Fe(III) clusters built with tripodal alcohol ligands

Article abstract of DOI:10.1016/j.poly.2005.08.034

Reaction of the tripodal ligand 1,1,1-tris(hydroxymethyl)ethane (H ₃thme) with a variety of Fe(III) precursors under both ambient and solvothermal conditions has led to the isolation of a number of new complexes ranging in nuclearity from 4 to 16. These include the rhomb or diamond-like clusters [Fe₄(thme)₂Cl₄(O₂CR) ₂(R′OH)₂] (1, R = Me, R′ = Et; 2, R = Ph, R′ = Me; 3, R = CMe₃, R′ = Me); the related nonanuclear and undecanuclear clusters [Fe₉O₄(O₂CCMe ₃)₁₃(thme)₂] (4) and [NEt₄][Fe ₁₁O₄(O₂CPh)₁₀(thme) ₄(dmhp)₂Cl₄] (5) (dmhp = 4,6-dimethyl-2- hydroxypyrimidine) and the octametallic and hexadecametallic wheels [Fe ₈(O₂CR)₁₂(thme)₄] (6, R = Ph; 7, R = CMe₃) and [Fe₁₆(EtO)₄(O₂CPh) ₁₆(Hthme)₁₂](NO₃)₄ (8). Magnetic studies reveal the presence of dominant antiferromagnetic interactions between the Fe centres in all eight complexes which results in the presence of small spin ground states for complexes 1-4 and 6-8. For 5 however, these stabilise an S = 11/2 spin ground state with an axial zero-field splitting parameter of D = -0.46 cm^-1. Single crystal hysteresis loops in magnetisation versus external magnetic field measurements show the temperature and sweep rate dependence indicative of single-molecule magnetism behaviour.

Full text of DOI:10.1016/j.poly.2005.08.034

Polyhedron 25 (2006) 325–333
www.elsevier.com/locate/poly
Fe(III) clusters built with tripodal alcohol ligands
a
a
a
a
Leigh F. Jones , David M. Low , Madeleine Helliwell , James Raftery ,
David Collison , Guillem Aromı , Joan Cano b,c, Talal Mallah ,
Wolfgang Wernsdorfer , Euan K. Brechin *, Eric J.L. Mcinnes
a
b
c
´
d
a,e,
,a
*
a
School of Chemistry, The University of Manchester, Oxford Road, Manchester M13 9PL, UK
b
`
Departament de Quı´mica Inorganica, Facultat de Quımica, Universitat de Barcelona, Diagonal 647, E-08028 Barcelona, Spain
c
´
Laboratoire de Chimie Inorganique, UMR CNRS 8613, Universite´ Paris-Sud, F-91405, Orsay, France
d
Laboratoire Louis Ne´el -CNRS, 38042 Grenoble, Cedex 9, France
School of Chemistry, The University of Edinburgh, West Mains Road, Edinburgh EH9 3JJ, UK
e
Received 21 April 2005; accepted 3 August 2005
Available online 29 September 2005
Dedicated to Prof. Malcolm Chisholm on the occasion of his 60th birthday.
Abstract
Reaction of the tripodal ligand 1,1,1-tris(hydroxymethyl)ethane (H3thme) with a variety of Fe(III) precursors under both ambient and
solvothermal conditions has led to the isolation of a number of new complexes ranging in nuclearity from 4 to 16. These include the
rhomb or diamond-like clusters [Fe4(thme)2Cl4(O2CR)2(R0OH)2] (1, R = Me, R0 = Et; 2, R = Ph, R0 = Me; 3, R = CMe3, R0 = Me);
the related nonanuclear and undecanuclear clusters [Fe9O4(O2CCMe3)13(thme)2] (4) and [NEt4][Fe11O4(O2CPh)10(thme)4(dmhp)2Cl4]
(5) (dmhp = 4,6-dimethyl-2-hydroxypyrimidine) and the octametallic and hexadecametallic wheels [Fe8(O2CR)12(thme)4] (6, R = Ph;
7, R = CMe3) and [Fe16(EtO)4(O2CPh)16(Hthme)12](NO3)4 (8). Magnetic studies reveal the presence of dominant antiferromagnetic
interactions between the Fe centres in all eight complexes which results in the presence of small spin ground states for complexes 1–4
and 6–8. For 5 however, these stabilise an S = 11/2 spin ground state with an axial zero-field splitting parameter of D = ꢀ0.46 cmꢀ1
.
Single crystal hysteresis loops in magnetisation versus external magnetic field measurements show the temperature and sweep rate depen-
dence indicative of single-molecule magnetism behaviour.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Keywords: Iron; Clusters; Single-molecule magnets
1. Introduction
this imparts a molecular magnetic memory effect that can
be observed as hysteresis loops in M versus H studies.
All SMMs made to date only exhibit this memory effect
at very low temperatures and there is therefore an intense
effort in synthesizing larger clusters in an attempt to gener-
ate larger spin ground states and higher energy barriers to
loss of magnetization [2]. Here, we describe the synthesis of
some new FeIII clusters with the tripodal ligand 1,1,1-
tris(hydroxymethl)ethane (H3thme). Tripodal ligands such
as H3thme have previously been used in the solvothermal
synthesis of oxo-vanadium and oxo-molybdenum clusters
but until recently, very rarely in the synthesis of paramag-
netic 3d transition metal clusters [3]. To date most cluster
Polymetallic cluster complexes of paramagnetic transi-
tion metal ions have attracted intense study since the dis-
covery that such molecules can display the phenomenon
of single-molecule magnetism (SMM) [1]. There is a barrier
to relaxation of magnetisation in SMMs due to the combi-
nation of a large ground state spin multiplicity and a signif-
icant negative zero-field splitting of that ground state and
*
Corresponding author. Tel.: +44 131 650 7545; fax: +44 131 650 4743.
