Mixed-valent dicobalt and iron-cobalt complexes with high-spin configurations and short metal-metal bonds

Article abstract of DOI:10.1021/ic400292g

Cobalt-cobalt and iron-cobalt bonds are investigated in coordination complexes with formally mixed-valent [M₂]³⁺ cores. The trigonal dicobalt tris(diphenylformamidinate) compound, Co₂(DPhF) ₃, which was previous

Full text of DOI:10.1021/ic400292g

Article
pubs.acs.org/IC
Mixed-Valent Dicobalt and Iron−Cobalt Complexes with High-Spin
Configurations and Short Metal−Metal Bonds
Christopher M. Zall,† Laura J. Clouston,† Victor G. Young, Jr.,† Keying Ding,† Hyun Jung Kim,†,‡
Danylo Zherebetskyy,†,‡ Yu-Sheng Chen,§ Eckhard Bill,*,∥ Laura Gagliardi,*,†,‡ and Connie C. Lu*,†
†Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455-0431, United States
‡Supercomputing Institute, and Chemical Theory Center, University of Minnesota, Minneapolis, Minnesota, 55455, United States
§Advanced Photon Source, Argonne, Illinois 60439, United States
∥Max Planck Institut fur Chemische Energiekonversion, Stiftstrasse 34-36, 45470, Mulheim an der Ruhr, Germany
̈
̈
S
* Supporting Information
ABSTRACT: Cobalt−cobalt and iron−cobalt bonds are
investigated in coordination complexes with formally mixed-
valent [M2]3+ cores. The trigonal dicobalt tris-
(diphenylformamidinate) compound, Co2(DPhF)3, which
was previously reported by Cotton, Murillo, and co-workers
(Inorg. Chim. Acta 1996, 249, 9), is shown to have an
energetically isolated, high-spin sextet ground-state by
magnetic susceptibility and electron paramagnetic resonance
(EPR) spectroscopy. A new tris(amidinato)amine ligand
platform is introduced. By tethering three amidinate donors
to an apical amine, this platform offers two distinct metal-
binding sites. Using the phenyl-substituted variant (abbreviated as LPh), the isolation of a dicobalt homobimetallic and an iron−
cobalt heterobimetallic are demonstrated. The new [Co2]3+ and [FeCo]3+ cores have high-spin sextet and septet ground states,
respectively. Their solid-state structures reveal short metal−metal bond distances of 2.29 Å for Co−Co and 2.18 Å for Fe−Co;
the latter is the shortest distance for an iron−cobalt bond to date. To assign the positions of iron and cobalt atoms as well as to
determine if Fe/Co mixing is occurring, X-ray anomalous scattering experiments were performed, spanning the Fe and Co
absorption energies. These studies show only a minor amount of metal-site mixing in this complex, and that FeCoLPh is more
precisely described as (Fe0.94(1)Co0.06(1))(Co0.95(1)Fe0.05(1))LPh. The iron−cobalt heterobimetallic has been further characterized
by Mossbauer spectroscopy. Its isomer shift of 0.65 mm/s and quadrupole splitting of 0.64 mm/s are comparable to the related
̈
diiron complex, Fe2(DPhF)3. On the basis of spectroscopic data and theoretical calculations, it is proposed that the formal [M2]3+
cores are fully delocalized.
also shown short M-M bonds.18−23 A classic example is the
diiron trigonal lantern complex, Fe2(DPhF)3 (where DPhF is
1. INTRODUCTION
Metal−metal bonds are interesting because of their potential
roles in heterogeneous and bioinorganic catalysis.1−3 Given the
diphenylformamidinate), reported by Cotton, Murillo, and co-
workers.22 Fe2(DPhF)3 has an exceptionally short iron−iron
large number of possible metal pairings and electronic
interactions, there is the potential to access a wide range of
electronic properties and chemical reactivities. In particular, one
sought-after application is the cooperative coupling of metals in
bond (2.23 Å) and an unusually high-spin ground state (S = 7/
2).24 These were ascribed to the close separation of the metal−
metal bonding orbitals engendered by the trigonal ligand
field.25 We recently reexamined this species and confirmed that
multielectron catalysis. Several M-M complexes have demon-
the octet state is a direct consequence of the π/π*, δ/δ*, and
strated redox reactions of up to four electrons using second and
third-row metals.4−7 Similar reactivity with first-row metals
would be valuable, as they are more earth-abundant and
environmentally benign.
σ* metal−metal d-orbitals being close in energy. Using
multiconfigurational wave function calculations, the effective
bond order (EBO)26,27 was determined to be slightly greater
than one (EBO = 1.15), which is consistent with the short Fe−
Fe bond distance.28
Bonds between first-row metals are less well-known than
their second- and third-row analogues.8 By using ligands that
stabilize low coordination numbers, that is, 2 to 3, short bonds
between first-row metals have been obtained, as demonstrated
in the dichromium systems with formal bond orders
approaching 5.9−17 Similar low-coordinate bimetallic complexes
of mid-to-late first-row transition metals have in many cases
We have since expanded our study of first-row metal−metal
interactions to cobalt. Dicobalt complexes lacking strongly π-
Received: February 4, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 1. Selected examples of structurally characterized dicobalt complexes with short Co−Co bond lengths (Å, in red). Ground spin states are
given.
acidic ligands like CO are uncommon and can exhibit extreme
spin states as shown in Figure 1. For instance, two of the
complexes with extremely short Co−Co bonds, reported by
Jones and Mindiola, are both dicobalt(I,I) systems, but Jones’ is
high-spin (S = 2) while Mindiola’s is low-spin (S = 0).18,29
Similarly, the tetragonal and trigonal “lantern” complexes
Co2(DPhF)4 and Co2(DPhF)3, both contain short Co−Co
distances, but the former is diamagnetic, while the latter was
found to be highly paramagnetic.19,23,30 Hayton and co-workers
have reported a dicobalt(II,II) ketimide complex with a Co−Co
bond distance of 2.41 Å that has a temperature-dependent
magnetic moment (0.68 to 3.5 μB), that would indicate weak
antiferromagnetic coupling of the two metal centers.31
Recently, Thomas and co-workers isolated dicobalt(I,I) and
dicobalt(II,I) species with reasonably short bond distances of
2.55 and 2.49 Å, respectively.32 An intermediate S = 1 state was
proposed for the dicobalt(I,I) species, while the dicobalt(II,I) is
a low-spin doublet. Additionally, Betley and co-workers have
isolated tricobalt clusters, [Co3]6+ and [Co3]7+, with low-spin
ground states that arise from strong metal−metal orbital
overlap within the trigonal metal core.33 Related tricobalt
complexes in analogous oxidation states but featuring a linear
arrangement of metal centers have been shown to undergo
thermally induced spin crossover.34−36 The wide range of spin
states and magnetic behavior observed within the relatively
small sample of Co−Co bonds illustrates the complexity of
bonding between late first-row metal centers.
purpose, we have designed the tris(amidinato)amine ligands
([LR]3−) shown in Figure 2. The multidentate nature of these
Figure 2. Tris(amidinato)amine ligands ([LR]3−).
ligands controls the coordination number, and the two planes
of donor atoms could allow sequential metalations to yield both
homo- and heterobimetallic complexes.
Using the phenyl-substituted derivative ([LPh]3−), we have
been able to isolate and study the dicobalt species, Co2(LPh). By
comparing Co2(DPhF)3 (1) with Co2(LPh) (3), we show that
adding an apical amine donor does not significantly perturb the
electronic structure of the dicobalt core. We also report a rare
iron−cobalt heterobimetallic, FeCo(LPh) (4). The isolation of
this complex demonstrates the utility of this ligand architecture
for gaining access to heterobimetallics. By comparing 4 to 3, we
show that the swapping of cobalt for iron retains the high-spin,
metal−metal bonded configuration of the other trigonal lantern
species. Finally, we have used X-ray anomalous scattering
experiments to determine the degree of Fe/Co mixing in
heterobimetallic 4. The occupancies of iron and cobalt in the
two metal-binding sites have been precisely quantified and are
in accord with the structural formula, (Fe0.94Co0.06)-
(Co0.95Fe0.05)LPh, where FeCoLPh represents the majority
species.
Even further complexity may be expected when mixed-metal
systems are considered. Heterobimetallic complexes are gaining
interest as potential organometallic catalysts,37 and late/late
combinations containing bona fide metal−metal bonds are
uncommon.38 Synthetic methods that selectively combine
different metal pairs within a fixed ligand environment would
further develop magnetostructural relationships, such as the
effect of the metal identity on the overall magnetism. For this
B
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Synthesis of Tris(2-benzimidoylchloroethyl)amine Hydrochloride.
Tris(2-benzamidoethyl)amine (7.6 g, 16.6 mmol) was dissolved in
CH2Cl2 (50 mL) in a 100 mL thick-walled flask. Phosphorus
pentachloride (11.8 g, 56.7 mmol) was added, the flask was sealed
with a Teflon stopper, and the mixture was refluxed at 50 °C. After 24
h, the volatiles were removed under vacuum. The resulting white
residue was washed with toluene (100 mL) and filtered, giving a fine
2. EXPERIMENTAL SECTION
General Considerations. Unless otherwise stated, all manipu-
lations were performed under a dinitrogen atmosphere in a VAC
Atmosphere glovebox. Standard solvents were deoxygenated by
sparging with dinitrogen and dried by passing through activated
alumina columns of a SG Water solvent purification system. N,N′-
diphenyformamidine (HDPhF) was purchased from Aldrich, dried at
60 °C under vacuum, and recrystallized from diethyl ether.
1
white powder (8.4 g, 85% yield). H NMR (300 MHz, CDCl3): δ
12.66 (br, 1H), 7.93 (d, J = 7.5 Hz, 6H), 7.44 (t, J = 7.5 Hz, 3H), 7.27
(t, J = 7.8 Hz, 6H), 4.34 (t, J = 4.8 Hz, 6H), 3.86 (quart, J = 3.9 Hz,
6H).
40
Benzylpotassium39 and KC8 were prepared according to literature
methods. Deuterated solvents were purchased from Cambridge
Isotope Laboratories, Inc., degassed via freeze−pump−thaw cycles,
dried over activated alumina, and stored over activated 4 Å molecular
sieves. All other reagents were purchased from Aldrich or Strem and
used without further purification. Elemental analyses were performed
by Complete Analysis Laboratories, Inc. (Parsippany, NJ). Inductively
Coupled Plasma-Optical Emission Spectrometry (ICP-OES) data were
collected at the University of Minnesota Earth Sciences Analytical
Geochemistry Lab using a Thermo Scientific iCAP 6500 dual view
instrument, with the addition of cesium as a matrix modifier and
yttrium as an internal standard. The average of two measurements is
reported.
