Synthesis and characterization of a new family of bi-, tri-, tetra-, and pentanuclear ferric complexes

Inorg. Chem. 2004, 43, 5053−5068

Synthesis and Characterization of a New Family of Bi-, Tri-, Tetra-, and

Pentanuclear Ferric Complexes

,†

Colette Boskovic,* Andreas Sieber,^†Gre´gory Chaboussant,^†Hans U. Gu1del,^†Ju1rgen Ensling,^‡

Wolfgang Wernsdorfer,^§Antonia Neels,^|Gael Labat,^|Helen Stoeckli-Evans,^|and Stefan Janssen

Departement fu¨r Chemie und Biochemie, UniVersita¨t Bern, Freiestrasse 3, Bern CH-3012,

Switzerland, Institut fu¨r Anorganische Chemie und Analytische Chemie, Johannes Gutenberg

UniVersita¨t, Staudingerweg 9, D-55099 Mainz, Germany, Laboratoire L. Neel-CNRS, BP 166,

25 AVenue des Martyrs, 38042 Grenoble Cedex 9, France, Laboratoire de Cristallographie,

Institut de Chimie, UniVersite´ de Neuchaˆtel, AVenue de BelleVaux 51,

C.P. 2, Neuchaˆtel CH-2007, Switzerland, and Laboratory for Neutron Scattering, PSI and ETHZ,

Villigen PSI CH-5232, Switzerland

Received March 25, 2004

Nine members of a new family of polynuclear ferric complexes have been synthesized and characterized. The

reaction of Fe(O CMe)₂with polydentate Schiff base proligands (H L) derived from salicylidene-2-ethanolamine,

2

followed in some cases by reaction with carboxylic acids, has afforded new complexes of general formulas [Fe₂-

(pic)₂(L)₂] (where pic^-is the anion of 2-picolinic acid), [Fe₃(O CMe)₃(L)₃], [Fe₄(OR)₂(O CMe)₂(L)₄], and [Fe₅O(OH)(O -

2

CR)₄(L)₄]. The tri-, tetra-, and pentanuclear complexes all possess unusual structures and novel core topologies.

Mo¨ssbauer spectroscopy confirms the presence of high-spin ferric centers in the tri- and pentanuclear complexes.

Variable-temperature magnetic measurements suggest spin ground states of S ) 0, 1/2, 0, and 5/2 for the bi-, tri-,

tetra-, and pentanuclear complexes, respectively. Fits of the magnetic susceptibility data have provided the magnitude

of the exclusively antiferromagnetic exchange interactions. In addition, an easy-axis-type magnetic anisotropy has

been observed for the pentanuclear complexes, with D values of approximately −0.4 cm^-1determined from modeling

the low-temperature magnetization data. A low-temperature micro-SQUID study of one of the pentanuclear complexes

reveals magnetization hysteresis at nonzero field. This is attributed to an anisotropy-induced energy barrier to

magnetization reversal that is of molecular origin. Finally, an inelastic neutron scattering study of one of the trinuclear

complexes has revealed that the magnetic behavior arises from two distinct species.

Introduction

ribonucleotide reductase contain similar (µ-oxo or µ-hy-

droxo)(µ-carboxylato)diiron cores, despite the disparate

functions that they perform (e.g., oxygen transport for Hr

and catalytic oxidation for MMO). On a larger scale, the

iron storage protein ferritin has a ferric oxyhydroxide core

containing up to 4500 Fe centers. This core is formed inside

the preassembled protein shell upon oxidation and hydrolysis

of Fe^II. Amino acids featuring carboxylate residues are

believed to play a key role in this process. As a result, a

number of polynuclear {Fe_x} (x ) 11, 12, 17, 19) complexes

featuring oxo, hydroxo, and carboxylato ligation have been

synthesized and studied as possible models for ferritin to

gain a better understanding of the biomineralization process

involved in the formation of its metal core.^2,3

The considerable current interest in polynuclear ferric

complexes arises from their relevance to diverse fields

ranging from bioinorganic chemistry to molecular magnetic

materials. In biological systems oxo-, hydroxo-, and car-

boxylato-bridged Fe centers play an important role in the

active sites of various non-heme Fe proteins.¹Proteins such

as hemerythrin (Hr), methane monoxygenase (MMO), and

* Author to whom correspondence should be addressed. Present ad-

dress: School of Chemistry, University of Melbourne, Parkville, Victoria

3010, Australia. E-mail: c.boskovic@unimelb.edu.au.

^†Universita¨t Bern.

^‡Johannes Gutenberg Universita¨t.

^§Laboratoire L. Neel-CNRS.

^|Universite´ de Neuchaˆtel.

PSI and ETHZ.

Oxo- and hydroxo-bridged diiron units typically display

antiferromagnetic exchange interactions.⁴For polynuclear Fe

(1) Crichton, R. R. Inorganic Biochemistry of Iron Metabolism; Hor-

wood: New York, 1991.

Published on Web 07/10/2004

Inorganic Chemistry, Vol. 43, No. 16, 2004 5053

Boskovic et al.

complexes ({Fe_x}, x > 2), uncompensated spins or spin

frustration associated with competing antiferromagnetic

interactions can lead to molecules with a nonzero magnetic

moment and interesting magnetic behavior. Low-temperature

Mo¨ssbauer spectroscopy in zero field of some of the

complexes proposed as models of ferritin reveals magnetic

splitting associated with slow relaxation of the magnetization

(reorientation of the magnetization direction).^2,5This has been

attributed to superparamagnetic behavior akin to that ob-

served for the ferritin core itself.⁶Such superparamagnetism

occurs for small particles of ordered materials (<100 Å).

These are of the dimension of a single magnetic domain,

where the small particle size renders the magnetic moment

unstable to thermal fluctuations. For such particles, the

energy barrier to magnetization relaxation depends on the

volume and shape of the particle. However, the thermal

scrambling of the magnetic dipole is offset, and slow

magnetic relaxation is observed below a characteristic

“blocking temperature”, which is dependent on the time scale

of the measurement technique.

bistability arising from this energy barrier suggests tremen-

dous potential for applications in information storage devices,

whereby a single molecule could act as the smallest possible

unit of magnetic memory. Furthermore, quantum tunneling

of the magnetization through the energy barrier has been

observed for a number of SMMs, and it has been proposed

that this may ultimately provide the superposition of logic

states necessary for SMMs to act as qubits in quantum

computers.¹⁴

With this in mind we have turned our attention to the

application of the Schiff base proligand salicylidene-2-

ethanolamine (H₂L¹) and its derivatives H₂L²and H₂L³(Chart

1) to the chemistry of iron.

Chart 1

Interestingly, similar slow magnetization relaxation has

been observed in the low-temperature zero-field Mo¨ssbauer

spectra of the ferric compounds [Fe₄(OMe)₆(dpm)₆] and

[Fe₈O₂(OH)₁₂(tacn)₆]Br₈.⁷However, for these species, the

slow relaxation has been attributed to a completely different

process. Both of these complexes possess a relatively high

spin ground state (S ) 5 and 10 for the tetra- and octanuclear

complexes, respectively) in addition to a significant magnetic

anisotropy. This results in an anisotropy-induced energy

barrier to thermal relaxation of the magnetization that is of

purely molecular origin.^8-10These compounds have been

described as “single-molecule magnets” (SMMs).¹¹In gen-

eral, for SMMs with a large-spin ground state (S) and a large

easy-axis-type magnetic anisotropy (D < 0, where D is the

axial zero-field splitting parameter), the energy barrier to

magnetization relaxation is given to first order by S²|D| or

(S²- 1/4)|D|, for integer and half-integer spins, respectively.

Other {Fe^III_x} (x ) 11, 17, 19) complexes have also been

identified as SMMs.^12,13For such materials the magnetic

Preliminary results of this study have been communicated

previously.^15,16Deprotonation of the proligands affords the

potentially tridentate mono- and bianionic ligands. Each

incorporates an ONO donor set and possesses the ability to

bind in both bridging and chelating modes, enhancing the

stability of the metal complexes formed. The ligands favor

binding in a meridional fashion, as has been observed in a

number of mono- and binuclear transition-metal species.¹⁷

In addition, we have recently reported that a family of

tetranuclear Mn complexes bearing (L¹)^2-and its derivatives

function as SMMs,¹⁸while a tetranuclear Fe^IIcomplex of

(L¹)^2-has also been reported to display SMM behavior.¹⁹

Finally, we have communicated the synthesis and preliminary

magnetic characterization of a tetranuclear Ni^IIcomplex of

(L¹)^2-,¹⁵while tetranuclear Cu and Mn complexes and

hexanuclear Mn complexes of the same ligand have been

reported previously.²⁰

(2) (a) Gorun, S. M.; Papaefthymiou, G. C.; Frankel, R. B.; Lippard, S.

J. J. Am. Chem. Soc. 1987, 109, 3337. (b) Taft, K. L.; Papaefthymiou,

G. C.; Lippard, S. J. Science 1993, 259, 1302. (c) Taft, K. L.;

Papaefthymiou, G. C.; Lippard, S. J. Inorg. Chem. 1994, 33, 1510.

(d) Caneschi, A.; Cornia, A.; Lippard, S. J.; Papaefthymiou, G. C.;

Sessoli, R. Inorg. Chim. Acta 1996, 243, 295.

(3) Powell, A. K.; Heath, S. L.; Gatteschi, D.; Pardi, L.; Sessoli, R.; Spina,

G.; Del Giallo, F.; Pieralli, F. J. Am. Chem. Soc. 1995, 117, 2491.

(4) Kurtz, D. M. Chem. ReV. 1990, 90, 585.

(5) Papaefthymiou, G. C. Phys. ReV. B 1992, 46, 366.

(6) St. Pierre, T. G.; Dickson, D. P. E.; Webb, J.; Kim, K. S.; Macey, D.

J.; Mann, S. Hyperfine Interact. 1986, 29, 1427.

(12) Jones, L. F.; Brechin, E. K.; Collison, D.; Helliwell, M.; Mallah, T.;

Piligkos, S.; Rajaraman, G.; Wernsdorfer, W. Inorg. Chem. 2003, 42,

6601.

(13) Goodwin, J. C.; Sessoli, R.; Gatteschi, D.; Wernsdorfer, W.; Powell,

A. K.; Heath, S. L. J. Chem. Soc., Dalton Trans. 2000, 1835.

(14) Leuenberger, M. N.; Loss, D. Nature 2000, 410, 789.

(15) (a) Boskovic, C.; Rusanov, E.; Stoeckli-Evans, H.; Gu¨del, H. U. Inorg.

Chem. Commun. 2002, 5, 881. (b) Basler, R.; Boskovic, C.; Chabous-

sant, G.; Gu¨del, H. U.; Murrie, M.; Ochsenbein, S. T.; Sieber, A.

ChemPhysChem 2003, 4, 910.

(7) (a) Barra, A.-L.; Debrunner, P.; Gatteschi, D.; Schulz, C. E.; Sessoli,

R. Europhys. Lett. 1996, 35, 133. (b) Caneschi, A.; Cianchi, L.; Del

Giallo, F.; Gatteschi, D.; Moretti, P.; Pieralli, F.; Spina, G. J. Phys.:

Condens. Matter 1999, 11, 3395. (c) Cianchi, L.; Del Giallo, F.; Spina,

G.; Reiff, W.; Caneschi, A. Phys. ReV. B 2002, 65, 064415.

(8) Barra, A. L.; Caneschi, A.; Cornia, A.; De Biani, F. F.; Gatteschi, D.;

Sangregorio, C.; Sessoli, R.; Sorace, L. J. Am. Chem. Soc. 1999, 121,

5302.

(9) Sangregorio, C.; Ohm, T.; Paulsen, C.; Sessoli, R.; Gatteschi, D. Phys.

ReV. Lett. 1997, 78, 4645.

(10) Gatteschi, D.; Sessoli, R.; Cornia, A. Chem. Commun. 2000, 725.

(11) Gatteschi, D.; Sessoli, R. Angew. Chem., Int. Ed. 2003, 42, 268 and

references therein.

