A redox series of aluminum complexes: Characterization of four oxidation states including a ligand biradical state stabilized via exchange coupling

Article abstract of DOI:10.1021/ja2015718

Electrophilic activation and subsequent reduction of substrates is in general not possible because highly Lewis acidic metals lack access to multiple redox states. Herein, we demonstrate that transition metal-like redox processes and electronic structure and magnetic properties can be imparted to aluminum(III). Bis(iminopyridine) complexes containing neutral, monoanionic, and dianionic iminopyridine ligands (IP) have been characterized structurally and electronically; yellow (IP)AlCl₃ (1), deep green (IP ^-)₂AlCl (2) and (IP^-)₂Al(CF ₃SO₃) (3), and deep purple [(IP^2-)Al] ^- (5) are presented. The mixed-valent, monoradical complex (IP ^-)(IP^2-)Al is unstable toward C-C coupling, and [(IP ^2-)Al]^2-(μ-IP-IP)^2- (4) has been isolated. Variable-temperature magnetic susceptibility and EPR spectroscopy measurements indicate that the biradical character of the ligand-based triplet in 2 is stabilized by strong antiferromagnetic exchange coupling mediated by aluminum(III): J = -230 cm^-1 for H = -2J(S _L(1)?S_L(2)). Coordination geometry-dependent (IP^-)-(IP^-) communication through aluminum(III) is observed electrochemically. The cyclic voltammogram of trigonal bipyramidal 2 displays successive ligand-based oxidation events for the two IP^1-/0 processes, at -0.86 and -1.20 V vs SCE. The 0.34 V spacing between redox couples corresponds to a conproportionation constant of K_c = 10 ^5.8 for the process (IP^-)₂AlCl + (IP) ₂AlCl → 2(IP^-)(IP)AlCl consistent with Robin and Day Class II mixed-valent behavior. Tetrahedral 5 displays localized, Class I behavior as indicated by closely spaced redox couples. Furthermore, CV's of 2 and 5 indicate that changes in the coordination environment of the aluminum center shift the potentials for the IP^1-/0 and IP^2-/1- redox couples by up to 0.9 V.

Full text of DOI:10.1021/ja2015718

ARTICLE
pubs.acs.org/JACS
A Redox Series of Aluminum Complexes: Characterization of Four
Oxidation States Including a Ligand Biradical State Stabilized via
Exchange Coupling
Thomas W. Myers,† Nasrin Kazem,† Stefan Stoll, R. David Britt, Maheswaran Shanmugam, and
Louise A. Berben*
Department of Chemistry, University of California, Davis, California 95616, United States
S Supporting Information
b
ABSTRACT: Electrophilic activation and subsequent reduction of sub-
strates is in general not possible because highly Lewis acidic metals lack
access to multiple redox states. Herein, we demonstrate that transition
metal-like redox processes and electronic structure and magnetic proper-
ties can be imparted to aluminum(III). Bis(iminopyridine) complexes
containing neutral, monoanionic, and dianionic iminopyridine ligands (IP)
have been characterized structurally and electronically; yellow (IP)AlCl3
(1), deep green (IPꢀ)2AlCl (2) and (IPꢀ)2Al(CF3SO3) (3), and deep
purple [(IP2ꢀ)Al]ꢀ (5) are presented. The mixed-valent, monoradical
complex (IPꢀ)(IP2ꢀ)Al is unstable toward CꢀC coupling, and [(IP2ꢀ)Al]2ꢀ
(μ-IPꢀIP)2ꢀ (4) has been isolated. Variable-temperature magnetic susceptibility and EPR spectroscopy measurements indicate that the
biradical character of the ligand-based triplet in 2 is stabilized by strong antiferromagnetic exchange coupling mediated by aluminum(III):
J = ꢀ230 cmꢀ1 for H = ꢀ2J(SL(1)
L(2)). Coordination geometry-dependent (IPꢀ)ꢀ(IPꢀ) communication through aluminum(III) is
^
^
^
S
3
observed electrochemically. The cyclic voltammogram of trigonal bipyramidal 2 displays successive ligand-based oxidation events for the
two IP1ꢀ/0 processes, at ꢀ0.86 and ꢀ1.20 V vs SCE. The 0.34 V spacing between redox couples corresponds to a conproportionation
constant of Kc = 105.8 for the process (IPꢀ)2AlCl þ (IP)2AlCl f 2(IPꢀ)(IP)AlCl consistent with Robin and Day Class II mixed-valent
behavior. Tetrahedral 5 displays localized, Class I behavior as indicated by closely spaced redox couples. Furthermore, CV’s of 2 and 5
indicate that changes in the coordination environment of the aluminum center shift the potentials for the IP1ꢀ/0 and IP2ꢀ/1ꢀ redox
couples by up to 0.9 V.
’ INTRODUCTION
binding and reduction at lower energies relative to transition
metal-only catalyst systems.5ꢀ7 In another example, highly elec-
trophilic aluminum trichloride has been shown to coactivate CO2
to yield isolable species such as AlCl3 CO2 PMes3, for which the
A defining characteristic of transition metal reactivity is the
ability of these metal centers to effect the activation and
subsequent redox transformations of substrates.1 It is also known
that moderately Lewis acidic ions such as the alkali or alkaline
earth metal ions, late transition ions Cu(I/II) or Zn(II), or highly
electrophilic main group metal centers such as Al(III) can facilitate
substrate activation.2 However, following substrate activation, sub-
sequent redox chemistry is in general not possible without addition
of a sacrificial oxidant or reductant, and so the investigation of
redox activity, redox transformations, or electrocatalytic reactions
of substrates by Lewis acidic metal complexes is not well-known.
Herein, we use redox-active ligands to demonstrate that transition
metal-like redox processes and electronic structure and magnetic
properties can be imparted to aluminum(III) complexes.
Recent work in main group chemistry concerning, for exam-
ple, catalytic hydrogenation,3 or activation of ethylene,4 has
demonstrated that transition metal-like reactivity can be observed.
In the context of small molecule chemistry, some examples of Lewis
acid-assisted activation are known for CO2 activation. For example,
addition of alkali metal ions to systems employing redox-active
transition metal electrocatalysts has been shown to facilitate CO2
3
3
solid-state structure was characterized crystallographically.8 How-
ever, to effect redox transformations at the immobilized CO2
molecule, external redox reagents must be employed, and hence
only stoichiometric transformations are possible; for example, the
sacrificial hydride donor ammonia borane in combination with
water could be used to effect reduction of CO2 to MeOH from
AlCl3 CO2 PMes3.8
3
3
Highly Lewis acidic aluminum is the most abundant metallic
element on earth, comprising 7ꢀ8% of the earth’s crust and as
such is an appealing element for use in large-scale applications
including catalysis. Aluminum is commonly found only in the þ3
oxidation state, and so facile redox transformations of small
molecules are not possible. Aluminum can be rendered redox-active
by employing a very harsh reducing agent such as potassium
metal and a ligand, which enforces a low coordination number.
Received: February 28, 2011
Published: May 13, 2011
r
2011 American Chemical Society
8662
dx.doi.org/10.1021/ja2015718 J. Am. Chem. Soc. 2011, 133, 8662–8672
|

Journal of the American Chemical Society
ARTICLE
Chart 1. Structures of Ligands Mentioned in the Text with
Their Abbreviations
and we have isolated four of these, which have S = 0 and S = 1 spin
states at room temperature. We show that mixed-valent, doublet
oxidation states are unstable toward radical coupling of the
iminopyridine ligands and present evidence that stabilization of
the biradical, triplet oxidation state at room temperature is due to
strong electronic and magnetic interactions mediated by the
aluminum(III) center between unpaired electrons on iminopyr-
idine ligands. In fact, the ground state may be more accurately
referred to as a singlet biradical state. We have also found that
substitution at the aluminum center can shift the position of the
ligand redox potentials by up to 0.9 V. Taken together, these
results indicate that appropriately chosen redox-active ligand sets
can impart magnetic and electronic properties reminiscent of the
transition metals to complement the highly electrophilic proper-
ties of a main group element.
