Synthesis and redox properties of triiron complexes featuring strong fe-fe interactions

Article abstract of DOI:10.1002/anie.201005198

A modular hexadentate polyamide ligand directs the assembly of triiron complexes (see structure). Crystallographic studies reveal significant metal-metal bonding interactions, which are enhanced upon oxidation. Copyright

Full text of DOI:10.1002/anie.201005198

DOI: 10.1002/anie.201005198
Polynuclear Complexes
Synthesis and Redox Properties of Triiron Complexes Featuring Strong
Fe–Fe Interactions**
Qinliang Zhao and Theodore A. Betley*
Polynuclear metalloenzymes mediate multielectron redox
processes extensively in biology. For example, polymetallic
assemblies are utilized in the cofactors of nitrogenases (FeMo,
metallic precursors, iron(II) provided easy entry into our
efforts at understanding the coordination modes of these
ligands.
VFe, Fe-only, P-cluster),[1] Mn O clusters in the oxygen-
The ligand (HL)H6 can be fully deprotonated with 3
equivalents of nBu2Mg in tetrahydrofuran to furnish the
trimagnesium complex [(HL)Mg3(THF)3] (1, Scheme 1).
ꢀ
evolving complex of photosystem II,[2] and the CuZ site in N2O
reductase.[3] The cluster nature of these assemblies aids
enzymatic function by mediating multielectron redox proc-
esses,[2b,4] providing multiple metal sites and coordination
modes for cooperative substrate binding,[5] and facilitating
electron transport between proteins.[6] The protein super-
structure provides the organizational template within which
the polynuclear core self-assembles and functions. In the
absence of a protein, we sought to generate a modular ligand
scaffold that could accommodate multiple transition metal
ions in the same proximal space, allowing us to explore the
reaction chemistry of clusters as molecular units.
Several factors limit the design and synthesis of multi-
nuclear assemblies, including nonselective reactivity, redox
instability, and ligand lability.[7] Cluster synthesis is largely a
self-assembly process where few means of control exist to
dictate the immediate coordination environments of each
metal center. An ideal ligand platform would permit control
over the coordination environments of the individual metal
sites within a multimetallic assembly, while directing the open
coordination sites to allow the transition metal ions to interact
cooperatively. We report herein our progress in synthesizing
trinuclear molecular units utilizing multidentate amine
ligands.[8]
To this end, the tris-amine 1,1,1-tris(aminomethyl)ethane
hydrochloric acid (tame·3HCl)[9] was derivatized with
o-fluoronitrobenzene by nucleophilic aromatic substitution
(8 equiv K2CO3, acetonitrile, 1208C, 48 h). Subsequent reduc-
tion of the bright orange trinitro species using H2 (50 psi) over
5% Pd/C afforded the hexamine MeC(CH2NHPh-o-NH2)3
((HL)H6) in good overall yield (84%). Within (HL)H6 each
binding site consists of an o-phenylenediamine subunit which
when doubly deprotonated can potentially bind a divalent
metal ion, providing an overall hexa-anionic core to stabilize
multiple metal ions. After screening of a number of divalent
Scheme 1. Conditions: a) 3 equiv nBu2Mg; b) 1.5 equiv [Fe2{N-
(SiMe3)2}4], 3 equiv PMe2R; c) Fc+PF6
.
ꢀ
Attempts to use 1 as a transmetalating reagent with metal
halide starting materials (e.g. FeCl2) yielded inconsistent
results (likely due to metal reduction), prompting us to pursue
other metal installation routes. Reaction of ligand (HL)H6
with 1.5 equivalents of [Fe2{N(SiMe3)2}4] and 3 equivalents of
a tertiary phosphine ligand in THF affords the stable triiron
complexes [(HL)Fe3(PMe2R)3] (R = Me 2a, Ph 2b) as dark
brown solids in good yields (85–89%).