(16) Boskovic, C.; Labat, G.; Neels, A.; Gu¨del, H. U. Dalton Trans. 2003,

3671.

(17) (a) Asada, H.; Ozeki, M.; Fujiwara, M.; Matsushita, T. Chem. Lett.

1999, 525. (b) Dey, M.; Rao, C. P.; Saarenketo, P. K.; Rissanen, K.

Inorg. Chem. Commun. 2002, 5, 924.

(18) Boskovic, C.; Bircher, R.; Tregenna-Piggott, P. L. W.; Gu¨del, H. U.;

Paulsen, C.; Wernsdorfer, W.; Barra, A. L.; Khatsko, E.; Neels, A.;

Stoeckli-Evans, H. J. Am. Chem. Soc. 2003, 125, 14046.

(19) Oshio, H.; Hoshino, N.; Ito, T. J. Am. Chem. Soc. 2000, 120, 2997.

(20) (a) Chumakov, Y. M.; Biyushkin, V. N.; Malinovskii, T. I.; Kulemu,

S.; Tsapkov, V. I.; Popov, M. S.; Samus, N. M. Koord. Khim. 1990,

16, 945. (b) Hoshino, N.; Ito, T.; Nihei, M.; Oshio, H. Inorg. Chem.

Commun. 2003, 6, 377.

5054 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

isolated by filtration and washed with hexane; yield 70%. A sample

for crystallography was maintained in contact with the mother liquor

to prevent the loss of interstitial solvent. Drying under vacuum at

100 °C afforded a fully desolvated sample. Anal. Calcd for [Fe₃(O₂-

CMe)₃(L³)₃], C₅₁H₄₈N₃Fe₃O₁₂: C, 57.65; H, 4.55; N, 3.95. Found:

C, 57.61; H, 4.64; N, 3.73. The solvation level was confirmed by

thermogravimetric analysis. Selected IR data (cm^-1): 1599 (m),

1565 (s), 1539 (s), 1491 (w), 1463 (m), 1439 (s), 1420 (m, sh),

1366 (w), 1359 (w), 1331 (s), 1261 (m), 1247 (m), 1144 (m), 1121

(w), 1061 (m), 1052 (m), 1042 (m), 1024 (w), 963 (w), 908 (m),

849 (m), 777 (w), 758 (m), 749 (m), 736 (w), 705 (m), 653 (w),

638 (w), 614 (m), 595 (m), 524 (w), 509 (w), 457 (w), 441 (w).

[Fe₄(OMe)₂(O₂CMe)₂(L³)₄] (5). Method 1. A concentrated

solution of crude 4 in MeOH was layered with Et₂O to produce

red-brown crystals together with an amorphous pale-colored

byproduct. These were isolated by filtration and washed with EtOH;

however, removal of the byproduct proved impossible. A sample

for crystallography was maintained in contact with the mother liquor

to prevent the loss of interstitial solvent. Selected IR data (cm^-1):

1598 (s), 1586 (s), 1564 (s), 1534 (s), 1491 (w), 1464 (m), 1441

(s), 1415 (m, sh) 1363 (w), 1340 (s), 1328 (m), 1260 (m), 1248

(m), 1238 (m), 1145 (m), 1120 (w), 1091 (w), 1059 (m), 1040 (m),

1026 (m), 961 (w), 905 (m), 848 (m), 775 (w), 747 (m), 725 (w),

701 (m), 693 (m), 653 (w), 636 (w), 611 (m), 587 (m), 556 (m),

533 (w), 508 (m), 492 (m), 453 (m).

Method 2. Fe(OAc)₂(0.43 g, 2.5 mmol) was added to a solution

of H₂L³(0.60 g, 2.5 mmol) in MeOH (40 cm³), and the resulting

mixture was stirred overnight, affording a red-brown precipitate.

This crude product was isolated by filtration and washed with

MeOH; yield 90%. IR spectroscopy indicated that this material was

identical with that obtained by method 1.

[Fe₄(OEt)₂(O₂CMe)₂(L³)₄] (6). Method 1. A concentrated

solution of crude 4 in EtOH was allowed to stand in a sealed vessel

to produce red-brown crystals. These were isolated by filtration

and washed with EtOH; yield 25%. The sample was dried overnight

under vacuum at 110 °C, but appears to be hygroscopic. Anal. Calcd

for [Fe₄(OEt)₂(O₂CMe)₂(L³)₄]‚H₂O, C₆₈H₇₀N₄Fe₄O₁₅: C, 58.06; H,

5.02; N, 3.98. Found: C, 57.94; H, 4.82; N, 3.77. The solvation

level was confirmed by thermogravimetric analysis. Selected IR

data (cm^-1): 1597 (s), 1586 (sh), 1568 (s), 1536 (s), 1491 (w),

1463 (m), 1441 (s), 1415 (m, sh) 1363 (w), 1340 (s), 1328 (m),

1260 (m), 1247 (m), 1146 (m), 1121 (w), 1082 (w), 1054 (m), 1040

(m), 1027 (m), 963 (w), 905 (m), 849 (m), 776 (w), 749 (m), 726

(w), 702 (m), 654 (w), 636 (w), 612 (m), 587 (m), 559 (m), 533

(w), 508 (w), 497 (m), 445 (m).

Experimental Section

Syntheses. All manipulations were performed under aerobic

conditions, using materials as received. H₂L¹was prepared as

described;^15aH₂L²and H₂L³were synthesized in an analogous

manner. picH is 2-picolinic acid.

[Fe₂(pic)₂(L¹)₂] (1). picH (0.15 g, 1.2 mmol) was added to a

solution of compound 2 (0.50 g, 0.60 mmol) in MeCN, and the

resulting mixture was stirred for several hours, affording a red

precipitate. This was recrystallized by layering a concentrated CH₂-

Cl₂solution with hexane to give dark red blocklike crystals. These

were isolated by filtration and washed with hexane; yield 90%. A

sample for crystallography was maintained in contact with the

mother liquor to prevent the loss of interstitial solvent. Drying under

vacuum at room temperature afforded a partially solvated sample.

Anal. Calcd for [Fe₂(pic)₂(L¹)₂]‚CH₂Cl₂, C₃₁H₂₈N₄Fe₂O₈Cl₂: C,

48.53; H, 3.68; N, 7.30. Found: C, 48.70; H, 3.72; N, 7.11. The

solvation levels were confirmed by thermogravimetric analysis.

Selected IR data (cm^-1): 1668 (s), 1654 (m), 1629 (s), 1600 (s),

1571 (w), 1546 (m), 1471 (m), 1448 (m), 1410 (m), 1347 (s), 1339

(s), 1305 (m), 1289 (s), 1257 (w), 1234 (w), 1199 (m), 1168 (m),

1148 (m), 1123 (w), 1093 (w), 1063 (w), 1046 (s), 1033 (m), 1023

(m), 936 (w), 899 (m), 850 (m), 796 (w), 761 (m), 707 (m), 694

(m), 647 (w), 618 (m), 519 (m), 481 (w), 467 (w), 440 (w), 420

(m).

[Fe₃(O₂CMe)₃(L¹)₃] (2). Fe(OAc)₂(0.43 g, 2.5 mmol) was added

to a solution of H₂L¹(0.41 g, 2.5 mmol) in EtOH (40 cm³), and

the resulting mixture was stirred overnight, affording a red-brown

precipitate. This was recrystallized by layering a concentrated CH₂-

Cl₂solution with hexane, to give dark red rodlike crystals. These

were isolated by filtration and washed with hexane; yield 75%. A

sample for crystallography was maintained in contact with the

mother liquor to prevent the loss of interstitial solvent. Drying under

vacuum at room temperature afforded a partially solvated sample.

Anal. Calcd for [Fe₃(O₂CMe)₃(L¹)₃]‚0.5CH₂Cl₂, C_33.5H₃₇N₃Fe₃O₁₂-

Cl₁: C, 45.90; H, 4.25; N, 4.79. Found: C, 46.11; H, 4.25; N,

4.41. Drying under vacuum at 80 °C afforded a fully desolvated

sample. Anal. Calcd for [Fe₃(O₂CMe)₃(L¹)₃], C₃₃H₃₆N₃Fe₃O₁₂: C,

47.51; H, 4.35; N, 5.04. Found: C, 47.52; H, 4.40; N, 5.07. The

solvation levels were confirmed by thermogravimetric analysis.

Selected IR data (cm^-1): 1639 (s), 1601 (m), 1560 (s), 1469 (m),

1449 (s), 1430 (s, sh), 1389 (m), 1337 (m), 1310 (m), 1219 (w),

1198 (w), 1148 (m), 1126 (w), 1066 (w), 1043 (m), 1030 (m), 931

(w), 901 (w), 878 (w), 857 (w), 794 (w), 759 (m), 739 (w), 660

(m), 643 (w), 613 (m), 560 (m), 545 (m), 412 (m).

[Fe₃(O₂CMe)₃(L²)₃] (3). Fe(OAc)₂(0.43 g, 2.5 mmol) was added

to a solution of H₂L²(0.45 g, 2.5 mmol) in EtOH (40 cm³), and

the resulting mixture was stirred overnight, affording a red-brown

precipitate. This crude product was isolated by filtration and washed

with EtOH; yield 85%. Drying under vacuum at room temperature

afforded a partially hydrated sample. Anal. Calcd for [Fe₃(O₂CMe)₃-

(L²)₃]‚2H₂O, C₃₆H₄₆N₃Fe₃O₁₄: C, 47.40; H, 5.08; N, 4.61. Found:

C, 47.22; H, 5.16; N, 4.12. The solvation level was confirmed by

thermogravimetric analysis. Selected IR data (cm^-1): 1601 (s), 1560

(s), 1543 (s), 1473 (w), 1440 (s), 1370 (w), 1327 (m), 1260 (w),

1237 (m), 1161 (w), 1136 (w), 1080 (w), 1067 (m), 1057 (m), 1021

(w), 984 (w), 928 (w), 879 (w), 848 (w), 752 (m), 659 (m), 615

(w), 604 (m), 592 (m), 561 (m), 524 (w), 490 (w), 419 (m).

Method 2. Fe(OAc)₂(0.43 g, 2.5 mmol) was added to a solution

of H₂L³(0.60 g, 2.5 mmol) in EtOH (40 cm³), and the resulting

mixture was stirred overnight, affording a red-brown precipitate.

This crude product was isolated by filtration and washed with EtOH;

yield 90%. IR spectroscopy indicated that this material was identical

with that obtained by method 1.

[Fe₅O(OH)(O₂CC₆H₄-p-NO₂)₄(L¹)₄] (7). A slurry of 2 (0.50 g,

0.60 mmol) in MeCN/toluene (1:1, 50 cm³) was treated with

p-NO₂C₆H₄CO₂H (0.30 g, 1.8 mmol) and evaporated to dryness.

The resulting residue was redissolved in toluene and evaporated to

dryness two more times before being recrystallized by layering a

concentrated solution in toluene with hexane, to give dark brown

crystals. These were isolated by filtration and washed with toluene;

yield 25%. A sample for crystallography was maintained in contact

with the mother liquor to prevent the loss of interstitial solvent.

Drying under vacuum at 45 °C afforded a fully desolvated sample.

Anal. Calcd for [Fe₅O(OH)(O₂CC₆H₄-p-NO₂)₄(L¹)₄], C₆₄H₅₃N₈-

Fe₅O₂₆: C, 47.18; H, 3.28; N, 6.88. Found: C, 47.13; H, 3.28; N,

[Fe₃(O₂CMe)₃(L³)₃] (4). Fe(OAc)₂(0.43 g, 2.5 mmol) was added

to a solution of H₂L³(0.60 g, 2.5 mmol) in MeCN (60 cm³), and

the resulting mixture was stirred overnight, affording a red-brown

precipitate. This was recrystallized by layering a concentrated CH₂-

Cl₂solution with hexane to give dark red crystals. These were

Inorganic Chemistry, Vol. 43, No. 16, 2004 5055

Boskovic et al.