For instance, using low-valent aluminum synthons, Roesky and
co-workers have reported two-electron oxidation reactions in-
volving an Al(I) to Al(III) transformation.9,10 These impressive,
albeit isolated cases further illustrate the difficulty of using
aluminum for facile and tunable redox chemistry. To investigate
the possibility of facile redox transformations via electrophilic
activation at an aluminum metal center, we are exploring the
chemistry of aluminum(III) complexes with redox-active ligands:
in principle, an aluminum complex of the form (L)2AlX (L =
redox active ligand) could give access to five successive redox
states if each ligand L can access three oxidation states, for
example, L, Lꢀ, and L2ꢀ. In addition, by judicious choice of
redox-active ligand, these redox states could be accessed at a wide
range of potentials, some of which may be very mild and allow
careful control of reactivity.
The majority of paramagnetic and potentially redox-active
aluminum(III) complexes feature N-donor ligands (Chart 1),
and, in general, a detailed understanding of their structural
properties, redox chemistry, and electronic and magnetic proper-
ties has not been elucidated. The earliest reported, redox-
active and paramagnetic complex of aluminum(III) was Al(bpy)3
(bpy = 2,20-bipyridyl).11 On the basis of the magnetic moment of
the complex, it was suggested that it is best described as an Al3þ
cation, with the unpaired electrons extensively delocalized on the
bpy ligands as three radical anions. The reported magnetic
moment for Al(bpy)3, of 2.32 μB is lower than expected, for
three unpaired electrons, and suggests that some antiferro-
magnetic coupling through aluminum is present. Subsequent
work by Kaim and others resulted in the isolation of related
dialkyl Al(III) complexes of the pyrazine, pyridine, and bipyr-
idine radical anions.12,13 Raston and co-workers studied the
reactivity of alanes and aluminum metal with dbdab (dbdab =
1,4-di-tert-butyl-1,4-diazabutadiene) resulting in the isolation of
Al(dbdab)2 and partially hydrogenated Al(dbdab)(dbdabH2).14
The EPR spectrum of Al(dbdab)2 shows hyperfine coupling to
four equivalent nitrogen atoms but not to the aluminum nucleus,
and thus the radical is most likely ligand centered and delocalized
over both ligands on the EPR time scale. In a study reported by
Barron and co-workers roughly 20 years ago, Al(dpt)3 (where
dpt = 1,3-diphenyltriazenido) was isolated and characterized.15
Recently, substituted bis(imino)pyridine (denoted I2P) com-
plexes of aluminum(III) have been isolated. Richeson and co-
workers have isolated a neutral ligand complex,16 and Gambarotta
and co-workers reported a reduced-ligand complex.17
’ RESULTS AND DISCUSSION
Synthesis of Aluminum Complexes 1ꢀ6. Iminopyridine
ligands and derivatives thereof have been employed as ligands
for Mg2þ 18
,
with divalent ions of the first row transition series
from Cr2þ to Zn2þ 19 and recently as ancillary ligands for Fe2þ
,
complexes, which are active for catalysis of hydrosilylation
reactions.20 In the examples of paramagnetic aluminum com-
plexes highlighted above, the metal center is most often coordi-
natively saturated by the redox-active ligands and has the form
L3Al (where L = bidentate ligand). In the present study, we focus
on the (2,6-diisopropylphenyl)-substituted iminopyridine ligand
because we find that its moderate size prevents formation of the
L3Al moiety and favors formation of four- and five-coordinate
complexes of aluminum with open coordination sites: IP2Al or
IP2AlX (where IP = 2,6-bis(1-methylethyl)-N-(2-pyridinyl-
methylene)phenylamine, and X is a monodentate ligand). In
this work, all three oxidation states of the iminopyridine ligand
can be stabilized at one metal center, and the doubly reduced
ligand is isolated as a nonbridging ligand in a monomeric
complex. We attribute both of these results to the stability of
the bonding obtained by the interaction between highly charged
Al3þ and strongly donating iminopyridine.
Using AlCl3 as a starting reagent, control of the number of
chloro ligands at the aluminum center in a series of product
complexes was achieved by limiting the number of equivalents of
Na used as reducing agent. We found that removal of chloride ion
from the AlCl3 coordination sphere as NaCl was necessary to
drive the reactions and was dependent on the presence of Naþ
ions generated during ligand reduction by sodium metal. If no
sodium is added to the reaction of AlCl3 with neutral IP, the five-
coordinate complex IPAlCl3 (1) could be isolated. A summary of
the preparation of complexes 2ꢀ5 is described in Scheme 1. A
dark green aluminum(III) complex, (IPꢀ)2AlCl (2) in which the
two IP ligands are each reduced by one electron, was prepared by
combination of AlCl3 with 2 equiv of IP and 2 equiv of sodium in
1,2-dimethoxyethane (DME). The corresponding triflate-sub-
stituted complex (IPꢀ)2Al(CF3SO3) (3) could be prepared from
Al(CF3SO3)3 in an analogous manner.
Synthesis of a mixed-valent, overall three-electron per alumi-
num center reduced complex, (IPꢀ)(IP2ꢀ)Al, was attempted. In
this case, direct three-electron reduction of the constituent AlCl3
and 2 equiv of IP using sodium, or reduction of 2 by 1 equiv of
sodium metal, both resulted in a three-electron per aluminum
reduced product. However, under both sets of reaction condi-
tions, dimerization of the complexes through the ligand backbone
In this work, we have undertaken the task to isolate bis-
(iminopyridine) aluminum(III) complexes of the form (IPnꢀ)AlꢀX,
in a wide range of oxidation states, and to thoroughly characterize
the magnetic and electronic properties of these molecules. We
find that five oxidation states can be observed electrochemically,
8663
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of Aluminum Complexes 1ꢀ5
has occurred to yield [(IP2ꢀ)Al]2(IPꢀIP)2ꢀ (4), preventing the
formation of a monometallic aluminum complex. In 4, each
aluminum center is coordinated by a two-electron reduced and a
one-electron reduced ligand; the one electron reduced ligand,
IPꢀ, has undergone CꢀC coupling through the formerly imine
carbon of IP. In addition to the metrics from single crystal X-ray
studies (vide infra), this observed reactivity provides further
evidence for the radical nature of the ligand IPꢀ upon one-
electron reduction. It is interesting to note that for complexes 2
and 3, in which each of the ligands are reduced by one electron,
we have never observed evidence for dimerization via CꢀC
coupling of the ligand. This reactivity difference suggests that the
ligand radical of 2 and 3 is energetically stabilized with respect to
the ligand radical of the one-electron reduced ligand generated
during the synthesis of 4. We propose that the stabilization of the
ligand radicals in 2 and 3 is achieved by antiferromagnetic
coupling of the unpaired electrons through the aluminum center,
which reduces the effective unpaired spin density residing at the
carbon atom of the imine functional group. Electrochemical and
magnetic susceptibility evidence for coupling between the un-
paired electrons is discussed below.
Preparation of a deep purple, overall four-electron reduced
complex, [(DME)3Na][(IP2ꢀ)2Al], 5, in which two ligands are
each reduced by two electrons was achieved by increasing the
ratio of IP to Na up to 1:2.5. In this case, use of DME as a solvent
was necessary to chelate the sodium countercation and prevent
coordination of sodium to one of the anionic pyridine rings.
When the reaction is performed in diethyl ether instead of DME,
the sodium countercation was observed coordinated to a pyr-
idine ring and two diethyl ether molecules in the solid-state
structure: [(Et2O)2Na][(IP2ꢀ)2Al], 6.
indicative of formation of a diamagnetic complex: peaks in the
aromatic region are shifted as compared to the free ligand but
remain sharp and consistent with chemical shifts located as
expected. The NMR spectra of 2 and 3 display peaks broadened
and shifted from the expected positions, which implies that in
each case the complexes are paramagnetic at room temperature.
Complexes 2 and 3 differ from each other only in the identiꢀty
of the fifth ligand at the aluminum center, Clꢀ or CF3SO3
,
respectively. Reflecting this similarity, many of the observed
proton resonances have similar chemical shifts. For example, the
proton resonances for the pyridine rings are observed at 8.6, 8.4,
and 8.2 ( 0.1 ppm in each case. For compounds 4ꢀ6, the NMR
spectra provide further evidence that the reduction events involved
in preparation of the complexes are localized on the ligands
rather than the metal; as expected for a two-electron reduction
per ligand, the observed chemical shifts corresponding to proton
resonances on the pyridine rings do not fall within the expected
range of aromatic proton resonances. For example, in 5, four
resonances fortheformerlypyridineprotonsareobservedintherange
5.91ꢀ4.65 ppm, as compared to the range of 8.16ꢀ7.63 ppm for the
aromatic pyridine protons of the neutral ligand in 1.