Storage of 2a and 2b at ꢀ358C in ether/hexanes and
saturated hexanes, respectively, afforded crystals suitable for
X-ray diffraction analysis. Solid-state structure determination
for both 2a and 2b verified three irons are bound within the
ligand platform, giving nearly isostructural complexes (Fig-
ure 1a and Figure S3 in the Supporting Information).[10] Each
of the ligand h2-amide residues bridge adjacent metal ions
oriented around a vertical axis of C3 symmetry. Each iron has
a local square pyramidal geometry, bound to four amides
forming an equatorial plane with the apical tertiary phosphine
[*] Dr. Q. Zhao, Prof. T. A. Betley
Chemistry & Chemical Biology, Harvard University
12 Oxford Street, Cambridge, MA 02138 (USA)
Fax: (+1)617-496-3792
E-mail: betley@chemistry.harvard.edu
[**] T.A.B. and Q.Z. thank the Camille and Henry Dreyfus Postdoctoral
Program in Environmental Chemistry and Harvard University for
funding.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201005198.
Angew. Chem. Int. Ed. 2011, 50, 709 –712
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
709

Communications
The origin of the observed redox reaction was of special
interest and we thus probed the complexes by a variety of
spectroscopic means. Complexes 2a, 2b, and 3 share a
common set of intense absorptions in the UV range with
shoulders near 250 and 290 nm (Figure S7). One major visible
absorption observed at 705 nm in the spectra for 2a and 2b
(e = 900–1000mꢀ1 cmꢀ1) intensifies substantially in the spec-
trum of triiron complex 3 (e = 4800mꢀ1 cmꢀ1). This shift is a
hallmark absorption attributed to ligand oxidation by Wie-
ghardt and co-workers, indicating some extent of ligand
participation in the oxidation of 2a.[17d,e]
Magnetic susceptibility data (Figure S8) for 2a and 2b
affords meff = 3.0 and 3.2 mB, respectively (consistent with
solution magnetic moments obtained by Evansꢁ method),
while 3 shows a larger magnetic moment (meff = 4.1 mB at
300 K). The observed magnetic moments (meff) for all the iron
complexes 2–3 gradually increase with increasing temper-
ature regardless of the applied magnetic field (0.5, 1, and 2 T).
The zero-field 57Fe Mꢂssbauer spectra of the triiron com-
plexes (Figure 2) reveal a single symmetric quadrupole
Figure 1. Solid-state molecular structures for a) 2a and b) 3 with the
thermal ellipsoids set at the 50% probability level (hydrogen atoms
omitted for clarity; Fe orange, C black, N blue, P magenta, F green).
Average bond lengths [ꢀ] for 2a (from 4 molecules in asymmetric
unit): Fe–Fe 2.2995(19), Fe–NA 1.984(8), Fe–NB 2.024(8), Fe–P
2.324(4), N–CAr 1.411(12), CAr–CAr 1.392(14). Bond lengths [ꢀ] for 3:
Fe1–Fe2 2.2499(4), Fe1–Fe3 2.2868(4), Fe2–Fe3 2.2863(5); average
bond lengths: Fe–N1,2,3 2.001(2), Fe–N4,5,6 1.986(2), Fe–P 2.363(1),
N–CAr 1.415(3), CAr–CAr 1.392(4).
ꢀ
trans to a diiron unit. The average Fe Fe distances of
ꢀ
2.299(2) ꢀ for 2a and 2b are shorter than the Fe Fe distance
found in metallic iron (2.48 ꢀ),[11] and consistently on the
short end for clusters featuring two or more irons (for Fen
dFe-Fe [ꢀ]: n = 2[12,13] 2.198–3.619; n ꢁ 3[14] 2.380–3.619). The
close Fe Fe distances suggest metal–metal bonding[15] and are
ꢀ
not solely due to ligand-enforced interactions. The Mg2+ in 1
share the same local coordination environment as the Fe ions
in 2a or 2b, but the Mg ions are on average separated by
2.847(1) ꢀ,[16] indicating the ligand can accommodate much
larger intermetallic spacing.
