Iron-arylimide clusters [Fem(NAr)nCl4]2- (m, n = 2, 2; 3, 4; 4, 4) from a ferric amide precursor: Synthesis, characterization, and comparison to Fe-S chemistry

Article abstract of DOI:10.1021/ic025825j

Tetrahedral FeCl[N(SiMe₃)₂]₂(THF) (2), prepared from FeCl₃ and 2 equiv of Na[N(SiMe₃)₂] in THF, is a useful ferric starting material for the synthesis of weak-field iron-imide (Fe-NR) clusters. Protonolysis of 2 with aniline yields azobenzene and [Fe₂(μ-Cl)₃(THF)₆]₂ [Fe₃(μ-NPh)₄Cl₄] (3), a salt composed of two diferrous monocations and a trinuclear dianion with a formal 2 Fe(III)/1 Fe(IV) oxidation state. Treatment of 2 with LiCl, which gives the adduct [FeCl₂{N(SiMe₃)₂}₂]^- (isolated as the [Li(TMEDA)₂]⁺ salt), suppresses arylamine oxidation/iron reduction chemistry during protonolysis. Thus, under appropriate conditions, the reaction of 1:1 2/LiCl with arylamine provides a practical route to the following Fe-NR clusters: [Li₂(THF)₇][Fe₃(μ-NPh) ₄Cl₄] (5a), which contains the same Fe-NR cluster found in 3; [Li(THF)₄]₂[Fe₃(μ-N-p-Tol) ₄Cl₄] (5b); [Li(DME)₃]₂[Fe₂(μ-NPh) ₂Cl₄] (6a); [Li₂(THF)₇][Fe₂(μ-NMes) ₂Cl₄] (6c). [Li(DME)₃]₂[Fe₄(μ ₃-NPh)₄Cl₄] (7), a trace product in the synthesis of 5a and 6a, forms readily as the sole Fe-NR complex upon reduction of these lower nuclearity clusters. Products were characterized by X-ray crystallographic analysis, by electronic absorption, ¹H NMR, and Moessbauer spectroscopies, and by cyclic voltammetry. The structures of the Fe-NR complexes derive from tetrahedral iron centers, edge-fused by imide bridges into linear arrays (5a,b; 6a,c) or the condensed heterocubane geometry (7), and are homologous to fundamental iron-sulfur (Fe-S) cluster motifs. The analogy to Fe-S chemistry also encompasses parallels between Fe-mediated redox transformations of nitrogen and sulfur ligands and reductive core conversions of linear dinuclear and trinuclear clusters to heterocubane species and is reinforced by other recent examples of iron- and cobalt-imide cluster chemistry. The correspondence of nitrogen and sulfur chemistry at iron is intriguing in the context of speculative Fe-mediated mechanisms for biological nitrogen fixation.

Full text of DOI:10.1021/ic025825j

Inorg. Chem. 2003, 42, 1211−1224
Iron−Arylimide Clusters [Fem(NAr)nCl4]2- (m, n ) 2, 2; 3, 4; 4, 4) from a
Ferric Amide Precursor: Synthesis, Characterization, and Comparison
to Fe−S Chemistry
Jeremiah S. Duncan, Tamim M. Nazif, Atul K. Verma, and Sonny C. Lee*
Department of Chemistry, Princeton UniVersity, Princeton, New Jersey 08544
Received June 27, 2002
Tetrahedral FeCl[N(SiMe3)2]2(THF) (2), prepared from FeCl3 and 2 equiv of Na[N(SiMe3)2] in THF, is a useful ferric
starting material for the synthesis of weak-field iron−imide (Fe−NR) clusters. Protonolysis of 2 with aniline yields
azobenzene and [Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4] (3), a salt composed of two diferrous monocations and a
trinuclear dianion with a formal 2 Fe(III)/1 Fe(IV) oxidation state. Treatment of 2 with LiCl, which gives the adduct
[FeCl2{N(SiMe3)2}2]- (isolated as the [Li(TMEDA)2]+ salt), suppresses arylamine oxidation/iron reduction chemistry
during protonolysis. Thus, under appropriate conditions, the reaction of 1:1 2/LiCl with arylamine provides a practical
route to the following Fe−NR clusters: [Li2(THF)7][Fe3(µ-NPh)4Cl4] (5a), which contains the same Fe−NR cluster
found in 3; [Li(THF)4]2[Fe3(µ-N-p-Tol)4Cl4] (5b); [Li(DME)3]2[Fe2(µ-NPh)2Cl4] (6a); [Li2(THF)7][Fe2(µ-NMes)2Cl4] (6c).
[Li(DME)3]2[Fe4(µ3-NPh)4Cl4] (7), a trace product in the synthesis of 5a and 6a, forms readily as the sole Fe−NR
complex upon reduction of these lower nuclearity clusters. Products were characterized by X-ray crystallographic
analysis, by electronic absorption, 1H NMR, and Mo¨ssbauer spectroscopies, and by cyclic voltammetry. The structures
of the Fe−NR complexes derive from tetrahedral iron centers, edge-fused by imide bridges into linear arrays (5a,b;
6a,c) or the condensed heterocubane geometry (7), and are homologous to fundamental iron−sulfur (Fe−S) cluster
motifs. The analogy to Fe−S chemistry also encompasses parallels between Fe-mediated redox transformations of
nitrogen and sulfur ligands and reductive core conversions of linear dinuclear and trinuclear clusters to heterocubane
species and is reinforced by other recent examples of iron− and cobalt−imide cluster chemistry. The correspondence
of nitrogen and sulfur chemistry at iron is intriguing in the context of speculative Fe-mediated mechanisms for
biological nitrogen fixation.
Introduction
cluster core during turnover. Experimental evidence3 for this
conjecture, however, is nonexistent.
We are exploring the chemistry of the Fe-N bond in
weak-field environments to gain insight into the proposed
enzymatic interactions of iron and nitrogen.4 Toward this
end, we have developed ferrous chemistry that accomplishes
The molecular description of biological nitrogen fixation
continues as a longstanding problem in inorganic chemistry.1
A unique weak-field [Fe7MoS9] cluster, the FeMo-cofactor,
is generally accepted as the site of dinitrogen reduction in
the best-studied form of the nitrogenase enzymes, yet the
means by which this cluster transforms substrate remains
very much a mystery. One line of speculation2,3 holds that
dinitrogen reacts as a bridging ligand to multiple iron sites
at the center of the cofactor, becoming, in effect, part of the
(2) Some competing mechanistic proposals: (a) Dance, I. J. Biol. Inorg.
Chem. 1996, 1, 581. (b) Sellmann, D.; Sutter, J. J. Biol. Inorg. Chem.
1996, 1, 587. (c) Coucouvanis, D. J. Biol. Inorg. Chem. 1996, 1, 594.
(d) Pickett, C. J. J. Biol. Inorg. Chem. 1996, 1, 601. (e) Leigh, G. J.
Eur. J. Biochem. 1995, 535, 171.
(3) Fe-mediated mechanisms have been the subject of considerable
computational study, as referenced and reviewed in Appendix 1 of
the following: (a) Lovell, T.; Li, J.; Liu, T.; Case, D. A.; Noodleman,
L. J. Am. Chem. Soc. 2001, 123, 12392. Also see: (b) Lovell, T.; Li,
J.; Case, D. A.; Noodleman, L. J. Am. Chem. Soc. 2002, 124, 4546.
(c) Zhong, S.-J.; Liu, C.-W. Polyhedron 1997, 16, 653. (d) Shestakov,
A. F. Russ. Chem. Bull. 1996, 45, 1827. (e) Machado, F. B. C.;
Davidson, E. R. Theor. Chim. Acta 1995, 92, 315. (f) Plass, W. J.
Mol. Struct. 1994, 315, 53.
* To whom correspondence should be addressed. E-mail: sclee@
chemvax.princeton.edu.
(1) Some recent reviews: (a) Christiansen, J.; Dean, D. R.; Seefeldt, L.
C. Annu. ReV. Plant Physiol. Plant Mol. Biol. 2001, 52, 269. (b) Rees,
D. C.; Howard, J. B. Curr. Opin. Chem. Biol. 2000, 4, 559. (c) Smith,
B. E. AdV. Inorg. Chem. 1999, 47, 159. (d) Howard, J. B.; Rees, D.
C. Chem. ReV. 1996, 96, 2965. (e) Burgess, B. K.; Lowe, D. J. Chem.
ReV. 1996, 96, 2983. (f) Eady, R. R. Chem. ReV. 1996, 96, 3013.
10.1021/ic025825j CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/23/2003
Inorganic Chemistry, Vol. 42, No. 4, 2003 1211

Duncan et al.
Scheme 1
Scheme 2
the 2-electron reduction of the N-N bond with concomitant
assembly of a ferric imide cluster (Scheme 1), a synthetic
transformation that might resemble the final stages of
dinitrogen reduction.5 The formation of iron-imide (Fe-
NR) clusters in this system and the possibility of cluster-
bound imides (or other anionic nitrogen fragments) as
intermediates in the latter stages of nitrogenase action have
prompted us to investigate the nature of Fe-NR clusters in
general.
We describe here a new synthetic approach to the Fe-
NR cluster manifold, with detailed characterization of the
reaction chemistry and products. Our findings provide insight
into the synthesis of Fe-NR clusters and reveal parallels
between Fe-NR and iron-sulfur (Fe-S) chemistry that may
have implications for the molecular description of biological
nitrogen fixation.
Results and Discussion
Synthetic Considerations. The preparative chemistry of
weak-field Fe-NR clusters is largely unexplored, comprising
just three published studies, all recent.4-7,8 In two accounts,
clusters were assembled from reactions of iron halides and
nitrogen anion (N-anion) reagents.6,7 Although the use of
these simple, accessible starting materials is synthetically
attractive, we have found that the combination of exchange-
labile, redox-active iron and aggressively reactive N-anions
gives complex, difficult-to-control chemistry.6
An alternative, redox-linked synthesis of ferric arylimide
clusters proceeds in good yield via bis(trimethylsilyl)amide
complexes of Fe(II) (Scheme 1).5 The [N(SiMe3)2]- ligand
promotes chemical selectivity in this system in two ways.
First, as a spectator ligand, its steric bulk enhances kinetic
control at the high-spin metal center. Second, as a functional
group, the coordinated amide serves as a powerful latent
base9 that can be replaced by other very basic ligands (e.g.,
imide) through protonolysis, thereby avoiding metathesis via
N-anions. With these tactics in mind, we investigated direct
routes to Fe-NR clusters using arylamines and sterically
hindered amide complexes of Fe(III), as summarized in
Scheme 2.
Ferric Amide Precursors. The ferric analogue of Fe-
[N(SiMe3)2]2 is trigonal planar Fe[N(SiMe3)2]3 (1),10 for
which two different syntheses are known. The original
preparation,10a,c conducted in arene solvents, works well
despite the nonintuitive11 2:3 FeCl3/Na[N(SiMe3)2] stoichi-
ometry. The other synthesis for tris(amide) 1sthe reaction
(4) Lee, S. C. Iron-Imide Clusters and Nitrogenase: Abiological Chemistry
of Biological Relevance? In Modern Coordination Chemistry: The
Legacy of Joseph Chatt; Leigh, G. J., Winterton, N., Eds.; Royal
Society of Chemistry: Cambridge, U.K., 2002; pp 278-287.
(5) Verma, A. K.; Lee, S. C. J. Am. Chem. Soc. 1999, 121, 10838.
(6) Verma, A. K.; Nazif, T. M.; Achim, C.; Lee, S. C. J. Am. Chem. Soc.
2000, 122, 11013.
(7) Link, H.; Decker, A.; Fenske, D. Z. Anorg. Allg. Chem. 2000, 626,
1567.
(8) A limited number of iron clusters with phosphiniminate ([R3PdN]-)
and sulfinimate ([R2SdN]-) core ligands are also known. Given that
the properties of these monoanionic ligand types differ considerably
from those of the dianionic imides of interest here, we exclude these
clusters from further consideration: (a) Riese, U.; Harms, K.; Pebler,
J.; Dehnicke, K. Z. Anorg. Allg. Chem. 1999, 625, 746. (b) Mai, H.-
J.; Wocadlo, S.; Kang, H.-C.; Massa, W.; Dehnicke, K.; Maichle-
Mo¨ssmer, C.; Stra¨hle, J.; Fenske, D. Z. Anorg. Allg. Chem. 1995, 621,
705. (c) Roesky, H. W.; Seseke, U.; Noltemeyer, M.; Sheldrick, G.
M. Z. Naturforsch. 1988, 43b, 1130. (d) Roesky, H. W.; Zimmer, M.;
Schmidt, H. G.; Noltemeyer, M. Z. Naturforsch. 1988, 43b, 1490.
(9) (a) Comparative pKa values in DMSO: MeOH, 29.0; n-BuSH, 17.0;
PhNH2, 30.6; PhOH, 18.0; PhSH, 10.3; Ph(H)NN(H)Ph, 26.2. Bor-
dwell, F. G. Acc. Chem. Res. 1988, 21, 456. (b) In THF:iPr2NH, 35.7.
(Me3Si)2NH, 25.8. Fraser, R. R.; Mansour, T. S. J. Org. Chem. 1984,
49, 3443. Also see correction: J. Org. Chem. 1984, 49, 5284.
1212 Inorganic Chemistry, Vol. 42, No. 4, 2003

