Persistent metal bis(hexafluoroacetylacetonato) complexes featuring a 2,2′-bipyridine substituted triarylamminium radical cation

Article abstract of DOI:10.1021/ic101471u

Herein, we describe the preparation of a 2,2′-bipyridine derivative containing a redox-active N,N′-(4,4′-dimethoxydiphenylamino) substituent (1), which readily coordinates M(hfac)₂ salts [M = Mn (2), Ni (3), Cu (4)] to generate stable, neutral, and pseudo-octahedral coordination complexes, which have been fully characterized. Cyclic voltammetry and spectroelectrochemical measurements on complexes 2-4 indicate stable one-electron oxidation processes, and the formation of persistent radical cation complexes. The neutral complexes (M = Mn or Ni) were subject to one-electron oxidation with NOPF₆ in acetonitrile, and magnetic moments of the resulting solutions were obtained using the Evans method at different temperatures. Our experimental results suggest that the first reported ferromagnetically coupled metal-triarylamminum radical cation complex is obtained when M = Mn²⁺, and antiferromagnetic coupling results when M = Ni²⁺. These results are supported by results from density functional theory calculations, which indicate that a π spin polarization mechanism for magnetic exchange coupling is operative in singly oxidized complexes, 2-4.

Full text of DOI:10.1021/ic101471u

Inorg. Chem. 2010, 49, 10183–10190 10183
DOI: 10.1021/ic101471u
Persistent Metal Bis(Hexafluoroacetylacetonato) Complexes Featuring
a 2,20-Bipyridine Substituted Triarylamminium Radical Cation
Sharew Adugna,† Kseniya Revunova,† Brandon Djukic,† Serge I. Gorelsky,‡ Hilary A. Jenkins,§ and
Martin T. Lemaire*,†
†Department of Chemistry, Brock University, St.Catharines, Ontario L2S 3A1, ‡Center for Catalysis Research and
§
Innovation, Department of Chemistry, University of Ottawa, Ottawa, Ontario K1N 6N5, and Department of
Chemistry, McMaster University, Hamilton, Ontario L8S 4M1
Received July 22, 2010
Herein, we describe the preparation of a 2,20-bipyridine derivative containing a redox-active N,N0-(4,40-dimethoxydiphenyl-
amino) substituent (1), which readily coordinates M(hfac)2 salts [M = Mn (2), Ni (3), Cu (4)] to generate stable, neutral, and
pseudo-octahedral coordination complexes, which have been fully characterized. Cyclic voltammetry and spectroelectro-
chemical measurements on complexes 2-4 indicate stable one-electron oxidation processes, and the formation of persistent
radical cation complexes. The neutral complexes (M = Mn or Ni) were subject to one-electron oxidation with NOPF6 in
acetonitrile, and magnetic moments of the resulting solutions were obtained using the Evans method at different temperatures.
Our experimental results suggest that the first reported ferromagnetically coupled metal-triarylamminum radical cation
complex is obtained when M = Mn2þ, and antiferromagnetic coupling results when M = Ni2þ. These results are supported by
results from density functional theory calculations, which indicate that a π spin polarization mechanism for magnetic exchange
coupling is operative in singly oxidized complexes, 2-4.
Introduction
difficulty in stable radical ligand synthesis and stability of the
resulting complexes. As a result, what remains virtually un-
explored is the use of radical ligands to create polynuclear com-
plexes exhibiting strong intracluster exchange coupling, which
seems like a promising route toward high-spin complexes
featuring ground spin states that are well isolated in energy
from other higher lying spin excited states.9 From a synthetic
perspective, some of the known families of stable radicals are
not easily amenable to the construction of larger polytopic
architectures that would be required in this regard. However,
there has been some progress in the creation of high-spin
molecules using radical ligands as linkers, with, for example,
charged radical species, such as semiquinones, and Shultz et al.
have reported a number of interesting polynuclear complexes
of tris(semiquinonyl) ligands.10 In this manner, we are inter-
ested in the exploration of the coordination chemistry of
substituted triarylamminium radical cations. Studies focused
on the preparation and electronic properties of these nitrogen-
centered radicals have been reported by Blackstock and co-
workers, Lahti and co-workers, and others, including a number
of recent reports which featured the isolation of high-spin poly-
triarylamminium radicals and other uses for triarylamminum
The metal-radical approach is a very well-established pro-
tocol geared toward the preparation of magnetic materials
featuring strong exchange coupling.1 The advantage provided
by radical ligands is the generation of typically stronger intra-
molecular magnetic exchange coupling than is obtained via
other superexchange mechanisms, whereby magnetic coupling
between metal ions is mediated by diamagnetic bridging
ligands. A large number of metal complexes have been reported
with a variety of coordinated stable radicals, including nitr-
oxides, verdazyls, thiazyls, and thioaminyls as well as other
charged radical species like semiquinones, iminosemiquinones,
and TCNE derivatives.2-8 The majority of reported metal-
radical complexes are monometallic, partly owing to the
*To whom correspondence should be addressed. E-mail: mlemaire@
brocku.ca.
(1) (a) Caneschi, A.; Gatteschi, D.; Sessoli, R.; Rey, P. Acc. Chem. Res.
1989, 22, 392–398. (b) Lemaire, M. T. Pure Appl. Chem. 2004, 76, 277–293.
