyridin-3-amine
(4)
2.5.7. 5,50-(4-Methoxyphenylazanediyl)dipicolinaldehyde (7)
Compound 6 (0.20 g, 0.60 mmol) was stirred with 2 equivalents
of p-anisidine (0.15 mg, 1.2 mmol) and dry MgSO4 (0.15 g) in DCM
(5 ml) for 2 days. The reaction mixture was then filtered and the
solvent was removed by rotary evaporation. Crude purification
was first carried out by column chromatography (silica gel, hex-
ane/EtOAc, 5:1). Further purification of 7 was accomplished by dis-
solving the chromatographed product into isopropanol (2 ml) and
precipitating it out with weak HCl solution. The product was fil-
tered off, shaken with 10% NaOH solution and then extracted into
ethyl acetate. The solvent was removed by rotary evaporation to
produce a yellow powder, 280 mg (85%). This product is mildly
susceptible to hydrolysis and was stored under an inert atmo-
sphere in the refrigerator. 1H NMR (300 MHz, CDCl3): d 8.56 (s,
2H), 8.42 (d, 2H, J = 2.6 Hz), 8.06 (d, 2H, J = 9.1 Hz), 7.45 (dd, 2H,
J = 8.3, 2.6 Hz), 7.30 (m, 4H), 7.13 (d, 2H, J = 9.1 Hz), 8.93 (m, 6H),
3.84 (s, 3H), 3.83 ppm (s, 6H). 13C NMR (150 MHz, CDCl3): d
158.7, 158, 157.4, 149, 144.1, 144, 143.4, 137.5, 128.8, 128,
122.6, 122.2, 115.6, 114.4, 55.5, 30.9 ppm. Anal. Calc. for
A Schlenk flask charged with 5-iodo-2-(1,3-dioxolan-2-yl)pyri-
dine (72 mg, 0.26 mmol), p-anisidine (16 mg, 0.13 mmol), palla-
dium (II) acetate (3 mg, 0.013 mmol), BINAP (16 mg, 0.026 mmol)
and Cs2CO3 (170 mg, 0.52 mmol) was evacuated and back-filled
with dinitrogen (3ꢁ) Next, dry toluene (2 ml) was syringed into
the flask and the reaction was stirred at 50 °C for 2 days. Conver-
sion of starting material was monitored by TLC (product Rf = 0.36
eluent, EtOAc). After complete consumption of starting material
was noted the reaction mixture was cooled to room temperature,
filtered and the solvent was removed by rotary evaporation. The
product was purified by column chromatography (hexane/EtOAc,
5:1) to produce a dark oil, 28 mg (40%). 1H NMR (300 MHz, CDCl3):
d 8.21 (d, 1H, J = 2.3 Hz), 7.33 (d, 1H, J = 8.7 Hz), 7.19 (dd, 1H, J = 8.3,
2.6 Hz), 7.06 (m, 2H), 6.86 (m, 2H), 5.77 (s, 1H), 5.65 (s, 1H), 4.10
(m, 4H), 3.79 ppm (s, 3H). 13C NMR (150 MHz, CDCl3): d 156.1,
147.1, 142.1, 137.6, 134, 123, 121.2, 121.2, 114.8, 103.8, 65.4,
55.5 ppm. HRMS, Calc. for
C15H16O3N2: 272.11609. Found:
272.11595. FT-IR (KBr): 3397 (w), 3267 (w), 2955 (w), 2890 (w),
2834 (w), 1593 (m), 1577 (m), 1511 (s), 1465 (w), 1440 (w),
1388 (w), 1330 (w), 1287 (w), 1244 (m), 1180 (w), 1132 (w),
1090 (m), 1032 (m), 987 (w), 943 (w), 829 (s), 751 (s), 518 cmꢀ1 (s).
C
33H29O3N5: C, 72.93; H, 5.34; N, 12.89. Found: C, 73.17; H, 5.58;
N, 12.73%. HRMS, Calc. for 33H29O3N5: 543.22704. Found:
C
543.22773. FT-IR (KBr): 3448 (w), 1618 (w), 1560 (w), 1506 (s),
1477 (m), 1319 (w), 1246 (m), 1260 (w), 1031 (w), 829 cmꢀ1 (w).
