Slow magnetic relaxation in a family of trigonal pyramidal iron(II) pyrrolide complexes

Article abstract of DOI:10.1021/ja105291x

We present a family of trigonal pyramidal iron(II) complexes supported by tris(pyrrolyl-α-methyl)amine ligands of the general formula [M(solv) _n][(tpa^R)Fe] (M = Na, R = tert-butyl (1), phenyl (4); M = K, R = mesityl (2), 2,4,6-triisopropylphenyl (3), 2,6-difluorophenyl (5)) and their characterization by X-ray crystallography, Moessbauer spectroscopy, and high-field EPR spectroscopy. Expanding on the discovery of slow magnetic relaxation in the recently reported mesityl derivative 2, this homologous series of high-spin iron(II) complexes enables an initial probe of how the ligand field influences the static and dynamic magnetic behavior. Magnetization experiments reveal large, uniaxial zero-field splitting parameters of D = -48, -44, -30, -26, and -6.2 cm^-1 for 1-5, respectively, demonstrating that the strength of axial magnetic anisotropy scales with increasing ligand field strength at the iron(II) center. In the case of 2,6-difluorophenyl substituted 5, high-field EPR experiments provide an independent determination of the zero-field splitting parameter (D = -4.397(9) cm^-1) that is in reasonable agreement with that obtained from fits to magnetization data. Ac magnetic susceptibility measurements indicate field-dependent, thermally activated spin reversal barriers in complexes 1, 2, and 4 of U_eff = 65, 42, and 25 cm^-1, respectively, with the barrier of 1 constituting the highest relaxation barrier yet observed for a mononuclear transition metal complex. In addition, in the case of 1, the large range of temperatures in which slow relaxation is observed has enabled us to fit the entire Arrhenius curve simultaneously to three distinct relaxation processes. Finally, zero-field Moessbauer spectra collected for 1 and 4 also reveal the presence of slow magnetic relaxation, with two independent relaxation barriers in 4 corresponding to the barrier obtained from ac susceptibility data and to the 3D energy gap between the M_S = ±2 and ±1 levels, respectively.

Full text of DOI:10.1021/ja105291x

Published on Web 12/08/2010
Slow Magnetic Relaxation in a Family of Trigonal Pyramidal
Iron(II) Pyrrolide Complexes
W. Hill Harman,†,§ T. David Harris,† Danna E. Freedman,† Henry Fong,†
Alicia Chang,† Jeffrey D. Rinehart,† Andrew Ozarowski,# Moulay T. Sougrati,|
Fernande Grandjean,| Gary J. Long,*,⊥ Jeffrey R. Long,*,† and
Christopher J. Chang*,†,‡,§
Department of Chemistry and the Howard Hughes Medical Institute, UniVersity of California,
Berkeley, California 94720, United States, Chemical Sciences DiVision, Lawrence Berkeley
National Laboratory, Berkeley, California 94720, United States, Department of Physics, B5,
UniVersity of Lie`ge, B-4000 Sart-Tilman, Belgium, Department of Chemistry, Missouri
UniVersity of Science and Technology, UniVersity of Missouri, Rolla,
Missouri 65409-0010, United States, and National High Magnetic Field Laboratory, Florida
State UniVersity, Tallahassee, Florida 32310, United States
Received June 16, 2010; E-mail: chrischang@berkeley.edu; jrlong@berkeley.edu; glong@mst.edu
Abstract: We present a family of trigonal pyramidal iron(II) complexes supported by tris(pyrrolyl-R-
methyl)amine ligands of the general formula [M(solv)n][(tpaR)Fe] (M ) Na, R ) tert-butyl (1), phenyl (4); M
) K, R ) mesityl (2), 2,4,6-triisopropylphenyl (3), 2,6-difluorophenyl (5)) and their characterization by X-ray
crystallography, Mo¨ssbauer spectroscopy, and high-field EPR spectroscopy. Expanding on the discovery
of slow magnetic relaxation in the recently reported mesityl derivative 2, this homologous series of high-
spin iron(II) complexes enables an initial probe of how the ligand field influences the static and dynamic
magnetic behavior. Magnetization experiments reveal large, uniaxial zero-field splitting parameters of D )
-48, -44, -30, -26, and -6.2 cm-1 for 1-5, respectively, demonstrating that the strength of axial magnetic
anisotropy scales with increasing ligand field strength at the iron(II) center. In the case of 2,6-difluorophenyl
substituted 5, high-field EPR experiments provide an independent determination of the zero-field splitting
parameter (D ) -4.397(9) cm-1) that is in reasonable agreement with that obtained from fits to magnetization
data. Ac magnetic susceptibility measurements indicate field-dependent, thermally activated spin reversal
barriers in complexes 1, 2, and 4 of Ueff ) 65, 42, and 25 cm-1, respectively, with the barrier of 1 constituting
the highest relaxation barrier yet observed for a mononuclear transition metal complex. In addition, in the
case of 1, the large range of temperatures in which slow relaxation is observed has enabled us to fit the
entire Arrhenius curve simultaneously to three distinct relaxation processes. Finally, zero-field Mo¨ssbauer
spectra collected for 1 and 4 also reveal the presence of slow magnetic relaxation, with two independent
relaxation barriers in 4 corresponding to the barrier obtained from ac susceptibility data and to the 3D
energy gap between the MS ) (2 and (1 levels, respectively.
according to the expression U ) S2|D|. Complexes exhibiting
Introduction
this behavior, known as single-molecule magnets, have garnered
Since the early 1990s, certain molecules have been shown
to exhibit an energy barrier to magnetic relaxation, thereby
enabling them to retain their magnetization after removal of an
applied field and thus act as nanoscopic classical magnets.1-5
This relaxation barrier arises due to a uniaxial magnetic
anisotropy (D) acting on a nonzero spin ground state (S),
considerable interest owing to their potential applications in
high-density information storage, quantum computing, and
magnetic refrigeration.6-15 However, in order for any such
applications to be realized, higher relaxation barriers must be
(4) Milios, C. J.; Vinslava, A.; Wernsdorfer, W.; Moggach, S.; Parsons,
S.; Perlepes, S. P.; Christou, G.; Brechin, E. K. J. Am. Chem. Soc.
2007, 129, 2754.
† Department of Chemistry, University of California, Berkeley.
‡ Howard Hughes Medical Institute, University of California, Berkeley.
§ Lawrence Berkeley National Laboratory.
# Florida State University.
(5) Yoshihara, D.; Karasawa, S.; Koga, N. J. Am. Chem. Soc. 2008, 130,
10460.
(6) Garanin, D. A.; Chudnovsky, E. M. Phys. ReV. B 1997, 56, 11102.
(7) Leuenberger, M. N.; Loss, D. Nature 2001, 410, 789.
(8) Jo, M. H.; Grose, J. E.; Baheti, K.; Deshmukh, M. M.; Sokol, J. J.;
Rumberger, E. M.; Hendrickson, D. N.; Long, J. R.; Park, H.; Ralph,
D. C. Nano Lett. 2006, 6, 2014.
| University of Lie`ge.
⊥ University of Missouri.
(1) Sessoli, R.; Tsai, H. L.; Schake, A. R.; Wang, S.; Vincent, J. B.;
Folting, K.; Gatteschi, D.; Christou, G.; Hendrickson, D. N. J. Am.
Chem. Soc. 1993, 115, 1804.
(9) Ardavan, A.; Rival, O.; Morton, J. J. L.; Blundell, S. J.; Tyryshkin,
A. M.; Timco, G. A.; Winpenny, R. E. P. Phys. ReV. Lett. 2007, 98,
57201.
(2) Sessoli, R.; Gatteschi, D.; Caneschi, A.; Novak, M. A. Nature 1993,
365, 141.
(3) Gatteschi, D.; Sessoli, R.; Villain, J. Molecular Nanomagnets; Oxford
University Press: Oxford, 2006.
(10) Bogani, L.; Wernsdorfer, W. Nat. Mater. 2008, 7, 179.
(11) Stamp, P. C. E.; Gaita-Arino, A. J. Mater. Chem. 2009, 19, 1718.
9
10.1021/ja105291x  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 18115–18126 18115

A R T I C L E S
Harman et al.
achieved. Indeed, despite the enormous effort aimed at generat-
ing single-molecule magnets with high relaxation barriers, to
date, no molecule has shown magnetic hysteresis above 10 K.
The vast majority of single-molecule magnets characterized
thus far have taken the form of multinuclear transition metal
cluster compounds.1-5 Recently, however, researchers have
uncovered slow magnetic relaxation in mononuclear lanthanide
and actinide complexes.16-20 In these complexes, the large spin-
orbit coupling of the f-block ion gives rise to a highly anisotropic
ground state. In particular, this discovery has led to mononuclear
lanthanide phthalocyanine complexes with very high relaxation
barriers. In principle, similar behavior should be attainable in a
mononuclear transition metal complex with a high-spin ground
state and uniaxial anisotropy. However, unlike their f-block
counterparts that display significant spin-orbit coupling largely
independent of ligand field effects, the orbital moment of
transition metal coordination compounds is frequently quenched
by geometric distortions.21-25 Furthermore, in the absence of
steric protection, transition metal ions often undergo coordinative
saturation to form low-spin complexes. Thus, the task of creating
mononuclear transition metal-based single-molecule magnets
requires enforcing coordination geometries that preserve a high-
spin ground state while minimizing or preventing anisotropy-
quenching structural distortions.
ment of single-molecule magnets are low-coordinate high-spin
iron(II) complexes, some of which have been reported to show
axial zero-field splitting magnitudes up to |D| ) 50 cm-1,38-43 as
in the case of the planar complex (ꢀ-diketiminate)FeCH3.38 In view
of these principles and the wealth of inorganic coordination
chemistry carried out by iron in heme and nonheme protein active
sites44 and their synthetic models,45-50 many laboratories, including
ours, have been interested in exploring the structure, magnetism,
and reactivity of iron complexes in lower-coordinate two-51-55 and
three-fold32,33,43,56-68 environments. In this context, we have
(31) Jenkins, D. M.; Di Bilio, A. J.; Allen, M. J.; Betley, T. A.; Peters,
J. C. J. Am. Chem. Soc. 2002, 124, 15336.
(32) England, J.; Martinho, M.; Farquhar, E. R.; Frisch, J. R.; Bominaar,
E. L.; Mu¨nck, E.; Que, L. Angew. Chem., Int. Ed. 2009, 48, 3622.
(33) Brown, S. D.; Peters, J. C. J. Am. Chem. Soc. 2005, 127, 1913.