(2) (a) Luneau, D.; Rey, P. Coord. Chem. Rev. 2005, 249, 2591–2611. (b)
Caneschi, A; Gatteschi, D; Rey, P. Prog. Inorg. Chem. 1991, 39, 331–429. (c)
Kaizaki, S. Coord. Chem. Rev. 2006, 250, 1804–1818.
(3) Koivisto, B. D.; Hicks, R. G. Coord. Chem. Rev. 2005, 249, 2612–2630.
(4) Preuss, K. E. Dalton Trans. 2007, 23, 2357–2369.
(5) Miura, Y.; Kato, I.; Teki, Y. Dalton Trans. 2006, 7, 961–966.
(6) Shultz, D. A. Comment Inorg. Chem. 2002, 23, 1–21.
(9) (a) Vostrikova, K. E.; Luneau, D.; Wernsdorfer, W.; Rey, P.;
Verdaguer, M. J. Am. Chem. Soc. 2000, 122, 718–719. (b) Caneschi, A.;
Gatteschi, D.; Laugier, J.; Rey, P.; Sessoli, R.; Zanchini, C. J. Am. Chem. Soc.
1988, 110, 2795–2799.
(10) Caneschi, A.; Dei, A.; Mussari, C. P.; Shultz, D. A.; Sorace, L.;
Vostrikova, K. E. Inorg. Chem. 2002, 41, 1086–1092.
(7) (a) Mukherjee, S.; Rentschler, E.; Weyhermueller, T.; Wieghardt, K.;
Chaudhuri, P. Chem. Commun. 2003, 15, 1828–1829. (b) Dei, A.; Gatteschi, D.;
Sangregorio, C.; Sorace, L.; Vaz, M. G. F. Inorg. Chem. 2003, 42, 1701–1706.
(8) (a) Miller, J. S. Dalton Trans. 2006, 23, 2742–2749. (b) Miller, J. S.;
Epstein, A. J. Chem. Commun. 1998, 13, 1319–1325.
r
2010 American Chemical Society
Published on Web 10/05/2010
pubs.acs.org/IC

10184 Inorganic Chemistry, Vol. 49, No. 21, 2010
Adugna et al.
are investigating complexes in which the triarylamminium spin
is close to the metal center in an effort to enhance the metal-
radical magnetic coupling. In this report, we describe the syn-
thesis and characterization of 1 and a series of neutral M(hfac)2
complexes (M2þ = Mn, Ni, Cu) containing 1, including the
electrochemical or chemical oxidation of these complexes to
afford triarylamminium radical cation complexes. The radical
cation complexes are characterized in part using spectroelec-
trochemical methods, and the magnetic properties of these
complexes (M2þ = Mn, Ni) are investigated in solution using
the Evans method. We propose a π spin polarization mecha-
nism to describe the observed intramolecular metal-radical
magnetic exchange coupling, which is corroborated by density
functional theory (DFT) calculations of the spin distributions
in each of the radical cation complexes.
Figure 1. Substituted triarylamminium radical cation as a SCU toward
high-spin polynuclear complexes.
Figure 2. Left:4-pyridyl-substitutedtriarylamminiumradicalcation(L)
used to prepare M(hfac)2L2 complexes (M = Mn2þ and Cu2þ). Right:
4-bis(2-picolyl)aminomethyl-substituted triarylamminium radical cation
used to prepare a Cu2þ complex.
radicals in materials science.11,12 Our approach is to use these
radical cations as “spin coupling units” (SCU) and create high-
spin polynuclear complexes making use of established
π-topology spin coupling rules in our design (Figure 1).13
However, the coordination chemistry of triarylamminium
radicals is nearly unexplored territory; reported complexes
featuring triarylamminium radical cations are few and, to our
knowledge, include only two published accounts. Bushby et al.
reported the generation of persistent M(hfac)2 (M = Mn, Cu)
complexes containing monodentate 4-pyridyl-substituted triary-
lamminium radical cations as models to explore the utility of
similar radicals toward 1-D coordination polymers (Figure 2).14
In all cases, weak antiferromagnetic metal-radical exchange
coupling was observed, owing in part to the large distance
separating the triarylamminium spin from the metal center. The
only other report was from Yano and included a Cu2þ complex
containing a 4-bis(2-picolyl)aminomethyl-substituted triary-
lamminium radical in which EPR measurements indicated a
very weak Cu2þ-radical magnetic interaction of an unknown
nature (Figure 2).15
Experimental Section
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-sensi-
tive reactions were carried out using standard Schlenk techni-
ques. 1H/13C NMR spectra were recorded on a Bruker Advance
300 MHz spectrometer with a 7.05 T Ultrashield magnet using
deuterated solvents. FT-IR spectra were recorded on a Shimadzu
IRAffinity spectrometer as thin films deposited by evaporation
of CH2Cl2 solutions on KBr plates. EI/FAB mass spectra were
obtained using a Kratos Concept 1S High Resolution E/B
mass spectrometer. UV-vis measurements were recorded on a
Shimadzu 3600 UV-vis-NIR spectrophotometer in CH2Cl2
solution using quartz cuvette cells. Elemental analyses were
carried out by Canadian Microanalytical Services, LTD, Surrey,
British Columbia, Canada, or Guelph Chemical Laboratories LTD,
Guelph, Ontario, Canada.
Electrochemical Measurements. Cyclic voltammetry (CV)
experiments were performed with a Bioanalytical Systems, Inc.