Iron-Arylimide Clusters [Fem(NAr)nCl4]2-
Table 1. Spectroscopic Data for Ferric Amide Complexes
1H NMR:b δ, ppm
THF or TMEDA
complex
Fe[N(SiMe3)2]3 (1)c
FeCl[N(SiMe3)2]2(THF) (2)
[Li(TMEDA)2][FeCl2{N(SiMe3)2}2] (3)
electronic abs:a λ, nm (ꢀM, L‚mol-1‚cm-1
)
SiMe3
345 (sh, 3500), 402 (5000), 498 (sh, 1500), 631 (1600)
267 (6500), 433 (4500), 536 (sh, 1700)
269 (6900), 410 (3800), 474 (sh, 2500)
33.5 (br)
15.5 (br)
14.6 (br)
43.0 (br), ca. 82 (v br)
2.3 (br), 2.7 (br sh)
a n-Pentane (1) or THF (2, 3) solution. b C6D6 solution, ca. 25 °C. c Our extinction coefficients are three times higher than literature values,10d,e due
1
perhaps to sample decomposition in the original measurements; to our knowledge, H NMR data have not been reported previously for 1.
of 1:3 FeCl3/Li[N(SiMe3)2] in THFsis not as straightforward
as the literature implies.10b We have observed that, at room
temperature, the reaction halts at a persistent deep red
solution color which is obviously not the dark green of 1;
reflux is required to establish the expected color. Workup
of the reaction at the intermediary red stage leads to the
isolation of highly crystalline, dark-red FeCl[N(SiMe3)2]2-
(THF) (2).
Complex 2 is obtained directly in high yield (ca. 75%) by
anion metathesis of FeCl3 with exactly 2 equiv of Na-
[N(SiMe3)2]12 in THF. The coordinated solvent is labile13
and easily replaced by chloride; thus, dissolution of 1 equiv
of LiCl in a deep red THF solution of bis(amide) 2 leads to
a lighter, transparent red solution, from which, in the presence
of TMEDA,14 crystalline [Li(TMEDA)2][FeCl2{N(SiMe3)2}2]
(3) can be isolated in quantity. Solution spectroscopic data
for all three mononuclear ferric amide species (1-3) are
presented for comparison in Table 1 and Figure 1. The color
change upon LiCl addition is evident in the electronic
Figure 1. Comparative electronic absorption spectra of mononuclear ferric
amide complexes in n-pentane (1) or THF (2, 3).
absorption spectrum of 3, which, in the visible range, is
hypsochromic and slightly hypochromic relative to that of
2 and 3;15-18 aside from the present ferric complexes,
however, we know of only four other examples of this ligand
sphere involving transition elements (Cr(II) and Mn(II)) later
than group 5.19
2. Both bis(amide) complexes are characterized by broad,
1
paramagnetically shifted H NMR signals, as expected for
isolated high-spin ferric centers, with the methyl resonance
of the anionic derivative located slightly upfield. The
sterically derived kinetic stability of the heteroleptic, tetra-
hedral [M{N(SiMe3)2}2(L)(L′)]z environment is well-docu-
mented in early metals (groups 2-5 and f-elements), and
Ti(III) and V(III) species exist that are isostructural to both
Crystallographic analyses of complexes 2 and 3 reveal
related distorted tetrahedral iron sites (Table 2) that differ
principally in the substitution of a THF ligand (Figure 2) by
chloride (Figure 3). The steric demand of the amide ligands
expands N-Fe-N and compresses Cl-Fe-O (2) or Cl-
Fe-Cl (3) angles relative to the tetrahedral ideal, although
replacement of the THF donor by the smaller, more distant
chloride relieves some congestion, resulting in less pro-
nounced angular distortions in 3. The Fe-N contacts of 2
are similar to terminal iron-amide bond distances in
tetrahedral environments20 and in 1 (1.92 Å);10h,i Fe-N and
(10) Syntheses: (a) Bu¨rger, H.; Wannagat, U. Monatsh. Chem. 94, 1963,
1007. (b) Alyea, E. C.; Bradley, D. C.; Copperthwaite, R. G. J. Chem.
Soc., Dalton Trans. 1972, 1580. (c) Bradley, D. C.; Copperthwaite,
R. G. Inorg. Synth. 1978, 18, 112. Spectroscopic and magnetic
properties: (d) Alyea, E. C.; Bradley, D. C.; Copperthwaite, R. G.;
Sales, K. D.; Fitzsimmons, B. W.; Johnson, C. E. Chem. Commun.
1970, 1715. (e) Alyea, E. C.; Bradley, D. C.; Copperthwaite, R. G.;
Sales, K. D. J. Chem. Soc., Dalton Trans. 1973, 185. (f) Bradley, D.
C.; Copperthwaite, R. G.; Cotton, S. A.; Sales, K. D.; Gibson, J. F. J.
Chem. Soc., Dalton Trans. 1973, 191. (g) Fitzsimmons, B. W.;
Johnson, C. E. Chem. Phys. Lett. 1974, 24, 422. Crystal structure:
(h) Bradley, D. C.; Hursthouse, M. B.; Rodesiler, P. F. Chem. Commun.
1969, 14. (i) Hursthouse, M. B.; Rodesiler, P. F. J. Chem. Soc., Dalton
Trans. 1972, 2100.
(11) Reaction at the limiting stoichiometry of 1:3 FeCl3/Na[N(SiMe3)2]
gives impurities that are difficult to separate from product.
(12) The use of the lithium salt is undesirable due to the formation of LiCl,
which can contaminate the product even after pentane extraction and
crystallization.
(13) Chemical exchange is evident upon doping of NMR solutions of 2
(C6D6) with THF: the resonances associated with bound THF
disappear and are replaced by broadened signals at the diamagnetic
shift positions of free THF. Despite the THF lability, 2 is recovered
unchanged after multiple, successive recrystallizations from n-pentane.
(14) Adduct formation is reversible and crystallization attempts in the
absence of TMEDA led to isolation of neutral 1.
(15) TiCl[N(SiMe3)2]2(THF): Scoles, L.; Gambarotta, S. Inorg. Chim. Acta
1995, 235, 375.
(16) [TiCl2{N(SiMe3)2}2]-: (a) Duchateau, R.; Gambarotta, S.; Beydoun,
N.; Bensimon, C. J. Am. Chem. Soc. 1991, 113, 8986. (b) Beydoun,
N.; Duchateau, R.; Gambarotta, S. J. Chem. Soc., Chem. Commun.
1992, 244. (c) Putzer, M. A.; Magull, J.; Goesmann, H.; Neumuller,
B.; Dehnicke, K. Chem. Ber. 1996, 129, 1401.
(17) VCl[N(SiMe3)2]2(THF): Berno, P.; Moore, M.; Minhas, R.; Gam-
barotta, S. Can. J. Chem. 1996, 74, 1930.
(18) [VCl2{N(SiMe3)2}2]-: Gerlach, C. P.; Arnold, J. Organometallics
1997, 16, 5148.
(19) Cr(II) (planar): (a) Bradley, D. C.; Hursthouse, M. B.; Newing, C.
W.; Welch, A. J. J. Chem. Soc., Chem. Commun. 1972, 567. Mn(II):
(b) Bradley, D. C.; Hursthouse, M. B.; Ibrahim, A. A.; Malik, K. M.
A.; Motevalli, M.; Moseler, R.; Powell, H.; Runnacles, J. D.; Sullivan,
A. C. Polyhedron 1990, 9, 2959. (c) Andruh, M.; Roesky, H. W.;
Noltemeyer, M.; Schmidt, H.-G. Z. Naturforsch., B 1994, 49, 31.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1213

Duncan et al.
Table 2. Selected Interatomic Distances (Å) and Angles (deg) in
FeCl[N(SiMe3)2]2(THF) (2) and [Li(TMEDA)2][FeCl2{N(SiMe3)2}2] (3)a
2 [X ) O(1)]
3 [X ) Cl(2/4)]
Fe(1/2)-N(1/3)
Fe(1/2)-N(2/4)
Fe(1/2)-Cl(1/3)
Fe(1/2)-X
1.902(2)
1.913(2)
2.2510(7)
2.043(2)
1.925(2)/1.918(2)
1.930(2)/1.921(2)
2.279(1)/2.276(1)
2.269(1)/2.283(1)
N(1/3)-Fe(1/2)-N(2/4)
N(1/3)-Fe(1/2)-Cl(1/3)
N(2/4)-Fe(1/2)-Cl(1/3)
N(1/3)-Fe(1/2)-X
123.16(9)
106.49(7)
114.96(6)
112.54(8)
98.96(8)
97.61(5)
120.2(1)/119.4(1)
103.90(9)/110.23(8)
111.38(8)/105.20(8)
110.47(9)/103.65(8)
105.15(9)/112.82(8)
104.79(4)/104.69(4)
N(2/4)-Fe(1/2)-X
Cl(1/3)-Fe(1/2)-X
Planarity
0.157(2)
0.002(2)
N(1/3)b
N(2/4)b
0.172(3)/0.157(3)
0.020(3)/0.017(3)
Figure 4. Cation structure of [Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4] (4‚
THF) with thermal ellipsoids (35% probability level) and selected atom
labels. Only one of two independent cations is presented. Selected distance
ranges [mean] (Å) for the dimers: Fe-O, 2.105(5)-2.172(5) [2.14(2)];
Fe-Cl, 2.455(2)-2.513(2) [2.48(2)]; Fe‚‚‚Fe, 3.051(2), 3.081(1) [3.07(2)].
a Divided entries refer to equivalent atoms and metrics of the two
independent complexes in 3 in the order given, e.g., N(1/3)-Fe(1/2)-N(2/
4) denotes two angles, N(1)-Fe(1)-N(2) and N(3)-Fe(2)-N(4); for the
unique molecule in 2, only the first atom number is used. b Perpendicular
displacement from the [Fe, Si, Si′] plane.
distortion appears to correlate with eclipsing strain (or its
avoidance) in the pyramidalized amide;21 the Me3Si-N-
SiMe3 wedge of this amide is almost coplanar with the THF
oxygen (in 2) or chloride (in 3) when viewed along the Fe-N
bond, while the planar amide adopts a perpendicular clinal
conformation with respect to the same fiducial group.
In preliminary reactivity assays, tris(amide) 1 was found
to be unsuitable as a protonolysis precursor, with test
reactions returning the starting complex along with intrac-
table, presumably polymeric material. This behavior suggests
that the steric stability of 1 is such that the initial products
of protonolysis, by virtue of reduced steric encumbrance,
react further and much faster relative to remaining 1. We
therefore turned our attention to the less-hindered, bis(amide)
complex 2.
Protonolysis by Aniline. In principle, the protonolysis of
2 with 1 equiv of aniline provides a plausible stoichiometric
route (eq 1) to a neutral Fe-NR heterocubane. In practice,
a more complicated chemistry ensues. The addition of aniline
to complex 2 in THF results in a rapid color change from
deep red to black, with formation of insoluble dark precipi-
tate. The THF-soluble components can be partitioned into
an orange, pentane-soluble fraction consisting almost exclu-
sively of azobenzene and a black, iron-containing solid that
is insoluble in both n-pentane and benzene. Slow crystal-
lization of this dark, THF-soluble material affords the cluster
salt [Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4] (4).
Figure 2. Structure of FeCl[N(SiMe3)2]2(THF) (2) with thermal ellipsoids
(50% probability level) and selected atom labels. The THF ligand is partly
disordered, and only one component is presented. For clarity in this and all
subsequent structural depictions, carbon ellipsoids are presented without
shading and hydrogen atoms are not shown.
4FeCl[N(SiMe3)2]2(THF) (2) + 4PhNH2 f
Fe4(µ3-NPh)4Cl4 + 8(Me3Si)2NH + 4THF (1)
Figure 3. Anion structure of [Li(TMEDA)2][FeCl2{N(SiMe3)2}2] (3) with
thermal ellipsoids (50% probability level) and selected atom labels. Only
one of two independent complexes is presented.
Crystallographic structure analysis discloses two different
cluster species in 4. The first type consists of two octahedral
iron centers, confacially bridged through three chloride
ligands and terminally ligated by six THF donors (Figure
4). This structure has been observed elsewhere as a diferrous
Fe-Cl distances in anionic 3 are slightly longer, as expected
from simple electrostatic considerations. In both complexes,
the nitrogen atom of one amide is near planar, while the
other amide nitrogen is somewhat pyramidalized. This
(21) The complexity of interactions and degrees of freedom within the
[M{N(SiMe3)2}2(L)(L′)]z coordination sphere precludes any simple
origin to the amide pyramidalization.
(20) Stokes, S. L.; Davis, W. M.; Odom, A. L.; Cummins, C. C.
Organometallics 1996, 15, 4521.
1214 Inorganic Chemistry, Vol. 42, No. 4, 2003