E-mail addresses: ebrechin@staffmail.ed.ac.uk (E.K. Brechin), eric.
mcinnes@manchester.ac.uk (E.J.L. Mcinnes).
0277-5387/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2005.08.034

326
L.F. Jones et al. / Polyhedron 25 (2006) 325–333
synthesis has involved ‘‘conventional’’ coordination chem-
istry techniques, i.e., solution chemistry under atmospheric
pressure and at temperatures limited to the boiling points
of common solvents. Solvothermal techniques allow the
application of high temperatures to reactions in low boiling
solvents and are also an excellent method for the prepara-
tion of pure, crystalline products [4]. In this work, we have
explored reactions of FeIII precursors with H3thme under
both ambient and solvothermal conditions and we observe
a very wide range of cluster topologies and nuclearities
depending on subtle changes in the reaction conditions.
re-dissolved in methanol (15 ml) and filtered. Yellow-
brown crystals grew upon diffusion of diethyl ether after
2–3 weeks. Yield: ꢁ17 %. Elemental Anal. Calc. for 3: C,
32.47; H, 4.77; Cl, 15.97; Fe, 25.16. Found: C, 31.89; H,
4.46; Cl, 16.50; Fe, 24.39%. IR (KBr) m /cmꢀ1): 2932 (w),
2863 (w), 1587 (m), 1561 (w), 1514 (s), 1445 (w), 1401 (s),
1177 (w), 1119 (m), 1037 (m), 1017 (s), 985 (w), 913 (w),
723 (s), 677 (m), 607 (m), 487 (s).
2.1.4. [Fe9O4(O2CCMe3)13(thme)2] (4)
[Fe3O(O2CCMe3)6(H2O)3](O2CCMe3) (0.30 g, 0.32 mmol)
and H3thme (0.038 g, 0.32 mmol) were dissolved in acetoni-
trile (6 ml) and placed in a 23 ml Teflon-lined autoclave.
This was then heated at a rate of 1.0 ꢁC minꢀ1 to 150 ꢁC
and maintained at this temperature for 12 h. Cooling to
room temperature at a rate of 0.05 ꢁC minꢀ1 resulted in
the formation of large orange/brown crystals. These were
washed with several portions of fresh acetonitrile and air
dried. Yield: ꢁ50 %. Elemental Anal. Calc. for 4: C,
44.75; H, 6.63; Fe, 20.80. Found: C, 45.13; H, 7.26; Fe,
20.93%. IR (KBr m/cmꢀ1): 29.62 (s), 2931 (m), 2903 (m),
2873 (m), 1567 (s), 1538 (s), 1484 (s),1459 (m) 1428 (s),
1393 (w), 1379 (s), 1362 (m), 1230 (s), 1153 (m), 1073 (s),
1015 (s), 966 (w), 939 (w), 919 (w), 896 (w), 788 (m), 608
(s), 441 (s).
2. Experimental
2.1. Synthesis
The complexes [Fe3O(O2CR)6(H2O)3]X (R = Me,
CMe3, Ph; X = Cl, NO3, O2CMe3) and [NEt4]2[Fe2OCl6]
were made as previously described [5].
2.1.1. [Fe4(thme)2Cl4(O2CMe)2(EtOH)2] (1)
FeCl3 Æ 6H2O (2.0 g, 7.39 mmol), H3thme (0.30 g,
2.46 mmol) and NaO2CMe (0.61 g, 7.39 mmol) were dis-
solved in acetonitrile (30 ml) with stirring. The resultant
yellow solution was stirred for 8 h, filtered and taken to
dryness under reduced pressure. The resulting residue was
then re-dissolved in MeOH (15 ml) and filtered. Yellow
crystals grew upon diffusion of ethanol during three weeks.
Yield: ꢁ17%. Elemental Anal. Calc. for 1: C, 45.86; H,
7.60; Fe, 15.99. Found: C, 45.91; H, 7.68; Fe, 16.02%. IR
(KBr; m/cmꢀ1): 2997 (w), 2953 (m), 2912 (w), 2862 (w),
1527 (s), 1461 (w), 1433 (s), 1399 (w), 1321 (m), 1208 (w),
1119 (m), 1016 (s), 979 (s), 917 (w), 866 (m), 722 (w), 663
(m), 608 (m), 551 (w), 502 (s), 413 (w).
2.1.5. [NEt4][Fe11O4(O2CPh)10(thme)4(dmhp)2Cl4] (5)
[NEt4][Fe2OCl6] (0.5 g, 0.83 mmol), NaO2CPh (0.24 g,
1.66 mmol) dmhp (0.103 g, 0.83 mmol) and H3thme
(0.1 g, 0.83 mmol) were added to MeCN (30 ml) and stirred
for 12 h and then filtered. The resulting orange solution
was allowed to evaporate yielding complex 5 in approxi-
mately 20% yield after one week. Elemental Anal. Calc.
for 5: C, 45.86; H, 4.24; N, 2.43. Found: C, 45.15; H,
4.23; N, 2.57%. IR (KBr m/cmꢀ1): 1590 (w), 1533 (s),
1480 (m), 1447 (w), 1394 (s), 1171 (w), 1123 (w) , 1047
(m), 1021 (m), 996 (m), 778 (w), 721 (m), 677 (w), 608
(w), 477 (m).