The cyclic voltammograms for both the free ligand (HL)H6
and the tris-magnesium complex [(HL)Mg3(THF)3] (1) feature
an irreversible oxidation wave with the peak anodic current
shifting from ꢀ22 mV for (HL)H6 to ꢀ530 mV for 1 versus
[(C5H5)2Fe+]/[(C5H5)2Fe] (Figure S6). The cyclic voltammo-
gram of 2a exhibits two quasi-reversible one-electron oxida-
tion processes at E1/2 = ꢀ1.59 and ꢀ1.04 V, followed by a
quasi-reversible two-electron oxidation process at ꢀ0.61 V,
with the open-circuit potential for 2a was consistently
measured to be ꢀ1.8 V. The last redox process is nearly
coincident with the irreversible oxidation event observed for
1 where the peak anodic current occurs at ꢀ530 mV for both
species.
Figure 2. Zero-field 57Fe Mꢁssbauer spectra obtained at 77 K for 2a
represented by black dots and spectral fit as solid blue line (d, jDEQ j
[mmsꢀ1]): 0.38, 1.03; and for 3 represented by black dots and spectral
fit as solid red line (d, jDEQ j [mmsꢀ1]): 0.28, 0.78.
Chemical oxidation of 2a with [(C5H5)2Fe]PF6 proceeds in
nearly quantitative yield (integration of ferrocene produced
versus internal standard) and the cationic triiron complex
[(HL)Fe3(PMe3)3]PF6 (3) could be isolated in 75% yield
following crystallization from THFat ꢀ358C (Scheme 1). The
solid-state molecular structure for 3 is shown in Figure 1b and
is very close to the structures of the neutral phosphine
complexes 2a and 2b. Like its neutral congeners, the bond
metrics within each of the o-phenylenediamide ligand
branches of 3 are characteristic of being aromatic, closed-
doublet for each triiron complex, including the cationic
complex 3 which has a lower isomer shift (d, j DEQ j
[mmsꢀ1]): 2a 0.38, 1.03; 2b 0.39, 0.96; 3 0.28, 0.78).[18] The
magnitude of the change in the isomer shift from 2a to 3 is
consistent with molecular oxidation distributed over multiple
Fe ions in iron–sulfur clusters.[19]
Complexes 2–3 are interesting because the overall mol-
ecule has a non-zero magnetic moment, comprised of three
irons with individual moments to consider, in addition to the
potential for ligand-based redox activity to add further
complexity to the overall electronic structure. The ligand
bond metrics are consistent with aromatic, closed-shell
ꢀ
ꢀ
shell dianions (N CAr 1.415(2); CAr CAr 1.392(2) ꢀ;
Table S3), unchanged from 2a within the error of the
measured bond lengths.[17] The most significant structural
change from 2a to 3 is the contraction of the triiron core:
ꢀ
dianions, and the extremely short Fe Fe separations indicate
the molecular spin state must be considered, where the 18
valence electrons populate molecular orbitals comprised of
the 15 Fe 3d orbitals akin to the orbital interactions described
ꢀ
ꢀ
ꢀ
Fe1 Fe2 2.2499(4); Fe1 Fe3 2.2868(4); Fe2 Fe3 2.2863(5) ꢀ
(Table S2).
710
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 709 –712

by Cotton for the bonding in [Re3Cl9(m2-Cl)3]3ꢀ.[20] Mixing the
frontier orbitals of the three iron centers produces four
bonding, three non-bonding, and five anti-bonding molecular
orbital combinations (illustrated in Figure S12). Populating
[7] a) P. Venkateswara Rao, R. H. Holm, Chem. Rev. 2004, 104,
527 – 560; b) S. C. Lee, R. H. Holm, Chem. Rev. 2004, 104, 1135 –
1158.