Iron-Arylimide Clusters [Fem(NAr)nCl4]2-
relative to the original reaction without LiCl. Differences
are immediately apparent: in the presence of LiCl, darkening
of the reaction solution to opacity is slower (minutes vs
seconds) and gives a different endpoint color (dark purple
vs black), while the formation of insoluble solid and
azobenzene is reduced to trace levels (<1% yield for
azobenzene). Spectroscopic and crystallographic character-
ization of this modified reaction system (vide infra) reveals
a product mixture consisting primarily of two Fe-NR
components, dinuclear [Fe2(µ-NPh)2Cl4]2- and trinuclear
[Fe3(µ-NPh)4Cl4]2-, in roughly equimolar ratios; in addition,
the tetranuclear heterocubane [Fe4(µ3-NPh)4Cl4]2- forms as
a minor byproduct (ca. 5 mol % relative to either the di- or
trinuclear species). Only the triiron complex, the least soluble
of the three clusters, has proven separable in quantity from
this mixture as deep-brown, crystalline [Li2(THF)7][Fe3(µ-
NPh)4Cl4] (5a).
In DME solvent, the diiron phenylimide cluster forms as
the predominant product (> 9:1 vs the triiron cluster),
precipitating directly en masse as purple-black microcrystals
of [Li(DME)3]2[Fe2(µ-NPh)2Cl4] (6a). The immediate re-
moval of poorly soluble 6a from this reaction system may
explain the limited formation of the trinuclear coproduct
relative to the THF system. By contrast, the reaction of 1:1
2/LiCl with p-toluidine gives an approximate 1:1 mixture
of the analogous di- and trinuclear clusters in either THF or
DME solvent; here, the solubilizing effect of the p-Me group
apparently allows sufficient solution concentrations to form
the larger cluster in either solvent. The lower solubility of
the triiron cluster again allows its bulk isolation by fractional
crystallization as black [Li(THF)4]2[Fe3(µ-N-p-Tol)4Cl4] (5b);
the diiron coproduct, formulated as Li2[Fe2(µ-N-p-Tol)2Cl4]
(6b), was identified in situ by NMR comparison with its
phenylimide congener 6a. The dinuclear Fe-NR core can
also be stabilized through steric effects. Thus, protonolysis
of 1:1 2/LiCl with mesidine in THF proceeds smoothly to
give deep green solutions consisting exclusively of the diiron
cluster, which can be crystallized in high yield as dark green
[Li2(THF)7][Fe2(µ-NMes)2Cl4] (6c). The absence of triiron
product is attributed to destabilizing steric contacts that would
exist between o-Me groups of adjacent mesitylimide ligands
in the trinuclear geometry. Last, without LiCl, the reactions
of p-toluidine or mesidine with 2 result in undefined iron-
containing species and significant formation of the corre-
sponding azoarenes.
Chemistry of Cluster Formation. The LiCl co-reagent
supplies soluble, inert countercation (Li+) and supplementary
ligand (Cl-) to the reaction system. While both components
contribute to the synthesis and isolation of tractable Fe-
NR species, the introduction of additional chloride appears
central in regulating cluster assembly. As observed in 3,
available chloride is expected to cap metal coordination sites
by replacing more labile solvent donors. The consequent
stabilization of coordination spheres and inhibition of ligand
exchange explain the slower rates for protonolysis, which
presumably requires precoordination of incoming amine, as
well as the near absence of insoluble, polymeric products.
The inhibition of Fe(III)/arylamine redox chemistry (i.e.,
Figure 5. Anion structure of [Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4] (4‚THF)
with thermal ellipsoids (35% probability level) and selected atom labels.
monocation.22 The other cluster in compound 4 is a linear
trinuclear array of tetrahedral iron centers with phenylimide
bridges and terminal chloride ligands (Figure 5), details of
which are presented later in comparison with other Fe-NR
clusters. Two independent clusters of the dinuclear type and
one trinuclear cluster occupy the asymmetric unit of the cell.
The assignment of cluster charge in 4 merits some
consideration. The two [Fe2(µ-Cl)3(THF)6] components are
metrically indistiguishable from the known diferrous mono-
cation,22 suggesting the same +1 charge for each on structural
grounds. If this is correct, the [Fe3(µ-NPh)4Cl4] constituent
must be a dianion with an unexpected 2 Fe(III)/1 Fe(IV)
formal oxidation state. Conclusive evidence for this charge
assignment comes from the isolation of [Fe3(µ-NPh)4Cl4]2-
as an unambiguous dilithium salt (6a; vide infra) with an
1H NMR signature identical to the predominant signal set
found in crude 4.
It is obvious that the reaction of the ferric amide precursor
with aniline extends beyond simple protonolysis. Although
the full mass balance for this reaction is unknown, the yield
of azobenzene is sufficient to reduce ca. 22% of the starting
Fe(III) to Fe(II); the formation of azobenzene as a significant
coproduct thus implicates aniline, probably via ligated anilide
or phenylimide, as the principal reductant in the system. By
itself, the unexpected redox behavior need not preclude this
synthesis as a viable route to Fe-NR chemistry. The redox-
derived cation in 4, however, is a source of chemically and
spectroscopically active iron that hinders the study of the
cluster of interest, and we therefore sought synthetic modi-
fications to obtain more tractable product. Lithium chloride
was effective for this purpose.
Practical Syntheses of Fe-NR Clusters. Treatment of a
THF solution of 1:1 2/LiCl (which presumably contains
[FeCl2{N(SiMe3)2}2]-, based on the preparation of 3)23 with
1 equiv of aniline gives significantly altered chemistry
(22) Janas, Z.; Sobota, P.; Lis, T. J. Chem. Soc., Dalton Trans. 1991, 2429.
(23) The direct use of 3 in cluster assembly was excluded in this study to
avoid chemistry involving TMEDA and iron.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1215

Duncan et al.
azoarene production) is also consistent with adduct formation,
since the resulting anionic chloroferrate complexes should
be less oxidizing than comparable neutral species. This is
demonstrated in the cyclic voltammetry of 2 and 3, which
shows grossly irreversible reductions for both but with the
peak potential of the anion shifted negative by 650 mV (0.1
c
M TBAP/MeCN, Ep vs SCE: 2, -0.40 V; 3, -1.05 V).
The use of LiCl does not eliminate metal-based redox
chemistry altogether. Of the three cluster types isolated, only
the diiron cluster represents a stoichiometrically exact
protonolysis product (eq 2), yet it forms exclusively only
with solubility or steric constraint. The triiron cluster, the
coproduct under general conditions, is more oxidized than
the starting materials, and its synthesis consequently entails
intervening redox chemistry. We can offer a balanced,
hypothetical reaction, using exact experimental reactant
ratios, that couples the one-electron oxidation of the trinuclear
cluster to the reduction of Fe(III) to Fe(II) (eq 3); the ligands
are assumed to segregate as in cluster salt 4 such that imide
donors ligate only to oxidized iron sites. Given the difficulties
in tracking mass balance for this chemistry, it is not
surprising that the minor24 Fe(II) product predicted by eq 3
escapes detection, particularly if it occurs in the form of
spectroscopically silent chloroferrous species.
Figure 6. Anion structure of [Li(DME)3]2[Fe2(µ-NPh)2Cl4] (6a) with
thermal ellipsoids (50% probability level) and selected atom labels. Atoms
with labels ending in “A” are generated by a crystallographic inversion
center.
Fe-NR Cluster Structures. The Fe-NR clusters in this
investigation all derive from tetrahedral iron centers that are
edge-fused either through µ2-imide bridges to form linear
di- and trinuclear arrays or through µ3-imide bridges to give
the condensed tetranuclear heterocubane geometry; terminal
chloride donors complete the ligand sets. The cluster types
can be viewed in terms of a shared [Fe2N2] rhombic motif
(where N is the imide nitrogen) that completely defines the
cores of the dinuclear clusters and constitutes the trimeric
and heterocubane cluster geometries through vertex and edge
fusion, respectively. The [Fe2N2] fragments are rigorously
planar for the dinuclear clusters due to lattice symmetry and
near-planar for the others. The rhombs also approach mutual
coplanarity with unhindered aryl subsituents as well, which
argues for some degree of extended conjugation between the
aryl π-orbitals, the imide lone pairs, and the iron centers.
On average, the [Fe2N2] polygons are almost square and
exhibit only small rhombic angular distortions, with the
exception of the trinuclear clusters described below. Cubane
7 has somewhat expanded [Fe2N2] subunits due to lengthened
Fe-N bonds from µ3-imide bridging.
The structures of the dinuclear phenylimide and mesityl-
imide clusters possess crystallographic inversion symmetry,
with a unique half-dimer in the asymmetric unit of 6a (Figure
6) and two independent half-dimers for 6c (Figure 7). The
three structures show comparable interatomic metrics (Table
3) and differ significantly only in the orientation of the imide
aryl substituents, which are coplanar with the [Fe2N2]
rhombic core in the phenyl cases but perpendicular in the
mesityl derivative due to the steric demand of the o-Me
groups. For one of the half-dimers in 6c (Figure 7), a partially
solvated [Li(THF)3]+ countercation interacts with one chlo-
ride ligand, leading to a moderately elongated Fe-Cl bond;
all other cluster dianions are accompanied by well-separated,
fully solvated counterions in the lattice.
2FeCl[N(SiMe3)2]2(THF) (2) + 2LiCl + 2ArNH2 f
Li2[Fe2(µ-NAr)2Cl4] + 4(Me3Si)2NH + 2THF (2)
4FeCl[N(SiMe3)2]2(THF) (2) + 4LiCl + 4ArNH2 f
Li2[Fe3(µ-NAr)4Cl4] + Li2[FeCl4] + 8(Me3Si)2NH +
4THF (3)
A small fraction of the Fe(II) is probably associated with
the limited formation of the cubane dianion [Fe4(µ3-NPh)4-
Cl4]2-, which is reduced by two electrons relative to the
neutral homologue expected from protonolysis alone (eq 1).
Reducing conditions should therefore favor production of
the cubane dianion, and indeed [Li(THF)4]2[Fe4(µ3-NPh)4-
Cl4] (7) is obtained as the sole soluble Fe-NR product when
an exogenous reductant (Zn dust) is added to solutions of
dinuclear 6a, trinuclear 5a, or crude 1:1 reaction mixtures
of both.
We note finally that the dinuclear and cubane cluster
dianions in 6a,c and 7 have also been prepared, in different
counterion form, from ill-defined reactions of iron chlorides
and N-anion reagents.7 These compounds were crystallo-
graphically characterized and possess Fe-NR cluster com-
ponents that are metrically equivalent to the analogous
structures described in the next section. Although synthetic
analyses and solution characterizations were completely
absent in the original account, we have ascertained, using
spectroscopic data derived herein, that the underlying
chemistry is quite complicated; the products reported do not,
in fact, reflect the distribution of species actually generated
by the reactions described.
Tetranuclear 7 (Figure 8) is one of several heterocubane
clusters known in Fe-NR chemistry. The [Fe4(µ3-NR)4]z
cores occur in a range of oxidation states: in addition to the
z ) 2+ state in 7,7 the z ) 4+ core has been reported with
arylimide bridges and terminal thiolate donors5 and the z )
3+, 4+, and 5+ states have been reported with tert-butyl-
imide bridges and terminal chloride donors (a terminal imide
donor is also present in the most oxidized example).6,7 Cluster
(24) If the di- and trinuclear clusters are assumed as the sole products,
forming in equimolar ratios according to eqs 2 and 3, one-sixth (17%)
of the total iron content is predicted to be in the ferrous state.
1216 Inorganic Chemistry, Vol. 42, No. 4, 2003