2.1.2. [Fe4(thme)2Cl4(O2CPh)2(MeOH)2] (2)
FeCl3 Æ 6H2O (2.0 g, 7.39 mmol), H3thme (0.30 g,
2.46 mmol) and NaO2CPh (1.06 g, 7.39 mmol) were dis-
solved in methanol (30 ml) with stirring. The yellow solu-
tion was stirred for 8 h, filtered and taken to dryness
under reduced pressure. The resulting residue was re-
dissolved in methanol (15 ml) and filtered. Yellow-brown
crystals grew upon diffusion of diethyl ether during three
weeks. Yield: ꢁ20 %. Elemental Anal. Calc. for 2: C,
37.00; H, 4.14; Cl, 14.56; Fe, 22.94. Found: C, 37.69; H,
3.61; Cl, 14.59; Fe, 22.25%. IR (KBr; m/cmꢀ1): 2976 (w),
2934 (w), 2867 (w), 1600 (m), 1562 (m), 1519 (m), 1493
(w), 1406 (s), 1178 (m), 1121 (m), 1070 (m), 1017 (s), 918
(w), 834 (w), 790 (w), 721 (m), 676 (w), 607 (w), 488 (s),
406 (w).
2.1.6. [Fe8(O2CPh)12(thme)4] (6)
(a) [Fe3O(O2CPh)6(H2O)3]Cl (0.30 g, 0.30 mmol) and
H3thme (0.036 g, 0.30 mmol) were dissolved in acetonitrile
and placed in a 23 ml Teflon-lined autoclave. This was then
heated at a rate of 1.0 ꢁC minꢀ1 to 150 ꢁC and maintained
at this temperature for 12 h. Cooling to room temperature
at a rate of 0.05 ꢁC minꢀ1 resulted in the formation of large
orange/brown crystals. These were washed with several
portions of fresh acetonitrile and air dried. Yield: ꢁ50%.
(b) [Fe3O(O2CPh)6(H2O)3](NO3) (0.5 g 0.49 mmol) and
H3thme (0.06 g, 0.49 mmol) were stirred in MeCN (30 ml)
for 12 h. The resulting precipitate was filtered, air-dried
and re-dissolved in CH2Cl2. Crystals formed in two days
after diffusion of diethyl ether into the solution in approx-
imately 30% yield.
2.1.3. [Fe4(thme)2Cl4(O2CCMe3)2(MeOH)2] (3)
FeCl3 Æ 6H2O (2.0 g, 7.39 mmol), H3thme (0.30 g,
2.46 mmol) and NaO2CCMe3 (0.92 g, 7.39 mmol) were dis-
solved in methanol (30 ml) with stirring. The yellow solu-
tion was stirred for 8 h, filtered and taken to dryness
under reduced pressure. The resulting residue was
Elemental Anal. Calc. for 6: C, 52.74; H, 4.09; Fe, 18.75.
Found: C, 52.73; H, 4.10; Fe, 18.86%. IR (KBr m/cmꢀ1):
3055 (w), 1598 (m), 1555 (m), 1530 (m), 1521 (m), 1429

L.F. Jones et al. / Polyhedron 25 (2006) 325–333
327
(s), 1307 (w), 1177 (m), 1152 w), 1065 (m), 1023 (w), 1011
(m), 838 (w), 718 (s), 676 (m), 623 (m), 588 (m), 477 (s),
442 (w).
2.1.7. [Fe8(O2CCMe3)12(thme)4] (7)
[Fe3O(O2CCMe3)6(H2O)3]Cl (0.30 g, 0.34 mmol) and
H3thme (0.041 g, 0.34 mmol) were dissolved in acetonitrile
and placed in a 23 ml Teflon-lined autoclave. This was then
heated at a rate of 1.0 ꢁC minꢀ1 to 150 ꢁC and maintained
at this temperature for 12 h. Cooling to room temperature
at a rate of 0.05 ꢁC minꢀ1 resulted in the formation of large
orange/brown crystals. These were washed with several
portions of fresh acetonitrile and air dried. Yield: ꢁ50 %.
Elemental Anal. Calc. C, 45.15; H, 6.82; Fe, 20.95. Found:
C, 45.13; H, 7.07; Fe, 20.99%. IR (KBr) m/cmꢀ1: 2962 (m),
2931 (w), 1567 (s), 1536 (s), 1484 (s), 1427 (s), 1379 (m),
1362 (m), 1231 (s), 1153 (w), 1071 (s), 1015 (m), 897 (w),
789 (w), 606 (m), 442 (s).
2.1.8. [Fe16(EtO)4(O2CPh)16(Hthme)12](NO3)4 (8)
[Fe3O(O2CPh)6(H2O)3](NO3) (0.5 g, 0.49 mmol) and
H3thme (0.06 g, 0.49 mmol) were stirred in MeCN (30 ml)
for 12 h. The resulting precipitate was filtered, air-dried
and re-dissolved in a 1:1 mixture of CH2Cl2 and EtOH.
Crystals formed after two days in approximately 15% yield.
Elemental Anal. Calc. C, 46.22; H, 4.74; N, 1.20. Found: C,
46.08; H, 4.41; N, 1.14%. Selected IR data (KBr pellet,
cmꢀ1): 1554 (s), 1528 (s) 1493 (m), 1409 (s), 1307 (w),
1176 (m), 1158 (w), 1069 (w), 1045 (m), 1025 (m), 877
(w), 717 (s).
2.2. X-ray crystallography and structure solution
All structures were solved by direct methods. The data
are summarised in Table 1.