Table 1. Crystallographic Data for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, and 5‚0.5MeOH

1‚CH₂Cl₂

2‚CH₂Cl₂

4‚CH₂Cl₂

5‚0.5MeOH

empirical formula

fw

space group

a, Å

b, Å

C₃₁H₂₈Cl₂Fe₂N₄O₈

767.17

C2/c

13.4247(12)

16.7179(16)

14.9140(11)

90

100.225(10)

90

C₃₄H₃₈Cl₂Fe₃N₃O₁₂

919.12

Pca2₁

17.9897(8)

13.8773(6)

31.0893(19)

90

7761.4(7)

8

C₅₂H₅₀Cl₂Fe₃N₃O₁₂

1147.40

P2₁/c

9.7371(11)

40.685(4)

13.1279(15)

90

106.802(9)

90

4978.6(9)

4

C

_66.5H₆₆Fe₄N₄O₁₄₅

1376.63

P2/c

17.3156(14)

8.0980(6)

25.702(3)

90

118.058(7)

90

3180.4(5)

2

c, Å

R, deg

â, deg

γ, deg

V, Å³

Z

3294.0(5)

4

T, K

153(2)

0.71073

1.547

1.099

2006

153(2)

0.71073

1.573

1.308

9064

153(2)

0.71073

1.531

1.037

3659

153(2)

0.71073

1.438

0.963

3850

λ, Å

F

_calcd, g cm^-3

µ, mm^-1

no. of obsd data [I > 2σ(I)]

R1^a

0.0316

0.0455

0.1026

0.0398

wR2

0.0571^b

0.1065^c

0.2083^d

0.0928^e

^aR1 ) ∑|F_o| - |F_c|/∑|F_o. ^bwR2 ) [∑w(F_o

F_c)²/∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.0257P)²], where P ) (F_o+ 2F_c)/3. ^cwR2 ) [∑w(F_o

F_c)²/

2

-

2

4

2

-

2

∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.1P)²], where P ) (F_o+ 2F_c)/3. ^dwR2 ) [∑w(F_o

F_c)²/∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.0974P)²], where P ) (F_o

+

4

2

-

2

4

2

-

2

4

2

2F_c)/3. ^ewR2 ) [∑w(F_o

F_c)²/∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.0552P)²], where P ) (F_o+ 2F_c)/3.

6.65. The lack of solvation was confirmed by thermogravimetric

analysis. Selected IR data (cm^-1): 1642 (s), 1616 (m), 1601 (m),

1563 (s), 1545 (s), 1524 (m), 1470 (m), 1450 (s), 1345 (s), 1317

(m), 1270 (w), 1230 (w), 1219 (w), 1200 (w), 1150 (m), 1128 (w),

1104 (w), 1076 (w), 1045 (m), 1015 8w), 978 (w), 931 (w), 902

(w), 877 (w), 863 (w), 828 (m), 795 (m), 759 (m), 725 (m),

642(w), 623 (m), 613 (m), 558 (m), 533 (m).

[Fe₅O(OH)(O₂CMe)₄(L²)₄] (8). A concentrated solution of crude

3 in CH₂Cl₂was layered with hexane to afford dark brown crystals.

These were isolated by filtration and washed with hexane; yield

80%. A sample for crystallography was maintained in contact with

the mother liquor to prevent the loss of interstitial solvent. Drying

under vacuum at room temperature and exposure to air afforded a

hydrated sample. Anal. Calcd for [Fe₅O(OH)(O₂CMe)₄(L²)₄]‚H₂O,

C₄₈H₅₉N₄Fe₅O₁₉: C, 45.21; H, 4.66; N, 4.39. Found: C, 45.16; H,

4.86; N, 4.34. The hydration level was confirmed by thermogravi-

metric analysis. Selected IR data (cm^-1): 1600 (s), 1560 (s), 1542

(s), 1473 (m), 1442 (s), 1420 (s), 1367 (m), 1324 (s), 1293 (m),

1259 (w), 1235 (m), 1163 (w), 1138 (w), 1100 (w), 1069 (m), 1019

(m), 983 (w), 922 (w), 880 (w), 846 (m), 757 (m), 670 (m), 656

(m), 601 (m), 563 (m), 523 (m), 494 (m), 421 (w).

[Fe₅O(OH)(O₂CPh)₄(L³)₄] (9). A solution of 4 (0.25 g, 0.24

mmol) in CH₂Cl₂/toluene (1:1, 50 cm³) was treated with PhCO₂H

(0.18 g, 1.5 mmol) and evaporated to dryness. The resulting residue

was redissolved in toluene and evaporated to dryness two more

times before being recrystallized by layering a concentrated solution

in toluene with hexane, to give dark brown crystals. These were

isolated by filtration and washed with toluene; yield 40%. A sample

for crystallography was maintained in contact with the mother liquor

to prevent the loss of interstitial solvent. Drying under vacuum at

room temperature followed by exposure to air afforded a hydrated

sample. Anal. Calcd for [Fe₅O(OH)(O₂CPh)₄(L³)₄] ]‚2H₂O, C₈₈H₇₇N₄-

Fe₅O₂₀: C, 59.05; H, 4.34; N, 3.13. Found: C, 59.11; H, 4.31; N,

2.79. The hydration level was confirmed by thermogravimetric

analysis. Selected IR data (cm^-1): 1619 (m), 1593 (s), 1540 (s),

1523 (s), 1492 (m), 1464 (m), 1441 (s), 1397 (s), 1326 (s), 1293

(w), 1242 (m), 1175 (w), 1146 (m), 1121 (w), 1093 (w), 1070 (m),

1052 (m), 1042 (m), 1026 (m), 956 (w), 903 (w) 848 (m), 777

(w), 753 (m), 725 (m), 718 (m), 702 (m), 690 (m), 676 (m), 639

(w), 621 (m), 598 (m), 556 (w), 539 (w), 504 (w), 464 (m), 438 (m).

X-ray Crystallography. The intensity data of compounds

1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚0.5MeOH, 7‚H₂O‚C₇H₈‚0.5C₆H₁₄,

8‚0.25H₂O‚4.5CH₂Cl₂, and 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O were col-

lected at 153 K on Stoe image plate diffraction systems I and II²¹

using Mo KR graphite-monochromated radiation. Compound 1‚

CH₂Cl₂employed image plate I, distance 70 mm, φ oscillation scans

0-199.5°, step ∆φ ) 1.5°, d_min- d_max) 12.45 - 0.81 Å.

Compound 2‚CH₂Cl₂employed image plate I, distance 70 mm, φ

oscillation scans 0-100°, step ∆φ ) 0.8°, d_min- d_max) 12.45 -

0.81 Å. Compound 4‚CH₂Cl₂employed image plate II, distance

140 mm, 120 frames with φ ) 0° and 0° < ω < 180° and 120

frames with φ ) 90° and 0° < ω < 180°, with the crystal oscillating

through 1.5° in ω. The resolution was d_min- d_max) 17.78 - 0.72

Å. Compound 5‚0.5MeOH employed image plate I, distance 70

mm, φ oscillation scans 0-180°, step ∆φ ) 1°, d_min- d_max

)

12.45 - 0.81 Å. Compound 7‚H₂O‚C₇H₈‚0.5C₆H₁₄employed image

plate II, distance 140 mm, 180 frames with φ ) 0° and 0° < ω <

180° and 44 frames with φ ) 90° and 0° < ω < 44°, with the

crystal oscillating through 1° in ω. The resolution was d_min- d_max

) 24.88 - 0.83 Å. Compound 8‚0.25H₂O‚4.5CH₂Cl₂and com-

pound 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O employed image plate I,

distance 70 mm, φ oscillation scans 0-180°, step ∆φ ) 1°, d_min

-

d_max) 12.45 - 0.81 Å. The structures of all compounds were

solved by direct methods using the program SHELXS-97²²and

refined using weighted full-matrix least-squares on F². The refine-

ment and all further calculations were carried out using SHELXL-

97.²³Crystallographic data for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚

0.5MeOH, 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and 9‚

1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O are given in Tables 1 and 2.

For compound 1‚CH₂Cl₂the H atoms were located from Fourier

difference maps and refined isotropically. For compounds 2‚CH₂-

Cl₂, 4‚CH₂Cl₂, and 5‚0.5MeOH the H atoms were included in

calculated positions and treated as riding atoms using SHELXL

default parameters. For compound 7‚H₂O‚C₇H₈‚0.5C₆H₁₄the posi-

tion of the hydroxo H atom was located from Fourier difference

maps and freely refined, while the remaining H atoms were included

in calculated positions and treated as riding atoms using SHELXL

default parameters. For compounds 8‚0.25H₂O‚4.5CH₂Cl₂and 9‚

(21) Stoe & Cie. IPDS Software; Stoe & Cie GmbH: Darmstadt, Germany,

2000.

(22) Sheldrick, G. M. SHELXS-97 Program for Crystal Structure Deter-

mination. Acta Crystallogr. 1990, A46, 467.

(23) Sheldrick, G. M. SHELXL-97, Program for Crystal Structure Refine-

ment; Universita¨t Go¨ttingen: Go¨ttingen, Germany, 1999.

5056 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

Table 2. Crystallographic Data for 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

7‚H₂O‚C₇H₈‚0.5C₆H₁₄

8‚0.25H₂O‚4.5CH₂Cl₂

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

empirical formula

fw

space group

a, Å

b, Å

C₇₄H₇₀Fe₅N₈O₂₇

1782.63

P1h

C_52.5H_66.5Cl₉Fe₅N₄O_18.25

1643.90

P1h

C_90.5H₈₂Cl₄Fe₅N₄O₂₀

1966.65

P1h

15.3288(9)

15.9751(10)

20.0357(13)

104.081(5)

99.318(5)

116.649(4)

4040.1(4)

2

14.6913(12)

23.5238(18)

23.4482(18)

62.310(8)

88.319(10)

72.627(9

6790.5(9)

4

14.4742(13)

15.4790(15)

21.1155(18)

87.069(11)

79.397(10)

69.702(10)

4360.9(7)

2

c, Å

R, deg

â, deg

γ, deg

V, Å³

Z

T, K

153(2)

λ, Å

0.71073

1.465

0.959

0.71073

1.608

1.465

0.71073

1.498

1.008

F

_calcd, g cm^-3

µ, mm^-1

no. of obsd data [I > 2σ(I)]

9976

13533

10027

R1^a

0.0588

0.0834

0.0584

wR2

0.1597^b

0.2130^c

0.1676^d

^aR1 ) ∑|F_o| - |F_c|/∑|F_o|. ^bwR2 ) [∑w(F_o

F_c)²/∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.1174P)²], where P ) (F_o+ 2F_c)/3. ^cwR2 ) [∑w(F_o

F_c)²/

F_c)²/∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.1021P)²], where P )

2

-

2

4

2

-

2

∑wF_o]^1/2; w ) 1/[σ²(F_o) + (0.1372P)²], where P ) (F_o+ 2F_c)/3. ^dwR2 ) [∑w(F_o

4

2

-

2

4

2

(F_o+ 2F_c)/3.

1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O the positions of the hydroxo H atoms

were derived from Fourier difference maps and refined with the

O-H distance constrained to the theoretical value, while the

remaining H atoms were included in calculated positions and treated

as riding atoms using SHELXL default parameters. The non-H

atoms were refined anisotropically for all compounds except for

one MeOH molecule (25% occupied) of 5‚0.5MeOH, C104 (50%

occupied) of 8‚0.25H₂O‚4.5CH₂Cl₂, and partially occupied MeOH

and H₂O atoms of 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O, which were

refined isotropically.

For compound 2‚CH₂Cl₂the structure was refined as a racemic

twin with a final refined BASF (batch scale factor) value of 0.40-

(2). A multiscan absorption correction was applied for compounds

4‚CH₂Cl₂and 5‚0.5MeOH using the MULABS routine in PLA-

TON,²⁴and an empirical absorption correction was applied for

compounds 8‚0.25H₂O‚4.5CH₂Cl₂and 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

using the DIFABS routine in PLATON.