Iron-Arylimide Clusters [Fem(NAr)nCl4]2-
Figure 7. Structure of [Li2(THF)7][Fe2(µ-NMes)2Cl4] (6c) with thermal
ellipsoids (50% probability level) and selected atom labels. Atoms with
labels ending in “A” are generated by a crystallographic inversion center.
One of two independent half-dimers is presented; the other dianion is well-
separated from [Li(THF)4]+ cations and has no Li‚‚‚Cl interactions.
Figure 8. Anion structure of [Li(DME)3]2[Fe4(µ3-NPh)4Cl4] (7‚DME) with
thermal ellipsoids (50% probability level) and selected atom labels.
Table 3. Selected Interatomic Distances (Å) and Angles (deg) in
[Li(DME)3]2[Fe2(µ-NPh)2Cl4] (6a) and [Li2(THF)7][[Fe2(µ-NMes)2Cl4]
(6c)a
Table 4. Selected Interatomic Distances (Å) and Angles (deg) in
[Li(DME)3]2[Fe4(µ3-NPh)4Cl4] (7‚DME)a
Fe(1)-N(1/2/3)
1.974(4)/1.986(4)/1.986(4)
1.980(4)/1.984(4)/1.978(4)
2.000(4)/1.976(4)/1.988(4)
1.968(4)/1.994(4)/1.978(4)
2.217(2)/2.223(2)/2.228(2)/2.228(2)
1.416(7)/1.404(6)/1.406(6)/1.404(7)
2.636(1)/2.608(1)/2.671(1)
2.689(1)/2.619(1)/2.681(1)
2.949(5)/2.988(6)/2.912(6)
2.918(6)/2.960(6)/2.916(6)
96.3(2)/98.0(2)/94.6(2)
96.1(2)/94.7(2)/96.7(2)
97.5(2)/93.8(2)/94.7(2)
94.9(2)/97.2(2)/94.5(2)
119.5(1)/124.3(1)/118.3(1)
122.8(1)/120.6(1)/119.6(1)
120.9(1)/122.2(1)/121.1(1)
119.3(1)/126.4(1)/118.1(1)
83.6(2)/82.0(2)/85.0(2)
83.2(2)/85.0(2)/83.0(2)
82.3(2)/84.3(2)/85.0(2)
6a
6c
Fe(2)-N(1/2/4)
Fe(1/2)-N(1/2)
Fe(1/2)-N(1A/2A)
Fe(1/2)-Cl(1/3)
Fe(1/2)-Cl(2/4)
Fe(1/2)‚‚‚Fe(1A/2A)
N(1/2)‚‚‚N(1A/2A)
N(1/2)-Cipso
1.880(2)
1.881(2)
2.2703(8)
2.2647(8)
2.5517(7)
2.763(4)
1.376(3)
1.867(2)/1.871(2)
1.884(2)/1.879(2)
2.2619(6)/2.2901(6)
2.3191(5)/2.2838(6)
2.5188(5)/2.5506(5)
2.779(3)/2.750(3)
1.389(2)/1.385(2)
Fe(3)-N(1/3/4)
Fe(4)-N(2/3/4)
Fe(1/2/3/4)-Cl(1/2/3/4)
N(1/2/3/4)-Cipso
Fe(1)‚‚‚Fe(2/3/4)
Fe(2/2/3)‚‚‚Fe(3/4/4)
N(1)‚‚‚N(2/3/4)
N(1/2)-Fe(1/2)-N(1A/2A)
Cl(1/3)-Fe(1/2)-Cl(2/4)
N(1/2)-Fe(1/2)-Cl(1/3)
N(1/2)-Fe(1/2)-Cl(2/4)
N(1A/2A)-Fe(1/2)-Cl(1/3)
N(1A/2A)-Fe(1/2)-Cl(2/4)
Fe(1/2)-N(1/2)-Fe(1A/2A)
Fe(1/2)-N(1/2)-Cipso
94.55(9)
107.00(3)
113.38(7)
113.87(7)
112.82(7)
115.12(7)
85.45(9)
137.0(2)
137.5(2)
95.63(7)/94.30(7)
106.01(5)/108.99(5)
117.99(5)/114.18(5)
110.16(5)/115.05(5)
115.76(5)/113.23(5)
112.26(5)/115.80(5)
84.37(7)/85.70(7)
141.6(1)/138.3(1)
134.0(1)/136.0(1)
N(2/2/3)‚‚‚N(3/4/4)
N(1/1/2)-Fe(1)-N(2/3/3)
N(1/1/2)-Fe(2)-N(2/4/4)
N(1/1/3)-Fe(3)-N(3/4/4)
N(2/2/3)-Fe(4)-N(3/4/4)
Cl(1)-Fe(1)-N(1/2/3)
Cl(2)-Fe(2)-N(1/2/4)
Cl(3)-Fe(3)-N(1/3/4)
Cl(4)-Fe(4)-N(2/3/4)
Fe(1/1/2)-N(1)-Fe(2/3/3)
Fe(1/1/2)-N(2)-Fe(2/4/4)
Fe(1/1/3)-N(3)-Fe(3/4/4)
Fe(2/2/3)-N(4)-Fe(3/4/4)
Cipso-N(1)-Fe(1/2/3)
Cipso-N(2)-Fe(1/2/4)
Cipso-N(3)-Fe(1/3/4)
Cipso-N(4)-Fe(2/3/4)
Fe(2/2/3)‚‚‚Fe(1)‚‚‚Fe(3/4/4)
Fe(1/1/3)‚‚‚Fe(2)‚‚‚Fe(3/4/4)
Fe(1/1/2)‚‚‚Fe(3)‚‚‚Fe(2/4/4)
Fe(1/1/2)‚‚‚Fe(4)‚‚‚Fe(2/3/3)
Fe(1A/2A)-N(1/2)-Cipso
Planarity and Torsion
0.027(3)
N(1/2)b
0.022(2)/0.015(2)
88.00(5)/89.24(4)
N(1/2)-Arc
5.15(3)
85.4(2)/82.9(2)/85.1(2)
a Divided entries refer to equivalent atoms and metrics of the two
independent half-dimers in 6c in the order given, e.g., N(1/2)-Fe(1/2)-
N(1A/2A) denotes two angles, N(1)-Fe(1)-N(1A) and N(2)-Fe(2)-
N(2A); for the single half-dimer in 6a, only the first atom number is used.
b Perpendicular displacement from the [Fe, Fe′, Cipso] plane. c Dihedral angle
between the planar [Fe2N2] rhomb and the least-squares-fitted aryl ring plane
linked at the specified nitrogen.
123.6(3)/131.7(3)/133.1(3)
129.1(3)/125.6(3)/133.7(3)
121.1(3)/133.4(3)/132.9(3)
126.0(3)/134.8(3)/125.9(3)
61.68(3)/59.14(3)/61.03(3)
58.65(3)/61.10(3)/60.67(3)
59.67(3)/60.64(3)/58.38(3)
59.77(3)/58.33(3)/60.95(3)
core metrics generally expand upon reduction across the
alkylimide series and upon subsitution of arylimide for alkyl-
imide; cubane 7 follows these trends in possessing the longest
Fe-N, Fe‚‚‚Fe, and Fe-Cl contacts of all (Table 4).
The linear trinuclear core found in 4 and 5a,b represents
a new Fe-NR geometry. The structures of 4 (Figure 5) and
5b (Figure 9) contain [Fe3(µ-NAr)4Cl4]2- clusters with near
identical core metrics (Table 5). In 5b, the presence of two
well-separated, [Li(THF)4]+ cations definitively establishes
the dianionic charge of the trinuclear cluster, giving a formal
oxidation state assignment of 2 Fe(III)/1 Fe(IV) to that
complex and supporting the charges assigned previously for
4. The µ2-imide bridging is asymmetric, with Fe-N contacts
about the central iron much shorter than those at the outer
iron centers, which are comparable to equivalent bond lengths
in the diferric dimers. The stabilization of the rare Fe(IV)
Planarity
0.057(2)
0.026(2)
0.072(1)
[Fe(1,2), N(1,2)]b
[Fe(1,3), N(1,3)]b
[Fe(1,4), N(2,3)]b
[Fe(2,3), N(1,4)]b
0.069(1)
0.026(2)
0.057(2)
[Fe(2,4), N(2,4)]b
[Fe(3,4), N(3,4)]b
a Divided entries refer to separate, related atoms and their associated
metrics in the order given, e.g., N(1/1/2)-Fe(1)-N(2/3/3) denotes three
angles, N(1)-Fe(1)-N(2), N(1)-Fe(1)-N(3), and N(2)-Fe(1)-N(3).
b Measured as the rms deviation from the least-squares-fitted [Fe2N2] rhomb
plane.
state by N-anion ligands is well-documented;6,25 the short
Fe-N distances might therefore indicate localization of the
(25) Some examples: (a) Ju¨stel, T.; Mu¨ller, M.; Weyhermu¨ller, T.; Kressl,
C.; Bill, E.; Hildebrandt, P.; Lengen, M.; Grodzicki, M.; Trautwein,
A. X.; Nuber, B.; Wieghardt, K. Chem.sEur. J. 1999, 5, 793. (b)
Cummins, C. C.; Schrock, R. R. Inorg. Chem. 1994, 33, 395. (c)
Kostka, K. L.; Fox, B. G.; Hendrich, M. P.; Collins, T. J.; Rickard,
C. E. F.; Wright, L. J.; Mu¨nck, E. J. Am. Chem. Soc. 1993, 115, 6746.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1217