Epsilon electrochemical workstation. Compounds 1-4 were dis-
solved in anhydrous solvent (CH3CN), filtered, and then deareated
by sparging with N2 gas for 20 min. Analyte concentrations were
approximately 10-3 M containing 0.1 M supporting electrolyte
(Bu4NPF6). A typical three-electrode setup was used, including a
glassy carbon working electrode, Ag/AgCl reference electrode, and
a platinum wire counter electrode. Ferrocene was used in all cases as
an internal standard to calibrate the reference electrode and was
oxidized at a potential of þ0.51 V in our setup. The scan rate for all
CV experiments was 100 mV/s or as otherwise stated in the text.
Spectroelectrochemical Measurements. Spectroelectrochem-
ical experiments were performed in CH2Cl2 solution (10-5 M in
2-4) containing approximately 0.1 M supporting electrolyte
(Bu4NPF6) using a Bioanalytical Systems, Inc. Epsilon electrochem-
ical workstation. A special thin layer quartz glass cuvette containing
the above solution in addition to a small platinum gauze working
electrode, platinum wire counter electrode, and silver wire
reference electrode was loaded into a Shimadzu UV-3600, and
a constant potential of 1.0 V (versus ferrocene) was applied.
UV-visible spectra were recorded in repeat mode (250-1000 nm),
Our group is interested in the exploration of the coordina-
tion chemistry of substituted triarylamminium radical cations
as a new class of redox-active ligands (RAL). Specifically, we
(11) (a) Blackstock, S. C.; Selby, T. D. Magnetic Properties of Organic
Materials; Lahti, P. M., Eds.; Marcel Dekker, Inc.: New York, 1999; pp 165-178.
(b) Murata, H.; Lahti, P. M. J. Org. Chem. 2007, 72, 4974–4977.
(12) (a) Sakurai, H.; Izuoka, A.; Sugawara, T. J. Am. Chem. Soc. 2000,
122, 9723–9734. (b) Murata, H.; Takahashi, M.; Namba, K.; Takahashi, N.;
Nishide, H. J. Org. Chem. 2004, 69, 631–638. (c) Fukuzaki, E.; Nishide, H.
J. Am. Chem. Soc. 2006, 128, 996–1001. (d) Kurata, T.; Koshika, K.; Kato, F.;
Kido, J.; Nishide, H. Polyhedron 2007, 26, 1776–1780. (e) Kurata, T.; Koshika,
K.; Kato, F.; Kido, J.; Nishide, H. Chem. Commun. 2007, 2986–2988.
(13) (a) Rajca, A. Chem. Rev. 1994, 94, 871–893. (b) Iwamura, H.; Koga, N.
Acc. Chem. Res. 1993, 26, 346–351. (c) Iwamura, H. Adv. Phys. Org. Chem.
1990, 26, 179–253.
(14) Bushby, R. J.; Kilner, C.; Taylor, N.; Williams, R. A. Polyhedron
2008, 27, 383–392.
(15) Yano, M.; Inoue, K.; Motoyama, T.; Azuma, Y.; Tatsumi, M.;
Yamauchi, O.; Oyama, M.; Sato, K.; Takui, T. Polyhedron 2005, 24, 2112–2115.
(16) (a) Sur, S. K. J. Magn. Reson. 1989, 82, 169–173. (b) Evans, D. F.
J. Chem. Soc. 1959, 2003–2005. (c) Chambers, J.; Eaves, B.; Parker, D.; Claxton,
R.; Ray, P. S.; Slattery, S. J. Inorg. Chim. Acta 2006, 359, 2400–2406.

Article
Inorganic Chemistry, Vol. 49, No. 21, 2010 10185
Table 1. Crystallographic Data for 4
dissolved in exactly 1.0 mL of acetonitrile was added, resulting in a
color change to dark green. The solution was stirred for approxi-
mately 10 min, and then a small portion was withdrawn by syringe
under N2 and loaded into a special narrow coaxial NMR tube
assembly. This narrow tube was inserted into another NMR tube
containing an approximately 1:1 mixture of CH3CN and CD3CN,
and H NMR spectra were immediately recorded at a series of
different temperatures. Mass susceptibility (χg) was calculated
from eq 1:
formula
fw
C34H23CuF12N3O6
861.09
dimensions (mm)
˚
0.266 ꢀ 0.135 ꢀ 0.095
a (A)
10.2632(16)
30.851(5)
˚
b (A)
1
˚
c (A)
11.4269(18)
monoclinic
90
104.555(3)
90
3502.0(10)
Pc
cryst syst
R (deg)
β (deg)
γ (deg)
volume (A )
space group
3Δν
3
χg
¼
ð1Þ
˚
4πνc
Z
4
0.738
where Δν is the difference in absorption frequency between
CH3CN in the outer and inner tubes, ν is the spectrometer
frequency (600.32 MHz), and c is the concentration (g/mL) of
the paramagnetic complex. Molar susceptibility was calculated in
the typical manner, and corrections for the diamagnetism of the
complex and solvent were calculated using Pascal’s constants.17
The temperature-dependent density changes of the acetonitrile
solvent were accounted for using published values, and the re-
quired corrections were made to the paramagnetic solution con-
centrations. EPR spectra were recorded as toluene solutions at
room temperature in quartz tubes on a BrukerElexsys E580 pulsed
and CW X-band (9 GHz) spectrometer.