(34) Diaconescu, P. L.; Arnold, P. L.; Baker, T. A.; Mindiola, D. J.;
Cummins, C. C. J. Am. Chem. Soc. 2000, 122, 6108.
(35) Mindiola, D. J.; Hillhouse, G. L. J. Am. Chem. Soc. 2001, 123, 4623.
(36) Melenkivitz, R.; Mindiola, D. J.; Hillhouse, G. L. J. Am. Chem. Soc.
2002, 124, 3846.
(37) Mindiola, D. J.; Hillhouse, G. L. J. Am. Chem. Soc. 2002, 124, 9976.
(38) Andres, H.; Bominaar, E. L.; Smith, J. M.; Eckert, N. A.; Holland,
P. L.; Münck, E. J. Am. Chem. Soc. 2002, 124, 3012.
(39) Boca, R. Coord. Chem. ReV. 2004, 248, 757.
(40) Merrill, W. A.; Stich, T. A.; Brynda, M.; Yeagle, G. J.; Fettinger,
J. C.; De Hont, R.; Reiff, W. M.; Schulz, C. E.; Britt, R. D.; Power,
P. P. J. Am. Chem. Soc. 2009, 131, 12693.
This task is ideally suited to molecular inorganic chemistry and
its focus on the design and synthesis of geometrically constrained
and sterically bulky ligands. Indeed, many metal complexes
developed within this paradigm display novel or enhanced mo-
lecular reactivity,26-30 as well as atypical spin states31 and bonding
configurations.33,32-37 Of particular promise toward the develop-
(41) Reiff, W. M.; Schulz, C. E.; Whangbo, M.-H.; Seo, J. I.; Lee, Y. S.;
Potratz, G. R.; Spicer, C. W.; Girolami, G. S. J. Am. Chem. Soc.
2008, 131, 404.
(42) Reiff, W. M.; LaPointe, A. M.; Witten, E. H. J. Am. Chem. Soc.
2004, 126, 10206.
(43) Popescu, C. V.; Mock, M. T.; Stoian, S. A.; Dougherty, W. G.; Yap,
G. P. A.; Riordan, C. G. Inorg. Chem. 2009, 48, 8317.
(44) Groves, J. T. J. Inorg. Biochem. 2006, 100, 434.
(45) Bukowski, M. R.; Koehntop, K. D.; Stubna, A.; Bominaar, E. L.;
Halfen, J. A.; Münck, E.; Nam, W.; Que, L. Science (Washington,
DC, U. S.) 2005, 310, 1000.
(12) Affronte, M.; Troiani, F.; Ghirri, A.; Candini, A.; Evangelisti, M.;
Corradini, V.; Carretta, S.; Santini, P.; Amoretti, G.; Tuna, F.; Timco,
G.; Winpenny, R. E. P. J. Phys. D Appl. Phys. 2007, 40, 2999.
(13) Manoli, M.; Johnstone, R. D. L.; Parsons, S.; Murrie, M.; Affronte,
M.; Evangelisti, M.; Brechin, E. K. Angew. Chem., Int. Ed. 2007,
46, 4456.
(14) Mannini, M.; Pineider, F.; Sainctavit, P.; Danieli, C.; Otero, E.;
Sciancalepore, C.; Talarico, A. M.; Arrio, M. A.; Cornia, A.;
Gatteschi, D.; Sessoli, R. Nat. Mater. 2009, 8, 194.
(15) Loth, S.; von Bergmann, K.; Ternes, M.; Otte, A. F.; Lutz, C. P.;
Heinrich, A. J. Nat. Phys. 2010, 6, 340.
(46) de Oliveira, F. T.; Chanda, A.; Banerjee, D.; Shan, X. P.; Mondal,
S.; Que, L.; Bominaar, E. L.; Münck, E.; Collins, T. J. Science
(Washington, DC, U. S.) 2007, 315, 835.
(47) Halfen, J. A.; Mahapatra, S.; Wilkinson, E. C.; Kaderli, S.; Young,
V. G.; Que, L.; Zuberbuhler, A. D.; Tolman, W. B. Science
(Washington, DC, U. S.) 1996, 271, 1397.
(48) Rohde, J. U.; In, J. H.; Lim, M. H.; Brennessel, W. W.; Bukowski,
M. R.; Stubna, A.; Mu¨nck, E.; Nam, W.; Que, L. Science (Wash-
ington, DC, U. S.) 2003, 299, 1037.
(16) Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y. J. Am.
Chem. Soc. 2003, 125, 8694.
(49) Grapperhaus, C. A.; Mienert, B.; Bill, E.; Weyhermu¨ller, T.;
Wieghardt, K. Inorg. Chem. 2000, 39, 5306.
(17) Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y. J.
Phys. Chem. B 2004, 108, 11265.
(50) Berry, J. F.; Bill, E.; Bothe, E.; George, S. D.; Mienert, B.; Neese,
F.; Wieghardt, K. Science (Washington, DC, U. S.) 2006, 312, 1937.
(51) Holland, P. L. Acc. Chem. Res. 2008, 41, 905.
(52) Bart, S. C.; Lobkovsky, E.; Bill, E.; Chirik, P. J. J. Am. Chem. Soc.
2006, 128, 5302.
(18) AlDamen, M. A.; Clemente-Juan, J. M.; Coronado, E.; Mart´ı-
Gastaldo, C.; Gaita-Ariño, A. J. Am. Chem. Soc. 2008, 130, 8874.
(19) AlDamen, M. A.; Cardona-Serra, S.; Clemente-Juan, J. M.; Coronado,
E.; Gaita-Ariño, A.; Marti-Gastaldo, C.; Luis, F.; Montero, O. Inorg.
Chem. 2009, 48, 3467.
(53) Sazama, G. T.; Betley, T. A. Inorg. Chem. 2010, 49, 2512.
(54) King, E. R.; Betley, T. A. J. Am. Chem. Soc. 2009, 131, 14374.
(55) King, E. R.; Betley, T. A. Inorg. Chem. 2009, 48, 2361.
(56) Thomas, C. M.; Mankad, N. P.; Peters, J. C. J. Am. Chem. Soc. 2006,
128, 4956.
(20) Rinehart, J. D.; Long, J. R. J. Am. Chem. Soc. 2009, 131, 12558.
(21) Edelstein, N. M.; Lander, G. H. In The Chemistry of the Actinide
and Transactinide elements, 3rd ed.; Morss, L. R., Edelstein, N. M.,
Fuger, J., Katz, J. J., Eds.; Springer: Dordrecht, 2006; Vol. 4.
(22) Orchard, A. F. Magnetochemistry; Oxford University Press: Oxford,
2003.
(57) Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2004, 126, 6252.
(58) Brown, S. D.; Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2003,
125, 322.
(23) Wybourne, B. G. Spectroscopic Properties of Rare Earths; Wiley:
New York, 1965.
(59) Mukherjee, J.; Lucas, R. L.; Zart, M. K.; Powell, D. R.; Day, V. W.;
Borovik, A. S. Inorg. Chem. 2008, 47, 5780.
(24) Rinehart, J. D.; Harris, T. D.; Kozimor, S. A.; Bartlett, B. M.; Long,
J. R. Inorg. Chem. 2009, 48, 3382.
(60) Larsen, P. L.; Gupta, R.; Powell, D. R.; Borovik, A. S. J. Am. Chem.
Soc. 2004, 126, 6522.
(25) Benelli, C.; Gatteschi, D. Chem. ReV. (Washington, DC, U. S.) 2002,
102, 2369.
(61) MacBeth, C. E.; Golombek, A. P.; Young, V. G.; Yang, C.; Kuczera,
K.; Hendrich, M. P.; Borovik, A. S. Science (Washington, DC, U. S.)
2000, 289, 938.
(26) Laplaza, C. E.; Cummins, C. C. Science (Washington, DC, U. S.)
1995, 268, 861.
(62) Mock, M. T.; Popescu, C. V.; Yap, G. P. A.; Dougherty, W. G.;
Riordan, C. G. Inorg. Chem. 2008, 47, 1889.
(27) Yandulov, D. V.; Schrock, R. R. Science (Washington, DC, U. S.)
2003, 301, 76.
(63) Scepaniak, J. J.; Young, J. A.; Bontchev, R. P.; Smith, J. M. Angew.
Chem., Int. Ed. 2009, 48, 3158.
(28) Castro-Rodriguez, I.; Nakai, H.; Zakharov, L. N.; Rheingold, A. L.;
Meyer, K. Science (Washington, DC, U. S.) 2004, 305, 1757.
(29) Fryzuk, M. D. Acc. Chem. Res. 2009, 42, 127.
(30) Knobloch, D. J.; Lobkovsky, E.; Chirik, P. J. Nat. Chem. 2010, 2,
30.
(64) Scepaniak, J. J.; Fulton, M. D.; Bontchev, R. P.; Duesler, E. N.; Kirk,
M. L.; Smith, J. M. J. Am. Chem. Soc. 2008, 130, 10515.
(65) Nieto, I.; Ding, F.; Bontchev, R. P.; Wang, H. B.; Smith, J. M. J. Am.
Chem. Soc. 2008, 130, 2716.
9
18116 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Slow Magnetic Relaxation
A R T I C L E S
Scheme 1
pursued hybrid ligand scaffolds that combine attributes of heme
and nonheme frameworks using trianionic tris(pyrrolyl-R-
methyl)amines.69 The addition of steric pickets to the [tpa]3-
platform serves to enforce an approximate three-fold symmetry
while preventing dimerization. Furthermore, the wide range of
potential ligand variants allows for facile tuning of the steric
and electronic properties of the corresponding metal complexes.