2.5.5. N-(6-(1,3-Dioxolan-2-yl)pyridin-3-yl)-6-(1,3-dioxolan-2-yl)-N-
2.5.8. [Mn(hfac)2]2(7) (8)
(4-methoxyphenyl)pyridin-3-amine (5)
Mn(hfac)2ꢂ3H2O (96.3 mg, 0.184 mmol) was dissolved in a mix-
ture of DCM (2 ml) and MeOH (ꢃ0.5 ml) and combined with a solu-
tion of 7 (50 mg, 0.092 mmol) in DCM (5 ml). The mixture was
stirred for 2 days at room temperature. The solvent volume was re-
duced to 2 ml and after several hours, precipitation of bright
brown-orange solid was observed. The product was collected by
vacuum filtration, and dried. Yield 68 mg (50%). Anal. Calc. for C53-
H33N5O11F24Mn2: C, 42.94; H, 2.25; N, 4.73. Found: C, 43.14; H,
2.09; N, 4.79%. MS (FAB ꢀ): m/z 1481 (Mꢀ, 0.3%), 262 [Mn(hfac),
2.5%], 207 (hfac, 100%). FT-IR (KBr): 3448 (s), 2364 (w), 1647 (m),
1617 (w), 1559 (w), 1500 (m), 1316 (w), 1256 (s), 1201 (m),
1146 (s), 1096 (m), 1032 (m), 797 (w), 664 (w), 583 cmꢀ1 (w).
A
Schlenk flask charged with 6-(1,3-dioxolan-2-yl)-N-
(4-methoxyphenyl)pyridin-3-amine (0.15 g, 0.55 mmol), 5-iodo-
2-(1,3-dioxolan-2-yl)pyridine (0.15 g, 0.55 mmol), palladium (II)
acetate (6.0 mg, 0.027 mmol), BINAP (34.3 mg, 0.055 mmol) and
Cs2CO3 (0.36 g, 1.1 mmol) was evacuated and back-filled with dini-
trogen (3ꢁ). Dry toluene (2 ml) was syringed into the flask and the
reaction was stirred at 95 °C for 3 days. The reaction mixture was
cooled to room temperature, filtered and the solvent was removed
by rotary evaporation. Column chromatography (hexane/EtOAc,
5:1) produced 5 as a pure brown oil 123 mg (53%). 1H NMR
(300 MHz, CDCl3): d 8.31 (d, 2H, J = 0 Hz), 7.36 (m, 4H), 7.04 (d,
2H, J = 8.7 Hz), 6.86 (d, 2H, J = 8.7 Hz), 5.79 (s, 2H), 4.11 (m, 8H),
3.79 ppm (s, 3H). 13C NMR (150 MHz, CDCl3): d 157.5, 143.8, 138,
135.3, 129.4, 127.6, 121.2, 120.9, 115.4, 103.3, 65.6, 55.5 ppm.
HRMS, Calc. for C23H23O5N3: 421.16377. Found: 421.16402. FT-IR
(KBr): 3423 (s), 3061 (w), 2965 (m), 2890 (m), 2362 (w), 2328
2.5.9. [Ni(hfac)2]2(7) (9)
Ni(hfac)2ꢂH2O (87 mg, 0.184 mmol) was dissolved in a mixture
of hexanes (3 ml) and DCM (3 ml) and combined with a DCM solu-
tion of 7 (50 mg, 0.092 mmol). After stirring at room temperature

K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
1121
for approximately 12 h the solvent volume was reduced to 2 ml
and the reaction mixture was cooled on ice. An orange-brown pre-
cipitate was collected by vacuum filtration and dried. Yield 75 mg
(55%). Anal. Calc. for C53H33N5O11F24Ni2: C, 42.73; H, 2.23; N, 4.70.
Found: C, 42.73; H, 1.98; N, 4.72%. MS (FAB ꢀ): m/z 1489 (Mꢀ,
0.3%), 472 [Ni(hfac)2, 26.3%], 207 (hfac, 100%). FT-IR (KBr): 3431
(s), 2921 (w), 2363 (w), 2339 (w), 1644 (m), 1617 (w), 1559 (w),
1507 (m), 1490 (w), 1424 (w), 1317 (w), 1256 (s), 1202 (m),
1149 (s), 1089 (m), 1059 (m), 1033 (m), 870 (w), 837 (w), 794
(w), 668 (w), 586 (w), 558 (w), 467 cmꢀ1 (w).