Modified Synthesis of Co2(DPhF)4. To a solution of HDPhF (11.0
g, 56.1 mmol) in tetrahydrofuran (THF, 100 mL) was added
benzylpotassium (7.3 g, 56.1 mmol) in THF (25 mL), forming a light
yellow solution immediately. The mixture was stirred for 6 h, and
solvent was removed under vacuum. The light yellow solid was rinsed
with pentane (3 × 5 mL) and dried under vacuum, giving a light
yellow crystalline powder (14.8 g, 85% yield). The 1H NMR spectrum
Synthesis of Tris(2-(N-phenylbenzamidinyl)ethyl)amine (H3LPh).
Aniline (5.3 g, 57 mmol) was dissolved in CH2Cl2 and cooled to −78
°C. A suspension of tris(2-benzimidoylchloroethyl)amine hydro-
chloride (5.24 g, 9.53 mmol) in CH2Cl2 was added dropwise, and
the reaction was allowed to slowly warm to rt overnight. The resulting
suspension was filtered, giving a fine, white powder, which was washed
with CH3CN (50 mL). After dissolving the powder in water, NaOH
(17 mL, 0.1 mol) was added, causing a large amount of white
precipitate to form. The precipitate was dissolved in CHCl3, washed
three times with water and once with brine, then dried with anhydrous
MgSO4. The solvent was removed under vacuum, and the resulting
yellow oil was recrystallized from diethyl ether to give a white solid.
The solids were dried overnight under vacuum at 60 °C, and finally
1
rewashed with Et2O to yield a white powder (5.5 g, 80% yield). H
NMR (500 MHz, CD3CN): δ 7.22 (t, bridgehead para-C-H, 3H), 7.12
(m, aryl,12H), 6.99 (t, bridgehead meta-C-H, 6H), 6.73 (t, apical para-
C-H, 3H), 6.52 (d, apical ortho-C-H, 6H), 5.53 (br, N-H, 3H), 3.56
(br, CH2, 6H), 2.99 (br, CH2, 6H). ESI-MS-TOF m/z: [M + H]+
calc’d for C45H46N7, 684.3815; found 684.3806. Anal. Calcd for
C45H45N7: C, 79.03; H, 6.63; N, 14.34. Found C 78.92; H 6.53; N
14.26.
1
of the solid is consistent with K(THF)[DPhF]: H NMR (C6D6): δ
8.81 (s, N−CH-N, 1H), 7.31 (t, J = 7.6 Hz, meta-CH, 4H), 6.96 (m,
ortho- and para-CH, 6H), 3.56 (THF, 4H), 1.39 (THF, 4H). To a
solution of K(THF)[DPhF] (4.4 g, 14.3 mmol) in THF (80 mL) was
added CoCl2(THF)1.5 (1.5 g, 6.3 mmol). The mixture was stirred at
room temperature (rt) for 2 h. The precipitate was filtered through
Celite (1 cm), and volatiles were removed under vacuum. The brown-
green solid was rinsed with diethyl ether (3 × 5 mL), and then dried
Synthesis of Tris(2-pivalamidoethyl)amine. This compound was
synthesized in a manner entirely analogous to that of tris(2-
benzamidoethyl)amine, above, starting from tren (3.1 mL, 20 mmol)
1
and pivaloyl chloride (7.6 mL, 62 mmol). Yield: 6.08 g (76%). H
NMR (500 MHz, CDCl3): δ 6.175 ppm (br. s, 3H, NH), 3.289 ppm
(dd, J = 6.25 Hz, J = 12.25 Hz, NH−CH2, 12H), 2.606 (t, J = 6 Hz,
NH−CH2−CH2, 12H), 1.194 (s, C(CH3)3, 27H). ESI-MS-TOF m/z:
[M + H]+ calc’d for C21H43N4O3, 399.3335; found: 399.3491.
Synthesis of Tris(2-pivalimidoylchloroethyl)amine Hydrochloride.
This compound was synthesized in a manner entirely analogous to
that of tris(2-benzimidoylchloroethyl)amine, starting from the tris-amide,
above, (5.813 g, 14.6 mmol) and phosphorus pentachloride (9.56 g, 46
1
under vacuum to give a brown powder (1.9 g, 65% yield). H NMR
(C6D6, 500 MHz): δ 8.40 (s, N−CH-N, 4H), 6.87 (t, J = 7.0 Hz, meta-
CH, 16H), 6.82 (t, J = 7.0 Hz, para-CH, 8H), 6.27 (d, J = 7.5 Hz,
ortho-CH, 16H).
Modified Synthesis of Co2(DPhF)3 (1). To a solution of
Co2(DPhF)4 (1.4 g, 1.5 mmol) in toluene (50 mL), KC8 (205 mg,
1.5 mmol) was added, resulting immediately in a black precipitate.
After 1 h, the mixture was filtered through a Celite plug, reduced to 10
mL, and cooled at −25 °C to give red-brown crystals (662 mg, 60%
1
mmol). Yield: 6.19 g (86%). H NMR (500 MHz, CDCl3): δ 12.978
(s, N-H+, 1H), 4.025 (t, J = 6 Hz, CH2−CH2−N-H+, 6H), 3.555 (dd, J
= 5.5 Hz and 4.5 Hz, CH2-CH2−N-H+, 6H), 1.227 (s, C(CH3)3,
27H).
1
yield). H NMR (C6D6, 500 MHz): δ 175.10 (N−CH-N, 3H), 13.00
(meta-CH, 12H), −26.19 (para-CH, 6H), −50.16 (ortho-CH, 12H).
UV−vis (toluene): λmax, nm (ε, M−1 cm−1): 545 (530), 752 (200),
∼1665 (100). The NIR values are only approximate because of the
artifacts created by imperfect subtraction of the solvent background.
Anal. Calcd for C39H33N6Co2: C 66.58; H 4.73; N 11.94. Found C
66.49; H 4.65; N 11.78.
Synthesis of Tris(2-(N-phenylpivalamidinyl)ethyl)amine (H3LtBu).
This compound was synthesized in a manner entirely analogous to
that of H3LPh above, starting from the tris-imidoyl chloride HCl salt,
above, (6.19 g, 12.6 mmol) and aniline (4.004 g, 4.3 mmol). Yield:
6.2g (80%). 1H NMR (500 MHz, CD3CN): δ 7.13 (t, J = 8 Hz, meta-
C-H, 6H), 6.80 (t, J = 7 Hz, para-C-H, 3H), 6.64 (d, J = 7.5 Hz, ortho-
C-H, 6H), 4.90 (br, NH, 3H), 2.59 (br, CH2, 6H), 2.08 (br, CH2, 6H),
1.16 (s, C(CH3)3, 27H). ESI-MS-TOF m/z: [M + H]+ calc’d for
C39H58N7, 624.4754; found: 624.4848.
Synthesis of Tris(2-benzamidoethyl)amine. Tris(2-aminoethyl)-
amine (tren) (3.2 mL, 21.4 mmol) and NEt3 (10.3 mL, 73.8 mmol)
were combined in a 250 mL RB flask with THF (100 mL) and cooled
to 0 °C under ambient atmosphere. Benzoyl chloride (7.8 mL, 67.1
mmol, diluted in 10 mL of THF) was added dropwise, forming a white
precipitate. After warming to rt overnight, the precipitate was filtered,
then dissolved in CHCl3, washed 4× with distilled water and once with
brine, then dried with anhydrous MgSO4. After filtering to collect the
supernatant and removing the solvent under vacuum, the resulting pale
yellow solid was stirred with diethyl ether and filtered to give a fine
white powder (7.7 g, 85% yield). 1H NMR (300 MHz, CDCl3): δ 7.60
(dd, J = 7.2 Hz and 1.2 Hz, 6H), 7.29 (tt, J = 7.5 and 1.2 Hz, 3H), 7.27
(br, 3H), 7.06 (t, J = 8.1 Hz, 6H), 3.55 (quart, J = 5.4 Hz, 6H), 2.72 (t,
J = 5.7 Hz, 6H). ESI-TOF-MS (MeOH) m/z: [M + H]+ calc’d for
C27H31O3N4, 459.24; found: 459.28, 481.27 [M + Na+], 497.27 [M +
K+], 497.27 [2 M + H+], 917.58 [2 M + Na+] 939.57.
Synthesis of K(THF)[CoLPh] (2). H3LPh (1.499 g, 2.2 mmol) was
dissolved in THF (180 mL) and cooled to −78 °C. Benzylpotassium
(895 mg, 6.87 mmol, in 10 mL of THF) was added dropwise over 5
min, during which time the solution turned bright yellow. The solution
was stirred for 15 min at −78 °C. CoCl2 (286 mg, 2.2 mmol) was then
added. The resulting cloudy green solution was allowed to slowly
warm to rt overnight. After 12 h, the mixture was filtered through
Celite, and the solvent was removed under vacuum. The solid was
washed with toluene (15 mL) and pentane (5 mL), and then dried
under vacuum, yielding a bright, lime-green powder (1.60 g, 90%
1
yield). H NMR (500 MHz, CD3CN): δ 70.2 (br, 6H), 59 (br, 6H),
15.2 (s, 6H), 13.3 (br, 6H), 4.36 (s, 6H), 0.41 (s, 3H), −1.86 (s, 3H)
−3.76 (s, 6H). UV−vis (THF): λmax, nm (ε, M−1 cm−1): 275 (30,900),
C
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Crystallographic Details for K(THF)[CoLPh] (2), K[CoLtBu] (2a), Co2LPh (3), and FeCoLPh (4)
2
2a
3
4
chemical formula
formula weight
crystal system
space group
a (Å)
C45H42CoKN7(C4H8O)
850.99
C39H54N7CoK
718.92
C45H42N7Co2
798.72
C45H42N7CoFe
795.64
monoclinic
P21/c
orthorhombic
Pbca
trigonal
trigonal
R3
̅
R3
̅
11.3407(5)
21.550(1)
17.9220(8)
90
20.568(2)
19.149(2)
24.290(2)
90
14.937(1)
14.937(1)
29.199(2)
90
14.956(2)
14.956(2)
29.250(4)
90
b (Å)
c (Å)
α (deg)
β (deg)
92.494(1)
90
90
90
90
γ (deg)
V (Å3)
90
120
120
4375(9)
9567(2)
8
5641.8(6)
6
5666(1)
6
Z
4
Dcalcd (g cm−3
)
1.292
0.998
1.411
1.399
λ (Å), μ (mm−1
)
0.71073, 0.53
173(2)
0.71073, 0.475
173(2)
0.71073, 0.925
173(2)
0.71073, 0.866
173(2)
T (K)
θ range (deg)
reflns collected
unique reflns
2.29 to 26.47
49826
1.68 to 27.59
92104
1.72 to 28.24
21580
1.72 to 27.50
13888
10090
11058
3083
2900
data/restraint/parameters
10090/92/714
0.0446, 0.1083
11058/0/439
0.0554, 0.1264
3083/0/163
0.0378, 0.0952
2900/0/163
0.0380, 0.0834
R1, wR2 (I > 2σ(I))
355 (20,350), 598 (80), 618 (80), 780 (12). Anal. Calcd for
C45H42N7CoK(OC4H8): C, 69.16; H, 5.92; N, 11.52. Found C, 69.22;
H, 6.00; N, 11.61.