For complexes 1, 2, and 6 all the non-hydrogen atoms
were refined anisotropically and the H atoms were included
in calculated positions. 3: the asymmetric unit contains half
a molecule and two disordered MeOH molecules. The
CMe3 group of the pivalates was disordered. All non-H
atoms were refined anisotropically except those of the dis-
ordered atoms of the solvent at lower occupancy. H atoms
were included in calculated positions except those bonded
to C1T and C2T which were omitted. 4: a number of the
CMe3 groups were disordered over two sites whose occu-
pancies were constrained to sum to 1.0 with restraints on
the geometric parameters. All non-H atoms were refined
anisotropically except the disordered atoms. H atoms were
included in calculated positions. 5: the asymmetric unit
contains half the molecule together with a number of sol-
vent fragments; there are two CH2Cl2 molecules and three
highly disordered fragments which were not recognisable
so the atoms were all defined as C with some at occupancy
0.5. The non-H atoms were refined anisotropically. H
atoms were included in calculated positions except those
of the disordered solvent. 7: the atoms C23–C25 and
C42, C43, C45 were disordered over two sites each whose

328
L.F. Jones et al. / Polyhedron 25 (2006) 325–333
occupancies were constrained to sum to 1.0. There were
two partially occupied MeCN solvent molecules one of
which was incomplete. All non-H atoms were refined aniso-
tropically except the disordered atoms and those of the sol-
vent. H atoms were included in calculated positions except
those of the solvent. 8: initially, the R value was very high
and twinning was suspected. A twin matrix which ex-
changes the h and k axes was applied (01010000ꢀ1),
and the twin fraction was found to be 0.44788. Accounting
for the twinning reduced the R value from over 30% to un-
der 10%. The asymmetric unit contains one quarter of the
molecule together with a nitrate and a disordered solvent
fragment, assumed to be CH2Cl2, and some water mole-
cules. The phenyl rings were constrained to ideal geometry.
Most of the non-H atoms were refined anisotropically ex-
cept those of the solvent fragments, using restraints. H
atoms were included in calculated positions except for the
solvent H atoms.
3. Crystal structure description
3.1. Structure of [Fe4(thme)2Cl4(O2CR)2(R0OH)2]
(1, R = Me, R0 = Et; 2, R = Ph, R0 = Me; 3, R = CMe3,
R0 = Me)
The core structures of complexes 1–3 (Fig. 1) are iso-
structural and consist of a central [Fe4O6]6+ planar rhomb
or diamond in which each of the oxygen atoms is derived
from bridging [thme]3ꢀ ligands. These sit one-above and
one-below the [Fe4] plane and each provide one l3-oxygen
arm that bridges between the two central FeIII ions and one
outer FeIII ion, and two l2-oxygen arms that each bridge
between an outer and central FeIII ion. These ions are also
bridged by the two carboxylates which adopt their com-
mon syn, syn l-mode. Each of the four FeIII ions contains
a terminal chloride ion, while the coordination spheres of
each of the two terminal Fe ions is completed by one R0OH
molecule. The Fe centres are in distorted octahedral geom-
etries and all are in the +3 oxidation state as confirmed by
a combination of charge balance and bond length consider-
ations and BVS calculations.
Fig. 1. The structure of complex 1 viewed perpendicular (top) and parallel
(bottom) to the [Fe4] plane.
with the [Fe4O6]6+ rhomb. The [Fe4O6]6+ unit and the
[Fe3O]7+ triangles are linked to each other through the thir-
teen carboxylate ligands that bind in three different modes:
the usual l, g1, g1, mode, the less common l, g2, g1 mode
3.2. Structure of [Fe9O4(O2CCMe3)13(thme)2] (4)
Complex 4 (Fig. 2) has a near planar rhomb-like array
of nine FeIII ions linked via a combination of oxide and
alkoxide ions to give a [Fe9(O2ꢀ)4(RO1ꢀ)6]13+ core. Closer
inspection reveals that the complex contains the identical
[Fe4O6]6+ rhomb (Fe6–Fe9, O25–O30) seen in complexes
1–3. To this rhomb are now attached four [Fe3O]7+ trian-
gular units which each share one Fe ion (Fe8, Fe9 or
Fe6) with the [Fe4O6]6+ rhomb. The two central [Fe3O]7+
triangular units (containing Fe2, Fe3, Fe9 and O6 and
Fe3, Fe4, Fe9 and O11) edge share to form a [Fe4O2]8+
rhomb. The two peripheral [Fe3O]7+ triangles (containing
Fe4, Fe5, Fe6 and O16, and Fe1, Fe2, Fe8 and O1) each
share one vertex with the [Fe4O2]8+ rhomb and one vertex
Fig. 2. The structure of complex 4 viewed perpendicular to the [Fe9]
ꢀplaneꢁ.

L.F. Jones et al. / Polyhedron 25 (2006) 325–333
329
(e.g., between Fe1 and Fe2), and as simple chelates (com-
pleting the coordination geometries of Fe(5) and Fe(7)).
All the Fe ions are in the +3 oxidation state and exhibit
distorted octahedral geometries, with the exception of
Fe(4), which is five-coordinate and in a distorted square
based pyramidal geometry.
While there are many examples of tetranuclear Fe aggre-
gates, complex 4 is only the fifth nonanuclear Fe cluster re-
ported [10] and complex 5 is only the second undecanuclear
Fe complex reported. The original [Fe11] has no structural
resemblance to 5 being based on a penta-capped trigonal
prism [5].
3.3. Structure of [NEt4][Fe11O4(O2CPh)10(thme)4-
(dmhp)2Cl4] (5)
3.4. Structure of [Fe8(O2CR)12(thme)4] (6, R = Ph; 7,
R = CMe3)
The structure of complex 5 (Fig. 3) closely resembles
that observed in 1–3 and 4. This time the structure contains
two identical [Fe4O6]6+ rhombs (containing Fe2, Fe4, Fe5
and Fe6 and symmetry equivalents), one at each ꢀendꢁ of
the molecule, linked to a central linear [Fe3O4]1+ unit (con-
taining Fe1, Fe3 and Fe1A). Alternatively the core can be
thought of as four vertex sharing [Fe4]n+ units in which the
two peripheral [Fe4O6]6+ rhombs are connected to two cen-
tral [Fe4O2]8+ rhombs (comprising Fe1, Fe2 and symmetry
equivalents) which themselves share one ꢀvertex (Fe3) form-
ing a [M7O4]13+ sub-unit. The 10 PhCO2 ligands bridge in
their familiar syn, syn l-mode both within and between the
various [Fe4] units. The coordination geometries of Fe2
and Fe4 are completed by a combination of terminal Clꢀ
ions and neutral dmhp ligands. These latter ligands are
protonated with the nitrogen atom hydrogen bonding to
a l2-oxygen of a thme3ꢀ ligand. The [M4] units seen in
all of the above complexes are common building blocks
in the structures of FeIII and MnIII clusters and the
[M7O4]13+ sub-unit is also a common feature in larger clus-
ters, including [Fe17]/[Fe19] [6], [Mn8] [7], and [Fe11] clusters
[8]. In fact the [Fe11] core is similar to a [Mn11] cluster re-
ported by Christou and co-workers [9], but in this case
the peripheral units are [Mn4] cubes not rhombs.