reactions where only slight modifications of the established meth-
ods were required [35–37]. The non-facile preparation of 4 and 5
required the use of palladium catalyzed C–N bond formation meth-
odology; so-called Buchwald–Hartwig conditions [38,39]. Tertiary
amine 5 could not be prepared cleanly in one step from anisidine
using two or more equivalents of 3 at low or elevated tempera-
tures. In all attempted reactions, significant amounts of both sec-
ondary and tertiary amine were produced, including also a large
proportion of dehalogenated 3. Unfortunately, we were unable to
separate the mixture of amines by column chromatography and
so attempted the preparation of 5 in two steps. Starting with a sin-
gle equivalent of 3, we could readily prepare pure secondary amine
4 with a reasonable yield. Reaction between 4 and a second equiv-
alent of 3 at elevated temperatures produced pure tertiary amine 5,
which was deprotected and reacted with two equivalents of anisi-
dine to produce the condensation product 7 as a hydrolytically
unstable yellow powder. Storage of 7 in an inert environment is
necessary to prevent decomposition.
Bimetallic complexes containing ligand 7 were readily pro-
duced by combining two equivalents of hydrated M(hfac)2 salts
(M = Mn, Ni, Cu) with 7 in CH2Cl2 solution (Fig. 2), which produced
analytically pure, microcrystalline precipitates of complexes 8
(Mn), 9 (Ni), or 10 (Cu). Complex 10 produced what appeared to
be X-ray diffraction quality crystals via slow evaporation from con-
centrated methanol solutions but in all our attempts to acquire an
X-ray structure, data collection was extremely weak due to poor
crystal diffraction. We are confident in the structural interpretation
noted in Fig. 2 based on all of our spectroscopic data acquired from
these complexes.
2.5.10. [Cu(hfac)2]2(7) (10)
Cu(hfac)2ꢂH2O (91.3 mg, 0.184 mmol) was dissolved in DCM
(3 ml) and added to DCM solution of 7 (50 mg, 0.092 mmol). After
stirring at room temperature for 12 h precipitation of a red-brown
microcrystalline solid was noted, which was isolated by vacuum
filtration and dried. Yield 80 mg (58%). Anal. Calc. for C53H33N5O11-
F24Cu2: C, 42.45; H, 2.22; N, 4.67. Found: C, 43.14; H, 2.09; N, 4.79%.
MS (FAB ꢀ): m/z 477 [Cu(hfac)2, 62%], 207 (hfac, 100%). FT-IR (KBr):
3432 (m), 2364 (w), 1648 (m), 1617 (w), 1582 (w), 1546 (m), 1508
(s), 1425 (w), 1368 (w), 1305 (w), 1256 (s), 1202 (s), 1147 (s), 1084
(w), 1032 (w), 943 (w), 896 (w), 836 (w), 794 (w), 741 (w), 667 (w),
586 (w), 527 (w), 419 cmꢀ1 (w).
2.5.11. Chemical oxidation of 7–10
About 15 mg each of 7–10 were dissolved in 2 or 3 ml of dry,
deoxygenated CH3CN in a 25 ml round bottom flask. One equiva-
lent of NOPF6 solution (CH3CN) was added by syringe to produce
in all cases but complex 10, a clear red solution. The solutions were
stirred under a dinitrogen atmosphere for 10 min and then the sol-
vent was removed in vacuo to produce solids that were immedi-
ately loaded into gel caps for squid magnetometry experiments.
3.2. Electronic spectra, electrochemistry and magnetic properties of 7–
10
3. Results and discussion
The electronic spectra of ligand 7 and complexes 8–10 were
investigated via UV–Vis absorbance spectroscopy and cyclic vol-
tammetry. The absorption spectra of 7–10 in CH3CN solution are
shown in Fig. 3. Ligand 7 features a dominant visible absorption
band centered at 410 nm, which is n-p⁄ in origin. Complexes 8–
10 each exhibit this same ligand centered transition but at longer
wavelengths in all cases relative to 7. No additional bands that
3.1. Ligand synthesis and coordination chemistry
The multi-step synthesis of ligand 7, starting from commer-
cially available 2,5-dibromopyridine, is outlined in Scheme 1. The
first three steps culminating in iodo-substituted 3 are literature
Scheme 1. Synthesis of 7.