[8] a) P. E. Baikie, O. S. Mills, Inorg. Chim. Acta 1967, 1, 55 – 60;
b) H. Kisch, P. Reißer, F. Knoch, Chem. Ber. 1991, 124, 1143 –
1148; c) A. L. Keen, M. Doster, H. Han, S. A. Johnson, Chem.
Commun. 2006, 1221 – 1223; d) J. A. Hatnean, R. Raturi, J.
Lefebvre, D. B. Leznoff, G. Lawes, S. A. Johnson, J. Am. Chem.
Soc. 2006, 128, 14992 – 14999; e) E. C. Brown, B. Johnson, S.
Palavicini, B. E. Kucera, L. Casella, W. B. Tolman, Dalton Trans.
2007, 3035 – 3042.
2
2
4
4
the 18 Fe
d
electrons [(1a1)s (2a1)s (1e)p (3a1)nb2(2e)nb
-
2
2
4
2
)s (2a1)s (1e)p (3a1)nb2(2e)nb4(1a2)s*2(3e)p*
] suggests a net
bond order of 2 within the triiron core, consistent with the
ꢀ
short Fe Fe bonds in 2a and 2b, and predicts triplet ground
state for the low-spin molecule. Furthermore, oxidation of 2a
ꢀ
would remove an electron from an Fe Fe antibonding orbital
[9] E. B. Fleischer, A. E. Gebala, A. Levey, P. A. Tasker, J. Org.
Chem. 1971, 36, 3042.
(removing an electron from the (1a2)s* orbital to give
the S = 3/2 formulation [(1a1)s (2a1)s (1e)p (3a1)nb2(2e)nb
-
2
2
4
4
ꢀ
[10] 1: C39H56Mg3N6O4, Mr = 745.83, triclinic, P1, a = 12.546(1), b =
(1a2)s*1(3e)p*2]), consistent with the observed contraction in
the triiron core observed in 3 and maintaining the electronic
equivalence of the iron ions within the molecule for both
neutral 2 and cationic 3.
12.834(2), c = 14.047(2) ꢀ, a = 95.038(2)8, b = 101.484(2)8, g =
118.428(2)8, V= 1904.8(4) ꢀ3, Z = 2, 1calcd = 1.300 Mgmꢀ3, m =
0.129 mmꢀ1
, R1 = 0.0467, wR2 = 0.1163, GOF = 1.024. 2a:
C32H51Fe3N6P3, Mr = 780.25, triclinic, P1, a = 12.6339(5), b =
17.0101(7), c = 17.4540(7) ꢀ, a = 91.360(3)8, b = 91.290(3)8, g =
108.680(2)8, V= 3550.3(2) ꢀ3, Z = 4, 1calcd = 1.460 Mgmꢀ3, m =
The new hexadentate ligand permits the isolation and full
characterization of all-ferrous triiron complexes. The trime-
1.377 mmꢀ1
,
R1 = 0.1029, wR2 = 0.1461, GOF = 1.002. 2b:
ꢀ
tallic complexes feature very short Fe Fe distances, consis-
ꢀ
C47H57Fe3N6P3, Mr = 966.45, triclinic, P1, a = 12.7214(5), b =
12.7230(4), c = 31.643(1) ꢀ, a = 82.402(2)8, b = 79.382(2)8, g =
60.162(2)8, V= 4361.6(3) ꢀ3, Z = 4, 1calcd = 1.472 Mgmꢀ3, m =
tent with strong bonding between the metal sites. The
surprisingly low chemical potential observed for the triiron
complex is attributed to the strong M–M interactions present.
The bonding within the triiron core can be described by
mixing of the frontier metal 3d orbitals, which predicts a
triplet ground-state for the low-spin complex and allows the
observed spectral data to be described in terms of the
molecular electronic structure. Work is currently underway to
investigate how the electronic structure can be varied by
ligand steric modification and ancillary ligand substitution.