The structure of both complex 6 and complex 7 (Fig. 4)
can be described as a wheel of eight FeIII ions bridged by a
combination of RCO2ꢀ and thme3ꢀ ligands. The RCO2ꢀ li-
gands bridge neighbouring FeIII ions in the usual syn, syn
l-manner while the triply-deprotonated thme3ꢀ ligands
show much greater flexibility in each bridging four iron
centres – each arm providing a l2-oxygen for adjacent met-
als. The iron ions all lie in distorted octahedral geometries
and are of four types: Fe1, Fe3, Fe5 and Fe7 are bridged to
three RCO2 and twꢀo thme3ꢀ ligands; Fe2 and Fe6 are
ꢀ
bound to four RCO2 and two thme3ꢀ ligands; and Fe4
and Fe8 bond to two RCO2 and two thme3ꢀ ligands.
ꢀ
The ‘‘Fe8’’ wheel is non-planar and perhaps best described
˚
as ladle-shaped with a diameter of approximately 8 A.
3.5. Structure of [Fe16(EtO)4(O2CPh)16-
(Hthme)12](NO3)4 (8)
The size of the wheel seen in complex 7 is doubled by the
addition of a coordinating alcohol in the crystallisation
step – use of EtOH produces the hexadecametallic
wheel [Fe16(EtO)4(O2CPh)16(Hthme)12](NO3)4. Complex 8
(Fig. 5) can be described as a wheel of sixteen FeIII ions
bridged by PhCO2ꢀ, Hthme2ꢀ and EtOꢀ ligands. The
Fig. 3. The structure of complex 5 viewed perpendicular to the [Fe11] ꢀplaneꢁ.

330
L.F. Jones et al. / Polyhedron 25 (2006) 325–333
gands occupy the central cavity of the wheel: the deproto-
nated arms again acting as l2-bridges, with the protonated
arm hydrogen bonding to the monodentate arm of an adja-
cent Hthme2ꢀ ligand (for example, O11–O15B 2.591(6) A)
˚
and to a water molecule above the centralꢀcavity of the
˚
wheel (O11–O1S, 2.804(6)) A. The four EtO ligands each
bridge two iron centres on the outside edge of the wheel.
The iron ions are all in distorted octahedral geometries
and bound to six oxygen atoms. For Fe(1) these are derived
from three PhCO2ꢀ, two Hthme2ꢀ ligands and one EtOꢀ.
For Fe(2) from two PhCO2ꢀ, three Hꢀthme2ꢀ ligands and
one EtOꢀ. For Fe(3) from one PhCO2 and five Hthme2ꢀ
ꢀ
ligands, and for Fe(4) from two PhCO2 and four
Hthme2ꢀ ligands. Again the wheel is non-planar and best
described as bowl-shaped with an approximate diameter
˚
of 16 A.
4. Magnetic measurements
4.1. Magnetism of [Fe4(thme)2Cl4(O2CR)2(R0OH)2]
(1–3)
Solid state DC susceptibility measurements of com-
plexes 1–3 were taken in the range of 2–300 K in a field
of 3 kG. The data for all three complexes is essentially
identical and here we limit our discussion to complex 1.
The vmT value of 10.61 cm3 molꢀ1 K at 300 K gradually
decreases with decreasing temperature to a minimum of
1.24 cm3 molꢀ1 K at 6 K before increasing slightly to
1.52 cm3 molꢀ1 K at 2 K. The spin only (g = 2) value ex-
pected for four non-interacting FeIII centres is approxi-
mately 17.5 cm3 molꢀ1 K. This behaviour is indicative of
antiferromagnetic exchange between the FeIII centres with
the vmT value at low temperature consistent with a ground
state spin of S = 0 or S = 1. Magnetisation data were col-
lected in the range of 0–5.0 T at 2 K. The magnetisation in-
creases beyond the expected saturation value for S = 1,
most likely because higher multiplicity states become more
stable as the magnetic field is increased due to the close
proximity of these states to the ground state. Because of
this no fit of the data was attempted.
Fig. 4. The structure of complex 6 viewed parallel to the [Fe8] ꢀplaneꢁ.
4.2. Magnetism of [Fe9O4(O2CCMe3)13(thme)2] (4)
Solid state DC susceptibility measurements of complex 4
were taken in the range 2–300 K in a field of 5 kG. The vmT
value of 28.3 cm3 molꢀ1 K at 300 K gradually decreases
with decreasing temperature approaching a value of
approximately 0.40 cm3 molꢀ1 K at 2 K. The room temper-
ature vmT value is smaller than the spin only value for nine
non-interacting high spin FeIII centres (39.38 cm3 molꢀ1 K)
indicating the presence of antiferromagnetic interactions.
The low temperature value is close to that expected for a
complex exhibiting a S = 1/2 spin ground state. Magnetisa-
tion data collected in the range of 0–5.0 T at 2 K are con-
sistent with a small spin ground state with the presence of
low-lying excited states of larger multiplicity that become
Fig. 5. The structure of complex 8 viewed perpendicular to the [Fe16
ꢀplaneꢁ.