Duncan et al.
Table 5. Selected Interatomic Distances (Å) and Angles (deg) in the
[Fe3(µ-NAr)4Cl4]2- Clusters of
[Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4]‚THF (4‚THF) and
[Li(THF)4]2[[Fe3(µ-N-p-Tol)4Cl4] (5b)a
4‚THF
5b
Fe(1)-N(1/2)
Fe(3)-N(3/4)
Fe(2)-N(1/2)
Fe(2)-N(3/4)
Fe(1)-Cl(1/2)
Fe(3)-Cl(3/4)
Fe(1/2)‚‚‚Fe(2/3)
N(1/3)‚‚‚N(2/4)
N(1/2)-Cipso
1.899(5)/1.892(6)
1.903(6)/1.903(6)
1.792(6)/1.797(5)
1.805(6)/1.793(6)
2.257(2)/2.262(2)
2.252(2)/2.247(2)
2.591(1)/2.582(1)
2.625(8)/2.652(9)
1.389(8)/1.383(9)
1.360(9)/1.379(9)
1.903(3)/1.896(3)
1.893(3)/1.886(3)
1.797(3)/1.799(3)
1.798(3)/1.805(3)
2.277(1)/2.266(1)
2.255(1)/2.246(1)
2.5963(7)/2.5814(7)
2.630(4)/2.637(5)
1.391(5)/1.392(5)
1.386(5)/1.384(5)
N(3/4)-Cipso
N(1/3)-Fe(1/3)-N(2/4)
87.7(2)/88.3(2)
87.6(1)/88.5(1)
Cl(1/3)-Fe(1/3)-Cl(2/4) 108.40(9)/106.75(9) 106.01(5)/108.99(5)
N(1)-Fe(1)-Cl(1/2)
N(2)Fe(1)-Cl(1/2)
117.1(2)/113.7(2)
113.3(2)/115.8(2)
114.1(2)/116.2(2)
115.1(2)/115.9(2)
94.0(3)/94.9(3)
117.1(3)/118.1(3)
118.5(3)/116.1(3)
89.1(2)/89.2(3)
88.2(3)/88.6(3)
143.0(5)/142.9(5)
127.8(5)/127.5(5)
144.2(5)/142.8(5)
127.6(5)/128.6(5)
179.56(6)
115.1(1)/116.1(1)
116.9(1)/114.9(1)
115.0(1)/115.0(1)
112.4(1)/115.9(1)
94.0(1)/94.1(2)
116.2(1)/119.6(1)
117.8(1)/117.2(1)
89.1(1)/89.2(1)
88.7(2)/88.7(2)
142.4(2)/143.1(3)
127.6(2)/131.3(3)
140.0(3)/138.8(2)
131.3(3)/132.4(3)
176.89(3)
N(3)-Fe(3)-Cl(3/4)
N(4)-Fe(3)-Cl(3/4)
N(1/3)-Fe(2)-N(2/4)
N(1)-Fe(2)-N(3/4)
N(2)-Fe(2)-N(3/4)
Fe(1)-N(1/2)-Fe(2)
Fe(2)-N(3/4)-Fe(3)
Fe(1)-N(1/2)-Cipso
Fe(2)-N(1/2)-Cipso
Fe(3)-N(3/4)-Cipso
Fe(2)-N(3/4)-Cipso
Fe(1)‚‚‚Fe(2)‚‚‚Fe(3)
Figure 9. Anion structure of [Li(THF)4]2[Fe3(µ3-N-p-Tol)4Cl4] (5b) with
thermal ellipsoids (50% probability level) and selected atom labels.
Planarity and Torsion
0.04(1)/0.05(1)
0.003(6)/0.02(1)
0.011
N(1/2)b
0.027(4)/0.020(4)
0.007(3)/0.011(3)
0.020
N(3/4)b
high-valent ion to the central site, which has the most
N-anion donors. For 5a, the diffraction data were sufficient
to resolve the trinuclear cluster and two solvated counterions,
but extensive disorder precluded a detailed analysis.26
Spectroscopic Properties. Characteristic spectroscopic
data for the Fe-NR clusters are summarized in Table 6. The
electronic absorption spectra of the clusters in solution
(Figure 10) are dominated by broad, intense charge-transfer
bands at near-ultraviolet to blue energies that tail across the
visible range. Molar absorptivities for the dinuclear clusters
are approximately half those of the higher nuclearity clusters,
and consequently, only the diiron clusters manifest distin-
guishable solution colors other than dark brown or black.
The Fe-NR clusters display well-defined, isotropically
shifted 1H NMR spectra that are consistent with ideal
molecular symmetries and unrestricted N-Ar rotation (Fig-
ure 11). The resonances for the phenylimide ligands were
assigned through integral ratios, which clearly identify the
p-H signal, and paramagnetic relaxation effects, which
associate the broadest peaks with the nuclei closest to the
paramagnetic center (i.e., o-H protons); comparison with
spectra of ring-substituted derivatives (i.e., 5b, 6b,c) further
supports the assignments. All aryl-H resonances evince
isotropic shifts27 that alternate in sign with alternant hydro-
carbon position (upfield for o- and p-H, downfield for m-H),
[Fe(1,2), N(1,2)]c
[Fe(2,3), N(3,4)]c
Fe(1)N2Fe(2)N2Fe(3)d
N(1/2)-Are
0.001
0.014
88.6(2)
10.7(2)/15.2(2)
1.9(3)/6.9(4)
88.3(1)
15.15(8)/14.76(8)
2.5(2)/4.8(2)
N(3/4)-Are
a Divided entries refer to separate, related atom pairs and their associated
metrics in the order given, e.g., N(1/3)-Fe(1/3)-N(2/4) denotes two angles,
N(1)-Fe(1)-N(2) and N(3)-Fe(3)-N(4). b Perpendicular displacement
from the [Fe, Fe′, Cipso] plane. c Rms deviation from the least-squares-fitted
[Fe2N2] rhomb plane. d Dihedral angle between vertex-fused, least-squares-
fitted [Fe2N2] rhomb planes. e Dihedral angle between the least-squares-
fitted [Fe2N2] rhomb and aryl ring planes linked at the specified nitrogen.
reverse sign upon substitution of aryl-Me for aryl-H and do
not attenuate with distance from the metal; these character-
istics indicate that the shifts are propagated predominantly
by a spin delocalization mechanism.28 The magnitudes of
the shifts fall into diagnostic ranges ([Fe3(µ-NAr)4Cl4]2-
.
[Fe2(µ-NAr)2Cl4]2- > [Fe4(µ3-NAr)4Cl4]2-) that easily dis-
tinguish the cluster types in solution.
The electronic properties of the iron centers were inves-
tigated further using zero-field Mo¨ssbauer spectroscopy.29
At 4.2 K, polycrystalline samples of dimeric 6a,c each show
a single, well-resolved quadrupole doublet with isomer shift
(27) (∆H/H0)iso ) (∆H/H0)dia - (∆H/H0)obs; (∆H/H0)dia in CD3CN, using
arylamines as diamagnetic references (ppm): PhNH2, 6.6 (o- and p-H),
7.08 (m-H); p-TolNH2, 2.18 (Me), 6.54 (o-H), 6.90 (m-H); MesNH2,
2.08 (o-Me), 2.14 (p-Me), 6.70 (m-H).
(26) Crystals of 5a were obtained in two forms; both diffracted poorly and
suffered from massive disorder of cation-associated solvent within a
large asymmetric unit (two independent complexes). Parameters for
crystals obtained by slow evaporation of THF at -30 °C: [Li(THF)4]2-
[Fe3(µ-NPh)3Cl4], space group P21 (No. 3), Z ) 4, a ) 14.486(1) Å,
b ) 18.495(1) Å, c ) 25.746(1) Å, â ) 90.437(4)°, V ) 6897.3(8)
Å3, temperature ) 200(2) K. For crystals obtained by THF/n-pentane
diffusion at -30 °C: [Li2(THF)7][Fe3(µ-NPh)3Cl4]‚3THF, space group
P21/c (No. 14), Z ) 8, a ) 21.882(2) Å, b ) 25.837(3) Å, c ) 24.657-
(3) Å, â ) 100.984(4)°, V ) 13685(3) Å3, temperature ) 200(2) K.
(28) (a) Horrocks, W. D., Jr. Analysis of Isotropic Shifts. In NMR of
Paramagnetic Molecules: Principles and Applications; LaMar, G. N.,
Horrocks, W. D., Jr., Holm, R. H., Eds.; Academic Press: New York,
1973; Chapter 4. (b) Holm, R. H.; Phillips, W. D.; Averill, B. A.;
Mayerle, J. J.; Herskovitz, T. J. Am. Chem. Soc. 1974, 96, 2109.
(29) (a) Mu¨nck, E. Aspects of 57Fe Mo¨ssbauer Spectroscopy. In Physical
Methods in Bioinorganic Chemistry; Que, L., Ed.; University Science
Books: Sausalito, CA, 2000; pp 287-319. (b) Gu¨tlich, P.; Link, R.;
Trautwein, A. Mo¨ssbauer Spectroscopy and Transition Metal Chem-
istry; Springer-Verlag: Berlin, 1978.
1218 Inorganic Chemistry, Vol. 42, No. 4, 2003

Iron-Arylimide Clusters [Fem(NAr)nCl4]2-
Table 6. Spectroscopic Data for Fe-NR Clusters
1H NMR:b δ, ppm
m-H
Mo¨ssbauer:c mm/s
cluster
[Fe3(µ-NPh)4Cl4]2- (5a)
electronic abs:a λ, nm (ꢀM, L‚mol-1‚cm-1
)
o-H/Me
p-H/Me
d
|∆EQ|
260 (48 000), 412 (18 000), 693 (sh, 4800)
ca. -105 (v br)
ca. -109 (v br)
-20.7 (br)
99 (br)
101 (br)
22.9
-104 (br)
0.51 (67%)
-0.21 (33%)
0.37 (67%)
-0.20 (33%)
0.44
1.70
0.84
1.13
0.85
1.21
[Fe3(µ-N-p-Tol)4Cl4]2- (5b)
[Fe2(µ-NPh)2Cl4]2- (6a)
[Fe2(µ-NMes)2Cl4]2- (6c)
[Fe4(µ3-NPh)4Cl4]2- (7)
256 (40 000), 432 (18 000), 700 (sh, 4800)
113 (br)
-22.8
16.7
268 (21 000), 298 (sh, 13 000), 399 (7300),
456 (7900), 524 (6700)
242 (22 300), 272 (21 800), 325 (sh, 12 600),
400 (sh, 11 500), 420 (13 200)
243 (45 000), 275 (sh, 35 000), 309 (sh, 28 000),
660 (4000)
13.9 (br)
22.5
0.51
0.57
1.38
1.32
-6.4 (br)
17.4
-8.8
a THF solution. b CD3CN solution, ca. 25 °C; cation/lattice-associated solvent signals were also evident: DME (6a, 7), 3.30 (s, Me), 3.47 (s, CH2); THF
(5a,b, 6c), 1.81 (m), 3.65 (m). c Data were collected at 4.2 K (6a,c, 7) or 200 K (5a,b), with isomer shifts referenced to Fe metal at room temperature; errors
in fit estimated at (0.02 mm/s.
Figure 10. Comparative electronic absorption spectra of Fe-NR clusters
in THF. The spectrum of 5b is almost indisinguishable from that of 5a and
is not shown.
Figure 11. Comparative 1H NMR spectra of Fe-NR clusters in CD3CN.
Signals in the diamagnetic region (δ 0-10 ppm) belong to residual protio
and quadrupole splitting indicative of high-spin Fe(III).
Under the same conditions, the mixed-valent cubane dianion
solvent, cation/lattice-associated solvent (DME or THF), and arylamines
arising from trace hydrolysis. Cluster 5a is visible as a minor contaminant
7 also presents a single symmetric quadrupole doublet of
narrow line width, demonstrating that a single, valence-
delocalized iron environment exists for this cluster on the
spectroscopic time scale; the isomer shift is greater than those
of the all-ferric dimers, as expected for a more reduced
(mean: +2.5) iron oxidation state. The Mo¨ssbauer spectra
of trinuclear 5a,b, in contrast, exhibit broad magnetic features
at 4.2 K that denote a paramagnetic ground state with inter-
mediate spin relaxation;30 these features collapse at 200 K
to give spectra suggestive of two broad, overlapping quad-
rupole doublets. These latter spectra can each be simulated
by a combination of two quadrupole doublets in a 2:1 ratio,
with isomer shifts characteristic of valence-localized, high-
spin Fe(III) and Fe(IV), respectively; this model must be
treated with caution, however, as the doublets are not fully
resolved at 200 K, but it is in accord with the oxidation state
distribution suggested by cluster charge and structure.
in the spectrum of 6a.
Redox Behavior. The redox properties of the Fe-NR
clusters were characterized by cyclic voltammetry in MeCN
solution (Table 7). Measurements were conducted using 0.1
M (n-Bu4N)Cl supporting electrolyte to suppress chloride
ligand dissociation; the use of a more typical, noncoordi-
nating electrolyte (0.1 M TBAP) led to ill-defined electro-
chemical responses. The following observations are noted:
(i) Redox processes for the di- and tetranuclear phenylim-
ide clusters (6a, 7) are irreversible. Voltammetric scans
through the initial irreversible reduction of dinuclear 6a lead
to the development of secondary redox features associated
with cubane 7; this behavior echoes the reductive core
conversion employed preparatively.
(ii) The diferric mesitylimide cluster 6c is oxidatively more
stable than its phenylimide counterpart (6a). The difference
is significant, with 6c displaying a reversible oxidation
process along with an irreversible reduction that is shifted
negative by 500 mV relative to the equivalent event in 6a.
(30) Yoo, S. J.; Angove, H. C.; Burgess, B. K.; Hendrich, M. P.; Mu¨nck,
E. J. Am. Chem. Soc. 1999, 121, 2534.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1219