μ (mm-1
T (K)
)
100(2)
10 497 (Rint = 0.0830)
964
0.0848
0.2239
independent reflns
number of params
R1 [F2 > 2σ(F2)]
wR2 (F2)a
a w = 1/[σ2(Fo2) þ (0.069P)2 þ 0.2902P] where P = (Fo þ 2Fc2)/3.
2
until subsequent runs no longer resulted in any significant
spectral changes.
X-Ray Structure Determination. Crystals of suitable size were
mounted on a glass fiber. Data were collected at room temperature
(100 K) on a SMART APEX II diffractometer with Mo KR
Computational Details
˚
radiation (λ = 0.71073 A) located at the McMaster Analytical
All DFT calculations were performed using the Gaussian
03packageusingtheB3LYPhybridfunctionalandtheTZVP
basis set for all atoms.18 Tight SCF convergence criteria were
used for all calculations. The converged wave functions were
tested to confirm that they correspond to the ground-state
surface. The evaluation of atomic charges and spin densities
was performed using the natural population analysis
(NPA).19 The analysis of molecular orbitals in terms of
fragment orbital contributions were carried out using the
AOMix program.20 Time-dependent DFT (TD-DFT) calcu-
lationsattheB3LYP/TZVPlevelwereperformedtocalculate
the absorption spectra as previously described.21
X-ray Diffraction Facility (MAX); see Table 1 for crystallographic
data for 4. Data were processed using APEX v2.2.0 and solved by
direct methods (SHELXS-97). Data were refined using least-
squares techniques on F2. The data were solved in the monoclinic
space group P2 and checked by Platon AddSymm for additional
symmetry. All non-hydrogen atoms were refined anisotropically
except for those that were disordered (see below). Hydrogen atoms
were generated and treated as riding on their constituent atoms with
updates after each cycle of refinement. A 2-fold twin law was
applied, with a resultant BASF of 0.13(2). The disordered groups
for molecule 1 (Cu1) include CF3 (F1, F2, and F3), rotationally
disordered over two positions in a 75:25 ratio, and the F atoms were
restrained with EADP. The other disordered portion of molecule 1
is the anisole ring starting with C18, with a ring shift and methyl
ether disorder of 80:20. These atoms were also restrained with
EADP and refined isotropically. Similarly, for molecule 2, there are
two CF3 groups with rotational disorder: F13, F14, and F15 at a
ratio of 56:44 and F22, F23, and F24, at a 41:59 ratio. One of the
anisole rings in this molecule was also disordered similar to that in
molecule 1, at 45:55. All disordered atoms in molecule 2 were also
refined isotropically. In the final stages of refinement, two peaks of
residual electron density were located. The first was associated with
O4 of the nondisordered ring of molecule 1, and the second with
C113 of the nondisordered anisole of molecule 2. Attempts to refine
these as atoms did not result in reasonable results, and it is assumed
that these peaks are possibly “shadows” from the heavy atoms in a
second crystal.
Synthesis
Preparation of 5-(4,40-dimethoxydiphenylamino)-2,20-
bipyridine (1). 5-Bromo-2,20-bipyridine22 (0.104 g, 0.442
mmol), N,N0-(4,40-dimethoxydiphenyl)amine (0.125 g,
0.546 mmol), Cs2CO3 (0.202 g, 0.620 mmol), palladium-
(II) acetate (0.0055 g, 0.0245 mmol), and BINAP (0.015 g,
0.024 mmol) were combined under N2 into a dry Schlenk
flask and slurried into dry toluene (20 mL). The solution
was purged with N2 for 20 min and then heated to 95 °C.
Total consumption of starting materials was noted after
35 h, at which time the reaction was cooled to room
temperature. The cooled reaction contents were filtered,
and the precipitate was washed with CHCl3. The filtrate
and washing were rotovaped down to a brown oil, which
Variable-Temperature Magnetic Susceptibility Measurements
and EPR Spectroscopy. Solid state variable-temperature magnetic
susceptibility measurements were recorded on a superconducting
quantum interference device (SQUID) magnetometer (Quantum
Design MPMS) with a 5.5 T magnet (temperature range 1.8-
400 K) in an external field of 0.5 T. Samples of 2-4 were carefully
weighed into gelatin capsules, which were loaded into plastic
straws, and attached to the sample transport rod. The magnetiza-
tion of the samples was scanned over a temperature range of 5 to
325 K and then back to 5 K (for 3 and 4), or from 5 to 325 K for 2.
Diamagnetic corrections to the paramagnetic susceptibilities were
accomplished using Pascal’s constants. Solution magnetic mo-
ments were obtained using the Evans method.15 In a typical
experiment, 29.1 mg of complex 2 was dissolved in exactly
4.0 mL of dry acetonitrile. Under N2, one equivalent of NOPF6
(17) Hatfield, W. E. Theory and Application of Molecular Paramagnetism;
Boudreaux, E. A., Mulay, L. N., Eds.; Wiley and Sons: New York, 1976; pp
491-495.
(18) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829–
5835.
(19) (a) Godbout, N.; Salahub, D. R.; Andzelm, J.; Wimmer, E. Can. J.
Chem. 1992, 70, 560–571. (b) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem.
Rev. 1988, 88, 899–926.
(20) Gorelsky, S. I. AOMix, version 6.46; University of Ottawa: Ottawa,
Canada, 2010.