1122
K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
The variable temperature magnetic properties of 8–10 were
investigated by SQUID magnetometry in a 5–300 K temperature
range, in an external field of 2000 Oe (Fig. 5A–C). For each complex,
the data were fitted [40] to numerical expressions derived from the
van Vleck equation for simple bimetallic complexes containing Mn,
Ni or Cu, based upon the following isotropic spin Hamiltonian:
^
^ ^
H ¼ JSMSM
ð1Þ
The best fit parameters for
q
8
(g = 2.0, JMnMn = ꢀ0.025 cmꢀ1
,
,
= 0.025, h = ꢀ0.4 K, R = 0.0086) and 9 (g = 2.26, JNiNi = ꢀ0.20 cmꢀ1
TIP = 900 ꢁ 10ꢀ6 cm3 molꢀ1 Kꢀ1, h = ꢀ0.9 K, R = 0.028) each include
very weak intra- and intermolecular anti-ferromagnetic coupling
terms. However, only very slightly poorer fits can also be obtained
with very small ferromagnetic coupling terms, so the actual sign of
the coupling is not unambiguous. It is best to describe complexes 8
and 9 as very weakly coupled systems. On the other hand, the best
fit parameters for complex 10 suggest a weak, but measureable fer-
romagnetic coupling may be operative been the copper ions
Fig. 2. Structures of complexes 8–10 (hfac is 1,1,1,5,5,5-hexafluoroacetylacetonato).
(g = 2.30, JCuCu = +2.0 cmꢀ1
,
q
= 0.020, TIP = 500 ꢁ 10ꢀ6 cm3 molꢀ1
Kꢀ1, h = ꢀ0.6 K, R = 0.0031). In an effort to confirm a ground triplet
state for 10, a low temperature (4 K) magnetization versus field
experiment was performed (Fig. 5D). While the magnetization does
not saturate up to 5 T, the experimental data points are similar to
the theoretical curve calculated for a triplet state at 4 K (assuming
a g parameter equal to that obtained from the fit of the magnetic
susceptibility data). The observed ferromagnetic coupling in 10
can be rationalized through a
(Fig. 6) [41,42]. The positive spin density on Cu (indicated as an
up arrow) is likely concentrated in a dx2 based molecular orbital;
p-spin polarization mechanism
Fig. 3. Electronic absorption spectra of 7–10 (CH3CN solution).
ꢀy2
as this orbital is orthogonal to the bridging ligand
p-system, spin
delocalization imparts spin density of the same sign on the coordi-
nated N atom from the pyridine ring. Alternation of the spin density
to the other Cu ion results in a ferromagnetic interaction.
could be attributed to metal-centered transitions (d–d) or metal–
ligand charge transfer transitions are observed in the visible region
of the spectra of these complexes.
Cyclic voltammograms were acquired for each of 7–10 in aceto-
nitrile solution and the CVs of 7 and 10, as examples, are displayed
in Fig. 4. In the oxidation scan, ligand 7 exhibits three closely
spaced and irreversible waves centered at +0.6, 0.9 and 1.2 V (ver-
sus ferrocene). The first and second waves likely represent produc-
tion of radical-cation and dication, respectively. The third wave is
of an unknown origin, but may represent a metal-centered oxida-
tion (for complexes 8–10), or oxidation of a decomposition product
formed from one of the first two oxidation processes. In the reduc-
tion scan, the CV of 7 is unremarkable, featuring an irreversible
imine reduction centered at ꢀ2.3 V. In complexes 8–10 the same
pattern of three irreversible oxidative processes are noted, but
are shifted to higher potentials relative to 7 (Fig. 4, right). The
waves are also more closely spaced than in 7, with the first two
waves separated by only approximately 100 mV. This is in stark
contrast to 11, in which M(hfac)2 coordination stabilized the radi-
cal-cation and produces reversible electrochemistry at potentials
lower than in the free ligand. Complexes 8–10 also exhibit irrevers-
ible cathodic waves at potentials centered at ꢀ1.6 and ꢀ2.4 V,
which represent hfac and imine reductions, respectively.