We recently disclosed the oxygen atom transfer chemistry of
the iron complexes [(tpaPh)Fe]- and [(tpaMes)Fe]-, demonstrating
intramolecular aromatic C-H hydroxylation by the former and
activation of nitrous oxide and intermolecular hydrogen atom
abstraction by the latter.69 In addition to this novel reactivity,
this ligand scaffold enforces a three-fold coordination geometry
about a high-spin S ) 2 iron center. Importantly, the resulting
electronic structure features three electrons in the 1e orbital set,
which leads to an unquenched orbital moment and thus the
potential for strong magnetic anisotropy. Indeed, we recently
demonstrated the efficacy of this strategy in our report of the
magnetic properties of K[(tpaMes)Fe].70 Magnetization measure-
ments on this compound revealed the presence of immense
uniaxial anisotropy, with an axial zero-field splitting parameter
of D ) -40 cm-1. Moreover, we demonstrated that this
anisotropy leads to slow relaxation effects under the presence
of a small applied dc field, with an effective relaxation barrier
of Ueff ) 42 cm-1, thereby providing the first example of a
mononuclear transition metal complex exhibiting single-
molecule magnet-like behavior.
The tunability of the [(tpaR)Fe]- platform affords the op-
portunity to expand this concept to other mononuclear transition
metal complexes. Herein, we report the design and synthesis
of a homologous series of [(tpaR)Fe]- complexes with various
aryl and alkyl substituents, where R ) tert-butyl (1), mesityl
(2), 2,4,6-triisopropylphenyl (3), phenyl (4), and 2,6-difluo-
rophenyl (5), as well as the structural, and magnetic properties
of this novel series of trigonal pyramidal iron(II) complexes.
Most importantly, we demonstrate the presence of strong
uniaxial magnetic anisotropy in the complexes that engenders
slow magnetic relaxation with barriers up to Ueff ) 65 cm-1 in
the case of 1. Moreover, our ability to vary the pendant
substituents of the ligand across the series has enabled us to
thoroughly examine the effect of factors such as ligand donor
strength and coordination geometry on governing magnetic
anisotropy and slow magnetic relaxation.
Results and Discussion
Synthesis of [tpaR] Ligands and Their Iron(II) Complexes.
Inspired by the initial report of tris(pyrrolyl-R-methyl)amine
(H3tpa)71 and a related indolyl system,72 we sought to expand
the steric and electronic properties of this platform into a family
of tripodal pyrrolide ligands. To this end, we prepared a series
of 2-alkyl- and 2-arylpyrroles and subjected these precursors
to Mannich reaction conditions adapted from the original
synthesis of H3tpa.71 Whereas in our hands 2-alkylpyrroles
containing pseudobenzylic protons (e.g., 2-methylpyrrole, 2-eth-
ylpyrrole) gave ill-defined polymeric products under these
reaction conditions, 2-tert-butylpyrrole (11) reacted smoothly
to generate the new alkyl substituted variant H3tpat-Bu (6) in
good yield. In the interest of facilitating the development of
this chemistry, we devised an improved synthesis of 2-tert-
butylpyrrole that provides access to large quantities of this
material (see Scheme 1). First, ethanolysis of 2-(trichloro-
acetyl)pyrrole affords ethyl pyrrole-2-carboxylate (16). Friedel-
Crafts alkylation of the pyrrole followed by saponification of
the ethyl ester and decarboxylation yields analytically pure 11
on scales upward of 30 g without the need for chromatography.
To further expand the utility of the [tpaR]3- ligand family, we
prepared several 2-arylpyrroles (2-mesitylpyrrole (12), 2-(2,4,6-
triisopropylphenyl)pyrrole (13) and 2-phenylpyrrole (14)) using
a published procedure for the direct Pd-catalyzed arylation of
sodium pyrrole73 and also extended this methodology to the
preparation of 2-(2,6-difluorophenyl)pyrrole (15) (see Scheme
1). All of these 2-arylpyrroles undergo triple Mannich conden-
sation to provide the respective ligands in moderate to good
yield. We have previously disclosed the syntheses of H3tpaMes
(7) and H3tpaPh (9),69 and the preparations of H3tpaTrip (8) and
its molybdenum complexes have been described recently.74
Vanadium complexes of [tpa ]
Mes 3- have also been reported.75,76
Deprotonation of the H3tpaR proligands in situ followed by
salt metathesis with FeCl2 in THF provides a general route to
anionic iron(II) complexes, as isolated in [M(solv)n][(tpaR)Fe]
(M ) Na, R ) tert-butyl (1), phenyl (4); M ) K, R ) mesityl
(2), 2,4,6-triisopropylphenyl (3), 2,6-difluorophenyl (5)) (see
(71) Shi, Y.; Cao, C.; Odom, A. L. Inorg. Chem. 2004, 43, 275.
(72) Tanski, J. M.; Parkin, G. Inorg. Chem. 2003, 42, 264.
(73) Rieth, R. D.; Mankad, N. P.; Calimano, E.; Sadighi, J. P. Org. Lett.
2004, 6, 3981.
(66) Govindaswamy, N.; Quarless, D. A.; Koch, S. A. J. Am. Chem. Soc.
1995, 117, 8468.
(67) Nguyen, D. H.; Hsu, H. F.; Millar, M.; Koch, S. A.; Achim, C.;
Bominaar, E. L.; Münck, E. J. Am. Chem. Soc. 1996, 118, 8963.
(68) Hwang, J.; Govindaswamy, K.; Koch, S. A. Chem. Commun.
(Cambridge, U. K.) 1998, 1667.
(74) Wampler, K. M.; Schrock, R. R. Inorg. Chem. 2007, 46, 8463.
(75) Betley, T. A.; Wu, Q.; Van Voorhis, T.; Nocera, D. G. Inorg. Chem.
2008, 47, 1849.
(69) Harman, W. H.; Chang, C. J. J. Am. Chem. Soc. 2007, 129, 15128.
(70) Freedman, D. E.; Harman, W. H.; Harris, T. D.; Long, G. J.; Chang,
C. J.; Long, J. R. J. Am. Chem. Soc. 2010, 132, 1224.
(76) Betley, T. A.; Surendranath, Y.; Childress, M. V.; Alliger, G. E.;
Fu, R.; Cummins, C. C.; Nocera, D. G. Philos. Trans. R. Soc. London,
Ser. B: 2008, 363, 1293.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18117

A R T I C L E S
Harman et al.
Scheme 2
Table 1. Summary of Bond Lengths (Å) and Angles (deg) for the
X-ray Structures of 1, 2, 4, and 5
Scheme 2). Counterion choice is dictated primarily by the
solubility and crystallinity of the product. Potassium hydride
proved to be a suitable base for the syntheses of 2, 3, and 5.
Sodium hydride was employed in the synthesis of 4 whereas
NaN(SiMe3)2 was used to furnish the desired sodium salt 1.
1
2
4
5
Fe1-N1
Fe1-N2
Fe1-N3
Fe1-N4
Fe1-Npyrrole
N2-Fe1-N3
N3-Fe1-N4
N4-Fe1-N2
Fe1-(N2,N3,N4)
Fe-Fe
2.144(1)
2.031(1)
2.172(2)
2.008(3)
2.041(2)
2.024(3)
2.024
117.36(9)
122.39(9)
115.3(1)
0.262
2.161(2)
2.016(1)
2.013(2)
2.019(2)
2.016
115.56(6)
120.22(7)
120.27(6)
0.233
2.196(2)
2.036(2)
2.042(3)
2.038(3)
2.039
121.6(1)
115.8(1)
116.6(1)
0.290
Solid-State Structures. Compounds 1-5 crystallize readily
as THF or DME solvates. Single crystal X-ray diffraction
measurements reveal four-coordinate, trigonal pyramidal iron
centers for each complex, as depicted in Figure 1. Although 3
crystallizes readily and in high yield, issues with twinning and
disorder have thus far impeded our efforts to obtain satisfactory
solutions to data collected on these crystals. Relevant bond
lengths and angles for 1, 2, 4, and 5 are listed in Table 1. The
Fe-Npyrrole distances range from 2.008(3) Å to 2.042(3) Å for
the series, consistent with anionic nitrogen ligation of high-
spin iron(II). The longer axial Fe-Namine distances vary over a
slightly larger range from 2.144(1) Å to 2.196(2) Å.
118.35(6)
0.263
9.482
10.651
8.927
9.277
at 300 K give values of ꢁMT ) 3.55, 3.32,70 3.68, 3.66, and
3.44 cm3mol/K for compounds 1-5, respectively (see Support-
ing Information Figures S1, S4, S9, and S14), confirming the
presence of a high-spin electron configuration and an S ) 2
spin ground state for each compound. Notably, each set of
ꢁMT data displays a downturn at low temperature, suggesting
the presence of significant zero-field splitting. To investigate
this possibility, we collected low-temperature magnetization
data at various applied dc fields for the compounds. The
resulting plot of reduced magnetization for 1, depicted in
Figure 2, reveals a series of nonsuperimposable isofield
curves that fall dramatically short of reaching the magnetiza-
tion saturation of 4.0 µB expected for an S ) 2 ground state
with g ) 2.0, confirming the presence of strong magnetic
The tert-butyl derivative 1 crystallizes in the cubic space
group P213 and is unique among iron(II) complexes 1-5 in
that it possesses crystallographically imposed three-fold sym-
metry at the iron center. The iron centers in 2, 4, and 5 exhibit
slight deviations from three-fold symmetry. The most pro-
nounced structural distortion is observed for 2, in which a
mesityl o-methyl group (C42) is rotated toward the iron center.
The Fe-Npyrrole bond directly opposite the iron-methyl close
contact is anomalously short at 2.008(3) Å.
Static Magnetic Properties. With structural data for this family
of trigonal pyramidal nonheme iron(II) complexes in hand, we
turned our attention to more comprehensively interrogating their
magnetic properties. Dc magnetic susceptibility measurements
Figure 1. Thermal ellipsoid plots of the trigonal pyramidal complexes in 1, 2, 4, and 5. Ellipsoids are shown at the 50% probability level. Orange, blue,
yellow, and gray ellipsoids represent Fe, N, F, and C atoms, respectively. Hydrogen atoms have been omitted for clarity.