X-ray Crystallographic Data Collection and Refinement of
the Structures. Single crystals of K(THF)[CoLPh] (2), K[CoLtBu
]
(2a), Co2LPh (3), and FeCoLPh (4) were grown by vapor diffusion of
Et2O or pentane into concentrated THF solutions of 2, 2a, 3, and 4 at
rt, respectively. Green blocks of 2 (0.60 × 0.30 × 0.20 mm3) and 2a
(0.50 × 0.38 × 0.18 mm3), a brown block of 3 (0.5 × 0.5 × 0.4 mm3),
and a red plate of 4 (0.36 × 0.20 × 0.06 mm3) were placed on the tip
of a glass capillary and mounted on a Bruker APEX-2 Platform CCD
diffractometer for data collection at 173(2) K. The data collection was
carried out using Mo−Kα radiation (graphite monochromator). The
data intensity was corrected for absorption and decay (SADABS).
Final cell constants were obtained from least-squares fits of all
measured reflections. The structure was solved using SHELXS-97 and
refined using SHELXL-97. A direct-methods solution was calculated
which provided most non-hydrogen atoms from the E-map. Full-
matrix least-squares/difference Fourier cycles were performed to locate
the remaining non-hydrogen atoms. All non-hydrogen atoms were
refined with anisotropic displacement parameters, respectively. Hydro-
gen atoms were placed ideally and refined as riding atoms with relative
isotropic displacement parameters. For 2a, the PLATON program,
SQUEEZE function, was used to remove disordered solvent
Synthesis of K(THF)[CoLtBu] (2a). This compound was synthesized
in a manner entirely analogous to that of K(CoLPhK) (2), above, using
H3LtBu (751 mg, 1.2 mmol), benzylpotassium (475 mg, 3.65 mmol),
1
and CoCl2(THF)1.5 (287 mg, 1.2 mmol). Yield: 812 mg (85%). H
NMR (500 MHz, CD3CN): δ 140.9 (br, 6H), 62.8 (br, 6H), 16.0 (br,
27H), 8.02 (s, 6H), 3.11 (s, THF, 1.3H), 2.74 (s, 6H), 1.17 (s, THF,
1.3H) −19.1 (s, 3H). Anal. Calcd for C39H54N7CoK(OC4H8): C,
65.29; H, 7.90; N, 12.39. Found C, 65.21; H, 7.96; N, 12.36.
Synthesis of Co2LPh (3). K(THF)[CoLPh] (325 mg, 0.382 mmol)
was dissolved in THF (120 mL) and cooled to −78 °C. Potassium
graphite (54 mg, 0.399 mmol) was added as a slurry in THF (3 mL),
causing the reaction mixture to turn dark yellow. After stirring for 5
min, CoBr2 (85 mg, 0.39 mmol) was added dropwise as a solution in
THF. The mixture was allowed to slowly warm to rt overnight. After
24 h, it was filtered to remove graphite, giving a dark brown filtrate,
which was pumped down under vacuum, taken up in toluene, filtered
through Celite, and dried to give a brown solid (285 mg, 90% yield).
1H NMR (500 MHz, THF-d8): δ 158.7 (br, 6H), 104.6 (br, 6H), 24.2
(s, 6H), 16.5 (s, 6H), 10.4 (br s, 6H), −0.09 (s, 3H), −37.1 (s, 3H),
−39.05 (br, 6H). 1H NMR (500 MHz, C6D6): δ 165.3 (br, 6H), 104.5
(br, 6H), 23.9 (s, 6H), 16.0 (s, 6H), 8.66 (br, 6H), −1.73 (s, 3H),
−34.4 (s, 3H), −35.1 (br, 6H). Vis-NIR (THF): λmax, nm (ε, M−1
cm−1): 480 sh (2200), 650 sh (430), 850 (150), 1140 (120), 1340 sh
(105). Anal. Calcd for C45H42N7Co2: C, 67.67; H, 5.30; N, 12.28.
Found C 67.58; H 5.24; N 12.17.
comprising two molecules of THF/Et2O per asymmetric unit.41
A
total of 672 electrons were removed in a total volume of 2976 Å3 per
unit cell, equally distributed between two positions at (0, 0, 0) and
(0.3, 0.3, 0.5) in the asymmetric unit. The number of electrons is
consistent with removal of 12 ether and 4 THF molecules per unit cell,
which is consistent with the 3:1 ratio observed by 1H NMR
spectroscopy. Crystallographic data are summarized in Table 1.
Anomalous Diffraction Data Collection and Refinement of
Fe/Co Occupancies in 4. A single crystal of 4 was mounted on a
glass fiber and cooled to 100 K using an Oxford Instruments Cryojet
cryostat. The Bruker D8 diffractometer, integrated with an APEX-II
CCD detector, was modified for synchrotron use at the Chem-
MatCARS 15-ID-B beamline at the Advanced Photon Source
(Argonne National Laboratory). Diffraction data were collected at
seven different energies with 0.3 s frames while manually attenuating
the beam to minimize overages of individual pixels. The scan at 30.0
keV (λ = 0.41328 Å), which is energetically well above the atomic
absorption energies, gave a least-squares refinement of all model
positional- and displacement parameters to 0.5 Å resolution. To
determine the compositions of Fe/Co at the two independent metal
sites, a total of six anomalous diffraction data sets were collected near
the absorption K-edges of Fe and Co (3 each): one preceding the Fe
K-edge at 7.062 keV (1.75567 Å); one at the Fe K-edge 7.112 keV (λ
Synthesis of FeCoLPh (4). K(THF)[CoLPh] (260 mg, 0.306 mmol)
was dissolved in THF (100 mL) and cooled to −78 °C. Potassium
graphite (45.2 mg, 0.334 mmol) was added as a slurry in THF (3 mL),
causing the reaction mixture to turn dark yellow. After stirring for 5
min, FeBr2 (72.0 mg, 0.334 mmol) was added dropwise as a solution
in THF. The mixture was allowed to slowly warm to rt overnight. After
12 h, it was filtered to remove graphite, giving a dark purple filtrate,
which was pumped down under vacuum, taken up in benzene, filtered
through Celite, and dried under vacuum to give a purple powder (225
1
mg, 90% yield). H NMR (500 MHz, C6D6): δ 172.6 (br, 6H), 127.0
(br, 6H), 23.13 (s, 6H), 11.40 (s, 6H), 6.94 (br s, 6H), 0.78 (s, 3H),
24.1 (s, 3H), −31.5 (br, 6H). Vis-NIR (THF): λmax, nm (ε, M−1
cm−1): 514 (3000), 1050 (175). Anal. Calcd for C45H42N7CoFe: C,
67.93; H, 5.32; N, 12.32. Found C 67.86; H 5.28; N 12.23. ICP-OES
(wt %): Fe, 5.5157; Co, 8.1954.
D
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
= 1.74332 Å); one following the Fe K-edge at 7.162 keV (1.73115 Å);
one preceding the Co K-edge at 7.659 keV (1.61882 Å); one at the Co
K-edge at 7.709 keV (λ = 1.60832 Å); and one following the Co K-
edge at 7.759 (1.59795 Å). The anomalous diffraction can readily
distinguish Fe/Co compositions at the two metal sites because of the
expected differences in the Δf′ and Δf″ values for these two elements,
as shown in Figure 3. Basically, Δf′ and Δf″ values of an element
controller. X-band EPR spectra were simulated using the ESIM
program written by Eckhard Bill. Mossbauer data were recorded on an
̈
alternating constant acceleration spectrometer. The minimum
experimental line width was 0.24 mm s−1 (full width at half-height).
The 57Co/Rh source (1.8 GBq) was positioned at rt inside the gap of
the magnet system at a zero-field position. Isomer shifts are quoted
relative to iron metal at 300 K.
Magnetic susceptibility data were measured from powder samples of
solid material in the temperature range 2−300 K by using a SQUID
susceptometer with a field of 1.0 T (MPMS-7, Quantum Design,
calibrated with standard palladium reference sample, error <2%).
Multiple-field variable-temperature magnetization measurements were
done at 1 T, 4 T, and 7 T also in the range 2 to 300 K with the
magnetization equidistantly sampled on a 1/T temperature scale. The
experimental data were corrected for underlying diamagnetism by use
of tabulated Pascal’s constants (χdia < 0),43,44 as well as for
temperature-independent paramagnetism (χTIP > 0).45 The latter
was adjusted such that χ*T was obtained constant above 50 K after
subtraction of χTIP. The susceptibility and magnetization data were
simulated with the program julX for exchange coupled systems.46 The
simulations are based on the usual spin-Hamiltonian operator for
mononuclear complexes with spin S:
⎡
⎤
2
1
3
2
2
̂
⇀ ⇀
̂
̂ ̂
S(S + 1) + E/D(Sx − S )
̂
H = gβ S ·B + D S
−
⎢
⎥
⎦
z
y
⎣
(1)
where g is the average electronic g value, and D and E/D are the axial
zero-field splitting and rhombicity parameters. Magnetic moments are
calculated after diagonalization of the Hamiltonian from the
Figure 3. Anomalous dispersion corrections to normal scattering
factors, including the real (Δf′) and imaginary (Δf″) components, for
Fe (red) and Co (blue) as a function of wavelength (Å). The four
dotted lines are the experimental wavelengths for the anomalous
experiments, which were selected to span the Fe and Co absorption
edge energies. The cross marks (×) indicate the Fe and Co anomalous
scattering factor values used in the least-squares refinement to
determine the metal occupancies.