3.3. Oxidations of 7–10
In an effort to produce the radical cation of ligand 7 and com-
plexes 8–10, chemical oxidations were carried out in acetonitrile
solution by careful addition of NO(PF6) (one equivalent) at room
temperature under anhydrous and aerobic conditions. Instantly, a
color change to deep cherry-red was noted in all reactions upon
oxidant addition. With the exception of complex 9, the visible
absorption spectra of the oxidation products are all very similar
and feature a longer wavelength band (relative to the neutral pre-
cursors) together with a higher energy shoulder (Fig. 7). Addition
of a second (or more) equivalent of oxidant results in no further
changes to the absorption spectra. Indeed, these absorption spectra
are very different from those produced by chemical oxidation of
M(hfac)2 complexes containing 11 and the spectra of other re-
ported triarylamminium radical cations and are in fact more repre-
sentative of spectra produced from reported diamagnetic
dicationic species [43,44]. To probe this further we attempted to
obtain the EPR spectra of oxidized 7 and 8 at room temperature
Fig. 4. Cyclic voltammograms of ligand 7 (left) and complex 10 (right) in CH3CN solution containing 10ꢀ3 M nBu4N(PF6). In all CV experiments that scan rate was 100 mV/s.

K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
1123
A
B
2
D
C
1.6
1.2
0.8
0.4
0
0
25000
50000
H (Oe)
Fig. 5. Plots A, B, and C: Effective magnetic moment vs. temperature plots for complexes 8 (h), 9 (4), and 10 (s) (2000 Oe external field). The lines through the points
represent best fits based on appropriate spin Hamiltonians with best fit parameters as described in the text. Plot D: Reduced magnetization vs. field data for complex 10 at
4 K. The solid line represents the theoretical curve for an S = 1 state at 4 K (g = 2.3).
Fig. 6. Proposed spin density distribution in 10.
Fig. 8. Proposed structure of dication complexes of 8 and 10.
in acetonitrile. We obtained no EPR spectrum from oxidized 7 and
the spectrum obtained from 8 (Fig. S1 in the Supporting Informa-
tion) was identical to that of the neutral precursor complex
(Fig. S2 in the Supporting Information). These results taken to-
gether suggest the production of rather stable dications by oxida-
to have structures as outlined in Fig. 8. Given the irreversible nat-
ure and meagre separation (less than 200 mV) between the pur-
ported oxidation waves for radical cation and dication observed
in the CV it is clear that isolating persistent radical cation com-
plexes is not possible with the given system.
tion of 7, 8, and 10; oxidation of
9 appears to result in
decomposition. The dications of complexes 8 and 10 are proposed
2
1.5
1
2
1.5
1
0.5
0.5
0
0
330
430
530
630
330
430
530
630
Wavelength (nm)
Wavelength (nm)
Fig. 7. Left: Visible absorption spectra of 7 (solid line) and 72+ (dashed line). Right: Visible absorption spectra of 8 (solid line) and 82+. Spectra were recorded at room
temperature in CH3CN.

1124
K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
10
8
best. DFT calculations for the dication of complex 8 suggest a sin-
glet ground state (Fig. S3).
3.4. DFT calculations
3.4.1. M(hfac)2-radical cation (7+Å) complexes (M = Ni, Mn)
6
For comparison, the structures, relative energies and spin den-
sity distributions were calculated for hypothetical bimetallic
M(hfac)2 (M = Ni, Mn) containing the radical cation of ligand 7. Cal-
culations were performed on the S = 11/2 and 9/2 states of the
Mn(hfac)2 complex, which corresponds to MnII–N radical ferro-
magnetic and anti-ferromagnetic coupling, respectively. The spin
density distributions for the two geometry-optimized structures
are shown in Fig. 10. The individual MnII ions are in the high-spin
configuration and NPA-derived atomic spin densities for the Mn
atoms are 4.74–4.75 for both S = 11/2 and 9/2 states, respectively
(Fig. 10). The MnII-ligand bonds in these complexes have fairly
small covalency [32] as can be seen from the Mn–O and Mn–N
bond orders (0.25–0.29 and 0.21–0.23, respectively). The complex
with the S = 9/2 spin that corresponds to the anti-ferromagnetical-
ly coupled MnII–N radical description is calculated to have a lower
energy than the complex with the ferromagnetic spin interactions
by 91 cmꢀ1. This is in contrast to the reported Mn(hfac)2 complex
with the radical cation of 11, which was demonstrated to exhibit a
weak ferromagnetic Mn-radical spin interaction, which was ratio-
4
2
0
40
80
120
160
Temperature (K)
Fig. 9. Effective magnetic moment vs. temperature plots for the oxidation products
of complexes 8 ( ) and 10 ( ) (2000 Oe external field).
