Influence of the redox active ligand on the reactivity and electronic structure of a series of Fe(TIM) complexes

Article abstract of DOI:10.1021/ic100586h

The redox properties of Fe and Zn complexes coordinated by an α-diimine based N₄-macrocyclic ligand (TIM) have been examined using spectroscopic methods and density functional theory (DFT) computational analysis. DFT results on the redox series of [Zn(TIM*)]ⁿ and [Fe(TIM*)]ⁿ molecules indicate the preferential reduction of the α-diimine ligand moiety. In addition to the previously reported [Fe(TIM*)]₂ dimer, we have now synthesized and characterized a further series of monomeric and dimeric complexes coordinated by the TIM ligand. This includes the five-coordinate monomeric [Fe(TIM*)I], the neutral and cationic forms of a monomeric phosphite adduct, [Fe(TIM*)(P(OPh) ₃)] and [Fe(TIM*)(P(OPh)₃)](PF₆), as well as a binuclear hydroxy-bridged complex, [{Fe(TIM*)}₂(μ-OH)] (PF₆). Experimental and computational data for these synthetic compounds denote the presence of ferrous and ferric species, suggesting that the α-diimine based macrocycles do not readily support the formation of formally low-valent (M⁰ or M^I) metal complexes as previously speculated. Magnetochemical, Moessbauer, electron paramagnetic resonance (EPR), and electronic spectral data have been employed to experimentally determine the oxidation state of the central metal ion and of the macrocyclic ligand (TIM*) in each compound. The series of compounds is described as follows: [Fe^II(TIM⁰)(CH₃CN ₂)]²⁺, S_Fe = S_T = 0; [Fe ^2.5(TIM^2.5-)]₂, S_T = 1; [{Fe ^III(TIM^2-)}₂(μ-OH)]⁺, S _Fe = 3/2, S_T = 0; [Fe^III(TIM^2-)I], S_Fe = 3/2, S_T = 1/2; [Fe^II(TIM ^2-)(P(OPh₃))], S_Fe = S_T = 0; and [Fe^II(TIM^1-)(P(OPh₃))]¹⁺/[Fe ^I(TIM⁰)(P(OPh₃))]¹⁺, S_T = 1/2. The results have been corroborated by DFT calculations.

Full text of DOI:10.1021/ic100586h

5
686 Inorg. Chem. 2010, 49, 5686–5700
DOI: 10.1021/ic100586h
Influence of the Redox Active Ligand on the Reactivity and Electronic Structure of a
Series of Fe(TIM) Complexes
†
Corinna R. Hess, Thomas Weyherm €u ller, Eckhard Bill, and Karl Wieghardt*
†
Max Planck Institut f u€ r Bioanorganische Chemie, 45470 M u€ lheim an der Ruhr, Germany. Current address:
Department of Chemistry, Durham University, South Road, Durham, DH1 3LE, United Kingdom.
Received March 29, 2010
The redox properties of Fe and Zn complexes coordinated by an R-diimine based N -macrocyclic ligand (TIM)
4
have been examined using spectroscopic methods and density functional theory (DFT) computational analysis.
DFT results on the redox series of [Zn(TIM*)] and [Fe(TIM*)] molecules indicate the preferential reduction of
n
n
the R-diimine ligand moiety. In addition to the previously reported [Fe(TIM*)] dimer, we have now synthesized
2
and characterized a further series of monomeric and dimeric complexes coordinated by the TIM ligand. This
includes the five-coordinate monomeric [Fe(TIM*)I], the neutral and cationic forms of a monomeric phosphite
adduct, [Fe(TIM*)(P(OPh) )] and [Fe(TIM*)(P(OPh) )](PF ), as well as a binuclear hydroxy-bridged complex,
3
3
6
[
{Fe(TIM*)} (μ-OH)](PF ). Experimental and computational data for these synthetic compounds denote the
2 6
presence of ferrous and ferric species, suggesting that the R-diimine based macrocycles do not readily support the
0
I
formation of formally low-valent (M or M ) metal complexes as previously speculated. Magnetochemical,
M €o ssbauer, electron paramagnetic resonance (EPR), and electronic spectral data have been employed to
experimentally determine the oxidation state of the central metal ion and of the macrocyclic ligand (TIM*) in
II
0
2þ
2
each compound. The series of compounds is described as follows: [Fe (TIM )(CH CN )] , SFe = S = 0;
3
T
2
.5
2.5-
III
2-
þ
III
2-
[
[
Fe (TIM )] , S = 1; [{Fe (TIM )} (μ-OH)] , S = 3/2, S = 0; [Fe (TIM )I], SFe = 3/2, S = 1/2;
2 T 2 T T
II II I 0
Fe (TIM )(P(OPh ))], S = S = 0; and [Fe (TIM )(P(OPh ))] /[Fe (TIM )(P(OPh ))] , S = 1/2. The
Fe
1-
2-
1þ
1þ
3
Fe
T
3
3
T
results have been corroborated by DFT calculations.
1
. Introduction
An extensive literature has been devoted to the chemistry
reactions. Most recently, cobalt tetraimine complexes have
been exploited as hydrogen production catalysts.
8,9
While the chemistry of the neutral macrocyclic R-diimine
complexes is well-established, it is far from complete. Studies
pertaining to ligand-centered redox processes of these mole-
cules are particularly lacking, in comparison with analogous
of metal complexes coordinated by synthetic tetradentate
N -macrocycles. These ligands provided simple synthetic
analogues of related biological molecules, such as hemes and
cobalamin. Unsaturated macrocyclic R-diimine metal com-
plexes in particular (e.g., tetraeneN * and TIM*, Scheme 1)
have proven to be broadly reactive, and have been targeted
for a wide range of chemical transformations including
dioxygen activation, CO reduction, and alkyl transfer
1
4
10,11
2
studies on porphyrin ligands.
Similar to the porphyrin
4
systems, however, evidence points to the redox non-innocence
of R-diimine-based macrocycles. Both Rose and Busch noted
the presence of low-lying ligand-centered π*-orbitals in Ni-
1,3
4,5
6,7
2
(
TIM) and Fe(tetraeneN ) complexes that might favor re-
4
12,13
II
0
2þ
duction of the macrocycle.
Reduction of [Ni (TIM )]
*
To whom correspondence should be addressed. E-mail: wieghardt@
mpi-muelheim.mpg.de. Phone: þ49 (0)208 306-3610. Fax: þ49 (0)208
06-3952.
1) Endicott, J. F.; Durham, B. In Coordination Chemistry of Macrocyclic
Compounds; Melson, G. A., Ed.; Plenum Press: New York, 1979; pp 393-460.
2) We will designate the macrocycles as TIM* and tetraeneN * when
3
(
(8) Hu, X.; Brunschwig, B. S.; Peters, J. C. J. Am. Chem. Soc. 2007, 129,
8988.
(
4
(9) Fihri, A.; Artero, V.; Razavet, M.; Baffert, C.; Leibl, W.; Fontecave,
used in a generic sense, irrespective of their true oxidation state (ranging
M. Angew. Chem., Int. Ed. 2008, 47, 564.
(
10) Synthesis and Organic Chemistry; Kadish, K. M., Smith, K. M., Guilard,
from 0 to 4-).
(
3) Wong, C.-L.; Switzer, J. A.; Balakrishnan, K. P.; Endicott, J. F. J. Am.
R., Eds.; Academic Press: New York, 2000; Vol. 1.
11) Floriani, C.; Floriani-Moro, R. In Inorganic, Organometallic and
(
Chem. Soc. 1980, 102, 5511.
4) Fujita, E.; Brunschwig, B. S.; Ogata, T.; Yanagida, S. Coord. Chem.
Rev. 1994, 132, 195.
5) Schmidt, M. H.; Miskelly, G. M.; Lewis, N. S. J. Am. Chem. Soc. 1990,
12, 3420.
(
Coordination Chemistry; Kadish, K. M., Smith, K. M., Guilard, R., Eds.;
Academic Press: New York, 2000; Vol. 3, pp 405-421.
(12) Lovecchio, F. V.; Gore, E. S.; Busch, D. H. J. Am. Chem. Soc. 1974,
96, 3109.
(13) Norman, J. G. J.; Chen, L. M.; Perkins, C. M.; Rose, N. J. Inorg.
Chem. 1981, 20, 1403.
(
1
(
(
6) Shi, S.; Bakac, A.; Espenson, J. H. Inorg. Chem. 1991, 30, 3410.
7) Bakac, A.; Espenson, J. H. Inorg. Chem. 1989, 28, 4319.
pubs.acs.org/IC
Published on Web 04/28/2010
r 2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5687
a
Scheme 1
Scheme 2. Oxidation Levels of R-Diimine Ligands
a
˚
Average bond distances given in red in A.
ligand-centered redox processes on the electronic structure
and reactivity of the Fe-containing compounds. Toward this
effort, we now have carried out detailed computational
studies on a series of monomeric Zn and Fe species coordi-
nated by the TIM* ligand. The results establish clear struc-
tural trends across the series, which assist in the electronic
structure characterization of M(TIM*) complexes. We also
have investigated the reactivity of the Fe dimer toward
several small molecules: I , PR , and O . The product Fe-
II
1- 1þ
to generate the monocation, [Ni (TIM )] , containing a
1-
12
ligand radical anion, (TIM ), has been established. In con-
trast, reduction of [Fe(tetraeneN )(CH CN) ] was reported
to yield low valent Fe and Fe species.
Recent studies have unequivocally established the redox
non-innocence of the fundamental building block of the
tetraeneN4 and TIM ligands, namely, the unsaturated
Redox studies on a variety of 4- or
-coordinate first row transition metal complexes coordi-
2þ
4
3
2
I
0
14
1
5-18
R-diimine unit.
5
2
3
2
(TIM*) complexes all comprise reduced TIM* ligands in
nated by iminopyridine and R-diimine ligands have demon-
strated the tendency of the organic molecules, rather than the
metal center, to acquire electrons. Spectroscopic, computa-
tional, and crystallographic data have provided clear evi-
dence for ligand radicals in these systems. Crystal structures
have proven particularly informative with regard to the
ligand oxidation state; concrete changes in the R-diimine
C-C and C-N bond distances are observed with each
successive reduction (Scheme 2). Consequently, a theoretical
various oxidation states (Scheme 3), again corroborating the
redox non-innocence of the macrocyclic diimines. The iso-
lation of low spin and intermediate spin ferrous and ferric
complexes points to the ability to stabilize a variety of unique
electronic structures. We have not found any evidence for
I
0
the presence of genuine low-valent Fe or Fe complexes in
7
8
the series of Fe(TIM*) compounds, with a genuine d or d
electronic configuration.
4
e- redox series is available to the macrocyclic R-diimines,
2
. Experimental Section
Syntheses were carried out under an inert atmosphere,
using standard Schlenk and glovebox techniques. Solvents
Fluka or Sigma-Aldrich) were dried and deoxygenated prior
to use. Reagents were purchased from Sigma-Aldrich.
Bis(pentamethylcyclopentadienyl)iron hexafluorophosphate,
[FeCp* ]PF , was synthesized by oxidation of FeCp* using
equiv of AgPF in acetonitrile. The resultant green solution
was filtered, and the solvent was removed in vacuo. The solid
product was used without further purification.
TIM and tetraeneN , which are expected to compete with
4
metal centered redox processes in their coordination com-
plexes. This factor is pertinent to the recent focus on Co-
containing tetraimine systems as hydrogen evolution cata-
lysts, since low-valent Co species have been invoked in the
catalytic cycle. Ligand based redox processes also provide
(
I
8
0
I
2
6
2
an alternative formulation of Busch’s low valent Fe and Fe
1
6
14
complexes. A greater understanding of the nature of the redox
processes in metal-coordinated R-diimine macrocycles may
provide key insight into the unique reactivities of these systems.
To address such issues, we have focused our studies on the
redox chemistry and reactivity of Fe complexes coordinated
by the TIM* ligand. We recently reported on the formation
of an unprecedented Fe-Fe dimer, [Fe(TIM*)] , generated
upon reduction of [Fe (TIM )(MeCN )] with sodium in
tetrahydrofuran (THF). The dimer has an S = 1 ground
state and possesses a weak unsupported Fe-Fe bond (bond
order= 0.5). Spectroscopic and computationalstudies on the
dimeric product clearly established the redox non-innocence
of the TIM* ligands; the presence of TIM
in [Fe(TIM*)] was deemed largely accountable for the
remarkable structure and magnetism of the dimer. These
findings prompted us to further investigate the redox chem-
istry of the Fe(TIM*) system and to examine the effects of
II
0
Synthesis. [Zn (TIM )Cl](PF ) (1). Compound 1 was pre-
6
II
0
4
pared from [Cu (TIM )]ZnCl according to a previously pub-
2
0
lished procedure. Singlecrystalswereobtainedbyslowdiffusion
of ether into a solution of 1 in acetonitrile. Anal. Calcd for
C H ClF N PZn: C, 34.01; H, 4.86; N, 11.33. Found: C,
34.05; H, 5.12; N, 11.33. H NMR (CDCl
2.25, (s, 12H), 2.97 (q, 2H), 3.63, (t, 2H), 3.86 (d, 2H).
14 24
6
4
2
1
II
0
2þ
2
3
), ppm: 2.07 (d, 2H),
19
II
0
[
prepared as previously described.
Fe (TIM )(MeCN) ](PF ) (2) and [Fe(TIM*)] (3) were
2 6 2 2
19
[
{Fe(TIM*)} (μ-OH)](PF ) (4). Compound 4 was prepared
2
6
from solutions of 2, as follows: 2.5 equiv of Na (as 5% Na/Hg
beads = 100 mg Na/Hg = 0.217 mmol Na) was added to a
suspension of 0.059 g (0.088 mmol) of 2 in ∼11 mL of THF.
After 24 h stirring, the crude mixture was filtered to remove Hg
and Na salts, and transferred to a Schlenk flask. Assuming that
2.5-
ligand radicals
2
II
0
all of the starting compound, [Fe (TIM )(MeCN)
2 6 2
](PF ) , was
converted to [Fe(TIM*)] , then 0.51 equiv of dioxygen (2.25 ꢀ
2
-5
1
0
2
mol O , added as 2.6 mL of air, 23 °C, 1 atm) was added via
(
14) Rakowski, M. C.; Busch, D. H. J. Am. Chem. Soc. 1975, 97, 2570–
571.
15) Lu, C. C.; Bill, E.; Weyherm u€ ller, T.; Bothe, E.; Wieghardt, K. J. Am.
Chem. Soc. 2008, 130.
16) Ghosh, M.; Weyherm u€ ller, T.; Wieghardt, K. Dalton Trans. 2008,
syringe, and the solution was stirred for 30 min, during which
time the color changed to purple. The sample was placed under
partial vacuum, and the solution was transferred to the glovebox
and filtered. Yield: 100% conversion, as determined by UV-vis,
using the absorbance at λmax = 511 nm. M o€ ssbauer analysis on
the crude solid product mixture also shows clean conversion
2
(
(
5
149.
(
17) Chlopek, K.; Muresan, N.; Neese, F.; Wieghardt, K. Chem.;Eur. J.
2
007, 13, 8390.
(
18) Muresan, N.; Chlopek, K.; Weyherm u€ ller, T.; Neese, F.; Wieghardt,
K. Inorg. Chem. 2007, 46, 5327.
19) Hess, C. R.; Weyherm u€ ller, T.; Bill, E.; Wieghardt, K. Angew. Chem.,
Int. Ed. 2009, 48.
(
(20) Jackels, S. C.; Ciavola, J.; Carter, R. C.; Cheek, P. L.; Pascarelli,
T. D. Inorg. Chem. 1983, 22, 3956.