⃗
eigenfunctions using the Hellman-Feynman theorem μ (B) = ⟨ψ |
⃗
i
i
⃗
dH/dB|ψi⟩. Intermolecular interactions were considered by using a
Weiss temperature, ΘW, as perturbation of the temperature scale, kT′ =
k(T − ΘW) for the calculation. Powder summations were done by
using a 16-point Lebedev grid. Because the program is presently not
dimensioned for individual spins larger than 5/2, we reproduced the
septet ground state of 4 by adopting two individual spins S1 = S2 = 3/2
with an arbitrarily chosen large, positive exchange coupling constant, J
= +200 cm−1, which is defined as:
change dramatically near the element’s absorption edge, but, for other
element(s), they remain relatively constant. Each of the anomalous
diffraction data sets thus provides a different view of the electrons
present at both sites. Since metal K-edges may shift slightly within
coordination complexes, the data collected at the metal K-edges are
less reliable than those collected above and below the K-edge energies.
Hence, only four of the six data sets were used in a least-squares
refinement to determine the Fe/Co occupancies at the two metal sites.
GSAS-II was employed for these least-squares refinements because it
allows multiple diffraction data sets as an input with subsequent
refinement using a common crystallographic model.42 The 30 keV data
was refined using a structural model of 4 that had been determined at
173 K (see above). The converged positional- and displacement
parameters were than frozen, so that only the Fe/Co occupancies of
the two independent sites could be refined simultaneously using the
six anomalous diffraction data sets. The refinement results shows
mixed occupancies of the metal elements at both sites with
significantly higher percentages of Fe at the top site (Fe 93.8%, Co
6.2%) and Co at the bottom site (Co 95.2%, Fe 4.8%) (see Supporting
Information). Hence, the precise structural formula of 4 is
(Fe0.94(1)Co0.06(1))(Co0.95(1)Fe0.05(1))LPh. Crystallographic details are as
follows: empirical formula, C45H42N7Co1.01Fe0.99; fw, 795.66; trigonal;
̂
̂
⇀ ⇀
̂
̂
̂
H = −2J[ S1· S2] + He,1 + He,2·
(2)
This approach is physically equivalent to a simulation with an
isolated total spin S = 3 for the bimetallic, because the resulting septet
ground state is energetically well isolated by an energy gap ΔE of 1200
cm−1 or more from the other calculated total spin manifolds (ΔE/k >
1700 K), such that Boltzmann population of other manifolds is
negligible up to ambient temperature (300 K).
Computational Methods. DFT Calculations. Bimetallic 4 was
studied using density functional theory (DFT) and the complete active
space self-consistent field (CASSCF) method,47 followed by a
multiconfigurational second-order perturbation theory (CASPT2)
method.48 Previous studies on similar systems have demonstrated
that this approach is successful in predicting accurate results for
ground and electronically excited states of bimetallic systems.49−52
Gas-phase geometry optimizations were performed for the various
possible spin states at the DFT level of theory with the use of the
Perdew−Burke−Ernzerhof (PBE) exchange-correlation functional53
within the TURBOMOLE 6.1 program package.54 For C and H atoms,
the double-ζ quality basis set def-SV(P) was used, whereas the triple-ζ
quality basis set def-TZVP was employed for N and additional
polarized functions where introduced by using def-TZVPP for Fe and
Co. The DFT calculations were performed with the broken symmetry
option (unrestricted calculations) and the resolution-of-the-identity
(RI) approximation for several spin states.55
R3; a,b = 14.9092(6) Å, c = 29.191(1) Å; α,β = 90°, γ = 120°; V,
̅
5619.4(5) Å3; Z, 6; Dcalcd, 1.411 g/cm3; μ, 0.205 mm−1; λ, 0.41328 Å;
T, 100(2) K; θ, 1.86−20.30°; reflns collected, 45843; unique reflns,
6123; data/restraint/parameters, 6123/0/163; R1,wR2 (I > 2σ(I)):
0.0377, 0.0891.
Physical Measurements. NMR spectra were collected on a
Varian Inova 500 MHz spectrophotometer. Visible and near-infrared
absorption data were collected on a Cary-14 spectrophotometer. UV-
wavelength absorption spectra were collected on a Cary 300 Bio UV−
visible spectrophotometer. Perpendicular-mode X-band EPR spectra
were recorded on a Bruker EPP 300 spectrometer equipped with an
Oxford ESR 910 liquid helium cryostat and an Oxford temperature
CASSCF/CASPT2 Calculations. All CASSCF/CASPT2 calculations
were performed with the MOLCAS-7.4 package56 using the DFT-
optimized structure. The relativistic all-electron ANO-RCC basis
sets57,58 were used for all elements. In all of these calculations, the
ANO-RCC-VDZP basis set was used for Fe, Co, and N while the
ANO-RCC-MB basis set was used for C and H. Scalar relativistic
effects were included by using the Douglas−Kroll−Hess Hamil-
E
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 1
Scheme 2
Scheme 3
tonian.59 The two-electron integral evaluation was simplified by
employing the Cholesky decomposition technique.60−62 To avoid
intruder states, an imaginary level shift of 0.2 au was used in the
CASPT2 calculations.63 A full valence complete active space consisting
14 d-electrons in 12 orbitals (14,12) was used. The 12 orbitals
consisted of the ten 3d orbitals of the two metal centers and two
virtual orbitals comprising the 4d Co orbitals.
significant Co−Co interaction. Of note, the loss of a ligand
upon reduction, as observed here, had not been reported.
We have been interested in developing a general synthetic
strategy to access other high-spin bimetallics, and for this
purpose, the tris(amidinato)amine ligands were designed. The
tris(amidine)amine proligands, H3LPh and H3LtBu, are, to our
knowledge, previously unknown. Their syntheses consist of
three relatively simple and good-yielding reactions from
commercially available tris(2-aminoethyl)amine, or tren. First,
benzoyl or pivaloyl chloride (3.4 equiv.) is added to tren in the
presence of excess triethylamine to generate the tris(amide)-
amine species. Second, the amide groups are transformed into
imidoyl chloride functionalities using phosphorus pentachloride
(Scheme 2). Third, nucleophilic substitution of aniline at each
imidoyl chloride group produces the neutral proligands, H3LPh
and H3LtBu after a basic workup.
3. RESULTS
Synthesis. Cotton, Murillo, and co-workers reported the
one-pot synthesis of Co2(DPhF)3 (1) from the Co(II)
precursor, CoCl2(HDPhF)2.23 Presumably, the reaction occurs
according to the balanced reaction,
4nBuLi
2CoCl2(HDPhF)2 ⎯⎯N⎯⎯a⎯B⎯⎯E⎯t⎯3⎯H→⎯ ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→⎯ Co2(DPhF)3
−78 to 0°C −78°C to rt
+ NaCl + 3LiCl + Li(DPhF) + 0.5H2 + BEt3
From the H3LPh proligand, the new dicobalt complex, Co2LPh
(3), is readily obtained after two sequential metalation reactions
(Scheme 3). In both metalations, the cobalt sources are simple
cobalt dihalides. Based on NMR analysis of the crude products,
CoCl2 reacts more cleanly than CoBr2 in the first metalation,
while CoBr2 is preferred in the second. The first metalation
involves deprotonation of the proligand with benzylpotassium
followed by a salt metathesis reaction with CoCl2 to yield the
monoanionic monocobalt complexes, K(THF)[CoLPh] (2).
The second metalation involves in situ reduction of 2 with KC8,
followed by the dropwise addition of a CoBr2 solution in THF.
A similar strategy is successful for obtaining the t-butyl
derivative, K(THF)[CoLtBu] (2a), though preparation of the
dicobalt counterpart Co2(LtBu) has been unsuccessful. Attempts
to make Co2LtBu gave a product with too many peaks in the
proton NMR spectrum, which is inconsistent with the desired,
wherein NaBEt3H acts as the reducing agent, nBuLi
deprotonates the amidine ligands, and the extra amidinate
ligand is lost as Li(DPhF). In our hands, this reaction gave
variable yields of 1, which is not so surprising given the
complicated set of elementary reactions that is needed to form
the mixed-valent bimetallic. To obtain 1 more reliably, we
developed the simplified procedure shown in Scheme 1 that
utilizes the known tetragonal Co(II)Co(II) analogue,
Co2(DPhF)4. By reducing Co2(DPhF)4 with one equivalent
of KC8, 1 can be generated cleanly and isolated in ∼60%
crystalline yield. K(DPhF) is presumed to be the byproduct.
The loss of an amidinate ligand from Co2(DPhF)4 has been
previously observed in a complementary ligand-abstraction
reaction.19 Specifically, the salt metathesis of Co2(DPhF)4 with
AgPF6 proceeds with loss of 0.5 equiv of [Ag(DPhF)]2 to the
trigonal dicobalt salt, [Co2(DPhF)3(CH3CN)2]PF6, wherein
the Co···Co separation of 2.885(1) Å is too long to have any
C3-symmetric Co2LtBu
.
F
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
The mononuclear cobalt species, 2, is a potential platform for
the synthesis of other bimetallics. Following a similar synthetic
protocol as that for 3, the heterobimetallic FeCoLPh 4 can be
prepared by using FeBr2 in the second metalation step.
NMR and Vis-NIR Data. The 1H NMR data for the trigonal
dicobalt complexes 1 and 3 is consistent with 3-fold or higher
symmetry (Figure 4). Compound 1 exhibits four resonances,
Figure 5. Vis-NIR spectra of dicobalt complexes 1 (in red), 3 (in
blue), and iron−cobalt 4 (in green) in toluene at rt. Inset shows the
NIR region (ε vs nm). An asterisk marks the onset of imperfect
background subtraction. See text for details.
M−1 cm−1, respectively), and an even weaker absorption in the
near-infrared (NIR) region at λmax ∼ 1665 nm (ε ∼ 100 M−1
cm−1). The electronic absorption spectrum of 1 is similar to
Fe2(DPhF)3, which exhibited four bands between 650 and 1250
nm. Based on symmetry and theoretical analyses, these
electronic transitions in Fe2(DPhF)3 were assigned as spin-
and dipole-allowed d-d excitations in a fully delocalized metal−
metal bonding scheme.28 By analogy, the Vis/NIR bands
exhibited by 3 are tentatively assigned as d-d transitions of the
dicobalt core.