Magnetic Measurements. Variable-temperature magnetic sus-

ceptibility measurements down to 1.8 K were performed with a

Quantum Design MPMS-XL susceptometer equipped with a 5 T

magnet. Data were collected on powdered crystals. Pascal’s

constants were used to estimate the diamagnetic correction for each

complex. Low-temperature magnetic measurements were performed

on single crystals using an array of micro-SQUIDS.²⁵Measurements

were performed on this magnetometer in the temperature range

0.04-7.0 K, with fields up to 1.4 T. The field can be applied in

any direction by separately driving three orthogonal coils. The

experimental susceptibility data were fit using the Levenberg-

Marquardt least-squares fitting algorithm, in combination with

MAGPACK, which employs matrix diagonalization methods and

the isotropic Heisenberg-Dirac-Van Vleck term as the leading

part of the exchange spin Hamiltonian.²⁶ø_Mvs T and ø_MT vs T

data were fit separately and the results combined to give the final

fitting parameters.

sample was sealed under helium in an aluminum container of size

40 × 40 × 3 mm³. The data treatment involved a calibration of

the detectors using a spectrum of vanadium metal and a background

subtraction using an empty can. The time-of-flight to energy

conversion and the data reduction were performed using the

standard program Nathan. Further data treatment employed the

program Igor-Pro 4.0.2.1 (Wave Metrics).

Mo1ssbauer Spectroscopy. A constant-acceleration-type Mo¨ss-

bauer spectrometer equipped with a 1024-channel analyzer operating

in the time scale mode, and a 25 mCi ⁵⁷Co/Rh source, was

employed. The isomer shifts reported here are relative to that of

R-Fe at room temperature. Spectra of the sample (thickness of about

5 mg of Fe/cm²) were collected between 293 and 4.2 K by means

of a combined He continuous flow/bath cryostat. The Mo¨ssbauer

spectra were analyzed using the computer program EFFINO.²⁷

Other Measurements. Infrared spectra (KBr disk) were recorded

on a Perkin-Elmer Spectrum One FTIR spectrometer. Elemental

analyses were performed at the Ecole d’inge´nieurs et d’architectes

de Fribourg, Switzerland.

Results

Syntheses. The transformations involving complexes 1-9

are summarized in Scheme 1.

Overnight treatment of a slurry of Fe(OAc)₂in EtOH with

1 equiv of H₂L¹or H₂L²affords the trinuclear complexes 2

and 3, respectively, while a similar reaction with H₂L³in

MeCN affords trinuclear 4. In each case the yields are high.

The reaction may be written as

3Fe(O₂CMe)₂+ 3H₂L f

[Fe₃(O₂CMe)₃L₃] + 3MeCO₂H + 3H⁺+ 3e

(2, L ) L¹; 3, L ) L²; 4, L ) L³) (1)

Inelastic Neutron Scattering (INS) Studies. INS spectra were

recorded on the time-of-flight spectrometer FOCUS at the SINQ

facility (PSI, Switzerland) with wavelengths of λ_i) 3.8 Å and λ_i

) 4.3 Å. Approximately 3 g of an undeuterated, polycrystalline

where Fe^IIis oxidized to Fe^IIIby oxygen from the air. It is

possible to obtain 2 and 4 in crystalline form by recrystal-

lization of the crude product from CH₂Cl₂/hexane, and these

complexes were structurally characterized by X-ray diffrac-

tion. Although the IR and elemental analysis data confirm

(24) Spek, A. L. Acta Crystallogr. 1990, A46, C34.

(25) Wernsdorfer, W. AdV. Chem. Phys. 2001, 118, 99.

(26) Borra´s-Almenar, J. J.; Clemente-Juan, J. M.; Coronado, E.; Tsukerblatt,

B. S. MAGPACK. Inorg. Chem. 1999, 38, 6081.

(27) Spiering, H.; Dea´k, L.; Bottya´n, L. Hyperfine Interact. 2000, 125, 197.

Inorganic Chemistry, Vol. 43, No. 16, 2004 5057

Boskovic et al.

Scheme 1

crystals of 5 proved impossible. In addition, crude and

crystalline samples of 5 and 6 are insoluble in common

organic solvents, which prevented attempts to obtain a pure

sample of 5 by further recrystallization. Thus, a pure bulk

sample of 5 was not available for elemental analysis or

SQUID measurements. In contrast, although crystals of 6

large enough for X-ray characterization could not be

obtained, microcrystalline samples are analytically pure and

the IR spectrum is essentially identical to that of crystals of

5, confirming that 6 is the ethoxo analogue of 5.

Carboxylate exchange was investigated for 2 and 4 in an

effort to produce analogous trinuclear complexes with

carboxylates other than acetate. A variety of carboxylates

were explored; however, it proved difficult to obtain pure

species. Nevertheless, reaction of a slurry of 2 in MeCN/

toluene with 3 equiv of p-NO₂C₆H₄CO₂H, followed by

repeated evaporation to dryness and redissolution in toluene

to remove the MeCO₂H via the toluene/MeCO₂H azeotrope,

affords pentanuclear 7 in reasonable yield. A similar treat-

ment of 4 with 6 equiv of PhCO₂H affords pentanuclear 9:

5[Fe₃(O₂CMe)₃(L)₃] + 12RCO₂H + 6H₂O f

3[Fe₅O(OH)(O₂CR)₄(L)₄] + 15MeCO₂H + 3H₂L

(7, L ) L¹, R ) p-NO₂C₆H₄; 9, L ) L³, R ) Ph) (5)

Again the new oxo and hydroxo ligands in the pentanuclear

species probably result from residual H₂O in the solvent.²⁸

Finally, the reaction of 2 with 3 equiv of picH affords the

binuclear species 1 in high yield:

that crude 3 is a similar trinuclear complex, recrystallization

of 3 from CH₂Cl₂/hexane affords the pentanuclear complex

8 in high yield:

5[Fe₃(O₂CMe)₃(L²)₃] + 6H₂O f

2[Fe₃(O₂CMe)₃(L¹)₃] + 6picH f

3[Fe₂(pic)₂(L¹)₂] + 6MeCO₂H (6)

3[Fe₅O(OH)(O₂CMe)₄(L²)₄] + 3MeCO₂H + 3H₂L²(2)

The additional oxo and hydroxo ligands in the pentanuclear

species apparently result from residual H₂O in the solvent.²⁸

A similar treatment of a slurry of Fe(OAc)₂in MeOH or

EtOH with 1 equiv of H₂L³gives the tetranuclear species 5

and 6, respectively, according to

In this case, although picH possesses a carboxylic acid

functionality, pic^-binds in a chelating manner through N

and O donor atoms, rather than in a bridging mode through

the carboxylate moiety. Presumably, this is due to the greater

stability provided to the resulting complex by chelation. It

seems likely that reaction of 2 with other simple chelating

ligands will result in similar binuclear complexes, although

this was not pursued further.

4Fe(O₂CMe)₂+ 4H₂L³+ 2ROH f

[Fe₄(OR)₂(O₂CMe)₂(L³)₄] + 6MeCO₂H + 4H⁺+ 4e

Structure Description. Labeled structural diagrams of the

complexes of interest are presented in Figures 1 (1), 2 (2

and 4), 3 (5), and 4 (7-9). The most relevant interatomic

distances and angles are available in Tables 3-5, while more

complete listings are available in the Supporting Information

(Tables S1-S7).

Complex 1 crystallizes in the monoclinic space group

C2/c. The asymmetric unit contains half of the molecule of

interest together with a disordered CH₂Cl₂molecule that is

50% occupied. The complex contains a {Fe^III₂(µ₂-O)₂}²⁺

core, where the Fe centers are bridged by the ethoxo-type O

atoms from the (L¹)^2-ligands (Figure 1). Crystallographic

C₂point symmetry is present, with the C₂axis passing

through the center of the {Fe^III₂O₂}²⁺core, coinciding with

the crystallographic b axis. The (L¹)^2-ligands coordinate in

a meridional bischelating fashion, and the remainder of the

peripheral ligation is provided by chelating pic^-ligands. The

(5, R ) Me; 6, R ) Et) (3)

Alternatively, recrystallization of 4 from MeOH/Et₂O affords

crystals of 5 that are suitable for X-ray diffraction, together

with a pale amorphous precipitate. Similarly, upon standing,

a saturated solution of 4 in EtOH produces tiny single crystals

of 6:

4[Fe₃(O₂CMe)₃(L³)₃] + 6ROH f

3[Fe₄(OR)₂(O₂CMe)₂(L³)₄] + 6MeCO₂H

(5, R ) Me; 6, R ) Et) (4)

However, removal of the precipitate that forms alongside

(28) Aromi, G.; Aubin, S. M. J.; Bolcar, M. A.; Christou, G.; Eppley, H.

J.; Folting, K.; Hendrickson, D. N.; Huffman, J. C.; Squire, R. C.;

Tsai, H.-L.; Wang, S.; Wemple, M. W. Polyhedron 1998, 17, 3005

and references therein.

5058 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

Table 3. Fe-O Distances (Å) for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚0.5MeOH, 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O^a

compd

Fe-O (alkoxo or phenoxo)

Fe-O (hydroxo)

Fe-O (oxo)

1‚CH₂Cl₂

2‚CH₂Cl₂

4‚CH₂Cl₂

1.999(2)-2.028(2)

1.943(5)-2.057(6)

1.918(8)-2.026(7)

1.965(2)-2.037(2)

1.948(3)-2.096(4)

1.956(5)-2.130(5)

1.981(4)-2.111(4)

5‚0.5MeOH

7‚H₂O‚C₇H₈‚0.5C₆H₁₄

8‚0.25H₂O‚4.5CH₂Cl₂

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

2.023(3)-2.098(3)

1.987(5)-2.072(4)

2.020(3)-2.103(3)

1.879(3)-1.930(3)

1.889(5)-1.937(5)

1.881(4)-1.925(4)

^aDistances to bridging O atoms only.

Table 4. Fe‚‚‚Fe Distances (Å) for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚0.5MeOH, 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

a

compd

via (µ-O)₂

3.124(2)

via µ₂-alkoxo

via µ₃-oxo

1‚CH₂Cl₂

2‚CH₂Cl₂

4‚CH₂Cl₂

3.409(2)-3.519(2)

3.409(2)-3.534(2)

3.544(2)

5‚0.5MeOH

3.111(2)

7‚H₂O‚C₇H₈‚0.5C₆H₁₄

8‚0.25H₂O‚4.5CH₂Cl₂

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

2.989(2)-3.205(2)

3.002(2)-3.218(2)

3.007(2)-3.219(2)

3.534(2)-3.555(2)

3.549(2)-3.591(2)

3.532(2)-3.592(2)

3.374(2)-3.384(2)

3.357(2)-3.396(2)

3.347(2)-3.388(2)

^aO may be from oxo, hydroxo, alkoxo, or phenoxo.