5
688 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
(
>90% purity, Supporting Information). Crystals of 4 were
to the γ-beam. Isomer shifts are quoted relative to iron metal at
300 K. Data simulation was carried out using the programs
“JulX” (SQUID), and “MFIT” (M o€ ssbauer) by E. Bill. The X-ray
powder diffraction measurement was performed on a Stoe STA-
DI P transmission diffractometer, equipped with a primary Ge
(111) monochromator and a linear position sensitive detector.
obtained by slow diffusion of ether into the crude product
mixture. Anal. Calcd for C28 Fe OP: C, 43.65; H,
.41; N, 14.54. Found: C, 43.99; H, 6.67; N, 14.09. UV-vis:
H
49
F
6
2 8
N
6
λ
-
1
-1
max, nm (ε M cm ) in THF: 277 (23,400), 511 (10,650), 668
(
3500), 744 (sh).
[
˚
Fe(TIM*)I] (5). I
2
(0.018 g, 0.071 mM) was added to a
The wavelength used was Mo KR1: 0.7093 A. Full powder
solution of 3 (0.043 g, 0.071 mM) in ∼5 mL of THF, whereupon
the color immediately turned purple. After several hours reac-
tion time, the solution was subsequently filtered, and the solvent
removed under vacuum. Yield: 82% Single crystals were ob-
tained by slow diffusion of ether into a saturated solution of
patterns were taken from 2 to 20° 2θ in steps of 0.01° 2θ. Data
were collected at room temperature.
X-ray Crystallographic Data Collection and Refinement of the
Structures. A colorless single crystal of 1, black crystals of 4, 5,
and 6 0.5 hexane, and a brown crystal of 7 were coated with
3
5
in THF. Anal. Calcd for C14
H24FeIN
4
: C, 39.00; H, 5.61; N,
perfluoropolyether, picked up with nylon loops, and were
immediately mounted in the nitrogen cold stream of the dif-
fractometer to prevent loss of solvent. A Bruker-Nonius Kappa-
CCD diffractometer was used for crystals of 4. Compounds 1, 5,
1
3.00. Found: C, 38.98; H, 5.57; N, 12.92. UV-vis: λmax nm
-
1
-1
(
ε M cm ) in THF: 403 (4800), 537 (8900), 674 (3900), 745 (sh).
Fe(TIM*)(P(OPh) )] (6). In a typical procedure, triphenyl-
[
3
phosphite (2.14 equiv, 47 μL = 0.18 mmol) was added to a
solution of 3 (0.051 g, 0.084 mmol) in 5 mL of THF, upon which
the color immediately turned a turquoise-green. The reaction was
allowed to stir overnight, and the solvent was removed under
vacuum. The crude product was redissolved in hexane, filtered to
remove any insoluble materials, and the dark turquoise filtrate
was concentrated ∼2-3 fold. Single crystals were obtained by
cooling (-19 °C) the saturated solution of the product in hexane.
Yield: 65% Anal. Calcd for C H FeN O P: C, 62.55; H, 6.40;
6 0.5 hexane, and 7 were measured on a Bruker APEX II dif-
fractometer. Diffractometers were equipped with a Mo-target
rotating-anode X-ray source and a graphite monochromator
3
˚
(Mo-KR, λ = 0.71073 A). Final cell constants were obtained
from least-squares fits of setting angles of several thousand
strong reflections. Intensity data were corrected for absorption
2
1
using intensities of redundant reflections using SADABS. The
structures were readily solved by Patterson methods and sub-
sequent difference Fourier techniques. The Bruker ShelXTL
2
2
3
2
39
4
3
N, 9.12. Found: C, 62.56; H, 6.35; N, 9.09.UV-vis: λmax nm
software package was used for solution, refinement, and art-
work of the structures. All non-hydrogen atoms were anisotro-
pically refined, and hydrogen atoms were placed at calculated
positions and refined as riding atoms with isotropic displace-
ment parameters, except for the protons of the bridging OH-
-
1
-1
(
(
(
ε M cm ) in hexane: 264 (12,500), 412 (9200), 450 (5600), 590
1
12,000), 892 (400). H NMR (THF-d ), ppm: 1.67, (s, 6H) 1.68
s, 6H), 2.06 (m, 2H), 3.08 (m, 2H), 3.31, (m, 4H), 3.62 (m, 4H),
6
6
.74 (d, 6H), 6.82 (t, 3H), 7.03 (t, 6H).
Fe(TIM*)(P(OPh) )](PF ) (7). To a solution of 6 (0.03 g,
6
.049mmol) in THF (10 mL) was added 1 equiv of [FeCp* ](PF )
groups in crystals of 4, which were located from the difference
map for both crystallographically independent cations. A PF
[
3
6
-
0
2
6
anion in compound 4 was found to be disordered over two
positions. A restrained split atom model was refined giving an
occupation ratio of 0.88:0.12. Anisotropic displacement para-
meters of corresponding disordered atoms were restrained to be
equal using EADP. P-F bond distances in the minor com-
ponent were restrained to be equal within errors applying the
SAME instruction of ShelXL97.
(
0.023 g, 0.049 mmol). The reaction mixture was stirred for 2 h,
during which time the color became dark purple. The solution
was filtered, and the solvent removed under vacuum. The crude
product was washed with hexane to remove FeCp* and any
unreacted starting materials. Yield: 50%. Single crystals could be
obtained from slow diffusion of hexane into a THF solution of
the product. Repeated attempts to obtain a reasonable elemental
analysis of this compound were unsuccessful; we suspect that this
complex is unstable in the commercially available pyrolysis
process. However, M o€ ssbauer analysis of the crude material
2
Density Functional Theory (DFT) Calculations. All DFT
2
3
calculations were performed with the ORCA program package.
The geometry optimizations of the complexes were performed at
2
4-26
the B3LYP
where geometry optimization at the BP86
level of DFT, except for compounds 5, 6, and 7,
level of DFT
(
Figure 12) obtained as described from the reaction of 6 with
2
4,27,28
FeCp*2 indicates clean conversion to a single Fe-containing
product. Thisis further supported byits X-ray powder diffraction
spectrum shown in Supporting Information, Figure S11, which is
identical with the simulated spectrum derived from the single
crystal analysis.
resulted in more accurate bond distances, by comparison to the
experimental values. Single-point calculations on the optimized
geometries were carried out using the B3LYP functional. This
hybrid functional often gives better results for transition metal
compounds than pure gradient-corrected functionals, especially
Physical Measurements. Elemental analyses were carried out
by the Microanalytical Laboratory, Kolbe & Springer, M u€ lheim
an der Ruhr, Germany. Electronic spectra were recorded on a
29
with regard to metal-ligand covalency. The all-electron Gauss-
ian basis sets were those developed by the Ahlrichs group.
30,31
Triple-ζ quality basis sets TZV(P) with one set of polarization
functions on the metals and on the atoms directly coordinated to
1
Perkin-Elmer double-beam spectrometer. H NMR spectra were
collected on a Varian Mercury 400 MHz instrument. Electro-
chemical measurements were carried out using an EG&G poten-
tiostat/galvanostat, glassy carbon working electrode, Ag/AgNO
reference electrode, and Pt wire as the counter electrode. Poten-
tials are reported versus ferrocenium/ferrocene. Magnetic sus-
ceptibility data were recorded on an MPMS Quantum Design
SQUID magnetometer (1.0 T, 2-300 K). Susceptibility data
were corrected for underlying diamagnetism using tabulated
Pascal’s constants. X-band electron paramagnetic resonance
30
the metalcenter wereused. Forthe carbonandhydrogenatoms,
slightly smaller polarized split-valence SV(P) basis sets were used,
that were of double-ζ quality in the valence region and contained
3
(21) SADABS, Vers. 2008/1; Bruker AXS Inc.: Madison, WI, 2003.
(22) ShelXTL, 6.14; Bruker AXS Inc.: Madison, WI, 2003.
(23) Neese, F. In An ab initio, DFT and Semiempirical Electronic Structure
Package, Version 2.7, Revision 0; Institut f €u r Physikalische und Theoretische
Chemie, Universit €a t Bonn: Bonn, Germany, 2009.
(
EPR) spectra were collected on a Bruker ESP 300 spectro-
(
(
24) Becke, A. D. J. Chem. Phys. 1986, 84, 4524.
25) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
meter. M o€ ssbauer data were recorded on alternating constant-
acceleration spectrometers. The minimum experimental line
(26) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(27) Perdew, J. P.; Yue, W. Phys. Rev. B 1986, 33, 8800.
-
1
width was 0.24 mm s (full width at half-height). The sample
temperature was maintained constant in an Oxford Instruments
Variox or an Oxford Instruments M o€ ssbauer-Spectromag 2000
cryostat; the latter was used for measurements in applied mag-
netic fields with the field at the sample oriented perpendicular
(
(
28) Perdew, J. P. Phys. Rev. B 1986, 33, 8822.
29) Neese, F.; Solomon, E. I. In Magnetism: From Molecules to Materials;
Miller, J. S., Drillon, M., Eds.; Wiley: New York, 2002; Vol. 4, p 345.
30) Sch €a fer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829.
(31) Sch €a fer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
(