1.137 mmꢀ1
,
R1 = 0.0804, wR2 = 0.0948, GOF = 1.006. 3:
C36H59F6Fe3N6OP4, Mr = 997.32, monoclinic, P21/n, a =
12.542(1), b = 20.036(2), c = 18.663(2) ꢀ, b = 107.673(1)8, V=
4468.7(9) ꢀ3, Z = 4, 1calcd = 1.482 Mgmꢀ3, m = 1.165 mmꢀ1, R1 =
0.0371, wR2 = 0.0914, GOF = 1.027. CCDC 765960 (1), 765961
(2a), 765962 (2b), 772302 (3) contain the supplementary
crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[11] Multiple Bonds Between Metal Atoms (Eds.: F. A. Cotton, C. A.
Murillo, R. A. Walton), 3rd ed., Springer, New York, 2005,
pp. 447 – 451.
Received: August 19, 2010
Keywords: iron complexes · metal cluster · polyamide ligand ·
[12] The 2009 update of the CSD was searched (WebCSD v1.0.4) to
extract data reported in this paper: webcsd.ccdc.cam.ac.uk.
[13] Shortest dFe–Fe for Fen, n ꢁ 3: a) E. J. Wucherer, M. Tasi, B.
Hansert, A. K. Powell, M.-T. Garland, J.-F. Halet, J.-Y. Saillard,
H. Vahrenkamp, Inorg. Chem. 1989, 28, 3564 – 3572; b) H.-R.
Gao, T. C. W. Mak, B.-S. Kang, B.-M. Wu, Y.-J. Xu, Y.-X. Tong,
X.-L. Yu, J. Chem. Res. Synop. 1996, 5, 186 – 187; c) M. Kim, J.
Han, Polyhedron 2007, 26, 2949 – 2956.
[14] Shortest dFe–Fe for diiron complexes: a) F. A. Cotton, L. M.
Daniels, L. R. Falvello, H. J. Matonic, C. A. Murillo, Inorg.
Chim. Acta 1997, 256, 269 – 275; b) F. A. Cotton, L. M. Daniels,
H. J. Matonic, C. A. Murillo, Inorg. Chim. Acta 1997, 256, 277 –
282; c) J. P. Blaha, B. E. Bursten, J. C. Dewan, R. B. Frankel,
C. L. Randolph, B. A. Wilson, M. S. Wrighton, J. Am. Chem.
Soc. 1985, 107, 4561 – 4562.
.
polynuclear complexes · redox reactions
[1] a) J. W. Peters, M. H. B. Stowell, S. M. Soltis, M. G. Finnegan,
M. K. Johnson, D. C. Rees, Biochemistry 1997, 36, 1181 – 1187;
b) S. M. Mayer, D. M. Lawson, C. A. Gormal, S. M. Roe, B. E.
Smith, J. Mol. Biol. 1999, 292, 871 – 891; c) O. Einsle, F. A.
Tezcan, S. Andrade, B. Schmid, M. Yoshida, J. B. Howard, D. C.
Rees, Science 2002, 297, 1696 – 1700.
[2] a) J. H. A. Nugent, A. M. Rich, M. C. W. Evans, Biochim.
Biophys. Acta Bioenerg. 2001, 1503, 138 – 146; b) K. N. Ferreira,
T. M. Iverson, K. Maghlaoui, J. Barber, S. Iwata, Science 2004,
303, 1831 – 1838; c) S. Iwata, J. Barber, Curr. Opin. Struct. Biol.
2004, 14, 447 – 453.
[3] a) K. Brown, K. Djinovic-Carugo, T. Haltia, I. Cabrito, M.
Saraste, J. J. G. Moura, I. Moura, M. Tegoni, C. Cambillau, J.
Biol. Chem. 2000, 275, 41133 – 41136; b) K. Brown, M. Tegoni,
M. PrudÞncio, A. S. Pereira, S. Besson, J. J. Moura, I. Moura, C.
Cambillau, Nat. Struct. Biol. 2000, 7, 191 – 195; c) P. Chen, S. D.