9
18118 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Slow Magnetic Relaxation
A R T I C L E S
Table 2. Summary of Magnetic Parameters for 1-5a
1
2
3
4
5
g
D
|E|max
Ueff
2.3(1)b
-48
0.4
2.2(1)
-44(4)
6
2.4(1)
-30(2)
4
2.4(1)
-26(2)
5
2.0(1)
-6.2
0.1
s
65
42
s
25
.
a All energies are given in cm-1 b See Experimental Section for
explanation of uncertainties.
trigonal plane of the molecule. Importantly, this large negative
value of D, in conjunction with the high-spin S ) 2 ground
state, demonstrates the potential of this type of mononuclear
transition metal complex to exhibit slow magnetic relaxation.
Indeed, a compound exhibiting these parameters could display
a maximal thermal relaxation barrier of U ) S2|D| ) 192 cm-1
,
Figure 2. Low-temperature magnetization data for 1 collected under various
applied dc fields. The black lines represent fits to the data.
which would be by far the record barrier for a transition metal
system.5
Zero-field splitting parameters extracted from fits to reduced
magnetization data collected for 1-5 are enumerated in Table
2 (see Supporting Information Figures S5, S10, and S15). For
all of the compounds, the best fits to the data give negative
values of D, indicative of a uniaxial anisotropy and thus the
possibility of slow magnetic relaxation. The series of D values
ranges from D ) -48 cm-1 for 1 to D ) -6.2 cm-1 in the
case of 5. Inspection of the trend across the series reveals a
dependence of the magnitude of D on Lewis base strength,
where the magnitude of D rises with increasing basicity of the
ligand. This observation suggests that the magnitude of axial
anisotropy may be related to the energy separation between the
1e (dxz and dyz) and 2e (dxy and dx –y ) orbitals, as the energy of
the 2e orbitals will increase with the σ-donating ability of the
ligand. Taken together, these data establish that the trigonal
pyramidal [(tpaR)Fe]- system offers a general platform for
obtaining large uniaxial zero-field splitting.
Dynamic Magnetic Properties. To investigate the potential
for slow magnetic relaxation in this homologous series of high-
spin iron(II) compounds, we collected variable-frequency ac
susceptibility data at multiple temperatures. In the absence of
an applied dc field, no ꢁM′′ signals were observed at frequencies
up to 1500 Hz and temperatures down to 1.8 K. This result is
somewhat unexpected, given the large uniaxial anisotropy and
S ) 2 spin ground states determined for the compounds through
static magnetic measurements. One explanation for the absence
of ꢁM′′ signals is that quantum tunneling of the magnetization
through the thermal relaxation barrier dominates other relaxation
pathways in the absence of an applied dc field. Such tunneling
processes may arise due to the presence of transverse magnetic
anisotropy in the compounds. For a molecule exhibiting
rigorously axial anisotropy, the wave functions corresponding
to each (MS pair do not overlap with one another, such that no
mixing can occur between the two.3,79 In this case, quantum
tunneling is a forbidden process. If a transverse component to
the magnetic anisotropy is introduced, however, mixing between
these wave functions occurs. This mixing then enables the
magnetization of the +MS level to tunnel through the anisotropy
barrier to the -MS level, such that the overall relaxation time
is fast. As has been observed for a number of previously reported
compounds, tunneling effects can drastically reduce the relax-
ation time of a single-molecule magnet.20,79-83 Indeed, the
anisotropy. To quantify this effect, the data were modeled
according to the following spin Hamiltonian:
ˆ
ˆ2
ˆ2
ˆ2
H ) DSz + E(Sx - Sy) + gisoµBS·H
(1)
where S is the spin ground state, D is the axial zero-field
splitting parameter, E is the transverse zero-field splitting
parameter, giso is the average g-factor, µB is the Bohr
magneton, and H is magnetic field. Here, application of a
Hamiltonian considering second-order anisotropy originating
from zero-field splitting faces some limitation, as very strong
first-order anisotropy can sometimes be better modeled by
directly considering spin-orbit coupling of the ground state.
Indeed, a recent report treated the magnetization data for
compound 2 in such a fashion, employing a ligand field
Hamiltonian that assumes the anisotropy arises from orbital
degeneracy of the iron(II) ion.77 However, this approach
assumes rigorous three-fold site symmetry at the iron(II)
center in order to preserve the orbital degeneracy. This is
clearly not a valid assumption in the case of compounds 2-5,
as the crystal structure reveals that each complex undergoes
a significant Jahn-Teller distortion to break the three-fold
symmetry. Moreover, despite its apparent crystallographic
C3 axis at 128 K, compound 1 may undergo a similar Jahn-
Teller distortion at lower temperature or exhibit a more subtle
distortion that is obscured by the thermal motion and three-
fold symmetry of the crystal lattice. Thus, the zero-field
splitting model likely provides a better physical description
of the magnetic behavior of these complexes than does a
model that assumes perfect three-fold symmetry and true
orbital degeneracy.
2
2
Fits to the data obtained using ANISOFIT 2.078 give axial
and transverse zero-field splitting parameters of D ) -48 cm-1
and |E| e 0.4 cm-1, respectively, with g ) 2.28. The presence
of such strong axial anisotropy arises from the unquenched
orbital angular momentum associated with a 1e32e2a11 electronic
configuration.70 In contrast to this strong uniaxial anisotropy,
the value of E is small, over 2 orders of magnitude smaller
than D. The presence of such a large |D/E| ratio likely arises
due to the crystallographic three-fold symmetry at the iron(II)
center, which minimizes the magnetic anisotropy within the
(77) Palii, A. V.; Clemente-Juan, J. M.; Coronado, E.; Klokishner, S. I.;
Ostrovsky, S. M.; Reu, O. S. Inorg. Chem. 2010, 49, 8073.
(78) Shores, M. P.; Sokol, J. J.; Long, J. R. J. Am. Chem. Soc. 2002, 124,
2279.
(79) Gatteschi, D.; Sessoli, R. Angew. Chem., Int. Ed. 2003, 42, 268.
(80) Friedman, J. R.; Sarachik, M. P.; Tejada, J.; Ziolo, R. Phys. ReV.
Lett. 1996, 76, 3830.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18119

A R T I C L E S
Harman et al.
smaller ratios of |D/E| observed for 2-5 support the hypothesis
that quantum tunneling provides a facile relaxation pathway for
the compounds. Somewhat surprisingly, despite exhibiting
crystallographic three-fold symmetry and a miniscule value of
E, compound 1 relaxes rapidly on the timescale of the ac
susceptibility experiment under zero applied dc field. In addition
to slight deviations from ideal three-fold symmetry at the iron(II)
center in 1 that may be undetectable in the crystal structure
due to thermal motion of the nitrogen atoms at 128 K, this rapid
relaxation in the absence of an applied field is also likely the
result of tunneling. In this context, the tunneling probability is
79,84
known to increase with decreasing MS
and this monoiron
system has a relatively small spin ground state of S ) 2. Finally,
as the magnetic measurements were performed on microcrys-
talline solids, fast relaxation in zero-field may be facilitated by
spin-spin interactions between neighboring iron(II) ions. This
hypothesis is consistent with the Mo¨ssbauer spectra obtained
for these systems, which, as discussed below, reveal a temper-
ature independence of the relaxation time below 10 K.
If quantum tunneling and/or spin-spin relaxation effects are
in fact responsible for shortcutting of the thermal relaxation
barriers in the iron(II) complexes, then application of a dc field
during the ac measurement should act to split the energies of
the (MS pairs, thereby eliminating tunneling as a facile
relaxation pathway and slowing down the relaxation dynamics.
Indeed, just such an experiment leads to slow relaxation effects
for all compounds with the exception of 5. For instance, data
collected for 1 under a 1500 Oe dc field reveal a set of
temperature-dependent peaks in the plot of ꢁM′′ vs ν (see Figure
3, bottom). In order to extract relaxation times from these peaks,
we constructed Cole-Cole plots from data collected in the
temperature range 1.8-6.8 K and fit them to a generalized
Debye model.85-87 For a single-molecule magnet, the relaxation
time (τ) follows a thermally activated relaxation process where
τ increases exponentially with decreasing temperature. Accord-
ingly, the corresponding plot of ln(τ) vs 1/T should feature a
linear region, with the slope of that line giving the relaxation
energy barrier. Indeed, the Arrhenius plot constructed for 1 (see
Figure 3, top) features a linear region at high temperature, with
a least-squares fit giving Ueff ) 65 cm-1 and τ0 ) 6.7 × 10-11
s. The value of τ0 provides a quantitative measure of the attempt
time of relaxation from the thermal phonon bath, and the value
obtained here is comparable to those found in single-molecule
magnets.1-5 In addition, this value of τ0 eliminates the possibility
that phonon bottleneck effects lead to the observed slow
relaxation.88
Figure 3. Bottom: Variable-frequency out-of-phase ac susceptibility data
for 1, collected under a 1500 Oe dc field at various temperatures. The solid
lines are guides for the eye. Top: Arrhenius plot constructed from data
obtained under a dc field of 1500 Oe. The dashed lines represent data fits
to an Orbach (blue), Raman (purple), and direct (green) process. The solid
red line represents a data fit to the three processes simultaneously.
applied dc field, the relaxation is dominated by spin-lattice
interactions.89,90 For instance, at high temperature, the relaxation
time exhibits a clear Arrhenius dependence, with ln(τ) increasing
linearly with 1/T (see Figure 3, upper, dashed blue line). This
region is likely dominated by an Orbach relaxation process,
sometimes referred to as thermally assisted quantum tunneling
of the magnetization. Here, a spin associated with the MS )
+2 level absorbs a phonon and is excited to the MS ) +1 level.