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5689
Scheme 3
3
1
a polarizing set of d-functions on the non-hydrogen atoms.
Auxiliary basis sets used to expand the electron density in the
resolution-of-the-identity (RI) approach were chosen, where
applicable, to match the orbital basis. The SCF calculations
3. Results and Discussion
3
2,33
DFT Computational Studies on the Zn- and Fe-TIM*
Redox Series. To simplify the structural analyses of
M(TIM*) complexes, we have carried out DFT computa-
tional studies on a redox series of monomeric Zn- and Fe-
coordinated TIM* species, as described below. The series,
-
8
-7
were tightly converged (1 ꢀ 10
E
h
in energy, 1 ꢀ 10
h
E in
-
7
the density change, and1ꢀ 10 inmaximumelementofthe DIIS
error vector). The geometry optimizations for all complexes were
carried out in redundant internal coordinates without imposing
symmetry constraints. In all cases the geometries were considered
2þ
beginning with the fully oxidized dicationic [M(TIM*)]
2-
through the dianionic [M(TIM*)] , disentangle the effect
ofthe metal ion and ligand oxidation state on the electronic
structure of metal-TIM* complexes.
-
6
converged after the energy change was less than 5 ꢀ 10
h
E , the
gradient norm and maximum gradient element were smaller than
-
4
-1
-4
-1
1
ꢀ 10
E
h
Bohr and 3 ꢀ 10
E
h
Bohr , respectively, and the
root-mean square and maximum displacements of all atoms
n
[
Zn(TIM*)] series. DFT computational studies were
-
3
-3
carried out, using the B3LYP functional, on each member
of a 4e- redox series of monomeric Zn(TIM*) species
shown in Scheme 3A. Geometry optimizations and elec-
tronic structure determinations were carried out for each
species with ground states as defined in Scheme 3A.
The Zn series illustrates the nature of ligand-centered
were smaller than 2 ꢀ 10 Bohr and 4 ꢀ 10 Bohr, respectively.
Computational results are described by the broken-symmetry
(
34,35
BS) approach.
Since several broken symmetry solutions to
the spin-unrestricted Kohn-Sham equations may be obtained,
the general notation BS(m,n) has been adopted, where m(n)
3
6
denotes the number of spin-up (spin-down) electrons at the two
interacting fragments. Canonical and corresponding orbitals
3
7
redox processes, virtually uncomplicated by the metal
ion, which in all cases comprises a Zn(II) ion (d , S =0).
3
8
were generated with the program Molekel. Non-relativistic
single-point calculations on the optimized geometry were carried
out to predict M o€ ssbauer spectral parameters (isomer shifts and
quadrupole splittings). These calculations employed the CP(PPP)
10
Zn
The low energy of the Zn d-orbitals precludes significant
overlap with the ligand orbitals; five doubly occupied 3d
metal orbitals were identified in each case that lie far below
the ligand centered valence orbitals in energy. The increas-
39
basis set for iron. The M o€ ssbauer isomer shifts were calculated
from the computed electron densities at the iron centers as
previously described.
4
0a
n
ing negative charge across the [Zn(TIM*)] series, thus,
We are aware of the fact that DFT does not allow for rigorous
handling of open-shell states since the total spin is not defined.
For cases with the maximum number of unpaired electrons one
can assume that the calculated state approaches the proper spin
eigenstate, except for some possible small spin contamination.
However, for an open-shell singlet, and also for other similar
states where many multiplets are possible for a given open-shell
configuration, the DFT solution is necessarily a mixture of
reflects the consecutive one electron reduction of the TIM*
macrocycle (Scheme 3B), as borne out by the DFT-derived
molecular orbital (MO) diagrams for the Zn(TIM*) series
(
Figure 1). Each reduction leads to the successive occu-
pancy of the ligand imine π* orbitals.
A clear trend in the calculated R-diimine C -C and
im
im
n
C
im-Nim bond distances is manifest across the [Zn(TIM*)]
40b,c
multiplets.
redox series: an increase in the Cim-Nim bonds lengths is
observed with each successive reduction, corresponding to
a decrease of the C -C bond distance (Figure 2). The
Nevertheless, we use DFT calculations here to explore the
basic electronic structure of a number of open-shell compounds,
in particular, to explain their spectroscopic properties. However, in
all cases the spin ground state of a given molecule was experimentally
determined. We do not attempt to infer details of the spin coupling
scheme from the present computational results.
im
im
trend is as expected in proceeding from the fully oxidized
II
0
2þ
II
Zn (TIM )] to the fully reduced [Zn (TIM )] . The
4- 2-
[
dicationic species, [Zn (TIM )] , consists of the neutral
II
0
2þ
0
TIM , with four C -N double bonds and two C -C
im
im
im
im
single bonds. The DFT derived MO diagram for the
dianionic species, [Zn (TIM )] , represents the fully
reduced macrocyle, TIM . The two doubly occupied
(
32) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Chem. Phys.
Lett. 1995, 240, 283.
33) Eichkorn, K.; Treutler, O.; Ohm, H.; H €a ser, M.; Ahlrichs, R. Chem.
II
4- 2-
€
4-
(
Phys. Lett. 1995, 242, 652.
ligand π* orbitals are antibonding with respect to the
C -N bonds and bonding with respect to the C -C
bonds, concurrentwiththe altered bonddistances, yielding
(34) Ginsberg, A. P. J. Am. Chem. Soc. 1980, 102, 111.
(35) Noodleman, L.; Peng, C. Y.; Case, D. A.; Mouesca, J. M. Coord.
im
im
im
im
Chem. Rev. 1995, 144, 199.
36) Kirchner, B.; Wennmohs, F.; Ye, S.; Neese, F. Curr. Opin. Chem.
Biol. 2007, 11, 134.
(
four C -N single bonds and two C =C double bonds
im
im
im
im
(
Scheme 3B). The calculated structures of the monocationic
(
(
(
(
37) Neese, F. J. Phys. Chem. Solids 2004, 65, 781.
38) Molekel; http:\\www.cscs.ch/molkel/.
39) Neese, F. Inorg. Chim. Acta 2002, 337, 181.
40) (a) Sinnecker, S.; Slep, L. D.; Bill, E.; Neese, F. Inorg. Chem. 2005, 44,
2þ
1- 1þ
2þ
3- 1-
[Zn (TIM )] and the monoanionic [Zn (TIM )]
complexes imply that the unpaired electron in each molecule
is delocalized over both diimine units of the TIM or
1-
2
(
245. (b) Ziegler, T.; Rauk, A.; Baerends, E. Theor. Chim. Acta 1977, 43, 261.
c) Illas, F.; Moreira, I.; Bofill, J.; Filatov, M. Theor. Chem. Acc. 2006, 116, 587.
3-
TIM ligand, since in both structures the two C -N
im
im