Duncan et al.
Table 7. Electrochemical Dataa for Fe-NR and Analogous Fe-S
core is known only with thiolate terminal ligation and strictly
in the all-ferric [Fe3(µ-S)4]+ state.35 While the substitution
of terminal thiolate ligands by chloride is well precedented
in Fe-S cluster chemistry34 and should be achievable in this
instance, there is no evidence that the more oxidized [Fe3-
(µ-S)4]2+ core is sustainable with either thiolate or chloride
ligation.
Clusters
cluster
Eox, V (∆Ep, mV)
Ered, V (∆Ep, mV)
[Fe3(µ-Q)4Cl4]2-
Q ) NPh (5a)
Q ) N-p-Tol (5b)
[Fe2(µ-Q)2Cl4]2-
Q ) NPh (6a)
Q ) NMes (6c)
Q ) Sd
+0.15 (80)
+0.04 (65)
-0.85 (qrevb)
-0.99 (qrevb)
not obsdc
+0.15 (110)
not reported
-1.48 (irrev)
<-1.7 (irrev)
-1.00 (irrev)
[Fe4(µ3-Q)4Cl4]2-
Q ) NPh (7)
Q ) Sd
-0.28 (irrev)
not reported
-1.57 (irrev)
-0.79 (90)
a Eox/red refer to the first oxidation/reduction potentials (E1/2 for reversible
and quasi-reversible couples, Epa or Epc for irreversible processes) measured
at 100 mV/s scan rate and reported vs SCE; see Experimental Section for
details. b Quasi-reversible behavior: for 5a, ∆Ep ) 140, 190, and 290 mV
c
a
at 20, 100, and 500 mV/s scan rates, respectively, with ip /ip ≈ 1 for all
c
a
scans; for 5b, ∆Ep ) 310, 440, and 550 mV and ip / ip ≈ 1.3, 2.0, and 2.9
for 10, 100, and 250 mV/s scan rates, respectively. c The onset of chloride
oxidation (from the supporting electrolyte) limits the upper scan range to
ca. +0.20 V vs SCE. d From ref 34a.
(iii) The trinuclear arylimide clusters 5a,b each exhibit a
quasi-reversible reduction and a reversible oxidation. The
reduction process is more reversible for the phenylimide
trimer 5a on the basis of peak current ratios and potential
separations. The potential for this couple, which presumably
represents the reduction of the trinuclear core to the all-ferric
state, is much lower than the oxidation potential of [FeCl4]2-
(-0.04 V vs SCE, reversible);31 these relative potentials
therefore support the thermodynamic feasibility of the
reaction outcome posited in eq 3, at least under the conditions
of these voltammetric measurements.
(iv) The reversible oxidation events observed for 5a,b and
6c denote the existence of well-defined 2-/1- redox couples
in these clusters. Although the nature of these oxidations,
whether ligand- or metal-centered, is unknown at present,
the reversibility of the couples and their modest oxidation
potentials suggest that the oxidized species might be acces-
sible for future study.
Fe-NR and Fe-S Chemistry: A Comparison. The Fe-
NR chemistry developed in this study bears striking similari-
ties to the well-established chemistry of Fe-S clusters. A
structural commonality is immediately evident: if we assume
congruency between imide and sulfide ligation, the three Fe-
NR core geometries become isostructural with fundamental
Fe-S cluster motifs32 found in both biological33 and synthetic
systems. The pairing is exact for the di- and tetranuclear
species, where equivalent Fe-NR and Fe-S clusters exist
with identical terminal ligation and cluster charge.34 For the
trinuclear cores, the homology is less precise, as the Fe-S
The relationship between Fe-NR and Fe-S reaction
systems also includes characteristic aspects of chemical
behavior. We have shown in this work and elsewhere that,
under suitable Fe-NR synthesis conditions, weak-field iron
species readily mediate both the 4-electron oxidative coupling
of arylamine to azoarene and the 2-electron reduction of
diarylhydrazines to iron-bound imide5 (eqs 4 and 5). These
interrelated redox transformations are equivalent in chemical
context to sulfur-based half-reactionssthe interconversions
of organodisulfide/thiolate and elemental sulfur/polysulfide/
sulfide (eqs 6 and 7)subiquitous in Fe-S chemistry.32,35,36
In addition, the reductive transformations of the di- and
trinuclear Fe-NPh clusters to cubane 7 fully replicate
chemical and electrochemical core conversions of isostruc-
tural Fe-S complexes (eqs 8 and 9).32a,34a,35
Ar(H)N-N(H)Ar + 2e- h 2[ArN]2- + 2H+
2ArNH2 h ArNdNAr + 4H+ + 4e-
(4)
(5)
RS-SR + 2e- h 2RS-
(6)
(7)
nS0 + 2e- h [Sn]2- h S2- (limit for n ) 1)
[Fe2(µ-S)2Cl4]2- + 2e- f [Fe4(µ3-S)4Cl4]2- + 4Cl- (8)
[Fe3(µ-S)4(SR)4]3- + FeCl2 + RS- f
(31) [FeCl4]2- was generated in situ by addition of a small amount of FeCl2
to a 0.1 M (n-Bu4N)Cl/MeCN electrochemical solution.
(32) (a) Beinert, H.; Holm, R. H.; Mu¨nck, E. Science 1997, 277, 653. (b)
Dance, I.; Fisher, K. Prog. Inorg. Chem. 1994, 41, 637 and references
therein.
(33) The terminal cysteinate-ligated di-, tri-, and tetranuclear Fe-S cluster
cores are biologically ubiquitous; the di- and tetranuclear clusters are
native to a variety of metalloproteins in all organisms, and the trinuclear
geometry has been observed as a stable, protein-bound isomer of a
native biometallocluster under specific denaturing conditions.
(34) (a) Wong, G. B.; Bobrik, M. A.; Holm, R. H. Inorg. Chem. 1978, 17,
578. (b) Bobrik, M. A.; Hodgson, K. O.; Holm, R. H. Inorg. Chem.
1977, 16, 1851.
[Fe4(µ3-S)4(SR)4]2- + 0.5RSSR + 2Cl- (9)
The occurrence of parallel nitrogen and sulfur chemistries
is curious inasmuch as imide and sulfide ligands, while
sharing the same formal charge, are not isoelectronic in a
(35) (a) Hagen, K. S.; Watson, A. D.; Holm. R. H. J. Am. Chem. Soc.
1983, 105, 3905. (b) Hagen, K. S.; Holm, R. H. J. Am. Chem. Soc.
1982, 104, 5496.
(36) Representative preparative examples: Christou, G.; Garner, C. D. J.
Chem. Soc., Dalton Trans. 1979, 1093.
1220 Inorganic Chemistry, Vol. 42, No. 4, 2003

Iron-Arylimide Clusters [Fem(NAr)nCl4]2-
resemblance to the nitrogenase FeMo-cofactor. The structure
of the latter derives primarily from macromolecular crystal-
lography39 and has yet to be attained in any synthetic Fe-S
system. The Co-NR cluster nevertheless faithfully repro-
duces essential distinguishing featuressthe unsaturated trigo-
nal metal centers, the open, nonrhombic cluster faces, and
the singular core frameworksof the cofactor geometry.40
Under specific circumstances, weak-field Fe-S clusters
are mirrored by cobalt-sulfur (Co-S) homologues with
matching geometries, ligands, and charge states (Figure
12).41-48 This homology depends not on electron counts but
Figure 12. Homologous weak-field Fe-S and Co-S cluster cores with
experimentally observed charges and mean oxidation states. The [M6(µ3-
S)8] geometry represents a strong-field environment50 that is included here
because of constitutional commonality (sulfide core, terminal phosphine
ligation) with the other, weak-field examples.
(39) (a) Kim, J.; Rees, D. C. Nature 1992, 360, 553. (b) Chan, M. K.;
Kim, J.; Rees, D. C. Science 1993, 260, 792. (c) Bolin, J. T.;
Campobasso, N.; Muchmore, S. W.; Morgan, T. V.; Mortenson, L.
E. In Molybdenum Enzymes, Cofactors, and Model Systems; Stiefel,
E. I., Coucouvanis, D., Newton, W. E., Eds.; ACS Symposium Series
535; American Chemical Society: Washington, DC, 1993; pp 186-
195. (d) Peters, J. W.; Stowell, M. H. B.; Soltis, S. M.; Finnegan, M.
G.; Johnson, M. K.; Rees, D. C. Biochemistry 1997, 36, 1181. (e)
Schindelin, N.; Kisker, C.; Sehlessman, J. L.; Howard, J. B.; Rees,
D. C. Nature 1997, 387, 370. (f) Mayer, S. M.; Lawson, D. M.;
Gormal, C. A.; Roe, S. M.; Smith, B. E. J. Mol. Biol. 1999, 292, 871.
(40) We believe the distinctive features of the cobalt cluster can be attributed
partly to the steric influence of the imide substituents, much as trigonal
planar iron is stabilized in 1 by its hindered amide ligands.
(41) [Fe3(µ3-S)]4+: Hagen, K. S.; Christou, G.; Holm, R. H. Inorg. Chem.
1983, 22, 309.
strict sense37 and differ considerably in essential properties
that include basicity,9 hard/soft character, bonding radii, and
steric demand. One definite consequence of these dissimilari-
ties is a marked stabilization of oxidized states in the imide
clusters relative to their sulfide analogues. This is clearly
evident in the first reduction potentials for analogous di- and
tetranuclear Fe-NPh/Fe-S cluster pairs, which are shifted
negative by 500-650 mV for the imide species (Table 7).
Increased oxidative stability is further demonstrated by the
trinuclear Fe-NR clusters, which are isolated in a state one
electron more oxidized than their all-ferric Fe-S counter-
parts, and in the reversible electrochemical oxidations
exhibited by these clusters and by the diferric cluster 6c;
the Fe-S clusters of interest here, by contrast, show no well-
defined anodic processes whatsoever. Despite the differences,
the qualitative chemical resemblance of nitrogen and sulfur
at iron remain intriguing, especially in light of mechanistic
proposals that invoke mixed Fe-S-N cluster cores as
defining intermediates in the enzymatic reduction of dini-
trogen.2,3
(42) [Co3(µ3-S)]4+: (a) Henkel, G.; Tremel, W.; Krebs, B. Angew. Chem.,
Int. Ed. Engl. 1983, 22, 318. (b) Also see ref 41.
(43) [Fe7(µ-S)3(µ3-S)3]3+: Noda, I.; Snyder, B. S.; Holm, R. H. Inorg.
Chem. 1986, 25, 3851.
(44) [Co7(µ-S)3(µ3-S)3]+,3+
: (a) Fenske, D.; Hachgenei, J.; Ohmer, J.
Angew. Chem., Int. Ed. Engl. 1985, 24, 706. (b) Jiang, F.; Huang, L.;
Lei, H.; Liu, H.; Kang, B.; Huang, Z.; Hong, M. Polyhedron 1992,
11, 361. (c) Jiang, F.; Lei, X.; Hong, M.; Huang, Z.; Kang, B.; Liu,
H. J. Organomet. Chem. 1993, 443, 229.
(45) [Fe8(µ4-S)6]4+,5+: (a) Pohl, S.; Saak, W. Angew. Chem., Int. Ed. Engl.
1984, 23, 904. (b) Pohl, S.; Opitz, U. Angew. Chem., Int. Ed. Engl.
1993, 32, 863. (c) Goh, C.; Segal, B. M.; Huang, J.; Long, J. R.; Holm,
R. H. J. Am. Chem. Soc. 1996, 118, 11844.
(46) [Co8(µ4-S)6]3+,4+: (a) Christou, G.; Hagen, K. S.; Holm, R. H. J. Am.
Chem. Soc. 1982, 104, 1744. (b) Christou, G.; Hagen, K. S.; Bashkin,
J. K.; Holm, R. H. Inorg. Chem. 1985, 24, 1010. (c) Hong, M.; Su,
W.; Cao, R.; Jiang, F.; Liu, H.; Lu, J. Inorg. Chim. Acta 1998, 274,
229.
A recently reported cobalt-imide (Co-NR) cluster,38
[Co8(µ3-NPh)6(µ-NPh)3(PPh3)2]-, is pertinent to the present
discussion. The octacobalt cluster is unmistakeable in its
(47) [Fe6(µ3-S)8]0-3+: (a) Cecconi, F.; Ghilardi, C. A.; Midollini, S. J.
Chem. Soc., Chem. Commun. 1981, 640. (b) Del Giallo, F.; Pieralli,
F.; Fiesoli, L.; Spina, G. Phys. Lett. 1983, 96A, 141. (c) Agresti, A.
Bacci, M.; Cecconi, F.; Ghilardi, C. A.; Midollini, S. Inorg. Chem.
1985, 24, 689. (d) Cecconi, F.; Ghilardi, C. A.; Midollini, S.; Orlandini,
A.; Zanello, P. J. Chem. Soc., Dalton Trans. 1987, 831. (e) Bencini,
A.; Uytterhoeven, M. G.; Zanchini, C. Int. J. Quantum Chem. 1994,
52, 903. (f) Bencini, A.; Ghilardi, C. A.; Midollini, S.; Orlandini, A.;
Russo, U.; Uytterhoeven, M. G.; Zanchini, C. J. Chem. Soc., Dalton
Trans. 1995, 963. (g) Goddard, C. A.; Long, J. R.; Holm, R. H. Inorg.
Chem. 1996, 35, 4347.
(37) Basic symmetry considerations dictate that, in isolation, imide and
sulfide have different degeneracies and relative orderings for their
respective frontier orbitals; there is therefore no theoretical basis to
expect a priori that the two ligand types will adopt the same bound
structures. Isoelectronic arguments can only be made a posteriori once
a shared ligation mode is evident and only for the complete metal-
ligand unit.
(38) Link, H.; Fenske, D. Z. Anorg. Allg. Chem. 1999, 625, 1878.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1221