The electronic absorption of 3 is the least distinctive,
resembling a broad curve that tapers gradually into the Vis/NIR
region. Four features can be distinguished: two shoulders at 480
and 650 nm, and two weak bands at 850 and 1140 nm (ε = 150,
120 M−1 cm−1, respectively). Complex 4 has intense
absorptions centered at 514 nm (ε ∼ 3,000 M−1 cm−1) and a
weak band at 1050 nm (ε = 175 M−1 cm−1). For 3 and 4, the
weak NIR bands are likely to be d-d transitions of a high-spin,
delocalized dicobalt and iron−cobalt core, respectively. The
visible transitions of 3 and 4 are highly intense and are likely to
be charge-transfer bands between the metal and the amidinate
ligands. The shifting of the band to higher energies for dicobalt
relative to iron−cobalt would be more consistent with metal-to-
ligand charge-transfer (MLCT).
X-ray Diffraction and Anomalous Scattering Studies.
En route to the M-Co bimetallics, the monocobalt anion,
K(THF)[CoLPh] 2, was isolated as a bright green solid. Single
crystals of 2 were grown by vapor diffusion of pentane into a
concentrated THF solution and analyzed by X-ray diffraction. A
single crystal of the t-butyl analogue, K[CoLtBu] 2a, has also
been analyzed. The solid-state structures of 2·THF and 2a are
shown in Figures 6 and 7. In both structures, the ligand is
tetradentate in a κNax,κ3Neq-coordination mode. The geometry
at the cobalt center is trigonal monopyramidal, which is typical
of Co(II) centers supported by related tris(amido)amine
ligands.64−66 The Co−Neq bond distances are slightly longer
in 2a (2.01 to 2.02 Å) than in 2 (1.96 to 1.98 Å). We were
gratified to observe that the three remaining nitrogen atoms of
the ligand remain unchelated with respect to cobalt. They do
bind the potassium countercation, creating a continuous, two-
dimensional network in the crystal lattice. Conceivably, these
Figure 4. 1H NMR (500 MHz, C6D6) spectra of dicobalt complexes 1
(top) and 3 (middle), as well as the iron−cobalt complex 4 (bottom).
The residual solvent resonance is indicated by an asterisk.
ranging from −50.2 to 175.1 ppm, where the most downfield
peak is assigned to the proton in the DPhF backbone (i.e.,
NCHN) based on its integration. The number of resonances is
consistent with the idealized D3h symmetry expected for 1. For
3, eight signals are observed, which is consistent with the
expected C3 symmetry in solution. The proton shifts range
from −35.1 to 165.3 ppm, but the most distinctive feature is the
pair of broad resonances at downfield shifts of 165.3 and 104.5
ppm. Each broad peak integrates to 6 protons. Because these
two resonances have no analogues among the peaks observed
for 1, they likely correspond to either the two methylene
groups of the tren backbone or the meta/ortho resonances of
the bridgehead phenyl group in the amidinate moiety, that is,
NCPhN.
The 1H NMR spectrum for the iron−cobalt species 4 is very
similar to that of dicobalt 3 (Figure 4). In comparing the two
bimetallics, the most shifted peaks are the two downfield
resonances at 172.6 and 127.0 ppm. The fact that these two
peaks are highly sensitive to the swapping of the top metal
center could suggest that they correspond to the bridgehead
phenyl group rather than the tren backbone.
The dicobalt complexes 1 and 3 are red-brown and brown,
respectively, while the iron−cobalt compound 4 is purple. As
shown in Figure 5, 1 has two relatively weak absorptions in the
visible (Vis) region at λmax = 545 and 750 nm (ε = 530, 200
G
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
donor. In addition, the cobalt atoms are significantly displaced
out of the planes defined by the equatorial nitrogen donors,
away from the axial donor.
Akin to 3, iron−cobalt 4 crystallizes in the same space group,
R3, with C symmetry (Figure 9, Table 2). Disorder that arises
̅
3
from metal-mixing can be challenging, if not impossible, to
detect by standard X-ray diffraction methods when the metals
have similar atomic numbers as in the case of 4. To differentiate
iron and cobalt, we turned to X-ray anomalous scattering
techniques, which have been shown to be effective at
quantifying metal-mixing disorder within single crystals.67,68
In these studies, several sets of X-ray diffraction data were
collected on a single crystal of 4, with the wavelength of the X-
ray source tuned to cover a range of energies near the K-edge
absorption energies of iron and cobalt. A final data set was
collected at significantly higher energy. These studies exploit
the fact that anomalous dispersion contributions to the
scattering factor change substantially near the scattering
atom’s absorption edge energy. Since edge energies are
significantly different even for similar metals such as iron and
cobalt, the occupancies of each metal at a specific crystallo-
graphic site can be determined. (Figure 3). We have
determined the ratios of Fe/Co in the two metal sites by
conducting least-squares refinement of four anomalous
diffraction data collected at the different wavelengths
simultaneously. This method varies from other literature
reports that calculate Fourier difference maps (or apparent
f′) at each wavelength.67,69 The current method has the benefit
of reporting refined values with standard uncertainties and
provides a straightforward scaling of the different anomalous
data sets.
Figure 6. Solid-state structure of 2 at 50% probability. Hydrogen
atoms and THF molecule (coordinated to K) are omitted for clarity.
Only the shortest K−N bond is shown. Selected bond distances (Å):
Co−N1 2.108(2), Co−N2 1.958(2), Co−N4 1.978(2), Co−N6
1.983(2), K−N7 2.845(2); selected bond angles (deg): N2−Co−N4
117.12(9), N4−Co−N6 122.32(9), N2−Co−N6 118.11(9), N1−
Co−N2 85.28(8), N1−Co−N4 84.52(7), N1−Co−N6 84.56(8).
The least-squares refinement of the metal occupancies gives a
precise structural formula of (Fe0.94(1)Co0. 06(1))-
(Co0.95(1)Fe0.05(1))LPh for 4, wherein a small degree of iron−
cobalt mixing occurs at both metal-binding sites. As expected
from the synthetic protocol, the cobalt is predominantly in the
“bottom” binding pocket, featuring the tren donor set, while the
iron is primarily located in the “top” binding pocket. Although
mixing of iron and cobalt is observed at both metal-binding
sites, the dominant metal accounts for 94 to 95% of the
occupancy at each site. This degree of selectivity is somewhat
remarkable considering the similarity of the ligand donors and
the known lability of high-spin metal centers (vide infra). Of
note, a recent study has exploited such lability to introduce
cobalt center(s) into a preassembled triiron core as a synthetic
route to iron−cobalt heterometallics.70
By taking the metal occupancies at the two binding sites and
assuming purely statistical mixing, we can predict the
percentages of each possible metal−metal pair. With this
assumption, the single crystal of 4 comprises 89.3% FeCoLPh,
5.9% CoCoLPh, 4.5% FeFeLPh, and 0.3% CoFeLPh. Unfortu-
nately, much higher data resolution would be needed in the
crystal structure determination to fully model this four-
component mixture. Hence, the bond distances and angles
reported for 4 reflect this composite, of which the majority
species is FeCoLPh. For simplicity, we discuss bond metrics of 4
with respect to its major component, but clearly, some
uncertainty exist in these values because of their composite
nature.
Figure 7. Solid-state structure of 2a at 50% probability. Hydrogen
atoms are omitted for clarity. Only the shortest K−N bond is shown.
Selected bond distances (Å): Co−N1 2.094(2), Co−N2 2.013(2),
Co−N4 2.018(2), Co−N6 2.022(2), K−N5 2.785(2); selected bond
angles (deg): N2−Co−N4 118.89(9), N4−Co−N6 119.15(9), N2−
Co−N6 119.99(9), N1−Co−N2 85.70(8), N1−Co−N4 85.11(9),
N1−Co−N6 85.14(9).
donors are available to bind a second metal center that can be
introduced in a second metalation step.
Cotton, Murillo, and co-workers previously reported the
molecular structure of 1, which crystallized in the triclinic space
group P1.19 Inspection of the Co−Neq bond lengths and the
̅
Neq−Co−Neq angles in 1 reveals a slight distortion from 3-fold
symmetry about the Co−Co axis (Table 2). A similar distortion
was observed in the diiron analogue Fe2(DPhF)3 and was
attributed to crystal packing forces.22 In contrast to 1, our
dicobalt 3 complex crystallizes in the trigonal space group R3,
̅
resulting in perfect C3 symmetry (Figure 8, Table 2).
One notable difference between the experimental structures
of 1 and 3 is the significant contraction in the Co−Co bond
length (ΔCo−Co = 0.09 Å) of 3. The shorter Co−Co bond
distance in 3 is surprising because one normally expects the
addition of an axial ligand to attenuate the metal−metal
interaction via the ligand’s trans influence. For example, it is
well-established that adding axial donors to dichromium
tetracarboxylates significantly increases the Cr−Cr bond
length.8 In 3, the presence of the axial donor and the
asymmetric nature of the amidinate donors differentiate the two
cobalt centers. The Co−Neq bond distances are significantly
shorter for the cobalt center that is bound to the axial amine
The solid-state structures of 3 and 4 are surprisingly similar.
For instance, all the metal−ligand bond distances are within
0.019(3) Å, despite the swapping of cobalt with iron in the top
binding pocket. The only significant change occurs in the M−
H
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2. Selected Distances (Å) and Angles (deg) for Experimental Bimetallic Structures: 1, 3, and 4
a
distances, angles
Co−Co, Å
Co−Neq, Å
Co2(DPhF)3, 119
Co2LPh, 3
bonds, angles
Fe−Co, Å
Co−Neq, Å
Fe0.99Co1.01LPh, 4
2.385(1)
1.932(5)
1.955(5)
1.959(6)
1.957(5)
1.978(5)
1.983(5)
2.2943(7)
1.923(2)
2.1846(4)
1.927(1)
2.041(2)
Fe−Neq, Å
2.0528(9)
Co−Nax, Å
Co-out of N3-plane, Å
2.135(3)
0.164
Co−Nax, Å
2.115(2)
0.116
0.090
Co-out of N3-plane
Fe-out of N3-plane
Neq−Co−Neq, deg
0.065
0.186
0.029
Neq−Co−Neq, deg
Neq−Co−Co, deg
Neq−C−Neq, deg
125.1(2)
119.2(2)
115.6(2)
120.1(2)
114.8(2)
124.7(2)
88.19(2)
88.24(2)
87.07(2)
90.06(2)
90.32(2)
91.30(2)
121.6(6)
122.7(6)
122.4(6)
119.28(2)
119.643(6)
119.18(2)
84.77(5)
94.89(5)
116.8(2)
Neq−Fe−Neq, deg
Neq−Fe−Co, deg
Neq−Co−Fe, deg
Neq−C−Neq, deg
119.982(1)
89.20(3)
93.44(3)
116.9(1)
a
Structure was determined from data collected using X-ray synchrotron radiation at 30 keV. See Experimental Section.