5
690 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
n
Figure 1. DFT derived (B3LYP) qualitative molecular orbital scheme for the [Zn(TIM)] redox series. The ligand centered frontier orbitals are depicted;
Zn d-orbitals lie well below in energy.
coupling through the S=0 metal center, in favor of a dia-
4
1
magnetic ground state.
The experimentally determined bond distances obtained
II
0
þ
from the crystal structure of the [Zn (TIM )(Cl)] complex
1) (vide infra) provide a reference for the calculated values,
and are in close agreement with the DFT-derived bond
(
II
0
2þ
lengths for the [Zn (TIM )] dication (Table 1). The slight
discrepancy is attributed to the presence of the fifth ligand
(the Cl ion) in 1. Indeed, the bond distances derived from
computational studies on the [Zn (TIM )Cl] moiety
-
II
0
þ
˚
˚
(
C -C (avg) = 1.528 A; C -N (avg) = 1.278 A) are
im
im
im
im
well matched with the experimentally determined values
˚
˚
(
C -C (avg) = 1.516(2) A;C -N (avg) = 1.278(1) A).
im
im
im
im
n
Fe(TIM*)] Series. An analogous redox series of Fe-
[
TIM*) species, namely, [Fe(TIM*)] f [Fe(TIM*)]
2
þ
1-
(
(
Scheme 4), also was investigated using DFT methods.
As illustrated in Figure 4, the Fe(TIM*) series is
Figure 2. DFT derived TIM R-diimine bond distances across the 4e-
n
considerably more complex than the Zn(TIM*) series.
The oxidation state of the ligand strongly influences the
spin state at the iron ion and reduction of the macrocycle
coincides with significant orbital rearrangements across
[
Zn(TIM)] redox series; (blue solid squares) R-diimine C-C bond
distances, (blue solid triangles) R-diimine C-N bond DFT calculated
green open circles) and crystallographically determined (green solid
(
1þ
circles) R-diimine C-C and C-N bond distances for [Zn(TIM)Cl]
also shown for comparison.
II
the series. Both low-spin and intermediate spin Fe ions
are observed. Yet despite the presence of the redox-active
iron ion, reduction also appears to be ligand based rather
than metal centered. Thus, the iron center remains diva-
lent across the series; we have not found any evidence
bond distances and the two C -N bond distances are
im
im
equivalent.
Both the S = 0 and S = 1 spin states of the neutral
II
Zn (TIM )] species were examined. The bond distances
2- 0
[
I
n
from computational studies to support an [Fe (TIM*)]
or an [Fe (TIM*)] species. This would suggest that
of the two structures are nearly identical, since the nature
of the occupied diimine π* orbitals is identical with respect
to the C-N and C-C bonding character. The triplet
diradical and closed-shell singlet species also are nearly
degenerate in energy; the calculated energy difference
between the two states was found to be a mere 3.8 kcal/
mol in favor of a triplet ground state. The near degeneracy
contrasts with other square planar metal complexes con-
taining ligand diradicals, where strong antiferromagnetic
interactionsbetween the two ligand-centeredspinsarisevia
0
n
þ
Busch’s [Fe(tetraeneN )] (OTf) (OTf = triflate anion)
4
0
and [Fe(tetraeneN )] complexes should be formulated as
4
1-
2
þ þ 1þ þ
Fe (tetraeneN )] rather than [Fe (tetraeneN )] ,
4 4
0
[
2þ
2-
0
0
0 0
4
and [Fe (tetraeneN )] rather than [Fe (tetraeneN )] .
DFT results for [Fe(TIM)(OTf)] demonstrate that the elec-
tronic structure is not perturbed by coordination of the anion
4
(Supporting Information); the five coordinate triflate complex
also consists of a divalent iron ion and an anionic (TIM )
ligand. Likewise, monomeric [Fe (tetraeneN )] provides
1-
2þ
2- 0
4
0
a more accurate description of the reported [Fe (tetraeneN )]
species, regardless of the total spin state of the molecule.
(
41) Blanchard, S.; Neese, F.; Bothe, E.; Bill, E.; Weyherm u€ ller, T.;
4
Wieghardt, K. Inorg. Chem. 2005, 44, 3636.

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5691
Table 1. Select M(TIM*) Bond Lengths from DFT Calculations and Crystallographic Values
Scheme 4
However, a description as a dimer is conceivable, in light of the
[
42
Fe(TIM*)] structure.
2
The trend in the R-diimine C -N and C -C bond
im
im
im
im
distances (Figure 3) parallels that of the Zn series. How-
ever, the values for the C -C bond distances all are
im
im
shorter than the analogous bond lengths in the Zn series
˚
(
Δd ∼ 0.03 A), while the C-N bond distances all are
˚
slightly longer (Δd ∼ 0.01 A). This effect is a consequence
of greater overlap of the Fe d orbitals with the ligand π*
orbitals, giving rise to enhanced π-backbonding effects.
The ligand π* orbitals possess greater electron density than
in the corresponding orbitals of the Zn(TIM*) species
stemming from the increased covalency of the Fe-N
versus Zn-N bonds. The largest variation in the R-diimine
Figure 3. DFT derived TIM R-diimine bond distances across the 3e-
n
[
Fe(TIM)] redox series; (red solid squares) R-diimine C-C bond dis-
tances, (red solid triangles) R-diimine C-N bond distances. DFT calcu-
lated (purple open circles) and crystallographically determined (purple
solid circles) R-diimine C-C and C-N bond distances for [Fe(TIM)
2þ
bond lengths occurs with the reduction of [Fe(TIM*)] to
1þ
[
Fe(TIM*)] , which corresponds to a change in the metal
spin state from low-spin (SFe = 0) to intermediate-spin
SFe = 1). The diimine bond distances also are dependent
2þ
(
MeCN)
2
]
also shown for comparison.
(
0
C -N distances in the calculated structure of [Fe(TIM*)] ,
on the total spin state of the molecule, as shown for the
singlet (S = 0) and triplet (S = 1) states of the [Fe -
im
im
II
2-
II
which indicate the presence of an Fe ion and a TIM ligand.
The experimental bond distances determined for [Fe-
TIM)(MeCN)2] (PF ) (C -C (avg) = 1.487(3) A;
C -N (avg) = 1.292(4) A) provide reference values
T
T
0
TIM )] and [Fe (TIM )] species. The triplet state of
2þ
II
2- 0
(
2
þ
II
Fe (TIM )] was found to be 34 kcal/mol lower in energy
0
2þ
˚
(
[
than the singlet state, as expected for a four-coordinate square
6 2
im
19
im
˚
im
im
6
planar molecule containing a divalent Fe ion (d , S = 1).
for the Fe series. The crystallographic values are in close
agreement with the calculated R-diimine bond distances
for the singlet [Fe (TIM )] species, and nearly identical
Fe
II
The triplet and singlet states of [Fe (TIM )] , however, are
2- 0
II
0
2þ
again nearly isoenergetic (ΔE(S = 0 - S = 1): 3.6 kcal/
T
T
II
to the geometry optimized values for the [Fe (TIM )-
0
mol). Given the near-degeneracy of the two structures, one
might expect the two-electron reduced [Fe(TIM*)] to be
2
þ
0
˚
(
(
MeCN)2] structure (C -C (avg) = 1.504 A;C -N -
im
im
im
im
˚
avg) = 1.291 A). The values for the C-C and C-N bond
particularly reactive. Formation of the [Fe(TIM*)] dimer
2
distances of the [Fe(TIM*)] dimer lie precisely in between
might in fact proceed via a pathway involving the neutral,
[
singly occupied dz
2
2þ
2-
0
the value calculated for the [Fe (TIM )] and [Fe -
2þ
0
Fe(TIM*)] , triplet diradical complex, which consists of a
3
- 1-
(TIM )] species (Table 1), in further support of the
presence of a mixed valent TIM
2
orbital and is thus primed for the forma-
tion of the Fe-Fe bond. Noteworthy are the C -C and
2
.5-
ligand. The explicit
im
im
trend in the diimine bond lengths illustrates the value of
crystallographic data in establishing the ligand oxidation
state in metal-coordinated R-diimine macrocyclic com-
plexes. However, the series also highlights scenarios where
(
42) The M o€ ssbauer parameters of Busch’s two-electron reduced complex
are very similar to those reported for [Fe(TIM*)] (δ = 0.13 mm/s, ΔE
.23 mm/s); Scarborough, C. C, Wieghardt, K., unpublished.
2
Q
=
2