George, I. Cabrito, W. E. Antholine, J. G. Moura, I. Moura, B.
Hedman, K. O. Hodgson, E. I. Solomon, J. Am. Chem. Soc. 2002,
124, 744 – 745.
[4] P. E. M. Siegbahn, Inorg. Chem. 2000, 39, 2923 – 2935.
[5] a) J. C. Fontecilla-Camps, J. Biol. Inorg. Chem. 1996, 1, 91 – 98;
b) U. Huniar, R. Ahlrichs, D. Coucouvanis, J. Am. Chem. Soc.
2004, 126, 2588 – 2601.
ꢀ
[15] Pauling predicts an Fe Fe single bond of 2.33 ꢀ. L. Pauling, The
Nature of the Chemical Bond, 3rd ed., Cornell University Press,
Ithaca, 1960, p. 403.
[16] The increased distance cannot be accounted for by the increased
ionic radius for Mg2+ alone; see R. D. Shannon, C. T. Prewitt,
Acta Crystallogr. Sect. B 1969, 25, 925 – 946.
[17] a) A. L. Balch, R. H. Holm, J. Am. Chem. Soc. 1966, 88, 5201 –
5209; b) L. F. Warren, Inorg. Chem. 1977, 16, 2814 – 2819; c) A.
Anillo, M. R. Diaz, S. Garcia-Granda, R. Obeso-Rosete, A.
Galindo, A. Ienco, C. Mealli, Organometallics 2004, 23, 471 –
481; d) E. Bill, E. Bothe, P. Chaudhuri, K. Chlopek, K. Herebian,
S. Kokatam, K. Ray, T. Weyhermꢃller, F. Neese, K. Wieghardt,
Chem. Eur. J. 2005, 11, 204 – 224; e) K. Chlopek, E. Bill, T.
Weyhermꢃller, K. Wieghardt, Inorg. Chem. 2005, 44, 7087 –
7098.
[6] B. K. Burgess, D. J. Lowe, Chem. Rev. 1996, 96, 2983 – 3012.
Angew. Chem. Int. Ed. 2011, 50, 709 –712
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
711

Communications
[18] Within range of both low-spin FeII or high-spin FeIII for mono-
f) J. S. Duncan, T. M. Nazif, A. K. Verma, S. C. Lee, Inorg.
Chem. 2003, 42, 1211 – 1224; g) J. S. Duncan, M. J. Zdilla, S. C.
Lee, Inorg. Chem. 2007, 46, 1071 – 1080.
and polynuclear systems, see a) H. Nasri, M. K. Ellison, C.
Krebs, B. H. Huynh, W. R. Scheidt, J. Am. Chem. Soc. 2000, 122,
10795 – 10804; b) H. Nasri, M. K. Ellison, B. Shaevitz, G. P.
Gupta, W. R. Scheidt, Inorg. Chem. 2006, 45, 5284 – 5290;
c) A. K. Verma, S. C. Lee, J. Am. Chem. Soc. 1999, 121,
10838 – 10839; d) A. K. Verma, T. N. Nazif, C. Achim, S. C.
Lee, J. Am. Chem. Soc. 2000, 122, 11013 – 11014; e) H. Link, A.
Decker, D. Fenske, Z. Anorg. Allg. Chem. 2000, 626, 1567 – 1574;
[19] N. Duprꢄ, P. Auric, H. M. J. Hendriks, J. Jordanov, Inorg. Chem.
1986, 25, 1391 – 1396.
[20] a) J. Bertrand, F. A. Cotton, W. A. Dollase, J. Am. Chem. Soc.
1963, 85, 1349 – 1351; b) F. A. Cotton, T. E. Haas, Inorg. Chem.
1964, 3, 10 – 17.
712
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 709 –712