Then, the spin tunnels from the MS ) +1 to MS ) -1 level
and subsequently relaxes to the MS ) -2 level. Note, though,
that this process cannot be the sole relaxation pathway operating
at these temperatures, because such a scenario would give a
relaxation barrier of Ueff ) 144 cm-1, the energy associated
with climbing from the MS ) +2 to MS ) +1 level. At low
temperature, ln(τ) also exhibits a linear dependence on 1/T, albeit
with a slope of nearly zero (see Figure 3, upper, dashed green
line). This region is likely dominated by ground-state tunneling
via a direct phonon-based relaxation process, as insufficient
thermal energy is available for a spin to relax via a thermally
assisted mechanism. In the case of 1, this process corresponds
to tunneling from the MS ) +2 to MS ) -2 level. Since the
ln(τ) data do not show a clear transition between a high
temperature Orbach region and a low temperature direct process
region, a Raman relaxation mechanism was considered in order
The large temperature range over which slow relaxation is
observed for 1 at 1500 Oe provides a comprehensive map of
the relaxation processes occurring within the molecule. At this
(81) Thomas, L.; Lionti, F.; Ballou, R.; Gatteschi, D.; Sessoli, R.; Barbara,
B. Nature 1996, 383, 145.
(82) Sangregorio, C.; Ohm, T.; Paulsen, C.; Sessoli, R.; Gatteschi, D. Phys.
ReV. Lett. 1997, 78, 4646.
(83) Lin, P. H.; Burchell, T. J.; Cle´rac, R.; Murugesu, M. Angew. Chem.,
Int. Ed. 2008, 47, 8848.
(84) Chudnovsky, E. M.; Tejada, J. Macroscopic Tunneling of the
Magnetic Moment; Cambridge University Press: Cambridge, UK,
1998.
(85) Cole, K. S.; Cole, R. H. J. Chem. Phys. 1941, 9, 341.
(86) Bo¨ttcher, C. J. F. Theory of Electric Polarisation; Elsevier: New York,
1952.
(87) Aubin, S. M. J.; Sun, Z. M.; Pardi, L.; Krzystek, J.; Folting, K.;
Brunel, L. C.; Rheingold, A. L.; Christou, G.; Hendrickson, D. N.
Inorg. Chem. 1999, 38, 5329.
(89) Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of
Transition Ions; Clarendon Press: Oxford, U. K., 1970.
(90) Carlin, R. L. Magnetochemistry; Springer-Verlag: Berlin, 1986.
(88) Schenker, R.; Leuenberger, M. N.; Chaboussant, G.; Loss, D.; Güdel,
H. U. Phys. ReV. B: Condens. Matter Mater. Phys. 2005, 72, 184403.
9
18120 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Slow Magnetic Relaxation
A R T I C L E S
to model the intermediate data (Figure 3, upper, dashed purple
line). In this process, the relaxation time scales with temperature,
intermediate between the exponential (Orbach) and direct
dependences on 1/T. The crossover between Arrhenius behavior
and direct tunneling through an intermediate Raman process is
prevalent among mononuclear and weakly exchange-coupled
multinuclear single-molecule magnets, yet previous fitting of
variable-temperature relaxation time data has included only the
Orbach region.20,70,83 To our knowledge, these data mark the
first time that the temperature dependence of the relaxation time
in a molecule has been fit to a non-Orbach mechanism (see
Figure 3, upper, solid red line).
As the temperature is decreased below 120 K, the spectral
profile of 1 broadens from a doublet to an asymmetric
absorption, ultimately reaching a broad complex shape at 4.2
K. The broadening of the spectra with decreasing temperature
results from the onset of slow magnetic relaxation down to 4.2
K, in accordance with the results of ac susceptibility experi-
ments. In order to quantify this effect, the spectra were fit
according to the Dattagupta and Blume formalism (see Sup-
porting Information for a detailed description of this fitting).93
Here, the relaxation of the magnetic hyperfine field, Heff, of the
iron(II) center is best modeled as occurring in 120° steps
perpendicular to the C3 axis of the molecule, along each
Fe-Npyrrole bond. Notably, the fit to the data reveals a hyperfine
field of Heff ) 5.31(6) T, a much smaller value than is commonly
observed for high-spin iron(II) complexes. This reduced value
is likely a result of the large unquenched orbital angular
momentum in 1, consistent with the strong magnetic anisotropy
determined through magnetization experiments. Indeed, a similar
phenomenon has been observed in the mixed-valence oxalates
(PPh4)[Fe2(ox)3] and (NBu4)[Fe2(ox)3].94
An Arrhenius plot of the relaxation times obtained from fitting
the spectra for 1 is shown in Figure 5. Between 4.2 and 10 K,
the relaxation time is essentially independent of temperature,
suggesting that the relaxation process is governed by spin-spin
relaxation and/or quantum tunneling under zero applied magnetic
field. In contrast, between 10 and 80 K, the relaxation time
demonstrates a strong temperature dependence, with a linear
least-squares fit providing a relaxation barrier of Ueff ) 9.8(5)
cm-1. This result indicates that a thermally activated Orbach
mechanism dominates the relaxation process above 10 K.
Interestingly, while a similar relaxation mechanism was ob-
served from ac susceptibility experiments, the relaxation barrier
of Ueff ) 9.8(5) cm-1 obtained from Mo¨ssbauer spectral fits is
much smaller than that of Ueff ) 65 cm-1 from magnetic
measurements. This difference may arise in large part due to
the presence of fast relaxation processes in zero field, such as
spin-spin relaxation and quantum tunneling, that serve to
shortcut the thermal relaxation processes. Nevertheless, these
data confirm that slow magnetic relaxation does indeed occur
for 1 in the absence of an applied magnetic field. Finally, it
should be noted that an adequate fit of the 4.2 and 10 K spectra
of 1 requires a second component of 13(1)% area with a faster
relaxation time of ca. 0.002 µs. This observation may indicate
that a small portion of the sample has a slightly different
coordination environment or a slightly different interaction with
the lattice and hence a differing extent of spin-lattice coupling.
Mo¨ssbauer spectra collected for compound 4 at selected
temperatures are shown in the right panel of Figure 4. Here,
the isomer shift of δ ) 0.858(2) mm/s and small negative
quadrupole splitting at 4.2 K confirm the presence of high-spin
iron(II) in a high-symmetry four-coordinate environment. In the
temperature range 80-220 K, the spectra consist of a narrow
doublet, indicative of fast magnetic relaxation. As in the case
of 1, simulations indicate that the hyperfine field of 4 must be
relaxing faster than 0.005 µs in this temperature range. As the
temperature is lowered, the doublet broadens to a complex peak
similar to that observed for 1 below 10 K. In contrast to that
observed for 1, however, the line shape continues to evolve until
We conducted similar ac susceptibility measurements on the
other [tpaRFe]- compounds (see Supporting Information Figures
S6-S8 and S11-S13). Arrhenius fits to relaxation times
extracted from Cole-Cole plots give relaxation barriers of Ueff
) 42 cm-1 for 2 and 25 cm-1 for 4. In the case of 3, we observe
a temperature- and frequency-dependent ꢁM′′ signal; however,
the corresponding relaxation times are not indicative of thermally
activated behavior. As such, no thermal relaxation barrier could
be obtained in the measured temperature range. With the
exception of 3, the magnitude of Ueff increases with increasing
value of D. The absence of slow relaxation observed for 5 is
likely a direct consequence of the low D value of the compound.
Mo¨ssbauer Spectroscopy. In view of the fast relaxation
processes that dominate the magnetization dynamics of the
iron(II) complexes under zero field at low temperature, we
carried out zero-field Mo¨ssbauer spectroscopy measurements
to probe these processes on a much faster time scale (ca. 0.01
µs). For these studies, we focused our attention on compounds
1, 2, and 4 in order to compare and contrast alkyl versus aryl
substitution on the ancillary tpa framework.91 Selected Mo¨ss-
bauer spectra of 1 obtained at various temperatures are shown
in the left panel of Figure 4. At 220, 180, and 120 K, the spectra
consist of a paramagnetic quadrupole doublet with an isomer
shift characteristic of iron(II) accounting for the majority of the
sample. A minor doublet is assigned to a small amount of an
iron(III)-containing impurity. The isomer shift of δ ) 0.85 mm/s
observed for 1 at 4.2 K is consistent with a high-spin iron(II)
ion in a four-coordinate environment.92 Notably, the quadrupole
splitting (∆EQ) observed for 1 is very small and positive. This
splitting arises from two contributions to the electric field
gradient, the lattice (qlat) and valence (qval) components. The
crystallographic three-fold symmetry of the iron complex in 1
generates a d-orbital manifold qualitatively similar to that of a
tetrahedron, a geometry for which qval is expected to be
essentially zero. In addition, the cubic symmetry of the lattice
in which 1 crystallizes likewise contributes to a highly uniform
electric field gradient. The presence of a simple quadrupole
doublet above 120 K indicates that the magnetic relaxation of
1 is fast relative to the Larmor precession time of the iron-57
nuclear magnetic moment (ca. 0.01 µs). Specifically, simulations
indicate that, for the hyperfine parameters observed for 1 at these
temperatures, the relaxation time must be less than 0.005 µs.
(91) The following discussion will emphasize the different relaxation times
observed for 1 and 4. The details of the relaxation model used to fit
the spectra and an expanded discussion of the hyperfine parameters
are given in the Supporting Information; the temperature dependence
of the spectral parameters is shown in Figure S17 and the spectral
fit parameters are given in Tables S1-S3.
(92) Reiff, W. M.; Long, G. J. In Mo¨ssbauer Spectroscopy Applied to
Inorganic Chemistry; Long, G. J., Ed.; Plenum Press: New York,
1984; Vol. 1, p 245.
(93) Dattagupta, S.; Blume, M. Phys. ReV. B: Condens. Matter Mater.
Phys. 1974, 10, 4540.
(94) Carling, S. G.; Visser, D.; Hautot, D.; Watts, I. D.; Day, P.; Ensling,
J.; Gütlich, P.; Long, G. J.; Grandjean, F. Phys. ReV. B: Condens.