(21) Gorelsky, S. I.; Lever, A. B. P. J. Organomet. Chem. 2001, 635, 187–
196.
(22) Fang, Y.-Q.; Hanan, G. S. Synlett 2003, 6, 852–854.

10186 Inorganic Chemistry, Vol. 49, No. 21, 2010
Adugna et al.
was chromatographed over neutral Al2O3 (1200 ꢀ 100),
eluting with 9:1 hex/EtOAc, followed by 7:3 hex/EtOAc
(Rf = 0.6) to collect the bright yellow zone. Pure fractions
were combined and concentrated to a white solid, yield:
130 mg (77%). 1H NMR (CDCl3): δ 8.63 (d, 1H, J = 4 Hz),
8.31 (d, 1H, J = 3 Hz), 8.28 (d, 1H, J = 8 Hz), 8.19 (d, 1H,
J = 9 Hz), 7.76 (dd, 1H, J = 8, 2 Hz), 7.32 (dd, 1H, J =
9, 3 Hz), 7.23 (dd, 1H, J = 8, 2 Hz), 7.12 (d, 4H, J = 9 Hz),
6.89 (d, 4H, J = 9 Hz), 3.83 ppm (s, 6H). 13C NMR
(CDCl3): δ 156.6, 156.0, 149.0, 147.4, 145.2, 140.9, 139.6,
136.7, 126.8, 126.3, 122.5, 121.0, 120.1, 115.0, 55.5 ppm.
Anal. Calcd for C24H21N3O2 (Found): C, 75.16 (74.82); H,
5.52 (5.61); N, 10.96% (10.89%). MS (EI þ): m/z 383 (Mþ,
100%), 368 (M - CH3, 30%). FT-IR (film, KBr): 3045 (w),
3001 (w), 2930 (w), 2833 (w), 1568 (m), 1541 (s), 1458 (s),
1435 (w), 1327 (m), 1242 (s), 1165 (w), 1033 (m), 827 (m),
739 (m), 638 (w), 581 cm-1 (w). UV-vis (CH2Cl2), λmax (ε):
352 nm (2.1 ꢀ 104 cm-1 M-1).
1508 (s), 1471 (s), 1442 (m), 1348 (w), 1254 (s), 1200 (s),
1144 (s), 1082 (m), 1031 (w), 833 (w), 791 (m), 665 (m),
586 cm-1 (w). UV-vis (CH2Cl2), λmax (ε): 394 nm (1.6 ꢀ
104 cm-1 M-1).
Results and Discussion
Ligand Synthesis, Coordination Chemistry, and Struc-
tural Studies. In an effort to explore the basic coordina-
tion chemistry of triarylamminium radical cations, we
have designed ligand 1, a simple 2,20-bipyridine derivative
featuring a N,N0-(4,40-dimethoxydiphenyl)amine substi-
tuent attached at the 5 position of the bipyridine frame-
work. Ligand 1 is easily prepared from 5-bromo-2,20-
bipyridine via palladium-catalyzed Buchwald-Hartwig
cross-coupling in toluene solution and has been fully
characterized (Scheme 1). A series of neutral coordina-
tion complexes containing 1 were prepared by combina-
tion of M(hfac)2 xH2O salts [M = Mn (2), Ni (3), Cu (4)]
3
Mn(hfac)2(1) (2). Mn(hfac)2 3H2O (0.22 g, 0.42 mmol)
3
with 1 in a hexanes/CH2Cl2 solution (Scheme 1). All of the
complexes are analytically pure microcrystals and feature
very similar FT-IR absorption spectra (Figures S4, S5,
and S6 in the Supporting Information). Slow evaporation
of methanol solutions containing 4 produced single crys-
tals suitable for X-ray diffraction studies. An ORTEP
diagram of the molecular structure of 4 is shown in Figure 3.
This structure features the anticipated Jahn-Teller
distorted, cis pseudo-octahedral coordination geometry
about Cu2þ, with 1 binding as a bidentate ligand through
the bipyridine ring nitrogen atoms. Coordinate bond
lengths and angles are listed in the Figure 2 caption and
are in the anticipated range, with two longer bonds and four
shorter. Other bond distances and angles within the hfac
ancillary ligands are unremarkable and are provided in the
Supporting Information. One structural feature of note is
the near planarity of the neutral three-coordinate nitrogen
atom (N3), featuring a bond angle sum of 359.8°. The bond
was dissolved in a solvent mixture containing hexanes
(5 mL) and MeOH (2 mL) and combined with a CH2Cl2
solution (5 mL) of 1 (0.15 g, 0.39 mmol). The solution was
stirred at room temperature for 1 h to produce a clear,
orange solution. The solution was filtered and left stand-
ing at room temperature. Over a period of days, orange,
rod-shaped crystals were observed, which were isolated
by vacuum filtration, yield: 0.14 g, (41%). Anal. Calcd for
C34H23N3O6F12Mn (Found): C, 47.88 (48.05); H, 2.72
(2.43); N, 4.93 (4.98). MS (FAB þ): 852 (Mþ, 2), 645 (M -
hfac, 50), 383 (1, 100%). FT-IR (film, KBr): 1651 (m),
1589 (w), 1557 (m), 1508 (s), 1470 (s), 1442 (m), 1344 (m),
1252 (s), 1196 (s), 1142 (s), 1082 (w), 1032 (w), 833 (w), 795
(m), 662 (m), 583 cm-1 (w). UV-vis (CH2Cl2), λmax (ε):
385 nm (1.8 ꢀ 104 cm-1 M-1).