Consistent with the intense deep green and purple colors of
2ꢀ6, the electronic absorption spectra obtained in hexanes
solution display a series of intense absorption bands (Figures 1
and S1). The electronic configuration of Al(III) is [Ne], and so
we attribute each of these absorption bands to ligand-based
πꢀπ* transitions. Similar assignments have been made for
previously studied Zn2þ complexes of iminopyridine ligands.19
The neutral free ligand, IP, displays one intense absorption at
370 nm (1500 L molꢀ1 cmꢀ1, Figure S1): this band increases in
intensity to 2000 L molꢀ1 cmꢀ1, and shoulders extend further
into the visible region as 1, followed by 2, equiv of sodium metal
are added to effect reduction of IP. Upon coordination of the
ligand to an aluminum center, this band increases significantly in
intensity. In complexes 2ꢀ5, a band at 370 nm is observed, which
Compounds 1ꢀ6 are soluble in alkanes, aromatic, and ether
solvents. The most reduced complexes, 5 and 6, decompose in
the presence of MeCN or dichloromethane. The NMR spectra of
the neutral and four-electron reduced complexes 1, 5, and 6 are
8664
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
decreases in intensity (ca. 38 700 down to 5500 L molꢀ1 cmꢀ1
)
was grown by chilling concentrated solutions at ꢀ30 °C for
about 1 week. Complexes 3, 4, and 5 were crystallized from
DME, and complexes 2 and 6 were crystallized from diethyl ether
(Table 1). The complexes are four or five coordinate and contain
two iminopyridine ligands (Tables 2 and 3): the five-coordinate
complexes contain an additional monodentate ligand. The five-
coordinate complexes (1, 2, and 3) are best described as distorted
trigonal pyramidal in geometry: τ values are all close to 1
(0.7151ꢀ0.8250) (Figures 2, S2, and S4).21 For complex 1,
the AlꢀNpy and AlꢀCl(3) bond vectors lie along the axial
direction of the molecule, and the NpyꢀAlꢀCl(3) bond angle
as the complex is successively reduced in one-electron steps
starting from one electron reduced per ligand in 2, down to two
electrons per ligand in 5. Further evidence for a ligand-based
assignment is that the absorption spectra of complexes 2 and 3 in
which the fifth coordination site on aluminum is occupied by a
chloro or triflate, respectively, are almost identical to each other
(Figure 1). In addition to the absorption at 370 nm, the dark
green complexes, 2 and 3, in which the ligands are each one-
electron reduced, exhibit a broad and less intense transition
(5000 L molꢀ1 cmꢀ1) at 700 nm attributed to πꢀπ* transitions.
Solid-State Structures of 1ꢀ6. Single crystals of 1 were
grown by diffusion of ether into MeCN. Each of the complexes
0
is 168.19(13)°. In 2 and 3, the AlꢀNpy and AlꢀNpy bond
0
vectors lie along the axial direction, and the NpyꢀAlꢀNpy bond
angles are 169.59(6)° and 176.49(5)° for 2 and 3, respectively.
0
The AlꢀNim, AlꢀNim , and AlꢀX bonds (X = Cl, triflate in 2 and
3, respectively) lie in the equatorial trigonal plane. The four-
coordinate complexes, 4ꢀ6 are pseudo tetrahedral: the
NimꢀAlꢀNpy bite angles of each ligand are pinched as compared
to perfect tetrahedral angles and range from 84.97(12)° for the
one-electron reduced ligand in 3 to 89.47(1)°, 87.87(17)°, and
88.57(9)° for the two-electron reduced ligands in 4ꢀ6(Figures 3, 4,
and S4). Complementing these pinched angles, the interligand
NꢀAlꢀN bond angles are more obtuse, ranging from 110.94(9)°
in 6 to 134.47(13)° in 4.
The bond distances and angles in the neutral ligand in complex
1 are unremarkable and serve as a reliable benchmark for
comparison with the bond lengths and angles in the one- and
two-electron reduced ligands in complexes 2ꢀ6 (Figure 5). In 2
and 3, the bond lengths and angles of the iminopyridine ligands
are consistent with reduction by one electron, and some general
trends in bond lengths and angles associated with this reduction
were observed. Metric parameters for 2 and 3 show the same
general trends, and so only a detailed discussion of 2 is presented
Figure 1. UVꢀvisible absorption spectra of complexes 2 (black), 3 (red),
and 5 (blue).
Table 1. Crystallographic Dataa for the Complexes (IP)AlCl3 DME (1), (IPꢀ)2AlCl (2), (IPꢀ)2Al(CF3SO3) DME (3),
3
3
[(IP2ꢀ)2Al2](μ-IPꢀIPꢀ) 2C6H6 (4), [(DME)3Na][(IP2ꢀ)2Al] 0.5DME (5), and [(Et2O)2Na] (IP2ꢀ)2Al (6)
3
3
1
2
3
4
5
6
formula
C
22H32N2O2Cl3
C36H44N4AlCl
C41H54N4AlF3O5S C84H100N8Al
C50H75.25N4O7AlNa C44H64N4O2AlNa
crystal size
formula weight, g molꢀ1 489.83
0.15 ꢁ 0.10 ꢁ 0.08 0.17 ꢁ 0.12 ꢁ 0.08 0.45 ꢁ 0.43 ꢁ 0.41 0.35 ꢁ 0.29 ꢁ 0.17 0.10 ꢁ 0.08 ꢁ 0.06
0.29 ꢁ 0.26 ꢁ 0.19
730.96
595.18
798.92
1275.68
894.36
space group
P21/n
P21/c
P21/c
P 1
P21/c
P21/c
a, Å
17.558(3)
9.835(18)
29.527(5)
90
15.796(3)
9.918(2
13.568(5)
18.277(7)
17.138(7)
90
14.273(9)
15.308(12)
19.838(12)
108.478(10)
99.313(10)
108.466(10)
3730.1(4)
2
13.862(5)
16.958(4)
22.251(8)
90
12.543(7)
14.039(8)
24.0531(14)
90
b, Å
c, Å
21.960(4)
90
R, deg
β, deg
90.172(6)
90
110.01(3)
90
102.473(7)
90
90.59(6)
90
97.574(10)
90
γ, deg
V, Å3
5098.5(16)
8
3232.8(11)
4
4149(3)
4
5231(3)
4
4198.3(4)
4
Z
T, K
F, calcd, g cmꢀ3
100 (2)
1.276
100 (2)
100 (2)
100 (2)
100 (2)
1.136
100 (2)
1.223
1.279
1.136
1.156
reflns collected/2θmax
47 377/52.04
80 818/58.88
8942/6893
387/0
47 376/61.28
12 082/9106
506/0
41 246/50.06
13 170/8337
787/0
33 537/42.98
5992/4522
618/108
0.71073/0.098
0.0739/1.019
0.1953
49 333/56.56
10 407/6836
477/0
unique reflns/ I > 2σ(I) 10 028/7208
parameters/restraints
λ, Å/μ (KR), cmꢀ1
R1/GOFb
wR2 (I > 2σ(I))b
residual density, e Åꢀ3
541/0
0.71073/0.415
0.0641/1.060
0.1685
0.71073/0.177
0.0487/1.042
0.1221
0.71073/0.160
0.0405/1.052
01412
0.71073/0.088
0.0687/1.219
0.2080
0.71073/0.099
0.0628/0.940
0.1912
þ1.089/ꢀ0.353
þ0.453/ꢀ0.274
þ0.452/ꢀ0.381
þ0.704/ꢀ0.436
þ1.550/ꢀ0.832
þ1.287/ꢀ0.511
a Obtained with graphite-monochromated Mo KR (λ = 0.71073 Å) radiation. b R1 = ∑ Fo| ꢀ Fc /∑|Fo|, wR2 = {∑[w(Fo ꢀ Fc2)2]/∑[w(Fo ) ]}1/2
.