]
ꢀ
PhCO2 ligands again bridge in their familiar l-mode:
above, below and in the ꢀplaneꢁ of the [Fe16] wheel. The
Hthme2ꢀ ligands are of two types. Eight ligands use their
two deprotonated arms as l2-bridges forming a nearly lin-
ear chain of three metals, with the third protonated arm
bonded only to the central iron center. Four Hthme2ꢀ li-

L.F. Jones et al. / Polyhedron 25 (2006) 325–333
331
populated with increasing field – as indicated by an
inflexion point (at ca. 2.2 T) and the continual rise in mag-
netization with applied field.
1
0.5
0
0.04 K
0.280 T/s
0.140 T/s
0.070 T/s
0.035 T/s
0.017 T/s
0.008 T/s
4.3. Magnetism of [NEt4][Fe11O4(O2CPh)10(thme)4-
(dmhp)2Cl4] (5)
-0.5
-1
Solid state dc magnetization measurements were per-
formed on 5 in the range 2–300 K in a field of 0.2 T be-
tween 300 and 50 K and 0.05 T between 50 and 2.0 K.
(Fig. 6). The room temperature vmT value of approxi-
mately 18.7 cm3 molꢀ1 K decreases to a minimum of
12.4 cm3 molꢀ1 K at 70 K and then rises to a maximum
of 15.4 cm3 molꢀ1 K at 12 K. This indicates that 5 has a
non-zero spin ground state and that the ground spin state
is higher than S = 9/2. In order to obtain the ground state
spin value magnetization data were collected in the ranges
of 2.0–6.0 K and 0.01–5.5 T (Fig. 6). The fits were made
simultaneously on the four different temperatures (2, 3, 4
and 6 K), assuming only the ground state is populated, giv-
-1.2 -0.8 -0.4
0.008 T/s
0
0.4
0.8
1.2
µ0 H (T)
a
1
0.5
0
0.04 K
0.10 K
0.15 K
0.2 K
0.3 K
0.4 K
0.5 K
0.6 K
0.7 K
0.8 K
0.95 K
1.05 K
1.2 K
-0.5
-1
ing a best fit of S = 11/2, g = 2.03 and D = ꢀ0.46 cmꢀ1
.
The maximum value of vmT at low temperature
(15.4 cm3 molꢀ1 K) and the fact that the magnetization
data at 2 and 6 K can be fit with one set of parameters is
a strong indication that the ground state is indeed
S = 11/2 and that the first excited state is rather high in en-
-1.2 -0.8 -0.4
0
0.4
0.8
1.2
b
µ0 H (T)
Fig. 7. Magnetisation (M) of 5 plotted as a fraction of maximum Ms vs.
applied magnetic field (l0H). The loops are shown at different field sweep
rates (top) and different temperatures (bottom).
20
18
16
14
12
10
ergy. The previously reported [Fe11] complex has an S =
1/2 spin ground state [5]. Given the nature of S and sign
of D, measurements were taken to see if 5 acts as a
SMM. Low-temperature (1.2–0.04 K) single crystal mag-
netic measurements were performed on 5 using a micro-
SQUID instrument equipped with three orthogonal field
that allow the magnetic field to be scanned in all directions
[11]. Below 1.2 K hysteresis loops are seen in magnetization
versus field studies whose coercivities increase with decreas-
ing temperature (Fig. 7). A detailed study of the field sweep
rate dependence of the hysteresis loops showed that the
hysteresis at non-zero fields is not due to a phonon bottle-
neck but due to slow relaxation because of the anisotropy
barrier. Slow relaxation is seen at H = 0 but the presence of
strong tunnelling does not allow for a reliable Arrhenius
plot. For all SMMs the barrier is reduced at H = 0 because
of the presence of tunnelling. Even for half-integer spin sys-
tems which in theory should not tunnel do so because of
coupling with the environment – dipolar coupling between
molecules, hyperfine coupling, spin–spin cross relaxation,
and other multibody quantum processes. For a system
with S = 11/2 and D = ꢀ0.46 cmꢀ1 for example, and
assuming an internal transverse field of approximately
10 mT, one can estimate a tunnel splitting of ca. 2.8 ·
10ꢀ6 K and thus a tunnel probability of P = 0.68 for a
sweep rate of 0.1 T sꢀ1 in good agreement with our
0
50 100 150 200 250 300
T(K)
60
50
40
30
20
10
0
0
0.5
1
1.5
2
2.5
3
H/T (Tesla/K)
Fig. 6. Thermal variation of vmT in ranges 300–50 K at 0.2 T and 50–2 K
at 0.05 T (top) and magnetization vs. H/T plot at 2 (h), 3 (s), 4 (n) and 6
(,) K (bottom) for complex 5.

332
L.F. Jones et al. / Polyhedron 25 (2006) 325–333
measurements. This has also been observed in a [Mn4] clus-
ter with S = 9/2 [7].
0.16
0.12
0.08
0.04
0.00
4.4. Magnetism of [Fe8(O2CR)12(thme)4] (6, R = Ph; 7,
R = CMe3) and [Fe16(EtO)4(O2CPh)16(Hthme)12]-
(NO3)4 (8)
[Fe16]
[Fe8]
The magnetic behaviour of compounds 6 and 7 are
essentially identical and thus here we limit the discussion
to complex 6. Magnetic studies of both 6 and 8 indicate
antiferromagnetic exchange between the metal centres.
For both 6 and 8 the value of vmT declines rapidly from
room temperature (ꢁ19, ꢁ28 cm3 K molꢀ1, respectively)
to a value approaching 0 cm3 K molꢀ1 at 1.8 K indicative
of S = 0 ground states for both complexes. The room tem-
perature vmT value is approximately half the expected va-
0
100
200
300
T / K
Fig. 8. Experimental and theoretical (solid lines) v vs. T curves for [Fe16
(squares) and [Fe8] (circles).