Ni(hfac)2(1) (3). Ni(hfac)2 2H2O (0.13 g, 0.26 mmol)
3
was slurried in 7 mL of hexanes and combined with a
CH2Cl2 solution (4 mL) of 1 (0.10 g, 0.26 mmol). The
solution was stirred at room temperature overnight to
produce a clear, golden solution. The solution was con-
centrated to a crusty golden residue, which was dissolved
into hot hexanes (12 mL). Upon cooling to room tem-
perature, a green microcrystalline powder precipitated,
which was isolated by vacuum filtration, yield: 0.14 g
(61%). Anal. Calcd for C34H23N3O6F12Ni (Found): C,
47.67 (47.86); H, 2.71 (2.78); N: 4.91 (4.95). MS (EI þ):
648 (M-hfac, 10%), 383 (1, 100%). FT-IR (film, KBr):
1643 (m), 1590(w), 1556 (m), 1506 (s), 1471 (s), 1441 (m),
1346 (m), 1256 (s), 1202 (s), 1148 (s), 1097 (m), 1034 (m),
833 (w), 791 (m), 671 (m), 586 cm-1 (m). UV-vis (CH2Cl2),
˚
to the bipyridine substituent, which is less than 1.4 A, is the
shortest of the three C-N bonds to N3. These features
suggest delocalization with the 2,20-bipyridine ring, which
bodes well for electronic communication between the radi-
cal cation and the coordinated metal ion in the one-electron
oxidized complex. Solutions of the radical cation complexes
of 2-4 can be generated by one electron oxidation of the
neutral precursors with NOPF6 as an oxidant. The stability
of these radical cation solutions is best described as persis-
tent under a N2 atmosphere, and the solutions gradually
decompose over the course of an hour, as indicated by
UV-visible spectroscopy. The stability of these solutions in
air is shorter, with decomposition noted over the course of
minutes.
Electronic Properties. The electronic properties of 1-4
have been investigated using a combination of UV-
visible spectroscopy and cyclic voltammetry (CV). The
UV-visible spectra of 1-4 are shown in Figure 4. The
spectrum of 1 is similar to other reported triarylamines
featuring a long wavelength nfπ* absorption at 350 nm
along with other higher energy transitions.23 The spectra
of complexes 2-4 are similar to one another with the long
λ
max (ε): 383 nm (1.8 ꢀ 104 cm-1 M-1).
Cu(hfac)2(1) (4). Cu(hfac)2 xH2O (0.068 g, 0.14 mmol)
3
was slurried in 7 mL of 2:1 hexanes/CH2Cl2 and combined
with a CH2Cl2 solution(2mL) of1(0.055 g, 0.14 mmol). The
solution was stirred at room temperature for 3 h to produce a
clear red-brown solution. The solution was concentrated to
dryness, producing a lime colored residue, which was dis-
solved into a minimum of hot methanol. The solution was
left open to the atmosphere, and slow evaporation provided
brown rod crystals, which were isolated by vacuum filtra-
tion, yield: 0.060 g (51%). Anal. Calcd for C34H23N3O6F12-
Cu (Found): C, 47.40 (47.30); H, 2.69 (2.74); N, 4.88 (4.87).
MS (FAB þ): 653 (M - hfac, 100), 446 (M - 2hfac, 30), 383
(1, 55%). FT-IR (film, KBr): 1653 (m), 1589(w), 1548 (m),
(23) (a) Amthor, S.; Noller, B.; Lambert, C. Chem. Phys. 2005, 316, 141–
152. (b) Bushby, R. J.; Taylor, N.; Williams, R. A. J. Mater. Chem. 2007, 17,
955–964. (c) Bushby, R. J.; Kilner, C. A.; Taylor, N.; Vale, M. E. Tetrahedron
2007, 63, 11458–11466.

Article
Inorganic Chemistry, Vol. 49, No. 21, 2010 10187
Scheme 1. Preparation of 1-4a
a Reagents and conditions: (a) 5 mol % Pd(PPh3)4, RT, 36 h. (b) 5.5 mol % Pd(OAc)2, BINAP, Cs2CO3, toluene, 95 °C, 36 h. (c) M(hfac)2 xH2O (M =
Mn, 2; Ni, 3; Cu, 4), hexanes/CH2Cl2, RT.