Table 5. Fe-O-Fe Bridging Angles (deg) for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚0.5MeOH, 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

a

compd

via (µ-O)₂

101.71(6)

via µ₂-alkoxo

via µ₃-oxo

1‚CH₂Cl₂

2‚CH₂Cl₂

4‚CH₂Cl₂

115.4(2)-124.7(2)

116.1(4)-125.7(4)

126.5(1)

123.6(2)-126.4(2)

125.3(3)-127.9(2)

122.6(2)-126.2(2)

5‚0.5MeOH

100.8(1)-102.9(1)

92.7(1)-104.5(1)

93.7(2)-105.8(2)

92.7(1)-104.4(2)

7‚H₂O‚C₇H₈‚0.5C₆H₁₄

8‚0.25H₂O‚4.5CH₂Cl₂

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

124.7(2)-125.9(2)

123.1(2)-126.0(3)

123.1(2)-126.4(2)

^aO may be from oxo, hydroxo, alkoxo, or phenoxo.

complexes each possess a {Fe^III₃(µ₂-O)₃}³⁺core, with the

three Fe atoms located at the vertexes of a scalene triangle

(Figure 2). Each pair of Fe centers is bridged by an ethoxo-

type O atom of (L¹)^2-or (L³)^2-for 2 and 4, respectively, in

-

addition to a µ₂-MeCO₂ligand. The Schiff base ligands

bind in a fashion similar to that observed in 1. The ligand

arrangement is such that two of the core O atoms and one

-

MeCO₂ligand lie on one side of the Fe₃plane, while the

-

third core O atom and the remaining two MeCO₂ligands

lie on the other. The {Fe₃O₃}³⁺cores and the molecules as

a whole all have C₁point symmetry. Although complexes 2

and 4 are essentially isostructural, inspection of the coordina-

tion geometries at the individual Fe sites reveals that 4 is

somewhat more distorted than 2. This can be quantified by

a consideration of the variation in Fe-X bond lengths, where

Fe1 of 4 displays a variation of 0.31 Å, versus the variation

of 0.17-0.22 Å evident in all of the other Fe centers in 2

and 4. Finally, triangular Fe^III₃clusters and in particular the

basic carboxylates [Fe₃O(O₂CR)₆(X)₃]⁺(X ) e.g., H₂O) are

well-known.²⁹However, complexes with a {Fe₃(µ₂-O)₃}³⁺

core with alkoxo-type O atoms and no central µ₃-bridge are

unknown. There are only two species reported in the

literature with the same core topology, one with a (µ-oxo)₃

core and one with (µ-hydroxo)(µ-oxo)₂.^30,31Both display no

additional ligands bridging the Fe centers.

Figure 1. Structure of 1 in 1‚CH₂Cl₂.

noncoordinated O atoms of the pic^-ligands participate in

an intermolecular O‚‚‚H-C hydrogen bond with the C16

atoms of the CH₂Cl₂molecules of crystallization, with a

C‚‚‚O separation of 3.25 Å.

Complex 2 crystallizes in the orthorhombic space group

Pca2₁, while complex 4 crystallizes in the monoclinic space

group P2₁/c. In addition to two molecules of CH₂Cl₂, the

asymmetric unit of 2 contains two independent trinuclear

complexes with enantiomeric configurations but different

interatomic distances and angles. The asymmetric unit of 4

contains a single trinuclear molecule together with one CH₂-

Cl₂molecule. All three trinuclear complexes are isostructural,

disregarding the differences in the Schiff base ligands. The

(29) Cannon, R. D.; White, R. P. Prog. Inorg. Chem. 1988, 36, 195.

Inorganic Chemistry, Vol. 43, No. 16, 2004 5059

Boskovic et al.

Figure 3. Structure of 5 in 5‚0.5MeOH.

Complexes 7-9 all crystallize in the triclinic space group

P1h. In addition to disordered solvent, the asymmetric units

of 7 and 9 contain one pentanuclear complex, while the

asymmetric unit of 8 contains two independent pentanuclear

molecules. All three complexes are isostructural, with similar

metric parameters and a {Fe₅(µ₃-O)₂(µ₂-O)₅)}⁺core (Figure

4). The five (µ₂-O) atoms are from the Schiff base ligands

and are ethoxo-type (7 and 8, O4, O6, O14, and O16; 9,

O4, O10, O12, and O13) or phenoxo-type (7 and 8, O5; 9,

O11), while the µ₃-bridges are oxo (O2) and hydroxo (O1)

groups. The core unit has a complex structure that, with some

imagination, can be described as an incomplete cubane

extended at one face by an incomplete adamantane unit. The

four carboxylato ligands provide additional bridges, three

binding in the typical µ₂-mode, while the fourth binds in a

terminal manner, displaying an intramolecular O‚‚‚H-O

hydrogen bond to the core hydroxo group, with O‚‚‚O

separations of 2.57, 2.55, and 2.57 Å for 7, 8 and 9,

respectively. Thus, the molecules are unusual in that they

contain oxo-, hydroxo- alkoxo-, and carboxylato- bridging

units. The peripheral ligation is completed by the Schiff base

ligands, with each bound in a fashion similar to that observed

in complexes 1, 2, 4, and 5. However, one Schiff base ligand

in each pentanuclear complex possesses a bridging phenoxo-

type O atom, which is not observed in the other complexes.

The hydroxo H atoms were located crystallographically in

all three structures. In addition, bond valence sum calcula-

tions are consistent with five Fe^IIIcenters for each complex.

A search of the Cambridge Crystallographic Database reveals

that the core unit evident in 7-9 is unprecedented for any

pentanuclear complex. Moreover, only three pentanuclear

Fe^IIIcomplexes have been reported previously, all with

structures very different from those of 7-9.³³

Figure 2. Structure of (a) one of the independent molecules of 2 in 2‚

CH₂Cl₂and (b) 4 in 4‚CH₂Cl₂.

Complex 5 crystallizes in the monoclinic space group

P2/c. The asymmetric unit contains half of the molecule of

interest together with a disordered MeOH molecule that is

25% occupied. The complex contains a {Fe^III₄(µ₂-O)₆} core,

where four of the bridging O atoms are ethoxo-type O atoms

from (L³)^2-and the other two are from additional µ₂-MeO^-

units (Figure 3). The Fe centers are located on the vertexes

of a rectangle, such that the sides of the rectangle are

alternately encompassed by one or two O atom bridges. The

two sides of the rectangle that contain only a single O atom

-

bridge each possess an additional µ₂-MeCO₂bridge. This

arrangement results in two long (3.5 Å) and two short (3.1

Å) Fe‚‚‚Fe separations. The (L³)^2-ligands bind in a fashion

similar to that observed in 1, 2, and 4. Crystallographic C₂

point symmetry is present, with the C₂axis passing through

the center of the {Fe^III₄O₄}⁴⁺core, coinciding with the

crystallographic a axis. Although numerous tetranuclear Fe

complexes are known,³²a search of the Cambridge Crystal-

lographic Database reveals that the core evident in 5 is

unprecedented in Fe chemistry.

Magnetic Studies. Variable-temperature magnetic sus-

ceptibility measurements were performed at 0.1 T on

powdered crystals of 1‚CH₂Cl₂, 2, 4, 6‚H₂O, 7, 8‚H₂O, and

(30) Zheng, H.; Zang, Y.; Dong, Y.; Young, V. G., Jr.; Que, L., Jr. J. Am.

Chem. Soc. 1999, 121, 2226.

(31) Furutachi, H.; Ohyama, Y.; Tsuchiya, Y.; Hashimoto, K.; Fujinami,

S.; Uehara, A.; Suzuki, M.; Maeda, Y. Chem. Lett. 2000, 1132.

(32) Murray, K. S. AdV. Inorg. Chem. 1995, 43, 261.

(33) (a) Mikuriya, M.; Nakadera, K. Chem. Lett. 1995, 213. (b) Mikuriya,

M.; Hashimoto, Y.; Nakashima, S. Chem. Commun. 1996, 295. (c)

O’Keefe, B. J.; Monnier, S. M.; Hillmeyer, M. A.; Tolman, W. B. J.

Am. Chem. Soc. 2001, 123, 339.

5060 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

Figure 5. ø_MT (]) and ø_M(O) vs T for 1‚CH₂Cl₂. The solid lines are fits

as described in the text.

Figure 6. Magnetic data for 2: (a) ø_MT (]) and ø_M(O) vs T and (b)

M(Nµ_Β)^-1vs HT^-1at 1.8 (3), 2.5 (O), 4 (4), and 8 K (]). The solid lines

are fits or simulations of the magnetic susceptibility or magnetization data

as described in the text. The dashed lines in (a) are the simulations of the

susceptibility data calculated using the parameters derived from INS

measurements (see the text).

2.0, suggesting an antiferromagnetic interaction. As the

temperature is decreased to 1.8 K, ø_MT decreases steadily to

∼0 cm³mol^-1K, consistent with an S ) 0 ground state. To

determine the magnitude of the exchange interaction, the

susceptibility data were fit with g fixed to 2.0 to the exchange

Hamiltonian

Figure 4. Structure of (a) 7 in 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, (b) one of the

independent molecules of 8 in 8‚0.25H₂O‚4.5CH₂Cl₂, and (c) 9 in

9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O.

Hˆ _ex) -2J(S_A‚S_B)

yielding J ) -7.3(2) cm^-1.

For 2 (Figure 6a) and 4 (Figure S1a), the value of ø_MT at

300 K is ∼8.2 cm³mol^-1K, which is significantly less than

the calculated spin-only value of 13.1 cm³mol^-1K for three

noninteracting Fe^IIIcenters with g ) 2.0. This and the

decrease in ø_MT as the temperature is decreased are consistent

with overall antiferromagnetic interactions in the molecule.

(7)

9‚2H₂O, in the temperature range 1.8-300 K. The data for

1‚CH₂Cl₂, 2, 6‚H₂O, and 8‚H₂O are plotted in Figures 5, 6a,

7a, and 8a, respectively, as ø_MT vs T and ø_Mvs T, together

with the fits of the experimental data. The data for 4, 7, and

9‚2H₂O are available in Figures S1, S2, and S3, respectively.

For 1‚CH₂Cl₂(Figure 5), ø_MT at 300 K is ∼7.0 cm³mol^-1

K, which is less than the calculated spin-only value of 8.8

cm³mol^-1K for two noninteracting Fe^IIIcenters with g )

Inorganic Chemistry, Vol. 43, No. 16, 2004 5061

Boskovic et al.

The values of ø_MT at 1.8 K are ∼0.35 cm³mol^-1K,

suggesting S ) 1/2 ground states (for an S ) 1/2 system

with g ) 2.0, ø_MT is calculated to be 0.38 cm³mol^-1K at

0 K). Crystallographically, the three Fe centers in complexes

2 and 4 are arranged at the vertexes of a scalene triangle,

resulting in the coupling scheme presented in Chart 2.

Chart 2

Figure 7. ø_MT (]) and ø_M(O) vs T for 6‚H₂O. The solid line is the sum

of the fits of the low-temperature data and the high-temperature data as

described in the text.

The data were fit with g fixed to 2.0 to the exchange

to a value of ∼0 cm³mol^-1K at 1.8 K, consistent with an

S ) 0 ground state. Correspondingly, ø_Mincreases gradually

before reaching a broad maximum at about 150 K and

decreasing gradually as the temperature is decreased to 4.5

K. Below 4.5 K, ø_Mincreases rapidly as the temperature is

decreased, which is attributed to the presence of a paramag-

netic impurity. Fitting the data below 4.5 K to the Curie law

suggests 0.19 mol % monomeric Fe^IIIimpurity. Crystallo-

graphically, the four Fe centers in complex 5 (and presum-

ably 6) are arranged at the vertexes of a rectangle, resulting

in the coupling scheme presented in Chart 3.

Hamiltonian

Hˆ _ex) -2J₁(S_A‚S_B) - 2J₂(S_B‚S_C) - 2J₃(S_A‚S_C) (8)

yielding J₁) -10.4(9) cm^-1, J₂) -9.4(7) cm^-1, and J₃)

-8.6(4) cm^-1for 2 and J₁) -10.7(4) cm^-1, J₂) -10.1(4)

cm^-1, and J₃) -9.0(3) cm^-1for 4. For both compounds,

the three competing antiferromagnetic interactions result in

a spin-frustrated system and an S ) 1/2 ground state. It was

found that a significant improvement was obtained in the fit

using three different exchange parameters rather than one

or two, which is in agreement with the inequivalence of the

three metal centers observed in the X-ray structures. The

similarity between the three coupling constants determined

for each complex is consistent with the similarity between

the three Fe‚‚‚Fe bridging geometries evident in the struc-

tures; however, this similarity precludes the assignment of

particular coupling constants to particular pairwise inter-

actions. In addition, the set of parameters determined for the

two complexes are the same within experimental error, which

is consistent with the similarity in the core topologies and

metric parameters. Finally, it is relevant to the discussion of

inelastic neutron scattering studies that follows to note that

the J values derived for 2 imply an S ) 1/2 excited state at

9.4 cm^-1above the S ) 1/2 ground state and four S ) 3/2

excited states at 19-48 cm^-1above the ground state. A

similar arrangement of energy levels is implied for 4.