j
d
The variable temperature magnetic properties of dication 8 and
10 were investigated by SQUID magnetometry over a 5–150 K tem-
perature range and indicate essentially no coupling between the
metal ion spins (Fig. 9). Given the uncertainty in the mass and pur-
ity of these products we did not attempt to fit these data but the
magnitude of the magnetic moments and curve profiles with
decreasing temperature suggest extremely weak interactions at
Fig. 10. Calculated spin density distributions in optimized structures for the S = 11/2 and 9/2 states of the radical cation of complex 8. The NPA-derived atomic spin densities
for the metal ions and the triarylamminium nitrogen are shown. Spin densities are shown with the isosurface contour value of 0.004.
Fig. 11. Calculated spin density distributions in optimized structures for the S = 5/2 and 3/2 states of the radical cation of complex 9. The NPA-derived atomic spin densities
for the metal ions and the triarylamminium nitrogen are shown. Spin densities are shown with the isosurface contour value of 0.004.

K. Revunova et al. / Polyhedron 52 (2013) 1118–1125
1125
nalized on the basis of a
p-spin polarization mechanism (sup-
References
ported by DFT calculations). The spin density distribution in the
bimetallic complex is different from Mn(hfac)211, including essen-
tially zero spin density on the coordinate N atom of the pyridine
ring or the C atom bonded to the amminium N atom.
[1] D.A. Dougherty, Acc. Chem. Res. 24 (1991) 88.
[2] O. Kahn, Inorg. Chim. Acta 62 (1982) 3.
[3] A. Rajca, Chem. Rev. 94 (1994) 871.
[4] H. Iwamura, J. Phys. Org. Chem. 11 (1998) 299.
[5] T. Glaser, H. Theil, M. Heidemeier, C. R. Chim. 11 (2008) 1121.
[6] R. Sessoli, A.K. Powell, Coord. Chem. Rev. 253 (2009) 2328.
[7] M. Murrie, Chem. Soc. Rev. 39 (2010) 2986.
[8] T. Glaser, Chem. Commun. 47 (2011) 116.
[9] M. Evangelisti, E.K. Brechin, Dalton Trans. 39 (2010) 4672.
[10] Y.-Z. Zheng, M. Evangelisti, R.E.P. Winpenny, Chem. Sci. 2 (2011) 99.
[11] Y. Wang, W. Li, S. Zhou, D. Kong, H. Yang, L. Wu, Chem. Commun. 47 (2011)
3541.
[12] A. Caneschi, D. Gatteschi, R. Sessoli, P. Rey, Acc. Chem. Res. 22 (1989) 392.
[13] M.T. Lemaire, Pure Appl. Chem. 76 (2004) 277.
[14] J.J. Kanamori, J. Phys. Chem. Solids 10 (1959) 87.
[15] J.B. Goodenough, J. Phys. Chem. Solids 6 (1958) 287.
[16] A. Caneschi, D. Gatteschi, P. Rey, Prog. Inorg. Chem. 39 (1991) 331.
[17] D. Luneau, P. Rey, Coord. Chem. Rev. 249 (2005) 2591.
[18] B.D. Koivisto, R.G. Hicks, Coord. Chem. Rev. 249 (2005) 2612.
[19] K.E. Preuss, Dalton Trans. 23 (2007) 2357.
Calculations were also performed on the S = 5/2 or 3/2 states of
the bimetallic Ni(hfac)2 complex (Fig. 11) of same radical cation li-
gand of 7. The individual NiII ions are in the high-spin configuration
and NPA-derived atomic spin densities for the Ni atoms are 1.65 for
both S = 5/2 and 3/2 states, respectively. Due to lower-energy
acceptor metal orbitals of the metal ion [32], the NiII-ligand bonds
have stronger covalency as can be seen from the Ni–O and Ni–N
bond orders (0.33–0.35 and 0.33–0.34, respectively) relative to
the MnII analog (see above). The energy of the S = 3/2 state (with
NiII-radical anti-ferromagnetic spin coupling) is lower by 56 cmꢀ1
relative to the S = 5/2 state (with the ferromagnetically coupled
ions).
[20] D.A. Shultz, Comment Inorg. Chem. 23 (2002) 1.
[21] J.S. Miller, Chem. Soc. Rev. 40 (2011) 3266.