5
692 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
n
Figure 4. DFT derived (B3LYP) qualitative molecular orbital scheme for the [Fe(TIM)] redox series. Sp. Ov. = DFT derived spatial overlap.
þ
precise electronic structure assignments become more difficult0.
The calculated R-diimine bond distances for the [Fe(TIM*)]
triplet diradical are comparable to the values obtained for the
of 2 (MeCN, ∼ -1.4 vs Fc /Fc) cleanly generates the one-
electron reduced product; the electronic spectrum coincides
=
with new bands in the region from 600-800 nm (λmax
1-
monoanionic [Fe(TIM*)] . Therefore, crystallographic evi-
dence alone would be insufficient to distinguish between the
trianionic TIM and the dianionic TIM ligand.
640 nm) and a blue-shift of the absorption band at 552 nm
that correspond to the starting complex (Figure 5, inset).
The first reduction is clearly reversible, given that the CV is
reproducible subsequent to coulometry, and reoxidation of
the sample restores the original absorption spectrum of 2.
The anisotropic EPR spectrum of the electrochemically
generated one-electron reduced product affords g values of
2.50, 2.07, and 2.0 (Supporting Information). The values
3-
2-
Reactivity. Synthesis and Characterization. Compound
was prepared according to the procedures described
1
by Jackels et al. for the synthesis of [Zn(MePhTIM)Cl]-
PF (MePhTIM=2,9-dimethyl-3,10-diphenyl-1,4,8,11-tetra-
6
2
0
azacyclo-tetradeca-1,3,8,10-tetraene). The procedure entails
the reductive transmetalation of [Cu(TIM)]ZnCl ; the lat-
ter complex, likewise, was prepared according to the pub-
again are in close agreement with those reported for the
EPR spectrum of the [Fe(tetraeneN *)] species (g= 2.302,
4
þ
4
43 1
I
2.074, 2.006), which was assigned to an Fe complex.
14
lished methods for the MePhTIM analogue. H NMR
illustrates that compound 1 is diamagnetic, as expected
upon coordination of the redox-inert Zn ion to the neutral
However, the EPR values also are entirely consistent
with an intermediate spin Fe center antiferromagneti-
cally coupled to the ligand π-radical anion ((TIM ) or
(tetraeneN4 )), which would signify ligand centered
reduction. The second reductive process observed in
the CV of 2 is irreversible, and its electrochemical
generation proved unsuccessful.
The one-electron reduced product could not be ob-
tained synthetically. As reported previously, chemical
reduction of [Fe(TIM*)(MeCN) ](PF ) using Na/Hg
II
0
1-
TIM ligand. The cyclic voltammogram (CV) in aceto-
nitrile exhibits a reversible one-electron reduction at -1.66 V
versus ferrocenium/ferrocene (Fc /Fc) (Supporting Infor-
1
-
þ
mation). This process is assigned as a ligand-centered reduc-
tion. The isolation of reduced Zn(TIM*) products (i.e.,
II
1-
þ
II
Zn (TIM )] or [Zn (TIM )] ) proved unsuccessful.
2- 0
[
The dication of octahedral compound 2, which consists
0
of the neutral TIM ligand, a low spin ferrous ion, and two
acetonitrile ligands in the trans-position, was the initial
choice of starting reagent to access and examine further
redox states of Fe(TIM*) complexes. The CV of 2 in
2
6 2
correspond to the two electron reduction of each Fe-
(TIM*) unit and yields the dimeric complex, [Fe(TIM*)]2
(3); no additional products, such as monomeric inter-
mediates preceding dimer formation, were observed.
The triplet ground state and the weak metal-metal
acetonitrile (0.2 M TBAPF , 25 °C; Figure 5) exhibits a
6
fully reversible one-electron oxidative wave at þ0.59 V
þ
versus Fc /Fc, as well as two closely spaced, one-electron
reductive processes at E1/2 = -1.13 V and -1.60 V. These
potentials agree nicely with the values reported for the
bond of [Fe(TIM*)] suggested to us that compound 3
2
should be highly reactive toward small molecules. In-
deed, the Fe-Fe dimer, composed of two reduced TIM
II
reduction of Busch’s [Fe (tetraeneN )(MeCN) ] (OTf)2
0
4
2þ
2.5
ligands ([Fe (TIM
-2.5
)] ), provides an ideal route to
2
2
1
complex. Controlled potential coulometry of a solution
4
further Fe(TIM*) complexes in various oxidation states,
via the reaction with dioxygen, iodine, and phosphines
(Scheme 5).
(
43) Coltrain, B. K.; Jackels, S. C. Inorg. Chem. 1981, 20, 2032.

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5693
Figure 5. Cyclic voltammogram of 2 in acetonitrile, 0.2 M TBAPF
(
6
, 25 °C, 100 mV/s. Inset: Absorption spectrum of controlled potential coulometry of 2
þ
-1.4 V vs Fc /Fc) to generate the one-electron reduced species.
a
2
Scheme 5. Synthesis of Compounds 4-7 from [Fe(TIM*)] (3)
a
Ligands have been drawn to more clearly describe the coordination geometry, and should not be taken as indicative of the the ligand oxidation state
0
or electronic structure of these compounds (i.e. the TIM ligand in 1-7 is NOT in the fully oxidized TIM form as depicted).
The reaction of [Fe(TIM*)] with dioxygen, in the
2
the data, assuming two independent spins of S = 1/2
(H = -2JS S ; S = S = 1/2), yields a moderately large
presence of a proton source (water), leads to the forma-
tion of the binuclear, hydroxo-bridged compound, [{Fe-
1
2
1
2
-
1
J value (J = -92.5 cm ) because of antiferromagnetic
coupling of the two Fe(TIM*) units. The rise in μeff values
from 50-300 K is indicative of an energetically low-lying
4
4
þ
(
cation denotes a two-electron oxidation of 3 by O . A
TIM*)} (μ-OH)](PF ) (4). The [{Fe(TIM*)} (μ-OH)]
2 6 2
2
color change of the solution from blue to violet-purple is
observed during the reaction of [Fe(TIM)] with air,
paramagnetic excited state for this compound (S = 1).
T
Compound 5, the neutral, five-coordinate compound,
[Fe(TIM*)I], is readily generated from the reaction of
[Fe(TIM*)] with 1 equiv of iodine. Each [Fe(TIM*)] unit
2
of 3 is oxidized by one electron, concomitant with coordi-
nation of the halide to the Fe center. The doublet ground
state (S = 1/2) of the iodo-complex was confirmed by
T
2
correspondent with a blue shift in the absorption band
of 3 at 578 nm, to 511 nm. Compound 4 is diamagnetic
below 50 K, as established from magnetic susceptibility
measurements (2-300 K, 1 T) (Figure 6). The best fit of
(
44) Dioxygen is added to the [Fe(TIM)]
2
solutions by injecting an
magnetic susceptibility measurements (Supporting Infor-
mation), which also indicate substantial intermolecular inter-
actions in the solid state (Weiss constant, θ = -21.9 K).
The crystalline solid of 5 is dark purple and exhibits absorp-
tion bands at 540 and 672 nm in the electronic spectrum.
appropriate volume of air, and the necessary protons were obtained from
moisture present; this method, as detailed in the Experimental Section,
proved to be the cleanest and most consistent route for the synthesis of
[
{Fe(TIM*)}
2 6
(μ-OH)](PF ). Attempts to use other proton sources, including
pyridinium-hexafluorophosphate, were unsuccessful.

5
694 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
Figure 7. Thermal ellipsoid representation (50%) of 1.
Figure 6. Magnetic susceptibility data (1 T, 2-300 K) for 4; experi-
mental = solid dots, simulation = solid line. Simulation with S = S
1
2
=
-1
-6
1/2, g = 2.0, J = -92.5 cm ,TIP = 1005 ꢀ 1e emu. The simula-
tion includes a small percentage (0.4%) of high spin (S = 2.5) impurity,
1
= g
2
shown as dashed line (PI), which accounts for the values of ueff > 0 from
2-40 K.
These features resemble the charge transfer (CT) bands
observed in the spectrum of a structurally related iodo-Fe(III)-
diiminosemiquinonate, which are characteristic of diimine
4
5
ligand diradicals.
The reaction of 3 with triphenylphosphite, likewise,
breaks the Fe-Fe bond and generates the neutral, five-
coordinate, [Fe(TIM*)(P(OPh) )] complex (6). Formally,
3
Figure 8. Thermal ellipsoid representation (50%) of 4.
this reaction does not represent a redox reaction, rather it
is an addition of 2 equiv of the phosphite to 1 equiv of the
dimer, 3. The H NMR illustrates that compound 6 is
OfnotearetheFe-P distances of compounds 6and 7: the
1
þ
Fe-P bond length of the cationic [Fe(TIM*)(P(OPh) )] is
3
˚
significantly longer (2.181(1) A) than the corresponding
diamagnetic. The CV of 6 (THF) exhibits a single rever-
þ
sible one-electron oxidative process at -1.2 V versus Fc /
axial bond of the neutral [Fe(TIM*)(P(OPh) )] molecule
3
˚
Fc (Supporting Information). On the basis of this poten-
tial, bis(pentamethylcyclopentadienyl)iron hexafluoro-
phosphate ([FeCp* ]PF ) was chosen as the oxidative
(Fe-P = 2.085(1) A). The same trend was previously
described in an analogous five-coordinate iron phosphine
4
5
2
6
complex coordinated by R-diimine ligands. The shorter
Fe-P bond of the neutral versus the cationic complex was
attributed to the better π-backbonding ability of a ferrous
versus ferric ion.
þ
reagent to generate the cationic [Fe(TIM*)(P(OPh)3)]
complex (7). As for compound 5, the oxidized phosphite
complex has a doublet ground state (Supporting Infor-
mation) and exhibits intense absorption bands in the
500-700 nm region of the electronic spectrum, giving
rise to the purple color of solutions of 7.
The hydroxo-bridged binuclear iron complex, 4, has a
“butterfly”-like structure: the Fe-O-Fe angle is 143°,
enforcing the V-shape with respect to the two macrocyclic
Solid State Structures. The solid state structures of
compounds 1 and 4-7 are shown in Figures 7-9. The
crystallographic data are listed in Table 2. The crystallo-
graphic data for compounds 2 and 3 has been described
N -planes. Two conformers of 4 are observed within the
4
unit cell, which differ slightly in the twist angle between
the two Fe(TIM*) units about the hydroxo group. Each
˚
Fe center is square pyramidal and displaced by 0.46-0.47 A
1
9
elsewhere.
from its N -cavity plane. The Fe-O bond distances (1.977(3)
4
˚ ˚
II
0
The Zn (TIM ) complex, 1, and the monomeric Fe(TIM*)
complexes 5-7, all are five-coordinate square pyramidal.
The Zn ion in 1 is displaced significantly (0.83 A) from the
A and 1.994(3) A) are in accord with the range of values
Com-
46-48
observed for (μ-hydroxo)diiron(III) complexes.
˚
pound 4 closely resembles the reported structure of a (μ-
4
6
N -macrocyclic cavity, while in the iron compounds, the
4
hydroxo)bis(octaethylporphyrinato)iron(III) complex, and
affords only the third example of a hydroxo-bridged diiron
complex lacking additional supporting bridging ligands.
proximity of the metal to the N plane is much closer
4
49
˚
(
0.4-0.47 A). In all structures, except 7, the TIM ligand
adopts a saddle-like structure, the two diimine units point-
ing downward, and thetwo methylene groups of the propyl
backbones directed upward along the metal-axial bond. In
the cationic [Fe(TIM*)(P(OPh)3] complex (7), these
methylene groups adopt a trans configuration.
(
46) Scheidt, W. R.; Cheng, B.; Safo, M. K.; Cukiernik, F.; Marchon,
J.-C.; Debrunner, P. G. J. Am. Chem. Soc. 1992, 114, 4420.
2þ
(47) Stamatatos, T. M.; Christou, A. G.; Jones, C. M.; O’Callaghan, B. J.;
Abboud, K. A.; O’Brien, T. A.; Christou, G. J. Am. Chem. Soc. 2007, 129, 9840.
(48) Jullien, J.; Juhasz, G.; Dumas, E.; Mayer, C. R.; Marrot, J.; Riviere,
E.; Bominaar, E. L.; Munck, E.; Secheresse, F. Inorg. Chem. 2006, 45, 6922.
(49) Evans, D. R.; Mathur, R. S.; Heerwegh, K.; Reed, C. A.; Xie, Z.
Angew. Chem., Int. Ed. 1997, 36, 1335.
(
45) Chlopek, K.; Bill, E.; Weyherm u€ ller, T.; Wieghardt, K. Inorg. Chem.
005, 44, 7087.
2