Variable-temperature magnetization measurements were

performed on 2 and 4 in the temperature range 1.8-8 K,

with fields up to 5 T. The data for 2 are presented in Figure

6b as plots of M(Nµ_B)^-1vs HT^-1, while the data for 4 are

available in Figure S1b. For both species, the M(Nµ_B)^-1

values at 5 T and 1.8 K of 0.98 and 0.93 for 2 and 4,

respectively, are consistent with S ) 1/2 ground states. It is

possible to reproduce the magnetization data for 2 using the

parameters determined from the fit of the susceptibility data

(Figure 6b). The simulation of the data for 4 is slightly less

satisfactory, which may be due to the presence of intermo-

lecular interactions.

Chart 3

The data were scaled to correct for the monomeric

impurity, and the data above 10 K were fit with g fixed to

2.0 to the exchange Hamiltonian

Hˆ _ex) -2J₁(S_A‚S_B+ S_C‚S_D) - 2J₂(S_B‚S_C+ S_A‚S_D) (9)

yielding the set of parameters J₁) -9.2(2) cm^-1and J₂)

-10.2(4) cm^-1, which result in an S ) 0 ground state. A

second set of parameters with J₁> 0 was found to reproduce

the data equally well, but was rejected because antiferro-

magnetic coupling is typically observed for ferric complexes

with bridging geometries similar to those evident in 5 (and

presumably 6).^34,35Although it was found that a significant

improvement was obtained in the fit using two different

exchange parameters rather than only one, the similarity in

(34) (a) Kato, M.; Yamada, Y.; Inagaki, T.; Mori, W.; Sakai, K.;

Tsubomura, T.; Sato, M.; Yano, S. Inorg. Chem. 1995, 34, 2645 and

references therein. (b) Neves, A.; de Brito, M. A.; Vencato, I.; Drago,

V.; Griesar, K.; Haase, W. Inorg. Chem. 1996, 35, 2360 and references

therein.

(35) (a) Chiari, B.; Piovesana, O.; Tarantelli, T.; Zanazzi, P. F. Inorg. Chem.

1984, 23, 3398. (b) Me´nage, S.; Que Jr, L. Inorg. Chem. 1990, 29,

4293 and references therein. (c) Taft, K. L.; Delfs, C. D.; Papaefthy-

miou, G. C.; Foner, S.; Gatteschi, D.; Lippard, S. J. J. Am. Chem.

Soc. 1994, 116, 823. (d) Raptopoulou, C. P.; Tangoulis, V.; Devlin,

E. Angew. Chem., Int. Ed. 2002, 41, 2386.

For 6‚H₂O (Figure 7), the value of ø_MT at 300 K is ∼10.3

cm³mol^-1K, which is substantially less than the calculated

spin-only value of 17.5 cm³mol^-1K for four noninteracting

Fe^IIIcenters with g ) 2.0. Again the behavior suggests overall

antiferromagnetic interactions, with ø_MT decreasing steadily

5062 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

the magnitudes of J₁and J₂precludes their individual

assignment to the two distinct pairwise interactions.

For 7, 8‚H₂O, and 9‚2H₂O (Figures S2a, 7a, and S3a,

respectively), the value of ø_MT at 300 K is ∼10 cm³mol^-1

K, which is considerably less than the calculated spin-only

value of 22 cm³mol^-1K for five noninteracting Fe^IIIcenters

with g ) 2.0. This and the decrease in ø_MT as the temperature

is decreased are consistent with overall antiferromagnetic

interactions. For the three compounds ø_MT decreases steadily

to ∼4.3 cm³mol^-1K at 35 K, where it levels out before

finally decreasing rapidly below 5 K. This suggests a spin

ground state of S ) 5/2 (for an S ) 5/2 system with g )

2.0, ø_MT is calculated to be 4.4 cm³mol^-1K at 0 K). The

low-temperature decrease in ø_MT is assigned to the effects

of zero-field splitting (ZFS) and/or antiferromagnetic inter-

molecular interactions. A consideration of the topologies of

the complexes affords the coupling scheme depicted in Chart

4.

Chart 4

Figure 8. Magnetic data for 8‚H₂O: (a) ø_MT (]) and ø_M(O) vs T and (b)

M(Nµ_Β)^-1vs HT^-1at 1.8 (3), 2.5 (O), 4 (4), and 8 K (]). The solid lines

are fits or simulations as described in the text.

two coupling constants, assuming that J₂) J₃, gave a

significantly poorer fit, while efforts to fit the data using

more than three coupling constants were not pursued due to

the adequacy of the fit already obtained and the desire to

avoid overparametrization. The sets of parameters determined

for the three complexes are very similar, which is consistent

with the similarity in the core topologies and metric

parameters.

Variable-temperature magnetization measurements were

performed on 7, 8‚H₂O, and 9‚2H₂O in the temperature range

1.8-8 K, with fields up to 5 T. The data for 8‚H₂O are

presented in Figure 8b as plots of M(Nµ_B)^-1vs HT^-1, while

the data for 7 and 9‚2H₂O are available in Figures S2b and

S3b, respectively. For the three species, the M(Nµ_B)^-1values

of ∼4.8 at 5 T and 1.8 K are consistent with S ) 5/2 ground

states. In each case the data deviate slightly from the Brillouin

function for S ) 5/2. This deviation can be reproduced upon

the incorporation of an axial ZFS term, D, and it is possible

to simulate the data using the ZFS Hamiltonian

On the basis of the nature of the Fe-O-Fe bonds, the

eight coupling constants that are strictly required can be

reduced to three, where J₁characterizes couplings through

a (µ-O)₂unit with Fe-O-Fe angles <106° and Fe‚‚‚Fe

distances of 2.9-3.3 Å, J₂characterizes couplings through

a single µ₂-alkoxo bridge with Fe-O-Fe angles of 123-

128° and Fe‚‚‚Fe distances of 3.5-3.6 Å, and J₃characterizes

couplings through a µ₃-oxo bridge with Fe-O-Fe angles

of 123-127° and Fe‚‚‚Fe distances of 3.3-3.4 Å (Tables 4

and 5). The ø_Mand ø_MT vs T data above 10 K were fit with

g fixed to 2.0 to the exchange Hamiltonian

Hˆ _ex) -2J₁(S_A‚S_B+ S_A‚S_C+ S_B‚S_C) - 2J₂(S_B‚S_D+

S_C‚S_D+ S_D‚S_E) - 2J₃(S_B‚S_E+ S_C‚S_E) (10)

yielding the parameters J₁) -3.0(3) cm^-1, J₂) -11.4(6)

cm^-1, and J₃) -46(1) cm^-1for 7, J₁) -3.5(3) cm^-1,

J₂) -12.6(5) cm^-1, and J₃) -53(5) cm^-1for 8‚H₂O, and

J₁) -2.2(3) cm^-1, J₂) -12.3(5) cm^-1, and J₃) -36(2)

cm^-1for 9‚2H₂O. Each set of parameters results in an overall

ground state of S ) 5/2 for the three complexes. A second

set of parameters with J₂, J₃< 0 and |J₂| > |J₃| was found

to reproduce the data equally well for the three complexes,

but was rejected because it has generally been observed that

µ-oxo bridges (characterized by J₃) mediate stronger anti-

ferromagnetic interactions than µ-alkoxo bridges (character-

ized by J₂).³⁴In addition, attempts to fit the data using only

Hˆ _ZFS) D(Sˆ_z²- S(S + 1)/3)

(11)

assuming axial anisotropy and a well-isolated S ) 5/2 ground

state. Two sets of simulation parameters have been deter-

mined for each compound that reproduce the data equally

well: g ) 1.99(1) and D ) -0.46(3) or 0.59(2) cm^-1for 7,

g ) 1.99(1) and D ) -0.45(3) or 0.52(4) cm^-1for 8‚H₂O,

and g ) 2.01(1) and D ) -0.35(2) or 0.42(2) cm^-1for

9‚2H₂O.

In an effort to determine which parameter set obtained

from the magnetization fits is correct and whether the D

Inorganic Chemistry, Vol. 43, No. 16, 2004 5063

Boskovic et al.

Figure 10. Energy-loss INS spectra of 2 at two temperatures. λ ) 3.8 Å,

and the total Q range is from 0.5 to 2.6 Å^-1. Inelastic focusing is at an

energy transfer of 27.5 cm^-1with an instrumental resolution of 1.6 cm^-1

.

The labels of the peaks are given in Table 6. The inset shows the

corresponding energy level and transition diagram based on an isosceles

triangle with J₁) J₂) J and ∆ ) J - J₃< 0.

Table 6. Observed Transition Energies from the INS Spectra of 2

label

energy (cm^-1

)

label

energy (cm^-1

)

I

II

III

20.4(3)

22.9(3)

27.8(3)

IV

V

30.2(3)

7.7(3)

Figure 9. Magnetization hysteresis loops for a single crystal of 8‚

0.25H₂O‚4.5CH₂Cl₂measured parallel to the easy axis (a) with a scan rate

of 0.035 T s^-1and different temperatures and (b) at 0.04 K and different

scan rates.

tunneling is generally observed due to coupling with the

environment through intermolecular dipolar coupling, hy-

perfine coupling, spin-spin cross relaxation, and other

processes.^11,12,37

parameter for the pentanuclear complexes is positive or

negative, low-temperature magnetization measurements were

performed on single crystals of 8‚0.25H₂O‚4.5CH₂Cl₂down

to 0.04 K using a micro-SQUID apparatus. If D < 0, the S

) 5/2 and D ) -0.45 cm^-1values suggest that 8 should

have an anisotropy barrier to magnetization reversal of

(S²- 1/4)|D| ≈ 2.7 cm^-1. This is potentially large enough

for the manifestation of slow relaxation of the magnetization

below 1 K. Figure 9 shows magnetization vs field measure-

ments parallel to the experimentally observed easy axis and

with the magnetization normalized to the saturation value

M_s. First, the experimental observation of an easy axis is

consistent with D < 0. In addition, at low temperatures

hysteresis loops are observed at nonzero magnetic fields

whose coercivity increases with decreasing temperature

(Figure 9a). Detailed investigations of the dependence of the

hysteresis loops on the sweep rate establish that this

hysteresis is not only due to a phonon bottleneck, where the

phonon exchange (thermal coupling) between the crystal and

its environment is retarded, hampering the spin relaxation,³⁶

but also due to slow relaxation of the magnetization

consistent with a molecular anisotropy barrier and D < 0.

However, the size of the barrier is small, and a reliable

Arrhenius plot cannot be obtained. The absence of hysteresis

at zero field is often observed for systems with relatively

low spin ground states and can be attributed to the presence

of rapid quantum tunneling.¹²Despite the fact that theoreti-

cally half-integer spin systems should not tunnel in zero field,

Inelastic Neutron Scattering Studies. INS has previously

proved to be a useful tool for the elucidation of the exchange

coupling in trinuclear Fe complexes with triangular arrange-

ments of the Fe centers, providing data that are complemen-

tary to those obtained from magnetic susceptibility measure-

ments.³⁸For compound 2 the energy levels derived from the

fit of the magnetic susceptibility data are such that transitions

between these levels should be of an energy whose magni-

tude is suitable for detection by INS. Figure 10 shows the

INS spectra of 2 at 1.5 and 18 K, corresponding to the sum

of all of the available detectors, i.e., the total Q range from

0.5 and 2.6 Å^-1. At 1.5 K a pair of peaks centered at around

21.6 cm^-1(Table 6) is observed, together with a pair of peaks

centered at around 29.0 cm^-1(less intense). At 18 K, the

intensity of the pair of peaks at 21.6 cm^-1is reduced whereas

the intensity of the pair of peaks at 29.0 cm^-1is essentially

unchanged. These peaks are also observed on the energy gain

side. The Q dependence (not shown) is consistent with both

pairs of peaks arising from magnetic transitions.