2
2 2
8665
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Table 2. Selected Average Interatomic Distances (Å) for the Complexes in (IP)AlCl3 DME (1), (IPꢀ)2AlCl (2),
3
(IPꢀ)2Al(CF3SO3) DME (3), [(IP2ꢀ)2Al2](μ-IPꢀIPꢀ) 2C6H6 (4), [(DME)3Na][(IP2ꢀ)2Al] 0.5DME (5), and
3
3
3
[(Et2O)2Na](IP2ꢀ)2Al (6)
AlꢀNim
AlꢀNpy
AlꢀXa
CimꢀNim
CimꢀCpy
CpyꢀNpy
CpyꢀC(3)
1
2.038(4)
1.915(14)
2.084(4)
2.009(15)
2.1847(19)
2.191(1)
1.857(12)
na
1.284(6)
1.354(2)
1.375(17)
1.453(7)
1.405(2)
1.347(6)
1.366(2)
1.358(16)
1.391(6)
1.412(2)
2
3
1.893(12)
1.824(3),b 1.833(2)c
1.981(13)
1.851(3),b 1.937(2)c
1.427(17)
1.344(5),b 1.520(4)c
1.405(18)
1.412(2),b 1.384(4)c
4
1.411(4),b 1.478 (4)c
1.401(4),b 1.350(4)c
5
1.844(4)
1.873(4)
na
1.414(6)
1.356(6)
1.399(6)
1.437(6)
6d
1.852(2)
1.858(2)
na
1.413(3)
1.355(3)
1.401(3)
1.442(3)
a X = Cl or O. b Two-electron reduced L. c One-electron reduced L. d Data not included for pyridine ring, which is coordinated to Naþ.
Table 3. Selected Average Angles (deg) for the Complexes in (IP)AlCl3 DME (1), (IPꢀ)2AlCl (2), (IPꢀ)2Al(CF3SO3) DME (3),
3
3
[(IP2ꢀ)2Al2](μ-IPꢀIPꢀ) 2C6H6 (4), [(DME)3Na][(IP2ꢀ)2Al] 0.5DME (5), and [(Et2O)2Na] (IP2ꢀ)2Al (6)
3
3
τa
NimꢀAlꢀNpy
NpyꢀAlꢀNim
NimꢀAlꢀNim
NpyꢀAlꢀNpy
0
0
0
1
0.8250
0.8180
0.8175
77.39(16)
81.46(6)
n/a
n/a
n/a
2
93.85(6), 93.61(6)
94.76(5), 96.50(5)
125.28(6)
127.44(5)
169.59(6)
176.49(5)
3
82.87(5)
4
84.97(12),c 89.47(11)b
134.47(13), 123.99(14), 131.27(12), 112.24(12)
125.87(18), 124.14(17)
123.55(12), 129.99(13)
122.26(18)
103.86(13), 105.96(11)
112.50(18)
5
87.87(17)
6d
88.57(9)
126.49(10), 129.39(9)
118.69(9)
110.94(9)
XꢀAlꢀNim
XꢀAlꢀNpy
ClꢀAlꢀCl
CpyꢀCimꢀNim
1
118.12(13), 120.34(13)
89.45(13), 86.38(12)
168.19(13)
95.21(5), 95.20(5)
89.55(6), 93.84(6)
n/a
118.70(8), 99.33(7)
117.7(4)
91.36(12)
99.16(7)
n/a
2
3
4
5
6
117.87(5), 116.85(5)
116.73(14)
115.11(11)
118.1(3),b 107.1(2)c
115.88(5), 116.59(5)
n/a
n/a
n/a
n/a
n/a
n/a
n/a
117.1(4)
n/a
n/a
117.1(2)
a τ values calculated for five-coordinate complexes. b Ligand reduced by two electrons. c Ligand reduced by one electron. d Data not included for pyridine
ring, which is coordinated to Naþ.
here. In the aluminum(III) coordination sphere of 2, the
0
NimꢀAl(1)ꢀNim bite angle has widened upon reduction from
77.47(16)° in 1 to 87.59(6)° in 2. Consistent with partial
localization of the negative charge on the imine nitrogen donor,
shortening of the AlꢀNim bond, as compared to the bond in
complex 1, by 0.185 Å to 1.913(1) Å is observed. Shortening of
the AlꢀNim bond is accompanied by other bond length changes
within the ligand framework consistent with reduction localized
on the imine functionality: the formerly CdN double bond
lengthens from 1.284(6) to 1.352(2) Å, and the carbon-to-
pyridine single bond, C(1)ꢀC(2) in Figure 2, shortens to
1.405(6) Å. The pyridine ring largely retains aromatic character,
and the bond lengths and angles deviate only very slightly from
those in 1.
In the ligand-based mixed-valent complex 4, two aluminum
(III) centers each have a local pseudotetrahedral coordination
environment and are bridged by iminopyridine ligands that have
dimerized. A CꢀC bond is formed by coupling of the ligand
radicals formed when IP is reduced by one electron (Scheme 1).
Comparison of the one-electron reduced and dimerized ligand
in 4 with the one-electron reduced ligands in 2 reveals that
dimerization further exaggerates the bond length changes asso-
ciated with the imine functional group upon one-electron
Figure 2. Structure of (IPꢀ)2AlCl, 2, in which the iminopyridine
ligands are reduced by one electron each. Red, blue, and green
ellipsoids represent Al, N, and Cl atoms, respectively; ellipsoids are
shown at the 40% probability level, and H atoms are omitted for
clarity.
8666
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Figure 3. Structure of [(IP2ꢀ)2Al](LꢀL), 4. Red and blue ellipsoids represent Al and N atoms, respectively; ellipsoids are shown at the 40% probability
level, and H atoms are omitted for clarity.
ligands. As mentioned above, the deviation from ideal tetrahedral
geometry in this structure results from pinched chelate bite
angles in the iminopyridine ligand. An almost 5° increase in bite
angle was observed on going from the neutral ligand in 1 to the
one-electron reduced ligand in 2. In the two-electron reduced
ligand in 5, we observe roughly a further 5° increase in the
NimꢀAlꢀNpy bite angle as compared to complex 2. AlꢀNim and
AlꢀNpy bond lengths contract by 0.225 and 0.258 Å, respec-
tively, as compared to the neutral ligand in 1, which is consistent
with the dianionic character of the two-electron reduced ligand.
Internal ligand bond lengths in 5 show features consistent with a
two-electron reduction per ligand. The formerly imine CꢀNim bond
lengthens further to 1.414(4) Å, and the formerly carbonꢀpyridine
bond C(2)ꢀC(6) shortens to 1.356(6) Å. The alternating bond
lengths in the formerly pyridine ring reflect a loss of aromaticity
upon two electron reduction; the bond lengths observed, listed
starting at the Npy atom, are 1.429(6), 1.437(6), 1.346(6),
1.440(6), 1.350(7), and 1.370(2) Å. In addition to the evidence
provided by the alternating bond lengths in the pyridine ring, the
anionic, and therefore nonaromatic character of the pyridine ring,
is verified by the solid-state structure of complex 6. Preparation of
6 was achieved by reaction and crystallization in the absence of the
chelating DME solvent, which was effective to separate the Naþ
Figure 4. Structure of the anionic complex [(IP2ꢀ)2Al]ꢀ in 5. Red and
countercation from the anionic complex in 5. With ether as the
blue ellipsoids represent Al and N atoms, respectively; ellipsoids are
shown at the 40% probability level, and H atoms are omitted for clarity.
solvent for preparation of 6, the sodium countercation is observed
bound to the face of one of the anionic, (formerly) pyridine rings
(Figure S4).
reduction; as expected, dimerization localizes the position of the
reduction site at the carbon of the imine functional group. The
formerly CdN double bond lengthens completely to a single
bond at 1.478(4) Å. Unlike in 2 and 3, the CimꢀCpy bond,
C(37)ꢀC(38) in 5, retains single bond character due to forma-
tion of the new CꢀC bond via radical coupling.
Further confirmation that the reduction events are localized at
the ligands rather than the metal comes from comparison with
the bond lengths reported for the handful of known iminopyr-
idine complexes (Table 4). For the first row transition metal
series reported by Wieghardt and co-workers, average distances
of the CꢀC and CꢀNimine bond lengths in the one-electron
reduced ligand are 1.409(5) and 1.340(5) Å, respectively. In
complexes 2ꢀ4, we find averages of 1.410(8) and 1.350(6) Å,
respectively, which are in close agreement with those previously
reported. The CꢀNpy bond lengths are not noticeably affected
Single crystals of a tetrahedral aluminum complex in which
each ligand is reduced by two electrons could be isolated from
reactions performed in DME. The negative charge on complex 5
is balanced by a sodium countercation chelated by three DME
8667
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Figure 5. Comparison of bond lengths in complexes 1, 2, and 5; the ligand oxidation states are IP, IPꢀ, and IP2ꢀ, respectively.
on the basis of the solid-state X-ray diffraction data and EPR
spectroscopy measurements (vide infra), that the unpaired spins
are located ∼5 Å apart and are canted roughly 120° with respect to
each other. It seems unlikely that through-space coupling could be
responsible for the observed strong antiferromagnetic coupling, and
we propose that it is most likely a result of an aluminum-mediated
superexchange pathway through empty 3p orbitals on Al(III).