Figure 9. Solid-state structure of Fe0.99Co1.01LPh 4 at 50% probability.
Hydrogen atoms are omitted for clarity. The pie charts depict the
percentages of iron and cobalt present at each metal binding site (as
determined by X-ray anomalous dispersion). Selected bond distances
(Å): Fe−Co 2.1846(4), Co−N2 1.927(1), Co−N1 2.115(2), Fe−N3
2.0528(9); selected bond angles (deg): N2−Co−N2′ 119.643(6),
N3−Fe−N3′ 119.982(1), N1−Co−N2 86.56(3), N1−Co−Fe 180.0,
N2−Co−Fe 93.44(3), N3−Fe−Co 89.20(3); torsion (deg): N2−Co−
Fe−N3 4.30(3).
Figure 8. Solid-state structure of 3 at 50% probability. Hydrogen
atoms are omitted for clarity. Selected bond distances (Å): Co1−Co2
2.2943(7), Co1−N2 1.923(2), Co1−N1 2.135(3), Co2−N3 2.041(2);
selected bond angles (deg): N2−Co1−N2′ 119.28(2), N3−Co2−N3′
119.18(2), N1−Co1−N2 85.11(5), N1−Co1−Co2 180.0, N2−Co1−
Co2 94.89(5), N3−Co2−Co1 84.77(5); torsion (deg): N2−Co1−
Co2−N3 4.57(7).
Co bond distance, which shortens from 2.2944(7) Å in 3 (M =
Co) to 2.1846(4) Å in 4 (M = Fe). Correlated with the shorter
M-M bond, the iron center is less displaced from the equatorial
N3-plane. If one considers that the covalent radius of iron is
larger than that of cobalt, then the M-Co bond distance is
expected to increase from 3 to 4, which is opposite of what is
observed. Hence, we ascribe the shorter Co−Fe bond to
depopulating a d-electron from an M−Co antibonding orbital,
which would increase the formal M-M bond order by 0.5. The
iron−cobalt complex 4 is an unusual example of iron−cobalt
bonding. Nearly all the structural examples of iron−cobalt
bonds in the Crystallographic Structural Database are within
carbonyl clusters, where the median Fe−Co bond distance is
2.55 Å.71 Only a single molecule has Fe−Co bonds reasonably
close to 4: a trinuclear Fe2Co core with a capping sulfide atom
(Fe−Co 2.225 and 2.296 Å).72
̈
Magnetic Measurements and EPR/Mossbauer Spec-
troscopy. Because spin states can be challenging to predict for
metal−metal bonds of late first-row transition metals, we were
interested in probing the effect of an apical donor and/or metal
identity on the electronic structure of these metal−metal
bonds. Owing to the high air-sensitivity of 1, the assignment of
its spin ground state had been based primarily on a theoretical
I
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
study.25 SCF-Xα-SW calculations on the truncated molecule
Co2(HNCHNH)3 favored the sextet spin.
We have investigated the variable temperature magnetic
susceptibility measurements of the dicobalt complexes 1 and 3,
as shown in Figure 10. From ∼50 to 290 K, the effective
Figure 10. Temperature dependence of the effective magnetic
moment, μeff, of 1 (green circles) and 3 (blue circles), at 1 T, from
2 to 290 K. The solid red lines represent the best fit. See text for
simulation parameters. The insets show the result of multifield
measurements recorded at 1, 4, and 7 T.
Figure 11. X-band EPR spectrum (dX″/dB) of 1 and 3 in toluene
glass shown in blue (1.0 mM, 20.0 K, frequency = 9.646 GHz,
modulation to 10 G, power = 2.01 mW). The spectrum was simulated
(shown in red) by adopting S = 5/2 with the following anisotropic
values: g = (2.243, 2.243, 2.12) for 1 and g = (2.201, 2.201, 2.215) for
3; line widths, W = (172.6, 172.6, 700) for 1 and (184.2, 184.2, 400)
for 3. Zero-field splitting parameters: for 1, D = 9 cm−1, E/D = 0.0509;
for 3, D = 6 cm−1, E/D = 0.0366 (D values fixed).
magnetic moment, μeff, is essentially temperature independent
with values of 6.92 and 6.39 μB for 1 and 3, respectively (χ*T =
5.99 cm3 K mol−1 for 1 and 5.11 cm3 K mol−1 for 3). These
plots are consistent with an energetically well-isolated sextet
state, which has a spin-only value of 5.92 μB. Hence, the data
can be fitted using a single spin S = 5/2 Hamiltonian for the
bimetallics 1 and 3. The deviation of the μeff versus T plots
from the spin-only value for 1 and 3 can be simulated by
adopting electronic g values different from g = 2; best fits are
obtained with g5/2 = 2.33 and 2.16. Because the presence of any
magnetic impurities can obscure true g-values in these static
susceptibility measurements, more reliable g-values are obtained
from EPR spectroscopy (vide infra).
Below 50 K, the effective magnetic moments of 1 and 3
deviate from the high-temperature limit because of the
combined effect of field saturation, zero-field splitting, and
weak intermolecular interactions, which together can be
reasonably simulated as shown in the inset of Figure 10.
Indeed, weak intermolecular ferromagnetic coupling is best
recognized as the small rise in μeff(T) around about 20 K. The
effect was approximated by using a mean-field approximation
with a Weiss-constant θw for parametrization. Best fits are
obtained with zero-field splitting parameter D5/2 = 9( 2) and
6( 3) cm−1 for the ground states of 1 and 3, respectively. The
corresponding values for θw are 1.5 K in both cases. We further
note that the intermolecular interactions should be weak as the
closest intermolecular M···M contacts are all greater than 8.5 Å
(Supporting Information, Figure 10).
simulated with the usual spin-Hamiltonian for the sextet spin
state, and the axial zero-field splitting was fixed to correspond
to the magnetic susceptibility measurements. (In the
simulations, any value of |D| > 2 cm−1 was found to be
consistent with the experimental spectra.) The electronic g
values were constrained to be axial, (g1 = g2), such that the
effects of weak rhombic splitting could be assigned to finite
values of E/D (see Figure caption 11). The average of the g-
values obtained from the EPR fits is 2.2 for both 1 and 3, which
is reasonably close in value to the average g-values used in the
magnetic susceptibility fits.
The sextet spin state is the highest spin configuration of a
formally mixed-valent Co(II)Co(I) core. In the related high-
spin diiron complex Fe (DPhF) , the Mossbauer spectra
̈
2
3
supported a fully delocalized Fe(1.5)Fe(1.5) configuration. By
analogy, Co2(DPhF)3 should also have delocalized dicobalt
core, Co(1.5)Co(1.5).
We anticipated that the iron−cobalt complex 4 would remain
high spin because of the weak trigonal ligand field. The
magnetic susceptibility data for 4 shows a temperature-
independent magnetic moment of 6.73 μB (χ*T = 5.68 cm3
K mol−1), which is consistent with S = 3 (Figure 12). The spin-
Hamiltonian simulation yields a fitted g-value of 2.00, an axial
zero-field splitting parameter, D3 = 6 cm−1, and θw = 3 K to
model the weak intermolecular ferromagnetic coupling.
Complex 4 coincidentally has a similar μeff value as 1 (perhaps
because of the increased g value for the latter). But unlike 1, 4 is
expected to have integer spin, and indeed, it is EPR silent,
which is consistent with the spin assignment.
The EPR spectra of 1 and 3 have also been collected in
frozen toluene at 20 K. Both complexes show similar axial
derivative signals at effective g-values around geff = 6 and 2,
which are typical of transitions within the ms = 1/2 Kramers
doublet of S = 5/2 systems with significant axial zero-field
splitting (D ≫ hν at X-band, i.e., 0.3 cm−1) and small-to-
vanishing rhombicity, E/D ≈ 0 (Figure 11). Under this
condition, the 3/2 and 5/2 Kramers doublets are EPR-
silent, and the shape of the spectra is independent of the actual
value of the zero-field splitting parameter D. The spectra were
As a cautionary note, the magnetic parameters reported for 4
have limited reliability because of the mixed metal composition.
Also, we have found that the composition of the bulk material
J
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 12. Temperature dependence of the effective magnetic
moment, μeff, of 4 at 1 T, from 2 to 290 K. The solid lines represent
the best fit for Stot = 3 spin with g = 2 (χTIP = 1200 × 10−6 emu). For
technical reasons (see Methods) the septet was simulated as the spin
ground state of two ferromagnetically coupled spins S1,2 = 3/2 with a
large coupling constant, J = +200 cm−1. The value renders the excited
states at 1200 cm−1 such that thermal population of excited states is
negligible, in accord with experiment. Formally the approach
corresponds to the coupling of high-spin Fe(I) and Co(II), but such
an interpretation is not meaningful since the septet state is
energetically isolated and does not unveil its origin. The inset shows
the result of a multifield measurement recorded at 1, 4, and 7 T.
Figure 13. Zero-field Mossbauer spectrum of 4 at 80 K. The red solid
̈
line represents the best fit using two Lorentzian lines of equal intensity
but different widths. Fitting parameters: δ = 0.65 mm/s, ΔEQ = 0.64
mm/s, line width = 0.5 mm/s for the left line and 1.08 mm/s for the
right one.
orbitals of two metal centers can combine maximally to form 1
σ, 2 π, and 2 δ metal−metal bonds. For Co2(HNCHNH)3, the
electronic configuration of the sextet ground state,
(σ)2(π)4(π*)4(σ*)1(δ)2(δ*)2, indicated a Co−Co bond order
of 0.5. The electronic structures of the dicobalt cores in 1 and 3
is expected to be similar since an apical donor should only
slightly perturb the energies of the σ and σ * MOs.