Matter Mater. Phys. 2002, 66, 104407.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18121

A R T I C L E S
Harman et al.
Figure 4. Variable temperature Mo¨ssbauer spectra of 1 and 4.
relaxation even slower than that observed in 1. To further
investigate this phenomenon, the spectra were modeled similarly
to those for complex 1 (see Supporting Information). Here,
however, acceptable fits of the relaxation spectra obtained in
the temperature range 12.5-20 K required two spectral com-
ponents with different relaxation times (see Figure 4). The need
for two relaxation times may indicate the presence of two
distinct or a small distribution of iron(II) coordination environ-
ments in 4, perhaps as a result of partial DME desolvation.
Similar to that observed for 1, the fits reveal an unusually small
hyperfine field of H ) 4.95(1) T, indicative of strong magnetic
anisotropy. The corresponding Arrhenius plots of relaxation time
for the two modes both show a thermally activated behavior
(see Figure 5), demonstrating that the hyperfine field of the
iron(II) center is relaxing via an Orbach process in this
temperature range. Least-squares fits to the two data sets give
relaxation barriers of 26(2) and 75(4) cm-1. Remarkably, the
barrier of 26(2) cm-1 is in accordance with that obtained from
ac susceptibility data obtained for compound 4 at much lower
Figure 5. Arrhenius plot of relaxation time, as obtained from Mo¨ssbauer
measurements, for 1 (red circles) and two relaxation processes for 4 (green and
blue triangles). Solid black lines represent linear least-squares fits to the data.
finally forming a well-resolved sextet below 10 K. The
observation of this sextet demonstrates the presence of magnetic
9
18122 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Slow Magnetic Relaxation
A R T I C L E S
temperatures under a dc field of 1500 Oe. In contrast, the second
process, corresponding to a barrier of 75(4) cm-1, was unde-
tected in our ac measurements with or without an applied dc
field. Note that the magnitude of this barrier corresponds exactly
to the energy separation between the |MS| ) 2 and |MS| ) 1
levels, considering the axial zero-field splitting parameter of D
) -26(2) cm-1 obtained from fitting magnetization data and
the expression ∆E ) |(MS2)2D| - |(MS1)2D|. This barrier may
thus represent a pure Orbach mechanism that is quenched at
low temperatures or upon application of a dc field. Owing to
the large separation between the MS ) (2 and (1 levels, there
is no substantial population of the MS ) (1 levels up to ca. 40
K, a temperature at which the relative Boltzmann population
of the MS ) (1 levels is 6%. Hence, a single sharp sextet is
observed for 4 at 4.2 K.
In contrast to the relaxation observed in the Mo¨ssbauer spectra
of 1 and 4, the spectra of 2 obtained between 4.2 and 270 K
consist of a paramagnetic doublet with a splitting that decreases
from ∆EQ ) 1.30 mm/s at 4.2 K to ∆EQ ) 0.67 mm/s at 270
K (see Supporting Information Figure S16). This splitting is
much larger than that observed for 1 and 4, and this difference
could result from the more distorted coordination environment
for the iron(II) center. The absence of slow magnetic relaxation
in 2, in stark contrast to the dynamic behavior observed in ac
magnetic susceptibility experiments, may stem from desolvation
in the compound, which loses two molecules of DME when
removed from its mother liquor. Such desolvation leads to a
loss of crystallinity, which likely decreases the average Fe · · ·Fe
distance, thereby facilitating fast magnetic relaxation through
spin-spin interactions.
High-Field Electron Paramagnetic Resonance Spectroscopy.
Finally, in order to probe directly the energy separation between
the MS levels and thus the zero-field splitting of the iron(II)
complexes, we carried out high-field EPR experiments on
polycrystalline samples of 1-5. Variable-field spectra collected
for the difluorophenyl-substituted compound 5 at 30 K and
various frequencies (see Figure 6) reveal the presence of up to
two transitions at each frequency, excluding the omnipresent
peak at g ≈ 2 that likely stems from a small impurity. The low-
field peak corresponds to the forbidden MS ) 2 f MS ) -2
transition, whereas the high-field peak corresponds to the
allowed MS ) -2 f MS ) -1 transition. We note that the
∆MS ) 4 transition in iron(II) complexes may sometimes be
observed even in X-band EPR spectra owing to the fact that
the MS ) +2 and -2 levels are split only by approximately
3E2/D, while splittings between other MS levels are of the order
of E or D. To quantify the energies of the observed transitions,
we used the frequency dependence of the peaks to construct a
plot of resonance field vs frequency (see Figure 7). The data
were then fit according to the following spin Hamiltonian:
Figure 6. Top: Variable-field EPR spectrum for 5 collected at 10 K and
224 GHz. The asterisk denotes an impurity positioned at g ≈ 2. Inset:
Expanded view of the high-field portion of the spectrum. Bottom: Spectra
for 5 collected at 30 K and frequencies of 56 (red), 112 (green), 224 (blue),
305 (purple), and 416 (magenta) GHz.
Figure 7. Resonant field vs frequency plot for 5, constructed from data
obtained at 30 K. Solid lines represent fits to the data, with x (red), y (blue),
and z (green) turning points, to give |D| ) 4.397(9) cm-1, |E| ) 0.574(9)
cm-1 and g ) 2.20.
ˆ
ˆ2
ˆ2
ˆ2
H ) µBH·g·S + D(Sz - S(S + 1)/3) + E(Sx - Sy)
(2)
The best fits to the data provide zero-field splitting parameters
of |D| ) 4.397(9) cm-1 and |E| ) 0.574(9) cm-1, with g ) 2.20.
These parameters are in reasonable agreement with those of D
) -6.2 cm-1 and |E| ) 0.1 cm-1 obtained from fits to reduced
magnetization data. In order to ascertain the sign of D for
compound 5, spectra were collected at 224 GHz at various
temperatures. As shown in Supporting Information Figure S18,
the intensity of the ∆MS ) 4 transition dramatically increases
as the temperature is decreased. In contrast, the intensity of the
allowed ∆MS ) 1 transition increases as the temperature is
increased. These observations indicate the presence of an MS
manifold in which the largest values of MS are lowest in energy,
corresponding to a negative D value. This uniaxial anisotropy
is consistent with the results from magnetization measurements.
High-field EPR spectra obtained for complexes 1-4 display
the forbidden ∆MS ) 4 transition, but we observe no well-
resolved peaks corresponding to allowed transitions at magnetic
fields up to 14 T and frequencies up to 600 GHz. Since a
transition corresponding to MS ) 2 f MS ) -2 does not
provide information regarding the separation between MS levels
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18123

A R T I C L E S
Harman et al.
operating at 300 or 400 MHz as noted. Chemical shifts are reported
in ppm and referenced to residual protiated solvent; coupling
constants are reported in Hz. Mass spectra and elemental analyses
were performed at the Mass Spectrometry and Microanalytical
Facilities at the University of California, Berkeley.
with different absolute quantum numbers, we were unable to
extract zero-field splitting parameters from these data. However,
the absence of observable allowed transitions in these spectra
is not surprising given the large axial zero-field splitting
parameters of these four [(tpaR)Fe]- compounds relative to 5.
For instance, according to fits to the magnetization data,
compound 4 exhibits the next smallest magnitude of D across
the series, with D ) -26(2) cm-1. The lowest-field allowed
transition expected for a molecule with this value occurs at ca.
12 T at 224 GHz. Moreover, this transition is located ca. 100
K higher in energy than the corresponding ∆MS ) 4 transition
at 12 T, which indicates that an experimental temperature of
100 K would be necessary to observe a peak. Unfortunately,
data obtained at such a high temperature show a very low signal/
noise ratio, such that any peak in the spectrum is likely lost in
the background.
Magnetic Measurements. Magnetic data were collected using
a Quantum Design MPMS-XL SQUID magnetometer. Measure-
ments for 1-5 were obtained for finely ground microcrystalline
powders restrained in a frozen eicosane matrix within polycarbonate
capsules. Dc susceptibility measurements were collected in the
temperature range 2-300 K under a dc field of 1000 Oe. Dc
magnetization measurements were obtained in the temperature range
1.8-10 K under dc fields of 1, 2, 3, 4, 5, 6, and 7 T. These data
were fit in the temperature range 1.8-3.0 K. In general, several
different values of E could be obtained and had little to no effect
on the goodness-of-fit, depending only on the input values for E.
As such, only the maximum values of E are reported. In addition,
in cases where multiple fits of similar quality provided slightly
different values of D, the average value is reported with the standard
deviation given in parentheses. Values of g are reported to only
two significant figures to be consistent with D values and to account
for any small errors in sample weighing. Ac susceptibility measure-
ments were obtained in the temperature range 1.7-6.8 K under a
4 Oe ac field oscillating at frequencies of 1-1488 Hz, under an
applied dc field of 1500 Oe. Dc magnetic susceptibility data were
corrected for diamagnetic contributions from the sample holder and
eicosane, as well as for the core diamagnetism of each sample
(estimated using Pascal’s constants).
Mo¨ssbauer Spectroscopy. The Mo¨ssbauer spectra of compounds
1, 2, and 4 have been measured between 4.2 and 220, 280, and
250 K, respectively, in a Janis Supervaritemp cryostat with a
constant-acceleration spectrometer that utilized a rhodium matrix
cobalt-57 source and was calibrated at 295 K with R-iron powder.
The Mo¨ssbauer spectral absorbers of 1, 2, and 4 contained 60(5),
85(5), and 35(5) mg/cm2, respectively, of powder mixed with boron
nitride; the errors are high because of the difficulty of preparing
the absorbers under an inert atmosphere. The statistical errors are
given in parentheses in the text and tables. However, more realistic
absolute errors for the isomer shifts are (0.005 mm/s, for the
quadrupole shifts and line widths are (0.01 mm/s, and for the
relative component areas are (1%.