Duncan et al.
on the mutual accessibility of shared oxidation states (this
is characteristic of weak-field 3d environments, where
metal-metal bonding and valence electron count have little
value in rationalizing cluster structure compared to other,
interrelated factors, such as oxidation state, ligand sphere,
and net charge). As a practical matter, redox levels are
dictated by cobalt,49 a metal that strongly favors the divalent
state under weak-field conditions,50 even in the presence of
imide donors.38 The relation of Co-S to Fe-S and Fe-S
to Fe-NR systems thus rationalizes the connection between
the [Co8(NPh)9]- core and the [Fe7MoS9]z cofactor, which
in turn strengthens the structural convergence of imide and
sulfide ligation already noted on iron. The argument for
oxidation state correspondence might also apply in the Co-
NR analogue, where the average cobalt ion is +2.13, and
the FeMo-cofactor, which is of uncertain charge; it is
interesting that a mean iron level of +2.14 (1 Mo(IV)/6 Fe-
(II)/1 Fe(III)) has recently been proposed51 for the resting
form of cofactor.
examples or the more condensed heterocubane geometry in
the tetranuclear case. The [Fe3(µ-NAr)4Cl4]2- cluster is
notable for its unusual 2 Fe(III)/1 Fe(IV) formal oxidation
state; structural and Mo¨ssbauer evidence suggest, but do not
prove, a localized Fe(IV) assignment to the central iron site
within the cluster. Distinctive isotropically shifted 1H NMR
spectra allow facile identification of the clusters in solution.
In a broader context, we find intriguing similarities of
structure and reactivity linking Fe-NR and Fe-S chemis-
tries. Weak-field Fe(II/III) complexes can mediate nitrogen
redox chemistry at the N-N and NdN bond levels, fully
reduced nitrogen incorporates readily as cluster-bound imide,
and nitrogen chemistry at these iron centers resembles sulfur
chemistry of Fe-S clusters. These observations reflect
intrinsic properties of N-anions in a weak-field iron cluster
environment and may therefore have bearing on chemical
possibilities in iron-mediated mechanisms of nitrogenase
action. The conspicuous parallels between abiological Fe-
NR (and Co-NR) clusters and their Fe-S homologues
provide a starting point for more detailed comparative
investigation.
Conclusions
The protonolysis of sterically hindered FeCl[N(SiMe3)2]2-
(THF) (2) by arylamines provides a new, serviceable route
to Fe-NR cluster chemistry, allowing the isolation of [Fe2(µ-
NAr)2Cl4]2-, [Fe3(µ-NAr)4Cl4]2-, and (with added reductant)
[Fe4(µ3-NAr)4Cl4]2- clusters in quantity under appropriate
conditions. The protonolysis reactions are accompanied by
Fe(III)-mediated redox events that oxidize arylamine to
azoarene with formation of Fe(II) and probably dispropor-
tionate some of the iron to the +II and (formal) +IV states;
the arylamine oxidation, which poses particular complications
in cluster synthesis, can be controlled by the introduction of
LiCl as a stoichiometric co-reagent.
Addendum
The structure of the FeMo-cofactor has very recently been
revised to include an interstitial light atom, believed to be
nitride, within the center of the cluster core.52 If correct, this
discovery has significant ramifications for the molecular
description of nitrogenase chemistry;53 indeed, the presence
of core-bound nitride may represent the first observational
evidence for direct Fe participation in nitrogenase action.
The existence of this unforeseen, biologically unprecedented
ligand provides a compelling and timely argument for further
exploration of iron/N-anion clusters as a tool to understand
fundamental aspects of cofactor-relevant chemistry.
The core frameworks of the Fe-NR clusters are con-
structed from tetrahedral iron centers and bridging imide
ligation, organized into linear arrays for the di- and trinuclear
Experimental Section
Preparation of Compounds. All operations were performed
under dry, anaerobic conditions (pure N2 atmosphere) using standard
protocols for the manipulation of air-sensitive compounds;54 bench-
top Schlenk operations were conducted in all-glass apparatus
whenever possible. Except for anaerobic storage and handling,
reagents were used as received (Acros, Aldrich, Cerac) unless
otherwise noted. Solvents and liquid reagents were distilled from
appropriate scavengers (ethers, n-pentane and benzene from sodium
benzophenone ketyl, toluene from Na metal, amine reagents from
CaH2) and then degassed and kept over 4 Å molecular sieves;
deuterated NMR solvents (Cambridge Isotope Laboratories) were
dried by storage over 4 Å molecular sieves. Diatomaceous earth
(Celite), LiCl, and alumina were dried/activated by heating at 150
°C (250-300 °C for alumina) for >12 h under dynamic vacuum.
Spectroscopic data for mononuclear ferric amide complexes and
Fe-NR clusters are summarized in Tables 1 and 6, respectively.
(48) [Co6(µ3-S)8]0,+: (a, b) Cecconi, F.; Ghilardi, C. A.; Midollini, S.;
Orlandini, A. Inorg. Chem. Acta 1982, 64, L47; 1983, 76, L183. (c)
Fenske, D.; Hachgenei, J.; Ohmer, J. Angew. Chem., Int. Ed. Engl.
1985, 24, 706. (d) Cecconi, F.; Ghilardi, C. A.; Midollini, S.; Orlandini,
A.; Zanello, P. Polyhedron 1986, 12, 2021. (e, f) Hong, M.; Huang,
Z.; Lei, X.; Wei, G.; Kang, B.; Liu, H. Inorg. Chim. Acta 1989, 159,
1; Polyhedron 1991, 10, 927. (g) Bencini, A.; Ghilardi, C.; Orlandini,
A.; Midollini, S.; Zanchini, C. J. Am. Chem. Soc. 1992, 114, 9898.
(h) Bencini, A.; Midollini, S.; Zanchini, C. Inorg. Chem. 1992, 31,
2132. (i) Jiang, F.-L.; Huang, Z.-Y.; Shi, J.-Q.; Lei, X.-J.; Hong, M.-
C.; Lui, H.-Q. Chin. J. Struct. Chem. (Jiegou Huaxue) 1993, 12, 312.
(j) Jiang, F.; Huang, X.; Cao, R.; Hong, M.; Liu, H. Acta Crystallogr.
1995, C51, 1275. (k) Cecconi, F.; Ghilardi, C. A.; Midollini, S.;
Orlandini, A.; Zanello, P.; Cinquantini, A,; Bencini, A.; Uytterhoeven,
M. G.; Giorgi, G. J. Chem. Soc., Dalton Trans. 1995, 3881.
(49) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. AdVanced
Inorganic Chemistry, 6th ed.; Wiley-Interscience: New York, 1999;
pp 814-835.
(50) The exception to the divalent trend occurs for the [M6(µ3-S)8] cores,
which have oxidation levels closer to trivalent; the correlation between
Fe/Co oxidation states persists nonetheless. Although the four cluster
types in Figure 12 share the same basic σ- and π-donor ligands (core
sulfides, terminal phosphines/halides/thiolates), the hexanuclear com-
plexes possess 5-coordinate, square-pyramidal metal sites that are
strong-field and low-spin (see refs 47f and 48 g,h); this explains the
stability of the oxidized states. The other structure types contain
4-coordinate, weak-field metal sites.
(52) Einsle, O.; Tezcan, F. A.; Andrade, L. A.; Schmid, B.; Yoshida, M.;
Howard, J. B.; Rees, D. C. Science 2002, 297, 1696.
(53) Lee, S. C.; Holm, R. H. Submitted for publication in Proc. Natl. Acad.
Sci. U.S.A.
(54) (a) Shriver, D. F.; Drezdon, M. A. The Manipulation of Air-SensitiVe
Compounds, 2nd ed.; Wiley-Interscience: New York, 1986. (b)
Experimental Organometallic Chemistry; Wayda, A. L., Darensbourg,
M. Y., Eds.; ACS Symposium Series 357; American Chemical
Society: Washington, DC, 1987.
(51) Lee, H. I.; Hales, B. J.; Hoffman, B. H. J. Am. Chem. Soc. 1997, 119,
11395.
1222 Inorganic Chemistry, Vol. 42, No. 4, 2003

Iron-Arylimide Clusters [Fem(NAr)nCl4]2-
Table 8. Crystallographic Dataa for FeCl[N(SiMe3)2]2(THF) (2), [Li(TMEDA)2][FeCl2{N(SiMe3)2}2] (3), [Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4]‚THF
(4‚THF), [Li(THF)4]2[Fe3(µ-N-p-Tol)4Cl4] (5b), [Li(DME)3]2[Fe2(µ-NPh)2Cl4] (6a), [Li2(THF)7][Fe2(µ-NMes)2Cl4] (6c), and
[Li(DME)3]2[Fe4(µ3-NPh)4Cl4]‚DME (7‚DME)
2
3
4‚THF
5b
6a
6c
7‚DME
formula
fw
space group P1h (No. 2)
Z
C
16H44N2OFeClSi4
C
24H68N6FeCl2Si4
C
76H124N4O13Fe7Cl10
C
60H92N4O8Fe3Cl4Li2
C
36H70N2O12Fe2Cl4Li2
C
46H78N2O7Fe2Cl4Li2
C
52H90N4O14Fe4Cl4Li2
484.18
686.88
P21/c (No. 14)
8
2047.23
P1h (No. 2)
2
1320.61
P21/n (No. 14)
4
990.28
C2/c (No. 15)
4
1038.48
P21/c (No. 14)
4
1374.31
P21/n (No. 14)
4
2
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
8.3297(4)
11.4333(6)
15.7784(8)
105.188(2)
92.640(2)
109.323(2)
1353.9(1)
19.2815(4)
22.4243(4)
19.5254(4)
13.4941(7)
14.7117(6)
23.627(1)
83.721(2)
88.776(2)
87.477(3)
4657.1(4)
1.46
200(2)
24.20
98.5
1.403
14.4481(7)
32.386(2)
14.4685(8)
29.231(1)
11.6925(5 )
19.0941(8 )
13.5238(4)
19.3851(6)
20.3408(7)
12.5052(6)
15.2469(8)
35.8276(19 )
96.7900(8)
91.610(2)
128.378(2 )
94.911(1)
96.323(2)
8383.1(2)
1.09
200(2)
24.96
99.8
0.623
6767.4(6)
1.30
200(2)
22.42
98.9
0.844
5115.9(4)
1.29
200(2)
26.01
99.8
0.827
5313.0(3)
1.30
110(2)
27.48
99.9
0.794
6789.5(6)
1.35
200(2)
23.00
99.9
1.014
F
calcd, g/cm3 1.19
temp, K
140(2)
26.42
θ
max, deg
tot. data,b % 99.5
µ, mm-1
R (wR2),c % 3.84 (9.01)
Sd
1.065
0.841
5.05 (12.25)
1.030
6.79 (13.38)
1.029
4.44 (10.97)
1.034
4.37 (10.78)
1.015
3.73 (8.00)
1.022
5.11 (14.86)
1.014
a Data collected with graphite-monochromatized Mo KR radiation (λ ) 0.71073 Å) using ω scans. b Percent completeness of (unique) data collection
2
2
2
within θmax limit. c Calculated for I > 2σ(I): R ) Σ||Fo| - |Fc||/Σ|Fo|, wR2 ) {Σ[w(Fo2 - Fc )2]/Σ[w(Fo )2]}1/2
.
d S ) goodness of fit ) [Σw(Fo2 - Fc )2/(n
- p)]1/2, where n is the number of reflections and p is the number of parameters refined.
The products described herein may decompose over time, even in
solid form under anaerobic conditions (e.g., 6c); compounds were
therefore stored at -30 °C to enhance stability. Sample instability/
reactivity probably explains difficulties in obtaining satisfactory
elemental analyses, which in some cases have given wildly
inconsistent values for duplicate analyses of the same crystalline
material; this problem has been noted previously for 1.10c,d
Fe[N(SiMe3)2]3 (1).10 A modified literature synthesis10a,c was
used. A solution of Na[N(SiMe3)2] (2.220 g, 12 mmol) in benzene
(50 mL) was slowly added to a stirred suspension of FeCl3 (1.300
g, 8.0 mmol) in benzene (15 mL) at 4 °C (ice bath), producing an
immediate dark green color. The mixture was allowed to warm to
ambient temperature, stirred overnight, and then evaporated to
dryness in vacuo. The residue was extracted with n-pentane (60
mL) and filtered; the filtrate was concentrated to ca. 5 mL and
cooled to -20 °C to yield dark green, crystalline 1, which was
isolated and dried in vacuo (0.836 g). Concentration and cooling
of the mother liquor gave a second crop (0.370 g). The isolated
yield (total: 1.206 g, 56% based on amide) is limited primarily by
high product solubility. The compound identity was confirmed by
unit cell determination10h,i (170 K): trigonal, a ) 16.191(4) Å, c
) 8.422(8) Å, V ) 1912(5) Å3.
FeCl[N(SiMe3)2]2(THF) (2). A cold (-2 °C NaCl/ice bath)
solution of Na[N(SiMe3)2] (24.86 g, 135 mmol) in THF (125 mL)
was added to a stirred suspension of FeCl3 (11.00 g, 67.8 mmol)
in THF (200 mL) at -2 °C. The deep red reaction mixture was
allowed to warm to ambient temperature, stirred overnight (>12
h), and then filtered through diatomaceous earth (Celite) to remove
NaCl; the filter bed was washed with THF to recover trapped
product. The filtrate and washings were combined, evaporated to
dryness in vacuo with gentle heating, then dried further under
dynamic vacuum (12 h) to remove entrained THF. The solid residue
was subjected to several cycles of n-pentane extraction and filtration,
after which the combined pentane extracts (ca. 850 mL) were
volume reduced to 650 mL and cooled to -20 °C for 6 h. The
resulting dark red, air-sensitive crystals were isolated by filtration
and dried under dynamic vacuum (12 h) at 40 °C to give 13.73 g
of 2. The filtrate was volume reduced further to 100 mL and then
30 mL and cooled to obtain additional crops (7.89 and 2.54 g,
respectively). Total yield: 24.16 g (73.6%). Anal. Calcd for C16H44-
ClFeN2OSi4: C, 39.69; H, 9.16; N, 5.79. Found: C, 39.63; H, 8.81;
N, 6.50.
[Li(TMEDA)2][FeCl2{N(SiMe3)2}2] (3). A mixture of 2 (0.968
g, 2.0 mmol) and LiCl (0.085 g, 2.0 mmol) was stirred in THF (20
mL) for 12 h. The resulting transparent red solution was evaporated
to give a viscous oil that was redissolved in N,N,N′,N′-tetrameth-
ylethylendiamine (TMEDA, 1.18 g, 10 mmol) and toluene (ca. 5
mL). This solution was maintained at ambient temperature for 2 d
to give red microcrystals that were collected, washed with cold
pentane (3 × 2 mL), and dried in vacuo to afford 1.08 g (78.6%)
of 3. Anal. Calcd for C24H68Cl2FeLiN6Si4: C, 41.97; H, 9.98; N,
12.24. Found: C, 40.88; H, 9.92; N, 12.36.
[Fe2(µ-Cl)3(THF)6]2[Fe3(µ-NPh)4Cl4] (4). A solution of aniline
(0.467 g, 5.0 mmol) in THF (10 mL) was added dropwise to a
stirred solution of 2 (2.422 g, 5.0 mmol) in THF (90 mL). The
resulting dark brown mixture was stirred for 19 h, filtered through
diatomaceous earth (Celite), and washed with several portions of
THF; a significant amount of THF-insoluble black precipitate was
evident at this stage. The combined black filtrate (ca. 150 mL) was
evaporated to dryness in vacuo and then extracted with n-pentane
(100 mL) to give a dark orange filtrate and a black, pentane-
insoluble powder. The filtrate was evaporated in vacuo to a dark
orange glass that was identified as azobenzene (isolated yield: 55
1
mg, 0.30 mmol) by H NMR and GC/MS analysis. The pentane-
insoluble black powder was dissolved in THF (75 mL) and then
filtered, and the black filtrate was concentrated to ca. 40 mL. Vapor
diffusion of n-pentane into this solution at -30 °C over 2 weeks
gave black microcrystalline 4, which was filtered out, washed with
n-pentane (ca. 10 mL), and dried under nitrogen flow. Yield: 0.275
g (18.5%, based on Fe and assuming 4‚THF crystal composition).
We have been unable to recrystallize 4; the compound was therefore
characterized by crystallographic and NMR analysis.
[Li2(THF)7][Fe3(µ-NPh)4Cl4] (5a). A mixture of 2 (0.968 g, 2.0
mmol) and LiCl (0.084 g, 2.0 mmol) was stirred in THF (20 mL)
for 2 h. Neat aniline (0.183 mL, 0.187 g, 2.0 mmol) was then added,
resulting in a solution color change solution (ca. 1-2 min) to purple-
black. The reaction mixture was stirred for 12 h, evaporated to
dryness, and extracted with n-pentane to remove azobenzene
(spectrophotometrically quantitated at 0.87% yield based on aniline),
giving 0.523 g of a glassy black powder. The solid material was
treated with DME (50 mL) and filtered to give a dark brown filtrate
and black precipitate. The precipitate was dissolved in THF (20
mL) and filtered; vapor diffusion of n-pentane into the filtrate at
-30 °C over 4 d gave brown needles of 5a that were isolated,
Inorganic Chemistry, Vol. 42, No. 4, 2003 1223