In modeling the data the molecular states are labeled as

|S_AB, S , with the intermediate spin BS_AB) BS_A+ BS_Band total

spin BS ) BS_AB+ BS_C(Chart 2). Calculations using the J values

(37) Boskovic, C.; Wernsdorfer, W.; Folting, K.; Huffman, J. C.; Hen-

drickson, D. N.; Christou, G. Inorg. Chem. 2002, 41, 5107.

(38) (a) Furrer, A.; Gu¨del, H. U. HelV. Phys. Acta 1977, 50, 439. (b)

Cannon, R. D.; Jayasooriya, U. A.; Wu, R.; arapKoske, S. K.; Stride,

J. A.; Nielson, O. F.; White, R. P.; Kearley, G. J.; Summerfield, D. J.

Am. Chem. Soc. 1994, 116, 11869. (c) Sowrey, F. E.; Tilford, C.;

Wocadlo, S.; Anson, C. E.; Powell, A. K.; Bennington, S. M.;

Montfrooij, W.; Jayasooriya, U. A. Cannon, R. D. J. Chem. Soc.,

Dalton Trans. 2001, 862.

(36) Chiorescu, I.; Wernsdorfer, W.; Mu¨ller, A.; Bogge, H.; Barbara, B.

Phys. ReV. Lett. 2000, 84, 3454.

5064 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

Table 7. Mo¨ssbauer Parameters for 2, 4, 7, and 8‚H₂O

determined from the fit of the susceptibility data for 2 and

eq 8 predict INS transitions to and from the |3, 1/2 ground

state at 19.4 cm^-1(|3, 1/2 T |4, 3/2 ) and 28.2 cm^-1(|3,

1/2 T |3, 3/2 ). INS peaks are observed at roughly these

positions and have been assigned accordingly. However, each

of these transitions occurs as a pair of peaks, the first as I

and II and the second as III and IV. This may be attributed

to the presence of two species in the sample, and indeed

two independent molecules with different metric parameters

were observed in the asymmetric unit in the single-crystal

X-ray structure (vide supra). Thus, INS, being a spectroscopic

technique, allows the two inequivalent complexes to be

distinguished, and the INS data were modeled assuming two

species. To avoid overparametrization, the observation of

only two peaks per complex requires the use of an isosceles

triangular model (Chart 2 and eq 8, J₁) J₂) J), rather than

the scalene model used for the susceptibility data. The energy

level diagram for such an antiferromagnetically coupled

isosceles triangle is given in the inset of Figure 10, together

with the peak assignment. Peaks I and III (Table 6) are

assigned to one species, while peaks II and IV belong to the

other species. A hot transition V (not shown) is clearly

observed at 7.7 cm^-1in a high-resolution spectrum. It

corresponds to the |3, 3/2 T |4, 3/2 transitions in both

isomers. The parameters which give the best simulation of

both the INS and susceptibility data are J₁) J₂) -9.27

cm^-1and J₃) -8.34 cm^-1for one isomer and J₁) J₂)

-10.08 cm^-1and J₃) -9.16 cm^-1for the second isomer.

Simulations of the magnetic susceptibility data using these

parameters are presented in Figure 6a and reproduce the

experimental data very well.

Mo1ssbauer Spectroscopy. ⁵⁷Fe Mo¨ssbauer spectra for

compounds 2, 4, 7, and 8‚H₂O were recorded at several

temperatures in the range 4.2-293 K, and an additional

spectrum was acquired for 8‚H₂O at 1.7 K. The spectra

obtained at temperatures >100 K (and at 4.2 K for 2) were

least-squares fit to Lorentzian lines assuming equal contribu-

tions from three (2 and 4) or five (7 and 8‚H₂O) Fe sites to

give the parameters presented in Table 7. These isomer shift

(δ) and quadrupole splitting (∆E_q) parameters are consistent

with the presence of only high-spin Fe^IIIcenters.³⁹The isomer

shift values are observed to decrease with increasing tem-

perature due to the second-order Doppler effect. The spectra

obtained at temperatures <100 K were not fitted for 4, 7,

and 8‚H₂O because of the increase in line width and loss of

resolution observed.

compd

site

T

(K)

δ

∆E_q

(mm s^-1

)

(mm s^-1

)

Γ^a

2a

2b

2c

293

135

4.2

293

135

4.2

293

135

4.2

0.45

0.53

0.55

0.42

0.51

0.59

0.42

0.49

0.46

0.42

0.48

0.42

0.49

0.41

0.46

0.41

0.48

0.39

0.45

0.40

0.51

0.43

0.50

0.43

0.52

0.42

0.46

0.49

0.41

0.46

0.49

0.42

0.45

0.47

0.43

0.46

0.47

0.40

0.46

0.50

0.49

0.52

0.79

0.80

0.93

1.12

1.11

1.05

0.72

0.77

0.87

0.92

1.32

1.35

0.49

0.58

0.74

0.73

0.79

0.74

1.37

1.40

1.83

1.89

0.45

0.48

0.47

0.70

0.71

0.69

0.72

0.74

0.72

1.48

1.53

1.49

1.66

1.72

1.75

0.38

0.30

0.29

0.38

0.30

0.29

0.38

0.30

0.29

0.36

0.30

0.36

0.30

0.36

0.30

0.33

0.34

0.30

0.29

0.34

0.30

0.29

0.34

0.30

0.29

0.34

0.30

0.29

0.34

0.30

0.29

4a

4b

4c

7a

7b

7c

7d

7e

8a

293

140

293

140

293

140

293

110

293

110

293

110

293

110

293

110

293

170

110

293

170

110

293

170

110

293

170

110

293

170

110

8b

8c

8d

8e

^aWidth at half-height.

all six Fe centers (two independent molecules), with little

variation in Fe-X bond lengths (from 0.17 to 0.21 Å). As

a result, assignment of the three sites to particular Fe centers

is not possible. However, as the X-ray and Mo¨ssbauer

measurements were performed on differently solvated spe-

cies, the metric parameters determined crystallographically

may differ from those present in the desolvated form that

was measured by Mo¨ssbauer spectroscopy. For 4, two of

the sites (4a and 4b) have relatively similar ∆E_qvalues, while

4c has a significantly larger value of ∆E_q. Inspection of the

X-ray structural data for 4‚CH₂Cl₂reveals that although all

three Fe centers have similar NO₅coordination environments,

Fe1 shows a significantly greater variation in Fe-X bond

lengths (0.31 Å) than either Fe2 (0.22 Å) or Fe3 (0.19 Å).

Although again the X-ray and Mo¨ssbauer measurements were

performed on differently solvated forms, the data are

consistent with the assignment of site 4c to Fe1, while the

specific assignment of the other two sites remains elusive.

The spectra for 7 (Figure S5) and 8‚H₂O (Figure 12) are

consistent with five overlapping quadrupole-split doublets,

which are better resolved for 8‚H₂O than for 7. The fitting

parameters for 7 and 8‚H₂O characterize five ferric sites with

similar δ values. For 7 the ∆E_qvalues of the five sites lie in

The spectra for 2 (Figure 11) and 4 (Figure S4) are

consistent with three overlapping quadrupole-split doublets

over the temperature range 4.2-293 K. The fitting param-

eters for both 2 and 4 characterize three ferric sites in each

complex with similar δ values. For 2, the three sites possess

three distinct ∆E_qvalues. The ∆E_qvalues reflect the

deviation from octahedral coordination geometry at each site,

with larger ∆E_qvalues associated with a less symmetric

environment. Inspection of the X-ray structural data for

2‚CH₂Cl₂reveals similar NO₅coordination environments for

(39) Murray, K. S. Coord. Chem. ReV. 1974, 12, 1.

Inorganic Chemistry, Vol. 43, No. 16, 2004 5065

Boskovic et al.

Figure 11. Variable-temperature Mo¨ssbauer spectra for 2. The solid lines

are the fits as described in the text.

the order 7a < 7b, 7c < 7d < 7e. The variations in Fe-X

bond lengths derived from the X-ray structure for 7‚H₂O‚

C₇H₈‚0.5C₆H₁₄are (Fe1) 0.23 Å, (Fe2) 0.18 Å, (Fe3) 0.17

Å, (Fe4) 0.22 Å, and (Fe5) 0.23 Å, with NO₅coordination

environments for all but Fe3, which has O₆coordination.

This is consistent with the tentative assignments 7a ) Fe3,

7b ) Fe2, 7c ) Fe4, and 7d, 7e ) Fe1, Fe5. For 8‚H₂O the

five sites can be broadly classified into three groups on the

basis of the ∆E_qvalues, which lie in the order 8a < 8b, 8c

< 8d, 8e. The variations in Fe-X bond lengths derived from

the X-ray structure for 8‚0.25H₂O‚4.5CH₂Cl₂are (Fe1) 0.23

Å, (Fe2) 0.22 Å, (Fe3) 0.17 Å, (Fe4) 0.26 Å, and (Fe5) 0.27

Å, with the same coordination environments as 7. This

suggests the assignments 8a ) Fe3, 8b, 8c ) Fe1, Fe2, and

8d, 8e ) Fe4, Fe5. However, again the X-ray and Mo¨ssbauer

measurements were performed on differently solvated forms

of 7 and 8, whose metric parameters may differ slightly.

Hence, the assignments are not unequivocal, although the

trends between the structural and Mo¨ssbauer data correlate

better for 8 than for 7.

Figure 12. Variable-temperature Mo¨ssbauer spectra for 8‚H₂O. For

temperatures above 100 K the solid lines are the fits as described in the

text. For the lower temperatures the solid lines correspond to the theoretical

spectra derived from the fitting parameters for the 110 K data.

measurements in the presence of a magnetic field would be

required to determine the nature of the complex relaxation

effects which give rise to this behavior. Finally, even down

to 1.7 K the spectrum of 8‚H₂O shows no magnetic splitting

due to slow magnetization relaxation associated with an

anisotropy barrier.

Discussion

The values derived from the fits of the magnetic suscep-

tibility data for the magnetic exchange interactions for

compounds 1‚CH₂Cl₂, 2, 4, 6‚H₂O, 7, 8‚H₂O, and 9‚2H₂O

are presented in Table 8, where the carboxylato bridges are

ignored for the purposes of categorizing the interactions.

With the exception of 6‚H₂O, these coupling constants can

all be assigned to particular superexchange pathways. For

6‚H₂O the similarity in the J values of the two coupling

constants precludes their specific assignment to the two

different bridging units evident in the structure. All of the

determined coupling constants are antiferromagnetic, which

The spectra for both 7 and 8‚H₂O broaden as the

temperature is decreased to 4.2 K, displaying a deviation

from the theoretical Lorentzian line shape. This is charac-

teristic of a decrease in the spin flip frequency, and as a

result the data below 110 K were not fitted. Curiously, the

spectrum of 8‚H₂O at 1.7 K displays a sharpening of the

lines; however, the origin of this is uncertain. Additional

5066 Inorganic Chemistry, Vol. 43, No. 16, 2004

Synthesis of a New Family of Ferric Complexes

Table 8. J (cm^-1) Determined for Compounds 1‚CH₂Cl₂, 2, 4, 6‚H₂O, 7, 8‚H₂O, and 9‚2H₂O

a

compd

via (µ-O)₂

-7.3(2)

via µ₂-alkoxo

via µ₃-oxo

P^b

1‚CH₂Cl₂

2.01

2

4

-8.6(4), -9.4(7), -10.4(9)

-9.0(3), -10.1(4), -10.7(4)

-9.2(2) or -10.2(4)^c

-11.4(6)

-12.6(5)

-12.3(5)

1.98, 2.00, 2.02

1.97, 1.99, 2.01

1.98, 1.99

1.90^d

6‚H₂O

-9.2(2) or -10.2(4)^c

-3.0(3)

7

-46(1)

-53(5)

-36(2)

8‚H₂O

9‚2H₂O

-3.5(3)

-2.2(3)

1.91^d

1.90^d

^aO may be from oxo, hydroxo, alkoxo, or phenoxo. ^bDefined in ref 40. ^cThe asignment of one value to (µ-O)₂and one to µ₂-alkoxo is uncertain. ^dOnly

applicable for the exchange interaction via µ₃-oxo.

is typical for these types of superexchange pathways.^4,34,35

Gorun and Lippard have determined an empirical correlation

between J (cm^-1) and the parameter P (Å), which is the

shortest superexchange pathway defined as the shortest

distance between the metal(s) and the bridging ligand(s), for

pairs of doubly or triply bridged high-spin Fe^IIIcenters:⁴⁰

smaller J values will result from the longer Fe-O and

Fe‚‚‚Fe distances that occur in the µ₃case. Finally, for 7,

8‚H₂O, and 9‚2H₂O, the coupling constant associated with

a single µ₃-oxo bridge is comparable to “body to wing”

coupling constants of -30 to -50 cm^-1determined for Fe^III

4

“butterfly” complexes, which involve a similar bridging

geometry.⁴¹

-J ) (8.763 × 10¹¹)exp(-12.663P)

(12)

The magnetic susceptibility data for the bi-, tri-. and

tetranuclear complexes can all be modeled satisfactorily

without considering the effects of ZFS or intermolecular

exchange interactions. However, for the pentanuclear com-

plexes the rapid decrease in ø_MT as the temperature ap-

proaches zero is consistent with the presence of one or both

of these phenomena. Although the structural and magnetic

measurements were performed on differently solvated forms

of these species, the structural data can give some insight

into possible intermolecular interactions. Typically, inter-

molecular interactions are either propagated through a

superexchange pathway involving intermolecular hydrogen

bonds or arise from dipolar (through-space) interactions.