]
lue for 8 (ꢁ35 cm3 K molꢀ1) and 16 (ꢁ70 cm3 K molꢀ1
)
non-interacting iron centres, indicating the existence of
strong antiferromagnetic interactions even at room temper-
ature. This is commonly found for iron-oxo clusters with
Fe–O–Fe angles between 100ꢁ and 120ꢁ [12]. Both 6 and
8 can be broken down into simple units of two metal ions
and the bridging ligands that connect them (Scheme 1). In
this scheme, J1, J2 and J3 are the magnetic coupling con-
stants between two FeIII ions for each of the three different
bonding interactions found. These are: one alkoxo and two
carboxylate bridging ligands (J1); two alkoxo and one car-
boxylate bridging ligands (J2); and two alkoxo bridging li-
gands (J3). In order to analyse the magnetic behaviour of 6
and 8 we have found exact analytical equations for the vT
product as a function of the temperature using the interac-
tion topologies shown in Scheme 1. These equations have
been developed following the classical spin approach and
the methodology previously described by Fischer for one-
dimensional systems [13]. This methodology can be applied
since the spin correlation paths generated by the interac-
tion schemes shown in Scheme 1 are large enough to ne-
glect spin auto-correlation contributions through an
entire loop [14]. The obtained analytical law is:
X
vT ¼ vT freeion
ci;j;kuivjwk;
ð1Þ
i;j;k
where u,v and w are the Langevin functions for J1, J2 and
J3, respectively, and ci,j,k are the coefficients of each term.
The best fits of the v versus T curves (Fig. 8) in the 300–
10 K temperature range are obtained with the following
parameters: g = 2.0, J1 = ꢀ22.2 cmꢀ1 and J2 = ꢀ8.5 cmꢀ1
for 6; and g = 2.0, J1 = ꢀ16.0 cmꢀ1, J2 = ꢀ9.1 cmꢀ1 and
J3 = ꢀ74.4 cmꢀ1 for 8.
5. Conclusions
The tripodal ligand H3thme has led to the isolation of a
series of new FeIII clusters. A simple combination of FeCl3,
NaO2CR and H3thme in a mixture of MeCN and MeOH
under ambient temperature and pressure gives rise to a ser-
ies of rhomb- or diamond-like [Fe4] clusters. The identical
[Fe4] units are seen as building blocks in the formation of
related [Fe9] and [Fe11] clusters, appearing once in the
[Fe9] cluster and twice in the [Fe11] cluster. The [Fe9] cluster
was made via the solvothermal reaction of [Fe3O(O2CR)-
(H2O)3]O2CR with H3thme in MeCN – the same reaction
at room temperature produces no isolable products.
R
R
O
O
O
O
R
O
R
O
O
R
R
O
O
R
J2
J3
J1
Fe
Fe
O
Fe
Fe
Fe
Fe
O
R
ꢀ
J1
Repeating this reaction but simply replacing the RCO2
J2
counter ion with NO3 or Clꢀ under the same conditions
ꢀ
J3
produces dramatically different compounds with the for-
mation of octametallic wheel-like clusters. The benzoate
(R = Ph) form of this cluster can also be made under ambi-
ent conditions, albeit in much lower yields and in lower
purity due to a tendency to co-crystallise with unreacted
starting material [Fe3O(O2CPh)6(H2O)3]NO3. The solvo-
thermal reaction gives much higher yield and pure crystal-
line product directly. The pivalate analogue (R = CMe3)
could only be isolated by a solvothermal route. The size
of the wheel formed can be doubled via the introduction
of alcohol to the reaction mixture to give a [Fe16] wheel.
J1
J2
Scheme 1. The different bonding interactions (J1, J2, J3) present in
complexes 6–8.

L.F. Jones et al. / Polyhedron 25 (2006) 325–333
333
(e) M.I. Khan, J. Zubieta, Prog. Inorg. Chem. 43 (1995) 1;
(f) R.C. Finn, J. Zubieta, J. Cluster Sci. 11 (2000) 4;
(g) K. Hegetschweiler, H. Schmalle, H.M. Streit, W. Schneider,
Inorg. Chem. 29 (1990) 3625;
(h) K. Hegetschweiler, H. Schmalle, H.M. Streit, V. Gramllicj, H.U.
Hund, I. Erni, Inorg. Chem. 31 (1992) 1299;
(i) K. Hegetschweiler, D. Gatteschi, A. Cornia, L. Hausherr-Primo,
V. Gramlich, Inorg. Chem. 33 (1996) 4414;
Previously reported cyclic [Fe8] compounds have contained
alkali metal ions within the central cavity of the wheel and
the [Fe16] complex is unprecedented. The above reactions
highlight the very wide range of product cluster nuclearities
and topologies that can result from subtle changes in reac-
tion conditions – for example, changes in counter-ion, or
solvothermal versus lower temperature routes.
(j) C. Boskovic, H.U. Gudel, G. Labat, A. Neels, W. Wernsdorfer,
¨
All of the Fe species display dominant antiferromagnetic
exchange interactions, but in the case of [Fe11] this stabi-
lises a rather large S = 11/2 spin ground state. Negative
zero-field splitting within this ground state gives rise to sin-
gle-molecule magnetism behaviour that is manifested in the
presence of temperature and sweep rate dependent hyster-
esis loops.
B. Moubaraki, K.S. Murray, Inorg. Chem. 44 (2005) 3181;
(k) A. Cornia, A.C. Fabretti, P. Garrisi, C. Mortalo, D. Bonacchi,
D. Gatteschi, R. Sessoli, L. Sorace, W. Wernsdorfer, A.L. Barra,
Angew. Chem., Int. Ed. 43 (2004) 1136.
[4] R.H. Laye, E.L.L. McInnes, Eur. J. Inorg. Chem. (2004) 2811, and
references therein.