On the other hand, the rarity of short Fe−Co bonds
prompted us to further investigate 4 using computational
methods, including DFT, CASSCF, and CASPT2 calculations.
The geometry of the full molecule was optimized at the DFT
level in several possible spin states, from S = 0 to S = 3. At all
three levels of theory, the septet was confirmed as the ground
spin state (Table 3). Also, the septet structure agrees quite well
can vary significantly from FeCoLPh. For a separate batch of 4
that was analyzed by inductively coupled plasma optical
emission spectrometry (ICP-OES), a composition consistent
with Fe0.79Co1.11LPh was determined. We suspect that the excess
of cobalt relative to iron arises from the dicobalt 3 impurity,
1
which was confirmed by H NMR spectroscopy.
To our knowledge, this is a unique spin state for an Fe−Co
bond; as mentioned above, virtually all known complexes with
an Fe−Co distance less than 3.02 Å (the sum of covalent radii
for high-spin Fe(II) and Co(II)) contain CO ligands, with most
of the remainder bearing strong-field NO or Cp-type ligands,
suggesting low-spin metal centers. A couple of exceptions are
an Fe(III)Co(III)Fe(III) complex, with a long Fe−Co distance
of 2.870(1) Å,73 found to have an S = 1 ground state, and a
salen-type Fe(II)−Co(II) complex with an S = 3/2 ground
state, from weak antiferromagnetic coupling across an even
longer M-M distance of 2.886(2) Å.74 Another rare exception is
a heterotrimetallic Fe2Co complex, wherein the [FeCo]5+
subunit, with a short Fe−Co distance of 2.29 Å, is proposed
to have a triplet subspin. Thus, 4 is a rare example of a high-
spin iron−cobalt bond.
Table 3. Calculated Relative Energies (kcal/mol) of 4 for All
Possible Spin States at DFT, CASSCF, and CASPT2 Levels
of Theory
DFT
CASSCF
CASPT2
Singlet
Triplet
Quintet
Septet
15.94
7.09
5.58
0
8.53
10.86
11.80
0
18.07
13.99
11.42
0
with the experimental structure (Supporting Information, Table
1). The only significant difference with respect to experiment is
the slight distortion from C3 symmetry.
Iron−cobalt 4 has also been characterized by Mossbauer
̈
spectroscopy as shown in Figure 13 (80 K, zero applied field).
The spectrum is broad and asymmetric, but a similar
phenomenon was previously observed with Fe2(DPhF)3. The
Because the wave functions of metal−metal bonds tend to be
multiconfigurational, CASSCF/CASPT2 calculations were used
to determine the electronic configuration of 4. The natural
orbitals comprising the valence 3d-electrons are shown in the
MO diagram in Figure 14. The MO diagram of 4, however,
lacks the δ and δ* MOs, which are replaced by localized dxy and
asymmetry in the Mossbauer doublet can be attributed to
̈
paramagnetic relaxation effects and/or the presence of a diiron
impurity. The isomer shift obtained from a provisional line fit, δ
= 0.65 mm/s, is identical to that found for Fe2(DPhF)3,
suggesting that the iron−cobalt complex is a fully delocalized
Fe(1.5)Co(1.5) core. The quadrupole splitting, ΔEQ = 0.64
mm/s, is slightly larger than that of Fe2(DPhF)3 (ΔEQ = 0.32
mm/s), reflecting the slightly larger electric field gradient upon
substituting iron with cobalt.
2
2
dx −y orbitals. The loss of the weak δ and δ* bonds is likely
because of the energetic mismatch between iron and cobalt 3d-
electrons. The ground-state wave function is dominated by one
main configuration (84.9%): (σ)2(π)4(π*)3(σ*)1(Co dxy)1(Co
1
1
1
2
2
2
2
d
x −y ) (Fe dxy) (Fe dx −y ) .
The main configuration predicts a formal Fe−Co bond order
Theoretical Calculations. Previously, the electronic
structure of the trigonal dicobalt compound was investigated
using SCF-Xα-SW and ab initio methods on the truncated
model, Co2(HNCHNH)3.25 In a trigonal ligand field, the d-
of 1, but a more accurate bond order can be calculated by
taking into consideration the contributing minor configurations.
For the total ground-state wave function, the natural orbital
occupation numbers become: (σ)1.88(π)3.92(π*)2.98(σ*)1.01(Co
K
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: ebill@gwdg.de (E.B.), gagliard@umn.edu (L.G.),
clu@umn.edu (C.C.L.).
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
Andrew Fielding, Andreas Gobels, and Bernd Mienert are
■
̈
acknowledged for assisting with the acquisition of spectroscopic
data. Professor Ted Betley is thanked for a helpful suggestion.
C.M.Z. and C.C.L. thank Prof. John Lipscomb and Prof. David
Blank for the use of their EPR and UV−vis-NIR instruments,
respectively. X-ray diffraction experiments were performed
using a crystal diffractometer acquired through an NSF-MRI
award (CHE-1229400). Computing support and resources
were provided by the Minnesota Supercomputing Institute, and
funding for this work was provided in part by the University of
Minnesota and the NSF (CHE-1254621). The computational
results (H.J.K., D.Z., and L.G.) are based on work supported by
the NSF (CHE-1212575). ChemMatCARS Sector 15 is
principally supported by the NSF/DOE (CHE-0822838).
Use of the Advanced Photon Source was supported by the
U.S. Department of Energy, Office of Science, Office of Basic
Energy Sciences, under Contract No. DE-AC02-06CH11357.
Figure 14. Qualitative MO diagram showing the natural orbitals for 4
that arise from CASSCF calculations. The dominating electronic
configuration (84.9%) is shown.
dxy)1.04(Co dx −y
)
1.03(Fe dxy)1.00(Fe dx −y
)
1.00(Co 4d)0.11(Co
2
2
2
2
4d)0.03, which gives an effective bond order of 0.905 that is
slightly lower than the formal bond order. Compared to
dicobalt, the metal−metal bond order of iron−cobalt formally
increases by half a π bond, which is consistent with the
contraction of the M-M bond distance that is experimentally
observed. On a related note, analysis of the orbital occupation
numbers can give insight into the distribution of the 14 valence
d-electrons between the iron and cobalt centers in 4. The d-
populations for iron and cobalt are 6.62 and 7.38, respectively
(Supporting Information,Table 2 ). Thus, the CAS calculations
support a highly delocalized Fe(1.5)Co(1.5) core.
REFERENCES
■
(1) Farmer, J. A.; Campbell, C. T. Science 2010, 329, 933.
(2) Blackburn, N. J.; Barr, M. E.; Woodruff, W. H.; van der Ooost, J.;
de Vries, S. Biochemistry 1994, 33, 10401.
(3) Lindahl, P. A. J. Inorg. Biochem. 2012, 106, 172.
(4) Nippe, M.; Goodman, S. M.; Fry, C. G.; Berry, J. F. J. Am. Chem.
Soc. 2011, 133, 2856.
(5) Kadis, J.; Shin, Y.-g. K.; Dulebohn, J. I.; Ward, D. L.; Nocera, D.
G. Inorg. Chem. 1996, 35, 811.
(6) Heyduk, A. F.; Macintosh, A. M.; Nocera, D. G. J. Am. Chem. Soc.
1999, 121, 5023.
(7) Dulebohn, J. I.; Ward, D. L.; Nocera, D. G. J. Am. Chem. Soc.
1990, 112, 2969.
(8) Multiple Bonds Between Metal Atoms, 3rd ed.; Cotton, F. A.,
Murillo, C. A., Walton, R. A., Eds.; Springer Science and Business
Media, Inc.: New York, 2005.
(9) Nguyen, T.; Sutton, A. D.; Brynda, M.; Fettinger, J. C.; Long, G.
J.; Power, P. P. Science 2005, 310, 844.
4. CONCLUSIONS
Since their initial report, dicobalt (1) and diiron trigonal
lanterns have been intriguing because of their short metal−
metal bond distance, unusual high-spin electronic configu-
rations, and mixed valency. Using a tris(amidinato)amine
ligand, the dicobalt analogue (3) and a rare, mixed iron−cobalt
complex (4) have been isolated. X-ray anomalous scattering
techniques have proven useful to quantify the level of metal
mixing in the heterobimetallic. While other bimetallics are
present, including the dicobalt and diiron species, the iron−
cobalt is the major component (89.3%). Collectively, the
bimetallic tris(amidinate) systems illustrate that adding an
apical amine donor or swapping one cobalt for iron (and vice
versa) does not change the high-spin nature of the electronic
ground state. Indeed, all the [M2]3+ cores herein are proposed
to be fully delocalized, that is, M(1.5)M(1.5). One significant
difference between dicobalt 3 to iron−cobalt 4 is the shortening
of the metal−metal bond distance from 2.29 to 2.18 Å,
respectively. The trend is consistent with a small gain in the
metal−metal bond order (formally half) upon removing one
antibonding d-electron from [Co2]3+ to obtain [FeCo]3+.
Finally, the two-step metalation strategy with the tris-
(amidinato)amine ligand is promising for gaining access to a
larger family of homo- and heterobimetallic complexes.
(10) Brynda, M.; Gagliardi, L.; Widmark, P.-O.; Power, P. P.; Roos,
B. O. Angew. Chem., Int. Ed. 2006, 45, 3804.
(11) La Macchia, G.; Aquilante, F.; Veryazov, V.; Roos, B. O.;
Gagliardi, L. Inorg. Chem. 2008, 47, 11455.
(12) La Macchia, G.; Li Manni, G.; Todorova, T. K.; Brynda, M.;
Aquilante, F.; Roos, B. O.; Gagliardi, L. Inorg. Chem. 2010, 49, 5216.
(13) Kreisel, K. A.; Yap, G. P. A.; Dmitrenko, O.; Landis, C. R.;
Theopold, K. H. J. Am. Chem. Soc. 2007, 129, 14162.
(14) Noor, A.; Wagner, F. R.; Kempe, R. Angew. Chem., Int. Ed. 2008,
47, 7246.
(15) Horvath, S.; Gorelsky, S. I.; Gambarotta, S.; Korobkov, I. Angew.
Chem., Int. Ed. 2008, 47, 9937.
(16) Tsai, Y.-C.; Hsu, C.-W.; Yu, J.-S. K.; Lee, G.-H.; Wang, Y.; Kuo,
T.-S. Angew. Chem., Int. Ed. 2008, 47, 7250.