Strong superexchange through unfilled cyanide π* orbitals or filled
2p orbitals of the isoelectronic O2ꢀ ion is well-documented.25 We
are not aware of prior investigations in which a superexchange
pathway through Al(III) is proposed.
EPR spectroscopy measurements confirm the ligand-based
biradical, and the antiferromagnetic coupling at low temperature
proposed for 2 based on variable temperature magnetic suscept-
ibility measurements. The 100 K X-band continuous wave (cw)
EPR spectrum of a dilute frozen solution of 2 features a pattern
consisting of four lines between 310 and 360 mT characteristic
of a triplet state (S = 1) as well as a single line due to a doublet
species (S = 1/2) (Figure 7). Both signals are centered at 335 mT,
g = 2.004(1), typical for carbon/nitrogen-centered delocalized
organic radicals. The triplet nature of the four-line spectrum is
confirmed by the additional presence of a weak ΔmS = (2
transition at half field (Figure 7, inset). The zero-field splitting
tensor for the triplet state isnearly axial, and a least-squares analysis
gives |D| = 0.0212(2) cmꢀ1 and |E| = 0.0006(1) cmꢀ1. The
intensity of the triplet spectrum decreases with decreasing tem-
perature, confirming that the triplet is an excited state that is
depopulated upon cooling. No EPR signals grow in upon cooling,
which indicates that the ground state is an EPR-silent singlet. This
spectrum is attributed to two antiferromagnetically coupled spins
(S1 = S2 = 1/2), confirming the ligand biradical nature of 2. From
the temperature dependence of the triplet spectrum, the triplet
state is significantly depopulated at 100 K (FigureS6): over 90% of
spins are in the EPR-silent singlet ground state, and, consistent
with this prediction, we are not able to observe EPR signals below
60 K.26
Table 4. Comparison of Ligand Backbone Bond Lengths for
One-Electron19 and Two-Electron22 Reduced Ligands from
This Work with Values Reported in the Literaturea
literature (Å)
current work (Å)
IPꢀ
CꢀC
1.409(5)
1.340(5)
1.366(6)
1.459(4)
1.404(4)
1.410(8)
1.350(6)
1.356(6)
1.414(4)
1.429(4)
CꢀNim
CꢀC
IP2ꢀ
CꢀNim
CꢀNpy
a CꢀNpy bond lengths do not change upon one-electron reduction and
so were not included for this oxidation state.
by one-electron reduction of the ligand. For the two-electron
reduced ligand, we found only one other set of data reported in
the literature, and this corresponds to a complex in which the
originally imine N atom bridges two magnesium centers.22 In this
case, the bond lengths for the ligand backbone, CꢀNim,
CimꢀCpy, and CpyꢀNpy, were given as 1.366(6), 1.459(4), and
1.404(4) Å, respectively. These are in close agreement with the
average bond lengths of the two independent ligands we observe
in complex 5: 1.356(6), 1.414(4), and 1.429(4) Å, respectively.
Electronic Structure. Magnetic susceptibility measurements
were performed on complexes 2, 4, and 5 (Figures 6 and S5).
Measurements were performed on multiple batches of all
samples discussed and gave consistent results. Data were col-
lected in an applied field of 1 T between 4 and 300 K. As expected,
complexes 4 and 5 are diamagnetic. The observed magnetic
moment at 300 K for complex 2 is 2.7 μB, which is consistent with
the presence of an unpaired electron on each of the iminopyridine
ligands, and with the apparent paramagnetism implied by the proton
NMR spectrum of 2. Temperature-dependent susceptibility measure-
ments indicate that the magnetic moment falls steadily from 2.7 μB to
0.39 μB as the temperature is lowered from 300 to 4 K, consistent
with strong antiferromagnetic coupling between the ligand radi-
cals. The data were fit using MAGFIT3.0 and assuming a spin
The triplet zero-field splitting D is determined mostly by the
magnetic dipoleꢀdipole interaction between the two unpaired
spins. Assuming localized point dipoles for each spin and
neglecting all quantum mechanical contributions,27 a rough
estimate of the distance r between the two spins can be obtained
from D = (3/2)(μ0/4π)g2β2rꢀ3. The resulting distance r ≈ 5 Å is
consistent with a ligand biradical with separate delocalization of
each spin over each ligand, without noticeable interligand
delocalization. The value of D is similar to those obtained in
other biradicals.28,29 The triplet spectrum does not saturate
under the conditions employed (up to 10 mW, 60ꢀ140 K). In
contrast, the central line saturates very easily (half-saturation
^
^
^
HamiltonianoftheformH = ꢀ2JSL(1) SL(2) and a valueofg = 2.0;
3
a value for the exchange coupling of J = ꢀ230 cmꢀ1 was obta-
ined.23 A contribution from temperature-independent paramag-
netism (TIP) was also observed: 2 ꢁ 10ꢀ3 emu.
In a previous report by Heyduk and co-workers, the observed
increase in magnetic susceptibility with temperature for a ligand-
based biradical system was modeled using purely a significant
contribution from TIP: 945 ꢁ 10ꢀ6 emu.24 This interaction was
attributed to through-space coupling, rather than a metal-mediated
phenomenon; the unpaired electrons were located on coplanar aryl
rings situated only 3.4 Å apart. In the present example, we estimate,
8668
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Figure 6. Magnetic susceptibility data for compound 2 in an applied
field of 0.1 T. The “b” represent the experimental data, and the red “ꢀ”
^
represents a fit to the data for 2 using a spin Hamiltonian of the form H =
ꢀ2JSL(1) SL(2). Fit parameters: g = 2.0, J = ꢀ230 cmꢀ1, and TIP = 2 ꢁ
^
^
3
10ꢀ3 emu.
Figure 8. Cyclic voltammogram for a 1 mM solution of complex 2 (top)
and of complex 5 (bottom), recorded in 0.3 M Bu4NClO4 THF solution.
Figure 7. X-band cw EPR spectrum of a frozen solution of 1 mM 1 in
2:1 toluene:chloroform. Experimental parameters: microwave fre-
quency 9.389 GHz, power 10 mW, temperature 100 K, modulation
amplitude 1 mT. Simulation parameters: S = 1, g = 2.004(1), D =
0.0212(2) cmꢀ1, E = 0.0006(1) cmꢀ1, residual peak-to-peak broadening
2.0(1) MHz. The asterisk indicates the doublet signal.
power P1/2 < 10 μW at 140 K), which is characteristic of
monoradicals. In addition, the central line intensity is propor-
tional to 1/T and supports the assignment to a doublet. Because
of the strong saturation of the doublet signal and the strong
depopulation of the triplet, it is not possible to reliably quantify
the relative concentration of the two species. Moreover, the
amount of monoradical varies between EPR sample preparations
from the same batch of solid sample. We attribute the mono-
radical signal to reaction of 2 with trace amounts of dioxygen
introduced during the preparation of the EPR samples and the
high sensitivity of 2 toward water and dioxygen.
Figure 9. Cyclic voltammogram for a 1 mM solution of complex 4
recorded in 0.3 M Bu4NClO4 THF solution.
Electrochemical Measurements. Cyclic voltammetry mea-
surements were obtained in 0.3 M Bu4NClO4 THF solutions of
complexes 2ꢀ5 (Figures 8, 9, and S6). In most cases, the
complexes gave rise to a series of four reversible redox couples
corresponding to all of the expected ligand redox events, two
events for each of the two ligands (Table 5).
8669
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Table 5. Electrochemical Potentials (V vs SCE) for Ligand Reduction Events Observed in Complexes 2 and 5
E°LL/LLꢀ
E°LLꢀ/LꢀLꢀ
ΔE°
E°LꢀLꢀ/LꢀL2ꢀ
E°LꢀL2ꢀ/L2ꢀL2ꢀ
ΔE°
b
2
5
(IPꢀ)2AlCl
[(IP2ꢀ)2Al]ꢀ
ꢀ0.86
ꢀ0.4a
ꢀ1.20
0.34
ꢀ2.51a
ꢀ1.17
∼0
0.19
b
∼0
ꢀ1.36
a Irreversible redox event. b One-electron processes overlap, and so only one potential is reported.