[5] (a) R.D. Cannon, R.P. White, Prog. Inorg. Chem. 36 (1988)
195;
(b) W.H. Armstrong, S.J. Lippard, Inorg. Chem. 24 (1985)
981.
[6] J.C. Goodwin, R. Sessoli, D. Gatteschi, W. Wernsdorfer, A.K.
Powell, S.L. Heath, J. Chem. Soc., Dalton Trans. (2000) 1835.
[7] (a) W. Wernsdorfer, S. Bhaduri, C. Boskovic, G. Christou, D.N.
Hendrickson, Phys. Rev. B. 65 (2002) 180403;
Acknowledgements
The authors thank Lloyds of London Tercentenary
Foundation, The Leverhulme Trust and the EPSRC.
(b) W. Wernsdorfer, S. Bhaduri, R. Tiron, D.N. Hendrickson, G.
Christou, Phys. Rev. Lett. 89 (2002) 197201.
References
[8] S.M. Gorun, G.C. Papaefthymiou, R.B. Frankel, S.J. Lippard, J.
Am. Chem. Soc. 109 (1987) 3337.
[1] R. Sessoli, D. Gatteschi, D.N. Hendrickson, G. Christou, MRS Bull.
25 (2000) 66.
[9] S. Perlepes, J.C. Huffman, G. Christou, Chem. Commun. (1991)
1657.
[2] (a) See for example D.N. Hendrickson, G. Christou, H. Ishimoto, J.
Yoo, E.K. Brechin, A. Yamaguchi, E.M. Rumberger, S.M.J. Aubin,
[10] (a) A. Bino, I. Shweky, S. Cohen, E.R. Banninger, S.J. Lippard,
Inorg. Chem. 37 (1998) 5168;
´
(b) W. Schmitt, C.E. Anson, B. Pilawa, A.K. Powell, Z. Anorg. Allg.
Chem. 628 (2002) 2443;
(c) M. Veith, F. Gra¨tz, V. Huch, Eur. J. Inorg. Chem. (2001)
367;
(d) A.K. Boudalis, B. Donnadieu, V. Nastopoulos, J.M. Clemente-
Juan, A. Mari, Y. Sanakis, J.-P. Tuchagues, S.P. Perlepes, Angew.
Chem., Int. Ed. 43 (2004) 2266.
Z. Sun, G. Aromı, Polyhedron 20 (2001) 1479;
(b) C.J. Milios, C.P. Raptopoulou, A. Terzis, F. Lloret, R. Vicente,
S.P. Perlepes, A. Escuer, Angew. Chem. Int. Ed. 43 (2003) 210;
(c) L.M. Wittick, K.S. Murray, B. Moubaraki, S.R. Batten, L.
Spiccia, K.J. Berry, Dalton. Trans. (2004) 1003;
(d) M. Murugesu, M. Habrych, W. Wernsdorfer, K.A. Abboud, G.
Christou, J. Am. Chem. Soc. 126 (2004) 4766;
(e) D.M. Low, L.F. Jones, A. Bell, E.K. Brechin, T. Mallah, E. Riviere,
S.J. Teat, E.J.L. McInnes, Angew. Chem., Int. Ed. 42 (2003) 3781;
(f) A.J. Tasiopoulos, A. Vinslava, W. Wernsdorfer, K.A. Abboud, G.
Christou, Angew. Chem., Int. Ed. 43 (2004) 2117.
[11] W. Wernsdorfer, Adv. Chem. Phys. 118 (2001) 99.
[12] (a) See for example K.L. Taft, S.J. Lippard, J. Am. Chem. Soc. 112
(1990) 9629;
(b) J. van Slageren, R. Sessoli, D. Gatteschi, A.A. Smith, M.
Helliwell, R.E.P. Winpenny, A. Cornia, A.-L. Barra, A.G.M. Jansen,
E. Rentscheler, G.A. Timco, Chem. Eur. J. 8 (2002) 277;
(c) O. Waldmann, R. Kock, S. Schromm, J. Schu¨lein, P. Mu¨ller, I.
Bernt, R.W. Saalfrank, F. Hampel, E. Balthes, Inorg. Chem. 40
(2001) 2986;
[3] (a) See for example S. Piligkos, G. Rajaraman, M. Soler, N.
Kirchner, J. van Slageren, R. Bircher, S. Parsons, H.-U. Gu¨del, J.
Kortus, W. Wernsdorfer, G. Christou, E.K. Brechin, J. Am. Chem.
Soc. 127 (2005) 5572;
(b) G. Rajaraman, M. Murugesu, M. Soler, W. Wernsdorfer, M.
Helliwell, S.J. Teat, G. Christou, E.K. Brechin, J. Am. Chem. Soc.
126 (2004) 15445;
(d) A. Cornia, M. Affronte, A.G.M. Jansen, G.L. Abbati, D.
Gatteschi, Angew. Chem., Int. Ed. 38 (1999) 2264.
[13] M.E. Fisher, Am. J. Phys. 32 (1964) 343.
[14] J. Cano, Y. Journaux, Monte Carlo simulation: a tool to analyse the
magnetic properties, in: J.S. Miller, M. Drillon (Eds.), Magnetism:
Molecules to Materials V, Wiley-VCH, Weinheim, Germany, 2005, p.
189.
´
`
(c) M. Moragues-Canovas, E. Riviere, L. Ricard, C. Paulsen, W.
Wernsdorfer, E.K. Brechin, T. Mallah, Adv. Mater. 13 (2004) 1101;
(d) R. Shaw, I.S. Tidmarsh, R.H. Laye, B. Breeze, M. Helliwell,
E.K. Brechin, S.L. Heath, M. Murrie, S. Ochsenbein, H.-U. Gudel,
¨
E.J.L. McInnes, Chem. Commun. (2004) 1418;