(17) Hsu, C.-W.; Yu, J.-S. K.; Yen, C.-H.; Lee, G.-H.; Wang, Y.; Tsai,
Y.-C. Angew. Chem., Int. Ed. 2008, 47, 9933.
(18) Jones, C.; Schulten, C.; Rose, R. P.; Stasch, A.; Aldridge, S.;
Woodul, W. D.; Murray, K. S.; Moubaraki, B.; Brynda, M.; La Macchia,
G.; Gagliardi, L. Angew. Chem., Int. Ed. 2009, 48, 7406.
ASSOCIATED CONTENT
* Supporting Information
Crystallographic data in CIF format. Additional NMR
spectroscopic characterization and computational data. This
■
S
L
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(19) Cotton, F. A.; Daniels, L. M.; Maloney, D. J.; Matonic, J. H.;
Murillo, C. A. Inorg. Chim. Acta 1997, 256, 283.
(20) Nguyen, T.; Merrill, W. A.; Ni, C.; Lei, H.; Fettinger, J. C.; Ellis,
B. D.; Long, G. J.; Brynda, M.; Power, P. P. Angew. Chem., Int. Ed.
2008, 47, 9115.
(21) Fohlmeister, L.; Liu, S.; Schulten, C.; Moubaraki, B.; Stasch, A.;
Cashion, J. D.; Murray, K. S.; Gagliardi, L.; Jones, C. Angew. Chem., Int.
Ed. 2012, 51, 8294.
(22) Cotton, F. A.; Daniels, L. M.; Falvello, L. R.; Murillo, C. A.
Inorg. Chim. Acta 1994, 219, 7.
(23) Cotton, F. A.; Daniels, L. M.; Maloney, D. J.; Murillo, C. A.
Inorg. Chim. Acta 1996, 249, 9.
(24) Cotton, F. A.; Daniels, L. M.; Falvello, L. R.; Matonic, J. H.;
Murillo, C. A. Inorg. Chim. Acta 1997, 256, 269.
(25) Cotton, F. A.; Feng, X.; Murillo, C. A. Inorg. Chim. Acta 1997,
256, 303.
(26) Brynda, M.; Gagliardi, L.; Roos, B. O. Chem. Phys. Lett. 2009,
471, 1.
(27) Roos, B. O.; Borin, A. C.; Gagliardi, L. Angew. Chem., Int. Ed.
2007, 46, 1469.
(28) Zall, C. M.; Zherebetskyy, D.; Dzubak, A. L.; Bill, E.; Gagliardi,
(52) Poineau, F.; Forster, P. M.; Todorova, T. K.; Gagliardi, L.;
Sattelberger, A. P.; Czerwinski, K. R. Inorg. Chem. 2010, 49, 6646.
(53) Perdew, J. P.; Zunger, A. Phys. Rev. B 1981, 23, 5048.
(54) A development of the University of Karlsruhe and
Forschungszentrum Karlsruhe GmbH, 1989−2007, TURBOMOLE
GmbH, since 2007: available from http://www.turbomole.com.
(55) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Chem.
Phys. Lett. 1989, 162, 165.
(56) Aquilante, F.; De Vico, L.; Ferre,
Å.; Pedersen, T.; Pitonak, M.; Reiher, M.; Roos, B. O.; Serrano-
Andres, L.; Urban, M.; Veryazov, V.; Lindh, R. J. Comput. Chem. 2010,
́
N.; Ghigo, G.; Malmqvist, P.-
́
31, 224.
(57) Roos, B. O.; Lindh, R.; Malmqvist, P.-Å.; Veryazov, V.;
Widmark, P. O. J. Phys. Chem. A 2004, 108, 2851.
(58) Roos, B. O.; Lindh, R.; Malmqvist, P.-Å.; Veryazov, V.;
Widmark, P. O. J. Phys. Chem. A 2005, 109, 6575.
(59) Hess, B. A. Phys. Rev. A 1986, 33, 3742.
(60) Aquilante, F.; Pedersen, T. B.; Lindh, R. J. Chem. Phys. 2007,
126, 11.
(61) Aquilante, F.; Pedersen, T. B.; Lindh, R.; Roos, B. O.; De Meras,
A. S.; Koch, H. J. Chem. Phys. 2008, 129, 8.
(62) Aquilante, F.; Malmqvist, P.-Å.; Pedersen, T. B.; Ghosh, A.;
Roos, B. O. J. Chem. Theory Comput. 2008, 4, 694.
(63) Forsberg, N.; Malmqvist, P.-Å. Chem. Phys. Lett. 1997, 274, 196.
(64) Lucas, R. L.; Zart, M. K.; Murkerjee, J.; Sorrell, T. N.; Powell, D.
R.; Borovik, A. S. J. Am. Chem. Soc. 2006, 128, 15476.
(65) Ray, M.; Hammes, B. S.; Yap, G. P. A.; Rheingold, A. L.;
Borovik, A. S. Inorg. Chem. 1998, 37, 1527.
L.; Lu, C. C. Inorg. Chem. 2012, 51, 728.
(29) Fout, A. R.; Basuli, F.; Fan, H.; Tomaszewski, J.; Huffman, J. C.;
Baik, M.-H.; Mindiola, D. J. Angew. Chem., Int. Ed. 2006, 45, 3291.
(30) Cotton, F. A.; Daniels, L. M.; Feng, X.; Maloney, D. J.; Matonic,
J. H.; Murilio, C. A. Inorg. Chim. Acta 1997, 256, 291.
(31) Lewis, R. A.; Morochnik, S.; Chapovetsky, A.; Wu, G.; Hayton,
T. W. Angew. Chem., Int. Ed. 2012, 51, 12772.
(32) Mathialagan, R.; Kuppuswamy, S.; De Denko, A. T.; Bezpalko,
M. W.; Foxman, B. M.; Thomas, C. M. Inorg. Chem. 2013, 52, 701.
(33) Fout, A. R.; Xiao, D. J.; Zhao, Q.; Harris, T. D.; King, E. R.;
Eames, E. V.; Zheng, S.-L.; Betley, T. A. Inorg. Chem. 2012, 51, 10290.
(66) Jones, M. B.; MacBeth, C. E. Inorg. Chem. 2007, 46, 8117.
(67) Freedman, D. E.; Han, T. H.; Prodi, A.; Muller, P.; Huang, Q.-
̈
Z.; Chen, Y.-S.; Webb, S. M.; Lee, Y. S.; McQueen, T. M.; Nocera, D.
G. J. Am. Chem. Soc. 2010, 132, 16185.
(68) Betley, T. A., personal communication.
(69) Wulf, R. Acta Crystallogr., Sect. A: Found. Crystallogr. 1990, 46,
681.
(34) Cler
Wang, X. Inorg. Chem. 2000, 39, 3065.
(35) Clerac, R.; Cotton, F. A.; Dunbar, K. R.; Lu, T.; Murillo, C. A.;
́
ac, R.; Cotton, F. A.; Dunbar, K. R.; Lu, T.; Murillo, C. A.;
́
(70) Eames, E. V.; Hernan
2013, 52, 5006.
́ ́
dez Sanchez, R.; Betley, T. A. Inorg. Chem.
Wang, X. J. Am. Chem. Soc. 2000, 122, 2272.
(36) Berry, J. F.; Cotton, F. A.; Lu, T.; Murillo, C. A. Inorg. Chem.
2003, 42, 4425.
(37) Cooper, B. G.; Napoline, J. W.; Thomas, C. M. Catal. Rev. 2012,
54, 1.
(38) Kuppuswamy, S.; Cooper, B. G.; Bezpalko, M. W.; Foxman, B.
M.; Powers, T. M.; Thomas, C. M. Inorg. Chem. 2012, 51, 1866.
(39) Bailey, P. J.; Coxall, R. A.; Dick, C. M.; Fabre, S.; Henderson, L.
(71) Allen, F. Acta Crystallogr., Sect. B 2002, 58, 380.
(72) WanTao, Z.; Go-Shin; Syu-Tin, S. Russ. J. Inorg. Chem. 1997, 42,
1107.
(73) Glaser, T.; Beissel, T.; Bill, E.; Weyhermuller, T.; Schunemann,
̈
̈
V.; Meyer-Klaucke, W.; Trautwein, A. X.; Wieghardt, K. J. Am. Chem.
Soc. 1999, 121, 2193.
(74) Matsumoto, K.; Sekine, N.; Arimura, K.; Ohba, M.; Sakiyama,
H.; Okawa, H. Bull. Chem. Soc. Jpn. 2004, 77, 1343.
́
C.; Herber, C.; Liddle, S. T.; Lorono-Gonzalez, D.; Parkin, A.; Parsons,
̃
S. Chem.Eur. J. 2003, 9, 4820.
(40) Bergbreiter, D. E.; Killough, J. M. J. Am. Chem. Soc. 1978, 100,
2126.
(41) Spek, A. L. Acta Crystallogr. 2009, D65, 148.
(42) Toby, B. H.; Von Dreele, R. B. J. Appl. Crystallogr. 2013, 46,
544.
(43) O’Connor, C. J. In Progress in Inorganic Chemistry; Lippard, S. J.,
Ed.; John Wiley & Sons, Inc.: New York, 1982; Vol. 29, p 203.
(44) Weast, R. C.; Astle, M. J. CRC Handbook of Chemistry and
Physics; CRC Press Inc.: Boca Raton, FL, 1979.
(45) Kahn, O. Molecular Magnetism; VCH Publishers, Inc.:
Weinheim, Germany, 1993.
(46) Bill, E. julX; available from: http://ewww.mpi-muelheim.mpg.
de/bac/logins/bill/julX_en.php.
(47) Roos, B. O.; Taylor, P. R.; Siegbahn, P. E. M. Chem. Phys. 1980,
48, 157.
(48) Andersson, K.; Malmqvist, P.-Å.; Roos, B. O. J. Chem. Phys.
1992, 96, 1218.
(49) Briggs, B. N.; McMillin, D. R.; Todorova, T. K.; Gagliardi, L.;
Poineau, F.; Czerwinski, K. R.; Sattelberger, A. P. Dalton Trans. 2010,
39, 11322−11324.
(50) Ferrante, F.; Gagliardi, L.; Bursten, B. E.; Sattelberger, A. P.
Inorg. Chem. 2005, 44, 8476.
(51) Gagliardi, L.; Roos, B. O. Inorg. Chem. 2003, 42, 1599.
M
dx.doi.org/10.1021/ic400292g | Inorg. Chem. XXXX, XXX, XXX−XXX