For complex 2, two successive one-electron oxidation events
corresponding to the IP0/1ꢀ couple on the different ligands are
observed at E1/2 = ꢀ0.86 and ꢀ1.20 V vs SCE (Figure 8, top).
The 0.34 V distance between these waves indicates that electro-
nic coupling between the two ligands and most likely through the
Al(III) center is occurring; the 0.34 V difference corresponds to a
comproportionation constant, Kc = 105.8, for the process
(IPꢀ)2AlCl þ [(IP)2AlCl]2þ f 2[(IPꢀ)(IP)AlCl]þ. This value
is similar to those usually associated with class II mixed-valent
compounds.30 In addition, two overlapping redox couples for the
two IP1ꢀ/2ꢀ events are observed at ꢀ2.21 V. The higher current
observed in the reducing direction for this process is attributed to
additional, less well-defined and irreversible ligand reduction
events: we observe these irreversible reduction events between
ꢀ2 and ꢀ3 V for other complexes in the (IPnꢀ)2Al series even
when the well-defined, reversible IP1ꢀ/2ꢀ couple is well-sepa-
rated from them and observed at more positive potentials (for
example, Figure 8, bottom). It is unlikely that these additional
irreversible reduction events correspond to redox activity asso-
ciated with the Al(III) center.
The redox chemistry of the mixed-valent complex 4 is
somewhat less well-defined (Figure 9). Two redox couples,
corresponding to successive one-electron and two-electron
oxidation processes, are observed at ꢀ1.37 and ꢀ0.57 V.
Presumably the first at ꢀ1.37 V involves the IP1ꢀ/2ꢀ couple of
the IP2ꢀ ligand, and the second involves concomitant IP0/1ꢀ
couples on both of the ligands. This second couple at ꢀ0.57 V
is barely reversible. At further negative potentials (ꢀ2.20 V),
we observed a multielectron process, which presumably
corresponds again to unidentified ligand-based, rather than
metal-centered events.
It is informative to compare the voltammograms for com-
plexes 2 and 5 (Figure 8, Table 5). Complex 5 displays a similar
pattern of ligand-based redox waves as described above for
complex 2. Starting at the rest potential of ꢀ1.8 V, successive
one-electron oxidation waves for the IP1ꢀ/2ꢀ couple on each
ligand are observed at ꢀ1.36 and ꢀ1.17 V. Subsequently, a two-
electron oxidation wave for the concurrent oxidation of both
ligands, IP0/1ꢀ, is observed at ꢀ0.4 V vs SCE. This concurrent
oxidation is in contrast to complex 2 for which Kc for the
successive ligand-based processes for IP0/1ꢀ was 105.8. One
possible explanation for this apparently contradictory result is
the different coordination numbers and geometries of complexes
2 and 5, which may lead to different strengths of electronic
communication between the two IP ligands. Of further interest in
comparing the voltammograms of 2 and 5, the redox potentials
for the IP1ꢀ/2ꢀ and IP0/1ꢀ processes in 5, respectively, are each
approximately 0.9 and 0.7 V positive of the corresponding
redox couple for 2. This shift in the redox potentials of the
redox-active ligands upon removal of chloride from the
coordination sphere of aluminum seems quite significant
and indicates that electronic effects associated with the ligand
coordination environment are transferred via the aluminum
center to the iminopyridine ligands.
’ CONCLUSION AND OUTLOOK
We have demonstrated that redox noninnocent ligands can be
employed to impart a rich redox reactivity and open shell
electronic structure to the nonredox-active and strongly Lewis
acidic aluminum(III) ion. Complexes of aluminum containing
the neutral, monoanionic, and dianionic iminopyridine ligand
have been characterized structurally and electronically. Strong
antiferromagnetic coupling of ligand electrons through the
aluminum center stabilizes the biradical complex (IPꢀ)2AlCl, 2,
toward the radical ligand CꢀC coupling, which we observe during
attempts to prepare mixed-valent doublet complexes. Moreover, we
have demonstrated that electrochemical potentials are sensitive to
changes in the aluminum coordination environment. Future work
will focus on exploiting the unique properties obtained from
the combination of ligand redox activity and an electrophilic
aluminum(III) center, toward chemical transformations.
’ EXPERIMENTAL SECTION
Physical Measurements. Elemental analyses were performed by
Columbia Analytical. 1H NMR spectra were recorded at ambient
temperature using a Varian 300 MHz spectrometer. Chemical shifts
were referenced to residual solvent. Assignment of peaks in the NMR
spectra of paramagnetic complexes was made wherever possible. For the
most part, assignments are not given in cases where the origin of the peak
was uncertain due to the paramagnetic nature of the complexes. The
abbreviation pyred indicates reduced (and dearomatized) pyridine rings.
Electrochemical measurements were recorded in a glovebox under a
dinitrogen atmosphere using a CH Instruments Electrochemical Analy-
zer, a glassy carbon working electrode, a platinum wire auxiliary
electrode, and an Ag/AgNO3 nonaqueous reference electrode. Reported
potentials are all referenced to the SCE couple and were determined
using decamethylferrocene as an internal standard. The number of
electrons passed in a given redox process was estimated by comparison
of the peak current with the peak current of decamethylferrocene
included as an internal standard. Magnetic measurements were recorded
using a Quantum Designs MPMS XL magnetometer at 0.1 T. The
sample was contained under nitrogen in a gelcap and suspended in the
magnetometer in a plastic straw. The magnetic susceptibility was
adjusted for diamagnetic contributions using the constitutive corrections
of Pascal’s constants. EPR measurements were performed on 100 μL
dilute solutions of the compound loaded into 4 mm OD quartz tubes in
the glovebox and then freezeꢀpumpꢀthawed and flame-sealed on a
Schlenk line. X-band continuous-wave EPR measurements were per-
formed at the CalEPR center at UC Davis, with a Bruker ECS106
X-band spectrometer equipped with a Bruker SHQ resonator, an EIP
548A frequency counter, and an Oxford liquid-helium cryostat. The
magnetic field was calibrated with a Bruker ER036TM teslameter.
Simulations were performed with EasySpin.31
X-ray Structure Determinations. X-ray diffraction studies
were carried out on a Bruker SMART 1000, a Bruker SMART APEXII,
and a Bruker SMART APEX Duo diffractometer equipped with a
CCD detector.32a Measurements were carried out at ꢀ175 °C using
Mo KR (λ = 0.71073 Å) radiation. Crystals were mounted on a glass
capillary or Kaptan Loop with Paratone-N oil. Initial lattice parameters
8670
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
were obtained from a least-squares analysis of more than 100 cen-
tered reflections; these parameters were later refined against all data.
Data were integrated and corrected for Lorentz polarization effects
using SAINT32b and were corrected for absorption effects using
SADABS2.3.32c
deep red solution was added, dropwise, to a stirred suspension of AlCl3
(0.25 g, 1.88 mmol) in DME (10 mL), and the mixture was stirred
vigorously for 36 h to afford a purple suspension. The purple precipitate
was collected, extracted into ether (4 ꢁ 20 mL), filtered through Celite,
and the solvent was removed in vacuo to afford a purple powder. A
further amount of product was obtained by cooling the concentrated
reaction filtrate (5 mL) at ꢀ25 °C. Yield: 0.56 g (55%). Crystals suitable
for X-ray diffraction were obtained by cooling a concentrated ether
solution of 4 at ꢀ25 °C for 1 week. 1H NMR (300 MHz, C6D6): δ 8.60
(s, 2H, py), 8.47 (2, J = 4.56, 2H, py), 8.28 (d, J = 7.8, 2H, py), 7.42
(s, 2H, py), 6.91 (d, J = 8.9, 8H, Ph), 6.63 (t, J = 6.1, 4H, Ph), 6.14 (dd,
J = 9.1, 4.8, 2H, pyred), 6.02 (d, J = 8.35 2H, pyred), 5.51 (dd, J = 8.78,
2H, pyred), 5.39 (s, 2H, imCH), 5.16 (d, J = 8.35, 2H, pyred), 4.46 (d,
J = 6.6, 4H, CH(CH3)2), 4.19 (d, J = 7.35, 2H, (bridging imCH),
Space group assignments were based upon systematic absences, E
statistics, and successful refinement of the structures. Structures were solved
by direct methods with the aid of successive difference Fourier maps and
were refined against all data using the SHELXTL 6.2 software package.32d
Thermal parameters for all non-hydrogen atoms were refined anisotropi-
cally. Hydrogen atoms, where added, were assigned to ideal positions and
refined using a riding model with an isotropic thermal parameter 1.2 times
that of the attached carbon atom (1.5 times for methyl hydrogens).