Inspection of the structural data for 7‚H₂O‚C₇H₈‚0.5C₆H₁₄,

8‚0.25H₂O‚4.5CH₂Cl₂, and 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O

indicates that there are no intermolecular hydrogen bonds

that link Fe₅units, and it seems unlikely that such bonds are

present in 7, 8‚H₂O, and 9‚2H₂O. In addition, the X-ray data

reveal closest intermolecular Fe‚‚‚Fe approaches of

The value of P has been calculated, where applicable, for

the present complexes and included in Table 8. P has not

been calculated for two of the three interactions for the

pentanuclear complexes 7, 8‚H₂O, and 9‚2H₂O, because as

an approximation a single coupling constant was used for

three different superexchange pathways mediated via (µ-O)₂

units, while the interactions through µ₂-alkoxo bridges are

associated with singly bridged Fe units. However, for the

interactions for which P was determined, the values of P

and J are in good agreement with the correlation deduced

by Gorun and Lippard. In addition, the J values determined

for complexes 1-9 are generally in good agreement with

literature values determined for corresponding binding modes

in other ferric complexes. In particular, for exchange

mediated by a single µ₂-alkoxo bridge, J varies from -8 to

-13 cm^-1, which is consistent with the literature values of

-6 to -26 cm^-1.³⁴Moreover, the coupling constants

determined for the (µ₂-alkoxo)₂bridges for 1‚CH₂Cl₂and

6‚H₂O of -7 to -11 cm^-1are consistent with the literature

values of -4 to -27 cm^-1.³⁵However, as mentioned above,

the J values obtained for the interactions mediated by a

(µ-O)₂bridge for 7, 8‚H₂O, and 9‚2H₂O actually characterize

the effect of three separate bridging units: one (µ₂-alkoxo)-

(µ₃-hydroxo)(µ₂-carboxylato), one (µ₂-phenoxo)(µ₃-hydroxo),

and one (µ₃-oxo)(µ₃-hydroxo). Thus, it is difficult to compare

the determined J values with those from the literature or

theory. The literature suggests that the values for the

(µ₂-alkoxo)(µ₃-hydroxo)(µ₂-carboxylato) and (µ₂-phenoxo)-

(µ₃-hydroxo) pathways should be similar, with previously

observed values of -7 to -12 cm^-1.³⁵Although examples

of (µ-oxo)(µ-hydroxo) bridges are relatively rare, one family

of ferric complexes with (µ₂-oxo)(µ₂-hydroxo) bridges is

reported to have J values of -27 to -57 cm^-1.³⁰Neverthe-

less, direct comparison of these species with the present

complexes is difficult, due to the different metric parameters

associated with the µ₃-bridging in 7-9 versus the µ₂-bridging

observed for the literature species. It is not improbable that

7.1-7.2

Å for all three complexes, while closest

centroid‚‚‚centroid (where the centroid is arbitrarily defined

as the center of the (µ-O)₂-bridged Fe₂unit) distances are

11.5, 9.8, and 9.6 Å for 7‚H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚

0.25H₂O‚4.5CH₂Cl₂, and 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O, re-

spectively. Treating the complexes as point dipoles with

S ) 5/2 and using the mean field approximation

E_dip) µ₀(gSµ_B)²/4πVk_Β

(13)

where V is the molecular volume (1698 Å³for 8‚0.25H₂O‚

4.5CH₂Cl₂), the energy associated with the dipolar inter-

actions (E_dip) is determined to be ∼0.01 cm^-1, which is

essentially negligible. Furthermore, experimentally no evi-

dence for strong intermolecular interactions is observed in

the low-temperature micro-SQUID magnetization measure-

(41) (a) McCusker, J. K.; Vincent, J. B.; Schmitt, E. A.; Mino, M. L.; Shin,

K.; Coggin, D. K.; Hagen, P. M.; Huffman, J. C.; Christou, G.;

Hendrickson, D. N. J. Am. Chem. Soc. 1991, 113, 3012. (b) Boudalis,

A. K.; Lalioti, N.; Spyroulias, G. A.; Raptopoulou, C. P.; Terzis, A.;

Bousseksou, A.; Tangoulis, V.; Tuchagues, J. P.; Perlepes, S. P. Inorg.

Chem. 2002, 41, 6474 and references therein.

(40) Gorun, S. M.; Lippard, S. J. Inorg. Chem. 1991, 30, 1625.

Inorganic Chemistry, Vol. 43, No. 16, 2004 5067

Boskovic et al.

ments of 8‚0.25H₂O‚4.5CH₂Cl₂, where a characteristic sig-

moidal shape is often associated with antiferromagnetic

intermolecular interactions.¹⁸Thus, it can be concluded that

the three pentanuclear complexes do not possess inter-

molecular exchange interactions of a significant magnitude.

In the absence of intermolecular interactions, the rapid

decrease in ø_MT as the temperature approaches zero for the

pentanuclear complexes may be attributed to the effect of

ZFS of the S ) 5/2 ground state. In addition, the magnetiza-

tion data can be modeled by considering an axial ZFS

parameter D, with D values of (0.3 to (0.5 cm^-1determined

for the three complexes. Low-temperature micro-SQUID

measurements on 8‚0.25H₂O‚4.5CH₂Cl₂confirm the presence

of an easy axis and therefore a negative D value. D values

of similar sign and magnitude have been determined previ-

ously for other polynuclear ferric complexes.^8-13The origin

of these has been attributed to a combination of the effects

of single-ion anisotropy of the Fe^IIIcenters and spin-spin

interactions. In particular, it has been suggested that a planar

arrangement of the Fe centers in polynuclear ferric complexes

(with axial symmetry and a non-zero spin ground state) is

best suited to provide easy-axis-type magnetic anisotropy

through spin-spin interactions.¹⁰This type of anisotropy is

necessary for SMM behavior, and indeed most of the reported

ferric SMMs do feature an essentially planar arrangement

of the Fe atoms.^8-13

The S ) 5/2 and D ≈ -0.4 cm^-1values suggest that the

pentanuclear complexes should have an anisotropy barrier

to magnetization reversal of (S²- 1/4)|D| ≈ 2.4 cm^-1.

Although this barrier is small, low-temperature micro-SQUID

measurements on 8‚0.25H₂O‚4.5CH₂Cl₂reveal hysteresis at

nonzero field, consistent with slow magnetization relaxation

associated with a molecular anisotropy barrier. The faster

time scale of Mo¨ssbauer spectroscopy allows for the

observation of slow relaxation at temperatures higher than

those associated with magnetic measurements. However,

Mo¨ssbauer spectra of 8‚H₂O down to 1.7 K display no

evidence of the magnetic splitting that would indicate such

slow relaxation. It is probable that measurements at even

lower temperatures are necessary to detect slow relaxation

associated with such a small energy barrier.

metric parameters (and enantiomeric configurations). The

parameters derived from modeling the INS data afford a

simulation of the susceptibility data that reproduces the

experimental data extremely well (Figure 6a). This demon-

strates the utility of the spectroscopic technique of INS in

this situation, where the presence of two distinct complexes

is clearly evident from the data. In contrast, magnetic

susceptibility measurements are bulk measurements which

can give no indication of the presence of multiple species

and which allow only a determination of J values that

represent an average of the values associated with the

different complexes.

Conclusions

Although it has previously been little used in such a

capacity, Fe(O₂CMe)₂has provided a starting point for the

synthesis of a whole new family of polynuclear complexes.

Following reaction with some simple polydentate Schiff base

proligands, oxidation by air, and, in some cases, reaction

with additional carboxylic acids, nine new bi-, tri-, tetra-,

and pentanuclear ferric complexes have been obtained. These

complexes feature unusual structures and novel core topolo-

gies. The exclusively antiferromagnetic exchange interactions

propagated through these iron-oxygen cores have afforded

spin ground states of 1/2 and 5/2 for the tri- and pentanuclear

ferric complexes, respectively. INS studies have allowed a

thorough exploration of the magnetic behavior of one of the

trinuclear complexes, confirming that the observed behavior

arises from two distinct species. Low-temperature magnetic

studies have revealed that the pentanuclear complexes

possess a small energy barrier to magnetization reversal that

results from the molecular anisotropy. Thus, these species

represent new examples of ferric complexes with a nonplanar

arrangement of Fe centers that display slow relaxation of

the magnetization associated with an anisotropy-induced

energy barrier of molecular origin.

Acknowledgment. This work was financially supported

by the Swiss National Science Foundation. Stefan Ochsen-

bein and Roland Bircher are acknowledged for their as-

sistance in collection of the INS data.

Supporting Information Available: X-ray crystallographic files

in CIF format for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚0.5MeOH, 7‚

H₂O‚C₇H₈‚0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and 9‚1.5H₂O‚2CH₂-

Cl₂‚0.5C₂H₄O, tables of selected interatomic distances and angles

for 1‚CH₂Cl₂, 2‚CH₂Cl₂, 4‚CH₂Cl₂, 5‚0.5MeOH, 7‚H₂O‚C₇H₈‚

0.5C₆H₁₄, 8‚0.25H₂O‚4.5CH₂Cl₂, and 9‚1.5H₂O‚2CH₂Cl₂‚0.5C₂H₄O,

plots of (a) ø_MT and ø_Mvs T and (b) M(Nµ_B))^-1vs HT^-1for 4, 7,

and 9‚2H₂O, and variable-temperature Mo¨ssbauer spectra for 4 and

7 (PDF). This material is available free of charge via the Internet

at http://pubs.acs.org.

An INS study of the trinuclear complex 2 was performed

to gain additional insight into the magnetic behavior. The

susceptibility data were fit assuming a single species and a

scalene triangular arrangement of the Fe centers. However,

from the INS data it was apparent that two distinct species

were responsible for the observed behavior, and the data were

modeled assuming two isomeric complexes, each with an

isosceles triangular arrangement of the Fe centers. This is

consistent with the observation of two independent molecules

in the crystallographic asymmetric unit, which have different

IC049600F

5068 Inorganic Chemistry, Vol. 43, No. 16, 2004

Article Doi

Synthesis and characterization of a new family of bi-, tri-, tetra-, and pentanuclear ferric complexes

DOI: 10.1021/ic049600f

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ic049600f

Full text of DOI:10.1021/ic049600f

Products guided by the article

R&D Labs maybe for 748781-56-8

Relevant to this article

Hot Product