General Methods for X-ray Crystallography. Crystals were
mounted on Kaptan or monofilament loops in Paratone-N hydro-
carbon oil. Air-sensitive samples were transferred from the glovebox
to Paratone-N and mounted quickly to avoid decomposition. All
data collection was performed on a Bruker (formerly Siemens)
SMART diffractometer/CCD area detector equipped with a low
temperature apparatus. Data integration was performed using
SAINT. Preliminary data analysis and absorption correction were
performed with the Bruker APEX2 software package. Structure
solution by direct methods was performed using SIR2004,96 and
the resulting solution was refined using SHELX. Hydrogen atoms
were included in calculated positions.97 Details of the data collection
and refinement for 1, 2, 4, and 5 can be found in Table 3.
High-Field Electron Paramagnetic Resonance Spectroscopy.
High-field, high-frequency EPR spectra at temperatures ranging
from ca. 3 to 290 K were recorded on a home-built spectrometer
at the Electron Magnetic Resonance facility of National High
Magnetic Field Laboratory. The setup of this instrument has been
described in detail previously. The instrument is a transmission-
type device in which microwaves are propagated in cylindrical
lightpipes. The microwaves are generated by a phase-locked
Virginia Diodes source, generating a frequency of 13 ( 1 GHz
and producing its harmonics of which the 2nd, 4th, 6th, 8th, 16th,
Concluding Remarks
In summary, we have described the synthesis and properties
of a novel family of nonheme trigonal pyramidal iron(II)
pyrrolide complexes exhibiting considerable magnetic anisotropy
and slow magnetic relaxation. These coordinatively unsaturated
high-spin S ) 2 iron(II) compounds have been characterized
by X-ray crystallography, static and dynamic magnetic suscep-
tibility measurements, Mo¨ssbauer spectroscopy, and high-field
EPR spectroscopy. Through systematic modification of a
conserved three-fold symmetric trispyrrolide [tpaR]3- core with
sterically demanding alkyl and aryl substituents, we reveal a
range of large uniaxial zero-field splittings and relaxation
barriers for the corresponding [(tpaR)Fe]-. In addition, high-
field EPR measurements provide direct independent evidence
for negative zero-field splittings in this mononuclear high-spin
iron(II) platform. Moreover, through the use of more sterically
encumbering and more Lewis basic alkyl pendants to prevent
distortions from three-fold symmetry and engender stronger axial
anisotropy, we have successfully increased the barrier to spin
inversion up to Ueff ) 65 cm-1 for compound 1. Taken together,
these studies establish the first class of mononuclear transition
metal complexes in which slow magnetic relaxation has been
observed, and the collective structural and spectroscopic data
provide a starting point for further studies into the reactivity
and magnetic properties of this promising family of bioinspired
coordination compounds.
Experimental Section
Synthetic Materials and Methods. Unless otherwise noted, all
manipulations were carried out at room temperature under an
atmosphere of dinitrogen in a Vacuum Atmospheres glovebox or
using Schlenk techniques. Pentane, dimethoxyethane (DME),
tetrahydrofuran (THF), and diethyl ether were deoxygenated by
sparging with dinitrogen and dried via Vacuum Atmospheres solvent
purification system. Diisopropyl ether was distilled from purple
sodium/benzophenone ketyl. Dry 1,2-dichloroethane was purchased
from Acros, and FeCl2 beads were purchased from Strem. Potassium
hydride was purchased as a suspension in mineral oil, washed with
pentane, and used as a dry solid inside the glovebox. Literature
procedures were used for the preparation of ethyl pyrrole-2-
carboxylate (6),95 K[(tpaMes)Fe] (2),69 tris((5-phenyl-1H-pyrrol-2-
yl)methyl)amine (H3tpaPh) (7),69 and tris((5-(2,4,6-triisopropylphenyl)-
1H-pyrrol-2-yl)methyl)amine (H3tpaTrip) (8).74 All other reagents
and solvents were purchased from chemical suppliers and used as
received. NMR spectra were recorded on Bruker spectrometers
(96) Burla, M. C.; Caliandro, R.; Camalli, M.; Carrozzini, B.; Cascarano,
G. L.; De Caro, L.; Giacovazzo, C.; Polidori, G.; Spagna, R. J. Appl.
Crystallogr. 2005, 38, 381.
(97) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2008,
64, 112.
(95) Bailey, D. M.; Johnson, R. E.; Albertson, N. F. Org. Synth. 1988,
50-9, 618.
9
18124 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Slow Magnetic Relaxation
A R T I C L E S
Table 3. Experimental Details for the X-ray Structures of Complexes 1, 2, 4, and 5
1
2a
4
5
formula
formula weight
T (K)
C31H47N4ONaFe
570.57
128(2)
C58H85N4O8KFe
1061.28
141(2)
C45H57N4O6NaFe
828.79
123(2)
C49H61N4O8F6KFe
1042.97
153(2)
λ (Å)
0.71069
cubic
0.71069
monoclinic
P21/c (#14)
17.625(5)
13.840(4)
24.177(6)
90
95.855(4)
90
5867(2)
4
0.71069
triclinic
P1 (#2)
0.71069
orthorhombic
P212121 (#19)
13.467(2)
18.827(4)
20.082(4)
90
crystal system
space group
a (Å)
P213 (#198)
14.7696(3)
14.7696(3)
14.7696(3)
90
11.3600(17)
12.0280(18)
17.8810(26)
76.897(2)
73.871(2)
69.850(2)
2190.03(21)
2
b (Å)
c (Å)
R (deg)
ꢀ (deg)
γ (deg)
V (Å3)
90
90
90
90
3221.86(1)
4
1.18
0.509
1224
5091.7(16)
4
1.36
0.455
2184
Z
F
calcd (g/cm3)
1.20
0.382
2280
1.26
0.407
880
µ(mm-1
F(000)
)
reflctns collected
ind reflctns (Rint)
Tmin/Tmax
18416
2223 (0.028)
0.87
26424
6050 (0.032)
0.85
16928
8139 (0.030)
0.92
33040
10400 (0.061)
0.88
data/restr/params
R, Rw, Rall
GOF
2223/0/130
0.028, 0.076, 0.031
1.10
6287/0/649
0.040, 0.047, 0.077
1.66
8139/0/520
0.063, 0.070, 0.12
1.01
10400/0/630
0.048, 0.092, 0.086
0.98
max shift/error
0.01
0.00
0.07
0.01
a Data taken from ref 70.
24th, and 32nd are available. A superconducting magnet (Oxford
Instruments) capable of reaching a field of 17 T was employed.
Least-Squares Fitting of the High-Field EPR Data.98 The
program used to fit the multifrequency data calculated the energies
of the sublevels within the S ) 2 spin ground state through
diagonalization of the spin Hamiltonian matrix using the House-
holder transformation.99 The resonance fields at each orientation
of the magnetic field defined by the polar angles Θ and Φ were
found by using an iterative procedure. Our instrument employs no
resonance cavity; thus, we observe resonances caused by micro-
waves whose Bmicrowave is parallel to Bmagnet in addition to the usual
EPR transitions, owing to the poorly defined microwave modes.
Upon finding by trial and error method the initial parameters g, D,
and E, all transitions observed in all experimental powder spectra
were identified as corresponding to the X, Y, or Z orientations. A
table whose rows contained frequency, Θ, Φ, and resonance field
was used as input data. The program attempted to minimize the
function:
The derivatives of the resonance fields, fkcalc, with respect to the
parameters p, had to be evaluated numerically. Finally, errors in
the best-fit parameters were estimated as:
ꢁ2
σi )
(H-1
)
ii
ꢀN - P
where N is the number of experimental resonance fields and P is
the number of fitted parameters.
Ethyl 5-tert-Butylpyrrole-2-carboxylate (9). This compound
was prepared by the modification of a literature procedure.100
A
dry, round-bottom flask was charged with a stir bar, 6 (30.0 g, 0.216
mol), and 2 L of anhydrous 1,2-dichloroethane. The flask was
purged with dinitrogen, and AlCl3 (60.5 g, 0.454 mol) was added
in one portion followed by the immediate addition of 2-chloro-2-
methylpropane (23.7 mL, 0.216 mol). The resulting mixture was
stirred under nitrogen for 2 h and then quenched in air by careful
addition to a saturated solution of aqueous NaHCO3 (2 L). Diethyl
ether (1 L) was added and the organic layer separated. The aqueous
layer was then further extracted with ether (2 × 250 mL). The
combined organic portions were dried over anhydrous Na2SO4 and
concentrated by rotary evaporation to yield a pale brown oil which
crystallized upon standing to give 37.5 g (89%) of an off-white
solid which was used without further purification. The spectral
properties of this material were identical to those reported in the
literature.101
N
ꢁ2 )
(f (calc) - f i(exp) 2
)
i
∑
i)1
by use of the Simplex method. After convergence had been
achieved, the Hessian matrix was calculated. The Hessian matrix
is formally a matrix containing the second derivatives of ꢁ2 with
respect to parameters p:
2-tert-Butylpyrrole (10). A slurry of 9 (65.5 g, 0.335 mol) and
powdered NaOH (67.0 g, 1.68 mol) in ethylene glycol (650 mL)
was brought to reflux with the aid of a heating mantle. After 6 h,
the reaction mixture was allowed to cool to room temperature,
diluted with water (800 mL), and extracted with CH2Cl2 (5 × 200
mL). The organic portions were dried over anhydrous Na2SO4 and
concentrated by rotary evaporation to yield a brown oil which was
vacuum distilled into a flask cooled to -78 °C to give 35.0 g (85%)
δ2ꢁ2
Hij )
δpiδpj
but, as it is generally done, it was calculated from:
δf ckalc δf ckalc
N
H )
∑
ij
(
)(
)
δpi
δpj
k)1
(100) Wood, J. E.; Wild, H.; Rogers, D. H.; Lyons, J.; Katz, M.; Caringal,
Y.; Dally, R.; Lee, W.; Smith, R. A.; Blum, C. U.S. Patent 6,187,799,
2001.
(101) Elder, T.; Gregory, L. C.; Orozco, A.; Pflug, J. L.; Wiens, P. S.;
Wilkinson, T. J. Synth. Commun. 1989, 19, 763.