Preparation of Compounds. All manipulations were carried out
using standard Schlenk or glovebox techniques under a dinitrogen
atmosphere. Unless otherwise noted, solvents were deoxygenated and
dried by thorough sparging with Ar gas followed by passage through an
activated alumina column. Deuterated solvents were purchased from
Cambridge Isotopes Laboratories, Inc. and were degassed and stored
over activated 3 Å molecular sieves prior to use. The compound 2,6-bis-
(1-methylethyl)-N-(2-pyridinylmethylene)phenylamine22 (abbreviated as
IP) was prepared according to literature procedures. All other reagents were
purchased from commercial vendors and used without further purification.
[(IP)AlCl3] (1). AlCl3 (0.13 g, 1 mmol) and IP (0.27 g, 1 mmol) were
stirred in DME (5 mL) for 10 h. Hexanes (10 mL) was added to afford
an orange-yellow precipitate, which was collected and washed with 5 mL
of DME (0.29 g, 72%). Crystals suitable for X-ray diffraction were
2
3.53 (hept, J = 6.75, 4H, CH(CH3)2), 1.14 (d, J = 6.9, 24H, CH(CH3) )
ppm. Anal. Calcd for C72H88AlN8: C, 77.27; H, 7.92; N, 10.01. Found:
C, 76.97; H, 8.19; N, 9.97. UVꢀvis spectrum (hexanes) λmax (εM): 238
(6390), 356 (5300), 424 (1510) nm (L molꢀ1 cmꢀ1). This compound is
diamagnetic.
[(DME)3Na][(IP2ꢀ)2Al] (5). Sodium metal (0.173 g, 7.5 mmol) and IP
(1.0 g, 3.75 mmol) were stirred with DME (10 mL) for 1 h. The resulting
deep red solution was added, dropwise, to a stirred suspension of AlCl3
(0.25 g, 1.88 mmol) in DME (10 mL), and the mixture was stirred
vigorously for 24 h to afford a deep purple suspension. The solvent was
removed in vacuo, and the purple residue was extracted into ether (4 ꢁ
20 mL), filtered through Celite, and the solvent wasremoved in vacuo to
afford 5 as a deep purple powder. A further amount of product was
obtained by cooling the concentrated reaction filtrate (5 mL) at ꢀ25 °C.
Yield: 0.98 g (62%). Crystals suitable for X-ray diffraction were obtained
by cooling a concentrated ether solution of 5 at ꢀ25 °C for 1 week. 1H
NMR (300 MHz, C6D6): δ 7.24 (t, J = 5.05, 2H, Ph), 7.12 (d, J = 4.65 2H,
Ph), 6.76 (d, J = 6.37, 2H, Ph), 5.91 (d, J = 9.6, 2H, py), 5.49 (s, 2H, imCH),
5.32 (dd, J= 9.1, 4.56, 2H, py), 4.65 (t, J=5.5,2H, py), 4.20(hept,J=6.7,2H,
CH(CH3)2), 2.95 (br, 12H, DME), 2.92 (br, 18H, DME), 1.48 (d, J = 7.2,
1
obtained by diffusion of ether into a CH3CN solution of 1. H NMR
(300 MHz, C6D6): 8.16 (d, J = 8.06, 2H, py), 7.63 (dd, J = 7.64, 2.93, 2H,
py), 7.18 (d, J = 8.09, 2H, Ph), 6.90 (br, 1H, Ph), 5.21 (s, 1H, imCH),
3.12 (hep, J = 6.5, 1H, CH(CH3)2), 1.19 (d, J = 6.9, 6H, CH(CH3)2).
Anal. Calcd for C18H23AlCl3N2: C, 53.95; H, 5.79; N, 6.99. Found: C,
53.82; H, 5.65; N, 6.79. UVꢀvis spectrum (hexanes) λmax (εM): 238
(12 100), 348 (br, 2030) nm (L molꢀ1 cmꢀ1).
[(IPꢀ)2AlCl] (2). Sodium metal (0.088 g, 3.85 mmol) and IP (1.0 g,
3.75 mmol) were stirred with DME (10 mL) for 1 h. The resulting deep
red solution was added, dropwise, to a stirred suspension of AlCl3 (0.25
g, 1.88 mmol) in DME (10 mL), and the mixture was stirred vigorously
for 24 h to afford a dark green suspension. The dark green precipitate
was collected, extracted into ether (4 ꢁ 20 mL), filtered through Celite,
and the solvent was removed in vacuo to afford a dark green powder. A
further amount of product was obtained by cooling the concentrated
reaction filtrate (5 mL) at ꢀ25 °C. Yield: 0.84 g (76%) of 2. Crystals
suitable for X-ray diffraction were obtained by cooling a concentrated
ether solution of 2 at ꢀ25 °C for 1 week. 1H NMR (300 MHz, C6D6): 8.60
(br, py), 8.47 (d, J = 4.33, py), 8.29 (br, py), 7.40 (br, py), 7.05ꢀ6.63 (m),
5.21 (d, J = 8.64), 4.51 (d, J = 4.68, 2H, imCH), 4.24 (d, J = 4.53), 3.71 (m,
2H, CH(CH3)), 1.24 (d, J = 6.59, 12H, CH(CH3)2). Anal. Calcd for
C36H44AlClN4: C, 72.64; H, 7.45; N, 9.41. Found: C, 72.10; H, 7.55; N,
9.32. UVꢀvis spectrum (hexanes) λmax (εM): 242 (20 647), 358 (20 875),
706 (1818) nm (L molꢀ1 cmꢀ1). μeff = 2.8 μB at 300 K.
2
2
6H, CH(CH3) ), 1.32 (d, J = 6.1, 6H, CH(CH3) ) ppm. Anal. Calcd for
C48H74AlN4NaO6: C, 67.58; H, 8.74; N, 6.57. Found: C, 66.80; H, 8.20; N,
6.79. UVꢀvis spectrum (hexanes) λmax (εM): 285 (11 150), 447 (1630) nm
(L molꢀ1 cmꢀ1). This compound is diamagnetic.
[(Et2O)2Na][(IP2ꢀ)2Al] (6). Compound 6 was prepared following the
same procedure described for compound 5. However, ether was used in
place of DME. Compound 6 was obtained as a purple powder (0.81 g,
59%). Crystals suitable for X-ray diffraction were obtained by cooling a
DME solution of the product at ꢀ25 °C for 1 week. 1H NMR (300 MHz,
C6D6): δ 7.23 (t, J = 5.1, 2H, Ph), 7.11 (d, J = 4.6, 2H, Ph), 6.87 (d, J =
6.6, 2H, Ph), 6.58 (m, 2H, py), 5.92 (d, J = 9.3, 2H, py), 5.51 (s, 2H, im
CH), 5.39 (m, 2H, py), 4.67 (s, 2H, py), 4.25 (hept, J = 6.6, 2H,
CH(CH3)2), 1.49 (d, J = 6.8, 6H, CH(CH3)2), 1.39 (d, J = 6.8, 6H,
CH(CH3)2) ppm. Anal. Calcd for C44H64AlN4NaO2: C, 72.30; H, 8.82;
N, 7.66. Found: C, 72.08; H, 8.97; N, 7.15.
[(IPꢀ)2Al(CF3SO3)] (3). Compound 3 was prepared following the
same procedure described for compound 2. However, Al(CF3SO3)3 was
used in place of AlCl3. Compound 3 was obtained as a dark green powder
(0.58 g, 44%). Crystals suitable for X-ray diffraction were obtained by cooling
a DME solution of the product at ꢀ25 °C for 1 week. 1H NMR (300 MHz,
C6D6): 8.58 (s, py), 8.47 (d, J = 4.54, py), 8.27 (d, J = 8.64, py), 7.43 (2, py),
7.07ꢀ6.87 (m, Ph), 6.63 (s, Ph), 4.50 (s, 2H, imCH)), 4.20 (s, Ph), 3.77
(m, 2H, CH(CH3)2) 1.24 (d, J = 7.97, 12H, CH(CH3)2) ppm. Anal. Calcd
for C37H44AlF3N4O3S: C, 62.70; H, 6.26; N, 7.90. Found: C, 62.58; H, 6.70;
N, 7.94. UVꢀvis spectrum (hexanes) λmax (εM): 241 (13 610), 356
(14 940), 661 (1490) nm (L molꢀ1 cmꢀ1). μeff = 3.4 μB at 300 K.