3
respectively. Increasing the scan rate from the original
rate of 100 mV/s did not improve the reversibility of either
anodic wave, which is attributed to formation of the
radical cation, followed by dication formation at a higher
potential. A minor feature is noted at approximately
þ0.7 V in the subsequent reduction cycle, which is likely
attributable to a decomposition product resulting from
the oxidation of 1. The cathodic scan revealed no facile
reductive processes for this ligand, as expected. These
observations are in line with electrochemical data re-
ported for other triarylamines. Anodic scans for com-
plexes 2-4 are similar to one another. Of note, the
formation of a radical cation (E0ox1) is a reversible process
in all cases, indicating that the radical cations are stabi-
lized by coordination to the metal ion. The potential for
radical cation formation is a little higher in the coordina-
Figure 3. ORTEP diagram of the molecular structure of 4. H atoms are
omitted for clarity. One of the two crystallographically independent
tion complexes than was observed in 1. In contrast to 2
and 3, the anodic waves in 4 are somewhat broader, and at
˚
molecules per unit cell is shown. Selected bond distances (A) and
scan rates slower0 than 1000 mV/s, a new irreversible
angles (deg): Cu(1)-O(2), 2.381(7); Cu(1)-O(1), 1.988(6); Cu(1)-O(3),
feature between E ox1 and E ox2 is clearly noted, which
disappears at faster scan rates (>200 mV/s). We tenta-
tively attribute this irreversible feature to an oxidation
side product that decomposes prior to detection at faster
scan rates. Dication formation is observed at potentials
slightly greater than þ1.0 V for each complex. Over
cathodic potentials greater than -1.0 V, a series of
irreversible processes are observed, which can be attrib-
uted to reductions involving the hfac and 2,20-bipyridine
ligands. The electrochemical properties of precursor
M(hfac)2 complexes have been investigated by Villamena
et al. and are well understood.24
1.978(6); Cu(1)-O(4), 2.294(7); Cu(1)-N(1), 2.008(7); Cu(1)-(N2),
2.007(7); O(1)-Cu(1)-N(1), 167.3(3); O(3)-Cu(1)-N(2), 164.7(3);
O(2)-Cu(1)-O(4), 155.8(2); O(1)-Cu(1)-N(2), 92.3(3); O(3)-Cu(1)-
N(1), 92.8(3); O(1)-Cu(1)-O(3), 96.1(3); N(1)-Cu(1)-N(2), 81.3(3).
Spectroelectrochemical Measurements. Spectroelectro-
chemical spectra (Figure 6; see Scheme 2 for the electro-
chemical generation scheme of complexes 2-4) were
recorded for complexes 2-4 by applying a potential of
approximately 1 V (versus silver wire) to solutions of the
complexes in CH2Cl2 (containing 0.1 M nBu4NPF6) while
simultaneously recording the time-dependent absorption
spectra between 250 and 1000 nm. In all cases, absorption
bands assigned to the starting neutral complexes de-
creased in absorption intensity (down arrows) over the
course of the oxidation while new bands appeared (up
arrows) which are assigned to the single electron oxidized
radical cation complexes. In particular, new bands at
wavelengths similar to 446, 615, and 820 nm have been
Figure 4. Room temperature UV-vis spectra of 1-4 in CH2Cl2.
wavelength band observed in 1 in all cases bathochromi-
cally shifted by approximately 30 nm.
The electrochemical properties of 1-4 were probed
with CV. The voltammograms are featured in Figure 5,
and important anodic data for all compounds are sum-
marized in Table 2. All potentials quoted are referenced to
the ferrocene/ferrocenium couple. The anodic scan of 1
features two quasi-reversible waves at þ0.5 and þ1.0 V,
(24) Villamena, F. A.; Horak, V.; Crist, D. R. Inorg. Chim. Acta 2003,
342, 125–130.

10188 Inorganic Chemistry, Vol. 49, No. 21, 2010
Adugna et al.
Figure 5. Cyclic voltammograms of 10-3 M solutions of 1-4 in CH3CN containing 0.1 M Bu4NPF6. In all cases, the scan rate was 100 mV/s.
Figure 6. Spectroelectrochemistry of 2 (left), 3 (center), and 4 (right) in CH2Cl2 at room temperature containing 0.1 M nBu4NPF6. A potential of þ1.0 V
was applied across a platinum gauze working electrode and silver wire pseudoreference electrode (platinum counter electrode). Black arrows represent the
increase (up pointing) or decrease (down pointing) in absorbance over the course of the oxidation.
Table 2. Selected Electrochemical Data for Compounds 1-4a
Scheme 2. Electrochemical Generation of Radical Cation Complexes
of 1 (M = Mn, 2; Ni, 3; Cu, 4)
Compound
E0ox1 (V)
E0ox2 (V)
1
2
3
4
þ0.5
þ1.0
þ1.0
þ1.1
þ1.1
þ0.6 (rev)
þ0.7 (rev)
þ0.6 (rev)
a All experiments were performed in CH3CN, containing 0.1 M
nBu4NPF6 supporting electrolyte, and all potentials are quoted versus
the ferrocene/ferrocenium reference.
observed in other reported free and coordinated triary-
lamminum radical cations. Isosbestic points are noted in
all spectra, indicating clean oxidative processess and the
presence of only two species in solution over the course of
the oxidation.
Variable-Temperature Magnetic Susceptibility and EPR
Spectroscopy. The solid-state temperature-dependent mag-
netic properties of neutral complexes 2-4 were investigated
using SQUID magnetometry. As anticipated, nearly ideal
Curie behavior is noted for all complexes with approxi-
mately temperature-independent magnetic moments that
are typical for isolated monometallic Mn, Ni, and Cu
complexes, with only very small deviations noted at the
lowest temperatures (Figure 7).