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5695
Figure 9. Thermal ellipsoid representation (50%) of compounds 5-7.
Table 2. Crystallographic Data for 1, 4, 5, 6 0.5Hexane, and 7
3
1
4
5
6 0.5 hexane
3
7
chem. formula
Fw
space group
C
14
H
24ClF
6
N
4
PZn
C
28
H
49
F
6
Fe
2
N
8
OP
C
431.12
14
H
24FeIN
4
C
657.58
35
H
46FeN
4
3
O P
C
32
H
4 3
39FeN O P
494.16
P4 2, No. 96
9.0688(6)
9.0688(6)
24.056(2)
90
90
90
1978.4(2)
4
100(2)
770.42
P2 /c, No. 14
13.1409(8)
13.7913(8)
37.013(3)
90
91.606(3)
90
6705.2(8)
8
100(2)
759.46
P2 /c, No. 14
13.572(2)
16.327(2)
15.444(2)
90
109.529(2)
90
3225.6(6)
4
100(2)
2
3 1
1
Pbcn, No. 60
9.9374(8)
14.8792(13)
10.7044(9)
90
90
90
1582.8(2)
4
100(2)
P1, No. 2
8.795(2)
9.128(2)
20.810(4)
89.791(4)
89.028(4)
78.168(4)
1635.0(6)
2
1
˚
a, A
˚
b, A
˚
c, A
R, deg
β, deg
γ, deg
˚
V, A
Z
T, K
100(2)
1.336
-3
F calcd, g cm
1.659
1.526
1.809
1.564
refl. collected/2Θmax
unique refl./I > 2σ(I)
no. of params/restr.
135983/70.00
4365/4304
126/0
0.71073/15.19
0.0193/1.127
0.0537
85789/52.00
13148/9326
873/23
0.71073/9.83
0.0528/1.016
0.1082
56015/72.96
3812/2718
94/0
0.71073/29.00
0.0238/1.101
0.0556
45386/61.98
10322/8745
402/0
0.71073/5.51
0.0388/1.082
0.0936
73290/55.00
7412/5317
437/0
0.71073/6.43
0.0372/1.051
0.0780
-
1
˚
λ, A /μ(KR), cm
R1 /goodness of fit
a
b
c
wR2 (I > 2σ (I))
residual density, e A
-
3
˚
þ0.56/-0.34
þ0.49/-0.60
þ1.56/-0.56
2 2
þ1.16/-0.54
þ0.81/-0.52
2 2 1/2
P
P
P
P
P
a
b
|. GoF = [w(F
2
1/2 c
2
2 2
c
Observation criterion: I > 2σ(I). R1 = ||F
o
| - |F
c
||/ |F
o
o
- F
c
) ]/(n - p)]
wR2 = [ w(F
o
- F
) / w(F
o
) ] where
2
2
2
2
o
2
c
w = 1/σ (F
o
) þ (aP) þ bP, P = (F
þ 2F
)/3.
þ
The crystallographically determined C-C and C-N
R-diimine bond distances of the M(TIM*) structures, 1-7,
provide the first clues regarding the oxidation state of the
ligand and, consequently, of the metal center. The bond
distances can be compared readily to the DFT-derived
Thecationic[Fe(TIM*)] unit of compounds 4, 5, and 7is
one-electron oxidized with respect to the [Fe(TIM*)] moiety
0
of the dimeric precursor compound (3). The R-diimine bond
distances of compound 4 (C-C: 1.409(6), C-N: 1.339(6);
Table 2) again are in accord with a TIM diradical dianion,
signifying the presence of two ferric ions. The bond distances
of the two TIM units are identical, ruling out a localized,
mixed valent species of class I. On the basis of the structural
data, a reasonable formulation of compound 4 is [{Fe -
(TIM )} (μ-OH)] . Likewise, the crystallographically de-
2-
n
ligand bond lengths of the monomeric [Zn(TIM*)] and
n
Fe(TIM*)] redox series. As discussed previously, the
2-
[
[
II
0
II
0
Zn (TIM )Cl](PF ) and [Fe (TIM )(MeCN) ](PF )
6
2
6 2
II
3þ
complexes serve as benchmarks for the dicationic [Fe -
0
TIM )] unit, with a singlet (S = 0) ground state. The
2þ
2-
þ
(
2
C -C and C -N bond distances of 1 and 2 (Table 1)
termined R-diimine bond distances of 5 (C-C: 1.417(2),
im
im
im
im
2-
are as expected for the C-C single bond and CdN double
bonds of the fully oxidized TIM ligand, containing four
C-N: 1.338(2); Table 2) specify a dianionic TIM ligand,
0
3þ
2-
and thus a ferric complex, represented by [Fe (TIM )I].
The ligand bond distances of compound 7 deviate signi-
ficantly from those of compounds 4 and 5 (Table 2). The
CdN double bonds.
Compound 6, akin to the [Fe(TIM*)] dimer, 3, is
2
0
˚
composed of the neutral [Fe(TIM*)] unit. The R-diimine
C -C distance at 1.442(4) A is significantly longer,
im
im
˚
˚
˚
C-C (1.407(2) A) and C-N (1.342(2) A) bond distances
while the C -N bond distance (1.317(3) A) is substan-
im
im
(
Table 2) are similar to the values obtained for 3. By
tially shorter than the corresponding bond lengths in
the structures of compounds 4 and 5 (Table 3). The
former values agree with the calculated distances for the
comparison to the DFT-derived bond distances of the
Fe(TIM*) series (Figure 3 and Table 1), the R-diimine
bond lengths of 6 are in agreement with the calculated
1-
monoanionic TIM species. Consequently, the crystallo-
graphic data implies that the cationic unit of 7 consists
2
-
values for the dianionic TIM ligand. The values are
compatible with either a low spin or an intermediate spin
Fe center.
1
of a ferrous ion coordinated by the TIM radical anion
-
II
II
([Fe (TIM )(P(OPh) )] . Oxidation of the neutral
1-
þ
3