(98) Aromi, G.; Telser, J.; Ozarowski, A.; Brunel, L. C.; Stoeckli-Evans,
H. M.; Krzystek, J. Inorg. Chem. 2005, 44, 187.
(99) Wilkinson, J. H. The algebraic eigenValue problem; Oxford Uni-
versity Press, USA, 1988.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18125

A R T I C L E S
Harman et al.
of a colorless crystalline solid. The spectral properties of this
material were identical to those reported in the literature.102
Tris(5-tert-butyl-1H-pyrrol-2-ylmethyl)amine, H3tpat-Bu (6).
To a solution of NH4Cl (1.50 g, 28.0 mmol) in an ethanol/water
mixture (1:1 v/v, 60 mL) was added aqueous formaldehyde (37 wt
%, 6.80 g, 83.7 mmol) followed by a solution of 10 (10.3 g, 83.7
mmol) in ethanol (30 mL). The resulting mixture was stirred for
three days at room temperature under nitrogen at which point a
fine white precipitate was collected by filtration, washed with a
small portion of ethanol, briefly air-dried, and dissolved in CH2Cl2
(200 mL). The solution was washed with 20% aqueous NaOH (200
mL) and the aqueous layer extracted further with CH2Cl2 (2 × 100
mL). The organic portions were combined, dried over anhydrous
Na2SO4, and concentrated to a pale yellow oil that was dried in
vacuo to give a pale yellow foam. The foam was triturated with 5
mL of hexanes and filtered to yield 7.1 g (60%) of the title
compound as a fine white powder. 1H NMR (400 MHz, CDCl3): δ
8.13 (s, 3 H), 5.94 (t, J ) 2.8 Hz, 3 H), 5.84 (t, J ) 2.8 Hz, 3 H),
3.52 (s, 6 H), 1.30 (s, 9 H); 13C NMR (CDCl3, 100 MHz) δ 142.1,
126.7, 107.7, 102.3, 49.8, 31.6, 30.8.
0.104 mmol), Pd2dba3 (17.4 mg, 0.0190 mmol), and 1-bromo-2,6-
difluorobenzene (0.974 g, 5.05 mmol) were added sequentially. The
reaction vessel was sealed, removed from the glovebox, and heated
in an oil bath at 100 °C for 48 h. After the mixture was cooled to
ambient temperature, diethyl ether (100 mL) and water (100 mL)
were added to the reaction, and the dark mixture was filtered through
Celite, washing the residue with diethyl ether (30 mL). The organic
portion of the filtrate was separated and the aqueous layer extracted
further with diethyl ether (3 × 75 mL). The combined organic
portions were dried over Na2SO4 and concentrated by rotary
evaporation. Purification of the resultant residue by column
chromatography on silica gel (5% ethyl acetate/hexanes) provided
0.485 g (54%) of the title compound as a slightly orange oil. The
spectral properties of this material were identical to those reported
in the literature.103
Tris((5-(2,6-difluorophenyl)-1H-pyrrol-2-yl)methyl)amine,
H3tpaDFP (16). To a solution of NH4Cl (0.183 g, 3.42 mmol) in
an ethanol/water mixture (1:1 v/v, 20 mL) was added 37 wt %
aqueous formaldehyde (0.836 g, 10.3 mmol) followed by a solution
of 15 (1.84 g, 10.3 mmol) in ethanol (15 mL). The resulting mixture
was stirred for three days at room temperature under nitrogen at
which point a fine white precipitate was collected by filtration,
washed with a small portion of ethanol, briefly air-dried, and
dissolved in CH2Cl2 (100 mL). The solution was washed with 20%
aqueous NaOH (100 mL) and the aqueous layer extracted further
with CH2Cl2 (2 × 75 mL). The organic portions were combined,
dried over anhydrous Na2SO4, and concentrated to a colorless oil
that crystallized on standing to yield 0.661 g (33%) of a colorless
solid. 1H NMR (400 MHz, CDCl3): δ 9.23 (s, 3 H), 7.08 (m, 3 H),
6.97 (t, J ) 22 Hz, 6 H), 6.84 (s, 3 H), 6.27 (s, 3 H) 3.71 (s, 6 H);
13C NMR (CDCl3, 100 MHz) δ 159.3 (dd, J ) 9, 246 Hz), 130.1
(s), 125.8 (t, J ) 12 Hz), 120.6 (s), 112.6 (s), 112.2 (d, J ) 27
Hz), 110.6 (t, J ) 15 Hz), 109.52 (s), 50.0 (s); 19F (CDCl3, 376
MHz) δ -112.6 (s). HRFABMS ([M + 1]+) m/z calcd for
C33H25N4F6 591.1983, found 591.1985.
Na[(tpat-Bu)Fe]·THF (1). To a solution of 11 (664 mg, 1.57
mmol) in THF (10 mL) was added solid NaN(SiMe3)2 (865 mg,
4.71 mmol). After the mixture was stirred for 2 h, the volatiles
were removed under reduced pressure, and the colorless residue
was redissolved in THF (10 mL). Solid FeCl2 (199 mg, 1.57 mmol)
was added and the resulting slurry stirred for 6 h. Precipitated
NaCl was removed by filtration over Celite, and the filtrate was
concentrated under reduced pressure to a colorless glaze. Redis-
solution of this residue in minimal THF followed by layering with
pentane deposited colorless tetrahedral crystals which were washed
with pentane and dried in vacuo to yield 456 mg (51%) of a
microcrystalline solid. Anal. Calcd for C31H47FeN4NaO: C, 65.26;
H, 8.30; N, 9.79. Found: C, 64.83; H, 8.43; N, 9.55.
K[(tpaTrip)Fe]·3 DME (3). To a stirring solution of 8 (514 mg,
0.597 mmol) in THF (10 mL) was added solid KH (144 mg, 3.59
mmol) in ca. 10 portions, resulting in vigorous effervescence. After
the mixture was stirred for 2 h, effervescence had ceased and the
slurry was filtered to remove excess KH. Solid FeCl2 (76 mg, 0.60
mmol) was added to the filtrate, and the resulting slurry was stirred
for 6 h. Precipitated KCl was removed by filtration over Celite,
and the filtrate was concentrated under reduced pressure to a
colorless glaze. Dissolving this residue in minimal 1:1 DME/i-Pr2O
and layering with pentane deposited colorless, block-shaped crystals
which were washed with pentane and dried in vacuo to yield 490
mg (67%) of a white powder. Anal. Calcd for C72H111FeKN4O6:
C, 70.67; H, 9.14; N, 4.58. Found: C, 70.44; H, 9.31; N, 4.81.
Na[(tpaPh)Fe]·3 DME (4). To a stirring solution of 7 (486 mg,
1.01 mmol) in THF (10 mL) was added solid NaH (145 mg, 604
mmol) in ca. 10 portions, resulting in vigorous effervescence. After
the mixture was stirred for 2 h, effervescence had ceased and the
slurry was filtered to remove excess NaH. Solid FeCl2 (128 mg,
1.01 mmol) was added to the filtrate, and the resulting slurry was
stirred for 6 h. Precipitated NaCl was removed by filtration over
Celite, and the filtrate was concentrated under reduced pressure to
a yellow glaze. Dissolution of this residue in a minimal amount of
THF followed by layering with DME deposited yellow crystals
which were washed with DME and dried in vacuo to afford 475
mg (57%) of yellow, microcrystalline solid. Anal. Calcd for
C45H57FeN4NaO6: C, 65.21; H, 6.93; N, 6.76. Found: C, 65.14; H,
6.91; N, 6.86.
K[(tpaDFP)Fe]·2 DME (5). To a stirring solution of 16 (698
mg, 1.18 mmol) in THF (10 mL) was added solid KH (284 mg,
7.09 mmol) in ca. 10 portions resulting in vigorous effervescence.
After the mixture was stirred for 2 h, effervescence had ceased
and the slurry was filtered to remove excess KH. Solid FeCl2 (150
mg, 1.18 mmol) was added to the filtrate, and the resulting slurry
was stirred for 6 h. Precipitated KCl was removed by filtration over
Celite, and the filtrate was concentrated under reduced pressure to
a yellow glaze. Dissolution of this residue in a minimal amount of
DME followed by layering with pentane deposited yellow crystals
which were washed with pentane and dried in vacuo to afford 575
mg (56%) of a yellow powder. Anal. Calcd for C41H41N4O4F6KFe:
C, 57.08; H, 4.79; N, 6.49. Found: C, 57.29; H, 4.56; N, 6.58.
Acknowledgment. We thank the Packard foundation (C.J.C.),
DOE/LBNL (403801 to C.J.C.), the NSF (CHE-0617063 to J.R.L.),
and the FNRS-Belgium (9.456595 and 1.5.064.05 to F.G.) for
funding this work. We also thank Arkema (W.H.H.) as well as Tyco
Electronics (T.D.H. and D.E.F.) for graduate fellowship support.
C.J.C. is an Investigator with the Howard Hughes Medical Institute.
The NHMFL is funded by the NSF through the Cooperative
Agreement No. DMR-0654118, the State of Florida, and the
DOE. F.G. thanks Prof. R. Cloots for the use of his controlled
atmosphere glove box.
2-(2,6-Difluorophenyl)-1H-pyrrole (15). This compound was
prepared by the modifcation of a literature procedure.73 In the
glovebox, sodium pyrrole (1.35 g, 15.2 mmol), ZnCl2 (2.07 g, 15.2
mmol), and THF (35 mL) were combined in a heavy-walled reaction
vessel and allowed to stir for 5 min. (Caution: vigorously
exothermic.) Then 2-(di-tert-butylphosphino)biphenyl (31.1 mg,
Supporting Information Available: Additional magnetic data,
Mössbauer spectra and experimental details. This information
is available free of charge via the Internet at http://pubs.acs.org.
JA105291X
(103) Kruse, C. G.; Bouw, J. P.; Vanhes, R.; Vandekuilen, A.; Denhartog,
(102) Spaggiari, A.; Vaccari, D.; Davoli, P.; Prati, F. Synthesis 2006, 995.
J. A. J. Heterocycles 1987, 26, 3141.
9
18126 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010