Article
The magnetic properties of the one-electron oxidized
Inorganic Chemistry, Vol. 49, No. 21, 2010 10189
complex. However, the magnetic moment of the singly
oxidized 2þ•PF6- complex is higher, approaching 6.4 μB
This value is also higher than the theoretical value anti-
cipated for noninteracting S = 5/2 Mn2þ and a coordi-
nated radical (S = 1/2), which is 6.16 μB and suggests the
presence of ferromagnetic coupling between Mn2þ and
the radical. While at low temperatures, a freshly prepared
solution of 3þ• was placed in the spectrometer, and an
identical Evans experiment was run from low to higher
temperatures. The initial magnetic moment of the sample
was 3.1 μB, which is a little lower than the anticipated
value (3.3 μB) for noninteracting S = 1 Ni2þ and a coordi-
nated radical (S = 1/2). Our data support the presence
of weak Ni2þ-radical antiferromagnetic coupling. For
comparison, Evans data were acquired for an acetonitrile
solution of neutral complex 3, and the data are very
similar to those obtained from the solid-state experiment.
The magnetic properties of oxidized 4 were not investi-
gated because the cyclic voltammetry results for this
complex indicated an unclean oxidation. Solution EPR
spectra of 4 and 4þ• recorded in toluene at 298 K are
included in the Supporting Information (Figure S7).
Spin Density Calculations. Density functional calcula-
tions at the B3LYP/TZVP level were used to calculate the
structures and spin density distributions for the ground
state singly oxidized radical cation complexes. The ground
state spin density distributions within the radical cation
complexes of 2-4 are shown in Figure 9. For complex 2þ•,
a substantial positive π spin density is noted at the ammi-
nium nitrogen atom, and the π spin distribution around each
atom of the adjacent bipyridine ring alternates negative then
positive. The 2,20-bipyridine coordinating nitrogen atom
features a significant amount of negative π spin density,
which results in positive dπ spin density at Mn2þ, which
produces an overall ferromagnetic interaction between the
radical and Mn2þ ion. On the other hand, complexes 3þ•,
and 4þ• feature metal ions with dσ spin density of the same
sign as the π spin density at the coordinate nitrogen atom,
and the net interaction with the radical in each case is anti-
ferromagnetic. These results are similar to the pioneering
studies of Iwamura et al., which elegantly demonstrated the
spin polarization mechanism behind the magnetic coupling
in a series of M(hfac)2 complexes featuring differentially
radical cations of 2 and 3 were measured in solution using
the Evans method (Figure 8). For comparison, the mag-
netic moments for neutral complexes 2 and 3 were also
measured by this technique. Data were recorded from 300
to 244 K for 2 and 2þ• and from 325 to 244 K for 3 and 3þ•
(low-temperature data collection is limited to tempera-
tures above the freezing point of acetonitrile). The data
for neutral 2 are very similar to those recorded in the solid
state, with a magnetic moment of 5.9 μB at 300 K, which is
the anticipated value for a magnetically dilute Mn2þ
Figure 7. Temperature dependence of μeff values for 2 (2), 3 ((), and
4 (9) (1000 Oe field).
Figure 8. Temperature dependence of μeff values for 2 (2), 2þ•PF6- (9),
3 (ꢀ), and 3þ•PF6- (() measured in solution by the Evans method.
Figure 9. Calculated spin density distributions for 2þ• (left), 3þ• (center), and 4þ• (right). NPA-derived atomic spin density is shown for most important
contributors. H atoms are not shown for clarity.

10190 Inorganic Chemistry, Vol. 49, No. 21, 2010
Adugna et al.
substituted pyridine or 2,20-bipyridine-containing nitroxide
radicals.25,26
of the triarylamminium radical cation on the 2,20-bipyridine
ring. For example, similar complexes to 3 and 4, but contain-
ing instead a 4- or 6-substituted 2,20-bipyridine triarylammi-
nium ligand, should result in ferromagnetic metal-radical
interactions (M = Ni or Cu). We are also investigating the
feasibility of enhancing the stability of the triarylamminium
cation complexes, through steric protection of the spin
density rich sites in an effort to isolate crystalline metal-
radical complexes to obtain solid-state structural and mag-
netic susceptibility data. These efforts are underway and will
be published in due course.
Conclusions
The results presented herein represent our initial efforts
toward exploring the coordination chemistry of triarylammi-
nium radical cations. We are actively designing ligands in
which the triarylamminium spin density is polarized toward
the metal center in an effort to engineer stronger metal-
radical exchange coupling. We have observed ferromagnetic
metal-radical coupling in oxidized complex 2, and we are
currently trying to exploit and enhance this interaction in
polymetallic complexes. Using this approach, we can engi-
neer ferromagnetic interactions with transition metals con-
taining dσ magnetic orbitals by simple synthetic adjustment
Acknowledgment. The authors acknowledge NSERC
(Discovery & RTI) and CFI (Leaders Opportunity Fund)
for financial support. We thank Razvan Simionescu
(Brock University) for help with the Evans method
experiments.
(25) (a) Kitano, M.; Ishimaru, Y.; Inoue, K.; Koga, M.; Iwamura, H.
Inorg. Chem. 1994, 33, 6012–6019. (b) Ishimaru, Y.; Kitano, M.; Kumada, H.;
Koga, N.; Iwamura, H. Inorg. Chem. 1998, 37, 2273–2280. (c) Kumada, H.;
Sakane, A.; Koga, N.; Iwamura J. Chem. Soc., Dalton Trans. 2000, 10, 911–914.
(26) Shimada, T.; Ishida, T.; Nogami, T. Polyhedron 2005, 24, 2593–2598.
Supporting Information Available: General spectroscopic
data and crystallographic data in CIF format. This material is
available free of charge via the Internet at http://pubs.acs.org.