5
696 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
Table 3. Crystallographically Determined, Average Ligand R-Diimine and Axial
Fe-Lax Bond Lengths for Compounds 4-7
˚
˚
˚
compound
C
im-Cim (avg), A
C
im-Nim (avg), A
Fe-Lax, A
4
1.409(6)
1.339(6)
Fe
Fe
A
-O: 1.977
-O: 1.994
B
5
6
7
1.417(2)
1.407(2)
1.442(4)
1.338(2)
1.342(2)
1.317(3)
L = I, 2.678
L = P, 2.085
L = P, 2.181
II
Fe (TIM )(P(OPh) )] (6) is thus ligand-centered. This
2-
[
contrasts with the synthesis of compounds 4 and 5, as both
3
of the latter compounds appear to contain an Fe(III)
center and TIM ligands.
2-
Spectroscopic Characterization. M o€ ssbauer and EPR
spectra (where applicable) were measured for compounds
4
pound (6) consists of the neutral [Fe(TIM*)] unit. The
-7. In this series, only the [Fe(TIM*)(P(OPh) )] com-
3
0
crystallographic data for the phosphite complex is con-
sistent with an [Fe (TIM )(P(OPh) )] formulation (vide
2þ
2-
3
supra). Two electronic structures are viable for this dia-
magnetic compound: [(FevV)(TIMVv)], with a low-spin,
SFe = 0, iron center and closed-shell, S = 0, ligand
L
dianion; or [(Fevv)(TIMVV)], composed of an intermediate
spin ferrous ion, SFe = 1, and a triplet, S = 1, radical
L
dianion. Antiferromagnetic interactions between the metal-
centered and ligand-centered spins in the latter formulation
give rise to the observed singlet, S = 0, ground state of
T
the molecule. The zero-field M o€ ssbauer spectrum of 6 at
8
0 K yields an isomer shift value of 0.14 mm/s and ΔE =
Q
1.99 mm/s (Supporting Information). These values are
similar to the M o€ ssbauer parameters obtained for related,
diamagnetic, 5-coordinate, Fe-phosphine complexes, con-
4
5,50
taining ligand radical dianions.
However, the isomer
shift and quadrupole splitting values do not unequivocally
distinguish between the two possibilities afforded by the
2þ
intermediate spin or low-spin Fe configurations described
above. The ambiguity regarding the metal spin state has
likewise prevented a precise assignment of the electronic
structure of the related Fe-phenylenediamine-phosphine,
Figure 10. Applied-field M o€ ssbauer spectra of 5, measured at 4.2 K,
1
-7 T. Experimental data shown as black triangles. Fit of the data (solid
blue line), S = 1/2, affords: δ = 0.28 mm/s, ΔE = 2.07 mm/s, η = 0.1,
= g = g = 2.00, [A/g ]: Axx = -6.03 T, Ayy = -32.44 T, Azz
II
2-
[
[
Fe (L
Fe (L
)(PR )], and Fe-aminothiophenol-phosphite,
3
T
Q
N,N
2-
N,S
II
45
)(P(OPh) ))] complexes. However, among
g
x
y
z
β
N N
=
3
II
2-
þ0.50 T. Isomer shift relative to R-Fe at 298 K.
the examples of [Fe (PR )(L•• )] complexes, the low-spin
iron configuration appears to be most prevalent.
The remaining compounds in the synthetic Fe(TIM*)
3
50
III
low-spin or intermediate spin Fe configuration. The
five-coordinate square pyramidal geometry at each Fe
center favors an SFe = 3/2 spin state.
series, compounds 4, 5, and 7, all contain the mono-
cationic [Fe(TIM*)] unit. As stated above, the crystallo-
þ
graphic data for [{Fe(TIM*)} (μ-OH)] (4) and [Fe-
The values for the isomer shift and quadrupole splitting
(δ = 0.28 mm/s and ΔE = 2.07 mm/s) determined from
2
(
TIM*)I] (5) point to the presence of ferric ions coordi-
Q
2
-
nated by the dianionic, TIM , ligands. The applied field
M o€ ssbauer spectrum of the μ-hydroxy dimer, 4, confirms
the singlet ground state of the binuclear complex, and
the M o€ ssbauer spectra of 5 are very similar to the data
obtained for 4. The comparable values indicate an iden-
tical electronic structure with respect to the iron center of
þ
yields values of δ = 0.27 mm/s, ΔE = 1.69 mm/s (4.2 K)
the shared monocationic [Fe(TIM*)] unit. Applied-field
Q
for the isomer shift and quadrupole splitting (Supporting
Information). The M o€ ssbauer spectrum of 4 is unlike the
spectra of analogous μ-hydroxo-bridged diiron(III) com-
M o€ ssbauer spectroscopy (1-7 T, 4.2 K) provides fur-
ther insight regarding the intrinsic spin state at the Fe
center of the paramagnetic iodo-complex (Figure 10). The
simulation of the magnetic spectra yields values for
the hyperfine coupling constants, A(S = 1/2)/g β =
þ
plexes (e.g., [(Fe(salen)) (OH)] : δ = 0.49 mm/s, ΔE =
2
Q
4
6
þ
-
0
0.97 mm/s; [(Fe(octaethylporphinato)) (OH)] : δ =
2
48
.40 mm/s, ΔE = 1.13 mm/s ), which all are character-
T
N N
Q
(-6.03, -32.44, þ0.50) T. The two large negative compo-
ized as high-spin ferric compounds. The isomer shift and
quadrupole values of compound 4 are in line with either a
nents and one smaller positive component of the hyper-
fine coupling tensor, denotes an intermediate spin (S
3
=
/2) ferric ion. The A-tensor values can be converted
Fe
5
0
to the intrinsic values for an intermediate spin iron
center using spin projection techniques, and thus afford
(
50) Ghosh, P.; Bill, E.; Weyherm u€ ller, T.; Wieghardt, K. J. Am. Chem.
Soc. 2003, 125, 3967.

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5697
a
Table 4. M o€ ssbauer and EPR Values for 5 and 7, and Related Compounds
a
For all compounds, the Applied-Field M o€ ssbauer was measured at 4.2 K and the EPR measured as frozen solutions, at 10 K. The EPR spectra of 5
b
and 7 were measured in THF. Further details can be found within the text or in the references indicated. β and γ values refer to Euler angles that
determine the rotation of the electric field gradient tensor with respect to the principal axes of zero-field splitting and g-tensor.
A(SFe = 3/2)/g β = (-3.62, -19.47, þ0.30) T, with A -
R = 0°, β = 30°, γ = 90°) leads to the observed splitting
pattern in the EPR spectrum of 5. The EPR data, likewise,
matches that of the related intermediate spin ferric-iodo com-
plexes (Table 4), in support of the shared electronic structure.
N N
iso
(
SFe = 3/2)/g β = -7.60 T. The isomer shift, quadrupole
N N
splitting, and hyperfine coupling constants all are in accord
with the M o€ ssbauer data determined for related five-coordi-
nate ferric-iodo complexes coordinated by ligand radical
þ
The remarkable outlier in the series of [Fe(TIM*)]
4
5,50,51
dianions.
characterized as intermediate spin ferric species. The electronic
These compounds, listed in Table 4, also were
compounds is the monocationic phosphite complex, 7.
Both the EPR and M o€ ssbauer spectra are in stark con-
trast to the data obtained for the related series of five-
structure of 5 is thus represented best by [Fe(vvv)(TIMVV)I],
with SFe = 3/2 and S = 1; antiferromagnetic coupling
3
þ
2-
þ
3þ
2-
þ
coordinate [Fe (L )] and [Fe (TIM )] complexes
listed in Table 4. The data provides additional evidence of
a unique electronic structure for 7. The crystallographic
data for 7 led to the prediction that the formation of the
monocationic phosphite complex involves the oxidation
of the TIM ligand in 6 rather than of the metal center.
Compound 7 would thus consist of a divalent iron center,
rather than a ferric ion, as established for 4 and 5. Two
electronic configurations again are plausible for a ferrous
L
between the unpaired metal- and ligand-centered spins
gives rise to the doublet, S = 1/2, ground state of the
T
3þ
2-
þ
molecule. By analogy, the [Fe (TIM )] units of com-
pound 4 are assigned an identical configuration.
The EPR spectrum of 5 (Figure 11, bottom) provides
further evidence of the electronic structure deduced from
the crystal structure and M o€ ssbauer spectroscopy. The
anisotropic spectrum displays g values of 2.3076, 2.102,
and 2.012. Hyperfine interactions with the iodo ligand
þ
description of 7, denoted by [Fe(vV)(TIMv)] and [Fe(vv)-
(TIMV)] . The former representation signifies a low spin
-
4
-1
þ
(
I= 5/2, A =(4, 11, -16) ꢀ 10 cm ) and a large electric
-
4
-1
quadrupole contribution (P = (6, -10, 4) ꢀ 10 cm
;
ferrous ion, with the unpaired spin resident in a ligand-
centered orbital. The latter configuration consists of an
2
þ
center, antiferromagnetically
intermediate spin Fe
coupled to the ligand radical anion. The EPR spectrum
(
51) Chun, H.; Bill, E.; Weyherm u€ ller, T.; Wieghardt, K. Inorg. Chem.
2
003, 42, 5612.

5
698 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
Figure 11. Bottom: EPR spectrum of 5, as frozen solution in THF/
toluene, 20 K, Microwave frequency 9.4303 GHz, 6 μW power, 5 G
modulation amplitude; Top: EPR spectrum of 7 as frozen solution in
THF, 15 K, Microwave frequency 9.4326 GHz, 0.1 mW power, 10 G
modulation amplitude.
of 7(Figure 11, top) displays a much tighter range of gvalues in
comparison to the EPR spectrum of 5, and is best fit with g=
5
2
2
.10, 2.08, and 2.02. However, the values are in accord with
2þ
either a low-spin or intermediate spin Fe description. The
isomer shift afforded by the M o€ ssbauer spectra is comparable to
the values observed for both 4 and 5. However, the very small
quadrupole splitting, ΔE = -0.35, is indicative of minimal
Q
charge asymmetry at the iron center and thus favors a low-spin
2þ
Figure 12. Zero-Field (0 T, 80 K) and applied-field M o€ ssbauer spectra
Fe species. The magnetic M o€ ssbauer spectrum (Figure 12)
for the cationic complex corroborates the low-spin iron
configuration. The small hyperfine coupling constants
(
4-7 T, 4.2 K) of 7. Experimental data shown as black squares. Fit of the
data (solid blue line), S = 1/2, affords: δ=0.17mm/s,ΔE =-0.35 mm/s,
η = 0.78, g = g = g = 2.00, [A/g β ]: A = -0.43 T, Ayy = þ1.05 T,
T
Q
x
y
z
N
N
xx
obtained from the fit of the data, A(S = 1/2)/g β =
Azz = þ0.20 T. Isomer shift relative to R-Fe at 298 K.
T
N N
(
-0.43, þ1.05, þ0.20) T, contrast with the values obtained
for compounds 4 and 5. The values indicate minimal
chargedensityattheironcenter, whichwouldbeconsistent
with the presence of a low-spin ferrous ion (SFe = 0) in 7.
DFT Computational Studies. Density functional theory
(
Supporting Information, Table S1). For the binuclear
3þ
2-
þ
[
{Fe (TIM )} (µ-OH)] complex, 4, substantial electron
2
delocalization over the various ligand-centered and metal-
centered orbitals, coincides with highly complex orbital
depictions that obscure a clear assignment of the electronic
structure of this binuclear complex. To obtain a simplified
molecular orbital (MO) picture for this compound, we
therefore calculated the electronic structure of the homo-
(
DFT) calculations (B3LYP functional) were carried out
on the synthetic Fe(TIM*) compounds, 4-7, to further
resolve aspects regarding their electronic structure. For all
compounds, the bond lengths of the geometry optimized
structures match well with the crystallographically deter-
mined bond distances, and the calculated M o€ ssbauer
parameters agree reasonably well with experimental values
logous mononuclear [Fe(TIM*)(OH)] species (S = 1/2),
T
which represents the core half-unit of 4, and likewise,
consists of the monocationic [Fe(TIM*)] species. The
þ
DFT-derived electronic structure is consistent with the
[Fe(vvv)(TIMVV)] configuration deduced from the crystal
structure and the spectroscopic data for 4. The qualitative
MO diagram, shown in Figure 13, reveals three unpaired
þ
(
52) The EPR spectrum appears to be concentration dependent: at higher
concentrations (> 100 μM), a minor second component is apparent in the
EPR spectrum. The origin of the observed spectral changes is not known. No
differences are observed in the M o€ ssbauer spectrum of solid samples of 7 vs
frozen solutions in THF, suggesting that the electronic structure of 7 remains
the same in solution.
electrons that reside in the metal-centered d , d , and d
xz yz
z2
orbitals, which account for the S = 3/2 spin state at the Fe