[(IP2-)Al](μ-IPꢀIPꢀ) (4). Sodium metal (0.13 g, 5.7 mmol) and IP
(1.0 g, 3.75 mmol) were stirred with DME (10 mL) for 1 h. The resulting
’ ASSOCIATED CONTENT
S
Supporting Information. CIF files for the structures of
b
1ꢀ6, and UVꢀvisible spectra of IP and 4. Depictions of the
solid-state structures of 1, 3, and 6. Cyclic voltammogram for IP,
magnetic susceptibility data for 4 and 5, and fits to variable-
temperature EPR data for 2. This material is available free of
charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
laberben@ucdavis.edu
8671
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672

Journal of the American Chemical Society
ARTICLE
Author Contributions
†These authors contributed equally.
(25) For example: (a) Crawford, V. H.; Richardson, H. W.; Wasson,
J. R.; Hodgson, D. J.; Hatfield, W. E. Inorg. Chem. 1976, 15, 2107. (b)
Ruiz, E.; Alemany, P.; Alvarez, S.; Cano, J. J. Am. Chem. Soc. 1997,
119, 1297.
(26) Bencini, A.; Gatteschi, D. EPR of Exchange Coupled Systems;
Springer: Berlin, 1990.
’ ACKNOWLEDGMENT
(27) Riplinger, C.; Kao, J. P. Y.; Rosen, G. M.; Kathirvelu, V.; Eaton,
G. R.; Eaton, S. S.; Kutateladze, A.; Neese, F. J. Am. Chem. Soc. 2009,
131, 10092.
(28) Gardiner, M. G.; Hanson, G. R.; Henderson, M. J.; Lee, F. C.;
Raston, C. L. Inorg. Chem. 1994, 33, 2456.
We would like to thank the University of California, Davis for
funding and Drs. J. Berg, P. Klavins, and J. Fettinger for assistance
with IR spectrometer, SQUID magnetometer, and X-ray crystal-
lography instrumentation.
(29) van Gastel, M.; Lu, C. C.; Wieghardt, K.; Lubitz, W. Inorg.
Chem. 2009, 48, 2626.
(30) Robin, M. B.; Day, P. Adv. Inorg. Chem. Radiochem. 1967,
10, 247.
(31) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42.
(32) (a) SMART Software Users Guide, Version 5.1; Bruker Analytical
X-Ray Systems, Inc.: Madison, WI, 1999. (b) SAINT Software Users
Guide, Version 7.0; Bruker Analytical X-Ray Systems, Inc.: Madison, WI,
1999. (c) Sheldrick, G. M. SADABS, Version 2.03; Bruker Analytical
X-Ray Systems, Inc.: Madison, WI, 2000. (d) Sheldrick, G. M. SHELXTL
Version 6.12; Bruker Analytical X-Ray Systems, Inc.: Madison, WI, 1999.
(e) International Tables for X-Ray Crystallography; Kluwer Academic
Publishers: Dordrecht, 1992; Vol. C.
’ REFERENCES
(1) Activation of Small Molecules; Tolman, W. B., Ed.; Wiley-VCH:
Weinheim, 2006.
(2) For example: Wright, A. M.; Wu, G.; Hayton, T. W. J. Am. Chem.
Soc. 2010, 132, 14336.
(3) Stephan., D. W. Chem. Commun. 2010, 46, 8526.
(4) Peng, Y.; Ellis, B. D.; Wang, X.; Fettinger, J. C.; Power, P. P.
Science 2010, 325, 1668.
(5) (a) Bhugun, I.; Lexa, D.; Saveant, J.-M. J. Phys. Chem. 1996,
100, 19981–19985. (b) Bhugun, I.; Lexa, D.; Saveant, J.-M. J. Am. Chem.
Soc. 1996, 118, 1769–76. (c) Hammouche, M.; Lexa, D.; Momenteau,
M.; Saveant, J. M. J. Am. Chem. Soc. 1991, 113, 8455–66.
(6) Gambarotta, S.; Arena, F.; Floriani, C.; Zanazzi, P. F. J. Am.
Chem. Soc. 1982, 104, 5082–92.
(7) Schmidt, M. H.; Miskelly, G. M.; Lewis, N. S. J. Am. Chem. Soc.
1990, 112, 3420–6.
(8) Menard, G.; Stephan, D. J. J. Am. Chem. Soc. 2010, 132, 1796–1797.
(9) Cui, C.; Roesky, H. W.; Schmidt, H.-G.; Noltemeyer, M.; Hao,
H.; Cimpoescu, F. Angew. Chem., Int. Ed. 2000, 39, 4274.
(10) Zhu, H.; Chai, J.; Chandrasekhar, V.; Roesky, H. W.; Magull, J.;
Vidovic, D.; Schmidt, H. G.; Noltemeyer, M.; Power, P. P.; Merrill, W. A.
J. Am. Chem. Soc. 2004, 126, 9472–9473.
(11) (a) Herzog, S.; Geisler, K.; Praekel, H. Angew. Chem., Int. Ed.
Engl. 1963, 2, 47. (b) Flamini, A.; Poli, N. Inorg. Chim. Acta 1988,
150, 149.
(12) (a) K€oster, R.; Benedikt, G.; Schr€otter Angew. Chem., Int. Ed.
Engl. 1964, 3, 514–515. (b) Lehmkuhl, H.; Fuchs, G.; K€oster, R.
Tetrahedron Lett. 1965, 29, 2511.
(13) (a) Kaim, W. J. Organomet. Chem. 1980, 201, C5–C8. (b) Kaim,
W. Chem. Ber. 1981, 114, 3789. (c) Kaim, W. Z. Naturforsch. 1982, 37B, 783.
(14) (a) Cloke, F. G. N.; Dalby, C. I.; Henderson, M. J.; Hitchcock,
P. B.; Kennard, C. H. L.; Lamb, R. L.; Raston, C. L. J. Chem. Soc., Chem.
Commun. 1990, 1394–1396. (b) Cloke, F. G. N.; Dalby, C. I.; Daff, J.;
Green, J. Dalton Trans. 1991, 181–184.
(15) Leman, J. T.; Barron, A. R.; Ziller, J. W.; Kren, R. M. Polyhedron
1989, 8, 1909.
(16) Jurca, T.; Dawson, K.; Mallov, I.; Burchell, T.; Yap, G. P. A.;
Risheson, D. S. Dalton Trans. 2010, 39, 1266.
(17) Scott, J.; Gambarotto, S.; Korobkov, I.; Knijnenburg, Q.; de
Bruin, B.; Budzelaar, P. H. M. J. Am. Chem. Soc. 2005, 127, 17204.
(18) Westerhausen, M.; Bollwein, T.; Makropoulos, N.; Schneiderbauer,
S.; Suter, S.; N€oth, H.; Mayer, P.; Piotrowski, H.; Polborn, K.; Pfitzner, A. Eur.
J. Inorg. Chem. 2002, 389–404.
(19) Lu, C. C.; Bill, E.; Weyherm€uller, T.; Bothe, E.; Wieghardt, K.
J. Am. Chem. Soc. 2008, 130, 3181.
(20) Wu, J. Y.; Stanzl, B. N.; Ritter, T. J. Am. Chem. Soc. 2010,
132, 13214–13216.
(21) Addison, A. W.; Rao, T. N.; Van Rijn, J. J.; Verschoor, G. C.
J. Chem. Soc., Dalton Trans. 1984, 1349.
(22) Laine, T. V.; Klinga, M.; Leskel€a, M. Eur. J. Inorg. Chem.
1999, 959.
(23) Schmitt, E. A. Ph.D. Thesis, University of Illinois at Urbana-
Champaign, Urbana, IL, 1995.
(24) Ketterer, N. A.; Fan, H.; Blackmore, K. J.; Yang, X.; Ziller, J. W.;
Baik, M.-H.; Heyduk, A. F. J. Am. Chem. Soc. 2008, 130, 4364.
8672
dx.doi.org/10.1021/ja2015718 |J. Am. Chem. Soc. 2011, 133, 8662–8672