[22] J.D. Rinehart, M. Fang, W.J. Evans, J.R. Long, Nat. Chem. 3 (2011) 538.
[23] S. Adugna, K. Revunova, B. Djukic, S.I. Gorelsky, H.A. Jenkins, M.T. Lemaire,
Inorg. Chem. 49 (2010) 10183.
[24] R.J. Bushby, C. Kilner, N. Taylor, R.A. Williams, Polyhedron 27 (2008) 383.
[25] M. Yano, K. Inoue, T. Motoyama, Y. Azuma, M. Tatsumi, O. Yamauchi, M.
Oyama, K. Sato, T. Takui, Polyhedron 24 (2005) 2112.
[26] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro,
M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M.
Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth,
P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman,
J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision A.02, Gaussian, Inc.,
Wallingford, CT, 2009.
4. Conclusions
We have described the multistep synthesis of a new ditopic tri-
arylamino-type ligand and the full characterization of bimetallic
MII(hfac)2 complexes (MII = Mn, Ni, Cu) containing this ligand. In
the neutral precursor complexes, very weak magnetic interactions
persist between the coordinated metal ions; including a weak fer-
romagnetic interaction mediated by spin polarization when
M = Cu. Chemical oxidation of the precursor complexes produces
persistent complexes of the dication of 7, which we could identify
by UV–Vis and EPR spectroscopies. Unfortunately, we could not
isolate complexes containing the radical cation of 7, perhaps owing
to the irreversible nature and poor separation of the first and sec-
ond oxidation waves observed in the ligand. In fact, DFT calcula-
tions on a bimetallic Mn(hfac)2 complex containing the dication
of 7 indicate a lower energy than the analogous radical-cation
complex. Currently, we are exploring the synthesis of related di-
and polytopic ligands but with structural features that enable
reversible oxidation to radical cation products.
[27] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
[28] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B37 (1988) 785.
[29] A. Schafer, C. Huber, R. Ahlrichs, J. Chem. Phys. 100 (1994) 5829.
[30] A.E. Reed, L.A. Curtiss, F. Weinhold, Chem. Rev. 88 (1988) 899.
[31] I. Mayer, Int. J. Quantum Chem. 29 (1986) 73.
[32] S.I. Gorelsky, L. Basumallick, J. Vura-Weis, R. Sarangi, B. Hedman, K.O. Hodgson,
K. Fujisawa, E.I. Solomon, Inorg. Chem. 44 (2005) 4947.
[33] S.I. Gorelsky, AOMix Version 6.6, University of Ottawa, Ottawa, Canada, 2012.
[34] S.I. Gorelsky, A.B.P. Lever, J. Organomet. Chem. 635 (2001) 187.
[35] X. Wang, P. Rabbat, P. O’Shea, R. Tillyer, E.J.J. Grabowski, P.J. Reider,
Tetrahedron Lett. 41 (2000) 4335.
Acknowledgements
[36] A. Landa, A. Minkkilä, G. Blay, K.A. Jørgensen, Chem. Eur. J. 12 (2006) 3472.
[37] T. Furuyama, M. Yonehara, S. Arimoto, M. Kobayashi, Y. Matsumoto, M.
Uchiyama, Chem. Eur. J. 14 (2008) 10348.
[38] A.S. Guram, R.A. Rennels, S.L. Buchwald, Angew. Chem., Int. Ed. 34 (1995) 1348.
[39] J. Louie, J.F. Hartwig, Tetrahedron Lett. 36 (1995) 3609.
[40] L.K. Thompson, Z. Xu, Magmun Version 4.1 ed., Memorial University of
Newfoundland, St. John’s, Canada, 20XX.
M.T.L. acknowledges NSERC (Discovery & RTI) and CFI (Leaders
Opportunity Fund) for financial support to carry out this research.
Mr. Samuel Mula is acknowledged for help with EPR experiments.
Appendix A. Supplementary data
[41] H.C. Longuet-Higgins, J. Chem. Phys. 18 (1950) 265.
[42] T. Glaser, M. Heidemeier, S. Grimme, E. Bill, Inorg. Chem. 43 (2004) 5192.
[43] S. Amthor, B. Noller, C. Lambert, Chem. Phys. 316 (2005) 141.
[44] R.J. Bushby, N. Taylor, R.A. Williams, J. Mater. Chem. 17 (2007) 955.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.poly.2012.06.059.