Article
Inorganic Chemistry, Vol. 49, No. 12, 2010 5699
Figure 13. DFT-derived (B3LYP) qualitative MO diagram for [Fe(TIM*)(OH)] and 5. Spatial overlap (Sp. Ov.) indicated; z-axis along axial Fe-OH and
Fe-I bonds.
center. Two singly occupied ligand centered orbitals also are
apparent. The non-integer values for the spatial overlaps
between the ligand and metal-centered SOMOs (singly occu-
pied MOs) are indicative of strong antiferromagnetic coupling,
which accounts for the doublet ground state of the molecule.
The DFT-derived molecular orbital diagram for [Fe(TIM*)I]
M o€ ssbauer data, denoting negligible charge asymmetry and
charge density at the metal center. The MO picture that is
most consistent with the spectroscopic data is shown in
Figure 14, and illustrates the lowest energy solution derived
from calculations (UKS) on the unoptimized or BP86 geo-
metry optimized structure. The calculated M o€ ssbauer para-
(
5, Figure 13) is virtually identical to that of the monomeric Fe-
hydroxy species, and describes an analogous [Fe(vvv)(TIMVV)I]
SFe = 3/2, S = 1, S = 1/2) configuration.
meters for this structure (δ= 0.19 mm/s, ΔE =-0.85 mm/s)
also are comparable to experimental values, whereas the
calculated quadrupole splitting value for the intermediate
Q
(
L
T
2þ
2-
DFT results on the neutral [Fe (TIM )(P(OPh ))] com-
3
spin configuration is much larger (ΔE = þ1.4-1.5 mm/s),
Q
pound indicate that, of the two possible electronic configura-
tions in question, [(FevV)(TIMVv)] or [(Fevv)(TIMVV)], the
former description, containing a low-spin Fe center, is
more accurate. The doubly occupied ligand-centered orbital
also denotes closed-shell character for the dianionic ligand
and of opposite sign. Three doubly occupied metal d-orbitals
are apparent in the MO scheme; the single electron that
accounts for the doublet, S = 1/2, ground state of the
2þ
T
molecule occupies a highly covalent orbital with significant
metal and ligand character. The high d-occupancy of this
SOMO (34% dz2, 50% d(total) based on orbital population
analysis) obscures a clear assignment of this orbital as either
metal- or ligand-centered. Thus, the electronic structure of 7
can be described as either a low-spin ferrous compound
(Figure 14). The electronic structure of 6 differs from the
calculated structure of the theoretical four-coordinate
0
Fe(TIM*)] species, which described an intermediate spin
[
2þ
Fe configuration (Figure 4). The presence of the axial
phosphite ligand appears to favor a low-spin iron center.
The results of DFT calculations on the cationic [Fe(TIM*)-
2þ þ 2þ
Fe (TIM )(P(OPh ))] , in which the Fe center is co-
3
1-
[
1-
ordinated by the TIM ligand radical anion, or alternatively,
þ
(
P(OPh ))] compound are less clear-cut, and do little to
3
I
as an Fe complex. This latter description, represented by
resolve the ambiguity regarding the electronic structure
of this distinctive complex. The DFT calculations carried
out on the crystal structure, as well as on the geometry
optimized (BP86 and B3LYP) structures, suggest two plau-
sible electronic configurations for 7, that are analogous to the
1þ þ
Fe (TIM )(P(OPh ))] , denotes the presence of the fully
3
0
[
0
oxidized, neutral form of the ligand (TIM ). In this scenario,
the substantial ligand character of the SOMO signifies back-
bonding from the diimine π* orbitals into the Fe dz
2
orbital.
π-backbonding would also account for the C -C and
þ
im
im
possible low spin ([Fe(vV)(TIMv)] ) or intermediate spin
(
C -N bond distances of the TIM* macrocycle in 7, which
þ
im
im
[Fe(vv)(TIMV)] ) ferrous structures inferred from the experi-
mental data, and that differ in energy by a mere 2-5 kcal/mol
Supporting Information). The bond distances of the BP86
deviate from the expected bond lengths for the neutral, fully
oxidized, TIM ligand. The experimental and DFT-derived
0
(
M o€ ssbauer values for 7 are comparable to the M o€ ssbauer
values reported for the five-coordinate monovalent Fe com-
and B3LYP optimized structures both are in close agreement
with the experimentally determined bond lengths for 7, with
the Fe-P distance more accurately replicated in the BP86
optimized structure. The lowest energy solution obtained
from calculations (BS2,1) of the B3LYP geometry optimized
tBu
plex, [PhTt ]Fe(CO) (δ = 0.21 mm/s, ΔE = 0.47 mm/s,
1
2
Q
I
00 K), in further support of the Fe formulation of 7.
53
Significant π-backdonation from the carbonyl ligands of the
tBu
reported [PhTt ]Fe(CO) also account for the low isomer
2
[Fe(TIM*)(P(OPh ))] structure represents the intermediate
spin ferrous species ([(Fevv)(TIMV)(P(OPh ))] (Supporting
3
þ
3
(53) Mock, M. T.; Popescu, C. V.; Yap, G. P. A.; Dougherty, W. G.;
Information). However, this description conflicts with the
Riordan, C. G. J. Am. Chem. Soc. 2008, 47, 1889.

5
700 Inorganic Chemistry, Vol. 49, No. 12, 2010
Hess et al.
Figure 14. DFT-derived (B3LYP) qualitative MO diagram for 6 and 7. Spatial overlap (Sp. Ov.) indicated; z-axis along axial Fe-P bonds.
shift values for this monovalent complex, in comparison with
other reported Fe compounds.
among the complexes we have examined. However, the
spectroscopic data and DFT results for compound 7 indicate
that this cationic Fe-phosphite complex possesses some Fe
I
54,55
n
Conclusions. The series of calculated [Zn(TIM*)] and
I
n
Fe(TIM*)] structures, and the synthetic Fe(TIM*) com-
[
character. For this complex, we cannot explicitly assign an
electronic configuration and oxidation states for this complex.
Notably, the series of compounds that we have examined
demonstrate that the R-diimine macrocyle is able to stabilize
a range of metal and ligand oxidation states. The results
suggest that the redox properties of these macrocyclic di-
imine compounds are “tunable”, dependent on the nature of
the axial ligands (and presumably also would be influenced
by pendant groups on the macrocycle). In general, ligand-
centered redox processes will always compete with metal-
centered reduction in these compounds, precluding the
formation of low valent metal complexes. However, given
the appropriate choice of ligand environment, formation of
pounds described in this work definitively illustrate the redox
non-innocence of the macrocyclic ligand, and further advocate
the need to reconsider the redox chemistry of diimine-based
ligands in the chemistry of their inorganic compounds. The
preferential reduction of the TIM ligand, versus of the metal
n
center, already is implied by DFT results on the [Fe(TIM*)]
species, which encompasses a range of ligand oxidation
0
states starting from the neutral TIM , through the trianionic
3-
TIM ; yet in each case, the metal center remains in the
divalent oxidation state regardless of the overall spin state of
the molecule. Among the synthetic compounds, 3-7, the
crystallographic and spectroscopic data also confirm the pre-
sence of reduced TIM ligands. The two-electron reduction of
I
an Fe complex in these type of systems is not entirely out of
[
Fe(TIM*)(MeCN) ](PF ) involves the macrocycle and yields
2
the question, as evidenced by our analysis of compound 7.
The electron storage potential of the macrocyclic ligand and
the tunable redox states of these compounds should be useful
features to exploit for reactivity. These aspects warrant
further investigation into the properties and reactivities of
metal coordinated R-diimine based macrocyclic compounds.
6 2
2.5þ
2.5-
19
the Fe-Fe dimer, [Fe (TIM )]2, as previously described.
The addition of a phosphine donor (P(OPh) ) to this dimeric
complex breaks the metal-metal bond, yet again, the resultant
neutral monomeric compound consists of a ferrous ion coordi-
nated by the dianionic TIM form of the ligand.
3
2-
Interestingly, unlike in the computational series of Fe(TIM*)
molecules, the oxidation of both 3and 7(via the reaction with I2,
O , or Fc*PF ) removes one electron from either the metal or
Acknowledgment. We thank the Fonds der Chemischen
Industrie of Germany for financial support. The authors
thank Petra H o€ fer, Heike Schucht, Bernd Mienert, Andreas
G o€ bels, Frank Reikowski, and Dr. C. Weidenthaler for
technical assistance.
2
6
the ligand center, depending on the nature of the axial ligands.
Thus, the monomeric iodo complex, [Fe (TIM )I], and the
III
2-
III
2-
μ-hydroxy dimer, [{Fe (TIM )} (μ-OH)](PF ), both con-
2
6
tain ferric ions coordinated by the dianionic ligand. In contrast,
the oxidation of 6 appears to be ligand centered.
Supporting Information Available: Additional spectroscopic
information, M o€ ssbauer, electrochemical, andmagnetic suscepti-
bility data, and results of DFT computational analysis, as noted
within the text. This material is available free of charge via the
Internet at http://pubs.acs.org.
The majority of the experimental data for the synthetic
Fe(TIM*) compounds establishes a ferrous or ferric descrip-
tion for these complexes, rather than as low valent Fe or Fe
I
0
1
4
0
species. No evidence exists at all for an Fe oxidation state
(55) Hendrich, M. P.; Gunderson, W.; Behan, R. K.; Green, M. T.; Mehn,
(
54) Mashiko, T.; Reed, C. A.; Haller, K. J.; Scheidt, W. R. Inorg. Chem.
M. P.; Betley, T. A.; Lu, C. C.; Peters, J. C. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 17107.
1
984, 23, 3192.