Duncan et al.
rinsed with THF (3 × 1 mL) followed by pentane (3 × 1 mL), and
dried in vacuo for 2 h. Yield: 0.222 g (31.8%, based on analytical
composition). Anal. Calcd for 5a-2THF, C44H60Cl4Fe3Li2N4O5: C,
50.42; H, 5.77; N, 5.34. Found: C, 50.50; H, 5.83; N, 4.94.
[Li(THF)4]2[Fe3(µ-N-p-Tol)4Cl4] (5b). A mixture of 2 (0.968
g, 2.0 mmol) and LiCl (0.084 g, 2.0 mmol) was stirred in
dimethoxyethane (DME, 20 mL) for 2 h. A DME solution (20 mL)
of p-toluidine (p-TolNH2, 0.107 g, 2.0 mmol) was then added,
resulting in darkening of the solution (ca. 1-2 min) to black. The
solution was stirred for 8 h and filtered to isolate a black precipitate
(0.288 g). Dissolution of this precipitate in THF (10 mL), followed
by vapor diffusion of n-pentane at -30 °C over 2 d, gave black,
crystalline 5b which was collected, rinsed with 1:1 THF/pentane
(5 mL) followed by pentane (3 mL), and then dried in vacuo for
1.5 h. Yield: 0.182 g (25.6%, based on analytical composition).
Anal. Calcd for 5b-3.5THF, C46H64Cl4Fe3Li2N4O4.5: C, 51.72; H,
6.04; N, 5.24. Found: C, 51.69; H, 5.55; N, 4.91.
[Li(DME)3]2[Fe2(µ-NPh)2Cl4] (6a). A mixture of 2 (0.121 g,
0.25 mmol) and LiCl (0.011 g, 0.25 mmol) was stirred in DME
(10 mL) for 2 h. Neat aniline (0.023 mL, 0.024 g 0.25 mmol) was
then added, resulting in darkening (ca. 2 min) to a purple-black
solution color. The mixture was stirred for 8 h and then filtered to
collect the purple-black, microcrystalline precipitate, which was
dried under N2 flow for 5 min. The precipitate is >90% pure 6a
by NMR analysis and contaminated only by 5a; the crystallization
characteristics of 5a and 6a prevented further purification of this
material. Yield: 0.109 g (88.1%).
[Li2(THF)7][Fe2(µ-NMes)2Cl4] (6c). A mixture of 2 (1.46 g, 3.0
mmol) and LiCl (0.128 g, 3.0 mmol) was stirred in THF (20 mL)
for 1 h. A THF solution (10 mL) of mesidine (2,4,6-trimethylaniline,
MesNH2, 0.408 g, 3.0 mmol) was then added slowly, resulting in
a color change to dark-green. The solution was stirred for 12 h
and then concentrated in vacuo to ca. 10 mL and diluted with 25
mL of pentane. After 4 h, the mixture was filtered to harvest the
microcrystalline precipitate, which was washed with pentane (3 ×
10 mL) and dried briefly in vacuo to afford dark green-gray 6c.
Yield: 1.107 g (71.1%). This compound decomposes both in
solution and in the solid state over a period of days at ambient
temperature; the solid product was therefore stored at -30 °C. Anal.
Calcd for 6c-5THF, C26H38Cl4Fe2Li2O2N2: C, 46.06; H, 5.65; N,
4.13. Found: C, 46.24; H, 6.20; N, 3.73.
[Li(DME)3]2[Fe4(µ3-NPh)4Cl4] (7). A mixture of 2 (0.968 g,
2.0 mmol) and LiCl (0.085 g, 2.0 mmol) was stirred in THF (20
mL) for 4 h. Neat aniline (0.183 mL, 2.0 mmol) was then added,
resulting in a darkening of the solution within 2 min (purple-black).
The mixture was stirred for 6 h, after which Zn dust (0.066 g, 1.0
mmol) was added, giving a brown-black solution after an additional
12 h. Solvent removal in vacuo yielded a black glass, which was
rinsed with 2:1 v/v DME/THF (80 mL) and filtered to remove a
small amount of brown tar. The filtrate was allowed to stand at
ambient temperature until small crystals formed (ca. 12 h) and then
concentrated in vacuo to 60 mL and diluted with 20 mL of
n-pentane. Storage at -30 °C for 2 d afforded black microcrystals
of 7, which were collected, rinsed with 1:1 v/v DME/n-pentane (5
mL) followed by n-pentane (5 mL), and dried in vacuo for 4 h.
Yield: 0.350 g (69.0% based on analytical composition). Anal.
Calcd for 7-3DME, C36H50Cl4Fe4Li2N4O6: C, 42.65; H, 4.97; N,
5.53. Found: C, 42.23; H, 5.01; N, 4.77.
THF, from the reaction system as black rods, and 5b, as black
needles); slow room-temperature evaporation of n-pentane (3, as
translucent orange-red rods) or 1:3 DME/THF (6a, as purple flat
needles).
General crystallographic procedures have been described else-
where.55 Essential crystallographic data for compounds in this work
are summarized in Table 8. Crystals of 6c decompose, perhaps from
desolvation, after a few hours of beam exposure under a nitrogen
stream at 200 K; data were therefore collected at 110 K. Disordered
coordinated and/or lattice solvent molecules were found in the
structures of 2, 4‚THF, 5b, 6a,c, and 7‚DME; appropriate models
and restraints were applied as needed. Specific details for individual
structure determinations are available as Supporting Information.
Electrochemical Analyses. Cyclic voltammograms were re-
corded on an Analytical Instrument Systems DLK-60 electrochemi-
cal analyzer using platinum working and counter electrodes and a
nonaqueous Ag/AgNO3 (0.01 M) reference electrode. The MeCN
solvent was distilled from CaH2 and then dried further by storage
over activated alumina, which was filtered off before use. For 2
and 3, measurements were conducted using a standard three-
electrode cell and 0.1 M (n-Bu4N)(ClO4) (TBAP) supporting
electrolyte in MeCN. For the Fe-NR clusters (5a,b, 6a,c, 7) and
[FeCl4]2-, a dual compartment cell was used to separate the analyte
solution (0.1 M (n-Bu4N)Cl/MeCN), containing the working and
counter electrodes, from the reference electrode (0.1 M TBAP/
MeCN); a fine porosity frit connected the two compartments.
Potentials were referenced internally to the ferrocene/ferrocenium
(Fc0/+) couple and then converted to the SCE scale (Fc+/0 ) +0.38
V vs SCE in 0.1 M TBAP/MeCN).56
Other Physical Measurements. NMR spectra were acquired
on Varian 300, 400, and 500 MHz spectrometers, with chemical
shifts referenced to residual protiosolvent signals. Solution absorp-
tion spectra were recorded on a Hewlett-Packard 8453 diode array
spectrophotometer; for the amide and imide complexes, concen-
trated solutions and 0.1 or 1 mm path length cells were necessary
to minimize decomposition. Mo¨ssbauer spectra were collected on
polycrystalline samples using a constant acceleration spectrometer.57
GC/MS data were obtained on a Hewlett-Packard 5890 Series II
gas chromatograph linked to a Hewlett-Packard 5971A mass
selective detector. Elemental analyses were performed by the
Microanalysis Laboratory at the School of Chemical Sciences,
University of Illinois, Urbana, IL 61801, or by Oneida Research
Services, Inc., Whitesboro, NY 13492.
Acknowledgment. We thank Dr. D. M. Ho (Princeton)
for crystallographic guidance, Prof. C. Achim (Carnegie-
Mellon) for helpful discussions and experimental assistance,
and Prof. R. H. Holm (Harvard) for access to a Mo¨ssbauer
spectrometer. The generous support of the Arnold and Mabel
Beckman Foundation (Beckman Young Investigator Award)
and the NSF (CAREER CHE-9984645) is gratefully ac-
knowledged.
Supporting Information Available: Crystallographic data
(CIF) and Mo¨ssbauer spectra (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
IC025825J
X-ray Crystallography. Single crystals suitable for X-ray
diffraction analysis were obtained from the following conditions
and solvents: cooling to -30 °C of n-pentane (2, as dark red
parallelopiped rods), THF (6c, as green needles), or DME (7‚DME,
as black blocks); THF/n-pentane vapor diffusion at -30 °C (4‚
(55) Kayal, A.; Ducruet, A. F.; Lee, S. C. Inorg. Chem. 2000, 39, 3696.
(56) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877.
(57) Details of Mo¨ssbauer instrumentation and methods: Cen, W.; Lee, S.
C.; Li, J.; MacDonnell, F. M.; Holm, R. H. J. Am. Chem. Soc. 1993,
115, 9515.
1224 Inorganic Chemistry, Vol. 42, No. 4, 2003